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
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-231581997410.1186/1471-2334-5-23Research ArticleAbacavir, efavirenz, didanosine, with or without hydroxyurea, in HIV-infected adults failing initial nucleoside/protease inhibitor-containing regimens Swindells Susan [email protected] Calvin J [email protected] Daniel S [email protected] Karen T [email protected] Qiming [email protected] Bonnie F [email protected] Jerry W [email protected] Gary E [email protected] Jaime E [email protected] NZTA4008 Study Team 1 HIV Clinic, University of Nebraska Medical Center, Omaha, NE, USA2 Community Research Initiative of New England, Boston, MA, USA3 North Star Medical Center, Chicago, IL, USA4 Infectious Diseases, Miriam Hospital, Providence, RI, USA5 GlaxoSmithKline, Research Triangle Park, NC, USA2005 8 4 2005 5 23 23 3 3 2004 8 4 2005 Copyright © 2005 Swindells et al; licensee BioMed Central Ltd.2005Swindells 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
Hydroxyurea (HU) is an immunomodulatory agent that has been documented to enhance the antiretroviral activity of nucleoside reverse transcriptase inhibitors, such as abacavir (ABC) and didanosine (ddI), and would be expected to improve virologic efficacy.
Methods
A 48-week, phase IV, multicenter, open-label, proof-of-concept clinical trial was conducted to evaluate second-line, protease inhibitor (PI)-sparing therapy with ABC/efavirenz (EFV)/ddI plus HU or without HU in HIV-infected subjects failing to achieve HIV-1 RNA ≤ 400 copies/mL after ≥ 16 weeks of treatment with lamivudine/zidovudine or lamivudine/stavudine, plus 1 or 2 PIs. Subjects were assigned to ABC (300 mg twice daily)/ EFV (600 mg once daily)/ ddI (400 mg once daily) plus HU (500 mg twice daily) (n = 30) or this regimen without HU (n = 24).
Results
Baseline mean HIV-1 RNA was 3.86 log10 copies/mL and CD4+ cell count was 345 cells/mm3. A similar percentage of subjects in the non-HU arm (58%) and HU arm (53%) completed the study. Intent-to-treat: missing = failure analysis showed no differences in proportions of subjects in the non-HU and HU arms achieving undetectable plasma HIV-1 RNA levels at week 24 (<400 copies/mL: 58% [14/24] vs 57% [17/30], P = 0.899; <50 copies/mL (50% [12/24] vs 47% [14/30], P = 0.780). Median change from baseline in CD4+ cell count in the non-HU and HU arms at week 48 was +114 cells/mm3 and -63 cells/mm3 (P = 0.007), respectively. Both regimens were generally well tolerated, although more subjects in the HU arm withdrew prematurely from the study due to adverse events (23% vs 4%). Four cases of possible ABC-related hypersensitivity were observed.
Conclusion
ABC/EFV/ddI was an effective and well-tolerated second-line regimen for nucleoside/PI-experienced HIV-infected subjects. The addition of HU blunted the CD4+ cell response, did not appear to enhance antiviral activity, and resulted in more treatment-limiting adverse events.
==== Body
Background
Combination antiretroviral therapy containing protease inhibitors (PIs) has contributed substantially towards delaying progression of HIV infection and decreasing morbidity and mortality [1,2]. However, PI-based regimens require strict adherence to ensure efficacy, and many of these regimens incur a high pill burden, complex dosing schedules, numerous drug interactions, and metabolic complications including lipodystrophy, hyperlipidemia, and elevated glucose levels. A high rate of virologic failure has been observed with PI-containing regimens in clinical practice [3]. In view of this, PI-sparing regimens using nucleoside reverse transcriptase inhibitors (NRTIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) are attractive for initial and rescue use.
Studies in which the NRTI abacavir (ABC) was substituted for a PI have shown maintenance of virologic suppression and CD4+ cell elevation, with reduction in plasma lipid levels [4-7]. Substitution of the once-daily NNRTI efavirenz (EFV) for PIs has also allowed simplification of therapy with maintenance of virologic suppression [8]. Hirschel et al [9] have shown that when patients are switched from PIs to EFV they experience less virologic failure over a 1-year follow-up period than non-switchers. These results support the use of ABC and EFV in PI-sparing regimens.
Hydroxyurea (HU) is an immunomodulatory agent that depletes intracellular deoxynucleotides by inhibiting ribonucleotide reductase, thereby enhancing the antiretroviral activity of NRTIs and possibly accelerating intracellular phosphorylation [10,11]. Synergy has been shown in vitro between HU and didanosine (ddI) [10], and in vivo (murine AIDS model) between HU and ABC [12]. Regimens employing HU 500 to 1000 mg daily in combination with ddI have been reported to produce marked virologic suppression for up to 2 years [13-23]. Combinations of HU with ABC or EFV have also shown promise as treatment options for antiretroviral-experienced subjects [23,24]. However, to date, limited data have been presented on the therapeutic outcome in HIV-infected subjects treated with both EFV and ABC in HU-containing regimens [24]. The objective of this proof-of-concept study (NZTA4008) was to evaluate the long-term efficacy, safety, and tolerability of second-line therapy with ABC, EFV, ddI, with or without HU, in subjects who had failed their initial antiretroviral regimens.
Methods
Subjects
Male or female subjects 13 years of age or older were eligible for enrollment in the study if they had a screening plasma HIV-1 RNA value between 400 and 100,000 copies/mL, a CD4+ cell count ≥ 100 cells/mm3, had experienced virologic failure (HIV-1 RNA ≥ 400 copies/mL) after at least 16 weeks of initial antiretroviral therapy with 1 or 2 PIs with lamivudine and either zidovudine or stavudine, and had not previously received treatment with non-nucleoside reverse transcriptase inhibitors (NNRTIs), ABC, or ddI. Subjects were eligible if they had a hemoglobin level >9.0 g/dL (for women) or >10.0 g/dL (for men), a neutrophil count >1,000 cells/mm3, a platelet count >75,000 cells/mm3, an estimated creatinine clearance >50 mL/min, a serum lipase < the upper limit of normal, a serum pancreatic amylase level <1.5 times the upper limit of normal, and levels of hepatic aminotransferases < 5 times the upper limit of normal within 2 weeks prior to the baseline visit. Subjects were not eligible for enrollment if they had been on non-suppressive initial antiretroviral therapy for >1.5 years, were pregnant or breastfeeding, were receiving immunomodulating agents, an immunotherapeutic vaccine, or cytotoxic chemotherapeutic agents within 8 weeks before study start, had a history of pancreatitis or peripheral neuropathy within 2 months before study start or had been diagnosed with acute hepatitis within 6 months before study start.
Study design
This was a Phase IV, randomized, open-label, proof-of-concept clinical trial that was conducted at 17 study sites in the United States between September 2, 1999 and April 27, 2001. The study protocol was approved by the institutional review boards at each study site, and all subjects provided written informed consent prior to their participation. Subjects were randomly assigned to receive one of two treatment regimens: ABC 300 mg twice daily, EFV 600 mg once daily, and ddI 400 mg once daily (the non-HU arm); or the ABC/EFV/ddI regimen with HU 500 mg twice daily (the HU arm). This study was planned for 48 weeks, but it concluded after the last subject completed 24 weeks of treatment because accrual was very slow due to the highly selective population sought. In the HU treatment arm, subjects were randomly assigned to start HU either at baseline or at week 8 in order to investigate whether delaying HU could prevent cytopenia and a decrease in CD4+ count.
ABC was supplied as 300-mg tablets of Ziagen® (Glaxo Wellcome, Research Triangle Park, NC), EFV as 200-mg capsules of Sustiva® (DuPont Pharmaceuticals, Wilmington, Delaware), ddI as non-enterically-coated 100-mg, 150-mg, and 200-mg tablets of Videx® (Bristol-Myers Squibb, Princeton, New Jersey), and HU as 200-mg, 300-mg, and 500-mg capsules of Droxia® (Bristol-Myers Squibb, Princeton, New Jersey). Doses of ddI were administered at least 30 minutes before a meal or 2 hours after a meal or snack. The once-daily dose of EFV was administered either in the morning or evening, with or without food. ABC and HU did not have specific dosing requirements regarding timing of doses with respect to meals.
Study procedures
Subjects were evaluated at screening, baseline (day 1), and at weeks 4, 8, 16, 24, 32, 36, 40, and 48. Subjects were evaluated upon premature discontinuation of the study and at 4 weeks post-study discontinuation. For subjects randomized to receive HU beginning at week 8, an additional study visit was scheduled at week 12 to perform safety evaluations. Plasma samples were collected at the study visits for virology, immunology, hematology, and clinical chemistry assessments that were performed by a central laboratory (Consolidated Laboratory Services, Van Nuys, California). Plasma HIV-1 RNA levels were assessed in blood samples at screening and at all study visits using both the Roche AMPLICOR PCR Standard 1.0 assay (lower limit of quantitation [LLOQ] 400 copies/mL) and the Roche PCR assay Amplicor HIV-1 MONITOR UltraSensitive Version 1.0 (LLOQ 50 copies/mL) (both assays from Roche Diagnostics, Branchburg, New Jersey). CD4+ cell counts were determined by flow cytometry. Urinalysis was performed at baseline and at weeks 4, 12, 16, 24, and 48.
Clinical adverse events and laboratory abnormalities were assessed and graded according to the standardized AIDS Clinical Trials Group (ACTG) toxicity grading scales wherever possible (grade 1 or mild to grade 4 or severe). In cases of suspected HU-related Grade 2 or higher hematologic toxicities, hyperamylasemia, or neuropathy, HU was discontinued until the toxicity returned to equal to or less than Grade 2. HU was restarted at a reduced dose and could be subsequently increased to full dose. If amylase levels did not decrease following interruption of ddI, all study medications were to be interrupted. Persistent, recurrent, or Grade 3 or higher hematologic toxicities, hyperamylasemia, or neuropathy required permanent discontinuation of HU. Doses of ddI could be replaced with an alternate NRTI. If pancreatitis was diagnosed, all study drugs were to be permanently discontinued and the subject discontinued from the study, with or without elevated amylase levels. If a subject presented with symptoms consistent with a possible ABC-related hypersensitivity reaction, including fever, skin rash, fatigue, and gastrointestinal symptoms such as nausea, vomiting, diarrhea, or abdominal pain, therapy with ABC was permanently discontinued.
Study endpoints
The primary efficacy endpoint was the proportion of subjects achieving plasma HIV-1 RNA levels <400 copies/mL at weeks 24 and 48. Secondary efficacy endpoints assessed at weeks 24 and 48 included the proportion of subjects achieving plasma HIV-1 RNA <50 copies/mL, change from baseline in HIV-1 RNA and CD4+ cell count, and proportion of subjects with at least a 50-cell increase in CD4+ cell count. During the 48-week study, assessments were made of the time to virologic failure and time to treatment failure. Virologic failure was defined by any of the following: HIV-1 RNA >400 copies/mL by week 24 of randomized treatment or repeated detection (>400 copies/mL) after initial suppression to undetectable levels (<400 copies/mL) or a 3-fold or greater increase in plasma HIV-1 RNA level from the nadir at week 8 or later not attributable to intercurrent infection or vaccination. Treatment failure was defined by one or more of the following: virologic failure, toxicity or other treatment-related withdrawal, or clinical disease progression (from CDC Categories A or B to Category C or death, or progression from Category C to death).
Statistical analysis
Subjects were randomly allocated to the HU and non-HU arms. Within the HU arm, subjects were randomized either to treatment with HU starting at study baseline or to HU starting at 8 weeks post-baseline. Randomization was stratified by screening HIV-1 RNA (<10,000 copies/mL and ≥ 10,000 copies/mL). The total number of subjects originally planned for this study was 150 (80 in the HU arm [40 starting HU at baseline and 40 at 8 weeks]) and 70 in the non-HU arm. However, very slow enrollment necessitated participation by fewer subjects. The proportions of subjects achieving HIV-1 RNA <400 copies/mL and <50 copies/mL at week 24 was calculated using an intent-to-treat: missing = failure (ITT: M=F) analysis. This analysis, which included all randomized subjects, regarded as a treatment failure any subject with missing values or who did not initiate treatment, changed treatment, or prematurely discontinued randomized treatment for any reason. The ITT: M=F analysis was not performed at week 48 because of potential bias due to early termination of the study after the last subject completed 24 weeks. An ITT: observed analysis was performed at both week 24 and week 48, which included all data from subjects seen at each specific visit. Comparisons of proportions were made using the Cochran Mantel Haenzel test. Change from baseline and average area under the curve minus baseline (AAUCMB) in HIV-1 RNA (log10 copies/mL) and change from baseline CD4+ cell count were analyzed using the two-sample t-test and Wilcoxon sum rank test, respectively. Time to virologic failure and time to treatment failure were analyzed using the Kaplan-Meier method and compared between the treatment arms using the log-rank test. The study was not powered to show statistical differences between the treatment arms. Differences were deemed statistically significant if the P-value was <0.05.
Results
Baseline characteristics and subject disposition
Baseline demographic and disease characteristics were similar between the treatment arms (Table 1). Most of the subjects (≥ 87%) were male, and approximately one-half were Caucasian. Baseline median HIV-1 RNA in the non-HU (n = 24) and HU arms (n = 30) was 3.93 and 3.90 log10 copies/mL, respectively, and median CD4+ cell counts were 291 and 326 cells/mm3, respectively. A higher percentage classified CDC Category B was included in the non-HU arm (42% vs 17%), and a higher percentage classified CDC Category A in the HU arm (63% vs 38%); one-fifth of the subjects in each treatment arm were Category C. Prior to the study, the most frequently used NRTIs in each arm had been lamivudine, d4T, and the lamivudine 150 mg/zidovudine 300 mg combination tablet (Combivir®, GlaxoSmithKline, Research Triangle Park, North Carolina), and the most frequently used PIs had been nelfinavir mesylate and indinavir sulfate.
Table 1 Baseline demographics and disease characteristics
Characteristic ABC/EFV/ddI (N = 24) ABC/EFV/ddI/HU (N = 30)
Age, years
Mean ± SD 39.5 ± 7.8 38.1 ± 8.4
Median (Range) 37 (29–62) 37 (26–59)
Gender, n (%)
Male 21 (88) 26 (87)
Female 3 (13) 4 (13)
Race, n (%)
Caucasian 13 (54) 16 (53)
African American 8 (33) 4 (13)
Hispanic 2 (8) 7 (23)
Other 1 (4) 3 (10)
CDC classification, n (%)
Category A 9 (38) 19 (63)
Category B 10 (42) 5 (17)
Category C 5 (21) 6 (20)
HIV-1 RNA, log10 copies/mL
Mean ± SD 3.86 ± 0.55 3.86 ± 0.67
Median (Range) 3.93 (2.86–4.73) 3.90 (2.73–4.82)
CD4 cell count, cells/mm3
Mean ± SD 345 ± 192 346 ± 167
Median (Range) 291 (67–805) 326 (53–794)
Prior antiretroviral treatment, n (%)
NRTIs 24 (100) 29 (97)
Lamivudine 16 (67) 19 (63)
Stavudine 11 (46) 18 (60)
Lamivudine/zidovudine combination tablet 10 (42) 10 (33)
Zidovudine 5 (21) 4 (13)
Zalcitabine 0 1 (3)
PIs 22 (92) 26 (87)
Nelfinavir mesylate 10 (42) 17 (57)
Indinavir sulfate 7 (29) 7 (23)
Saquinavir 3 (13) 2 (7)
Ritonavir 1 (4) 2 (7)
Lopinavir + ritonavir 1 (4) 1 (3)
Amprenavir 1 (4) 0
Premature withdrawal from study, n (%) 10 (42) 14 (47)
Adverse eventa 1 (4) 7 (23)
Consent withdrawn 2 (8) 0
Protocol-defined virologic failure 5 (21) 1 (3)
Lost to follow-up 1 (4) 3 (10)
Protocol violation 0 1 (3)
Other 1 (4) 2 (7)
Note: ABC, abacavir; ddI, didanosine; EFV, efavirenz; HU, hydroxyurea; NRTIs, nucleoside reverse transcriptase inhibitors; PIs, protease inhibitors; SD, standard deviation.
aAdverse events that led to premature study withdrawal in the HU arm were diarrhea, dizziness, headaches, vomiting, rash on chest, insomnia (1 patient); pancreatitis (1); decrease in concentration, exacerbation of depression, nightmares (1); fatigue (1); flushing, fatigue, vomiting, nausea, palpitations (1); dizziness, incoherence (1); and possible ABC-related hypersensitivity reaction (1). The one patient in the non-HU arm who withdrew prematurely from the study did so because of fatal asphyxia, which was not considered related to drug treatment.
Of the subjects in the HU arm, 17 started HU at baseline and 13 at week 8. A similar percentage of subjects in the non-HU arm (58% [14/24]) and HU arm (53% [16/30]) completed the study. The reasons for premature withdrawal from the study are given in Table 1. A greater proportion of subjects in the HU arm withdrew prematurely due to adverse events (23% vs 4%), whereas a greater proportion in the non-HU arm withdrew due to protocol-defined virologic failure.
Virologic response
At week 24, the proportion of subjects in the non-HU and HU treatment arms who achieved HIV-1 RNA <400 copies/mL was not significantly different, according to the ITT: M=F analysis (58% [14/24] and 57% [17/30], respectively; P = 0.899, Fig. 1) and ITT: observed analysis (67% [14/21] and 89% [17/19], respectively; P = 0.081, Fig. 2). Differences also were not observed at week 48, although the proportion of subjects with HIV-1 RNA <400 copies/mL tended to be higher in the non-HU arm in the ITT: observed analysis (91% [10/11] and 80% [8/10], respectively; P = 0.512).
Figure 1 Proportion of subjects in the non-HU and HU arms who achieved plasma HIV-1 RNA <400 copies/mL and <50 copies/mL in the intent-to-treat: missing = failure analyses. HU = hydroxyurea.
Figure 2 Proportion of subjects in the non-HU and HU arms who achieved plasma HIV-1 RNA <400 copies/mL and <50 copies/mL in the intent-to-treat: observed (B) analyses. HU = hydroxyurea.
Results with the 50-copy/mL assay paralleled those with the 400-copy/mL assay. At week 24, the proportion of subjects in the non-HU and HU treatment arms who achieved HIV-1 RNA <50 copies/mL was not significantly different in the ITT: M=F analysis (50% [12/24] and 47% [14/30], respectively; P = 0.780, Fig. 1) or the ITT: observed analysis (57% [12/21] and 74% [14/19], respectively; P = 0.301, Fig. 2). Differences also were not observed at week 48, although the proportion of subjects with HIV-1 RNA <50 copies/mL tended to be higher in the non-HU arm in the ITT: observed analysis (82% [9/11] and 50% [5/10], respectively; P = 0.137). No significant differences between the treatment regimens were observed in proportions of subjects achieving HIV-1 RNA <400 and <50 copies/mL in the protocol-specified subgroups of subjects with baseline HIV-1 RNA >10,000 copies/mL or between 400–10,000 copies/mL (data not shown). Median decrease in HIV-1 RNA from baseline tended to be greater in the HU arm than the non-HU arm at week 24 (-2.10 vs -1.45 log10 copies/mL, P = 0.070), but not week 48 (-2.05 vs -2.12 log10 copies/mL, P = 0.453). Median AAUCMB in HIV-1 RNA in the HU and non-HU arms was not different at week 24 (-1.57 vs -1.38 log10 copies/mL, P = 0.571) or week 48 (-1.66 vs -1.41 log10 copies/mL, P = 0.585).
The time to virologic failure did not differ between the non-HU arm and HU arm (P = 0.808, Fig. 3). In several subjects, a time to virologic failure of 0 was observed due to their lack of virologic response (continued or dropped out) during the first 24 weeks. No statistically significant difference was noted in the time to treatment failure between the two treatment arms (P = 0.418), although the non-HU arm demonstrated slightly longer survival time to treatment failure than the HU arm (Fig. 4).
Figure 3 Time to virologic failure. ABC = abacavir; ddI = didanosine; EFV = efavirenz; HU = hydroxyurea. Virologic failure was defined by any of the following: HIV-1 RNA >400 copies/mL by week 24, repeated detection (>400 copies/mL) after initial suppression to undetectable levels (<400 copies/mL) or a 3-fold or greater increase in plasma HIV-1 RNA level from the nadir at week 8 or later.
Figure 4 Time to treatment failure. ABC = abacavir; ddI = didanosine; EFV = efavirenz; HU = hydroxyurea. Virologic failure was defined by any of the following: HIV-1 RNA >400 copies/mL by week 24, repeated detection (>400 copies/mL) after initial suppression to undetectable levels (<400 copies/mL) or a 3-fold or greater increase in plasma HIV-1 RNA level from the nadir at week 8 or later.
Immunologic response
CD4+ cell counts increased steadily and modestly over 48 weeks in the non-HU arm, whereas they decreased slightly in the HU arm as a whole, fell markedly below baseline in subjects who started HU at the beginning of the study, and increased modestly between Weeks 8 and 40 in subjects who started HU at study week 8 (Fig. 5) Median change from baseline in CD4+ cell count in the non-HU and HU arms was +94 cells/mm3 and -17 cells/mm3 (P = 0.028), respectively, at week 24, and +114 cells/mm3 and -63 cells/mm3 (P = 0.007), respectively, at week 48. At week 48, the median CD4+ cell count was 403 cells/mm3 in the non-HU arm and 249 cells/mm3 in the HU arm (P = 0.168). The proportion of subjects achieving at least a 50-cell increase above baseline in CD4+ cell counts was greater in the non-HU arm at week 24 (55% vs 24%, P = 0.092) and week 48 (73% vs 18%, P = 0.030).
Figure 5 Median change from baseline in CD4+ cell counts in non-HU arm, total HU arm, and HU arm that began HU at baseline and at week 8. BL = baseline; HU = hydroxyurea.
Safety
The number of subjects experiencing adverse events was similar in both treatment arms. Grade 1–4 drug-related adverse events reported by ≥ 5% of subjects are presented in Table 2. The majority of clinical adverse events were gastrointestinal or neurological in nature and mild to moderate in intensity. Gastrointestinal discomfort/pain, nausea, nausea/vomiting, headache, neuropathy, dizziness, and malaise/fatigue were reported more frequently in the HU arm. The incidence of treatment-limiting adverse events was also higher in the HU arm than the non-HU arm (17 [57%] vs 5 [21%]. Events that occurred once included nausea and vomiting (4 subjects), nausea (3), diarrhea (2), malaise/fatigue (2), dizziness (2), headache (2), allergic reaction (2), and skin rashes (2) in the HU arm, and nausea (1), diarrhea (1), allergic reaction to medicinal substance (1), disorders of lipid metabolism (1), and depressive disorders (1) in the non-HU arm (some subjects experienced more than 1 adverse event). Serious adverse events in the non-HU and HU arms included possible ABC-related hypersensitivity reactions (1 and 3, respectively), appendicitis (1 each), diarrhea (1 in HU arm), bronchitis (1 in HU arm), cognitive function disorders (1 in HU arm), pneumonia (1 in non-HU arm), and asphyxia resulting in death (1 in non-HU arm; not considered related to study drugs). There were no differences between the non-HU and HU arms regarding the incidence of Grade 3 or 4 laboratory toxicities, which included abnormalities in triglycerides (3 and 2 subjects, respectively), amylase (1 and 2), creatine kinase (2 and 1), alanine aminotransferase (0 and 1), and glucose (1 and 0). No significant hematologic toxicities were observed. The HU arm had only slightly lower total lymphocyte counts (mean ± SD 27.7 ± 7.6 cells/μL vs 30.3 ± 10.3 cells/μL) and white blood cell counts (5.4 ± 1.7 cells/μL vs 5.9 ± 1.6 cells/μL) than the non-HU arm.
Table 2 Drug-related adverse events reported by ≥ 5% of subjects
Event by body system ABC/EFV/ddI (N = 24) ABC/EFV/ddI/HU (N = 30)
n (%) n (%)
Ear, nose, and throat
Nasal signs and symptoms 2 (8) 0
Endocrine and metabolic
Lack of appetite 1 (4) 3 (10)
Lipid metabolism disorders 2 (8) 2 (7)
Weight problems 0 2 (7)
Gastrointestinal
Abdominal distension 1 (4) 2 (7)
Diarrhea 5 (21) 6 (20)
Gaseous symptoms 0 2 (7)
GI discomfort and pain 0 3 (10)
Nausea 3 (13) 9 (30)
Nausea and vomiting 1 (4) 6 (20)
Musculoskeletal
Arthralgia 1 (4) 2 (7)
Neurology
Abnormal dreams 5 (21) 4 (13)
Cognitive function 2 (8) 1 (3)
disorders 2 (8) 1 (3)
Dizziness 1 (4) 4 (13)
Headache 1 (4) 5 (17)
Memory effects 0 2 (7)
Neuropathy 1 (4) 4 (13)
Sleep disorders 0 2 (7)
Non-site specific
ABC hypersensitivity 1 (4) 3 (10)
Malaise and fatigue 2 (8) 6 (20)
Psychiatric
Depressive disorders 2 (8) 1 (3)
Skin
Nail disorders 0 2 (7)
Skin rashes 3 (13) 3 (10)
Note: ABC, abacavir; ddI, didanosine; EFV, efavirenz; HU, hydroxyurea
Discussion
In this prospective proof-of-concept clinical trial, combination therapy with ABC/EFV/ddI maintained modest suppression of HIV-1 RNA levels and increases in CD4+ cell counts through 48 weeks of therapy in a significant proportion of subjects who previously failed to respond to their initial NRTI/PI-containing antiretroviral regimens. Thus, this study showed that changing from a PI-based regimen to an NNRTI-based regimen for virologic failure is associated with a favorable outcome in some patients. While cross-study comparisons are limited by differences in subject populations, comparisons of the virologic findings of NZTA4008 with those from studies in subjects with similar baseline HIV disease characteristics and antiretroviral experience can suggest the relative therapeutic usefulness of ABC/EFV/ddI in this subject population. Thus, ABC/EFV/ddI resulted in more subjects achieving undetectable HIV-1 RNA levels and a more pronounced CD4+ cell count increase at 48 weeks than has been reported with d4T/ABC/EFV/ddI in antiretroviral-experienced subjects who had failed on PI-containing HAART [24].
The addition of HU to the triple regimen did not enhance virologic suppression with ABC/EFV/ddI over the entire 48-week study period, although in the ITT: observed analysis (but not the ITT: M = F analysis), a tendency for greater suppression in the HU group was observed at 24 weeks (HIV-1 RNA <400 copies/mL: 89% vs 67% [non-HU]; <50 copies/mL: 74% vs 57%; P >0.05). This trend reversed by week 48, with a slightly higher proportion of subjects in the non-HU arm achieving undetectable HIV-1 RNA at that time according to both assays. In contrast, Lafeuillade et al [24] found that when HU 500 mg twice daily was added to an ABC/EFV/ddI/d4T regimen, the significantly enhanced virologic suppression over ABC/EFV/ddI/d4T alone observed at week 24 was maintained at week 48. As HU selectively depletes a greater number of purine, rather than pyrimidine, nucleotides in vitro [10], theoretically greater inhibition of HIV-1 would be expected with purine analog reverse transcriptase inhibitors, such as ddI [25,26]. Where the effect of HU on HAART in antiretroviral-experienced subjects has been compared to the effect in antiretroviral-naïve subjects, a greater HU-potentiating effect on virologic suppression was seen in the -experienced group [21]. It is noteworthy that not all studies of antiretroviral-experienced subjects have shown even short-term enhancement of virologic suppression when HU is added to HAART regimens. Indeed, no change in virologic response was observed by Gonzalez et al [27] in their case-control study in subjects receiving HU-containing HAART (HU dose: 500 mg twice daily) (n = 59) versus non-HU-containing HAART (n = 57) over a median of 18 weeks.
Subsequent to our study, Lori et al [28] showed in RIGHT702 that HU, administered at the low dosage of 600 mg daily (lower than that in our study and most other clinical trials), had a better efficacy and safety profile than that seen with higher dosages. RIGHT702 was a randomized, controlled clinical trial in 115 HIV-infected patients comparing the efficacy and safety of HU at three different daily doses (600, 800–900, or 1200 mg/day) given as once-daily, twice-daily, or three-times-daily regimens with ddI and d4T. A pairwise comparison demonstrated a significantly greater proportion of patients on 600 mg daily than 800–900 mg daily attaining HIV-1 RNA <400 copies/mL at week 24 (primary endpoint) (P = 0.027) and week 48 (P = 0.03), and HIV-1 RNA <50 copies/mL at week 24 (P = 0.013) and week 48 (P = 0.028). HIV-1 RNA area under the plasma concentration-time curve (AUC) at week 24 (P = 0.016) and week 48 (P = 0.001) was also lower in the 600 mg daily groups. The twice-daily dosing interval groups were superior to the once-daily group for all virologic endpoints; however, for the CD4+ count there was a tendency favoring the once-daily dosing. The most efficacious combination of total daily dose and dosing interval for the primary endpoint was HU 300 mg twice daily (P = 0.017). The total daily dose groups and the dosing interval groups were quite comparable with respect to adverse events. However, one case of lethal pancreatitis occurred in the HU 1,200 mg/day group.
Our study evaluated the effect on CD4+ count of delayed HU treatment (until 8 weeks post-baseline) compared to HU treatment initiated at the start of the study. HU had a cytopenic effect on the CD4+ cell count and blunted the CD4+ response to ABC/EFV/ddI. This effect was diminished if the addition of HU was delayed from baseline until week 8. Rutschmann et al [22] previously compared the effect of immediate versus delayed (by 12 weeks) addition of HU 500 mg twice daily to a HAART regimen (ddI plus d4T), but they did not assess comparative CD4+ cell effects. The cytopenic effect of HU on CD4+ cell counts has been well documented in other studies of HU, especially those in which ddI was given concurrently [29]. However, a few studies have reported little or no reduction in CD4+ cell counts, or even increases [30-32]. The cytopenic effect appears to be a dose-related rather than a duration-related phenomenon [26]. A strategy of delaying HU administration until several weeks after initiation of a new HAART regimen may have potential value in the treatment of HIV-infected subjects, especially those with low CD4+ cell counts pre-treatment.
ABC/EFV/ddI was generally well tolerated, with the primary adverse events being GI in nature. The addition of HU to this regimen did not affect the incidence of rashes, depressive disorders, diarrhea, cognitive function disorders, or possible ABC-related hypersensitivity reactions, although its addition did increase the incidence of nausea, nausea/vomiting, lack of appetite, headache, dizziness, neuropathy, and malaise/fatigue. Other studies have shown that HU at a dose of ≥500 mg twice daily produces adverse GI events that may be additive to those associated with other concurrently administered drugs [26]. In combination with ddI, the incidence of peripheral neuropathy can also be expected to rise [33]. The addition of HU 500 mg twice daily decreased the tolerability of the regimen, with a higher proportion of subjects discontinuing due to treatment-limiting toxicities. Other studies in which subjects received HU 1000 mg/day have similarly reported a high dropout rate due to adverse events [34,35]. Thus, Biron et al [35] found that approximately one-quarter of subjects receiving HU/ddI-containing antiretroviral therapy discontinued treatment within 12 months. However, unlike in our study, early withdrawal in these other studies was due primarily to hematologic toxicity (pancytopenia, neutropenia, anemia), elevations in amylase or liver function tests, or pancreatitis. In contrast to other long-term studies that evaluated HU in combination with ddI and d4T, we did not observe a greater incidence of hematologic toxicities in the HU arm compared with the non-HU arm. This may be due in part to the relative lack of myelosuppression with ABC and EFV [36,37].
This study had several limitations in that it comprised a small sample size (especially at week 48: 10 and 11 in the HU and non-HU arms, respectively), included participants with relatively low baseline HIV-1 RNA values, involved differing numbers of subjects in the treatment arms, and had a high withdrawal rate. HU was randomized against a combination that was highly suppressive, and this could be viewed as a formidable situation in which to verify increased efficacy. As HU was added to the 3-drug combination regimen rather than substituted for one of the regimen components, the occurrence of additional adverse events in the HU arm compared to the non-HU arm is not surprising. The use of non-enteric-coated ddI in our study may have been responsible for greater safety concerns than would have been the case had an enteric-coated ddI formulation been given. Indeed, in an in vitro study, Foli et al [38] showed that HU increases mitochondrial toxicity when given with high doses of ddI. Once-daily non-enteric-coated formulation of ddI (which was used in the present study) results in higher maximal blood concentrations compared to the same doses of enteric-coated ddI, thus making mitochondrial toxicity more likely. Our study was also limited because it was not powered to show significant differences between subjects who received HU versus those who did not.
Conclusion
In conclusion, the results of this study showed that in subjects who have failed initial NRTI/PI-containing regimens, ABC/EFV/ddI may result in modest virologic suppression and increases in CD4+ cell counts. Although no additional enhancement of virologic response was seen over 48 weeks when HU was given concurrently with ABC/EFV/ddI, the above mentioned limitations of this study make it difficult to make generalizable conclusions about the ultimate value of HU in a HAART regimen. Recent data reported for HU and ddI suggest that future efficacy/safety studies of HU in HAART regimens should evaluate an HU dosage no greater than 600 mg daily and use an enteric-coated rather than a non-enteric-coated formulation of ddI if the latter drug is to be administered concurrently.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JEH, QL, BFP, and JWS conceived the study design. SS, CJC, DSB, and KTT reviewed and approved the study design. QL provided the statistical methods for the study and performed the statistical analysis of the results. JEH, QL, BFP, and JWS wrote, reviewed and edited the protocol. GEP drafted the manuscript and evaluated hydroxyurea data previously published in antiretroviral studies described in Background and Discussion. SS, CJC, DSB, KTT, JEH, QL, BFP, and JWS reviewed and edited the manuscript. SS, CJC, DSB, and KTT enrolled study subjects. BFP, JWS, and JEH monitored the study. JEH, QL, BFP, and JWS evaluated the clinical data from the study. BFP set up the study at study sites. JEH contributed to secure funding. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Preliminary results of this study were presented in Poster 1918 at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Chicago, Illinois, September 22–25, 2001; Poster 965 at the 39th Annual Meeting of the Infectious Diseases Society of America, San Francisco, California, October 25–28, 2001; and Poster 4587 at the XIV International AIDS Conference, Barcelona, Spain, July 7–12, 2002. This study was funded by GlaxoSmithKline.
The investigators participating in the NZTA4008 study team are as follows:
Stephen Becker, MD, Pacific Horizon Medical Center, San Francisco, CA; Daniel Pearce, DO, St. Luke Medical Group, San Diego, CA; William C. Woodward, DO, Anderson Clinical Research, Reading, PA; David Paar, MD, University of Texas Medical Center, Galveston, TX; Timothy P. Cooley, MD, Boston Medical Center, Boston, MA; Daniel Seekins, MD, St. Joseph's Comprehensive Research Center, Tampa, FL; Joe Eron, MD, University of North Carolina at Chapel Hill, Chapel Hill, NC; Stephen Green, MD, Hampton Roads Medical Specialists, Hampton, VA; Donna Mildvan, MD, Beth Israel Medical Center, New York, NY; Charles Schleupner, MS, MD, New Hanover Regional Medical Center, Wilmington, NC; and Gary Richmond, MD, Broward General Medical Center, Fort Lauderdale, FL. The authors thank all of the patients who participated in this study. In addition, they gratefully acknowledge the study coordinators and research nurses for their care of the study patients and conduct of the study, Denise Wayne for study monitoring, Sandy Griffith for study management, Ilisse Minto for data management, Tom Soeder and Brian Wine for statistical programming, Michael Stevens of Bristol Myers Squibb for providing hydroxyurea and didanosine, Betsy Dusak of DuPont for providing efavirenz, Randall Lanier and Jim Demarest of GlaxoSmithKline for performing the virologic and immunologic analyses, and Belinda Ha for manuscript writing and editorial assistance.
==== Refs
Gulick RM Mellors JW Havlir D Eron JJ Gonzalez C McMahon D Richman DD Valentine FT Jonas L Meibohm A Emini EA Chodakewitz JA Deutsch P Holder D Schleif WA Condra JH Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy N Engl J Med 1997 337 734 739 9287228 10.1056/NEJM199709113371102
Hammer SM Squires KE Hughes MD Grimes JM Demeter LM Currier JS Eron JJ Feinberg JE Balfour HH Deyton LR Chodakewitz JA Fischl MA Phair JP Pedneault L Nguyen B-Y Cook JC for The AIDS Clinical Trials Group 320 Study Team A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less N Engl J Med 1997 337 725 733 9287227 10.1056/NEJM199709113371101
Bartlett JA DeMasi R Quinn J Moxham C Rousseau F Overview of the effectiveness of triple combination therapy in antiretroviral-naïve HIV-1 infected adults AIDS 2001 15 1369 1377 11504958 10.1097/00002030-200107270-00006
Opravil M Hirschel B Lazzarin A Furrer H Chave J-P Yerly S Bisset LR Fischer M Vernazza P Bernasconi E Battegay M Ledergerber B Günthard H Howe C Weber R Perrin L for the Swiss HIV Cohort Study A randomized trial of simplified mantenance therapy with abacavir, lamivudine, and zidovudine in human immunodeficiency virus infection J Infect Dis 2002 185 1251 1260 12001042 10.1086/340312
Katlama C Fenske S Gazzard B Lazzarin A Clumeck N Mallolas J Lafeuillade A Mamet J-P Beauvais L on behalf of the AZL30002 European study team TRIZAL study: switching from successful HAART to Trizivir™ (abacavir-lamivudine-zidovudine combination tablet): 48 weeks efficacy, safety and adherence results HIV Med 2003 4 79 86 12702127 10.1046/j.1468-1293.2003.00139.x
Clumeck N Goebel F Rozenbaum W Gerstoft J Staszewski S Montaner J Johnson M Gazzard B Stone C Athisegaran R Moore S on behalf of the CNA30017 Study Team Simplification with abacavir-based triple nucleoside therapy versus continued protease inhibitor-based highly active antiretroviral therapy in HIV-1-infected patients with undetectable plasma HIV-1 RNA AIDS 2001 15 1517 1526 11504984 10.1097/00002030-200108170-00009
Pulvirenti J Goodwin D Slater L Rodriguez A Fleming J Williams V Kauf T Hernandez J Simplification of protease inhibitor (PI)-containing HAART regimens with abacavir (ABC) maintains viral suppression and favorable adherence in HIV-1-infected adults (COL30305) 39th Annual Meeting of the Infectious Diseases Society of America, San Francisco, California Poster 689. October 25–28, 2001
Staszewski S Morales-Ramirez J Tashima KT Rachlis A Skiest D Stanford J Stryker R Johnson P Labriola DF Farina D Manion DJ Ruiz NM for The Study 006 Team Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults N Engl J Med 1999 341 1865 1873 10601505 10.1056/NEJM199912163412501
Hirschel B Flepp M Bucher H Heiner C Zellweger C Telenti A Wagels T Bernasconi E Ledergerber B the Swiss HIV Cohort Switching from protease inhibitors to efavirenz: difference in efficacy and tolerance among risk groups: a case:control study from the Swiss HIV Cohort AIDS 2002 16 381 385 11834949 10.1097/00002030-200202150-00009
Lori F Malykh A Cara A Sun D Weinstein JN Lisziewicz J Gallo RC Hydroxyurea as an inhibitor of human immunodeficiency virus-type 1 replication Science 1994 266 801 805 7973634
Bianchi V Borrella S Calderazzo F Ferraro P Chieco Bianchi L Reichard P Inhibition of ribonucleotide reductase by 2-substituted deoxycytidine analogs: possible application in AIDS treatment Proc Natl Acad Sci USA 1994 91 8403 8407 8078894
Sumpter LR Inayat MS Elford HL Gallicchio VS Novel ribonucleotide reductase inhibitors (RRIs) trimidox (TX), didox (DX) and hydroxyurea (HU) enhance the anti-viral activity of abacavir (ABC) in vivo XIV International AIDS Conference, Barcelona, Spain Abstract A10131. July 7–12, 2002
Foli A Lori F Maserati R Tinelli C Minoli L Lisziewicz J Hydroxyurea and didanosine is a more potent combination than hydroxyurea and zidovudine Antivir Ther 1997 2 31 38 11322264
Lori F Rosenberg E Lieberman J Foli A Maserati R Seminari E Alberici F Walker B Lisziewicz J Hydroxyurea and didanosine long-term treatment prevents HIV breakthrough and normalizes immune parameters AIDS Res Hum Retrovir 1999 15 1333 1338 10515148 10.1089/088922299310034
Rodriguez CG Vila J Capurro AF Maidana MM Boffo Lissin LD Combination therapy with hydroxyurea versus without hydroxyurea as first line treatment options for antiretroviral-naïve patients HIV Clin Trials 2000 1 1 8 11590492
Vila J Biron F Nugier F Vallet T Peyramond D 1-Year follow-up of the use of hydroxycarbamide and didanosine in HIV infection Lancet 1996 348 203 204 8684186 10.1016/S0140-6736(05)66157-0
Biron F Lucht F Peyramond D Fresard A Vallet T Nugier F Grange J Malley S Hamedi-Sangsari F Vila J Anti-HIV activity of the combination of didanosine and hydroxyurea in HIV-1-infected individuals J Acquir Immune Defic Syndr Hum Retrovirol 1995 10 36 40 7648282
Hellinger JA Iwane MK Smith JJ Fleishman AN Torres RA Schrader S Perez G Cohen CJ Skowron G Giordano MF Accetta G Cooper EC Frost KR A randomized study of the safety and antiretroviral activity of hydroxyurea combined with didanosine in persons infected with human immunodeficiency virus type 1 J Infect Dis 2000 181 540 547 10669337 10.1086/315231
Goujard C Boue F Rousseau C Mamet JP Antoun Z Delfrasssy JF A pilot study evaluating tolerance and efficacy of abacavir (ABC), amprenavir (APV), didanosine (DDI) and hydroxyurea (HU) in pretreated HIV-1 infected adults: CNAF3011 7th European Conference on Clinical Aspects and Treatment of HIV Infection Lisbon, Portugal Poster 461. October 23–27, 1999
Arrizabalaga J Iribarren JA Pinilla J Rodriguez-Arrondo FJ von Wichman MA Labarga P Riera M Aliaga L Aguirrebengoa K Prospective, multicenter study of ddI plus hydroxyurea (HU) plus efavirenz (EFV) plus protease inhibitor (PI) salvage therapy. One-year follow-up. Correlation of viral outcome and genotypic mutations XIII International AIDS Conference, Durban, South Africa Poster WePeB4164. July 9–14, 2000
Johnson V Delaugerre C Hazelwood J Berzins B Belsey E Rhodes R Calvez V Katlama C Murphy R for the International 3d Study Team Viral resistance results from the international, multicenter, randomized, placebo-controlled, 48-week study of hydroxyurea (HU) or placebo combined with efavirenz (EFV), didanosine (ddl) and stavudine (d4T) in treatment-naive and – experienced patients 8th Conference on Retroviruses and Opportunistic Infections, Chicago, Illinois Abstract 445. February 4–8, 2001
Rutschmann OT Vernazza PL Bucher HC Opravil M Ledergerber B Telenti A Malinverni R Bernasconi E Fagard C Leduc D Perrin L Hirschel B the Swiss HIV Cohort Study Long-term hydroxyurea in combination with didanosine and stavudine for the treatment of HIV-1 infection. Swiss HIV Cohort Study AIDS 2000 14 2145 2151 11061656 10.1097/00002030-200009290-00011
Zala C Salomon H Ochoa C Kijak G Federico A Perez H Montaner JSG Cahn P Higher rate of toxicity with no increased efficacy when hydroxyurea is added to a regimen of stavudine plus didanosine and nevirapine in primary HIV infection J Acquir Immune Defic Syndr 2002 29 368 673 11917241
Lafeuillade A Hittinger G Chadapaud S Maillefet S Rieu A Poggi C The HYDILE trial: efficacy and tolerance of a quadruple combination of reverse transcriptase inhibitors versus the same regimen plus hydroxyurea or hydroxyurea and interleukin-2 in HIV-infected patients failing protease inhibitor-based combinations HIV Clinical Trials 2002 3 263 271 12187499
Lori F Gallo RC Hydroxyurea and AIDS: an old drug finds new application? AIDS Res Hum Retrovir 1995 11 1149 1151 8573369
Gibbs MA Sorensen SJ Hydroxyurea in the treatment of HIV-1 Ann Pharmacother 2000 34 89 93 10669190 10.1345/aph.19004
Gonzalez OY Jovanovich JF Mayers DL Musher DM The lack of therapeutic benefit of adding hydroxyurea to highly active antiretroviral therapy (HAART) J Acquir Immune Defic Syndr 2002 30 363 364 Letter 12131575
Lori F Pollard R Shalit P Blick G Peterson D Lisziewicz J A low hydroxyurea dose (600 mg daily) is found optimal in a randomized, controlled trial (RIGHT 702) XIVth International AIDS Conference, Barcelona, Spain Abstract TuPeB4465. July 6–12, 2002
Romanelli F Pomeroy C Smith KM Hydroxyurea to inhibit human immunodeficiency virus-1 replication Pharmacotherapy 1999 19 196 204 10030769 10.1592/phco.19.3.196.30913
Rossero R Asmuth DM Grady JJ McKinsey DS Green S Andron L Pollard RB Hydroxyurea in combination with didanosine and stavudine in antiretroviral-experienced HIV-infected subjects with a review of the literature Intl J STD & AIDS 2003 14 350 355 10.1258/095646203321605576
Zala C Salomon H Gun A Rouleau D Pampurro S Perez H Cahn P Montaner JSG A pilot randomized comparative study of stavudine, didanosine, nevirapine with or without hydroxyurea in primary HIV infection 6th Conference on Retroviruses and Opportunistic Infections, Chicago, Illinois Abstract 399. February 1–5, 1999
Youle M Mocroft A Johnson M Tyrer M Madge S Dykhoff A Drinkwater A Surrogate marker responses to multidrug combinations comprising hydroxyurea, efavirenz, double protease inhibitors, and nucleoside analogues in protease inhibitor failures 6th Conference on Retroviruses and Opportunistic Infections, Chicago, Illinois Abstract. February 1–5, 1999
Moyle GJ Sadler M Peripheral neuropathy with nucleoside antiretrovirals. Risk factors, incidence and management Drug Safety 1998 19 481 494 9880091
Havlir DV Gilbert PB Bennett K Collier AC Hirschff MS Tebas P Adams EM Wheat LJ Goodwin D Schnittman S Holohan MK Richman DD the ACTG 5025 Study Team Effects of treatment intensification with hydroxyurea in HIV-infected patients with virologic suppression AIDS 2001 15 1379 1388 11504959 10.1097/00002030-200107270-00007
Biron F Ponceau B Bouhour D Boibieux A Verrier B Peyramond D Long-term safety and antiretroviral activity of hydroxyurea and didanosine in HIV-infected patients J Acquir Immune Defic Syndr 2000 25 329 336 11114833 10.1097/00042560-200012010-00006
Hervey PS Perry CM Abacavir – a review of its clinical potential in patients with HIV infection Drugs 2000 60 447 479 10983741
Adkins JC Noble S Efavirenz Drugs 1998 56 1371 1379
Foli A Benvenuto F Piccinini G Bareggi A Cossarizza A Lisziewicz J Lori F Direct analysis of mitochondrial toxicity of antiretroviral drugs AIDS 2001 15 1687 1694 11546944 10.1097/00002030-200109070-00012
| 15819974 | PMC1090576 | CC BY | 2021-01-04 16:28:14 | no | BMC Infect Dis. 2005 Apr 8; 5:23 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-23 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-251581998510.1186/1471-2334-5-25Research ArticleEffective antiprotease-antibiotic treatment of experimental anthrax Popov Serguei G [email protected] Taissia G [email protected] Svetlana [email protected] Raymond S [email protected] Rebecca [email protected] Karl J [email protected] Vikas [email protected] Charles [email protected] Ken [email protected] Advanced Biosystems, Inc., Manassas, VA, USA2 National Center for Biodefense, George Mason University, Manassas, VA, USA3 Potomac Hospital, Woodbridge, VA, USA4 Center for Biomedical Genomics & Informatics, Department of Molecular & Microbiology, George Mason University, Manassas, VA, USA5 Current affiliation: National Center for Biodefense, George Mason University, Manassas, VA, USA2005 8 4 2005 5 25 25 6 1 2005 8 4 2005 Copyright © 2005 Popov et al; licensee BioMed Central Ltd.2005Popov 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
Inhalation anthrax is characterized by a systemic spread of the challenge agent, Bacillus anthracis. It causes severe damage, including multiple hemorrhagic lesions, to host tissues and organs. It is widely believed that anthrax lethal toxin secreted by proliferating bacteria is a major cause of death, however, the pathology of intoxication in experimental animals is drastically different from that found during the infectious process. In order to close a gap between our understanding of anthrax molecular pathology and the most prominent clinical features of the infectious process we undertook bioinformatic and experimental analyses of potential proteolytic virulence factors of B. anthracis distinct from lethal toxin.
Methods
Secreted proteins (other than lethal and edema toxins) produced by B. anthracis were tested for tissue-damaging activity and toxicity in mice. Chemical protease inhibitors and rabbit immune sera raised against B. anthracis proteases were used to treat mice challenged with B. anthracis (Sterne) spores.
Results
B. anthracis strain delta Ames (pXO1-, pXO2-) producing no lethal and edema toxins secrets a number of metalloprotease virulence factors upon cultivation under aerobic conditions, including those with hemorrhagic, caseinolytic and collagenolytic activities, belonging to M4 and M9 thermolysin and bacterial collagenase families, respectively.
These factors are directly toxic to DBA/2 mice upon intratracheal administration at 0.5 mg/kg and higher doses. Chemical protease inhibitors (phosphoramidon and 1, 10-phenanthroline), as well as immune sera against M4 and M9 proteases of B. anthracis, were used to treat mice challenged with B. anthracis (Sterne) spores. These substances demonstrate a substantial protective efficacy in combination with ciprofloxacin therapy initiated as late as 48 h post spore challenge, compared to the antibiotic alone.
Conclusion
Secreted proteolytic enzymes are important pathogenic factors of B. anthrasis, which can be considered as effective therapeutic targets in the development of anthrax treatment and prophylactic approaches complementing anti-lethal toxin therapy.
==== Body
Background
Inhalation anthrax is a severe, often fatal disease characterized by systemic spread of the challenge agent, Bacillus anthracis, which is capable of causing severe damage to host tissues and organs. Multiple hemorrhagic lesions in the mediastinum, mediastinal lymph nodes, bronchi, lungs, heart, spleen, liver, intestines, kidneys, adrenal glands, and/or central nervous system are typically found upon postmortem examination of patients who succumbed to inhalation anthrax. The most dramatic and potentially life-threatening changes were observed in the vascular system with a diffuse vasculitis extending from moderate sized arteries and veins down to the capillary level. The vasculitis was often associated with vessel destruction, especially of the smallest vessels, and was typically accompanied by massive necrosis in some tissues [1-3]. It is widely believed that anthrax lethal toxin (LT) secreted by proliferating bacteria is a major cause of death in man and in several other susceptible animal species [4]. However, the pathology of intoxication in experimental animals is drastically different from that found during the natural infectious process. Recent extensive analyses in mice and rats challenged with a highly purified lethal toxin [5,6] confirmed earlier observations [7] that toxin activity caused no gross pathology and almost solely manifested in hypoxic liver failure. In addition to lethal toxin, the hemorrhagic and other tissue-damaging factors elaborated by B. anthracis could play important virulence-enhancing roles but these factors have not yet been characterized. Early publications using culture filtrates of B. anthracis assumed that the observed effects of secreted substances were caused entirely by LT [8,9]. It is also possible that these other factors could themselves exhibit a direct lethal effect.
The capacity of bacteria to cause destruction of tissues, as well as other pathological consequences such as degradation of immunoglobulins, cytokines and complement, release of inflammatory mediators, or activation of host proteolytic enzymes, is attributed to a wide variety of secreted proteases [10], however in the case of B. anthracis the proteases of this microorganism, other than lethal factor, have attracted little attention in the scientific literature. The current study aimed to carry out an initial characterization of certain B. anthracis proteolytic enzymes and to obtain evidence on their possible pathogenic role in anthrax.
Methods
Microbial strains
The non-encapsulated Bacillus anthracis strain 34F2 (Sterne) [pXO1+, pXO2-] obtained from the Colorado Serum Company (Boulder, CO) was used in animal challenge experiments. The 50% lethal doses (LD50s) by the intraperitoneal (i.p.) route were established earlier (Popov et al., 2004) and the LD50 value for intraperitoneal challenge for DBA/2 mice was found to be 3 × 106 spores per mouse. The non-encapsulated, atoxigenic strain of B. anthracis (delta Ames) [pXO1-, pXO2-] was kindly provided by Dr. J. Shiloach (National Institutes of Health, Bethesda, MD). B. cereus strain ATCC #11778 and B. subtilis strain #23857 were purchased from American Type Culture Collection (Manassas, VA).
Mice
Female DBA/2 mice (9 weeks old) were obtained from Taconic (Germantown, NY) and were used throughout the study.
Reagents
The following substances were used in this study: ciprofloxacin (ICN Biomedicals, lot no.4913F), phosphoramidon disodium salt, and 1,10-phenanthroline (Sigma, MO), EDTA (GibcoBRL, CA), soybean trypsin inhibitor from Glycine max (Sigma, MO), thermolysin (EC 3.4.24.27) from Bacillus thermoproteolyticus (Sigma, MO). The fluorescently labeled casein and collagen type I for determination of proteolytic acivity were from Molecular Probes (OR). Zymogram gels were from Invitrogen (Carlsbad, CA). Lethal factor (LF) and protective antigen (PA) were from List Biological Laboratories (CA).
Preparation of secreted proteins
Secreted substances were prepared by culturing B. anthracis (delta Ames) in LB media overnight. Cells were removed by centrifugation at 8000 g, and the supernatant was sterilized by filtration through 0.22 μm cellulose acetate filtration system (Corning, NY) and further concentrated 50-fold using Amicon Ultra15 centrifugal filter devices (10 K cut-off pore size) (Millipore, MA). The proteins were used immediately after preparation or were stored at 4°C for several days. Protein content was determined using Bradford reagent (Bio-Rad) with bovine serum albumin as standard. Slow reduction in the hemorrhagic activity was found upon storage within a week.
Fractionation of culture supernatants
1 ml of B. anthracis culture supernatant (BACS) was loaded onto the size-exclusion Superdex column (25 × 60, Pharmacia Biotech) and was eluted with PBS (pH 7.4) with a flow rate of 2 ml/min. Fractions of eluate were concentrated to equal volumes using Centricon devices (Millipore, MA) with a 10 K cut-off pore size.
Hemorrhages in femoral artery region
Mice were anesthetized by intraperitoneal injection of Avertin (2,2,2 tribromethanol, Aldrich) and 100 μl of secreted proteins (20 to 100 μg) were subcutaneously (s.c.) injected into the femoral artery region for observation of hemorrhagic changes after 3 to 15 h. In order to record hemorrhagic changes animals were anesthetized by i.p. injection of Avertin and the fur over the femoral artery region was removed to allow observation of a 1.5 to 2.5 cm2 area of skin. It was photographed, and the size of the hemorrhagic spot was measured. In the experiments on the inhibition of hemorrhagic effect the secreted proteins were preincubated with specific antisera or protease inhibitors for 30 min on ice.
Generation of antibodies against B. anthracis MPs
The Invitrogen (CA) custom service was used to obtain rabbit polyclonal sera against peptides conjugated with kallikrein (Table 1). Two animals were immunized by each conjugate. All six rabbit sera had ELISA titers ranging from 100,000 to 200,000. For generation of murine polyclonal antibodies against the M4 protease (BA3442) the C-terminal part of the gene encoding amino acids 248 to 532 was cloned into pTrcHis2 TOPO TA cloning vector (Invitrogen, CA). The recombinant protein containing a 6 × His tag was expressed in E. coli and purified using the Ni-NTA resin (Quiagen, CA). Mice were immunized with 50 μg of the protein emulsified in a complete Freund's adjuvant and were given two booster immunizations using an incomplete adjuvant with 2 week intervals. Serum was collected after two weeks since the last boost injection. In the skin hemorrhagic test described above, 30 μl of serum was able to completely suppress the hemorrhage caused by 30 μl of BACS.
Table 1 Sera against B. anthracis proteases
Serum # Protease family Protein Gene number Antigen Designation
1 M4 Elastase-like neutral protease BA3442 Recombinant polypeptide corresponding to the fragment 248–532. M4EL
2 M9 Collagenase BA0555, BA3299, BA3584 HEFTHYLQGRYEVPGL spanning the region of active center M9Coll
3 M4 Neutral protease BA5282, BA0599 DVIGHELTHAVTE spanning the region of active center M4AC
4 M4 Neutral protease BA2730 ADYTRGQGIETY distant from the active center M4EP
Intratracheal delivery of B. anthracis secreted proteins
Mice were anesthetized by i.p. injection of Avertin and a 24G angiogenic catheter (BD Biosciences, CA) was inserted into the trachea. 50 μl of experimental mixture, containing 10 to 100 μg of culture supernatant proteins were slowly injected through the catheter connected to a microsyringe. The angiogenic catheter was removed and animals were left for further observation. The untreated control group received the same volume of phosphate-buffered saline (PBS). A control group of 3 animals was injected with 50 μl of PBS solution of lethal toxin (100 μg PA+100 μg LF). In all experiments the rate of breathing was recorded every 10 min during the first 3 h following injection, and animals were observed for survival for 7 days.
Detection of proteolytic activity of culture supernatants
Proteolitic activities of supernatants were measured using EnzChek Protease assay kit (Molecular Probes) and EnzChek Gelatinase/Collagenase assay kit (Molecular Probes) according to the manufacturer's protocols. Briefly, 5 μl of different dilutions of supernatant in 45 μl of digestion buffer were mixed with 50 μl of fluorescein-labeled substrate (casein or gelatin) and fluorescence intensity was measured at different time points using 485 nm excitation and 530 nm emission wavelengths.
Treatment of spore-challenged mice
Mice used in all experiments were maintained under proper conditions with a 12-h light/dark cycle in accordance with IACUC standards in the animal facility of the Biocon, Inc. (Rockville, MD). Mice received food and water ad libitum. Groups of 10 mice were randomly assigned for challenge and were observed for survival and signs of disease. The animals were inoculated i.p. by 1 × 107 spores per mouse of Sterne strain. Treatment (i.p.) with phenanthroline (30 mg/kg), phosphoramidon (10 mg/kg), or rabbit sera (5 or 25 mg/kg) was carried out individually for each substance or in combination with ciprofloxacin at (50 mg/kg) once a day started at different time points post spore challenge and continued for 10 days. In all experiments the animals were monitored for survival for at least 12 days after termination of treatment.
Statistical analysis
Kaplan-Meier open-end survival analysis was performed to compare results between treatment groups. Statistical significance was established as P < 0.05 using log-rank test.
Results
Genomic analysis of B. anthracis secreted proteins as potential virulence factors
In order to evaluate a pathogenic potential attributed to the B. anthracis proteins other than known lethal and edema toxins we used a nontoxigenic and nonencapsulated strain of B. anthracis (delta Ames), which is a parental Ames strain cured of both plasmids, pXO1 and pXO2. The substances secreted by vegetative B. anthracis cells seem to be the most promising candidates, as is the case for many bacterial toxins [10]. Analysis of the chromosome sequence of the B. anthracis Ames strain revealed a variety of potential virulence-enhancing factors, including collagenases, phospholipases, hemolysins, proteases and other enterotoxins identified based on their sequence homology with pathogenic factors in other bacterial species [11]. The B. cereus group, which includes B. anthracis, B. thuringiensis and B. cereus, has an expanded number of predicted secreted proteins relative to nonpathogenic B. subtilis [11]. These B. cereus group-specific genes represent the ancestral adaptations to a pathogenic lifestyle by the common ancestor, which was quite similar to B. cereus. Our attention was attracted to the group of proteases that are encoded on the B. anthracis chromosome, shared in common with B. cereus but absent or relatively rare in the genomes of nonpathogenic bacteria, such as B. subtilis and B. halodurans. A large number of these proteases fall into clan MA (classified according to the MEROPS system [12]), which among others includes thermolysin-like enzymes of the M4 family. Metalloproteases (MPs) from several bacterial species belonging to this family are capable of causing massive internal hemorrhages and other life-threatening pathologies [10,13-16].
Whole genome analyses also indicated collagenolytic proteases of the M9B family as potentially having pathogenic functions. Eleven protease families are present in B. anthracis and B. cereus but absent in B. subtilis. Six of the eleven subfamilies encode MPs. Three of the latter, namely M6, M9B, and M20C subfamilies, are encoded on the bacterial chromosomes. Members of the M6 peptidase family are usually annotated as "immune inhibitors" because in B. thuringiensis they can inhibit the insect antibacterial response [17]. The M20C peptidase subfamily represents exopeptidases [18] that are the unlikely cause of tissue destruction or internal bleeding. Based on the above analysis, this study focused on the M4 family thermolysin/elastase-like neutral proteases and the M9 family collagenases as the candidate virulence-enhancing factors of B. anthracis Aimes strain.
Hemorrhagic, caseinolytic and gelatinolytic activities of anthrax proteases
The proteins secreted by three Bacillus species (B. anthracis, B. cereus and B. subtilis) into culture media were prepared by successive steps of inoculation of the culture media with spores, overnight incubation at 37°C, removal of bacterial cells by centrifugation, sterilization of the supernatant by filtration through 0.22 μ filter and further 50-fold concentration using ultrafiltration devices Amicon Ultra 15 (Millipore, MA) with a 10 KDa cutoff size. The SDS-PAGE gel separation (Fig. 1A) demonstrates the protein content in the concentrated B. anthracis culture supernatant (designated as BACS) used in our animal tests. Similar procedures were used to prepare culture supernatants for B. cereus ATCC #11778 and B. subtilis ATCC #23857 (designated BCCS and BSCS, respectively). Proteolytic activities of BACS are readily detected by zymography using casein or gelatin (denatured collagen) (Fig. 1D). A major band of gelatinase activity corresponds to molecular mass of about 100 KDa, whereas a collagenase activity is represented by about 55 KDa proteins.
Figure 1 SDS-PAGE of BACS fractions separated on size exclusion column (A) and Western blots: fractions (A) with specific antisera a-M4EL (B), BACS with a-M4AC (C, left lane), BACS with a-M4EP (C, right lane), BACS with a-M9Coll (D, right lane), and zymograms of caseinolytic and gelatinolytic activities of BACS (D, left and center lanes, correspondingly). Molecular masses (KDa) of the marker proteins are indicated by arrows. In A, s denoted BACS, and numbers above correspond to column fractions. In zymogram gels 3 μl of BACS were loaded.
The concentrated culture supernatants were tested in mice. Upon subcutaneous administration, mice developed hemorrhages of different intensity within several hours (Fig. 2A, B). BCCS showed the highest activity followed by BACS, while BSCS was completely inactive. Chemical inhibitors such as phosphoramidon (potent chelating inhibitor of thermolysin and other M4 bacterial metallo-endopeptidases [19]), EDTA (specific for a broad range of MPs) and soybean trypsin inhibitor (SBTI, reversible competitive inhibitor of trypsin and other trypsin-like proteases such as chymotrypsin, plasmin and plasma kallikrein [10]) effectively abrogated the hemorrhagic affect of BACS (Fig. 2C). The murine serum raised against the recombinant protein corresponding to the mature form of the M4-type thermolysin-like neutral protease of B. anthracis (gene identification number, BA 3442 according to [11]) was also effective in suppressing the hemorrhagic effect in the skin test. In negative control experiments, neither naïve murine serum nor three irrelevant murine sera against B. anthracis candidate pathogenic factors, hemolysins O, A and B [20] showed anti-hemorrhagic activity (data not shown). Additional control experiments demonstrated that under the conditions of our test the hemorrhagic activity of thermolysin from B. thermophilicus was detectable in a dose range from 10 to 100 μg, similar to that for BACS. In contrast to BACS, the inhibitors displayed only partial protection in the case of BCCS (Fig. 2C). Overall, these results correlate with the experimental data that culture supernatants obtained from fully-virulent toxigenic strain of B. anthracis were less toxic to mice compared to B. cereus ones [9,21].
Figure 2 Hemorrhagic activity of culture supernatants (A), its graphic representation (B) and inhibition with chemical inhibitors (C). 100 μl of secreted proteins (20 to 100 μg) were subcutaneously (sc) injected into the femoral artery region for observation of hemorrhagic changes. The fur over the femoral artery region was removed to allow observation of a 1.5 to 2.5 cm2 area of skin. Control mice were administered equal amount of phosphate-buffered saline. Error bars indicate standard deviations.
Generation of antibodies against B. anthracis MPs
Obvious complexity of the BACS protein composition prompted us to develop specific means of detection and inhibition of its components. For this purpose several high-titer immune sera were raised in mice and rabbits using the antigens listed in Table 1. The sera were used in Western blots of BACS proteins. When the proteins were directly separated in the SDS-PAGE for subsequent transfer to the nitrocellulose membrane, the resulting blots were of low intensity indicative of proteolytic degradation during the electrophoresis (Fig. 1A and 1B, left lanes). In order to avoid this complication the BACS was fractionated according to the molecular masses of its components on the Superdex size exclusion column in the presence of EDTA as a chelating agent. Analysis of the column fractions in SDS-PAGE showed a complex pattern of proteins bands (Fig. 1). Multiple proteins with a broad spectrum of molecular masses seem to be highly associated and migrate through the column as high molecular mass complexes. Several factors, such as the presence of multiple precursor and mature protein forms resulting from specific proteolytic maturation, along with nonspecific proteolytic products, can potentially contribute to the complexity of the fractions' composition. Western blot experiments with column fractions revealed several discrete bands recognized by antibodies (Fig. 1). The M4 proteases are represented by several major bands at about 50 KDa, as well as by the bands at about 40 and 20 KDa. These bands probably correspond to different maturation forms of proteases, including the enzymes lacking signal peptides, and mature enzyme forms. The M9 collagenases are detected as a major band with a molecular mass of about 98 kDa which is close to the estimated mass of the pro-enzymes, however the major gelatinase enzymatic activity corresponds to the 55 kDa proteins in the BACS.
Acute toxicity of B. anthracis culture supernatants
Although bacterial proteases are well known pathogenic factors, little information is available regarding their acute toxicity. We tested BACS in mice using intratracheal administration into the lungs because hemorrhagic mediastinitis and lung edema typically precede the lethal outcome in late anthrax. Therefore, lung damage may be considered as a probable death-causing factor. Mice were given different doses of BACS (10 μg to 40 μg of total protein) and were observed daily for lethality. Fig. 3 shows that all mice challenged with different protein doses died within 2 to 3 days, while the highest dose caused 80% mortality on day 1. For histopathological examination mice were given 100 μg of BACS protein. All animals died within 3 to 4 hours. Postmortem harvested lungs revealed focal intraalveolar acute hemorrhage, which was from minimal to moderately severe with no endothelial cell damage or vasculitis, and mild patchy congestion of medium-size blood vessels. There was evidence of focal platelet accumulation located within areas of hemorrhage or within vessels. In a control experiment, lethal toxin at a comparable dose (100 μg LF, 100 μg PA) caused neither mortality nor hemorrhage, and in fact, produced no significant identifiable histopathological changes. In an additional control experiment, 100 μg of BSCS protein demonstrated no lethality (data not shown).
Figure 3 Survival of mice upon intratracheal injection of BACS.
Protection of mice against anthrax using protease inhibitors
Effective suppression of the hemorrhagic activity of BACS with chemical inhibitors prompted us to test their protective effect against B. anthracis infection. Two common chemical inhibitors were tested in this study: phosphoramidon and 1, 10-phenanthroline. As discussed above, phosphoramidon inhibits the hemorrhagic effect of BACS. It is a potent inhibitor of thermolysin and other bacterial metallo-endopeptidases, but it does not inhibit trypsin, papain, chymotrypsin or pepsin and weakly inhibits collagenase [19]. Phenanthroline is a potent chelating inhibitor of M4 MPs, such as pseudolysin, as well as matrix MPs [10]. We first tested the above inhibitors for their capacity against the bulk of gelatinase and caseinase activity of BACS in vitro. It was found that up to 1 mM phosphoramidon did not inhibit the BACS gelatinase activity, while 50% of caseinolytic activity was inhibited at the concentration of 0.1 mM (IC50). In contrast, phenanthroline was more potent with gelatin as substrate (IC50 0.01 mM), compared to casein (IC50 0.1 mM).
The B. anthracis delta Ames strain is avirulent in mice, and therefore for the spore challenge experiments we used a toxigenic Sterne strain. We have previously reported the successful application of an adjunct therapy against anthrax infection targeting both bacterial multiplication and host response to infection by using a combination of antibiotic and caspase inhibitors [22]. The same principle was used in the current study because inhibition of secreted pathogenic factors is not expected to directly interfere with bacterial multiplication and therefore may not be fully protective. In order to target both the bacteria and the proteolytic factors we used a combination therapy where antibiotic administration was complemented by the administration of a protease inhibitor. We were also interested in the efficacy of delayed treatment initiated after certain periods of time following spore challenge. It is a practically relevant scenario because patients generally seek medical help after the onset of symptoms, and in other patients treatment begins after a certain period of time required to confirm the exposure. In addition, a delayed ciprofloxacin therapy in murine model is only partially protective when currently recommended human antibiotic doses (adjusted for body weight) are used in mice [22]. These conditions of treatment allowed us to test if a combination approach could lead to a synergistic enhancement in survival.
Results of three independent experiments are presented in Figs. 4 and 5. Mice were challenged intraperitoneally (i.p.) with about 1 × 107 of B. anthracis Sterne spores. Treatment with a single daily dose of ciprofloxacin (50 mg/kg, i.p.) began immediately after challenge, as well as at 24 h or 48 h post challenge, and continued for 10 days. In our conditions the ciprofloxacin treatment initiated immediately after spore challenge was only 70% effective in preventing death. The survival rate after a 24 h delay in antibiotic administration produced a sharp decline to 20% but remained statistically reliable (compared to untreated group, p = 0.015). After a 48 h delay the antibiotic was completely ineffective (p = 0.23). The inhibitor treatment without antibiotic was not able to improve survival, however the combination of ciprofloxacin with inhibitors displayed a synergistic increase in protection, especially notable in the case of phenanthroline. The group receiving phenanthroline/ciprofloxacin treatment delayed by 24 h, demonstrated 70% protection of animals, compared to only a 20% survival in the group with ciprofloxacin alone (p = 0.03 for these groups). The 48 h-delayed regimen resulted in a statistically reliable 30% protection (relative to the untreated spore-challenged group, p < 0.05), in contrast to ciprofloxacin alone (relative to the untreated spore-challenged group, p = 0.23). There is a similar trend in the efficacy of the combination phosporamidon/ciprifloxacin therapy, compared to ciprofloxacin alone, however the observed differences are less reliable (p > 0.05).
Figure 4 Protection of mice against B. anthracis (Sterne) infection by administration of ciprofloxacin and its combination with phosphoramidon for 10 days beginning 24 h and 48 h post spore challenge.
Figure 5 Protection of mice against B. anthracis (Sterne) infection by administration of ciprofloxacin or its combination with phenanthroline for 10 days beginning 24 h and 48 h post spore challenge.
Protection of mice against B. anthracis using anti-protease sera
As in the experiments with inhibitors, mice were challenged intraperitoneally (i.p.) with about 30 LD50 of B. anthracis Sterne spores. Treatment with a single daily dose of ciprofloxacin (50 mg/kg, i.p.) began at 24 h post challenge and continued for 10 days. The immune sera (each pulled from two rabbits) were administered once daily at a concentration of 25 mg/ml (i.p.). The sera displayed substantial differences in their protective effect (Fig. 6). The anti-M4 serum against the epitope(s) of the active center displayed the highest protection (60%; p = 0.038 compared to naïve serum). The rest of the immune sera, namely the anti-collagenase serum anti-M9Coll and the anti-neutral protease serum anti-M4EP behaved similar to the naïve serum. All three latter sera demonstrated a borderline protection from 10% to 30%, compared to untreated mice (p = 0.028, 0.084, and 0.052, respectively), however the difference in survival between them was statistically unreliable (p > 0.41).
Figure 6 Post exposure efficacy of hyperimmune rabbit sera in mice challenged with B. anthracis (Sterne). Treatment with sera and ciprofloxacin was initiated 24 h post exposure and continued for 10 days once daily.
A combination treatment with both antibiotic and all studied immune sera, administered at the dose of 25 mg/kg was strongly synergistic and protected from 80 to 100% mice with high statistical significance (p = 0.0024, 0.0058, and 0.016 for anti-M9, anti-M4EP, and anti-M4AC, correspondingly, compared to the effect of naïve serum-antibiotic combination). A lower serum dose (5 mg/kg) showed a similar pattern of protection, however the effect of combination treatment was reduced to 70%. Some synergistic protection was also noticed in the case of naïve serum (perhaps due to its natural bactericidal properties), although it is statistically unreliable compared to antibiotic-alone group.
Discussion
In order to be highly virulent, any pathogenic microbe is required to possess the means to effectively establish and further propagate the infectious process. Distinct virulence factors may be necessary to fulfill these requirements at different stages of the disease. B. anthracis is a recently emerged highly virulent pathogen, which acquired two plasmids pXO1 and pXO2 compared to the genetically similar but opportunistic pathogen B. cereus. These plasmids encode for the lethal toxin (LT) and capsule genes, respectively [4]. LT is a secreted protein, which has long been considered as a major late virulence factor. It appears in the circulation at the septisemic stage of the disease shortly before death [23,24]. LT is the only anthrax protein, which was found toxic in experimental animals [23,24], and this discovery essentially abrogated further research efforts on other potential virulence factors. However, it has long been known that systemic toxicity of LT is low [24,25], and that the histopathology of LT intoxication differs considerably from that found in clinical anthrax infection [5,6]. It is especially notable that postmortem examination of victims from the Sverdlovsk accident and those autopsied following the 2001 U.S. anthrax attacks [2,4] revealed hemorrhagic thoracic lymphadenitis and necrotizing hemorrhagic mediastinitis in all patients. About half of the Sverdlovsk victims additionally had hemorrhagic meningitis [2]. These severe life-critical symptoms had not been noticed in the intoxicated animals.
Recent data, however, suggest a new important role of LT as a disease-establishing virulence factor playing an immunosuppressive role within alveolar macrophages at the early stages of inhalation anthrax [for review, see ref. [26]]. The immunosuppressive capacity of LT is limited, and therefore LT may not completely inhibit the pro-inflammatory host response [27]. Nevertheless, the LT activity does help anthrax spores to survive in the hostile environment, which contributes to the decrease in the infectious dose. It has been demonstrated that LT caused apoptotic death of macrophages and that its inhibition decreased survival of B. anthracis spores engulfed by macrophages [28]. These data suggested that the major roles of LT and ET might actually be to provide a more hospitable environment for the pathogen intra-phagocytic survival. This hypothesis helps explain the low virulence of the plasmid-free strains of both B. anthracis and B. cereus upon spore challenge [29] in spite of the highly toxic proteolytic capacity of these species.
While only fragmented data have been reported on the existence of B. anthracis chromosome-encoded virulence factors [20,30,31], it is well established that B. cereus produces a variety of pathogenic determinants, including a necrotizing enterotoxin, an emetic toxin, extracellular proteases, phospholipases and hemolysins [32]. B. cereus is capable of causing serious and sometimes lethal infections such as sepsis, pneumonia, meningitis, endocarditis, wound and ocular infections, especially in immunocompromised individuals [32-35]. A highly virulent isolate of B. cereus has recently been identified which contains a plasmid 99.6% similar to pXO1 [36]. This finding is consistent with the point of view that the B. cereus genetic background is sufficient for high virulence when it is complemented with an infection-establishing virulence factor such as LT. Complete sequencing of the B. anthracis and B. cereus genomes confirmed their close relationship suggested previously [37] and allowed us to suggest new candidate virulence factors for B. anthracis, specifically the proteases of the M4 and M9 families. Structurally similar proteolytic factors in other pathogenic microorganisms are known to be involved in inactivation of complement factors [38], cleavage of serum protease inhibitors [39], activation of blood coagulation system [40], invasiveness into the host tissue [41], and development of hemorrhages [13].
We demonstrate here that secreted metalloproteases (MPs) of B. anthracis can digest protein substrates such as casein and gelatin in vitro, and can induce a hemorrhagic process in our test subjects, in vivo. Both of these activities are inhibited by chemical inhibitors of M4 and M9 MPs, such as EDTA, phosphoramidon and 1, 10-phenanthroline. Consistent with this, the hyper-immune mouse serum against M4 family thermolysin/elastase-like enzyme is capable of inhibiting the hemorrhagic effect of BACS (we currently investigate the in vitro inhibiting activity of the hyper-immune sera used in this report). The tissue-damaging action of this type of enzymes is well known [10]. For example, pseudolysin, an elastase of Pseudomonas aeruginosa, destroys arterial elastic laminae in systemic infection, causes lung damage with hemorrhages and necrosis, and induces septic shock through activation of the host kinin cascade [42-44].
In the present study we used an intratracheal (i.t.) administration to demonstrate that tissue destructive and hemorrhagic properties of BACS could cause a lethal effect at 0.5 to 3 mg/kg doses of total protein (10 to 60 μg per 20 g DBA/2 mouse) within a few days or even hours (Fig. 3). Histopathological examination confirmed our observations of life-threatening severe bleeding upon administration of BACS. It has been reported that P. aeruginosa elastase induced an immediate lethal shock in guinea pigs upon an i.v. injection at a similar dose of 1.2 mg/kg [44]. Compared to BACS, the LT is non-toxic upon an i.t. administration. There is no mortality in a control group of LT-treated mice (200 μg/mouse, i.t.). Recently, the toxicity of highly purified LT was re-evaluated in BALB/CJ and C57BL/6J mice, and it was found that doses from 5 to 12.5 mg/kg (i.v.) were required for up to 90% mortality in 5 days [6]. We conclude that secreted proteins of B. anthracis, in addition to LT and ET, have high pathogenic potential and should be considered as important virulence factors.
We have previously suggested a combination antibiotic-antitoxin approach to anthrax therapy and for the first time demonstrated an increased efficacy of ciprofloxacin treatment in a murine model when it was combined with a caspase inhibitor administration [22]. Up to now, there has been no other report on any anti-LT treatment during the anthrax infectious process. In the present study, we used two new combination therapies, namely the antibiotic-protease inhibitor and the antibiotic-antiserum ones. Both of them also proved beneficial, compared to antibiotic alone. It is especially important to note that we used the delayed treatment regiments, which are inherently less effective, compared to prophylactic drug administration.
Currently, a considerable effort is directed towards development of specific LT inhibitors (see for example [45-47]) while their efficacy in the treatment of the infectious process has not been reported. In contrast, the inhibitors we used (phosphoramidon and phenanthroline) to model anthrax therapy are not considered as LT inhibitors [47], however in our experiments both of these substances increased survival during ciprofloxacin therapy, which was initiated at 24 h and 48 h post challenge. The phenanthroline-ciprofloxacin combination administered 24 h post challenge protected 70% of the mice, compared to 20% for antibiotic alone. Late stages of anthrax are especially difficult to treat [4]. In our model the 48 h post challenge time approximately correlates with the period of typical overt anthrax symptoms in patients because 30% of mice die within the next 24 h, a situation that is similar to that clinically observed in natural human infection. In these circumstances, our therapy with phenanthroline-ciprofloxacin was 30% protective, while ciprofloxacin alone was ineffective. It is worth pointing out that the high extent of protection (70% after 24 h delay) conferred by a combination of antibiotic with phenanthroline argues in favor of the substantial role of secreted proteases as death-causing factors, although their contribution to mortality relative to LT is currently unknown. We have previously reported that in similar experimental conditions the caspase inhibitor YVAD, capable of protecting macrophages against LT-induced apoptosis, improved survival of DBA/2 mice by 30%. One may therefore expect that a triple component therapy, such as ciprofloxacin-phosphoramidon-caspase inhibitor, might be completely protective. Experiments in this direction are in progress. In connection with the question on the role of LT, we also currently investigate a spectrum of neutralizing activity of immune sera used in this study.
Conclusion
Overall, our data demonstrate that B. anthracis cultivated in culture media secrets a number of proteolytic virulence factors, including those with hemorrhagic, caseinolytic and gelatinolytic activities. These factors in most respects are distinct from LT, including the mode of their expression under aerobic conditions (LT requires bicarbonate for its expression in vitro [48]), their molecular targets, as well as a high virulent potency upon intratracheal administration. Chemical inhibitors of these factors as well as immune sera raised against them in rabbits demonstrate a substantial protective efficacy in combination with antibiotic therapy. Our findings outline a new direction in the development of anthrax therapeutic approaches, and close a substantial gap between the understanding of anthrax molecular pathology and the most prominent clinical features of its infectious process. Complexity of the BACS composition with regard to the number and specificity of proteolytic enzymes suggests a multitude of their potential virulent mechanisms that need to be explored further.
Competing interests
SP, TP, SH, KA, KJF, CB and VC are co-authors of pending patent applications related to the content of the manuscript. The authors have also received salaries from George Mason University and/or Advanced Biosystems, Inc., both of which hold an interest in these patents. The authors declare they have no other competing interests.
Authors' contributions
SP developed the research plan, directed experimental work and was principal writer of the manuscript. TP analyzed the protease sequences, designed peptides for immunization, and carried out the in vitro studies. SH carried out all animal experiments and participated in the in vitro characterization of culture supernatants. RW and RM conducted a histopathological evaluation of tissue specimens. KF carried out bioinformatics analyses and participated in manuscript preparation. KA conceived of the study, developed its initial design, along with CB and VC participated in research coordination and helped to draft 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 the U.S. Department of Defense grant DAMD17-03-C-0122. The authors thank Lavin Joseph and Chris Bradburne for help in carrying out experiments.
==== Refs
Abramova FA Grinberg LM Yampolskaya OV Walker DH Pathology of inhalational anthrax in 42 cases from the Sverdlovsk outbreak of 1979 Proc Natl Acad Sci U S A 1993 90 2291 2294 8460135
Grinberg LM Abramova FA Yampolskaya OV Walker DH Smith JH Quantitative pathology of inhalational anthrax I: quantitative microscopic findings Mod Pathol 2001 14 482 495 11353060 10.1038/modpathol.3880337
Vasconcelos D Barnewall R Babin M Hunt R Estep J Nielsen C Carnes R Carney J Pathology of inhalation anthrax in cynomolgus monkeys (Macaca fascicularis) Lab Invest 2003 83 1201 1209 12920249 10.1097/01.LAB.0000080599.43791.01
Inglesby TV O'Toole T Henderson DA Bartlett JG Ascher MS Eitzen E Friedlander AM Gerberding J Hauer J Hughes J McDade J Osterholm MT Parker G Perl TM Russell PK Tonat K Anthrax as a biological weapon, 2002: updated recommendations for management JAMA 2002 287 2236 2252 11980524 10.1001/jama.287.17.2236
Cui X Moayeri M Li Y Li X Haley M Fitz Y Correa-Araujo R Banks SM Leppla SH Eichacker PQ Lethality during continuous anthrax lethal toxin infusion is associated with circulatory shock but not inflammatory cytokine or nitric oxide release in rats Am J Physiol Regul Integr Comp Physiol 2004 286 R699 709 14715494
Moayeri M Haines D Young HA Leppla SH Bacillus anthracis lethal toxin induces TNF-alpha-independent hypoxia-mediated toxicity in mice J Clin Invest 2003 112 670 682 12952916 10.1172/JCI200317991
Klein F Walker JS Fitzpatrick DF Lincoln RE Mahlandt BG Jones WI JrDobbs JP Hendrix KJ Pathophysiology of anthrax J Infect Dis 1966 116 123 138 4956203
Eckert NJ Bonventre PF In vivo effects of Bacillus anthracis culture filtrates J Inf Dis 1963 112 226 232
Ezepchuk IuV Bobkova ED Prosvetova NK [On the toxic and various biochemical properties of filtrates of B. anthracis and cereus] Zh Mikrobiol Epidemiol Immunobiol 1969 46 19 23 [Article in Russian] 4981336
Supuran CT Scozzafava A Clare BW Bacterial protease inhibitors Med Res Rew 2002 22 329 372 10.1002/med.10007
Read TD Peterson SN Tourasse N Baillie LW Paulsen IT Nelson KE The genome sequence of Bacillus anthracis Ames and comparison to closely related bacteria Nature 2003 423 81 6 12721629 10.1038/nature01586
Barrett AJ Bioinformatics of proteases in the MEROPS database Curr Opin Drug Discov Devel 2004 7 334 341 15216937
Miyoshi S Nakazawa H Kawata K Tomochika K Tobe K Shinoda S Characterization of the hemorrhagic reaction caused by Vibrio vulnificus metalloprotease, a member of the thermolysin family Infect Immun 1998 66 4851 4855 9746589
Okamoto T Akaike T Suga M Tanase S Horie H Miyajima S Ando M Ichinose Y Maeda H Activation of human matrix metalloproteinases by various bacterial proteinases J Biol Chem 1997 272 6059 6066 9038230 10.1074/jbc.272.9.6059
Sakata Y Akaike T Suga M Ijiri S Ando M Maeda H Bradykinin generation triggered by Pseudomonas proteases facilitates invasion of the systemic circulation by Pseudomonas aeruginosa Microbiol Immunol 1996 40 415 423 8839427
Shin YH Akaike T Khan MM Sakata Y Maeda H Further evidence of bradykinin involvement in septic shock: reduction of kinin production in vivo and improved survival in rats by use of polymer tailored SBTI with longer t1/2 Immunopharmacology 1996 33 369 373 8856189 10.1016/0162-3109(96)00088-4
Lovgren A Zhang M Engstrom A Dalhammar G Landen R Molecular characterization of immune inhibitor A, a secreted virulence protease from Bacillus thuringiensis Mol Microbiol 1990 4 2137 2146 2089225
Biagini A Puigserver A Sequence analysis of the aminoacylase-1 family. A new proposed signature for metalloexopeptidases Comp Biochem Physiol B Biochem Mol Biol 2001 128 469 481 11250542 10.1016/S1096-4959(00)00341-9
Komiyama T Suda H Aoyagi T Takeuchi T Umezawa H Studies on inhibitory effect of phosphoramidon and its analogs on thermolysin Arch Biochem Biophys 1975 171 727 731 1200647 10.1016/0003-9861(75)90085-5
Klichko VI Miller J Wu A Popov SG Alibek K Anaerobic induction of Bacillus anthracis hemolytic activity Biochem Biophys Res Commun 2003 303 855 862 12670489 10.1016/S0006-291X(03)00440-6
Bonventre PF Eckert NJ The biologic activities of Bacillus anthracis and Bacillus cereus culture filtrates Am J Pathol 1963 43 201 211 14076029
Popov SG Popova TG Grene E Klotz F Cardwell J Bradburne C Jama Y Maland M Wells J Nalca A Voss T Bailey C Alibek K Systemic cytokine response in murine anthrax Cell Microbiol 2004 6 225 233 14764106 10.1046/j.1462-5822.2003.00358.x
Smith H Keppie J Observations on experimental anthrax: Demonstration on a specific lethal factor produced in vivoi by Bacillus anthracis Nature 1954 173 869 870 13165673
Smith H Keppie J Stanley JL The chemical basis of the virulence of Bacillus anthracis. V. The specific toxin produced by B. anthracis in vivo Br J Exp Pathol 1955 36 460 472 13269658
Friedlander AM Tackling anthrax Nature 2001 414 160 161 11700539 10.1038/35102660
Agrawal A Pulendran B Anthrax lethal toxin: a weapon of multisystem destruction Cell Mol Life Sci 2004 61 2859 2865 15558214 10.1007/s00018-004-4251-4
Pickering AK Osorio M Lee GM Grippe VK Bray M Merkel TJ Cytokine response to infection with Bacillus anthracis spores Infect Immun 2004 72 6382 6389 15501768 10.1128/IAI.72.11.6382-6389.2004
Popov SG Villasmil R Bernardi J Grene E Cardwell J Wu A Alibek D Bailey C Alibek K Lethal toxin of Bacillus anthracis causes apoptosis of macrophages Biochem Biophys Res Commun 2002 293 349 355 12054607 10.1016/S0006-291X(02)00227-9
Guidi-Rontani C levy M Ohayon H Mock M Fate of germinated Bacillus anthracis spores in primary murine macrophages Mol Microbiol 2001 42 931 938 11737637 10.1046/j.1365-2958.2001.02695.x
Mikshis NI Stepanov AS Shevchenko OV Eremin SA The intrastrain heterogeneity of the causative agent of anthrax Zh Mikrobiol Epidemiol Immunobiol 1999 3 78 9 [Article in Russian] 10851999
Stepanov AS Mikshis NI Eremin SA Bolotnikova MF Contribution of determinants, located in Bacillus anthracis chromosomes, in realizing the pathogenic properties of the pathogen Mol Gen Mikrobiol Virusol 1999 1 20 23 [Article in Russian] 10190106
Drobniewski FA Bacillus cereus and related species Clin Microbiol Rev 1993 6 324 338 8269390
Akesson A Hedstrom SA Ripa T Bacillus cereus: a significant pathogen in postoperative and post-traumatic wounds on orthopaedic wards Scand J Infect Dis 1991 23 71 77 1902983
David DB Kirkby GR Noble BA Bacillus cereus endophthalmitis Br J Ophthalmol 1994 78 577 580 7918273
Turnbull PC French TA Dowsett EG Severe systemic and pyogenic infections with Bacillus cereus Br Med J 1977 1 1628 1629 406000
Hoffmaster AR Ravel J Rasko DA Chapman GD Chute MD Marston CK De BK Sacchi CT Fitzgerald C Mayer LW Maiden MC Priest FG Barker M Jiang L Cer RZ Rilstone J Peterson SN Weyant RS Galloway DR Read D Popovic T Fraser CM Identification of anthrax toxin genes in a Bacillus cereus associated with an illness resembling inhalation anthrax Proc Natl Acad Sci U S A 2004 101 8449 8454 15155910 10.1073/pnas.0402414101
Helgason E Okstad OA Caugant DA Johansen HA Fouet A M Mock Hegna I Kolsto J Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis – one species on the basis of genetic evidence Appl Environ Microbiol 2000 66 2627 30 10831447 10.1128/AEM.66.6.2627-2630.2000
Cleary PP Prahbu U Dale JB Wexler DE Handley J Streptococcal C5a peptidase is a highly specific endopeptidase Infect Immun 1992 60 5219 5223 1452354
Ishihara K Miura T Kuramitsu HK Okuda K Characterization of the Treponema denticola prtP gene encoding a prolyl-phenylalanine-specific protease (dentilisin) Infect Immun 1996 64 5178 5186 8945563
Kaminishi H Hamatake H Cho T Tamaki T Suenaga N Fujii T Hagihara Y Maeda H Activation of blood clotting factors by microbial proteinases FEMS Microbiol Lett 1994 121 327 332 7926688 10.1016/0378-1097(94)90312-3
Sodeinde OA Subrahmanyam YV Stark K Quan T Bao Y Goguen JD A surface protease and the invasive character of plague Science 1992 258 1004 1007 1439793
Kon Y Tsukada H Hasegawa T Igarashi K Wada K Suzuki E Arakawa M Gejyo F The role of Pseudomonas aeruginosa elastase as a potent inflammatory factor in a rat air pouch inflammation model FEMS Immunol Med Microbiol 1999 25 313 321 10459586 10.1016/S0928-8244(99)00101-7
Maeda H Akaike T Sakata Y Maruo K Role of bradykinin in microbial infection: enhancement of septicemia by microbial proteases and kinin Agents Actions Suppl 1993 42 159 165 8356921
Khan MM Yamamoto T Araki H Ijiri Y Shibuya Y Okamoto M Kambara T Pseudomonal elastase injection causes low vascular resistant shock in guinea pigs Biochim Biophys Acta 1993 1182 83 93 8347690
Mourez M Kane RS Mogridge J Metallo S Deschtelets P Sellman BR Whitesides GM Collier RJ Designing a polyvalent inhibitor of anthrax toxin Nat Biotech 2001 19 958 961 10.1038/nbt1001-958
Min D-H Tang W-J Mrksich M Chemical screening by mass spectrometry to identify inhibitors of anthrax lethal factor Nat Biotech 2004 22 717 720 10.1038/nbt973
Turk BE Wong TY Schwarzenbacher R Jarrell ET Leppla SH Collier RJ Liddington RC Cantley LC The structural basis for substrate and inhibitor selectivity of the anthrax lethal factor Nat Struct Mol Biol 2004 11 60 66 14718924 10.1038/nsmb708
Wright GG Hedberg MA Slein JB Studies on immunity in anthrax. III. Elaboration of protective antigen in a chemically-defined, non-protein medium J Immunol 1954 72 263 269 13163399
| 15819985 | PMC1090577 | CC BY | 2021-01-04 16:28:15 | no | BMC Infect Dis. 2005 Apr 8; 5:25 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-25 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-261581999510.1186/1471-2334-5-26Research ArticleAbsence of association between angiotensin converting enzyme polymorphism and development of adult respiratory distress syndrome in patients with severe acute respiratory syndrome: a case control study Chan KC Allen [email protected] Nelson LS [email protected] David SC [email protected] Grace TY [email protected] Alan KL [email protected] Stephen SC [email protected] Rossa WK [email protected] Nelson [email protected] KW [email protected] YM [email protected] Paul KS [email protected] YK [email protected] ST [email protected] WC [email protected] Owen [email protected] YM Dennis [email protected] The Centre for Emerging Infectious Diseases, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong2 Department of Chemical Pathology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong3 Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong4 Department of Microbiology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong5 Department of Medicine and Geriatrics, Princess Margaret Hospital, Hong Kong2005 9 4 2005 5 26 26 11 1 2005 9 4 2005 Copyright © 2005 Chan et al; licensee BioMed Central Ltd.2005Chan 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
It has been postulated that genetic predisposition may influence the susceptibility to SARS-coronavirus infection and disease outcomes. A recent study has suggested that the deletion allele (D allele) of the angiotensin converting enzyme (ACE) gene is associated with hypoxemia in SARS patients. Moreover, the ACE D allele has been shown to be more prevalent in patients suffering from adult respiratory distress syndrome (ARDS) in a previous study. Thus, we have investigated the association between ACE insertion/deletion (I/D) polymorphism and the progression to ARDS or requirement of intensive care in SARS patients.
Method
One hundred and forty genetically unrelated Chinese SARS patients and 326 healthy volunteers were recruited. The ACE I/D genotypes were determined by polymerase chain reaction and agarose gel electrophoresis.
Results
There is no significant difference in the genotypic distributions and the allelic frequencies of the ACE I/D polymorphism between the SARS patients and the healthy control subjects. Moreover, there is also no evidence that ACE I/D polymorphism is associated with the progression to ARDS or the requirement of intensive care in the SARS patients. In multivariate logistic analysis, age is the only factor associated with the development of ARDS while age and male sex are independent factors associated with the requirement of intensive care.
Conclusion
The ACE I/D polymorphism is not directly related to increased susceptibility to SARS-coronavirus infection and is not associated with poor outcomes after SARS-coronavirus infection.
==== Body
Background
The outbreak of the severe acute respiratory syndrome (SARS) has made a great impact to the health care systems around the world. The pandemic affected over 8000 individuals and resulted in 774 deaths worldwide [1]. Several clinical parameters, including male sex [2,3], age of over 60 years [2,3], elevated lactate dehydrogenase activity [2-4], low platelet count [2] and high viral load on presentation [5], have been identified to be predictive of the severity of the disease in affected individuals. Moreover, it has been postulated that genetic variations of the host and the virus may account for the individual difference in the susceptibility to the infection and the severity of the disease. With regard to viral factors, it has been shown that there is no significant difference in the genetic sequences of viruses causing the two major outbreaks in Hong Kong, namely the Prince of Wales Hospital and Amoy Gardens outbreaks, despite the significant difference in the mortality rates and diarrheal rates of the two cohorts [6]. Furthermore, several association studies have been conducted to investigate the possible contribution of host genetic factors in the determination of the susceptibility and prognosis of SARS-coronavirus infection. Thus, certain human leukocyte antigen subtypes have been shown to be more prevalent in SARS patients [7] and in those who had poorer outcomes [8]. On the other hand, the polymorphism in the angiotensin converting enzyme II gene, coding for a functional receptor of the SARS-coronavirus, is not associated with the susceptibility or outcome of SARS [9]. Recently, it has also been reported that the deletion of the 287 bp Alu repeat (D allele) in intron 16 of the ACE gene is associated with hypoxemia in SARS patients [10]. However, there are several limitations to this previous study. First, only 44 SARS patients were studied. Second, hypoxemia was arbitrarily defined as requiring oxygen supplementation. Moreover, patients who died were excluded from the study. These factors may be potential confounders to a genetic association study.
Therefore, in this study, we investigated the association of the ACE insertion/deletion (I/D) polymorphism of the 287 bp Alu repeat to the susceptibility to SARS and the development of adult respiratory distress syndrome (ARDS) with a larger population.
Methods
Study population
This study was reviewed and approved by the Ethical Committee of the Prince of Wales Hospital, Hong Kong. Patients who were admitted to the hospitals of the New Territories East cluster of Hong Kong for the treatment of SARS were recruited retrospectively. The recruitment of patients depended on the availability of blood samples. All patients, including survivors and deceased patients, with available blood samples were recruited. For genetically related SARS patients, only the index case (the first individual who developed symptoms) was recruited. All patients were of Chinese ethnicity and fulfilled the World Health Organisation case definition of probable SARS [11]. Three hundred and twenty-six healthy individuals undergoing routine health check were recruited as controls. The control subjects were recruited before the SARS epidemic and none of them had respiratory symptoms. All control subjects were ethnical Chinese and were not genetically related.
Definition of adult respiratory distress syndrome (ARDS) and patients with severe disease
The association between genotype and disease outcome was studied in the SARS patients. Two categories of patients were considered as having a severe disease: (1) patients who developed ARDS; and (2) patients who required admission to the intensive care unit (ICU). A patient was classified as having ARDS if he or she fulfilled all criteria of the joint American/European Consensus for ARDS [12], including: (1) acute onset of respiratory distress; (2) presence of bilateral infiltrates on chest X-ray; (3) having a ratio of arterial partial pressure of oxygen to inspired fractional oxygen concentration (PaO2/FiO2) of less than 26.8 kPa and absence of clinical evidence of left heart failure.
ACE Genotyping
DNA was extracted from whole blood using a QIAamp DNA Blood Mini Kit (Qiagen) with the 'Blood and Body Fluid Spin protocol' as recommended by the manufacturer. ACE I/D genotypes were determined by polymerase chain reaction amplification. The forward and reverse primers were 5'-CTGGAGACCACTCCCATCCTTTCT-3' and 5'-GATGTGGCCATCACATTCGTCAGAT-3', respectively. Reactions were set up in a volume of 25 μl containing 0.1 μM of each primer, 1X buffer II (Applied Biosystems), 2 mM MgCl2, 0.2 mM of each dNTP, 1.25 U Taq polymerase (AmpliTaq Gold DNA polymerase, Applied Biosystems) and 20 ng DNA. After initial denaturation at 95°C for 12 min, the reaction mixtures were subjected to 40 cycles of 94°C for 1 min, 58°C for 1 min and 72°C for 1 min, and a final extension at 72°C for 5 min. This method yielded amplification products of 480 bp for the I allele and 192 bp for the D allele. The products were electrophoresed and visualized in 2% agarose gels with ethidium bromide.
Statistical analysis
Statistical analyses were performed using SigmaStat, Ver. 3.0; SPSS. Disease associations were compared by chi-square tests. Univariate and multivariate logistic regression analyses were performed to identify predictors of ARDS or the outcome of SARS.
Results
Demographics
One hundred and forty SARS patients (67 males, 73 females) and 326 healthy individuals (172 males, 154 females) were recruited. The mean ages of the SARS patients and control subjects were 39.9 and 42.5 years, respectively (p = 0.93). Seventeen of the 140 SARS patients developed ARDS during the course of their illness. The demographic data of the SARS patients who had or had not developed ARDS are summarized in table 1. Patients who developed ARDS were significantly older than those who did not develop ARDS (p < 0.001). There was no significant difference in gender, smoking habits, hepatitis B status and the presence of comorbidity between the two groups. Thirty-five patients required intensive care and sixteen died. Patients who required intensive care were significantly older than those with milder disease.
Table 1 Demographic data of SARS patients. The patients are categorized by the development of ARDS (ARDS vs non-ARDS). All patients who had developed ARDS required intensive care. The numbers listed in the table represent the number of patients (except the numbers list in the first row which signify age).
ARDS non-ARDS
Age mean (range) 59.1 years (24–83) 36.9 years (21–76)
M:F 9/8 58/65
Smoking 2 5
HbsAg positive 0 7
Comorbidities
chronic obstructive pulmonary disease 0 3
Ischemic heart disease 5 2
Cerebral vascular disease 3 4
Cancer 0 2
Diabetes mellitus 1 6
Chronic renal disease 0 1
Liver cirrhosis 0 1
Genotypic and allelic frequencies of ACE I/D polymorphism
The genotypic distributions and allelic frequencies of ACE I/D polymorphism in the SARS patients and control subjects are shown in table 2. The genotypic distributions of the SARS patients and the healthy control subjects follow the Hardy-Weinberg equilibrium using chi-square analysis. There was no significant difference in the genotypic distributions (χ2 value = 1.43, df = 2, p = 0.489) and allelic frequencies (χ2 value = 0.000624, df = 1, p = 0.980) of the two groups. Among the SARS patients, we further analyzed the genotypic distributions and allelic frequencies of ACE I/D polymorphism in patients who developed ARDS and in those who did not develop ARDS in the course of their illness. The results are shown in table 3a. There was no significant difference in the genotypic distributions (χ2 value = 4.361, df = 2, p = 0.113) and allelic frequencies (χ2 value = 1.376, df = 1, p = 0.241) between the two groups. Besides, there was also no significant difference in the genotypic distributions (χ2 value = 2.489, df = 2, p = 0.288) and allelic frequencies (χ2 value = 1.424, df = 1, p = 0.233) between patients who did or did not require intensive care. The results are shown in table 3b.
Table 2 Genotypic and allelic frequencies of ACE I/D polymorphism in SARS patients and control subjects
SARS cases Controls
n = 140 n = 326
Genotypic frequency
DD 15 (11%) 43 (13%)
ID 65 (46%) 133 (41%)
II 60 (43%) 150 (46%)
χ2 value = 1.43, df = 2, p = 0.489
Allelic frequency
D 95 (34%) 219 (34%)
I 185 (66%) 433 (66%)
χ2 value = 0.000624, df = 1, p = 0.980
Table 3a Genotypic and allelic frequencies of ACE I/D polymorphism in SARS patients who had developed or had not developed ARDS
ARDS Non-ARDS
n = 17 n = 123
Genotypic frequency
DD 2 (12%) 13 (10%)
ID 4 (23%) 61 (50%)
II 11 (65%) 49 (40%)
χ2 value = 4.361, df = 2, p = 0.113
Allelic frequency
D 8 (24%) 87 (35%)
I 26 (76%) 159 (65%)
χ2 value = 1.376, df = 1, p = 0.241
Table 3b Genotypic and allelic frequencies of ACE I/D polymorphism in SARS patients who had or had not required intensive care (ICU vs non-ICU)
ICU Non-ICU
n = 35 n = 105
Genotypic frequency
DD 3 (9%) 12 (11%)
ID 13 (37%) 52 (50%)
II 19 (54%) 41 (39%)
χ2 value = 2.489, df = 2, p = 0.288
Allelic frequency
D 19 (27%) 76 (36%)
I 51 (73%) 134 (64%)
χ2 value = 1.424, df = 1, p = 0.233
Logistic regression on the development of ARDS or requirement of intensive care
In the univariate analysis, we did not detect any significant difference in the number of D alleles in the ACE polymorphism between patients who did and did not develop ARDS (p = 0.169, OR = 0.549 (95% CI: 0.23–1.29). Following multivariate logistic regression analysis, age was found to be the only significant factor that determined the development of ARDS in SARS patients (table 4a). In the multivariate analysis for the requirement of intensive care, we have shown that age and male sex are associated with the requirement of intensive care (table 4b).
Table 4a Logistic regression on the development of ARDS in SARS patients
Variables Univariate analysis Multivariate analysis
Odds ratio (95% CI) p value Odds ratio (95% CI) p value
ACE (no. of D allele) 0.64 (0.28 – 1.47) 0.30 0.74 (0.27 – 1.98) 0.46
Age 1.08 (1.04 – 1.12) <0.01 1.10 (1.04 – 1.16) <0.01
Male sex 1.26 (0.46 – 3.48) 0.65 0.22 (0.04 – 1.10) 0.06
Smoking 3.15 (0.56 – 17.67) 0.19 1.03 (0.11 – 9.65) 0.98
HbsAg 1 (0 – ∞) 0.99 1 (0.00 – ∞) 0.99
Co-morbidities 6.15 (2.11 – 17.96) <0.01 1.61 (0.36 – 7.26) 0.53
Table 4b Logistic regression on the requirement of intensive care in SARS patients
Variables Univariate analysis Multivariate analysis
Odds ratio (95% CI) p value Odds ratio (95% CI) p value
ACE (no. of D allele) 0.69 (0.38 – 1.27) 0.23 0.77 (0.39 – 1.51) 0.46
Age 1.06 (1.03 – 1.08) <0.01 1.06 (1.02 – 1.09) <0.01
Male sex 4.50 (1.92 – 10.58) <0.01 3.20 (1.25 – 8.19) 0.02
Smoking 2.37 (0.50 – 11.14) 0.28 0.68 (0.10 – 4.52) 0.69
HbsAg 2.37 (0.50 – 11.14) 0.28 2.73 (0.52 – 14.43) 0.24
Co-morbidities 6.15 (2.11 – 17.96) <0.01 1.61 (0.36 – 7.26) 0.53
Discussion
The possible contribution of host genetic factors to the susceptibility and outcome of SARS-coronavirus infection has been investigated through several association studies [7-10]. In contrast to a recent report showing an association between the presence of the D allele of the ACE gene and hypoxemia in SARS patients [10], we have shown that the I/D polymorphism of the ACE gene is associated with neither increased susceptibility to SARS-coronavirus infection nor progression to ARDS once infected. In multivariate logistic regression analysis, we have identified that age is the only significant factor associated with the development of ARDS while age and male sex are independently associated with the requirement of intensive care in SARS patients. Our findings are consistent with other published reports [2,3].
There are several possible explanations for the discrepancies in our conclusion and that by Itoyama et al [10] concerning the association between ACE polymorphism and the outcome of SARS. First, the inclusion of subjects within the same family and exclusion of deceased patients by the previous study might cause potential bias, especially when the frequency of the DD genotype was reported to be as low as 6% in control subjects [10]. In this study, we have only included the index patient if more than one member in a family developed SARS. Second, we have used a well defined endpoint of ARDS instead of the requirement of supplemental oxygen. SARS infection commonly leads to respiratory distress and over 80% of patients were given supplemental oxygen during the course of their illness in our cohort. Therefore, it seems to be more appropriate to use ARDS instead of the requirement of oxygen supplement to define the severity of SARS. As ARDS is the more severe end of the spectrum of disease progression, any potential association between genotype and disease progression would become even more obvious when the most severe cases were considered. Similarly, the disease outcome was not associated with ACE I/D genotype when we also used another broader definition for severe disease after SARS infection (requiring intensive care or death).
Previous studies on Caucasian populations have suggested that the presence of the D allele of the ACE gene is associated with increased incidence of ARDS [13]. This effect has been postulated to be related to the higher enzyme activity in individuals with DD genotype [15]. However, it is unclear whether these observations can also be seen in Chinese as the frequencies of DD genotype and D allele of the ACE gene are much lower in Chinese than in Caucasian subjects [13,16]. Furthermore, the SARS-coronavirus characteristically affects the pneumocytes, and the formation of multinucleated pneumocytes and intrabronchial fibrogranulation (bronchiolitis obliterans organizing pneumonia-like lesions) are commonly observed in the lung biopsies of SARS patients in addition to the typical pathological changes of ARDS [17]. Therefore, it is possible that the pathogenesis and genetic factors predisposing to SARS-related ARDS may be different from ARDS resulted from other respiratory illnesses.
Previous reports have highlighted the inconsistency of the results of genetic association studies for complex diseases [18,19]. This inconsistency may be attributable to the difference in the genetic composition of the studied population and study design. Here, we showed that both susceptibility and disease outcome of SARS infection were not associated with ACE I/D polymorphism among Chinese patients in contrast to the recent report studying Vietnamese patients [10]. The sample size was definitively larger in our study. Two different better-defined criteria were used as indicators of severe disease progression, yet no association was found between disease severity and ACE I/D genotype. The D allele which was the hypothetical high risk allele [13], did not show any sign of over-representation in the subgroups of patients with severe disease.
Conclusion
Our analysis indicates that ACE I/D polymorphism is not directly related to poor outcomes after SARS-coronavirus infection in Chinese.
List of abbreviations used
ACE: angiotensin converting enzyme
SARS: severe acute respiratory syndrome
ARDS: adult respiratory distress syndrome
ACE I/D polymorphism: angiotensin converting enzyme insertion/deletion polymorphism
Competing interests
YMDL, KCAC, RWKC, SCCC and YKT have filed patent applications on aspects concerning the genomics and detection of the SARS-coronavirus.
Authors' contributions
KCAC, NLST, GTYC, and YMDL have contributed in the preparation of the manuscript and the overall study design. RWKC, SSCC, YKT, PKSC and YMS have contributed in the data analysis and conducting the experiments. DSCH, AKLW, NL, KWC, PKSC, STL, WCY and OT have contributed in the collection and analysis of clinical data from the patients.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The work is supported by the Research Fund for the Control of Infectious Disease (RFCID) from the Health, Welfare and Food Bureau of the Hong Kong SAR Government. We thank Coral Lee, WB Lui and Katherine Chow for technical support.
==== Refs
WHO Cumulative number of reported probable cases of SARS http://wwwwhoint/csr/sars/country/2003_07)11/en/ 2004
Tsui PT Kwok ML Yuen H Lai ST Severe acute respiratory syndrome: clinical outcome and prognostic correlates Emerg Infect Dis 2003 9 1064 1069 14519241
Leung GM Hedley AJ Ho LM Chau P Wong IO Thach TQ Ghani AC Donnelly CA Fraser C Riley S Ferguson NM Anderson RM Tsang T Leung PY Wong V Chan JC Tsui E Lo SV Lam TH The epidemiology of severe acute respiratory syndrome in the 2003 Hong Kong epidemic: an analysis of all 1755 patients Ann Intern Med 2004 141 662 673 15520422
Chan MH Wong VW Wong CK Chan PK Chu CM Hui DS Suen MW Sung JJ Chung SS Lam CW Serum LD1 isoenzyme and blood lymphocyte subsets as prognostic indicators for severe acute respiratory syndrome J Intern Med 2004 255 512 518 15049886 10.1111/j.1365-2796.2004.01323.x
Ng EKO Hui DS Chan KCA Hung EC Chiu RWK Lee N Wu A Chim SSC Tong YK Sung JJ Tam JS Lo YMD Quantitative analysis and prognostic implication of SARS coronavirus RNA in the plasma and serum of patients with severe acute respiratory syndrome Clin Chem 2003 49 1976 1980 14633867 10.1373/clinchem.2003.024125
Chim SSC Tsui SK Chan KCA Au TC Hung EC Tong YK Chiu RWK Ng EK Chan PKS Chu CM Sung JJ Tam JS Fung KP Waye MM Lee CY Yuen KY Lo YMD Genomic characterisation of the severe acute respiratory syndrome coronavirus of Amoy Gardens outbreak in Hong Kong Lancet 2003 362 1807 1808 14654320 10.1016/S0140-6736(03)14901-X
Ng MH Lau KM Li L Cheng SH Chan WY Hui PK Zee B Leung CB Sung JJ Association of human-leukocyte-antigen class I (B*0703) and class II (DRB1*0301) genotypes with susceptibility and resistance to the development of severe acute respiratory syndrome J Infect Dis 2004 190 515 8. Epub 2004 Jul 07. 15243926 10.1086/421523
Lin M Tseng HK Trejaut JA Lee HL Loo JH Chu CC Chen PJ Su YW Lim KH Tsai ZU Lin RY Lin RS Huang CH Association of HLA class I with severe acute respiratory syndrome coronavirus infection BMC Med Genet 2003 4 9 12969506 10.1186/1471-2350-4-9
Chiu RWK Tang NL Hui DS Chung GT Chim SSC Chan KCA Sung YM Chan LY Tong YK Lee WS Chan PK Lo YMD ACE2 gene polymorphisms do not affect outcome of severe acute respiratory syndrome Clin Chem 2004 50 1683 1686 15331509 10.1373/clinchem.2004.035436
Itoyama S Keicho N Quy T Phi NC Long HT Ha le D Ban VV Ohashi J Hijikata M Matsushita I Kawana A Yanai H Kirikae T Kuratsuji T Sasazuki T ACE1 polymorphism and progression of SARS Biochem Biophys Res Commun 2004 323 1124 1129 15381116 10.1016/j.bbrc.2004.08.208
WHO Case Definitions for Surveillance of Severe Acute Respiratory Syndrome (SARS) http://wwwwhoint/csr/sars/casedefinition/en/ 2004
Bernard GR Artigas A Brigham KL Carlet J Falke K Hudson L Lamy M Legall JR Morris A Spragg R The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination Am J Respir Crit Care Med 1994 149 818 824 7509706
Marshall RP Webb S Bellingan GJ Montgomery HE Chaudhari B McAnulty RJ Humphries SE Hill MR Laurent GJ Angiotensin converting enzyme insertion/deletion polymorphism is associated with susceptibility and outcome in acute respiratory distress syndrome Am J Respir Crit Care Med 2002 166 646 650 12204859 10.1164/rccm.2108086
Zhao B Moochhala SM Tham S Lu J Chia M Byrne C Hu Q Lee LK Relationship between angiotensin-converting enzyme ID polymorphism and VO(2max) of Chinese males Life Sci 2003 73 2625 2630 12967686 10.1016/S0024-3205(03)00608-8
Costerousse O Allegrini J Lopez M Alhenc-Gelas F Angiotensin I-converting enzyme in human circulating mononuclear cells: genetic polymorphism of expression in T-lymphocytes Biochem J 1993 290 33 40 8382480
Koh WP Yuan JM Sun CL van den Berg D Seow A Lee HP Yu MC Angiotensin I-converting enzyme (ACE) gene polymorphism and breast cancer risk among Chinese women in Singapore Cancer Res 2003 63 573 578 12566298
Tse GM To KF Chan PK Lo AW Ng KC Wu A Lee N Wong HC Mak SM Chan KF Hui DS Sung JJ Ng HK Pulmonary pathological features in coronavirus associated severe acute respiratory syndrome (SARS) J Clin Pathol 2004 57 260 265 14990596 10.1136/jcp.2003.013276
Colhoun HM McKeigue PM Davey Smith G Problems of reporting genetic associations with complex outcomes Lancet 2003 361 865 872 12642066 10.1016/S0140-6736(03)12715-8
Daly AK Day CP Candidate gene case-control association studies: advantages and potential pitfalls Br J Clin Pharmacol 2001 52 489 499 11736857 10.1046/j.0306-5251.2001.01510.x
| 15819995 | PMC1090578 | CC BY | 2021-01-04 16:28:15 | no | BMC Infect Dis. 2005 Apr 9; 5:26 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-26 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-271582321210.1186/1471-2334-5-27Research ArticleHigh versus standard doses interferon-alpha in the treatment of naïve chronic hepatitis C patients in Taiwan: a 10-year cohort study Yu Ming-Lung [email protected] Chia-Yen [email protected] Shinn-Cherng [email protected] Li-Po [email protected] Ming-Yen [email protected] Zu-Yau [email protected] Ming-Yuh [email protected] Liang-Yen [email protected] Jung-Fa [email protected] Wen-Yu [email protected] Wan-Long [email protected] Hepatobiliary Division, Department of Internal Medicine, Kaohsiung Medical University Hospital, No. 100, Tzyou 1st Rd, Kaohsiung 807, Kaohsiung, Taiwan2 Department of Occupational Medicine, Kaohsiung Municipal HsiaoKang Hospital, No. 482, Shan-Ming Rd, Kaohsiung 812, Kaohsiung, Taiwan2005 12 4 2005 5 27 27 25 11 2004 12 4 2005 Copyright © 2005 Yu et al; licensee BioMed Central Ltd.2005Yu 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
Interferon-alpha monotherapy is effective in less than one-third patients with chronic hepatitis C. The dose-effect, tolerability and durability of interferon-alpha treatment and its long-term effect on the prevention of cirrhosis and hepatocellular carcinoma in naïve Taiwanese patients with chronic hepatitis C have not been well investigated. We conducted the present cohort study treated with high and standard interferon-alpha to illustrate the issues.
Methods
We performed a long-term virologic and histological follow-up of 214 chronic hepatitis C patients treated with interferon-alpha, 3 million units (3-MU, n = 80) or 6-MU (n = 134) thrice weekly for 24 weeks, in Taiwan between 1992 and 2001.
Results
There was no difference in the incidence of discontinuation between 3-MU and 6-MU groups (4/80, 5.0% versus 10/134. 7.5%). The 6-MU group had similar incidence of adverse events with the 3-MU group, except that 6-MU group had significantly higher incidence of psychological manifestations, mainly presented as irritability. The rates of sustained virological response (SVR) were significantly higher in 6-MU regimen (37.1%) than in 3-MU regimen (23.7%, p < 0.05) in per protocol analysis. Based on multivariate analysis, baseline viral load was strongly associated with SVR, followed by hepatitis C virus genotype, interferon-alpha regimen, and liver fibrosis. A histological improvement in necroinflammatory activity, but not in fibrosis was observed in the follow-up biopsy performed 0.5–5.5 years (mean: 1.9 years, n = 51) after end-of-treatment. Among patients without SVR, there was more activity improvement in 6-MU group. The durability of SVR was 100% (18/18) and 97.8% (45/46) for 3-MU and 6-MU group, respectively, in a mean follow-up period of 6.81 years (5.25–9.18 years). For 163 baseline non-cirrhotic patients, 9 of 84 (10.7%) non-responders and 3 of 79 (3.8%) sustained responders progressed to cirrhosis during a mean follow-up period of 5.52 and 5.74 years, respectively (p = 0.067, Kaplan-Meier survival analysis, log-rank test). For all 200 patients, hepatocellular carcinoma was detected in 12 of 113 (10.6%) non-responders and one of 87 (1.1%) sustained responders during a mean follow-up period of 5.67 and 5.73 years, respectively (p < 0.01, Kaplan-Meier survival analysis, log-rank test).
Conclusion
We confirm the dose effect of interferon-alpha in chronic hepatitis C. Six-MU regimen had better efficacy than 3-MU regimen in virologic and histological responses. Both regimens had good tolerability and durability in Taiwan. Sustained response could reduce the incidence of cirrhotic change and hepatocarcinogenesis.
==== Body
Background
Hepatitis C virus (HCV) is the major etiologic agent in parenterally transmitted non-A non-B hepatitis and frequently causes persistent infection leading to chronic liver disease and primary hepatocellular carcinoma (HCC)[1,2]. Interferon-alpha (IFN-α) was the first approved therapy in the 1980s but resulted in a sustained virological response in only 8–20% of chronic hepatitis C patients treated with a standard regimen of IFN-α monotherapy, 3 million units (MU) thrice weekly for 24 weeks [3-8]. A number of factors have been considered in terms of their potential to predict the response to IFN-α therapy. These include infection with non-genotype 1b, lower levels of viremia and the absence of cirrhosis, which have been currently reported to be associated with a better response[2,9,10]. Several studies in Japan and Western countries have shown that higher dose (5–10 MU) of IFN could improve the efficacy on chronic hepatitis C[5,9,11]. Reviewing the Medline, however, only one limited study on the efficacy of IFN therapy for naïve patients of chronic hepatitis C has been reported before 1997 in Taiwan[4]. Furthermore, the dose effect, tolerability and durability of high-dose IFN therapy for naïve Taiwanese chronic hepatitis C patients have never been reported.
Since 1992, we have treated a number of Taiwanese patients with chronic hepatitis C with a standard IFN regimen, 3 MU thrice weekly for 24 weeks, and since 1995, with 6 MU thrice weekly for 24 weeks. We have followed these patients closely for 5–10 years, and we report herein the end-of-treatment and sustained response, and results of long-term follow-up to both regimens, as well as the dose effect, tolerability and durability of high-dose IFN therapy.
Methods
Patients
Two hundred and fourteen naïve chronic hepatitis C patients were enrolled in the study. All were positive for HCV antibodies (second-generation, enzyme-linked immunosorbent assay, Abbott, North Chicago, IL) and serum HCV RNA for at least 6 months. Patients with a concurrent hepatitis B virus, alcohol abuse (≧80 mL ethanol per day), overt hepatic failure, a current or past history of psychiatric condition, pregnancy, or with evidence of hepatocellular carcinoma were excluded. Two pathologists assessed all biopsy results, which were taken before IFN treatment, without knowledge of patients' clinical or laboratory data. Disease activity grade and fibrosis stage were quantitatively scored according to the histological activity index described by Knodell et al[12] Cirrhosis was diagnosed histologically in the presence of a stage 4 of fibrosis or in the absence of a liver biopsy, by a compatible ultrasonographic and clinical picture. A follow-up liver biopsy was performed in 51 patients. The present study was approved by the ethics committee of Kaohsiung Medical University Hospital. After they had given their informed consent, all patients were treated with recombinant IFN-α 2a (n = 40), IFN-α 2b (n = 102) or lymphoblastoid IFN-α n1 (n = 72), given intramuscularly, in a dose of 3-or 6-MU thrice weekly for 24 weeks. Group allocation was chronological and consecutive rather than randomized, because, before 1994, all patients were given a 3-MU and then a 6-MU IFN-α regimen. Fourteen patients (4 in 3-MU group, 5.0% and 10 in 6-MU group, 7.5%), who received IFN-α therapy less than 12 weeks, were excluded from the present study (Table 1). The presence of HCV RNA in the serum was assessed every three months for 12 months, and then every 12 months. End-of-treatment virological responder (ETVR) was defined as patients showing clearance of HCV RNA at the end of treatment. Others were defined as non-ETVR. Sustained virological responder (SVR) was defined as patients showing clearance of HCV RNA by at the end-of-treatment and 6 months after end-of-treatment. The others were classified as non-SVR. Histological improvement and worsening was defined as a ≧2-point decrease and increase in the total necroinflammatory scores between paired biopsies, respectively.
Table 1 Incidence of discontinuation and adverse events
3-MU group, No. (%) 6-MU group, No. (%) P value
Patient number 80 134
Discontinuation 4 (5.0) 10 (7.5) NS
Adverse event 3 (3.8) 6 (4.5) NS
Insufficient response 1 (1.3) 1 (0.7) NS
Laboratory abnormality 0 (0) 1 (0.7) NS
Economic problem 0 (0) 2 (1.5) NS
Adverse event
Flu-like symptoms† 49 (61.3) 85 (63.4) NS
Gastrointestinal manifestations‡ 19 (23.8) 35 (26.1) NS
Psychological manifestations§ 34 (41.0) 74 (55.2) 0.028
Alopecia 13 (16.3) 30 (22.4) NS
Dermatological manifestations¶ 11 (13.8) 19 (14.2) NS
Note: 3-MU group: interferon-α 3 million units thrice weekly for 24 weeks.
6-MU group: interferon-α 6 million units thrice weekly for 24 weeks.
NS: not significant.
† Including fatigue, headache, pyrexia, myalgia, and rigors.
‡ Including nausea, vomiting, anorexia, and diarrhea.
§Including irritability, depression, and insomnia.
¶Including dermatitis, and pruritus.
Detection/quantification of serum HCV RNA and genotyping
Detection of serum HCV RNA was performed using a standardized automated qualitative reverse transcription polymerase chain reaction assay (RT-PCR, COBAS AMPLICOR Hepatitis C Virus Test, version 2.0; Roche, Branchburg, NJ, USA). The detection limit was 50 IU/mL. HCV genotypes 1a, 1b, 2a, 2b and 3a were determined by amplification of the core region using genotype-specific primers described by Okamoto et al[13]. Serum HCV RNA levels were measured by using the branched DNA assay (Quantiplex HCV RNA 2.0, Bayer, Emeryville, CA), performed strictly in accordance with the manufacturer's instructions. The quantification range was 0.2 to 120 million equivalents (Meq) of HCV RNA per ml.
Statistical analyses
Frequency was compared between groups using the chi-square test with Yate's correction or Fisher's exact test. Group means were compared using the Student t-test. Serum HCV RNA levels were expressed as the mean ± standard deviation after logarithmic transformation of original values. Stepwise logistic regression was used to analyze factors associated with IFN-α response. Comparisons of paired liver histology were carried out with two-sample Wilcoxon's signed rank test. Preventive effects of antiviral therapy on progression of chronic HCV infection to liver cirrhosis (LC) and HCC were analyzed by using Kaplan-Meier survival analysis.
Results
Adverse effect
There was no difference of incidence of discontinuation of treatment between 3-MU and 6-MU groups (Table 1). The incidence of adverse events was similar between two groups in the presentation of flu-like symptoms, gastrointestinal and dermatological manifestations, and alopecia. However, patients in 6-MU group had significantly higher rate of psychological manifestation than those in 3-MU group did (74/134, 55.2% versus 34/80, 41.0%, p = 0.028, Table 1).
SVR and factors predicting SVR
Seventy-six patients received 3-MU IFN-α and 124 received 6-MU, thrice weekly for 24 weeks. All were followed-up for at least 6 months after end of treatment. Baseline clinical and laboratory data were not different in the two groups (Table 2). At the end of IFN-α therapy, 36 patients (47.4%) in 3-MU group and 103 (83.1%) had ETVR, the difference being highly significant (P < 0.0001). Of those with ETVR, 50% (18/36) in 3-MU group and 55.3% (57/103) in 6-MU group had reappearance of serum HCV RNA within 6 months after end-of-treatment. The difference was not significant. The SVR was significantly higher in 6-MU group (37.1%, 46/124) than in 3-MU group (23.7%, 18/76, P < 0.05).
Table 2 Baseline demographic and clinical features of 200 chronic hepatitis C patients with interferon-alpha therapy
Total 3-MU group† No. (%) 6-MU group‡ No. (%)
Patient Number 200 76 124
Gender
Male 103 34(44.7) 69(55.6)
Female 97 42(55.3) 55(44.4)
Age (year) 47.3 ± 10.4 45.3 ± 11.7
History of transfusion
No 141 56(73.7) 85(68.5)
Yes 59 20(26.3) 39(31.5)
Liver histopathology
Total score of necroinflammatory activity 4.48 ± 2.56 4.23 ± 2.46
Fibrosis score
F3–4 70 30(39.5) 40(32.3)
F0–2 130 46(60.5) 84(67.7)
Pretreatment ALT value (U/L) 117.4 ± 102.8 104.1 ± 95.4
Pretreatment HCV RNA level (log equivalent/mL) 5.97 ± 0.59 6.01 ± 0.68
HCV genotype
1b 80 27(35.5) 53(42.7)
Non-1b 120 49(64.5) 71(57.3)
Interferon preparation
Recombinant IFN-α-2a 37 10(13.2) 27(21.8)
Recombinant IFN-α-2b 95 36(47.4) 59(47.6)
Lymphoblastoid IFN-α-n1 68 30(39.5) 38(30.6)
† 3-MU group: interferon-α 3 million units thrice weekly for 24 weeks.
‡ 6-MU group: interferon-α 6 million units thrice weekly for 24 weeks.
In univariate analysis, SVR was not related gender, age, history of transfusion, the necroinflammatory activity of liver histopathology, or preparations of IFN-α (Table 3). A significant positive association for SVR was found for pretreatment alanine aminotransferase value and IFN-α dose, and a negative association with severe fibrosis (score 3 or 4), infected with HCV genotype 1b and pretreatment serum levels of HCV RNA. Based on multivariate logistic regression analysis, pretreatment HCV RNA level was strongly associated with SVR, followed by infected HCV genotype, IFN-α regimen, and severity of liver fibrosis (table 4).
Table 3 Factors associated with sustained viral response to interferon-alpha therapy
Total No. Non-SVR† No. (%) SVR† No. (%) P value
Patient Number 200 136(68.0) 64(32.0)
Gender NS‡
Male 103 69(70.0) 34(33.0)
Female 97 67(69.1) 30(30.9)
Age (year) 47.6 ± 10.0 44.2 ± 13.5 NS
History of transfusion NS
No 141 97(68.8) 44(31.2)
Yes 59 39(66.1) 20(33.9)
Liver histopathology
Total score of necroinflammatory activity 4.10 ± 2.59 4.21 ± 2.31 NS
Fibrosis score <0.01
F3–4 70 57(81.4) 13(18.6)
F0–2 130 79(60.8) 51(39.2)
Pretreatment ALT value (IU/L) 87.8 ± 86.5 143.5 ± 153.5 <0.01
Pretreatment serum level of HCV RNA§ 6.17 ± 0.61 5.62 ± 0.56 <0.0001
HCV genotype
1b 80 63(78.8) 17(21.3) <0.01
Non-1b 120 73(60.8) 47(39.2)
Interferon preparation NS
Recombinant IFN-α-2a 37 29(78.4) 8(21.6)
Recombinant IFN-α-2b 95 62(65.3) 33(34.7)
Lymphoblastoid IFN-α-n1 68 45(66.2) 23(33.8)
Interferon regimen <0.05
3-MU group¶ 76 58(76.3) 18(23.7)
6-MU group¶ 124 78(62.1) 46(37.1)
† SVR: sustained virological response.
‡ NS: not significant.
§Presented as log (equivalent/mL).
¶ 3-MU group: interferon-α 3 million units thrice weekly for 24 weeks; 6-MU group: interferon-α 6 million units thrice weekly for 24 weeks.
Table 4 Multivariate logistic regression analysis of factors associated with sustained viral response to interferon-alpha therapy in chronic hepatitis C patients
Variable† Odds ratio 95% CI‡
Pretreatment HCV RNA level§ Per 1 log increase 0.179 0.096–0.336
HCV genotype 1b = 1; non-1b = 0 0.324 0.151–0.695
IFN-α regimen 6-MU = 1; 3-MU = 0¶ 2.310 1.070–4.990
Stage of fibrosis F3–4 = 1; F0–2 = 0 0.408 0.104–0.892
† Factors included, age, sex, history of transfusion, pretreatment serum level of alanine aminotransferase and hepatitis C virus RNA, hepatitis C virus genotype, total necroinflammatory activity score, fibrosis score, interferon preparation, and interferon dose.
‡ CI: confidence interval.
§ HCV: hepatitis C virus.
¶ 3-MU group: interferon-α 3 million units thrice weekly for 24 weeks; 6-MU group: interferon-α 6 million units thrice weekly for 24 weeks.
Dose effect on SVR according to virological risk factors
Since HCV genotype and baseline viral loads were important predictors for SVR to IFN-α, we divided 200 patients into three virological unfavorable risk groups according to the infected HCV genotype and pretreatment HCV RNA levels[14]: low-risk group, absence of both virological risk factors (genotype 1b and pretreatment HCV RNA levels > 0.65 Meq/mL); medium-risk group, presence of only one virological risk factor; and high-risk group, presence of both two virological risk factors. The 6-MU group had significantly higher ETVR than the 3-MU group in all of the three risk groups (p < 0.01, <0.001, and <0.01 for low-, medium-, and high-risk group, respectively, Table 5). The relapsed rate did not differ between 3-MU and 6-MU groups whatever the risk groups. In medium-risk group, patients treated with 6-MU had significantly higher SVR than those with 3-MU (29.5% vs. 7.7%, p < 0.05). However, SVR did not differ between 3-MU and 6-MU groups both in low- and high-risk groups.
Table 5 Dose effect of interferon-alpha in chronic hepatitis C patients Dose effect of interferon-alpha on end-of-treatment virological response, virological relapse, and sustained virological response in chronic hepatitis C patients with low-, medium-, and high- virological risk
Low-risk patients, n/N (%)† Medium-risk patients, n/N (%)‡ High-risk patients, n/N (%)§
3-MU group 6-MU group 3-MU group 6-MU group 3-MU group 6-MU group
End-of-treatment virological response 14/19 (73.7)* 32/32 (100)* 19/39 (48.7)** 50/61 (82.0)** 4/18 (22.2)* 21/31 (67.7)*
Virological relapse 1/14 (7.1) 8/32 (25.0) 16/19 (84.0) 32/50 (64.0) 2/4 (50.0) 17/21 (81.0)
Sustained virological response 13/19 (68.4) 24/32 (75.0) 3/39 (7.7)* 18/61 (29.5)* 2/18 (11.1) 4/31 (12.9)
† Low-risk patients, neither genotype 1b nor pretreatment HCV RNA levels>0.65 Meq/mL.
‡ Medium-risk patients, either genotype 1b or pretreatment HCV RNA levels>0.65 Meq/mL.
§High-risk patients, both genotype 1b and pretreatment HCV RNA levels>0.65 Meq/mL.
*, P < 0.01; **, P < 0.001.
Paired histological examination
A follow-up liver biopsy was performed 0.5–5.5 years (mean ± SD, 1.9 ± 1.4 years) after end of treatment in 51 patients. A histological improvement with a significant decrease in the mean scores for all necroinflammatory activity (periportal, intralobular, portal inflammation and total necroinflammatory score), but not for fibrosis was observed in the follow-up biopsy (Table 6). The necroinflammatory activity improved in 68.2%, remained stable in 27.3%, and worsened in 4.6% for patients achieved a SVR, in contrast to 37.9%, 34.5%, and 27.6%, respectively, for those with non-SVR (P < 0.05, Fig. 1). Among patients achieving a SVR, there was no difference in histological response between 3-MU and 6-MU groups. Among patients without SVR, there was less frequently worsening of necroinflammatory activity (17.4%) and more activity improvement (47.8%) in 6-MU group, when compared with 3-MU group (66.7% and 0%, respectively, P < 0.01).
Table 6 Mean liver histological scores before and after interferon-alpha treatment in 51 Taiwanese chronic hepatitis C patients
Before treatment Follow-up† P value‡
Periportal inflammation 1.25 ± 1.47 0.84 ± 1.27 0.0419
Intralobular necrosis 1.02 ± 1.24 0.49 ± 0.92 0.0129
Portal inflammation 1.88 ± 1.24 1.16 ± 1.32 0.0030
Total necroinflammatory score 4.08 ± 2.86 2.51 ± 3.06 0.0037
Fibrosis 1.12 ± 1.09 1.10 ± 1.22 0.8485
† Mean follow-up time was 1.9 year (range 0.5–5.5 years) after end of treatment.
‡ Wilcoxon's signed rank test.
Figure 1 Impact of sustained virological response (SVR) to interferon-alpha therapy on necroinflammatory activity of liver histology. Impact of SVR to interferon-alpha therapy (left side) and of interferon regimens (3-MU versus. 6-MU groups, middle and right side) on necroinflammatory activity progression between paired liver histology.
Long-term follow-up for viral status and the development of LC and HCC
During the follow-up period, reappearance of serum HCV RNA was found in one patient, who has achieved SVR in 6-MU group, at the 57th month. The durability of SVR was 100% and 97.8% for 3-MU and 6MU group, respectively, in a mean follow-up period of 6.81 years (5.25–9.18 years). The SVR remained similar among three IFN preparations at the end of the 10 years.
Of 136 non-responders in the present study, 50 received a second course of IFN-based treatment. Three of 11 (27.3%) with IFN-α monotherapy and 20 of 39 (51.3%) with IFN-α/ribavirin combination therapy achieved a second SVR. Clinical evaluation, liver biochemistry, α-fetoprotein, and abdominal sonography were performed every 3–6 months for patients without baseline cirrhosis, and every 2–3 months for those with baseline cirrhosis during the follow-up period.
For 163 baseline non-cirrhotic patients, 9 of 84 (10.7%) non-responders and 3 of 79 (3.8%) sustained responders progressed to cirrhosis during a mean follow-up period of 5.52 and 5.74 years, respectively. The difference was borderline significant by using Kaplan-Meier survival analysis (p = 0.067, figure 2, left side). For all 200 patients, HCC was detected in 12 of 113 (10.6%) non-responders and one of 87 (1.1%) sustained responders during a mean follow-up period of 5.67 and 5.73 years, respectively. The difference was significant by using Kaplan-Meier survival analysis (p < 0.01, figure 2, right side).
Figure 2 Cumulative incidence of liver cirrhosis (LC) and hepatocellular carcinoma (HCC) among patients treated with interferon-alpha. Left side, cumulative incidence of LC among 163 baseline non-cirrhotic patients with sustained virological response (SVR, solid line) and without SVR (dotted line) to interferon-alpha therapy. P = 0.067 by the log-rank test. The number of LC events and patients at risk at each time point are shown below the graph. Right side, cumulative incidence of HCC among 200 patients with SVR (solid line) and without SVR (dotted line) to interferon-alpha therapy. P < 0.01 by the log-rank test. The number of HCC events and patients at risk at each time point are shown below the graph.
Discussion
In the present study, we have investigated the dose effect of IFN-α therapy on a group of Taiwanese patients with chronic hepatitis C in a 10-year follow-up. Our results demonstrated that 6 MU IFN-α, thrice weekly for 24 weeks, had better efficacy (2.3-fold) than 3 MU, thrice weekly for 24 weeks, in both virological and histological responses. Improved efficacy was mainly seen in subgroups of patients with one of the two unfavorable virologic factors, the medium-risk group, in the current study. The 6-MU regimen was as tolerable as standard 3-MU regimen for Taiwanese patients. Both regimens had good durability of SVR.
Treatment with IFN-α was the first approved therapy but a SVR could be achieved in only 8–20% of chronic hepatitis C patients treated with a standard regimen of IFN-α monotherapy (3 MU thrice weekly for 24 weeks) [3-8]. Therefore, how to improve the efficacy of IFN-α therapy is the concern of many studies. Until recently, 48 weeks of IFN-α, 3 MU thrice weekly was recommended for the treatment of chronic hepatitis C[5]. Concerning the suffering from long-term IFN therapy, we conducted higher dose of IFN regimen, 6 MU thrice weekly for 24 weeks, which total dosage was equal to that of 3 MU, thrice weekly for 48 weeks. In the present study, 6-MU regimen was observed to achieve significantly higher rates of ETVR and SVR than 3-MU regimen, but had no benefit on reducing the relapse rate. In contrast to the duration effect on improvement of SVR by reducing the relapse rate after end-of-treatment, our results implicated the dose effect on the initial response of HCV to IFN therapy[5,7,8]. High dose regimen as well as induction therapy could result in early clearance of HCV viremia, which is very important in the prediction of IFN response[10,15,16].
In the present study, the incidence of discontinuation and most of the major adverse events were similar between standard dose and higher dose of IFN therapy, despite that the higher dose of IFN regimen had significantly higher incidence of psychological manifestations. However, the psychological manifestations, mainly presented as irritability, could be controlled well with minor tranquillizers.
Consistent with previous reports[2,3,9,10,14], we confirmed the associations of pretreatment serum HCV RNA levels, HCV genotype 1b and presence of cirrhosis with response to IFN-α treatment in chronic hepatitis C patients by using stepwise logistic regression model. According to previous studies for Taiwanese patients, we divided our patients into three virological risk groups based on the presence/absence of HCV genotype 1b and the baseline viral loads greater than the cut-off point, 0.65 Meq/mL, or not[14]. As observed in the present study, 6-MU regimen could achieve significantly higher rate of ETVR but could not decrease the relapse rate in all of the three risk groups. Although our logistic regression model predicted a 2.3-times increase in probability for a SVR to occur in the 6-MU regimen versus the 3-MU regimen, the significantly better efficacy of 6-MU regimen on SVR was only observed in patients with one virological risk factor. Therefore, how to enhance the ETVR for the medium- and high-risk groups and to reduce the relapse rates for all of the three risk groups will be the key to improve SVR. Extended therapy and/or combination with ribavirin might be considered to overcome the unfavorable virological risk factors[16,17]. Since the unfavorable virological risk factors are unchangeable, tailored regimens of IFN with or without ribavirin combination therapy according to the virological status are the only way to improve the efficiency of HCV eradication, as recommended in recent consensus on management of chronic hepatitis C[18,19].
In accordance with previous studies [20-22], a significant decrease in necroinflammatory activity of liver histology was observed in the current study. However, our material was unable to show whether the IFN treatment resulted in cessation of fibrogenesis, regardless of IFN regimen and virological response (data not shown). Indeed, no firm conclusion can be drawn about whether the fibrosis remaining at long-term follow-up was irreversible and established or whether it will diminish with even longer follow-up[22]. Further analysis of histological response, stratified by IFN regimen and SVR, we found that 6-MU regimen could achieve histological improvement not only in SVR but also in non-SVR. By contrast, most of non-SVR treated with 3-MU regimen had histological worsening. These results presumed the prominent benefit of high dose IFN on the treatment of chronic hepatitis C.
Similar to other reports[23,24], the long-term prognosis was excellent in this well-defined material of SVR. Serum HCV RNA was persistently undetectable in 63 of 64 patients with SVR. The only one patient, who had reappearance of serum HCV RNA concomitant with abnormal ALT level at the 57th month of follow-up, is a resident of an HCV hyperendemic area, Tzukuan[25]. Whether this is reinfection[26] or recurrence[23,24] of HCV remains to be clarified by phylogenetic analysis of viral genome. Also our results confirm that IFN therapy, when associated with response, reduces the incidence of LC and HCC among chronic hepatitis C patients[27].
A greater efficacy in treatment of chronic hepatitis C was observed in combination with IFN-α and ribavirin[24], which has been available in Taiwan since August 1998 and becomes to be the recommendations for chronic hepatitis C patients[18,19]. More recently, combination of pegylated IFN-α plus ribavirin shows more effective and convenient and may replace the current standard of IFN-α plus ribavirin[28,29]. The results of current study could provide decision-making information for future therapeutic strategies of individualizing dose and duration of standard or pegylated IFN-α treatment in combination with ribavirin according to the baseline virological predictors to improve the efficiency of HCV eradication. However, pegylated IFN-α and ribavirin combination therapy is expensive and might carry potential side effects. The present higher dose regimen of IFN monotherapy might be suggested for patients who are contraindicated to ribavirin and/or pegylated IFN-α therapy[30].
Conclusion
We confirm the dose effect of IFN in chronic hepatitis C. Six-MU regimen had better efficacy than standard 3-MU regimen in both of virological and histological responses, and was as tolerable as 3-MU regimen. Nearly all sustained virological responders had a durable response and the sustained response had a preventive effect on the development of HCV-associated LC and HCC. Since the unfavorable predictors, high pretreatment HCV RNA levels, infected with genotype 1b, and presence of cirrhosis, are unchangeable, adjustment of IFN dose and duration according to the unfavorable factors is important to achieve a better efficacy/risk ratio. Our results could provide decision-making information for future therapeutic strategies of individualizing dose and duration of standard or pegylated IFN-α treatment in combination with ribavirin according to baseline predictors.
List of abbreviations
Hepatitis C virus: HCV; interferon-alpha: IFN-α; million units: MU; liver cirrhosis: LC, hepatocellular carcinoma: HCC, end-of-treatment virologic responder: ETVR, sustained virologic responder: SVR; reverse transcription polymerase chain reaction assay: RT-PCR; million equivalent: Meq.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MLY carried out the laboratory work and molecular virologic studies, participated in the collection and clinical evaluation of patients, performed the statistical analysis, and drafted the manuscript. CYD, SCC and LPL participated in the laboratory work and molecular virologic studies, and in the collection and clinical evaluation of patients. MYH, ZYL, MYH, JFT and LYW, participated in the collection and clinical evaluation of patients. WYC participated in the design of the study. WLC 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
This study was supported by the Taiwan Liver Research Foundation.
==== Refs
Alter MJ Margolis HS Krawczynski K Judson FN Mares A Alexander WJ Hu PY Miller JK Gerber MA Sampliner RE The natural history of community-acquired hepatitis C in the United States. The Sentinel Counties Chronic non-A, non-B Hepatitis Study Team N Engl J Med 1992 327 1899 1905 1280771
Lauer GM Walker BD Hepatitis C virus infection N Engl J Med 2001 345 41 52 11439948 10.1056/NEJM200107053450107
Martinot-Peignoux M Boyer N Pouteau M Castelnau C Giuily N Duchatelle V Auperin A Degott C Benhamou JP Erlinger S Predictors of sustained response to alpha interferon therapy in chronic hepatitis C J Hepatol 1998 29 214 223 9722202 10.1016/S0168-8278(98)80006-8
Hwang SJ Chan CY Lu RH Wu JC Lee SD Randomized controlled trial of recombinant interferon-alpha 2b in the treatment of Chinese patients with chronic hepatitis C J Interferon Cytokine Res 1995 15 611 616 7553231
Poynard T Leroy V Cohard M Thevenot T Mathurin P Opolon P Zarski JP Meta-analysis of interferon randomized trials in the treatment of viral hepatitis C: effects of dose and duration Hepatology 1996 24 778 789 8855176
Lindsay KL Therapy of hepatitis C: overview Hepatology 1997 26 71S 77S 9305668 10.1002/hep.510260713
Carithers RLJ Emerson SS Therapy of hepatitis C: meta-analysis of interferon alfa-2b trials Hepatology 1997 26 83S 88S 9305670 10.1002/hep.510260715
Thevenot T Regimbeau C Ratziu V Leroy V Opolon P Poynard T Meta-analysis of interferon randomized trials in the treatment of viral hepatitis C in naive patients: 1999 update J Viral Hepat 2001 8 48 62 11155152 10.1046/j.1365-2893.2001.00271.x
Shiratori Y Kato N Yokosuka O Imazeki F Hashimoto E Hayashi N Nakamura A Asada M Kuroda H Tanaka N Arakawa Y Omata M Predictors of the efficacy of interferon therapy in chronic hepatitis C virus infection. Tokyo-Chiba Hepatitis Research Group Gastroenterology 1997 113 558 566 9247476
Yu ML Chuang WL Dai CY Chen SC Lin ZY Hsieh MY Wang LY Chang WY Clinical evaluation of the automated COBAS AMPLICOR HCV MONITOR test version 2.0 for quantifying serum hepatitis C virus RNA and comparison to the quantiplex HCV version 2.0 test Journal Of Clinical Microbiology 2000 38 2933 2939 10921954
Hakozaki Y Shirahama T Katou M Nakagawa K Oba K Mitamura K A controlled study to determine the optimal dose regimen of interferon-alpha 2b in chronic hepatitis C Am J Gastroenterol 1995 90 1246 1249 7639224
Knodell RG Ishak KG Black WC Chen TS Craig R Kaplowitz N Kiernan TW Wollman J Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis Hepatology 1981 1 431 435 7308988
Okamoto H Tokita H Sakamoto M Horikita M Kojima M Iizuka H Mishiro S Characterization of the genomic sequence of type V (or 3a) hepatitis C virus isolates and PCR primers for specific detection J Gen Virol 1993 74 ( Pt 11) 2385 2390 7504073
Yu ML Chuang WL Chen SC Lin ZY Hsieh MY Wang LY Chang WY Clinical application of the Quantiplex HCV RNA 2.0 and Amplicor HCV Monitor assays for quantifying serum hepatitis C virus RNA Journal Of Clinical Pathology 1999 52 807 811 10690168
Neumann AU Lam NP Dahari H Gretch DR Wiley TE Layden TJ Perelson AS Hepatitis C viral dynamics in vivo and the antiviral efficacy of interferon-alpha therapy Science 1998 282 103 107 9756471 10.1126/science.282.5386.103
Layden TJ Principles of interferon induction therapy Am J Med 1999 107 71S 73S 10653462 10.1016/S0002-9343(99)00387-3
Chemello L Bonetti P Cavalletto L Talato F Donadon V Casarin P Belussi F Frezza M Noventa F Pontisso P Randomized trial comparing three different regimens of alpha-2a-interferon in chronic hepatitis C. The TriVeneto Viral Hepatitis Group Hepatology 1995 22 700 706 7657272 10.1016/0270-9139(95)90286-4
Consensus statements on the prevention and management of hepatitis B and hepatitis C in the Asia-Pacific region. Core Working Party for Asia-Pacific Consensus on Hepatitis B and C J Gastroenterol Hepatol 2000 15 825 841 11022822 10.1046/j.1440-1746.2000.02324.x
EASL International Consensus Conference on hepatitis C. Paris, 26-27 February 1999. Consensus statement J Hepatol 1999 31 Suppl 1 3 8 10622553 10.1016/S0168-8278(99)80367-5
Poynard T Bedossa P Chevallier M Mathurin P Lemonnier C Trepo C Couzigou P Payen JL Sajus M Costa JM A comparison of three interferon alfa-2b regimens for the long-term treatment of chronic non-A, non-B hepatitis. Multicenter Study Group N Engl J Med 1995 332 1457 1462 7739681 10.1056/NEJM199506013322201
Marcellin P Boyer N Gervais A Martinot M Pouteau M Castelnau C Kilani A Areias J Auperin A Benhamou JP Degott C Erlinger S Long-term histologic improvement and loss of detectable intrahepatic HCV RNA in patients with chronic hepatitis C and sustained response to interferon-alpha therapy Ann Intern Med 1997 127 875 881 9382365
Shiratori Y Imazeki F Moriyama M Yano M Arakawa Y Yokosuka O Kuroki T Nishiguchi S Sata M Yamada G Fujiyama S Yoshida H Omata M Histologic improvement of fibrosis in patients with hepatitis C who have sustained response to interferon therapy Ann Intern Med 2000 132 517 524 10744587
Larghi A Tagger A Crosignani A Ribero ML Bruno S Portera G Battezzati PM Maggioni M Fasola M Zuin M Podda M Clinical significance of hepatic HCV RNA in patients with chronic hepatitis C demonstrating long-term sustained response to interferon-alpha therapy J Med Virol 1998 55 7 11 9580879 10.1002/(SICI)1096-9071(199805)55:1<7::AID-JMV2>3.0.CO;2-7
McHutchison JG Gordon SC Schiff ER Shiffman ML Lee WM Rustgi VK Goodman ZD Ling MH Cort S Albrecht JK Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. Hepatitis Interventional Therapy Group N Engl J Med 1998 339 1485 1492 9819446 10.1056/NEJM199811193392101
Yu ML Chuang WL Chen SC Dai CY Hou C Wang JH Lu SN Huang JF Lin ZY Hsieh MY Tsai JF Wang LY Chang WY Changing prevalence of hepatitis C virus genotypes: Molecular epidemiology and clinical implications in the hepatitis C virus hyperendemic areas and a Tertiary Referral Center in Taiwan Journal of Medical Virology 2001 65 58 65 11505444 10.1002/jmv.2001
Kao JH Lai MY Chen PJ Chen DS Probable reinfection with hepatitis C virus in a chronic hepatitis C patient with a sustained response to combination therapy J Formos Med Assoc 2001 100 824 828 11802523
Yoshida H Shiratori Y Moriyama M Arakawa Y Ide T Sata M Inoue O Yano M Tanaka M Fujiyama S Nishiguchi S Kuroki T Imazeki F Yokosuka O Kinoyama S Yamada G Omata M Interferon therapy reduces the risk for hepatocellular carcinoma: national surveillance program of cirrhotic and noncirrhotic patients with chronic hepatitis C in Japan. IHIT Study Group. Inhibition of Hepatocarcinogenesis by Interferon Therapy Ann Intern Med 1999 131 174 181 10428733
Manns MP McHutchison JG Gordon SC Rustgi VK Shiffman M Reindollar R Goodman ZD Koury K Ling M Albrecht JK Peginterferon alfa-2b plus ribavirin compared with interferon alfa-2b plus ribavirin for initial treatment of chronic hepatitis C: a randomised trial Lancet 2001 358 958 965 11583749 10.1016/S0140-6736(01)06102-5
Fried MW Shiffman ML Reddy KR Smith C Marinos G Goncales FLJ Haussinger D Diago M Carosi G Dhumeaux D Craxi A Lin A Hoffman J Yu J Peginterferon alfa-2a plus ribavirin for chronic hepatitis C virus infection N Engl J Med 2002 347 975 982 12324553 10.1056/NEJMoa020047
Keating GM Curran MP Peginterferon-alpha-2a (40kD) plus ribavirin: a review of its use in the management of chronic hepatitis C Drugs 2003 63 701 730 12656650
| 15823212 | PMC1090579 | CC BY | 2021-01-04 16:28:15 | no | BMC Infect Dis. 2005 Apr 12; 5:27 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-27 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-291585751510.1186/1471-2334-5-29Case ReportTuberculous dilated cardiomyopathy: an under-recognized entity? Agarwal Ritesh [email protected] Puneet [email protected] Anshu [email protected] Nandita [email protected] Dheeraj [email protected] Department of Pulmonary Medicine, Post-Graduate Institute of Medical Education and Research, Sector-12, Chandigarh-160012, India2 Department of Histopathology, Post-Graduate Institute of Medical Education and Research, Sector-12, Chandigarh-160012, India2005 27 4 2005 5 29 29 3 3 2005 27 4 2005 Copyright © 2005 Agarwal et al; licensee BioMed Central Ltd.2005Agarwal 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
Tuberculosis (TB) is a common public health problem in many parts of the world. TB is generally believed to spare these four organs-heart, skeletal muscle, thyroid and pancreas. We describe a rare case of myocardial TB diagnosed on a post-mortem cardiac biopsy.
Case presentation
Patient presented with history suggestive of congestive heart failure. We describe the clinical presentation, investigations and outcome of this case, and review the literature on the involvement of myocardium by TB.
Conclusion
Involvement of myocardium by TB is rare. However it should be suspected as a cause of congestive heart failure in any patient with features suggestive of TB. Increasing recognition of the entity and the use of endomyocardial biopsy may help us detect more cases of this "curable" form of cardiomyopathy.
==== Body
Background
Tuberculosis (TB) is generally believed to spare these four organs-heart, thyroid, pancreas and skeletal muscle. Involvement of myocardium by TB is rare, and generally occurs in conjunction with pericardial involvement. Isolated myocardial TB is a rare finding, and definitive diagnosis during life requires a myocardial biopsy. Herein we describe a patient who presented with features suggestive of congestive heart failure, and was finally diagnosed to have myocardial TB on a post-mortem cardiac biopsy.
Case presentation
A 25-year old female presented to the emergency department with a one week history of low grade fever, increasing cough and dyspnea on exertion. At presentation she had dyspnea at rest, orthopnea and paroxysmal nocturnal dyspnea. Around two and a half months before the present illness, she had low grade fever, anorexia and weight loss, cough with expectoration and hemoptysis. She was investigated at another center and a sputum smear for acid-fast bacilli was positive. She was started on antituberculous therapy with which she improved symptomatically. She experienced weight gain, and her appetite became normal. A sputum smear performed after two months for acid-fast bacilli was negative. On examination, the patient was conscious and afebrile with a pulse rate of 128 beats/minute, blood pressure of 100/60 mm Hg and a respiratory rate of 38/minute. She had bilateral pitting pedal edema. Examination of the cardiovascular system revealed tachycardia, elevated jugular venous pressure, diffuse apical impulse and left ventricular third heart sound. Auscultation of the chest showed bibasal inspiratory crackles. She also had tender hepatomegaly and free fluid in the abdomen. Her oxygen saturation was 85% on pulse oximetry. Arterial blood gases showed type 1 respiratory failure [(FiO2 0.5) – pH 7.48, PaO2 8.2 kPa, PaCO2 3.6 kPa, HCO3 21 mEq/L]. She was started on supplemental oxygen, intravenous morphine, nitroglycerin, furosemide and dobutamine, oral captopril, stress ulcer and deep venous thrombosis prophylaxis. Chest X-ray (Figure 1) done two months ago showed right upper lobe consolidation with normal cardiac silhouette. A chest radiograph (Figure 2) performed at admission revealed cardiomegaly and bilateral alveolar opacities. Electrocardiogram revealed sinus tachycardia and non-specific ST-T changes in the lateral leads. Echocardiography was performed which revealed global hypokinesia, enlarged left atrium and left ventricle, mild mitral regurgitation, severe left ventricular systolic dysfunction with an ejection fraction of 20%; left ventricular end systolic and end diastolic dimensions were 45 and 51 mm respectively. Right atrium and right ventricle were dilated with associated mild tricuspid regurgitation and pulmonary artery systolic pressure of 47 mm Hg. A provisional diagnosis of viral myocarditis was made and she was shifted to respiratory intensive care unit where noninvasive ventilation (NIV) was initiated with continuous positive airway pressure at 10 cm H2O. Arterial line was placed through the right radial artery and continuous electrocardiographic monitoring was performed. As the patient had no improvement with NIV and was becoming hypoxemic and agitated, she was given intravenous fentanyl and oral endotracheal intubation was performed. She was mechanically ventilated with assist/control mode at tidal volumes of 350 mL, rate of 20, PEEP of 10 cm H2O, and FiO2 of 1. Her peak and plateau pressures were 35 and 28 cm H2O respectively. She was sedated intermittently with intravenous fentanyl. Blood cultures, mycoplasma and legionella serology were sent and patient was also started on intravenous azithromycin. Other treatment measures and antituberculous therapy were continued. Biochemical investigations revealed hypoalbuminemia and prerenal azotemia; liver function tests, complete blood, coagulation profile and urinalysis were normal. With mechanical ventilation, oxygen saturation improved and oxygen requirements decreased to 0.4. She had recurrent episodes of nonsustained ventricular tachycardia and amiodarone infusion was started. However on the second day of admission she had a sudden episode of ventricular fibrillation and despite all resuscitative measures the patient could not be revived. Her husband did not give consent for an autopsy; however he agreed for a post-mortem cardiac biopsy, which revealed multi-focal areas of caseous myocardial necrosis, Ziehl-Neelson stain for acid-fast bacilli was positive (Figure 3). HIV serology received postmortem was nonreactive.
Figure 1 Chest radiograph showing right upper lobe consolidation (cardiac silhouette is normal).
Figure 2 Chest radiograph bilateral alveolar opacities and cardiomegaly.
Figure 3 Photomicrograph demonstrates large areas of necrosis with interspersed myocytes showing degeneration and regeneration [Hematoxylin & Eosin, original magnification × 33]. Inset shows photomicrograph demonstrating acid-fast bacilli [Ziehl-Neelson stain, original magnification × 33].
Conclusion
Involvement of heart in tuberculosis (TB) occurs in one to two percent of patients with tuberculosis [1,2]. The most common site involved is pericardium, and tuberculous involvement of the myocardium is exceedingly rare. The introduction of effective antituberculous therapy has further decreased the incidence [3]. The earliest report of myocardial TB was in 1664 by Maurocordat and second report in 1761 by Morgagni [4]. The myocardium can be affected either by direct extension or by retrograde lymphatic drainage from mediastinal nodes; direct spread from tuberculous pericarditis can also occur [5]. Moreover, during the hematogenous phase of dissemination of primary TB, any and every tissue and organ in the body is liable to seeding by mycobacteria and consequent pathological changes.
Myocardial TB is often not diagnosed during life, but if suspected the diagnosis can be established by an endomyocardial biopsy [6]. Three types of myocardial involvement have been described viz. tuberculomas of the myocardium with central caseation (seen in our patient), miliary tubercles of the myocardium complicating generalized miliary disease and the uncommon diffuse infiltrative type associated with tuberculous pericarditis [7]. The right heart, particularly the right atrium, is most often affected, probably because of the frequent involvement of the right mediastinal lymph nodes with consequent involvement of the myocardium [8], although right ventricle [3] and left ventricle [4] have been found to be involved most frequently in different series. Myocardial TB can manifest in various forms. Rhythm disturbances include supraventricular arrhythmias [3,6], ventricular arrhythmias [9] or varying degrees of conduction blocks [10], and sudden cardiac death is also described [5,11]. Right ventricular outflow tract obstruction [3,12,13], ventricular aneurysm [4,13], ventricular pseudoaneurysm [14], aortic insufficiency [15], coronary arteritis [10,13], or congestive heart failure [6,16,17] have also been described in literature. Recently, magnetic resonance imaging has also been used in the diagnosis of myocardial TB [18].
Antitubercular drugs are the cornerstone of therapy [6], and surgery [12,14-16] is indicated only in complicated cases. Paradoxically, our patient presented with features of congestive heart failure, and echocardiography showed severe left ventricular systolic dysfunction inspite of being on anti-tuberculous therapy which had resulted in bacteriological improvement of the pulmonary lesions. The reason for this is not clear, but one likely reason is probably enhanced immunogenicity of the host to tubercle bacilli. This has been called as a 'paradoxical response' [19]. In this regard, at least theoretically, glucocorticoids may have a role along with antituberculous therapy.
In conclusion, although myocardial involvement by tuberculosis is rare, it should be suspected as a cause of congestive heart failure in any patient with features suggestive of TB, as cases of myocardial TB almost always show evidence of TB at other sites [6]. Increasing recognition of the entity, and the use of endomyocardial biopsy may help us detect more cases of this "curable" form of cardiomyopathy especially in areas of high prevalence of TB. It may not be out of place to state.... Look and ye shall find....
Competing interests
The author(s) declare that they have no competing interests.
Authors' Contributions
RA was involved in patient care and drafting the manuscript
PM was involved in patient care and editing the manuscript
AA was involved in histopathological examination and editing the manuscript
NK was involved in histopathological examination
DG conceived the study and was also involved in patient care
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Anders JM Tuberculosis of the myocardium JAMA 1902 39 1081 1086
Fowler NO Tuberculous pericarditis JAMA 1991 266 199 203 10.1001/jama.266.1.99
Kapoor OP Marcarenhas E Rananaware MM Gadgil RK Tuberculoma of the heart. Report of 9 cases Am Heart J 1973 86 334 340 4199393 10.1016/0002-8703(73)90042-2
Rose AG Cardiac tuberculosis: a study of 19 patients Arch Pathol Lab Med 1987 111 422 426 3566473
Wallis PJW Branfoot AC Emerson PA Sudden death due to myocardial tuberculosis Thorax 1984 39 155 6701827
Bali HK Wahi S Sharma BK Anand IS Datta BN Wahi PL Myocardial tuberculosis presenting as restrictive cardiomyopathy Am Heart J 1990 120 703 706 2389712 10.1016/0002-8703(90)90036-W
Horn H Saphir O The involvement of the myocardium in tuberculosis: a review of the literature and report of three cases Am Rev Tuberc 1935 32 492 506
Maeder M Ammann M Rickli H Schoch OD Fever and night sweats in a 22-year-old man with a mediastinal mass involving the heart Chest 2003 124 2006 2009 14605080 10.1378/chest.124.5.2006
Behr G Palin HC Temperley JM Myocardial tuberculosis Br Med J 1977 1 951 851799
Kinare SG Interesting facets of cardiovascular tuberculosis Indian J Surg 1975 37 144 151
Kinare SG Deshmukh MM Complete atrioventricular block due to myocardial tuberculosis. Report of a case Arch Pathol 1969 88 684 687 5357722
Rawls WJ Shuford WH Logan WD Hurst JW Schlant RC Right ventricular outflow tract obstruction produced by a myocardial abscess in a patient with tuberculosis Am J Cardiol 1968 21 738 745 4230731 10.1016/0002-9149(68)90274-9
Kinare SG Bhatia BI Tuberculous coronary arteritis with aneurysm of the ventricular septum Chest 1971 60 613 616 5126198
Halim MA Mercer EN Guinn GA Myocardial tuberculoma with rupture and pseudoaneurysm formation: successful treatment Br Heart J 1985 54 603 604 4074594
Soyer R Brunet A Chavalier B Leroy J Morere M Redonnet M Tuberculous aortic insufficiency. Report of a case with successful surgical treatment J Thorac Cardiovasc Surg 1981 82 254 256 7253687
Wilbur EL Myocardial tuberculosis: a case of congestive cardiac failure Am Rev Tuberc 1938 38 769 776
Danbauchi SS Odigie VI Rafindadi AH Kalayi GD Mohammed I Tuberculous myocarditis: A case report Niger Postgrad Med J 2001 8 199 202 11922029
Jagia P Gulati GS Sharma S Goyal NK Gaikwad S Saxena A MRI features of tuberculoma of the right atrial myocardium Pediatr Radiol 2004 34 904 907 15221236 10.1007/s00247-004-1222-8
Chambers ST Hendrickse WA Record C Rudge P Smith H Paradoxical expansion of intracranial tuberculomas during chemotherapy Lancet 1984 2 181 184 6146749 10.1016/S0140-6736(84)90478-1
| 15857515 | PMC1090580 | CC BY | 2021-01-04 16:28:15 | no | BMC Infect Dis. 2005 Apr 27; 5:29 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-29 | oa_comm |
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-101585423010.1186/1471-2199-6-10Research ArticleRelative transcript quantification by Quantitative PCR: Roughly right or precisely wrong? Skern Rasmus [email protected] Petter [email protected] Frank [email protected] Marine Genome Biology, Institute of Marine Research, N-5817 Bergen, Norway2005 26 4 2005 6 10 10 1 12 2004 26 4 2005 Copyright © 2005 Skern et al; licensee BioMed Central Ltd.2005Skern 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
When estimating relative transcript abundances by quantitative real-time PCR (Q-PCR) we found that the results can vary dramatically depending on the method chosen for data analysis.
Results
Analyses of Q-PCR results from a salmon louse starvation experiment show that, even with apparently good raw data, different analytical approaches [1,2] may lead to opposing biological conclusions.
Conclusion
The results emphasise the importance of being cautious when analysing Q-PCR data and indicate that uncritical routine application of an analytical method will eventually result in incorrect conclusions. We do not know the extent of, or have a universal solution to this problem. However, we strongly recommend caution when analysing Q-PCR results e.g. by using two or more analytical approaches to validate conclusions. In our view a common effort should be made to standardise methods for analysis and validation of Q-PCR results.
==== Body
Communication
Reverse transcription (RT) followed by quantitative polymerase chain reaction (Q-PCR) is at present the most sensitive method for transcript abundance measurement. However, there are many sources of errors, both when purifying RNA, performing the RT reaction and during the PCR setup [3,4]. Q-PCR utilises optical measurement of generated amplicons to survey PCR amplifications. It is common to derive the initial template concentration from the number of amplification cycles required for a signal to reach a threshold chosen by the investigator [1,2,5]. In relative quantification the expression of a target gene is stated relative to a standard gene, which is assumed to be constitutively and uniformly expressed. One popular approach, the 2-ΔΔCT method, assumes ≈100% efficient target and standard gene PCR reactions given that the results conform to certain criteria [1,5]. In recognition of the fact that PCR efficiencies may vary between runs or between target and standard genes, other numerous methods have emerged that calculate template concentrations using amplification simulations or PCR efficiencies derived from CT values or fluorescence data [2,6-9]. We here present the results of a case study showing that the interpretation of results may vary dramatically with the chosen method for data analysis.
We have analysed results from a salmon lice (Lepeophtheirus salmonis) starvation experiment using the 2-ΔΔCT method [1] and the "DART method" adjusting for PCR efficiency differences [2]. When analysed using the 2-ΔΔCT method, our results show that LsTryp1 transcript levels decrease following starvation and return to normal adult level when the louse subsequently gets access to food (Fig. 1). The inclinations obtained when plotting ΔCT or CT against log RNA concentration do not indicate significant differences in PCR efficiencies between LsTryp1 and eEF1α (Table 1). When analysed using the "DART method" the results indicate that LsTryp1 transcript levels decrease 2–3 fold when lice are starved and remain low when lice subsequently get access to food (Fig. 1). The PCR efficiencies, calculated from at least 3 points for each reaction [2], indicated significant differences in PCR efficiency between eEF1α and LsTryp1 in starved and refed lice but not in unstarved lice (Table 1).
By intuition it appears that surveying PCR efficiencies using several measured fluorescence points from each PCR reaction, as done using the "DART method", is superior to using one point from each reaction, as done when comparing ΔCT values using the 2-ΔΔCT method. However, since PCR efficiencies calculated using the "DART method" exceed 100% in some instances, it is clear that this approach also has weaknesses. In the present example (Fig. 1) we would not have more confidence in one method than the other unless we had data from supplementary methods (e.g. microarrays) to support this. Consequently these data indicate that LsTryp1 transcript levels decrease when lice are starved, which is in accordance with the alleged digestive function of the encoded protein [10]. However, since the result varies between the "DART-method" and the 2-ΔΔCT method, we are unable to determine how transcription is regulated after lice resume feeding. Thus, despite the fact that both the 2-ΔΔCT method and the "DART-method" are theoretically sound given a number of assumptions [1,2], we may be mislead when these assumptions are not fulfilled.
All strategies for analysing Q-PCR data are based on a number of assumptions, and due to experimental errors none or few of these assumptions will be fulfilled entirely. Unfortunately, it is not always obvious when assumptions are broken to a degree that invalidates the conclusions. Since the sources of potential problems are diverse, no simple solution is available. Therefore we do not offer a universal analytical approach that can be applied to any given set of data and ensure a correct conclusion. Rather, we suggest investigators to urge caution when analysing results and hope that future discussions will lead to a more unified approach to Q-PCR data analysis and improved reliability of published results.
Methods
Salmon lice (Lepeophtheirus salmonis) were reared as earlier described [10]. After development to the adult stage, 15 lice were removed with forceps from their anaesthetised (80 μg/ml benzocaine) salmon hosts (Salmo salar) and 3 lice were stored in RNA later (Ambion). The remaining 12 lice were starved in incubators with flowing seawater for 14 days. After starvation, 3 lice were sampled and stored as described above, and the remaining 9 lice were put in a tank with uninfected salmon where they could settle on their salmon hosts and resume feeding. After 15 days on their new hosts 3 lice were sampled and stored as described above. The experimental procedures were carried out in accordance with national regulations for use of animals in scientific research.
The transcript levels of LsTryp1 [10] and the reference gene eEF1α [11] in 1 selected unstarved, starved and refed lice were determined by quantitative real time PCR carried out with 3 parallels at 5 sequential 2-fold dilutions as previously described [10]. The RNA purification protocol is previously described [10] and cDNA syntheses were performed using MultiScribe™ according to the manufacturers recommendations (Applied Biosystems). The Q-PCR results were analysed by the 2-ΔΔCT method as earlier described [10] and a method adjusting for PCR efficiency differences described by Peirson et al. [2]. The latter analysis was performed partially in the DART-PCR Excel spreadsheet [2]. When using the 2-ΔΔCT method, at least 2 parallels were required at each dilution. Parallels were removed when the CT value differed more than 0.3 (CT<32) or 0.4 (CT = 32) from the most similar parallel at the same dilution. At least 4 dilutions were required for each stage. The resulting data were calibrated to unstarved lice and analysed as described by Kvamme et al.[10]. When using the "DART-method", dilutions were removed when PCR efficiency differed significantly (one way ANOVA, α = 0.05) from the other dilutions. The signal corresponding to the initial template concentration (R0) was derived using the average PCR efficiency for LsTryp1 and eEF1α when the PCR efficiencies were not significantly different (one way ANOVA, α = 0.05). When the PCR efficiency differed significantly, R0 was calculated using individual gene specific mean efficiencies. The mean R0 for each dilution of LsTryp1 was normalised to corresponding eEF1α values. The normalised R0 values were calibrated to the values for unstarved lice. 95% confidence intervals (CI) were derived from normalised R0 values.
Authors' contributions
RS and PF conceived the study and designed the Q-PCR assay. RS carried out the assay and analyses, and wrote the first draft of the communication. PF contributed to development of the communication. FN provided expert input for the writing and supervised the study.
Acknowledgements
We would like to thank Dr. Mette Mauritzen, Dr. Kevin Glover and Bjørn Olav Kvamme for constructive criticism of the communication. This work has been supported by Intervet International bv and the Norwegian Research Council.
Figures and Tables
Figure 1 Q-PCR analysis. Transcript levels from the same Q-PCR runs analysed using the 2-ΔΔCT method and the DART-PCR Excel Spreadsheet. Error bars indicate 95% confidence intervals.
Table 1 Quality assessment of the retrieved data. Quality assessment of the retrieved data. For the 2-ΔΔCT results the table shows inclination and R2 for ΔCT plotted against log RNA concentration and inclinations and R2 for CT plotted against log RNA concentration for eEF1α and LsTryp1. For the DART-PCR results the table shows PCR-efficiencies for eEF1α and LsTryp1 calculated by DART-PCR and the p-value (one-way ANOVA) for the hypothesis that there is no difference between the efficiencies.
2-ΔΔCT results DART-PCR results
ΔCT vs. log [RNA] (R2) eEF1α CT vs. log [RNA] (R2) LsTryp1 CT vs. log [RNA] (R2) eEF1α LsTryp1 ANOVA
Unstarved -0.080 (0.16) -2.7664 (0.95) -2.7949 (0.99) 0.901 0.902 0.9863
Starved 0.065 (0.06) -2.8815 (0.99) -2.8171 (0.93) 0.839 0.491 0.0004
Refed -0.052 (0.07) -3.3109 (0.99) -3.3845 (0.99) 0.883 1.095 0.0004
==== Refs
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
Peirson SN Butler JN Foster RG Experimental validation of novel and conventional approaches to quantitative real-time PCR data analysis Nucleic Acids Res 2003 31 e73 12853650 10.1093/nar/gng073
Bustin SA Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems J Mol Endocrinol 2002 29 23 39 12200227 10.1677/jme.0.0290023
Freeman WM Walker SJ Vrana KE Quantitative RT-PCR: pitfalls and potential Biotechniques 1999 26 112 125 9894600
Livak KJ ABI Prims 7700 Sequencer detection system. User bulletin 2. PE Applied Biosystems 1997 1 36
Pfaffl MW A new mathematical model for relative quantification in real-time RT-PCR Nucleic Acids Res 2001 29 2002 2007 10.1093/nar/29.9.e45
Pfaffl MW Horgan GW Dempfle L Relative expression software tool (REST) for group-wise comparison and statistical analysis of relative expression results in real-time PCR Nucleic Acids Res 2002 30 1 10 11752241 10.1093/nar/30.9.e36
Liu W Saint DA Validation of a quantitative method for real time PCR kinetics Biochem Biophys Res Commun 2002 294 347 353 12051718 10.1016/S0006-291X(02)00478-3
Liu W Saint DA A new quantitative method of real time reverse transcription polymerase chain reaction assay based on simulation of polymerase chain reaction kinetics Anal Biochem 2002 302 52 59 11846375 10.1006/abio.2001.5530
Kvamme BO Skern R Frost P Nilsen F Molecular characterisation of five trypsin-like peptidase transcripts from the salmon louse (Lepeophtheirus salmonis) intestine Int J Parasitol 2004 34 823 832 15157765 10.1016/j.ijpara.2004.02.004
Frost P Nilsen F Validation of Reference Genes for Transcription Profiling in the Salmon Louse, Lepeophtheirus Salmonis, by Quantitative Real-Time Pcr Veterinary Parasitology 2003 118 169 174 14651887 10.1016/j.vetpar.2003.09.020
| 15854230 | PMC1090581 | CC BY | 2021-01-04 16:22:50 | no | BMC Mol Biol. 2005 Apr 26; 6:10 | utf-8 | BMC Mol Biol | 2,005 | 10.1186/1471-2199-6-10 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-201585750810.1186/1471-2180-5-20Research ArticleProliferative activity of extracellular HIV-1 Tat protein in human epithelial cells: expression profile of pathogenetically relevant genes Bettaccini Alessia A [email protected] Andreina [email protected] Roberto S [email protected] Fulvio [email protected] Antonio Q [email protected] Dipartimento di Scienze Cliniche e Biologiche, Università dell' Insubria, Varese, Italy2 Dipartimento di Oncologia, Università di Pisa, Pisa, Italy2005 27 4 2005 5 20 20 23 12 2004 27 4 2005 Copyright © 2005 Bettaccini 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
Tat is being tested as a component of HIV vaccines. Tat activity has been mainly investigated on cells of lymphoid/hematopoietic lineages. HIV-1, however, is known to infect many different cells of both solid organs and mucosal surfaces. The activity of two-exon (aa 1–101) and synthetic (aa 1–86) Tat was studied on mammary and amniotic epithelial cells cultured under low serum conditions.
Results
small concentrations of Tat (100 ng/ml) stimulated cell proliferation. Tat antibodies neutralized the mitogenic Tat activity. Changes of gene expression in Tat-treated cells were evaluated by RT-PCR and gene-array methods. Within 4 hours of treatment, exposure to Tat is followed by up-regulation of some cell cycle-associated genes (transcription factors, cyclin/cdk complexes, genes of apoptotic pathways) and of genes relevant to HIV pathogenesis [chemokine receptors (CXCR4, CCR3), chemotactic cytokines (SDF-1, RANTES, SCYC1, SCYE1), IL6 family cytokines, inflammatory cytokines, factors of the TGF-beta family (TGFb, BMP-1, BMP-2)]. Up-regulation of anti-inflammatory cytokines (IL-10, IL-19, IL-20), a hallmark of other persistent viral infections, was a remarkable feature of Tat-treated epithelial cell lines.
Conclusion
extracellular Tat is mitogenic for mammary and amniotic epithelial cells and stimulates the expression of genes of pathogenetic interest in HIV infection. These effects may favor virus replication and may facilitate the mother-to-child transmission of virus.
==== Body
Background
Numerous extracellular roles for the HIV-1 transactivator protein (Tat) have emerged from experiments showing that Tat is taken up by cultured cells, enters the nucleus, and transactivates genes linked to the HIV LTR [1].
Early experiments showing that recombinant Tat was able to inhibit antigen-induced, but not mitogen-induced, proliferation of peripheral blood mononuclear cells indicated Tat as a viral immunosuppressant [2]. This view has been confirmed by subsequent observations indicating that: a) Tat down-regulates HLA class I [3] and class II [4] expression in T-lymphocytes and macrophages, b) Tat represses transcription of the mannose receptor [5] a key molecule in the early response to invading pathogens, c) Tat suppresses CD26-dependent T cell proliferation [6], d) Tat down-regulates TCR/CD3 surface complexes [7], e) Tat induces suppressive levels of alpha interferon in T cells [8], and f) Tat causes human monocytes to secrete interleukin-10 (IL-10) an anti-inflammatory cytokine [9].
Extracellular Tat may stimulate virus replication and disease either through receptor-mediated signal transduction or after internalization and transport to the nucleus [10]. The basic domain of Tat protein (aa 45–56) is required for the binding to the viral transactivation-responsive element (TAR) and for importingextracellular Tat into the cell [11]. Low concentrations of extracellular Tat – similar to those measured in HIV-infected humans, i.e. 2–40 ng/ml [12] – have been shown to promote angiogenesis [13], to stimulate growth and motility of cultured Kaposi sarcoma cells [14], to induce monocyte chemotaxis through the synthesis of platelet-activating factor [15], to stimulate vascular permeability and the recruitment of mononuclear cells [16]. Thus, Tat has been shown to act through several pathways to establish an adequate cellular environment enhancing viral replication.
The multiple pathogenetic activities of extracellular Tat in HIV-infected individuals are unclear, but vaccination with native Tat or detoxified derivatives is being tested to prevent pathogenic events that would otherwise lead to a spreading infection [17,18]. The rationale for this approach is that Tat is conserved among different virus strains [19] and that vaccination might also protect uninfected bystander cells. Human antibodies to Tat correlate with nonprogression of HIV disease [20,21] and have been shown to reduce virus production in HIV-infected cell lines [22]. Remarkably, monkeys immunized with Tat and other HIV antigens were protected from challenge with simian/human immunodeficiency virus [18,23].
For using biologically active Tat as an immunogen, it is desirable to better define its activity on a wide variety of human cells. In this context, it should be considered that most studies have been carried out on cultured cells of lymphoid origin [5,24], whereas HIV-1 is capable of infecting numerous types non-lymphoid cells derived from solid tissues [reviewed in [25]]. As an example, HIV has been shown to affect the viability and functions of lung fibroblasts [26], vascular endothelial cells [27], kidney glomerular mesangial cells [28], as well as epithelial cells derived from the mammary gland [29], intestine [30], and kidney [31].
Our laboratory has shown that primary cultures of mammary epithelial cells (MEC) – that are possibly implicated to the mother-to-child transmission of infection – do express surface CD4, galactosylceramide and CD26 and can be infected by HIV-1 [29]. Infection was associated with the up-regulation of surface HLA class-II molecules and CD26, while the expression of CD4, tissue-specific markers, adhesion molecules and growth factor receptors was downregulated by the virus. Cytopathic effect ensued. Hormones (T3, alpha-estradiol and prolactin) enhanced HIV replication, possibly by stimulating cell proliferation.
Here we analyzed the effects of exogenous Tat protein on epithelial cells of mammary and amniotic origin. It was found that the HIV-1 transactivator induces the above cells to proliferate and that – a few hours after treatment – expression of surface receptors, cytokines and growth factors of pathogenetic relevance in HIV infection is up-regulated.
Results
Cell proliferation in response to exogenous Tat
Preliminary experiments with a variety of epithelial cell lines have shown that only marginal proliferative responses to Tat were observed when cells were grown in complete medium (i.e., medium containing 10% FBS). In contrast, experiments performed using LSM revealed a marked proliferative effect of Tat on epithelial cells. The activity of various doses of Tat was evaluated with two methods: 1) the XTT assay and 2) cell counts. As shown in Figure 1, the XTT assay showed significant proliferative responses when MCF-7 and MEC-1 cell lines were exposed for 2 days to doses of Tat ≥ 50 ng/ml. Equivalent results have been obtained with recombinant Tat101aa and synthetic Tat86aa protein preparations (as detailed in the Materials and Methods section). Thus, in subsequent experiments cell cultures were exposed to 100 ng/ml of Tat for different periods of time. In addition to the XTT method, cell proliferation has been measured by direct microscopic counts of cell monolayers. An example of the results is given in Figure 2. Different numbers of adherent cells can be observed in cultures of MCF-7 and MEC-1 cells either untreated or treated with Tat 100 ng/ml for 36 h. Comparable results have been obtained with several different epithelial cell lines. As shown in Figure 3, Tat consistently induced the proliferation of four different lines of mammary epithelial cells and of the amniotic AV-3 cell line. In these experiments it was also shown that a monoclonal anti-Tat antibody was capable of neutralizing Tat proliferative activity. Using the AV-3 cell line, the Tat-neutralizing activity of a rabbit polyclonal antibody was also demonstrated (bottom panel). Unrelated monoclonal and polyclonal antibodies to coxsackievirus B4 failed to neutralize Tat activity (data not shown). The hormone-dependent spontaneously immortalized MCF-10A line gave the lowest response to Tat, whereas consistently higher responses were observed with MCF-7 mammary cells and AV-3 amniotic cells. Thus, these two cell lines have been investigated in more detail to study the influence of exogenous Tat on gene expression profile.
RT-PCR analysis: Tat modulation of HIV-1 receptors, growth factors, and cytokine transcripts
Semiquantitative RT-PCR was used to evaluate transcript expression in cell cultures exposed to Tat. Primers for cytokines and VEGF were used in conjunction with a series of published primers (Table 1) to detect alterations of receptor, growth factor and cytokine transcript expression. Different time points were used (i.e., 1, 3, 6 and 12 h post-treatment). As shown by transcript analysis of selected HIV-1 receptors/co-receptors, gene up-regulation was observed early after Tat treatment (i.e., 3–6 h). Results relative to the MCF-7 and MEC-1 cell lines are reported in Table 2. Comparable results have been obtained with other mammary cell lines (data not shown). CD4 is expressed at low-levels in all investigated cell lines. Tat failed to influence CD4 expression. CXCR4, CCR3, and CCR4 transcripts were consistently up-regulated after 3–6 h treatment with Tat. In contrast, CCR1, CCR2, and CCR5 transcripts were not expressed. As shown in Figure 4, up-regulation of VEFG (mainly the 165 isoform) was observed in MCF-7 cells early after Tat exposure (1–3 h). This was concomitant with the up-regulation of VEGF receptor-2 (Flk-1/KDR). This result indicates the possibility of an autocrine loop contributing to Tat-stimulated cell proliferation. VEGF up-regulation was also observed in the other epithelial cell lines using the gene-array technique. Up-regulation of IL-6 and IL-6 receptor has also been detected in MCF-7 and AV-3 cells (data not shown). IL-6 has been shown to contribute to epithelial cell differentiation in both thyroid and the mammary gland [32]. Tat did not influence the expression of IL-8, a neutrophil-attracting cytokine produced by mammary epithelial cells [32]. Exposure to Tat for 4 h was followed by up-regulation of TGF-beta1 and TGF-beta2 in MCF-7 and AV-3 cells. Real time RT-PCR confirmed that the expression of IL-6, IL-6-R, TGF-beta1, and VEGF (isoform 165) was enhanced from 2.4- to 6.5-fold in cells treated with Tat for 4 h. As seen by RT-PCR, expression of CPSF3 [33] was not modified by 1 to 12 h treatment with Tat (data not shown).
Gene array studies. Genes up-regulated by Tat in both MCF-7 and AV-3 cell lines
Gene-arrays were used to compare gene expression in control untreated and Tat-treated cell cultures. Tables 3 and 4 summarize the results of cultures of MCF-7 and AV-3 cells exposed to Tat for 4 h. Genes that were up-regulated ≥1.5-fold in both cell lines are reported. Up-regulation is expressed as change (n-fold) for each cell line. The last column shows the average up-regulation in the two cell lines. Fifty of 96 genes (52%) related to the cell cycle were up-regulated from 1.7- to 4.6-fold. Twenty-four of 167 genes (14.4%) coding for growth factors, cytokines and their receptors were up-regulated from 1.6- to 3.7-fold.
The up-regulation of a wide spectrum of genes involved in cell cycle regulation that is presented here for two epithelial cell lines confirms the observed proliferative effect of Tat on non-lymphoid cell lines. As shown in Table 3, Tat treatment was associated with increased expression of: a) transcription factors (E2F, E2F-6, MPP2); b) cyclins (cyclin E1, E2, F, B, B2, C, D1, G, H); c) cyclin-dependent kinases and associated proteins (cdk6, cdk7, cdk8, protein regulating cytokinesis-1, cks1p9, cks2); d) cell division cycle proteins cdc27, cdc34, cdc37, cdc45); e) genes involved in DNA repair (MRE11A, nibrin, Hus1, RAD17, RAD50); f) cullins and associated proteins (cul-1, RBX1, Skp1, Skp2, Cul-2, Cul-3, Cul-4A, Cul-4B); g) nuclear factors associated with DNA replication (Ki67, PCNA, rpa, MCM2, MCM4, MCM5, MCM7); h) factors involved in cell cycle arrest (c-abl, chk1, cyclin G2, p18,, P21Cip1, p27Kip1); i) apoptosis-related genes (mdm2, bax).
Tat treatment was also associated with the increased transcription of selected growth factors, cytokines and receptors. Table 4 shows that the following genes were up-regulated in both MCF-7 and AV-3 cells: a) cytokines of the IL-6 superfamily and a signal-transducing molecule (IL-6, IL-11, GP130); b) inflammatory cytokines (LT-b, LTbR, TNFR1, IL-18, MIF); c) chemokines and their receptors (SDF-1, CXCR4, RANTES, CCR3, SCYC1, SCYE1); d) members of the PDGF/VEGF growth factor family (PDGFa, VEGF, VEGF-B, VEGF-C); e) factors belonging to the TGF-beta family [(TGFb1, bone morphogenetic proteins-1 and -2 (BMP1, BMP2)]; f) anti-inflammatory cytokines of the IL-10 family (IL-10, IL-19, IL-20). Exposure to Tat of the AV-3 and MCF-7 cell lines was not associated with downregulation of any of the tested genes.
Discussion
Our experiments confirm that low concentrations of extracellular Tat are capable of affecting cell proliferation and cellular physiology. We have shown that different epithelial cell lines cultured under low-serum conditions are induced to proliferate by doses of extracellular Tat of at least 50 ng/ml. The Tat dose used in these experiments (100 ng/ml) is close to that measured in the plasma of some HIV-1-infected patients [2–40 ng/ml; [12]]. The values reported [12] could be underestimated because local concentrations of Tat in infected tissues are probably higher as Tat may be sequestered by anti-Tat antibodies and by glycosaminoglycans. Tat produced in HIV-infected cells is released extracellularly and is able to stimulate the growth of Kaposi sarcoma cells at concentrations <1 ng/ml [34]. In the quoted paper, activation of cells by extracellular Tat required concentrations of 100 ng/ml or more. Extracellular Tat has also been shown to enhance the proliferation of germinal center B cells [35] and renal podocytes [36]. Tat is also involved in the progression of tumors arising in Tat-transgenic mice [37]. In lymphoid and neural cells Tat has also been shown to inhibit cell growth and cause apoptosis [38]. Pathways leading to cell death depend on the cell lineage and involve activation of death receptor pathways (i.e., TNF-alpha, Fas, TRAIL), chemokine receptor signaling, cytokine dysregulation, caspase activation, calcium mobilization, and loss of mitochondrial membrane potential.
In HIV replication cycle, Tat is known to interact with the transcription factor IIF and thereby to increase the activity of cdk7, thus allowing RNA polymerase II to transcribe the Tat activation region of the virus LTR [1]. Up-regulation of cdk7 by exogenous Tat has been demonstrated also in this study. Using Tat-transfected cell lines [33], Tat was shown to stimulate mRNA processing through up-regulation of the cleavage and polyadenylation specificity factor 3 (CPSF3). Up-regulation of CPSF3 was held responsible for stimulating viral and cellular gene expression. In contrast, in our experimental model and within 1 to 12 h exposure, Tat failed to induce CPSF3 up-regulation. Cell conditions different from those employed by us may be operative in Tat-expressing cell lines [33].
Increased expression of several transcription factors and cyclin/cdk complexes has been observed in Tat-treated epithelial cells. Genes involved in different phases of the cell cycle are activated, including genes (e.g., bax) linked to cell cycle arrest and negative modulation or apoptosis. These seemingly opposite effects of Tat on cell cycle regulators require some additional comments. Tat may either act directly on the transcription of negative regulators of cell cycle, or their up-regulation may be part of a negative feed back loop generated by an excessive increase of cell cycle inducers. Within this frame, it is of interest to note that extracellular Tat can inhibit p53 functions, thus providing a candidate mechanism through which HIV-1 might contribute to cell cycle dysregulation and malignant transformation [39]. Indeed, though the expression of p53 was not increased in epithelial cells exposed to Tat, the expression of mdm2 (a protein that binds to p53 and represses the p53-induced transcription activation) was strongly up-regulated. This is in contrast to recent findings showing repression of the mdm2 gene promoter in cells transfected with Tat constructs [33]. p21Cip1 – an important factor whose expression is regulated by p53 and that blocks cell replication in G1 by inhibiting CDK-cyclin complexes – was strongly up-regulated in Tat-treated cells. Thus, further studies are needed to assess the balance between activation and inhibition of cell cycle genes produced by Tat in cells of different lineages. These studies might help designing innovative antiviral therapies [40]. Measurements of epithelial cell growth in response to extracellular Tat could represent auseful model in this context.
As shown in Kaposi sarcoma cells [13,41], our studies confirm that extracellular Tat up-regulates the expression of VEGF receptor-2 (Flk-1/KDR). Tat also enhanced the expression of multiple angiogenic stimuli (PDGFa, VEGF-165, VEGF-B, and VEGF-C). This was seen both by RT-PCR analysis and gene arrays. In addition to stimulating angiogenesis, these factors are also capable of stimulating epithelial cell growth by autocrine mechanisms. In renal mesangial cells, persistent HIV replication was shown to induce expression of PDGFa/b and TGFb [28], cytokines that mediate mesangial cell proliferation and extracellular matrix deposition.
RT-PCR and gene array studies revealed the increased expression of some chemokine receptors serving as HIV coreceptors (CXCR4, CCR3, and to a minor extent CCR4). CXCR4 is the high affinity receptor for SDF-1, while CCR3 binds RANTES. The enhanced expression of CXCR4 and CCR3 in epithelial cells exposed to Tat suggests that their susceptibility to T- and M-tropic viruses may be increased since these cells express basal levels of CD4 and are permissive to HIV-1 [29]. Enhanced expression of HIV co-receptors has also been shown in monocytes [42], basophils [43], and erythroid cells [44]. Expression of SDF-1, RANTES and other chemotactic cytokines (SCYC1 and SCYE1) was also increased. The ability of Tat to signalthrough chemokines and their receptors is supposed to attract lymphoid cells toward virus producing cells, thus favoring the spread ofinfection. Up-regulation of the IL6 family factors (IL6, IL11, GP130) may promote cell growth [45]. Increased expression of the inflammatory cytokine LT-b (member of the TNF family) may co-operate with IL6 in the reactivation of HIV in CD4 T lymphocytes harboring a latent provirus [46]. Up-regulation of IL18 is also of interest since this cytokine stimulates virus replication. Increased IL18 levels have been reported in HIV-infected patients [47]. Up-regulation of MIF expression may contribute to recruit macrophages at sites of inflammation.
The significance of the increased expression of BMPs is unclear. BMPs have been identified as factors that induce ectopic bone and cartilage production. BMPs constitute the largest subfamily of the TGFb superfamily of growth factors and exert pleiotropic biological effects ranging from regulation of early developmental processes to organogenesis and differentiation of bone, cartilage, epidermal and neural tissues [48]. The increased expression of TGFb found in Tat-treated epithelial cells is a common feature of HIV-1-infected cells and is thought to be an important determinant of immunosuppression [49]. Expression of interferon-alpha (a putative immunosuppressive factor) could not be detected in epithelial cells [8].
Expression of the anti-inflammatory cytokines IL-10, IL-19, and IL-20 appears a remarkable feature of epithelial cell lines. In particular, high-level expression of IL-10 has been detected not only in MCF-7 and AV-3 cells, but also in other mammary and thyroid cell lines. IL-10 is a TH-2 cytokine supposed to limit HIV-1 replication in vivo through inhibiting the secretion of inflammatory cytokines (IL-1, TNF-alpha, IL-6, IL-8, IL-12) by lymphoid cells [10]. IL-10 receptor transcripts (IL-10Ra, IL-10Rb) have not been detected in Tat-treated epithelial cells. IL-19 and IL-20 are cytokines identified as IL-10 homologs that share the same receptor complex, indicating that the biological activities of the two cytokines overlap. Both appear to regulate development and proper functioning of the skin. Because the simultaneous presence of both receptor subunits (IL-20R1 and IL-20R2) is requiredfor IL-19 activity in a cell, only tissues that express bothsubunits represent potential targetsfor IL-19. The skin, testis, ovary, heart, lung, muscle, placenta, adrenal gland, small intestine, and salivary gland appear to express both receptors [50]. Expression of both IL-19 receptorsubunits in skin is of a particular interest since IL-19 is up-regulated in psoriatic skin. As for IL-10, the pattern of expression of IL-19 and IL-20 indicates that these cytokines may be involved in the regulationof inflammatory responses. Up-regulation of IL-10 has been associated with the mother-to-child transmission of human cytomegalovirus [51], with peak replication of feline immunodeficiency virus in vivo [52], and in HIV-infected patients [53]. Up-regulation of IL-19 mRNA has been observed in Epstein-Barr virus-transformed lymphocytes [54].
Conclusion
Extracellular Tat has a mitogenic effect on mammary and amniotic epithelial cells and stimulates the expression of genes of pathogenetic significance in HIV infection. Of particular interest is that Tat enhances the expression of anti-inflammatory cytokines of the IL-10 family. These multiple activities may also help explain the mother-to-child transmission of HIV.
Methods
Cell cultures
Four lines of mammary epithelial cells have been used: the spontaneously immortalized MCF-10A cells (non-tumorigenic), the carcinoma-derived MCF-7 (tumorigenic, hormone-dependent) cells, the MEC-1 (SV40-immortalized, non-tumorigenic) and MEC-2 (SV40-immortalized, tumorigenic, hormone-independent) cells. MEC-1 and MEC-2 cell lines have been obtained at our laboratory (32). In addition, the AV-3 line of amniotic origin was used. Cells were cultured in a 1:1 mixture of Dulbecco's modified Eagle's medium and F12 medium (DMEM/F12) supplemented with 10% heat-inactivated fetal bovine serum (FBS; HyClone Laboratories, Logan, UT), 4 mM L-glutamine, 1 mM Na-pyruvate, penicillin and gentamicin. Cell cultures were incubated at 37°C in a humidified atmosphere containing 5% CO2.
Unless otherwise specified, tissue culture reagents and chemicals were from Sigma Chemical Co. (St. Louis, MO). Plasticware was obtained from Falcon (Oxnard, CA). Molecular biology reagents were from Applied Biosystems (Monza, Italy).
Tat treatment of cells grown in low-serum-medium (LSM)
Purified recombinant HIV-1 two-exon Tat produced in E. coli (aa 1-101) was obtained from Intracel (London, UK). Synthetic Tat (aa 1–86) was chemically synthesized following established procedures (4). Tat was dissolved at 10 μg/ml in DMEM/F12 containing 1 mM dithio-L-threitol. Tat preparations were shown to be free of endotoxin contamination by the limulus amebocyte lysate assay (BioWhittaker Inc., Walkersville, MD). Aliquots were frozen at -70°C. The two preparations gave equivalent results and were equally neutralized both by a monoclonal antibody (1 μg/ml clone tMab-B, IgG1; Intracel) and by rabbit anti-Tat serum (Intracel).
For studying the effects of Tat on cell growth and gene expression, cells were plated in complete medium for 24 h and – before Tat treatment – incubated overnight in LSM to limit the activity of serum growth factors. LSM consisted of DMEM/F12 containing 0.5% SITE-3 supplement (Sigma), 0.3% dialyzed FBS (10,000 M.W. cut-off), 0.3% human serum albumin, human transferrin (5 μg/ml), Na-selenite (5 ng/ml). Cell proliferation and gene expression profiles were compared in control cultures devoid of Tat and in cultures treated with different concentrations of the Tat protein (10–500 ng/ml). For cell proliferation assays, cultures were exposed to Tat for 1 to 4 days. For analyzing gene expression profiles, cultures were exposed to Tat for 1 to 12 hr.
Cell proliferation assays
A) Cell counts
Five to 7 × 104 epithelial cells were plated for 24 hrs in wells of 24-well plates (1.9 square cm) with 0.5 ml of DMEM/F12 containing 10% FCS. Cultures were then incubated overnight in LSM. The medium was substituted with LSM containing either no Tat or increasing Tat concentrations (10–500 ng/ml). Cultures were examined 24, 48, 72, and 96 h after treatment with an inverted microscope equipped with a 37°C incubation chamber. Images were acquired with a 10× objective from 6–8 random fields per culture and recorded with a digital camera. Cell counts were obtained with a computerized image analysis system (Image DB; Amplimedical, Mira, Italy) and expressed as number of cells/well. Each bar represents the average of three different tests. In Figure 3, standard deviations of the mean are not reported, but were within 12% of mean values.
In some experiments, the growth-stimulating activity of Tat was neutralized with a monoclonal antibody to Tat (clone tMAb-B, 1 μg/ml) or with Tat rabbit antiserum (1:300 – 1:600 dilution). As a specificity control, a non-relevant neutralizing mAb to coxsackievirus B4 (clone 356.1, 3 μg/ml) and rabbit neutralizing serum to coxsackievirus B4 (1:300) were used. Control antibodies did not influence the mitogenic activity of Tat.
B) XTT reduction assay
The assay is based on the ability of viable cells to cleave the tetrazolium ring of the XTT sodium salt (Sigma) generating a water-soluble orange compound. Briefly, 1–3 × 104 epithelial cells per well were cultured for 24 hrs in flat-bottom 96-well microtiter plates using DMEM/F12 containing 10% FCS. Cultures were then incubated overnight in LSM to free cells from growth factors present in complete medium. The medium was then substituted with LSM without phenol red containing either no Tat or increasing Tat concentrations (10–500 ng/ml). When stimulated cultures were sub-confluent (1–3 days after treatment), XTT (200 μg/ml) was added to each well and cultures were further incubated for 3 hrs. Absorption values were measured at 450 nm with an ELISA reader. Tests were done in quadruplicate. As reported for cell count assays, the Tat neutralizing activity of tMAb-B and Tat rabbit antisera was also evaluated by the XTT assay.
Analysis of gene expression by RT-PCR
Semiquantitative RT-PCR was used to evaluate the expression of selected mRNA transcripts in cell cultures grown with LSM either in the absence or in the presence of Tat (100 ng/ml). Cultures in T-25 flasks were treated with Tat for different times (1, 3, 6, 12 h). Total RNA was extracted from 3–6 × 106 cells by the guanidinium thiocyanate method (Life Technologies, Gaithersburg, MD). Experiments were run in duplicate. Total RNA was treated for 1 h at 37°C with DNAse I 2 U/ml in a buffer containing RNAse inhibitor and MgCl2 2 mM. DNAse I was inactivated for 5 min at 90°C. cDNA was obtained from 2 μg of RNA with Mo-MLV reverse transcriptase in conjunction with random hexamer primers (Clontech, Palo Alto, CA). Cytokine-specific primer pairs obtained by Clontech (Cytokine MAPPing Amplimers) were used to amplify a variety of cytokines. Reagents from Maxim Biotech (San Francisco, CA) were used to detect transcripts of vascular endothelial growth factor (VEGF 121, 165, and 189 isoforms) and the VEGF receptor-2 (Flk/KDR). A series of published primers specific to HIV receptors, growth-factors, and a factor involved in pre-mRNA maturation were used (Table 1). Expression of the following transcripts was analyzed: IL-6, IL-6-R, IL-8, TGF-beta1, TGF-beta2, CD4, CXCR4, CCR1, CCR2, CCR3, CCR4, CCR5, CPSF3 (cleavage and polyadenylation specificity factor-3). The Applied Biosystems model 2400 thermal cycler was used for PCR reactions using AmpliTaq Gold polymerase (2 units) in a final volume of 50 μl. Samples were denatured at 94°C for 10 minutes before cycling for 21–25 cycles. Amplicons were analyzed on 2% agarose gel using ethidium bromide staining and were photographed on a transilluminator (Kodak Image Station440). Amplicons were quantified using the Kodak 1D 3.5 software using beta-actin and GAPDH transcripts to normalize the data. For selected genes (IL-6, IL-6-R, TGF-beta1, VEGF isoform 165), the relative gene expression obtained by microarray technology was confirmed using real-time PCR. A real-time PCR instrument (Smart Cycler; Cepheid, Sunnyvale, CA) was used together with SYBR Green PCR Master Mix and AmpliTaq Gold DNA polymerase obtained from Applied Biosystems. Reactions were carried out in a final volume of 25 μl. No-amplification control tubes containing samples, but no enzyme were included in each run to exclude the presence of fluorescent contaminants. Conditions consisted of an initial 10-min hold at 95°C followed by 40 amplification cycles. Real-time data were collected during the extension step of each cycle. Amplification of human beta-actin was used as a positive control and for normalizing data. Upon normalization with Ct values of beta-actin amplification, threshold cycle numbers (Ct) obtained for Tat-treated cultures were compared to those obtained for untreated control cultures.
Analysis of gene expression by gene arrays
GE Array™ Q series membranes (HS-001, HS-003, and HS-015 non-radioactive kits; Superarray, Bethesda, MD) were used to characterize gene expression profiles associated with Tat treatment. Membranes contained 96 cDNA tetraspots of human genes associated with specific biological pathways (cell cycle, common cytokines, inflammatory cytokines and receptors). pUC18 DNA was used as a negative control, and housekeeping genes (beta-actin, cyclophillin A, ribosomal protein L13a, GAPDH) as positive controls. To evaluate the pattern of induced or suppressed gene expression in response to Tat treatment, side-by-side hybridizations with samples from untreated and Tat-treated cultures were performed. Total RNA was extracted from confluent cell monolayers grown with LSM in T-25 flasks and treated or not with Tat. Preliminary experiments with Tat treatment given from 1 to 12 hr, indicated that 4 h was the optimal treatment time. Biotinylated probe synthesis was performed using mixtures of primers proper to each gene set. Before probe synthesis, RNA samples were analyzed and quantified by agarose gel electrophoresis. RNA samples (1–5 μg) combined with the primer mix were added with a prewarmed labeling mix containing 50 units of Mo-MLV reverse transcriptase, 20 units of RNase inhibitor and a dNTP mix containing 2 nmol of biotin-16-dUTP (Roche Applied Science, Monza, Italy) and incubated at 42°C for 90 min. GE Array membranes were prehybridized for 1 h with GEAhyb buffer containing 100 μg/ml heat-denatured salmon sperm DNA (Life Technologies) to prevent non-specific hybridization. Membranes were hybridized overnight at 60°C with denatured cDNA probes using the kit hybridization buffer and hybridization bottles rotating at 10 rpm. After extensive washing (60°C) at low and high stringency conditions in rotating hybridization bottles (10 rpm), membranes were incubated for 10 min with alkaline phosphatase-conjugated streptavidin (AP-streptavidin 1:7.500). Gene expression was detected by chemiluminescence using the AP substrate CDP-Star. Chemiluminescence signals were recorded on X-OMAT film (Kodak, Rochester, NY) using exposure times of 0.5 to 8 minutes. After development, X-OMAT films were scanned with a high-definition scanner (Coolscan 4000 ED, Nikon, Tokyo, Japan). Dedicated software (Gene-Analyzer, Superarray Inc.) was used for densitometric analysis. Experiments on MCF-7 and AV-3 cells were repeatedthree times.
Authors' contributions
AAB carried out tissue culture work, molecular methods, performed statistical analysis, drafted the manuscript giving critical contributions to the analysis of data and literature.
AB carried out tissue culture work, molecular methods, performed statistical analysis drafted the manuscript giving critical contributions to the analysis of data and literature.
RSA carried out immunological experiments, gave critical contributions to the analysis of data.
FB carried out tissue culture experiments, designed molecular reagents, gave critical contributions to the analysis of data.
AQT conceived the study, set up the methods, helped perform experiments, supervised data acquisition, and prepared the manuscript.
Acknowledgements
Work supported by the Banca del Monte di Lombardia (BML, Milan, Italy). RSA was supported by the Istituto Superiore di Sanità (Rome, Italy; National Research Project on AIDS No. 40D.01) and the European Vaccine Effort against AIDS (EUROVAC contract No. QLK2-CT-1999-01321). Dr. A. Bettaccini is a recipient of a fellowship from the Medical School of the University of Insubria, Varese, Italy. Dr. A. Baj is a Ph.D. student of the University of Pavia, Italy.
Figures and Tables
Figure 1 Growth of mammary epithelial cells exposed to different doses of Tat101aa. Cells were cultured in low-serum medium. Growth was measured by the XTT assay after 48 h of Tat treatment. Each bar represents the average of 3 different tests + SD. *, P < 0.05, as compared with untreated cell cultures.
Figure 2 Monolayers of MCF-7 and MEC-1 cells cultured in low-serum medium and treated or not with extracellular Tat101aa (100 ng/ml for 36 hours). Increased numbers of cells can be observed in Tat-treated cultures as compared to untreated control cultures. Phase contrast; microscopic fields taken with a 10× objective.
Figure 3 Neutralization by anti-Tat antibodies of the proliferative response to Tat of five different human epithelial cell lines cultured in low-serum medium. Cells were either left untreated (open bars) or exposed to 100 ng/ml Tat86aa (closed bars) at time 0. Cells were counted by microscopy at day-0, -1 and -2 after plating. Antibody tMAb-B (1 μg/ml) was mixed with Tat before treatment (red bars) or given alone (green bars). Bars represent the mean of three wells. Experiments on AV-3 cells (bottom panel) have also demonstrated the Tat-neutralizing activity of a rabbit polyclonal anti-Tat antibody: anti-Tat plus Tat (yellow bars) and anti-Tat alone (gray bars). Results are expressed as number of cells/well. Each bar represents the average of 3 different tests. Standard deviations of the mean are not reported, but were within 12% of the mean.
Figure 4 Semiquantitative RT-PCR of VEGF family genes in MCF-7 cells cultured in low-serum medium and exposed for different times to extracellular Tat86aa. Increased expression of the VEGF receptor-2 (Flk-1/KDR) and VEGF isoform 165. To a lower extent, the VEGF isoform-121 was also up-regulated. The GAPDH signal was used as a control. MW, DNA molecular weight ladder.
Table 1 Primers used to evaluate the expression of mRNA transcripts by RT-PCR.
Gene GenBank Accession No. Primer 5'-3' sequence
beta-actin
NM001101
Fwd ATCTGGCACCACACCTTCTACAATGAGCTGCG
Rvs CGTCATACTCCTGCTTGCTGATCCACATCTGC
CD4
BC025782
Fwd GTGAACCTGGTGGTGATGAGAGC
Rvs GGGGCTACATGTCTTCTGAAACCGGTG
CXCR-4
BC020968
Fwd CTGAGAAGCATGACGGACAAGTACAG
Rvs CAACAGCTTCCTTGGCCTCTGACT
CCR-1
NM001295
Fwd GGAAACTCCAAACACCACAGAGGA
Rvs AAGATCTCGCTGTACAAGCCTGTG
CCR-2
NM000647
Fwd CTCTCCCATTGTGGGCTCACTCTG
Rvs GCAAACACAGCATGGACAATAGCC
CCR-3
AF026535
Fwd CTATGATGACGTGGGCCTGCTC
Rvs AAGATCTCGCTGTACAAGCCTGTG
CCR-4
NM005508
Fwd CACCAAAGAAGGCATCAAGGCAT
Rvs TGCCACTGTAAAAGCCCACCAAG
CCR-5
NM000579
Fwd CCTGATAAACTGCAAAAGGCTGAAG
Rvs AGCAAACACAGCATGGACGACAG
IL-6
NM000600
Fwd ATGAACTCCTTCTCCACAAGCGC
Rvs GAAGAGCCCTCAGGCTGGACTG
IL-6R
X12830
Fwd CATTGCCATTGTTCTGAGGTTC
Rvs AGTAGTCTGTATTGCTGATGTC
IL-8
NM000584
Fwd ATGACTTCCAAGCTGGCCGTGGCT
Rvs TCTCAGCCCTCTTCAAAAACTTCTC
TGF-beta 1
NM000660
Fwd GCCCTGGACACCAACTATTGCT
Rvs GGACGGGGATGTAAACCTCGGA
TGF-beta 2
M19154
Fwd GATTTCCATCTACAAGACCACGAGGGACTTGC
Rvs GCTTACCGAGAGGAAGCTACATTGACTACGAC
CPFS3
AF171877
Fwd AATGGCTGGCAAACCCTTCTAATG
Rvs CATCGTCTTCACTTCCCTCTTCACA
Table 2 Expression of HIV receptors and co-receptors in mammary cells exposed to exogenous Tat protein (100 ng/ml)1.
Cell line CD4 CXCR4 CCR1 CCR2 CCR3 CCR4 CCR5
MCF-7 (untreated) + + - - -/+ -/+ -
MCF-7 + Tat
1 h + ++ - - + + -
3 h + ++ - - ++ ++ -
6 h + ++ - - ++ ++ -
MEC-1 (untreated) + + - - -/+ -/+ -
MEC-1 + Tat
1 h + ++ - - + + -
3 h + ++ - - ++ ++ -
6 h + ++ - - ++ + -
1. Expression of mRNA transcripts has been evaluated by semi-quantitative RT-PCR as reported in the Materials and Methods section. Transcript expression is reported as negative (-) or positive. Band intensity is indicated with 1+, 2+, or 3+.
Table 3 Expression profile of cell cycle-related genes in epithelial cell lines exposed for 4 h to exogenous Tat protein (100 ng/ml).
Gene name GenBank accession No. Description Fold change in expression1
MCF-7 cell line mean ± SD AV-3 cell line mean ± SD Mean of the two cell lines
P21Waf1 (p21Cip1)
L47233
Cyclin-dependent kinase inhibitor 1A (p21, Cip1) 4.2 ± .8 5.0 ± .8 4.6
Cdk7
NM 001799
Cyclin-dependent kinase 7 3.8 ± .9 5.2 ± .7 4.5
Cul1
NM 003592
Cullin 1 2.7 ± .6 5.3 ± 1.1 4.0
Cullin-Cul3
NM 003590
Cullin 3 2.6 ± .7 5.3 ± 1.3 3.9
CDC37
U63131
Cell division cycle 37 (S. cerevisiae) homolog 1.9 ± .3 5.9 ± .9 3.9
cyclinD1
M64349
Cyclin D1 (PRAD1 parathyroid adenomatosis 1) 1.8 ± .3 5.8 ± .7 3.8
Cks1p9
NM 001826
CDC28 protein kinase 1 2.4 ± .5 4.9 ± 1.2 3.7
Cullin-Cul4A
AF077188
Cullin 4A 2.2 ± .4 4.8 ± 1.1 3.5
Cks2
NM 001827
CDC28 protein kinase 2 1.6 ± .5 5.2 ± 1.1 3.4
CDC45-like1
NM 003504
Cell division cycle 45-like (S. cerevisiae) 5.1 ± 1.4 1.6 ± .4 3.4
CyclinH
U11791
Cyclin H 4.0 ± .6 2.6 ± 1.0 3.3
p27Kip1
U10906
Cyclin-dependent kinase inhibitor 1B (p27, Kip1) 1.8 ± .2 4.8 ± 1.0 3.3
Cdk8
X85753
Cyclin-dependent kinase 8 1.6 ± .3 4.8 ± .6 3.2
CDC34
L22005
Ubiquitin-conjugating enzyme, cell division cycle 34 3.0 ± .5 3.3 ± .6 3.1
Chk1
AF016582
CHK1 (checkpoint, S. pombe) homolog 1.6 ± .3 4.7 ± .9 3.1
Cyclin G
X77794
Cyclin G1 3.2 ± .5 3.0 ± .3 3.1
CyclinG2
L49506
Cyclin G2 2.7 ± .6 3.3 ± .5 3.0
Cdc27
NM 001256
Cell division cycle 27 2.7 ± .8 3.2 ± .7 3.0
p18
U17074
p18 (cdk4 inhibitor) cyclin-dependent kinase inhibitor 2C (p18, inhibits CDK4) 2.3 ± .5 3.6 ± 1.0 3.0
E2F
U47677
Transcription factor 1 3.8 ± .9 1.9 ± .6 2.9
RAD17
NM 002873
RAD17 (S. pombe) homolog 2.3 ± 1.0 3.3 ± .8 2.8
RAD50
U63139
RAD50 (S. cerevisiae) homolog 2.4 ± .4 3.1 ± .8 2.8
Cullin-Cul2
U83410
Cullin 2 1.5 ± .5 3.8 ± .8 2.7
E2F-6
AF059292
E2F transcription factor 6 2.9 ± .6 2.3 ± .8 2.6
cyclinB2
NM 004701
Cyclin B2 1.8 ± .8 3.2 ± .5 2.5
CyclinF
U17105
Cyclin F 3.5 ± .8 1.5 ± .6 2.5
PRC1
NM 003981
Protein regulator of cytokinesis1 2.0 ± .4 3.0 ± .5 2.5
Cullin-Cul4B
AB014595
Cullin 4B 2.0 ± .3 2.9 ± .3 2.5
cyclinE1
M73812
Cyclin E1 2.0 ± .4 2.9 ± .3 2.5
Hus1
NM 004507
HUS1 (S. pombe) checkpoint homolog 1.8 ± .8 3.1 ± .2 2.5
MPP2
U74613
Human putative M phase phosphoprotein 2 (MPP2) mRNA 1.8 ± .6 3.1 ± .6 2.5
Ki67 (MKI67)
X65550
Antigen identified by monoclonal Ab Ki-67 1.5 ± .3 3.3 ± .4 2.4
Skp1
U33760
Cyclin A / CDK2-associated p19 (Skp1) 1.5 ± .3 3.2 ± .6 2.4
Skp2
U33761
Human cyclin A / CDK2-associated p45 (Skp2) 2.0 ± .3 2.7 ± .6 2.4
PCNA
J04718
Proliferating cell nuclear antigen 1.9 ± .5 2.8 ± .5 2.4
Bax
L22474
Bcl2-associated X protein 1.6 ± .5 3.0 ± .7 2.3
cyclinE2
NM 004702
Cyclin E2 1.5 ± .4 3.1 ± .3 2.3
CyclinB
M25753
Cyclin B1 1.7 ± .3 3.1 ± .5 2.4
CyclinC
M74091
G1/S-specific Cyclin C 1.6 ± .4 3.0 ± .6 2.3
MCM5 (CDC46)
NM 006739
Minichromosome maintenance deficient (S. cerevisiae) 5 (cell division cycle 46) 1.8 ± .2 2.7 ± .4 2.3
MRE11A
U37359
Meiotic recombination (S. cerevisiae) 11 homolog A 1.5 ± .7 2.9 ± .5 2.2
Cdk6
X66365
Cyclin-dependent kinase 6 2.0 ± .3 2.3 ± .9 2.2
Nibrin
AF058696
Nijmegen breakage syndrome 1 (nibrin) 1.8 ± .4 2.5 ± .5 2.2
Rbx1
NM 014248
Homo sapiens ring-box protein 1 (RBX1) mRNA 1.5 ± .7 2.8 ± .6 2.2
MCM2
D83987
Minichromosome maintenance deficient (S. cerevisiae) 2 (mitotin) 1.5 ± .3 2.7 ± .3 2.1
MCM4 (CDC21)
X74794
Minichromosome maintenance deficient (S. cerevisiae) 4 2.0 ± .4 1.8 ± .2 1.9
MCM7 (cdc47)
D28480
Minichromosome maintenance deficient (S. cerevisiae) 7 1.5 ± .5 2.3 ± .7 1.9
c-abl
X16416
V-abl Abelson murine leukemia viral oncogene homolog 1 1.6 ± .5 1.9 ± .6 1.7
Mdm2
Z12020
Mouse double minute 2, human homolog of p53-binding protein 1.6 ± .6 1.8 ± .6 1.7
Rpa
L07493
Replication protein A3 (14 Kd) 1.6 ± .5 1.8 ± .4 1.7
1. Mean densitometric values are expressed for each gene as fold difference in cultures exposed to Tat / control cultures not exposed to Tat. Negative (pUC18) and positive controls (GAPDH, cyclophillin A, RPL13A, beta-actin) were used to normalize the data. Data are reported only for genes that were up-regulated by Tat at least 1.5-fold in both cell lines.
Table 4 Expression profile of cytokine-related genes in epithelial cell lines exposed for 4 h to exogenous Tat protein (100 ng/ml).
Gene name GenBank accession No. Description Fold change in expression1
MCF-7 cell line mean ± SD AV-3 cell line mean ± SD Mean of the two cell lines
MIF
NM 002415
Macrophage migration inhibitory factor (glycosylation-inhibiting factor) 5.2 ± .9 2.1 ± 1.8 3.7
IL-19
NM 013371
Interleukin 19 4.0 ± 1.1 2.4 ± .3 3.2
IL-13RA1
NM 001560
Interleukin 13 receptor. alpha 1 4.1 ± 1.2 2.1 ± .6 3.1
GP130
NM 002184
IL-6 signal transducer (gp 130, oncostatin M receptor) 3.9 ± 1.4 2.0 ± .3 3.0
CCR3
NM 001837
Chemokine (CC motif) receptor 3 4.2 ± 1.1 1.6 ± .3 2.9
IL-11
M57765
Interleukin 11 3.9 ± .7 1.7 ± .4 2.8
SCYE1
NM 004757
Small inducible cytokine subfamily E. member 1 (endothelial monocyte-activating) 2.6 ± .5 3.0 ± .6 2.8
BMP2
NM 001200
Bone morphogenetic protein 2 3.4 ± .5 2.0 ± .5 2.7
IL-18
NM 001562
Interleukin 18 (interferon-gamma-inducing factor) 2.2 ± .6 3.1 ± .9 2.7
TGFb1
X02812
Transforming growth factor beta 1 (Camurati-Engelmann disease) 2.8 ± .5 2.1 ± .4 2.5
SCYC1
U23772
Chemokine (C motif) ligand 1 2.5 ± .6 1.8 ± .4 2.2
BMP1
NM 006128
Bone morphogenetic protein 1 2.7 ± 1.0 1.5 ± .5 2.1
VEGF-C
X94216
Vascular endothelial growth factor C 2.5 ± 1.1 1.6 ± .2 2.0
IL-15
AF031167
Interleukin 15 2.0 ± .5 2.0 ± .3 2.0
TNFR1
M33294
Tumor necrosis factor receptor superfamily. member 1A 2.3 ± .7 1.7 ± .5 2.0
VEGF-B
U48801
Vascular endothelial growth factor B 2.2 ± .5 1.8 ± .1 2.0
CXCR4
NM 003467
Chemokine (CXC motif) receptor 4 2.1 ± .5 1.5 ± .5 1.8
IL-6
M14584
Interleukin 6 (interferon, beta 2) 2.0 ± .6 1.6 ± .2 1.8
IL-10
M57627
Interleukin 10 (high-level expression in both cell lines) 1.9 ± .5 1.7 ± .4 1.8
LT-b
NM 002341
Lymphotoxin beta (TNF superfamily, member 3) 1.8 ± .6 1.6 ± .8 1.7
PDGFa
X06374
Platelet-derived growth factor alpha polypeptide 1.7 ± .7 1.7 ± .5 1.7
LTbR
L04270
Lymphotoxin beta receptor (TNFR superfamily, member 3) 1.7 ± .2 1.6 ± .4 1.7
SCYA5 / RANTES
NM 002985
Chemokine (CC motif) ligand 5 1.6 ± .3 1.7 ± .5 1.6
IL-20
NM 018724
Interleukin 20 1.6 ± .7 1.6 ± .5 1.6
1. Mean densitometric values are expressed for each gene as fold difference in cultures exposed to Tat / control cultures not exposed to Tat. Negative (pUC18) and positive controls (GAPDH, cyclophillin A, RPL13A, beta-actin) were used to normalize the data. Data are reported only for genes that were up-regulated by Tat at least 1.5-fold in both cell lines.
==== Refs
Huigen MC Kamp W Nottet HS Multiple effects of HIV-1 trans-activator protein on the pathogenesis of HIV-1 infection Eur J Clin Invest 2004 34 57 66 14984439 10.1111/j.1365-2362.2004.01282.x
Viscidi RP Mayur K Lederman HM Frankel AD Inhibition of antigen-induced lymphocyte proliferation by Tat protein from HIV-1 Science 1989 246 1606 1608 2556795
Weissman JD Brown JA Howcroft TK Hwang J Chawla A Roche PA Schiltz L Nakatani Y Singer DS HIV-1 Tat binds TAFII250 and represses TAFII250-dependent transcription of major histocompatibility class I genes Proc Natl Acad Sci USA 1998 95 11601 11606 9751712 10.1073/pnas.95.20.11601
Tosi G De Lerma Barbaro A D'Agostino A Valle MT Megiovanni AM Manca F Caputo A Barbanti-Brodano G Accolla RS HIV-1 Tat mutants in the cysteine-rich region down-regulate HLA class II expression in T lymphocytic and macrophage cell lines Eur J Immunol 2000 30 19 28 10602023 10.1002/1521-4141(200001)30:1<19::AID-IMMU19>3.3.CO;2-0
Caldwell RL Egan BS Shepherd VL HIV-1 Tat represses transcription from the mannose receptor promoter J Immunol 2000 165 7035 7041 11120831
Wrenger S Hoffmann T Faust J Mrestani-Klaus C Brandt W Neubert K Kraft M Olek S Frank R Ansorge S Reinhold D The N-terminal structure of HIV-1 Tat is required for suppression of CD26-dependent T cell growth J Biol Chem 1997 272 30283 30288 9374514 10.1074/jbc.272.48.30283
Willard-Gallo KE Furtado M Burny A Wolinski SM Down-modulation of TCR/CD3 surface complexes after HIV-1 infection is associated with differential expression of the viral regulatory genes Eur J Immunol 2001 31 969 979 11298321 10.1002/1521-4141(200104)31:4<969::AID-IMMU969>3.0.CO;2-2
Zagury D Lachgar A Chams V Fall LS Bernard J Zagury JF Bizzini B Gringeri A Santagostino E Rappaport J Feldman M Burny A Gallo RC Interferon alpha and Tat involvement in the immunosuppression of uninfected T cells and C-C chemokine decline in AIDS Proc Natl Acad Sci USA 1998 95 3851 3856 9520456 10.1073/pnas.95.7.3851
Badou A Bennasser Y Moreau M Leclerc C Benkirane M Bahraoui E Tat protein of human immunodeficiency virus type 1 induces interleukin-10 in human peripheral blood monocytes: implication of protein kinase C-dependent pathway J Virol 2000 74 10551 10562 11044099 10.1128/JVI.74.22.10551-10562.2000
Shin HD Winkler C Stephens JC Bream J Young H Goedert JJ O'Brien TR Vlahov D Buchbinder S Giorgi J Rinaldo C Donfield S Willoughby A O'Brien SJ Smith MW Genetic restriction of HIV-1 pathogenesis to AIDS by promoter alleles of IL10 Proc Natl Acad Sci USA 2000 97 14467 14472 11121048 10.1073/pnas.97.26.14467
Fawell S Seery J Daikh Y Moore C Chen LL Pepinsky B Barsoum J Tat-mediated delivery of heterologous proteins into cells Proc Natl Acad Sci USA 1994 91 664 668 8290579
Xiao H Neuveut C Tiffany HL Benkirane M Rich EA Murphy PM Jeang KT Selective CXCR4 antagonism by Tat: implications for in vivo expansion of coreceptor use by HIV-1 Proc Natl Acad Sci USA 2000 97 11466 11471 11027346 10.1073/pnas.97.21.11466
Albini A Soldi R Gianciuglio D Giraudo E Benelli R Primo L Noonan D Salio M Camussi G Rockl W Bussolino F The angiogenesis induced by HIV-1 tat protein is mediated by the Flk-1/KDR receptor on vascular endothelial cells Nat Med 1996 2 1371 1375 8946838 10.1038/nm1296-1371
Biancone L Cantaluppi V Boccellino M Bussolati B Del Sorbo L Conaldi PG Albini A Toniolo A Camussi G Motility induced by human immunodeficiency virus-1 Tat on Kaposi's sarcoma cells requires platelet-activating factor synthesis Am J Pathol 1999 155 1731 1739 10550329
Del Sorbo L De Martino A Biancone L Bussolati B Conaldi PG Toniolo A Camussi G The synthesis of platelet-activating factor modulates chemotaxis of monocytes induced by HIV-1 Tat Eur J Immunol 1999 29 1513 1521 10359105 10.1002/(SICI)1521-4141(199905)29:05<1513::AID-IMMU1513>3.3.CO;2-O
Arese M Ferrandi C Primo L Camussi G Bussolino F HIV-1 Tat protein stimulates in vivo vascular permeability and lymphomononuclear cell recruitment J Immunol 2001 166 1380 1388 11145723
Ensoli B Cafaro A HIV-1 Tat vaccines Virus Res 2002 82 91 101 11885958 10.1016/S0168-1702(01)00393-8
Mooij P Nieuwenhuis IG Knoop CJ Doms RW Bogers WM Ten Haaft PJ Niphuis H Koornstra W Bieler K Kostler J Morein B Cafaro A Ensoli B Wagner R Heeney JL Qualitative T-helper responses to multiple viral antigens correlate with vaccine-induced immunity to simian/human immunodeficiency virus infection J Virol 2004 78 3333 3342 15016855 10.1128/JVI.78.7.3333-3342.2004
Butto S Fiorelli V Tripiciano A Ruiz-Alvarez MJ Scoglio A Ensoli F Ciccozzi M Collacchi B Sabbatucci M Cafaro A Guzman CA Borsetti A Caputo A Vardas E Colvin M Lukwiya M Rezza G Ensoli B Tat Multicentric Study Group. Sequence conservation and antibody cross-recognition of clade B human immunodeficiency virus (HIV) type 1 Tat protein in HIV-1-infected Italians, Ugandans, and South Africans J Infect Dis 2003 188 1171 1180 14551888 10.1086/378412
Zagury JF Sill A Blattner W Lachgar A Le Buanec H Richardson M Rappaport J Hendel H Bizzini B Gringeri A Carcagno M Criscuolo M Burny A Gallo RC Zagury D Antibodies to the HIV-1 Tat protein correlated with nonprogression to AIDS: a rationale for the use of Tat toxoid as an HIV-1 vaccine J Hum Virol 1998 1 282 292 10195253
Re MC Vignoli M Furlini G Gibellini D Colangeli V Vitone F La Placa M Antibodies against full-length Tat protein and some low-molecular-weight Tat-peptides correlate with low or undetectable viral load in HIV-1 seropositive patients J Clin Virol 2001 21 81 89 11255101 10.1016/S1386-6532(00)00189-X
Moreau E Hoebeke J Zagury D Muller S Desgranges C Generation and characterization of neutralizing human monoclonal antibodies against human immunodeficiency virus type 1 Tat antigen J Virol 2004 78 3792 3796 15016898 10.1128/JVI.78.7.3792-3796.2004
Cafaro A Caputo A Fracasso C Maggiorella MT Goletti D Baroncelli S Pace M Sernicola L Koanga-Mogtomo ML Betti M Borsetti A Belli R Akerblom L Corrias F Butto S Heeney J Verani P Titti F Ensoli B Control of SHIV-89.6P-infection of cynomolgus monkeys by HIV-1 Tat protein vaccine Nat Med 1999 5 643 650 10371502 10.1038/9488
Vene R Benelli R Noonan DM Albini A HIV-Tat dependent chemotaxis and invasion, key aspects of tat mediated pathogenesis Clin Exp Metastasis 2000 18 533 538 11688957 10.1023/A:1011991906685
Levy JA Levy JA Effect of HIV on various tissues and organ systems in the host HIV and the pathogenesis of AIDS 1998 Washington DC: ASM Press 189 228
Dolei A Serra C Arca MV Tilocca F Riva E Antonelli G Dianzani F Toniolo A Infectious virus with reduced cytopathogenicity resulting from persistent infection of normal lung fibroblasts by HIV type 1 strains AIDS Res Hum Retroviruses 1994 10 1089 1095 7826696
Conaldi PG Serra C Dolei A Basolo F Falcone V Mariani G Speziale P Toniolo A Productive HIV-1 infection of human vascular endothelial cells requires cell proliferation and is stimulated by combined treatment with interleukin-1 beta plus tumor necrosis factor-alpha J Med Virol 1995 47 355 363 8636703
Conaldi PG Bottelli A Wade-Evans A Biancone L Baj A Cantaluppi V Serra C Dolei A Toniolo A Camussi G HIV-persistent infection and cytokine induction in mesangial cells: a potential mechanism for HIV-associated glomerulosclerosis AIDS 2000 14 2045 2047 10997410 10.1097/00002030-200009080-00021
Toniolo A Serra Conaldi PG Basolo F Falcone V Dolei A Productive HIV-1 infection of normal human mammary epithelial cells AIDS 1995 9 859 866 7576319
Fantini J Yahi N Chermann JC Human immunodeficiency virus can infect the apical and basolateral surfaces of human colonic epithelial cells Proc Natl Acad Sci U S A 1991 88 9297 9301 1718004
Conaldi PG Biancone L Bottelli A Wade-Evans A Racusen LC Boccellino M Orlandi V Serra C Camussi G Toniolo A HIV-1 kills renal tubular epithelial cells in vitro by triggering an apoptotic pathway involving caspase activation and Fas up-regulation J Clin Invest 1998 102 2041 2049 9854039
Basolo F Fiore L Calvo S Falcone V Conaldi PG Fontanini G Caligo AM Merlo G Gluzman Y Toniolo A Defective interleukin six expression and responsiveness in human mammary cells transformed by an adeno 5/SV40 hybrid virus Br J Cancer 1996 73 1356 1361 8645579
Calzado MA Sancho R Munoz E Human immunodeficiency virus type 1 Tat increases the expression of cleavage and polyadenylation specificity factor 73-kilodalton subunit modulating cellular and viral expression J Virol 2004 78 6846 54 15194760 10.1128/JVI.78.13.6846-6854.2004
Ensoli B Buonaguro L Barillari G Fiorelli V Gendelman R Morgan RA Wingfield P Gallo RC Release, uptake, and effects of extracellular human immunodeficiency virus type 1 Tat protein on cell growth and viral transactivation J Virol 1993 67 277 287 8416373
Lefevre EA Krzysiek R Loret EP Galanaud P Richard Y Cutting edge: HIV-1 Tat protein differentially modulates the B cell response of naive, memory, and germinal center B cells J Immunol 1999 163 1119 1122 10415004
Conaldi PG Bottelli A Baj A Serra C Fiore L Federico G Bussolati B Camussi G Human immunodeficiency virus-1 tat induces hyperproliferation and dysregulation of renal glomerular epithelial cells Am J Pathol 2002 161 53 61 12107089
Altavilla G Caputo A Trabanelli C Brocca Cofano E Sabbioni S Menegatti MA Barbanti-Brodano G Corallini A Prevalence of liver tumours in HIV-1 tat-transgenic mice treated with urethane Eur J Cancer 2004 40 275 283 14728943 10.1016/j.ejca.2003.08.025
Regulier EG Reiss K Khalili K Amini S Rappaport J Zagury JF Katsikis PD T-cell and neuronal apoptosis in HIV infection: implications for therapeutic intervention Int Rev Immunol 2004 23 25 59 14690854
Harrod R Nacsa J Van Lint C Hansen J Karpova T McNally J Franchini G Human immunodeficiency virus type-1 Tat/co-activator acetyltransferase interactions inhibit p53Lys-320 acetylation and p53-responsive transcription J Biol Chem 2003 278 12310 12318 12501250 10.1074/jbc.M211167200
Sadaie MR Mayner R Doniger J A novel approach to develop anti-HIV drugs: adapting non-nucleoside anticancer chemotherapeutics Antiviral Res 2004 61 1 18 14670589 10.1016/j.antiviral.2003.09.004
Ganju RK Munshi N Nair BC Liu ZY Gill P Groopman JE Human immunodeficiency virus tat modulates the Flk-1/KDR receptor, mitogen-activated protein kinases, and components of focal adhesion in Kaposi's sarcoma cells J Virol 1998 72 6131 6137 9621077
Huang L Bosch I Hofmann W Sodroski J Pardee AB Tat protein induces human immunodeficiency virus type 1 (HIV-1) coreceptors and promotes infection with both macrophage-tropic and T-lymphotropic HIV-1 strains J Virol 1998 72 8952 8960 9765440
De Paulis A De Palma R Di Gioia L Carfora M Prevete N Tosi G Accolla RS Marone G Tat protein is an HIV-1-encoded beta-chemokine homolog that promotes migration and up-regulates CCR3 expression on human Fc epsilon RI+ cells J Immunol 2000 165 7171 7179 11120849
Gibellini D Re MC Vitone F Rizzo N Maldini C La Placa M Zauli G Selective up-regulation of functional CXCR4 expression in erythroid cells by HIV-1 Tat protein Clin Exp Immunol 2003 31 428 435 10.1046/j.1365-2249.2003.02095.x
Basolo F Calvo S Fiore L Conaldi PG Falcone V Toniolo A Growth-stimulating activity of interleukin 6 on human mammary epithelial cells transfected with the int-2 gene Cancer Res 1993 53 2957 2960 8100482
Ghose R Liou LY Herrmann CH Rice AP Induction of TAK (cyclin T1/P-TEFb) in purified resting CD4(+) T lymphocytes by combination of cytokines J Virol 2001 75 11336 11343 11689614 10.1128/JVI.75.23.11336-11343.2001
Ahmad R Sindhu ST Toma E Morisset R Ahmad A Elevated levels of circulating interleukin-18 in human immunodeficiency virus-infected individuals: role of peripheral blood mononuclear cells and implications for AIDS pathogenesis J Virol 2002 76 12448 12456 12438570 10.1128/JVI.76.24.12448-12456.2002
Massague J Integration of Smad and MAPK pathways: a link and a linker revisited Genes Dev 2003 17 2993 2997 14701870 10.1101/gad.1167003
Reinhold D Wrenger S Kahne T Ansorge S HIV-1 Tat: immunosuppression via TGF-beta1 induction Immunol Today 1999 20 384 385 10431160 10.1016/S0167-5699(99)01497-8
Blumberg H Conklin D Xu WF Grossmann A Brender T Carollo S Eagan M Foster D Haldeman BA Hammond A Haugen H Jelinek L Kelly JD Madden K Maurer MF Parrish-Novak J Prunkard D Sexson S Sprecher C Waggie K West J Whitmore TE Yao L Kuechle MK Dale BA Chandrasekher YA Interleukin 20: discovery, receptor identification, and role in epidermal function Cell 2001 104 9 19 11163236 10.1016/S0092-8674(01)00187-8
Yamamoto-Tabata T McDonagh S Chang HT Fisher S Pereira L Human cytomegalovirus interleukin-10 downregulates metalloproteinase activity and impairs endothelial cell migration and placental cytotrophoblast invasiveness in vitro J Virol 2004 78 2831 40 14990702 10.1128/JVI.78.6.2831-2840.2004
Avery PR Hoover EA Gamma interferon/interleukin 10 balance in tissue lymphocytes correlates with down modulation of mucosal feline immunodeficiency virus infection J Virol 2004 78 4011 9 15047817 10.1128/JVI.78.8.4011-4019.2004
Granelli-Piperno A Golebiowska A Trumpfheller C Siegal FP Steinman RM HIV-1-infected monocyte-derived dendritic cells do not undergo maturation but can elicit IL-10 production and T cell regulation Proc Natl Acad Sci USA 2004 101 7669 74 15128934 10.1073/pnas.0402431101
Wolk K Kunz S Asadullah K Sabat R Cutting edge: immune cells as sources and targets of the IL-10 family members? J Immunol 2002 168 5397 5402 12023331
| 15857508 | PMC1090582 | CC BY | 2021-01-04 16:03:39 | no | BMC Microbiol. 2005 Apr 27; 5:20 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-20 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-221586971610.1186/1471-2180-5-22Research ArticleInterferon-γ promotes abortion due to Brucella infection in pregnant mice Kim Suk [email protected] Dong Soo [email protected] Kenta [email protected] Hidefumi [email protected] Hiroshi [email protected] Masahisa [email protected] Department of Applied Veterinary Science, Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido 080-8555, Japan2 Research Center for Protozoan Diseases, Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido 080-8555, Japan3 Department of Pathological Science, Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido 080-8555, Japan4 Department of Development and Medical Technology, Graduate School of Medicine, The University of Tokyo, Hongo, Tokyo 113-0033, Japan2005 4 5 2005 5 22 22 3 12 2004 4 5 2005 Copyright © 2005 Kim 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 mechanisms of abortion induced by bacterial infection are largely unknown. In the present study, we investigated abortion induced by Brucella abortus, a causative agent of brucellosis and facultative intracellular pathogen, in a mouse model.
Results
High rates of abortion were observed for bacterial infection on day 4.5 of gestation, but not for other days. Regardless of whether fetuses were aborted or stayed alive, the transmission of bacteria into the fetus and bacterial replication in the placenta were observed. There was a higher degree of bacterial colonization in the placenta than in other organs and many bacteria were detected in trophoblast giant cells in the placenta. Intracellular growth-defective virB4 mutant and attenuated vaccine strain S19 did not induce abortion. In the case of abortion, around day 7.5 of gestation (period of placental development), transient induction of IFN-γ production was observed for infection by the wild type strain, but not by the virB4 mutant and S19. Neutralization of IFN-γ, whose production was induced by infection with B. abortus, served to prevent abortion.
Conclusion
These results indicate that abortion induced by B. abortus infection is a result of transient IFN-γ production during the period of placental development.
==== Body
Background
Brucellosis is a widespread and economically important infectious disease of animals and humans caused by members of the genus Brucella. Brucella spp. are small gram-negative, facultative intracellular pathogens that cause abortion, retained placenta and infertility in numerous domestic and wild mammals, and a disease known as undulant fever in humans [1,2]. Infection in humans is almost exclusively due to zoonosis, either through direct contact with infected animals or from contaminated dairy products [3]. The mouse model, particularly that using the unpregnant mouse, has been used extensively to study some aspects of the pathogenesis of brucellosis [4]. While brucellosis is known to primarily affect the reproductive tract in the natural host and has been much studied, little is known regarding the cellular and molecular mechanisms of Brucella infection in the pregnant mouse [5]. There have been no studies on the induction of abortion by Brucella infection in the pregnant mouse.
A key aspect of the virulence of Brucella is its ability to proliferate within professional and non-professional phagocytic host cells, thereby successfully bypassing the bactericidal actions of phagocytes, which is thought to explain its virulence and ability to cause chronic infections [6,7]. The molecular mechanisms and genetic basis for its intracellular survival and replication, however, are not understood completely. Recent studies with non-professional phagocyte HeLa cells show that Brucella inhibits phagosome-lysosome fusion and transits through an intracellular compartment that resembles autophagosomes. Bacteria replicate in a different compartment, containing protein markers normally associated with the endoplasmic reticulum, as shown by confocal microscopy and immunogold electron microscopy [8-10].
Pregnancy leads to a generalized suppression of the adaptive immune system, typified by significantly decreased cell-mediated immunity and reduced T helper cell (Th) 1 responsiveness [11-13]. This immunosuppressed state prevents maternal rejection of the fetus but has the unfortunate consequence of increasing maternal susceptibility to certain infectious agents [14,15]. Immunity against B. abortus is principally mediated by cellular immune responses since it is an intracellular pathogen, and involves antigen-specific T-cell activation of CD4 and CD8 T cells and humoral responses. Protection of the host against B. abortus infection is thought to be mediated primarily by a Th1 type of immune response [16]. For many other intracellular bacterial and protozoan pathogens, it has been shown that interferon-γ (IFN-γ) is an important component of Th1 immune responses and contributes to control through its ability to stimulate macrophages to kill more microbes. The role of IFN-γ in the control of B. abortus infections has been demonstrated by supplementing BALB/c mice with recombinant IFN-γ, when such treatment resulted in a 10-fold decrease in the number of bacteria at 1 week after infection [17]. It has also been shown that neutralizing endogenous IFN-γ by administering anti-IFN-γ monoclonal antibodies results in a decrease in control [16]. Despite these results, however, the role of cytokines in the control of B. abortus infection in the pregnant mouse is still unknown.
In the present study, we investigated the pathogenesis of B. abortus in the pregnant mouse and established a mouse model for abortion induced by B. abortus infection. We noted that IFN-γ plays an important role in the induction of abortion by B. abortus infection.
Results
B. abortus infection causes abortion in pregnant mice
In order to construct a mouse model of abortion due to B. abortus infection, the numbers of aborted fetuses in infected mice were counted on day 18.5 of gestation. For infection with wild type B. abortus on day 4.5 of gestation, all fetuses, except for those of 1 mouse (No. 2), were aborted (98.4% abortion) (Table 1) (Fig. 1). For infection on days 3.5, 6.5 and 9.5 of gestation, the majority of fetuses were still alive (28.1–37.3% abortion), and no abortion was observed for infection on day 14.5 of gestation (Table 1) (Fig. 1). The endpoint number of delivered mice in pregnant mice infected on day 4.5 of gestation was 2, but that in pregnant mice infected on the other days of gestation or in uninfected pregnant mice was 16. Replicating bacteria were observed in both aborted fetuses and those that lived, and there were no significant differences in bacterial infection rates between live (average 3.43 log CFU) and aborted fetuses (average 4.10 log CFU) (Table 1). These experiments were performed at least three times with the same results.
B. abortus infection predominantly in placenta
It is well known that the placenta plays an integral role in the pathogenesis of congenital infections. However, the placental factors controlling these processes are not completely understood. To investigate the role of the placenta in B. abortus infection, we examined the placental colonization of the wild type strain. Replicating bacteria were observed in both the placentas of live and aborted fetuses, and the difference between the bacterial growth rate in the placenta of live (average 7.12 log CFU) and aborted fetuses (average 7.17 log CFU) was not significant (Table 1). The bacterial growth rate in the spleen was lower than that in the placenta in all cases (Table 1). In order to check whether B. abortus colonizes the placenta predominantly, bacterial growth in the liver, lung and kidney were measured in addition to that in the spleen and placenta for infection on day 6.5 of gestation. As expected, colonization by B. abortus was much greater in the placenta than in the other organs (Fig. 2). Immunostaining of bacteria infected placenta specimens revealed that bacteria were present in numerous trophoblast giant cells (TGC) along the periphery of the placenta. Bacteria were also observed in neutrophils associated with the regressing deciduas capsularis or free within the newly formed uterine lumen adjacent to this region of the placenta (Fig. 3). Replicating bacteria were observed in TGC in both the placentas of live and aborted fetuses (data not shown).
Type IV secretion system of B. abortus contributes to abortion indirectly
As the internalization and intracellular replication of B. abortus are controlled by the type IV secretion system [18,19], on day 4.5 of gestation, pregnant mice were infected with the virB4 mutant, which does not have the ability of intracellular replication [20,21]. The inoculation of pregnant mice with the virB4 mutant did not induce abortion. Bacterial growth, however, was observed in the fetus, placenta and spleen (Table 2). We also observed that attenuated live vaccine strain S19 did not induce abortion (Table 2), even though in contrast to the virB4 mutant it has been seen to have replication capability in macrophages [22,23]. These results suggest that the type IV secretion system may not contribute to abortion directly when mice were infected with B. abortus, because S19 has intact virB genes.
Transient increase in IFN-γ in pregnant mice induced by B. abortus infection
Though fetal tissues possess paternal antigens, the fetus is thought to be insulated from rejection by the maternal immune system. The precise cellular and molecular mechanisms underlying the maintenance of pregnancy and the induction of abortion, however, remain unclear. As recent studies have shown that Th1 cytokines, such as interferon-γ (IFN-γ) and tumor necrosis factor α (TNF-α), induce abortion [24], we decided to determine if IFN-γ and TNF-α contribute to the induction of abortion by B. abortus infection and measured the production of IFN-γ and TNF-α by ELISA in pregnant mice infected on day 4.5 of gestation. A large amount of induced IFN-γ production was observed at 3 days after infection with the wild type strain (day 7.5 of gestation) but this production of IFN-γ then decreased rapidly (Fig. 4A). A similar pattern of IFN-γ induction was observed after infecting unpregnant mice with the wild type strain. However, neither the virB4 mutant nor the attenuated live vaccine strain S19 induced IFN-γ production (Fig. 4A). In addition, no significant induction of TNF-α production was observed in the infected pregnant mice (Fig. 4B). These results imply that IFN-γ contributes to abortion due to B. abortus infection, and increased production of IFN-γ around day 7.5 of gestation may be critical, but that TNF-α is not significantly involved.
Preventing abortion by neutralizing IFN-γ production
To determine if abortion is prevented by neutralizing the IFN-γ produced as a result of bacterial infection, pregnant mice were inoculated with monoclonal anti-IFN-γ antibody 1 day before infection with B. abortus, which was done on day 4.5 of gestation. Thereafter, the serum levels of IFN-γ and abortion rates were recorded. A large amount of induced IFN-γ production was observed 3 days after infection (day 7.5 of gestation) in the pregnant mice and unpregnant control for mice without antibody, and this production was inhibited in mice inoculated monoclonal anti-IFN-γ antibody for both pregnant mice and the unpregnant control (Fig. 5A). Also, while abortions were observed in the untreated pregnant mice as before, none were seen in pregnant mice inoculated with anti-IFN-γ antibody (Fig. 5B). These results indicate that IFN-γ production contributes to abortion due to B. abortus infection in pregnant mice.
Discussion
In the present study we investigated abortion induced by B. abortus infection in a mouse model. In a previous study, Tobias et al. observed no abortion in pregnant BALB/c mice infected with B. abortus virulent strain 2308 on day 9 of gestation, but the bacteria multiplied in the placenta and could be identified in trophoblast giant cells (TGC) [5]. In our mouse model, high abortion rates were seen with bacterial infection on day 4.5 of gestation and lower rates were observed when infection was on day 9.5 of gestation. It thus seems that the day of infection during gestation is critical for induction of abortion in mouse models, and this may explain the difference between our results and those of the previous study. Through inoculation with wild type B. abortus on day 4.5 of gestation, an increase in the production of IFN-γ was induced around day 7.5 of gestation. It is thought that the serum IFN-γ level in this period of gestation is relevant to the maintenance of the fetus in the pregnant mice and when the amount of IFN-γ increases during this critical period of pregnancy, abortion may occur [13]. And we observed that abortion due to bacterial infection was indeed prevented when the amount of IFN-γ decreased during the critical period due to neutralization with anti-IFN-γ monoclonal antibody. Cytokine profiles have been observed to have physiological implications for pregnancy, perhaps the most profound being the detrimental affect of T helper cell (Th) 1 cytokines such as IFN-γ and TNF-α, and associated cells and mediators. Thus, natural killer (NK) cells, IFN-γ, TNF-α and nitric oxide (NO) can be deleterious to the well-being of the fetoplacental unit, and capable of inducing abortion and fetal resorption [13]. Our results indicated that the Th1 cytokine IFN-γ play a key role in pregnancy.
Our study and that of Tobias et al. showed that B. abortus replicated preferentially in the TGC in the placenta [5]. The mechanism for this is unclear. Erythritol, a sugar alcohol synthesized in the ungulate placenta and known to stimulate the growth of virulent strains of B. abortus, has been credited with the preferential localization of this bacterium within the placenta of ruminants [25,26]. Although very low concentrations of erythritol are found in the placentas of rodents [25], B. abortus will localize and proliferate preferentially in their placenta [[5,27], this study]. While the growth of attenuated vaccine strain S19 was inhibited by erythritol, but its growth was still preferentially localized in the ruminant placenta [26]. Two placental regions can be recognized: a junctional zone and a labyrinth zone. The junctional zone is comprised of stem cells and three differentiated cell types: TGC, spongiotrophoblast cells, and glycogen cells. TGC arise by endoreduplication, are situated at the maternal-placental interface, and are one of the major endocrine cells of the placenta [28]. TGC are polyploidy cells that play a crucial role during implantation and in remodeling of the embryonic cavity, avoidance of maternal immune rejection and promotion of maternal blood flow to the implantation site [29]. TGC should have been hijacked by B. abortus so that cell functions were not exhibited completely, this leading to abortion due to inhibition of implantation and placental development. However, we observed replicating bacteria in the TGC in both placentas of live and aborted fetuses. TGC, spongiotrophoblasts and the labyrinth zone are established from day 8.5 to 10.5 of gestation and comprise the definitive placenta that persists until the end of gestation [29]. Therefore, abortion-inducing factors should act just before day 8.5 of gestation, adding to bacterial infection in the TGC. As abortion occurs in pregnant mice when IFN-γ production is induced by the bacterial infection around day 7.5 of gestation, IFN-γ may influence the development of TGC. In this regard, IFN-γ has been shown to inhibit the cells of the trophoblast lineage outgrowth and trophoblast cell invasion to be accelerated in mice with genetic deficiency in the IFN-γ or IFN-γ receptor [30]. Presumably, IFN-γ production induced by bacterial infection inhibits the development of TGC and B. abortus invades the remaining small number of TGC around day 7.5 of gestation. TGC completely lose all of their function due to the effect of IFN-γ and bacterial replication inside them.
Natural killer (NK) cells are large granular lymphocytes found in peripheral blood and also in the maternal decidua during pregnancy. The actions of NK cells on trophoblast lineage cells are likely mediated by NK cell secretory products, including cytokines, and may be direct or indirect. Uterine NK cells produce several cytokines and are the primary source of INF-γ in the metrial grand [30,31]. IFN-γ has been implicated as a major mediator of uterine NK cell function during pregnancy [32,33]. Trophoblast cells are among a variety of different IFN-γ targets, and in vitro trophoblast cell differentiation, survival, and outgrowth are affected by IFN-γ [30,34]. IFN-γ is a crucial cytokine for Brucella immunity [35,36]. The major function of IFN-γ in Brucella immunity is the stimulation of bactericidal activity in macrophages, host cells of Brucella. However, the function of IFN-γ is more diverse than the induction of bactericidal function and includes the stimulation of antigen presentation through class I and class II MHC molecules, the orchestration of leukocyte-endothelium interactions, the effects on cell proliferation and apoptosis, as well as stimulation and repression of a variety of genes whose functional significance remains obscure [37]. Though a previous study found that little role for NK cells in controlling brucellosis in unpregnant mice when NK cells were depleted by administering monoclonal antibody despite the activation of these cells following infection [38], NK cells may play a role in controlling brucellosis in pregnant mice because the modulation of host immune responses differs between pregnant mice and unpregnant normal mice [39].
The virB4 mutant, which is defective intracellular growth, did not cause abortion in pregnant mice, suggesting that the type IV secretion system of B. abortus contributes to abortion induced by bacterial infection. Our results also showed that the virB4 mutant replicated in the placenta. Presumably, virB4 mutant replicates extracellulary in the placenta due to maternal immunosuppression during pregnancy [40]. Wild type B. abortus internalizes into macrophages by swimming on the cell surface with generalized membrane ruffling for several minutes, a process termed "swimming internalization", after which the bacteria are enclosed by macropinosomes [41]. In this period, the phagosomal membrane continues to maintain a dynamic state. Lipid raft-associated molecules, such as glycosylphosphatidylinositol (GPI)-anchored proteins, GM1 ganglioside and cholesterol, have been shown to be selectively incorporated into macropinosomes containing B. abortus. While late endosomal marker lysosomal-associated membrane protein (LAMP)-1 and host cell transmembrane proteins are excluded from the macropinosomes. The disruption of lipid rafts on macrophages markedly inhibits the type IV secretion system-dependent macropinocytosis and intracellular replication [21]. Thus, B. abortus may internalize into TGC by lipid rafts-dependent pathway like that of macrophages, and such internalization and intracellular replication of B. abortus may contribute to induction of IFN-γ in pregnant mice. Our recent study showed that B. abortus Hsp60 (a member of the GroEL family of chaperonins) was secreted on the bacterial surface through the type IV system and that it can interact with the cellular prion protein (PrPC) on macrophages [41]. This interaction is important for establishment of B. abortus infection. As prion protein accumulates in the placenta [42], PrPC may be involved in the internalization of B. abortus into TGC.
Many vertically transmitted pathogens are detrimental to the welfare of livestock. Among these are Listeria monocytogenes, Chlamydophila abortus, Toxoplasma gondii, Neospora caninum, bovine viral diarrhoea virus (BVDV) and Border disease virus (BDV). These diverse organisms, ranging from bacteria to apicomplexan protozoa and RNA viruses, share some common features. They give rise to mild and often inapparent disease in non-pregnant animals but cause abortion in ruminants and also in human beings in some cases. However, the disease processes and mechanisms by which they cause abortion are different [39]. IFN-γ and TNF-α are central to host immunity against listeriosis, but can compromise fetal survival [13]. Pregnant mice mount lower levels of protective Th 1 cytokines and are unable to eliminate the pathogen. This leads to severe necrotizing hepatitis and bacteria-induced placental necrosis, increasing the incidence of postimplantation loss and a poor pregnancy response [43].
Conclusion
B. abortus appears to be able to exploit the very precise regulatory processes that control immune responsiveness during pregnancy. Our observations suggest that the generation of inflammatory responses in the maternal periphery and at the maternofetal interface are detrimental for fetal survival and this has important implications for pathogen control during pregnancy. Therefore, vaccination strategies to control B. abortus have to be carefully designed and implemented appropriately to prevent the generation of inappropriate inflammatory responses that might affect the outcome of pregnancy. Thus, using our mice model, we should investigate the mechanism of abortion due to B. abortus infection in more detail and develop on methods for the prevention of abortion that are safer than current control programs.
Methods
Bacterial strains and mice
All B. abortus derivatives were from 544 (ATCC23448) smooth virulent B. abrotus biovar 1 strains. Ba598 (544ΔvirB4) and B. abortus vaccine strain S19 were used in this study [20]. B. abortus strains were maintained as frozen glycerol stocks and cultured on Brucella broth (Becton Dickinson) or Brucella broth containing 1.5% agar.
Mice
Six to ten-week-old ICR female mice were individually mated to 6- to 10-week-old ICR male mice. The parent mice were obtained from CLEA Japan. Day 0.5 of gestation was the day the vaginal plug was observed. The normal gestational time for these mice is 19 days.
Virulence in pregnant mice
Groups of four or five pregnant mice were infected intraperitoneally with approximately 104 CFU of brucellae in 0.1 ml saline at the indicated days of gestation. On day 18.5 of gestation, their fetus, placenta, spleen, liver, lungs, and kidneys were removed and homogenized in saline. Tissue homogenates were serially diluted with PBS and plated on Brucella agar to count the number of CFU in each spleen. Fetuses were determined to be alive if there was a heartbeat, and dead if there was no heartbeat.
Histology
Placentas not used for bacterial culture were fixed in situ within uteri in 10% neutral buffered formalin and processed routinely for histologic examination and scoring using Meyer's hematoxylin stain. Specific labeling of B. abortus in placental sections was performed using the Dako EnVision System with anti-B. abortus rabbit serum [21].
Cytokine measurement
The level of serum IFN-γ and TNF-α were measured for infected and uninfected virgin and pregnant ICR mice. Groups of five mice were infected intraperitoneally with approximately 104 CFU of brucellae in 0.1 ml saline on day 4.5 of gestation, and blood was collected at 1, 3, 5, 7 or 10 days after infection. Blood was collected at the same time periods for uninfected mice. On day 18.5 of gestation, their uteruses were removed and a judgment was made as to whether mice were pregnant or not. Serum levels of IFN-γ and TNF-α were measured with an enzyme linked immunosorbent assay (ELISA) kit (PIERCE Endogen) according to the instructions of the manufacturer.
In vivo depletion of IFN-γ
IFN-γ was neutralized in the mice by administering anti-mouse IFN-γ monoclonal antibody (clone HB170) in vivo using 1 mg of antibody in a volume of 0.5 ml intraperitoneally 24 h before infection. Control mice were given 1 mg of normal rat IgG in 0.5 ml according to the same injection schedule as the corresponding anti-IFN-γ monoclonal antibody treatment groups. Bacterial infection was done as described above. Blood was collected at 1, 3, 5, 7 or 10 days after infection, and the serum levels of IFN-γ and TNF-α were measured with an ELISA kit as described above. On day 18.5 of gestation, their uteruses were removed and a judgment was made as to whether mice were pregnant or not. Fetuses were determined to be alive if there was a heartbeat, and dead if there was no heartbeat.
Statistical analysis
All statistical analysis were calculated using Student t test.
Authors' contributions
SK carried out all virulence assays and analysis of the data. DSL and KW carried out mice mating and maintenances. HF carried out pathological experiments. HS and MW conceived of the study and participated in its design and coordination.
Acknowledgements
We thank Drs. Neeraj Rana and Alexander Cox for critical reading of the manuscript, and Dr Yoshitaka Omata for valuable discussion. This work was supported, in part, by grants from Special Coordination Funds for Promoting Science and Technology, Ministry of Education, Culture, Sports, Science, and Scientific Research (16017207 and 16790250), and the Japan Society for the Promotion of Science, and a grant from the Akiyama Fundation.
Figures and Tables
Figure 1 B. abortus infection causes abortion in pregnant mice. Groups of four pregnant mice were infected with bacteria intraperitoneally at each gestation time point. On day 18.5 of gestation, their fetuses were removed. Fetuses were determined to be alive if there was a heartbeat, and dead if there was no heartbeat. Fetuses underlined were aborted.
Figure 2 B. abortus infection predominantly in placenta. Groups of four pregnant mice were infected with bacteria intraperitoneally on day 6.5 of gestation. On day 18.5 of gestation, their placenta, spleen, liver, lungs, and kidneys were removed and homogenized in saline. Tissue homogenates were serially diluted with PBS and plated on Brucella agar to count the number of CFU in each organ. Statistically significant differences bacterial numbers between placenta and spleen are indicated by asterisks (*, P < 0.01).
Figure 3 B. abortus predominantly invades into trophoblast giant cells in placenta. Placentas not used for bacterial culture were fixed in situ within uteri in 10% neutral buffered formalin and processed routinely for histologic examination and scoring using Meyer's hematoxylin stain. Specific labeling of B. abortus in placental sections was performed using the Dako EnVision System with anti-B. abortus rabbit serum and replicated bacteria are shown in brown. Normal (A and B) and infected placenta (C and D) are shown. Panels (B) and (D) indicate magnified images in panels (A) and (B). Arrows show trophoblast giant cells.
Figure 4 Transient increase in IFN-γ in pregnant mice induced by B. abortus infection. The IFN-γ (A), and TNF-α (B) serum levels were measured in each mouse by ELISA at the indicated numbers of days after infection. The mean and SE for groups of five mice are shown.
Figure 5 Preventing abortion by neutralizing IFN-γ. IFN-γ was neutralized in mice by administering anti-mouse IFN-γ monoclonal antibody (clone HB170) in vivo using 1 mg of antibody in a volume of 0.5 ml intraperitoneally 24 h before infection. Control mice were given 1 mg of normal rat IgG in 0.5 ml according to the same injection schedule as the corresponding anti-IFN-γ monoclonal antibody treatment groups. (A) IFN-γ in serum measured by ELISA. The mean and SE for groups of five mice are shown (B) Number of fetuses. Statistically significant differences between the untreated control and antibody treated mice are indicated by asterisks (*, P < 0.01).
Table 1 B. abortus infection in pregnant mice.
Fetus Placenta Spleen
Days of gestation Mouse No. Alive Aborted Alive Aborted
No. Weight (g)* Bacterial No. (log)* No. Weight (g)* Bacterial No. (log)* Weight (g)* Bacterial No. (log)* Weight (g)* Bacterial No. (log)* Weight (g) Bacterial No. (log)
3.5 1 14 1.21 (±0.10) 3.82 (± 0.36) 1 0.99 3.84 0.12 (±0.02) 7.48 (±0.24) 0.13 7.51 0.65 5.93
2 0 - - 14** - - - - - - 0.46 6.11
3 11 1.32 (±0.13) 3.11 (±0.24) 1 1.13 3.27 0.15 (±0.01) 7.52 (±0.35) 0.14 7.39 0.67 6.12
4 16 1.42 (±0.11) 2.18 (±0.13) 0 - - 0.13 (±0.02) 7.55 (±0.41) - - 0.32 5.62
4.5 1 0 - - 16** - - - - - - 0.55 5.30
2 1 1.33 3.32 14 0.82 (±0.26) 3.56 (±0.27) 0.14 7.42 0.18 (±0.02) 5.36 (±0.23) 0.82 6.12
3 0 - - 17** - - - - - - 0.51 5.14
4 0 - - 15** - - - - - - 0.92 6.11
6.5 1 12 1.12 (±0.14) 4.21 (±0.31) 5 1.00 (±0.22) 4.18 (±0.20) 0.15 (±0.01) 7.35 (±0.24) 0.11 (±0.01) 7.18 (±0.22) 0.39 6.12
2 16 1.42 (±0.21) 3.23 (±0.33) 2 1.00 (±0.14) 4.13 (±0.31) 0.12 (±0.02) 4.72 (±0.32) 0.09 (±0.02) 7.33 (±0.41) 0.18 5.41
3 0 - - 15 0.61 (±0.18) 5.22 (±0.27) - - 0.11 (±0.01) 7.24 (±0.38) 1.22 6.43
4 12 1.38 (+0.10) 4.14 (+0.29) 1 0.95 4.36 0.14 (±0.02) 7.52 (±0.31) 0.11 7.43 0.47 6.19
9.5 1 0 - - 14** - - - - - - 0.70 6.12
2 10 1.23 (±0.16) 3.53 (±0.14) 5 0.93 (±0.10) 4.11 (±0.24) 0.11 (±0.01) 7.52 (±0.32) 0.08 (±0.02) 7.55 (±0.31) 0.65 6.13
3 15 1.02 (±0.08) 4.29 (±0.21) 3 1.02 (±0.21) 4.23 (±0.19) 0.11 (±0.01) 7.54 (±0.26) 0.12 (±0.02) 7.53 (±0.27) 0.46 6.11
4 12 1.43 (±0.19) 3.14 (±0.16) 0 - - 0.14 (±0.02) 7.35 (±0.38) - - 0.26 5.19
14.5 1 12 1.61 (±0.18) 4.17 (±0.19) 0 - - 0.11 (±0.01) 7.22 (±0.29) - - 0.18 6.15
2 14 1.32 (±0.12) 3.42 (±0.21) 0 - - 0.11 (±0.01) 7.11 (±0.31) - - 0.29 6.81
3 10 1.72 (±0.21) 2.24 (±0.14) 0 - - 0.14 (±0.01) 6.13 (±0.27) - - 0.14 5.42
4 11 1.43 (±0.14) 3.17 (±0.17) 0 - - 0.11 (±0.02) 7.21 (±0.39) - - 0.26 6.91
Cont. 1 15 1.45 (±0.11) - - - - 0.13 (±0.01) - - - 0.20 -
2 11 1.59 (±0.16) - - - - 0.12 (±0.01) - - - 0.14 -
3 14 1.63 (±0.12) - - - - 0.13 (±0.02) - - - 0.15 -
*Average number of fetuses or placentas.
**Number of mummifications.
Table 2 Infection of pregnant mice with B. abortus virB4 mutant or vaccine strain S19 on day 4.5 of gestation.
Fetus Placenta Spleen
Strain Mouse No. Alive Aborted Alive Aborted
No. Weight (g)* Bacterial No. (log)* No. Weight (g)* Bacterial No. (log)* Weight (g)* Bacterial No. (log)* Weight (g)* Bacterial No. (log)* Weight (g) Bacterial No. (log)
virB4 mutant 1 15 1.51 (±0.21) 3.25 (±0.16) 0 - - 0.11 (±0.01) 7.51 (±0.26) - - 0.16 5.41
2 16 1.43 (±0.18) 3.24 (±0.15) 0 - - 0.14 (±0.01) 5.15 (±0.27) - - 0.20 5.38
3 16 1.62 (±0.13) 1.14 (±0.08) 0 - - 0.12 (±0.02) 5.52 (±0.34) - - 0.15 4.80
4 13 1.51 (±0.15) 2.52 (±0.12) 0 - - 0.12 (±0.02) 7.25 (±0.36) - - 0.21 5.61
5 14 1.72 (±0.17) 3.64 (±0.24) 0 - - 0.14 (±0.01) 7.48 (±0.33) - - 0.18 5.41
S19 1 14 1.43 (±0.14)) 2.54 (±0.24) 0 - - 0.13 (±0.01) 7.52 (±0.27) - - 0.50 7.12
2 14 1.45 (±0.18) 2.24 (±0.20) 0 - - 0.10 (±0.01) 7.51 (±0.35) - - 0.20 6.56
3 15 1.35 (±0.14) 3.23 (±0.16) 0 - - 0.10 (±0.02) 7.58 (±0.41) - - 0.21 6.63
4 12 1.34 (±0.17) 3.17 (±0.27) 1 1.28 3.23 0.11 (±0.01) 7.64 (±0.21) 0.10 7.39 0.26 6.53
5 14 1.35 (±0.12) 3.26 (±0.12) 0 - - 0.10 (±0.01) 7.62 (±0.29) - - 0.24 6.96
*Average number of fetuses or placentas.
==== Refs
Delrue RM Martinez-Lorenzo M Lestrate P Danese I Bielarz V Mertens P De Bolle X Tibor A Gorvel JP Letesson JJ Identification of Brucella spp. genes involved in intracellular trafficking Cell Microbiol 2001 3 487 497 11437834 10.1046/j.1462-5822.2001.00131.x
Detileux PG Deyoe BL Cheville NF Entry and intracellular localization of Brucella spp. in Vero cells: fluorescence and electron microscopy Vet Pathol 1990 27 317 328 2122572
Zavala I Nava A Guerra J Quiros C Brucellosis Infect Dis Clin N Am 1994 8 225 241
Enright FM Nielsen K and Duncan JR The pathogenesis and pathibiology of Brucella infection in domestic animals Animal Brucellosis 1990 CRC Press, Boca Raton, FL 301 320
Tobias L Cordes DO Schurig GG Placental pathology of the pregnant mouse inoculated with Brucella abortus strain 2308 Vet Pathol 1993 30 119 129 8470334
Pizarro-Cerda J Moreno E Sanguedolce V Mege JL Gorvel JP Virulent Brucella abortus prevents lysosome fusion and is distributed within autophagosome-like compartments Infect Immun 1998 66 2387 2392 9573138
Ugalde RA Intracellular lifestyle of Brucella spp. common genes with other animal pathogens, plant pathogens, and endosymbionts Microbes Infect 1999 1 1211 1219 10580277 10.1016/S1286-4579(99)00240-3
Pizarro-Cerda J Meresse S Parton RG van der Goot G Sola-Landa A Lopez-Goni I Moreno E Grovel JP Brucella abortus transits through the autophagic pathway and replicates in the endoplasmic reticulum of nonprofessional phagocytes Infect Immun 1998 66 5711 5724 9826346
Comerci DJ Martinez-Lorenzo MJ Sieira R Gorvel JP Ugalde RA Essential role of the VirB machinery in the maturation of the Brucella abortus-containing vacuole Cell Microbiol 2001 3 159 168 11260139 10.1046/j.1462-5822.2001.00102.x
Dorn BR Dunn WA JrProgulske-Fox A Bacterial interactions with the autophagic pathway Cell Microbiol 2002 4 1 10 11856168 10.1046/j.1462-5822.2002.00164.x
Weinberg ED Pregnancy-associated immune suppression: risks and mechanisms Microb Pathog 1987 3 393 397 3332911 10.1016/0882-4010(87)90009-X
Wegmann TG Lin H Guilbert L Mosmann TR Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a Th2 phenomenon? Immunol Today 1993 14 353 356 8363725 10.1016/0167-5699(93)90235-D
Raghupathy R Th1-type immunity is incompatible with successful pregnancy Immunol Today 1997 18 478 482 9357139 10.1016/S0167-5699(97)01127-4
Sano M Mitsuyama M Watanabe Y Nomoto K Impairment of T cell-mediated immunity to Listeria monocytogenes in pregnant mice Microbiol Immunol 1986 30 165 176 3086677
Krishnan L Guilbert LJ Russell AS Wegmann TG Mosmann TR Belosevic M Pregnancy impairs resistance of C57BL/6 mice to Leishmania major infection and causes decreased antigen-specific IFN-gamma response and increased production of T helper 2 cytokines J Immunol 1996 156 644 652 8543816
Zhan Y Cheers C Endogenous gamma interferon mediates resistance to Brucella abortus infection Infect Immun 1993 61 4899 4901 8406893
Stevens MG Pugh GW JrTabatabai LB Effects of γ-interferon and indomethacin in preventing Brucella abortus infections in mice Infect Immun 1992 60 4407 4409 1398950
O'Callaghan D Cazevieille C Allardet-Servent A Boschiroli ML Bourg G Foulongne V Frutos P Kulakov Y Ramuz M A homologue of the Agrobacterium tumefaciens VirB and Bordetella pertussis Ptl type IV secretion systems is essential for intracellular survival of Brucella suis Mol Microbiol 1999 33 1210 1220 10510235 10.1046/j.1365-2958.1999.01569.x
Sieira R Comerci DJ Sanchez DO Ugalde RA A homologue of an operon required for DNA transfer in Agrobacterium tumefaciens is required in Brucella abortus for virulence and intracellular multiplication J Bacteriol 2000 182 4849 4855 10940027 10.1128/JB.182.17.4849-4855.2000
Watarai M Makino S-I Shirahata T An essential virulence protein of Brucella abortus, VirB4, requires an intact nucleoside-triphosphate-binding domain Microbiology 2002 184 1439 1446
Watarai M Makino S-I Fujii Y Okamoto K Shirahata T Modulation of Brucella-induced macropinocytosis by lipid rafts mediates intracellular replication Cell Microbiol 2002 4 341 356 12067319 10.1046/j.1462-5822.2002.00195.x
Bosseray N Plommet M Brucella suis S2, Brucella melitensis Rev. 1 and Brucella abortus S19 living vaccines: residual virulence and immunity induced against three Brucella species challenge strains in mice Vaccine 1990 8 462 468 2123586 10.1016/0264-410X(90)90247-J
Jones SM Winter AJ Survival of virulent and attenuated strains of Brucella abortus in normal and gamma interferon-activated murine peritoneal macrophages Infect Immun 1992 60 3011 3014 1612769
Quinn HE Ellis JT Smith NC Neospora caninum: a cause of immune-mediated failure of pregnancy? Trend Parasitol 2002 18 391 394 10.1016/S1471-4922(02)02324-3
Smith H Williams AE Pearce JH Keppie J Harris-Smith PW Fitz-George RB Witt K Foetal erythritol: a cause of the localization of Brucella abortus Foetal erythritol: in bovine contagious abortion Nature 1962 193 47 49 13914250
Keppie J Williams AE Witt K Smith H The role of erythritol in tissue localization of the brucellae Br J Exp Pathol 1965 46 104 108 14295553
Payne JM Changes in the rat placenta and foetus following experimental infection with Brucella abortus J Pathol Bacteriol 1957 73 141 154
Soares MJ Chapman BM Rasmussen CA Dai G Kamei T Orwig KE Differentiation of trophoblast endocrine cells Placenta 1996 17 277 289 8829210 10.1016/S0143-4004(96)90051-X
Cross JC Genetic insights into trophoblast differentiation and placental morphogenesis Semin Cell Dev Biol 2000 11 105 113 10873707 10.1006/scdb.2000.0156
Ain R Canham LN Soares MJ Gestation stage-dependent intrauterine trophoblast cell invasion in the rat and mouse: novel endocrine phenotype and regulation Dev Biol 2003 260 176 190 12885563 10.1016/S0012-1606(03)00210-0
Ashkar AA Croy BA Interferon-γ contributes to the normalcy of murine pregnancy Biol Reprod 1999 61 493 502 10411532
Ashkar AA Di Santo JP Croy BA Interferon γ contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy J Exp Med 2000 192 259 269 10899912 10.1084/jem.192.2.259
Ashkar AA Croy BA Functions of uterine natural killer cells are mediated by interferon gamma production during murine pregnancy Semin Immunol 2001 13 235 241 11437631 10.1006/smim.2000.0319
Athanassakis I Papadimitriou L Bouris G Vassiliadis S Interferon-γ induces differentiation of ectoplacental cone cells to phenotypically distinct trophoblasts Dev Comp Immunol 2000 24 663 672 10831800 10.1016/S0145-305X(00)00023-9
Baldwin CL Winter AJ Macrophages and Brucella Immunol Ser 1994 60 363 380 8251581
Splitter G Oliveira S Carey M Miller C Ko J Covert J T lymphocyte mediated protection against facultative intracellular bacteria Vet Immunol Immunopathol 1996 54 309 319 8988876 10.1016/S0165-2427(96)05703-0
Boehm U Klamp T Groot M Howard JC Cellular responses to interferon-γ Annu Rev Immunol 1997 15 749 795 9143706 10.1146/annurev.immunol.15.1.749
Fernandes DM Benson R Baldwin CL Lack of a role for natural killer cells in early control of Brucella abortus Infect Immun 1995 63 4029 4033 7558315
Entrican G Immune regulation during pregnancy and host-pathogen interactions in infectious abortion J Comp Path 2002 126 79 94 11944996 10.1053/jcpa.2001.0539
Medawar PB Some immunological and endocrinological problems raised by the evolution of viviparity in vertebrates Symp Soci Exp Biol Med 1953 7 320 338
Watarai M Kim S Erdenebaatar J Makino S Horiuchi M Shirahata T Sakaguchi S Katamine S Cellular prion protein promotes Brucella infection into macrophages J Exp Med 2003 198 5 17 12847134 10.1084/jem.20021980
Tuo W O'Rourke KI Zhuang D Cheevers WP Spraker TR Knowles DP Pregnancy status and fetal prion genetics determine PrPSc accumulation in placentomes of scrapie-infected sheep Proc Natl Acad Sci USA 2002 99 6310 6315 11959902 10.1073/pnas.072071199
Abram M Schluter D Vuckovic D Wraber B Doric M Deckert M Murine model of pregnancy-associated Listeria monocytogenes infection FEMS Immunol Med Microbiol 2003 35 177 182 12648834 10.1016/S0928-8244(02)00449-2
| 15869716 | PMC1090583 | CC BY | 2021-01-04 16:03:39 | no | BMC Microbiol. 2005 May 4; 5:22 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-22 | oa_comm |
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-121582631110.1186/1471-2288-5-12Research ArticleWho wants to join preventive trials? – Experience from the Estonian Postmenopausal Hormone Therapy Trial [ISRCTN35338757] Hovi Sirpa-Liisa [email protected] Matti [email protected] Piret [email protected] Mati [email protected] Elina [email protected] National Research and Development Centre for Welfare and Health, STAKES, Health and Social Services, PO Box 220, FI-00531 Helsinki, Finland2 School of Public Health, University of Tampere, FI-33014 Tampere, Finland3 Department of Epidemiology and Biostatistcs, National Institute for Health Development, Hiiu 42, 11619 Tallinn, Estonia4 Estonian Center of Excellence in Behavioral and Health Sciences, Tartu-Tallinn, Estonia2005 12 4 2005 5 12 12 13 9 2004 12 4 2005 Copyright © 2005 Hovi et al; licensee BioMed Central Ltd.2005Hovi 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 interest of patients in participating in randomized clinical trials involving treatments has been widely studied, but there has been much less research on interest in preventive trials. The objective of this study was to find out how many women would be interested in a trial involving postmenopausal hormone therapy (PHT) and how the women's background characteristics and opinions correlated to their interest.
Methods
The data come from recruitment questionnaires (n = 2000) sent to women in Estonia in 1998. A random sample of women aged 45 to 64 was drawn from the Population Registry. The trial is a two-group randomized trial comparing estrogen-progestogen therapy with placebo or no drugs. A brief description of the study was attached to the questionnaires. Women were not told at this stage of the recruitment which group they would be assigned to, however, they were told of the chance to receive either hormone, placebo or no treatment.
Results
After two reminders, 1312 women (66%) responded. Eleven percent of the women approached (17% of the respondents) were interested in joining the trial, and 8% wanted more information before deciding. When the 225 women who stated clearly that they were interested in joining and the 553 women who said they were not interested were compared, it was found that interested women were younger and, adjusting for age, that more had given birth; in other respects, the sociodemographic characteristics and health habits of the interested women were similar to those of the non-interested women. The interested women had made more use of more health services, calcium preparations and PHT, they were more often overweight, and more had chronic diseases and reported symptoms. Interested women's opinions on the menopause were more negative, and they favoured PHT more than the non-interested women.
Conclusion
Unlike the situation described in previous reports on preventive trials, in this case Estonian women interested in participating in a PHT trial were not healthier than other women. This suggests that trials involving PHT are more similar to treatment trials than to preventive trials. In a randomized controlled trial, more information should be obtained from those women who decline to participate.
==== Body
Background
Randomized controlled trials (RCTs) are needed so as to avoid bias in scientific research. There are many studies and reviews on factors that promote or hinder interest in participating in treatment trials – both in regard to patients and physicians [see e.g. [1-4]]. Men, patients who are older, less educated and from lower socioeconomic backgrounds, non-whites, smokers, and persons lacking adequate social support are more willing to participate in clinical trials. Furthermore, they tend to be more severely ill than those who do not participate. Identified obstacles include patients' and physicians' disapproval of patients' serving as research subjects, a lack of altruistic motives, distrust of the medical profession, a lack of knowledge of what is required of trial participants, and preference for a certain treatment. Many patients do not understand the reasons why treatments should be allocated at random, and this is an important reason why patients choose not to join randomized clinical trials. [5-8]. Disinterest on the part of patients and physicians in participating in clinical trials constitutes a threat to the generalizability of RCTs [9].
There has been much less research on participation in trials studying preventive measures (preventive trials). The data available thus far suggest that there is a difference in the types of people who join preventive trials and treatment trials. Participants in preventive trials tend to be better off than non-participants as regards socioeconomic situation, health habits and health [10-14]. However, most evaluative studies have failed to document adequately the characteristics of persons who were eligible but did not participate [1]. Less is known of physicians' motivation as regards including or encouraging people to participate in preventive trials [15]. Preventive drug therapy, and thus trials involving such drugs, are likely to increase in the future, and more information is needed on who wants to take part in preventive trials. Such data are useful in order to increase the recruitment rate as well as to interpret trial results.
By means of a mailed questionnaire in the Estonian Postmenopausal Hormone Therapy (EPHT)-trial, we recruited healthy 45–64-year-old Estonian women for studying the long-term (5-year) health effects of PHT. The trial investigated the immediate effects of PHT on well-being and symptoms, impacts on the experience of the climacteric, on aging and partner relationships, and influences on the use of health services. Furthermore, the trial investigated the placebo effect and trial effect by means of the design as well as their effect on recruitment and compliance. The object of this paper is to report on how many women were interested in participating in such a trial and how sociodemographic characteristics, health, health habits, health services utilization and opinions regarding the menopause and aging influence women's interest.
Methods
The trial is a two-group randomized trial comparing oestrogen-progestogen therapy to placebo or no drugs carried out in Estonia. The participants were allocated to four random arms forming two groups: the blind group was given an active drug or a placebo, and the non-blind group was given an active drug or no treatment. Women were informed that drugs would be provided for five years. A random sample of women aged 45 to 64 was drawn from the Population Registry and questionnaires were sent to the women. A brief description of the study was attached to the questionnaire and women were invited to participate in the trial if they were found to be eligible. In the letter, women were told about the sampling; that the trial investigated the health problems of perimenopausal and older women in Estonia, especially the long-term effect of hormonal replacement therapy after cessation of periods; and that the drugs would be provided for five years. In this first letter, women were not presented with the trial design in detail, but they were told that the women will be randomly divided into groups of hormone treatment, placebo, or no tablets. Further, they were told that physician examinations will be provided annually, and the possible risks and benefits of PHT were explained. The study plan had been accepted in the Committee of Medical Ethics in Tallinn. The first 2000 women, in the sample of spring 1998, were sent a more detailed questionnaire, which provided more information on the respondents. After two reminders, the response rate was 66% (n = 1 312).
The questions used in this article were structured, with fixed alternatives. As regards current health, women were asked to choose between very good, good, average, poor or very poor; "very good" and "good" were later combined to "good". To measure health status, women were asked if they had or had had chronic diseases, such as cancer (breast, uterus, ovary), myocardial infarction, cardiac failure, hypertension, stroke, thromboses, liver diseases, renal failure, diabetes or icterus. In the case of other symptoms, women were given a checklist of 18 different symptoms or health problems and asked to choose all that they had experienced within the past two weeks. In the case of health habits, questions on current smoking, alcohol consumption and exercise were asked. The choices for current smoking were: no, yes every now and then, and yes daily, how many cigarettes per day. The choices for alcohol consumption were not at all, low, moderate, fairly high and high. Exercise in leisure-time was elicited using the choices not at all, a little, some, a lot, a large amount; "a lot" and "a large amount" were later combined to "plenty of exercise". Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters.
Women's opinions on the menopause were elicited using the statements: "The menopause is a normal phase in a woman's life, and usually it does not need treatment by a doctor", and "A woman does not loose her femininity during the menopause". Women's opinions concerning PHT were elicited using the statements: "PHT effectively prevents osteoporosis ", "PHT should be given to all middle-aged women with (menopausal) symptoms", and "PHT should be given to most postmenopausal women". The possible answers were: I totally agree, I agree somewhat, I don't know, I disagree somewhat, and I totally disagree. In the analyses, the choices "I totally agree" and "I agree somewhat " were later combined to "agree".
After the previous statements came the remark 'the next questions deal with the menopause and PHT use. Those with normal periods and who do not use PHT may stop here'. This mistake resulted in our many missing answers concerning PHT use (see Table 2), and the percentages in Table 2 are given in two ways.
Table 2 Comparison of health service utilization of women interested and not interested in participating in a randomized PHT preventive trial in Estonia, proportion (%) of women, and age-adjusted odds-ratios (OR).1)
Interested Non-interested OR (95% CI)1)
(n = 225) (n = 553)
% %
Physician visit in past year 76 68 1.53 (1.06–2.20)
Gynaecologist visit in past year 53 39 1.40 (1.01–1.94)
Used calcium drugs in last 2 weeks 32 23 1.49 (1.04–2.13)
Used PHT at some time 2) 26 11 2.62 (1.58–4.35)
Used PHT at some time 3) 16 8 2.05 (1.26–3.35)
1) Reference group: non-interested
2) Excluding missing values, interested n = 90 (40%), non-interested n = 143 (26%), see Methods
3) Including missing values in the denominator
Open-ended questions were also used to ask women what kind of positive and negative features they associated with the menopause. The proportions giving positive or negative aspects are also reported in this article.
Testing the statistical significance of medians was done using Mann-Whitney's U test. Age-adjusted odds ratios (OR) with 95% confidence intervals (CI) were calculated by logistic regression, using interested women as the reference group. SAS 8.0 was used in the analyses.
Results
Altogether 2000 questionnaires were sent and after two reminders, 1312 women (66%) responded (Table 1). Non-respondent women were somewhat older and were somewhat more often residents of the capital than were respondents.
Table 1 Distribution of women by their interest in participating in a randomized PHT preventive trial in Estonia and comparisons of the background characteristics of women according to their interest in participating. 1)
Interested Want more information Do not know Non-interested All respondents No reply
(n = 225) (n = 163) (n = 371) (n = 553) (n = 1312) (n = 688)
(% = 11) (% = 8) (% = 19) (% = 28) (% = 66) (% = 34)
Median age, years 2) 51 53 53* 56*** 54 53*
Lives in the capital, % 68 64 67 62 65 76 3)
Married or cohabiting, % 62 64 63 58 61
≥ 12 years of education, % 64 62 60 61 61
In employment, % 78 72 73 67 70
Given birth, % 92 92 89 854) 87
1) Adjusted for age
2) Statistically significant: * p < 0.05, *** p < 0.0001 compared to the "interested" group
3) Age adjusted OR 0.64 (CI 0.46 0.90), reference group "interested"
4) Age adjusted OR 2.00 (CI 1.16 3.46), reference group "interested"
Of the 1312 respondents, 17% wanted to participate in the trial. Using the whole sample of 2000 women as a basis for calculation, 11% were interested in participating in the trial, 42% did not want to participate,12% wanted to receive more information before deciding, and 28% gave no answer ('do not know'). (Table 1).
The socioeconomic background characteristics of these four groups were very similar, but women in the "do not know" and "non-interested" groups were older than interested women. After adjustment for age, it was found that fewer non-interested than interested women had given birth. When the women who had given a clearly positive or negative answer to the question concerning their interest in participating were compared, it was found that interested women had had more contacts with the health-care system – as measured by visits to a physician and a gynaecologist during the previous year – than had non-interested women. (Table 2). Interested women had also used calcium drugs more often, and they more often reported using PHT at some time than did non-interested women.
Health habits, smoking, alcohol use, and exercise did not vary by interest in participating (Table 3). In both groups, more than half were mildly overweight, but heavy overweight was somewhat more prevalent in those interested. Interested women more often had some chronic disease, such as hypertension, cardiac failure or diabetes. (Table 3). There was no difference in regard to subjective current health, or in the proportion of women who had experienced hot flashes, but interested women more frequently reported depression and sleeplessness.
Table 3 Comparison of health habits and health of women interested and non-interested in participating in a randomized PHT preventive trial in Estonia, proportion (%) of women, and age-adjusted odds-ratios (OR).
Interested Non-interested OR (95% CI) 1)
(n = 225) (n = 553)
% %
Daily smoker 18 14 0.87 (0.56–1.34)
No alcohol 13 20 1.40 (0.88–2.20)
Plenty of exercise in leisure-time 25 28 1.13 (0.78–1.62)
BMI 25–29.9 2) 55 59 0.96 (0.69–1.34)
BMI ≥ 30 2) 26 21 0.68 (0.47–0.99)
Current health good 24 24 1.27 (0.87–1.86)
Some chronic disease 73 67 0.70 (0.49–1.00)
Tiredness 50 45 0.83 (0.60–1.15)
Irritability 29 23 0.80 (0.56–1.16)
Depression 26 19 0.65 (0.45–0.96)
Headache 36 32 0.96 (0.68–1.34)
Sweating 32 27 0.78 (0.55–1.11)
Hot flashes 28 22 0.74 (0.51–1.07)
Sleeplessness 26 20 0.67 (0.46–0.97)
No symptoms 2 9 4.04 (1.57–10.44)
1) Reference group: "interested"
2) kg/m2
Fewer women who were interested in participating in the trial agreed that the menopause is a normal phase, and more gave negative aspects of the menopause than did non-interested women (Table 4). A more favourable opinion of PHT was held by interested than by non-interested respondents.
Table 4 Comparison of opinions on aging and attitudes to PHT of women interested and non-interested in participating in a randomized PHT preventive trial in Estonia, proportion (%) of women, and age-adjusted odds-ratios (OR).1)
Interested Non-interested OR (95% CI) 2)
(n = 225) (n = 553)
% %
Menopause is a normal phase 46 56 0.68 (0.50–0.94)
Women do not loose femininity in menopause 52 51 1.01 (0.73–1.39)
PHT prevents osteoporosis 27 9 4.27 (2.75–6.64)
PHT should be given to all women with symptoms 32 12 3.20 (2.16–4.74)
PHT should be given to all postmenopausal women 17 3 5.88 (3.18–10.88)
Gave positive aspects of menopause 20 17 1.38 (0.92–2.08)
Gave negative aspects of menopause 39 28 1.82 (1.30–2.57)
1) Missing values excluded from the denominator. The proportion of missing values varied from 17 to 21% in the "interested" group, and from 23 to 26% in the "non-interested" group.
2) Reference group: "non-interested"
The "want more information" group was similar to the group of interested women in regard to the variables studied, but they had had fewer gynaecologist appointments in the last 12 months (40% vs. 53%). The "do not know" group exhibited a clearly greater difference from interested women: fewer of them had ever made use of PHT and calcium drugs; they suffered less from irritability, depression, joint pain, sleeplessness, sweating and hot flashes; they had had fewer appointments with gynaecologists because of menopausal symptoms; and they reported fewer both positive and negative aspects of the menopause.
Discussion
In the case of many background characteristics, the interested and non-interested women were similar to each other, or the differences between them were small. The major differences between them were in age, health, use of health services, experience and attitude towards the menopause. When compared to non-interested respondents, interested women were younger, and they suffered from poorer health in terms of chronic diseases, more reported symptoms, and more visits to a physician. Interested women had more negative experiences with the menopause and a more positive attitude to PHT, and they had also more often used PHT than had non-interested women.
As in previous preventive trials, the interested women were younger than those who were not interested [1]. But in contrast to Britton et al. [1], who found that interested women had a more healthy lifestyle, we did not find differences in health habits – except with regard to overweight. The use of health services positively correlated with women's willingness to join this trial. It seems that women's contacts to health services may increase their willingness to join a trial; or use of health services may result from their poorer health: interested women had more chronic diseases and more symptoms. Or women expected some benefits from the trial. In an imaginary trial of PHT, the advantages that the women expected from the treatment was more important than how benefits were described [16].
Interested women's negative experiences with the menopause and positive attitude to PHT, as well as a positive attitude towards PHT on the part of gynaecologists [17], may have influenced women's interest in the PHT trial. By contrast, a fifth of the respondents had no opinion concerning PHT. Its use is still infrequent in Estonia, and knowledge of PHT is likely to have been scant.
A limitation for generalizing the study results may result from our particular trial design and target group. However, the trial design of blind and non-blind groups was not presented in the invitation letter. Women were not told at this stage of the recruitment which group they would be assigned to, however, they were told of the chance to receive either hormone, placebo or no treatment. This trial concerned only mid-aged women, and different factors may influence men, or young and old people. The instruction to stop filling in the questionnaire if the women had regular menstruation and no use for hormone therapy resulted in mainly postmenopausal women being included in this report; PHT use in Estonia was low at the time of the questionnaire.
When one is recruiting participants for a randomized controlled trial, more information should be obtained from those who do not enter the trial. Ellenberg [18] argues that information involving the selection process should be obtained at each stage of selection, beginning with the screening of potential participants and proceeding to the final enrolment of those who agree to take part. This process may establish some basis for judging limits when one is generalizing results of an intervention trial. In the present population-based study, the characteristics of persons who did not return the questionnaire remain largely unknown. We know that they were somewhat older and were more often residents of the capital than were those who were interested in joining the trial. Only 52% of people living in the capital have Estonian as their home language, and language problems in this area may explain the lower response rate to our Estonian-language questionnaires.
This preventive randomized controlled trial differed from previous preventive trials in that interested women had more chronic diseases and symptoms. In this respect they were more similar to the participants of treatment trials, in which interested persons tend to be sicker rather than more interested. Possibly some women did not regard our trial as a preventive trial but wished for better care or treatment than they would receive outside the trial – as is the case in treatment trials [3]
List of abbreviations used
Estonian Postmenopausal Hormone Therapy trial EPHT trial
Postmenopausal hormone therapy PHT
Randomized controlled trial RCT
Women's International Study of long-Duration Oestrogen after Menopause WISDOM
Odds ratio OR
Confidence Interval CI
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SLH participated in the design and the conducting of the study, acquired and analysed the data, and drafted the manuscript; MH participated in the trial design and gave critical comments on the manuscript; PV participated in the study design; MR participated in the study design and acquisition of the data; EH conceived the study and participated in its design, participated in the acquisition of the data and gave critical comments on the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Britton A McKee M Black N McPherson K Sanderson C Bain C Threats to applicability of randomised trials: exclusions and selective participation Journal of Health Services & Research Policy 1999 4 112 121 10387403
Ross S Grant A Counsell C Gillespie W Russell I Prescott R Barriers to participation in randomised controlled trials: a systematic review.[see comment] Journal of Clinical Epidemiology 1999 52 1143 1156 10580777 10.1016/S0895-4356(99)00141-9
Ellis PM Attitudes towards and participation in randomised clinical trials in oncology: a review of the literature Annals of Oncology 2000 11 939 945 11038029 10.1023/A:1008342222205
Ruffin MT 4thBaron J Recruiting subjects in cancer prevention and control studies Journal of Cellular Biochemistry - Supplement 2000 34 80 83 10762019
Ellis PM Dowsett SM Butow PN Tattersall MH Attitudes to randomized clinical trials amongst out-patients attending a medical oncology clinic Health Expect 1999 2 33 43 11281873 10.1046/j.1369-6513.1999.00028.x
Fallowfield LJ Jenkins V Brennan C Sawtell M Moynihan C Souhami RL Attitudes of patients to randomised clinical trials of cancer therapy Eur J Cancer 1998 34 1554 1559 9893627 10.1016/S0959-8049(98)00193-2
Featherstone K Donovan JL Random allocation or allocation at random? Patients' perspectives of participation in a randomised controlled trial Bmj 1998 317 1177 1180 9794849
Snowdon C Garcia J Elbourne D Making sense of randomization; responses of parents of critically ill babies to random allocation of treatment in a clinical trial Soc Sci Med 1997 45 1337 1355 9351153 10.1016/S0277-9536(97)00063-4
Torgerson DJ Klaber-Moffett J Russell IT Patient preferences in randomised trials: threat or opportunity? J Health Serv Res Policy 1996 1 194 197 10180870
Hunninghake DB Darby CA Probstfield JL Recruitment experience in clinical trials: literature summary and annotated bibliography Control Clin Trials 1987 8 6S 30S 3326716 10.1016/0197-2456(87)90004-3
Davies G Pyke S Kinmonth AL Effect of non-attenders on the potential of a primary care programme to reduce cardiovascular risk in the population. Family Heart Study Group Bmj 1994 309 1553 1556 7819899
Naslund GK Fredrikson M Hellenius ML de Faire U Characteristics of participating and nonparticipating men in a randomized, controlled diet and exercise intervention trial Scand J Prim Health Care 1994 12 249 254 7863142
Yeomans-Kinney A Vernon SW Frankowski RF Weber DM Bitsura JM Vogel VG Factors related to enrollment in the breast cancer prevention trial at a comprehensive cancer center during the first year of recruitment Cancer 1995 76 46 56 8630876
Pacala JT Judge JO Boult C Factors affecting sample selection in a randomized trial of balance enhancement: the FICSIT Study J Am Geriatr Soc 1996 44 377 382 8636580
Shelton BJ Wofford JL Gosselink CA McClatchey MW Brekke K Conry C Wolfe P Cohen SJ Recruitment and retention of physicians for primary care research J Community Health 2002 27 79 89 11936759 10.1023/A:1014598332211
Wragg JA Robinson EJ Lilford RJ Information presentation and decisions to enter clinical trials: a hypothetical trial of hormone replacement therapy Soc Sci Med 2000 51 453 462 10855931 10.1016/S0277-9536(99)00477-3
Hovi SL Karttunen T Karro H Hemminki E Comparison of Estonian and Finnish physicians' opinions of menopause and hormone therapy Maturitas 2004 49 107 113 15474754 10.1016/j.maturitas.2003.11.004
Ellenberg JH Selection bias in observational and experimental studies Statistics in Medicine 1994 13 557 567 8023035
| 15826311 | PMC1090584 | CC BY | 2021-01-04 16:32:51 | no | BMC Med Res Methodol. 2005 Apr 12; 5:12 | utf-8 | BMC Med Res Methodol | 2,005 | 10.1186/1471-2288-5-12 | oa_comm |
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-181581118610.1186/1471-2474-6-18Research ArticleA 2-year prospective study of patient-relevant outcomes in patients operated on for knee osteoarthritis with tibial osteotomy W-Dahl Annette [email protected] Sören [email protected] Ewa M [email protected] Department of Orthopedics, Lund University Hospital, SE-221 85 Lund, Sweden2005 5 4 2005 6 18 18 11 11 2004 5 4 2005 Copyright © 2005 W-Dahl et al; licensee BioMed Central Ltd.2005W-Dahl 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
Tibial osteotomy is a treatment for younger and/or physically active patients suffering from uni-compartmental knee osteoarthritis. The open wedge osteotomy by the hemicallotasis technique includes the use of external fixation. The use of external fixation has several advantages, as early mobilization and the opportunity for optimal correction. However, the hemicallotasis technique has also been described as a cumbersome procedure for the patient. The aim of this study was to prospectively evaluate patient-relevant outcomes during the first 2 post-operative years. Especially the treatment period, during which external fixation was used, was closely monitored.
Methods
In an uncontrolled study, fifty-eight consecutive patients, 30 men and 28 women (mean age 54 years) were operated on by the hemicallotasis technique were evaluated with the patient-relevant outcome measure Knee injury and Osteoarthritis Outcome Score (KOOS) preoperatively, during the treatment with external fixation, one week after removal of the external fixation, at 6 months, and at one and two years postoperatively.
Results
At the 2-year postoperative follow-up, all subscales of the KOOS were improved (p < 0.001), mostly in pain (41–80 on a 0–100 worst to best scale) and knee-related quality of life (21–61 on a 0–100 worst to best scale), compared to the preoperative status. Significant improvements in pain and other symptoms, function of daily life and quality of life were seen already during the treatment period (mean 98 ± 18 days) with the external fixation. More demanding functions such as kneeling, squatting, jumping and running, were improved first after extraction of the external fixation device and the pins.
Conclusion
Tibial osteotomy by the hemicallotasis technique yields large improvement in self-rated pain, function and quality of life, which persists over two years. Surprisingly, large improvements occurred already during the immediate post-operative period when the external fixation was still used.
==== Body
Background
Tibial osteotomy is a treatment for younger and/or physically active patients suffering from uni-compartmental knee osteoarthritis (OA) [1]. Traditional closed wedge osteotomy is regarded as the golden standard for osteotomy techniques. The closed wedge osteotomy is technically difficult and the degree of correction achieved may be unpredictable. In many cases the rehabilitation time and sick leave period is long [2]. Hemicallotasis osteotomy (HCO) is an open wedge technique, which implies a successive correction of the angle deformity of the knee postoperatively under radiographic control (Figure 1). The advantages of the HCO include an easy surgical technique and improved possibility to achieve the planned correction [3-5]. As the technique is based on the use of an external fixator just below the knee joint, early mobilization and daily living as well as recreational sports activities are possible [1,3]. HCO can be performed in medial and lateral knee OA. However, HCO has also been described as a cumbersome procedure for the patient as well as for the surgeon, due to frequent minor complications requiring frequent follow-ups [4,6]. Little is known from the patients' perspective regarding treatment with external fixation and the outcome of HCO. This information is valuable in helping patients make treatment choices and preparing them for the treatment period.
Figure 1 Hemicallotasis osteotomy a. After correction b. The Orthofix T-garche
The purpose was, in a prospective study, to evaluate pain, function and quality of life during two years in patients operated on for knee OA with tibial osteotomy by the hemicallotasis technique. Especially the treatment period, during which external fixation was used, was closely monitored.
Methods
Patients
58 consecutive patients (30 men and 28 women), mean age 54 years (36–69) were operated on for knee OA by HCO at the Department of Orthopaedics, Lund University Hospital, Sweden were included in a uncontrolled study. Patient characteristics are given in Table 1 and a flow chart of the study design is given in Figure 2.
Table 1 Patient characteristics
All N = 58 Men n = 30 Women n = 28
Age mean (years) 54 54 55
SD 7 8.7 4
BMI mean 29 28.7 29.5
SD 4.5 4 5
Medial OA (n) 49 27 22
Pre HKA mean (degrees) 170 170 171
SD 4.5 5 3.5
Lateral OA (n) 9 3 6
Pre HKA median (degrees) 189 188 189.5
IQR 184–194 184–190 184–194
Known knee trauma (n) 27 16 11
mean age (years) 51 50 53
SD 7 8 5
Not known knee trauma (n) 28 12 16
mean age (years) 58 58 56
SD 5 7 4
Physically active earlier in life (n) 45 25 20
Level of physical activity (n)
recreational 23 13 10
competitive 12 11 1
Joint loading in sport activity (n)a
low 21 7 14
medium 12 6 6
high 12 12 0
Physically active the year prior surgery (n) 22 14 8
Smokers (n) 14 6 8
Working (n) 51 26 25
Full time (n) 44 26 18
Part time (n) 7 0 7
Disability pension (n) 3 0 3
Retired (n) 4 4 0
BMI = Body Mass Index, OA = Osteoarthritis, Pre HKA = Preoperative radiographic Hip-Knee-Ankle angle, IQR = Inter Quartile Range aaccording to Buckwalter & Lane [11].
Figure 2 Flow chart of the study design.
Preoperative information
When a patient was recommended HCO, written and verbal information was given by a specially trained nurse in our outpatient clinic for patients treated with external fixation.
Operation
Four conical pins were inserted, 2 hydroxyapatite (HA)-coated pins in the metaphyseal bone and 2 standard pins (Orthofix® Bussolengo, Italy) in the diaphyseal bone. A 5 – 7-cm longitudinal skin incision for the osteotomy was placed ventral to the tibial tuberosity. After a transverse incision of the periosteum, an osteotomy was performed. The periosteum was sutured and the wound was closed. After draping the pin sites and the incision, the fixator (Orthofix® T-garche) was mounted. For valgus deformity, the surgical procedure was the same except that a fibulotomy was performed 10 – 15 cm below the head of the fibula [7]. The surgical procedure took about 30 minutes. In most cases the patients were discharged the same day.
Postoperative treatment
Once a week during the treatment period the patients visited the outpatient clinic for pin site care, initiation and follow-up of the correction. The correction started 7–10 days postoperatively. The patient made the correction by adjusting one quarter of a turn 4 times per day on a distractor placed at the external fixator, the first turn in the morning and the last turn, not later than 5 pm to avoid pain at night. The patients got instructions how to slow down the correction in case of unacceptable pain. When the desired correction, 4° valgus for the varus knee and 0–2° varus for the valgus knee, was achieved as determined by radiographic hip-knee-ankle angle (HKA-angle) of the knee, the instrument was locked. Eight weeks postoperatively the instrument was dynamised, unlocking the fixation partly to allow micro movement of the osteotomy at weight bearing, in order to stimulate bone healing. About 12 weeks postoperatively, the first healing control by radiography and ultrasound examination was done. If the osteotomy healing was satisfying the patient made a weight bearing test; i.e., walking (with or without crutches) for an extended period of time varying from some hours to some days without the instrument but still with the pins in situ. If no symptoms arouse, the pins were removed at the out patient clinic. If the patient developed symptoms, the fixation was applied again for another 2–4 weeks.
Postoperative exercise
Exercise and physical therapy were recommended to the patients when receiving the preoperative information. Immediately after surgery the patients were informed by a physiotherapist. Full weight-bearing and free mobilization was allowed postoperatively. Physiotherapy was prescribed individually and related to the needs for each patient. The patient and physiotherapist decided jointly the length of the rehabilitation period needed.
Outcome measure
KOOS (Knee injury and Osteoarthritis Outcome Score) [8,9] was used as the primary outcome measure. KOOS is a 42-item self-administrated knee-specific questionnaire based on the WOMAC Index [10]. KOOS was developed to be used for short- and long-term follow-up studies of knee injury and knee OA, and comprises five subscales: Pain, Symptoms, Activities of Daily Life (ADL), Sports and Recreation Function (Sport/Rec) and knee-related Quality of Life (QOL). Standardized answer options are given (5 Likert boxes), and each question gets a score from 0 to 4. A score from 0 to 100 is calculated for each subscale; 100 represent the best results. The KOOS can be downloaded from . A difference of 10 points was considered a clinically significant difference [11].
Physical activity level was categorized as competitive sports, recreational sports or no sport at all. Regular walks were classified as recreational sports. At baseline, information was obtained regarding physical activity level earlier in life and during the year prior to surgery. According to Buckwalter and Lane [12], the activities performed were divided into high (e.g. team handball, high mileage running, football-soccer, rugby, water skiing), medium (e.g. canoeing, horse-riding, downhill skiing, basketball), and low (e.g. swimming, golf, cycling, walking) joint loading. The patient's previously known knee injuries (meniscus and/or anterior cruciate ligament injury) were documented. The patient's current working status (working full time or part-time, disability pension or retired) was documented.
Complications such as a loose pin, delayed healing, pseudoarthrosis, septic arthritis, interrupted treatment, deep venous thrombosis and nervous injury were recorded. A loose pin was defined as a pin, which could be removed by hand without use of a wrench. Pin site infection was classified according to the Checketts- Otterburns classification [13]. Grades 1–3 are classified as minor infections and respond to proper treatment and external fixation can be continued. Major infections include soft tissue and/or bone and the treatment by external fixation may be interrupted. Grade 1 infections were not classified, as a complication as this grade of infection did not required any antibiotic or surgical treatment. Paracetamol and Tramadol were prescribed as analgesics. At the weekly visit, the patients reported the analgesic consumption during the previous week to the specialist nurse at the out patient clinic.
The patients were defined as non-smokers if they, at the preoperative visit, reported that they never had smoked or stopped smoking since more than 6 month. The radiographic hip-knee-ankle (HKA)-angle was determined after the correction was performed and at the 2-year follow-up. A HKA-angle of <180 degrees was indicated varus alignment of the knee.
Design
The KOOS questionnaire was distributed to the 58 patients by mail and returned in a pre-stamped envelope.
The patients were asked to answer the Swedish version of KOOS LK 1.0, preoperatively, one week after removal of the external fixator and the pins, 6 months, one year and two years postoperatively. 50 of the patients were also asked to answer the KOOS during the treatment with external fixator at 4, 7 and 10 weeks postoperatively. The evaluation time points during the treatment with external fixator reflected the different phases of the treatment: a) correction, b) after correction, when the instrument was locked and c) the final stage of the treatment.
The study was approved by the Ethics Committee at the Medical Faculty, Lund University.
Statistical analysis
The underlying data obtained from questionnaires such as the KOOS are ordinal, which implies the use of non-parametric statistics. However, means and standard deviations are often given instead of medians and inter-quartile ranges for this type of questionnaire data. Postoperative change over time were assessed by Friedman's test, and changes between two postoperative measurements were assessed by Wilcoxon's rang test for each of the five KOOS subscales. The level of significance was set to p 0.05. The influence of 7 potential predictor variables (age, BMI, sex, acceptable HKA-angle at the 2-year follow-up, smoking and complications) on improvements at the 2 year follow-up compared to baseline in pain and knee-related QOL was analyzed by means of linear regression analysis. First, the influence of each potential predictor was assessed in simple regression analysis. Those predictors that implied p < 0.20 were considered further in a multivariate regression analysis.
Results
52/58 patients (90 %) were evaluated at the 2-year follow-up (Fig 2). Due to geographic reasons, 14 patients (8 medial OA, 6 lateral OA) were not followed up by radiography including the HKA-angle at 2 years. The mean correction time was 24 (SD 12) days and the mean time in external fixation was 98 (SD 18) days. The desired HKA-angle was achieved in 57/58 patients (septic arthritis, n = 1) after the correction phase. The mean HKA-angle after correction phase was 183 ± 1 degrees for the varus knees (n = 49) and 180 ± 2 degrees for the valgus knees (n = 9). 2-year postoperatively the mean HKA-angle was 182 ± 3.5 degrees in the varus knees and mean 177 ± 3 degrees in the valgus knees.
Complications
During the treatment period in external fixation, 16/58 patients had complications (Table 2). 1/6 patients with a loose pin at the time of removal of the fixator and pins, had clinical pin site infection grade 2 according to the Checketts-Otterburns classification [13] at some point during the treatment. One patient developed septic arthritis and did not achieve the desired correction and had revision surgery by total knee replacement 20 month after the HCO Two patients developed deep venous thrombosis and were successfully treated by Warfarin. One patient with delayed healing developed pseudoarthrosis and healed after additional surgery. One patient lost the achieved correction after extraction of the fixator and pins but required no additional surgery.
Table 2 Complications during the 2 years follow-up.
Complication Patients (n)
Pin site infection grade 2 [13] 5
Loose pin at removal of fixator/pins 6
Septic arthritis 1
Deep venous thrombosis 2
Loss of correction 1
Pseudoarthrosis 1
Work level
At the 2-year follow-up 40/52 (77%) patients were working. 10 of the patients (equal number of men and women) had decreased their time spent working compared to preoperatively (Table 1). Of these, 2 patients had received disability pension due to other reasons (fibromyalgia, n = 1 and one low back pain, n = 1) and one patient was on sick leave due to surgery in the contra lateral knee). Two patients had retired during the follow-up period.
Improvement during the 2 year-follow-up
At the 2-year follow-up, the patients reported a mean pain reduction of 40/100 points compared to baseline (p < 0.001). Parallel improvement was seen in function and quality of life as determined by the KOOS (p < 0.001) (Table 3). Since 6/58 patients were unavailable for 2 year-follow-up an ITT (Intension To Treat) analysis using the last obtained KOOS value was performed as a substitute for the missing 2 year follow-up value. The ITT analyses yield very similar results.
Table 3 Mean scores of the KOOS preoperatively to two years postoperatively
Treatment period Follow-ups
KOOS subscale (mean, SD) Preop (N = 58) 4 weeks (N = 49) 7 weeks (N = 47) 10 weeks (N = 47) One week after extraction (N = 56) 6 month (N = 51) one year (N = 51) two years (N = 52)
Pain 41 51 59 62 71 74 75 80
17 22 22 20 21 20 20 20
Symptom 50 55 61 63 68 74 75 80
18 19 19 19 19 20 20 20
Activities of daily life 48 49 57 62 71 75 79 80
19 20 19 16 20 21 20 19
Sport and recreation function 9 2 6 5 11 20 30 29
12 5 10 9 15 24 29 28
Knee related quality of life 21 24 33 32 43 52 56 61
14 19 21 18 22 25 27 25
KOOS = Knee injury and Osteoarthritis Outcome Score, 0–100 worst to best scale.
Improvement during the treatment with external fixation
The major improvements in pain and ADL function were obtained during the treatment with the external fixator, thereafter further but smaller improvements were seen until the 2-year follow-up as exemplified in Figure 3. Already at four weeks postoperatively, a significant pain reduction (p = 0.02) was seen (Table 3). There was also a significant reduction of the analgesic consumption at week 4, paracetamol (p < 0.001) and tramadol (p < 0.001) compared to the first postoperative week (Fig 4.). Reductions of the percentages of patients reporting both activity-related pain and pain at rest were seen. Preoperatively, 90% of the patients reported moderate to extreme pain during walking (activity-related pain, question 5 in KOOS subscale pain) compared to 70 % at week 4, 55% at week 7 and 36% at extraction of the external fixation and pins. Preoperatively, 55% of the patients reported moderate to extreme pain at night (pain at rest, question 7 in KOOS subscale pain) compared to 55% at week 4, 38% at week 7 and 8% at extraction of the external fixation and pins.
Figure 3 KOOS Pain mean scores (± 95%CI) over time (0–100, worst to best).
Figure 4 Average analgesic consumption during the treatment period (n = 50).
At 7 weeks a significant improvement compared to preoperatively was seen with regard to symptoms (p = 0.002), activities of daily living (p = 0.003) and knee-related quality of life (p < 0.0001) (Table 3).
Improvements after removal of the external fixation
Compared to one week after removal of the external fixation, significant improvements were seen in knee-related quality of life (p = 0.005) at the 6-month follow-up. At the one-year follow-up, improvement was seen in sport and recreation function (p = 0.0001) which assessed more demanding functions such as kneeling, squatting, jumping and running. At the 2-year follow-up improvement were seen in the subscales pain (p = 0.01), symptom (p = 0.02) and ADL (p = 0.02) (Table 3).
Predictors of poor improvement
Worse preoperative pain and complications were predictors of poor improvement in the KOOS subscale pain (Table 4). Age, BMI, sex, acceptable HKA-angle two years postoperatively and smoking were not predictors for poor improvement in pain. Performing the same analysis for poor improvement in the KOOS subscale knee-related QOL, there were no predictors of poor improvement.
Table 4 Linear regression preformed to determind predictors of poor changes preoperatively to the 2 years follow-up of the KOOS subscales pain.
Univeriate analysis Multivariate analysisII
Predictor Variable Na Cb (95% CI) pc R2 adjd Cb (95% CI) pc
Age (year) 52 0.27 (-0.62–1.17) 0.54 0% Not included
BMI (kg/m2) 48 0.62 (-0.89–2.12) 0.41 0% Not included
Preop pain (0–100) 52 '-0.61 (-0.92–-0.29) 0.0003 21% '-0.6 (-0.9–-0.2,8) 0.0004
Sex 52 1.42 (-10.77–13.6) 0.82 0% Not included
0: female 26
1: male 26
Acceptable 2 year HKA-angle 42 8.78 (-12.81–30.38) 0.41 0% Not included
0: no 7
1: yes 37
Smoker 52 '-0.35 (-15.27–14.56) 0.96 0% Not included
0: no 40
1: yes 12
Complication 51 '-15.8 (-16.6–7.94) 0.06 0.5% '-14.7 (-29.4–0.08) 0.05
0: no 43
1: yes 8
aNumber of patients included
bEstimated change of changes in preoperative until 2 years postoperative pain per unit change predictor variable; 95% confidence interval (CI) within paranthesis.
cP value for effect of predictor variable.
cFraction of changes of preopertive until 2 years postoperative pain variance explaind by each predictor (adjusted value).
Discussion
This study shows large improvements in self-rated pain, function and quality of life at 2 years for patients operated on for knee OA by tibial osteotomy using the hemicallotasis technique. Surprisingly, substantial improvements were seen already during the immediate postoperative period when the external fixation was still used.
To our knowledge, this follow-up study is the first evaluating the patients' perspective of the HCO including the treatment period. Clinical scores have been used by Magyar et al in a randomized study comparing close wedge osteotomy and HCO, and by Gerdhem et al when evaluating the HCO [2,6]. Both studies showed significant improvements and good to excellent results as evaluated by the Hospital for Special Surgery Score (HSS), at 2 years and at 12 to 28 months respectively. Our results are in line with these previous reports.
Magyar et al [2] found no differences in clinical scores between two methods of high tibial osteotomy and remarked that the clinical scores used, seemed to be too blunt to detect the differences in younger active patients. For this reason we choose an outcome measure validated for younger or physically active subjects with knee OA. The Knee injury and Osteoarthritis Outcome Score (KOOS) evaluates pain, other symptoms and activities of daily living but also includes sport and recreational activities and quality of life, dimensions that have been shown to be more sensitive in younger and/or physically active with knee OA than the more commonly used Western Ontario and MacMaster Universities Osteoarthritis Index (WOMAC) [14]. The patients filled in the questionnaires themselves in their homes. In studies using this administration mode and frequent follow-ups a high number of dropouts could be expected. In the present study the 2-year follow-up rate was 90%, which must be considered high.
Most of the improvements in all subscales of the KOOS, except for sport/recreational function, were obtained during the treatment period, when using the external fixation. The KOOS questionnaire was sensitive enough to detect significant changes over just a few weeks. Four weeks postoperatively, pain related to the correction was common, but the pain gradually diminished. Seven weeks postoperatively, the correction was completed and the external fixator was locked. When visiting the outpatient clinic, the patients told that they felt improvements almost day by day, as confirmed by the improvements in all KOOS subscales except for sports/recreation function. During the later part of the treatment, patients gained more knee stability and gradually decreased the use of crutches, but were still prevented from certain activities due to the external fixator. This clinical improvement was detected by the KOOS, especially the subscale ADL. One week after finishing the treatment with external fixator knee-related QOL was further improved, probably due to the extraction of the fixator and the pins. The similar decrease of number of patients with moderate to extreme pain during walking and at rest reported during the treatment in external fixation, indicate that the treatment by the HCO effected activity-induced pain as well as pain at rest.
Most probably the early pain reduction seen is due to the gradually corrected alignment of the leg. Alternative explanations to the early improvements seen include decrease of intraosseous pressure [15,16].
The gender distribution in our study was almost even reflecting that knee injuries, and thus post-traumatic OA due to knee injuries, are more common in men [17], whereas elderly women more commonly have OA. The patients with known knee injury in this study were mostly men (28%). They were, on average, 8 years younger than men without known knee injury, and had performed more high/medium joint-loading sport activities. This reflects that patients with joint injury have an increased risk of developing knee OA requiring surgery [17-19].
Patients having tibial osteotomy in the current study and patients having knee arthroplasty in a previous study report similar preoperative pain and function [20]. This is remarkable, taking into account the tibial osteotomy patient being on average 17 years younger. It should also be noted that patients developing OA at a younger age often have high demands of their knee function in both working life and leisure time.
Long-term results have been shown to depend on the achieved correction of the healed osteotomy [21-24]. 57/58 patients in our study achieved the intended correction. The mean HKA-angle at the 2-year follow-up was acceptable in 38/52 patients. These results are comparable or even better than studies with similar evaluation period [2,5,6].
As a predictor for poor improvements in pain over time, worse preoperative pain accounted for 21% of the variance. This is a factor that should be taken into consideration when selecting patients for high tibial osteotomy. This may indicate that operating earlier would give a better result as also discussed for total joint replacement [25].
Ten patients had complications and not unexpectedly, these complications were predictors of poor improvement in pain over time.
Conclusion
Our study showed that tibial osteotomy by the hemicallotasis technique yields large improvement in self-rated pain, function, and quality of life, which persists over two years. Surprisingly, most improvements were seen already during the immediate post-operative period when the external fixation was still used. This new knowledge should be incorporated in the information to the patients to help them make treatment choices regarding knee OA and in the pre-operative information of the hemicallotasis technique.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AWD, STL and ER designed the study. AWD collected the data, analyzed the data and drafted the manuscript. STL and ER revised the manuscript. All three authors read and approved of the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Financial support was obtained from the Swedish Rheumatism Association, the Swedish Research Council, the Swedish National Centre for Research in Sports, the Zoega Medical Foundation, and Lund University Hospital.
==== Refs
Magyar G Toksvig-Larsen S Lindsttrand A Open wedge tibial osteotomy by callus distraction in gonarthrosis. Operative technique and early results in 36 patients Acta Orthop Scand 1998 69 147 51 9602772
Magyar G Ahl TL Vibe P Toksvig-Larsen S Lindstrand A Open-wedge osteotomy by hemicallotasis or the closed-wedge technique for osteoarthritis of the knee. A randomised study of 50 operations J Bone Joint Surg Br 1999 81 444 8 10872363 10.1302/0301-620X.81B3.8925
Calista F Pegreffi P High tibial osteotomy: osteotomy in minus or hemicallotasis with EAF? Chir Organi Mov 1996 81 155 63 8968118
Magyar G Toksvig-Larsen S Lindstrand A Hemicallotasis open-wedge osteotomy for osteoarthritis of the knee. Complications in 308 operations J Bone Joint Surg Br 1999 81 449 51 10872364 10.1302/0301-620X.81B3.8926
Klinger HM Lorenz F Harer T Open wedge tibial osteotomy by hemicallotasis for medial compartment osteoarthritis Arch Orthop Trauma Surg 2001 121 245 7 11409551 10.1007/s004020000216
Gerdhem P Abdon P Odenbring S Hemicallotasis for medial gonarthrosis: a short-term follow-up of 21 patients Arch Orthop Trauma Surg 2002 122 134 8 11927993 10.1007/s004020100323
Magyar G Osteotomy for gonarthrosis – with special references to the tibial callus distraction technique Department of Orthopedics, University Hospital, Lund 1999 Lund university Lund 49
Roos EM Roos HP Lohmander LS Ekdahl C Beynnon BD Knee Injury and Osteoarthritis Outcome Score (KOOS) – development of a self-administered outcome measure J Orthop Sports Phys Ther 1998 28 88 96 9699158
Roos EM Roos HP Ekdahl C Lohmander LS Knee injury and Osteoarthritis Outcome Score (KOOS) – validation of a Swedish version Scand J Med Sci Sports 1998 8 439 48 9863983
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 40 3068365
Roos EM Lohmander LS The Knee injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis Health Qual Life Outcomes 2003 1 64 14613558 10.1186/1477-7525-1-64
Buckwalter JA Lane NE Athletics and osteoarthritis Am J Sports Med 1997 25 873 81 9397280
Checketts RG Otterburn M Mac Eachern AG Pin track infection; definition, incidence and prevention Supplement to International Journal of Orthopaedic Trauma 1993 3 16 18
Roos EM Roos HP Lohmander LS WOMAC Osteoarthritis Index – additional dimensions for use in subjects with post-traumatic osteoarthritis of the knee. Western Ontario and MacMaster Universities Osteoarthritis Cartilage 1999 7 216 21 10222220 10.1053/joca.1998.0153
Arnoldi CC Lemperg K Linderholm H Intraosseous hypertension and pain in the knee J Bone Joint Surg Br 1975 57 360 3 1158947
Simkin PA Bone pain and pressure in osteoarthritic joints Novartis Found Symp 2004 260 179 86 discussion 186-90, 277-9. 15283450
Roos H Adalberth T Dahlberg L Lohmander LS Osteoarthritis of the knee after injury to the anterior cruciate ligament or meniscus: the influence of time and age Osteoarthritis Cartilage 1995 3 261 7 8689461
Lohmander LS Roos H Knee ligament injury, surgery and osteoarthrosis. Truth or consequences? Acta Orthop Scand 1994 65 605 9 7839844
Roos H Lindberg H Gardsell P Lohmander LS Wingstrand H The prevalence of gonarthrosis and its relation to meniscectomy in former soccer players Am J Sports Med 1994 22 219 22 8198190
Roos EM Toksvig-Larsen S Knee injury and Osteoarthritis Outcome Score (KOOS) – validation and comparison to the WOMAC in total knee replacement Health Qual Life Outcomes 2003 1 17 12801417 10.1186/1477-7525-1-17
Coventry MB Ilstrup DM Wallrichs SL Proximal tibial osteotomy. A critical long-term study of eighty-seven cases J Bone Joint Surg Am 1993 75 196 201 8423180
Odenbring S Egund N Hagstedt B Larsson J Lindstrand A Toksvig-Larsen S Ten-year results of tibial osteotomy for medial gonarthrosis. The influence of overcorrection Arch Orthop Trauma Surg 1991 110 103 8 2015131 10.1007/BF00393883
Valenti J Calvo RR Lopez R Canadell J Long term evaluation of high tibial valgus osteotomy Int Orthop 1990 14 347 9 2076916 10.1007/BF00182642
Shoji H Insall J High tibial osteotomy for osteoarthritis of the knee with valgus deformity J Bone Joint Surg Am 1973 55 963 73 4760103
Nilsdotter AK Petersson IF Roos EM Lohmander LS Predictors of patient relevant outcome after total hip replacement for osteoarthritis: a prospective study Ann Rheum Dis 2003 62 923 30 12972468 10.1136/ard.62.10.923
| 15811186 | PMC1090585 | CC BY | 2021-01-04 16:32:03 | no | BMC Musculoskelet Disord. 2005 Apr 5; 6:18 | utf-8 | BMC Musculoskelet Disord | 2,005 | 10.1186/1471-2474-6-18 | oa_comm |
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-201585049710.1186/1471-2474-6-20Study ProtocolThe effect of glucosamine sulphate on osteoarthritis: design of a long-term randomised clinical trial [ISRCTN54513166] Rozendaal Rianne M [email protected] Bart W [email protected] Harrie [email protected] Elian J [email protected] Osch Gerjo JVM [email protected] Abida Z [email protected] Jan AN [email protected] Sita MA [email protected] Department of General Practice, Erasmus MC, University Medical Centre Rotterdam, the Netherlands2 Department of Orthopaedics, Erasmus MC, University Medical Centre Rotterdam, the Netherlands3 Department of Radiology, Erasmus MC, University Medical Centre Rotterdam, the Netherlands2005 26 4 2005 6 20 20 17 3 2005 26 4 2005 Copyright © 2005 Rozendaal et al; licensee BioMed Central Ltd.2005Rozendaal 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
Pharmacological treatment for osteoarthritis (OA) can be divided into two groups: symptom-modifying drugs and disease-modifying drugs. Symptom-modifying drugs are currently the prescription of choice for patients with OA, as disease-modifying drugs are not yet available in usual care. However, there has recently been a lot of debate about glucosamine sulphate (GS), a biological agent that is thought to have both symptom-modifying and disease-modifying properties. This assumption has yet to be proved.
The objective of this article is to present the design of a blind randomised clinical trial that examines the long-term symptom-modifying and disease-modifying effectiveness of GS in patients with hip OA. This trial is ongoing and will finish in March 2006.
Methods/design
Patients with hip OA meeting the ACR-criteria are randomly allocated to either 1500 mg of oral GS or placebo for the duration of two years. The primary outcome measures, which are joint space narrowing (JSN), and change in the pain and function score of the Western Ontario McMaster Universities Osteoarthritis index (WOMAC), are determined at baseline and after two years of follow-up during the final assessment. Intermediate measures at three-month intervals throughout the trial are used to study secondary outcome measures. Secondary outcome measures are changes in WOMAC stiffness score, quality of life, medical consumption, side effects and differences in biomarker CTX-II.
==== Body
Background
Pharmacological treatment for osteoarthritis (OA) can be divided into two groups: symptom-modifying drugs and disease-modifying drugs [1]. Symptom-modifying drugs are at present the prescription of choice for patients with OA. Drugs in this group are: simple analgesics (such as acetaminophen) and non-steroidal anti-inflammatory drugs (NSAIDs). Both acetaminophen and NSAIDs are effective in relieving symptoms of OA. In more severe stages of the disease NSAIDs are more effective, however, they are also the cause of serious side effects [2].
Disease-modifying drugs (i.e. drugs which alter disease progression) are not yet available in usual care. Although there has recently been a lot of debate about some biological agents that are thought to have both symptom-modifying and disease-modifying properties, results from previous trials have not been convincing. Of these biological agents, glucosamine sulphate seems to be the most promising.
Glucosamine sulphate (GS) has been shown to be an effective symptom-modifying agent, with effect sizes ranging from moderate to high [3,4]. In four trials that compared GS and NSAIDs, GS was found to be as effective as, or slightly more effective than NSAIDs [4]. Together with the fact that no serious adverse events have been reported concerning GS [3,4] this implies that GS may be a good alternative to NSAIDs. However, due to publication bias and due to quality issues in the trials studying GS, it may well be that reported effect sizes are exaggerated. A more recent trial studying the effect of GS did not find a difference between GS and placebo [5]. Also, several other uncertainties exist concerning the symptom-modifying properties of GS. For example, most trials were only of short-term duration (e.g. mean 6.25 weeks [4]) and it is therefore not possible to draw conclusions about the long-term efficacy. Another problem is that the mechanism behind the improvement of symptoms due to GS is not known. If GS directly influences the remaining cartilage, it would seem plausible that the symptomatic effect is greater in people with mild to moderate OA than in people with more severe OA, because there is more cartilage remaining in the first group. However, this possible difference in effect between different stages of OA has not been tested yet in a randomised clinical trial (RCT). These uncertainties make further study into the magnitude of long-term symptom-modifying effects in different stages of the disease justified.
Concerning disease-modifying effects, two recent long-term (three years) trials in patients with knee OA did report some evidence that GS affected the progression of OA [6,7]. Progressive joint space narrowing in the narrowest medial compartment of the tibiofemoral joint was used to define progression of knee OA (as recommended by a task force of the OA research society [8]). Whereas joint space narrowing (JSN), had significantly progressed in the placebo groups, it had not in the groups that were taking GS. This implies that daily intake of GS acted against progression of OA. However, these results are controversial, because both trials lacked appropriate and standardised protocols for taking radiographs. Although it is not likely that this influenced the results much [9], it is necessary to reproduce them in a study with well-standardised protocols. These two long-term trials both looked at the effect of GS on knee OA, no trial has been or is being performed yet that looks at its effect on hip OA.
Based on the above, we designed a long-term trial to answer our main question: Does glucosamine sulphate favourably modify progression of osteoarthritis? Because there still is uncertainty about the symptom-modifying properties of GS we will also try to answer three secondary questions: Does GS have the same effects in all stages of OA? What is the long-term cost-effectiveness of addition of GS to usual care? And, does GS prevent the onset or progression of OA in the contralateral hip joint? Additional to these clinical questions, the data will be used to look at changes on cell-level caused by GS to learn more about its possible mechanisms of action. All results derived from this trial will be published using our International Standardised Randomised Controlled Trial Number (ISRCTN).
In this article we will present the detailed protocol of the trial. This trial is ongoing; at the moment all patients are included and have passed the first 9 months of the follow-up period.
Methods/design
Study design
This study is a randomised, blinded, placebo-controlled trial. All actors in this trial, who may cause bias, are blinded to treatment allocation: the patient, who is the assessor of the symptomatic outcomes, the researcher, who is the assessor of the objective outcomes, and the caregiver. The analyses will also be performed blind. The study design was approved by the Medical Ethics Committee at the Erasmus MC – university medical centre Rotterdam. All patients gave written informed consent.
Patient selection
General practitioners in the Rotterdam area agreed to search their electronic medical record for patients diagnosed with hip OA and for patients with symptoms associated with hip OA (i.e. persistent hip pain in combination with NSAID use). These patients are contacted by their general practitioner and informed about the trial. For more information, patients can forward their contact details to the researchers. These patients then receive an extensive information folder containing all the information needed to make an informed decision about participation in the study. This folder has been reviewed and approved by the medical ethics committee.
The information folder also contains an informed consent form, which patients need to fill out if they want to participate in the study. Patients who give written informed consent are contacted by phone for a preliminary check of the inclusion and exclusion criteria. People meeting these criteria are invited to the research centre of Erasmus MC for a baseline-measurement, during which the criteria can be checked more precisely.
In- and exclusion criteria
Patients are eligible for inclusion when they meet one of the ACR criteria for hip OA [10]. Patients that have already undergone hip replacement surgery or those on the waiting list for joint replacement are not included in the study. Neither eligible patients with a Kellgren & Lawrence (K-L) score of 4 [11], nor people with renal and/or hepatic disease, diabetes mellitus or a disabling co-morbidity are included. Finally, patients unable to understand Dutch questionnaires are excluded from participation.
Sample size
The sample size was calculated primarily to detect clinically relevant differences in radiological progression of the affected joints between the two groups (treatment and placebo) after two years of follow up. To detect a difference of 0.25 mm in radiological progression (SD 0.5) between the intervention and placebo groups (power 80%, alpha 5%, one-tailed testing) after two years of follow-up, 63 patients with hip osteoarthritis are needed per group. These calculations are based on an average change of 0.33 mm in joint space (SD 0.5) during one year of follow-up of patients with hip osteoarthritis [12].
Fewer patients are needed to detect relevant clinical differences: to detect a difference of 25% in pain (Western Ontario McMaster Universities Osteoarthritis index (WOMAC)) with one-tailed testing, a power of 80%, and alpha 5% (Mean 4.83, SD 2.25 [13]) 55 patients per group are needed. To detect the same difference in function (WOMAC) (mean 4.81, SD 2.18 [13]) 51 patients are needed per group.
As we expect a 20% loss to follow-up, we need to include 150 patients. However, to create options for studying effect-modification by type and severity of osteoarthritis, we oversized this trial to 220 patients (110 in each group).
Intervention
Patients who participate in the trial are randomised to either GS or a placebo for the duration of two years. To ensure a daily intake of 1500 mg GS, they are required to take two pills each day. The GS and placebo pills are identical in taste and appearance and were delivered in identical plastic bottles. This will ensure true blinding of the patients and of the researchers. Blinding of the patients will be tested after two years; if people can guess what sort of pills they were taking, this might have an influence on the subjective measures. This will therefore be taken into account in the analysis of the data.
The Department of Nutritional Sciences at Numico Research BV manufactured the pills used in this trial.
Randomisation
Following informed consent and baseline assessments, patients are allocated to the intervention or control group using a blinded randomisation list. The randomisation list contains four different strata and is randomised per block of six numbers. This list was generated with a computer by an independent researcher. This researcher also handled labelling the pill-bottles with the randomisation numbers. The researchers involved in this project received all bottles after they were labelled, ensuring blinding to treatment allocation. The randomisation list with the key to treatment allocation will be kept in a safe until the end of the trial. To be able to perform the analyses blinded, the allocation to treatment A and treatment B will be provided, but not the key to A and B.
People are assigned to one of the four different strata on the basis of the Kellgren-Lawrence score of the hips, knees and hands. A researcher (RMR) will score all the radiographs according to the Kellgren-Lawrence score. The outline of the four different strata is given in table 1. Once the correct stratum is established at baseline, the patient is given the subsequent unique four-digit randomisation number from his/her stratum on the randomisation list. This number is used for labelling study materials and data. By stratifying, patients are optimally distributed to GS and placebo in the different strata, which makes comparing people with mild OA to people with moderate-severe OA, and comparing people with local OA to people with generalised OA possible. In this way, we will be able to study whether effect of treatment depends on severity or localisation of OA.
Table 1 Outline of the randomisation strata
Hip radiograph Knee and hand radiographs Type
Group 1 K-L score < 2 K-L score < 2 for hands and knees mild + localised
Group 2 K-L score < 2 K-L score ≥ 2 for hands and/or knees mild + generalised
Group 3 K-L score ≥ 2 K-L score < 2 for hands and knees moderate/severe + localised
Group 4 K-L score ≥ 2 K-L score ≥ 2 for hands and/or knees moderate/severe + generalised
Measurements
Data for the primary and secondary outcome measures are being collected at different time-points throughout the trial. An overview of the timing of the measurements and the outline of the primary and secondary outcome measures is given in table 2.
Table 2 Timing of measurements and outline of primary and secondary outcome measures
0 m
B.A. 3 m
Q 6 m
Visit 9 m
Q 12 m
Visit 15 m
Q 18 m
Visit 21 m
Q 24 m
F.A.
Primary outcome measures
Joint space width x x
Pain score (WOMAC) x x
Function score (WOMAC) x x
Secondary outcome measures
Subchondral bone quality x x
Stiffness score (WOMAC) x x x x x x x x x
Quality of life (EuroQol EQ-5D) x x x x x x x x x
Medical consumption x x x x x x x x x
Side effects x x x x x x x x
CTX-II x x x x x
Possible confounders/Effect modifiers
Type of OA (localised – generalised) x
Radiological severity x x
Joint function x x
Age x
Gender x
Activity level x x x x x x x x x
Co-interventions x x x x x x x x x
Compliance (BMQ) x x x x x x x x
Compliance (pill count) x x x x
Note: 0 m: 0 month of follow up, 3 m: 3 months of follow up etc. B.A.: baseline assessment. Visit: 6 monthly visit. F.A.: final assessment
In brief, the trial starts for every patient with a baseline assessment at the research centre. At the end of this assessment, patients receive a supply of GS or placebo sufficient for seven months. After the baseline assessment, patients will receive a questionnaire every three months, which has to be returned to the researchers, except from those at 6, 12 and 18 months after baseline, which will be collected by the researchers during a home visit. After two years, patients return to the research centre for the final assessment, which marks the end of the trial. The collection of the outcome measures is described in the following sections.
Radiographs
Radiographs are taken during the baseline assessment and during the final assessment two years later.
At baseline, three anteroposterior (AP) radiographs are taken, one of the pelvis, one of both knees, and one of both hands. All radiographs are used to establish what stratum the subject belongs to (table 1). The radiographs of hands and knees will not be used to determine outcome measures and will therefore not be repeated at follow up. As follows from table 1, people with knee and/or hand OA are stratified to one of the 'generalised OA' groups (2 or 4).
A highly standardised protocol is used to make the weight-bearing, AP pelvic radiographs at baseline and follow-up, allowing for a correct measurement of our primary outcome variable: joint space narrowing. The patients' feet are positioned alongside a frame, which was designed to ensure 15° internal rotation of the hips. A second frame (no internal rotation) is available for patients with severe mobility restrictions of the hips. The frame used during the baseline radiograph of a patient will also be used two years later for his/her follow up radiograph. Patients are asked to stand upright. If present, flexion in hips or knees is recorded. Protocol for the pelvic X-ray further states that focus-to-film distance should be 130 cm and that the X-ray beam should be centred on the superior aspect of the pubic symphysis. The X-rays are digitised.
The X-rays from baseline and final assessment will be analysed side by side. The minimal joint space width (JSW) will be identified from the baseline X-ray by assessing four different points: medial, axial, superior and lateral [14]). The researcher will also identify a point that appears to be the minimal JSW. From these five points the actual minimal JSW will be determined. This point will be used to measure changes in joint space width over the two-year follow up period.
Dexa-scan
During the baseline and final assessment, a Dual Energy X-ray Absorptiometry (DEXA) scan will be used to make scans of the pelvis. A frame similar to the one used for the radiograph of the hips is also used to make the DEXA-scan, ensuring 15° internal rotation of the hips. The scan will be used to study quantitative changes in subchondral bone density both of the affected joint and of the contra-lateral joint. Subchondral bone-density alterations might indicate osteoarthritic progression. This long-term trial can be helpful to determine whether pre-clinical OA can be recognised from a DEXA-scan. And, if so, whether GS prevents the onset or progression of OA in the pre-clinical stage.
Physical examination
A physical examination is carried out at baseline and is repeated during the final assessment. At baseline, this test is first of all used to check part of the inclusion criteria. Various tests are also carried out to check for co-existing musculoskeletal disorders. Findings from the physical examination will be used as baseline characteristics, and to register clinical signs and joint function after two years of follow up. Joint function is established by assessing pain due to joint motion, and by measuring limitation of joint motion with a two-arm goniometer.
Questionnaires
Throughout the study, patients will fill out a total of nine questionnaires. The first during the baseline assessment, followed by a questionnaire every three months in the following two years (including the last one during the final assessment).
The baseline questionnaire is used to measure different patient characteristics (age, gender, race, social status, Body Mass Index (BMI)), disease related characteristics (localisation of symptoms, duration of symptoms, family history) and co-morbidities. Of these characteristics, BMI and co-morbidities will be monitored throughout the trial.
Three validated instruments are used in all nine questionnaires: the WOMAC questionnaire will be used to establish severity of clinical status. It contains subscales for pain, stiffness and function. The WOMAC questionnaire is extensively validated and recommended for clinical assessment in osteoarthritis trials by the WHO [15]. The EuroQol (EQ-5D) will be used to measure quality of life, because of the usefulness of this scale in cost-effectiveness analysis [16,17]. The cost-effectiveness analysis will also be based on employment status, sick leave, changes in work-tasks or other work-related adjustments, and on medical consumption. The SQUASH questionnaire is used to measure load level in work and sports [18].
In the eight follow-up questionnaires, patients will be asked to answer questions about alterations in their symptoms (i.e. whether they improved or deteriorated), which will be measured with a 7-point Likert scale. Also, compliance to treatment is measured with the Brief Medication Questionnaire (BMQ) [19].
Laboratory assessments
At baseline, two samples of blood are collected. The first to measure the erythrocyte sedimentation rate (ESR), which is used for the inclusion criteria (ACR-criteria). The second sample is stored at -20°C to create options for future DNA-research, for which patients gave separate written informed consent.
Throughout the study, we will collect samples of second-morning void urine of all patients. In urine a marker of cartilage degradation can be found, called CTX-II. In the Rotterdam study [20] this marker was found to be predictive of radiological progression of hip and knee OA. It may therefore be used to assess the effect of treatment on the progression of OA. Urine samples will be collected at baseline and once every six months during follow up. At the end of the study a total of five samples will be available from every patient. These urine samples are stored at -80°C. If promising new markers are discovered during the course of the study, these can also be included in the analysis.
Half-yearly visits
Every six months one of the researchers will visit the patients at home. The main reason for this visit is to provide the patient with new pills (sufficient for seven months). To be able to calculate compliance to treatment, the pills remaining of the previous supply will be collected. The amount of remaining pills combined with the score on the compliance questionnaire (BMQ) will give a good indication of the actual amount of pills the patient has been taking. Finally, a sample of second-morning void urine on the day of the visit will be collected.
Analyses
The researchers will be aware of allocation to treatment A or B at the time of the statistical analyses, but will not know which group received GS and which group received the placebo. All analyses will take place after the trial has finished, no intermediate analyses will be performed.
Success of randomisation and normality of outcome measures will be checked before actual analyses are done. Differences in the primary outcome measures JSN and WOMAC (pain and function) between the intervention and placebo group will be analysed on the basis of the 'intention to treat' principle using linear regression models. Additionally a per-protocol analysis will be done. When it turns out that randomisation was (partially) unsuccessful, we will adjust for differences in prognosis. Using baseline characteristics, we can identify factors that influence outcome of the study. Factors that change the outcome with 10% will be regarded as confounders and will therefore be added to the regression-model.
A cost-effectiveness analysis will be performed from a social and a patient perspective, looking at differences in direct and indirect health care cost between the two groups (GS and placebo). If the trial does not show a difference in disease parameters (WOMAC) and quality of life (EuroQol) between the GS and the placebo group, the analysis will be reduced to a cost minimisation analysis. This form of analysis evaluates the efficacy of treatment based solely on direct and indirect costs. If the study does find a positive difference in disease parameter and/or quality of life a cost-effectiveness ratio can be determined with on the one hand the costs and savings and on the other hand the disease-specific parameters and also quality of life.
Current status
A total of 40 GP's were found willing to participate in the study. They sent a total of 600 letters to inform possible eligible patients of the study. We received 417 requests for additional information and thus sent an equal amount of information folders. Of these 417 people 250 returned a written informed consent. Eventually 222 people entered the study. Of the 28 people that did not enter the study, most did not meet the inclusion criteria and a few people changed their mind and withdrew their informed consent before randomisation.
We started including patients at the end of September 2003 and the last patient was included on March 15th of 2004. This means the study will run until March 2006. The first results will be available around September 2006.
Competing interests
The Department of Nutritional Sciences of Numico Research BV provided the pills used in this study free of charge. Numico Research BV was not involved in the design of this trial nor will they be involved in any other aspect of the trial.
Authors' contributions
SMABZ and HW conceived of the study and developed the design of this randomised clinical trial. SMABZ participated in writing the article, HW contributed to its content. BWK, JANV and GJVMvO contributed to the design of the study and the content of the article. EJU and AZG contributed to the content of the article. RMR is conducting the research, participated in the completion of the study design and wrote the article.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Lequesne M Brandt K Bellamy N Moskowitz R Menkes CJ Pelletier JP Altman R Guidelines for testing slow acting drugs in osteoarthritis J Rheumatol Suppl 1994 41 65 71; discussion 72-3 7799389
Towheed TE Judd MJ Hochberg MC Wells G Acetaminophen for osteoarthritis Cochrane Database Syst Rev 2003 CD004257 12804508
McAlindon TE LaValley MP Gulin JP Felson DT Glucosamine and chondroitin for treatment of osteoarthritis: a systematic quality assessment and meta-analysis Jama 2000 283 1469 1475 10732937 10.1001/jama.283.11.1469
Towheed TE Anastassiades TP Shea B Houpt J Welch V Hochberg MC Glucosamine therapy for treating osteoarthritis Cochrane Database Syst Rev 2001 CD002946 11279782
McAlindon T Formica M LaValley M Lehmer M Kabbara K Effectiveness of glucosamine for symptoms of knee osteoarthritis: results from an internet-based randomized double-blind controlled trial Am J Med 2004 117 643 649 15501201 10.1016/j.amjmed.2004.06.023
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
Pavelka K Gatterova J Olejarova M Machacek S Giacovelli G Rovati LC Glucosamine sulfate use and delay of progression of knee osteoarthritis: a 3-year, randomized, placebo-controlled, double-blind study Arch Intern Med 2002 162 2113 2123 12374520 10.1001/archinte.162.18.2113
Hochberg MC Altman RD Brandt KD Moskowitz RW Design and conduct of clinical trials in osteoarthritis: preliminary recommendations from a task force of the Osteoarthritis Research Society J Rheumatol 1997 24 792 794 9101520
Pavelka K Bruyere O Rovati LC Olejarova M Giacovelli G Reginster JY Relief in mild-to-moderate pain is not a confounder in joint space narrowing assessment of full extension knee radiographs in recent osteoarthritis structure-modifying drug trials Osteoarthritis Cartilage 2003 11 730 737 13129692 10.1016/S1063-4584(03)00166-3
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
Kellgren JH Lawrence JS Radiological assessment of osteo-arthrosis Ann Rheum Dis 1957 16 494 502 13498604
Dougados M Gueguen A Nguyen M Berdah L Lequesne M Mazieres B Vignon E Radiological progression of hip osteoarthritis: definition, risk factors and correlations with clinical status Ann Rheum Dis 1996 55 356 362 8694574
Angst F Aeschlimann A Steiner W Stucki G Responsiveness of the WOMAC osteoarthritis index as compared with the SF-36 in patients with osteoarthritis of the legs undergoing a comprehensive rehabilitation intervention Ann Rheum Dis 2001 60 834 840 11502609
Reijman M Hazes JM Pols HA Bernsen RM Koes BW Bierma-Zeinstra SM Validity and reliability of three definitions of hip osteoarthritis: cross sectional and longitudinal approach Ann Rheum Dis 2004 63 1427 1433 15479891 10.1136/ard.2003.016477
Dieppe PA Recommended methodology for assessing the progression of osteoarthritis of the hip and knee joints Osteoarthritis Cartilage 1995 3 73 77 7584320
van Agt HM Essink-Bot ML Krabbe PF Bonsel GJ Test-retest reliability of health state valuations collected with the EuroQol questionnaire Soc Sci Med 1994 39 1537 1544 7817218 10.1016/0277-9536(94)90005-1
Dolan P Modeling valuations for EuroQol health states Med Care 1997 35 1095 1108 9366889 10.1097/00005650-199711000-00002
Wendel-Vos GC Schuit AJ Saris WH Kromhout D Reproducibility and relative validity of the short questionnaire to assess health-enhancing physical activity J Clin Epidemiol 2003 56 1163 1169 14680666 10.1016/S0895-4356(03)00220-8
Svarstad BL Chewning BA Sleath BL Claesson C The Brief Medication Questionnaire: a tool for screening patient adherence and barriers to adherence Patient Educ Couns 1999 37 113 124 14528539 10.1016/S0738-3991(98)00107-4
Reijman M Hazes JM Bierma-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
| 15850497 | PMC1090586 | CC BY | 2021-01-04 16:32:03 | no | BMC Musculoskelet Disord. 2005 Apr 26; 6:20 | utf-8 | BMC Musculoskelet Disord | 2,005 | 10.1186/1471-2474-6-20 | oa_comm |
==== Front
BMC NeurolBMC Neurology1471-2377BioMed Central London 1471-2377-5-81582900510.1186/1471-2377-5-8Research ArticleValidation of multi-stage telephone-based identification of cognitive impairment and dementia Crooks Valerie C [email protected] Linda [email protected] Diana B [email protected] Helena [email protected] Vicki [email protected] Department of Research & Evaluation, Kaiser Permanente Southern California, 100 S. Los Robles, Pasadena, CA91101, USA2 Alzheimer's Disease Research Center, University of Southern California, 3715 McClintock Ave., Los Angeles, CA90089-0191, USA2005 13 4 2005 5 8 8 28 12 2004 13 4 2005 Copyright © 2005 Crooks et al; licensee BioMed Central Ltd.2005Crooks 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
Many types of research on dementia and cognitive impairment require large sample sizes. Detailed in-person assessment using batteries of neuropyschologic testing is expensive. This study evaluates whether a brief telephone cognitive assessment strategy can reliably classify cognitive status when compared to an in-person "gold-standard" clinical assessment.
Methods
The gold standard assessment of cognitive status was conducted at the University of Southern California Alzheimer Disease Research Center (USC ADRC). It involved an examination of patients with a memory complaint by a neurologist or psychiatrist specializing in cognitive disorders and administration of a battery of neuropsychologic tests. The method being evaluated was a multi-staged assessment using the Telephone Interview of Cognitive Status-modified (TICSm) with patients and the Telephone Dementia Questionnaire (TDQ) with a proxy. Elderly male and female patients who had received the gold standard in-person assessment were asked to also undergo the telephone assessment. The unweighted kappa statistic was calculated to compare the gold standard and the multistage telephone assessment methods. Sensitivity for classification with dementia and specificity for classification as normal were also calculated.
Results
Of 50 patients who underwent the gold standard assessment and were referred for telephone assessment, 38 (76%) completed the TICS. The mean age was 78.1 years and 26 (68%) were female. When comparing the gold standard assessment and the telephone method for classifying subjects as having dementia or no dementia, the sensitivity of the telephone method was 0.83 (95% confidence interval 0.36, 1.00), the specificity was 1.00 (95% confidence interval 0.89,1.00). Kappa was 0.89 (95% confidence interval 0.69, 1.000). Considering a gold-standard assessment of age-associated memory impairment as cognitive impairment, the sensitivity of the telephone approach is 0.38 (95% confidence interval 0.09, 0.76) specificity 0.96 (CI 0.45, 0.89) and kappa 0.61 (CI 0.37, 0.85).
Conclusion
Use of a telephone interview to identify people with dementia or cognitive impairment is a promising and relatively inexpensive strategy for identifying potential participants in intervention and clinical research studies and for classifying subjects in epidemiologic studies.
==== Body
Background
Epidemiologic studies of dementia generally require a large number of subjects. Longitudinal study designs that start with subjects who are cognitively intact and follow them to ascertain the development of dementia are often needed to reduce bias. Although physician and neuropsychological evaluations are considered the gold standard for the diagnosis of dementia, the high cost of these methods makes them infeasible for population-based epidemiological surveys. Administration of mental status screening tests and functional questionnaires by telephone offers a less costly alternative. These methods minimize participant burden, permit standardization and use less costly personnel. However, few studies have examined the reliability and validity of telephone-based methods in correctly identifying cognitive impairment and dementia [1].
This study evaluates whether a brief multi-stage telephone-based cognitive test and functional questionnaire strategy can reliably classify cognitive status when compared to a gold standard in-person clinical assessment. In this study, the gold standard was a classification derived by in-person neurological and neuropsychological evaluation at a university-based Alzheimer Disease Research Center. The staged telephone assessment of cognitive status has been used in the Kaiser Permanente Women's Memory Study, a large ongoing epidemiologic study of dementia in a sample of 3,681 women 75 and older. The same staged model was used in the veteran twin study by Gallo and Breitner [2] and a similar staged model for assessment of dementia and cognitive impairment was used in the Cache County study [3].
Methods
Overview
The study compared the classification of elderly patients into 3 categories (dementia, cognitive impairment or cognitively unimpaired) using two different assessment approaches. The gold standard reference classification was performed at the University of Southern California Alzheimer Disease Research Center (USC ADRC) and involved clinical examination of patients by a neurologist and a team experienced in the diagnosis of dementia and cognitive impairment. The second approach relied on information from telephone interviews, the Telephone Interview of Cognitive Status-modified (TICSm) [4,5], and a computer-assisted telephone version of the Dementia Questionnaire [6]. In this study we refer to this interview as the Telephone Dementia Questionnaire (TDQ). The comparison study was reviewed and approved by the Kaiser Permanente Southern California and University of Southern California Institutional Review Boards (IRBs).
Subjects and procedures
The subjects in this study were enrolled in a longitudinal study of normal aging, cognitive impairment, and Alzheimer-type dementia at the University of Southern California (USC) Alzheimer Disease Research Center (ADRC). Initial evaluations involve clinical examination of patients by a Board-certified neurologist and a licensed neuropsychologist experienced in the assessment and diagnosis of dementia. The neuropsychological battery was comprised of 20 tests of memory, language, executive control, and construction. In addition, scales of mood and psychiatric symptoms and screenings for visual or auditory impairment were administered to the subject. A rating of functional ability was provided by a collateral informant. See Appendix I for a list of the tests given.
Patients evaluated at the USC ADRC are classified as possible or probable AD, mild cognitive impairment (MCI), age-associated memory impairment (AAMI), or normal [7]. One patient in this study had a diagnosis of mixed dementia because she had AD and a history of stroke. Table 1 shows the criteria used to classify patients in each of these four categories.
Table 1 Diagnostic criteria for Alzheimer disease and cognitive impairment
Probable Alzheimer disease By criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) [6] Dementia by clinical examination, and documented with the Mini-Mental Status Exam (MMSE) and the Short-Blessed functional impairment questionnaire, with progressive deficits in 2 or more cognitive domains.
• Absence of other disorders that could account for the progressive dementia symptoms
Possible Alzheimer disease By NINCDS-ADRDA criteria :
• Dementia, as defined above;
• Presence of
1) variations in the onset, presentation or clinical course;
- or -
2) other disorder(s) sufficient to produce dementia, but not thought to be the cause of the dementia in this case
Mild Cognitive Impairment • Subjective memory complaint
• Normal general cognitive function by MMSE
• Standard score of -1.5, below the same-age norm, on the long delay free recall item of the California Verbal Learning Test (CVLT)
Age-Associated Memory Impairment • Has both Subjective memory complaint and normal general cognitive function by MMSE
• And either:
1) Standard score of -1.0, below the norm for 20–24 year- olds, on the long delay free recall item of the CVLT
- or -
2) Clinical Dementia Rating score of 0.5 - or -
3) Significant decline from previous performance on cognitive test(s) while still scoring in normal range
Follow-up evaluation is comprised of an interval history, depression, behavior and functional questionnaire and neuropsychological testing. Follow-up evaluation is conducted annually for subjects who are classified as cognitively impaired, demented or cognitively intact but over age 80 years and is repeated every 2 years for normal controls less than 80 years old. The results of the follow-up evaluation and neuropsychological testing are reviewed by a team of neurologists and psychologists and the diagnostic classification is updated accordingly. For this study, the most recent classification was used.
Ninety-eight subjects (representing normal, cognitively impaired and dementia cases) at the USC ADRC during the past 2 years were invited to participate in this study. They were contacted by letter by the USC ADRC (with a return opt-in postcard) to obtain initial permission to contact for a possible telephone interview. Postcard or verbal permission to contact was obtained in 50 cases.
Kaiser Permanente investigators mailed an informed consent form to the subject or proxy. Shortly thereafter, a telephone contact was made. Participants read and discussed the consent with the research assistant and were instructed to sign and return the written consent form by mail. If consent was obtained, telephone contact information for these participants (and their consenting proxies for patients with dementia) and appointment times were provided to the survey research firm that conducts the telephone cognitive assessments.
Interview responses were reviewed using the procedures described below. Both the telephone interviewers and the experts making the classification of cognitive status based on the TDQ were blinded to the results of the USC ADRC assessment until all subjects in the study had been classified.
Women's Memory Study assessment of dementia and cognitive status
The first step in the Women's Memory Study assessment of dementia and cognitive status is administration of the 23-question TICSm, which takes from 8 to 15 minutes to administer. The TICSm is strongly correlated with the Mini-Mental State Exam [8,5,10], a more widely used brief cognitive screen. The TICSm has been previously validated [4] and the psychometric properties of the computer-assisted TICSm used in the Women's Memory Study has also been evaluated [11]. Prior studies have shown that the TICSm has a high sensitivity in the detection of dementia [2,12] but a low positive predictive value [4]. Similar to an earlier study [2] TICsm cut-off scores are used to classify individuals as having no cognitive impairment (TICS score > 27) or possible cognitive impairment (TICSm score ≤ 27).
For those classified as possibly impaired based on the TICSm score, an attempt is made to do a second stage cognitive assessment in which the Telephone Dementia Questionnaire (TDQ), is administered to a proxy. The TDQ is a previously validated instrument [13,14] that takes about 19 minutes to complete and asks the proxy up to 48 questions about the subject's cognitive function in several domains (memory, fluency, comprehension, orientation). The TDQ alone, when compared to antemortem clinical exams, was found to have 92.8% sensitivity for dementia and a specificity of 89.5% [15]. In previous work by Gallo and Breitner [2] when used with the TICSm, the specificity of cognitive classification was .99. Two trained assessors independently review the TDQ responses and use this information and pre-defined criteria to classify individuals in one of three categories: 1) definite dementia; 2) no or minimal cognitive impairment; or 3) cognitive impairment without definite evidence of dementia.
TDQ classification in the dementia category requires memory deficits and multiple impairments in another cognitive domain and at least one deficit in an additional cognitive domain. Classification in the category of "no or minimal impairment" requires no more than two "yes" responses to questions about problems with memory. Classification as cognitive impairment without definite evidence of dementia requires a scatter of deficits that were not sufficient for a dementia classification but were too many to be classified as no dementia. Overall, classification of dementia based on the two-stage telephone classification is "conservative" in that it requires multiple impairments in memory plus 2 additional cognitive domains and functional impairment.
The reviewers then make a consensus TDQ classification in the same three categories after discussion of the independent assessments. When there is agreement among the reviewers, no further review is done. When the reviewers disagree, a third trained reviewer, a neuropsychologist or a neurologist, assesses the TDQ responses and makes a final classification. The test reliability of the consensus TDQ has been measured on an on-going basis by having the reviewers reassess TDQs selected at random and blinded to the initial assessment. The kappa coefficient was 0.85 for the consensus assessments.
Analysis
The analysis compared the USC ADRC final diagnosis of cognitive status and the Women's Memory Study classification based on telephone surveys. The unweighted kappa statistic was calculated comparing the ADRC and the staged telephone classification. As this was the principal aim of the Women's Memory Study method, the first comparison included dementia and no dementia classifications only.
Other comparisons were based on two alternate assumptions about patients who had been classified by the USC ADRC as having Age-Associated Memory Impairment (AAMI). The first assumption was that these patients were normal and the second was that they had cognitive impairment. Sensitivity for classification with dementia and specificity for classification as normal were also calculated under both sets of assumptions. Ninety-five percent confidence intervals for kappa were estimated as described in Fleiss [16]. Confidence intervals for sensitivity and specificity are exact intervals calculated using StatXact 5 software [17].
Results
Fifty patients who completed the USC ADRC evaluations initially consented to participate in this study. We were able to complete the TICSm assessments on 38 (76%) of these patients. The mean age of participants was 78.1 years (SD ± 8.0) and 26 (68%) were female. Eight of the 38 patients scored below 28 on the TICSm. We then completed TDQ interviews for 8 of the 8 proxies (100%).
Table 2 shows the final USC ADRC diagnoses in four ADRC categories and the telephone survey classifications in three categories for all 38 cases.
Table 2 Comparison of results of USC ADRC and telephone survey-based classification.
USC ADRC Classification
AD or Mixed Dementia MCI AAMI Normal All
Classification by Telephone Assessment Dementia 5 0 0 0 5
Cognitive Impairment 0 2 1 1 4
No or Minimal Impairment 1 2 3 23 29
All 6 4 4 24 38
AD Alzheimer's disease
MCI mild cognitive impairment
AAMI age-associated memory impairment
Table 3 shows the classifications broken down into two categories for both the USC ADRC and telephone classifications: dementia and not dementia. When using this staged classification method, the sensitivity of the telephone method to detect dementia was 0.83 (95% CI 0.36,1.00), and the specificity to classify as not dementia was 1.00 (95% CI 0.89,1.00). Kappa was 0.89 (95% CI 0.69,1.00).
Table 3 Comparison of results of USC ADRC "gold standard" and telephone survey-based classification of dementia and no dementia
Gold Standard Classification
Dementia Not Dementia Total
Classification by Telephone Assessment Dementia 5 0 5
Not Dementia 1 32 33
All 6 32 38
Telephone Assessment
Sensitivity for dementia 5/6 = 0.83 (95% confidence interval 0.36, 1.00)
Specificity for not dementia 32/32 = 1.00 (95% confidence interval 0.89, 1.00)
Kappa 0.89 (95% confidence interval 0.69, 1.00)
Tables 4 and 5 show both the ADRC and telephone classifications in three categories under two sets of assumptions about whether individuals with AAMI are properly classified as normal or cognitively impaired. In both tables, whether AAMI is classified as normal or cognitively impaired, sensitivity to detect dementia is 0.83 (95% C.I. 0.35, 0.99). In Table 4 under the assumption that AAMI is normal, the specificity of the telephone approach is 0.93 (95% C.I. 0.77, 0.99), while the sensitivity for cognitive impairment is 0.50 (95% C.I. 0.07, 0.93). Kappa is 0.67 (95% C.I. 0.45, 0.89). In Table 5 under the assumption that AAMI is cognitive impairment, the specificity of the telephone classification is 0.96 (95% C.I. 0.79, 1.00) and the sensitivity for cognitive impairment is 0.38 (95% C.I. 0.09, 0.76). Kappa is 0.61 (95% C.I. 0.37, 0.85).
Table 4 Comparison of results of USC ADRC and telephone survey-based classification in three categories. Assumption 1: AAMI (age-associated memory impairment) considered normal.
Gold Standard Classification
AD or Mixed Dementia Cognitive Impairment Normal All
Classification by Telephone Assessment Dementia 5 0 0 5
Cognitive Impairment 0 2 2 4
No or Minimal Impairment 1 2 26 29
All 6 4 28 38
Telephone Assessment
Sensitivity for dementia 5/6 = 0.83 (95% confidence interval 0.35, 0.99)
Sensitivity for cognitive impairment 2/4 = 0.50 (95% confidence interval 0.07, 0.93)
Specificity for normal 26/28 = 0.93 (95% confidence interval 0.77, 0.99)
Kappa 0.67 (95% confidence interval 0.45, 0.89)
Table 5 Comparison of results of USC ADRC and telephone survey-based classification in three categories. Assumption 2: AAMI (age-associated memory impairment) considered cognitive impairment.
Gold Standard Classification
AD or Mixed Dementia Cognitive Impairment Normal All
Classification by Telephone Assessment Dementia 5 0 0 5
Cognitive Impairment 0 3 1 4
No or Minimal Impairment 1 5 23 29
All 6 8 24 38
Telephone Assessment
Sensitivity for dementia 5/6 = 0.83 (95% confidence interval 0.35, 0.99)
Sensitivity for cognitive impairment 3/8 = 0.38 (95% confidence interval 0.09, 0.76)
Specificity for normal 23/24 = 0.96 (95% confidence interval 0.79, 1.00)
Kappa 0.61 (95% confidence interval 0.37, 0.85)
Table 6 gives the TICSm scores and other clinical information for patients with definite disagreement between the ADRC and the telephone classification. These included the single patient classified as having dementia by the ADRC and as unimpaired based on the telephone assessment; the two patients classified as having no or minimal impairment in the telephone assessment and Mild Cognitive Impairment (MCI) by the ADRC; and the one patient classified with cognitive impairment in the telephone assessment who was classified as normal by the USC ADRC.
Table 6 TlCSm scores and other clinical information for patients with definite disagreement between USC ADRC classification and telephone survey classification.
USC ADRC Classification Telephone Survey Classification TICSm Score Clinical History
Dementia No/minimal impairment 29 --
MCI No/minimal impairment 28 History of seizures
MCI No/minimal impairment 30 History of cardiovascular disease
Normal Cognitive impairment 24 Anxiety and depression
MCI mild cognitive impairment
Discussion
Our study shows that assessment of dementia status in elderly subjects by a multi-stage telephone method has good agreement with a comprehensive in-person dementia assessment. Sensitivity and specificity for classification of cognitively normal or no dementia were also good. The one case of dementia classified falsely as not impaired in the telephone assessment may have been due to the high educational level (BA degree) and IQ (estimated premorbid Verbal IQ of 120) of this subject. Our approach does not make adjustments for education in the TICSm cutpoint, therefore dementia in persons with high premorbid education and/or IQ may be missed.
Note that dementia in this study was confined to AD dementia and the definition of dementia required multiple problems with memory. This study does not specifically address the accuracy of the telephone methods for detecting other types of dementia which are not characterized by predominant memory impairment.
On the other hand, sensitivity and specificity of the telephone approach for cognitive impairment was low, especially if one considers AAMI as a form of cognitive impairment. Even in clinical settings, however, the criteria for measuring and defining mild cognitive impairment (MCI) are not uniform or standardized [18,19]. Consensus on the definition of MCI would be critical for clinicians and researchers since a number of studies have suggested that MCI is a prodrome or precursor to dementia. Progression or conversion from MCI to dementia can range from 6–25% annually [20,21] to 23 – 47% over 2.6 years [22]. Clinic-based studies of MCI suggest more uniform progression to dementia than population-based ones where the classification is more unstable [23]. Regardless of assessment method used, the instability of the MCI classification can be seen in studies where one third [24,25] to one half revert back to normal cognition [26]. These findings could explain some of the classification differences between our two assessment methods.
Reliance on the telephone assessment would misclassify some subjects. In general, misclassification in epidemiologic studies causes a bias to the null when it is non-differential [27]. In large population studies, because screening test scores often detect possible dementia or cognitive impairment, the percentage of subjects invited for in-person clinical and neuropsychologic examinations can be substantial. Failure to participate in clinical examinations also has the potential to cause bias especially if failure to attend is related to cognitive impairment. In the Cache County study, for example, 31% of subjects with an indication of dementia did not return for the requested clinical work-up [3]. In comparing in-person assessment of cognitive status using gold-standard neuropyschologic testing with telephone assessment, bias due to misclassification must be balanced against bias due to non-response.
Epidemiologic studies that involve in-person assessments of cognition are costly compared with those that rely solely on telephone assessment. We estimate that the telephone strategy used here costs about $38.00 per person assessed when the percentage of those assessed who "fail" the TICSm (score ≤ 27) is 25%. The costs of in-person assessments are generally 20 times that cost.
The costs of in-person assessments also affected the number of cases in this analysis. The small number of patients who were willing to be interviewed for the in-person assessments as well as follow-up telephone interviews is a further limitation. In our previous work we have found that there tends to be a bias toward non-participation for the cognitively impaired and demented [28].
Conclusion
Use of the multi-stage telephone strategy to identify people who have dementia or cognitive impairment is promising as a way to identify potential participants in intervention or clinical research. Use of the strategy in dementia detection studies is less costly, does not require representative samples, can be generally smaller than epidemiologic studies, and can be designed to validate eligibility prior to study enrollment.
Competing interests
None of the authors have competing personal or financial interests in the interpretation or presentation of this data.
Authors' contributions
VCC & DBP contributed to all aspects of design, analyses and implementation and interpretation of study, and drafts, revisions and critical review of paper.
LC & HC contributed to implementation of study, interpretation of data and critical review of paper.
VC contributed to analyses and critical review of paper.
All have given final approval of this submission.
Appendix I. USC ADRC Test Battery
Screening
Mini Mental State Examination
Short Blessed Test
Estimated Premorbid IQ
American Version of the Nelson Adult Reading Test (AMNART)
Language
Boston Naming Test
Phrase repetition – Boston Diagnostic Aphasia Exam
Oral Word Association Test
Animal Naming
Token Test – Multilingual Aphasia Examination
Construction
Rey-Osterrieth Complex Figure, copy
Judgment of Line Orientation
Clock Drawing
Block Design -WAIS- R
Memory
Logical Memory I and II- Wechsler Memory Scale III
California Verbal Learning Test – II
Faces I and II – Wechsler Memory Scale III
Rey-Osterrieth Complex Figure 3' Delay
Executive Control
Digit Span, forward and backward – WAIS-R
Trail Making, A and B
Number-Letter Sequencing – Wechsler Memory Scale III
Repeated patterns
Emotional Status
Geriatric Depression Scale
Symptoms Checklist 90-R
Functional Ability
Clinical Dementia Rating Scale (completed by a collateral informant)
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by grants R01-AG-14745 and P50-AG-05142 from the National Institute of Aging.
==== Refs
Gifford DR Cummings JL Evaluating dementia screening tests: methodologic standards to rate their performance. Editorial Neurology 1999 52 224 227 9932934
Gallo JJ Breitner JC Alzheimer's disease in the NAS-NRC Registry of aging twin veterans, IV. Performance characteristics of a two-stage telephone screening procedure for Alzheimer's dementia Psychol Med 1995 25 1211 1219 8637951
Zandi PP Carlson MV Plassman BL Welsh-Bohmer KA Steffens DC Breitner JCS Hormone replacement therapy and the incidence of Alzheimer Disease in older women the Cache County Study JAMA 2002 288 2123 2129 12413371 10.1001/jama.288.17.2123
Welsh KA Breitner JCS Magruder-Habib KM Detection of dementia in the elderly using telephone screening of cognitive status Neuropsychiatry Neuropsychol Behav Neurol 1993 6 103 110
Plassman BL Newman TT Welsh KA Helms M Breitner JCS Properties of the telephone Interview for Cognitive Status: application in epidemiological and longitudinal studies Neuropsychiatry Neuropsychol Behav Neurol 1994 7 235 241
Silverman JM Breitner JC Mohs RC Davis KL Reliability of the family history method in genetic studies of Alzheimer's disease and related dementias Am J Psychiatry 1986 143 1279 1282 3766791
McKhann G Drachman D Folstein M Katzman R Price D Stadlan EM Clinical diagnosis of Alzheimer's disease: Report of the NINCDS-ADRDA work group under the auspices of the Department of Health and Human Services Task Force on Alzheimer's disease Neurology 1984 34 939 944 6610841
Folstein MF Folstein SE McHugh PR "Mini-mental state": a practical method for grading the cognitive state of patients for the clinician J Psychiatr Res 1975 12 189 98 1202204 10.1016/0022-3956(75)90026-6
Brandt J Spencer M Folstein M The Telephone Interview for Cognitive Status Neuropsychiatry, Neuropsychol Behav Neurol 1998 1 111 117
De Jager CA Budge MM Clarke R Utility of TICS-m for the assessment of cognitive function in older adults Int J Geriatr Psychiatry 2003 18 318 24 12673608 10.1002/gps.830
Buckwalter JG Crooks VC Petitti DB A psychometric analysis of a computer-assisted administration of the Telephone Interview for Cognitive Status-modified J Clin and Exp Neuropsychol 2002 24 168 175 11992199
Beeri MS Werner P Davidson M Schmindler J Silverman J Validation of the modified telephone interview for cognitive status (TICS-m) in Hebrew Int J Geriatr Psychiatry 2003 18 381 6 12766912 10.1002/gps.840
Kawas C Segal J Stewart WF Corrada M Thal LJ A validation study of the dementia questionnaire Arch Neurol 1994 51 901 906 8080390
Li G Aryan M Silverman JM Haroutunian V Perl DP Birstein S Lantz M Marin DB Mohs RC Davis KL The validity of the family history method for identifying Alzheimer disease Arch Neurol 1997 54 634 640 9152121
Ellis RJ Jan K Kawas C Koller WC Lyons KE Jeste DV Hansen LA Thal L Diagnostic validity of the dementia questionnaire for Alzheimer disease Arch Neurol 1998 55 360 5 9520010 10.1001/archneur.55.3.360
Fleiss JL Statistical Methods for Rates and Proportions 1981 2 New York, John Wiley & Sons
CYTEL Software Corporation StatXact 5 2001 Cambridge, MA, CYTEL Software Corporation
Luis CA Lowenstein DA Acevedo A Barker WW Duara R Mild cognitive impairment: directions for future research Neurology 2003 61 438 44 12939414
DeCarli C Mild cognitive impairment: prevalence, prognosis, aetiology, and treatment Lancet Neurol 2003 2 15 21 12849297 10.1016/S1474-4422(03)00262-X
Petersen RC Stevens JC Ganguli M Tangalos EG Cummings JL DeKosky ST Practice parameter:early detection of dementia: mild cognitive impairment (an evidence-based review). Report of the Quality Standards Subcommittee of the American Academy of Neurology Neurology 2001 56 1133 1142 11342677
Bruscoli M Lovestone S Is MCI really just early dementia? A systematic review of conversion studies Int Psychogeriatr 2004 6 129 40 15318760 10.1017/S1041610204000092
Busse A Bischkopf J Reidel-Heller SG Angermeyer MC Mild cognitive impairment: prevalence and incidence according to different diagnostic criteria. Results of the Leipzig Longitudinal Study of the Aged (LEILA75+) Br J Psychiatry 2003 182 449 54 12724250 10.1192/bjp.182.5.449
Davis HS Rockwood K Conceptualization of mild cognitive impairment: a review Int J Geriatr Psychiatry 2004 19 313 9 15065223 10.1002/gps.1049
Fisk JD Merry HR Rockwood K Variations in case definition affect prevalence but not outcomes of mild cognitive impairment Neurology 2003 61 1179 84 14610117
Larrieu S Letenneur L Orgogozo JM Fabrigoule C Amieva H Le Carret N Barberger-Gateau P Dartigues JF Incidence and outcome of mild cognitive impairment in a population-based prospective cohort Neurology 2002 59 1594 9 12451203
Ganguli M Dodge HH Shen C DeKoskey ST Mild cognitive impairment, amnestic type: an epidemiologic study Neurology 2004 63 115 21 15249620
Rothman KJ Greenland S Modern Epidemiology 1998 2 Philadelphia, PA, Lippincott-Raven
Petitti DB Buckwalter JG Crooks VC Chiu V Prevalence of dementia in users of hormone replacement therapy as defined by prescription data J Gerontol A Biol Sci Med Sci 2002 57 M532 M538 12145368
| 15829005 | PMC1090587 | CC BY | 2021-01-04 16:28:53 | no | BMC Neurol. 2005 Apr 13; 5:8 | utf-8 | BMC Neurol | 2,005 | 10.1186/1471-2377-5-8 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-231581998810.1186/1471-2202-6-23Research ArticleBifrontal transcranial direct current stimulation slows reaction time in a working memory task Marshall Lisa [email protected]ölle Matthias [email protected] Hartwig R [email protected] Jan [email protected] University of Lübeck, Department of Neuroendocrinology H23a, Ratzeburger Allee 160, 23538 Lübeck, Germany2 University of Kiel, Department of Neurology, Niemannsweg 147, 24105 Kiel, Germany2005 8 4 2005 6 23 23 20 12 2004 8 4 2005 Copyright © 2005 Marshall 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
Weak transcortical direct current stimulation (tDCS) applied to the cortex can shift the membrane potential of superficial neurons thereby modulating cortical excitability and activity. Here we test the possibility of modifying ongoing activity associated with working memory by tDCS. The concept of working memory applies to a system that is capable of transiently storing and manipulating information, as an integral part of the human memory system. We applied anodal and cathodal transcranial direct current (tDCS) stimulation (260 μA) bilaterally at fronto-cortical electrode sites on the scalp over 15 min repeatedly (15 sec-on/15 sec-off) as well as sham-tDCS while subjects performed a modified Sternberg task.
Results
Reaction time linearly increased with increasing set size. The slope of this increase was closely comparable for real and sham stimulation indicating that our real stimulation did not effect time required for memory scanning. However, reaction time was slowed during both anodal and cathodal stimulation as compared to placebo (p < 0.05) indicating that real stimulation hampered neuronal processing related to response selection and preparation.
Conclusion
Intermittent tDCS over lateral prefrontal cortex during a working memory task impairs central nervous processing related to response selection and preparation. We conclude that this decrease in performance by our protocol of intermittent stimulation results from an interference mainly with the temporal dynamics of cortical processing as indexed by event-related sustained and oscillatory EEG activity such as theta.
==== Body
Background
Application of transcranial direct current stimulation (tDCS) to the cortex has been shown to shift the membrane potential of superficial neurons in a de- or hyperpolarizing direction, and to modulate spontaneous neuronal activity as well as the processing of afferent signals (for reviews see [1-3]). Anodal stimulation, i.e., stimulation with electrodes of positive charge, causes via an extracellular negative sink in underlying neural tissue a depolarization of the membrane potential, whereas cathodal stimulation hyperpolarizes the neurons. A recent in vitro study using a cortical slice preparation showed that pyramidal cells oriented perpendicular to the cortical surface, i.e., closely parallel to the DC field, can be readily polarized [4]. At the behavioral level in animals anodal stimulation of the cortical surface has been associated with facilitation of an unconditioned response [5-7] and improved learning [8,9]. In humans, anodal polarization increased excitability measures of the motor and visual cortex, improved motor learning, decreased response latencies and increased verbal fluency [2,3,10].
The present study analizes the ability of tDCS, with electrodes overlaying lateral prefrontal cortex, to modulate activity within cortical networks underlying working memory using the Sternberg paradigm [11]. Working memory refers to a set of basic mental operations which define the ability to hold an item of information transiently in mind, in order to recall, manipulate and/or associate this information to other ideas and incoming new information. Working memory is assumed to involve a number of subsystems [12]. Sustained neuronal activity up to several tens of seconds in the prefrontal cortex, but also in other areas e.g., parietal cortex is hereby an essential correlate of working memory operations [13,14]. The generation of this persistent activity has been modeled mostly through recurrent excitatory networks, although it may also rely on specific changes in intrinsic membrane properties [15-18].
Working memory and the effect of memory load are well assessed in the Sternberg paradigm which requires subjects to maintain in memory a set of initially presented items and compare these with a probe [11]. Subsequently, individual items are presented and subjects have to perform a fast button press response indicating whether the presented item belongs to the set or not. Changes in reaction time as a function of memory load are attributed to a manipulation of working memory [11,19,20]. In the EEG and MEG frontal theta and gamma activity are enhanced during the retention and scanning periods of working memory [21-24]. An event-related sustained negative potential shift recorded from fronto-cortical scalp locations also represents a close electrophysiological correlate of memory load [25,26].
Results
Reaction time
Reaction time was increased for both tDCS polarities as compared to placebo (anodal vs. sham-tDCS: F[1,11] = 8.99, p < 0.01; cathodal vs. sham-tDCS F[1,11] = 7.17, p < 0.05; for the main effect of stimulation: F[2,22] = 4.56, p < 0.05). Reaction times in response to anodal vs. cathodal tDCS, however, did not differ significantly (F[1,11] = 0.04, p > 0.8). Fig. 1 reveals mean reaction times of the Sternberg task for all three conditions and memory set sizes. In the Sternberg task reaction time increases linearly with the number of items in the memory set. This linear increase in reaction time with increased memory load corresponds to the slope of the reaction time function. The reaction time at the zero intercept of this function is attributed to processes independent of set size such as response selection and preparation [11,20]. Since slopes and zero intercepts of the reaction time functions for anodal and cathodal tDCS conditions did not differ significantly statistics are given for collapsed data only. As is typical for the Sternberg task [11] reaction time increased for larger set sizes, i.e. for increased memory load (see Fig. 1; F[2,22] = 65.04, p < 0.001, for the main effect of set size). A major finding is that slopes did not differ between tDCS conditions and sham-tDCS (Mean ± SEM of positive responses, tDCS: 29.6 ± 3.8 ms/item, sham-tDCS: 31.2 ± 4.1 ms/item; negative responses, tDCS: 32.1 ± 3.9 ms/item, sham-tDCS: 32.3 ± 4.9 ms/item, F[1,11] = 0.19, p > 0.6 and F[1,11] = 0.43, p > 0.5, for the main effect of polarity and response type, respectively) indicating that tDCS did not effect time required for memory scanning. The above data also show that slopes for negative and positive probes did not differ statistically in line with Sternberg's hypothesis that a serial exhaustive search processes is taking place [11]. For both negative and positive response types zero intercepts for tDCS were higher (slower reaction times) than for sham-tDCS (F[1,11] = 11.77, p < 0.01) reflecting an increased time required for response selection and preparation with tDCS. Positive responses revealed higher zero intercepts (slower reaction times) than negative responses (F[1,11] = 22.38, p < 0.001, Fig. 1). For positive responses the zero intercepts were at 421.3 ± 9.8 ms for tDCS and at 395.9 ± 11.8 ms for sham-tDCS, and for negative responses at 371.7 ± 12.2 ms for tDCS and at 343.3 ± 9.7 ms for sham-tDCS. The faster reaction time to negative than positive probes is due to the greater probability of negative probes [27].
In supplementary analyses with the same Sternberg task repetitive anodal tDCS applied unilaterally rather than bilaterally at left and right fronto-cortical regions (F3, F4) during separate sessions also showed a tendency to reduce reaction time as compared to sham stimulation (F[1,9]= 3.18, p = 0.10, for the main effect of stimulation, with data collapsed for F3- and F4-tDCS vs. sham-tDCS, 594.2 ± 41.5 vs. 581.5 ± 42.1 msec, respectively). As in the main experiment slopes did not differ between tDCS conditions and sham-tDCS (p > 0.7).
Error rate
tDCS did not affect error rate (F[2,22] = 0.88, p > 0.42). Subjects performed the task with only few errors, whereby the mean number of incorrect responses to positive probes was larger than to negative probes (1.11 ± 0.32 vs. 0.43 ± 0.09, p < 0.01). Nevertheless, the overall error rate across all conditions was low, 3.1 %. For positive responses only, there was a statistical trend for an interaction between set size and response type showing an increase in the number of errors with working memory load (F[2,22] = 2.79, p < 0.10). Errors for the set size of 4 items (2.22 ± 0.40) occurred more frequent than errors for the set size of one (1.11 ± 0.36) and two items (1.5 ± 0.33, p < 0.05).
Discussion
This study shows that both cathodal as well as anodal tDCS polarization applied at lateral prefronto-cortical locations slowed reaction time in a Sternberg item recognition task equally for all set sizes tested. Despite overall slower reaction time during tDCS the slope for reaction time with increasing set size was not significantly altered by tDCS. This suggests, that contrary to our expectation tDCS influenced more directly processes related to response selection than operations in working memory. However, attempts to localize underlying processes and structures suggest these processes may not be as separable as originally suggested by Sternberg [11,20]. The absence of an effect of tDCS on error rate in our study is attributed to the rather low error rate across all conditions presumably associated with a "floor" effect.
Interestingly, the slowing effect of tDCS on reaction time contrasts findings in another study in which anodal tDCS improved reaction time [28]. However, the two studies differ in several aspects: Firstly, in the location of stimulation. The facilitatory effect on reaction time found by Elbert and co-workers was obtained when anodal tDCS was applied at the vertex close to the supplementary motor area. It may therefore be argued that the direction of intracortical current flow induced by our bilateral anodal tDCS at fronto-lateral sites was responsible for the impairment of reaction time. Yet, reversing current flow in our study, i.e., cathodal tDCS also impaired reaction time. Thus, the difference in stimulation location is probably not the sole reason for the discrepancy in the effects. Second, our stimulation protocol differed from that of Elbert and co-workers. In our study stimulation occurred in an intermittent manner, and was not event-related. The finding that both tDCS polarities had a decremental effect on reaction time furthermore suggests that this effect is to be attributed to the intermittent application mode of both anodal (the constant current alternated every 30 sec between zero and positive polarity) and cathodal (alternating between zero and negative current) tDCS. This intermittent tDCS protocol was chosen, on the one hand, for safety reasons. In order to stimulate during the complete Sternberg task a duration of 15 min was required. However studies to date have not used or evaluated safety effects of continuous tDCS for such a long duration. On the other hand, exactly the same intermittent stimulation protocol has been applied effectively in a previous study of ours resulting in improved sleep-dependent consolidation of memory [29]. Thus, the presumed fluctuating shifts in the membrane potentials of cortical neurons induced by our intermittent tDCS protocol could interfere with temporal dynamics of processing. A critical role of a finely tuned temporal pattern of neuronal processing is suggested by the event-related or time-locked slow negative cortical potentials during working memory operations including response selection and preparation [30-33]. Moreover, as mentioned above working memory operations including response selection are associated with modulations in oscillatory activity, in particular enhanced theta activity [21-24,34-37], and thus with the underlying finely tuned intercellular cortical activity [14,19,38]. The ability of tDCS to influence oscillatory cortical network activity has been reported previously [29,39]. With regard to theta activity, some recent findings lend themselves to speculation that the decreasing influence of tDCS on these oscillations involves a detrimental effect on cholinergic afferent input to the upper neocortical layer I [4,40,41].
Taken together, interference with endogenous EEG rhythms and/or slow potential activity linked to task associated network activity is suggested to have caused the poorer response latency with tDCS. Here, it should be underlined that the cortical networks underlying response selection which were presumably influenced by tDCS rather than other working memory operations are functions involved in many other types of cognitive tasks. Furthermore, common regions of the human frontal lobe are recruited by different cognitive demands [42]. The low spatial selectivity of tDCS and the different engagement of frontal sub-regions in working memory operations including response selection may be a confounding factor in our study. Our results, however, are in line with findings of studies on working memory with fMRI and rTMS, methods of much higher spatial resolution than tDCS, indicating that the dorsolateral prefrontal cortex is specifically involved in response selection [43-45]. The dorsolateral prefrontal cortex is concurrently involved in the active manipulation of information within working memory, but in humans the neuronal networks of these functions appear not to be identical [45]. The effect on response selection and preparation rather than working memory in the present study may thus have been related as well to the specific location of the stimulating electrodes and direction of current flow within the stimulated cortical region. Also, during the task networks underlying response related processes may have shown greater susceptibility (e.g. increased neuronal excitability) to tDCS. On the other hand, it cannot be excluded that increased memory load exceeding the relatively low load used in our task may have rendered the working memory networks more susceptible to the modulatory effects of tDCS.
Conclusion
Our data show that tDCS applied over the dorsolateral prefrontal cortex rather than affecting working memory interfered with response selection and preparation. This result is attributed to a disturbance by intermittent tDCS of endogenous task-related cortical oscillatory activity. Both stimulation polarities caused fluctuating shifts in the membrane potentials of cortical neurons independent of the endogenous activity which seems to have interfered with the time-locked selection and/or generation of the response. The new finding is that tDCS modulated the excitability and activity of cortical networks which are involved in response selection and probably also in other types of cognitive tasks. Whether intermittent tDCS exerts a facilitatory or suppressing effect on central nervous information processing may depend upon the behavioral context, and associated requirements regarding the temporal dynamics of processing within the underlying networks.
Methods
Subjects and procedure
Twelve subjects (6 females) aged 19 to 27 years, free of medication and non-smokers participated in the experiment proper. Subjects with any of the following (or the history of) were excluded: epilepsy, paroxysms, cognitive impairments, mental, hormonal, metabolic or circulatory disorders. All participants gave written informed consent. The experiments were approved by the ethics committee of the University of Lübeck and were conducted in accordance with the Declaration of Helsinki. Subjects were tested on three conditions: anodal, cathodal and placebo stimulation, in separate sessions according to a double-blind crossover design. Sessions were separated by an interval of at least 1 week. Subjects were seated comfortably in front of a monitor presenting a modified visual Sternberg task. First a set of 1, 2 or 4 items (letters of the alphabet) was presented to memorize. After subjects pressed a button the memory set disappeared and after 2 sec a series of probe stimuli was presented. Every probe was paired with a mildly alerting tone (1000 Hz, 20 ms) to maintain the same high level of alertness throughout the experiment and subjects had to indicate whether or not the item was a member of the memorized set by pressing a "yes" or a "no" button with the index finger of their dominant hand. After this button press the probe was replaced with a feedback light for 200 ms informing the subject whether their response was correct (green) or incorrect (red). The next probe stimulus appeared after an interval of 1 sec. Instructions stressed both speed and accuracy, but underscored the importance of high accuracy. Six blocks of 65 trials were presented, whereby the first 15 trials in every block served as practice. Each memory set size was presented twice. Order of set size was randomized between subjects, but remained the same within a session and subject. Of the probe stimuli 25% required a positive response ("yes") and 75% a negative response ("no"). Presentation length of memory set and probe stimuli were terminated by the subjects' response. Only correct trials with reaction times not exceeding mean reaction times ± 2 SD were used for analyses.
Transcranial direct current stimulation (tDCS)
For tDCS two pairs of electrodes (8 mm diameter) were applied bilaterally at two fronto-lateral locations (F3 and F4 of the international 10:20 system, [46]), and at the two mastoids, in accordance with the dorsolateral prefrontal cortical activations reported in imaging studies involving working memory [47-49]. Anodal tDCS (positively charged electrode) or cathodal tDCS (negatively charged electrode) were applied intermittently (current strength: 260 μA; 15 sec-on/15 sec-off; two sec rise and fall time) over a period of 15 minutes by a battery driven constant current stimulator. Intermittent stimulation was employed to be comparable with a previous study [29]. In the placebo session electrodes were applied as in the stimulation sessions, but the stimulator remained off. Subjects did not report having felt the stimulation.
In a supplementary analyses conducted to assess whether left sided frontal tDCS was more effective than right-sided stimulation, tDCS was applied unilaterally at the left (F3) and right (F4) prefrontal locations. These two sessions and the sham-tDCS session were separated by at least one week.
Statistical analysis
Statistical analyses of reaction time and error rate were based on ANOVA with repeated measures for polarity of stimulation (anodal, cathodal and placebo), response type (negative, positive) and set size (1, 2, 4 items). Mean reaction time as a function of memory set size was plotted. For the least-squares regression lines of these reaction time functions, slope and zero intercept were calculated and analyzed with an additional ANOVA with repeated measures factors for polarity and response type. The Greenhouse-Geisser method was used to correct for non-sphericity. Subsequently, pairwise contrasts were used to analyze reaction time and error rate (p < 0.05).
List of abbreviations
ANOVA analyses of variance
DLPFC dorsolateral prefrontal cortex
EEG electroencephalography
MEG magnetoencephalography
rTMS rapid transmagnetic stimulation
tDCS transcranial direct current stimulation
VLPFC ventrolateral prefrontal cortex
Authors' contributions
LM and MM participated in the planning and execution of the study and performed the data analysis. LM drafted the manuscript. The manuscript was subsequently revised by MM, HS and JB, and all four authors gave final approval.
Acknowledgements
We wish to thank Alexander Kraepalis for help in acquiring data, Horst Koller and Marek Zelazny of the University of Lübeck, Electronics Department, for building the constant current stimulator. This study was supported by the DFG (MA 2053, BO 854).
Figures and Tables
Figure 1 Relation between mean reaction time and size of memory set. Positive (top) and negative responses (bottom) are depicted during placebo (dotted line), anodal (continuous line) and cathodal (dashed line) tDCS stimulation (n = 12). For post-hoc t Tests data of both stimulation sessions were collapsed since they were practically identical (see text). Symbols indicate differences in reaction time between placebo and stimulation. ** p < 0.01, * p < 0.05, t p < 0.1.
==== Refs
Terzuolo CC Bullock TH Measurement of imposed voltage gradient adequate to modulate neuronal firing. Proc Natl Acad Sci U S A 1956 42 687 694
Nitsche MA Liebetanz D Antal A Lang N Tergau F Paulus W Modulation of cortical excitability by weak direct current stimulation--technical, safety and functional aspects Suppl Clin Neurophysiol 2003 56 255 276 14677403
Priori A Brain polarization in humans: a reappraisal of an old tool for prolonged non-invasive modulation of brain excitability Clin Neurophysiol 2003 114 589 595 12686266 10.1016/S1388-2457(02)00437-6
Bikson M Inoue M Akiyama H Deans JK Fox JE Miyakawa H Jefferys JG Effects of uniform extracellular DC electric fields on excitability in rat hippocampal slices in vitro J Physiol 2004 557 175 190 14978199 10.1113/jphysiol.2003.055772
Morrell F Naitoh P Effect of cortical polarization on a conditioned avoidance response. Exp Neurol 1962 6 507 523 10.1016/0014-4886(62)90075-4
Rusinov VS The dominant focus: Electrophysiological investigations 1973 New York, Plenum Press
Yamaguchi K Hori Y Long lasting retention of cortical dominant focus in rabbit Med J Osaka Univ 1975 26 39 50 1228391
Albert DJ The effects of polarizing currents on the consolidation of learning. Neuropsychologia 1966 4 65 77 10.1016/0028-3932(66)90021-2
Rosen SC Stamm JS Transcortical polarization: facilitation of delayed response performance by monkeys Exp Neurol 1972 35 282 289 4624141 10.1016/0014-4886(72)90154-9
Iyer MB Mattu U Grafman J Wassermann EM Direct current polarization of the human prefrontal cortex facilitates verbal fluency Society for Neuroscience 2004 Program No. 226.1 Washington, DC: Society for Neuroscience, 2004. Online
Sternberg S High-speed scanning in human memory Science 1966 153 652 654 5939936
Goldman-Rakic PS Regional and cellular fractionation of working memory Proc Natl Acad Sci U S A 1996 93 13473 13480 8942959 10.1073/pnas.93.24.13473
Babiloni C Babiloni F Carducci F Cincotti F Vecchio F Cola B Rossi S Miniussi C Rossini PM Functional frontoparietal connectivity during short-term memory as revealed by high-resolution EEG coherence analysis Behav Neurosci 2004 118 687 697 15301596 10.1037/0735-7044.118.4.687
Passingham D Sakai K The prefrontal cortex and working memory: physiology and brain imaging Curr Opin Neurobiol 2004 14 163 168 15082320 10.1016/j.conb.2004.03.003
Compte A Brunel N Goldman-Rakic PS Wang XJ Synaptic mechanisms and network dynamics underlying spatial working memory in a cortical network model Cereb Cortex 2000 10 910 923 10982751 10.1093/cercor/10.9.910
McCormick DA Shu Y Hasenstaub A Sanchez-Vives M Badoual M Bal T Persistent cortical activity: mechanisms of generation and effects on neuronal excitability Cereb Cortex 2003 13 1219 1231 14576213 10.1093/cercor/bhg104
Seamans JK Nogueira L Lavin A Synaptic basis of persistent activity in prefrontal cortex in vivo and in organotypic cultures Cereb Cortex 2003 13 1242 1250 14576215 10.1093/cercor/bhg094
Wang XJ Tegner J Constantinidis C Goldman-Rakic PS Division of labor among distinct subtypes of inhibitory neurons in a cortical microcircuit of working memory Proc Natl Acad Sci U S A 2004 101 1368 1373 14742867 10.1073/pnas.0305337101
Jensen O Lisman JE An oscillatory short-term memory buffer model can account for data on the Sternberg task J Neurosci 1998 18 10688 10699 9852604
Sternberg S Memory-scanning: mental processes revealed by reaction-time experiments Am Sci 1969 57 421 457 5360276
Gevins A Smith ME McEvoy L Yu D High-resolution EEG mapping of cortical activation related to working memory: effects of task difficulty, type of processing, and practice Cereb Cortex 1997 7 374 385 9177767 10.1093/cercor/7.4.374
Sarnthein J Petsche H Rappelsberger P Shaw GL von Stein A Synchronization between prefrontal and posterior association cortex during human working memory Proc Natl Acad Sci U S A 1998 95 7092 7096 9618544 10.1073/pnas.95.12.7092
Jensen O Tesche CD Frontal theta activity in humans increases with memory load in a working memory task Eur J Neurosci 2002 15 1395 1399 11994134 10.1046/j.1460-9568.2002.01975.x
Schack B Vath N Petsche H Geissler HG Moller E Phase-coupling of theta-gamma EEG rhythms during short-term memory processing Int J Psychophysiol 2002 44 143 163 11909647 10.1016/S0167-8760(01)00199-4
Kotchoubey BI Jordan JS Grozinger B Westphal KP Kornhuber HH Event-related brain potentials in a varied-set memory search task: a reconsideration Psychophysiology 1996 33 530 540 8854740
Pelosi L Hayward M Blumhardt LD Which event-related potentials reflect memory processing in a digit-probe identification task? Brain Res Cogn Brain Res 1998 6 205 218 9479072 10.1016/S0926-6410(97)00032-3
Krinchik EP Koster WG The probabiliy of a signal as a determinant of reaction time. Acta Psychologica 30, Attention and Performance II 2004 Amsterdam, The Netherlands, North-Holland Publishing Company 27 36
Elbert T Lutzenberger W Rockstroh B Birbaumer N The influence of low-level transcortical DC-currents on response speed in humans Int J Neurosci 1981 14 101 114 7263138
Marshall L Mölle M Hallschmid M Born J Transcranial direct current stimulation during sleep improves declarative memory J Neurosci 2004 24 9985 9992 15525784 10.1523/JNEUROSCI.2725-04.2004
Wang Y Tian S Wang H Cui L Zhang Y Event-related potentials in a No-go task involving response-tendency conflict Clin Electroencephalogr 2002 33 82 85 12025736
Suchan B Zoppelt D Daum I Frontocentral negativity in electroencephalogram reflects motor response evaluation in humans on correct trials Neurosci Lett 2003 350 101 104 12972163 10.1016/S0304-3940(03)00879-6
Starr A Dong CJ Michalewski HJ Brain potentials before and during memory scanning Electroencephalogr Clin Neurophysiol 1996 99 28 37 8758968 10.1016/0921-884X(96)95147-4
Lang W Starr A Lang V Lindinger G Deecke L Cortical DC potential shifts accompanying auditory and visual short-term memory Electroencephalogr Clin Neurophysiol 1992 82 285 295 1372549 10.1016/0013-4694(92)90108-T
Shibata T Shimoyama I Ito T Abla D Iwasa H Koseki K Yamanouchi N Sato T Nakajima Y The synchronization between brain areas under motor inhibition process in humans estimated by event-related EEG coherence Neurosci Res 1998 31 265 271 9809585 10.1016/S0168-0102(98)00046-7
Raghavachari S Kahana MJ Rizzuto DS Caplan JB Kirschen MP Bourgeois B Madsen JR Lisman JE Gating of human theta oscillations by a working memory task J Neurosci 2001 21 3175 3183 11312302
Tesche CD Karhu J Theta oscillations index human hippocampal activation during a working memory task Proc Natl Acad Sci U S A 2000 97 919 924 10639180 10.1073/pnas.97.2.919
Serrien DJ Pogosyan AH Cassidy MJ Brown P Anticipatory cortico-cortical interactions: switching the task configuration between effectors Exp Brain Res 2004 154 359 367 14618284 10.1007/s00221-003-1667-2
Buzsaki G Theta oscillations in the hippocampus Neuron 2002 33 325 340 11832222 10.1016/S0896-6273(02)00586-X
Antal A Varga ET Kincses TZ Nitsche MA Paulus W Oscillatory brain activity and transcranial direct current stimulation in humans Neuroreport 2004 15 1307 1310 15167555
Mechawar N Cozzari C Descarries L Cholinergic innervation in adult rat cerebral cortex: a quantitative immunocytochemical description J Comp Neurol 2000 428 305 318 11064369 10.1002/1096-9861(20001211)428:2<305::AID-CNE9>3.0.CO;2-Y
Blatow M Rozov A Katona I Hormuzdi SG Meyer AH Whittington MA Caputi A Monyer H A novel network of multipolar bursting interneurons generates theta frequency oscillations in neocortex Neuron 2003 38 805 817 12797964 10.1016/S0896-6273(03)00300-3
Duncan J Owen AM Common regions of the human frontal lobe recruited by diverse cognitive demands Trends Neurosci 2000 23 475 483 11006464 10.1016/S0166-2236(00)01633-7
Hadland KA Rushworth MF Passingham RE Jahanshahi M Rothwell JC Interference with performance of a response selection task that has no working memory component: an rTMS comparison of the dorsolateral prefrontal and medial frontal cortex J Cogn Neurosci 2001 13 1097 1108 11784448 10.1162/089892901753294392
Rowe JB Toni I Josephs O Frackowiak RS Passingham RE The prefrontal cortex: response selection or maintenance within working memory? Science 2000 288 1656 1660 10834847 10.1126/science.288.5471.1656
Koch G Oliveri M Torriero S Carlesimo GA Turriziani P Caltagirone C rTMS evidence of different delay and decision processes in a fronto-parietal neuronal network activated during spatial working memory Neuroimage 2005 24 34 39 15588594 10.1016/j.neuroimage.2004.09.042
Jasper HH The ten-twenty electrode system of the International Federation. Electroencephalographalogr Clin Neurophysiol 1958 10 371 375
Veltman DJ Rombouts SA Dolan RJ Maintenance versus manipulation in verbal working memory revisited: an fMRI study Neuroimage 2003 18 247 256 12595179 10.1016/S1053-8119(02)00049-6
Herwig U Satrapi P Schonfeldt-Lecuona C Using the international 10-20 EEG system for positioning of transcranial magnetic stimulation Brain Topogr 2003 16 95 99 14977202 10.1023/B:BRAT.0000006333.93597.9d
Zhang JX Leung HC Johnson MK Frontal activations associated with accessing and evaluating information in working memory: an fMRI study Neuroimage 2003 20 1531 1539 14642465 10.1016/j.neuroimage.2003.07.016
| 15819988 | PMC1090588 | CC BY | 2021-01-04 16:03:48 | no | BMC Neurosci. 2005 Apr 8; 6:23 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-23 | oa_comm |
==== Front
BMC Nucl MedBMC Nuclear Medicine1471-2385BioMed Central London 1471-2385-5-11582320610.1186/1471-2385-5-1Research ArticleSodium pertechnetate (Na99mTcO4) biodistribution in mice exposed to cigarette smoke Valenca Samuel S [email protected] Elaine AC [email protected] Gláucio F [email protected] Mário [email protected] Luís Cristóvão [email protected] Departamento de Histologia e Embriologia, Universidade do Estado do Rio de Janeiro, Av. Prof. Manoel de Abreu 444 3° andar – Rio de Janeiro, RJ – 20551-170 Brasil2 Departamento de Biofísica e Biometria, Universidade do Estado do Rio de Janeiro, Av. Prof. Manoel de Abreu 444 3° andar – Rio de Janeiro, RJ – 20551-170 Brasil2005 11 4 2005 5 1 1 26 7 2004 11 4 2005 Copyright © 2005 Valenca et al; licensee BioMed Central Ltd.2005Valenca 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 biological effects of cigarette smoke are not fully known. To improve our understanding of the action of various chemical agents, we investigated the biodistribution of sodium pertechnetate (Na99mTcO4) in mice exposed to cigarette smoke.
Methods
Fifteen BALB/c male mice were exposed to the smoke of nine whole commercial cigarettes per day, 3 times/day, for up to 10 days to whole body exposure in a chamber. A control group of 5 BALB/c male mice was sham-smoked. One day later, the exposed and control groups of mice received (7.4 MBq/0.3 ml) of Na99mTcO4 before being killed at 30 min. Bones, brain, heart, intestine, kidney, liver, lungs, muscle, pancreas, spleen, stomach, testis and thyroid were weighed and these organs and blood radioactivity recorded with a gamma counter. The percentage per gram of tissue of injected dose (%ID/g) was determined for each organ.
Results
Cigarette smoke significantly decreased (p < 0.05) the %ID/g in red blood cells, bone, kidney, lung, spleen, stomach, testis and thyroid of the exposed mice.
Conclusion
The toxic effects of cigarette smoke reduced the Na99mTcO4 biodistribution.
==== Body
Background
Tobacco can be smoked in cigarettes, cigars, pipe, water pipes, or chewed. It is as an important cash crop, and has vast economical impact affecting the livelihoods of the farmers growing it, the companies that manufacture it, and the healthcare system that deals with the consequences of using it. Cigarette smoking (CS), the most popular method of smoking tobacco, is one of the most prevalent social habits practiced worldwide today [1]. The World Health Organization estimated that almost 1.1 billion people are smokers. Smoking has been identified as the leading preventable cause of death and disability in the world [2,3].
The development of diagnostic tools using radioisotopes is widely used in almost all hospitals nationwide. In nuclear medicine practice, physicians set norms for morphological or physiological function for each organ by diagnosing a large number of patients. For every procedure, there is diagnostic data for a range of normal variations familiar to physicians. Nuclear medicine practitioners interpret diseases on the basis of deviations from these limits [4]. Radioisotopes provide vital information to help diagnosis and therapy of various medical diseases. Data on tissue shape, function and localization within the body are relayed by use of one of various radionuclides, which can either be a free chemical species or covalently bound to part of a larger organic or inorganic moiety. These images are generated by the distribution of radioactive decay of the nuclide [5]. Technetium-99m (99mTc) is the most frequently used radionuclide in diagnostic nuclear medicine procedures for a wide variety of diseases [6]. Various radiochemicals have been labelled with this nuclide. 99mTc has a half-life of 6 h and emits γ-rays of 140 keV with an abundance of 90%. 99mTc is primarily obtained from 99Mo/99mTc generator and is eluted as sodium pertechnetate (Na99mTcO4). In this chemical form, it can be used to study brain and thyroid [7]. Intravenously administered 99mTc-pertechnetate is loosely bound to plasma proteins and rapidly moves out of the intravenous compartment. The plasma half-time clearance is ~30 min. Approximately 30% of the administered dose is excreted within 24 h. The total urinary and faecal excretion of 99mTc activity is about 50% in 3 days and up to 70% after 8 days. 99mTc is also trapped by the thyroid gland and it passes into the small intestine. However in the brain, the blood-brain barrier prevents Na99mTcO4from entering brain cells [6,7].
The biodistribution and kinetics of radiochemicals can be altered by a variety of chemical agents, as is widely known [8,9]. Without knowing that these chemical agents are being used, the images can be affected to give poor organ visualization, possibly necessitating repeating the procedure, and resulting in unnecessary irradiation or even misdiagnosis [10]. The effect of CS on the biodistribution of radiochemicals has not been fully evaluated. The clearance of 99mTc-pentetic acid aerosols is markedly increased in both sarcoidosis and other inflammatory lung diseases. However, the limitations of 99mTc-pentetic acids include the fact that cigarette smoking will also cause increased clearance [11]. Moreover, the pulmonary uptake of 99mTc-HMPAO (hexamethylprophylenoaminooxin) induced by smoking appears to be partially reversible after the cessation of smoking [12]. We evaluated the biodistribution as an experimental model to understand the action of various chemical agents [13-16]. This study deqals with possible changes in Na99mTcO4 biodistribution induced by the effects of CS in an animal model.
Methods
Adult male BALB/c mice were housed, 5 per cage, in a controlled environment room, with light/dark cycle conditions (12 h light/12 h dark, lights on at 6 p.m.), for an acclimatization period of >3 weeks at an ambient temperature was kept at 25 ± 2°C. The animals had free access to water and food.
Male BALB/c mice (n = 15), weighting 20–22 g, were exposed to smoke-air mixture of commercial filtered Virginia cigarettes, 3 times/day for 1 (CS1d), 5 (CS5d) or 10 (CS10d) days by whole body exposure in an inhalation chamber. A control group (n = 5), sham-smoke (exposed to environment air), was also used. This protocol was performed twice with a total of 40 animals, which were weighed at the end of the experiment.
Each group of mice was placed in the inhalation chamber (40 cm long, 30 cm wide and 25 cm high), inside an exhaustion chapel. A cigarette was coupled to a plastic 60 ml syringe so that puffs could be drawn in and subsequently expelled into the exposure chamber. We aspirated one litre of smoke from one cigarette with this syringe (20 puffs of 50 ml) and immediately injected the puff into the chamber. The 5 animals of each group were maintained in this smoke-air condition for 6 min (~ 3 %), and the inhalation chamber was opened, by removing its cover, and the smoke evacuated for 1 min by exhaustion of the chapel. This cigarette exposition was repeated three times (3 × 6 min) with intervals of 1 min (exhaustion). We repeated this procedure 3 times per day (morning, lunch time and afternoon) resulting in a 54 min of CS exposition of 9 cigarettes. This protocol has been described by us elsewhere [17] and was an experimental protocol approved by the Instituto de Biologia Roberto Alcantara Gomes Animal Research Ethics Committee – UERJ.
Twenty-four hours after exposure, the mice received 0.3 ml of Na99mTcO4 (7.4 MBq) via the ocular plexus, and 30 min later, they were rapidly killed. Heart perfusion was performed with saline at constant pressure of 25 cm H2O to clear organs of blood for 5 min. The organs were isolated (bones, brain, heart, intestine, kidney, liver, lungs, muscle, pancreas, spleen, stomach, testis and thyroid), their weights determined, and the organs and blood sample radioactivity of Na99mTcO4 measured by a gamma-counter NaI (TI) (Cobra Auto-gamma, Packard Instrument Co.; Downers Grove, Illinois, USA). The percentage per gram of tissue of injected dose (%ID/g) was determined for each organ. Statistical analyses involved one-way ANOVA, followed by the Tukey-Kramer Multiple Comparisons Test, with the significance level being p < 0.05. InStat Graphpad software was used to perform statistical analysis (GraphPad InStat version 3.00 for Windows 95, GraphPad Software, San Diego Ca, USA).
Results
The effects of CS on the biodistribution of Na99mTcO4 in isolated organs can first be seen in Table 1, which shows that CS decreased the %ID/g in red blood cells, bone, kidney, lung, spleen, stomach, testis and thyroid of the exposed mice. The greatest decrease was at CS10d in kidney (p < 0.001), spleen and testis (p < 0.01). In the stomach, CS exposure changes the %ID/g for CS1d, CS5d and CS10d (p < 0.05) were comparable with the controls. Table 2 shows that CS did not change the %ID/g of Na99mTcO4 significantly in brain, heart, intestine, liver, muscle and pancreas of exposed animals.
Table 1 Organs in which CS significantly changed (%ID/g) Na99mTcO4 biodistribution (mean ± SD).
CS effects of Na99mTcO4 biodistribution in mice
Organs Groups
Control Mean ± SD CS1d Mean ± SD CS5d Mean ± SD CS10d Mean ± SD
Red blood cells 3.63 ± 0.24 2.73 ± 0.14* 2.33 ± 0.23** 2.74 ± 0.26*
Bone 0.26 ± 0.07 0.24 ± 0.05 0.15 ± 0.07 0.13 ± 0.05*
Kidney 0.89 ± 0.15 0.80 ± 0.10 0.62 ± 0.09** 0.51 ± 0.05***
Lung 1.72 ± 0.68 1.33 ± 0.28 1.28 ± 0.12 0.86 ± 0.04*
Spleen 0.36 ± 0.12 0.34 ± 0.06 0.22 ± 0.05 0.17 ± 0.05**
Stomach 5.66 ± 1.49 3.53 ± 0.59* 3.24 ± 1.24* 3.37 ± 0.77*
Testis 0.25 ± 0.06 0.19 ± 0.07 0.16 ± 0.03 0.12 ± 0.01**
Thyroid 1.07 ± 0.88 0.80 ± 0.47 0.47 ± 0.38 0.28 ± 0.23*
Statistical significance: (*) p < 0.05, (**) p < 0.01, (***) p < 0.001 when compared with control.
Table 2 Organs in which CS unchanged (%ID/g) Na99mTcO4 biodistribution (mean ± SD).
CS effects of Na99mTcO4 biodistribution in mice
Organs Groups
Control Mean ± SD CS1d Mean ± SD CS5d Mean ± SD CS10d Mean ± SD
Brain 0.11 ± 0.02 0.14 ± 0.04 0.11 ± 0.01 0.10 ± 0.03
Heart 0.53 ± 0.24 0.61 ± 0.17 0.35 ± 0.12 0.48 ± 0.18
Intestine 0.91 ± 0.52 0.47 ± 0.21 0.72 ± 0.18 0.55 ± 0.39
Liver 2.32 ± 1.30 1.60 ± 0.21 2.07 ± 0.89 1.20 ± 0.37
Muscle 0.40 ± 0.35 0.14 ± 0.06 0.24 ± 0.21 0.09 ± 0.03
Pancreas 0.33 ± 0.28 0.16 ± 0.07 0.11 ± 0.03 0.13 ± 0.05
Discussion
Cigarette smoke is a complex mixture of chemicals containing more than 4000 different constituents. Some of the compounds identified include pyridine alkaloids, such as nicotine, as well as ammonia, acrolein, phenols, acetaldehyde, N-nitrosamine; aromatic hydrocarbons (such as benzopyrene), combustion gases (e.g. carbon monoxide, nitrogen oxides, and hydrogen cyanide), trace metals, α-emitter radioactive elements such as polonium, radium, and thorium. Most of these compounds are produced by pyrolysis and distillation in the zone immediately behind the lighted tip of a cigarette, where the temperature reaches 950°C [18,19]. It has been estimated that undiluted mainstream smoke contains as much as 30 mg of tar and 5 mg of nicotine in an unfiltered regular tar cigarette or 0.5 mg tar and 0.05 mg nicotine in a filtered low-tar cigarette [20].
Radionuclides have been used to investigate diseases related to smoking [21,22]. In a recent study, Iwado et al. [23] demonstrated a decrease in the vasomotor response by PET (positron emission tomography) in the endothelium, in smokers. Moreover, acute CS can diminish cerebral blood flow, as was shown with technetium-99m-labelled ethylcysteine in single-photon emission tomography (SPET) on 10 healthy human volunteers [24]. However, there were no changes in biodistribution in the brain in our study. However, the blood-brain barrier prevents Na99mTcO4 from entering normal brain, and the radioactivity detected may represent a minimal amount of Na99mTcO4 from blood in brain circulation not cleared with our perfusion protocol.
Our experimental model showed that CS could affect %ID/g of Na99mTcO4 distribution in certain organs but not others. These changes vary according to the time of exposure; CS10d organs had lower %ID/g when compared to CS5d and CS1d, except for red blood cells, where there was lower %ID/g in CS5d. In this study, CS affected Na99mTcO4 %ID/g in red blood cells, bone, kidney, lung, spleen, stomach, testis and thyroid. It is noteworthy that some of these organs have also been associated with CS-induced cancer in humans [25].
Cigarette smoke can interfere with the ability of bone cells to participate in repairing and remodelling events [26]. This could be one of the mechanisms leading to the development of osteoporosis. Pathological analysis of the kidney by Cigremes et al. [27] has shown severe degeneration of this tissue with advanced hydrophobic degeneration of kidney tubules in the CS exposed animals.
Kidney radioactivity decreased with CS exposure as a function of time. There are two different effects of CS; first, less availability of Na99mTcO4 due to diminished blood flow to the kidney; and the second is increased permeability or secretion of Na99mTcO4. Further studies to determine the levels of Na99mTcO4 in the urine following CS exposure may conclude whether the decrease in radioactivity is as a result of one of these above changes.
Significant dependence of alveolar deposition on flow rate, but not lung function, was found in young non-symptomatic cigarette smokers and in older non-smokers with 99mTc-polystyrene particles [28]. This could explain the significant decrease of Na99mTcO4 %ID/g in mouse lung at CS10d. Lungs are the main target of cigarette smoke [17], and a change in biodistribution of Na99mTcO4 in this experimental design was anticipated. Sopori et al. [29] have reported results on spleen cells from animals subjected to heavy doses of cigarette smoke which show a significant reduction in the natural killer cell-mediated proteolysis activity and a decreased response to concanavalin A.
There is a higher incidence of gastric cancer in smokers than non-smokers and cigarette smoking has also a strong association with colon cancer [30]. The uptake of Na99mTcO4 was significantly reduced in the stomach in our study, although not in the intestine. Technetium99m is secreted via the stomach mucous and transferred into the intestine, which leads to an anticipated reduction of its radioactivity in intestine. Non-significant reduction of 99mTc activity in the intestine is probably due to a great variability in radioactivity measured in this organ among animals.
Rajpurkar et al. [31] analysed chronic cigarette smoke on rat testis, and observed apoptosis, which may be one of the pathogenic mechanisms responsible for defective spermatogenesis in these animals, Fisher et al. [32] showed that chronic smokers have higher thyroxin levels and a lower thyroid-stimulating hormone level than non-smokers or former smokers. These results corroborate the effects on specific organs by CS, which validates this experimental design as a method to seek alterations of Na99mTcO4 biodistribution in mice.
The knowledge of radionuclide biodistribution due to chemical agents will help physicians avoid misinterpreting images and, thus from making incorrect diagnoses. Such information also helps to prevent or treat possible adverse reactions to radionuclides [15]. Therefore, developing models to study the interactions of chemical agents and radiopharmaceuticals will be of considerable help in understanding this problem. CS has many compounds and smoking habits vary in the population; a direct association between the CS effects and the biodistribution of 99mTc has not previously been reported. However, we cannot comment on how CS changes the biodistribution in other radionuclides, radiochemicals and radiopharmaceuticals. The use of laboratory animals under controlled experimental conditions provides a good model to test the hypothesis that CS can change their biodistribution.
When comparing the sensitivity of CS in BALB/c mice to Na99mTcO4 biodistribution, we cannot exclude the possibility that C57BL6 mice will give better results in the future. When Na99mTcO4 is used as a powerful diagnostic tool in nuclear medicine for examining patients to assess the brain and thyroid, CS can cause alterations in the readings.
Conclusion
CS induced changes of Na99mTcO4 biodistribution in exposed BALB/c mice. Na99mTcO4 distribution on kidney, lung, spleen, testis and thyroid decreased progressively with CS time exposure from 1 to 10 days. The Na99mTcO4 distribution in stomach and red blood cells was affected from the first day of CS exposure.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SSV – carried out the experimental procedure with the mouse exposed to CS, performed the statistical analysis and write the manuscript.
EACL – carried out the experimental procedure with Na99mTcO4 (7.4 MBq) via the ocular plexus.
GFD – determined the radioactivity of the Na99mTcO4 in a counter NaI (TI)
MB-F – conceived of the study and participated in its design and coordination.
LCP – 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
This research was supported by Fundação de Amparo a Pesquisa do Estado do Rio de Janeiro (FAPERJ), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
==== Refs
Rahman I MacNee W Oxidant/antioxidant imbalance in smokers and chronic obstructive pulmonary disease Thorax 1996 51 348 350 8733482
Peto R Lopez AD Boreham J Thun M Heath CJr Mortality from tobacco in developed countries: indirect estimation from national vital statistics Lancet 1992 339 1268 1278 1349675 10.1016/0140-6736(92)91600-D
World Health Organization Mortality Statistics, World Health Statistics Annual 1997–1999
Hom RK Katzenellenbogen JA Technetium-99m labeled Receptor-specific small-molecule radiopharmaceuticals: recent developments and encouraging results Nucl Med Biol 1997 24 485 498 9316075 10.1016/S0969-8051(97)00066-8
Hladik J Snozzi M Bachofen R Incorporation of reaction centers into submitochondrial particles resulting in light induced electron transfer Biochem Biophys Res Commun 1987 14 170 177 2823802 10.1016/0006-291X(87)91091-6
Early PJ Sodee DB Principles and practice of nuclear medicine 1995 MosebyYear Book: London
Saha GB Go RT MacIntyre WJ Radiopharmaceucticals for cardiovascular imaging Int J Rad Appl Instrum B 1992 19 1 20 1577608
Hesslewood S Leung E Drug interactions with radiopharmaceuticals Eur J Nucl Med 1994 21 348 356 8005160 10.1007/BF00947972
Sampson CB Adverse reactions and drug interactions with radiopharmaceuticals Drug Safety 1993 8 280 294 8481215
Lentle BC Scott JR Iatrogenic alterations in radionuclide biodistribution Semin Nucl Med 1979 9 131 143 482949
Baughman RP Fernandez M Radionuclide imaging in interstitial lung disease Curr Opin Pulm Med 1996 2 376 379 9363171
Shih WJ Lee JK Coupal J Lai YL Grunwald F Biersack H Diminished Tc-99m HMPAO pulmonary uptake in ex-smokers Clin Nucl Med 1995 20 788 791 8521654
Capriles PVSZ Dias APM Costa TEMM Oliveira MBN Faria MVC Moura EG Abreu BAL Bernardo-Filho M Effect of eggplant (solanum melongena) extract on the in vitro labeling of blood elements with technetium-99m and on the biodistribution of sodium pertechnetate in rats Cell Mol Biol (Noisy-le-grand) 2002 48 771 776 12619974
Feliciano GD Lima EAC Pereira MJDS Oliveira MBS Moreno SRF Mattos DMM Jales RL Bernardo-Filho M Effect of a chayotte (sechium edule) extract on the labeling of red blood cells and plasma proteins with technetium-99m: in vitro and in vivo studies Cell Mol Biol (Noisy-le-grand) 2002 48 751 755 12619970
Gomes ML Braga ACS Mattos DMM Freitas RS Boasquevisque E Bernardo-Filho M The effect of mitomycin-C on the biodistribution of 99Tcm-MDP in Balb/c mice Nucl Med Commun 1998 19 1177 1179 9885808
Gomes ML Mattos DM Freitas RS Dire GF Lima EA Souza SM Bernardo-Filho M Evaluation of the effect of mitomycin-C on the bioavailability of technetium-99m-labelled sodium pyrophosphate in mice Cell Mol Biol (Noisy-le-grand) 2002 48 757 759 12619971
Valenca SS da Hora K Castro P Moraes VG Carvalho L Porto LC Emphysema and metalloelastase expression in mouse lung induced by cigarette smoke Toxicol Pathol 2004 32 351 356 15204978 10.1080/01926230490431466
Rodgman A Smith CJ Perfetti TA The composition of cigarette smoke: a retrospective, with emphasis on polycyclic components Hum Exp Toxicol 2000 19 573 595 11211997 10.1191/096032700701546514
Skwarzec B Ulatowski J Struminska DI Borylo A Inhalation of 210Po and 210Pb from cigarette smoking in Poland J Environ Radioact 2001 57 221 230 11720371 10.1016/S0265-931X(01)00018-2
Hoffmann D Hoffmann I The changing cigarette, 1950–1995 J Toxicol Environ Health 1997 50 307 364 9120872 10.1080/009841097160393
Meeder JG Blanksma PK van der Wall EE Anthonio RL Willemsen AT Pruim J Vaalburg W Lie KI Long-term cigarette smoking is associated with increased myocardial perfusion heterogeneity assessed by positron emission tomography Eur J Nucl Med 1996 23 1442 1447 8854839 10.1007/BF01254465
Shih WJ Gruenwald F Biersack HJ Berger R Brandenburg S Coupal J Ryo UY Tc-99m HMPAO diffuse pulmonary uptake demonstrated in cigarette smokers Clin Nucl Med 1991 16 668 672 1934830
Iwado Y Yoshinaga K Furuyama H Ito Y Noriyasu K Katoh C Kuge Y Tsukamoto E Tamaki N Decreased endothelium-dependent coronary vasomotion in healthy young smokers Eur J Nucl Med Mol Imaging 2002 29 984 990 12173010 10.1007/s00259-002-0818-1
Yamamoto Y Nishiyama Y Monden T Satoh K Ohkawa M A study of the acute effect of smoking on cerebral blood flow using (99m)Tc-ECD SPET Eur J Nucl Med Mol Imaging 2003 30 612 614 12589479
De Flora S D'Agostini F Balansky R Camoirano A Bennicelli C Bagnasco M Cartiglia C Tampa E Longobardi MG Lubet RA Izzotti A Modulation of cigarette smoke-related end-points in mutagenesis and carcinogenesis Mutat Res 2003 523–524 237 252
Liu X Kohyama T Kobayashi T Abe S Kim HJ Reed EC Rennard SI Cigarette smoke extract inhibits chemotaxis and collagen gel contraction mediated by human bone marrow osteoprogenitor cells and osteoblast-like cells Osteoporos Int 2003 14 235 242 12730796
Cigremis Y Turkoz Y Akgoz M Sozmen M The effects of chronic exposure to ethanol and cigarette smoke on the level of reduced glutathione and malondialdehyde in rat kidney Urol Res 2004 32 213 218 15205856 10.1007/s00240-004-0406-x
Agnew JE Pavia D Clarke SW Factors affecting the alveolar deposition of 5 microns inhaled particles in healthy subjects Clin Phys Physiol Meas 1985 6 27 36 3987208 10.1088/0143-0815/6/1/003
Sopori ML Gairola CC DeLucia AJ Bryant LR Cherian S Immune responsiveness of monkeys exposed chronically to cigarette smoke Clin Immunol Immunopathol 1985 36 338 344 4017293 10.1016/0090-1229(85)90054-6
Shin VY Wang HY Liu ES Koo MW Cho CH Differential effects of cigarette smoke extracts on cell proliferation in gastric and colon cells Cancer Invest 2003 21 200 207 12743985 10.1081/CNV-120016416
Rajpurkar A Jiang Y Dhabuwala CB Dunbar JC Li H Cigarette smoking induces apoptosis in rat testis J Environ Pathol Toxicol Oncol 2002 21 243 248 12435077
Fisher CL Mannino DM Herman WH Frumkin H Cigarette smoking and thyroid hormone levels in males Int J Epidemiol 1997 26 972 927 9363517 10.1093/ije/26.5.972
| 15823206 | PMC1090589 | CC BY | 2021-01-04 16:30:51 | no | BMC Nucl Med. 2005 Apr 11; 5:1 | utf-8 | BMC Nucl Med | 2,005 | 10.1186/1471-2385-5-1 | oa_comm |
==== Front
BMC Nucl MedBMC Nuclear Medicine1471-2385BioMed Central London 1471-2385-5-21583109810.1186/1471-2385-5-2Research ArticleComparison of SPECT bone scintigraphy with MRI for diagnosis of meniscal tears Tahmasebi Mohammad-naghi [email protected] Mohsen [email protected] Masoud [email protected] Ali [email protected] MD, Orthopedic Department, Shariati hospital, Tehran University of medical sciences, Northern Kargar St. 14114, Tehran, Iran2 MD, Nuclear Medicine Research Center, Shariati hospital, Tehran University of medical sciences, Northern Kargar St. 14114, Tehran, Iran2005 14 4 2005 5 2 2 2 1 2005 14 4 2005 Copyright © 2005 Tahmasebi et al; licensee BioMed Central Ltd.2005Tahmasebi 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
Scintigraphy has been considered as competitive to MRI, but limited data are available on the accuracy of single photon emission tomography (SPECT) compared with MRI for the assessment of meniscal tears. Our objective was to assess the value of SPECT in comparison to MRI.
Methods
Between January 2003 and March 2004, sixteen patients were studied with both modalities and the accuracy rates of SPECT scan results, and MRI findings in the diagnosis of meniscal tears were compared. Arthroscopy was the gold standard.
Results
The respective sensitivity rate, specificity rate, and positive and negative predictive accuracies of MRI were 89%, 94%, 93%, and 79% and for SPECT those were 78%, 94%, 94%, and 88%. There was good agreement on the presence or absence of tears between two modalities (κ statistic = 0.699).
Conclusion
SPECT and MRI are both valuable imaging techniques. SPECT is a useful alternative when MRI is unavailable or unsuitable and it is beneficial when more possible accuracy is desired (such as when MRI results are either inconclusive or conflict with other clinical data).
==== Body
Background
In diagnosing meniscal tears, MRI is a sensitive and specific tool and has become the procedure of choice in these affections. In fact, MRI is the most commonly used noninvasive imaging method for diagnosing meniscal tears, but its limits are also acknowledged [1].
On the other hand, although it is not widely carried out in clinical practice, nuclear medicine procedures have also been used in diagnosing meniscal tears and some authors have demonstrated the usefulness of SPECT in the assessment of knee injuries [2-8].
In particular, recent data [9] have demonstrated a higher specificity and accuracy for 99mTc MDP SPECT than those for MRI. As a result, scintigraphy has been considered as competitive to MRI, and also it could give complementary informations which are commonly derived from scintigraphy but not available from MRI.
There are not so much published works concerning meniscal tears and bone SPECT [10-14]. Only few studies have compared bone scintigraphy and MR imaging and limited data are available on the accuracy of SPECT compared with that of MRI for the assessment of meniscal tears. Therefore, it was emphasized that further work should be undertaken to evaluate the role of SPECT as a screening test for the evaluation of knee disorders.
This study reports the results of a recent prospective evaluation of MRI and SPECT bone scintigraphy and compares them with the results of arthroscopy as the gold standard test for the diagnosis of meniscal tears.
Methods
Patients
Between January 2003 and March 2004 sixteen consecutive patients (13 men, 3 women), aged 15–52 yr (31 ± 10 yr, Mean ± S.D.), who were referred to our orthopedic surgeon were entered in this prospective study. Subjects were selected on the basis of positive history and clinical signs suggestive of meniscal tears. MRI and SPECT bone scintigraphy of both knees were obtained from all subjects. The time interval between the SPECT and MRI examinations was 1 to 3 weeks (mean time interval, 2.4 weeks). None of the patients had trauma or additional invasive therapeutic interventions between the SPECT and MRI scans. Patients also underwent arthroscopy of the affected knees.
MRI
All studies were performed using a scanner (IGE Medical Systems, Signa Herza, Milwaukee, WI) with a 1.5 Tesla magnet. The knees were placed in an extended position with approximately 15° of external rotation. The imaging protocol included sagittal multiecho (repetition time msec/echo time msec, 2,500–3,600/20–120), coronal T1-weighted (600/12), coronal multiecho (2,500–3,000/17–119), and transverse gradient-echo or turbo T2-weighted sequences with a slice thickness of 4.5 mm, no interslice gap, and a matrix of 256 × 256. MRI results were reported by a radiologist experienced in MRI of the knees.
Bone scintigraphy
A commercial MDP preparation (Myoview; Amersham International) was used. The labeling and quality control procedures were performed according to the manufacturer's instructions.
Scans were performed on a Vertex dual head ADAC camera. All patients received 750 MBq (20 mCi) 99mTc methylene diphosphonate (99mTc-MDP) by injection, and 3 hours later, anterior, posterior, medial and lateral views of both knees were obtained. SPECT was performed after securing the knees with a band around the tibiae and the legs straightened, with the same dual-head gamma camera, equipped with a pair of low energy, high resolution collimators. Images were acquired in a 128 × 128 matrix at 64 steps, 40 s each step. Data were processed by back projection and filtered by Hanning 0.8 filter. Images were reconstructed and displayed in all three orthogonal planes. Two experienced nuclear medicine physicians familiar with knee SPECT scans interpreted the findings of knee SPECT and the final diagnoses were reached by consensus. Blinded to other informations both MRI and bone SPECT were reported as definite or probable meniscal tear (positive) and normal or non-specific (negative). The positive criterion for meniscal tears was tibial plateau activity on the planar image with at least a half crescent of peripheral tibial plateau uptake on SPECT[12]. Other abnormal scintigraphic patterns were considered as non-specific. The positive criteria for meniscal tears in the MRI were abnormal morphology of the meniscus on one or more MR images and/or abnormal increased signal in that area on fat saturated proton density or T2-weighted images. Arthroscopy was performed by our experienced arthroscopist who already knows the results of the MRI and SPECT bone scans at the time of the examination.
Statistical evaluation
Results were analyzed on a per-meniscus basis. Using arthroscopy as a gold standard, the results of each modality were analyzed for sensitivity, specificity, negative predictive value and positive predictive value. Differences between these performance indices in the two modalities were evaluated with the McNemar test. P < 0.05 was considered to be statistically significant.
All patients gave informed consent to participate in this study, which was approved by the committee on ethics at the faculty of medicine, university of Tehran.
Results
A total of 32 menisci, including 16 left and 16 right menisci, in 16 patients (table 1) were assessed. According to the arthroscopic results, tears were present in 18 (56%) menisci (table 2), of these thirteen tears were in the medial menisci, five tears were in the lateral menisci, ten tears were in the left menisci and eight tears were in the right menisci.
Table 1 Patients' characteristics and results of MRI, SPECT and Arthroscopy.
Patient no. Sex Age MRI SPECT Arthroscopy
1 M 34 LM LM LM
2 M 15 NL NL NL
3 M 19 RM RM RM
4 F 52 LM LM/LL LM/LL
5 M 45 RM RM RM
6 M 35 LM LM LM
7 M 30 LL LL LL
8 F 50 RM D.I.U RM
9 M 26 RM/RL RL RM/RL
10 M 24 RM RM RM
11 M 26 LM LM/LL LM
12 M 28 RM RM RM
13 M 25 LL NL LM
14 M 30 LM/LL LM LM/LL
15 F 32 RL RL RL
16 M 21 LM LM LM
* M = Male, F = Female, D. I. U = Diffusely Increased Uptake, RL = Right Lateral, RM = Right Medial, LL = Left Lateral, LM = Left Medial, NL = Normal.
Table 2 Arthroscopy results by meniscus.
Meniscus Positive Negative Total
Left/Medial 7 1 8
Left/Lateral 3 5 8
Right/Medial 6 2 8
Right/Lateral 2 6 8
Total 18 14 32
By consensus, observers detected 15 meniscal tears at SPECT readings, of which one was falsely positive (table 3). One knee showed generalized increased uptake on bone SPECT images, in which the exact anatomical location of the pathologic process could not be determined and this finding was categorized as false-negative for meniscal tear. Overall there were only four false-negative SPECT scans.
Table 3 Bone SPECT readings by meniscus.
Meniscus Positive Negative Total
Left 9 7 16
Right 6 10 16
Total 15 17 32
MRI detected 17 tears (table 4). However one of them was false positive. Two meniscal tears were missed by MRI. Among the tears detected on SPECT images, only one tear was not detected by MRI. On the other hand, three meniscal tears were not depicted on SPECT images, which were detectable on MR images.
Table 4 MRI reading by meniscus.
Meniscus Positive Negative Total
Left 9 7 16
Right 8 8 16
Total 17 15 32
Overall, MRI had a positive predictive value of 94% (16/17) and SPECT had 93% (14/15). MRI had a negative predictive value of 88% (15/17) and SPECT had 79% (15/19).
Table 5 shows the sensitivities, specificities, positive predictive values, and negative predictive values for each imaging modality. The two-tailed p value of these differences equals 0.683, which by conventional criteria, is considered to be not statistically significant. MRI and SPECT results were further compared on another per meniscus basis according to whether SPECT findings were positive or negative (table 6). In a total of 17 knees which underwent imaging with both modalities, meniscal tear was found in 17 menisci with MRI and in 15 menisci with SPECT. Concordant positive results were reported in 13 menisci. In four menisci, MRI depicted additional tears. The κ value, as a measure for agreement between SPECT and MRI, revealed that despite differences between methods in sensitivity and specificity for the detection of meniscal tears, there was still good overall agreement (κ statistic = 0.699, with standard error = 0.101).
Table 5 A comparison of diagnostic ability of SPECT and MRI in diagnosis of meniscal tears.
MRI SPECT
Sensitivity 89% 78%
Specificity 94% 94%
Positive predictive value 93% 94%
Negative predictive value 79% 88%
Table 6 Summary of knee SPECT and MRI results.
Test Results No. of menisci
Positive SPECT vs positive MRI 13
Positive SPECT vs Negative MRI 2
Negative SPECT vs Positive MRI 4
Negative SPECT vs Negative MRI 13
Conclusion
MRI has become the radiologic procedure of choice for the diagnosis of meniscal tears. Recently SPECT also has been used for assessment of knee pathologies and has been documented to have a higher sensitivity than MRI. One of the most widely referenced studies is that of Ryan PJ et al. [9], in which 100 patients with undiagnosed knee pain were studied by clinical examination, MRI, SPECT bone scintigraphy and arthroscopy. The authors found the accuracy of MRI and SPECT in detecting meniscal tears to be comparable. Using arthroscopy as a gold standard, both MRI and SPECT showed high diagnostic ability to detect meniscal tears, with respective sensitivity rate, specificity rate, and positive and negative predictive accuracies of 80%, 71%, 84% and 71% for MRI and 84%, 80%, 88% and 76% for SPECT. Some meniscal tears were detected by MRI alone (n = 5), or SPECT alone (n = 8). These authors concluded that SPECT bone scintigraphy is a suitable alternative to MRI to detect meniscal tears. It was also noted that the comparable diagnostic ability of SPECT bone scintigraphy implies that it can be used successfully when MRI is unavailable or unsuitable.
Even-Sapir and colleagues reported on 94 patients with suspected ACL/meniscal tear, or both who underwent SPECT followed by arthroscopy (n = 74), magnetic resonance imaging (n = 37), or both [15]. Tears of the medial meniscus were diagnosed by arthroscopy in 43 patients. SPECT images detected increased uptake in the medial tibial plateau with a positive predictive value of 78% and a negative predictive value of 83%. These authors suggest that bone SPECT is valuable in acute knee trauma for assessment of ACL, meniscal tears or both and for detection of associated bone injury.
In another study of patients with chronic knee pain Collier et al., found a high sensitivity of SPECT for the detection of meniscal tears, although specificity was less good [16].
In one of the initial studies, Murray et al. [11] found a SPECT sensitivity of 88% and specificity of 87% in patients with acute knee pain. They concluded that with respect to meniscal tears a negative bone scan can obviate the need for arthroscopy.
In the recent study done by Vellala RP et al. [17] the role of SPECT bone scan for the diagnosis of knee lesions in routine clinical practice was evaluated. Fourty consecutive case records were examined in patients who underwent a SPECT scan prior to knee arthroscopy in routine clinical practice. The accuracy of clinical examination, SPECT scan results, and arthroscopic findings (as the gold standard) in diagnosing knee lesions were compared. The sensitivity of SPECT scans in detecting medial meniscal and lateral meniscal lesions was 77% and 14%, respectively. The specificities for the same structural lesions were high at 89% and 94%, respectively. The authors concluded that SPECT bone scan appears to be useful in the diagnosis of knee pathology in routine practice and in selecting patients for arthroscopy, especially most useful for the diagnosis of medial meniscal tears.
Our results are not very different from the above-mentioned researches. The present prospective study demonstrated that MRI was only slightly superior to SPECT for detection of meniscal tears. However, the difference did not reach statistical significance. In our study SPECT revealed the majority of lesions seen on arthroscopy and MRI. Also there was a tear that was missed by MRI but adequately diagnostic by SPECT. Similarly there were some tears in which SPECT was negative but MRI showed the tears. It is not yet well determined that why some tears are missed by one modality and are detected by the other. However, it seems that in the presence of high clinical suspicion and negative MRI results (as the primary modality), SPECT can be helpful with detecting MRI negative tears. Despite of this fact, the detection of more tears at SPECT compared with MRI, did not lead to altered decision for treatment (i.e. conservative vs. arthroscopic treatment). The major reason is that the difference in detection of lesion was predominantly in patients in whom another MRI-positive tear was detected. These patients are usually candidate for arthroscopy and therefore differences in the number of tears do not influence management.
In general, the findings of our study and that of previous studies suggest that examination with SPECT as well as with MRI can be used as a basis for the assessment of patients suspected for meniscal tears. However, both MRI & SPECT have various relative advantages and disadvantages. In general, SPECT is less costly than MRI because it involves lower capital equipment costs. SPECT is also widely available. The major limitation with the use of SPECT is the radiation exposure, the potential harm of which is poorly understood.
MRI also has some advantages over SPECT of the knees too. Most significantly, no ionizing radiation is used. MRI also has some important drawbacks, however. In most regions it is a more expensive than SPECT and has more contraindications and scheduling difficulties. Some authors concluded that MRI, except in certain circumstances, is an expensive and unnecessary diagnostic test in patients with suspected meniscal and ACL pathology (may be due to many false positive MRI reports) [18,19].
These facts in addition to all of the above-mentioned research results indicate that SPECT and MRI are both valuable advanced imaging techniques but the absence of radiation exposure may make MRI preferable for the workup of patients suspected of having meniscal tears. Therefore it seems that all patients suspected for meniscal tears can be evaluated with MRI. However, SPECT has clear advantages when more possible accuracy is desired when MRI results are either inconclusive or conflict with other clinical data (i.e. SPECT should be performed if MRI is negative but there are clinical evidences of meniscal tear). SPECT may be available alternative when MRI is unavailable or unsuitable. This approach must be addressed in larger series of patients and a larger prospective study is currently being performed to confirm these data and approach.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MNT participated in the design of the study and carried out the arthroscopies. MS participated in the interpretation of the scintigraphic results. MM participated in its design and coordination, supervised the acquisition process and participated in the interpretation of the scintigraphic results. AG supervised the acquisition process, interpreted the scintigraphic results, performed the statistical analysis and drafted the manuscript. All authors read and approved the final manuscript.
Figure 1 Magnified coronal T2-weighted MRI image of the right knee showing a medial meniscal tear.
Figure 2 SPECT images of the same patient (presented in Figure 1.) showing a crescent of increased activity in the medial tibial plateau, which is scintigraphically characteristic feature of a meniscal tear.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was carried out with the sponsorship of Tehran University of medical sciences. We are indebted to Dr. Mohammad Eftekhari, Dr. Armaghan Fard Esfahani, Dr. Davoud Beigi, and Dr. Babak Fallahi (Nuclear Medicine research center, Shariati Hospital) and Dr. Ali Radmehr (Department of radiology, Shariati Hospital) for their consultations throughout the investigation and their suggestions. Thanks are also extended to the technologists at our hospital (especially Mr. S. M. A. Abdollahzadeh) for data acquisition.
==== Refs
Rath E Richmond JC The menisci: basic science and advances in treatment Br J Sports Med 2000 34 252 7 10953895 10.1136/bjsm.34.4.252
Lorberboym M Ami DB Zin D Nikolov G Adar E Incremental diagnostic value of 99mTc methylene diphosphonate bone SPECT in patients with patellofemoral pain disorders Nucl Med Commun 2003 24 403 10 12673169 10.1097/00006231-200304000-00010
Yildirim M Gursoy R Varoglu E Oztasyonar Y Cogalgil S 99mTc-MDP bone SPECT in evaluation of the knee in asymptomatic soccer players Br J Sports Med 2004 38 15 8 14751939 10.1136/bjsm.2002.000695
Gupta SM Foster CR Kayani N Usefulness of SPECT in the early detection of avascular necrosis of the knees Clin Nucl Med 1987 12 99 102 3493874
Chung HW Kim YH Hong SH Kim SS Chung JK Seong SC Kang HS Indirect signs of anterior cruciate ligament injury on SPET: comparison with MRI and arthroscopy Nucl Med Commun 2000 21 651 8 10994669 10.1097/00006231-200007000-00009
Cook GJ Fogelman I Lateral collateral ligament tear of the knee: appearances on bone scintigraphy with single-photon emission tomography Eur J Nucl Med 1996 23 720 2 8662108 10.1007/BF00834536
Cook GJ Ryan PJ Clarke SE Fogelman I SPECT bone scintigraphy of anterior cruciate ligament injury J Nucl Med 1996 37 1353 6 8708771
al-Janabi MA The role of bone scintigraphy and other imaging modalities in knee pain Nucl Med Commun 1994 15 991 6 7715899
Ryan PJ Reddy K Fleetcroft J A prospective comparison of clinical examination, MRI, bone SPECT, and arthroscopy to detect meniscal tears Clin Nucl Med 1998 23 803 6 9858289 10.1097/00003072-199812000-00002
Grevitt MP Taylor M Churchill M Allen P Ryan PJ Fogelman I SPECT imaging in the diagnosis of meniscal tears J R Soc Med 1993 86 639 41 8258798
Murray IP Dixon J Kohan L SPECT for acute knee pain Clin Nucl Med 1990 15 828 40 2292159
Ryan PJ Bone SPECT of the knees Nucl Med Commun 2000 21 877 85 11130328 10.1097/00006231-200010000-00001
So Y Chung JK Seong SC Sohn YJ Kang HS Lee DS Lee MC Usefulness of 99Tcm-MDP knee SPET for pre-arthroscopic evaluation of patients with internal derangements of the knee Nucl Med Commun 2000 21 103 9 10717910 10.1097/00006231-200001000-00017
Ryan PJ Chauduri R Bingham J Fogelman I A comparison of MRI and bone SPET in the diagnosis of knee pathology Nucl Med Commun 1996 17 125 31 8778636
Even-Sapir E Arbel R Lerman H Flusser G Livshitz G Halperin N Bone injury associated with anterior cruciate ligament and meniscal tears: assessment with bone single photon emission computed tomography Invest Radiol 2002 37 521 7 12218448 10.1097/00004424-200209000-00007
Collier BD Johnson RP Carrera GF Isitman AT Veluvolu P Knobel J Hellman RS Barthelemy CR Chronic knee pain assessed by SPECT: comparison with other modalities Radiology 1985 157 795 802 3877315
Vellala RP Manjure S Ryan PJ Single photon emission computed tomography scanning in the diagnosis of knee pathology J Orthop Surg (Hong Kong) 2004 12 87 90 15237128
Rose NE Gold SM A comparison of accuracy between clinical examination and magnetic resonance imaging in the diagnosis of meniscal and anterior cruciate ligament tears Arthroscopy 1996 12 398 405 8863996
Terry GC Tagert BE Young MJ Reliability of the clinical assessment in predicting the cause of internal derangement of the knee Arthroscopy 1995 11 568 76 8534299
| 15831098 | PMC1090590 | CC BY | 2021-01-04 16:30:51 | no | BMC Nucl Med. 2005 Apr 14; 5:2 | utf-8 | BMC Nucl Med | 2,005 | 10.1186/1471-2385-5-2 | oa_comm |
==== Front
BMC OphthalmolBMC Ophthalmology1471-2415BioMed Central London 1471-2415-5-81581397210.1186/1471-2415-5-8Technical AdvanceAutomated analysis of digital fundus autofluorescence images of geographic atrophy in advanced age-related macular degeneration using confocal scanning laser ophthalmoscopy (cSLO) Deckert A [email protected] S [email protected] J [email protected] A [email protected] FG [email protected] U [email protected] Institute of Medical Biometry and Informatics (IMBI), University of Heidelberg, Im Neuenheimer Feld 305, 69120 Heidelberg, Germany2 Department of Ophthalmology, University Eye Hospital, Abbe-Straß2, 53127 Bonn, Germany3 Department of Ophthalmology, University Eye Hospital, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany4 Institute of Medical Information, Biometry, and Epidemiology (IBE), University of Munich, Marchioninistrasse 15, 81377 Munich, Germany2005 6 4 2005 5 8 8 28 10 2004 6 4 2005 Copyright © 2005 Deckert 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
Fundus autofluorescence (AF) imaging using confocal scanning laser ophthalmoscopy (cSLO) provides an accurate delineation of areas of geographic atrophy (GA).
Automated computer-assisted methods for detecting and removing interfering vessels are needed to support the GA quantification process in longitudinal studies and in reading centres.
Methods
A test tool was implemented that uses region-growing techniques to segment GA areas. An algorithm for illuminating shadows can be used to process low-quality images. Agreement between observers and between three different methods was evaluated by two independent readers in a pilot study. Agreement and objectivity were assessed using the Bland-Altman approach.
Results
The new method (C) identifies vascular structures that interfere with the delineation of GA. Results are comparable to those of two commonly used procedures (A, B), with a mean difference between C and A of -0.67 mm2 (95% CI [-0.99, -0.36]), between B and A of -0.81 mm2, (95% CI [-1.08, -0.53]), and between C and B of 0.15 mm2 (95% CI [-0.12, 0.41]). Objectivity of a method is quantified by the mean difference between observers: A 0.30 mm2 (95% CI [0.02, 0.57]), B -0.11 mm2 (95% CI [-0.28, 0.10]), and C 0.12 mm2 (95% CI [0.02, 0.22]).
Conclusion
The novel procedure is comparable with regard to objectivity and inter-reader agreement to established methods of quantifying GA. It considerably speeds up the lengthy measurement process in AF with well defined GA zones.
==== Body
Background
Age-related macular degeneration (AMD) is the most common cause of legal blindness among industrialized nations in the population aged 50 years and above [1-4]. Besides choroidal neovascularization and detachments of the retinal pigment epithelium, geographic atrophy (GA) of retinal pigment epithelium (RPE) is a common cause of severe visual loss in patients with AMD [5-7]. Changes in time can be documented by fundus autofluorescence images (AF) mediated by RPE lipofuscin accumulations and its spatial distribution over retinal areas, obtained in vivo using a confocal scanning laser ophthalmosocope (Heidelberg Retina Angiograph (HRA), Dossenheim, Germany) [8-10]. Areas of GA are usually associated with a well-defined zone of decreased autofluorescence due to the absence of fluorophores residing in RPE lipofuscin granules [11,12]. The deduction of clinically relevant information from these pictures is a complex process which should be optimised and automated, especially in the context of multicenter studies.
Generally, AF are recorded using a confocal scanning laser ophthalmoscope (Heidelberg Retina Angiograph, HRA, Heidelberg Engineering, Germany; which includes the Heidelberg Eye Explorer (HEE) software package). The images are immediately digitised and processed using a flexible frame processor and subsequently displayed on a computer screen. Corresponding with funduscopically visible atrophic areas, fundus intensity of AF is markedly decreased [13].
With method A, atrophic areas are outlined on the screen using the mouse-driven cursor of the HEE software program. The areas are then measured and the data exported manually to an Excel spreadsheet by cut and paste. This completely manual, mouse-driven method A is time-consuming and can exaggerate subjective impressions. Mistakes can occur as a result of the error-prone interface between the user's hand and the computer mouse [13]. This implies that the accuracy of mouse-driven contour painting depends not only on subjective impressions, but also on the user's dexterity.
With method B, the images are exported as bitmap files from the HEE program. Interfering vascular structures, which appear as dark atrophic areas, are manually repainted white using the mouse-driven paintbrush of Microsoft Paint. The modified images are then transferred to Global Lab Image 2 and the remaining dark areas are measured using a threshold procedure tool. The resulting data are exported to Microsoft Excel manually by cut and paste [13]. This semi-automated method B only allows an interpretation using high-quality AF images. It requires the manual, mouse-driven whitewashing of interfering vessels that touch upon, or extend into, the GA area. However, as the vessels exhibit grey levels that are similar to atrophic patches, automated segmentation is not possible [13]. Both methods rely on the circumstantial handling of different software tools that have not been specifically adapted to the problems of GA measurements.
We have developed a novel customized image analysis test tool that includes an adapted algorithm for automated identification of interfering vascular structures, and compared this new method with the previous ones.
Methods
The test tool combines all steps of GA area measurement, including the automatic export of data into an Excel spreadsheet for further analysis. The method is based on region-growing, instead of segmentation with a threshold value similar to that of method B, allowing the reader to sort out non-atrophic areas with similarly diminished grey values as in actual GA areas.
After selecting the GA area by moving the mouse cursor and clicking on the region of interest, a first segmentation is started using a default parameter value calculated by the image's mean grey value. Automatic correction of the generated contour is possible by adjusting the associated parameter value with a relocatable graphical element. Holes within the detected GA area can be identified and further GA areas in the same image can be integrated.
For segmentation of GA areas with interfering vascular structures, the tool includes an optional algorithm for detecting and whitewashing such structures [14]. The number of detected vessels depends on the individual settings of the parameters for vessels diameter, length and cross-linking. Incorrect contour segments of the GA caused by whitewashed vessel stubs can reach into the GA area and are corrected, either automatically with a default parameter or by user interaction.
The algorithm eliminates vessel stubs within the GA area and corrects the contour segments in these regions. One disadvantage of this process is that small contour jags which were correctly detected a step before now disappear, and instead the contour is minimally smoothed and widened. As a result the process produces correct contour segments in passages from GA area to vessels and incorrectly widened contour segments otherwise. A new segmentation step produces the correct contour without interfering vessels.
Using well-aligned relocatable convex and radial hulls facilitates fine-tuning of the actual delineation and measurement of GA areas. These tools enable alternative contour finding in difficult contour segments with tangentially interfering vessels. Poor quality images often contain large shady outer areas causing erupting segmentation of GA areas before identifying the true borders of the GA. This can be countered by the method's option for illuminating such areas of shade. The original image is shown in a separate window. An electronic magnifier is integrated to support contour finding in difficult sections as well as for small GA areas.
Validation
Validation is based on the material collected in a previous study [13], with published data from the right eye only. Data for both eyes are available [13] and are used for validation. The same readers evaluated the same material, but using the new method. The images were evaluated in random order and the readers had no access to the previous results based on methods A and B. With the unpublished material from 2001, our validation sample consists of 10 left and 24 right eyes. The Bland-Altman Design for method comparison studies [15] was applied. The limits of agreement (REF) were calculated for each comparison. Nonparametric tests for paired and unpaired observations as well as ANOVA (Analysis of Variance) were used for statistical assessment. The significance level was set to α = 0.05.
The study was a sub-study of the FAM study which followed the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of the University of Heidelberg. Informed consent was obtained from the patients prior to recruitment into the study.
Results
Table 1 shows the results for the 40 eyes included in the study. Method C produced values which lay between those of methods A and B. The quantification of the agreement following Bland and Altman [15] between methods B and A, C and A, and B and C is given in the upper half of table 2 (mean difference, 95% confidence interval, REF). Differences in the quantified area are presented as box-plots in figure 1 and figure 2. Using the Friedman test [16], a global distinction between methods could be shown for each of the readers (reader 1 p < 0.001, reader 2 p < 0.001). The Wilcoxon test gives significant distinctions for reader 2 between method A and B (p < 0.001) and between A and C (p < 0.001). No significant difference could be found between methods B and C (p = 0.467). For reader 1, the results were similar, except that methods B and C also do not differ significantly (A/B p < 0.001; A/C p < 0.001; B/C p = 0.027). The investigation of objectivity by the Bland-Altman design is given in the lower half of table 2 (mean difference, 95% confidence interval, REF). The box plots presented in figure 2 show a slightly larger bias between the readers for method C in comparison with method B and less bias in comparison with method A. Differences between the readers is established for all three methods by the Wilcoxon test (A: p = 0.007; B: p = 0.012; C: p = 0.018). Combining agreement between readers, the Friedman test shows global significant distinctions between all methods (p < 0.001). An analysis of variance provided a deeper insight into the cause of the observed reader effect, which is not homogeneous for all methods. The ANOVA approach establishes an interaction between both effects. The significant interaction term (p < 0.01) demonstrates the clear dependence of a method's result on the reader using it.
Discussion
Comparability, repeatability, and objectivity are crucial factors to consider when developing a new method. Comparability with at least two methods has to be determined; objectivity calls for a multitude of readers; and repeatability requires multiple measurements by the same reader using each method. An appropriate study would randomly allocate image and reader in a cross-over design.
Bland-Altman plots visualize comparability, repeatability, and objectivity in terms of the limits of agreement in which 95% of data should appear. Furthermore, the data within these borders must lie within the region of clinically irrelevant differences. Therewith, verification of a clinically comparable quality can be obtained. The measurements for method comparison have to be performed at the same time to exclude subjective effects caused by changing user criteria rather than by the methods themselves.
Additionally, improved effectiveness must be demonstrated to justify replacement of methods [15]. Effectiveness is defined here as the number of successfully assessed AF images and the time required to complete the assessment procedure. In 2001 [13] no exact measurements of reading duration were taken. However, both readers had the impression that method C speeds up the reading process. Furthermore, low-quality images were excluded from the readings in 2001 [13]. Some of these pictures could be handled with method C. Again both readers agreed that method C was better suited to evaluate low-quality images. A further study has to put these subjective impressions into an appropriate objective framework to demonstrate effectiveness.
We conclude that method C is not inferior to the two commonly methods used in measuring GA areas. This has been shown by the comparability of the measured values in the Bland-Altman design. Thereby, the inter-methods comparisons align with the degree of decision freedom for each method. Method A [13] enables the setting of each contour pixel individually with no relation between contour pixels. Method B [13] has only one degree of decision freedom using one threshold value for the whole contour, and method C has fewer degrees of decision freedom than A, but more than B: it allows for the individual exclusion of non-atrophic holes or a fine-tuning of critical contour-sections. In accordance with the degrees of decision freedom, the mean values of the new method C lie between those of method A and B (data not shown). Altogether, method C is similar more to method B than to method A (see table 2, A-C, B-C). Increasing the number of degrees of decision freedom increases the influence of subjectivity. It also gives a larger impact of user's competence.
Method B has only one degree of decision freedom and should theoretically be the most objective. But the bias in objectivity of method C is, first of all, a result of more than one degree of decision freedom (in comparison with method B). Therefore, the readers were trained in using method C. This made the objectivity of method C not significantly inferior to method B. The remaining bias of method C is redeemed by fewer outliers and less dispersion. Even within method C, there was a clear difference in how the readers interpreted the same AF images (see figure 3 and figure 4).
There are subjective inter-reader effects and effects caused by changing assessment criteria within the same reader over a period of years due to increasing experience or changing criteria, but not due to the method used. This influences the bias in method comparison and therewith the statements about measurement error. Considering this confounder, method comparison can only show that method C is no worse than methods A and B. The effect of the influence by the subjectivity of readers is most distinct in method A (see table 2).
Heterogeneity in agreement between the readers over the methods was demonstrated by a significant interaction term (reader by method) in an ANOVA model. There is a loss of histological information due to the process of producing AF images from the retina by confocal laser scanning ophthalmoscopy. Several underlying factors for alterations in grey values in the AF images, mainly in the border region of GA areas, force the reader to interpret by his subjective impressions, more or less supported by the interaction between him and the method.
Conclusion
The study is a pilot study, yet it has provided important information for the further design and testing of method C. Method C has been developed and launched in consideration of several viewpoints. One aim has been to reduce the influence of manual skills and the number of procedural steps, while another important goal was not to restrict the competence of the medical user.
Method C can be implemented to quickly assess both unproblematic AF images and, with the additional accessory tools, difficult AF images. For example, the presence of shadows within the marginal areas of the images appears to be the main reason for poor quality. Presenting a processed image with illuminated shadows together with the original image facilitates segmentation by different algorithms. Furthermore, robust algorithms have been integrated to rectify segmentation and to eliminate interfering vessels in an effective way. Hence, a large proportion of GAs can be measured in a short time using vessel detection and default parameters.
The individual nature of each AF image and the wide variation of possible combinations of features suggest that it would be impossible to develop a clearly defined method that could measure GA areas with only a few procedural steps. Thus, it seems to be recommendable to represent in the future a combination of several methods, with adaptations depending on the quality of AF images, in one tool. However, in attempting to decide upon a particular combination of method parts, some basic questions arose during the process of developing and testing method C, which have still to be clarified.
The selection of further algorithms and methods will depend on the definition of the clinical relevance of differences between GA areas. Consequently, care should be taken that no clinical relevance is attached to artificial differences due solely to repeated use of a method or to different users. Furthermore, a method comparison using the Bland-Altman design will only be meaningful if there are well-defined limits of agreement and clinically relevant bounds [15]. For a method to be useful, the probability of differences between repeated measures and objectivity must not transgress this bound significantly. If two methods produce absolute differences below this limit with non-significant probabilities, then the method with low dispersion but nearer the bound should be preferred. It should be favoured over a method with less bias but larger dispersion. The reason is that a uniform bias could be eliminated by training or considered in the statistical evaluation. With regard to objectivity, it is also important to know whether high objectivity by one threshold value like in method B limits precise contours and facilitates medically correct detection of GA borders. If the precision in imaging the GA borders is important for quantification of disease progression, a segmentation algorithm like region-growing with more degrees of decision freedom should be used. For the decision, clinical relevance in method comparison should be the main criteria, too.
We conclude with an example of successful segmentation after whitewashing the vessels, presented in figure 5. The first segmentation produced erupted contours (first image, green line) caused in each case by interfering vessels. After automated whitewashing with corrected GA entries of white vessel stubs the accurate GA contour is found (second image, green line). Fine tuning of this contour within the wider red one is possible.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AD has developed the new method, performed the study and composed the manuscript. SS and JJ performed the evaluation of the pictures in the validation study. AB and FH have given support in ophthalmologic issues and have provided patients data and ophthalmologic devices. UM initialised and supervised the project, he also contributed to the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Supported by the Deutsche Forschungsgemeinschaft (DFG) Research Priority Program SPP 1088: MA 1723/1-1 and HO 1926/1-2
Figures and Tables
Figure 1 Box-plot method comparison. Reader adjusted mean difference in the quantified GA area [mm2] between methods: A-B, A-C, B-C.
Figure 2 Box-plot reader comparison. Method induced mean difference [mm2] between two readers.
Figure 3 Example for different interpretation with method C. Reader 1 defines a non-atrophic area within the GA in a second step (filled white area) which will not be considered for the calculation of the area size.
Figure 4 Example for different interpretation with method C. This example shows differences in the perception of what is a GA. Reader 2 declares additionally small GA spots.
Figure 5 Example for image processing. Interfering vessels prevent correct segmentation as shown in the first image. Whitewashing the interfering vessels allows a correct segmentation of the GA.
Table 1 The complete GA measurement data set
method reader 1 (mm2) reader 2 (mm2)
A B C A B C
ID right eyes
7 8,86 8,335 8,737 8,94 8,286 8,711
8 8,08 5,209 6,913 7,01 5,759 6,34
12 10,17 9,42 9,372 9,73 9,368 9,965
13 10,1 9,163 9,462 10,48 9,228 9,174
16 21,58 19,821 21,093 21,44 20,407 20,908
28 2,65 2,352 2,15 2,49 2,29 2,44
29 9,47 8,861 9,163 9,5 9,07 8,672
53 7,17 6,606 6,864 7,04 6,595 6,688
65 11 9,868 10,226 10,82 9,488 10,316
67 2,22 2,029 2,193 2,19 2,029 2,143
68 12,07 10,916 11,218 11,54 10,785 10,874
89 12,23 11,3 10,466 11,1 10,117 10,259
90 4,36 4,095 4,174 4,63 4,339 4,114
96 10,82 8,832 10,312 10,1 10,18 9,947
97 27,1 23,976 23,533 22,95 22,868 23,769
102 23,69 20,433 21,489 22,6 21,023 21,05
109 10,7 9,032 10,153 10,39 9,112 10,04
113 12,01 11,115 11,283 11,67 10,911 11,384
144 3,86 3,555 3,576 3,84 3,653 3,649
146 2,95 2,695 2,718 2,97 2,677 2,764
147 1,8 1,504 1,478 1,92 1,504 1,449
167 2,82 2,396 2,438 3,01 2,396 2,62
176 4,03 3,671 3,936 3,98 3,677 3,869
182 1,68 1,328 1,433 1,69 1,418 1,484
ID left eyes
7 23,54 22,78 22,186 23,38 22,993 22,026
8 15,23 12,86 15,228 14,8 13,644 14,661
12 3,2 2,795 2,968 2,94 2,793 2,662
13 12,78 12,682 12,588 13,4 13,009 12,501
16 24,34 22,687 21,023 24,43 22,677 20,855
29 10,3 9,139 8,65 10,35 9,477 8,201
67 0,66 0,536 0,512 0,6 0,965 0,528
89 11,62 11,156 11,459 11,53 10,954 10,798
99 11,62 10,887 11,259 11,46 10,826 11,001
120 6,23 4,594 4,856 6 5,781 5,15
ID in 2001 excluded eyes
77 6,843 6,783
86 7,665 6,977
92 21,719 21,223
123 7,038 7,005
161 4,67 4,475
186 5,474 5,254
Table 2 Quality measures to assess method comparison
Mean difference (mm2) 95% confidence interval Limits of agreement
Agreement A-B 0.81 [-0.53; 1.08] [-0.76; 2.37]
A-C 0.67 [0.36, 0.99] [-1.13; 2.47]
B-C -0.15 [-0.41; 0.12] [-1.68; 1.38]
Objectivity A 0.30 [0.02; 0.57] [-1.27; 1.86]
B -0.11 [-0.28; 0.07] [-1.10; 0.88]
C 0.12 [0.02; 0.22] [-0.44; 0.68]
==== Refs
Bressler NM Bressler SB Fine SL Age-related macular degeneration Surv Ophthalmol 1988 32 375 413 2457955 10.1016/0039-6257(88)90052-5
Klein R Wang Q Klein BE Moss SE Meuer SM The relationship of age-related maculopathy, cataract and glaucoma to visual acuity Invest Ophthalmol Vis Sci 1995 36 182 191 7822146
Leibowitz HM Krueger DE Maunder LR Milton RC Kini MM Kahn HA Nickerson RJ Pool J Colton TL Ganley JP Loewenstein JI Dawber TR The Framingham Eye Study monograph: an ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a general population of 2631 adults, 1973–1975 Surv Ophthalmol 1980 24 335 610 7444756
Wormald R Assessing the prevalence of eye disease in the community Eye 1995 9 674 6 8849531
Holz FG Wolfensberger TJ Piguet B Gross-Jendroska M Wells JA Minassian DC Chisholm IH Bird AC Bilateral drusen in age-related macular degeneration – prognosis and risk factors Ophthalmology 1994 101 1522 1528 8090455
Maguire P Vine AK Geographic atrophy of the retinal pigment epithelium Am J Ophthalmol 1986 102 621 625 3777083
Sunness JS Bressler NM Tian Y Alexander J Applegate CA Measuring geographic atrophy in advanced age-related macular degeneration Invest Ophththalmol Vis Sci 1999 40 1761 1769
Rückmann AV Fitzke FW Bird AC Distribution of fundus autofluorescence with a scanning laser ophthalmoscope BR J Ophthalmol 1995 79 407 412 7612549
Rückmann AV Fitzke FW Bird AC Fundus autofluorescence in age-related macular disease imaged with a laser scanning ophthalmoscope Invest Ophthalmol Vis Sci 1997 38 478 486 9040481
Solbach U Keilhauer C Knabben H Wolf S Imaging of retinal autofluorescence in patients with age-related macular degeneration Retina 1997 17 385 389 9355185
Holz FG Bellmann C Margaritidis M Schutt F Otto TP Volcker HE Patterns of increased in vivo fundus autofluorescence in the junctional zone of geographic atrophy of the retinal pigment epithelium associated with age-related macular degeneration Graefes Arch Clin Exp Ophthalmol 1999 237 145 152 9987631 10.1007/s004170050209
Rückmann AV Fitzke FW Bird AC In vivo fundus autofluorescence in macular dystrophies Arch Ophthalmol 1997 115 609 615 9152128
Schmitz-Valckenberg S Jorzik J Unnebrink C Holz FG Analysis of digital scanning laser ophthalmoscopy fundus autofluorescence images of geographic atrophy in advanced age-related macular degeneration Graefes Arch Clin Exp Ophthalmol 2002 240 73 78 11933894
Steger C An unbiased detector of curvilinear structures Technical Report FGBV-96-03, Technical University Munich, research team image processing 1996
Bland JM Altman DG Measuring agreement in method comparison studies Stat Methods Med Res 1986 8 135 160 10.1191/096228099673819272
Sachs L Angewandte Statistik 2004 11 Berlin: Springer-Verlag
| 15813972 | PMC1090591 | CC BY | 2021-01-04 16:03:50 | no | BMC Ophthalmol. 2005 Apr 6; 5:8 | utf-8 | BMC Ophthalmol | 2,005 | 10.1186/1471-2415-5-8 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-311581117910.1186/1471-2458-5-31Research ArticleYou might as well smoke; the misleading and harmful public message about smokeless tobacco Phillips Carl V [email protected] Constance [email protected] Brian [email protected] University of Texas Medical School, Center for Clinical Research and Evidence Based Medicine, Houston, USA2 University of Texas School of Public Health, Houston, USA3 Center for Philosophy, Health, and Policy Sciences, Inc., 10923 Atwell Dr, Houston, TX 77096, USA2005 5 4 2005 5 31 31 25 10 2004 5 4 2005 Copyright © 2005 Phillips et al; licensee BioMed Central Ltd.2005Phillips 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
Compared to smoking cigarettes, use of Western smokeless tobacco (ST) products is associated with a very small risk of life-threatening disease (with estimates in the range of a few percent of the risk from smoking, or even less). This means that smokers can realize substantial health benefits by switching to ST, an obvious substitute. But consumers and policy makers have little chance of learning that ST is much less dangerous than smoking because popular information provided by experts and advocates overstates the health risks from ST relative to cigarettes.
Methods
To examine the extent of this overstatement in one medium, we conducted a systematic review of websites containing information about ST and health risks. We examined the content of 316 relevant websites identified by a Google search.
Results
We found that when any substantive information about the risk from ST is given, the risk is almost universally conflated with the risk from cigarettes. Accurate comparative risk information was quite rare, provided by only a handful of websites, all appearing low in our search results (i.e., of low popularity and thus unlikely to be found by someone searching for information). About 1/3 of the websites, including various authoritative entities, explicitly claimed that ST is as bad as or worse than cigarettes. Most of the other sites made statements that imply the risks are comparable.
Conclusion
Through these websites, and presumably other information provided by the same government, advocacy, and educational organizations, ST users are told, in effect, that they might as well switch to smoking if they like it a bit more. Smokers and policy makers are told there is no potential for harm reduction. These messages are clearly false and likely harmful, representing violations of ethical standards.
==== Body
Background
The negative health consequences of smoking cigarettes are well known. What is less well known is that not all tobacco products create similar levels of risk. In particular, use of Western smokeless tobacco (ST) is substantially less harmful than smoking cigarettes. This should not be surprising, given that ST use does not expose the body to the harmful combustion products and assault on the lungs that result from smoking. But even many health experts do not realize there is a major difference, perhaps because of repeated messages about "tobacco" (usually referring just to cigarettes), which imply that all products made from this plant have the same health implications.
ST is usually only linked to one life-threatening disease, oral cancer (OC), and even that association may not apply to the types of products that increasingly dominate ST use in the West [1]. Claims of OC risk are largely based on a single study [2] and are contradicted by a substantial portion of the evidence about modern moist snuff [3-6]. Claims are sometimes also made about links to cardiovascular disease and pancreatic cancer, though the evidence supporting these claims is even thinner and more equivocal. The lack of clear evidence of a strong association with any diseases is not due to lack of research; there have been extensive attempts to find health risks from ST, including in Swedish populations where prevalence of use is high. While it is impossible to ever rule out small associations between an exposure and a disease, there is ample evidence to rule out, with a very high degree of confidence, the possibility that the combined risk of life threatening diseases due to ST use is anything close to that from smoking.
Even if we were to believe the commonly cited estimates for the risk of OC from ST, that risk is still lower than the estimated risk due to smoking for OC alone (a very small fraction of the total risk from smoking). If we further allow for the possibility that ST creates small, yet-undetected risks for some other diseases, the risks from the many diseases caused by smoking still clearly dwarf possible risk from ST. The most frequently repeated estimate conservatively puts the risk of premature mortality from ST use at 2% of that from cigarettes [7,8]. The Royal College of Physicians recently stated that the risk from ST might be as low as 1/1000 that from cigarettes [9]. The Royal College and another recent high-profile report suggest an upper bound estimate of 1/10 the risk [10], but the available epidemiology suggests that the true value is extremely unlikely to be this large. Whatever the exact magnitude, the conclusion must be that cigarettes are considerably more harmful than ST.
This comparison is more than a matter of curiosity or perspective; the products are obvious substitutes. Using products that provide nicotine (which is, in itself, fairly benign) need not be very harmful – it is smoking that is very harmful. Among the several things a smoker can do to eliminate most of the risk from his nicotine use (e.g., quitting nicotine entirely or using pharmaceutical nicotine products), switching to ST is unique in allowing continued consumption of nicotine using a product for which there is a history of consumer demand. For some smokers, this switch – a "harm reduction" strategy – offers the best chance of changing their behavior to eliminate the huge risk from smoking. Calls for such a strategy are increasing in popularity among advocates and in the media.
However, if smokers are unaware of the difference in risk between the two products they are obviously unlikely to switch from cigarettes to ST to reduce their risk. Moreover, if current ST users are unaware of the difference, they will believe that they might as well smoke, possibly resulting in a terribly unhealthy change of products. Casual empiricism and what data exists suggest that consumers are largely unaware that there is a large (or any) difference in risk, and popular information reinforces that lack of understanding. Inaccurate statements by experts, government organizations, and advocates that appear in the press and educational materials frequently reinforce the mistaken belief that risks are similar. To expand on casual observations about such misinformation and to systematically assess its prevalence in one forum – web pages offering health advice about ST – we conducted the systematic review presented in this paper.
Unethical messages
A preeminent tenet of modern health and medical ethics is the right of individuals to make fully-informed autonomous decisions, and the accompanying obligation of health experts, clinicians, and policy makers to provide the information and permit the autonomy. Systematic provision of inaccurate comparative risk information violates these principles.
The provision of such misinformation can be attributed to ignorance of the science, deliberate misrepresentation under the belief that it is for people's own good, or deliberate misrepresentation for other reasons. Each of these is an ethical violation. Intentionally misleading people about health information, even for their own good, is considered ethically unacceptable in all but the most extreme emergent cases, and in the present case there are compelling arguments that this misleading information does more harm than good. But even when ignorance is the explanation, it must be considered unethical. When an individual or organization actively endeavors to provide authoritative information and portray it as accurate, particularly when such an entity is considered a respected authority on health science, ignorance of well-established scientific truth is unethical negligence.
Kozlowski and O'Conner recently challenged the U.S. Centers for Disease Control and Prevention (CDC) and the U.S. Substance Abuse and Mental Health Services Administration (SAMHSA) over the content of their websites, which included clearly false claims that ST poses a similar health risk to cigarettes [11]. Kozlowski and O'Conner reported that following their protests, CDC changed their website (SAMHSA did not), though the change was merely from out-and-out falsehood ("Is smokeless tobacco safer than cigarettes? NO WAY!") to a literally true statement that still misleads readers, a tactic discussed below. Our review lets us assess how pervasive such false and misleading comparisons are.
Methods
To examine the extent of systematic overstatement of the risks from ST in one medium, we conducted a systematic review of popular sources of information, looking at websites that implicitly purport to deliver a public service message about ST and health risks. (We have previously presented preliminary results of this research along with more extensive background [12,13].)
With an increasing portion of those seeking health information turning to the web,[14,15] information found there provides a good measure of what people might learn. While websites do not contain all popularly available information, many people searching for information on this topic would start with a web search and most organizations that have a stated position on the topic, particularly those actively trying to influence popular opinion through other media, have a web page that reflects their claims. Thus, the information in web pages is likely to be representative of all information reaching the average consumer.
We performed a Google search for [tobacco AND cancer AND (smokeless OR snuff OR dip OR spit OR chew OR chewing)], the latter disjunction covering most of the synonyms for "smokeless". We conducted the search on 3 May 2003 and stored the results offline so they would not change when re-accessed. The search reported 763 results (after Google's algorithms eliminated many, but not all, multiple similar hits), which we used as our dataset. While the nature of the Google search makes it impossible to describe the exact sampling properties that yielded this population, and thus we would hesitate to do any formal statistical analysis, it is safe to conclude that it contains all of the most popular websites that address the subject, as well as a large portion of the less popular ones. This is sufficient for the present analysis.
We were interested in public service sites (as implicitly self-defined, without an attempt on our part to judge what constitutes a genuine public service) or health advice/information sites which state an entity's opinion about the health risks from ST. Our protocol eliminated: sites that were selling tobacco products or methods for quitting tobacco (with one exception discussed below); news of the day; search engine, web maps, and other sites that just provide links; sites from South Asia (because the products used there contain major ingredients other than tobacco and the epidemiologic research suggests they create greater oral health problems than Western moist snuff); and scientific literature (scholarly papers, journals, and conference abstracts). Eliminating these and the double counting from organizations that were duplicated in the Google results yielded 316 web presences in our population. The search terms limited the results to English language sites. The vast majority were U.S. entities, with a handful from the U.K., Canada, and other countries.
For each included website, we searched the entire website (not just the page hits from the Google search) for statements about the health effects of ST and collected the results. We ignored information that was clearly not the position of the sponsoring organizations (e.g., quotations of material they were disagreeing with; postings on message boards). We initially reviewed the websites between 4 May 2003 and 11 June 2003 and printed out relevant pages. We audited our results and expanded the collected data during the period 25 September 2003 through 8 December 2003. To maximize consistency, multiple researchers viewed each website, and ambiguous codings were discussed by the group.
The ordering of the websites in our list is important because those that are earlier are more popular (specifically, are more often linked to from other sites) and are much more likely to be found and accessed by someone doing a search. In the results presented below, the hit number is a website's ranking within in our list of 763, with lower numbers being higher ranked (earlier in the list). For websites that generated multiple hits, we present the highest ranked hit unless otherwise noted.
Results and Discussion
The risk from ST is widely conflated with the risk from cigarettes on websites that provide health advice and information. Almost every website had statements that played up the health risks from ST without caveat, making it difficult for consumers to recognize the huge contrast with cigarettes. The quantitative claims of health risks from ST were very often beyond a worst-case-scenario interpretation of the scientific literature. A large portion of websites directly stated or implied that the risks from ST and cigarettes are similar.
As noted above, the most salient feature of the comparative risks of smoking and ST is how different they are, a message that is buried deeply in the websites that inform the public on this topic.
Very little accurate comparative risk information
Very few websites provided accurate information. Two organizations, ASH (Action on Smoking and Health) in the U.K. and the American Council on Science and Health (ACSH) in the U.S. were the most prominent sources of accurate comparative risk information. Their highest appearances in our search were hits 96 and 93, respectively, meaning they would likely not be found by someone seeking information on ST, since most people seldom read beyond the first few tens of hits [16,17]. Moreover, it was not until hit 491 that ASH's major harm reduction statement [10], probably the most prominent current call to consider harm reduction for cigarettes, appears. ACSH's harm reduction message appears at hit 120 [16]. The earlier hits were actually anti-harm-reduction statements, apparently presented by those organizations to acknowledge other positions (these ranked higher that the organizations' actual stated positions because more other websites linked to those pages).
Brad Rodu, a professor of pathology and dentistry at the University of Alabama at Birmingham (UAB), is the longtime leading advocate of the use of ST as a harm reduction strategy for smoking. His pages at the UAB website provide comprehensive information on the topic, but this was only hit 625 on our list [18]. A commercial site (included because of its extensive health message) for a quit-smoking product, hit 408, is a mirror of an old version of Rodu's UAB pages, posted under Rodu's name with his permission [19]. Despite his numerous writings on the topic, the only earlier entries on our list that would lead to Rodu's work were hit 276, one of his op-eds in the news archives from an anti-tobacco organization, and the aforementioned ACSH hit 120 [16].
Only three other sites mentioned that ST use is not as bad as cigarettes, and they offered little more than mentions [20-22]. Astonishingly, we were unable to find any other statements about the much lower risk of ST compared to smoking. No high-ranking sites provided the information tobacco users would need to make choices based on which product is safer. Notably, no site from the most prolific source of information, the U.S. government, provided such information (excluding a few scholarly or technical papers that can be downloaded from the sites but are not presented as the government's message to consumers). Indeed, U.S. government agencies consistently provided misleading information, as did popular medical advice sites and the best-known advocacy groups.
Misleading comparative risk information
The most prevalent messages were those that would tend to convince readers that the health risk from ST is comparable to that from smoking.
We identified 237 of the remaining 309 websites in our population as discussing the risks of smoking and ST in proximity to each other. Most of the other 72 sites either contained very little substance (often just a passing mention that ST poses health risks), appeared very low in our results, or both, so these numbers tends to understate how common the juxtaposition of health claims about cigarettes and ST is. Any juxtaposition of health claims about the two products that does not make clear the very different absolute risk, even if it makes no explicit comparison, implies to readers that the risks are comparable. Most websites did more than merely juxtapose; they made specific statements that reinforced this implication.
Explicit claims of equal risk
We identified 108 websites that claimed that the risks from ST are as bad as or worse than those from smoking. Most often this took the form of an explicit statement that ST is not safer than smoking. It is worth noting that this is equivalent to saying that you are better off, or at least no worse off, deciding to smoke rather than use ST.
Examples include various authoritative entities
• American Cancer Society: "Some people believe that using smokeless tobacco is safer than smoking. This is not true." [23]
• World Health Organization: "There is also a prevalent myth that it is less dangerous than smoking. The reality is that smokeless tobacco is just as addictive and fatal as cigarettes." [24]
• U.S. Department of Health and Human Services (the statement noted by Kozlowski and O'Conner): "Q. Isn't smokeless tobacco safer to use than cigarettes? A. No." [25]
Implicit claims that ST is worse than cigarettes
Of the 108 websites making claims that ST is as bad or worse than cigarettes, 26 suggested that ST is worse than smoking by likening the risks and then identifying differences that exclusively favor smoking.
A typical example appears in the second highest-ranking website from our search, the Academy of General Dentistry: "Isn't it safer than smoking? Absolutely not. Some wrongly believe that spit tobacco is safer than smoking cigarettes. But spit tobacco is more addictive because it contains higher levels of addictive nicotine than cigarettes and can be harder to quit than cigarettes. One can of snuff delivers as much nicotine as 60 cigarettes." Though there is no explicit claim that ST is worse – the explicit claim is simply that it is no better – the comparisons that follow imply that it is better to smoke than to use ST.
Implicit claims of equal risk
Of the websites not making explicit claims that ST is as bad as or worse than cigarettes, 100 made statements that directly imply that risks from ST are comparable to those of smoking, while another 29 simply juxtaposed the two risks without suggesting there are differences. (Most of those that made explicit claims also included some of these implicit claims.)
There are various literally true statements that are apparently intended to dissuade readers from the (accurate) belief that ST is safer than smoking. Some might argue that such statements do not violate ethical rules that prohibit lying. On the other hand (as has been widely discussed regarding recent U.S. government policy in other arenas), clearly misleading statements that are carefully crafted to be literally true are arguably worse than literally false statements. They suggest that the authors know the truth and believe that it is sufficiently clear that they should maintain a plausible claim they are not contradicting it, but are nevertheless trying to get people to believe a falsehood.
The most popular types of literally-true-but-misleading information were comparisons with smoking that characterize ST as "not a safe substitute to smoking cigarettes" or "not harmless," or statements that "there is no safe tobacco." (The former of these is quoted from the 1986 U.S. Surgeon General's report [27] or the similar warning on 1/3 of the units of ST products sold in the U.S.) We identified 62 websites (among those not making an explicit claim of equivalent risk) that made one or more such claims. Since basically nothing is perfectly safe, these statements are literally true, but the comparison implies more than the literal interpretation, "it would not eliminate every last bit of risk to switch from cigarettes to ST." Saying "ST is not a safe alternative" without any hint of the fact that it is immensely safer implies that there is no benefit in switching from smoking to ST or, equivalently, no increased risk in switching from ST to smoking.
We identified 55 websites where ST and smoking risks were combined in lists of health effects or attributable risk (this excludes those that make explicit claims of equivalent risks, but overlaps with the 62 in the previous paragraph), either by conjunction or by using the word "tobacco" in contexts where it refers to both products. A popular U.S. health advice site, Virtual Hospital, states, "Both cigarettes and smokeless tobacco are harmful to your child's health," followed immediately by detailing the known health effects of smoking [28]. The U.S. National Library of Medicine's consumer advice site, MedlinePlus, under the heading "Tobacco use, smoking and smokeless tobacco," states, "Tobacco and its various components increase the risk of cancer (especially in the lung, mouth, larynx, esophagus, bladder, kidney, pancreas, and cervix), heart attacks and strokes, and chronic lung disease" [29]. Absent a statement of specific or comparative risks, this tends to imply that the components of the conjunction contribute similarly to all the claimed outcomes. These conjunctions are particularly common in the later Google hits, which only briefly mention ST, often in a broad discussion of behavioral risk factors, suggesting that most brief presentations of the health effects of ST conflate the exposure with smoking.
The U.S. National Cancer Institute (NCI) had the largest number of search hits (all of the first 4 and 16 others). We found no literally false claims in the NCI websites. However, they did make many literally true misleading claims, including saying "not a safe alternative" [30-32], and lumping together attributable risk from ST for oral cancer with the (many times greater) attributable risk from cigarettes [32,33]. A particularly misleading conjunction is, "Smoking tobacco, using smokeless tobacco, and being regularly exposed to environmental tobacco smoke are responsible for one-third of all cancer deaths in the United States each year" [34]. Even the worst-case scenario for claims about the risk from ST would make it responsible for in the order of 1/1000 of this attributable risk.
Relative popularity
The imbalance of good and bad information is somewhat worse if we focus on the hits from earlier in the list (i.e., the ones more likely to be found and accessed). Looking at the first 90 hits, chosen because (at the default ten hits displayed per page) those would be appear before a searcher saw a link to ASH, ACSH, or any accurate comparative risk information, yields 44 websites in our population. Those include 13 that claim ST is as bad or worse than cigarettes and 19 others that use one of the rhetorical devices to imply the risks are similar.
Conclusion
Even though we were aware that the available popular information was skewed before we undertook the systematic review, we were astonished to find the near ubiquity of misinformation and how unlikely consumers are to find accurate comparative risk information. Websites provide a substantial and growing portion of the knowledge consumers get about health issues, including the health implications of tobacco use. We expect that the mix of information about ST we found is similar to that provided in pamphlets, public service messages, and other popular media since the organizations represented in our websites are the same ones that provide that information. A recent study reporting some information about pamphlets tends to confirm this hypothesis [35], as does our monitoring of popular press stories (unpublished). Thus, the pattern of misinformation we found very likely represents a comparable pattern of misinformation reaching people from all popular sources.
The negative health implications of preventing people from realizing that ST is relatively safe should not be underestimated. ST users are told, in effect, that they might as well switch to smoking if they find they like it a bit more. The larger population of smokers is told that they cannot switch to a safer form of tobacco, a message that is often characterized as "quit or die." It is extremely difficult for anyone to deliver a harm reduction message in the face of the widespread misperception that is fueled by the misinformation. At this point, we can only speculate about how many smokers would take advantage of this opportunity to reduce their risk by two orders of magnitude or more.
Health advocates, particularly those in public service, have an affirmative ethical duty to tell the truth. It is difficult to justify keeping the truth from people, even when knowing it might be harmful; it is clearly unjustified when it would be beneficial.
Abbreviations
ST = smokeless tobacco
OC = oral cancer
CDC = U.S. Centers for Disease Control and Prevention
ASH = Action on Smoking and Health
ACSH = American Council on Science and Health
UAB = University of Alabama at Birmingham
NCI = U.S. National Cancer Institute
Competing interests
Phillips is the director of and Guenzel was (at the time of writing) employed by the non-profit research institute, Center for Philosophy, Health, and Policy Sciences, Inc. CPHPS is the recipient of an unrestricted gift from U.S. Smokeless Tobacco Company for support of research of Dr. Carl V. Phillips. This research was investigator-initiated. USSTC did not influence the content or see the study results before they were publicly released. Phillips and Wang have received consulting fees from USSTC related to litigation. Phillips is the recipient of an unrestricted research grant from USSTC at the University of Alberta where he will soon be employed.
Authors' contributions
CVP designed the study and analysis, participated in data collection and cleaning, and wrote the manuscript (based partially on earlier poster versions). CW assisted in the design of the study, participated in data collection and analysis, and oversaw the data cleaning and auditing. BG was the primary data collector, contributed to study design, participated in data cleaning and analysis, and coauthored the earlier poster versions of the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Foulds J Ramstrom L Burke M Fagerstrom K Effect of smokeless tobacco (snus) on smoking and public health in Sweden Tob Control 2003 12 349 359 14660766 10.1136/tc.12.4.349
Winn DM Blot WJ Shy CM Pickle LW Fraumeni JF Snuff dipping and oral cancer among women in the southern United States N Engl J Med 1981 304 745 749 7193288
Schildt EB Eriksson M Hardell L Magnuson A Oral snuff, smoking habits and alcohol consumption in relation to oral cancer in a Swedish case-control study Int J Cancer 1998 77 341 346 9663593 10.1002/(SICI)1097-0215(19980729)77:3<341::AID-IJC6>3.0.CO;2-O
Lewin F Norell SE Johansson H Gustavsson P Wennerberg J Bjorklund A Rutqvist LE Smoking tobacco, oral snuff, and alcohol in the etiology of squamous cell carcinoma of the head and neck. A population-based case-referent study in Sweden Cancer 1998 82 1367 1275 9529030 10.1002/(SICI)1097-0142(19980401)82:7<1367::AID-CNCR21>3.0.CO;2-3
Rodu B Cole P Smokeless tobacco use and cancer of the upper respiratory tract Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2002 93 511 515 12075196
Phillips CV Publication Bias in situ Am J Epidemiol 2003 157 S59 Accessed June 6, 2004
Rodu B For Smokers Only University of Alabama 2004 Accessed August 18, 2004
Rodu B For Smokers Only: How smokeless tobacco can save your life 1995 Sulzburger & Graham Pub Co
Royal College of Physicians of London Protecting smokers, saving lives: The case for a tobacco and nicotine regulatory agency 2002 Salisbury: Sarum Colorview Group Accessed August 18, 2004
Bates C Fagerstrom K Jarvis M Krunze M McNeill A Ramstrom L European Union policy on smokeless tobacco: A statement in favour of evidence-based regulation for public health. Action on Smoking and Health Tob Control 2003 12 360 367 14660767 10.1136/tc.12.4.360
Kozlowski LT O'Conner RJ Apply federal research rules on deception to misleading health information: an example on smokeless tobacco and cigarettes Public Health Reports 2003 118 187 192 12766212
Phillips CV Wang C Guenzel B Daw C Smokeless tobacco and oral cancer, the curious history of a "fact" Am J Epidemiol 2003 157 S13
Phillips CV Guenzel B Internet Information About Health Risks From Smokeless Tobacco. Poster, 10th Annual Meeting for Society for Research on Nicotine and Tobacco Phoenix, Arizona 2003 Accessed June 6, 2004
Fox S Rainie L The online health care revolution: How the Web helps Americans take better care of themselves Pew Internet & American Life Project: Online life report 2000 Accessed December 18, 2003
Rideout V Generation Rx.com: How Young People Use The Internet For Health Information. The Kaiser Family Foundation The Henry J Kaiser Family Foundation 2001
ACSH American Council on Science and Health Accessed May 18, 2003
ASH US Tobacco acknowledges smokeless tobacco causes oral cancer. Action on Smoking and Health Smoking News and Documents Accessed October 20, 2003
Rodu B Pathology Primary Faculty Listing University of Alabama Department of Pathology Accessed October 17, 2003
Rodu B For smokers only: The research. DrRodu.com For Smokers Only Accessed October 17, 2003
University of Wisconsin A pitch for tobacco. Why Files Accessed June 17, 2004
Utah Education Network Prevention Dimensions: Give me a reason. Lesson plan Accessed May 24, 2003
Socialstyrelsen Health Hazards of Moist Snuff. Socialstyrelsen English translation Accessed October 13, 2003
ACS Quitting Smokeless Tobacco. American Cancer Society prevention and early detection Accessed October 20, 2003
WHO Smokeless tobacco. World Health Organization Accessed October 16, 2003
SAMHSA Frequently asked questions about substance abuse. U.S. Department of Health and Human Services Frequently Asked Questions Accessed December 6, 2003
Surgeon General's Advisory Committee on the Health Consequences of Using Smokeless Tobacco. The health consequences of using smokeless tobacco 1986 U S National Institutes of Health Publication 86 2874
Virtual Hospital. For Patients Smokeless Tobacco. Virtual Hospital: Cancer: Smokeless Tobacco Accessed May 15, 2003
U.S. National Library and the National Institute of Health: Medline Plus Health Information. Smokeless Tobacco. Medline Plus Health Information Accessed May 29, 2003
NCI Smokeless Tobacco. National Cancer Institute Cancer Control & Population Sciences Accessed October 14, 2003
NCI Smokeless Tobacco: especially for kids. National Cancer Institute Cancer Control & Population Sciences Accessed October 14, 2003
NCI Smokeless Tobacco and Cancer: Questions and Answers. National Cancer Institute Cancer Facts Accessed May 30, 2003
NCI Oral Cancer Prevention. National Cancer Institute Oral Cancer (PDQ): Prevention December 17, 2003
NCI What You Need To Know About Cancer – An Overview. National Cancer Institute Cancer Information Accessed December 6, 2003
Waterbor JW Adams RM Robinson JM Crabtree FG Accortt NA Gilliland MJ Disparities between public health educational materials and the scientific evidence that smokeless tobacco use causes cancer J Cancer Educ 2004 19 17 28 15059752 10.1207/s15430154jce1901_08
| 15811179 | PMC1090592 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Apr 5; 5:31 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-31 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-321581117810.1186/1471-2458-5-32Research ArticleMid-term Body Mass Index increase among obese and non-obese individuals in middle life and deprivation status: A cohort study Lyratzopoulos Georgios [email protected] Patrick [email protected] Richard F [email protected] Margaret [email protected] Philip S [email protected] Directorate of Clinical Services and Public Health, Norfolk, Suffolk and Cambridgeshire Strategic Health Authority, Fulbourn, UK2 Evidence for Population Health Unit, University of Manchester, Manchester, UK3 Former Stockport Healthcare NHS Trust, Stockport, UK4 Stockport NHS Trust, Stockport, UK2005 5 4 2005 5 32 32 29 11 2004 5 4 2005 Copyright © 2005 Lyratzopoulos et al; licensee BioMed Central Ltd.2005Lyratzopoulos 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 UK, obesity is associated with a clear socioeconomic gradient, with individuals of lower socioeconomic status being more likely to be obese. Several previous studies, using individual measures of soecioeconomic status, have shown a more rapid increase in Body Mass Index (BMI) over time among adults of lower socioeconomic status. We conducted a study to further examine whether ecologically defined deprivation status influences within-individual BMI change during middle life, as the answer to this question can help determine optimal preventive strategies both for obesity per se, and its' associated socioeconomic disparities.
Methods
Anonymised records of participants to the Stockport population-based cardiovascular disease risk factor screening programme were analysed. Individuals aged 35–55 who had a first screening episode between 1989 and 1993, and a subsequent screening episode were included in the study. Deprivation status was defined using quintiles of the Townsend score. Mean annual BMI change by deprivation group was calculated using linear regression. Subsequently, deprivation group was included in the model as an ordinal variable, to test for trend. The modelling was repeated separately for individuals who were obese (BMI < 30) and non-obese at the time of first screening. In supplementary analysis, regression models were also adjusted for baseline BMI.
Results
Of 21,976 women and 19,158 men initially screened, final analysis included just over half of all individuals [11,158 (50.8%) women and 9,831 (51.3%) men], due to the combined effect of loss to follow-up and incomplete BMI ascertainment. In both sexes BMI increased by 0.19 kg/m2 annually (95% Confidence Intervals 0.15–0.24 for women and 0.16–0.23 for men). All deprivation groups had similar mean annual change, and there was no evidence of a significant deprivation trend (p = 0.801, women and 0.892, men). Restricting the analysis to individuals who were non-obese at baseline did not alter the results in relation to the lack of a deprivation effect. When restricting the analysis to individuals who were obese at baseline however, the findings were suggestive of an association of BMI increase with higher deprivation group, which was further supported by a significant association when adjusting for baseline BMI.
Conclusion
In the study setting, the BMI of non-obese individuals aged 35–55 was increasing over time independently of deprivation status; among obese individuals a positive association with higher deprivation was found. The findings support that socioeconomic differences in mean BMI and obesity status are principally attained prior to 35 years of age. Efforts to tackle inequalities in mean BMI and obesity status should principally concentrate in earlier life periods, although there may still be scope for focusing inequality reduction efforts on obese individuals even in middle life.
==== Body
Background
Increased attention is being paid to the "epidemic of obesity" in post-industrialised countries [1-3]. Since the 1980s, there has been a rapid rise in the proportion of the population who are obese, and in the mean values of Body Mass Index (BMI) and other anthropometric measurements associated with obesity. One of the striking characteristics of the epidemic is that it appears to affect most age groups, including children [4-8].
Among individuals of middle age, obesity is one of the cardiovascular risk factors that demonstrate a clear socioeconomic gradient, with individuals of lower socioeconomic status being more likely to be obese [9,10]. At least four previous studies have shown that individuals with lower socioeconomic status experience a more rapid BMI increase during adulthood [11-14]. These previous studies measured socioeconomic status directly, using either job grade [11], own or parental occupation-defined social class [12,13], employment status [12], or educational attainment [12-14]. These studies also included individuals whose minimum baseline age was 25 years or younger [11-14]. It is therefore important to further examine whether cross-sectional socioeconomic gradients in BMI progress continuously throughout the life course, including during middle life in particular, or whether periods in earlier life are mainly responsible for BMI socioeconomic disparities in later life. As cardiovascular risk greatly increases from middle to late life, determining optimal strategies for the prevention of obesity and its' associated socioeconomic disparities in middle-aged individuals is of great importance to public health and healthcare policy-making.
We therefore examined mean annual BMI change among participants of a primary care based cardiovascular risk factor screening programme co-ordinated by a UK Health Authority. The Stockport Cardiovascular Disease Risk Factor Screening Programme was originally introduced in 1989. Using a call-recall system operated by the Stockport Health Authority, residents registered with a Stockport General Practitioner (GP) and aged between 35 and 60 were invited five-yearly to book a screening appointment. During the 1989–1993 period, about 10.8% of all patients that were registered with a GP were excluded from initial screening invitation, as they were already known to suffer from hypertension (3.91%), diabetes (1.23%) and conditions including history of any cardiovascular disease, and terminal illness (6.64%) (see Additional File 1).
In relation to all Stockport residents aged 35–60 enumerated at the 1991 Census the population coverage was 53.7% for women and 47.5% for men, however true coverage among invitees was higher as the denominator used in this calculation also includes those individuals excluded from screening invitation (see also Additional File 1). Coverage was significantly higher among least deprived groups in men (but not in women), however the magnitude of this difference was small (i.e. coverage of 49.4%, 48.3%, 48.4%, 46.5% and 44.8% for the least to the most deprived quintile group respectively, p < 0.001 (see Additional File 1 for details of the method used for this calculation). Because exclusions due to already known (treated) hypertension or cardiovascular disease can be hypothesized to be more frequent among more deprived individuals (due to a higher prevalence of hypertension and cardiovascular disease in those individuals), the true coverage among deprived invitees may be higher to that calculated, as the denominator used includes "excluded" cases. Individual data on risk factor levels were collated by the Health Authority and anonymised into a usable electronic dataset, used in the present study. An evaluation of the quality and utility of this dataset for research purposes was carried out in 2002. The evaluation concentrated on examining data completeness, and whether or not there were systematic differences in completeness of data items by individual patient characteristics. Also, whether the sample of screened individuals was concordant to the socio-demographic characteristics of the Stockport population of similar age. The evaluation concluded that the quality of the data was excellent [15].
Methods
Individuals who attended for an initial screening during the first five-year cycle of the screening programme (1989–1993) and who also attended for screening on a subsequent occasion to the end of 1999 were included in the study. Analysis was restricted to individuals who at the time of the first screening were aged 55 years or younger, as older individuals would have not usually been scheduled for a subsequent 5-yearly screening episode (the upper age limit for participation to the Programme being 60).
Measurements
On both screening occasions height was measured to the nearest centimetre and weight to the nearest kilogram. Weight scales and stadiometers of variable types were used, available at the Stockport General Practice surgeries participating in the scheme. Protocols were in place and training was provided about measurement of weight and height by a visiting nurse screening facilitator, whose role was to quality assure and co-ordinate the implementation of the screening activity. Weight was measured with light clothing and without shoes. Body Mass Index was calculated by dividing weight in kilograms by height in meters squared. Other measurements are described in Additional File 1.
Information on deprivation status was based on the Townsend deprivation score of the enumeration district of residence (1991 Census) [16]. The Townsend deprivation score measures deprivation at a small area level (Census Enumeration District), using information from four separate Census-based variables: unemployement (unemployed residents aged over 16 as a percentage of all economically active residents aged over 16), overcrowding (households with one person per room and over as a percentage of all houselhods), non-car ownership (households with no car as a percentage of all households), and non-home ownership (households not owning their own home as percentage of all households) (s ). Five deprivation quintiles were defined, using the participant's distribution of the Townsend score. The range of Townsend scores among study participants was -7.12 to +10.9 (mean: -1.69, median: -2.39, standard deviation: 2.88, quintile-defining points: -4.10, -2.93, -1.77, +0.61). The range for the population of England and Wales as a whole was -7.55 to +11.8 (mean: 0, median: -0.65, standard deviation: 3.39, quintile-defining points: -3.1, -1.55, 0.39 and 3.13).
Statistical analysis
Whether "loss-to-follow-up" (i.e. lack of second screening) was differential was examined using multiple logistic regression. The dependent binary variable was attendance (or not) for a second screening; and the independent variables were age, presence of risk factor values above cut-off points and deprivation group. Similarly, whether completeness of BMI ascertainment (i.e. whether BMI was measured on both the first and the second screening) was differential was examined using multiple logistic regression.
The significance of baseline deprivation group differences in obesity status (BMI >30) and mean BMI value was assessed with simple logistic and linear regression respectively, entering deprivation group as a continuous variable.
To examine mean annual BMI change during follow-up, a linear regression model was fitted with change in BMI between the initial and second screening as the dependent variable and follow-up time in years (taken as screening interval between the two screening episodes) as the independent variable of interest, adjusting for age at initial screening (model 1). In this model, the co-efficient for follow-up time denotes the mean (age-adjusted) annual BMI change per year of follow-up. The model was subsequently fitted to each deprivation group stratum separately (model 2), producing five different co-efficients for follow-up time, denoting the mean (age-adjusted) annual BMI change for each deprivation group. To test for a statistically significant effect of deprivation on BMI change, all individuals were subsequently included in one model and deprivation group entered as a continuous variable (and age- and follow-up period-adjusted) (model 3). In this model, the co-efficient for deprivation group denotes the difference in BMI change for each one level increase in deprivation group.
Stratified analysis by baseline obesity status
The analysis was repeated stratifying by baseline obesity status, i.e. sequentially restricting the analysis to individuals who were obese and non-obese at baseline (BMI values of 30+ or <30 respectively).
Baseline BMI adjustment
Furthermore, as per previous research on the subject, [11-14] analysis was repeated adjusting for baseline BMI (i.e. at the first screening) in all models.
Results
In total there were 21,976 women and 19,158 men aged 35–55 who had a first screening episode during 1989 and 1993, of whom 16,932 women (77%) and 13,812 men (72.1%) also had a second screening. Deprivation status was available for 99.8% of all cases. The mean screening interval (follow-up period) was 4.79 years for women and 4.83 years for men. Persons lost to follow-up were significantly more likely to be obese, hypertensive, to have high cholesterol, and to be current smokers, younger and more deprived (Table 1).
Table 1 Baseline characteristics and completeness of follow-up (second screening). Significance levels from multiple regression models.
Baseline characteristic Women (n = 21,976) Men (n = 19,158)
Initial screen only Two screens p Initial screen only Two screens p
Mean age (years) 44.5 44.9 <0.001 44.3 44.7 <0.001
% current smokers 39.9 36.4 0.018 53.4 48.0 <0.001
% with SBP>140 mmHg 25.5 18.2 <0.001 33.2 24.8 <0.001
% with DBP>90 mmHg 15.3 9.4 * 23.9 17.0 *
% cholesterol > 6.5 mmol/l 25.5 22.6 0.019 36.7 29.8 <0.000
% BMI >30 kg/m2 17.7 11.7 <0.000 15.9 11.9 0.001
% in the two most deprived groups (4 and 5) 43.6 37.3 <0.001** 42.7 36.8 <0.001**
SBP: Systolic Blood Pressure; DBP: Diastolic Blood Pressure
*Significant in multiple logistic regression model excluding SBP but omitted from presented model to avoid over-adjustment, as highly correlated with SBP.
** Deprivation group (quintile) entered as a continuous variable.
Among individuals with two screening episodes 11,158 women (65.9%) and 9,831 men (71.2%) had dual (i.e. in both screens) BMI ascertainment. Individuals were significantly more likely to have complete BMI ascertainment if they were more deprived; and for men only, if they were current smokers and of younger age (Table 2). Hypertensive men and women, and obese men were significantly less likely to have complete ascertainment. High cholesterol at baseline did not have a significant effect.
Table 2 Baseline characteristics and completeness of "dual" BMI ascertainment (i.e. BMI measurement on both screening episodes). Significance levels from multiple regression models.
Women (n = 16,932) Men (n = 13,812)
Without "dual" BMI ascertainment Dual BMI ascertainment p Without "dual" BMI ascertainment Dual BMI ascertainment p
Mean age (years) 45.3 44.6 0.717 45.3 44.2 0.001
% current smokers 14.3 36.1 0.980 18.6 44.4 0.024
% with SBP>140 mmHg 23.0 15.7 <0.001 30.9 22.3 <0.001
% with DBP>90 mmHg 12.0 8.1 * 21.7 15.1 *
% cholesterol > 6.5 mmol/l 7.6 15.2 0.064 11.5 21.2 0.253
% BMI >30 kg/m2 4.8 11.4 0.150 6.0 11.2 <0.001
% in the two most deprived groups (4 and 5) 29.6 41.2 <0.001** 31.7 38.8 0.006**
SBP: Systolic Blood Pressure; DBP: Diastolic Blood Pressure
*Significant in multiple logistic regression model excluding SBP but omitted from presented model to avoid over-adjustment, as highly correlated with SBP.
** Deprivation group (quintile) entered as a continuous variable.
From the original cohort, final analysis included just over half of all individuals (50.8% of women and 51.3% of men), due to the combined effect of loss to follow-up and incomplete BMI ascertainment. The combined effect of differential loss to follow-up and BMI ascertainment completeness in relation to deprivation group is summarised in Additional File 2.
Among individuals with dual BMI ascertainment, at the initial screening, 11.4% of women and 11.2% of men were obese (BMI > 30 kg/m2). The percentage of women who were obese was 7.8%, 8.6%, 9.3%, 13.1% and 17.6% for the least to most deprived quintiles, respectively (p < 0.001). For men the respective percentages were 8.8%, 9.3%, 11.2%, 11.4% and 15.7% (p < 0.001). The overall mean BMI was 24.79 kg/m2 for women and 25.69 kg/m2 for men. The mean levels of BMI for women (initial screening) were 24.15 kg/m2, 24.35 kg/m2, 24.67 kg/m2, 25.05 kg/m2 and 25.74 kg/m2 for the least to most deprived quintiles, respectively (p < 0.001); and for men 25.41 kg/m2, 25.56 kg/m2, 25.79 kg/m2, 25.70 kg/m2 and 26.07 kg/m2 respectively (p < 0.001).
The average annual increase in BMI among women was 0.19 kg/m2 (95% Confidence Intervals [CI]: 0.15–0.24), and 0.19 kg/m2, 0.21 kg/m2, 0.21 kg/m2, 0.13 kg/m2 and 0.24 kg/m2 for least to most deprived quintiles respectively (Table 3 and Figure 1). In men, the average annual increase was also 0.19 kg/m2 (CI: 0.16–0.23) and 0.22 kg/m2, 0.16 kg/m2, 0.15 kg/m2, 0.24 kg/m2 and 0.18 kg/m2 for least to most deprived quintiles respectively. There were no significant trends in mean annual BMI change across deprivation groups (p = 0.801, women, 0.891, men).
Table 3 Mean annual Body Mass Index change, adjusted for age and follow-up time.
All individuals
Women Men
Mean LCI UCI p Mean LCI UCI p
All (Women: n = 11,158, Men: n = 9,831) Model 1
All 0.19 0.15 0.24 <0.001 0.19 0.16 0.23 <0.001
Model 2
Affluent 0.19 0.09 0.29 <0.001 0.22 0.15 0.30 <0.001
2 0.21 0.11 0.32 <0.001 0.16 0.09 0.23 <0.001
3 0.21 0.13 0.30 <0.001 0.15 0.08 0.22 <0.001
4 0.13 0.05 0.20 0.001 0.24 0.18 0.31 <0.001
Deprived 0.24 0.14 0.34 <0.001 0.18 0.11 0.26 <0.001
Model 3
Trend 0.00 -0.04 0.03 0.801 0.00 -0.03 0.03 0.891
Stratified analysis to non-obese individuals
BMI < 30 at first screening (Women: n = 9,891, Men: n = 8,729) Model 1
All 0.19 0.15 0.23 <0.001 0.16 0.13 0.19 <0.001
Model 2
Affluent 0.19 0.10 0.27 <0.001 0.18 0.12 0.24 <0.001
2 0.24 0.15 0.32 <0.001 0.12 0.05 0.18 <0.001
3 0.20 0.12 0.28 <0.001 0.13 0.07 0.19 <0.001
4 0.13 0.05 0.20 0.001 0.21 0.15 0.27 <0.001
Deprived 0.22 0.13 0.31 <0.001 0.13 0.06 0.21 <0.001
Model 3
Trend 0.00 -0.03 0.03 0.954 -0.01 -0.04 0.01 0.352
Stratified analysis to obese individuals
BMI 30+ at first screening (Women: n = 1,267, Men: n = 1,102) Model 1
All 0.16 -0.03 0.36 0.107 0.34 0.19 0.49 <0.001
Model 2
Affluent 0.24 -0.59 1.07 0.570 0.39 -0.08 0.87 0.106
2 0.04 -0.62 0.71 0.902 0.30 -0.06 0.66 0.105
3 0.29 -0.18 0.75 0.229 0.25 -0.12 0.63 0.185
4 0.09 -0.18 0.37 0.505 0.39 0.10 0.69 0.008
Deprived 0.28 -0.07 0.63 0.120 0.38 0.11 0.65 0.005
Model 3
Trend 0.18 -0.01 0.37 0.057 0.17 0.00 0.33 0.049
Model 1. Adjusted for age and follow-up time, with all participants included. Reported co-efficient of Model 1 denotes mean annual BMI change (kg / m2 / year).
Model 2. As for model 1, but stratified by deprivation group. Reported co-efficients denote the mean annual BMI change for each deprivation group.
Model 3. As for model 1, but deprivation group entered as ordinal variable, with all individuals included. Reported co-efficient denotes the additional mean annual BMI change for each one level increase in deprivation group (e.g. from deprivation group i to i+1).
LCI: Lower Confidence Interval; UCI: Upper Confidence Interval, BMI: Body Mass Index.
Figure 1 Mean annual change in BMI (age- and follow-up- adjusted), women.
Analysis stratified by baseline obesity status
Restricting the analysis to individuals who were not obese at the time of first screening produced trivial changes to the results by deprivation group, and once more there was no significant deprivation trend in mean annual BMI change (p = 0.954, women, p = 0.352, men) (Table 3). The small sample size limits the precision of estimates obtained when restricting analysis to individuals who were obese at baseline, however a borderline significant effect of deprivation on BMI change was found in test for trend for both men and women (p = 0.057 for women, and 0.049 for men).
Adjustment for Baseline BMI
When adjusting for baseline BMI value (i.e. at screening episode 1) results in models 1 (mean annual change for all individuals) and 2 (mean annual change by deprivation group) change little (Table 4). However, for women only, the test for trend (model 3) shows a significantly higher mean annual BMI increase with higher deprivation quintile.
Table 4 Mean annual Body Mass Index change, adjusted for age, follow-up time and baseline BMI value (screening episode 1).
All individuals
Women Men
Mean LCI UCI p Mean LCI UCI p
All (Women: n = 11,158, Men: n = 9,831) Model 1
All 0.18 0.14 0.22 <0.001 0.18 0.14 0.21 <0.001
Model 2
Affluent 0.18 0.08 0.28 <0.001 0.19 0.12 0.26 <0.001
2 0.21 0.11 0.31 <0.001 0.14 0.07 0.21 <0.001
3 0.21 0.12 0.29 <0.001 0.13 0.06 0.20 <0.001
4 0.12 0.04 0.19 0.003 0.23 0.17 0.29 <0.001
Deprived 0.21 0.11 0.30 <0.001 0.18 0.10 0.25 <0.001
Model 3
Trend 0.04 0.01 0.07 0.016 0.01 -0.01 0.04 0.331
Stratified analysis to non-obese individuals
BMI < 30 at first screening (Women: n = 9,891, Men: n = 8,729) Model 1
All 0.19 0.15 0.23 <0.001 0.16 0.13 0.19 <0.001
Model 2
Affluent 0.18 0.10 0.27 <0.001 0.17 0.11 0.24 <0.001
2 0.24 0.15 0.32 <0.001 0.12 0.05 0.18 <0.001
3 0.20 0.12 0.27 <0.001 0.12 0.07 0.18 <0.001
4 0.13 0.05 0.20 0.001 0.21 0.15 0.27 <0.001
Deprived 0.21 0.12 0.29 <0.001 0.13 0.06 0.20 <0.001
Model 3
Trend 0.01 -0.02 0.04 0.619 -0.01 -0.04 0.01 0.331
Stratified analysis to obese individuals
BMI 30+ at first screening (Women: n = 1,267, Men: n = 1,102) Model 1
All 0.16
0.03 - 0.36 0.107 0.34 0.19 0.49 <0.001
Model 2
Affluent 0.24 -0.50 0.98 0.522 0.48 0.13 0.83 0.008
2 0.29 -0.25 0.82 0.293 0.21 -0.12 0.54 0.208
3 0.32
0.14 - 0.78 0.172 0.15 -0.16 0.47 0.336
4 0.08 -0.20 0.36 0.570 0.34 0.06 0.61 0.016
Deprived 0.20 -0.12 0.53 0.222 0.46 0.23 0.70 <0.001
Model 3
Trend 0.20 0.03 0.37 0.024 0.21 0.07 0.36 0.003
Models 1–3 notes: As explained at footnote of Table 3, but all models adjusted for baseline BMI
LCI: Lower Confidence Interval; UCI: Upper Confidence Interval, BMI: Body Mass Index
When stratifying the (baseline-BMI-adjusted) analysis to individuals who were non-obese at baseline a significant effect of deprivation status on mean annual BMI change could not be shown (Table 4). Inversely, a significant effect of deprivation on BMI change was found among individuals who were obese at baseline (p = 0.024 for women and 0.003 for men).
Discussion
This study shows increasing BMI trends among adult UK individuals of both sexes who were followed up for a mean of 4.8 years chiefly during the late 1980s to mid-1990s. Although in cross-sectional analysis participants manifest the well-known socioeconomic pattern of obesity, a deprivation status effect on mean annual BMI change could not be shown, except for individuals who were obese at baseline. This finding is suggestive that the socioeconomic gradients in mean BMI and obesity in middle aged individuals are principally determined by factors operating prior to age 35, with the exception of obese individuals, in whom deprivation group differentials on BMI increase appear to continue even during middle age.
The results of this study in part contrast with previous research demonstrating that lower socioeconomic status is associated with accelerated weight gain during adulthood [11-13], or at least during part of adulthood [14]. When interpreting such differences, methodological differences in a wider range of factors, including the study population, the age of participants, the length of follow-up, the measurement (and number of measurements) of BMI, the measurement of socioeconomic status, and analytical approaches used should be taken into consideration (Table 5). In our opinion the key difference between the present and the above studies relates to the minimum baseline age of participants, which in this study is considerably higher (35 years) compared to all previous studies. The findings of the present study do not suggest that socioeconomic differentials in BMI change do not exist in general, rather that among non-obese individuals these may cease to exist after the age of 35.
Table 5 Comparative features of previous and current study about effect of socioeconomic status on BMI change.
Ref. Population (% coverage) No of individuals included in final analysis (% women) Min-Max age at baseline Min-Max follow-up (mean follow- up) No of BMI measurements/method SES measurement Whether baseline BMI adjustment (or otherwise taken into account) in analysis Main findings in relation to SES effect on BMI change Comment
11 Subset of "Whitehall II" civil servant cohort study (73%, actual coverage higher as ~4% of invited persons had moved) 2,466 W
5,507 M
(30.9%) 25 y – 25 y ~25 y – ~25 y (~25 y) 2 / 1st recalled, 2nd measured Employment grade (I-III) Yes Significant SES effect, particularly among those with largest BMI increase (i.e. > 6 kg/m2) Individuals who lost weight / BMI during follow-up were excluded.
12 Subset of Malmo Diet and Cancer Study, excluding those with history of cancer, heart attack and stroke inter alia.
Initial "invited" sample random. (NR) 5,464 W (100%) 20 y – 20 y 25 y – 53 y (36.6 y) 2 / 1st recalled, 2nd measured Employment status
Own occupational group Paternal (bread-winner) occupational group Educational attainment Yes Significant SES effect, for all different SES measures
13 Subset of the Medical Research Council National Survey of Health and Development Cohort Study (socially stratified cohort of 1946 newborns) 2,659 M + W (% W not explicitly reported in this study, originally cohort 47.5%) 20 y – 20 y 6 y (f-up 1)-23 y (f-up 4) (NR) 4 / First 2 recalled (some indication of underestimate), last 2 measured Paternal Social Class at age 14 Also Educational attainment Yes Significant childhood SES effect, even adjusting for educational attainment
14 Finnish Twin Cohort Study (89% to 1st questionnaire, follow-up q'rres coverage of 84% and 77%) 2,482 monozygotic and 5,113 dizygotic twin pairs (56% of participants were W) 18 y – 60 y 6 y (f-up 1)
15 y (f-up 2) (6 y and 15 y for f-up 1–2 respectively 3 / All self-reported (validation study proves good validity) Educational attainment Yes Significant SES effect for BMI change between 1975–1981, but no effect between 1981 and 1990
PS Borough residents (53.7% W 47.5% M, actual coverage higher as 10.8% "excluded" cases also included in denominator) 11,158 W
9,831 M (53.1% W) 35 y – 55 y 1 y – 10 y (4.8 y) 2 / Both measured Ecological (based on small area deprivation measurements) Yes Null SES effect for non-obese individuals, significant SES effect for obese individuals models adjusting for baseline BMI Stratification of analysis by baseline obesity status
Ref: Reference, No: Number, BMI: Body Mass Index, SES: Socioeconomic Status, PS: Present study, W: Women, M: Men, F-up: Follow-up, NR: Not reported, y: years.
A crucial question in relation to the generalisability of the study findings is whether screening participants were representative of the Stockport population as a whole. It has to be borne in mind that persons invited for screening excluded an important proportion of individuals with known hypertension, diabetes and other cardiovascular conditions, i.e. a "healthier" cohort of patients compared to the general population. Overall population coverage was approximately 50%, and reassuringly there were either non-significant (women) or significant but small (men) differences of coverage by deprivation group. Although the retrospective and indirect nature of the assessment of coverage (see Additional File 1) has to be borne in mind, we nevertheless believe that our sample is unlikely to have been substantially different to the overall eligible Stockport population of 35 and 55 year olds. Conservatively, it can as a minimum be assumed that the study is representative of longitudinal BMI changes in middle-aged individuals who at the time of invitation for the first screening were not known not to have an established diagnosis of hypertension, diabetes or cardiovascular disease.
An important proportion of individuals who attended for first screening were "lost to follow-up" and such individuals were more likely to be hypertensive or to have high cholesterol. This most likely reflects the fact that under the operating protocols of the Programme, such individuals were meant to be excluded from further screening invitations, in order to be offered "usual" clinical care for the management of their high cardiovascular risk. The fact that a relatively greater proportion of "lost to follow-up" individuals were obese and deprived could have potentially biased the study results. Nevertheless restricting the analysis to participants who were not obese at the initial screening did not alter the principal findings in relation to lack of a deprivation effect on BMI change. Conservatively, in stratified analysis, it could be assumed that the study is representative of longitudinal BMI changes in non-obese (at baseline) individuals.
Dual BMI ascertainment among study participants with two screening episodes was incomplete, and also differential. Individuals that could be perceived to be at higher cardiovascular risk exhibited both higher and lower levels of complete (dual) ascertainment. For example, more deprived individuals and men who were current smokers were more likely to have dual BMI ascertainment, but the opposite was also true for hypertensive individuals, and obese men. Given the non-uniform way by which cardiovascular risk status appears to have influenced dual BMI ascertainment it could be hypothesised that overall this did not introduce an important degree of bias.
Socioeconomic status can be measured directly (i.e. by measuring individuals' income, occupation or education) or indirectly, using area-based measures (i.e. based on the predominant characteristics of the population of a small area) [17]. An area-based measurement (Townsend deprivation score) was used in this study, in common with other previous UK research in the field of socioeconomic inequalities in health, because information about individual measures of socioeconomic status in participants was incomplete, and might have been less accurate. Although in theory the use of a direct individual marker of socioeconomic status may have been preferable, area-based UK deprivation indices have been shown to predict poor health outcomes at the individual level [18], including coronary heart disease [19]. The fact that the area level used was small (Census Enumeration District) diminishes the potential for misclassification error, and a recent study has shown that Enumeration District-based Townsend scores can be a valid measure of individual deprivation [20]. Although Stockport as a whole is less deprived than England and Wales, the range of Townsend deprivation score among study participants was large and comparable to the national range. It is worth noting that at baseline the study participants demonstrate a clear socioeconomic pattern of obesity. This observation supports the validity of deprivation status as an indicator of socioeconomic status among the study's participants.
Was the observed BMI increase among the study participants "natural", or could the dynamics of BMI change of the cohort have been altered by the screening process per se – and if so to what degree? If the latter is true, it is theoretically possible that the observed deprivation group differential in BMI increase among obese individuals were due to those least deprived being differentially more able to more change their dietary or energy expending behaviour following the baseline screening episode attendance. It is impossible to answer the above questions with certainty directly from the information available, and given the uncontrolled nature of the study. We however believe that the findings are much more likely to describe the "natural" BMI increase experience of the cohort, rather than a healthcare-mediated effect for three reasons. Firstly, it is apparent from the findings that screening appears not to have been overall effective in halting a further increase in BMI among obese individuals (Tables 3,4), and this would minimise the theoretical potential for a differential effect of screening by deprivation group. Secondly, although it is not possible to know with certainty how individuals found to be obese at baseline were managed (e.g. advice to lose weight), given the time of the study (1989–1993) it is likely that obesity management would have been given lesser priority as a cardiovascular risk factor, compared with hypertension, smoking and high cholesterol. Thirdly, the evidence overall suggests that, at least in the past, primary care based cardiovascular risk factor screening interventions were of limited effectiveness, [21] although this meta-analysis did not include results about a potential effect of advice on physical activity [22].
Prospective use of routinely collected data has been advocated as a method to help support surveillance and monitoring of risk factor trends in the population [23]. Although a previous meta-analysis or primary studies reported between 1977 and 1996 has questioned the overall efficacy of multiple cardiovascular risk factor screening as a means of preventing the development of cardiovascular disease [21], a potentially important secondary use of population-based screening programmes is as public health surveillance tools, to monitor population risk factor trends. This study proves that indeed there is a potential for using such data for population health monitoring and surveillance, as suggested by both the recent "Wanless Report" [24] and the subsequent UK Department of Health White Paper on Public Health [25]. It is of note that the design and conduct of a population-based cohort study of similar magnitude would have been associated with very considerable resource implications.
Conclusion
During the study period and in the study setting, the BMI of middle-aged individuals of both sexes residing in a UK district was increasing by 0.19 kg/m2 per year. The increase was similar across individuals of all deprivation groups for non-obese individuals, but was significantly higher among more deprived individuals who were also obese at baseline. The findings support that socioeconomic differences in mean BMI and obesity status are principally attained prior to 35 years of age, although among obese individuals those differentials may be further augmented during middle life. Efforts to tackle inequalities in obesity status should therefore principally concentrate in earlier life periods. Increasing BMI trends in individuals of middle age are likely to have important detrimental and multiplicative effects on the overall population burden of cardiovascular risk factors and population health in future years.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GL identified the research question with support from RFH. GL has developed the statistical methodology, with supervisory support by PMcE. PL and MH have over a number of years helped run the Stockport Cardiovascular Risk Factor Screening Programme and collect and collate data that enabled the analysis to be carried out. GL wrote the first draft of the manuscript. All authors contributed to the writing of the paper and read and approved the final manuscript. The work leading to this report has been carried out as part fulfillment for the study for the degree of MD, University of Manchester, for GL. RFH is GL's supervisor for the named degree and P McElduff GL's advisor.
Figure 2 Mean annual change in BMI (age- and follow-up- adjusted), men.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Supplementary information about the Stockport Cardiovascular Risk Factor Screening Programme 1989–1993. This file provides additional information about operational aspects of the Stockport Cardiovascular Risk Factor Screening Programme, including about exclusions from screening invitations, measurements of risk factors, and overall population coverage.
Click here for file
Additional File 2
Synthesis of information presented in Tables 1,2, in relation to deprivation group. This file provides information about the combined effect of loss to follow up and incomplete "dual" (i.e. on both screening episodes) ascertainment of BMI by deprivation group, synthesising relevant information from the Results section and Tables 1 and 2
Click here for file
Acknowledgements
We would wish to acknowledge the help, advice and practical support of the following individuals: Prof. Deborah Baker, Dr Stephen Watkins, Ms Jane Jefferson, Ms Jane Pilkington, Dr Gill Greenhough, Ms Jane Bowdenleigh, Dr Gill Burroughs, Ms Bernadette Ryan-Wooley, Ms Barbara Shallaker, Mr Dan Byrne, Ms Eleanor Bannister.
==== Refs
The United Kingdom Parliament Ninth Report: Tackling Obesity In England accessed November 2004
Silventoinen K Sans S Tolonen H Monterde D Kuulasmaa K Kesteloot H Tuomilehto J WHO MONICA Project Trends in obesity and energy supply in the WHO MONICA Project Int J Obes Relat Metab Disord 2004 28 710 8 15007395 10.1038/sj.ijo.0802614
Lindstrom M Isacsson S-O Merlo J Increasing prevalence of overweight, obesity and physical inactivity. Two population-based studies 1986 and 1994 Eur J Public Health 2003 13 306 312 14703316 10.1093/eurpub/13.4.306
Bundred P Kitchener D Buchan I Prevalence of overweight and obesechildren between 1989 and 1998: population based series of cross sectional studies BMJ 2001 322 326 8 11159654 10.1136/bmj.322.7282.326
Wardle J Jarvis MJ Steggles N Sutton S Williamson S Farrimond H Cartwright M Simon AD Socioeconomic disparities in cancer-risk behaviours in adolescence: baseline results from the Health and Behaviour in Teenagers Study (HABITS) Preventive Medicine 2003 35 721 730 12744916 10.1016/S0091-7435(03)00047-1
Reilly JJ Dorosty AR Epidemic of obesity in UK children Lancet 1999 354 1874 5 10584727 10.1016/S0140-6736(99)04555-9
Rudolf MC Sahota P Barth JH Walker J Increasing prevalence of obesity in primary school children: cohort study BMJ 2001 322 1094 1095 11337437 10.1136/bmj.322.7294.1094
Chinn S Rona RJ Prevalence and trends in overweight and obesity inthree cross sectional studies of British Children, 1974–94 BMJ 2001 322 24 26 11141148 10.1136/bmj.322.7277.24
Molarius A Sidell JC Sans S Tuomilehto J Kuulasmaa K Educational level, relative body weight and changes in their association over 10 years: an international perspective from the WHO MONICA project American J Public Health 2000 90 1260 1268
Boreham R Erns B Falaschetti E Hirani V Primatesta P Risk factors for cardiovascular disease Health Survey for England-Cardiovascular Disease 1999 3 UK: The Stationery Office
Martikainen PT Marmot MG Socioeconomic differences in weight gain and determinants and consequences of coronary risk factors Am J Clin Nutr 1999 69 719 26 10197574
Lahmann PH Lissner L Gullberg B Berglund G Socioedemographic factors associated with long-term weight gain, current body fatness and central adiposity in Swedish women Int J Obes Relat Metab Diord 2000 24 685 94 10.1038/sj.ijo.0801219
Hardy R Wadsowrth M Kuh D The influence of childhood weight and soeciocnomic status change in adult body mass index in a British national birth cohort Int J Obes Relat Metab Disord 2000 24 725 34 10878679 10.1038/sj.ijo.0801238
Silventoinen K Sarlio-Lahteenkorva S Koskenvuo M Lahelma E Kaprio J Effect of environmental and genetic factors on education-asssociated disparities in weight and weight gain: a study of Finnish adult twins Am J Clin Nutr 2004 80 815 22 15447885
Baker D Hann M Evaluation of the Stockport Cardiovascular Disease Screening Database as a tool for research and evaluation projects National Primary care Research and Development Centre, University of Manchester 2002
Townsend P Phillimore P Beattie A Health and deprivation: inequality and the north 1988 London: Croom Helm
Liberatos P Link BG Kelsey JL The measurement of social class in epidemiology Epidemiolo review 1988 10 87 121
Carstairs V Deprivation indices: their interpretation and use in relation to health J Epidemiol Community Health 1995 49 S3 8 8594130
Woodward M Small area statistics as markers for personal social status in the Scottish Heart Health study J Epidemiol Community Health 1996 50 570 6 8944867
Adams J Ryan V White M How accurate are Townsend Deprivation Scores as predictors of self-reported health? A comparison with individual level data J Public Health 2005 27 101 6 10.1093/pubmed/fdh193
Ebrahim S Smith GD Systematic review of randomised controlled trials of multiple risk factor interventions for preventing coronary heart disease BMJ 1997 314 1666 74 9193292
Batty D Review of interventions to prevent heart disease BMJ 1997 315 1468 9418119
Bundred P Kitchener D Buchan I Prevalence of overweight and obese children between 1989 and 1998: population based series of cross sectional studies BMJ 2001 322 326 8 11159654 10.1136/bmj.322.7282.326
Wanless D Treasury HM Securing good health for the whole population: Final report-February 2004 Accessed March 2005
Department of Health Choosing Health. Making healthy choices easier 2004 HMSO, London Accessed March 2005
| 15811178 | PMC1090593 | CC BY | 2021-01-04 16:28:58 | no | BMC Public Health. 2005 Apr 5; 5:32 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-32 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-331581118910.1186/1471-2458-5-33Research ArticleAttitudes to and management of fertility among primary health care physicians in Turkey: An epidemiological study Hassa Hikmet [email protected] Unal [email protected] Ilhami [email protected] Selma [email protected] Alaeddin [email protected] Medical Faculty, Gyneocology Department, Osmangazi University, 26480 Meselik-Eskisehir, Turkey2 Medico-Social Center, Osmangazi University, 26480 Meselik-Eskisehir, Turkey3 Medical Faculty, Family Medicine Department, Osmangazi University, 26480 Meselik-Eskisehir, Turkey4 Medical Faculty, Public Health Department, Osmangazi University, 26480 Meselik-Eskisehir, Turkey2005 5 4 2005 5 33 33 21 12 2004 5 4 2005 Copyright © 2005 Hassa et al; licensee BioMed Central Ltd.2005Hassa 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 subject of infertility has taken its place in the health sector at the top level. Since primary health care services are insufficient, most people, especially women, keep on suffering from it all over the world, namely in underdeveloped or developing countries.
The aim of this study was to determine primary care physicians' opinions about the approach to infertility cases and their place within primary health care services (PHCSs).
Methods
The study was conducted between October 2003 and April 2004. The study group comprised 748 physicians working in PHCSs. They were asked to fill in a questionnaire with questions pertaining to infertility support, laboratory and treatment algorithms, as well as the demographic characteristics. The data was evaluated using the chi square test, percentage rates and a logistic regression model.
Results
The multivariate analyses showed that having a previous interest in infertility and having worked for a postgraduate period of between 5–9 years and ≥10 years were the variables that most positively influenced them in their approach to cases of infertility (p < 0.05, each one).
Just 28.7% of the physicians indicated that they believed cases of infertility could be evaluated at the primary care level. The most frequently proposed reason for indicating 'difficulty in practice' (n = 533) was inadequate provision of equipment in PHCSs (55.7%). The physicians reported that they were able to perform most of the supportive treatments and proposals (between 64.6%–87.7%). The most requested laboratory investigations were the instruction of patients in taking basal body temperatures and semen analysis (89.7% and 88.7%, respectively). The most preferential course of treatment was that of sexually transmitted diseases (95.5%).
Conclusion
It is clear that not enough importance is attached to the provision of care to infertile couples within PHCSs. This leads us to conclude that an integration of infertility services in primary care would be appropriate after strengthening the PHCSs.
==== Body
Background
Infertility affects between 8–20% of couples around the world at some stage during their reproductive years. As the figures suggest, it has become a common health and communal problem in terms of not only the couples involved, but also their health team and the social environment [1-3].
The level and patterns of infertility apparently vary widely, being found less in developed countries, and more in underdeveloped or developing countries [4].
Although the damage that infertility causes to health is generally emotional, it can also have a negative effect on an individual's health, quality of life, and life expectancy [5]. Thus, social and psychological support as well as medical diagnosis and treatment is of great importance while treating couples with infertility [6].
The likelihood of a couple presenting to their primary health care physicians (PHCPs) for the first time with fertility concerns, i.e. wanting to conceive a child but having difficulty, at some point during their reproductive lives is high [7]. PHCPs must decide on the circumstances under which these patients should be referred to specialist reproductive health centers [8,9]. Even when PHCPs refer these cases to specialists, they are still expected to provide counseling and information for patients in every aspect of Assisted Reproductive Technologies, such as in vitro fertilization and intracytoplasmic techniques [10]. This places PHCPs at the heart of all issues relating to infertility [3,6,11,12].
The 21st World Health Assembly (1968) expressed the opinion that every family should have the opportunity of obtaining information and advice on problems connected with family planning, including fertility and infertility, and suggested the integration of these services within primary health care services (PHCSs) [13]. Similarly, in the Alma Ata Declaration of 1978, emphasis was placed on the World Health Organization's role in strengthening PHCSs and the correlation of infertility services with primary health care (PHC) [14].
As with other types of disease, not enough importance has been given to the provision of infertility services by PHC in Turkey, a state reflected in most developing countries [15]. Since Turkey lacks a PHC physician system, it has yet to form a standard in other disciplines, as well as in its approach to cases of infertility. However, in developed countries, PHCPs are in a unique position to provide patient education, offer ongoing counseling, and give psychological help to couples who experience problems with fertility [16].
The purpose of this study was therefore to contribute to the formation of a model for the integration of infertility services within PHCSs, and to both determine PHCPs' opinions about the approach to infertility cases and their place within PHCSs.
Methods
Sampling and subjects
This study was carried out in 7 cities, namely Eskisehir, Konya, Bilecik, Bolu, Duzce, Kutahya and Afyon, selected by a random sampling method from 18 cities in Central Anatolia, Turkey during 6 months between October 2003 and April 2004. 1,212 PHCPs were working in 110 PHCSs in the seven cities during the study. PHCPs were recruited randomly during routine visits to local health institutions. The sample capacity to constitute the study was estimated as 384 physicians, assuming that the occurring frequency of infertility was 50%. In the study, since the cluster sampling method was used as the sampling method, the sample volume was determined as 768, multiplying by two the number of 384. These physicians were then further selected by a random sampling method as proportional with the number of the physicians in those cities. All 748 of the physicians contacted agreed to fill in the questionnaires (100%). A total of 20 physicians could not participate in the study due to the taking leave or being absent at the time of the study.
The interview schedules
The dates on which the study would be conducted were determined in cooperation with government health officials in the cities concerned. The PHCSs were visited with an official of health authority. The physicians completed questionnaires in the presence of a member of the research team, with the researchers being on hand to explain any questions that the physicians found incomprehensible.
Development of the questionnaires
In the first part of the questionnaire, the physicians were asked to state their demographic characteristics. The second part of the questionnaire, which was prepared with the benefit of reference to previous studies [17-20] and according to infertility treatment algorithms, included three phases. In the first phase, PHCPs were asked to reply to questions concerning to circumstances of providing supportive treatment and proposals to infertile individuals; in the second phase, which laboratory support was requested; and in the third phase, the things to be done during treatment.
Data analysis
The data were evaluated by SPSS with the chi square (χ2) test and percentage (%) ratios where possible. The evaluation of primary care infertile cases by PHCPs was accepted as the dependent variable. Independent variables, which can affect the dependent variable, were determined using a logistic regression model, which is used as a multivariate analysis method. The variables which produced significance of p < 0.05 in the univariable analyse were applied to the model. The physician's age, sex, working time, and the status of having previously evaluated a case of infertility were chosen as independent variables.
Informed consent for the study
Local authorities, such as the Osmangazi University School of Medicine and the health authorities in the cities concerned, approved this study. All physicians gave their informed consent prior to their inclusion in the study.
Results
The average age of the physicians included was 33.4 ± 5.5. The rates of the male and female PHCPs were 52.1% (390) and 47.9% (358), respectively. The average working period of the physicians was 8.6 ± 5.2 years.
When the physicians were asked about the proportion of applications to receive advice for an infertility concern from within the total number of applications, 124 (16.6%) responded 0%; 491 (65.6%) for 1–3%; 95 (12.7%) for 4–5%; and 38 (5.1%) for 6% and over.
Just 28.7% of the physicians indicated that infertile cases could be evaluated at PHC. The most frequently stated reason given with an indication of 'practice difficulty' (n = 533) was a lack of devices/equipment at PHCSs (55.7%). Detailed data on the approach to patients with infertility problems is shown in Table 1.
Table 1 The physicians' attitudes about evaluation of infertile cases at primary care and the reasons that were put forward by those who indicated that there was practice difficulty
Attitudes n = 748(100%)† 95% CI
Those believing that infertile cases could be evaluated at primary care 215 (28.7) 25.7–31.7
Those believing that infertile cases could be evaluated at primary care but that the application would prove difficult 76 (10.2) 8.0–12.4
Those believing that infertile cases cannot be evaluated at primary care level 457 (61.1) 59.3–62.9
Reasons proposed by those indicating that there was difficulty in application* n = 533(100%)† 95% CI
Supply of logistics is inadequate (lack of device and equipment at primary care) 297 (55.7) 51.5–59.9
Only specialists should evaluate 179 (33.6) 31.6–35.6
My level of education in this field is insufficient 149 (30.0) 26.1–33.9
There is not enough time at primary care 46 (8.6) 6.2–11.0
It is a loss of time for patient 43 (8.1) 5.8–10.4
†Since more than one reason is proposed, the proportion exceeds 100%. * Those who believe that while infertile cases could be evaluated at primary care the practice could prove difficult (n = 76) and those who believe that the practice would not be possible at primary care (n = 457)
Table 2 shows the opinions of PHCPs related to the evaluation of cases with infertility. With the exception of items 11., 12. and 13. (p > 0.05, each one), more physicians indicated that PHC was an appropriate place for evaluation than did those who did not. These majorities were significant in 6 of the 7 items (items 1., 3., 4., 5., 6. and 7.) related to support, treatment and proposals. However, in the group related to the use of laboratory investigation, only four of the twelve items showed a significance (items 8., 9., 14., and 16.). The opinions of the physicians on treatment were significant for all the items (p < 0.05, each one).
Table 2 Primary health care physicians' opinions related to evaluation of infertile cases at primary care
Physicians' opinions concerning supportive treatment and proposals for infertile couples Physicians' opinions related to evaluation of infertile cases at primary care Total
Appropriate* n = 291(38.9%) Inappropriate** n = 457(61.1%) n(%) 748(100) p
1.I could administer rubella prophylaxis 209(72.2) 295(64.6) 504(67.4) 0.030
2.I could begin folic acid support 262(90.0) 401(87.7) 663(88.6) 0.336
3.I can encourage couples to avoid cigarettes, alcohol and drug abuse 258(88.7) 376(82.3) 634(84.8) 0.018
4.I can resolve obesity problems 251(86.3) 348(76.1) 599(80.1) 0.001
5.I can prevent testicular hypertermia by advising on appropriate clothing to be worn 266(91.4) 375(82.1) 641(85.7) 0.000
6.I can inform of coit order 269(92.4) 381(83.4) 650(86.9) 0.000
7.I can investigate the distress that childlessness causes 255(87.6) 349(76.4) 604(80.7) 0.000
Physicians' opinions concerning request for laboratory investigations at primary care level in evaluating infertility in couples
8.I have semen analyses done 258(88.7) 368(80.5) 626(83.7) 0.003
9.I evaluate one value progesterone hormone for ovulation between the 22. and 24. days of cyclus 217(74.6) 309(67.6) 526(70.3) 0.042
10.I request an evaluation of FSH, LH, E2 and Prolactin on the 2. and 4. days of the male's cycle 193(66.3) 292(63.9) 485(64.8) 0.498
11.I perform an ultrasonic folliculometric ovulation follow-up 118(40.5) 204(44.6) 322(43.0) 0.271
12.I diagnoses policystic ovary disease by means of folliculometric measure 117(40.2) 211(46.2) 328(43.9) 0.109
13.I diagnoses uterus anomalies by ultrasound 123(42.3) 226(49.5) 349(46.7) 0.055
14.I teach the patient to measure basal body temperature 261(89.7) 373(81.6) 634(84.8) 0.003
15.I investigate thyroid functions if the result of a physical exam is positive 241(82.8) 363(79.4) 604(80.7) 0.252
16.I investigate prolactine levels if there is a history of galactore history or if a physical exam is positive 238(81.8) 346(75.7) 584(78.1) 0.050
17.I have patient's adrenal hormones investigated if the results of hirsutismus or a physical exam are positive 226(77.7) 331(72.4) 557(74.5) 0.109
18.I evaluate vaginal or urethral discharge by microscope in cases with complaint 194(66.7) 284(62.1) 478(63.9) 0.209
19.I ask whether or not hysterosalpingographic study has previously been conducted and if tubes were open 194(66.7) 300(65.6) 494(66.0) 0.774
Physicians' opinions about treatment of infertile cases
20. I can treat sexually transmitted diseases 278(95.5) 412(90.2) 690(92.2) 0.007
21.I can guide patients I have diagnosed as infertile to a higher healthcare level, and can correlate a follow- up treatment for patients with that center 270(92.8) 388(84.9) 658(88.0) 0.001
22.I can administer clomiphen citrate treatment for ovulation 107(36.8) 136(29.8) 243(32.5) 0.046
23.I can administer bromocriptin in cases with hyperprolactinemia 123(42.3) 142(31.1) 265(35.4) 0.002
24.I can perform the treatment of hyperandrogenemia 92(31.6) 103(22.5) 195(36.6) 0.006
25.I can administer metphormine derived drugs for cases with policystic ovary disease 111(38.1) 137(30.0) 248(33.2) 0.021
26. I can administer gonodotrophinler for ovulation 94(32.3) 113(24.7) 207(27.7) 0.024
27.I can perform hormonal treatment for male infertility 94(32.3) 102(22.3) 196(26.2) 0.002
28. I can perform insemination with a males split ejaculation 59(20.3) 56(12.3) 115(15.4) 0.003
*Those who believed that infertile cases could be evaluated at primary care (n = 215) and Those who believed that while infertile cases could be evaluated at primary care the practice could prove difficult (n = 76), ** Those who think that infertile cases could not be evaluated at primary care (n = 457)
The result model of the logistic regression analyse is presented in Table 3. 40.2% (164/408) of the physicians working in cities with infertility centers, and 37.4% (127/340) of those who did not, reported that they saw appropriate the evaluation of infertile couples (p = 0.427). This proportion was reported at 45.4% (49/108) for physicians who had received education related to reproductive health after graduating, whereas the figure was 37.8% (242/640) for those who had received no education (p = 0.136); 36.2% (71/196) for physicians having graduated from medical faculties with have infertility centers, but 39.9% (220/552) for those who had not (p = 0.370) (Unshown data).
Table 3 Multivariate analysis results for the relationship between appropriate and inappropriate behaviour to infertile cases according to the demographic characteristics of the primary health care physicians
Demographics Those seen as appropriate n = 291(38.9%) Those seen as inappropriate n = 457(61.1%) Total n = 748 (100%) Odds Ratio (95% Confidence Interval) p
Sex
Female 123(34.4) 235(65.6) 358(47.9) 1
Male 168(43.1) 222(56.9) 390(52.1) 1.27 (0.93–1.74) 0.126
Age groups
Age 24–29 years 63(30.3) 145(69.7) 208(27.8) 1
Age 30–39 years 183(40.4) 270(59.6) 453(60.6) 0.784 (0.473–1.30) 0.345
Age ≥40 years 45(51.7) 42(48.3) 87(11.6) 0.974 (0.475–1.99) 0.942
Post graduation period
1–4 years 43(25.3) 127(74.7) 170(22.7) 1
5–9 years 107(38.1) 174(61.9) 281(37.6) 1.785 (1.04–3.05) 0.034
≥10 years 141(47.5) 156(52.5) 297(39.7) 2.418 (1.298–4.502) 0.005
Previously dealing with the management of an infertile couple
No 123(35.6) 177(64.4) 345(46.1) 1
Yes 168(41.7) 280(58.3) 403(53.9) 1.870 (1.368–2.554) 0.034
The independent variable most influencing the physicians' opinions on this subject was having previously dealt with such a case during their occupational career (p < 0.05) and having postgraduate experience in the field of PHC for a period of 5–9 years and ≥10 years (p < 0.05, each one).
46.1% (345/748) of the PHCPs reported that they had been involved in the evaluation/treatment of an infertile couple or couples. While 4.3% (15/345) of the physicians reported having had no involvement in the evaluation of infertile couples, the data provided by the majority of the physicians (95.7%, 330/345) for the procedures most frequently administered to patients are given in table 4. According to this data, more than half of the procedures consisted of the referral of patients to a higher tier health center (57.6%), the provision of information and counseling (58.5%), followed by requesting that laboratory and radiological investigations be done (27.9%).
Table 4 The practices of physicians who had been involved in provision of services to infertile cases until the time of the study
Total n = 330(100%)*
Provision of information and counseling (concerning iron deficiency, menstruation disorders, folic acid deficiency, ovulation periods, alcohol-cigarette-coffee drinking, coit order, behaviour pertaining to raising sperm quality, basal body temperature) 193(58.5)
Reference to a secondary healthcare center 190(57.6)
Request laboratory and radiological investigations (PRL, E2, FSH, LH, Thyroid tests, USG, Sperm analysis) in order to diagnose the reason for infertility (Galactorhea, hirsutism, polycystic ovary syndrome) 92(27.9)
Treatment of sexually transmitted diseases 36(10.9)
Psychological support to decrease the distress of infertile couple 35(10.6)
Case history and physical examination 28 (8.5)
Hormonal treatment for infertility (gonadotropine treatment, drug treatment for ovulation, bromocriptine treatment, polycystic ovary syndrome treatment, hyperandrogen treatment) 17 (5.2)
Performance of a follow-up for patients returning to primary care after referral to a higher health center 15 (4.5)
*Since the physicians indicated that they had performed more than one treatment, the total proportion exceeds 100%.
6.9% (23/330) of the physicians indicated that they had not experienced any difficulties during the approach to their patients. The various difficulties of 307 (93.1%) of the physicians stating having experienced a problem during consultation are shown in Table 5. According to the information provided, the most widely reported difficulty physicians had experienced was an inadequacy of logistics (35.2%), followed by feelings of despondency on the patient's part (32.2%).
Table 5 Difficulties physicians had experienced during approach
n = 307(100%)*
Inadequate supply of logistics (lack of device-equipment-staff) 108(35.2)
Patient's lack of harmony (psychological problems, illogical behaviour, couples not attending together, low cultural level of couples, couples remained unconvinced) 99(32.2)
Management mistakes/errors (the absence of communication between primary health care and hospitals, the absence of a comfortable atmosphere in which patients with infertility can speak) 72(23.5)
Physician's lack of education/knowledge-physician's lack of post-graduate education 71(23.1)
Patients' not having health insurance, money problems of patients 27 (8.8)
Lack of belief and confidence in the primary health care physician 17 (5.5)
*Since more than one reason is proposed, the proportion exceeds 100%.
Table 6 reports suggestions forwarded by PHCPs' for the provision of better service to infertile couples. The physicians most commonly recommended the encompassment of infertility matters in family planning services, as well as the introduction of an on-going postgraduate education programme (73.1% and 70.1%, respectively).
Table 6 Proposals forwarded by primary health care physicians for the provision of better service to infertile cases
Practices n = 748(100%)*
Inclusion of the subject of infertility in Family planning services 547(73.1)
A planned post-graduate continuing education programme 524(70.1)
A strengthening of routes of communication between primary care and other care services 520(69.5)
Improvement of laboratory conditions 503(67.2)
More support for this subject in graduation education 398(53.2)
*Since more than one reason is proposed, the proportion exceeds 100%.
Discussion
This study is the first large scale Turkish epidemiological study conducted to learn PHCP's management of infertile patient, their attitudes to fertility treatment, and their recommendations for the provision of better service in the future.
In our study, a near absence of awareness on the part of doctors that nearly 1/5 (16.6%) of all patients entering PHC applied for consultation for fertility problems, and that about 2/3 (65.6%) of the same physicians reported that they believed that the issue of infertility did not have a place during routine examination is a cause for concern.
In this study, the proportion of the physicians who viewed the evaluation of infertile cases as appropriate in PHCSs was 38.9%. This proportion was reported at around 27.0% in a study by Ittner et al (2000) [21]. That the proportions are less than 50% shows that the physicians in different countries have similar thoughts about the evaluation of infertility in PHC.
In the current study, most physicians reported that they could perform a large part (67.4%-88.6%) of the necessary supportive treatments and consultations directed at fertility in the preconceptional period. In parallel to this finding, those physicians who had been involved in cases of infertility during their working life stated that they had performed 82.4% of supportive treatment and recommendations, including offering psychological support (10.9%), the taking of patient histories and physical examinations (8.5%), and the following- up of patients returning after having been referred to higher-level healthcare centers (4.5%) (Table 4). These results reveal a balance between what the physicians reported they did and what they would do. In the study by Heyes et al (2004) [22], a proportion of between 79.1% and 86.7% of PHCPs indicated the practicality of performing support treatment and recommendations and that the most suitable location was within PHC, whereas the proportion of physicians stating that a hospital setting was the most appropriate place for the performance of these duties was only 1.2%.
In this study, the physicians reported that the area of supportive treatment and recommendations for treatment they least preferred was concerning rubella prophylaxis (67.4%). By means of a suggestion for this low proportion, we can offer the explanation that rubella vaccination in Turkey is not included in routine vaccination schema, and that, therefore, physicians do not have much knowledge of this vaccination. However, Turkish studies on cord blood indicate that 15% of those who will become mothers are sensitive to rubella [23]. This demonstrates the need for fortification of knowledge on rubella prophylaxis at the level of the physicians.
In this study, a great majority of physicians reported that in terms of the provision of laboratory support they were able to perform an investigation of semen analyses and of a one-value progesterone hormone for ovulation (83.7% and 70.3%, respectively). However, areas that they felt less confident in were performing such investigations as folliculometric ovulation follow-up by ultrasound, the diagnosis of polycystic ovarian disease by means of folliculometric measurement, and the diagnosis by ultrasound of uterine anomalies (between 43% and 46.7%). These findings are consistent with those of other studies on distribution of duties performed by PHCPs [5,11,15].
In this study, the opinions of both sets of physicians on the evaluation of cases of infertility in PHCSs were positive as expected. The proportions of the physicians responding that they were able to perform treatment of sexually transmitted diseases and referral and follow-up of patients diagnosed as infertile were rather high (92.2% and 88.0%, respectively), whereas those reporting being able to perform more heavy treatments, regarded as secondary care duties, such as ovulation induction was rather low (between 36.6% and 15.4%). These proportions show that PHCPs working in the region have sufficient reproductive health knowledge and that when physicians are presented with appropriate conditions they are able to perform follow-up to and treatment of infertility.
The proportion of physicians in this study reporting that they could evaluate vaginal or urethral discharge through microscopic investigation was less than that of physicians able to offer treatment for sexually transmitted diseases (63.9% and 92.2%, respectively). This result is important from the viewpoint that the approach offered to patients by physicians in the region is to directly use empirical treatment without performing detailed investigations, and that there was a big gap in adequacy of laboratory equipment available in PHCSs.
In this survey, it was determined that approximately two thirds of the physicians were not comfortable performing procedures usually offered in secondary care, such as performing insemination with a split ejaculation (items 22. to 28.). In line with our study results, some guides [24] suggested that if criteria related to secondary care, such as those mentioned above were present, such patients should be referred to higher-level health centers.
In our study, a rather low proportion of physicians reported that they were able to administer clomiphen citrate and bromocriptin (32.5% and 35.4%, respectively) to their patients. In some studies [25], although emphasizing that these medicines could be administered at PHC level, the fact that this approach is rather new and not approved all over the world, may indicate why the physicians in our study may have marked these alternatives less.
Physicians reported that having a previous interest in infertility and having worked for a longer postgraduate period were the important independent variables that most positively influenced them in their approach to cases of infertility. This finding shows that, as the time passed after graduation and the number of infertility cases dealt with increases, so do the physicians' levels of confidence.
In this survey, more than half of the physicians who had been involved in cases of infertility (57.6%) reported that they directly referred infertile patients to a higher-level healthcare provider. The result of a study conducted in Germany is in line with our study result, were the proportion was reported as 65% [21]. In the same study [21], it was found that 43% of the physicians did not know of any patient in their practice who had had fertility problems. In a similar study [26], a majority of the physicians placed infertility within the realm of fertility specialists.
In the present study, the most frequent reason seen for physicians reporting the referral of infertile patients was a lack of logistics in the PHCSs (35.2%, 108/307). This reason, taking into consideration the health care system in our country, reveals that even when physicians want to perform treatment of patients in the primary healthcare system, the system is unable to deal with the demand. In a study by Heyes et al (2004) [22], the most reported reason for experiencing difficulties while evaluating cases at primary care level was due to service access problems and resource constraints
In our study, the second most frequently reported reason was the patient's lack of harmony (32.2%). This finding is consistent with other study results [21,26-28]. The study by Ittner et al (2000), [21] in particular reported that almost the same proportion of physicians emphasized the influence of fertility problems on personality, sexuality, social acceptance, and mental health status.
A great proportion of the difficulties that physicians experienced during their approach arose from reasons not directly connected to the patient. If lack of knowledge on the part of the physician, inadequacy in the supply of logistics, and managerial errors were to be removed, an increase in information being given to the patient and a 32.2% decrease in patients experiencing feelings of despondency may be anticipated.
Of the difficulties that physicians reported having encountered during consultation, a further important reason that emerged was lack of communication between health institutions (23.5%). In our country, since there is still no coordination between health institutions, this absence negatively affects the follow-up of patients with infertility. If higher tier health centers were to inform PHCSs of the treatment administered, PHCPs would have an increase in self-confidence.
In our study, three-quarters of all the PHCPs while reporting in the recommendations section mentioned the importance of education both before graduation and post graduation. In a similar way to the reports of other studies [22,29], the emphasis was placed on providing across the board training to all primary health care workers in order to enable them to deliver this care more confidently and raise awareness and skills.
Conclusion
The concern held by many that cases are not given due care and attention, particularly care offered to couples with infertility, within PHCSs [26-30], also appeared in this study. Our conclusion, based on the recommendations provided by physicians working in PHC, was that the implementation of an education programme, thus boosting self-confidence, would have the greatest impact on changing the attitudes of physicians towards cases of infertility.
Since we have not yet been able to find any study conducted on this subject in Turkey, we recommend that further comprehensive research is needed in order to expose PHCPs' thoughts and attitudes towards infertility.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HH conceived of the study, AU participated in its design and coordination, sequence alignment, collected the data and drafted the manuscript, UI participated in its coordination and collected the data, MS participated in the design of the study, performed the statistical analyses and collected the data, UA participated in its 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 thank all the PHCPs from the cities who participated in this survey, and the health authorities of the cities who helped in the collection of data.
==== Refs
World Health Organization Report of the Meeting on the Prevention of Infertility at the Primary Health Care Level 1984 WHO, Geneva 12–16 December 1983. WHO/MCH/84.4.
Zargar AH Wani AI Masoodi SR Laway BA Salahuddin M Epidemiologic and etiologic aspects of primary infertility in the Kashmir region of India Fertil Steril 1997 68 637 43 9341602 10.1016/S0015-0282(97)00269-0
Whitman-Elia GF A primary care approach to the infertile couple J Am Board Fam Pract 2001 14 33 45 11206691
Cates W Farley TM Rowe PJ Worldwide patterns of infertility: is Africa different? Lancet 1985 2 596 598 2863605 10.1016/S0140-6736(85)90594-X
Jones HW Tomer JP The infertile couple N Engl J Med 1993 329 1710 5 8166801 10.1056/NEJM199312023292307
Meniru GI Hetcht BR Hopkins MP Comprehensive preventive health care in female-factor infertility: issues, problems and proposals Prim Care Update Ob/Gyns 2002 9 44 50 10.1016/S1068-607X(01)00102-0
Collins JA A couple with infertility JAMA 1995 274 1159 64 7563488 10.1001/jama.274.14.1159
Grimshaw J Emslie C Managing infertility-the GP's role Practitioner 1994 238 586 90 8072942
Hanson MA Dumesic DA Initial evaluation and treatment of infertility in a primary care setting Mayo Clin Proc 1998 73 681 5 9663199
Schenker JG Ezra Y Complications of assisted reproductive techniques Fertil Steril 1994 61 411 22 8137959
Wilkes S Jones K Retrospective review of the prevalence and management of infertility in women on one practice over a 5-year period Br J Gen Pract 1995 45 75 7 7702886
Brundage SC Preconception health care Am Fam Physician 2002 65 2507 14 1521-2 12086240
World Health Organization. Twenty-first World Health Assembly, Part I-Resolotions and decisions, annexes. WHA21:43-Health aspects of populationdynamics. Official Records of the World Health Organization 1968 168 21 22
World Health Organization Primary health care. International Conference on Primary health care. International Conference on Primary Health Care, Alma-Ata, USSR, 6–12 September 1978. A joint report by the World Health Organization and the Executive Director of the United Nations Children's Fund, Geneva 1978
Rowe PJ Clinical aspects of infertility and the role of health care services Reproductive Health Matters 1999 7 103 111 10.1016/S0968-8080(99)90119-2
Whitman-Elia GF Baxley EG A primary care approach to the infertile couple J Am Board Fam Pract 2001 14 33 45 11206691
Elstein M Vaughan Williams CA Shalet SM Clinical algorithms: infertility BMJ 1987 294 1550 3111630
Emslie C Grimshaw J Templeton A Do clinical guideliness improve general practice management and referral of infertile couples BMJ 1993 306 1728 31 8280213
Blackwell RE A proposed algorithm for evaluation and treatment of infertility: update Infertil Reprod Med 1995 7 1 8
Morell V Basic infertility assessment Prim Care 1997 24 195 204 9016736
Ittner E Himmel W Kochen M German Family Physicians' attitudes toward care of involuntarily childless patients Fam Med 2000 32 119 25 10697771
Heyes T Long S Mathers N Preconception Care Fam Prac 2004 21 22 27 10.1093/fampra/cmh106
Cakici C Aka N Yorulmaz S Acar N Gokmen B Gebelerde Rutin Olarak Toksoplazma, Rubella ve Sitomegalovirus Taramasý Yapilmali midir? Jinekoloji-Obstetrik 1995 5 14 18
Belisle S Belisle S, Pierson R Introduction. CFAS Consensus document for the investigation of infertility by first-line physicians 2002 Montreal, Que: Canadian Fertility and Andrology Society 8 9 Accessed 2003 September 12.
Smith LF The role of primary care in infertility management Hum Fertil (Camb) 2003 6 S9 S12 12869770
Weingarten MA Reinitz A Hart J Attitudes to primary care gynaecology among family physicians and gynaecologists in Israel Scand J Prim Health Care 1992 10 36 41 1589662
Bagshawe A Taylor A ABC of subfertility. Counselling BMJ 2003 327 1038 40 14593042 10.1136/bmj.327.7422.1038
Marvel MK Morphew PK Levels of family involvement by resident and attending physicians Fam Med 1993 25 26 30 8454120
Wallace M Hurwitz B Preconception care: who needs it, who wants it, and how should it be provided? Br J Gen Pract 1998 48 963 966 9624765
Smilkstein G The family in family medicine revisited, again J Fam Pract 1994 39 527 31 7798855
| 15811189 | PMC1090594 | CC BY | 2021-01-04 16:28:57 | no | BMC Public Health. 2005 Apr 5; 5:33 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-33 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-361583314010.1186/1471-2458-5-36Research ArticlePredictors of glycemic control among patients with Type 2 diabetes: A longitudinal study Benoit Stephen R [email protected] Regina [email protected] Athena [email protected] Ming [email protected] University of California, San Diego – San Diego State University Preventive Medicine Residency, Dept. of Family and Preventive Medicine, MC-0811, 9500 Gilman Drive, La Jolla, CA 92093-0811, USA2 The Whittier Institute for Diabetes, Scripps Health, 9894 Genesee Ave., La Jolla, CA 92037, USA3 San Diego State Graduate School of Public Health, 5500 Campanile Drive, San Diego, CA 92182-4162, USA2005 17 4 2005 5 36 36 25 10 2004 17 4 2005 Copyright © 2005 Benoit et al; licensee BioMed Central Ltd.2005Benoit 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
Diabetes is the sixth leading cause of death and results in significant morbidity. The purpose of this study is to determine what demographic, health status, treatment, access/quality of care, and behavioral factors are associated with poor glycemic control in a Type 2 diabetic, low-income, minority, San Diego population.
Methods
Longitudinal observational data was collected on patients with Type 2 diabetes from Project Dulce, a program in San Diego County designed to care for an underserved diabetic population. The study sample included 573 patients with a racial/ethnic mix of 53% Hispanic, 7% black, 18% Asian, 20% white, and 2% other. We utilized mixed effects models to determine the factors associated with poor glycemic control using hemoglobin A1C (A1C) as the outcome of interest. A multi-step model building process was used resulting in a final parsimonious model with main effects and interaction terms.
Results
Patients had a mean age of 55 years, 69% were female, the mean duration of diabetes was 7.1 years, 31% were treated with insulin, and 57% were obese. American Diabetes Association (ADA) recommendations for blood pressure and total cholesterol were met by 71% and 68%, respectively. Results of the mixed effects model showed that patients who were uninsured, had diabetes for a longer period of time, used insulin or multiple oral agents, or had high cholesterol had higher A1C values over time indicating poorer glycemic control. The younger subjects also had poorer control.
Conclusion
This study provides factors that predict glycemic control in a specific low-income, multiethnic, Type 2 diabetic population. With this information, subgroups with high risk of disease morbidity were identified. Barriers that prevent these patients from meeting their goals must be explored to improve health outcomes.
==== Body
Background
Approximately 13 million people have been diagnosed with diabetes in the United States and an additional 5.2 million do not yet know they have the disease [1]. Of these people, 90–95% have Type 2 diabetes [2]. Diabetes has been among the top ten leading causes of death in the United States since 1932 and is now the sixth leading cause of death [3]. Because of the magnitude of the burden of disease, the Healthy People 2010 objectives include goals of reducing diabetes-related deaths and increasing the monitoring frequency of glucose control and chronic complications [4]. The United Kingdom Prospective Diabetes Study and Kumamoto study confirm that improved glucose control reduces the microvascular complications in Type 2 diabetes such as retinopathy, nephropathy, and neuropathy [5,6]. Because of these findings, new standards of care and new models of health care delivery have emerged [7].
Disease management programs that incorporate group patient education, nutrition consultation, case management as well as close clinical care have been effective [8]. Project Dulce, an initiative of Community Health Improvement Partners, the Council of Community Clinics, and The Whittier Institute for Diabetes was started in 1998 in San Diego County. The program has a multifaceted approach and focuses on providing care to minority groups that often lack access to medical services.
Although it is known that improved glycemic control improves microvascular outcomes, less is known about the factors that influence control. Harris et al. [9] examined racial and ethnic differences in glycemic control in patients with Type 2 diabetes using the Third National Health and Nutrition Examination Survey (NHANES III) and found that black women, Mexican-American men, those treated with insulin or oral antiglycemic medications, and patients over 60 years of age had poorer glycemic control. Shorr et al. [10] studied the relationship between age and glycemic control and found no significant differences between age groups. Nichols et al. [11] found that age, body mass index (BMI) and emotional distress were significantly related to glycemic control in a health maintenance organization population in Oregon. Blaum et al. [12] found that disease duration, C peptide levels, poor self-care, and failure to receive diet recommendations were related to control in a mostly white, primary care population in Michigan. Project Dulce, however, is a distinct population of low-income, multiethnic patients with a high proportion of Hispanics and Asians. Therefore, it is useful to study patient characteristics associated with glycemic control in this unique setting. Prior studies have not accounted for fluctuations in glycemic control over time. We used a longitudinal data analysis approach to account for glycemic variation and thus maximized the amount of information that can be drawn from the data.
Methods
Data study sample
Project Dulce is a nurse-based diabetes disease management system in San Diego, California [13]. Patients with diabetes are referred to Project Dulce by primary care providers. Once the patient is referred, the nurse educator conducts an initial assessment and follows the American Diabetes Association (ADA) standards of appropriate physical and laboratory exams and referrals to specialists (7). Hemoglobin A1C (A1C) is monitored quarterly and lipids, urine microalbumin, thyroid stimulating hormone (TSH), and retinal exams are completed yearly or more frequently as needed. At each visit, height, weight, blood pressure, foot exams, and glucometer results are reviewed.
The nurse educator is the case manager and follows-up on missed patient appointments and identifies individual service and access needs of his/her panel of patients. The nurse also communicates with the primary care physician regarding clinical care issues. Project Dulce Dieticians see patients referred by the nurse educators. The program is active in seventeen sites including community clinics and hospital ambulatory care centers throughout San Diego County. Project Dulce uses the same procedures and supervision at each site and tracks patients with the Diabetes Electronic Management System (DEMS) software. The database contains demographic, health status, treatment, laboratory, and behavioral factors for each patient and collects the information over time. This study included data from July 18, 2000 to October 7, 2002 and was approved by the Institutional Review Board of San Diego State University.
Eligibility criteria
For purposes of this analysis, we selected patients with Type 2 diabetes, reducing the population size from 1,728 to 1,357. To avoid bias and ensure that the study population was actively participating in the Project Dulce program, inclusion criteria were established. The patient required: 1) at least two A1C values at least six months apart, 2) participation in the program for at least six months, and 3) at least three Project Dulce provider visits. Of the 1,357 Type 2 diabetes patients in the database, 573 met these criteria.
Measures
a. Demographics
Demographic variables included gender, age, race/ethnicity, and primary language. All were of low socioeconomic status. For purposes of this study, five racial/ethnic categories were created: Hispanic, Asian (including Indian), black, white, and other.
b. Glycemic control
A1C is a laboratory value that indicates glycemic control over a 2 to 3 month period; values less than 7% are considered optimal. A1C was our outcome of interest and was evaluated over time by examining the patients' A1C laboratory results over a 24 month period. Since Project Dulce follows the ADA recommendations of checking A1C values every 3 months, values were placed in 3 month block intervals, using the patients' initial provider visit as the reference starting point. A1C laboratory data on individual patients was not always precisely three months apart so approximations were necessary. Since A1C is an indicator of glycemic control over a 2 to 3 month period, we used a plus or minus 1.5 month approximation. For example, a three-month lab was considered an A1C measurement 1.5 to 4.5 months after the initial Project Dulce visit. A baseline A1C value was considered a measurement between 2.8 months before the initial Project Dulce visit to 1.5 months after the visit. Since a two year time period was of interest in this study, only A1C values falling within 2.8 months of the initial visit or 25.5 months after this visit were included in the analysis. If more than one A1C value was available in a particular 3 month time block, the first measurement within that block was used.
c. Diabetes severity
The difference in dates of the patient's initial Project Dulce visit and the diabetes diagnosis date estimated disease duration in years. Medicines used for glucose control (insulin, sulfonylureas, metformin, glitazones, alpha glucosidase inhibitors, meglitinides) were categorized into three levels: 1) insulin alone or insulin with oral agents, 2) more than one oral agent but no insulin, and 3) one oral medication or no medication at all. Since a patient's pharmacotherapy changed over time, we created a coding strategy. If the patient used insulin at any point over the two year study period, he/she was placed in the insulin category. Similarly, if the patient ever used more than one oral medication but never used insulin, the patient was placed in the more than one oral agent category.
d. Health status
Clinical characteristics considered included systolic (SBP) and diastolic (DBP) blood pressure, total and HDL cholesterol, urine microalbumin-to-creatinine ratio, and BMI. Mean values of the clinical variables were used over the appropriate time period. In univariate analysis, cutpoints were created based on ADA guidelines [7]. Hypertension was considered SBP or DBP greater than or equal to 130 mm Hg and 80 mm Hg, respectively. Total cholesterol or HDL greater than equal to 200 mg/dl and 45 mg/dl, respectively were defined as dyslipidemia. Urine microalbumin-to-creatinine levels between 30 and 299 ug/mg was considered microalbuminuria and greater than or equal to 300 ug/mg was considered clinical albuminuria. BMI greater than or equal to 30 kg/m2 was considered obese.
e. Access / quality of care
Most of the patients in Project Dulce have County Medical Services, an insurance program funded by San Diego County to care for the medically indigent adult population (MIA). The remainder are uninsured and pay out-of-pocket to enroll in the program or are covered by Medicare, Medicaid, or private insurance. For purposes of this study insurance status was categorized as uninsured, MIA, or insured (insured = Medicare, Medicaid or private insurance). The number of provider visits, duration in the program, and whether the patient was seen by a Project Dulce nutritionist was also recorded.
f. Behavioral factors
Behavioral factors in the model included smoking and the number of Project Dulce diabetes education classes attended.
Descriptive and Univariate analysis
The number and proportion of patients were recorded for each variable within demographic, diabetes severity, health status, access/quality of care, and behavioral factor groups. In addition, mean A1C values were compared across levels of each variable. Univariate analysis using a t-test or One-Way ANOVA was used to assess significant differences in mean A1C. If significant differences were found in ANOVA, the Duncan function in SAS 8.1 was used to asses individual differences.
Mixed effects model
Mixed effects models were used to assess glycemic control by analyzing the repeated measure data of A1C values. The A1C values were skewed and therefore log transformed in order to meet the normal distribution assumption. Several correlation structures including Compound Symmetry, Unstructured, First-order Autoregressive, and Toeplitz were assessed for each model. We used Akaike's Information Criterion (AIC) to select the appropriate correlation structure [14]. All model fittings were implemented using SAS PROC MIXED and the model with the smallest AIC was considered the best fit [15].
Univariate associations were performed to assess the best functional form of the variables. Continuous variables were assessed as linear and curve-linear with the addition of quadratic terms. Using a hierarchical model building process, clusters of variables were added in, one-by-one. All models included baseline A1C and time (in months) since these two variables were considered essential to control for in assessing glycemic control in the longitudinal format. In each model, the AIC of the best-fitted correlation structure was noted.
All variables significant in one of the hierarchical models at an alpha level of 0.15 were placed together in a separate model. Finally, a parsimonious mean effects model was created, leaving only variables significant at the alpha level of 0.05. Variable by time interaction terms were entered into the parsimonious mean effects model in a clustered process. Significant interaction terms at the alpha level of 0.05 were then placed together with the parsimonious model. Once the model variables were finalized, correlation structures for fixed and random effects were verified.
Results
Table 1 shows the study population characteristics and univariate associations of factors with glycemic control.
Table 1 Population characteristics and univariate associations of factors with mean A1C. Project Dulce, 2000–2002 (N = 573)
n (%) Mean A1C (%) p value
Demographic factors
Gender, n (%) 0.99
1. Female 392 (68.7%) 7.63
2. Male 179 (31.4%) 7.63
Age, n (%) Mean = 55.4 ± 10.1 < 0.0001
1. < 50 years 149 (26.0%) 7.89 1 > 2, 3
2. 50–65 355 (62.0%) 7.57
3. > 65 69 (12.0%) 7.36
Ethnicity, n (%) < 0.0001
1. Hispanic 304 (53.3%) 7.84 2 > 3, 4
2. Black 39 (6.8%) 7.96 4 < 2
3. Asian 100 (17.5%) 7.06 3 < 1,2, 5
4. Other 11 (1.9%) 7.46
5. White 116 (20.4%) 7.55
Primary Language, n (%) 0.18
1. Not English 302 (52.8%) 7.58
2. English 270 (47.2%) 7.68
Diabetes severity
Diabetes duration, n (%) Mean = 7.1 ± 7.1 < 0.0001
1. < 1 year 122 (21.9%) 7.04 4 > 3 > 2 > 1
2. 1 – 5 years 195 (34.9%) 7.45
3. 6 – 10 years 93 (16.7%) 7.77
4. > 10 years 148 (26.5%) 8.19
Medicine, n (%) < 0.0001
1. Insulin alone or insulin + oral agents 177 (30.9%) 8.32 1 > 2 > 3
2. > 1 oral agent (no insulin) 284 (49.6%) 7.58
3. No medicine or 1 oral agent 112 (19.5%) 6.47
Health status
Systolic blood pressure, n (%) Mean = 125.2 ± 11.9 0.33
1. < 130 mm Hg 404 (70.5%) 7.65
2. ≥ 130 mm Hg 169 (29.5%) 7.57
Diastolic blood pressure, n (%) Mean = 72.1 ± 6.6 0.19
1. < 80 mm Hg 506 (88.3%) 7.64
2. ≥ 80 mm Hg 67 (11.7%) 7.48
Total cholesterol, n (%) Mean = 187.6 ± 36.0 < 0.0001
1. < 200 mg/dl 386 (68.2%) 7.45
2. ≥ 200 mg/dl 180 (31.8%) 8.03
HDL, n (%) Mean = 45.3 ± 12.0 0.27
1. ≤ 45 mg/dl 322 (56.6%) 7.73
2. > 45 mg/dl 247 (43.4%) 7.59
Urine Microalbumin / creatinine, n (%) Mean = 139.9 ± 480.7 < 0.0001
1. < 30 ug/mg 356 (63.6%) 7.40 3 > 2 > 1
2. 30 – 299 ug/mg 160 (28.6%) 7.84
3. ≥ 300 ug/mg 44 (7.8%) 8.55
Body Mass Index, n (%) Mean = 32.5 ± 7.5 0.003
1. < 30 kg/m2 246 (43.2%) 7.50
2. ≥ 30 kg/m2 323 (56.8%) 7.72
Access/quality of care
Insurance, n (%) < 0.0001
1. Uninsured 169 (29.5%) 8.10 1 > 2, 3
2. County Medical Services 249 (43.5%) 7.39
3. Insurance 155 (27.0%) 7.50
Number of provider visits, n (%) Mean = 10.2 ± 4.4 0.002
1. 3 – 6 118 (20.6%) 7.72 4 > 2, 3
2. 7 – 10 225 (39.3%) 7.56
3. 11 – 15 166 (29.0%) 7.51
4. > 15 64 (11.1%) 7.93
Duration in program, n (%) Mean = 15.7 ± 5.5 0.0001
1. 6 – 12 months 212 (37%) 7.88
2. > 12 – 24 months 361 (63%) 7.53
Seen by nutritionist, n (%) 0.32
1. Yes 475 (82.9%) 7.61
2. No 98 (17.1%) 7.71
Behavioral Factors
Smoking habit, n (%) 0.29
1. Current 63 (12.5%) 7.58
2. Past 162 (32.0%) 7.60
3. Never 281 (55.5%) 7.72
Diabetes Classes Attended, n (%) Mean = 1.5 ± 3.0 0.002
1. 0 358 (74.3%) 7.60 2 > 1 > 3
2. 1–4 42 (8.7%) 7.88
3. 5 + 82 (17.0%) 7.31
Multiple comparison tests in ANOVA were done with the Duncan function in SAS 8.1
a. Demographics
There were more females (68.7%) than males (31.3%). The mean age was 55.4 years and the younger group had a higher mean A1C (7.9%) than the other two age groups. Hispanics represented 53.3% of the study sample and Asians had lower mean A1C values (7.1%) than Hispanics (7.8%), blacks (8.0%), and whites (7.6%). The majority of the patients (52.8%) used a language other than English as their primary language.
b. Diabetes severity
The mean duration of diabetes was 7.1 years and increasing duration of disease resulted in progressively higher mean A1C values. Insulin users comprised 30.9% of the study population and had higher mean A1C values (8.3%) than multiple oral medication users (7.6%) and those on one oral agent or no medication (6.5%).
c. Health status
Mean systolic and diastolic blood pressures were within ADA target recommendations (less than 130 and 80 mm Hg, respectively) with 70.5% and 88.3% of the study population, respectively, meeting the goals. Mean total cholesterol was 187.6 mg/dl and those with lower total cholesterol (less than 200 mg/dl) had lower mean A1C values (7.5%) than those with higher total cholesterol levels (8.0%). Patients with clinical albuminuria comprised 7.8% of the study population and had higher mean A1C values (8.6%) than those with microalbuminuria (7.8%) and those with no microalbuminuria (7.4%). The majority (56.8%) of the patients were obese and they had higher mean A1C values (7.7%) than those who were not obese (7.5%).
d. Access / quality of care
The largest (43.5%) group of patients were enrolled in San Diego County Medical Services followed by the uninsured (29.5%). Uninsured patients had higher mean A1C values (8.1%) than those with insurance (7.5%) or County Medical Services (7.4%). Most (63.0%) of the study patients were enrolled in Project Dulce over one year and this group had lower mean A1C values (7.5%) compared to the group enrolled for one year or less (7.9%). Patients with greater than 15 provider visits had higher mean A1C values (7.9%) than those with less provider visits. The majority (82.9%) of the patients had seen a Project Dulce nutritionist.
e. Behavioral factors
Current smokers comprised 12.5% of the study population. Patients who attended 5 or more Project Dulce diabetes classes had lower (7.3) mean A1C values than those who attended no classes (7.6) and those attending 1 to 4 classes (7.9).
Table 2 provides the regression results for the multivariate mixed effects model analysis which illustrates the factors associated with glycemic control. Insurance status, disease duration, pharmacotherapy, and cholesterol level were significantly associated with glucose control. Fluctuation in mean A1C over time also differed by age (age*month). Using the insured as the reference group, the uninsured had a 5.2% higher A1C level. Patients who had diabetes over ten years had a 15.3% higher A1C level compared to those who had diabetes less than one year. Similarly, patients who had diabetes six to ten years and one to five years had significantly higher A1C values compared to those with diabetes less than one year. Patients who required insulin had a 22.4% higher A1C and those who required more than one oral medication had a 12.0% higher A1C compared to hose who used one oral medication or no medication at all. On average, for every 0.65 mmol/l (25 mg/dl) increase in total cholesterol, the A1C value was 2.6% higher.
Table 2 Multivariate mixed effects model to assess characteristics associated with glycemic control. Project Dulce, 2000–2002 (N = 555)
Estimate p value Translation*
Baseline A1C 0.06050 <0.0001
Month (0 – 24) <0.0001
Insurance 0.003
Uninsured 0.02198 0.06 5.2% increase in A1C 1
County Medical Services (MIA) -0.01300 0.20
Insured (ref) - -
Diabetes duration <0.0001
> 10 years 0.06169 <0.0001 15.3% increase in A1C
6 – 10 years 0.03555 0.008 8.5% increase in A1C
1 – 5.9 years 0.03253 0.003 7.8% increase in A1C
< 1 year (ref) - -
Medicine <0.0001
Insulin alone or insulin + oral agents 0.08768 <0.0001 22.4% increase in A1C
> 1 oral agent (no insulin) 0.04930 <0.0001 12.0% increase in A1C
No medicine or 1 oral agent (ref) - -
Total cholesterol (0.65 mmol/l (25 mg/dl) interval) 0.01115 <0.001 2.6% increase in A1C
Age * month <0.001
* Formula for calculating change in A1C = 10(estimate) - 1
110(0.02198) - 1 = 0.052
How mean A1C fluctuated over time differently for various age groups is best interpreted with a plot. Although age was a continuous variable in this analysis, for purposes of interpretation, we created three categories. Figure 1 demonstrates that the less than 50 years of age group initially declined in A1C from baseline to three months but then slowly rose during the next fifteen months. The two older age groups had lower mean A1C values that fluctuated during those fifteen months.
Figure 1 Fluctuation in mean A1C values over time by age group. Project Dulce, 2000–2002
Conclusion
Univariate analysis indicates that multiple variables are associated with glycemic control. Age, race/ethnicity, disease duration, medication, number of Project Dulce visits, duration in Project Dulce, total cholesterol, microalbumin-to-creatinine ratio, BMI, insurance status, and the number of diabetes classes attended were all significant. However, after controlling for baseline A1C, time and other demographic, disease severity, health status, and access/quality of care factors, only age, insurance status, disease duration, pharmacotherapy, and total cholesterol were significant in the final model with main effects or two-way interaction terms.
The association of insurance status with glycemic control contradicts previous studies. Harris et al. [9] did not find an association between glycemic control and insurance coverage or socioeconomic status using NHANES III data, a representative sample of the U.S. population. Similarly, the Michigan community study [12], a study of blacks and whites in South Carolina [16], and a study of whites and Mexican-Americans in Texas [17] did not find an association between glycemic control and socioeconomic status. Our population, however, is distinct. It is a multiracial/ethnic population and all have a low socioeconomic status. Within this low-income population, the uninsured had poorer glycemic control. They represent a subgroup of patients that struggle to find care for a disease that requires close monitoring. Their disease state may have been out-of-control before entering Project Dulce and thus more difficult to gain control. They may also have lacked self-care skills or basic knowledge of diabetes since their care in the past was likely sporadic. Perhaps, factors such as dietary practices, physical exercise, and education level were important predictors and differed in this subgroup [18,19]. Providers may have difficulty procuring medication and equipment for this group of patients. All of these factors could help explain the discrepancy but were not accounted for in this study.
Study findings have differed on the association of glycemic control and disease duration. Similar to Blaum et al. [12] and contradictory to Nichols et al. [11], we found that the longer someone had been diagnosed with diabetes, the harder it was to maintain glycemic control. Although self-care skills could improve with longer duration of disease, resistance to medication and the need for higher doses or additional medications increase over time. Insulin use is also a factor of disease severity and was a predictor of poorer glycemic control in this study. The mean A1C value (8.3%) of insulin users in our study was equivalent to the mean value of insulin users in the NHANES III data [9]. They also found insulin users to have poorer glucose control.
Among health status factors, high total cholesterol was associated with poorer glycemic control. Since patients with diabetes are already at high risk for cardiovascular disease, this finding reinforces the need to aggressively screen and treat elevated cholesterol.
Although other health status factors were not associated with glycemic control in multivariate analysis, it is important to assess the health status of Project Dulce patients compared to other populations. Harris [20] studied health status and outcomes using NHANES III. Blood pressure was elevated (greater than or equal to 140/90 mm Hg) in 55% to 65% of the population compared to Project Dulce's 30% for SBP (greater than 130 mm Hg) and 12% for DBP (greater than 80 mm Hg). Total cholesterol was greater than or equal to 200 mg/dl in 62% to 69% of the NHANES III population compared to 32% of the Project Dulce population. Cigarette smokers compreised 18% to 24% of NHANES III and 13% of Project Dulce. The prevalence of microalbuminuria and obesity, however, was higher in the Project Dulce population than NHANES III, 36.4% vs. 26% to 30% and 57% vs. 34% to 54%, respectively. The fact that a higher proportion of Project Dulce patients compared to a representative sample of the U.S. population were meeting ADA blood pressure and cholesterol recommendations suggests the positive impact and importance of community disease management programs in low-income, multiracial/ethnic communities.
Prior studies have demonstrated race/ethnicity as a predictor of glycemic control with higher proportions of poorly controlled patients among black women and Mexican-American men [9]. In univariate analysis, our study found that Asians had better glycemic control than Hispanics, blacks, and whites. However, this relationship disappeared in multivariate analysis after taking other factors into account. Perhaps the reason why our finding differs from other studies is that regardless of race/ethnicity, all study patients were of low socioeconomic status. Prior studies examined race/ethnicity in populations with differing socioeconomic status levels.
Similar to results from Shorr et al.'s study [10] using NHANES III data, our study found that in multivariate analysis, age was not a significant main effect in predicting glucose control. However, the significant age by time interaction term (age*month) indicates that A1C patterns over time differed between age groups. Figure 1 shows that while the 50 to 65 and 65 and over age groups' A1C values fluctuated over time, the younger age group's A1C values steadily rose. Nichols et al. [11] also found poorer metabolic control among the younger age group. Since our longitudinal analysis accounted for fluctuations in A1C values, we were able to study A1C pattern differences. It would be interesting to see if this same A1C pattern difference among age groups exists in a representative sample of the U.S. population.
The strength of the current study was the use of mixed effects models. This is the first study that used a longitudinal approach to find factors associated with glycemic control. Incorporating repeated measures over time accounts for fluctuations in glucose control and maximizes the amount of information that can be drawn from the data. Another advantage was the size and diversity of the population which included large numbers of Hispanic, Asian, and white patients, far more diverse than studies using representative samples of the U.S. population.
While the race/ethnic population was diverse, the socioeconomic status of the population was not. Most of the patients were of very low income which limits the generalizability of the study results. Missing data was also a limitation. Missing quarterly A1C values was common but mixed effects models still yield unbiased estimates provided that the missing data was missing at random (MAR) [14].
Finally, multiple factors affect glycemic control. The mixed effects model incorporated demographic, disease severity, health status, access/quality of care, and behavioral factors but these are just some of the possible factors that affect glycemic control. Psychological and biological factors, self-care skills, knowledge of disease and education level, diet, exercise, other comorbid diseases, etc. were not explained by this model. Nichols et al.'s [11] study found that only 9% of the variability in glycemic control was explained by the factors in their model and suggested that personal characteristics may not explain a lot of differences in glycemic control among patients with Type 2 diabetes.
This study identified patients with poorer glycemic control in Project Dulce. The findings should not be generalized to all patients with Type 2 diabetes but can be applied to racial/ethnically diverse, low-income populations. Those who were uninsured, had diabetes for a longer period of time, used insulin or multiple oral agents, or had high cholesterol had poorer glycemic control. The younger population also lagged behind others. Secondarily, this study showed that a high proportion of the patients were meeting ADA's blood pressure and cholesterol recommendations, suggesting that community disease management programs in low-income populations can be effective and may contribute to improved health outcomes.
This study provides a useful methodology to assess disease management systems that collect longitudinal data. It does not provide answers to why patients are not optimally controlled but does provide a starting point from which to investigate and address the obstacles that prevent patients with diabetes from reaching their metabolic targets.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SB designed and wrote the study and analyzed the data. MJ participated with the design, analysis and interpretation of the data and revision of the paper. RF participated in the design and writing of the paper. AT acquired the data and participated in the design and revision of the paper. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
National Institute of Diabetes and Digestive and Kidney Diseases National Diabetes Statistics fact sheet: general information and national estimates on diabetes in the United States, 2003 Bethesda, MD: US Department of Health and Human Services, National Institutes of Health, 2003 Rev ed Bethesda, MD: US Department of Health and Human Services, National Institutes of Health 2004 Accessed February 5, 2005.
Bishop DB Zimmerman BR Roesler JS Brownson RC, Remington PL, Davis JR Diabetes Chronic Disease Epidemiology and Control 1998 2 Washington, DC: The American Public Health Association 421 464
Arias E Smith BL Deaths: preliminary data for 2001 National vital statistics reports 2003 51 Hyattsville, MD: National Center for Health Statistics
Dept of Health and Human Services (US) Health people 2010: Understanding and improving health 2000 2 Washington, DC: US Government Printing Office
U.K. Prospective Diabetes Study Group Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33) Lancet 1998 352 837 853 9742976 10.1016/S0140-6736(98)07019-6
Ohkubo Y Kishikawa H Araki E Miyata T Isami S Motoyoshi S Kojima Y Furuyoshi N Shichiri M Intensive insulin therapy prevents the progression of diabetic microvascular complications in japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study Diabetes Res Clin Pract 1995 28 103 117 7587918 10.1016/0168-8227(95)01064-K
American Diabetes Association Standards of medical care for patients with diabetes mellitus Diabetes Care 2003 26 33 50
Norris SL Nichols PJ Caspersen CJ Glasgow RE Engelgau MM Jack L Isham G Snyder SR Carande-Kulis VG Garfield S Briss P McCulloch D The effectiveness of disease and case management for people with diabetes Am J Prev Med 2002 22 15 38 11985933 10.1016/S0749-3797(02)00423-3
Harris MI Eastman RC Cowie CC Flegal KM Eberhardt MS Racial and ethnic differences in glycemic control of adults with type 2 diabetes Diabetes Care 1999 22 403 408 10097918
Shorr RI Franse LV Resnick HE Di Bari M Johnson KC Pahor M Glycemic control of older adults with type 2 diabetes: Findings from the Third National Health and Nutrition Examination Survey, 1988–1994 J Am Geriatr Soc 2000 48 264 267 10733051
Nichols GA Hillier TA Javor K Brown JB Predictors of glycemic control in insulin-using adults with type 2 diabetes Diabetes Care 2000 23 273 277 10868850
Blaum CS Velez L Hiss RG Halter JB Characteristics related to poor glycemic control in niddm patients in community practice Diabetes Care 1997 20 7 11 9028685
Philis-Tsimikas A Walker C Improved care for diabetes in underserved populations J Ambulatory Care Management 2001 24 39 43
Verbeke G Molenberghs G Linear mixed models for longitudinal data 2000 New York: Springer
Ngo L Brand R Model selection in linear mixed effects models using sas proc mixed: Technical proceedings SAS Users Group International 1997 22 1 6
Eberhardt MS Lackland DT Wheeler FC German RR Teutsch SM Is race related to glycemic control? An assessment of glycosylated hemoglobin in two South Carolina communities J Clin Epidemiol 1994 47 1181 1189 7722552 10.1016/0895-4356(94)90105-8
Haffner SM Hazuda HP Stern MP Patterson JK Van Heuven WAJ Fong D Effect of socioeconomic status on hyperglycemia and retinopathy levels in Mexican Americans with NIDDM Diabetes Care 1989 12 128 134 2702895
Schillinger D Grumbach K Piette J Wang F Osmond D Daher C Palacios J Sullivan GD Bindman AB Association of health literacy with diabetes outcomes JAMA 2002 288 475 482 12132978 10.1001/jama.288.4.475
Goldman DP Smith JP Can patient self-management help explain the SES health gradient? PNAS 2002 99 10929 10934 12140364 10.1073/pnas.162086599
Harris MI Racial and ethnic differences in health care access and health outcomes for adults with type 2 diabetes Diabetes Care 2001 24 454 459 11289467
| 15833140 | PMC1090595 | CC BY | 2021-01-04 16:28:57 | no | BMC Public Health. 2005 Apr 17; 5:36 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-36 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-421584514410.1186/1471-2458-5-42Research ArticleCompulsory and recommended vaccination in Italy: evaluation of coverage and non-compliance between 1998-2002 in Northern Italy Stampi Serena [email protected] Rita [email protected] Isa [email protected] Franca [email protected] Department of Medicine and Public Health, Division of Hygiene, University of Bologna, Via San Giacomo 12, Bologna, Italy2 Bologna Health Unit, Community Pediatric Service, Bologna East District, Via Zanolini, 2, Bologna, Italy2005 21 4 2005 5 42 42 13 1 2005 21 4 2005 Copyright © 2005 Stampi et al; licensee BioMed Central Ltd.2005Stampi 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
Since vaccinations are an effective prevention tool for maintaining the health of society, the monitoring of immunization coverage allows us to identify areas where disease outbreaks are likely to occur, and possibly assist us in predicting future outbreaks. The aim of this study is the investigation of the coverage achieved for compulsory (diphtheria, tetanus, polio, hepatitis B,) and recommended (pertussis, Haemophilus influenzae, measles-mumps-rubella) vaccinations between 1998 and 2002 in the municipality of Bologna and the identification of the subjects not complying with compulsory and recommended vaccinations.
Methods
The statistics regarding vaccinal coverage were elaborated from the data supplied by the Bologna vaccinal registration system (1998–2000) and the IPV4 program (2001–2002). To calculate the coverage for compulsory vaccinations and cases of non-compliance reference was made to the protocol drawn up by the Emilia Romagna Regional Administration. The reasons for non-compliance were divided into various categories
Results
In Bologna the levels of immunization for the four compulsory vaccinations are satisfactory: over 95% children completed the vaccinal cycle, receiving the booster for anti-polio foreseen in their 3rd year and for anti-dyphteria, tetanus, pertussis at 6 years. The frequency of subjects with total non-compliance (i.e. those who have not begun any compulsory vaccinations by the age of one year) is generally higher in Bologna than in the region, with a slight increase in 2002 (2.52% and 1.06% in the city and the region respectively). The frequency of the anti-measles vaccination is higher than that of mumps and rubella, which means that the single vaccine, as opposed to the combined MMR (measles-mumps-rubella) was still being used in the period in question.
The most common reason for non compliance is objection of parents and is probably due to reduction of certain diseases or anxiety about the possible risks.
Conclusion
In Bologna the frequency of children aged 12 and 24 months who have achieved compulsory vaccination varied, in 2002, between 95% and 98%. As regards recommended vaccinations the percentage of coverage against Haemophilus influenzae is 93.3%, while the levels for measles, mumps and pertussis range from 84% to approx. 92%. Although these percentages are higher if compared to those obtained by other Italian regions, every effort should be made to strengthen the aspects that lead to a successful vaccinal strategy.
==== Body
Background
Vaccinations are an effective prevention tool for maintaining the health of the individual and society as a whole. In Italy, however, certain diseases that are preventable with vaccines whose efficiency and safety have been fully tested and proven, still present a public health issue.
Monitoring immunization coverage allows us to identify areas where disease outbreaks are likely to occur, and possibly assist us in predicting future outbreaks. Also, monitoring can help keep immunization coverage high. One of the main aims of public health services throughout the world is therefore to obtain reliable data about the level of immunization with a view to supporting the implementation and monitoring of vaccination programs [1].
Vaccinal coverage is normally referred to in terms of the percentage of subjects vaccinated out of the total target population. In pediatric age the target population is made up of the cohorts of children who were under one year old as of 31 December in a given locality, while the vaccinal state is defined as the number of doses required to bring about immunity. To determine vaccinal coverage, the WHO (World Health Organization), on a yearly basis, considers the percentage of children who have received three doses of vaccines against diphtheria, tetanus, polio, hepatitis B, pertussis and Haemophilus influenzae (Hib), and one dose of anti-measles, mumps and rubella (MMR) by the age of 24 months.
In Italy the vaccination coverage of newborns (i.e. the relationship between the number of doses of vaccine administered at a certain age in a given geographical area and the resident population of the same age) is estimated every year by the regional authorities. The data are then sent to the Ministry of Health. In order to have a more detailed picture of the national situation (to identify areas lacking coverage, to confirm coverage where present and to obtain further details, for example, about reasons for delay in application), a cluster survey, called ICONA, was made in 1998 by means of interviews carried out at the subjects' home. The results showed a considerable difference between the coverage achieved for compulsory (diphtheria, tetanus, polio, hepatitis B) and for recommended (pertussis, Hib, measles, mumps ands rubella) vaccines [2].
After five years the national survey was repeated (ICONA 2003) [3], with the particular aim of checking the data supplied by regional authorities to the Ministry and investigating the reasons for non-compliance. The study revealed that while the situation had improved, there were still differences between one region and another.
In the wake of the approval of the National Health Plan (NHP) to eliminate measles and congenital Rubella (13 November 2003) and considering the current attempt to overcome the existing division between compulsory and recommended vaccinations, we decided to undertake a study with the following objectives:
1. To investigate the coverage achieved for compulsory and recommended vaccinations between 1998 (the year when the NHP set the objective of 95% coverage) [4] and 2002 in the municipality of Bologna (capital of Emilia-Romagna region), which coincides with the East and West Districts of the city's health service;
2. To identify the categories of subjects not complying with compulsory and recommended vaccinations
Methods
In Bologna the vaccination of subjects under 18 years of age is carried out almost exclusively by the Community Pediatric Service (which combines the former pediatric surgeries and the Nursery and Infant School Medical Service). In the current make-up of Bologna Health Services, this operative Unit is situated within the Department of Primary Care, and is divided into two districts: Bologna East and Bologna West.
Until 2000 the registration of vaccinations was carried out by the Public Health Department by means of a computerized system run by the Municipality of Bologna. The office had access to registration records (they could check the issuing of vaccination certificates), but not to the more detailed data processing system (search for cases of non-compliance, new residents etc).
Details of births or new residents under the age of 6 years were passed on to the subject's local pediatric consultation centre by the municipal Registry Office, with the issuing of a vaccination booklet. At the end of the year the Office printed a list of non-compliances and sent it to the respective Pediatric Centres for a further check. Finally, full details were sent yearly to the Regional offices by the Community Pediatric Service. During the year health workers could check the extent of vaccinal coverage only by means of the paper documents (files) sent at the same time as the vaccination booklets. The only way that the employees who registered the vaccinations on the files could find out about subjects who had moved or died was by going directly to the municipal registry office.
Since the beginning of 2001 the registration of vaccinations has been run exclusively by the Local Health Services of the city of Bologna, using a special computer program (called IPV4), "for the purposes of the management of vaccinal registration, the municipality periodically transmits details of population variations (births, immigrations, emigrations, deaths) to the Local Health Offices" (art. 160, paragraph 3 of the Municipal Hygiene Regulations).
Alongside the installation of the vaccinal registration program, a project was launched to set up an on-line link between the various centres administrating the vaccines at pediatric age. This allows the immediate registration of vaccinal data at the moment of vaccination and the possibility to consult in real time the vaccinal situation of each subject, including details of any migration and access to the so-called "notes" (possible contra-indications, previous adverse reactions, refusals etc).
The great advantage of the revised system is that it is now possible to monitor vaccinal coverage at any time, obtain information pertaining to moves or deaths, and contact residents who have not been vaccinated.
The statistics regarding vaccinal coverage were elaborated from the data supplied by the vaccinal registration system of the Municipality of Bologna (1998–2000) and the IPV4 program (2001–2002).
To calculate the coverage for compulsory vaccinations and cases of non-compliance reference was made to the protocol drawn up by the Emilia Romagna Regional Administration:
compulsory vaccinations at 12 months
residents who reached one year during the year of reference and received at least 2 doses of each compulsory vaccine by the age of 12 months;
compulsory vaccinations at 24 months
residents who reached two years during the year of reference and received at least 3 doses of each compulsory vaccine by the age of 24 months.
subjects totally or partially non-complying
subjects who did not receive any compulsory vaccination by the age of one year or received only one dose of the compulsory vaccinations.
The reasons for non-compliance are categorized matching those used in the collection of regional data for coverage:
- living abroad: children resident in Bologna but living abroad for a long period
- refusals or objectors: children whose parents were explicitly against at least one compulsory vaccination
- immigrants: children from other regions of Italy or from other countries
- delays: children who received at least one vaccination after the pre-established dates
- other: children with health contra-indications, momentarily exempt for health reasons, not traceable or without fixed abode.
For all recommended vaccinations (anti pertussis, anti Hib, anti measles-mumps-rubella) the extent of coverage was calculated at 24 months and also at 13 years for anti-measles and anti-rubella.
Results
The levels of immunization for the four compulsory vaccinations (taken as a whole) achieved during the five-year period in question are shown in table 1. In the municipality of Bologna these levels are satisfactory (>95%), but are lower than the overall levels reached in Emilia Romagna, although from the year 2000 the percentage of children vaccinated in the region also shows a slight decrease. Interestingly, the statistics differ from one district of the city to another: in Bologna East non-compliance tends to be higher.
Table 1 Levels of vaccinal coverage at 12 months (2 doses) and 24 months (3 doses) for compulsory vaccinations in the municipality of Bologna and the Emilia Romagna region
12 months 24 months
resident population resident population
N % N %
1998
Municipality of Bologna 2561 97.11 2398 94.33
Emilia Romagna 30585 98.70 29662 98.40
1999
Municipality of Bologna 2464 96.15 2560 95.90
Emilia Romagna 31567 98.41 30916 98.30
2000
Municipality of Bologna 2630 97.30 2429 96.62
Emilia Romagna 32423 98.40 31727 98.10
2001
Municipality of Bologna 2755 97.10 2619 96.33
Emilia Romagna 34728 98.05 32866 98.01
2002
Municipality of Bologna 2741 95.81 2706 96.01
Emilia Romagna 34969 97.64 34950 97.45
Data for the assessment of each separate vaccination are available only from 2001 (table 2). There is a clear trend towards a decline in the percentage of coverage in Bologna, which becomes even more evident in 2002. There are also differences between the compulsory vaccinations: anti-DT has the highest coverage, anti-polio intermediate and anti-hepatitis B the lowest. Regarding the vaccination against Hepatitis B, the percentage of subjects in Bologna who had completed the cycle by the age of 13 years passed from 95.0% in 1988 to 96.4% in 2002. In the same period the mean national levels of coverage were more than one percent lower, with an opposite trend to that of the Emilia Romagna region.
Table 2 Levels of vaccinal coverage at 12 months (2 doses) and 24 months (3 doses) for compulsory vaccinations in the municipality of Bologna and the Emilia-Romagna region
12 months 24 months
DPT Polio Hepatitis B DPT Polio Hepatitis B
% % % % % %
2001
Municipality of Bologna 98.7 98.5 97.1 98.3 97.9 96.5
Emilia Romagna 98.5 98.4 98.1 98.5 98.4 98.0
Italy 96.1 96.0 94.7
2002
Municipality of Bologna 96.7 96.7 95.9 97.5 97.4 96.0
Emilia Romagna 98.2 98.2 97.7 98.1 98.0 97.5
Italy* 96.9 96.7 95.7
* estimated data at 31 October 2003
In the city of Bologna, as in the region taken as a whole, over 95% children completed the vaccinal cycle, receiving the booster for anti-polio foreseen in their 3rd year and for anti-DPT at 6 years. Once again there is a marked difference between the two Districts considered (Polio: East vs West = 95.8% vs 98.3%; DPT East vs West = 93.45% vs 96.8%).
The frequency of subjects with total non-compliance (i.e. those who have not begun any compulsory vaccinations by the age of one year) is generally higher in Bologna (especially in the East District) than in the region, with a slight increase in 2002 (2.52% and 1.06% in the city and the region respectively) (table 3).
Table 3 Frequency of subjects with total non-compliance (*) for compulsory vaccinations in the municipality of Bologna (Bologna East and Bologna West), and in the Emilia Romagna region
1998 1999 2000 2001 2002
% % % % %
Bologna East 0.65 1.74 1.14 0.89 2.67
Bologna West 1.12 0.17 0.76 0.36 2.34
Municipality of Bologna 0.88 0.97 0.95 0.62 2.52
Emilia Romagna 0.40 0.62 0.52 0.78 1.06
(*) subjects that have not started any compulsory vaccination by the age of one year
The vaccination registration system provides us with information about the reasons for non-compliance for the compulsory vaccinations: some can be attributed to the conditions of the children (postponements or exemptions for medical reasons, living abroad), others to the parents (untraceable, objectors-refusals). At 12 months the most common reason for non-compliance is "parents who object to and refuse the vaccinations" (1.01–1.56% of children resident in Bologna) (table 4). At 24 months the reasons vary more from year to year (table 5). In 2001 and 2002 the frequency of parents who stated they were against the anti-hepatitis B vaccination (both at 12 months and 24 months) was higher than that of the other compulsory vaccinations.
Table 4 Frequency of non-compliance for compulsory vaccinations at 12 months (2 doses) divided by category
resident population living abroad no fixed abode refusals objectors immigrated delays other*
N. % % % % % %
1998
Municipality of Bologna 2561 0.59 0.08 1.01 0.23 0.66 0.31
Emilia Romagna 30585 0.22 0.06 0.35 0.18 0.27 0.13
1999
Municipality of Bologna 2464 0.93 0.24 1.30 0.49 0.28 0.61
Emilia Romagna 31567 0.32 0.07 0.57 0.22 0.30 0.11
2000
Municipality of Bologna 2630 0.00 0.11 1.37 0.61 0.42 0.19
Emilia Romagna 32423 0.21 0.05 0.22 0.57 0.37 0.19
2001
Municipality of Bologna 2755 0.11 0.04 1.56 0.22 0.54 0.43
Emilia Romagna 34728 0.20 0.02 0.91 0.20 0.39 0.22
2002
Municipality of Bologna 2741 0.62 0.11 1.39 0.91 0.33 0.84
Emilia Romagna 34969 0.21 0.06 1.06 0.24 0.42 0.37
* postponed or exempted for medical reasons, untraceable, etc.
Table 5 Frequency of non-compliance for compulsory vaccinations at 24 months (3 doses) divided by category
resident population living abroad no fixed abode refusals objectors immigrated delays other*
N. % % % % % %
1998
Municipality of Bologna 2398 0.71 0.12 1.00 0.25 2.04 1.54
Emilia Romagna 29662 0.24 0.09 0.29 0.22 0.51 0.22
1999
Municipality of Bologna 2560 1.09 0.43 0.47 0.78 1.05 0.27
Emilia Romagna 30916 0.27 0.07 0.35 0.22 0.37 0.11
2000
Municipality of Bologna 2429 0,00 0.08 0.54 1.19 0.86 0.70
Emilia Romagna 31727 0.20 0.00 0.20 0.40 0.50 0.2
2001
Municipality of Bologna 2619 0.04 0.19 1.26 0.38 1.07 0.53
Emilia Romagna 32866 0.25 0.05 0.60 0.26 0.58 0.25
2002
Municipality of Bologna 2706 0.00 0.04 1.55 0.59 1.07 0.74
Emilia Romagna 34950 0.21 0.04 1.00 0.23 0.69 0.38
* postponed or exempted for medical reasons, untraceable,
Data regarding the coverage at 24 months for the recommended vaccinations confirm the decline in the vaccination of children in Bologna as compared to the rest of the region (table 6), a trend already noted for the compulsory vaccinations.
Table 6 Levels of vaccinal coverage at 24 months and 13 years for recommended vaccinations in the municipality of Bologna (BO) and the Emilia-Romagna region (ER)
1998 1999 2000 2001 2002
BO ER BO ER BO ER BO ER BO ER
% % % % % % % % % %
24 months
vaccine against
pertussis 95.7 96.8 94.2 96.1 95.1 96.7 94.9 96.8 94.8 96.8
H. influenzae 41.8 50.6 45.0 66.8 65.9 81.0 86.9 90.6 93.3 95.2
measles 79.3 88.2 82.4 89.1 83.5 90.4 86.4 90.7 90.2 92.3
mumps - - 78.8 84.2 83.3 89.4 85.1 90.1 88.8 91.7
rubella 77.2* 85.2* 78.6 88.0 83.4 89.4 85.6 90.1 88.8 91.7
13 years
vaccine against
measles 75.1 83.8 72.5 89.1 85.8 84.9 90.1 87.4 92.6 91.3
rubella 84.7* 85.2* 49.9 60.8 72.6 66.3 76.1 73.4 84.2 80.6
* data available for females only
The underlying trend for such coverage tends to vary over the five-year period. The level of protection against pertussis remains more or less constant, even though it falls a little below the regional average. However, a steady rise can be seen for the other recommended vaccinations, especially evident in the case of anti-Hib (41.8% and 50.6% of children resident in Bologna and Emilia Romagna respectively in 1998, against 93.3% and 95.2% in 2002). The notable increase in anti-Hib vaccination can probably be attributed to the availability of a combined vaccine.
At 12 months a similar consistent increase in the number of subjects receiving the anti-Hib vaccination can again be seen in Bologna (from 26.7% to 91.6%) and throughout the whole of the region (from 40.9% to 96.0%). In Bologna the target percentage of 95% coverage for the recommended vaccinations has yet to be reached.
Table 6 also shows how the frequency of the anti-measles vaccination is higher than that of mumps and rubella, which means that the single vaccine, as opposed to the combined MMR, was still being used in the period in question. From the year 2000, the levels of coverage reached in Bologna for anti-measles and anti-rubella at 13 years of age would seem to indicate a greater attention to these vaccinations in Bologna than in other parts of the region. Unlike the findings for the compulsory and recommended vaccinations, no difference was found between the two Districts of the city.
Discussion
The principal infant vaccinations have now been included in the Essential Levels of Assistance that the Region must guarantee free of charge to all citizens. Despite this, however, the levels of coverage for compulsory and recommended vaccinations have not yet reached homogeneous levels. In the municipality of Bologna the frequency of children aged 12 and 24 months who have been vaccinated against DTP, tetanus, polio, hepatitis B and pertussis varied, in 2002, between 95% and 98%. As far as the recommended vaccinations are concerned, the percentage of coverage against Hib is 93.3%, while the levels for measles, mumps and pertussis range from 84% to approx. 92%. Although these percentages are generally lower than the regional levels, they can nevertheless be considered fairly good if compared to those obtained by certain other Italian regions [5] and Italy as a whole. The results achieved at a national level for vaccinations against polio, DTP and hepatitis B were respectively 96.6%, 95.3% and 94.1% in 2000 and 96.7%, 96.9% and 95.7% in 2002 [3]. Moreover, the anti-Hib vaccination was given to 54.7%, 71.2% and 84.4% of children in Italy respectively in the years 2000, 2001 and 2002, while the mean coverage for MMR was 74.1,%, 76.5% and 81.1% (Ministry of Health). In other countries [1,6,7] the rate is sometimes even lower: in France, for example, the Hepatitis B vaccine was administered to only 40.1% of children of equal age [8], even though the vaccine against measles, mumps and rubella at 24 months was given to 93%.
The data for Bologna in 2002 show a general rise in the level of coverage for recommended vaccinations in comparison to previous years, especially for anti-Hib. However, even though an increase can be seen in the levels of anti-measles and rubella and the situation in Bologna is on average better than that reported by the rest of the country, the target coverage of 95% envisaged by the National Plan for the Elimination of Measles and Rubella has not yet been achieved. The objectives set by the Plan, to be reached by 2007, are:
- to reduce and maintain the incidence of Congenital Rubella Syndrome (CRS) at levels below 1 case in every 100.000 live-borns.
- to achieve and maintain the elimination of measles at a national level, interrupting its indigenous transmission.
It should be stressed that the inadequate vaccinal coverage achieved for anti-measles does not interrupt the circulation of the virus and may even bring about a greater probability of complications and outbreaks; it prolongs the interval between epidemics and delays the outbreaks until later in adult life, as presumably occurred in the Campania region [9]. It has also been shown that low coverage rates tend to enhance epidemics associated with imported viruses [10]. In 2002 there was quite a large outbreak of measles in Italy. The geographic distribution of cases strictly coincided with that of the vaccine coverage, which is lower in southern Italy [11].
As far as the data on non-compliance (compulsory vaccinations) is concerned, it is interesting to note that the frequency is higher in Bologna than the regional average and, in particular, Bologna East has a higher percentage than Bologna West both at 12 and 24 months. The percentage of total non-compliance is also greater in the East district than the West. Furthermore, the number of children in this category is steadily increasing: passing from 0.88% in 1998 (regional mean 0.40%) to 2.52% in 2002 (regional mean 1.06%).
The most common reason for non compliance is objection of parents. Objection to compulsory vaccinations, especially in Bologna (but not throughout the whole of Italy), is on the rise, even though vaccinal coverage remains above 95%. There is an increasing tendency not to comply with vaccinal obligations, probably because the disappearance (diphtheria, polio) or limited incidence (tetanus, hepatitis B) of certain diseases has led some people to believe that the relative vaccines are obsolete and there is less concern about the risk of contracting infectious diseases. However, recent reports [12] indicate that Italy could risk the importation of the wild polio virus on account of the continuous exchanges with North African countries which are currently in danger of an epidemic due to their negative behaviour towards the vaccination, thus allowing the virus to pass into otherwise polio-free areas.
A further reason for the refusal of vaccination is the anxiety about the possible risks associated with receiving the vaccines. In other words there is the pressure, especially in industrialised countries, from organised groups opposing the practice of vaccination and spreading alarmist and often inaccurate information [13,14]. For example, there is absolutely no scientific evidence that the presence of thiomerosal in vaccines can be linked to the increase in autism [15] or that there is any link between autism and the anti-measles vaccination [16,17]. If the tendency not to vaccinate children becomes more pronounced over the coming years, it could result in serious consequences for public and individual health. It should be remembered that many children still die from preventable infectious diseases and connected pathologies not only in the so-called "poor" countries, but also in other nations. In Europe, the political, social and economic changes that took place following the fall of the Berlin Wall have led to a decline in the levels of immunization in many regions of Eastern and Central Europe, resulting in thousands of new cases of diseases such as diphtheria [18]. For this reason the level of dissent needs to be constantly monitored. The computerized system adopted in Bologna can, in fact, help improve the identification of non-vaccinated subjects and verify the phenomenon of immigration.
Conclusion
In Bologna the frequency of children aged 12 and 24 months who have achieved compulsory vaccination varied, in 2002, between 95% and 98%; As regards recommended vaccinations the percentage of coverage against Hib is 93.3%, while the levels for measles, mumps and pertussis range from 84% to approx. 92%. Although these percentages are higher if compared to those obtained by other Italian regions, it is necessary to take into account that in Italy, during the 1998–2002 period the mean number of cases of measles, rubella and mumps was 5456, 3685 and 24663 respectively [19]. Therefore, every effort should be made to strengthen the aspects that lead to a successful vaccinal strategy: the continuous improvement in the quality of services, the provision of accurate information to parents and, finally, the training of health workers – in accordance with the suggestions of the WHO – to monitor adverse reactions and maintain an open and constant dialogue with the media about the problems of vaccine safety [18,20].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SS participated in the design and coordination of the study, in the interpretation of results, in revising and editing the manuscript. RR participated in the conception of the study, in the interpretation of the results and in writing the first draft. IR collected the data and performed statistical analyses. FZ participated in the interpretation of the results, the writing of the manuscript and assisted with the statistical analyses and surveys
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Murray CJL Shengelia B Gupta N Moussavi S Tandou A Thierren M Validity of reported vaccination coverage in 45 Countries Lancet 2003 362 1022 1027 14522532 10.1016/S0140-6736(03)14411-X
Salmaso S Rota MC Ciofi degli Atti M Tozzi AE Kreidl P e ICONA Study Group Infant immunization coverage in Italy by cluster survey estimates Bull World Health Organ 1999 77 843 851 10593033
Ministry of Health, General Management for Prevention, Office III, Infectious Diseases Vaccination Coverage in Italy 2004
DPR 23 luglio 1998 Public Health and Hygiene Approval of the National Health Plan for the three-year period 1998–2000 GU n. 288 10 december 1998 (Suppl Ord n. 201)
Angelillo IF Ricciardi G Rossi P Pantisano P Langiano E Pavia M Mothers and vaccination: knowledge, attitudes and behaviour in Italy Bull World Health Organ 1999 77 224 229 10212512
Antona D Bussière E Guignon N Badeyan G Lèvy-Bruhl D Vaccine coverage of pre-school age children in France in 2000 Eurosurveillance Monthly Arch 2003 8 139 144
Swennen B Van Damme P Vellinga A Coppieters Y Depoorter AM Analysis of factors influencing vaccine uptake: perspectives from Belgium Vaccine 2002 20 55 57 10.1016/S0264-410X(02)00132-9
Gaudelus J Ovetchkine P Cheimol J De Courson F Allaert FA Suivi des recommendations vaccinales des nourrissans de 0 à 24 mois: è propos d'une enquete en medicine libérale Arch Pediatr 2003 10 871 786 10.1016/S0929-693X(03)00409-3
CDC Measles epidemic attributed to inadequate vaccination coverage, Campania, Italy, 2002 MMWR Weekly 2003 52 1044 1047
Coughlans S. Connel J Cohen B Jin L Hall WW Suboptimal Measles-Mumps-Rubella vaccination coverage facilitates an imported measles outbreak in Ireland Clin Infect Dis 2002 35 84 86 12060880 10.1086/340708
Echarlod E Marcoz G Sudano L Vaccine coverage and measles outbreak in the Valle d'Aosta: the reality and criticality of a small region Il Giornale della Vaccinazione 2004 4 10 14
Fara GM Patti AM What strategies in an Italy without Polio? Journal of Prev Med Hyg 2004 45 89
Committee on Community Health Services and Committee on Practice and Ambulatory Medicine Increasing immunization coverage Pediatrics 2003 112 993 996 14523201 10.1542/peds.112.4.993
Leask J McIntyre Public opponents of vaccination: a case study Vaccine 2003 21 4700 4703 14585678 10.1016/S0264-410X(03)00520-6
Verstraeten T Davis RL De Stefano F Lieu TA Rhodes PH Black SB Shienfield H Chen RT for the vaccine safety Datalink Team Safety of Thimerosal-containing vaccines. A two phase study of computerized health maintenance organization database Pediatrics 2003 112 1039 1048 14595043 10.1542/peds.112.2.e98
Chen W Landau S Sham P Fombonne E No evidence for links between autism, MMR and measles Virus Psychol Med 2004 34 543 553 15259839 10.1017/S0033291703001259
Smeeth L Cook C Fombonne E Heavey L Rodrigues LC Smith PG Hall AJ MMR vaccination and pervasive developmental disorders: a case-control study Lancet 2004 364 963 969 2004 (11 September) 15364187 10.1016/S0140-6736(04)17020-7
Panà A Current situation of vaccines and immunization of the population Ig Sanita Pubb 2003 58 379 388
Carreri V The commitment of the regions concerning diseases preventable with vaccines Il Giornale della Vaccinazione 2004 5 14 16
Jacobson RM Vaccine Safety Immunol Allergy Clin North Am 2003 23 589 603 14753382 10.1016/S0889-8561(03)00090-0
| 15845144 | PMC1090596 | CC BY | 2021-01-04 16:28:57 | no | BMC Public Health. 2005 Apr 21; 5:42 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-42 | oa_comm |
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-221586970710.1186/1471-244X-5-22Research ArticleThe effects of amisulpride on five dimensions of psychopathology in patients with schizophrenia: a prospective open- label study Herrera-Estrella Miguel [email protected] Rogelio [email protected] Ana [email protected] Isabel [email protected] Fray Bernardino Alvarez Psychiatric Hospital, Mexico City, Mexico2 Psychiatry Department, National Institute of Neurology and Neurosurgery, Mexico City, Mexico3 Clinical Research Division, National Institute of Psychiatry, Mexico City, Mexico4 Carracci Medical Group, Mexico City, Mexico5 Sanofi- Synthelabo, Mexico City, Mexico2005 3 5 2005 5 22 22 9 12 2004 3 5 2005 Copyright © 2005 Herrera-Estrella et al; licensee BioMed Central Ltd.2005Herrera-Estrella 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 efficacy of antipsychotics can be evaluated using the dimensional models of schizophrenic symptoms. The D2/D3-selective antagonist amisulpride has shown similar efficacy and tolerability to other atypical antipsychotics. The aim of the present study was to determine the efficacy of amisulpride on the dimensional model of schizophrenic symptoms and tolerability in latin schizophrenic patients.
Method
Eighty schizophrenic patients were enrolled and 70 completed a prospective open-label 3-month study with amisulpride. The schizophrenic symptoms, psychosocial functioning and side-effects were evaluated with standardized scales.
Results
The patients showed significant improvement in the five dimensions evaluated. Amisulpride (median final dose 357.1 mg/d) was well-tolerated without treatment-emergent extrapyramidal side-effects.
Conclusion
Amisulpride showed efficacy on different psychopathological dimensions and was well tolerated, leading to consider this drug a first line choice for the treatment of schizophrenia.
==== Body
Background
The treatment of schizophrenia has shown important improvements since the introduction of new antipsychotics [1]. These drugs, also called atypical or second generation antipsychotics (SGAs) bring the possibility of a better quality of life for patients affected with schizophrenia, because they have been associated with a better efficacy over negative symptoms, probably less cognitive impairment and lower probability of extrapyramidal symptoms (EPS) [2], which is one of their main advantages.
The SGAs increase the release of dopamine in the prefrontal cortex and the hippocampus [3]. This effect of SGAs is critical to improve negative symptoms and cognition, and to decrease the EPS. The principal hypothesis of their mechanism of action has been attributed to the antagonism of 5-HT2A receptors coupled to weaker antagonism of dopamine D2 receptors [4,5]. Their effect as 5-HT1A receptor agonist has also been suggested to contribute to an atypical antipsychotic profile [6].
Nevertheless, amisulpride represents an important contrast on the theory of the 5HT2A receptor antagonism. Amisulpride, is a benzamide with high affinity for dopamine D2 and D3 receptors without affinity for serotonin, muscarinic or alpha-adrenergic receptors [7]. Amisulpride also shows selectivity for dopamine receptors in limbic and hippocampal structures, rather than striatal region [8].
At low doses (100–300 mg/d), amisulpride binds preferentially on D2/D3 presynaptic autoreceptors [8,9], increasing dopaminergic transmission in the prefrontal cortex, which is believed to be associated with improvement of primary negative symptoms. Doses between 400–800 mg/d result in antagonism of postsynaptic dopamine receptors [8], leading to an improvement of positive symptoms of schizophrenia with less EPS development.
The limbic selectivity of amisulpride is similar to the observed with clozapine, and is secondary to its high affinity for D3 receptors and the short isoform of the D2 receptor, which are highly distributed in these regions [10]. This selectivity has been documented in animal models [7]. In addition, PET studies have shown that receptor D2 occupancy in striatal regions is around 14% when amisulpride is prescribed at doses between 50–100 mg/d [11]. A decreased amisulpride plasma concentration induces a low percentage of occupancy in striatal and increased occupancy in limbic regions [12].
The improvement of negative symptoms has been documented in clinical studies with doses between 50–100 mg/d [13,14]. Until now, amisulpride is the only antipsychotic that has shown scientific evidence of its efficacy over the primary negative symptoms of schizophrenia [15-18]. Several clinical trials have shown that amisulpride has similar efficacy and better tolerability in comparison to haloperidol and flupentixol [19-22] as well as similar efficacy and safety when compared to olanzapine and risperidone [23-27]. Additionally, amisulpride has shown a positive effect over depressive symptoms [28,29] and the cognitive impairment [30,31] related to schizophrenia.
These data support that amisulpride is also an 'atypical' antipsychotic despite having a different receptor-affinity profile. However, there is a lack of studies about the efficacy and tolerability of amisulpride in latin populations. We decided to perform a 3-month open trial to determine these parameters on a sample of Mexican patients, using the five-factor model of schizophrenic psychopathology, a previously determined useful strategy for the evaluation of drug efficacy [32].
Method
Subjects were consecutively recruited from the inpatient and outpatient services from 7 centers in Mexico between August 2001 and January 2003. The study was conducted in accordance with the Good Clinical Practices and the Declaration of Helsinki. The study protocol was approved by an Institutional Review Board and Ethics Committee at each center. Written informed consent was obtained after the procedures had been fully and detailed explained to patients.
The sample comprised men and nonpregnant, nonlactating women aged 18 to 65 years who met diagnostic criteria of schizophrenia (as per DSM-IV), including patients on their first-episode psychosis (defined as their first psychiatric admission due to psychosis with a maximum duration of untreated illness of 5 years), patients with acute psychotic symptoms (defined by baseline PANSS total score of at least 60 with a minimum score of 4 on at least 2 items of the positive subscale) due to antipsychotic treatment noncompliance and patients requiring treatment switch due to lack of efficacy (positive and/or negative symptom persistence) or presence of adverse events. Patients were excluded if they had history of bipolar disorder, high risk for suicide or agitation/violence, concomitant medical or neurological illness (as per review of systems, and general physical examination) and DSM-IV defined current substance abuse or a history of substance dependence in the last six months.
Assessments
Diagnoses were based on the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) [33] according to DSM-IV criteria. Efficacy was rated using the dimensional models of schizophrenic symptoms, based on confirmatory 5 factor analysis (positive, negative, cognitive, excitement and depression/anxiety factors) of the Positive and Negative Syndrome Scale (PANSS) (each item rated from 1 to 7) in Mexican population [34], the Clinical Global Impression Scale (severity and improvement subscales rated from 0 to 7) [35] and the Calgary Depression Scale for Schizophrenia (CDSS) [36] for depressive symptoms. The efficacy of amisulpride was examined on other dimensions of schizophrenia, such as psychosocial functioning, with the Psychosocial Aptitude Rating Scale (PARS) (10 items rated from 0 to 10, where a total score of "0" denotes the least healthy end of the functioning range, and "100" the healthiest end) [37]. Tolerability was assessed with the Barnes Akathisia Scale (BAS) for akathisia [38], the Simpson Angus Scale (SAS) for EPS [39] and the Abnormal Involuntary Movements Scale (AIMS) for dyskinesia [40]. For each patient, the same investigator rated these measures at baseline, day 15, day 30 and day 90.
Treatment
This was an open label study using flexible doses (100 mg/d – 1200 mg/d) of amisulpride. Patients with predominantly negative symptoms, defined as a greater score on the negative than on the positive subscale in the PANSS, a negative subscale score >20, at least moderate global negative symptom score (items score as 4) and absence of depression [41] received an initial dose of 100 mg/d – 300 mg/d, while patients with predominantly positive symptoms (defined as a greater score on the positive than on the negative subscale in the PANSS) [41] received an initial dose ≥400 mg/d. During the 3-month treatment phase, the drug doses were adjusted based on clinical considerations, treatment response and side effects. Patients with lack of clinical response and/or presence of adverse events with the antipsychotic they were taking before were switched to amisulpride tapering off the previous antipsychotic agent and gradually titrating amisulpride to the established therapeutic dose. Biperiden (2–12 mg/d) was prescribed for the treatment of EPS, lorazepam (0.5–4 mg/d) for anxiety and akathisia, and antidepressant medication was allowed for the treatment of depressive symptoms.
Analysis
Demographic and clinical characteristics description was done with frequencies and percentages for categorical variables and with means and standard deviations (S.D.) for continuous variables. The effect of treatment was assessed using repeated measures ANOVA for continuous contrasts. Where appropriate, paired t-test (two tailed) were conducted to examine the differences. For the analysis of patients with predominantly positive symptoms and predominantly negative symptoms, changes from baseline to endpoint in PANSS dimensions and the PARS score were examined and compared using an analysis of covariance (ANCOVA) model adjusted for baseline scores. The improvement criteria a priori specified in the study required a ≥ 20% decrease from baseline in PANSS total score to study endpoint. However, we also examined the results with ≥ 30% decrease from baseline in PANSS total score as the criterion for response, as the latter may be more clinically meaningful. Chi square test (x2) was used for the comparison of improvement and response criteria between patients with predominantly positive symptoms and predominantly negative symptoms. The significance level for tests was established at p = 0.01 to decrease the Type 1 error as a consequence of multiple comparisons.
Results
A total of 80 patients were included, their demographic and clinical characteristics are described in table 1. The 3-month treatment phase was completed by 70 (87.5%) patients. The most frequent reason for premature withdrawal was the lack of clinical response (n = 6). The other reasons for discontinuation from the study were the lost to follow-up (n = 3), serious adverse events (severe EPS in one patient and overdosing in one patient) and non-compliance (n = 1). Non-compliance was determined when patients discontinued their treatment for a period longer than 1 week. Only three patients with first-episode were discontinued from the study. The data from these 70 patients remaining in the study is reported.
Table 1 Demographic and clinical characteristics of the sample
n (%)
Gender
Male 60 (75)
Female 20 (25)
Schizophrenia subtype
Paranoid 40 (50)
Undifferentiated 37 (46.2)
Disorganized 2 (2.5)
Residual 1 (1.3)
Inpatients 20 (25)
Outpatients 60 (75)
Switched from previous antipsychotic 48 (60)
Noncompliance 20 (25)
First Episode 12 (15)
Mean (S.D.)
Age 33.1 (10.3)
Age at onset 22.6 (6.7)
Length of illness (yr.) 10.5 (8.2)
PANSS total score 90.9 (23.8)
The mean initial dose of amisulpride was 335.7 mg/day (S.D. = 147.4, median 350 mg/day) and the mean dose at the end of the study was 357.1 mg/day (S.D. = 159.3, median 400 mg/day). The amisulpride doses were not different between first episode and chronic patients.
Efficacy data
Treatment with amisulpride was associated with a highly significant improvement on the five factors and on the total score of the PANSS. The improvement was observed from the end of the second week with a substantial reduction in the severity of the symptoms, patients reached a high level of improvement evaluated with CGI-I and this level was maintained at the end of third month. The depressive symptoms evaluated with CDSS were mild to moderate at baseline, the CDSS total score decreased significantly through the study (Table 2).
Table 2 Effects of treatment on PANSS factors, depressive symptoms and the clinical global impression.
Baseline Week 2 Week 4 Week 12 Statistics p value
PANSS Factors
Positive 25.7 (8.6) 18.2 (6.6) 14.7 (7.1) 13 (6.6) F(3,77) = 33 <0.001
Negative 26.8 (8.3) 19.3 (7.1) 16.3 (7.3) 13.8 (6.4) F(3,77) = 41.1 <0.001
Cognitive 19.8 (6.2) 14 (5.2) 12.8 (5.2) 11.2 (5) F(3,77) = 24.4 <0.001
Excitement 9.4 (4.1) 6.1 (2.8) 5.6 (3) 5.2 (2) F(3,77) = 17.6 <0.001
Depression/Anxiety 9.7 (4.3) 7.3 (2.9) 6.4 (3) 5.4 (2.6) F(3,77) = 21.6 <0.001
Total Score 91.5 (23.8) 78.4(16.9) 56 (21.9) 48.8 (20.1) F(3,77) = 43.6 <0.001
CDSS 3.9 (4.9) 2.4 (3) 1.7 (2.9) 1 (2.7) F(3,77) = 9.6 <0.001
CGI
Severity 4.7 (1.1) 3.4 (1.1) 2.8 (1.2) 2.3 (1.4) F(3,77) = 54.2 <0.001
Improvement 2.5 (0.9) 2.2 (1) 2.3 (1.4) F(2,78) = 2.1 0.12
Mean (S.D.)
PANSS, Positive and negative syndrome scale
CDSS, Total score of the Calgary depression scale for schizophrenia
CGI, Clinical global impression scale
Amisulpride treatment also improved the social functioning measured with the PARS (baseline = 47.6, S.D. = 18.6 vs 3-month = 73.9, S.D. = 17.2, t = 10.3, df = 69, p < 0.001) and none was an inpatient at the end of the study.
A higher frequency of patients with predominantly positive symptoms was found (n = 40, 57%). All negative symptom subjects were classified as having the undifferentiated subtype of schizophrenia according to DSM-IV criteria. The mean final doses were 343 mg/day (S.D. = 143.1, median 350 mg/day) for the predominantly negative group and 392.5 mg/day (S.D. = 183.1, median 400 mg/day) for the predominantly positive subjects. Amisulpride was associated to similar improvement in all the five-factors of the PANSS. The mean baseline and changes from baseline to endpoint in PANSS dimensions scores for both groups are shown in Table 3. The predominantly positive patients showed a lower psychosocial functioning at baseline when compared to the predominantly negative group (t = 3.58, df 68, p = 0.001), however, no statistically significant differences emerged between groups at the end of the study (predominantly negative = 72.5 SD = 20.19; predominantly positive = 75.07 SD = 14.74, t = 0.61, df 68, p = 0.53).
Table 3 Mean change from baseline to endpoint in secondary efficacy measures
PANSS Factors Predominantly Negative (n = 30) Predominantly Positive (n = 40) Statistic* Between Groups Statistic* During the Follow-up
Mean SD Mean SD
Positive
Baseline 19.5 6.8 30.4 6.8 F = 1.09, df 1, p = 0.29 F = 79.58, df 1, p < 0.001
Mean Change -7.2 10.1 -16.7 9.6
Negative
Baseline 30.8 8.6 23.8 6.8 F = 0.16, df 1, p = 0.68 F = 80.18, df 1, p < 0.001
Mean Change -16.3 11.2 -10.5 8.1
Cognitive
Baseline 18.5 6.7 20.7 5.7 F = 0.01, df 1, p = 0.91 F = 111.29, df 1, p < 0.001
Mean Change -7.1 9.4 -9.6 7.3
Excitement
Baseline 8.0 4.0 10.4 3.9 F = 2.63, df 1, p = 0.10 F = 304.76, df 1, p < 0.001
Mean Change -3.1 4.7 -4.9 4.7
Depression-Anxiety
Baseline 8.9 4.2 10.3 4.2 F = 0.16, df 1, p = 0.68 F = 171.77, df 1, p < 0.001
Mean Change -3.6 4.9 -4.7 5.1
Total score
Baseline 85.8 25.1 95.8 22.1 F = 0.05, df 1, p = 0.82 F = 94.80, df 1, p < 0.001
Mean Change -37.4 35.3 -46.6 28.4
* Based on analysis of covariance adjusted for baseline score.
Response rate analysis
The improvement rate in the total sample was 82.9% using the >20% reduction in PANSS total score while response rate was 75.7% using the ≥30% reduction in PANSS total score. Similar improvement rates were found in the predominantly negative patients (73.3%) and in the predominantly positive patients (90.0%) (x2 = 3.35, df 1, p = 0.07). In the same way, similar response rates were found in both groups (predominantly negative 66.7% vs. predominantly positive 82.5%) (x2 = 2.33, df 1, p = 0.12).
Tolerability data
Amisulpride significantly decreased the EPS and abnormal involuntary movements at the end of the study. There were no significant differences in akathisia's severity during the study, although it remained as mild. There was a trend toward a slight weight increase with the use of amisulpride during the 3-month treatment phase (Table 4).
Table 4 Summary of tolerability data during the treatment phase.
Baseline Week 2 Week 4 Week 12 Statistics p value
SAS 4.3 (5.9) 2.2 (4) 2.1 (3.7) 2 (3.7) F (3,77) = 5.1 <0.001
BAS 2.2 (3.2) 1.5 (2.8) 1.3 (2.2) 1.3 (2.2) F (3,77) = 2.2 0.09
AIMS 3.8 (5.9) 2.5 (4.8) 2.1 (3.9) 1.8 (4.1) F (3,77) = 5.0 <0.001
Weight (kg,) 70.7 (13.5) 70.5 (12.9) 70.9 (13.2) 71.4 (13.2) F (3,77) = 2.3 0.08
Mean (S.D.)
SAS, Total score of the Simpson Angus scale
BAS, Total Score of the Barnes akathisia scale
AIMS, Total score of the abnormal involuntary movements scale
A total of 10 patients (12.5%) required the use of biperiden (4.0 mg/day, S.D. = 3.1) for EPS control at baseline, only 2 patients (2.8%) continued on anticholinergic medication (1.5 mg/day, S.D. = 0.70) at the end of the study. In addition 2 patients (2.5%) required antidepressants (paroxetine and reboxetine, respectively) at the beginning of the study but none at the end of the study. Sixteen 16 patients (20%) required lorazepam for anxiety or insomnia at baseline, but only 3 (4.2%) required this drug at the end of the study.
Akathisia was the main adverse event reported during the study (n = 4), followed by headache (n = 1), insomnia (n = 1) and amenorrhea (n = 1). All side effects were reported as mild.
Discussion
The factor-analysis studies of the PANSS produced five factors (positive, negative, cognitive, excitability/hostility and depression/anxiety), this solution fits better to the multidimensional model described for the psychopathology of schizophrenia [42-44]. Our results demonstrated a substantial benefit with amisulpride on the five dimensions of psychopathology based on confirmatory factor analysis of the PANSS in Mexican schizophrenic patients. This confirms the properties of amisulpride as an atypical antipsychotic given its efficacy and low frequency of EPS showed in this study. The findings of this study resemble previous reports of the efficacy of amisulpride among the dimensions obtained from factor analysis of BPRS [22,28,29].
Amisulpride also improved the psychosocial functioning of the patients and was associated with discharge from the hospital in subjects who were initially inpatients. These two variables could be associated with the drug's efficacy, better tolerability, or both. It has been observed that atypical antipsychotics improve the functioning and reduce the global negative symptoms; amisulpride has shown similar efficacy to the reported of risperidone [23,26] and olanzapine [24,27] in the treatment of these symptoms.
Similarly to previous reports, amisulpride improved the depressive [22,29] and cognitive symptoms [30,31]. This finding also increase the evidence of the atypical properties of amisulpride, despite its action as D2/D3 antagonist without affinity with other neurotransmitter systems.
In the current study, amisulpride produced substantial and significant reductions in psychopathology during the follow-up both in patients classified with predominantly negative or predominantly positive symptoms, without statistically significant difference between the groups. This finding could be explained by the dual pharmacodynamic effect of amisulpride [7-9]. Although both improvement and clinical response were observed in the groups according to predominant symptoms; patients with predominantly positive symptoms showed a trend to improve in higher rates.
The mean dose used in this study was lower than the doses reported in other studies which included patients with acute exacerbation of positive symptoms (600 mg/day) [45], given that the present sample included patients with predominantly negative symptoms, which have been reported to respond to lower doses of amisulpride. However, the final doses for patients with predominantly negative symptoms and predominantly positive symptoms did not differ, probably because some patients from the first group showed an increase on the severity of positive symptoms during the follow up. The most frequently prescribed dose was 400 mg/day for the treatment of patients with acute symptoms and first episode of psychosis.
Amisulpride was well tolerated, the need for concomitant medication was reduced at the end of the study and the reported adverse events were mild. The total weight increase was 0.7 kg during the 3-month treatment phase. This increase was similar to the reported in other studies [46] and lower to the 1.4 kg increase reported in previous meta-analysis [47]. This finding could be explained by the fact that in 20 patients (28.6%) a mean 2.1 kg of weight decrease was registered. Indeed, it has been demonstrated in randomized 6 months follow-up studies that amisulpride is associated with less weight gain than risperidone or olanzapine [25,27].
The observed decrease on the abnormal involuntary movements at the end of the study could also be explained by the treatment switch. Although only one patient reported an adverse event associated with hyperprolactinemia, this adverse event could not be rejected since in this study prolactin levels were not regularly measured. Amisulpride tended to increase prolactin levels when compared with other atypical antipsychotics [48,49]. Some studies reported few adverse events related with hiperprolactinemia [50], this is probably due to the decrease of prolactin levels with the long term treatment [51].
The findings of this study were similar to those observed in other international clinical trials with amisulpride, but should be considered with caution since one of its limitations was the open label design with a short term follow-up. Another limitation was that the outpatients outnumbered the inpatients, since the high cost of treatment in Mexico does not permit to afford the inpatient services. This sample distribution could not allow the results to be generalized to other countries where subjects with moderate or severe symptoms are treated as inpatients. Further studies should include adequate instruments for the assessment of cognitive symptoms as well as laboratory measures of prolactin, glucose and lipids levels.
Conclusion
The efficacy and tolerability profile of amisulpride in a Mexican population of schizophrenic patients is similar to that reported with other second generation antipsychotics, leading to consider this drug as first line for the treatment of schizophrenia.
Competing interests
This study was supported by Sanofi-Synthelabo. MHE has been member of advisory board of Bristol-Myers Squibb and has received funding from Pfizer Inc, Eli Lilly, Bristol-Myers Squibb and Sanofi-Synthelabo and he has served on the speakers bureau of Sanofi-Synthelabo, Bristol-Myers Squibb, Janssen Cilag and AstraZeneca Pharmaceuticals. RA has been a consultant and member of advisory board of Bristol-Myers Squibb and has received funding from Bristol-Myers Squibb, Sanofi-Synthelabo, Janssen Cilag and Theodore and Vada Stanley Foundation; he has served on the speakers bureau of Sanofi-Synthelabo, Bristol-Myers Squibb, and Janssen Cilag. IST is affiliated to Sanofi-Synthelabo.
Authors' contributions
RA, MHE and IST drafted the manuscript. AF performed the statistical analysis. RA, MHE and IST participated in the study 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 gratefully acknowledge the participation of the researchers from de Mexican Observational Study of Schizophrenia with Amisulpride (MOSSA): Juan Ignacio Rosales, Wazcar Verduzco, Rodrigo Garnica, Andrés López, Alejandra Victoria. We also wish to acknowledge the Sanofi-Synthelabo Research Team, in particular Margarita Murrieta-Aguttes and Robert Manfredi for their critical review and contributions to this study. This study was sponsored by Sanofi-Synthelabo.
==== Refs
Lieberman JA Atypical antipsychotic drugs as a first-line treatment of schizophrenia: a rationale and hypothesis Journal of Clinical Psychiatry 1996 57 68 71 8941173
Meltzer HY What's atypical about atypical antipsychotic drugs? Curr Opin Pharmacol 2004 4 53 57 15018839 10.1016/j.coph.2003.09.010
Meltzer HY McGurk SR The effects of clozapine, risperidone, and olanzapine on cognitive function in schizophrenia Schizophr Bull 1999 25 233 255 10416729
Meltzer HY Matsubara S Lee JC The ratios of serotonin2 and dopamine2 affinities differentiate atypical and typical antipsychotic drugs Psychopharmacol Bull 1989 25 390 392 2576319
Kapur S Seeman P Does fast dissociation from the dopamine d(2) receptor explain the action of atypical antipsychotics?: A new hypothesis Am J Psychiatry 2001 158 360 369 11229973 10.1176/appi.ajp.158.3.360
Ichikawa J Ishii H Bonaccorso S Fowler WL O'Laughlin IA Meltzer HY 5-HT(2A) and D(2) receptor blockade increases cortical DA release via 5-HT(1A) receptor activation: a possible mechanism of atypical antipsychotic-induced cortical dopamine release J Neurochem 2001 76 1521 1531 11238736 10.1046/j.1471-4159.2001.00154.x
Schoemaker H Claustre Y Fage D Rouquier L Chergui K Curet O Oblin A Gonon F Carter C Benavides J Scatton B Neurochemical characteristics of amisulpride, an atypical dopamine D2/D3 receptor antagonist with both presynaptic and limbic selectivity J Pharmacol Exp Ther 1997 280 83 97 8996185
Perrault G Depoortere R Morel E Sanger DJ Scatton B Psychopharmacological profile of amisulpride: an antipsychotic drug with presynaptic D2/D3 dopamine receptor antagonist activity and limbic selectivity J Pharmacol Exp Ther 1997 280 73 82 8996184
Scatton B Claustre Y Cudennec A Oblin A Perrault G Sanger DJ Schoemaker H Amisulpride: from animal pharmacology to therapeutic action Int Clin Psychopharmacol 1997 12 Suppl 2 S29 36 9218165
Malmberg A Jackson DM Eriksson A Mohell N Unique binding characteristics of antipsychotic agents interacting with human dopamine D2A, D2B, and D3 receptors Mol Pharmacol 1993 43 749 754 8099194
Martinot JL Paillere-Martinot ML Poirier MF Dao-Castellana MH Loc'h C Maziere B In vivo characteristics of dopamine D2 receptor occupancy by amisulpride in schizophrenia Psychopharmacology (Berl) 1996 124 154 158 8935811
Xiberas X Martinot JL Mallet L Artiges E Canal M Loc'h C Maziere B Paillere-Martinot ML In vivo extrastriatal and striatal D2 dopamine receptor blockade by amisulpride in schizophrenia J Clin Psychopharmacol 2001 21 207 214 11270918 10.1097/00004714-200104000-00013
Loo H Poirier-Littre MF Theron M Rein W Fleurot O Amisulpride versus placebo in the medium-term treatment of the negative symptoms of schizophrenia Br J Psychiatry 1997 170 18 22 9068769
Colonna L Saleem P Dondey-Nouvel L Rein W Long-term safety and efficacy of amisulpride in subchronic or chronic schizophrenia. Amisulpride Study Group Int Clin Psychopharmacol 2000 15 13 22 10836281
Paillere-Martinot ML Lecrubier Y Martinot JL Aubin F Improvement of some schizophrenic deficit symptoms with low doses of amisulpride Am J Psychiatry 1995 152 130 134 7802104
Boyer P Lecrubier Y Puech AJ Dewailly J Aubin F Treatment of negative symptoms in schizophrenia with amisulpride Br J Psychiatry 1995 166 68 72 7894879
Speller JC Barnes TR Curson DA Pantelis C Alberts JL One-year, low-dose neuroleptic study of in-patients with chronic schizophrenia characterised by persistent negative symptoms. Amisulpride v. haloperidol Br J Psychiatry 1997 171 564 568 9519098
Danion JM Rein W Fleurot O Improvement of schizophrenic patients with primary negative symptoms treated with amisulpride. Amisulpride Study Group Am J Psychiatry 1999 156 610 616 10200742
Moller HJ Boyer P Fleurot O Rein W Improvement of acute exacerbations of schizophrenia with amisulpride: a comparison with haloperidol. PROD-ASLP Study Group Psychopharmacology (Berl) 1997 132 396 401 9298518 10.1007/s002130050361
Wetzel H Grunder G Hillert A Philipp M Gattaz WF Sauer H Adler G Schroder J Rein W Benkert O Amisulpride versus flupentixol in schizophrenia with predominantly positive symptomatology -- a double-blind controlled study comparing a selective D2-like antagonist to a mixed D1-/D2-like antagonist. The Amisulpride Study Group Psychopharmacology (Berl) 1998 137 223 232 9682999 10.1007/s002130050614
Puech A Fleurot O Rein W Amisulpride, and atypical antipsychotic, in the treatment of acute episodes of schizophrenia: a dose-ranging study vs. haloperidol. The Amisulpride Study Group Acta Psychiatr Scand 1998 98 65 72 9696517
Muller MJ Wetzel H Benkert O Differential effects of high-dose amisulpride versus flupentixol on latent dimensions of depressive and negative symptomatology in acute schizophrenia: an evaluation using confirmatory factor analysis Int Clin Psychopharmacol 2002 17 249 261 12177587 10.1097/00004850-200209000-00005
Peuskens J Bech P Moller HJ Bale R Fleurot O Rein W Amisulpride vs. risperidone in the treatment of acute exacerbations of schizophrenia. Amisulpride study group Psychiatry Res 1999 88 107 117 10622347 10.1016/S0165-1781(99)00075-X
Martin S Ljo H Peuskens J Thirumalai S Giudicelli A Fleurot O Rein W A double-blind, randomised comparative trial of amisulpride versus olanzapine in the treatment of schizophrenia: short-term results at two months Curr Med Res Opin 2002 18 355 362 12442883 10.1185/030079902125001128
Sechter D Peuskens J Fleurot O Rein W Lecrubier Y Amisulpride vs. risperidone in chronic schizophrenia: results of a 6-month double-blind study Neuropsychopharmacology 2002 27 1071 1081 12464464 10.1016/S0893-133X(02)00375-5
Hwang TJ Lee SM Sun HJ Lin HN Tsai SJ Lee YC Chen YS Amisulpride versus risperidone in the treatment of schizophrenic patients: a double-blind pilot study in Taiwan J Formos Med Assoc 2003 102 30 36 12684609
Mortimer A Martin S Loo H Peuskens J A double-blind, randomized comparative trial of amisulpride versus olanzapine for 6 months in the treatment of schizophrenia Int Clin Psychopharmacol 2004 19 63 69 15076013 10.1097/00004850-200403000-00002
Muller MJ Wetzel H Eich FX Rein W Puech A Benkert O Dose-related effects of amisulpride on five dimensions of psychopathology in patients with acute exacerbation of schizophrenia J Clin Psychopharmacol 2002 22 554 560 12454554 10.1097/00004714-200212000-00004
Peuskens J Moller HJ Puech A Amisulpride improves depressive symptoms in acute exacerbations of schizophrenia: comparison with haloperidol and risperidone Eur Neuropsychopharmacol 2002 12 305 310 12126869 10.1016/S0924-977X(02)00031-7
Vaiva G Thomas P Llorca PM Dupont S Cottencin O Devos P Mazas O Rascle C Steinling M Goudemand M SPECT imaging, clinical features, and cognition before and after low doses of amisulpride in schizophrenic patients with the deficit syndrome Psychiatry Res 2002 115 37 48 12165366
Wagner M Quednow BB Westheide J Schlaepfer TE Maier W Kuhn KU Cognitive improvement in schizophrenic patients does not require a serotonergic mechanism: randomized controlled trial of olanzapine vs amisulpride Neuropsychopharmacology 2005 30 381 390 15578006 10.1038/sj.npp.1300626
Marder SR Davis JM Chouinard G The effects of risperidone on the five dimensions of schizophrenia derived by factor analysis: combined results of the North American trials J Clin Psychiatry 1997 58 538 546 9448657
First MB Spitzer RL Gibbon M Williams JBW Structured Clinical Intereview for DSM-IV Axis I Disorders (SCID), Clinician Version 1996 Washington, DC, American Psychiatric Press
Fresan A De la Fuente-Sandoval C Loyzaga C Garcia-Anaya M Meyenberg N Nicolini H Apiquian R A forced five-dimensional factor analysis and concurrent validity of the Positive and Negative Syndrome Scale in Mexican schizophrenic patients Schizophr Res 2005 72 123 129 15560957 10.1016/j.schres.2004.03.021
Clinical Global Impressions (CGI). In Guy W. ECDEU Assessment Manual of Psychopharmacology. US Department of Health, Education and Welfare Publication (ADM) 76-338. Rockville, Md: National Institute of Mental Health, 1976: 218-222
Addington D Addington J Maticka-Tyndale E Joyce J Reliability and validity of a depression rating scale for schizophrenics Schizophr Res 1992 6 201 208 1571313 10.1016/0920-9964(92)90003-N
Robert P Braccini T Vitali P Darcourt G Psycho-social aptitude rating scale (PARS): Presentation and validation Psychologie Medicale 1987 19 1761 1765
Barnes TR A rating scale for drug-induced akathisia Br J Psychiatry 1989 154 672 676 2574607
Simpson GM Angus JW A rating scale for extrapyramidal side effects Acta Psychiatr Scand 1970 212 11 19
Abnormal Involuntary Movement Scale (AIMS). In Guy W. ECDEU Assessment Manual of Psychopharmacology. US Department of Health, Education and Welfare Publication (ADM) 76-338. Rockville, Md: National Institute of Mental Health, 1976: 534-537
Potkin SG Alva G Fleming K Anand R Keator D Carreon D Doo M Jin Y Wu JC Fallon JH A PET study of the pathophysiology of negative symptoms in schizophrenia. Positron emission tomography Am J Psychiatry 2002 159 227 237 11823264 10.1176/appi.ajp.159.2.227
Lindenmayer JP Grochowski S Hyman RB Five factor model of schizophrenia: replication across samples Schizophr Res 1995 14 229 234 7766534 10.1016/0920-9964(94)00041-6
Lancon C Aghababian V Llorca PM Auquier P Factorial structure of the Positive and Negative Syndrome Scale (PANSS): a forced five-dimensional factor analysis Acta Psychiatr Scand 1998 98 369 376 9845175
Fitzgerald PB de Castella AR Brewer K Filia K Collins J Davey P Rolfe T Kulkarni J A confirmatory factor analytic evaluation of the pentagonal PANSS model Schizophr Res 2003 61 97 104 12648740 10.1016/S0920-9964(02)00295-5
Lecrubier Y Azorin M Bottai T Dalery J Garreau G Lemperiere T Lisoprawski A Petitjean F Vanelle JM Consensus on the Practical Use of Amisulpride, an Atypical Antipsychotic, in the Treatment of Schizophrenia Neuropsychobiology 2001 44 41 46 11408792 10.1159/000054913
Leucht S Wagenpfeil S Hamann J Kissling W Amisulpride is an "atypical" antipsychotic associated with low weight gain Psychopharmacology (Berl) 2004 173 112 115 14647963 10.1007/s00213-003-1721-6
Taylor DM McAskill R Atypical antipsychotics and weight gain--a systematic review Acta Psychiatr Scand 2000 101 416 432 10868465 10.1034/j.1600-0447.2000.101006416.x
Stanniland C Taylor D Tolerability of atypical antipsychotics Drug Saf 2000 22 195 214 10738844
Mortimer AM How do we choose between atypical antipsychotics? The advantages of amisulpride Int J Neuropsychopharmacol 2004 7 Suppl 1 S21 5 14972081 10.1017/S1461145704004134
Coulouvrat C Dondey-Nouvel L Safety of amisulpride (Solian): a review of 11 clinical studies Int Clin Psychopharmacol 1999 14 209 218 10468313
Schlosser R Grunder G Anghelescu I Hillert A Ewald-Grunder S Hiemke C Benkert O Long-term effects of the substituted benzamide derivative amisulpride on baseline and stimulated prolactin levels Neuropsychobiology 2002 46 33 40 12207145 10.1159/000063574
| 15869707 | PMC1090597 | CC BY | 2021-01-04 16:33:02 | no | BMC Psychiatry. 2005 May 3; 5:22 | utf-8 | BMC Psychiatry | 2,005 | 10.1186/1471-244X-5-22 | oa_comm |
==== Front
BMC Womens HealthBMC Women's Health1472-6874BioMed Central London 1472-6874-5-41582321110.1186/1472-6874-5-4Study ProtocolA cluster randomized controlled trial of a behavioral intervention to facilitate the development and implementation of clinical practice guidelines in Latin American maternity hospitals: the Guidelines Trial: Study protocol [ISRCTN82417627] Althabe Fernando [email protected] Pierre [email protected] Eduardo [email protected]án José M [email protected] Nora [email protected] Linda [email protected] Norman [email protected] Nancy [email protected] the Guidelines Trial Group 1 Perinatal Research Unit (PRU), Montevideo, Uruguay2 School of Public Health and Tropical Medicine, Tulane University, New Orleans, Louisiana, USA3 Institute of Clinical Effectiveness and Health Policy (IECS), Buenos Aires, Argentina4 RTI International, North Carolina, USA5 Center for Research for Mothers and Children, National Institute of Child Health & Human Development, USA2005 11 4 2005 5 4 4 9 3 2005 11 4 2005 Copyright © 2005 Althabe et al; licensee BioMed Central Ltd.2005Althabe 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
A significant proportion of the health care administered to women in Latin American maternity hospitals during labor and delivery has been demonstrated to be ineffective or harmful, whereas effective interventions remain underutilized. The routine use of episiotomies and the failure to use active management of the third stage of labor are good examples.
Methods/Design
The aim of this trial is to evaluate the effect of a multifaceted behavioral intervention on the use of two evidence-based birth practices, the selective use of episiotomies and active management of the third stage of labor (injection of 10 International Units of oxytocin). The intervention is based on behavioral and organizational change theories and was based on formative research. Twenty-four hospitals in three urban districts of Argentina and Uruguay will be randomized. Opinion leaders in the 12 intervention hospitals will be identified and trained to develop and implement evidence-based guidelines. They will then disseminate the guidelines using a multifaceted approach including academic detailing, reminders, and feedback on utilization rates. The 12 hospitals in the control group will continue with their standard in-service training activities. The main outcomes to be assessed are the rates of episiotomy and oxytocin use during the third stage of labor. Secondary outcomes will be perineal sutures, postpartum hemorrhages, and birth attendants' opinions.
==== Body
Background
In Latin American maternity hospitals, a significant proportion of the health care administered to women during labor and delivery has been demonstrated to be ineffective or harmful, whereas proven effective interventions are underutilized. For example, routine episiotomy use has been documented to be useless and even harmful [1], yet the episiotomy rate in vaginal births of primiparous women in Latin American hospitals is 92 %. [2]. The caesarean section rate in Latin American countries is above 25% for the region as a whole, resulting in approximately 850,000 unnecessary cesarean sections performed each year [3]. Simultaneously, many of the birth practices verified as beneficial are not routinely used, e.g. active management of the third stage of labor has been proven effective to prevent postpartum hemorrhage, one of the leading causes of maternal deaths in the developing world [4]. In a survey in 19 maternity services of Uruguay and Argentina, 17 (89.5%) used expectant management as the standard form of care (unpublished observation). Another practice that has been proven to be beneficial and without adverse effects is a companion accompanying women in labor and delivery [5]; however, only 21% of women had a companion in the survey of the 19 hospitals mentioned above. The administration of enemas at admission in labor, perineal shaving, and systematic intravenous infusion are other examples of practices not supported by scientific evidence which are still in widespread use in Latin American hospitals [6,7].
Barriers to the adoption of evidence based birth practices
Why, despite scientific evidence and active dissemination of scientific information, are harmful or unnecessary procedures (or both) still used, whereas other beneficial procedures are ignored?
It is postulated that a wide variety of barriers can hinder practitioners from adhering to evidence-based practices, thus a theoretical framework may help delineate these barriers and possibly help target specific interventions to address appropriate practice patterns. Cabana et al. published a comprehensive systematic review about the barriers to physician adherence to practice guidelines and developed a theoretical approach to the barriers [8].
They reviewed 76 articles including five qualitative studies and data from 120 surveys to identify possible barriers to guideline adherence. Barriers were classified into seven general categories: barriers affecting physician knowledge (lack of awareness and lack of familiarity); those affecting attitudes (lack of agreement, lack of self-efficacy, lack of outcome expectancy and the inertia of previous practices); and those affecting behavior (external barriers).
One study included in Cabana's review was a survey of 38 childbirth-related organizations regarding the use of Effective Care in Pregnancy and Childbirth[9]. The survey asked what the organizations considered the main barriers to effective transfer of research information.. Every organization specifically cited the practitioners' failure to keep up with the literature, a lack of resources, and the low value placed on research evidence as major obstructions in the implementation of research findings.
Making research information accessible and comprehensible to practitioners is an important mechanism to increasing awareness and familiarity with best clinical practices; however, it is rarely independently sufficient to overcome other categories of barriers and to effectively change practice patterns [10].
Interventions to change professional practice
Many methods designed to change medical behavior have been developed and used in industrialized countries. Several systematic reviews on that subject concluded that there are no "magic bullets" to change professional behavior and that the best approach is to combine several strategies, such as the use of local opinion leaders, convening workshops, providing outreach visits (academic detailing), and reminders, as well as the use of audits and a provision for feedback. [11,12]. A more recent review did not support this conclusion providing evidence that some single interventions might be as effective as the combined strategies, but also acknowledging that there is a need for further and improved research regarding such interventions thus providing a strong rationale and theoretical basis for the selection of their components [13,14].
Among the trials considered in the above reviews were strategies directed at behavior change of birth attendants in North America and Europe, including distribution of educational materials, employing local opinion leaders, and the use of audits and feedback directed at increasing the number of vaginal births after cesarean section [15]; nurse opinion leaders to reduce rates of epidural analgesia by increasing the amount of support to women in labor [16]; educational outreach visits to introduce Cochrane Reviews to facilitate evidence-based use of specific birth practices [17]; employing local opinion leaders, grand round lectures, chart reminders, group interactive discussions, and audits to increase the use of corticosteroids for enhancing fetal lung maturation prior to a preterm birth [18].
Very few trials of a similar nature have been performed in developing countries, and among these results were inconclusive [19,20]. To our knowledge, no randomized controlled trial has been performed to evaluate the effectiveness of an intervention to implement evidence-based birth practices in Latin American countries.
Our hypothesis is that a multifaceted intervention designed to increase birth attendant concern about the effectiveness of routinely used birth practices, to provide them with resources and skills to access, interpret, and develop evidence-based clinical guidelines, and to establish mechanisms facilitated through key hospital leaders, to implement the guidelines and sustain them over time, will increase the use of evidence-based birth practices in Latin American hospitals. The primary objective of this trial is to evaluate the effect of this multifaceted behavioral intervention on the use of two evidence-based birth practices, the selective use of episiotomies and active management of the third stage of labor (injection of 10 IU of oxytocin). Secondary objectives are to evaluate the effect of the intervention on rates of post-partum hemorrhage, use of perineal sutures and to document readiness to change among birth attendants.
Additionally, through this process we intend to improve the research capacities of a network of Latin American hospitals, and thus increase their ability to perform quality local and collaborative research studies.
Methods/design
The study design is a two-arm cluster randomized controlled trial with hospitals as units of randomization. Hospitals will be randomized to receive a multifaceted behavioral intervention directed at developing and implementing guidelines about episiotomy use and management of the third stage of labor, or to a control group that will continue with their usual in-service training activities. Data will be collected from eligible patients at three data collection periods, one baseline period before randomization, a post-intervention period just after the end of the intervention, and a long-term follow up one year after the end of the intervention (fig 1).
Figure 1 Trial Profile
Participating hospitals
Twenty four public hospitals in Argentina and Uruguay that had at least 1000 vaginal deliveries per year, and had no explicit policy for selective episiotomy or active management of third stage of labor were invited to participate. Those pre-selected hospitals that agreed to participate, collected baseline data. From the 24 hospitals, only 19 will be randomized because the analysis of their baseline data confirmed episiotomy rates in women with single vaginal births higher than 20%, and active management of the third stage of labor 25% or less.
Randomization procedures
Since only 19 hospitals will participate in the intervention trial, random allocation may not provide adequate balance in the groups regarding important prognostic variables [21]. We will use a minimization procedure to assure balance between intervention and control hospitals on five important prognostic variables: 1) baseline episiotomy and active management rates, 2) residency programs, 3) country (Argentina-Uruguay), 4) hospital size (number of births per year), and 5) region within the country. Minimization has the advantage that it can match small numbers of similar units with respect to several unit characteristics [22], and may be considered methodologically equivalent to randomization [23].
Because all hospitals will enter the study at the same time, all prognostic variables are known in advance. A computer algorithm will be written to generate a large number of allocation sequences and to select only those sequences that minimize the imbalance between the two groups of hospitals. One of those sequences will be randomly selected by the computer and will be used to assign hospitals to the control or intervention groups. Both analysis of the baseline data and the allocation procedure will be done by an independent data center (Research Triangle Institute, North Carolina, USA) and their results communicated to the study coordinating unit. Thus, there will be a clear separation between the generator of the intervention allocation and the study coordination [24].
The intervention
General outline
The intervention is based on the stages of change and organizational change theories and tailored by formative research [25-27]. Opinion leaders in the 12 intervention hospitals will be identified by their peers through a specific questionnaire and trained in a five day workshop to develop and implement evidence-based guidelines. They will then use a specified multifaceted approach to disseminate, implement, and maintain the guidelines in their hospitals. The 18-month intervention will include training to perform the evidence-based best practices, use of a web site, academic detailing, reminders, and feedback on utilization rates. Those components which will form the intervention are noted in Table 1 and are described in detail below.
Table 1 Components of the intervention
Selection of opinion leaders
- Teams of 3–6 birth attendants per hospital
- Selected by peer nomination
Interactive workshops
- Opinion leaders teams will participate in a 5 day workshop
- Objectives:
- To learn the need of an evidence based clinical practice
- To develop simple evidence based guidelines about episiotomy use and management of the third stage of labor
- To identify the barriers for the adoption of those guidelines at the hospital level
- To learn how to overcome barriers and to implement the guidelines
- To adapt and organize the dissemination and implementation of the guidelines in their hospitals
Academic detailing
- Dissemination of the guidelines to hospital birth attendants in small groups and individual discussions
- Identification of barriers to implement the guidelines
- Adaptation and organization of implementation activities working closely with birth attendants.
Training on how to comply with the recommended practices
- Training in manual abilities with videos, anatomical models and patients. One day workshop.
Reminders
- Placing reminders of selective episiotomy and active management of the third stage of labor in labor and delivery wards, clinical records, and surgical packages.
Audit & feed back
- Monthly reports of hospital episiotomy and active management rates to be distributed to every birth attendant.
Information technology
- Each hospital in the intervention group will receive a computer with internet access
- A specific web site will be developed to facilitate the access to study manuals and guidelines, sources of evidence-based health care literature (Reproductive Health Library, Cochrane Library), and communication among hospitals and study coordinators
Formative research
With the aim of refining the intervention strategy and better integrating the intervention into the hospitals' routine, a qualitative study was carried out during the preparatory phase of the trial. We conducted three in-depth interviews with upper level physician administrators (responsible for the clinical and administrative management of the departments), three focus groups with second level physicians (responsible for the individual provision of clinical care to users) and three focus groups with midwives. Additionally we conducted three focus groups with pregnant women to explore their perspectives regarding the health care they receive and their expectations of the birth practices routinely used.
Selection of opinion leaders
Each intervention hospital will select a team of opinion leaders who will work collaboratively during the intervention period. Each team will be composed by three to six professionals per hospital, with representation by obstetricians, residents and midwives. The department chair will be invited to designate one professional not included in the nominated team. Peer nomination will identify opinion leader team members among the professional staff of the maternity hospital using a previously validated sociometric questionnaire [28].
Guideline development and implementation workshops
All opinion leader teams will participate in a regional five-day workshop. There the teams will learn about the scientific basis for evidence based clinical practice, they will develop simple evidence based guidelines for episiotomy use and for the management of the third stage of labor. They will additionally suggest the main barriers for the implementation of those guidelines, and will learn and practice ways to overcome such barriers in the context of the planned intervention. Finally, each team will adapt the planned guideline dissemination and implementation activities to the characteristics of their own hospital, and will develop an organizational approach directed at the interventions.
The workshops will be given in five consecutive days, in a selected site outside the hospital walls, and will be conducted by trainers with previous experience in conducting related workshops in Latin American countries that will act as tutors, together with the country coordinators. Each tutor will guide a group with no more than 10 participants.
Web portal
Each hospital in the intervention group will receive a computer, access to the internet, the last issue of the WHO Reproductive Health Library [29] and access to the Cochrane Library in Spanish [30]. A study web site will be developed to facilitate the easy access to guidelines, workshop manuals, medical literature and electronic libraries and databases. This tool will be presented to the teams during the workshops, and they will later introduce the portal to each intervention hospital staff as part of the dissemination activities. The portal will be for the exclusive use of the professional staff at the intervention hospitals. A password access system will be in place during the intervention.
Training on manual skills
After completion of training, each opinion leader team will participate in a one-day workshop on clinical management skills (how to assist a delivery without episiotomy, and active management of the third stage of labor) using videos, anatomical models, and patients, conducted by the country coordinator or a trainer with previous experience. This workshop is based on the JHPIEGO workshop [31] currently being organized in several Latin American countries.
Dissemination of guidelines at the hospital and identification of barriers
Each opinion leader team disseminates the guidelines among the professional staff at its respective hospitals. They will primarily utilize academic detailing as the strategy for this purpose [32]. The team will present and discuss the guidelines with all birth attendants, providing to each one the opportunity to discuss them and to identify specific barriers; the birth attendants will adapt the planned activities for overcoming the barriers and develop an implementation timetable. Finally, the team will present the web portal, and will select a group of potential early adopters (volunteers) who appear enthusiastic for participation in the implementation of the guidelines.
Implementation and maintenance of guidelines
The objective of this component is to facilitate deliveries by birth attendants according to the recommendations developed in the workshops. Three strategies will be carried out to achieve this aim:
a) Training on how to do the recommended practices: all birth attendants will receive one training session given by the opinion leaders, on anatomical models and on patients.
b) Placing reminders: there will be short messages that remind birth attendants to consider the two evidence-based birth practices of interest:
• Active management reminders: to be placed in the partograph and as posters in the delivery ward.
• Selective episiotomy reminders: to be placed on the surgical package for assisting delivery and in the partograph.
The text of the reminders will be prepared by the opinion leader teams according to the designed recommendation and will take into account the characteristics and barriers of each specific hospital.
c) Monthly reporting of the episiotomy rates and the use of oxytocin for management of the third stage of labor at the hospital level: This information will be produced by their Perinatal Information System (PIS) [33] and distributed to birth attendants.
The intervention will be pilot tested in two hospitals similar to the hospitals that will participate in the trial.
Activities in control hospitals
Birth attendants from hospitals in the control group will receive no intervention after randomization other than their standard in-service training activities and standard sources of information. We intend to provide birth attendants with the guideline development component, a computer and bibliography following the completion of the study.
Outcome measures
The primary outcomes are aimed at changing the birth attendants' behavior relative to the rate of episiotomy and the rate of injection use of 10 I.U. of oxytocin during third stage of labor. Secondary clinical outcomes will be studied to estimate the health impact of the intervention, including the rate of perineal sutures use and the incidence of postpartum hemorrhage >500 ml. These outcomes will be assessed in singleton vaginal deliveries. We will also measure the provider's readiness to change with respect to episiotomies and management of third stage of labor with a specific questionnaire.
All outcomes will be assessed for a three-month period at three time points: at baseline (before randomization), after the end of the intervention, and the primary outcomes one year after the end of the intervention to measure the long term effects of the intervention. Interim analysis will be performed as required by the Data and Safety Monitoring Committee.
Process measures
Process data will be collected during the intervention at the active intervention hospitals, with the objective of allowing for early detection and correction of implementation problems. This will allow detection of implementation problems if the intervention is not effective or facilitation of the intervention if it is effective. Three categories of process measures will be considered: program inputs, guideline implementation activities, and birth attendants' reactions.
Data collection
Although the hospital will be the unit of analysis, we shall collect individual patient data in order to assure data quality. The goal is to accumulate data on approximately 300–500 deliveries from each hospital. All the hospitals in the project use a standard perinatal clinical record form (PIS form) [33]. This form includes the data of the obstetric history, prenatal care, labor, delivery, and neonatal outcomes. To obtain the additional necessary data to calculate the outcome variables, we will implement a modified version of the PIS clinical record that will add some variables at the bottom of the form to be addressed during the data collection periods. This approach will reduce to a minimum extra activities associated with data collection by the birth attendants during the study.
Blood loss during the third stage of labor (after delivery) will be measured from all vaginal deliveries during the data collection periods. A plastic, transparent, collection drape will be employed. At the end of the blood collection period, the amount of blood loss will be measured by pouring the contents of the drape into a calibrated pitcher. The amount of blood loss will be recorded in the data form together with the time (beginning and end) of the blood collection.
To measure provider's readiness to change, a self-administered questionnaire was designed and will be completed by all physicians and midwives at all participating hospitals, before randomization and at the end of the intervention.
Data management
The data collection system to determine outcomes will be independent from the implementation of the intervention. Given the nature of the intervention, we cannot blind the randomization, and data collectors will likely know if they are part of an intervention/control hospital. In addition, intervention hospitals will monitor and record clinical data (i.e., episiotomy rate and active management of the third stage of labor) through their routine data collection system. As a consequence, it is expected that the proposed intervention will improve the capacity of intervention hospitals to collect and review clinical data with the potential of introducing bias in the outcome assessment. To minimize this bias, the data collection system will as much as possible be isolated from the intervention instruments.. In practical terms, this means that data collectors at the hospital level, the data supervisors and the associated computer software used for data collection will not be associated with intervention activities.
Each participating hospital will receive a computer with an Internet access to be used for data collection purposes. A custom developed data management software program will be installed in each computer and will be used by trained and certified in-hospital data collectors, for distributed data entry, data validation and data transmission to the coordinating unit. The system provides a secure environment for confidential medical information and the access to it will be limited to authorized individuals. No personal identifiers will be transmitted from the hospital to the data coordinating unit
Sample size
The sample size was estimated based upon the hospital being the unit of analysis. A preliminary analysis of data from Argentinean hospitals [2] documented a baseline frequency of episiotomies in vaginal deliveries of 42%, with a standard deviation of 11%. To protect against changes in the standard deviation [15], we have based our calculations on a standard deviation of 15%. Thus, we need 18 hospitals (9 intervention and 9 control) to identify a decrease of episiotomy rates from 40% to 20%, with a 0.05 significance level and 80% power. Considering that in public hospitals in Argentina and Uruguay about 25% of deliveries are seen among primiparae, and 75% among multiparae, a reduction of the general episiotomy rate from 40% to 20% could be achieved by reducing the rate among primiparae from 80% to 40%, and the rate among multiparae from 26% to 13%. The rates after this intervention would thus be similar to the rates achieved in a previous trial [34] and are therefore potentially attainable. Moreover, the sample size of 18 hospitals will allow identification of smaller decreases of episiotomy rates among primiparae vaginal deliveries, from 80% to 60%, with a 0.05 significance level and 80% power.
We expect the use of oxytocin during the third stage of labor to increase from 10% to 50%. However, assuming a baseline frequency of oxytocin use of 10% and a standard deviation of 5%, a sample size of 18 hospitals will provide a power > 95% to identify an increase in oxytocin use from 10% to 20%, with a 0.05 significance level. Assuming a baseline incidence rate of post-partum hemorrhage of 15% and a standard deviation of 5%, a sample size of 18 hospitals will provide a power of 84% to identify a reduction in hemorrhage from 15% to 8%.
Hospitals will initiate baseline data collection to obtain information that will be used to assess the inclusion criteria before randomization. A total of 24 hospitals were included to allow for hospital drop-outs.
Data analysis
Inference will be primaryly directed at the cluster (hospital) level. All analyses addressing the study research questions will use the "intention to treat" principle, comparing the outcome in the intervention group to the nonintervention group.
The outcome variables will be the percentage of primary and secondary outcomes measured during the three months following the end of the intervention. Mantel Haenszel summary risk ratios combining the individual ratios for each hospital in the intervention group and non-intervention group will be computed. The intervention and non-intervention groups will then be compared at the group level using Student's t-test on the logarithms of the summary risk ratio [17].
In addition, the data collected at the individual level will be used to explore the potential for confounding of the main effects of the intervention due to imbalances arising from the group randomization. Such multi-level analyses will use mixed model techniques (hierarchical linear models) [35,36].
For the analysis of provider's "readiness to change," the data will be summarized in terms of medians and interquartile ranges. Non-parametric statistics will be used to compare the pre- and post-intervention differences between the control and intervention groups.
Ethical aspects
This is a cluster randomized trial with an intervention targeted at the cluster (hospital) level; the behavioral intervention will be directed to the entire group of birth attendants at each intervention hospital. Given that the purpose of the study is to improve the quality of care at the hospitals, this should not imply any additional risk for women and their children. Birth attendants are the subjects of the trial, although the main outcomes will be measured in terms of the patient. All responsible hospital authorities are to be provided written agreements to participate in advance of randomization [37,38]. Birth attendants in the intervention hospitals will receive a fact sheet with information as to the format, length, and purpose of the training intervention. In addition, birth attendants who are nominated as opinion leaders in the intervention hospitals will also provide written agreements to participate in that role.
The necessary data for outcome measurement is routinely collected at the hospitals, and no personal identifiers will be transmitted with in the data to the coordinating unit; thus there will be need no need to obtain informed consent from patients.
All birth attendants will provide written consent before filling out the 'Readiness to Change' questionnaire and all measures will be taken to keep their identities confidential
The protocol was submitted and approved by the IRBs of the following institutions: University of North Carolina at Chapel Hill; Tulane University; Research Triangle Institute; Pan American Health Organization; School of Medicine of the University of the Republic in Uruguay; the University Hospital of Montevideo, Uruguay; and the Argentinean Society for Clinical Research.
The trial is registered at the NIH clinical trials register (; Identifier: .NCT00070720) and at the Current Controlled Trials Register ; ISRCTN82417627).
Abbreviations
IU: international units
WHO: World Health Organization
PIS: Perinatal Information System
IRB: Institutional review board
NIH: National Institutes of Health
NICHD: National Institute of Child Health and Human Development
Competing interests
The author(s) declare that they have no have competing interests.
Authors' contributions
P Buekens, J Belizán and F Althabe had the original idea and designed the intervention and the first protocol. F Althabe, E Bergel, P Buekens, and J Belizán wrote the final protocol, in collaboration with N Kropp, L Wright, N Goco, N Moss, and the rest of the Guidelines Trial Group.
Appendix 1 - Guidelines Trial Group
Principal Investigators: P Buekens, JM Belizán
Trial Coordinating Unit: J Belizán (Principal Investigator), F Althabe (Trial Coordinator), E Bergel (Statistician and Data Coordinator), M Delgado (Programmer), A Ciganda (Data Manager), G Tomasso (Intervention Manager), A Codazzi (Research Assistant), M Colomar (Research Assistant).
Intervention development team: P Buekens, JM Belizán, F Althabe, M Campbell, G Sotero, G Tomasso, ML Cafferata, B Dugan, S Cohen.
Data management system design team: E Bergel, M Delgado, A Ciganda
Statistical support: E Bergel, S Bandiwala, Q Yao, T Hartwell, H Chakraborty
Countries' Coordinating teams:
Argentina: A Blake, A Karolinski (Country Coordinators); AM Bonotti, A del Pino, A Sánchez, M Walker (Data Supervisors).
Uruguay: G Sotero (Country Coordinator), A Ciganda (Data Supervisor).
Formative research team: M Belizán, M Campbell, A Codazzi, M Colomar, A Meier.
Global Network for Women's and Children's Health Research supporting team: LWright (Director), T Hartwell (PI Data Center), N Moss (Program Officer), N Kropp (Protocol Manager), N Goco (Former Protocol Manager).
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study is funded by the National Institute of Child Health & Human Development, National Institutes of Health, USA, within the Global Network for Women's and Children's Health Research (grant number: 1 U01 HD40477-01). Additional support is given by the Bill and Melinda Gates Foundation, the School of Public Health and Tropical Medicine, Tulane University, and the Pan American Health Organization, World Health Organization.
==== Refs
Carroli G Belizán J Episiotomy for vaginal birth (Cochrane Review) The Cochrane Library 2000 Oxford: Update Software
Althabe F Belizan JM Bergel E Episiotomy rates in Latin-American primiparous women. Hospital based descriptive study BMJ 2002 324 945 6 11964339 10.1136/bmj.324.7343.945
Belizán JM Althabe F Barros F Alexander S Rates and implications of Caesarean section in Latin America: ecological study B M J 1999 319 1397 402
Prendiville WJ Elbourne D McDonald S Active versus expectant management in the third stage of labour (Cochrane Review) The Cochrane Library 2000 Oxford: Update Software
Hodnett ED Gates S Hofmeyr GJ Sakala C Continuous support for women during childbirth (Cochrane Review) The Cochrane Library 2004 Oxford: Update Software
Cuervo LG Rodríguez MN Delgado MB Enemas during labour (Cochrane Review) The Cochrane Library 2004 Oxford: Update Software
Basevi V Lavender T Routine perineal shaving on admission in labour (Cochrane Review) The Cochrane Library 2004 Oxford: Update Software
Cabana MD Rand CS Powe NR Wu AW Wilson MH Abboud PA Rubin HR Why don't physicians follow clinical practice guidelines? A framework for improvement JAMA 282 1458 65 1999 Oct 20 10535437 10.1001/jama.282.15.1458
Lomas J Making clinical policy explicit. Legislative policy making and lessons for developing practice guidelines Int J Technol Assess Health Care 1993 9 11 25 Winter 8423109
NHS Effective Health Care Getting Evidence into Practice 1999 5
Oxman AD Thomson MA Davis DA Haynes RB No magic bullets: a systematic review of 102 trials of interventions to help health care professionals deliver services more effectively or efficiently C M A J 1995 153 1423 31
Bero LA Grilli R Grimshaw JM Harvey E Oxman AD Thomson MA Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings. The Cochrane Effective Practice and Organization of Care Review Group B M J 1998 317 465 8
Grimshaw J Eccles M Walker A Thomas R Changing physician's behaviour: what works and thoughts on getting more things to work J Contin Educ Health Prof 2002 22 237 43 12613059 10.1002/chp.1340220408
Grol R Grimshaw J From best evidence to best practice: effective implementation of change in patients' care Lancet 2003 362 1225 30 14568747 10.1016/S0140-6736(03)14546-1
Lomas J Enkin M Anderson G Hannah W Vayda E Singer J Opinion leaders vs. audit and feedback to implement practice guidelines JAMA 1991 265 2202 2207 2013952 10.1001/jama.265.17.2202
Hodnett ED Kaufmann K O'Brien-Pallas L Chipman M Watson-MacDonell J Hunsberger W A strategy to promote research-based nursing care: effects on childbirth outcomes Res Nurs Health 1996 19 13 20 8552798 10.1002/(SICI)1098-240X(199602)19:1<13::AID-NUR2>3.0.CO;2-O
Wyatt JC Paterson-Brown S Johanson R Altman DG Bradburn MJ Fisk NM Randomised trial of educational visits to enhance use of systematic reviews in 25 obstetric units B M J 1998 317 1041 6
Leviton LC Goldenberg RL Baker CS Schwartz RM Freda MC Fish LJ Cliver SP Rouse DJ Chazotte C Merkatz IR Raczynski JM Methods to encourage the use of antenatal corticosteroid therapy for fetal maturation: a randomized controlled trial JAMA 1999 28 46 52 9892450 10.1001/jama.281.1.46
Santoso B Suryawti S Prawaitasari JE Small group intervention vs formal seminar for improving appropriate drug use Soc Sci Med 1996 42 1163 8 8737434 10.1016/0277-9536(95)00390-8
Ross-Degnan D Soumerai S Goel PK Bates J Makhulo J Dondi N Adi D Ferraz-Tabor L Hogan R The impact of face-to-face educational outreach on diarrhoea treatment in pharmacies Health Pol Plann 1996 11 308 18
Treasure T MacRae KD Minimisation is much better than the randomised block design in certain cases BMJ 1999 318 263 4 9915749
Altman DG Schulz KF Moher D Egger M Davidoff F Elbourne D for the CONSORT Group The revised CONSORT statement for reporting randomized trials: explanation and elaboration Ann Intern Med 2001 134 663 694 11304107
Moher D Schulz KF Altman DG for the CONSORT Group The CONSORT statement: revised recommendations for improving the quality of reports of parallel group randomized trials JAMA 2001 285 1987 91 11308435 10.1001/jama.285.15.1987
Donner A Piaggio G Villar J Methodological considerations in the design of the WHO antenatal care randomised controlled trial Paediatr Perinat Epidemiol 1998 12 59 74 9805723 10.1046/j.1365-3016.1998.00007.x
Prochaska J Velicer W Rossi J Goldstein M Marcus B Rakowski W Fiore C Harlow L Redding C Rosenbloom D Rossi S Stages of change and decisional balance for 12 problem behaviors Health Psychol 1994 13 39 46 8168470 10.1037//0278-6133.13.1.39
Rogers EM Diffusion of Innovations 1983 3 New York: The Free Press
Beyer JM Trice HM Implementing Change: Alcoholism Policies in Work Organizations 1978 New York: Free Press
Hiss RG MacDonald R Wayne K Davis Identification of physician Educational Influences in Small community hospital 1978 The University of Michigan Medical School
World Health Organization, Department of Reproductive Health and Research, UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction The WHO Reproductive Health Library N°5 World Health Organization, Geneva 2002
The Cochrane Library Plus en Español 2003 Oxford: Update Software Ltd., Barcelona: Infoglobal Suport SL
Sullivan RL Clinical Training Skills for Reproductive Health Professionals 1998 2 JHPIEGO, Baltimore
Borbas C Morris N McLaughlin B Asinger R Gobel F The Role of Clinical Opinion Leaders in Guideline Implementation and Quality Improvement Chest 2000 118 24S 32S 10939996 10.1378/chest.118.2_suppl.24S
Schwarcz R Díaz AG Fescina R Díaz JL Martell M Simini F The Perinatal Information System I: the simplified perinatal clinical record J Perinat Med 1987 15 9 9
Argentine Episiotomy Trial Collaborative Group Routine versus selective episiotomy: a randomised controlled trial Lancet 1993 342 1517 18 7902901
Littell RC Milliken GA Stroup WW Wolfinger RD SAS System for Mixed Models 1996 SAS Institute, Inc., Cary, NC
Bryk AS Raudenbush SW Hierarchical Linear Models: Applications and Data Analysis Methods 1992 Newbury Park, CA: Sage
Hutton JL Are distinctive ethical principles required for cluster randomized controlled trials? Stat Med 2001 20 473 88 11180314 10.1002/1097-0258(20010215)20:3<473::AID-SIM805>3.0.CO;2-D
Edwards SJ Braunholtz DA Lilford RJ Stevens AJ Ethical issues in the design and conduct of cluster randomised controlled trials BMJ 1999 318 1407 9 10334756
| 15823211 | PMC1090598 | CC BY | 2021-01-04 16:30:35 | no | BMC Womens Health. 2005 Apr 11; 5:4 | utf-8 | BMC Womens Health | 2,005 | 10.1186/1472-6874-5-4 | oa_comm |
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-241580788810.1186/1475-925X-4-24ResearchMathematical model describing erythrocyte sedimentation rate. Implications for blood viscosity changes in traumatic shock and crush syndrome Ismailov Rovshan M [email protected] Nikolai A [email protected] Higmat [email protected] Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15213, USA2 Center for Cancer and Immunology Research, Children's Research Institute, Washington, DC, USA3 Institute for Biomedical Sciences/Program in Molecular and Cellular Oncology, Washington, DC, USA4 Institute of Mechanics and Seismic Stability of Structures, Academy of Science of Uzbekistan, Tashkent, Uzbekistan2005 4 4 2005 4 24 24 24 1 2005 4 4 2005 Copyright © 2005 Ismailov et al; licensee BioMed Central Ltd.2005Ismailov 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 erythrocyte sedimentation rate (ESR) is a simple and inexpensive laboratory test, which is widespread in clinical practice, for assessing the inflammatory or acute response. This work addresses the theoretical and experimental investigation of sedimentation a single and multiple particles in homogeneous and heterogeneous (multiphase) medium, as it relates to their internal structure (aggregation of solid or deformed particles).
Methods
The equation system has been solved numerically. To choose finite analogs of derivatives we used the schemes of directional differences.
Results
(1) Our model takes into account the influence of the vessel wall on group aggregation of particles in tubes as well as the effects of rotation of particles, the constraint coefficient, and viscosity of a mixture as a function of the volume fraction. (2) This model can describe ESR as a function of the velocity of adhesion of erythrocytes; (3) Determination of the ESR is best conducted at certain time intervals, i.e. in a series of periods not exceeding 5 minutes each; (4) Differential diagnosis of various diseases by means of ESR should be performed using the aforementioned timed measurement of ESR; (5) An increase in blood viscosity during trauma results from an increase in rouleaux formation and the time-course method of ESR will be useful in patients with trauma, in particular, with traumatic shock and crush syndrome.
Conclusion
The mathematical model created in this study used the most fundamental differential equations that have ever been derived to estimate ESR. It may further our understanding of its complex mechanism.
==== Body
Introduction
The erythrocyte sedimentation rate (ESR) is a simple and inexpensive laboratory test that is widespread in clinical practice for assessing the inflammatory or acute response [1]. The ESR has also been found to be of clinical significance in the follow-up and prognosis of non-inflammatory conditions, such as prostate cancer [2], coronary artery disease [3], and stroke [4]. In addition, the ESR can be used in the diagnosis of inflammatory conditions [4,5] as well as in the prognosis of non-inflammatory conditions [6]. Some examples of recent applications of the ESR may include sickle cell disease and bacterial otitis media [7,8]. The ESR has been shown to be elevated in 55% of patients with otitis media [7]. Those with elevated ESR have been shown to have a much higher risk for recurrence [7]. In sickle cell anemia, the ESR is usually low in the absence of a painful crisis [8]. A low ESR is an intrinsic property of the sickle red blood cell rheology [9,10]. Certainly, the ESR is only one parameter among others that a clinician can use in the diagnosis and follow up of the above diseases.
Sedimentation of particles, in particular erythrocytes, in a Newtonian fluid (plasma), has been studied by many investigators based on the theory and, therefore, model of interpenetrating motion of two-phase medium that take into account aggregation of erythrocytes [11,12]. We aimed to investigate group precipitation of particles, such as erythrocytes, in a two-phase medium (plasma), both theoretically and experimentally. We added the influence of the vessel wall on group precipitation of particles in tubes, as well as the effects of rotation of particles, the constraint coefficient, and viscosity of a mixture as a function of the volume fraction. The theory has also taken into consideration certain experimental coefficients such as: the coefficient of interaction between the fluid and particles, the aggregation coefficient, the constraint coefficient of phases, the coefficient of viscosity of the mixture, and the coefficient of rotation of a particle. The equation system has been solved numerically. To choose finite analogs of derivatives we used the schemes of directional differences.
Methods
Stokes[13] was the first who derived an equation for non-steady-state flow when he was linearizing the equation of motion of a viscous incompressible fluid. In that work, Stokes developed a theory of resistance for a falling spherical body. The relationship that he derived is called Stokes' formula:
F = 6πμ aV, (1)
Where μ represents viscosity of the fluid, a – the radius of the sphere, V velocity of the fall, and F resistance force.
Albert Einstein investigated the disturbances caused by a particle suspended in a flow with a constant velocity gradient [13,14]. He developed a theory of resistance to shear for a suspension of small spherical particles in a continuous fluid medium. He proved theoretically that an increase in viscosity of a fluid carrying solid particles is connected to the volume fraction of the particles via a proportionality coefficient:
μ = μ0 (1 + 2.5 f2), (2)
where μ0 represents viscosity of the fluid, and f2 concentration of particles.
Up to now, the Einstein formula of viscosity of suspension has been the foundation for most theories describing behavior of a suspension in a shear flow [13,14]. Most studies deal with precipitation of a single particle or multiple diluted particles in a Newtonian viscous fluid and offer various corrective parameters for the Stokes formula [13,14]. For instance, Ozeen [13,14] deduced an approximate solution of equations for a flow of spheres that served as a basis for the formula
where NRe = aVρ/μ is the Reynolds number.
Subsequently, some problems related to non-Newtonian behavior of fluids [13,14] were also investigated. Casuell and Schwarz applied Ericson and Rivlin's model for a slow flow of a non-Newtonian fluid [13,14]. Applying the method of jointed expansions that uses corresponding non-Newtonian terms for both "internal" and "external" expansions to the formula of spherical flow, they derived the following formula:
where φ is a compound expression dependent on non-Newtonian parameters that are constant for any given fluid under isothermal conditions. Lesli also investigated a slow spherical flow, using the Oldroid model [13,14] for non-Newtonian conditions. He concluded that the non-Newtonian term is proportional to V3.
The above studies refer to the precipitation or flow of a single spherical particle. However, concentrated mixtures with a large number of particles interacting during precipitation or in a flow are affected not only by the Stokes force shown in equations (1), (3), and (4), but also by other forces given in equation [14]:
where is a buoyancy force, – frictional force or Stokes force resulting from the viscous force during interaction between phases, determined by the difference between velocities (slippage) u1 - u2, size – a, by the quantity and shape of inclusions, as well as by the physical properties of phases; p-pressure difference, χ(m) – coefficient of constraint, ρ1 and ρ2 are densities of the first and the second phase and are determined by the theory of multiphase medium as reduced densities, K(μ) – coefficient of phase interaction, μ1 and μ2 – viscosity of the first and second phase (where first and second phases correspond to plasma and erythrocytes respectively), f2 – concentration of the second phase, – force related to the effect of "added masses" and caused by the accelerated motion of inclusions (particles) relative to the carrying (viscous fluid) medium, when disturbances arise at a distance on the order of the size of particles; – force of additional effect on particles, due to gradients in the area of average velocities in the liquid phase (the Magnus or Zhukovskiy force). This force is a result of the difference between the densities of phases and the difference between pressures on the opposite sides of a flowing particle.
The influence of the shape of particles, their multiplicity, and some other factors in the expressions for forces , , are taken into account in coefficients K(μ), χ(m), χ(r). As a first approximation, the data on precipitation of a single particle or of a body of a corresponding shape can be applied here, disregarding the direct influence of particles on each other, but taking into account the constraint of the flow of fluid around particles that is caused by the multiplicity of particles [11,14,15].
In order to account for the Magnus effect connected to , it is necessary to take into account the rotation of particles, and in the general case, also the corresponding kinetic parameter ω2, which represents the velocity of rotation that is independent of the field u2 [14]:
where – coefficient accounting for the shape of particles ( = 4π/3 for spherical particles with radius a). In addition to convection, the value of N (the number of particles per unit volume) may change, due to the processes of crushing, adhesion, aggregation of particles, and formation of new ones that are defined by value ψ in the equation [14]:
In case of the absence of crushing, adhesion, and aggregation of particles or formation of new ones, i.e. when ψ = 0, as well as under the condition of non-compressibility of the material (of particles), that is ρ2 = constant α then is a constant. Moreover, the equation of conservation of mass of the second phase[14] leads to . And N, number of particles per unit volume, can be expressed as [14]:
or [14]
∂N/∂t + ∇Nu2 = 0. (9)
It should be noted that in the case of a moderate volume fraction of particles in a fluid, the mixture of two incompressible phases cannot be considered in terms of Navier – Stokes' equations, even in the absence of relative rotary and radial motion [14]. This is because viscosity of suspensions and emulsions is determined using viscosity-meters based on the measurement of a mass consumption of mixture Q through a tube, with diameter d and length L, dependent on pressure ∇P [16-18]. If a Poiseuile flow of a Newtonian fluid is the case, then a linear connection between Q and ∇P holds true. It follows, then, that with low concentrations, when f2 < 0.05, the resulting values of μef agree with the Einstein formula (2), but when f2 > 0.05 then values μef exceed values obtained from the equation [14]:
Additionally, there is considerable variation among different authors and among different combinations of phases. This variation apparently reflects the non-Newtonian nature of concentrated viscous disperse mixtures and insufficiency of values ρ and μ for determining their mechanical properties. In this regard, laboratory experiments have to be carried out for each mixture and real devices over a range of operating parameters to determine the loss of pressure when applying different rheological models, in particular, the model of a viscous fluid with an effective coefficient of viscosity.
One should keep in mind that when f2 ≥ 0.1, in addition to the shape and size of solid particles, their material may also have an effect on effective viscosity and other rheological characteristics of the mixture. Apparently, this is due to irregular distribution of particles, collision of particles with each other, and solid walls [16-18].
The above considerations lead to an examination of group precipitation of particles, when f2 ≥ 0.1 in the interpenetrating model of two- or multiphase medium. The first report of such an approach was presented in [11], where researchers used the theory of interpenetrating motion of two-phase mediums, taking into account the aggregation of particles, in this case erythrocytes, in a fluid layer limited by a free border x = L from the top and a solid border x = 0. In those studies, precipitation of particles in a fluid was considered based on the following assumptions: the motion of particles in a fluid is taking place in the strictly vertical direction under the influence of gravity; the inertial force is negligibly small; viscosity appears only during interaction between particles and the fluid surrounding them; spontaneous disintegration is absent; the column of sedimentation of particles is divided into two main zones: clean fluid and settling particles. The influence of the wall and interaction between particles were ignored. Studies [19,20] showed that the radius of the tube has an effect on sedimentation of particles, in particular, erythrocytes. From this it follows that the group precipitation of particles, such as erythrocytes, depends appreciably on border conditions, in this case on conditions of adhesion and impact of particles on the vessel wall [19,20].
In addition, in most studies the rotation of particles caused by force was not taken into account, and neither were the force of interaction between adjacent particles or the force arising from accelerated movement of particles. Thus, in the present work, we examine the group precipitation of particles in a liquid mixture with forces and , as well determination of the constraint coefficient and of viscosity of the mixture μm as functions of the volume fraction of particles. We also describe the effect of the vessel wall on the process of group precipitation of particles in a multiphase medium.
Determination of the coefficient of constraint and viscosity of a mixture as a function of the volume fraction of particles
Problems of constrained precipitation of solid and deformed particles in a fluid are the principal part of the theory of two-phase flows and are of great importance for such hydrodynamic processes as: mass of exchange, separation by gravity, oil refining, and precipitation of erythrocytes in blood [11,12,16-26].
Studies in this field usually deal with high- and low-concentration mixtures. The laws of constrained precipitation of moderately concentrated mixtures have not been well investigated. The foundation of precipitation of a single particle in an infinite fluid is the Stokes law. According to that law, the frictional force arising from the motion of spherical particles with diameter d and velocity V in a medium of viscosity μ is expressed by formula
The gravitational force acting on a particle depends on its specific gravity, that is
where ρ1;ρ2;g – the density of a fluid, density of the particle, and acceleration of gravity respectively.
– buoyancy force, and, in (11)
– frictional force.
Due to force the particle accelerates its motion. Aside from gravitational force, the particle is affected by frictional force directed oppositely; its value increases with the increasing velocity, according to the Stokes law. This means force is canceled out by gravitational force . Therefore, the movement proceeds with constant velocity Vc that is determined by the equations (11) and (12):
Sometimes one has to deal with precipitation of a large number of particles in concentrated mixtures. Formulae for the velocity of precipitation of particles as a function of the concentration and velocity of a single particle in an infinite fluid can be derived using some statements of the interpenetrating model [14] and the Euler equation [14]. From these formulas, an equation for low concentrations of particles can be easily deduced; the equation is consistent with the experimental data [27]. Let us assume that in volume V there are two phases with different values of specific gravity. Then particles with greater specific gravity start moving down canals that are being formed; this holds true for the displacing phase, i.e. the process of mutual penetration is taking place.
Let us determine the hydraulic radius of a canal as
If the canal is cylindrical then
From (14) and (15) the effective diameter of the canal is equal to:
where S – cross-section area of the canal, P – perimeter of the canal. Multiplying both numerator and denominator by diameter d (in this case d corresponds to the diameter of a single erythrocyte) and by volume V and also taking into account [21] that external concentration equals volume concentration, it follows:
where Su – surface of the particles; f1 – volume fraction of the first phase. But:
(where N – number of particles per unit of volume equal to ), therefore (16), (17):
(where is a coefficient of shape for spherical particles).
For particles of irregular shape:
( – form coefficient).
We can divide the continuity equation:
V1S = V1iS1
by S, and multiply both numerator and denominator by diameter d (in this case d corresponds to the diameter of a single erythrocyte) and by volume V and also taking into account [21] that external concentration equals volume concentration. Then it follows:
V1 = f1V1i (19)
The flow of fluid between particles by analogy with the flow of fluid in tubes can be expressed as criterion equations [23]:
or [23]:
In the processes of precipitation when concentration of inclusions is significant and sizes of particles are small, and in the processes of industrial filtration as well, a laminar flow is very often observed, for which m = -1, n = 1. Then [23]:
If one takes into account (18) and (19), the last equation can be expressed as Coseni Carman for constrained precipitation in a laminar flow:
Let's divide equation (23) by the number of particles per unit of volume and derive the resistance force of the fluid acting on one particle as:
(χ – resistance coefficient for precipitation of a set of particles).
It is known that the resistance force experienced by a particle during precipitation in a fluid is:
Since for particles suspended in a fluid:
F* = Fc
Then from (24) and (25) it follows:
where β – the ratio of velocity to hydraulic size;
χc – resistance coefficient for precipitation of a single particle (25) in an infinite fluid. From (24) when f1 → 1 it follows:
when values of Reynolds numbers are small:
Therefore, it can be supposed that the basic mechanism is:
From the equations (26) and (27), we can derive:
where
According to the Stokes law, in a laminar flow we have:
figuring the last equation into equation (30), we can derive
In accordance with our experimental data when Reynolds numbers are small, constant C = 5.
Equation (29) when f1 → 1 can be expressed as
β = 1 - cf2. (30)
The difference between the coefficients of equation (30) and the coefficients calculated for small concentrations [22] is the factor of 0.34. If we do not ignore the fraction of velocity resulting from the proximity of two spheres, which is true for small concentrations of inclusions, then equation (30) coincides with the Bachelor formula [22]. The change in the velocity of particles depends on the concentration calculated from (29); a wide range of concentrations is given in Figure 2.
Figure 2 The relative velocity of sedimentation as a function of the concentration of particles. Horizontal axis: volume fraction of phase two; vertical axis: β, a change in the relative velocity.
From the equation of equilibrium of forces, taking into account the coefficient of constraint, it follows (13) [14]:
Based on (31) and (32) we have:
or taking into account (29)
(φf – coefficient describing constrained precipitation).
Assuming precipitation of a particle in suspension with viscosity μm, density ρm, we can express the equilibrium equation [21] as
ρm = f1ρ 1i + f2 ρ2i. (36)
Allowing for (32) and V1 = 0, it follows
and from (28) we can obtain
When f1 → 1, but C = 2.5, we have the Einstein formula (10):
From the calculation given in Figure 1, it follows that formula (38) agrees with experimental data (up to f2 = 0.5, when C = 2.5) obtained in a previous study [27] on the change of viscosity of suspension for a wide range of fluids, sizes, and materials of solid discrete particles.
Figure 1 The dependence of a change in relative viscosity on the concentration of particles. Horizontal axis: volume fraction of solids; vertical axis: viscosity (mNsm-2)
The above dependence of relative velocity and viscosity of the mixture on volume fraction agrees with the experimental data [27]. Let us examine formulae (33) and (34), and consider the possibility of determining an analytic calculation formula for the constraint coefficient and effective viscosity, and their dependence on the concentration of particles. In (33) and (34) the formula for the relative velocity of the precipitation of particles has been derived as:
where V2 – velocity of constrained sedimentation of disperse particles;
Vc – velocity of sedimentation of a single particle (f2 ≈ 0) in a motionless fluid (V1 = 0);
f2 – volume fraction of disperse phase f1 + f2 = 1;
c – dimensionless numerical coefficient.
As is known[14], from the equilibrium equation of a two-phase medium, allowing for constraint motion of particles
We can derive an equilibrium equation of gravitational force, buoyancy force, and Stokes force acting on the disperse phase as:
The parameters correspond to the parameters in the study [14]. The second equation (42) describes a rest condition of the mixture; that is, the motion of dispersed particles in one direction accompanies the motion of fluid in the opposite direction.
From the equation (42) for precipitation of a single particle (f0 ≈ 0) in (φj = 0) in the at-rest mixture, we can derive the Stokes formula, however the velocity of constrained precipitation is:
From the equations (40) and (43), the constraint coefficient can be determined as:
From the analysis of the experimental data for small volume fractions (f2 < 0.005), it follows that there are three types of functions [14] for the determination of a constraint coefficient corresponding to different types fluids in which spherical particles are precipitating:
Comparing equations (44) – (47), we can conclude that formula (44) deduced from [28] is similar to formula (46).
When and c ≈ 5 ÷ 6; that is, under conditions (44), a change of constraint coefficient is described, depending on the volume fraction of randomly dispersed particles. As described previously [14], when the concentration of the disperse phase in a mixture increases, numerical values of volume and shear viscosity start to differ from each other, and the motion of the mixture cannot be described by Navier – Stokes equations [14].
When concentrations (f2 < 0.005) and values of viscosity agree with the Einstein formula, and when concentrations determined in the laboratory are high, then the viscosities considerably exceed the values derived from the formula for moderate and concentrated viscous disperse mixtures of non-Newtonian nature. In addition, the rheological parameters are insufficient for the determination of the mechanical properties of those moving mixtures.
In particular, one of the models of viscous fluid is a model taking into account the effective coefficient of viscosity, as an additional rheology parameter. Moreover, it is assumed that the effective coefficient of viscosity (viscosity of a mixture) reflects not only a shape and size of solid particles, but also their material and irregularity of their center of gravity.
Following the approach described in[14] for precipitation of particles in a fluid with viscosity μm, we can derive the following equality
Using equation (48) and allowing for equation (43) [28], it follows that
μm = φf μ1 (49)
We can also substitute equation (44) into equation (49), and from the equations (45)–(47) and (49) we can derive [28]
We can summarize here: equation (45) has been derived from the consideration of the cell model for uniform distribution of particles; equation (46) deals with the random allocation of particles; and equation (47) is the case of significant formation of clumps and alignment of particles in chains [14]. Therefore, equations (50) allow one to determine viscosity of a mixture for all three types of distribution of particles. Thus, the formulae for calculation of the constraint coefficient and viscosity of a mixture as functions of the volume fraction of the disperse phase have been fully described.
Forces in a two-phase medium that affect group precipitation, in particular, precipitation of erythrocytes in blood
As follows from the above, there can be four basic interacting forces when particles are precipitating in a mixture:
1. The buoyancy force occurring due to the difference between densities of phases;
2. The frictional force or Stokes force; it results from interaction between the two phases;
3. The force of "added masses" arising from either accelerated or constrained motion of particles;
4. The force of additional effect on particles which is due to gradients in the area of average velocities in the liquid phase (the Magnus or Zhukovski force).
These forces should be taken into consideration when creating a mathematical model of group precipitation in a multiphase medium.
The first force, buoyancy force, is defined by formula:
where f2 – volume fraction of the second phase (particles).
ΔP – gradient of pressure, or the difference of densities of phases.
The second force, frictional force, or Stokes force, is defined by the difference of velocities (slippage) |V1 - V2|, the size, quantity and shape of inclusions, as well as physical properties of the phases. From [11] we can write down the following equation for the frictional force (5), (6):
Instead of viscosity of plasma μ1φ(f2) we can use the viscosity of the mixture determined in formula (50)
Expression (53) has been derived while applying the cell model for uniform distribution of particles, expression (54) deals with irregular distribution of particles, and expression (55) is the case of significant formation of aggregations and alignment of particles in chains. Therefore, from the equations (54) and (55), one can determine the viscosity of the mixture for all three types of distribution of particles in a fluid. Instead of function ψ(f2) from [11] we can use .
The third force: the force is related to the action of "added masses," i.e. this force arises from either constrained or accelerated motion of particles:
where constraint coefficient χ(m) from formulae (44), (45) and (46) can be defined as [14]:
Horizontal movement of particles under the influence of gravitational force can be explained, if one considers the flow of precipitating particles. In a homogeneous mixture with a zero average flow rate [24],
f1V1 + f2V2 = 0 (60)
Which means that in each point, the volume flow of the precipitating solid phase accompanies upward motion of the fluid. This phenomenon has been considered in the previous paragraph. The following is the formulae for forces of the additional effect on particles, caused by gradients in the area of average velocities in the liquid phase:
In addition, coagulation or aggregation of multiple particles can take place, which is described by these equations (7), (8), (9) [14]:
Where N – number of units, ψ – total velocity of aggregation [14].
ψ = KN2, (63)
where k – aggregation coefficient.
Based on the above, the mathematical model for precipitation of particles in a two-phase medium can be formulated.
The mathematical model for group precipitation in a two-phase medium (mixture) in a flat tube
Let us consider precipitation of particles in a fluid on the basis of the hydrodynamic theory. Experiments on precipitation of erythrocytes in plasma carried out in cylindrical tubes of various diameters have shown that the radius of the tube significantly affects the process of precipitation [29].
It follows then that group precipitation of particles such as erythrocytes, depends on border conditions, in this case the conditions of adhesion. The small radius of a tube defines the area of velocities of flow of the liquid phase during precipitation of a particle. Therefore, particles experience a differential of pressure from the surrounding fluid. That is why, in this study, the fluid (carrying phase) is considered a viscous phase.
Let us consider precipitation of a particle taking place between vertical parallel walls. We will place X axis between the walls pointing upward, Y axis also parallel and equidistant from the walls but horizontal, and Z axis is horizontal and perpendicular to the walls. Let us assume that the horizontal length of the tube is 2 h, and its height is L. We will also assume that the motion is not taking place along Z axis. This leaves us with two-dimensional movement of particles. Group precipitation of the particles in the mixture can be described mathematically as the motion of two-phase interpenetrating mediums according to the theory of Rakhmatullin [21].
It follows that equations for the two-phase flow in a Cartesian space coordinate system can be expressed as [30]:
for the second phase [30]:
Continuity equations for each phase are expressed as [30]:
and the equation of numerical concentration is then [14]
Assuming that the components of velocity V and W are negligibly small compared to component U, and that the horizontal length of the canal is negligibly small compared to the height, that is, V ≪ U, W ≪ U and L ≫ 2h, we can then estimate some terms and simplify the system of equations ; instead of U1, U2 we will use u1 and u2
where μm – effective coefficient of viscosity of the mixture that is derived from the formula:
or for specific cases
Where μ1 – viscosity of plasma.
We can write down the continuity equation as
The equation of numerical concentration then becomes:
where N – number of aggregates, ψ – total velocity of aggregate formation.
Three forces determine the force of interaction between phases.
The frictional force can be derived from formula (52)
which takes into account the number of aggregates N. Equation (78) also allows an investigator to take into account the changes in the number of particles per unit of volume, the average distance between them, the radius of particles, volume of the disperse phase, etc. These geometrical parameters have a significant effect on the kinematics of the aggregation processes.
From [11] we can derive the expression for the total velocity of aggregate formation
φ = -KN2. (79)
From expression (78) when , and based on the formula from [31] we can obtain
here – coefficient of shape for the particles in question.
Instead of the force of "added masses" we can use formula (56), namely
where χ(m) – coefficient of constraint, which is derived from one of the following formulae
The Magnus or Zhukovski force is defined by the expression
Therefore, the closed equation system to solve the problem has been obtained:
Using equations (74) and (75) we can obtain
Since the precipitating particles cannot flow through the bottom of the tube, the following holds true:
f1u1 + f2u2 = 0. (88)
Equation (88) gives us the rest condition of the mixture. Let us use the force of interaction between phases given in [11] to derive the closed equation system. First, it is necessary to formulate initial and border conditions.
At t = 0 u1 = 0, u2 = 0, (89)
(N0 is the number of particles in 1 mm3, a – the radius of a particle)
This problem can be solved numerically. We will use the following initial data:
Analysis of the experimental and numerical results and their comparison
Figure 3 shows the vertical distribution of concentrations of the disperse phase at various time intervals. More precisely, it shows the concentration at various points along X axis at various intervals of time. In figures 3, different curves correspond to the distribution of concentration in the tube after t = 5 min, t = 25 min, t = 60 min from the beginning of the process. It was noted that the change of concentration of the disperse phase along the height of the tube at various intervals of time, corresponds to the decrease of volume fraction of the disperse phase in the upper part and its increase in the lower part of the tube. Moreover, aggregation of particles reinforces the precipitation of erythrocytes.
Figure 3 Curves of the change of concentration of the disperse phase along the height of the tube at various moments of time. Curves 1 and 2 when t = 60 min, 3 and 4 when t = 25 min, 5 and 6 when t = 5 min. Curves 1, 3 and 5 correspond to precipitation without border conditions, and 2, 4 and 6 with border conditions. Aggregation coefficient K = 10-5 for all curves. Horizontal axis: concentration of the second phase (f2); vertical axis: distance (millimeters)
As one can see, when aggregation coefficient κ is 10-5 the effect is observed, if we compare the curves in figure 3. It should be noted that the wall layer has a considerable effect on precipitation of erythrocytes. Since particles bounce off the walls, the aggregation of particles in the central part of the tube is enhanced. As a result, a big aggregate of particles precipitates faster than a single particle. This effect was described in the theoretical studies, as well as shown experimentally. In order to determine the influence of the vessel wall, we placed the blood of a patient in capillaries of various diameters. During precipitation, erythrocytes descended 18 mm after one hour.
When they descended 11 mm, but when by 6 mm.
That means that the influence of the vessel wall on precipitation of erythrocytes is considerable in capillary tubes of lesser diameter. Also, according to figure 3, the curves of precipitation are different at various time intervals.
Figure 4 shows the curve of precipitation of particles as a function of time, for various values of aggregation coefficient K, both in the theoretical (curves 1–3, 5, 6) and experimental studies (curve 4). According to Figure 4, theoretical curves are consistent with the experimental data.
Figure 4 Curves of the change in the borders of separation between fluid and mixture as a function of time, when values of total velocity of aggregation are different. k : 1 - k = 5·10-7; 2 - k = 1·10-6; 3 - k = 5·10-7; 4-experimental curve corresponds to rheumatism; 5 - k = 1·10-5; 6 - k = 5·10-5. Dotted curves match the results of equations (86), with the border conditions (89) and (90). Horizontal axis: time (minutes); vertical axis: the height of the tube, millimeters.
Discussion
This work addresses the theoretical and experimental investigation of single and group precipitation of particles in homogeneous and heterogeneous (multiphase) medium, as it relates to their internal structure (coagulation or aggregation of solid and/or deformed particles). Such processes cannot be described by means of a single velocity model of a non-Newtonian fluid. Precipitation of particles in a viscous homogeneous relaxing medium and the motion of flow of the fluid are directed oppositely, due to the gravitational force. These processes can be described by means of a two-velocity interpenetrating, two-phase (or multiphase) model.
In general, ESR has a complex mechanism where, on the one hand, an increase in the concentration of erythrocytes leads to a decrease in β, the change in the relative velocity of sedimentation (figure 2). On the other hand, such an increase may result in an increase in the quantity of rouleaux. Our mathematical model takes into account the influence of the vessel wall on group precipitation of particles in tubes, and also viscosity of the mixture, constraint, and rotation of particles. This model can also describe ESR, using the coefficient of the velocity of erythrocyte precipitation.
As mentioned earlier, ESR measurement remains the method of choice in evaluating different clinical conditions [5]. In our view, the ESR could also be useful in patients with the traumatic shock and traumatic crush syndrome. In general, conditions such as hemorrhage, trauma, and burns may result in blood viscosity changes [32]. Erythrocyte aggregation caused by trauma, burns, and hemorrhage was described in detail by Knisely [33].
In the 21st century, at the time of high-speed transportation and motor-vehicle accidents, critical states, such as traumatic shock, have become more widespread. Traumatic shock results in generalized vasoconstriction, and therefore, this condition may result in microcirculatory disorders that have been associated with increased blood viscosity [34,35]. Traumatic shock, due to increased blood viscosity, may further decrease the cardiac output [34].
Theoretically, any trauma may result in extensive intravascular erythrocyte aggregation, irrespective of the presence or absence of traumatic shock [36]. However, generalized intravascular blood slowing and stasis, as well as erythrocyte aggregation, is much more severe in the presence of traumatic shock [36]. The analysis of the certain rheological changes caused by the experimental traumatic shock was conducted by Tatarishvili et al [35]. In total, 40 adult male laboratory rats were used, of which 20 were used for experiments and 20 served as controls [35]. Kennon's method was used to produce traumatic shock. Erythrocyte aggregability was measured using the "Georgian technique"[37]. Traumatic shock resulted in significant changes in all investigated blood rheological parameters (erythrocyte aggregability, deformation, and systemic hematocrit). The experimental group of rats showed an increase in erythrocyte aggregability, which was more than 2 times higher compared to control animals. However, the hematocrit and erythrocyte deformability decreased in experimental rats compared to control animals [37]. Despite the small sample size, these results were statistically significant [37]. Tatarishvili et al. concluded that "the blood rheological disorders in the microcirculation related in the present experiments should be considered as a significant factor determining the severity of the damage to the tissues during the development of the traumatic shock" [37].
As mentioned earlier, the ESR measurement might also become a useful diagnostic test in patients with crush syndrome. Disasters, such as earthquakes, are a key source of victims with crush syndrome. Earthquakes in Armenia in 1988 [38] and in Japan in 1995 have caused multiple cases of traumatic crush syndrome. [39] Victims of such earthquakes have been trapped for prolonged periods. The other possibility for the crush injury and crush syndrome is, for example, the severe and transient pressure that occurs when a limb is run over by a heavy vehicle. Finally, crush syndrome may also affect patients with stroke or drug overdose [40]. Such patients could crush a part of the body with their own weight while unconscious. The lower limbs and, less frequently, upper limbs are predominant sites of injury in victims with traumatic crush syndrome. Traumatic crush syndrome caused by a crush injury to the torso has also been described [39]. Rheological changes caused by traumatic crush syndrome have been studied in rats that were exposed to 6 hours of compression of soft tissues on the thigh. Narcotized rats with crush syndrome were shown to have increased blood viscosity. In these animals, total peripheral vascular resistance was shown to correlate well with changes in blood viscosity at different shear rates. [41]. Rats' blood viscosity was predominantly determined by erythrocyte aggregation at low shear. [41] These rats were also divided into two groups: low resistant and high resistant to shock. Low resistant animals exhibited high hematocrit and plasma viscosity and decreased deformability of erythrocytes. On the contrary, blood viscosity increased, independent of shear rates (10 – 300 sec -1). Severe hemoconcentration and blood hyperviscosity developed in all low resistant animals. High resistant rats developed sufficient hemodynamics and oxygen supply to tissues. On the other hand, high resistant rats exhibited less significant hemoconcentration and an increase in blood viscosity [42].
According to our current research, increased ESR in patients with traumatic shock and crush syndrome is due to an increased number of rouleaux. Our understanding of changes in the ESR in traumatic shock or in crush syndrome could be summarized as follows: crush syndrome or traumatic shock (due to hemorrhage) may result in an increase in peripheral resistance[34,43-45]. Increased peripheral resistance may result in a decrease in cardiac output, which subsequently leads to a decrease in blood flow velocity and, thus, a decrease in yield velocity and shear stress[46] A decrease in blood flow velocity, due to trauma, may result in a decrease of erythrocyte adhesiveness within rouleaux [46], since even a relatively high yield velocity (300 m-1) did not result in a significant decrease in blood viscosity [41,42]. The damage of particles at different erythrocyte adhesiveness within rouleaux has been studied in the previous research [47]. From those experiments it follows that the damage is proportional to yield velocity [46]. On the other hand, an increased ESR in patients with traumatic shock and crush syndrome is due to an increased number of rouleaux, however, as it has been shown earlier, the quantity of rouleaux is largely dependent on the quantity of erythrocytes [47]. Going back to the beginning of the discussion, those rats that had lower survival (second group) had, according to our hypothesis, a larger quantity of rouleaux formation, due to an initially larger quantity of erythrocytes [41,42].
Conclusion
The mathematical model created in our study accounts for the influence of the vessel wall on group precipitation of particles in tubes, viscosity of the mixture, constraint and rotation of particles. The determined viscosity of a mixture and constraint of particles depend on volume fraction. In theoretical studies, the difference of curves of precipitation, during the initial and later intervals of time, has been established.
The theoretical curves of precipitation of particles at various values of aggregation coefficient K show that an increase of K increases the velocity of precipitation. Likewise, the presence of the tube wall enhances precipitation of particles. The Magnus force and the force connected to acceleration of particles in a relatively liquid phase do not have a significant effect on the precipitation of particles. From our experimental and theoretical data, we can conclude that the behavior of the curve of velocity of erythrocytes depends on the timing of observations (interval of time is 5 min). Time-course measurements of erythrocyte precipitation more accurately reflect the state of the human body than the known determination of velocity of erythrocytes.
Determination of the ESR should be conducted at certain intervals of time, i.e. in a series of periods not exceeding 5 minutes each. Differential diagnosis of various pathological conditions when ESR is applied can be conducted using the aforementioned timed method of determining ESR. Different pathological states, when this method is used, would have different initial blood viscosity parameters. We have shown that an increase in blood viscosity during trauma results from an increase in rouleaux formation. We have shown that the time-course method of ESR will be relevant in patients with trauma, in particular, with traumatic shock and crush syndrome. Therefore the number of rouleaux is dependent on the initial quantity of erythrocytes and is increasing sharply with the decreasing yield velocity [47] which may result from increased peripheral resistance and low cardiac output from traumatic shock or crush syndrome.
==== Refs
Saadeh C The erythrocyte sedimentation rate: old and new clinical applications South Med J 1998 91 220 225 9521358
Johansson JE Sigurdsson T Holmberg L Bergstrom R Erythrocyte sedimentation rate as a tumor marker in human prostatic cancer. An analysis of prognostic factors in 300 population-based consecutive cases Cancer 1992 70 1556 1563 1516006
Gillum RF Mussolino ME Makuc DM Erythrocyte sedimentation rate and coronary heart disease: the NHANES I Epidemiologic Follow-up Study J Clin Epidemiol 1995 48 353 361 7897457 10.1016/0895-4356(94)00156-K
Chamorro A Vila N Ascaso C Saiz A Montalvo J Alonso P Tolosa E Early prediction of stroke severity. Role of the erythrocyte sedimentation rate Stroke 1995 26 573 576 7709399
Sox HC JrLiang MH The erythrocyte sedimentation rate. Guidelines for rational use Ann Intern Med 1986 104 515 523 3954279
van den Hoogen HM Koes BW van Eijk JT Bouter LM On the accuracy of history, physical examination, and erythrocyte sedimentation rate in diagnosing low back pain in general practice. A criteria-based review of the literature Spine 1995 20 318 327 7732468
Del Beccaro MA Mendelman PM Inglis AF Richardson MA Duncan NO Shugerman RP Acute-phase reactants and acute bacterial otitis media Am J Dis Child 1992 146 1037 1039 1514547
Ballas SK The sickle cell painful crisis in adults: phases and objective signs Hemoglobin 1995 19 323 333 8718691
Akinola NO Stevens SM Franklin IM Nash GB Stuart J Rheological changes in the prodromal and established phases of sickle cell vaso-occlusive crisis Br J Haematol 1992 81 598 602 1390248
Lawrence C Fabry ME Erythrocyte sedimentation rate during steady state and painful crisis in sickle cell anemia Am J Med 1986 81 801 8 3776987 10.1016/0002-9343(86)90349-9
Losev ES A physical model of gravitational erythrocyte sedimentation Biofizika 1992 37 1057 1062 1298346
Normatov TD Khusanov IN Determination of constraint coefficients and viscosity of mixture in dependence on volume fraction of particles Nauka 2001 5 35 38
Happel J Brenner H Low Reynolds Number Hydrodynamics with Special Applications to Particulate Media 1986 Dordrecht: Martinus Nijholl Publishers
Nigmatullin RI Basic mechanics of multiphase medium 1978 Moscow: Nauka
Faizulaev DF Shakirov AA Abidov S Khusanov IN Investigation of precipitation of particles by photoelectrical method. Registration of concentrations Problems of mechanics 1974 130 133
Pedley TJ The fluid mechanics of large blood vessels 1980 Cambridge: Cambridge University Press
Regurer SA Lectures for biological mechanics 1980 Moscow: MGU
Pavlovski UN Regurer SA Skobeleva JM Hydrodynamics of blood 1970 Moscow: VINITI
Minz DM About weight of grain layer in rising flow of fluid Reports RAN 1952 1 17 20
Navruzov K Akhmedzhanova GD Yakubov BS Khakberdiev DB The motion of non Newtonian fluid in tubes with allowing for a wall layer Academy of Science of Uzbekistan 1989 2 53 56
Rakhmatullin KA Foundations of gas dynamics of mutually penetrable flows of compressible medium. (Russian) Prikladnaya Matematika Mekhanika 1956 20 184 195
Batchelor GK An introduction to fluid dynamics 1967 Cambridge: Cambridge University Press
Malinovskaya TA Separation of suspension in industry of limited synthesis 1971 Moscow: Nauka
Navruzov KN Khakberdiev ZB Dynamics of Non-Newtonian fluids 2000 Tashkent: Nauka
Podrabinek PA Kamenskii II On the Gulden size of erythrocytes Laboratornoye delo 1967 8 482 484
Gavalov SM Mechanism of fractional erythrocyte sedimentation rate Sov Med 1957 21 62 66 13519206
Thomas DG Transport Characteristics of Suspension: VIII. A Note on the Viscosity of Newtonian Suspensions of Uniform Spherical Particles J Coll Sci 1965 20 267 77 10.1016/0095-8522(65)90016-4
Khusanov H Sedimentation of particles in fluid Nauka 1986 120 130
Caro CG The mechanics of the circulation 1978 Oxford: Oxford University Press
Faizullaev FD Laminar motion of multiphase media in conduits 1969 New York: Consultants Bureau
Nigmatullin RI Mechanics of heterogeneous medium 1978 Moscow: Nauka
Bergentz SE Gelin LE Rudenstam CM Zederfeldt B The viscosity of whole blood in trauma Acta Chir Scand 1963 126 289 293 14070600
Knisely MH Elliot TW Block EH Sludged blood in traumatic shock Arch Surg 1945 51 220
Seligman AM Frank HA Fine J Traumatic shock. XII. Hemodynamic effects of alterations of blood viscosity in normal dogs and in dogs in shock J Clin Invest 1946 25 1 21 16695288
Tatarishvili J Momtselidze N McHedlishvili G Blood rheological abnormalities in the microcirculation during experimental traumatic shock Clin Hemorheol Microcirc 2004 30 407 410 15258374
Heimbeckersh RO Bigelow WG Intravascular agglutination of erythrocytes (sludged blood) and traumatic shock Surgery 1950 28 461 473 14776654
Mchedlishvili G Gobejishvili L Mamaladze A Momtselidze N Varazashvili M Microcirculatory stasis induced by hemorheological disorders: further evidence Microcirculation 1999 6 97 106 10466112 10.1038/sj.mn.7300028
Collins AJ Kidney dialysis treatment for victims of the Armenian earthquake N Engl J Med 1989 320 1291 1292 2710215
Oda J Tanaka H Yoshioka T Iwai A Yamamura H Ishikawa K Matsuoka T Kuwagata Y Hiraide A Shimazu T Sugimoto H Analysis of 372 patients with Crush syndrome caused by the Hanshin-Awaji earthquake J Trauma 1997 42 470 475 9095115
Shaw AD Sjolin SU McQueen MM Crush syndrome following unconsciousness: need for urgent orthopaedic referral BMJ 1994 309 857 859 7695702
Chernysheva GA Plotnikov MB Smol'yakova VI Relationship between rheological and hemodynamic changes in rats with crush syndrome Bull Exp Biol Med 2000 130 1048 1050 11182812 10.1023/A:1002827230305
Chernysheva GA Plotnikov MB Smol'yakova VI Aliev OI Ulanova EV Sutormina TG Course of shock in rats with different resistance to shockogenic trauma during Crush syndrome Bull Exp Biol Med 2001 132 946 949 11782788 10.1023/A:1013654925156
Haugan A Kirkebo A Local blood flow changes in the renal cortex during tourniquet and burn shock in rats Circ Shock 1984 14 147 157 6509728
Cryer HG Mavroudis C Yu J Roberts AM Cue JI Richardson JD Polk HC Jr Shock, transfusion, and pneumonectomy. Death is due to right heart failure and increased pulmonary vascular resistance Ann Surg 1990 212 197 201 2375651
Carleton SC Cardiac problems associated with burns Cardiol Clin 1995 13 257 262 7614515
Ismailov RM Shevchuk NA Schwerha J Keller L Khusanov H Blunt trauma to large vessels: a mathematical study Biomed Eng Online 2004 3 14 15153246 10.1186/1475-925X-3-14
Ismailov RM Mathematical model of blunt injury to the vascular wall via formation of rouleaux and changes in local hemodynamic and rheological factors. Implications for the mechanism of traumatic myocardial infarction Theor Biol Med Model 2005 2 13 15799779 10.1186/1742-4682-2-13
| 15807888 | PMC1090599 | CC BY | 2021-01-04 16:37:37 | no | Biomed Eng Online. 2005 Apr 4; 4:24 | utf-8 | Biomed Eng Online | 2,005 | 10.1186/1475-925X-4-24 | oa_comm |
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-101585750910.1186/1475-2867-5-10ReviewRel/Nuclear factor-kappa B apoptosis pathways in human cervical cancer cells Shehata Marlene F [email protected] Division of Basic Sciences, Faculty of Medicine, Memorial University of Newfoundland, St. John's, A1B 3V6, Canada2005 27 4 2005 5 10 10 9 1 2005 27 4 2005 Copyright © 2005 Shehata; licensee BioMed Central Ltd.2005Shehata; 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.
Cervical cancer is considered a common yet preventable cause of death in women. It has been estimated that about 420 women out of the 1400 women diagnosed with cervical cancer will die during 5 years from diagnosis. This review addresses the pathogenesis of cervical cancer in humans with a special emphasis on the human papilloma virus as a predominant cause of cervical cancer in humans. The current understanding of apoptosis and regulators of apoptosis as well as their implication in carcinogenesis will follow. A special focus will be given to the role of Rel/NF-κB family of genes in the growth and chemotherapeutic treatment of the malignant HeLa cervical cells emphasizing on Xrel3, a cRel homologue.
Cervical cancerHeLa cellsNF-κBXrel3cisplatinapoptosisanti-apoptosis
==== Body
Introduction
A. Oncogenesis
The process of oncogenesis or carcinogenesis fundamentally emerges from defects in the balance between the activity of proto-oncogenes, which promote cell proliferation, and tumor suppressor genes, which regulate the cell cycle. It is known that DNA damage and repair occurs normally in every living cell. When the rate of DNA damage exceeds that of repair, accumulation of DNA damage and defects might trigger the initiation of cancer [[1-3] and [4]].
Uterine cervical cancer is a serious gynecologic malignancy in women. There are two main types of cervical cancer, squamous cell cancer and adenocarcinoma, based on the type of cells that become cancerous. Cervical cancer is initiated when the combined action of a group of carcinogens cause the normal, physiological events associated with cervical metaplastic transformation to go awry and cause the formation of pre-malignant dysplasia [5]. Poor prognosis is usually associated with positive pelvic lymph nodes, indicating that the tumor cells have become metastatic [6].
Recent studies have demonstrated that estrogen, which is the female sex hormone, might have a contributory role in increasing vaginal epithelium proliferation and thus promoting the malignant transformation of the squamous and columnar cells at the junction of the cervical and vaginal epithelium [7]. Infection by the Human Papilloma Virus, HPV, is a necessary requirement for cervical cancer, but not all women infected by this virus develop cervical cancer [8]. Some HPV infections, for instance are associated with benign proliferation or wart formation.
B. Human Papilloma Virus (HPV)
HPVs are small DNA viruses that are known to be the most common etiological agents in cervical cancer [9]. More than 100 types of HPVs have been discovered, isolated and studied (See Table 1) [10]. HPVs are implicated in the mucosal and epithelial infections that may range from a benign lesion to a malignant carcinoma [4]. HPV has also been reported to be associated with anal and genital cancers [11]. Preliminary findings suggested their involvement in some head and neck cancers as well [10].
Table 1 Naturally occurring cancers associated with papillomaviruses [10, 13].
Species Cancer Predominant viral types
Humans Skin carcinomas HPV-5, -8
Lower genital tract cancers HPV-16, -18, -31, -33
Malignant progression of respiratory papillomas HPV-6, -11
Cattle Alimentary-tract carcinoma HPV-4
Eye and skin carcinoma Not characterized
Sheep Skin carcinoma Not characterized
Cottontail rabbit Skin carcinoma Cotton rabbit papillomavirus (CRPV)
The high risk HPV 16 and HPV 18 are associated with malignant transformation and carcinogenesis in 85% of the diagnosed cervical cancer cases [4]. Recent studies have shown that 13 different types of HPV are associated with carcinogenesis [3]. The most widely known factors associated with HPV are the E6 and E7 oncoproteins, which interact with p53 and Rb tumor suppressors respectively [2]. The interaction of E6 and E7 with these cellular proteins results in their suppression [9], thus disrupting the normal physiological process of programmed cell death in response to DNA damage (See Figure 1) [12]. In the presence of carcinogens, therefore, the accumulation of DNA damage without apoptosis is presumed to lead to cancer.
Figure 1 A schematic representation of RB/p53 interactions to regulate cell cycle and apoptosis. Cell cycle transition from G1-S phase is mediated by RB interactions with the E2F transcription factor family, which is considered an important regulator of the cell cycle. Growth factors lead to the phosphorylation of RB in late G1 phase by cdk/cyclin. This is followed by the release of E2F, allowing transcriptional activation of E2F target genes, which promotes S-phase entry and cell proliferation. HPV E7 and Simian Virus 40 (SV40) promote the release of E2F from RB, whereas HPV E6 and the dominant negative, DN-p53 inhibit p53 activity leading to cell proliferation.
It should be made clear that viral infection by itself does not cause cancer. It is the interaction of the viral genome with host genes that disrupts the normal cell cycle and transforms the cell into a pre-malignant state. For instance, some viruses might interact with specific genes (like tumor suppressor genes mentioned above) in the host cells, switching some systems on or off, thus leaving the cell free to divide in an uncontrolled way and raising the risk of cancer [13].
Other cellular proteins may be affected by HPV infection as well. For instance, cervical cancer cell lines showed overexpression of the anti-apoptotic protein BAG-1, which might contribute to its malignant proliferation [14]. Therefore, understanding the molecular mechanisms leading to this disease will be of importance for generating means for its early detection and possible prevention and treatment.
C. Apoptosis
Apoptosis, or programmed cell death, is orchestrated by a highly organized group of signaling pathway proteins [[15] and [16]]. Apoptosis can be triggered by a variety of events the cell may face. For example, exposures to X-rays, ultraviolet light and chemotherapeutic drugs are factors that can initiate the process of apoptosis [15].
One mechanism to protect an organism from the consequences of accumulated DNA damage involves a class of protein-splitting enzymes called caspases, which are activated upon detection of DNA damage and eventually cause cell death [17]. The control of programmed or physiological cell death acts as a protective mechanism for the organism because accumulation of DNA damage without concomitant repair could lead to the development of cancer, while unregulated apoptosis can cause autoimmune diseases.
The process of apoptosis is essential in stopping the uncontrolled proliferation of cells [18]. Any defects in this dynamic process may eventually lead to the development of benign proliferative lesions or even malignant tumors [18]. Apoptosis can be initiated via specific receptors that are members of the tumor necrosis factor (TNF) receptor superfamily [19]. Such "Death receptors" are activated via binding of specific ligands [19] and once they are activated, they can initiate apoptosis. However, the role of these receptors is not restricted to the initiation of apoptosis, but also includes other functions that differ from apoptosis and sometimes counteract apoptosis such as the recruitment and subsequent ligand binding of growth factors such as the Nerve Growth Factor (NGF) [19].
Another apoptotic pathway involves the mitochondria. Under stress, these essential organelles can release cytochrome c into the cytoplasm of the cell [20]. This cytochrome c release is a possible activator for caspases by the recruitment of procaspase-9, which undergoes conformational changes that leads to the activation of downstream, effector caspases [20].
Apoptosis acts as a double-edged sword. Despite its importance in restricting cell proliferation and maintaining constant cell number, excessive apoptosis is associated with stroke, Alzheimer's disease and other neurodegenerative disorders [21]. Damaged neurons in these disorders commit suicide inappropriately. Alzheimer's disease, for instance, was found to be associated with a genetic component that involves mutations in the chromosomes (1, 14, and 21) as well as the tau gene on chromosome 17 and results in unscheduled or unregulated death of brain cells [22].
Understanding the details and the signaling pathways of this phenomenon might be helpful in manipulating and intervening in the process of apoptosis. Apoptosis is required to restrict cell proliferation and to maintain a constant cell number. Attempts to suppress apoptosis, however, may be useful to treat neurodegenerative disorders, while attempts to activate apoptosis may be useful in disorders involving overproliferation.
D. Regulators of apoptosis
Many genes have been implicated in enhancing or inhibiting the process of apoptosis. They act by different mechanisms that ultimately contribute to either tumor suppression or progression, respectively. Four of the most important factors that regulate apoptosis are p53, the caspases, the Bcl-2 family of proteins and PARP [[23,24], and [25]].
1. Caspases
Caspases are the executioners of cell death. They receive the signals that enable them to initiate apoptosis. Cells undergoing apoptosis exhibit fragmentation of DNA, condensation of the chromatin, budding of the cell membrane and the formation of apoptotic bodies by dissociation of the cell and its constituents into membrane-enclosed vesicles [[17] and [26]]. All caspases share a similar structure that consists of three domains: an NH2-terminal peptide (prodomain), a large subunit (approximately 20 kD) and a small subunit (approximately 10 kD) [27]. Caspases are expressed as procaspases, which undergo cleavage to the 2 subunits mentioned above [28]. Cleavage of caspases is a sign of active apoptosis. The large and small subunits then associate to form a heterodimer [[29] and [23]].
The exact mechanism of action of caspases is still unknown. However, several studies have shown that caspases exert both direct and indirect actions on the cell [30]. The direct action of caspases can be exemplified by their ability to act on cell structural integrity by destroying the nuclear lamina [30] and cleaving the proteins responsible for regulating the cytoskeleton [[30] and [31]]. The indirect action of caspases is via their ability to inhibit the proteins that promote cell survival and growth [27]. Among these proteins is the Bcl-2 family of proteins, which are cleaved by caspases resulting in inactivation of the Bcl-2 proteins and the release of a fragment that has a direct apoptotic effect [30].
Caspases-8, -9, and -10 are known to initiate the caspase activation cascade. However, caspases-3, -6 and -7 propagate the cascade and are activated by the proteolytic cleavage process mediated by other upstream caspases in the caspase cascade pathway [27].
2. Bcl-2 family of proteins
The Bcl-2 family of proteins has several members with various functions [18]. The Bcl-2 gene family comprises pro-apoptotic and anti-apoptotic proteins sharing one or more Bcl-2 homology (BH) domains [23]. The gene family is made up of 3 main groups [27]. Group I includes the anti-apoptotic members similar to Bcl-2. Group II includes the pro-apoptotic members like Bax and Bak, while group III comprises a diverse collection of proteins that resemble one another structurally, but not necessarily functionally. Bcl-2, BAG-1 and Bcl-xL provide a cell survival function [[14] and [32]]. However, Bax, which promotes apoptosis translocates to the mitochondrial membrane and releases cytochrome c, which can initiate the apoptotic cascade [33]. It also competes for binding with Bcl-2 and with other members of the Bcl-2 superfamily of proteins [33]. Such heterodimerization between anti-apoptotic and pro-apoptotic members of this family is very common and is considered a regulatory mechanism for the decision to undergo apoptosis [23]. Thus, the balance between Bcl-2 and Bax is essential for the determination of the apoptotic potential of the cells, in which high apoptotic activity is often associated with a low Bcl-2/Bax level ratio [23].
BAG-1 has been shown to provide an anti-apoptotic effect. Its overexpression in cervical cancer suppressed apoptosis both independently and by increasing Bcl-2 protective activity, which further increased the resistance of cervical carcinoma to the effect of DNA-damaging agents [14].
3. p53
The tumor suppressor protein p53 has numerous functions (See Figure 1.2) [34]. Its principle role, however, is as a transcriptional regulator required for the expression of a number of genes involved in cell cycle regulation and apoptosis. The gene encoding p53 can be mutated in many forms of cancer including cervical, uterine, adenocarcinoma, adrenal and colorectal cancers. In cervical cancer, mutation patterns of p53 may vary from point mutation to deletion to base-pair alteration, however 30% of the cases showed a higher percentage of Guanine-Cytosine complementary base pairs compared to the Adenine-Thymine complementary base pairs suggesting that alteration in the base-pairing sequence is the major mutation pattern recognized in p53 [34]. A recent clinical study showed that the overexpression of p53 in cisplatin-treated tumors might be associated with resistance of the tumor to further cell death and apoptosis [35,36].
MDM2 is a p53-regulated protein that has a role in the translocation of p53 from the nucleus and enhances its proteosomal degradation [22]. Therefore, increased levels of MDM-2 and subsequent low levels of p53 are associated with increased cell growth and proliferation. The p53 tumor suppressor protein can also be targeted for degradation by the E6 oncogene of the Human Papilloma virus (HPV), thus promoting neoplastic proliferations (refer to Figure 1) [34].
4. PARP
Poly (ADP-ribose) polymerase, PARP, has recently been found to promote cell death, but the exact mechanism of action of PARP remains largely obscure. Many cellular enzymes were found to contain the PARP catalytic subunit, but they have different cellular localizations. Because PARP activation consumes much cellular energy, detection of abnormally high levels of PARP in cells might indicate excessive energy consumption and cellular exhaustion [22]. PARP is also known as an apoptosis-inducing factor and high levels of PARP are detected following DNA damage. Thus, this group of enzymes might also be involved in DNA repair, as well as apoptotic responses of the cells (summarized in Figure 2) [22].
Figure 2 Summary of the mechanism of action of the tumor suppressor protein, p53. After DNA damage, the tumor suppressor protein, p53, will be upregulated causing cell cycle arrest and enhancing DNA repair. However, in cases of irreversible DNA damage, p53 has been shown to transcriptionally repress the antiapoptotic gene Bcl-2, while it upregulates the pro-apoptotic proteins Bax and Fas. This in turn, promotes apoptosis. During apoptosis loss of the integrity of the mitochondrial membrane is followed by release of cytochrome c into the cytosol, this in turn leads to activation of caspase cleavage. Bax has been shown to contain p53-binding sites in its promoter site and is upregulated in response to DNA damage and increased p53 [36].
E. Rel/NF-κB family
The first identified member of the nuclear factor-kappa (κ)B (Rel/NF-κB) family was a protein found to be associated with a decameric oligonucleotide sequence in the enhancer element of the immunoglobulin kappa light chain in B-lymphocytes [[37,38], and [39]]. The Rel/NF-κB family is now known to be made up of a plethora of transcriptional regulators which share a 300 amino acid terminal domain called the Rel homology domain (RHD) [40,41]. This RHD comprises the DNA binding domain, nuclear localization signal (NLS), dimerization domains and the IκB binding domain (See Figure 3) [42,43]. Members of this family include [37]:
Figure 3 NF-κB and IκB proteins. A schematic representation of various domains in (Rel)/nuclear factor of kappa B (NF-κB) proteins including the Rel Homology Domain, RHD, which comprises the DNA binding domain, nuclear localization signal (NLS), dimerization domains and the IκB binding domain. (Rel)/nuclear factor of κB (NF-κB) proteins include those that do not require proteolytic processing and those that do require proteolytic processing. The first group consists of: RelA (known as p65), c-Rel and RelB and the second group includes NF-κB1 (known as p105) and NF-κB2 (known as p100), which further produce p50 and p52 proteins, respectively. These two groups dimerize, the most commonly detected NF-κB dimer is p50 – RelA. RelA is responsible for most of NF-κB transcriptional activity due to the presence of a strong transcriptional activation domain. p50 – c-Rel dimers are less abundant. Both p50 – RelA and p50 – c-Rel dimers are regulated by interactions with the inhibitor of κB (IκB) proteins, which cause their cytoplasmic localization. RelB, however, mostly associates with p100 and the p100 – RelB dimers are exclusively cytoplasmic. Proteolytic processing of p100 results in the release of p52 – RelB dimers, which translocate to the nucleus. RelB, unlike RelA and c-Rel, can function as an activator or repressor (Reproduced with permission from Nature Reviews Cancer (Vol 2, No. 4, pp 301#150;310 copyright (2002) Macmillan Magazines Ltd.).
1. NF-κB, including p50, p65, p105 (mice devoid of p65 generated by targeted "knock-out" gene disruption resulted in defects in fetal development localized to the spleen and liver. However, knock-out mice devoid of p105/p50 expression showed no defects during their development).
2. Lyt-10 (p100), including p100 and p52, which are required in spleen development.
3. c-rel (knock-out mice showed defects in the proliferation of B and T cells).
4. relB (knock-out mice showed defects in thymus development).
5. Dorsal, which is involved in the formation of the dorsal-ventral axis of the fruit fly Drosophila [42].
Rel/nuclear factor of kappa B (NF-κB) proteins include those that do not require proteolytic processing and those that do require proteolytic processing. The first group consists of: RelA (known as p65), c-Rel and RelB. The second group includes NF-κB1 (known as p105) and NF-κB2 (known as p100), which further produce p50 and p52 proteins, respectively (See Figure 3). Members of these two groups pair with each other with the most commonly detected NF-κB being a heterodimer of p50 and RelA. RelA is responsible for most of NF-κB's transcriptional activity due to the presence of a strong transcriptional activation domain at its C-terminus. p50-c-Rel dimers are less abundant.
Both p50-RelA and p50-c-Rel dimers are regulated by interactions with the inhibitor of κB (IκB) proteins, which cause their cytoplasmic localization. RelB, however, mostly associates with p100 and the p100-RelB dimers are exclusively cytoplasmic. Proteolytic processing of p100 results in the release of p52-RelB dimers, which then translocate to the nucleus. RelB, unlike RelA and c-Rel, can function as an activator or repressor [37,43]. Of the above-mentioned proteins, only p50 and p52 are produced from the cytoplasmic precursors p105 and p100, in the presence of ATP as an energy source [43]. However, the other members contain trans-activation domains and can act as activators or inhibitors of transcription based on dimers containing or lacking trans-activation domains (See Figure 3) [44].
The functions of the Rel/NF-κB family of proteins are strongly related to the target genes that contain the response elements for the protein [[37,45,46], and [47]]. For example, κB response elements are localized in IL-2, IL-2R, Ig κ and MHC Classes (I) and (II) genes, and here the Rel/NF-κB family of proteins function in modulating the immune system responses by binding to these target sequences and recruiting other immune system and inflammatory reaction mediators (See Figure 4 and Table 2). However, the Rel/NF-κB family of proteins is also directly involved in inflammatory reactions and acute phase responses when the κB binding sites are found in the regulatory sequences for the IL-1, IL-6, TNF-α, TNF-β and serum amyloid A protein genes. Also, the Rel/NF-κB family of proteins is involved in viral infections when the κB sites are found in the HIV-LTR, SV 40, CMV and adenovirus. Other functions of Rel/NF-κB proteins include growth regulation, immune system responses and cell adhesion molecules (see Table 2).
Figure 4 The steps involved in the activation of NF-κB family of transcription factors (Reproduced from Ponnappan, 1998, Feb 01; 3:d152-68 with permission from Frontiers in Bioscience). Activators of NF-κB like TNF-alpha, PMA, UV or LPS activate the NF-κB inducible kinase, which in turn phosphorylates at least IKK1 (I kappa B kinase-alpha) and sometimes IKK2 (I kappa B kinase-beta) in the I kappa B-kinase complex. Activators of NF-κB may directly activate the kinase complex as well. This may be followed by phosphorylation of the p105/p65 complex by the kinase complex, which is in turn followed by ubiquitination, proteasomal degradation and the nuclear translocation of NF kappa B. Inside the nucleus; NF kappa B promotes the transcription of immune response genes. The "??????" indicates the possibility of lowered translocation and consequent activation of NF-κB, which occurs in various diseases.
Table 2 Localization of κB binding motifs in the body suggests the functions of Rel/NF-κB [37].
κB sites Related functions
IL-2, IL-2R, Igκ, MHC Classes I and II Immune system reaction and responses.
IL-1, IL-6, TNF-α, TNF-β, serum amyloid A protein Inflammatory reactions and acute phase responses.
HIV-LTR, SV 40, CMV, adenovirus Viral infections
Rel/NF-κB family (NF-κB1, NF-κB2, c-rel, RelB) Immune system responses.
p53, c-Myc, Ras, pRB1 Growth regulation.
IκB-α, IκB-γ, p105, p100 and Bcl-3 IκB family members
I-CAM, V-CAM, E-selectin, ELAM1 Cell adhesion molecules
Activation of NF-κB transcription involves the translocation of NF-κB proteins to the nucleus as illustrated in Figure 4 [[37,42,43,48] and [49]]. The factors involved in the transcriptional activation of different members of the Rel/NF-κB family are mentioned in Table 3.
Table 3 The factors associated with activation of NF-κB transcription factor [37] and [42].
• Cytokines (TNF-α, IL-1, IL-2, IL-6)
• Bacterial lipopolysaccharides
• Phytohemagglutinin (PHA)
• Cross-linking surface CD2, CD3, CD28 and T-cell receptors.
• Proteins secreted by viruses, for example, tax, X, E1A
• Viral infections, for example, HIV-1, Hepatitis B, HSV, HTLV-1
• Antigenic stimulants for the T and B-cells receptors
• Ultraviolet light exposure
• X-irradiation
• Nitric oxide
• Hydrogen peroxide and other oxidizing agents
• Calcium ionophores
F. IκB inhibitor system
The multiple targets of Rel/NF-κB proteins and their multiple modes of regulation indicate that this family possesses diversity in function. Interestingly, their major mode of regulation appears to be well conserved through the IκB inhibitor system. IκB is a protein of 60–70 kDa [[38] and [43]]. The IκB inhibitor system comprises seven molecules IκB-α, IκB-β, IκB-γ, Bcl-3, p105, p100 and I κB R [45]. The inhibitor of κB (IκB) kinase (IKK) complex is composed of two catalytic subunits, IKKα and IKKβ, and one regulatory subunit, IKKγ.
The IκB inhibitor system regulates NF-κB (p50, p65) by retaining it as a complex in the cytoplasm [54]. As a result, the NF-κB family members remain in the cytoplasm in an inactive form. In response to stimuli such as tumour-necrosis factor-α (TNF-α), CD40 ligand (CD40L), interleukin-1 (IL-1) or lipopolysaccharide (LPS), the IKKβ subunit is activated, and phosphorylates the IκB proteins (bound to the NF-κB heterodimers) at two conserved serines. This phosphorylation event triggers the ubiquitin-dependent degradation of IκB by the 26S proteasome, resulting in the nuclear translocation of RelA-p50 (or c-Rel-p50) heterodimers and transcriptional activation of target genes (See Figure 5). In response to other stimuli, such as the TNF family members lymphotoxin B (LTβ) and BAFF, IKKα is activated to induce the phosphorylation of p100 (bound to RelB) at two serine residues at its carboxyl terminus. This phosphorylation event triggers the ubiquitin-dependent degradation of the carboxy-terminal half of p100, releasing its amino-terminal half, the p52 polypeptide, which together with its heterodimer partner, RelB, translocates to the nucleus to activate transcription [[38,43], and [45]] (See Figure 5 and Table 4).
Figure 5 Illustration of the Rel/NF-κB pathway. In response to stimuli such as tumour-necrosis factor-α (TNF-α), CD40 ligand (CD40L), interleukin-1 (IL-1) or lipopolysaccharide (LPS), the IKKβ subunit is activated, and phosphorylates the IκB proteins (bound to the NF-κB heterodimers) at two conserved serines. This phosphorylation event triggers the ubiquitin-dependent degradation of IκB by the 26S proteasome, resulting in the nuclear translocation of RelA – p50 (or c-Rel – p50) heterodimers and transcriptional activation of target genes [48].
Table 4 Steps involved in Rel/NF-κB activation [42].
Steps involved in Rel/NF-κB activation
1. Exposure to a stimulus that activates NF-κB such as UV light.
2. Degradation of IκB or its inactivation by phosphorylation by means of protein kinases (McKenzie et al., 2000).
3. Dissociation of the complex between NF-κB family members and IκB inhibitor system.
4. Translocation of NF-κB proteins to the nucleus.
5. DNA binding of NF-κB proteins.
6. Transcriptional induction by NF-κB proteins.
Rel/NF-κB family members also cooperate with other transcriptional regulators such as the non-Rel/NF-κB protein Ets-1. Recent data has provided evidence that physical interaction between Ets and NF-kappaB proteins is required for the transcriptional activity of the HIV-1 and HIV-2 enhancers [47]. These interactions represent a potential target for the development of novel immunosuppressive and antiviral therapies.
G. Role of NF-κB in apoptosis and cell survival
The dual role of NF-κB in enhancing or inhibiting apoptosis and cell death has attracted much attention in regard to its role in carcinogenesis.
NF-κB involvement in apoptosis
The role of Rel/NF-κB proteins in apoptosis has been well studied [[50] and [51]]. NF-κB, for instance, was found to be activated following TNF-α-induced apoptosis in several cell lines [52]. Treatment of cell lines derived from acute B-cell leukemia and human thymocytes with etoposide was found to activate NF-κB and this activation occurred prior to the initiation of apoptosis [44]. Further evidence supporting the involvement of NF-κB in apoptosis is the presence of NF-κB binding sites in the genes encoding IL-1β converting enzyme protease, c-myc, and TNFα, which are all involved in apoptosis and cell death [[44] and [53]]. Also, several studies showed that p65 is involved in apoptosis. This was based on an original observation whereby inhibition of apoptosis was achieved by overexpression of a dominant-negative p65 protein [54].
Rel/NF-κB role in cell protection and survival
The role of Rel/NF-κB proteins in cell survival is generally associated with their ability to upregulate the expression of myc [[55] and [56]]. Myc is a protein that mediates the transcriptional activation of cyclin A and cyclin D3, which are cell cycle regulators. A decrease in the myc protein concentration in the cell has been associated with apoptosis. Another pathway that leads to high myc levels is through the stimulation of CD40, a member of TNF receptor family, which results in NF-κB activation [57] and whose stimulation has been implicated in cell survival and protection [57].
The role of Rel/NF-κB as a factor in both cancerous and normal cell survival is well documented. Recent results, for instance, have shown NF-κB to be activated in the early malignant transformation of mammary cells of the breast [58]. Furthermore, NF-κB is constitutively active in pancreatic adenocarcinoma in humans [59], in T-cell leukemia cells [60], in human breast cancer [61] and in head and neck squamous cell carcinoma cell lines [62].
In non-cancer cells, NF-κB was reported to be essential for the growth and survival of sympathetic nerve cells independently of the de novo protein synthesis [63]. Further evidence of severe liver degeneration was associated with lack of NF-κB activation [64]. This was based on the death of murine embryonic fibroblasts that lack detectable NF-κB DNA binding activity in response to TNF-α, LPS, IL-1 and do not show IκB kinase activity required for NF-κB activation [64].
The anti-apoptotic activity of Rel/NF-κB can be regulated by other proteins. For instance, the X chromosome-linked inhibitor of apoptosis (XIAP) induces NF-κB activation by increasing the nuclear translocation of its p65 subunit [56]. In addition, CD95, which is known as Fas and possesses an apoptotic effect, was found to stimulate NF-κB degradation by caspases [66], but when an antibody against CD95 was used, caspases were inhibited and the inducibility of NF-κB was restored [66].
H. Xrel3
Xrel3 encodes an embryonic protein found to be related to the rel family of proteins. The Xrel3 gene is present in the genome of the amphibian, Xenopus laevis and is expressed in and is essential for the normal development of the head of Xenopus laevis embryos [67]. Xrel3 is also normally expressed in the otocysts and notochord of the embryonic larval stages [67]. Interestingly, Xrel3 overexpression has been implicated in the development of epidermal tumors in embryos [[14] and [67]], but little is known about how these tumors form, or whether they have similar properties to human tumors. Investigating whether the Xrel3 protein had properties that could contribute to human cancer has been explored by many researchers [[68,69] and [70]]. By applying what is known about the role of Xrel3 in embryos to human cell lines, it may be possible to uncover new knowledge about the mechanism of Rel/NF-κB activity in general.
In addition to its ability to cause embryonic tumor formation, the rationale for studying the effects of Xrel3 in human cervical cancer cells has basis in practicality. When a DNA vector encoding tagged-Xrel3 was transiently transfected into HeLa cells, Xrel3 protein constitutively localized in the nuclei, suggesting its ability to be active constantly in mammalian cells [68]. In addition, HeLa cells do not normally express Rel/NF-κB, so the transfection of Xrel3 into these cells gives the opportunity to study the activity of an interesting Rel/NF-κB protein in a negative background [[69] and [70]]. Therefore, even though Xrel3 is not a mammalian gene, its homology to the mammalian Rel/NF-κB family indicates that it may serve as a good model for gene regulation by this family enabling us to understand the mechanism of action of the Rel/NF-κB family of transcriptional regulators in cancer cells.
I. Cancer Chemotherapy
Many chemical agents are used in the treatment of cancer. Some can be used alone in single therapy and others have to be combined or added to other regimens for an effective outcome. The five groups of single chemotherapeutic agents are: alkylating agents, antimetabolites, plant derivatives, antitumor antibiotics and the miscellaneous group which contains the platinums, procarbazine, mitotane and gallium nitrate.
Platinums
The platinum-containing compounds are carboplatin, cisplatin and oxaliplatin. This group of chemotherapeutic drugs is very effective in monotherapy regimens [71]. They are the most active agents in the treatment of ovarian and cervical cancers. However, they are associated with three major drawbacks [72]:
1. Severe toxicity in the form of nephrotoxicity, ototoxicity, myelosuppression and peripheral neuropathy.
2. Narrow range of tumors upon which they are effective.
3. The development of resistance after a short period of treatment.
New approaches are now designed in an attempt to expand the mechanism of action of platinums. This is done by developing a new generation of platinum-containing compounds that exhibited a broader spectrum of activity on different tumors, lower toxicity potential as well as delayed resistance to treatment [71].
Cisplatin
Cis DiamminedichloroplatinII (cisplatin) is one of the platinum-containing anti-cancer agents. It can be recognized from its chemical name that the cis form is the active form of the drug. The trans form was found to possess no biologic activity [72]. The mechanism of action of cisplatin is similar to the alkylating agents, but it is not identical. Cisplatin works by promoting DNA cross-linking and chelation. Recent clinical studies have shown that improved cytotoxicity of cisplatin can be attained by increasing the exposure time of the tumor to the drug [73].
J. Rel/NF-κB and Chemoresistance
Many researches have attempted to investigate the role that NF-κB family might have in chemotherapeutic resistance. Activation of the Rel/NF-κB was found to be associated with chemotherapeutic resistance by suppressing the apoptotic potential of the chemotherapeutic drug. Recent data demonstrate that the protection from apoptosis induced in response to carbonyloxycamptothecin (CPT-11) treatment is effectively inhibited by the transient inhibition of NF-κB in a variety of human colon cancer cell lines [74]. This might be due to the cell survival effects associated with the upregulation of Rel/NF-κB family as previously mentioned. In addition, genetic manipulation aimed at inhibiting Rel/NF-κB, was found to cause sensitization of different tumor cells, like lung cancer cells, to the effect of chemotherapeutic drugs [75]. This makes the Rel/NF-κB family an attractive set of proteins to study in chemoresistant tumors.
The urge for overcomming the resistance encountered by the prolonged usage of chemotherapeutic drugs necessitates a deeper understanding of the underlying pathways that favors cell survival and inhibits apoptosis. An investigation of the upregulation of NF-κB might be a promising field of study in this regard since Rel/NF-κB activation has been associated with chemoresistance. Regarding chemotherapy, cisplatin can be used as a monotherapy without any adjuvant chemotherapeutic drugs. Cisplatin is also used in the treatment of gynecologic cancers like ovarian cancers [72]. Previous studies have shown that the apoptotic effect induced by chemotherapy in cervical cancer involves the apoptosis factor p53 and the HPV-E6 oncogenes and might be enhanced or attenuated depending on the platinum carrier ligand [76].
Recent studies by Shehata et al. investigated the effect of Xrel3 overexpression on the growth of HeLa cells with and without chemotherapeutic treatment [[69] and [70]]. Results showed that Rel/NF-κB might be a possible cause of chemotherapeutic resistance encountered in cervical cancer cells. This was observed at low doses of cisplatin treatment, where larger population of the malignant HeLa cells was present as compared to control cells transfected with an empty vector. However, at high concentrations of cisplatin, the upregulated Xrel3 enhanced apoptosis synergistic with cisplatin. This implies that Xrel3, a cRel homologue, possesses a dual apoptotic and antiapoptotic effect based on the degree of stress the cell might be facing.
==== Refs
Anonymous National Cancer Institute of Canada Canadian Cancer Statistics 2003
Furomoto HIM Irahara Human papilloma virus (HPV) and cervical cancer J Med Invest 2002 49 124 133 12323001
Munoz N Bosch FX Sanjose S Herrero R Castellsague X Shah KV Snijders PJ Meijer CJ International Agency for Research on Cancer Multicenter Cervical Cancer Study Group Epidemiologic classification of human papillomavirus types associated with cervical cancer N Engl J Med 2003 348 518 527 12571259 10.1056/NEJMoa021641
Garland SM Human papillomavirus update with a particular focus on cervical disease Pathology 2002 34 213 224 12109780 10.1080/00313020212469
Josefson D Mild cervical dysplasia often reverts to normal BMJ 1999 318 420 9974451
Kim CJ Jeong JK Park M Park TS Park TC Namkoong SE Park JS HPV oligonucleotide microarray-based detection of HPV genotypes in cervical neoplastic lesions Gynecologic Oncology 2003 89 210 217 12713982 10.1016/S0090-8258(02)00069-0
Park JS Rhyu JW Kim CJ Kim HS Lee SY Kwon YI Namkoong SI Sin HS Um SJ Neoplastic change of squamo-columnar junction in uterine cervix and vaginal epitheluim by exogenous estrogen in hpv-18 URR E6/E7 transgenic mice small star, filled Gynecologic Oncology 2003 89 360 368 12798696 10.1016/S0090-8258(02)00106-3
Castellsague X Bosch FX Munoz N Environmental co-factors in HPV carcinogenesis Virus Research 2002 89 191 199 12445659 10.1016/S0168-1702(02)00188-0
Ghim SJ Basu PS Jenson A Cervical Cancer: Etiology, Pathogenesis, Treatment, and Future Vaccines Asian Pac J Cancer 2002 3 207 214
Sisk EA Robertson ES Clinical implications of human papillomavirus infection Front Biosci 2002 1 e77 e84 Finzer,P, Lemarroy AA, Rosl F (2002). The role of human papillomavirus oncoproteins E6 and E7 in apoptosis. Cancer Letters, 188, 15-24 11861218
Heilmann V Kreienberg R Molecular biology of cervical cancer and its precursors Curr Womens Health Rep 2002 2 27 33 12112980
Finzer P Lemarroy AA Rosl F The role of human papillomavirus oncoproteins E6 and E7 in apoptosis Cancer Letters 2002 188 15 24 12406543 10.1016/S0304-3835(02)00431-7
zur Hausen A van Beek J Bloemena E ten Kate FJ Meijer CJ van den Brule AJ No role for Epstein-Barr virus in Dutch hepatocellular carcinoma: a study at the DNA, RNA and protein levels J Gen Virol 2003 84 1863 1869 12810881 10.1099/vir.0.19217-0
Yang X Hao Y Ferenczy A Tang SC Pater A Overexpression of Anti-apoptotic Gene BAG-1 in Human Cervical Cancer Experimental Cell Research 1999 247 200 207 10047462 10.1006/excr.1998.4349
Miller LJ Marx J Apoptosis Science 1998 281 1301 1304 10.1126/science.281.5381.1301
Fiers W Beyaert R Declercq W Vandenabeele P More than one way to die: apoptosis, necrosis and reactive oxygen damage Oncogene 1999 18 7719 7730 10618712 10.1038/sj.onc.1203249
Thornberry NA Lazebnik Y Caspases: Enemies Within Science 1998 281 1312 1316 9721091 10.1126/science.281.5381.1312
Nicholson DW From bench to clinic with apoptosis-based therapeutic agents Nature Insight 2000 407 810 816
Ashkenazi A Dixit VM Death receptors: Signalling and Modulation Science 1998 281 1305 1308 9721089 10.1126/science.281.5381.1305
Green DR Reed JC Mitochondria and Apoptosis Science 1998 281 1309 1312 9721092 10.1126/science.281.5381.1309
Yuan J Yanker BA Apoptosis in the nervous system Nature Insight 2000 407 802 809
Rich T Allen RLWAH Defying death after DNA damage Nature Insight 2000 407 777 783
Reed JC Double identity for proteins of the Bcl-2 family Nature 1997 387 773 776 9194558 10.1038/42867
Packham G Brimmell M Cleveland JL Mammalian cells express two differently localized Bag-1 isoforms generated by alternative translation initiation Journal of Biochemistry 1997 328 807 813
Green DR Apoptotic Pathways: Paper Wraps Stone Blunts Scissors Cell 2000 102 1 4 10929706 10.1016/S0092-8674(00)00003-9
Savill J Fadok V Corpse clearance defines the meaning of cell death Nature Insight 2000 407 784 788
Hengartner MO The biochemistry of apoptosis Nature Insight 2000 407 770 776
Meier P Finch A Evan G Apoptosis in development Nature Insight 2000 407 796 801
Boise LH Gottschalk AR Quintans J Thonpson CB Bcl-2 and Bcl-2-related proteins in apoptosis regulation Curr Top Microbiol Immunol 1995 200 107 121 7634826
Evan G Littlewood T A Matter of Life and Cell Death Science 1998 281 1317 1322 9721090 10.1126/science.281.5381.1317
Boyd JM Gallo GJ Elangovan B Houghton AB Malstrom S Avery BJ Ebb RG Subramanian T Chittenden T Lutz RJ Bik, a novel death-inducing protein shares a distinct sequence motif with Bcl-2 family proteins and interacts with viral and cellular survival-promoting proteins Oncogene 1995 11 1921 1928 7478623
Stuart JK Myszka DG Joss L Mitchell RS McDonald SM Xie Z Takayama S Reed JC Ely KR Characterization of Interactions between the Anti-apoptotic Protein BAG-1 and Hsc70 Molecular Chaperones The Journal of Biological Chemistry 1998 273 22506 22514 9712876 10.1074/jbc.273.35.22506
Ferrer I Planas AM Signaling of cell death and cell survival following focal cerebral ischemia: life and death struggle in the penumbra J Neuropathol Exp Neurol 2003 62 329 339 12722825
Rosenthal AN Ryan A Al-Jejani RM Storey A Harwood CA Jacobs IJ p53 codon 72 polymorphism and risk of cervical cancer in UK Lancet 1998 352 871 874 9742979 10.1016/S0140-6736(98)07357-7
Nakayama K Takebayashi Y Nakayama S Hata K Fujiwaki R Fukumoto M Miyazaki K Prognostic value of overexpression of p53 in human ovarian carcinoma patients receiving cisplatin Cancer Letters 2003 192 227 235 12668287 10.1016/S0304-3835(02)00686-9
Famuboni AK Graveling AJ Markey AL Minns FC Patel A Apoptosis 2002
Verma IM Stevenson JK Schwarz EM Antwerp DV Rel/NF-κB/IκB family: intimate tales of association and dissociation Genes and Development 1995 9 2723 2735 7590248
Baeuerle PA Baltimore D IκB: A Specific Inhibitor of the NF-κB Transcription Factor Science 1988 242 540 546 3140380
Govind S Control of development and immunity by Rel transcription factors in Drosophila Oncogene 1999 18 6875 6887 10602463 10.1038/sj.onc.1203223
Castranova V Chen F Shi X Demers LM New Insights into the Role of Nuclear Factor-κB, a Ubiquitous Transcription Factor in the Initiation of Diseases Clinical Chemistry 1998 45 7 17 9895331
Gerondakis S Grossmann M Nakamura Y Pohl T Grumont R Genetic approaches in mice to understand Rel/NF-κB and IκB function: transgenics and knockouts Oncogene 1999 18 6888 6895 10602464 10.1038/sj.onc.1203236
Liou HC Baltimore D Regulation of the NF-κB /rel transcription factor and IκB inhibitor system Current Opinion in Cell biology 1993 5 477 487 8352966 10.1016/0955-0674(93)90014-H
May MJ Ghosh S Signal transduction through NF-κB Immunology Today 1998 19 80 88 9509763 10.1016/S0167-5699(97)01197-3
Beg AA Baldwin AS Jr The IκB proteins: multifunctional regulators of Rel/NF-κB trancription factors Genes and Development 1993 7 2064 2070 8224838
Garcia GE Xia Y Chen S Wang Y Ye RD Harrison JK Bacon KB Zerwes HG Feng L NF-κB-dependent fractalkine induction in rat aortic endothelial cells stimulated by IL-1beta, TNF-alpha, and LPS Journal of Leukocyte Biology 2000 67 577 584 10770292
Henkel T Machleidt T Alkalay I Kronke M Ben-Neriah Y Baeuerle PA Rapid proteolysis of I kappa B-alpha is necessary for activation of transcription factor NF-kappa B Nature 1993 365 182 185 8371761 10.1038/365182a0
Bassuk AG Anandappa RT Leiden JM Physical interactions between Ets and NF-kappaB/NFAT proteins play an important role in their cooperative activation of the human immunodeficiency virus enhancer in T cells Journal of Virology 1997 71 3563 3573 9094628
Hiscott J Kwon H Genin P Hostile takeovers viral appropriation of the NF-kB pathway J Clin Invest 2001 107 143 151 11160127
Shain KH Jove R Olashaw NE Constitutive RelB Activation in v-Src-Transformed Fibroblasts: Requirement for I κB Degradation Journal of Cellular Biochemistry 1999 73 237 247 10227387 10.1002/(SICI)1097-4644(19990501)73:2<237::AID-JCB10>3.0.CO;2-H
Hogerlinden MV Rozell BL Richter LA Toftgard R Squamous Cell Carcinomas and Increased Apoptosis in Skin with Inhibited Rel/Nuclear Factor-κB Signaling Cancer research 1999 59 3299 3303 10416581
Manna SK Mukhopadhyay A Aggarwal BB Resveratrol Suppresses TNF-Induced Activation of Nuclear Transcription Factors NF-κB, Activator Protein-1, and Apoptosis: Potential Role of Reactive Oxygen Intermediates and Lipid Peroxidation The Journal of Immunology 2000 164 6509 6519 10843709
Miyamoto S Maki M Schmitt MJ Hatanaka M Verma IM Tumor necrosis factor alpha-induced phosphorylation of I kappa B alpha is a signal for its degradation but not dissociation from NF-kappa B Proc Natl Acad Sci U S A 1994 91 12740 12744 7809113
Collins T Read MA Neish AS Whitley MZ Thanos D Maniatis T Transcriptional regulation of endothelial cell adhesion molecules: NF-kappa B and cytokine-inducible enhancers FASEB J 1995 9 899 909 7542214
Higgins KA Coleman TA McComas WA Perez JR Dorshkind K Sarmiento UM Rosen CA Narayanan R Antisense inhibition of the p65 subunit of NF-kappa B blocks tumorigenicity and causes tumor regression Proc Natl Acad Sci U S A 1993 90 9901 9905 8234333
Foo SY Nolan GP NF-κB to the rescue Trends Genet 1999 15 229 235 10354583 10.1016/S0168-9525(99)01719-9
Chen F Castranova V Shi X New Insights into the Role of Nuclear Factor-κB in Cell Growth Regulation American Journal of Pathology 2001 159 387 397 11485895
Madrid LV Wang CY Guttridge DC Schottelius AJG Baldwin AS JrMayo MW Akt Suppresses Apoptosis by Stimulating the Transactivation Potential of the RelA/p65 Subunit of NF-κB Molecular and Cellular Biology 2000 20 1626 1638 10669740 10.1128/MCB.20.5.1626-1638.2000
Kim DW Sovak MA Zanieski G Nonet G Romieu-Mourez R Lau AW Hafer LJ Yaswen P Stampfer M Rogers AE Russo J Sonenshein GE Activation of NF-κB/Rel occurs early during neoplastic transformation of mammary cells Carcinogenesis 2000 21 871 879 10783306 10.1093/carcin/21.5.871
Wang W Abbruzzese JL Evans DB Chiao PJ Overexpression of urokinase-type plasminogen activator in pancreatic adenocarcinoma is regulated by constitutively activated RelA Oncogene 1999 18 4554 4563 10467400 10.1038/sj.onc.1202833
Mori N Fujii M Ikeda S Yamada Y Tomonaga M Ballard DW Yamamoto N Constitutive Activation of NF-κB in Primary Adult T-Cell Leukemia Cells Blood 1999 2360 2368 10090947
Cogswell PC Guttridge DC Funkhouser WK Baldwin AS Jr Selective activation of NF-κB subunits in human breast cancer: potential roles for NF-κB2/p52 and for Bcl-3 Oncogene 2000 19 1123 1131 10713699 10.1038/sj.onc.1203412
Ondrey FG Dong G Sunwoo J Chen Z Wolf JS Crowl-Bancroft CV Mukaida N Waes CV Constitutive Activation of Transcription Factors NF κB, AP-1, and NF-IL6 in Human Head and Neck Squamous Cell Carcinoma Cell Lines that Express Pro-inflammatory and Pro-angiogenic Cytokines Molecular Carcinogenesis 1999 26 119 129 10506755 10.1002/(SICI)1098-2744(199910)26:2<119::AID-MC6>3.0.CO;2-N
Maggirwar SB Sarmiere PD Dewhurst S Freeman RS Nerve Growth Factor-Dependent Activation of NF-κB Contributes to Survival of Sympathetic Neurons The Journal of Neuroscience 1999 18 10356 10365 9852573
Rudolph D Yeh WC Wakeham A Rudolph B Nallainathan D Potter J Elia AJ Mak TW Severe liver degeneration and lack of NF-κB activation in NEMO/IKKγ-deficient mice Genes and Development 2000 854 862 10766741
Stehlik C Rainer de Martin Kumabashiri I Schmid JA Binder BR Lipp J Nuclear Factor (NF) – κB -regulated X-chromosome-linked iap Gene Expression Protects Endothelial Cells from Tumor Necrosis Factor α-induced Apoptosis J Exp Med 1998 188 211 216 9653098 10.1084/jem.188.1.211
Ravi R Bedi A Fuchs EJ Bedi A CD95 (Fas)-induced Caspase-mediated Proteolysis of NF-κB Cancer research 1998 58 882 886 9500443
Lake BB Ford R Kao KR Xrel3 is required for head development in Xenopus laevis Development 2001 128 263 273 11124121
Green A Effect of Xrel3 on TGF-β Signalling Pathway in Human Cervical Cancer cells 2003 Memorial University of Newfoundland
Shehata M Shehata M Shehata F Pater A Dual apoptotic effect of Xrel3 c-Rel/NF-kappaB homolog in human cervical cancer cells Cell Biol Int 2004 28 895 904 15566959 10.1016/j.cellbi.2004.09.002
Shehata M Shehata M Shehata F Pater A Apoptosis Effects of Xrel3 c-Rel/Nuclear Factor-Kappa B homolog in Human Cervical Cancer Cells Cell Biol Int
Fuertes MA Castilla J Alonso C Perez JM Novel concepts in the development of platinum antitumor drugs Curr Med Chem Anti-Canc Agents 2002 2 539 551 12678734 10.2174/1568011023353958
Reedijk J New Clues for platinum antitumor chemistry: Kinetically controlled metal binding to DNA National Academy of Sciences 2003 100 3611 3616 10.1073/pnas.0737293100
Markman M Intraperitoneal antineoplastic drug delivery: rationale and results The Lancet Oncology 2003 4 277 283 12732164 10.1016/S1470-2045(03)01074-X
Cusack JC JrLiu R Baldwin AS Jr Inducible Chemoresistance to 7-Ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin (CPT-11) in Colorectal Cancer Cells and a Xenograft Model Is Overcome by Inhibition of Nuclear Factor-κB Activation Cancer Research 2000 60 2323 2330 10811101
Jones DR Broad M Madrid LV Baldwin AS JrMayo MW Inhibition of NF-κB sensitizes non-small cell lung cancer cells to chemotherapy-induced apoptosis Ann Thorac Surg 2000 70 930 936 11016336 10.1016/S0003-4975(00)01635-0
Koivusalo R Krausz E Ruotsalainen P Helenius H Hietanen S Chemoradiation of Cervical Cancer Cells: Targeting Human Papillomavirus E6 and p53 Leads to Either Augmented or Attenuated Apoptosis Depending on the Platinum Carrier Ligand Cancer research 2002 62 7364 7371 12499281
Ponnappan U Regulation of transcription factor NFkappa B in immune senescence Frontiers in Bioscience 1998 3 D152 168 9445466
| 15857509 | PMC1090600 | CC BY | 2021-01-04 16:40:10 | no | Cancer Cell Int. 2005 Apr 27; 5:10 | utf-8 | Cancer Cell Int | 2,005 | 10.1186/1475-2867-5-10 | oa_comm |
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-111585750410.1186/1475-2867-5-11Primary ResearchIn vitro analysis of the invasive phenotype of SUM 149, an inflammatory breast cancer cell line Hoffmeyer Michaela R [email protected] Kristin M [email protected] Suranganie F [email protected] Department of Molecular Cell and Developmental Biology, University of Texas Austin, 205 West 24th Street, Austin, Texas, 78712, USA2 Department of Biomedical Engineering, University of Texas at Austin, 1 University Station Stop C0800, Austin, TX 78712, USA2005 27 4 2005 5 11 11 15 11 2004 27 4 2005 Copyright © 2005 Hoffmeyer et al; licensee BioMed Central Ltd.2005Hoffmeyer 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
Inflammatory breast cancer (IBC) is the most lethal form of locally invasive breast cancer known. However, very little information is available on the cellular mechanisms responsible for manifestation of the IBC phenotype. To understand the unique phenotype of IBC, we compared the motile and adhesive interactions of an IBC cell line, SUM 149, to the non-IBC cell line SUM 102.
Results
Our results demonstrate that both IBC and non-IBC cell lines exhibit similar adhesive properties to basal lamina, but SUM 149 showed a marked increase in adhesion to collagen I. In vitro haptotaxis assays demonstrate that SUM 149 was less invasive, while wound healing assays show a less in vitro migratory phenotype for SUM 149 cells relative to SUM 102 cells. We also demonstrate a role for Rho and E-cadherin in the unique invasive phenotype of IBC. Immunoblotting reveals higher E-cadherin and RhoA expression in the IBC cell line but similar RhoC expression. Rhodamine phalloidin staining demonstrates increased formation of actin stress fibers and larger focal adhesions in SUM 149 relative to the SUM 102 cell line.
Conclusion
The observed unique actin and cellular architecture as well as the invasive and adhesive responses to the extracellular matrix of SUM 149 IBC cells suggest that the preference of IBC cells for connective tissue, possibly a mediator important for the vasculogenic mimicry via tubulogenesis seen in IBC pathological specimens. Overexpression of E-cadherin and RhoA may contribute to passive dissemination of IBC by promoting cell-cell adhesion and actin cytoskeletal structures that maintain tissue integrity. Therefore, we believe that these findings indicate a passive metastatic mechanism by which IBC cells invade the circulatory system as tumor emboli rather than by active migratory mechanisms.
==== Body
Background
With an average five-year post-recovery survival rate of 45%, inflammatory breast cancer (IBC) is the most lethal and aggressive form of locally advanced breast cancer [1]. The lethality of IBC stems from its highly invasive nature. Diagnosis of IBC is often complicated by lack of a palpable precursor lesion commonly associated with breast cancer. Moreover, the correct diagnosis is hindered by inflammatory-like symptoms such as redness, warmth, and edema. Characteristic of IBC is a change in breast skin texture, similar to that of an orange, due to extensive invasion of the dermal lymphatics by IBC tumor cell emboli. These complications contribute to IBC lethality in that by the time a proper diagnosis is made, the cancer has aggressively infiltrated the surrounding tissue and lymphatics system, leading to a lowered patient prognosis [2]. Complicating treatment of this deadly form of breast cancer is that very little information about the cellular mechanisms responsible for the unique IBC phenotype is known.
Cancer cell invasion through the basal lamina and subsequent metastasis involves multiple steps including intravasation through the surrounding tissue into the lymphatic or vascular systems. Transient adhesion to extracellular matrix (ECM) components as well as modification of cell shape by reorganization of the actin cytoskeleton is required for cancer cell infiltration into the adjacent tissue. The Rho GTPases regulate actin cytoskeletal rearrangements, and are thus likely candidates for involvement in cancer cell invasion and metastasis [3,4]. Further evidence for a relationship between cancer cell mobilization and dysregulation of Rho GTPases is seen in the overexpression of Rho proteins in numerous invasive human cancers. The recent discovery of the overexpression of the Rho isoform RhoC by IBC tumors has been implicated in the physiological mechanisms of this poorly characterized form of breast cancer [5]. RhoC was demonstrated to be overexpressed in metastatic tumors of pancreatic adenocarcinoma patients [6], murine melanomas [7], and in the patient-derived IBC cell line SUM 149 [5]. Transient inhibition of RhoC in IBC cells by treatment with farnesyl transferase inhibitors reduced invasion and motility in vitro [8]. Recently it was reported that RhoC overexpression in mammary epithelial cells resulted in a significant increase in cell migration [9], mediated by the MAPK pathway [10]. These findings led us to hypothesize that RhoC overexpression may promote the highly invasive phenotype of IBC and contribute to the uniquely aggressive phenotype exhibited by IBC.
Another unique feature of IBC is the overexpression of E-cadherin, a transmembrane protein involved in cell-cell adhesion, which is generally lost in highly invasive cancers. It seems somewhat paradoxical that such an aggressive cancer that overexpresses proteins involved in actin cytoskeleton rearrangement and promotion of migration (i.e., RhoC) also overexpresses cell-cell junction proteins such as E-cadherin [11-15]. The literature thus far seems to hold to two schools of thought about the contradictory protein expression seen in IBC. One tends to support the idea that E-cadherin expression fluctuates with disease progression and decreases as IBC cells become invasive [15]. The second school supports the theory of passive metastasis by IBC [11,12]. In passive metastasis, strong tumor cell-cell adhesions are maintained during dissemination that proceeds via vasculogenesis through secretion of differentiation factors by the tumor cells causing de novo vessel formation [16]. This results in a cancer cell cluster within the vessel, reminiscent of the IBC tumor emboli seen in IBC histology. Furthermore, RhoC overexpression in human mammary epithelial cells has been shown to increase production of angiogenic factors, some of which might mediate passive or active metastasis [17].
The IBC phenotype has mystified clinicians due to its inflammatory-like symptoms. However IBC symptomology is not considered to be a true immunoreaction, but rather a consequence of cancer cell invasion to the lymphatics system. The mechanism by which IBC invades is unclear and further experimentation with IBC models is required to clarify the exact mechanism by which this form of breast cancer is disseminated. Using the SUM 149 IBC cell line, we have examined the adhesive and migratory capacities in an effort to understand the invasive behavior of IBC for future experimentation with in situ imaging of IBC in animal models. SUM 149 was compared to a control cell line, SUM 102, which was selected because it shares a deletion in the LIBC (lost in inflammatory breast cancer) gene with the SUM 149 cell line but reportedly expresses RhoC mRNA at low levels [5]. We show that SUM 149 is less invasive and adhesive to basal lamina components in vitro than SUM 102, and that SUM 149 expresses more Rho proteins and E-cadherin. These data shows that SUM 149 is not highly motile and therefore possibly not actively invasive, suggesting passive metastasis as the mechanism of IBC dissemination.
Results
Endogenous Levels of Rho
Figure 1A presents the relative protein levels of the various Rho isoforms in the IBC cell line SUM 149 versus the non-IBC cell line SUM 102. Previous investigators have reported overexpression of RhoC mRNA in IBC cells compared to SUM 102 [5]. To verify overexpression of RhoC at the protein level, we preformed Western blots on cell lysates using a RhoC polyclonal antibody (Santa Cruz Biotechnology, CA). We found no significant difference in RhoC protein levels between the SUM 149 and the SUM 102 cell lines. However, immunoblotting revealed a significant difference in Rho (A, B, and C), with the IBC cell line expressing much higher Rho protein levels. We then examined RhoA protein levels and found a significant overexpression of RhoA in the IBC cell line. This finding is interesting considering that RhoA has been shown to play a vital role in actomyosin-mediated contractility [21,24].
Figure 1 1A. Rho GTPase protein expression levels in the IBC cell line SUM 149 versus SUM 102. 1B. F-actin and focal adhesion distribution in SUM 149 and SUM 102 human breast cancer cell lines. 1A. Equal protein amounts were separated by 10% SDS-PAGE, transferred to nitrocellulose, and probed for RhoA, RhoC, and Rho (A, B, and C). Images are representative of at least three independent experiments with actin serving to verify equal protein loading. 1B. SUM 149 and SUM 102 cells were starved in unsupplemented F-12 Hams media for 24 hours and stimulated for 10 minutes with PBS (control), EGF (50 ng/ml), or FBS (5%). Cells were stained with rhodamine phalloidin (red) to visualize F-actin and anti-phosphotyrosine (green) to visualize focal adhesion. Micrographs were taken at 1000× magnification. Images are representative for at least three independent experiments.
Subcellular Distribution of Focal Adhesions and Filamentous Actin
Because RhoA is involved in actin stress fiber and focal adhesion formation, we stained the cells with rhodamine phalloidin to visualize F-actin and anti-phosphotyrosine to visualize focal adhesions. Figure 1B demonstrates F-actin and focal adhesion distribution in both cell lines. SUM 149 displayed larger focal adhesions and more actin stress fibers than the SUM 102 cell line, as might be expected from the high levels of RhoA in the SUM 149 cell line. Upon stimulation of quiescent cells with EGF or FBS, the SUM 102 cells formed large membrane ruffles (lamellipodia). However, stimulation by both EGF and FBS seemed to have little effect on the actin cytoskeleton of the SUM 149 cells. An increase in focal adhesion was seen in the SUM 149 cells after stimulation with EGF, but no clear cell polarization was observed.
Adhesion to Extracellular Matrix Proteins
Invasion and metastasis is a multi-step process in which cells must break local connections, move through the basal lamina, survive in circulation, and reestablish cellular attachment at distant sites. Clearly, many of these steps involve interaction with ECM components. Transient adhesion to the ECM in conjunction with cytoskeletal rearrangements are requirements for cell motility. To examine the ability of the breast cancer cell lines under study to adhere to the various ECM proteins, we performed adhesion assays (Figure 2). Here, we demonstrate that both SUM 149 and SUM 102 cells have similar adhesive properties on laminin, the major component of the basal lamina. A slight increase in adhesive properties for the IBC cells was observed compared to the SUM 102 cells on collagen IV. However, a marked increase in adhesion to collagen I, the major component of connective tissue, was seen for the SUM 149 cell line. Taken together, this data suggest that the exacerbated invasive phenotype seen in IBC is not due to differences in adhesive properties to basal lamina components, but may indicate a preference of IBC cells to the connective tissue, through which these cells must invade before entering circulation. Furthermore, attachment to connective tissue components maybe important for vasculogenic mimicry via tubulogenesis, as seen in IBC pathological specimens [25-27].
Figure 2 Adhesion of human breast cancer cells to extracellular matrix proteins. SUM 149 and SUM 102 cells (105) were plated on glass coverslips coated with laminin (50 μg/ml), collagen I (10 μg/ml), or collagen IV (10 μg/ml) and allowed to adhere for 15 minutes. Micrographs were taken at 400× magnification. Adherent cells were quantified in 10 random microscopic fields. Data are expressed as mean ± SEM of at least three independent experiments.
Haptotaxis Stimulated Invasion
An aggressively infiltrative cancer must invade surrounding tissue by movement through the ECM. Haptotaxis, or cell movement toward ECM proteins, was assayed in vitro and is presented in Figure 3. SUM 149 cells were significantly less invasive into laminin (basal lamina component) and collagen I (connective tissue component) after 24 hours than the SUM 102 cell line. Therefore, SUM 149 was less invasive when assayed in a manner that requires individual cell movement by an active motile mechanism through a membrane with 8 μm pores.
Figure 3 Haptotaxis of human breast cancer cells to extracellular matrix proteins. SUM 149 and SUM 102 cells (105) were placed into the top chamber of a Costar well coated with laminin (50 μg/ml) or collagen I (10 μg/ml) and allowed to invade for 24 hours. Invasive cells were stained with propidium iodide and quantified in 10 random microscopic fields. Micrographs were taken at 200× magnification. Data are expressed as mean ± SEM of at least three independent experiments.
Subcellular Distribution of Filamentous Actin Subsequent to Cellular Polarization
To induce cell polarization and migration that does not involve active migration of individual cells but rather collective cell migration over a wound edge, we performed wound healing assays as described in [28]. A confluent monolayer of cells was wounded, leading to the release of chemotractant signals by the cells at the wound edge, thus mimicking cell motility cues in vivo (Figure 4). Here, we present that the IBC cell line SUM 149 was less responsive to the cell-derived migration signals after 7.5 hours. By this time, the SUM 102 cell line was much more invasive into the wound space and had nearly closed the wound entirely. Thus, active migration as a sheet of cells is not likely the mechanism by which IBC is disseminated.
Figure 4 Human breast cancer cell migration in response to wounding. SUM 149 and SUM 102 cells were grown to confluency on glass coverslips and wounded with a sterile razor blade. Closure of the wound was monitored over 7.5 hours. Cells were stained with rhodamine phalloidin to visualize F-actin reorganization in response to cell migration. Arrows indicate the wound edge. Micrographs were taken at 1000× magnification. Images are representative of at least three independent experiments.
Endogenous Expression of E-cadherin
An interesting and perplexing characteristic of IBC is the expression of E-cadherin by this invasive form of breast cancer. Usually the loss of E-cadherin correlates with increased invasive and metastatic potential [29]. To verify E-cadherin expression in the IBC cell line SUM 149, we performed immunofluorescence experiments (Figure 5A). Both SUM 149 and SUM 102 show E-cadherin staining localized to the shared margins between neighboring cells, however this staining is much more intense in the SUM 149 cell line. Immunoblotting analysis likewise demonstrates E-cadherin expression by both cell lines with higher levels of E-cadherin expressed in the SUM 149 cells (Figure 5B).
Figure 5 5A. E-cadherin distribution in SUM 149 and SUM 102 human breast cancer cells. 5B. E-cadherin protein expression levels in the IBC cell lineSUM 149 versus SUM 102. 5A. Cells were grown to 60% confluency and stained with anti-E-cadherin (green). Arrows indicate cell-cell adhesions containing E-cadherin. Micrographs were taken at 1000× magnification. Images are representative for at least three independent experiments. 5B. Cell lysates were separated by 8% SDS-PAGE, transferred to nitrocellulose, and probed for E-cadherin. Images are representative of at least three independent experiments with actin serving to verify equal protein loading.
Discussion
IBC is a unique and highly aggressive form of locally advanced breast cancer with distinct clinical presentation. We hypothesized that upregulated expression of RhoC, as reported by others to be characteristic of IBC, contributes to the unusual pathological presentation of IBC. For the first time, we have compared the actin architecture, invasive, and adhesive properties of the IBC cell line SUM 149 with a cell line reported to express less RhoC mRNA compared to SUM 149 but share a deletion in LIBC [5]. Using a commercially available specific antibody to RhoC, we report that RhoC is not overexpressed at the protein level by the IBC cell line SUM 149. Interestingly, we confirmed overexpression of RhoA by utilizing an anti-RhoA specific antibody. However, post-transcriptional regulation of RhoC expression may account for the observed discrepancy. It is possible that our results do not agree with the reported mRNA expression due to specificity problems with the commercially developed antibodies. Furthermore, we demonstrate that, compared to SUM 102, SUM 149 is less invasive and migratory, and displays impaired adhesion to basal lamina components but strong adhesion to connective tissue proteins.
The role of the Rho protein in cancer cell invasion is somewhat controversial. RhoA is known to be involved in cell contractility, both in the formation of bundled actin fibers and through the activation of Rho kinase and subsequent activation of myosin light chain [30]. Such contractile cells have previously been shown to be less motile [31]. However, Rho overexpression has been documented in various human cancers such as bladder and ovarian, and correlates with lymph node invasion, metastasis, and poor patient prognosis [32,33]. Overexpression of RhoC by human mammary epithelial cells increased invasion, motility, and anchorage independent growth, similar to SUM 149 [9]. Expression of dominant negative Rho T19N has been demonstrated to block melanoma cell invasion [34]. Some investigators report that Rho overexpression has little impact on invasion and cell motility, while others demonstrate a positive correlation between Rho expression and cell migration capacity [35-37]. Rho is required for cell body contraction and tail retraction during directed cell motility, while active Rac and Cdc42 are required for lamellipodia and filopodia extension at the leading edge [30]. Thus, invasive potential is considered to be a balance between Rac, Cdc42, and Rho activities. Overexpression or activation of one of these Rho GTPases will shift this balance and result in a cellular phenotype dominated by the actin structure promoted by the activated Rho GTPase [38]. SUM 149 may display reduced invasion and migration in vitro compared to SUM 102 due to the overexpression of RhoA alone, thus masking the motile effects of Rac and Cdc42.
Another aspect that makes IBC so remarkable is that this form of aggressive breast cancer maintains strong E-cadherin expression [11-15]. Typically, loss of E-cadherin expression correlates with progression to metastatic disease since cancer cells must break inter-cell adhesions before attaining a motile phenotype [29]. Here, we demonstrate that the SUM 149 model of IBC maintains strong E-cadherin expression in culture, as seen in other IBC xenograft models and IBC pathological specimens. Previous reports indicate the E-cadherin axis is also complete and functional [11]. IBC histology reveals an extensive invasion of E-cadherin positive tumor cell emboli within the dermal lymphatics [11-15]. The expression of E-cadherin may be critical for invasion in that IBC is thought by some to be passively disseminated, an invasion mechanism that necessitates cell-cell attachment [12]. In this scenario, tumor cells maintain strong cell-cell connections and enter circulation via vasculogenesis around a tumor cell embolus. Others hold that E-cadherin expression varies with the malignant stage of the disease, and is lost during invasion but reestablished once tumor cells invade the vasculature [15]. The finding reported here, in which the IBC cell line SUM 149 was less invasive and adhesive in vitro compared to the reportedly less aggressive breast cancer cell line SUM 102, seems to support an alternative mode for IBC dissemination from classic actin cytoskeleton- mediated cell motility.
The high expression levels of both E-cadherin and RhoA by SUM 149 may contribute to the uniquely invasive phenotype of IBC. However, signaling via E-cadherin to Rho is unclear with E-cadherin-mediated Rho activation and inhibition reported in a cell line specific manner [39]. Dominant negative RhoA expression in EL and nEαCL cells has been reported to reduce E-cadherin activity [40]. During the embryonic development of stratified epithelium, it was found that α-catenin, Rho, and Rho kinase were vital for coordinated tissue movement. In this sense, cells maintain tissue architecture via cadherin binding but move as a unit through actin reorganization mediated by Rho and its downstream effector Rho kinase [41]. A parallel argument could be made for the dissemination of IBC, in which tightly bound tumor cells move as a coordinated front. This possibility was tested in a wound healing assay, in which we found that the SUM 149 cells do not polarize or move into the wound after 7.5 hours, suggesting that this form of invasion is not the mechanism by which IBC is dissemination.
Conclusion
Thus, our results demonstrate that the IBC cell line SUM 149 is less invasive than a similar cell line, SUM 102, which expresses less Rho. This finding seems to support an alternate mode of dissemination for IBC than that of the classic invasive model, in which individual cells break local attachments and move through the ECM via actin cytoskeleton remodeling. A previously hypothesized mode of invasion for IBC, termed passive metastasis, would then seem the likely candidate (Figure 6A). In passive metastasis, vasculogenesis, stimulated by secreted differentiation factors, occurs around a tumor cell embolus that has maintained strong cell-cell attachments [16]. IBC is known to secrete angiogenic and possibly also vasculogenic growth factors, such as VEGF, bFGF, IL-6, and IL-8 [17]. Vasculogenic tubule formation by melanoma cells has been shown to be dependent on cadherin expression [42]. E-cadherin positive tumor cell emboli located within the dermal lymphatics are typically found in IBC histological specimens [15]. Three-dimensional culture of SUM 149 cells in matrigel results in IBC cell spheroids that are reminiscent of the tumor cell emboli seen in pathology (Figure 6B). The probability that IBC cells invade the circulatory system by a passive metastasis mechanism as tumor emboli rather than by active migratory mechanisms is being tested by in vivo image analysis of fluorescent protein tagged SUM 149 mammary tumors in SCID mice. This investigation could drastically change the course of IBC treatment and identify new therapeutic targets specific for this form of breast cancer.
Figure 6 6A. Model of passive metastasis. 6B. SUM 149 cell spheroids expressing RFP. 6A. Tumor cells secrete unknown differentiation factors, stimulating vasculogenesis and resulting in a cluster of tumor cells, termed embolus, located within the de novo formed vessel. The embolus maintains cell-cell attachments as it moves through the vessel and lodges within a dermal lymph node. 6B. 100× confocal projection image (10 images, Z = 20 μm) of SUM 149 cell spheroids stably expressing RFP in 3-dimensional matrigel culture after 5 days.
Methods
Cell Culture
The SUM cell lines used for the study have been recently developed from pleural effusions of breast cancer patients [18,19] and are generous gifts of Dr. Stephen Ethier, The University of Michigan, MI. SUM 149 is an IBC cell line that lacks expression of the gene LIBC and overexpresses RhoC [5]. SUM 102, developed from a minimally invasive human breast carcinoma [20] will be used as a model for non-IBC human breast cancer cells. SUM 149 cells were cultured in F-12 Hams (Gibco™, CA) supplemented with 5% fetal bovine serum (Tissue Culture Biologicals, CA), insulin, and hydrocortisone. SUM 102 cells were cultured in F-12 Hams (Gibco™, CA) supplemented with 5% bovine serum albumin (BSA), epidermal growth factor, T3, ethanolamine, and sodium selenite.
Adhesion Assays
Cell adhesion assays were performed according to [21]. Briefly, glass coverslips (Fisher Scientific, TX) were coated with 50 μg/ml laminin (Gibco BRL, MD), 10 μg/ml of collagen I (BD Biosciences, MA), 10 μg/ml of collagen IV (BD Biosciences, MA) and incubated overnight at 4°C. The coverslips were blocked for 1 hour with 1% heat-denatured BSA (Sigma Chemical Corporation, MO) in PBS. Cells (105) were placed on coverslips and allowed to adhere for 15 minutes. Non-adherent cells were removed by washing. The adherent cells were fixed in 3.7% formaldehyde (Sigma Chemical Corp., MO) and stained for F-actin as described below to aid in quantification. The number of cells per coverslip was quantified with a 40× phase contrast objective.
Haptotaxis Invasion Assay
Cell invasion assays were performed as described in [22]. Modified Boyden chambers (tissue culture treated, 6.5 mm diameter, 10 μm thickness, 8 μm pores, Transwell®, Costar Corp., Cambridge, MA) were coated on the upper surface (invasion), of the membrane with 50 μg/ml laminin, 10 μg/ml collagen I, or 10 μg/ml collagen IV overnight at 4°C and then placed into the lower chamber containing 500 μl culture media with 10% fetal bovine serum (FBS). Serum starved cells (105) were added to the upper surface of each migration chamber and allowed to migrate to the underside of the membrane for 24 hours (invasion). The non-migratory cells on the upper membrane surface were removed with a cotton swab, and the migratory cells attached to the bottom surface of the membrane stained with propidium iodide (CalBioChem-Novabiochem Corp., CA). The number of invasive cells per membrane was counted with an Olympus upright fluorescence microscope with a 40× objective.
Wound Healing Assay
Cells were first grown to a confluent monolayer, wounded with a sterile razor blade and allowed to migrate for 7.5 hours before fixing, permeabilizing, and blocking. Cells were then stained for F-actin as described below and visualized using an Olympus upright fluorescence microscope.
Immunofluorescence Microscopy
For focal adhesion and F-actin staining, cells were cultured on coverslips until they reached 60% confluency and starved for 24 hours in unsupplemented F-12 Hams. Cells were then stimulated with 50 ng/ml epidermal growth factor (EGF), 5% FBS, or PBS control for 10 minutes, fixed in 3.7% formaldehyde (Sigma Chemical Corp., MO), permeabilized with 0.2% Triton X-100 (Sigma, MO), and blocked with 5% goat serum (Gibco™, CA), and 5% BSA (Sigma Chemical Corp., MO) in PBS. Cells were stained with rhodamine phalloidin (Molecular Probes Inc., OR) to visualize F-actin, and a mouse monoclonal anti-phosphotyrosine antibody, clone 4G10 (Upstate Biotechnology, NY), followed by FITC-conjugated goat anti mouse IgG (ICN Biomedicals Inc., CA) to visualize the focal adhesions. Phosphotyrosine staining to is commonly utilized to visualize focal adhesions [23]. For E-cadherin staining, cells were cultured until 60% confluency, fixed in methanol at -20°C for 15 minutes, and blocked with 5% goat serum (Gibco™, CA) and 5% BSA (Sigma Chemical Corp., MO) in PBS. Cells were stained with a mouse monoclonal anti-E-cadherin antibody, clone G-10 (Santa Cruz Biotechnology, CA) followed by FITC-conjugated goat anti-mouse IgG (ICN Biomedicals Inc., CA). Cells were imaged using an Olympus upright fluorescence microscope with Spot Advanced digital camera software, Version 2.2.1 (Diagnostic Instruments Inc., MI).
Immunoblotting
Cells were cultured to confluency on 6 cm plates, trypsinized and the pellet washed in 1X PBS. The cell pellet was then lysed in 1% NP-40 lysis buffer. Equal amounts of protein, as determined by Bio-Rad (Hercules, CA) total protein assay, were then separated by 10% SDS-PAGE gel for Rho (A, B, and C), RhoA, and RhoC, or 8% SDS-PAGE gel for E-cadherin. Cellular proteins were then transferred to a nitrocellulose membrane. Membranes were blocked with 4% milk 0.05% Tween and probed with rabbit polyclonal anti-Rho (A, B, and C) (Upstate Biotechnology, NY), mouse monoclonal anti-RhoA (Santa Cruz Biotechnology, CA), goat polyclonal anti-RhoC (Santa Cruz Biotechnology, CA), or mouse monoclonal anti-E-cadherin (Santa Cruz Biotechnology, CA) followed by horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce Endogen, IL) for Rho or alkaline phosphatase conjugated goat anti mouse antibody for E-cadherin (Pierce Endogen, IL). Rho immunoblots were detected with the Super Signal West Femto-Substrate chemiluminescence kit (Pierce Endogen, IL) and Kodak Biomax MR film (Fisher Scientific, TX). E-Cadherin immunoblots were detected with NBT/BCIP alkaline phosphatase substrate (Pierce Endogen, IL).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MRH participated in design of the study, carried out the assays, including their quantification and interpretation, and was the primary author of the manuscript. KMW participated in the cell culture, assisted in the assays and their quantification and interpretation, and helped to draft and revise the manuscript. SFD was responsible for the conception of the project, advise and training on experimental design and procedures, aided in the analysis and interpretation of data, and helped revise the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Dr. Rebecca Richards-Kortum for critical input during the preparation of this manuscript. This investigation was supported by NIH/NCI CA83957-01A1 and University of Texas Biomedical Engineering Seed Grant to S.D., DOD/US Army BC031906 to M.H, and University of Texas Cooperative Society Awards to KW.
==== Refs
Levine PH Steinhorn SC Ries LG Aron JL Inflammatory breast cancer: the experience of the surveillance, epidemiology, and end results (SEER) program J Natl Cancer Inst 1985 74 291 297 3856043
Lopez MJ Porter KA Inflammatory breast cancer Surg Clin North Am 1996 76 411 429 8610272
Schmitz AAP Govek EE Ötner BB van Aelst L Rho GTPases: signaling, migration, and invasion Exp Cell Res 2000 261 1 12 11082269 10.1006/excr.2000.5049
Hall A Nobes CD Rho GTPases: molecular switches that control the organization and dynamics of the actin cytoskeleton Philos Trans R Soc Lond B Biol Sci 2000 355 965 970 11128990 10.1098/rstb.2000.0632
van Golen KL Davies S Wu ZF Wang YF Bucana CD Root H Chandrasekharappa S Strawderman M Ethier SP Merajver SD A novel putative low-affinity insulin-like growth factor-binding protein, LIBC (lost in inflammatory breast cancer), and RhoC GTPase correlate with the inflammatory breast cancer phenotype Clin Cancer Res 1999 5 2511 2519 10499627
Suwa H Ohshio G Imamura T Wantanabe 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
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
van Golen KL Bao L DiVito MM Wu Z Prendergrast GC Merajver SD Reversion of RhoC GTPase-induced inflammatory breast cancer phenotype by treatment with a farnesyl transferase inhibitor Mol Cancer Ther 2002 1 575 583 12479217
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
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
Tomlinson JS Alpaugh ML Barsky SH An intact overexpressed E-cadherin/α, β-Catenin axis characterizes the lymphovascular emboli of inflammatory breast carcinoma Cancer Res 2001 61 5231 5241 11431364
Alpaugh ML Tomlinson JS Kasraeian S Barsky SH Cooperative role of E-cadherin and sialyl-Lewis X/A-deficient MUC1 in the passive dissemination of tumor emboli in inflammatory breast carcinoma Oncogene 2002 21 3631 3643 12032865 10.1038/sj.onc.1205389
Alpaugh ML Tomlinson JS Shao ZM Barsky SH A novel human xenograft model of inflammatory breast cancer Cancer Res 1999 59 5079 5084 10537277
Colpaert CG Vermeulen PB Benoy I Soubry A van Roy F van Beest P Goovaerts G Dirix LY van Dam P Fox SB Harris AL van Marck EA Inflammatory breast cancer shows angiogenesis with endothelial proliferation rate and strong E-cadherin expression Br J Cancer 2003 10 718 725 12618881 10.1038/sj.bjc.6600807
Kleer CG van Golen KL Braun T Merajver SD Persistent E-cadherin expression in inflammatory breast cancer Mod Pathol 2001 14 458 464 11353057 10.1038/modpathol.3880334
Cavallaro U Christofori G Cell adhesion in tumor invasion and metastasis: loss of the glue is not enough Biochimica et Biophysica Acta 2001 1552 39 45 11781114 10.1016/S0304-419X(01)00038-5
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
Ethier SP Kokeny KE Ridings JW Dilts CA erbB family receptor expression and growth regulation in a newly isolated human breast cancer cell line Cancer Res 1996 56 899 907 8631031
Ethier SP Human breast cancer cell lines as models of growth regulation and disease progression J Mammary Gland Biol Neoplasia 1996 1 111 121 10887484
Sartor CI Role of epidermal growth factor receptor and STAT-3 activation in autonomous proliferation of SUM-102PT human breast cancer cells Cancer Res 1997 57 978 987 9041204
Maddox AS Oegema K Closing the GAP: A Role for RhoA GAP in Cytokinesis Mol Cell 2003 11 846 848 12718869 10.1016/S1097-2765(03)00151-5
Klemke RL Leng J Molander R Brooks PC Vuori K Cheresh DA CAS/Crk coupling serves as a "molecular switch" for induction of cell migration J Cell Biol 1998 140 961 72 9472046 10.1083/jcb.140.4.961
Schoenwaelder SM Burridge K Evidence for a calpeptin-sensitive protein-tyrosine phosphatase upstream of the small GTPase Rho. A novel role for the calpain inhibitor calpeptin in the inhibition of protein-tyrosine phosphatases J Biol Chem 1999 274 14359 67 10318859 10.1074/jbc.274.20.14359
Sward K Mita M Wilson DP Deng JT Susnjar M Walsh MP The role of RhoA and Rho-associated kinase in vascular smooth muscle contraction Curr Hypertens Rep 2003 5 66 72 12530938
Shirakawa K Wakasugi H Heike Y Watanabe I Yamada S Saito K Konishi F Vasculogenic mimcry and psuedo-comedo formation in breast cancer Int J Cancer 2002 99 821 828 12115483 10.1002/ijc.10423
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
Shirakawa K Tsuda H Heike Y Kato K Asada R Inomata M Sasaki H Kasumi F Yoshimoto M Iwanaga T Konishi F Terada M Wakasugi H Absence of endothelial cells, central necrosis, and fibrosis are associated with aggressive inflammatory breast cancer Cancer Res 2001 61 445 451 11212228
Nobes CD Hall A Rho GTPases control polarity, protrusion, and adhesion during cell movement J Cell Biol 1999 144 1235 1244 10087266 10.1083/jcb.144.6.1235
Okegawa T Li Y Pong RC Hsieh JT Cell adhesion proteins as tumor suppressors J Urol 2002 167 1836 43 11912444 10.1097/00005392-200204000-00091
Ridley AJ Rho GTPases and cell migration J Cell Sci 2001 114 2713 2722 11683406
Rottner K Hall A Small JV Interplay between Rac and Rho in the control of substrate contact dynamics Curr Biol 1999 9 640 648 10375527 10.1016/S0960-9822(99)80286-3
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 41 12855641
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 70 12808121
Clark EA Golub TR Lander ES Hynes RO Genomic analysis of metastasis reveals an essential role for RhoC Nature 2000 406 532 5 10952316 10.1038/35020106
Stam JC Michiels F van der Kammen RA Moolenaar WH Collard JG Invasion of T-lymphoma cells: cooperation between Rho family GTPases and lysophospholipid receptor signaling EMBO J 1998 17 4066 4074 9670021 10.1093/emboj/17.14.4066
Itoh K Yoshioka K Akedo H Uehata M Ishizaki T Narumiya S An essential part for Rho-associated kinase in the transcellular invasion of tumor cells Nat Med 1999 5 221 225 9930872 10.1038/5587
O'Connor KL Nguyen BK Mercurio AM RhoA function in lamellae formation and migration is regulated by the alpha6beta4 integrin and cAMP metabolism J Cell Biol 2000 148 253 258 10648558 10.1083/jcb.148.2.253
Moorman JP Luu D Wickham J Bobak DA Hahn CS A balance of signaling by Rho family small GTPases RhoA, Rac1, and Cdc42 coordinates cytoskeletal morphology but not cell survival Oncogene 1999 18 47 57 9926919 10.1038/sj.onc.1202262
Braga VMM Cell-cell adhesion and signalling Curr Opin Cell Biol 2002 14 546 556 12231348 10.1016/S0955-0674(02)00373-3
Fukata M Kaibuchi K Rho-family GTPases in cadherin-mediated cell-cell adhesion Nat Rev Mol Cell Biol 2001 2 887 897 11733768 10.1038/35103068
Vaezi A Bauer C Vasioukhin V Fuchs E Actin cable dynamics and Rho/Rock orchestrate a polarized cytoskeletal architecture in the early steps of assembling a stratified epithelium Dev Cell 2002 3 367 381 12361600 10.1016/S1534-5807(02)00259-9
Hendrix MJC Softer EA Meltzer PS Gardner LMG Hess AR Kirschmann DA Schatteman GC Seftor REB Expression and functional significance of VE-cadherin in aggressive human melanoma cells: Role in vasculogenic mimicry Proc Nat Acd Sci USA 2001 98 8018 8023 10.1073/pnas.131209798
| 15857504 | PMC1090601 | CC BY | 2021-01-04 16:40:10 | no | Cancer Cell Int. 2005 Apr 27; 5:11 | utf-8 | Cancer Cell Int | 2,005 | 10.1186/1475-2867-5-11 | oa_comm |
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-111585750410.1186/1475-2867-5-11Primary ResearchIn vitro analysis of the invasive phenotype of SUM 149, an inflammatory breast cancer cell line Hoffmeyer Michaela R [email protected] Kristin M [email protected] Suranganie F [email protected] Department of Molecular Cell and Developmental Biology, University of Texas Austin, 205 West 24th Street, Austin, Texas, 78712, USA2 Department of Biomedical Engineering, University of Texas at Austin, 1 University Station Stop C0800, Austin, TX 78712, USA2005 27 4 2005 5 11 11 15 11 2004 27 4 2005 Copyright © 2005 Hoffmeyer et al; licensee BioMed Central Ltd.2005Hoffmeyer 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
Inflammatory breast cancer (IBC) is the most lethal form of locally invasive breast cancer known. However, very little information is available on the cellular mechanisms responsible for manifestation of the IBC phenotype. To understand the unique phenotype of IBC, we compared the motile and adhesive interactions of an IBC cell line, SUM 149, to the non-IBC cell line SUM 102.
Results
Our results demonstrate that both IBC and non-IBC cell lines exhibit similar adhesive properties to basal lamina, but SUM 149 showed a marked increase in adhesion to collagen I. In vitro haptotaxis assays demonstrate that SUM 149 was less invasive, while wound healing assays show a less in vitro migratory phenotype for SUM 149 cells relative to SUM 102 cells. We also demonstrate a role for Rho and E-cadherin in the unique invasive phenotype of IBC. Immunoblotting reveals higher E-cadherin and RhoA expression in the IBC cell line but similar RhoC expression. Rhodamine phalloidin staining demonstrates increased formation of actin stress fibers and larger focal adhesions in SUM 149 relative to the SUM 102 cell line.
Conclusion
The observed unique actin and cellular architecture as well as the invasive and adhesive responses to the extracellular matrix of SUM 149 IBC cells suggest that the preference of IBC cells for connective tissue, possibly a mediator important for the vasculogenic mimicry via tubulogenesis seen in IBC pathological specimens. Overexpression of E-cadherin and RhoA may contribute to passive dissemination of IBC by promoting cell-cell adhesion and actin cytoskeletal structures that maintain tissue integrity. Therefore, we believe that these findings indicate a passive metastatic mechanism by which IBC cells invade the circulatory system as tumor emboli rather than by active migratory mechanisms.
==== Body
Background
With an average five-year post-recovery survival rate of 45%, inflammatory breast cancer (IBC) is the most lethal and aggressive form of locally advanced breast cancer [1]. The lethality of IBC stems from its highly invasive nature. Diagnosis of IBC is often complicated by lack of a palpable precursor lesion commonly associated with breast cancer. Moreover, the correct diagnosis is hindered by inflammatory-like symptoms such as redness, warmth, and edema. Characteristic of IBC is a change in breast skin texture, similar to that of an orange, due to extensive invasion of the dermal lymphatics by IBC tumor cell emboli. These complications contribute to IBC lethality in that by the time a proper diagnosis is made, the cancer has aggressively infiltrated the surrounding tissue and lymphatics system, leading to a lowered patient prognosis [2]. Complicating treatment of this deadly form of breast cancer is that very little information about the cellular mechanisms responsible for the unique IBC phenotype is known.
Cancer cell invasion through the basal lamina and subsequent metastasis involves multiple steps including intravasation through the surrounding tissue into the lymphatic or vascular systems. Transient adhesion to extracellular matrix (ECM) components as well as modification of cell shape by reorganization of the actin cytoskeleton is required for cancer cell infiltration into the adjacent tissue. The Rho GTPases regulate actin cytoskeletal rearrangements, and are thus likely candidates for involvement in cancer cell invasion and metastasis [3,4]. Further evidence for a relationship between cancer cell mobilization and dysregulation of Rho GTPases is seen in the overexpression of Rho proteins in numerous invasive human cancers. The recent discovery of the overexpression of the Rho isoform RhoC by IBC tumors has been implicated in the physiological mechanisms of this poorly characterized form of breast cancer [5]. RhoC was demonstrated to be overexpressed in metastatic tumors of pancreatic adenocarcinoma patients [6], murine melanomas [7], and in the patient-derived IBC cell line SUM 149 [5]. Transient inhibition of RhoC in IBC cells by treatment with farnesyl transferase inhibitors reduced invasion and motility in vitro [8]. Recently it was reported that RhoC overexpression in mammary epithelial cells resulted in a significant increase in cell migration [9], mediated by the MAPK pathway [10]. These findings led us to hypothesize that RhoC overexpression may promote the highly invasive phenotype of IBC and contribute to the uniquely aggressive phenotype exhibited by IBC.
Another unique feature of IBC is the overexpression of E-cadherin, a transmembrane protein involved in cell-cell adhesion, which is generally lost in highly invasive cancers. It seems somewhat paradoxical that such an aggressive cancer that overexpresses proteins involved in actin cytoskeleton rearrangement and promotion of migration (i.e., RhoC) also overexpresses cell-cell junction proteins such as E-cadherin [11-15]. The literature thus far seems to hold to two schools of thought about the contradictory protein expression seen in IBC. One tends to support the idea that E-cadherin expression fluctuates with disease progression and decreases as IBC cells become invasive [15]. The second school supports the theory of passive metastasis by IBC [11,12]. In passive metastasis, strong tumor cell-cell adhesions are maintained during dissemination that proceeds via vasculogenesis through secretion of differentiation factors by the tumor cells causing de novo vessel formation [16]. This results in a cancer cell cluster within the vessel, reminiscent of the IBC tumor emboli seen in IBC histology. Furthermore, RhoC overexpression in human mammary epithelial cells has been shown to increase production of angiogenic factors, some of which might mediate passive or active metastasis [17].
The IBC phenotype has mystified clinicians due to its inflammatory-like symptoms. However IBC symptomology is not considered to be a true immunoreaction, but rather a consequence of cancer cell invasion to the lymphatics system. The mechanism by which IBC invades is unclear and further experimentation with IBC models is required to clarify the exact mechanism by which this form of breast cancer is disseminated. Using the SUM 149 IBC cell line, we have examined the adhesive and migratory capacities in an effort to understand the invasive behavior of IBC for future experimentation with in situ imaging of IBC in animal models. SUM 149 was compared to a control cell line, SUM 102, which was selected because it shares a deletion in the LIBC (lost in inflammatory breast cancer) gene with the SUM 149 cell line but reportedly expresses RhoC mRNA at low levels [5]. We show that SUM 149 is less invasive and adhesive to basal lamina components in vitro than SUM 102, and that SUM 149 expresses more Rho proteins and E-cadherin. These data shows that SUM 149 is not highly motile and therefore possibly not actively invasive, suggesting passive metastasis as the mechanism of IBC dissemination.
Results
Endogenous Levels of Rho
Figure 1A presents the relative protein levels of the various Rho isoforms in the IBC cell line SUM 149 versus the non-IBC cell line SUM 102. Previous investigators have reported overexpression of RhoC mRNA in IBC cells compared to SUM 102 [5]. To verify overexpression of RhoC at the protein level, we preformed Western blots on cell lysates using a RhoC polyclonal antibody (Santa Cruz Biotechnology, CA). We found no significant difference in RhoC protein levels between the SUM 149 and the SUM 102 cell lines. However, immunoblotting revealed a significant difference in Rho (A, B, and C), with the IBC cell line expressing much higher Rho protein levels. We then examined RhoA protein levels and found a significant overexpression of RhoA in the IBC cell line. This finding is interesting considering that RhoA has been shown to play a vital role in actomyosin-mediated contractility [21,24].
Figure 1 1A. Rho GTPase protein expression levels in the IBC cell line SUM 149 versus SUM 102. 1B. F-actin and focal adhesion distribution in SUM 149 and SUM 102 human breast cancer cell lines. 1A. Equal protein amounts were separated by 10% SDS-PAGE, transferred to nitrocellulose, and probed for RhoA, RhoC, and Rho (A, B, and C). Images are representative of at least three independent experiments with actin serving to verify equal protein loading. 1B. SUM 149 and SUM 102 cells were starved in unsupplemented F-12 Hams media for 24 hours and stimulated for 10 minutes with PBS (control), EGF (50 ng/ml), or FBS (5%). Cells were stained with rhodamine phalloidin (red) to visualize F-actin and anti-phosphotyrosine (green) to visualize focal adhesion. Micrographs were taken at 1000× magnification. Images are representative for at least three independent experiments.
Subcellular Distribution of Focal Adhesions and Filamentous Actin
Because RhoA is involved in actin stress fiber and focal adhesion formation, we stained the cells with rhodamine phalloidin to visualize F-actin and anti-phosphotyrosine to visualize focal adhesions. Figure 1B demonstrates F-actin and focal adhesion distribution in both cell lines. SUM 149 displayed larger focal adhesions and more actin stress fibers than the SUM 102 cell line, as might be expected from the high levels of RhoA in the SUM 149 cell line. Upon stimulation of quiescent cells with EGF or FBS, the SUM 102 cells formed large membrane ruffles (lamellipodia). However, stimulation by both EGF and FBS seemed to have little effect on the actin cytoskeleton of the SUM 149 cells. An increase in focal adhesion was seen in the SUM 149 cells after stimulation with EGF, but no clear cell polarization was observed.
Adhesion to Extracellular Matrix Proteins
Invasion and metastasis is a multi-step process in which cells must break local connections, move through the basal lamina, survive in circulation, and reestablish cellular attachment at distant sites. Clearly, many of these steps involve interaction with ECM components. Transient adhesion to the ECM in conjunction with cytoskeletal rearrangements are requirements for cell motility. To examine the ability of the breast cancer cell lines under study to adhere to the various ECM proteins, we performed adhesion assays (Figure 2). Here, we demonstrate that both SUM 149 and SUM 102 cells have similar adhesive properties on laminin, the major component of the basal lamina. A slight increase in adhesive properties for the IBC cells was observed compared to the SUM 102 cells on collagen IV. However, a marked increase in adhesion to collagen I, the major component of connective tissue, was seen for the SUM 149 cell line. Taken together, this data suggest that the exacerbated invasive phenotype seen in IBC is not due to differences in adhesive properties to basal lamina components, but may indicate a preference of IBC cells to the connective tissue, through which these cells must invade before entering circulation. Furthermore, attachment to connective tissue components maybe important for vasculogenic mimicry via tubulogenesis, as seen in IBC pathological specimens [25-27].
Figure 2 Adhesion of human breast cancer cells to extracellular matrix proteins. SUM 149 and SUM 102 cells (105) were plated on glass coverslips coated with laminin (50 μg/ml), collagen I (10 μg/ml), or collagen IV (10 μg/ml) and allowed to adhere for 15 minutes. Micrographs were taken at 400× magnification. Adherent cells were quantified in 10 random microscopic fields. Data are expressed as mean ± SEM of at least three independent experiments.
Haptotaxis Stimulated Invasion
An aggressively infiltrative cancer must invade surrounding tissue by movement through the ECM. Haptotaxis, or cell movement toward ECM proteins, was assayed in vitro and is presented in Figure 3. SUM 149 cells were significantly less invasive into laminin (basal lamina component) and collagen I (connective tissue component) after 24 hours than the SUM 102 cell line. Therefore, SUM 149 was less invasive when assayed in a manner that requires individual cell movement by an active motile mechanism through a membrane with 8 μm pores.
Figure 3 Haptotaxis of human breast cancer cells to extracellular matrix proteins. SUM 149 and SUM 102 cells (105) were placed into the top chamber of a Costar well coated with laminin (50 μg/ml) or collagen I (10 μg/ml) and allowed to invade for 24 hours. Invasive cells were stained with propidium iodide and quantified in 10 random microscopic fields. Micrographs were taken at 200× magnification. Data are expressed as mean ± SEM of at least three independent experiments.
Subcellular Distribution of Filamentous Actin Subsequent to Cellular Polarization
To induce cell polarization and migration that does not involve active migration of individual cells but rather collective cell migration over a wound edge, we performed wound healing assays as described in [28]. A confluent monolayer of cells was wounded, leading to the release of chemotractant signals by the cells at the wound edge, thus mimicking cell motility cues in vivo (Figure 4). Here, we present that the IBC cell line SUM 149 was less responsive to the cell-derived migration signals after 7.5 hours. By this time, the SUM 102 cell line was much more invasive into the wound space and had nearly closed the wound entirely. Thus, active migration as a sheet of cells is not likely the mechanism by which IBC is disseminated.
Figure 4 Human breast cancer cell migration in response to wounding. SUM 149 and SUM 102 cells were grown to confluency on glass coverslips and wounded with a sterile razor blade. Closure of the wound was monitored over 7.5 hours. Cells were stained with rhodamine phalloidin to visualize F-actin reorganization in response to cell migration. Arrows indicate the wound edge. Micrographs were taken at 1000× magnification. Images are representative of at least three independent experiments.
Endogenous Expression of E-cadherin
An interesting and perplexing characteristic of IBC is the expression of E-cadherin by this invasive form of breast cancer. Usually the loss of E-cadherin correlates with increased invasive and metastatic potential [29]. To verify E-cadherin expression in the IBC cell line SUM 149, we performed immunofluorescence experiments (Figure 5A). Both SUM 149 and SUM 102 show E-cadherin staining localized to the shared margins between neighboring cells, however this staining is much more intense in the SUM 149 cell line. Immunoblotting analysis likewise demonstrates E-cadherin expression by both cell lines with higher levels of E-cadherin expressed in the SUM 149 cells (Figure 5B).
Figure 5 5A. E-cadherin distribution in SUM 149 and SUM 102 human breast cancer cells. 5B. E-cadherin protein expression levels in the IBC cell lineSUM 149 versus SUM 102. 5A. Cells were grown to 60% confluency and stained with anti-E-cadherin (green). Arrows indicate cell-cell adhesions containing E-cadherin. Micrographs were taken at 1000× magnification. Images are representative for at least three independent experiments. 5B. Cell lysates were separated by 8% SDS-PAGE, transferred to nitrocellulose, and probed for E-cadherin. Images are representative of at least three independent experiments with actin serving to verify equal protein loading.
Discussion
IBC is a unique and highly aggressive form of locally advanced breast cancer with distinct clinical presentation. We hypothesized that upregulated expression of RhoC, as reported by others to be characteristic of IBC, contributes to the unusual pathological presentation of IBC. For the first time, we have compared the actin architecture, invasive, and adhesive properties of the IBC cell line SUM 149 with a cell line reported to express less RhoC mRNA compared to SUM 149 but share a deletion in LIBC [5]. Using a commercially available specific antibody to RhoC, we report that RhoC is not overexpressed at the protein level by the IBC cell line SUM 149. Interestingly, we confirmed overexpression of RhoA by utilizing an anti-RhoA specific antibody. However, post-transcriptional regulation of RhoC expression may account for the observed discrepancy. It is possible that our results do not agree with the reported mRNA expression due to specificity problems with the commercially developed antibodies. Furthermore, we demonstrate that, compared to SUM 102, SUM 149 is less invasive and migratory, and displays impaired adhesion to basal lamina components but strong adhesion to connective tissue proteins.
The role of the Rho protein in cancer cell invasion is somewhat controversial. RhoA is known to be involved in cell contractility, both in the formation of bundled actin fibers and through the activation of Rho kinase and subsequent activation of myosin light chain [30]. Such contractile cells have previously been shown to be less motile [31]. However, Rho overexpression has been documented in various human cancers such as bladder and ovarian, and correlates with lymph node invasion, metastasis, and poor patient prognosis [32,33]. Overexpression of RhoC by human mammary epithelial cells increased invasion, motility, and anchorage independent growth, similar to SUM 149 [9]. Expression of dominant negative Rho T19N has been demonstrated to block melanoma cell invasion [34]. Some investigators report that Rho overexpression has little impact on invasion and cell motility, while others demonstrate a positive correlation between Rho expression and cell migration capacity [35-37]. Rho is required for cell body contraction and tail retraction during directed cell motility, while active Rac and Cdc42 are required for lamellipodia and filopodia extension at the leading edge [30]. Thus, invasive potential is considered to be a balance between Rac, Cdc42, and Rho activities. Overexpression or activation of one of these Rho GTPases will shift this balance and result in a cellular phenotype dominated by the actin structure promoted by the activated Rho GTPase [38]. SUM 149 may display reduced invasion and migration in vitro compared to SUM 102 due to the overexpression of RhoA alone, thus masking the motile effects of Rac and Cdc42.
Another aspect that makes IBC so remarkable is that this form of aggressive breast cancer maintains strong E-cadherin expression [11-15]. Typically, loss of E-cadherin expression correlates with progression to metastatic disease since cancer cells must break inter-cell adhesions before attaining a motile phenotype [29]. Here, we demonstrate that the SUM 149 model of IBC maintains strong E-cadherin expression in culture, as seen in other IBC xenograft models and IBC pathological specimens. Previous reports indicate the E-cadherin axis is also complete and functional [11]. IBC histology reveals an extensive invasion of E-cadherin positive tumor cell emboli within the dermal lymphatics [11-15]. The expression of E-cadherin may be critical for invasion in that IBC is thought by some to be passively disseminated, an invasion mechanism that necessitates cell-cell attachment [12]. In this scenario, tumor cells maintain strong cell-cell connections and enter circulation via vasculogenesis around a tumor cell embolus. Others hold that E-cadherin expression varies with the malignant stage of the disease, and is lost during invasion but reestablished once tumor cells invade the vasculature [15]. The finding reported here, in which the IBC cell line SUM 149 was less invasive and adhesive in vitro compared to the reportedly less aggressive breast cancer cell line SUM 102, seems to support an alternative mode for IBC dissemination from classic actin cytoskeleton- mediated cell motility.
The high expression levels of both E-cadherin and RhoA by SUM 149 may contribute to the uniquely invasive phenotype of IBC. However, signaling via E-cadherin to Rho is unclear with E-cadherin-mediated Rho activation and inhibition reported in a cell line specific manner [39]. Dominant negative RhoA expression in EL and nEαCL cells has been reported to reduce E-cadherin activity [40]. During the embryonic development of stratified epithelium, it was found that α-catenin, Rho, and Rho kinase were vital for coordinated tissue movement. In this sense, cells maintain tissue architecture via cadherin binding but move as a unit through actin reorganization mediated by Rho and its downstream effector Rho kinase [41]. A parallel argument could be made for the dissemination of IBC, in which tightly bound tumor cells move as a coordinated front. This possibility was tested in a wound healing assay, in which we found that the SUM 149 cells do not polarize or move into the wound after 7.5 hours, suggesting that this form of invasion is not the mechanism by which IBC is dissemination.
Conclusion
Thus, our results demonstrate that the IBC cell line SUM 149 is less invasive than a similar cell line, SUM 102, which expresses less Rho. This finding seems to support an alternate mode of dissemination for IBC than that of the classic invasive model, in which individual cells break local attachments and move through the ECM via actin cytoskeleton remodeling. A previously hypothesized mode of invasion for IBC, termed passive metastasis, would then seem the likely candidate (Figure 6A). In passive metastasis, vasculogenesis, stimulated by secreted differentiation factors, occurs around a tumor cell embolus that has maintained strong cell-cell attachments [16]. IBC is known to secrete angiogenic and possibly also vasculogenic growth factors, such as VEGF, bFGF, IL-6, and IL-8 [17]. Vasculogenic tubule formation by melanoma cells has been shown to be dependent on cadherin expression [42]. E-cadherin positive tumor cell emboli located within the dermal lymphatics are typically found in IBC histological specimens [15]. Three-dimensional culture of SUM 149 cells in matrigel results in IBC cell spheroids that are reminiscent of the tumor cell emboli seen in pathology (Figure 6B). The probability that IBC cells invade the circulatory system by a passive metastasis mechanism as tumor emboli rather than by active migratory mechanisms is being tested by in vivo image analysis of fluorescent protein tagged SUM 149 mammary tumors in SCID mice. This investigation could drastically change the course of IBC treatment and identify new therapeutic targets specific for this form of breast cancer.
Figure 6 6A. Model of passive metastasis. 6B. SUM 149 cell spheroids expressing RFP. 6A. Tumor cells secrete unknown differentiation factors, stimulating vasculogenesis and resulting in a cluster of tumor cells, termed embolus, located within the de novo formed vessel. The embolus maintains cell-cell attachments as it moves through the vessel and lodges within a dermal lymph node. 6B. 100× confocal projection image (10 images, Z = 20 μm) of SUM 149 cell spheroids stably expressing RFP in 3-dimensional matrigel culture after 5 days.
Methods
Cell Culture
The SUM cell lines used for the study have been recently developed from pleural effusions of breast cancer patients [18,19] and are generous gifts of Dr. Stephen Ethier, The University of Michigan, MI. SUM 149 is an IBC cell line that lacks expression of the gene LIBC and overexpresses RhoC [5]. SUM 102, developed from a minimally invasive human breast carcinoma [20] will be used as a model for non-IBC human breast cancer cells. SUM 149 cells were cultured in F-12 Hams (Gibco™, CA) supplemented with 5% fetal bovine serum (Tissue Culture Biologicals, CA), insulin, and hydrocortisone. SUM 102 cells were cultured in F-12 Hams (Gibco™, CA) supplemented with 5% bovine serum albumin (BSA), epidermal growth factor, T3, ethanolamine, and sodium selenite.
Adhesion Assays
Cell adhesion assays were performed according to [21]. Briefly, glass coverslips (Fisher Scientific, TX) were coated with 50 μg/ml laminin (Gibco BRL, MD), 10 μg/ml of collagen I (BD Biosciences, MA), 10 μg/ml of collagen IV (BD Biosciences, MA) and incubated overnight at 4°C. The coverslips were blocked for 1 hour with 1% heat-denatured BSA (Sigma Chemical Corporation, MO) in PBS. Cells (105) were placed on coverslips and allowed to adhere for 15 minutes. Non-adherent cells were removed by washing. The adherent cells were fixed in 3.7% formaldehyde (Sigma Chemical Corp., MO) and stained for F-actin as described below to aid in quantification. The number of cells per coverslip was quantified with a 40× phase contrast objective.
Haptotaxis Invasion Assay
Cell invasion assays were performed as described in [22]. Modified Boyden chambers (tissue culture treated, 6.5 mm diameter, 10 μm thickness, 8 μm pores, Transwell®, Costar Corp., Cambridge, MA) were coated on the upper surface (invasion), of the membrane with 50 μg/ml laminin, 10 μg/ml collagen I, or 10 μg/ml collagen IV overnight at 4°C and then placed into the lower chamber containing 500 μl culture media with 10% fetal bovine serum (FBS). Serum starved cells (105) were added to the upper surface of each migration chamber and allowed to migrate to the underside of the membrane for 24 hours (invasion). The non-migratory cells on the upper membrane surface were removed with a cotton swab, and the migratory cells attached to the bottom surface of the membrane stained with propidium iodide (CalBioChem-Novabiochem Corp., CA). The number of invasive cells per membrane was counted with an Olympus upright fluorescence microscope with a 40× objective.
Wound Healing Assay
Cells were first grown to a confluent monolayer, wounded with a sterile razor blade and allowed to migrate for 7.5 hours before fixing, permeabilizing, and blocking. Cells were then stained for F-actin as described below and visualized using an Olympus upright fluorescence microscope.
Immunofluorescence Microscopy
For focal adhesion and F-actin staining, cells were cultured on coverslips until they reached 60% confluency and starved for 24 hours in unsupplemented F-12 Hams. Cells were then stimulated with 50 ng/ml epidermal growth factor (EGF), 5% FBS, or PBS control for 10 minutes, fixed in 3.7% formaldehyde (Sigma Chemical Corp., MO), permeabilized with 0.2% Triton X-100 (Sigma, MO), and blocked with 5% goat serum (Gibco™, CA), and 5% BSA (Sigma Chemical Corp., MO) in PBS. Cells were stained with rhodamine phalloidin (Molecular Probes Inc., OR) to visualize F-actin, and a mouse monoclonal anti-phosphotyrosine antibody, clone 4G10 (Upstate Biotechnology, NY), followed by FITC-conjugated goat anti mouse IgG (ICN Biomedicals Inc., CA) to visualize the focal adhesions. Phosphotyrosine staining to is commonly utilized to visualize focal adhesions [23]. For E-cadherin staining, cells were cultured until 60% confluency, fixed in methanol at -20°C for 15 minutes, and blocked with 5% goat serum (Gibco™, CA) and 5% BSA (Sigma Chemical Corp., MO) in PBS. Cells were stained with a mouse monoclonal anti-E-cadherin antibody, clone G-10 (Santa Cruz Biotechnology, CA) followed by FITC-conjugated goat anti-mouse IgG (ICN Biomedicals Inc., CA). Cells were imaged using an Olympus upright fluorescence microscope with Spot Advanced digital camera software, Version 2.2.1 (Diagnostic Instruments Inc., MI).
Immunoblotting
Cells were cultured to confluency on 6 cm plates, trypsinized and the pellet washed in 1X PBS. The cell pellet was then lysed in 1% NP-40 lysis buffer. Equal amounts of protein, as determined by Bio-Rad (Hercules, CA) total protein assay, were then separated by 10% SDS-PAGE gel for Rho (A, B, and C), RhoA, and RhoC, or 8% SDS-PAGE gel for E-cadherin. Cellular proteins were then transferred to a nitrocellulose membrane. Membranes were blocked with 4% milk 0.05% Tween and probed with rabbit polyclonal anti-Rho (A, B, and C) (Upstate Biotechnology, NY), mouse monoclonal anti-RhoA (Santa Cruz Biotechnology, CA), goat polyclonal anti-RhoC (Santa Cruz Biotechnology, CA), or mouse monoclonal anti-E-cadherin (Santa Cruz Biotechnology, CA) followed by horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce Endogen, IL) for Rho or alkaline phosphatase conjugated goat anti mouse antibody for E-cadherin (Pierce Endogen, IL). Rho immunoblots were detected with the Super Signal West Femto-Substrate chemiluminescence kit (Pierce Endogen, IL) and Kodak Biomax MR film (Fisher Scientific, TX). E-Cadherin immunoblots were detected with NBT/BCIP alkaline phosphatase substrate (Pierce Endogen, IL).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MRH participated in design of the study, carried out the assays, including their quantification and interpretation, and was the primary author of the manuscript. KMW participated in the cell culture, assisted in the assays and their quantification and interpretation, and helped to draft and revise the manuscript. SFD was responsible for the conception of the project, advise and training on experimental design and procedures, aided in the analysis and interpretation of data, and helped revise the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Dr. Rebecca Richards-Kortum for critical input during the preparation of this manuscript. This investigation was supported by NIH/NCI CA83957-01A1 and University of Texas Biomedical Engineering Seed Grant to S.D., DOD/US Army BC031906 to M.H, and University of Texas Cooperative Society Awards to KW.
==== Refs
Levine PH Steinhorn SC Ries LG Aron JL Inflammatory breast cancer: the experience of the surveillance, epidemiology, and end results (SEER) program J Natl Cancer Inst 1985 74 291 297 3856043
Lopez MJ Porter KA Inflammatory breast cancer Surg Clin North Am 1996 76 411 429 8610272
Schmitz AAP Govek EE Ötner BB van Aelst L Rho GTPases: signaling, migration, and invasion Exp Cell Res 2000 261 1 12 11082269 10.1006/excr.2000.5049
Hall A Nobes CD Rho GTPases: molecular switches that control the organization and dynamics of the actin cytoskeleton Philos Trans R Soc Lond B Biol Sci 2000 355 965 970 11128990 10.1098/rstb.2000.0632
van Golen KL Davies S Wu ZF Wang YF Bucana CD Root H Chandrasekharappa S Strawderman M Ethier SP Merajver SD A novel putative low-affinity insulin-like growth factor-binding protein, LIBC (lost in inflammatory breast cancer), and RhoC GTPase correlate with the inflammatory breast cancer phenotype Clin Cancer Res 1999 5 2511 2519 10499627
Suwa H Ohshio G Imamura T Wantanabe 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
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
van Golen KL Bao L DiVito MM Wu Z Prendergrast GC Merajver SD Reversion of RhoC GTPase-induced inflammatory breast cancer phenotype by treatment with a farnesyl transferase inhibitor Mol Cancer Ther 2002 1 575 583 12479217
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
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
Tomlinson JS Alpaugh ML Barsky SH An intact overexpressed E-cadherin/α, β-Catenin axis characterizes the lymphovascular emboli of inflammatory breast carcinoma Cancer Res 2001 61 5231 5241 11431364
Alpaugh ML Tomlinson JS Kasraeian S Barsky SH Cooperative role of E-cadherin and sialyl-Lewis X/A-deficient MUC1 in the passive dissemination of tumor emboli in inflammatory breast carcinoma Oncogene 2002 21 3631 3643 12032865 10.1038/sj.onc.1205389
Alpaugh ML Tomlinson JS Shao ZM Barsky SH A novel human xenograft model of inflammatory breast cancer Cancer Res 1999 59 5079 5084 10537277
Colpaert CG Vermeulen PB Benoy I Soubry A van Roy F van Beest P Goovaerts G Dirix LY van Dam P Fox SB Harris AL van Marck EA Inflammatory breast cancer shows angiogenesis with endothelial proliferation rate and strong E-cadherin expression Br J Cancer 2003 10 718 725 12618881 10.1038/sj.bjc.6600807
Kleer CG van Golen KL Braun T Merajver SD Persistent E-cadherin expression in inflammatory breast cancer Mod Pathol 2001 14 458 464 11353057 10.1038/modpathol.3880334
Cavallaro U Christofori G Cell adhesion in tumor invasion and metastasis: loss of the glue is not enough Biochimica et Biophysica Acta 2001 1552 39 45 11781114 10.1016/S0304-419X(01)00038-5
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
Ethier SP Kokeny KE Ridings JW Dilts CA erbB family receptor expression and growth regulation in a newly isolated human breast cancer cell line Cancer Res 1996 56 899 907 8631031
Ethier SP Human breast cancer cell lines as models of growth regulation and disease progression J Mammary Gland Biol Neoplasia 1996 1 111 121 10887484
Sartor CI Role of epidermal growth factor receptor and STAT-3 activation in autonomous proliferation of SUM-102PT human breast cancer cells Cancer Res 1997 57 978 987 9041204
Maddox AS Oegema K Closing the GAP: A Role for RhoA GAP in Cytokinesis Mol Cell 2003 11 846 848 12718869 10.1016/S1097-2765(03)00151-5
Klemke RL Leng J Molander R Brooks PC Vuori K Cheresh DA CAS/Crk coupling serves as a "molecular switch" for induction of cell migration J Cell Biol 1998 140 961 72 9472046 10.1083/jcb.140.4.961
Schoenwaelder SM Burridge K Evidence for a calpeptin-sensitive protein-tyrosine phosphatase upstream of the small GTPase Rho. A novel role for the calpain inhibitor calpeptin in the inhibition of protein-tyrosine phosphatases J Biol Chem 1999 274 14359 67 10318859 10.1074/jbc.274.20.14359
Sward K Mita M Wilson DP Deng JT Susnjar M Walsh MP The role of RhoA and Rho-associated kinase in vascular smooth muscle contraction Curr Hypertens Rep 2003 5 66 72 12530938
Shirakawa K Wakasugi H Heike Y Watanabe I Yamada S Saito K Konishi F Vasculogenic mimcry and psuedo-comedo formation in breast cancer Int J Cancer 2002 99 821 828 12115483 10.1002/ijc.10423
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
Shirakawa K Tsuda H Heike Y Kato K Asada R Inomata M Sasaki H Kasumi F Yoshimoto M Iwanaga T Konishi F Terada M Wakasugi H Absence of endothelial cells, central necrosis, and fibrosis are associated with aggressive inflammatory breast cancer Cancer Res 2001 61 445 451 11212228
Nobes CD Hall A Rho GTPases control polarity, protrusion, and adhesion during cell movement J Cell Biol 1999 144 1235 1244 10087266 10.1083/jcb.144.6.1235
Okegawa T Li Y Pong RC Hsieh JT Cell adhesion proteins as tumor suppressors J Urol 2002 167 1836 43 11912444 10.1097/00005392-200204000-00091
Ridley AJ Rho GTPases and cell migration J Cell Sci 2001 114 2713 2722 11683406
Rottner K Hall A Small JV Interplay between Rac and Rho in the control of substrate contact dynamics Curr Biol 1999 9 640 648 10375527 10.1016/S0960-9822(99)80286-3
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 41 12855641
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 70 12808121
Clark EA Golub TR Lander ES Hynes RO Genomic analysis of metastasis reveals an essential role for RhoC Nature 2000 406 532 5 10952316 10.1038/35020106
Stam JC Michiels F van der Kammen RA Moolenaar WH Collard JG Invasion of T-lymphoma cells: cooperation between Rho family GTPases and lysophospholipid receptor signaling EMBO J 1998 17 4066 4074 9670021 10.1093/emboj/17.14.4066
Itoh K Yoshioka K Akedo H Uehata M Ishizaki T Narumiya S An essential part for Rho-associated kinase in the transcellular invasion of tumor cells Nat Med 1999 5 221 225 9930872 10.1038/5587
O'Connor KL Nguyen BK Mercurio AM RhoA function in lamellae formation and migration is regulated by the alpha6beta4 integrin and cAMP metabolism J Cell Biol 2000 148 253 258 10648558 10.1083/jcb.148.2.253
Moorman JP Luu D Wickham J Bobak DA Hahn CS A balance of signaling by Rho family small GTPases RhoA, Rac1, and Cdc42 coordinates cytoskeletal morphology but not cell survival Oncogene 1999 18 47 57 9926919 10.1038/sj.onc.1202262
Braga VMM Cell-cell adhesion and signalling Curr Opin Cell Biol 2002 14 546 556 12231348 10.1016/S0955-0674(02)00373-3
Fukata M Kaibuchi K Rho-family GTPases in cadherin-mediated cell-cell adhesion Nat Rev Mol Cell Biol 2001 2 887 897 11733768 10.1038/35103068
Vaezi A Bauer C Vasioukhin V Fuchs E Actin cable dynamics and Rho/Rock orchestrate a polarized cytoskeletal architecture in the early steps of assembling a stratified epithelium Dev Cell 2002 3 367 381 12361600 10.1016/S1534-5807(02)00259-9
Hendrix MJC Softer EA Meltzer PS Gardner LMG Hess AR Kirschmann DA Schatteman GC Seftor REB Expression and functional significance of VE-cadherin in aggressive human melanoma cells: Role in vasculogenic mimicry Proc Nat Acd Sci USA 2001 98 8018 8023 10.1073/pnas.131209798
| 15854226 | PMC1090602 | CC BY | 2021-01-04 16:40:10 | no | Cancer Cell Int. 2005 Apr 26; 5:9 | latin-1 | Cancer Cell Int | 2,005 | 10.1186/1475-2867-5-9 | oa_comm |
==== Front
Filaria JFilaria Journal1475-2883BioMed Central London 1475-2883-4-21581712410.1186/1475-2883-4-2ResearchCombined Utilisation of Rapid Assessment Procedures for Loiasis (RAPLOA) and Onchocerciasis (REA) in Rain forest Villages of Cameroon Wanji Samuel [email protected] Nicholas [email protected] Mathias [email protected] Siker SJ [email protected] Mark J [email protected] Peter [email protected] University of Buea, Faculty of Science, Department of Life Sciences, P.O. Box 63, Buea, Cameroon2 Research Foundation in Tropical Diseases and Environment (REFOTDE), P.O. Box 474, Buea, Cameroon3 Filariasis Research Laboratory, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK4 Tropical Medicine Research Station, P.O. Box 55, Kumba, Cameroon2005 7 4 2005 4 2 2 30 7 2004 7 4 2005 Copyright © 2005 Wanji et al; licensee BioMed Central Ltd.2005Wanji 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
Individuals with high microfilarial loads of Loa loa are at increased risk of neurologic serious adverse (SAE) events following ivermectin treatment against onchocerciasis. RAPLOA (Rapid Assessment Procedure for loiasis), a newly developed rapid assessment procedure for loiasis that relates the prevalence of key clinical manifestation of loiasis (history of eye worm) to the level of endemicity of the infection (prevalence of high intensity), is a very useful tool to identify areas at potential risk of L. loa post ivermectin treatment encephalopathy. In a perspective of treatment decision making in areas of co-endemicity of loiasis/onchocerciasis, it would be advantageous (both in time and cost savings) for national onchocerciasis control programmes to use RAPLOA and the Rapid epidemiologic assessment for onchocerciasis (REA), in combination in given surveys. Since each of the two rapid assessment tools have their own specificities, the workability of combining the two methods needed to be tested.
Methods
We worked in 10 communities of a forest area presumed co-endemic for loiasis and onchocerciasis in the North-West Province of Cameroon where the mass-treatment with ivermectin had not been carried out. A four-step approach was used and comprised: (i) generating data on the prevalence and intensity of loiasis and onchocerciasis in an area where such information is scarce; (ii) testing the relationship between the L. loa microfilaraemia prevalence and the RAPLOA prevalence, (iii) testing the relationship between the O. volvulus microfiladermia prevalence and the REA prevalence, (iv) testing the workability of combining RAPLOA/REA by study teams in which a single individual can perform the interview for RAPLOA and the nodule palpation for REA.
Results
The microfilaraemia prevalence of loiasis in communities ranged from 3.6% to 14.3%. 6 (0.61%) individuals had L. loa microfilarial loads above 8000 mf/ml but none of them attained 30,000 mf/ml, the threshold value above which the risk of developing neurologic SAE after ivermectin treatment is very high. None of the communities surveyed had RAPLOA prevalence above 40%. All the communities had microfiladermia prevalence above 60%. The microfiladermia results could be confirmed by the rapid epidemiologic method (nodule palpation), with all the 10 communities having REA prevalence above 20%. For the first time, this study has demonstrated that the two rapid assessment procedures for loiasis and onchocerciasis can be carried out simultaneously by a survey team, in which a single individual can administer the questionnaire for RAPLOA and perform the nodule palpation for REA.
Conclusion
This study has: (i) Revealed that the Momo valley of the North West province of Cameroon is hyperendemic for onchocerciasis, but is of lower level of endemicity for L. loa. (ii) Confirmed the previous relationships established between RAPLOA and the L. loa microfilaraemia prevalence in one hand and between the REA and the O. volvulus microfiladermia prevalence in another hand (iii) Shown that RAPLOA and REA could be used simultaneously for the evaluation of loiasis and onchocerciasis endemicity in areas targeted by the African Programme for onchocerciasis Control for community-directed treatment with ivermectin (CDTI).
==== Body
Background
For several years mass treatment with ivermectin has been used to control onchocerciasis. The community directed distribution of annual doses of ivermectin introduced through the African Programme for Onchocerciasis Control (APOC) is the key component of this programme. In this Community-directed treatment with ivermectin (CDTI), the community itself is in charge of designing and implementing the ivermectin distribution [1].
The mass distribution of ivermectin is always preceded by the mapping of the target area, using rapid epidemiologic mapping of onchocerciasis (REMO), which takes into consideration specific spatio-epidemiological characteristics of onchocerciasis (spatial distribution of vectors in breeding sites along rivers) and the Rapid epidemiologic assessment for onchocerciasis (REA), which is based on the estimation of the prevalence of onchocercal nodules in adult males using simple palpation [2]. The prevalence of palpable nodules in adult males is almost half the prevalence of microfiladermia in the total population (nodule prevalence of 20% corresponds to microfilaria prevalence of 40–60%) [3,4]. Communities with nodule prevalence of 20% and above are eligible for CDTI.
The large-scale distribution of ivermectin was successfully introduced into Cameroon until neurologic serious adverse events (SAEs) in individuals with high L. loa microfilaraemia after ivermectin treatment were reported [[5,6] and [7]]. The SAEs are characterized by progressive neurologic decline and encephalopathy within a few days of taking ivermectin. In some cases, this can result in death or chronic disability. A recent retrospective analysis of L. loa encephalopathy temporally related to treatment with Mectizan® (PLERM – 'Probable' or 'Possible' L. loa Encephalopathy temporally Related to treatment with Mectizan®), revealed that 97% of cases so far declared in Africa came from southern Cameroon, with 93% of them being individuals treated with ivermectin for the first time [8]. The main clinical signs and symptoms associated with the PLERM are altered mental status, incontinence, difficulty standing up or walking, dysarthria, fever, diarrhoea, headache and feverishness.
These neurologic SAEs have been observed mainly in areas where L. Loa and O. volvulus are co-endemic and this has hampered the advancement of the mass treatment with ivermectin in the forested areas. There was therefore an urgent need for a rapid method to identify communities where individuals are at risk of developing neurologic SAEs, before the implementation of mass treatment with ivermectin.
A study carried out in Cameroon and Nigeria in 2001 supported by the UNDP/World bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR) and APOC led to the development of a Rapid Assessment Procedure for loiasis (RAPLOA) [9,10]. From the results of this study it was found that communities in which more than 40% of individuals were reported to have experienced the sub-conjunctival migration of the adult L. loa, confirmed by a photograph of the worm in the eye, with the most recent episode lasting between 1–7 days, had a L. loa microfilaraemia prevalence of 20% or above. In such communities 5% of individuals harbour more than 8000 mf/ml and are exposed to significant increase risk of occurrence of functional impairment after ivermectin treatment [6]; meanwhile, 2% of them carry more than 30,000 mf/ml and have a risk of serious neurological reactions following ivermectin treatment [5].
It is now recommended by the Mectizan® Expert Committee and the Technical Consultative Committee of APOC (MEC/TCC) that before commencing ivermectin distribution in areas suspected, or known to be endemic for loiasis, RAPLOA should be undertaken to assess the prevalence of L. loa [11].
RAPLOA is a newly developed tool and its full programmatic implementation requires that it should be validated, in several independent studies, in sites different from where it was originally developed. Furthermore, it will be advantageous, (both in time and cost savings) for the National onchocerciasis control programmes in Africa, to be able to carry out RAPLOA and REA using single research teams to assess the level of endemicity of the two infections during a given survey. Data from such surveys could be used to facilitate rapid decision making to enable the most effective treatment [12]. For instance, if both RAPLOA and REA are negative, or if RAPLOA is positive and REA is negative, there should be no mass treatment with ivermectin. If REA is positive and RAPLOA is negative, the CDTI can be implemented safely; but if REA is positive and RAPLOA is also positive, mass treatment can be carried out with precautions (possibility of identifying early warning signs of encephalopathy and possibility of therapeutic intervention).
Given that each of the two rapid assessment tools have their own specificities (time factor, mode of assessment, selection of participants and sample size requirements), the workability of combining the two methods, needed to be evaluated. It is in this light that we recently carried out a study in 10 communities, selected from a forest area of North-west Cameroon, an area which has not yet undergone large scale distribution of ivermectin, but which is earmarked for CDTI and where O. volvulus and L. loa are co-endemic. The general objective of this study was to explore the feasibility of combining RAPLOA and REA in a given survey by a single research team. The specific objectives were to (i) provide information on the prevalence and intensity of loiasis and onchocerciasis in an area where data was scarce, (ii) validate RAPLOA in different sites from where it was originally developed and (iii) to reinforce the validity of nodule palpation in an area of co-endemicity L. loa/O. volvulus, (iv) to test the workability of combining RAPLOA/REA by study teams in which a single individual can perform the interview for RAPLOA and the nodule palpation for REA.
Methods
Study site
The study was carried out in 10 villages located in the Batibo Health District, North West province of Cameroon between March and May 2003. The villages were selected from the Widikum subdivision located in the Momo valley and included: Bifang, Ebendi, Ngalla, Ambombo, Eka, Dinku, Oche1, Oche 2, Mbullam and Olorunti. They extend between latitude 5° N 43 – 5° N 54 and between longitude 9° E 41 – 9° E 44. The vegetation is a degraded forest in which the primary rain forest has been completely replaced by the oil palm, which constitutes the main commercial crop. The inhabitants of these villages are mainly farmers. The climate is tropical with two seasons: the rainy season which lasts from mid March to mid November and the dry season from mid November to mid March. The main rivers are river Momo and Tanjo, which are tributaries of river Manyu (Figure 1).
Figure 1 Map of the Widikum area showing the geographical location of the villages surveyed.
Study population
A census was conducted in each of the villages surveyed to estimate the population size. The study population consisted of males and females aged 15 years and above who have been resident in the village for a minimum of five consecutive years and who have not taken antifilarial treatment for a minimum period of one year. All eligible members of the community and who consented to participate were enrolled into the study. The investigation was carried out according to a protocol approved by the ethical committee of the Research Foundation in Tropical Diseases, and Environment, Buea and the Tropical Medicine Research Station Kumba.
Conduct of the rapid assessment procedures for loiasis and onchocerciasis
Organization of work
A form designed to collect data was divided into four sections: the first section was for the identification of participants, the second and third sections for the collection of RAPLOA and REA data with special reference on the starting and ending time for both exercises (RAPLOA and REA). The last section was for the parasitological results of skin snips and thick blood films. In each community surveyed, a team of three technicians moved from one household to another to register eligible participants, administer the RAPLOA questionnaire and carried out a Rapid epidemiological assessment (REA) by nodule palpation. Each participant was then referred to the parasitological post situated at the centre of the village for blood collection and skin snipping.
Administration of RAPLOA questionnaire
The Rapid Assessment Procedure for loiasis was based on the restricted definition of the eye worm; the past experience of eye worm, confirmed by a photograph of L. loa adult worm in the white part of the eye and with the duration of the most recent episode being between 1 to 7 days [9,10]. The questionnaires were administered in the English language and where required, interpreters from the community assisted in the interview process according to the RAPLOA guidelines [13]
Nodule palpation (REA)
The REA was based on nodule palpation. After undergoing the RAPLOA interview, every patient was examined by the same technician for the presence of Onchocerca nodules according to previous studies [4,14].
Parasitological examinations
These involved the thick blood film to search for L. loa microfilariae (mf) and the skin snip for O. volvulus microfilariae.
Thick blood film
The thick blood films were prepared from a standardised 50 μl finger prick blood collected between 10:00 and 16:00 hours using a 75 μl non-heparinised capillary tube [15,16,10].
Collection and processing of skin snips
Two skin snips were taken with a sterile corneo-scleral punch (Holth 2 mm) from the two iliac crests of each patient. The skin snips were placed in two separate wells of a microtitration plate containing 100 μl of normal saline and covered with parafilm. Between each patient the corneo-scleral punches were sterilised by dipping in hypochlorous solution, followed by 70% ethanol and finally rinsed in distilled water. After 24 hours, a drop of 2% formalin was added to each well to preserve the microfilariae. Microfilariae were counted later using an inverted microscope.
Expression of results
The results were entered in Epi Info (Version 6.03, 1996) and analysed using SPSS 10.1 for Windows. The RAPLOA prevalence was expressed as the proportion of individuals whose answers to the questionnaire complied with the restricted definition of eye worm. The REA prevalence was expressed as the proportion of individuals with at least one palpable O. volvulus nodule.
The arithmetic mean and William's mean were used to express the intensity of the infection in study villages. The arithmetic mean was determined as the average count of mf/ml or mf/skin snip in each village. The William's mean was determined as the geometric mean calculated after having added 1 to each count [17]. The Chi-square test of trend was used to compare prevalence between communities whereas the non-parametric Kruskal and Wallis test was used to assess differences in the intensities of the infection in different communities. Scatter diagrams were plotted with the following cut-off points defined in previous studies. RAPLOA40: Prevalence of loiasis, as determined by the restricted definition of history of eye worm, above which the risk of SAEs post ivermectin treatment is increased. MFLOA20: Prevalence of L. loa, determined by thick film method, above which individuals within the community are at risk of SAEs post mass ivermectin treatment. REA20: Prevalence of onchocerciasis, determined by nodule palpation, above which the large-scale treatment with ivermectin is highly desirable. MFDONCHO60: Prevalence of onchocerciasis, determined by skin snipping, above which the large-scale treatment with ivermectin is most urgent. The RAPLOA results generated from this study were superimposed on the same scatter plot with the original data generated during the development of the RAPLOA study for comparison.
Results
Prevalence and intensity of loiasis
Table 1 gives the prevalence and intensity of L. loa microfilariae as well as the RAPLOA prevalence in study villages. A total of 977 individuals both males and females were enrolled in the study. The prevalence of L. loa microfilariae varied from one village to another. The lowest prevalence (3.36%) was observed at Ebendi whereas the highest prevalence (14.29 %) was observed at Mbullam. RAPLOA prevalence also varied from one village to another, ranging from 9.38 % in Oche 1 to 31.03% in Oche 2.
Table 1 Prevalence and intensity (mf/ml) of L. loa determined by RAPLOA and by the Thick blood film techniques.
Prevalence (%) Intensity (mf/ml)
Village Population* No. Examined RAPLOA Thick blood film Arithmetic Mean Mf +ve & -ve individuals Arithmetic Mean Mf +ve individuals William's means > 15 William's means > 20 (CMFL)
Ambombo 250 49 24.49 6.12 1.63 27.67 1.22 1.25
Bifang 1889 141 21.99 3.55 81.99 2313 1.25 1.30
Mbullam 65 21 19.05 14.29 86.67 608.67 2.36 2.21
Dinku 516 159 30.82 10.69 1.3.77 970.79 1.83 2.02
Ebendi 260 120 13.33 3.36 8.00 240.00 1.18 1.14
Eka 1079 96 14.58 9.38 418.96 4468.89 1.87 1.96
Ngalla 765 138 16.67 7.25 232.46 3208.00 1.53 1.55
Oche 1 83 32 9.38 12.50 293.13 2345.00 2.08 2.13
Oche 2 77 29 31.03 10.34 13.10 126.67 1.609 1.93
Olorunti 707 192 20.31 4.69 119.27 2544.44 1.39 1.40
Total 5691 977
Average 20.47 6.86 139.08 2028.06 1.50 1.53
* Total population recorded during a census conducted in study villages; CMFL: Community microfilarial load. Mf +ve: Microfilaraemic, MF -ve: amicrofilaraemic
The arithmetic means of L. loa in the entire study population ranging from 8 mf/ml of blood in Ebendi to 418.96 mf/ml of blood in Eka. In microfilaraemic individuals they ranged from 27.67 mf/ml in Ambombo to 4468.89 mf/ml in Eka. The William's mean of microfilarial loads also differed from one village to another, ranging from 1.18 mf/ml to 2.36 mf/ml in the entire study population and from 1.14 mf/ml to 2.21 mf/ml in individuals with ≥ 20 years.
Prevalence and intensity of onchocerciasis
The prevalence and intensity of O. volvulus as determined by REA and Skin snips are indicated in Table 2. The prevalence of palpable nodules varied significantly (p < 0.001) from one village to another. These ranged from 20.83% in Eka to 65% in Ebendi. The prevalence of onchocerciasis determined by the skin snips was high in all the 10 villages. It varied from 66.67% at Bifang to 95.24% at Dinku. The William's means varied from 4.08 at Bifang to 17.38 at Dinku. In individuals with age ≥ 20 years, it ranges from 4.19 mf/ml in Bifang to 18.58 in Mbullam.
Table 2 Prevalence and Intensity of O. volvulus determined by REA and Skin snip
Prevalence (%) Intensity (mf/Skin snip)
Village Population* No. Examined REA Skin snip Arithmetic Mean Mf +ve & -ve individuals Arithmetic Mean Mf +ve individuals William's means ≥ 15 William's means ≥ 20 (CMFL)
Ambombo 250 49 40.82 69.39 14.57 21.00 5.07 4.76
Bifang 1889 141 38.30 66.67 12.54 18.80 4.06 4.19
Mbullam 65 21 23.81 95.24 65.26 68.52 13.38 14.60
Dinku 516 159 54.72 94.34 32.45 34.40 17.31 18.58
Ebendi 260 120 65.00 85.83 19.70 22.95 7.76 7.82
Eka 1079 96 20.83 72.92 14.13 19.37 5.24 5.69
Ngalla 765 138 56.52 76.09 18.45 24.25 6.85 7.70
Oche 1 83 32 21.88 87.50 11.25 12.86 5.41 6.52
Oche 2 77 29 55.17 86.21 28.07 32.56 10.53 10.49
Olorunti 707 192 39.06 86.46 31.68 36.43 10.57 10.89
Total 5691 977
Average 45.04 81.37 23.06 28.31 7.99 8.22
* Total population recorded during a census conducted in study villages, CMFL : Community microfilarial load ; Mf +ve: Microfiladermic, MF -ve: amicrofiladermic individuals & -ve
Knowledge of eye worm, attitude and practice in study communities
The eye worm was well known in these communities. In Bifang, Ebendi, Ngalla, Ambombo and Eka it is called "Embele", in Dinku, Oche 1, Oche 2 and Mbullam it is called "ntembele". In Olorunti it is called "etembele". These names were generally descriptive. It was noticed that communities have "Eye Worm specialists", who are capable of removing adult L. loa as they migrate across the subconjunctiva.
Relationship between the RAPLOA and L. loa microfilaraemia prevalence
Figure 2a summarises the relationship between the parasitological prevalence and RAPLOA prevalence. All the villages surveyed had RAPLOA prevalences less than 40%. L. loa microfilaraemia prevalences were less than 20 % in all the communities. The parasitological prevalences are in agreement with the RAPLOA predictions. These results fitted very well when superimposed on the original data generated during the development of RAPLOA (Figure 3).
Figure 2 a: Relationship between the RAPLOA and L. loa microfilaraemia prevalences. Dotted lines represent the thresholds levels above which there is increased risk of neurologic SAEs. b Relationship between the REA and the microfiladermia prevalences. Dotted lines represent the thresholds levels above which mass treatment with ivermectin is most urgent (microfiladermia) or highly desirable (REA). c Relationship between RAPLOA and REA prevalences. A: Large-scale treatment with low risk of neurologic SAEs; B: Large scale treatment with high risk of neurologic SAEs; C: No large-scale treatment with ivermectin D: No large-scale treatment with ivermectin. Dotted lines represent the thresholds levels above which the risk of neurologic SAEs is high
Figure 3 Relationship between RAPLOA prevalence and Loa microfilaraemia: (▲) New data, (o) Data generated during RAPLOA development.
Relationship between the REA prevalence and microfiladermia prevalence
Figure 2b summarises the relationship. All the communities surveyed had REA prevalence above 20% and prevalence of microfiladermia greater than 60%. On average, the skin snip prevalences (81.37%) were nearly twice as high as the REA prevalences (45.04%). There is good agreement between the two sets of prevalence, indicating that the study area has high endemicity for onchocerciasis. Some communities with REA close to 20% had high microfiladermia prevalence (> 70 %).
Combining RAPLOA and REA
This study has revealed that an examiner, using a single recording form with specific sections for REA and RAPLOA can take 8–10 minutes to examine an individual for loiasis and onchocerciasis in tandem. With three technicians performing the two tasks at the household's level, it took on average 4 hours to cover one community with a sample size of 80 people per day. Not-withstanding the difficulties of movement in the study area due to bad roads or lack of passable roads, we could examine during this survey on average two communities with an average sample size of 80 individuals per day. In the village Olurunti, working from 10:00 AM to 3:00 PM we examined 192 individuals. Figure 2c summarises the relationship between REA (%) and RAPLOA (%). All the communities surveyed had REA above 20% and RAPLOA less than 40% indicating that CDTI could be conducted within them safely.
Discussion
Endemicities of onchocerciasis and loiasis in the study area
This study has revealed that the Widikum health area situated in the Momo valley is an area hyperendemic for onchocerciasis. All the communities surveyed had microfiladermia prevalences above 60%. The intensity of infection as expressed by the arithmetic means and the William geometric means were relatively low. None of the communities had an arithmetic mean above 100 mf/skin snip. This lack of close relationship between the prevalence and intensity of infection could be explained by the existence of passive treatment with ivermectin in the area. Even though the CDTI was not in place in this area at the time of this survey, it could be noticed that almost 25% of individuals involved in the study stated they had taken ivermectin 2 to 3 years before the study (unpublished observations). Since the effect of ivermectin can last up to 3 years, this could have contributed to the reduction of microfilarial density. Indeed, the Momo valley is close to the South-West Province where the CDTI has been going on for over 5 years and some people admitted getting treatment from the South-West project.
The infection level of L. loa was relatively low. None of the communities surveyed had L. loa microfilariae prevalence (mf-prevalence) greater than 20%. According [16], for practical purposes, communities with L. loa mf-prevalence less than 10% should be considered as hypo endemic and communities with mf-prevalence between 10% and 20% should be considered as meso-endemic. The intensity of L. loa infection in microfilaraemic patients as expressed by the arithmetic mean did not exceed 4500 mf/ml, contrasting with the observations made in the Lekié division [16], Central Province of Cameroon, where some communities examined for L. loa had intensities up to 22791 mf/ml of blood. It should be noted that the Lékié division is the area where most of the cases of Loa-related encephalopathies were observed. The results of this study also contrast with the observations made in the Ntem valley (a forested savannah zone in the North-west province of Cameroon) [17], where in a survey carried out in10 communities, L. loa intensities (arithmetic mean of total number of patients examined) of up to 9995 mf/ml were observed in some communities. However, the findings of the present study were very similar to the observations made in the forest villages of south-west Province of Cameroon [18].
Relationship between the Rapid assessment procedure prevalences and the parasitological prevalences
– RAPLOA prevalence versus Loa-microfilaria prevalence
None of the communities surveyed had RAPLOA prevalence above 40% (threshold value established by the RAPLOA development study) [10]. Also none of the communities had L. loa microfilareamia above 20% (threshold prevalence above which more than 5% of adults living in the community are at risk of adverse-related responses with ivermectin) [16]. There is therefore a good agreement between the results from the Thick blood film technique and the results from RAPLOA, since the two sets of findings have reached the same conclusion that the study area is not hyperendemic for loiasis and that the risk of serious adverse events after ivermectin treatment may be very low in these communities. In this light, this independent study has validated the rapid assessment procedure for loiasis (RAPLOA) in predicting the level of endemicity of loiasis and the risk of neurologic serious adverse events after ivermectin treatment in a given area. It may not be necessary therefore to strengthen monitoring by a medical team, after a campaign of mass treatment with ivermectin. Indeed, a mass treatment with ivermectin, carried out in the 10 communities and involving some 4000 eligible individuals did not show any neurologic serious adverse events related to ivermectin treatment (unpublished data)
- REA prevalence versus Onchocerca microfiladermia prevalence
All the villages surveyed had prevalences of palpable nodules above 20% (threshold defined as proportion above which a community should be included in the CDTI programme). These REA results were confirmed by skin snip results; all the communities examined had microfiladermia prevalences above 60%. These two sets of results confirm that the Momo valley, characterized by important fast flowing rivers with excellent breeding sites for Simulium damnosum s.l, is an area of high endemicity of onchocerciasis. The results of this study are in agreement with previous studies [4,14] and constitute a further validation of the REA to predict the level of endemicity of onchocerciasis. Although the validation of REA has been done before, this is the first time such validation is being done in an area of co-endemicity of onchocerciasis/loiasis.
Combining REA and RAPLOA and its implications for the onchocerciasis control programmes' activities in Africa
The recent MEC/TCC recommendation that before introducing the mass distribution of ivermectin for onchocerciasis control in an area suspected or known to be endemic for loiasis, RAPLOA should be carried out to assess the prevalence of L. loa, has necessitated the evaluation of the possibility of combining REA and RAPLOA, the two assessment procedures for onchocerciasis and loiasis endemicity.
The two tools have several similarities in their methodological approaches: they are carried out on individuals (adults or nearly-adults) of both sexes, who have been resident in a community for a long period of time, their sample size requirements are not conflicting, the exercises of interview and nodule palpation are very simply to execute. However, the question of whether a single examiner could perform the two tasks in succession, with a high level of efficiency, as measured by the ratio output/time, arose. In other words, how long can it take an examiner to assess a patient for the two filarial infections? This study has shown that it can take 8 to 10 minutes. By extrapolating this basic result to a survey team made up of 5 individuals who can carry out the two procedures in tandem, it is possible to examine 80 subjects in less than three hours. This will allow a good team to evaluate 3–4 communities per day under optimum conditions (good mobilisation, community accessibility). Indeed, we were able in the present study to examine 192 individuals in the village of Olurunti from 10:00 AM to 3: 00 P.M with a team of 3 persons.
The possibility of simultaneously carrying out at the community level RAPLOA and REA for assessing loiasis and onchocerciasis prevalences will have a great impact on the planning and implementation of activities of onchocerciasis control programmes in areas of co-endemicity with loiasis. There will be a gain both in time and in cost savings, since a single well organized survey in a given area will be enough to assess the level of endemicity of the two filarial species. More important will be the fact that, data from such surveys could be used rapidly in treatment decision making. For instance, if both RAPLOA and REA are negative (prevalence <40% and <20% respectively), or if RAPLOA is positive (prevalence >40%) and REA is negative, there should be no mass treatment with ivermectin. If REA is positive (prevalence >20%) and RAPLOA is negative, the CDTI can be implemented safely. But if REA is positive and RAPLOA is also positive, mass treatment can be carried out following MEC/TCC guidelines. The results of this combined survey RAPLOA/REA clearly indicate that the CDTI could be implemented safely in the Momo valley of the North-west province of Cameroon. Indeed, a CDTI project, which is now in its second year of activities in this area, has not so far registered any neurologic SAE (unpublished data).
List of Abbreviations
APOC: African Programme for Onchocerciasis Control
CDTI: Community-Directed Treatment with Ivermectin
SAEs: Severe Adverse Effects
MEC/TCC: Mectizan Expert Committee/Technical Consultative Committee
REMO: Rapid Epidemiological Mapping of Onchocerciasis
CMFL: Community Microfilarial Load
RAPLOA: Rapid Assessment Procedure for Loiasis
RAPLOA40: Loiasis prevalence, as determined by RAPLOA, above which the risk of serious adverse effect is increased
MFLOA20: Loiasis prevalence, as determined by thick blood film method, above which the risk of serious adverse effect is increased
REA: Rapid Epidemiological Assessment of Onchocerciasis
REA20: Onchocerciasis prevalence, as determined by REA, above which the large-scale treatment with ivermectin is highly desirable
MFD-ONCHO60: Onchocerciasis prevalence, as determined by skin snipping, above which the large-scale treatment with ivermectin is most urgent
Mf-prevalence: microfilarial prevalence
PLERM: 'Probable' or 'Possible' Loa Loa Encephalopathy temporally related to treatment with Mectizan
Authors' contributions
SW: Participated in the design of the study, collection and processing of the data, drafting and editing of the manuscript.
TN: Participated in the collection, processing and analysis of data.
YSSJ: Participated in the collection and processing of data.
EM: Participated in the collection and processing of the data
MJT: Collection of data and review of the manuscript,
EP: Participated in the design of the study, data collection and review of the manuscript.
Acknowledgements
We wish to acknowledge the active participation of the following persons Mr Jato Isaac, Bernard Syankalbe, Mboussi Kassang, Tchuassi David, Tanga Chrysantus. The chief medical officer, Batibo health district, the chief of health post of Olurunti. We also wish to thank the members of the communities involved in this study. We thank Dr Hans Remme for his technical assistance. This study received financial support from the European Union grant, Wolbachfil, ICA4-CT2002-10051.
==== Refs
TDR Community directed treatment with ivermectin: Report of a Multi-country study TDR/FT/RP/961 WHO Geneva 1996
Ngoumou P Walsh JF Mace JM A rapid mapping technique for the prevalence and distribution of onchocerciasis: a Cameroon case study Annals of Tropical Medicine and Parasitology 1994 88 463 474 7979636
TDR Methods for community diagnosis of onchocerciasis to Guide ivermectin based control in Africa TDR/AFT/RP/961 WHO, Geneva 1992
Taylor HR Duke BO Munoz B The selection of communities for treatment of onchocerciasis with ivermectin Trop Med Parasitol 1992 43 267 70 1293734
Chippaux JP Boussinesq M Gardon J Gardon-Wendel N Ernould JC Severe adverse reaction risks during mass treatment with ivermectin in loiasis-endemic areas Parasitology Today 1996 12 448 450 15275280 10.1016/0169-4758(96)40006-0
Gardon J Gardon-Wendel N Demanga-Ngangue Kamgno J Chippaux JP Boussinesq M Serious reactions after mass treatment of onchocerciasis with ivermectin in an area endemic for L. loa infection Lancet 1997 350 18 22 9217715 10.1016/S0140-6736(96)11094-1
Boussinesq M Gardon J Gardon-Wendel N Kamgno J Ngoumou P Chippaux JP Three probable cases of L. loa encephalopathy following ivermectin treatment for onchocerciasis American Journal of Tropical Medicine and Hygiene 1998 58 461 469 9574793
Twum-Danso NA L. loa encephalopathy temporally related to ivermectin administration reported from onchocerciasis mass treatment programs from 1989 to 2001: implication for the future Filaria Journal 2003 2 S7 14975064 10.1186/1475-2883-2-S1-S7
Wanji S editor Rapid Assessment Procedures for loiasis. Report of a Multi-centre study UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Disease TDR/IDE/RP/RAPL/011 2001
Takougang I Meremikwu M Wanji S Yenshu EV Aripko B Lamlenn SB Eka BL Enyong P Meli J Kale O Remme JH Rapid assessment method for prevalence and intensity of L. loa infection Bull World Health Organ 2002 80 852 8 12481206
MEC/TCC Guidelines Recommendations for the treatment of Onchocerciasis with Mectizan® in areas co-endemic for Onchocerciasis and Loiasis 2004
Addiss GD Richard R Nana AY Twum-Danso Frank OR A framework for Directed-making for mass distribution of Mectizan® in Areas Endemic for L. loa Filaria Journal 2003 2 S9 14975066 10.1186/1475-2883-2-S1-S9
UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Disease, Guidelines for Rapid assessment of Loa loa TDR/IDE/RAPLOA/021 2002
Kipp W Bamhuhiiga J Validity of nodule palpation in a Simulium neavei-transmitted onchocerciasis area in Uganda American Journal of Tropical Medicine and Hygiene 2002 67 28 131 12363060
Sasa M Microfilaria survey methods and analysis of survey data in filariasis control programmes Bull World Health Organ 1967 37 629 65 4873190
Boussinesq M Gardon J Kamgno J Pion SDS Gardon-Wendel N Chippaux JP Relationship between the prevalence and intensity of L. loa infection in the Central province of Cameroon Annals of Tropical Medicine and Parasitology 2001 95 495 507 11487371 10.1080/00034980120073184
Wanji S Tendongfor N Esum M Atanga S and Enyong P Heterogeneity in the prevalence and intensity of loiasis in contrasting bio-ecological zones of Cameroon Tran R Soc Trop Med Hyg 2003 97 182 187 10.1016/S0035-9203(03)90114-3
Wanji S Tendongfor N Esum M Ndindeng S Enyong P Epidemiology of Concomitant infections due to L. loa, Mansonella perstans and Onchocerca volvulus in rain forest villages of Cameroon Medical Microbiol Immunol (Berl) 2003 192 15 21
| 15817124 | PMC1090603 | CC BY | 2021-01-04 16:38:24 | no | Filaria J. 2005 Apr 7; 4:2 | utf-8 | Filaria J | 2,005 | 10.1186/1475-2883-4-2 | oa_comm |
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-61583679110.1186/1742-9994-2-6EditorialFrontiers in Zoology speeds up Heinze Jürgen [email protected] Diethard [email protected] Biologie I, Universität Regensburg, D-93040 Regensburg, Germany2 Institute for Genetics, Weyertal 121, D-50931 Cologne, Germany2005 18 4 2005 2 6 6 4 4 2005 18 4 2005 Copyright © 2005 Heinze and Tautz; licensee BioMed Central Ltd.2005Heinze and Tautz; 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.
==== Body
Introduction
Only a few months after its official launch by BioMed Central, Frontiers in Zoology has become a widely-read journal attracting submissions of reviews and original papers on research from all fields of zoology. To ensure that peer-reviewed research is Open Access, i.e. universally and freely available online to everyone and that it is archived in internationally recognised free repositories [1], from April 1, 2005, authors of articles accepted for publication will be asked to pay an article-processing charge (APC) of £ 330.
Traditionally, readers pay to access articles, either through subscriptions or by paying a fee each time they download an article. For many journals, subscription costs have increased over the last couple of years. Journal costs have always been a problem for researchers in some countries, but the recent cost escalation together with cuts in funding for universities have almost globally resulted in libraries describing to fewer journals [2]. The availability of results to readers therefore has become more and more limited. Although traditional journals publish authors' work for free (unless there are page or colour charges), having to pay to access articles limits how many can read, use and cite them.
The advantages of Open Access
As summarized by Slade et al. [3], the Open Access policy generally changes the way in which articles are published. First, all articles become freely and universally accessible online, and so an author's work can be read by anyone at no cost. Second, the authors hold copyright for their work and grant anyone the right to reproduce and disseminate the article, provided that it is correctly cited and no errors are introduced [1]. Third, a copy of the full text of each article is permanently archived in an online repository separate from the journal. Articles in Frontiers in Zoology are archived in PubMed Central [4], the US National Library of Medicine's full-text repository of life science literature, and also in repositories at the University of Potsdam [5] in Germany, at INIST [6] in France and in e-Depot [7], the National Library of the Netherlands' digital archive of all electronic publications.
Open Access has four broad benefits for science and the general public. First, authors are assured that their work is disseminated to the widest possible audience, given that there are no barriers to access their work. This is accentuated by the authors being free to reproduce and distribute their work, for example by placing it on their institution's website. It has been shown that free online articles are more highly cited because of their easier availability [8]. Second, the information available to researchers will not be limited by what their library can afford, and the widespread availability of articles will enhance literature searching [9]. Third, the results of publicly funded research will be accessible to all taxpayers and not just those with access to a library with a subscription. Note that this public accessibility may become a legal requirement in the USA if the proposed Public Access to Science Act is made law [10]. Fourth, a country's economy will not influence its scientists' ability to access articles because resource-poor countries (and institutions) will be able to read the same material as wealthier ones (although creating access to the internet is another matter [11]).
Article-processing charges
APCs will allow continued Open Access to all articles published in Frontiers in Zoology. Authors are asked to pay £ 330 if their article is accepted for publication. Waiver requests will be considered on a case-by-case basis, by the Editor-in-Chief. Authors can circumvent the charge by getting their institution to become a 'member' of BioMed Central, whereby the annual membership fee covers the APCs for all authors at that institution for that year. Current members include the Deutsche Zoologische Gesellschaft, who initiated and supports this journal, Harvard, Princeton and Yale universities, all UK universities, CNRS, INRA, and Institute Pasteur, Max Planck Society, and many others [12]. Publishing in Frontiers in Zoology and other BMC journals is completely free of costs for all whose institutions are listed as BioMed Central Institutional Members. Please note that no charge is made for articles that are rejected after peer review. Many funding agencies have also realized the importance of Open Access publishing and have specified that their grants may be used directly to pay APCs [13].
The APC pays for efficient and thorough peer review, for the article to be freely and universally accessible in various formats online, and for the processes required for inclusion in PubMed and archiving in PubMed Central, e-Depot, Potsdam and INIST. Although some authors may consider £ 330 expensive, it must be remembered that Frontiers in Zoology does not levy additional page or colour charges on top of this fee, which can easily exceed £330. With the article being online only, any number of colour figures and photographs can be included, at no extra cost. Given that colour prints are often essential for the documentation of results from research on morphology, neurobiology, taxonomy etc. and extremely valuable also in other discipline of zoology, Frontiers in Zoology provides a suitable medium for the broad dissemination of high-quality research from all areas of zoological research.
Although several journals now offer free access to their articles online, this is different from Open Access (as defined by the Bethesda Statement [14]). Journals often delay free access for 6–12 months, and even when the full text is available, readers are not allowed to reproduce and/or disseminate the work because of restrictions imposed by the copyright policy. That said, Frontiers in Zoology is not alone in the move to Open Access funded by APCs. The British Medical Journal has recently announced that it cannot continue to provide free access to its website [14] and is considering various sources of revenue, including APCs [16]. Also, the Public Library of Science has set up new Open Access journals, and have elected to set APCs of US$1500 for each accepted article [17]. Given that the Public Library of Science has used television advertising to promote journals [10], the high profile of these journals will raise awareness of Open Access and encourage researchers in all disciplines to understand and accept Open Access, with APCs as an acceptable method to fund it.
Conclusion
By providing a forum for Open Access, APCs will enable Frontiers in Zoology to continue to publish attractive and important papers on outstanding research from all fields of zoology that will be accessible to everyone worldwide. We believe this change will support the current renaissance of integrative zoology as a modern and vigorous field of research, and we hope you will support this progress by submitting your next article to this Open Access journal.
==== Refs
BioMed Central Open Access Charter
Tamber PS Is scholarly publishing becoming a monopoly? BMC News and Views 2000 1 1
Slade E Tamber PS Vincent J-L Critical Care's move to fund open access Crit Care 2003 7 1 2 12617727 10.1186/cc2326
PubMed Central
Potsdam
INIST
e-Depot
Lawrence S Free online availability substantially increases a paper's impact Nature 2001 411 521 11385534 10.1038/35079151
Velterop J Should scholarly societies embrace Open Access (or is it the kiss of death)? Learned Publishing 2003 16 167 169 10.1087/095315103322110932
Open Access law introduced
Tan-Torres Edejer T Disseminating health information in developing countries: the role of the internet BMJ 2000 321 797 800 11009519 10.1136/bmj.321.7264.797
BioMed Central Institutional Members
Which funding agencies explicitly allow direct use of their grants to cover article processing charges?
Bethesda Statement on Open Access Publishing
Delamothe T Smith R Paying for bmj.com BMJ 2003 327 241 242 10.1136/bmj.327.7409.241
Smith R The BMJ will experiment with the 'author pays' model (Rapid response to BMJ 2003;327:241-2)
Public Library of Science to launch new free-access biomedical journals with $9 million grant from the Gordon and Betty Moore Foundation
| 15836791 | PMC1090604 | CC BY | 2021-01-04 16:38:34 | no | Front Zool. 2005 Apr 18; 2:6 | utf-8 | Front Zool | 2,005 | 10.1186/1742-9994-2-6 | oa_comm |
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-71584768610.1186/1742-9994-2-7ResearchBehavioral and neuroendocrine consequences of social subjugation across adolescence and adulthood Ferris Craig F [email protected] Tara [email protected] Ross [email protected] Center for Comparative Neuroimaging, University of Massachusetts Medical School, Worcester, Massachusetts, USA2005 22 4 2005 2 7 7 3 1 2005 22 4 2005 Copyright © 2005 Ferris et al; licensee BioMed Central Ltd.2005Ferris 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
Social subjugation is a very significant and natural stressor in the animal kingdom. Adult animals defeated and subjugated during establishment of dominance hierarchies or territorial encounters can be highly submissive in future agonistic interactions. While much is know about the biological and behavioral consequences of winning and losing fights in adulthood, little is known about adolescence; a developmental period noted for impulsivity and heightened agonistic behavior. The present studies were undertaken to determine if the behavioral and neuroendocrine consequences of social subjugation are comparable in adolescent versus adult Syrian golden hamsters (Mesocricetus auratus). Male siblings were studied from adolescence into adulthood following exposure to counterbalanced episodes of either a benign stressor, i.e., isolation in a novel cage, or the more severe stressor of social subjugation.
Results
As adults, hamsters with a history of social subjugation in adolescence show high levels of aggression toward intruders as compared to siblings subjugated in adulthood. Sibling controls subjugated in adulthood are highly submissive with little or no aggressive behavior. However, when subjugated in adulthood, hamsters with the earlier history of subjugation are no different than their sibling controls, i.e., adult subjugation promotes submissive behavior. Sexual motivation is high in adult hamsters with adolescent subjugation and testosterone levels remained stable over adulthood. In contrast, sibling controls subjugated in adulthood show lower levels of sexual motivation and reduced levels of testosterone. Release of cortisol during agonistic encounters is blunted in animals subjugated in adolescence but not adulthood. Measures of anxiety are reduced in hamsters with adolescent subjugation as compared to their sibling controls.
Conclusion
These data demonstrate a pronounced difference in behavior and neuroendocrinology between adolescent and adult hamsters in their response to social subjugation and suggest adolescence is a resilient period in development.
==== Body
Background
Social subjugation is a natural stressor in the animal kingdom with long-term behavioral consequences. Adult male rhesus monkeys that fight for dominance status and lose are relegated to the lowest social rank displaying highly submissive behavior [1]. Social subjugation in adult male talapoin monkeys reduces social activity and sexual behavior even in the absence of dominant conspecifics [2]. Defeated adult mice and rats display less aggressive and more submissive behavior [3-5]. Individually housed adult hamsters will routinely attack and bite an equal or smaller sized intruder placed into their home cage. However, following repeated defeat by a dominant conspecific, a resident hamster will be defensive or fearful of equal sized non aggressive intruders [6-8].
Social subjugation in adult animals can dramatically alter stress and reproductive hormones levels. There are many reports across species that continuous subjugation results in the dysregulation of the stress response resulting in elevated and protracted levels of stress hormone [8-13]. High basal levels of glucocorticoids are associated with a depressed immune system, diminished reproductive function, bone loss, and increased fear, anxiety and depression [14]. So the reduced aggressive and sexual behavior and heightened submission observed in defeated animals may be due to, in part, dysregulation of stress hormone release. The diminution in offensive aggression and sexual activity may result also from a general reduction in circulating levels of testosterone that normally present in subjugated animals as compared to their dominant male conspecifics [1,15-17].
While the behavioral and neuroendocrine consequences of social subjugation in adult animals are well documented, less in know about the effects of aggressive encounters in the adolescent period. Adolescence is defined as a period of pronounced physical, cognitive and emotional growth. This period usually begins just before puberty and ends in early adulthood with sexual maturity, social awareness and independence [18]. The animal reported on here, the Syrian golden hamster (Mesocricetus auratus) has a developmental period analogous to adolescence. In the wild, hamsters wean around P-25 (postnatal day 25), leave the home nest, forage on their own, establish nest sites, and defend their territory [19,20]. Hamsters can begin to establish dominance hierarchies as early as P-35 [21], and have a minimal breeding age of 42 days [22]. Androgen levels start to rise between P-28 and P-35 [23,24]. Thus between P-25 and P-42, as hamsters achieve independence from the maternal nest, they increase their weight and size, reach full sexual maturity and reproductive competence and establish social relationships. This period between P-25 and P-42 is designated as adolescence in golden hamsters. Hamsters in the wild are solitary and live in their own isolated burrows [19,20]. Thus, animals studied in the laboratory setting can be individually housed after weaning, an experimental feature that eliminates the confounding variable of group interactions.
Previous work on social subjugation in adolescent hamsters reported unexpected behavioral changes [25]. Male golden hamsters weaned at P-25, were exposed daily to aggressive adults from P-28 to P-42, and tested for aggression as young adults several days later after the cessation of stress. Animals with a history of social subjugation showed a context-dependent alteration in their aggressive behavior. They showed little or no aggression toward intruders of comparable age and size. However, when confronted by a smaller, younger intruder they were exceedingly aggressive, displaying short attack latencies and high number of bites as compared to sibling controls that were not subjugated during adolescence. Given the dire consequences of social subjugation in adulthood, i.e., low social status, submissive behavior, decreased reproductive activity, higher risk for disease, etc. could these hamsters with a history of adolescent subjugation be able to compete for dominant status as adults? The present studies were undertaken to test the hypothesis that adolescence is a resilient development period, immune to the behavioral consequences of repeated social subjugation.
Results
Agonistic Behaviors
The effects of social history on biting behavior and flank marking are shown in Figure 1. There was a significant main effect in the latency to bite intruders between AS (adolescent subjugation) and AI (adolescent isolation) hamsters (F(1,18) = 23.4, p < 0.01). There was also a significant main effect for repeated test trials across the six month period (F(3,54) = 74.15, p < 0.001). When tested as young adults at 48 days of age, AS hamsters on average bite intruders in just over two min while AI hamsters took almost six min to bite (p < 0.01). This difference in bite latency was even more pronounced when animals were tested around 108 days of age. However, at this time AS hamsters were exposed to isolation stress prior to testing, while their AI siblings were subjugated. Average bite latencies were reduced to less than 40 sec in AS hamsters follow adult isolation but delayed by over 8 min in AI hamsters following adult subjugation (p < 0.01). Indeed, the first episode of adult subjugation dramatically inhibited biting behavior regardless of subsequent life histories as noted at 168 and 222 days of age.
Figure 1 Measures of Aggressive Behavior: Shown are the mean scores (+ SEM) for bite latency, bites and flank marks for hamsters initially subjugated in adolescence (black bar) as compared to their siblings isolated in adolescence (open bar) across a changing social history of adult subjugation (AS) or adult isolation (AI). (* p <.05; ** p < .01).
The profile of biting attacks complimented the bite latency data. There was a significant main effect in the frequency of biting attacks toward intruders between AS and AI hamsters (F(1,18) = 26.9, p < 0.001). There was also a significant main effect for test trials (F(3,54) = 37.87, p < 0.001). When tested as young adults at 48 days of age, AS hamsters displayed almost twice as many bites as their AI siblings (p < 0.05). This difference in biting attacks was even more pronounced when hamsters were tested around 108 days of age. Adult subjugation dramatically reduced biting attacks as compared to adult isolation (p < 0.01). This result was confirmed when animals were tested around 168 days of age. Adult subjugation of AS hamsters eliminated biting attacks. Following the first episode of adult subjugation there was never any recovery of biting behavior toward equal sized intruders in either the AS or AI groups.
There was a significant main effect for retreats (F(1,18) = 13.7, p < 0.01) and test trials (F(3,54) = 27.6, p < 0.001) (data not shown). AS hamsters tested at 48 days of age showed no signs of retreating from intruders as compared to their AI siblings (p < 0.01). At 108 days of age, AI hamsters were highly submissive as compared to their earlier behavior and to the behavior of their AS siblings (p < 0.01). Adult subjugation of AS hamsters also resulted in highly submissive behavior. This submissive behavior following the first episode of adult subjugation persisted through 168 and 222 days of age for both AS and AI groups regardless of life histories.
There was a striking similarity between the behavioral pattern of flank marking and the frequency of biting over the life history of both AS and AI groups. There was a significant main effect for the number of flank marks (F(1,18) = 42.6, p < 0.001) and test trials (F(3,54) = 44.5 p < 0.001). At 48 days of age, AS hamsters displayed over twice the number of flank marks as their AI siblings (p < 0.01). This high level of flank marking persisted at 108 days of age following the stress of adult isolation; however, their AI siblings for the first time displayed little or no flank marking (p < 0.01). Similarly, AS hamsters, when exposed to their first episode of subjugation as adults ceased to flank mark in the presence of an intruder.
Locomotor Activity, Anxiety, and Sexual Motivation
Data showing the effect of life history on locomotor activity in an open field, anxiety, and sexual motivation are shown in Figure 2. There was no significant main effect for motor activity (F(1,18) = 0.18, p > 0.5); however there was a significant main effect for trials over the six month testing period (F(3,54) = p < 0.001). While motor activity was similar for both AS and AI groups, there was a significant increase in activity at the older ages.
Figure 2 Measures of General Behaviors: Shown are the mean scores (+ SEM) for seed finding (time in seconds), motor activity (number of quadrants traversed in one min) and mount latency for hamsters initially subjugated in adolescence (black bar) as compared to their siblings isolated in adolescence (open bar) across a changing social history of adult subjugation (AS) or adult isolation (AI). (* p <.05; ** p < .01)
Stress-induced anxiety, as measured by the latency to find hidden seeds, was significantly different between the AS and AI groups (F(1,18) = 100.6, p < 0.001). There was also a significant main effect on seed finding across test trials (F(3,54) = 75.8, p < 0.001). There were no significant differences in seed finding across the life history of AS hamsters. In contrast, AI hamsters took almost four min to find the seeds as compared to about 40 sec for their AS siblings (p < 0.01). At 108 and 168 days of age the latency to find seeds was reduced but still significantly higher than the AS group (p < 0.01). It was not until the final and second adult subjugation that the latency to find seeds was comparable between AS and AI groups.
The latency to mount a receptive female as a measure of sexual motivation was significantly different between AS and AI groups (F(1,18) = 7.47, p < 0.05). However, there was no significant difference in mounting latency within each group across test trials (F(3,54) = 2.03, p > 0.1). At 168 days of age, AS hamsters, took less time to mount a receptive female than their AI siblings (p < 0.05). At 222 days of age, AS and AI groups showed the same latency in mounting behavior.
Neuroendocrine Measures
Data showing changes in blood levels of cortisol, testosterone and the androgen-sensitive flank glands are shown in Figure 3. The levels of cortisol between the AS and AI groups were significantly different (F(1,18) = 25.37, p < 0.0001). There was also a change of cortisol levels over time (F(3,54) = 9.32, p < 0.01). At 48 days of age both AS and AI siblings showed no appreciable release of stress hormone following an agonistic encounter with an adult intruder. This same blunted or suppressed stress response was observed for AS hamsters when they were tested at 108 days of age against an adult intruder (p < 0.01). In contrast, their AI siblings subjugated as adults showed significantly greater cortisol release following the interaction with an adult intruder. This pronounced difference disappeared at ages 168 after both groups had been exposed to adult subjugation. However, AI hamsters exposed to their second experience of adult subjugation at 222 days of age showed the highest cortisol levels of any sampling period (P < 0.05).
Figure 3 Measures of Steroid Hormone Changes: Shown are the mean plasma levels of cortisol and testosterone and the diameter of the androgen-sensitive flank glands (+ SEM). Measures are collected over a changing social history from hamsters initially subjugated in adolescence (black bar) as compared to their siblings isolated in adolescence (open bar). AS – adult subjugation; AI – adult isolation (* p <.05; ** p < .01)
The levels of testosterone between the AS and AI groups were significantly different (F(1,18) = 9.43, p < 0.01). There was also a significant change in testosterone across the test trials (F(3,54) = 74.1, p < 0.001). At 48 days of age, AS and AI siblings showed comparable levels of testosterone that were significantly lower than those measured at 108 days of age and older. At 108 days of age and older, AS hamsters showed no significant change in testosterone. While AI hamsters showed high levels of testosterone at 108 days of age following adult subjugation, their levels were significantly reduced at 168 and 222 days of age following experiences of isolation and subjugation, respectively (p < 0.01).
The size of the androgen-sensitive flank glands was significantly different between groups (F(1,18) = 8.15, p < 0.01) and across test trials (F(3,54) = 11.2, p < 0.001). AS siblings presented with the same sized flank glands over the course of their life history. In contrast, their AI siblings showed variable and significantly different sized flank glands across their life history. At 168 days of age their glands were significantly larger than at any age of the AS group's life history (p < 0.01).
Discussion
The anticipated changes in behavior with social subjugation noted in the literature for adult animals were not apparent in adolescent hamsters. For twelve consecutive days AS hamsters were placed into the home cage of a novel, larger, more aggressive hamster. As predicted, they were threatened and attacked each day over the thirty min encounter. Ten days after the cessation of subjugation the "table was turned" and these AS hamsters, now as young adults, became the residents and novel, larger hamsters became the intruders. In this context, and despite the social history of subjugation, AS hamsters were very aggressive. This display of aggression toward adult intruders was only heightened at three and half months of age as these AS hamsters grew into full maturity. In contrast, their AI siblings were far less aggressive as young adults, avoiding and retreating from adult intruders. When subjugated as fully mature adults, these AI hamsters took on a stable, submissive behavioral profile with no signs of aggression toward adult intruders. Adult subjugation was equally effective in promoting submissive behavior in AS hamsters. While adolescent subjugation favored future aggressive behavior, it did not protect these hamsters from the behavioral consequences of losing fights as adults. When socially subjugated as adults their fate was the same as their AI siblings – a stable, submissive behavioral profile.
Motor activity in the open field was essentially no different between AS and AI siblings over their life history. The latency to mount a receptive female as a measure of sexual motivation was not disrupted in either sibling group; although, there was a modest but significant delay in mounting time by the AI hamsters following their adult subjugation as compared to their AS siblings. The seed finding as a measure of anxiety showed an interesting pattern over the social history of the two sibling groups. AS hamsters are less anxious and attend to finding seeds in their home territory. Their AI siblings are more anxious but after repeated social subjugation in adulthood show the same attention to finding seed. From these observations, it would seem that socially subjugated male hamsters maintain behaviors necessary for survival despite a stable submissive phenotype that would limit their ability to compete for mates.
In an earlier paper, Delville et al. [25] reported social subjugation in adolescent hamsters increased biting attacks toward smaller intruders with reduced generalized aggression toward conspecifics of equal size. The study was designed to determine whether early traumatic stress in the form of social subjugation fostered inappropriate aggressive behavior. Indeed, the short bite latency and excessive number of bites toward non-threatening prepubescent submissive hamsters was judged to be inappropriate as compared to controls with no history of early abuse. However, when control and subjugated animals were tested against adult male intruders the level of biting attacks was very low or absent in both groups. Consequently, offensive aggression was reduced to the less specific behavioral measure of "attacks", or aggressive approaches and not as bite latency and number of bites as reported here. While there are probably many reasons for the high level of biting behavior toward adult intruders in our study versus the relative absence of biting attacks in the other, two possible explanations stand out. First, is the choice of the adult intruder. Delville and colleagues prescreened all male hamsters in adolescence for aggressive behavior. If animals showed any submissive predisposition they were eliminated from the population or used as small prepubescent intruders. Hence, the adult male intruders in their study may have presented as being more aggressive than typical males, discouraging biting attacks from the residents. The second explanation may be the value placed on the defense of the home territory by the resident. In our studies, resident's experienced temporary food deprivation associated with seed finding and access to a receptive female on the days prior to presentation of an unknown adult intruder. Under these circumstances, the level of offensive aggression toward intruders may be heightened as resident's fight to defend a territory perceived to have high reproductive potential and limited resources.
In a more recent study, Wommack and Delville [26] reported hamsters 42 days of age with a history of adolescent subjugation were highly submissive in the home territory of the larger more aggressive hamsters. These results are not surprising, as some level of reduced aggression or submission would be expected when smaller male hamsters are placed into the home cage of larger conspecifics [27]. In contrast, the present study has experimental and control siblings displaying offensive aggression to defend their own territory from adult male intruders. Given the context-dependent nature of aggressive responding it is difficult to compare the data between the two studies.
The anticipated suppression of plasma testosterone and elevation in stress hormone in adult animals following social subjugation did not occur in adolescent hamsters. Early adolescent stress followed by exposure to episodes of social and environmental stressors over adulthood produced an unexpected pattern of steroid hormone release in response to social conflict. The first unanticipated finding was the low plasma levels of testosterone (<0.5 ng/ml) in 48 day old AS and AI siblings following an aggressive encounter with an adult male intruder. Levels for this age are usually 2–4 ng/ml or near maximum for adult animals on long light/dark cycles [24,28]. Indeed, in our own laboratory we measured testosterone levels of 1–2 ng/ml in unstressed P-45 hamsters [29]. Both AS and AI siblings showed aggressive behavior toward the intruders. As noted above, the AS hamsters were far more aggressive than their AI siblings. While golden hamsters do not need testosterone to develop aggressive behavior [21], this steroid hormone augments aggressive responding and favors dominance behavior [30,31].
The low levels of testosterone during transition from adolescence to young adulthood may represent suppression of the hypothalamic-pituitary-gonadal axis caused by the stress of social subjugation or isolation in a novel environment. However, this is unlikely since both AS and AI groups showed normal adult-sized flank glands at 48 days of age. The flank glands are androgen sensitive; their size and pigmentation have been correlated with plasma testosterone levels [31]. Perhaps the low levels of testosterone are transient, decreasing over the 15 min period from start of social interaction to blood sampling. A significant decrease in plasma testosterone following defeat was noted in Siberian dwarf hamsters [32] and guinea pigs [33] within 10–15 min of the social encounter. Huhman and coworkers [8] noted a tendency for hamsters to show lower levels of testosterone 15–20 min after a single defeat but the trend was not significant. This robust suppression of testosterone levels following social conflict in 48 day old hamsters was not unique to their social history since both AS and AI siblings showed the same endocrine response. However, what is common to both social histories is adolescent stress.
Interestingly, fully mature hamsters tested at 108 days of age showed stable plasma levels of testosterone (ca. 3–4 ng/ml) after social conflict despite two opposite agonistic behaviors. AS hamsters were extremely aggressive, while their AI siblings were highly submissive. The endocrine response of the AI hamsters was unexpected since social subjugation in adult rodents reduces androgen levels [8,9]. Indeed, it was only after repeated exposure to episodes of subjugation and isolation stress over adulthood, did AI hamsters show significantly reduced (ca. 2 ng/ml) levels of testosterone that correlated with their submissive behavior. Why this reduction in testosterone was not observed following their first experience of social subjugation is unknown. Perhaps the isolation stress during adolescence provided some "immunity" to future stressors, sparing the hypothalamic-pituitary-gonadal axis. However, this early resilience is lost as hamsters are continuously exposed to chronic episodes of social and environmental stressors over adulthood. In contrast, AS hamsters show stable levels of testosterone over their adult social history. Although they show submissive behavior following subjugation in adulthood their testosterone levels are unchanged following defeat. Unlike their AI sibling, their "immunity" to chronic episodes of social and environmental stressors over adulthood persists.
Instead of the anticipated elevation in cortisol levels associated with social conflict, both AS and AI siblings showed a suppressed cortisol response (<0.3 ug/dL) following an aggressive encounter at 48 days of age. When fully mature at three and half months of age, AS hamsters showed the same blunted stress response following a successful attack on an adult intruder. This high aggression in the face of a blunted cortisol response is not unlike that reported by Halász and coworkers in rats [34]. Following adult subjugation, AI hamsters showed increased cortisol release associated with submissive behavior toward intruders. Their AS siblings showed the same release of cortisol, but only after adult subjugation.
Because of the 7–8 month duration of this study and the logistics of controlling for each developmental time point, there were limitations in the experimental design. Ideally, at each of the four test periods, 48, 108, 168, and 222 a sibling control group without any history of adolescent stress (subjugation or isolation) would have aided in interpreting developmental changes in behavior. However, it was not possible to get the required number (n = 40) male siblings from the same dams that provided AS and AI groups. This deficiency not withstanding, the behavioral results following adolescent subjugation and that following adult subjugation corroborate much of the previous work on hamsters using experimental animals with no history of winning and losing fights [6-8]. What are unique about these studies are the developmental consequences of repeated agonistic encounters in which the out come of winning or losing is predetermined.
Conclusion
Adolescence is a time of enhanced physical growth, driven, in part, by rising testosterone levels culminating in sexual maturity. Around P-25 male hamsters leave the maternal nest to forage for food and live alone in their own burrow. Losing fights might be very common for adolescent hamsters because they have not developed fully the size, strength and social behaviors to compete for food, mates and territory. If the results of repeated defeat in adolescence were a stable submissive phenotype toward all competitors then these hamsters might never have the opportunity to mate. The data presented here show that adult hamsters with a history of subjugation in adolescence can be very aggressive toward male conspecifics. Hence, adolescence may be a resilient period in development, protecting animals until they reach adulthood.
Methods
The breeding stock of Syrian golden hamsters was obtained from Harlan Sprague-Dawley Laboratories (Indianapolis, IN). All animals were housed individually in Plexiglas cages (24 cm × 24 cm × 20 cm), maintained on a reverse light:dark cycle (14:10; lights on at 19:00 hr) and provided food and water ad libitum. All animals were acquired and cared for in accordance with the guidelines published in the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publications No. 85-23, Revised 1985). The protocols used in this study were in compliance with the regulations of the Institutional Animal Care and Use Committee at the University of Massachusetts Medical School.
Female golden hamsters were bred in the animal facility at the University of Massachusetts Medical School. On P-23 male hamsters from each litter were individually housed. A diagram of the experimental procedure is provided in Fig 4. Four male subjects were sampled from each of five dams. These four were divided equally between either a adolescent subjugation (AS) group or adolescent isolation (AI) group. On P-26, the AS group was exposed to the stress of threat and attack by placing them into the home cage of a novel larger, experienced fighter for 30 min/day for 12 consecutive days while the AI group was exposed to the stress of a novel environment by placing them into a new clean cage for 30 min/day for 12 consecutive days. Handling and exposure of hamsters to a novel environment is a stressor as measured by release of cortisol [35]. Hence, these AI siblings control for the daily exposure to a novel environment of their AS siblings. It should be noted, hamsters exposed to subjugation have few if any physical injuries. Biting attacks very seldom cause in wounds.
Figure 4 Schematic Diagram of Experimental Procedure: Shown are daily schedules of subjugation stress (red, subjugation) and isolation stress (blue, novel cage) from birth to postnatal day 222 for two groups of 10 hamsters each, sampled from 5 dams. Ten days after the cessation of each stress paradigm there was a three day test period (test) during which hamsters were screened for behavior and bleed for steroid hormone measures.
On P-48, ten days after the cessation of stress, animals were tested and scored for a battery of behaviors over three consecutive days. As young adults these animals were again exposed to the stress of threat and attack of a large experienced fighter or isolation in a novel environment; however, the stressors were reversed. AS animals exposed to social subjugation during adolescence were now exposed to isolation while AI animals exposed to isolation in adolescence were exposed to social subjugation. These studies were begun ca. three months of age because age-dependent increases in aggression are stabile at this time [21]. On P-108, ten days after the cessation of stress, animals were again tested and scored for a battery of behaviors. The procedure was repeated twice more at approximately two month intervals for each group of animals reversing the stressor at each time.
All behavioral tests were performed during the first four hrs of the dark phase under dim red illumination, video taped and scored by an independent observer blind to the history of the animals. Screening began with seed finding a model for screening anti-anxiety drugs in golden hamsters [36]. Briefly, hamsters are deprived of food overnight. The following day they are exposed to the additional stress of being taken from their home cage and placed in a novel environment for a few minutes. During their absence from the home cage, sunflower seeds are hidden under the bedding in one of the corners. When returned to the home cage, hamsters routinely scramble along the walls for 3–4 min before settling down, locating and eating the seeds. However, animals treated with the traditional anxiolytics e.g., chlordiazepoxide, fluoxetine or buspirone find seeds in less than 20 sec. [36]. Seed finding was scored as the latency to find and eat or pouch a single seed.
Following seed finding, hamsters were tested for a sexual motivation by introducing a receptive, estrus female into their home cage. Females would routinely assume the lordotic posture and males were timed for their latency to mount and first intromission. Intromission was characterized by a constant rate of pelvic thrusting. On the second day animals were tested for general motor activity in an "open field." Animals were placed into a large clean Plexiglas cage (48 × 32 × 40 cm) devoid of any bedding. This open field was delineated into equal quadrants by tape on the underside of the cage. Animals were scored for motor activity by counting the number of quadrants traversed in 1 min.
On the third behavioral screening day, animals were tested for agonistic behavior in a resident/intruder paradigm. A male intruder of approximately the same size, weight and age was introduced into the home cage of the test animal and the resident scored for latency to bite the intruder, total number of bites, contact time, retreats and flank marks over a 10 min test period as previously described [37]. Flank marking is a form of olfactory communication in which a hamster arches its back and rubs pheromone producing flank glands against objects in the environment [38]. Flank marking frequency is greatly enhanced during aggressive encounters and is particularly robust in dominant animals initiating and winning fights [39].
Immediately after the end of resident/intruder encounter animals were lightly anesthetized with isoflurane and a venous blood sample of ca. 0.3 – 0.4 ml was collected by orbital eye bleed from the resident. The introduction of the hamster to the isoflurane followed by the sampling took less than 1 min. The samples was heparinized, centrifuged and assayed for testosterone and cortisol using a solid phase 125I radioimmunoassay (ICN Biomedical, Inc. Costa Mesa, CA). A total of four blood samples were collected from each animal over the 8–9 month study.
The data for all measures from both AS and AI groups over the four test periods were analyzed with a repeated measures two-way ANOVA followed by Newman-Keuls for post hoc comparisons.
Acknowledgements
This work was funded by a grant from the National Institute of Mental Health, Program in Behavioral and Integrative Neuroscience MH058700. The contents of this review are solely the responsibility of the authors and do not necessarily represent the official views of the NIMH.
==== Refs
Rose RM Berstein IS Gordon TP Consequences of social conflict on plasma testosterone levels in rhesus monkeys Psychosom Med 1975 37 50 61 1168355
Eberhart JA Yodyingyuad U Keverne EB Subordination in male talapoin monkeys lowers sexual behaviour in the absence of dominants Physiol Behav 1985 35 673 677 4080831 10.1016/0031-9384(85)90395-6
Frishknecht HR Seigfreid B Waser PG Learning of submissive behavior in mice: A new model Behav Processes 1982 7 235 245 10.1016/0376-6357(82)90038-9
Van de Poll NE DeJonge F Van Oyen HG Van Pelt J Aggressive behaviour in rats: effects of winning or losing on subsequent aggressive interactions Behav Processes 1982 7 143 155 10.1016/0376-6357(82)90023-7
Williams J Lierle DM Effects of repeated defeat by a dominant conspecific on subsequent pain sensitivity, open-field activity, and escape learning Anim Learn Behav 1988 16 477 485
Potegal M Huhman K Moore T Meyerhoff J Conditioned defeat in the Syrian golden hamster (Mesocricetus auratus) Behav Neural Biol 1993 60 93 102 8117243 10.1016/0163-1047(93)90159-F
Ferris CF Axelson JF Martin AM Roberge LR Vasopressin immunoreactivity in the anterior hypothalamus is altered during the establishment of dominant/subordinate relationships between hamsters Neuroscience 1989 29 675 683 2739905 10.1016/0306-4522(89)90140-1
Huhman KL Moore TO Ferris CF Mougey EH Meyerhoff JL Acute and repeated exposure to social conflict in male golden hamsters: Increases in plasma POMC-peptides and cortisol and decreases in plasma testosterone Horm Behav 1991 25 206 216 1648544 10.1016/0018-506X(91)90051-I
Bronson FH Eleftheriou BF Chronic physiological effects of fighting in mice Gen Comp Endocrinol 1964 4 9 14 14116580 10.1016/0016-6480(64)90033-4
Louch CD Higginbotham M The relation between social rank and plasma corticosterone levels in mice Gen Comp Endocrinol 1967 8 441 444 6068098
Ely DL Henry JP Neuroendocrine response patterns in dominant and subordinate mice Horm Behav 1978 10 156 169 29002 10.1016/0018-506X(78)90005-3
Eberhart JA Keverne EB Meller RE Social influences on circulating levels of cortisol and prolactin in male talapoin monkeys Physiol Behav 1983 30 361 369 6683410 10.1016/0031-9384(83)90139-7
Raab A Dantzer R Michaud B Mormede P Taghzouti K Simon H Le Moal M Behavioral, physiological and immunological consequences of social status and aggression in chronically coexisting resident-intruder dyads of male rats Physiol Behav 1986 36 223 228 3960994 10.1016/0031-9384(86)90007-7
McEwen BA Protective and damaging effects of stress mediators N Engl J Med 1998 338 171 179 9428819 10.1056/NEJM199801153380307
Coe CL Mendoza SP Levine S Social status contrains the stress response in the squirrel monkey Physiol Behav 1979 23 633 638 116264 10.1016/0031-9384(79)90151-3
Eberhart JA Keverne EB Meller RE Social influences on plasma testosterone levels in male talapoin monkeys Horm Behav 1980 14 247 266 7191826 10.1016/0018-506X(80)90033-1
Sapolsky RM Stress-induced suppression of testicular function in the wild baboon:role of glucocorticoids Endocrinol 1985 116 2273 2278
Spear LP The adolescent brain and age-related behavioral manifestations Neurosci Biobehav Rev 2000 24 417 463 10817843 10.1016/S0149-7634(00)00014-2
Dieterlen F Das Verhalten des Syrischen Goldhamsters (Mesocricetus auratus Waterhouse) Z Tierpsychol 1959 16 47 103
Schoenfeld TA Leonard CM Siegel HI Behavioral development in the Syrian golden hamster The Hamster: Reproduction and Behavior 1985 Plenun Press, New York 289 318
Whitsett JM The development of aggressive and marking behavior in intact and castrated male hamsters Horm Behav 1975 6 47 57 1168164 10.1016/0018-506X(75)90022-7
Festing MFW Hamsters The UFAW Handbook on the Care and Management of Laboratory Animals 1979 4 (Universities Federation for Animal Welfare, ED.), Williams & Wilkins, Baltimore, MD 242 256
Miller LL Whitsett JM Vandenbergh JG Colby DR Physical and behavioral aspects of sexual maturation in male golden hamsters J Comp Physiol Psychol 1977 91 245 259 858817
Vomachka AJ Greenwald GS The development of gonadotropin and steroid hormone patterns in male and female hamsters from birth to puberty Endocrinology 1979 105 960 966 477608
Delville Y Melloni RH JrFerris CF Behavioral and neurobiological consequences of social subjugation during puberty in golden hamsters J Neurosci 1998 18 2667 2672 9502824
Wommack JC Delville Y Repeated social stress and the development of agonistic behavior: individual differences in coping responses in male golden hamsters Physiol Behav 2003 80 303 308 14637229 10.1016/j.physbeh.2003.08.002
Payne AP Swanson HH The effect of sex hormones on the agonistic behavior of the male golden hamster (Mesocricetus auratus Waterhouse) Physiol Behav 1972 8 687 691 5064527 10.1016/0031-9384(72)90096-0
Bartke A Klemcke H Amador A Effects of testosterone, progesterone and cortisol on pituitary and testicular function in male golden hamsters with gonadal atrophy induced by short photoperiods J Endocrinol 1981 90 97 102 7264515
Ferris CF Shtiegman K King JA Voluntary ethanol consumption in male adolescent hamsters increases testosterone and aggression Physiol Behav 1998 63 739 744 9617993 10.1016/S0031-9384(97)00533-7
Vandenbergh JG The effects of gonadal hormones on the aggressive behaviour of adult golden hamsters Anim Behav 1971 19 589 594 5167836
Drickamer LC Vandenberrgh JG Colby DR Predictors of dominance in male golden hamster (Mesocricetus auratus) Anim Behav 1973 21 557 563 4795935
Castro WLR Matt KS The importance of social condition in the hormonal and behavioral responses to an acute social stressor in male Siberian dwarf hamster (Phodopus sungorus) Horm Behav 1997 32 209 216 9454672 10.1006/hbeh.1997.1423
Sachser N Short-term responses of plasma norepinephrine, epinephrine, glucocorticoid and testosterone titers to social and non-social stressors in male guinea pigs of different social status Physiol Behav 1987 39 11 20 3562643 10.1016/0031-9384(87)90338-6
Halasz J Liposits Z Kruk MR Haller J Neural background of glucocorticoid dysfunction-induced abnormal aggression in rats: involvement of fear- and stress-related structures Eur J Neurosci 2002 15 561 569 11876784 10.1046/j.0953-816x.2001.01883.x
Weinberg J Wong R Adrenocortical responsiveness to novelty in hamsters Physiol Behav 1986 37 669 672 3774898 10.1016/0031-9384(86)90170-8
King JA Messenger T Ferris CF Seed finding in golden hamsters: a potential animal model for screening anxiolytic drugs Neuropsychobiology 2002 45 150 155 11979066 10.1159/000054956
Ferris CF Melloni RH JrKoppel G Perry KW Fuller RW Delville Y Vasopressin/serotonin interactions in the anterior hypothalamus control aggressive behavior in golden hamsters J Neurosci 1997 17 4331 4340 9151749
Johnston RE Scent marking by male golden hamsters (Mesocricetus auratus). II. The role of the flank gland scent in the causation of marking Z Tierpsychol 1975 37 138 144 1146424
Ferris CF Axelson JF Shinto L Albers HE Scent marking and the maintenance of dominant/subordinate status in male golden hamster Physiol Behav 1987 40 661 664 3671532 10.1016/0031-9384(87)90114-4
| 15847686 | PMC1090605 | CC BY | 2021-01-04 16:38:34 | no | Front Zool. 2005 Apr 22; 2:7 | utf-8 | Front Zool | 2,005 | 10.1186/1742-9994-2-7 | oa_comm |
==== Front
Harm Reduct JHarm Reduction Journal1477-7517BioMed Central London 1477-7517-2-61583311610.1186/1477-7517-2-6ResearchUse of the femoral vein ('groin injecting') by a sample of needle exchange clients in Bristol, UK Maliphant John [email protected] Jenny [email protected] Bristol Drugs Project, 11 Brunswick Square, Bristol, BS2 8PE, UK2 Dept Pharmacy & Pharmacology, University of Bath, Bath, BA2 7AY, UK2005 15 4 2005 2 6 6 4 2 2005 15 4 2005 Copyright © 2005 Maliphant and Scott; licensee BioMed Central Ltd.2005Maliphant and Scott; 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
Use of the femoral vein for intravenous access by injecting drug users (IDUs) (commonly called 'groin injecting') is a practice that is often observed but on which little is written in the literature. The purpose of this study was to describe self-reported data from a sample of groin injectors on the natural history and rationale regarding their groin injecting, to inform future research and the development of appropriate harm reduction strategies.
Methods
A convenience sample of groin injectors willing to participate in a semi-structured interview were recruited through the Bristol Drugs Project Harm Reduction Service. The interviews were conducted over the period of one week. Data on transition to groin injecting, rationale for use and incidence of problems were collected.
Results
Forty seven IDUs currently injecting in their femoral vein ('groin') were interviewed, 66% (n = 31) male and 34% (n = 16) female. Their mean age was 31 yrs (range 17 to 50 yrs; SD = 7.7). The mean length of time since first injecting episode was 9.6 yrs (range 6 mths to 30 yrs; SD = 7.0). The mean length of time since use of the groin began was 2.6 years (range 1 mth to 15 yrs; SD = 3.3). The mean length of time between first injection and first use of the groin was 7.0 yrs (SD = 7.0). One person had used no other area for venous access prior to using the groin, nine people had used one, nine people had used two, 10 people had used three, five people had used four and 13 people had used more than four areas. The main reason given for starting to inject in the groin was that 'no other sites were left'. However further discussion identified this meant no other convenient sites were accessible. Practises such as the rotation of injecting sites, as advocated in many harm reduction leaflets, were reported to be difficult and unreliable. The risk of missing the vein and subsequently losing the 'hit' was considered high. Use of the non-dominant hand to administer injections was problematic and deterred rotation between arms. The groin site was reported to be convenient, provide quick access, with little mess and less pain than smaller more awkward veins. The formation of sinuses over time facilitated continued use of the groin. Approximately two thirds of participants had experienced difficulty gaining IV access at their groin. Common problem included scar tissue occlusion, swelling and pain. Some reported infections and past history of deep vein thrombosis.
Conclusion
Use of the groin was perceived to be convenient by the study group. Problems following safer injecting advice were identified, including dexterity difficulties leading to fear of losing the 'hit'. Health problems at the groin site did not deter use. These results suggest further qualitative work is needed to explore the difficulties in following safer injecting advice in more detail and inform the development of more appropriate advice. Further quantitative work is necessary to establish the prevalence of groin injecting amongst IDUs and the incidence of associated problems. There is a need for a longitudinal study to examine the relationship between injecting technique and loss of patency of veins. If protective factors could be identified, evidence-based safer injecting advice could be established to preserve peripheral veins and reduce use of the groin site, which is high risk and associated with serious adverse consequences.
==== Body
Background
The physical health complications of injecting drug use are well documented in the literature (e.g.[1-6]). The injections used by injecting drug users (IDUs) are non-sterile and not subject to quality control. This, coupled with frequent, chronic venous administration is associated with damage to the vascular structure. Vascular damage commonly begins with thrombophlebitis, leading to vein sclerosis and loss of patency [1,3], rendering the vein unusable. This leads to the IDU seeking other useable points of intravenous (IV) access.
There is little research literature reporting patterns of vascular access in IDUs. If the natural history of IDU injecting patterns was better understood, effective strategies to protect vascular health could be explored. A paper by Darke et al [7] describes a pattern of use of various injecting sites over time, identified amongst a sample of injecting drug users in Sydney, Australia. The authors report most injectors began their injecting careers using the cubital fossa (inner crook of the arm), with a pattern of progression through forearm (after a median of two yrs from first injection), upper arm (3.5 yrs from first injection), hand (4 yrs from first injection), neck, feet, leg (all 6 yrs after first injection) and finally groin and peripheral digits (both 10 yrs from first injection). This suggests that use of the groin amongst this sample was reserved as a 'last resort' with other points of access being selected first.
The use of the femoral vein in the groin by IDUs is of concern. It is linked with increased risk of vascular complications such as deep vein thrombosis (DVT), leg ulcers and vascular insufficiency. Its close proximity to the femoral artery and nerve also poses the risk of inadvertent trauma to these sites. Arterial injection is associated with arterial spasm and arterial thrombus formation [8].
The literature contains reports of adverse consequences from use of the groin site but little qualitative study of the factors that motivate this practice. The primary purpose of this study was to inform service development at Bristol Drugs Project (BDP) and compare the findings with that of Darke et al5. However the findings of this work are of interest to the wider harm reduction community because little is written in the literature about groin injecting. This study begins to shed some light on factors that motivate this practice. It also suggests future areas for research around groin injecting in order to inform the development of evidence-based safer injecting advice.
Methods
Location
This study was undertaken in Bristol, which is the largest city (pop. 382,000) in the South West region of England, UK. Participants were clients of BDP, which is a voluntary sector drug service. BDP is the only needle exchange and harm reduction agency in the city, but there are also pharmacy-based exchanges.
Recruitment
A convenience sampling method was used. Willing participants were recruited from attendees at BDP needle exchange base which is a fixed site service, and the Mobile Harm Reduction Service, which is a vehicle providing outreach needle exchange services across the city. Data was collected over a period of one week in 2004 by the same interviewer in all cases. All clients who used the needle exchange services staffed by the interviewer were invited to take part in the study. Participants were guaranteed strict confidentiality and data was collected anonymously.
Data collection
Data was gathered using a short semi structured interview based on a series of questions derived from previous discussions amongst needle exchange clients and staff. It explored injecting history and whether the person had ever or was currently experiencing problems using their groin. The study was reviewed and approved by the BDP management board. The interview was conducted after the needle exchange transaction was completed. Verbal consent was obtained. Data was recorded on a tick-box data collection form by the interviewer and by additional note writing.
Analysis
Data was coded and input into SPSS for Windows (v. 12, SPSS inc. Illenois, 2003) for analysis where appropriate. Descriptive data was analysed to identify emergent themes.
Results
Incidence of use of the groin and demographics
The interview took approximately 10 minutes. A total of 92 clients were interviewed as part of the wider review and 47 (51%) of these were currently injecting in their groin. None of those who were not injecting in the groin presently had ever done so in the past. Of those injecting in their groin, 66% (n = 31) were male and 34% (n = 16) were female. The mean age of the groin injectors was 31 yrs (SD = 7.7), with the youngest being 17 and the oldest being 50 yrs. Twelve (26%) of the groin injectors were between 17–24 yrs, 29 (62%) were between 25 and 39 yrs and 6 (13%) were between 40–50 yrs.
Length of time injecting
The mean length of time since first injecting episode was 9.6 yrs (SD = 7.0), with the shortest time since first injecting episode being 6 months and the longest time being 30 yrs. Seventeen (36%) had been injecting 5 yrs or less, 12 (26%) had been injecting between 6 and 10 yrs, 10 (21%) had been injecting 11–15 yrs, 6 (13%) had been injecting 16–20 yrs, one (2%) had been injecting for 25 years and one (2%) for 30 years.
Length of time of groin injecting
The mean length of time since use of the groin site began was 2.6 years (SD = 3.3), with the shortest time being 0.08 years (1 month) and the longest time being 15 years. The mode length of time was 2 years, reported by 6 people (13%).
Time from first injecting episode to first use of the groin
The mean length of time between first injection and first use of the groin for IV access was 7.0 yrs (SD = 7.0). Two people (4%) had been using the groin for IV access since they first began injecting, both were male. One had tried to inject into the arms unsuccessfully prior to using the groin, so switched to the groin straight away. This person was very thin with no visible veins, so chose to use the groin to ensure IV access. The other person had not tried any other injecting sites prior to the groin, as all his associates who were injecting at the time were using the groin, encouraging him not to attempt to try any others first. Twenty five people (53%) had begun using the groin within 5 years from their first injection and 11 (23%) had begun using the groin 5 or more yrs but less then 10 years since their first injection. Eleven people (23%) had begun using the groin 10 or more yrs since their first injection. The longest time between first injection and use of the groin was 23 yrs in a male who had begun injecting 25 yrs ago but only started using the groin 2 yrs ago.
Areas used prior to the groin
People were asked to report which areas they had injected into prior to using the groin. One (2%) person had used no other area for venous access prior to using the groin and had been using this site for 15 years. Nine (19%) people had used one other area and in all cases this was the cubital fossa (inner crook of the arm). Nine (19%) people had used two areas prior to the groin, all of these cases had used the cubital fossa, seven had also used site(s) on the legs, one had used the foot and one had used the neck. Ten (21%) people had used three areas prior to using the groin. Again in all cases the cubital fossa had been used, nine of the ten had used sites on the legs, six had used the feet and five had used the neck. Five (11%) people reported injecting in four areas prior to the groin. All had used the same four areas, which were the cubital fossa, legs, feet and neck. Thirteen (28%) people had used more than four areas prior to the groin and classed themselves as having used 'everywhere'.
Why did you start injecting in your groin?
The main reason given for starting to inject in the groin was that 'no other sites were left'. However as many people had not tried all other sites, this was probed further. Further discussion found that in the majority of cases, no other convenient sites were perceived to be accessible. Many reported that practises such as the rotation of injecting sites, as advocated in many harm reduction publications, were found to be difficult and unreliable. The risk of 'losing' an injection (missing IV access) through poor injecting technique was considered to be too big a risk, presumably because subcutaneous and intramuscular drug absorption does not provide the same euphoria. Use of the non-dominant hand to administer injections was also reported to be difficult and deter rotation of injecting sites between arms, or require third party assistance. The groin site was reported by most to be convenient, provide quick access, with little mess and less pain than smaller more awkward veins. The gradual formation of sinuses in the groin over time was reported to further facilitate continued use of this site.
Drugs injected into the groin and equipment used
The most common drug injected by the group was heroin, used by 46 (98%) of interviewees. Nineteen people (40%) injected crack cocaine and eight (17%) injected amphetamine. Twenty four people (51%) currently injected one main drug only into the groin, with 23 injecting heroin and one injecting amphetamine. Twenty people (43%) injected two main drugs into the groin, for 16 of these people their main drugs were heroin and crack cocaine. The remaining four injected heroin and amphetamine. Three people injected three main drugs into the groin and for all these were heroin, crack cocaine and amphetamine. No other drugs were reported to be injected into the groin within the group.
The most common injecting equipment used to access the groin was detachable 1 ml syringes with orange needles (0.5 × 25 mm, 25G) used by 33 people (70%), 11 people (23%) used the same syringes with blue needles (0.6 × 30 mm, 23G) and one person (2%) used green needles (0.8 × 40 mm, 21G). Seven people (15%) used 1 ml insulin syringes. Numbers exceed 100% as four people regularly used more than one type of equipment for groin access.
Condition of the groin site and history of access problems
Participants were asked whether they were currently or had in the past experienced any problems gaining IV access using the groin site. They were also asked to self-assess the current condition of their groin based on a five point Likert scale: 'very poor', 'poor', 'OK', 'good' or 'very good'.
Approximately one third of people reported never having had a problem gaining IV access at their groin site (n = 16, 34%). This group comprised 11 males and five females. Five described the current state of their groin site as 'ok', five said it was 'good' and six said it was 'very good'. Their mean length of use of the groin site was 1.1 yrs (SD = 1.2). Approximately two thirds of people had experienced problems with IV access at the groin site on one or more occasions in the past, or were currently experiencing problems (n = 31, 66%). This group comprised 20 males and 11 females. When asked to describe the current condition of their groin two said it was 'very poor', seven said it was 'poor', 14 described it as 'ok', three said it was 'good' and five said it was 'very good'. Their mean length of use of the groin site was 3.3 yrs (SD = 3.8).
When asked to describe the types of access and health problems experienced, a common problem reported was hardened scar tissue occluding the site. This was said to be difficult to penetrate and a cause of needles bending and breaking, causing some people to select longer, thicker needles. Another common problem was swellings in the groin area, accompanied at times by pain. Some people reported infections at the injecting site. Some reported having experienced 'blood clots' or 'DVT' (deep vein thrombosis). It is unknown whether these had been medically diagnosed or treated.
Discussion
It is of interest that in the overall sample (n = 92) all those who had tried groin injecting (n = 47) continued to do so, despite two thirds having experienced problems with access and a range of health problems at the site. These included reports of infections, hardened tissue, swelling and DVT, which is consistent with problems described in the literature. Further exploration of factors that discourage use of the groin amongst non groin injectors would be of interest. Comparisons between vein health and injecting practices of groin injectors and those who do not inject in the groin would establish protective factors. A longitudinal study is necessary for this to establish the factors that protect and damage the patency of veins. Modifiable protective factors could inform safer injecting advice. The average length of time of groin injecting amongst the sample was 2.6 years with the longest time being 15 years, illustrating that this site of access can be used for considerable time. The formation of sinuses around hardened scar tissue was seen to be an advantage and facilitated continued use, despite posing the risk of breaking needles. Further work to explore responses to problems with the groin site and long-term consequences would be of interest.
Darke et al [7] reported an average time of 10 years from first injection to use of the groin site in their sample of IDUs in Sydney. In this study the average length of time from first injection to use of the groin site was 7 yrs, with more than half (53%) of participants having begun groin injecting within 5 yrs. One theory for the earlier use of the groin site in Bristol is that the use of acidifiers such as citric acid, necessary to dissolve the brown base heroin common in Western Europe, shortens the usable life of veins. Acidifiers are not used in Australia as street heroin is in a soluble form. However further work would be needed to confirm is this theory is true.
A pattern of use of various sites prior to the groin site was found in the majority in this study, similar to the findings of Darke et al [7]. However in this study there were some for whom rapid progression to use of the groin occurred, for example nine participants had only used one site, the cubital fossa, prior to the groin and a further nine had only used two sites. The qualitative data gathered from the sample illustrated that the groin was viewed as 'easy-to-use' with more security of delivering the injection intravenously than other, more awkward sites. This study showed that the groin site was favoured over others for utility and convenience. Other useable sites did potentially exist in many, such as the cubital fossa of the dominant arm, but were viewed as difficult to use and risked loss of the injection. A need for safer injecting information was highlighted in many in this study and practice within the agency has been developed to address this. Future work should examine decision making around use of the groin and whether information on health risks coupled with support to access more awkward peripheral veins can deter use of the groin. However caution is needed not to promote use of sites that require third party assistance, as this may reduce the level of control the IDU has over the injecting process and increase the risks of transmission of hepatitis C and other blood-borne pathogens.
Several points can be learned from this work to inform the delivery of harm reduction messages to this client group:
1. Recognition of the importance of utility and convenience when selecting an injecting site. Had this study not enquired about previous sites used and probed those who said they had 'no other sites' left, it may have been wrongly assumed that they did indeed have no useable sites. The identification of the 'utility and convenience' factor and the difficulties in using the non dominant hand for drug injecting has prompted the authors to consider the implementation of structured safer injecting training for IDUs, in order to deter use of the potentially high risk groin site. Such training, run by experienced nurses or anaesthetists, could develop injecting skills amongst IDUs in order to improve injecting techniques and promote the use of other available peripheral sites on the upper limbs. Such services could be integrated within safer injecting facilities.
2. The data on choice of injecting equipment is encouraging, as the majority of participants used detachable needles, which are intended for intravenous access. Most also chose to use short needles (orange), which it is believed to reduce the extent of vascular assault when injecting. This practice has been promoted by needle exchange workers locally. However the identified use of insulin syringes and longer needles (e.g. blue) in some is of concern. Insulin syringes are fragile and intended for subcutaneous use only hence carry a risk of breaking, especially if scar tissue forms that is tough to penetrate. Longer needles may increase the assault on the vascular system or increase the risk of injuring the surrounding nerves or arteries.
3. The contribution of the quantities, frequency of injecting and poly drug use to vascular damage should be studied. Due to tolerance to the effects of many psychoactive drugs, injectors often use increasing quantities and inject with increasing frequency as their injecting careers lengthen. Some also progress to poly drug use. Just over half of this sample (51%) injected one drug only, and in all but one cases this was heroin with the remaining case being amphetamine. However of the remaining 49% (n = 23) who injected more than one drug, 19 were injecting crack cocaine. Cocaine is a potent local anaesthetic and could potentially increase the risks of using the groin site (and other sites) due to lack of sensation on injecting. Further work is needed to quantify and explore these risks and also to assess the longevity of intravenous access in relation to single and poly drug use and frequency of use.
This study focused on the practices of groin site injectors in Bristol using the BDP needle exchange services, illustrating past and current injecting practices, identifying learning points for safer injecting advice delivery and future research. A convenience sample was used and the sample size was dictated by willingness to participate in the given time frame of the study. The results should not be extrapolated to the rest of the UK or those not in contact with BDP.
Conclusion
This study found, amongst a convenience sample of IDUs in Bristol, that the average time from first injection to use of the groin site was 7 yrs, with the majority having begun use of this site within five years. Reasons for use of the groin site centred on utility, convenience, ease of use and reduced risk of losing the euphoric effects due to extravenous delivery. Several key points were generated to inform the BDP harm reduction service, including the idea of developing safer injecting workshops for IDUs and encouragement that messages on use of equipment were successful. Several areas for future research have been prompted by this study. Groin site injecting is a risky practice that appears to have had little mention in the literature other than the reporting of case studies. The health risks are significant therefore further work to better understand this practice amongst IDUs and how to deter it would be of benefit.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
The questionnaire was developed on the basis of discussion between the two authors and staff at BDP. JM, who is employed by BDP, conducted the interviews. JS entered and analysed all the data and drafted the first version of this paper. Both authors contributed towards the revision of this paper and production of the final draft.
Acknowledgements
Sincere thanks go to all the service users who took part in the interviews and to Dr M Weiss, University of Bath for her comments on an earlier draft.
==== Refs
Cherubin C Sapira JD The Medical Complications of Drug Addiction and the Medical Assessment of the Intravenous Drug User: 25 years later Annals of Internal Medicine 1993 119 1017 1028 8214979
Selwyn P Illicit Drug Use Revisited; What a Long, Strange Trip It's Been Annals of Internal Medicine 1993 119 1044 1046 8214983
Stein MD Medical Complications of Intravenous Drug Use Journal of General Internal Medicine 1990 5 249 257 2187962
Haverkos H Lange R Serious Infections Other Than Immunodeficiency Virus Among Intravenous Drug Abusers The Journal Of Infectious Diseases 1990 161 894 902 2182730
Levine DP Levine DP, Sobell JD Skin and Soft Tissue Infections in Intravenous Drug Users Infections in Intravenous Drug Users 1991 Oxford University Press, New York 183 207
Morrison A Elliott L Gruer L Injecting-related Harm and Treatment-seeking Behaviour Among Injecting Drug Users Addiction 1997 92 1349 1352 9489051 10.1046/j.1360-0443.1997.9210134911.x
Darke S Ross J Kaye SL Physical Injecting Sites Among Injecting Drug Users in Sydney, Australia Drug and Alcohol Dependence 2001 62 77 82 11173170 10.1016/S0376-8716(00)00161-7
Woodburn KR Murie JA Vascular Complications of Injecting Drug misuse British Journal of Surgery 1996 83 1329 34 8944446
| 15833116 | PMC1090606 | CC BY | 2021-01-04 16:36:50 | no | Harm Reduct J. 2005 Apr 15; 2:6 | utf-8 | Harm Reduct J | 2,005 | 10.1186/1477-7517-2-6 | oa_comm |
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-261583313810.1186/1477-7525-3-26ResearchDevelopment and validation of the Chinese Quality of Life Instrument Leung Kwok-fai [email protected] Feng-bin [email protected] Li [email protected] Ji-qian [email protected] Kelvin [email protected] Li-zhu [email protected] Department of Occupational Therapy, Queen Elizabeth Hospital, Kowloon, Hong Kong SAR, China2 The First Affiliation Hospital, Guangzhou University of Traditional Chinese Medicine, Guangzhou, China3 Research & Development Division, School of Chinese Medicine, Hong Kong Baptist University, Hong Kong SAR, China4 School of Public Health, Sun Yat-Sen University, Guangzhou, China2005 16 4 2005 3 26 26 21 9 2004 16 4 2005 Copyright © 2005 Leung et al; licensee BioMed Central Ltd.2005Leung 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
This paper describes the development of the Chinese Quality of Life Instrument (ChQOL) which is a self-report health status instrument. Chinese Medicine relies very much on asking subjective feelings of patients in the process of diagnosis and monitoring of treatment. For thousands of years, Chinese Medicine practitioners have accumulated a good wealth of experiences in asking questions about health of their patients based on the concept of health in Chinese Medicine. These experiences were then transformed into questions for the ChQOL. It is believed that ChQOL can contribute to the existing Patient Report Outcome measures. This paper outlines the concept of health and disease in Traditional Chinese Medicine, the building of the conceptual framework of the ChQOL, the steps of drafting, selecting and validating the items, and the psychometric properties of the ChQOL.
Methods
The development of the ChQOL was based on the concept of health in Traditional Chinese Medicine with a theory driven approach. Based on the results of literature review, the research team developed an initial model of health which encompassed the concept of health in TCM. An expert panel was then invited to comment and give suggestions for improvement of the initial model. According to their suggestions, the model was refined and a set of initial items for the ChQOL was drafted. The refined model, together with the key domains, facets and initial items of the ChQOL were then mailed to a sample of about 100 Chinese medicine practitioners throughout Mainland China for their comments and advice. A revised set of items were developed for linguistic testing by a convenience sample consisting of both healthy people and people who attended Chinese Medicine treatment. After that, an item pool was developed for field-testing. Field test was conducted on a convenience sample of healthy and patient subjects to determine the construct validity and psychometric properties of the ChQOL.
Results
Construct validity was established by various methods, i.e. the internal consistency in all facets and domains were good; the correlation between facets to domain, and domains to overall ChQOL correlation were high; confirmatory factor analysis showed that the structure fitness of all facets, domain and overall structure were good with CFI > 0.9. Test-retest reliability was also good, especially in the domain scores with ICC value ranging from 0.83 to 0.90. No ceiling or floor effect was noted which indicated that ChQOL can be applied to subjects with a wide range of health status. Most facet scores, domain scores and the overall CHQOL scores were able to discriminate groups of subjects with known differences in health status. The ChQOL had mild positive convergence with the other generic health related QOL measures, i.e. the WHOQOL-100 and the SF-36, with moderate correlations.
Conclusion
In conclusion, the study indicated that the ChQOL is conceptually valid with satisfactory psychometric properties. It can provide additional information on health and QOL on top of the existing generic health related QOL measures. Furthermore, it forms basis for further testing and applications in clinical trials.
Quality of lifeself-reported health statustheory driven approachYing-yang, structure fitness, validation, psychometric propertiesthe Chinese Quality of Life instrument
==== Body
Background
Traditional Chinese Medicine (TCM) has been practiced in China for thousands of years. Throughout history, millions and millions of Chinese people have been cured by various forms of TCM, including the use of Chinese herbal medicine, acupuncture, and Qigong, etc. In the recent few decades, the use of Chinese medicine (CM) is becoming more popular on its own, and also as a kind of complementary treatment to Western Medicine throughout the world [1]. In the West, much attention is paid on the use of CM in the alleviation of painful symptoms and undesirable side effects of radical pharmaceutical and radiological treatments.
TCM practitioners do not rely on biomedical and radiological examinations for their diagnosis and treatment of patients. (Modern CM practitioners may take biomedical and radiological examinations as reference) They rely heavily on observable signs and subjective feelings of patients. The four fundamental techniques that have been used by CM practitioners for thousands of years for diagnosis and treatment are: observation (looking for visible signs), auscultation and olfaction (listening and smelling), interrogation (asking questions), and palpation (felling the pulse) [2,3]. Throughout the centuries, CM practitioners have accumulated valuable experiences on capturing very detailed and accurate information about the signs, symptoms and feelings of their patients. We believed that these experiences may be useful in applying in health related quality of life (HRQOL) measures and other forms of patient reported outcome measures.
Traditionally, Western Medicine and Chinese Medicine are using very different ways in testing their clinical efficacy. Randomized clinical trial (RCT) technique is a key method adopted in Western Medicine. In Chinese Medicine, clinical efficacies are observed, categorized and recorded descriptively in the literature. In the past fifty years, many scholars in Mainland China and in the Western countries in North America, Europe and Australia, etc, have been studying CM with RCT methods. The scientific study on the efficacy of CM using RCT method is becoming more and more popular. QOL indicators were used as secondary outcome indicators in many of these studies. However, the efficacies of CM on QOL were not particularly evident in these studies. There is a general impression among CM practitioners that the currently available QOL instruments may not be sensitive enough to detect the health changes that is regarded as important in CM. For examples, in CM, appetite and digestion, routine of urination and bowel, facial and lips colour, spirit in the eyes, and adaptation to climates and seasons are very important indicators of health status. However, these indicators are usually not included in common HRQOL measures.
A group of CM practitioners and QOL researchers in Hong Kong and Mainland China came together and worked towards a mission of developing a HRQOL instrument, the Chinese Quality of Life Instrument (ChQOL), which embedded the concept of health in TCM. We believed that the experience of CM might further enhance the advancement of quality of life assessment and research. We also believed that a HRQOL instrument that was developed basing on the concept of health in TCM might be more sensitive to the efficacy of CM. Besides adopting a theory driven approach, we had applied a set of standardized procedures that was commonly recognized and accepted in Western Medicine in the development of the ChQOL [4].
Method
Establish the initial ChQOL structure
Based on the results of literature review on the concept of health and disease in TCM, the research team developed an initial model of the ChQOL which encompassed the concept of health in TCM. The model consisted of the essential domain of HRQOL. Then, an expert panel consisted of several CM scholars from different fields of practice in Chinese medicine were formed. These experts came from various areas of Chinese Medicine, including clinical experts, pharmaceutical experts and academics who study theories of TCM. A half-day expert panel meeting was held. The panel was asked to comment on the initial domain model, and to proposed strategies on developing facets under the domains. The results of the penal meeting were used to produce the domain and facet structure that guided the drafting of items.
Draft items
Items of the ChQOL were supposed to be indicators of health in TCM. They were written in wordings that were commonly used in the communication between CM practitioners and their patients in soliciting clinical information. The items were phrased in form of questions asking about intensity, frequency, capacity or satisfaction of signs, symptoms, and feelings where appropriate. Subjects could rate the items with a set of five-point ordinal scales with descriptors. For each of the facet, 4 to 8 items would be drafted.
Review of the drafted items by clinician
The drafted items, the domain and facet structure of the ChQOL, were then mailed to a convenience sample of about 100 CM practitioners throughout Mainland China for their comments and advice. They were asked to give their comments in a standard reply form indicating if they agreed on the domains, facets and the items. They were asked to give suggestions to improve the items and to add new items.
With the input from these CM practitioners, a revised set of items were developed for the cognitive debriefing step. A small convenience sample consisted of both healthy people and patients, who were receiving CM treatment, were recruited. They were asked to comment on the linguistic and semantic clarity of the items. They were also asked to suggest improvement in the wordings of the items. After that, an item pool was developed for field-testing.
Select response scales
The 5-points response scales in the Chinese version WHOQOL-100 were adopted in the ChQOL. There were several types of response scales that suit intensity, frequency, capacity, and satisfaction questions. The scales were developed in a way that the 3 descriptors between the two anchor points lie approximately at the 25%, 50% and 75% of the scale. The WHOQL-100 responses scales were assumed to be interval scale for data analysis [5,6].
Facet and domain scores
The response scales ranged from 1 to 5. In most items, score 1 referred to the lowest QOL and 5 referred to the highest QOL. In some items, reverse of polarity might be required before formulating the facet scores.
In formulating the facet scores, two methods were tested. The first method assumed that the weights of the items in the same facet were the same. Raw facet score was derived by summing the item scores in the facet and then transformed to a 0 -100 facet scale. The same way was used to derive the domain scores. In the second method, the coefficients between the items and the corresponding facets, obtained from the structural equations, as the weights for the items. Raw facet scores were derived by summing up the product of individual item score and the corresponding coefficient. The raw facet score was transformed to a 0 -100 facet scale. The same ways were used to derive the domain scores.
Field test
A convenience sample of healthy and patient subjects was recruited to test the domain and facet structure and to examine psychometric properties of the ChQOL. In the field test, all the subjects were asked to fill in a set of questionnaire which consisted of the items in field test version of the ChQOL, the Chinese version SF-36 [7,8], and the Chinese version WHOQOL-100 [7,9]. The subjects were asked to comment on their own health status by answering a single question on self-perceived health status, i.e. rate their health status as "very bad" or "bad" or "neither good nor bad" or "good" or 'very good" Further, they were also asked to comment on their illness condition by responding "having no illness" or "having a stable chronic illness" or "having an acute illness". Subjects were asked to provide personal particulars as well. Assistances were provided to those who had difficulties in reading and writing due to various reasons.
Build the final ChQOL structure
Exploratory factor analyses using Principle component extraction method with Varimax rotation and Eigenvalue >1, were used to examine the items in each of the domains. The purpose was to select items that best represent the facets, and to reduce items that did not fit in well with other items in any facets under the domain. Confirmatory factor analyses were then conducted to confirm the item-facet and facet-domain structure of the domains. Comparative fit index and the Chi square value were reported to show the structure fitness. The facet scores were used to test the overall facet-domain structure of the ChQOL.
Examine psychometric properties of the CHQOL
A series of analysis were conducted to test the psychometric properties of the ChQOL. Distributions of the facets and domain scores were examined and reported in terms of range, means, and standard deviation. The existence of ceiling and floor effect were also tested. Correlation between facet and domain scores, and inter-domain scores were examined. Pearson correlation coefficients were reported. Internal consistency of the facets was tested by examining the Cronbach's alpha value. Test-retest reliability of the facet and the domain scores were examined by a sub-sample within 2 days and the ICC(1,1) values were reported. Convergent validity was established by correlating the domain scores of the ChQOL with the 6 domain scores of the WHOQOL-100 and the 8 facet scores of the SF-36. ANOVA tests were used to examine the discrimination validity of the ChQOL. It was performed by checking if the domain scores of the ChQOL can discriminate the self-perceived health status, and the illness status of the subjects.
Study conceptual overlapping with the WHOQOL-100 and SF-36
Exploratory factor analyses were conducted with the facet scores of the ChQOL and two commonly used generic HRQOL measures separately. The two measures were the Chinese version WHOQOL-100, and the Chinese version SF-36. The purpose was to examine if there were overlapping of the facets at the domain level in the three measures.
Statistical software
SPSS version 11.0 was used in data analysis [10]. The software EQS version 6.5 b was used in examining structure fitness of various structure models. [11].
Results
Concepts of Health in TCM
The first task of the study was to explore the theoretical bases for the ChQOL by reviewing both ancient and modern CM literatures. We could not find an established theory on "concept of health in Chinese Medicine" in the TCM literature. Health state is being described in various TCM theories in different contexts. Our conceptual model for the ChQOL had to be deducted from our understanding of the core concepts of TCM.
The fundamental theoretical basis of TCM originates from the ancient Chinese philosophy of Yin-yang [2,3]. The theory of Yin-yang can be understood as a conceptual tool for man to understand the order, principles and relationship of the universe, nature, objects and every day life phenomena. The nature, objects and day-to-day phenomena can be classified into two opposite but complementary categories, which are understood from two opposite perspectives, i.e. Yin and Yang. Although Yin and Yang are opposite to each other, they are mutually dependent and mutually transformable. Further, Yin and Yang are mutually restrictive and interactive, and in equilibrium with each other. Examples in nature are: earth to sky, moon to sun, water to fire, coldness to warmth, descending to ascending, static to motion, material to function, etc. Examples of Yin-yang in human are: woman to man, body organ to body function, physical form to spirit, etc. Examples in disease states are: exterior to interior syndrome, cold to heat syndrome and deficiency to excessive syndrome, etc.
Under the theory of Yin-yang, health is a state of harmonization of the "Physical Form" and "Spirit" (), harmonization of "Man" and "Society" (), and harmonization of "Man" and "Nature" (or environment) () [12].
Physical form (Xing, ) refers to the bodily structure of the human beings. Spirit (Shen, ) has many meanings. It can also refer to the general outward manifestations of the life activities and the mental activities, which include consciousness and reasoning of human mind. It can also refer to the functional manifestations of the changes in the body and their intrinsic laws. In accordance with CM, the physical form of the body is the bases for producing the spirit, while the spirit can regulate and control the body. They depend on each other for their existence, and the unity of the body and spirit is the main assurance of one's survival [2].
The concept of the Seven-Emotion refers to various human emotions. It can be regarded as emotional consequences of the interactions between an individual with the external environment, including both human and natural environment. It can also be the manifestation of the internal health status of an individual. According to this theory, human being has 7 basic emotions, which are: joy (), anger(), worry(), pensiveness(), grief(), fear(), and anxiety(). The 7-emotions theory stresses that emotion and body are linked and mutually affected. Therefore, emotions can be the determinants of health as well as the indicators of health in a different context [2,13]. In TCM literature, the relationship between emotion and health was elaborated in details. Liver relates to anger, heart to joy, spleen to pensiveness, lung to worry, and kidney relates to fear. Emotion disturbances due to stress coming from outside of the body affect the corresponding viscera organs and results in illnesses. For example, anger impairs liver and over-joy impairs heart. Worry leads to abnormality of lung and spleen. Pensiveness affects heart and spleen. Grief causes problems in heart and lung. Fear gives rise to diseases of heart, kidney and liver. Fright affects heart and liver. In fact, worry, pensiveness, fright and fear usually appear at the same time and causing multi-organs or systemic diseases.
Concepts of diseases in TCM
In the TCM literature, there are much more description on diseases rather than health. Disease state is the result of breaking down of the equilibrium of Yin and Yang. In the disease state, there is an excess or deficiency of either Yin or Yang. In describing a disease, two major theories are used. They are the 8-Principle Syndrome Differentiation Theory () and the Visceral Syndrome Differentiation Theory () [2]. These two theories can be viewed as clinical reasoning tools for classifying specific health states and syndromes, and for diagnosing illnesses. The 8-Principle Syndrome Differentiation Theory classifies syndromes by using 4 different dimensions, which are: Yin or Yang syndrome (), exterior or interior syndrome (), cold or heat syndrome () and deficiency or excessive syndrome (). The body organs or body functions that are affected by these syndromes are classified by the Visceral Syndrome Differentiation Theory. By using these two classification systems on various syndromes, a diagnosis can be derived, and treatment can be planned accordingly.
Treatment in CM is a process of regulating and re-establishing the balance of Yin and Yang through reducing the redundancy of Yin or Yang () and reinforcing the deficiency of Yin or Yang () [3]. Herbal medicine, acupuncture, and Qi Gong are used as major treatment modalities in CM.
Theoretical framework of the ChQOL
In this study, we adopted the concept of health rather than the concept of disease in TCM as the theoretical base of the ChQOL. The concept of positive health was used to form the domain-facet structure of the ChQOL. The research team did not use the 8-Principal Syndrome Differentiation Theory and the Visceral Syndrome Differentiation Theory to construct the initial structure of the ChQOL. These 2 theories might be good for developing disease specific modules since they elaborate more on the diseases rather than health.
The research team had proposed an initial structure model for the ChQOL with 4-domains (Figure 1). The four domains were: the harmonization between Physical Form and Spirit, harmonization between Man and Society, harmonization between Man and Nature, and the Emotions. Theory on harmonization of Physical Form and Spirit was the core domain because it is much richer in its coverage and content. Since the conceptual relationship of the domains was not clear, several possible models could be developed. At the initial stage, the four domains were arranged in parallel with each other.
Figure 1 Initial structure model of the ChQOL for the purpose of drafting of items
An expert penal consisted of 5 CM experts were invited to comment on the initial model. There were two CM clinicians, one CM pharmaceutical expert, and one academic who was an expert in TCM theory. A half day panel meeting was conducted. In the meeting, the penal members supported the initial model in principle. The penal opined that the domain on the harmonization of Physical Form and Sprit could be viewed as two related but separate domains. It was because of the fact that quite a number of facets could be proposed under each of these two domains; and the two domains could provide some unique information about the general health of an individual.
Based on the discussion in the penal meeting, the research team had revised the domain structure of the ChQOL. The original domain on Physical Form and Spirit was separated into two domains. Therefore, the revised model had 5 domains. In TCM literature, there was no clear description of the hierarchical relationship of various health concepts and body functions, i.e. there were no clear classification of domain and facets for health. The team had proposed facets for the domains basing on their interpretation of TCM literature and their clinical experiences. These facets were important indicators for reflecting the equilibrium of Yin and Yang in the respective domains.
In the Physical Form domain, 8 facets were proposed, which were: body build, complexion, sleep, appetite and digestion, stamina, breathing, sex, bowel and urination. In the Spirit domain, 4 facets were proposed, which were: thinking, verbal expression, consciousness, and spirit of the eyes. In the Emotion domain, 5 facets were proposed, which were: joy, anger, worry & pensiveness, grief, fear & anxiety. No facets were developed under the domain on harmonization of Man and Society, and harmonization of Man and nature.
Item pool
Based on the revised domain and facets, we started to draft items for the ChQOL. The team was successful in drafting many items for the Physical Form domain, Spirit domain and Emotion domain. The numbers of items drafted were 33, 13, and 23 in the Physical Form, Sprit, and Emotion domain respectively (Table 1).
Table 1 Number of items under the facets in various stage of development
Number of items
Domains and facets Initial draft by the research team After review by 76 TCM practitioners After cognitive debriefing After factor analysis
Physical Form Domain 33 37 36 20
Body build 7 7 7 0
Complexion 3 5 5 4
Sleep 4 4 4 3
Stamina 3 4 4 6
Appetite & digestion 6 7 7 4
Breathing 1 2 1 0
Sex 1 2 2 0
Bowel & urine 4 2 2 0
Adaptation to climate 4 4 4 3
Spirit Domain 13 16 16 12
Consciousness 3 5 4 3
Thinking 5 4 5 5
Spirit of the eye 2 3 3 2
Verbal expression 3 4 4 2
Emotion Domain 23 25 26 18
Joy 4 5 6 4
Anger 4 5 5 5
Worry & Pensiveness 7 5 5 0
Grief 4 5 5 0
Fear & Anxiety 4 5 5 3
Depressed mood 0 0 0 6
Total: 69 78 78 50
However, in the process of item drafting, it was found that CM practitioners did not ask many questions related to the domain on Harmonization between Man and Society, This might be due to the fact that CM practitioners focus more on health and diseases rather than social relationship in their day-to-day clinical practice. The research team managed to draft some questions on this domain. However, the items were not usual asked by CM practitioner in their practice; and further, these items were very much similar to the items in many existing QOL instruments, which did not show much connection to CM. Therefore, the team decided not to use these items because this might deviated from the original intention of using the genuine knowledge and experiences of TCM in developing the ChQOL.
For the domain on the harmonization between Man and Nature, only 4 items related to adaptation to seasons and climate were drafted. These items could form a domain. However, in CM, adaptation to seasons and climate could also be regarded as a facet under the Physical Form domain. Therefore, the research team decided to re-group these items as a facet under the Physical Form domain so that the overall ChQOL structure could be simplified. This facet was named as adaptation to climate. As a result, there were 9 facets under the Physical Form domain.
The conceptual structure of ChQOL was then revised as a three domains structure, which were the Physical Form domain, the Spirit domain, and the Emotion domain, with a total of 69 items (Table 1). By dropping the domain on Harmonization of Man and Society and Harmonization between Man and Nature, the coverage of ChQOL became narrower and focused more on health status rather than a broader sense of quality of life.
Clinician review on the ChQOL structure and the drafted items
The refined domains and facets model and the 69 items of the ChQOL, were then mailed to a sample of about 100 Chinese Medicine practitioners for their comments and advice. 76 sets of comment were received in a standard reply form. Nearly all members in the group agreed on the domains and facets structure of the ChQOL. Some of them expressed reservations but could not make concert suggestions. With their input, the domains and facets structure of the ChQOL were further confirmed. Many of the CM practitioners made suggestions on refining certain items. Some pointed out that some items might have different interpretations in different people, some were confusing, some asked two things in one question, etc. The team reviewed each suggestion and many of them were adopted to revise the original items. Some additional questions for the facets were suggested. All items that could expand the scope of coverage for the facets were adopted. As a result, a total of 78 items were then developed for the cognitive debriefing step.
Cognitive Debriefing
Cognitive Debriefing was performed with a convenience sample of 15 adult subjects that included both healthy people and people attending Chinese Medicine treatment. There were 6 healthy subjects and 9 patients in the group. About 2/3 of them were male. They were divided into 3 groups conveniently and a focus group session was arranged for each group. In the cognitive debriefing session, they were asked to answer the 78 questions and then to think aloud how they understood the items and commented on the linguistic and semantic clarity of the items. They were also asked to suggest ways to improve the items. Many items were revised to make them more understandable to lay people. There were two items asking two independent things in one question. Those items were split into two items. There were another two items found written quite badly, and were not able to be improved. Therefore, they were removed from the item pool. As a result, the total number of item in the item pool remained as 78. (Table 1)
Demographics of the sample in the field study
A sample of 273 subjects was recruited for the field study. About 63% of the subjects were recruited from provinces in Southern China and the rest from Northern China. About half of the subjects were male. The mean age was 40 years, ranging from 18 to 68 years. Most of the subjects came from the ethnic group of Han. About 66% were married. About 42% had completed secondary or technical school, and about 29% had completed university or above. There were about 18% workers, 7% farmers, 14% students, 24% professionals. About 62% were living in city, 12% in county, 9% in town and 16% in rural areas. Of the total sample, about 33% were recruited from out-patient clinics and 37% from in-patient services of Chinese Medicine hospitals. The other 30% were healthy subjects recruited conveniently in the community. (Table 2) There were no statistically significant difference among the three groups in gender, age, marital status, education level, occupation and living area. The self-reported health status was significantly better in the healthy group as compared with the patient groups.
Table 2 Demographic characteristic of the sample of the field test
Number of subjects (%)
Demographics Healthy Subjects Patient subjects All Subjects
Out-patients In-patients
No. of subjects 80 (29.3%) 91 (33.3%) 102 (37.3%) 273 (100%)
Source of sampling:
Northern China 33 (41.3%) 34 (37.4%) 34 (33.3%) 101 (37.0%)
Southern China 47 (58.8%) 57 (62.6%) 68 (66.7%) 172 (63.0%)
Gender
Male 40 (50.0%) 46 (50.5%) 51 (50.0%) 137 (50.2%)
Female 40 (50.0%) 45 (49.5%) 51 (50.0%) 136 (49.8%)
Age (Years)
Mean (SD) 38.8 (14.1) 39.1 (14.0) 40.6 (14.6) 39.6 (14.1)
range 18–68 18–65 19–67 18–68
Ethnic group
Han 74 (92.5%) 84 (92.3%) 95 (93.1%) 253 (92.7%)
Hui 4 (5.0%) 6 (6.6%) 5 (4.9%) 15 (5.5%)
Others 2 (2.5%) 1 (1.1%) 2 (2.0%) 5 (1.8%)
Marital status
Single 30 (37.5%) 26 (28.6%) 21 (20.5%) 77 (28.2%)
Married 48 (60.0%) 60 (65.9%) 71 (69.6%) 179 (65.6%)
Partnered 0 (0.0%) 1 (1.1%) 5 (4.9%) 6 (2.2%)
Separated/Divorced 1 (1.25%) 1 (1.1%) 1 (1.0%) 3 (1.1%)
Widowed 1 (1.25%) 2 (2.2%) 2 (2.0%) 5 (1.8%)
Missing 0 (0.0%) 1 (1.1%) 2 (2.0%) 3 (1.1%)
Education level
Primary school 2 (2.5%) 6 (6.6%) 10 (9.8%) 18 (6.6%)
Middle school 9 (11.25%) 28 (30.8%) 21(20.6%) 58 (21.2%)
High school 9 (11.25%) 12 (13.2%) 26 (25.5%) 47 (17.2%)
Technical school 5 (6.25%) 9 (9.9%) 13 (12.7%) 27 (9.9%)
College diploma or degree 11 (13.75%) 15 (16.5%) 16 (15.7%) 42 (15.4%)
University degree or above 44 (55.00%) 20 (22.0%) 15 (14.7%) 79 (28.9%)
Other 0 (0.0%) 1 (1.1%) 1 (1.0%) 2 (0.7%)
Occupation:
Worker 10 (12.5%) 18 (19.8%) 22 (21.6%) 50 (18.3%)
Farmer 2 (2.5%) 8 (8.8%) 10 (9.8%) 20 (7.3%)
Professionals 32 (40.0%) 21 (23.1%) 12 (11.8%) 65 (23.8%)
Student 19 (23.8%) 13 (14.3%) 7 (6.9%) 39 (14.3%)
Unemployed 4 (5.0%) 4 (4.4%) 14 (13.7%) 22 (8.1%)
Other 11 (13.8%) 25 (27.5%) 33 (32.4%) 69 (25.3%)
Missing 2 (2.5%) 2 (2.2%) 4 (3.9%) 8 (2.9%)
Living Arrangements:
City 62 (77.5) 51 (56.0%) 56 (54.9%) 169 (61.9%)
County 4 (5.0%) 14 (15.4%) 14 (13.7%) 32 (11.7%)
Town 1 (1.3%) 10 (11.0%) 13 (12.7%) 24 (8.8%)
Rural area 13 (16.3%) 14 (15.4%) 16 (15.7%) 43 (15.8%)
Missing 0 (0.0%) 2 (2.2%) 3 (2.9%) 5 (1.8%)
Self-reported health status:
Very poor 0 (0.0) 2 (2.2%) 12 (11.8%) 14 (5.1%)
Poor 3 (3.8%) 15 (16.5%) 26 (25.5%) 44 (16.1%)
Neither poor nor good 20 (25.0%) 40 (44.0%) 38 (37.3%) 98 (35.9%)
Good 48 (60.0%) 30 (33.0%) 22 (21.6%) 100 (36.6%)
Very good 8 (10.0%) 4 (4.4%) 4 (3.9%) 16 (5.9%)
Missing 1 (1.3%) 0 (0.0%) 0 (0.0%) 1 (0.4%)
The final ChQOL structure
A series of exploratory factor analysis and confirmatory factor analysis were used to test the appropriateness and structure fitness of the initial ChQOL model. The first round of exploratory factor analysis was done on the 52 items under the Physical Form domain and the Spirit domain. A two-factor structure, explaining 38% of the total variance, was derived. Factor 1 consisted mostly of items under the Physical Form domain, and factor 2 consisted of items under the Spirit domain. We had excluded items with factor loading smaller than 0.4 on either factor. Items that loaded more or less the same on both factors were also excluded. There were 20 items remained under the factor on Physical Form, and 12 items under the factor on Spirit.
Further exploratory factor analysis was conducted on the two domains independently. Factor analysis of the 20 items under the Physical Form domain resulted in 5 factors. The factors corresponded to the facets on stamina, complexion, appetite and digestion, sleep, and adaptation to climate respectively. The other facets under the Physical Form domain, including body builds, breathing, sex, bowel and urination were not supported by the data. No item under the Physical Form domain was further reduced in this step.
Factor analysis on the 12 items in the Spirit domain supported all the 4 facets in the domain. No item was reduced from the domain in this step.
Exploratory factor analysis on the 26 items under the Emotions domain resulted in four factors. The date supported the facet on joy and anger. Items under facets on worry, pensiveness and grief combined to form a factor that could be summarized as depressed mood. Some items under anxiety were included in the facet on fear. As a result, 8 items were removed and 18 items remained in the four facets, i.e. joy, anger, depression mood, and fear.
Item-facet and facet-domain structure fitness
A series of confirmatory factor analysis on item-facet fitness were done. Nearly all the item-facet structures were supported with Comparative Fitness Index (CFI) exceeding 0.9. We therefore concluded that all the item-facet structures of the ChQOL were supported (Table 3).
Table 3 Structure fitness of facets, domain and the overall structure
CHQOL facets Number of items CFI Item-facet Chi sq Degree of freedom
Ch1.1 Complexion 4 0.928 38.07 1
Ch1.2 Sleep 3 0.946 13.73 2
Ch1.3 Stamina 6 0.994 11.70 8
Ch1.4 Appetite & digestion 4 0.985 7.29 1
Ch1.5 Adaptation to climate 3 0.820 38.42 2
Ch2.1 Consciousness 3 0.967 10.13 2
Ch2.2 Thinking 5 0.982 20.02 4
Ch2.3 Spirit of the eyes 2 1.000 0.00 0
Ch2.4 Verbal expression 2 1.000 0.00 0
Ch3.1 Joy 4 0.990 6.97 1
Ch3.2 Anger 5 0.961 20.09 4
Ch3.3 Depressed mood 6 0.982 19.04 8
Ch3.4 Fear & Anxiety 3 0.991 3.67 2
ChQOL domains No. of facets CFI Facet-domain Chi sq Degree of freedom
Physical Form Domain 5 0.943 295.03 159
Vitality & Spirit Domain 4 0.956 133.99 45
Emotion Domain 4 0.872 437.36 126
Facet and domain fitness of the 3 domains were also supported by confirmatory factor analysis. The CFI of the Physical form and the Sprit domain were 0.943 and 0.956 respectively. The CFI of the facet and domain structure of the Emotion domain was 0.872. (Table 3)
Domain-overall ChQOL structure fitness
We had proposed the 3 possible domain structures of the ChQOL (Figure 2). All of them were compatible with the concept of health in TCM. We used the domain scores to test these three possible structures. The CFI of all the three models had exceeded 0.9 and were satisfactory. Model 1 was a 3-level structure consisting of domains and sub-domains. Since the composite domain on Physical Form and Spirit did not provide much interpretable information on top of the two separate Physical form domain and Spirit domain, model 1 was aborted. Model 2 was also a 3-level structure. It was a complex model and the relationship of the domains was difficult to interpret. Hence, this model was not adopted. Model 3 is a simple 2-level model. The CFI of this model was 0.932. It was therefore adopted as the final structure. The domain and facets of the final model were shown in Figure 3. The items of the ChQOL in the original Chinese and tentative English translation were listed in Appendix 1 (Additional file 1).
Figure 2 Possible final structure models of the ChQOL
Figure 3 Final domains and facets structure of the ChQOL
Psychometric properties of the ChQOL
Distribution of scores
The mean facet scores ranged from 55 in the Anger facet to 79 in the Fear facet. Mean domain scores ranged from 61 in the Physical Form domain to 69 in the Spirit domain. No ceiling and floor effects were noted. (Table 4)
Table 4 Distributions of the facets & domain scores (N = 273)
ChQOL facets & Domains No. of items / facets / domains Mean scores (SD) Floor effect (%) Ceiling effect (%)
ChQOL facets
Ch1.1 Complexion 4 56.07 (18.989) 0.7 2.2
Ch1.2 Sleep 3 59.59 (22.511) 0.7 5.5
Ch1.3 Stamina 6 56.85 (19.468) 1.1 1.1
Ch1.4 Appetite & digestion 4 65.32 (18.725) 0.7 5.9
Ch1.5 Adaptation to climate 3 64.83 (19.255) 0.4 6.6
Ch2.1 Consciousness 3 71.54 (17.983) 0.4 9.6
Ch2.2 Thinking 5 64.16 (17.326) 1.1 3.3
Ch2.3 Spirit of the eyes 2 66.91 (18.508) 1.1 4.0
Ch2.4 Verbal expression 2 74.36 (18.557) 0.4 14.7
Ch3.1 Joy 4 63.26 (17.229) 0.7 2.6
Ch3.2 Anger 5 54.54 (17.788) 0.4 0.7
Ch3.3 Depressed mood 6 69.84 (18.595) 0.4 2.2
Ch3.4 Fear & Anxiety 3 79.49 (19.306) 0.4 23.4
ChQOL domains
Physical form 5 60.57 (14.715) 0.4 0.4
Spirit 4 69.21 (15.745) 0.4 0.7
Emotion 4 66.78 (14.721) 0.4 0.7
Overall ChQOL 3 65.51 (12.764) 0.4 0.4
Facet to domain and inter-domain correlation
Facet to domain correlation ranged from 0.71 to 0.78 in the Physical Form domain, 0.83 to 0.91 in the Spirit domain, and 0.77 to 0.89 in the Emotion domain. The domain to overall ChQOL score correlation ranged from 0.56–0.78 (Table 5). There were moderate correlations among the domain scores. Correlation between the Physical Form and the Spirit domain, and the emotion domain were 0.56 and 0.61 respectively. Correlation between the Spirit domain and the Emotion domain was 0.54. There were high correlations between the overall ChQOL score and the three domain scores. The correlation coefficient ranged from 0.84 to 0.85.
Table 5 Correlation between facets and domain and overall ChQOL
ChQOL facets Pearson correlation coefficients
Physical Form Spirit Emotion Overall ChQOL
Ch1.1 Complexion 0.736** 0.553** 0.417** 0.672**
Ch1.2 Sleep 0.730** 0.264** 0.452** 0.564**
Ch1.3 Stamina 0.783** 0.534** 0.596** 0.752**
Ch1.4 Appetite & digestion 0.763** 0.400** 0.409** 0.619**
Ch1.5 Adaptation to climate 0.705** 0.366** 0.383** 0.572**
Ch2.1 Consciousness 0.518** 0.889** 0.541** 0.777**
Ch2.2 Thinking 0.498** 0.908** 0.513** 0.766**
Ch2.3 Spirit of the eyes 0.521** 0.833** 0.396** 0.700**
Ch2.4 Verbal expression 0.411** 0.846** 0.427** 0.673**
Ch3.1 Joy 0.446** 0.534** 0.774** 0.691**
Ch3.2 Anger 0.490** 0.405** 0.795** 0.662**
Ch3.3 Depressed mood 0.562** 0.434** 0.886** 0.739**
Ch3.4 Fear & Anxiety 0.468** 0.358** 0.774** 0.639**
ChQOL Domains
Physical form 1.000 --- --- 0.852**
Spirit 0.562** 1.000 --- 0.838**
Emotion 0.609** 0.542** 1.000 0.845**
** Correlation is significant at the 0.01 level (2-tailed)
Reliabilities
The results on internal consistencies of the facets were all good. Cronbach's alpha values of the facets ranged from 0.71 in the Verbal Expression facet to 0.90 in the Thinking facet. Test-retest reliability study was conducted on 56 subjects within 2 days. The ICC(1,1) values ranged from 0.68 to 0.84 in the facet scores, and from 0.83 to 0.87 in the domain scores. The ICC(1,1) value for the overall ChQOL score was 0.90. (Table 6)
Table 6 Internal consistency and test-retest reliability of ChQOL
ChQOL facets Cronbach's alpha N = 273 ICC N = 56
Ch1.1 Complexion 0.8601 0.75
Ch1.2 Sleep 0.7425 0.68
Ch1.3 Stamina 0.8484 0.82
Ch1.4 Appetite & digestion 0.8183 0.74
Ch1.5 Adaptation to climate 0.7403 0.84
Ch2.1 Consciousness 0.7949 0.67
Ch2.2 Thinking 0.9022 0.78
Ch2.3 Spirit of the eyes 0.7534 0.70
Ch2.4 Verbal expression 0.7083 0.81
Ch3.1 Joy 0.8707 0.71
Ch3.2 Anger 0.8012 0.80
Ch3.3 Depressed mood 0.8409 0.76
Ch3.4 Fear & Anxiety 0.7280 0.80
ChQOL domains
Physical form --- 0.83
Spirit --- 0.87
Emotion --- 0.87
Overall ChQOL --- 0.90
Convergent Validity
Convergent and concurrent validities were studied by correlating the three domain scores of the ChQOL with the self-report health status (5-point scale ranging from very good to very bad), the six domains of the WHOQOL-100, and the eight domain scores of the SF-36.
The correlation between the three domain scores and the self reported health status were fair. The correlation coefficients were 0.56 in the Physical Form domain, 0.39 in the Spirit domain and 0.63 in the Emotion domain. All correlations were all statistically significant.
The three domain scores of the ChQOL showed fair to moderate correlation with the six domain scores of the WHOQOL-100. All the correlations were statistically significant, except the correlation between Physical Form domain in the ChQOL and Spirituality and Personal Belief domain of the WHOQOL-100. The Physical Form domain had higher correlations with the Physical domain (r = 0.73) and Level of Independence domain (r = 0.63), and had lower correlation between the Social (r = 0.31), Environmental (r = 0.36) and Spirituality domain (r = 0.08) of the WHOQOL-100. The Spirit domain had only fair correlation with all the six WHOQOL-100 domains, with correlation coefficient ranging from 0.19 in the Spirituality domain to 0.47 in the Psychological domain of the WHOQOL-100. The Emotion domain showed higher correlation with the Psychological domain (r = 0.64) and the Physical domain (r = 0.52), and lower correlation with the other 4 domains in the WHOQOL-100. The overall ChQOL score demonstrated significant correlation with all the 6 WHOQOL-100 domains. The correlation was moderate with the Physical (r = 0.64), Psychological (r = 0.63) and Level of Independence (r = 0.55) domains, and fair with the Social (r = 0.41) and the Environmental (r = 0.43) domain, and low with the Spirituality and Personal Belief domain (r = 0.21) of the WHOQOL-100. (Table 7)
Table 7 Correlation between ChQOL domains and WHOQOL-100 domains
ChQOL domains WHOQOL-100 domains
Physical Psychological Level of independent Social Environmental Spirituality
Physical form 0.730** 0.481** 0.627** 0.306** 0.358** 0.084
Vitality & Spirit 0.397** 0.470** 0.399** 0.274** 0.291** 0.187**
Emotions 0.516** 0.642** 0.382** 0.462** 0.449** 0.263**
Overall ChQOL 0.644** 0.628** 0.554** 0.406** 0.430** 0.208**
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
The three domain scores of the ChQOL showed fair to moderate correlation with all the 8 facet scores of the SF-36. All the correlations were statistically significant except the correlation between the Spirit domain of the ChQOL with the Role Emotion facet in the SF-36. The correlation coefficients had ranged from 0.16 to 0.65. The overall ChQOL score revealed moderate correlation with General Health, Vitality, and Mental Health of the SF-36, yielding correlation coefficients of 0.55, 0.53 and 0.52 respectively. (Table 8)
Table 8 Correlation between ChQOL domains and SF-36 facets
ChQOL domains SF-36 facets
PF RP BP GH VT SF RE MH
Physical form 0.400** 0.423** 0.398* 0.610** 0.512** 0.501** 0.170** 0.377**
Vitality & Spirit 0.233** 0.237** 0.161** 0.379** 0.320** 0.251** 0.070 0.298**
Emotion 0.279** 0.317** 0.188** 0.416** 0.522** 0.408** 0.246** 0.647**
Overall ChQOL 0.359** 0.386** 0.294** 0.553** 0.533** 0.458** 0.192** 0.520**
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
PF = physical function; RP = role physical; BP = bodily pain; GH = General health; VT = vitality; SF = social function; RE = Role emotional; MH = mental health.
Discrimination validity
The examinations on the ability of the mean facet, domain and overall ChQOL scores in discriminating different groups of subjects with known differences in health status were studied using ANOVA tests, with gender and age adjusted, were conducted. The total sample was categorized in groups by three different ways. Firstly, the sample was categorized as healthy subjects, out-patients and in-patients groups. This category showed statistically significant differences in the mean scores in 19 out of 13 facets, and in the Physical Form domain, the Emotion domain, and the overall ChQOL scores among these three groups. (Table 9)
Table 9 Differences between facet and domain scores in healthy and patient adjusted for gender and age
Mean facet / domain scores
ChQOL facets Healthy subjects N = 80 Out-patients N = 91 In-patients N = 102 F P-value1
Mean SD Mean SD Mean SD
Ch1.1 Complexion 61.67 15.58 56.52 17.92 51.27 21.14 6.791 0.001** 5
Ch1.2 Sleep 67.29 19.88 58.24 22.00 54.74 23.50 7.338 0.001** 3
Ch1.3 Stamina 65.79 15.84 57.48 16.90 49.26 21.17 17.655 0.000** 2
Ch1.4 Appetite & digestion 69.53 17.38 67.79 18.23 59.80 19.02 7.223 0.001** 4
Ch1.5 Adaptation to climate 67.93 17.30 65.84 17.83 61.52 21.46 2.263 0.106
Ch2.1 Consciousness 73.33 16.31 72.80 17.52 68.98 19.48 1.488 0.228
Ch2.2 Thinking 67.81 14.54 64.34 17.23 61.13 18.94 3.162 0.044* 5
Ch2.3 Spirit of the eyes 70.89 15.07 68.54 16.70 62.38 21.43 4.998 0.007** 5
Ch2.4 Verbal expression 76.58 15.93 74.86 18.68 72.18 20.20 1.152 0.317
Ch3.1 Joy 66.72 15.75 63.80 17.44 60.05 17.72 3.372 0.036* 5
Ch3.2 Anger 59.44 17.95 54.18 17.44 51.03 17.25 5.197 0.006** 5
Ch3.3 Depressed mood 73.02 17.69 72.07 17.06 65.36 19.86 5.355 0.005** 4
Ch3.4 Fear & Anxiety 82.92 17.78 80.31 19.02 76.06 20.30 3.143 0.045* 5
ChQOL domains
Physical form 66.64 12.54 61.18 13.04 55.32 15.85 14.087 0.000** 2
Spirit 72.02 12.97 70.14 15.00 66.20 17.85 2.954 0.054
Emotion 70.52 14.52 67.59 14.30 63.13 14.90 6.269 0.002** 5
Overall ChQOL 69.84 11.26 66.30 11.73 61.50 13.61 9.781 0.000** 4
**significant at the 0.01 level (2-tailed)
* significant at the 0.05 level (2-tailed)
1Scheffe and Dunnett T3 tests are used for post-hoc analysis.
2The difference is significant for the group (1,2), (1,3) and (2,3).
3The difference is significant for the group (1,2), (1,3).
4The difference is significant for the group (1,3), (2,3).
5The difference is significant for the group (1,3).
Secondly, the sample was categorized into another three groups according to their self-report health status, i.e. very poor and poor health, neither poor nor good health, and good and very good health. There were statistically significant differences in all the 13 facet scores, all the 3 domain scores and the overall ChQOL scores among these three groups (Table 10).
Table 10 Differences between facet and domain scores in subjects of different self reported health status adjusted for gender and age
Mean & SD of facet / domain scores
ChQOL Facets Poor & Very poor (N = 58) Neither poor nor good (N = 98) Good & Very good (N = 116) F P-value
Mean SD Mean SD Mean SD
Ch1.1 Complexion 43.14 21.75 54.17 16.74 64.13 15.13 30.767 0.000** 2
Ch1.2 Sleep 46.26 23.58 58.50 19.48 66.95 21.28 19.272 0.000** 2
Ch1.3 Stamina 37.00 19.78 56.55 13.55 66.83 15.62 68.083 0.000** 2
Ch1.4 Appetite & digestion 52.05 20.73 66.20 15.12 70.96 17.18 24.416 0.000** 3
Ch1.5 Adaptation to climate 57.33 21.86 62.16 18.07 70.80 17.11 10.824 0.000** 4
Ch2.1 Consciousness 61.64 21.34 69.76 17.24 77.80 13.90 18.749 0.000** 2
Ch2.2 Thinking 52.84 22.07 62.04 12.96 71.51 14.25 28.468 0.000** 2
Ch2.3 Spirit of the eyes 57.11 24.01 65.56 15.34 72.83 15.38 15.500 0.000** 4
Ch2.4 Verbal expression 66.16 22.58 73.32 17.74 79.31 15.38 9.960 0.000** 4
Ch3.1 Joy 55.17 19.80 61.54 15.45 68.64 15.51 13.609 0.000** 4
Ch3.2 Anger 46.47 18.66 54.18 15.79 58.88 17.71 10.692 0.000** 3
Ch3.3 Depressed mood 59.55 21.46 69.94 18.87 74.71 14.42 16.907 0.000** 3
Ch3.4 Fear & Anxiety 72.27 24.55 79.17 18.00 83.26 16.36 8.310 0.002** 5
ChQOL domains
Physical form 47.16 16.00 59.52 10.74 68.08 11.69 57.891 0.000** 2
Spirit 59.44 20.44 67.60 12.94 75.37 12.08 23.723 0.000** 2
Emotion 58.37 16.83 66.21 13.74 71.37 12.43 19.210 0.000** 2
Overall ChQOL 54.99 14.92 64.46 9.55 71.63 10.01 46.131 0.000** 2
** Significant at the 0.01 level (2-tailed)
* Significant at the 0.05 level (2-tailed)
1Scheffe and Dunnett T3 tests are used for post-hoc analysis.
2The difference is significant for the group (1,2), (1,3) and (2,3).
3The difference is significant for the group (1,2), (1,3).
4The difference is significant for the group (1,3), (2,3).
5The difference is significant for the group (1,3).
Thirdly, the sample was categorized according to their self-report disease conditions, i.e. healthy, having stable chronic illness and having acute illness that required active treatment. There were statistically significant differences in 10 out of the 13 facet scores, the Physical Form domain, the Emotion domain, and the overall ChQOL score. (Table 11)
Table 11 Differences between facet and domain scores in subjects of different disease conditions adjusted for gender and age
Mean & SD of facet / domain scores
ChQOL Facets Healthy (N = 76) Stable chronic illness (N = 56) Receiving active treatment (N = 139) F P-value
Mean SD Mean SD Mean SD
Ch1.1 Complexion 64.67 15.82 57.11 15.49 50.99 20.29 13.656 0.000** 2
Ch1.2 Sleep 68.09 20.52 55.51 20.86 56.35 23.13 7.696 0.000** 3
Ch1.3 Stamina 67.45 16.12 57.29 14.22 50.73 20.54 19.459 0.000** 4
Ch1.4 Appetite & digestion 71.71 17.27 65.29 16.47 61.51 19.39 7.252 0.001** 2
Ch1.5 Adaptation to climate 69.00 16.81 65.63 16.44 61.93 21.00 2.604 0.076
Ch2.1 Consciousness 74.34 17.04 71.21 16.10 70.08 19.11 1.209 0.300
Ch2.2 Thinking 68.55 16.73 63.04 14.42 62.23 18.43 3.090 0.047* 2
Ch2.3 Spirit of the eyes 71.83 16.19 66.29 13.47 64.48 20.87 3.380 0.036* 2
Ch2.4 Verbal expression 77.33 17.52 73.66 16.63 73.11 19.85 1.022 0.361
Ch3.1 Joy 67.43 17.33 62.61 13.82 61.24 18.15 3.153 0.044* 2
Ch3.2 Anger 60.79 18.04 52.14 15.34 52.23 17.92 6.849 0.001* 3
Ch3.3 Depressed mood 74.01 16.46 69.72 16.64 67.42 20.12 4.112 0.017* 2
Ch3.4 Fear & Anxiety 83.99 17.09 81.70 15.68 76.08 21.25 5.765 0.004** 2
ChQOL domains
Physical form 68.41 12.78 60.16 10.96 56.30 15.33 17.518 0.000** 3
Spirit 72.89 14.23 68.66 13.24 67.48 17.16 2.432 0.090
Emotion 71.56 13.67 66.54 12.34 64.24 15.63 7.096 0.001* 2
Overall ChQOL 71.11 11.34 65.09 9.44 62.67 13.72 10.705 0.000** 3
** Significant at the 0.01 level (2-tailed)
* Significant at the 0.05 level (2-tailed)
1Scheffe and Dunnett T3 tests are used for post-hoc analysis.
2The difference is significant for the group (1,3).
3The difference is significant for the group (1,2), (1,3).
4The difference is significant for the group (1,2), (1,3) and (2,3).
Conceptual overlapping with the WHOQOL-100 and SF-36
Further exploration was then conducted to examine the degree of overlapping between the ChQOL and two other well established generic HrQOL instruments, i.e. the WHOQOL-100 and the SF36. In the exploratory factor analysis on the 13 facet scores of the ChQOL and the 24 facet scores of the WHOQOL-100, seven factors were generated that explained about 66% of the total variance in the data set. (Table 12). Factor one consisted of 4 out of 5 facets in the Psychological domain, all the 3 facets in the Social Relationship domain, 2 facets in the Environment domain and the only facet in the Spirituality and Personal beliefs domain of the WHOQOL-100. This explained about 14% of the total variance of the data. Factor two consisted of all the 3 facets in the Physical domain, and all the 4 facets in the Level of Independence domain of the WHOQOL-100, and explained about 11.7% of the total variance. Factor three consisted of all the 4 facets in the Spirit domain and 1 out of 5 facets of the Physical Form domain of the ChQOL, and explained 11.3% of the variance. Factor four consisted of 6 out of 8 facets in the Environment domain of the WHOQOL-100, and explained 10.4% of the total variance. Factor five consisted of all the 4 facets in the Emotion domain in the ChQOL and the Negative feeling facet in the Psychological domain of the WHOQOL-100, and explained 8.7% of the total variance. The 2 factors that made up mostly of facets of the ChQOL, i.e. factor 3 and 5, explained about 20% of the total variance, i.e. 1/3 of the total variances that could be explained by all the seven factors.
Table 12 Exploratory Factor analysis on the 24 facets of WHOQOL-100 and the 13 facets of ChQOL
Facets Component
1 2 3 4 5 6 7
F2.1 Positive Feelings .663
F2.2 Thinking, Memory & Concentration .604
F2.3 Self-esteem .712
F2.4 Bodily Image and Appearance .521
F4.1 Personal Relationships .779
F4.2 Social Support .643
F4.3 Sexual Activity .607
F5.5 Acquiring New Information & Skills .563 .509
F5.6 Participation in Leisure .675
F6.1 Spirituality/ Religion/ Personal Beliefs .533
F1.1 Pain & discomfort -.604
F1.2 Energy and Fatigue .655
Ch1.3 Stamina .604
F3.1 Mobility .768
F3.2 Activities of Daily Living .723
F3.3 Dependence on Medication -.723
F3.4 Working Capacity .650
Ch1.1 Complexion .535
Ch2.1 Consciousness .807
Ch2.2 Thinking .869
Ch2.3 Spirit of Eyes .776
Ch2.4 Verbal Expression .786
Ch1.5 Adaptation to climate
F5.1 Physical Safety and Security .547
F5.2 Home Environment .730
F5.3 Financial Resources .704
F5.4 Availability & quality of health & social care .676
F5.7 Physical Environment .711
F5.8 Transport .763
Ch3.1 Joy .476
Ch3.2 Anger .664
Ch3.3 Depressed mood .765
Ch3.4 Fear .681
F2.5 Negative Feelings -.570
F1.3 Sleep and Rest .776
Ch1.2 Sleep .745
Ch1.4 Appetite & digestion .580
Variance (%) 13.98 11.74 11.29 10.38 8.68 6.28 3.63
Cumulative (%) 13.98 25.72 37.01 47.40 56.08 62.37 66.00
Extraction Method: Principal Component Analysis; Eigenvalue >1; Varimax rotation with Kaiser normalization; sorted by factor names; factor loading <0.5 suppressed
Facet labels begin with F are WHOQOL-100 facets
Facet labels begin with Ch are ChQOL facets
Similar exploratory factor analysis was conducted on the 8 facets of the SF-36 and the 13 facets of the ChQOL. Four factors which explained about 64.5% of total variance of the data were identified (Table 13). Factor one consisted of 7 out of the 8 facets of the SF-36, and explained about 18.8% of the total variance. Only the facet on mental health was not included. Factor two consisted of all the 4 facets of the Spirit domain of the ChQOL, and explained about 17.8% of the total variance. Factor three consisted of all the 4 facets of the Emotion domain of the ChQOL and the Mental Health facet of the SF-36, and explained about 16.0% of the total variance. Factor four consisted of all the 5 facets of the Physical Form domain of the ChQOL, and explained 11.8% of the total variance. The three factors that consisted of facets mainly from ChQOL, i.e. factor 2, 3 and 4 could explain more than 2/3 of the total variance.
Discussion
We have established the item-facet-domain structure of the ChQOL based on our understanding of the concept of health in TCM. An item pool consisting of 78-items was developed for field testing. The ChQOL structure was validated and the items in the ChQOL reduced to 50-items. Psychometric properties of the ChQOL were good. Construct validity was established by various methods, i.e. the internal consistency in all facets were good; the correlation between facets to domain, and domains to overall ChQOL correlation were high; confirmation factor analysis showed that the structure fitness of all facets, domains and overall structure were good. Test-retest reliability was also good, especially in the domain scores with ICC value ranging from 0.83 to 0.90. No ceiling or floor effect was noted. This indicated that ChQOL could be applied to subjects with a wide range of health status. Most facet scores, domain scores and the overall ChQOL scores were able to discriminate groups of subjects with known differences in health status. The ChQOL had mild positive convergence with the other generic health related QOL measures, i.e. the WHOQOL-100 and the SF-36, with moderate correlations. This suggested that the ChQOL was measuring similar but not the same concept as compared with the two measures.
The findings showed that the ChQOL encompasses several aspects of health that are not well covered by the two existing generic HRQOL measures, i.e. the WHOQOL-100 and SF-36. Some of the examples are: the facets on complexion, appetite and digestion, adaptation to climate, spirit of eyes, and verbal expression. The ChQOL categorizes physical aspect of health into two domains, the Physical Form domain and the Spirit domain. The ChQOL contains a domain on emotion that covers various kinds of moods. This domain captures information that is affective in nature. Other psychological mechanisms, for example, self-esteem, that are more cognitive in nature are however not included.
Results indicated that there was only mild overlapping in the WHOQOL-100 and the ChQOL at the domain level, i.e. the WHOQOL-100 domains did not cover the important facets in the ChQOL and vice versa. ChQOL did contribute additional information on top of the WHOQOL-100. There were even less overlapping between SF-36 and ChQOL and both together could provide a wider range of information on people's health.
In the process of drafting items for the ChQOL, we decided to exclude the domain on Harmonization of Man and Nature, and the Harmonization of Man and Society from the ChQOL, because we were not able to draft items basing on how CM practitioners ask questions on these two domains. Since the social domain and environmental domain of life are two essential components in QOL, we may explore ways of developing these two domains to further widen the coverage of the ChQOL. To develop domain on social and environment, we may have to explore into the traditional and contemporary Chinese culture rather then Chinese medicine.
Since the ChQOL consists of 50 items and requires at least 10 to 20 minutes for a person to response to all the questions. We will explore the need and feasibility of developing a shorter version of the ChQOL for screening purpose.
We speculate that the ChQOL is more sensitive to health changes brought about by Chinese Medicine. We are planning for the next phase of study on testing the sensitivity of ChQOL in the detection of health changes attained by Chinese Medicine.
Limitations
We attempted to develop an instrument based on the concept of health in TCM. However, due to many methodological limitations, the structure of the ChQOL was not able to fully reflect the concept of health in TCM. The ChQOL was establish in form of a questionnaire and this could only capture information, i.e. signs, symptom, emotions and feelings, that could be consciously aware by the subjects. There might be some other important information, e.g. the pulse, which could only be assessed by CM practitioners not included in the CHQOL.
The items in the ChQOL came mainly from the conversation between CM practitioners and their patients. These questions relate mostly to diseases state rather to healthy state. This might make the items in the ChQOL biased toward disease rather than health.
We were trying to use methods that were commonly used in developing QOL measures in Western medicine in the development of the ChQOL. A linear overall-domain-facet-item structure was assumed. However, this might not be a good and sufficient representation of the concept of equilibrium of Ying and Yang in Chinese medicine. The ChQOL was reflecting the concept of equilibrium of Yin and yang in health at the item level, i.e. the rating on a particular item is reflecting the equilibrium of the corresponding facet and domain, but not at the facet or domain level. We may further explore ways to manifest the concept of equilibrium at the facet and domain level. New statistical methods and new ways of conceptualizing and developing QOL instrument may be needed to achieve this.
We relied very much on the use of exploratory and confirmatory factor analysis in establishing and testing of the structure of the ChQOL. However, the sample size was relatively small to ensure a robust structure. Further studies on the construct validity of the ChQOL are necessary.
Conclusion
We have developed a new generic, self-reported health status instrument based on a well established theory on health in Chinese Medicine. This is one of the very few instruments which were developed based on a clear and explicit theory of health. Field test results indicated that the structure of the ChQOL was valid and the psychometric properties were good. It could provide additional information of health on top of some other generic health related QOL instruments. It is ready for use in clinical trial especially in studies related to the use of Chinese Medicine.
Authors' contributions
Leung was the principal investigator of the study. He assumed overall responsibility of coordination and the design of the study, building the structure of the ChQOL, drafting items, training of interviewers, data analysis and interpretation, and drafting of this article. Liu contributed in the formulation of the concept of health in Chinese Medicine, drafting of items, coordinate data collection and interpretation of field test results. Zhao helped in literature review, development of the structure of ChQOL, drafting of items, and the implementation of the study. She was authorized to use part of her work in this study in her PhD thesis. Fang gave advice on the design of the study, data analysis and interpretation of the results. Chan helped in supervising the work of Zhao, building the structure of the ChQOL, and interpretation of results. Lin contributed in formulating the concept of health and helped in data collection.
Table 13 Factor analysis of the 8 facets of SF-36 and the 13 facets of ChQOL
Facets Component
1 2 3 4
SF-RP Role Physical .824
SF-SF Social Functioning .798
SF-BP Bodily Pain .702
SF-VT Vitality .643
SF-GH General Health .614
SF-PF Physical Functioning .604
SF-RE Role Emotional .592
Ch2.1 Consciousness .836
Ch2.2 Thinking .857
Ch2.3 Spirit of Eyes .802
Ch2.4 Verbal Expression .800
Ch3.1 Joy .657
Ch3.2 Anger .664
Ch3.3 Depressed mood .786
Ch3.4 Fear .622
SF- Mental Health .785
Ch1.1 Complexion .527 .488
Ch1.2 Sleep .677
Ch1.3 Stamina .481
Ch1.4 Appetite& digestion .711
Ch1.5 Adaptation to climate .559
Variance (%) 18.86 17.79 15.97 11.82
Cumulative (%) 18.86 36.66 52.63 64.45
Extraction Method: Principal Component Analysis; Eigenvalue >1; Varimax rotation with Kaiser normalization; sorted by facet names; factor loading <0.5 suppressed.
Facet labels begin with SF are SF-36 domains
Facet labels begin with Ch are ChQOL facets
Supplementary Material
Additional File 1
Appendix 1. The items of the ChQOL in their original Chinese and tentative English translation
Click here for file
==== Refs
Chan K Lee H The way forward for Chinese medicine 2002 London and New York: Taylor & Francis Inc
Beijing University of Traditional Chinese Medicine Basic theories of Traditional Chinese Medicine 1998 Beijing, Academy Press
Beijing University of Traditional Chinese Medicine Diagnostics of Traditional Chinese Medicine 1998 Beijing, Academy Press
Scientific Advisory Committee of the Medical Outcome Trust Assessing health status and quality-of-life instruments: Attributes and review criteria Quality of Life Research 2002 11 193 205 12074258 10.1023/A:1015291021312
WHOQOL Study Protocol Division of Mental Health World Health Organization, Geneva MNH/PSF/93.9.
The WHOQOL Group The World Health Organization Quality of Life Assessment (WHOQOL) : Position Paper from the World Health Organization Soc Sci Med 1995 41 1403 1409 8560308 10.1016/0277-9536(95)00112-K
Fang JQ ed Quality of life assessment and application 2000 Beijing: Beijing Medical University Press (Chinese)
Wang SH Li LM Li J The application of the SF-36 Foreign Medical Sciences, Section of Social Medicine 2001 18 4 8 (Chinese)
Hao YT Fang JQ Li CX Shi ML The World Health Organization Quality of Life Assessment (WHOQOL) Chinese version Foreign Medical Sciences, Section of Social Medicine 1999 16 118 122 (Chinese)
Norusis MJ SPSS 11.0 guide to data analysis Upper Saddle River 2002 NJ: Prentice Hall
Byrne B Structural equation modeling with EQS and EQS/Windows: basic concepts, applications, and programming / Thousand Oaks: Sage Publications 1994
Zhao L Chan K The health concept in Chinese medicine Medicine and Philosophy 2003 24 58 59 (Chinese)
Zhao L Liu FB Leung KF Fang JQ Lin LZ Chan K Discussion on the theory and structure model of the Chinese Quality of Life Instrument Chinese Journal of Clinical Rehabilitation 2004 8 3132 3134 (Chinese)
| 15833138 | PMC1090607 | CC BY | 2021-01-04 16:38:14 | no | Health Qual Life Outcomes. 2005 Apr 16; 3:26 | utf-8 | Health Qual Life Outcomes | 2,005 | 10.1186/1477-7525-3-26 | oa_comm |
==== Front
Health Res Policy SystHealth Research Policy and Systems1478-4505BioMed Central London 1478-4505-3-51583678910.1186/1478-4505-3-5ResearchChanges in characteristics of veterans using the VHA health care system between 1996 and 1999 Liu Chuan-Fen [email protected] Matthew L [email protected] Anne EB [email protected] Department of Veterans Affairs, Puget Sound Health Care System, Health Services Research and Development, USA2 University of Washington, School of Public Health, USA2005 18 4 2005 3 5 5 11 2 2004 18 4 2005 Copyright © 2005 Liu et al; licensee BioMed Central Ltd.2005Liu 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 Department of Veterans Affairs' Veterans Health Administration (VHA) provides a health care safety net to veterans. This study examined changes in characteristics of veterans using the VHA health care system between 1996 and 1999 when VHA implemented major organizational changes to improve access of ambulatory care and to provide care to more veterans.
Methods
The study used two cross-sectional samples of the Medical Expenditures Panel Survey (MEPS), a national representative survey, in 1996 and 1999. The 1996 MEPS survey included 1,944 veterans and the 1999 MEPS survey included 1,974 veterans. There were 534 veterans and 740 veterans who used VHA services in 1996 and 1999, respectively.
Results
The proportion of veterans using the VHA system increased from 12.4% in 1996 to 14.6% in 1999. In both years, veterans were more likely to use VHA care if they were older, male, less educated, uninsured, unemployed, and in fair or poor health status. Only two variables, marital status and income, were different between the two years. Married veterans were more likely to use VHA care in 1999, but not in 1996. Veterans with higher incomes had greater odds of using VHA care in 1996, but there was no significant association between income and VHA use in 1999.
Conclusion
Characteristics of VHA users did not fundamentally change despite the reorganization of VHA health care delivery system and changes in eligibility and enrollment policy. The VHA system maintains its safety net mission while attracting more veterans.
==== Body
Background
Veterans have access to a health care system unavailable to most Americans – the Veterans Health Administration (VHA), part of the Department of Veterans Affairs. This large, integrated health care system provided care to about 4.5 million veterans in 2002. In 2001, the total veteran population was estimated at 25.3 million, accounting for about 10% of the U.S. total population [1].
The VHA has implemented significant organizational changes in recent years as mandated by the Veterans Eligibility Reform Act of 1996 [2]. The reorganization of the VHA aimed to shift the focus of care from inpatient to outpatient settings, to improve access to ambulatory care, and to provide care to more veterans. Prior to 1996, only inpatient care was mandated and outpatient care could not be provided unless there was an antecedent inpatient admission for the same problem. Legislative changes were necessary to allow VHA practitioners to provide more ambulatory care when appropriate and to free up inpatient resources that could be used to care for more veterans in outpatient settings [3].
While the VHA was shifting from inpatient to outpatient care, the VHA system was also reorganized from four large regional units into 22 Veterans Integrated Service Networks (VISNs), which became the primary budgetary and organizational unit for the health care facilities in each network. As of 2003, there are 21 VISNs. In addition, VHA was allowed by Congress to establish Community-Based Outpatient Clinics (CBOCs), linked to existing local medical centers, to improve access for veterans in suburban and rural areas [4,5]. The number of CBOCs increased from 10 in 1996 to 228 in 1999 [5], and to over 800 in 2003 [6]. VHA has transformed from a confederation of individual medical centers and clinics focused primarily on inpatient care to a fully integrated health care system that promotes primary and ambulatory care [7]. Over the 1996–2000 period, the annual number of ambulatory care visits has increased by 35%, the number of acute care hospital beds has dropped by 35%, inpatient admissions have dropped by 36%, and total bed days of care have been reduced by 68% [2].
Organizational reforms were accompanied by VHA eligibility and enrollment changes. The VHA system's historic mission has primarily been to serve as a safety net for veterans with service-related disability or low income [2]. In 1999, the VHA system was opened to all veterans based on a new enrollment system based on the Veterans Eligibility Reform Act of 1996. In the current eligibility system, vulnerable veterans who have been eligible to receive free care continue to receive free care. For many of these veterans, the VHA continues to represent an important safety net because many of them are indigent and/or vulnerable. A new group of veterans have been receiving VHA care for the first time in large numbers, including veterans without service connected disability, veterans whose level of service connected disability does not entitle them to compensation and their income is above the VHA means test threshold. These (Priority Group 7) veterans are required to make co-payments for certain types of VHA care.
On October 1, 1998, the VHA began an enrollment program that registered veterans using the health care system and assigned them a primary care provider. The enrollment program and other changes were implemented to support two goals established in the 1998–2003 Strategic Plan for the Department of Veterans Affairs: 1) to increase the number of users of the VHA health care system by 20% by 2003, and 2) to increase the percent of the operating budget obtained from revenues sources that are not appropriated by Congress, to 10% of the total [8]. VHA facilities were encouraged to generate additional revenues from sources other than Congressional appropriations and took two approaches to obtaining non-appropriated revenues. First, VHA facilities have improved billing to private insurers for veterans with private insurance while using VHA care [9]. The second approach was to attract more insured veterans. Non-elderly veterans with insurance are likely to be different from the historic population of veteran users of VHA care, because many are likely to be healthy enough to be employed and insured. In addition to generating non-appropriated revenues, increased enrollment of insured non-elderly veterans may change the overall risk profile of VHA users. Elderly veterans with insurance are almost all enrolled in Medicare, and may or may not be healthier than existing users of VHA care. The VHA has had to balance the conflicting challenges of continuing its health care safety net mission and, at the same time, attracting new users with insurance who can generate non-appropriated revenues.
Prior to these eligibility, enrollment and organizational changes, veterans seeking care at the VHA were known to be sicker than the general population [10,11]. The policies and organizational changes may attract a broader group of veterans entering the VHA system. The VHA has been a safety net for many indigent veterans and could be a new source of care for the 90% of veterans who do not seek VHA care in a given year [10,12]. In prior studies, researchers have examined VHA user or enrollee characteristics from administrative databases or the Survey of VA Enrollees [12,13]. However, there is to date no research assessing changes in veteran characteristics between veterans who use VHA services (VHA users) and those who do not use VHA services (VHA nonusers) during the period 1996–1999, with its major policy and organizational changes in the VHA.
In this study, we assess whether there were changes in demographic characteristics, health insurance coverage, and health status among veterans who used the VHA system between 1996 and 1999 when VHA underwent major organizational changes in health care delivery using the Medical Expenditure Panel Survey (MEPS). We hypothesize that these policy changes, especially the change in the enrollment policy, would attract more insured and healthier veterans to the VHA system. MEPS data, comprised of a nationally representative sample of the civilian, non-institutionalized U.S. population data, provides a unique opportunity to examine the characteristics of VHA users and VHA nonusers among veterans.
Methods
This analysis compared the characteristics of veterans who used VHA care and those who did not use VHA care using two cross-sectional samples in MEPS in 1996 and 1999.
Data Sources
The data sources were the 1996 and 1999 Household Component (HC) of the MEPS, which is co-sponsored by the Agency for Healthcare Research and Quality and the National Center for Health Statistics. We used the 1996 and 1999 Full Year Consolidated Data Files from the MEPS HC survey, which contained questions on demographic characteristics, health conditions, health status, having usual source of care, health insurance coverage, income, and employment status. The files also captured information about use and expenditures of medical services for office and hospital-based care, home health care, dental services, vision aids, and prescribed medicines. The two files covered the calendar years 1996 and 1999, respectively. For a detailed description of the MEPS sampling and survey methods, see Cohen [14].
Study Sample
Survey respondents were identified as veterans by answering affirmatively to the following question: "Did you ever serve in the Armed Forces?" The MEPS surveys included 1,944 veterans in 1996 and 1,974 veterans in 1999. The number of veterans represented 9.2% and 8.5% of the MEPS sample for 1996 and 1999, consisting of 26.6 and 25.7 million veterans in 1996 and 1999, respectively. These estimates are similar to the Census estimates, which were 26.5 million in 1996 and 26.1 million in 1999 [15,16].
Variable Definitions
VHA User and VHA Nonuser
VHA users were defined as veterans who used VHA care in 1996 and 1999, respectively. We used MEPS expenditure data to identify VHA users, based on non-zero medical expenditures paid by VHA facilities in 1996 or 1999. VHA nonusers were defined as veterans who did not have any medical expenditure paid by VHA. The definition of VHA users we used in this analysis is similar to the one used by VHA administratively to allocate resources among VHA facilities. There were 534 and 740 VHA users among veterans in the 1996 and 1999 MEPS surveys, respectively.
Health Insurance Status
Veterans were defined as entirely insured if they had either public or private insurance coverage that included both hospital and physician care throughout the entire year for 1996 and 1999, respectively. Veterans were defined as partially uninsured if they had insurance coverage for some segment of 1996 or 1999, or entirely uninsured in a year. Uninsured veterans were split about half partially uninsured and half entirely uninsured. We combined partially uninsured and entirely uninsured veterans into the uninsured group because two groups of veterans had similar characteristics.
Health Status
Two measures, perceived physical health status and mental health status, were used. Both health status measures were rated as excellent, very good, good, fair, or poor.
Employment Status
Veterans were categorized into three employment status groups – unemployed, retired, and employed. Veterans were defined as unemployed if they were unemployed throughout one entire year; retired veterans were those who did not work in a year due to retirement; and employed veterans were defined as those who worked at least for a part of a year.
Usual Source of Care
Veterans were defined as having a usual source of care if they answered yes on the question "Did you have a usual source of care provider?"
Population Weights
Additional adjustment to the MEPS population weights from the full year sample are needed to accurately estimate VHA users and VHA nonusers, because the MEPS weights do not take into account the status of VHA user versus nonuser. MEPS would allow us to generate accurate estimates for the national population of veterans for December 31, 1996 and December 31, 1999, but not for VHA users and VHA nonusers. Therefore, we applied three additional adjustments to the MEPS population weights. First, we adjusted the proportions of VHA users and nonusers in 1996 and 1999, respectively. The MEPS overestimated the numbers of VHA users and underestimated the number of VHA nonusers. Based on MEPS, there were 7.2 million VHA users in 1996 and 9.8 million in 1999, while there were 2.9 million users in 1996 and 3.4 million in 1999 based on VHA workload data [18]. Second, we adjusted the age distribution of veterans based on the Census data, because the MEPS over-weighted veterans under 55 years old and under-weighted those 55 or older. Third, we adjusted the age distribution of VHA users based on the age distribution of VHA users, because the MEPS under weighted VHA users under 55 years old, and over weighted VHA users 55 or older.
Statistical Analysis
To examine the differences in socio-demographic characteristics, health status, and having usual source of care between VHA users and non-users in 1996 and 1999, as well as among VHA users between 1996 and 1999, we used bivariate analyses that were conducted using Pearson Chi-Square or Wald tests. A multivariate logit analysis was used to further explore factors associated with using VHA care in 1996 and 1999, respectively. All analyses were adjusted for the complex survey design used in the MEPS using survey analysis techniques in STATA SE 8 [17].
Results and Discussion
Proportion of Veterans Using the VHA System
The percentage of veterans using the VHA system increased from 12.4% in 1996 to 14.6% in 1999 (p = 0.03). The result indicates a significant gain in VHA market share among veterans. This result is consistent with VHA data showing that the number of VHA users increased by about 17% during the same time period (2.9 million in 1996 and 3.4 million in 1999) [16].
Characteristics of VHA Users versus Nonusers
Table 1 describes characteristics of VHA users and VHA nonusers in 1996 and 1999. In 1996, there were statistically significant differences between VHA users and nonusers in marital status, education, incomes, general and mental health status, employment status, and having usual source of care. As expected, VHA users were more likely to be not married and unemployed (12% for VHA users versus 5% for VHA nonusers) than VHA nonusers. Compared to VHA nonusers, users had lower incomes ($24,854 versus $30,222) and lower education level (26% with college or higher versus 42%). In addition, compared to VHA nonusers, more VHA users reported having fair or poor health status (29% versus 12%) as well as fair or poor mental health status (13% versus 4%). The majority of VHA users and nonusers were insured (74% versus 87%) partly because most elderly veterans (age ≥ 65) have Medicare coverage. Among veterans under age 65, VHA users were less likely to have insurance than VHA nonusers (60% versus 81%, p < 0.01). However, VHA users were more likely to have a usual source of care than non-users (87% versus 79%).
Table 1 Characteristics of VHA users and non-users in 1996 and 1999
1996 1999
Characteristics VHA Users VHA Nonusers VHA Users VHA Nonusers
MEPS Sample Size 534 1410 740 1234
Mean Age 58 57 59 57
%Male 94 96 93 95
%White 86 89 87 88
%Marrieda 67 76 72 72
%Living in the MSA Area 77 80 76 80
Employment Status
%Retired 31 28 29 25
%Unemployed in the entire year a,b 12 5 15 5
%Insured a,b 74 87 81 87
%Insured for veterans < 65 yearsa,b 60 81 67 81
Annual Income (in 1999 dollar) a,b $24,854 $30,222 $27,593 $33,151
(standard error) (1,429) (875) (1,145) (783)
Education a
%Less than high school 24 16 22 20
%High school 50 51 51 50
%College degree or higher 26 33 27 30
Perceived Health Statusa,b,c
%Excellent or very good 45 65 42 61
%Good 26 23 34 28
%Fair or poor 29 12 24 11
Perceived Mental Health Statusa,b
%Excellent or very good 63 75 61 73
%Good 24 20 29 23
%Fair or poor 13 4 10 4
%Having Usual source of Carea,b 87 79 89 78
a Indicates a significant difference at p < 0.05 level between users and non-users in 1996
b Indicates a significant difference at p < 0.05 level between users and non-users in 1999
c Indicates a significant difference at p < 0.05 level among users between 1996 and 1999
d Indicates a veteran who had health insurance throughout the year
Patterns similar to those in 1996 were observed in 1999 regarding annual income, general and mental health status, employment status, insurance status, and having usual source of care. In 1999, there were no significant differences in marital status and education between VHA users and VHA nonusers.
Characteristics of VHA Users between 1996 and 1999
In Table 1, we compare characteristics of VHA users between 1996 and 1999. There was a statistically significant difference in perceived health status among VHA users between 1996 and 1999, while the remaining characteristics were not significantly different among VHA users between the two years. Compared to 1996 VHA users, 1999 VHA users were less likely to report health status as fair or poor (29% in 1996 versus 24% in1999), and more likely to report health status as good (34% in 1999 versus 26% in 1996).
Factors Associated with Using VHA Services
In Table 2, we summarize factors associated with using VHA services, based on logit regressions for 1996 and 1999, respectively. In 1996, age, gender, education, income, employment status, lack of insurance coverage, and perceived health status are significantly associated with the probability of using VHA care among veterans. Compared to veterans in the age group 18 – 44, the odds of using the VHA system increased with age (age 45 – 54 OR = 4.92, p < 0.01; age 55–64 OR = 4.67, p < 0.01; age ≥ 65 OR = 8.60, p < 0.01). Male veterans were more likely to use the VHA system than female veterans (OR = 20.55, p < 0.01). Veterans with high school education were more likely to use the VHA system than those with less than high school education (OR = 1.78, p < 0.01). The odds of using the VHA system significantly increased with income (OR = 1.01, p < 0.01).
Table 2 Significant Factors Associated with Using VHAServices by Year Based on a Logistic Analysis1
1996 1999
Variable Odds Ratio 95% CI Odds Ratio 95% CI
Age Group (index group = 18–44)
45–54 4.92** 3.43–7.05 4.43** 2.81–6.98
55–64 4.67** 3.21–6.80 4.43** 2.91–6.74
≥ 65 8.60** 5.47–13.53 9.71** 6.52–14.46
Gender (Male = 1; Female = 0) 20.55** 12.69–33.30 15.8** 7.91–31.55
Race (White = 1; 0 otherwise) 1.08 0.79–1.47 1.12 0.84–1.50
Marital Status (Married = 1; 0 otherwise) 1.00 0.74–1.34 1.54** 1.17–2.03
Living in MSA (MSA = 1; 0 otherwise) 0.89 0.69–1.14 0.87 0.70–1.08
Education (index group = less than high school)
High school 1.78** 1.31–2.40 1.60** 1.20–2.14
College or more 0.97 0.67–1.42 1.14 0.80–1.62
Income (in $1,000) 1.01** 1.00–1.02 1.00 0.99–1.01
Employment Status (index group = employed)
Retired 1.70** 1.14–2.52 1.44* 1.06–1.92
Unemployed 1.68** 1.15–2.45 1.61** 1.07–2.41
Uninsured2 (index group = insured) 2.15** 1.6–2.90 1.78** 1.27–2.50
Perceived health status (index group = excellent or very good)
Good 1.42* 1.07–1.89 1.71** 1.36–2.15
Fair or poor 2.12** 1.56–2.89 2.26** 1.69–3.03
Perceived mental health status (index group = excellent or very good)
Good 0.93 0.69–1.25 0.94 0.74–1.19
Fair or poor 1.30 0.87–1.96 1.04 0.70–1.55
Note: * p < 0.05; ** p < 0.01
1. The dependent variable is the VHA user status with 1 indicating VHA users and 0 indicating nonusers.
2. Indicates veterans who were uninsured throughout the year, or a portion of the year.
Compared to employed veterans, both retired and unemployed veterans were more likely to use the VHA system. The odds ratio was 1.70 for retired veterans (p < 0.01) and 1.68 for unemployed veterans (p < 0.01). Furthermore, the odds ratio of using the VHA system for uninsured veterans was 2.15 (p < 0.01) compared to insured veterans. Finally, the odds of using VHA care increased as self-reported health status worsened. Compared to veterans in excellent or very good health status, the odds ratio of using the VHA system was 1.42 (p < 0.05) for those in good health status and 2.12 (p < 0.01) for those in fair or poor health status.
Similar veteran characteristics associated with using VHA care found in 1996 were also found in 1999, with the exception of marital status and income. Marital status was not a significant factor in 1996, but married veterans were more likely to use VHA than unmarried veterans in 1999 (OR = 1.63, p < 0.01), controlling for other factors. In contrast, income was not a significant factor in 1999, while it was significantly associated with using VHA care in 1996.
To further examine the changes in characteristics between the two years, we conducted a sensitivity analysis by pooling the data from 1996 and 1999 to examine the interactions between the year indicator and individual characteristics. The analysis shows that the results are consistent with those in the separate logit regression analyses.
Conclusion
We provided an overview of changes in health insurance status and demographic characteristics of veterans using the VHA health care system during a period when VHA implemented major organizational changes between 1996 and 1999. The VHA system provides a special safety net for a significant number of veterans that is not available to most Americans. The proportion of veterans using the VHA system increased significantly from 12.4% in 1996 to 14.6% in 1999. This pace has continued as the number of veterans treated in the VHA health care system increased by 36% between 1998 and 2002 [18]. In January 2003, VHA established a policy to stop enrolling Priority Group 7 veterans after January 16, 2003 due to the increase in enrollment and lack of resources.
The results indicate that VHA users characteristics did not change substantially following the major reorganization and policy changes. Only two variables, marital status and income, are different between the two years. Married veterans were more likely to use VHA care in 1999, but not in 1996. In addition, the odds of using VHA care increased with income in 1996, while income was not statistically significant in 1999. As a result, we reject the hypothesis that more veterans with health insurance and healthier veterans use VHA care after the major reforms of 1996. The fact that we did not observe changes in VHA user characteristics may suggest that sizeable numbers of veterans needed care but were not eligible to receive it through VHA prior to the changes in the VHA systems in 1996.
Our results suggest that being uninsured, unemployed, in poor health status remain significant factors associated with using the VHA system in the both years. Even though VHA users more likely to be uninsured than VHA nonusers, VHA users were more likely to report having a usual source of care than VHA nonusers. Therefore, VHA appears to have continued its mission as a safety net to serve vulnerable veterans who were in poor health, uninsured, or unemployed.
In addition, the findings indicate that a majority of VHA users had alternative insurance coverage, such as Medicaid, Medicare, or private health insurance. This may have important implications for quality of care and coordination of care [13]. For patients with access to VHA and other health care systems, there is an increasing need for communication between VHA and non-VHA providers to ensure continuity and quality of care. Further research on strategies of sharing information among providers in VHA and other health care systems is needed to improve care for patients using both VHA and non-VHA care.
This study has two notable limitations. The MEPS study sample, including the veteran sample, is potentially biased because the households in MEPS were recruited via mail or telephone contact from the sample of the National Health Interview Survey. Homeless veterans, a highly vulnerable group, are likely to be under-sampled because they have no permanent residence or telephone access. Second, this is a cross-sectional study rather than a cohort study. Therefore, we were unable to evaluate the discrete impact of individual policy change between 1996 and 1999. However, using the two cross-sectional MEPS samples provides a unique opportunity to examine the differences in the characteristics of VHA veteran users and non-users during the period when VHA experienced major policy and organizational changes.
In summary, the VHA health care system is the nation's largest integrated health care system, providing a special health care safety net for indigent and vulnerable veterans. We observed that characteristics of VHA users did not change following the reorganization of the VHA health care delivery system and changes in eligibility and enrollment policy. VHA appears to have continued its mission as a safety net to serve vulnerable veterans who were in poor health, uninsured, or unemployed after major organizational changes.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CFL, MLM, and AEBS designed the conceptual framework for the study. CFL designed the method, conducted data analysis, and drafted the manuscript. CFL, MLM, and AES interpreted the results.
Acknowledgements
This research was supported by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service Project. Drs. Liu, Maciejewski, and Sales are presently investigators at the VA Puget Sound Health Care System's HSR&D Center of Excellence. The views expressed in this report are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or the Health Services Research and Development Service or the University of Washington.
==== Refs
Statistical Abstract of the United States City: US Census Bureau 2002
Kizer KW Demakis JG Feussner JR Reinventing VA health care: systematizing quality improvement and quality innovation Med Care 2000 38 I7 16 10843266 10.1097/00005650-200006001-00002
Kizer KW Pane GA The "new VA": delivering health care value through integrated service networks Ann Emerg Med 1997 30 804 7 9398778
Fortney JC Borowsky SJ Hedeen AN Maciejewski ML Chapko MK VA community-based outpatient clinics: access and utilization performance measures Med Care 2002 40 561 9 12142771 10.1097/00005650-200207000-00002
Chapko MK Borowsky SJ Fortney JC Hedeen AN Hoegle M Maciejewski ML VanDeusen Lukas C Evaluation of the Department of Veterans Affairs community-based outpatient clinics Med Care 2002 40 555 60 12142770 10.1097/00005650-200207000-00001
Department of Veterans Affairs Facts about the Department of Veterans Affairs Press Release 2003
Department of Veterans Affairs Resource Allocation for Veterans' Health Care Press Release 1999
Department of Veterans Affairs Department of Veterans Affairs Strategic Plan, Fiscals Years 1998–2003 Office of the Assistant Secretary for Policy Planning 1997
Department of Veterans Affairs Medical Care Cost Recovery Press Release 2002
Wilson NJ Kizer KW The VA health care system: an unrecognized national safety net Health Aff (Millwood) 1997 16 200 4 9248165 10.1377/hlthaff.16.4.200
Agha Z Lofgren RP VanRuiswyk JV Layde PM Are patients at Veterans Affairs medical centers sicker? A comparative analysis of health status and medical resource use Arch Intern Med 2000 160 3252 7 11088086 10.1001/archinte.160.21.3252
Ashton CM Souchek J Petersen NJ Menke TJ Collins TC Kizer KW Wright SM Wray NP Hospital use and survival among Veterans Affairs beneficiaries N Engl J Med 2003 349 1637 46 14573736 10.1056/NEJMsa003299
Shen Y Hendricks A Zhang S Kazis LE VHA enrollees' health care coverage and use of care Med Care Res Rev 2003 60 253 67 12800686 10.1177/1077558703060002007
Cohen J Design and Methods of the Medical Expenditure Panel Survey Household Component Agency for Health Care Policy and Research 1997
Statistical Abstract of United States US Census Bureau 1997
Statistical Abstract of United States US Census Bureau 2000
STATA Reference Manual College Station, TX 2003
Department of Veterans Affairs VA Health Care, Systemwide Workload, FY1997-FY2002 2003
| 15836789 | PMC1090608 | CC BY | 2021-01-04 16:37:16 | no | Health Res Policy Syst. 2005 Apr 18; 3:5 | utf-8 | Health Res Policy Syst | 2,005 | 10.1186/1478-4505-3-5 | oa_comm |
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-41582901410.1186/1479-5868-2-4ResearchDiets and selected lifestyle practices of self-defined adult vegetarians from a population-based sample suggest they are more 'health conscious' Bedford Jennifer L [email protected] Susan I [email protected] Human Nutrition, University of British Columbia, 2205 East Mall, Vancouver, BC, Canada2005 13 4 2005 2 4 4 15 12 2004 13 4 2005 Copyright © 2005 Bedford and Barr; licensee BioMed Central Ltd.2005Bedford and Barr; 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
Few population-based studies of vegetarians have been published. Thus we compared self-reported vegetarians to non-vegetarians in a representative sample of British Columbia (BC) adults, weighted to reflect the BC population.
Methods
Questionnaires, 24-hr recalls and anthropometric measures were completed during in-person interviews with 1817 community-dwelling residents, 19–84 years, recruited using a population-based health registry. Vegetarian status was self-defined. ANOVA with age as a covariate was used to analyze continuous variables, and chi-square was used for categorical variables. Supplement intakes were compared using the Mann-Whitney test.
Results
Approximately 6% (n = 106) stated that they were vegetarian, and most did not adhere rigidly to a flesh-free diet. Vegetarians were more likely female (71% vs. 49%), single, of low-income status, and tended to be younger. Female vegetarians had lower BMI than non-vegetarians (23.1 ± 0.7 (mean ± SE) vs. 25.7 ± 0.2 kg/m2), and also had lower waist circumference (75.0 ± 1.5 vs. 79.8 ± 0.5 cm). Male vegetarians and non-vegetarians had similar BMI (25.9 ± 0.8 vs. 26.7 ± 0.2 kg/m2) and waist circumference (92.5 ± 2.3 vs. 91.7 ± 0.4 cm). Female vegetarians were more physically active (69% vs. 42% active ≥4/wk) while male vegetarians were more likely to use nutritive supplements (71% vs. 51%). Energy intakes were similar, but vegetarians reported higher % energy as carbohydrate (56% vs. 50%), and lower % protein (men only; 13% vs. 17%) or % fat (women only; 27% vs. 33%). Vegetarians had higher fiber, magnesium and potassium intakes. For several other nutrients, differences by vegetarian status differed by gender. The prevalence of inadequate magnesium intake (% below Estimated Average Requirement) was lower in vegetarians than non-vegetarians (15% vs. 34%). Female vegetarians also had a lower prevalence of inadequate thiamin, folate, vitamin B6 and C intakes. Vegetarians were more likely than non-vegetarians to consider various health conditions and food/nutrition concerns when choosing foods.
Conclusion
In this population-based study, evidence was obtained to indicate that vegetarians appear more 'health conscious' than non-vegetarians, although specific differences were not always consistent by gender. Additional population-based studies are required to determine if the observed gender differences exist in other populations.
DietvegetarianHealth behaviorFood habitsHealth attitudes and behaviors.
==== Body
Background
Interest in the dietary habits of vegetarians emerges from research suggesting that vegetarians have a lower prevalence of various chronic diseases that currently plague the developed world [1,2]. It has been hypothesized that these findings are due to vegetarians' dietary habits, which more closely follow recommendations for healthy eating [3], and also to their lower BMI [4]. Recently, however, it has been found that mortality rates did not differ among vegetarians and similar 'health conscious' omnivores [5,6] despite vegetarians' lower age-adjusted BMI [5-7]. This suggests that other lifestyle behaviors commonly observed in health conscious individuals may be responsible for the observed beneficial health effects. Yet, it is important to note that the majority of reports of vegetarians' dietary intakes and lifestyle behaviors are from convenience samples.
Few population-based studies have examined vegetarians' dietary intake and habits. Kennedy and coworkers used population representative data from the Continuing Survey of Food Intakes by Individuals (CSFII) to gain insights into the dietary patterns of vegetarians [8]. Specifically, they compared the intakes of those who consumed meat on data collection days to those who did not. However, many of those who did not consume meat were likely not 'vegetarian' as the proportion of those grouped as vegetarian by this method (15%) was much higher than those who self-identified as vegetarian in the same sample (2.5%) [9].
CSFII data were also analyzed to compare self-identified vegetarians to non-vegetarians [9]. Adult vegetarians were found to have lower BMI as well as lower intakes of total fat, saturated fat and cholesterol and higher intakes of fiber and fruits [9]. However, differences by gender or age were not examined, and if age or gender differences in the prevalence of vegetarian status exist, as has been observed in other population representative studies of vegetarians [10], these could confound the results. For example, if vegetarians were more likely than non-vegetarians to be younger and female this could impact group BMIs and intakes of nutrients that differ by gender or age. In addition, differences in lifestyle behaviors were not assessed. This could be important to consider as it has been suggested that other lifestyle behaviors may be the determinant of differences observed in health conditions by vegetarian status [5,6].
Thus, the aim of the present analysis was to describe and compare the demographics, lifestyle behaviors, dietary intake, nutritive supplement use, and food and nutrition concerns of male and female self-defined vegetarians and non-vegetarians from a population-based representative sample of adults from the province of British Columbia, Canada.
Methods
The data used for this analysis were collected as part of the British Columbia Nutrition Survey (BCNS). Details of the methodology used for the BCNS including sampling strategies, survey instruments, and data entry and analysis are described in detail elsewhere [11], and are summarized below.
Participants
Adults aged 19 to 84 years living in BC were recruited for the BCNS using the BC Health Registry – a central repository of individuals who receive health services in BC. Exclusion criteria included those living in care or correctional facilities, military bases, or Indian Reserves, as well as pregnant and lactating women. Less than 3% of the population was excluded on these grounds. The sample was stratified by age, sex and geographical region. The study protocol was approved by the University of British Columbia's Behavioral Research Ethics Board, and written informed consent was obtained from all participants.
Measures
The BCNS included a 24-hour recall; questions on food habits and choices, physical activity and demographics; and anthropometric measurements.
24-Hour Recall
To obtain information on dietary intake each individual completed a 24-hour recall conducted by personal interview using the open-ended, multiple-pass technique for which each participant recalled all foods and beverages consumed on the previous day (midnight to midnight). Food models and household measures were used to estimate portion sizes. In one third of the sample, a second 24-hour recall was conducted on a different weekday at least one week following the first recall. Supplement data were also obtained during the 24-hour recall. Participants were first asked whether they took any nutritional supplements yesterday followed by a question about whether they took any supplements within the past month that differed from the ones taken yesterday. The frequency (daily, weekly or monthly) and the number or amount of each supplement were also recorded. When possible, brand names and the drug identification number (DIN) were recorded.
Food and Nutrition Habits
Participants were read a list of health-related reasons for choosing and avoiding foods to determine nutrition and food concerns. Vegetarian status was assessed by asking participants if they considered themselves to be a vegetarian. Those who answered 'yes' were also asked if they ever ate animal products including dairy, eggs, fish/seafood, poultry and red meat.
Physical Activity
Questions from previously validated instruments were asked to obtain information on physical activity, including the frequency of mild, moderate and strenuous activity; and motivational readiness for exercise [12,13].
Demographic Questionnaire
Questions regarding age, gender, marital status, education and income were asked to characterize the sample.
Anthropometrics
Weight, without shoes, hats or any heavy clothing or items, was measured using a weekly calibrated electronic scale. Height was measured using a setsquare and measuring tape, and girths were assessed using a measuring tape. Weight and height were measured and recorded once; waist circumferences was measured and recorded at least twice. BMI was calculated from weight and height (kg/m2).
Procedure
Eligible residents who chose not to participate were asked to complete a non-response questionnaire to determine if non-responders differed from those who agreed to participate. Those who agreed to participate were interviewed in person by trained interviewers. Interviews lasted approximately 90 minutes and most took place in participants' homes.
Analysis
Response rate
In a large population survey such as the BCNS, the response rate depends on whether individuals that could not be contacted (unresolved cases) were eligible to participate in the survey. Lower and upper bounds for the response rate can be calculated based on the assumption that all unresolved cases are eligible (lower bound), or that all unresolved cases are ineligible (upper bound).
Nutrient intake data
All data were sent to the Bureau of Nutritional Sciences at Health Canada and were entered into the Nutrition Survey System, a software program that uses the Canadian Nutrient File and a recipe database adapted from the United States Department of Agriculture CSFII. The Canadian Nutrient File was updated to reflect fortification of grain products with folic acid that began in Canada in 1998. Data on nutritional supplements were entered using the DIN or by name and/or nutrient content. The method of estimating the distribution of usual intakes from food sources alone and from food sources and supplements used the data from duplicate 24-hour recalls to remove within-person variability from population distributions of nutrient intakes, yielding an adjusted distribution of usual intakes for age-sex groups [11,14]. Then the monthly supplement data were expressed per day and added to the adjusted usual intake distribution. The proportion of this distribution that fell below the Estimated Average Requirement (EAR) was used to estimate the prevalence of inadequate nutrient intakes in an age-sex group, for nutrients with an EAR and a symmetrical requirement distribution [14]. This analysis could not be conducted for vitamin A as the EAR is expressed in new Retinol Activity Equivalents, whereas intake data were in Retinol Equivalents. Food intake data were also expressed as number of servings from the food groups included in Canada's Food Guide to Healthy Eating (CFGHE).
Statistical analysis
All data were weighted to reflect the BC population based on gender, age, and geographical region, and were analyzed using the Statistical Package for Social Sciences (SPSS; v11.0, Chicago, Ill., 2002). One-way ANOVA with age as a covariate was used to analyze parametric data (demographics, nutrient intakes) and Chi-square analysis was used to analyze categorical data (lifestyle behaviors, supplement usage, nutrition concerns) between vegetarians and non-vegetarians. The Mann-Whitney test was used to compare differences by vegetarian status in supplemental nutrient intake as the data were not normally distributed. Non-parametric tests were also applied to nutrient intake data that were not normally distributed. However, findings did not differ from those of parametric analysis and thus ANOVA was applied to all nutrients for consistency, and because ANOVA allowed consideration of effects of covariates. The data were also examined to assess whether demographic differences between vegetarians and non-vegetarians (other than age and sex) may have influenced the results. The significance level was set at p = 0.05 for all statistical measures.
Results
Response rate lower and upper bounds were 42% and 52%, respectively [11]. Approximately 66% of those who declined to participate completed the non-respondent survey. Using this information, it was found that BCNS participants were less likely to smoke (17% vs. 23%) and more likely to use vitamin/mineral supplements (66% vs. 60%) and hold university degrees than non-participants (14% vs. 9%). However, in comparison to the general BC population, BCNS participants had a similar prevalence of smoking, although study participants were more educated (21% vs. 13% completed university) and more men were married (64% vs. 55%). Because of the difference in educational attainment between survey participants and the population, the effect of this variable on nutrient intakes was examined. Educational attainment was associated only with the intakes of vitamin C and vitamin B12: it appeared that those with more education had higher intakes of vitamin C and lower intakes of vitamin B12 [11].
Of the 1817 participants, 5.8% (n = 106) identified themselves as vegetarian. Those who identified themselves as vegetarian were asked if they 'ever' ate various animal products. It appeared that the majority of self-identified vegetarians did not adhere rigidly to a vegetarian dietary pattern: In terms of tissue protein, 74.9% consumed fish and/or seafood at least occasionally, while 57.6% consumed poultry and 22.4% consumed red meat at least occasionally. Dairy products were used at least occasionally by 97.3% and eggs by 92.3%.
Demographics
The demographic characteristics of vegetarian and non-vegetarian participants are presented in Table 1. Vegetarians tended to be younger than non-vegetarians (p = 0.057), and the age group distribution differed significantly, with more vegetarians falling in the 19 to 30 year range. Groups differed in sex distribution with women representing over 70% of vegetarians and half of non-vegetarians. Although the majority of both groups was married, there was a significant difference in marital status distribution, with vegetarians more likely to be single. In addition, vegetarians were significantly more likely to be of low income status although there were no differences in education level. Accordingly, data were examined to assess the effects of marital status and income status.
Table 1 Participant demographics by vegetarian status
Non-vegetarian (n = 1711) Vegetarian (n = 106) Test Statistic1 P value
Age (years, mean ± SE) 44.8 ± 0.4 41.5 ± 1.8 F = 3.63 0.057
Age Group χ2 = 9.80 0.020
19 – 30 years 24.8% 38.1%
31 – 50 years 40.9% 31.4%
50 – 70 years 24.3% 20.0%
71+ years 10.1% 10.5%
Sex (% female) 49.2% 70.8% χ2 = 18.58 <0.001
Marital status χ2 = 17.55 0.001
Single 22.5% 35.2%
Married 62.7% 50.5%
Widowed 5.3% 10.5%
Divorced/separated 9.5% 3.8%
Education level χ2 = 0.45 0.800
Secondary or less 33.3% 32.1%
Technical or some university 47.6% 46.2%
University graduate 19.1% 21.7%
Low income 22.8% 37.2% χ2 = 9.28 0.002
Weight (kg, mean ± SE)
Men 83.1 ± 0.6 82.2 ± 3.0 F = 0.10 0.754
Women 67.9 ± 0.6 62.5 ± 1.8 F = 7.92 0.005
BMI (kg/m2, mean ± SE)
Men 26.7 ± 0.2 25.9 ± 0.8 F = 0.81 0.346
Women 25.7 ± 0.2 23.1 ± 0.7 F = 13.66 <0.001
BMI category – men χ2 = 0.92 0.821
Underweight (< 18) 0.2% 0.0%
Normal weight (18-<25) 48.8% 51.6%
Overweight (≥ 25-<30) 31.6% 35.5%
Obese (≥ 30) 19.4% 12.9%
BMI category – women χ2 = 16.34 0.001
Underweight (< 18) 1.1% 0.0%
Normal weight (18-<25) 59.6% 83.1%
Overweight (≥ 25-<30) 21.2% 12.7%
Obese (≥ 30) 18.1% 4.2%
Waist Circ. (cm, mean ± SE)
Men 91.7 ± 0.4 92.5 ± 2.3 F = 0.11 0.740
Women 79.8 ± 0.5 75.0 ± 1.5 F = 9.66 0.002
Health Condition – Men (% yes)
Diabetes 5.5 3.2 χ2 = 0.31 0.580
Heart disease 6.1 19.4 χ2 = 8.59 0.003
Stroke 0.8 3.2 χ2 = 1.99 0.158
High blood pressure 11.6 6.5 χ2 = 0.79 0.375
High cholesterol 14.7 9.7 χ2 = 0.61 0.434
Cancer 4.5 6.5 χ2 = 0.27 0.606
Osteoporosis 0.7 0.0 χ2 = 0.22 0.643
None of the above 70.9 63.3 χ2 = 0.81 0.370
Health Condition – Women (% yes)
Diabetes 4.9 1.3 χ2 = 1.97 0.160
Heart disease 3.7 4.0 χ2 = 0.02 0.889
Stroke 1.9 0.0 χ2 = 1.45 0.228
High blood pressure 15.3 6.7 χ2 = 4.15 0.042
High cholesterol 11.3 6.8 χ2 = 1.44 0.230
Cancer 8.3 1.3 χ2 = 4.70 0.030
Osteoporosis 6.1 8.0 χ2 = 0.44 0.506
None of the above 67.7 78.4 χ2 = 3.62 0.057
1. Statistical analysis consisted of ANOVA (vegetarian versus non-vegetarian, with age as a covariate) for continuous variables. Chi-square analysis was used for categorical variables.
Female vegetarians had a significantly lower mean age-adjusted body weight and mean BMI than non-vegetarians, as well as a lower waist circumference. Low income status and marital status did not affect these variables. In addition, vegetarian women were significantly less likely to be classified as overweight or obese (17% vs. 40%). Conversely, for males, weight, BMI, and BMI category distribution were very similar between vegetarians and non-vegetarians, with approximately 50% of both groups classified as overweight or obese. There were no significant differences in age-adjusted waist circumference between vegetarian and non-vegetarian men.
The prevalence of certain health conditions differed by vegetarian status. Male vegetarians had a higher prevalence of heart disease while female non-vegetarians were more likely to report cancer and hypertension.
Lifestyle behaviors
Female vegetarians were more likely than non-vegetarians to report moderate to strenuous physical activity four or more times weekly (69% vs. 42%, χ2 = 21.69, p < 0.001), and more women vegetarians than non-vegetarians were in the 'action' or 'maintenance' stages of motivational readiness for exercise (76% vs. 53%, χ2 = 21.67, p < 0.001). Although single women in the sample as a whole were more active than women who were married, widowed, divorced or separated, when age was considered marital status did not affect physical activity level. Low income status was not associated with physical activity in women. In contrast, male vegetarians and non-vegetarians did not differ in the amount of weekly exercise: the majority of both groups participated in moderate to strenuous physical activity less than four times a week (55% vs. 52% respectively, χ2 = 0.18, p = 0.913). Men were also similar in terms of the distribution of the stage of motivational readiness for exercise (χ2 = 1.78, p < 0.776). On the other hand, smoking status differed between vegetarians and non-vegetarians for both men (3% vs. 18%, χ2 = 4.15, p < 0.05) and women (0% vs. 18%, χ2 = 15.85, p < 0.001).
Supplement use and intakes
The majority of all groups reported nutritive supplement use. Among men, significantly more vegetarians than non-vegetarians reported using supplements (71% vs. 51%, χ2 = 4.76, p = 0.029). However, for women, the difference was not significant with 76% of vegetarians and 68% of non-vegetarians reporting supplement use (χ2 = 1.88, p = 0.170). On the other hand, female vegetarians who used supplements reported using a higher number of supplements than non-vegetarians (3.5 ± 0.4 (mean ± SE) vs. 2.3 ± 0.1, F = 15.33, p < 0.001), while the numbers used by males were similar (1.3 ± 0.4 vs. 1.5 ± 0.1, F = 0.40, p = 0.53).
The proportion of individuals using supplements of many nutrients differed significantly by vegetarian status (data not shown). More vegetarians than non-vegetarians of both sexes used a supplement containing the B vitamins. However, for other nutrients, the results were split along gender lines: more female vegetarians used supplements of all other vitamins/minerals except for vitamin E and calcium; whereas among males, there were no additional significant differences by vegetarian status.
Differences in supplemental nutrient intake were also evident between vegetarians and non-vegetarians who used supplements (data not shown). Among women, vegetarians had significantly higher median supplemental intakes of calcium, iron, magnesium, potassium, niacin, folic acid and vitamins A, D and B12. Among men, vegetarians had a significantly higher median supplemental intake of vitamin C.
Energy and nutrient intakes
Age-adjusted energy and nutrient intakes from food are presented by vegetarian status and gender (see Additional File 1). Income status did not affect nutrient intakes [11]. There were no significant differences in energy intake between vegetarians and non-vegetarians, but energy distribution differed significantly by vegetarian status for both sexes. Compared to non-vegetarians, both male and female vegetarians consumed significantly more energy as carbohydrate. Among men, vegetarians had a significantly lower proportion of energy from protein. Conversely, female vegetarians had a significantly lower percentage of energy from fat.
Male and female vegetarians had significantly higher intakes of fiber, magnesium and potassium. Female vegetarians had significantly higher intakes of carbohydrate (g), phosphorus, thiamin, pantothenic acid, vitamin B6, and folate and lower intakes of saturated fat and sodium. Conversely, male vegetarians had significantly higher intakes of vitamin C and calcium, and lower intakes of protein (g), niacin and cholesterol.
The prevalence of inadequate intakes of selected nutrients by vegetarian status and gender are shown in Figure 1. These data are based on combined intake from food plus supplements. No differences were observed in the prevalence of inadequate intakes of vitamin B12 or zinc, but significantly more non-vegetarians had intakes below the EAR for magnesium in both men and women. There were also significant differences for both genders in the prevalence of inadequacies for thiamin, although in both cases the prevalence of inadequacy was <10%: for men, more vegetarians were below the EAR, while for women, non-vegetarians were more likely to be below the EAR. Female non-vegetarians were also significantly more likely to be below the EAR for vitamin C, vitamin B6 and folate.
Figure 1 Prevalence of nutrient inadequacies by vegetarian status and gender for selected nutrients. The prevalence of nutrient inadequacy was estimated by determining the proportion of the usual intake distribution (from food plus supplements) that was below the Estimated Average Requirement (EAR). * Prevalence of inadequacy higher in non-vegetarians (p < 0.05). NV = non-vegetarians, V = vegetarians.
Intake based on Canada's Food Guide to Healthy Eating
Analyses of dietary intake based on servings of CFGHE food groups by vegetarian status and gender, adjusted for age, are presented in Table 2. Vegetarians of both genders had a significantly higher number of servings of fruits and vegetables. Only female vegetarians had a significantly higher number of servings of grain products while only male vegetarians had a significantly higher number of servings of milk products and a significantly lower number of servings of meat and alternatives.
Table 2 Canada's Food Guide servings and percentage of participants meeting minimum recommendations by vegetarian status and gender.
Food Group Servings (Mean ± SE) % Meeting Recommendation
Non-vegetarian Vegetarian Non-vegetarian Vegetarian
Grain Products1
♂ 7.7 ± 0.16 7.9 ± 0.81 70.5 67.7
♀ 5.0 ± 0.09 5.9 ± 0.31** 45.5 57.3*
Fruit and Vegetables1
♂ 5.3 ± 0.13 7.3 ± 0.70** 45.5 64.5*
♀ 4.5 ± 0.12 5.6 ± 0.40** 34.0 40.0
Milk and Milk Products2
♂ 1.7 ± 0.06 3.0 ± 0.29*** 33.5 67.7***
♀ 1.4 ± 0.04 1.5 ± 0.14 25.3 33.3
Meat and Alternatives3
♂ 4.5 ± 0.12 1.9 ± 0.6*** 78.1 51.6***
♀ 2.7 ± 0.07 2.6 ± 0.24 57.8 45.3*
* p < 0.05, ** p < 0.01, *** p < 0.001
1 Minimum serving recommendation is 5 servings.
2 Minimum serving recommendation is 2 servings.
3 Minimum serving recommendation is 2 servings; servings calculated as 50 g equivalent
The proportions of participants meeting the minimum number of CFGHE servings by vegetarian status and gender are also displayed in Table 2. Vegetarians of both genders were less likely to meet the minimum recommendations for meat and alternatives. Among women, vegetarians were more likely to meet the minimum servings of grain products, while among men, vegetarians were more likely to meet recommendations for fruits and vegetables as well as milk products.
Food and nutrition concerns
For both genders, vegetarians were significantly more likely to report 'maintaining/improving health' as a consideration when choosing/avoiding foods than non-vegetarians (men: 100% vs. 65%, p < 0.001; women: 93% vs. 77%, p = 0.001). Male vegetarians were more likely than non-vegetarians to also consider heart disease (77% vs. 38%, p < 0.001) and high blood pressure (45% vs. 25%, p = 0.013) when choosing/avoiding foods. Female vegetarians were more likely than non-vegetarians to also consider cancer (41% vs. 30%, p = 0.05), osteoporosis (61% vs. 38%, p < 0.001) and food allergies/intolerances (43% vs. 30%, p = 0.026), and were less likely than non-vegetarians to consider weight gain (46% vs. 61%, p = 0.013) when choosing/avoiding foods. Finally, more non-vegetarians than vegetarians reported that they did not consider any of the aforementioned factors (maintaining/improving health, heart disease, cancer, osteoporosis, high blood pressure, weight gain, food allergies/intolerances) when choosing/avoiding foods (men: 26% vs. 0%, p = 0.001; women: 12% vs. 1%, p = 0.004).
Vegetarians of both genders were more likely than non-vegetarians to report choosing foods because of the nutrients they contain (men: 73% vs. 53%, p = 0.025; women: 88% vs. 68%, p < 0.001), and to report avoiding foods because of their fat content (men: 77% vs. 59%, p = 0.041; women: 84% vs. 72%, p = 0.023). Male vegetarians were more likely than non-vegetarians to also consider the type of fat (63% vs. 36%, p = 0.003) and the amount of unsaturated fat (63% vs. 28%, p < 0.001) when choosing foods and to avoid foods because of their cholesterol (55% vs. 36%, p = 0.032) and saturated fat (58% vs. 38%, p = 0.021) content. Female vegetarians were more likely than non-vegetarians to also report avoiding foods because of their salt content (57% vs. 45%, p = 0.043). On the other hand, more non-vegetarians than vegetarians reported that, when choosing foods, they did not consider any of nutrient content, type of fat, amount of unsaturated fat, or fiber content (men: 34% vs. 16%, p = 0.035; women: 19% vs. 3%, p < 0.001). When avoiding foods, more female non-vegetarians reported that they did not consider any of the fat, salt, cholesterol, sugar or saturated fat content (16% vs. 5%, 0 = 0.014).
Discussion
The purpose of this study was to examine and compare the dietary habits and lifestyle behaviors of self-defined vegetarians and non-vegetarians from a population-based representative sample of BC adults. Approximately 6% of the sample, weighted to reflect the BC population, reported being vegetarian. The findings of this study suggest that the dietary habits, lifestyle behaviors, and food-choice motivations of self-defined vegetarians differ from those of non-vegetarians, and that there may be variation between men and women which has not previously been examined in population-based studies.
Several aspects of our results warrant additional consideration, one of which is the small proportion of self-identified vegetarians who adhered rigidly to diets free from animal flesh. Occasional use of seafood, poultry, or meat by a majority of those who consider themselves to be vegetarian has also been reported in other studies [9,15]. If a strict definition of vegetarianism had been used, the prevalence in our study would be less than 1.5% rather than close to 6%. Despite basing our analysis on respondents' self-definition, we still observed a number of differences in nutrient intake and lifestyle behavior. At some level, this validates respondents' self-identification as vegetarian.
Evidence for a higher level of 'health consciousness' among vegetarians in our sample was provided by findings of increased use of nutrient supplements, higher intakes of several nutrients (fiber, magnesium, potassium), higher intakes of fruits and vegetables, a considerably lower prevalence of smoking, and among women, higher physical activity and a lower BMI. Many of these findings have been reported in other studies, although most reports from convenience samples have not found differences in smoking or exercise behavior by vegetarian status [7,16-21]. It is likely that convenience sampling resulted in recruitment of more 'health conscious' participants and therefore did not detect differences. Thus our findings provide population-level support for the concept that vegetarians have healthier lifestyle practices than the general population of non-vegetarians.
Vegetarians were also more likely to consider 'maintaining/improving health' when choosing/avoiding foods, to choose foods for the nutrients they contain and to avoid foods for their fat content. These findings provide additional evidence of health consciousness, and are consistent with research reporting that health concerns and benefits are a primary reason for adopting a vegetarian lifestyle [22,23], although we did not assess motivation for adopting a vegetarian diet. They are also consistent with a population-based study in the Netherlands that found vegetarians were more likely to report health considerations when purchasing food [10]. That study, however, did not report nutrient intakes.
A novel aspect of our analysis was that, in addition to assessing differences in nutrient intakes, we also compared the prevalence of inadequate nutrient intakes using the EAR cut-point method [14]. As assessed by the proportions with total usual nutrient intakes below the EAR, vegetarians were less likely to have an inadequate intake of magnesium, and female vegetarians were also less likely to have inadequate intakes of folate, vitamin C, thiamin and vitamin B6. Although there were no differences by vegetarian status in the proportions with zinc intakes below the EAR, this may not be an accurate reflection of zinc adequacy, as the requirement for dietary zinc may be as much as 50% greater for vegetarians [24]. Similarly, iron requirements of vegetarians are estimated to be 80% greater than those of non-vegetarians [24]. However, the adequacy of iron intakes was not assessed in our study because the iron requirement distribution is skewed, and therefore the EAR cut-point method cannot be used to estimate the prevalence of inadequacy [14]. Finally, although adequacy of vitamin B12 intakes is often identified as a concern for vegetarians, in our sample the prevalence of inadequate intakes was similar by vegetarian status. This is likely due to the fact that almost all vegetarians used dairy products and eggs, as well as to the high prevalence of B vitamin supplementation among vegetarians.
Although our vegetarian sample was small, our results provide suggestive evidence of gender differences. For example, vegetarian women had a lower age-adjusted BMI and waist circumference, and a lower prevalence of overweight/obesity, while no differences were seen between vegetarian and non-vegetarian men. This may have been due to the higher frequency of physical activity reported by vegetarian women (but not men), as energy intake did not differ by vegetarian status for either sex. Reports from convenience samples often suggest that vegetarians have lower BMI and/or a lower rate of obesity [2,7,22,25-27]. Conversely, other convenience samples, in which energy intakes and physical activity were similar between vegetarians and non-vegetarians, did not detect differences in BMI between groups [16-19,28,29]. In the population-based CSFII, self-identified vegetarians had lower energy intakes and age-adjusted BMI [9]. However, a major limitation of that report was that analyses were not conducted by gender. Accordingly, if vegetarians were more likely to be female, as observed in our sample and another population-based sample [10], vegetarians' mean energy intake and BMI would appear to be lower because of women's lower mean energy intakes and BMI.
The distribution of macronutrient intakes also provided suggestive evidence of gender differences. Carbohydrate as a percentage of energy was higher among vegetarians, as was also found in the CSFII vegetarian analysis [9] and the majority of convenience sample studies [18,22,26,27,30,31]. Other studies have also reported lower percentage energy from fat [8,9,22,27,32] and protein [8,18,22,27,28,30-32]. In our sample, only male vegetarians had a lower proportion of energy from protein and only female vegetarians consumed less energy from fat.
We also observed gender differences in motivations for choosing/avoiding foods. Only male vegetarians were more likely to report considering heart disease and high blood pressure when choosing/avoiding foods and to report avoiding foods because of their cholesterol or saturated fat content. This is consistent with the higher prevalence of heart disease among the male vegetarians in our sample, who we speculate may have chosen to follow a vegetarian diet as a result of heart disease. Because we did not assess motivation for adopting a vegetarian diet, this cannot be ascertained, and in any case, the study's cross-sectional design precludes causal inferences. Female vegetarians, on the other hand, were not more concerned about heart disease, but were more likely to consider cancer, osteoporosis and food allergies/intolerances when choosing/avoiding foods and to avoid foods because of their salt content. They were also less likely to consider weight gain when choosing/avoiding foods. It has been suggested that some young women may adopt a vegetarian lifestyle in an effort to lose weight [33,34]; however, this does not appear to be true for our population-based sample.
While our findings suggest that variation by gender may exist in vegetarians' dietary habits and lifestyle behaviors, the study limitations should be acknowledged. First, although the sample was considered representative of the province of British Columbia, it was not nationally representative, which means that inferences cannot be made about the Canadian population. Also, the response rate, although typical of other studies of this kind, was not optimal. Second, the absolute number of self-identified vegetarians was small and therefore caution must be used when interpreting the apparent gender differences. We had limited power to detect gender-by-vegetarian status interactions. Finally, data on dietary intake and lifestyle behaviors were based on self-reports, and it is known that dietary intakes are underreported [35]. This would be problematic if differences existed in the extent of underreporting by vegetarian status. However, based on similar reported energy intakes of the two groups, it appears unlikely that differential underreporting occurred.
We do not believe that our observations of higher 'health consciousness' among vegetarians were confounded by other differences between vegetarian and non-vegetarian groups. First, although the prevalence of vegetarianism was higher among women than men, we conducted analyses separately by gender. Second, vegetarians tended to be younger than non-vegetarians, so age was included as a covariate in nutrient intake and anthropometric analyses. Third, although vegetarians were more likely to be single and to report low-income status, consideration of these differences did not affect our observations.
Conclusion
Taken together, these population-based findings add further support to the concept that adult vegetarians are more health-conscious than non-vegetarians, and that this difference extends to food choice and nutrition concerns. Additional population-based studies comparing dietary habits and lifestyle behaviors by vegetarian status and gender are needed to determine if gender differences observed in our representative sample exist in other populations in the developed world.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JB performed the statistical analyses and drafted the manuscript. SB conceived of the study and participated in its design, and helped to draft the manuscript. Both authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Dietary intake (age-adjusted, mean ± SE) by vegetarian status and gender A
Click here for file
Acknowledgements
The BC Nutrition Survey was supported by the BC Ministry of Health Services and Health Canada.
==== Refs
Appleby PN Thorogood M Mann JI Key TJ The Oxford vegetarian study: an overview Am J Clin Nutr 1999 70 525S 531S 10479226
Brathwaite N Fraser HS Modeste N Broome H King R Obesity, diabetes, hypertension, and vegetarian status among Seventh-Day Adventists in Barbados: preliminary results Ethn Dis 2003 13 34 39 12723010
Segasothy M Phillips PA Vegetarian diet: panacea for modern lifestyle diseases? QJM 1999 92 531 544 10627874 10.1093/qjmed/92.9.531
Key TJ Davey GK Appleby PN Health benefits of a vegetarian diet Proc Nutr Soc 1999 58 271 275 10466166
Key TJ Fraser GE Thorogood M Appleby PN Beral V Reeves G Burr ML Chang-Claude J Frentzel-Beyme R Kuzma JW Mann J McPherson K Mortality in vegetarians and nonvegetarians: detailed findings from a collaborative analysis of 5 prospective studies Am J Clin Nutr 1999 70 516S 524S 10479225
Key TJ Appleby PN Davey GK Allen NE Spencer EA Travis RC Mortality in British vegetarians: review and preliminary results from EPIC-Oxford Am J Clin Nutr 2003 78 533S 538S 12936946
Spencer EA Appleby PN Davey GK Key TJ Diet and body mass index in 38000 EPIC-Oxford meat-eaters, fish-eaters, vegetarians and vegans Int J Obes Relat Metab Disord 2003 2 728 734 12833118 10.1038/sj.ijo.0802300
Kennedy ET Bowman SA Spence JT Freedman M King J Popular diets: correlation to health, nutrition, and obesity J Am Diet Assoc 2001 101 411 420 11320946 10.1016/S0002-8223(01)00108-0
Haddad EH Tanzman JS What do vegetarians in the United States eat? Am J Clin Nutr 2003 78 626S 632S 12936957
Hoek AC Luning PA Stafleu A de Graaf C Food-related lifestyle and health attitudes of Dutch vegetarians, non-vegetarian consumers of meat substitutes, and meat consumers Appetite 2004 42 265 272 15183917 10.1016/j.appet.2003.12.003
Forster-Coull L Barr SI Levy-Milne R British Columbia Nutrition Survey. Report on Energy and Nutrient Intakes 2004 Health Canada, British Columbia Ministry of Health Planning, & University of British Columbia
Godin G Shephard RJ A simple method to assess exercise behavior in the community Can J Appl Sport Sci 1985 10 141 146 4053261
Marcus BH Simkin LR The stages of exercise behavior J Sports Med Phys Fitness 1993 33 83 88 8350613
Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes: Applications in Dietary Assessment 2000 Washington DC: National Academy Press
Barr SI Chapman GE Perceptions and practices of self-defined current vegetarian, former vegetarian and nonvegetarian women J Am Diet Assoc 2002 102 354 360 11902368 10.1016/S0002-8223(02)90083-0
Ball MJ Bartlett MA Dietary intake and iron status of Australian vegetarian women Am J Clin Nutr 1999 70 353 358 10479197
Ball MJ Ackland ML Zinc intake and status in Australian vegetarians Br J Nutr 2000 83 27 33 10703461
Barr SI Broughton TM Relative weight, weight loss efforts and nutrient intakes among health-conscious vegetarian, past vegetarian and nonvegetarian women ages 18 to 50 J Am Coll Nutr 2000 19 781 788 11194532
Outila TA Karkkainen MU Seppanen RH Lamberg-Allardt CJ Dietary intake of vitamin D in premenopausal, healthy vegans was insufficient to maintain concentrations of serum 25-hydroxyvitamin D and intact parathyroid hormone within normal ranges during the winter in Finland J Am Diet Assoc 2000 100 434 441 10767899 10.1016/S0002-8223(00)00134-6
Davey GK Spencer EA Appleby PN Allen NE Knox KH Key TJ EPIC-Oxford: lifestyle characteristics and nutrient intakes in a cohort of 33 883 meat-eaters and 31 546 non meat-eaters in the UK Public Health Nutr 2003 6 259 269 12740075 10.1079/PHN2002430
Locong A Nutritional status and dietary intake of a selected sample of young adult vegetarians J Can Diet Assoc 1986 47 101 106
Huang Y Lin W Cheng C Su K Nutrient intakes and iron status of healthy young vegetarians and nonvegetarians Nutr Res 1999 19 663 674 10.1016/S0271-5317(99)00031-7
Kim EHJ Schroeder KM Houser RF Dwyer JT Two small surveys, 25 years apart, investigating motivations of dietary choice in 2 groups of vegetarians in the Boston area J Am Diet Assoc 1999 99 598 601 10333784 10.1016/S0002-8223(99)00147-9
Institute of Medicine, Food and Nutrition Board Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc 2000 Washington DC: National Academy Press
Appleby PN Thorogood M Mann JI Key TJ Low body mass index in non-meat eaters: the possible roles of animal fat, dietary fibre and alcohol Int J Obes Relat Metab Disord 1998 22 454 460 9622343 10.1038/sj.ijo.0800607
Janelle KC Barr SI Nutrient intakes and eating behavior scores of vegetarian and nonvegetarian women J Am Diet Assoc 1995 95 180 186 7852684 10.1016/S0002-8223(95)00045-3
Millet P Guilland JC Fuchs F Klepping J Nutrient intake and vitamin status of healthy French vegetarians and nonvegetarians Am J Clin Nutr 1989 50 718 727 2801576
Harman SK Parnell WR The nutritional health of New Zealand vegetarian and non-vegetarian Seventh-day Adventists: selected vitamin, mineral and lipid levels N Z Med J 1998 111 91 94 9577459
Hunt IF Murphy NJ Henderson C Food and nutrient intake of Seventh-day Adventist women Am J Clin Nutr 1988 48 850 851 3414593
Faber M Gouws E Benade AJ Labadarios D Anthropometric measurements, dietary intake and biochemical data of South African lacto-ovovegetarians S Afr Med J 1986 69 733 738 3715645
Lee HY Woo J Chen ZY Leung SF Peng XH Serum fatty acid, lipid profile and dietary intake of Hong Kong Chinese omnivores and vegetarians Eur J Clin Nutr 2000 54 768 773 11083485 10.1038/sj.ejcn.1601089
Wilson AK Ball MJ Nutrient intake and iron status of Australian male vegetarians Eur J Clin Nutr 1999 53 189 194 10201799 10.1038/sj.ejcn.1600696
Gilbody SM Kirk SF Hill AJ Vegetarianism in young women: another means of weight control? Int J Eat Disord 1999 26 87 90 10349588 10.1002/(SICI)1098-108X(199907)26:1<87::AID-EAT11>3.0.CO;2-M
McLean JA Barr SI Cognitive dietary restraint is associated with eating behaviors, lifestyle practices, personality characteristics and menstrual irregularity in college women Appetite 2003 40 185 92 12781168 10.1016/S0195-6663(02)00125-3
Subar AF Kipnis V Troiano RP Midthune D Schoeller DA Bingham S Sharbaugh CO Trabulsi J Runswick S Ballard-Barbash R Sunshine J Schatzkin A Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: the OPEN study Am J Epidemiol 2003 158 1 13 12835280 10.1093/aje/kwg092
| 15829014 | PMC1090609 | CC BY | 2021-01-04 16:38:04 | no | Int J Behav Nutr Phys Act. 2005 Apr 13; 2:4 | utf-8 | Int J Behav Nutr Phys Act | 2,005 | 10.1186/1479-5868-2-4 | oa_comm |
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-101584769810.1186/1476-072X-4-10ReviewGIDEON: a comprehensive Web-based resource for geographic medicine Berger Stephen A [email protected] Department of Geographic Medicine, Tel Aviv Medical Center, 6 Weitzman Street, Tel Aviv 64239, Israel2005 22 4 2005 4 10 10 16 4 2005 22 4 2005 Copyright © 2005 Berger; licensee BioMed Central Ltd.2005Berger; 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.
GIDEON (Global Infectious Diseases and Epidemiology Network) is a web-based computer program designed for decision support and informatics in the field of Geographic Medicine. The first of four interactive modules generates a ranked differential diagnosis based on patient signs, symptoms, exposure history and country of disease acquisition. Additional options include syndromic disease surveillance capability and simulation of bioterrorism scenarios. The second module accesses detailed and current information regarding the status of 338 individual diseases in each of 220 countries. Over 50,000 disease images, maps and user-designed graphs may be downloaded for use in teaching and preparation of written materials. The third module is a comprehensive source on the use of 328 anti-infective drugs and vaccines, including a listing of over 9,500 international trade names. The fourth module can be used to characterize or identify any bacterium or yeast, based on laboratory phenotype. GIDEON is an up-to-date and comprehensive resource for Geographic Medicine.
==== Body
Introduction
As of 2005, the world is confronted by 338 generic infectious diseases, scattered in a complex fashion across over 220 countries and regions. Each new day confronts health care workers with unexpected outbreaks, epidemics and heretofore unknown pathogens. Over 2,000 named bacteria, viruses, fungi and parasites are known to cause human disease; and are confronted by 328 anti-infective agents and vaccines. Experts working in Health Geographics share an obvious and immediate need for comprehensive and timely data on the status of infection around the globe. A recent outline of GIDEON addressed uses for the Infectious Diseases clinician [1]. This review will focus on the Global Health aspect of the program.
In 1990, we initiated a project to design computer systems to follow all diseases, outbreaks, pathogens and drugs. The initial DOS-based program was written in Paradox for floppy disks, later evolving through a compact disk-based program in Windows. A commercial web-based program was eventually released under the name, GIDEON (Global Infectious Diseases and Epidemiology ON-line, Gideon Informatics, Inc, Los Angeles, California) at . The current version is available on CD (updated every three months) or web subscription (updated every week).
The program consists of four modules: Diagnosis, Epidemiology, Therapy and Microbiology. Program modules of peripheral interest in Health Geographics (Therapy and Microbiology) will be discussed only briefly.
Diagnosis Module
The Diagnosis module is designed to generate a ranked differential diagnosis based on signs, symptoms, laboratory tests, incubation period, nature of exposure and country of disease origin. Figure 1 depicts the data entry screen for a patient suffering from fever and joint pain following a trip to Indonesia. The lower 'Personal notes' box is used to record additional case data, and can be written in the user's own language. The differential diagnosis list for this case (figure 2) indicates that this patient may be suffering from Chikungunya. The appearance of many diseases on the list indicates that the user failed to enter all positive, and negative findings. For example, the fact that cough was absent would have reduced the likelihood of the second disease listed (Mycoplasma infection) and increased the statistical probability of Chikungunya.
Figure 1 GIDEON Diagnosis module. Clinical data entered for a patient suffering from joint pain and fever following a trip to Indonesia.
Figure 2 Differential diagnosis for case in Figure 1. "Why not" and 'Compare' options are available at upper left.
At this point, the user can generate a hard copy or e-mail report, access a table comparing the clinical features of the diseases listed, or examine the ranking or omission of specific diseases. If the user clicks on a specific disease name, clinical and epidemiological data on the disease in question are depicted (figure 3). The differential diagnosis list is generated by a Bayesian formula which compares the product of disease-incidence and symptom incidence, for all compatible infectious diseases. In the above example, a number of diseases known to occur in Indonesia were capable of producing fever, and joint pain. The statistical likelihood of Chikungunya in this case can be computed by a simple Bayesian formula, as follows:
Figure 3 Epidemiological background on the status of Chikungunya in Indonesia. Additional options access the descriptive epidemiology ('General' tab), clinical features, clinical images, etc for this disease.
C = Chikungunya, P = probability or incidence, S = observed symptoms
P-(C/S) = probability of Chikungunya, given these symptoms
D2, D3, Dn = other diseases compabible with this clinical scenario
Two spreadsheets in the GIDEON database respectively follow the incidence of all symptoms for every disease, and the incidence of all diseases for every country. When a clinical case is "entered" into GIDEON, the program identifies all compatible diseases and ranks their relative likelihoods as determined by the above formula, ie: P-(C/S) vs. P-(D2/S) vs. P-(D3/S) ... vs. P-(Dn/S).
A blinded study of 500 cases conducted by this author found that the correct diagnosis was listed in the differential list in 94.7% of cases, and was ranked first in 75% [2]. A second study of hospitalized patients in Boston found that the correct diagnosis was listed in only 69%, and was ranked first in 60% [3]. It is likely that inclusion in the differential diagnosis list may be more important than disease ranking in such systems [4].
A "Bioterrorism" option generates the differential diagnosis for diseases associated with suspected bioterror scenarios. In Figure 4, "<bioterrorism simulator>" has been substituted for Indonesia, given the above constellation of fever, joint pain, etc. The resulting differential diagnosis lists Ebola (42.9% probability), followed by Crimean-Congo hemorrhagic fever (12.6% probability). A similar "Worldwide" option can be used to explore all of the worlds diseases consistent with given clinical features, and access text on the global status for individual diseases.
Figure 4 Data entry screen for a bioterrorism scenario.
In theory, data entry by users can be monitored at the server level for purposes of surveillance. For example, if one or more users in China were to enter cases of fatal pneumonia, a "red-flag" at any monitoring agency (i.e., the World Health Organization) could indicate the possible appearance of SARS – long before submission of specimens or reporting of the case to local authorities. Similarly, the appearance of multiple cases of "dysentery" by users in a given community could indicate a possible outbreak of shigellosis.
Epidemiology module
The Epidemiology module presents detailed country-specific information on the status of each disease, both globally and within each relevant country. The current version contains over two million words in 12,000 notes. All data are derived from Health Ministry publications, peer-review journals, standard textbooks, WHO and CDC websites and data presented at conferences. The user may also access over 30,000 graphs which follow disease incidence, rates and other numerical data. The main Epidemiology screen is shown in Figure 5. Note that the user can append custom "personal notes" – in any national language or font- regarding the status of every disease in their own institution. Such notes would be accessible by all colleagues using GIDEON on the local network.
Figure 5 Epidemiology module, main screen. The 'images' tab has been pressed, to access thumbnail images of Plague. These can be maximized and copied to PowerPoint, etc. Note addition of 'Personal notes' by the user, at lower right.
Maps which depict the global distribution of each disease can be accessed through the 'Distribution' tab (Figure 6). Text outlining country-specific data for the disease (Figure 7) is available through either a list of countries displayed in this module, or by clicking the relevant 'red dot' on the map. These text boxes also include data sets which automatically generate incidence / rate graphs (Figure 8), a chronological account of all disease outbreaks, and numbered reference links to relevant journal publications and reports of ongoing outbreaks from ProMed . A separate 'Graphs' option allows the user to generate custom-made graphs comparing multiple disease rates, or rates in multiple countries. (Figure 9).
Figure 6 Epidemiology module. Map depicting the global distribution of plague. Specific map areas can be expanded, and all elements can be copied for reproduction as necessary. Country-specific notes regarding plague appear when corresponding red dots are clicked.
Figure 7 Plague in Tanzania. Clicking on relevant data sets will generate incidence and rates graphs. Note several numbered links to journal publications.
Figure 8 Plague – Worldwide incidence and rates per 100,000.
Figure 9 Graph contrasting AIDS rates among user-selected countries.
Additional tabs access the descriptive epidemiology and clinical background of each disease. Synonym tabs generate lists of alternative terms for diseases and countries in Spanish, German, Norwegian, etc. Historical data record the incidence of individual diseases and significant outbreaks spanning decades. An additional "Fingerprint" option generates a list of diseases compatible with any set of epidemiological parameters. For example, in Figure 10 we see that ten parasitic diseases are transmitted by fish in Japan.
Figure 10 Epidemiology "Fingerprint" query (upper left screen). How many different parasitic diseases can be acquired by eating fish in Japan?
Therapy
The Therapy module follows the pharmacology and application of all drugs and vaccines used in Infectious Diseases. The current version contains 264 generic drugs and 64 vaccines. Various sub-modules present the mechanism of action; pharmacology, dosages, drug-drug interactions, contraindications, spectrum, and susceptibility testing standards. An international synonym lists contains over 9,500 trade names. As in other modules, users may add custom notes in their own language for each drug or vaccine: prices, resistance patterns, local trade names, etc.
Microbiology
The Microbiology option is similar to the Diagnosis module. Users may enter any combination of phenotypic tests, and obtain a ranked probability list of compatible bacteria. The current version incorporates more than 1,300 taxa. The Microbiology module is also designed to list the phenotype, prior names, ecology and disease association for any organism, or compare the phenotypes of any combination of organisms selected by the user.
Assessment
Since the graphic and mapping functions of GIDEON treat individual countries as whole units, data presentations lack a certain degree of "granularity." Thus, the differential diagnosis of fever in Venezuela will include malaria, even if the patient is living outside of the endemic, southern region. This problem is corrected to a large extent by text in the associated country-specific notes and the general knowledge base of the treating physician. In theory, the manufacturer could follow the incidence of each disease for every state, district, province and oblast; but variability would still exist according to occupation, rural vs. urban setting, season, etc.
An additional problem relates to the availability and quality of valid epidemiological data. Disease reporting varies widely from country to country. For example, AIDS reporting statistics from sub-Saharan Africa are generally inadequate. Where necessary, the spreadsheets used by GIDEON record published true estimates rather than questionable reports. In other instances, Health Ministry data conflict with reports of the World Health Organisation, a fact which is recorded in relevant GIDEON country notes. Occasionally, major diseases are not reported at all. For example, several recent cases of cholera in Japan originated from Thailand; but Thailand has not officially reported a single case in many years. Where possible, the GIDEON data base relies on published best estimates, and at times 'educated guesses' when data are entirely lacking. Thus, there are few published data for disease incidence in Togo, and the program is forced to rely on publications for neighboring Ghana.
The reader is referred to the GIDEON website for an extensive listing of data sources, published reviews, technical background and pricing information.
Competing interests
The author serves as a salaried Scientific Advisor to Gideon Informatics, Inc.
==== Refs
Edberg SC Global infectious diseases and epidemiology network (GIDEON): A Web-based Program for Diagnosis and Informatics in Infectious Diseases Clin Infect Dis 2005 40 123 126 15614701 10.1086/426549
Berger S Blackman U Computer program for diagnosing and teaching geographic medicine J Travel Med 1995 2 199 203 9815388
Ross JJ Shapiro DS Evaluation of the computer program GIDEON for the diagnosis of fever in patients admitted to a medical service Clin Infect Dis 1998 26 766 767 9524864
Kassirer JP A report card on computer-assisted diagnosis – the grade: C N Engl J Med 1994 330 1824 5 8190163 10.1056/NEJM199406233302512
| 15847698 | PMC1090610 | CC BY | 2021-01-04 16:39:06 | no | Int J Health Geogr. 2005 Apr 22; 4:10 | utf-8 | Int J Health Geogr | 2,005 | 10.1186/1476-072X-4-10 | oa_comm |
==== Front
J Negat Results BiomedJournal of Negative Results in Biomedicine1477-5751BioMed Central London 1477-5751-4-41581997510.1186/1477-5751-4-4ResearchDoes socioeconomic status affect mortality subsequent to hospital admission for community acquired pneumonia among older persons? Vrbova Linda [email protected] Muhammad [email protected] Rahim [email protected] Liisa [email protected] Ross EG [email protected] Department of Public Health Sciences, University of Toronto, McMurrich Building, 12 Queen's Park Crescent W, Toronto, ON, M5S 1A8, Canada2 Institute for Clinical Evaluative Sciences, 2075 Bayview Avenue, Toronto, ON, M4N 3M5, Canada3 Health Policy Management and Evaluation, University of Toronto, McMurrich Building, 2nd Floor, 12 Queen's Park Crescent West, Toronto, ON, Ma5S 1A8, Canada4 Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, ON, M5S 2S2, Canada5 Department of Family and Community Medicine, University of Toronto, 256 McCaul Street, 2nd Floor, Toronto, ON, M5T 2W5, Canada6 Primary Care Research Unit, Department of Family and Community Medicine, Sunnybrook and Women's College Health Sciences Centre, 2075 Bayview Avenue, Toronto, ON, M4N 3M5, Canada2005 8 4 2005 4 4 4 11 8 2004 8 4 2005 Copyright © 2005 Vrbova et al; licensee BioMed Central Ltd.2005Vrbova 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
Low socioeconomic status has been associated with increased morbidity and mortality for various health conditions. The purpose of this study was twofold: to examine the mortality experience of older persons admitted to hospital with community acquired pneumonia and to test the hypothesis of whether an association exists between socioeconomic status and mortality subsequent to hospital admission for community-acquired pneumonia.
Methods
A population based retrospective cohort study was conducted including all older persons patients admitted to Ontario hospitals with community acquired pneumonia between April 1995 and March 2001. The main outcome measures were 30 day and 1 year mortality subsequent to hospital admission for community-acquired pneumonia.
Results
Socioeconomic status for each patient was imputed from median neighbourhood income. Multivariate analyses were undertaken to adjust for age, sex, co-morbid illness, hospital and physician characteristics. The study sample consisted of 60,457 people. Increasing age, male gender and high co-morbidity increased the risk for mortality at 30 days and one year. Female gender and having a family physician as attending physician reduced mortality risk.
The adjusted odds of death after 30-days for the quintiles compared to the lowest income quintile (quintile 1) were 1.02 (95% CI: 0.95–1.09) for quintile 2, 1.04 (95% CI: 0.97–1.12) for quintile 3, 1.01 (95% CI: 0.94–1.08) for quintile 4 and 1.03 (95% CI: 0.96–1.12) for the highest income quintile (quintile 5). For 1 year mortality, compared to the lowest income quintile the adjusted odds ratios were 1.01 (95% CI: 0.96–1.06) for quintile 2, 0.99 (95% CI: 0.94–1.04) for quintile 3, 0.99 (95% CI: 0.93–1.05) for quintile 4 and 1.03 (95% CI: 0.97–1.10) for the highest income quintile.
Conclusion
Socioeconomic status is not associated with mortality in the older persons from community-acquired pneumonia in Ontario, Canada.
==== Body
Introduction
Community-acquired pneumonia (CAP) is a substantial cause of mortality, morbidity, and health services utilization in the older persons [1]. In Canada pneumonia and influenza are, together, the leading cause of death from infectious disease and sixth leading cause of death overall. In Canada, the annual hospitalization for pneumonia and influenza is 1,358 per 100,000, and in Ontario 1,283 per 100,000 [2]. The high morbidity and mortality associated with CAP makes understanding its epidemiology a research priority.
Current health research increasingly recognizes the existence and contribution of broader determinants of health in explaining differences in health status and health outcomes among populations. Socioeconomic status is an important influence on morbidity and mortality [3,4]. Access to large databases in Canada has allowed for the examination of the relationship of socioeconomic status to specific health outcomes. Recently, Canadian studies have revealed that those with lower socioeconomic status experience higher mortality and morbidity after myocardial infarction [5] and stroke [6]. Such mortality gradients are problematic in a publicly funded health care system, indicating potential problems with unequal access to care. There are few published studies investigating the relation between socioeconomic factors and pneumonia. Previous studies of the association between SES and pneumonia have looked at different endpoints of pneumonia, and used various SES measures, yielding conflicting results.
There is no consensus as to which factors contribute the most to increasing mortality risk from pneumonia, notably whether it is age or co-morbidity that is the deciding variable [7-9]. No studies to date have examined the independent effects of age, gender, co-morbidity and SES on mortality after CAP. This study examines the mortality experience hypothesis of whether there is an association between socioeconomic status and mortality after community-acquired pneumonia in older persons in Ontario, Canada, controlling for age, gender, co-morbidity and other factors.
Methods
Study Design and Data Sources
A cohort of patients diagnosed with pneumonia in Ontario hospitals were assembled for 6 years, from April 1, 1995 to March 31, 2001. The inclusion criteria of the cohort were: "most responsible" diagnosis of pneumonia and influenza (codes 480–487 of the International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM][10], age greater than 65 and less than 105 and resident of Ontario. Unpublished data indicates that influenza codes (487) are infrequently used and account for less than .05% of the sample. The most frequently used codes are 485 and 486 which are for pneumonia with no specific isolated causative organism
In order to rule out readmissions, patients who were admitted for pneumonia in the previous 12 months were excluded. Furthermore, in order to focus solely on community-acquired pneumonia, patients transferred from another health-care institution or long-term facility were also excluded.
Hospital discharge abstracts were drawn from the Canadian Institute of Health Information (CIHI) database. The abstracts contained information pertaining to the index admission, age and gender, physician and hospital characteristics, demographic characteristics and co-morbid illnesses of patients, as well as in-hospital mortality. The Ontario Registered Persons Database provided the 30-day and 1 year mortality, both in and out of hospital. Algorithms used to link data across databases have proven reliability and validity.
Administrative databases used do not contain personal income data of the individual patients. They do, however, include the Forward Sortation Area (FSA) (the first three digits of the postal code), which was used to impute the patient' s median neighborhood income from the 1996 Canadian Census. Of the 504 FSAs in Ontario, the median neighborhood income for 11 was suppressed by Statistics Canada due to small sample size.
Statistical Analysis
Median neighborhood income was broken down into quintiles for analysis. Baseline data across socioeconomic quintiles were compared using the Cochrane-Mantel-Haenszel chi-square for the categorical data, and weighted linear regression for continuous data. Kaplan-Meier survival curves were created to illustrate 30-day and 1 year mortality of the cohort by income quintile. Cox proportional hazards and logistic regression was used to determine the relation of median neighbourhood income to 30-day and 1-year mortality, adjusting for potentially confounding variables known or suspected to influence mortality risk: age, gender, co-morbidity (Charlson index score ≥ 1), specialty of attending physician and hospital status (teaching or non-teaching).
All statistics was done using SAS software (version 8.2), survival curve graphs were done using Microsoft Excel (version 9.0.0.3822).
Results
Baseline Data
The cohort consisted of 61,086 people, of whom 60,457 could be assigned a SES quintile (99% of the cohort), and hence could be used in the analyses. Table 1 shows the baseline characteristics of the population. There was a significant difference (p < 0.0001) in co-morbidity among the classes. The higher social classes had the higher co-morbidity (Charlson score >1) of cancer and ischemic heart disease, but there was no difference in the prevalence of chronic lung disease, chronic renal failure or congestive heart failure. The higher social classes were admitted more often to teaching hospitals,(as compared to community hospitals) and were attended by specialists (internal medicine, respirology) more frequently than general practitioners (see Table 1). Table 2 indicates no difference was found in length of stay across the income quintiles. Persons from the higher income quintiles were more likely to be discharged home.
Table 1 Baseline Characteristics of Pneumonia Patients According to Neighbourhood Median Income
Characteristics Income Quintile P-value
1 2 3 4 5
N = 15057 N = 14655 N = 12268 N = 10310 N = 8167
Neighborhood Income ($) 0.0039
Median 16433 18572 20256 22982 26786
Interquartile Range 15408–16988 18056–19201 20043–21020 22186–24258 25899–29710
Mean Age (yr) 78.03+/-
7.65 77.94+/-
7.55 78.07+/-
7.59 78.29+/-
7.58 78.51+/-
7.72 0.0370
Female sex (%) 47.23 48.29 47.21 49.20 49.44 0.0007
Comorbid conditions (%)
Chronic Lung Disease 6.91 7.55 6.33 7.59 6.61 0.5382
Congestive Heart Failure 4.30 4.79 4.73 4.76 3.87 0.4348
Ischemic Heart Disease 11.60 12.90 14.11 14.31 13.22 <0.0001
Peripheral Vascular Disease 1.34 1.30 1.36 1.26 1.04 0.1003
Chronic Renal Failure 0.90 0.87 1.04 1.09 0.73 0.9210
Diabetes 6.97 8.02 7.44 7.38 6.09 0.0199
Cancer 2.89 3.43 3.73 3.77 4.04 <0.0001
Charlson score > 1 (%) 28.76 28.98 30.91 31.74 31.50 <0.0001
Specialty of Attending Physician (%): <0.0001
General Practice 54.78 62.27 56.26 43.96 42.07
Internal Medicine 26.69 23.82 26.18 34.24 35.14
Respirology 6.16 5.43 5.98 7.19 8.90
Other 12.37 8.47 11.57 14.62 13.89
Teaching Hospital (%) 17.63 11.31 17.12 23.66 23.58 <0.0001
Table 2 Pneumonia Treatment and Outcomes According to Quintile of Median Neighborhood Income
Outcome/Treatment N (%) Income Quintile P-value
1 2 3 4 5
N = 15057 N = 14655 N = 12268 N = 10310 N = 8167
Length of Stay 0.1448
Mean +/- SD (days) 9.41 +/-
13.0 8.95 +/-
12.5 9.54 +/-
13.9 9.59 +/-
14.0 9.96 +/-
14.0
Median (Interquartile Range) 6 (4–10) 6 (4–10) 6 (4–10) 6 (4–10) 6 (4–11)
Acute Length of Stay 0.2362
Mean +/- SD (days) 8.39 +/-
8.52 7.93 +/-
7.58 8.29 +/-
8.12 8.37 +/-
8.44 8.86 +/-
9.60
Median (Interquartile Range) 6 (4–10) 6 (4–10) 6 (4–10) 6 (4–10) 6 (4–11)
Discharge Destination <0.0001
Acute Care Hospital 433 (2.88) 284 (1.94) 224 (1.83) 128 (1.24) 99 (1.21)
Chronic Care Hospital 263 (1.75) 365 (2.49) 285 (2.32) 211 (2.05) 140 (1.71)
Rehabilitation Hospital 57 (0.38) 106 (0.72) 92 (0.75) 102 (0.99) 93 (1.14)
Nursing Home 361 (2.40) 289 (1.97) 229 (1.87) 187 (1.81) 174 (2.13)
Home Care Program 2101 (13.95) 1890 (12.90) 1754 (14.30) 1561 (15.14) 1123 (13.75)
Home 11648 (77.36) 11484 (78.36) 9511 (77.53) 7998 (77.58) 6455 (79.04)
Mortality
The Kaplan-Meier survival curves for pneumonia mortality resulted in similar findings for both the 30-day mortality and the 1-year mortality (see Fig 1 and 2).
Figure 1 30 day mortality after initial admission for pneumonia, survival curves for social class quintiles from lowest (I) to highest (V)
Figure 2 1 year mortality after initial admission for pneumonia, survival curves for social class quintiles from lowest (I) to highest (V).
Multivariate modelling with logistic regression resulted in no significant difference in mortality (both 30-day and 1 year) across the income quintiles after adjustment for age, gender, co-morbidity (Charlson index score ≥ 1), specialty of attending physician and hospital teaching status (see Table 3 Cox models were completed, but the assumptions of the model violated. Odds Ratios were similar to the logistic models). The odds of death after 30-days for the quintiles compared to the lowest income quintile (quintile 1) were 1.02 (95% CI: 0.95–1.09) for quintile 2, 1.04 (95% CI: 0.97–1.12) for quintile 3, 1.01 (95% CI: 0.94–1.08) for quintile 4 and 1.03 (95% CI: 0.96–1.12) for the highest income quintile (quintile 5). The results were very similar for 1 year mortality, where, compared to the lowest income quintile the odds ratios were 1.01 (95% CI: 0.96–1.06) for quintile 2, 0.99 (95% CI: 0.94–1.04) for quintile 3, 0.99 (95% CI: 0.93–1.05) for quintile 4 and 1.03 (95% CI: 0.97–1.10) for the highest income quintile.
Table 3 Odds of Dying after Initial Admission for Pneumonia after 30 Days and 1 Year, Adjusted for Gender, Age, Specialty of Attending Physician, Hospital Teaching Status and Socio-economic Status
Characteristic Odds Ratio (65% CI) P-value
30-Day Mortality
Characteristics of Patient
Female Sex 0.78 (0.74–0.82) <0.0001
Charlson Index 1.81 (1.71–1.91) <0.0001
Age 65–74* 1.00
75–84 1.55 (1.46–1.65) <0.0001
85–105 3.02 (2.83–3.21) <0.0001
Income Quintile 1* 1.00
2 1.02 (0.95–1.09) 0.64
3 1.04 (0.97–1.12) 0.23
4 1.01 (0.94–1.08) 0.86
5 1.03 (0.96–1.12) 0.41
Characteristics of Hospital
Teaching Hospital (vs. non) 0.88 (0.83–0.94) 0.0003
Specialty of Attending Physician
General Practitioner 0.65 (0.60–0.70) <0.0001
Internal Medicine 0.93 (0.87–1.00) 0.073
Respirology 0.84 (0.75–0.94) 0.0016
Other* 1.00
1-Year Mortality
Characteristics of Patient
Female Sex 0.68 (0.65–0.70) <0.0001
Charlson Index 2.09 (2.01–2.18) <0.0001
Age 65–74* 1.00
75–84 1.45 (1.39–1.52) <0.0001
85–105 2.86 (2.73–3.01) <0.0001
Income Quintile 1* 1.00
2 1.01 (0.96–1.06) 0.72
3 0.99 (0.94–1.04) 0.73
4 0.99 (0.93–1.05) 0.70
5 1.03 (0.97–1.10) 0.31
Characteristics of Hospital
Teaching Hospital (vs. non) 1.01 (0.96–1.06) 0.67
Specialty of Attending Physician
General Practitioner 0.67 (0.63–0.71) <0.0001
Internal Medicine 0.82 (0.77–0.87) <0.0001
Respirology 0.78 (0.72–0.86) <0.0001
Other* 1.00
(* = reference group)
Women had lower odds of dying for both 30-day and 1-year mortality respectively (OR = 0.780, 95% CI: 0.744–0.818; OR = 0.68, 95% CI: 0.65–0.70) than men. The middle and oldest age groups (75–84, 85+) had higher odds of dying than the lowest age group (65–74) (30-day mortality: OR = 1.55, 95% CI: 1.47, 1.65; OR = 3.017, 95% CI: 2.83, 3.21; 1-year mortality: OR = 1.55, 95% CI: 1.39, 1.52; OR = 2.866, 65% CI: 2.73–3.01). The presence of another illness (Charlson co-morbidity index > = 1) significantly increased the mortality (OR = 1.81, 95% CI: 1.72, 1.91; OR = 2.094, 95% CI: 2.01–2.18). The specialty of the attending physician was also significant: compared to other types of physicians, treatment by a general practitioner had the highest protective effect on 30-day mortality (OR = 0.65, 95%CI = 0.599–0.696). Respirologist care was protective (OR = 0.83, 95% CI = 0.75, 0.94), while internal medicine practitioners did not have significant protective effects. One-year mortality was significantly affected by all three physician specialty groups studied; compared to other types of physicians, treatment by a general practitioner had the highest protective effect (OR = 0.67, 95% CI = 0.63–0.71), then respirologist (OR = 0.79, 95% CI = 0.72–0.86), then internal medicine practitioners (OR = 0.82, 95% CI: 0.77–0.87). Hospital teaching status was significant for 30-day mortality (OR = 0.89, 95% CI: 0.83–0.95) but not for 1-year mortality.
Discussion
There is increasing evidence from diverse observational studies that low SES is associated with adverse health outcomes [11]. The findings of this study do not support the existence of an association between socioeconomic status and mortality subsequent to CAP. The study results indicate that community acquired pneumonia is a condition associated with high mortality, and that gender, age and co-morbidity most significantly influence outcome. The study results are similar to a recent study by Kaplan et al. reporting a 33.6% mortality rate in survivors of CAP [12]. The results underscore the high prevalence, resource intensity and mortality associated with CAP, particularly in older persons [1].
The strengths of this study are its population base, large sample size, accurate linkages and detailed follow up. The study captured all pneumonia admissions for those over 65 in the province of Ontario. The pneumonia diagnostic codes (ICD-9 codes 480–487) were the same as in previous studies [13-15].
The study is limited by the use of proxy measures for socioeconomic status, namely income data from median neighborhood income. Currently there is considerable debate as to the proper measure of SES and controversy as to whether area level measures are valid for imputing SES. The measure of SES status employed in this study is identical to that used in other published studies demonstrating significant associations with mortality and access to healthcare services for myocardial infarction and stroke[5,6]. The use of area-level information applying to individuals forces consideration of the ecological fallacy. However, others have argued for the validity of using income quintiles as a proxy for socioeconomic status [16-19]. Mustard found that ecologic measures of income are highly correlated to individual income. Hence use of such proxy measures is justified when individual level data is not available [20].
There have been conflicting results concerning the relationship between socioeconomic status and pneumonia (both diagnosis and outcome). Wood [13] found an increased relative risk (RR 2.3 95% CI: 1.4–4.0) for lower social class quintiles and pneumonia and bronchitis mortality. Stelianides found that the duration of hospitalization was 5.9 days longer for low SES patients as compared to high SES patients (p < 0.003), but found no differences in mortality or ICU admission [14]. Singh and Siahpush [15] found a relative risk of 2.69 (p < 0.05) for the lowest compared to the highest income group with pneumonia and influenza mortality. Other studies, looking at pneumonia diagnosis and SES found no relation between the two [21,22]. Our study found differences in the process of care, in that higher SES patients were more likely to be treated by specialists and in academic teaching centres, but not in outcomes, as mortality and length of stay were not significantly different between SES levels. Interestingly, as a secondary outcome, those with family physicians had lower mortality than those without, suggesting that provision of primary care has a protective effect. This finding bears further exploration. As well, as in other studies [5,6], academic health sciences centres had better mortality outcomes for acute care.
The relation between SES and health is not completely understood, but theories abound. Among the explanations for the relation found between disease outcomes and socioeconomic status relates to equitable access to health services as well as more negative lifestyle and environmental exposures (higher rates of smoking, worse air quality). How can we explain that our data does not corroborate past findings or theories? One possible explanation may lie in the nature of the management of pneumonia. Myocardial infraction and stroke, where differences in outcomes and SES in this population have been reported, increasingly rely on the provision of timely and specialized technology, diagnosis and management. Management of pneumonia is, for the most part, a relatively low technology process. The majority of patients were cared for by primary care providers. Hence access to care seems relatively unproblematic in this cohort. However, pneumonia remains, as it was in Osler's day, a potent force of mortality and socioeconomic status provides no advantage or protection.
Conclusion
In this population based, retrospective cohort study of older persons in the province of Ontario, socioeconomic status was not a factor in increasing the risk of death subsequent to hospital admission for community acquired pneumonia. Male gender, age and co-morbid illness significantly increase both 30 day and one year mortality. Female gender is associated with significantly reduced risk. Having a primary care provider and being cared for in an academic health sciences centre also reduced the mortality risk.
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
RU initiated the idea for the study. LJ and MM co-wrote the grant with RU. LV wrote the first draft of the article and carried out the statistical analysis. RM provided intellectual input to the study design and analysis. All contributed intellectual input into the study. All participated in the revision of drafts and approve of the final draft.
Acknowledgements
This study was funded by an Integrated Health Research Term Grant entitled Respiratory Infections in Older Adults, from the Canadian Institutes of Health Research. Ross Upshur is supported by a New Investigator Award from the Canadian Institutes of Health Research and a Research Scholar Award from the Department of Family and Community Medicine, University of Toronto. The authors would like to thank Shari Gruman for her expert assistance in the preparation of this manuscript.
==== Refs
Kaplan V Angus DC Griffin MF Clermont G Scott Watson R Linde-Zwirble WT Hospitalized community-acquired pneumonia in the older persons: age- and sex-related patterns of care and outcome in the United States Am J Respir Crit Care Med 2002 165 766 772 11897642
Health Canada Chapter 7: Infectious Diseases Respiratory Disease in Canada 2001 Ottawa, Canada
Evans RG Barer M Marmor T Why Are Some People Healthy and Others Not? The Determinants of Health in Populations 1994 New York: Aldine de Gruyter
Evans RG Stoddart GL Producing health, consuming health care Soc Sci Med 1990 31 1347 1363 2126895 10.1016/0277-9536(90)90074-3
Alter DA Naylor CD Austin P Tu JV Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction N Engl J Med 1999 341 1359 1367 10536129 10.1056/NEJM199910283411806
Kapral MK Wang H Mamdani M Tu JV Effect of socioeconomic status on treatment and mortality after stroke Stroke 2002 33 268 273 11779921 10.1161/hs0102.101169
Brancati FL Chow JW Wagener MM Vacarello SJ Yu VL Is pneumonia really the old man's friend? Two-year prognosis after community-acquired pneumonia Lancet 1993 342 30 33 8100295 10.1016/0140-6736(93)91887-R
Koivula I Sten M Makela PH Risk factors for pneumonia in the older persons Am J Med 1994 96 313 320 8166149 10.1016/0002-9343(94)90060-4
Hedlund JU Ortqvist AB Kalin ME Granath F Factors of importance for the long term prognosis after hospital treated pneumonia Thorax 1993 48 785 789 8211867
Commission on Professional and Hospital Activities International Classification of Diseases, 9th rev, clinical modification: ICD-9-CM 1992 Ann Arbor, Mich
Feinstein JS The relationship between socioeconomic status and health: a review of the literature Milbank Q 1993 71 279 322 8510603
Kaplan V Clermont G Griffin MF Kasal J Watson RS Linde-Zwirble WT Angus DC Pneumonia: Still the Old Man's Friend? Arch Intern Med 2003 163 317 323 12578512 10.1001/archinte.163.3.317
Wood E Sallar AM Schechter MT Hogg RS Social inequalities in male mortality amenable to medical intervention in British Columbia Soc Sci Med 1999 48 1751 1758 10405014 10.1016/S0277-9536(99)00081-7
Stelianides S Golmard JL Carbon C Fantin B Influence of socioeconomic status on features and outcome of community- acquired pneumonia Eur J Clin Microbiol Infect Dis 1999 18 704 708 10584896 10.1007/s100960050382
Singh GK Siahpush M All-cause and cause-specific mortality of immigrants and native born in the United States Am J Public Health 2001 91 392 399 11236403
Roos NP Mustard CA Variation in health and health care use by socioeconomic status in Winnipeg, Canada: does the system work well? Yes and no Milbank Q 1997 75 89 111 9063301 10.1111/1468-0009.00045
Smith GD Hart C Watt G Hole D Hawthorne V Individual social class, area-based deprivation, cardiovascular disease risk factors, and mortality: the Renfrew and Paisley Study J Epidemiol Community Health 1998 52 399 405 9764262
Krieger N Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology Am J Public Health 1992 82 703 710 1566949
Carr-Hill R Rice N Is enumeration district level an improvement on ward level analysis in studies of deprivation and health? J Epidemiol Community Health 1995 49 S28 29 8594129
Mustard CA Derksen S Berthelot JM Wolfson M Assessing ecologic proxies for household income: a comparison of household and neighbourhood level income measures in the study of population health status Health Place 1999 5 157 171 10670997 10.1016/S1353-8292(99)00008-8
Farr BM Woodhead MA Macfarlane JT Bartlett CL McCraken JS Wadsworth J Miller DL Risk factors for community-acquired pneumonia diagnosed by general practitioners in the community Respir Med 2000 94 422 427 10868703 10.1053/rmed.1999.0743
Farr BM Bartlett CL Wadsworth J Miller DL Risk factors for community-acquired pneumonia diagnosed upon hospital admission. British Thoracic Society Pneumonia Study Group Respir Med 2000 94 954 963 11059948 10.1053/rmed.2000.0865
| 15819975 | PMC1090611 | CC BY | 2021-01-04 16:37:37 | no | J Negat Results Biomed. 2005 Apr 8; 4:4 | utf-8 | J Negat Results Biomed | 2,005 | 10.1186/1477-5751-4-4 | oa_comm |
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-71582630010.1186/1476-511X-4-7ReviewThe role of Mediterranean diet in the epidemiology of metabolic syndrome; converting epidemiology to clinical practice Panagiotakos Demosthenes B [email protected] Evangelos [email protected] Department of Nutrition and Dietetics, Harokopio University, Athens, Greece2005 12 4 2005 4 7 7 30 3 2005 12 4 2005 Copyright © 2005 Panagiotakos and Polychronopoulos; licensee BioMed Central Ltd.2005Panagiotakos and Polychronopoulos; 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.
Metabolic syndrome is a collection of associated conditions such as dyslipidemia, high blood pressure, impaired glucose tolerance and tendency to develop fat around the abdomen. It is now well known that individuals with the metabolic syndrome are at high risk for atherosclerosis and, especially, coronary heart disease. However, it has been suggested that people with the metabolic syndrome may benefit from aggressive lifestyle modification, through diet and exercise. In this review we summarize scientific evidence regarding the effect of Mediterranean diet on the development of metabolic syndrome.
metabolic syndromedietlifestyle modificationMediterranean diet
==== Body
The metabolic syndrome is a collection of associated conditions such as dyslipidemia, high blood pressure, impaired glucose tolerance, and abdominal fat. It has first been described in 1988, and it is now widely adopted that it is a health situation that promotes atherosclerosis [1]. Each of the associated conditions has an independent effect, but clustering together they become synergistic, making the risk of developing atherosclerosis greater. Moreover, many investigators have shown a direct association between the prevalence of the syndrome with increased risk of cardiovascular disease and diabetes. Because each independent factor of the metabolic syndrome can increase the patient's cardiovascular risk, an integrated, comprehensive approach is indicated for patients with the syndrome. Treatment of metabolic syndrome is primarily based on Therapeutic Lifestyle Change, implementing weight-loss diets and exercise programmes to increase physical activity [2].
What is the metabolic syndrome?
Definitions of the metabolic syndrome are confusing. Since 1999 several investigators and Organizations have suggested different definitions. Nevertheless, all agree that the characteristics of the metabolic syndrome include atherogenic dyslipidemia, a prothrombotic state, insulin resistance, hypertension, abdominal obesity, as well as elevated microalbuminuria, increased fibrinogen, decreased plasminogen activator, elevated plasminogen activator inhibitor-1, increased blood viscosity, and increased uric acid [1]. Each abnormality promotes atherosclerosis independently, but when clustered together, these metabolic disorders are increasingly atherogenic and enhance the risk of cardiovascular morbidity and mortality. Moreover, the metabolic syndrome has been associated with the development of stroke, type-2 diabetes, diabetic nephropathy, retinopathy, and distal neuropathy. In addition, the National Cholesterol Education Program Adult Treatment Panel (NCEP ATP) III [2] presents the metabolic syndrome as an enhancer of cardiovascular risk beyond elevated low-density lipoprotein cholesterol.
Epidemiology of the Metabolic Syndrome
A substantial proportion of individuals living in Western nations are afflicted with multiple metabolic abnormalities. A recent report estimated that 115 million people in US, Japan, France, Germany, Italy, Spain and United Kingdom suffer from the metabolic syndrome [3]. Moreover, it is estimated that at least 47 million Americans have this condition, while by the year 2010 the number of US citizens that have this condition it is estimated to be between 50 to 75 million [1]. Ferrannini et al. [4] reported that more than 70% of adults have at least one of the major characteristics of the metabolic syndrome. The ATTICA Study investigators [5] recently reported that the prevalence of the metabolic syndrome was 25% in men and 15% in women from Greece. The prevalence of the metabolic syndrome in a south Mediterranean population (i.e. Greece) was similar with the prevalence of the syndrome in a sample of 8814 American men and women from the 3rd National Health and Nutrition Examination Survey (1988–1994) (i.e. 22%) [1]. In a Korean population the prevalence of metabolic syndrome was 29.0% in men and 16.7% in women aged 30–80 years [6], while in another study of the same country the prevalence of the syndrome was 13% in both en and women [7]. Hu et al. [8] studying 6156 men and 5356 women without diabetes from 11 prospective European cohorts reported that the age-standardized prevalence of the metabolic syndrome was 15.7% in men and 14.2% in women. The prevalence of metabolic syndrome in a Portuguese sample was 27.0% in women and 19.1% in men [9]. Moreover almost all these studies observed that the prevalence of the syndrome increases with age. Differences in genetic background, dietary habits, levels of physical activity, population age and sex structure, levels of over- and under-nutrition, may influence the prevalence of both the metabolic syndrome and its components worldwide. Nevertheless, all these data suggest that the prevalence of the syndrome is high due to increasing obesity and sedentary lifestyles; and reflect the growing necessity for therapeutic intervention.
Recently the NCEP ATP III suggested for Therapeutic Lifestyle Changes [2] in order to reduce the prevalence of the metabolic syndrome. These changes included the consumption of low-saturated diet (<7% of total fat) and the adoption of a physically active lifestyle. In this review we focus our interest of the effect of diet and exercise on the prevalence of the metabolic syndrome. We particularly study the effect of a traditional diet, the Mediterranean diet, on the components of the syndrome, in relation to exercise.
Diet and metabolic syndrome
Among several factors, related to lifestyle habits that could influence cardiovascular risk the beneficial effect of diet has already been underlined [10-12]. During the last decades there is increasing scientific evidence that there are protective health effects from diets, which are high in fruits, vegetables, legumes and whole grains and which include fish, nuts, and low-fat dairy products. Such diets need not to be restricted in total lipid intake fat as long as there is not an excess of energy intake over expenditure calories and emphasize predominantly vegetable oils that there are low in saturated fats and partially hydrogenated oils [13]. Foods are composed from six basic nutrients: (a) carbohydrates, (b) fat, (c) proteins, (d) vitamin, (e) minerals and (f) water. Its balance on daily dietary patterns has more or less related to the prevalence of many metabolic disorders, like hypertension, dyslipidemia, obesity and diabetes, as well as to increased risk of atherosclerotic disease.
The main purpose of carbohydrates is to provide energy for the body. Although the body can make energy from other sources, like fats and even proteins, the consumption of carbohydrates is special because it can provide energy without the use of oxygen, it is required for explosive types of work, it is the preferred energy for the brain, and the energy derived from carbohydrates is necessary to burn fats. However, carbohydrate consumption has been branded as the culprit in weight gains, obesity, diabetes, and a number of other diseases. Although carbohydrate consumption has been related to such health problems, it has been suggested that it is the excess consumption of the wrong carbohydrates that causes these health problems and not the proper carbohydrates consumed at proper amounts. Moreover, metabolic consequences of carbohydrates depend not only on their quantity but also on their quality. However, other people suggest that carbohydrates of any type may be bad for some people, particularly those with metabolic syndrome. The use of glycemic index as a measure of how fast a food raises your blood sugar is controversial [14]. Major health organizations discourage the use of the glycemic index in nutritional therapy, despite the fact that the concept has now been widely adopted by many international organizations. Although the evaluation of the glycemic index needs to be further improvements, especially within different populations, the concept of glycemic load is attractive because it captures both the quality and quantity of carbohydrates as well as potential interactions between them. Large amounts of complex carbohydrates, like such potatoes, breads, corn, etc. are recommended by several specialists. The Diet, Nutrition, and the Prevention of Chronic Diseases report of the World Health Organization recommend that 55% to 60% of the daily caloric intake should be obtained from carbohydrates. Forty-five to fifty percent of these calories should come from complex carbohydrates, and natural sugars found in fresh fruits and vegetables and no more than 10% from refined and processed sugars [13]. Nevertheless, the consumption of carbohydrates in people with the metabolic syndrome remains controversial and needs further investigation.
Fiber is an organic compound found in plants. It is found in the skin of fruits, seeds, leaves, stems and roots. High-fiber diets have received considerable attention in recent years, due to their connection with lower incidence of several metabolic disorders like blood pressure, diabetes, obesity, as well as heart disease and colon cancer.
Fat is a general term used that refers to oils, fats and waxes. There are two types of dietary fats: a) saturated and b) unsaturated. The unsaturated fats are further divided into the monounsaturated and polyunsaturated fats. Usually the daily energy intake consists 30% of fat, but no more than 10% of these calories should come from saturated (animal fats). The rest 20% should come from unsaturated (vegetable) oils. However, whatever the fat intake, certain oils must be included in the diet, like the essential fatty acids. They are polyunsaturated fats derived mostly from vegetable oils such as safflower oil, corn oil, olive oil and soybean oil. Lack of these oils in one's diet will cause series illness.
Proteins, like carbohydrates and fats, contain atoms of carbon, oxygen and hydrogen. In addition, proteins contain nitrogen. Proteins are made up of molecules called amino acids. These amino acids are strung together in a specific order and make a complete protein. Contrary to the beliefs of many, we only need relatively small amounts of protein for good health. The requirements for adults are 0.8 grams per kilogram of body weight.
Vitamins are organic compounds that are necessary for normal metabolism. They are manufactured in the green leaves and roots of plants, except vitamin B-12 that is mainly found in animal foods and products. It is a fact that vitamin deficiency leads to poor health and even death, while vitamin overdose can also lead to health problems. Thus, taking more vitamins than recommended will be of no benefit and may even be harmful. Vitamins like B-6 and B-12 have been associated with lower risk of cardiovascular disease, but this association is still under controversy.
Finally, minerals are metallic elements that play an important role in the metabolism and proper function of the body. Minerals occur freely in nature, and they are found in water, plants, and soil. Most of the major minerals (salt, potassium and iron) have been associated with some metabolic disorders and human health. Excessive consumption of salt causes hypertension in many people or aggravates existing hypertension. On the other hand potassium helps get rid of excess salt in the body. Foods reach in potassium includes bananas and all fruits, vegetables, beans and nuts (unsalted). Iron plays a key role in metabolism, performance and health. It is involved in the transport of oxygen to the tissues, and in a number of biochemical reactions responsible for energy production. Iron is absorbed by the body from iron found in the food consumed. Iron deficiency results in a reduction in the size of red blood cells, and the amount of hemoglobin they contain. The daily iron requirements for males are 10 mg, while females requires almost double. It is common essence that more or less all nutrient components are necessary for human health, however dietary habits should have a balanced intake of the aforementioned food components.
The Mediterranean diet
Although different regions in the Mediterranean basin, have their own diets (like Spain, France, Italy, Greece, Cyprus, etc) it is legitimate to consider these as variants of a single entity, the Mediterranean diet [15]. During the past decades a large body of evidence related adherence to a Mediterranean diet with all causes mortality, prevalence of some metabolic disorders (like obesity, and high blood pressure), as well as incidence of coronary heart disease and various types of cancer [16-20].
This dietary pattern is mainly characterized by daily olive oil consumption. Olive oil is important not only just because it has several beneficial properties, but because it allows the consumption of large quantities of vegetables in the form of salads and equally large quantities of legumes in the form of cooked foods, too. Thus, it might be convenient, if not wholly accurate, defining Mediterranean diet as the dietary pattern found in the olive growing areas of the Mediterranean region, in the late 1950s and early 1960s, when the consequences of World War II were overcome, but the fast-food culture had not yet invaded the area. Other essential components of the Mediterranean diet are wheat, olives and grapes, and their various derivative products. Total lipid intake may be high, (around or in excess of 40% of total energy intake) in Greece, or moderate, (around 30% of total energy intake) in Italy. In all instances, however, the ratio of monounsaturated to saturated fats is much higher than in other places of the world, including northern Europe and North America. Specifically this dietary pattern could be described by the following components: (a) daily consumption: of non refined cereals and products (whole grain bread, pasta, brown rice, etc), vegetables (2–3 servings/day), fruits (4–6 servings/day), olive oil (as the main added lipid) and non-fat or low fat dairy products (1–2 servings/day), (b) weekly consumption: of potatoes (4–5 servings/week), fish (4–5 servings/week), olives, pulses, and nuts (>4 servings/week), and more rare poultry (1–3 servings/week), eggs and sweets (1–3 servings/week) and (c) monthly consumption: of red meat and meat products (4–5 servings/month) [15]. It is, also, characterized by moderate consumption of wine (1–2 wineglasses/day), mainly wine during the meals, and high monounsaturated-to-saturated fat ratio (>2). Additionally, although intake of milk is moderate, the consumption of cheese and yogurt is relatively high. Feta cheese is regularly added to salads and accompanies vegetable stews [15].
The relationship between Mediterranean diet and metabolic syndrome has been investigated in several studies. Several aspects of this dietary pattern have shown a beneficial effect on the development of the metabolic syndrome or its components. Particularly, fish and omega – 3 fatty acids intake, which are essential components of the Mediterranean diet, have been associated with a lower risk of cardiovascular disease. Kris – Etherton et al. in a recent review article summarize epidemiological studies and randomised clinical trials that showed a considerable of effect of fish consumption on cardiovascular system [21]. They concluded that eicosapentaenoic and docosahexaenoic acids supplementation ranging from 0.5 to 1.8 gr per day (either as fatty fish or supplements) significantly reduces subsequent cardiac and all-cause mortality. The mechanisms are not fully understood, but there are strong evidences to support that omega-3 fatty acids intake are associated with lower levels of blood pressure, triglycerides, reduced endothelial activation, and other factors associated with the metabolic syndrome [21]. Moderate alcohol consumption, has been associated with low prevalence of the metabolic syndrome. In particular Rosell et al. [22], in a cross-sectional survey of about 4200 middle aged men and women from Stockholm county in Sweden, observed that the metabolic syndrome was significantly more common in non-drinkers (20%), and less common among wine drinkers (8%), compared to a group with low alcohol intake. Moreover, after various adjustments made the investigators observed 40% lower odds ratio for the metabolic syndrome only in women who consumed wine. Many researchers related Mediterranean diet with improvements in blood pressure and lipid profile (especially LDL cholesterol and triglycerides), decreased risk of thrombosis (i.e. fibrinogen levels), improvement in endothelial function, and insulin resistance, reduction in plasma homocysteine concentrations and decrease in ventricular irritability [23-27]. Particularly, Esposito et al. [24] in a randomized clinical trial among men and women with the metabolic syndrome instructed 90 patients in the intervention group to follow a Mediterranean-style diet (i.e. crease daily consumption of whole grains, fruits, vegetables, nuts, and olive oil), while patients in the control group to follow a prudent diet (carbohydrates, 50%–60%; proteins, 15%–20%; total fat, <30%). After two years, patients consuming the Mediterranean diet had significantly reduced serum concentrations of C-reactive protein, interleukin – 6, insulin resistance, as well as improved endothelial function score. Recently, Psaltopoulou et al. [26] from the Greek cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC) study that included about 20000 healthy people from all regions of the country, observed that greater adherence to the Mediterranean diet was significantly inversely associated with both systolic and diastolic blood pressure. Particularly, intakes of olive oil, vegetables, and fruit were inversely associated with both systolic and diastolic blood pressure, whereas cereals, meat and meat products, were positively associated with arterial blood pressure. Furthermore, antioxidants represent a common element in the Mediterranean pattern and an antioxidant action provides a plausible explanation for the apparent benefits. In a recent review article, Vissers et al. [28] searched the MEDLINE database for the years 1966–2002 and observed that phenol-rich olive oil lowers oxidisability of ex vivo low-density lipoprotein particles or lowers markers in urine of oxidative processes in the body. However, these effects of phenols were not found in human studies. It is known that wild edible greens frequently eaten in the form of salads and pies contain very high quantities of flavonoids. Although there is no direct evidence that these antioxidants are central to the benefits of the Mediterranean diet, indirect evidence from epidemiological data and the increasing understanding of their mechanisms of action suggest that antioxidants may play a major role. Finally, adoption of the Mediterranean diet has been associated with reduced inflammation process, too. In particular, Chrysohoou et al. studied 1514 men and 1528 women from the ATTICA study and observed that greater adherence to the Mediterranean diet (upper tertile of the diet score) was associated with 20% lower C-reactive protein, 17% lower interleukin-6, 15% lower homocysteine, 14% lower white blood cell and 6% lower fibrinogen, as compared to those who were in the lowest tertile, after various adjustments were made [29]. All these findings provided scientific evidences regarding the pathways that relate diet and cardiovascular diseases.
A potential explanation of the beneficial effect of this dietary pattern on human health is because it is low in saturated fat, high in monounsaturated fat, mainly from olive oil, high in complex carbohydrates, from legumes, and high in fibre, mostly from vegetables and fruits. The high content of vegetables, fresh fruits, cereals and olive oil, guarantee a high intake of b-carotene, vitamins C and E, polyphenols and various important minerals. These key elements have been suggested to be responsible for the beneficial effect of diet on human health, and especially cardiovascular disease.
In nutritional epidemiology, interest has shifted from the study of single nutrients or foods, to the study of food groups and, more recently, dietary patterns. The single-nutrient approach fails to account for the interactions between nutrients and does not take into consideration that some nutrients are correlated between them. In this context, the study of dietary patterns, and that of the Mediterranean dietary pattern, has considerable interest. Recently, based on a cross sectional survey in an urban area in Greece (the ATTICA Study), we showed that adherence to this dietary pattern has been associated with 20% lower odds of having the metabolic syndrome, irrespective of age, sex, physical activity status, lipids and blood pressure levels [5]. Williams et al. [30] supported the hypothesis that dietary patterns rich in salads, vegetables, fruits, fish, pasta and rice and low intake in monounsaturated fats are associated with glucose intolerance and various other features of the metabolic syndrome. In particular, he studied 802 middle aged subjects that underwent an oral glucose-tolerance test, and observed that a healthy balanced diet close to the Mediterranean diet was associated with reduced central obesity, fasting plasma glucose, 120 min non-esterified fatty acid and triacylglycerol, and positively correlated with HDL-cholesterol. These results provided further evidence for the recommendation of a healthy balanced diet as one of the main components of chronic disease prevention. The CARDIO2000 study investigators studying 848 patients with a first event of an acute coronary syndrome and 1078 people without any evidence of cardiovascular disease, from all Greek areas, focused their interest in the relationship between Mediterranean diet and cardiovascular risk in people with the metabolic syndrome. They observed that the adoption of the Mediterranean diet was associated with a 35% reduction of the coronary risk in subjects with the metabolic syndrome, after adjusting for various potential confounders [31]. More recently, Esposito et al. [24] showed that a Mediterranean-style diet seems effective in reducing the prevalence of the metabolic syndrome and its associated cardiovascular risk. Particularly, at the end of the follow up 44% patients in the intervention group still had features of the metabolic syndrome, compared to 87% patients in the control group. This may lead to a 51% risk reduction of having the syndrome due to this dietary intervention.
Conclusion
Dietary approaches to treating and preventing metabolic syndrome vary, but nearly all experts agree that clinical parameters are greatly improved through various dietary changes and body weight control. Whether adoption of Mediterranean diet is a recipe for the metabolic syndrome seems to have a scientific basis; however, it needs further investigation by randomised clinical trials. Nevertheless, health care professionals need to help people understand the benefits from the introduced dietary patterns and support them to adopt these lifestyle characteristics.
==== Refs
Hansen BC The metabolic syndrome X Ann NY Acad Sci 1999 892 1 24 10842649
Executive Summary of the 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) JAMA 2001 285 2486 2497 11368702 10.1001/jama.285.19.2486
Ford ES Giles WH Dietz WH Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey JAMA 2002 287 356 359 11790215 10.1001/jama.287.3.356
Ferrannini E Natali A Essential hypertension, metabolic disorders, and insulin resistance Am Heart J 1991 121 1274 1282 2008856 10.1016/0002-8703(91)90433-I
Panagiotakos DB Pitsavos CH Chrysohoou C Skoumas J Tousoulis D Toutouza M Toutouzas PK Stefanadis C The Impact of Lifestyle Habits on the Prevalence of the Metabolic Syndrome among Greek adults from the ATTICA study Am Heart J 2004 147 106 112 14691427 10.1016/S0002-8703(03)00442-3
Oh JY Hong YS Sung YA Barrett-Connor E Prevalence and factor analysis of metabolic syndrome in an urban Korean population Diabetes Care 2004 27 2027 2032 15277435
Lee WY Park JS Noh SY Rhee EJ Kim SW Zimmet PZ Prevalence of the metabolic syndrome among 40,698 Korean metropolitan subjects Diabetes Res Clin Pract 2004 65 143 149 15223226 10.1016/j.diabres.2003.12.007
Hu G Qiao Q Tuomilehto J Balkau B Borch-Johnsen K Pyorala K DECODE Study Group Prevalence of the metabolic syndrome and its relation to all-cause and cardiovascular mortality in non-diabetic European men and women Arch Intern Med 2004 164 1066 1076 15159263 10.1001/archinte.164.10.1066
Santos AC Lopes C Barros H Prevalence of metabolic syndrome in the city of Porto Rev Port Cardiol 2004 23 45 52 15058146
Kannel WB McGee DL Gordon T A general cardiovascular risk profile: The Framingham Study Am J Cardiol 1976 38 46 51 132862 10.1016/0002-9149(76)90061-8
Appel L Moore T Obarzanek E Vollmer W Svetkey L Sacks F Bray G Vogt T for the DASH Collaborative Research Group A Clinical Trial of the Effects of Dietary Patterns on Blood Pressure N Engl J Med 1997 336 1117 24 9099655 10.1056/NEJM199704173361601
Wood D Established and emerging cardiovascular risk factors Am Heart J 2001 141 49 57 10.1067/mhj.2001.109951
World Health Organization Study Group Diet, Nutrition, and the Prevention of Chronic Diseases Geneva, Switzerland: World Health Organization; Technical Report Series, 916, 2003
Schulze MB Hu FB Dietary approaches to prevent the metabolic syndrome Diabetes Care 2004 27 613 614 14747248
Trichopoulou A. Lagiou P Healthy traditional Mediterranean diet – An expression of culture, history and lifestyle Nutr Rev 1997 55 383 389 9420448
Trichopoulou A Lagiou P Kuper H Trichopoulos D Cancer and Mediterranean dietary traditions Cancer Epidemiol Biomarkers Prev 2000 9 869 873 11008902
Kris-Etherton P Eckel R Howard B Jeor S Bazzare T Lyon Diet Heart Study. Benefits of a Mediterranean-Style, National Education Program/AHA Step I Dietary Pattern on Cardiovascular Disease Circulation 2001 103 1823 1825 11282918
Panagiotakos DB Pitsavos C Chrysohoou C Tzioumis K Papaioannou I Stefanadis C Toutouzas PK The Association of Mediterranean Diet With Lower Risk of Acute Coronary Syndromes, in Hypertensive Subjects Int J Cardiol 2002 19 141 147 11853900 10.1016/S0167-5273(01)00611-8
Trichopoulou A Costacou T Bamia C Trichopoulos D Adherence to a Mediterranean diet and survival in a Greek population N Engl J Med 2003 348 2599 2608 12826634 10.1056/NEJMoa025039
Martinez-Gonzalez MA Fernandez-Jarne E Serrano-Martinez M Mediterranean diet and reduction in the risk of a first acute myocardial infarction: an operational healthy dietary score Eur J Nutr 2002 41 153 160 12242583 10.1007/s00394-002-0370-6
Kris-Etherton P Harris W Appel L Fish Consumption, Fish Oil, Omega-3 Fatty Acids, and Cardiovascular Disease Circulation 2002 106 2747 2757 12438303 10.1161/01.CIR.0000038493.65177.94
Rosell M De Faire U Hellenius ML Low prevalence of the metabolic syndrome in wine drinkers – is it the alcohol beverage or the lifestyle? Eur J Clin Nutr 2003 57 227 234 12571653 10.1038/sj.ejcn.1601548
Knapp HW Dietary fatty acids in human thrombosis and hemostasis Am J Clin Nutr 1997 65 1687S 1698S 9129511
Esposito K Marfella R Ciotola M Di Palo C Giugliano F Giugliano G D'Armiento M D'Andrea F Giugliano D Effect of a Mediterranean-style diet on endothelial dysfunction and markers of vascular inflammation in the metabolic syndrome: a randomized trial JAMA 2004 292 1440 1446 15383514 10.1001/jama.292.12.1440
Ruit-Gutierrez V Muriana FJG Guerrero A Plasma lipids, erythrocyte membrane lipids and blood pressure of hypertensive women after ingestion of dietary oleic acid from two different sources J Hypertens 1996 14 1483 1490 8986934
Psaltopoulou T Naska A Orfanos P Trichopoulos D Mountokalakis T Trichopoulou A Olive oil, the Mediterranean diet, and arterial blood pressure: the Greek European Prospective Investigation into Cancer and Nutrition (EPIC) study Am J Clin Nutr 2004 80 1012 1018 15447913
Assmann G de Basker G Bagnara S International Consensus statement on olive oil and the Mediterranean diet: implications for health in Europe Eur JCancer Prev 1997 6 418 421 9466113
Vissers MN Zock PL Katan MB Bioavailability and antioxidant effects of olive oil phenols in humans: a review Eur J Clin Nutr 2004 58 955 565 15164117 10.1038/sj.ejcn.1601917
Chrysohoou C Panagiotakos DB Pitsavos C Das UN Stefanadis C Adherence to the Mediterranean diet attenuates inflammation and coagulation process, in healthy adults: the ATTICA study J Am Col Cardiol 2004 44 152 158 10.1016/j.jacc.2004.03.039
Williams DE Prevost AT Whichelow MJ Cox BD Day NE Wareham NJ A cross-sectional study of dietary patterns with glucose intolerance and other features of the metabolic syndrome Br J Nutr 2000 83 257 266 10884714
Panagiotakos DB Pitsavos C Kokkinos P Chrysohoou C Vavuranakis M Stefanadis C Toutouzas P The Adoption of Mediterranean Diet Attenuates the Development of Acute Coronary Syndromes in People with the Metabolic Syndrome Nutr J 2003 2 2 12773206 10.1186/1475-2891-2-2
| 15826300 | PMC1090612 | CC BY | 2021-01-04 16:39:19 | no | Lipids Health Dis. 2005 Apr 12; 4:7 | utf-8 | Lipids Health Dis | 2,005 | 10.1186/1476-511X-4-7 | oa_comm |
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-191584770410.1186/1475-2875-4-19ResearchEndemic malaria: an 'indoor' disease in northern Europe. Historical data analysed Huldén Lena [email protected]én Larry [email protected]övaara Kari [email protected] Department of Applied Biology, Faculty of Agriculture and Forestry, University of Helsinki, Finland2 Finnish Museum of Natural History, University of Helsinki, Finland2005 25 4 2005 4 19 19 20 1 2005 25 4 2005 Copyright © 2005 Huldén et al; licensee BioMed Central Ltd.2005Huldén 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
Endemic northern malaria reached 68°N latitude in Europe during the 19th century, where the summer mean temperature only irregularly exceeded 16°C, the lower limit needed for sporogony of Plasmodium vivax. Because of the available historical material and little use of quinine, Finland was suitable for an analysis of endemic malaria and temperature.
Methods
Annual malaria death frequencies during 1800–1870 extracted from parish records were analysed against long-term temperature records in Finland, Russia and Sweden. Supporting data from 1750–1799 were used in the interpretation of the results. The life cycle and behaviour of the anopheline mosquitoes were interpreted according to the literature.
Results
Malaria frequencies correlated strongly with the mean temperature of June and July of the preceding summer, corresponding to larval development of the vector. Hatching of imagoes peaks in the middle of August, when the temperature most years is too low for the sporogony of Plasmodium. After mating some of the females hibernate in human dwellings. If the female gets gametocytes from infective humans, the development of Plasmodium can only continue indoors, in heated buildings.
Conclusion
Northern malaria existed in a cold climate by means of summer dormancy of hypnozoites in humans and indoor transmission of sporozoites throughout the winter by semiactive hibernating mosquitoes. Variable climatic conditions did not affect this relationship. The epidemics, however, were regulated by the population size of the mosquitoes which, in turn, ultimately was controlled by the temperatures of the preceding summer.
==== Body
Background
Endemic malaria was declining in western Europe from the 18th century onwards, but in the 19th century it was still common in the north and north-east of Europe. In the 1930s, endemic malaria finally disappeared from most parts of Europe, with the exception of occasional epidemics during the Second World War [1]. There is very little modern research on the historical conditions of the disease or even its historical distribution. Most published studies of historical distribution do not include the cases in Finland [2]. Imported cases during the Second World War, however, have attracted some attention [3].
Northern malaria has usually been connected with Plasmodium vivax. After infection, the sporozoites enter human hepatocytes and develop into hypnozoites. These remain in dormancy for a variable time, and the patient may occasionally suffer from malaria for up to four years and have symptoms every second month [4]. Sokolova & Snow [5] report that Plasmodium falciparum occurred in Archangelsk and Vologda in the northern part of the Soviet Union in the 1930s. It is usually assumed that P. falciparum lacks a dormant stage. In the Finnish medical reports the symptoms of malaria included tertian, quotidian and quartan malaria. The tertian form was the most common and Sievers [6] connected this form with P. vivax. The possible historical occurrence of other Plasmodium species in Finland remains uncertain.
The temperatures needed for the sporogony have been studied in laboratory conditions [4]. Garnham [7,8] limits the range of malaria to summer temperatures. The northern range of the occurrences of endemic malaria has been estimated to coincide with a summer isotherm of 16°C. However, even a superficial examination shows that malaria also occurred in the northern parts of Sweden and Finland, where the summer temperature was considerably below 16°C.
The present study aims to establish that malaria can spread in a cold climate, even when the outside temperature is lower than needed for the sporogony of the Plasmodium. Because of the high quality of available historical sources, of demographic, medical and temperature data, Finland is a very suitable area for the study. The study focused on the period 1800–1870, since the only effective drug, quinine, was not then frequently used by the common people. Its wide use would have biased the statistics of the causes of death. High resolution temperature data are also available for that period.
Methods
The known distribution of three Anopheles species in Finland was published by Utrio [9]. The distribution unit of 50 × 50 km2 gives a reasonably good resolution of the distribution of these species in Finland. Additional information from Sweden and Estonia was also used for comparison [10,11]. Life cycles, behaviour and the phenology of the mosquito has been interpreted according to the literature [12-18] and field samplings in the south-west archipelago by the first author. The data of historical malaria cases were collected and tested by historical methods. Although the historical data can be questioned, it should be emphasized that the Finnish sources give a good picture of the actual situation, and that the statistics are of a better quality than in most other countries. Causes of death were especially recorded by ministers in hundreds of parishes and, although there probably are some mistakes, most of the entries were correct.
There are two different kinds of source for collecting the records of malaria cases in Finland during the years 1800–1870: 1) reports by the district physicians and 2) the causes of death that were recorded by the local ministers. The physicians wrote an annual report to the medical board in Helsinki about the health conditions in their district. At the end of the century a separate report on epidemics was added. Malaria is mentioned in both kinds of reports. The reports had a free format and the number of sick patients was only rarely mentioned. Usually the physician formulated an opinion of the severity of the actual epidemics. Many districts were so large that the physician could hardly know the actual situation in the more remote parts, especially in the north. The archival series of physicians' reports include reports from the beginning of the 19th century. The oldest reports contained very general information. They improved in 1857 when the number of districts increased to 50 [19]. Malaria was mentioned in 542 reports between 1826 and 1870. Sievers [6], who studied malaria in Finland, based his opinion solely on these reports. Sievers' interpretation was compared with the original documents, that can be examined in the National Archives in Helsinki, Finland. The reports cannot be used for quantitative analyses, but they mention the years in which they were epidemics and these coincide with the epidemic years in the death statistics.
From 1749, the minister had to record the cause of death for every diseased parish member in the burial records. The parish burial registers can therefore be used for detailed statistics on malaria. For this study these local records were used. The minister often attended the deathbed and, in contrast to the district physician, actually saw the patient. Ministers were probably very familiar with the typical symptoms of malaria, as it was a common disease. They often had some medical knowledge because the training in medicine for students of theology had developed rapidly duringthe second half of the 18th century. The minister was not only responsible for the spiritual guidance of his parish members, but also for their bodily welfare [20]. In the present study, only those cases which the ministers recorded as malaria (5,431 death cases) have been taken into account. In the parish records, several words were used for malaria. Frossa, fråssa, frossfeber were the most common and other frequently used words were kallfeber, kallsot, kylfeber, skälvan, skälvasot, tertian, tredjedagsfrossa, vardagsfrossa, skärgårdsfeber, växelfeber, omväxlingsfeber, intermittent feber and malaria. The district physicians used frossa, tredjedagsfrossa, vardagsfrossa, tertian, quartana or quotidiana. These words were used in both Swedish and Finnish medical research from the 18th to the 20th century along with skärgårdsfeber [6,20-27]. The use of the above-mentioned words was strictly separated from the use of other words for diseases, which included fever or resembled malaria like hetsig feber, flussfeber, feber, scharlakansfeber, älta and förkylning. The consistent use of these malaria words is also confirmed by comparative correlation analyses. The statistics for cause of death is most representative for the period of 1800–1850. The data are collected from digitalized parish records, which were made available on the Internet by the Finnish Genealogical Society through Dec. 31st, 2003 [28]. The project is not yet finished, but most of the material before 1850 has been digitalized. The period of 1850–1900 is not very representative and only material from a few parishes is included. In the present study all cases recorded in southern Finland until 1870 are included but the last decade is under-represented in the material. It must be stressed that the records only show the number of deaths due to malaria but that the relation between the number of infected and the death rate is uncertain.
The longest Finnish temperature record from Helsinki, starting from 1829, is used as the basis for climatic correlations for the years 1829–1870. Temperatures from St. Petersburg (Russia) (1805–1870), Tornedalen (Sweden)(1818–1870) and Stockholm (Sweden)(1800–1870) are used as supporting data for the extended period of 1800–1870 (Figure 1). These are the best high resolution temperature data available at the present time [29-32].
Figure 1 Map of malaria study area and locations of temperature series.
Results
The distribution and life cycle of anophelines in Finland
Three Anopheles species have been reported from Finland [9]. Anopheles beklemishevi has a northern distribution in Finland, while the other common species, Anopheles messeae, is dominant in the southern part of the country [17]. Anopheles claviger is recorded from the Aland Islands and was also found by the authors in the south-western archipelago. In December 2004 and January 2005 flying specimens of an unidentified Anopheles (cf. messeae) were observed in summer cottages in two different localities in southern Finland.
Because of many taxonomical problems in the Anopheles genus, there is conflicting information on the details of the life cycles of the separate species. A general feature of the northern species is that the adult female hibernates. There are possibilities of a summer generation, at least in warm summers as in 1901 in Finland (possibly A. messeae)[9,12]. A. claviger hibernates as larvae [33] and, at least in England, seems to be bivoltine [34]. In Finland, the majority of the hibernating females reach the adult stage in the middle of August. The male is short lived and dies soon after mating. The female may suck blood or honey-dew before seeking shelter in sheds and houses for hibernation. Because of the cold climate the female cannot leave the hibernation site before spring. In the traditional agricultural society of Finland many of the mosquito females spent the main part of their adult life together with man and domestic animals. In warm conditions the female may take several blood meals during the hibernation, but it will not lay eggs before the spring [13].
A. messeae has been considered an important vector of historical endemic malaria in several parts of Europe and was also an important vector in the Soviet Union [4,15]. In the northern parts of Russia and along the Baltic coast, A. beklemishevi was an important vector [35]. A. claviger has also locally been reported as a vector [15]. Some recent research, however, has questioned A. messeae as a vector of malaria [36]. Jaenson & Ameneshewa [37] reported that blood-feeding was not a prerequisite for hibernation of A. messeae. This does not exclude the possibility of repeated indoor blood-feeding in warm indoor conditions. There is still unresolved and conflicting information on the feeding, resting and hibernation habits of the female (exophagy or endophagy, zoophily or anthropophily, exophily or endophily, complete diapause or semiactive winter habit). The anophelines in the northern region, however, must be flexible in their habits to survive the strong annual and seasonal fluctuations in temperature (-51°C to +33°C in Finland and rain/snow). In any case, one or more species, distributed over the whole of Finland, have been vectors of malaria in Finland. Yet it is likely that one or more additional species, still not reported from Finland, will be found as four additional species have been reported from Estonia and Sweden [10,11]. For the time being it is not possible to define which mosquito species was important for the malaria transmission in Finland.
Endemic malaria in Finland
The Anopheles species, which have been vectors of malaria in Finland, presumably have existed here since prehistoric times. It is, however, probable that malaria reached Finland only during the 17th century. There are several cases of malaria in the Mälaren region to the west of Stockholm in Sweden during that time. Both 1692 and 1693 are reported as years of severe fever [38], which spread rapidly also to the east. A total of 1,803 persons died of malaria in the western parts of Finland and in the south-western archipelago during the years 1751–1773 [23]. Haartman [21] reports severe epidemics in the region of Turku in the years 1774–1777 and the physician F.W. Radloff mentioned that malaria was very common in the Aland Islands in 1795 [39].
From 1800 to 1870, there were at least 5,431 death from malaria in Finland. During this period the Finnish population grew from 832,700 to 2,032,700 people [40]. In a few epidemics, the mortality can be estimated to 0.85–3.0% [6]. The number of deaths from malaria increased when the number of malaria cases increased and thus a change in virulence of the Plasmodium is not to be assumed. The total number of malaria cases in the middle of the 19th century may tentatively be estimated to be 100,000–300,000 or about 7–20% of the whole population.
Although the source value of both the death statistics and the medical reports can be discussed separately, they are strengthened when compared with each other. As sources they are completely independent, and they partly overlap the period 1850–1870, showing similar trends. The district physicians reported 174 local epidemics during 1854–1862 and during the same period the deaths from malaria rose to 1,687 cases. In the years 1873–74 and 1877, the number of epidemics also rose and the same trend can be seen in the parish death records.
The malaria epidemics and temperature
The worst malaria year in Finland was 1862. The eastern parts of Finland were particularly affected, including Karelia. Several physician's reports reveal how severe the situation was. In the district of Mikkeli over 4,000 persons became ill. Moreover, in the district of Joensuu 4,000 persons suffered, in Rautalampi several thousands and in the worst parishes in the districts of Viipuri and Muola every third person became ill. In eastern Finland the summer was cold and damp. Malaria epidemics even broke out in the north and in Sotkamo in the Kajana district more than a hundred people became ill.
The correlation between temperature and the number of malaria deaths was tested on annual, seasonal and monthly levels. The seasons were interpreted as follows: winter (DJF), spring (MAM), summer (JJA) and autumn (SON). The malaria cases were tested against the temperatures of both the current malaria year and the preceding year. In Table 4 the calculations of monthly correlations were repeated with a square root transformation of the number of malaria cases to avoid biases caused by strong peaks during epidemics. Malaria deaths peak late in the spring but for completeness the calculations are extended further to the autumn of the malaria year. The results are presented in Tables 1,2,3,4 and Figure 2.
Table 1 Annual mean temperature correlated with malaria cases in 1800–1870.
Year correlated with malaria cases in 1800–1870 Source of temperature series
Helsinki 1830–1870
41 years Tornedalen 1818–1870
53 years Stockholm 1800–1870
71 years St. Petersburg 1806–1870
65 years
malaria year-1 0.1364 0.1133 0.1238 0.0958
malaria year 0.0996 -0.1177 -0.0822 -0.1164
Table 2 Seasonal mean temperature correlated with malaria cases in 1800–1870
Season correlated with malaria cases in 1800–1870 Source of temperature series
Helsinki 1830–1870
41 years Tornedalen 1818–1870
53 years Stockholm 1800–1870
71 years St. Petersburg 1806–1870
65 years
malaria year-1; winter -0.1932 -0.1209 -0.1652 -0.0789
malaria year-1; spring -0.0091 -0.0228 0.0248 -0.0312
malaria year-1; summer 0.5120 *** 0.5208 *** 0.4708 *** 0.3981 ***
malaria year-1; autumn 0.1218 0.1621 0.1055 0.1598
malaria year; winter -0.1254 -0.0743 -0.1214 -0.1543
malaria year; spring 0.1140 0.0880 0.0223 0.0546
malaria year; summer -0.0740 -0.0332 0.0201 -0.0439
malaria year; autumn -0.0831 -0.0819 -0.0211 -0.0797
Table 3 Monthly mean temperature correlated with malaria cases in 1800–1870.
Month of preceding year correlated with malaria cases in 1800–1870 Source of temperature series
Helsinki 1830–1870
41 years Tornedalen 1818–1870
53 years Stockholm 1800–1870
71 years St. Petersburg 1806–1870
65 years
malaria year-1; May 0.1487 0.0269 0.0511 0.1338
malaria year-1; June 0.4301 ** 0.5289 *** 0.3372 ** 0.3587 **
malaria year-1; July 0.4628 ** 0.4135 ** 0.4543 *** 0.2888 *
malaria year-1; August 0.2672° 0.1570 0.2740 * 0.1793
malaria year-1; September 0.1156 0.0655 0.0637 0.0678
malaria year-1; October 0.1566 0.0200 0.1326 0.2657 *
malaria year-1; November 0.1462 0.1958 0.0000 0.0220
malaria year-1; December 0.0037 -0.002 -0.1144 -0.1221
Table 4 Monthly mean temperature correlated with square root transformed malaria cases in 1800–1870
Month of preceding year correlated with square root transformed malaria cases in 1800–1870 Source of temperature series
Helsinki 1830–1870
41 years Tornedalen 1818–1870
53 years Stockholm 1800–1870
71 years St. Petersburg 1806–1870
65 years
malaria year-1; May 0.1512 0.0770 0.0623 0.1389
malaria year-1; June 0.4407 ** 0.5739 *** 0.4049 *** 0.3780 **
malaria year-1; July 0.4801 ** 0.4511 *** 0.5081 *** 0.3275 **
malaria year-1; August 0.2428 0.1174 0.2610 * 0.1657
malaria year-1; September -0.1479 0.0469 0.0824 0.0280
malaria year-1; October 0.1591 0.0322 0.1487 0,2145°
malaria year-1; November 0.1428 0.2048 0.0039 0.0137
malaria year-1; December 0.0190 0.0055 -0.1049 -0.1250
Figure 2 Malaria deaths in southern Finland and mean temperatures of the preceding summer in Helsinki and Tornedalen 1830–1870.
Annual mean temperatures were obviously not significant in the correlations. On the seasonal level the summer of the preceding year is highly significant (0.1% risk level). In all other seasons the risk level was more than 10%. Renkonen [26] tried to show a correlation between malaria epidemics and the temperature of April or spring. The correlation coefficient for both April and the entire spring were close to zero and thus Renkonen's hypothesis can be rejected.
June and July of the preceding summer were very significant with a faint "tail" in August. The weak correlation of October temperatures in the St. Petersburg series (5% and 10% risk level respectively in Tables 3,4) can safely be considered stochastic noise with no support from the other series.
Death caused by what were called 'intense fever', hetsig feber (48853 cases), cold, influenza, flussfeber (7670 cases), fever, feber (64604 cases) and cold, förkylning (1381 cases) were tested for correlations with both temperatures and with malaria in 1800–1850, but no correlations were found. The words expected to indicate malaria are well-defined and separated from other kind of fevers and they can safely be analysed as a homogeneous disease.
The correlation is strongest with the season when the vector is in the larval stage. Thus the number of malaria cases correlates with the warmest time of the summer when there is no Plasmodium present in connection with the mosquito. The conclusion is that fluctuations in summer temperatures regulate the number of mosquitoes reaching the adult stage and not the development of Plasmodium. Because malaria epidemics covariate with the summer temperatures, the sporogony inside the mosquito may be treated as a constant parameter. Thus the sporogony of Plasmodium in Finland must have taken place in reasonably stable temperature conditions.
The sporogony is temperature dependant and the duration of the sporogony of P. vivax is 7–8 days in 30°C, 8–10 days in 28°C, 15–16 days in 20–21°C and 55 days near 16°C with development stopping completely below 16°C [4]. When the mosquito reached the adult stage in August, the temperature was already falling. If the mosquito took a blood meal from a carrier of malaria late in the summer or at the beginning of autumn, the malaria parasite would have had no time to complete its sporogony in the vector in outside conditions. August was the warmest month the adult mosquito female encountered during its life. The mean temperatures of August and May during the years 1829–1870 were in Helsinki 15.53°C (range 12.3–20.6°C) and 7.34°C respectively and the corresponding temperatures in Tornedalen 13.15°C (range 9.9–16.7°C) and 4.53°C. These values give a realistic picture of the temperature conditions within the area where endemic malaria occurred in Finland. The summer season in Finland was within a wide range clearly unsuitable for the development of Plasmodium (Figure 3).
Figure 3 Mean temperature of August in Helsinki and Tornedalen 1829–1870.
The principal explanation for continuation of endemic malaria must be the choice of hibernation shelter by the female mosquito, within a warm environment in close proximity to humans. Such places in previous centuries were animal shelters and human dwellings. An indoor transmission cycle was first suggested by Robert Koch in the 1890's [25] and Swellengrebel and de Buck [41] gave a detailed description of indoor infection by hibernating mosquitoes in the Netherlands. The situation in Finland was different. Sivén [25] noted during the epidemic in Helsinki in 1902, that many infants who got malaria in the spring, were born in the winter. In the years 1750–1850 there were about 40–50 corresponding death cases. As a consequence hibernating mosquito females were active and capable of infecting humans during the whole winter.
The relative proportion of annual deaths in different age groups of the total population was compared in Figures 4,5,6, for malaria and seven other diseases in the years 1750–1850. The age group structure of malaria deaths is very similar to some other fevers and approaches that of the total population. Deviations in the total population are mainly explained by increased (>50%) omissions of the cause of death in children under two years old in the parish records. Malaria seems to spread with equal probability among all age groups. If the majority of infections had occurred during a limited time in late summer or the beginning of autumn, the resulting age group distribution of malaria deaths would have been different. In the traditional agricultural society, certain age groups used to live apart during the summer, usually in unheated buildings. In October-November, decreasing temperatures forced them to move back into heated buildings [42]. Thus, the most conservative explanation is an extended cold season infection time when all age groups congregate in heated buildings.
Figure 4 Relative proportions of annual age classes (of deaths) of the total population, malaria and seven other diseases in 1750–1850 (110,538 cases without age data was excluded), in the 0–25 year age group. The number of cases are as follows: (1) total population 1,143,547; (2) malaria 3,128; (3) hetsig feber (intense fever) 38,368; (4) feber (fever) 32,426; (5) flussfeber (cold, influenza) 4,681; (6) scharlakansfeber (scarlet fever) 10,202; (7) smittkoppor (smallpox) 100,768; (8) lungsot (phthisis) 13,674; (9) venerisk sjukdom (venereal disease) 462
Figure 5 Relative proportions of annual age classes (of deaths) of the total population, malaria and seven other diseases in 1750–1850, in the 26–50 year age group. The number of cases are as follows: (1) total population 238,047; (2) malaria 2,554; (3) hetsig feber (intense fever) 22,325; (4) feber (fever) 16,038; (5) flussfeber (cold, influenza) 3,690; (6) scharlakansfeber (scarlet fever) 95; (7) smittkoppor (smallpox) 902; (8) lungsot (phthisis) 33,950; (9) venerisk sjukdom (venereal disease) 364
Figure 6 Relative proportions of annual age classes (of deaths) of the total population, malaria and seven other diseases in 1750–1850, in the 51–80+ year age group. The number of cases are as follows: (1) total population 413,775; (2) malaria 3,364; (3) hetsig feber (intense fever) 25,560; (4) feber (fever) 21,816; (5) flussfeber (cold, influenza) 4,773; (6) scharlakansfeber (scarlet fever) 36; (7) smittkoppor (smallpox) 66; (8) lungsot (phthisis) 31,122; (9) venerisk sjukdom (venereal disease) 172
Hernberg [43,44] studied the malaria epidemic in Finland during the years 1941–1945, resulting from infections acquired during the war. He concluded that P. vivax has a variable incubation time of 5–12 months or more. Another conclusion was, that the spring peak of malaria was a result of an infection in the previous year, because he assumed that no mosquitoes were present during the winter. He obviously did not note Sivén's observation of winter infections. Winter infections were probably the reason why Renkonen tried to correlate the infections with the mean temperature of March [26].
According to Swellengrebel and de Buck [41], the main infection time of malaria in the Netherlands was from the middle of August to the middle of November and there were practically no infections after November because of degenerating sporozoites. The degeneration of sporozoites late in the autumn was due to low temperature in unheated bedrooms. In Finland, however, hibernating mosquitoes have been able to infect humans during the whole winter because of heated bedrooms. The importance of the dormancy of the Plasmodium is pronounced in the summer, as the parasite needs a prolonged dormant stage in humans to survive the summer. The summer temperature conditions are unstable and highly variable and there are no or too few anopheline females during a considerable part of the warm season. The winter temperature conditions are stable inside human dwellings and the mosquito female can repeatedly produce sporozoites and infect humans. Some of the sporozoites immediately initiate schizogony while most of them remain as dormant hypnozoites in the hepatocytes. Another source for schizonts may be gradual activation of dormant hypnozoites from the preceding winter. As a consequence, during the erythrocytic cycle some gametocytes are produced throughout the winter in high latitudes. It has been suggested that hormonal changes in the host, controlled by the changing light-cycle, induce mass activation of the hypnozoites in the spring [45]. The spring peak of hypnozoites could be of no importance for the Plasmodium in Finland. Some of the hypnozoites, however, were activated only in the following autumn or later. Thus, paradoxically, summer dormancy was crucial for maintaining the Plasmodium until the next generation of anopheline females were present. Such a pattern had already been suggested, in parts, by Reiter [1].
Discussion
Endemic malaria in Finland was independent of outdoor temperature conditions. The infection could not spread in Finland during the summer because possible summer generations of mosquitoes are irregular and short-lived and the female lays eggs immediately after its blood meal. The hibernating generation of adult anophelines hatches in the late summer when the temperature is too low for the sporogony to be completed. After mating they seek shelter in human dwellings to hibernate. During the main part of its adult life the female was inside a house together with humans. The hatched female may take its first blood meal outdoors but sporogony can only be completed indoors. Endemic malaria could thus only exist in Finland as an 'indoor' disease.
The spreading of malaria is best explained by movements of a large number of working people. Large working sites in Finland – the construction of the Saimaa canal in 1845–54, the lowering of lakes in the 19th century, the building of the railway in 1862–71, etc. – brought temporary populations to uninfected areas. Some of the workers may have been carriers of Plasmodium. During the last weeks of the working season the adult anopheline females appeared and took their first blood meals and some of them became infected by Plasmodium. The working site was dispersed for the winter and the workers returned to their home villages and towns. The infected anophelines sought shelter in the houses of the local population, and the following spring the whole household was infected. A district physician reported from Ostrobothnia in 1845 and 1846 that malaria was common among carpenters, who had been working in the south of Finland [6]. The number of infected anophelines also grew significantly during the winter when uninfected females fed on infected humans. The cold season lasted for almost 8 months and since the incubation period for P. vivax could be 3–4 weeks in 18°C, it is probable that if the infection reached the household in September, all the members and the entire hibernating Anopheles population were infected by the following spring.
Conclusion
The present study of the frequency of northern malaria cases and temperature data establishes the relation between malaria and climate. Endemic malaria existed exclusively as an indoor disease independently of climate. The temperature of the preceding summer regulated the amplitude of malaria cases. Infected people spread malaria into new regions, where mosquito populations transferred malaria locally. The mosquito-human interaction in human dwellings then maintained the introduced malaria in the new region.
P. vivax was the principal disease factor in Scandinavia and Finland. P. falciparum was not limited by winter conditions, but the lack of a summer dormant stage has been crucial in a cool maritime summer in north-western Europe. The continental summer is usually warm which could explain occasional outbreaks of P. falciparum malaria in northern Russia in the 1920's and 1930's. The final decline of malaria in the 20th century in Finland is not explained by the current results. A statistical analysis of the changing demographic and ecological conditions is needed for a definitive explanation of that issue.
Authors' contributions
LeH drafted the manuscript and collected the historical data. LaH contributed the temperature data and participated in the design of the study. KH conceived the study and participated in its design. All authors read and approved the final manuscript.
Figure 7 Deaths from malaria in relation to mean temperature of the preceding summer in Tornedalen 1803–1870.
Acknowledgements
This study was supported by grants from Svenska Litteratursällskapet i Finland.
==== Refs
Reiter P Climate change and mosquito-borne diseases Environ Health Perspect 2001 109 141 161 11250812
Wernsdorfer WH Kreier JP The importance of malaria in the world Malaria 1980 1 Academic Press, New York 1 93
Bruce-Chwatt LJ de Zulueta J The rise and fall of malaria in Europe A historico-epidemiological study 1980 Oxford University Press, Oxford
Kettle DS Medical and veterinary entomology 1995 2 CAB International, Cambridge
Sokolova MI Snow KR Malaria vectors in European Russia Eur Mosq Bull 2002 12 1 6
Sievers R Om frossan i Finland Finska Läkaresällskapets handlingar 1891 33 563 734
Garnham PCC Wernsdorfer WH, McGregor I Malaria parasites of man: life-cycles and morphology (excluding ultrastructure) Malaria Principles and Practice of Malariology 1988 1 Churchill Livingstone, Edinburgh 61 96
Garnham PCC Kreier JP Malaria in its various vertebrate hosts Malaria 1980 1 Academic Press, New York 95 144
Utrio P Geographic distribution of mosquitoes (Diptera, Culicidae) in eastern Fennoscandia Notulae Entomol 1979 59 105 123
Dahl C Taxonomy and geographic distribution of Swedish Culicidae (Diptera, Nematocera) Entomol Scand 1977 8 59 69
Remm HJ Materialy po faune i ekologii komarov (Diptera, Culicidae) Estonskoi SSR Entomologicheskoe Obozrenie 1957 36 148 160
Levander KM Mittheilungen über Anopheles claviger Fabr. in Finland Acta Societatis pro Fauna et Flora Fennica 1902 21 1 30
Ekblom T Ströman R Geographical and biological studies of Anopheles maculipennis in Sweden from an epidemiological point of view Kungliga Svenska Vetenskapsakademiens handlingar Ser 3 1932 11 1 113
Ekblom T Les races suédoises de l'Anopheles maculipennis et leur rôle épidémiologique Bull Soc Path Exot 1938 31 647 655
Gutsevich AV Monchadskii AS Shtakelberg AA Mosquitoes. Family Culicidae Fauna of the USSR Diptera 1974 3 1 408 (Jerusalem, English translation)
Utrio P Distribution of mosquitoes (Diptera, Culicidae) attracted by man in Finland in early July Notulae Entomol 1978 58 107 114
Lokki J Saura A Korvenkontio P Ulmanen I Diagnosing adult Anopheles mosquitoes Aquilo Ser Zool 1979 20 5 12
Becker N Petric D Zgomba M Boase C Dahl C Lane J Kaiser A Mosquitoes and their control 2003 New York, Boston, Dordrecht, London, Moscow, Kluwer Academic/Plenum Publishers
Hirvonen L Lääketieteen historia 1987 Hanko, Recallmed Oy
Lindroth S Svensk lärdomshistoria Frihetstiden 1978 Stockholm, Norstedts Förlag AB
Haartman J Om orsakerna til de svåra Frossor, som åren 1774, 75, 76 och 1777 kring Åbo och annorstädes grasserat Kongliga Vetenskaps Academiens Nya Handlingar 1781 2 30 49
Haartman J Florin JC Om skärgårds febren omkring Åbo. Til disputations ämne för stipendiarierne 1781 Thesis. University of Åbo
Rabbe FJ Några anteckningar om frossan i Finland samt botemedlen deremot Finska Läkaresällskapets handlingar 1856 6 185 216
Stigzelius LH Anteckningar om Malaria epidemierna på Åland under sednaste århundrade 1864 PhD thesis. University of Helsinki, Faculty of Medicine
Sivén VO Om frossan i Helsingfors i belysning af den moderna myggteorin Finska Läkaresällskapets handlingar 1904 47 99 134
Renkonen KO Über das Vorkommen von Malaria in Finnland Ada Med Scand 1944 69 261 275
Lagerkranz G Svenska sjukdomsnamn i gångna tider Stockholm 2001 Sveriges Släktforskarförbund
Finnish Genealogical Society
Camuffo D Jones PD Improved Understanding of Past Climatic Variability from Early Daily European Instrumental Sources 2002 Dordrecht/Boston/London, Kluwer Academic Publisher
Polyakov IV Alekseev GV Bekryaev RV Bhatt U Colony RL Johnson MA Karklin VP Makshtas AP Walsh D Yulin AV Observationally based assessment of polar amplification of global warming Geophys Res Lett 2002 29 1878 (doi:10.1029/2001GL011111) 10.1029/2001GL011111
Moberg A Bergström H Homogenization of Swedish Temperature Data. Part III: The Long Temperature Records from Uppsala and Stockholm Int J Clim 1997 17 667 700 10.1002/(SICI)1097-0088(19970615)17:7<667::AID-JOC115>3.3.CO;2-A
Klingbjaer P Moberg A A composite monthly temperature record from Tornedalen in northern Sweden, 1802–2002 Int J Clim 2003 23 1465 1494 10.1002/joc.946
Utrio P Identification key to Finnish mosquito larvae (Diptera, Culicidae) Annales Agriculturae Fenniae 1976 15 128 136
Service MW The biology of Anopheles claviger Meigen (Diptera: Culicidae) in southern England Bull Entomol Res 1973 63 347 359
White GB Systematic reappraisal of the Anopheles maculipennis complex Mosq Syst 1978 7 303 344
Takken W Geene R Adam W Jetten TH van der Velden JA Distribution and dynamics of larval populations of Anopheles messeae and A. atroparvus in the delta of the rivers Rhine and Meuse, the Netherlands Ambio 2001 31 212 218 12164130
Jaenson TG Ameneshewa B Prehibernation diet and reproductive condition of female Anopheles messeae in Sweden Med Vet Entomol 1991 5 243 252 1768915
Linder J Tanckar om Frossan och Kin-Kina barken 1717 Stockholm, Joh. L. S. Horn
Radloff FW Beskrifning öfver Åland 1795 Åbo
Vattula K The Economic History of Finland 3 Historic statistics 1983 Helsinki, Kustannusosakeyhtiö Tammi
Swellengrebel NH de Buck A Malaria in the Netherlands 1938 Amsterdam, Scheltema & Holkema Ltd
Sirelius UT Suomen kansanomaista kulttuuria 1921 Helsinki, Kustannusosakeyhtiö Otava
Hernberg CA The epidemiology in malaria tertiana in Finland during the years 1941–1945 Acta Med Scand 1947 77 342 360
Hernberg CA Clinical observations on malaria tertiana in Finland, and on difference between autumn and spring malaria Ada Med Scand 1947 78 428 451
Paul REL Diallo M Brey PT Mosquitoes and transmission of malaria parasites – not just vectors Malar J 2004 3 39 15533243 10.1186/1475-2875-3-39
| 15847704 | PMC1090613 | CC BY | 2021-01-04 16:37:30 | no | Malar J. 2005 Apr 25; 4:19 | utf-8 | Malar J | 2,005 | 10.1186/1475-2875-4-19 | oa_comm |
==== Front
Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central London 1743-7075-2-91581713310.1186/1743-7075-2-9ResearchDifferential effect of obesity on bone mineral density in White, Hispanic and African American women: a cross sectional study Castro Jonathan P [email protected] Linda A [email protected] John J [email protected] Surender K [email protected] John [email protected] Joshua [email protected] Irina [email protected] Irina [email protected] Vincent [email protected] Gul [email protected] Leon [email protected] Lina [email protected] Sara [email protected] Nilofar [email protected] Pramodini [email protected] Rangnath [email protected] Hans [email protected] Samy I [email protected] Department of Medicine, State University of New York Downstate Medical Center, Brooklyn, New York 11203, USA2 Division of Endocrinology, Diabetes and Hypertension, State University of New York Downstate Medical Center, Brooklyn, New York 11203, USA3 Scientific Computing and Statistics Center, State University of New York Downstate Medical Center, Brooklyn, New York 11203, USA2005 7 4 2005 2 9 9 25 2 2005 7 4 2005 Copyright © 2005 Castro et al; licensee BioMed Central Ltd.2005Castro 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.
Osteoporosis is a major public health problem with low bone mass affecting nearly half the women aged 50 years or older. Evidence from various studies has shown that higher body mass index (BMI) is a protective factor for bone mineral density (BMD). Most of the evidence, however, is from studies with Caucasian women and it is unclear to what extent ethnicity plays a role in modifying the effect of BMI on BMD.
A cross sectional study was performed in which records of postmenopausal women who presented for screening for osteoporosis at 2 urban medical centres were reviewed. Using logistic regression, we examined the interaction of race and BMI after adjusting for age, family history of osteoporosis, maternal fracture, smoking, and sedentary lifestyle on BMD. Low BMD was defined as T-score at the lumbar spine < -1.
Among 3,206 patients identified, the mean age of the study population was 58.3 ± 0.24 (Years ± SEM) and the BMI was 30.6 kg/m2. 2,417 (75.4%) were African Americans (AA), 441(13.6%) were Whites and 348 (10.9%) were Hispanics. The AA women had lower odds of having low BMD compared to Whites [Odds ratio (OR) = 0.079 (0.03–0.24) (95% CI), p < 0.01]. The odds ratio of low BMD was not statistically significant between White and Hispanic women. We examined the interaction between race and BMD. For White women; as the BMI increases by unity, the odds of low BMD decreases [OR = 0.9 (0.87–0.94), p < 0.01; for every unit increase in BMI]. AA women had slightly but significantly higher odds of low BMD compared to Whites [OR 1.015 (1.007–1.14), p <0.01 for every unit increase in BMI]. This effect was not observed when Hispanic women were compared to Whites.
There is thus a race-dependent effect of BMI on BMD. With each unit increase in BMI, BMD increases for White women, while a slight but significant decrease in BMD occurs in African American women.
==== Body
Background
Osteoporosis is a chronic bone disease characterized by low bone mass and microarchitectural disruption, leading to bone fragility and an increased susceptibility to fractures [1,2]. As a result of increased life expectancy in the population as well as greater public awareness, osteoporosis has evolved from an often overlooked disease entity into a recognized health problem approaching epidemic proportions. In the United States, the disease affects 7.8 million women aged 50 or older accounting for over 1.5 million fractures per year with annual health care expenditures exceeding $13 billion dollars. The National Osteoporosis Foundation reports that close to 22 million women aged 50 or older have low bone mass with projections thru year 2010 and 2020 reaching almost 26 million and over 30 million respectively [3-5]. Third National Health and Nutrition Examination Survey (NHANES III) reported 50%-68% estimated national prevalence of low BMD observed among women aged 50 years or older [6]. The National Osteoporosis Risk Assessment (NORA) study, the largest study of postmenopausal osteoporosis conducted in the United States, with more than 200,000 women aged 50 or older, demonstrated low BMD in almost half of the study population [7]. These numbers, by any standard, portend substantial societal and economic burdens that underscore the importance of disease prevention.
Preventing osteoporosis, a multifactorial disease, resides not only in recognizing its risk factors, but also in identifying potentially modifiable determinants of bone mineral density (BMD), the surrogate measure for osteoporosis [8]. As a result of mounting evidence suggesting beneficial effects on bone, increased body weight has emerged in recent years as a potential modifier of osteoporosis risk [7,9-13]. Likewise, body mass index (BMI), a height-adjusted derivative of body weight, has been accorded the same attention [14-16]. Evidence from the NORA study also demonstrated decreased odds of osteoporosis with increasing BMI and reported BMI as a BMD- protective factor [7]. Among postmenopausal women of Caucasian descent, sufficient evidence has been garnered to suggest that moderate obesity is a protective factor in osteoporosis [7]. In this high risk group, BMI has been shown to positively correlate with BMD [7,17,18]. As BMI increases, BMD increases while the rate of bone loss decreases [19]. The exact physiologic mechanisms that underlie these beneficial effects are unknown but mechanical loading on weight-bearing bones and estrogen synthesis in adipose tissue have been suggested as possible mechanisms [18,20]. Furthermore, studies evaluating the effect of low BMI on BMD have shown unfavourable results. In the EPIC study, early postmenopausal women in the lowest tertiles of BMI were shown to have baseline BMD's that were nearly 12 percent lower and experienced over 2-fold increase in bone loss at two years when compared to the highest tertiles of percentage of body fat or BMI [21]. This indicates that a low BMI is an important risk factor for osteoporosis by predisposing to lower peak bone mass and accelerated bone loss. These studies, however, were mainly done on Caucasian women. It is unclear as to what extent; ethnicity plays a role in modifying the effect of BMI on BMD. National and regional surveys in United States demonstrate significant ethnic differences with higher incidence of hip fracture in Caucasian women compared with African American and Mexican American women [22,23]. Similarly, evidence from the NORA study shows increased odds of osteoporosis for Asian and Hispanic women compared to White women as well as decreased odds for African American women. Although prevalence of osteoporosis was higher among Asian and Hispanic women than among Whites, the likelihood of fracture was no different for Hispanics and was in fact lower in Asians [7]. Since, race is an important determinant of body structure and by extension, of body weight and BMI; it seems logical to assume that racial differences in anthropometric constitution may partially explain differences in osteoporosis risk.
To further clarify the beneficial effects of an increasing BMI on BMD, we evaluated the race-dependent effect of BMI on BMD by comparing postmenopausal women from three different racial backgrounds: African American, Hispanic and Caucasian. We hypothesized that, given their inherent advantage of having denser bones compared to Caucasian women, African American or Hispanic women, should have further reduction in osteoporosis risk in the presence of an increasing BMI [24-26].
Methods
From December 2002 to December 2003, data for this cross-sectional study were obtained by reviewing the clinic records of 3,206 women, aged 50 years or older, screened for osteoporosis at 2 urban centres located in Brooklyn, New York. A standardized data entry sheet was utilized to collect data on patient demographics, and risk factors for osteoporosis. At both sites, bone mineral densities were measured by dual x-ray absorptiometry (DXA) using Hologic® QDR4200. T-scores of non-Caucasian patients were reported by comparing BMD against race-appropriate normative data-bases to adjust for the effect of racial differences. T-scores, as defined by the World Health Organization, is the number of standard deviations above or below the mean BMD value in young adults of the same sex and race [27].
Low BMD was defined as T-score at the lumbar spine <-1. Data was analyzed using SPSS® version 11.5. In comparing two groups for continuous variables, analysis of variance with adjustment for multiple measures employing Benferoni's technique was used with results presented as the mean ± SEM. After adjusting for age, family history of osteoporosis, maternal fracture, smoking, steroid use and sedentary lifestyle, the effect of race on interaction of BMI and BMD was assessed by logistic regression method.
Results
A total of 3,206 women were screened for postmenopausal osteoporosis with their clinic records subsequently reviewed (Table 1). 2,417 (75.4%) were African Americans (AA), 441(13.6%) were Whites, and 348 (10.9%) were Hispanics, reflecting the predominance of AA as the main ethnic group in the community examined. Mean age (± SEM) for the entire cohort was 58.3 ± 0.24. White women were significantly older with a mean age (± SEM) of 60.4 ± 0.62 years while Hispanic women were the youngest at age 56.7 ± 0.67 (p < 0.01). The mean BMI (± SEM) was 30.7 ± 0.87 kg/m2; indicating that overall, the group examined was mildly obese. There was no significant difference in BMI among the three groups (p = 0.38). AA women had significantly higher T-scores and BMD (p < 0.01) at both the lumbar spine and hip when compared to the other groups. After adjusting for age, family history of osteoporosis, maternal fracture, smoking, and sedentary lifestyle, AA women had significantly lower odds of having low BMD compared to Whites [Odds ratio (OR) = 0.079 (0.03–0.24) (95% CI), p < 0.01]. The odds ratio of low BMD was not significant between White and Hispanic women (Figure 1). On examining the race-dependent effect of an increasing BMI on BMD, for every unit increase in BMI, White women had lower odds of having low BMD [OR = 0.9 (0.87–0.94), p < 0.01] while AA women had higher odds of having low BMD [OR = 1.015 (1.007–1.14), p < 0.01] when compared to Whites (Figure 2). This effect was not observed when Hispanic women were compared to Whites.
Table 1 Demographic characteristics, T-scores, and BMD by race
Whites African Americans Hispanics Total P*
N 441 (13.6%) 2,417 (75.4%) 348 (10.9%) 3,206
Age (years) 60.4 ± 0.62 58.5 ± 0.28 56.7 ± 0.67 58.6 ± 0.24 <0.01
BMI (kg/m2) 27.9 ± 0.30 31.3 ± 1.16 30.2 ± 0.41 30.7 ± 0.87 NS
T-score
Hip -1.4 ± 0.06 -0.8 ± 0.04 -0.9 ± 0.06 -0.9 ± 0.03 <0.01
Lumbar spine -1.3 ± 0.02 -1.4 ± 0.01 -1.5 ± 0.07 -1.4 ± 0.07 NS
BMD (gm/cm2)
Hip 0.777 ± 0.07 0.923 ± 0.01 0.845 ± 0.01 0.899 ± 0.04 <0.01
Lumbar spine 0.905 ± 0.01 0.988 ± 0.01 0.880 ± 0.01 0.965 ± 0.03 <0.01
Figure 1 Odds ratio of Low BMD in African American, Hispanic and White Women
Figure 2 Effect of Race on the Interaction between BMI and BMD
Using general linear model (GLM), we also assessed BMI, Race and BMI × Race as factors for BMD. Significant main effect was found for BMI (p < 0.01), Race (p < 0.01) and BMI × Race (p < 0.01), indicating that the relationship between BMI and BMD was different between the races.
To further examine the linear relationship between BMI and BMD for each race (figure 3), regression parameters were estimated for each race with specific contrasts between each pair. Significant differences were found between Whites and AA women, Hispanic and AA women, but not between Whites and Hispanics (p = 0.8), (Figure 3).
Figure 3 Effect of Race and Obesity on BMD. Bone mineral density (BMD) as it relates to body mass index (BMI). Note the different slopes for each ethnic group. Slopes were significantly different for African American (AA) women and White women p = 0.001. There was no significant difference between Hispanic and White women p = 0.8.
Discussion
Our study is the first to investigate the effect of race on the interaction of BMI and BMD. The AA women in our study had significantly lower odds of having low BMD, a finding consistent with the results of NORA study [7] and with previous reports that AA women begin menopause with higher BMD and have lower rates of bone loss after menopause [25,26]. On the other hand, Hispanic women were shown to have the same risk of having low BMD as the White women in the study. Although our findings are at variance with the results from NORA study which showed increased odds for osteoporosis for Hispanic women compared with White women, they more closely reflect the observations of similar fracture risk in Hispanic and White women in NORA study [7].
As anticipated, our study demonstrated that in white women, an increasing BMI was associated with slightly lower odds of having a low BMD. This BMD protective effect of increasing BMI and obesity is similar to the findings of other studies including the NORA study [7]. However, when race was introduced as an independent variable to examine its effect on BMD, we were surprised to find that among the AA women, an increase in BMI was associated with a slight albeit significant increase in the odds of having a low BMD. This effect was not observed when Hispanic women were compared to whites. This finding was both unexpected and counterintuitive as an increasing BMI has generally been thought to result in higher BMD. In the NORA study which included more than 18000 minority women, no such correlation between the BMI and BMD was observed in the African American population [7]. However, NORA study also had a predominant Caucasian population (89.7%) and African American women accounted for only 3.9% of study cohort, unlike our study which had predominant African American population.
It is unclear why the protective effect of an increasing BMI is lost in AA women as the present study has not been designed to answer this question. However, the results of our study highlight the need for further studies in this area to confirm our findings and to determine the possible mechanisms that underlie such race-dependent effect of BMI on BMD. Furthermore, previously established risk factors for postmenopausal osteoporosis should be applied with caution to non-white populations as these risk factors were ascertained from studies in which women from minority groups have been underrepresented. As such, studies to establish risk factors for different ethnic groups should be undertaken to attain a more representative estimation of osteoporosis risk that hopefully will lead to an individualized treatment based on ethnic differences.
Conclusion
Interaction of the BMI and BMD is complex and race is an important modifier of this interaction. Though AA women are less likely to have low BMD than White women, the increase in BMD with BMI seen in Caucasians was not seen in the African Americans. More studies are needed to ascertain why the protective effect of an increasing BMI is lost in AA women and to determine the possible mechanisms that underlie such race-dependent effect of BMI on BMD. Furthermore, previously established risk factors for postmenopausal osteoporosis should be applied to non-white populations with caution as these risk factors were ascertained from studies in which women from minority groups have been underrepresented.
Abbreviations
BMI = body mass index; BMD = bone mineral density; DXA = dual x-ray absorptiometry, OR = odds ratio, CI = confidence interval, AA = African American
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JPC, SA and JN participated in literature search and manuscript preparation, LJ, JS and GB carried out literature search and data collection, JS, IR, VP, LC, LP, SC, NG, PG and RM participated in the data collection and management. HG performed statistical analysis and participated in data presentation, SIM conceived of the study, participated in its design, coordination and statistical analysis, helped to draft the manuscript and in final presentation. All authors read and approved the final manuscript.
Acknowledgements
This study was supported by a grant from Procter and Gamble pharmaceuticals and by grants from the NIH, K12HD043428, BIRCWH to, JJS, JN and SIM and also by grant support for the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medications (DREAM) study from the Population Health Research Institute (PHRI) at McMaster University (Hamilton, Ontario, Canada) to JN and SIM. This work was presented in an abstract form and won the the Young Investigator Award (to Jonathan Castro) from International Society for Clinical Densitometry, January 2004, Miami Florida.
==== Refs
Consensus development conference: diagnosis, prophylaxis, and treatment of osteoporosis Am J Med 1993 94 646 650 8506892 10.1016/0002-9343(93)90218-E
Fulton JP New guidelines for the prevention and treatment of osteoporosis. National Osteoporosis Foundation Med Health R I 1999 82 110 111 10193144
Melton LJ Thamer M Ray NF Chan JK Chesnut CH Einhorn TA Johnston CC Raisz LG Silverman SL Siris ES Fractures attributable to osteoporosis: report from the National Osteoporosis Foundation J Bone Miner Res 1997 12 16 23 9240721
Riggs BL Melton LJ The worldwide problem of osteoporosis: insights afforded by epidemiology Bone 1995 17 505S 511S 8573428 10.1016/8756-3282(95)00258-4
Ray NF Chan JK Thamer M Melton LJ Medical expenditures for the treatment of osteoporotic fractures in the United States in 1995: report from the National Osteoporosis Foundation J Bone Miner Res 1997 12 24 35 9240722
Looker AC Wahner HW Dunn WL Calvo MS Harris TB Heyse SP Johnston CCJ Lindsay R Updated data on proximal femur bone mineral levels of US adults Osteoporos Int 1998 8 468 489 9850356 10.1007/s001980050093
Siris ES Miller PD Barrett-Connor E Faulkner KG Wehren LE Abbott TA Berger ML Santora AC Sherwood LM Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women: results from the National Osteoporosis Risk Assessment Jama 2001 286 2815 2822 11735756 10.1001/jama.286.22.2815
Melton LJ Epidemiology of osteoporosis: predicting who is at risk Ann N Y Acad Sci 1990 592 295 306; discussion 334-45 2197948
Kirchengast S Peterson B Hauser G Knogler W [Significance of body weight status for bone density in the elderly and very old] Z Gerontol Geriatr 2001 34 313 318 11584716 10.1007/s003910170055
Valtuena Martinez S [Obesity and osteoporosis: effect of weight variation on bone mass] Nutr Hosp 2002 17 Suppl 1 49 54 11928536
Kirchengast S Knogler W Hauser G Protective effect of moderate overweight on bone density of the hip joint in elderly and old Austrians Anthropol Anz 2002 60 187 197 12161964
Cifuentes M Johnson MA Lewis RD Heymsfield SB Chowdhury HA Modlesky CM Shapses SA Bone turnover and body weight relationships differ in normal-weight compared with heavier postmenopausal women Osteoporos Int 2003 14 116 122 12730781
Rico H Arribas I Casanova FJ Duce AM Hernandez ER Cortes-Prieto J Bone mass, bone metabolism, gonadal status and body mass index Osteoporos Int 2002 13 379 387 12086348 10.1007/s001980200043
Felson DT Zhang Y Hannan MT Anderson JJ Effects of weight and body mass index on bone mineral density in men and women: the Framingham study J Bone Miner Res 1993 8 567 573 8511983
Murillo-Uribe A Aranda-Gallegos JE Rio de la Loza-Cava MF Ortiz-Luna G Mendoza-Torres LJ Santos-Gonzalez J [Relation between body mass index and bone mineral density in a sample population of Mexican women] Ginecol Obstet Mex 1998 66 267 271 9737066
Turner LW Kendrick O Perry BA Study classifies low body mass index as risk factor for osteoporotic fracture among older U.S. women J Ark Med Soc 1999 96 138 141 10500379
Albala C Yanez M Devoto E Sostin C Zeballos L Santos JL Obesity as a protective factor for postmenopausal osteoporosis Int J Obes Relat Metab Disord 1996 20 1027 1032 8923160
Ribot C Tremollieres F Pouilles JM Bonneu M Germain F Louvet JP Obesity and postmenopausal bone loss: the influence of obesity on vertebral density and bone turnover in postmenopausal women Bone 1987 8 327 331 2966634 10.1016/8756-3282(87)90062-7
Ribot C Tremollieres F Pouilles JM The effect of obesity on postmenopausal bone loss and the risk of osteoporosis Adv Nutr Res 1994 9 257 271 7747670
Tremollieres FA Pouilles JM Ribot C Vertebral postmenopausal bone loss is reduced in overweight women: a longitudinal study in 155 early postmenopausal women J Clin Endocrinol Metab 1993 77 683 686 8370689 10.1210/jc.77.3.683
Ravn P Cizza G Bjarnason NH Thompson D Daley M Wasnich RD McClung M Hosking D Yates AJ Christiansen C Low body mass index is an important risk factor for low bone mass and increased bone loss in early postmenopausal women. Early Postmenopausal Intervention Cohort (EPIC) study group J Bone Miner Res 1999 14 1622 1627 10469292
Anderson JJ Pollitzer WS Ethnic and genetic differences in susceptibility to osteoporotic fractures Adv Nutr Res 1994 9 129 149 7747663
Silverman SL Madison RE Decreased incidence of hip fracture in Hispanics, Asians, and blacks: California Hospital Discharge Data Am J Public Health 1988 78 1482 1483 3177728
Bohannon AD Osteoporosis and African American women J Womens Health Gend Based Med 1999 8 609 615 10839646
Kessenich CR Osteoporosis and African-American women Womens Health Issues 2000 10 300 304 11077212 10.1016/S1049-3867(00)00065-7
Escalante A del Rincon I Epidemiology and impact of rheumatic disorders in the United States Hispanic population Curr Opin Rheumatol 2001 13 104 110 11224734 10.1097/00002281-200103000-00003
Genant HK Cooper C Poor G Reid I Ehrlich G Kanis J Nordin BE Barrett-Connor E Black D Bonjour JP Dawson-Hughes B Delmas PD Dequeker J Ragi Eis S Gennari C Johnell O Johnston CCJ Lau EM Liberman UA Lindsay R Martin TJ Masri B Mautalen CA Meunier PJ Khaltaev N Interim report and recommendations of the World Health Organization Task-Force for Osteoporosis Osteoporos Int 1999 10 259 264 10692972 10.1007/s001980050224
| 15817133 | PMC1090614 | CC BY | 2021-01-04 16:37:46 | no | Nutr Metab (Lond). 2005 Apr 7; 2:9 | utf-8 | Nutr Metab (Lond) | 2,005 | 10.1186/1743-7075-2-9 | oa_comm |
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-131584016910.1186/1477-7827-3-13ResearchModulation of MAA-induced apoptosis in male germ cells: role of Sertoli cell P/Q-type calcium channels Barone Fortunata [email protected] Salvatore [email protected]'Agostino Angela [email protected] Biotechnology Unit Casaccia Research Center, ENEA, 00060 Rome, Italy2 Department of Histology and Medical Embryology, University of Rome "La Sapienza", 00161 Rome, Italy2005 19 4 2005 3 13 13 24 1 2005 19 4 2005 Copyright © 2005 Barone et al; licensee BioMed Central Ltd.2005Barone 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.
Spontaneous germ cell death by apoptosis occurs during normal spermatogenesis in mammals and is thought to play a role in the physiological mechanism limiting the clonal expansion of such cell population in the male gonad. In the prepubertal rat testis, the most conspicuous dying cells are pachytene spermatocytes, which are also the primary target of the apoptosis experimentally induced by the methoxyacetic acid (MAA). Since we have recently reported that Sertoli cells, the somatic component of the seminiferous epithelium, regulate not only germ cell viability and differentiation but also their death, we have further investigated the mechanism involved in such a control.
In this paper we have used the protein clusterin, produced by Sertoli cells and associated with tissue damage or injury, as indicator of germ cell apoptosis in rat seminiferous tubules treated with MAA in the presence or in the absence of omega-agatoxin, a specific inhibitor of P/Q type voltage-operated calcium channels (VOCC's). We performed both a qualitative analysis of clusterin content and germ cell apoptosis by immunofluorescence experiments and a quantitative analysis by in situ end labelling of apoptotic germ cells followed by flow cytometry. The results obtained demonstrate that Sertoli cells modulate germ cell apoptosis induced by methoxyacetic acid also throughout the P/Q-type VOCC's.
==== Body
Background
Spermatogenesis in mammals is a complex process dependent upon the hormonal stimulation as well as the dynamic interactions between the Sertoli cells, the somatic component of the seminiferous epithelium, and the germ cells. In fact, in the mammalian testis germ cells, differentiating from spermatogonia to mature spermatozoa, are in close contact with Sertoli cells which supply the nutrients and the hormonal signals essential for successful spermatogenesis. Spontaneous death of germ cells occurs normally during spermatogenesis by the process of apoptosis [1], leading to a loss of up to 75% of the potential number of spermatozoa [2], probably as a physiological mechanism limiting the clonal expansion of germ cells and the spermatozoa release. Such a process has stimulated a number of studies in vivo and in vitro by investigators working in the area of male reproductive endocrinology and toxicology [3-5], with the aim to define the cellular and molecular mechanisms of both spontaneous and toxicant-induced germ cell apoptosis. Among the most commonly toxicants inducing injuries of the male reproductive system [6-8] the 2-methoxyethanol glycol (2-ME), a major bioproduct of the paint industry, causes severe testicular lesions in many mammalian species, including men [9]. By using the methoxyacetic acid (MAA), the proximate toxic metabolite of the 2-ME, germ cell death can be induced in vitro in seminiferous tubule cultures [10]. In 18–21 days old rats, that have not yet completed the spermatogenetic process, the MAA-induced cell death concerns a large proportion of pachytene spermatocytes. We have recently shown that such MAA-induced apoptosis is significantly prevented by co-treatment with nifedipine and ω-conotoxin [11], which block, respectively, L-type and N-type voltage-operated calcium channels (VOCC's). Ca++ channels in many different cell types activate upon membrane depolarisation and mediate Ca++ influx in response to action potentials and sub-threshold depolarising signals. Ca++ entering the cells through VOCC's serves as second messenger of electric signalling, initiating intracellular events such as contraction, secretion, synaptic transmission, and gene expression.
L-type and N-type VOCC's are present on the Sertoli cell plasma membrane [12] and mainly localised at the level of contact surface between Sertoli cells and pachytene spermatocytes in the adluminal compartment of the seminiferous epithelium [12]. Such calcium channels are responsible for the substantial Ca++ influx in rat Sertoli cells [13] and play a role in laminin-dependent [Ca++]i raise in Sertoli cells and in Sertoli cell secretory process [14,15].
Besides L-and N-type VOCC's, on Sertoli cell plasma membrane is present a third type of Ca2+ channels, the P/Q type [12], specifically blocked by the ω-agatoxin IVA, a peptide isolated from the venom of Agenolanopsis aperta. These channels are localised in the plasma membranes of Sertoli cells adjacent to the basal lamina, at the level of the blood-testis barrier, which selects the substances that can reach the lumen of the seminiferous tubule. P/Q type channels have been shown in neurons where they are primarily responsible for Ca2+ entry and subsequent release of fast neurotransmitters at synapses [18]. Furthermore, several endocrine cell types express these particular Ca2+ channels, still often believed to be expressed only in the nervous system. Pancreatic cells [19], pituitary cells [20], adrenal medullary chromaffin cells [21], as well assmall cell lung carcinoma cells [22] and insulin secreting cell lines [23] express P/Q type Ca2+ channels that participate in the control of their secretory activity.
Clusterin is a ubiquitously expressed heterodimeric glycoprotein that is the major protein produced by cultured rat Sertoli cells [16]. A common theme found in several tissues is the association of clusterin with tissue damage or injury. It has been shown [11,17] that, in MAA-induced apoptosis, Sertoli cell-derived clusterin is very early accumulated in the cytoplasm of dying germ cells at a specific stage of differentiation, i.e. pachytene spermatocytes.
In the present study we used the protein clusterin as a marker of apoptosis to verify whether Sertoli cell P/Q-type VOCC's are involved in the modulation of MAA-induced germ cell death. The presence of VOCC's in the seminiferous epithelium only on Sertoli cell plasma membrane and not on germ cells is very intriguing. Which is the function of such channels in spermatogenesis?
Because of their role in mediating exocytosis and of their peculiar localisation at the level of blood-testis barrier, we would investigate whether P/Q type channels, as well as L- and N-types, were involved in the apoptotic process of germ cells, modulating Sertoli cell response to MAA injury, clusterin secretion and accumulation in pachytene spermatocytes and consequently germ cell death.
Methods
Chemicals
ω-agatoxin was purchased from Peptide Institute, Louisville, KY. ω-conotoxin GVIA was purchased by Bachem, UK. Other chemicals, unless otherwise specified, were of the purest grade available from Sigma (St Louis, MO).
Animals
Male Wistar rats were used in all experiments. Animals were housed in accordance with the guidelines for animal care of University of Rome "La Sapienza" and were killed humanely by asphyxiation with CO2 before organ removal. The testes were excised from immature Wistar rats aged from 18 to 21 days, where the spermatogenesis is still uncompleted, washed and immediately processed after excision as described below.
Sertoli cell in vitro cultures
Sertoli cell monolayers were prepared from the testes of 18–21-day-old Wistar rats according to previously described methods [11,12]. Briefly, after excision, testes were mechanically decapsulated, reduced in small fragments and resuspended in equal volume of Eagle's Minimum Essential Medium (MEM) (Life Technologies, Inc, Grand Islands, NY). Fragments were digested with 0.25 % trypsin (Difco Laboratories, Detroit, MI) 10 μg/ml DNAase I (Roche Molecular Biochemicals, Mannheim, Germany) for 30 min at 32°C, then washed with Hank's Balanced Salt Solutions (HBSS) (Sigma, St Louis MO) and digested with HBSS supplemented with 0.1 % collagenase A plus 10 μg/ml DNAase I (Roche Molecular Biochemicals, Mannheim, Germany) to remove interstitial tissue and peritubular cells. The cell suspension was then centrifuged at 15 × g for 5 min. Pellets were resuspended 1:10 in serum-free MEM with Earle's salt's (Life Technologies, Inc, Grand Islands, NY). Cells were then plated in 3.5 cm Petri dishes and incubated at 32°C in a controlled atmosphere of 95 % air and 5 % CO2. On day 3 of culture, cell monolayers were subjected to hypotonic treatment [14] to completely remove endogenous germ cells and were allowed to recover for 24 hours. Experiments were performed on day 4 of culture.
Treatment with MAA
A 500 mM stock solution of MAA >97 % pure (Fluka, CH9471 Buchs SG, Schweiz) was prepared by dilution in culture medium and the pH was adjusted to 7.4 with 1 N NaOH. This MAA stock was then diluted to a final concentration of 5 mM in the culture medium.
Treatment with calcium channel blockers
Stock solutions were prepared with appropriate vehicles (H2O, 100 % ethanol or 100 % DMSO) according to their solubility. The final concentrations of ethanol and DMSO in the cultures were ≤0.1 %. Vehicle-only controls were included in all experiments. No differences between the medium-only controls and the vehicle-only controls was observed during the experiments.
Seminiferous tubule in vitro cultures
Testes from 18–21 day-old rats were decapsulated and digested under gentle shaking at room temperature in MEM containing 1 mg/ml collagenase A (Roche Molecular Biochemicals, Mannheim, Germany). After dispersion of the interstitium the tubular mass was rinsed in MEM, then stretches of tubules were dissected by means of sharp needles and carefully transferred to 3.5 cm culture dishes in 300 μl of medium. The tubules were incubated for 10 min at 32°C in a humidified chamber under an atmosphere containing 5 % CO2. At the end of the incubation time, the medium was replaced by 600 μl of medium to be exposed to different treatments.
SDS-PAGE and immunoblotting
Cultured Sertoli cells were scraped off the culture dishes, pelleted and washed with PBS. Pellets were resuspended in lysis buffer (150 mM NaCl, 1 % SDS, 20 mM Tris-HCl, 1 mM EDTA, mammalian Protease Inhibitor Cocktail (Sigma, St Louis, MO), pH 8), and incubated 10 min at room temperature. Samples were then centrifuged at 10, 000Xg for 10 min. Samples containing 20–40 μg protein were loaded onto 10 % Tris-Glycine Bis-Acrylamide gels, run under reducing conditions and blotted onto 0.45 μm nitrocellulose membrane (Novex). After an overnight block with 5 % (v/v) skimmed milk in PBS, the blots were immunostained for 1.5 hr with goat polyclonal antibody raised against a peptide mapping at the carboxy terminus of clusterin of mouse origin (500 ng/ml) (Santa Cruz Biotechnology Inc.). Following incubation in alkaline phosphatase-conjugated rabbit anti-goat IgG (Sigma, St Louis, MO) (1:1000), immunoreactive bands were detected by enhanced chemiluminescence methodology using ECL-Plus, according to manufacturer's instructions (Amersham Biosciences Europe GmbH). All antibodies were diluted in 2 % skimmed milk in PBS (v/v) and the blot was washed extensively with PBS between each stage. Immunoblots were scanned and digitised images were quantitatively analysed by densitometry with AIDA 2.0 program.
Immunoprecipitation experiments
Immunoprecipitation experiments were carried out to examine whether the VOCC's block of MAA- treated cultured Sertoli cells could modulate clusterin secretion. Sertoli cells were labelled in culture with 50 μCi/ml of 35S-methionine (NEN Life Science Products, spec. activ. 1175 Ci/mmol) for 6 hrs in the presence of 5 mM MAA plus the ω-agatoxin (1 μM). At the end of the incubation, the media were removed and processed as described. For immunoprecipitation, we used protein-GSepharose beads (Amersham Pharmacia Biotech) which were first resuspended in RIPA buffer (150 mM NaCl, 0.1 % SDS, 0.5% sodium deoxycholate, and 1 % NP40), washed twice and then incubated for 3 hr with a goat polyclonal antibody raised against a peptide mapping at the carboxy terminus of clusterin of mouse origin (500 ng/ml) (Santa Cruz Biotechnology Inc.), which was used for immunocytochemistry, as well. Following several washes with RIPA buffer, beads were incubated with Sertoli cell culture media overnight at 4°C with continuous agitation. After 1000Xg centrifugation for 1 min and supernatant removal, pellets were washed several times and then resuspended in Laemmli sample buffer (20 % glycerol; 10 % β-mercaptoethanol; 5 % SDS; 0.2 M Tris-HCl, pH 6.8; and 0.4 % bromophenol blue). The samples were boiled for 5 min and run on a 12% SDS-polyacrylamide gel followed by fluorography. Digitised images were quantitatively analysed by densitometry with AIDA 2.0 program.
Immunolocalisation of clusterin in seminiferous tubule cryosections
The immunohistochemical localization of clusterin protein was performed in cryosections (6 μm) of seminiferous tubules cultured in the presence of MAA, MAA plus ω-agatoxyn or withouth any substance added (control). Slides were first fixed with 4 % paraformaldehyde for 10 min at 4°C, washed twice with PBS and pre-incubated with PBS plus 4 % BSA for 20 min. at room temperature. Sections were then treated with 0.1 % Triton and 0.1 % sodium citrate on ice for 5 min and stained with a goat polyclonal antibody raised against a peptide mapping at the carboxy terminus of clusterin of mouse origin (500 ng/ml) (Santa Cruz Biotechnology Inc.), for 60 min at 4°C in a close, humidified chamber. Then, the cryosections were washed twice with PBS plus 1% BSA and incubated with a FITC-conjugated rabbit anti-goat secondary antibody (Sigma, St Louis, MO), washed with PBS and photographed by a Zeiss fluorescence microscope with an Axiocam imaging system.
Apoptosis detection of germ cells in seminiferous tubule cryosections
Apoptosis was detected in seminiferous tubule cryosections (6 μm) by using an apoptosis detection system, the In Situ Cell Death Detection Kit (Roche Molecular Biochemicals, Mannheim, Germany). The system end-labels the fragmented DNA of apoptotic germ cells using modified terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assay. It consists in the incorporation of biotynylated nucleotides at the 3'-OH DNA ends, using the enzyme terminal deoxynucleotidyl transferase (TdT). The assay was performed according to manufacturer's instructions. Briefly, sections were rehydrated, washed in PBS and incubated at 37°C for 1 hr with a TUNEL reaction mixture. This allows direct detection of fragmented DNA in apoptotic cells with fluorescein-12-dUTP by means of the enzyme TdT. Negative controls were processed in an identical manner, except that TdT was omitted. As positive controls, sections from control seminiferous tubules were treated with DNAase I for 10 min before performing the end labeling reaction. After the necessary washes, sections were covered with glycerol and glass coverslips.
Flow Cytometry
For quantitative analysis of germ cells apoptosis levels, cellular suspension prepared from seminiferous tubules was analyzed by flow cytometry using an apoptosis detection system (Roche Molecular Biochemicals, Mannheim, Germany). Seminiferous tubules cultured cultured in the presence of MAA, MAA plus ω-agatoxyn or withouth any substance added (control) were digested with 0.25 % trypsin and 0.02 % EDTA and testicular cell suspension was permeabilized with a permeabilization solution (0.1 % Triton 100X in 0.1 % sodium citrate) for 2 min on ice. Then cells were resuspended in 50 μl of TUNEL reaction mixture or label solution as negative control for 60 min at 37°C, according to manufacturer's instructions. The cells were resuspended in a final volume of 500 μl in PBS and analyzed with a Coulter Epics XL flow cytometer (Beckman Coulter).
For detection of clusterin levels, cell suspension from seminiferous epithelium was washed once with cold PBS plus 1 % BSA before incubation with antibodies. Then, the cells were fixed with 1 % cold paraformaldehyde for 5 min, whereas 0.1 % Triton and 0.1 % sodium citrate in PBS was used for permeabilization. For detection of intracellular clusterin, samples were then incubated overnight at 4°C with a goat polyclonal antibody raised against a peptide mapping at the carboxy terminus of clusterin of mouse origin (500 ng/ml) (Santa Cruz Biotechnology Inc.). Thereafter, cells were washed twice with PBS plus 1 % BSA and incubated 30 min in ice with a FITC-conjugated rabbit anti-goat secondary antibody, washed twice with PBS plus 1% BSA, and analysed. Cells were gated by using forward versus side scatter to exclude dead cells and debris. Fluorescence of 104 cells/sample was acquired in logarithmic mode for visual inspection of the distributions, and in linear mode to quantify the expression of the relevant molecules by calculating the mean fluorescence intensity. Besides the morphological analysis, DNA was stained with propidium iodide for ploidy analysis to evaluate the purity of each germ cell fraction, as previously described [24].
Statistics
Student's t test was used for statistical comparison between means where applicable. The statistic software used was the SPSS version 9.0 (Chicago, Illinois).
Results
Effects of MAA and ω-agatoxin in rat Sertoli cells in culture
In order to use the clusterin as marker of P/Q-type VOCC's-modulated apoptosis it was necessary to verify whether intracellular clusterin level, which is stimulated by MAA in cultured rat Sertoli cells [17,11], was modulated by P/Q-type Ca2+ channels. With the aim to answer this question, Sertoli cells from 18–21 day-old Wistar rats were cultured for 6 hrs with any drug added (control) or in the presence of 5 mM MAA and of 5 mM MAA plus the specific inhibitor of the P/Q-type Ca2+ channels, the ω-agatoxin (1 μM). The samples were analysed by SDS-PAGE followed by Western blot and immunostaining with polyclonal antibodies raised against clusterin. Fig. 1 shows that the MAA-increased amount of clusterin was considerably reduced in the samples treated with MAA plus ω-agatoxin. Fig. 2 shows a modulation also of the secreted clusterin in the culture medium: Sertoli cells, cultured in the presence of MAA, secrete a major amount of clusterin respect to the control; the block of the P/Q channels by ω-agatoxin reduces such a secretion to the basal values.
Figure 1 (Top panel) Western blotting of rat Sertoli cell lysates treated with polyclonal antibodies raised against clusterin. Representative data of n = 3 independent experiments. (Lower panel) Quantification of immunoblot by densitometry. The amount of 40-kDa was quantified by integrating the volume as described under "Methods". Results are expressed as mean +/- SEM of three determinations. 1. control; 2. MAA; 3. MAA plus ω-agatoxin (p < 0.001).
Figure 2 (Top panel) SDS-PAGE of immunoprecipitated culture medium from rat Sertoli cells labelled with 35S-methionine. Representative data of n = 3 independent experiments. (Lower panel) Quantification of fluorography by densitometry. Results are expressed as mean +/- SEM of three determinations. 1. control; 2. MAA; 3. MAA plus ω-agatoxin (p < 0.001).
Effects of MAA and ω-agatoxin on germ cell apoptosis in cultured rat seminiferous tubules
With the aim to study the effect of MAA and ω-agatoxin on the apoptosis in the seminiferous epithelium, seminiferous tubules were cultured for 6 and 12 hrs in the presence of MAA (5 mM), and MAA plus ω-agatoxin (1 μM). The controls were performed in absence of any drug added to the culture medium. The time of 6 hrs was chosen because is the minimum time required to identify the clusterin in the cytoplasm of pachytene spermatocytes [11,17], while 12 hrs is the minimum time necessary, in our experimental conditions, to detect apoptotic cells in cultured seminiferous tubules.
At the end of the culture an aliquot of seminiferous tubules were immunostained with antibodies raised against clusterin or processed by TUNEL, in order to visualise, respectively, the effect of MAA and ω-agatoxin on clusterin localisation and on germ cell apoptosis in histologic cryosections. The remaining aliquot of seminiferous tubules was trypsinized in order to quantitatively measure the clusterin content and the apoptosis by the flow cytometry analysis. The clusterin content was evaluated after immunostaining of cells with antibodies raised against clusterin while the apoptosis was determined by TUNEL.
The immunofluorescence analysis of the histologic cryosections(Fig. 3) shows that, following 6 hrs of culture, the treatment with MAA not only increases the content of clusterin in Sertoli cells (double arrows) but evidences the presence of clusterin also in germ cells (single arrows) which is reduced in the presence of ω-agatoxin. In parallel, the quantitative flow cytometry analysis confirmed these results. Following 6 hrs of culture, clusterin content is increased in MAA-treated seminiferous tubules (grey histogram 2) and it is reduced in seminiferous tubules treated with MAA plus the ω-agatoxin (grey histogram 3). Apoptosis is not appreciable (black histograms), in agreement with literature data in vitro and in vivo [10,11], due to the short time of MAA-treatment. In Fig. 4 the immunofluorescence analysis of seminiferous tubules cultured for 12 hrs shows that the apoptosis of pachytene spermatocytes (single arrows) is enhanced in the presence of MAA and is decreased by the VOCC's inhibitor. Fig. 4 also shows the analysis performed by flow cytometry: ω-agatoxin (histograms 3) does decrease the MAA-induced germ cell apoptosis (histogram 2). Fig. 5 shows a comparison among ω-agatoxin, nifedipine and ω-conotoxin effects on MAA-induced apoptosis. The inhibitory effect due to the block of P/Q type channels appears less pronounced (mean +/- SEM: 51 +/- 2.8) than that of L (nifedipine-sensitive) (mean +/- SEM: 30 +/- 1.4)and N (ω-conotoxin-sensitive) (mean +/- SEM: 33.5 +/- 3.5) type VOOC'S. Expressed as percentage ofinhibition of the MAA response, the block due to ω-agatoxin is of 85%, while nifedipine and ω-conotoxin blocks are of 91% and 90 %, respectively.
Figure 3 (Top panel, a) Immunofluorescence analysis of histologic cryosections of seminiferous tubules cultured for 6 h and immunostained with antibodies raised against clusterin. 1. control (Bar 25x); 2. MAA (Bar 10x); 3. MAA plus ω-agatoxin (Bar 25x). Single arrow indicates pachytene spermatocytes; double arrow indicates Sertoli cells. Representative data of n = 3 independent experiments. (Lower panel, b) Quantitative analysis by flow cytometry of testicular cell population prepared from seminiferous tubules cultured for 6 h. Grey bar represents clusterin content (p < 0.001). Black bar represents apoptosis levels determined by TUNEL. Data are expressed as mean +/- SEM of three independent experiments. 1. control; 2. MAA; 3. MAA plus ω-agatoxin.
Figure 4 (Top panel, a) Immunofluorescence analysis of histologic cryosections of seminiferous tubules cultured for 12 h and processed by TUNEL. 1. control (Bar 20x); 2. MAA (Bar 40x); 3. MAA plus ω-agatoxin (Bar 20x). (Lower panel, b) Quantitative analysis by flow cytometry of testicular cell population prepared from seminiferous tubules cultured for 12 h. Bar represents apoptosis levels determined by TUNEL (p < 0.001). Data are expressed as mean +/- SEM of three independent experiments. 1. control; 2. MAA; 3. MAA plus ω-agatoxin.
Figure 5 Comparison of VOCC's inhibitors effects on MAA-induced apoptosis of testicular cell population prepared from seminiferous tubules cultured for 12 h. Histograms represent apoptosis levels determined by TUNEL (p < 0.001). Data are expressed as mean +/- SEM of three independent experiments. 1. control; 2. MAA; 3 MAA plus ω-conotoxin; 4. MAA plus nifedipine; 5. MAA plus ω-agatoxin.
Discussion
A growing body of evidence demonstrates that both spontaneous (during normal spermatogenesis) and increased germ cell death triggered by various regulatory stimuli [25-29] in rats occur via apoptosis. However, the mechanisms by which these proapoptotic stimuli activate germ cell death are not well understood. We have recently proposed a calcium-mediated mechanism underlying spermatocyte apoptosis induced by MAA [11]. It has been shown that in rat seminiferous epithelium Sertoli cells can modulate germ cell apoptosis throughout L-type and N-type VOCC's and that the contacts between Sertoli cells and germ cells are necessary for such a modulation [11].
In the present study we demonstrate that Sertoli cells are involved in MAA-induced germ cell apoptosis also throughout VOCC's blocked by ω-agatoxin, i.e. of P/Q type. The block of such VOCC's inhibits the accumulation of Sertoli cell-produced clusterin in the cytoplasm of germ cells, consequent to the MAA treatment, and prevents pachytene spermatocyte apoptosis. The data presented here, however, show that, from a quantitative point of view, the inhibition of germ cell death is less pronounced respect to that induced by nifedipine and ω-conotoxin, suggesting a synergistic role of P/Q type calcium channels. These results fit well with the peculiar localization of P/Q type calcium channels in the seminiferous epithelium.
The previous identification of P/Q channels on the plasma membrane of Sertoli cells in the zone adjacent to the basal lamina and their involvement in the modulation of protein secretion [12], suggested a role at the level of the blood-testis barrier (BTB), which selects the substances that can reach the lumen of the seminiferous tubule. At the level of BTB the Ca2+ plays a relevant role. It is known that the Sertoli cell barrier can be disrupted and resealed by manipulating [Ca2+] in the culture medium [31]. In the rat, the entire process of germ cell development, except for the early phase of spermatogenesis, is segregated from the systemic circulation by the BTB. This means that whatever substance would have to pass through the Sertoli cell cytoplasm to reach the late meiotic and post-meiotic germ cells. P/Q-type channels should represent the first "sentinel"gates of Sertoli cells and could constitute a first inhibitory mechanism respect to L-and N-type channels, which are localized deeply in the seminiferous epithelium, near the lumen of the tubule. Furthermore, the ascertainment that the block of P/Q-type VOCC's inhibits germ cell apoptosis less than L- and N-type should suggest a gradual, spatial control of germ cell death.
The decrease of intracytoplasmic Sertoli cell clusterin and the inhibition of its accumulation into pachytene spermatocytes, following P/Q channels block, suggest that calcium channels control germ cell apoptosis throughout an indirect mechanism. The protective effect of calcium channel blockers against MAA-induced spermatocyte apoptosis is probably mediated by the inhibition of clusterin transfer from Sertoli cells to germ cells. This hypothesis agrees with recent data from literature which demonstrate that the intracellular form of clusterin could be cytotoxic [32-34]. Clusterin accumulation in germ cells could be a selection mechanism for cells secondarily injured and committed to death. On the other hand, being Sertoli cell the direct target of MAA, the injured cell should be not more able to support a further differentiation of germ cells and should limit the number of spermatocytes. This latter hypothesis agrees with recent results obtained in MAA-treated rats where the expression of both Sertoli cell androgen receptor protein and androgen binding protein are significantly altered [35]. It is known, in fact, that the diminution in androgen receptor is concomitant to a decrease of intratesticular androgen levels and is accompanied by significant germ cell apoptosis [7,36]. Altered expression of ABP in transgenic mice is associated with relevant apoptosis of pachytene spermatocytes [37].
Furthermore, in this paper we demonstrate a role of P/Q-type VOCC's specific for the control of male germ cell death. In other cell types, i.e. cultured cortical neurons, exposed to amyloid β protein in order to induce apoptosis, the block of L-type voltage-sensitive Ca++ channels attenuates neuronal apoptosis while the block of N- and P/Q type channels has no effect [38].
Our work confirm and extend previous data showing that the VOCC's localized exclusively on Sertoli cell plasma membrane modulate Sertoli cell response to injuries that could damage germ cell differentiation. Such intratesticular regulatory mechanism controlling germ cell apoptosis adds a further piece of the puzzle concerning the knowledge of the apoptotic program in the testis [39].
Authors' contributions
FB was responsible for experimental activity. She performed cell cultures, immunofluorescence and flow cytometry analysis, partecipated in the analysis of data and in drafting the manuscript. SA collaborated in the immunofluorescence analysis of seminiferous tubule cryosections and provided valuable suggestions during writing the manuscript. AD was responsible for design and coordination of the study. She analyzed the data and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work has been supported by a grant from MIUR 2002 and by contributes of Agilent Technologies and Fondazione Cassa di Risparmio di Roma.
==== Refs
Blanco-Rodriguez J Martinez-Garcia C Spontaneous germ cell death in the seminiferous tubule of the adult rat testis: assessment of the germ cell types that exhibit the capability to undergo apoptosis during spermatogenesis Cell proliferation 1996 29 13 31 8603107 10.1046/j.1365-2184.1996.d01-5.x
de Rooij DG Janssen JM Regulation of the density of spermatogonia in the seminiferous epithelium of the chinese hamster:I. Undifferentiated spermatogonia Anat Rec 1987 217 124 130 3578832 10.1002/ar.1092170203
Darzynkiewicz Z Juan G Li JX Gorczyca W Murakami T Taganos F Cytometry in cell necrobiology: analysis of apoptosis and accidental cell death (necrosis) Cytometry 1997 27 1 4 9000580 10.1002/(SICI)1097-0320(19970101)27:1<1::AID-CYTO2>3.0.CO;2-L
Krishnamurthy H Weinbauer GF Aslam H -Hei Yeung C Niesclag E Quantification of apoptotic testicular germ cells in normal and methoxyacetic acid-treated mice as determined by flow cytometry J Androl 1998 19 710 717 9876022
Linder RE Klinefelter GR Strader LF Suarez JD Roberts NL Dyer CJ Spermatotoxicity of dibromoacetic acid in rats after 14 daily exposure Reprod Toxicol 1994 8 251 259 8075514 10.1016/0890-6238(94)90009-4
Blount BC Silva MJ Claudilli SP Needham LL Pirkle JL Sampson EJ Lucier GW Jackson RJ Brock JW Levels of seven urinary phthalate metabolites in a human reference population Environ Health Perspect 2000 108 979 982 11049818
Hikim AP Wang C Leung A Swerdloff RS Involvement of apoptosis in the induction of germ cell degeneration in adult rats after gonadotropin-releasing hormone antagonist treatment Endocrinology 1995 136 2770 2775 7750502 10.1210/en.136.6.2770
Klinefelter GR Strader LF Suarez JD Roberts NL Bromochloroacetic acid exerts qualitative effects on rat sperm: implications of a novel biomarker Toxicol Sci 2002 68 164 173 12075119 10.1093/toxsci/68.1.164
Hardin BD Reproductive toxicity of the glycol ethers Toxicology 1983 27 91 102 6351353 10.1016/0300-483X(83)90014-8
Li LH Chapin R 2-Metoxyacetic acid (MAA)-induced spermatocyte apoptosis in human and rat testes: An in vitro comparison J Androl 1996 17 538 549 8957698
Barone F Aguanno S D'Alessio A D'Agostino A Sertoli cell modulates MAA-induced germ cell apoptosis throughout voltage-operated calcium channels FASEB J 2004 18 353 354 14656996
Fragale A Aguanno S Kemp M Revees M Price K Beattie R Craig P Volsen P Sher E D'Agostino A Identification and cellular localization of voltage-operated calcium channels in immature rat testis Mol Cell Endocrinol 2000 162 25 33 10854695 10.1016/S0303-7207(00)00213-6
D'Agostino A Menè P Stefanini M Voltage-gated calcium channels in rat Sertoli cells Biol Reprod 1992 46 414 418 1377510
Taranta A Teti A Stefanini M D'Agostino A Immediate cell signal induced by laminin in rat Sertoli cells Matrix Biol 2000 19 11 18 10686421 10.1016/S0945-053X(99)00054-2
Taranta A Morena AR Barbacci E D'Agostino A Omega-Conotoxin-sensitive Ca2+ voltage-gated channels modulate protein secretion in cultured rat Sertoli cells Mol Cell Endocrinol 1997 126 117 123 9089649 10.1016/S0303-7207(96)03973-1
Griswold MD Protein secretion of Sertoli cells Int Rev Cytol 1988 110 133 156 3053498
Clark AM Maguire SM Griswold MD Accumulation of clusterin/sulfated glycoprotein-2 in degenerating pachytene spermatocytes of adult rats treated with methoxyacetic acid Biol Reprod 1997 57 837 846 9314588
Catterall WA Structure and regulation of voltage-gated Ca2+ channels Annu Rev Cell Dev Biol 2000 16 521 555 11031246 10.1146/annurev.cellbio.16.1.521
Davalli AM Biancardi E Pollo A Socci C Pontiroli AE Pozza G Clementi F Sher E Carbone E Dihydropyridine-sensitive and in sensitive voltage-operated calcium channels participate in the control of glucose-induced insulin release from human pancreatic beta cells J Endocrinol 1996 150 195 203 8869586 10.1677/joe.0.1500195
Sher E Clementi F Omega-conotoxin sensitive voltage-operated calcium channels in vertebrate cells Neuroscience 1991 42 301 307 1654532 10.1016/0306-4522(91)90376-Y
Kim SJ Lim W Kim J Contribution of L- and N- type calcium currents to exocytosis in rat adrenal medullary chromaffin cells Brain Res 1995 675 289 296 7796141 10.1016/0006-8993(95)00085-5
Codignola A Tarroni P Clementi F Pollo A Lavallo M Carbone E Sher E Calcium channels subtypes controllino serotonin release from human small-lung carcinoma cell lines J Biol Chem 1993 268 26240 26247 8253745
Magnelli V Pollo A Sher E Carbone E Block of non L-, non N-type Ca2+ channels in rat insulinoma RINm5F cells by ω-agatoxin IVA and ω-conotoxin MVIIC Pflugers Arch 1995 429 762 771 7603830 10.1007/BF00374799
Telford WG King L Fraker PJ Rapid quantitation of apoptosis in pure and heterogeneous cell populations using flow cytometry J Immunol Methods 1994 172 1 16 8207258 10.1016/0022-1759(94)90373-5
Sinha Hikim AP Wang C Leung A Swerdloff RS Involvement of apoptosis in the induction of germ cell degeneration in adult rats after gonadotropin-releasing hormone antagonist treatment Endocrinology 1995 136 2770 2775 7750502 10.1210/en.136.6.2770
Sinha Hikim AP Rajavashisth TB Sinha Hikim I Lue Y Bonavera JJ Leung A Wang C Swerdloff RS Significance of apoptosis in temporal and stage-specific loss of germ cells in the adult rat after gonadotropin deprivation Biol Reprod 1997 57 1193 1201 9369187
Blanco-Rodriguez J Martinez-Garcia C Apoptosis precedes detachment of germ cells from the seminiferous epithelium after hormone supprression by short-term oestradiol treatment of rats Int J Androl 1998 21 109 115 9675620 10.1046/j.1365-2605.1998.00109.x
Lue Y Sinha Hikim AP Wang C Im M Leung A Swerdloff RS Testicular heat exposure enhances the suppression of spermatogenesis by testosterone in rats: the two-hit approach to male contraceptive development Endocrinology 2000 141 1414 1424 10746645 10.1210/en.141.4.1414
Lee J Richburg JH Shipp EB Meistrich ML Boekelheide K The Fas system, a regulator of testicular germ cell apoptosis, is differentially up-regulated in Sertoli cell versus germ cell injury of the testis Endocrinology 1999 140 852 858 9927315 10.1210/en.140.2.852
Fawcett DW Ultrastructure and the function of the Sertoli cell Handbook of Physiology 1975 V Washington DC, Am Physiol Soc, Male Reproductive System 21 55
Grima J Wong CSC Zhu LJ Zong SD Cheng CY Testin secreted by Sertoli cells is associated with the cell surface and its expression is correlated with the disruption of Sertoli cell-germ cell junctions but not the inter-Sertoli tight junctions J Biol Chem 1998 273 21040 21053 9694856 10.1074/jbc.273.33.21040
O'Sullivan J Whyte L Drake J Tenniswood M Alterations in post-translational modification and intracellular trafficking of clusterin in MCF-7 cells during apoptosis Cell Death Differ 2003 10 914 927 12867999 10.1038/sj.cdd.4401254
Criswell T Klokov D Beman M Lavik JP Boothman DA Repression of IR-inducible clusterin expression by the p53 tumor suppressor protein Cancer Biol Ther 2003 2 372 380 14508108
Kaighn ME Vayssiere JL Rincheval V Intracellular clusterin causes juxtanuclear aggregate formation and mitochondrial alteration J Cell Sci 2003 116 3109 3121 12799419 10.1242/jcs.00619
Tirado OM Martinez ED Rodriguez OC Danielsen M Selva DM Reventos J Munell F Suarez-Quian CA Methoxyacetic acid disregulation of androgen receptor and androgen-binding protein expression in adult rat testis Biol Reprod 2002 68 1437 1446 12606434 10.1095/biolreprod.102.004937
Nandi S Banerjee PP Zirkin BR Germ cell apoptosis in the testes of Sprague Dawley rats following testosterone withdrawal by ethane 1,2-dimethanesulfonate administration: relationship to Fas? Biol Reprod 1999 61 70 75 10377033
Selva DM Tirado OM Toran N Suarez-Quian CA Reventos J Munell F Meiotic arrest and germ cell apoptosis in androgen-binding protein transgenic mice Endocrinology 2000 241 1168 1177 10698194 10.1210/en.141.3.1168
Yagami T Ueda K Asakura K Kuroda T Hata S Sakaeda T Kambayashi Y Fujimoto M Effects of endothelin B receptor agonists on amyloid beta protein (25-35)-induced neuronal cell death Brain Res 2002 948 72 80 12383957 10.1016/S0006-8993(02)02951-7
Sinha Hikim AP Lue Y Diaz-Romero M Yen PH Wang C Swerdloff RS Deciphering the pathways of germ cell apoptosis in the testis J Steroid Biochem Mol Biol 2003 85 175 182 12943702 10.1016/S0960-0760(03)00193-6
| 15840169 | PMC1090615 | CC BY | 2021-01-04 16:37:13 | no | Reprod Biol Endocrinol. 2005 Apr 19; 3:13 | utf-8 | Reprod Biol Endocrinol | 2,005 | 10.1186/1477-7827-3-13 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-321581998610.1186/1465-9921-6-32ReviewInterstitial lung disease in children – genetic background and associated phenotypes Hartl Dominik [email protected] Matthias [email protected] Pediatric Pneumology, Childrens' hospital of the Ludwig-Maximilians-University, Munich, Germany2005 8 4 2005 6 1 32 32 9 1 2005 8 4 2005 Copyright © 2005 Hartl and Griese; licensee BioMed Central Ltd.2005Hartl and Griese; 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.
Interstitial lung disease in children represents a group of rare chronic respiratory disorders. There is growing evidence that mutations in the surfactant protein C gene play a role in the pathogenesis of certain forms of pediatric interstitial lung disease. Recently, mutations in the ABCA3 transporter were found as an underlying cause of fatal respiratory failure in neonates without surfactant protein B deficiency. Especially in familiar cases or in children of consanguineous parents, genetic diagnosis provides an useful tool to identify the underlying etiology of interstitial lung disease. The aim of this review is to summarize and to describe in detail the clinical features of hereditary interstitial lung disease in children. The knowledge of gene variants and associated phenotypes is crucial to identify relevant patients in clinical practice.
interstitial lung diseasechildrensurfactant-protein CABCA3mutations
==== Body
Review
Interstitial lung disease in children
Interstitial lung disease (ILD) in children represents a heterogeneous group of rare respiratory disorders characterised by restrictive lung disease and diffuse pulmonary infiltrates. Affected children usually present with dry cough, dyspnoea and tachypnoea. Physical findings show elevated resting respiratory rate and rales (or crackles) mostly more pronounced in the basal segments of the lungs. In children suffering from severe ILD, chronic hypoxemia, cyanosis, finger clubbing and growth failure occur [1,2]. High-resolution computed tomography provides detailed diagnostic information about the extent and distribution of the lung pathology and has become the most relevant imaging technique for ILD in children [3]. However, the diagnostic gold standard is still the lung biopsy where the characteristic histological findings of inflammation in the pulmonary interstitium with wall thickening by inflammatory cells and/or fibrosis can be found. At present, no pathognomonic laboratory criteria for the diagnosis of ILD in children are available [4]. ILD is associated with pulmonary inflammation which can resolve completely, partially or can progress to fibrosis. The factors determining the progression of ILD are not well understood but an interaction between genetic background and environmental modifiers is suggested [1,5-7].
In children, ILD is less frequent and comprises a broader spectrum of disorders with a more variable clinical course compared to adults. In addition, there are special disease entities which predominantly occur in children, like the recently described pulmonary interstitial glycogenosis [8], the neuro-endocrine cell hyperplasia of infancy [9] and the genetic disorders of surfactant metabolism [10]. Most of the information on ILD in childhood is based on studies performed in adults. In the recent years, however, there is increasing evidence that ILD in children differs substantially from adult ILD [2]. PediatricILD occur in the context of lung development and are most frequent diagnosed in the first year of life when alveolar development is taking place [11,12]. The combination of epidemiological rarity, clinical heterogenity and genetic background makes the classification of ILD in children difficult. Thus, studies up to now included only limited numbersof patients, except of one recent study encompassing 185 children [13], and no randomised controlled trials have been performed to evaluate treatment strategies.
Mutations in surfactant protein B (SP-B) cause fatal respiratory distress in neonates [14]. Recently, mutations in the SP-C gene and in the ATP-binding cassette protein A3 (ABCA3) gene were found to be associated with pediatric ILD [15-17]. SP-C gene variants presented as a variable clinical phenotype. While children with SP-C mutations showed mostly non-specific interstitial pneumonitis (NSIP) except of two cases with combined pulmonary alveolar proteinosis (PAP), adults with SP-C mutations suffered from usual interstitial pneumonitis (UIP) or desquamative interstitial pneumonitis (DIP) or remained asymptomatic. ABCA3 mutations were associated with fatal respiratory failure in neonates without SP-B or SP-C deficiency and with non-fatal ILD in one older child. These infants showed a diagnosis of PAP, DIP or chronic pneumonitis of infancy (CPI) [18].
The aim of this review is to summarize and describe in detail the reported clinical courses of hereditary ILD in children. At first, the role of surfactant protein B in ILD is described. It is delineated how SP-B deficiency results in lung disease in mouse models and in humans. Reported single cases with partial SP-B deficiency are discussed in detail. Similarly, the role of surfactant protein C in ILD is reviewed and a model of surfactant protein C associated lung disease is described. Finally, the roles of ABCA3 and of further genes involved in surfactant protein expression, lung morphogenesis and lamellar body synthesis are discussed. The diversity of gene variants and associated phenotypes illustratesthe overlap between genetic defect, environmental influence and the resulting respiratory disease. Knowledge on the clinical features of hereditary ILD is crucial to identify relevant patients for genetic analysis in clinical practice.
Surfactant
The alveolus of the human lung tends to collapse at the end of the expiration due to high surface tension. A phospholipid film, named surfactant, holds alveolar stability at the air-liquid interface by reducing surface tension. Surfactant is a mixture of lipids and proteins, which contains 70 – 80% phospholipids, 10% protein and about 10% of neutral lipids. When surfactant is lacking at birth, the alveoli collapse, the alveolo-capillary membranes become leaky and intra-alveolar hyaline membranes occur, termed as respiratory distress syndrome (RDS). Surfactant is synthesized and secreted by type II epithelialcells. Surfactant proteins can be divided into the hydrophobic proteins SP-B and SP-C and the hydrophilic proteins SP-A and SP-D. While SP-A and SP-D play essential roles in the pulmonary immune defense, SP-B and SP-C lower the surface-tension and facilitate formation and stabilization of the surfactant film [19,20].
Surfactant protein B
Structure and biosynthesis
Human SP-B is encoded on the short arm of chromosome 2 and spans about 10 kilobases. The SP-B gene contains 11 exons, with the 11th exon being untranslated, and is transcribed into a ~2000 base pair mRNA [21-23]. Corticosteriods were found to enhance the SP-B transcription [24], while Transforming Growth Factor-β (TGF-β), NO and insulin inhibited the transcriptional process [25,26]. The SP-B gene is expressed in type II cells and in non-ciliated bronchiolar epithelial cells, but only type II cells are able to fully process the inital translation product, pro-SP-B, to functionally active, mature SP-B. In a first step, the SP-B mRNA is translated into a preproprotein of 381-amino acids [23,27-29]. After removal of the first 23 amino acids, the remaining proprotein (proSP-B) is proteolytically processed at the amino-, and carboxy-terminal ends to the mature SP-B of 79 amino acid. The cysteine protease, cathepsin H and the aspartyl protease, Napsin A, highly expressed in type II epithelial cells, have been implicated recently in processing of the SP-B and SP-C propeptides [30-32].
Surfactant protein B deleted mice
After birth, SP-B deleted mice showed normal respiratory efforts but failed to inflate the lungs, were cyanotic and died rapidly due to severe respiratory failure [33]. Histologically, their lungs showed a normal morphogenesis without signs of inflammation. Pulmonary function testing revealed decreased lung volumesand an absence of residual volume at end-expiration, similar to neonates with RDS. Furthermore, no proSP-B or SP-B protein were detected and no tubular myelin and lamellar bodies were found in lungs of these mice. Interestingly, an aberrant form of proSP-C of 12-kd and abundance of SP-A and SP-C mRNA were detected in these lungs, but fully processed SP-C protein was found to be decreased. Type II alveolar cells of these mice contained notypical lamellar bodies and showed enlarged multivesicularbodies with small lipid vesicles, which suggests a role for SP-B in the packing of surfactantphospholipids into lamellar bodies [34,35]. These findings in SP-B deficient mice indicate thatSP-B is critical for reduction of surface tension in the alveolus, for the proteolytic processing of proSP-C, for the arrangementof surfactant phospholipids in lamellar bodies and for lung functionduring the early neonatal period.
Surfactant protein B deficiency in human lung disease
SP-B deficiency was the first reported genetic cause of lethal RDS in infants [36]. A full-term newborn died in the neonatal period from unclear lung disease. An open lung biopsy revealed the histopathological features of congenital PAP, a pathology which had been observed previously in several full-term neonates with severe respiratory failure, mostly with a family history of neonatal respiratory disease [37], suggesting an autosomal-recessive inheritance. In lung tissue of the patient, mature SP-B was undetectable whereas SP-A and proSP-C were regularly found. Genetic analysis revealed that the affected infant was homozygous for a frameshift mutation in codon 121 (termed 121ins2 mutation) in the SP-B gene [38,39]. The 121ins2 loss-of-function mutation is suggested to account for up to two-thirds of the mutant alleles identified in the SP-B locus. At least 27 loss of function mutations were found in the SP-B gene that resulted in neonatal RDS [40-48]. Since most of these mutations were unique to the index families the sensitivity of genetic counseling is limited. The incidence of hereditary SP-B deficiency is not known, but an allele frequency of 1 per 1000–3000 individuals for the 121ins2 mutation is suggested [45,47-49]. The phenotype of infants with hereditary SP-B deficiency is described as typically full-term neonates with severe RDS who are refractory to standard therapies, includingventilation and surfactant replacement. Chest X-rays usually show diffuse ground glass pattern in both lung fields consistent with a diagnosis of hyaline membrane disease. PAP was found as the predominant pulmonary phenotype of the first identified infants with SP-B deficiency, which was not the case for all infants with this genetic disorder. Recently, a study of neonates and children with severe unexplained respiratory distress and PAP revealed that the majority of cases with PAP was not linked to SP-B gene mutations [48].
Histologically, SP-B deficiency in humans has been associated with PAS-positive eosinophilic material in the alveoli, epithelial cell desquamation, enlarged alveolar macrophages with lamellar inclusions and abundant accumulation of SP-A [50,50]. Since SP-B was found to be essential for the proteolytic processing of proSP-C [34,51], newborns with hereditarySP-B deficiency show aberrantly processed SP-C in theintra-alveolar lumen [50,52-54]. Furthermore, an abnormal surfactant activity and altered levels of phosphatidylglycerol were observed in SP-B deficient patients [42,52,55].
Although SP-B deficient neonates usually present with respiratory failure in the first 24 to 48 hours of life, the clinical diagnosis can sometimes be delayed since affected infants may show initially mild symptoms and do not require ventilation or further medical support for some time. The lung disease however is rapidly progressive and fatal respiratory failure finally sets in within 3–6 months, even with maximal medical efforts [56] and lung transplantation is suggested as the only effective treatment [57]. Despite a possible, transient beneficial effect, exogenous surfactant in SP-B deficient infants was found to be largely ineffective [58,59]. This lack of response of surfactant replacement in SP-B deficient infants may be due to an additional function of SP-B precursors, to the accumulation of aberrant SP-C which may inhibit surface film formation or due to an impairment of SP-B recycling [10].
Partial surfactant protein B deficiency
There is increasing evidence that some SP-B mutations may result in milder pulmonary phenotypes. In mice with heterozygous disrupted SP-B production, the reduced SP-B synthesis led to chronic lung damage when exposed to hyperoxia [60]. Exogenous SP-B corrected this oxygen-induced pulmonary dysfunction [61].
So far, four patients with partial SP-B deficiency are described in the literature. Characteristics of these patients are described in table 1 (Additional file 1).
Case 1 (mutation: G135S)
A male neonate presented during the first hours of life with respiratory distress and required mechanical ventilation. The patient developed persistent pulmonary hypertension and was treated with continous oxygen therapy. Biochemical analyses of his tracheal aspirate fluid revealed a complete absence of SP-B and an aberrant form of SP-C at 12 kd, suggesting the diagnosis of hereditary SP-B deficiency. A genetic analysis demonstrated a point mutation in exon 5 on one SP-B allele and the mother was found to be the carrier. Histological sections of the lung revealed PAP. The authors suggested that this patient had a transient SP-B deficiency caused by a SP-B gene variant. The patient developed chronic lung disease and was alive at 3 years of age requiring oxygen supplementation [62]
Case 2 (mutation: 121ins2/R236C)
The most severe course of partial SP-B deficiency was a full-term infant who developed tachypnea shortly after birth. A chest radiogram showed bilateral pneumonitis. Because of poor oxygenation, the infant received extracorporeal membrane oxygenation (ECMO) and remained oxygen dependent for 9 months. The patient showed a beneficial response to corticosteroids and withdrawals of steroid therapy worsened his status dramatically. At 8 months of age, no SP-B was detected in BAL and a genetic analysis revealed a heterozygous state for the 121ins2 mutation. The other allele carried a point mutation (R236C). The mother was identified as the carrier for the 121ins2 and the father of the R236C allele. Since his pulmonary situation deteriorated rapidly, he was accepted for lung transplantation but died at the age of 9 months of unexplained sudden respiratory failure and cardiac arrest.
Cases 3 & 4 (mutation: c.479G>T)
Two unrelated children homozygous for a mutation in exon 5 (479G→T) of the SP-B gene have been reported. The genetic defect resulted in reduced SP-B levels and an unclear lung disease. The first patient was born at 40 weeks of gestation and showed RDS at 8 hours of age. Furthermore, he developed a spontaneouspneumothorax at 18 hours of age. He required continous oxygen supplementation and dexamethasone seemed to have no influence onhis respiratory symptoms. At 5 weeks of life, he underwent an open lungbiopsy. Hisrespiratory function declined progressively and he underwentbilateral lung transplantation at 4 months of age.
The second patient was a full-term female who developed RDS shortly after birth. The patient did not require mechanical ventilation but had a persistentoxygen requirement. At 2 months ofage, an open lung biopsy was performed which showed interstitialfibrosis and PAP. Diuresis, steroid therapy and cyclophosphamide therapy had no beneficial effect on her pulmonary symptoms and she developed pulmonary hypertension andright ventricular hypertrophy. At 6 months ofage she was discharged with home oxygen therapy and has been stable for several years with persistent oxygen requirement [63].
The histological phenotype of partial SP-B deficiency included accumulationof extracellular proteins, atypical macrophages, epithelial-celldysplasia and pulmonary fibrosis. However it remained unclear whether these pathologicalfindings were caused by impaired surface tension, by accumulationof abnormal SP-B or by other factors that may causethese forms ofILD. The patients with partial SP-B deficiency were bearing uncommon mutations leading either to the production of small amounts of SP-B or to proSP-B which was not further processed to mature SP-B [62-64]. A critical SP-B level required for physiological lung function has been suggested, which is supported by experiments in mice which developed lung disease when SP-B levels were genetically downregulated below 20–25% of physiological values [65]. Since SP-B levels in the reported cases were not evaluated during the course of disease, the association between levels of SP-B and respiratory status could not be assessed.
Surfactant protein B polymorphisms
Polymorphisms in the SP-B gene or an interaction between SP-A and SP-B gene products may act as genetic determinants for the susceptibility to RDS [66-68]. The SP-B gene contains two polymorphisms, the Ile131Thr locus and a polymorphic site in intron 4. A length variation in intron 4 has been proposed to associate withRDS independently [69] and in combination with an SP-A allele [70]. The Ile131Thr polymorphism affects a putative N-linked glycosylation site of proSP-B which may play a role in the genetic susceptibility to RDS in premature neonates. Furthermore, the SP-B Ile131Thr polymorphism may be linked to an unknown genetic element that increases the risk of RDS in cooperation with SP-A allelic variants [66,68]. A large study of premature neonates found no association of the two SP-B polymorphisms with RDS, but showed that the association between SP-A alleles and RDS was dependenton the SP-B Ile131Thr genotype [71]. A consanguineous pedigree with 14 infants dying of neonatal RDS had biochemically a SP-B deficiency. Molecular analyses showed nine SP-B polymorphisms but none of them could explain the observed SP-B deficiency. Interestingly, in lung tissue of these patients aberrant SP-B mRNA was found [72]. A further study of SP-A and SP-B genetic variants in singletons and twins found no direct genetic influence on the risk of RDS [73].
Surfactant protein C
Structure and biosynthesis
SP-C is the only surfactant-associated protein that isexclusively expressed in typeII pneumocytes [74]. The SP-C gene is encoded on the short arm of human chromosome 8 and contains 6 exons. The mature SP-C is encoded in exon 2 and the sixth exon is being untranslated. SP-C is synthesizedby proteolytical processing of the larger precursor protein proSP-C [54]. The SP-C gene is transcribed into an 900 bp mRNA and is translated into the 197 amino acid proprotein (proSP-C). Alternative splicing atexon 5 results in a shorter transcript of 191-amino acid proprotein but the physiological role of the splice variant is not known [75]. The SP-C precursor protein isrouted together with the SP-B precursor protein to multivesicularbodies, where both proteins are processed and packaged into lamellarbodies for secretion into the alveolus [51]. ProSP-C is extensively processed posttranslationally, similar to proSP-B. Recently, the proteases cathepsin H [30] and Napsin A [76] were found to be involved in the first N-terminal processing step of proSP-C. Furthermore, the SP-C peptide is modified posttranslationally by palmitoylation at cysteine residues at positions 5 and 6 [54]. The mature SP-C peptide of 35-amino-acids is stored in the lamellar body from where it is secreted into the alveolar space.
Surfactant protein C deleted mice
While SP-B was shown to play a critical role in neonatal lung function, SP-C seems to be involved in a different way.
Swiss black outbred mice [35]
The initial strain of SP-C knock-out mice (Swiss black outbred) was viable and had normal lung function. Lung tissues showed subtle pulmonary abnormalities with regular lamellar bodies, reduced viscoelasticity and decreased stability of surfactant bubbles. SP-C mRNA was not detected in the lungs of these mice and no mature SP-C protein was found. BAL fluid levels of SP-A, SP-B and SP-D were similar compared to wild-type mice and pulmonary structures were normal with no indication of inflammation. The number of type II pneumocytes was similar compared to wild type and well-organized lamellar bodies and tubular myelin were detected in the SP-C -/- strains. However, lung mechanics showed a lower hysteresivity (mechanical coupling between tissue resistance and elastance) in SP-C -/- mice at the end-expiratory pressure (low PEEP), suggesting that SP-C stabilizes the phospholipid film at reduced lung volumes. The surface activity of large aggregate surfactant from SP-C -/- mice was similar to SP-C wild-type mice, but the stability of small bubbles was decreased in surfactant of SP-C -/- mice.
This SP-C -/- genetic model showed marked differences compared to the SP-B knock-out model and suggests, (1) that SP-C is not essential for the synthesis or secretion of pulmonary surfactant, (2) that SP-C plays an important role in the stabilization of phospholipid packing at small bubble radius and (3) that SP-C is involved in the recruitment of phospholipids from the subphase to the surface film. Thus, SP-C may play a more important role in the stabilizationof surfactant rather than in its synthesis [35].
129/Sv strain mice [77]
Interestingly, when these SP-C-deficient mice were generated in a different genetic background (129/Sv strain) they developed a more severe lung disease. While lung structure was normal at birth, a progressive air space enlargement, emphysema, macrophage infiltration, abnormal tissue lipid accumulation and pulmonary fibrosis was observed within 2 months after birth. At the age of 6 months, parenchymal infiltrates and type II cell hyperplasia were found. Furthermore, these lungs showed activation of metalloproteases, fragmentationof elastin fibers and infiltration of myofibroblasts in thealveolar septa, similar to findings ofILD in humans. The pulmonary alterations increased progressively with age and at one year of age, the mice developed pneumonitis and emphysema. Analysis of lung mechanics revealed that lung volumes were significantly increased in SP-C -/- mice as compared to wild type mice, consistent with the histologically observed emphysema. Biochemical analyses showed normal levels of SP-B while levels of SP-A and SP-D were increased.
This knock-out mouse model shows that lack of SP-C causes chronic lung disease depending on the genetic background suggesting that SP-C associated lung disease may be triggered by environmental factors and/or modifier genes [77].
Surfactant protein C deficiency in human lung disease
There is increasing evidence linking SP-C deficiency to various types of ILD in humans [10,78,79]. While SP-B deficiency generally results in fatal neonatal lung disease, abnormal expression of SP-C has been shown to present as a much more variable phenotype [78,80-89]. Amin et al. found an association between abnormalities in SP-C expression and familial ILD. Three family members with chronic ILD had no detectable levels of SP-C and decreased levels of SP-A and SP-B in their BAL fluid. Immunostaining showed a decreased expression of proSP-C in the alveolar epithelial cells, but no mutations in the exons of SP-B or SP-C were found. Thus the underlying mechanism responsible for the lack of SP-C in these patients remained unknown [90].
Table 2 (Additional file 1) summarizes reported cases of ILD associated with SP-C gene variants. Since the clinical courses showed a large variance, each report is described separately.
Case report on SP-C deficiency with mutation c.460+1 G>A / Δexon 4 [83]
The first report on SP-C deficiency was a female infant whose family history showed a three-generation history of ILD inherited in an autosomal-dominant way. The mother was given a diagnosis of DIP at one yearof age and she had been treated with glucocorticoids until the age of 15 years. Her father had diedfrom an unclear chronic lung disease without a diagnosis. The women delivered a full-term baby with no respiratory symptoms at birth. This infant developed tachypnea and cyanosis at six weeks of age and chest radiographs showed interstitial involvement. An open-lung biopsy was performed andshowed NSIP. Mature SP-C could not be detected. Sequencing of the SP-C gene revealed a single base donor splice site mutation with substitution of the first base of intron 4 (c.460+ 1G > A) leading to the skipping of exon 4 and a shortened SP-C proprotein. This mutation was identified on only one allele of the SP-C gene suggesting an autosomal-dominant way of inheritance. The patient's mother also had reduced expression of proSP-C and the same SP-C mutation was found. The girlwas treated with oxygen and glucocorticosteroids andher respiratory symptoms improved. After delivery the lung disease of the mother worsened and she died from respiratoryfailure.
Furthermore, Nogee et al. analyzed the SP-C gene sequence in 34 infants with chronic lungdisease of unknown etiologyand found heterozygous mutations in 11 of 34 patients. Similar to the initially described patient, these mutations led to skippingof exon 4, expression of an aberrant proSP-C and decreased expression of normal SP-C. In 10 infants, with 6 showing a family history of ILD, missense SP-C mutations resulting in amino-acid substitutions (P30L, I73T, G100V, Y104H, P115L, I126R, T187N and L188R)and a frameshift mutation (140delA) associated with expressionof a stable transcript, were identified. Recently, neonates with severe RDS carrying mutations in the SP-C gene were described [91]. Exon 4 deleted SP-C constructs concentrated in perinuclear aggregates which have been shown to be toxic for the developing mouse lung [92]. Furthermore, co-transfection with wild-type and exon 4 deleted SP-C constructs resulted in a pattern similar to the mutant SP-C construct, suggesting a dominant-negative mechanism in heterozygous SP-C mutations [93].
Familial SP-C deficiency with mutation c.588 T>A / L188Q [81]
Thomas reported a large kindred of 16 individuals with adults suffering from UIP and children from cellular NSIP. Genetic analyses found a heterozygous T>A transversion at exon 5 + 128 which led to a substitution of a glutamine for leucine at the highly conserved amino acid position 188 of the C-terminus of proSP-C, a region which is essential for intracellular protein trafficking. Immunostaining showed an aberrant subcellular localization of SP-C precursor protein, dense fibrosis and distorted cellular architecture with alveolar type II cell atypia and abnormal lamellar bodies. These findings were similar to those which have been described in type II cellular death or apoptosis [94,95]. When mouse type II cells (MLE cells) were transfected with this SP-C mutation, a similar morphology was found with atypical inclusions resembling lamellarbodies. Furthermore, expression of the mutant protein resulted ina threefold increase of cytotoxicity compared to the wild-type.
Since the identical SP-C mutation caused different pathologic phenotypes (UIP, NSIP) in affected relatives, it has been suggested that SP-C dependent lung diseases may represent pleiotropic manifestations of the same underlying pathogenesis and the mutation may result indifferent histological patterns dependent on the age of manifestation. It is also possible that in this family, cellular NSIP in the affected children occurred as a precursor lesion to UIP in adulthood, but the discrimination of theseentities is discussed controversial [96-98]. Interestingly, two members of the index family with heterozygous SP-C mutations were unaffected. Therefore, the mutation in this kindred showed incomplete penetrance and secondary genetic and/or environmental modifiers may affectthe disease progress.
Case report on SP-C deficiency with mutation in exon 3 [86]
A full-term female of two parents without a history of pulmonary disease thrived normally until the age of 3 months. Then, she developed respiratory symptoms with diffuse interstitial infiltrates. The airway surfactant contained decreased levels of mature SP-B and SP-C. An open lung biopsy at 6 months of age revealed an interstitial pneumonitis without an underlying cause. Minimum pulmonary surface tension was increased. SP-C mRNA was present in normal size and amount, but proSP-C was aggregated within type II cells in a compartment separate from SP-B. The cells contained both normal and disorganized lamellar bodies. A genetic analysis was performed and revealed a 9-bp deletion in exon 3 of the SP-C gene. This in-frame mutation spanning codons 91–93 was present on one allele only. Since the parents did not carry this deletion a de novo mutation was suggested. The patient's lung function and physical capacity declined dramatically and a bilateral lung transplantation was performed with 14-months of age. The patient's course ameliorated and at 30 months she was breathing ambient air and was gaining weight.
Case reports on SP-C deficiency with mutation c.243 T>C / I73T and c.525 G>A / R176Q [80]
In a group of 34 sporadic and familial cases with RDS, in which SP-B mutations had been excluded, two distinct heterozygous SP-C missense mutations were found. The first case was a full-term boy with a normal neonatal course who developed slight respiratory symptoms at the age of 1 month. At age of 9 months, he failed to thrive and his dyspnoea worsened countinously, so he required oxygen supplementation. A subsequent lung biopsy showed PAP combined with ILD. PAP has not yet been described before in patients with SP-C gene variants. Treatment with repetitive whole-lung lavage and systemic corticosteroids and azathioprine resulted in an amelioration of his pulmonary symptoms. Since biochemical BAL fluid analysis showed normal levels of SP-B and SP-C but aberrant forms of proSP-C, a genetic analysis of the SP-C gene was performed. A missense mutation (c.243 T > C) in exon 3 of the SP-C gene was identified, which resulted in the substitution of threonine for a highly conserved isoleucine (I73T). This mutation was not detected in the parents. Similar to the other SP-C mutations, this mutation was heterozygous, consistent with dominant inheritance [88]. In contrast to the initial patient of Nogee [83], where no SP-C could be detected, in this patient mature SP-C protein was present together with an abnormally processed proSP-C suggesting that the mutation altered the processing of proSP-C but allowed the production of mature SP-C.
Brasch et al [88] characterizedthe intracellular trafficking of the I73T mutation in vitro using a EGFP/proSP-CI73T fusion protein. When transfected into A549 cells, the mutant proSP-C was routed to early endosomes which is in contrast to previously described SP-C mutations which led to the formation of intracellular proSP-C aggregates [80,93].
Recently, a further de novo I73T mutation was found in a full-term infant with PAP and ILD showing a more severe clinical course as compared to the previous case with this mutation. After recurrent bacterial infections, the patient died from respiratory failure at the age of 19 months [99].
The second SP-C mutation of Tredano et al. was found in two patients with PAP from the endogamous white settler population of Réunion Island in which unexplained respiratory distress (URD) has an unexpectedly high prevalence. These two patients had a similar age at the onset of disease of 9 months. However, one infant died from lung disease at age of 18 months, while the other one showed a mild ILD, which was treated by repetitive therapeutic BALs. Both patients showed an abnormal proSP-C, and genetic analysis of the SP-C gene revealed a 2125G-A transition in exon 5, leading to an arginine 167 to glutamine change (R167Q) in the C-terminus of the precursor protein, which is crucial for proper folding and trafficking of proSP-C [93].
Case report on SP-C deficiency with mutation g.1509 G>A / E66K [82]
A full-term male showed no post-natal respiratory problems until day 13 when he was admitted to a local hospital with tachypnea, cyanosis and hypoxemia. Chest X-ray showed diffuse bilateral pulmonary infiltrates and he required mild ventilatory support. A lung biopsy showed marked thickening of alveolar septa, mild interstitial chronic inflammation and hyperplasia of type II pneumocytes. The histological diagnosis of PAP and NSIP was esthablished. Immunostaining yielded detectable levels of SP-A, SP-B, SP-C and SP-D in BAL fluid. A DNA sequence analysis found a heterozygous mutation in exon 2 of the SP-C gene (1509 G>A) leading to substitution of a lysine for glutamic acid at codon 66 (E66K). Histologically, proSP-C expression was detected in small endosome-like vesicles. Further in vitro trafficking studies were performed using the EGFP/hSP-CE66K fusion protein and showed, that the proSP-CE66K protein accumulated in early endosomes in contrast to Δexon4 and L188Q mutant constructs which formed intracellular aggregates.
Surfactant protein C polymorphisms
Two single nucleotide polymorphisms (SNPs) in the SP-C coding sequence resulting in 4 allelic variants have been described. One is localized in codon 138 that encodes threonine or asparagines and the other one in codon 186 that codes for serine or asparagine. The allelic frequencies or the effect of these SNPs on proSP-C processing or function are not known. Several other SNPs in non-coding regions of the SP-C gene have been reported, but also their functional relevance remained unclear [100-102]. Lahti et al. analyzed the association between exonic SP-C gene polymorphisms and their susceptibility to RDS [103]. The study showed that SP-C polymorphisms were associated with RDS and with premature birth but the strength of association varied depending on to the gender of the infants. Recently, Lawson et al found 10 SNPs in the SP-C gene sequence of 13 of 135 analyzed adults with sporadic UIP and NSIP but not in controls. Thus, the majority of sporadic cases of ILD in adulthood may not be associated with SP-Cgene variants [104].
Immunomodulatory functions of surfactant protein C
While the collectins SP-A and SP-D are well known to participate in innate immune response, a possible role of SP-C in pulmonary immunity has been elucidated only recently. Inhalationof bacterial lipopolysaccharide (LPS) induces acute airway inflammation [105] and long-term exposure to LPS increases the risk of developing chronic lung disease [106]. It has been shown, that the hydrophobic regions of SP-C bind Lipid A, a portion of LPS, in a specific and competitive way. Furthermore, SP-C was found to interact with CD14, the cellular LPS pattern recognition receptor which plays a central role in bacterial-induced inflammation. The SP-C/CD14 interaction blocked the binding of LPS to CD14 and inhibited the effect of LPS on macrophages (Figure 1). This resulted in a decrease of pro-inflammatory cytokine (TNF-α) and NO release by alveolar and peritoneal macrophages in response to LPS [107]. However, an in vitro study showed that SP-C didnot influence the direct binding of LPS to CD14, suggesting that other receptors on macrophages, e.g. toll-like receptor 4 or the complement receptor CD11/CD18 may be more essential in this mechanism [108].
Figure 1 Pathways in the pathogenesis of surfactant protein-C associated interstitiallung disease (according to Beers 2004 [74]). LPS: lipopolysaccharide; MØ: macrophage
These findings indicate thatSP-C plays a role in innatelung defense by preventing LPS-induced pulmonary inflammation. SP-C may either neutralize inhaledLPS or may act as a shuttle to allow the binding of LPS to distinct receptors on alveolarmacrophages. These immunological findings may explain why exogeneous SP-C showed a beneficial effect in a porcine model of LPS-induced lung injury [109].
Pathogenesis of surfactant protein C associated lung disease
The current model of SP-C associated lung disease suggests two mechanisms how aberrant forms of SP-C may cause interstitial lung disease.
• Heterozygous mutations in the SP-C gene produce an aberrant proSP-C protein. This aberrant protein interacts with the regular intracellular pathway of SP-C biosynthesis. Consequently, this leads to inhibition of functionally active SP-C production and results in a SP-C deficiency in the alveolus.
• When aberrant proSP-C is produced in large amounts, it cannot be handled by the protein degradation pathway and induces cellular injury and pulmonary inflammation resulting in interstitial lung disease. Figure 1 depicts several hypothesized pathways in the pathogenesis of SP-C associated ILD according to Beers [78].
SP-C associated lung disease may be caused either by a lack of mature SP-C, by the accumulation of aberrant, toxic proSP-C or by both mechanisms. In humans with SP-C mutations, additional factors, e.g. viral infections, hypoxia or drugs, may trigger the accumulation of misfolded proprotein as the SP-C clearance pathway may operate at its limiting capacity. Since SP-C is suggested to be involved in thealveolar reuptake of surfactant particles and plays a role in innate immunity, it remains unclear which specific activity causes the lung diseasein patients with SP-C mutations.
BRICHOS domain
Recently, a conserved domain of ~100 amino acids, termed BRICHOS domain, has been identified in several previously unrelated proteins. These proteins include the BRI family of proteins (linked to familial British and Danish dementia), chondromodulin-I associated with chondrosarcoma, CA11 protein associated with stomach cancer and the distal region of proSP-C (F94-I197) associated with ILD. In most of these proteins, the BRICHOS domain is localized in the propeptide region that is removed during proteolytic processing [110]. Mutations in the SP-C gene can be classified depending on the involvement of the BRICHOS domain (Table 2, Additional file 1, and more detailed in [78]). In vitro trafficking experiments showed that constructs of BRICHOS domain mutations, e.g. hSP-CΔexon4 [83] and hSP-CL188Q [81], form large toxic intracellular aggregates, termed aggresomes [82,93]. In contrast, mutations outside the BRICHOS domain, as hSP-CE66K and hSP-CI73T, led to aberrant proSP-C accumulation in a constitutive endosomal pathway which may inhibit the cellular surfactant recycling system [78,82,88]. Clinically, mutations in the BRICHOS domain resulted in a more severe pulmonary phenotype, mostly associated with death in the neonatal period, while non-BRICHOS mutations led to a milder clinical course.
Amyloid fibril formation
Amyloid fibril formation represents a potential mechanism how misfolded higher-order structures of SP-C proteins may cause epithelial cell damage. At least 20 incurable diseases are associated with amyloid fibrils including Alzheimer's disease and spongiform encephalopathy [111]. The metastable α-helical domain of SP-C was found to transform under certain circumstances into energetical favored β-sheet aggregates, which polymerize to an amyloid-like fibril formation [112]. Such amyloid-like fibrils were isolatedfrom the BAL fluids of a PAP patient and in low amounts of three healthy controls [113]. The pathophysiological relevance of these findings remains unclear but suggest that conditions, which alter the secondary structure of SP-C in terms of predisposing to intracellular fibril formation, may have deleterious effects on type II cells. These findings support the model of a conformational lung disease due to a misfolded SP-Cprotein [78]. Destabilizing point mutations may enhance the transition of regular SP-C structure to pathological forms [78].
ATP-Binding cassette (ABC) transporters
ABCA3
The ATP-binding cassette (ABC) transporters aremembrane proteins that utilize the energy of ATP hydrolysis to translocate solutes across cellular membranes. Fourteen ABC genes have been associated withdistinct genetic diseases in humans including Tangier-disease and cystic fibrosis [114,115]. ABCA3 is a 1704 amino acid proteinwhichis highly expressed in type II epithelial cells at the limiting membraneof lamellar bodies. The structureand localization of the ABCA3 transporter suggest a potentialrole in intracellular lipid homeostasis of lamellar bodies [116]. Mutations in the ABCA3 gene are the most recently described inborn error of surfactant metabolism [16]. Shulenin et al. identified mutations in the ABCA3 as a cause of fatal RDS in newborn infants and theunderlying cause of chronic ILD in one child living up to 6 years of age. In 21 infantswith severe neonatal surfactant deficiency in which SP-B and SP-C mutations had been excluded, the coding regions of the ABCA3 genewere sequenced. Furthermore, lung tissue from four patients was analyzed. In 16 of the 21 patients, nonsense, frameshift and splice sites mutations were found. These patients included five consanguineous families. The consanguinity and a family historyof neonatal respiratory distress suggest that the lung disease associatedwith mutations in the ABCA3 gene were inherited in an autosomal-recessive manner. A mutation on one allele was identified in two patients who showed a similarclinical presentation as compared to the remaining patients. Therefore the authors speculate that these children may have a second mutation within introns or in regulatory regions which was not detected by sequencing the coding regions. The radiological findings and clinical symptoms of infants with ABCA3 mutations were similar to newborns with SP-B deficiency. One infant with a missense mutation on one allele was still alive at six years of age showing lung histology of a ILD/DIP, suggesting that some ABCA3 mutations are not fatal and may result in a milder course of disease. Histological examination in these patients showed type II cells hyperplasia and accumulation of alveolar macrophageswith proteinaceous materialand interstitial thickening, indicating DIP and neonatal PAP. Electron microscopy revealed abnormally small lamellar bodies indicating that the ABCA3 transporter is critical for proper formation of the lamellar bodies. A defectivetransport of lamellar-body associated lipid components may result in abnormal processing and structure of surfactant. Furthermore, the mutant ABCA3 transporter may transfer lipids which are deleteriousto the function of alveolar surfactant. These findings suggest that a high percentage of unclear casesof lethal neonatal lung disease may be due to mutations in the ABCA3 gene and affected families could benefit from genetic counseling.
Since there are no therapies known for the lung disease due to ABCA3 mutations, the majority of affected newborns die from respiratory failure in the neonatal period. Except for one patient, the phenotype of non-fatal lung disease caused by ABCA3 mutations is unknown and requires screening for ABCA3 mutations in neonates and children with severe ILD where no underlying cause can be found [16].
ABCA1
A recent in vitro study showed that surfactant synthesis and export are associated with a basolateral lipid efflux pathway which is mediated by ABCA1, suggesting a role for ABCA1 in the regulation of cellular content and composition of pulmonary surfactant [117]. Targeted disruption of the ABCA1 locus resulted in distinctive pulmonary features [118]. Macroscopically, multiple pale foci were found in the lungs which increased in severity withage, affecting <10% of the parenchyma in 7 months-old miceto 30% in 18 months-old mice. These lesionsconsisted of foamy hypertrophic type II cells, alveolar macrophages and cholesterol clefts. Furhermore, alveolar septae were thickened by aggregates of lymphocytes and plasma cells. In more severe lesions, pulmonary architecture collapsed due to hypertrophic and hyperplastic type II pneumocytes. Oil red staining showed lipid accumulation within alveolar cells and electron microscopy identified these cells as type II pneumocytes and intraalveolar macrophages. The possible role of ABCA1 in ILD is unknown and there are no reports on ABCA1 mutations associated with lung disease in humans.
Genes involved in lung morphogenesis, surfactant protein expression and lamellar body synthesis
Besides SP-B, SP-C and ABCA3, there is a number of genes involved in lung morphogenesis, surfactant expression and lamellar body function which may also play a role in ILD in children.
Lung morphogenesis is a complex process mediated by numerous transcription factors [11,119], growth factors [120,121] and other signalling molecules that coordinatealveolar cell proliferation and differentiation. Factors playing important rolesin lung morphogenesis include thyroid transcription factor-1(TTF-1) [122-124], GATA-6 [125,126], Foxa2 [127], Hoxb5 [128], TGF-β via SMADs [129,130], CREB [131], Glucocorticoid Receptor [132], HNF-3β [133], N-myc [134] and members of the Gli family [119]. Knock-out mice of these genes have provided insight into the function during lung development [119,135]. For example, Foxa2 expressionwas found to be restricted to subsets of respiratory epithelial cells and was co-expressed with TTF-1, CCSP, SP-A, SP-B, SP-C and SP-D [127,136]. In a mouse model, deletion of Foxa2 in pulmonary epithelialcells inhibited the differentiationof respiratory epithelial cells and resulted in neonatal respiratoryfailure with hyaline membranes as typically found in preterm infants with RDS.
TTF-1
TTF-1 is a member of the Nkx2.1 family which is involved in thyroid and lung epithelial-specific gene expression [124,137,138]. Co-transfection assays showed that TTF-1 activity is stimulatedby HNF-3β and GATA-6 which are co-expressed with TTF-1 in the developing lung [139]. While TTF-1 serves overall as an important regulator of surfactant protein gene expression [140], a regulatory role for HNF-3β has been reported only for SP-B [141] and CCSP [142] expression. Furthermore, TTF-1 and HNF-3β were shown to functionally interact in the regulation of the SP-B and CCSP promoter activity [122,143]. TTF-1 deficient pulmonaryepithelial cells failed to express mature SP-B, SP-C and CCSP [124,144]. Deletion of TTF-1 in mice caused thyroid and lungabnormalities with tracheal-esophageal fistulaand dysgenesis of the peripheral lung resulting in respiratoryfailure at birth [145,146]. In humans, mutations in the TTF-1 gene were associated with chorea, hypothyroidism and respiratory distress in newborns [147-149]. Furterhmore TTF-1 gene variants were found in older children with choreoathetosis, hypothyroidism and respiratory distress [150].
Growth factors and cytokines
Various growth factors and cytokines have been described to regulate surfactant protein gene expression in the fetal lung [121]. Epidermal growth factor (EGF) was found to increase SP-C gene expression [151], while TGF-β and TNF-α inhibited surfactant protein expression. In mice homozygous for a targeted deletion of the EGF receptor, defects in surfactant function and in SP-A and SP-C expression were observed [152]. The inhibitory effect of TGF-β on surfactant protein gene expression in the fetal lung was found to be mediated at the transcriptional level [26] through SMAD3 interactions with TTF-1 and HNF-3β [153].
Lamellar body associated molecules
Lamellar bodies have an acidic interior (pH≈5.5) and contain lysosomal enzymes (e.g. cathepsins C and H) and lamellar body associated membrane proteins (LAMP; e.g. LAMP1 [154], LAMP2 [155] and LAMP3 [156]). Recently, the dendritic cell-lysosomal associated membrane protein (DC-LAMP) was detected intracellularly and at the cell surface of type II pneumocytes. Since type II cells constitutively express MHC class II molecules, it is suggested that they may act as antigen-presentingcells. Since DC-LAMP is expressed on type II cells and on dendritic cells, a relationship between DC-LAMP and MHC II function in type II pneumocytes has been suggested [157,158]. Recently, a study showed that the proteins t-SNARE, syntaxin 2 and SNAP-23 were expressed in alveolar type II cells and were associated with the plasma and lamellar body membrane. SNAP23 was found to be required for the fusion of lamellar bodies with the plasma membrane. An antisense oligonucleotide for syntaxin 2 inhibited surfactant secretion. These results suggest that syntaxin 2 and SNAP-23 are required for regulated surfactant exocytosis in type II cells [159].
The characterization of molecules involvedin surfactant protein gene expression, lung morphogenesis and lamellar body synthesis expands the current model how a network of genes may cause, trigger or promote lung disease. However, most of these genes causedlethal lung malformations rather than chronic lung diseases and were only studied in animal models [11]. Nevertheless, the potential role of these genes in pediatric lung disease is still unclear and requires further investigation. Insights on how these molecules are involved in human lung physiology may reveal new candidate genes for identifying mutations associated with ILD in children.
Hermansky-Pudlak syndrome
In the autosomal-recessive disorder Hermansky-Pudlak-Syndrome (HPS), most prevalent among Puerto Ricans, the biogenesis of lysosomes and lysosome-related organelles is affected. In humans, six different forms of HPS have been described referring to the genes HPS 1 to 6. In patients with HPS, the molecular defect leads to hypopigmentation, bleeding disorders and fibrotic lung disease, mostly UIP. The incidence and severity of pulmonary involvement in HPS patients varies widely but the forms HPS1 and HPS4 were most likely associated with ILD [160-162]. Lung tissue from HPS patients with ILD exhibits giant lamellar bodies in type II epithelial cells, patchy fibrosis and airspace enlargement [163].
Conclusion
Surfactant protein B
The lethal lung disease in SP-B deficient mice and humans demonstrates the crucial role for SP-B in neonatal lung function. The accumulation of aberrant proSP-C in SP-B deficiency strongly suggests an important intracellular role for SP-B in the processing of SP-C. The clinical phenotype of hereditary SP-B deficiency presents typically as full-term neonates with severe respiratory distress syndrome refractory to standard therapies, includingventilation and surfactant replacement. Lung transplantation is regarded as the only effective treatment. Chest X-rays are similar to those of premature infants with respiratory distress syndrome showing diffuse ground glass pattern. Alveolar proteinosis is the predominant pathological phenotype. The SP-B 121ins2 mutation is suggested to account for up to two-thirds of the mutant alleles identified in the SP-B gene. The overall incidence of hereditary SP-B deficiency can be estimated from the allele frequency of the 121ins2 mutation of 1 per 1000–3000 individuals. Of great interest for clinical practice are some cases with partial SP-B deficiency. These patients showed less severe clinical phenotypes and prolonged survival into childhood with chronic lung disease and oxygen dependency as compared to fully SP-B deficient patients. Thus, it is likely that partial SP-B deficiency may be the underlying cause in some children with unexplained chronic lung disease.
Surfactant protein C
SP-C mutations were found to be associated with ILD in neonates and children with sporadic or autosomal-dominant inheritance. The majority of these mutations were missenses or deletions leading to skipping of exon 4 or substitution of highly conserved amino acids in the region of the proSP-C molecule. In vitro studies support the hypothesis that the identified SP-C mutations were causally related to the associated lung disease. The phenotype covers a wide range of clinical features and age of onset. Reported symptoms include respiratory distress with dyspnoea, tachypnea, cyanosis, asthmatic bronchitis and failure to thrive. The histological diagnosis was mostly NSIP, in some cases combined with PAP. Adults with SP-C mutations showed UIP or DIP while other subjects remained asymptomatic. The phenotype of lung disease associated with SP-C mutations may represent pleiotropic manifestations of the same underlying pathogenesis. Affected individuals may carry a genetic risk for interstitial lung disease, which becomes apparent in dependence of the genetic background, i.e. modifier genes and environmental influences. These findings have essential implicationsfor the diagnosis of ILD in children as genetic diagnoses can be made and underlying etiologies of familiar cases of ILD may be identified. In children with ILD, where no underlying cause can be found, the patient's DNA, BAL and lung tissue and the parents' DNA should be conserved for genetical, biochemical and histological analysis. Since the incidence and prevalence of ILD associated with SP-C gene mutations are unknown it is necessary to analyse larger subsets of clinically well defined patient cohorts.
ABCA3
Mutations in the ATP-binding cassette transporter ABCA3 gene are the most recently described hereditary disorders of surfactant metabolism. In 16 of 21 patients, who had no mutations in the SP-B or SP-C gene, mutations in ABCA3 were found inherited in an autosomal-recessive manner. Clinical symptoms and radiological findings of infants with ABCA3 mutations were similar to newborns with SP-B deficiency. Histological diagnoses included DIP, neonatal PAP and CPI. Interestingly, one infant with a missense mutation on one allele was still alive at six years of age showing lung histology of DIP, suggesting that ABCA3 mutations may also result in a milder course of disease. These findings indicate that a high percentage of unclear casesof lethal neonatal lung disease may be due to mutations in the ABCA3 gene and affected families could benefit from genetic counseling. Except for one patient, the phenotype of non-fatal lung disease caused by ABCA3 mutations is unknown and requires screening for ABCA3 mutations in neonates and children with severe ILD where no underlying cause can be found.
Table 3 (Additional file 1) gives an overview of pediatric ILD associated with surfactant protein deficiency or ABCA3 mutations. In full-term neonates with unexplained, severe respiratory distress, refractory to therapeutical efforts, genetic analysis of the SP-B and ABCA3 coding sequences are warranted. In children with progressive interstitial lung disease, where no underlying cause can be found, a genetic analysis of the SP-C gene may be useful, especially in cases with a family history of unclear lung disease.
Abbreviations
ABCA1 ATP-binding cassette transporter A1
ABCA3 ATP-binding cassette transporter A3
CPI chronic pneumonitis of infancy
DIP desquamative interstitial pneumonitis
GM-CSF granulocyte-macrophagecolony-stimulating factor
ILD interstitial lung disease
LAMP lamellar body associated membrane protein
LPS lipopolysaccharide
NSIP non-specific interstitial pneumonitis
PAP pulmonary alveolar proteinosis
RDS respiratory distress syndrome
SNP single nucleotide polymorphism
SP surfactant protein
TTF-1 thyroid transcription factor-1
URD unexplained respiratory distress
UIP usual interstitial pneumonitis
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DH and MG wrote the manuscript and read and approved the final version.
Supplementary Material
Additional File 1
Table 1. Pathological features and clinical phenotypes of reported cases with partial SP-B deficiency. Table 2. Pathological features and clinical phenotypes of reported cases with mutations in the SP-C gene. Table 3. Overview: Pediatric interstitial lung disease associated with surfactant protein deficiency or ABCA3 mutations
Click here for file
Acknowledgements
We thank Markus Woischnik, PhD, for critical revisions of the manuscript.
==== Refs
Hilman BC Chronic interstitial lung disease Clinical Pediatrics 1998 37 701 702 9825217
Fan LL Langston C Pediatric interstitial lung disease – Children are not small adults American Journal of Respiratory and Critical Care Medicine 2002 165 1466 1467 12045116 10.1164/rccm.2204012
Copley SJ Coren M Nicholson AG Rubens MB Bush A Hansell DM Diagnostic accuracy of thin-section CT and chest radiography of pediatric interstitial lung disease American Journal of Roentgenology 2000 174 549 554 10658741
Hilman BC Diagnosis and treatment of ILD Pediatric Pulmonology 1997 23 1 7 9035192
du Bois RM Focusing on diffuse (interstitial) lung disease: a rapidly evolving field Respir Res 2002 3 16 11980585 10.1186/rr165
Fan LL Kozinetz CA Factors influencing survival in children with chronic interstitial lung disease American Journal of Respiratory and Critical Care Medicine 1997 156 939 942 9310017
du Bois RM The genetic predisposition to interstitial lung disease Chest 2002 121 14S 20S 11893655 10.1378/chest.121.3_suppl.14S-a
Canakis AM Cutz E Manson D O'Brodovich H Pulmonary interstitial glycogenosis – A new variant of neonatal interstitial lung disease American Journal of Respiratory and Critical Care Medicine 2002 165 1557 1565 12045133 10.1164/rccm.2105139
Deterding RR Fan LL Morton R Hay TC Langston C Persistent tachypnea of infancy (PTI) – A new entity Pediatric Pulmonology 2001 72 73 11886155
Nogee LM Alterations in SP-B and SP-C expression in neonatal lung disease Annual Review of Physiology 2004 66 601 623 14977415 10.1146/annurev.physiol.66.032102.134711
Whitsett J A lungful of transcription factors Nature Genetics 1998 20 7 8 9731520 10.1038/1654
Langston C Kida K Reed M Thurlbeck WM Human-Lung Growth in Late Gestation and in the Neonate American Review of Respiratory Disease 1984 129 607 613 6538770
Clement A Allen J Corrin B Dinwiddie R le Pointe HD Eber E Task force on chronic interstitial lung disease in immunocompetent children European Respiratory Journal 2004 24 686 697 15459150 10.1183/09031936.04.00089803
Nogee LM Hull WM Whitsett JA Wert SE Surfactant protein (SP) B deficiency in neonates with severe respiratory failure Pediatric Research 1996 39 2044
Nogee LM Abnormal expression of surfactant protein C and lung disease American Journal of Respiratory Cell and Molecular Biology 2002 26 641 644 12034561
Shulenin S Nogee LM Annilo T Wert SE Whitsett JA Dean M ABCA3 gene mutations in newborns with fatal surfactant deficiency New England Journal of Medicine 2004 350 1296 1303 15044640 10.1056/NEJMoa032178
Pantelidis P Veeraraghavan S du Bois RM Surfactant gene polymorphisms and interstitial lung diseases Respir Res 2001 3 14 11806849 10.1186/rr163
Shulenin S Nogee LM Annilo T Wert SE Whitsett JA Dean M ABCA3 Gene Mutations in Newborns with Fatal Surfactant Deficiency N Engl J Med 2004 350 1296 1303 15044640 10.1056/NEJMoa032178
Rooney SA The surfactant system of the lung 1985 471 502
Griese M Pulmonary surfactant in health and human lung diseases: state of the art European Respiratory Journal 1999 13 1455 1476 10445627
Hawgood S Derrick M Poulain F Structure and properties of surfactant protein B Biochimica et Biophysica Acta-Molecular Basis of Disease 1998 1408 150 160 10.1016/S0925-4439(98)00064-7
Nogee LM Genetics of the hydrophobic surfactant proteins Biochimica et Biophysica Acta-Molecular Basis of Disease 1998 1408 323 333 10.1016/S0925-4439(98)00078-7
Cochrane CG Revak SD Pulmonary Surfactant Protein-B (Sp-B) – Structure-Function-Relationships Science 1991 254 566 568 1948032
Beers MF Shuman H Liley HG Floros J Gonzales LW Yue N Surfactant Protein-B in Human Fetal Lung – Developmental and Glucocorticoid Regulation Pediatric Research 1995 38 668 675 8552432
Mendelson CR Boggaram V Hormonal-Control of the Surfactant System in Fetal Lung Annual Review of Physiology 1991 53 415 440 2042967 10.1146/annurev.ph.53.030191.002215
Beers MF Solarin KO Guttentag SH Rosenbloom J Kormilli A Gonzales LW TGF-beta 1 inhibits surfactant component expression and epithelial cell maturation in cultured human fetal lung American Journal of Physiology-Lung Cellular and Molecular Physiology 1998 19 L950 L960
Hawgood S Surfactant protein B: Structure and function Biology of the Neonate 2004 85 285 289 15218284 10.1159/000078169
Weaver TE Whitsett JA Structure and Function of Pulmonary Surfactant Proteins Seminars in Perinatology 1988 12 213 220 3041604
Whitsett JA Nogee LM Weaver TE Horowitz AD Human Surfactant Protein-B – Structure, Function, Regulation, and Genetic-Disease Physiological Reviews 1995 75 749 757 7480161
Brasch F ten Brinke A Johnen G Ochs M Kapp N Muller KM Involvement of cathepsin H in the processing of the hydrophobic surfactant-associated protein C in type II pneumocytes American Journal of Respiratory Cell and Molecular Biology 2002 26 659 670 12034564
Ueno T Linder S Na CL Rice WR Johansson J Weaver TE Processing of pulmonary surfactant protein B by napsin and cathepsin H Journal of Biological Chemistry 2004 279 16178 16184 14766755 10.1074/jbc.M312029200
Brasch F Ochs M Kahne T Guttentag S Schauer-Vukasinovic V Derrick M Involvement of napsin A in the C- and N-terminal processing of surfactant protein B in type-II pneumocytes of the human lung Journal of Biological Chemistry 2003 278 49006 49014 13129928 10.1074/jbc.M306844200
Clark JC Wert SE Bachurski CJ Stahlman MT Stripp BR Weaver TE Targeted Disruption of the Surfactant Protein-B Gene Disrupts Surfactant Homeostasis, Causing Respiratory-Failure in Newborn Mice Proceedings of the National Academy of Sciences of the United States of America 1995 92 7794 7798 7644495
Vorbroker DK Profitt SA Nogee LM Whitsett JA Aberrant Processing of Surfactant Protein-C in Hereditary Sp-B Deficiency American Journal of Physiology-Lung Cellular and Molecular Physiology 1995 12 L647 L656
Glasser SW Burhans MS Korfhagen TR Na CL Sly PD Ross GF Altered stability of pulmonary surfactant in SP-C-deficient mice Proceedings of the National Academy of Sciences of the United States of America 2001 98 6366 6371 11344267 10.1073/pnas.101500298
Nogee LM deMello DE Dehner LP Colten HR Brief Report – Deficiency of Pulmonary Surfactant Protein-B in Congenital Alveolar Proteinosis New England Journal of Medicine 1993 328 406 410 8421459 10.1056/NEJM199302113280606
de Blic J Pulmonary alveolar proteinosis in children Paediatr Respir Rev 2005 5 316 322 15531257 10.1016/j.prrv.2004.07.001
Nogee LM Garnier G Dietz HC Singer L Murphy AM deMello DE A Mutation in the Surfactant Protein-B Gene Responsible for Fatal Neonatal Respiratory-Disease in Multiple Kindreds Journal of Clinical Investigation 1994 93 1860 1863 8163685
Stahlman MT Gray MP Falconieri MW Whitsett JA Weaver TE Lamellar body formation in normal and surfactant protein B-deficient fetal mice Laboratory Investigation 2000 80 395 403 10744075
Bahuau M Stuhrmann A Christodoulou J Lorant T Houdayer C Elion J Analysis of founder effect in SFTPB 1549C -> GAA (121 ins2) mutation American Journal of Human Genetics 2002 71 369
Tredano M de Blic J Griese M Fournet JC Elion J Bahuau M Clinical biological and genetic heterogeneity of the inborn errors of pulmonary surfactant metabolism Clinical Chemistry and Laboratory Medicine 2001 39 90 108 11341756 10.1515/CCLM.2001.018
deMello DE Nogee LM Heyman S Krous HF Hussain M Merritt A Molecular and Phenotypic Variability in the Congenital Alveolar Proteinosis Syndrome-Associated with Inherited Surfactant Protein-B Deficiency Journal of Pediatrics 1994 125 43 50 8021783
Nogee LM Murphy AM Dietz HC Identification of A Common Mutant Allele in the Pulmonary Surfactant Protein-B (Sp-B) Gene in Patients with Neonatal Alveolar Proteinosis American Journal of Human Genetics 1993 53 93
Nogee LM Hamvas A Murphy AM Dietz HC Molecular Diagnosis of Hereditary Surfactant Protein-B (Sp-B) Deficiency Pediatric Research 1994 35 A243
Nogee LM Wert SE Proffit SA Hull WM Whitsett JA Allelic heterogeneity in hereditary surfactant protein B (SP-B) deficiency American Journal of Respiratory and Critical Care Medicine 2000 161 973 981 10712351
Tredano M van Elburg RM Kaspers AG Zimmermann LJ Houdayer C Aymard P Compound SFTPB 1549C -> GAA (121ins2) and 457delC heterozygosity in severe congenital lung disease and surfactant protein B (SP-B) deficiency Human Mutation 1999 14 502 509 10571948 10.1002/(SICI)1098-1004(199912)14:6<502::AID-HUMU9>3.0.CO;2-C
Tredano M Griese M de Blic J Lorant T Houdayer C Cartault F Analysis of SFTPB in 40 sporadic or familial neonatal and pediatric cases with severe unexplained respiratory distress American Journal of Human Genetics 2002 71 526
Tredano M Griese M de Blic J Lorant T Houdayer C Schumacher S Analysis of 40 sporadic or familial neonatal and pediatric cases with severe unexplained respiratory distress: Relationship to SFTPB American Journal of Medical Genetics Part A 2003 119A 324 339 12784301 10.1002/ajmg.a.20058
Cole FS Hamvas A Rubinstein P King E Trusgnich M Nogee LM Population-based estimates of surfactant protein B deficiency Pediatrics 2000 105 538 541 10699106 10.1542/peds.105.3.538
deMello DE Heyman S Phelps DS Hamvas A Nogee L Cole S Ultrastructure of Lung in Surfactant Protein-B Deficiency American Journal of Respiratory Cell and Molecular Biology 1994 11 230 239 8049084
Weaver TE Conkright JJ Functions of surfactant proteins B and C Annual Review of Physiology 2001 63 555 578 11181967 10.1146/annurev.physiol.63.1.555
Beers MF Hamvas A Moxley MA Gonzales LW Guttentag SH Solarin KO Pulmonary surfactant metabolism in infants lacking surfactant protein B American Journal of Respiratory Cell and Molecular Biology 2000 22 380 391 10696076
Guttentag SH Ballard PL Gonzales LW Solarin KO Beers MF Hamvas A Expression of surfactant components in infants deficient in surfactant protein B Journal of Investigative Medicine 1999 47 172A
Weaver TE Synthesis, processing and secretion of surfactant proteins B and C Biochimica et Biophysica Acta-Molecular Basis of Disease 1998 1408 173 179 10.1016/S0925-4439(98)00066-0
Nogee LM Surfactant protein-B deficiency Chest 1997 111 S129 S135
Nogee LM Genetic mechanisms of surfactant deficiency Biology of the Neonate 2004 85 314 318 15218289 10.1159/000078171
Hamvas A Nogee LM Mallory GB Spray TL Huddleston CB August A Lung transplantation for treatment of infants with surfactant protein B deficiency Journal of Pediatrics 1997 130 231 239 9042125
Hamvas A Nogee LM August A Mallory GB Spray TL Colten HR Surfactant Protein-B Deficiency – Prompt Genetic Diagnosis and Treatment with Lung Transplantation Clinical Research 1994 42 A348
Hamvas A Cole FS deMello DE Moxley M Whitsett JA Colten HR Surfactant Protein-B Deficiency – Antenatal Diagnosis and Prospective Treatment with Surfactant Replacement Journal of Pediatrics 1994 125 356 361 8071741
Tokieda K Iwamoto HS Bachurski C Wert SE Hull WM Ikeda K Surfactant protein-B-deficient mice are susceptible to hyperoxic lung injury American Journal of Respiratory Cell and Molecular Biology 1999 21 463 472 10502556
Tokieda K Ikegami M Wert SE Baatz JE Zou Y Whitsett JA Surfactant protein B corrects oxygen-induced pulmonary dysfunction in heterozygous surfactant protein B-deficient mice Pediatric Research 1999 46 708 714 10590028
Klein JM Thompson MW Snyder JM George TN Whitsett JA Bell EF Transient surfactant protein B deficiency in a term infant with severe respiratory failure Journal of Pediatrics 1998 132 244 248 9506635
Dunbar AE Wert SE Ikegami M Whitsett JA Hamvas A White FV Prolonged survival in hereditary surfactant protein B (SP-B) deficiency associated with a novel splicing mutation Pediatric Research 2000 48 275 282 10960490
Ballard PL Nogee LM Beers MF Ballard RA Planer BC Polk L Partial Deficiency of Surfactant Protein-B in An Infant with Chronic Lung-Disease Pediatrics 1995 96 1046 1052 7491219
Melton KR Nesslein LL Ikegami M Tichelaar JW Clark JC Whitsett JA SP-B deficiency causes respiratory failure in adult mice American Journal of Physiology-Lung Cellular and Molecular Physiology 2003 285 L543 L549 12639841
Hallman M Haataja R Marttila R Surfactant proteins and genetic predisposition to respiratory distress syndrome Seminars in Perinatology 2002 26 450 460 12537318
Makri V Hospes B Stoll-Becker S Borkhardt A Gortner L Polymorphisms of surfactant protein B encoding gene: modifiers of the course of neonatal respiratory distress syndrome? European Journal of Pediatrics 2002 161 604 608 12424586 10.1007/s00431-002-1046-1
Haataja R Hallman M Surfactant proteins as genetic determinants of multifactorial pulmonary diseases Annals of Medicine 2002 34 324 333 12452477 10.1080/078538902320772089
Floros J Veletza SV Kotikalapudi P Krizkova L Karinch AM Friedman C Dinucleotide repeats in the human surfactant protein-B gene and respiratory distress syndrome Biochem J 1995 305 583 590 7832777
Kala P Ten Have T Nielsen H Dunn M Floros J Association of pulmonary surfactant protein A (SP-A) gene and respiratory distress syndrome: Interaction with SP-B Pediatric Research 1998 43 169 177 9475280
Haataja R Ramet M Marttila R Hallman M Surfactant proteins A and B as interactive genetic determinants of neonatal respiratory distress syndrome Human Molecular Genetics 2000 9 2751 2760 11063734 10.1093/hmg/9.18.2751
Lin Z Pearson C Chincilli P Pietschmann SM Pison U Floros J Polymorhisms of human SP-A, SP-B and SP-D genes: associations of SP-B Thr131Ile with ARDS Clin Genet 2000 58 181 191 11076040 10.1034/j.1399-0004.2000.580305.x
Marttila R Haataja R Guttentag S Hallman M Surfactant protein A and B genetic variants in respiratory distress syndrome in singletons and twins American Journal of Respiratory and Critical Care Medicine 2003 168 1216 1222 12947025 10.1164/rccm.200304-524OC
Phelps DS Floros J Localization of Pulmonary Surfactant Proteins Using Immunohistochemistry and Tissue Insitu Hybridization Experimental Lung Research 1991 17 985 995 1769356
Glasser SW Korfhagen TR Perme CM Pilotmatias TJ Kister SE Whitsett JA 2 Sp-C Genes Encoding Human Pulmonary Surfactant Proteolipid Journal of Biological Chemistry 1988 263 10326 10331 2839484
Brasch F Johnen G Schauer-Vukasinovic V ten Brinke A Kapp N Napsin A is involved in the processing of proSP-C Am J Respir Crit Care Med 2001 163 560A
Glasser SW Detmer EA Ikegami M Na CL Stahlman MT Whitsett JA Pneumonitis and emphysema in sp-C gene targeted mice Journal of Biological Chemistry 2003 278 14291 14298 12519727 10.1074/jbc.M210909200
Beers MF Mulugeta S Surfactant Protein C Biosynthesis and Its Emerging Role in Conformational Lung Disease Annual Review of Physiology 2004 19
Pantelidis P Veeraraghavan S du Bois RM Surfactant gene polymorphisms and interstitial lung diseases Respiratory Research 2001 3
Tredano M Griese M Brasch F Schumacher S de Blic J Marque S Mutation of SFTPC in infantile pulmonary alveolar proteinosis with or without fibrosing lung disease American Journal of Medical Genetics Part A 2004 126A 18 26 15039969 10.1002/ajmg.a.20670
Thomas AQ Lane K Phillips J Prince M Markin C Speer M Heterozygosity for a surfactant protein C gene mutation associated with usual interstitial pneumonitis and cellular nonspecific interstitial pneumonitis in one kindred American Journal of Respiratory and Critical Care Medicine 2002 165 1322 1328 11991887 10.1164/rccm.200112-123OC
Stevens PA Pettenazzo A Brasch F Mulugeta S Baritussio A Ochs M Nonspecific Interstitial Pneumonia, Alveolar Proteinosis, and Abnormal Proprotein Trafficking Resulting from a Spontaneous Mutation in the Surfactant Protein C Gene Pediatric Research 2004 19
Nogee LM Dunbar AE Wert SE Askin F Hamvas A Whitsett JA A mutation in the surfactant protein C gene associated with familial interstitial lung disease New England Journal of Medicine 2001 344 573 579 11207353 10.1056/NEJM200102223440805
Nogee LM Dunbar AE Wert S Askin F Hamvas A Whitsett JA Mutations in the surfactant protein C gene associated with interstitial lung disease Chest 2002 121 20S 21S 11893657 10.1378/chest.121.3_suppl.20S
Hamvas A Nogee L Beers M Gonzales L White F Wert S Abnormal surfactant composition and function associated with a mutation in the surfactant protein-C (SP-C) gene Pediatric Research 2003 53 573A 12612217 10.1203/01.PDR.0000055770.07236.98
Hamvas A Nogee LM White FV Schuler P Hackett BP Huddleston CB Progressive lung disease and surfactant dysfunction with a deletion in surfactant protein C gene American Journal of Respiratory Cell and Molecular Biology 2004 30 771 776 14656744 10.1165/rcmb.2003-0323OC
Cameron HS Hamvas A Whitsett JA Wert SE Nogee LM A common mutation in the surfactant protein C (SP-C) gene responsible for lung disease Pediatric Research 2003 53 570A
Brasch F Griese M Tredano M Johnen G Ochs M Rieger C Interstitial lung disease in a baby with a de novo mutation in the SFTPC gene European Respiratory Journal 2004 24 30 39 15293602 10.1183/09031936.04.00000104
Bahuau M Griese M Marque S Brasch F Schumacher S de Blic J Mutation of SFTPC in infantile pulmonary alveolar proteinosis with or without fibrosing lung disease American Journal of Human Genetics 2003 73 584
Amin RS Wert SE Baughman RP Tomashefski JF Nogee LM Brody AS Surfactant protein deficiency in familial interstitial lung disease Journal of Pediatrics 2001 139 85 92 11445799 10.1067/mpd.2001.114545
Nogee L Shulenin S Annilo T Wert S Ikegami M Whitsett J Mutations in the ABCA3 Gene Are a Frequent Cause of Fatal Neonatal Surfactant Deficiency. American Journal of Respiratory and Critical Care Medicine International Conference Abstracts 2004 [B85] 2004
Bridges JP Wert SE Nogee LM Weaver TE Expression of a human surfactant protein C mutation associated with interstitial lung disease disrupts lung development in transgenic mice Journal of Biological Chemistry 2003 278 52739 52746 14525980 10.1074/jbc.M309599200
Wang WJ Mulugeta S Russo SJ Beers MF Deletion of exon 4 from human surfactant protein C results in aggresome formation and generation of a dominant negative Journal of Cell Science 2003 116 683 692 12538769 10.1242/jcs.00267
Allen RT Hunter WJ Agrawal DK Morphological and biochemical characterization and analysis of apoptosis Journal of Pharmacological and Toxicological Methods 1997 37 215 228 9279777 10.1016/S1056-8719(97)00033-6
Sutherland LM Edwards YS Murray AW Alveolar type II cell apoptosis Comparative Biochemistry and Physiology A-Molecular & Integrative Physiology 2001 129 267 285 10.1016/S1095-6433(01)00323-3
Stillwell PC Norris DG Oconnell EJ Rosenow EC Weiland LH Harrison EG Desquamative Interstitial Pneumonitis in Children Chest 1980 77 165 171 7353410
King TE Diagnostic Advances in Idiopathic Pulmonary Fibrosis Chest 1991 100 238 241 2060352
Katzenstein ALA Myers JL Idiopathic pulmonary fibrosis – Clinical relevance of pathologic classification American Journal of Respiratory and Critical Care Medicine 1998 157 1301 1315 9563754
Percopo S Cameron HS Nogee LM Pettinato G Montella S Santamaria F Variable phenotype associated with SP-C gene mutations: fatal case with the I73T mutation European Respiratory Journal 2004 24 1072 1073 15572558 10.1183/09031936.04.00092304
Glasser SW Korfhagen TR Perme CM Pilotmatias TJ Kister SE Whitsett JA 2 Sp-C Genes Encoding Human Pulmonary Surfactant Proteolipid Journal of Biological Chemistry 1988 263 10326 10331 2839484
Warr RG Hawgood S Buckley DI Crisp TM Schilling J Benson BJ Low-Molecular-Weight Human Pulmonary Surfactant Protein (Sp5) – Isolation, Characterization, and Cdna and Amino-Acid-Sequences Proceedings of the National Academy of Sciences of the United States of America 1987 84 7915 7919 3479771
Hatzis D Deiter G deMello DE Floros J Human Surfactant Protein-C – Genetic Homogeneity and Expression in Rds – Comparison with Other Species Experimental Lung Research 1994 20 57 72 8181452
Lahti M Marttila R Hallman M Surfactant protein C gene variation in the Finnish population – association with perinatal respiratory disease European Journal of Human Genetics 2004 12 312 320 14735158 10.1038/sj.ejhg.5201137
Lawson WE Grant SW Ambrosini V Womble KE Dawson EP Lane KB Genetic mutations in surfactant protein C are a rare cause of sporadic cases of IPF Thorax 2004 59 977 980 15516475 10.1136/thx.2004.026336
Nightingale JA Rogers DF Hart LA Kharitonov SA Chung KF Barnes PJ Effect of inhaled endotoxin on induced sputum in normal, atopic, and atopic asthmatic subjects Thorax 1998 53 563 571 9797755
Augusto LA Synguelakis M Espinassous Q Lepoivre M Johansson A Chaby R Cellular antiendotoxin activities of lung surfactant protein C in lipid vesicles American Journal of Respiratory and Critical Care Medicine 2003 168 335 341 12773323 10.1164/rccm.200212-1440OC
Augusto LA Synguelakis M Johansson J Pedron T Girard R Chaby R Interaction of pulmonary surfactant protein C with CD14 and lipopolysaccharide Infection and Immunity 2003 71 61 67 12496149 10.1128/IAI.71.1.61-67.2003
Perera PY Mayadas TN Takeuchi O Akira S Zaks-Zilberman M Goyert SM CD11b/CD18 acts in concert with CD14 and Toll-like receptor (TLR) 4 to elicit full lipopolysaccharide and taxol-inducible gene expression Journal of Immunology 2001 166 574 581
Lutz C Carney D Finck C Picone A Gatto LA Paskanik A Aerosolized surfactant improves pulmonary function in endotoxin-induced lung injury American Journal of Respiratory and Critical Care Medicine 1998 158 840 845 9731014
Sanchez-Pulido L Devos D Valencia A BRICHOS: a conserved domain in proteins associated with dementia, respiratory distress and cancer Trends in Biochemical Sciences 2002 27 329 332 12114016 10.1016/S0968-0004(02)02134-5
Johansson J Molecular determinants for amyloid fibril formation: lessons from lung surfactant protein C Swiss Medical Weekly 2003 133 275 282 12844270
Johansson J Membrane properties and amyloid fibril formation of lung surfactant protein Biochemical Society Transactions 2001 29 601 606 11498036 10.1042/BST0290601
Gustafsson M Thyberg J Naslund J Eliasson E Johansson J Amyloid fibril formation by pulmonary surfactant protein C Febs Letters 2004 31 138 142
Schmitz G Kaminski WE ABC transporters and cholesterol metabolism Frontiers in Bioscience 2001 6 D505 D514 11229879
Bodzioch M Orso E Klucken T Langmann T Bottcher L Diederich W The gene encoding ATP-binding cassette transporter 1 is mutated in Tangier disease Nature Genetics 1999 22 347 351 10431237 10.1038/11914
Yamano G Funahashi H Kawanami O Zhao LX Ban N Uchida Y ABCA3 is a lamellar body membrane protein in human lung alveolar type II cells Febs Letters 2001 508 221 225 11718719 10.1016/S0014-5793(01)03056-3
Zhou JM You Y Ryan AJ Mallampalli RK Upregulation of surfactant synthesis triggers ABCA1-mediated basoateral phospholipid efflux Journal of Lipid Research 2004 45 1758 1767 15210848 10.1194/jlr.M400179-JLR200
McNeish J Aiello RJ Guyot D Turi T Gabel C Aldinger C High density lipoprotein deficiency and foam cell accumulation in mice with targeted disruption of ATP-binding cassette transporter-1 Proceedings of the National Academy of Sciences of the United States of America 2000 97 4245 4250 10760292 10.1073/pnas.97.8.4245
Costa RH Kalinichenko VV Lim L Transcription factors in mouse lung development and function American Journal of Physiology-Lung Cellular and Molecular Physiology 2001 280 L823 L838 11290504
Warburton D Bellusci S Del Moral PM Kaartinen V Lee M Tefft D Growth factor signaling in lung morphogenetic centers: automaticity, stereotypy and symmetry Respir Res 2003 4 5 12818006 10.1186/1465-9921-4-5
Kumar VH Ryan RM Growth factors in the fetal and neonatal lung Frontiers in Bioscience 2004 9 464 480 14766383
Bohinski RJ DiLauro R Whitsett JA The Lung-Specific Surfactant Protein-B Gene Promoter Is A Target for Thyroid Transcription Factor-1 and Hepatocyte Nuclear Factor-3, Indicating Common Factors for Organ-Specific Gene-Expression Along the Foregut Axis Molecular and Cellular Biology 1994 14 5671 5681 8065304
Ikeda K Clark JC ShawWhite JR Stahlman MT Boutell CJ Whitsett JA Gene Structure and Expression of Human Thyroid Transcription Factor-I in Respiratory Epithelial-Cells Journal of Biological Chemistry 1995 270 8108 8114 7713914 10.1074/jbc.270.44.26460
Stahlman MT Gray ME Whitsett JA Expression of thyroid transcription factor-1 (TTF-1) in fetal and neonatal human lung Journal of Histochemistry & Cytochemistry 1996 44 673 678 8675988
Yang HH Lu MM Zhang LL Whitsett JA Morrisey EE GATA6 regulates differentiation of distal lung epithelium Development 2002 129 2233 2246 11959831 10.1242/dev.00150
Shaw-White JR Bruno MD Whitsett JA GATA-6 activates transcription of thyroid transcription factor-1 Journal of Biological Chemistry 1999 274 2658 2664 9915795 10.1074/jbc.274.5.2658
Wan HJ Xu Y Ikegami M Stahlman MT Kaestner KH Ang SL Foxa2 is required for transition to air breathing at birth Proceedings of the National Academy of Sciences of the United States of America 2004 101 14449 14454 15452354 10.1073/pnas.0404424101
Volpe MV Vosatka RJ Nielsen HC Hoxb-5 control of early airway formation during branching morphogenesis in the developing mouse lung Biochimica et Biophysica Acta-General Subjects 2000 1475 337 345 10.1016/S0304-4165(00)00087-8
Massague J TGF-beta signal transduction Annual Review of Biochemistry 1998 67 753 791 9759503 10.1146/annurev.biochem.67.1.753
Li CG Zhu NL Tan RC Ballard PL Derynck R Minoo P Transforming growth factor-beta inhibits pulmonary surfactant protein B gene transcription through SMAD3 interactions with NKX2.1 and HNF-3 transcription factors Journal of Biological Chemistry 2002 277 38399 38408 12161428 10.1074/jbc.M203188200
Cao GT Ni XH Jiang M Ma YS Cheng HP Guo LC Molecular cloning and characterization of a novel human cAMP response element-binding (CREB) gene (CREB4) Journal of Human Genetics 2002 47 373 376 12111373 10.1007/s100380200053
McKenna NJ Lanz RB O'Malley BW Nuclear receptor coregulators: Cellular and molecular biology Endocrine Reviews 1999 20 321 344 10368774 10.1210/er.20.3.321
Zhou L Dey CR Wert SE Yan C Costa RH Whitsett JA Hepatocyte nuclear factor-3 beta limits cellular diversity in the developing respiratory epithelium and alters lung morphogenesis in vivo Developmental Dynamics 1997 210 305 314 9389455 10.1002/(SICI)1097-0177(199711)210:3<305::AID-AJA10>3.0.CO;2-9
Moens CB Auerbach AB Conlon RA Joyner AL Rossant J A Targeted Mutation Reveals A Role for N-Myc in Branching Morphogenesis in the Embryonic Mouse Lung Genes & Development 1992 6 691 704 1577267
Whitsett JA Zhou L Use of transgenic mice to study autocrine-paracrine signaling in lung morphogenesis and differentiation Clinics in Perinatology 1996 23 753 8982569
Wan HJ Kaestner KH Ang SL Ikegami M Finkelman FD Stahlman MT Foxa2 regulates alveolarization and goblet cell hyperplasia Development 2004 131 953 964 14757645 10.1242/dev.00966
Zeng X Yutzey KE Whitsett JA Thyroid transcription factor-1, hepatocyte nuclear factor-3 beta and surfactant protein A and B in the developing chick lung Journal of Anatomy 1998 193 399 408 9877295 10.1046/j.1469-7580.1998.19330399.x
Reynolds PR Mucenski ML Whitsett JA Thyroid transcription factor (TTF)-1 regulates the expression of midkine (MK) during lung morphogenesis Developmental Dynamics 2003 227 227 237 12761850 10.1002/dvdy.10304
Shaw-White JR Bruno MD Whitsett JA GATA-6 activates transcription of thyroid transcription factor-1 Journal of Biological Chemistry 1999 274 2658 2664 9915795 10.1074/jbc.274.5.2658
Bruno MD Bohinski RJ Huelsman KM Whitsett JA Korfhagen TR Lung Cell-Specific Expression of the Murine Surfactant Protein-A (Sp-A) Gene Is Mediated by Interactions Between the Sp-A Promoter and Thyroid Transcription Factor-I Journal of Biological Chemistry 1995 270 6531 6536 7896788 10.1074/jbc.270.12.6531
Bohinski RJ Dilauro R Whitsett JA The Lung-Specific Surfactant Protein-B Gene Promoter Is A Target for Thyroid Transcription Factor-1 and Hepatocyte Nuclear Factor-3, Indicating Common Factors for Organ-Specific Gene-Expression Along the Foregut Axis Molecular and Cellular Biology 1994 14 5671 5681 8065304
Bingle CD Hackett BP Moxley M Longmore W Gitlin JD Role of Hepatocyte Nuclear Factor-3-Alpha and Hepatocyte Nuclear Factor-3-Beta in Clara Cell Secretory Protein Gene-Expression in the Bronchiolar Epithelium Biochemical Journal 1995 308 197 202 7755566
Ikeda K ShawWhite JR Wert SE Whitsett JA Hepatocyte nuclear factor 3 activates transcription of thyroid transcription factor 1 in respiratory epithelial cells Molecular and Cellular Biology 1996 16 3626 3636 8668179
Mendelson CR Alcorn JL Gao E The Pulmonary Surfactant Protein Genes and Their Regulation in Fetal Lung Seminars in Perinatology 1993 17 223 232 8140435
Kimura S Hara Y Pineau T FernandezSalguero P Fox CH Ward JM The T/ebp null mouse thyroid-specific enhancer-binding protein is essential for the organogenesis of the thyroid, lung, ventral forebrain, and pituitary Genes & Development 1996 10 60 69 8557195
Minoo P Su GS Drum H Bringas P Kimura S Defects in tracheoesophageal and lung morphogenesis in Nkx2.1(-/-) mouse embryos Developmental Biology 1999 209 60 71 10208743 10.1006/dbio.1999.9234
Breedveld GJ van Dongen JWF Danesino C Guala A Percy AK Dure LS Mutations in TITF-1 are associated with benign hereditary chorea Human Molecular Genetics 2002 11 971 979 11971878 10.1093/hmg/11.8.971
Devriendt K Vanhole C Matthijs G de Zegher F Deletion of thyroid transcription factor-1 gene in an infant with neonatal thyroid dysfunction and respiratory failure New England Journal of Medicine 1998 338 1317 1318 9565498 10.1056/NEJM199804303381817
Iwatani N Mabe H Devriendt K Kodama M Miike T Deletion of NKX2.1 gene encoding thyroid transcription factor-1 in two siblings with hypothyroidism and respiratory failure Journal of Pediatrics 2000 137 272 276 10931427 10.1067/mpd.2000.107111
Krude H Schutz B Biebermann H von Moers A Schnabel D Neitzel H Choreoathetosis, hypothyroidism, and pulmonary alterations due to human NKX2-1 haploinsufficiency Journal of Clinical Investigation 2002 109 475 480 11854319 10.1172/JCI200214341
Nielsen HC Martin A Volpe MV Hatzis D Vosatka RJ Growth factor control of growth and epithelial differentiation in embryonic lungs Biochemical and Molecular Medicine 1997 60 38 48 9066980 10.1006/bmme.1996.2560
Miettinen PJ Berger JE Meneses J Phung Y Pedersen RA Werb Z Epithelial Immaturity and Multiorgan Failure in Mice Lacking Epidermal Growth-Factor Receptor Nature 1995 376 337 341 7630400 10.1038/376337a0
Li CG Zhu NL Tan RC Ballard PL Derynck R Minoo P Transforming growth factor-beta inhibits pulmonary surfactant protein B gene transcription through SMAD3 interactions with NKX2.1 and HNF-3 transcription factors Journal of Biological Chemistry 2002 277 38399 38408 12161428 10.1074/jbc.M203188200
Viitala J Carlsson SR Siebert PD Fukuda M Molecular-Cloning of Cdnas Encoding Lamp-A, A Human Lysosomal Membrane Glycoprotein with Apparent Mr Almost-Equal-to 120,000 Proceedings of the National Academy of Sciences of the United States of America 1988 85 3743 3747 3131762
Konecki DS Foetisch K Schlotter M Lichterkonecki U Complete Cdna Sequence of Human Lysosome-Associated Membrane Protein-2 Biochemical and Biophysical Research Communications 1994 205 1 5 7999007 10.1006/bbrc.1994.2620
Metzelaar MJ Wijngaard PLJ Peters PJ Sixma JJ Nieuwenhuis HK Clevers HC Cd63-Antigen – A Novel Lysosomal Membrane Glycoprotein, Cloned by A Screening-Procedure for Intracellular Antigens in Eukaryotic Cells Journal of Biological Chemistry 1991 266 3239 3245 1993697
Akasaki K Nakamura N Tsukui N Yokota S Murata S Katoh R Human dendritic cell lysosome-associated membrane protein expressed in lung type II pneumocytes Archives of Biochemistry and Biophysics 2004 425 147 157 15111122 10.1016/j.abb.2004.02.042
Salaun B Saint-Vis B Pacheco N Pacheco Y Riesler A Isaac S CD208/dendritic cell-lysosomal associated membrane protein is a marker of normal and transformed type II pneumocytes American Journal of Pathology 2004 164 861 871 14982840
Abonyo BO Gou DM Wang PC Narasaraju T Wang ZX Liu L Syntaxin 2 and SNAP-23 are required for regulated surfactant seeretion Biochemistry 2004 43 3499 3506 15035620 10.1021/bi036338y
Huizing M Anikster Y Gahl WA Hermansky-Pudlak syndrome and Chediak-Higashi syndrome: Disorders of vesicle formation and trafficking Thrombosis and Haemostasis 2001 86 233 245 11487012
Avila NA Brantly M Premkumar A Huizing M Dwyer A Gahl WA Hermansky-Pudlak syndrome: Radiography and CT of the chest compared with pulmonary function tests and genetic studies American Journal of Roentgenology 2002 179
Gahl WA Shotelersuk V Brantly M Huizing M Dell'Angelica EC Bonifacino JS Hermansky-Pudlak syndrome (HPS) as a genetically heterogeneous disease: Clinical, molecular, and cell biological aspects Pediatric Research 1999 45 138A 9890622
Nakatani Y Nakamura N Sano J Inayama Y Kawano N Yamanaka S Interstitial pneumonia in Hermansky-Pudlak syndrome: significance of florid foamy swelling/degeneration (giant lamellar body degeneration) of type-2 pneumocytes Virchows Archiv 2000 437 304 313 11037352 10.1007/s004280000241
| 15819986 | PMC1090616 | CC BY | 2021-01-04 16:36:27 | no | Respir Res. 2005 Apr 8; 6(1):32 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-32 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-331582320810.1186/1465-9921-6-33ResearchAirway allergen exposure stimulates bone marrow eosinophilia partly via IL-9 Sitkauskiene Brigita [email protected]ådinger Madeleine [email protected] Apostolos [email protected] Anna-Karin [email protected] Raimundas [email protected]ötvall Jan [email protected] The Lung Pharmacology Group, Department of Respiratory Medicine and Allergology, Institute of Internal Medicine, Göteborg University, Guldhedsgatan 10A, 413 46 Gothenburg, Sweden2 Department of Pulmonology and Immunology, Kaunas University of Medicine, Eiveniu 2, 50009 Kaunas, Lithuania3 Lab of Pulmonology, Institute for Biomedical Research, Kaunas University of Medicine, Eiveniu 4, 50009 Kaunas, Lithuania2005 11 4 2005 6 1 33 33 25 11 2004 11 4 2005 Copyright © 2005 Sitkauskiene et al; licensee BioMed Central Ltd.2005Sitkauskiene 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
Interleukin (IL)-9 is a Th2-derived cytokine with pleiotropic biological effects, which recently has been proposed as a candidate gene for asthma and allergy. We aimed to evaluate the therapeutic effect of a neutralizing anti-IL-9 antibody in a mouse model of airway eosinophilic inflammation and compared any such effect with anti-IL-5 treatment.
Methods
OVA-sensitized Balb/c mice were intraperitoneally pretreated with a single dose (100 μg) of an anti-mouse IL-9 monoclonal antibody (clone D9302C12) or its vehicle. A third group was given 50 μg of a monoclonal anti-mouse IL-5 antibody (TRFK-5) or its vehicle. Animals were subsequently exposed to OVA on five days via airways. Newly produced eosinophils were labelled using 5-bromo-2'-deoxyuridine (BrdU). BrdU+ eosinophils and CD34+ cell numbers were examined by immunocytochemistry. After culture and stimulation with OVA or PMA+IC, intracellular staining of IL-9 in bone marrow cells from OVA-exposed animals was measured by Flow Cytometry. The Mann-Whitney U-test was used to determine significant differences between groups.
Results
Anti-IL-9 significantly reduced bone marrow eosinophilia, primarily by decrease of newly produced (BrdU+) and mature eosinophils. Anti-IL-9 treatment also reduced blood neutrophil counts, but did not affect BAL neutrophils. Anti-IL-5 was able to reduce eosinophil numbers in all tissue compartments, as well as BrdU+ eosinophils and CD34+ progenitor cells, and in all instances to a greater extent than anti-IL-9. Also, FACS analysis showed that IL-9 is over-expressed in bone marrow CD4+ cells after allergen exposure.
Conclusions
Our data shows that a single dose of a neutralizing IL-9 antibody is not sufficient to reduce allergen-induced influx of newly produced cells from bone marrow to airways. However, in response to allergen, bone marrow cells over-express IL-9. This data suggest that IL-9 may participate in the regulation of granulocytopoiesis in allergic inflammation.
==== Body
Background
Airway eosinophilic inflammation is a predominant feature of asthma. Eosinophils are believed to be involved in several features of asthma through the release of cationic granule proteins, reactive oxygen radicals, a variety of cytokines and bronchoconstrictive mediators [1-4]. The regulation of the eosinophil in asthma is considered to be orchestrated by the Th2-cell, which can release a range of Th2-cytokines, particularly interleukin (IL)-5 [5-8]. These cytokines regulate eosinophil growth, differentiation and activation.
IL-9 is another Th2-derived cytokine with pleiotropic biological effects on various types of cells. It acts as a growth factor for T cells, a maturation factor for B cells, and as a proliferation and differentiation factor for mast cells and hematopoietic progenitors [9-13]. Recently, it has been suggested that IL-9 might play a role in allergy [14-22]. Evidence from both murine and human mapping studies shows that IL-9 is a candidate gene for asthma [17,18]. Moreover, the expression of IL-9 and its receptor is increased in allergic asthma [19-21]. It was also shown that eosinophils have the capacity to synthesize and release IL-9 [23]. Evidence, that in vitro, IL-9 prolongs eosinophil survival, as well as IL-5 mediated differentiation and maturation [24,25], suggests that IL-9 may potentiate eosinophil function in vivo. In vivo, locally instilled IL-9 increases eosinophil count in BAL fluid [16]. Moreover, IL-9 transgenic mice were found to display significantly enhanced eosinophilic inflammation [15]. Eosinophils develop from CD34+ progenitor cells, and it has been suggested that IL-9 alone may upregulate the expression of IL-5 Rα on human CD34+ cord blood progenitor cells [24].
Despite increasing evidence, that IL-9 may be involved in allergy, only few studies have been performed using an IL-9 antagonist to test its possible therapeutic efficacy in the treatment of allergic inflammation. Furthermore, the possible regulatory effect of IL-9 on newly produced eosinophils and CD34+ progenitor cells has not been documented. In our study, we aimed to examine the effect of a neutralizing monoclonal anti-IL-9 antibody on different tissue compartments in vivo, in a model of allergic eosinophilic inflammation, and relate this effect to newly produced eosinophils (labelled with 5-bromo-2'-deoxyuridine; BrdU) and CD34+ cells. Furthermore, we compared the effect of anti-IL-9 with that of anti-IL-5.
Methods
Animals
Male Balb/c mice, 5 to 6 weeks old, were obtained from B&K Universal AB (Sollentuna, Sweden). All animals were maintained under conventional animal housing conditions and provided with food and water ad libitum. The experimental protocol was approved by the Animal Ethics Committee in Gothenburg, Sweden.
Sensitization and allergen exposure protocol
All mice were sensitized by intraperitoneal (i.p.) injection with 8 μg of ovalbumin (OVA; Sigma-Aldrich Sweden AB, Tyresö, Sweden) adsorbed to 4 mg of aluminum hydroxide (Al(OH)3; Sigma) in 0.5 ml of phosphate-buffered saline (PBS). A booster dose of the OVA-Al(OH)3 mixture was given on a second occasion, five days after the first injection. Starting eight days after the second sensitization, the animals were briefly anesthetized using aerosolized Isoflurane (Baxter, Deerfield, Ill), and exposed to 100 μg OVA in 25 μl of PBS by intranasal (i.n.) administration on five consecutive days.
Pretreatment with anti-cytokine antibodies and treatment with BrdU
Animals were pretreated i.p. with a single dose (100 μg/animal) of purified hamster anti-mouse IL-9 monoclonal antibody (clone D9302C12; Pharmingen, San Diego, CA, USA) or its isotype control, purified hamster IgG (clone G235-2356; Pharmingen). In parallel groups, animals were treated either with a single dose (50 μg/animal) of a monoclonal antibody to mouse IL-5 (clone TRFK-5; R&D Systems Europe Ltd., Barton Lane, Abingdon, UK) or its isotype control, rat IgG1 (clone R3-34; Pharmingen) in 0.5 ml of PBS, 30 min. before the OVA exposure. Secondly, the combined treatment of anti-IL-9 and anti-IL-5 antibodies was tested. Animals were treated with both anti-IL-9 (100 μg/animal) and anti-IL-5 (50 μg/animal), either anti-IL-9 (100 μg/animal) and rat IgG1 (50 μg/animal), or anti-IL-5 (50 μg/animal) and purified hamster IgG (100 μg/animal), or both isotype controls together (100 μg of purified hamster IgG and 50 μg of rat IgG1). Newly produced inflammatory cells were labelled with a thymidine analogue, 5'-bromo-2'-deoxyuridine (BrdU; Boehringer Mannheim Scandinavia AB, Bromma, Sweden), at a dose of 1 mg in 0.25 ml of PBS i.p. twice (7 h apart) on day one and three of allergen exposure (total dose 4 mg/animal).
Cell collection and samples processing
Samples were collected 24 h after the last allergen exposure. The animals were anesthetized with an i.p. mixture of xylazine (130 mg/kg) and ketamine (670 mg/kg). When in adequately deep anaesthesia, the chest was opened and samples of blood, bronchoalveolar lavage fluid (BALf) and bone marrow were taken.
Blood was obtained by penetration of the right ventricle of the heart with a needle. BAL was performed through the trachea with a cannula by instillation of 0.25 and 0.2 ml of ice-cold PBS. BALf of each animal was pooled; approximately 0.4 ml of the instilled fluid was consistently recovered. Finally, one femur was excised and cut at the epiphyses. Bone marrow cells were removed by perfusion of the femur with 2 ml of PBS. BALf and bone marrow cell suspension were kept on ice until further processing.
Cytospin of blood was obtained by taking 200 μl of blood and mixing it with 800 μl 2 mM EDTA (Sigma) in PBS. The red blood cells were lysed in 0.1% potassium bicarbonate and 0.83% ammonium chloride for 15 min. at 4°C as described previously [8]. The white blood cells were re-suspended in PBS with 0.03% bovine serum albumin (BSA; Sigma). Serum was obtained from the remaining volume of blood by centrifugation at 3000 rpm for 10 min. at 4°C.
BAL and bone marrow cell suspension were centrifuged at 1000 rpm for 10 min. at 4°C. Supernatants were aspirated and cells were re-suspend in PBS with 0.03% BSA. The total number of cells in BAL, bone marrow and blood was determined using standard hematological procedures. Cytospins of BALf, bone marrow and blood cells were prepared and stained with May-Grünwald-Giemsa for differential cell counts. Cell differentiation was determined by counting 300–400 cells using a light microscope (Zeiss Axioplan 2; Carl Zeiss, Jena, Germany). The cells were identified by standard morphological criteria, and bone marrow mature and immature eosinophils were determined by nuclear morphology, May-Grünwald-Giemsa staining properties, and cytoplasmic granulation, as previously described [7,8]. Cytospin preparations for immunocytochemistry were air-dried and stored at -80°C until further examination.
Immunocytochemical detection of BrdU-labelled eosinophils
Intranuclear BrdU was detected in BAL, bone marrow and blood cytospins using a mouse monoclonal antibody against BrdU. Cytospin preparations taken out of the freezer were fixed overnight in 4% paraformaldehyde, and subsequently washed with tris-buffer saline (TBS) and subjected to digestion in trypsin (No.232-650-8; Sigma) diluted in distillated H2O at 37°C for 20 min. The slides were then incubated in 4 M HCl for 15 min. to denature the DNA, followed by neutralization in Holmes Borate-Borax buffer (1.24% H3BO3 in distillated H2O, pH 8.5) for 10 min. The samples were treated with peroxidase blocking solution (No. S2023; DAKO) for 1 h. Non-specific binding sites were blocked with 5% normal rabbit serum (No. X0902; DAKO) for 15 min. Subsequently, the slides were incubated with 2.5 μg/ml rat anti-BrdU antibody (clone BU1/75; No. MAS 250p; Harlan-Sera Lab) or isotype control, rat IgG2a (clone R35-95; Pharmingen) in incubation buffer (0.5% BSA/TBS) for 1 h. After washing in 0.05% Tween/TBS and TBS, the slides were incubated with 1:50 dilution of rabbit F(ab')2 anti-rat Ig-HRP (No.6130-05, SBA) and 2% normal mouse serum (No.X0910; DAKO) for 1 hour. After further washing, the staining with DAB substrate-chromogen system (No. K3466; DAKO) was developed for 10–15 min. by monitoring in microscope. The slides were counterstained with Mayer's hematoxylin for 30 sec. and eosin for 2 min., dehydrated and mounted in Mountex. All slides were evaluated on light microscope in random fields of view. Cells with any nuclear brown staining together with the pink staining in cytoplasma were counted as BrdU-labelled (BrdU+) eosinophils.
Immunocytochemical detection of CD34+ progenitor cells
Cytospin preparations from blood, BAL and bone marrow were fixed in 2% formaldehyde for 10 min. All incubations were performed at room temperature (RT). After washing in TBS, endogenous biotin was blocked with DAKO Biotin Blocking system. The preparations were incubated with a monoclonal biotinylated rat anti-mouse CD34 antibody (0.5 μg/ml; clone RAM34, BD Biosciences) or its isotype control (biotinylated Rat IgG2a, BD Biosiences) for 2 h. After further washings, the preparations were incubated with Streptavidin Alkaline-Phosphatase (DAKO) for 45 min. Bound antibodies were visualized with Vector Red Alkaline Phosphatase Substrate kit (Vector Laboratories Inc. Burlingame, CA, USA). The preparations were washed in Tris-HCl buffer (100 mM, pH 8.2–8.5), rinsed in distilled H2O and counterstained with Mayer's hematoxylin for 30 sec., dehydrated and mounted in Mountex. Three hundred cells were counted in random fields of view.
Measurement of IL-5 and IL-9
IL-5 concentration in serum and BALf from anti-IL-9 or its control treated animals was measured using a commercial murine IL-5 enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Inc, Abingdon, UK) according to the manufacturer's instruction. The detection limit for IL-5 was 5 pg/ml.
Quantification of IL-9 was measured in BALf, serum and fresh bone marrow supernatant (the femur was perfused twice with 1 ml of PBS) from OVA-sensitized and exposed for 5 days to OVA or PBS (control) animals by ELISA method. ELISA plate (Maxisorp F96, NUNC) was incubated with 100 μl of purified rat anti-mouse IL-9 antibody (No.551218, BD) at a concentration of 1 μg/ml in coating buffer (0.1 M NaHCO3, pH 9.5) at 4°C for 18 h. All incubations were performed in dark. After washings with PBS containing 0.05% Tween 20, non-specific bindings were blocked with assay diluent (10% fetal calf serum, FCS, in PBS) for 1 h at RT. After further washings, bone marrow supernatant, BALf and serum samples (100 μl of each) were incubated in duplicate at 4°C for 18 h. A standard curve was generated by using serial dilutions of recombinant mouse IL-9 (No.551867, BD). After washing, 100 μl of biotinylated anti-mouse IL-9 antibody (No. 554473, BD) at a concentration 0.5 μg/ml in the assay diluent was added and incubated at RT for 1 h. After further washings, the plates were incubated with 1:6000 dilution of Extr-Avidin HRP (No. E2886, Sigma) for 30 min. After another washing, 100 μl of TMB substrate solution was added and incubated for 30 min. The reaction was terminated by adding 50 μl of stop solution (1.0 M H2SO4) and assessed with an ELISA reader (type-349; Labsystems, Stockholm, Sweden) at 450 nm. The detection limit for IL-9 was 125 pg/ml.
Culture of bone marrow cells
Bone marrow cells were harvested from OVA-sensitized and exposed for six days animals, after 4 h of the last exposure. Cells were cultured in a 48-well plate in RPMI 1640 culture medium, complemented with 10% FCS, 1% penicillin-streptomycin, 1% sodium pyruvate and 2 mM L-glutamine (all obtained from Sigma) in a concentration of 1 × 106 cells/500 μl of medium at 37°C in an atmosphere containing 5% CO2. Cells were cultured only with plain medium (for negative control), either stimulated with 100 μg/ml of OVA or stimulated with phorbol myristate acetate (PMA) and calcium ionophore (end conc. 2 ng/ml and 1 μg/ml respectively, both obtained from Sigma). After 2 h, Brefeldin A solution was added in a final concentration of 10 μg/ml, and incubated for additional 6 h.
FACS analysis of bone marrow cells for IL-9 intracellular staining
After incubation, bone marrow cells were harvested, washed and double-stained (CD4 surface/ IL-9 intracellular), using a standard saponin protocol [26] with some modifications. Unspecific binding was blocked with 2% mouse sera (Dako) for 15 min. The cells were thereafter incubated with a phycoerythrin-conjugated anti-CD4 (clone H129.19, Pharmingen) and its isotype-matched control for 30 min. at 4°C. Surface immunostained cells were fixed in 4% paraformaldehyde at room temperature for 10 min., followed two washings in 1% FCS/PBS. After resuspention in 2 ml of SAP Buffer (0.1% saponin and 0.05 NaN3, w/v in HBSS, Sigma), the cells were incubated with a biotin-conjugated anti-mouse IL-9 monoclonal antibody (clone D9302C12, Pharmingen) at RT for 40 min., then with streptavidin-FITC for 20 min. Finally, the cells were washed with 1% FCS/PBS, resuspended in the same Buffer, and analyzed using a FACScan flow cytometer (Becton Dickinson, Mountain View, CA). Ten thousand cells were computed in a list mode and analyzed using the CellQuest Software (Becton Dickinson).
Statistics
Results are presented as mean values ± SEMs. The Mann-Whitney U-test was employed for comparison of data between groups. Value of p < 0.05 was considered as statistically significant.
Results
Cellular changes in BAL
Total cell numbers in BAL were not different between anti-IL-9 and its control IgG treated groups (Figure 1a). There was no significant difference in eosinophil, neutrophil, and lymphocyte numbers in BAL. Animals treated with anti-IL-9 had significantly higher number of macrophages in BAL (Figure 1a) and significantly lower percentage of eosinophils in comparison to its control treated animals (63.4 ± 6.7 vs 80.1 ± 1.4 % of total cells in BAL, p = 0.024). Anti-IL-5, in comparison to control IgG1 treated mice, significantly decreased the total cell number in BAL, mainly due to decrease in eosinophil number (Figure 1b).
Figure 1 Cellular profiles of BAL-cells in OVA-sensitized and on five days exposed to OVA Balb/c mice. Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. BAL was performed 24 hours after the last OVA-exposure. Data are shown as mean ± SEM (n = 5–10 per group). *p < 0.05 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Immunocytochemical stainings of BAL cells for BrdU and CD34 are illustrated in Figure 2. Anti-IL-9 did not significantly change newly produced (BrdU+) eosinophil numbers in BAL (Figure 3a), but decreased the percentage of BrdU- eosinophils in BAL (11.5 ± 1.4 vs 16.6 ± 1.5 % of total cells in BAL, p = 0.024). The number of BrdU+ BAL eosinophils was extensively reduced by anti-IL-5 (Figure 3b).
Figure 2 Photomicrographs (original magnification: ×1000) of BAL cells immunocytochemical analysis of staining positively for BrdU (brown nuclear, a) and CD34 (red, b) in sensitized and exposed to OVA on five consecutive days mice. a) 1- BrdU+ eosinophil, 2- BrdU- eosinophil, 3- BrdU+ neutrophil. b) 1- CD34+ granulocyte, 2- CD34- granulocyte.
Figure 3 Effect of treatment with the neutralizing antibodies on BAL eosinophils stained positive for BrdU and unstained cells in OVA-sensitized and on five days exposed to OVA Balb/c mice. BrdU was injected on first and third day of OVA exposure period. Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. Data are shown as mean ± SEM (n = 5–10 per group). **p < 0.01 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Anti-IL-9 did not affect the increase in BAL CD34+ granulocytes induced by OVA-exposure, but slightly (p = 0.048) decreased the number of BAL CD34- granulocytes (Figure 4a). Numbers of BAL CD34+ and CD34- granulocytes were significantly reduced with anti-IL-5 treatment (Figure 4b).
Figure 4 Effect of treatment with the neutralizing antibodies on allergen-induced changes of BAL granulocytes stained for CD34 in Balb/c mice (sensitized and on five days exposed to OVA). Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. Data are shown as mean ± SEM (n = 5–10 per group). **p < 0.01 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Blood cells
As compared with the control antibody, treatment with anti-IL-9 significantly decreased the number of blood neutrophils, but without effect on other cell types or total blood cell numbers (Figure 5a). Anti-IL-5 treatment significantly decreased blood eosinophil number (Figure 5b).
Figure 5 Blood cellular profiles in OVA-sensitized and on five days exposed to OVA Balb/c mice. Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. Blood was taken 24 hours after the last OVA-exposure. Data are shown as mean ± SEM (n = 5–10 per group). *p < 0.05, **p < 0.01 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Immunocytochemical analysis of intranuclear BrdU staining in blood cells showed significant decrease in BrdU+ eosinophils from animals treated with anti-IL-9, compared to its control treated animals (Figure 6a). Also anti-IL-9 significantly reduced the number of BrdU+ blood neutrophils (0.7 ± 0.1 vs 1.3 ± 0.2 × 106/ml blood, p < 0.05). Treatment with anti-IL-5 decreased BrdU+ and BrdU- blood eosinophil numbers (Figure 6b), without any effect on BrdU+ blood neutrophils.
Figure 6 Changes of BrdU-staining in blood eosinophils in sensitized and exposed to allergen (OVA; 100 μg intranasally on five consecutive days) Balb/c mice after the treatment with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. BrdU was injected on first and third day of OVA exposure period. Data are shown as mean ± SEM (n = 5–10 per group). *p < 0.05 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Anti-IL-9 treatment and anti-IL-5 treatment did not significantly affect allergen induced increase in total blood CD34+ cell numbers (data not shown).
Bone marrow eosinophils and CD34+ cells
Anti-IL-9 treatment reduced bone marrow eosinophilia with significant effect on mature eosinophils (Figure 7a). Anti-IL-5 treatment decreased mature and immature bone marrow eosinophils (Figure 7b).
Figure 7 Effect of treatment with the neutralizing antibodies on bone marrow (BM) eosinophil content profiles in OVA-sensitized and on five days exposed to OVA Balb/c mice. Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. Figure shows both eosin-staining cells with immature and with mature morphology. Data are shown as mean ± SEM (n = 5–10 per group). *p < 0.05, **p < 0.01 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Anti-IL-9 also decreased BrdU+ eosinophil number in bone marrow, but did not affect BrdU- eosinophils (Figure 8a). Anti-IL-5 treatment reduced both BrdU+ and BrdU- eosinophils in bone marrow (Figure 8b).
Figure 8 BrdU-staining in bone marrow (BM) eosinophils in OVA-sensitized and on five days exposed to OVA Balb/c mice. BrdU was injected on first and third day of OVA exposure period. Animals were pretreated with a) anti-IL-9 antibody or its isotype control IgG (100 μg/animal); or with b) anti-IL-5 antibody or its isotype control IgG1 (50 μg/animal) 30 minutes before the first exposure. Data are shown as mean ± SEM (n = 5–10 per group). *p < 0.05, **p < 0.01 compared with the corresponding isotype control antibody (Mann-Whitney U test).
Anti-IL-9 treatment did not significantly change bone marrow CD34+ cell numbers, as compared with control treated animals (p = 0.08). Anti-IL-5 significantly reduced CD34+ cell numbers in bone marrow, as compared with its isotype control treatment (26.7 ± 2.7 vs 16.5 ± 2.7 % of total cells, p = 0.03).
Effect of anti-IL-9 treatment on IL-5
Anti-IL-9 treatment, as compared to its isotype control treatment, did not significantly change IL-5 levels in serum (8.7 ± 2.3 vs 30.5 ± 16.7 pg/ml respectively, p = 0.24) and BALf (20.8 ± 6.2 vs 54.6 ± 22.2 pg/ml respectively, p = 0.16).
IL-9 production in bone marrow cells after repeated allergen exposure
The concentration of IL-9, measured by ELISA, in bone marrow supernatant from allergen-exposed animals was 369 ± 247 pg/ml, but was not detectable in serum and BALf from the same animals, as well as in PBS-exposed animals. IL-5 in bone marrow supernatant from the same animals was not detectable by ELISA.
Flow cytometry, using intracellular staining of IL-9, revealed that bone marrow cells from sensitized and OVA-exposed for six days animals after in vitro stimulation with OVA and PMA together with calcium ionophore (IC), had increased expression of IL-9 in comparison with baseline (1.04 and 1.84 times fold respectively). This expression was more pronounced in the bone marrow CD4+ cells (Figure 9). There was no substantial difference between the percent of bone marrow cells expressing CD4 at baseline (plain medium), or after stimulation with OVA, or PMA+IC (9.1%, 8.5% and 6.2% respectively).
Figure 9 Intracellular expression of IL-9, detected by flow cytometry in the bone marrow CD4+ cells from OVA-sensitized and exposed for 6 days animals, cultured in vitro with OVA or PMA (phorbol myristate acetate) + IC (calcium ionophore). Data are expressed as time-fold increase of the mean fluorescence intensity (MFI) of the baseline expression (culture with plain medium).
Effects of combination of anti-IL-5 and anti-IL-9
The effects of combined treatment with anti-IL-5 and anti-IL-9 are described in Table I. No additive effect of anti-IL-9 was observed above the effect of anti-IL-5.
Table 1 The effect of combined treatment with anti-IL-9 and anti-IL-5 or their isotype control on cell profiles in differenttissue compartments
IsC of anti-IL9 + IsC of anti-IL-5 Anti-IL-9 + Anti-IL-5 Anti-IL-5 + IsC of anti-IL-9 Anti-IL-9 + IsC of anti-IL-5
Bone marrow eosinophils(%): 19.3 ± 2.5 3.6 ± 0.6** 3.6 ± 0.7** 18.4 ± 2.2
Immature eos. (%) 13.4 ± 2.1 2.3 ± 0.4** 2.3 ± 0.4** 14.8 ± 2.1
Mature eos. (%) 5.9 ± 0.6 1.2 ± 0.3** 1.4 ± 0.5** 3.7 ± 0.4*
Blood cells (total no., ×106/ml): 1.9 ± 0.2 1.5 ± 0.6 1.3 ± 0.3 1.7 ± 0.2
Neutrophils (×106/ml) 0.7 ± 0.1 0.8 ± 0.3 0.56 ± 0.1 0.7 ± 0.1
Eosinophils (×106/ml) 0.3 ± 0.1 0.1 ± 0.03* 0.1 ± 0.02* 0.3 ± 0.04
Monocytes (×106/ml) 0.3 ± 0.04 0.2 ± 0.04 0.2 ± 0.1 0.2 ± 0.03
Lymphocytes (×106/ml) 0.5 ± 0.1 0.4 ± 0.2 0.5 ± 0.1 0.5 ± 0.1
BAL cells (total no., ×104/ml): 109.3 ± 1.9 44.5 ± 7.6 45.9 ± 3.3 92.4 ± 8.8
Macrophages (×104/ml) 18.9 ± 1.9 11.8 ± 1.0 11.7 ± 0.3 16.9 ± 3.4
Neutrophils (×104/ml) 5.4 ± 1.7 3.8 ± 0.7 4.6 ± 1.0 4.9 ± 1.2
Eosinophils (×104/ml) 67.9 ± 15.1 20.0 ± 6.3* 18.8 ± 2.1* 54.8 ± 6.0
Lymphocytes (×104/ml) 16.8 ± 3.2 8.5 ± 1.8 10.7 ± 1.2 15.5 ± 4.5
Balb/c mice sensitized to OVA were subjected to repeated allergen exposure (OVA; 100 μg i.n. on five consecutive days).
Neutralizing anti-IL-9 (100 μg/animal i.p.) and anti-IL-5 (50 μg/animal i.p.) treatment or their isotype controls treatment, or single antibody in a combination with respective isotype control, were given before the allergen exposure. BAL was performed 24 hours after the last allergen exposure. Data represent mean ± SEM (n = 5 per point). * p < 0.05 and ** p < 0.01 compared with the isotype control antibodies treated group. IsC – isotype control.
Discussion
Our study was designed to determine whether treatment with a monoclonal anti-IL-9 antibody affects the allergic inflammation in a mouse model of airway eosinophilic inflammation and compared any such effect with the results of anti-IL-5 antibody treatment. The monoclonal anti-IL-9 antibody, given at a dose of 100 μg before allergen exposure, did not significantly reduce allergen-induced airway eosinophilia, but consistently reduced bone marrow eosinophilia, by a reduction of newly produced (BrdU+) and mature bone marrow eosinophils. IL-9 was expressed in bone marrow CD4+ cells after allergen exposure, which was more pronounced after stimulation of the cells with allergen or PMA+IC. The monoclonal IL-5 antibody, given at a dose of 50 μg before allergen exposure, reduced eosinophil numbers in all tissue compartments, as well as BrdU+ eosinophils and CD34+ progenitor cells. The effect of anti-IL-5 was significantly greater than the effect of anti-IL-9, and no additive effect of anti-IL-9 was observed beyond the effect of anti-IL-5.
In our study, the neutralizing anti-IL-9 antibody treatment was given as a single dose administered intraperitoneally in sensitized animals before the allergen exposure. Despite variable effects of IL-9 in in vitro studies of allergic cellular processes [23-25], the treatment with a neutralizing anti-IL-9 antibody in vivo in our study did not protect airways from the abundant eosinophilia in BAL, arguing against anti-IL-9 having a potential as an anti-inflammatory treatment in established allergic disease. This result is in contrast with a previous report by Cheng and colleagues [27], which showed that a polyclonal anti-IL-9 antibody at a dose of 20 μg significantly inhibited airway eosinophilic inflammation. The fact that the antibody used in that study was polyclonal may suggest that the inhibitory effect was not solely dependent on inhibition of IL-9. Treatment with a monoclonal anti-IL-9-antibody during the sensitization process attenuates the development of airway allergic inflammation and airway hyperresponsiveness in mice exposed to allergen at a later time, implying a role of IL-9 in the development of allergy [28]. However, the treatment was given in very high concentration (200 μg) as four doses (total 800 μg/animal) during the sensitization period, which would correspond to very high doses of antibody treatment in a clinical situation. Furthermore, clinical treatment of allergic disease is given to individuals that have already been sensitized to one or several allergens, why any effect of a neutralizing antibody during sensitization is unlikely to be helpful clinically. Importantly, our finding of no or limited inhibitory effect of BAL eosinophilia by anti-IL-9 is in agreement with a report showing that sensitized IL-9KO mice [29] respond with normal airway inflammation after allergen exposure.
Despite the lack of effect of inhibition of airway eosinophilia by anti-IL-9, we consistently observed an inhibitory effect of the treatment on the increase in bone marrow eosinophils after allergen exposure. Airway allergic inflammation is to a great extent the result of the recruitment of newly produced cells from bone marrow via blood under the influence of allergen. Newly produced eosinophils, labeled with BrdU, to a substantial degree contribute to the allergen-induced inflammatory process, which results in an accumulation of eosinophils in the airways [7,8,30]. Anti-IL-9 treatment significantly decreased the number of mature eosinophils in bone marrow, and reduced newly produced eosinophils (BrdU+) in bone marrow as well as in blood. This supports the hypothesis that IL-9 participates in the eosinophilopoiesis and the release of eosinophils from bone marrow to blood. However, the regulatory effect of IL-9 seems to be of a subordinate magnitude compared to IL-5, since the effects induced by neutralization of IL-5 were of a much greater magnitude.
Eosinophils develop from CD34+ progenitor cells that undergo terminal differentiation normally within the bone marrow and primarily under the influence of IL-5, but also with the support of other cytokines. In several animal studies, an increase in the CD34+ population is observed in the bone marrow, circulation and airways after the allergen exposure [7,8,30,31], as well as in human studies of atopic subjects, regardless of asthmatic status [32,33]. In vitro IL-9 has shown a capacity to act on hematopoietic progenitors to enhance eosinophil development when added to CD34+ cells cultured with a mixture of IL-3 and IL-5 [12,13]. In our study in vivo, we failed to observe any effect of anti-IL-9 on CD34+ cells, and no significant effects were observed on IL-5 levels either in serum or in BALf. However, the anti-IL-5 antibody showed pronounced capacity to reduce an allergen-induced increase in CD34+ cells in both bone marrow and BAL. Again, our data therefore suggest that IL-9 has a subordinate role compared to IL-5 in eosinophilopoiesis.
Lymphocytes, especially circulating lymphocytes, are considered to be the major cellular source of IL-9 [19]. We found that bone marrow CD4+ cells taken from allergen-exposed animals, over-expressed IL-9 after in vitro stimulation with OVA or PMA and calcium ionophore. This finding further supports a role of IL-9 in allergen-induced bone marrow responses, especially in enhanced granulocytopoiesis, although the degree of the response and the exact importance remain to be elucidated.
We also found some degree of reduction of blood neutrophils after allergen exposure in animals treated with anti-IL-9. Primarily, a reduction of BrdU+ neutrophils was observed, which may imply a role of IL-9 in the production of neutrophils in the bone marrow. This study was not designed to specifically evaluate the role of IL-9 in neutrophilia, but implies that such studies may be of interest. This is further supported by the finding that IL-9 can induce IL-8 production in a concentration depend manner [34].
We cannot exclude the possibility that even higher doses of the neutralizing monoclonal IL-9 antibody could have been more effective in attenuating eosinophilia. Overall, however, the degree of contribution of IL-9 in allergen-induced airway eosinophilia seems to be much smaller than that of IL-5, since the treatment with the single dose of the anti-IL-5 antibody (50 μg) significantly inhibited bone marrow, blood and airway eosinophilia, but the higher dose of anti-IL-9 (100 μg) was much less effective. Furthermore, the fact that IL-9KO mice respond with similar degree of inflammation after sensitization and allergen exposure as wild type mice [29] further supports a subordinate role of IL-9 in allergen induced airway eosinophilia.
Taken together, our data indicate that IL-9 may have some degree of regulatory effect on eosinophilia in the bone marrow and blood, perhaps via eosinophilopoiesis, since newly produced eosinophils were reduced in these compartments. However, no inhibitory effect on airway eosinophils was observed after anti-IL-9 treatment, which argues that the development of such a drug will be unsuccessful in treating allergic airway eosinophilia.
List of abbreviations
BAL bronchoalveolar lavage
BrdU 5-bromo-2'-deoxyuridine
IC calcium ionophore
IL interleukin
i.n. intranasal(ly)
i.p. intraperitoneal(ly)
OVA ovalbumin
PBS phosphate-buffered saline
PMA phorbol myristate acetate
RT room temperature
Authors' contributions
BS carried out the major part of the animal experiments, the immunocytochemistry, and participated in the writing of the manuscript.
MR carried out part of the animal experiments, participated in the sequence alignment, and performed some statistical analysis.
AB carried out the cell culture and FACS analysis.
AKJ carried out the immunoassays of IL-9, and participated in the sequence alignment.
RS participated in the study design and in the sequence alignment.
JL conceived the study, participated in its design and coordination of the study, and participated in the writing of the manuscript.
Acknowledgements
We are grateful to Ms Margareta Sjöstrand and Ms Carina Malmhäll for technical assistance, and to Professor Bengt-Eric Skoogh for helpful discussion.
This work was supported in part by the Swedish Heart Lung Foundation, and the Swedish National Research Council (K2001-71X-13492-02B).
==== Refs
Bousquet J Chanez P Lacoste YJ Barneon G Chavanian N Enander I Venge P Ahlstedt S Simony-Lafontaine J Godard P Michel FB Eosinophilic inflammation in asthma N Engl J Med 1990 323 1033 1039 2215562
Frigas E Motojima S Gleich GJ The eosinophilic injury to the mucosa of the airways in the pathogenesis of bronchial asthma Eur Respir J Suppl 1991 13 123s 135s 1953909
Gleich GJ Mechanisms of eosinophil-associated inflammation J Allergy Clin Immunol 2000 105 651 663 10756213 10.1067/mai.2000.105712
Redington AE Howarth PH Airway wall remodeling in asthma Thorax 1997 52 310 312 9196510
Foster PS Hogan SP Ramsay AJ Matthaei KI Young IG Interleukin 5 deficiency abolishes eosinophilia, airways hyperreactivity, and lung damage in a mouse asthma model J Exp Med 1996 183 195 201 8551223 10.1084/jem.183.1.195
Wang J Palmer K Lötvall J Milan S Lei XF Matthaei KI Gauldie J Inman MD Jordana M Xing Z Circulating, but not local lung, IL-5 is required for the development of antigen-induced airways eosinophilia J Clin Invest 1998 102 1132 1141 9739047
Tomaki M Zhao LL Lundahl J Sjöstrand M Jordana M Linden A O'Byrne P Lötvall J Eosinophilopoiesis in a murine model of allergic airway eosinophilia: involvement of bone marrow IL-5 and IL-5 receptor α1 J Immunol 2000 165 4040 4050 11034415
Sitkauskiene B Johansson A-K Sergejeva S Lundin S Sjöstrand M Lötvall J Regulation of bone marrow and airway CD34+ eosinophils by IL-5 Am J Respir Cell Mol Biol 2004 30 367 378 12920051 10.1165/rcmb.2002-0311OC
Renauld JC Kermouni A Vink A Louahed J Van Snick J Interleukin-9 and its receptor: involvement in mast cell differentiation and T cell oncogenesis J Leukoc Biol 1995 57 353 360 7884304
Van Snick J Goethals A Renauld JC Van Roost E Uyttenhove C Rubira MR Moritz RL Simpson RJ Cloning and characterization of a cDNA for a new mouse T cell growth factor (P40) J Exp Med 1989 169 363 368 2521242 10.1084/jem.169.1.363
Vink A Warnier G Brombacher F Renauld JC Interleukin 9-induced in vivo expansion of the B-1 lymphocyte population J Exp Med 1999 189 1413 1423 10224281 10.1084/jem.189.9.1413
Bourette RP Royet J Mouchiroud G Schmitt E Blanchet JP Murine interleukin 9 stimulates the proliferation of mouse erythroid progenitor cells and favours the erythroid differentiation of multipotent FDCP-mix cells Exp Hematol 1992 20 868 873 1628705
Matsuzawa S Sakashita K Kinoshita T Ito S Yamashita T Koike K IL-9 enhances the growth of human mast cell progenitors under stimulation with stem cell factor J Immunol 2003 170 3461 3467 12646606
Temann UA Ray P Flavell RA Pulmonary overexpression of IL-9 induces Th2 cytokine expression, leading to immune pathology J Clin Invest 2002 109 29 39 11781348 10.1172/JCI200213696
McLane MP Haczku A van de Rijn M Weiss K Ferrante V MacDonald D Renauld JC Nicolaides NC Holroyd KJ Levitt RC Interleukin-9 promotes allergen-induced eosinophilic inflammation and airway hyperresponsiveness in transgenic mice Am J Respir Cell Mol Biol 1998 19 713 720 9806735
Levitt RC McLane MP MacDonald D Ferrante V Weiss C Zhou T Holroyd KJ Nicolaides NC IL-9 pathway in asthma: new therapeutic targets for allergic inflammatory disorders J Allergy Clin Immunol 1999 103 5 Pt 2 S485 S491 10329852
Nicolaides NC Holroyed KJ Ewart SL Eleff SM Kiser MB Dragwa CR Sullivan CD Grasso L Zhang LY Messler CJ Zhou T Kleeberger SR Buetow KH Levitt RC Interleukin-9: a candidate gene for asthma Proc Natl Acad Sci USA 1997 94 13175 13180 9371819 10.1073/pnas.94.24.13175
Doull IJ Lawrence S Watson M Begishvili T Beasley RW Lampe F Holgate T Morton NE Allergic association of gene markers on chromosomes 5q and 11q with atopy and bronchial hyperresponsiveness Am J Respir Crit Care Med 1996 153 1280 1284 8616554
Erpenbeck VJ Hohlfeld JM Volkmann B Hagenberg A Geldmacher H Braun A Krug N Segmental allergen challenge in patients with atopic asthma leads to increased IL-9 expression in bronchoalveolar lavage fluid lymphocytes J Allergy Clin Immunol 2003 111 1319 1327 12789235 10.1067/mai.2003.1485
Shimbara A Christodoulopoulos P Soussi-Gounni A Olivenstein R Nakamura Y Levitt RC Nicolaides NC Holroyd KJ Tsicopoulos A Lafitte JJ Wallaert B Hamid QA IL-9 and its receptor in allergic and nonallergic lung disease: increased expression in asthma J Allergy Clin Immunol 2000 105 108 115 10629460 10.1016/S0091-6749(00)90185-4
Bhathena PR Comhair SA Holroyd KJ Enzurum SC Interleukin-9 receptor expression in asthmatic airways in vivo Lung 2000 178 149 160 10871433 10.1007/s004080000018
Zhou Y McLane M Levitt RC Interleukin-9 as a therapeutic target for asthma Respir Res 2001 2 80 84 11686869 10.1186/rr42
Gounni AS Nutku E Koussih L Aris F Louahed J Levitt RC Nicolaides NC Hamid Q IL-9 expression by human eosinophils: regulation by IL-1β and TNF-α J Allergy Clin Immunol 2000 106 460 466 10984364 10.1067/mai.2000.109172
Gounni AS Gregory B Nutku E Aris F Latifa K Minshall E North J Tavernier J Levit R Nicolaides N Robinson D Hamid Q Interleukin-9 enhances interleukin-5 receptor expression, differentiation, and survival of human eosinophils Blood 2000 96 2163 2171 10979962
Louahed J Zhou Y Maloy WL Rani PU Weiss C Tomer Y Vink A Renauld JC Van Snick J Nicolaides NC Levitt RC Haczku A Interleukin 9 promotes influx and local maturation of eosinophils Blood 2001 97 1035 1042 11159534 10.1182/blood.V97.4.1035
Falchetti R Di Francesco P Lanzilli G Gaziano R Casalinuovo IA Palamara AT Ravagnan G Garaci E Determination of cytokine co-expression in individual splenic CD4+ and CD8+ T cells from influenza virus-immune mice Immunology 1998 95 346 351 9824496 10.1046/j.1365-2567.1998.00608.x
Cheng G Arima M Honda K Hirata H Eda F Yoshida N Fukushima F Ishii Y Fukuda T Anti-interleukin-9 antibody treatment inhibits airway inflammation and hyperreactivity in mouse asthma model Am J Respir Crit Care Med 2002 166 409 416 12153980 10.1164/rccm.2105079
Kung TT Luo B Crawley Y Garlisi CG Devito K Minnicozzi M Egan R Kreutner W Chapman RW Effect of anti-mIL-9 antibody on the development of pulmonary inflammation and airway hyperresponsiveness in allergic mice Am J Respir Cell Mol Biol 2001 25 600 605 11713102
McMillan S Bishop B Townsend MJ McKenzie AN Lloyd CM The absence of interleukin 9 does not affect the development of allergen-induced pulmonary inflammation nor airway hyperreactivity J Exp Med 2002 195 51 57 11781365 10.1084/jem.20011732
Johansson AK Sergejeva S Sjostrand M Lee JJ Lotvall J Allergen-induced traffic of bone marrow eosinophils, neutrophils and lymphocytes to airways Eur J Immunol 2004 34 3135 3145 15384047 10.1002/eji.200425043
Sergejeva S Johansson AK Malmhall C Lotvall J Allergen exposure-induced differences in CD34+ cell phenotype: relationship to eosinophilopoietic responses in different compartments Blood 2004 103 1270 1277 14512324 10.1182/blood-2003-05-1618
Sehmi R Howie K Sutherland DR Schragge W O'Byrne PM Denburg JA Increased levels of CD34+ hemopoietic progenitor cells in atopic subjects Am J Respir Cell Mol Biol 1996 15 645 655 8918371
Robinson DS Damia R Zeibecoglou K Molet S North J Yamada T Kay AB Hamid Q Presence of CD34+/IL-5+ eosinophil precursor cells in airways of asthmatic individuals J Allergy Clin Immunol 1999 103 9 13
Abdelilah SG Latifa K Esra N Lisa C Bouchaib L Nicolaides NC Levitt RC Hamid Q Functional expression of IL-9 receptor by human neutrophils from asthmatic donors: role in IL-8 release J Immunol 2001 166 2768 2774 11160343
| 15823208 | PMC1090617 | CC BY | 2021-01-04 16:36:26 | no | Respir Res. 2005 Apr 11; 6(1):33 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-33 | oa_comm |
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-161582630310.1186/1742-4682-2-16ResearchBreakdown of accommodation in nerve: a possible role for persistent sodium current Hennings Kristian [email protected] Lars [email protected] Ole K [email protected] Center for Sensory-Motor Interaction (SMI), Aalborg University. Frederik Bajers Vej D3-203, 9220 Aalborg Ø, Denmark2005 12 4 2005 2 16 16 8 12 2004 12 4 2005 Copyright © 2005 Hennings et al; licensee BioMed Central Ltd.2005Hennings 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
Accommodation and breakdown of accommodation are important elements of information processing in nerve fibers, as they determine how nerve fibers react to natural slowly changing stimuli or electrical stimulation. The aim of the present study was to elucidate the biophysical mechanism of breakdown of accommodation, which at present is unknown.
Results
A model of a space-clamped motor nerve fiber was developed. It was found that this new model could reproduce breakdown of accommodation when it included a low-threshold, rapidly activating, persistent sodium current. However, the phenomenon was not reproduced when the persistent sodium current did not have fast activation kinetics or a low activation threshold.
Conclusion
The present modeling study suggests that persistent, low-threshold, rapidly activating sodium currents have a key role in breakdown of accommodation, and that breakdown of accommodation can be used as a tool for studying persistent sodium current under normal and pathological conditions.
==== Body
Background
Accommodation is important for information processing in nerve fibers, as it determines whether, and how frequently, slowly-changing natural and artificial stimuli are translated into action potentials. Hill's theory of accommodation in nerve has been one of the most influential theories in this area [1]. A prediction of this theory is that a linearly rising current requires a certain critical slope in order to excite nerve fibers. Although this critical slope has been demonstrated in experimental preparations [2,3], it has not been found under normal physiological conditions [4,5]. Instead, nerve fibers have been shown to exhibit breakdown of accommodation; that is, a long-duration slowly rising current excites nerve fibers at a nearly constant intensity no matter how slowly this intensity is approached [4,5]. A critical slope has only been found for depolarized nerve fibers, and Hill's theory of accommodation has been shown only to be applicable to such fibers [6]. Accommodation and breakdown of accommodation were the foci of several studies before the invention of the voltage-clamp, since prior to this innovation it was one of the few methods by which membrane kinetics could be studied. Since the invention of the voltage-clamp and later the patch-clamp some fifty years ago, the concept of breakdown of accommodation has been virtually absent from the scientific literature [7]. However, the biophysical mechanism responsible for breakdown of accommodation is still unknown; and as will be shown in this paper, a model that only contains transient sodium channels (i.e. currents that activate and deactivate rapidly in response to membrane depolarization) is unable to reproduce the phenomenon. Persistent (no inactivation) and late (slow inactivation) sodium channels have been identified in large dorsal ganglion neurons [8] and it has been found that these channels are needed for modeling latent addition in motor and sensory nerve fibers (i.e. threshold changes to short sub-threshold stimuli [9]). This suggests that persistent or late sodium channels are present in both motor and sensory myelinated nerve fibers and have fast activation kinetics that can initiate action potentials. The present study was undertaken to study the hypothesis that persistent sodium channels create a "threshold region" of membrane depolarization that cannot be exceeded without the generation of an action potential. Thus, it is suggested that persistent sodium channels are the cause of breakdown of accommodation.
The results in the present paper were based on a model of a space-clamped nerve fiber. This model included a persistent sodium channel based on the work of Bostock and Rothwell (1997) [9]. This channel was defined from the transient sodium channel with the following modifications: a) inactivation was removed (a persistent channel); b) the time-constant was slowed by a factor of two (time-constant); and c) the kinetics was displaced so that the channel was activated at a membrane potential 20 mV more negative than is required to activate the transient channel (voltage shift) [9].
Results
Model validation
The structure of the model and the choice of parameters allowed it to reproduce four sets of independent experimental data: threshold electrotonus, recovery cycle, latent addition and breakdown of accommodation (see Figure 1). The model was found to have a strength-duration time constant of 133.2 μs, which is similar to the experimental recorded value for the median nerve (139 ± 59 μs [10]). Furthermore, the model simulated breakdown of accommodation (see Figure 1D). The initial critical slope of the model was found to be 17.3 rheobase/s, which is lower than the experimentally recorded value (21.2 ± 3.72 rheobase/s) for ulnar nerves. The breakdown of accommodation seen in the model was likewise greater than observed in the ulnar nerve, and was closer to that observed in sensory nerve fibers [5]. The accommodation curve flattened out and remained near 2 rheobases when the time-constant of the current rise was greater than ~150 ms.
Figure 1 Comparison of the new model with experimental data for: A) threshold electrotonus [40], B) recovery cycle [41], C) latent addition [10], and D) accommodation curve [5]. In threshold electrotonus, a sub-threshold conditioning pulse of 100 ms duration is used to alter the threshold of a test stimulus delayed with respect to the onset of the conditioning pulse. In the recovery cycle, the nerve fiber is excited by a supra-threshold stimulus and the threshold of a test stimulus is determined at inter-stimulus intervals (TISI) of 2 ms to 100 ms. In latent addition, a short duration sub-threshold conditioning stimulus is used to alter the threshold of a test stimulus; the onset of the test stimulus is delayed with regard to the onset of the conditioning stimulus. In the accommodation curve, the threshold of stimuli of the form IS(1-e-tτ) was determined, where τ was the time-constant of the current rise. In A, the bold line is the initial critical slope, which was estimated from the first four points in the accommodation curve where it is approximately a straight line. Experimental range: a) minimum and maximum of the experimental range, b) and c) mean ± standard deviation.
Breakdown of accommodation
In order to study the relationship between the properties of persistent sodium channels and breakdown of accommodation, one parameter at a time (number of channels, voltage shift, time constant) was changed and its influence on breakdown of accommodation was assessed (see Figure 2). When the voltage shift was decreased from -20 mV to -10 mV or -0 mV, it was still possible to create breakdown of accommodation by increasing the number of persistent sodium channels to 3.75% (-10 mV) or 15% (-0 mV) of the total number of sodium channels (see Figure 3A). However, for a voltage shift of -0 mV, the membrane potential did not return to the resting potential after the generation of an action potential (see Figure 3A).
Figure 2 Relationship between the properties of persistent sodium channels and breakdown of accommodation: A) Number of persistent sodium channels (number of persistent sodium channels: 1.0%, 1.5%, 2.0%, 2.5%, and 3.0%). B) Voltage shift of the kinetics of the persistent sodium channels relative to the transient sodium channels (voltage shift: -0 mV, -5 mV, -10 mV, -15 mV, and -20 mV). C) Time constant of persistent sodium channel activation (time-constant slowed by a factor of: 10, 6.66, 4.5, 3.0, 2.0, and 1.0). A thick line and bold number indicates the default model.
Figure 3 Relationship between voltage shift of persistent sodium channels relative to transient sodium channels and the shape of the action potential. For voltage shifts of -10 mV and -0 mV, the number of persistent sodium channels is set to a value (3.75% and 15%, respectively) that would produce approximately the same degree of breakdown of accommodation as the default model (voltage shift of -20 mV). A) The shape of the action potentials. B) The accommodation curves for the three voltage shifts of -20 mV●, -10 mV■, and -0 mV▲.
Threshold responses to linearly rising stimuli
The threshold responses to linearly rising stimuli were significantly modified by the presence of persistent sodium channels (see Figure 4). (A threshold response is a response to a stimulus with intensity equal to the excitation threshold of the nerve fiber). Without persistent sodium channels, the threshold response to a linearly rising current of 20 ms duration did not occur at the end of the stimulus but had a latency of 5.04 ms (see Figure 4A). With 2.5% persistent sodium channels, the threshold response occurred at the end of the stimulus (see Figure 4B). This difference was found for all linearly rising currents tested that had stimulus durations in the range 1 ms to 200 ms; when the model had 2.5 % persistent sodium channels the threshold responses were always observed at the end of the stimulus, whereas without persistent sodium channels the longest latency of the threshold response was 5.04 ms (see Figure 4C). For both models, with and without persistent sodium channels, a non-linear response always occurred when the membrane was depolarized to a certain threshold value. This non-linear response initially occurred at the end of the stimulus, and with persistent sodium channels it resulted in an action potential. Without persistent sodium channels, it only resulted in an action potential when the stimulus intensity was sufficient for the response to occur with a latency of 5.04 ms or less.
Figure 4 Responses of the new model without (A) and with (B) persistent sodium channels to a linearly rising current; (C) the latencies of the threshold responses for the models without (bold line) and with (thin line) persistent sodium channels. In subfigures (A) and (B) the responses for each model are shown for increasing stimulus intensities: A) 0.950, 0.975, 1.00, and 1.025 excitation threshold, and B) 0.925, 0.950, 0.975, and 1.00 excitation threshold. Threshold responses are drawn with bold lines.
Threshold electrotonus
A proportional relationship between the threshold change of a test stimulus and the underlying electrotonic changes in the membrane potential is a fundamental requirement for threshold electrotonus. Such a relationship was found for the model with 2.5% persistent sodium channels (see Figure 5). However, when the persistent sodium channels were removed from the model, the relationship between threshold and membrane potential broke down. This was tested with the conditioning current at an intensity of 40% for the model with persistent sodium currents, and 30.5% for the model without persistent sodium current. These two conditioning current intensities produced similar membrane depolarizations in the two models, which enabled the effect of the persistent sodium channels to be assessed. The relationship between threshold and membrane potential has also been found to break down when nerve fibers are depolarized because of a long-duration conditioning current or ischaemia [6]. Consequently, had the membrane depolarizations not been matched in the two models, the loss of the relationship between threshold and membrane potential may have been attributable to the greater membrane depolarization in the model without persistent sodium channels.
Figure 5 Electrotonus (A) and threshold electrotonus (B) of the new model with (thick line) and without (thin line) persistent sodium channels. The intensities of the conditioning currents were 40% and 30.5% of the threshold of the test stimulus alone for the model with and without persistent sodium channels, respectively.
Discussion
We have used a model of a space-clamped motor nerve fiber to provide evidence for a link between persistent sodium currents and breakdown of accommodation. The model demonstrated that these channels might be the cause of breakdown of accommodation, as their inclusion enabled the model to reproduce the phenomenon (see Figure 1D). It also demonstrated that such channels are likely to be low-threshold and rapidly activating (see Figure 2). The low-threshold property is further supported by the fact that although breakdown of accommodation can be reproduced by high-threshold persistent sodium channels, in this case it results in an action potential that does not return to the resting potential (see Figure 3).
Experimental evidence for the role of persistent sodium current in breakdown of accommodation
Persistent, late sodium currents have been observed in large dorsal root ganglion cells. These current were found to have a low threshold and fast activation kinetics and were therefore expected to modulate membrane excitability by amplifying and prolonging depolarization from a generator potential or an external electrode [8,11]. Indirect evidence has been obtained for the presence of such channels in both large diameter sensory nerve fibers and motor nerve fibers [9], and they can produce regenerative currents that facilitate action potential generation. Persistent sodium channels have been shown to amplify otherwise sub-threshold depolarization, thereby initiating action potentials [12]. Furthermore, acidification and alkalization within the physiological range have been found respectively to decrease and increase persistent and late sodium currents [13]. This pH-dependence of the late sodium current correlates well with experimental observations of breakdown of accommodation. Hence, breakdown of accommodation has been found to decrease during ischaemia and increase during hyperventilation [5]. Furthermore, when nerve fibers are depolarized with a polarizing current, there is a decrease in the threshold to triangular stimuli [14]. This suggests that it is not membrane depolarization per se that causes loss of breakdown of accommodation and the presence of a critical slope for slowly rising stimuli. In the present study, the loss of breakdown of accommodation is explained by loss of the persistent sodium current, such as would be caused by ischaemic depolarization due to acidification. Consequently, the present study predicts that the critical slope found by [2] was caused by ischaemic acidification and not membrane depolarization.
The effect of persistent sodium channels on threshold responses
In the present study, when the models with and without persistent sodium currents were stimulated by linearly rising currents, non-linear responses always resulted in action potentials when the model exhibited breakdown of accommodation. However, without breakdown of accommodation, non-linear responses only resulted in action potentials when they occurred within a "critical latency" from the onset of the stimulus. Without a "critical latency", which is the case with breakdown of accommodation, the threshold response occurred at the cessation of a long linearly rising stimulus. Consequently, the threshold for such stimuli is nearly constant regardless of their duration. However, when there is a "critical latency", the membrane potential needs to reach the voltage threshold within this "critical latency" for the nerve fiber to fire an action potential. The critical slope will then be proportional to the voltage threshold for which a non-linear response occurs divided by the "critical latency"; i.e. a decrease in "critical latency" results in an increased critical slope.
Model limitations
The present model simplifies existing knowledge of neuronal morphology and the distribution of ion channels. With regard to ion channels, only two potassium channels and two sodium channels were included in the model, but at least five distinct potassium channels [15] and three persistent and late sodium channels [11] have been identified, besides the classical transient sodium channel [16]. Unfortunately, current knowledge of the potassium and sodium channels in motor nerve fibers does not provide enough detail to allow modeling of them all. For example, the channel densities and kinetics are not known for all five potassium channels [15], and the kinetic data we have for the slow and fast potassium channels are likely to represent amalgamations of several channel species into single stereotypes [16]. Consequently, the present model is based on an amalgamation of distinct channels into stereotypes and the detailed geometrical structures of motor nerve fibers into a gross equivalent electrical circuit. The parameters of the model were based on experimental current- and voltage-clamp recordings whenever possible and the results obtained were found to be in line with experimental work. Consequently, these simplifications appear justified and therefore provide a basis for studying the biophysical properties of breakdown of accommodation. This assumption is supported by previous work where models have provided insights into biophysical mechanisms [9,17,18]. However, the internodal leak resistance (RIL) in particular was not based on experimental data, but was instead set by trial and error to a value that would enable the model to reproduce known experimental data. This approach was used since few experimental data on the internodal leak resistance are available. There are only modeling data on the periaxonal resistivity [19], but further modeling data suggest that the longitudinal conductance of the myelin sheaths has to be taken into account in determining the internodal leak resistance [20]. Consequently, an internodal leak resistance based solely on the width of the periaxonal space is likely to be an underestimate. The unknown resistivity of the periaxonal space presents further difficulties in obtaining a value for internodal leak resistance on the basis of experimental data alone. For these reasons we believe that the present approach was justified.
Alternative explanations of breakdown of accommodation
An alternative explanation for breakdown of accommodation could be the gating mode of the transient sodium channel [21]. The present paper follows the convention of assuming that activation and inactivation are two independent processes (i.e. the formalism of [22]). Today, it is known that activation and inactivation are inter-dependent, and that most transient sodium channels will go through an open state before entering an inactivated state [21]. This difference between the Hodgkin and Huxley formalism and recent knowledge of transient sodium channel function may have a synergistic role in breakdown of accommodation. Hence, a transient sodium channel with little inactivation before channel opening would not permit a critical slope and loss of breakdown of accommodation. However, this explanation remains unproven and would not change the conclusion of the present study, that persistent and late sodium channels can cause breakdown of accommodation. The interdependence of transient channel activation and inactivation may change the densities of persistent sodium channels needed for creating breakdown of accommodation, and thus there may be synergism between transient and persistent sodium channels.
A second explanation may be m-h overlap in the activation/inactivation kinetics of the transient sodium channel. For transient sodium channels, there is a region of membrane depolarization in which a persistent sodium current is generated l [23]. This is caused by channel activation while the membrane is still not sufficiently depolarized for all the channels to be inactivated, a phenomenon that has been termed m-h overlap. A theoretical study has demonstrated that the original squid axon model of Hodgkin and Huxley has breakdown of accommodation as a result of m-h overlap [23]. In this paper and other studies [9,24], persistent sodium channels are modeled as discrete channels. However, this does not imply that they are physically different from transient sodium channels. Three discrete persistent and late sodium currents have been identified on the basis of inactivation kinetics [11] in addition to the classical transient sodium current [16], but only one sodium channel Nav (1.6) has been found in the nodes of Ranvier in large peripheral nerve fibers [25]. This may suggest that persistent and late sodium currents are not generated specifically, but instead by transient sodium channels that operate in a gating mode with no or slowed channel inactivation. The modeling of persistent sodium current as created by persistent sodium channels does not provide evidence for the existence of such channels, only evidence that persistent sodium current can lead to breakdown of accommodation. Consequently, such persistent sodium current may be created by m-h overlap. However, in studies on persistent sodium currents, it has been argued that m-h overlap is not consistent with the observed kinetics [8,11]. Evidently, in mammalian nerve fibers, the persistent sodium current is most likely not generated by m-h overlap; but the study of [23] suggests that m-h overlap may be important for the persistent sodium current and breakdown of accommodation in squid axons.
Conclusion
The present modeling study has demonstrated that persistent sodium currents can create a "threshold region" for membrane depolarization that cannot be exceeded without the generation of an action potential. Thus, a persistent sodium current may be the underlying biophysical mechanism for the breakdown of accommodation to slowly rising currents, which are observed under normal physiological conditions in mammalian nerve fibers [4,5]. This suggests that accommodation curves can be used as a tool for studying persistent sodium currents under normal and pathological conditions.
Methods
Electrical model of a motor nerve fiber
The structure of the model of the space-clamped motor nerve fiber was based on previous models used for studying the accommodative properties of such fibers [9,26,27]. The present model represents a motor nerve fiber by the electrical equivalent circuit shown in Figure 6. The geometry of the node and internode was based on studies on the morphology of cat ventral spinal roots. The geometrical parameters were taken from cats of 1–11 years of age for a motor nerve fiber with a diameter of 14 μm (see Table 1). The nodal, internodal and myelin capacitances in the electrical equivalent circuit were calculated on the basis of these geometrical parameters and experimentally estimated capacitances per square micrometer (see Table 2 and Figure 6). The internodal leak resistance (Ril) and nodal resting potential were set by trial and error rather than calculated from geometrical and electrical parameters (see section entitled 'Validation', below).
Table 1 Geometrical parameters
Inter-nodal length (L) 1.37 mm [45]
Inter-nodal diameter (di) 8.8 μm [46]
Nodal diameter (dn) 3.5 μm [47]
Nodal length (l) 1 μm [48]
Number of myelin lamella (N) 141 [46]
Table 2 Electrical parameters
Nodal capacitance (cn) 2 μF/cm2 [39]
Internodal capacitance (ci) 1 μF/cm2 [49]
Myelin capacitance (cm) 0.1 μF/cm2 [50]
Figure 6 Equivalent circuit for a space-clamped motor neuron. The model consisted of a node and an internode. Both the node and the internode contained non-linear current sources, which were calculated from equilibrium potentials and conductances. Channel types and maximum ionic conductances: node, transient sodium (Nat, 276nS), persistent sodium (Nap, 7.1nS), fast potassium (Kf, 4.1nS), and slow potassium (Ks, 17.4nS); internode, slow potassium (Ks, 87.1nS) and leak conductance (L, 1.7nS). The linear parameters of the model were: Cn, nodal capacity (0.22pF), Ci, internodal capacity (379pF), Cm, capacity of the myelin sheath (0.17pF), and Ril, internodal leak resistance (41 MΩ).
Ionic currents
Five major ionic currents have been identified in myelinated nerve fibers as necessary for modeling a wide variety of experimental data: the transient sodium current (Nat) for modeling the action potential [22], and the persistent sodium current (Nap) for modeling latent addition [9] and the recovery cycle [24]. Fast (Kf) and slow (Ks) potassium currents have been shown to explain accommodation to depolarizing conditioning currents [28]. Accommodation to hyperpolarizing currents can be explained by a hyperpolarization-activated cation conductance (IH), which is also thought to limit hyperpolarization in nerve fibers after they have conducted a train of impulses [28,29].
Transient and persistent sodium channels were included in the node, but following the work of [9] they were omitted from the internode for simplicity. The hyperpolarization-activated cation conductance was omitted from the model as it does not influence the response of nerve fibers to depolarizing stimuli [14]. Based on the work of [28,30,31]., the slow potassium current was included in the node as well as the internode. There is evidence for the localization of fast potassium channels in the paranode [32-35] As the paranode was not included in the present model, it was impossible to include fast potassium channels at this location. Instead, the approach used by [36] was applied and the fast potassium channels were included in the node. The ionic currents were described as being generated through membrane conductances (see Figure 6). The sodium conductances and slow potassium conductance in the node were based on single channel conductances and channel densities. Single channel conductances of 13pS and 8pS were used for the sodium channels and slow potassium channels, respectively [37]. The nodal densities for the sodium and slow potassium channels were set to 1000 channels/μm2 [37] and 100 channels/μm2 [38], respectively. The ion conductance of the fast potassium current was based on the work of [16], who found a fast potassium conductance of 15nS and a capacitive load of 1.4pF on the nodal membrane. The conductance of the fast potassium current was set from an estimate of the membrane area [16], which was based on the nodal capacitance in experimental data and the nodal capacitance per square micrometer [39].
The nodal resting potential was kept stable by a current leak to the internode, and the internodal resting potential was determined from this relationship. The internodal resting potential was kept stable by a small internodal sodium leak conductance. The nodal persistent sodium conductance was set by the fraction of nodal sodium channels that would be persistent. Therefore, the total number of nodal sodium channels was kept constant for all simulations.
Membrane kinetics
The non-linear membrane dynamics were based on human data [16]. The ionic current was given as: transient sodium current iNat = GNatm3h(E-ENa), persistent sodium current iNap = GNapp3(E-ENa), fast potassium current iKf = GKfn4(E-EK), and slow potassium current iKs = GKss(E-EK). The fractional activations (m, h, p, n and s) were given by the differential equation:
dx/dt = αx(1-x)-βxx, for x = m, h, p, n, s
where αm, αp, αn, αs = A(E-B)/(1-exp((B-E)/C)); βm, αh, βp, βn, βs = A(B-E)/(1-exp((E-B)/C)), βh = A/(1+exp((B-E)/C)) (see Table 3 for the constant: A, B, and C) and E is the membrane potential. The rate constants (αx and βx) where scaled by appropriate Q10 factors to a temperature of 37°C (see Table 3). All membrane kinetics were the same as those of [16] except the kinetics for the slow potassium current (see section entitled 'Validation', below). In order to allow the model to reproduce threshold electrotonus, it was necessary to modify the slow potassium channels. The kinetics were changed in order to slow the channel activation and to lower the fraction of open slow potassium channels at the resting potentials for the node and internode.
Table 3 Rate constants
Q10 A B C
ms-1 mV mV
αm 2.2 1.86 -18.4 10.3
βm 2.2 0.086 -22.7 9.16
αh 2.9 0.0336 -111.0 11.0
βh 2.9 2.3 -28.8 13.4
αp 2.2 0.93 -38.4 10.3
βp 2.2 0.043 -42.7 9.16
αn 3.0 0.00798 -93.2 1.1
βn 3.0 0.0142 -76.0 10.5
αs 3.0 0.0122 -12.5 16.9
βs 3.0 0.000736 -80.1 12.6
Validation
The model was validated with four sets of experimental data: threshold electrotonus [40], recovery cycle [41], latent addition [10] and accommodation curve [5] (slope and breakdown of accommodation). Threshold electrotonus is important as it provides insight into internodal conductances in human subjects in vivo, and it is promising for providing insight into disease mechanisms in neurological disorders [42]. In threshold electrotonus, sub-threshold currents are used to alter the nodal and internodal membrane potentials. The change in threshold to a test stimulus is measured during the sub-threshold current, and this pattern of threshold alternations is termed threshold electrotonus [43]. The recovery cycle is a series of threshold fluctuations following an action potential. It is obtained by stimulating with a supra-threshold conditioning pulse and estimating the threshold with a subsequent test stimulus at various inter-stimulus intervals [43]. The threshold is usually tracked up to 200 ms after the conditioning pulse, during which time it goes through the absolute refractory period, relative refractory period, supernormal period and subnormal period. During the refractory and subnormal periods the threshold is increased, whereas it is decreased during the supernormal period [42,44]. Latent addition is obtained in the same manner as the recovery cycle [9,10]. The difference between the recovery cycle and latent addition is that the conditioning pulse is sub-threshold in latent addition but super-threshold in the recovery cycle. The strength-duration time constant τ was determined from the latent addition curve by fitting the function S2 = 100 - 90e-s/τ to the simulated data, where S2 is the threshold of the test stimulus and s is the delay between the sub-threshold conditioning stimulus and the test stimulus. Eleven delays, equally spaced between 0.0 ms and 1.0 ms, were used in this fit (see [10] for a more detailed description of the estimation of the strength-duration time constant using latent addition). The accommodation curve is a plot of the threshold current as a function of the time-constant of current rise for exponentially rising stimuli [5]. Exponentially rising stimuli have the form IS(1-exp(-t/τ)), where τ is the time-constant of current rise.
Five parameters were adjusted in order to fit the model to these experimental data: the nodal resting potential, the internodal leak resistance, the internodal slow potassium conductance, the nodal persistent sodium conductance and the kinetics of the slow potassium channel. Throughout the paper, modeling data are presented as superimposed on the corresponding experimental ranges, a method taken from [24].
Implementation
The model was implemented in C and integrated by Euler's method with a time step of 2 μs. The model was interfaced with Matlab 6.0 as a mex function, and m-functions were written to estimate measurements of axonal excitability. The excitability measurements were based on a binary search algorithm, which determined the excitation threshold with an accuracy of 0.1pA. An action potential was identified if the nerve fiber was depolarized to -30 mV with a rate of rise of more than 60 mV/ms. Stimulation was achieved by an intracellularly-injected current in the node.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KH contributed extensively in all phases of the present study. OKA and LAN contributed to the planning of the study and to the discussion of the results.
==== Refs
Hill AV Excitation and accommodation in nerve Proc R Soc 1936 119 305 355
Lucas K On the rate of variation of the exciting current as a factor in electrical excitation J Physiol (Lond) 1907 36 253 274 16992906
Vallbo AB Accommodation related to inactivation of the sodium permeability in single myelinated nerve fibers from Xenopus laevis Acta Physiol Scand 1964 61 429 444 14209259
Bernhard CG Granit R Skoglund CR The breakdown of accommodation – nerve as a model sense-organ J Neurophysiol 1942 5 55 68
Kugelberg E Accommodation in human nerves Acta Physiol Scand 1944 8 1 115
Baker M Bostock H Depolarization changes the mechanism of accommodation in rat and human motor axons J Physiol 1989 411 545 561 2614732
Bostock H Waxman SG, Kocsis JD, Stys PK Mechanisms of accommodation and adaptation in myelinated axons The Axon 1995 Oxford University Press, Inc 311 327
Baker MD Bostock H Low-threshold, persistent sodium current in rat large dorsal root ganglion neurons in culture J Neurophysiol 1997 77 1503 1513 9084615
Bostock H Rothwell JC Latent addition in motor and sensory fibres of human peripheral nerve J Physiol 1997 498 277 294 9023784
Panizza M Nilsson J Bradley JR Rothwell JC Hallett M The time constants of motor and sensory peripheral nerve fibers measured with the method of latent addition Electroencephalogr Clin Neurophysiol 1994 93 147 154 7512921 10.1016/0168-5597(94)90078-7
Baker MD Bostock H Inactivation of macroscopic late Na+ current and characteristics of unitary late Na+ currents in sensory neurons J Neurophysiol 1998 80 2538 2549 9819261
Baker MD Selective block of late Na(+) current by local anaesthetics in rat large sensory neurones Br J Pharmacol 2000 129 1617 1626 10780966 10.1038/sj.bjp.0703261
Baker MD Bostock H The pH dependence of late sodium current in large sensory neurons Neuroscience 1999 92 1119 1130 10426550 10.1016/S0306-4522(99)00058-5
Bostock H Baker M Grafe P Reid G Changes in excitability and accommodation of human motor axons following brief periods of ischaemia J Physiol 1991 441 513 535 1816385
Reid G Scholz A Bostock H Vogel W Human axons contain at least five types of voltage-dependent potassium channel J Physiol 1999 518 681 696 10420006 10.1111/j.1469-7793.1999.0681p.x
Schwarz JR Reid G Bostock H Action potentials and membrane currents in the human node of Ranvier Pflugers Arch 1995 430 283 292 7675638 10.1007/BF00374660
Bostock H The strength-duration relationship for excitation of myelinated nerve: computed dependence on membrane parameters J Physiol 1983 341 59 74 6312032
Rattay F Analysis of models for external stimulation of axons IEEE Trans Biomed Eng 1986 33 974 977 3770787
Halter J Clark J A distributed-parameter model of the myelinated nerve fiber J Theor Biol 1991 148 345 382 2016898
Stephanova DI Myelin as longitudinal conductor: a multi-layered model of the myelinated human motor nerve fibre Biol Cybern 2001 84 301 308 11324341 10.1007/s004220000213
Patlak J Molecular kinetics of voltage-dependent Na+ channels Physiol Rev 1991 71 1047 1080 1656476
Hodgkin AL Huxley AF A quantitative description of membrane current and its application to conduction and excitation in nerve Bull Math Biol 1952 52 25 71 2185861
Jakobsson E Guttman R The standard Hodgkin-Huxley model and squid axons in reduced external Ca++ fail to accommodate to slowly rising currents Biophys J 1980 31 293 297 7260290
McIntyre CC Richardson AG Grill WM Modeling the excitability of mammalian nerve fibers: influence of afterpotentials on the recovery cycle J Neurophysiol 2002 87 995 1006 11826063
Caldwell JH Schaller KL Lasher RS Peles E Levinson SR Sodium channel Na(v)1.6 is localized at nodes of Ranvier, dendrites, and synapses Proc Natl Acad Sci U S A 2000 97 5616 5620 10779552 10.1073/pnas.090034797
Bostock H Baker M Reid G Changes in excitability of human motor axons underlying post-ischaemic fasciculations: evidence for two stable states J Physiol 1991 441 537 557 1667800
Bostock H Burke D Hales J Differences in behaviour of sensory and motor axons following release of ischaemia Brain 1994 117 225 234 8186950
Baker M Bostock H Grafe P Martius P Function and distribution of three types of rectifying channel in rat spinal root myelinated axons J Physiol 1987 383 45 67 2443652
Pape H Queer current and pacemaker: the hyperpolarization-activated cation current in neurons Annu Rev Physiol 1996 58 299 327 8815797 10.1146/annurev.ph.58.030196.001503
Kocsis JD Eng DL Gordon TR Waxman SG Functional differences between 4-AP and TEA-sensitive potassium channels in mammalian axons Neurosci Lett 1987 75 193 198 2437499 10.1016/0304-3940(87)90296-5
Roper J Schwarz JR Heterogeneous distribution of fast and slow potassium channels in myelinated rat nerve fibres J Physiol 1989 416 93 110 2558178
Eng DL Gordon TR Kocsis JD Waxman SG Development of 4-AP and TEA sensitivities in mammalian myelinated nerve fibers J Neurophysiol 1988 60 2168 2179 2853208
Vabnick I Trimmer JS Schwarz TL Levinson SR Risal D Shrager P Dynamic potassium channel distributions during axonal development prevent aberrant firing patterns J Neurosci 1999 19 747 758 9880595
Foster RE Conners BW Waxman SG Rat optic nerve: Electrophysiological, pharmacological, and anatomical studies during development Dev Brain Res 1982 3 361 376 10.1016/0165-3806(82)90005-0
Kocsis JD Waxman SG Hildebrand C Ruiz JA Regenerating mammalian nerve fibers: Changes in action potential waveform and firing characteristics following blockage of potassium conductance Proc R Soc Lond B 1982 217 277 287
Wesselink WA Holsheimer J Boom HB A model of the electrical behaviour of myelinated sensory nerve fibres based on human data Med Biol Eng Comput 1999 37 228 235 10396827
Scholz A Reid G Vogel W Bostock H Ion channels in human axons J Neurophysiol 1993 70 1274 1279 7693885
Safronov BV Kampe K Vogel W Single voltage-dependent potassium channels in rat peripheral nerve membrane J Physiol (Lond) 1993 460 675 691 8387591
Frankenhauser B Huxley AF The Action Potential In The Myelinated Nerve Fibre of Xenopus Laevis As Computed On The Basis Of Voltage Clamp Data J Physiol 1964 171 302 315 14191481
Yang Q Kaji R Hirota N Kojima Y Takagi T Kohara N Effect of maturation on nerve excitability in an experimental model of threshold electrotonus Muscle Nerve 2000 23 498 506 10716759 10.1002/(SICI)1097-4598(200004)23:4<498::AID-MUS7>3.0.CO;2-U
Kiernan MC Burke D Andersen KV Bostock H Multiple measures of axonal excitability: a new approach in clinical testing Muscle Nerve 2000 23 399 409 10679717 10.1002/(SICI)1097-4598(200003)23:3<399::AID-MUS12>3.0.CO;2-G
Burke D Kiernan MC Bostock H Excitability of human axons Clin Neurophysiol 2001 112 1575 1585 11514239 10.1016/S1388-2457(01)00595-8
Bostock H Cikurel K Burke D Threshold tracking techniques in the study of human peripheral nerve Muscle Nerve 1998 21 137 158 9466589 10.1002/(SICI)1097-4598(199802)21:2<137::AID-MUS1>3.0.CO;2-C
Stys PK Ashby P An automated technique for measuring the recovery cycle of human nerves Muscle Nerve 1990 13 750 758 2385261 10.1002/mus.880130814
Nilsson I Berthold CH Axon classes and internodal growth in the ventral spinal root L7 of adult and developing cats J Anat 1988 156 71 96 3417553
Berthold CH Nilsson I Rydmark M Axon diameter and myelin sheath thickness in nerve fibres of the ventral spinal root of the seventh lumbar nerve of the adult and developing cat J Anat 1983 136 483 508 6885614
Rydmark M Nodal axon diameter correlates linearly with internodal axon diameter in spinal roots of the cat Neurosci Lett 1981 24 247 250 7279292 10.1016/0304-3940(81)90165-8
Rydmark M Berthold CH Electron microscopic serial section analysis of nodes of Ranvier in lumbar spinal roots of the cat: a morphometric study of nodal compartments in fibres of different sizes J Neurocytol 1983 12 537 565 6619904 10.1007/BF01181523
Bostock H Sears TA The internodal axon membrane: electrical excitability and continuous conduction in segmental demyelination J Physiol (Lond) 1978 280 273 301 690876
Tasaki I New measurements of the capacity and the resistance of the myelin sheath and the nodal membrane of the isolated frog nerve fiber Am J Physiol 1955 181 639 650 13238615
| 15826303 | PMC1090618 | CC BY | 2021-01-04 16:39:25 | no | Theor Biol Med Model. 2005 Apr 12; 2:16 | utf-8 | Theor Biol Med Model | 2,005 | 10.1186/1742-4682-2-16 | oa_comm |
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-161582630310.1186/1742-4682-2-16ResearchBreakdown of accommodation in nerve: a possible role for persistent sodium current Hennings Kristian [email protected] Lars [email protected] Ole K [email protected] Center for Sensory-Motor Interaction (SMI), Aalborg University. Frederik Bajers Vej D3-203, 9220 Aalborg Ø, Denmark2005 12 4 2005 2 16 16 8 12 2004 12 4 2005 Copyright © 2005 Hennings et al; licensee BioMed Central Ltd.2005Hennings 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
Accommodation and breakdown of accommodation are important elements of information processing in nerve fibers, as they determine how nerve fibers react to natural slowly changing stimuli or electrical stimulation. The aim of the present study was to elucidate the biophysical mechanism of breakdown of accommodation, which at present is unknown.
Results
A model of a space-clamped motor nerve fiber was developed. It was found that this new model could reproduce breakdown of accommodation when it included a low-threshold, rapidly activating, persistent sodium current. However, the phenomenon was not reproduced when the persistent sodium current did not have fast activation kinetics or a low activation threshold.
Conclusion
The present modeling study suggests that persistent, low-threshold, rapidly activating sodium currents have a key role in breakdown of accommodation, and that breakdown of accommodation can be used as a tool for studying persistent sodium current under normal and pathological conditions.
==== Body
Background
Accommodation is important for information processing in nerve fibers, as it determines whether, and how frequently, slowly-changing natural and artificial stimuli are translated into action potentials. Hill's theory of accommodation in nerve has been one of the most influential theories in this area [1]. A prediction of this theory is that a linearly rising current requires a certain critical slope in order to excite nerve fibers. Although this critical slope has been demonstrated in experimental preparations [2,3], it has not been found under normal physiological conditions [4,5]. Instead, nerve fibers have been shown to exhibit breakdown of accommodation; that is, a long-duration slowly rising current excites nerve fibers at a nearly constant intensity no matter how slowly this intensity is approached [4,5]. A critical slope has only been found for depolarized nerve fibers, and Hill's theory of accommodation has been shown only to be applicable to such fibers [6]. Accommodation and breakdown of accommodation were the foci of several studies before the invention of the voltage-clamp, since prior to this innovation it was one of the few methods by which membrane kinetics could be studied. Since the invention of the voltage-clamp and later the patch-clamp some fifty years ago, the concept of breakdown of accommodation has been virtually absent from the scientific literature [7]. However, the biophysical mechanism responsible for breakdown of accommodation is still unknown; and as will be shown in this paper, a model that only contains transient sodium channels (i.e. currents that activate and deactivate rapidly in response to membrane depolarization) is unable to reproduce the phenomenon. Persistent (no inactivation) and late (slow inactivation) sodium channels have been identified in large dorsal ganglion neurons [8] and it has been found that these channels are needed for modeling latent addition in motor and sensory nerve fibers (i.e. threshold changes to short sub-threshold stimuli [9]). This suggests that persistent or late sodium channels are present in both motor and sensory myelinated nerve fibers and have fast activation kinetics that can initiate action potentials. The present study was undertaken to study the hypothesis that persistent sodium channels create a "threshold region" of membrane depolarization that cannot be exceeded without the generation of an action potential. Thus, it is suggested that persistent sodium channels are the cause of breakdown of accommodation.
The results in the present paper were based on a model of a space-clamped nerve fiber. This model included a persistent sodium channel based on the work of Bostock and Rothwell (1997) [9]. This channel was defined from the transient sodium channel with the following modifications: a) inactivation was removed (a persistent channel); b) the time-constant was slowed by a factor of two (time-constant); and c) the kinetics was displaced so that the channel was activated at a membrane potential 20 mV more negative than is required to activate the transient channel (voltage shift) [9].
Results
Model validation
The structure of the model and the choice of parameters allowed it to reproduce four sets of independent experimental data: threshold electrotonus, recovery cycle, latent addition and breakdown of accommodation (see Figure 1). The model was found to have a strength-duration time constant of 133.2 μs, which is similar to the experimental recorded value for the median nerve (139 ± 59 μs [10]). Furthermore, the model simulated breakdown of accommodation (see Figure 1D). The initial critical slope of the model was found to be 17.3 rheobase/s, which is lower than the experimentally recorded value (21.2 ± 3.72 rheobase/s) for ulnar nerves. The breakdown of accommodation seen in the model was likewise greater than observed in the ulnar nerve, and was closer to that observed in sensory nerve fibers [5]. The accommodation curve flattened out and remained near 2 rheobases when the time-constant of the current rise was greater than ~150 ms.
Figure 1 Comparison of the new model with experimental data for: A) threshold electrotonus [40], B) recovery cycle [41], C) latent addition [10], and D) accommodation curve [5]. In threshold electrotonus, a sub-threshold conditioning pulse of 100 ms duration is used to alter the threshold of a test stimulus delayed with respect to the onset of the conditioning pulse. In the recovery cycle, the nerve fiber is excited by a supra-threshold stimulus and the threshold of a test stimulus is determined at inter-stimulus intervals (TISI) of 2 ms to 100 ms. In latent addition, a short duration sub-threshold conditioning stimulus is used to alter the threshold of a test stimulus; the onset of the test stimulus is delayed with regard to the onset of the conditioning stimulus. In the accommodation curve, the threshold of stimuli of the form IS(1-e-tτ) was determined, where τ was the time-constant of the current rise. In A, the bold line is the initial critical slope, which was estimated from the first four points in the accommodation curve where it is approximately a straight line. Experimental range: a) minimum and maximum of the experimental range, b) and c) mean ± standard deviation.
Breakdown of accommodation
In order to study the relationship between the properties of persistent sodium channels and breakdown of accommodation, one parameter at a time (number of channels, voltage shift, time constant) was changed and its influence on breakdown of accommodation was assessed (see Figure 2). When the voltage shift was decreased from -20 mV to -10 mV or -0 mV, it was still possible to create breakdown of accommodation by increasing the number of persistent sodium channels to 3.75% (-10 mV) or 15% (-0 mV) of the total number of sodium channels (see Figure 3A). However, for a voltage shift of -0 mV, the membrane potential did not return to the resting potential after the generation of an action potential (see Figure 3A).
Figure 2 Relationship between the properties of persistent sodium channels and breakdown of accommodation: A) Number of persistent sodium channels (number of persistent sodium channels: 1.0%, 1.5%, 2.0%, 2.5%, and 3.0%). B) Voltage shift of the kinetics of the persistent sodium channels relative to the transient sodium channels (voltage shift: -0 mV, -5 mV, -10 mV, -15 mV, and -20 mV). C) Time constant of persistent sodium channel activation (time-constant slowed by a factor of: 10, 6.66, 4.5, 3.0, 2.0, and 1.0). A thick line and bold number indicates the default model.
Figure 3 Relationship between voltage shift of persistent sodium channels relative to transient sodium channels and the shape of the action potential. For voltage shifts of -10 mV and -0 mV, the number of persistent sodium channels is set to a value (3.75% and 15%, respectively) that would produce approximately the same degree of breakdown of accommodation as the default model (voltage shift of -20 mV). A) The shape of the action potentials. B) The accommodation curves for the three voltage shifts of -20 mV●, -10 mV■, and -0 mV▲.
Threshold responses to linearly rising stimuli
The threshold responses to linearly rising stimuli were significantly modified by the presence of persistent sodium channels (see Figure 4). (A threshold response is a response to a stimulus with intensity equal to the excitation threshold of the nerve fiber). Without persistent sodium channels, the threshold response to a linearly rising current of 20 ms duration did not occur at the end of the stimulus but had a latency of 5.04 ms (see Figure 4A). With 2.5% persistent sodium channels, the threshold response occurred at the end of the stimulus (see Figure 4B). This difference was found for all linearly rising currents tested that had stimulus durations in the range 1 ms to 200 ms; when the model had 2.5 % persistent sodium channels the threshold responses were always observed at the end of the stimulus, whereas without persistent sodium channels the longest latency of the threshold response was 5.04 ms (see Figure 4C). For both models, with and without persistent sodium channels, a non-linear response always occurred when the membrane was depolarized to a certain threshold value. This non-linear response initially occurred at the end of the stimulus, and with persistent sodium channels it resulted in an action potential. Without persistent sodium channels, it only resulted in an action potential when the stimulus intensity was sufficient for the response to occur with a latency of 5.04 ms or less.
Figure 4 Responses of the new model without (A) and with (B) persistent sodium channels to a linearly rising current; (C) the latencies of the threshold responses for the models without (bold line) and with (thin line) persistent sodium channels. In subfigures (A) and (B) the responses for each model are shown for increasing stimulus intensities: A) 0.950, 0.975, 1.00, and 1.025 excitation threshold, and B) 0.925, 0.950, 0.975, and 1.00 excitation threshold. Threshold responses are drawn with bold lines.
Threshold electrotonus
A proportional relationship between the threshold change of a test stimulus and the underlying electrotonic changes in the membrane potential is a fundamental requirement for threshold electrotonus. Such a relationship was found for the model with 2.5% persistent sodium channels (see Figure 5). However, when the persistent sodium channels were removed from the model, the relationship between threshold and membrane potential broke down. This was tested with the conditioning current at an intensity of 40% for the model with persistent sodium currents, and 30.5% for the model without persistent sodium current. These two conditioning current intensities produced similar membrane depolarizations in the two models, which enabled the effect of the persistent sodium channels to be assessed. The relationship between threshold and membrane potential has also been found to break down when nerve fibers are depolarized because of a long-duration conditioning current or ischaemia [6]. Consequently, had the membrane depolarizations not been matched in the two models, the loss of the relationship between threshold and membrane potential may have been attributable to the greater membrane depolarization in the model without persistent sodium channels.
Figure 5 Electrotonus (A) and threshold electrotonus (B) of the new model with (thick line) and without (thin line) persistent sodium channels. The intensities of the conditioning currents were 40% and 30.5% of the threshold of the test stimulus alone for the model with and without persistent sodium channels, respectively.
Discussion
We have used a model of a space-clamped motor nerve fiber to provide evidence for a link between persistent sodium currents and breakdown of accommodation. The model demonstrated that these channels might be the cause of breakdown of accommodation, as their inclusion enabled the model to reproduce the phenomenon (see Figure 1D). It also demonstrated that such channels are likely to be low-threshold and rapidly activating (see Figure 2). The low-threshold property is further supported by the fact that although breakdown of accommodation can be reproduced by high-threshold persistent sodium channels, in this case it results in an action potential that does not return to the resting potential (see Figure 3).
Experimental evidence for the role of persistent sodium current in breakdown of accommodation
Persistent, late sodium currents have been observed in large dorsal root ganglion cells. These current were found to have a low threshold and fast activation kinetics and were therefore expected to modulate membrane excitability by amplifying and prolonging depolarization from a generator potential or an external electrode [8,11]. Indirect evidence has been obtained for the presence of such channels in both large diameter sensory nerve fibers and motor nerve fibers [9], and they can produce regenerative currents that facilitate action potential generation. Persistent sodium channels have been shown to amplify otherwise sub-threshold depolarization, thereby initiating action potentials [12]. Furthermore, acidification and alkalization within the physiological range have been found respectively to decrease and increase persistent and late sodium currents [13]. This pH-dependence of the late sodium current correlates well with experimental observations of breakdown of accommodation. Hence, breakdown of accommodation has been found to decrease during ischaemia and increase during hyperventilation [5]. Furthermore, when nerve fibers are depolarized with a polarizing current, there is a decrease in the threshold to triangular stimuli [14]. This suggests that it is not membrane depolarization per se that causes loss of breakdown of accommodation and the presence of a critical slope for slowly rising stimuli. In the present study, the loss of breakdown of accommodation is explained by loss of the persistent sodium current, such as would be caused by ischaemic depolarization due to acidification. Consequently, the present study predicts that the critical slope found by [2] was caused by ischaemic acidification and not membrane depolarization.
The effect of persistent sodium channels on threshold responses
In the present study, when the models with and without persistent sodium currents were stimulated by linearly rising currents, non-linear responses always resulted in action potentials when the model exhibited breakdown of accommodation. However, without breakdown of accommodation, non-linear responses only resulted in action potentials when they occurred within a "critical latency" from the onset of the stimulus. Without a "critical latency", which is the case with breakdown of accommodation, the threshold response occurred at the cessation of a long linearly rising stimulus. Consequently, the threshold for such stimuli is nearly constant regardless of their duration. However, when there is a "critical latency", the membrane potential needs to reach the voltage threshold within this "critical latency" for the nerve fiber to fire an action potential. The critical slope will then be proportional to the voltage threshold for which a non-linear response occurs divided by the "critical latency"; i.e. a decrease in "critical latency" results in an increased critical slope.
Model limitations
The present model simplifies existing knowledge of neuronal morphology and the distribution of ion channels. With regard to ion channels, only two potassium channels and two sodium channels were included in the model, but at least five distinct potassium channels [15] and three persistent and late sodium channels [11] have been identified, besides the classical transient sodium channel [16]. Unfortunately, current knowledge of the potassium and sodium channels in motor nerve fibers does not provide enough detail to allow modeling of them all. For example, the channel densities and kinetics are not known for all five potassium channels [15], and the kinetic data we have for the slow and fast potassium channels are likely to represent amalgamations of several channel species into single stereotypes [16]. Consequently, the present model is based on an amalgamation of distinct channels into stereotypes and the detailed geometrical structures of motor nerve fibers into a gross equivalent electrical circuit. The parameters of the model were based on experimental current- and voltage-clamp recordings whenever possible and the results obtained were found to be in line with experimental work. Consequently, these simplifications appear justified and therefore provide a basis for studying the biophysical properties of breakdown of accommodation. This assumption is supported by previous work where models have provided insights into biophysical mechanisms [9,17,18]. However, the internodal leak resistance (RIL) in particular was not based on experimental data, but was instead set by trial and error to a value that would enable the model to reproduce known experimental data. This approach was used since few experimental data on the internodal leak resistance are available. There are only modeling data on the periaxonal resistivity [19], but further modeling data suggest that the longitudinal conductance of the myelin sheaths has to be taken into account in determining the internodal leak resistance [20]. Consequently, an internodal leak resistance based solely on the width of the periaxonal space is likely to be an underestimate. The unknown resistivity of the periaxonal space presents further difficulties in obtaining a value for internodal leak resistance on the basis of experimental data alone. For these reasons we believe that the present approach was justified.
Alternative explanations of breakdown of accommodation
An alternative explanation for breakdown of accommodation could be the gating mode of the transient sodium channel [21]. The present paper follows the convention of assuming that activation and inactivation are two independent processes (i.e. the formalism of [22]). Today, it is known that activation and inactivation are inter-dependent, and that most transient sodium channels will go through an open state before entering an inactivated state [21]. This difference between the Hodgkin and Huxley formalism and recent knowledge of transient sodium channel function may have a synergistic role in breakdown of accommodation. Hence, a transient sodium channel with little inactivation before channel opening would not permit a critical slope and loss of breakdown of accommodation. However, this explanation remains unproven and would not change the conclusion of the present study, that persistent and late sodium channels can cause breakdown of accommodation. The interdependence of transient channel activation and inactivation may change the densities of persistent sodium channels needed for creating breakdown of accommodation, and thus there may be synergism between transient and persistent sodium channels.
A second explanation may be m-h overlap in the activation/inactivation kinetics of the transient sodium channel. For transient sodium channels, there is a region of membrane depolarization in which a persistent sodium current is generated l [23]. This is caused by channel activation while the membrane is still not sufficiently depolarized for all the channels to be inactivated, a phenomenon that has been termed m-h overlap. A theoretical study has demonstrated that the original squid axon model of Hodgkin and Huxley has breakdown of accommodation as a result of m-h overlap [23]. In this paper and other studies [9,24], persistent sodium channels are modeled as discrete channels. However, this does not imply that they are physically different from transient sodium channels. Three discrete persistent and late sodium currents have been identified on the basis of inactivation kinetics [11] in addition to the classical transient sodium current [16], but only one sodium channel Nav (1.6) has been found in the nodes of Ranvier in large peripheral nerve fibers [25]. This may suggest that persistent and late sodium currents are not generated specifically, but instead by transient sodium channels that operate in a gating mode with no or slowed channel inactivation. The modeling of persistent sodium current as created by persistent sodium channels does not provide evidence for the existence of such channels, only evidence that persistent sodium current can lead to breakdown of accommodation. Consequently, such persistent sodium current may be created by m-h overlap. However, in studies on persistent sodium currents, it has been argued that m-h overlap is not consistent with the observed kinetics [8,11]. Evidently, in mammalian nerve fibers, the persistent sodium current is most likely not generated by m-h overlap; but the study of [23] suggests that m-h overlap may be important for the persistent sodium current and breakdown of accommodation in squid axons.
Conclusion
The present modeling study has demonstrated that persistent sodium currents can create a "threshold region" for membrane depolarization that cannot be exceeded without the generation of an action potential. Thus, a persistent sodium current may be the underlying biophysical mechanism for the breakdown of accommodation to slowly rising currents, which are observed under normal physiological conditions in mammalian nerve fibers [4,5]. This suggests that accommodation curves can be used as a tool for studying persistent sodium currents under normal and pathological conditions.
Methods
Electrical model of a motor nerve fiber
The structure of the model of the space-clamped motor nerve fiber was based on previous models used for studying the accommodative properties of such fibers [9,26,27]. The present model represents a motor nerve fiber by the electrical equivalent circuit shown in Figure 6. The geometry of the node and internode was based on studies on the morphology of cat ventral spinal roots. The geometrical parameters were taken from cats of 1–11 years of age for a motor nerve fiber with a diameter of 14 μm (see Table 1). The nodal, internodal and myelin capacitances in the electrical equivalent circuit were calculated on the basis of these geometrical parameters and experimentally estimated capacitances per square micrometer (see Table 2 and Figure 6). The internodal leak resistance (Ril) and nodal resting potential were set by trial and error rather than calculated from geometrical and electrical parameters (see section entitled 'Validation', below).
Table 1 Geometrical parameters
Inter-nodal length (L) 1.37 mm [45]
Inter-nodal diameter (di) 8.8 μm [46]
Nodal diameter (dn) 3.5 μm [47]
Nodal length (l) 1 μm [48]
Number of myelin lamella (N) 141 [46]
Table 2 Electrical parameters
Nodal capacitance (cn) 2 μF/cm2 [39]
Internodal capacitance (ci) 1 μF/cm2 [49]
Myelin capacitance (cm) 0.1 μF/cm2 [50]
Figure 6 Equivalent circuit for a space-clamped motor neuron. The model consisted of a node and an internode. Both the node and the internode contained non-linear current sources, which were calculated from equilibrium potentials and conductances. Channel types and maximum ionic conductances: node, transient sodium (Nat, 276nS), persistent sodium (Nap, 7.1nS), fast potassium (Kf, 4.1nS), and slow potassium (Ks, 17.4nS); internode, slow potassium (Ks, 87.1nS) and leak conductance (L, 1.7nS). The linear parameters of the model were: Cn, nodal capacity (0.22pF), Ci, internodal capacity (379pF), Cm, capacity of the myelin sheath (0.17pF), and Ril, internodal leak resistance (41 MΩ).
Ionic currents
Five major ionic currents have been identified in myelinated nerve fibers as necessary for modeling a wide variety of experimental data: the transient sodium current (Nat) for modeling the action potential [22], and the persistent sodium current (Nap) for modeling latent addition [9] and the recovery cycle [24]. Fast (Kf) and slow (Ks) potassium currents have been shown to explain accommodation to depolarizing conditioning currents [28]. Accommodation to hyperpolarizing currents can be explained by a hyperpolarization-activated cation conductance (IH), which is also thought to limit hyperpolarization in nerve fibers after they have conducted a train of impulses [28,29].
Transient and persistent sodium channels were included in the node, but following the work of [9] they were omitted from the internode for simplicity. The hyperpolarization-activated cation conductance was omitted from the model as it does not influence the response of nerve fibers to depolarizing stimuli [14]. Based on the work of [28,30,31]., the slow potassium current was included in the node as well as the internode. There is evidence for the localization of fast potassium channels in the paranode [32-35] As the paranode was not included in the present model, it was impossible to include fast potassium channels at this location. Instead, the approach used by [36] was applied and the fast potassium channels were included in the node. The ionic currents were described as being generated through membrane conductances (see Figure 6). The sodium conductances and slow potassium conductance in the node were based on single channel conductances and channel densities. Single channel conductances of 13pS and 8pS were used for the sodium channels and slow potassium channels, respectively [37]. The nodal densities for the sodium and slow potassium channels were set to 1000 channels/μm2 [37] and 100 channels/μm2 [38], respectively. The ion conductance of the fast potassium current was based on the work of [16], who found a fast potassium conductance of 15nS and a capacitive load of 1.4pF on the nodal membrane. The conductance of the fast potassium current was set from an estimate of the membrane area [16], which was based on the nodal capacitance in experimental data and the nodal capacitance per square micrometer [39].
The nodal resting potential was kept stable by a current leak to the internode, and the internodal resting potential was determined from this relationship. The internodal resting potential was kept stable by a small internodal sodium leak conductance. The nodal persistent sodium conductance was set by the fraction of nodal sodium channels that would be persistent. Therefore, the total number of nodal sodium channels was kept constant for all simulations.
Membrane kinetics
The non-linear membrane dynamics were based on human data [16]. The ionic current was given as: transient sodium current iNat = GNatm3h(E-ENa), persistent sodium current iNap = GNapp3(E-ENa), fast potassium current iKf = GKfn4(E-EK), and slow potassium current iKs = GKss(E-EK). The fractional activations (m, h, p, n and s) were given by the differential equation:
dx/dt = αx(1-x)-βxx, for x = m, h, p, n, s
where αm, αp, αn, αs = A(E-B)/(1-exp((B-E)/C)); βm, αh, βp, βn, βs = A(B-E)/(1-exp((E-B)/C)), βh = A/(1+exp((B-E)/C)) (see Table 3 for the constant: A, B, and C) and E is the membrane potential. The rate constants (αx and βx) where scaled by appropriate Q10 factors to a temperature of 37°C (see Table 3). All membrane kinetics were the same as those of [16] except the kinetics for the slow potassium current (see section entitled 'Validation', below). In order to allow the model to reproduce threshold electrotonus, it was necessary to modify the slow potassium channels. The kinetics were changed in order to slow the channel activation and to lower the fraction of open slow potassium channels at the resting potentials for the node and internode.
Table 3 Rate constants
Q10 A B C
ms-1 mV mV
αm 2.2 1.86 -18.4 10.3
βm 2.2 0.086 -22.7 9.16
αh 2.9 0.0336 -111.0 11.0
βh 2.9 2.3 -28.8 13.4
αp 2.2 0.93 -38.4 10.3
βp 2.2 0.043 -42.7 9.16
αn 3.0 0.00798 -93.2 1.1
βn 3.0 0.0142 -76.0 10.5
αs 3.0 0.0122 -12.5 16.9
βs 3.0 0.000736 -80.1 12.6
Validation
The model was validated with four sets of experimental data: threshold electrotonus [40], recovery cycle [41], latent addition [10] and accommodation curve [5] (slope and breakdown of accommodation). Threshold electrotonus is important as it provides insight into internodal conductances in human subjects in vivo, and it is promising for providing insight into disease mechanisms in neurological disorders [42]. In threshold electrotonus, sub-threshold currents are used to alter the nodal and internodal membrane potentials. The change in threshold to a test stimulus is measured during the sub-threshold current, and this pattern of threshold alternations is termed threshold electrotonus [43]. The recovery cycle is a series of threshold fluctuations following an action potential. It is obtained by stimulating with a supra-threshold conditioning pulse and estimating the threshold with a subsequent test stimulus at various inter-stimulus intervals [43]. The threshold is usually tracked up to 200 ms after the conditioning pulse, during which time it goes through the absolute refractory period, relative refractory period, supernormal period and subnormal period. During the refractory and subnormal periods the threshold is increased, whereas it is decreased during the supernormal period [42,44]. Latent addition is obtained in the same manner as the recovery cycle [9,10]. The difference between the recovery cycle and latent addition is that the conditioning pulse is sub-threshold in latent addition but super-threshold in the recovery cycle. The strength-duration time constant τ was determined from the latent addition curve by fitting the function S2 = 100 - 90e-s/τ to the simulated data, where S2 is the threshold of the test stimulus and s is the delay between the sub-threshold conditioning stimulus and the test stimulus. Eleven delays, equally spaced between 0.0 ms and 1.0 ms, were used in this fit (see [10] for a more detailed description of the estimation of the strength-duration time constant using latent addition). The accommodation curve is a plot of the threshold current as a function of the time-constant of current rise for exponentially rising stimuli [5]. Exponentially rising stimuli have the form IS(1-exp(-t/τ)), where τ is the time-constant of current rise.
Five parameters were adjusted in order to fit the model to these experimental data: the nodal resting potential, the internodal leak resistance, the internodal slow potassium conductance, the nodal persistent sodium conductance and the kinetics of the slow potassium channel. Throughout the paper, modeling data are presented as superimposed on the corresponding experimental ranges, a method taken from [24].
Implementation
The model was implemented in C and integrated by Euler's method with a time step of 2 μs. The model was interfaced with Matlab 6.0 as a mex function, and m-functions were written to estimate measurements of axonal excitability. The excitability measurements were based on a binary search algorithm, which determined the excitation threshold with an accuracy of 0.1pA. An action potential was identified if the nerve fiber was depolarized to -30 mV with a rate of rise of more than 60 mV/ms. Stimulation was achieved by an intracellularly-injected current in the node.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KH contributed extensively in all phases of the present study. OKA and LAN contributed to the planning of the study and to the discussion of the results.
==== Refs
Hill AV Excitation and accommodation in nerve Proc R Soc 1936 119 305 355
Lucas K On the rate of variation of the exciting current as a factor in electrical excitation J Physiol (Lond) 1907 36 253 274 16992906
Vallbo AB Accommodation related to inactivation of the sodium permeability in single myelinated nerve fibers from Xenopus laevis Acta Physiol Scand 1964 61 429 444 14209259
Bernhard CG Granit R Skoglund CR The breakdown of accommodation – nerve as a model sense-organ J Neurophysiol 1942 5 55 68
Kugelberg E Accommodation in human nerves Acta Physiol Scand 1944 8 1 115
Baker M Bostock H Depolarization changes the mechanism of accommodation in rat and human motor axons J Physiol 1989 411 545 561 2614732
Bostock H Waxman SG, Kocsis JD, Stys PK Mechanisms of accommodation and adaptation in myelinated axons The Axon 1995 Oxford University Press, Inc 311 327
Baker MD Bostock H Low-threshold, persistent sodium current in rat large dorsal root ganglion neurons in culture J Neurophysiol 1997 77 1503 1513 9084615
Bostock H Rothwell JC Latent addition in motor and sensory fibres of human peripheral nerve J Physiol 1997 498 277 294 9023784
Panizza M Nilsson J Bradley JR Rothwell JC Hallett M The time constants of motor and sensory peripheral nerve fibers measured with the method of latent addition Electroencephalogr Clin Neurophysiol 1994 93 147 154 7512921 10.1016/0168-5597(94)90078-7
Baker MD Bostock H Inactivation of macroscopic late Na+ current and characteristics of unitary late Na+ currents in sensory neurons J Neurophysiol 1998 80 2538 2549 9819261
Baker MD Selective block of late Na(+) current by local anaesthetics in rat large sensory neurones Br J Pharmacol 2000 129 1617 1626 10780966 10.1038/sj.bjp.0703261
Baker MD Bostock H The pH dependence of late sodium current in large sensory neurons Neuroscience 1999 92 1119 1130 10426550 10.1016/S0306-4522(99)00058-5
Bostock H Baker M Grafe P Reid G Changes in excitability and accommodation of human motor axons following brief periods of ischaemia J Physiol 1991 441 513 535 1816385
Reid G Scholz A Bostock H Vogel W Human axons contain at least five types of voltage-dependent potassium channel J Physiol 1999 518 681 696 10420006 10.1111/j.1469-7793.1999.0681p.x
Schwarz JR Reid G Bostock H Action potentials and membrane currents in the human node of Ranvier Pflugers Arch 1995 430 283 292 7675638 10.1007/BF00374660
Bostock H The strength-duration relationship for excitation of myelinated nerve: computed dependence on membrane parameters J Physiol 1983 341 59 74 6312032
Rattay F Analysis of models for external stimulation of axons IEEE Trans Biomed Eng 1986 33 974 977 3770787
Halter J Clark J A distributed-parameter model of the myelinated nerve fiber J Theor Biol 1991 148 345 382 2016898
Stephanova DI Myelin as longitudinal conductor: a multi-layered model of the myelinated human motor nerve fibre Biol Cybern 2001 84 301 308 11324341 10.1007/s004220000213
Patlak J Molecular kinetics of voltage-dependent Na+ channels Physiol Rev 1991 71 1047 1080 1656476
Hodgkin AL Huxley AF A quantitative description of membrane current and its application to conduction and excitation in nerve Bull Math Biol 1952 52 25 71 2185861
Jakobsson E Guttman R The standard Hodgkin-Huxley model and squid axons in reduced external Ca++ fail to accommodate to slowly rising currents Biophys J 1980 31 293 297 7260290
McIntyre CC Richardson AG Grill WM Modeling the excitability of mammalian nerve fibers: influence of afterpotentials on the recovery cycle J Neurophysiol 2002 87 995 1006 11826063
Caldwell JH Schaller KL Lasher RS Peles E Levinson SR Sodium channel Na(v)1.6 is localized at nodes of Ranvier, dendrites, and synapses Proc Natl Acad Sci U S A 2000 97 5616 5620 10779552 10.1073/pnas.090034797
Bostock H Baker M Reid G Changes in excitability of human motor axons underlying post-ischaemic fasciculations: evidence for two stable states J Physiol 1991 441 537 557 1667800
Bostock H Burke D Hales J Differences in behaviour of sensory and motor axons following release of ischaemia Brain 1994 117 225 234 8186950
Baker M Bostock H Grafe P Martius P Function and distribution of three types of rectifying channel in rat spinal root myelinated axons J Physiol 1987 383 45 67 2443652
Pape H Queer current and pacemaker: the hyperpolarization-activated cation current in neurons Annu Rev Physiol 1996 58 299 327 8815797 10.1146/annurev.ph.58.030196.001503
Kocsis JD Eng DL Gordon TR Waxman SG Functional differences between 4-AP and TEA-sensitive potassium channels in mammalian axons Neurosci Lett 1987 75 193 198 2437499 10.1016/0304-3940(87)90296-5
Roper J Schwarz JR Heterogeneous distribution of fast and slow potassium channels in myelinated rat nerve fibres J Physiol 1989 416 93 110 2558178
Eng DL Gordon TR Kocsis JD Waxman SG Development of 4-AP and TEA sensitivities in mammalian myelinated nerve fibers J Neurophysiol 1988 60 2168 2179 2853208
Vabnick I Trimmer JS Schwarz TL Levinson SR Risal D Shrager P Dynamic potassium channel distributions during axonal development prevent aberrant firing patterns J Neurosci 1999 19 747 758 9880595
Foster RE Conners BW Waxman SG Rat optic nerve: Electrophysiological, pharmacological, and anatomical studies during development Dev Brain Res 1982 3 361 376 10.1016/0165-3806(82)90005-0
Kocsis JD Waxman SG Hildebrand C Ruiz JA Regenerating mammalian nerve fibers: Changes in action potential waveform and firing characteristics following blockage of potassium conductance Proc R Soc Lond B 1982 217 277 287
Wesselink WA Holsheimer J Boom HB A model of the electrical behaviour of myelinated sensory nerve fibres based on human data Med Biol Eng Comput 1999 37 228 235 10396827
Scholz A Reid G Vogel W Bostock H Ion channels in human axons J Neurophysiol 1993 70 1274 1279 7693885
Safronov BV Kampe K Vogel W Single voltage-dependent potassium channels in rat peripheral nerve membrane J Physiol (Lond) 1993 460 675 691 8387591
Frankenhauser B Huxley AF The Action Potential In The Myelinated Nerve Fibre of Xenopus Laevis As Computed On The Basis Of Voltage Clamp Data J Physiol 1964 171 302 315 14191481
Yang Q Kaji R Hirota N Kojima Y Takagi T Kohara N Effect of maturation on nerve excitability in an experimental model of threshold electrotonus Muscle Nerve 2000 23 498 506 10716759 10.1002/(SICI)1097-4598(200004)23:4<498::AID-MUS7>3.0.CO;2-U
Kiernan MC Burke D Andersen KV Bostock H Multiple measures of axonal excitability: a new approach in clinical testing Muscle Nerve 2000 23 399 409 10679717 10.1002/(SICI)1097-4598(200003)23:3<399::AID-MUS12>3.0.CO;2-G
Burke D Kiernan MC Bostock H Excitability of human axons Clin Neurophysiol 2001 112 1575 1585 11514239 10.1016/S1388-2457(01)00595-8
Bostock H Cikurel K Burke D Threshold tracking techniques in the study of human peripheral nerve Muscle Nerve 1998 21 137 158 9466589 10.1002/(SICI)1097-4598(199802)21:2<137::AID-MUS1>3.0.CO;2-C
Stys PK Ashby P An automated technique for measuring the recovery cycle of human nerves Muscle Nerve 1990 13 750 758 2385261 10.1002/mus.880130814
Nilsson I Berthold CH Axon classes and internodal growth in the ventral spinal root L7 of adult and developing cats J Anat 1988 156 71 96 3417553
Berthold CH Nilsson I Rydmark M Axon diameter and myelin sheath thickness in nerve fibres of the ventral spinal root of the seventh lumbar nerve of the adult and developing cat J Anat 1983 136 483 508 6885614
Rydmark M Nodal axon diameter correlates linearly with internodal axon diameter in spinal roots of the cat Neurosci Lett 1981 24 247 250 7279292 10.1016/0304-3940(81)90165-8
Rydmark M Berthold CH Electron microscopic serial section analysis of nodes of Ranvier in lumbar spinal roots of the cat: a morphometric study of nodal compartments in fibres of different sizes J Neurocytol 1983 12 537 565 6619904 10.1007/BF01181523
Bostock H Sears TA The internodal axon membrane: electrical excitability and continuous conduction in segmental demyelination J Physiol (Lond) 1978 280 273 301 690876
Tasaki I New measurements of the capacity and the resistance of the myelin sheath and the nodal membrane of the isolated frog nerve fiber Am J Physiol 1955 181 639 650 13238615
| 15831096 | PMC1090619 | CC BY | 2021-01-04 16:39:28 | no | J Transl Med. 2005 Apr 14; 3:16 | latin-1 | J Transl Med | 2,005 | 10.1186/1479-5876-3-16 | oa_comm |
==== Front
J Transl MedJournal of Translational Medicine1479-5876BioMed Central London 1479-5876-3-191586213010.1186/1479-5876-3-19CommentaryThe not-for-profit form and translational research: Kerr revisited? Joiner Keith A [email protected] University of Arizona College of Medicine, Office of the Dean, University of Arizona Health Sciences Center, 1501 N. Campbell Avenue, Room 2205, Tucson, Arizona, 85724, USA2005 29 4 2005 3 19 19 11 2 2005 29 4 2005 Copyright © 2005 Joiner; licensee BioMed Central Ltd.2005Joiner; 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.
Translational research conducted in academic health centers is confounded by the organizational structure in which the work is performed. Investigators must obtain research funding and appropriate recognition as a part of a research team in a not-for-profit environment which has more readily rewarded basic work, and individual accomplishments. What results is a unique form of conflict of interest, best understood by relating the basic principles underlying the not-for-profit form to the conduct of translational research in the AHC setting.
==== Body
Introduction
In his classic article entitled "On the Folly of Rewarding A while Hoping for B", Kerr outlines a common phenomenon in organizational management and behavior [1]. Employees receive explicit or implicit incentives (rewards) to carry out one set of activities (A), while the organization hopes that the same employees will also or instead carry out a different set of activities (B). Conflict of interest, not of the conventional type [2], is created between what is rewarded and what is desired. This situation often helps to explain the dichotomy between desired and actual results. Kerr cites teaching in higher educational settings as a prime example. Faculty are expected by the organization to be devoted and excellent teachers, yet are rewarded for endeavors other than teaching.
The premise of this article is that the same dichotomy operates for translational research in AHC. The dichotomy emanates largely from the fact that AHC translational research is typically conducted in the not-for-profit (NPO) setting [3-7]. The dearth of physician-scientists working at the translational interface is one consequence.
Forces Promoting the Not-for-Profit Form
In her book, "Strategic Management for Nonprofit Organizations", Sharon Oster suggests three forces which promote the existence of the not-for-profit form [3]. A paramount consideration is contract failure. Many if not most NPO produce goods or services which are difficult to evaluate in terms of quality. Under these circumstances, it is both problematic and inadvisable to write or enforce a contract between the provider and the stakeholders/consumer/buyer. Instead, provision of services must be based on trust between these two entities. This is the most direct manifestation of public trust. A second force promoting NPO is public failure, typically meaning that the government cannot or will not provide the service, even though the service is considered a public good. The last feature is worker sorting. Individuals are attracted to work for NPO based on the altruistic mission of the organization. Employees often serve as volunteers or at compensation levels below the market, if compared with for profit enterprises.
These three forces are inextricably linked to the most commonly appreciated feature of NPO: the non-distribution constraint, which in turn is linked to the special tax status. The non-distribution constraint states that "any financial surplus generated by operations cannot be distributed to those in control of the operation", which includes the stakeholders (in theory the public) and the employees of the NPO. While this does not preclude NPO from having positive cash flows, increasing net assets, or growing balances in endowment accounts, funds from these accounts cannot be proportionately distributed in the form of bonuses, or other direct benefits.
To what extent do non-profit Academic Health Centers (AHC) conform to these criteria and to what extent do they diverge? Answers to this question vary for the clinical, educational, and research missions. This article will not address the former two areas, and will approach only a subset of the latter – basic and disease-oriented research (as defined by Goldstein and Brown [8]), with a particular focus on translational research. Patient-oriented research, and specifically therapeutic trials, are best considered under other paradigms.
Mission Statements and Research in AHC
How can the goal of the research mission within AHC best be summarized? For most NPO, the mission statement represents the most succinct and important expression of the mission. A review of web sites for more than 100 AHC revealed that only 1 in 5 had readily accessible and explicit institutional mission statements. From these mission statements, the statement referable to research was evaluated. More explicit vision statements accompanied the mission statements in some cases. A nearly constant theme was that research should be translated into advantage for human health. While this informal survey suggests that AHC should be more attentive to explicit statements of mission and vision, it also demonstrates the expected and hoped for mandate that translational investigation be the goal of research in the not-for-profit AHC setting.
Translational Research and Contract Failure in AHC
If the research mission of AHC is to make basic discoveries which are translated into advantage for maintaining or improving human health, does contract failure exist for the translational research mission? Stakeholders (the public) cannot easily judge the quality of scientific research conducted by individual investigators, nor can they readily determine the likelihood that advances will be translated into new diagnostics, therapeutics or preventative strategies. By these criteria, the conditions for contract failure are met.
Given the existence of contract failure, to what extent does the non-profit form provide reassurances that individual investigators will neither provide a sub-standard product, nor work towards goals which are not mission-based.? With regard to the first issue, the guarantees for quality are strict, since investigators are scrutinized by their peers for the quality of the research conducted.
With regard to the second issue, the answer is much more ambiguous when considering research funded through the National Institutes of Health or other federal sources (Department of Veteran's Affairs, Department of Defense). There is a widely recognized tendency for review groups to fund well written applications for which the hypothesis being addressed is important, the preliminary data are compelling, but most importantly, the proposed experiments are likely to be successful and can be completed in the time frame of the award (personal observation). Direct applicability to human health may be discussed, but rarely is used as an essential criteria in the funding decision. Highly creative approaches, which the investigator believes provide the best chance for making a real impact for human health, will typically not be funded without substantial preliminary data to indicate that they will be successful. The complexities of conducting translational research, including insuring sufficient access to patients and volunteers, meeting regulatory challenges, conducting research in an only partially controlled setting, and dealing with expensive infrastructure, all preclude preparation of applications which have the crispness and apparent sophistication of more basic proposals. Similarly, extremely applied experiments, which the scientist believes will most rapidly translate into an advantage for human health, are often poorly received because they are not considered sufficiently innovative. (Of note, in the for-profit arena, these latter constraints are minimized, given the imperative to establish marketable products). While funding agencies and AHC are vocal advocates of translational research, reward systems are structured to support less risky basic activities. Kerr is at work.
These contentions cannot be readily proved or disproved, because there is no set of uniform criteria by which to make these judgments. The most relevant data comes from comparison of NIH funding success rates for clinical vs. non-clinical research. Two analyses [9,10], spanning an era (1997–2004) where substantial emphasis was placed on funding clinical research, revealed substantially higher success rates for non-clinical than for clinical applications. More general perspectives reach the same conclusion [11-13] Some of the above assertions are recognized by the National Institutes of Health (NIH). As stated in the description of high risk research in the NIH Roadmap , "the NIH peer-review process is oriented to fund so-called 'low-risk' proposals that advance well-established areas of science. This leaves many more speculative, or 'high-risk', proposals without an obvious mechanism of NIH support". One proposed solution is the Directors Pioneer award . The Pioneer award provides up to $500,000 per year of direct costs for 5 years for investigators to explore ideas that are considered risky at their inception, yet for which strong preliminary data is lacking.
In contrast to the mandate for funding the Pioneer award, the usually unstated "logic" of review groups for more common funding mechanisms (e.g. R01) is different. The assumption is made that allocating funds towards "safer" projects, even when quite basic and without direct connection to human disease, is ultimately the best way to accomplish the mission, since it is probable that those discoveries will ultimately contribute directly or indirectly to more applied solutions.
Irrespective of whether this "logic" is or can ultimately be validated, the premise of this manuscript is that it creates a conflict of interest for translationally-oriented investigators. This conflict is typified in published papers and in applications for research funding. It is commonplace for the introduction and discussion/conclusion of grant applications and manuscripts to make reference to the relevance of the work for advancing human health. While often done quite tangentially, this addresses the mission of central importance to the public and arguably to the AHC. If the author were to state explicitly that the likelihood was extremely low that the proposed or completed research would ultimately have any significant relevance to human health (or would even be cited by other authors), the application or manuscript would be seriously compromised. In contrast, throughout the remainder of the application or manuscript (and hence in the work itself), the investigator is compelled to focus on a "mission" which is different – convincing fellow scientists that the work is valid, irrespective of its relationship to the broader mission of importance to the public. If instead, the requirement at each step of the work was to make the connection to human disease, the form of the experiments would be often need to be altered in a fashion precluding the rigorous proof expected of high level scientific investigation.
In the for-profit arena, by contrast, a focus on product development with direct applicability to human health necessarily dominates the investigative and development process. Funding decisions are made internally. Preparation of peer-reviewed manuscripts is not an expected end-goal of much of the investigation. Research form follows desired organizational outcome more directly.
Translational Research and Public Failure in AHC
Does public failure exist for the translational research mission of AHC?
Leaving aside the legitimate debate as to whether advancements in biomedical research in AHC are public goods, it is a reasonable assumption that government cannot or will not directly provide the vast array of biomedical advances desired by the public. What is and what should be the linkage between government and NPO in addressing this public failure? NPO often serve as a mechanism for delivery of public services, in partnership with the government. In these situations (for example education, transportation, environment), the government establishes the services to be provided, and contracts with NPO to provide those services.
Is this model applicable to biomedical research? In some respects yes, but in many respects no. By far the largest provider of funds for biomedical research in the United States is the federal government, through the National Institutes of Health. Who determines the criteria for providing the funds to the investigators, and what is the role of government in insuring that there is an applied outlet for the work [14]? The substantive scientific conditions, which are the dominant consideration in determining the likelihood of funding success, are established by the investigators themselves, in the form of study sections and review groups. These panels typically consist of investigators who have obtained NIH funding, usually in a somewhat related area, and who will generally apply for additional NIH grant support in the future. In other words, the grant is effectively awarded by the investigators for the investigators, a notion far quite far astray from the concept of public failure. No matter how supportive the investigators may be of the government's overarching goal, it is both human nature and part of one's training to support "good science" (the kind of science they consider themselves to do and to be rewarded for) over other factors. While the administrative conditions for the award are set by the NIH, while budget allocations to individual institutes or for particular initiatives do influence the scope of the funded research, and while NIH program officers have some discretion in insuring that grant awards fulfill program goals, it is not even clear that budget allocations faithfully match the burden of disease or government intent [15-17].
These considerations are not meant to call into question the rigor or validity of the review process, both of which are held in high regard. Similarly, it remains the case that the public recognizes the value of basic research, temporizing the extent of "public failure". Nonetheless, this perspective indicates that the hope for more and better translational research is not faithfully aligned with the reward systems for most investigators. Based on the considerations discussed above with regard to contract failure, it illustrates the potential conflict of interest between the government and the public on the one hand, and the government and the academic research community on the other [14,18,19].
Translational Research and Worker Sorting in AHC
The pure concept of worker sorting suggests that biomedical investigators work in AHC based on the organizational mission, and at some personal sacrifice, at least financially. Is this formulation largely correct for the translational research enterprise? Only partially.
Investigators wishing to establish an independent career in research will nearly always opt to do so in AHC. The attraction of doing so is not primarily the mission of the organization, but rather the freedom of choice and the breadth of opportunities which are provided to excel. This is particularly true for bona fide translational research, which is most easily conducted in the environment of an AHC. Such opportunities, particularly for more junior investigators, are simply not present in for-profit research enterprises. In AHC, investigators strive to gain both recognition and support, for purposes of career advancement and career stability. If such individuals respond to a mission statement at all, it is likely to be to the professional organization/s of which they are members or occasionally to a much broader mission statement which transcends the institution. Investigators can even be likened to mini-religious organizations, with a fervent zeal for their own research which dominates other considerations [20]. The institutional mission and organizational structure of the AHC may even conflict with the individual mission statement of the investigator, the latter of which often promotes and condone individual pursuits but may or may not be aligned with those of the organization [21]. At the same time, the rewards to the investigator for conducting translational research transcend some of these considerations, and include improved patient recruitment, with attendant increases in patient volume and clinical income, increased personal recognition, and funds from licensing agreements for intellectual property.
Programmatic and team-based investigation is increasingly recognized as the most rapid and efficient structure for scientific progress for many if not most questions – i.e. to accomplish the research mission. In the for-profit-setting, rewards can be readily distributed based on the performance of the organization, and/or of individual teams within the organization. In AHC, academic promotion, research funding and professional recognition are far more readily achieved from an individual standpoint. Once again leaving aside the cogent arguments supporting both sides of this dilemma, individual accomplishment is most readily rewarded, while funding agencies, scientific organizations, AHC and the public hope that team-based research will predominate.
There is a growing body of literature describing the impediments to interdisciplinary research (IDR) in AHC, and suggesting strategies to remove those impediments [22]. Based on a survey of over 300 respondents, the five most important recommendations for institutions to facilitate IDR were a.) fostering a collaborative environment, b.) providing faculty incentives and organizing hiring policies and tenure processes to support IDR, c.) providing seed funding, d.) creating cross-department budgeting mechanisms, and e.) generating strategic plans that promote IDR.
Of importance, NIH Roadmap initiatives have established a series of awards which facilitate interdisciplinary research . These new awards provide funding for training of scientists in interdisciplinary strategies, for creation of interdisciplinary centers, and for conferences which catalyze collaboration between the life and physical sciences. In turn, AHC must respond by reconfiguring the reward system (promotions, professional recognition) to accurately value team-based research. Otherwise a new manifestation of conflict of interest will result.
This requires new strategies to appropriately identify team members, to evaluate their contributions to the team, and to adequately reward them for these contributions. While detailed lists of recommendations [22] are beyond the scope of this manuscript, selected budgetary policies are reflective of the philosophy: a.) crediting a percentage of indirect costs from all projects to support cross departmental infrastructure for IDR, b.) providing seed money, staff and space to support IDR, c.) creating a campus-wide inventory of equipment to enhance sharing of across facilities.
Conclusion
The "misalignments" described here derive from genuine uncertainties about the best approaches to meeting the research mission in AHC. They also emanate from a reward system in AHC (and more broadly in universities) which recognizes and rewards individual accomplishments more than group performance. These conflicts are not necessarily unhealthy, and can be described as balancing of competing priorities.
New initiatives outlined in the NIH Roadmap favor funding for programmatic, team-based research. Concurrently, AHC must configure the reward system to appropriately and accurately value contributions of team members. Doing so will minimize conflict of interest, and will allow AHC to more faithfully meet their mission.
List of Abbreviations
AHC, academic health center; NIH, national institutes of health; NPO, not-for-profit organization
Competing interests
The author(s) declare that they have no competing interests.
==== Refs
Kerr S An Academy Classic On the folly of rewarding A while hoping for B Academy of Management Executive 1995 9
Parks MR Disis ML Conflicts of interest in translational research J Transl Med 2004 2 28 15301694 10.1186/1479-5876-2-28
Oster SM Strategic Management for Nonprofit Organizations Theory and Cases 1995 , Oxford University Press
Hansmann HB The role of nonprofit enterprise The Yale Law Journal 1980 89 835 898
Unterman I Davis R Strategic Management of Not-For-Profit Organizations 1984 New York,
Salamon LM America's Non-Profit Sector: A Primer 1992 New York, The Foundation Center
Ott JS The Nature of the Nonprofit Sector 2000 , Westview Press
Goldstein JL Brown MS The clinical investigator: Bewitched, bothered, and bewildered-but still beloved. J Clin Invest 1997 99 2803 2812 9185499
Kotchen TA Lindquist T Malik K Ehrenfeld E NIH peer review of grant applications for clinical research Jama 2004 291 836 843 14970062 10.1001/jama.291.7.836
Williams GH Wara DW Carbone P Funding for patient-oriented research. Critical strain on a fundamental linchpin JAMA 1997 278 227 231 9218670 10.1001/jama.278.3.227
Sung NS Crowley WFJ Genel M Salber P Sandy L Sherwood LM Johnson SB Catanese V Tilson H Getz K Larson EL Scheinberg D Reece EA Slavkin H Dobs A Grebb J Martinez RA Korn A Rimoin D Central challenges facing the national clinical research enterprise Jama 2003 289 1278 1287 12633190 10.1001/jama.289.10.1278
Snyderman R The clinical researcher--an "emerging" species Jama 2004 291 882 883 14970069 10.1001/jama.291.7.882
Crowley WFJ Sherwood L Salber P Scheinberg D Slavkin H Tilson H Reece EA Catanese V Johnson SB Dobs A Genel M Korn A Reame N Bonow R Grebb J Rimoin D Clinical research in the United States at a crossroads: proposal for a novel public-private partnership to establish a national clinical research enterprise Jama 2004 291 1120 1126 14996782 10.1001/jama.291.9.1120
Stokes DE Pasteur's Quadrant: Basic Science and Technological Innovation 1997 Washington, D.C., The Brookings Institution
Gross CP Anderson GF Powe NR The relation between funding by the National Institutes of Health and the burden of disease N Engl J Med 1999 340 1881 1887 10369852 10.1056/NEJM199906173402406
Varmus H Evaluating the burden of disease and spending the research dollars of the National Institutes of Health N Engl J Med 1999 340 1914 1915 10369857 10.1056/NEJM199906173402411
IOM Scientific opportunities and public needs: improving the priority setting and publich input at the National Institutes of Health. 1998 Washington, D.C., Institute of Medicine; National Academy Press
Martin J Research in biomedicine. Is anyone representing/advocating the public interest? Eur J Public Health 2001 11 458 459 11766492 10.1093/eurpub/11.4.458
Brown LP Bair AH Meier PP Does federal funding for breastfeeding research target our national health objectives? Pediatrics 2003 111 e360 4 12671152 10.1542/peds.111.4.e360
Levinsky NG Nonfinancial conflicts of interest in research N Engl J Med 2002 347 759 761 12213950 10.1056/NEJMsb020853
von Schroeder HP The altruistic medical researcher: gone and forgotten? Ann R Coll Physicians Surg Can 1997 30 353 358 12378741
National Academy of Sciences NAEIM Facilitating Interdisciplinary Research 2005 Washington, D.C., National Academies Press
| 15862130 | PMC1090620 | CC BY | 2021-01-04 16:39:28 | no | J Transl Med. 2005 Apr 29; 3:19 | utf-8 | J Transl Med | 2,005 | 10.1186/1479-5876-3-19 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-271582319910.1186/1743-422X-2-27ResearchPhysicochemical characterization of vibriophage N5 Sen Anindito [email protected] Amar N [email protected] Division of Electron Microscopy, National Institute of Cholera and Enteric Diseaess, P-33, C.I.T. Road, Scheme- XM, Beleghata, Kolkata- 700010. India2 (Present Address)Laboratory of Structural Biology, Room 1504, Building 50, NIAMS/NIH Bethesda, MD, 20852, USA2005 11 4 2005 2 27 27 20 12 2004 11 4 2005 Copyright © 2005 Sen and Ghosh; licensee BioMed Central Ltd.2005Sen and Ghosh; 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.
Phage N5 is one of the phages of Vibrio cholerae serovar O1 biotype El Tor (Ghosh, A. N., Ansari, M. Q., and Dutta, G. C. Isolation and morphological characterization of El Tor cholera phages. J. Gen. Virol. 70: 2241–2243, 1989). In the present communication the growth curve, molecular weight and confirmation of the genome, partial denaturation map and restriction endonuclease digestion pattern have been determined. Partial denaturation map indicates that the genome has non-permuted / invariant sequence. Presence of cohesive ends has also been documented.
Vibriophage N5DNAVibrio choleraeElectron MicroscopyPartial denaturation mappingBacteriophage
==== Body
Vibrio cholerae, the causative agent of cholera in humansis classified into two serotypes: O1 and nonO1 [1]. The O1 strains are divided into two biotypes: Classical and El Tor. Before 1961 most epidemics had been caused by the classical biotype. However after 1961 the El Tor strains became the main causative agent of the cholera. Phage typing has proved to be useful and successful tool to tract down the spread of this dreadful disease. Vibriophage has also proved to be useful in studying the host chromosomes [2]. In the present work we report physicochemical characterization of an ElTor vibriophage N5 that was isolated from the sewage water samples of Calcutta, India [3].
The N5 phage was isolated from the sewage water of Calcutta [3] and was propagated on MAK 757, a Vibrio cholerae O1 ElTor strain. One-step growth curve of this phage was determined following the method described in Adams [4]. About 105 cells of freshly cultured MAK 757 were infected with N5 phage at an m.o.i. of 0.1. An aliquot was withdrawn at every 5 minutes and titrated for the total number of phages. A total of 8 × 108 plaque forming units are generated after 50–55 minutes from the time of infection. The eclipse period is nearly 8–10 minutes.
A phage lysate of N5 was prepared on soft agar (1% Nutrient Agar, pH 7.4, 0.5% NaCl, 1.5% Agar, HiMedia laboratories, Mumbai, India) overlay using freshly cultured MAK 757 (m.o.i of 0.01) as the propagating strain [4]. A few drops of chloroform were added to the freshly prepared phage lysate to remove bacterial content in it. The phage lysate (nearly 109phages/ml) was subjected to ultracentrifugation at 35,000 r.p.m. for 1 h and 30 mins in a Sorval T 865 rotor and a phage pellet was obtained. The phage pellet was resuspended in 1 ml of 50 mM – Tris-HCl pH 7.5, 20 mM – MgCl2 (TM buffer) to concentrate and the phage was stored at 4°C. The phage was purified on a sucrose step gradient of 10% to 40% as described previously [5,6] using a Sorval TW 668 swing-out rotor at revolution speed of 35,000 r.p.m. for 1 h and 15 minutes. The purified phage pellet l was re-suspended in 1 ml of TM buffer and stored at 4°C. The final concentration the resuspension was nearly 1011 phages/ml.
The N5 phage has an isomeric head with an extremely short non-contractile tail (figure 1a inset). The diameter (distances between the opposite apices) of the of the head is nearly equal to 71.5 ± 1.5 nm and the length of the tail is equal to 12.2 ± 1.9 nm. The tails are so short that in many of the phages, when observed under electron microscope, the tails are not visible because of the improper orientations on the specimen support film or breakage during phage preparation. N5 phage belongs to the 'podoviridae' family according to international committee for the taxonomy of viruses (1982).
Figure 1 Electron microscopic analysis of vibriophage N5 virion morphology and DNA structure. Panel A: Electron micrograph of vibriophage N5 stained with 2% uranyl acetate. Bar: 40 nm. Panel B: Electron micrograph of N5 DNA mounted on the grid by Kleinschmidt's technique. The arrows show the free ends of the DNA indicating that it is linear. Bar: 300 nm.
The high-titer purified phage lysate (1011/ml) was mixed with equal volume 0.0625 M Tris-HCl (pH 6.8) along with 1% sodium dodecyl sulphate (SDS) 15% glycerol, 1% Beta-mercaptoethanol and bromophenol blue. The solution was then incubated at 100°C for 3 minutes. SDS-polyacrylamide gel electrophoresis (PAGE) was performed by the method of Laemmli [7] as adopted by Sambrook et al., [8] for obtaining the SDS-PAGE pattern of the structural proteins of the N5 phage. The apparent molecular masses of the vibriophage N5 polypeptides were evaluated by SDS- PAGE 12.5% step-gel electrophoresis (figure 2a). From figure 2a we find four major bands of sizes 51 kDa, 37 kDa, 25 kDa and 18 kDa respectively. The major component had a molecular size of 51 kDa (approx). However, two small bands of 15 kDa and 12 kDa also visible in the step SDS-PAGE step-gel electrophoresis along with several other extremely faint bands that appeared in figure 2a which are probably bacterial debris.
Figure 2 Structural proteins of vibriophage N5 and restriction enzyme analysis of vibriophage N5 DNA. Panel A: SDS-PAGE patterns of structural proteins of purified N5 Vibriophage. Panel B: Restriction endonuclease digestion pattern of the N5 vibriophage DNA.
A comparison of these results with the SDS-PAGE of the N4 phage [6] reveals that the molecular masses of the N5 polypeptides are slightly higher than that of N4 phage. Nevertheless, the second band of 37 kDa in N5 (figure 2a) is very much of the same size as of the 36 kDa of the N4 phage. In fact the SDS-PAGE of the structural proteins of N5 vibriophage closely resembles the SDS-PAGE pattern of another vibrio El Tor phage e5 [9], which also has a 51-kDa polypeptide as a major component.
N5 vibriophage DNA was extracted from the phage using phenol-chloroform method described in Sambrook et al., [8] and dialyzed against 20 mM NaCl, 5 mM EDTA, (pH 7.4) buffer. The N5 DNA was spread using protein monolayer technique of Kleinschmidt et al. [10] with the modifications described in Inman [11]. About 500 ng/ml of N5 DNA was mixed up with 50 ng/ml of pBR322 marker DNA along with 0.067 M Na2CO3, 0.0107 M EDTA, 50% formamide (Sigma), and 0.01% cytochrome c (Sigma) at pH 7.4. The hypophase was double distilled water. The DNA-bound protein was picked up on carbon coated nickel grids, stained with uranyl acetate and was subjected to rotary shadow with platinum. Figure 1b shows a N5 DNA. The two ends (marked by arrows) are clearly visible thus indicating that the N5 DNA is linear. The length of the N5 phage DNA is computed to be 40.7 ± 0.7 kb as compared with pBR322 DNA of length 4.36 kb used as a marker (not seen in figure 1). This is very much similar to the size of the linear N4 vibriophage DNA that has a size of 40.4 ± 0.1 kb [6].
Partial denaturation of the N5 vibriophage DNA was carried out as described previously [4,8] A high pH buffer was prepared that contained 34% formaldehyde, 10 mM Na2CO3, 1 mM EDTA and suitable amount of NaOH to make up the pH to 10.9. About 7 μl of N5 DNA was gently mixed with 3 μl of the high pH buffer and was incubated at 37 ± 1°C for 15 minutes. The final solution was then mixed with formamide and cytochrome c to a final concentration of 50% and 0.01% respectively. Partial denatured vibriophage N5 DNA molecules were obtained (not shown in figure). The DNA molecules were arranged in a linear fashion according to their denaturation sites and a partial denaturation map was constructed (figure 3a). A weight average, denaturation histogram (figure 3b) of these maps was plotted to visualize the average denaturation pattern of the total N5 DNA molecules. It is quiet apparent from the histogram that there are at least 5 major denaturation sites. There is always a denaturation site at one end (by convention to the right-hand end, Inman, [11]) of the DNA molecule irrespective of the degree of denaturation. The other denaturation sites are at postions 26%, 65 %, 75 % and 91% from the left hand end respectively and a minor peak at the 55% position (figure 3b). The result shows that the vibriophage N5 DNA has a non-permuted and unique sequence. Comparison with the denaturation map of N4 [6] reveals that N5 had denaturation site at one end (right hand end) while N4 DNA has a denaturation sites at both ends. However the vibriophage D10 DNA has denaturation peaks at the same locations as that in N5 phage. In this respect denaturation map of N5 DNA is similar to that of D10 DNA [5].
Figure 3 Partial denaturation maps of vibriophage N5 DNA. Panel A: Vibriophage N5 DNA was subjected to partal denaturation. Each line represents one double stranded DNA molecule. The denaturation sites along each DNA molecule are shown by small solid rectangles. Panel B: Histogram average of partial denaturation maps of N5 DNA.
Since the N5 DNA is non-permuted it was expected that the DNA might have cohesive ends [12]. In order to test whether the N5 DNA has cohesive ends 3 μl of N5 DNA (500–800 μg/ml) was mixed with 2 mM Tris, 0.2 mM EDTA buffer (pH 8.4) along with 20 mM Tris, 2 mM EDTA buffer (pH 8.4), 50 mM Na2CO3 and 50% of formamide [13].
The mixture was left for incubation at room temperature for 72 hrs. After about 48 hrs 2 μl of Tris-EDTA buffer {0.1 M Tris + 10 mM EDTA (pH 8.4)} was added (to maintain the pH of the mixture) to the mixture and left for another 24 hrs of incubation at room temperature. After the completion of 72 hrs of incubation 3 μl of cytochrome c was added to a final concentration of 0.01% and was spread on double-distilled water. After examining about 10 DNA molecules it was found that molecules have length nearly twice the native length of the N5 DNA. The average length of these 20 molecules is 79 ± 0.8 kb while is twice the native length mentioned earlier i.e. 40.7 ± 0.7 kb. This confirms that N5 DNA has cohesive ends.
Restriction endonuclease digestion of the N5 vibriophage was carried out with the help of the procedure recommended by manufactures ("Genie", India). The enzymes used were: Eco RI, Sal I, Bam H1, Bgl II, Pst I, Bgl I, Ass I, Sma I, Hind III, Hpa II, Eco RV, Acc I, Hae III and Xba I. Restriction endonuclease digestion pattern of the N5 phage DNA revealed that the N5 DNA is double stranded. The N5 phage DNA was resistant to Eco RI, Sal I, Bam H1, Bgl II, Pst I, Bgl I, Ass I and Sma I. It is worth mentioning here that the DNA of vibrio El Tor typing phage 'e5' is also resistant to first five restriction endonucleases mentioned above [9]. However both e5 and N5 phage DNAs have restriction sites of Hpa II (figure 2b and [9]). It is also observed that Hind III gives rise to 6 kb, 2 kb, and 1.3 kb common fragments in N5 and N4 [6] but N5 DNA has an additional fragment of 21 kb which is absent in N4 DNA.
Acknowledgements
Authors are thankful to Dr. S. K. Bhattacharya, Director of the institute, for his interest and encouragement in this study.
==== Refs
Mukherjee S Principles and practice of typing Vibrio cholerae Methods Microbiol 1978 12 51 115
Guidolin A Manning PA Genetics of Vibrio cholerae and its bacteriophages Microbiol Rev 1987 51 285 298 3299030
Ghosh AN Ansari MQ Dutta GC Isolation and morphological characterization of El Tor cholera phages J Gen Virol 1989 70 2241 2243 2769238
Adams MH Bacteriophages 1959 Interscience publishers, Inc. New York
Chakrabarti BK Chattopadhyay DJ Ghosh AN Vibriophage D10 contains non-permutated DNA with cohesive ends J Gen Virol 1993 74 2749 2752 8277281
Ghosh AN Chakrabarti BK Chattopadhyay DJ Sil S Vibriophage N4 DNA is nonpermutated and terminally redundant Can J Microbiol 1995 41 842 845
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 1970 227 680 685 5432063
Sambrook J Fritsch EF Maniatis T Ford N, Nolan C, Ferguson M Bacteriophage λ growth, purification and DNA extraction.2.60–2.81, and SDS-polyacrylamide gel electrophoresis of proteins, 18.47–18.59 Molecular cloning: a laboratory manual 1989 2 Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York
Basu R Ghosh AN Ghosh A Biophysical characterization of Vibrio El Tor typing phage e5 FEMS Microbiol Lett 1993 106 9 16 8095039 10.1016/0378-1097(93)90049-8
Kleinschmidth AK Lang D Jacherts D Zhan RK Darstellung und Längenmessungen des gesamten Desoxyriobnuclein-Säure-Inhaltes von T2 bakteriophagen Biochim Biophys Acta 1962 61 857 869 14033431
Inman RB Griffith JD Partial denatuartion mapping of DNA determined by electron microscopy Electron microscopy in biology 1982 2 John Wiley and sons, New York 237 271
Tye BK Huberman JA Botstein D Non-random circular permutation of phage P22 DNA J Mol Biol 1974 85 501 528 4853363 10.1016/0022-2836(74)90312-X
Davis R Simon M Davidson N Electron microscope for hetroduplex methods for mapping regions of base sequence homology in nucleic acids Methods Enzymol 1971 21 413 428
| 15823199 | PMC1090621 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 Apr 11; 2:27 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-27 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-371584016410.1186/1743-422X-2-37ResearchTransmission of human hepatitis C virus from patients in secondary cells for long term culture Revie Dennis [email protected] Ravi S [email protected] David [email protected] Nickolas [email protected] Rafat [email protected] Cheryl [email protected] Richard [email protected] Ann S [email protected] John G [email protected] S Zaki [email protected] Department of Biology, California Lutheran University, Thousand Oaks, California, USA2 California Institute of Molecular Medicine, Ventura, California, USA3 Institute of Molecular Medicine & Technology, Huntington Hospital, Pasadena, California, USA4 Center for Women's Well Being, Camarillo, California, USA5 Community Memorial Hospital, Ventura, California, USA6 Ventura County Hematology-Oncology Specialists, Oxnard, California, USA7 Ventura County Medical Center, Ventura, California, USA8 University of Southern California, Los Angeles, California, USA2005 19 4 2005 2 37 37 6 4 2005 19 4 2005 Copyright © 2005 Revie et al; licensee BioMed Central Ltd.2005Revie 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.
Infection by human hepatitis C virus (HCV) is the principal cause of post-transfusion hepatitis and chronic liver diseases worldwide. A reliable in vitro culture system for the isolation and analysis of this virus is not currently available, and, as a consequence, HCV pathogenesis is poorly understood. We report here the first robust in vitro system for the isolation and propagation of HCV from infected donor blood. This system involves infecting freshly prepared macrophages with HCV and then transmission of macrophage-adapted virus into freshly immortalized B-cells from human fetal cord blood. Using this system, newly isolated HCV have been replicated in vitro in continuous cultures for over 130 weeks. These isolates were also transmitted by cell-free methods into different cell types, including B-cells, T-cells and neuronal precursor cells. These secondarily infected cells also produced in vitro transmissible infectious virus. Replication of HCV-RNA was validated by RT-PCR analysis and by in situ hybridization. Although nucleic acid sequencing of the HCV isolate reported here indicates that the isolate is probably of type 1a, other HCV types have also been isolated using this system. Western blot analysis shows the synthesis of major HCV structural proteins. We present here, for the first time, a method for productively growing HCV in vitro for prolonged periods of time. This method allows studies related to understanding the replication process, viral pathogenesis, and the development of anti-HCV drugs and vaccines.
==== Body
Introduction
The global public health impact of chronic HCV infection and consequent liver disease continues to grow in numbers. It has been estimated that there are over 170 million carriers of HCV worldwide, with an increasing incidence of new infections [1]. In the United States, an estimated 1% to 5% of the 2.7 million individuals that are currently chronically infected will die due to the HCV infection [2].
Although HCV has proven to be very difficult to grow in vitro, HCV-RNA has been detected in cell cultures of a variety of cell types, the presence of positive-strand HCV-RNA persists for periods ranging from a few days to several months, albeit with no evidence of infectious virus [3-6]. The recent creation of HCV-RNA replicons has contributed to a better understanding of some of the molecular events, particularly gene expression [7-9]. However, studies using parts of a virus can only give limited insights about the infectious process and pathogenesis of a specific genotype. For the development of effective rational therapies and the production of protective vaccines, a reproducible in vitro system for the isolation and replication of HCV from patients is critical.
We report here that the isolation and long-term replication of HCV in vitro. Since this is the first experience with actively replicating HCV in vitro, some of the results shown here may not fit the current concepts using systems that do not replicate infectious virus.
Materials and Methods
Infection of cultured cells with sera from HCV infected patients
HCV infected patient serum (minimum of 104 genome equivalents/ml) was filtered through 0.45 μ filters (Fisher Scientific) and frozen in 1 ml aliquots at -70°C. A fresh vial of frozen serum was used for every new transmission experiment. The cells were infected using 500 μl of thawed donor serum [10,11].
Generation of macrophages
Macrophages were generated from human cord blood mononuclear cells (CBMCs) by treating with Phorbol-12-myristate-13-acetate (PMA, 5 ng/ml in complete medium) [12]. A majority of the cells that adhered to the plastic were positive for non-specific esterase and phagocytosis, which are established markers for all macrophages. Multiple flasks (Falcon 3108 and 3109) were prepared in all cases to be used separately either for infection with HCV sera or for coculture with the infected patient's peripheral blood mononuclear cells (PBMC). The non-adherent cells contained approximately 60% CD19 and CD20 positive B-cells, with T-cells and monocytes accounting for the remainder. The cells that did not stain for macrophage-specific markers or phagocytosis were designated as non-committed lymphoid cells, and then infected either with HCV using 500 μl sera or cocultured with PBMC from the same patient.
Infection of macrophages with HCV
The macrophages were first treated overnight with polybrene (5 ng/ml) and then infected either with 500 μl of sera or cocultured with the PBMC from the same patient (Fig. 1A). These infected macrophages were incubated overnight at 37°C in a 5% CO2 atmosphere. Media were changed and the cultures were continued for another six days with change of media on day four.
Figure 1 Isolation of HCV from human patients. (A) Isolation scheme for the replication of HCV in vitro. (B) History of transmission of the specimen donated from HCV infected patient #081. Fresh macrophages were infected by using cell-free serum or cocultured with HCV infected PBMC from the blood of patient #081. Human T-cells (112 A), B-cells (112 B) or the non-committed lymphoid cells (112 AB) were then either infected by cell-free transmission of HCV from cell culture supernatant from macrophages or cocultured with HCV infected macrophages. Similarly freshly transformed cord blood B-cells (PCLB 1°) were infected by cell free transmission from previously infected B-cell (112 B) culture supernatant. Uninfected transformed B-cells (PCLB T1-T4) were infected by serial, cell-free transmission from filtered PCLB 1° culture supernatant. Neuronal precursor cells were infected by cell free transmission of HCV from filtered #081 culture supernatant.
Generation of immortalized B-cells
To create immortalized B-cells, cord blood mononuclear cells (CBMC) were stimulated with pokeweed mitogen (PWM, 5 μg/ml in complete culture medium), and then infected with transforming Epstein-Barr virus (EBV). These immortalized B-cells did not produce EBV [13,14].
Preparation of cell culture supernatants
Media taken from the cultures of infected macrophages were centrifuged at 500 × g for 10 minutes. The supernatants were then filtered through a 0.45 μ filter to remove extraneous material. The filtered supernatant is referred to as the cell culture supernatant.
Cell free transmission of HCV
The target cells were pretreated overnight with polybrene (5 ng/ml). A 500 μl aliquot of cell culture supernatant was used for infecting each of the target cells.
Design of positive- and negative-strand primers
In order to identify HCV-RNA, nested primers for each strand from the 5' untranslated region (UTR) were designed by CIMM using the default parameters of the DNASTAR PrimerSelect program (Table 1).
Table 1 Primers used to analyze HCV
Primer Strand Sequence (5' to 3')1
HCV 9.1 positive gac act cca cca tag atc act c
HCV 9.2 positive cat gat gca cgc tct acg aga c
HCV 10.1 positive ctg tga gga act act gtc ttc acg cag
HCV 10.2 positive cac tcg caa cca ccc tat cag
HCV 1 negative act gtc ttc acg cag aag cgt cta gcc at
HCV 2 negative cga gac ctc ccg ggg cac tcg caa gca ccc
HCV 3 negative acg cag aaa gcg tct agc cat ggc gtt agt
HCV 4 negative tcc cgg ggc act cgc aag cac cct atc agg
HutLA2 positive ggg ccg ggc atg aga cac gct gtg ata aat gtc
1The primers were designed with the program PrimerSelect (DNASTAR) using conserved HCV sequences downloaded from GenBank.
Detection of positive- and negative-strand HCV-RNA by nested RT-PCR assay
Total RNA was extracted from infected cell culture supernatants harvested 5 days after a change of media (Tri Reagent LS, Molecular Research Center Inc. Cincinnati, OH). A 269 base pair region was amplified by nested RT-PCR from the highly conserved 5'-UTR of the HCV genome.
The positive strand assay was performed using a 10 μl aliquot of the total extracted RNA was reverse transcribed using the primer HCV 9.2 with the MMLV Reverse Transcriptase (Promega Corp. Madison, WI) or with the Sensiscript Reverse Transcriptase (Qiagen Inc. Valencia, CA) according to the manufacturers' instructions. A 5 μl aliquot of the cDNA was then amplified by nested PCR using HCV 9.1 and HCV 9.2 as the outside primers, followed by amplification of 5 μl of the first PCR product using HCV 10.1 and HCV 10.2 as the inner primers.
The negative strand assay was performed by using the Oligotex Direct mRNA purification kit (Qiagen Inc.) to extract RNA from the cells. A 10 μl aliquot of the RNA was reverse transcribed using the HCV1 primer with the Thermoscript Reverse transcriptase (Invitrogen) according to manufacturer's instructions. Nested PCR amplification was then carried out on a 5 μl aliquot of the cDNA using HCV1 and HCV2 as the outer primers, followed by amplification of 5 μl of the first PCR product using the HCV3 and HCV4 as the nested primers under standard PCR conditions.
For each PCR, forty cycles of amplification were performed with the following temperature profiles: 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min for the outer primer set and 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min for the inner primer set.
Detection of positive-strand HCV-RNA by Real-time RT-PCR
The total extracted RNA was solubilized in 10 μl of RNase-free water and then reverse transcribed using the primer HCV 10.2 with the MMLV Reverse Transcriptase. A 5 μl aliquot of the cDNA was then amplified by real-time PCR, using HCV 10.1 and HCV 10.2 primers on the Rotor-Gene 200 amplification system (Corbett Research, Australia) and the SYBR Green I fluorescent dye (BioWhittaker Molecular Applications, Rockland, ME), using the manufacturers' instructions. An in vitro transcribed RNA from the HCV 5'-UTR was utilized as the standard. Forty cycles of amplification were performed with the following temperature profile: 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min.
Detection of HCV-RNA by in situ hybridization
Approximately 6 × 104 cells were centrifuged (Cytospin II, Shandon, Pittsburgh, PA) onto RNase-free Poly-L-lysine coated slides (Fisher Scientific, Pittsburgh, PA), forming a uniform well spread monolayer of cells. These cells were fixed and desiccated with ethanol. Cells were then rehydrated with 1× SSC buffer and treated for protein digestion with proteinase K (Fisher) for permeation and retention. Hybridization of the probes to the cells was performed overnight at 56°C. After overnight hybridization, to minimize the amount of unhybridized probes, cells were washed three times with formamide followed by one wash with RNAse A, and then one wash with RNAse-free buffer. Depending upon the batch of reagents, the slides were coated with liquid emulsion (K5 Liquid Emulsion, Ilford Imaging, UK) and exposed for 10–15 days. After exposure, the slides were developed with Kodak D19 developer (Eastman Kodak Company, Rochester, NY) and fixed using the Ilford Hypam Fixer (Ilford Imaging, UK). The developed slides were then stained with Wright-Gimsa Stain (EM Diagnostics Systems, Gibbstown NJ) and mounted with permount. The probes, used for in situ hybridizations, were prepared by cloning a DNA sequence corresponding to the 5' untranslated region (5'-UTR), nucleotides 55–308, of HCV RNA into pGEM-T Easy vector (Promega Corp. Madison, WI). S35-labeled probes, complementary to the positive- or negative-strand of HCV-RNA, were generated by in vitro transcription in the presence of a 35S rUTP (Amersham Biosciences, England) using the appropriate RNA polymerases as supplied by the manufacturer (Promega Corp. Madison, WI) and purified through Sephadex G50 [11].
Detection of HCV-RNA by fluorescence microscopy
An indirect immunofluorescence (IF) assay was used [11]. Cells were washed for 10 minutes three times with phosphate-buffered saline (PBS), resuspended in PBS, deposited on Teflon-coated slides, air-dried, and fixed in cold acetone for 10 minutes. Patients' heat-inactivated sera (56°C for 30 minutes and then clarified by centrifugation) was added to the fixed cells, and incubated at 37°C for 40 minutes. They were then washed with PBS, air-dried, and stained with FITC-conjugated anti-human IgG for 40 minutes. The cells were again washed, air-dried, counter-stained with Evans Blue for 5 minutes and mounted with IF mounting solution.
Kinetics of HCV production in vitro to determine the optimum day for harvesting positive-strand CIMM-HCV RNA
On day zero, a CIMM-HCV cell culture was taken out of liquid nitrogen, resuspended, separated into seven flasks of approximately 106 cells each, and fresh media was added to each flask. The initial concentration of the virus in the media therefore starts at zero viral particles. For each of the next seven days, one flask was harvested and assayed for the positive- and negative-strands of HCV-RNA using nested RT-PCR.
Genotyping of CIMM-HCV RNA
RNA from cell culture supernatants was amplified via nested RT-PCR using the positive-strand RT-PCR assay primer set as described before. Products of the RT-PCR were cloned into the PCR 4.1 cloning vector (Invitrogen Corp. Carlsbad, CA). Plasmid DNA was isolated from individual clones and sequenced on an ABI 377 automated DNA sequencer using a Dye Terminator Sequencing Kit (Applied Biosystems, Foster City, CA).
Purification of Immunoglobulin (IgG) from HCV infected patient sera
Serum from patient #081 was applied to an Affi-Gel II Protein A column (Bio-Rad Laboratories, Hercules, CA), and the IgG fraction was eluted. Purified IgG were concentrated by Microcon 50 columns (Millipore Corp., Billerica, MA) and stored at -20°C.
Extraction of viral proteins from cell culture supernatants
Total proteins were precipitated from 1 ml of cell culture supernatant or patient serum with the TRI REAGENT (Molecular Research Center, Inc. Cincinnati, OH). The ethanol washed protein pellet was solubilized into 200–500 μl of 1% SDS by incubating at 55°C for 10 minutes. Any remaining insoluble subcellular particles were removed by centrifugation at 14000 × g for 10 minutes at 4°C. Proteins were quantified using the Bradford Protein Assay (Sigma-Aldrich Corp. St. Louis, MO) and frozen (-20°C).
Dot-blot and Western blot analyses
For the dot-blot assay, 2 μl of various protein samples (undiluted to 10-3) were diluted to 25 μl using TBS and were dot blotted onto a nitrocellulose membrane (0.22 μ, Micron Separations Inc. Westboro, MA). For the Western analysis, proteins were separated by SDS-PAGE under non-reducing conditions and transferred to nitrocellulose membranes (Bio-Rad Labs). The membranes were blocked with 2% non-fat milk in 20 mM TBS, 500 mM NaCl, 0.02% Tween 20 for 1 hour. The samples were then incubated with purified IgG (1:1000 dilution) for 2–4 hours at room temperature. Antibody binding was detected by incubation with alkaline phosphatase-conjugated goat anti-human antibodies followed by color development (Bio-Rad) [15].
Accession numbers of HCV sequences used for genotyping
The 5' UTR sequence was obtained using the 10.1 and 10.2 primers (Table 1), and has the accession number DQ010313. The partial sequence of the NS5B region was obtained using the C-anti and the reverse complement of C1A primers, and has the accession number DQ010314.
Results
Nine years ago, we undertook to isolate infectious HCV from patients and to grow such isolates in vitro. Our initial experiments to develop an in vitro system of HCV replication were performed as previously reported by many investigators using a large variety of established cell lines comprising of various cell types [16]. These included human transformed liver cells in addition to Hela, CEM, H9, Jurkat, Molt 3, Molt 4, U937, P3HR1, Raji, Daudi, human foreskin fibroblast (ATCC, Bethesda, MD). All of these cell types could be infected by the reported methods, with the exception of human foreskin fibroblasts, which was uninfectable (Table 2). Results from these efforts did not prove to be reproducible for the sustained replication of HCV. Although we were able to detect positive and negative-strand (replicative) RNA for HCV in a few B-cells, liver cells, and monocytoid cells, none of these standard cell lines produced infectious HCV that could be transmitted into uninfected cells. These freshly infected cell cultures eventually became negative for HCV-RNA, while the uninfected cells grew. We now know from our experience that HCV behaves as a lytic virus, with up to 20% cell death in infected cultured cells. Infected B-cells form enlarged cells which eventually die without further replication (Fig. 2C). Cell line U937, despite its monocytic nature and the presence of detectable positive- and negative-strand HCV-RNA, had very low levels of viral RNA expression.
Table 2 Summary of HCV transmission experiments with various hematopoetic and liver cells
Short term Long term
A. T-cells1 + -
B. B-cells2 + +
C. Monocytes/macrophages3 + -
D. Neuronal precursors4 + +
E. Liver cells5
Kupffer's cells + -
Hepatocyte + +/-
1T-cells isolated from human fetal chord blood.
2B-cells immortalized by infection with transforming EBV.
3Monocyte/Macrophages, adherent cells stimulated with PMA.
4Recently isolated neuronal cells from fetal human brain.
5Freshly isolated liver cells from liver biopsies. Kupffer's cells are liver macrophages and Hepatocytes are liver endothelial cells.
Figure 2 In vitro propagation of HCV in cultured cells. Morphology of neuronal precursor cells infected with HCV. (A) T (telencephalon, suspension cells that grow in clumps and also can adhere to plastic), and (B) M (metencephalon, primarily adherent cells that develop neuronal processes). (C) Freshly transformed B-cells co-cultured with HCV infected macrophages. None of these cells have definitive cytopathic effects when compared with uninfected cells.
Because our initial experiments provided no significant improvement over the previously reported findings, we used a different approach for HCV isolation. We noted higher levels of HCV-RNA in infected macrophages compared to other infected cells. This was analogous to the infection of similar cells with human immunodeficiency virus (HIV-1) [17]. Therefore, we initiated the use of freshly isolated macrophages and other cells. We tested a variety of cell types from different origins for infectivity with HCV: endothelial cells from fresh fetal umbilical cord, mononuclear cells from fetal cord blood, CBMC, PBMC, and Kupffer's cells and hepatocytes from fresh liver biopsies. These freshly obtained cells were infectable and expressed both the positive- and the negative-strands of HCV-RNA.
Further experiments were designed that used macrophages as the intermediate host. The results from macrophage cultures were most encouraging. A number of researchers have previously used B-cells as a target [4,6]. Therefore, we decided to combine the macrophages with B-cells into one system. It also became apparent that in order to carry the transmitted virus for an extended period of time in vitro, long-lived B-cells were required. We opted in favor of freshly immortalized B-cells because they were free of various adventitious agents such as mycoplasma and other cellular contamination. Retransmissions were achieved by using the culture supernatants obtained from the macrophages and the B-cells prepared in our laboratories.
Transmission of HCV isolates
In order to show that our system could be used to grow HCV for extended periods, we tested each isolate at regular intervals by RT-PCR and retransmission into fresh cells (Table 3). Due to the large number of samples that were tested, HCV isolation and long term replication were carried out in several phases: short term cultures (positive for HCV up to 10 weeks), medium term cultures (positive for 10–23 weeks), or extended term cultures (positive for over 23 weeks). Experiments using either human patient sera or PBMC were equally able to infect macrophages that could be used in cell-free transmission of HCV. We did not compare the levels of virus produced by these two methods. An example of a long term positive cell culture is isolate #081. This isolate was obtained from similarly numbered serum from donor #081. Isolate #081 has been maintained in culture for over one hundred thirty weeks. This is designated as the index isolate: CIMM-HCV. This isolate has been propagated in different cell types such as enriched B-cells, T-cells, and non-committed lymphoid cells obtained from fresh blood by both co-culture and cell-free methods. Serial transmissions to freshly transformed B-cells were performed by cell-free methods for further analysis (Figure 1B). The first transfer of HCV from macrophages to target cells is designated as T1. A transfer from the T1 culture to fresh target cells is designated T2. Transfers of isolates have been carried out as many as four times (T4), such as isolate PCLBT4. Cell culture supernatants were harvested at least every month and assayed for positive-strand HCV-RNA by nested RT-PCR analysis (Table 3). Nested PCR has been used as a diagnostic method by many researchers [18-20], and was used in order to eliminate false positives. Due to the consistently positive nested PCR and sequential biological transmission assays over a period of many months, the isolated HCV was considered to be replicating and infectious virus.
Table 3 History of HCV positivity for CIMM-HCV isolates1
Sample
Month #081 112 A 112 B 112 AB PCLB T4
1 -
2 -
3 +
4 + - -
5 + + +
6 ND + +
7 ND + + -
8 ND + + + -
9 ND + + + +
10 ND ND ND ND -
11 ND - - ND -
12 ND + + ND +
13 ND + + ND +
14 ND + + ND +
15 ND + + - +
16 ND + + + +
17 ND + + + +
18 ND + + + +
19 ND + + + +
20 ND + + + +
21 ND + + + +
22 ND + + + +
23 ND + + + +
24 ND + + + +
25 - ND ND ND ND
26 + ND ND ND ND
27 + ND ND ND ND
28 + ND ND ND ND
29 + ND ND ND ND
30 + ND ND - -
31 + ND ND + +
1Each sample represents a monthly harvest of cell culture supernatants that were tested for the presence of human HCV positive-strand RNA and the cell free transmission to fresh target cells. Each individual sample was stored in liquid nitrogen at various time points throughout our testing. CIMM-HCV has been carried for over 12 months as a primary culture and over 31 months as a transmitted virus into other cell types including T-Cells (112A), B-Cells (112B), non-committed lymphoid cells (112AB) and 4th serial transmission into immortalized cord B-cells (PCLB T4). Primary cells are the first B-cells infected with HCV isolated from the macrophages.
Our results suggest that there is no significant difference between using patient sera or PBMC as a source of the infectious agent, but there were no attempts made to quantitate the levels of infectious virus in the primary samples (serum or cells). Since only one cell producing infective virus can be enough to achieve transmission, both methods can be used to successfully culture HCV.
Host range of HCV isolates
CIMM-HCV is maintained in one cell type: freshly transformed B-cells. In order to establish the host range of this isolate, a large number of cell types were tested for HCV propagation as described before. In addition to B-cells and macrophages, neuronal precursors could also be infected. These neuronal cells are very similar to macrophages, and they became a significant producer of infectious HCV (Table 4). Neuronal T cells grow in large non-adherent and adherent clumps and the M cells are generally adherent and form neuronal cell-like processes. They survived HCV infection better than B-cells in terms of cell viability (Figs. 2A and 2B). Cell-free CIMM-HCV was transmitted to our two neuronal cell types, T (telencephalon) and M (metencephalon), which subsequently showed replication of transmissible infectious virus (experiment 244). Virus from these cells was subsequently transmitted to fresh T and M neuronal cell cultures in experiment 248 and from 248 to 260 (Table 4). These retransmissions were similar to the ones performed for B-cells (Table 3). Infections of neuronal cells were repeated several times with similar results with respect to HCV production. We have since transmitted this HCV from experiments 260 to 273 and 273 to 277 (data not shown).
Table 4 Transmission of human HCV in neuronal precursor cells1
Month
Sample 1 2 3 4 5 6 7
#081 + + + + + + +
244 M - + + + + + +
244 T - + + + + + +
248 M - + + + + + +
248 T - + + + + + +
260 M + + + + + + +
260 T + + + + + + +
1The neuronal precursor cells were isolated from human fetal brain by Dr. Olag Kopyov (see acknowledgment). They were designated T (telencephalon, suspension cells) and M (metencephalon, adherent cells). Each sample represents a monthly harvest of cell culture supernatant that were tested for the presence of positive-strand HCV RNA.
Testing the HCV isolation system using additional patients
In order to take advantage of the system developed in our laboratories, we obtained 156 samples from patients who volunteered to donate their blood. Of these, 151 were peripheral blood specimens from HCV infected patients and 5 were from uninfected controls. All specimens were acquired with the approval of the Institutional Review Board (IRB) and donors' informed consent. The HCV-infected specimens were obtained from 109 Caucasians, 37 Hispanics and 5 African Americans. The uninfected controls were from 2 Caucasians, 2 Hispanics, and 1 African-American. The participants included 108 males and 48 females. All specimens were freshly processed within an hour of blood drawing. Repeat samples were obtained from 77 of the original patients in order to confirm our initial results. Thirty-three of these 151 HCV-infected patients were co-infected with HIV-1, and the remainder of the donors had hematological malignancies or other cancers. HCV was isolated with 75% efficiency from these 151 specimens. In the case of co-infected patients, the failure to isolate HCV was commonly due to rapid cell death. No HCV was ever isolated from the 5 uninfected controls. This high rate of isolation of HCV shows that this system is useful in obtaining HCV from a variety of individual patients for further analysis.
Determination of optimum day for harvesting HCV for RNA extraction
In order to determine the optimum day for harvesting the highest accumulation of positive-strand RNA, the kinetics of HCV production was measured using nested PCR. For each of the next seven days, flasks were harvested and assayed for the positive- and negative-strands of HCV-RNA using nested PCR. An example of our results is shown in Fig. 3A. While day 5 showed the greatest accumulation of positive-strand of HCV-RNA, the levels of the negative-strand inside the cells on all seven days remained unchanged (Fig 3A). There was no significant increase in cell numbers during the experiment.
Figure 3 Detection of positive- and negative-strand HCV-RNA in infected cell cultures via RT-PCR. Posititve strands were assayed using the cell culture supernatant while the negative strands were assayed using total RNA purified from the cells. (A) Determining the optimum day to harvest HCV for RNA extraction and analysis. Approximately one million cells of culture #081 were divided into seven flasks and incubated. One flask was harvested on each of the following days and assayed for positive and negative strand RNA. (B) Quantitation of molecules of positive-strand HCV-RNA per ml of cell culture supernatant via real-time RT-PCR. (C) Positive- and negative-strand HCV-RNA in different cells infected with CIMM-HCV. Lane 1 CIMM-HCV, lane 2 T-cells (112 A), lane 3 B-cells (112 B), lane 4 non-committed lymphoid cells (112 AB), lane 5 the 4th serial transmission into immortalized cord B-cells (PCLB T4), lane 6 T-cells (200 A), lane 7 B-cells (200 B), lane 8 uninfected B-cells, lane 9 HCV infected patient serum, and lane 10 negative PCR control.
In an experiment performed simultaneously, the positive-strand HCV-RNA in the cell culture supernatants was analyzed quantitatively by real-time RT-PCR. As expected, on day zero there was no measurable HCV-RNA. On day one, the measurable number of copies of HCV-RNA was 3,200, which increased during the experiment to approximately 27,000 copies per ml on day 5 and then decreased from thereon (Fig. 3B). This data was consistent with the pattern obtained using the nested RT-PCR assay shown in Fig. 3A. Note that the data for Figures 3A and 3B are from using the same samples. The optimum day of harvesting this isolate of HCV was on day 5. Other isolates have produced similar growth curves (data not shown).
Seven isolates were tested by nested RT-PCR to show that the results from Figure 3A were reproducible. The presence of the expected PCR products demonstrated that on day 5, both positive- and negative-strands of HCV-RNA were present in our system (Figure 3C). This experiment shows both replication and extracellular production of the virus. This indicates that harvesting RNA on day 5 will permit reproducible results.
Detection of HCV-RNA by in situ hybridization
We analyzed our HCV infected cells by performing in situ hybridizations to visualize the percentage of infected cells and the locations of the HCV-specific strands [21]. The uninfected cells used as a control did not hybridize to either negative or positive strand probes (Figs. 4C and 4D). In all cases, the numbers of background grains were light. Hybridization with the probe for the positive-strand produced a halo-like appearance around the periphery of the infected cells (Fig. 4E). A strong signal for the negative strands of HCV-RNA was seen confined within the cells, possibly in the cytoplasm (Figs. 4A and 4B). Fluorescence microscopy of infected cell cultures showed a similar result (Fig. 4F). Although approximately 5% of the cells appeared strongly positive, this may have been an underestimate due to: (1) cell lysis of infected cells in culture; and (2) the loss of cells that attach to the filter cards used in preparing the cytospin slides. Hybridization to both the positive- and negative-strands of HCV-RNA suggests replication and production of HCV. Since most of the cells do not appear positive, the positivitity that was observed is not just a result of non-specific staining of cells. Results of the in situ hybridizations are consistent with the nested RT-PCR assay described above. A majority of the infected cells appear to be large; however, there were a significant number of smaller cells that also gave positive signals above background. By comparison, neither the enlarged cells nor the small ones in the control population showed any positive signal (Fig. 4C and 4D). We believe that the small, infected cells produce virus and probably progressively enlarge and die, as trypan blue dye exclusion tests showed that these cells eventually died. Similar phenomena are observed in human immunodeficiency virus (HIV) and HHV-6 infected cell cultures [22].
Figure 4 Detection of HCV RNA in cultured cells by in situ hybridization with S35-labeled RNA probes and detection of HCV protein by fluorescence microscopy. (A, B) Infected B-cells hybridized with labeled positive-strand RNA probe. (C) Freshly transformed uninfected B-cells showing no significant hybridization. (D) Picture of freshly transformed uninfected B-cells showing no significant hybridization and having a wider field of view than (C). (E) Infected B-cells hybridized with labeled negative-strand RNA probe. (F) HCV infected cells: PCLBT4 treated with human polyclonal IgG purified from the serum of patient 081 and stained with goat anti-human IgG conjugated with fluorescein isothiocyanate (FITC). PCLBT4 is the fourth consecutive transfer of HCV to freshly immortalized B-cells.
Genotyping of the CIMM-HCV isolate
Based on sequence analysis, HCV has been classified into six major genotypes and a series of subtypes [23]. The highly conserved 5' untranslated region (5'-UTR), routinely used for RT-PCR detection of HCV-RNA, exhibits considerable genetic heterogeneity [24] and shows polymorphism between types and subtypes. This genetic heterogeneity of the 5'-UTR has been utilized for the genotyping of HCV [19,25-29], therefore, the 5'-UTR of CIMM-HCV was cloned and sequenced. Based on sequence homology searches, CIMM-HCV was similar to genotype 1a.
In order to spot check the genome of CIMM-HCV, we tested most of the previously published primers [30-33]. We, however, found that many of these primers did not lead to RT-PCR products from our isolate, including CD 2.10 [31], CD 5.10 [31], CD 5.20 [31], A5310 [33], and A6306 [33]. This may be due to the heterogeneity of HCV RNA [18]. It is also possible that parts of our isolate may differ significantly from the previously reported sequences. We have included here the sequence from part of the NS5B gene of CIMM-HCV, which is located near the 3' end of the genome. This sequence is most similar to HCV of genotype 1a/2a.
Although the culture system described here is capable of isolating HCV from approximately 75% of infected patients, this process may select more competent and infectious virus. Our analysis of sequences from the 5' UTR region shows in one case that the blood of a patient and the isolate in culture are both of type 1b. There were no significant differences in the sequences of the patient and the isolate in this region (Revie, Alberti, and Salahuddin, manuscript in preparation).
Reactivity of the polyclonal IgG purified from infected patient sera
To determine the reactivity of the purified polyclonal IgG, various dilutions of the total protein preparations from cell culture supernatants were analyzed. A positive reaction was noted with homologous serum proteins using CIMM-HCV obtained from the B-cell supernatant, supernatants from neuronal cells (from transmission experiment 260), and commercially available HCV core antigen (ViroGen Corp. Watertown, MA) (Fig. 5A). There was no reaction with NS4 as well as the uninfected cell culture supernatants. These results show that IgG purified from patient's sera specifically detects HCV virion proteins, particularly Core antigen, and that the virus grown in culture reacts with antibodies from patient's sera.
Figure 5 Dot blot analysis of HCV proteins binding to IgG from patient sera. HCV proteins were obtained from tissue culture fluid. Protein preparations were serially diluted (1, 10-1, 10-2, 10-3) and were dot blotted onto a nitrocellulose membrane. These blots were then treated with patient antibodies. The figure shows (A) Reactions of patient IgG against dilutions of IgG depleted patient sera, CIMM-HIV cell culture supernatants from different cell lines (HCV infected B-cells, HCV infected human neuronal precursor T and M cells), or commercial antigens (NS4 and Core antigen), or uninfected B-cells. (B) Reactions of patient IgG against dilutions of various HCV isolates grown in vitro as described before. All HCV isolates were from the first transfer to fresh B cells (T1) except for isolates 314T2 (second transfer) and PCLBT4 (fourth transfer). These infected cells have been in culture for varying periods of time, including over three years for PCLBT4.
Six different independent HCV isolates (081T1, 112T1, 238T1, 313T1, 314T2, and PCLBT4) were tested against polyclonal antibodies from patient 238 using a dot blot (Figure 5B). This was performed in order to determine if these isolates reacted similarly to the previous experiment shown in Figure 5A. The patient antibodies reacted with all of these isolates, as well as to commercial Core antigen and NS4. The amount of undiluted NS4 used here was 2 μg. This shows that all of these HCV isolates are producing HCV proteins, and that even a fourth transfer (T4) of one isolate into freshly transformed B-cells still produces reactive HCV proteins (PCLBT4). Each of these isolates has been passaged in culture many times.
Analysis of HCV proteins
The HCV genome encodes a polyprotein which is subsequently processed into a number of mature structural and nonstructural moieties [34]. In order to determine whether the replicating CIMM-HCV was producing major HCV proteins, Western blot analyses using non-reducing conditions were performed. The polyclonal IgG detected a series of proteins (i) in the HCV positive patient sera and (ii) in the infected cell culture supernatant (Figs. 6A and 6B). Proteins of 140, 75, 50, 37, 32, 27 and 25 kDa were detected in these samples. The polyclonal IgG also gave a positive reaction with the commercially obtained recombinant core antigen (lane 5, Fig 6A). This core antigen has β-galactosidase fused at the N-terminus and is thus approximately 140 kDa in size, as reported by the manufacturer.
Figure 6 Western blot analysis of HCV proteins under non-reducing conditions. Proteins were isolated from cell culture supernatants. (A) Analysis of large molecular weight proteins from the cell culture supernatants of CIMM-HCV in various cell lines. Lane 1 uninfected transformed B-cells, lane 2 transformed B-cells infected with HCV, lane 3 neuronal precursor cells derived from the metencephalon infected with HCV, lane 4 total protein from HCV positive sera and lane 5 commercially engineered HCV core antigen (ViroGen). Other smaller bands in the Core antigen may be due to breakdown products or other contaminating proteins. (B) Analysis of low molecular weight proteins of CIMM-HCV cultured in various cell lines. Lane 1 total protein from HCV positive sera, lane 2 transformed B-cells infected with HCV, lane 3 neuronal precursor cells derived from the metencephalon infected with HCV, and lane 4 uninfected transformed B-cells.
There are two highly glycosylated envelope proteins, E1 (32 and 35 kDa) and E2 (70 kDa) [35-39]. A band at approximately ~ 140 kDa was seen in all of the HCV isolates (Fig. 6A). This band has been seen by other researchers [40,41], and may have resulted from the multimerization of core, E1 and E2, or homodimerization of E2. The E1 and E2 proteins are known to form non-covalently linked heterodimers under non-reducing conditions [36,40,42]. The Core and E1 proteins also bind to each other [43,44], and. possibly form HCV protein and host cellular protein complexes as well.
Proteins in the 32 and 37 kDa range were also detected in the Western blots (Fig. 6B). These bands are consistent with the known sizes of E1. An approximately 75 kDa protein was also detected in all of the infected samples that were analyzed (Fig. 6A). This protein corresponds to the known molecular weight of E2.
In all the HCV isolates that were tested, a major protein band of approximately 50 kDa was seen (Fig. 6B). This was perhaps due to the incomplete processing of the precursor polyprotein [45]. Since the band was present only when protein purified from infected cell culture supernatants were used for the analyses, the band is therefore related to HCV proteins.
Bands of approximately 25 and 27 kDa were also detected on the Western blots (Fig. 6B). The host cell signal peptidases cleave the N-terminal region of the precursor polypeptide to produce the HCV core protein [37,46,47]. The HCV core protein is reported to range between 16 and 25 kDa in size. However, it is possible that the size differences that have been previously reported may be due to differences in processing of the HCV core protein [37,45,48,49]. We believe that the Core antigen, as expressed by our isolates of HCV, may be larger than has been previously described.
CIMM-HCV grown in three different cell lines produced the same pattern of bands. Although we were using polyclonal IgG extracted from patient sera, we detected HCV proteins that have been reported by other researchers.
Discussion
Our cell culture system comprises freshly prepared macrophages and immortalized B-cells for the isolation of HCV. The isolations are performed using both co-cultures and cell-free methods. Freshly prepared macrophages in our system are essential as an intermediate host for HCV isolation. We have not explored the mechanism of infection or replication of HCV, but macrophages from a variety of sources appear to have a positive role in this process. It is possible that macrophages either concentrate the HCV particles or modify the HCV sufficiently to enable them to infect other cell types. For example, changes in the glycosylated envelope proteins E1 and E2 [36,38,39] could affect the infectious capability of the progeny virus, as well as define its host range. In bacteria, infecting phage can have their DNA modified by the host modification system. This allows the phage to escape the restriction system, thus enabling better infection of new hosts. It is therefore not too difficult to imagine that various types of animal cells could produce modified versions of infecting viruses. These viruses may preferentially infect different cell types. It is also possible that the macrophages reduce or eliminate defective HCV that are found in the patient's blood that may interfere with starting a productive culture system. The macrophages would thus be acting as a gatekeeper, only allowing intact HCV to be cultured and disallowing defective ones that could interfere with the culturing.
As stated before, we discovered that neuronal precursor cells can be infected with CIMM-HCV and, in turn, are significant producers of infectious virus. We have repeatedly transmitted several HCV isolates into these neuronal cells. These cells are similar to other macrophages both in staining characteristics and in functional assays. They are growth factor dependent and grow well in vitro, and have been in culture for over two years. Macrophages from other sources, e.g. Kupffer's cells from liver, get infected, but after a few weeks gradually lose virus production. HCV-RNA, however, can be detected for several weeks. This loss of virus production may be related to the maturation, cytostasis, and eventual death of Kupffer's cells. Similar experiments were performed with freshly cultured endothelial cells obtained from human umbilical cord, because these are related to hepatocytes from liver, which are also endothelial cells. The results from these cells also showed a pattern of virus production similar to Kupffer's cells.
For almost all human viruses, there is a well-observed phenomenon of asynchrony of expression, with a few notable exceptions such as HIV. For human Herpes viruses, no matter which host cell they use, only a minority of cells produce replicating viruses [50,51]. We do not know why less than 5% of the cells in our system are productive. Human immunodeficiency viruses such as HIV-1 show active replication in less than 20% of infected T-cells if freshly isolated leukocytes are used [52]. The system we have described uses all freshly isolated cells. It is very likely that a specific receptor may be the limiting factor in determining the number of infectable cells.
We determined that the amount of HCV-RNA in the cell culture supernatant rises until day 5, then gradually declines. Therefore, the fifth day after subculturing is optimum for harvesting of HCV-RNA. As shown in Figure 3C, harvesting RNA on day 5 allows reproducible detection of HCV-RNA. Changes in the overall levels of HCV-RNA may reflect the sum of the RNA production and RNA destruction in culture. This observed periodicity in the positive-strand therefore, may be due to: (1) slowing of the replication process of the infected cells or from production of an inhibitor in situ; or (2) lysis of infected cells causing destruction of the virus and its RNA, e.g. by released proteases and ribonucleases, or (3) both of the above mentioned possibilities. The measured level of the negative strand in the cells remained stable. Since the number of cells and the percentage of cells that are infected don't change during the culturing of the virus, the observed stability of the negative-strand inside the cells is understandable.
Although the majority of patient samples described above are of type 1b, our analysis of CIMM-HCV sequences show it is type 1a. In the 5'UTR region there is very little sequence difference between types 1a and 1b. Changing of one base can cause the sequence to more closely resemble type 1b. It is possible that only a small subset of HCV in a patient is actually infectious, and therefore our system best typed it as type 1a. A complete sequence of the CIMM-HCV genome will reveal its identity. It is possible that a chronic infection in patients may develop mutants that resemble another type. However, we have additional isolates that are consistent with other genotypes such as 1b (data not shown). The sequences that we have reported here indicate that both ends of the viral genome are present in our cultures. We are currently in the process of sequencing the entire CIMM-HCV genome in order to better determine its genotype as well as to compare it to the currently published HCV genomes.
Since this is the first in vitro system for culturing HCV, we have been able to make initial observations regarding replication of the virus. Further studies related to HCV replication and pathogenesis are in progress. The in situ hybridization results seen in Figure 4 suggest that the positive strand of HCV is synthesized at or migrates to the plasma membrane and that the negative strand remains in the cytoplasm. This observation can only be made in a dynamic system with actively replicating virus. This suggestion is supported by recent reports that RNA-dependent RNA polymerase contains a transmembrane segment which is anchored in the membrane [53]. Non-structural proteins and positive strand RNA have also been found associated with the plasma membranes [54]. These results suggest that the site where HCV is fully assembled is probably in or near the plasma membrane of the infected cells. Probably HCV-RNA is synthesized in the cytoplasm and migrates to the plasma membrane for the final assembly. The completed virion is then released into the extracellular space.
We believe that a chronically infected patient makes antibodies to their own virus. Hence, the polyclonal IgG is specific for that virus. This is confirmed by our dot blot analysis of several different HCV isolates. Since these isolates had undergone as many as four separate transfers into fresh cells and many serial passages, the system that we describe here maintains stable isolates in culture, producing HCV-specific protein.
Similarly, data from Western blot analysis of the HCV proteins in the cell culture supernatants show that all of the expected major structural proteins are present. No specific antibody bindings were seen in the samples from uninfected cell culture supernatants. Taken together, these results suggest that there is production of HCV specific proteins in CIMM-HCV infected cell cultures. This data shows that the virus that was grown in culture contains epitopes of all of the major structural proteins that react with antibodies purified from the patient. This is strong evidence that the virus grown in culture is not significantly different from the HCV growing in the patient. In addition, one isolate that was cultured in several different cells lines showed a consistent pattern of bands on Western blots. This is additional and strong evidence that the bands result from HCV proteins. We have concluded that the replicating HCV is a stable virus.
In addition to the molecular analysis which establishes that our cells are producing HCV virions, the serial transmission of HCV to fresh uninfected cells via cell-free culture supernatants establishes biological evidence of infectious virus (Figure 1B: PCLBT1-PCLBT4). Since this virus is infectious in vitro and all of the major proteins appear to be present, the virus that has been grown in culture most likely contains the entire genome. We do not believe that there is a large amount of defective HCV present in our system. Although our standard assay is to amplify the 5' UTR region, we have been able to obtain sequences from other regions of the genome. This, coupled with the evidence for the presence of the major HCV structural proteins, is strong evidence that the entire genome is present. In addition, we have been able to use the HutLA2 primer (Table 1), which is complementary to a region near the 3' end of the positive strand, to produce cDNA. Using our standard primers from the 5' UTR region to amplify the cDNA, we were able to detect HCV positive-strand RNA (data not shown). This suggests that a significant proportion of the RNA present contains both the 3' and 5' ends of the genome. In addition, work in progress suggests that a representative virus population grows in this system, suggesting that a particular genotype is not selected (Revie, Alberti, and Salahuddin, manuscript in preparation).
Our system has allowed us to reproducibly isolate HCV from a majority of patients and in a few cases these cell cultures have been carried for over 130 weeks. Our ability to detect HCV for these extended periods shows that we are not just detecting virus that has been diluted from the initial sample. For the initial infection, 50 μl of serum was added to 2 ml of media, and then the media was changed once a week for months. A conservative estimate of the dilution of the virus at 130 weeks would be over 10130-fold. The amount of HCV produced from this system was sufficient to conduct biological, molecular, and immunological investigations. We are continuing to pursue other investigations in this field.
The analogy between macrophage-initiated in vitro propagation of HIV and HCV is rather remarkable. The dendritic cell-specific ICAM-grabbing non-integrins (DC-SIGN) can bind HIV, and protect it for protracted periods to concentrate and deliver the virions to cause infection of T-cells in trans with high efficiency [55,56]. The structural basis for selective recognition of oligosaccharides on virion envelope proteins by DC-SIGN and DC-SIGR may indeed be a common pattern by which HIV and HCV are concentrated for in vitro transmission to their respective susceptible cells [57,58]. Unlike CD4 for HIV, the HCV receptor CD81 is currently a subject of serious discussions [58-61].
We would like to point out that our system allows us to produce significant quantities of HCV, which made these studies possible. Having a reliable and long-term growth system for HCV in cell culture will facilitate in vitro studies and also aid in the production of rational drugs and vaccines. This culture system will, therefore, allow researchers in the field of HCV and liver disease to perform a wide variety of further analyses that can help in understanding the life cycle of HCV and the mechanisms of pathologies induced in human hosts.
Declaration of Competing Interests
All intellectual rights are reserved by the California Institute of Molecular Medicine (CIMM), and all aspects of this work were performed by CIMM. There are no competing interests between California Lutheran University or any other body and CIMM.
Authors' contributions
SZS and RK performed the biological work and the isolations, transmissions, and retransmissions of HCV. JGP, ASK, CG and RR performed the clinical work, recruitment of patients, and procurement of specimens. DR, RSB, DB, and NC performed the molecular work.
Acknowledgements
The California Institute of Molecular Medicine would like to thank to Drs. Richard Green and Terry L. Cole of Community Memorial Hospital, Ventura, CA, Drs. Rosemary McIntyre and Parsa of Hematology & Oncology Specialists of Oxnard, Ventura, CA, Drs. Kip Lyche and Ronald Chochinov of Ventura County Medical Center, Ventura, CA, Binom Galloway, Phlebotomist, Immunology Clinic, Ventura County Medical Center, Ventura, CA, the nursing staff of Labor and Delivery at Community Memorial Hospital, Ventura, CA and Hematology & Oncology Specialists, Ventura, CA for their continued efforts and support in advancement of HCV research. The authors also thank S. Ramakrishnan for his efforts and Dr. Joan Wines of California Lutheran University for help in preparation of this manuscript. Authors greatly appreciate the gift of transforming Epstein-Barr virus (B-95/8) from the late Dr. Meihan Nonoyama of Showa-Teikyo International Research Center, Florida and the kind donation of neuronal precursor cells for our research by Dr. Olag Kopyov of the California Neuroscience Institute, St. John Regional Medical Center, Oxnard, CA. Authors are also very grateful to Dr. Len Adelman of University of Southern California for allowing us to use the Roto-Gene 2000 amplification system. Part of this study was supported by NSF grant #0116789 to Dr. Dennis Revie.
==== Refs
World Health Organization Hepatitis C: Global prevalence Weekly Epidemiol Rec 1997 72 341 344
Salomon JA Weinstein MC Hammitt JK Goldie SJ Cost-effectiveness of Treatment for Chronic Hepatitis C Infection in an Evolving Patient Population JAMA 2003 290 228 237 12851278
Iacovacci S Sargiacomo M Parolini I Ponzetto A Peschle C Carloni G Replication and multiplication of hepatitis C virus genome in human foetal liver cells Res Virol 1993 144 275 279 8210708
Morsica G Tambussi G Sitia G Novati R Lazzarin A Lopalco L Mukenge S Replication of hepatitis C virus in B lymphocytes (CD19+) Blood 1999 94 1138 1139 10454799
Shimizu YK Iwamoto A Hijikata M Purcell RH Yoshikura H Evidence for in vitro replication of hepatitis C virus genome in a human T-cell line Proc Natl Acad Sci USA 1992 89 5477 5481 1319062
Sung V Shimodaira S Doughty A Picchio G Can H Yen T Lindsay K Levine A Lai M Establishment of B-Cell lymphoma cell Lines persistently infected with hepatitis C virus in vivo and in vitro : the apoptotic effects of virus infection J Virol 2003 77 2134 2146 12525648
Blight KJ Kolykhalov AA Rice CM Efficient initiation of HCV RNA replication in cell culture Science 2000 290 1972 1974 11110665
Ikeda M Yi M Li M Lemon SM Selectable subgenomic and genome-length dicistronic RNAs derived from an infectious molecular clone of the HCV-N strain of hepatitis C virus replicate efficiently in cultured Huh7 cells J Virol 2002 76 2997 3006 11861865
Lohmann V Korner F Koch J Herian U Theilmann L Bartenschlager R Replication of subgenomic hepatitis C virus RNAs in a hepatoma cell line Science 1999 285 110 113 10390360
Ablashi DV Lusso P Hung CL Salahuddin SZ Josephs SF Liana T Kramarsky B Biberfeld P Markham PD Gallo RC Utilization of human hematopoietic cell lines for the propagation and characterization of HBLV (human herpesvirus-6) Int J Cancer 1988 42 787 791 3053468
Salahuddin SZ Ablashi DV Markham PD Josephs SF Sturzenegger S Kaplan M Halligan G Biberfeld P Wong-Staal F Kramarsky B Gallo RC Isolation of a new virus, HBLV, in patients with lymphoproliferative disorders Science 1986 234 596 601 2876520
Salahuddin SZ Nakamura S Biberfeld P Kaplan MH Markham PD Larsson L Gallo RC Angiogenic properties of Kaposi's sarcoma-derived cells after long-term culture in vitro Science 1988 242 430 433 2459779
Fingeroth JD Weiss JJ Tedder TF Strominger JL Biro PA Fearon DT Epstein-Barr virus receptor of human B lymphocytes is the C3d receptor CR2 Proc Natl Acad Sci USA 1984 81 4510 4514 6087328
Lusso P Ablashi DV Luka J Ablashi DV, Krueger GRF, Salahuddin SZ Interaction between HHV-6 and other viruses Human Herpesvirus-6 Epidemiology, Molecular Biology and Clinical Pathology 1992 4 Amsterdam, The Netherlands: Elsevier 121 133
Josephs SF Salahuddin SZ Ablashi DV Schachler F Wong-Staal F Gallo RC Genomic analysis of the human B lymphotropic virus Science 1986 234 601 603 3020691
Kato N Shimotohno K Systems to culture hepatitis C virus Curr Top Microbiol Immunol 2000 242 261 278 10592664
Moriuchi H Moriuchi M Combadiere C Murphy PM Fauci AS CD8+ T-cell-derived soluble factor(s), but not beta-chemokines RANTES, MIP-1 alpha, and MIP-1 beta, suppress HIV-1 replication in monocyte/macrophages Proc Natl Acad Sci USA 1996 93 15341 15345 8986813
Bukh J Purcell RH Miller RH Importance of primer selection for the detection of hepatitis C virus RNA with the polymerase chain reaction assay Proc Natl Acad Sci USA 1992 89 187 191 1309604
Krekulova L Rehak V Wakil AE Harris E Riley LW Nested restriction site-specific PCR to detect and type hepatitis C virus (HCV): a rapid method to distinguish HCV subtype 1b from other genotypes J Clin Microbiol 2001 39 1774 1780 11325989
Tanaka E Ohue C Aoyagi K Yamaguchi K Yagi S Kiyosawa K Alter HJ Evaluation of a new enzyme immunoassay for hepatitis C virus (HCV) core antigen with clinical sensitivity approximating that of genomic amplification of HCV RNA Hepatology 2000 32 388 93 10915747
Moldvay J Deny P Pol S Brechot C Lamas E Detection of hepatitis C virus RNA in peripheral blood mononuclear cells of infected patients by in situ hybridization Blood 1994 83 269 273 8274741
Lusso P Ablashi DV, Krueger GRF, Salahuddin SZ Target cells for infection Human Herpesvirus-6 Epidemiology, Molecular Biology and Clinical Pathology 1992 4 Amsterdam, The Netherlands: Elsevier 25 36
Simmonds P Holmes EC Cha TA Chan SW McOmish F Irvine B Beall E Yap PL Kolberg J Urdea MS Classification of hepatitis C virus into six major genotypes and a series of subtypes by phylogenetic analysis of the NS-5 region J Gen Virol 1993 74 2391 2399 8245854
Bukh J Purcell RH Miller RH Sequence analysis of the 5' noncoding region of hepatitis C virus Proc Natl Acad Sci USA 1992 89 4942 4946 1317578
Chan SW McOmish F Holmes EC Dow B Peutherer JF Follett E Yap PL Simmonds P Analysis of a new hepatitis C virus type and its phylogenetic relationship to existing variants J Gen Virol 1992 73 1131 1141 1316939
Davidson F Simmonds P Ferguson JC Jarvis LM Dow BC Follett EA Seed CR Krusius T Lin C Medgyesi GA Survey of major genotypes and subtypes of hepatitis C virus using RFLP of sequences amplified from the 5' non-coding region J Gen Virol 1995 76 1197 1204 7730804
O'Brien CB Henzel BS Wolfe L Gutekunst K Moonka D cDNA sequencing of the 5' noncoding region (5' NCR) to determine hepatitis C genotypes in patients with chronic hepatitis C Dig Dis Sci 1997 42 1087 1093 9149068
Stuyver L Rossau R Wyseur A Duhamel M Vanderborght B Van Heuverswyn H Maertens G Typing of hepatitis C virus isolates and characterization of new subtypes using a line probe assay J Gen Virol 1993 74 1093 1102 8389799
White PA Zhai X Carter I Zhao Y Rawlinson WD Simplified hepatitis C virus genotyping by heteroduplex mobility analysis J Clin Microbiol 2000 38 477 482 10655331
Chayama K Lau JY Genotyping Hepatitis C Virus by Type-Specific Primers for PCR based NS5 Region Hepatitis C Protocols, of Methods in Molecular Medicine 1998 19 Totowa, New Jersey, USA: Humana Press 165 173
Koylkhalov AA Reed KE Rice CM Lau JY Cloning and Assembly of Complex Libraries of Full-Length cDNA clones Hepatitis C Protocols, of Methods in Molecular Medicine 1998 19 Totowa, New Jersey, USA; Humana Press 289 301
Norder H Berrgström A Uhnoo I Aldén J Weiss L Czajkowski J Magnius L Confirmation of Nosocomial Transmission of Hepatitis C Virus by Phylogenetic Analysis of the NS5-B Region J Clin Micro 1998 36 3066 3069
Rispeter K Li M Lechner S Zibert A Roggendorf M Cloning and characterization of a complete open reading frame of hepatitis C virus genome in only two cDNA fragments J Gen Virol 1997 78 2751 2759 9367360
Grakoui A McCourt DW Wychowski C Feinstone SM Rice CM Characterization of the hepatitis C virus-encoded serine proteinase: determination of proteinase-dependent polyprotein cleavage sites J Virol 1993 67 2832 2843 8386278
Blanchard E Brand D Trassard S Goudeau A Roingeard P Hepatitis C virus-like particle morphogenesis J Virol 2002 76 4073 4079 11907246
Dubuisson J Hsu HH Cheung RC Greenberg HB Russell DG Rice CM Formation and intracellular localization of hepatitis C virus envelope glycoprotein complexes expressed by recombinant vaccinia and Sindbis viruses J Virol 1994 68 6147 6160 8083956
Hijikata M Kato N Ootsuyama Y Nakagawa M Shimotohno K Gene mapping of the putative structural region of the hepatitis C virus genome by in vitro processing analysis Proc Natl Acad Sci USA 1991 88 5547 5551 1648221
Lanford RE Notvall L Chavez D White R Frenzel G Simonsen C Kim J Analysis of hepatitis C virus capsid, E1, and E2/NS1 proteins expressed in insect cells Virology 1993 197 225 235 8212557
Deleersnyder V Pillez A Wychowski C Blight K Xu J Hahn YS Rice CM Dubuisson J Formation of native hepatitis C virus glycoprotein complexes J Virol 1997 71 697 704 8985401
Spaete RR Alexander D Rugroden ME Choo QL Berger K Crawford K Kuo C Leng S Lee C Ralston R Thudium K Tung JW Kuo G Houghton M Characterization of the hepatitis C virus E2/NS1 gene product expressed in mammalian cells Virology 1992 188 819 830 1316682
Dubuisson J Rice CM Hepatitis C virus glycoprotein folding: disulfide bond formation and association with calnexin J Virol 1996 70 778 786 8551615
Op De Beeck A Voisset C Bartosch B Ciczora Y Cocquerel L Keck Z Foung S Cosset FL Dubuisson J Characterization of functional hepatitis C virus envelope glycoproteins J Virol 2004 78 2994 3002 14990718
Lo SY Selby MJ Ou JH Interaction between hepatitis C virus core protein and E1 envelope protein J Virol 1996 70 5177 5182 8764026
Matsumoto M Hwang SB Jeng KS Zhu N Lai MM Homotypic interaction and multimerization of hepatitis C virus core protein Virology 1996 218 43 51 8615040
Yasui K Wakita T Tsukiyama-Kohara K Funahashi SI Ichikawa M Kajita T Moradpour D Wands JR Kohara M The native form and maturation process of hepatitis C virus core protein J Virol 1998 72 6048 6055 9621068
Harada S Watanabe Y Takeuchi K Suzuki T Katayama T Takebe Y Saito I Miyamura T Expression of processed core protein of hepatitis C virus in mammalian cells J Virol 1991 65 3015 3021 1709694
Selby MJ Choo QL Berger K Kuo G Glazer F Eckart M Lee C Chien D Kuo C Houghton M Expression, identification and subcellular localization of the proteins encoded by the hepatitis C viral genome J Gen Virol 1993 74 1103 1113 8389800
Lo SY Selby M Tong M Ou JH Comparative studies of the core gene products of two different hepatitis C virus isolates: two alternative forms determined by a single amino acid substitution Virology 1994 199 124 131 8116235
Yeh CT Lo SY Dai DI Tang JH Chu CM Liaw YF Amino acid substitutions in codons 9–11 of hepatitis C virus core protein lead to the synthesis of a short core protein product J Gastroenterol Hepatol 2000 15 182 191 10735543
Nahmias AJ Ashman RB de-Thé G, Henle W, Rapp F The Immunology of Primary and Recurrent Herpesvirus Infection: An Overview Oncogenesis and Herpesviruses III Part 2: Cell-Virus Interactions, Host Response to Herpesvirus Infection and Associated Tumours, Role of Co-factors 1978 Lyon, France: International Agency for Research on Cancer 659 673
Stevens JG de-Thé G, Henle W, Rapp F Persistent, Chronic, and Latent Infections by Herpesviruses: A Review Oncogenesis and Herpesviruses III Part 2: Cell-Virus Interactions, Host Response to Herpesvirus Infection and Associated Tumours, Role of Co-factors 1978 Lyon, France: International Agency for Research on Cancer 675 685
Gallo RC Salahuddin SZ Popovic M Shearer GM Kaplan M Haynes BF Palker TJ Redfield R Oleske J Safai B White G Foster P Markham PD Frequent Detection and Isolation of Cytopathic Retroviruses (HTLV-III) from Patients with AIDS and at Risk for AIDS Science 1984 224 500 503 6200936
Ivashkina N Wolk B Lohmann V Bartenschlager R Blum HE Penin F Moradpour D The hepatitis C virus RNA-dependent RNA polymerase membrane insertion sequence is a transmembrane segment J Virol 2002 76 13088 13093 12438637
Gosert R Egger D Lohmann V Bartenschlager R Blum HE Bienz K Moradpour D Identification of the hepatitis C virus RNA replication complex in Huh- 7 cells harboring subgenomic replicons J Virol 2003 77 5487 5492 12692249
Geijtenbeek TB Kwon DS Torensma R van Vliet SJ van Duijnhoven GC Middel J Cornelissen IL Nottet HS KewalRamani VN Littman DR Figdor CG van Kooyk Y DC-SIGN, a dendritic cell-Specific HIV-1 binding protein that enhances trans-infection of T cells Cell 2000 100 587 597 10721995
McDonald D Wu L Bohks SM KewalRamani VN Unutmaz D Hope TJ Recruitment of HIV and its receptors to dendritic cell T-cell junctions Science 2003 300 1295 1297 12730499
Feinberg H Mitchell DA Drickamer K Weis WI Structural basis for selective recognition of oligosaccharides by DC-SIGN and DC-SIGNR Science 2001 294 2163 2166 11739956
Pohlmann S Zhang J Baribaud F Chen Z Leslie GJ Lin G Granelli-Piperno A Doms RW Rice CM McKeating JA Hepatitis C virus glycoproteins interact with DC-SIGN and DC-SIGNR J Virol 2003 77 4070 4080 12634366
Masciopinto F Freer G Burgio VL Levy S Galli-Stampino L Bendinelli M Houghton M Abrignani S Uematsu Y Expression of human CD81 in transgenic mice does not confer susceptibility to hepatitis C virus infection Virology 2002 304 187 196 12504561
Pileri P Uematsu Y Campagnoli S Galli G Falugi F Petracca R Weiner AJ Houghton M Rosa D Grandi G Abrignani S Binding of Hepatitis C virus to CD81 Science 1998 282 938 941 9794763
Sasaki M Yamauchi K Nakanishi T Kamogawa Y Hayashi N In vitro binding of hepatitis C virus to CD81-positive and -negative human cell lines J Gastroenterol Hepatol 2003 18 74 79 12519228
| 15840164 | PMC1090622 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 Apr 19; 2:37 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-37 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-411584769710.1186/1743-422X-2-41ResearchComparison of amplification enzymes for Hepatitis C Virus quasispecies analysis Polyak Stephen J [email protected] Daniel G [email protected] Michael A [email protected] James Y [email protected] Margaret C [email protected] Karen L [email protected] Herbert L [email protected] Bisceglie Adrian M [email protected] William M [email protected] Chihiro [email protected] David R [email protected] HALT-C Trial Group [email protected] Virology Division, Department of Laboratory Medicine, University of Washington, Seattle, WA, USA2 Department of Microbiology, University of Washington, Seattle, WA, USA3 Department of Pathobiology, University of Washington, Seattle, WA, USA4 Department of Biostatistics, University of Washington, Seattle, WA, USA5 Department of Medicine, University of Washington, Seattle, WA, USA6 Department of Pediatrics, University of Washington, Seattle, WA, USA7 Division of Gastrointestinal and Liver Diseases, University of Southern California, Los Angeles, CA, USA8 Liver-Biliary-Pancreatic Center and the General Clinical Research Center, University of Connecticut Health Center, Farmington, CT, USA9 Division of Gastroenterology and Hepatology, Saint Louis University School of Medicine, St. Louis, MO, USA10 Division of Digestive and Liver Diseases, University of Texas Southwestern Medical Center, Dallas, TX, USA2005 22 4 2005 2 41 41 14 4 2005 22 4 2005 Copyright © 2005 Polyak et al; licensee BioMed Central Ltd.2005Polyak 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
Hepatitis C virus (HCV) circulates as quasispecies (QS), whose evolution is associated with pathogenesis. Previous studies have suggested that the use of thermostable polymerases without proofreading function may contribute to inaccurate assessment of HCV QS. In this report, we compared non-proofreading (Taq) with proofreading (Advantage High Fidelity-2; HF-2) polymerases in the sensitivity, robustness, and HCV QS diversity and complexity in the second envelope glycoprotein gene hypervariable region 1 (E2-HVR1) on baseline specimens from 20 patients in the HALT-C trial and in a small cohort of 12 HCV/HIV co-infected patients. QS diversity and complexity were quantified using heteroduplex mobility assays (HMA).
Results
The sensitivities of both enzymes were comparable at 50 IU/ml, although HF-2 was more robust and slightly more sensitive than Taq. Both enzymes generated QS diversity and complexity scores that were correlated (r = 0.68; p < 0.0001, and r = 0.47; p < 0.01; Spearman's rank correlation). QS diversity was similar for both Taq and HF-2 enzymes, although there was a trend for higher diversity in samples amplified by Taq (p = 0.126). Taq amplified samples yielded complexity scores that were significantly higher than HF-2 samples (p = 0.033). HALT-C patients who were HCV positive or negative following 20 weeks of pegylated IFN plus ribavirin therapy had similar QS diversity scores for Taq and HF-2 samples, and there was a trend for higher complexity scores from Taq as compared with HF-2 samples. Among patients with HCV and HIV co-infection, HAART increased HCV QS diversity and complexity as compared with patients not receiving therapy, suggesting that immune reconstitution drives HCV QS evolution. However, diversity and complexity scores were similar for both HF-2 and Taq amplified specimens.
Conclusion
The data suggest that while Taq may overestimate HCV QS complexity, its use does not significantly affect results in cohort-based studies of HCV QS analyzed by HMA. However, the use of proofreading enzymes such as HF-2 is recommended for more accurate characterization of HCV QS in vivo.
hepatitis C virusHALT-Cquasispecieshypervariable regionE2
==== Body
Background
HCV exists as quasispecies (QS) in infected individuals, consisting of a predominant viral variant and related, yet genetically distinct minor variants [1]. The study of HCV QS has historically focused on the hypervariable region 1 (HVR1) of the second envelope (E2) glycoprotein gene [2,3], the most variable region of the HCV genome. Early studies revealed that E2-HVR1 is a target of neutralizing antibodies [4-10]. Immune pressure is thought to be chiefly responsible for the fixation of the mutations in this region of the E2 gene.
HCV QS have been analyzed in many different patient cohorts. HVR1 QS evolution may reflect progression of liver disease [11-15]. HCV QS also reflect the outcome of acute HCV infection [16], and responses to antiviral therapy [17]. More recent studies have investigated the effect of HCV/HIV co-infection on HCV QS dynamics [18,19]. In the current study, we analyzed 20 baseline samples from the HALT-C trial and 12 HCV-HIV co-infected patient samples. The HALT-C study is a randomized multi-center clinical trial to assess the effects of long-term pegylated interferon-α (peg-IFN) therapy on the progression of liver fibrosis and development of decompensated liver disease in hepatitis C patients who are non-responders to prior pegylated IFN plus ribavirin therapy [20,21].
Viral QS have been analyzed by many techniques. Cloning and sequencing is the gold standard. Electrophoretic mobility-based assays, including single strand conformation polymorphism analysis (SSCP) and heteroduplex mobility analysis (HMA) allow determination of HCV QS heterogeneity without the need for sequencing (reviewed in [22]). HMA was originally described for analyzing the sequence heterogeneity of the envelope gene of human immunodeficiency virus (HIV) [23,24]. HMA involves hybridization of a radioactive probe generated from a QS variant to either heterogeneous PCR reaction products derived by direct PCR amplification from clinical specimens, or to homogeneous HVR1 sequences derived from cloned QS variants (clonal frequency analysis, CFA; [25]). Hybridizations between the probe and various target sequences result in the formation of double stranded DNA molecules (heteroduplexes) that produce shifts when the hybrids are separated on non-denaturing polyacrylamide gels. The shifts are determined by comparison to a homoduplex probe control (probe hybridized to itself). The extent of the heteroduplex shift compared to the homoduplex control is proportional to the degree of sequence divergence between the two DNA molecules. We have shown an excellent correlation between genetic diversity and complexity (number of variants) of individual QS variants derived by HMA as compared with standard cloning and sequencing of the HVR1 [11-13,25,26]. Moreover, HMA can be applied to studies of HCV QS evolution, in the context of therapy, transmission, and pathogenesis [11,12,25-29].
Taq polymerase is the enzyme used in most studies of HCV QS analysis. However, given that Taq lacks proofreading activity, mathematical debates have been raised to suggest that QS evolution is overestimated by Taq induced mutations [30]. Indeed, in one report, use of Taq increased the proportion of minor QS variants [31], and we have found that, in general, QS diversity is lower if proof-reading enzymes are used instead of Taq [32]. But the question arises as to whether Taq induced mutations affect the outcomes/conclusions in cohort-based studies on HCV QS. Thus, in the current study, we compared Taq and HF-2 polymerases in terms of sensitivity and robustness, and in terms of evaluating HCV QS diversity and complexity as assessed by CFA of the E2-HVR1 on baseline specimens from 20 patients in the HALT-C trial and 12 HCV-HIV co-infected patients receiving or not receiving HAART.
Results
Clinical characteristics of the HALT-C and HCV-HIV co-infected patients are depicted in Tables 1 and 2. For HALT-C patients, all patients were infected with HCV genotype 1, with 10/20 (50%) being genotype 1a, and 7/20 (35%) being genotype 1b. Subtype designations were not obtained for 2/20 (10%) genotype 1 patients, and 1 patient was designated as type 1a or 1b. 16/20 (80%) patients were male. The average age of patients was 48.9 years, and average viral load was 106.56 IU/ml (3.6 × 106 IU/ml). For the HCV/HIV co-infected patients, all patients were infected with HCV genotype 1, with 10/12 (83%) being genotype 1a, and 2/12 (17%) being genotype 1b. Eight of 12 patients (67%) were male. The average age of the co-infected patients was 44 years, and average viral load was 106.53 IU/ml (3.3 × 106 IU/ml).
Table 1 Clinical and virological characteristics of the 20 HALT-C patients.
Patient Age Gender Genotype Serum HCV RNA at W00 (log10) Diversity Complexity
Taq HF-2 Taq HF-2
1 52 M 1a 6.55 0.9953 1.0000 7 1
2 54 F 1b 6.49 0.9905 1.0000 5 1
3 51 M 1a/b 6.99 0.9958 1.0000 5 1
4 42 M 1a 6.81 0.9877 0.9894 6 2
5 56 M 1a 7.13 0.9979 1.0000 4 1
6 42 M 1 6.55 0.9903 0.9912 8 6
7 73 M 1b 6.78 0.9892 0.9877 7 5
8 45 M 1 6.37 0.9193 0.9805 7 6
9 53 M 1a 7.06 0.9954 1.0000 2 1
10 47 M 1b 6.35 0.9845 0.9994 4 3
11 19 M 1b 5.72 0.9964 0.9994 2 2
12 49 M 1a 7.06 0.9979 0.9986 3 3
13 42 F 1b 6.26 0.9993 0.9781 2 2
14 42 M 1a 7.30 0.9810 0.9685 8 9
15 61 M 1b 5.85 0.9858 0.9624 4 5
16 54 M 1b 6.93 0.9047 0.9409 6 9
17 45 F 1a 6.72 0.9962 0.9917 2 5
18 47 F 1a 6.14 0.9966 0.9984 4 2
19 50 M 1a 6.37 0.9962 0.9968 4 3
20 53 M 1a 5.81 0.989 0.9943 6 5
AVG 48.85 6.56 0.9845 0.9889 4.8 3.6
HCV genotypes were determined by INNO-LiPA. HCV RNA was quantified with the Roche COBAS Monitor assay, and is expressed as log10 IU/mL. QS heterogeneity was assessed using the clonal frequency analysis (CFA) technique. Diversity scores represent the average heteroduplex mobility ratios (HMR) for either Taq or HF-2 enzymes. Complexity scores represent the total number of distinct gel shift variants analyzed by CFA. W00 represents the week 0 or baseline sample. AVG represents average.
Table 2 Clinical and virological characteristics of the 12 HCV/HIV co-infected patients.
Patient Age Gender Genotype Serum HCV RNA (log10) Diversity Complexity
Taq HF-2 Taq HF-2
1 49 F 1a 6.37 0.9856 0.9616 4 5
2 40 M 1a 5.96 0.9987 0.9978 2 2
3 36 F 1b 5.53 1.0000 1.0000 1 1
4 38 M 1a 6.75 0.9313 0.9715 8 11
5 42 M 1a 6.94 1.0000 1.0000 1 1
6 40 M 1b 5.55 0.9912 0.9997 5 2
7 41 M 1a 6.39 0.9909 1.0000 6 1
8 45 F 1a 6.25 0.9856 0.9468 4 7
9 47 M 1a 5.68 0.9945 0.9996 2 2
10 41 M 1a 5.92 0.9575 0.9430 10 10
11 61 F 1a 6.83 0.9389 0.9283 7 6
12 51 M 1a 6.99 0.9890 0.9638 2 7
AVG 44.2 6.53 0.9803 0.9760 4.33 5.0
HCV genotypes were determined by INNO-LiPA, while HCV RNA was quantified with the Roche COBAS Monitor assay, and is expressed a log10IU/mL. QS heterogeneity was assessed using the clonal frequency analysis (CFA) technique. Diversity scores represent the average heteroduplex mobility ratios (HMR) for either Taq or HF-2 enzymes. Complexity scores represent the total number of distinct gel shift variants analyzed by CFA. AVG represents average.
Figure 1 presents the sensitivity of the E2-HVR1 PCR using Taq versus HF-2 polymerases. Serial dilutions of the HCV international standard were run through the assay, and PCR products were visualized on agarose gels. As shown in Fig. 1, Taq produced amplification products at all dilutions, with 1 of 2 duplicate samples giving a signal at a dilution of 50 IU/ml. However, the HF-2 enzyme produced more PCR product than Taq at all dilutions and both replicates were positive at 50 IU/ml. Below each lane is the result of Roche COBAS Amplicor qualitative RT-PCR testing of the same serum specimens, which are scored as positive (+) or negative (-). At 50 IU/ml the Amplicor test was positive for 1 of the 2 duplicates. The data indicate that the sensitivities of the Taq and HF-2 enzymes were similar to the Amplicor assay, and the HF-2 enzyme was slightly more robust and sensitive than Taq polymerase.
Figure 1 Comparison of the sensitivity of the E2-HVR1 PCR using Taq and HF-2 enzymes. RNA was extracted from duplicate serial dilutions of a WHO HCV standard and RT-PCR was performed with Taq and HF-2 enzymes. The dilutions corresponded to 50,000 (5K), 1,000 (1K), 500, 100, and 50 IU/ml, and are indicated above each lane. The position of the 176 bp E2-HVR1 is indicated with arrows. MW represents the 100 base pair DNA molecular weight marker. Below each lane is the result of testing of the same dilution of the standard with the Roche COBAS Amplicor assay. The result of this test gives a positive (+) or negative (-) result.
The 2 enzymes produced QS diversity and complexity scores that were correlated (Linear regression analysis: R-squared: 0.4113, p < 0.0001 for diversity, and R-squared: 0.311, p < 0.001 for complexity; Spearman's rank correlation test: r= 0.68, p < 0.0001 for diversity, and r = 0.47, p < 0.01 for complexity). Figure 2 depicts representative clonal frequency analyses (CFA) of HALT-C baseline samples from patient 9 (figure 2A), patient 6 (figure 2B) and patient 16 (figure 2C), amplified by both Taq and HF-2 enzymes. The first lane of each gel represents the homoduplex probe control, obtained by hybridizing the radiolabeled probe to its non-labeled self. The homoduplex (designated as "HD" in the figure) serves as the reference point for all comparisons of individual clonal gel shift patterns. The second lane of each gel represents the heterogenous (ie non-clonal) PCR product, which contains all the QS variants amplified from the serum sample, and is designated as "H" in the figure. For each CFA, a shift control is also included, which involves the hybridization of the probe hybridized to a different HVR1 PCR product. In all cases, shift controls produced clearly identifiable shifts, indicating the hybridization reaction and electrophoresis was successful (data not shown). An internal control is also derived from a comparison of the heterogenous PCR product ("H") versus CFA gel shift patterns. If the gel shift patterns of the individual clones are representative of the heterogenous PCR product, the CFA was successful [11,26,27].
Figure 2 Representative autoradiograms of clonal frequency analyses of HALT-C patients with low (patient 9, panel A), intermediate (patient 6, panel B) and high (patient 16, panel C) QS diversity and complexity. E2-HVR1 RT-PCRs were performed with Taq and HF-2 enzymes. PCR products were cloned as described in the Materials and Methods, and individual colonies were picked and re-amplified. Lane 1 represents the homoduplex (HD) control and represents the probe hybridized to itself. Lane 2 represents the heteroduplex profile of the heterogenous (ie not cloned) E2-HVR1 PCR product and is designated "H". Panels D and E are graphical summaries of HMR and Complexity in the 3 patients.
As shown in Figure 2A, patient 9 had very few discernible gel shifts, and the shifts themselves were not very distinct from the homoduplex probe control. The same pattern was observed when either Taq or HF-2 was used. Indeed, HMR scores for Taq and HF-2 (0.9954 vs. 1.0000) were similar, as were the complexity scores for samples analyzed by Taq and HF-2 (2 vs.1 variants). The flat line nature of the CFA pattern was not due to a failure of the hybridization reaction, because the shift control (the same probe hybridized to a different HVR1 PCR product) produced a significant gel shift (data not shown). Moreover, the same pattern was observed when the heterogenous PCR product (prior to cloning) was hybridized to the probe (Figure 2A, lane 2). The results indicate that this patient had minimal QS heterogeneity. In contrast, patient 6 had clearly identifiable gel shifts when compared to the probe (Figure 2B). The same pattern was observed whether the reaction was performed with Taq or HF-2 enzyme. Again, Taq and HF-2 scores were similar both for HMR (0.9903 vs 0.9912 (Taq vs. HF-2)) and complexity (8 vs. 6 variants (Taq vs. HF-2)). The gel shift pattern observed with CFA was similar to the pattern observed when heterogenous PCR product from the same time point was hybridized to the probe (Figure 2B, lane 2). Patient 16 displayed even more marked QS heterogeneity (Figure 2C). This was quantitated both in terms of HMR (0.9047 vs. 0.9409 (Taq vs. HF-2)) and complexity (6 vs. 9 variants, (Taq vs. HF-2)). The bar graphs in Figures 2D and 2E summarize the diversity (HMR) and complexity data, and confirm the progressively increasing QS diversity and complexity in patients 9, 6, and 16. Note that increased QS diversity is reflected as a decrease in HMR.
Furthermore, Taq gave lower HMRs indicative of higher diversity in all 3 patients, and higher complexity scores in 2 of the 3 patients. The data suggest that Taq may overestimate HCV QS genetic diversity and complexity when analyzed by HMA.
To further examine this issue, CFA was performed for 20 HALT-C baseline specimens, and QS diversity scores and complexity scores were determined. These data are summarized in Table 1 and depicted graphically in Figure 3. Figure 3A depicts the HMR results, while Figure 3B depicts the complexity scores for the 20 HALT-C patients analyzed by both enzymes. The average HMR for Taq amplified samples was 0.9845 (range 0.9047–0.9993), while the average HMR for the HF-2 enzyme was 0.9889 (range 0.9407–1.000). Taq samples appeared somewhat more diverse than HF-2 samples. However, this trend did not reach statistical significance (p = 0.126). Similarly, the average complexity for Taq amplified samples was 4.8 (range 2–8), while the average complexity for the HF-2 enzyme was 3.6 (range 1–9). QS complexity scores were higher in samples amplified by Taq as compared with samples amplified by HF-2 (p = 0.033).
Figure 3 Comparison of quasipecies genetic diversity (assessed by HMR, Panel A) and complexity (panel B) values in samples processed with Taq or HF-2 polymerases in HALT-C baseline specimens. The box plots represent the means and ranges of the HMR and complexity scores for 400 HVR1 clones amplified by each polymerase, for a total of 800 clones (20 patients × 20 clones/patient × 2 enzymes). Error bars represent standard deviations. Significance values above each panel were derived from Wilcoxon Signed Ranks tests.
Cumulatively, the data indicate that Taq overestimates HCV QS genetic diversity and complexity. However, what is not clear is whether this tendency for overestimation by Taq impacts QS scores in clinical studies.
To address this issue, we grouped the 20 HALT-C patients according to whether they were serum positive or negative for HCV RNA following 20 weeks of pegylated IFN plus ribavirin therapy. By this criterion, 6 patients were HCV RNA negative and 14 patients were HCV RNA positive at week 20. Figure 4 presents the HMR and complexity scores for the 2 groups of patients and demonstrates that HMR and complexity scores were similar among patients who were HCV RNA negative or positive following 20 weeks of peg-IFN plus ribavirin, regardless of the enzyme used. There was a trend for HF-2 samples generating lower complexity scores as compared to Taq samples, but the difference was not significant. Note also that the lower complexity scores from HF-2 as compared with Taq processed samples among week 20 responders and non-responders mirrored the results when all patients were considered together (Figure 3).
Figure 4 Comparison of quasipecies genetic diversity (assessed by HMR, Panel A) and complexity (panel B) scores generated with Taq or HF-2 polymerases, in 20 HALT-C samples who were HCV RNA negative (N = 6) or HCV RNA positive (N = 14) at week 20 (W20) of pegylated IFN plus ribavirin therapy. The box plots represent the means and ranges of the HMR and complexity scores for 400 HVR1 clones amplified by each polymerase, for a total of 800 clones (20 patients × 20 clones/patient × 2 enzymes). Error bars represent standard deviations. Wilcoxon Signed Ranks tests determined that the differences between enzymes and patient groups were not statistically significant.
To further determine if the choice of amplification enzyme affects results in patient cohort studies, we compared QS diversity and complexity on samples from 12 HCV/HIV co-infected patients treated or not treated with highly active anti-retroviral therapy (HAART). As shown in Figure 5, patients receiving HAART had increased QS diversity (shown as a decrease in HMR in panel A) and complexity (panel B) as compared to patients who did not receive HAART. This trend was apparent regardless of whether Taq and HF-2 were used. The increase in diversity and complexity scores upon HAART did not reach statistical significance. Moreover, HMR and complexity scores were not significantly different between Taq and HF-2 samples. In fact, HAART-samples amplified by HF-2 showed a trend for higher complexity scores (complexity = 4) as compared with Taq samples (complexity = 3.2). These data suggest that although Taq may overestimate QS complexity, it likely does not mask potentially important clinical associations when HCV QS are analyzed by HMA.
Figure 5 Comparison of quasipecies genetic diversity (assessed by HMR, Panel A) and complexity (panel B) values processed with Taq or HF-2 polymerases, in 12 HCV/HIV co-infected samples treated (N = 7) or not treated (N = 5) with HAART. The box plots represent the means and ranges of the HMR and complexity scores for 240 HVR1 clones amplified by each polymerase, for a total of 480 clones (12 patients × 20 clones/patient × 2 enzymes). Error bars represent standard deviations. Wilcoxon Signed Ranks tests determined that all differences were not statistically significant.
Discussion
In the current investigation, we found that the sensitivity of Taq and HF-2 enzymes in amplifying the E2-HVR1 were similar to the qualitative Roche COBAS Amplicor RT-PCR assay. The HF-2 enzyme generated more PCR product and was more sensitive than Taq. For HALT-C samples, QS diversity, expressed as an HMR, was similar for both Taq and HF-2 enzymes, although Taq tended to give higher diversity scores. HCV QS complexity was significantly higher in samples amplified by Taq as compared with samples amplified by HF-2. In contrast, in HCV/HIV co-infected samples, diversity and complexity scores were similar for both enzymes. The data from this limited cohort suggest that QS results are not significantly influenced by choice of polymerase when using the CFA method.
Our data indicate that Taq induced errors provide inflated estimates of HCV QS diversity and complexity. The data are in accord with previous mathematical models of HCV QS mutations [30], and a recent study that demonstrated that Taq induced errors can generate minor QS variants [31]. However, our results showed only a trend for increased HMR for Taq amplified HALT-C samples as well as for HMR and complexity scores for HCV/HIV co-infected samples. Based on the current results, it would seem that Taq and proof-reading enzymes are acceptable for HCV QS analyzed by CFA. However, further studies on larger patient cohorts need to be performed.
Genetic evolution in the E2-HVR1 is believed to reflect selective forces imposed on this domain by neutralizing antibodies [4-10]. It is also possible that cell mediated immune responses may impart selective pressures on HVR1. In the face of immune pressure, the plasticity of the HVR1 may allow the virus to persist. However, recent studies suggest that the HVR1, although it is the most variable region in the HCV genome, has certain constraints in its structure. It is intriguing that key basic residues are highly conserved, which maintains the chemicophysical properties and conformation of the HVR1 [33]. Because HVR1 is an exposed domain on the E2 protein, these data suggest that the positively charged HVR1 is involved in interactions with negatively charged molecules such as lipids, proteins, or glycosaminoglycans (GAGs). As such, the HVR1 may interact with GAGs facilitating host cell recognition and attachment [33]. In this regard, it has recently been shown that high pretreatment HCV QS diversity and complexity at baseline are associated with non-response to pegylated IFN ribavirin in HALT-C patients [34].
HAART therapy appeared to increase HCV QS diversity and complexity, regardless of the enzyme used, but this did not reach statistical signficance in this small cohort. These data are consistent with recent reports that suggest that HAART-induced immune reconstitution drives HCV QS evolution [18,19]. Additional prospective trials on larger patient cohorts are required to further examine the relationships between immune pressure, HCV QS evolution, and liver disease progression in patients co-infected with HIV and HCV.
Materials and methods
Patients
Twenty patients who met entry criteria for the HALT-C trial were included in the current study. Serum samples from baseline (week 0; (W00)), prior to the start of IFN therapy, were analyzed. Written, informed consent was provided by all patients, following institutional and trial specified IRB regulations. The design and conduct of the HALT-C trial have been described [21]. Briefly, in the initial or lead-in course of therapy, all patients are treated with pegylated interferon (Pegasys, Roche) plus ribavirin for a period of 24 weeks, with virologic assessment at 20 weeks to assess whether or not they are virological responders. Patients who are responders at week 20 continue treatment and receive a full course of 48 weeks of therapy. Those who remain HCV-positive in serum at 20 weeks are randomized to receive either continued low-dose pegylated interferon (without further ribavirin) or no further treatment beyond careful observation for the ensuing 3.5 years. Immunology and virology ancillary studies including QS analysis are aimed at increasing our understanding of the complex interactions of virus and host in this debilitating and widespread disease [34].
Samples from twelve patients co-infected with HCV and HIV were also analyzed. Seven of these patients received HAART, while 5 patients did not. All 32 patients were infected with HCV of genotype 1.
E2-HVR1 RT-PCR Amplification and Cloning
HCV RNA was extracted from 100 μL of patient sera using HCV RNA isolation columns (Qiagen). RNA was resuspended in DEPC-treated water, and converted into cDNA using oligonucleotide primers and AMV reverse transcriptase, as described previously [26]. The same cDNA sample was then amplified using either Taq (Perkin Elmer, Wellesley, MA) or Advantage High Fidelity 2 (HF-2) (Clontech; Mountain View, CA) polymerases, using a nested PCR reaction as described [26]. To provide a hot start, AmpliWax (Perkin Elmer) beads were used to separate Taq enzyme and primers, whereas the HF-2 enzyme mixture contains an anti-Taq antibody. The primers generated the expected second round product of 176 base pairs. PCR products were excised and purified with the QiaEx purification system (Qiagen, Valencia, CA), ligated into pCR2 vector, and transformed into TOP10 cells as described [11].
Sensitivity of E2-HVR1 PCR Assay
To determine the sensitivity of the E2-HVR1 PCR assay using Taq versus HF-2 polymerases, serial dilutions of an HCV World Health Organization (WHO) international standard [35] were prepared from 50,000 to 50 International Units/milliliter (IU/ml). RNA was extracted, converted into cDNA and amplified by nested PCR as described above. PCR products were separated on agarose gels. Aliquots of WHO standards were also run through the Roche Amplicor assay for internal comparison.
HCV RNA quantitation and genotyping
HCV RNA was qualitatively measured by RT-PCR using the Roche COBAS Amplicor HCV test version 2.0 (RocheMolecular Systems, Branchburg, NJ), and quantitatively using Roche COBAS Amplicor HCV Monitor v. 2.0 (Roche Molecular Systems, Branchburg, NJ). Genotyping was performed using the INNO-LiPA HCV II Kit (Bayer Diagnostics, Emeryville, CA) All assays were performed according to manufacturer's specifications.
Clonal Frequency Analysis (CFA)
For each cloned HVR1 PCR product, 20 colonies were picked directly into tubes for re-amplification of the second round PCR product. Thus, for each patient, 20 PCR products representing 20 individual QS clones derived from the baseline serum were analyzed. PCR products were visualized on ethidium bromide-stained agarose gels, and one HVR1 PCR product was randomly selected, purified as described above, and end-labeled with 32P ATP and T4 polynucleotide kinase. Unincorporated label was separated with Centrisep columns (Princeton Separations, Adelphia, NJ), and the labeled DNA probe was eluted from the column. Labeled probe was hybridized directly to each amplified PCR product for 2 hours at 55°C as described [11]. Hybrids were separated on 6% non-denaturing polyacrylamide MDE gels (Cambrex, Baltimore, MD) and visualized by autoradiography as described [11].
Data Analysis
QS complexity was determined by counting the total number of unique gel shift patterns. QS genetic diversity was determined by deriving the average heteroduplex mobility of all clones relative to the homoduplex probe control. A heteroduplex mobility ratio (HMR) was calculated by dividing the distance in millimeters (mm) from the origin of the gel to the heteroduplex by the distance in mm from the origin to the homoduplex control. In cases where both strands of the heteroduplex were clearly distinguishable, the average of the distance of each strand of the heteroduplex was used to calculate heteroduplex mobility [25]. The HMRs for all variants in the population were averaged to provide the final HMR value. Non-parametric Wilcoxon Signed Ranks tests were used to compare the differences in QS complexity and diversity scores between the different enzymes. Linear regression analysis and Spearman's rank correlation tests were also used to determine the correlation of Taq and Ad-HF2 measurements of QS diversity and complexity.
Acknowledgements
This is publication number 9 from the HALT-C Trial Group. Financial support: This study was supported by the National Institute of Diabetes & Digestive & Kidney Diseases and the National Institute of Allergy and Infectious Diseases under contract numbers N01 DK92318, DK92319, DK92321, DK92324, DK92325, DK92326, and DK92328, with supplemental funds from the National Cancer Institute, the National Center for Minority Health and Health Disparities and by General Clinical Research Center grants from the National Center for Research Resources, National Institutes of Health (grant numbers M01 RR00043, RR00633, RR06192). Additional funding to conduct this study was provided by Roche Laboratories, Inc.
Authors with no financial relationships to disclose are: SJP, DGS, MAA, JYD, and CM. Financial relationships of authors with Roche Laboratories/Hoffmann-La Roche, Inc., are as follows: MCS is a consultant and speaker's bureau member; KLL is a consultant and receives research support; HLB is a consultant, speaker's bureau member, and receives research support; AMDB is a consultant, speaker's bureau member, and receives research support; WML is a speaker's bureau member and receives research support; and DRG receives educational grant support.
The authors gratefully acknowledge the HALT-C study investigators, research staff, and participants. The members of the HALT-C Trial Group who contributed to the performance of the clinical study are: Dawn Bombard, RN; Minjun Chung; Maureen Cormier, NP-C; Nicole Crowder, LVN; Michael Doherty, MS; Rivka Elbein, RN; Donna Giansiracusa, RN, BSN; Carol B. Jones, RN; Michelle Kelley, ANP; Debra King, RN; Rachel Life, BS; Peter F. Malet, MD; Savant Mehta, MD; Susan L. Milstein, RN; Patricia Osmack; Amanda Reeck; Marisol Serrano; Rohit Shankar; Judy Thompson, RN; and Elizabeth Wright, PhD.
==== Refs
Martell M Esteban JI Quer J Genesca J Weiner A Esteban R Guardia J Gomez J Hepatitis C virus (HCV) circulates as a population of different but closely related genomes: quasispecies nature of HCV genome distribution J Virol 1992 66 3225 3229 1313927
Hijikata M Kato N Ootsuyama Y Nakagawa M Ohkoshi S Shimotohno K Hypervariable regions in the putative glycoprotein of hepatitis C virus Biochem Biophys Res Commun 1991 175 220 228 1847805
Weiner AJ Brauer MJ Rosenblatt J Richman KH Tung J Crawford K Bonino F Saracco G Choo QL Houghton M Variable and hypervariable domains are found in the regions of HCV corresponding to the flavivirus envelope and NS1 proteins and the pestivirus envelope glycoproteins Virology 1991 180 842 848 1846505 10.1016/0042-6822(91)90104-J
Farci P Alter HJ Wong DC Miller RH Govindarajan S Engle R Shapiro M Purcell RH Prevention of hepatitis C virus infection in chimpanzees after antibody-mediated in vitro neutralization Proc Natl Acad Sci U S A 1994 91 7792 7796 7519785
Higashi Y Kakumu S Yoshioka K Wakita T Mizokami M Ohba K Ito Y Ishikawa T Takayanagi M Nagai Y Dynamics of genome change in the E2/NS1 region of hepatitis C virus in vivo Virology 1993 197 659 668 8249288 10.1006/viro.1993.1641
Kato N Sekiya H Ootsuyama Y Nakazawa T Hijikata M Ohkoshi S Shimotohno K Humoral immune response to hypervariable region 1 of the putative envelope glycoprotein (gp70) of hepatitis C virus J Virol 1993 67 3923 3930 7685404
Okamoto H Kojima M Okada S Yoshizawa H Iizuka H Tanaka T Muchmore EE Peterson DA Ito Y Mishiro S Genetic drift of hepatitis C virus during an 8.2-year infection in a chimpanzee: variability and stability Virology 1992 190 894 899 1325713 10.1016/0042-6822(92)90933-G
van Doorn LJ Capriles I Maertens G DeLeys R Murray K Kos T Schellekens H Quint W Sequence evolution of the hypervariable region in the putative envelope region E2/NS1 of hepatitis C virus is correlated with specific humoral immune responses J Virol 1995 69 773 778 7815542
Weiner AJ Geysen HM Christopherson C Hall JE Mason TJ Saracco G Bonino F Crawford K Marion CD Crawford KA Evidence for immune selection of hepatitis C virus (HCV) putative envelope glycoprotein variants: potential role in chronic HCV infections Proc Natl Acad Sci U S A 1992 89 3468 3472 1314389
Ziebert A Schreier E Roggendorf M Antibodies in human sera specific to hypervariable region 1 of Hepatitis C virus can block viral attachment Virology 1995 208 653 661 7538251 10.1006/viro.1995.1196
Gretch DR Polyak SJ Wilson JJ Carithers RL Perkins JD Corey L Tracking hepatitis C virus quasispecies major and minor variants in symptomatic and asymptomatic liver transplant recipients J Virol 1996 70 7622 7631 8892882
Sullivan DG Wilson JJ Carithers RL Perkins JD Gretch DR Multigene tracking of hepatitis C virus quasispecies after liver transplantation: correlation of genetic diversification in the envelope region with asymptomatic or mild disease patterns Journal of Virology 1998 72 10036 10043 9811742
Wilson JJ Polyak SJ Day TD Gretch DR Characterization of simple and complex hepatitis C virus quasispecies by heteroduplex gel shift analysis: correlation with nucleotide sequencing J Gen Virol 1995 76 1763 1771 9049381
Lyra AC Fan X Lang DM Yusim K Ramrakhiani S Brunt EM Korber B Perelson AS Di Bisceglie AM Evolution of hepatitis C viral quasispecies after liver transplantation Gastroenterology 2002 123 1485 1493 12404223
Fan X Di Bisceglie AM Diversification of hypervariable region 1 of hepatitis C virus after liver transplantation J Med Virol 2003 70 212 218 12696107 10.1002/jmv.10380
Ray SC Wang YM Laeyendecker O Ticehurst JR Villano SA Thomas DL Acute hepatitis C virus structural gene sequences as predictors of persistent viremia: Hypervariable region 1 as a decoy J Virol 1999 73 2938 2946 10074143
Pawlotsky JM Hepatitis C virus genetic variability: pathogenic and clinical implications Clin Liver Dis 2003 7 45 66 12691458 10.1016/S1089-3261(02)00065-X
Mao Q Ray SC Laeyendecker O Ticehurst JR Strathdee SA Vlahov D Thomas DL Human immunodeficiency virus seroconversion and evolution of the hepatitis C virus quasispecies J Virol 2001 75 3259 3267 11238852 10.1128/JVI.75.7.3259-3267.2001
Babik JM Holodniy M Impact of highly active antiretroviral therapy and immunologic status on hepatitis C virus quasispecies diversity in human immunodeficiency virus/hepatitis C virus-coinfected patients J Virol 2003 77 1940 1950 12525628 10.1128/JVI.77.3.1940-1950.2003
Shiffman ML Di Bisceglie AM Lindsay KL Morishima C Wright EC Everson GT Lok AS Morgan TR Bonkovsky HL Lee WM Peginterferon alfa-2a and ribavirin in patients with chronic hepatitis C who have failed prior treatment Gastroenterology 2004 126 1015 1023 15057741 10.1053/j.gastro.2004.01.014
Lee WM Dienstag JL Lindsay KL Lok AS Bonkovsky HL Shiffman ML Everson GT Di Bisceglie AM Morgan TR Ghany MG Evolution of the HALT-C Trial: pegylated interferon as maintenance therapy for chronic hepatitis C in previous interferon nonresponders Control Clin Trials 2004 25 472 492 15465617 10.1016/j.cct.2004.08.003
Forns X Bukh J The molecular biology of hepatitis C virus. Genotypes and quasispecies Clin Liver Dis 1999 3 693 716 vii 11291246
Delwart EL Shpaer EG Louwagie J McCutchan FE Grez M Rubsamen WH Mullins JI Genetic relationships determined by a DNA heteroduplex mobility assay: analysis of HIV-1 env genes Science 1993 262 1257 1261 8235655
Delwart EL Sheppard HW Walker BD Goudsmit J Mullins JI Human immunodeficiency virus type 1 evolution in vivo tracked by DNA heteroduplex mobility assays J Virol 1994 68 6672 6683 8084001
Polyak SJ Faulkner G Carithers RL Corey L Gretch DR Assessment of hepatitis C virus quasispecies heterogeneity by gel shift analysis: correlation with response to Interferon therapy J Infect Dis 1997 175 1101 1107 9129072
Polyak SJ McArdle S Liu SL Sullivan DG Chung MJ Hofgartner WT Carithers RL McMahon BJ Mullins JI Corey L Evolution of hepatitis C virus quasispecies in hypervariable region 1 and the putative interferon sensitivity-determining region during interferon therapy and natural infection J Virol 1998 72 4288 4296 9557719
Gerotto M Sullivan DC Polyak SJ Chemello L Cavalletto L Pontisso P Alberti A Gretch DR Effect of retreatment with interferon alone or interferon plus ribavirin on hepatitis C virus quasispecies diversification in nonresponder patients with chronic hepatitis C Journal Of Virology 1999 73 7241 7247 10438811
Gerotto M Dal P-F Sullivan DG Chemello L Cavalletto L Polyak SJ Pontisso P Gretch DR Alberti A Evidence for sequence selection within the non-structural 5A gene of hepatitis C virus type 1b during unsuccessful treatment with interferon-alpha J Viral Hepatitis 1999 6 367 372 10.1046/j.1365-2893.1999.00166.x
Sullivan DG Kim SS Wilson JJ Stehman-Breen C Gretch DR Investigating hepatitis C virus heterogeneity in a high prevalence setting using heteroduplex tracking analysis J Virol Methods 2001 96 5 16 11516484 10.1016/S0166-0934(01)00303-2
Smith DB McAllister J Casino C Simmonds P Virus 'quasispecies': Making a mountain out of a molehill? Journal Of General Virology 1997 78 1511 1519 9225023
Mullan B Kenny-Walsh E Collins JK Shanahan F Fanning LJ Inferred hepatitis C virus quasispecies diversity is influenced by choice of DNA polymerase in reverse transcriptase-polymerase chain reactions Anal Biochem 2001 289 137 146 11161307 10.1006/abio.2000.4946
Nousbaum J Polyak SJ Ray SC Sullivan DG Larson AM Carithers RL JrGretch DR Prospective characterization of full-length hepatitis C virus NS5A quasispecies during induction and combination antiviral therapy J Virol 2000 74 9028 9038 10982347 10.1128/JVI.74.19.9028-9038.2000
Penin F Combet C Germanidis G Frainais PO Deleage G Pawlotsky JM Conservation of the conformation and positive charges of hepatitis C virus E2 envelope glycoprotein hypervariable region 1 points to a role in cell attachment J Virol 2001 75 5703 5710 11356980 10.1128/JVI.75.12.5703-5710.2001
Rothman AL Morishima C Bonkovsky HL Polyak SJ Ray R Di Bisceglie AM Lindsay KL Malet PF Chang M Gretch DR Associations among clinical, immunological, and viral quasispecies measurements in advanced chronic hepatitis C Hepatology 2005 41 617 625 15726656 10.1002/hep.20581
Saldanha J Lelie N Heath A Establishment of the first international standard for nucleic acid amplification technology (NAT) assays for HCV RNA. WHO Collaborative Study Group Vox Sang 1999 76 149 158 10341329 10.1159/000031040
| 15847697 | PMC1090623 | CC BY | 2021-01-04 16:38:59 | no | Virol J. 2005 Apr 22; 2:41 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-41 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-421585049010.1186/1743-422X-2-42ResearchIntracellular localization of Crimean-Congo Hemorrhagic Fever (CCHF) virus glycoproteins Haferkamp Sebastian [email protected] Lisa [email protected] Tino F [email protected] Heinz [email protected] Ramon [email protected] University of Texas Medical Branch, Department of Pathology, Center for Biodefense and Emerging Infectious Diseases, 301 University Boulevard, Galveston, Texas, 77555-0609 USA2 Special Pathogens Program, National Microbiology Laboratory, Health Canada, CA-R3E 3R2 Winnipeg, Canada3 Stiftung Juliusspital Wuerzburg, 97070 Wuerzburg, Germany4 Department of Medical Microbiology, University of Manitoba, 543-730 William Avenue, Winnipeg, R3E 0W3 Canada2005 25 4 2005 2 42 42 17 2 2005 25 4 2005 Copyright © 2005 Haferkamp et al; licensee BioMed Central Ltd.2005Haferkamp 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
Crimean-Congo Hemorrhagic Fever virus (CCHFV), a member of the genus Nairovirus, family Bunyaviridae, is a tick-borne pathogen causing severe disease in humans. To better understand the CCHFV life cycle and explore potential intervention strategies, we studied the biosynthesis and intracellular targeting of the glycoproteins, which are encoded by the M genome segment.
Results
Following determination of the complete genome sequence of the CCHFV reference strain IbAr10200, we generated expression plasmids for the individual expression of the glycoproteins GN and GC, using CMV- and chicken β-actin-driven promoters. The cellular localization of recombinantly expressed CCHFV glycoproteins was compared to authentic glycoproteins expressed during virus infection using indirect immunofluorescence assays, subcellular fractionation/western blot assays and confocal microscopy. To further elucidate potential intracellular targeting/retention signals of the two glycoproteins, GFP-fusion proteins containing different parts of the CCHFV glycoprotein were analyzed for their intracellular targeting. The N-terminal glycoprotein GN localized to the Golgi complex, a process mediated by retention/targeting signal(s) in the cytoplasmic domain and ectodomain of this protein. In contrast, the C-terminal glycoprotein GC remained in the endoplasmic reticulum but could be rescued into the Golgi complex by co-expression of GN.
Conclusion
The data are consistent with the intracellular targeting of most bunyavirus glycoproteins and support the general model for assembly and budding of bunyavirus particles in the Golgi compartment.
==== Body
Background
Crimean-Congo hemorrhagic fever virus (CCHFV) is a member of the genus Nairovirus, one of five genera in the family Bunyaviridae [1]. Bunyaviruses are enveloped particles with a tripartite, single stranded RNA genome of negative polarity [2-4]. The three genome segments encode four structural proteins: the RNA-dependent RNA polymerase (L protein) is encoded by the large (L) segment, the glycoproteins (GN and GC; previously referred to as G1 and G2) are encoded by the medium (M) segment, and the nucleocapsid protein (N) is encoded by the small (S) segment [2-4].
The virus glycoproteins are likely to play an important role in the natural tick-vertebrate cycle of the virus as well as for the high pathogenicity in humans. Indeed, a highly variable mucin-like region at the amino terminus of the CCHFV glycoprotein precursor has recently been identified, a unique feature of nairoviruses within the family Bunyaviridae [5]. A similar serine-threonine-rich domain has been associated with increased vascular permeability and development of hemorrhages in Ebola hemorrhagic fever [6].
The Nairovirus genus includes 34 described viruses and is divided into seven different serogroups [1]. Only three viruses are known to cause disease: CCHFV, Dugbe virus, and Nairobi sheep disease virus. CCHFV is an arthropod-borne pathogen and the causative agent of a serious form of hemorrhagic fever [7-9] with mortality rates ranging from 15 to 60% [10-17]. The virus is endemic in parts of Africa, Southeastern Europe and Asia as far east as western China [16,18,19]. The geographic distribution of CCHFV infections corresponds most closely with the distribution of Hyalomma ticks, suggesting their principal vector role [18,20,21]. Hyalomma ticks normally feed on a variety of livestock (sheep, goats, cattle, and ostriches), large wild herbivores, hares, and hedgehogs, which can become infected with CCHFV [13,18,22,23]. In contrast to human infections, infection in these animals generally results in inapparent or subclinical disease but generates viremia levels capable of supporting virus transmission to uninfected ticks [10,18,21,23-25]. Transmission to humans occurs either by bites from infected ticks or direct contact with blood or tissues of infected livestock. Nosocomial infections are common [26] and represent a major problem in health care institutions [27].
The widespread geographical distribution of CCHFV, its ability to produce severe human disease with high mortality rates, and fears about its intentional use as a bioterrorism agent makes CCHFV an extremely important human pathogen and a worldwide public health concern. Case management and intervention strategies would greatly benefit from knowledge of the biology and pathogenesis of the virus.
Recently, the expression strategy and biosynthesis of the CCHFV glycoproteins have been studied in more detail including the identification of precursor cleavage sites and the determination of the exact N termini of the two major cleavage products, GN and GC [5,28]. SKI-1, also responsible for the proteolytic processing of the Lassa virus glycoprotein precursor [29], has been identified as the cellular protease responsible for the processing step that generates the N-terminus of mature GN. Another yet unidentified protease is required for GC processing. However, the exact C-terminus of GN could not yet been determined. Two cleavage sites have been predicted for this processing step, one at amino acid position 808 (RKLL) and the other at 940/944 (KKRKK) favouring the cellular proteases SKI-1 and furin, respectively, as the responsible proteases [28].
Bunyaviruses are known to bud from Golgi membranes and the budding site seems to be defined by an retention of the glycoproteins GN and GC at that particular site [3,4]. From a number of studies which have addressed the mechanisms of Golgi targeting and retention, one can conclude that the N-terminal located glycoprotein appears to carry the appropriate signal(s) [30-40] So far, no studies have investigated Golgi targeting and retention of nairovirus glycoproteins.
In this study we cloned the complete M segment ORF of CCHFV, strain IbAr10200, into different expression plasmids. Expression and intracellular localization of the glycoproteins GN and GC were studied and compared to glycoproteins generated by virus infection. Using recombinant fusion proteins between the green fluorescence protein (GFP) and CCHFV glycoproteins, the Golgi targeting/retention signal could be mapped to a hydrophobic region within the cytoplasmic domain of the GN protein.
Results
Sequence determination of the full-length CCHFV M segment
The complete M segment nucleotide sequences of two different sources of CCHFV, strain IbAr10200, was determined and compared to previously published sequences [GenBank: U39455]. Several nucleotide changes resulting in amino acid changes in the glycoprotein precursor were identified (Table 1). In two different CCHF viral RNA samples eight amino acid changes and two silent nucleotide changes could be detected. Four additional amino acid changes were found in sample #2 as well as four silent nucleotide changes not leading to any amino acid alteration. CCHFV RNA sample #1 showed two additional unique amino acid changes.
Table 1 Sequence comparison of available CCHFV IbAr10200 M segment sequences
CCHF IbAr10200 Nucleotide changes compare to U39455 (vRNA position*) Amino acid changes compare to U39455*
Samples 1 and 2 83: C → T 1671: Gly → Arg
713: T → C 1461: Ser → Gly
1621: T → C 1158: Glu → Gly
2263: G → T 944: Thr → Lys
2926: C → T 723: Arg → Lys
2964: A → G 710: – (silent)
3512: C → T 528: Val → Ile
3550: A → G 515: Phe → Ser
4044: G → T 350: – (silent)
4981: T → G 38: His → Pro
Sample 2 3425: T → C 557: Asn → Asp
3427: C→ A 556: Cys → Phe
3429: G → A 555: – (silent)
3435: T → A 555: – (silent)
3441/42: GT → AG 551: Asp → Ala
3444: A → T 550: – (silent)
4247: G → A 282: – (silent)
4610: T → G 162: Thr → Pro
Sample 1 2760/61: TA → AT 778: Ile → Asn
4684: G → A 137: Ser → Phe
* based on [GenBank: U39455] position numbering
Furthermore, we determined the sequences of the exact ends of the M segment using an RNA ligation approach. Beside constructs with nucleotide deletions due to RNA degradation prior to RNA ligation several full-length sequences were determined, demonstrating the expected homologous RNA ends compare to the CCHF S and L segments (Fig. 1). Especially the first and last nine nucleotides of the CCHF M vRNA segment showed high complementarity to the L and S segment ends (bold and italicized nucleotides in Fig. 1), confirming their role as important cis-acting elements for RNA polymerase binding [41,42].
Figure 1 Schematic presentation of the predicted base-paired ends of the CCHFV M vRNA segment. The first nine nucleotides as well as nucleotides at position 11 at both RNA ends are highly conserved within the three genome segments (bold, italics), whereas the first thirteen nucleotides at the 3' and 5' vRNA ends are inverted complementary and can form base-paired terminal regions.
Expression of CCHFV glycoproteins
Based on the recently published N-terminal sequence determination of mature CCHFV glycoproteins [5] and using the above described determined CCHFV M segment sequence (sample #1), expression plasmids for both glycoproteins GN and GC as well as for the glycoprotein precursor (GPC) were generated. Since the C-terminus of CCHFV GN has not yet been determined ([28]; S. Nichol: pers. communication) two constructs were generated containing an N-terminal Influenza HA-tag for detection: pCMV CCHF GN "short" (GNs) and pCMV CCHF GN "long" (GNl). Glycoprotein expression was first analyzed by immunoblot using CCHFV-specific polyclonal or HA-tag antibodies. The CCHF full-length glycoprotein precursor construct (pCAGGS CCHFV GPC) was successfully expressed and correctly processed into the cleavage fragments GC and GN (Fig. 2, lane 2). Molecular weights (GC, 37 kDa and GN, 75 kDa) as determined by immunoblot analysis were in accordance with those of the GC and GN expressed in CCHFV-infected VeroE6 cells (Fig. 2, lane 1). CMV-driven HA-GNs and HA-GNl expression resulted in a protein of approximately 75 kDa (Fig. 2, lanes 3 and 4), similar to authentic GN glycoprotein seen in CCHFV-infected cells (Fig. 2, lane 1). Expression of chicken β-actin-driven GC resulted in a product of approximately 37 kDa, again similar to GC expression in CCHV-infected cells (Fig. 2, lane 5). The data demonstrates that each glycoprotein can be authentically expressed individually from separate plasmids (e.g., pCMV GNs, pCMV GNl and pCAGGS GC) as well as from a clone encoding the GPC precursor (pCAGGS GPC) using polyclonal CCHFV-specific and HA-tag antibodies (Fig. 2). Expression could also be confirmed using CCHFV-specific GC and GN antipeptide antibodies which were kindly provided by S. Nichol, CDC) (data not shown).
Figure 2 Western blotting of CCHFV glycoproteins after transfection of different expression plasmids. BHK-21 cells were either infected with CCHFV (lane 1) or transfected with individual CCHFV glycoprotein expression plasmids (lanes 2–5). Twenty-four hours post infection/transfection cells were harvested and cell lysate used for western blotting, using immune serum specific against CCHFV IbAr10200. Protein bands with expected sizes were detected and confirmed successful expression of CCHFV glycoproteins.
Intracellular localization of CCHFV glycoproteins
Indirect immunofluorescence assays (IFA) were initially performed to analyze the cellular localization of CCHFV glycoproteins. For this, different CCHFV glycoprotein expression plasmids were individually transfected into BHK-21 or 293T cells and 24 to 48 h post transfection the cells were fixed with acetone/methanol or paraformaldehyde for intracellular (Fig. 3A) or surface immunofluorescence analysis (Fig. 3B), respectively. HA-specific monoclonal antibodies were used to detect the two forms of individually expressed N-terminal HA-tagged GN (Fig. 3A: b, c, g, h) and CCHFV GC-specific antibodies were used to monitor β-actin promoter-driven GC expression products (Fig. 3A: a, f). In addition, a CCHFV-specific antiserum was used to detect GN and GC expression from full-length glycoprotein precursor construct pCAGGS GPC (Fig. 3A: d, i) as well as in CCHFV-infected cells (Fig. 3A: e, j). In all cases GN and GC were detected intracellular but never on the cell surface (Figs. 3A and 3B). Mock-infected and -transfected cells were used as negative controls (data not shown). Two different cell lines were used to exclude artificial cell type-specific localization pattern of CCHFV glycoproteins.
Figure 3 Indirect immunofluorescence assays of CCHFV glycoproteins after transfection of different expression plasmids. BHK-21 and 293T cells were transfected with CCHFV glycoprotein expression plasmids. Twenty-four hours post transfection cells were fixed and stained with CCHFV-specific or anti-HA antibodies. A: Cells fixed with methanol/acetone allow analyses of intracellular proteins. Polyclonal antibodies against CCHFV GC were used for GC detection (a, f). CCHFV GN expression from two different CMV-driven expression plasmids was analyzed using anti-HA tag antibodies (b, c, g, h). GN and GC expressed from the GPC were studied using polyclonal anti-CCHFV antibodies (d, i) as well as specific antipeptide antibodies against GN and GC (data not shown). CCHFV-infected cells served as controls (e, j). B: Cells fixed with paraformaldehyde were analyzed for CCHFV G expression on cellular surfaces. GC was stained using anti-CCHFV GC antibodies, GN with anti-HA tag, and mature CCHFV proteins derived from the GPC with anti-CCHFV antibodies. No clearly visible staining correlates with no detectable surface expression.
In a next step we tried to specify the intracellular localization of CCHFV GN and GC glycoproteins expressed from plasmids encoding either the individual glycoproteins or the precursor GPC. Intracellular staining pattern of CCHV-infected cells as well as cells expressing the CCHFV precursor GPC revealed a Golgi complex staining pattern independent of the antibodies used for detection of the individual glycoproteins (Fig. 3A: d, e, i, j). Subsequently, we analyzed the intracellular localization of individually expressed GN and GC. Whereas individually expressed GN showed a Golgi complex localization (Fig. 3A: b, c, g, h), individually expressed GC accumulated in the perinuclear region of the cell indicative of ER localization (Fig. 3A: a, f). Confirmation for these results were achieved by co-immunofluorescence analyzed on a confocal microscope using CCHFV glycoprotein-specific or HA-specific antibodies and either antibodies directed against the ER-specific marker molecule calreticulin or direct staining of the Golgi region with BODIPY-TR C5 ceramides (Fig. 4). Again, CCHFV GN expression from the two expression plasmids pCMV GNs and pCMV GNl overlapped with Golgi staining (Fig. 4: e, f), whereas GC expression overlapped with that of calreticulin (Fig. 4: d). However, co-expression of both CCHFV glycoproteins either from the glycoprotein precursor plasmid or from simultaneous transfection of the two expression plasmids resulted in Golgi targeting of both glycoproteins (Fig. 4: c, g) strongly indicating that GN drives the Golgi localization and that GC needs to interact with GN in order to be transported out of the ER.
Figure 4 Intracellular co-localization studies of CCHFV glycoproteins and cellular compartment markers. BHK-21 cells were either infected with CCHFV or transfected with individual CCHFV glycoprotein expression plasmids. Twenty-four hours post infection/transfection cells were fixed with methanol/acetone and subsequently an indirect immunofluorescence assay was performed. DAPI-stained (for nucleus localization), Golgi or ER compartment stained and CCHFV G expression pictures were taken individually and subsequently merged to study the intracellular CCHFV G localization. a: CCHFV-infected cells stained for CCHF GN and GC proteins; b and c: CCHFV GPC expression plasmid-transfected cells stained for GN or GC, respectively; d: Cells transfected with CCHFV GC expression plasmid stained with GC-specific antibodies and co-stained with ER compartment marker calreticulin; e: CCHFV GNs-transfected cells stained with HA tag-specific antibodies; f: CCHFV GNl-transfected cells stained with HA tag-specific antibodies; g: CCHFV GC expression pattern after co-transfection of CCHFV GNl and GC.
To further strengthen the association of CCHFV glycoproteins with intracellular membrane-containing compartments such as ER and Golgi complex, we performed subcellular fractionation experiments. This method allows the separation of soluble proteins from membrane-associated proteins. CCHFV-infected cells were used for comparison (Fig. 5: lane 1). As expected all expressed CCHFV glycoproteins were exclusively found in the pellet fractions, which contain membrane-associated proteins. This confirms the intracellular localization of these proteins with membrane structures and together with the co-immunofluorescence data confirms either ER or Golgi localization (Fig. 5: lanes 2 to 5). To evaluate the described approach control experiments using either the soluble CCHFV N proteins or the Golgi marker Mannosidase II were performed. As expected CCHF N protein was exclusively found in the soluble fraction, whereas the Golgi marker protein was only detected in the membrane-associate fraction.
Figure 5 Subcellular fractionation studies of expressed CCHFV glycoproteins. CCHFV-infected or CCHFV G expression plasmid-transfected BHK-21 cells were used for a subcellular fractionation study to determine if CCHFV G proteins are membrane associated. The results shown are summarized from two independent experiments. M: membrane fraction, S: soluble fraction, N: CCHF nucleoprotein, Man II: Mannosidase II. For quantification individual protein bands were analyzed and compared using the software Quantity One (BioRad).
Signals for intracellular targeting of CCHFV glycoproteins
After determining the intracellular localization of the CCHFV glycoproteins, we next were interested to determine the signals for intracellular targeting. For this, we generated GFP-fusion proteins containing different fragments of the GC or GN proteins attached to GFP. On the basis of published data obtained with other bunyaviruses we expected Golgi localization signals rather within the transmembrane or cytoplasmic domains than in the ectodomain [3,4]. A CMV-driven GFP expression plasmid (pHL2823, Flick and Hobom, unpublished) was used as a cloning vector for fusing different regions of the CCHFV glycoproteins to the C-terminus of the GFP. Firstly, the different PCR-amplified GN cytoplasmic domain fragments (Table 2) were cleaved with BsmBI and inserted into pHL2823 after BamHI/XbaI endonuclease treatment. In an alternative approach a signal peptide (Ig κ-chain signal of the pDisplay vector) was fused to the GFP N-terminus to allow entry into the secretory pathway. Secondly, the GN transmembrane domain (TM I) was inserted using a hybridized oligonucleotide linker (RF372/RF373: GATCCTTTGGCTATGT AATAACCTGCATACTTTGCAAGGCCATTTTTTACTTGTTAATAATTGTTGGATAAT/ CTAGATTATCCAACAATTATTAACAAGTAAAAAATGGCCTTGCAAAGTATGCAGGTTATTACATAGCCAAAG). The expression of the resulting constructs GFP-GNA, GFP-GNB, GFP-GNC, GFP-GND, GFP-GNE, GFP-GNF, GFP-GNG, GFP-GNH, and GFP-GNI (Fig. 6A) was first verified by immunoblot (data not shown). All constructs expressed GFP-fusion proteins of expected sizes and were subsequently used in co-localization studies. For this two different cell lines (BHK-21 and 293T), for comparison purposes, were transfected with the different plasmid DNAs and GFP fluorescence localization was analyzed using UV-microscopy.
Table 2 Features and construction details of different GFP-CCHFV G fusion proteins
Construct Oligo-nucleotide primer Restriction endonuclease PCR-fragment length Glycoprotein fragment [nt] Fragment length
GFP-G2I RF372
RF373 BamHI
XbaI 72 bp Oligonucleotide linker 3177-3113
64 nt
GFP-G2A RF364
RF363 BsmBI 296 bp 3114-2854
290 nt
GFP-G2B RF364
RF365 BsmBI 332 bp 3114-2818
296 nt
GFP-G2C RF364
RF366 BsmBI 401 bp 3114-2749
365 nt
GFP-G2D RF364
RF367 BsmBI 443 bp 3114-2707
407 nt
GFP-G2E RF364
RF368 BsmBI 512 bp 3114-2638
476 nt
GFP-G2F RF364
RF369 BsmBI 779 bp 3114-2371
743 nt
GFP-G2G RF364
RF370 BsmBI 848 bp 3114-2302
812 nt
GFP-G2H RF364
RF371 BsmBI 995 bp 3114-2155
959 nt
GFP-G1A RF361
RF362 BamHI
XbaI 210 bp 411-220
191 nt
GFP-G1B RF362
RF378 BglII
XbaI 288 bp 489-220
69 nt
* based on [GenBank: U39455] position numbering
Figure 6 GFP-CCHFV G fusion proteins to identify Golgi and ER localization signals. BHK-21 and 293T cells were transfected with individual GFP-CCHFV G fusion proteins. Twenty-four hours post transfection cells were analyzed via UV microscopy. A: Schematic presentation of different GFP-CCHFV G fusion proteins. Green labeled parts represent the GFP protein, red and blue labeled parts show the fragments of the CCHFV G cytoplasmic tails C-terminal fused to the GFP gene. White boxes symbolize the predicted hydrophobic transmembrane domains (TM). Numbers represent the amino acid position from the CCHFV GPC; B: BHK-21 cell analyses; C: 293T cell analyses.
The fusion protein GFP-GNI, containing the TM I of CCHF GN was expressed in the cell cytoplasm in both used cell lines (Figs. 6B and 6C: b) similarly to GFP expressed from the basic vector pHL2823 (Figs. 6B and 6C: a). In case of the signal peptide-containing GFP fusion protein a diffuse staining consistent with the distribution throughout the secretory system was observed (data not shown). Based on this result we conclude that the transmembrane domain TM I does not contain any intracellular targeting signal.
The fusion proteins GFP-GNA and GFP-GNB showed a similar cytoplasmic expression pattern (Figs. 6B and 6C: c, d). GFP-GNA contains the first 87 amino acids from the cytoplasmic domain including the RKLL motif at position 808, which is a predicted protease cleavage motif for generating the C-terminus of the mature GN protein, whereas GFP-GNB has 99 amino acids fused to the GFP C-terminus, corresponding to the first GN cytosolic tail fragment, which is followed by a second hydrophobic region predicted as a potential transmembrane domain 2 (TM II) (compare Fig. 6A). Interestingly, the fusion proteins GFP-GNC, GFP-GND, GFP-GNE, GFP-GNF, GFP-GNG, and GFP-GNH, which contain longer fragments of the predicted GN cytoplasmic domain including additional predicted hydrophobic transmembran regions (Fig. 6A), showed an increased level of similarity to the intracellular pattern of GNl, which contained the entire GN cytoplasmic domain up to the determined mature GC start (Figs. 6B and 6C: e-j). The switch from a diffuse staining pattern to a Golgi complex localization is caused by the addition of TM II to the first 99 amino acids of the cytoplasmic domain resulting in GFP-fusion proteins containing 122 amino acids of the predicted GN cytoplasmic domain (Fig. 6A). These results demonstrate that the Golgi targeting signal is not located within the first 99 amino acids of the GN cytoplasmic domain. However, the addition of an additional hydrophobic 23 amino acid stretch (TM II) result in a co-localization of the GFP-fusion protein with the Golgi complex marker mannosidase II (Fig. 7), demonstrating that a Golgi localization signal is located within the predicted TM II.
Figure 7 Intracellular co-localization analysis of GFP-CCHFV G fusion proteins. 293T cells were transfected with a GFP-CCHFV GN fusion protein construct encoding 123 amino acids from the GN cytoplasmic tail (GFP GNC). Twenty-four hours later cells were fixed with methanol/acetone. A subsequent indirect immunofluorescence using the Golgi compartment marker mannosidase II was performed and analyzed via UV microscopy. The merged picture clearly demonstrates the Golgi localization of the fusion protein.
The Golgi localization signal was further analyzed with two more GFP-fusion proteins containing only the 23 amino acids from the predicted TM II directly fused to the C-terminus of GFP. To determine if a specific primary sequence within TM II was recognized as a signal or rather the hydrophobic character of this region was crucial to target GFP to the Golgi complex, the 23 amino acids were fused in two different orientations (Fig. 6A). BHK-21 (Fig. 6B) and 293T (Fig. 6C) cells were transfected with these constructs and GFP expression and intracellular localization were analyzed. Both GFP-fusion proteins showed specific Golgi complex localization demonstrating that TM II contains a Golgi localization signal and that the orientation of the primary amino acid sequence is not important for GFP translocation (Figs. 6B and 6C: k, l). GFP fusion proteins containing either the predicted GC TM (GFP GCA) or cytoplasmic domain (GFP GCB) showed perinuclear staining, suggesting ER localization (Figs. 6B and 6C: m, n).
Subsequent analyses of expressed GFP-GN fusion proteins with subcellular fractionation approaches were performed to confirm the association of the fusion proteins with cellular membranes and to demonstrate the transition of intracellular localization from a diffuse cytoplasmic to a Golgi complex region pattern (Fig. 8). For this, membrane-associated cellular proteins were separated from soluble proteins and the different fractions analyzed via immunoblot using GFP-specific antibodies. As expected, constructs GFP GNI, GFP GNA, and GFP GNB, containing only the TM I, 87 or 99 amino acids from the predicted GN cytoplasmic domain, respectively, were only detected in the soluble fraction (Fig. 8a; only shown for GNB), whereas GFP GNC and constructs with longer parts from the GN cytoplasmic domain including the TM II region were detected mainly in the pellet fraction containing membrane-associated proteins (Fig. 8b). Constructs with longer fragments of the GN cytoplasmic domain, including additional TM regions, were exclusively detected within the pellet fraction (e.g., GFP GNG; Fig. 8c). These results confirmed our previous findings that the addition of GN TM II results in a change of intracellular protein localization and seems to mediate targeting to Golgi membranes.
Figure 8 Subcellular fractionation studies of expressed GFP-CCHFV G fusion proteins. Expression plasmids for GFP-CCHFV glycoprotein fusion protein were transfected into BHK-21 cells and used for a subcellular fractionation study to determine if fusion proteins containing different parts from the CCHFV GN cytoplasmic domain are soluble or membrane associated. A GFP-specific antibody was used for the immunoblot. Representative results from three constructs are shown. M: membrane fraction, S: soluble fraction.
Discussion
Enveloped viruses, which do not acquire their lipoprotein coat by budding through the plasma membrane bud at internal membranes, including the inner nuclear membrane (herpesviruses;[43], the ER (flaviviruses and rotaviruses; [44,45], the intermediate compartment (ERGIC) (coronaviruses and poxviruses; [46,47], and the Golgi complex (rubellaviruses, coronaviruses, and bunyaviruses; [4,46,48]. Usually, the accumulation of the viral surface proteins in the specific intracellular compartment determines the assembly and budding site of the virus. This intracellular accumulation is dependent on certain compartment-specific retention or retrieval signals.
For almost all bunyaviruses assembly and budding takes place in the Golgi region [4,46,48]. However, so far no common motifs could be identified for signals within bunyaviral glycoproteins resulting in Golgi targeting and accumulation. Indeed, even the locations of such signals within bunyaviral glycoproteins are different. For the phlebovirus Uukuniemi (UUK), a Golgi retention signal could be identified within the membrane-proximal half (aa1040) of the 81 aa long cytoplasmic domain of GN [30,31,38]. In contrast, for the phlebovirus Punto Toro, such signals were mapped to the transmembrane domain (TM) and the adjacent amino acids of the GN cytoplasmic domain [36,37]. A similar localization was recently described for the Golgi retention signal in the GN of the phlebovirus Rift Valley Fever (RVF) virus GN [34]. Notable, for the Old World hantavirus Hantaan (HTN) it was reported that the conformation of the GN/GC complex might play a more important role for Golgi accumulation than an actual primary sequence motif [39].
While extensive studies have been performed regarding intracellular budding sites and glycoprotein accumulation for members of the genera Orthobunyavirus, Phlebovirus, Hantavirus and Tospovirus [30,32,34-36,39,40], nothing is known for members of the genus Nairovirus. Here we demonstrated, for the first time, that the CCHFV GN protein is membrane associated and contains a Golgi localization motif. In addition we have mapped this signal to a hydrophobic region (TM II) within the predicted cytoplasmic tail [5]. Co-expressed GN and GC result in a specific Golgi accumulation and co-localization using specific Golgi markers, whereas individual expressed GC is retained in the ER. These results imply that the two CCHFV glycoproteins have to interact and form hetero-oligomers for a proper Golgi transport of GC.
GFP-fusion proteins containing different portions of the CCHF GN glycoprotein allowed mapping of the Golgi targeting sequence within the cytoplasmic domain. Interestingly, we located the signal downstream of the predicted protease cleavage site RKLL at position 808 of the CCHFV precursor GPC, responsible for generating the C-terminus of the mature GN protein [5,28], implying that this cleavage site might not be used during the maturation process of GN. Furthermore, we could demonstrate that the addition of only the hydrophobic region from the predicted TM II within the GN cytoplasmic domain targeted a GFP-fusion protein to the Golgi complex. This shows that the 23 amino acids of TM II are sufficient and necessary for targeting GFP to the Golgi region, whereas the first 99 amino acids from the cytoplasmic domain and the TM I domain do not contribute to Golgi targeting.
The results obtained from the GFP-GN fusion proteins seem contradictory to the studies with the GN expression plasmid. IFA data combined with confocal microscopy co-localization studies of cells transfected with GNs expression plasmids demonstrated a clear Golgi complex staining (Fig. 3A: b, g; Fig. 4e). Since GNs contains only the first 87 amino acids of the predicted cytoplasmic domain without the predicted TM II sequence, we expected that the corresponding GFP-fusion protein GFP-GNA would show similar intracellular localization. However, the diffuse staining throughout the cytoplasm of transfected cells demonstrates that the first 87 amino acids are not sufficient to target the GFP to the Golgi complex (GFP-GNA; Figs. 6B and 6C: c). A possible explanation for this discrepancy is the existence of a second Golgi localization signal located within the GN ectodomain. Such a signal would be the reason for the Golgi localization pattern of GNs, whereas GFP-GNC and fusion proteins containing longer fragments of the predicted GN cytoplasmic domain localize to the Golgi region because of a Golgi localization signal located in TM II.
CCHFV GC protein expressed by its own retained in the ER and did not relocate into the Golgi complex. Interestingly, similar to all described GC proteins of phleboviruses CCHFV GC proteins also contain a lysine-based ER retrieval signal (KKXX; [49] within the predicted cytoplasmic domain. In case of single expressed GC protein this signal is most likely responsible for the ER localization of the protein, even though GFP-CCHF GCA fusion proteins containing only the predicted TM showed perinuclear staining pattern (Figs. 6B and 6C: m). However, co-expression with GN protein leads to interaction between these two proteins most likely resulting in masking of the ER retrieval signal and an accumulation of the heterodimer in the Golgi complex, due to the Golgi retention signal(s) located on GN (Fig. 3A: d, e, i, j; Fig. 4g). A similar phenomenon with conflicting transport/targeting signals was previously described for the rubellavirus E1 and E2 proteins [48,50,51].
Conclusion
In conclusion, we were able to express CCHF GN and GC glycoproteins individually as well as from the precursor GPC. GN could be localized to the Golgi compartment, whereas GC was found in the ER. Co-expression of both proteins resulted in Golgi rescue of GC, indicating that proper interaction between GN and GC is important for transportation of the heterodimer out of the ER. The potential Golgi targeting signal could be localized to a hydrophobic region within the cytoplasmic domain in the GN protein. Furthermore, our results suggest that additional signals could be localized within the GN ectodomain.
Further characterization of the CCHFV GN Golgi retention signals could provide helpful information to understand the proteolytic cleavage event(s) of the GPC and the glycoprotein maturation process. The different CCHFV G expression plasmids might show also useful for the generation of virus-like particles (VLPs) as well as for identification of interaction sites between the viral glycoproteins and the ribonucleoproteins.
The identification of the potential budding site(s) of nairoviruses and the detailed analysis of the Golgi localization signal of the CCHFV GN protein will allow subsequent studies for targeting the glycoprotein accumulation during the development of antiviral strategies or even for rational vaccine design.
Methods
Cells and virus
BHK-21 (baby hamster kidney), 293T (human embryonic kidney), VeroE6 (African green monkey kidney) and SW13 cells (human adenocarcinoma cells)(American Type Culture Collection) were grown on plastic dishes in Glasgow (BHK-21), Eagle's minimal essential (293T, VeroE6), or Leibovitz L15 (SW13) medium, respectively, supplemented with 5 to 10% fetal calf serum, 2 mM L-glutamine, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml (Invitrogen). The CCHFV, strain IbAr10200, isolated in 1970 from ticks (Hyalomma excavatum) in Nigeria (Sokoto), kindly provided by Special Pathogens Branch, Centers for Disease Control and Prevention, Atlanta (T. G. Ksiazek), was used for all experiments. The CCHFV stocks were prepared on SW13 cells by infection of T162 cell culture flasks with a 1:100 dilution. Supernatant was collected three days post infection (p.i.), clarified from cell debris by low speed centrifugation (3,000 × g, 10 min, 4°C), and aliquots were stored in liquid nitrogen. Virus titers were determined either by plaque assay or 50% tissue culture infectious dose assay (TCID50).
Sequence determination of the full-length CCHFV M segment
Total RNA was isolated 7 days post infection from VeroE6 cells infected with CCHFV (1:1000 dilution of virus stock; 10-3 pfu, RNA sample #1). Additional CCHFV RNA was kindly provided by J. Smith, USAMRIID, Alphavax, Durham, N.C. (RNA sample #2). CCHFV specific M segment vRNA or cRNA molecules were reverse transcribed using the primers CCHF M1 (TCTCAAAGAAATAGTGGCGGCACGCAGTC) or CCHF M2 (TCTCAAAGAAATACTTGCGGCACGTCAGT) for the reverse transcription reaction, respectively. The resulting cDNA molecules were used as templates for subsequent PCR reactions producing overlapping PCR fragments covering the entire CCHFV M segment. PCR products were inserted into pCR4 using the TOPO TA cloning kit (Invitrogen). Prior to sequence determination, positive clones were screened by PCR technology (primer TOPO F: AGCTCGGATCCACTAGTAACG and TOPO R: ATGCTCGAGCGGCCGCCAGTG) and restriction enzyme digest (EcoRI). For vRNA and cRNA-based constructs three of the cloning plasmids were sequenced using primers specific for the M segment ORF. The sequence results were aligned to the genebank sequence U39455 using the Align Plus 5 program of the Clone Manager Professional Suite 6 (Scientific & Educational Software). Determined nucleotide exchanges and the corresponding amino acid differences are listed in Table 1. For sequence determination of the M segment ends CCHFV specific vRNA and cRNA molecules were ligated using T4 RNA ligase (Roche) prior to the reverse transcription reaction (vRNA: M32: AGAACCAGAGGCCTGTTCAA, cRNA: M33: AAGGTGTCTGTGCCGGTTGT). Subsequent PCR amplification with primers CCHF M34 (AATACTAGTCTAAT CCACTGGCTGGTGTT) and M35 (AATGAATTCTGCCGAACTGTTCTCTAC) generated fragments containing both segment ends. PCR products were inserted into pCR4 (Invitrogen) for direct sequence determination as described above. In total, 12 cRNA/mRNA and 37 vRNA clones were analyzed using T7 and T3 promoter-specific primers.
CCHFV glycoprotein expression plasmids
Based on the recently published N-terminal sequence determination of mature CCHFV glycoproteins [5], expression plasmids for both glycoproteins were generated. In case of the CCHFV GN two constructs were generated since the C terminus of the mature GN is not yet experimentally determined: pCMV CCHF GN "short" (GNs) contains the GN-ORF from pos. 519 to 807, preceding the predicted C-terminal cleavage site RKLL at position 808 (44). pCMV CCHF GN "long" (GNl) consists of pos. 519 to 1040 extending the GN-ORF to the experimentally determined N-terminal end of GC. The PCR fragments (GNs: RF346/352, GNl: RF353/352) were inserted after BsmBI endonuclease treatment into pDisplay (Invitrogen) previously digested with BglII/XmaI digest, resulting in CMV-driven (human cytomegalovirus immediate early promoter and enhancer) expression plasmids for CCHFV GN. The Ig κ-chain signal peptide sequence and the hemagglutinin A (HA) epitope of the pDisplay vector were used for correct intracellular processing and detection, respectively. The CCHFV GC was PCR-amplified using primers CCHF M9 (AGTTGGTCTAGCCAATGTGTG) and RF351 (AATCGTCTCAAATTCATGGAGAC AGACACACTCCTGCTATGGGTACTGCTGCTCTGGGTTCCAGGTTCCACTGGTGACTTCCTAGATAGTACAGCTAAAGGCATG) (pos. 1041 to 1684).
BsmBI- and XhoI-restricted PCR fragments were inserted into the plasmid pCAGGS/MCS [52] prior digested with EcoRI/XhoI digest, resulting in a chicken β-actin-driven expression plasmid for CCHFV GC. For correct intracellular processing of the CCHFV GC we inserted the Ig κ-chain signal peptide of the pDisplay vector via forward oligonucleotide primer RF351.
Different expression strategies (CMV-, chicken β-actin-driven) were used for the different CCHFV glycoproteins to yield maximum expression levels.
Transfection
CCHFV glycoprotein expression plasmid DNA was transfected into subconfluent BHK-21 or 293T cells (3 × 106) using 2 to 4 μg of the respective plasmid and 8 μl of liposome plus buffer (LipofectAMINE PLUS; Life Technologies, Invitrogen) mixed in serum-free MEM and incubated for 15 min at room temperature. After addition of 12 μl of liposome reagent, incubation was continued for a further 15 min. The cells were incubated at 37°C with the DNA-Lipofectamine mixture for 3 h. To determine the efficiency of transfection, plasmid pHL2823, expressing enhanced GFP (EGFP) under the CMV immediate early promoter and enhancer (R. Flick and G. Hobom, unpublished), was transfected similarly. After further incubation for 20–24 h in MEM containing 2% FCS, the transfected cells were fixed and CCHFV glycoprotein expression levels determined using indirect immunofluorescence assays (IFA).
Indirect immunofluorescence assay
293T or BHK-21 cells grown on coverslips within a 6-well dish were transfected as described above. After 20 to 44 h, cycloheximide (final concentration of 0.18 mM) was added when indicated to inhibit further protein synthesis. The cells were incubated for an additional 2 to 5 h and then washed with phosphate buffer saline (PBS, pH 7.5) and fixed in methanol:acetone (50:50, V/V) for 20 min at -20°C. Permeabilization was omitted by fixation with paraformaldehyde when surface-expressed proteins were to be detected. After fixation, cells were washed with PBS and blocked for at least 30 min with PBS containing 5 % bovine serum albumin (BSA). Poly- or monoclonal antisera were diluted in PBS containing 1 % BSA and incubated for 1 h at room temperature. After several washes with PBS, goat anti-rabbit or mouse immunoglobulin secondary antibodies conjugated to fluorescein isothiocyanate (FITC) or tetramethyl rhodamin isothiocyanate (TRITC) were incubated with the cells for 45 to 60 min at room temperature. Procedures were repeated for double labeling with a different antiserum and fluorescent probe, and at the end of the procedure the slides were washed with PBS overnight.
Intracellular localization of the glycoproteins GN was determined by co-localization with commercially available organelle-specific fluorescent dyes (Molecular Probe Inc., Oregon, USA): BODIPY-TR C5 ceramide was selected as an indicator of the Golgi region. In addition Golgi (mannosidase II; Sigma) and ER-specific (Calreticulin; Sigma) monoclonal or polyclonal antibodies were used.
Confocal Microscopy
Sample preparation and immunocytochemical staining were the same as for wide-field fluorescence microscopy. The fluorescence staining patterns were analysed with a ZEISS LSM 510 UV META laser scanning confocal microscope (Jena, Germany) equipped with a Coherent Enterprise II 81 mW Argon UV laser, a Lasos 30 mW Argon laser, and 5 mW HeNe laser. Images were acquired with a C-apochromat 63/1.2 corr. water-immersion lens. FITC-stained proteins were imaged with excitation at 488 nm and with a 505 to 530 nm bandpass emission filter. Golgi marker BODIPY-TR C5 ceramide were imaged with excitation at 543 nm and with a 570 to 655 nm bandpass emission. DAPI-stained DNA was imaged with excitation at 364 nm and emission through a 385 to 470 bandpass filter. Merged pictures for analysis of intracellular co-localization were generated using Zeiss LSM Image Brower 3.2 software.
Membrane Fractionation
Alkaline carbonate extraction was performed on BHK-21 cells 24–48 h post transfection. The protocol described in Current Protocols in Cell Biology Online, John Wiley & Sons, Inc. was followed. Briefly, BHK-21 cells were transfected with individual constructs as described before. At 24 to 48 h post transfection, supernatant was removed and cells were washed three times with PBS followed by an additional washing step with 100 ml NaCl. Occasionally, the transfected cells would detach from the plate thus, the non-adherent cells were isolated between washes by microcentrifugation (2 min at 1000 × g). Remaining cells were scraped (adherent) or resuspended (non-adherent) into 1 ml of ice-cold 100 mM sodium carbonate, pH 11.5 and homogenized (five strokes) in a 2 ml Dounce homogenizer. The homogenate was then incubated for 30 min on ice and 1 ml of sodium carbonate was added to attain the necessary volume for subsequent ultracentrifugation (2 ml). The homogenate was then centrifuged for 60 min at 50.000 rpm using a TLS-55 rotor (Beckman) at 4°C. Following centrifugation, the supernatant was transferred to a fresh tube and concentrated three to five times. The pellet was resuspended in 250 μl of sodium carbonate. Pellet and supernatant fractions were then mixed with 4× SDS-PAGE sample buffer containing β-mercaptoethanol and run on SDS-PAGE. Protein gels were then transferred to PVDF transfer membrane (Amersham Biosciences) using a Trans-blot SD semi-dry transfer apparatus (Bio-Rad). Proteins were subsequently visualized by immunoblot.
Western Blot
Following transfer, the blot was blocked overnight in 5 % skim milk + 0.1 % Tween. The following morning, the blot was washed three times with PBS/0.1 % Tween then incubated with the primary antibody, e.g. anti-GFP (Oncogene) at a 1:2000 dilution in PBS for 1 h at room temperature with rocking. The blot was then washed three times with PBS/0.1 % Tween and incubated with the secondary antibody, goat anti rabbit HRP (Sigma) at a 1:30.000 dilution in PBS for 1 h at room temperature with rocking. The blot was then washed with PBS/0.1 %Tween three times, followed by three washes with PBS. Proteins were visualized using the ECL+plus Western Blotting Detection system (Amersham Biosciences).
Authors' contributions
SH carried out the described cloning work, confocal microscopy studies and the GFP-fusion protein analysis. LF carried out the membrane fractionation. TS revised the manuscript critically. HF helped to draft the manuscript and revised it critically. RF designed the study and coordinated and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to Stuart T. Nichol (Special Pathogens Branch, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, Atlanta, GA) for providing antibodies against the CCHF virus glycoproteins GC and GN. We also thank Sherif R. Zaki (Molecular Pathology and Ultrastructure Activity, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, Atlanta, GA) for providing anti-CCHFV IbAr10200 immune serum. CCHF viral stocks and viral RNA were kindly provided by Special Pathogens Branch, Centers for Disease Control and Prevention, Atlanta (T. G. Ksiazek) and by J. Smith, USAMRIID, Alphavax, Durham, N.C.
Images were captured using the laser confocal scanning microscope at the Infectious Disease and Toxicology Optical Imaging Core (OIC) facility with the assistance of Eugene Knutson, manager, and supervision of Thomas Albrecht, director, at the Department of Microbiology and Immunology, University of Texas Medical Branch (UTMB). We also like to thank Thomas Bednarek at the UTMB, Department of Pathology Imaging Laboratory, for his assistance with the illustrations.
S.H. was supported by the Boehringer Ingelheim Foundation, and performed this work in partial fulfillment of the requirements for a Ph.D. degree from the Julius-Maximilians University Wuerzburg.
==== Refs
van Regenmortel MHV Fauquet CM Bishop DML Carstens EB Estes MK Lemon SM Maniloff J Mago MA McGeoch DJ Pringle CR Wicknen RB 7th report of the International Committee of Taxonomy of Viruses Virus Taxonomy 2000 599 621
Elliott RM Schmaljohn CS Collett MS Bunyaviridae genome structure and gene expression Curr Top Microbiol Immunol 1991 169 91 141 1935231
Schmaljohn CS Le Duc JW Collier LH Bunyaviridae Topley and Wilson's Microbiology and Microbial Infections 1998 9th London , Edward Arnold 601 628
Schmaljohn CS Hooper JW Knipe DMPMH Bunyaviridae: the viruses and their replication Fields virology 2001 Philadelphia , Lippincott Williams and Wilkins 1581 1602
Sanchez AJ Vincent MJ Nichol ST Characterization of the glycoproteins of Crimean-Congo hemorrhagic fever virus J Virol 2002 76 7263 7275 12072526 10.1128/JVI.76.14.7263-7275.2002
Yang ZY Duckers HJ Sullivan NJ Sanchez A Nabel EG Nabel GJ Identification of the Ebola virus glycoprotein as the main viral determinant of vascular cell cytotoxicity and injury Nat Med 2000 6 886 889 10932225 10.1038/78645
Casals J Antigenic similarity between the virus causing Crimean hemorrhagic fever and Congo virus Proc Soc Exp Biol Med 1969 131 233 236 5770109
Chumakov MP Sokolov AACMPKAA A new tick-borne disease-Crimean hemorrhagic fever Crimean Hemorrhagic Fever (Acute Infectious Capillary Toxicosis) 1945 Simferopol, Russia , Izd. Otd. Primorskoi Armii
Chumakov MP Contribution to 30 years of investigation of Crimean haemorrhagic fever Med Virol 1974 22 5 18
Gonzalez JP Camicas JL Cornet JP Wilson ML Biological and clinical responses of west African sheep to Crimean-Congo haemorrhagic fever virus experimental infection Res Virol 1998 149 445 455 9923021 10.1016/S0923-2516(99)80013-2
Leshchinskaya EV Clinical picture of Crimean hemorrhagic fever (in Russian) (in English: NAMRU3-1856) 1965 7 226 236 (in English: NAMRU3-1856)
Oldfield EC Wallace MR Hyams KC Yousif AA Lewis DE Bourgeois AL Endemic infectious diseases of the Middle East Rev Infect Dis 1991 13 Suppl 3 S199 217 1901665
Shepherd AJ Swanepoel R Cornel AJ Mathee O Experimental studies on the replication and transmission of Crimean-Congo hemorrhagic fever virus in some African tick species Am J Trop Med Hyg 1989 40 326 331 2494900
Swanepoel R Porterfield JS Nairovirus Infections Exotic Viral Infections 1995 London
Swanepoel R Gill DE Shepherd AJ Leman PA Mynhardt JH Harvey S The clinical pathology of Crimean-Congo hemorrhagic fever Rev Infect Dis 1989 11 Suppl 4 S794 800 2749111
Watts DMTGKKJLHH Monath TP Crimean-Congo hemorrhagic fever The arboviruses: epidemiology and ecology 1988 Boca Raton, Florida , CRC Press Inc. 177 222
Yen YC Kong LX Lee L Zhang YQ Li F Cai BJ Gao SY Characteristics of Crimean-Congo hemorrhagic fever virus (Xinjiang strain) in China Am J Trop Med Hyg 1985 34 1179 1182 2422968
Hoogstral H The epidemiology of tickborne Crimean-Congo hemorrhagic fever in Asia, Europe and Africa (Review) J Med Entomol 1979 15 307 417 113533
Schwarz TF Nitschko H Jager G Nsanze H Longson M Pugh RN Abraham AK Crimean-Congo haemorrhagic fever in Oman Lancet 1995 346 1230 7475685 10.1016/S0140-6736(95)92936-3
Logan TM Linthicum KJ Bailey CL Watts DM Moulton JR Experimental transmission of Crimean-Congo hemorrhagic fever virus by Hyalomma truncatum Koch Am J Trop Med Hyg 1989 40 207 212 2493203
Sparagano OA Allsopp MT Mank RA Rijpkema SG Figueroa JV Jongejan F Molecular detection of pathogen DNA in ticks (Acari: Ixodidae): a review Exp Appl Acarol 1999 23 929 960 10737729 10.1023/A:1006313803979
Zeller HG Cornet JP Camicas JL Experimental transmission of Crimean-Congo hemorrhagic fever virus by west African wild ground-feeding birds to Hyalomma marginatum rufipes ticks Am J Trop Med Hyg 1994 50 676 681 8024058
Fagbami AH Tomori O Fabiyi A Isoun TT Experimantal Congo virus (Ib -AN 7620) infection in primates Virologie 1975 26 33 37 814708
Swanepoel R Leman PA Burt FJ Jardine J Verwoerd DJ Capua I Bruckner GK Burger WP Experimental infection of ostriches with Crimean-Congo haemorrhagic fever virus Epidemiol Infect 1998 121 427 432 9825796 10.1017/S0950268898001344
Fisher-Hoch SP McCormick JB Swanepoel R Van Middlekoop A Harvey S Kustner HG Risk of human infections with Crimean-Congo hemorrhagic fever virus in a South African rural community Am J Trop Med Hyg 1992 47 337 345 1524147
Mayers DL Exotic virus infections of military significance. Hemorrhagic fever viruses and pox virus infections Dermatol Clin 1999 17 29 40, vii-viii 9986994 10.1016/S0738-081X(98)00065-0
Vincent MJ Sanchez AJ Erickson BR Basak A Chretien M Seidah NG Nichol ST Crimean-Congo hemorrhagic fever virus glycoprotein proteolytic processing by subtilase SKI-1 J Virol 2003 77 8640 8649 12885882 10.1128/JVI.77.16.8640-8649.2003
Lenz O ter Meulen J Klenk HD Seidah NG Garten W The Lassa virus glycoprotein precursor GP-C is proteolytically processed by subtilase SKI-1/S1P Proc Natl Acad Sci U S A 2001 98 12701 12705 11606739 10.1073/pnas.221447598
Andersson AM Melin L Bean A Pettersson RF A retention signal necessary and sufficient for Golgi localization maps to the cytoplasmic tail of a Bunyaviridae (Uukuniemi virus) membrane glycoprotein J Virol 1997 71 4717 4727 9151865
Andersson AM Pettersson RF Targeting of a short peptide derived from the cytoplasmic tail of the G1 membrane glycoprotein of Uukuniemi virus (Bunyaviridae) to the Golgi complex J Virol 1998 72 9585 9596 9811692
Chen SY Compans RW Oligomerization, transport, and Golgi retention of Punta Toro virus glycoproteins J Virol 1991 65 5902 5909 1920619
Chen SY Matsuoka Y Compans RW Golgi complex localization of the Punta Toro virus G2 protein requires its association with the G1 protein Virology 1991 183 351 365 1905078 10.1016/0042-6822(91)90148-5
Gerrard SR Nichol ST Characterization of the Golgi retention motif of Rift Valley fever virus G(N) glycoprotein J Virol 2002 76 12200 12210 12414959 10.1128/JVI.76.23.12200-12210.2002
Lappin DF Nakitare GW Palfreyman JW Elliott RM Localization of Bunyamwera bunyavirus G1 glycoprotein to the Golgi requires association with G2 but not with NSm J Gen Virol 1994 75 ( Pt 12) 3441 3451 7996137
Matsuoka Y Chen SY Compans RW A signal for Golgi retention in the bunyavirus G1 glycoprotein J Biol Chem 1994 269 22565 22573 8077205
Matsuoka Y Chen SY Holland CE Compans RW Molecular determinants of Golgi retention in the Punta Toro virus G1 protein Arch Biochem Biophys 1996 336 184 189 8951050 10.1006/abbi.1996.0547
Pettersson RF Andersson A Melin L Mapping a retention signal for Golgi localization of a viral spike protein complex Cold Spring Harb Symp Quant Biol 1995 60 147 155 8824387
Shi X Lappin DF Elliott RM Mapping the Golgi targeting and retention signal of Bunyamwera virus glycoproteins J Virol 2004 78 10793 10802 15367646 10.1128/JVI.78.19.10793-10802.2004
Spiropoulou CF Goldsmith CS Shoemaker TR Peters CJ Compans RW Sin Nombre virus glycoprotein trafficking Virology 2003 308 48 63 12706089 10.1016/S0042-6822(02)00092-2
Flick R Elgh F Pettersson RF Mutational analysis of the Uukuniemi virus (Bunyaviridae family) promoter reveals two elements of functional importance J Virol 2002 76 10849 10860 12368328 10.1128/JVI.76.21.10849-10860.2002
Flick R Flick K Feldmann H Elgh F Reverse genetics for crimean-congo hemorrhagic fever virus J Virol 2003 77 5997 6006 12719591 10.1128/JVI.77.10.5997-6006.2003
Mettenleiter TC Herpesvirus assembly and egress J Virol 2002 76 1537 1547 11799148 10.1128/JVI.76.4.1537-1547.2002
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
Suzuki H A hypothesis about the mechanism of assembly of double-shelled rotavirus particles Arch Virol Suppl 1996 12 79 85 9015104
Krijnse-Locker J Ericsson M Rottier PJ Griffiths G Characterization of the budding compartment of mouse hepatitis virus: evidence that transport from the RER to the Golgi complex requires only one vesicular transport step J Cell Biol 1994 124 55 70 8294506 10.1083/jcb.124.1.55
Moss B Knipe DMPMH Poxviridae: the viruses and their replication Fields virology 2001 Philadelphia , Lippincott Williams and Wilkins 2849 2871
Hobman TC Zhao B Chan H Farquhar MG Immunoisolation and characterization of a subdomain of the endoplasmic reticulum that concentrates proteins involved in COPII vesicle biogenesis Mol Biol Cell 1998 9 1265 1278 9614173
Jackson MR Nilsson T Peterson PA Identification of a consensus motif for retention of transmembrane proteins in the endoplasmic reticulum Embo J 1990 9 3153 3162 2120038
Hobman TC Woodward L Farquhar MG The rubella virus E1 glycoprotein is arrested in a novel post-ER, pre-Golgi compartment J Cell Biol 1992 118 795 811 1500424 10.1083/jcb.118.4.795
Hobman TC Woodward L Farquhar MG The rubella virus E2 and E1 spike glycoproteins are targeted to the Golgi complex J Cell Biol 1993 121 269 281 8468347 10.1083/jcb.121.2.269
Niwa H Yamamura K Miyazaki J Efficient selection for high-expression transfectants with a novel eukaryotic vector Gene 1991 108 193 199 1660837 10.1016/0378-1119(91)90434-D
| 15850490 | PMC1090624 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 Apr 25; 2:42 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-42 | oa_comm |
==== Front
Aust New Zealand Health PolicyAustralia and New Zealand Health Policy1743-8462BioMed Central London 1743-8462-2-81584017010.1186/1743-8462-2-8ResearchRoutine measurement of outcomes in Australia's public sector mental health services Pirkis Jane [email protected] Philip [email protected] Tim [email protected] Adam [email protected] David [email protected] Rosemary [email protected] Program Evaluation Unit, School of Population Health, The University of Melbourne, Melbourne, Australia2 Queensland Centre for Mental Health Research, The University of Queensland, Brisbane, Australia3 New South Wales Institute of Psychiatry, Sydney, Australia4 Strategic Data Pty Ltd, Melbourne, Australia2005 19 4 2005 2 8 8 15 2 2005 19 4 2005 Copyright © 2005 Pirkis et al; licensee BioMed Central Ltd.2005Pirkis 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.
Objective
This paper describes the Australian experience to date with a national 'roll out' of routine outcome measurement in public sector mental health services.
Methods
Consultations were held with 123 stakeholders representing a range of roles.
Results
Australia has made an impressive start to nationally implementing routine outcome measurement in mental health services, although it still has a long way to go. All States/Territories have established data collection systems, although some are more streamlined than others. Significant numbers of clinicians and managers have been trained in the use of routine outcome measures, and thought is now being given to ongoing training strategies. Outcome measurement is now occurring 'on the ground'; all States/Territories will be reporting data for 2003–04, and a number have been doing so for several years. Having said this, there is considerable variability regarding data coverage, completeness and compliance. Some States/Territories have gone to considerable lengths to 'embed' outcome measurement in day-to-day practice. To date, reporting of outcome data has largely been limited to reports profiling individual consumers and/or aggregate reports that focus on compliance and data quality issues, although a few States/Territories have begun to turn their attention to producing aggregate reports of consumers by clinician, team or service.
Conclusion
Routine outcome measurement is possible if it is supported by a co-ordinated, strategic approach and strong leadership, and there is commitment from clinicians and managers. The Australian experience can provide lessons for other countries.
Mental healthoutcomesroutine outcome measurementAustralia
==== Body
Introduction
Internationally, there is an increasing emphasis on routine outcome measurement in mental health. A push to improve quality of care for consumers has prompted interest in monitoring outcomes at an individual level, and financial pressures and a need to demonstrate value-for-money have led to the use of aggregate reports that allow comparisons between services [1]. In the United States, there are examples of routine outcome measurement being 'rolled-out' across mental health services in entire states, such as the Ohio Mental Health Consumer Outcomes System [2]. In Europe, there are also some examples of individual services monitoring outcomes, as in the South Verona Outcomes Project [3] and the MECCA Study [4], but the routine collection of outcome data has not extended to larger areas.
Australia's commitment to routine outcome measurement is evidenced in its National Mental Health Strategy [5-7]. Since its inception in 1992, the continued improvement of the quality and effectiveness of the treatment of people with a mental illness has been a key objective of the Strategy. The Strategy has recognised that this objective can only be achieved through the development of sound information to support service planning and delivery, and consequently the systematic implementation of routine outcome measurement in all public sector mental health services is one of its priorities.
State/Territory governments and the Australian Government are collaborating in a coherent national approach. For their part, all States/Territories have signed Agreements that require them to routinely submit two sets of data from public sector mental health services to the Australian Government. Firstly, they are required to submit de-identified, patient-level outcome data, referred to as the 'National Outcomes and Casemix Collection' (NOCC) [8]. These outcome data are collected via a range of instruments that incorporate clinician and consumer perspectives on a range of mental health related constructs (e.g., symptomatology, level of functioning, degree of disability) relevant to adults, children/adolescents and older people. The idea is that administration of these instruments at specific points in time will allow services to monitor changes in individual consumers and in groups of consumers, and, ultimately, to make comparisons with similar consumers in like services. Table 1 shows the specific instruments that comprise the NOCC dataset.
Table 1 Data comprising the NOCC collection
Adults Older persons Children and adolescents
Clinician-rated Principal and additional diagnoses √ √ √
Mental health legal status √ √ √
Health of the Nation Outcomes Scale (HoNOS) [16] √
Health of the Nation Outcomes Scale 65+ (HoNOS65+) [17] √
Health of the Nation Outcomes Scale for Children and Adolescents (HoNOSCA) [18] √
Life Skills Profile 16 (LSP-16) [19, 20] √ √
Resource Utilisation Groups – Activities of Daily Living Scale (RUG-ADL) [21] √
Focus of Care [20] √ √
Children's Global Assessment Scale (CGAS) [22] √
Factors Influencing Health Status (FIHS) [20] √
Consumer-rated Mental Health Inventory (MHI) [23] or Behaviour and Symptom Identification Scale (BASIS-32) [24] or Kessler-10 Plus (K-10+) [25] √ √
Consumer- and parent-rated Strengths and Difficulties Questionnaire (SDQ)[26] √
Source: Department of Health and Ageing (2003) [27]
Secondly, States/Territories are required to submit data on inpatient episodes of care and community contacts, termed the 'National Minimum Data Set – Mental Health Care' (NMDS) [9-11]. These data provide information on resource use by consumers, and, when combined with the above outcome data will promote the development and informed use of casemix to understand the role of provider variation in differences between agencies' costs and outcomes. For the purposes of the current paper, however, the focus is routine outcome measurement only, rather than casemix development.
For its part, the Australian Government has established three Expert Groups (Adult, Child/Adolescent, and Older Persons) to advise on the implementation and use of routine outcome data in mental health services. It has also provided resources to support training in the use of outcome measures, and arrangements to receive, process, analyse and report on outcome data submitted by States/Territories.
The latter arrangements have been established through the 'Australian Mental Health Outcomes and Classification Network' (AMHOCN), a consortium contracted from late 2003 to provide national leadership in the development of outcome measurement in mental health. AMHOCN is pursuing a work program with three components, each being undertaken by a different member of the consortium: data management (Strategic Data Pty Ltd, Victoria); analysis and reporting (Queensland Centre for Mental Health Services Research, The University of Queensland, Queensland); and training and service development (New South Wales Institute of Psychiatry, New South Wales).
An immediate concern for AMHOCN was determining States/Territories' progress with respect to 'rolling out' routine outcome measurement, so each State/Territory was invited to participate in a consultation with AMHOCN. This paper reports on the findings from these consultations.
Method
The consultations occurred in March/April 2004. The intention was to seek a range of views, rather than to try to achieve a representative sample, and States/Territories were asked to nominate relevant stakeholders. They could choose whomever they wished, but they were advised to consider including policy-makers and technical personnel from central mental health units and mainstream health information sections, as well as consumers/carers. Most States/Territories sent representatives from all of these groups, and many also sent service managers, clinicians, individuals responsible for supporting routine outcome measurement at a site level, and members of the Expert Groups. In total, 123 individuals attended the consultations: 10 from New South Wales; 21 from Victoria; 28 from Queensland; 23 from Western Australia; 17 from South Australia; six from both Tasmania and the Australian Capital Territory; and 12 from the Northern Territory. Many 'wore several hats', rendering it difficult to provide a breakdown of their roles.
The consultations sought answers to questions regarding progress in four domains: (a) data collection systems and infrastructure; (b) training and retraining of staff; (c) the implementation of routine outcome measurement; and (d) and analysis, reporting and use of data.
The majority of consultations took place over a full day, with the shortest being half a day. All of the consultations began with a brief presentation from AMHOCN, and then elicited information from participants. Some States/Territories chose to split the consultation in two, inviting policy makers, planners and clinicians to attend one session and technical personnel to attend the other. Some States/Territories gave formal presentations responding to specific questions; others took a more informal approach. In some cases, the information presented at the consultations was supplemented by a written response.
Each consultation was transcribed. The transcription was combined with any other written material (e.g., formal responses and presentations), and examined at a global level to identify major themes within each domain. Individual responses were classified according to these themes. Each State/Territory was given the opportunity to comment on the accuracy of the written interpretation of the consultation.
Results
States/Territories' progress regarding data collection systems and infrastructure
Ultimately, States/Territories are aiming to have streamlined data collection systems that allow the outcome data collected via the NOCC dataset to be linked to the admitted and non-admitted activity data in the NMDS. This will allow outcome data to be 'attached' to given inpatient and community episodes of care. This has advantages in terms of allowing outcomes for consumers to be 'tracked' across episodes, and is necessary for progressing casemix development work that requires outcome data and resource use data to be combined within episodes.
All States/Territories have developed data collection systems, or are in the final stages of doing so. For some, this has involved 'starting from scratch'; for others it has required modifications to existing systems. For example, the systems used in Queensland to capture admitted and non-admitted NMDS information did not have the functionality to incorporate outcome measures, so an additional system was developed to do so. By contrast, in the Australian Capital Territory, the system used by all community teams to collect non-admitted NMDS data, was modified to collect outcome data and extended to inpatient services, where it runs alongside a separate patient administration system for the collection of inpatient activity data.
States/Territories differ in terms of the number of systems that are currently involved in the collection of routine outcome data. The simplest scenario is one where outcome measurement functionality has been added to an existing system for recording activity in community mental health settings, and has been extended into inpatient settings (as with the system in the Australian Capital Territory, described above). This also occurs in Victoria, Tasmania and the Northern Territory. Other States rely on as many as four statewide systems to collect NOCC and NMDS information, sometimes with further degrees of complexity between areas or metropolitan/country settings.
Linking NOCC and admitted and non-admitted NMDS datasets is impeded in most States/Territories by the lack of a unique identifier. Typically, linkage is only possible for parts of the data (usually NOCC and non-admitted NMDS data) and/or by conducting quite complex record linkage tasks. The exception is the Northern Territory, which has a client master index that allocates each consumer a unique identifier that allows him/her to be 'tracked' across episodes, across services, and over time. Other States/Territories are working towards improvements, but have some way to go. Western Australia's data collection system has a unique identifier that will allow episodes of care to be attributed to the same individual, regardless of location or time, but its 'roll-out' is not yet completed. Queensland and New South Wales have plans to reconcile their unique identifier systems via specific projects. This will mean that States will assign a unique identifier to a given individual that he or she will 'carry' across all health services, including mental health services, but this will not occur in the near future.
States/Territories have differing levels of infrastructure to support the NOCC and NMDS collections. Human resources vary, with some States/Territories having a number of personnel deployed to train and support clinicians and managers, and others relying on one or two core individuals. So, for example, Queensland has Zonal Outcomes Co-ordinators and Mental Health Information Support Officers providing 'on the ground' support, whereas Tasmania has a small, centrally-located team performing the same function. Physical resources also vary, with some States/Territories having sophisticated online data entry systems (e.g., the Australian Capital Territory), others relying on batch entry of paper-based forms (e.g., Tasmania), and still others using a combination of the two (e.g., New South Wales and South Australia).
States/Territories' progress regarding training and retraining of staff
All States/Territories have implemented comprehensive training programs and have trained substantial proportions of their mental health workforces in routine outcome measurement. According to stakeholders, well over 7,000 clinicians and managers across Australia have received direct training, and possibly half as many again have received training under a train-the-trainer model. This figure is consistent with that of 10,000 reported by the Department of Health and Ageing, which is estimated to represent approximately 60% the public sector mental health workforce [12].
The direct training approach is seen as having the benefit of consistency, while the train-the-trainer approach is seen as fostering capacity building and being less labour intensive and cheaper. Some States/Territories have considered accrediting trainers, so that the advantages of both approaches can be combined. Managers are also more commonly being recruited as trainers, as part of a move to secure their commitment in leading the change process. South Australia has been innovative here, building capacity by training staff as trainers through the Certificate 4 in Workplace Training and Assessment, and investing in training in content knowledge around outcome measures in the NOCC collection. In this way, South Australia has addressed some of the difficulties inherent in more standard train-the-trainer approaches.
Many States/Territories are now beginning to consider issues of ongoing training and support. High levels of staff turnover in some States/Territories mean that there are new staff who have not been trained, and lags between training and implementation in some jurisdictions have resulted in skills being lost. In addition, many States/Territories are recognising the need for a second wave of training that goes beyond how to use the outcome measures and focuses more on how to interpret the results of specific measures (at individual and aggregate levels).
Some States/Territories have implemented ongoing training strategies. Western Australia, for example, has begun refresher training. Tasmania has implemented a second round of training, focusing on the outcome measures that were not covered in the original training (i.e., the LSP-16 and the BASIS-32). Queensland has established an ongoing training program that emphasises sustainability, clinical utility and building capacity, and involves its Zonal Outcomes Co-ordinators modelling for clinicians how outcome data can be used in clinical management. Most other jurisdictions have plans in place to implement a second wave of training that focuses on the clinical and management utility of outcome measurement.
Novel, clinician-focused approaches, such as the use of vignettes and interactive case studies in Victoria and Western Australia, have underpinned the initial and ongoing training in many States/Territories. Training has also typically involved the development of resources (e.g., guides and glossaries for specific measures, consumer/carer brochures), many of which are located on individual State/Territory websites.
States/Territories' progress regarding the implementation of routine outcome measurement
States/Territories are now implementing routine outcome measurement, albeit with very variable degrees of progress. By May 2004, Victoria had provided data for 2000–01, 2001–02 and 2002–03; New South Wales for 2001–02 and 2002–03; Tasmania for 2001–02; and Western Australia, Queensland and the Northern Territory for 2002–03 (partial year only in the latter two). For South Australia and the Australian Capital Territory, 2003–04 data will constitute the first report. Within States/Territories, there is considerable patchiness in terms of coverage, compliance and completeness. There is variability by setting (with community services generally having higher coverage than inpatient services) and by outcome measure (with clinician-rated measures being completed to a greater extent than consumer-rated measures). Strong leadership at all levels has been associated with high levels of overall performance in terms of implementation.
Beyond initial training and rollout, some States/Territories have considered how to sustain and build upon current efforts with regard to routine outcome measurement. There is recognition by these (and other) States/Territories that unless routine outcome measurement becomes embedded in the process of clinical care, it will not be seen as a priority by clinicians and managers. So, for example, in New South Wales outcome measurement has been embedded in a standard protocol, which involves triage, assessment, review and discharge documentation. Specifically, a suite of clinical modules has been developed that not only includes an outcomes module but also includes the incorporation of outcome measures into the process of care. For example, the collaborative care planning module encourages collaboration between the clinician and consumer, and prompts review of the clinician-rated HoNOS and the consumer-rated K-10. The process of embedding outcome measurement within the clinical process of care is enhanced by providing clinical interpretations of given scores on particular measures. All New South Wales Area Mental Health Services will have the same modules, produced as standard medical record stationery for use within clinical files.
States/Territories' progress regarding analysis and reporting of data
Some States/Territories have also begun to consider how best to provide feedback to staff. There is recognition that without appropriate and timely feedback in the form of relevant reports that shed light on clinical and management issues, the current momentum will falter and data quality and comprehensiveness will be jeopardised. Feedback in the form of reports is required at a variety of levels. Some States/Territories have developed individual-level reports that allow clinicians to profile an individual consumer's scores on a range of outcome measures, either at a single point in time or over time. For example, in the Australian Capital Territory, the data capture system produces an electronic management plan, similar to the New South Wales module described above, which incorporates areas that the clinician and consumer might want to address, given the consumer's profile on the outcome measures. Similarly, in Western Australia, HoNOS scores of greater than 2 on Items 1 (Overactive, aggressive, disruptive or agitated behaviour) and 2 (Non-accidental self injury) trigger a risk assessment, and an alert is registered on the system.
Other States/Territories are generating aggregate-level reports about compliance. For instance, Western Australia generates Statewide compliance reports that are distributed to mental health services every six weeks, and the Office of Mental Health works with services that are experiencing difficulties with compliance to review the systems in place for monitoring the NOCC collection.
A few States/Territories have started producing some rudimentary, aggregate-level reports that provide information about groups of consumers under the care of a given clinician, team or service. Tasmania, for example, has produced monthly reports for its Southern Region, which include aggregate-level data on average HoNOS scores at admission, review and discharge. Some States/Territories have begun to consider how best to provide these reports to areas and services. New South Wales, for example, has conducted a project involving workshops in all area health services, using their own data to demonstrate the clinical and management utility of the information. A similar process has been undertaken in Queensland.
A range of factors has hampered efforts at analysis and reporting to date. These include resource issues (e.g., lack of personnel and technological constraints), data quality, a lack of clarity about which reports will have greatest clinical and management utility, and the absence of relevant normative and/or benchmarking data.
Discussion
Summary of findings
Australia has made an impressive start to nationally implementing routine outcome measurement in mental health services, although it still has a long way to go. All States/Territories have established data collection systems, although some are more streamlined than others. Significant numbers of clinicians and managers have been trained in the use of routine outcome measures, and thought is now being given to ongoing training strategies. Outcome measurement is now occurring 'on the ground'; all States/Territories will be reporting data for 2003–04, and a number have been doing so for several years. Having said this, there is considerable variability regarding data coverage, completeness and compliance. Some States/Territories have gone to considerable lengths to 'embed' outcome measurement in day-to-day practice. To date, reporting of outcome data has largely been limited to reports profiling individual consumers and/or aggregate reports that focus on compliance and data quality issues, although a few States/Territories have begun to turn their attention to producing aggregate reports of consumers by clinician, team or service.
Study limitations
Several limitations must be borne in mind in interpreting the above findings. Firstly, the study was dependent upon States/Territories selecting the most appropriate stakeholders to attend the consultations. Guidance was provided, but it is possible that States/Territories were more inclined to invite those who were in favour of routine outcome measurement, and that the views of some key stakeholders were missed. In particular, the perspective of 'coalface' clinicians was not well captured. Anecdotal reports suggest that there is some apathy, cynicism and resistance towards outcome measurement among this group. Secondly, the study relied almost exclusively on subjective reports from the stakeholders who were present at the consultations. Standard qualitative methodologies were used to record and analyse their responses, but there were few opportunities for their views to be checked against any objective measures. Finally, routine outcome measurement is moving at a considerable pace in Australia, and further progress has been made since the time of the study. The study therefore provides a conservative picture of the status quo.
Interpretation of findings
These limitations aside, some key messages emerge from the study. Specifically, it shows that routine outcome measurement is possible if it is supported by a co-ordinated, strategic approach and strong leadership. Equally important is commitment from clinicians who are involved in the day-to-day collection of the outcome data, and from managers who must make it a priority within their services. Stakeholders in the current study repeatedly stressed that this commitment will only be sustained in the long term if clinicians and managers value routine outcome measurement. Feedback to these groups in the form of reports tailored to their specific needs is crucial, and has been identified by others as necessary for maintaining momentum [2-4,13,14].
AMHOCN clearly has a role in taking routine outcome measurement to its next level. As a priority, AMHOCN is specifying a reporting framework for providing feedback to States/Territories. This involves considerations of the nature and form of data that AMHOCN itself will provide, as well as guidance to States/Territories about their own reporting. Several principles are guiding this process, in an effort to ensure that feedback has maximum clinical and management utility, and occurs as quickly as possible. Specifically, feedback should take the form of reports that are relevant and useful at a range of levels (e.g., individual, team, service and State/Territory). The precise nature of the reports should be informed by an iterative process, where relevant recipients are given the opportunity to comment on reports, and subsequent reports are modified accordingly. Reports should provide reference points that allow individual scores to be compared with normative data, and service profiles to be benchmarked against those of their peers. For now, reports will be based on NOCC data alone, in recognition of the difficulties in linking NOCC and NMDS data. This has implications for defining episodes of care, but provides scope for much to be done regarding reporting outcome data in a manner that is useful for clinicians and managers.
AMHOCN is also helping to consolidate the existing State/Territory training efforts. Specifically, it is working towards: developing and disseminating resources that fill particular gaps; helping States/Territories to streamline their training and re-training packages in a way that balances national consistency against the unique requirements of the local context; fostering the skills and knowledge required for interpreting and reflecting upon the meaning of outcome data, at a range of levels; encouraging information-sharing across the board, taking advantage of its 'birds eye view' to identify good ideas and approaches in given States/Territories and promote them in others; exploring processes for accrediting trainers, ensuring that national accreditation is consistent with and complementary to any existing accreditation efforts; and engaging, nurturing and supporting clinical leaders, champions and innovators.
Conclusion
Australia has consistently been regarded as a world leader in routine mental health outcome measurement [15]. It is acknowledged that Australia still has some way to go before routine outcome measurement is 'bedded down', and issues of data coverage, completeness and compliance are fully addressed. However, its achievements regarding national implementation are significant, and may provide lessons for other countries.
Sources of financial support
This work was funded by the Health Priorities and Suicide Prevention Branch of the Australian Government's Department of Health and Ageing.
==== Refs
Slade M The use of patient-level outcomes to inform treatment Epidemiologia e Psichiatria Sociale 2002 11 20 27 12043430
Brower LA The Ohio Mental Health Consumer Outcomes System: reflections on a major policy initiative in the US Clinical Psychology and Psychotherapy 2003 10 400 406 10.1002/cpp.386
Ruggeri M Feasibility, usefulness, limitations and perspectives of routine outcome assessment: the South Verona Outcome Project Epidemiologia e Psichiatria Sociale 2002 11 177 185 12451964
Priebe S McCabe R Bullenkamp J Hansson L Rossler W Torres-Gonzales F Wiersma D The impact of routine outcome measurement on treatment processes in community mental health care: approach and methods of the MECCA study Epidemiologia e Psichiatria Sociale 2002 11 198 204 12451967
Australian Health Ministers National Mental Health Plan 1992 Canberra, Australian Government Publishing Service
Australian Health Ministers Second National Mental Health Plan 1997 Canberra: Mental Health Branch, Commonwealth Department of Health and Family Services
Australian Health Ministers National Mental Health Plan (2003-2008) 1998 Canberra, Australian Government
Department of Health and Ageing National Outcomes and Casemix Collection: Overview of clinical measures and data items. Version 1.02 2002 Canberra, Commonwealth Department of Health and Ageing
Australian Institute of Health and Welfare Admitted patient mental health care National Minimum Data Set. National Health Data Dictionary, Version 12. AIHW Cat. No. HWI 49. 2003 Canberra, Australian Institute of Health and Welfare
Australian Institute of Health and Welfare Community mental health establishments National Minimum Data Set. National Health Data Dictionary. Version 12 2003 Canberra, Australian Institute of Health and Welfare
Australian Institute of Health and Welfare Community mental health care National Minimum Data Set. National Health Data Dictionary. Version 12 2003 Canberra, Australian Institute of Health and Welfare
Department of Health and Ageing National Mental Health Information Priorities, 2nd Edition Canberra, Australian Government Forthcoming
Callaly T Hallebone EL Introducing the routine use of outcomes measurement to mental health services Australian Health Review 2001 24 43 50 11357741
Rock D Combrinck J Groves A Issues associated with the implementation of routine outcome measures in public mental health serivces Australasian Psychiatry 2001 9 43 46 10.1046/j.1440-1665.2001.00303.x
Slade M Routine outcome assessment in mental health services Psychological Medicine 2002 32 1339 1343 12455932 10.1017/S0033291701004974
Wing JK Beevor AS Curtis RH Park SB Hadden S Burns A Health of the Nation Outcome Scales (HoNOS). Research and development British Journal of Psychiatry 1998 172 11 18 9534825
Burns A Beevor A Lelliott P Wing J Blakey A Orrell M Mulinga J Hadden S Health of the Nation Outcome Scales for elderly people (HoNOS 65+) British Journal of Psychiatry 1999 174 424 427 10616609
Gowers SG Harrington RC Whitton A Lelliott P Beevor A Wing J Jezzard R Brief scale for measuring the outcomes of emotional and behavioural disorders in children. Health of the Nation Outcome Scales for children and Adolescents (HoNOSCA) British Journal of Psychiatry 1999 174 413 416 10616607
Parker G Rosen A Emdur N Hadzi-Pavlovic D The Life Skills Profile: Psychometric properties of a measure assessing function and disability in schizophrenia Acta Psychiatrica Scandinavica 1991 83 145 152 2017913
Buckingham W Burgess P Solomon S Pirkis J Eagar K Developing
Casemix a Classification for Mental Health Services. Volume 1: Main Report 1998 Canberra, Commonwealth Department of Health and Family Services
Fries BE Schneider DP Foley WJ Gavazzi M Refining a case-mix measure for nursing homes: Resource utilization groups (RUG-III) Medical Care 1994 32 668 685 8028403
Schaffer D Gould MS Brasic J Ambrosini P Fisher P Bird H Aluwahlia S A children's global assessment scale (CGAS) Archives of General Psychiatry 1983 40 1228 1231 6639293
Veit CT Ware JE The structure of psychological distress and well-being in general populations Journal of Consulting & Clinical Psychology 1983 51 730 742 6630688 10.1037//0022-006X.51.5.730
Eisen SV Dickey B Sederer LI A self-report symptom and problem rating scale to increase inpatients' involvement in treatment. Psychiatric Services 2000 51 349 353 10686242 10.1176/appi.ps.51.3.349
Kessler RC Andrews G Colpe LJ Hiripi E Mroczek DK Normand SLT Walters EE Zaslavsky AM Short screening scales to monitor population prevalences and trends in non-specific psychological distress Psychological Medicine 2002 32 959 976 12214795 10.1017/S0033291702006074
Goodman R The Strengths and Difficulties Questionnaire: A research note Journal of Child Psychology & Psychiatry & Allied Disciplines 1997 38 581 586
Department of Health and Ageing Mental Health National Outcomes and Casemix Collection: Overview of clinician rated and consumer self-report measures, Version 1.50 2003 Canberra, Department of Health and Ageing
| 15840170 | PMC1097711 | CC BY | 2021-01-04 16:38:28 | no | Aust New Zealand Health Policy. 2005 Apr 19; 2:8 | utf-8 | Aust New Zealand Health Policy | 2,005 | 10.1186/1743-8462-2-8 | oa_comm |
==== Front
Biomed Digit LibrBiomedical Digital Libraries1742-5581BioMed Central London 1742-5581-2-31582901010.1186/1742-5581-2-3EditorialGood old days? Greenberg Charles J [email protected] Harvey Cushing/John Hay Whitney Medical Library, Yale University, 333 Cedar Street, PO Box 208014, New Haven, CT, 06850-8014, USA2005 13 4 2005 2 3 3 8 4 2005 13 4 2005 Copyright © 2005 Greenberg; licensee BioMed Central Ltd.2005Greenberg; 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.
Alternative models of subsidizing scholarly publishing and dissemination have germinated and gathered momentum in the fertile soil of dissatisfaction. Like the stubborn spring dandelion that needs but a small crack in the sidewalk to flower boldly, the first flowers of Open Access in library literature, including Biomedical Digital Libraries, have sensed their opportunity to change the existing paradigm of giving away our scholarship and intellectual property, only to buy it back for the privilege of knowing it can be read. Will biomedical digital library and informatics researchers understand their role in a new era of Open Access simply by desiring an immediate uninhibited global audience and recognizing the necessity of open access peer-reviewed literature to become self-sufficient?
==== Body
Background
The role that the publisher, librarian, and faculty will play in resolving the crisis has not been determined. Because of the potential impact of reduced collections on the mission of the health sciences center library, each group has a stake in working to contain the rise in scholarly journal prices and to stabilize the volatile journal market [1].
These words, written in 1990 before the invention of the graphical web browser or Adobe Acrobat™ [2], predicted today's journal subscription-dominated budgets, but perhaps did not anticipate the astonishing library costs of maintaining both print and electronic scholarship simultaneously. Is it any wonder that alternative models of subsidizing publishing and dissemination have germinated and gathered momentum in the fertile soil of dissatisfaction? Like the stubborn spring dandelion that needs but a small crack in the sidewalk to flower boldly, the first flowers of open access in library literature, including Biomedical Digital Libraries, have sensed their opportunity to change the paradigm: giving away our scholarship and intellectual property, only to buy it back for the privilege of knowing it can be read. Will you, the biomedical digital library and informatics researcher, understand their role in a new era of Open Access simply by desiring an immediate uninhibited global audience and recognizing the necessity of open access peer-reviewed literature to become self-sufficient?
Quality
Biomedical Digital Libraries, published by BioMed Central [3], is universally and freely available online to everyone; its authors retain copyright, and it is archived in at least one internationally recognised free repository [4]. Biomedical Digital Libraries however, has taken this further, by making all its content Open Access. In order to provide tangible benefits – no page or digital content limits or colour charges, unlimited image and data storage, preparation and inclusion in redundant global digital repositories – from 01 April, 2005, authors of articles accepted for publication in Biomedical Digital Libraries will be asked to pay an article-processing charge (APC) of £330 (€480, US$615). Authors from BioMed Central member institutions will not incur any APC. Many notable funding agencies authorize and encourage use of their grant resources to pay an APC [5]. The Editor-in-Chief will consider a limited number APC waiver requests per year on a case-by-case basis. Priorities for waivers include authors in countries served by the Health InterNetwork (HINARI)[6] or individual circumstance.
Creating a New Tradition
Librarians and other information professionals are no doubt more familiar than most with the traditional access model: readers pay to view articles, either through library or personal subscriptions or by paying a fee each time they download an article. Escalating journal subscription costs have resulted in libraries subscribing to fewer journals [7]. Although scholarly journals traditionally publish authors' work for free (unless there are page or colour charges!) in exchange for obtaining permanent and exclusive rights to publish articles, subscriptions limit who and how many can immediately read, use and cite the research. However we feel about the level of profits from scholarly publishing, no one can deny that there are unavoidable costs in the quality production and timely distribution of peer-reviewed biomedical library research.
A Time to Change
Biomedical Digital Libraries' Open Access policy changes the way in which peer-reviewed library research is published: following blinded peer review, all accepted articles are accessible online without the barrier of subscription or cost, and the author(s) retain copyright for their work and grant anyone the right to reproduce and disseminate the article [8]. Shortly thereafter, a copy of the full text of each article is permanently archived in an online repository separate from the journal. Biomedical Digital Libraries' articles are archived in PubMed Central [4], the US National Library of Medicine's full-text repository of life science literature, and also in repositories at the University of Potsdam [9] in Germany, at INIST [10] in France and in e-Depot [11], the National Library of the Netherlands' digital archive.
APCs will support the future publication of biomedical library and informatics articles with conscious attention to peer review and immediate impact on professional practice. open access to all of Biomedical Digital Libraries' articles. Although some authors may consider an APC of £330 (€480, US$615) to be an economic hardship, it must be remembered that Biomedical Digital Libraries does not levy additional processing charges on top of this fee. With the article being digital from birth, any amount of colour figures, photographs, and digitized video recording can be included with submissions.
Open or Free?
Many libraries under increasing budgetary constraints feel indebted to websites like Free Medical Journals [13] which offers immediate assessment of which journals are releasing free content after a subscription embargo. Although a growing number of journals now offer delayed free access to their articles online, this is different from Open Access (as defined by the Bethesda Statement [14]). Journals with subscription revenue often delay free access for 6–12 months, and even when the full text is available, readers are not allowed to reproduce and/or disseminate the work because of restrictions imposed by the copyright policy. Biomedical Digital Libraries is not alone in the move to Open Access funded by APCs: the Public Library of Science (PLOS) has set up new Open Access journals, and have elected to set APCs of US$1500 for each accepted article, with institutional membership discounts available [15]. The high profile of PLOS in mainstream public media will raise awareness of Open Access and encourage researchers in all disciplines to understand and accept Open Access, with APCs as an acceptable revenue stream.
It's your [copy] right!
By providing an unimpeded forum for authors that wish to maintain their intellectual property, APCs will enable Biomedical Digital Librariesto serve the individual content creator in the global library and informatics communities. We believe electronic publishing formats will benefit library and informatics practice and aid information science research, and we hope you will submit your next article to our protocol for blind peer review, which has without a doubt improved the organization and presentation of every submission so far.
Declare Your Interest
More than 40 volunteer editorial staff members of Biomedical Digital Libraries solicit their peers and colleagues for evidence-based, quality research. In our editorial model, both authors and editorial reviewers declare their competing interests, especially any related to the APC. If you would like advice on what to declare, and how to declare it, please feel free to ask us prior to submission. Many authors have some form of competing interest, conscious or not. A declaration of a competing interest reminds the reader that an awareness of potential influence can help to distinguish an appearance of a relationship from actual manipulative intent.
Competing interests
The Editor declares his affiliation with one of 519 BioMed Central institutional members in 40 countries that provide a reasonable alternative to direct individual APC revenue.
Abbreviations
APC article-processing charge
HINARI Health InterNetwork
INIST INstitut de l'Information Scientifique et Technique (Institute for Scientific and Technical Information)
PLOS Public Library of Science
==== Refs
Hafner AW Podsadecki TJ Whitely WP Journal pricing issues: an economic perspective Bull Med Libr Assoc 1990 78 217 223 2203496
Adobe's Sarah Rosenbaum: From Acrobat 1.0 to Acrobat 6.0
BioMed Central
PubMed Central
Which funding agencies explicitly allow direct use of their grants to cover article processing charges?
Health InterNetwork (HINARI)
Figure: Current serial titles data boxplots with medians, ranges, and first-to-third quartile ranges for selected editions, 1977/78 to 2000/01 J Med Libr Assoc 2003 91 186 202 12883578
Creative Commons Attribution 2.0
Potsdam
INIST
e-Depot
BioMed Central Institutional Members
Free Medical Journals
Bethesda Statement on Open Access Publishing
Public Library of Science to launch new free-access biomedical journals with $9 million grant from the Gordon and Betty Moore Foundation
| 15829010 | PMC1097712 | CC BY | 2021-01-04 16:38:25 | no | Biomed Digit Libr. 2005 Apr 13; 2:3 | utf-8 | Biomed Digit Libr | 2,005 | 10.1186/1742-5581-2-3 | oa_comm |
==== Front
BMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 1472-6882-5-91581119110.1186/1472-6882-5-9Research ArticleAnti-obesity effects of chikusetsusaponins isolated from Panax japonicus rhizomes Han Li-Kun [email protected] Yi-Nan [email protected] Masayuki [email protected] Hiromichi [email protected] Yoshiyuki [email protected] Faculty of Environmental and Symbiotic Sciences, Prefectural University of Kumamoto, Tsukide, Kumamoto 862-8502, Japan2 Department of Chinese Material Medicine, Chinese Material Medicine College of Jilin Agricultural University, Changchun-City, Jilin 130118, China3 Department of Pharmacognosy, Kyoto Pharmaceutical University, Yamashina, Kyoto 607-8412, Japan4 Division of Biochemistry, Department of Molecular and Cellular Biology, School of Medicine, Ehime University, Shitsukawa, Toon-City, Ehime 791-0295, Japan2005 6 4 2005 5 9 9 8 1 2005 6 4 2005 Copyright © 2005 Han et al; licensee BioMed Central Ltd.2005Han 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 rhizomes of Panax japonicus are used as a folk medicine for treatment of life-style related diseases such as arteriosclerosis, hyperlipidemia, hypertension and non-insulin-dependent diabetes mellitus as a substitute for ginseng roots in China and Japan. Obesity is closely associated with life-style-related diseases. This study was performed to clarify whether chikusetsusaponins prevent obesity induced in mice by a high-fat diet for 9 weeks.
Methods
We performed two in vivo experiments. In one, female ICR mice were fed a high-fat diet with or without 1 or 3% chikusetsusaponins isolated from P. japonicus rhizomes for 9 weeks. In the other, lipid emulsion with or without chikusetsusaponins was administered orally to male Wistar rats, and then the plasma triacylglycerol level was measured 0.5 to 5 h after the orally administered lipid emulsion. For in vitro experiments, the inhibitory effects of total chikusetsusaponins and various purified chikusetsusaponins on pancreatic lipase activity were determined by measuring the rate of release of oleic acid from triolein in an assay system using triolein emulsified with lecithin.
Results
Total chikusetsusaponins prevented the increases in body weight and parametrial adipose tissue weight induced by a high-fat diet. Furthermore, consumption of a high-fat diet containing 1 or 3% total chikusetsusaponins significantly increased the fecal content and triacylglycerol level at day 3 compared with the high-fat diet groups. Total chikusetsusaponins inhibited the elevation of the plasma triacylglycerol level 2 h after the oral administration of the lipid emulsion. Total chikusetsusaponins, chikusetsusaponin III, 28-deglucosyl-chikusetsusaponin IV and 28-deglucosyl-chikusetsusaponin V inhibited the pancreatic lipase activity.
Conclusion
The anti-obesity effects of chikusetsusaponins isolated from P. japonicus rhizomes in mice fed a high-fat diet may be partly mediated through delaying the intestinal absorption of dietary fat by inhibiting pancreatic lipase activity. The present study clearly indicated that the saponin fractions of P. japonicus rhizomes had a significant anti-obesity action and supports the traditional usage as a substitute drug for ginseng roots.
==== Body
Background
The rhizomes of Panax japonicus C.A. Meyer (Japanese name; Chikusetsuninjin), have been used as a substitute for Ginseng roots (the roots of Panax ginseng C.A. Meyer). On the other hand, Ginseng roots are used as the remedy for life-style-related diseases such as arteriosclerosis hyperlipidemia, hypertension and non-insulin-dependent diabetes mellitus in China, Korea, Japan and Europe and there are a number of reports on the pharmacological studies of Ginseng roots [1-8]. It has been reported that the rhizomes of Panax japonicus have anti-ulcer action and fibrinolysis [9,10]. However, the investigations of pharmacological effects of Panax japonicus rhizomes on life-style-related diseases such as obesity, arteriosclerosis hyperlipidemia, hypertension and non-insulin-dependent diabetes mellitus have been not thoroughly reported. Obesity is one of the fastest-growing major diseases in many areas of the world including Europe, the United States and Japan. Obesity results from an imbalance between energy intake and expenditure. Obesity is closely associated with life-style-related diseases such as hyperlipidemia, hypertension, arteriosclerosis and non-insulin-dependent diabetes mellitus and with increased risk of coronary heart disease [11]. It has been reported that variations in total energy intake and diet composition are important in the regulation of metabolic processes [12,13]. Furthermore, it has been suggested that dietary fat promotes body fat storage more effectively than dietary carbohydrate. Thus, inhibition of the digestion and absorption of dietary fat is a key to treating obesity. Dietary fat is not directly absorbed from the small intestine unless it has been subjected to the action of pancreatic lipase [14]. In this study, we examined the effects of total chikusetsusaponins of P. japonicus rhizomes on obesity induced by long-term feeding of a high-fat diet.
Methods
Materials
Pancreatic lipase was purchased from Sigma Chemical Co. (St Louis, MO). Triglyceride E- and Total Cholesterol E-test kits were purchased from Wako Pure Chemical Co. (Osaka, Japan). Laboratory pellet chow was purchased from CLEA Japan (Osaka, Japan). Beef tallow, casein, vitamin and mineral mixtures were purchased from Oriental Yeast Co. Ltd (Tokyo, Japan). Orlistat (a lipase inhibitor) was purchased from Hong Kong Market. Other chemicals were of reagent grade.
Plant materials
The rhizomes of P. japonicus were obtained from Tochimoto Tennkaido Co. Ltd. (Osaka, Japan). Voucher specimens (No. PJ020817) are deposited in the Faculty of Environmental and Symbiotic Sciences, Prefectural University of Kumamoto and the Second Department of Medical Biochemistry, School of Medicine, Ehime University.
Preparation of total saponins and five chikusetsusaponins from the rhizomes of P. japonicus
Chikusetsusaponins were isolated using by a modification of previous reported method [15,16]. The dried rhizomes of P. japonicus (1 kg) were extracted three times by refluxing with methanol (3 L). After removal of the solvent from the methanol solution under reduced pressure, the extract (370 g) was subjected to chromatography on reversed-phase highly porous polymer [DIAION HP-20 (4 kg), Mistubishi Chemical Ind. Ltd.: elution with H2O (800 mL), MeOH (500 mL), and CHCl3 (300 mL)] to provide the three fractions such as H2O eluate (235 g), MeOH eluate (130 g), and CHCl3 eluate (5 g). The fraction eluted with methanol was designated the "total chikusetsusaponins". The MeOH eluate (total saponin fraction, 110 g) was separated by column chromatography on silica gel [silica gel G (2 kg), Merck, CHCl3-MeOH-H2O] to furnish four fractions and the fractions 2 and 3 were purified by HPLC (YMC-pack ODS, MeOH-H2O-AcOH) to give chikusetsusaponins III (13.5 g), IV (6.2 g) and V (54.5 g), 28-deglucosylchikusetsusaponins IV (15 mg) and V (110 mg). 28-Deglucosylchikusetsusaponins IV and V were also obtained by alkaline hydrolysis of chikusetsusaponins IV and V, respectively. The isolated five compounds were identified by comparison of their physical data ([α]D, IR, 1H- and 13C-NMR) with reported values [15,16]. The purity of each compound was over 95% based on the analysis by HPLC (Fig. 1).
Figure 1 The structure of various chikusetsusaponins
Diet compositions
Previously, we reported that varying the casein concentration (22, 31, 34 and 36 % diets) in high-fat diet containing 40 % beef tallow did not affect body weight or parametrial adipose tissue weight [17]. Therefore, we substituted 1 or 3% total chikusetsusaponins isolated from P. japonicus rhizomes for an equivalent mass of casein in the high-fat diets shown in Table 1. To avoid autooxidation of the fat components, food was stored at -30°C. If the diet except fat from high-fat diet was used as control diet, carbohydrate (e.g. corn starch or sugar) instead of fat was added to the diet; and consequently this diet is high-carbohydrate diet. Black et al. reported that feeding high-carbohydrate diet containing 73% kcal corn starch for 16 weeks elevated the α-disaccharidase activity in small intestine [18]. Therefore, it seems likely that the high-carbohydrate diet may affect carbohydrate metabolism. In this study, we used laboratory chow diet as a control diet.
Table 1 Composition of experimental high-fat diets
Total chikusetsusaponins (g/100 g diet)
High-fat diet (HF) HF + 1% chikusetusaponins HF + 3% chikusetsusaponins HF + 0.012% orlistat HF + 0.024% orlistat
Beef tallow 40 40 40 40 40
Corn starch 10 10 10 10 10
Sugar 9 9 9 9 9
Mineral mixture 4 4 4 4 4
Vitamin mixture 1 1 1 1 1
Casein 36 35 33 36 36
Total saponins 0 1 3 0 0
Orlistat 0 0 0 0.012 0.024
Animals
Female ICR strain mice (3 weeks old) and male Wistar King strain rats (6 weeks old) were obtained from CLEA Japan (Osaka, Japan) and Charles River Japan (Yokohama, Japan), respectively, and housed for 1 week under a 12 h/12 h light/dark cycle in a temperature- and humidity-controlled room.
The animals were given free access to food and water. After adaptation to the lighting conditions for 1 week, the healthy animals were used in these experiments. The experimental protocol was approved by the Animal Studies Committees of Ehime University and Kumamoto Prefectural University.
Estimation of body and parametrial adipose tissue weights, plasma triacylglycerol and total cholesterol in mice fed a high-fat diet for 9 weeks
Female ICR mice (3 weeks old) were divided into four groups that were matched for body weight, after 1 week of being fed laboratory pellet chow ad libitum. The control group continued to be fed laboratory pellet chow ad libitum. The mice consumed the high-fat diet or the high-fat diet containing 1 and 3% total chikusetsusaponins or 0.012 and 0.024% orlistat (positive control) for 9 weeks. The body weight of each mouse was measured once a week and the total amount food consumed was recorded 3 times per week. After the mice had been fed these diets for 9 weeks, blood was taken from each mouse by venous puncture under anesthesia with diethyl ether; the mice were then killed with an overdose of diethyl ether. Experiments were performed in a ventilated room. The plasma was prepared and frozen at -80°C until analysis. The liver and parametrial adipose tissue were dissected and weighed. Liver triacylglycerol and total cholesterol concentrations were measured using Wako Triglyceride E-Test and Total Cholesterol E-Test kits.
Fat excretion in feces of mice
Female ICR mice (3 weeks old) were housed for 1 week in a room maintained at 25 ± 1°C with 60% relative humidity and given free access to standard laboratory pellet chow and water. The mice consumed the high-fat diet or the high-fat diet containing 1% or 3% total chikusetsusaponins, or 0.025% orlistat for 3 days. Wet weight and triacylglycerol content in the feces obtained during the last 24 h were measured using the Wako Triglyceride E-Test kit.
Plasma triacylglycerol levels after oral administration of lipid emulsions to rats
Male Wistar rats were also housed for 1 week in the above same conditions. After rats had been deprived of food overnight, they were orally administered 1 mL of a lipid emulsion consisting of corn oil (3 mL), cholic acid (40 mg) and cholesteryl oleate (1 g) plus physiological saline (3 mL), the lipid emulsion (1 mL) plus total chikusetsusaponins (final concentration 1000 mg/kg body) or the lipid emulsion plus orlistat (final concentration 45 mg/kg body). Blood samples were taken from the tail vein 0, 0.5, 1, 2, 3, 4 and 5 h after administration of the lipid emulsion with or without total saponins or orlistat using a capillary tube (heparinized), and centrifuged at 5500 × g for 5 min in a Model KH-120 M (Kubota Co., Osaka, Japan) centrifuge to obtain the plasma. The plasma triacylglycerol concentration was determined using a Wako Triglyceride E-Test kit.
In vitro pancreatic lipase activity
Lipase activity in the porcine pancreas was assay as described previously [19]. Enzyme activity was expressed as μmoles of oleic acid released per mL of reaction mixture per min.
Statistical analysis
All values are expressed as means ± s.e. Data were analyzed by one-way ANOVA, and then differences among means were analyzed using the Fisher's protected LSD test. Differences were considered significant at P < 0.05.
Results
Fat excretion in feces of mice fed a high-fat diet with or without total chikusetsusaponins
Consumption of the high fat diet for 3 days reduced the feces weight (0.14 ± 0.035 g/mouse/day) at day 3 compared with that of the control group (1.04 ± 0.017 g /mouse/day). Mice fed the high-fat diet plus 3% total chikusetsusaponins, or 0.025% orlistat for 3 days had significantly higher triacylglycerol level at day 3 compared with the high-fat diet group (Table 2).
Table 2 Effects of total chikusetsusaponins and orlistat (a lipase inhibitor) on fat excretion on day 3 in feces of mice fed a high-fat diet
Animal No. Feces weight (g) Triacyglycrol in feces (μmol/g feces)
Laboratory chow pellet 13 1.04 ± 0.017* 5.05 ± 0.70*
High-fat diet (HF) 13 0.14 ± 0.035 36.66 ± 1.03
HF +
1% Total chikusetsusaponins 7 0.26 ± 0.012* 41.64 ± 3.48
HF +
3% Total chikusetsusaponins 7 0.31 ± 0.049 53.42 ± 4.86*
HF +
0.025 % orlistat (lipase inhibitor) 7 0.22 ± 0.040 394.38 ± 65.39*
Values are means ± S.E. of 7 – 13 mice. * Significantly different from the high-fat diet group, P < 0.05.
Food consumption; body, parametrial adipose tissue and liver weights; and triacylglycerol content in the livers of mice fed high-fat diet with or without chikusetsusaponins for 9 weeks
The mean food consumption per week per mouse was significantly different between the control group and high-fat diet groups, being 409.4 ±7.0 KJ in the control group and 611.0 ± 41.6 KJ in the high-fat diet group?? but was not significantly different among the high-fat diet groups, high-fat plus 1% and 3% saponin-diet groups, and high-fat plus 0.012% orlistat-diet groups, being 611.0 ± 41.6 KJ (high-fat diet), 708.7 ± 34.6 KJ (high-fat diet plus 1% chikusetsusaponin), 686.0 ± 30.0 KJ (high-fat diet plus 3% chikusetsusaponin) and 693.1 ± 53.8 KJ (high-fat diet plus 0.012% orlistat). A high-fat plus 0.024% orlistat diet significantly reduced the body weight to 4 weeks compared to both the normal and the high-fat diet groups, but among 15 mice of high-fat plus 0.024% orlistat-diet groups, 10 mice died by the severe diarrhea until 5 weeks by the feeding high-fat diet containing 0.024% orlistat (data not shown). Orlistat is a potent inhibitor of pancreatic lipase; therefore, the absorption of dietary fat from small intestine was strongly inhibited by feeding high-fat diet containing 0.025% orlistat (Table 2). Mice treated with high-fat diet plus 0.024% orlistat for a long term died by severe diarrhea, but high-fat diet plus 0.012% orlistat-fed mice did not cause diarrhea (data not shown). This finding suggests that feeding level of 0.024% orlistat in the high-fat diet to mice for a long term may cause the physiological toxicity through the continuous inhibition of pancreatic lipase. Feeding high-fat plus 0.012% orlistat diet had no effect on the adipose tissue weight. In the present study, the discrepancy between anti-obesity action and inhibitory action of pancreatic lipase by orlistat could not sufficiently be clarified; therefore, this discrepancy is needed to clarify in future studies. Figure 2 shows the changes in body weight of the groups during the experiment. Consuming a high-fat diet for 9 weeks caused significant increases in body weight at 1 to 9 weeks compared to that of the control group (laboratory pellet chow). Consuming a high-fat diet supplemented with 1% or 3% chikusetsusaponins significantly suppressed the increase in body weight during the experimental period compared to the high-fat diet, but the high-fat diet plus 0.012% orlistat did not affect. The final parametrial adipose tissue weights of the groups are shown in Table 3. The final parametrial adipose tissue weight was significantly increased by consumption of a high-fat diet compared to the control diet, and that in animals with a high-fat diet containing 1% or 3% chikusetsusaponins was significantly reduced as compared to that in the high-fat diet group. Furthermore, mice fed a high-fat diet long term developed fatty liver, with increases of the liver weight and an accumulation of hepatic triacylyglycerol compared to the control group. Consumption of a high-fat diet containing 1% or 3% chikusetsusaponins, or 0.012% orlistat reduced the liver weight and hepatic triacylglycerol content compared to consumption of a high-fat diet group (Table 3).
Table 3 Effects of total chikusetsusaponins and orlistat (a lipase inhibitor) on liver weight, hepatic triacylglycerol, and adipose tissue weight in mice fed a high-fat diet for 9 weeks.
Animal No. Liver weight (g/100 g body weight) Hepatic triacylglycerol (μmol/g) Parametrial adipose tissue weight (g)
Laboratory chow pellet 26 5.20 ± 0.46* 22.3 ± 2.3* 0.71 ± 0.1*
High-fat diet (HF) 26 6.35 ± 0.76 141.8 ± 21.8 1.59 ± 0.3
HF +
1% Total saponins 11 6.13 ± 0.79 101.5 ± 6.8* 0.87 ± 0.1*
HF +
3% Total saponin 11 5.79 ± 0.47* 83.3 ± 6.7* 0.74 ± 0.1*
HF + 0.012% Orlistat 15 4.00 ± 0.11* 101.1 ± 4.8* 2.03 ± 0.12
Values are means ± S.E. of 11 – 26 mice. *Significantly different from the high-fat diet group, P < 0.05.
Figure 2 Effects of chikusetsusaponins and orlistat (a lipase inhibitor) on body weight in mice fed a high-fat diet for 9 weeks. Open circles, normal groups; solid circles, high-fat diet groups; open squares, high-fat plus 1% total chikusetsusaponin diet groups; solid squares, high-fat plus 3% total chikusetsusaponin diet groups; solid triangle, high-fat plus 0.012% orlistat diet. Values are means ± S.E. of 11–26 mice. * Significantly different from the high-fat diet group, P < 0.05.
Total chikusetsusaponins reduced the elevation of rat plasma triacylglycerol levels after oral administration of lipid emulsion to rats
The plasma triacylglycerol level (2.98 ± 0.31 mM, n = 4) at 2 h after orally administered lipid-emulsion was significantly reduced to 1.66 ± 0.28 mM (n = 4) and 0.93 ± 0.16 mM (n = 4), respectively, by the oral administration of total chikusetsusaponins (1000 mg/kg) and orlistat (45 mg/kg) (Fig. 3).
Figure 3 Effects of total chikusetsusaponins and orlistat (a lipase inhibitor) on rat plasma triacylglycerol levels after oral administration of a lipid emulsion. Open circles, lipid emulsion; solid circles, lipid emulsion plus total chikusetsusaponins (1000 mg/kg) and open squares, lipid emulsion plus orlistat (a lipase inhibitor) (45 mg/kg). Values are means ± S.E. of 4 rats. * Significantly different from lipid-emulsion-only treated group, P < 0.05.
Effects of total chikusetsusaponins and various purified chikusetsusaponins on pancreatic lipase activity in vitro
As shown in Fig. 4a, methanol extracts, total chikusetsusaponins and the positive control orlistat dose dependently inhibited pancreatic lipase activity in an assay system using triolein emulsified with lecithin, but water and CHClS3-eluate fraction had no effect. We then examined the various chikusetsusaponins isolated from P. japonicus rhizomes. Chikusetsusaponin III and 28-deglucosylchikusetsusaponins IV and V dose-dependently inhibited the pancreatic lipase activity at concentrations of 125 to 500 μg/mL (Fig. 4b).
Figure 4 Effects of total chikusetsusaponins, various chikusetsusaponins and orlistat (a lipase inhibitor) on pancreatic lipase activity. Values are means ± S.E. of 4 experiments. a) Solid circles, methanol extract; open circles, total chikusetsusaponins; open squares, orlistat (a lipase inhibitor) was used as a positive control. b) Open circles, chikusetsusaponin III; solid circles, chikusetsusaponin IV; open squares, 28-deglucosyl-chikusetsusaponin IV; solid squares, 28-deglucosyl-chikusetsusaponin V.
Discussion
There are a number of reports showing that ginseng roots are clinically anti-allergic, anti-hypertensive, hypoglycemic, anti-thrombotic, and anti-arteriosclerotic [7,8]. Recently, obesity is increasing in advanced countries including Europe, the United States and Japan. Obesity is closely associated with life-style-related diseases such as hyperlipidemia, hypertension, arteriosclerosis and non-insulin-dependent diabetes mellitus and with increased risk of coronary heart disease [11]. Although the rhizomes of P. japonicus are used as a substitute drug for ginseng roots in Japan, the pharmacological effects of these rhizomes on life-style-related diseases have been not clarified. A methanol extract of these rhizomes and chikusetsusaponins such as chikusetsusaponins III, IV and V derived from P. japonicus rhizomes have been shown to promote the fibrinolysis [10]. Yamahara et al. [9] reported that the saponin fractions and chikusetsusaponin III of P. japonicus exert anti-ulcer effects. However, there have hitherto been no reports on the inhibitory effects of chikusetsusaponins isolated from P. japonicus rhizomes on obesity induced by consumption of a high-fat diet. Dietary fat can increase body weight and adiposity in humans and animals more effectively than dietary carbohydrate [17,20-22]. Dietary fat is not directly absorbed from the intestine unless it has been subjected to the action of pancreatic lipase [14]. Therefore, the application of pancreatic lipase inhibitor was examined earlier as a treatment for high-fat diet-induced obesity in humans. It has been reported that a pancreatic lipase inhibitor, orlistat prevented obesity and hyperlipidemia through the enhancement of fat excretion in feces and the inhibition of pancreatic lipase [23]. We found that feeding high-fat diet containing 40% fat content caused obesity [17,19], but the high-fat diet containing 25% fat content slightly increased the body weight (data not shown). Anai et al. reported that feeding high-fat diet containing 60% fat for 2 weeks slightly increased the body weight [24]. Furthermore, it has been reported that feeding high-fat diet containing 58% kcal fat for 16 weeks caused obesity [18]. In the present study, to examine the effects of chikusetsusaponins on high-fat diet-induced obesity, we used the obesity model induced by feeding high-fat diet containing 40% fat for 9 weeks. We found that the administration of total chikusetsusaponins isolated from P. japonicus significantly suppressed the increase in body weight in mice fed a high-fat diet containing 40% beef tallow for 9 weeks. The treatment with total chikusetsusaponins also significantly reduced the final parametrial adipose tissue weight compared to that of the high-fat diet group. These inhibitions did not depend on decreased food or energy intake, because there was no significant difference between these in the groups fed the high-fat diet and the high-fat diet containing 1 or 3 % total chikusetsusaponin. Next, we examined the effects of total chikusetsusaponins on plasma triacylglycerol concentrations after oral administration of a lipid emulsion in rats, and found that the total chikusetsusaponins reduced the elevation of plasma triacylglycerol levels as measured by an oral lipid emulsion tolerance test. This finding suggests that total chikusetsusaponins may inhibit the uptake of dietary fat. In fact, total chikusetsusaponins strongly inhibited the pancreatic lipase activity, and therefore, we attempted to isolate substance(s) from total saponins that inhibit pancreatic lipase activity. Five chikusetsusaponins were isolated from total saponin fractions and identified as chikusetsusaponins III, IV and V and 28-deglucosyl-chikusetsusaponins IV and IV by the direct comparison with authentic samples. Among these five saponins, chikusetsusaponin III and 28-deglucosylchikusetsusaponins IV and V inhibited the pancreatic lipase activity, most strongly. The contents of total chikusetsusaponins in the MeOH extract are about 35%, and the contents of chikusetsusaponins III, IV and V in the total saponin fraction are about 10.4, 4.8 and 41.9%, respectively. On the other hand, the content of 28-deglucosylchikusetsusaponins IV and V in the total saponins are about 0.01 and 0.08%, respectively. Chikusetsusaponins IV and V are rapidly converted to 28-deglycosyl-forms by the treatment with alkaline solution. Therefore, it seems likely that the above 28-deglucosylchikusetsusaponin derive from chikusetsusaponins IV and V through the small intestine after orally administered chikusetsusaponins IV and V. Consequently, it is suggested that their metabolites (28-deglucosyl form) exhibit the inhibition of pancreatic lipase activity.
Conclusion
Total chikusetsusaponins isolated from P. japonicus may prevent high-fat-diet-induced increases in body weight and fat storage in adipose tissue by inhibiting intestinal absorption of dietary fat through the inhibition of pancreatic lipase activity, and the active components were identified here as chikusetsusaponins III and IV, 28-deglucosyl-chikusetsusaponins IV and V. The present study clearly indicated that the saponin fractions of P. japonicus rhizomes had a significant anti-obesity action and supports the traditional usage as a substitute drug for ginseng roots. Hence it might help in preventing obesity complications and serve as good adjuvant in the present armamentarium of anti-obesity drugs.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
The experimental design was made by L-K. Han, H. Okuda and Y. Kimura
The experiments of animal models and enzyme activity were worked by L-K. Han, H. Okuda and Y. Kimura. The isolation and structural determination of various chikusetsusaponins from Panax japonicus rhizomes were worked by L-K. Han, Y-N. Zheng and M. Yoshikawa.
The paper was written by L-K. Han and Y. Kimura.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Shibata S Tanaka O Shoji J Saito H Wagner H, Hikino H, Farnsworth NR Chemistry and pharmacology of Panax Economic and Medical Plant Research 1985 1 London, Orland, San-Diego, New York, Toronto, Montreal, Sydney, Tokyo, Academic Press 217 287
Yamamoto M Ha H, Yamamoto M, Joo CN, Kuo S-C Current status Ginseng research in Japan Proceedings of the International Symposium on Ginseng Research Taichuu; Life Medical Research Institute, China Medical College 12 26
Nocerino E Amato M Izzo AA The aphrodisiac and adaptogenic properties of ginseng Fitoterapia 2000 71 51 55
Inoue M Wu CZ Dou DQ Chen YJ Ogihara Y Lipoprotein lipase activity by red ginseng saponins in hyperlipidemia model animals Phytomedicine 1999 6 257 265 10589445
Chung SH Choi CG Park SH Comparisons between white ginseng radix and rootlet for antidiabetic activity and mechanism in KKAy mice Arch Pharm Res 2001 2 214 218 11440080
Kiefer D Pantuso T Panax ginseng Am Fam Physician 2003 68 1539 1542 14596440
Han KH Choe SC Kim HS Sohn DW Nam KY Oh BH Lee MM Park YB Choi YS Seo JD Lee YW Effects of red ginseng on blood pressure in patients with essential hypertension and white coat hypertension Am J Chin Med 1998 36 199 209 9799972
Sotaniemi EA Haapakoski E Rautio A Ginseng therapy in non-insulin-dependent diabetic patients Diabetes Care 1995 18 1373 1375 8721940
Yamahara J Kubomura Y Miki K Fujimura H Anti-ulcer action of Panax japonicus rhizome J Ethnopharmacol 1987 19 95 101 3586698 10.1016/0378-8741(87)90141-3
Matsuda H Samukawa K Fukuda S Shiomoto H Tung CN Kubo M Studies of Panax japonicus fibrinolysis Planta Med 1989 55 18 21 2717685
Leonhardt M Hrupka B Langhans W New approaches in the pharmacological treatment of obesity Eur J Nutr 1999 38 1 13 10338682 10.1007/s003940050040
Hill JO Melanson EL Wyatt HT Dietary fat intake and regulation of energy balance: implications for obesity J Nutr 2000 130 284S 288S 10721889
Leslie WS Lean MEJ Baillie HM Hankey CR Weight management: a comparison of existing dietary approaches in a work-site setting 2002 26 1469 1475 12439649
Verger R Borgstrom B, Brockman HL Pancreatic lipase 1984 Elsevier, Amsterdam 83 105
Konzo N Shoji J Nagumo N Komatsu N Studies on the constituents of Panax japonica rhizome. I. Isolation and purification of the saponins (1) Yakugaku Zasshi 1968 88 325 329
Konzo N Shoji J Nagumo N Komatsu N Studies on the constituents of Panax japonica rhizome. II. The structure of chikusetsusaponin IV and some observation on the structural relationship with araloside A Yakugaku Zasshi 1969 89 846 850 5817239
Han L-K Kimura Y Okuda H Reduction in fat storage during chitin-chitosan treatment in mice fed a high-fat diet Int J Obesity 1999 23 174 179 10.1038/sj.ijo.0800806
Black BL Croom J Eisen EJ Petro AE Edwards CL Surwit RS Differential effects of fat and sucrose on body composition in A/J and C57BL/6 mice Metabolism 1998 47 1354 1359 9826212 10.1016/S0026-0495(98)90304-3
Han L-K Takaku T Li J Kimura Y Okuda H Anti-obesity action of oolong tea Int J Obesity 1999 23 98 105 10.1038/sj.ijo.0800766
Lim K Shimomura Y Suzuki M Romsos DR, Himms-Hagen J, Suzuki M Obesity: Dietary factors and control 1991 Karger, Basel, Swittzerland 181 190 1934568
Astrup A Buemann B Western P Toubro S Raben A Christiensen NJ Obesity is an adaptation to a high-fat diet: evidence from cross-sectional study Am J Clin Nutr 1994 59 350 355 7993398
Portillo MP Simon E Garcia-Calonge MA Del Barrio AS Effects of high-fat diet on lipolysis in isolated adipocytes from visceral and subcutaneous WAT Eur J Nutr 1999 38 177 182 10502029 10.1007/s003940050059
Drent ML Larsson I William OT Quaade F Czubayko F von Bergmann K Strobel W Sjostrom L van der Veen EA Orlistat (Ro 18-0647), a lipase inhibitor, in the treatment of human obesity: a multiple dose study Int J Obesity 1995 19 221 226
Anai M Funaki M Ogihara T Kanda A Onishi Y Sakoda H Inukai K Nawano M Fukushima Y Yazaki Y Kikuchi M Oka Y Asano T Enhanced insulin-stimulated activation of phosphatidylinositol 3-kinase in the liver of high-fat-fed rats Diabetes 1999 48 158 169 9892238
| 15811191 | PMC1097713 | CC BY | 2021-01-04 16:31:46 | no | BMC Complement Altern Med. 2005 Apr 6; 5:9 | utf-8 | BMC Complement Altern Med | 2,005 | 10.1186/1472-6882-5-9 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1101586562510.1186/1471-2105-6-110Research ArticleAb initio identification of putative human transcription factor binding sites by comparative genomics Corà D [email protected] C [email protected] C [email protected] Cunto F [email protected] P [email protected] M [email protected] Dipartimento di Fisica Teorica dell'Università degli Studi di Torino and INFN, Via P. Giuria 1 – I 10125 Torino, Italy2 LGPD-IBDM, Université de la Méditerranée / CNRS, Campus de Luminy Case 907 – F-13288 Marseille Cedex 9, France3 Max-Planck-Institute for Molecular Genetics, Ihnestrasse 73 – D-14195 Berlin, Germany4 Dipartimento di Genetica, Biologia e Biochimica dell'Università di Torino, Via Santena 5 bis – I-10126 Torino, Italy2005 2 5 2005 6 110 110 2 12 2004 2 5 2005 Copyright © 2005 Corà 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
Understanding transcriptional regulation of gene expression is one of the greatest challenges of modern molecular biology. A central role in this mechanism is played by transcription factors, which typically bind to specific, short DNA sequence motifs usually located in the upstream region of the regulated genes. We discuss here a simple and powerful approach for the ab initio identification of these cis-regulatory motifs. The method we present integrates several elements: human-mouse comparison, statistical analysis of genomic sequences and the concept of coregulation. We apply it to a complete scan of the human genome.
Results
By using the catalogue of conserved upstream sequences collected in the CORG database we construct sets of genes sharing the same overrepresented motif (short DNA sequence) in their upstream regions both in human and in mouse. We perform this construction for all possible motifs from 5 to 8 nucleotides in length and then filter the resulting sets looking for two types of evidence of coregulation: first, we analyze the Gene Ontology annotation of the genes in the set, searching for statistically significant common annotations; second, we analyze the expression profiles of the genes in the set as measured by microarray experiments, searching for evidence of coexpression. The sets which pass one or both filters are conjectured to contain a significant fraction of coregulated genes, and the upstream motifs characterizing the sets are thus good candidates to be the binding sites of the TF's involved in such regulation.
In this way we find various known motifs and also some new candidate binding sites.
Conclusion
We have discussed a new integrated algorithm for the "ab initio" identification of transcription factor binding sites in the human genome. The method is based on three ingredients: comparative genomics, overrepresentation, different types of coregulation. The method is applied to a full-scan of the human genome, giving satisfactory results.
==== Body
Background
Understanding transcriptional regulation of gene expression is one of the greatest challenges of modern molecular biology. A central role in this mechanism is played by transcription factors (TF), which typically bind to specific, short DNA sequence motifs. These motifs are usually located in the upstream region of the regulated genes, although it is possible to find them also in the introns and in the 3' downstream region. They are often overrepresented, and appear in multiple copies inside the regulatory regions to form modules of cooperating items.
In these last years, the study of gene regulation has undergone a deep change of perspective [1,2]. While past studies usually dealt with individual regulatory interactions, it has become by now clear that the only way to understand the regulatory activity of the genome is to directly address the complex, combinatorial nature of the whole ensemble of TFs.
The identification of the cis-binding sequences and of the related TF's is a mandatory preliminary step toward this goal.
To this end it is becoming more and more important to construct tools able to
- address the problem on a genome wide scale
- keep under control the number of false positives to avoid an excessive increase of the noise to signal ratio
- use as input the statistical properties of the DNA sequences, thus avoiding, as far as possible, any other a priori assumption on the binding motifs.
However, the study cannot be based exclusively on the statistical features of the DNA regions presumably involved in transcriptional regulation, but must be complemented with independent information about gene regulation. In this respect three important sources of information may be used: the functional annotations collected in public databases, gene expression data on a global scale, and the so called 'phylogenetic footprinting' [3].
In fact large functional annotation databases and large-scale expression data provide a wealth of information about coregulation. This is a crucial point, since coregulated genes are likely to share similar transcriptional regulatory mechanisms. At the same time in these last years a growing interest has been attracted by the 'phylogenetic footprinting', i.e. the idea that functional sequences are preferentially conserved over the course of evolution by selective pressure. Comparison of orthologous gene sequences has been for a long time a standard tool in genomic analysis. Recently this comparative approach has been extended also to non coding regions, thanks to the progress of the sequencing programs. It is by now accepted that these non-coding conserved regions have an important regulatory role [4-9].
Several computational method have been proposed in the last few years to identify TF binding sites. These can be classified into two separate groups: enumerative methods, including the one we will present in this paper, explore all possible motifs up to a certain length (see for example [10-17]). The other large group consists of local search algorithms, including expectation maximization and various flavours of Gibbs sampling (see e.g. [18-21]).
We discuss here a simple and powerful approach for the identification of cis-regulatory motifs which fulfils the above requirements. It can be tuned to keep the number of false positives under control, and it allows us to study the transcriptional regulation of more than 10,000 genes of the human genome (which is a good approximation of a genome wide scale). The method is based on an "ab initio" study of the statistical properties of the regulatory sequences of the genes of interest. Together with the discussion of the method itself, we apply it to a full-genome analysis of the human case.
In particular in this paper, as a first step, we concentrated on the upstream sequences of the human genome (to be defined more precisely below), while we plan to extend these same tools to the downstream and intron regions in the future.
Results and Discussion
Our proposed approach is based on three main ingredients:
(1) human-mouse genomic comparison
(2) statistical analysis of "motifs" (short DNA sequences) that are overrepresented in evolutionarily conserved regions upstream of orthologous genes
(3) two complementary "filters" to infer coregulation: the distribution of Gene Ontology annotation terms and the results of a set of microarray experiments
The approach based on steps (2) and (3) above was successfully applied to the search for regulatory binding sites in yeast [14,15]. The human-mouse genetic comparison is crucial in extending the method to higher eukaryotes, since it is expected to greatly improve the signal/noise ratio by selecting for analysis those portions of the upstream regions that are more likely to be functionally relevant. Other algorithms taking advantage of phylogenetic footprinting to detect transcription factor binding sites have been published in [22,23].
As a final result of our analysis we obtain a set of motifs which survive one or both of the above filters, which we consider as our candidate binding sequences. The final step is then to cluster together these words to obtain consensus binding site sequences. This step allows, at the end of the whole process, to recover the intrinsic variability or regulatory motifs, that we know to be one of the most important features of binding sequences in higher eukaryotes.
A flow-chart of our algorithm is depicted in Figure 1.
The total number of motifs analyzed, after identifying each motif with its reverse complement and disregarding the self-overlapping ones, was 40,484. With the false discovery rate set at 10%, 373 of these turned out to be significant with at least one choice of scoring matrix (PAM1 or PAM10) and filter (Gene Ontology or microarray). 105 different Gene Ontology terms were involved, and 57 microarray time-points.
All the significant associations between motifs and Gene Ontology terms and microarray time points, for both scoring systems, are reported in Supplementary Table 1 [see Additional file 1 suptab1.txt]. The total number of such associations is 800, meaning that each motif was found, on average, about twice. Even if the various filters and scoring methods cannot be considered independent of each other, this fact is certainly encouraging in terms of the robustness of the method. These results are summarized in Tab. 1.
While the high degree of superposition between the results found with different choices is a clear indication of the robustness of our approach, the lists obtained are still significantly different: therefore the use of different filters/scoring matrices is useful in expanding the range of regulatory interactions explored by the algorithm.
In the supplementary table 2 we also provide the lists of the genes included in the sets found significant by one or more filter [see Additional file 2].
Not surprisingly, in many instances several motifs, often very similar to each other, are associated to the same GO term or microarray timepoint. For all Gene Ontology terms and microarray point associated to one or more motifs, we constructed, when possible, a consensus binding sequence from the motifs associated to the term as explained in the Materials and Methods section. The results are presented in Tab. 2, 3, 4, for the three branches of the Gene Ontology and in Tab. 5 for microarray time-points. For the latter, the consensus obtained is reported only when its length is at least 4. In many cases, in fact, the large number of motifs significantly associated to a microarray time-point causes the clustering algorithm to produce very short and rather uninformative consensus sequences. These results were produced by considering the PAM1 and PAM10 results together.
If our method is really able to identify genuine transcription factor binding sites, we would expect to find, among the surviving sets, at least some of the TF binding sites that are known to regulate the transcription of target genes through multiple occurrences in their promoters.
We focus here on some major examples.
E2Fs are transcription factors well known for their ability to regulate DNA replication by binding multiple sites in the promoters of the target genes [24]. Since the most abundant subpopulation of sets surviving the GO filter display a strong overrepresentation of DNA replication-related terms, it would be reasonable to expect that many of them are E2F binding sites. This was indeed the case, as the motifs TTGGCGC associated to many significant sets perfectly matched experimentally determined E2F binding sites as well as the consensus sequence found in the TRANSFAC[25] database. Significantly, some of these words were identified not only by the GO filter, but also by the microarray filtering scheme, confirming that our method is very robust in identifying the binding sites of this particular transcription factor.
It is interesting to see whether these motifs are found in the conserved parts of the upstream regions of experimentally verified targets of regulation by E2F. From the TRANSFAC database we identified 8 such targets, 6 of which are included in the PAM1 version of the CORG database: CAV1, MYC, DHFR, E2F2, RBL1 and CDC6. In five cases at least one of the motifs we find matching the E2F consensus can be found in the conserved part of the upstream regions, and in four cases many instances of the motifs are found: we find a total of 33 occurrences in the upstream region of MYC, 14 for E2F2, 11 for RBL1 and 11 for CDC6. Only one occurrence is found for DHFR, and none for CAV1. Similar results are found using the PAM10 version of CORG.
By performing these analyses, we observed that motifs characterized by the annotations 'chromatin' and 'nucleosome assembly', although obviously related to DNA replication, could not be reconciled with E2F binding sites, but included instead the motif AGAGCCTT and several similar ones. Since most of the annotated genes in the sets encoded for histone proteins, we speculated that these consensus could be part of a critical control element involved in the production of histones during DNA replication. One of the best known such elements is an evolutionary conserved inverted repeat found in the 3' untranslated region of histone mRNAs, controlling their stability during the cell cycle [26]. Surprisingly, our consensus sequence matched this element, raising the problem of how a 3' located regulatory element could be identified by our method. The reason is that histone genes form tight clusters in different chromosomal locations, and the distance between the initiator codon is in many cases below the 15000 bp limit used by our algorithm. Although of serendipitous nature, this result underscores two important points. The first is that our method is able to identify not only regular transcription factor binding sites, but also other less conventional regulatory elements characterized by a motif repetition. The second is that our approach could be systematically extended to other gene regions, such as the 3' untranslated and the introns.
The highly heterogeneous annotation associated to the sets surviving the GO filter strongly suggests that our method can potentially identify relevant binding sites for known and/or unknown transcription factors in the promoter of groups of genes involved in a wide variety of biological processes, such as tissue and organelle-specific transcription.
For instance, we identified many sets significantly enriched for genes involved in muscle development and/or functions (see Tab. 6). Interestingly, one of them (AGCAGG, associated to the term "sarcomere") is compatible with the binding site of the well known muscular master genes MyoD and Myf5 [27] as represented by a mixture of the TRANSFAC matrices M00184 and M0001, while the others had no significant match in the TRANSFAC database.
Another example are the different motifs associated with the annotations "endoplasmic reticulum", "protein transport" and "intracellular protein transport"(Tab. 7). Three of them (ACGTG, CCACGTCA and GACGTGGC) with known binding sites of ATF6 (TRANSFAC matrix M00483), a strongly conserved transcription factor involved in endoplasmic reticulum function [28]. The others don't show significant overlapping with TRANSFAC, suggesting that they are new putative cis elements important for regulating ER genes.
It is important to notice that in some instances, even though no hypothesis on the precise transcription factor can be formulated, it is at least possible to conjecture the general structural class to which the TF belongs. For example, the word GGGGGGGT, associated with the annotation "organogenesis", is consistent with the binding sites of many zinc finger transcription factors, such as Zic1, Zic3 and MZF1 [29], thus suggesting that some of the genes in the set are transcriptional target for a member of this particular family of transcription factors.
It is interesting to investigate the distribution of the distance of the motifs identified by our algorithm from the TSS of the corresponding gene. For all motifs found significant and for all genes in which the motif is overrepresented we computed the distance between the locations in which the motif is found and the TSS of the gene. All these data are represented as a histogram in Fig. 2. The motifs are very obviously concentrated near the TSS. This fact suggests that the choice to cut at 15,000 bp the length of the upstream regions considered is unlikely to decrease the signal significantly. The data shown are for PAM1, but the ones for PAM10 do not differ in any significant way.
Taken together, these results suggest that our approach has the potential to identify new critical regulatory elements for genes involved in a wide variety of biological processes.
Conclusion
We have discussed a new algorithm for the "ab initio" identification of transcription factor binding sites in the human genome. The method is based on three ingredients:
- the so called phylogenetic footprinting, i.e. the idea that functional sequences are preferentially conserved over the course of evolution by selective pressure.
- the overrepresentation criterion, i.e. the observation that binding sequences are usually overrepresented in the upstream region of the genes that are regulated by the corresponding transcription factors.
- the coregulation test, i.e. the use of coregulation (detected by using GO categories or microarray data) as a criterion to select the true positive binding sequences
Experience with yeast [14,15] suggests that our method is characterized by a low rate of false positives but, presumably, a rather high number of false negatives. The reason for this is that the basic ingredients of our analysis are motifs defined as completely specified sequences. This requires the motifs to be overrepresented, in the upstream region of a gene in order to be selected for our analysis and thus limits our candidates to a subset of all possible motifs. Our method is therefore complementary with respect to the standard approaches to binding sequences identification which use weighted matrices instead of completely specified motifs, since these typically have problems in detecting the true positive signals from the statistical background noise.
The variability of the motifs, which is a fundamental feature of Eukariotic binding sequences and is neglected at the beginning of our algorithm is recovered at the end thanks to the careful consensus reconstruction discussed in the previous section. These are the major novelties of our approach.
We consider as an encouraging validation test of our procedure the fact that several known TF binding sequences are found with our method. This makes us confident about the reliability of the other candidates that we found. Needless to say, these should be validated with suitable experimental tests. Indeed we think our "ab initio" approach could be of value as a preliminary test for any experimental search of binding sequences.
Several improvements of the present algorithm are possible. In particular it would be interesting to extend our analysis to other regions besides the 5' upstream one (the results on the control element of histones discussed above clearly indicate that this would be a fruitful research direction). In this respect the most natural candidates are the 3' downstream regions and the first intronic interval.
The method could be extended without major modifications to motifs with gaps, as considered in [40]. Extension to longer motifs would also be important: however the extension of the algorithm to motifs significantly longer than the ones considered here should probably take into account motif variability from the start, which would in turn imply a significant increase in computational complexity. We are currently investigating some possible ways of overcoming this problem.
Similarly, it would be important to address the combinatorial nature of transcriptional regulation by studying the correlation of overrepresented words along the lines discussed for instance in [30,31]. It is only by looking at the intricated network of interactions as a whole that one can hope to understand the collective behaviours leading to the tight and impressively efficient regulation of gene expression in higher eukaryotes and in particular in mammalians. It also clear that the algorithm can in principle be applied to any pair of closely related organisms.
We plan to address these issues in future work.
Methods
Construction of the new release of the CORG database
Definition of upstream regions and conserved non-coding blocks:
An upstream region is a sequence window that contains 5' genomic DNA extending from the start of translation of each individual transcript. The maximal size of an upstream region is taken to be 15 kbp. This upper bound stems from the observation that most promoter regions are less than 15,000 bp away from the start of translation [32]. Evidently, upstream regions may be smaller since they are bounded by the size of the intergenic region under consideration. Given this definition, upstream regions of different transcripts of the same gene or transcripts belonging to neighbouring genes could overlap. This is taken into account when compiling the conservation information by cutting the upstream region short of 15 kbp when necessary. All man and mouse DNA sequences were retrieved from the NCBI genome assemblies (NCBI33 and mNCBI30). Gene annotations were obtained from the EnsEMBL databases (release 17).
Orthologous man/mouse upstream regions were scanned for significant local similarities. We prefer a local alignment approach over a global one. That is we do not constrain the arrangement of putative regulatory modules. We denote these similarities as Conserved Non-coding sequence Blocks (CNBs). CNBs are computed with an implementation of the algorithm of Waterman and Eggert [33], which extends the well known Smith-Waterman algorithm to suboptimal alignments. The two scoring matrices used in the computation are derived from the Kimura two-parameter model and are normalized to a distance of 1 PAM and 10 PAM, respectively. The two matrices yield alignments of differing stringency with an expected level of identity of 99% for 1 PAM vs. 90.7 % for 10 PAM. Gap penalties were set to 11× match score for opening a gap and 0.1× match score for extending one. On average 8% (1 PAM) vs. 18% (10 PAM) of each upstream region (excluding repeats) is covered with CNBs.
An assessment of statistical significance of alignment scores was introduced to discriminate "true" from random alignments. Waterman and Vingron [34] showed that scores of local suboptimal alignments follow approximately the order statistics of a Poisson distribution. This facilitates the calculation of p-values by simulating random scores. We applied a P-value cutoff of 0.001.
Further details on the derivation of the data set can be found in [35]. Most of the data are part of the CORG database and can be accessed via the website .
Construction of the sets
The first step in the algorithm is the construction of sets of genes associated to all possible motifs. In this work a motif is defined as a short (5–8 bps), completely specified DNA sequence. The set associated to the motif m consists of all genes such that m is overrepresented, in the sense defined below, in the CNBs upstream of the genes. Motifs are always read on both strands, and therefore the sets associated to a motif and to its reverse complement coincide by definition. All genes for which one ore more CNBs were available were examined and assigned to one or more sets: 11,265 genes are included in the PAM1 version of CORG, and 13,294 in the PAM10 one.
The CORG database includes many rather long entries, up to several hundred bps. It is likely that many of these are actually exons. Since the inclusion of long exons would decrease our signal/noise ratio, we discarded all entries of length greater than 200 bps. We also eliminated multiple overlapping entries so that, as a final result of this preliminary step, each nucleotide in each conserved upstream region has exactly the same statistical weight. With this choice we end up with a total of 389560 distinct CNBs in the PAM10 case out of which 9155 (2.3%) have a length greater than 200 bp and, according to the strategy discussed above, were discarded in the following steps of our analysis. In the PAM1 case we find a total of 203417 CNBs out of which 3408 (1.7%) have a length greater than 200 b. As expected the proportion of CNBs larger than 200 bp decreases as the stringency of the alignments increases.
The definition of overrepresentation of a motif is the same that we used in Ref. [14] and [15], and was originally introduced in Ref.[16]. It is based on the frequency f(m) of the motif in all the CNBs contained in the database. For each gene we count the occurrences of m in the CNBs associated to the gene; then we compute the probability P of finding as many or more occurrences, based on a binomial distribution. The parameters of the binomial distribution are chosen as follows: f(m) is the success probability at each trial, and the number of trials is equal to the number of motifs that can be read in the CNBs associated to the gene. The use of the binomial approximation is based on the assumption of independence between successive trials. While rigorously speaking such assumption is never correct, it leads to serious errors only for periodic motifs, that are likely to be repeated several times in a row on the sequence. Therefore we did not include in our analysis the motifs that can be found repeated (possibly as their reverse complement) at a distance of 1, 2 or 3 bps (for example, respectively, CCCCC, ACGTA, CATCA).
If P < 0.01 we include the gene in the set labeled by the motif m. Notice that no biological significance is ascribed to these sets before they are selected for evidence of coregulation as explained below: therefore the choice of the cutoff on P can be arbitrarily lenient. Based on previous experience, we set the cutoff at P = 0.01. As it can be expected from the number of genes analyzed, essentially all possible motifs turn out to be overrepresented in some genes with this cutoff; however only a small fraction of them are selected by the GO and microarray filters and thus identified as candidate binding sites.
At this point we have thus obtained, for each possible motif m, a set of genes such that m is overrepresented in the CNBs of the genes in the set. The next step consists in looking for evidence of coregulation of the genes included in each set.
The Gene Ontology filter
As a first filter to select the sets whose genes are functionally related, and hence likely co-regulated, we analyze the prevalence of Gene Ontology (GO) annotations terms [36] in each set. For all GO terms associated to the genes of a set we perform an exact Fisher's test to determine whether the term appears in the set significantly more often than expected by chance. More precisely the Fisher's test gives us the probability P of obtaining an equal or greater number of genes annotated to the term in a set made of the same number of genes, but selected at random from the database. If P is statistically significant, then we can postulate the existence of a correlation between the overrepresentation of the motif m labeling the set and the functional characterization of the genes in the set, and hence include m in the list of candidate binding sites found by the algorithm.
Since this test is performed for all GO term and all sets, multiple testing is certainly an important issue. It is made rather non-trivial by the fact that the tests made on different GO terms are far from being independent of each other (think for example of testing the same set of genes for overrepresentation of the terms "cell cycle" and "DNA replication"). We chose to approach this issue with a safe, brute-force method based on random sampling, previously used in Ref. [15]: we generated a large sample of randomly selected gene sets of the typical size of our sets and we used it to estimate the number of false discoveries to be expected as a function of the cutoff on P-values. This allowed us to tune the cutoff on the Fisher's test P-values to obtain the desired value of the FDR (False Discovery Rate), which for the results we present is 10%.
The Microarray filter
An alternative and complementary filter to select candidate binding sites of our method uses microarray data. The assumption is that the distribution of expression values of a set of co-regulated genes is significantly different from the distribution of expression values of the whole genome. We used microarray data from the Stanford human cell-cyle experiment [37], consisting of 114 microarrays. We used the labels available in the raw data file (Unigene identifier and HUGO symbol when available) to make the correspondence with the Ensembl clusters used to identify the genes in the sets.
For each set and each microarray experiment, the comparison between the distribution of the data for the set and for the wholes genome was performed using the non parametric Kolmogorov-Smirnov (KS) test on the distribution of log(R), where R is the red/green normalized ratio. The goal is to identify the sets (and thus the corresponding motifs) showing an expression pattern significantly different from the background distribution (i.e. from the whole genome expression pattern for that particular microarray experiment). The non parametric KS test is the best suited tool for this type of analysis since it makes it possible to compare the expression levels measured in a given experiment without any a-priori assumption about the distribution of the data. Moreover the KS test looks for significant differences in the whole distribution (the statistic used is the largest difference between the cumulative distributions): therefore, at least in principle, it is able to detect subtle differences which would not be detected by tests based, for example, simply on the average expression level. However, like most non-parametric tests, the KS test is generally less potent than parametric tests, and hence requires a very strong signal to turn out significant. In particular, it is more likely to be successful in detecting the differential expression of large sets of genes, like our own. The KS test on expression data was previously used to identify candidate binding sites in Ref. [17].
Finally we evaluated the False Discovery Rate (FDR) by using the standard Benjamini-Hochberg method [38], setting a FDR threshold of 10%.
Construction of consensus sequences for the binding sites
In many cases several words, similar to each other, are found to be significantly associated to the same Gene Ontology term, or to the same microarray experiment. In such cases it is natural to assemble such words into a consensus sequence for the candidate binding site. This was systematically done for each Gene Ontology term in the following way: all the words associated to the same Gene Ontology term were aligned using the wconsensus [39] package. We selected the wconsensus results in the following way: the best matrix found by wconsensus was accepted if its expected frequency was less than 0.001; other matrices were also accepted if they exceeded such significance and they were generated from motifs that did not enter the previously accepted matrices. Therefore the algorithm is in principle capable of generating more than one consensus from a group of motifs. However in practice this never happens in our case: either one or no consensus sequence was produced for each group of motifs.
The same approach gives less satisfactory results when applied to the words associated to the same microarray experiment, the reason being that several microarray experiments turn out to be associated to a large number of rather different motifs, which cannot produce a single, meaningful consensus sequence. This is hardly surprising based on the results of the same approach applied to yeast [14], where the analysis of a rather small set of microarray experiments revealed many unrelated binding sites. Indeed genes regulated by several different transcription factor can be expected to show differential expression in the same experimental conditions. For several time-points, the best consensus was a three-letter sequence of dubious informative value. Only for six time-points we obtained a consensus of length 4 or higher.
Additional material and raw data are available at:
Authors' contributions
DC: implementation of most of the algorithms and also participation in the project design.
CH: microarray validation of the motifs.
CD: construction of the CORG database.
FDC: assessment of the biological significance of the results.
PP: contribution to the planning of the work and to the development of the algorithms used in the construction and analysis of the sets.
MC: supervision of the project.
All the authors contributed to the manuscript writing.
Supplementary Material
Additional File 1
The complete list of significant motif/GO term and motif/microarray time-point associations. The first column is the motif (to be considered coinciding with its reverse complement); the second column is the scoring matrix; the third is the type of filter ("GO" for Gene Ontology or "MA" for microarray); the fourth is the GO term or the microarray experiment; for GO terms, the fifth column contains the GO branch ("C": cellular component; "F": molecular function; "P": biological process); the sixth column is -log10 of the P-value of the test determining the significativit of the motif (Fisher's test for Gene Ontology, Kolmogorov-Smirnov for microarrays).
Click here for file
Additional File 2
The complete list of sets corresponding to the significant motifs. Each gene in each significant set is represented by its EnsEmbl ID.
Click here for file
Acknowledgements
P.P. is a Lagrange Fellow of the I.S.I. foundation. We thanks Enrico Curiotto for helpful discussion.
Figures and Tables
Figure 1 Flow-chart of the algorithm
Figure 2 Histogram of the distance from the TSS of the motifs found significant by the algorithm (see text).
Table 1 Number of significant motifs found with the four scoring matrix/filter combinations and their intersection. The third line contains the number of motif identified using both the PAM1 and PAM10 scoring system. The third column shows the number of motif identified by both the Gene Ontology and the microarray filter.
GO MA GO & MA
PAM1 139 61 29
PAM10 93 181 55
PAM1 & PAM10 42 38 17
Table 2 Consensus binding sites corresponding to GO terms in the biological process branch of the Gene Ontology. For each GO term we display either the consensus sequence obtained from wconsensus, or the longest motif associated to the term if the consensus sequence was not significant enough as defined in the text. The third column is the logarithm of the expected frequency of the alignment as given by wconsensus, if exists. The fourth column contains the number A of motifs which were used in the alignment and the total number B of motifs associated to the term in the format A/B. For this table the data obtained with PAM1 and PAM10 are considered together.
actin filament-based process GGGATTA - 1/1
ATP metabolism CCGTCCC - 1/1
biosynthesis CGCACG - 1/1
cell growth and/or maintenance CTTCA - 1/1
cell motility AGGGG - 1/2
defense response AGGAA - 1/1
development CCCC -32,3389 16/16
DNA metabolism TTCCCGC -35,3236 6/7
DNA replication and chromosome cycle TTCCCGCG -17,6184 4/4
DNA replication initiation GCGCGAAA - 1/1
enzyme linked receptor protein signaling pathway AGGGGG - 1/1
epidermal differentiation AGGCA - 1/1
frizzled-2 signaling pathway GCTGGAGA - 1/1
glycoprotein catabolism CTGACCTA - 1/1
heterophilic cell adhesion CTAAACTC - 1/1
immune response GAAAC - 1/1
intracellular protein transport CCACGTC -7,62462 2/2
L-amino acid transport ACTTTG - 1/1
macromolecule catabolism GACTC - 1/1
metabolism CGGAAG - 1/2
metabolism CGGGCCCG - 1/2
mitotic cell cycle TCCCGCCA - 1/1
muscle development CCAAG - 1/1
negative regulation of cell growth AACGACT - 1/1
nucleobase\, nucleoside\, nucleotide and nucleic acid metabolism AACGG - 1/4
nucleosome assembly GGCTCT -92,8905 27/40
organogenesis ACCCCCCC - 1/2
perception of chemical substance TCTAA - 1/1
phototransduction AAGRGGCC -12,0169 6/6
pinocytosis CTTACGA -7,62462 2/2
potassium ion transport CCAAG - 1/1
protein biosynthesis CGGAAG - 1/1
protein transport CCCAG - 1/1
regulation of apoptosis CATAG - 1/1
regulation of protein kinase activity AAAAG - 1/1
regulation of translation CGTGCTTC - 1/1
ribonucleotide metabolism CTTGATCC - 1/1
RNA localization ACGCCG - 1/1
synaptogenesis AGCGCCAC - 1/1
transcription CCGAG - 1/1
transcription\, DNA-dependent CCGAG - 1/2
translation ACTTCCGG - 1/1
two-component signal transduction system (phosphorelay) CACACGGG - 1/1
vision AATCCCT - 1/1
Table 3 Consensus binding sites corresponding to GO terms in the cellular component branch of the Gene Ontology. For each GO term we display either the consensus sequence obtained from wconsensus, or the longest motif associated to the term if the consensus sequence was not significant enough as defined in the text. The third column is the logarithm of the expected frequency of the alignment as given by wconsensus, if exists. The fourth column contains the number A of motifs which were used in the alignment and the total number B of motifs associated to the term in the format A/B. For this table the data obtained with PAM1 and PAM10 are considered together.
actin cytoskeleton AGGAC - 1/1
chromatin GGCTC -9,99174 3/3
chromosome GGCGGGAA - 1/2
chromosome\, pericentric region CAAATAGA - 1/1
clathrin-coated vesicle ATGGCA - 1/1
collagen GGACC - 1/1
COPI-coated vesicle CTCAGAG - 1/1
cytosol CGAAAGC - 1/2
cytosolic large ribosomal subunit (sensu Eukarya) CGGAGGAG - 1/3
cytosolic ribosome (sensu Eukarya) TTTCCG -11,6127 4/5
endoplasmic reticulum GACGTGGC - 1/4
eukaryotic 43S preinitiation complex CGGAAAA - 1/2
eukaryotic 48S initiation complex GGGCGGAA - 1/1
eukaryotic translation initiation factor 3 complex CACCTCCG - 1/4
external encapsulating structure GTATCTA - 1/1
extracellular matrix CAAATG - 1/2
extracellular space GGGAA - 1/1
fibrillar collagen ACCCT - 1/1
Golgi lumen CAACAT -8,99019 3/4
heterogeneous nuclear ribonucleoprotein complex AATGGCG - 1/4
inner membrane ACCGGCT - 1/1
integral to membrane ATCTCTG - 1/4
integral to nuclear inner membrane ACCTGAG - 1/2
intracellular CGGAAGCG -23,1425 5/15
lytic vacuole GATTCA - 1/1
membrane CCTGGC - 1/6
minor (U12-dependent) spliceosome complex ATTGCG - 1/1
mitochondrial inner membrane presequence translocase complex ACGGGAA - 1/2
mitochondrion AAGTTGC - 1/2
muscle fiber CCTCAG - 1/1
muscle myosin CAGAG - 1/1
muscle thin filament tropomyosin TCCTCCA - 1/1
nuclear chromatin ATTGAG - 1/1
nucleosome GGCTCT -85,6578 28/45
nucleus CACCAATC - 1/5
plasma membrane CTCCC - 1/1
replisome TCCCGCCA - 1/1
ribonucleoprotein complex CSGAA -18,8768 6/8
ribosome CGTGTAG - 1/3
sarcomere AGCAGG - 1/2
small ribosomal subunit GGCGGAA - 1/2
synaptic vesicle ACCAGAAT - 1/1
synaptonemal complex GGTCTTA - 1/1
vesicle coat ACTGCCT - 1/1
voltage-gated calcium channel complex CCTCCC - 1/1
Table 4 Consensus binding sites corresponding to GO terms in the molecular function branch of the Gene Ontology. For each GO term we display either the consensus sequence obtained from wconsensus, or the longest motif associated to the term if the consensus sequence was not significant enough as defined in the text. The third column is the logarithm of the expected frequency of the alignment as given by wconsensus, if exists. The fourth column contains the number A of motifs which were used in the alignment and the total number B of motifs associated to the term in the format A/B. For this table the data obtained with PAM1 and PAM10 are considered together.
calcium-activated potassium channel activity GCCACA - 1/1
chemoattractant activity GAATTTCC - 1/1
G-protein coupled receptor activity AATAG - 1/1
ligand-dependent nuclear receptor activity CAGGG - 1/1
nucleic acid binding CGGGAG - 1/2
pancreatic ribonuclease activity AACTACTC - 1/1
phosphatidylinositol-4\, 5-bisphosphate 3-kinase activity AAGGA - 1/1
retinoic acid receptor activity ACCCA - 1/1
RNA binding ATGGCG - 1/1
serine-type endopeptidase activity CAGAGGG - 1/1
single-stranded DNA binding AAACC - 1/1
surfactant activity ACTCACCC - 1/1
translation factor activity\, nucleic acid binding CGGAAG - 1/1
uncoupling protein activity GACGTAGC - 1/1
Table 5 Consensus binding sites corresponding to microarray time-points. Only time-points for which the clustering algorithm produced a consensus sequence of length 4 or more are shown.
Timepoint consensus sequences used ln (expected freq.)
t = 23 CTGG 4/7 -7,99646
t = 50 CCMCA 5/15 -9,71859
t = 61 SCCAGG 12/43 -18,6948
t = 89 CWGGG 17/23 -11,1386
t = 100 CCCWG 12/31 -12,5918
t = 107 CGGM 13/14 -14,7383
Table 6 Words associated to "muscle development" and related terms
AGCAGG sarcomere
CCAAG sarcomere
CCAAG muscle development
TCCTCCA muscle thin filament tropomyosin
Table 7 Words associated to "endoplasmic reticulum", "protein transport" and "intracellular protein transport"
AAGTTGG endoplasmic reticulum
AATCGGC endoplasmic reticulum
ATCAGCG endoplasmic reticulum
CGCAG endoplasmic reticulum
GACGTGGC endoplasmic reticulum
ACGTG intracellular protein transport
CCACGTCA intracellular protein transport
GACGTGGC intracellular protein transport
CCCAG protein transport
==== Refs
Wassermann WW Sandelin A Applied bioinformatics for the identification of regulatory elements Nat Rev Genet 2004 5 276 87 15131651 10.1038/nrg1315
Pennacchio LA Rubin EM Genomic strategies to identify mammalian regulatory sequences Nat Rev Genet 2001 2 100 109 11253049 10.1038/35052548
Hardison R Conserved non-coding sequences are reliable guides to regulatory elements Trends Genet 2000 16 369 372 10973062 10.1016/S0168-9525(00)02081-3
Duret L Dorkeld F Gautier C Strong conservation of non-coding sequences during vertebrates evolution: potential involvement in post-transcriptional regulation of gene expression Nucleic Acid Res 1993 21 2315 2322 8506129
Loots GG Locksley RM Blankespoor CM Wang ZE Miller W Rubin EM Frazer KA Identification of a coordinate regulator of interleukins 4, 13, and 5 by cross-species sequence comparisons Science 2000 288 136 140 10753117 10.1126/science.288.5463.136
Goettgens B Barton L Gilbert J Bench A Sanchez M Bahn S Mistry S Grafham D McMurray A Vaudin M Amaya E Bentley D Green A Sinclair A Analysis of vertebrate scl loci identifies conserved enhancers Nat Biotechnol 2000 18 181 186 10657125 10.1038/72635
Flint J Tufarelli C Peden J Clark K Daniels R Hardison R Miller W Philipsen S Tan-Un K McMorrow T Frampton J Alter B Frischauf A Higgs D Comparative genome analysis delimits a chromosomal domain and identifies key regulatory elements in the alpha globin cluster Hum Mol Genet 2001 10 371 382 11157800 10.1093/hmg/10.4.371
Lenhard B Sandelin A Mendoza L Engström P Jareborg N Wasserman WW Identification of conserved regulatory elements by comparative genome analysis J Biol 2003 2 13 12760745 10.1186/1475-4924-2-13
Zhang Z Gerstein M Of mice and men: phylogenetic footprinting aids the discovery of regulatory elements J Biol 2003 2 11 12814519 10.1186/1475-4924-2-11
Sinha S Tompa M Discovery of novel transcription factor binding sites by statistical overrepresentation Nucleic Acids Res 2002 30 5549 12490723 10.1093/nar/gkf669
Sinha S Tompa M YMF: A program for discovery of novel transcription factor binding sites by statistical overrepresentation Nucleic Acids Res 2003 31 3586 12824371 10.1093/nar/gkg618
Birnbaum K Benfey PN Shasha DE cis element/transcription factor analysis (cis/TF): a method for discovering transcription factor/cis element relationships Genome Research 2001 11 1567 11544201 10.1101/gr.158301
Wolfsberg TG Gabrielian AE Campbell MJ Cho RJ Spouge JL Landsman D Candidate regulatory sequence elements for cell cycle-dependent transcription in Saccharomyces cerevisiae Genome Research 1999 9 775 10447512
Caselle M Di Cunto F Provero P Correlating overrepresented upstream motifs to gene expression: a computational approach to regulatory element discovery in eukaryotes BMC Bioinformatics 2002 3 7 11876822 10.1186/1471-2105-3-7
Cora' D Di Cunto F Provero P Silengo L Caselle M Computational identification of transcription factor binding sites by functional analysis of sets of genes sharing overrepresented upstream motifs BMC Bioinformatics 2004 5 57 15137914 10.1186/1471-2105-5-57
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 281 827 42 1998 Sep 4 9719638 10.1006/jmbi.1998.1947
Jensen LJ Knudsen S Automatic discovery of regulatory patterns in promoter regions based on whole cell expression data and functional annotation Bioinformatics 2000 16 326 33 10869030 10.1093/bioinformatics/16.4.326
Lawrence CE Reilly AA An expectation maximization (EM) algorithm for the identification and characterization of common sites in unaligned biopolymer sequences Proteins 1990 7 41 2184437
Thijs G Marchal K Lescot M Rombauts S De Moor B Rouze P Moreau Y A Gibbs sampling method to detect overrepresented motifs in the upstream regions of coexpressed genes J Comput Biol 2002 9 447 12015892 10.1089/10665270252935566
Thompson W Rouchka EC Lawrence CE Gibbs Recursive Sampler: finding transcription factor binding sites Nucleic Acids Res 2003 31 3580 12824370 10.1093/nar/gkg608
Hughes JD Estep PW Tavazoie S Church GM Computational identification of cis-regulatory elements associated with groups of functionally related genes in Saccharomyces cerevisiae J Mol Biol 2000 296 1205 10698627 10.1006/jmbi.2000.3519
Sandelin A Wasserman WW Lenhard B ConSite: web-based prediction of regulatory elements using cross-species comparison Nucleic Acids Res W249 52 2004, Jul 1
Prakash A Blanchette M Sinha S Tompa M Motif discovery in heterogeneous sequence data Pac Symp Biocomput 2004 348 59 14992516
Ohtani K DeGregori J Nevins JR Regulation of the cyclin E gene by transcription factor E2F1 PNAS 1995 92 12146 50 8618861
Matys V Fricke E Geffers R Gossling E Haubrock M Hehl R Hornischer K Karas D Kel AE Kel-Margoulis OV Kloos DU Land S Lewicki-Potapov B Michael H Munch R Reuter I Rotert S Saxel H Scheer M Thiele S Wingender E TRANSFAC: transcriptional regulation, from patterns to profiles Nucleic Acids Res 2003 31 374 12520026 10.1093/nar/gkg108
Sittman DB Graves RA Marzluff WF Structure of a cluster of mouse histone genes Nucleic Acids Res 1993 11 6679 97
Wentworth BM Donoghue M Engert JC Berglund EB Rosenthal N Paired MyoD-binding sites regulate myosin light chain gene expression PNAS 1991 88 1242 6 1847512
Wang Y Shen J Arenzana N Tirasophon W Kaufman RJ Prywes R Activation of ATF6 and an ATF6 DNA binding site by the endoplasmic reticulum stress response J Biol Chem 2000 275 27013 20 10856300
Mizugishi K Aruga J Nakata K Mikoshiba K Molecular properties of Zic proteins as transcriptional regulators and their relationship to GLI proteins J Biol Chem 276 2180 8 2001 Jan 19 11053430 10.1074/jbc.M004430200
Sudarsanam P Pilpel Y Church GM Genome-wide cooccurrence of promoter elements reveals a cis-regulatory cassette of rRNA transcription motifs in S. cerevisiae Genome Research 2002 12 1723 12421759 10.1101/gr.301202
Chiang DerekY Moses AlanM Manolis Kellis Lander EricS Eisen MichaelB Phylogenetically and conserved word pairs associated with gene-expression changes in yeasts Genome Biology 2003 4 R43 12844359 10.1186/gb-2003-4-7-r43
Dieterich C Cusack B Wang H Rateitschak K Krause A Vingron M Annotating regulatory DNA based on man-mouse genomic comparison Bioinformatics 2002 18 S84 12385988
Waterman MS Eggert M A new algorithm for best subsequence alignmnents with application to tRNA-rRNA comparison J Mol Biol 1997 197 723 728 10.1016/0022-2836(87)90478-5
Waterman MS Vingron M Sequence comparison significance and Poisson approximation Statistical Science 1994 9 367 381
Dieterich C Wang H Rateitschak K Luz H Vingron M CORG: a database for Comparative Regulatory Genomics Nucleic Acid Res 2003 31 55 57 12519946 10.1093/nar/gkg007
The Gene Ontology Consortium Gene Ontology: tool for the unification of biology Nature Genetics 2000 25 25 29 10802651 10.1038/75556
Whitfield ML Sherlock G Saldanha AJ Murray JI Ball CA Alexander KE Matese JC Perou CM Hurt MM Brown PO Botstein D Identification of genes periodically expressed in the human cell cycle and their expression in tumors Mol Biol Cell 2002 13 1977 2000 12058064 10.1091/mbc.02-02-0030.
Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing J R Statist Soc 1995 B57 289
Stormo G Hartzell GW Identifying protein-binding sites from unaligned DNA fragments PNAS 1989 86 1183 1187 2919167
van Helden J Rios AF Collado-Vidas J Discovering regulatory elements in non-coding sequences by analysis of spaced dyads Nucleic Acid Res 2000 15 1808 18 10.1093/nar/28.8.1808
| 15865625 | PMC1097714 | CC BY | 2021-01-04 16:02:48 | no | BMC Bioinformatics. 2005 May 2; 6:110 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-110 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-921581998910.1186/1471-2105-6-92Research ArticleShortest triplet clustering: reconstructing large phylogenies using representative sets Sy Vinh Le [email protected] Haeseler Arndt [email protected] Heinrich-Heine-Universität Düsseldorf, WE Informatik, Universitätstr. 1, D-040225 Düsseldorf, Germany2 Forschungszentrum Jülich, Germany2005 8 4 2005 6 92 92 29 11 2004 8 4 2005 Copyright © 2005 Sy Vinh and von Haeseler; 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
Understanding the evolutionary relationships among species based on their genetic information is one of the primary objectives in phylogenetic analysis. Reconstructing phylogenies for large data sets is still a challenging task in Bioinformatics.
Results
We propose a new distance-based clustering method, the shortest triplet clustering algorithm (STC), to reconstruct phylogenies. The main idea is the introduction of a natural definition of so-called k-representative sets. Based on k-representative sets, shortest triplets are reconstructed and serve as building blocks for the STC algorithm to agglomerate sequences for tree reconstruction in O(n2) time for n sequences.
Simulations show that STC gives better topological accuracy than other tested methods that also build a first starting tree. STC appears as a very good method to start the tree reconstruction. However, all tested methods give similar results if balanced nearest neighbor interchange (BNNI) is applied as a post-processing step. BNNI leads to an improvement in all instances. The program is available at .
Conclusion
The results demonstrate that the new approach efficiently reconstructs phylogenies for large data sets. We found that BNNI boosts the topological accuracy of all methods including STC, therefore, one should use BNNI as a post-processing step to get better topological accuracy.
==== Body
Background
Reconstructing the evolutionary relationships among species based on their genetic information is one of the primary objectives in phylogenetic analysis. In recent years, numerous heuristics to reconstruct phylogenies for large data sets have been proposed [1-11]. In addition, parallel tree-reconstruction programs have been implemented [12-15].
To date, distance-based methods introduced by Cavalli-Sforza and Edwards [16] and by Fitch and Margoliash [17] appear most appropriate to reconstruct phylogenies based on thousands of sequences. These methods are a compromise between computational speed and topological accuracy [1,3,5-7] and run typically in O(n3) time for n sequences [1,3,5] or in O(n2) for recently suggested approaches [6,7]. Clustering algorithms form a major class of distance-based methods [18]. They do not have an explicit objective function that needs to be optimized. They rather group sequences (or taxa) iteratively to reconstruct a distance-based phylogenetic tree. UPGMA is a popular method to infer phylogenies with the constraint that a molecular clock is imposed on the evolutionary process. Other clustering approaches have been proposed to relax the molecular clock assumption [1,3,5,19-21].
An attempt to boost the accuracy and to reduce the computational burden is the introduction of k-representative set concepts [10,11]. k-representative sets consist of at most k elements but retain the most important information from whole sets. In this paper, we extend our original approach [10] by introducing a more natural k-representative set concept. In a nutshell, representative sets are regarded as components to construct shortest triplets, each of which comprises three closely related sequences from three k-representative sets. The collection of shortest triplets serves as building block for a new distance-based clustering method called shortest triplet clustering algorithm (STC).
Results
Simulations were run on a PC cluster with 16 nodes. Each node has two 1.8 GHz processors and 2 GB RAM. Seq-Gen [22] was used to evolve sequences along trees using the Kimura two-parameter model [23] with a transition/transversion ratio of 2.0. We generated 100 simulated data sets of 500 sequences each with sequence lengths 500, 1000 and 2000 nucleotides (nt), respectively. As one model tree, we used the rbcl gene tree with diameter 0.36 substitutions per site as inferred from an alignment of 500 rbcl-genes [10]. We call this the rbcl-simulation.
In a second experiment, the so-called large simulation, tree topologies were drawn from the Yule-Harding distribution [24], and edge lengths were drawn from an exponential distribution and subsequently rescaled such that the mean diameter of the tree was either 0.1, 0.5, 1.0, or 1.5. For each value of the diameter we generated 100 trees with 1000 sequences and 100 trees with 5000 sequences. Thus, a total of 800 trees were used.
Finally, we tested the accuracy and runtime of the STC and compared it with six other commonly used distance-based methods. More specifically, we investigate the performance of the Neighbor-Joining method (NJ) [1] implemented in PAUP* 4.0 [25], BIONJ [3], Weighbor 1.2 [5], Harmony Greedy Triplet and Four Point Condition (HGT/FP) [7] as well as Greedy Minimum Evolution (GME) and Balanced Minimum Evolution (BME) [6]. Unfortunately, no distance-based program is available for the disc-covering method [4]. All methods were combined with DNADIST version 3.5 [26] and pairwise distances were corrected for multiple hits according to the model used in the simulation. Moreover, we examined the performance of all methods when the balanced nearest neighbor interchange (BNNI) [6] is used as a post-processing step.
Further, to illustrate the performance of STC we re-analyzed the 96-taxon alignments of sequence length 500 nt, that were analyzed in [6] and available at . The 6000 trees were split into three groups called "slow" (0.2 substitutions per site), "moderate" (0.4 substitutions per site) and "fast" (1.0 substitutions per site). We call this the re-analyzed simulation.
The accuracy of a tree reconstruction method for a simulated data set is measured by the Robinson and Foulds (RF) distance [27] between the inferred tree and the model tree used to generate the data set. The RF distance between two trees is the number of bi-partitions present in one of the two trees but not the other, divided by the number of possible bi-partitions. Thus, the smaller the RF distance between two trees the closer are their topologies. In other words, the smaller the RF distance is between the inferred tree and the model tree the higher is the topological accuracy of the tree reconstruction method.
In the following we discuss the results of the rbcl-simulation, and the large simulation and the re-analyzed simulation.
rbcl-simulation
Table 1 shows that the STC outperforms all other methods analyzed in terms of topologlcal accuracy. For instance, the RF distance between the STC-tree and the model tree is on average 0.177 (with respect to the sequence length of 500 nt) and better than NJ (0.190), slightly better than the second best method BME (0.184) and much better than HGT/FP (0.512). Table 1 also demonstrates that all tested methods including STC give higher topologlcal accuracy when the sequence length is increased. However, Table 2 shows that other methods in combination with BNNI outperform STC without BNNI. The combination of STC and BNNI shows similar performance as the combinations of NJ (BIONJ, Weighbor) and BNNI and, slightly better results than the combination of GME (HGT/FP) and BNNI.
Large simulation
Due to the increase in runtime, Weighbor could not be tested. Table 3 and 4 show that STC gives better results than the other methods independent of the diameter. All methods display a decrease in accuracy when the number of sequences changes from 1000 to 5000. As shown in Table 5 and 6, BNNI boosts the accuracy of all methods including STC. All methods give similar results when being used together with BNNI.
Re-analyzed simulation
Except for STC, the accuracies for the other methods displayed in Table 7 and 8 were taken from [6]. Table 7 shows that STC outperforms the other methods in terms of topological accuracy with the exception that Weighbor is slightly better than STC with respect to the slow simulation group. If BNNI is applied, all methods exhibit an almost identical performance (see Table 8).
Another look at the performance
Instead of looking at the average RF distance, we suggest to take a closer look at the simulated data. For each simulated data set, that is subjected to the STC and six other tree reconstruction methods mentioned above, we compute the RF distance between the reconstructed tree and the model tree for all methods. Figure 1 illustrates the results for the large simulation when comparing STC with NJ (left column) and STC with the second best method BME (right column). In each diagram specified by the number of taxa and reconstruction methods, 400 points are displayed, that resulted from 100 simulations for each of the tree-diameters (0.1, 0.5, 1.0 and 1.5). Although four tree-diameters were studied only two clouds of points are discernible, where the cloud in the north-east corner of each diagram represents the simulations with the tree-diameter 0.1. The remaining 300 points gather in the south-west cloud because the RF-distances from trees with diameter 0.5, 1.0, 1.5 are not substantially different from each other (see Table 3 and 4). More precisely, the horizontal and vertical axes indicate the RF distances of STC and NJ (or BME), respectively. Each point in the graph presents the RF distance for a simulated data set. Points above the dotted line are examples where the RF distance of the STC-tree is less than the RF distance of the NJ-tree or BME-tree. Thus, the STC gives higher topological accuracy than NJ or BME with respect to the simulated data set. For example, Figure 1a illustrates the comparison between STC and NJ with respect to 1000 taxa data sets. 379 out of 400 points are above the diagonal, thus, STC gives better results than NJ in about 95% of the simulations. For the remaining 21 alignments (points), two methods showed the same RF distance. Finally, we found 19 points below the diagonal in which case NJ outperforms STC. For the large simulation (5000 taxa), NJ is worse than STC in all cases. However, the second best method BME is better than STC in 11% and 5% of the cases with respect to 1000 and 5000 sequence data sets.
Figure 2 shows the same analysis for the rbcl simulation. It shows that with increasing sequence length the cloud of points moves towards zero. From Figure 2 we learn that in some instances NJ (or BME) performs better (with regard to the RF distance) than STC, i.e. 20%, 12%, 8% (or 34%, 17%, 14%) of the simulations for sequence lengths 500, 1000 and 2000 nt, respectively.
Similar results hold for the other methods. These results are summarized in Table 9 where we show the percentage of simulations in which STC is at least as good as the other methods.
Again, if BNNI is applied we observe that no substantial difference among the various approaches. The accuracy of the methods is mostly determined by BNNI (see Table 10).
Conclusion
We are presenting k-representative sets which allow us to design a fast and accurate method to reconstruct phylogenies from large data sets with 1000 or more taxa. Simulations show that STC gives better results than other tested methods in terms of topological accuracy. However, if BNNI is introduced as a subsequent optimization step, the differences in the performance disappear. All methods show more or less the same accuracy. Thus, one should apply BNNI to improve the topological accuracy.
The time to reconstruct a tree of up to 1000 sequences is not really an issue for all tested distance-based methods, with the exception of Weighbor. Weighbor needed about 19 minutes to reconstruct a tree with 500 sequences, thus it is only applicable to data sets with up to some hundred sequences. For data sets with up to 1000 sequences, the remaining methods needed less than one minute to output a tree, thus the difference between methods in terms of runtime is not significant. For data sets with 5000 sequences, STC (GME, HGT/FP or BME) with BNNI took about 2.0 (2.5, 3.0 or 3.5) minutes to reconstruct a tree. NJ (BIONJ) with BNNI were slower and consumed approximately six minutes to output a tree. In short, the combination of STC and BNNI efficiently reconstruct trees for large data sets in both terms of topological accuracy and runtime.
Finally, we did not systematically evaluate the impact of the number of representatives k. We present some preliminary results for k = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100. Figure 3 shows that the RF distance of STC decreases when k grows from 1 to 5. This proves our intuition that a too small number of triplets leads to an inaccurate estimate of path lengths and edge lengths. When k ranges from 5 to 10, the RF distance remains more or less unchanged. For k ≥ 10, the RF distance increases steadily indicating a loss of accuracy. The decrease in accuracy is explained by the inclusion of triplets with large distances which include noise and disturb the reconstruction. Thus, we chose k = 5 as a good compromise between the accuracy and computational complexity for all data sets. That is, the practical complexity of the STC algorithm is only O(n2).
Methods
In this section we introduce a new clustering algorithm to reconstruct phylogenies based on distance matrices.
Additive distances
Let S = {s1, s2,..., sn} be a set of n objects (typically contemporary sequences/taxa), let D = D(uv) be a distance matrix where D(uv) is the distance between two objects u and v.
Definition 1
The distance matrix D is additive if and only if it satisfies the four-point condition [28]: for any quartet {u, v, w, x},
D(uv) + D(wx) ≤ max{D(uw) + D(vx), D(ux) + D(vw)}.
In this case, the objects s ∈ S are related by a tree T = (V, E) where V is the set of vertices such that S ⊂ V and E = {{v1, v2}|v1, v2 ∈ V} is the set of edges. A vertex with one adjacent edge is called a leaf, all other vertices are called internal nodes. We let L ⊂ V be the leaf set of the tree T. Note that we typically require L ⊆ S in the phylogenetic setting.
If D is additive, then there exists a map and a length function such that
for all u, v ∈ S where p(φ (u), φ(v)) is the unique path connecting φ(u) and φ(v) in T and denotes the distances between vertices in T (cf. [29]). ℓ(e) is called edge length of the edge e. To avoid unnecessary complication, we consider only one-to-one maps from S on the leaf set L of T. If D is additive, the reconstruction of tree T and ℓ is trivial. If D is not additive, methods are available that try to fit a tree T to D with respect to an objective function (cf. [30]). Thus, in the following we consider arbitrary distance matrices and we want to reconstruct a tree together with a length function .
Estimating edge lengths using triplets
We consider a subset X of S, then induces a map on a subtree of T such that the relationships of objects in X are displayed by the subtree with leaf set φ(X). The complement S0(X) = S - X we will call the unclassified object set, because the relationships of objects in S0(X) to X is not known from the subtree. Note that we will use S0 instead of S0(X) if X is clear from the context.
Let denote Tr = (Vr, Er) a rooted tree with root r and leaf set Lr, and let Sr be a subset of S such that φ(Sr) = Lr. For convenience, we use Sr and Lr interchangeably.
Now, we consider the most simple edge length estimation problem. That is, we would like to estimate the edge lengths for a triplet tree {a, b, c} with distance matrix D (see Figure 4a). Edge lengths are estimated as follows
Now consider a rooted Tr with the inferred tree-like metric . The rooted tree Tr consists of two rooted subtrees and (see Figure 4b). For convenience, we will use Ti instead of if ri is clear from the context. The leaf set Sr = {S1 ∪ S2} where Sr ⊂ S and S0 = S - Sr is not represented in Tr. Then we can compute
for each triplet (s0, s1, s2) ∈ (S0 × S1 × S2).
With (s1r1) and (s2r2) we denote the known distances of s1 and s2 to their roots r1 and r2. Thus, we can compute for each triplet {s0, s1, s2} the lengths ℓ(r1r) and ℓ(r2r) as
Note that, if D is additive and T1, T2 are isometric subtrees of T, the lengths ℓ(r1r) and ℓ(r2r) do not depend on the choice of the triplet {s0, s1, s2}.
Regardless of additivity considerations, we may define the average length for a fixed s0 ∈ S0 as
We can estimate edge lengths ℓ(r1r) and ℓ(r2r) by using all possible triplets as
Recovering a tree from a distance matrix
The largest path length criterion
We want to reconstruct a tree T = (V, E) with respect to a distance matrix D such that DT represents D. To this end, we use triplets and the notation of a rooted tree Tr together with Equations 4 and 5.
Our algorithm starts with the observation that if we take an arbitrarily rooted tree Tm with m ∈ S and length function , then there must be a pair of leaves (neighboring leaves) that share an immediate most recent common ancestor mrca which is farthest away from the root m with respect to . In Figure 5, the pair (3, 4) satisfies this condition, we say this pair fulfills the largest path length criterion. The largest path length criterion easily generalizes to arbitrarily rooted subtrees Ti and Tj of Tm, where all descendants from the roots of Ti and Tj are in the vertex sets Vi or Vj, respectively.
Let be the set of rooted subtrees of Tm (each leaf l ∈ Lm is considered as a rooted subtree Tl). Now consider two disjoint rooted subtrees Ti and Tj of Tm where i, j ∈ Vm. Then the distance ℓ(m, mrca|mSiSj) from the mrca of Ti and Tj to m is computed according to Equation 4, where Si and Sj are the leaf sets of Ti and Tj, respectively. Then we pick
(, ) = argmax{ℓ(m, mrca|mSiSj) | Ti, Tj ∈ } (6)
as a pair of neighbors (if we detect more than one pair, we randomly select one). By construction, (, ) fulfills the largest path length criterion.
If D is additive, ℓ(m, mrca|mSiSj) is exactly the path length from the mrca of (Ti, Tj) to m. In other words, the path length from the mrca of (, ) to m is large stand (, ) is a true neighboring pair. However, in real applications D is rarely additive, therefore the root m is selected so as to avoid noise from stochastic errors involved with large distance estimates [17]. To this end, m is selected such that the distance from the farthest object to root m is minimal,
med = argminm'∈S{max{D(m'x)|x = 1,..., n}} (7)
med is called a median object.
Moreover, to reduce the computational complexity of finding a pair of neighbors (, ) using Equation 6, we store for each Ti ∈ its potential neighbor Ti' ∈ such that
Ti' = argmax{ℓ(med, mrca|med, Si, Sj)|Tj ∈ }. (8)
Now the neighboring pair (, ) fulfilling the largest path length criterion is determined as follows
(, ) = argmax{ℓ(med, mrca|med, Si, Si')|Ti ∈ }. (9)
In the following, we present a natural clustering algorithm to reconstruct trees based on distance matrices and the largest path length criterion
Clustering Algorithm
• Initial step: Find the median object med using Equation 7. Set = {T1,..., Tn} - {Tmed}. Find for each Ti ∈ its potential neighbor Ti' ∈ using Equation 8.
• Selection step (largest path length criterion): Find the neighboring pair (, ) using Equation 9.
• Agglomeration step: Combine and into a new rooted tree with root i0j0, and estimate new edge lengths of using Equation 5. Delete and and add to . Find the potential neighbor for the new rooted tree . using Equation 8, and replace Ti' for each Ti ∈ by if is its potential neighbor.
• Stopping step: If || > 1 goto the Selection step, otherwise output the tree.
This algorithm is similar to approaches described elsewhere [19-21], however, an essential difference is that we estimate path lengths and edge lengths by using triplets.
Local rearrangement
The heart of the clustering algorithm is the largest path length criterion, at which the path length from the mrca of (Ti, Tj) to med is estimated by ℓ(med, mrca|med, Si, Sj) using Equation 4. Thus, as path length we take the average of the lengths obtained from at most O(n2 triplets {med, si, sj} ∈ med × Si × Sj. This average may not be the representative estimate of the true path length. Moreover the root med may be too far way from the mrca and this leads to an inaccurate estimate of the path length.
To take these problems into account, we extend the clustering algorithm. To this end, imagine the algorithm has clustered Ti and Tj with corresponding disjoint leaf sets Si, Sj ⊂ S (we have finished the agglomeration step). Thus, we have created the newly rooted tree T{ij} with leaf set Sij = {Si ∪ Sj} and the set of unclassified objects S0(Sij) = S - Sij. In the following, we describe the nearest neighbor interchange operation around the root of Ti upon condition that Ti consists of two rooted subtrees Tx, Ty (Figure 6a). First, we estimate average path lengths from the unclassified object set S0(Sij) to the mrca of (Tx, Ty), (Tx, Tj) and (Ty, Tj) as
For convenience, we will use ℓ(S0(Sij)|SxSy) instead of ℓ(S0(Sij)SxSy|SxSy). We now use the average path lengths from Equation 10 to decide which pair of subtrees among (Tx, Ty), (Tx, Tj) and (Ty, Tj) is preferred. More specifically, if
ℓ(S0(Sij)|SxSy) ≥ max{ℓ(S0(Sij)|SxSj), ℓ(S0(Sij)|SySj)}
we stick to the suggested grouping of Tx and Ty (see Figure 6a). Otherwise, if ℓ(S0(Sij)|SxSj) or ℓ(S0(Sij)|SySj) is larger than the remaining average path lengths, we swap Ty and Tj or Tx and Tj as displayed in Figure 6b or 6c, respectively. Note that, this decision can be considered as a correction of the largest path length criterion by taking all possible triplets into account. We call the correction the largest average path length criterion.
We now explain the preorder traversal procedure [31] to reconstruct the rooted tree Ti using the nearest neighbor interchange operation based on the largest average path length criterion (Ti is a subtree of T{ij} = (Ti, Tj)):
Preorder traversal procedure (Ti)
• Step 1: If Ti is a single leaf, return.
• Step 2: Otherwise, Ti consists of two subtrees Tx and Ty. Do the nearest neighbor interchange operation around the root of Ti based on the largest average path length criterion (Equation 10). If Tx and Tj (or Ty and Tj) were exchanged, estimate new edge lengths using Equation 5.
• Step 3: Apply the preorder traversal procedure to two rooted subtrees of Ti.
Representative sets and shortest triplets
For a set S of sequences (or taxa), the (genetic) distance matrix D is typically not additive due to stochastic errors [17]. Larger distances between two sequences are less accurately estimated. This leads to a low performance of both the clustering algorithm and the preorder traversal procedure for divergent data sets.
In earlier work [10,11], we have presented simple representative concepts to reduce stochastic error involved in large distances. Here, we extend our work by introducing the so-called k-representative set concept. We use now genetic distances instead of topological distances (all edges have length 1). Our motivation is to reduce the computational complexity and to exclude objects far away from the root under consideration. In the clustering algorithm, the path length from the mrca of (Ti, Tj) to med (see Figure 7) can be estimated by two approaches. The first method picks randomly one pair (si, sj) ∈ Si × Sj then computes
The second approach takes the average distance
Both approaches suffer from noise. Estimating the path length using Equation 11 may be inaccurate because it randomly picks a pair (si, sj) which may not be really representative. Equation 12 may be problematic, especially since it might be susceptible to noise, due to the possibility of including long- distances with large stochastic errors.
To overcome these problems, we select only min(k, |Si|) and min(k, |Sj|) closest leaves to the root of Ti and Tj with respect to the path length, respectively. To illustrate, for k = 3 we pick {1, 2} from Ti and {4, 5, 6} from Tj in Figure 7.
We now define as the set of min(k, |Si|) closest leaves to the root of Ti. is called the k-representative leaf set. Hereafter, we estimate similar to Equation 4 the path length from the mrca of (Ti, Tj) to med as
which is only based on the k-representative leaf sets. Now we can perform the clustering algorithm with reduced complexity. However, we also want to improve the preorder traversal procedure. The average path length from the unclassified object set S0(Sij) to the mrca of (Ti, Tj) is estimated by Equation 10 which also suffers from noise. To overcome this problem, we select only min(k, |S0(Sij)|) unclassified objects closest to the root of tree T{ij} with respect to distances where s0 ∈ S0(Sij). We call the subset, denoted (Sij), k-representative unclassified object set.
We now define a shortest triplet, which contains three representatives of the three k-representative sets. By construction, are close to the root of T{ij} and close to each other. Therefore, the edge length estimates based on shortest triplet {} are less susceptible to estimation errors.
We now rewrite Equation 10 to estimate the average path length from the representative unclassified object set (Sij) to the mrca of (Ti, Tj) using only shortest triplets as
In short, the preorder traversal procedure uses only shortest triplets to estimate path lengths as well as edge lengths.
Shortest triplet clustering algorithm (STC)
We introduce now the shortest triplet clustering algorithm by combining the clustering algorithm, the local rearrangement, the k-representative sets, and the shortest triplets approach.
Shortest triplet clustering algorithm (STC)
• Initial step:
- (i): Find the median object med using Equation 7.
- (ii): Set = {T1,..., Tn} - {Tmed} and for each Ti ∈ its representative leaf set = {i}.
- (iii): Find for each Ti ∈ its potential neighbor Ti' ∈ using Equation 8.
• Selection step (largest path length criterion): Find the neighboring pair (, ) using Equation 9.
• Agglomeration step:
- (i): Combine and into a new rooted tree with root i0j0, and estimate new edge lengths of using Equation 5 based on shortest triplets.
- (ii): Compute the k-representative leaf set of . based on k-representative leaf sets and of and , respectively.
- (iii): Compute the k-representative unclassified object set of .
- (iv): Delete and and add to .
- (v): Find the potential neighbor for the new rooted tree using Equation 8 based on representative sets, and replace Ti' for each Ti ∈ by if is its potential neighbor.
• Local rearrangement step: Apply the preorder traversal procedure to the rooted subtrees and of the new rooted tree based on only shortest triplets.
• Stopping step: If || > 1, goto Selection step, otherwise output the tree.
The complexity of STC
Now we briefly describe the complexity of the STC. At the initial step, (i), (ii), and (iii) are done in O(n2), O(n) and O(n2) time, respectively. Thus, the complexity of the initial step is O(n2). The selection step is done in O(n). At the agglomeration step, (i), (ii), (iii), (iv), and (v) are done in O(k3), O(k), O(nk2), O(1), and O(nk2) time, respectively. Thus, the complexity of the agglomeration step is O(nk2 + k3). Finally, we are estimating the complexity of the preorder traversal procedure based on only shortest triplets. Step 1 is done in constant time. Step 2, the nearest neighbor interchange operation around the root of Ti costs O(k3). Estimating new edge lengths is done in O(k3) time. Re-computing the k-representative leaf set of Ti based on k-representative leaf sets of its rooted subtrees Tx and Ty costs O(k) time. Finally, re-computing the k-representative unclassified object set (Si) of Ti based on the k-representative leaf set of Tj and the k-representative unclassified object set (Sij) of T{ij} is done in O(k) time. Thus, the complexity of step 2 is O(k3). Step 3 is done in constant time. Step 1, step 2, and step 3 are repeated O(n) times so the complexity of the preorder traversal procedure is O(nk3).
In the STC algorithm, the selection step, the agglomeration step and the local rearrangement step are repeated (n - 2) times so the overall complexity of the STC algorithm is O(n2k3). Practically, we chose k = 5 as a good compromise between the accuracy and computational complexity for all data sets. That is, the practical complexity of the STC algorithm is only O(n2).
Authors' contributions
Both authors participated in the design of the study and performed the analysis. LSV implemented the algorithms. Both authors wrote and approved the final manuscript.
Acknowledgements
We would like to express special thanks to Heiko Schmidt for his technical support. We thank Gunter Weiss, Ingo Ebersberger, Tanja Gesell and Jutta Buschbom for carefully reading the manuscript. We acknowledge the use of supercomputing resources of the ZAM/NIC at the Research Center Jiilich. We thank three anonymous referees for helpful comments.
Figures and Tables
Figure 1 The comparisons of topological accuracy between STC, NJ and BME for the large simulation. Each point in the graph presents the Robinson and Foulds (RF) distance for a simulated data set. Points above the dotted line are examples where the RF distance of the STC-tree is less than the RF distance of the NJ-tree or BME-tree. Thus, the STC gives higher topological accuracy than NJ or BME with respect to the simulated data set.
Figure 2 The comparisons of topological accuracy between STC, NJ and BME for the rbcl simulation. Each point in the graph presents the Robinson and Foulds (RF) distance for a simulated data set. Points above the dotted line are examples where the RF distance of the STC-tree is less than the RF distance of the NJ-tree or BME-tree. Thus, the STC gives higher topological accuracy than NJ or BME with respect to the simulated data set.
Figure 3 The impact of the number of representatives k. The RF distance of STC decreases when k grows from 1 to 5. When k ranges from 5 to 10, the RF distance remains more or less unchanged. For k ≥ 10, the RF distance increases steadily indicating a loss of accuracy.
Figure 4 On the left, estimation of edge lengths ℓ(ar|abc), ℓ(br|abc) and ℓ(cr|abc) of the triplet tree {a, b, c}. On the right, estimation of path length ℓ(s0r|s0s1s2) and edge lengths ℓ(r1r|s0s1s2), ℓ(r2r|s0s1s2) based on the triplet tree {s0, s1, s2}.
Figure 5 The tree is rooted at leaf 5. In the tree, leaves 3 and 4 with the largest path length from their most recent common ancestor to the root 5 are neighbors.
Figure 6 Reconstruction of new rooted tree T{ij} using the the preorder traversal procedure based on the largest average path length criterion. If (Tx, Ty) is the neighboring pair, we stick to the suggested grouping of Ti and Tj (see Figure 6a). Otherwise, if (Tx, Tj) or (Ty, Tj) is the neighboring pair, we switch to the trees displayed in Figure 6b or 6c, respectively.
Figure 7 we select only min(k, |Si|) and min(k, |Sj|) closest leaves to the root of Ti and Tj with respect to the path length, respectively, i.e. for k = 3 we pick {1, 2} from and {4, 5, 6} from . The leaf set {1, 2} (or {4, 5, 6}) is the 3-representative leaf set of the rooted subtree . (or ).
Table 1 The average Robinson and Foulds distance of 100 simulated data sets of 500 sequences each with sequence lengths 500, 1000 and 2000 nt (rbcl simulation). Methods are used without BNNI.
sequence length NJ BIONJ Weighbor HGT/FP GME BME STCk = 5
500 .190 .188 .194 .512 .240 .184 .177
1000 .100 .098 .099 .409 .144 .096 .088
2000 .049 .048 .050 .313 .082 .046 .040
Table 2 The average Robinson and Foulds distance of 100 simulated data sets of 500 sequences each with sequence lengths 500, 1000 and 2000 nt (rbcl simulation). Methods are used with BNNI.
sequence length NJ BIONJ Weighbor HGT/FP GME BME STCk = 5
500 .162 .162 .162 .166 .163 .163 .162
1000 .079 .079 .079 .079 .080 .079 .079
2000 .035 .035 .035 .036 .036 .035 .035
Table 3 The average Robinson and Foulds distance of 100 simulated data sets of 1000 taxa for each tree diameter 0.1, 0.5, 1.0 and 1.5 and with sequence length 1000 nt (large simulation). Methods are used without BNNI.
number sequences NJ BIONJ HGT/FP GME BME STCk = 5
1000 (0.1) .146 .146 .378 .168 .143 .139
1000 (0.5) .093 .089 .193 .126 .075 .066
1000 (1.0) .094 .090 .188 .132 .074 .062
1000 (1.5) .097 .091 .182 .138 .073 .061
Table 4 The average Robinson and Foulds distance of 100 data sets of 5000 taxa for each tree diameter 0.1, 0.5, 1.0 and 1.5 and with sequence length 1000 nt (large simulation). Methods are used without BNNI.
number sequences NJ BIONJ HGT/FP GME BME STCk = 5
5000 (0.1) .178 .179 .442 .207 .173 .170
5000 (0.5) .109 .105 .210 .156 .084 .072
5000 (1.0) .107 .102 .192 .155 .073 .064
5000 (1.5) .112 .106 .188 .164 .072 .063
Table 5 The average Robinson and Foulds distance of 100 simulated data sets of 1000 taxa for each tree diameter 0.1, 0.5, 1.0 and 1.5 and with sequence length 1000 nt (large simulation). Methods are used with BNNI.
number sequences NJ BIONJ HGT/FP GME BME STCk = 5
1000 (0.1) .137 .137 .137 .137 .137 .138
1000 (0.5) .061 .061 .061 .061 .061 .061
1000 (1.0) .057 .057 .057 .057 .057 .056
1000 (1.5) .055 .055 .055 .055 .055 .055
Table 6 The average Robinson and Foulds distance of 100 data sets of 5000 taxa for each tree diameter 0.1, 0.5, 1.0 and 1.5 and with sequence length 1000 nt (large simulation). Methods are used with BNNI.
number sequences NJ BIONJ HGT/FP GME BME STCk = 5
5000 (0.1) .168 .168 .168 .168 .168 .168
5000 (0.5) .066 .066 .066 .066 .066 .066
5000 (1.0) .057 .057 .057 .057 .057 .057
5000 (1.5) .055 .055 .055 .055 .055 .055
Table 7 The average RF distance of the 96-taxon alignments of sequence length 500 nt, that were analyzed in [6]. The 6000 trees were split into three groups called "slow" (0.2 substitutions per site), "moderate" (0.4 substitutions per site) and "fast" (1.0 substitutions per site). Except for STC, the accuracies for the other methods were taken from [6]. Methods are used without BNNI.
number sequences NJ BIONJ Weighbor HGT/FP GME BME STCk = 5
96 (slow) .183 .180 .178 .512 .199 .186 .179
96 (moderate) .136 .134 .129 .480 .158 .137 .125
96 (fast) .115 .112 .103 .465 .144 .117 .102
Table 8 The average RF distance of the 96-taxon alignments of sequence length 500 nt, that were analyzed in [6]. The 6000 trees were split into three groups called "slow" (0.2 substitutions per site), "moderate" (0.4 substitutions per site) and "fast" (1.0 substitutions per site). Except for STC, the accuracies for the other methods were taken from [6]. Methods are used with BNNI.
number sequences NJ BIONJ Weighbor HGT/FP GME BME STCk = 5
96 (slow) .173 .173 .173 .175 .173 .173 .173
96 (moderate) .119 .118 .118 .123 .118 .118 .116
96 (fast) .090 .090 .091 .098 .091 .090 .090
Table 9 The percentage of cases where STC is at least as good as other tested methods in terms of RF distance. The number in parentheses is the percentage of cases where STC is equally good as other tested methods. Methods are used without BNNI.
number sequences NJ BIONJ Weighbor HGT/FP GME BME
96 (500 nt) 68 (16) 65 (15) 57 (16) 100 (0) 73 (10) 70 (14)
500 (500 nt) 80 (4) 76 (4) 88 (3) 100 (0) 100 (0) 66 (1)
500 (1000 nt) 88 (3) 79 (4) 84 (4) 100 (0) 100 (0) 83 (6)
500 (2000 nt) 92 (6) 90 (4) 92 (3) 100 (0) 100 (0) 86 (9)
1000 (1000 nt) 95 (2) 95 (1) n.d. 100 (0) 100 (0) 89 (15)
5000 (1000 nt) 100 (0) 99 (0) n.d. 100 (0) 100 (0) 95 (1)
Table 10 The percentage of cases where STC is better than other tested methods in terms of RF distance. The number in parentheses is the percentage of cases where STC is worse than other tested methods. Methods are used with BNNI.
number sequences NJ BIONJ Weighbor HGT/FP GME BME
96 (500 nt) 9 (8) 8 (8) 10 (10) 12 (10) 10 (8) 10 (9)
500 (500 nt) 34 (37) 35 (39) 35 (36) 59 (29) 46 (33) 41 (39)
500 (1000 nt) 22 (19) 17 (23) 18 (22) 23 (28) 30 (20) 24 (20)
500 (2000 nt) 10 (13) 8 (7) 10 (8) 9 (8) 12 (10) 7 (10)
1000 (1000 nt) 30 (28) 27 (29) n.d. 28 (22) 30 (24) 28 (27)
5000 (1000 nt) 48 (40) 42 (44) n.d. 45 (45) 52 (37) 43 (43)
==== Refs
Saitou N Nei M The Neighbor – joining Method: A New Method for Reconstructing Phylogenetic Trees Mol Biol Evol 1987 4 406 425 3447015
Strimmer K von Haeseler A Quartet Puzzling: A Quartet Maximum – Likelihood Method for Reconstructing Tree Topologies Mol Biol Evol 1996 13 964 969
Gascuel O BIONJ: An Improved Version of the NJ Algorithm Based on a Simple Model of Sequence Data Mol Biol Evol 1997 14 685 695 9254330
Huson DH Nettles SM Warnow TJ Disk-Covering, a Fast-Converging Method for Phylogenetic Reconstruction J Comput Biol 1999 6 369 386 10582573 10.1089/106652799318337
Bruno WJ Socci ND Halpern AL Weighted Neighbor Joining: A Likelihood Based-Approach to Distance-Based Phylogeny Reconstruction Mol Biol Evol 2000 17 189 197 10666718
Desper R Gascuel O Fast and Accurate Phylogeny Reconstruction Algorithms Based on the Minimum-Evolution Principle J Comput Biol 2002 9 687 706 12487758 10.1089/106652702761034136
Csürös M Fast Recovery of Evolutionary Trees with Thousands of Nodes J Comput Biol 2002 9 277 297 12015882 10.1089/10665270252935467
Guindon S Gascuel O A Simple, Fast and Accurate Algorithm to Estimate Large Phylogenies by Maximum Likelihood Syst Biol 2003 52 696 704 14530136 10.1080/10635150390235520
Stamatakis A Ludwig T Meier H RAxML-III: A fast program for maximum likelihood-based inference of large phylogenetic trees Bioinformatics 2005 21 456 463 15608047 10.1093/bioinformatics/bti191
Vinh LS von Haeseler A IQPNNI: Moving fast through tree space and stopping in time Mol Biol Evol 2004 21 1565 1571 15163768 10.1093/molbev/msh176
Vinh LS Schmidt HA von Haeseler A PhyNav: A Novel Approach to Reconstruct Large Phylogenies Proceedings of the 28th Annual German Classification Society Conference (GfKl 2004) 2004 Dortmund, Germany
Charleston MA Hitch-Hiking: A Parallel Heuristic Search Strategy, Applied to the Phylogeny Problem J Comput Biol 2001 8 79 91 11339908 10.1089/106652701300099137
Brauer MJ Holder MT Dries LA Zwickl DJ Lewis PO Hillis DM Genetic Algorithms and Parallel Processing in Maximum-Likelihood Phylogeny Inference Mol Biol Evol 2002 19 1717 1726 12270898
Schmidt HA Strimmer K Vingron M von Haeseler A TREE-PUZZLE: Maximum likelihood phylogenetic analysis using quartets and parallel computing Bioinformatics 2002 18 502 504 11934758 10.1093/bioinformatics/18.3.502
Schmidt HA von Haeseler A Baxevanis AD, Davison DB, Page RDM, Stormo G, Stein L Maximum Likelihood Analysis Using TREE-PUZZLE Current Protocols in Bioinformatics 2003 New York, USA: Wiley and Sons 6.6.1 6.6.25
Cavalli-Sforza L Edwards AWF Phylogenetic analysis: Models and estimation procedures Am J Hum Genet 1967 19 233 257 6026583
Fitch W Margoliash E Construction of Phylogenetic trees Science 1967 155 279 284 5334057
Hartigan AJ Clustering Algorithms 1975 John Wiley and Sons, Inc
Farris J On the phenetic approach to vertebrate classification 1977 17 823 850
Klotz NKRB LC Mitchell RM Calculation of evolutionary trees from sequence data Proc Natl Acad Sci USA 1979 76 4516 4520 291984
Li WH Simple method for constructing phylogenetic trees from distance matrices Proc Natl Acad Sci USA 1981 78 1085 1089 6940127
Rambaut A Crassly NC Seq-Gen: An application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees Comput Appl Biosci 1997 13 235 238 9183526
Kimura M A Simple Method for Estimating Evolutionary Rates of Base Substitutions through Comparative Studies of Nucleotide Sequences J Mol Evol 1980 16 111 120 7463489
Harding EF The probabilities of rooted tree-shapes generated by random bifurcation Adv Appl Prob 1971 3 44 77
Swofford DL Olsen GJ Waddell PJ Hillis DM Hillis DM, Moritz C, Mable BK Phylogeny Reconstruction Molecular Systematics 1996 2 Sunderland, Massachusetts: Sinauer Associates 407 514
Felsenstein J PHYLIP (Phylogeny Inference Package) version 3.5c Department of Genetics, University of Washington, Seattle 1993
Robinson DR Foulds LR Comparison of phylogenetic trees Mathematical Biosciences 1981 53 131 147 10.1016/0025-5564(81)90043-2
Buneman P Hodson, Lendall, Tautu The recovery of trees from measures of dissimilarity Mathematics in the archaeological and historical sciences 1971 Edinburgh: Edinburgh university press
Semple C Steel M Phylogenetics 2003 OXFORD univerity press
Felsenstein J Inferring Phylogenies 2004 Sunderland, Massachusetts: Sinauer Associates
Aho AV Hopcroft JE Ullman JD The Design and Analysis of Computer Algorithms 1974 Addison-Wesley Publishing Company
| 15819989 | PMC1097715 | CC BY | 2021-01-04 16:02:48 | no | BMC Bioinformatics. 2005 Apr 8; 6:92 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-92 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-991583110510.1186/1471-2105-6-99Research ArticleImproved profile HMM performance by assessment of critical algorithmic features in SAM and HMMER Wistrand Markus [email protected] Erik LL [email protected] Center for Genomics and Bioinformatics, Karolinska Institutet, S-17177 Stockholm, Sweden2005 15 4 2005 6 99 99 1 2 2005 15 4 2005 Copyright © 2005 Wistrand and Sonnhammer; 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
Profile hidden Markov model (HMM) techniques are among the most powerful methods for protein homology detection. Yet, the critical features for successful modelling are not fully known. In the present work we approached this by using two of the most popular HMM packages: SAM and HMMER. The programs' abilities to build models and score sequences were compared on a SCOP/Pfam based test set. The comparison was done separately for local and global HMM scoring.
Results
Using default settings, SAM was overall more sensitive. SAM's model estimation was superior, while HMMER's model scoring was more accurate. Critical features for model building were then analysed by comparing the two packages' algorithmic choices and parameters. The weighting between prior probabilities and multiple alignment counts held the primary explanation why SAM's model building was superior. Our analysis suggests that HMMER gives too much weight to the sequence counts. SAM's emission prior probabilities were also shown to be more sensitive. The relative sequence weighting schemes are different in the two packages but performed equivalently.
Conclusion
SAM model estimation was more sensitive, while HMMER model scoring was more accurate. By combining the best algorithmic features from both packages the accuracy was substantially improved compared to their default performance.
==== Body
Background
Computational protein sequence homology detection has become a central component in genome analysis. Today sequences of unknown function are routinely searched against databases of known proteins and this has become an important aid for sequence annotation and to guide laboratory experiments. Without the development of software tools for the detection of protein homology from amino acid sequence this would not have been possible. Such homology detection tools aim to find similarities between related proteins and to score them above the noise level. Different methods have varying degrees of success and it has been shown that profile-based methods, which consider information from a number of sequences, perform better than pairwise methods[1]. In particular, profile Hidden Markov models (profile HMMs)[2,3] have generated good results, and are today employed by several databases. Pfam[4] and Superfamily[5] for example, are large collections of protein families where each family is represented by a profile HMM. The profile HMM is a probabilistic model of a multiple sequence alignment of the family and is used to represent the family in database searches.
Several aspects of profile HMM technology have been further developed since its initial conception. Various schemes for sequence weighting have been proposed[6,7] and different null models have been studied[8]. The introduction of Dirichlet mixtures to model prior distributions[9,10] constituted a major step forward. The maximum likelihood technique employed to estimate such prior distributions has, however, been shown to be problematic for transition priors[11]. Discriminative training has been incorporated into model building and been shown to give improved performance[12,13]. Methods that incorporate phylogenetic information directly into the profile HMM and bypass the need for ad-hoc sequence weighting, have been developed and proved promising for homology detection [14,15]. Secondary structure prediction has been combined with profile HMMs into a probabilistic framework for more accurate fold recognition[16,17]. Finally, explicitly including knowledge about the taxonomic distribution of protein domains has proved to enhance protein domain discovery[18], as has the incorporation of knowledge about the likelihood of certain domain combinations[19]. These examples of HMM improvements are only a few and by no means a complete listing.
Two widely used profile HMM packages are SAM[2] and HMMER[3]. It is of interest for users to know the relative performance of the programs, and for profile HMM developers to know the key factors for good performance. Madera and Gough contributed the most recent and still most thorough comparison of the two programs[20]. The authors divided their analysis into the two main steps of profile HMM homology detection: model building and database searching. Model building involves converting a multiple alignment of the family into a probabilistic model, while database searching involves scoring a sequence to the profile HMM. The two steps are independent and by using a program to convert HMMER models into SAM format and vice versa, Madera and Gough were able to separately evaluate the building and scoring performance of the two programs. SAM model building was found to be clearly superior to HMMER model building, while no conclusion could be made concerning the scoring algorithms. E-value calculation, low complexity masking and time consumption were also investigated, but neither of the packages stood out as clearly better than the other.
Profile HMMs often model complete protein domains while real proteins may contain several domains. It therefore makes sense to look for a local match of the protein to the HMM. "Global/local matching" forces the entire HMM to match a part of the sequence. This is often the desired mode if the HMM is built from one domain and the entire domain can be expected also in other proteins, perhaps in combination with other domains. In contrast, "local/local matching" means that a part of the HMM is matched locally to the sequence. The choice between the two modes depends on the application. Local/local searches can find fragmentary protein sequences that would get poor scores if they were forced to match the entire model. However, in case the query sequences contain full domains, the sensitivity of the global/local mode should be better.
Madera and Gough compared the packages only for local/local mode. The first of the two objectives of this article is to extend their analysis to global/local searches. The second objective is to find the key features for profile HMM performance with a particular focus on model building. Profile HMM estimation involves choices concerning for example sequence weighting, prior probabilities, and model architecture, and the two programs approach these issues differently. By comparing SAM and HMMER run with non-default settings and with parameters borrowed from each other, we show which model parameters are crucial for profile HMM performance.
The article has the following structure. First we introduce the SAM and HMMER packages and explain the role of the model building components that we will investigate. Second we go through the test procedure, the data sets and the performance measure that we use. In the result section we compare the packages and analyse the influence of algorithmic components and parameters on the HMM performance in terms of sensitivity and specificity.
Profile HMMs – HMMER, SAM and relevant parameters
The SAM package comes from the University of California Santa Cruz. The package includes the SAM-T2K iterative procedure to generate multiple alignments and HMMs starting from a single sequence[21]. Another feature is "multi-track HMMs", to process more than just the primary sequence data. Secondary structure information can, for example, be incorporated in a probabilistically sound way for better modelling[17,22]. This article will not evaluate these two additional features (both lacking in HMMER), but deals with traditional profile HMMs built from multiple alignments. The SAM package is used by the SUPERFAMILY[5] database. HMMER is developed by Sean Eddy and is open-source, well documented and easy to use. HMMER and the protein family database Pfam have co-evolved, but today HMMER is the engine also in other databases, including TIGRFAMs[23] and SMART[24].
In this study we used SAM version 3.4 and HMMER version 2.3.2.
HMM architecture and construction
A profile HMM is a probabilistic model of a multiple alignment of related proteins. The alignment is modeled using a series of nodes (roughly one per alignment column) each composed of three states: match, insert and delete. Match and insert states emit amino acids with probabilities learned during model estimation while delete states are quiet. Insertions and deletions with respect to the HMM are modeled by insert and delete states and transition probabilities to them. The original profile HMM architecture allowed transitions between all states, which gives 3 × 3 = 9 possible transitions for each node. SAM has kept this architecture, while HMMER since version 2.0 allows seven transitions only. In the HMMER architecture transitions from insert to delete states, and vice versa, are forbidden.
Both HMMER and SAM allow the user to "label" columns in the multiple alignment to tell the program which columns should be seen as match/delete states and which should be seen as insert states. In case such information is missing, SAM will assign every column to a match state and produce an HMM with one node per column. In HMMER an algorithm will assign columns to match or insert states so as to maximize the posterior probability of the aligned sequences, given the model. Compared to SAM, this normally results in shorter models since some of the columns are assigned to insert states.
Prior probability alternatives
Profile HMM parameters are estimated by combining the observed data (the multiple alignment) and a prior over probability distributions. If the multiple alignment were a good representation of the underlying protein family, there would be less need for using a prior. However, this is often not the case, primarily because the alignment includes too few sequences. The strength of priors is in compensating for small sample size and to distribute probability also to unseen events.
SAM and HMMER both use Dirichlet mixtures to model emission prior probabilities[9,10]. A Dirichlet mixture consists of a number of Dirichlet components, which are distributions over probability parameters. Each component typically captures a specific feature of columns observed in multiple sequence alignments, i.e. hydrophobicity or polarity, but also the degree of conservation to certain amino acids. During model building, the Dirichlet components are weighted probabilistically for each column in the multiple alignment (optimally given the amino acid frequencies of the column) and combined with the observed amino acid frequencies to obtain the posterior emission probabilities. The default emission prior in SAM is currently a mixture of 20 components, while HMMER's default is a mixture of nine components. Transition prior probabilities are modeled by single distributions in both SAM and HMMER, but differ in two ways. First because SAM employs nine parameters and HMMER only seven, i.e. one per transition. Second because SAM assigns a higher prior probability to deletions and insertions than HMMER.
Sequence weighting: relative and total weights
The weight assigned to a sequence determines its influence on the final HMM. Without sequence weighting, a high redundancy among the sequences would make the model skewed and it would not recognize under-represented sequences. The sequence weights are calculated in a two-step process: relative weights are first determined and then scaled to sum to the total weight ("effective sequence number" in HMMER terminology), which is calculated separately.
The relative weights determine the influence of one sequence relative to the others. There are several algorithms for relative sequence weighting and common for all of them is to assign less weight to similar sequences and more to the divergent but still trusted family members. The HMMER default algorithm derives weights from a sequence tree relating the sequences[25], while SAM uses an unpublished algorithm based on relative entropy (SAM documentation). Relative weights do not sum to any particular value, but are scaled to add up to the total weight. The total weight thus governs the weight of the multiple sequence alignment relative to the prior probabilities. HMMER and SAM calculate the total weight in two very different ways. HMMER applies an algorithm that groups sequences by single-linkage clustering and counts the number of clusters above a specified level of identity. SAM scales the weights according to an entropy target that specifies the number of bits per column to save during model building, i.e. the information content of the final model compared to a background model.
Global or local scoring
A sequence can be scored locally to the entire profile HMM (global/local) or to a part of it (local/local). In HMMER, the search mode is specified in the HMM at the time of model building. Two HMMs can thus be built from the same alignment, one global/local and one local/local, and both are scored using the same algorithm. SAM estimates only one HMM for each multiple alignment, and the search mode is instead specified at the time of model scoring.
A SCOP/Pfam based benchmark
Data sets
A large number of studies have used the SCOP[26] structural classification for evaluating the performance of sequence homology detection methods[13,18,20,27,28]. SCOP is a database classifying all protein domains of known structure into a hierarchical order of four levels: class, fold, superfamily and family. Two domains belong to the same family if they have a clear common evolutionary origin revealed either by a minimum of 30% sequence identity or very similar structure and function. Two domains belong to the same superfamily but different families, if a common evolutionary origin is not obvious from sequence identity, but probable from an analysis of structure and functional features. The fold level is grouping all superfamilies and families that have a common pattern of secondary structure elements. Finally, the class level divides domains into large classes based on secondary structure components.
In this work we evaluated the performance of profile HMMs for homology detection at the superfamily level. We wanted to avoid conditioning the results on the use of a particular program to generate the multiple alignment. Following Coin et al[18], we therefore developed a test set that combines the high quality Pfam alignments and the SCOP classification. Pfam is a database of protein families based on sequence similarity rather than structural similarity. A manually curated sequence alignment is provided for each family, as is a profile HMM to search for homologs. We used the ASTRAL data set filtered to a maximum of 40% sequence identity for the SCOP sequence classification[29]. ASTRAL is a database derived from SCOP and provides sequence files filtered to various levels.
To generate the test dataset, Pfam families were linked to the superfamily level in the SCOP classification. We kept all Pfam families that contain sequences that belong to one and only one SCOP superfamily. We also required that the SCOP domain definition for at least one of the sequences spanned the entire Pfam domain. Using Pfam 15.0 and the ASTRAL dataset this gave a test set of exactly 1400 families. We imposed two extra conditions: that the Pfam seed alignment had at least 10 sequences and that the average sequence length was above 30 amino acids. This gave 1009 families from which we extracted every second family in alphabetical order to get a large enough but yet computationally feasible set of 505 families. All in all, the dataset contains 9 411 positive and 2 842 994 negative test sequences.
Test procedure
The test procedure was the following. We built profile HMMs from the seed alignments of the 505 Pfam families. We scored the entire ASTRAL dataset to the HMMs and classified the matches from their SCOP classification. Matches to the SCOP superfamily mapped to the query Pfam family were classified as true hits. Matches to a different SCOP fold were classified as false hits, while matches to the same SCOP fold as the query but a different superfamily were ignored. For each HMM, the searching generated a list of hits that we sorted on E-values. We went through this list from top to bottom and for each level of false positives we recorded the number of true positives. This gave a plot of true positives versus false positives, which is the standard way of displaying results from this type of tests.
Results and discussion
Default settings
SAM and HMMER were compared for both local/local and global/local mode. We first ran the packages in local/local mode using all default settings, except that SAM scoring was performed by the Viterbi algorithm. Figure 1a shows that SAM performed considerably better than HMMER; building and scoring using SAM detected more true positives than building and scoring using HMMER, and this was true across all error rates. However, the best results were obtained when SAM models were converted to HMMER models and scored by HMMER. In contrast, HMMER models converted to SAM models followed by SAM scoring produced the worst results. It is thus fair to say that model estimation is considerably better done by SAM, while scoring is better done by HMMER.
Figure 1b shows the corresponding results for global/local mode. Here SAM and HMMER produced nearly identical results and no conclusion could be drawn as to which is the better package. Splitting the performance into building and scoring, it became evident that HMMER scoring was the best choice for both HMMER and SAM models. Consequently, SAM's model building was more accurate than HMMER's. In agreement with local/local mode, SAM model estimation was superior, but for global/local mode this advantage was fully compensated by HMMER's more accurate scoring program.
It is striking how much worse HMMER performed in local/local mode compared to global/local. In contrast, the SAM results for local/local mode were very close to those for global/local mode. Remember that the two packages have solved the issue of global or local scoring in different ways: while HMMER has two separate models and one way of scoring, SAM has one model and two ways of scoring. Could it be that the HMMER local/local model architecture, rather than the actual parameter estimation, is causing the poor local/local performance? If this were the case, SAM models converted to HMMER format and configured for local/local scoring should be less accurate than SAM models converted to HMMER format and configured for global/local scoring. This was not the case in our test (compare the SAM-HMMER curves in Figure 1a and 1b). Instead, the reason must be poorer model building in HMMER than in SAM, affecting local/local models more than global/local models.
To conclude this section we note that although HMMER proved comparable to SAM for global/local scoring, SAM is the preferred package as it performed much better in local/local mode. SAM produced better models, but lost some of the advantage due to an inferior scoring program. While HMMER model building was underperforming overall, local/local models proved particularly poor. In what follows we will seek explanations to these differences by analysing the effect of relevant model estimation parameters and algorithmic choices.
Prior probability options
The default SAM amino acid emission prior (recode3.20comp) has more than twice the number of free parameters compared to the default HMMER emission prior (20 and 9 component mixtures, respectively). We ran HMMER using recode3.20comp on our test. This gave an increase in performance both for global/local and local/local models, showing that the emission prior is important in explaining why SAM model building is more sensitive (Figure 2).
We also investigated the role of the transition prior. This is not as straightforward since the HMMER transition prior has only seven parameters and the SAM prior has nine; the delete-insert and insert-delete transitions are non-existing in the HMMER architecture. Nevertheless, we ran HMMER with the SAM transition prior ignoring the two superfluous insert-delete parameters, and SAM with the HMMER transition prior plus the insert-delete parameters from the SAM prior. In local/local mode we could see almost no effect of using a foreign transition prior (Figure 3a), but in global/local mode the performance deteriorated considerably (Figure 3b).
The test of prior probabilities thus revealed that the SAM emission prior is an important factor to explain SAM model building superiority, while the default transition priors were program-specific for global/local mode. In all subsequent HMMER experiments, we used the SAM emission prior in order to reduce the difference in parameter settings.
Sequence weighting
Sequence weighting involves (a) the relative weight assigned to each sequence and (b) the total weight given to all sequences as a group. While the relative weights determine the influence of one sequence relative to the others, the total weight gives the influence of the sequences vis-à-vis the prior probabilities. In addition, SAM model building involves a filter that reduces the number of training sequences such that no two sequences have more than 80% sequence identity. Excluding the filter had no important impact on SAM results (data not shown), hence the filter was removed in subsequent runs.
We analysed whether differences in sequence weighting could be a source for package-specific results. First we turned off both the relative weighting scheme and the total weight calculation in both packages. The effect of these changes is that each sequence gets a weight of 1.0, such that all sequences will be equally important and the total weight will be the number of training sequences. These changes had a negative effect on both packages, but the effect was much worse for SAM (data not shown). Two conclusions could be drawn: 1) Sequence weighting does play a major role for performance, and 2) the SAM weighting procedure is more important for performance than the HMMER weighting.
Would SAM weighting work better also for HMMER? To answer this question the HMMER code was modified to read sequence weights from file, with the option to rescale those weights according to HMMER's total weight calculation. We let SAM generate weights and used them in HMMER model building. For local/local models this had a very large effect and sensitivity improved greatly when using SAM weights instead of HMMER weights (Figure 4a).
We next analysed what makes SAM weights better: the relative weighting algorithm or the total weight calculation. In order to answer this we needed to isolate the effect of the two weighting components. We let HMMER read SAM weights from file but rescaled them by the HMMER total weight; in this way HMMER was run using the SAM relative weighting algorithm but the HMMER total weight calculation. Performance dropped to a level comparable to all-HMMER weighting (Figure 4a), which indicates that the SAM total weight calculation is the crucial factor. To verify this we implemented our own version of the SAM "bits saved" method for total weight calculation in the HMMER code (see Methods). We used the SAM default target value of 0.5 bits saved relative to the background distribution. Using HMMER relative weighting and the "bits saved" method produced as good results as using SAM weights. The conclusion is that the SAM "bits saved" method for calculating the total weight is much better than the HMMER method and a main source of the difference in performance, while the schemes for relative weight calculation are essentially equivalent.
The previous tests were all done for local/local scoring. For global/local scoring the picture was less clear. Running HMMER with SAM weights in global/local mode decreased performance (Figure 4b) compared to using HMMER weights, i.e. a result opposite to what we saw for local/local mode. However, when we also added the SAM transition prior, in addition to the SAM emission prior used for all runs, the results were improved (Figure 4b). Remember that the SAM transition prior earlier proved far from optimal for global/local HMMER usage (Figure 3b). Apparently the transition prior and the total weight cannot be specified independently in order to obtain sensitive global/local HMMs.
In global/local mode, SAM sequence weights thus gave more accurate HMMER models provided they were combined with the SAM transition prior. We again split this effect into relative weighting and total weight calculation, and as for local/local scoring, the improvement was entirely due to the SAM method for total weight calculation (Figure 4b).
The total weight calculation emerges from this study as a very important component in profile HMM building. The higher the total weight, the larger will be the influence of the multiple alignment on the HMM, at the expense of the prior probabilities. Is SAM performing better because it assigns more weight or less weight to the multiple alignment, compared to HMMER? To answer this question, we investigated the output of the SAM and HMMER methods for total weight calculation on our test set of 505 Pfam families. SAM produced an average total weight of 11.8 and HMMER an average total weight of 39.8. Profile HMMs by HMMER are thus relatively more determined by the multiple alignment, while SAM gives a stronger influence to the prior probabilities. HMMER's weak performance in our test together with these numbers suggest that HMMER might overfit its models to the training data.
In conclusion, SAM sequence weighting proved more accurate than HMMER weighting. The difference was entirely due to SAM's method for total weight calculation, while the methods for relative weighting seemed to be of equivalent quality. The choice of transition prior had no influence on local/local searches. However, for global/local models, the transition prior employed and the method for calculating total weights could not be chosen independently. The best performance was obtained using SAM total weight and the SAM transition prior. However, if HMMER's transition prior was employed, the HMMER total weight calculation was more appropriate.
Choosing match nodes
HMMER labels columns in the multiple alignment as "match" or "insert" nodes based on an automatic procedure where the overall probability of the sequences is maximised. SAM has no such algorithm and treats every column as a match column, in case nothing else is specified in the alignment. We turned off the HMMER automatic algorithm and made it assign every column to match/delete states, as is SAM default. This had a slightly negative effect on performance, suggesting that the HMMER automatic algorithm is sensible and gives some improvement (data not shown).
Model scoring
As seen in Figure 1, HMMER model scoring is more accurate than SAM's. We believe that the principal reason for the difference lies in the used null model. Both packages calculate log-odds scores, that tell how much better the sequence matches the family-trained model than the null model. The simplest null model is based on the average amino acid frequencies in protein coding sequences. By default, HMMER and SAM use more advanced alternatives designed to compensate for the effect of biased sequence composition. This occurs when a sequence gets a relatively high score only because its overall amino acid composition is close to that of the modelled domain. HMMER compensates for biased composition by correcting the score using a second null model which is calculated as the average over all emission probabilities of the states in the target sequence's path through the model. SAM on the other hand uses the score of the reversed sequence as null model score. Unfortunately, the reversed sequence null model is compulsory for SAM's E-value calculation, hence we were unable to investigate the effect of the null model. In any case we can only improve the free HMMER code, which already seems to have the superior method.
Comparison to earlier work
Madera and Gough carried out a similar benchmark of HMMER and SAM[20]. The authors analyzed local/local mode only and concluded that SAM was better at model building while the results for model scoring were not clear as different tests generated different results. Our study agrees with theirs on model building, but not for model scoring where our results indicate that HMMER is more accurate. From where does this discrepancy stem? Madera and Gough ran the test the way a non-expert user would, i.e. with all default settings. This means forward scoring (sum of all paths) for SAM and Viterbi scoring (single best path) for HMMER (personal communication, Martin Madera). On our test set and for both packages, forward scoring was more accurate than Viterbi scoring, but is a slower algorithm. If the authors had compared similar scoring algorithms they would most likely have concluded that HMMER scoring performs better.
Conclusion
We have presented a comparison of the SAM and HMMER packages. SAM stands out as the better package for building HMMs; particularly so for local/local searches. SAM loses some of this advantage due to a slightly worse performing search algorithm, and for global/local mode the HMMER package was actually at par with SAM. However, if default settings are applied, SAM should be the preferred package.
We furthermore sought for the key factors in profile HMM estimation by analysing what makes SAM build more sensitive HMMs. SAM's emission prior proved clearly superior to HMMER's. The relative sequence weighting schemes of the two packages, although different, proved to perform essentially equivalently. The main effect, however, was due to how prior probabilities and multiple alignment counts are combined. The total sequence weight, which determines the degree of faith in the observed data relative to the prior probabilities, seems to be much better handled by the SAM package. It is generally correct to say that, compared to HMMER, SAM puts more belief in the prior and less in the observed alignment. Our results suggest that HMMER is overfitting models to the observed data while SAM is better utilising the Dirichlet mixture's capability to extrapolate observed amino acids to the underlying distribution.
By dissecting the importance of the different components in HMM building and scoring, we were able to combine the best features of HMMER and SAM into a modified HMMER program that is superior to both programs. The code for this and the test used in the study is freely available from the authors via ftp . Profile HMMs are used by many databases that have a large influence on genome annotation. Improvements to the profile HMM technology will therefore be of potentially large importance, which should render the results presented here valuable for many genome projects.
Looking ahead, a recent development is profile HMM – profile HMM scoring[30] which has showed significantly higher sensitivity than ordinary profile HMM to sequence searches as well as profile – profile[31,32] searches. Profile HMM – profile HMM searches can detect remote homology between two protein families. Alternatively, a multiple alignment can be constructed automatically around a single query sequence; from this a profile HMM can be built and used in a search against for example the Pfam database. The inclusion of homologs in the search improves sensitivity, and one can speculate that profile HMM – profile HMM searches gradually will out-compete profile HMM to sequence searches. Profile HMM estimation, however, is a fundamental issue also for this novel technology and we expect there is room for improvements.
Methods
Total weight calculation
In order to assess the importance of different methods for total weight calculation, we implemented a version of the SAM "bits saved" method in the HMMER code.
This method is unpublished by the original inventors but explained in an article by Edgar and Sjölander[33]. The number of bits saved is the relative entropy between the final HMM and an HMM defined from the Dirichlet mixture prior only (the background model), and is written as:
PD(a) is the background probability of emitting amino acid a, Pj(a) is the emission probability vector of match state j in the HMM and M is the number of match states in the HMM. The summation is over all match states j and all amino acids a. The background probability vector PD(a) is defined by applying the Dirichlet mixture to a zero count vector. During profile HMM estimation, the total weight is adjusted iteratively until the number of bits saved matches a target entropy specified by the user.
Programs and settings
For the comparisons in this paper we used the HMMER programs hmmbuild to build models and hmmsearch to score sequences. To build SAM models we used the w0.5 script, which is recommended for constructing models for homology detection at the superfamily level. The SAM program hmmscore was used to score sequences to SAM models. In order to achieve proper E-values, all models were calibrated using the HMMER program hmmcalibrate and the SAM hmmscore program. For scoring, we employed the Viterbi algorithm. The "forward" algorithm, which is SAM's default, generally produces better results but is much slower and hardly suitable for large scale database searches.
Conversion between model formats
In order to isolate the performance of the build and search components of SAM and HMMER, we converted HMMs models between the two model formats. Models built by one program can thus be used for searching by both packages' scoring programs independently. Also, models from both packages can be converted to the same model format and scored using the same scoring program. For the conversion we used a program developed by Madera and Gough, with our own minor modifications. Since the original code only converts between SAM models and local/local HMMER models, we extended it to also allow conversions to global/local HMMER models.
The number of allowed transitions differs between the two model formats, which makes a complete mapping between them impossible. The conversion program handles this as follows. In SAM to HMMER conversions, the two extra SAM delete-insert transition parameters are simply omitted, causing a small information loss. In HMMER to SAM conversions, no information is lost, and converting a model from HMMER to SAM format and again back to HMMER format will re-create the original model, provided it is configured for the same score mode. According to Madera and Gough, testing indicated that the two extra SAM transitions are redundant and the information loss in SAM to HMMER conversions of no importance[34].
Authors' contributions
MW wrote the code for the analysis, designed the test set and performed all experiments. ES participated in the design of the study. Both authors collaborated in writing the final version.
Acknowledgements
This work was supported by grants from Pfizer Inc. and the Swedish Knowledge Foundation. We thank Carsten Daub and Lukas Käll for valuable comments on the manuscript.
Figures and Tables
Figure 1 Model building and database searching performance of SAM and HMMER. (A) Local/local mode. (B) Global/local mode. The Viterbi algorithm was used for all searches. Otherwise, default settings were used. The model building program is mentioned first in the legend, and the scoring program second. 'SAM-SAM' means 'Building the HMM using SAM; Searching using SAM', 'HMMER-SAM' means 'Building the HMMs using HMMER; Searching using SAM', etc.
Figure 2 Analysis of emission prior probabilities. HMMER models were estimated using the SAM emission prior recode3.20.comp. Both local/local and global/local models were improved compared to using the HMMER default emission prior. Part of the SAM – HMMER performance difference can therefore be assigned to the emission prior.
Figure 3 Analysis of transition prior probabilities. (A) Local/local mode. (B) Global/local mode. SAM and HMMER models were built using their own default transition prior, and with default transition prior from the other program. The transition priors seem to be specific to their programs as they degrade the performance of the other program.
Figure 4 Analysis of relative and total sequence weight calculation methods. (A) Local/local mode. (B) Global/local mode. HMMER was run with relative and total sequence weights produced by HMMER or SAM in different combinations. In local/local mode the benefit of SAM's total weight is strong. In global/local mode, the benefit is less pronounced and dependent on using SAM's transition prior. With our implementation of SAM's "bits saved" method for total weight calculation, HMMER performed about as well as using weights estimated by SAM. SAM's recode3.20.comp emission prior was used for all model building.
==== Refs
Park J Karplus K Barrett C Hughey R Haussler D Hubbard T Chothia C Sequence comparisons using multiple sequences detect three times as many remote homologues as pairwise methods J Mol Biol 1998 284 1201 1210 9837738 10.1006/jmbi.1998.2221
Hughey R Krogh A Hidden Markov models for sequence analysis: extension and analysis of the basic method Comput Appl Biosci 1996 12 95 107 8744772
Eddy SR Profile hidden Markov models Bioinformatics 1998 14 755 763 9918945 10.1093/bioinformatics/14.9.755
Bateman A Coin L Durbin R Finn RD Hollich V Griffiths-Jones S Khanna A Marshall M Moxon S Sonnhammer EL Studholme DJ Yeats C Eddy SR The Pfam protein families database Nucleic Acids Res 2004 32 (Database issue) D138 41 14681378 10.1093/nar/gkh121
Gough J Karplus K Hughey R Chothia C Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure J Mol Biol 2001 313 903 919 11697912 10.1006/jmbi.2001.5080
Krogh A Mitchison G Maximum entropy weighting of aligned sequences of proteins or DNA Proc Int Conf Intell Syst Mol Biol 1995 3 215 221 7584440
Karchin R Hughey R Weighting hidden Markov models for maximum discrimination Bioinformatics 1998 14 772 782 9918947 10.1093/bioinformatics/14.9.772
Barrett C Hughey R Karplus K Scoring hidden Markov models Comput Appl Biosci 1997 13 191 199 9146967
Brown M Hughey R Krogh A Mian IS Sjolander K Haussler D Using Dirichlet mixture priors to derive hidden Markov models for protein families Proc Int Conf Intell Syst Mol Biol 1993 1 47 55 7584370
Sjolander K Karplus K Brown M Hughey R Krogh A Mian IS Haussler D Dirichlet mixtures: a method for improved detection of weak but significant protein sequence homology Comput Appl Biosci 1996 12 327 345 8902360
Wistrand M Sonnhammer EL Transition priors for protein hidden Markov models: an empirical study towards maximum discrimination J Comput Biol 2004 11 181 193 15072695 10.1089/106652704773416957
Eddy SR Mitchison G Durbin R Maximum discrimination hidden Markov models of sequence consensus J Comput Biol 1995 2 9 23 7497123
Wistrand M Sonnhammer EL Improving profile HMM discrimination by adapting transition probabilities J Mol Biol 2004 338 847 854 15099750 10.1016/j.jmb.2004.03.023
Mitchison GJ Durbin R Tree-based maximal likelihood substitution matrices and hidden Markov models Journal of Molecular Evolution 1995 41 1139 1151 10.1007/BF00173195
Qian B Goldstein RA Detecting distant homologs using phylogenetic tree-based HMMs Proteins 2003 52 446 453 12866055 10.1002/prot.10373
Hargbo J Elofsson A Hidden Markov models that use predicted secondary structures for fold recognition Proteins 1999 36 68 76 10373007 10.1002/(SICI)1097-0134(19990701)36:1<68::AID-PROT6>3.0.CO;2-1
Karchin R Cline M Mandel-Gutfreund Y Karplus K Hidden Markov models that use predicted local structure for fold recognition: alphabets of backbone geometry Proteins 2003 51 504 514 12784210 10.1002/prot.10369
Coin L Bateman A Durbin R Enhanced protein domain discovery using taxonomy BMC Bioinformatics 2004 5 56 15137915 10.1186/1471-2105-5-56
Coin L Bateman A Durbin R Enhanced protein domain discovery by using language modeling techniques from speech recognition Proc Natl Acad Sci U S A 2003 100 4516 4520 12668763 10.1073/pnas.0737502100
Madera M Gough J A comparison of profile hidden Markov model procedures for remote homology detection Nucleic Acids Res 2002 30 4321 4328 12364612 10.1093/nar/gkf544
Karplus K Barrett C Hughey R Hidden Markov models for detecting remote protein homologies Bioinformatics 1998 14 846 856 9927713 10.1093/bioinformatics/14.10.846
Karplus K Karchin R Barrett C Tu S Cline M Diekhans M Grate L Casper J Hughey R What is the value added by human intervention in protein structure prediction? Proteins 2001 Suppl 5 86 91 11835485 10.1002/prot.10021
Haft DH Selengut JD White O The TIGRFAMs database of protein families Nucleic Acids Res 2003 31 371 373 12520025 10.1093/nar/gkg128
Letunic I Copley RR Schmidt S Ciccarelli FD Doerks T Schultz J Ponting CP Bork P SMART 4.0: towards genomic data integration Nucleic Acids Res 2004 32 D142 4 14681379 14681379 10.1093/nar/gkh088
Gerstein M Sonnhammer EL Chothia C Volume changes in protein evolution J Mol Biol 1994 236 1067 1078 8120887 10.1016/0022-2836(94)90012-4
Murzin AG Brenner SE Hubbard T Chothia C SCOP: a structural classification of proteins database for the investigation of sequences and structures J Mol Biol 1995 247 536 540 7723011 10.1006/jmbi.1995.0159
Lindahl E Elofsson A Identification of related proteins on family, superfamily and fold level J Mol Biol 2000 295 613 625 10623551 10.1006/jmbi.1999.3377
Brenner SE Chothia C Hubbard TJ Assessing sequence comparison methods with reliable structurally identified distant evolutionary relationships Proc Natl Acad Sci U S A 1998 95 6073 6078 9600919 10.1073/pnas.95.11.6073
Chandonia JM Hon G Walker NS Lo Conte L Koehl P Levitt M Brenner SE The ASTRAL Compendium in 2004 Nucleic Acids Res 2004 32 D189 92 14681391 14681391 10.1093/nar/gkh034
Soding J Protein homology detection by HMM-HMM comparison Bioinformatics 2005 21 951 960 15531603 10.1093/bioinformatics/bti125
Sadreyev R Grishin N COMPASS: a tool for comparison of multiple protein alignments with assessment of statistical significance J Mol Biol 2003 326 317 336 12547212 10.1016/S0022-2836(02)01371-2
Yona G Levitt M Within the twilight zone: a sensitive profile-profile comparison tool based on information theory J Mol Biol 2002 315 1257 1275 11827492 10.1006/jmbi.2001.5293
Edgar RC Sjolander K SATCHMO: sequence alignment and tree construction using hidden Markov models Bioinformatics 2003 19 1404 1411 12874053 10.1093/bioinformatics/btg158
Madera MGJ A conversion program between SAM and HMMER
| 15831105 | PMC1097716 | CC BY | 2021-01-04 16:02:48 | no | BMC Bioinformatics. 2005 Apr 15; 6:99 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-99 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-361581998110.1186/1471-2407-5-36Research ArticleMutation analysis of genes that control the G1/S cell cycle in melanoma: TP53, CDKN1A, CDKN2A, and CDKN2B Soto José Luis [email protected] Carmen M [email protected] Salvio [email protected]ópez-Nevot Miguel Ángel [email protected] Servicio de Análisis Clínicos e Inmunología, Hospital Universitario Virgen de las Nieves, Avenida Fuerzas Armadas N°2, 18014 Granada, Spain2 Servicio de Dermatología, Hospital Universitario San Cecilio, 18014 Granada, Spain2005 8 4 2005 5 36 36 24 12 2004 8 4 2005 Copyright © 2005 Soto 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 role of genes involved in the control of progression from the G1 to the S phase of the cell cycle in melanoma tumors in not fully known. The aim of our study was to analyse mutations in TP53, CDKN1A, CDKN2A, and CDKN2B genes in melanoma tumors and melanoma cell lines
Methods
We analysed 39 primary and metastatic melanomas and 9 melanoma cell lines by single-stranded conformational polymorphism (SSCP).
Results
The single-stranded technique showed heterozygous defects in the TP53 gene in 8 of 39 (20.5%) melanoma tumors: three new single point mutations in intronic sequences (introns 1 and 2) and exon 10, and three new single nucleotide polymorphisms located in introns 1 and 2 (C to T transition at position 11701 in intron 1; C insertion at position 11818 in intron 2; and C insertion at position 11875 in intron 2). One melanoma tumor exhibited two heterozygous alterations in the CDKN2A exon 1 one of which was novel (stop codon, and missense mutation). No defects were found in the remaining genes.
Conclusion
These results suggest that these genes are involved in melanoma tumorigenesis, although they may be not the major targets. Other suppressor genes that may be informative of the mechanism of tumorigenesis in skin melanomas should be studied.
==== Body
Background
The transition from phase G1 to S of the cell cycle is controlled by sequential activation of cyclin/Cdk complexes (Cyclin-dependent kinases) [1]. Active cyclin/Cdk complexes phosphorylate and inactivate members of the retinoblastoma protein (Rb) family, which are negative regulators of G1 and S-phase progression, leading to the induction of E2F-regulated gene expression and cell proliferation. Inhibitors of cyclin/Cdk complexes, by binding to these complexes, negatively regulate cell cycle progression [2].
Two families of Cdk-inhibitors (CKI) control the actions mediated by cyclin/Cdk complexes. p21 (also called WAF1, and CDKN1A; MIM# 116899) [3] is the founding member of the Cip/Kip family of CKI, which also includes p27 [4] and p57 [5]. Another class of Cdk inhibitors, the so-called INK4 proteins (named for their ability to inhibit cdk4), specifically target the cyclin D-dependent kinases [6]. To date, four INK4 proteins have been identified: the founding member p16INK4a (CDKN2A; MIM# 600160) [7], and three other closely related genes designated p15INK4b (CDKN2B; MIM# 600431) [8], p18INK4c (MIM# 603369) [9] and p19INK4d (MIM# 600927) [9].
In response to irradiation and chemotherapy, p53 protein (MIM# 191170) is stabilised and mediates apoptosis and cell cycle arrest. Whereas the mechanisms of p53-dependent apoptosis are not well understood, p53-dependent cell cycle arrest is known to be primarily mediated by p21, a potent inhibitor of cyclin-dependent kinases that is transactivated by p53 and p73 [10]. In addition to p21, several other cell cycle regulators are induced by p53, such as GADD45 and members of the 14-3-3 family [11].
The TP53 suppressor gene and Cdk-inhibitors such as CDKN1A, CDKN2A, and CDKN2B are targets of tumoral process in different types of tumors [12,13]. Mutations in the TP53 gene occur frequently in skin tumors as basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) [14]. In human melanoma, TP53 mutations are apparently not commonly detected [15,16], and consist mainly of C to T transitions located on dipyrimidine sites originated by UV radiation [17]. In contrast, CDKN2A is deleted or mutated in human sporadic melanomas and derived cell lines [18], and it appears to be the predisposing mutation in some familial melanoma kindreds [19]. A low incidence of mutations has been described for the CDKN2B gene in sporadic melanoma tumors [20]; however, no structural defects have been detected in the CDKN1A gene in human melanoma.
In order to investigate the role of the genes involved in the control of G1/S phase cell cycle progression in human melanomas, the aim of our study was to determine the presence of mutations in TP53, CDKN1A, CDKN2A and CDKN2B genes in primary and metastatic melanomas and melanoma cell lines.
Methods
Tumor samples
Thirty-nine specimens of skin melanoma were obtained from the Department of Surgery at the Hospital Universitario San Cecilio of Granada, Spain (Table 1). Melanoma tumors were dissected from normal tissues in fresh samples under sterile conditions, and tumor tissues were frozen in liquid nitrogen and stored at -80°C until DNA isolation. DNA was obtained from peripheral blood from each patient. The following 9 melanoma cell lines were included in this study: MZ2-MEL, MEL-3.0, MEL-2.2, and Mi-13443 were provided by Dr. T. Boon (Ludwig Institute of Cancer Research, Brussels, Belgium); and M31-L, M42-L, M52-L, M34-L, and M59-L were established in our laboratory as described previously [21]. The clinical and pathological characteristics of primary melanoma tumors and derived metastases are described in Table 1. Of the 39 tumors studied, 14 were primary (36%) while the rest were metastatic (18 ganlionar metastases and 4 subcutaneous metastases).
DNA isolation
DNA was isolated from tumor samples and peripheral blood lymphocytes with the MicroTurboGen Genomic DNA Isolation Kit (Invitrogen, San Diego, California) and the Quiagen kit (Wetsburg, Leusden, The Netherlands) respectively.
Mutation analysis of TP53, CDKN1A, CDKN2A, and CDKN2B genes
Point mutations were detected by changes in single-stranded conformational polymorphism (SSCP) of DNA amplified by polymerase chain reaction (PCR), as described by Orita et al [22] with slight modifications [23]. TP53 exons 2–11, CDKN1A exon 2, CDKN2B exon 2, and CDKN2A exons 1–2 were amplified. The sequences of the primers used and fragments (bp) amplified are described in Table 2. All TP53 primers used were provided by Clontech (Human p53 Amplier Panels, Palo Alto, CA). A portion of TP53 intron 1, exon 2, intron 2 and exon 3 was amplified using the primers PU2 (sense) and PD3 (antisense). CDKN1A exon 2 was amplified in two overlapping fragments with the following primer pairs: p21-L1/p21-R1 and p21-L2/p21-R2 (Table 2).
Amplifications were performed using 100 ng genomic DNA and α32P-dCTP (300 Ci/mmol) in a final volume of 25 μl. The PCR conditions for TP53 exons 2–11 were as follows: 35 cycles at 95°C/30 s, 66°C/45 s and 72°C/1.5 min, with a 10 min extension after the last cycle. CDKN1A exon 2, CDKN2B exon 2, and CDKN2A exons 1–2 were amplified under the same PCR conditions: in a touchdown PCR procedure the temperature of the reaction was lowered by 1°C every second cycle from 68°C to 60°C, at which temperature 30 cycles were carried out.
Amplified samples (2.5 μl) were mixed with 9 μl of sequencing stop solution (USB, Cleveland, OH, USA), 1.5 μl of 0.08 N NaOH, and 15 μl of 0.1% SDS, denatured for 10 min at 95°C, and the samples were quickly cooled in dry ice. Samples of 3 μl were loaded onto a 6% non-denaturing acrylamide gel containing 10% glycerol, and run at room temperature for 4 h at 22 W. Gels were dried at 80°C under vacuum and exposed to x-ray films for 4–16 h.
DNA sequencing
Asymmetric PCR reactions were purified from agarose gels and reamplified. PCR products were cloned in the PCR 4-TOPO vector using the TOPO TA Cloning Kit for sequencing (Invitrogen, Groningen, The Netherlands). After transformation several clones were picked and sequenced. Sequence analysis was carried out with the Sequenase DNA Sequencing kit (USB), using α35S-dATP (DuPont-NEN, Boston, MA) incorporation. Aliquots of the reaction mixture were run on a 6% denaturing acrylamide gel.
Results
Intronic single nucleotide polymorphisms and heterozygous point mutations in the TP53 gene
Of a total of 39 melanoma tumors and 9 melanoma cell lines studied by PCR-SSCP, we detected defects in the TP53 gene in 8 of 39 (20.5%) melanoma tumors, and did not find any alteration in melanoma cell lines (Table 3). Mutation analysis showed three novel heterozygous single point mutations in the TP53 sequence and four different single nucleotide polymorphisms, three of which have not been described to date. All novel single point mutations and polymorphisms were compared with the IARC (International Agency for Research on Cancer) TP53 Mutation Database .
The G to C transversion found at position 11827 in TP53 intron 2 in M4, M7, M38 and M53 melanoma tumors was previously described by Oliva et al [24] (Figure 1). In this study, we found 3 new single nucleotide polymorphisms located at intron 1 and 2 of the TP53 gene when we compared genomic DNA from melanoma tumors and DNA from autologous PBLs with control PBLs (Figure 1). The C to T transition was found at position 11701 of TP53 intron 1 in melanoma tumors M53 and M71 (Figure 2); a C insertion was found at position 11818 of TP53 intron 2 in melanoma tumors M7, M42, M53 and M71; and a C insertion was found at position 11875 of TP53 intron 2 in melanoma tumors M7, M38, M42, M53, and M71 (Table 3).
A heterozygous C deletion in TP53 exon 10 at position 172628 produced a stop codon and truncated the p53 protein in melanoma tumor M34. Melanoma tumors M42 and M43 showed heterozygous single point mutations at TP53 introns 1 and 2. The C to T transition at position 11701 in intron 1, observed in melanoma tumor M42, contrasted with the absence of this transition in autologous PBLs. In contrast, a C insertion was found at position 11818 in intron 2 of TP53 in autologous PBLs, but this polymorphism was not seen in melanoma tumor M43 (Tale 3).
Mutation analysis of CDKN1A, CDKN2A, and CDKN2B Cdk (Cyclin-dependent kinases) inhibitors genes
Two heterozygous alterations in CDKN2A exon 1 were observed in melanoma tumor M13 one of which novel, whereas no defects were seen in the CDKN1A and CDKN2B genes. The G to A transition produced a stop codon at position 149 [25]; and the novel T to C transition at position 298 resulted in substitution of proline for leucine (Table 3).
Discussion
Polymorphisms versus mutations in the TP53 gene in human melanoma
The major carcinogenic agent in most skin cancers is well established as solar ultraviolet light [26]. This is absorbed in DNA, with the formation of UV-specific dipyrimidine photoproducts. About 50% of all skin cancers exhibit TP53 mutations [17], and these mutations are characterised by a specific signature attributed to the UVB part of the solar spectrum. The impact of UVB radiation can be clearly inferred from the characteristic point mutations in TP53 found in human SCC and BCC, consisting of C to T or CC to TT transitions at dipirymidine sites [27]. These findings contrast with the situation in human melanomas, in which TP53 mutations are not commonly detected. The results of the present study support earlier findings that such mutations are indeed infrequent in this type of tumour. The influence of UVB radiation in these mutations is not clear. TP53 mutations in primary melanoma tumors induced by UVB radiation have been described previously by Zerp et al [16]. However, in the tumors examined in our study we did not find TP53 alterations originated by UVB radiation (Table 3). In contrast, we detected two new mutations located in intronic sequences (C deletion at position 11818 in intron 2, and C to T transition at position 11701 in intron 1) (Table 3) and the novel C deletion at codon 350 of TP53 exon 10 which produced a stop codon and truncated protein. More than 90% of the mutations reported in non-melanoma skin cancers and different types of tumors are clustered between exons 4 and 8 [28]. This region is highly conserved throughout evolution and contains the DNA-binding domain of p53, which is essential for its activity [29]. This contrasts with the trans-activation domain (encodes by exons 2 and 3) and the regulatory region (encodes by exons 9 to 11), where few mutations have been described. Therefore, the TP53 single nucleotide polymorphisms detected in these melanoma tumors appear to be their most frequent characteristic.
At least twelve intronic polymorphisms have been described in the human TP53 gene. These include between others a VNTR (variable number tandem repeat) region [30] and HaeIII restriction fragment length polymorphism (RFLP) [31] in intron 1, a G to C transversion in intron 2 [24], a 16 bp duplication in intron 3 (5'-gacctggagggctggg-3') [32], a MspI RFLP in intron 6 (G to A transition at 61 bp downstream of exon 6) [33,34], a G to C transversion at 37 bp upstream to exon 7 [35], an ApaI RFLP in intron 7 [36], and A to T transversion in intron 10 [37]. The melanoma tumors and melanoma cell lines studied here showed the G to C transversion at position 11827 of TP53 intron 2, previously described by Oliva et al [24] in four melanoma tumors (M4, M7, M38, and M53), and three new single nucleotide polymorphisms: C to T transition at position 11701 of TP53 intron 1; C insertion at position 11818 of TP53 intron 2; and C insertion at position 11875 of TP53 intron 2 (Figure 1). Moreover, we found three new heterozygous single point mutations in the TP53 gene (Table 3), the incidence of mutations detected in the TP53 gene accounted for only 7.7% (3 of 39 melanoma tumors) in contrast to 18% (7 of 39 melanoma tumors) of single nucleotide polymorphisms found in these tumors.
Associations between cancer phenotypes and inherited TP53 intronic polymorphisms have been observed in studies of epithelial cancers including ovarian, breast, colon, thyroid, nasopharyngeal, lung cancer, and thyroid [34,35,38-40]. The frequency of G to C transition at position 11827 in intron 2 of TP53 gene (3 of 39 melanoma tumors, 7.7%) is low compared to the frequency of the A1 allele (G to C transition at position 11827) described previously in Caucasian individuals [41]. The polymorphisms that we detected in these melanoma tumors may play a role in the risk of developing skin melanoma in these patients.
Alterations in cyclin-Cdk inhibitors: CDKN1A, CDKN2A, and CDKN2B
Our results revealed the low incidence of mutations in cyclin-Cdk inhibitors in both melanoma tumors and melanoma cell lines. We detected only two mutations in exon 1 of the CDKN2A gene, both in melanoma tumor M13: G to A transition at position 149 producing a stop codon [25], and the novel missense mutation Leu298 (leucine → proline). The low incidence of CDKN2A mutations found in primary melanomas is concordant with previous reports [42,43]. In contrast, melanomas cell lines show a high incidence of mutations in CDKN2A, with homozygous deletion being the main mechanism of alteration [42]. These results suggest that in sporadic melanoma tumors, CDKN2A might not be a target of mutation, whereas in familial melanoma this mutation accounts for approximately 10% of all cases of tumors.
Conclusion
We conclude that although none of the cell cycle regulators analysed here can be singled out as a main mutation target for melanoma tumorigenesis, G1/S checkpoint defects are one of the significant factors in the development of melanoma tumors. However, this influence appear to be low in tumors, unlike the situation in melanoma cell lines. Other suppressor genes will be investigated to identify the main targets in the pathogenesis of melanoma.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JLS carried out the molecular genetic studies of SSCP and DNA sequencing. CMC participated in the manuscript drafted. SS participated in the design of the study. MALN participated in the design, coordination and helped to draft 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 study was supported by FISS grant 021542 from the Spanish Ministry of Health. We thank K. Shashok for checking the use of English in the manuscript.
Figures and Tables
Figure 1 Single nucleotide polymorphisms detected in intron 1 and intron 2 sequences of the TP53 gene in melanoma tumors. Each polymorphism and its positions are indicated by an asterisk (*), to determinate the presence of the polymorphisms the DNA sequence from melanoma tumors and autologous PBLs were compared with control PBLs.
Figure 2 New single nucleotide polymorphism found in intron 1 of the TP53 gene (C>T transition) at position 11701. (A) PCR-SSCP analysis of melanoma tumors. The arrow indicates the shifted band in melanoma tumor M71. DC, denaturing control; NDC, non-denaturing control. (B) DNA sequence of melanoma tumor M71 and autologous PBLs showing the C to T base change (shown by an asterisk *) at position 11701 compared with control PBLs.
Table 1 Melanoma tumor samples
Tumor aHistopathology bBreslow (mm) Clark cTumor sample
M3 SSM 5.5 IV P
M4a NM - - Sm
M4b NM - - Sm
M5 - - - P
M6 - - - Nm
M7a SSM 4 IV Nm
M7b SSM 4 IV Sm
M8 SSM 0.5 II P
M13 SSM 3.9 III P
M18 NM 5 V Nm
M19 - - - Nm
M21 SSM 3.5 III P
M22 - - - Nm
M23 SSM 9 IV Nm
M24 ALM - - Nm
M31 SSM 3 IV Nm
M32 - - - Nm
M34 SSM 16 V P
M37 SSM 1.8 III Nm
M38 NM 9 V P
M40 NM 3.4 IV Nm
M42 NM 1.5 IV Nm
M43 SSM 2.5 IV P
M44a NM 10 IV P
M44b NM 10 IV Sm
M45 - - - Nm
M46 - - - P
M49 - - - P
M50 - - - Nm
M52 LMM - - Nm
M53 - - - Nm
M55 ALM - V Nm
M56 LMM 1 III P
M59 NM 10.1 III P
M60 SSM 3 III P
M71 NM - - P
M72 SSM 0.6 III P
M73 SSM 2.5 III P
M74 - - - Nm
aSSM (superficial spreading melanoma), NM (nodular melanoma), LMM (lentigo maligna melanoma), ALM (acral lentigo melanoma). bBreslow vertical tumor thickness. cP, primary melanoma; Nm, lymph node metastases; Sm, subcutaneous metastases.
Table 2 Oligonucleotide primer sequences
TP53, exons 2 to 11 (GeneBank Accession: X54156)
Exon Primer Sequence 5'→3' Fragment (bp)
PU2 (sense) TCCTCTTGCAGCAGCCAGACTGC
PD3 (antisense) AACCCTTGTCCTTACCAGAACGTTG
4 PU4 (sense) CACCCATCTACAGTCCCCCTTGC 307
PD4 (antisense) CTCAGGGCAACTGACCGTGCAAG
5 PU5 (sense) CTCTTCCTACAGTACTCCCCTGC 211
PD5 (antisense) GCCCCAGCTGCTCACCATCGCTA
6 PU6 (sense) GATTGCTCTTAGGTCTGGCCCCTC 182
PD6 (antisense) GGCCACTGACAACCACCCTTAACC
7 PU7 (sense) GTGTTATCTCCTAGGTTGGCTCTG 139
PD7(antisense) CAAGTGGCTCCTGACCTGGAGTC
8 PU8 (sense) ACCTGATTTCCTTACTGCCTCTTGC 200
PD8 (antisense) GTCCTGCTTGCTTACCTCGCTTAC
9 PU9 (sense) GCCTCTTTCCTAGCACTGCCCAAC 102
PD9 (antisense) CCCAAGACTTAGTACCTGAAGGGTG
10 PU10 (sense) TGTTGCTGCAGATCCGTGGGCGT 130
PD10 (antisense) GAGGTCACTCACCTGGAGTGAGC
11 PU11 (sense) TGTGATGTCATCTCTCCTCCCTGC 153
PD11 (antisense) GGCTGTCAGTGGGGAACAAGAAGT
CDKN1A, exon 2 (GeneBank Accession: AF497972)
p21-L1 (sense) GATGTCCGTCAGAACCCATG 258
p21-R1 (antisense) TGCCTCCTCCCAACTCAT
p21-L2 (sense) ATGAGTTGGGAGGAGGCA 181
p21-R2 (antisense) ATGCTGGTCTGCCGCCGTT
CDKN2A, exons 1(GeneBank Accession: U12818) and 2 (GeneBank Accession: U12819)
1 MTS1-L1 (sense) GAAGAAAGAGGAGGGGCTG 340
MTS1R1 (antisense) GCGCTACCTGATTCCAATTC
2 p16-L2 (sense) GTCATGATGATGGGCAGC 307
p16-R2 (antisense) CTGAGGGACCTTCCGCG
CDKN2B, exon 2 (GeneBank Accession: AF513858)
MTS2-L2 (sense) TAAGTTTAACCTGAAGGTGG 500
MTS2-R (antisense) GGGTGGGAAATTGGGTAAG
Table 3 Single nucleotide polymorphisms (SNPs) and mutations (M) in TP53, CDKN1A, CDKN2A, and CDKN2B genes
TP53
Tumor
M4 11827G>C (intron 2) (SNP)
11875insC (intron 2) (SNP)
M7 11827G>C (intron 2) (SNP)
11818insC (intron 2) (SNP)
11875insC (intron 2) (SNP)
M38 11827G>C (intron 2) (SNP)
11875insC (intron 2) (SNP)
M42 11701C>T (intron 1) (M)
11818insC (intron 2) (SNP)
11875insC (intron 2) (SNP)
M43 11875insC (intron 2) (SNP)
11818delC (intron 2) (M)
M53 11701C>T (intron 1) (SNP)
11827G>C (intron 2) (SNP)
11818insC (intron 2) (SNP)
11875insC (intron 2) (SNP)
M71 11701C>T (intron 1) (SNP)
11818insC (intron 2) (SNP)
11875insC (intron 2) (SNP)
M34 17628delC (codon 350: CTC→CTA, exon 10), stop codon at the beginning of exon 11 (M)
CDKN2A, exon 1
M13 g.149G>A, TGG (Trp)→TGA (stop codon) (M)
g.298T>C, CTC (Leu)→CCC (Pro) (M)
CDKN1A, and CDKN2B No alterations found
==== Refs
Ekholm SV Reed SI Regulation of G1 cyclin-dependent kinases in the mammalian cell cycle Curr Opin Cell Biol 2000 12 676 684 11063931 10.1016/S0955-0674(00)00151-4
Sherr CJ Roberts JM CDK inhibitors: positive and negative regulators of G1-phase progression Genes Dev 1999 13 1501 1512 10385618
El-Deiry W Tokino T Velculescu V Levy D Parsons R Trent J Lin D Mercer WE Kinzler K Vogelstein B WAF1, a potential mediator of p53 tumor suppression Cell 1993 75 817 825 8242752 10.1016/0092-8674(93)90500-P
Toyoshima H Hunter T p27, a novel inhibitor of G1 cyclin/cdk protein kinase activity, is related to p21 Cell 1994 78 67 74 8033213 10.1016/0092-8674(94)90573-8
Lee MH Reynisdottir I Massagué J Cloning of p57KIP2, a cylin-dependent kinase inhibitor with unique domain structure and tissue distribution Genes Dev 1995 9 639 649 7729683
Ruas M Peters G The p16INKa/CDKN2A tumor suppressor and its relatives Biochim Biophys Acta 1998 1378 F115 F177 9823374 10.1016/S0304-419X(98)00017-1
Serrano M Hannon GJ Beach D A new regulatory motif in cell cycle control causing specific inhibition of cyclin D/CDK4 Nature 1993 366 704 707 8259215 10.1038/366704a0
Hannon GJ Beach D p15INK4b is a potential effector of TGFβ-induced cell cycle arrest Nature 1994 371 257 261 8078588 10.1038/371257a0
Hirai H Roussel MF Kato J Ashmun RA Sherr CJ Novel INK4 proteins, p19 and p18, are specific inhibitors of the cyclin D-dependent kinases CDK4 and CDK6 Mol Cell Biol 1995 15 2672 2681 7739547
Lohrum M Vousden K Regulation and function of the p53-related proteins: same family, different rules Trends Cell Biol 2000 10 197 202 10754563 10.1016/S0962-8924(00)01736-0
Rich T Allen RL Wyllie AH Defying death after DNA damage Nature 2000 407 777 783 11048728 10.1038/35037717
Yonghas T Quian H Chuanyuan L Yandell DW Deletions and point mutations of p16, p15 genes in primary tumors and tumors cell lines Chin Med Sci J 1999 14 200 205 12894891
Kawamura M Ohnishi H Guo SX Sheng XM Minegishi M Hanada R Horibe K Hongo T Kaneko Y Bessho F Yanagisawa M Sekiga T Hayashi Y Alterations of the p53, p21, p16, p15 and RAS genes in childhood T-cell acute lymphoblastic leukemia Leuk Res 1999 23 115 126 10071127 10.1016/S0145-2126(98)00146-5
Moles JP Moyret C Guillot B Jeanteur P Guilhou JJ Theillet C Basset-Seguin N p53 gene mutations in human epithelial skin cancers Oncogene 1993 8 583 588 8437842
Hartmann A Blaszyk H Cunningham JS McGovern RM Schroeder JS Helander S Pittelkow MR Sommer SS Kovach JS Overexpression and mutations of p53 in metastatic malignant melanomas Int J Cancer 1996 67 313 317 8707401 10.1002/(SICI)1097-0215(19960729)67:3<313::AID-IJC1>3.0.CO;2-U
Zerp SF Elsas van A Peltenburg LTC Schrier PI P53 mutations in human cutaneous melanoma correlate with sun exposure but are not always involved in melanogenesis Br J Cancer 1999 79 921 926 10070891 10.1038/sj.bjc.6690147
Giglia-Mari G Sarasin A TP53 mutations in human skin cancers Hum Mutat 2003 21 217 228 12619107 10.1002/humu.10179
Bartkova J Lukas J Guldberg P Alsner J Kirkin AF Zeuthen J Bartek J The p16-cyclin D/Cdk4-pRb pathway as a functional unit frequently altered in melanoma pathogenesis Cancer Res 1996 56 5475 5483 8968104
Liggett WH JrSidransky D Role of the p16 tumor suppressor gene in cancer J Clin Oncol 1998 16 1197 1206 9508208
Matsumura Y Nishigori C Yagi T Imamura S Takebe H Mutations of p16 and p15 tumor suppressor genes and replication errors contribute independently to the pathogenesis of sporadic malignant melanoma Arch Dermatol Res 1998 290 175 180 9617435 10.1007/s004030050286
Kirkin AF Petersen TR Olsen AC Li L thor Straten P Zeuthen J Generation of human-melanoma-specific T lymphocyte clones defining novel cytolytic targets with panels of newly established melanoma cell lines Cancer Immunol Immunother 1995 41 71 76 7656272 10.1007/s002620050202
Orita M Suzuki Y Sekiya T Hayashi K Rapid and sensitive detection of point mutations and DNA polymorphisms using polymerase chain reaction Genomics 1989 5 874 879 2687159 10.1016/0888-7543(89)90129-8
Prosser J Detecting single-base mutations TIBTECH 1993 11 238
Oliva MR Saez GT Latres E Cordon-Cardo C A new polymorphic site in intron 2 to TP53 characterizes LOH in human tumors by PCR-SSCP Diagn Mol Pathol 1995 4 54 58 7735557
Holland EA Schmid H Kefford RF Mann GJ CDKN2A (P16(INK4a)) and CDK4 mutation analysis in 131 Australian melanoma probands: effect of family history and multiple primary melanomas Genes Chromosomes Cancer 1999 25 339 348 10398427 10.1002/(SICI)1098-2264(199908)25:4<339::AID-GCC5>3.0.CO;2-H
Cleaver JE Crowley E UV damage, DNA repair and skin carcinogenesis Front Biosci 2002 7 d1024 1043 11897551 11897551
Ziegler A Leffell DJ Kunala S Sharma HW Gailani M Siman JA Halperin AJ Baden HP Shapiro PE Bale AE Brash DE Mutation hotspots due to sunlight in the p53 gene of non-melanoma skin cancers Proc Natl Acad Sci USA 1993 90 4216 4220 8483937
Soussi T The p53 tumour suppressor gene: a model for molecular epidemiology of human cancer Mol Med Today 1996 2 32 37 8796849 10.1016/1357-4310(96)88756-9
Mendoza-Rodriguez CA Cerbon MA Tumor suppressor gene p53: mechanisms of action in cell proliferation and death Rev Invest Clin 2001 53 266 273 11496714
Hahn M Serth J Fislage R Wolfes H Allhoff E Jonas V Pingoud A Polymerase chain reaction detection of a highly polymorphic VNTR segment in intron 1 of the human p53 gene Clin Chem 1993 39 549 550 8448878
Ito T Seyama T Hayashi T Mizuno T Iwamoto KS Tsuyama N Dohi K Nakamura N Akiyama M Hae III polymorphism in intron 1 of the human p53 gene Human Genet 1994 93 222 8112754 10.1007/BF00210619
Lazar V Hazard F Bertin F Janin N Bellet D Bressac B Simple sequence repeat polymorphism within the p53 gene Oncogene 1993 8 1703 1705 8502490
Peller S Kopilova Y Slutzki S Halevy A Kvitko K Rother V A novel polymorphism in intron 6 of the human p53 gene: a possible association with cancer predisposition and susceptibility DNA Cell Biol 1995 14 983 990 8534372
Mavridou D Gornall R Campbell IG Eccles DM TP53 intron 6 polymorphism and the risk of ovarian and breast cancer Br J Cancer 1998 77 676 677 9484829
Hillebrandt S Streffer C Demidchik EP Biko J Reiners C Polymorphisms in the p53 gene in thyroid tumours and blood samples of children from areas in Belarus Mutat Res 1997 381 201 207 9434876
Prosser J Condie A Biallelic ApaI polymorphism of the human p53 gene (TP53) Nucleic Acids Res 1991 19 4799 1679931
Buller RE Skilling JS Kaliszewski S Niemann T Anderson B Absence of significant germline p53 mutations in ovarian cancer patients Gynecol Oncol 1995 58 368 374 7672703 10.1006/gyno.1995.1244
Sjalander A Birgander R Athlin L Stenling R Rutegard J Beckman L Beckman G p53 germline haplotypes associated with increased risk for colorectal cancer Carcinogenesis 1995 16 1461 1464 7614678
Birgander R Sjalander A Rannug A Alexandrie AK Sundberg MI Seidergard J Tornling G Beckman G Beckman L p53 polymorphisms and haplotypes in lung cancer Carcinogenesis 1995 16 2233 2236 7554081
Birgander R Sjalander A Zhou Z Fan C Beckman L Beckman G p53 polymorphisms and haplotypes in nasopharyngeal cancer Hum Hered 1996 46 49 54 8825463
Ge H Lam WK Lee J Wong MP Fu KH Yew WW Lung ML Detection and evaluation of p53 intron 2 polymorphism in lung carcinomas in Hong Kong Int J Cancer 1996 69 120 124 8608979 10.1002/(SICI)1097-0215(19960422)69:2<120::AID-IJC9>3.0.CO;2-2
Ruiz A Puig S Lynch M Castel T Estivell X Retention of the CDKN2A locus and low frequency of point mutations in primary and metastatic cutaneous malignant melanoma Int J Cancer 1998 76 312 316 9579564 10.1002/(SICI)1097-0215(19980504)76:3<312::AID-IJC4>3.0.CO;2-Y
Sauroja I Smeds J Vlaykova T Kumar R Talve L Hahka-Kemppinen M Punnonen K Jansen CT Hemminki K Pyrhonen S Analysis of G(1)/S checkpoint regulators in metastatic melanoma Genes Chromosomes Cancer 2000 28 404 414 10862049 10.1002/1098-2264(200008)28:4<404::AID-GCC6>3.0.CO;2-P
| 15819981 | PMC1097717 | CC BY | 2021-01-04 16:03:04 | no | BMC Cancer. 2005 Apr 8; 5:36 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-36 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-401583313610.1186/1471-2407-5-40Research ArticleRare mutations predisposing to familial adenomatous polyposis in Greek FAP patients Mihalatos Markos [email protected] Angela [email protected] Hans [email protected] Voula [email protected] Aristidis [email protected] Alexander [email protected] Konstantinos [email protected] John K [email protected] Ioannis [email protected] George [email protected] Niki J [email protected] Georgios [email protected] Molecular Biology Research Center HYGEIA – «Antonis Papayiannis», Athens2 Center for Human and Clinical Genetics, Leiden University Medical Center, The Netherlands3 Cytogenetics Laboratory, Department of Genetics and Molecular Biology, Mitera Maternity and Surgical Center, Athens, Greece4 Hygeia Ofthalmos, Diagnostic and Therapeutic Center of Athens HYGEIA S.A., Athens, Greece5 Surgical Clinic, General State Hospital of Athens "G. Gennimatas", Athens, Greece6 Gastroenterology Department, Euromedica – Engephalos Diagnostic Center of Athens, Greece7 Gastroenterology Department, Peripheral General Hospital of Nikaia, Greece8 Gastroenterology Department Diagnostic and Therapeutic Center of Athens HYGEIA S.A., Athens Greece9 AHEPA Hospital, Aristotle University of Thessaloniki, Thessaloniki Greece10 Department of Pathology, Medical School, University of Ioannina, Ioannina, Greece2005 15 4 2005 5 40 40 20 1 2005 15 4 2005 Copyright © 2005 Mihalatos 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
Familial Adenomatous Polyposis (FAP) is caused by germline mutations in the APC (Adenomatous Polyposis Coli) gene. The vast majority of APC mutations are point mutations or small insertions / deletions which lead to truncated protein products. Splicing mutations or gross genomic rearrangements are less common inactivating events of the APC gene.
Methods
In the current study genomic DNA or RNA from ten unrelated FAP suspected patients was examined for germline mutations in the APC gene. Family history and phenotype were used in order to select the patients. Methods used for testing were dHPLC (denaturing High Performance Liquid Chromatography), sequencing, MLPA (Multiplex Ligation – dependent Probe Amplification), Karyotyping, FISH (Fluorescence In Situ Hybridization) and RT-PCR (Reverse Transcription – Polymerase Chain Reaction).
Results
A 250 Kbp deletion in the APC gene starting from intron 5 and extending beyond exon 15 was identified in one patient. A substitution of the +5 conserved nucleotide at the splice donor site of intron 9 in the APC gene was shown to produce frameshift and inefficient exon skipping in a second patient. Four frameshift mutations (1577insT, 1973delAG, 3180delAAAA, 3212delA) and a nonsense mutation (C1690T) were identified in the rest of the patients.
Conclusion
Screening for APC mutations in FAP patients should include testing for splicing defects and gross genomic alterations.
==== Body
Background
Germline mutations in the APC gene are the causative factor for the Familial Adenomatous Polyposis syndrome which follows an autosomal dominant inheritance pattern [1,2]. FAP patients develop multiple to thousands of colonic polyps before their second decade of life.
If polyps are left untreated they give rise to colorectal cancer in almost all cases. The syndrome has two forms, one referred to as classic FAP and one milder form referred to as attenuated FAP (AFAP) [3]. Most FAP patients develop certain extracolonic features [3,4]. Congenital Hypertrophy of the Retinal Pigment Epithelium (CHRPE) is present in the majority of FAP patients. Nearly 94% of APC germline mutations lead to a truncated protein product. Thirty three percent of APC mutations represent nonsense point mutations, 6% small insertions and 55% small deletions [5].
Recently, a number of studies have demonstrated the presence of less common APC mutations in FAP patients, such as large genomic rearrangements or splicing affecting mutations [6,7]. It is generally noted that large rearrangements are more likely to produce a classic FAP phenotype, while splicing mutations tend to result in attenuated FAP (AFAP) [8-13].
Large rearrangements account for up to one third of identified mutations in the BRCA1 and hMLH1 and hMSH2 genes which are responsible for other distinct or related types of cancer, respectively [14,15]. In APC the percentages are lower but according to recent reports may reach as much as 10% of identified mutations [16]. Recently, MLPA has been used in our laboratory, to successfully identify large deletions in the hMSH2 and BRCA1 genes [17,18].
Splice site mutations are a common inactivating event for several human genes estimated to account for about 15% of the mutations [19]. In the ATM and NF1 genes they represent about 50% of the identified mutations [20,21]. Most frequent targets of splice site mutations are the consensus GT and AG donor and acceptor dinucleotides respectively. However, several splice site mutations altering other exonic or intronic nucleotides neighboring the consensus sites, have been described [12,21,22]. Screening the corresponding genes' cDNA elucidates most splicing defects.
Nevertheless, these studies point out the need for more intensive mutation screening than coding region and splice junction sequencing. Finally, these studies address the issue that such uncommon mutations may have been underestimated due to the methods of mutation screening that have been generally used in the past decade.
In the present study, two different cases of FAP families carrying two uncommon APC mutations are presented: a splice site mutation on the fifth nucleotide of intron 9 and a large genomic deletion from one chromosome 5 encompassing around 250 Kbp, starting in intron 5 of the APC gene and extending beyond exon 15. We also further support our ongoing study on FAP presenting another four frameshift mutations of the APC gene identified in four unrelated patients.
Finally, we propose a simple algorithm for the APC genetic testing of FAP patients based on our experience after screening 96 individuals from 25 FAP families from our Greek cohort.
Methods
Patients
Patients were referred to our center and selected for genetic testing as previously described [4]. Ethical approval was obtained from the HYGEIA – Diagnostic and Therapeutic Center of Athens S.A. – hospital's advisory committee (ref.27931/23.10.2000) and all patients signed informed consent.
DNA and RNA isolation
Genomic DNA and RNA were purified from peripheral blood leukocytes using standard extraction protocols, as already described previously [4].
PCR amplification
All 16 exons (1–15, 10A) including splice junctions of APC were amplified by PCR (Polymerase Chain Reaction) as already described elsewhere [4,23]. The primers used were taken from other studies [1,4].
dHPLC analysis
The WAVE DNA Fragment Analysis System (Transgenomic, Inc., USA) and associated WAVE-Maker™ software were used as previously described [23].
Sequence analysis
Purification of the PCR products and automated cycle sequencing for both strands was performed on the ABI Prism® 310 Genetic Analyzer as previously described [4,23].
RT-PCR
First strand synthesis was performed by denaturing approximately 500 – 1000 ng total RNA and random hexamers (5 μM final concentration) for 4 min at 70°C, followed by snap freezing on ice and addition of dNTPs (0.5 mM final concentration), 1 U/μl recombinant RNase inhibitor (Invitrogen, The Netherlands) and 200 U MMLV reverse transcriptase (Invitrogen, The Netherlands). The mixture was incubated at 37°C for 1 hour followed by denaturation of the enzymes at 95°C for 5 min. 4 μl of cDNA were used for subsequent PCR amplification. The primers used for PCR were described by Varesco et al (1994) [22] and amplify the cDNA region between exons 8 and 12 of the APC gene.
Multiplex Ligation-dependent Probe Amplification (MLPA)
MLPA was carried out using the P043 APC kit (MRC-Holland, Netherlands) as instructed by the manufacturer. Fragment analysis was carried out on the ABI Prism® 310 Genetic Analyzer using TAMRA-500 provided by the manufacturer, as size standard. A peak pattern of 31 peaks ranging in size from 130 to 400 nt is obtained [24].
Conventional cytogenetic analysis
Metaphase chromosome preparations were obtained from peripheral blood lymphocytes using standard techniques. Conventional cytogenetic analysis was carried out using high resolution GTG banding (1200 bands). Cells were analyzed, karyotyped and captured by computer imaging (Cytovision system, Image Analysis, Applied Imaging, Sunderland, UK).
FISH analysis
FISH analysis was performed as already described by Dauwerse JG et al (1992) [25]. Cosmid probes cAPC1, cAPC2, cAPC3 for the APC gene and MCBCD, for the MCC gene on band 5q22.2 were simultaneously hybridized with the telomere PACs GS-240G13 (5qter) and GS-24H17 (5pter) for identification of chromosome 5 [26,27].
Results
Alternative splicing
In family A the proband was a young man at the age of 29 with a phenotype showing more than a hundred polyps throughout the colon, particularly in the rectum, sigmoid and descending colon. No polyps were found in the esophagus, duodenum and ileum, except for one polypoid adenomatous lesion in the stomach. The polyps were adenomatous, smaller than 5 millimeters and some tubular and serrated adenomas were also observed. All lesions showed low or medium grade dysplasia while some serrated adenomas were mildly hyperplastic. Funduscopic examination showed the absence of CHRPE.
dHPLC screening of the APC gene revealed an aberrant elution profile for the amplicon encompassing both exons 9 and 9A. Subsequent sequencing of that amplicon confirmed the presence of the IVS9+5G>A mutation (Fig. 1a). In order to examine the consequences of the G to A transition, RT-PCR was performed from total RNA extracted from peripheral blood lymphocytes of the proband. Total RNA from normal individuals and other FAP patients with identified mutations in other regions of the APC gene, were used as normal controls for this RT-PCR. RT-PCR products were separated by non denaturing 6% PAGE and visualized by EtBr staining (Fig. 1b). Two normal transcripts were amplified in all samples representing the two normal spliced forms of exon 9: 9 and 9A (Fig. 1b). A novel shorter transcript was only amplified from the proband's RNA (Fig. 1b). The corresponding band of 336 bp was purified and sequenced. Sequence analysis revealed an alternate transcript in which the whole of exon 9 was missing and exon 8 was directly joined to exon 10. This abnormal splicing results in a frameshift and creates a STOP signal at the 16th codon of exon 10 (Fig. 1c).
The proband's phenotype was resembling AFAP with a late age of onset. The proband's siblings were asymptomatic but younger than 25 years old while the proband's parents were of normal phenotype. In order to exclude AFAP from these individuals, they were screened by dHPLC for the IVS9+5G>A mutation. All individuals were of normal genotype, indicating that the mutation is de-novo.
Large genomic rearrangements
In family B the proband was a young man at the age of 25 with a phenotype of more than a hundred polyps throughout his colon. The patient's mother and maternal uncle died at the age of 46 and 30 from CRC and liver cancer – probably metastatic, respectively. The patient's maternal grandmother had died from CRC at the age of 45 (Fig. 2). Polyps were also present in the stomach, duodenum and ileum. The polyps were tubular adenomas, ranging in size from 5 to 15 millimeters and some with a diameter of more than 10 mm were pediculus. All lesions showed low or medium grade dysplasia. Funduscopic examination showed the presence of CHRPE for this patient.
dHPLC and sequencing of the whole APC coding region, including splice junctions showed only the known G4479A polymorphism for which the patient was homozygous. Considering these results and the severity of the patient's phenotype, MLPA was performed to screen for genomic rearrangements which might had been undetected with the conventional techniques. MLPA analysis showed the deletion of exons 6 to 15 from one allele, in the patient's lymphocytes, from two different blood samples (Fig. 3). The proband's asymptomatic brother tested normal.
A large deletion, starting from intron 5 and ending downstream of the last APC exon, was revealed. In order to narrow down the 3' breakpoint, karyotype analysis was performed on the patient's peripheral blood lymphocytes. Karyotyping was negative excluding a deletion longer than 20 Mbp. Subsequent FISH analysis confirmed that a large portion of the APC gene was missing from one chromosome 5 (data not shown). Further FISH analysis showed that the 3' breakpoint of the deletion is lying in the MCC gene near its 3' end (data not shown). The 5' breakpoint lies somewhere in intron 5 which has a size of about 12 Kbp. These data indicate that the deletion encompasses about 250 Kbp.
APC frameshift and nonsense mutations
Four frameshift mutations (1577insT, 1973delAG, 3180delAAAA and 3212delA) of the APC gene were identified in four unrelated FAP patients and the C1690T nonsense mutation was identified in two unrelated patients. To our knowledge 1577insT is a novel mutation.
In two patients no mutation was identified in the APC either screening for small insertions, deletions or point mutations or for large genomic rearrangements.
Discussion
Alternative splicing and exon skipping
In the current study the IVS9+5G>A mutation was identified and investigated. Our patient's phenotype is resembling more to AFAP than to FAP, with small polyps, particularly in the right colon. A novel transcript is produced in the carrier's peripheral blood lymphocytes not present in normal controls. Normal transcripts are also present in the patient's sample, but result in lower amplification when compared to the corresponding ones of the normal samples (Fig. 1a). The IVS9+5G>A mutation does not alter the splicing donor site but changes a very conserved nucleotide. Varesco et al, (1994) [22] have documented the transversion of G to T (IVS9+5G>T) at the same point to be pathogenic. The IVS9+5G>T mutation causes inefficient exon skipping, thus producing a milder phenotype leading to attenuated FAP. In both cases, the predicted truncated protein product of about 35 kDa is not large enough to stay stable and produce a dominant negative effect. The absence of CHRPE from this patient is in agreement with the findings from Olschwang et al (1993) [28] who concluded that patients with mutations before exon 9 do not develop CHRPE.
Most mutations observed around the donor splice site, that cause aberrant splicing, occur at positions -1, +1 or +2 of the intron [19]. There are few reports of mutations altering position +5 [12,21,22]. However, the substitution of the G residue at position +5, as well as at position +1, would be expected to produce a significant reduction of the base pairing stability between the 5' splice site and its complementary region of the U1 snRNA [29,30]. U1 snRNA binding is a crucial step for the correct folding of pre-mRNA prior to cleavage and ligation at the spliceosome.
Moisio et al (2002) [12], also report the IVS4+5G>C mutation to cause skipping of exon 4 of the APC at the mRNA level and resulting in AFAP. Based on the data presented here and in earlier studies [12,22] we conclude that the IVS9+5G>A mutation is responsible for this patient's AFAP phenotype and that the +5G of the donor site is a highly conserved and functionally important nucleotide.
Large genomic rearrangements of the APC
In this FAP family the patient's phenotype is classic FAP, in agreement with other studies [8-10] supporting that whole gene APC deletions produce the classic FAP rather than attenuated FAP. Considering MLPA and FISH results, the size of the identified deletion has a length of about 250 Kbp.
The presence of CHRPE in this patient is in contradiction to the general observation that patients with mutations before exon 9 do not develop CHRPE. Davies et al (1995) [31] have also reported a patient with CHRPE and a mutation affecting splicing at the donor splice site of intron 4. Wallis et al (1994) [32] state that the presence of CHRPE depends on the size of the truncated APC protein. This leaves a possibility of a fusion transcript as the result of our patient's deletion to be clarified by further experiments such as possibly 3' Rapid Amplification of cDNA Ends (RACE).
Having analysed 25 FAP families in the past 3 years, one family (4%) carrying a large deletion at the APC locus was identified. This supports the notion that large genomic events are not rare in APC but more data is needed in order to calculate the percentage among Greek patients. Recently, the notion that rearrangements are not so rare for the APC but may account for more than 10% of its mutation spectrum, is supported by studies in "APC negative" FAP patients. Su et al (2000) [6], report a percentage of 13% and Cao et al (2001) [16] report 9% of FAP patients carrying large rearrangements. It is possible that the true percentage of these mutations is underestimated due to the limitations of conventional mutation scanning used until now.
Frameshift and nonsense APC mutations
Three small deletions, one small insertion and a nonsense mutation were identified in this study, supporting the fact that most APC mutations cause frameshift and usually delete than insert nucleotides. The identified mutations are predicted to produce truncated protein products introducing STOP codons: 1577insT at codon 535, C1690T at codon 564, 1973delAG at codon 672 while 3180delAAAA and 3212delA at codon 1125.
Genetic testing for adenomatous polyposis
In the past 3 years we have screened 25 families with adenomatous polyposis [4,23]. To minimize time, cost and effort, we use a simple algorithm to perform genetic test for FAP patients. First, we carefully examine the patient's phenotype in order to classify it as classic FAP or AFAP.
For classic FAP we start dHPLC screening of the APC from the MCR (Mutation Cluster Region), followed by screening of the remaining 5' region of exon 15. Then we screen the remaining exons with the following order: exons 11 to 14, exons 1 to 10A and finally the remaining exon 15. MLPA is performed in order to identify possible gross genomic alterations affecting the APC gene. For AFAP patients, screening follows the order: exons 1- 10A, 3' region of exon 15, exons 11 – 14, 5' region and MCR of exon 15. If no mutation is detected we finally perform the MLPA test.
Even following the above algorithm there is a small number of FAP patients in whom no pathogenic mutation can be detected in the APC gene. These cases include patients carrying unclassified variants of these genes such as I1307K, E1317Q and D1822V in the APC gene. Large epidemiologic studies and functional studies with transgenic animals are needed to characterize these mutations. Also, prior to examining these cases for promoter mutations or expression defects, reviewing the patient and family history is strongly recommended in order to investigate the possibility that other genes are involved. Recently, numerous studies point out the need for screening the MYH gene for biallelic germline mutations in "APC negative" FAP patients [33-35]. In these patients, polyposis arises from multiple somatic mutations that accumulate in the APC gene due to the biallelic inactivation of the MYH.
Conclusion
In conclusion, the combination of the above methods; PCR, dHPLC, RT-PCR, sequencing and MLPA is a good tool for successful genetic testing of FAP patients. It allows the identification of common but also uncommon or rare mutations of the APC gene. Splice site defects or gross genomic alterations may have been underestimated by the older testing methods and require cDNA screening and gene rearrangement testing.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MM carried out the mutation screening, identification and interpretation of the results and preparation of the manuscript and figures, AA contributed in the mutation screening, the preparation of the manuscript and set up the MLPA, HD performed the FISH experiments and their interpretation, VV performed the karyotyping, AP performed the funduscopic examination, AK, KP, ID, JT, GF and NA provided the patient material, diagnosis and managment and GN 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 wish to thank Dr. C. Breukel and Dr. R. Fodde for kindly providing the cosmid probes for FISH. Written consent was obtained from the patients or their relatives for publication of study. This work is supported by the STAVROS NIARCHOS FOUNDATION FOR CHARITY and the "HYGEIA" Diagnostic & Therapeutic Center of Athens S.A., Greece.
Figures and Tables
Figure 1 A. DNA sequence chromatogram at the splice junction between exon 9 and intron 9. Donor site is in a box and IVS9+5G>A mutation is indicated by an arrow. Genomic structure from the region including exons 8 and 10 with exons in boxes and introns printed as broken lines. B. The two normal splice variants 9 (exons 8-9-9A-10), 9A (exons 8-9A-10) and the abnormal variant missing exon 9 (exons 8–10). RT-PCR products on 6% PAGE are indicated by horizontal arrows. M: 100 bp DNA Ladder (New England Biolabs), p: patient sample and n: normal samples. The 336 bp abnormal product is quite evident in the patient sample. C. Chromatogram showing sequence of the abnormal transcript. Vertical line in chromatogram showing the abnormal junction between exons 8 and 10 while the box indicates the STOP signal created by the frameshift.
Figure 2 Family tree of the FAP family carrying a gross APC deletion. Though no sample is available for individuals I:3, II:3 and II:4, phenotype, cause and age of death indicate the presence and direct transmission of the deletion identified in the proband (III:2). (CRC: Colorectal Cancer, Ca Li: Liver cancer).
Figure 3 Results from MLPA experiment. Chart of normalized copy number of the MLPA probes target sequences, representing APC exons and control regions on other chromosomes, versus size of amplified products. Two different DNA samples from the same patient present only one copy per cell for exons 6 to 15 supporting the presence of a large APC deletion from one chromosome.
==== Refs
Groden J Thliveris A Samowitz W Carlson M Gelbert L Albertsen H Joslyn G Stevens J Spirio L Robertson M Identification and characterization of the familial adenomatous polyposis coli gene Cell 1991 66 589 600 1651174 10.1016/0092-8674(81)90021-0
Kinzler KW Nilbert MC Su LK Vogelstein B Bryan TM Levy DB Smith KJ Preisinger AC Hedge P McKechnie D Identification of FAP locus genes from chromosome 5q21 Science 1991 253 661 5 1651562
Fearnhead NS Britton MP Bodmer WF The ABC of APC Hum Mol Genet 2001 10 721 33 11257105 10.1093/hmg/10.7.721
Mihalatos M Danielides I Belogianni J Harokopos E Papadopoulou E Kalimanis G Tsiava M Triantafillidis JK Kosmidis PA Fountzilas G Basdanis G Agnantis NJ Yannoukakos D Nasioulas G Novel mutations of the APC gene in Greek familial adenomatous polyposis patients Cancer Genet Cytogenet 2003 141 65 70 12581900 10.1016/S0165-4608(02)00723-9
Beroud C Soussi T APC gene: database of germline and somatic mutations in human tumors and cell lines Nucleic Acids Res 1996 24 121 124 8594558 10.1093/nar/24.1.121
Su LK Steinbach G Sawyer JC Hindi M Ward PA Lynch PM Genomic rearrangements of the APC tumor-suppressor gene in familial adenomatous polyposis Hum Genet 2000 106 101 7 10982189 10.1007/s004390051016
Charames GS Cheng H Gilpin CA Hunter AG Berk T Bapat B A novel aberrant splice site mutation in the APC gene J Med Genet 2002 39 754 7 12362034 10.1136/jmg.39.10.754
Sieber OM Lamlum H Crabtree MD Rowan AJ Barclay E Lipton L Hodgson S Thomas HJ Neale K Phillips RK Farrington SM Dunlop MG Mueller HJ Bisgaard ML Bulow S Fidalgo P Albuquerque C Scarano MI Bodmer W Tomlinson IP Heinimann K Whole-gene APC deletions cause classical familial adenomatous polyposis, but not attenuated polyposis or "multiple" colorectal adenomas Proc Natl Acad Sci U S A 2002 99 2954 8 11867715 10.1073/pnas.042699199
Su LK Kohlmann W Ward PA Lynch PM Different familial adenomatous polyposis phenotypes resulting from deletions of the entire APC exon 15 Hum Genet 2002 111 88 95 12136240 10.1007/s00439-002-0758-7
Ponz de Leon M Varesco L Benatti P Sassatelli R Izzo P Scarano MI Rossi GB Di Gregorio C Gismondi V Percesepe A de Rosa M Roncucci L Phenotype-genotype correlations in an extended family with adenomatosis coli and an unusual APC gene mutation Dis Colon Rectum 2001 44 1597 604 11711730
Samowitz WS Thliveris A Spirio LN White R Alternatively spliced adenomatous polyposis coli (APC) gene transcripts that delete exons mutated in attenuated APC Cancer Res 1995 55 3732 4 7641184
Moisio AL Jarvinen H Peltomaki P Genetic and clinical characterisation of familial adenomatous polyposis: a population based study Gut 2002 50 845 50 12010888 10.1136/gut.50.6.845
Spirio L Green J Robertson J Robertson M Otterud B Sheldon J Howse E Green R Groden J White R Leppert M The identical 5' splice-site acceptor mutation in five attenuated APC families from Newfoundland demonstrates a founder effect Hum Genet 1999 105 388 98 10598803 10.1007/s004390051121
Montagna M Dalla Palma M Menin C Agata S De Nicolo A Chieco-Bianchi L D'Andrea E Genomic rearrangements account for more than one-third of the BRCA1 mutations in northern Italian breast/ovarian cancer families Hum Mol Genet 2003 12 1055 61 12700174 10.1093/hmg/ddg120
Wagner A Barrows A Wijnen JT van der Klift H Franken PF Verkuijlen P Nakagawa H Geugien M Jaghmohan-Changur S Breukel C Meijers-Heijboer H Morreau H van Puijenbroek M Burn J Coronel S Kinarski Y Okimoto R Watson P Lynch JF de la Chapelle A Lynch HT Fodde R Molecular analysis of hereditary nonpolyposis colorectal cancer in the United States: high mutation detection rate among clinically selected families and characterization of an American founder genomic deletion of the MSH2 gene Am J Hum Genet 2003 72 1088 100 12658575 10.1086/373963
Cao X Eu KW Seow-Choen F Zhao Y Cheah PY Topoisomerase-I- and Alu-mediated genomic deletions of the APC gene in familial adenomatous polyposis Hum Genet 2001 108 436 42 11409872 10.1007/s004390100492
Apessos A Mihalatos M Danielidis I Kalimanis G Agnantis NJ Triantafillidis JK Fountzilas G Kosmidis PA Razis E Georgoulias VA HeCOG HeHeGI Nasioulas G hMSH2 is the most commonly mutated MMR gene in a cohort of Greek HNPCC patients Br J Cancer 2005 92 396 404 15655560
Belogianni I Apessos A Mihalatos M Razi E Labropoulos S Petounis A Gaki V Keramopoulos A Pandis N Kyriacou K Hadjisavvas A Kosmidis P Yannoukakos D Nasioulas G Characterization of a novel large deletion and single point mutations in the BRCA1 gene in a Greek cohort of families with suspected hereditary breast cancer BMC Cancer 4 61 2004 Sep 7 15353005 10.1186/1471-2407-4-61
Krawczak M Reiss J Cooper DN The mutational spectrum of single base-pair substitutions in mRNA splice junctions of human genes: causes and consequences Hum Genet 1992 90 41 54 1427786
Teraoka SN Telatar M Becker-Catania S Liang T Onengut S Tolun A Chessa L Sanal O Bernatowska E Gatti RA Concannon P Splicing defects in the ataxia-telangiectasia gene, ATM: underlying mutations and consequences Am J Hum Genet 1999 64 1617 31 10330348 10.1086/302418
Ars E Serra E Garcia J Kruyer H Gaona A Lazaro C Estivill X Mutations affecting mRNA splicing are the most common molecular defects in patients with neurofibromatosis type 1 Hum Mol Genet 9 237 47 2000 Jan 22 10607834 10.1093/hmg/9.2.237
Varesco L Gismondi V Presciuttini S Groden J Spirio L Sala P Rossetti C De Benedetti L Bafico A Heouaine A Mutation in a splice-donor site of the APC gene in a family with polyposis and late age of colonic cancer death Hum Genet 1994 93 281 6 8125478 10.1007/BF00212023
Mihalatos M Apessos A Triantafillidis JK Kosmidis PA Fountzilas G Agnantis NJ Yannoukakos D Nasioulas G Evaluation of dHPLC in mutation screening of the APC gene in a Greek FAP cohort Anticancer Res 2003 23 2691 5 12894559
Schouten JP McElgunn CJ Waaijer R Zwijnenburg D Diepvens F Pals G Relative quantification of 40 nucleic acid sequences by multiplex ligation-dependent probe amplification Nucleic Acids Res 30 e57 2002 Jun 15 12060695 10.1093/nar/gnf056
Dauwerse JG Jumelet EA Wessels JW Saris JJ Hagemeijer A Beverstock GC van Ommen GJ Breuning MH Extensive cross-homology between the long and the short arm of chromosome 16 may explain leukemic inversions and translocations Blood 1992 79 1299 304 1536953
Knight SJ Lese CM Precht KS Kuc J Ning Y Lucas S Regan R Brenan M Nicod A Lawrie NM Cardy DL Nguyen H Hudson TJ Riethman HC Ledbetter DH Flint J An optimized set of human telomere clones for studying telomere integrity and architecture Am J Hum Genet 2000 67 320 32 10869233 10.1086/302998
Van der Luijt RB Tops CM Khan PM van der Klift HM Breukel C van Leeuwen-Cornelisse IS Dauwerse HG Beverstock GC van Noort E Snel P Molecular, cytogenetic, and phenotypic studies of a constitutional reciprocal translocation t(5;10)(q22;q25) responsible for familial adenomatous polyposis in a Dutch pedigree Genes Chromosomes Cancer 1995 13 192 202 7669739
Olschwang S Tiret A Laurent-Puig P Muleris M Parc R Thomas G Restriction of ocular fundus lesions to a specific subgroup of APC mutations in adenomatous polyposis coli patients Cell 75 959 68 1993 Dec 3 8252631 10.1016/0092-8674(93)90539-3
Krainer AR Maniatis T Multiple factors including the small nuclear ribonucleoproteins U1 and U2 are necessary for pre-mRNA splicing in vitro Cell 1985 42 725 36 2996774 10.1016/0092-8674(85)90269-7
Shapiro MB Senapathy P RNA splice junctions of different classes of eukaryotes: sequence statistics and functional implications in gene expression Nucleic Acids Res 1987 15 7155 74 3658675
Davies DR Armstrong JG Thakker N Horner K Guy SP Clancy T Sloan P Blair V Dodd C Warnes TW Severe Gardner syndrome in families with mutations restricted to a specific region of the APC gene Am J Hum Genet 1995 57 1151 8 7485167
Wallis YL Macdonald F Hulten M Morton JE McKeown CM Neoptolemos JP Keighley M Morton DG Genotype-phenotype correlation between position of constitutional APC gene mutation and CHRPE expression in familial adenomatous polyposis Hum Genet 1994 94 543 8 7959691 10.1007/BF00211023
Jones S Emmerson P Maynard J Best JM Jordan S Williams GT Sampson JR Cheadle JP Biallelic germline mutations in MYH predispose to multiple colorectal adenoma and somatic G:C – >T:A mutations Hum Mol Genet 11 2961 7 2002 Nov 1 12393807 10.1093/hmg/11.23.2961
Sieber OM Lipton L Crabtree M Heinimann K Fidalgo P Phillips RK Bisgaard ML Orntoft TF Aaltonen LA Hodgson SV Thomas HJ Tomlinson IP Multiple colorectal adenomas, classic adenomatous polyposis, and germ-line mutations in MYH N Engl J Med 348 791 9 2003 Feb 27 12606733 10.1056/NEJMoa025283
Marra G Jiricny J Multiple colorectal adenomas – is their number up? N Engl J Med 348 845 7 2003 Feb 27 12606740 10.1056/NEJMe030002
| 15833136 | PMC1097718 | CC BY | 2021-01-04 16:03:04 | no | BMC Cancer. 2005 Apr 15; 5:40 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-40 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-411583678310.1186/1471-2407-5-41Research ArticleExpression of a novel carbonic anhydrase, CA XIII, in normal and neoplastic colorectal mucosa Kummola Laura [email protected]ämäläinen Jonna M [email protected]ä Jyrki [email protected]ä Antti J [email protected] Juha [email protected] Tuomo [email protected] Seppo [email protected] Institute of Medical Technology, University of Tampere, and Tampere University Hospital, Tampere, Finland2 Institute of Dentistry, University of Helsinki and Research Institute of Military Medicine, Central Military Hospital, Helsinki, Finland3 Department of Surgery, University of Oulu, Oulu, Finland4 Department of Pathology, University of Oulu, Oulu, Finland5 Department of Clinical Chemistry, University of Oulu, Oulu, Finland2005 18 4 2005 5 41 41 13 12 2004 18 4 2005 Copyright © 2005 Kummola 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
Carbonic anhydrase (CA) isozymes may have an important role in cancer development. Some isozymes control pH homeostasis in tumors that appears to modulate the behaviour of cancer cells. CA XIII is the newest member of the CA gene family. It is a cytosolic isozyme which is expressed in a number of normal tissues. The present study was designed to investigate CA XIII expression in prospectively collected colorectal tumor samples.
Methods
Both neoplastic and normal tissue specimens were obtained from the same patients. The analyses were performed using CA XIII-specific antibodies and an immunohistochemical staining method. For comparison, the tissue sections were immunostained for other cytosolic isozymes, CA I and II.
Results
The results indicated that the expression of CA XIII is down-regulated in tumor cells compared to the normal tissue. The lowest signal was detected in carcinoma samples. This pattern of expression was quite parallel for CA I and II.
Conclusion
The down-regulation of cytosolic CA I, II and XIII in colorectal cancer may result from reduced levels of a common transcription factor or loss of closely linked CA1, CA2 and CA13 alleles on chromosome 8. Their possible role as tumor suppressors should be further evaluated.
==== Body
Background
Carbonic anhydrases (CAs) are a growing family of zinc-containing metalloenzymes that have an important role in maintaining the pH homeostasis by catalysing the reversible hydration of carbon dioxide, CO2 + H2O ↔ HCO3 - + H+. This family includes at least twelve enzymatically active isozymes, whose subcellular localization, tissue distribution and kinetic properties differ. Five of the isozymes are cytosolic (CA I, II, III, VII and XIII), four are membrane associated (CA IV, IX, XII and XIV), two are mitochondrial (CA VA and VB) and one is a secretory form (CA VI) [1-6]. CAs contribute to several important biological functions, such as respiration, bone resorption, renal acidification, gluconeogenesis and formation of cerebrospinal fluid and gastric acid [1,7].
It is typical for solid tumors to create a pH gradient between the intracellular and extracellular compartments by acidifying the surrounding tissue to lower pH value than in adjacent normal tissue. Extracellular acidification has been associated with many tumor progression effects, such as loss of intercellular adhesion, increased metastasis and migration, increased rate of mutations and reduced DNA repair and upregulation of growth factors and proteases [4]. In order to maintain the pH gradient, tumor cells express ion transport proteins, including vacuolar H+-ATPase, Cl-/HCO3 - exchanger and Na+/H+ exchanger [8,9]. Previously, lactic acid has been considered a major source of acidity in tumors [4]. More recent studies have indicated, however, that carbon dioxide also makes a significant contribution to the tumor acidity, which strongly implies to the importance of CA activity in oncogenesis. This hypothesis has been indirectly supported by the recent findings that several CA inhibitors may suppress the invasion capacity of malignant cell lines [4].
CA XIII is a novel enzyme, which is expressed in several human tissues including salivary glands, kidney, small intestine, colon, uterus and testis. Modelling of CA XIII protein revealed that is a globular molecule, and it has the highest (about 60%) identity with cytosolic isozymes CA I, II and III. Therefore, CA XIII is clustered as cytosolic/intracellular isoform. In the previous report, we described the production of recombinant mouse CA XIII protein, which was used in detailed activity and inhibition studies [2,10]. CA XIII demonstrated moderate CA catalytic activity with kcat/Km of 4.3 × 107 M-1s-1, and kcat of 8.3 × 104 s-1 at 25°C and pH 7.5 [2]. Interestingly, CA XIII showed unique inhibitory properties compared to CA I, II and IX [10]. The clinically used sulfonamides including acetazolamide, methazolamide, dichlorophenamide, and dorzolamide demonstrated potent CA XIII inhibition, with Ki's in the range of 17–23 nM. Another group of inhibitors, the sulfanilyl-derived compounds, obtained by reaction of aminosulfonamides with 4-acetamido-benzenesulfonyl chloride showed very potent CA XIII inhibitory properties, with Ki's in the range of 1.3–2.4 nM, whereas these compounds were generally much less effective as inhibitors of isozymes I, II, and IX.
Different CA isozymes have become attractive targets in studies aimed at molecular mechanisms of carcinogenesis. CA IX and XII have been clearly associated with certain types of cancer [11-21]. Also cytosolic isozymes CA I and II have been investigated in some tumors including colorectal cancer [22,23]. Since there was no previous data on CA XIII in tumors, we collected a series of colorectal neoplasias and adjacent normal tissues, and analyzed them by an immunohistochemical staining method for the expression of this novel isozyme.
Methods
Antibodies
The rabbit anti-mouse/human CA XIII antibody has been described in a recent article [2]. It was raised against a synthetic peptide (AC-) DGDQQSPIEIKTKEC (-COOH). The affinity and specificity of the antibody was confirmed by ELISA titre analysis, western blotting and immunohistochemistry. In immunohistochemical staining, CA XIII has shown a distinctly different distribution pattern compared to the localization of CA I or II [2]. This finding indicates that the anti-CA XIII antibody does not cross-react with CA I or II under the present staining conditions. Rabbit antisera against human CA I and II have been produced and characterized earlier [24].
Immunohistochemistry
The normal tissue samples from the large intestine and the corresponding benign and/or malignant neoplastic samples were prospectively collected at Oulu University Hospital (Oulu, Finland). The normal samples were excised separately 20 centimeters proximally from the tumor resection site. The study was approved by the Ethics Committee of Oulu University Hospital and performed according to the guidelines of the Declaration of Helsinki.
32 distinct areas (both normal tissue and pathological lesions) of the human colorectal mucosa were examined from 12 patients. They consisted of 11 separate samples of histologically normal human colon or rectum and 17 colorectal lesions, including 6 adenomas and 11 adenocarcinomas. In 4 samples, normal tissue was found adjacent to the neoplasm and was included in the analyses. The grade of dysplasia was moderate in 3 and grave in 3 adenomas. The group of 11 malignant colorectal tumors consisted of 4 well-differentiated and 4 moderately differentiated adenocarcinomas, and 3 adenocarcinomas with a mucinous component. It should be noted that each patient had either one or more (1–3) lesions which were analyzed.
The specimens were fixed in Carnoy's fluid (absolute ethanol + chloroform + glacial acetic acid 6:3:1) for 6 hours. The samples were then dehydrated, embedded in paraffin in a vacuum oven at 58°C, and sections of 5 μm were placed on gelatine-coated microscope slides The CA isozymes were immunostained by the biotin-streptavidin complex method, employing the following steps: (a) pre-treatment of the sections with undiluted cow colostral whey (Biotop Oy, Oulu, Finland) for 30 min and rinsing in phosphate-buffered saline (PBS), (b) incubation for 1 hr with primary antibodies diluted 1:100 in 1% bovine serum albumin -PBS (BSA-PBS), (c) incubation for 1 hr with biotinylated goat anti-rabbit IgG (Zymed laboratories inc., South San Fransisco, CA) diluted 1:300 in 1% BSA-PBS, (d) incubation for 30 min with streptavidin-horseradish peroxidase (HRP) conjugate (Zymed laboratories) diluted 1:750 in PBS and (e) incubation for 2 min in DAB solution containing 4,5 mg 3,3'-diaminobenzidine tetrahydrochloride (Fluka, Buchs, Switzerland) in 7,5 ml PBS + 2,5 μl 30% H2O2. The sections were washed three times for 10 min in PBS after incubation steps b and c, and four times for 5 min in PBS after step d. All the incubations and washings were carried out at room temperature. The sections were finally mounted in Neo-Mount (Merck, Darmstadt, Germany). The stained sections were examined and photographed with Zeiss Axioskop 40 microscope (Carl Zeiss, Göttingen, Germany). The intensity of immunostaining was scored on a scale of 0 to 3 as follows: 0, no reaction; 1, weak reaction; 2, moderate reaction; 3, strong reaction.
Results and discussion
Carbonic anhydrase XIII is a recently discovered enzyme, which is expressed in a number of different mammalian tissues. Structural modelling revealed that it is most closely related to cytosolic isozymes, CA I, II, and III with about 60% sequence identity, and therefore, this isozyme was clustered to the group of cytosolic/intracellular isoforms. Even though the previous paper described the distribution of CA XIII in several normal tissues, no data has been published to date on its expression in neoplastic tissues. To get a better insight into the changes in the expression levels that occur during the development of neoplasia, we analyzed both normal and pathological colorectal specimens from the same individuals.
Figures 1 and 2 show some images of CA XIII immunostaining in the colorectal tumors and normal mucosa. Reactions for CA I and II are shown for comparison. The immunostaining scoring results are shown in Figure 3. The results indicated that CA XIII immunoreactions generally follow a similar pattern with CA I and II, i.e. the signals became weaker along with increasing dysplasia and malignancy grades. There were a few exceptions, and particularly, CA II expression was relatively high in adenocarcinomas with a mucinous component, while CA I and XIII were decreased like in other adenocarcinomas (Fig. 2).
Previous studies have indicated that CA IX and XII are overexpressed in certain tumors, including colorectal adenomas and carcinomas [18,23]. They have been functionally linked to extracellular acidification in tumors that appears to regulate the invasive behaviour of tumor cells. Consequently, expression of CA IX has been associated with poor prognosis of some malignancies [19,20,25,26]. Based on the present and previous results [23], cytosolic isozymes I, II, and XIII do not probably play a significant role in the regulation of pH in tumors. Instead, they may be important factors for the normal intestinal physiology. This suggestion raises an important question why at least three cytosolic isozymes are expressed in the normal colorectal mucosa. Even though we cannot answer to this question at this stage, it is of interest that these CAs represent isoforms with different kinetic or inhibitory properties. CA II exhibits the highest and CA I and CA XIII moderate carbon dioxide hydratase activities [2]. Because of different biochemical characteristics, it is unlikely that all these isozymes could contribute to the same pH regulation process. If this was the case, one could suggest that the enzymes with low or moderate activities should represent some sort of "junk" proteins in the proteome. However, there are other possible explanations which could be considered more sound in a biological system. First, several CA isozymes have emerged during evolution, because CA enzymatic activity is required for very fundamental reactions in the body. If one isozyme was functionally defective, the others could compensate the missing one. These compensatory mechanisms would partly explain why different CA deficiencies typically produce no or only mild phenotypic changes [27-29].
Another possible role of cytosolic CAs could be linked to a hypothetical tumor suppressor function. It is notable that CA1, CA2 and CA13 genes are closely linked locating in the q arm of chromosome 8. Simultaneous loss of expression could result from a deletion or point mutations within regulatory regions or alleles specifying these genes. The down-regulation could also result from reduced levels of a common transcription factor. Regardless of the mechanism, three cytosolic CA isozymes are affected in colorectal tumors, and it would be tempting to propose that normal CA I, II and/or XIII proteins could have a tumor suppressor function in epithelial cells, and deletion of those alleles could be an important factor in the cascade of events contributing to colorectal carcinogenesis. Even though colorectal cancer has been associated with gain of chromosome 8q in a number of previous studies, and no losses have been reported [30-33], there is some evidence that 8q21 contains a yet unidentified tumor suppressor gene [34,35]. However, larger sets of tumors will be required to define explicitly, whether these CA genes are involved in genetic alterations of colorectal tumors.
Conclusion
CA XIII expression was clearly decreased in colorectal tumors compared to the normal tissue in a pattern similar to CA I and II. Loss of expression of closely linked CA1, CA2 and CA13 genes may result from reduced levels of a transcription factor or loss of alleles specifying these genes.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in the design of the study. JS, TK and AJK collected the tissue samples. LK and SP drafted the manuscript. JMH and SP characterized the antibodies. LK performed the immunohistochemical staining. LK, JK and SP analyzed the staining results. All authors read, modified 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 grants from Sigrid Juselius Foundation and Academy of Finland (200969).
Figures and Tables
Figure 1 Immunohistochemical staining of CA I, II and XIII in the normal colonic mucosa and tumors. The samples of normal mucosa, adenoma (grave dysplasia) and well differentiated adenocarcinoma were collected from one patient. The strongest immunoreaction is located in the enterocytes of the normal mucosa (arrows). Original magnification × 200.
Figure 2 CA expression in the normal colonic mucosa and adenocarcinoma with a mucinous component. CA II staining intensity remained at a relatively high level in the carcinoma sample (arrows), whereas the intensities for other isozymes were decreased compared to the normal epithelium.
Figure 3 A graphic summary of CA immunostaining in the normal colorectal mucosa and tumor samples. The coloured lines represent the staining indices in individual patients. The dash lines indicate the mean values of staining intensity.
==== Refs
Parkkila S Parkkila AK Carbonic anhydrase in the alimentary tract. Roles of the different isozymes and salivary factors in the maintenance of optimal conditions in the gastrointestinal canal Scand J Gastroenterol 1996 31 305 317 8726296
Lehtonen J Shen B Vihinen M Casini A Scozzafava A Supuran CT Parkkila AK Saarnio J Kivela AJ Waheed A Sly WS Parkkila S Characterization of CA XIII, a novel member of the carbonic anhydrase isozyme family J Biol Chem 2004 279 2719 2727 14600151 10.1074/jbc.M308984200
Mori K Ogawa Y Ebihara K Tamura N Tashiro K Kuwahara T Mukoyama M Sugawara A Ozaki S Tanaka I Nakao K Isolation and characterization of CA XIV, a novel membrane-bound carbonic anhydrase from mouse kidney J Biol Chem 1999 274 15701 15705 10336468 10.1074/jbc.274.22.15701
Pastorekova S Zavada J Carbonic anhydrase IX (CA IX) as a potential target for cancer therapy Cancer Ther 2004 2 245 262
Shah GN Hewett-Emmett D Grubb JH Migas MC Fleming RE Waheed A Sly WS Mitochondrial carbonic anhydrase CA VB: differences in tissue distribution and pattern of evolution from those of CA VA suggest distinct physiological roles Proc Natl Acad Sci U S A 2000 97 1677 1682 10677517 10.1073/pnas.97.4.1677
Karhumaa P Leinonen J Parkkila S Kaunisto K Tapanainen J Rajaniemi H The identification of secreted carbonic anhydrase VI as a constitutive glycoprotein of human and rat milk Proc Natl Acad Sci U S A 2001 98 11604 11608 11553764 10.1073/pnas.121172598
Sly WS Hu PY Human carbonic anhydrases and carbonic anhydrase deficiencies Annu Rev Biochem 1995 64 375 401 7574487 10.1146/annurev.bi.64.070195.002111
Montcourrier P Silver I Farnoud R Bird I Rochefort H Breast cancer cells have a high capacity to acidify extracellular milieu by a dual mechanism Clin Exp Metastasis 1997 15 382 392 9219726 10.1023/A:1018446104071
Lee AH Tannock IF Heterogeneity of intracellular pH and of mechanisms that regulate intracellular pH in populations of cultured cells Cancer Res 1998 58 1901 1908 9581831
Lehtonen JM Parkkila S Vullo D Casini A Scozzafava A Supuran CT Carbonic anhydrase inhibitors. Inhibition of cytosolic isozyme XIII with aromatic and heterocyclic sulfonamides: a novel target for the drug design Bioorg Med Chem Lett 2004 14 3757 3762 15203157 10.1016/j.bmcl.2004.04.106
Zavada J Zavadova Z Pastorekova S Ciampor F Pastorek J Zelnik V Expression of MaTu-MN protein in human tumor cultures and in clinical specimens Int J Cancer 1993 54 268 274 8486430
Liao SY Brewer C Zavada J Pastorek J Pastorekova S Manetta A Berman ML DiSaia PJ Stanbridge EJ Identification of the MN antigen as a diagnostic biomarker of cervical intraepithelial squamous and glandular neoplasia and cervical carcinomas Am J Pathol 1994 145 598 609 8080042
McKiernan JM Buttyan R Bander NH Stifelman MD Katz AE Chen MW Olsson CA Sawczuk IS Expression of the tumor-associated gene MN: a potential biomarker for human renal cell carcinoma Cancer Res 1997 57 2362 2365 9192809
Vermylen P Roufosse C Burny A Verhest A Bosschaerts T Pastorekova S Ninane V Sculier JP Carbonic anhydrase IX antigen differentiates between preneoplastic malignant lesions in non-small cell lung carcinoma Eur Respir J 1999 14 806 811 10573225 10.1034/j.1399-3003.1999.14d14.x
Liao SY Aurelio ON Jan K Zavada J Stanbridge EJ Identification of the MN/CA9 protein as a reliable diagnostic biomarker of clear cell carcinoma of the kidney Cancer Res 1997 57 2827 2831 9230182
Bartosova M Parkkila S Pohlodek K Karttunen TJ Galbavy S Mucha V Harris AL Pastorek J Pastorekova S Expression of carbonic anhydrase IX in breast is associated with malignant tissues and is related to overexpression of c-erbB2 J Pathol 2002 197 314 321 12115877 10.1002/path.1120
Saarnio J Parkkila S Parkkila AK Pastorekova S Haukipuro K Pastorek J Juvonen T Karttunen TJ Transmembrane carbonic anhydrase, MN/CA IX, is a potential biomarker for biliary tumours J Hepatol 2001 35 643 649 11690711 10.1016/S0168-8278(01)00193-3
Saarnio J Parkkila S Parkkila AK Haukipuro K Pastorekova S Pastorek J Kairaluoma MI Karttunen TJ Immunohistochemical study of colorectal tumors for expression of a novel transmembrane carbonic anhydrase, MN/CA IX, with potential value as a marker of cell proliferation Am J Pathol 1998 153 279 285 9665489
Giatromanolaki A Koukourakis MI Sivridis E Pastorek J Wykoff CC Gatter KC Harris AL Expression of hypoxia-inducible carbonic anhydrase-9 relates to angiogenic pathways and independently to poor outcome in non-small cell lung cancer Cancer Res 2001 61 7992 7998 11691824
Chia SK Wykoff CC Watson PH Han C Leek RD Pastorek J Gatter KC Ratcliffe P Harris AL Prognostic significance of a novel hypoxia-regulated marker, carbonic anhydrase IX, in invasive breast carcinoma J Clin Oncol 2001 19 3660 3668 11504747
Wykoff CC Beasley NJ Watson PH Turner KJ Pastorek J Sibtain A Wilson GD Turley H Talks KL Maxwell PH Pugh CW Ratcliffe PJ Harris AL Hypoxia-inducible expression of tumor-associated carbonic anhydrases Cancer Res 2000 60 7075 7083 11156414
Mori M Staniunas RJ Barnard GF Jessup JM Steele GDJ Chen LB The significance of carbonic anhydrase expression in human colorectal cancer Gastroenterology 1993 105 820 826 8359652
Kivela AJ Saarnio J Karttunen TJ Kivela J Parkkila AK Pastorekova S Pastorek J Waheed A Sly WS Parkkila TS Rajaniemi H Differential expression of cytoplasmic carbonic anhydrases, CA I and II, and membrane-associated isozymes, CA IX and XII, in normal mucosa of large intestine and in colorectal tumors Dig Dis Sci 2001 46 2179 2186 11680594 10.1023/A:1011910931210
Parkkila AK Parkkila S Juvonen T Rajaniemi H Carbonic anhydrase isoenzymes II and I are present in the zona glomerulosa cells of the human adrenal gland Histochemistry 1993 99 37 41 8468192 10.1007/BF00268018
Koukourakis MI Giatromanolaki A Sivridis E Simopoulos K Pastorek J Wykoff CC Gatter KC Harris AL Hypoxia-regulated carbonic anhydrase-9 (CA9) relates to poor vascularization and resistance of squamous cell head and neck cancer to chemoradiotherapy Clin Cancer Res 2001 7 3399 3403 11705854
Hoskin PJ Sibtain A Daley FM Wilson GD GLUT1 and CAIX as intrinsic markers of hypoxia in bladder cancer: relationship with vascularity and proliferation as predictors of outcome of ARCON Br J Cancer 2003 89 1290 1297 14520462 10.1038/sj.bjc.6601260
Kendall AG Tashian RE Erythrocyte carbonic anhydrase I: inherited deficiency in humans Science 1977 197 471 472 406674
Kim G Lee TH Wetzel P Geers C Robinson MA Myers TG Owens JW Wehr NB Eckhaus MW Gros G Wynshaw-Boris A Levine RL Carbonic anhydrase III is not required in the mouse for normal growth, development, and life span Mol Cell Biol 2004 24 9942 9947 15509796 10.1128/MCB.24.22.9942-9947.2004
Ortova Gut MO Parkkila S Vernerova Z Rohde E Zavada J Hocker M Pastorek J Karttunen T Gibadulinova A Zavadova Z Knobeloch KP Wiedenmann B Svoboda J Horak I Pastorekova S Gastric hyperplasia in mice with targeted disruption of the carbonic anhydrase gene Car9 Gastroenterology 2002 123 1889 1903 12454846 10.1053/gast.2002.37052
Diep CB Teixeira MR Thorstensen L Wiig JN Eknaes M Nesland JM Giercksky KE Johansson B Lothe RA Genome characteristics of primary carcinomas, local recurrences, carcinomatoses, and liver metastases from colorectal cancer patients Mol Cancer 2004 3 6 14977426 10.1186/1476-4598-3-6
Douglas EJ Fiegler H Rowan A Halford S Bicknell DC Bodmer W Tomlinson IP Carter NP Array comparative genomic hybridization analysis of colorectal cancer cell lines and primary carcinomas Cancer Res 2004 64 4817 4825 15256451
Richter H Slezak P Walch A Werner M Braselmann H Jaramillo E Ost A Hirata I Takahama K Zitzelsberger H Distinct chromosomal imbalances in nonpolypoid and polypoid colorectal adenomas indicate different genetic pathways in the development of colorectal neoplasms Am J Pathol 2003 163 287 294 12819033
Aragane H Sakakura C Nakanishi M Yasuoka R Fujita Y Taniguchi H Hagiwara A Yamaguchi T Abe T Inazawa J Yamagishi H Chromosomal aberrations in colorectal cancers and liver metastases analyzed by comparative genomic hybridization Int J Cancer 2001 94 623 629 11745455 10.1002/ijc.1522
Varon R Gosse-Brun S Bignon YJ Sperling K Uhrhammer N Nijmegen breakage syndrome gene (NBS1) is not the tumor suppressor gene at 8q21.3 involved in colorectal carcinoma Oncol Rep 2002 9 709 711 12066197
Uhrhammer N Bay JO Gosse-Brun S Kwiatkowski F Rio P Daver A Bignon YJ Allelic imbalance at NBS1 is frequent in both proximal and distal colorectal carcinoma Oncol Rep 2000 7 427 431 10671697
| 15836783 | PMC1097719 | CC BY | 2021-01-04 16:03:03 | no | BMC Cancer. 2005 Apr 18; 5:41 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-41 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-421584770210.1186/1471-2407-5-42Research ArticleThe prognostic value of the hypoxia markers CA IX and GLUT 1 and the cytokines VEGF and IL 6 in head and neck squamous cell carcinoma treated by radiotherapy ± chemotherapy De Schutter Harlinde [email protected] Willy [email protected] Erik [email protected] Laurence [email protected] Robert [email protected] Sandra [email protected] Department of Radiation Oncology, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium2 Department of Pathology, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium3 Department of Radiology, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium2005 25 4 2005 5 42 42 15 12 2004 25 4 2005 Copyright © 2005 De Schutter 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
Several parameters of the tumor microenvironment, such as hypoxia, inflammation and angiogenesis, play a critical role in tumor aggressiveness and treatment response. A major question remains if these markers can be used to stratify patients to certain treatment protocols.
The purpose of this study was to investigate the inter-relationship and the prognostic significance of several biological and clinicopathological parameters in patients with head and neck squamous cell carcinoma (HNSCC) treated by radiotherapy ± chemotherapy.
Methods
We used two subgroups of a retrospective series for which CT-determined tumoral perfusion correlated with local control. In the first subgroup (n = 67), immunohistochemistry for carbonic anhydrase IX (CA IX) and glucose transporter-1 (GLUT-1) was performed on the pretreatment tumor biopsy. In the second subgroup (n = 34), enzyme linked immunosorbent assay (ELISA) was used to determine pretreatment levels of the cytokines vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6) in serum. Correlation was investigated between tumoral perfusion and each of these biological markers, as well as between the markers mutually. The prognostic value of these microenvironmental parameters was also evaluated.
Results
For CA IX and GLUT-1, the combined assessment of patients with both markers expressed above the median showed an independent correlation with local control (p = 0.02) and disease-free survival (p = 0.04) with a trend for regional control (p = 0.06).
In the second subgroup, IL-6 pretreatment serum level above the median was the only independent predictor of local control (p = 0.009), disease-free survival (p = 0.02) and overall survival (p = 0.005).
Conclusion
To our knowledge, we are the first to report a link in HNSCC between IL-6 pretreatment serum levels and radioresistance in vivo. This link is supported by the strong prognostic association of pretreatment IL-6 with local control, known to be the most important parameter to judge radiotherapy responses.
Furthermore, the combined assessment of CA IX and GLUT-1 correlated independently with prognosis. This is a valuable indication that a combined approach is important in the investigation of prognostic markers.
==== Body
Background
Different biological and clinicopathological parameters next to the commonly used tumor and nodal stage have been proposed to help stratifying patients to certain treatment protocols. Specifically, molecular properties of the tumor microenvironment such as hypoxia, inflammation and angiogenesis are currently investigated to serve as a target for therapy[1].
The aim of the present research was to evaluate potential prognostic markers in a retrospective series of HNSCC treated by radiotherapy ± chemotherapy. A recently published report on a retrospective series of patients with a primary HNSCC treated by radiotherapy ± chemotherapy and investigated by dynamic contrast-enhanced perfusion-computer tomography (perfusion-CT), indicated that low tumoral perfusion predicted poor local control (LC), when stratified according to the median perfusion value [2]. This patient group was now used for further exploration.
Our first hypothesis was that CT-determined tumoral perfusion could be a surrogate marker for hypoxia, which is an important negative prognostic factor in radiation treatment of HNSCC [3,4].
For this purpose, we evaluated the correlation of tumoral perfusion with the expression of carbonic anhydrase IX (CA IX) and glucose transporter-1 protein (GLUT-1) using immunohistochemistry. These antigens are possible intrinsic markers of hypoxia, both mainly controlled by the hypoxia inducible factor-1 (HIF-1) pathway. Their clinical relevance to assess the expression pattern and the relation with hypoxia has been suggested by others for several tumors [5-8].
CA IX is a transmembrane carbonic anhydrase involved in the reversible hydration of carbon dioxide to carbonic acid. It is a widely investigated protein, overexpressed in many tumor types [9]. In HNSCC, its expression is associated with worse prognosis [6,10].
GLUT-1 mediates cellular glucose uptake and thus facilitates anaerobic glycolysis. This protein is largely undetectable in normal epithelium and benign tumors, but is expressed in several types of human carcinomas such as HNSCC and associated with poor prognosis [5,7,8,11].
Correlation was assessed between tumoral perfusion and each of these intrinsic markers, as well as between CA IX and GLUT-1 mutually. The prognostic value of these immunohistochemical parameters was also evaluated.
Our second hypothesis was that the CT-determined tumoral perfusion could be linked with the measurement of vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6).
VEGF is a pro-angiogenic cytokine known to stimulate the proliferation and migration of endothelial cells and to promote vessel permeability. It is one of the hypoxia-responsive genes, upregulated by increased levels of HIF-1 [12,13]. VEGF serum levels are elevated in patients with HNSCC compared to healthy controls [12,14,15]. A negative prognostic role for circulating VEGF serum levels has been indicated for laryngeal carcinoma [15].
IL-6 is a proinflammatory cytokine produced by inflammatory cells as well as tumor cells [14,16]. In cancer patients, it has a multifunctional role concerning systemic alterations like cachexia as well a loco-regional impact on for example tumor cell motility and intercellular adhesion. Although mainly involved in inflammatory reactions, its expression can be induced by the transcription factor nuclear factor-kappaB (NF-κB) under hypoxic conditions [17]. Increased serum IL-6 levels have been detected in several tumor types, having a prognostic meaning for some of them [18-20]. The possible involvement of IL-6 in radioresistance was mentioned in several in vitro publications [21-23].
Available serum samples were analysed to evaluate VEGF and IL-6 at pretreatment time using an enzyme linked immunosorbent assay (ELISA).
We intended to evaluate the prognostic value of these serum cytokines, as well as the possible link between tumoral perfusion and the cytokine levels. This link could be established by the hypoxia mediated HIF-1 or NF-κB pathways.
The present work results on the possible prognostic significance of different biological and clinicopathological parameters studied in a retrospective study of patients with HNSCC. Furthermore, the relationship with CT-determined tumoral perfusion was investigated.
Methods
Patients and tissues
Between March 1995 and January 2000, patients with a primary HNSCC treated by radiotherapy ± chemotherapy were investigated by perfusion-CT to evaluate tumoral perfusion. The largest axial tumor diameter was selected and dynamic CT sequences were taken during injection of a contrast agent bolus. Perfusion was correlated with the carotid artery [2].
The original patient population was divided into two subgroups. For the first subgroup of 67 patients (group A), we had the disposal of formalin-fixed, paraffin-embedded biopsies, taken at the time of diagnostic endoscopy. Serial sections, 5 μm thick, were stained for hematoxylin and eosin (H&E) to allow evaluation of the representativity of the biopsy specimen, and immunohistochemically for CA IX and GLUT-1.
For the second subgroup of 34 patients (group B), we had the disposal of serum samples taken at pretreatment time and stored at -20°C until time of ELISA-analyses for VEGF and IL-6. For 24 patients of this subgroup, we had the disposal of other patient-related parameters, such as weight loss, performance status, and sedimentation rate determined at the same pretreatment time.
Histological diagnosis was made by an independent pathologist. Clinical data regarding local recurrence, regional recurrence, disease-free and overall survival were available for a minimum follow-up of four years. Patients' characteristics and treatment schedules are shown in Table 1.
Immunohistochemistry – Group A
CA IX
After dewaxing and rehydrating, an endogenous peroxidase block (0.3% H2O2 in methanol) was applied for 30 minutes. Slides were rinsed with Phosphate Buffered Saline (PBS) and Phosphate Buffer Saline with 0.4% Tween 20 (PBST) and then incubated with a mouse monoclonal anti-human CA IX antibody (M75, 1/50 dilution in PBS; Bayer, USA) [24] for 30 minutes at room temperature. After 3 × 5 min rinses in PBST, incubation with a secondary anti-mouse labeled polymer/horse radish peroxidase conjugate solution (HRP, 30 minutes at room temperature) was carried out before rinsing 3 more times. Visualization was performed by diaminobenzidine (DAB) substrate (DAKO, Substrate Chromogen System). After rinsing in PBS, sections were counter-stained with Harris hematoxylin (Merck). Slides were then dehydrated and mounted with "micromount mounting medium" (Surgipath 03731). Substitution of the primary antibody with PBS was used as a negative control. To exclude nonspecific binding of the primary antibody, we performed the same protocol using an irrelevant antibody of the same species (monoclonal mouse anti-human Smooth Muscle Actin, DAKO, 1/100 dilution). This showed no binding (data not shown). As a positive control, we used normal gastric mucosa. Representative examples of positive and negative staining results are shown in Figure 1 and Figure 2, respectively.
GLUT-1
After dewaxing and rehydrating, antigen retrieval was performed by microwaving sections immerged in 10 mM citrate buffer (pH 6). After cooling down, sections were rinsed with Tris Buffered Saline, pH 7,5 with 0.1% Tween 20 (TBST) and an endogenous peroxidase block (3% H2O2 in destilled water) was applied for 15 minutes. Slides were rinsed with TBST and incubated with "Protein Block Serum Free" (DAKO) for 20 minutes to block nonspecific binding of immunoglobulins to the tissue. In a next step, sections were incubated with a polyclonal rabbit anti-human GLUT-1 antibody (DAKO, 1/200 dilution in TBS) at 4°C overnight. After 3 × 5 minutes rinses with TBST, incubation with a secondary anti-rabbit labeled polymer/HRP conjugate solution (30 minutes at room temperature) was carried out before rinsing 3 more times. Visualization, counter-staining, dehydrating and mounting was done as for the CA IX staining. Omitting of the primary antibody served as a negative control. To exclude nonspecific binding of the primary antibody, we performed the same protocol using an irrelevant antibody of the same species (polyclonal rabbit anti-human S-100, DAKO, 1/300 dilution). This showed no binding (data not shown). Erythrocytes, present in every section, served as internal positive control for GLUT-1. Representative examples of positive and negative staining results are shown in Figure 3 and Figure 4, respectively.
Quantification technique
All morphologic investigations were performed by the same investigator, who was blinded to the clinical and follow-up data.
The representativity of the biopsies was evaluated on the H&E slices. All the tissue figuring in the biopsy was classified into one of the following compartments: viable invasive tumor cells, carcinoma in situ, normal epithelium, necrotic tissue, blood vessels, collagenous connective tissue, muscle tissue, inflammatory connective tissue, and a "residu". The proportion of each of these compartments was measured by means of a point counting system. At an optical magnification of 100 × and using a grid with 10 matched points, at least 100 points per tissue section were counted as hitting one or another compartment. The proportion of points fitting each compartment yielded directly the volume proportions of the corresponding compartments.
CA IX and GLUT-1 immunostained sections were semiquantitatively scored. The fraction of cells stained was expressed as a percentage of the total tumor area, both fractions being estimated by light microscopic inspection of between 1 and 25 low power fields (×40) per patient.
Serum VEGF and IL-6 immunoassay – Group B
ELISA kits for specific cytokines (VEGF and IL-6) were used according to the manufacturer's protocol (Quantikine Human VEGF/IL-6 Immunoassay, R&D Systems, Minneapolis, MN). Serum samples were incubated in duplicate on microtiter plates coated with a specific monoclonal antibody for 2 hours at room temperature. The plates were washed to remove unbound antibody. After incubation of a conjugate solution, a substrate solution was added. Color development was stopped after 20–25 minutes, depending on the assay. A microplate reader was used to determine colorimetric densities at 570 nm and 450 nm for each sample. The readings at 570 nm were subtracted from those at 450 nm to get the optical density for each sample. Final results were calculated from a standard curve generated by a form parametric logistic curve fit. Results were expressed in pg/ml.
Statistics
The association between discrete and continuous variables was tested by the Mann Whitney's U-test or Kruskall Wallis test as appropriate. Correlations between continuous variables were obtained using Spearman's rank correlation. Survival curves were constructed using the Kaplan-Meier method, all time intervals being calculated from the date of the first radiotherapy session. Individual factors were evaluated for their predictive value by the log rank test. The Cox proportional hazard model was used for multiple regression analysis. All tests were two-sided, using a significance level of 0.05. All statistics were done using Statistica software 6.
Results
Group A: immunohistochemistry
Representativity of the biopsies
On average, the biopsies were composed of viable tumor cells (52.4%, range 9.76%–92.13%), carcinoma in situ (1.3%), normal epithelium (2.0%), necrotic tissue (2.8%), blood vessels (6.0%), collagenous connective tissue (17.8%), muscle tissue (1.4%), inflammatory connective tissue (14.0%), and a "residu" (mucous glands i.e.; 2.8%). Because of their high tumor content the biopsies were considered as representative to score tumoral expression of CA IX and GLUT-1.
Biological and clinicopathological parameters
The CT-determined perfusion rate ranged from 12.4 ml/min/100 g to 274.4 ml/min/100 g with a median value of 87.9 ml/min/100 g (n = 67). The percentage of CA IX positivity ranged from 0 to 87.8% with a median of 17.14%, and the GLUT-1 from 0 to 100% with a median of 67.5%. Both stainings were almost exclusively membranous, except for some aspecific cytoplasmatic CA IX staining.
There was no significant correlation between the CT-determined tumoral perfusion and percentage positivity for CA IX or GLUT-1.
We also found no significant correlation between tumor characteristics such as tumor stage, nodal stage and differentiation on one hand and perfusion-CT, CA IX and GLUT-1 percentages on the other hand.
Correlation with outcome
In univariate analysis, the prognostic value of different clinicopathological and biological parameters towards local control (LC), regional control (RC), disease-free survival (DFS) and overall survival (OS) was evaluated.
The tumor stage (T3-4 vs. T1-2) predicted a poorer RC (48.7 vs. 73.3% at 2 years; p = 0.03) and OS (43.7 vs. 64.6% at 5 years; p = 0.002). N3 stage also seemed to predict poorer outcome (LC: 9.8 vs. 44.8% at 2 years, p = 0.04; RC: 17.6 vs. 62.7% at 2 years, p = 0.0009; DFS: 0 vs. 37.9% at 5 years, p = 0.03; OS: 0 and 56.7% at 5 years, p = 0.002). No significant correlation was found between tumor differentiation and outcome.
Poorer tumor perfusion stratified according to the median value was linked with poorer LC (33.2 vs. 48.2 % at 2 years, p = 0.03) and DFS (26.4 vs. 38.6% at 5 years, p = 0.03).
For the CA IX and GLUT-1 percentages we could not find a significant correlation with outcome. Nevertheless, a trend towards poorer LC (31.4 vs. 44.6% at 2 years p = 0.08) and DFS (27.8 vs. 34.7% at 5 years, p = 0.08) was seen in patients with CA IX as well as GLUT-1 values above the median.
The multivariate analysis included tumor stage (T3-4 vs. T1-2), nodal stage (N0-1-2 vs. N3), differentiation grade, perfusion (≥ median) and combined CA IX and GLUT-1 percentages (CA IX and GLUT-1 ≥ median).
As shown in Table 2, the combined CA IX and GLUT-1 percentage was an independent prognostic factor for LC (p = 0.02) and DFS (p = 0.04) with a trend for RC (p = 0.06). N3 stage was a strong independent predictor of LC (p = 0.006), RC (p = 0.0001), DFS (p = 0.0001) and OS (p = 0.0002). Higher tumor stage predicted poorer overall survival (p = 0.006). The perfusion showed no clear correlation with LC (p = 0.09).
Group B: serum immunoassay
Biological and clinicopathological parameters
The CT-determined tumoral perfusion rate ranged from 29.6 ml/min/100 g to 263.6 ml/min/100 g with a median value of 88.7 ml/min/100 g (n = 34). IL-6 values ranged from 0.4 pg/ml to 57.3 pg/ml with a median of 5.4 pg/ml. VEGF values ranged from 73.5 pg/ml to 1613.1 pg/ml with a median of 326.9 pg/ml.
There was no significant correlation between the perfusion-CT measurements, IL-6 and VEGF values. However, Figure 5 shows that the perfusion was borderline inversely related to VEGF (p = 0.06) suggesting lower perfusion being associated with higher serum VEGF levels.
Patients with higher nodal stage showed significantly more elevated IL-6 serum levels (p = 0.04). Other tumor characteristics such as tumor stage and differentiation grade were not correlated with either perfusion-CT, serum VEGF or serum IL-6 levels.
Correlation with outcome
In univariate analysis, the prognostic value of IL-6, VEGF and clinicopathological parameters towards LC, RC, DFS and OS was evaluated.
As clear from Figures 6 and 7, the IL-6 serum level, stratified according to the median, was found to have a predictive value towards LC (9.3 vs. 70.4% at 2 years, p = 0.03), DFS (8.6 vs. 57.7% at 5 years, p = 0.05) and OS (11.9 vs. 74.9% at 5 years, p = 0.0009). Tumor stage, nodal stage and tumoral perfusion were found to have only a borderline prognostic value, probably due to the small sample size in this group (data not shown). We found no correlation between pretreatment cytokine serum levels and other patient-related parameters such as weight loss, performance status and sedimentation rate determined at the same pretreatment time.
The multivariate analysis included tumor stage (T3-4 vs. T1-2), nodal stage, perfusion (≥ median), VEGF serum level (≥ median) and IL-6 serum level (≥ median). Table 3 shows that the IL-6 serum level was the only independent predictor of LC (p = 0.009), DFS (p = 0.02) and OS (p = 0.005).
Discussion
The tumor microenvironment plays a critical role in tumor aggressiveness and treatment response. Several microregional features have been shown to be relevant for treatment outcome, such as hypoxia, inflammation and angiogenesis. The major question stays whether these markers can be used in an integrated way to stratify patients to certain treatment options [1].
Starting from a retrospective series of patients with HNSCC treated by radiotherapy ± chemotherapy, we investigated the prognostic significance of several biological and clinicopathological parameters. Earlier established results showed that for this whole patient group, CT-determined tumoral perfusion was correlated with poorer LC [2].
Two subgroups were now used for further analysis. In the first subgroup of 67 patients, immunohistochemistry for CA IX and GLUT-1 was performed on a representative paraffin-embedded pretreatment biopsy. In the second subgroup of 34 patients, ELISA was used to determine pretreatment VEGF and IL-6 serum levels. The correlation between these parameters mutually, as well as their prognostic value, was investigated.
Hypothesizing that CT-determined tumoral perfusion could be an inverse marker for hypoxia (less perfusion relates to more hypoxia), we assessed the correlation between tumoral perfusion, the intrinsic hypoxia markers CA IX and GLUT-1, and the serum levels of VEGF and IL-6, known to be upregulated under certain microenvironmental conditions like hypoxia.
Besides a borderline inverse association between perfusion and VEGF, no significant correlation was found in our study. There are three possible explanations in support of this finding.
First, it is likely that the dynamic CT measures rather acute hypoxia, while CA IX and GLUT-1, as well as VEGF and IL-6, are more indicative for chronic hypoxia.
Second, the CT-determined tumoral perfusion in this patient group was a measurement that took place through one region of interest, i.e. the maximum diameter of the tumoral region. On the other hand, CA IX and GLUT-1 immunohistochemistry was performed on one tumor biopsy taken at the periphery of the tumor field, and VEGF and IL-6 values were measured as systemic serum levels.
Third, the relationship between perfusion and oxygenation has shown to be in a precious balance [1,25].
The borderline association found between perfusion and VEGF can be explained through the hypoxia-hypothesis. Nevertheless, an opposite finding would also have been possible, associating the angiogenetic property directly with perfusion.
The further analysis of the biological and clinicopathological parameters showed that patients with higher nodal stage expressed significantly higher IL-6 serum levels. This experience is in accordance with previous reports linking elevated IL-6 levels to extensive disease [20], and more specifically, to lymph node involvement [26]. It is not surprising since this multifunctional cytokine plays an important role in promoting tumorigenesis.
Concerning the prognostic value of the clinicopathological parameters, we found a correlation with outcome for known factors like tumor stage (RC and OS) and N3 stage (LC, RC, DFS, OS). Of these, only the nodal stage appeared to behave as an independent prognostic factor for RC and DFS.
Concerning the prognostic value of the biological parameters, poorer tumor perfusion predicted poorer LC and DFS, as expected [2]. This correlation was lost in the smallest subgroup, probably due to a too small patient number in this group. Furthermore, in multivariate analysis, the perfusion only kept a borderline prognostic value for outcome.
For CA IX and GLUT-1, the combined assessment of patients with both markers expressed above the median did indicate an independent correlation with worse LC, RC and DFS. This result demonstrates that a combined assessment gives additive information, though we have to note that the prognostic value of each of both markers is not clearly established, given prognostic [5-8,27] as well as non-prognostic [28] publications.
In the second subgroup, IL-6 pretreatment serum level stratified according to the median seemed to be the only independent predictor of LC, DFS and OS. VEGF pretreatment levels were not correlated with prognosis.
The prognostic value of IL-6 has been described in many tumor types [18], but not yet in HNSCC treated by radiotherapy ± chemotherapy. In our multivariate analysis, we found a strong prognostic impact of the IL-6 pretreatment serum level on local control. This might indicate that there is indeed a link between IL-6 and radioresistance, as already shown in vitro by others [21-23].
VEGF- reports in literature are a bit conflicting, with papers demonstrating [15] and denying [29] a prognostic role for VEGF serum levels at pretreatment time.
In this study, due to methodological reasons, we performed serum and immunohistochemical analyses for two subgroups instead of the whole patient group.
The advantage of using serum samples over immunohistochemistry on biopsies is that the ELISA test is quickly and simple to perform. Furthermore, the technique is not too invasive and quite established.
On the other hand we must be aware that HNSCC cells may not be the only source of the elevated serum cytokine levels in patients. The serum levels may also depend in part upon individual host inflammatory responses and conditions like cachexia, and certainly not upon hypoxia only. In our patient group however, we did not see a correlation between VEGF or IL-6 serum levels and sedimentation rates, determined at the same pretreatment time. Furthermore, there are some pitfalls in the measurement of these cytokines [30]. Concerning VEGF, the isoform specificity and platelet interaction may be of influence. The use of banked frozen serum for analysis may also be perceived as a shortcoming. Because of the retrospective character of this study, we could not control the methods used for taking and storing the serum samples.
Conclusion
In conclusion, we found a prognostic value in this series of patients with HNSCC treated by radiotherapy ± chemotherapy for some of the investigated parameters, with a major role for IL-6. To our knowledge, we are the first to report a link in HNSCC between IL-6 pretreatment serum levels and radioresistance in vivo. This link is supported by the strong prognostic association of pretreatment IL-6 with local control, known to be the most important parameter to judge radiotherapy responses. This finding has to be validated in other, preferably prospective, studies.
Furthermore, the combined assessment of CA IX and GLUT-1 correlated independently with prognosis. This is a valuable indication that a combined approach is important in the investigation of prognostic markers.
We believe that additional to the standard TNM classification this combined assessment of several biological and clinicopathological parameters will show most appropriate to stratify patients to certain treatment protocols. The efforts made to search for prognostic factors in radiotherapy settings will stretch towards a more efficient combination of radiotherapy with other treatment strategies like chemotherapy, radiosensitizers, bioreductive drugs and biological targeting. Therefore, further research will be necessary.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HDS carried out the immunohistochemistry, ELISA's and the acquisition and analysis of data. She also drafted the manuscript. WL participated in the conception of the study and helped to draft the manuscript. EV assisted in the immunohistochemistry performance and scoring technique. LG participated in the scoring process, and the analysis of data. RH performed perfusion-CT and collected data. SN conceived of the study, participated in its design and coordination and helped to draft 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
- We thank Dr S Pastorekova and Dr J Pastorek (Institute of Virology of the Slovak Academy of Science of the Slovak Republic) and Dr J Zavada (Institute of Molecular Genetics, Academy of Science of the Czech Republic) for the use of M75 MAb for research purposes.
- Harlinde De Schutter is a research fellow of the "Fonds voor Wetenschappelijk Onderzoek – Vlaanderen" (FWO).
- This work was supported by a FWO grant (FWO 1.5.131.04N).
Figures and Tables
Figure 1 Example of immunohistochemistry for CA IX. Example of biopsy stained for CA IX (×100). This tumor was scored highly positive.
Figure 2 Example of immunohistochemistry for CA IX. Example of biopsy stained for CA IX (×100). This tumor showed a small positive percentage.
Figure 3 Example of immunohistochemistry for GLUT-1. Example of biopsy stained for GLUT-1 (×40). This tumor was scored highly positive.
Figure 4 Example of immunohistochemistry for GLUT-1. Example of biopsy stained for GLUT-1 (×40). This tumor showed a small positive percentage.
Figure 5 Correlation between CT-determined tumoral perfusion and pretreatment VEGF serum levels. There was a trend for poorer CT-determined tumoral perfusion to correlate with higher pretreatment VEGF serum levels (rS = -0.32, p = 0.06; n = 34).
Figure 6 Kaplan-Meier curve local control analysis. There was a significant difference in local control between patients with pretreatment IL-6 serum levels above and below the median (9.3 vs. 70.4% at 2 years, p = 0.03).
Figure 7 Kaplan-Meier curve overall survival analysis. There was a highly significant difference in overall survival between patients with pretreatment IL-6 serum levels above and below the median (11.9 vs. 74.9% at 5 years, p = 0.0009)
Table 1 Patient characteristics, tumor characteristics and treatment modalities
Group A (n = 67) Group B (n = 34)
Age (years) mean (min-max) 56 (36–82) 58 (42–79)
Sex M 60 31
F 7 3
cT stage T1 2 1
T2 17 13
T3 27 7
T4 21 13
cN stage N0 15 11
N1 12 3
N2 31 17
N3 9 3
Tumor site glottic 3 2
supraglottic 16 10
hypopharynx 20 7
oropharynx 24 14
oral cavity 2 0
nasopharynx 2 1
Radiotherapy 66 Gy/ 2 Gy 20 4
70 Gy/1.4 Gy 4 1
70 Gy/2 Gy 41 (2+ cisplatinum) 28 (1+cisplatinum)
80.5 Gy/1.15 Gy 2 (1+cisplatinum) 1 (+ cisplatinum)
Patient characteristics, tumor characteristics and treatment modalities are shown for Group A and Group B, both subgroups of the original study of CT-determined tumoral perfusion in patients with HNSCC treated by radiotherapy ± chemotherapy. Cisplatinum was given at a dose of 100 mg/m2 on days 1, 22 and 43.
cT: clinical tumor stage. cN: clinical nodal stage. M: male. F: female. Gy: Gray.
Table 2 Multivariate analysis for Group A
Local Control Regional Control Disease-free Survival Overall Survival
Parameter RR 95% CI P value RR 95% CI P value RR 95% CI P value RR 95% CI P value
T1-2 vs. T3-4 1.21 0.58–2.53 0.62 2.74 0.99–7.62 0.05 1.40 0.70–2.83 0.34 3.85 1.46–10.18 0.006
N 0-1-2 vs. N3 3.75 1.47–9.60 0.006 7.68 2.72–21.70 0.0001 5.70 2.34–13.85 0.0001 6.05 2.32–15.77 0.0002
Differentiation 1.50 0.90–2.49 0.12 1.64 0.89–3.02 0.11 1.49 0.93–2.40 0.12 0.81 0.45–1.47 0.49
Perfusion > median 0.55 0.28–1.09 0.09 1.07 0.46–2.51 0.87 0.61 0.32–1.15 0.12 0.65 0.48–0.88 0.28
GLUT-1 & CA IX > median 2.33 1.12–4.82 0.02 2.24 0.97–11.58 0.06 2.06 1.05–4.07 0.04 1.53 0.65–3.60 0.33
Multivariate analysis (Cox proportional hazard model) for Group A towards local control, regional control, disease-free survival and overall survival. Significant P values are marked in bold. RR: relative risk. CI: confidence interval.
Table 3 Univariate and multivariate analysis for pretreatment IL-6 serum levels (Group B)
Local control Regional control Disease-free survival Overall survival
RR 95% CI P value RR 95% CI P value RR 95% CI P value RR 95% CI P value
Univariate 0.98 1.39–9.78 0.03 1.29 0.37–4.53 0.68 3.24 1.25–8.50 0.05 8.30 2.44–28.50 0.0009
Multivariate 4.06 1.40–11.70 0.009 1.80 0.40–8.17 0.45 3.39 1.22–9.39 0.02 7.61 1.82–31.80 0.005
Univariate and multivariate analysis (Cox proportional hazard model) for IL-6 (> or < median) towards local control, regional control, disease-free survival and overall survival. Significant P values are marked in bold. RR: relative risk. CI: confidence interval.
==== Refs
Bussink J Kaanders JH van der Kogel AJ Tumor hypoxia at the micro-regional level: clinical relevance and predictive value of exogenous and endogenous hypoxic cell markers Radiother Oncol 2003 67 3 15 12758235 10.1016/S0167-8140(03)00011-2
Hermans R Meijerink M Van den BW Rijnders A Weltens C Lambin P Tumor perfusion rate determined noninvasively by dynamic computed tomography predicts outcome in head-and-neck cancer after radiotherapy Int J Radiat Oncol Biol Phys 2003 57 1351 1356 14630273 10.1016/S0360-3016(03)00764-8
Brizel DM Dodge RK Clough RW Dewhirst MW Oxygenation of head and neck cancer: changes during radiotherapy and impact on treatment outcome Radiother Oncol 1999 53 113 117 10665787 10.1016/S0167-8140(99)00102-4
Nordsmark M Overgaard J A confirmatory prognostic study on oxygenation status and loco-regional control in advanced head and neck squamous cell carcinoma treated by radiation therapy Radiother Oncol 2000 57 39 43 11033187 10.1016/S0167-8140(00)00223-1
Airley R Loncaster J Davidson S Bromley M Roberts S Patterson A Hunter R Stratford I West C Glucose transporter glut-1 expression correlates with tumor hypoxia and predicts metastasis-free survival in advanced carcinoma of the cervix Clin Cancer Res 2001 7 928 934 11309343
Koukourakis MI Giatromanolaki A Sivridis E Simopoulos K Pastorek J Wykoff CC Gatter KC Harris AL Hypoxia-regulated carbonic anhydrase-9 (CA9) relates to poor vascularization and resistance of squamous cell head and neck cancer to chemoradiotherapy Clin Cancer Res 2001 7 3399 3403 11705854
Kunkel M Reichert TE Benz P Lehr HA Jeong JH Wieand S Bartenstein P Wagner W Whiteside TL Overexpression of Glut-1 and increased glucose metabolism in tumors are associated with a poor prognosis in patients with oral squamous cell carcinoma Cancer 2003 97 1015 1024 12569601 10.1002/cncr.11159
Mineta H Miura K Takebayashi S Misawa K Araki K Misawa Y Ueda Y Prognostic value of glucose transporter 1 expression in patients with hypopharyngeal carcinoma Anticancer Res 2002 22 3489 3494 12552944
Ivanov S Liao SY Ivanova A Danilkovitch-Miagkova A Tarasova N Weirich G Merrill MJ Proescholdt MA Oldfield EH Lee J Zavada J Waheed A Sly W Lerman MI Stanbridge EJ Expression of hypoxia-inducible cell-surface transmembrane carbonic anhydrases in human cancer Am J Pathol 2001 158 905 919 11238039
Wykoff CC Beasley NJ Watson PH Turner KJ Pastorek J Sibtain A Wilson GD Turley H Talks KL Maxwell PH Pugh CW Ratcliffe PJ Harris AL Hypoxia-inducible expression of tumor-associated carbonic anhydrases Cancer Res 2000 60 7075 7083 11156414
Mellanen P Minn H Grenman R Harkonen P Expression of glucose transporters in head-and-neck tumors Int J Cancer 1994 56 622 629 8314336
Dunst J Stadler P Becker A Kuhnt T Lautenschlager C Molls M Haensgen G Tumor hypoxia and systemic levels of vascular endothelial growth factor (VEGF) in head and neck cancers Strahlenther Onkol 2001 177 469 473 11591020
Forsythe JA Jiang BH Iyer NV Agani F Leung SW Koos RD Semenza GL Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1 Mol Cell Biol 1996 16 4604 4613 8756616
Chen Z Malhotra PS Thomas GR Ondrey FG Duffey DC Smith CW Enamorado I Yeh NT Kroog GS Rudy S McCullagh L Mousa S Quezado M Herscher LL Van Waes C Expression of proinflammatory and proangiogenic cytokines in patients with head and neck cancer Clin Cancer Res 1999 5 1369 1379 10389921
Teknos TN Cox C Yoo S Chepeha DB Wolf GT Bradford CR Carey TE Fisher SG Elevated serum vascular endothelial growth factor and decreased survival in advanced laryngeal carcinoma Head Neck 2002 24 1004 1011 12410536 10.1002/hed.10163
St John MA Li Y Zhou X Denny P Ho CM Montemagno C Shi W Qi F Wu B Sinha U Jordan R Wolinsky L Park NH Liu H Abemayor E Wong DT Interleukin 6 and interleukin 8 as potential biomarkers for oral cavity and oropharyngeal squamous cell carcinoma Arch Otolaryngol Head Neck Surg 2004 130 929 935 15313862 10.1001/archotol.130.8.929
Muraoka K Shimizu K Sun X Zhang YK Tani T Hashimoto T Yagi M Miyazaki I Yamamoto K Hypoxia, but not reoxygenation, induces interleukin 6 gene expression through NF-kappa B activation Transplantation 1997 63 466 470 9039941 10.1097/00007890-199702150-00023
Srivani R Nagarajan B A prognostic insight on in vivo expression of interleukin-6 in uterine cervical cancer Int J Gynecol Cancer 2003 13 331 339 12801265 10.1046/j.1525-1438.2003.13197.x
Chung YC Chang YF Serum interleukin-6 levels reflect the disease status of colorectal cancer J Surg Oncol 2003 83 222 226 12884234 10.1002/jso.10269
Michalaki V Syrigos K Charles P Waxman J Serum levels of IL-6 and TNF-alpha correlate with clinicopathological features and patient survival in patients with prostate cancer Br J Cancer 2004 90 2312 2316 15150588
Brouazin-Jousseaume V Guitton N Legue F Chenal C GSH level and IL-6 production increased in Sertoli cells and astrocytes after gamma irradiation Anticancer Res 2002 22 257 262 12017299
Miyamoto Y Hosotani R Doi R Wada M Ida J Tsuji S Kawaguchi M Nakajima S Kobayashi H Masui T Imamura M Interleukin-6 inhibits radiation induced apoptosis in pancreatic cancer cells Anticancer Res 2001 21 2449 2456 11724306
Tamatani T Azuma M Ashida Y Motegi K Takashima R Harada K Kawaguchi S Sato M Enhanced radiosensitization and chemosensitization in NF-kappaB-suppressed human oral cancer cells via the inhibition of gamma-irradiation- and 5-FU-induced production of IL-6 and IL-8 Int J Cancer 2004 108 912 921 14712497 10.1002/ijc.11640
Pastorekova S Zavadova Z Kostal M Babusikova O Zavada J A novel quasi-viral agent, MaTu, is a two-component system Virology 1992 187 620 626 1312272 10.1016/0042-6822(92)90464-Z
Vaupel P Kelleher DK Hockel M Oxygen status of malignant tumors: pathogenesis of hypoxia and significance for tumor therapy Semin Oncol 2001 28 29 35 11395850 10.1053/sonc.2001.25398
Tartour E Deneux L Mosseri V Jaulerry C Brunin F Point D Validire P Dubray B Fridman WH Rodriguez J Soluble interleukin-2 receptor serum level as a predictor of locoregional control and survival for patients with head and neck carcinoma: results of a multivariate prospective study Cancer 1997 79 1401 1408 9083163 10.1002/(SICI)1097-0142(19970401)79:7<1401::AID-CNCR18>3.0.CO;2-3
Hui EP Chan AT Pezzella F Turley H To KF Poon TC Zee B Mo F Teo PM Huang DP Gatter KC Johnson PJ Harris AL Coexpression of hypoxia-inducible factors 1alpha and 2alpha, carbonic anhydrase IX, and vascular endothelial growth factor in nasopharyngeal carcinoma and relationship to survival Clin Cancer Res 2002 8 2595 2604 12171890
Kaanders JH Wijffels KI Marres HA Ljungkvist AS Pop LA van den Hoogen FJ de Wilde PC Bussink J Raleigh JA van der Kogel AJ Pimonidazole binding and tumor vascularity predict for treatment outcome in head and neck cancer Cancer Res 2002 62 7066 7074 12460928
Dietz A Rudat V Conradt C Weidauer H Ho A Moehler T Prognostic relevance of serum levels of the angiogenic peptide bFGF in advanced carcinoma of the head and neck treated by primary radiochemotherapy Head Neck 2000 22 666 673 11002321 10.1002/1097-0347(200010)22:7<666::AID-HED5>3.0.CO;2-9
Jelkmann W Pitfalls in the measurement of circulating vascular endothelial growth factor Clin Chem 2001 47 617 623 11274009
| 15847702 | PMC1097720 | CC BY | 2021-01-04 16:03:05 | no | BMC Cancer. 2005 Apr 25; 5:42 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-42 | oa_comm |
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-191583109410.1186/1471-2121-6-19Methodology ArticleAutomated measurement of cell motility and proliferation Bahnson Alfred [email protected] Charalambos [email protected] Douglas [email protected] Lei [email protected] Tongying [email protected] Donna [email protected] Hui [email protected] Hong [email protected] Julie [email protected] Tao [email protected] Raymond [email protected] Lex [email protected] Automated Cell, Inc. 390 William Pitt Way, Pittsburgh, PA, 15238 USA2 University of Pittsburgh Cancer Institute, Research Pavilion at The Hillman Cancer Center, 5117 Center Ave, Pittsburgh, PA, 15213-1863 USA2005 14 4 2005 6 19 19 30 6 2004 14 4 2005 Copyright © 2005 Bahnson et al; licensee BioMed Central Ltd.2005Bahnson 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
Time-lapse microscopic imaging provides a powerful approach for following changes in cell phenotype over time. Visible responses of whole cells can yield insight into functional changes that underlie physiological processes in health and disease. For example, features of cell motility accompany molecular changes that are central to the immune response, to carcinogenesis and metastasis, to wound healing and tissue regeneration, and to the myriad developmental processes that generate an organism. Previously reported image processing methods for motility analysis required custom viewing devices and manual interactions that may introduce bias, that slow throughput, and that constrain the scope of experiments in terms of the number of treatment variables, time period of observation, replication and statistical options. Here we describe a fully automated system in which images are acquired 24/7 from 384 well plates and are automatically processed to yield high-content motility and morphological data.
Results
We have applied this technology to study the effects of different extracellular matrix compounds on human osteoblast-like cell lines to explore functional changes that may underlie processes involved in bone formation and maintenance. We show dose-response and kinetic data for induction of increased motility by laminin and collagen type I without significant effects on growth rate. Differential motility response was evident within 4 hours of plating cells; long-term responses differed depending upon cell type and surface coating. Average velocities were increased approximately 0.1 um/min by ten-fold increases in laminin coating concentration in some cases. Comparison with manual tracking demonstrated the accuracy of the automated method and highlighted the comparative imprecision of human tracking for analysis of cell motility data. Quality statistics are reported that associate with stage noise, interference by non-cell objects, and uncertainty in the outlining and positioning of cells by automated image analysis. Exponential growth, as monitored by total cell area, did not linearly correlate with absolute cell number, but proved valuable for selection of reliable tracking data and for disclosing between-experiment variations in cell growth.
Conclusion
These results demonstrate the applicability of a system that uses fully automated image acquisition and analysis to study cell motility and growth. Cellular motility response is determined in an unbiased and comparatively high throughput manner. Abundant ancillary data provide opportunities for uniform filtering according to criteria that select for biological relevance and for providing insight into features of system performance. Data quality measures have been developed that can serve as a basis for the design and quality control of experiments that are facilitated by automation and the 384 well plate format. This system is applicable to large-scale studies such as drug screening and research into effects of complex combinations of factors and matrices on cell phenotype.
==== Body
Background
Cell-matrix interactions are key components of many physiological processes in health and disease. Frequently these interactions result in changes in cellular motility, morphology, and/or growth, and so quantitation of these changes is useful for comparing matrix and soluble factor effects and for assessing sensitivity of cells to varying concentrations of these factors [1,2]. A variety of methods are used to measure cell migration, including most commonly the transwell assay [3] often modified by fluorescence quantitation [4], and less commonly the under-agarose migration assay [5], the soft-agarose drop method [6], the phagokinetic track motility assay for phagocytic cell types [7], wound healing [8], and time-lapse video microscopy.
Although the transwell assay has been applied to random migration [9], video time-lapse microscopy provides advantages by yielding actual speeds of individual cells and additional features of motion, e.g. persistence [10]. The video time-lapse approach has been applied since the late 1930's using film-projected images and manual methods for tracking cell paths for determination of velocities [11]. The introduction of video imaging and computer-assisted methods of tracking have aided this approach [12,13]. However, even with computer-assisted methods, analysis of video time-lapse images can be labour intensive, particularly if the data have been gathered over extended time periods, and the opportunities for human fatigue and inadvertent selection using such methods may introduce bias. Moreover, the normal cell culture environment must be maintained during imaging. Although sophisticated studies are being conducted within special chambers [14] or under mineral oil [15], an ideal system would incorporate acquisition of images simultaneously from multiple wells under normal culture conditions maintained throughout the entire experiment.
A fully automated system for acquiring and analysing time-lapse images over extended time periods from multiple wells within 384 well plates has been developed [16-21]. The system includes an electronically controlled incubated cell culture environment for continuous monitoring over extended periods. Automated image analysis is used to determine cell morphological properties and cell location, and proprietary algorithms are used to construct cell paths or tracks through time, yielding magnitude and direction of motion. Cell proliferation is also monitored based upon processing of the same set of images. The data set provides a rich source of unbiased quantitative information about cell behaviour that can be accessed at the individual cell level and filtered in a manner similar to gating in flow cytometry. Here we apply this technology to examine responses of osteoblast-type cells to surface coatings of extracellular matrix compounds that may be involved in osteoblast differentiation and growth. This study is part of an ongoing effort seeking to uncover underlying factors that influence osteoblasts in the process of osteogenesis and in the dysregulation of this process leading to osteoporosis.
Results
Differential cell velocity on different ECM coatings
KM101 cells and MG-63 cells were chosen as models for this study because of their potential for differentiation into cells with osteoblast-like phenotypes. KM101 cells, from primary human bone marrow stroma [22], have been shown to secrete bone-type alkaline phosphatase upon differentiation [Julie Goff, personal observations]. The MG-63 cell line, originating from a human osteosarcoma [23], exhibits characteristics of bone forming cells including high levels of 1,25-(OH)2D3-responsive alkaline phosphatase activity and osteoblast-like regulated synthesis of osteocalcin and collagen type I [24]. Responses in cell motility may reflect functional changes that accompany osteoblast migration into areas of newly forming bone.
For KM101 cells and to a lesser extent for MG-63 cells, increased migratory activity on laminin coated surfaces in comparison with the other extracellular matrix compounds, collagen type I, collagen type IV, and mock-coated plastic, is evident when viewing time lapse video sequences [see Additional file 1]. Individual cell tracks were automatically constructed by the software algorithms (see Methods), and the tracks are seen to be longer on laminin I coated surfaces than on collagen type I, collagen type IV, or mock-coated plastic for KM101 cells over the same time period (Figure 1). Track length over equal time intervals correlates with velocity, and so these images demonstrate faster migration of KM101 cells on laminin-coated surfaces. It is not clearly evident from visual inspection whether less-pronounced effects are occurring in either cell type with the other surface coatings, and to what extent, if any, the effects are changing over time.
Figure 1 Example tracks from KM101 cells on different surface coatings. Single time-point images shown were acquired at 36 hours elapsed time from KM101 cells seeded into a 384 well plate pre-treated with 10 ug/ml solutions of (A) laminin, (B) collagen I, (C) collagen IV, (D) and mock coated. Tracks (green lines) have been superimposed on the images to indicate the paths of cells or clusters of cells over the previous set of 15 images (7.5 hours). Track continuity is broken in cases where two objects collide, and tracks are reinitiated upon object separation. Longer tracks reflect uninterrupted tracks and higher cell velocities. Within a track, each straight-line segment represents the distance between object centroid positions in successive images taken 30 minutes apart.
Data selection for single live-cell velocity
Automated multi-well image acquisition and analysis provides the potential to examine biological response for multiple cell types, compounds, and doses simultaneously and continuously over extended periods of time, but the data must be judiciously selected. From each image, multiple objects were segmented and tracked, and a set of filtering criteria was uniformly applied from the outset, based upon logical principles (see Methods). In addition, most view-fields included surface imperfections or adherent particles that were segmented as objects, some with areas within the range of single cells. Such objects are easily ignored and often go unnoticed when using manually assisted methods of analysis, but they present challenges for automated image processing. Stationary objects became particularly apparent when object positions were plotted over time in 3 dimensions (Figure 2). Tracks from stationary objects form nearly straight lines along the time dimension, whereas live cell tracks exhibit more arbitrary excursion in the x,y plane. It is also evident from this figure that stationary objects were often segmented and tracked throughout the extended time period of the experiment, and their cumulative contribution to the data was therefore more significant than was apparent from a 2D view of accumulated tracks such as that shown in Figure 1.
Figure 2 3D visualization of tracks over time. The 2 dimensional culture surface from within an imaged well is represented in the plane of the paper with the time dimension (labelled as "scan") represented by the third dimension coming up toward the reader. More distant time is perceived to be deeper beneath the surface of the paper using exaggerated perspective. Cell object positions appear at each scan where objects met the combined criteria of having an area less than 3000 sq microns and a track length of at least 3 segments. Position markers are colored according to fence size with blue indicating smaller, and red indicating larger fence size. Stationary objects are readily recognized by persistently straight "vertical" paths toward the reader through time, because their x, y coordinates changed very little. A total elapsed time of approximately 75 hours is represented from the "bottom" to the "top" of the 3D representation. Four different surface coatings are shown as indicated. This display was created using Spotfire visualization software (), an important tool for accessing, analysing and reporting on the multi-parametric data from the automated system.
For a given track, the maximum excursion in either the x or y direction is called the "fence size", and an additional filter based upon fence size was applied to exclude data from stationary objects. In order to establish fence size filter levels and to verify minimum loss of live-cell data, we devised a program to manually review and score track segments according to whether they originated from live cells or from stationary non-cell objects. For both cell types, fence sizes of 20 microns or less appeared to include the majority of non-cell objects without appreciable inclusion of live cells, and so filtering was conservatively applied at this level (Figure 3). The larger MG-63 cell area limit is associated with greater segmentation error that is in turn associated with larger fence sizes; thus the 20 micron filtering criterion was more of a compromise for slowly moving MG-63 cells than for KM101 cells. Having implemented these data filtering measures, we present overall data summaries from three experiments. It should be mentioned that this approach would not be valid in cases where significant numbers of live cells remained stationary on the culture surface.
Figure 3 Fence size and live-cell versus non-cell objects. Individual track segments were manually scored from image sequences according to whether they originated from live-cell or non-cell objects. A custom viewing device permitted selection of a maximum fence size for tracks to be included in the display; in addition, only segments from tracks connecting 4 or more points were displayed. With each increasing step in fence size, the entire image sequence consisting of more than 100 images was re-scored. This process was repeated for 4 representative wells for each cell type. The cumulative total counts for live-cell (blue) and non-cell (pink) objects are plotted versus fence size on the horizontal axis for both cell types, KM101 (left) and MG-63 (right). Cell debris appeared to be more abundant from MG-63 cells after plating, but the wider range of permissible area for MG-63 cells would be expected to include also more non-cell objects of larger area. Objects of larger area would be expected to have greater segmentation and centroid position variability (see Segmentation/Outline Error) resulting in greater associated fence size. For MG-63 cells, the counting of non-cell objects was discontinued at fence sizes of 40 and above because it became simply too demanding to distinguish and count both live and non-cell objects.
Kinetics of velocity response
Automated quantitation demonstrated that cells responded to the higher concentration laminin coatings with increased cell velocity over an elapsed time period up to three days, at which time in one of three experiments cells exceeded the 30% confluence cut-off for velocity measurement (Figure 4). A significant but less pronounced effect for collagen type I coating in comparison to mock-treated plastic and collagen type IV coating was also evident. Both cell types exhibited higher initial velocities on laminin and collagen type I surface coatings compared with plastic, even during the initial 4-hour time interval.
Figure 4 Kinetics of velocity response. Velocity measurements were averaged for KM101 cells (A) and MG-63 cells (B) at coating concentrations of 10 ug/ml (Left Panels) and 100 ug/ml (Right Panels). Each point represents the average for 8 images (acquired every 30 minutes over four-hour time intervals) from triplicate wells in each of 3 experiments. Surface coatings are indicated as follows: laminin (triangles), collagen I (open squares), collagen IV (closed squares), and mock-coated plastic (open circles). Linear trend lines were added, and statistical analysis was applied as described (Methods). For KM101 cells at dose level 10 ug/ml (upper left panel), the estimated velocity model is: Collagen I: v = 0.3327+0.0015*time; Collagen IV: v = 0.2641-0.00002*time; Laminin: v = 0.4532+0.0018*time; Plastic: v = 0.2938-0.0007*time. The p-values for the intercepts and slopes are all significantly positive (p < 0.0001) except that the slope for Collagen4 is nonsignificant (p = 0.9217). The overall test for the equivalence of the starting velocity is significant (p < 0.0001), which means not all the starting velocities are equal, and the Bonferroni adjusted pair wise tests show that all the initial velocities are significantly different. The overall test for the equivalence of the slope is significant (p < 0.0001), which means that not all the slopes are equal, and the Bonferroni adjusted pair wise tests show that only Collagen I and Laminin have equal slope (p = 0.5100). For KM101 cells at coating concentrations of 100 ug/ml (upper right panel), the p-values for the intercepts and slopes are all significant. The overall test for the equivalence of the initial velocity is significant (p < 0.0001), which means not all the initial velocities are equal, and the Bonferroni adjusted pairwise tests show that only Collagen I and Collagen IV have the same initial velocity (p = 0.0202). The overall test for the equivalence of the slope is significant (p < 0.0001), which means not all the slopes are equal, and the Bonferroni adjusted pairwise tests show that only Collagen IV and Plastic have the same slope (p = 0.8705). For MG63 cells at dose level 10 ug/ml (lower left panel), the p-values for the intercepts and slopes are all significant (all p < 0.0001). The overall test for the equivalence of the initial velocity is significant (p < 0.0001), which means not all the initial velocities are equal, and all the Bonferroni adjusted pair wise tests show significant difference. The overall test for the equivalence of the slope is significant (p < 0.0001), which means not all the slopes are equal. The Bonferroni adjusted pair wise tests show that Collagen I and Laminin have the same slopes (p = 0.5176), Collagen IV and Plastic have the same slope (p = 0.6623), and other pairs have significantly different slopes. For MG63 cells at 100 ug/ml (lower right panel), the p-values for the intercepts and slopes are all significantly positive (p < 0.0001). The overall test for the equivalence of the starting velocity is significant (p < 0.0001), which means not all the starting velocities are equal, and the Bonferroni adjusted pair wise tests show that only Collagen IV and Plastic have the same initial velocity (p = 0.0428).). The overall test for the equivalence of the slope is significant (p = 0.0084), which means that not all the slopes are equal, and the Bonferroni adjusted pair wise tests show that only Collagen IV and Plastic have significantly different slope (p = 0.0042), others are not significant.
A conspicuous feature of these results is that MG-63 cells accelerated throughout the time period regardless of surface treatment and even on mock treated plastic, whereas KM101 cells exhibited constant or slowly declining velocities on mock coated plastic and on collagen type IV coated surfaces. The continued acceleration of MG-63 cells may be verified visually in time-lapse images [see Additional file 2], but this behaviour was not perceptible to the unprepared eye prior to quantitative analysis. The kinetic data output from the automated system provides an obvious advantage in this regard for screening and discovery purposes over non-kinetic methods such as the transwell assay. Statistically significant effects were observed for collagen IV in comparison to plastic based upon the entire data set from three experiments over 80 hours elapsed time (see Figure 4 legend), even though measurement values frequently overlapped between these two groups. Once again, sufficient data are provided by the automated system to demonstrate moderate effects that are not likely to be significant using other methods.
Velocity dose-response
Dose-response curves indicate that KM101 cell velocity increased approximately 0.1 um/min with each 10 fold increase in laminin coating solution concentration across the range of concentrations tested, whereas for collagen type I, the cell velocity appeared to maximize at approximately 0.1 um/min above control levels with the 10 ug/ml coating solution (Figure 5). For MG-63 cells, more modest increases in velocity were observed. Average velocity appeared to maximize at levels approximately 0.15 um/min and 0.07 um/min above control levels with the 10 ug/ml coating solution concentrations for laminin and collagen type I, respectively. Average velocity levels were not significantly increased with collagen type IV coating solutions above those observed for mock-coated plastic for either cell type.
Figure 5 Velocity dose-response. Velocity measurements for single cells were averaged over the time period from 24 to 48 hours and were normalized based upon the average velocity of cells on mock-coated plastic in each of three experiments (see Methods). Velocities are plotted versus the coating concentration of the extracellular matrices laminin (triangles), collagen I (open squares), collagen IV (closed squares), and mock-coated plastic (open circles) for KM101 cells (A) and MG-63 cells (B). Error bars indicate the standard error for three experiments. Identical mock-coated plastic controls were included for each concentration; in this case, since the concentration values for plastic are bogus, these results illustrate random variation for the replicate means.
Velocity measurement heterogeneity
There are many possible ways to visualize and to quantitatively summarize the data. For the kinetic and dose-response analyses above, a velocity measurement was included for each object at every time point when the set of criteria were fulfilled, i.e. the object area was in the specified range for each cell type, the track connected across at least 4 points, the track fence size was greater than 20 microns, and the total area occupied by cells was less than 30% of the total image area. The population heterogeneity of these measurements under different treatments over a 24-hour time period is illustrated in Figure 6. The distributions are positively skewed and appear unimodal for KM101 cells but exhibit evidence for bimodality for MG-63 cells, suggesting the possible continued presence of a subpopulation of stationary non-cell objects in the measurements. Under the higher area-filtering limit for MG-63 cell objects, larger fences would be expected for stationary non-cell objects; but attempts to increase the fence-size filter resulted in significant loss of records from cell-like objects without eliminating the bimodality (not shown). Velocity measurements up to 2.1 um/min were recorded. Each measurement represents an average over at least three segments (1.5 hours) up to as long as 10 segments (5 hours); the relative contribution from tracks of each size for each bar in the histograms is indicated by colours (see Figure 6 legend). It is evident that a significant proportion of tracks were longer than 10 segments and that shorter tracks were more numerous in the velocity range associated with live-cells (≥ 0.2 um/min) as opposed to the range associated with non-cell stationary objects (0.1 um/min; see Segmentation/Outline Error below). At similar cell densities, more rapidly migrating cells are more likely to terminate tracks by colliding or leaving the viewfield [see Additional file 3].
Figure 6 Velocity distributions. Histograms show changes in the velocity distribution profiles in response to surface coatings for the different extracellular matrices laminin, collagen IV and collagen I (all at 10 ug/ml coating concentration) in comparison to mock-coated plastic for KM101 cells (A) and MG-63 cells (B). Velocity measurements on individual objects were filtered according to criteria for single cells as described (see Methods) and were compiled over an elapsed time from 24 to 48 hours from three experiments. The horizontal axes indicate velocity measurement bins from which the number of objects included within each bin are determined as shown on the vertical axis (faint horizontal lines represent 1000 objects). The different colours within each bar indicate cell track length (from bottom to top: grey indicates tracks of 11 or more segments, violet: 10 segments, light-blue: 9 segments, bright-blue: 8, green: 7, black: 6, yellow: 5, dark-blue: 4, and red: 3 segments).
The individual cell velocity distribution modes were noticeably shifted toward higher velocities with laminin and to a lesser extent with collagen type I surface coatings, in comparison with mock-coated plastic and collagen type IV. These histograms provide a more descriptive picture of the nature of the response of the cell populations to the surface treatments than that provided by measures of central tendency alone, without invoking complex statistical models.
To provide sufficient numbers of objects for reliable averaging, triplicate wells were coated and seeded for each treatment/dose/cell type combination. Single cell counts, i.e. the average number of objects identified in each image that met the area-range criteria for single cells, and for which velocity measurements were included in the kinetic, dose-response, and velocity distribution histograms, are shown in Figure 7. There was a tendency for more motile cells to break free of clusters, i.e. to "scatter" [see Additional file 1], and this tendency gave rise to and is reflected in the greater number of single cells on laminin coated surfaces. For KM101 cells, single cell numbers eventually declined as crowding from cell growth contributed to a greater proportion of cells in clusters. For MG-63 cells, slightly increasing numbers of single cells were observed despite growth toward increasing levels of confluence; this increase correlates with, and possibly resulted from, accelerating motility [see Figure 4 and Additional file 2].
Figure 7 Single cell numbers. The average number of objects that met the criteria for measurement of velocity (see Methods) over 4-hour time intervals for KM101 cells (A) and MG-63 cells (B) at extra-cellular matrix coating concentrations of 10 ug/ml (Left Panels) and 100 ug/ml (Right Panels). Each point represents the average for 8 images from triplicate wells in each of 3 experiments. These numbers represent the typical sample size, n, from the populations of measurements shown in Figure 6, from which a mean velocity is calculated for each well at each time point. The observed dispersion of these mean velocities is shown in Figures 9 and 10.
Alternative approaches to data analysis can be used to examine the heterogeneity of velocity measurements at the individual cell level over time, and to ask such questions as: Do subsets of fast and slow cells exist, or do all cells show a wide range of speeds over time? How long can a rapidly moving cell maintain rapid motion? Is there evidence that individual cells require resting periods? How does cell-cycle phase affect velocity? An example of velocity histories at the individual cell level is shown in Figure 8. More specific analysis at this level may form the basis of a subsequent report; findings of the current study afford an ability to assign significance levels to the instantaneous peaks and troughs of such non-smoothed data (see below).
Figure 8 Individual cell velocities over time. Instantaneous velocity for individual KM101 cells is plotted for each time point (30 minute intervals) across time periods that extend in each case from the initiation to the termination of individual cell tracks. Cell tracks from four wells with different surface coatings (Collagen I – blue, Collagen IV – red, Laminin – yellow, Plastic – green) are shown; coating concentrations were 100 ug/ml in each case. The number at the top of each frame indicates the "unique track index" by which the tracks are identified. Not all surface coatings are represented in each frame, because in many cases tracks that were initiated did not meet criteria for track length, fence size, or object area (see Methods). These plots show the magnitude of individual non-smoothed velocity vectors, indicating cell motion plus random contribution from stage noise and cell-outline error at each 30-minute time point. In many cases the same cell may terminate in one frame and reinitiate in another due to collision.
Technical precision and quality control
The performance capabilities of the automated system, including inter-well variation that is associated with plate manufacture, pipeting, and other operations, may be examined in terms of the technical precision of measurements between replicate wells within experiments (Figure 9). The abundance of data that can be automatically gathered without constraint over time (in contrast to manual analysis by graduate students), leads to robust precision estimates, shown here to be proportional to velocity measurement values.
Figure 9 Precision between replicates. Precision (right-hand y axis scale) of the mean velocity from all cells in an image at a single time point is represented by the square root of the average variance from triplicates ranked according to mean velocity for either KM101 cells (square symbols – solid line) or MG-63 cells (round symbols – dashed line)(See Methods section). The bars in the background represent the number of triplicate sets that were included in each velocity bin (x axis) for determining average variance for KM101 cells (black bars) and MG-63 cells (white bars). The reliability of the precision estimates is reinforced by the smooth correlation of precision estimate with mean velocity across much of the range. Note how precision estimates became more "unstable" as the number of triplicate sets (bars in background) decreased at higher mean velocities.
Since the velocity data are not normally distributed (bounded on the left and skewed toward the right), we examined histograms for standard deviation from triplicate sets as an approach to graphical estimation of quality control limits (Figure 10). For MG-63 cells, all replicate data appeared to be "well contained" by an upper bound of 0.26 um/min, with only two outliers (at ≥ 0.4 um/min) among almost 8,000 triplicate sets. For KM101 cells, 99.7% of the data were contained within an upper limit of 0.32 um/min, leaving 33 questionable sets out of almost 10,000 triplicate sets. Manual examination of some of these sets revealed the presence of surface blemishes in close proximity to each other that gave rise to oscillating tracks between them in images from one well of the set. The regularity in these false tracks may provide a mechanism for filtering them out of the larger data set in future experiments. These data establish a benchmark for monitoring quality in future studies and for inter-experiment and inter-lab comparisons.
Figure 10 Graphical quality control limits. The full population of triplicate sets shown in Figure 9 is here represented as a histogram of the individual standard deviations for velocities from triplicate wells of KM101 cells (black bars) and MG-63 cells (white bars). Presented in this fashion, the data are useful for graphically estimating quality control limits and identifying outliers. Data from images of MG-63 cells are shown to be "well behaved" in comparison to KM101 cells, which exhibit a number of outliers to the right of the "normal" limit expected for the tail. Arrows indicate positions of "3 sigma" limits that are based on the observed percentage (99.7%) of observations (rather than on a more complex transformation of the data according to a mathematical model). In the case of MG-63 cells, the limit (block arrow) excludes several adjacent triplicate sets that are properly part of the distribution. Two certified MG-63 outliers are not readily visible on this chart at 0.40–0.41 um/min. In the case of KM101 cells, the "3-sigma" limit (solid arrow) more haphazardly cuts off a chain of outliers. As described in the text, one source of KM101 outliers was identified by manual viewing of tracked objects in the image sequences, and may lead to a rational filtering method that would improve the data on both sides of the control limit for KM101 cells. Since the average variance for triplicate velocity measurements correlated to some extent with average velocity (see Figure 9), a more refined approach to quality control would involve first ranking the data according to average velocity. Presumably the next step would be to determine whether a single member of the outlying triplicate sets should be discarded in favour of preserving the remaining two measurements in each case.
Segmentation/outline error
A surrogate for the empirical evaluation of the total variability associated with both stage noise (i.e. imprecision in returning to the same view-field between acquiring sequential images) and object segmentation variability (i.e. imprecision in "drawing" cell outlines) would ideally consist of inert cell-like objects that remained stationary on the surface. In Figure 3 above, we showed that tracks associated with fence sizes below 20 microns arose predominantly from stationary non-cell objects of similar area to single cells, and here we consider using these objects, excluded from the live cell analysis, as logical surrogates for estimating system variability. The average velocities of such objects at each fence size were nearly identical for all three experiments, demonstrating system stability with respect to stage noise and image segmentation parameters (Figure 11).
Figure 11 Velocity at small fence sizes. Average velocity (lines with symbols), individual maximum velocity (*), and minimum velocity (×) measurements are shown for objects associated with tracks of small fence sizes from three experiments (see legend). Average velocity was calculated as the mean of all velocity measurements from objects meeting cell specific criteria for area with tracks consisting of 3 or more segments, pooled according to fence size and experiment. Consistency between experiments reflects consistency in the rate at which stage noise and object segmentation variability are exhibited. By limiting fence size, we also limit the maximum possible displacement, and we would expect this to be reflected in the maximum observable velocity. It is. Consider tracks with a fence size of 20 microns (10 pixels); this is near the maximum observable displacement for these objects between any two time-points*. Since the images were taken at 30 minute intervals, the corresponding maximum permissible velocity is 20/30 um/min, or 0.67 um/min, which corresponds to data from all three experiments at a fence size of 20 um. These data, in addition to manually constructed images (see Methods), confirm soundness within the system for accurate velocity measurement. (*Fence size is not strictly the maximum observable displacement. A slightly more sophisticated definition of fence size would use the diagonal from maximum x and y values, which would result in slightly greater fence sizes for objects displaced in random directions.)
A clearer picture of the proportional contribution to the average velocities from non-cell and live-cell objects is shown by the velocity distributions for fence sizes below 20 microns versus those above 20 microns, respectively (Figure 12). The system velocity output at the individual object level for stationary non-cell objects is seen to cluster about a mode of 0.1 um/min, and approximately 90% of the measurements fall between 0 and 0.23 um/min. Taken in conjunction with Figure 4, these data support an estimate of approximately ± 0.1 um/min for the uncertainty in velocity measurements due to stage noise in combination with object segmentation variability.
Figure 12 Velocity histograms at small fence sizes. The same data that gave rise to the averages and ranges in Figure 11 are represented in histograms here after having been grouped according to whether each velocity measurement was associated with a track of fence size smaller than 20 microns (red bars) or fence size from 20 to 50 microns (blue bars). Measurements were also grouped according to whether the well contained KM101 cells (A) or MG-63 cells (B). A total of over 600,000 measurements arising from three experiments are included. These data should be interpreted in conjunction with evidence presented in Figure 3 for the proportions of non-cell and live-cell objects associated with the group distributions shown here. Note that the MG-63 distribution for the larger fence size population is bimodal, further supporting the dichotomy of object type.
Comparison of automated tracking versus manual tracking
Comparison of the results from automated tracking with manual tracking further validated the accuracy of the system, the efficacy of the fence size filtering method, and the individual fidelity of cell tracks from example image sets. Velocity histograms prepared from all data obtained by both methods show that similar treatment effects were obtained (Figure 13). Moreover, if we assume that the velocity distributions acquired from manual tracking are more truly "accurate" because only known live-cells were tracked, then this comparison further supports the validity of the "fence-size filter" method to discriminate non-cell stationary objects. This is because removal of the red, small-fence-size data from the histogram bars for automated tracks (Panel B) brings them more in line with the manual track histograms (Panel A). Thus, this comparison shows that we are not excluding appreciable numbers of stationary live cells by applying the fence-size filter, and that the automated data for velocity is very similar in quality to that obtained manually.
Figure 13 Velocity histograms by manual and automated tracking. Velocity measurements were acquired through manual tracking (Panel A) or through automated tracking (Panel B) from image sets from four wells as described in the text. Different coating compounds are indicated along the y-axis (all 10 ug/ml). Red portions of the bars for automated tracking indicate measurements that came from tracks with fence sizes of less than 11 pixels (i.e. 20 um or less motion in the x or y direction over the length of the track). These tracks (red portions) were selected for exclusion based upon the supposition, supported by Figure 3, that such tracks arose predominantly from non-cell stationary objects. Note in particular how subtraction of the red portions from the 0.0, 0.1, and 0.2 um/min bins for Collagen IV and Collagen I from the automated track histogram (panels B) would bring the remaining data (blue portions) more in line with the profile of data from manual tracks (Panel A).
To more specifically compare manual versus automated tracking on a point-by-point basis, we developed an algorithm to match object positions from automated analysis with manually tracked cell locations from each image. A necessary component of this comparison is determining the precision of each method, and so both methods were performed in replicate on the identical image sets. Standard statistical methods of comparison were applied to the matched velocity data. Regression analysis for a perfectly precise system would yield slope equal to one and intercept equal to zero, producing a 45° line of identity in an x-y plot. Such a result was obtained for replicate analyses by automated tracking, i.e. almost 20,000 individual cell locations were identically analysed when the original "jpeg" images were reprocessed. When comparing the test system versus a reference system, an observed slope less than one and intercept greater than one for a regression line would imply that the dependent variable is "accurate" in a mid-region, but is somewhat overestimated at low values and underestimated at high values. These features were evident from regression analysis of automated versus manual tracking, but also for the second replicate versus the first replicate manual tracking exercise, pointing out limitations in the precision of manual tracking as a reference method (Figure 14).
Figure 14 Correlation between manual and automated velocity measurement. Velocity from the automated method (Panel A) or from the second manual replicate trial (Panel B) were plotted in reference to corresponding matched velocities from the first manual tracking exercise (y axis = dependent variable, x axis = independent variable). Increasing frequency of points is indicated by density contour lines and red shaded intensity (color scale). The dashed line indicates the line of identity and the solid line indicates the regression line for each comparison. Regression statistics were as follows: For automated tracking (y) versus the first manual tracking (x) (Panel A), the slope and intercept were 0.79 ± 0.02 and 0.03 ± 0.01, respectively with Pearson's correlation, r, equal to 0.81 ± 0.01. For the second (y) versus the first (x) manual tracking (Panel B), the slope and intercept were 0.84 ± 0.02 and 0.07 ± 0.01, respectively with Pearson's correlation equal to 0.86 ± 0.01.
To better understand the magnitude of these errors in practice, we examined the actual differences in locating cell positions by the automated and manual tracking methods. Our data set consisted of 716 cell positions that were co-located by automated tracking and by the first and second manual exercises according to the matching algorithm (see Methods). No single position can be considered the "true" position; however, the differences between replicates can be used to estimate precision, S, according to the relationship, S = 1/v2 times the standard deviation of the differences, where the differences retain the sign; for example, if the second comparison measurement is larger than the first, the sign is negative. By this approach, for the first and second manual exercises, the precision was 1.91 and 2.03 pixels in the x and y coordinates, respectively. For the comparison between automated tracking and manual tracking the precision ranged between 2.04 to 2.28, but when automated tracking was compared with the mean from the two manual tracking exercises, the precision improved to 1.87 and 1.91 pixels in the x and y coordinates, respectively. Thus, the automatically determined cell positions were closer to the mean positions from the two manual determinations than any individually determined position was to any other. Under the settings used, a distance of 2 pixels corresponded to an error of less than 1/32 inch, easily within the realm of human error for mouse-clicking on a monitor.
Finally, the time-resolved fidelity of individual tracks may be examined in detail in 3D scatter plots of the x and y locations over time (Figure 15). Here we show all cell positions for which the first manual tracking matched automated tracking and/or the second manual tracking exercise. Replicate manually tracked cell positions are represented along the full length of most tracks. Automatically tracked cell positions show more frequent gaps resulting from merging and branching, but they are nevertheless represented along the majority of the positions for most tracks. Altogether, these data provide ample support for the accuracy and fidelity of the automated tracking method.
Figure 15 3D visualization of matched cell tracks. Cell positions from the first manual tracking exercise (vertical blue markers) that matched either the second manual tracking exercise (horizontal yellow markers) or the automated tracking (red globes) were plotted in 3D with x and y coordinates in the plane of the paper and with increasing elapsed time (scan = 30 minutes) represented by the z axis coming "up" toward the viewer (similar to Figure 2). Lines connect markers by unique Track ID numbers, so that detailed features of track continuity may be examined. For clarity, only cell positions from one example culture well are shown (mock-coated plastic).
Cell proliferation
As cell numbers increased over time, the total area of the segmented objects naturally increased, and initially we used this sum divided by the average area of separate individual cells for estimation of total cell numbers per view-field. Such numbers are seen to increase exponentially, and the slope of log-transformed cell number can be used to estimate doubling times for comparison of growth rates (Figure 16). The linearity of the log scale plots and the high density of data points permit estimation of area-doubling times, over time periods as short as one to two days during the "clean" exponential growth phase, which persisted for 4 to 6 days. However, quantitative analysis of cell growth rates by the area-based method of estimating cell numbers was set-aside when we compared manually scored cell numbers with area-based estimates (Figure 17). The area-based method increasingly underestimated actual cell numbers with increasing levels of cell confluence.
Figure 16 Growth curves. Area-based cell numbers (see Methods) were plotted versus time (hours) over the course of the three experiments performed for this study. Data from wells with extra-cellular matrix compound coating concentrations of 10 ug/ml (left panels) and 100 ug/ml (right panels) are shown for both cell types, KM101 cells (first and third rows) and MG-63 cells (second and fourth rows) using a normal scale (top two rows) and log scale (bottom two rows) for area-based cell numbers along the vertical axis. A datum point is plotted for each image, i.e. at every 30 minute interval. Each point represents the average from a set of triplicate wells. During the third experiment (far right hand set of curves in each chart), disengagement of the coupling for the mechanical focus was discovered and repaired; the data from that period of malfunction were removed, as evident in the 27 hour gap before the end of the third experiment in each chart. Area-doubling rates obtained from the average of the four treatments for each of the three experiments are printed above the log-scale charts. These area-doubling rates (doublings/day) are the slopes obtained from linear regression of log base 2 cell numbers versus time across time periods selected from each experiment during which the log scale curves were linear – at least 48 hours in each case. Area-doubling rates for the first experiment were significantly greater than the doubling rates for the second and third experiments (t-Test, p < 0.01).
Figure 17 Ratio of area-based cell counts to manual cell counts. Area-based cell counts were automatically estimated by dividing the total area of segmented objects in each image by the average area of a single cell for each cell type, KM101 (red squares) and MG-63 cells (green circles). Manual counts were scored using a custom program that marks cells when clicked on, and increments a total count. A total of 90 images were analysed for each cell type across a range of confluence levels. The individual counts were grouped and averaged according to manual count levels; maximum values corresponded to 100% confluence for each cell type. Trend lines were based upon an exponential fit.
Regardless of the non-linearity of area-based cell numbers, the monitoring of cell growth based upon total cell area provided extremely valuable information. Within experiments, the log-based slopes were similar for the different treatments, with the pronounced exception of laminin and KM101 cells in the second experiment, even when starting with different initial seeding densities and when absolute cell numbers differ considerably over time as was evident in the three experiments performed for this study. A faster doubling rate for all treatments in the first experiment contrasted with slower doubling in the second and third experiments for both cell types. Treatment-related patterns occurred within experiments that were not evident between experiments, and even without quantitative analysis it is clear that no reliable pattern of surface-treatment effect on cell growth can be found. Hazards are evident, however, if conclusions were to be drawn from a more limited number of experiments, even with convincing within-experiment data.
One can speculate on possible causes for the between and within-experiment patterns shown here. For the second experiment, seeding densities were unusually low, and an atypical initial period of cell death was evident in the time-lapse videos during the first day after plating (data not shown) and is shown also by the unusual "lag phase" in the log-based charts for both KM101 and MG-63 cells in the second experiment. Typically, healthy cells under normal conditions begin exponential growth immediately upon re-plating, as evident in the linear log-based graphs for the first and third experiments from the point of initiation up to the point of confluence. The departure evident for the second experiment in the growth charts holds implications for variability between experiments in the velocity data presented above (Figure 5). Standardization of procedures and testing for robustness are called for. The value of our system for assisting in such efforts is prominently evident in these charts, apart from the question of treatment-related effects on growth in this study. In future studies, consistency in growth parameters could be made a prerequisite for acceptability of results for other parameters; here the results serve to demonstrate system potential.
Discussion
This study demonstrates the applicability of a fully automated image acquisition and analysis system for quantitative measurement of cell motility and for monitoring cell proliferation in a relatively high-throughput manner. Simultaneous comparison of multiple treatment variables in a 364 well plate format over extended periods of time is possible. Here we show how various surface coatings of extracellular matrix compounds can induce increased velocity in two types of osteogenic cell lines, how these responses vary over time periods of more than three days, and how they vary by surface coating concentration, i.e. dose-response. We show that laminin and to a lesser extent collagen type I surface treatments brought about increased migratory activity, while collagen type IV did not, in comparison to mock-coated plastic for KM101 cells and MG-63 cells. We show further how the extensive multiparametric output can be used to selectively isolate data of interest for improvement of bio-informational content and for characterization of system performance.
We believe our system is the first to accomplish full automation of time-lapse motion analysis of cells in culture, broadening the scope of application well beyond the practical limits imposed by manually interactive methods. Image sets consisting of hundreds to thousands of scans from up to 384 locations over extensive time periods (days to weeks) can be acquired and batch processed, first to segment cell-like objects and clusters of such objects, and subsequently to construct tracks for the segmented objects over time. Quantitative data is seamlessly exported to a Sequel Server Data Base from which results are presented in various forms as shown here. The relational data base is absolutely necessary for handling the large volume of multi-parametric data – over three million records for this study, including one for each object at each time point containing more than 40 fields of positional, textural, morphological, and motility-related data. Additional fluorescence measurements are also commonly used [19-21].
Our analysis contrasts in some respects with the exquisite optical characteristics and advanced analytical details of migratory activity achieved in systems where parallel plane surfaces are provided by specially constructed chambers [e.g. [25,26]]. Such non-conventional culture environments overcome the meniscus problem, but are not readily adaptable to simultaneous analysis of multiple treatment conditions in a high throughput manner. Most importantly for screening and discovery purposes, our system preserves the essential features of trajectory analysis that were emphasized by Friedl and co-workers for evaluating subtle changes that reveal treatment effects on cell locomotion [27,28].
To support the introduction of fully automated cell tracking, we have included a thorough analysis of measurement variability, system errors, and accuracy. Biological differences between individual cells dominated the velocity measurement heterogeneity for the cells chosen for this study, while stage noise, cell segmentation, and cell positioning were shown to be relatively minor sources of error. For more slowly moving cells, i.e. those with velocities less than 0.4 um/min, the latter relatively minor errors will result in large coefficients of variation under current conditions. Rational data-filtering methods, such as exclusion of data from non-cell stationary objects based upon "fence size" and exclusion of data from objects with areas outside of single-cell ranges (see Methods), yielded further improvements to the detectability of biological response from single cells. The accuracy and validity of these approaches were supported by comparison of the automated data with data from manually tracked cells. A significant feature of this analysis was the finding that, in general, automatically determined cell positions were closer to the mean positions from manual tracking than the replicate manual positions were to each other. Thus a major limitation of this comparison was the imprecision of manual tracking.
On the other hand, the perfect precision of automated tracking represents a two-sided coin. It will precisely repeat any errors, posing an obstacle to data control based upon monitoring internal consistency. At the same time, small variations in processing variables can be readily applied, tested, and evaluated. For example, in current experiments, we are exploring the possibility of obtaining better "average" cell positions by acquiring replicate non-identical images closely spaced in time, e.g. one minute apart, for characterizing motion over longer intervals, typically 30-minutes. Such an approach should improve precision and segmentation reliability at the individual cell level, while simultaneously providing a foundation for routine control based on inherent data consistency. The ability to perform automated batch processing will in due course be utilized more broadly for quality control and for optimization of image processing variables. Advantage can be leveraged from perfect precision by applying and evaluating systematic changes to variables of a contrast enhancement algorithm, for example, followed by repeated cell segmentation and cell track processing of representative image sets, such that cell track lengths, in terms of linked segments, would be maximized for non-collision-related scenarios. Automated routines for such optimization are not yet implemented, but are entirely feasible.
In seeking higher levels of automation, however, it is important to continue to manually monitor for and evaluate consistency between the raw data input and the processed data output. For each cell type, and for varying treatment conditions, processing variables are checked and evaluated. With each experiment, we manually verify that cell segmentation and cell track formation proceed reliably throughout representative regions of the image sets. The processing software presents the operator with cell outlines and tracks as shown in Figure 1 [see also Additional file 3], and Spotfire™ visualization software is used extensively to verify consistency in the quantitative data. Tracking patterns dominated by short fragments indicate, for example, problems requiring adjustment in image processing variables. Representative image sets are first examined for such problems before initiating batch processing of the full experiment; but it is not impractical to automatically reprocess image sets from an entire experiment when evidence demonstrates a need for it. It should be pointed out that most of the refinements in data filtering, thresholding, and so on for this study were conducted at the level of data base querying, not at the level of image processing or track construction. Whereas image processing requires hours to complete for a full experiment image set, data base queries can be run and re-run in a matter of minutes, and so the effect of changing thresholds or "gates" may be readily examined and interpreted using statistics or visualization software such as Spotfire™. Part of the beauty of automated analysis lies in our ability to make incremental improvements in both image processing and data querying approaches over time, leading to increasingly more accurate and reliable data. In contrast, manual analysis will remain decidedly imprecise and variable over time.
Quantitative results from other published studies further support the general accuracy of our system. For example, Huttenlocher et al. reported speeds ranging from 0.2 to 1.5 um/min for individual CHO cells plated on fibrinogen coated surfaces and later reported average speeds of approximately 0.3 um/min for myoblasts on fibronectin using manually-assisted computer based methods [9]. Similarly, Ware et al. reported maximal average velocities of a 3T3-derived cell line of approximately 1.5 um/min and presented histograms for individual cell velocities ranging from near-zero to approximately 2.5 um/min on a mixed extracellular matrix derivative in the presence and absence of epidermal growth factor [10]. It has been shown that experimental conditions such as pH and temperature, among others, can have a profound influence on motility [29], nevertheless, reported velocities for adherent cells in other published studies are consistent with our findings.
Determining cell proliferation using area-based measurement was shown to be vulnerable to non-linearity between area and cell numbers. But area-based growth curves exhibit exponential characteristics of high quality, and so automated quantitative monitoring of total cell area can be highly informative and useful. Image processing algorithms for segmentation of individual cells within colonies or clusters using transmitted light will yield more reliable data as improvement continues with auto-focusing and contrast optimisation. The use of nuclear fluorescence staining is a reliable alternative for automated cell counting, but known stains are toxic and therefore were not used in this study. Application of a mathematical adjustment factor based upon 'calibration' with manual or fluorescent counts might be considered; the risk here is that treatment effects could be masked, so this approach has not been used.
Velocity enhancement of osteogenic cells with extracellular matrix compounds provided a useful illustration for this pilot investigation of the automated system. It should be pointed out that collagen extra-cellular matrices consist of a 3D network of polymerised fibers, and that the 2D non-polymerised surface coatings used here may not reveal representative cellular response to the normal polymerised form of the extracellular matrix. An unexpected finding from this study was the continued acceleration of MG-63 cells on mock-coated plastic as well as on the extracellular matrix compound coated surfaces in contrast to KM101 cells that did not accelerate on mock-coated plastic or collagen type IV coated surfaces. Trypsinization cleaves cell adhesion receptors at the time of replating, and it is possible that the re-synthesis and transport of such receptors played a role in the kinetics of the responses shown in this study. The effects of such recovery would not seem to explain the continued acceleration of MG-63 cells through several cell cycles, however. Migration inducing-factors secreted by osteoblasts have been reported [6,30], and it is possible that an effect from such a factor is involved here. Alternatively, MG-63 cells may secrete extra-cellular matrix proteins that gradually accumulate on the surface to enhance migration rates.
Conclusion
This study demonstrates performance of a system for fully automated, time-resolved, and high throughput analysis of cell migration and proliferation, applied to continuously growing cultures in standard 384-well plates. This format permits the study and comparison of cellular responses to a large number of different treatment conditions, as shown here for three different extracellular matrix coatings on cultureware plastic for two different osteoblast-type cell lines. The simultaneous determination of multiple parameters at the individual cell level affords opportunities for data-base filtering and mining in order to select biologically relevant information. The automated velocity analysis has been demonstrated to be accurate within a level of uncertainly that is imposed by manual analysis of cell velocity from time-lapse image sets. Data quality and system performance measures have been developed that can serve as a basis for the design and quality control of large scale experiments for which this system is broadly applicable.
Methods
Cells and cell culture
MG-63 cells were obtained from ATCC and were grown according to recommended protocol (ATCC). Specifically, MG-63 cells were cultured in Minimum Essential Medium (Gibco/Invitrogen) with 2 mM L-glutamine and Earle's BSS adjusted to contain 1.5 g/L sodium bicarbonate, 0.1 mM non-essential amino acids, and 1.0 mM sodium pyruvate with 10% heat-inactivated fetal bovine serum (FBS). KM101 cells, a human bone marrow stromal cell line [22] were grown in 10% FBS in Iscove's Modified Dulbecco's Medium (Gibco/Invitrogen). Cells were cultured at 37°C in humidified atmosphere with 5% CO2. Prior to plating, cells were trypsinized with 0.25% trypsin (Gibco) and resuspended at a density of 2000 cell/ml. Aliquots of 60 ul were seeded into 384 well plates (Costar, black wall) yielding approximately 5 to 15 individual cells in the camera viewfield at the outset of imaging.
The plate layout included triplicate wells for each compound at each coating concentration for both cell lines. All operations were performed using multi-tip pipettors so that inter-well variations within-treatments were minimized to the greatest possible extent. Although the ideal plating pattern would be fully random, for practical purposes, advantage was taken of the interspacing of wells that occurs using 96 well multi-tip pipettors for applying solutions from 96 well "seed" plates into every-other well of the 384 well plate, producing "checker-board" patterns that intermixed treatments, doses, and uniform mock-coated control wells across the plate. Cell suspensions were pipetted from solution basins uniformly across each row.
Extracellular matrix surface coating
Mouse laminin I (Cultrex), mouse collagen type IV (Cultrex) and collagen type I (Sigma) were stored and reconstituted according to the manufacturer's instructions. Dilutions for coating were performed either in sterile water for collagen types I and IV or in Iscove's Modified Dulbecco's Medium (IMDM; Gibco/Invitrogen) without added serum for laminin I. Control wells were mock-coated using IMDM without added serum. Aliquots of 10 ul of solution were added to each well of a 384-well plate (Costar), and the plate was incubated for 1 hour at 37C. The collagen solutions were aspirated, and collagen-coated wells were allowed to air dry for about 20 minutes in a laminar flow hood. The laminin and mock-coated wells were then aspirated and all wells were rinsed with IMDM and aspirated prior to addition of cell suspensions.
Cell culture imaging cystem
The cell culture imaging system consists of a custom made environmentally controlled biochamber on an electronically controlled x-y motorized stage driven by stepper motor drive systems (Ludl Electronics, Ltd.) mounted upon an inverted microscope (Nikon TE 300) with electronically controlled motorized focus. The stage moves precisely to each well in the multi-well plate or to any number of locations within each well with a positioning repeatability of ± 1.5 um over the longest distance traveled by the stage. The biochamber temperature is controlled via heating cartridges with temperature feedback loops, and the humidity and CO2 content are controlled via commercial sensors to control feedback loops to a water reservoir heater and low pressure CO2 solenoid valve, respectively. The glass windows for illumination and microscope imaging are specifically heated through an electrically conductive indium tin oxide (ITO) coating using feedback temperature regulation so that condensation does not occur on these surfaces. A custom instrument control program, written in Visual Basic and C++, integrates control of the microscope stage, focus, optical filters, shutters, camera, fluidics, image storing functions, thermal zones, and subsystems through a specialized serial interface board with eight RS-232 connections.
Video time-lapse imaging and analysis
Images were acquired at 30 minute intervals with a 10× objective on a Nikon TE300 inverted microscope with a Photometrics SenSys high resolution (7 × 9 mm, 1036 × 1318 pixel chip) CCD camera (Roper Scientific). Image sequences were processed using a custom software program that identifies and records the location and morphological characteristics of cell-like objects (see Cell Segmentation and Outline Determination). The centroids of cell-like objects are then linked through sequential images to construct "tracks" that trace the route of individual cells according to proprietary algorithms [see Cell tracking algorithm]. The information accumulated during the processing is represented in probabilistic form so that the decision-making process does not have to be "black-and-white" (e.g. is the given object a cell or not, or is the given track the right track or not?), but is postponed until the end of the decision-making chain allowing for corrections for "lost" cells.
In each experiment for this study, 96 or more wells were imaged for at least 3 days, yielding more than a million records, each containing multiple measurements acquired for each object at each imaging time-point, i.e. every 30 minutes. These measurements fall into the categories of motility (scalar and vector forms of velocity, linearity, measures of deviation and frequency of oscillation of cell paths, track size and boundary), morphology and texture (area, perimeter, elongation, eccentricity, roughness, intensity and variation of intensity), summary statistics (object counts, cell counts, apoptotic frequency when fluorescence vital staining is applied) and complex parameters such as cell motility persistence [2], proximity analysis (cell-cell interaction, frequency and duration), division detection, growth rate, and viability (not all are applicable).
Cell segmentation and outline determination
The optical characteristics for inverted light microscopy of 384-well plates present challenges for robust segmentation of live cells. The meniscus obviates phase contrast; brightfield images are low in contrast and require significant processing. So a succession of filters is used to increase the difference (signal to noise ratio) between the background and foreground (cell-like objects), followed by an efficient region-growing operation that segments cell-like objects from the background. Heterogeneity of illumination across the image is reduced using local histogram equalization. Variations in illumination between images (across time) are handled using histogram matching. Background variations are smoothed using anisotropic filtering and adaptive median filtering, preserving cell detail and texture. Finally with brightfield microscopy, the cell boundary produced by the cytoplasmic membrane easily blends into the surrounding background, so a unique set of gradient variation and texture filters is applied to enhance the cell outline. Following filter-based enhancement, a region-growing operation identifies contiguous areas of cell-like or background-like pixels to segment cell-like objects from the background. A still more involved cell boundary determination can be achieved via active contour techniques (snakes) at the operator's discretion.
Cell tracking algorithm
The time-lapse interval in multi-well experiments is dependent upon practical considerations including the cell-type specific rate of motion, the total number of wells, the rates of stage movement, camera operation, and so on. For very fast cells such as T cells, imaging is performed on subsets of wells on a rotational basis in order to achieve intervals short enough for reliable track construction. At longer intervals, and particularly for view-fields containing many similar cells, cell track linking across the interval becomes increasingly unreliable because cells change shape and direction frequently, and as their paths converge, incorrect links may be assigned probabilities equal to or greater than correct ones.
Tracking is achieved by linking the matched cell-like objects between consecutive images in a probabilistic manner using a succession of increasingly stringent criteria. First, for each cell-like object, a set of candidate matches is chosen from within maximum speed and acceleration limits. Within this set, cell-like objects (blobs) are assigned to tracks based on match probabilities. Blobs are compared using multiple features such as location displacement, size difference, eccentricity changes, grayscale intensity IQR changes and normalized cross-correlation of the respective image portions. Converting distances into probabilities is done using a Gaussian probability density function. Because we assume that the features are independent, we can also assume that the resulting probabilities are independent. Therefore we can combine the different probabilities,
into the matching probability P(matchij | i, j), where matchij is a candidate match between track i and blob j and dn is the distance for feature n.
Cell objects in one image may compete simultaneously for multiple matches to different cells in the next image. A rule-based algorithm develops tracks based upon the values of the matching probability and tracking scenario. "Tracking scenario" includes the recent history and circumstances such as cell merging and splitting. For example, when two tracked cells that are similar collide, the merged object cannot be assigned to either track so both tracks are terminated. When two merged cells subsequently split apart, two new tracks are initiated because the matching of cells before and after the collision is ambiguous [see Additional file 3]. The cell tracking algorithm is very efficient in comparison to the cell segmentation; both are completed at the rate of 32 images/min, for 658 × 517 images on an Intel P4 2.8 GHz PC with 512MB RAM.
Cell velocities
Both magnitude and directional velocity information are output from the linked positions of objects in sequential images. In this study, and for general investigation, we use a scalar average across several time points to smooth variation due to many factors. This average velocity represents the actual distance travelled, as determined by the movement of the centroid of the cell, divided by the elapsed time. Track lengths with fewer than 3 segments were not considered, and a maximum of 10 segments were included such that, for tracks longer than 10 segments, the velocity represents a running average. An exception to this method was used for "instantaneous velocity", as shown in Figure 8, where calculation was based upon displacement divided by time for the single track segment between the previous and current image. In order to exclude objects containing multiple cells in clusters or colonies, only velocity measurements for objects with areas less than three standard deviations above the mean area for each cell type were considered. The mean cell area and standard deviation was 1800 +/- 400 and 3100 +/- 900 square microns for KM101 and MG-63 cells, respectively. The lower limit for object area was 1200 square microns in both cases, based upon optimal visual exclusion of non-cell objects during set up of image processing variables. Velocity measurements were not included when the total area occupied by cells was greater than 30% of the viewfield area, i.e. when the cells were greater than 30% confluent. Finally, in order to filter out plate surface imperfections and adherent particles that gave rise to cell-like objects, data were excluded from tracks with fence sizes of less than 20 microns (See Results).
The technical accuracy of the imaging processing and data conversion steps were verified manually by constructing idealized image sequences with objects "seeded" at known pixel distances in order to generate known velocities based upon typical magnification and binning settings. Microscope optical magnification levels have been verified and calibrated with image pixel dimensions using a standard reticle.
Technical precision and quality control
Technical precision is here defined as the square root of the average variance for sets of triplicate velocity measurements (each measurement representing the mean velocity for all cells in the image) at single time points. Since absolute precision tended to increase in value for wells with cells of higher average velocity, and since there were sufficient data for analysis across the full range of velocities, we calculated summary precision values for triplicate sets ranked according to their measurement means (Figure 9). These data may be interpreted to indicate that the uncertainty in the mean velocity for a single well at a single time point ranged from approximately ± 0.08 um/min at the 0.2 um/min mean velocity level (CV = 40%) to approximately ± 0.12 at the 0.8 um/min level (CV = 15%) for KM101 cells, and it was slightly better for MG-63 cells. Sources of this measurement variability include the irreducible variation expected from random sampling from populations shown in Figure 6, taking into account the number, n, of cells in each sample, i.e. the number of cells imaged in each view-field as indicated in Figure 7.
Incidentally, the bin-based population variance, in which these precision estimates are rooted, should be employed when using the t-Test, rather than the variance of individual sets of replicates, to evaluate the significance of differences observed within experiments between triplicate means, i.e. for testing whether cell velocity was affected by an experimental treatment compared to a control treatment. The reason for this is that the sample mean and sample variance are independent when sampling from a normally distributed population. In other words, counter-intuitively, the mean of a set of three replicates that are widely spaced is likely to be as close to the "true value" as the mean of a set of replicates that are very closely spaced. The abundance of measurements made available by automation supports the validity of this claim, and would allow for slight adjustment of this principle when the data depart from normality, as it does here.
Stage noise
At the most fundamental level, because the stage mechanism re-centers each well into view after each time interval, images and data derived from them are subject to random errors associated with slight misalignment of the culture plate at each scan. The limits of this misalignment were determined by expanding images and manually tracking the motion of highlighted features of small imperfections on the culture surface throughout example image sequences. Such features were confined within a boundary of 3 square pixels over greater than 100 sequential images in all three experiments. Assuming Gaussian statistics, a limit of less than 1 pixel was estimated for the standard deviation of alignment of a single-pixel object, i.e. "stage noise". Under the magnification (10×) and binning conditions (2 × 2) used in these experiments, one pixel corresponds to 2 microns; this displacement for an object within a 30 minute time interval yields an upper limit for velocity of approximately 0.07 um/min. This calculation represents an upper limit to the contribution of stage noise, because in practice, stage-positioning error is expected to be a relatively small component of the total error that includes image-processing variability in determining cell outlines and object centroids.
Comparison of manual and automated tracking
Manual tracking was performed with a custom viewing program that enabled the user to store x and y coordinates by clicking on cells in sequential images with a computer-mouse. The 517 × 658 pixel images were displayed at approximately 7 × 8.5 inches on the monitor (approximately 75 pixels/inch) with a zooming option. A total of 27 cells were manually tracked from each of the four selected treatments. (For this exercise, a 27-cell limit was imposed by the nature of data output into a Microsoft Excel worksheet). The "rules" for manually placing the "centroid" and for terminating or initiating tracks were somewhat discretionary, e.g. the author continued to manually track cells through periods of contact with other cells, even though such scenarios were expected to involve track termination and re-initiation by the automated tracking algorithm [see Additional file 3].
For comparison of tracks on an object-by-object, point-by-point basis, we developed an algorithm to match objects from automated image analysis with manually tracked cells from each image. This algorithm identified and tabulated data from objects with automatically located centroids that fell within 10 pixels (20 um) of the x and y coordinates of the manually located cell positions that were determined by clicking with the mouse pointer. Following the initial comparison, both the manual tracking and the automated analysis were repeated on the four identical image sets, and the algorithm was applied to identify matched objects between the replicate manual and automated operations as well. Regression analysis and t-Test calculations were performed using R Project for Statistics [V. 2.0, see ].
Cell numbers and growth rates
Individual cells within colonies and cell clusters are not reliably recognized by current automated software using brightfield imaging. Instead, colonies and clusters are segmented as individual objects, and the areas of these objects provide a basis for estimating cell numbers. Doubling rates are calculated using linear regression of log transformed area-based cell numbers over time. Briefly, the slope of Log2(Cell Number) versus Time equals the doubling rate. As shown in this study, however, cell numbers do not correlate linearly will total cell area, and so doubling rates based upon an exponential growth model for area were called "area-doubling" rates.
Experimental design, normalization, and statistical analysis
A randomized complete block design [31] was used so that all informative factors (cell-line, compound, and dose) could be individually evaluated and separated from nuisance factors contributed by between-experiment variation and technical variability as evaluated between replicate wells within each experiment.
The behaviour of the cell velocities over time was analysed using 4-hour intervals and linear regression. First, velocity measurements were averaged for wells with the same cell line and treatment. Each data-point represents the average for 8 images (acquired every 30 minutes over 4-hour time intervals) from triplicate wells in each of 3 experiments. Second, a linear trend line was fitted to each profile. Finally the intercept (initial average velocity) and the slope (behaviour of average velocity over time) were compared across cell-line and treatment. The model fitting was done using SAS Proc Mixed. The significance of coefficients in these models is tested with student's t test. Overall tests for the equivalence of starting velocities are performed using the Chi-square test. Tests for equivalence of slopes are also performed. Pair-wise comparison with Bonferroni adjustment is also employed to see differences within each pair of treatments.
The dose-response of the cell-velocities was analysed within a 24-hour period. There was considerable variation between experiments as measured by the intra-class correlation (ICC), or ratio of variance from individual factors to the total variance [32]. Therefore, prior to combining across experiments, all data for velocity and growth rates were normalized using an additive model based upon the difference between mock-coated plastic wells for each experiment. That is, the normalized measurements, Mn, were calculated from the original measurements, Mi, as follows: Mn = (Mi - Pi) + Pe, where Pe equalled the overall mean measurement for mock-coated wells for all experiments, and Pi equalled the individual means for mock-coated wells within experiments. Means were estimated using Tukey's biweight single-step M-estimator [33]. The ICC for informative factors increased from 0.41 to 0.89, while that for the major nuisance factor, between-experiment variability, decreased from 0.58 to 0.09. The proportional contribution from technical variability (replication error) increased slightly from 0.005 to 0.011 due the decrease in overall variance after normalization.
Authors' contributions
AB assisted with design and implementation of the experiments, analysis of data and drafting of the manuscript. CA performed statistical analysis, assisted with data analysis and normalization, and contributed to software. DK designed and built the prototype as well as contributed to the software control, data analysis, and drafting of the manuscript. LQ contributed to design and description of the segmentation and tracking algorithms, construction of video examples, and custom software devices for data validation. TS designed algorithms for object matching (manual vs automated tracking) and queries for extraction of data. HW performed and summarized statistical analysis of the kinetic data. DS and YH carried out cell culture and monitoring of the automated system. JG and TC participated in design of the study and monitoring of the system. RH and LC supervised and coordinated the study. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Time-lapse motility and proliferation of KM101 and MG-63 cells. Images from four wells were combined to show the differential effects of laminin (upper left), collagen type I (upper right), collagen type IV (lower left) and mock-coated plastic (lower right). Red outlines indicate the perimeters of the cell "objects" from which object areas and centroid positions are derived. Green lines indicate tracks that are established between successive centroid positions that define the individual cell paths. In this image sequence, track segments are erased following 10 hours elapsed time. The total elapsed time of 114 hours for each cell type shows cells from shortly after seeding to near confluence. Images were acquired at 30-minute intervals and are displayed in the video at a rate of 6 images per second.
Click here for file
Additional File 2
Acceleration of MG-63 cells. Time-lapse videos show MG-63 cells on laminin-coated (10 ug/ml, first image sequence "I04") and mock-coated plastic (second image sequence "J03"). Images were acquired at 30-minute intervals and are displayed in the video at a rate of 6 images per second. Green lines indicate automatically generated tracks connecting successive centroid positions of objects. Track segments are erased after 15 images, so the maximum track length represents 7.5 hours' migration. A total of 118 hours elapsed time is shown for each image sequence.
Click here for file
Additional File 3
Tracking scenarios. This tracking example sequence was taken from the set of images that are shown in the upper left panel of Movie 1 in order to illustrate in more detail how the tracking algorithm handles "scenarios" such as the colliding and splitting of cells. A total elapsed time of 35 hours (70 scans) is depicted, and specific details are given in the video to explain example scenarios. Original image quality is not preserved in the videos due to necessary compression.
Click here for file
Acknowledgements
We thank Julie Glowacki and Joel Greenberger for inspiration in the development of the automated system and in the performance of this study. We thank William Golding for assistance with statistical issues, Darrin Sabol for assistance with image processing, Yifang Song for manual cell counting, and Lori McKenzie and Kris Sachsenmeier for technical advice. The author thanks Sallie Boggs for seed ideas and John Wolf, Peter Friedl, and the anonymous reviewers for helpful comments on the manuscript. This investigation was supported by the National Institutes of Health grant number 1 RO1 EB 001051-01A2 from the National Institute of Biomedical Imaging and Bioengineering to R.K.H.
==== Refs
Dimilla PA Stone JA Quinn JA Albelda SM Lauffenburger DA Maximal migration of human smooth muscle cells on fibronectin and type IV, collagen occurs at an intermediate attachment strength J Cell Biol 1993 122 729 737 8335696 10.1083/jcb.122.3.729
Lauffenburger DA Horwitz AF Cell migration: a physically integrated molecular process Cell 1996 84 359 369 8608589 10.1016/S0092-8674(00)81280-5
Boyden S The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear leucocytes J Exp Med 1962 115 453 466 13872176 10.1084/jem.115.3.453
Spessotto P Giacomello E Perri R Improving fluorescence-based assays for the in vitro analysis of cell adhesion and migration Mol Biotechnol 2002 20 285 304 11936258 10.1385/MB:20:3:285
Nelson RD Quie PG Simmons RL Chemotaxis under agarose: a new and simple method for measuring chemotaxis and spontaneous migration of human polymorphonuclear leukocytes and monocytes J Immunol 1975 115 1650 1656 1102606
Silva J Beckedorf A Bieberich E Osteoblast-derived oxysterol is a migration-inducing factor for human breast cancer cells J of Biol Chem 2003 278 25376 25385 12734199 10.1074/jbc.M301233200
Albrecht-Buehler G The phagokinetic tracks of 3T3 cells Cell 1977 11 359 404 10.1016/0092-8674(77)90057-5
Providence KM Kutz SM Staiano-Coico L Higgins PJ PAI-1 gene expression is regionally induced in wounded epithelial cell monolayers and required for injury repair J Cell Physiol 2000 182 269 280 10623891 10.1002/(SICI)1097-4652(200002)182:2<269::AID-JCP16>3.0.CO;2-E
Huttenlocher A Ginsberg MH Horwitz AF Modulation of cell migration by integrin-mediated cytoskeletal linkages and ligand-binding affinity J Cell Biol 1996 134 1551 1562 8830782 10.1083/jcb.134.6.1551
Ware MF Wells A Lauffenburger DA Epidermal growth factor alters fibroblast migration speed and directional persistence reciprocally and in a matrix-dependent manner J Cell Sci 1998 111 2423 2432 9683636
Lewis WH Locomotion of Lymphocytes Bulletin of the Johns Hopkins Hospital 1931 49 29 36
Huttenlocher A Lakonishok M Kinder M Wu S Truong T Knudsen KA Horwitz AF Integrin and cadherin synergy regulates contact inhibition of migration and motile activity J Cell Biol 1998 141 515 526 9548728 10.1083/jcb.141.2.515
Shutt DC Daniels KJ Carolan EJ Hill AC Soll DR Changes in the motility, morphology, and F-actin architecture of human dendritic cells in an in vitro model of dendritic cell development Cell Motil Cytoskeleton 2000 46 200 21 10913967 10.1002/1097-0169(200007)46:3<200::AID-CM5>3.0.CO;2-M
Cox EA Sastry SK Huttenlocher A Integrin-mediated adhesion regulates cell polarity and membrane protrusion through the Rho Family of GTPases Mol Biol Cell 2001 12 265 277 11179414
Maheshwari G Wells A Griffith LG Lauffenburger DA Biophysical integration of effects of epidermal growth factor and fibronectin on fibroblast migration Biophys J 1999 76 2814 2823 10233097
Greenberger J Goff JP Epperly M Shields DS Yanez-Hanley K Bahnson A Koebler DK Houck RK Huard J Expansion of hemapoietic stem cells in vitro as a model system for human tissue engineering Orthop Clin North Am 2000 31 499 510 10882474
Greenberger J Goff JP Epperly M Shields DS Yanez-Hanley K Bahnson A Koebler D Houck RK Huard J Lewandrowski K-U Combinatorial cell culture applications to tissue engineering Tissue Engineering and Biodegradable Equivalents: Scientific and Clinical Applications 2002 New York: Marcel Dekker Inc
Greenberger JS DiMilla PA Domach MM Houck RK Method and apparatus for holding cells Patent No: US 6,008,010, Dec. 28, 1999
Bahnson A Koebler D Athanassiou H Houck R Sachsenmeier K Qian L Method and apparatus for monitoring of proteins and cells Patent No: US 6,830,931 B2, Dec. 14, 2004
Craggs G Berry J Tyson KL Sachsenmeier KF Loader JA Boyd R Poyser HR Page MJ Adam PJ Terrett JA NSE2, a protein up-regulated in breast cancer, exhibits multiple oncogenic properties Appl Genomics and Proteomics 2003 2 101 100
Adam PJ Berry J Loader JA Tyson KL Craggs G Smith P De Belin J Steers G Pezzella F Sachsenmeier KF Stamps AC Herath A Sim E O'Hare MJ Harris AL Terrett JA Arylamine N -acetyltransferase-1 is highly expressed in breast cancers and conveys enhanced growth and resistance to etoposide in vitro Mol Cancer Res 2003 1 826 835 14517345
FitzGerald TJ Santucci MA Harigaya K Woda B McKenna M Sakakeeny MA Pierce JH Kase K Holland CA Greenberger JS Radiosensitivity of permanent human bone marrow stromal cell lines: effect of dose rate Int J Radiat Oncol Biol Phys 1988 15 1153 1159 3182348
Billiau A Edy VG Heremans H Van Damme J Desmyter J Georgiades JA De Somer P Human interferon: mass production in a newly established cell line, MG-63 Antimicrob Agents Chemother 1977 12 11 15 883813
Lajeunesse D Kiebzak GM Frondoza CG Sacktor BV Regulation of osteocalcin secretion by human primary bone cells and by the human osteosarcoma cell line MG-63 Bone Miner 1991 14 237 250 1657256 10.1016/0169-6009(91)90025-U
Heid PJ Voss E Soll DR 3D-DIASemb: A computer-assisted system for reconstructing and motion analyzing in 4D every cell and nucleus in a developing embryo Developmental Biology 2002 245 329 347 11977985 10.1006/dbio.2002.0631
Peckham M Miller G Wells C Zicha D Dunn GA Specific changes to the mechanism of cell locomotion induced by overexpression of beta-actin J Cell Sci 2001 114 1367 1377 11257002
Friedl P Noble PB Zanker KS Lymphocyte locomotion in three-dimensional collagen gels. Comparison of three quantitative methods for analysing cell trajectories J Immunol Methods 1993 165 157 165 7901283 10.1016/0022-1759(93)90341-4
Friedl P Brocker EB Reconstructing leukocyte migration in 3D extracellular matrix by time-lapse videomicroscopy and computer-assisted tracking Methods Mol Biol 2004 239 77 90 14573911
Hartmann-Peterson R Walmod PS Berezin A Berezin V Bock E Individual cell motility studied by time-lapse video recording: Influence of experimental conditions Cytometry 2000 40 260 270 10918277 10.1002/1097-0320(20000801)40:4<260::AID-CYTO2>3.0.CO;2-J
Giumciuglio D Cai T Filanti C Manduca P Albini A Effect of osteoblast supernatants on cancer cell migration and invasion Cancer Lett 1995 97 69 74 7585480 10.1016/0304-3835(95)03955-V
Tabachnick BG Fidell LS Using Multivariate Statistics 2000 Boston: Pearson Allyn & Bacon
Carrasco JL Jover L Estimating the generalized concordance correlation coefficient through variance components Biometrics 2003 59 849 858 14969463 10.1111/j.0006-341X.2003.00099.x
Wilcox RR Introduction to Robust Estimation and Hypothesis Testing 1997 New York: Academic Press
| 15831094 | PMC1097721 | CC BY | 2021-01-04 16:39:11 | no | BMC Cell Biol. 2005 Apr 14; 6:19 | utf-8 | BMC Cell Biol | 2,005 | 10.1186/1471-2121-6-19 | oa_comm |
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-201584770110.1186/1471-2121-6-20Research ArticleIdentification of a novel Rev-interacting cellular protein Kramer-Hämmerle Susanne [email protected] Francesca [email protected] Christian [email protected] Horst [email protected] Michelle [email protected] Thomas [email protected] Volker [email protected] Ruth [email protected] Institute of Molecular Virology, GSF-National Research Center for Environment and Health, Ingolstädter Landstraße 1, D-85764 Neuherberg, Germany2 Department of Experimental Medicine, University of Rome Tor Vergata, Via Montpellier 1, Rome 00133, Italy3 Genomatix Software GmbH, Landsbergerstr. 6, D-80339 München, Germany2005 24 4 2005 6 20 20 30 11 2004 24 4 2005 Copyright © 2005 Kramer-Hämmerle et al; licensee BioMed Central Ltd.2005Kramer-Hämmerle 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
Human cell types respond differently to infection by human immunodeficiency virus (HIV). Defining specific interactions between host cells and viral proteins is essential in understanding how viruses exploit cellular functions and the innate strategies underlying cellular control of HIV replication. The HIV Rev protein is a post-transcriptional inducer of HIV gene expression and an important target for interaction with cellular proteins. Identification of Rev-modulating cellular factors may eventually contribute to the design of novel antiviral therapies.
Results
Yeast-two hybrid screening of a T-cell cDNA library with Rev as bait led to isolation of a novel human cDNA product (16.4.1). 16.4.1-containing fusion proteins showed predominant cytoplasmic localization, which was dependent on CRM1-mediated export from the nucleus. Nuclear export activity of 16.4.1 was mapped to a 60 amino acid region and a novel transport signal identified. Interaction of 16.4.1 with Rev in human cells was shown in a mammalian two-hybrid assay and by colocalization of Rev and 16.4.1 in nucleoli, indicating that Rev can recruit 16.4.1 to the nucleus/nucleoli. Rev-dependent reporter expression was inhibited by overexpressing 16.4.1 and stimulated by siRNAs targeted to 16.4.1 sequences, demonstrating that 16.4.1 expression influences the transactivation function of Rev.
Conclusion
These results suggest that 16.4.1 may act as a modulator of Rev activity. The experimental strategies outlined in this study are applicable to the identification and biological characterization of further novel Rev-interacting cellular factors.
==== Body
Background
The human immunodeficiency virus (HIV) Rev protein is a small (116 amino acids) post-transcriptional activator of expression of incompletely spliced and unspliced HIV mRNAs. Since these HIV transcripts direct production of progeny virions, Rev is a crucial factor in HIV replication (for overview see [1]). Rev interacts with HIV mRNAs by binding to a structured RNA element called the RRE (Rev response element). Rev offsets the activities of inhibitory sequences (INS) in HIV-1 mRNAs [2,3] and promotes their export to the cytoplasm. Once in the cytoplasm, Rev may also stimulate production of viral proteins on translational level (reviewed in [4]).
Rev characteristically localizes to the nucleus, where it accumulates in nucleoli. However, a proportion of the Rev molecules expressed in a cell continuously shuttles between nucleus and cytoplasm by using active transport mechanisms both for entry into and exit from the nucleus.
Mutational analyses of the Rev protein have identified various functionally important regions, indicating that Rev is organized into modular domains (Fig. 1A; reviewed in [5] and [6]). The N-terminal domain of Rev contains an arginine-rich motif (ARM; amino acids 35 to 50) with dual functions as a nuclear localization signal (NLS) and RNA binding domain. Sequences flanking the ARM (amino acid regions 12 to 29 and 52 to 60) direct multimerization of Rev. The C-terminal domain of Rev, also known as activation domain, contains a leucine-rich motif (amino acid region 75 to 83) that functions as a nuclear export signal.
Figure 1 Analysis of interaction of 16.4.1 with Rev. (A) Schematic overview of domains of Rev. Locations of functional regions are taken from [5]. Numbering of amino acids (aa) is based on the Rev sequence of HIV-1 isolate HXB-2. MI and MII: Regions that direct multimerization of Rev. NLS: nuclear localization signal; NES: nuclear export signal. (B) Identification of amino acid positions in Rev required for interaction with 16.4.1. Bait proteins containing Rev or various mutants of Rev were analysed for interaction with 16.4.1-prey. Numbers indicate positions of amino acid changes in Rev mutants. Mutations are as follows: RevM4: Y23 to D23, S25 to D25, N26 to L26 [45]; Rev M10BL: L78 to D78, E79 to L79 and insertion of EDLP between L81 and T82 [47]; Rev M5: R38 to D38, R39 to L39 [45]; RevSLT40: I59 to D59, L60 to D59 [46]. Interaction is indicated by growth of yeast transformants under selective conditions (≥ 500 transformants per plate). Results represent four independent experiments. The red box marks the amino acid positions in Rev required for interaction with 16.4.1 and the location of the putative 16.4.1-interaction region of Rev (aa 38 to 60). (C) Identification of regions of 16.4.1 required for interaction with Rev. Prey proteins containing various regions of the 16.4.1 domain were analysed for interaction with Rev-bait. Interaction was analysed by growth of yeast transformants under selective conditions. Results represent three independent experiments. +++ 700–1400 transformants per plate; + 200–400 transformants per plate; – no transformants. 16.4.1 regions comprising amino acids 2 to 133 and 39 to 171 interacted with Rev with similar efficiency as full-length 16.4.1, whereas regions 2 to 73 and 74 to 171 showed weaker interaction. No interaction was observed with regions 2 to 38 and 134 to 171. The red arrow indicates the putative Rev-interaction region of 16.4.1 (aa 39 to 133).
Biochemical analyses indicate that Rev directly binds the nuclear transport receptors Importin β and CRM1/Exportin 1 [7-9]. Interaction of Rev with CRM1/Exportin 1 was confirmed by two-hybrid assays in yeast [10] and in human cells [11,12]. Together with results from various other experimental approaches (reviewed in [6]), these observations have led to the concept that import of Rev into the nucleus is mediated by interaction of the ARM/NLS with Importin and export of the Rev-RNA complex from the nucleus by interaction of the Rev-NES with CRM1/Exportin 1.
Various other Rev-interacting cellular factors have been identified by using Rev or segments of Rev for yeast two-hybrid screening of cDNA libraries or for biochemical purification of interacting factors from cell extracts.
Cellular factors shown to interact with the ARM of Rev include p32 [13,14] and B23 [15,16]. Human p32 was recently reported to block splicing of Rev-dependent HIV transcripts [17]. The nucleolar protein B23 was shown to stimulate nuclear import of Rev [18] and counteract aggregation of Rev [19]in vitro.
The C-terminal domain of Rev interacts with various human nucleoporins, including hRIP/hRab, NLP-1, Nup98, and Nup214 [20-26]. Other factors shown to interact with this domain of Rev are eIF-5A [27] and the nuclear kinesin-like protein REBP [28]. hRIP/hRab, Nup 98 and eIF-5A interact with CRM1 as well as Rev [10,25,26,29], suggesting that Rev can associate with CRM1 in multifactorial complexes in which CRM1 "bridges" the interaction of Rev with other factors. Rev-CRM1 complexes containing hRIP/hRab or eIF-5A may be crucial for Rev-dependent export of HIV RNAs, since eIF-5A and hRIP/hRab have been shown to be essential for Rev-directed RNA export in Xenopus oocytes and in human cells, respectively [30,31].
Nuclear export of Rev has proven to be exemplary for many viral and cellular factors (reviewed in [32,33]). Since the discovery of leucine-rich signals in Rev [34] and in the cellular regulatory factor PKIα [35], these sequences have been shown to mediate the export of numerous factors from the nucleus by CRM1/Exportin1 [36]. The drug Leptomycin B (LMB), first shown to block nuclear export of Rev [37], proved to be a potent inhibitor of CRM1-dependent export [38,39] and is now widely used to identify transport substrates of CRM1. Elucidating interactions of Rev with cellular factors is highly relevant to understanding pathogenicity of HIV and may have an impact on the design of therapeutic anti-HIV strategies. The functional diversity of Rev and its activities in both nuclear and cytoplasmic compartments of the cell suggest the existence of still unidentified Rev-interacting factors. Therefore we reasoned that screening of a human cDNA library with Rev as "bait" should lead to isolation of novel Rev-interacting human factors. Of particular interest would be the identification of unknown human gene products, since their interaction Rev would not only be relevant for Rev function but would also provide a key for biological characterisation of these novel factors.
Here we identify a human cDNA that encodes a novel protein (16.4.1) that interacts specifically with Rev via sequences in the N-terminal half of Rev. We show that 16.4.1 is exported from the nucleus by CRM1 and localizes to the cytoplasm. In Rev-expressing cells, 16.4.1 is recruited to nucleoli. 16.4.1 has a negative effect on Rev function in a Rev-reporter assay. These results suggest that 16.4.1 can act as a modulator of Rev function.
Results
Identification of novel HIV-1 Rev-interacting proteins
To identify novel Rev-interacting proteins, we screened a library of cDNAs derived from the human Jurkat T-cell line with full-length Rev as bait in a yeast two-hybrid system. Repeated selection procedures led to isolation of two library plasmids (11.5.1 and 16.4.1) encoding specific interactors of Rev.
Sequence analyses and data base comparisons revealed that the 936 bp insert in plasmid 11.5.1 is identical with a segment of a 1543 bp cDNA encoding human DNA binding protein B (dbpB; NCBI accession number BC002411) [40]. The predicted coding sequences in the 11.5.1 cDNA comprise the C-terminal 139 amino acids of the dbpB protein (324 amino acids; NCBI accession number M24070). Several biological activities have been attributed to dbpB, including binding to DNA [41] and RNA [42,43] and regulation of transcription [44].
The other library plasmid 16.4.1 contained a 696 bp insert of which a region of over 450 nucleotides showed strong similarity to a sequence within a human fetal heart cDNA (NCBI accession number W67699). In the fetal heart cDNA the matching region encompasses a predicted open reading frame. Alignment of the 16.4.1 and the fetal heart cDNA sequences yielded a sequence encoding a hypothetical 171 amino acid 16.4.1 protein. Since interaction with Rev is the first biological activity associated with this gene product, we analysed interaction of Rev with the 16.4.1 protein in more detail.
To investigate which regions of Rev contribute to interaction with the 16.4.1 protein, we analysed the capacity of various known mutants of Rev to interact with 16.4.1 in the yeast two-hybrid assay. The amino acid exchanges in these mutants map to regions associated with major biological properties of Rev (Fig. 1B), including multimerization (RevM4 [45] and RevSLT40 [46]), RNA binding and nuclear localization/accumulation (RevM5 [45]) and nuclear export of Rev (RevM10BL [47]). Expression of LexA-Rev-mutant bait proteins in yeast transformants was confirmed by Western blot analysis with polyclonal antibodies against Rev (data not shown). As positive control for Rev interaction, interaction analysis was performed with LexA-Rev bait and B42AD-Rev prey, confirming oligomerization of wildtype Rev molecules with each other (data not shown).
While Rev mutants RevM4 and Rev M10BL were capable of interacting with 16.4.1, no interaction was observed with Rev mutants RevM5 and RevSLT40 (Fig. 1B). These results indicate that amino acid residues R38 or R39 of the ARM and I59 or L60 of the multimerization region II (MII) are required for interaction of Rev with the 16.4.1 protein. Furthermore, they suggest that the 16.4.1 interacting sequences in Rev are located between aa positions 38 and 60.
For more detailed study of the interaction of the 16.4.1 protein with Rev, yeast two-hybrid analysis was performed with various segments of the 16.4.1 cDNA as prey and wildtype Rev as bait (Fig. 1C). Amino acid regions of 16.4.1 extending from position 2 to 133 and from position 39 to 171 showed similar Rev-binding capacity as full-length 16.4.1 protein. In contrast, both the N-terminal region (2 to 38) and the C-terminal region (134 to 171) of 16.4.1 failed to interact with Rev. While 16.4.1 protein fragments from position 2 to 73 or position 74 to 171 clearly interacted with Rev, interactions were weaker than that of full-length 16.4.1. These results indicate that the Rev-interacting region of the 16.4.1 protein is located between amino acid positions 39 and 133 and that, within this region, sequences N- and C-terminal of position 73 contribute to interaction with Rev.
Interaction of the 16.4.1 protein with Rev, CRM1 and itself in human cells
The interaction of the 16.4.1 protein with Rev in yeast raises the question whether the 16.4.1 protein can also interact with Rev in human cells. It was also of interest whether 16.4.1 is capable of interacting with human CRM1, since CRM1 has been shown to interact with several Rev-associated factors (see Background).
We addressed these issues with a mammalian two-hybrid assay, in which the interaction of a protein fused to the Gal4 DNA-binding domain with a second protein fused to the VP16-activator domain induces transcription of a luciferase reporter gene from a synthetic promoter (for details see Materials and Methods). Rev was fused to VP16 (VP16-Rev) to avoid unspecific interactions between the acidic VP16 domain [48] and the basic Rev protein (estimated pI = 9.93; MacVector calculation). Functionality of VP16-Rev was demonstrated (data not shown) in a Rev-reporter assay [3]. For interaction analysis, HEK293 cells were cotransfected with expression plasmids for VP16-Rev and Gal4-16.4.1 fusion proteins and the reporter plasmid pG5luc. As shown in Fig. 2, a ≈11-fold mean induction of luciferase activity was observed in 14 independent transfection experiments. Assessment of interaction of 16.4.1 with human CRM1 in cells coexpressing Gal4-16.4.1 and VP16-hCRM1 revealed a ≈41-fold mean induction of luciferase activity export (n = 7) (Fig. 2). Self-interaction of the 16.4.1 domain was analysed by coexpressing Gal4-16.4.1 and VP16-16.4.1, resulting in ≈12-fold mean induction of luciferase activity (n = 6).
Figure 2 Interaction of 16.4.1 with HIV-1 Rev, hCrm1 and with itself in human cells. 16.4.1 interactions in human cells were analysed with a mammalian two-hybrid assay in which the interaction of a protein fused to the Gal4 DNA-binding domain with a second protein fused to the VP16-activator domain induces transcription of a luciferase reporter gene. HEK293 cells were cotransfected with pBIND-16.4.1 plasmid for expression of Gal4-16.4.1 fusion protein, a pACT plasmid for expression of the indicated VP16-fusion protein and with the pG5luc reporter plasmid. Parallel cultures were cotransfected with pG5luc and the pBIND and pACT vectors to determine basal expression of the luciferase gene. Cells were lysed 48 hours after transfection and luciferase activities determined. Bars indicate the mean fold- induction of luciferase activity over basal expression ± SEM (standard error of the mean) and represent at least 6 independent transfection experiments. Cells coexpressing Gal4-16.4.1 and VP16 fused with Rev (grey bar), hCRM1 (black bar) or 16.4.1 (vertically striped bar) domains showed significantly stronger induction of luciferase production than cells coexpressing Gal4-16.41 and unfused VP16 (diagonally striped bar). Statistical analysis was performed by two-tailed Mann-Whitney U test.
In all three cases, induction of luciferase activity was significantly (p < 0.04) increased over induction levels obtained in control assays with unfused VP16 and Gal4-16.4.1 (3.3-fold; n = 7).
These results indicate that the 16.4.1 domain is capable of interacting with Rev as well as with the export receptor CRM1 and of forming homo-oligomers in human cells.
Cytoplasmic localization of 16.4.1 is CRM1/Exportin 1 dependent
Comparison of the sequence in the 16.4.1 cDNA with the fetal heart cDNA indicated that the 16.4.1 sequence was incomplete at its 5' terminus. To generate a full-length (171 aa) 16.4.1 coding sequence, nucleotides encoding the first 8 N-terminal amino acids derived from the predicted open reading frame of the fetal heart cDNA were inserted upstream of the 16.4.1 cDNA. To analyse subcellular localization of the 16.4.1 protein, cells were transfected with plasmids directing expression of fusion proteins containing full-length 16.4.1 or various segments of 16.4.1. Those fusion proteins contained either a N-terminal IgG1 tag or a C-terminal GFP tag. The full-length IgG1-16.4.1 fusion protein (IgG1-2-171) was located mainly in the cytoplasm of HeLa cells (Fig. 3A). IgG1 fusion proteins with 16.4.1 regions extending from amino acid position 2 to 133, 39 to 171 and 74 to 171 showed similar predominantly cytoplasmic localization (Fig. 3A). In contrast, IgG1 fusion proteins with the N-terminal region (2 to 38) or the C-terminal region (134 to 171) of 16.4.1 were apparent in both nucleus and cytoplasm, similar to unfused IgG1. These results demonstrate that the 16.4.1 protein is capable of cytoplasmic accumulation and suggest that sequences directing cytoplasmic localization of the 16.4.1 protein are located between amino acid positions 74 to 133.
Figure 3 CRM1-dependent cytoplasmic localization of 16.4.1. HeLa cells were transfected with plasmids directing expression of IgG1-16.4.1 or 16.4.1-GFP fusion proteins and subcellular distribution of tagged proteins analysed 24 hours later in fixed cells. (A) Subcellular distribution of IgG1 fusion proteins containing full length 16.4.1 or various segments of 16.4.1. IgG1-16.4.1 proteins were detected by immunocytochemistry with a Cy3-conjugated anti human IgG1 antibody. A schematic diagram of the IgG1-16.4.1 fusion proteins and a summary of their localization behavior are shown at the top. Representative images of cells expressing IgG1 fusion proteins containing the indicated amino acid regions of 16.4.1 are shown below. IgG1 fusion proteins containing amino acids 2–171 (full-length), 2–133, 39–171 and 74–171 of 16.4.1 were predominantly cytoplasmic, whereas fusion proteins with amino acids 2–38 or 134–171 and unfused IgG1 were both cytoplasmic and nuclear. Scale bars: 20 μm. (B) Disruption of predominantly cytoplasmic localization of 16.4.1-GFP by treatment of cells with the CRM1-inhibitor Leptomycin B (LMB). HeLa cells were transiently transfected with plasmids for expression of 16.4.1-GFP, PKIα-GFP, Rev(52–116)-GFP and unfused GFP. Cells were treated with LMB (5 nM) for two hours. Representative images of the subcellular distribution of the GFP fusion proteins in untreated (-) and LMB treated cells (+) are shown at the top. Symbols in the graph indicate the nuclear proportion of fluorescence (%) in individual cells and horizontal lines and numbers the median of the cell population. LMB treatment increased the median nuclear proportion of 16.4.1-GFP from 25% to 44%. LMB had a similar effect on localization of GFP fusion proteins containing PKIα or the carboxyterminal region of Rev (aa 52–116), which are known transport substrates of CRM1. In contrast, LMB only marginally affects subcellular distribution of unfused GFP. Scale bars: 10 μm.
The 16.4.1-GFP fusion protein showed similar cytoplasmic localization as IgG1-16.4.1 (Fig. 3B). Quantitative evaluation of subcellular distribution of GFP fluorescence [49] revealed that only 25% of total fluorescence was contained in the nuclei of 16.4.1-GFP expressing cells. This localization is comparable to that of GFP fusion proteins containing PKIα (PKIα-GFP) or the carboxyterminal half of Rev (Rev(52–116)-GFP), which localize to 23% and 25%, respectively, in the nucleus (Fig. 3B). PKIα and the carboxyterminal half of Rev contain well-characterized recognition signals for CRM1/Exportin 1-dependent export [36]. Similar cytoplasmic localization of 16.4.1-GFP and interaction of 16.4.1 with CRM1/Exportin 1 in human cells (Fig. 2) raised the possibility that cytoplasmic localization of 16.4.1-GFP at steady state may involve nuclear export of 16.4.1 by CRM1/Exportin 1. Therefore we analysed the effect of Leptomycin B (LMB), an inhibitor of CRM1-dependent nuclear export [11,37,38] on subcellular distribution of 16.4.1-GFP. LMB treatment significantly increased the nuclear proportion of 16.4.1-GFP from 25% to 44%. LMB-induced nuclear redistribution was similar in cells expressing PKIα-GFP and Rev(52–116)-GFP, whose nuclear proportion increased to 49% and 46%, respectively. Quantitative analysis demonstrated that 45% of unfused GFP localized to the nucleus, in agreement with its known capacity to diffuse throughout the cell [50]. LMB had no significant effect on subcellular distribution of unfused GFP.
These results indicate that cytoplasmic localization of 16.4.1 involves nuclear export by CRM1/Exportin1. Amino acid region 74 to 133 of 16.4.1 seems to be crucial for those transport processes.
Identification of a candidate nuclear export signal in 16.4.1
To further characterize the involvement of the amino acid region 74–133 in cytoplasmic localization of 16.4.1, we assessed subcellular distribution of GFP fusion proteins containing this region of 16.4.1. Cells expressing a GFP fusion protein with a single copy of aa 74–133 of 16.4.1 contained a higher proportion of nuclear fluorescence (38%, Fig. 4) than cells expressing GFP fusion proteins with full-length 16.4.1 (25%, Fig. 3B). However, GFP-fusion proteins containing two copies of region 74 to 133 of 16.4.1 in tandem showed similar cytoplasmic localization (Fig. 4; 27% nuclear proportion) as full-length 16.4.1-GFP (Fig. 3B; 25%). Treatment of cells with LMB raised nuclear proportions of GFP fusion proteins with one or two copies of 16.4.1 region 74–133 to similar levels as full-length 16.4.1-GFP. These results suggest that the region between amino acid positions 74 and 133 contains a CRM1/Exportin 1 dependent nuclear export signal, which can act in a cumulative manner.
Figure 4 Analysis of CRM1-dependent nuclear export of amino acid region 74 to 133 of 16.4.1. HeLa cells were transfected with plasmids directing expression of GFP fusion proteins containing 16.4.1 region 74–133 in single copy or in tandem (74–133)2. Representative images of the subcellular distribution of the GFP fusion proteins in untreated (-) and LMB treated cells (+) are shown at the top. Symbols in the graph indicate the nuclear proportion of fluorescence (%) in individual cells and horizontal lines and numbers the median of the cell population. Scale bars: 10 μm. Inhibition of CRM1 by LMB treatment increases nuclear proportion of GFP fusion proteins containing aa residues 74 to 133 of 16.4.1. GFP fusion proteins containing two copies of region 74 to 133 show stronger cytoplasmic localization than GFP fusion proteins with a single copy. These results indicate that region 74 to 133 of 16.4.1 is a substrate for CRM1-dependent export, which is recognized more efficiently in tandem than as a single copy.
Examination of the hypothetical amino acid sequence of region 74 to 133 revealed a clustering of leucine and isoleucine residues between amino acid 86 and 105 (Fig. 5A, shaded in grey). To analyse whether region 86 to 105 of the 16.4.1 protein functions as a nuclear export signal, we compared its translocation capacities with the Rev-NES in a previously described microinjection assay [51] (Fig. 5B). In this assay, peptides bearing the candidate transport sequences are linked to fluorescently labeled bovine serum albumin (BSA). These potential transport substrates are coinjected into the nucleus with unlinked BSA labeled with a different fluorescent color that serves as injection control. Two hours later, cells are fixed and the percentage of each fluorescent label in the nuclear compartment of individual cells determined. The relative translocation activity signifies the ratio of fluorescence of the transport substrate to the fluorescence of the injection control. Selective export of the transport substrate from the nucleus yields relative translocation activities < 1, as demonstrated for a transport substrate containing the NES of Rev (Fig. 5B and [51]). A substrate containing the 16.4.1-derived sequence also yielded a relative translocation activity < 1 (Fig. 5B). These results indicate that region 86 to 105 of 16.4.1 sequence can function as a nuclear export signal.
Figure 5 Functional analysis of a nuclear export signal in 16.4.1. (A) Depicted is the sequence (nucleotides and predicted amino acids) of the 16.4.1 protein investigated here. The sequence encoding amino acids 1–8 are derived from the fetal heart cDNA W67699. The region between amino acid residue 74 and 133 showing CRM1-dependent nuclear export activity (Fig. 4) is underscored. The amino acid sequence between residues 86 and 105 (shaded in grey) contains several Leucine and Isoleucine residues representing a candidate nuclear export signal. (B) Functional characterisation of the Leucine-Isoleucine rich sequence of 16.4.1. Comparison of the translocation activities of 16.4.1 region 86–105 and the Rev-NES in a microinjection-based transport assay [51]. Transport substrates were generated by conjugating peptides containing region 86–105 of 16.4.1 or region 73–84 of Rev with bovine serum albumin (BSA) labeled with a red fluorescent dye. Transport substrates were coinjected into the nucleus with an injection control consisting of unconjugated BSA labeled with a different fluorescent dye (e.g. green). The proportion of each fluorescent label in the nucleus of the injected cell was determined and the ratio of fluorescence of the transport substrate to fluorescence of the injection control calculated. This ratio represents the relative translocation activity of the transport substrate and is indicated in the graph. Nuclear export activity yields ratios <1 as demonstrated for the Rev NES. Transport substrates containing amino acid region 86–105 of 16.4.1 also yielded a relative translocation activity <1, indicating that this region of 16.4.1 can function as a nuclear export signal.
To further characterize this nuclear export signal in 16.4.1 we took advantage of a collection of weight matrices (M1-M7) derived for recognition of NES by bioinformatics (Blossom similarity matrix). These matrices recognized 48 out of 75 signals of a published NES database [36] at a default threshold of 0.84 in the context of their native proteins. No match was obtained upon scanning of the 16.4.1 amino acid sequence with these matrices at default threshold. This indicates that the 16.4.1 sequence is distinct from the 48 NES represented by the matrices. However, rescanning of the 16.4.1 sequence at a lower threshold (0.74) yielded a single match for matrix M5 (0.78), comprising amino acids 92–99 of 16.4.1 (core NES). At default threshold the same matrix recognized a specific group of NES that includes the NES of Stat1 and p65RelA (Fig. 6A). However this matrix did not recognize the NES of PKIα or Rev, which were recognized by different matrices. An artificial 16.4.1 NES sequence containing leucine instead of isoleucine residues at positions 99 and 101 was recognized by matrix M5 above default score (0.86) but by no other matrices, even at reduced thresholds.
Figure 6 Influence of the nuclear export signal on cytoplasmic localization of 16.4.1. (A) Sequences identified in 16.4.1 and in proteins contained in a database of experimentally verified NES by a common weight matrix. Protein sequences in NESbase version 1 [36] were analysed with a collection of 7 weight matrices. These identified 48 of 75 NES in their native protein context at default threshold (0.84). Analysis of the 16.4.1 sequence with reduced threshold settings (0.74) yielded a single match with matrix M5 but not with any other matrix. This match identified the 16.4.1 core NES (shown at top). Sequences in proteins of the NES database identified by M5 at default threshold are shown below. All sequence matches map to experimentally verified NES. (B) To assess functionality of the novel 16.4.1 NES in the context of the full length protein, a 16.4.1 mutant was generated in which L92, I97 and I99 were replaced by alanine residues. HeLa cells were transfected with plasmids for expression of GFP fusion proteins containing wild type (16.4.1-GFP) or mutated (16.4.1(NESmut)-GFP) 16.4.1 and subcellular distribution of GFP fusion proteins were evaluated by quantitative fluorescence microscopy. Cells expressing 16.4.1(NESmut)-GFP showed a significantly higher nuclear proportion of fluorescence (34%) than cells expressing wild type 16.4.1-GFP (27%). This effect was significant (p < 0.005). Scale bars: 10 μm.
Finally we investigated whether the candidate transport signal also shows nuclear export activity in the context of the complete 16.4.1 protein. As shown in figure 6B, the leucine and two isoleucine residues of the 16.4.1 core NES were changed to Alanin and the subcellular distribution of the 16.4.1(NESmut)-GFP was compared to the wildtype 16.4.1 fused to GFP. The mutant 16.4.1-GFP fusion protein localized to significantly higher levels in the nucleus than wildtype 16.4.1-GFP (34% versus 27%). However, the nuclear proportion of the mutant 16.4.1-GFP remained below that of unfused GFP (Fig. 3B), indicating residual nuclear export of the mutant 16.4.1-GFP.
In summary, combined computational and functional analyses indicate that amino acid residues 86 to 105 act as a nuclear export signal, with amino acids 92 to 99 constituting a potential core NES. Mutational analysis indicates that the leucine/isoleucine of the 16.4.1 core NES contribute to but are not sole determinants of cytoplasmic localization of 16.4.1.
Colocalization of 16.4.1 and Rev
This report demonstrates interaction of 16.4.1 and Rev in yeast and mammalian two-hybrid assays (Figs. 1 and 2). In these approaches, candidate interaction partners are artificially targeted to the nucleus to measure interaction-dependent reporter gene expression.
To analyse whether 16.4.1 and Rev interact under conditions in which they retain their natural localization behavior, we analysed cells coexpressing 16.4.1 and Rev for colocalization of both proteins.
To this end, we first established a HeLa cell line stably expressing 16.4.1-GFP and a corresponding control cell line expressing unfused GFP. The expression of 16.4.1-GFP for more than 20 passages did not affect cell growth monitored by measurement of growth curves and did not lead to cell toxicity detectable as release of lactate dehydrogenase (LDH) or ATP into cell culture supernatants (data not shown). Furthermore, long-term expression did not alter the predominantly cytoplasmic localization of 16.4.1-GFP.
For colocalization studies, HeLa 16.4.1-GFP cells and control HeLa-GFP cells were transfected with a plasmid directing expression of Rev-CFP (cyan fluorescent protein) fusion proteins. Transfected cells were subjected to epifluorescence microscopy and Z-stacks were collected. Images were processed by deconvolution and multichannel unmixing, allowing separate evaluation of the spatial distribution of GFP and CFP signals. Over 25 cells were analyzed. Multichannel unmixing is a recently developed technique for separate detection of fluorochromes that exhibit significant spectral overlap in conventional fluorescence microscopy setups, such as CFP and GFP (for a review see [52]). Fig. 7A shows examples of cells expressing 16.4.1-GFP either alone (white arrow) or together with Rev-CFP (red arrow). 16.4.1-GFP was only visible in the nucleoli of cells co-expressing Rev-CFP but not in cells lacking Rev-CFP (see also Fig. 3). Cells coexpressing 16.4.1-GFP and Rev-CFP showed stronger nucleoplasmic GFP fluorescence than HeLa 16.4.1-GFP cells lacking Rev-CFP. Rev-CFP retained typical nuclear/nucleolar localization [49] when coexpressed with 16.4.1-GFP, indicating that 16.4.1-GFP does not influence localization of Rev-CFP. Control imaging of HeLa cells expressing GFP either alone (Fig. 7B, top panel) or together with Rev-CFP (Fig. 7B, bottom panel) showed that presence of Rev-CFP did not influence the GFP signal and that the CFP signal was apparent only in cells expressing Rev-CFP. These results verified separation of Rev-CFP and GFP signals by the multichannel unmixing routine and confirmed that the CFP-tag in Rev-CFP does not affect localization of GFP.
Figure 7 Nucleolar colocalization of 16.4.1-GFP and Rev-CFP. HeLa 16.4.1-GFP cells and control HeLa GFP cells were transiently transfected with a Rev-CFP expression plasmid. Nuclei were counterstained with Hoechst 33342. 3D-Maximum image projections of Z-stacks were deconvolved and processed by multichannel unmixing, applying identical parameters to all images (see Material and Methods). Left images represent the GFP-channel (green), middle images the CFP-channel (pseudocolored red) and right images either an interference contrast photo (A) or a single z-slice of a phase contrast image (B). (A) Exemplary images of over 25 analyzed cells demonstrating nucleolar colocalization of 16.4.1-GFP and Rev-CFP. The left image shows HeLa-16.4.1-GFP cells either without Rev-CFP (white arrow), or with concurrent Rev-CFP expression (red arrow). The nucleolar 16.4.1-GFP signal is only visible in cells coexpressing Rev-CFP. Cells with extremely high expression levels of Rev-CFP were excluded from analysis. Scale bars: 10 μm. (B) Controls for separation of GFP and CFP and signals by multichannel unmixing. Images of HeLa cells stably expressing GFP (HeLa-GFP) either alone or together with Rev-CFP are shown in the upper and lower panels, respectively. The intracellular distribution of the GFP signal is not influenced by coexpression of Rev-CFP. Conversely, a nuclear/nucleolar CFP signal is detected only in HeLa-GFP cells coexpressing Rev-CFP. These results confirm that multichannel unmixing eliminates spectral crosstalk between CFP and GFP channels. Scale bars: 10 μm.
These results indicate that Rev is capable of directing 16.4.1 to nucleoli and provide further evidence for interaction of Rev and 16.4.1 in human cells.
Influence of 16.4.1 on Rev functions
To investigate the influence of 16.4.1 on Rev function, we analysed the effect of IgG1-16.4.1 and 16.4.1-GFP fusion proteins on transactivation capacity of Rev using a previously described Rev-reporter assay [3]. The mRNA synthesized from the reporter gene in this assay contains a region coding for red fluorescent protein (RFP) and a non-coding region with HIV-1 derived sequence elements mediating Rev-responsiveness. These consist of multiple INS from the HIV-1 gag gene and the RRE from the HIV-1 env gene. Rev activity is measured by quantification of RFP reporter positive cells by flow cytometry using the gating strategy depicted in Fig. 8A (for further details see figure legend). Experiments were performed in 293T cells because of the high transfection efficiencies achieved in these cells.
Figure 8 16.4.1 influences transactivation capacity of Rev in human cells. Panel (A) depicts the reduction of Rev activity obtained by cotransfection of 0.5 μg plasmids directing expression of 16.4.1 fusion proteins and 0.1 μg rev expression plasmids. The graph shows mean values and standard deviations of 5 independent experiments carried out in 293T cells expressing either IgG1-16.4.1 or 16.4.1-GFP in combination with Rev-GFP or Rev-CFP. FACS analysis was used to quantify the number of cells expressing the Rev-dependent RFP reporter encoded by pLRed(INS)2R (see Material and Methods) within the population of transfected cells. The dot plots above the bars (FL1 = green fluorescence, FL3 = red fluorescence) represent examples for evaluation of FACS data for fully functional Rev (left), inhibited Rev function by 16.4.1 coexpression (middle) and background expression of the reporter in the absence of Rev (right). The non-transfected, fluorescence negative 293T cell population is represented in grey and excluded from analyses. 293T cells containing the inactive reporter gene (no red fluorescence) and expressing GFP as transfection control were used to define the transfected cell population (R2, right dot plot). Cells transfected with all plasmids required for activation of reporter gene expression (expression plasmids encoding Rev, Tat and the reporter RFP) as well as for expression of the GFP transfection control contain a population of cells expressing both green and red fluorescent proteins defined as R3. The ratio of R3 to R2 represents the proportion of transfected cells showing Rev-dependent reporter expression (i.e. Rev activity). Rev activity in the absence of 16.4.1 was set at 100% (left bar of the graph). Coexpression of 16.4.1 fusion proteins diminished Rev activity to approximately 50% (middle bar). (B) The graph shows the result of a single experiment analysing the effect of different amounts of 16.4.1-GFP expression on Rev-CFP activity. Rev-CFP activity in the absence of 16.4.1-GFP was set at 100%. Coexpression of different amounts of 16.4.1-GFP reduced Rev activity to 35.1% (0.5 μg pC16.4.1sg143) or 12.8% (1 μg pC16.4.1sg143) demonstrating a dose dependent effect of the 16.4.1-GFP on Rev activity.
The transactivation capacity of Rev in the absence of exogenous 16.4.1 was set at 100%. The result of 5 independent experiments demonstrate an approximately 50% reduction of Rev activity by coexpression of 16.4.1 fusion proteins (Fig. 8A). A dose-dependent effect of 16.4.1-GFP expression on Rev activity was observed (Fig. 8B). No effect was observed for numerous other gene products of a human cDNA-library tested in this assay (Wolff et al, manuscript in preparation).
In the experiment above we showed that overexpression of 16.4.1-GFP exhibited a negative effect on the transactivation capacity of HIV-1 Rev in human cells. Isolation of 16.4.1 from a human cDNA library suggests that 16.4.1 proteins may be produced in human cells. To target expression of native 16.4.1 we decided to use RNA interference. To identify inhibitors of 16.4.1 expression we analysed several candidate siRNAs targeted to sequences within the 16.4.1 coding region and a negative control siRNA that recognizes sequences located upstream of the 16.4.1 coding region. An exemplary experiment is shown in Fig. 9A, B. HeLa 16.4.1-GFP cells were transfected with siRNAs and the effects on expression of 16.4.1-GFP monitored by flow cytometry (Fig. 9A and 9B). 16.4.1-GFP expression levels in RNAi transfected cells were determined relative to those in untransfected cells in 40.000 cells by FACS analysis. siRNA-16.4.1 reduced mean relative expression levels of 16.4.1-GFP to 36%. A similar effect was observed for a positive control siRNA that silences GFP (data not shown). The negative control siRNA (siRNA-nsp) only moderately diminished mean relative expression of 16.4.1-GFP to 81%. A similarly moderate reduction was observed for mock-transfected cells (data not shown) indicating that this is caused by the RNA transfection procedure. Analysis of the inhibitory effect of siRNA-16.4.1 on 16.4.1-GFP expression in three additional experiments yielded a mean relative expression of 16.4.1-GFP of 30.7% + 4.7 (standard deviation), confirming the inhibitory effect of this siRNA on 16.4.1-GFP.
Figure 9 Influences of endogenous 16.4.1 proteins on transactivation capacity of Rev in human cells. (A and B) Downregulation of 16.4.1-GFP expression in HeLa cells by siRNAs. HeLa cells stably expressing 16.4.1-GFP were transfected with a pre-synthesized siRNA directed against a 16.4.1 specific sequence (siRNA-16.4.1) or a non-specific siRNA (siRNA-nsp). Expression of 16.4.1-GFP was quantified by flow cytometry in 40.000 cells. Panel A shows an overlay of GFP-expressing cell populations in untransfected cells (green curve) and after transfection with the 16.4.1-specific siRNA (blue) and the non-specific (red). 16.4.1-GFP expression in untransfected cells was set at 100% and the percentage of down regulation of green fluorescence indicated. siRNA-16.4.1 reduced expression levels of GFP-tagged 16.4.1 to 36% which was similar to the silencing effect of si-GFP which was analyzed as control (not shown). Higher levels of 16.4.1-GFP expression (81%) were observed in cells transfected with the unspecific siRNA-nsp. (C) Enhancement of Rev transactivation capacity by siRNAs targeted against 16.4.1. 293T cells were transfected with siRNA-16.4.1 or siRNA-nsp and Rev activities compared with mock-transfected cells (i.e. cells treated with the RNAi transfection reagent lacking RNA). Transfection of siRNA-16.4.1 increased Rev activity by 17% whereas the unspecific siRNA-nsp increased Rev activity by only 1%.
Subsequently we investigated the effect of siRNA-16.4.1 and the negative control siRNA-nsp in 293T cells in the Rev-reporter assay described above (Fig. 9C). The negative control siRNA (siRNA-nsp) had no effect on Rev transactivation capacity compared to mock transfected cells (1% increase of Rev activity). In contrast, siRNA-16.4.1 increased Rev transactivation capacity by 17% compared to mock-transfected controls. A specific enhancing effect of siRNA-16.4.1 was observed in three independent experiments. These results indicate that endogenous 16.4.1 gene products are capable of modulating Rev activity.
Expression of 16.4.1 proteins
Database searches identified several cDNAs of various lengths that contain the complete 16.4.1 sequence within a predicted open reading frame [For overview see additional file 1: Figure A1 and additional file 2: Table A1]. These are derived from various human tissues and cells. The predicted molecular masses of the hypothetical proteins encoded by these cDNAs range from ~145 kDa to 18.5 kDa. This suggests existence of several human 16.4.1 protein species encoded by various cDNAs, rather than a single 16.4.1 protein generated from a single cDNA.
Production of 16.4.1 proteins has not been reported so far and is currently under investigation in our laboratory (Kramer-Hämmerle et al., in preparation). In the context of this ongoing study, a monoclonal antibody (Mab) was generated against recombinant 16.4.1. Indirect immunofluorescence of HeLa cells transfected with the IgG1-16.4.1 expression plasmid revealed a cytoplasmic staining pattern that was indistinguishable from that obtained with antibodies against IgG1 [see additional file 3: Figure A2, A) a, for image]. The 16.4.1 Mab, but not the secondary antibody, also stained untransfected HeLa cells, yielding a cytoplasmic, granular pattern [see additional file 3, Figure A2, A) b for image]. In Western Blot analysis of HeLa-16.4.1-GFP cells, the 16.4.1 Mab recognized a band with the predicted molecular mass of ~45 kDa for the 16.4.1-GFP fusion protein as well as additional proteins [see additional file 3, Figure A2, B]. These results confirm specific recognition of 16.4.1 antigens by the 16.4.1 Mab and indicate expression of endogenous 16.4.1 proteins.
The 16.4.1 Mab has also been used to analyze cells from different lines and primary tissues, including brain and peripheral blood mononuclear cells. A staining pattern similar to that in HeLa cells was observed for 4 out 5 cell lines analyzed by indirect immunofluorescence (not shown). Western blot pagesanalysis of the cell lines/tissues investigated so far yielded a total of 4 distinct bands, ranging in size from > 150 kDa to < 30 kDa [for summary see additional file 3, Table A2]. The occurrence of these bands depended on the cell line/tissue investigated. These results confirm expression of 16.4.1 proteins in human cells and tissues and suggest cell-specific expression patterns of 16.4.1 proteins.
Future studies are directed at identifying the full range of 16.4.1 protein species with a panel of antibodies and characterizing 16.4.1 expression patterns on cDNA and protein levels in various cell types.
Discussion
In this study we identified a cDNA encoding a novel cellular gene product (16.4.1) that interacts with the HIV-1 Rev protein in yeast and mammalian cells. In human cells, 16.4.1 is a substrate for CRM1-dependent export and shows predominant cytoplasmic localization. Colocalization of Rev and 16.4.1 was observed in nuclei, particularly in the nucleoli, of cells expressing both proteins. Overexpression and RNAi experiments indicate that 16.4.1 can influence transactivation function of Rev.
Comparison of cytoplasmic localization properties of 16.4.1, Rev and PKI
We demonstrate that 16.4.1 interacts with CRM1 (Fig. 2) and shows similar Leptomycin B sensitive, cytoplasmic localization behaviour as PKIα and the carboxyterminal half of Rev (Fig. 3B) which are known substrates for CRM1-dependent export [53]. These results indicate that cytoplasmic localization of 16.4.1 involves CRM1-mediated nuclear export.
A nuclear export signal was mapped in 16.4.1 (Fig 5). Mutation of the Leucine/Isoleucine residues of the 16.4.1 NES only partially inhibited nuclear export of full-length 16.4.1, whereas leucine/isoleucine residues in the NES have been shown to be essential for export of Rev and PKIα [35,53]. This suggests differences in the export functions of the 16.4.1 NES and the NES of Rev and PKIα. This conclusion is supported further by the bioinformatics analysis, which showed that the group of NES sequences recognized by the same matrix as the 16.4.1 transport signal did not include the NES of Rev or PKIα (Fig. 6).
GFP fusion proteins containing a single copy of the 16.4.1 NES showed weaker cytoplasmic localization than GFP fusion proteins with tandem copies of this region or full-length 16.4.1-GFP (Fig. 4). This suggests that cytoplasmic localization of 16.4.1 does not depend solely on the functionality of a single copy of the 16.4.1 NES. The formation of homo-oligomers of 16.4.1, as shown by mammalian two-hybrid analysis (Fig. 2), could allow cooperative activity of multiple 16.4.1 NES. In addition, sequences beyond the NES could also contribute to cytoplasmic localization, for example by increasing cytoplasmic retention of 16.4.1. Sequences beyond the NES of 16.4.1 could also promote interactions with export-enhancing co-factors, several of which have been identified so far. These include the Ran-binding protein 3 (RanBP3) [54,55], NXT1 [56] and eukaryotic initiation factor 5A (eIF-5A). eIF-5A was demonstrated to be involved in export of Rev-like NES but not of the PKIα-NES [30,57], suggesting the existence of substrate-specific export cofactors. Future studies will be directed at identifying cellular interaction partners of the 16.4.1 protein and investigating their influence on its export activity.
Interactions of 16.4.1 and Rev
In this study we show that 16.4.1 and Rev are capable of influencing biological properties of one another.
In cells expressing 16.4.1 and Rev, Rev can alter localization properties of 16.4.1 by recruiting 16.4.1 to the nucleus, in particular nucleoli. This is shown by colocalization of Rev and 16.4.1 in the nucleoli of cells expressing both proteins (Fig. 7). Cytoplasmic localization of 16.4.1 suggests that 16.4.1 interacts with Rev in the cytoplasm and is then translocated together with Rev to the nucleus and to nucleoli. The region of 16.4.1 that mediates interaction with Rev (amino acids 39–133, Fig. 1) contains the 16.4.1 NES (amino acids 86–105). Thus CRM1 could "bridge" interaction of 16.4.1 with Rev. CRM1-mediated interaction with Rev has been observed for several cellular proteins proposed to function as cofactors for nuclear export of Rev (see Background). However, amino acid positions of Rev essential for interaction with 16.4.1 are located outside the Rev NES, and an export-deficient NES-mutant of Rev (RevM10Bl) was capable of interacting with 16.4.1 (Fig. 1). This suggests that 16.4.1 does not function as an essential cofactor for nuclear export of Rev.
We have demonstrated that overexpression of 16.4.1 inhibits transactivation function of Rev (Fig. 8). The molecular mechanism underlying this inhibitory effect is unclear. A possible model to explain an inhibitory effect of 16.4.1 on Rev activity is that 16.4.1 recruited to nucleoli by Rev promotes association of Rev and CRM1 in inactive complexes. The strong interaction of 16.4.1 with CRM1 may increase the amount of CRM1 associated with Rev to inhibitory levels. In support of this model, experimental evidence has been obtained demonstrating that (1) Rev associates with CRM1 in nucleoli, influencing its mobility, (2) high levels of CRM1 inhibit Rev activity and (3) Rev is capable of recruiting other CRM1-interacting factors to nucleoli that are capable of inhibiting Rev activity [26,58,59]. This model will be investigated in future experiments.
The RNAi experiments suggest that endogenously expressed 16.4.1 gene products can also affect Rev function. As expected, the stimulatory effect of RNAi-mediated inhibition of 16.4.1 expression was small, since Rev is known to function efficiently in 293T cells [60]. We attempted to study the long-term effect of inhibition of endogenous 16.4.1 on Rev function by establishing cell lines stably expressing siRNA against 16.4.1. However, this approach was not feasible because of cell death after 2–3 weeks of expression of 16.4.1 siRNAs. This indicates that 16.4.1 gene products are crucial for cell viability. On the other hand, overexpression of 16.4.1 is well tolerated as demonstrated by the establishment of a cell line stably expressing 16.4.1-GFP.
The physiological role of interaction of Rev with 16.4.1 is not clear yet and may be positive or negative, depending on the levels of expression of 16.4.1. At low levels 16.4.1 proteins may act as a molecular chaperones of Rev, counteracting the strong tendency of Rev to aggregate with itself and/or preventing incorrect interactions with other cellular proteins. The occurrence of cytoplasmic cellular factors that inhibit Rev multimerization is suggested by a recent report demonstrating only weak Rev-Rev interaction in the cytoplasm of living cells [61]. At high concentrations, 16.4.1 may decrease transactivation function of Rev, for example by sequestering Rev in inactive complexes in nucleoli. Inactivation of Rev by 16.4.1 could play a role in protecting the cells from Rev-mediated cytotoxicity [62].
Conclusion
HIV-1 infection of human cells involves various interactions between cellular and viral factors (reviewed in [63]). Some cell types (e.g. astrocytes) can control HIV-1 replication demonstrating the impact of cellular factors on HIV infection [60,64]. Identification of cellular factors that are able to interfere with viral replication will contribute to understanding of cellular defence mechanisms against viral intruders and may also lead to identification of new targets for therapeutic approaches for virus restriction. Using HIV-1 Rev as "bait" we were able to identify a previously undescribed cellular interaction partner of an HIV-1 protein, 16.4.1. The 16.4.1 protein is exported from the nucleus by CRM1 and accumulates in the cytoplasm. An important feature of 16.4.1 is its ability to impair transactivation capacity of Rev, although both proteins localize to different cellular compartments. Conversely, Rev is capable of affecting localization of the 16.4.1 protein by recruiting it to the nuclei/nucleoli of eukaryotic cells. Because of its properties we suggest naming the 16.4.1 protein "Risp" (Rev interacting shuttling protein).
Data base analyses and preliminary studies with a specific monoclonal antibody suggest that human proteins with Risp sequences are expressed in various human cell types including HIV-1 target cells. The objective of future studies will be to characterize Risp proteins, their cellular functions and their influence on Rev activity and HIV-1 replication in different HIV-1 target cells.
This study represents a further step toward elucidating the network of host cell factors that interact with the HIV-1 Rev protein and influence its functions. This study also illustrates the power of viral proteins as tools for identification and biological characterization of novel cellular factors. Use of similar experimental strategies as presented here will help to gain deeper understanding of virus-cell interactions.
Methods
Plasmids
The inserts of all plasmid constructs used in this study were verified by sequence analysis (Sequiserve, Germany).
Expression plasmids for yeast two-hybrid analysis
pEG202 [65] was used to express bait proteins containing various Rev sequences fused to the LexA DNA binding domain in yeast. pEG202 also expresses the yeast selection marker His3.
For construction of pEG202-sRev, pBsRev [64,66] was used as template for PCR amplification to generate the rev sequence of HIV-1 isolate HXB2; the PCR product was inserted into the EcoRI site of pEG202. The same procedure was used for construction of bait plasmids pEG202-RevM4, pEG202-RevM10BL, pEG202-RevM5, pEG202-RevSLT40, using pcTat-RevM4 [46], pCsRevM10BLsg143 [60], pcRevM5 [45] and pcRevSLT40 [46] as PCR templates. Bait plasmids used as controls for unspecific interaction contained the wildtype rev sequence in antisense orientation (pEG202-sRev(antisense)) or encode a bait protein unrelated to Rev (pEG202-LexCD2).
pJG4-5 and its derivative pJG4-6 was used for galactose-inducible expression of prey proteins containing the protein of interest fused to the NLS of SV40 T antigen, and the B42 transcriptional activator [65]. pJG4-5/6 also express the yeast selection marker Trp1.
pSH18-34 reporter plasmid contains 8 LexA operators that direct expression of the lacZ reporter gene.
The expression library used for two-hybrid screening contained cDNAs from human Jurkat T-cells inserted into the EcoRI/XhoI sites of pJG4-5 [67].
For mapping of Rev-interacting regions of 16.4.1, sequences encoding full-length 16.4.1 or various fragments of 16.4.1 were generated by PCR amplification using pC16.4.1sg143 as template and primers adding a 5' MluI site and a 3' NotI site. PCR products were inserted into pJG4-6 cleaved with MluI and NotI.
Plasmids pJG4-5, pJG4-6, pEG202, pSH18-34 and the Jurkat T-cell cDNA expression library were kindly provided by Waldemar Kolanus, University of Bonn, Germany.
Expression plasmids for mammalian two-hybrid analysis
Protein interactions in human cells were analysed with the CheckMate™ Mammalian Two-Hybrid System (Promega, Madison, USA), which uses pACT and pBIND vectors and the G5luc reporter plasmid. pACT and pBIND direct expression of fusion proteins containing the transcriptional activation domain of Herpes virus simplex VP-16 (pACT) or the Gal4 DNA binding domain (pBIND) at the N-terminus and potential interactor domains at the C-terminus. pG5luc contains 5 Gal4-binding motifs and a minimal promoter for inducible expression of the firefly luciferase reporter gene.
pACT and pBIND expression plasmids were constructed by PCR amplification of coding sequences from plasmid templates with primers adding restriction sites for insertion into the multiple cloning regions of the target vectors. The rev sequence was generated from pEG202-sRev and inserted into the SalI site of pACT (pACT-Rev). 16.4.1 sequence was generated from clone DKFZp434O171Q (RZPD, Berlin, Germany;[68]) and inserted into the BamHI sites of pACT and pBIND (pACT-16.4.1 and pBIND-16.4.1). The human CRM1 sequence was amplified from pChCRM1sg143 [49] and inserted into the BamHI site of pACT (pACT-hCRM1).
Plasmids encoding GFP-tagged proteins
The vector pFRED143 contains a humanized version of a strong fluorescent GFP mutant (sg143) under the control of the CMV immediate early promoter [69]. pC-sg143 plasmids were constructed by using the cloning strategy described in [60], involving insertion of protein coding sequences without translational start and stop codons in frame with gfp-sequences into pFRED at a unique NheI site located immediately downstream of codon 1 of the GFP ORF.
The 16.4.1 sequence in pJG4-5 contains a 163 amino acid reading frame which is terminated by a stop codon but lacks an initiation codon. A potential translational initiation codon was identified 24 nucleotides upstream of and in frame with the 16.4.1 sequence in a human fetal cDNA (Gene Bank Accession Nr. W67699). For construction of pC16.4.1sg143, the 16.4.1 sequence in pJG4-5 was amplified with a 5' primer incorporating sequences encoding amino acids 2–7 into the PCR product, which was inserted into the NheI site of pFRED143. Sequence analysis of pC16.4.1sg143 verified formation of a single open reading frame by 16.4.1 and GFP sequences, resulting in expression of a protein with a predicted molecular mass of approximately 46 kDa.
For construction of pC16.4.1(74–133)sg143 the sequence encoding amino acids 74 to 133 was generated by PCR from pC16.4.1sg143 and inserted into the NheI site of pFRED143. pC16.4.1(74–133)2sg143 contains tandem sequences encoding amino acids 74 to 133 of 16.4.1.
The plasmid pC16.4.1(NESmut)sg143 expresses a mutant 16.4.1-GFP fusion protein in which L92, I97 and I99 within the core nuclear export of 16.4.1 are replaced by alanine residues. pC16.41(NESmut)sg143 was constructed by PCR based site directed mutagenesis from pC16.4.1sg143.
For construction of pCRev(52–116)sg143, sequences encoding amino acids 52–116 of Rev were amplified from pCsRevsg143 [60] and inserted into the BspEI site of the pFRED143 variant pFRED143BspEI.
pCPKIα sg143 directs expression of GFP-tagged human PKI and was constructed as described in [49].
Plasmids encoding IgG1-tagged proteins
pIg was kindly provided by Waldemar Kolanus and was used for construction of plasmids directing expression of 16.4.1 fusion proteins with an N-terminal IgG1 tag. pIg contains CH2 and CH3 domain segments from human IgG1 cDNA in the mammalian expression vector pRK5 [67,70]. Sequences encoding 16.4.1 or various regions of 16.4.1 were amplified from pC16.4.1sg143. PCR products were inserted into the MluI/NotI sites of the pIg polylinker, resulting in construction of the following plasmids: pIgG1-16.4.1, pIgG1-16.4.1(2–38), pIgG1-16.4.1(2–133), pIgG1-16.4.1(38–171), pIgG1-16.4.1(74–171) and pIgG1-16.4.1(134–171).
Plasmids used in Rev activity assay
pLRed(INS)2R reporter plasmid was constructed in a similar manner as described for pLRed(p17/p24INS)R [3]. Briefly, a DNA fragment with HIV-1 gag-sequences containing INS 1 and 2 (nucleotides 379–1424 of the HXB2 genome) was isolated from pB37R [71] by ClaI-digestion and inserted into ClaI-digested and dephosphorylated pLRedR [3]. pLRed(p17/p24INS)R contains two copies of the HIV-1 gag sequences in sense orientation. This plasmid directs Rev and Tat-dependent expression of red fluorescent protein (RFP).
pL3Tat contains the HIV-1 tat gene under the control of the HIV-1 LTR [64].
pCsRev-CFP was established by replacement of the GFP encoding sequence of pCsRevsg143 mentioned above with the coding sequence for cyan fluorescent protein (CFP).
Yeast two-hybrid assay
The yeast interaction trap was performed essentially as described in [65,67], using yeast strain EGY48 which contains the LEU2 gene under the control of LexA operators and has defective leu2, his3, trp1 and ura3 genes. The pEG202-sRev bait plasmid was transformed into yeast strain EGY48 bearing the pSH18-34 reporter plasmid. This "bait" strain was transformed with the Jurkat T-cell library contained in the yeast expression plasmid pJG4-5. Criteria for protein-protein interactions were growth on medium containing galactose and lacking uracil, histidine, tryptophane and leucine, no growth in the same medium containing glucose instead of galactose, and expression of beta-galactosidase. Forty-six yeast transformants were obtained that grew on selective medium and expressed the lacZ reporter gene. cDNAs from those yeast clones were purified by passaging through E. coli KC8 and retested for interaction with pEG202-sRev but not with control bait plasmids pEG202-sRev(antisense) and pEG202-LexCD2 to confirm specific interaction.
Mammalian two-hybrid assay
Mammalian two-hybrid assay was performed in HEK293 cells, using the CheckMate™ Mammalian Two-Hybrid System (Promega, Madison, USA).
HEK293 cells were cotransfected with pBIND and pACT constructs for expression of VP16 and Gal4 proteins fused to potential interactor domains (1 μg each) and with the pG5luc reporter plasmid (2 μg). For each interactor assay, parallel transfections were performed with G5luc and pBIND and pACT vectors expressing Vp16 and Gal4 without interactor domains to determine background expression of the luciferase gene. Two days after transfection cells were lysed, and firefly luciferase activity (in relative light units per second, RLU/s) quantified using the Luciferase Reporter Assay System (PROMEGA) and the ORION I Microplate Luminometer (Berthold Detection Systems, Pforzheim, Germany). The total amount of protein in cell lysates was quantified using the BCA Protein Assay Reagent Kit (Pierce Chemical Co., Rockford, USA) and luciferase activity (RLU/s) standardized to 1mg of total protein in the cell lysate. Values are expressed as fold-induction of luciferase activity over basal expression levels.
Cell culture, transfection and Leptomycin B treatment
HeLa and HEK293 cells were maintained in Dulbecco's Modified Eagle Medium containing 2 nM Glutamax I (Life Technologies, Karlsruhe, Germany) and 10% fetal calf serum (Seromed, Berlin, Germany).
All transfection experiments were performed in 35-mm-diameter dishes (BD Biosciences, Bedford, MA, USA). Cells were seeded at a density of 1 × 105 cells per dish one day prior to transfection and cultured for 24 h after transfection. HEK293 cells were transfected by calcium phosphate coprecipitation using the CellPhect kit (Pharmacia, Freiburg, Germany). Transfection of HeLa cells was performed with the FuGENE™6 Transfection Reagent (Roche Diagnostics, Mannheim, Germany) using 500 ng plasmid DNA per dish.
Leptomycin B (LMB, Sigma-Aldrich, Munich, Germany) treatments were performed 24 hours after transfection at a concentration of 5 nM LMB for 2 hours. For microinjection experiments, LMB was added at a concentration of 10 nM 2 hours prior to injection.
For evaluation and quantification of fluorescence, cells were fixed with 4% paraformaldehyde (PFA) for 30 minutes at room temperature; nuclei were stained with Hoechst 33343 (Molecular Probes Europe BV, Leiden, Netherlands) for 10 minutes.
Toxic influences of long-term expression of 16.4.1-GFP in HeLa cells were assessed by CytoTox-ONE™ and CellTiter-Glo™ cell viability assays (Promega, Madison, USA) according to manufacturer's instructions.
The cell line HeLa 16.4.1-GFP expresses 16.4.1-GFP constitutively and was established by transfection with pC16.4.1sg143 followed by G418-selection (80 μg/ml). Non-fluorescent antibiotic resistant cells were excluded by FACS sorting.
Immunological methods
Immunocytochemistry
Cells were fixed with PFA (4%) and permeabilized with TritonX-100 (0.2%).
IgG1-16.4.1 fusion proteins were detected by direct staining with a Cy3-conjugated goat anti-human IgG1 antibodies (Dianova, Hamburg, Germany), diluted 1:200 in phosphate buffered saline containing 1% BSA. For detection of 16.4.1 antigens, a monoclonal antibody against 16.4.1 (Kramer-Hämmerle et al. in preparation) was used as primary antibody and a Cy3-conjugated goat anti-rat antibody (Dianova, Hamburg, Germany) as secondary antibody.
Western Blot
Whole cell lysates were prepared with RIPA buffer containing protease inhibitors (protease inhibitor cocktail; Sigma, Taufkirchen, Germany) and separated on either precast 4–12% Bis-Tris or 3–8% Tris-Acetate gradient gels (Invitrogen, Karlsruhe, Germany). After transfer onto nitrocellulose membranes, proteins were probed with primary polyclonal rabbit antibodies against GFP (Chemicon Europe, Hofheim, Germany) or with the 16.4.1-specific monoclonal antibody and with HRP-conjugated secondary goat anti-rabbit or anti-rat antibodies (Dianova, Hamburg, Germany). Protein bands were detected by an enhanced chemiluminescence system (Perbio Science, Bonn, Germany).
Quantitative fluorescence microscopy
Microscopy of cells expressing fluorescent proteins and quantitative analysis of subcellular distribution of fluorescence was performed as described [49]. Images for quantification were taken at 32-fold magnification with flexible exposure times and evaluated by IPlab software for fluorescence values below pixel saturation. Each cell group was photographed as phase contrast and fluorescence images for Hoechst 33342 and GFP.
3D deconvolution and widefield multichannel unmixing microscopy
Fluorescence microscopic imaging, 3D deconvolution and widefield multichannel unmixing was carried out with a computer-controlled Zeiss Axiovert 200 M research microscope ("Cell Observer") with scanning stage (Carl Zeiss, Goettingen, Germany) and Software AxioVision 4.2 (Carl Zeiss Vision, Hallbergmoos, Germany). Images of 2% PFA-fixed specimen were acquired using a Zeiss 40×/1.3 Plan Neofluar objective and Zeiss filter sets No. 44 (GFP), 49 (Hoechst33342) and 47 (CFP). Z-stacks with 100 optical sections at 0.325 μm intervals were captured with a Zeiss AxioCam HRm CCD-Camera with full resolution of 1388 × 1044 pixels. Deconvolution of fluorescence images [72] was carried out with AxioVision 4.2 software using a constrained iterative algorithm and auto linear normalization. Subsequently widefield multichannel unmixing was performed on the deconvolved image-stacks to correct for fluorescence bleedthrough of Hoechst-, CFP- and GFP- signals. Three reference samples with either one of the three fluorochromes were prepared, reference measurements were performed and a 3 × 3 matrix was generated that was used to unmix the sample image-stack. Processed images were then arranged for presentation and exported with AxioVision 4.2 software.
Microinjection experiments
Compounds for microinjection were generated and microinjections were performed as described previously [49,51]. Briefly, bovine serum albumin (BSA) was first labeled with Alexa-red (excitation 568 nm, Molecular probes) and subsequently conjugated to the following peptides:
(1.) HIV-1 Rev-NES represented by -CGG-LQPPLERLTLD and (2.) 16.4.1 amino acids 86 to 105 -CG-KTLESNLFDDNIDIFADLTV. Both peptides were synthesized by Sigma-Genosys. For coinjections unconjugated BSA labeled with Alexa-green (excitation 488 nm, Molecular probes) was used. Peptide solutions with a concentration of 1 mg/ml were injected into the nuclear compartment of HeLa cells. Two hours after injection, cells were fixed with 4% PFA, images of green and red channels were taken and the percentages of red (BSA-conjugated peptides) and green (unconjugated BSA) fluorescent signals were determined. The ratio between red and green fluorescence in the nuclear and cytoplasmic compartments indicates the relative translocation activity of the peptide.
Rev activity assay
The Rev activity assay was performed essentially as described previously [3]. Transfections were performed in 6-well plates using FUGENE™6 according to the manufacturer's protocol with 1 μg pLRed(INS)2R, 0.2 μg pL3Tat, 0.1 μg pCsRev-CFP and 0.1 μg pFRED143. To evaluate the influence of 16.4.1-GFP on Rev-activity, pC16.4.1sg143 (0.5 μg or 1 μg) was added to the transfection mixture. Expression of fluorescent fusion proteins was checked by microscopy 24 hours after transfection and cells were analysed by flow cytometry. Typically 100,000 cells of each transfected well were analysed with a Becton Dickinson FACSCalibur flow cytometer equipped with a 488 nm argon laser and controlled by the software CellQuest Pro. GFP-fluorescence was analysed in channel FL-1 and RFP-fluorescence in FL-3. The percentage of reporter positive cells (RFP) in the transfected cell population (GFP) was determined.
RNAi interference in combination with Rev activity assay
For down-regulation of gene expression, 16.4.1 specific (siRNA-16.4.1) and non-specific (siRNA-nsp) and a GFP-specific siRNA were designed and synthesized by Qiagen. Transfections were carried out using RNAiFect transfection reagent (Qiagen) according to supplier's protocol. Target cells were seeded one day prior to transfection in 6-well plates, 5 μg siRNA per well were used for each transfection. Silencing effects of GFP-fusion proteins were determined by FACS analyses 48 hours after transfection. For the combined RNAi-Rev activity experiments, siRNA transfection was performed 24 h prior to DNA transfection.
Bioinformatics
In silico identification and analysis of sequences were performed using the databases and bioinformatics tools of NCBI [73] and of Genomatix Software GmbH [74].
Similarities between nuclear export signals in the NES database NESbase 1.0 [75] were analysed with a set of amino acid weight matrices adapted from the MatInspector algorithm [76] using the BLOSSOM similarity matrix values to account for conserved amino acid substitutions. Reading frames were predicted with the tool ATGpr [77].
Statistics
Statistical analysis of data was carried out with the GraphPadPRISM program (GraphPad Software Inc., San Diego, CA, USA). Significances of differences between data sets were determined by calculating two-tailed P values by the Mann-Whitney U test.
Authors' contributions
SKH performed expression and localization studies of Risp; functional analyses of nuclear export of Risp and influences on Rev-activity and was primarily responsible for writing of the manuscript. FCS and CB carried out the yeast and mammalian two-hybrid analyses including design and construction of expression plasmids and final evaluation of the data. HW developed the Rev reporter assay, carried out 3D deconvolution and multichannel unmixing microscopy and participated in analysis of experimental data. MV performed the experiments regarding RNA interference. TW carried out the in silico analysis of the NES signals. VE contributed with substantial discussions and participated on interpretation of the results. RBW was the principal investigator and was responsible for designing the study, supervising the experimental work and for writing the manuscript together with SKH.
Supplementary Material
Additional File 1
Schematic representation of cDNAs with 16.4.1-coding sequences in predicted open reading frames. The scheme shows various cDNAs with 16.4.1 sequences identified by BLAST search of Entrez databases ([73] (indicated by lines). Open reading frames (ORF) were predicted with the ATGpr program [77]. Bars indicate the locations of the predicted ORFs within the cDNAs. The positions of the potential translation initiation and stop codons are marked by M and *, respectively. Regions with 16.4.1 encoding sequences are labeled yellow. All accession numbers are from the GenBank database.
Click here for file
Additional File 2
Sizes of hypothetical 16.4.1 cDNAs and proteins. The table indicates the lengths of the cDNAs shown in Additional file 1: Figure A1 and the calculated sizes of the proteins they are predicted to encode.
Click here for file
Additional File 3
Analysis of expression of 16.4.1 proteins with a monoclonal antibody directed against 16.4.1. Figure A2 shows expression of 16.4.1 proteins in HeLa cells. (A) Immunohistochemical analysis. Expression of 16.4.1 was detected in HeLa cells transfected with pIgG-16.4.1 (panel a) and in non-transfected HeLa cells (panel b) by indirect immunofluorescence with the monoclonal antibody against 1.6.4.1 and Cy3-labelled secondary antibodies. Panel c shows lack of reactivity of non-transfected HeLa cells with the secondary antibody. All images were taken with the same exposure times (300 ms). (B) Western blot analyses. Additional to 16.4.1-GFP several proteins are detected in HeLa cells expressing 16.4.1-GFP after transfection of pC16.4.1sg143. Proteins in whole-cell lysates were separated by electrophoresis in gradient gels (4–12% and 3–8%), and blotted onto nitrocellulose membranes. 16.4.1 proteins were detected with the monoclonal 16.4.1 antibody and a secondary antibody conjugated with horse radish peroxidase (HRP). For detection of the 16.4.1-GFP fusion protein, the membranes were stripped and reprobed with polyclonal antibodies against GFP. Table A2 lists proteins recognized by the monoclonal antibody against 16.4.1 in different human cell lines and primary human tissues by Western blot analyses. HeLa (cervix carcinoma) and 293T (embryonal kidney) cell lines are non-neural cells and U138MG, U251MG and 85HG66 human glioblastoma cell lines. Human cortical brain tissue and peripheral blood mononuclear cells were also investigated.
Click here for file
Acknowledgements
We are very grateful to Waldemar Kolanus, University of Bonn, for providing us with the yeast strain, plasmids and cDNA library for yeast two-hybrid analysis and with the mammalian expression plasmid pIgG1. We thank Alexandra Ludvigsen for her contributions to the yeast two-hybrid screening, Kerstin Lux and Andrea Brebeck for cloning of the M2H constructs and the Rev-CFP expression plasmid and Petra Willnauer for performing microinjections. We are especially grateful to Elisabeth Kremmer for establishment of the rat monoclonal antibody against 16.4.1. We thank Kerstin Cartharius (Genomatix) for adapting the MatInspector algorithm to protein sequences.
This work was partially supported by the Deutsche Forschungsgemeinschaft (F.C.S.; SFB 464).
==== Refs
Freed EO Martin MA Knipe DM and Howley PM HIVs and their replication Fields Virology 2001 2 Philadelphia, Lippincott Williams and Wilkins 1971 2041
Zolotukhin AS Michalowski D Bear J Smulevitch SV Traish AM Peng R Patton J Shatsky IN Felber BK PSF acts through the human immunodeficiency virus type 1 mRNA instability elements to regulate virus expression Mol Cell Biol 2003 23 6618 6630 12944487 10.1128/MCB.23.18.6618-6630.2003
Wolff H Brack-Werner R Neumann M Werner T Schneider R Integrated functional and bioinformatics approach for the identification and experimental verification of RNA signals: application to HIV-1 INS Nucleic Acids Res 2003 31 2839 2851 12771211 10.1093/nar/gkg390
Boris-Lawrie K Roberts TM Hull S Retroviral RNA elements integrate components of post-transcriptional gene expression Life Sci 2001 69 2697 2709 11720075 10.1016/S0024-3205(01)01360-1
Kjems J Askjaer P Rev protein and its cellular partners Adv Pharmacol 2000 48 251 298 10987094
Hauber J Nuclear export mediated by the Rev/Rex class of retroviral Trans-activator proteins Curr Top Microbiol Immunol 2001 259 55 76 11417127
Henderson BR Percipalle P Interactions between HIV Rev and nuclear import and export factors: the Rev nuclear localisation signal mediates specific binding to human importin-beta J Mol Biol 1997 274 693 707 9405152 10.1006/jmbi.1997.1420
Truant R Cullen BR The arginine-rich domains present in human immunodeficiency virus type 1 Tat and Rev function as direct importin beta-dependent nuclear localization signals Mol Cell Biol 1999 19 1210 1217 9891055
Askjaer P Jensen TH Nilsson J Englmeier L Kjems J The specificity of the CRM1-Rev nuclear export signal interaction is mediated by RanGTP J Biol Chem 1998 273 33414 33422 9837918 10.1074/jbc.273.50.33414
Neville M Stutz F Lee L Davis LI Rosbash M The importin-beta family member Crm1p bridges the interaction between Rev and the nuclear pore complex during nuclear export Curr Biol 1997 7 767 775 9368759 10.1016/S0960-9822(06)00335-6
Bogerd HP Echarri A Ross TM Cullen BR Inhibition of human immunodeficiency virus Rev and human T-cell leukemia virus Rex function, but not Mason-Pfizer monkey virus constitutive transport element activity, by a mutant human nucleoporin targeted to Crm1 J Virol 1998 72 8627 8635 9765402
Hakata Y Yamada M Mabuchi N Shida H The carboxy-terminal region of the human immunodeficiency virus type 1 protein Rev has multiple roles in mediating CRM1-related Rev functions J Virol 2002 76 8079 8089 12134013 10.1128/JVI.76.16.8079-8089.2002
Luo Y Yu H Peterlin BM Cellular protein modulates effects of human immunodeficiency virus type 1 Rev J Virol 1994 68 3850 3856 8189522
Tange TO Jensen TH Kjems J In vitro interaction between human immunodeficiency virus type 1 Rev protein and splicing factor ASF/SF2-associated protein, p32 J Biol Chem 1996 271 10066 10072 8626563 10.1074/jbc.271.17.10066
Fankhauser C Izaurralde E Adachi Y Wingfield P Laemmli UK Specific complex of human immunodeficiency virus type 1 rev and nucleolar B23 proteins: dissociation by the Rev response element Mol Cell Biol 1991 11 2567 2575 2017166
Szebeni A Herrera JE Olson MO Interaction of nucleolar protein B23 with peptides related to nuclear localization signals Biochemistry 1995 34 8037 8042 7794916 10.1021/bi00025a009
Zheng YH Yu HF Peterlin BM Human p32 protein relieves a post-transcriptional block to HIV replication in murine cells Nat Cell Biol 2003
Szebeni A Mehrotra B Baumann A Adam SA Wingfield PT Olson MO Nucleolar protein B23 stimulates nuclear import of the HIV-1 Rev protein and NLS-conjugated albumin Biochemistry 1997 36 3941 3949 9092824 10.1021/bi9627931
Szebeni A Olson MO Nucleolar protein B23 has molecular chaperone activities Protein Sci 1999 8 905 912 10211837
Stutz F Izaurralde E Mattaj IW Rosbash M A role for nucleoporin FG repeat domains in export of human immunodeficiency virus type 1 Rev protein and RNA from the nucleus Mol Cell Biol 1996 16 7144 7150 8943370
Stutz F Neville M Rosbash M Identification of a novel nuclear pore-associated protein as a functional target of the HIV-1 Rev protein in yeast Cell 1995 82 495 506 7634338 10.1016/0092-8674(95)90438-7
Kiss A Li L Gettemeier T Venkatesh LK Functional analysis of the interaction of the human immunodeficiency virus type 1 Rev nuclear export signal with its cofactors Virology 2003 314 591 600 14554087 10.1016/S0042-6822(03)00531-2
Bogerd HP Fridell RA Madore S Cullen BR Identification of a novel cellular cofactor for the Rev/Rex class of retroviral regulatory proteins Cell 1995 82 485 494 7634337 10.1016/0092-8674(95)90437-9
Fritz CC Zapp ML Green MR A human nucleoporin-like protein that specifically interacts with HIV Rev Nature 1995 376 530 533 7637788 10.1038/376530a0
Farjot G Sergeant A Mikaelian I A new nucleoporin-like protein interacts with both HIV-1 Rev nuclear export signal and CRM-1 J Biol Chem 1999 274 17309 17317 10358091 10.1074/jbc.274.24.17309
Zolotukhin AS Felber BK Nucleoporins nup98 and nup214 participate in nuclear export of human immunodeficiency virus type 1 Rev J Virol 1999 73 120 127 9847314
Ruhl M Himmelspach M Bahr GM Hammerschmid F Jaksche H Wolff B Aschauer H Farrington GK Probst H Bevec D Eukaryotic initiation factor 5A is a cellular target of the human immunodeficiency virus type 1 Rev activation domain mediating trans-activation J Cell Biol 1993 123 1309 1320 8253832 10.1083/jcb.123.6.1309
Venkatesh LK Gettemeier T Chinnadurai G A nuclear kinesin-like protein interacts with and stimulates the activity of the leucine-rich nuclear export signal of the human immunodeficiency virus type 1 rev protein J Virol 2003 77 7236 7243 12805422 10.1128/JVI.77.13.7236-7243.2003
Rosorius O Reichart B Kratzer F Heger P Dabauvalle MC Hauber J Nuclear pore localization and nucleocytoplasmic transport of eIF-5A: evidence for direct interaction with the export receptor CRM1 J Cell Sci 1999 112 ( Pt 14) 2369 2380 10381392
Hofmann W Reichart B Ewald A Muller E Schmitt I Stauber RH Lottspeich F Jockusch BM Scheer U Hauber J Dabauvalle MC Cofactor requirements for nuclear export of Rev response element (RRE)- and constitutive transport element (CTE)-containing retroviral RNAs. An unexpected role for actin J Cell Biol 2001 152 895 910 11238447 10.1083/jcb.152.5.895
Sanchez-Velar N Udofia EB Yu Z Zapp ML hRIP, a cellular cofactor for Rev function, promotes release of HIV RNAs from the perinuclear region Genes Dev 2004 18 23 34 14701878 10.1101/gad.1149704
Cullen BR Nuclear mRNA export: insights from virology Trends Biochem Sci 2003 28 419 424 12932730 10.1016/S0968-0004(03)00142-7
Ossareh-Nazari B Gwizdek C Dargemont C Protein export from the nucleus Traffic 2001 2 684 689 11576444 10.1034/j.1600-0854.2001.21002.x
Fischer U Huber J Boelens WC Mattaj IW Luhrmann R The HIV-1 Rev activation domain is a nuclear export signal that accesses an export pathway used by specific cellular RNAs Cell 1995 82 475 483 7543368 10.1016/0092-8674(95)90436-0
Wen W Meinkoth JL Tsien RY Taylor SS Identification of a signal for rapid export of proteins from the nucleus Cell 1995 82 463 473 7634336 10.1016/0092-8674(95)90435-2
la Cour T Gupta R Rapacki K Skriver K Poulsen FM Brunak S NESbase version 1.0: a database of nuclear export signals Nucleic Acids Res 2003 31 393 396 12520031 10.1093/nar/gkg101
Wolff B Sanglier JJ Wang Y Leptomycin B is an inhibitor of nuclear export: inhibition of nucleo-cytoplasmic translocation of the human immunodeficiency virus type 1 (HIV-1) Rev protein and Rev-dependent mRNA Chem Biol 1997 4 139 147 9190288 10.1016/S1074-5521(97)90257-X
Kudo N Wolff B Sekimoto T Schreiner EP Yoneda Y Yanagida M Horinouchi S Yoshida M Leptomycin B inhibition of signal-mediated nuclear export by direct binding to CRM1 Exp Cell Res 1998 242 540 547 9683540 10.1006/excr.1998.4136
Fornerod M Ohno M Yoshida M Mattaj IW CRM1 is an export receptor for leucine-rich nuclear export signals [see comments] Cell 1997 90 1051 1060 9323133 10.1016/S0092-8674(00)80371-2
Sakura H Maekawa T Imamoto F Yasuda K Ishii S Two human genes isolated by a novel method encode DNA-binding proteins containing a common region of homology Gene 1988 73 499 507 2977358 10.1016/0378-1119(88)90514-8
Horwitz EM Maloney KA Ley TJ A human protein containing a "cold shock" domain binds specifically to H-DNA upstream from the human gamma-globin genes J Biol Chem 1994 269 14130 14139 8188694
Shen Q Wu R Leonard JL Newburger PE Identification and molecular cloning of a human selenocysteine insertion sequence-binding protein. A bifunctional role for DNA-binding protein B J Biol Chem 1998 273 5443 5446 9488664 10.1074/jbc.273.10.5443
Stenina OI Shaneyfelt KM DiCorleto PE Thrombin induces the release of the Y-box protein dbpB from mRNA: a mechanism of transcriptional activation Proc Natl Acad Sci U S A 2001 98 7277 7282 11390977 10.1073/pnas.121592298
Lloberas J Maki RA Celada A Repression of major histocompatibility complex I-A beta gene expression by dbpA and dbpB (mYB-1) proteins Mol Cell Biol 1995 15 5092 5099 7651426
Malim MH Bohnlein S Hauber J Cullen BR Functional dissection of the HIV-1 Rev trans-activator--derivation of a trans-dominant repressor of Rev function Cell 1989 58 205 214 2752419 10.1016/0092-8674(89)90416-9
Thomas SL Oft M Jaksche H Casari G Heger P Dobrovnik M Bevec D Hauber J Functional analysis of the human immunodeficiency virus type 1 Rev protein oligomerization interface J Virol 1998 72 2935 2944 9525614
Stauber R Gaitanaris GA Pavlakis GN Analysis of trafficking of Rev and transdominant Rev proteins in living cells using green fluorescent protein fusions: transdominant Rev blocks the export of Rev from the nucleus to the cytoplasm Virology 1995 213 439 449 7491768 10.1006/viro.1995.0016
Grapes M O'Hare P Differences in determinants required for complex formation and transactivation in related VP16 proteins J Virol 2000 74 10112 10121 11024140 10.1128/JVI.74.21.10112-10121.2000
Neumann M Afonina E Ceccherini-Silberstein F Schlicht S Erfle V Pavlakis GN Brack-Werner R Nucleocytoplasmic transport in human astrocytes: decreased nuclear uptake of the HIV Rev shuttle protein J Cell Sci 2001 114 1717 1729 11309202
Ribbeck K Gorlich D Kinetic analysis of translocation through nuclear pore complexes Embo J 2001 20 1320 1330 11250898 10.1093/emboj/20.6.1320
Demart S Ceccherini-Silberstein F Schlicht S Walcher S Wolff H Neumann M Erfle V Brack-Werner R Analysis of nuclear targeting activities of transport signals in the human immunodeficiency virus Rev protein Exp Cell Res 2003 291 484 501 14644169 10.1016/j.yexcr.2003.07.011
Zimmermann T Rietdorf J Pepperkok R Spectral imaging and its applications in live cell microscopy FEBS Lett 2003 546 87 92 12829241 10.1016/S0014-5793(03)00521-0
Henderson BR Eleftheriou A A comparison of the activity, sequence specificity, and CRM1-dependence of different nuclear export signals Exp Cell Res 2000 256 213 224 10739668 10.1006/excr.2000.4825
Lindsay ME Holaska JM Welch K Paschal BM Macara IG Ran-binding protein 3 is a cofactor for Crm1-mediated nuclear protein export J Cell Biol 2001 153 1391 1402 11425870 10.1083/jcb.153.7.1391
Nemergut ME Lindsay ME Brownawell AM Macara IG Ran-binding protein 3 links Crm1 to the Ran guanine nucleotide exchange factor J Biol Chem 2002 277 17385 17388 11932251 10.1074/jbc.C100620200
Black BE Holaska JM Levesque L Ossareh-Nazari B Gwizdek C Dargemont C Paschal BM NXT1 is necessary for the terminal step of Crm1-mediated nuclear export J Cell Biol 2001 152 141 155 11149927 10.1083/jcb.152.1.141
Elfgang C Rosorius O Hofer L Jaksche H Hauber J Bevec D Evidence for specific nucleocytoplasmic transport pathways used by leucine-rich nuclear export signals Proc Natl Acad Sci U S A 1999 96 6229 6234 10339570 10.1073/pnas.96.11.6229
Daelemans D Costes SV Lockett S Pavlakis GN Kinetic and molecular analysis of nuclear export factor CRM1 association with its cargo in vivo Mol Cell Biol 2005 25 728 739 15632073 10.1128/MCB.25.2.728-739.2005
Ristea S Dobbelstein M Roth J Rev protein of human immunodeficiency virus type 1 and cellular exportin 1 protein relocalize each other to a subnucleolar structure AIDS Res Hum Retroviruses 2000 16 857 865 10875611 10.1089/08892220050042792
Ludwig E Silberstein FC van Empel J Erfle V Neumann M Brack-Werner R Diminished rev-mediated stimulation of human immunodeficiency virus type 1 protein synthesis is a hallmark of human astrocytes J Virol 1999 73 8279 8289 10482578
Daelemans D Costes SV Cho EH Erwin-Cohen RA Lockett S Pavlakis GN In vivo HIV-1 Rev multimerization in the nucleolus and cytoplasm identified by fluorescence resonance energy transfer J Biol Chem 2004 279 50167 50175 15294891 10.1074/jbc.M407713200
Miyazaki Y Takamatsu T Nosaka T Fujita S Martin TE Hatanaka M The cytotoxicity of human immunodeficiency virus type 1 Rev: implications for its interaction with the nucleolar protein B23 Exp Cell Res 1995 219 93 101 7628555 10.1006/excr.1995.1209
Trkola A HIV-host interactions: vital to the virus and key to its inhibition Curr Opin Microbiol 2004 7 407 411 15358260 10.1016/j.mib.2004.06.002
Neumann M Felber BK Kleinschmidt A Froese B Erfle V Pavlakis GN Brack-Werner R Restriction of human immunodeficiency virus type 1 production in a human astrocytoma cell line is associated with a cellular block in Rev function J Virol 1995 69 2159 2167 7884864
Gyuris J Golemis E Chertkov H Brent R Cdi1, a human G1 and S phase protein phosphatase that associates with Cdk2 Cell 1993 75 791 803 8242750 10.1016/0092-8674(93)90498-F
D'Agostino DM Ciminale V Pavlakis GN Chieco-Bianchi L Intracellular trafficking of the human immunodeficiency virus type 1 Rev protein: involvement of continued rRNA synthesis in nuclear retention AIDS Res Hum Retroviruses 1995 11 1063 1071 8554903
Kolanus W Nagel W Schiller B Zeitlmann L Godar S Stockinger H Seed B Alpha L beta 2 integrin/LFA-1 binding to ICAM-1 induced by cytohesin-1, a cytoplasmic regulatory molecule Cell 1996 86 233 242 8706128 10.1016/S0092-8674(00)80095-1
RZPD - Deutsches Ressourcenzentrum für Genomforschung GmbH
Stauber RH Horie K Carney P Hudson EA Tarasova NI Gaitanaris GA Pavlakis GN Development and applications of enhanced green fluorescent protein mutants Biotechniques 1998 24 462 6, 468-71 9526659
Zeitlmann L Knorr T Knoll M Romeo C Sirim P Kolanus W T cell activation induced by novel gain-of-function mutants of Syk and ZAP-70 J Biol Chem 1998 273 15445 15452 9624129 10.1074/jbc.273.25.15445
Schneider R Campbell M Nasioulas G Felber BK Pavlakis GN Inactivation of the human immunodeficiency virus type 1 inhibitory elements allows Rev-independent expression of Gag and Gag/protease and particle formation J Virol 1997 71 4892 4903 9188551
Schaefer LH Schuster D Herz H Generalized approach for accelerated maximum likelihood based image restoration applied to three-dimensional fluorescence microscopy J Microsc 2001 204 99 107 11737543 10.1046/j.1365-2818.2001.00949.x
National Center for Biotechnology Information
Genomatix-Understanding gene regulation
NESbase 1.0
Quandt K Frech K Karas H Wingender E Werner T MatInd and MatInspector: new fast and versatile tools for detection of consensus matches in nucleotide sequence data Nucleic Acids Res 1995 23 4878 4884 8532532
A program for identifying the initiation codons in cDNA sequences
| 15847701 | PMC1097722 | CC BY | 2021-01-04 16:39:11 | no | BMC Cell Biol. 2005 Apr 24; 6:20 | utf-8 | BMC Cell Biol | 2,005 | 10.1186/1471-2121-6-20 | oa_comm |
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-231586212910.1186/1471-2121-6-23Research ArticleA novel EB-1/AIDA-1 isoform, AIDA-1c, interacts with the Cajal body protein coilin Xu Hongzhi [email protected] Michael D [email protected] Department of Biochemistry, The University of Mississippi Medical Center Jackson, MS 39216-4505, USA2005 29 4 2005 6 23 23 31 1 2005 29 4 2005 Copyright © 2005 Xu and Hebert; licensee BioMed Central Ltd.2005Xu and Hebert; 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
Cajal bodies (CBs) are nuclear suborganelles that play a role in the biogenesis of small nuclear ribonucleoproteins (snRNPs), which are crucial for pre-mRNA splicing. Upon nuclear reentry, Sm-class snRNPs localize first to the CB, where the snRNA moiety of the snRNP is modified. It is not clear how snRNPs target to the CB and are released from this structure after their modification. Coilin, the CB marker protein, may participate in snRNP biogenesis given that it can interact with snRNPs and SMN. SMN is crucial for snRNP assembly and is the protein mutated in the neurodegenerative disease Spinal Muscular Atrophy. Coilin knockout mice display significant viability problems and altered CB formation. Thus characterization of the CB and its associated proteins will give insight into snRNP biogenesis and clarify the dynamic organization of the nucleus.
Results
In this report, we identify a novel protein isoform of EB-1/AIDA-1, termed AIDA-1c, that interacts with the CB marker protein, coilin. Northern and nested PCR experiments reveal that the AIDA-1c isoform is expressed in brain and several cancer cell lines. Competition binding experiments demonstrate that AIDA-1c competes with SmB' for coilin binding sites, but does not bind SMN. When ectopically expressed, AIDA-1c is predominantly nuclear with no obvious accumulations in CBs. Interestingly, another EB-1/AIDA-1 nuclear isoform, AIDA-1a, does not bind coilin in vivo as efficiently as AIDA-1c. Knockdown of EB-1/AIDA-1 isoforms by siRNA altered Cajal body organization and reduced cell viability.
Conclusion
These data suggest that specific EB-1/AIDA-1 isoforms, such as AIDA-1c, may participate in the regulation of nucleoplasmic coilin protein interactions in neuronal and transformed cells.
==== Body
Background
Studies conducted to understand nuclear organization and function have revealed that the nucleus contains a myriad of dynamic, highly organized domains, territories and bodies [1-3]. The Cajal body (CB) is one such structure and contains the highest concentration of small nuclear ribonucleoproteins (snRNPs) in the nucleus [4,5]. SnRNPs are a vital part of the pre-mRNA processing machinery, necessary for both splice site recognition and the splicing reaction. More precisely, the catalyst for the splicing reaction is probably the small nuclear ribonucleic acid (snRNA) component of the snRNP [6]. The human U1, U2, U4, U5 and U6 snRNAs share the intriguing characteristic in that they contain numerous post-transcriptionally modified nucleotides, namely pseudouridines and 2'-O-methyl groups [7]. These modifications appear to be important for efficient splicing of RNA and, in the case of U2, are essential for spliceosomal assembly [8]. In addition to base modification, U1, U2, U4 and U5 snRNAs undergo an elaborate maturation pathway in which they are exported to the cytoplasm. In the cytoplasm, these snRNAs are substrates for the survival of motor neurons (SMN) protein [9]. Mutations in SMN cause the neurodegenerative disorder Spinal Muscular Atrophy [10]. SMN is essential for the assembly of core Sm proteins (B/B', D3, D1, D2, E, F, and G) onto the snRNA [11], and participates in additional modification steps and the import of nascent snRNPs into the nucleus [12,13].
After their transport into the nucleus, newly formed snRNPs accumulate first in the CB and then localize to interchromatin granule clusters, which are also known as speckles [14]. Recent work from the groups of Kiss and Bertrand has shown that CBs contain small guide RNAs, termed small Cajal body-specific RNAs (scaRNAs), which localize exclusively to the CB and guide modifications of U1, U2, U4 and U5 snRNA by pseudouridylation and 2'-O-methylation [15,16]. Additional work from this same group demonstrates that these snRNAs are indeed subject to base modification in the CB after import from the cytoplasm [17]. For example, short fragments of Sm snRNAs are modified when they are targeted to the CB, but are not modified when targeted to the nucleolus. Other work has shown that telomerase RNA shares structural similarities to scaRNAs and is also enriched within the Cajal body [18,19]. It is possible, therefore, that CBs may play a role in the assembly of the telomerase holoenzyme in addition to its clear role in snRNA modification. Still additional work implicates CBs in U4/U6 assembly [20,21], U4/U6.U5 tri-snRNP recycling [22], U2 snRNP biogenesis [23] and U3 small nucleolar RNA transport [24].
The marker protein for the CB is considered to be coilin, but the majority (70%) of coilin is actually diffusely localized throughout the nucleoplasm [25]. Although coilin is conserved among vertebrates and homologues are present in plants, a corresponding gene has not been identified in worms and flies [4]. Nevertheless, CB-like structures are present in a variety of insects, including Drosophila [26,27]. The accumulation of snRNPs within CBs is contingent upon coilin [28-30] and characterization of coilin knockout mice demonstrates that inbred strains have significant viability and fertility defects [29]. Cell lines derived from coilin knockout mice do not have CBs, but instead have residual CBs which fail to accumulate snRNPs or SMN [29]. Thus while not an essential protein, coilin likely plays an important role in coordinating the activities of various types of RNA processing machineries into a common nuclear structure [4,31,32].
Coilin has been shown to interact directly with SMN and various Sm proteins and can compete with SmB' for binding sites on SMN [30]. The interaction between coilin and SMN has been found to be contingent upon several arginine/glycine (RG) dipeptide repeats present within coilin [30]. These arginines are symmetrically dimethylated, increasing the affinity of SMN for coilin [33,34]. In the nucleus, SMN localizes in most cell types and cell lines to CBs [1]. Curiously, SMN exists in some cell lines and fetal tissues in discrete nuclear structures termed "gems" (for Gemini of Cajal bodies) [35,36]. The presence of SMN gems correlates with a decrease in coilin methylation [33,34].
It is not yet understood how nascent snRNPs are targeted to the CB for modification. Also unclear is how snRNPs are released from the CB after modification of the snRNA moiety takes place. Moreover, the function of coilin in the nucleoplasm, where the majority of the protein is found, remains elusive. In this work, we identify a novel protein isoform of the EB-1/AIDA-1 gene, AIDA-1c, as a coilin interacting protein. Expression analysis demonstrates that AIDA-1c is present in neuronal tissue and cancer cell lines. AIDA-1c competes with SmB' for coilin binding sites and localizes predominantly to the nucleoplasm when ectopically expressed. Other EB-1/AIDA-1 isoforms, in particular AIDA-1a, have been shown to interact with the amyloid-β protein precursor (AβPP) intracellular domain (AID) [37,38]. All known EB-1/AIDA-1 isoforms contain a conserved phosphotyrosine binding (PTB) domain, indicating a role in signal transduction and/or protein:protein interaction [39,38]. Interestingly, AIDA-1a does not interact with coilin in vivo as efficiently as AIDA-1c. These findings suggest that specific EB-1/AIDA-1 isoforms, such as AIDA-1c, may participate in the regulation of nucleoplasmic coilin protein interactions, and thereby indirectly affect CB activity in neuronal and transformed cells.
Results and discussion
Isolation of AIDA-1c
To broaden our understanding of Cajal bodies and coilin, we conducted a yeast two-hybrid screen with coilin as bait. We chose a human brain cDNA library for this assay because we are interested in assessing CB protein dynamics in neuronal tissue, which is the cell type primarily affected in Spinal Muscular Atrophy. The C-terminal 214 aa of coilin, which contains the RG box, was the bait for the screen. This region of coilin mediates the interaction between coilin and SMN. Additionally, C214 of coilin is sufficient to interact with Sm proteins [30]. Consequently, other proteins that bind C214 may regulate the interplay between coilin/SMN and coilin/Sm and thus play a role in CB dynamics. The isolation of proteins that interact with this important region of coilin is therefore a key first step towards understanding how RNPs are targeted to, modified within and released from the CB.
Several preys were recovered from the screen, one of which appeared to be a partial coding sequence of an EB-1 protein isoform. Transcription of EB-1 is activated in t(1:19) pre-B cell acute lymphocytic leukemia [39]. This translocation produces the chimeric oncoprotein E2a-Pbx1, leading to the aberrant expression of three tissue-specific genes, EB-1 among them. Recent work has identified other isoforms emanating from this gene and have found that at least one of these proteins associates with the Amyloid-β Protein Precursor (AβPP) Intracellular Domain (AID) and hence are termed AIDA proteins [38]. Alterations in AβPP processing are thought to underlie the pathology of Alzheimer's disease. All known isoforms of the EB-1/AIDA-1 gene contain a phosphotyrosine binding (PTB) domain, including the apparently truncated protein we isolated from the two-hybrid screen. PTB domains were initially characterized as protein interaction modules that bind to regions containing phosphotyrosine [40]. However, these domains have also been found to be important for protein:protein interaction in the absence of phosphotyrosine, possibly serving as a platform for a broad range of associations [41]. In addition to the PTB domain, some isoforms of EB-1/AIDA-1 contain other protein interaction domains such Sterile Alpha Motifs (SAM) [42] and ankyrin repeats [43].
Since the cDNA clone obtained from the two-hybrid screen against coilin(C214) appears to be incomplete (it lacks a translational start sequence), we conducted PCR on a human brain cDNA library using a forward primer specific to the translation start of another EB-1/AIDA-1 isoform, AIDA-1a, and a reverse primer specific to the 3' end of the isolated cDNA. Sequencing of the resultant product demonstrated that we amplified a novel EB-1/AIDA-1 isoform, which we have termed AIDA-1c (Figure 1A). AIDA-1c is 426 amino acids length and has a predicted molecular weight of 48 kDa. We have not tested if AIDA-1c binds the AID domain of the amyloid-β protein precursor. Alignment of the protein sequence of AIDA-1c with AIDA-1a demonstrates that the C-terminus of AIDA-1c differs from AIDA-1a (Figure 1B). Furthermore, while the N-terminus of AIDA-1c is identical to AIDA-1a, AIDA-1c contains an internal deletion between the SAM and PTB domains not found in AIDA-1a or other known isoforms. We could not amplify a product that contained the translational start of AIDA-1b and the 3' exon of AIDA-1c.
Figure 1 AIDA-1c is a member of the EB-1/AIDA-1 protein family. (A) Schematic amino acid alignment of some known EB-1/AIDA-1 isoforms. Accession numbers AY281131 (AIDA-1b) and AY281132 (AIDA-1a) were reported by Ghersi et al., 2004a. AF145204 (EB-1) was reported by [39] and does not contain a translational start site. AIDA-1c (AY753193) is from this study. AIDA-1b is predicted to be approximately 1250 aa and contains ankrin repeats (denoted by 'A'), Eph-related SAM domains ('S') and a phosphotyrosine-binding domain ('PTB'). Boxes with shading or hatching represent variation from the AIDA-1b sequence. (B) Amino acid alignment of AIDA-1a versus AIDA-1c.
To get an approximate idea as to the expression profile of AIDA-1c mRNA, we conducted a multi-tissue Northern blot using a probe derived from the unique 3' end of AIDA-1c (Figure 2A). This probe should therefore discriminate between AIDA-1c and other isoforms, unless they share the same 3' exon and untranslated region as found for AIDA-1c. A band of approximately 4 kbp in length corresponding to the AIDA-1c message is detected only in neuronal tissue (recall that a brain cDNA library was used in the isolation of AIDA-1c). To explore if AIDA-1c expression can be detected in cell lines, we conducted nested PCR using primers specific to the 3' end of AIDA-1c on cDNA made from various cancer lines. As shown in Figure 2B, AIDA-1c expression can be detected in all six cell lines. We also observed expression of AIDA-1c in HeLa cells, as assessed by RT-PCR (Figure 2B, lower panel). These results show that AIDA-1c is detected in cancer cells and is expressed at the highest level in brain among normal tissues. However, since nested- and RT-PCR are much more sensitive techniques than Northern blotting, we cannot exclude the possibility that low levels of this isoform are present in normal tissues.
Figure 2 AIDA-1c expression is highest in neuronal tissue and is present in cancer lines. (A) A Multi-Tissue Northern blot (Clontech) was probed with an AIDA-1c specific probe. Re-probing of the same blot with a β-actin probe (bottom panel) is shown, verifying the presence of RNA in each lane. (B) (Upper panel) Nested PCR using AIDA-1c specific primers was conducted on cDNA obtained from a variety of cancer cell lines: SK-MEL-2 (melanoma), NIH:OVCAR-3 (ovarian), HCT-15 (colon), A-549 (non small cell lung carcinoma), MDA-MB-231 (breast), MCF-7 (breast). The positive control lane used AIDA-1c cDNA as the template. (Lower panel) HeLa expresses AIDA-1c. HeLa RNA was subjected to RT-PCR using an AIDA-1c specific primer. No product is observed in the reaction lacking reverse transcriptase (-RT).
A 47kDa. protein reactive to AIDA-1 antibodies interacts with the C-terminus of coilin
Since the two-hybrid result indicates that coilin interacts with a member of the EB-1/AIDA-1 protein family, possibly AIDA-1c, we decided to screen HeLa lysate for EB-1/AIDA-1 proteins that interact with coilin. As a first step, we monitored the expression of EB-1/AIDA-1 proteins from HeLa cytoplasmic and nuclear extract using anti-AIDA-1 antibodies (Zymed, South San Francisco, CA) [37,38]. This antibody was made using a peptide whose sequence is also present in AIDA-1c, and has been confirmed to specifically recognize a number of proteins in human brain extract [37]. By immunoblotting of HeLa lystate, anti-AIDA-1 detects several proteins in both the cytoplasmic and nuclear fractions (Figure 3A). In particular, three prominent proteins, approximately 172, 75, and 47 kDa. in size, are detected by anti-AIDA-1 in nuclear extract. A doublet migrating around 62 kDa. in size is also detected, albeit faintly. To test if any of these proteins interact with coilin, which is also nuclear, we subjected HeLa nuclear lysate to immunoprecipitation with anti-AIDA-1 antibodies, followed by SDS-PAGE and Western blotting with anti-coilin antibodies. We could not detect endogenous coilin association with any AIDA-1 protein (our unpublished results). However, ectopically expressed myc-tagged coilin can be specifically co-immunoprecipitated by anti-AIDA-1 antibodies, but not by normal rabbit serum (Figure 3B). The interaction between coilin and EB-1/AIDA-1 proteins is specific, but likely to be weak or transient given that the input lane accounts for 1/20th the amount of lysate used in the immunoprecipitation reactions (Figure 3B, lane 1).
Figure 3 A 47 kDa. EB-1/AIDA-1 protein interacts with coilin. (A) Cytoplasmic and nuclear extracts from HeLa cells were subjected to Western blotting with anti-AIDA-1 antibodies. (B) Myc-coilin interacts in vivo with EB-1/AIDA-1 nuclear proteins. HeLa cells were transfected with myc-coilin and nuclear extract was subjected to immunoprecipitation (IP) with anti-AIDA-1 antibodies followed by SDS-PAGE and Western blotting with anti-myc antibodies. Normal rabbit serum (NRS) was used as a negative IP control. The input lane accounts for 1/20 the amount of lysate used in the IP reactions. (C) HeLa and melanoma cell lysates were incubated with a GST-tagged coilin fragment (C214) or GST alone, followed by Western blotting with anti-AIDA-1 antibodies (upper panels). Pulldowns using HeLa lysate are shown in lanes 3 and 4. Melanoma cell lysate was used in the reactions shown in lanes 5 and 6. Only the 47 kDa. EB-1/AIDA-1 protein binds to the coilin fragment. The blot was subsequently probed with anti-GST antibody to verify the loading of the GST and GST-C214 beads (lower panel). The input lanes account for 1/20 the amount of lysate used in the pulldown reactions. (D) Purified T7-tagged AIDA-1c protein was run as a control to mark the position of the 47 kDa. protein from nuclear extract (NE). The blot was probed with anti-AIDA-1 antibody.
To corroborate this result and ascertain which EB-1/AIDA-1 protein associates with coilin, we incubated total HeLa or SK-MEL-2 lysate with GST alone or GST fused to the C-terminal 214 aa of coilin. This region of coilin was used in the two-hybrid screen and interacts with Sm proteins and SMN. After extensive washes, the beads were subjected to SDS-PAGE and Western blotting with anti-AIDA-1 antibodies. No proteins reactive to anti-AIDA-1 antibodies were recovered over GST beads, but a 47 kDa. protein was retained by the C-terminal 214 aa coilin fragment in both HeLa and SK-MEL-2 reactions (Figure 3C, lanes 4 and 6 upper panel). Thus the C-terminus of coilin mediates interaction with the 47 kDa. EB-1/AIDA-1 nuclear protein. Of all the known EB-1/AIDA-1 isoforms, only AIDA-1c and a putative protein found in the GenBank database (accession number AF164792) closely match the size of the protein that interacts with the C-terminus of coilin. To confirm that AIDA-1c migrates through SDS-PAGE gels with an apparent size of 47 kDa., we purified a His-T7-tag fusion of AIDA-1c and ran this on a gel with HeLa nuclear extract. Probing with anti-AIDA-1 demonstrates that His-T7-AIDA-1c has a mobility of around 52 kDa. (Figure 3D), indicating that an untagged protein would migrate at 47–48 kDa. (the His-T7 tag and linker add approximately 4 kDa.).
AIDA-1c interaction with coilin in vivo and in vitro
To establish if AIDA-1c interacts with coilin, as suggested from the immunoprecipitation and two-hybrid experiments described above, we conducted co-immunoprecipitation experiments on lysate from HeLa cells transfected with empty GFP vector or GFP-AIDA-1c using anti-GFP antibodies followed by Western blotting with anti-coilin antibodies. We also tested another nuclear EB-1/AIDA-1 isoform, AIDA-1a, for its ability to associate with coilin in vivo. As shown in Figure 4A, coilin specifically interacts with GFP-AIDA-1c, but not GFP alone, despite the large amount of immunoprecipitated GFP (the anti-GFP signal in lane 4 (lower panel) has been underexposed six-fold relative to the signals in lane 5 and 6). YFP-AIDA-1a associates with coilin above background levels, but below that found for AIDA-1c (Figure 4A, upper panel, compare the signal in lane 6 to that in lane 5). Reprobing of the same blot verified that both FP-AIDA-1c and FP-AIDA-1a were expressed and immunoprecipitated equally (lower panel, compare the anti-GFP signal in lane 5 to that in lane 6). Therefore, we conclude that AIDA-1a does not form a complex with coilin as efficiently as AIDA-1c.
Figure 4 AIDA-1c associates with coilin in vivo and interacts directly with the C-terminus of coilin in vitro. (A) Lysates obtained from cells transfected with GFP-only, GFP-AIDA-1c or YFP-AIDA-1a [38] were immunoprecipitated with anti-GFP antibodies followed by Western analysis with anti-coilin antibodies (upper panel). Co-immunoprecipitated coilin can be observed most clearly in the AIDA-1c IP lane (lane 5). The blot was reprobed with anti-GFP antibodies to monitor the expression and immunoprecipitation of the GFP-tagged proteins (lower panel). The anti-GFP signal in lane 4 (bottom panel) has been underexposed six-fold relative to the signal in lanes 5 and 6. The expression level of GFP-AIDA-1c and YFP-AIDA-1a is low, resulting in no signal for these proteins in the input lanes (not shown). Input lanes account for 1/30 the amount of lysate used in the IP reactions. (B) AIDA-1c direct interaction with coilin is salt-sensitive. Purified coilin was incubated with GST-AIDA-1c in binding buffer containing various concentrations of NaCl. After the beads were washed, Western analysis was conducted using anti-coilin antibodies (upper panel). The blot was then stained with Ponceau S to verify that approximately equal amounts of GST-AIDA-1c were used in each reaction (lower panel). The input lane accounts for 20% of the protein used in the pulldown reactions. (C) AIDA-1c directly interacts with the C-terminus of coilin. Purified T7-tagged AIDA-1c was incubated with GST or GST-tagged coilin fragments containing either the C-terminal 156 or 214 amino acids of coilin. The blot was subsequently stained with Ponceau S to verify that approximately equal amounts of GST-fusions were used in the assay (lower panel). The input lane accounts for 20% of the protein used in the pulldown reactions. (D) AIDA-1c competes with Sm proteins for coilin binding sites. A competition experiment is shown wherein individual reactions were set-up containing a fixed amount of GST-coilin fragment (GST-C214) and T7-SmB' with increasing amounts T7-AIDA-1c as indicated. As can been seen in lane 6, as more AIDA-1c binds to GST-C214, less SmB' is recovered. Lane 2 shows the binding of either AIDA-1c (top) or SmB' (bottom) to GST-C214 in the absence of competitor. Input lanes account for 20% of the protein used in the 1X reactions.
To determine if the interaction between AIDA-1c and coilin is direct, we conducted GST-pulldown assays using purified recombinant proteins. We found that soluble full-length coilin is recovered by AIDA-1c fused to GST (Figure 4B, lane 2). In support of our previous observations suggesting that the interaction between AIDA-1c and coilin is weak or transient, coilin binding is abolished when the assay was carried out in the presence of increasing salt concentrations (Figure 4B, lanes 3–5). These results indicate that the association between coilin and AIDA-1c is specific and electrostatic bonds are important for this interaction.
To verify that the coilin fragment used in the two-hybrid assay directly interacts with AIDA-1c, we conducted GST-pulldown experiments with bacterially expressed AIDA-1c incubated with GST alone or GST fused to aa 362–576 of coilin (C214). We found that AIDA-1c, detected using anti-AIDA-1 antibodies, interacts directly with the coilin fragment but not with GST alone (Figure 4C, upper panel, compare lane 4 to lane 3). The binding site for AIDA-1c on coilin was further explored using an additional coilin C-terminal fragment (aa 420–576, C156) that lacks the coilin RG box (found between aa 392–420). Recovery of AIDA-1c is observed in the reaction using GST-C156 of coilin (Figure 4C, upper panel, lane 2).
AIDA-1c competes with Sm proteins for coilin binding
We have previously shown that the C-terminal 214 aa of coilin, in addition to AIDA-1c binding, also mediates the interaction with SMN and Sm proteins, such as SmB' [30]. While the interaction between coilin and SMN involves the coilin RG box, which is subject to modification by symmetrical dimethylation [33,34], the binding site for Sm proteins on coilin has not been defined. However, competition binding experiments show that Sm proteins and SMN may have overlapping binding sites on coilin [30]. Since AIDA-1c binds within the same C-terminal 214 residues of coilin as do SMN and Sm proteins, we investigated whether AIDA-1c may compete with Sm proteins for coilin binding sites. AIDA-1c can effectively compete with SmB' for coilin binding sites (Figure 4D, compare the level of SmB' recovered in lane 2 to that in lane 6). We did not observe a direct interaction between SMN and AIDA-1c (our unpublished observations). Thus, the findings that AIDA-1c competes with Sm proteins and binds downstream of the coilin RG box suggest that Sm proteins and SMN have overlapping but distinct binding sites on coilin. The interaction studies also demonstrate that one function of AIDA-1c may be to modify coilin protein dynamics.
Cellular localization of ectopically expressed AIDA-1c
Given that AIDA-1c interacts with coilin, and that a 47kDa. protein which is possibly AIDA-1c is found in nuclear extract, we were interested in assessing the cellular localization of AIDA-1c by transient transfection into HeLa cells. As shown in Figure 5, GFP-tagged AIDA-1c is localized diffusely in the nucleus, excluding nucleoli, with some staining in the cytoplasm. We did not detect significant enrichment of AIDA-1c in CBs (our unpublished observations). A similar localization pattern has been observed for AIDA-1a [37,38]. Therefore, interactions between AIDA-1c and coilin may be taking place in the nucleoplasm. Indeed, while coilin is highly enriched in the CB, the majority (70%) of the protein is diffusely localized throughout the nucleoplasm [25]. To determine the localization of endogenous EB-1/AIDA-1 proteins, we stained HeLa cells with antibodies to EB-1/AIDA-1 proteins. Despite the fact that anti-AIDA-1 works well by immunoblotting, this antibody does not detect nuclear proteins in HeLa cells by immunofluorescence [37,38] (our unpublished observations). We are currently developing an antibody to the unique C-terminus of AIDA-1c in order to specifically detect this isoform.
Figure 5 Ectopically expressed AIDA-1c is predominantly nuclear. HeLa cells transfected with GFP-fused to AIDA-1c were processed and stained with DAPI. GFP-AIDA-1c is predominantly localized in the nucleus.
Knockdown of EB-1/AIDA-1 proteins by siRNA reduces cell viability and disrupts Cajal bodies
The studies described above suggest that a subset of EB-1/AIDA-1 proteins may play a role in Cajal body protein dynamics. To further explore this possibility, siRNA was used to target all known EB-1/AIDA-1 proteins. For this experiment, cells were treated with EB-1/AIDA-1 siRNA for 72 hours, followed by nuclear lysate preparation and Western blotting using anti-AIDA-1 antibodies. Three prominent bands reactive to anti-AIDA-1 are detected in nuclear lysate from mock treated cells 172, 75, and 47 kDa. in size (Figure 6A, top panel). The amounts of these proteins is significantly reduced in extract obtained from cells treated with EB-1/AIDA-1 siRNA. In fact, the 47 kDa. protein, which we believe is AIDA-1c, is reduced by 3.5 fold as assessed by densitometric analysis. Reprobing of the same blot with anti-SMN demonstrates that an approximately equal amount of nuclear protein was loaded in each lane.
Figure 6 EB-1/AIDA-1 siRNA disrupts Cajal body structure and causes cell death. (A) Endogenous EB-1/AIDA-1 protein levels are reduced upon treatment with EB-1/AIDA-1 siRNA. HeLa cells were treated with EB-1/AIDA-1 siRNA (100 nM) for 72 hours, followed by nuclear extraction and Western blotting using anti-AIDA-1 antibodies. The film with the 47 kDa. protein was exposed for a shorter period of time. The blot was reprobed with anti-SMN to verify that equivalent levels of extract were loaded on the gel. (B) Coilin and snRNP localization is disrupted upon treatment with actinomycin D. Panel A shows the normal localization of coilin in CBs (arrows), along with the DAPI stain of the same cell. Panel B shows the normal localization of snRNPs to CBs (arrows) and speckles (arrowheads). Panels C and D are representative images of cells treated with actinomycin D. Note that coilin is relocalized to the periphery of the nucleolus (panel C, arrows) and speckles are enlarged (panel D, arrowheads) in the presence of actinomycin D. (C) HeLa cells exposed to EB-1/AIDA-1 siRNA (100 nM for 72 hours) were subjected to immunofluorescence analysis with antibodies to coilin (red) and snRNPs (Y12) (green). The overlay is shown in the far right panel. Arrows mark altered Cajal body structure. The bottom panels show a closer view of another cell with disrupted CB organization and co-localization of coilin and snRNPs.
During these experiments, we noticed an increased rate of cell death in cells treated with EB-1/AIDA-1 siRNA compared to mock treated. We quantified this phenomenon by conducting experiments with mock-, EB-1/AIDA-1-, or control siRNA-treatment followed by cell counting using the trypan blue exclusion method. The control siRNA used in this study is functional in that it is processed by the RISC machinery, but is non-targeting (Dharmacon Research). Thus this siRNA is an excellent control for off-target effects. In fact, there is no reduction in cell viability when using the control siRNA. In contrast, only 45% (+/- 4%) of cells are viable upon treatment with EB-1/AIDA-1 siRNA.
Since EB-1/AIDA-1 siRNA causes cell death and a subset of EB-1/AIDA-1 isoforms interact with the Cajal body protein coilin, we investigated if nuclear organization, in particular Cajal body structure, was disrupted upon treating cells with EB-1/AIDA-1 siRNA. The normal localization of coilin in CBs is shown in Figure 6B (panel A, arrows), as is the typical distribution of snRNPs in speckles and CBs (panel B, arrowheads and arrows, respectively). In contrast, CBs and speckles display altered structures in the presence of the transcription inhibitor actinomycin D (Figure 6B, panels C and D, [44]). Specifically, CBs relocalize to nucleolar caps (panel C, arrows) and speckles enlarge (panel D, arrowheads). Cells exposed to EB-1/AIDA-1 siRNA have noticeable changes in nuclear organization compared to mock- or control siRNA-treated cells. Notably, coilin and CBs are reorganized in a manner reminiscent to that observed when treating cells with actinomycin D (Figure 6C). In contrast, mock or control treated cells are indistinguishable from normal cells with regard to CB organization (not shown). One noticeable difference between the phenotype of EB-1/AIDA-1 siRNA versus actinomycin D treated cells is that snRNPs, as assessed by Y12 staining, are co-localized with coilin in EB-1/AIDA-1 siRNA cells (Figure 6C, arrows). Previous reports [44] and our work (Figure 6B, panel D) have shown that snRNPs do not co-localize with coilin in actinomycin D treated cells. It is conceivable, therefore, that the observed cell death upon treating cells with EB-1/AIDA-1 siRNA is the result of altered Cajal body organization rather than simply an arrest of transcription. Given that several EB-1/AIDA-1 proteins are likely to be targeted by the EB-1/AIDA-1 siRNA, we cannot ascertain the specific function of AIDA-1c in these experiments. However, if AIDA-1c does participate in snRNP trafficking through the CB, then we would expect to observe altered snRNP transport when AIDA-1c levels are reduced. The presence of snRNPs in coilin-containing nucleolar cap structures formed by EB-1/AIDA-1 siRNA treatment is therefore consistent with a role of AIDA-1c in snRNP displacement from the CB.
Conclusion
The isolation and partial characterization of AIDA-1c demonstrates that specific EB-1/AIDA-1 protein isoforms may regulate the dynamics of coilin protein interactions in the nucleoplasm. Little is known about the role of nucleoplasmic coilin, even though this fraction accounts for the majority of the protein [25]. For example, coilin could interact with snRNPs in the nucleoplasm and bring them to the CB. Coilin association with AIDA-1c in the nucleoplasm might therefore affect the protein composition of the CB indirectly. If correct, then we would expect that the specific disruption of AIDA-1c expression would alter some aspect of snRNP biogenesis or CB function. Current experiments are underway to test this hypothesis.
It is noteworthy that AIDA-1c expression can be detected in both neuronal and cancer cells given that Cajal bodies are prominent in these cells [1]. Curiously, certain neurodegenerative disorders such as Spinal Muscular Atrophy (SMA) dramatically affect neuronal cells even though the mutant protein (SMN) is ubiquitously expressed. Why is neuronal tissue so susceptible to disease proteins? One possible explanation is that neuronal cells, which are known to undergo a high degree of alternative splicing [45], cannot tolerate deficiencies in splicing. It may be that a range of neurodegenerative disorders, through a variety of mechanisms, affect the level of neuronal alternative splicing by disrupting Cajal body function. Consequently, in addition to other cellular processes which are disturbed in some neurodegenerative disorders, inadequate levels of snRNPs may exacerbate the deleterious effects of the disease protein in neuronal tissue. Interestingly, CBs are not normally found in every human cell but are present in immortalized or transformed cells [36,46]. The induction of CBs in cancer cells may reflect on the need for increased levels of snRNPs. Given the presence of CBs in neuronal and cancer cells, it would not be surprising if various factors were upregulated in these cell types to meet their RNP demands. Indeed, EB-1 mRNA is normally expressed in brain and testis but is dramatically induced in marrow from patients having a specific acute lymphocytic leukemia [39]. Likewise, we have found that AIDA-1c expression can be detected both in brain and certain cancer cell lines. We propose that AIDA-1c expression in neuronal and cancer cell lines may facilitate snRNP biogenesis by modulating the interaction between coilin and snRNPs.
Methods
Yeast two-hybrid screen and plasmid construction
A human brain cDNA library cloned into the prey vector pACT2 and pretransformed into the yeast strain Y187 (Clontech) was mated with the strain PJ69-2A harboring the bait vector pAS2-1-coilin(362–576, C214) per the manufacturer's instructions. After mating, the yeast were plated onto medium lacking tryptophan, leucine, histidine and adenine (to select for bait, prey and protein interaction). Colonies were picked after several days of incubation and the prey plasmid was isolated and transformed into PJ69-2A containing pAS2-1-C214 or control bait plasmids to confirm the specificity of the interaction. Restriction digests and sequencing revealed that nine different prey cDNAs were recovered, one of which appeared to encode a truncated EB-1/AIDA-1 protein and was termed CAP36C. The sequence for CAP36C has been submitted to GenBank and given the accession number AY356353. For bacterial expression, cDNAs were cloned into pET28a (Novagen), generating a His-T7-tag fusion, or pGEX-2T/ pGEX-3X (Amersham Pharmacia), generating a fusion to GST. Additional constructs were generated employing standard molecular biological techniques or have been described previously [47,30,33].
Isolation of AIDA-1c
AIDA-1c was isolated by conducting PCR on a human brain cDNA library (Clontech) using a forward primer specific to the translation start of another EB-1/AIDA-1 isoform, AIDA-1a (accession number AY281132), and a reverse primer specific to the 3' exon of AIDA-1c. The sequence for AIDA-1c has been submitted to GenBank and given the accession number AY753193. AIDA-1c was cloned into various expression vectors by standard molecular biological techniques. We could not amplify a product that contained the translational start of AIDA-1b and the 3' exon of AIDA-1c.
Northern analysis
A Multi-Tissue Northern blot (Clontech) was hybridized with a 32P-radiolabled cDNA probe for AIDA-1c according to the manufacturer's instructions. The probe was generated by PCR using CAP36C DNA as template and primers designed to amplify a fragment of AIDA-1c downstream of the PTB domain, including the 3' untranslated region.
Nested- and RT-PCR
The Nested PCR assay was carried out in the DNA Engine PTC-200 Peltier Thermal Cycler (MJ Research, Watertown, MA) to assess the expression of AIDA-1c in cultured cancer cell lines. The cDNA from six cancer cell lines were kindly provided by Dr. Laree Hiser (The University of Mississippi Medical Center): SK-MEL-2 (melanoma), NIH: OVCAR-3 (ovarian cancer), HCT-15 (colon cancer), A-549 (non-small cell lung cancer), MDA-MB-231 (breast cancer) and MCF-7 (breast cancer). Two pairs of nested primers flanking the AIDA-1c specific sequence on the 3' end of the AIDA-1c cDNA were used to amplify the sequence. Plasmid containing the AIDA-1c cDNA and water served as positive and negative controls for the reaction, respectively. Each reaction had 30 cycles and, after the first reaction, the product was diluted 50 fold to serve as the template for the second pair of primers. For RT-PCR reactions, DNA-free RNA was isolated from HeLa cells using the RNAqueous-4PCR DNA-free RNA isolation kit (Ambion, Austin, TX). The RNA was reverse-transcribed with the Thermoscript RT-PCR kit (Invitrogen) using an antisense primer for 45 minutes at 50°C. The antinsense primer binds the 3' UTR of AIDA-1c. A forward primer which binds to the 3' coding sequence of AIDA1c was used in combination with the antisense primer to amplify the corresponding cDNA frangment (25 cycle reaction).
Antibodies
A polyclonal anti-AIDA-1 antibody was purchased from Zymed (South San Francisco, CA) A monoclonal antibody to coilin [48] was a gift from Greg Matera (Case Western Reserve University). We also utilized a commercially available coilin mnonclonal antibody (ab11822) from Abcam (Cambridge, MA). Anti-T7 (1:1000) was obtained from Novagen (Madison, WI). Antibodies to GFP were from BD Biosciences, Palo Alto, CA (polyclonal, catalog number 632459) or Roche, Indianapolis, IN (monoclonal, catalog number 1814460). Myc and GST antibodies were obtained from Santa Cruz Biotechnology, Santa Cruz, CA
Cell Culture, Transfection, Immunofluorescence, and Immunoprecipitation
HeLA cells were cultured, transfected with Superfect (Qiagen, Valencia, CA) and processed as described [47,30,33]. For immunofluorescence, cells were grown on chamberslides and fixed with 4% paraformaldeyde followed by permeabilizaton with 0.5% Triton X-100. The permeable cells were then blocked with 10% normal goat serum for 30 min and probed with the corresponding primary and secondary antibodies for 30 min each. For co-immunoprecipitation experiments, the protocol detailed in [47] was followed with a few modifications. Briefly, cells were lysed in a modified RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 2 mM EDTA plus 1 tablet of complete protease inhibitor cocktail (Roche, Mannheim, Germany) per 50 ml lysis solution). The lysates were then repeatedly passed through a 25 gauge needle to shear DNA or passed through a Qiashredder column (Qiagen) and subject t incubation with the appropriate antibodies for 1 hour at 4°C with gentle shaking. Following the addition of protein A or G Sepharose (Amersham Pharmacia, Uppsala, Sweden) for 1 hour with gentle shaking, the beads were washed with 1 ml lysis solution 5 times. The beads were resuspended in 5X SDS loading buffer, boiled, and subject to SDS-PAGE and Western blotting as described [47]. Input lanes typically account for 1/20 to 1/50 the amount of lysate used in the immunoprecipitation reactions. For co-immunoprecipitation experiments on nuclear protein, nuclear extract was made using the NE-PER kit from Pierce, followed by dilution of the extract to 1 ml with mRIPA and immunoprecipitaton as described above. Where indicated, actinomycin D was used at 2.5 μg/ml for two hours.
In vitro binding assays
GST- and His-tagged constructs, after transformation into E. coil BL21(DE3)pLysS cells, were indiced and purified as described [30]. In a binding reaction, approximately 1 ug of His-T7-tagged protein was incubated with 1 ug of the GST-fusion protein in 1 ml of mRIPA plus 2 mM DTT. After incubation for 1 hour at 4°C with gentle inversion, the beads were washed 5 times (1 ml each) with mRIPA plus DTT, resuspended in 15 ul 5X SDS loading buffer, boiled and subjected to SDS-PAGE. Anti-T7 (1:1000; Novagen, Madison, WI) was used to detect SmB' and anti-AIDA-1 was used to detect AIDA-1c. Competition experiments were performed as described in [30], except that increasing amounts of AIDA-1c were used. Input lanes account for 1/5 the amount of lysate used in the pulldown reactions. Additional NaCl was added to the binding buffer where indicated.
SiRNA knockdowns and viability assays
SiRNAs to the clone obtained in the two-hybrid screen described above (accession number AY356353) were obtained from Dharmacon Research, as was control siRNA (no. D-001210-01-05). The siRNA should target all known EB-1/AIDA-1 isoforms. SiRNA was introduced into HeLa cells using Lipofectamine 2000 per the manufacturers' directions (Invitrogen). To assess the knockdown of endogenous EB-1/AIDA-1 proteins, EB-1/AIDA-1 siRNA was added to cells at a final concentration of 100 nM, followed by 72 hours of incubation. Nuclear extract was then made using the NE-PER kit from Pierce, followed by Western blotting with anti-AIDA-1 and anti-SMN antibodies. Densitometric analysis of the films was conducted using ImageQuant software. To monitor the viability of cells treated with EB-1/AIDA-1 siRNA, equal numbers of cells were seeded into three wells of a six-well plate. One well was mock treated (no siRNA), while the other two received 100 nM of control or EB-1/AIDA-1 siRNA. After 72 hours of incubation, the cells were counted and scored for viability using Trypan blue. The experiment was done in triplicate.
Authors' contributions
HX made the DNA constructs and conducted most of the experiments. MDH conceived the experiments, coordinated the study and wrote the paper.
Acknowledgements
We acknowledge Brad Somers for his excellent technical assistance. We thank P. Szymczyk (Case Western Reserve University, Cleveland, OH) for assistance with the Northern blot and G. Matera (Case Western Reserve University, Cleveland, OH) for communicating data before publication and the coilin monoclonal antibody gift. We also wish to thank E. Ghersi and L. D'Adamio (Albert Einstein College of Medicine, NY) for providing us with the YFP-AIDA-1a construct. A Research Grant to M.D.H. from the Muscular Dystrophy Association supported this work.
==== Refs
Matera AG Nuclear bodies: multifaceted subdomains of the interchromatin space Trends Cell Biol 1999 9 302 9 10407409 10.1016/S0962-8924(99)01606-2
Spector DL Nuclear domains J Cell Sci 2001 114 2891 3 11686292
Bubulya PA Spector DL "On the move"ments of nuclear components in living cells Exp Cell Res 2004 296 4 11 15120987 10.1016/j.yexcr.2004.03.018
Gall JG The centennial of the Cajal body Nat Rev Mol Cell Biol 2003 4 975 80 14685175 10.1038/nrm1262
Matera AG Cajal bodies Curr Biol 2003 13 R503 12842022 10.1016/S0960-9822(03)00438-X
Villa T Pleiss JA Guthrie C Spliceosomal snRNAs: Mg(2+)-dependent chemistry at the catalytic core? Cell 2002 109 149 52 12007401 10.1016/S0092-8674(02)00726-2
Massenet S Branlant C A limited number of pseudouridine residues in the human atac spliceosomal UsnRNAs as compared to human major spliceosomal UsnRNAs RNA 1999 5 1495 503 10580478 10.1017/S1355838299991537
Yu YT Shu MD Steitz JA Modifications of U2 snRNA are required for snRNP assembly and pre-mRNA splicing EMBO J 1998 17 5783 95 9755178 10.1093/emboj/17.19.5783
Will CL Luhrmann R Spliceosomal UsnRNP biogenesis, structure and function Curr Opin Cell Biol 2001 13 290 301 11343899 10.1016/S0955-0674(00)00211-8
Gubitz AK Feng W Dreyfuss G The SMN complex Exp Cell Res 2004 296 51 6 15120993 10.1016/j.yexcr.2004.03.022
Pellizzoni L Yong J Dreyfuss G Essential Role for the SMN Complex in the Specificity of snRNP Assembly Science 2002 298 1775 9 12459587 10.1126/science.1074962
Mouaikel J Narayanan U Verheggen C Matera AG Bertrand E Tazi J Bordonne R Interaction between the small-nuclear-RNA cap hypermethylase and the spinal muscular atrophy protein, survival of motor neuron EMBO Rep 2003 4 616 22 12776181 10.1038/sj.embor.embor863
Narayanan U Ospina JK Frey MR Hebert MD Matera AG SMN, the spinal muscular atrophy protein, forms a pre-import snRNP complex with snurportin1 and importin beta Hum Mol Genet 2002 11 1785 95 12095920 10.1093/hmg/11.15.1785
Sleeman JE Lamond AI Newly assembled snRNPs associate with coiled bodies before speckles, suggesting a nuclear snRNP maturation pathway Curr Biol 1999 9 1065 74 10531003 10.1016/S0960-9822(99)80475-8
Kiss AM Jady BE Darzacq X Verheggen C Bertrand E Kiss T A Cajal body-specific pseudouridylation guide RNA is composed of two box H/ACA snoRNA-like domains Nucleic Acids Res 2002 30 4643 9 12409454 10.1093/nar/gkf592
Darzacq X Jady BE Verheggen C Kiss AM Bertrand E Kiss T Cajal body-specific small nuclear RNAs: a novel class of 2'-O-methylation and pseudouridylation guide RNAs EMBO J 2002 21 2746 56 12032087 10.1093/emboj/21.11.2746
Jady BE Darzacq X Tucker KE Matera AG Bertrand E Kiss T Modification of Sm small nuclear RNAs occurs in the nucleoplasmic Cajal body following import from the cytoplasm EMBO J 2003 22 1878 88 12682020 10.1093/emboj/cdg187
Richard P Darzacq X Bertrand E Jady BE Verheggen C Kiss T A common sequence motif determines the Cajal body-specific localization of box H/ACA scaRNAs EMBO J 2003 22 4283 93 12912925 10.1093/emboj/cdg394
Zhu Y Tomlinson RL Lukowiak AA Terns RM Terns MP Telomerase RNA accumulates in Cajal bodies in human cancer cells Mol Biol Cell 2004 15 81 90 14528011 10.1091/mbc.E03-07-0525
Stanek D Rader SD Klingauf M Neugebauer KM Targeting of U4/U6 small nuclear RNP assembly factor SART3/p110 to Cajal bodies J Cell Biol 2003 160 505 16 12578909 10.1083/jcb.200210087
Stanek D Neugebauer KM Detection of snRNP assembly intermediates in Cajal bodies by fluorescence resonance energy transfer J Cell Biol 2004 166 1015 25 15452143 10.1083/jcb.200405160
Schaffert N Hossbach M Heintzmann R Achsel T Luhrmann R RNAi knockdown of hPrp31 leads to an accumulation of U4/U6 di-snRNPs in Cajal bodies EMBO J 2004 23 3000 9 15257298 10.1038/sj.emboj.7600296
Nesic D Tanackovic G Kramer A A role for Cajal bodies in the final steps of U2 snRNP biogenesis J Cell Sci 2004 117 4423 33 15316075 10.1242/jcs.01308
Boulon S Verheggen C Jady BE Girard C Pescia C Paul C Ospina JK Kiss T Matera AG Bordonne R PHAX and CRM1 are required sequentially to transport U3 snoRNA to nucleoli Mol Cell 2004 16 777 87 15574332 10.1016/j.molcel.2004.11.013
Lam YW Lyon CE Lamond AI Large-scale isolation of Cajal bodies from HeLa cells Mol Biol Cell 2002 13 2461 73 12134083 10.1091/mbc.02-03-0034
Gall JG Tsvetkov A Wu Z Murphy C Is the sphere organelle/coiled body a universal nuclear component? Dev Genet 1995 16 25 35 7758244
Yannoni Y White K Association of the neuron-specific RNA binding domain-containing protein ELAV with the coiled body in Drosophila neurons Chromosoma 1997 105 332 341 9087375 10.1007/s004120050192
Bauer DW Gall JG Coiled bodies without coilin Mol Biol Cell 1997 8 73 82 9017596
Tucker KE Berciano MT Jacobs EY LePage DF Shpargel KB Rossire JJ Chan EK Lafarga M Conlon RA Matera AG Residual Cajal bodies in coilin knockout mice fail to recruit Sm snRNPs and SMN, the spinal muscular atrophy gene product J Cell Biol 2001 154 293 307 11470819 10.1083/jcb.200104083
Hebert MD Szymczyk PW Shpargel KB Matera AG Coilin forms the bridge between Cajal bodies and SMN, the spinal muscular atrophy protein Genes Dev 2001 15 2720 9 11641277 10.1101/gad.908401
Gall JG Cajal bodies: the first 100 years Annu Rev Cell Dev Biol 2000 16 273 300 11031238 10.1146/annurev.cellbio.16.1.273
Ogg SC Lamond AI Cajal bodies and coilin – moving towards function J Cell Biol 2002 159 17 21 12379800 10.1083/jcb.200206111
Hebert MD Shpargel KB Ospina JK Tucker KE Matera AG Coilin methylation regulates nuclear body formation Dev Cell 2002 3 329 37 12361597 10.1016/S1534-5807(02)00222-8
Boisvert FM Cote J Boulanger MC Cleroux P Bachand F Autexier C Richard S Symmetrical dimethylarginine methylation is required for the localization of SMN in Cajal bodies and pre-mRNA splicing J Cell Biol 2002 159 957 69 12486110 10.1083/jcb.200207028
Liu Q Dreyfuss G A novel nuclear structure containing the survival of motor neurons protein EMBO J 1996 15 3555 3565 8670859
Young PJ Le TT Dunckley M Nguyen TM Burghes AH Morris GE Nuclear gems and Cajal (coiled) bodies in fetal tissues: nucleolar distribution of the spinal muscular atrophy protein, SMN Exp Cell Res 2001 265 252 61 11302690 10.1006/excr.2001.5186
Ghersi E Noviello C D'Adamio L Amyloid-beta protein precursor (AbetaPP) intracellular domain-associated protein-1 proteins bind to AbetaPP and modulate its processing in an isoform-specific manner J Biol Chem 2004 279 49105 12 15347684 10.1074/jbc.M405329200
Ghersi E Vito P Lopez P Abdallah M D'Adamio L The intracellular localization of amyloid beta protein precursor (AbetaPP) intracellular domain associated protein-1 (AIDA-1) is regulated by AbetaPP and alternative splicing J Alzheimers Dis 2004 6 67 78 15004329
Fu X McGrath S Pasillas M Nakazawa S Kamps MP EB-1, a tyrosine kinase signal transduction gene, is transcriptionally activated in the t(1;19) subset of pre-B ALL, which express oncoprotein E2a-Pbxl Oncogene 1999 18 4920 9 10490826 10.1038/sj.onc.1202874
Bork P Margolis B A phosphotyrosine interaction domain Cell 1995 80 693 4 7534213 10.1016/0092-8674(95)90347-X
Zwahlen C Li SC Kay LE Pawson T Forman-Kay JD Multiple modes of peptide recognition by the PTB domain of the cell fate determinant Numb EMBO J 2000 19 1505 15 10747019 10.1093/emboj/19.7.1505
Schultz J Ponting CP Hofmann K Bork P SAM as a protein interaction domain involved in developmental regulation Protein Sci 1997 6 249 53 9007998
Sedgwick SG Smerdon SJ The ankyrin repeat: a diversity of interactions on a common structural framework Trends Biochem Sci 1999 24 311 6 10431175 10.1016/S0968-0004(99)01426-7
Carmo-Fonseca M Pepperkok R Carvalho MT Lamond AI Transcription-dependent colocalization of the U1, U2, U4/U6 and U5 snRNPs in coiled bodies J Cell Biol 1992 117 1 14 1532583 10.1083/jcb.117.1.1
Grabowski PJ Black DL Alternative RNA splicing in the nervous system Prog Neurobiol 2001 65 289 308 11473790 10.1016/S0301-0082(01)00007-7
Spector DL Lark G Huang S Differences in snRNP localization between transformed and nontransformed cells Mol Biol Cell 1992 3 555 569 1535243
Hebert MD Matera AG Self-association of coilin reveals a common theme in nuclear body localization Mol Biol Cell 2000 11 4159 71 11102515
Almeida F Saffrich R Ansorge W Carmo-Fonseca M Microinjection of anti-coilin antibodies affects the structure of coiled bodies J Cell Biol 1998 142 899 912 9722604 10.1083/jcb.142.4.899
| 15862129 | PMC1097723 | CC BY | 2021-01-04 16:39:11 | no | BMC Cell Biol. 2005 Apr 29; 6:23 | utf-8 | BMC Cell Biol | 2,005 | 10.1186/1471-2121-6-23 | oa_comm |
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-291583678710.1186/1471-2148-5-29Research ArticleMitochondrial haplotype diversity in the tortoise species Testudo graeca from North Africa and the Middle East van der Kuyl Antoinette C [email protected] Donato LP [email protected] Fokla [email protected] Dept. of Human Retrovirology, Academic Medical Center, University of Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands2 Centro CARAPAX, CP34, 58024 Massa Marittima (Gr), Italy2005 18 4 2005 5 29 29 21 12 2004 18 4 2005 Copyright © 2005 van der Kuyl et al; licensee BioMed Central Ltd.2005van der Kuyl 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
To help conservation programs of the endangered spur-thighed tortoise and to gain better insight into its systematics, genetic variation and evolution in the tortoise species Testudo graeca (Testudines: Testudinidae) was investigated by sequence analysis of a 394-nucleotide fragment of the mitochondrial 12S rRNA gene for 158 tortoise specimens belonging to the subspecies Testudo graeca graeca, Testudo graeca ibera, Testudo graeca terrestris, and a newly recognized subspecies Testudo graeca whitei. A 411-nucleotide fragment of the mitochondrial D-loop was additionally sequenced for a subset of 22 T. graeca, chosen because of their 12S gene haplotype and/or geographical origin.
Results
Haplotype networks generated by maximum-likelihood and neighbor-joining analyses of both the separate and the combined sequence data sets suggested the existence of two main clades of Testudo graeca, comprising Testudo graeca from northern Africa and Testudo graeca from the Turkey and the Middle East, respectively.
Conclusion
Mitochondrial DNA haplotyping suggests that the tortoise subspecies of T. g. graeca and T. g. ibera are genetically distinct, with a calculated divergence time in the early or middle Pleistocene. Other proposed subspecies could not clearly be recognized based upon their mt haplotypes and phylogenetic position, and were either part of the T. g. graeca or of the T. g. ibera clade, suggesting that genetic evidence for the existence of most of the 15 proposed subspecies of T. graeca is weak.
==== Body
Background
Testudo graeca Linnaeus 1758, or the spur-thighed tortoise, is an endangered species with a broad distribution range. It can be found in northern Africa (e.g. Morocco, Algeria, Tunisia and Libya), the Middle East (e.g. Israel, Lebanon, Jordan, Syria, and Iraq), Europe (Bulgaria, Romania, Turkey, Greece, and multiple introductions into Spain and Greece), and in Asia (e.g. Armenia, Azerbaijan, Georgia, Turkmenistan, Iran, and possibly Afghanistan). Surprisingly, T. graeca is absent from Egypt [1]. Speciation and subspeciation in the genus Testudo are still highly debated, and are notoriously complicated in Testudo graeca. Four subspecies have been recognized based upon morphology [2] and are generally accepted, but many more have (often unofficially) been postulated ([3,4] and several others). The latest revision of T. graeca systematics now includes 15 subspecies [5]: T. g. graeca Linnaeus 1758, T. g. anamurensis Weissinger 1987, T. g. antakyensis Perälä 1996, T. g. armeniaca Chkhikvadze & Bakradze 1991, T. g. buxtoni Boulanger 1920, T. g. cyrenaica Pieh & Perälä 2002, T. g. floweri Bodenheimer 1935, T. g. ibera Pallas 1814, T. g. nabeulensis Highfield 1990, T. g. nikolskii Chkhikvadze &Bakradze 2002, T. g. perses Perälä 2002, T. g. soussensis Pieh 2001, T. g. terrestris Forskål 1775, and T. g. zarudnyi Nikolsky 1896. An overview of Testudo graeca taxonomy is also available from the EMBL reptile database [6]. However, many novel descriptions have not been extended with population research or DNA analysis, and thus the contribution to the classification of T. graeca remains unclear. Two dwarf forms of T. graeca that developed in North Africa have been described as a new species or even another genus: Testudo flavominimaralis and Furculachelys nabeulensis [3]. A rather large, highly domed form, originally named Testudo whitei Bennett, 1836, was moved into a new genus: Furculachelys whitei [7]. This latter classification could not be confirmed by mitochondrial DNA analysis [8].
Of the generally accepted subspecies, T. g. graeca of northern Africa, T. g. ibera from the Republic of Georgia, Bulgaria, North-eastern Greece, Turkey (with the exception of the Black Sea coast), Iran, and northern Iraq, and T. g. terrestris from southern Turkey, Syria, Lebanon, Jordan and Israel, are the best described. The fourth subspecies, T. g. zarudnyi, is restricted to central and eastern Iran and Afghanistan, while T. g. nikolskii was reported to inhabit dry subtropic ecosystems along the north eastern Black Sea coast [9]. T. g. graeca has been introduced in recent times in southern Spain and on the Spanish island of Mallorca (see [10]). A morphological study suggested that the first three subspecies should be elevated to full species level [11]. Furthermore, it was recently hypothesized based on genetic analysis that T. graeca of western Morocco constitutes a separate subspecies [8,10], provisionally named Testudo graeca whitei. Morphologically it was assigned full species status in an earlier study [7]. T. graeca shows variable colouring and patterning, which are greatly influenced by environmental factors and make subspecies identification difficult [12], especially in captive animals of unknown provenance. DNA analysis could also be helpful in subspecies determination [13].
T. graeca numbers are steadily declining in the wild, partly due to agricultural developments and partly because the species is one of the most popular tortoises on the pet market, although its protected status and associated regulations have lessened much of the trade. For example export from Morocco has been illegal since 1978, and import into the European Community has been forbidden since 1984. Because the systematics of the species complex is unclear, and the genetics of wild T. graeca populations has not been investigated, conservation and reintroduction programs are difficult to establish. T. graeca can be bred in captivity, and captive offspring could thus be used to replace wild populations (or be used for the pet market to protect wild animals from illegal trade).
Using samples obtained from 158 Testudo graeca, fragments of the mitochondrial (mt) 12S rRNA gene and D-loop sequence were analysed to estimate haplotype diversity and genetic distances in this tortoise complex. The 12S rRNA gene has previously been used to examine genetic variation in T. graeca from Morocco and from Spain [10]. In this study, three 12S rRNA gene haplotypes were found in 19 tortoises from six locations in these countries. The mt 12S rRNA gene was also used to elucidate phylogenetic relationships in the genus Testudo [8]. In the latter study a total of ten 12S haplotypes were detected in 28 T. graeca specimens originating from 9 locations in Africa and Asia.
Results and discussion
Mitochondrial haplotype variation in Testudo graeca
The 12S rRNA fragment amplified was 393–394 nucleotides in length in our Testudo graeca data set, and 10 positions were variable (including a single nt deletion in one individual). A total of fourteen mt 12S haplotypes were observed in 158 unrelated individuals (Table 1), and six haplotypes were seen only once (Table 2). Five of these haplotypes differed by a single nt change from more frequent haplotypes; one had a single nt deletion. The single nt differences cannot easily be attributed to PCR artifacts, as the analysis was done by direct sequencing of PCR fragments. The D-loop fragment amplified was 410–411 nucleotides in length in the twenty-two selected T. graeca, and 20 positions were variable (including a single nt insertion in one individual). A total of 13 mt D-loop haplotypes were observed in 22 T. graeca (Table 3), of which 6 were seen only once. Four of these differed by a single nt change from other, more frequent haplotypes, while one had a single nt insertion, and one differed by 5 nt from the nearest haplotype. The combined data set contained 16 different mt haplotypes.
Table 1 Nucleotide variation in a 12S rRNA gene fragment of captive and wild-caught Testudo graeca
15 27 85 183 194 216 219 224 244 253 326 328 339 Haplotype
C T A A G A T C C T G T G Tg1.0
C T A A G A T C C T G T A Tg1.1
C T A C G A T C C T G T G Tg2.0
C T A C G A T C C T G C G Tg2.1
C T A C G A T T C T G T G Tg3.0
C T A C G A T T C - G T G Tg3.1
C C A C A G T C C T G T G Tg4.1
C C A C G G T C C T G T G Tg4.2
C C A C G A T C T T G T G Tg4.3
C C A C G A T C C T G T G Tg5.0
T C G C G A C C C T G T G Tg6.0
T C G C G A C C C T A T G Tg6.1
T T G C G A C C C T G T G Tg6.2
T A G C G A C C C T G T G Tg7.0
Table 2 12S rRNA haplotypes in captive and wild-caught Testudo graeca.
12S haplotype Number of individuals Country of origin Subspecies
Tg1.0 33 Morocco (Al-Hoceima)/Tunisia/unknown/Algeria Testudo graeca graeca
Tg1.1 1 Algeria Testudo graeca graeca
Tg2.0 33 Libya/Tunisia/Italy (Sardinia)/unknown Testudo graeca graeca
Tg2.1 1 Algeria Testudo graeca graeca
Tg3.0 5 Unknown Testudo graeca graeca
Tg3.1 1 Unknown Testudo graeca graeca
Tg4.1 5 Unknown Testudo graeca whitei
Tg4.2 8 Western Morocco (Laroche)/unknown Testudo graeca whitei
Tg4.3 1 Unknown Testudo graeca whitei
Tg5.0 17 West-Morocco (Laroche)/unknown Testudo graeca whitei
Tg6.0 34 Lebanon/Israel/Turkey/Italy (Sardinia)/Bulgaria/unknown Testudo graeca terrestris
Tg6.1 1 Unknown Testudo graeca terrestris?
Tg6.2 1 Turkey Testudo graeca anamurensis?
Tg7.0 17 Bulgaria/Russia/unknown Testudo graeca ibera
Table 3 Nucleotide variation in a mt D-loop fragment of 22 Testudo graeca
Tortoise no. Nucleotide position Origin 12S haplotype
1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3
4 4 5 9 1 7 7 9 0 0 3 4 5 9 1 1 2 3 7 7
8 9 8 9 3 0 4 3 1 8 6 4 8 4 3 6 2 2 5 7
TGKL T T - C A T G A A G A G A A A C A A G C Morocco (Al-Hoceima) Tg1.0
703888 T T - C A T G A A G A G A A A C A A G C Tunisia Tg1.0
704089 T T - C A T G A A G A G A A A C A A G C Morocco Tg1.0
SNA1 T T - C A T G A A G A G G A A C A A G C Algeria Tg1.0
TGG0042 A C - T A T G A A G G G A A A C A A G C Tunisia (Sibkur) Tg2.0
TGG0043 A C - T A T G A A G G G A A A C A A G C Tunisia (Sibkur) Tg2.0
TGG0041 A C - T A T G A A G G G A A A C A A G C Tunisia Tg2.0
A2SN A C - T A T G A A G G G A A A C A G G C Algeria Tg2.0
TGGLY1 A C - C A T A A A G A A A A G C A A G C ? Tg3.0
TGLKX A C - C A T A A A G A A A A G C A A G C ? Tg3.1
TGLM1 A C - C A T G A A G A A A A A C A A A C Morocco (Laroche) Tg5.1
TGLM2 A C - C A T G A A G A A A A A C A A A C Morocco (Laroche) Tg4.2
703157 A C - C A T G A A G A A A A A C A A A C ? Tg4.1
TGTURBO A C - C A T G A A G A A A A A C A A A C ? Tg5.0
F1621 A C T C C T G A A G G A A A A T A A A C Lebanon Tg6.0
HZTGIAN A C - C C T G A A G G A A G A T A A A C South-Turkey Tg6.0
TGRW5 A C - C C T G A A G G A A G A T A A G T Israel (Hasharon Park) Tg6.0
TGRW4 A C - C C T G A A G G A A G A T A A A T Israel (Golan Heights) Tg6.0
TGRW1 A C - C C T G A A G G A A G A T A A A T Israel (Petah Tiqwa) Tg6.0
TGRW3 A C - C C T G A T G G A A G A T A A A T Israel (Golan Heights) Tg6.0
TGRW2 A C - C C T G A A A G A A A A T A A G T Israel (Petah Tiqwa) Tg6.0
703837 A C - C C C G G A A G G A A A T G A G C ? Tg7.0
Phylogenetic reconstitutions
The 12S rRNA gene data set consisted of 394 total characters of which 377 were constant and 8 were parsimony informative. The D-loop data set consisted of 411 total characters of which 391 were constant and 13 were parsimony informative. Both data sets were analysed separately, as well as combined, where the 3'end of the 12S sequence was joined to the 5'end of the D-loop sequence (= 805 nt of total sequence). Analyses of both fragments resulted in a comparable clustering of haplotypes (not shown). Only the analysis of the combined data set will be discussed here. Two main clusters, present in both the ML (Fig. 1) and NJ (Fig. 2) trees, each supported by bootstrap values of 93 in the NJ tree, consisted of T. graeca from Morocco, Tunisia and Algeria (subspecies T. g. graeca and T. g. whitei), and T. graeca from Israel, Lebanon, and Turkey (T. g. terrestris and T. g. ibera), respectively. Four T. g. graeca subclades are detected with both ML and NJ methods, all four receive high support (bootstrap value ≥ 85) with the NJ method. These lineages could represent proposed subspecies, but unfortunately detailed descriptions or geographic origins of the animals sequenced were frequently lacking. The most divergent subclade (comprising numbers 703157, TGLM2, TGTURBO, and TGLM1) corresponds with the haplotype from T. g. whitei [8] and to the West Moroccan haplotype of Álvarez et al. [10]. Within T. g. terrestris / ibera no subclades could be confidently resolved. One wild-caught animal described as T. graeca nikolskii did harbour 12S haplotype Tg7.0, which is common in T. g. ibera. In three animals described as putative T. graeca anamurensis, three different 12S haplotypes were found: Tg6.0, Tg7.0 and Tg6.2. Tg6.0 is the common haplotype of T. g. terrestris, while Tg7.0 is encountered in animals found north of Tg6.0, which are described as T. g. ibera. Most likely, two of these three animals are actually terrestris or ibera, also because their geographic origins are unknown. Since T. graeca anamurensis has a limited distribution in the south of Turkey, the unique haplotype Tg6.2 (differing by 2 nt from its nearest neighbour) could eventually belong to this putative subspecies. To confirm this finding, additional wild-caught animals should be analysed. It is possible that additional mt haplotypes are present in animals from the northeastern part of the range of T. graeca (e.g. the Balkans and Greece), but unfortunately no such samples were available.
Figure 1 ML tree of combined 12S rRNA (394 nt) and D-loop (411 nt) sequences from Testudo graeca. Corresponding sequences from Geochelone sulcata were used as outgroup. Provenance of tortoise samples is indicated. The Ln likelihood of the tree = -1708.32.
Figure 2 PNJ tree of combined 12S rRNA (394 nt) and D-loop (411 nt) sequences from Testudo graeca. Provenance of the samples is indicated. Five hundred bootstrap replicates have been analysed. Numbers shown are bootstrap confidence levels (BCL). Haplotype designations are indicated.
Nucleotide distances
P-distances were calculated to estimate the average amount of haplotype divergence within clades and between the five well-supported clades. The average distance for the whole data set was 0.010, with the average p-distance between the Tg6/7 haplotypes being 0.004. The four clades characterized by haplotypes Tg1, Tg2, Tg3, and Tg4/5 (T. g. graeca) showed less within-clade variation (0.000–0.002), with an overall average distance of 0.007. The average p-distance between haplotypes Tg6/7 and the other haplotypes was 0.014. So, the highest variation in this data set is found between haplotypes Tg6/7 versus all the other haplotypes, with a lower variation within the five supported clades.
Dating of mitochondrial DNA lineages
T. graeca 12S rRNA gene sequences from different locations grouped together when compared to 12S fragments from other tortoise species, suggestive of a single species [8]. At least 16 closely related but distinct mitochondrial lineages are found in T. graeca over the main part of its distribution range, suggestive of a relatively recent, rapid radiation of mitochondrial haplotypes in this species. Earlier, depending on the 12S rRNA substitution rate, we have calculated the divergence of T. g. graeca and T. g. ibera to occur between 0.3 and 2.3 million years ago [8]. Based upon fossil evidence, Mlynarski [14] dates the emergence of T. cf. graeca to the early Pleistocene, with the modern subspecies arising in the Holocene. The mitochondrial DNA data agree with a recent origin of T. graeca mt lineages (this paper and [8]). Although the exact dating of lineage separation depends upon the (unknown) substitution rate of tortoise mitochondrial DNA, the foundation of the modern lineages is most likely in the Pleistocene (between 0.1 and 1.8 mya). Especially the very clear separation of T. g. graeca and T. g. ibera is compatible with an earlier origin of these subspecies than the Holocene.
MtDNA variation in T. graeca is rather large compared with the related species T. hermanni, probably because no losses occurred due to Pleistocene Ice Age selection as a consequence of the geographic distribution of T. graeca compared with that of the European T. hermanni [8]. This variety in mitochondrial lineages in both T. g. graeca and T. g. ibera, although not representing subspecies per se, could be at the bottom of the confusing taxonomy of T. graeca. The mtDNA variation observed corresponds with ongoing evolution, and is most likely accompanied by nuclear gene evolution or allele selection that could result in subtle morphological changes, especially when there is simultaneous pressure by environmental conditions.
It is interesting to note that T. graeca 12S fragment is somewhat less variable than the D-loop fragment analysed, a finding compatible with the situation in mammals, where rRNA genes evolve slower than synonymous sites and the variable parts of the D-loop [15]. Also, strong rate variability among mammalian species was detected for the D-loop region [15], making this region less suitable for interspecies comparisons. The average amount of nucleotide substitutions between the T. g. graeca and T. g. ibera haplotype sets is 0.011 subst/site for the 12S gene, and 0.017 subst/site for the D-loop fragment.
In Testudines, the overall mtDNA sequence divergence rate has been estimated at 0.4–0.6% per million years [16,17]. Others, however, did not find this 2.5–3 × reduced rate for the 12S rRNA gene in turtles [18] and other reptiles [19]. Palkovacs et al. [20] detected an accelerated rate in Malagasy Pyxis, which they attributed to small body size and a short generation time in this species. In contrast, mammalian body size, generation time, and metabolic rate were not found to influence "taxon specific" mtDNA evolutionary rates [15]. If we assume that the divergence of T. g. graeca and T. g. ibera occurred approximately in the middle of the Pleistocene (± 1 million years ago), the nucleotide substitution rate in T. graeca is estimated to be around 1.1% per million years for the 12S gene, and 1.7 % per million years for the D-loop fragment (= mammalian rate). If we assume the time point of divergence to be earlier (e.g. 1.8 mya, the beginning of the Pleistocene), substitution rates decrease to 0.6% per million years for the 12S gene, and 0.9 % per million years for the D-loop fragment (=decreased reptile rate).
Some authors (e.g. [11]) have proposed to elevate T. g. graeca and T. g. ibera to full species status (T. graeca and T. ibera). Full species status is compatible with our mtDNA data, and would certainly simplify T. graeca taxonomy. Full species status would also be compatible with geography, T. graeca being confined to North Africa, and T. ibera to the remaining regions in Europe, the Middle East and countries belonging to the former Soviet Union, with Egypt serving as a corridor where neither T. graeca nor T. ibera is found. However, as the evolutionary history of a species can be confused by hybridisation, introgression, and incomplete lineage sorting [21], analysis of nuclear markers would be advisable before full-species status can be reached.
Conclusion
Based upon fragments of the mitochondrial 12S rRNA gene and a D-loop fragment, 16 different haplotypes were detected in animals of the tortoise species Testudo graeca obtained from several locations in North Africa, the Middle East and Turkey. Phylogenetic analysis suggested a clear separation between the recognized subspecies T. g. graeca and T. g. ibera, which a calculated divergence time in the early or middle Pleistocene. Many other putative subspecies were not clearly recognizable from their mt haplotypes, and clustered either with T. g. graeca or T. g. ibera haplotypes. Awaiting analysis of nuclear markers, we propose full species status for T. g. graeca (T. graeca) and T. g. ibera (T. ibera), based upon their divergent mtDNA lineages, with subspecies to be defined.
Methods
Tortoise samples
A total of 158 Testudo graeca samples from supposedly unrelated individuals (verified only for captive animals)were obtained from Henk Zwartepoorte (Blijdorp Rotterdam Zoo, Rotterdam, The Netherlands), Donato Ballasina (CARAPAX, Massa Marittima, Italy), Richard Gibson (Durrell Wildlife Conservation Trust, Jersey, UK), Ron Winkler (University of Leiden, Leiden, The Netherlands), Adrie van Zanten (Dierenpark Amersfoort, Amersfoort, The Netherlands), and from several individual members of the Nederlandse Schildpadden Vereniging (Waalwijk, The Netherlands). Samples were obtained from both captive animals of unknown provenance as well as wild-caught specimens from known geographic origin (Table 1).
Amplification and sequencing
Tortoise mucosal cell samples were obtained by gently scraping the inside of the mouth with a cotton swab or plastic scraper, a non-invasive and highly efficient method for obtaining a sample for DNA analysis. In a few cases blood samples were taken. DNA was extracted by a procedure using silica and guanidine thiocyanate [22]. Amplification of 394 nucleotides of the mt 12SrRNA gene was done with the primer set 12S-L01091 / 12S-H01478 [23]; the numbering refers to the position in the human mtDNA. PCR primers were extended with T7 and SP6 promoter sequences, respectively, to facilitate direct sequencing of the PCR product. Amplification of 411 nucleotides of the mt D-loop sequence was done with a primer set consisting of upstream primer myt001 (5' GAGAAAGACTTAAACCTTC 3'), and downstream primer myt003 (5' GACAAAACAACCAAAGGCCAG 3'), based upon corresponding sequences from Geochelone nigra (Genbank accession numbers AF192942 – AF192964). Primer myt001 is located in the mt tRNAPro gene, primer myt003 in the D-loop sequence. PCR primers myt001 and myt003 were extended with -21M13 and M13RP sequences, respectively, to facilitate direct sequencing of the PCR product.
PCR amplifications were done using the following protocol: denaturation 5 min 95°C, amplification 35 cycles of 1 min 95°C, 1 min 55°C, 2 min 72°C, followed by an extension of 10 min 72°C. Sometimes 40 cycles were needed to obtain a D-loop fragment.
Sequencing was performed in both directions using a PE – Applied Biosystems 377 automated sequencer, using the Dyenamic Direct Cycle Sequencing Kit and the Dyenamic Energy Transfer Dye Primer Set from Amersham Int. (UK), following the manufacturers protocols. Mitochondrial D-loop and 12S fragment sequences used in the analyses are available as supplementary information (see additional file 1).
Phylogenetic analysis
Obtained sequences were aligned manually. Neighbor joining (NJ) trees of the sequences and reference sequences were constructed using the Kimura 2-parameter method and the NJ option in the MEGA package [24]. Bootstrap confidence levels were calculated for 500 replicates of the NJ tree. A single nucleotide insertion in a single D-loop sequence was treated as additional information, and used in pair-wise comparison. Furthermore, the data were analysed using Felsenstein's maximum-likelihood (ML) method as implemented in the BioEdit program, version 4.7.8 (kindly provided by Tom Hall, Department of Microbiology, North Carolina State University, 1999). The option DNAML version 3.5c was used with the transition / transversion ratio set at 2.0. Due to computational time, no bootstrap levels were calculated for the ML tree. P-distances were calculated with the distance option of the MEGA program.
Authors' contributions
ACvdK conceived and designed the study, analysed the sequences, and drafted the manuscript. DLPhB initiated the studies on DNA variation in tortoise species. FZ carried out the PCR reactions, and performed the sequence analysis.
Supplementary Material
Additional File 1
Testudo graeca mitochondrial nucleotide alignments Nucleotide sequences of Testudo graeca mitochondrial D-loop and 12S fragments
Click here for file
Acknowledgements
The author thanks Xianghong Chen, and Jolanda Maas for assistance with sequencing.
==== Refs
Buskirk JR Of the absence of spur-thighed tortoises, Testudo graeca, from Egypt Chelonian Conserv Biol 1996 2 118 120
Ernst CH Barbour RW Turtles of the world 1989 Washington, D.C., Smithsonian Institution Press 266
Highfield AC Tortoises of North Africa: Taxonomy, nomenclature, phylogeny, and evolution with notes on field studies in Tunisia J Chelon Herpetol 1990 1 1 56
Highfield AC Martin J A revision of the Testudines of North Africa, Asia and Europe J Chelon Herpetol 1989 1 1 12
Guyot-Jackson G (Ed.) Numéro spécial Testudo Manouria 2004 7 1 52
The EMBL Reptile Database 2005
Highfield AC Martin J Testudo whitei Bennett 1836: a new light on an old carapace- Gilbert White's Selborne tortoise rediscovered J Chelon Herpetol 1989 1 13 22
van der Kuyl AC Ph Ballasina DL Dekker JT Maas J Willemsen RE Goudsmit J Phylogenetic relationships among the species of the genus Testudo (Testudines: Testudinidae) inferred from mitochondrial 12S rRNA gene sequences Mol Phylogenet Evol 2002 22 174 183 11820839 10.1006/mpev.2001.1052
Leontyeva O Pereshkolnik S Demin S The status of the Mediterranean tortoise (Testudo graeca nikolskii) in northern Caucasus Russian Conservation News 1995 4 26
Alvarez Y Mateo JA Andreu AC Diaz-Paniagua C Diez A Bautista JM Mitochondrial DNA haplotyping of Testudo graeca on both continental sides of the Straits of Gibraltar J Hered 2000 91 39 41 10739123 10.1093/jhered/91.1.39
Gmira S Une nouvelle espèce de tortue Testudininei (Testudo kenitrensis n. sp.) de l `inter Amirien-Tensiftien de Kénitra (Maroc) C R Acad Sci Paris, Serie II 1993 316 701 707
Lambert M Ballasina D Tortoise situation in Northern Africa Red data book on Mediterranean chelonians 2004 Bologna, Italy, Edagricole 1 19
van der Kuyl AC van Gennep DR Dekker JT Goudsmit J Routine DNA analysis based on 12S rRNA gene sequencing as a tool in the management of captive primates J Med Primatol 2000 29 309 317 11198014 10.1034/j.1600-0684.2000.290501.x
Mlynarski M Notes on the amphibian and reptilian fauna of the Polish Pliocene and early Pleistocene Acta Zool Cracov 1962 15 177 192
Pesole G Gissi C De Chirico A Saccone C Nucleotide substitution rate of mammalian mitochondrial genomes J Mol Evol 1999 48 427 434 10079281
Lamb T Lydeard C A molecular phylogeny of the gopher tortoises, with comments on familial relationships within the Testudinoidea Mol Phylogenet Evol 1994 3 283 291 7697187 10.1006/mpev.1994.1036
Caccone A Amato G Gratry OC Behler J Powell JR A molecular phylogeny of four endangered Madagascar tortoises based on MtDNA sequences Mol Phylogenet Evol 1999 12 1 9 10222157 10.1006/mpev.1998.0594
Seddon JM Baverstock PR Georges A The rate of mitochondrial 12S rRNA gene evolution is similar in freshwater turtles and marsupials J Mol Evol 1998 46 460 464 9541541
Mindell DP Knight A Baer C Huddleston CJ Slow rates of molecular evolution in birds and the metabolic rate and body temperature hypotheses Mol Biol Evol 1996 13 422 426
Palkovacs EP Gerlach J Caccone A The evolutionary origin of Indian Ocean tortoises (Dipsochelys) Mol Phylogenet Evol 2002 24 216 227 12144758 10.1016/S1055-7903(02)00211-7
Ballard JW Whitlock MC The incomplete natural history of mitochondria Mol Ecol 2004 13 729 744 15012752 10.1046/j.1365-294X.2003.02063.x
Boom R Sol CJ Salimans MM Jansen CL Wertheim-van Dillen PM van der Noordaa J Rapid and simple method for purification of nucleic acids J Clin Microbiol 1990 28 495 503 1691208
Kocher TD Thomas WK Meyer A Edwards SV Paabo S Villablanca FX Wilson AC Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers Proc Natl Acad Sci U S A 1989 86 6196 6200 2762322
MEGA3: Integrated Software for Molecular Evolutionary Genetics Analysis and Sequence Alignment 2005
| 15836787 | PMC1097724 | CC BY | 2021-01-04 16:37:16 | no | BMC Evol Biol. 2005 Apr 18; 5:29 | utf-8 | BMC Evol Biol | 2,005 | 10.1186/1471-2148-5-29 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-531583313910.1186/1471-2164-6-53Research ArticleMicroarray karyotyping of commercial wine yeast strains reveals shared, as well as unique, genomic signatures Dunn Barbara [email protected] R Paul [email protected] Gavin [email protected] Dept. of Genetics, Stanford University Medical Ctr., Stanford, CA 94305-5120, USA2 Dept. of Biological Sciences, Stanford University, Stanford, CA 94305-5020, USA2005 16 4 2005 6 53 53 8 10 2004 16 4 2005 Copyright © 2005 Dunn et al; licensee BioMed Central Ltd.2005Dunn 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
Genetic differences between yeast strains used in wine-making may account for some of the variation seen in their fermentation properties and may also produce differing sensory characteristics in the final wine product itself. To investigate this, we have determined genomic differences among several Saccharomyces cerevisiae wine strains by using a "microarray karyotyping" (also known as "array-CGH" or "aCGH") technique.
Results
We have studied four commonly used commercial wine yeast strains, assaying three independent isolates from each strain. All four wine strains showed common differences with respect to the laboratory S. cerevisiae strain S288C, some of which may be specific to commercial wine yeasts. We observed very little intra-strain variation; i.e., the genomic karyotypes of different commercial isolates of the same strain looked very similar, although an exception to this was seen among the Montrachet isolates. A moderate amount of inter-strain genomic variation between the four wine strains was observed, mostly in the form of depletions or amplifications of single genes; these differences allowed unique identification of each strain. Many of the inter-strain differences appear to be in transporter genes, especially hexose transporters (HXT genes), metal ion sensors/transporters (CUP1, ZRT1, ENA genes), members of the major facilitator superfamily, and in genes involved in drug response (PDR3, SNQ1, QDR1, RDS1, AYT1, YAR068W). We therefore used halo assays to investigate the response of these strains to three different fungicidal drugs (cycloheximide, clotrimazole, sulfomethuron methyl). Strains with fewer copies of the CUP1 loci showed hypersensitivity to sulfomethuron methyl.
Conclusion
Microarray karyotyping is a useful tool for analyzing the genome structures of wine yeasts. Despite only small to moderate variations in gene copy numbers between different wine yeast strains and within different isolates of a given strain, there was enough variation to allow unique identification of strains; additionally, some of the variation correlated with drug sensitivity. The relatively small number of differences seen by microarray karyotyping between the strains suggests that the differences in fermentative and organoleptic properties ascribed to these different strains may arise from a small number of genetic changes, making it possible to test whether the observed differences do indeed confer different sensory properties in the finished wine.
==== Body
Background
Winemakers have long noted that different strains of wine yeasts, even when used to ferment the same juice under identical conditions, can yield very different wines in terms of sensory characteristics [1,2], presumably as a result of variations in the strains' fermentative properties [3-7]. Previous studies have demonstrated genetic diversity among both commercial [8] and wild [9]Saccharomyces cerevisiae wine yeast strains, and it has been hypothesized that this genetic diversity may, at least in part, be a root cause of their differing fermentative and sensory qualities [10-16]. In addition to a large amount of diversity in mitochondrial DNA [17], some wine yeast strains have been shown to be genetically unstable [18,19], with varying ploidy levels [20,21], multiple chromosomal aneuploidies [21-23], and chromosome length polymorphisms [21,22,24-26]. In addition, the genomes of a few wine yeast strains appear to have arisen from a hybridization event between two species, S. cerevisiae and S. bayanus[27], similar to that of the lager yeast S. pastorianus [28].
Many studies have shown that extensive genome rearrangements occur in organisms during adaptation to changing environments [29-31]. It is also likely that any source of cellular stress, such as an environmental insult, may trigger a "genome shock" response [32,33], leading to genome rearrangements. Although some genome changes observed in yeast populations are on a small scale (one to a few nucleotides), many changes appear to be on the scale of entire genes or even whole chromosomes, resulting in alterations of their copy numbers [31,34]. These features may be a result of uneven mitotic crossing over [35,36], gene conversion or break-induced replication [18,37], or Ty transposon-mediated chromosomal translocations[25].
Wine-making has been an ongoing human activity for several millennia [38-40]. Although the wine strains we have examined in this paper are commercial strains that have been propagated industrially for only the last ~40 years, each of their progenitors were distinct wild strains that had presumably been selected for many hundreds, perhaps thousands, of years to confer specific organoleptic traits. Likewise the laboratory S288C strain has been propagated as a pure culture for ~70 years; while its exact background is unknown and may include both wild and baking yeasts, it is almost certain that S288C did not derive from any commonly-used wine yeast strains [41]. At a generation time of 1.5 hours, however, 70 years can represent thousands (possibly hundreds of thousands) of generations of divergence between commercial wine yeasts and their laboratory "cousins". The genomic differences that exist among these different wine yeast strains, and between wine and laboratory yeasts, may have arisen in response to the unique selective pressures that each has encountered during its propagation. Detailed characterization of these differences may thus shed light on which biochemical pathways and cellular processes play important roles in determining the specific fermentative qualities (and the resultant wine sensory characteristics) of a particular wine yeast strain.
The recent development of DNA microarrays has enabled an unprecedented level of genome-scale research on both gene expression patterns [42,43], and also on genomic DNA copy number changes and genome rearrangements through the "array-CGH" ("aCGH") technique [44,45], which we will refer to as "microarray karyotyping" or "array karyotyping". This technique can be used to determine copy number changes for every gene of a given species (whose genome has been sequenced) in relation to any other strain of that species, giving information on whole or partial chromosome aneuploidies, non-reciprocal translocations and isolated gene deletions or amplifications. Using human cDNA microarrays, this technique has helped elucidate genomic copy number changes in cancer cells during tumor progression [46-48]; using S. cerevisiae-based microarrays, it has also been employed to discover non-reciprocal chromosomal translocations that occurred in yeasts evolved to tolerate low glucose concentrations [31]. A number of previous papers have demonstrated that aCGH data accurately reflects genome changes. For example, Pollack et al.[45], using human cDNA microarrays, demonstrated good correlation of copy number of known human X-chromosome aneuploid lines (from 0 to 5 X chromosomes per diploid genome) with aCGH data. Likewise for yeast microarrays, standard PCR [49,50] or quantitative real-time PCR [51] has been used to validate either deletions or amplifications predicted by aCGH data. Also, Dunham et al. [31] and Winzeler et al. [52] have used DNA sequencing to validate rearrangements (with associated copy number changes) or single-nucleotide polymorphisms, respectively, to corroborate their aCGH results.
Until recently, studies of the genetic diversity and instability of wine yeasts (reviewed by [8,50]) were performed using either classical genetic means, pulsed-field gel analysis, or by molecular techniques such as PCR, Southern blotting and restriction-fragment length polymorphisms. These studies did not, however, utilize whole-genome platforms such as microarrays, and thus the results were confined to only certain regions of the genome. Global gene expression patterns in wine yeast strains, deduced by microarray analysis, have been recently reported [53-55], but these studies did not include microarray karyotyping of the wine yeasts.
Microarray karyotyping has also been used to explore the genomic diversity of different S. cerevisiae strains [50,52,56-58] as well as the genomic architectures (relative to S. cerevisiae) of the hybrid organism S. pastorianus [51] and the sensu stricto group of closely-related Saccharomyces species [59]. However, none of these studies have utilized microarray karyotyping to assay the genomes of commercially-used wine yeasts for both copy number increases and decreases relative to the sequenced S288c laboratory strain. As part of larger gene expression studies, Lashkari et al. [57] and Primig et al. [58] generated array karyotype data, relative to S288C, for the non-S288C-derived laboratory S. cerevisiae strains Y55 and SK1 respectively; however these authors did not analyze any wine strains by array karyotyping. Filter-based whole-ORF genome arrays have also been used to investigate global gene expression patterns in a wine yeast [60]. These authors also re-hybridized a filter array using labeled genomic DNA to obtain gene copy numbers changes in the wine strain relative to the laboratory strain, but the data as presented did not show many interpretable changes in the wine strain other than a decrease in Ty1 copy number relative to the laboratory strain; a small amount of the same data generated by [60], restricted to chromosome VIII, was also presented in a separate paper [50].
Winzeler et al. [52] performed a type of array karyotyping on the Y55 strain, as well as on a wine strain, by using high-density short (25-mer) oligo arrays; however, their technique only allowed the detection of complete deletions at the whole gene or chromosome level, i.e., they could not detect amplification or hemizygous depletion of genes or chromosomes (e.g., the increase or decrease of just one copy of a gene or chromosome in a diploid cell, giving either three or one copies, respectively). It is important to note that whole-ORF arrays allow hybridization of even a fairly diverged gene (20 – 25% nucleotide divergence) to the target ORF on the microarray due to the long length of the target. Such divergence is much more likely to occur in non-S288C strains such as wine strains. 25-mer oligo arrays, however, are exquisitely sensitive to single-nucleotide changes; this is useful in some regards, but less so when widely-diverged strains are being analyzed. Infante et al. [56] performed a very detailed microarray karyotyping study of the genomic differences between two "flor" yeasts used in the production of sherry wines; however, they only compared the two "flor" yeasts to each other, and not to the sequenced S288C laboratory strain. Finally, both Edwards-Ingram et al. [59] and Bond et al. [51] used S. cerevisiae spotted ORF microarrays to look at the whole genome organization of other species that are closely related to S. cerevisiae, but did not assess the genomes of wine yeasts.
We describe here the use of microarray karyotyping to explore global changes in the genomic DNA of four commonly-used commercial wine S. cerevisiae strains relative to each other and to the sequenced S. cerevisiae laboratory strain (S288C), including the analysis of three independent isolates of each of the four wine strains. We report on the characterization of genomic similarities and differences between the wine strains and the S288C laboratory strain, as well as between the different wine strains themselves and between different independent isolates of the same strain.
Results and discussion
Microarray-based karyotyping
In the microarray karyotyping protocol employed here, genomic DNA from an experimental strain (i.e., a commercial wine yeast strain) was labeled with the fluorescent dye Cy5 (red), while genomic DNA from a wild-type (WT) reference strain (the sequenced haploid laboratory S. cerevisiae strain S288C) was labeled with the fluorescent Cy3 dye (green). The two labeled samples were mixed, then competitively hybridized to a spotted DNA microarray in which each spot is composed of PCR-amplified DNA corresponding to one gene of the reference S288C yeast S. cerevisiae [31,61]. Under standard conditions [31,61], hybridization to the microarray spots will generally occur between DNA molecules with 75% or more sequence identity. All experiments were performed in duplicate, i.e., the labeled DNA was hybridized to two separate microarrays. Note that the raw data for all microarray hybridizations reported in this paper can be downloaded from the website as well as from the Stanford Microarray Database [62,63].
Assuming that the experimental strain's genome has not experienced significant sequence divergence (i.e., greater than 25%) from that of the laboratory reference strain – a supposition which appears to be true for most S. cerevisiae wine yeast strains, which have an average of greater than 99% identity to the S288C strain [64] – any DNA copy number change in a single gene within the experimental strain will be detected by a deviation of the R/G ratio for that gene from the expected 1:1 ratio. If there are proportionally more copies of the gene in the experimental strain relative to the reference strain, the R/G ratio will thus be higher than 1:1 (ideally, it will be 2:1 for a duplication event, 3:1 for a triplication, etc.). Likewise, if there has been a deletion of the gene in the experimental strain relative to the reference strain, or if there are more copies of the gene in the reference strain than in the experimental strain (i.e., the experimental strain has experienced a "depletion" of the gene in terms of copy number), the R/G ratio will be less than 1:1 (ideally, it will be 1:2 for a single duplication in the reference strain relative to the wine strain, or for a deletion in the experimental strain of a single-copy gene in the reference strain, and so on). Thus, with only one hybridization, the DNA copy number of all 6,000 genes of a given strain can be determined relative to the reference strain.
Because microarray karyotyping detects proportional changes in DNA copy number relative to a reference strain, genome rearrangements that result in net copy number alterations can be easily detected with this method, including gene duplications or deletions, as well as larger-scale rearrangements such as non-reciprocal translocations, loss or amplification of large chromosomal regions, and ploidy changes for a single entire chromosome (aneuploidy). These large-scale rearrangements are seen as large regions of the chromosomes for which a similar bias in the ratios exists for all the genes in the region. Thus, despite the fact that complete genomic sequencing of every wine yeast strain is impractical, a great deal of information about the genomic structure of wine strains relative to laboratory strains and to each other can be achieved by microarray karyotyping.
Strains used for analysis
We chose four different commercial wine yeast strains for this study, using independent isolates from each of three different commercial and/or academic sources as shown in Table 1. The four yeast strains used were Montrachet, Pasteur Red, Pasteur Champagne, and Prise de Mousse (see Table 1 and Methods for complete list). These are among the most commonly used commercially-produced yeast strains in winemaking. Each strain has distinct fermentative qualities, and each is thought to bring unique flavor and other sensory components to the finished wine product [65,66]. Independent isolates of each strain were obtained from two major wine yeast producers, Lalvin and Red Star, as well as from the wine yeast strain collection at Univ. of Calif., Davis (Table 1).
Table 1 Wine Yeast Strains used in Study
Wine Yeast Type a Lab Name b Product Name c Source d Wine Use e
1. Montrachet GSY1 Montrachet Lalvin White, Red
GSY2 Montrachet Red Star
GSY3 Montrachet (UCD522) U.C. Davis
2. French Red GSY4 Pasteur Red Lalvin Red
GSY5 French Red (UCD725) U.C. Davis
GSY6 French Red (UCD904) U.C. Davis
3. Past. Champagne GSY7 Champagne Lalvin Sparkling
GSY8 Past. Champ. Red Star
GSY9 Past. Champ. (UCD595) U.C. Davis
4. Prise de Mousse GSY10 EC1118 Lalvin White, Red
GSY11 Premier Cuvee Red Star
GSY12 Prise de Mousse (UCD819) U.C. Davis
aGeneral category of wine yeast. bName used in laboratory collection. cName used by source (manufacturer); numbers in parentheses indicate the U.C. Davis strain collection name. dCommercial source of strain; see Methods for contact information. eGeneral type(s) of wines for which strains are used.
The genomes of all four wine yeast strains are highly similar to that of S. cerevisiae S288C
We performed microarray karyotyping, as described in Methods, on each of the wine yeast strains. Because some of the strains are listed by suppliers as either S. bayanus or S. cerevisiae/S. bayanus, we also performed microarray karyotyping experiments using genomic DNA from a known type strain of S. bayanus to determine what possible signal may be generated when hybridizing S. bayanus genomic DNA to S. cerevisiae microarrays. We performed a similar experiment using S. cerevisiae DNA as a control (a "self-self" hybridization).
Figure 1 shows a graphical representation of the microarray karyotyping data for the S. cerevisiae and S. bayanus strains presented as "karyoscope" diagrams, made using the Java TreeView program [67]. To make a karyoscope diagram the hybridization ratio for each gene is mapped onto its corresponding position on each chromosome of the reference strain of S. cerevisiae. The height of each red or green vertical bar is proportional to the log2 of the red:green (R/G) ratio for a gene; if the ratio is greater than 1 (i.e., a positive log2 value), the bar will be red and drawn above the chromosome; a red bar thus represents an over-representation (amplification) of that gene in the wine strain relative to S288C. For R/G ratios less than 1 (i.e., a negative log2 value), the bar will be green and drawn below the chromosome; a green bar thus represents an under-representation (deletion or depletion, i.e., lower copy number) of that gene in the wine strain relative to S288C.
Figure 1 Karyoscope views of S288C-S288C and S. bayanus-S288C microarray hybridizations. Microarray hybridizations were performed as described in the text. In both panels the S. cerevisiae laboratory strain S288C was used as a reference sample, labeled with Cy3 dye (green). In panel A, the Cy5-labeled (red) DNA was also S288C, giving a self-self hybridization. In panel B, the Cy-5 labeled DNA was S. bayanus. The labeled DNAs were competitively hybridized to spotted microarrays bearing full-length ORFs from the S288C S. cerevisiae strain; the data thus obtained is displayed here in graphical form as a karyoscope. Red bars indicate red:green ratios above 1.0 and are graphed on a log scale; green bars indicate red:green ratios below 1.0, also graphed on a log scale. The chromosomes are shown in each panel in numerical order with chromosome 1 at the top and chromosome 16 at the bottom, and are aligned by their centromeres, with their left arms extending to the left. Both karyoscopes are drawn to the same scale, i.e., the bar heights in panels A and B proportionally represent the same amplitude of change.
In the "self-self" hybridization control, the genomic DNA from the reference S. cerevisiae laboratory strain was labeled independently with Cy3 and Cy5 and then competitively hybridized against itself. We observed robust hybridization across the entire array, and after a global mean normalization (see Methods) there were equal red and green hybridization signals (R/G ratios of 1), with no obvious deviations for any particular gene, as expected (Fig. 1A). We found, however, that the S. bayanus DNA did not hybridize to the arrays in the same manner as did the S. cerevisiae genomic DNA. The S. bayanus hybridization signals (Fig 1B) were significantly weaker and much more variable across the array, as compared to the S. cerevisiae self-self hybridization (Fig 1A).
For all 12 of the wine yeast isolates, we observed hybridization patterns much more similar to that of the "self-self" experiment, i.e., with robust signals across most of the array, and the majority of ratios near a value of 1.0. This indicates that each of the wine strains possesses an essentially complete complement of the S. cerevisiae genome, with no observed aneuploidy. Note that we cannot rule out that there may be additional genomic DNA in these wine strains that comes from other Saccharomyces species (i.e., if the strain is a hybrid organism) whose hybridization signal would be overshadowed by the signal generated by the S. cerevisiae genomic complement. However, we do not think this is likely, as control spots on the microarrays containing S. bayanus genomic DNA did not show a hybridization signal, which we consistently observe when using known S. cerevisiae – S. bayanus hybrid yeasts (B. Dunn, unpublished). Thus, we do not expect any potential non cerevisiae DNA in any of these wine strains to have any great effect on our observed ratios.
Genomic differences between wine yeast strains and the laboratory strain
We observed a distinct set of genomic variations that occur in every wine strain relative to the laboratory strain; these are shown in Figure 2, and an abbreviated list of these variations is given in Table 2. Figure 2 represents a consensus karyoscope-type plot generated by the program CGH-Miner [68], where the data from the 4 wine strains (taken as averages of the data for the 3 isolates of each strain) have been consolidated into statistically-significant amplifications or depletions relative to multiple self-self hybridizations of the laboratory S288C strain.
Table 2 "Commercial Wine Strain Signature" Gene Lista
Sig. amplified in at least 3 strains Montrachet French Red Champagne Prise de Mousse
YAR066W 0.84 1.43 0.62 0.19
YAR068W 1.05 1.67 0.91 0.52
YAR069C 1.11 1.57 1.09 0.59
YAR070C 1.14 1.62 1.16 0.63
YAR071W PHO11 1.01 1.37 1.26 0.57
YAR073W IMD1 0.99 1.41 1.34 0.56
YAR075W 0.98 1.47 1.46 0.56
YPL276W 0.59 1.36 1.64 0.75
YPL275W FDH2 0.63 1.39 1.68 0.78
YPL274W SAM3 0.58 1.06 1.31 0.66
YPL273W SAM4 0.52 0.81 1.04 0.57
YPL272C 0.49 0.59 0.79 0.47
Sig. depleted in at least 3 strains Montrachet French Red Champagne Prise de Mousse
YAR007C RFA1 -0.44 -0.73 -0.80 -0.31
YAR008W SEN34 -1.05 -1.53 -1.31 -0.69
YAR014C BUD14 -1.03 -1.25 -0.99 -0.65
YAR015W ADE1 -0.54 -0.63 -0.64 -0.38
YBL007C SLA1 -0.51 -0.61 -0.81 -0.39
YBL006C LDB7 -0.86 -0.99 -1.02 -0.66
YBL005W PDR3 -0.88 -1.07 -1.10 -0.67
YBR013C -0.97 -1.58 -1.04 -0.45
YHR052W CIC1 -0.39 -0.89 -1.40 -0.78
YHR053C CUP1-1 -0.46 -1.32 -1.91 -1.07
YHR054C -0.56 -1.65 -2.16 -1.29
YHR055C CUP1-2 -0.56 -1.43 -1.98 -1.26
YHR056C RSC30 -0.49 -1.10 -1.52 -0.96
YJR024C -0.59 -0.62 -0.49 -0.30
YJR025C BNA1 -0.91 -1.13 -0.84 -0.53
YJR026W -1.47 -1.92 -1.41 -0.90
YJR030C -0.95 -1.65 -1.32 -0.67
YJR031C GEA1 -0.41 -0.90 -0.78 -0.32
YJR032W CPR7 -0.16 -0.31 -0.39 -0.15
YJR033C RAV1 -0.14 -0.32 -0.39 -0.11
YML047C PRM6 -0.46 -0.89 -0.91 -0.25
YML046W PRP39 -0.51 -1.16 -1.19 -0.31
YML043C RRN11 -0.49 -0.99 -1.05 -0.33
YML042W CAT2 -0.83 -1.33 -1.16 -0.51
YML041C VPS71 -0.93 -1.35 -1.18 -0.58
YML038C YMD8 -0.84 -1.39 -1.16 -0.46
YML037C -0.47 -0.97 -0.93 -0.23
YMR043W MCM1 -0.30 -0.28 -0.30 -0.23
YMR044W IOC4 -0.81 -0.98 -0.82 -0.59
YMR046W-A -1.14 -1.58 -1.29 -0.78
YMR047C NUP116 -0.70 -0.96 -0.88 -0.56
YMR048W CSM3 -0.69 -0.89 -0.75 -0.46
YMR049C ERB1 -1.07 -1.45 -1.03 -0.66
YMR052W FAR3 -1.17 -1.87 -1.28 -0.63
YMR052C-A -0.69 -1.26 -0.89 -0.36
YOL165C AAD15 -1.09 -1.03 -0.74 -0.80
YOL164W -1.33 -1.24 -0.96 -0.80
a The data shown in this table are the actual mean log2(R/G) ratios for each gene, calculated by averaging the values from the six arrays (duplicate arrays of each of the three isolates of each strain) hybridized per wine strain. This list represents the wine-specific subset of the total set of genes with significant copy number changes, as described in the text. Values shown in bold text were determined by the CGH-Miner program to be significantly depleted or amplified as indicated. Values not significantly depleted or amplified are shown in non-bold italicized text.
Figure 2 Changes in Gene Copy Number Shared by all Wine Strains: the "Commercial Wine Yeast Signature". A consensus plot pooling the results of all the wine strains relative to the lab strain S288C was generated by the program CGH-Miner [68]. This plot is similar to a karyoscope in that it displays values along each chromosome, and the chromosomes are shown in numerical order from top to bottom. However, in this plot, a determination of which regions were statistically significantly altered in copy number for each wine strain (i.e., averaged array values for each of the three isolates within a strain) relative to the laboratory S288C strain were determined. For each of these regions the number of wine strains showing a significant change were then added together to give a metric of how well-shared among the wine strains that particular change was. This is shown as tall red bars for significantly amplified regions among all wine yeast strains relative to the laboratory strain, and as tall green bars for significantly depleted regions in all the wine strains relative to the laboratory strain. Each gray "shadow line" above or below the chromosome represents 20% of the number of wine strains showing a significant change in that region (see legend). The group of changes comprising the tallest (>75%) red and green bars represents the "commercial wine strain signature"; these are listed in Table 2.
The most distinctive genomic variations seen in Fig. 2 are the many prominent green bars which are seen on most of the chromosomes. These shared green bars represent genes or regions in the wine strains, taken together as a whole, that are depleted in copy number relative to the laboratory strain. In most cases these shared depleted regions consist of genes that are moderately or highly repeated in the laboratory strain, with the majority representing Ty1 elements (a yeast retrotransposon family). Also depleted are the tandemly-repeated ASP3 (asparaginase) genes on chromosome 12, and the tandemly-repeated ENA (sodium transport) genes on chromosome 4; note: general descriptions and functions for genes described in this paper were obtained from the Saccharomyces Genome Database (SGD) [69,70]. The CUP1 (copper binding) tandemly-repeated genes are also significantly depleted in 3 of the 4 strains (Fig. 2, Table 2) and additionally show some intra-strain variation in copy number (Fig. 5B); depletion of the CUP1 genes has been observed previously in a wine strain [50]. Because these sequences are all moderately repetitive and highly conserved, we can only interpret these results to indicate that in all of the wine strains we investigated, there are either fewer or no copies of these repeated genes relative to the laboratory strain.
Because the CGH-Miner program utilizes a moving-average method for calculating significance, copy number changes of single isolated genes are not detected. However, inspection of clustered data and karyoscopes allowed us to identify single genes whose copy numbers are highly altered relative to the reference strain. Among these, we found that the single-copy drug response genes SNQ2 and PDR15 are depleted in all of the wine strains we examined (Fig 5B). In addition, PDR3, a single-copy zinc-finger transcription factor involved in pleiotropic resistance to drugs [71], appears to be entirely deleted in all four wine strains due to the very low R/G ratios it exhibits (Table 2, Fig. 5B). This gene resides adjacent to one of the Ty1 elements that is depleted in the wine strains. Additional altered-copy number genes identified by inspection of karyoscopes are listed in Fig. 5B.
In addition to the shared depleted regions, there are shared regions that appear to be slightly, but consistently, amplified in all four wine yeast strains; the most prominent is located at the right end of chromosomes 1; other regions are on the right arms of chromosomes 2 and 12, and the left end of chromosome 16 (Fig. 2, Table 2). These shared amplified regions contain some genes that are members of homologous gene families in the reference strain, such as the IMD genes (involved in GMP production) and PHO genes (acid phosphatases), but they also include unique genes such as SAM3 and SAM4 (S-adenosyl methionine transport and S-AM metabolism, respectively), PET8 (mitochondrial S-AM transport), RDS1 (response to xenobiotic drugs, including fungicides, herbicides, and pesticides), TPO1 (multidrug and polyamine transporter and drug detoxification), AAD3 (aryl-alcohol dehydrogenase), ADH7 (alcohol dehydrogenase), and FDH1 and FDH2 (formate dehydrogenase) (Table 2, Fig. 5B). Also occurring in an amplified region is the un-named and uncharacterized gene YAR068W (Table 2, Fig. 5B). This gene codes for a membrane protein which when overexpressed results in resistance to the drug ketoconazole [72]; additionally, it has been found to have one of the most elevated Ka/Ks ratios among the genomes of the Saccharomyces sensu stricto group [73], signifying that it may be a gene that is actively undergoing positive selection [74]. A homologue to this gene, YHR214W-A, is also amplified in most of the wine yeast strains.
These results indicate that all of the wine strains that we examined have experienced a characteristic increase or decrease in copy number, relative to that of the laboratory strain, of the genes listed in Table 2. Although some of these changes, particularly the low copy number (or absence) of the ENA and ASP3 genes and of Ty1 elements, are consistently observed in other industrial and non-S288C laboratory yeast strains ([51,52,57-60], B. Dunn, unpublished), most of the other changes, particularly the amplifications, are novel and thus far appear to be unique to commercial wine yeasts.
Genes whose copy number changes are shared by all wine strains
Interestingly, there are many transporter and permease genes, including some involved in drug response, that show altered copy numbers in the all the wine strains relative to the laboratory strain. As described above, PDR3, PDR15, RDS1, SNQ2, TPO1, and YAR068W show alterations in their copy numbers that are shared by all the wine strains. In addition, HXT9 and HXT11 are depleted in three of the four strains (Fig. 5B), and AYT1 is depleted in 2 strains and amplified in the other two. All of these genes are involved in some aspect of drug response [71,72,75-78].
It thus appears that a major group of shared genomic differences found among all the wine strains is composed of genes, especially transporters and permeases, involved in drug resistance pathways. Since vineyards are often treated with herbicides, pesticides, and fungicides, wine yeasts may be exposed to these agents on a routine basis and their genomes may have evolved to better tolerate such exposure. Alternatively – and perhaps more likely given the fact that these commercial strains may have been isolated prior to the widespread use of herbicides, pesticides and fungicides – the variations in transporter gene copy number seen here may have arisen to fine-tune the fermentation properties of the wine yeasts by altering the types and amounts of fermentable substrates that are taken up. These variations may also reflect an increased tolerance to, or excretion of, plant phenolic compounds. Thus, any changes in drug response pathways in these strains may be a secondary effect and not due to direct selection for drug resistance or sensitivity. This view is supported by reports that alterations in PDR1 and PDR3 gene expression, as well as mutations in the ABC transporter genes that they regulate, result in better fermentation performance by sake yeasts [79,80].
Inter-strain variation
We also observed inter-strain genomic variations that are specific to each wine strain. Examples of such inter-strain genomic variation are highlighted as regions circled in black on the Karyoscopes in Figure 3A–D. For an idea of scale in these karyoscopes, the tallest bars on the figure represent ratios of approximately 2:1 (2-fold enhanced) for red, and 1:16 (16-fold depleted) for green. Each panel in Fig. 3 represents the average of the hybridization log ratios for all three independent isolates of the given strain; since each isolate's DNA was hybridized in duplicate, each panel thus represents the average of six arrays, making the data very robust. Some examples of inter-strain variation are also shown in more detail in Figure 5A, in which the hybridization data are represented as colored boxes, with the most intense green representing severe depletion (R/G ratios much less than 1), black representing no change in copy number (R/G ratio of approximately 1), and the most intense red representing significant amplification (R/G ratios greater than 1).
Figure 3 Wine strain karyoscopes. Microarray ratios from each of the three isolates of a given wine strain were averaged to yield an "average" microarray karyotype for each wine strain. Chromosomes 9 through 16 for the 4 strains are shown in panels A – D as follows: A. "Mont" = Montrachet (average of strains GSY1-3). B. "PDM" = Prise de Mousse (average of strains GSY10-12). C. "Champ" = Pasteur Champagne (average of strains GSY7-9). D. "Red" = French Red (average of strains GSY4-6). Circled regions highlight areas that vary in characteristic ways between the different wine strains. All karyoscopes are drawn to the same scale.
One distinct inter-strain variation is seen in the copy numbers of HXT (hexose transporter) genes. As shown by the black-circled regions on chromosomes 9 and 10 in Fig. 3A–D, and in the top panel of Fig. 5A, the Prise de Mousse isolates show no depletion of the HXT9, HXT11, and HXT12 genes, whereas the Pasteur Red isolates show a slight depletion of these genes. On the other hand, the Montrachet and Champagne isolates show extreme depletion of these genes. Since these three HXT genes are all virtually identical to each other and can cross-hybridize, this result is most likely explained by a loss of one or more of the three genes in the depleted strains. It is impossible, however, to know which particular gene(s) are missing without further investigation.
Another class of inter-strain differences is of particular interest because the differences are seen in genes that are single-copy in the laboratory strain. Depletions of these genes in a given wine strain as measured by array karyotyping may thus reflect a true deletion of that gene relative to the other wine strains, rather than just a decrease in copy number. Three instances of this type of inter-strain variation were found: the AYT1 gene (chromosome 12), the ENB1 gene (chromosome 15), and the un-characterized ORF YPL257W (chromosome 16). The circled regions on the relevant chromosomes in Fig. 3A–D highlight these single-gene strain differences, while Fig. 5A shows the data in more detail for these three genes. AYT1 encodes an acetyltransferase that was first characterized in Fusarium and subsequently identified in S. cerevisiae by homology [75]. It plays a role in de-toxifying endotoxins of the tricothecene family in Fusarium, but its function in S. cerevisiae is unknown. Although it has been shown that the S. cerevisiae AYT1 gene product can acetylate tricothecene in vivo, cells that are deficient for this gene show no increased sensitivity to the compound, and in fact show no phenotype at all [75]. ENB1 encodes an enterobactin transporter of the major facilitator superfamily involved in iron uptake [81], while YPL257W has no known function, although its predicted product has homology to membrane proteins.
Intra-strain variation
We observed relatively little genomic variation between independent isolates of each of the different strains. A major exception to this was seen, however, among the three different isolates of Montrachet. As shown in Figure 4, the Montrachet isolate obtained from Lalvin (GSY1) has a fairly extensive region near the left end of chromosome 7 that exhibits copy-number depletion (relative to the S288C reference strain), seen as a distinct cluster of green bars (a thick black line underlines this region). This region spans approximately 37 contiguous genes, all of which show a ratio lower than 1 (shown as green bars). This is in contrast to the other two Montrachet isolates (GSY2 and GSY3 from Red Star and UCD, respectively), which show no copy-number depletion in this region (Fig. 4). Because so many contiguous genes are involved in this block of genes, and because almost all of the genes exhibit a similar direction and magnitude of deviation, we believe that this region has been lost from either one or both copies of chromosome 7 in the (presumably diploid) Lalvin isolate, yielding a net depletion in the copy numbers of these genes relative to the reference strain. It is formally possible that sequence divergence in the region caused the decreased hybridization to the reference strain, but this is unlikely because sequence divergence rarely, if ever, occurs in such a dense and homogeneous pattern, nor is it ever constrained to only one region of the genome. Since there are at least two essential genes in this region (see below), it is most likely that only one of the chromosomes in this presumably diploid strain has lost this region.
It is likely that the region of depletion extends out to the end of the chromosome, despite the fact that the two distal-most genes appear as red (amplified) bars; both genes are sub-telomeric genes that are repeated in the genome and it is thus not possible to determine whether the chromosome 7 copies are specifically present or not. Therefore, the event that caused the depletion of this large region of chromosome 7 in the Lalvin isolate could have been a non-reciprocal translocation resulting in the net loss of the distal end of chromosome 7. This is supported by the fact that the inner (proximal) boundary of the depleted region occurs exactly at a tRNA (tV(AAC)G3). It has been shown that non-reciprocal translocations and other types of gross chromosomal rearrangement events that result in better-adapted genotypes occur frequently during adaptive evolutionary growth, and that these rearrangements are almost always bounded by tRNAs or Ty1 elements [31,56,82-84].
Figure 4 Karyoscopes of Chromosome 7 from individual Montrachet isolates. The larger black bar on the left, placed under the left end of chromosome 7 from the Lalvin Montrachet isolate (GSY1), shows a 37-ORF region that has been depleted (or deleted) with respect to the S288C laboratory strain and with respect to the chromosome 7's of each of the other two Montrachet isolates. The smaller black bar on the right, under the UCD522 (GSY3) chromosome, indicates a region of 3 ORFs (including MAL11 and MAL13) that are present at "wild-type" (S288C) copy number in the UCD522 strain but are depleted or deleted in the other two Montrachet strains. Note that this same MAL region is also present at "wild-type" copy number in UCD725, but is depleted or deleted in all remaining wine yeast strains. Again, all karyoscopes are drawn to the same scale.
Figure 5 Unique combinations of genomic differences between different strains. Panel A: Selected portions of the averaged data shown in Karyoscope form in Figure 3 are shown here in color-block form. To generate this figure the averaged data for all four wine strains were sorted for the Montrachet strain from highest green values (greatest negative number) to highest red values (greatest postitive number) and visualized as a "clustergram" using Java TreeView (note that the data were sorted, but however, were not clustered). The topmost (most green) 25 genes are shown in the top panel; the lower two panels show the sorted data surrounding two of the genes (AYT1 in the middle panel, YPL257W in the lowest panel) whose copy number varies greatly on a strain-to-strain basis. Red asterisks mark the genes whose copy numbers vary in unique combinations between the different strains. Panel B shows a matrix of gene copy number changes for each of the 12 wine strain isolates, along with the drug response profiles of each strain isolate at the bottom. Within the copy number data portion, a dark red cell with a "++" symbol indicates a red:green log ratio greater than 0.8, light red with "+" indicates a ratio between 0.2 and 0.8, and gray with "+/-" indicates a ratio between 0.2 and -0.2. Likewise dark green with "--" indicates a red:green log ratio less than -0.8, light green with "-" indicates a ratio between -0.2 and -0.8.
The depleted region of Chromosome 7 in the Lalvin Montrachet isolate extends from Open Reading Frame ("ORF") YGL226W through YGL261W. Included in this region are three genes, two essential and one non-essential, involved in protein-nucleus export (BRR6, CSE1, KAP114), two genes involved in cell cycle control (SAP4, DOC1), as well as genes involved in secretion (SEC15), and respiratory growth (HAP2, MTO1). Also included in this region is a transporter protein for zinc uptake (ZRT1). Of possible significance in terms of winemaking are genes that are involved in glucose/fructose (HXK2) [55] and alcohol (ADH4) metabolism [85], as well as one gene (YGL258W) involved in velum formation in a flor (Sherry) yeast strain. Interestingly, YGL258W and ADH4 are both induced under zinc-deficient conditions [86,87], which may have significance with respect to the associated depletion of ZRT1. If significant fermentative or organoleptic differences exist between the Lalvin Montrachet isolate and the other two Montrachet isolates, it would be interesting to determine whether adding back genes from the depleted region (e.g. on a stable single-copy CEN plasmid) to the Lalvin isolate would cause its fermentative and/or organoleptic characteristics to now mimic those of the other two isolates.
A second intra-strain difference we observed is a depletion of four genes near the right end of chromosome 7 that is seen in both the Lalvin and Red Star Montrachet isolates but not in the UCD522 Montrachet isolate (Fig. 4, shorter black bar underlining the region at the right telomere). This group of four genes includes a maltose permease gene (MAL11) and a maltose-activator protein (MAL13); loss of these genes usually indicates a defect in maltose fermentation, but they also appear to be involved in drug response pathways since deletions of MAL11 lead to nystatin sensitivity [88]. We observed that this same group of MAL genes is depleted in all of the other strain isolates except in the French Red isolate UCD725 (GSY5). In other words, of all 12 isolates that we studied, only UCD522 (GSY3) Montrachet and UCD725 (GSY5) French Red do not show a depletion of this MAL region relative to the reference S288C strain (Fig. 3A–D, Fig. 5B).
Finally, two genomic differences were seen in the UCD725 French Red isolate (GSY5) when compared to the other two French Red strains (GSY4 and GSY6). First, as mentioned just above, the same group of MAL genes that is present at wild-type copy number in the UCD522 Montrachet isolate is also present in this UCD725 isolate, but is depleted in the other two French Red isolates. Secondly, YPL257W is depleted in the French Red strains GSY4 and GSY6, but is present at wild-type copy number in GSY5 (data not shown). This gene, encoding a putative membrane protein, is one of the set of genes that exhibit a characteristic copy number genotype between the different wine yeast strains (see below), but it apparently varies in its genotype among the French Red isolates. Overall, however, aside from the fairly small differences among the Montrachet and French Red strains, there is remarkably little intra-strain genomic variation among the strains we studied, indicating that these wine strains are relatively stable genetically, at least as assayed by microarray karyotyping.
Transporters and permeases are part of the shared group of copy number differences, and also show inter- and intra-strain differences
Figure 5B shows a matrix of relative copy numbers for each of the 12 individual wine strain isolates. The first group of genes shown in the matrix represents all of the Major Facilitator Superfamily (MFS) transporter genes for which a large copy number deviation relative to the reference strain (defined as log2 (R/G) value greater than 0.8 or less than -0.8, corresponding to a 1.74 fold change) was seen for at least one isolate. A "-" or "--" sign indicates moderate or large depletion of the gene's copy number, respectively; assuming that these are diploid strains, a "-" would indicate a loss of one of the two copies and a "--" would indicate loss of both copies. Likewise, "+" or "++" indicates moderate or large amplification of the gene, respectively, which would indicate an increase from 2 to 3 copies for "+" and to 4 or more copies for "++". Remarkably, of the total of 27 MFS transporter genes annotated as such in the Saccharomyces Genome Database [69,70], almost half (13) show large deviations (as defined above) in copy number among the wine strain isolates; all 13 are shown in Fig. 5B. Most MFS transporters show a depletion in copy number relative to the wild-type strain; TPO1, which can transport cycloheximide and other drugs [78], is the major exception, showing amplification in all strains.
Figure 5B also shows copy number data for known permeases and other transporters, as well as any genes annotated as having a role in drug response that show a large deviation in at least one isolate. Finally, genes which do not fall into a transporter/membrane protein class, but which nonetheless stood out in the clustered data as being significantly altered in copy number relative to S288C are included (e.g., the CUP1 locus and members of the S-adenosylmethione metabolism pathway). At the bottom of the figure are shown the results of the drug halo assays for each of the isolates (see below).
Drug resistance/sensitivity phenotypes in wine strains
The prevalence of genes involved in drug response and drug resistance occurring in both the shared group of copy number changes, as well as in inter- and intra-strain differences, directed us to investigate the drug response profiles of the wine yeast strains. We chose three different drugs with which to assay the cells, representing three different general classes of drugs: cycloheximide, "CYH" (a protein synthesis inhibitor); clotrimazole, "CTZ" (member of the azole class of antifungal agents which inhibit ergosterol synthesis); and sulfomethuron methyl, "SMM" (a sulfonylurea herbicide). As described in Methods, halo assays were performed on all 12 wine strains listed in Table 1 as well as on six other control yeast strains mentioned in Methods. These additional six strains represent either control wild-type strains (FY1679, S288C, RW2802) or strains with single-gene deletions of various drug resistance genes whose copy numbers had been observed to be altered in the wine yeast strains: an AYT1 deletion strain (ScAYT1Δ; [75]), a PDR5 deletion strain (JG436; [75]), and a YPL257W deletion strain (GSY28).
Figures 6A and 6B show the actual halos on the plates, while Fig. 6C tabulates the halo results in chart form. The most notable feature of the chart is that the majority of the 12 wine strain isolates (Fig. 6A, 6C) are very sensitive to SMM compared to the WT diploid laboratory strain. However, there are three exceptions to the SMM sensitivity, with two of the three Montrachet isolates, Lalvin Montrachet (GSY1) and Red Star Montrachet (GSY2), as well as the UCD725 French Red isolate (GSY5), showing much less sensitivity or even low resistance to SMM.
Figure 6 Drug resistance phenotypes of wine strains. Halo assays were performed as described in the text to test the response of the wine strains to various drugs. Panels A and B show photographs of the halos; on the far left of each panel are shown the halo assays for the WT-diploid (FY1679, a S288C-based diploid, see Methods), against which all other strains were compared. CYH = 500 ng cycloheximide; CTZ = 5 μg clotrimazole; SMM = 20 μg sulfomethuron methyl. Panel A: strains are as described in Methods and Table 1; "WT-Dip" refers to FY1679. Panel B: Strains are as described in Methods; "WT-Dip" refers to FY1679; "YPL257 Δ" to GSY28; "ayt1 Δ" to ScAYT1Δ, "pdr5-" to JG436; and "WT-Hap" to S288C. Panel C: Graphical results of halo assays. Annular area (i.e., area of total halo minus area of disc) was calculated for each halo, then expressed as log2 of the ratio of the annular area to that of the WT-Dip (FY1679) annular area.
It can also be seen in Figs. 5B, 6A and 6C that the individual isolates of Champagne (GSY7-9) and Prise de Mousse (GSY10-12) wine strains show more consistency in their drug response profiles than do the individual isolates of Montrachet (GSY1-3) and French Red (GSY4-6). Interestingly, it is the individual isolates of both the Montrachet and the French Red wine strains that also showed the most variability in their genome structures (Fig. 5B).
Halo assays for the control yeast strains are shown in Fig. 6B. As expected, the pdr5 strain, known to be hypersensitive to drugs, is extremely sensitive to CYH and to CTZ. It is very resistant to SMM, however, with no halo at all forming around the SMM disc. The ayt1Δ strain's response to CTZ or CYH is similar to that of the WT diploid, but like the pdr5 strain it is very resistant to SMM. The YPL257WΔ strain shows moderate resistance to SMM compared to the WT diploid, with no variant response to CTZ or CYH. Note that both of the haploid wild-type control strains, S288C ("WT-Hap" in figure) and RW2802, show greater resistance to SMM than does the WT diploid strain; S288C is moderately resistant and RW2802 is extremely resistant, with no halo detected in the presence of SMM. This may indicate that haploid strains are for some reason more resistant to SMM than diploid strains. It also appears that strain differences (i.e., compare RW2802 and S288C) may affect the degree of resistance to SMM.
Conclusion
We have found that each of the three independent isolates of the four commercially-produced wine strains we examined (i.e., 12 strains total) are all S. cerevisiae strains, based on their array karyotypes showing that all have genomes highly similar to the standard laboratory S. cerevisiae strain (S288C). Two of the wine strains we examined – Prise de Mousse (also known as EC1118 or Premier Cuvée), and Pasteur Champagne – were listed in company catalogs as S. cerevisiae var. bayanus or S. bayanus, and the other two as S. cerevisiae var. cerevisiae or S. cerevisiae (e.g., see [89]). Our hybridization data thus show that there is great similarity between S. cerevisiae var. bayanus and S. cerevisiae var. cerevisiae, supporting the view that variety or race designations within S. cerevisiae should be abolished if there is no overwhelming phenotypic or molecular evidence to support their existence; this conclusion was also reached by Fernandez-Espinar et al. in a study comparing different isolates of commercial wine yeasts by restriction fragment polymorphisms, pulsed-field gels, and PCR analysis [90].
Contrary to previous results [23], none of the strains we examined showed any whole chromosome aneuploidies for S. cerevisiae chromosomes, although two of the strains we examined (UCD522 and UCD595) were identical to those used in that study. We know from our work with hybrid yeast strains that array karyotyping can easily detect whole chromosome aneuploidies (B. Dunn, unpublished), so we believe that the wine strains we examined in this work indeed do not contain any whole chromosome aneuploidies; we do not know the basis for the discrepancy other than that the previous study used indirect (genetic) means to detect the aneuploidies and it is possible that some of the markers behaved aberrantly.
We also found that the wine strains we examined all shared in common a set of genome copy number changes relative to the standard S288C strain. Many of these changes are either in Ty1 transposons, where copy number changes can be explained by differences in transposition rates, or in tandemly-repeated genes, where gene copy number changes can be explained by unequal cross-overs [35,36]. In fact, copy number depletion relative to the S. cerevisiae S288C strain of Ty1 elements and the tandemly-repeated ASP3 and ENA regions have been noted previously in various non-S288C yeast strains [51,52,57-60], B. Dunn, unpublished); they therefore appear to be changes shared by many, if not all, non-S288C yeast strains; although they are thus not wine strain specific, they are shared by all the wine strains we examined.
We therefore propose a "commercial wine strain signature" that comprises those genes whose copy number is altered in all the wine yeast isolates we examined relative to the sequenced S288C strain, but which excludes any of the genes whose copy numbers are altered in ale or beer strains or in other non-S288C laboratory strains such as SK1 and Y55. The final list of genes in the "commercial wine strain signature" is given in Table 2; it is comprised of a subset of those genes for which a statistically significant change in copy number, either depletion or amplification, occurred in at least 3 of 4 strains, as determined by the CGH-Miner program. All Ty1 genes, the ENA1, 2 and 5 genes, and genes from the ASP3 region on Chromosome XII have been eliminated from this list since they appear to be copy number changes shared by all non-S288C yeast strains and are not wine-specific.
Also of interest are the genes that show inter- or intra-genic variation, shown in Fig. 5B. These genes consist mainly of sets of transporter or sugar utilization genes residing at or near the telomeres. These genes can undergo a type of inter-chromosomal unequal cross-over or gene-conversion event through dispersed sub-telomeric sequences shared among the ends of the chromosomes [91], and thus can undergo changes in copy number relatively facilely. Although many other non-wine, non-S288C strains have been observed to have alterations (predominantly deletions) in sub-telomeric sugar utilization genes [52], none of them show the exact pattern of the "commercial wine strain signature". This is not surprising, since it might be expected that the exact suite of transporter and sugar utilization genes would vary in defined patterns among yeasts, depending on the particular substrate (e.g., grape juice, malted barley, laboratory defined media, or wheat flour) on which they had been selected for good growth performance. In fact, the wine strains are unique in having several sets of sub-telomeric genes, most notably those at the right end of Chromosome I (YAR064W – YAR075W), amplified relative to the S288C strain. On a blind assessment of a panel of karyoscopes generated from an ongoing study (B. Dunn, unpublished), which included a set of 6 additional commercial wine strains, as well as beer strains (one ale and one lager), laboratory non-S288C strains (Y55 and SK1), and S288C itself, we were easily able to identify the wine strains as a set distinct from the other strains. The assignment as a "wine" strain was based mainly on the presence of the amplified region at the right end of chromosome I, as well as small amplified telomeric regions on both ends of Chromosome III. In addition, a subset of the wine strains exhibited a characteristic depletion of either AYT1, ENB1 or YPL257W as discussed below. We thus believe that the list presented in Table 2 is a bona fide "commercial wine strain signature", at least for commercially-produced wine strains; the amplification of the right end of Chromosome I, which has never been reported before, is probably the most defining aspect of the "commercial wine strain signature". These particular genomic changes may be a result of long-term selection for strains that demonstrate good fermentative behavior and good organoleptic characteristics during wine-making. We expect that there may be slight changes made to the list, as more wine yeast array karyotype data are accumulated. We are planning to perform similar studies with "wild" wine yeast strains, as well as with more commercial wine yeasts in order to refine the list, if necessary.
We also observed internally-located single-copy genes (AYT1, ENB1, YPL257W) whose copy numbers vary either as part of the "commercial wine strain signature" or in inter- or intra-strain variation. It is possible that these genes have diverged in their sequences relative to the reference strain to a degree that hybridization no longer occurs, but it is more likely that rather than accumulation of multiple changes localized to just a few isolated genes (which would have to be the result of an extremely strong positive selection) among the whole genome, a more drastic change (either a deletion or amplification) has occurred to the gene. The mechanism by which this might happen is not known but could be explained by the presence of tandem repeats on either side of the gene, thus enabling loop-out or unequal cross-over events.
Many of the changes in genomic copy number in the "commercial wine strain signature", and in inter- and intra-strain differences, involve genes that encode transporter proteins or genes involved in drug resistance phenotypes in laboratory yeasts. Drug sensitivity results for the wine strains show that most are highly sensitive to SMM as compared to the laboratory strain, although it appears that Montrachet strains may not exhibit as high a level of SMM sensitivity as the other strains (Fig. 5B). When the results of the drug sensitivity tests are compared with the copy number genotypes of the transporter, permease, drug response and other variable genes (Fig. 5B), it is apparent that the situation is complex; among all of the genes that vary in copy number in these strains, only one gene's copy number status correlates well with an easily identifiable drug response – CUP1, the copper-binding yeast metallothionein gene. It appears that the nine wine strains (i.e., all except GSY1, 2, and 5) which are highly depleted for the CUP1 gene (which is tandemly-repeated as 2 copies in the haploid S288C) show hypersensitivity to sulfomethuron methyl (Fig. 5B). Interestingly, the CUP1 gene has been implicated as being highly-expressed in an azole-resistant yeast strain [92], indicating that it may play a role in drug response as well as in copper detoxification. Additionally, pdr13 deletion mutants are hypersensitive to copper [93]; both studies support our present findings that there is a link between the CUP1 locus and drug sensitivity.
There is also a good correlation between the overall observed variability of the copy number genotypes of the wine isolates among each wine strain and the variability (or lack thereof) of the drug responses; the Montrachet and French Red strains showed the most variability in gene copy numbers among their individual isolates, and likewise these two strains show the most variability in the drug response profiles of the isolates (Fig. 5B).
As mentioned above, we observed less inter-strain and intra-strain variation, especially aneuploidies, than would be expected based on previously published studies [23]. However, the strains we studied were all commercial and/or academic strains that are maintained under fairly constant growth conditions, while many of the previous studies showing genome instability and genomic diversity in wine strains were performed with "wild" yeast strains, i.e., those strains that occur spontaneously in uninoculated wine fermentations. It should also be noted that in this study we have analyzed a set of strains that all originated from France. It is possible that as we expand our studies either to wine yeast strains whose origins are outside of this geographic region, or to "wild" yeast strains, there will be greater genomic differences. Nevertheless, specific inter-strain genomic differences among the four wine strains were found that uniquely identify each strain. The three Montrachet isolates are depleted for ENB1 and the HXT9, HXT11, and HXT12 genes, but not depleted for AYT1 or YPL257W. The Prise de Mousse isolates show exactly the opposite characteristics, being depleted for AYT1 and YPL257W, but not ENB1 or HXT9, HXT11, and HXT12. Two of the three Pasteur Red isolates are depleted for YPL257W, and all are partially depleted for HXT9, HXT11, and HXT12, but not depleted for AYT1 or ENB1. Finally, the Champagne isolates are depleted for all of these genes (Fig. 5B). These results suggest the possibility of establishing a database that uniquely identifies every major wine yeast strain based on its array karyotype. This could allow rapid identification of wine strains in, for example, un-inoculated wine fermentations, or in suspected contaminated fermentations, etc., by the comparison of karyotypes from the unknown yeasts to those of the known yeasts.
Some amount of intra-strain variation was found among the French Red isolates and especially in the Montrachet isolates, in which a stretch of 37 genes on the left end of Chromosome VII was depleted in the Lalvin isolate relative to the other two isolates and to the S288C strain; the magnitude of this particular change was somewhat unexpected, and may indicate that wine yeast producers should be more vigilant in periodically checking the genomic integrity of their strains. Microarray karyotyping appears to be a method that can easily be used to check the genomic integrity of such strains.
Overall, however, the amount of both inter- and intra-strain genomic variation was low among the wine yeast strains, indicating firstly that these strains are closely related, and secondly, that they are quite genetically stable. This implies that these strains can generally be propagated under various conditions without fear of large-scale genomic rearrangements occurring. Our characterization of a "commercial wine strain signature", as well as the determination of unique signatures for various of wine yeast strains, may make it possible to use array karyotyping to help winemakers identify unknown yeasts in spontaneous fermentations, to detect contaminating yeasts during fermentation, and to identify the origins of novel yeast strains. Our results also allow the inference that the differences in the fermentation and organoleptic properties ascribed to these different strains may arise from a small number of genetic changes; it must be noted, though, that the microarray karyotyping technique only allows the detection of genetic changes (primarily copy number changes) at the whole-gene level, not single-nucleotide changes, of which there may be many. However, we can easily test, by adding back deleted genes into the genome, e.g., whether the observed copy number differences between the strains do indeed confer the different sensory properties in the finished wine that have been noted by winemakers.
Methods
Strains used
Table 1 shows a list of the commercial wine yeast strains used in this study, including the manufacturer ("Source") of each strain. UCD-numbers in parentheses under "Product Name" refer to the Univ. of Calif. Davis Department of Viticulture and Enology strain collection number. All UCD strains were purchased directly from U. C. Davis (Davis, CA, USA), all Lalvin strains were acquired from Vinquiry (Windsor, CA, USA), and all Red Star strains were purchased from "The Wine Lab" (Napa, CA, USA, [94]).
Six additional S. cerevisiae strains were used as controls for halo assays:
JG436: MATa, pdr5::Tn5, leu2, ura3, met5 (obtained from N. Alexander; [75]);
ScAYT1Δ: MATa, ayt1Δ, ura3, met5 (obtained from N. Alexander; [75]);
RW2802: MATa, leu2, ura3, met5 (obtained from N. Alexander; [75]);
GSY28: MATa, YPL257WΔ, his3Δ1, leu2Δ0, met15Δ0, ura3Δ0 (YeDelStrain #1035 [95]);
FY1679: MATa/MATα, leu2Δ1/+, +/his3Δ200, trp1Δ63/+, ura3-52/ura3-52 (S288C-based diploid strain obtained from laboratory of F. Winston; it is the result of a cross between FY23 and FY73; [96]);
S288C: MATα (obtained from laboratory of G. Fink).
The S288C strain has been previously described [41] and carries no auxotrophic markers; it represents one of the most widely-used laboratory "wild-type" yeast strains and is also the strain background for the sequenced S. cerevisiae strain. The S. bayanus strain used in Figure 1 was obtained from Cletus Kurtzman via Tom Pugh; it is the type strain of S. bayanus (NRRL Y-12624).
Microarray karyotyping protocol
In all karyotyping experiments described, we used microarrays onto which had been spotted PCR products corresponding to full-length ORFs from the S288C strain of S. cerevisiae, analogous to previously described arrays [97]. The DNA used in the hybridizations was prepared as described [98]. The reference DNA in all experiments was prepared from S288C (the S288C haploid strain; see above). We used S288C as the reference DNA since this is the strain that was sequenced and from which the PCR products used on the microarrays were generated; the Cy3-labeled DNA should thus give a green signal in every channel. After isolation the DNAs were then directly labeled with fluorescently-tagged nucleotides (Amersham, Piscataway, NJ, USA), either Cy3-dUTP for the reference strain or Cy-5 dUTP for the wine strains, using Klenow polymerase (NEB, Beverly, MA, USA) and random hexamers in a random-priming method as described [61]. After labeling, the reactions were heat-inactivated, the experimental (Cy5-labeled) and reference (Cy3-labeled) DNAs were mixed, purified away from unincorporated label using Microcon-30 filters (Millipore, Billerica, MA, USA), and then hybridized to the microarrays at 65°C as previously described [99]. All experiments were performed using duplicate arrays, where we hybridized the same labeling reaction to each of the two arrays. Arrays were scanned with an Axon 4000A scanner and the data were extracted with GenePix (Molecular Devices Corp., Union City, CA, USA) software as described [97]. The data were deposited in the Stanford Microarray Database [62,63] where they can be retrieved for further analysis; the raw data for all microarray hybridizations reported in this paper can also be downloaded from the site: . For all data analysis described, the array data were filtered first to eliminate manually flagged bad spots, then filtered to include spots for which the ratio of green (Cy3) channel mean intensity to Median background intensity was > 1.5 and the ratio of red (Cy5) channel normalized mean intensity to Median background intensity was > 1.0. The filtering was fairly permissive in order to allow truly deleted genes in the wine yeasts (i.e., no red signal at all) to be detected. All data are presented as the average of the values from the duplicate arrays. Karyoscopes were generated by the Java TreeView program [67].
Note that prior to the data filtering described above, all array data were normalized by setting the average log fluorescence hybridization ratio of all array elements to a value of zero; differences in hybridization intensity due to ploidy differences are therefore eliminated. Thus, although most wine strains are of diploid or higher ploidy, the normalization process allows direct comparison to the haploid reference strain so that only changes in gene copy number relative to the haploid state are detected. For example, if a gene has a copy number of two in the WT haploid strain and a copy number of four in the WT diploid strain, the red:green (R/G) hybridization ratio will be 1.0 (and the log ratio will be zero) since the haploid-relative copy number is the same between both strains. If, however, a diploid strain (labeled in red) has experienced a depletion in the copy number so that it now has only two copies of the gene, the R/G hybridization ratio will be 0.5 because the copy number relative to the haploid state has been lowered.
CGH-Miner
The CGH-Miner program [68] was installed and employed as described in the CGH-Miner User guide and Manual [100]. The karyscope shown in Fig. 2 was generated by the program using the parameters for BAC analysis because it gives the smallest size (three) of the moving-window for data smoothing. Eight separate S288C self-self hybridizations were used as "NN" controls in the CGH-Miner program, giving a robust baseline to determine the shared copy number changes in the wine strains. When the CGH-Miner program [68] generated the consensus plot shown in Fig. 2, it simultaneously created a file showing the genes that are significantly altered in copy number relative to "normal" (self-self) arrays; an abbreviated list of these genes with significantly-altered copy number are given in Table 2.
Halo assays
For each of the 18 yeast strains described above, 2 ml of liquid YPD medium [98] supplemented with adenine (50 mg/l final) and tryptophan (50 mg/l final) was inoculated with cells from frozen glycerol stocks. The cultures were grown at 25°C overnight until saturated. Cultures were measured for cell density using a Coulter Counter (Beckman-Coulter, Fullerton, CA, USA), then resuspended at 1 × 106 cells/ml in 4 ml of top agar (= 0.7% agar, 0.82X YPD plus adenine and tryptophan as above) which had been held at 42°C. After quickly vortexing, the mixture was poured smoothly onto YPD+Ade+Trp plates. The top agar was allowed to cool and harden, then 6.5 mm-diameter filter discs pre-loaded with the desired amount of drug were placed onto the surface of the top agar. The amounts of each drug loaded onto the filter discs were: 500 ng cycloheximide (Sigma-Aldrich, St. Louis, MO, USA; 2 μl of 250 μg/ml solution in water), 5 μg clotrimazole (Sigma; 4 μl of 1.25 mg/ml solution in acetone) and 20 μg sulfomethuron-methyl (DuPont, Wilmington, DE, USA; 8 μl of 2.5 mg/ml solution in acetone). Control filter discs containing 8 μl of either water or acetone were also placed on the plates; halos were never observed around these control discs. The plates were then incubated at 25°C for 2 – 3 days after which photographs of the plates were taken. Final halo measurements used for the charts in Fig. 5 were calculated by first measuring the diameter of the total halo (including the central filter disc) and calculating its area, then subtracting the area of the filter disc to give the annular area (i.e., the area of the annulus defined by the outer edge of the disc to the outer boundary of the halo). The annular area for each experimental yeast strain was then divided by that of the strain FY1679 (S288C-based diploid) to yield an annular area ratio for each drug. Finally, the log2 of the annular area ratio was calculated and represent the y-axis values shown in Fig. 5B, where positive values indicate greater sensitivity to the drug compared to the WT-diploid control, and negative values indicate greater resistance to the drug compared to WT.
Abbreviations
CGH: Comparative Genome Hybridization
CTZ: clotrimazole
CYH: cycloheximide
S-AM: S-adenosylmethionine
SGD: Saccharomyces Genome Database
SMM: sulfomethuron methyl
WT: Wild-type
Authors' contributions
BD planned and designed the experiments, performed the experiments and the data analysis, wrote the main draft of the paper, and generated the figures. RPL aided in the experimental design and in editing the manuscript. GS provided guidance over the entire project, and also aided in figure making, data analysis and editing of the manuscript. All authors read and approved the final manuscript.
Additional material
The raw data files for all microarray hybridizations presented in this paper may be downloaded at:
Acknowledgements
We thank Nancy Alexander for providing us with the ayt1, pdr5, and RW2802 yeast strains and Tom Pugh for providing the S. bayanus strain used in Fig. 1. We also thank Pei Wang for providing interactive support for the CGH-Miner program and Katja Schwartz for critical reading of the manuscript. Finally, we thank our three anonymous reviewers, all of whom helped improve the manuscript with their comments and insight. This work was supported by NSF ADVANCE grant DBI-0340856 to BD.
==== Refs
Romano P Fiore C Paraggio M Caruso M Capece A Function of yeast species and strains in wine flavour Int J Food Microbiol 2003 86 169 180 12892932 10.1016/S0168-1605(03)00290-3
Torrea D Ancin C Content of biogenic amines in a Chardonnay wine obtained through spontaneous and inoculated fermentations J Agric Food Chem 2002 50 4895 4899 12166978 10.1021/jf011653h
Patel S Shibamoto T Effect of different strains of Saccharomyces cerevisiae on production of volatiles in Napa Gamay wine and Petite Sirah wine J Agric Food Chem 2002 50 5649 5653 12236692 10.1021/jf020337f
Lilly M Lambrechts MG Pretorius IS Effect of increased yeast alcohol acetyltransferase activity on flavor profiles of wine and distillates Appl Environ Microbiol 2000 66 744 753 10653746 10.1128/AEM.66.2.744-753.2000
Nadal D Colomer B Pina B Molecular polymorphism distribution in phenotypically distinct populations of wine yeast strains Appl Environ Microbiol 1996 62 1944 1950 8787392
Martinez-Rodriguez AJ Carrascosa AV Martin-Alvarez PJ Moreno-Arribas V Polo MC Influence of the yeast strain on the changes of the amino acids, peptides and proteins during sparkling wine production by the traditional method J Ind Microbiol Biotechnol 2002 29 314 322 12483471 10.1038/sj.jim.7000323
Martinez-Rodriguez AJ Carrascosa AV Polo MC Release of nitrogen compounds to the extracellular medium by three strains of Saccharomyces cerevisiae during induced autolysis in a model wine system Int J Food Microbiol 2001 68 155 160 11545216 10.1016/S0168-1605(01)00486-X
Benitez T Martinez P Codon AC Genetic constitution of industrial yeast Microbiologia 1996 12 371 384 8897417
Mortimer RK Evolution and variation of the yeast (Saccharomyces) genome Genome Res 2000 10 403 409 10779481 10.1101/gr.10.4.403
Fleet GH Yeast interactions and wine flavour Int J Food Microbiol 2003 86 11 22 12892919 10.1016/S0168-1605(03)00245-9
Ganga MA Pinaga F Valles S Ramon D Querol A Aroma improving in microvinification processes by the use of a recombinant wine yeast strain expressing the Aspergillus nidulans xlnA gene Int J Food Microbiol 1999 47 171 178 10359487 10.1016/S0168-1605(98)00202-5
Gardner N Rodrigue N Champagne CP Combined effects of sulfites, temperature, and agitation time on production of glycerol in grape juice by Saccharomyces cerevisiae Appl Environ Microbiol 1993 59 2022 2028 8357243
Michnick S Roustan JL Remize F Barre P Dequin S Modulation of glycerol and ethanol yields during alcoholic fermentation in Saccharomyces cerevisiae strains overexpressed or disrupted for GPD1 encoding glycerol 3-phosphate dehydrogenase Yeast 1997 13 783 793 9234667 10.1002/(SICI)1097-0061(199707)13:9<783::AID-YEA128>3.0.CO;2-W
Remize F Sablayrolles JM Dequin S Re-assessment of the influence of yeast strain and environmental factors on glycerol production in wine J Appl Microbiol 2000 88 371 378 10747217 10.1046/j.1365-2672.2000.00964.x
Romano P Suzzi G Mortimer R Polsinelli M Production of high levels of acetoin in Saccharomyces cerevisiae wine yeasts is a recessive trait J Appl Bacteriol 1995 78 169 174 7698951
Romano P Suzzi G Turbanti L Polsinelli M Acetaldehyde production in Saccharomyces cerevisiae wine yeasts FEMS Microbiol Lett 1994 118 213 218 8020744 10.1016/0378-1097(94)90506-1
Martinez P Codon AC Perez L Benitez T Physiological and molecular characterization of flor yeasts: polymorphism of flor yeast populations Yeast 1995 11 1399 1411 8585323
Ramirez M Vinagre A Ambrona J Molina F Maqueda M Rebollo JE Genetic instability of heterozygous, hybrid, natural wine yeasts Appl Environ Microbiol 2004 70 4686 4691 15294803 10.1128/AEM.70.8.4686-4691.2004
Longo E Vezinhet F Chromosomal rearrangements during vegetative growth of a wild strain of Saccharomyces cerevisiae Appl Environ Microbiol 1993 59 322 326 8439158
Guijo S Mauricio JC Salmon JM Ortega JM Determination of the relative ploidy in different Saccharomyces cerevisiae strains used for fermentation and 'flor' film ageing of dry sherry-type wines Yeast 1997 13 101 117 9046092 10.1002/(SICI)1097-0061(199702)13:2<101::AID-YEA66>3.0.CO;2-H
Miklos I Varga T Nagy A Sipiczki M Genome instability and chromosomal rearrangements in a heterothallic wine yeast J Basic Microbiol 1997 37 345 354 9373949
Nadal D Carro D Fernandez-Larrea J Pina B Analysis and dynamics of the chromosomal complements of wild sparkling-wine yeast strains Appl Environ Microbiol 1999 65 1688 1695 10103269
Bakalinsky AT Snow R The chromosomal constitution of wine strains of Saccharomyces cerevisiae Yeast 1990 6 367 382 2220073
Naumov GI Naumova ES Antunovics Z Sipiczki M Saccharomyces bayanus var. uvarum in Tokaj wine-making of Slovakia and Hungary Appl Microbiol Biotechnol 2002 59 727 730 12226732 10.1007/s00253-002-1077-6
Rachidi N Barre P Blondin B Multiple Ty-mediated chromosomal translocations lead to karyotype changes in a wine strain of Saccharomyces cerevisiae Mol Gen Genet 1999 261 841 850 10394922 10.1007/s004380050028
Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chromosomal DNA polymorphism of wine strains of Saccharomyces cerevisiae Curr Genet 1992 22 1 7 1611665 10.1007/BF00351734
Naumov GI Naumova ES Masneuf I Aigle M Kondratieva VI Dubourdieu D Natural polyploidization of some cultured yeast Saccharomyces sensu stricto: auto- and allotetraploidy Syst Appl Microbiol 2000 23 442 449 11108025
Tamai Y Momma T Yoshimoto H Kaneko Y Co-existence of two types of chromosome in the bottom fermenting yeast, Saccharomyces pastorianus Yeast 1998 14 923 933 9717238 10.1002/(SICI)1097-0061(199807)14:10<923::AID-YEA298>3.3.CO;2-9
Adams J Oeller PW Structure of evolving populations of Saccharomyces cerevisiae: adaptive changes are frequently associated with sequence alterations involving mobile elements belonging to the Ty family Proc Natl Acad Sci U S A 1986 83 7124 7127 3018758
Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation and major chromosomal changes in populations of Saccharomyces cerevisiae Curr Genet 1992 22 13 19 1611666 10.1007/BF00351736
Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig F Botstein D Characteristic genome rearrangements in experimental evolution of Saccharomyces cerevisiae Proc Natl Acad Sci U S A 2002 99 16144 16149 12446845 10.1073/pnas.242624799
McClintock B The significance of responses of the genome to challenge Science 1984 226 792 15739260
Watson JD Origin of concatemeric T7 DNA Nat New Biol 1972 239 197 201 4507727
Hughes TR Roberts CJ Dai H Jones AR Meyer MR Slade D Burchard J Dow S Ward TR Kidd MJ Friend SH Marton MJ Widespread aneuploidy revealed by DNA microarray expression profiling Nat Genet 2000 25 333 337 10888885 10.1038/77116
Perez-Ortin JE Querol A Puig S Barrio E Molecular characterization of a chromosomal rearrangement involved in the adaptive evolution of yeast strains Genome Res 2002 12 1533 1539 12368245 10.1101/gr.436602
Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombination and genetic changes in Saccharomyces cerevisiae during wine fermentation Appl Environ Microbiol 2000 66 2057 2061 10788381 10.1128/AEM.66.5.2057-2061.2000
Aguilera A Chavez S Malagon F Mitotic recombination in yeast: elements controlling its incidence Yeast 2000 16 731 754 10861900 10.1002/1097-0061(20000615)16:8<731::AID-YEA586>3.0.CO;2-L
Frankel R Wine and Oil Production in Antiquity in Israel and Other Mediterranean Countries. Jsot/ASOR Series 1998 10 Sheffield, UK , Sheffield Academic Press
McGovern PE Ancient Wine: The Search for the Origins of Viniculture. 2003 Princeton , Princeton University Press
Cavalieri D McGovern PE Hartl DL Mortimer R Polsinelli M Evidence for S. cerevisiae fermentation in ancient wine J Mol Evol 2003 57 Suppl 1 S226 32 15008419 10.1007/s00239-003-0031-2
Mortimer RK Johnston JR Genealogy of principal strains of the yeast genetic stock center Genetics 1986 113 35 43 3519363
Eisen MB Brown PO DNA arrays for analysis of gene expression Methods Enzymol 1999 303 179 205 10349646
Holloway AJ van Laar RK Tothill RW Bowtell DD Options available--from start to finish--for obtaining data from DNA microarrays II Nat Genet 2002 32 Suppl 481 489 12454642 10.1038/ng1030
Pinkel D Landegent J Collins C Fuscoe J Segraves R Lucas J Gray J Fluorescence in situ hybridization with human chromosome-specific libraries: detection of trisomy 21 and translocations of chromosome 4 Proc Natl Acad Sci U S A 1988 85 9138 9142 2973607
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 10.1038/12640
Albertson DG Profiling breast cancer by array CGH Breast Cancer Res Treat 2003 78 289 298 12755488 10.1023/A:1023025506386
Kashiwagi H Uchida K Genome-wide profiling of gene amplification and deletion in cancer Hum Cell 2000 13 135 141 11197775
Pollack JR Sorlie T Perou CM Rees CA Jeffrey SS Lonning PE Tibshirani R Botstein D Borresen-Dale AL Brown PO Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors Proc Natl Acad Sci U S A 2002 99 12963 12968 12297621 10.1073/pnas.162471999
Lashkari DA McCusker JH Davis RW Whole genome analysis: experimental access to all genome sequenced segments through larger-scale efficient oligonucleotide synthesis and PCR Proc Natl Acad Sci U S A 1997 94 8945 8947 9256414 10.1073/pnas.94.17.8945
Perez-Ortin JE Garcia-Martinez J Alberola TM DNA chips for yeast biotechnology. The case of wine yeasts J Biotechnol 2002 98 227 241 12141989 10.1016/S0168-1656(02)00134-7
Bond U Neal C Donnelly D James TC Aneuploidy and copy number breakpoints in the genome of lager yeasts mapped by microarray hybridisation Curr Genet 2004 45 360 370 15103502 10.1007/s00294-004-0504-x
Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DR Zhou Y Hartl DL Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 163 79 89 12586698
Backhus LE DeRisi J Brown PO Bisson LF Functional genomic analysis of a commercial wine strain of Saccharomyces cerevisiae under differing nitrogen conditions FEMS Yeast Res 2001 1 111 125 12702356 10.1016/S1567-1356(01)00019-8
Erasmus DJ van der Merwe GK van Vuuren HJ Genome-wide expression analyses: Metabolic adaptation of Saccharomyces cerevisiae to high sugar stress FEMS Yeast Res 2003 3 375 399 12748050 10.1016/S1567-1356(02)00203-9
Rossignol T Dulau L Julien A Blondin B Genome-wide monitoring of wine yeast gene expression during alcoholic fermentation Yeast 2003 20 1369 1385 14663829 10.1002/yea.1046
Infante JJ Dombek KM Rebordinos L Cantoral JM Young ET Genome-wide amplifications caused by chromosomal rearrangements play a major role in the adaptive evolution of natural yeast Genetics 2003 165 1745 1759 14704163
Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C Hwang SY Brown PO Davis RW Yeast microarrays for genome wide parallel genetic and gene expression analysis Proc Natl Acad Sci U S A 1997 94 13057 13062 9371799 10.1073/pnas.94.24.13057
Primig M Williams RM Winzeler EA Tevzadze GG Conway AR Hwang SY Davis RW Esposito RE The core meiotic transcriptome in budding yeasts Nat Genet 2000 26 415 423 11101837 10.1038/82539
Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI Oliver SG Comparative genomic hybridization provides new insights into the molecular taxonomy of the Saccharomyces sensu stricto complex Genome Res 2004 14 1043 1051 15173111 10.1101/gr.2114704
Hauser NC Fellenberg R Gil R Bastuk S Hoheisel JD Perez-Ortin JE Whole genome analysis of a wine yeast strain Comparative and Functional Genomics 2001 2 69 79 10.1002/cfg.73
Genomic DNA Labeling Protocol
[http://cmgm.stanford.edu/pbrown/protocols/4_genomic.html]
Stanford Microarray Database [http://smd.stanford.edu/]
Sherlock G Hernandez-Boussard T Kasarskis A Binkley G Matese JC Dwight SS Kaloper M Weng S Jin H Ball CA Eisen MB Spellman PT Brown PO Botstein D Cherry JM The Stanford Microarray Database Nucleic Acids Res 2001 29 152 155 11125075 10.1093/nar/29.1.152
Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybrids between Saccharomyces sensu stricto yeast species found among wine and cider production strains Appl Environ Microbiol 1998 64 3887 3892 9758815
Egli CM Edinger WD Mitrakul CM Henick-Kling T Dynamics of indigenous and inoculated yeast populations and their effect on the sensory character of Riesling and Chardonnay wines J Appl Microbiol 1998 85 779 789 9830113 10.1046/j.1365-2672.1998.00521.x
Kunkee RE Vilas MR Toward an understanding of the relationship between yeast strain and flavor production during vinifications: Flavor effects in vinifications of a nondistinct variety of grapes by several strains of wine yeast Vitic Enol Sci 1994 49 46 50
Saldanha AJ Java treeview--extensible visualization of microarray data Bioinformatics 2004 20 3246 3248 15180930 10.1093/bioinformatics/bth349
Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method for calling gains and losses in array CGH data Biostatistics 2005 6 45 58 15618527 10.1093/biostatistics/kxh017
Saccharomyces Genome Database [http://www.yeastgenome.org/]
Cherry JM Adler C Ball C Chervitz SA Dwight SS Hester ET Jia Y Juvik G Roe T Schroeder M Weng S Botstein D SGD: Saccharomyces Genome Database Nucleic Acids Res 1998 26 73 79 9399804 10.1093/nar/26.1.73
Delaveau T Delahodde A Carvajal E Subik J Jacq C PDR3, a new yeast regulatory gene, is homologous to PDR1 and controls the multidrug resistance phenomenon Mol Gen Genet 1994 244 501 511 8078477 10.1007/BF00583901
Launhardt H Hinnen A Munder T Drug-induced phenotypes provide a tool for the functional analysis of yeast genes Yeast 1998 14 935 942 9717239 10.1002/(SICI)1097-0061(199807)14:10<935::AID-YEA289>3.3.CO;2-0
Kellis M Patterson N Endrizzi M Birren B Lander ES Sequencing and comparison of yeast species to identify genes and regulatory elements Nature 2003 423 241 254 12748633 10.1038/nature01644
Hurst LD The Ka/Ks ratio: diagnosing the form of sequence evolution Trends Genet 2002 18 486 12175810 10.1016/S0168-9525(02)02722-1
Alexander NJ McCormick SP Hohn TM The identification of the Saccharomyces cerevisiae gene AYT1(ORF-YLL063c) encoding an acetyltransferase Yeast 2002 19 1425 1430 12478589 10.1002/yea.924
Nourani A Wesolowski-Louvel M Delaveau T Jacq C Delahodde A Multiple-drug-resistance phenomenon in the yeast Saccharomyces cerevisiae: involvement of two hexose transporters Mol Cell Biol 1997 17 5453 5460 9271421
Akache B Turcotte B New regulators of drug sensitivity in the family of yeast zinc cluster proteins J Biol Chem 2002 277 21254 21260 11943786 10.1074/jbc.M202566200
do Valle Matta MA Jonniaux JL Balzi E Goffeau A van den Hazel B Novel target genes of the yeast regulator Pdr1p: a contribution of the TPO1 gene in resistance to quinidine and other drugs Gene 2001 272 111 119 11470516 10.1016/S0378-1119(01)00558-3
Mizoguchi H Yamauchi T Watanabe M Yamanaka H Nishimura A Hanmoto H Different missense mutations in PDR1 and PDR3 genes from clotrimazole-resistant sake yeast are responsible for pleiotropic drug resistance and improved fermentative activity J Biosci Bioeng 2002 93 221 227 16233191 10.1263/jbb.93.221
Watanabe M Mizoguchi H Nishimura A Disruption of the ABC transporter genes PDR5, YOR1, and SNQ2, and their participation in improved fermentative activity of a sake yeast mutant showing pleiotropic drug resistance J Biosci Bioeng 2000 89 569 576 16232800 10.1016/S1389-1723(00)80059-6
Heymann P Ernst JF Winkelmann G A gene of the major facilitator superfamily encodes a transporter for enterobactin (Enb1p) in Saccharomyces cerevisiae Biometals 2000 13 65 72 10831226 10.1023/A:1009250017785
Codon AC Benitez T Korhola M Chromosomal polymorphism and adaptation to specific industrial environments of Saccharomyces strains Appl Microbiol Biotechnol 1998 49 154 163 9534255 10.1007/s002530051152
Umezu K Hiraoka M Mori M Maki H Structural analysis of aberrant chromosomes that occur spontaneously in diploid Saccharomyces cerevisiae: retrotransposon Ty1 plays a crucial role in chromosomal rearrangements Genetics 2002 160 97 110 11805048
Yu X Gabriel A Reciprocal translocations in Saccharomyces cerevisiae formed by nonhomologous end joining Genetics 2004 166 741 751 15020464 10.1534/genetics.166.2.741
Sakurai M Tohda H Kumagai H Giga-Hama Y A distinct type of alcohol dehydrogenase, adh4+, complements ethanol fermentation in an adh1-deficient strain of Schizosaccharomyces pombe FEMS Yeast Res 2004 4 649 654 15040954 10.1016/j.femsyr.2003.12.009
Higgins VJ Rogers PJ Dawes IW Application of genome-wide expression analysis to identify molecular markers useful in monitoring industrial fermentations Appl Environ Microbiol 2003 69 7535 7540 14660410 10.1128/AEM.69.12.7535-7540.2003
Lyons TJ Gasch AP Gaither LA Botstein D Brown PO Eide DJ Genome-wide characterization of the Zap1p zinc-responsive regulon in yeast Proc Natl Acad Sci U S A 2000 97 7957 7962 10884426 10.1073/pnas.97.14.7957
Giaever G Chu AM Ni L Connelly C Riles L Veronneau S Dow S Lucau-Danila A Anderson K Andre B Arkin AP Astromoff A El-Bakkoury M Bangham R Benito R Brachat S Campanaro S Curtiss M Davis K Deutschbauer A Entian KD Flaherty P Foury F Garfinkel DJ Gerstein M Gotte D Guldener U Hegemann JH Hempel S Herman Z Jaramillo DF Kelly DE Kelly SL Kotter P LaBonte D Lamb DC Lan N Liang H Liao H Liu L Luo C Lussier M Mao R Menard P Ooi SL Revuelta JL Roberts CJ Rose M Ross-Macdonald P Scherens B Schimmack G Shafer B Shoemaker DD Sookhai-Mahadeo S Storms RK Strathern JN Valle G Voet M Volckaert G Wang CY Ward TR Wilhelmy J Winzeler EA Yang Y Yen G Youngman E Yu K Bussey H Boeke JD Snyder M Philippsen P Davis RW Johnston M Functional profiling of the Saccharomyces cerevisiae genome Nature 2002 418 387 391 12140549 10.1038/nature00935
Welcome to Lesaffre Yeast Corporation | Wine Yeast Division | Products http://www.lesaffreyeastcorp.com/wineyeast/products.html
Fernandez-Espinar MT Lopez V Ramon D Bartra E Querol A Study of the authenticity of commercial wine yeast strains by molecular techniques Int J Food Microbiol 2001 70 1 10 11759747 10.1016/S0168-1605(01)00502-5
Dunn B Szauter P Pardue ML Szostak JW Transfer of yeast telomeres to linear plasmids by recombination Cell 1984 39 191 201 6091911 10.1016/0092-8674(84)90205-8
Barker KS Pearson MM Rogers PD Identification of genes differentially expressed in association with reduced azole susceptibility in Saccharomyces cerevisiae J Antimicrob Chemother 2003 51 1131 1140 12697649 10.1093/jac/dkg217
Kim DY Song WY Yang YY Lee Y The role of PDR13 in tolerance to high copper stress in budding yeast FEBS Lett 2001 508 99 102 11707276 10.1016/S0014-5793(01)03038-1
The Wine Lab http://www.thewinelab.com/
Saccharomyces Genome Deletion Project [http://www sequence.stanford.edu/group/yeast_deletion_project/deletions3.html]
Winston F Dollard C Ricupero-Hovasse SL Construction of a set of convenient Saccharomyces cerevisiae strains that are isogenic to S288C Yeast 1995 11 53 55 7762301
DeRisi JL Iyer VR Brown PO Exploring the metabolic and genetic control of gene expression on a genomic scale Science 1997 278 680 686 9381177 10.1126/science.278.5338.680
Treco D Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl R Preparation of Yeast DNA Current Protocols in Molecular Biology Current Protocols in Molecular Biology 2003 John Wiley & Sons 13.11 14686320
Hybridization of Arrays [http://cmgm.stanford.edu/pbrown/protocols/5_hyb_yeast.html]
CGH Miner Users Guide and Manual [http://www stat.stanford.edu/~wp57/CGH-Miner/CGH-Miner-Manual.pdf]
| 15833139 | PMC1097725 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 Apr 16; 6:53 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-53 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-561584272910.1186/1471-2164-6-56Research ArticleType 1 Diabetes in the Spanish Population: additional factors to Class II HLA-DR3 and -DR4 Urcelay Elena [email protected] José L [email protected] la Calle Hermenegildo [email protected]ínez Alfonso [email protected]éndez Julián [email protected] José M [email protected] Carlos [email protected]ández-Arquero Miguel [email protected] la Concha Emilio G [email protected] Immunology Department, Hospital Universitario San Carlos, Madrid, Spain2 Endocrinology Department, Hospital Ramón y Cajal, Madrid, Spain3 Diabetes Unit, Hospital Universitario San Carlos, Madrid, Spain4 Department of Paediatrics, Hospital Universitario San Carlos, Madrid, Spain2005 20 4 2005 6 56 56 15 11 2004 20 4 2005 Copyright © 2005 Urcelay et al; licensee BioMed Central Ltd.2005Urcelay 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 Major Histocompatibility Complex is the main genetic contributor to susceptibility to type 1 diabetes (T1D); genome-wide scans have consistently mapped increased predisposition to this region. The highest disease risk has been associated with HLA-DR3 and HLA-DR4. In particular, the DR3-positive ancestral haplotype 18.2 was reported as highly diabetogenic. We aimed to corroborate whether this haplotype increases the susceptibility conferred by the DQ2-DR3 alleles in a Mediterranean population. We also searched for additional susceptibility factors to the classic DQ2-DR3 and DQ8-DR4.
Results
Genetic MHC markers were analysed in a case-control study with 302 T1D patients and 529 ethnically matched controls. DR3-TNFa1b5 carrier rate was significantly higher in DR3-positive heterozygous T1D patients than in DR3-positive heterozygous controls (p = 0.0019; odds ratio OR [95% confidence interval CI] = 2.26 [1.3–3.93]). This data was confirmed analysing the allelic frequency, which includes the information corresponding to the DR3-homozygous individuals (p = 0.001; OR = 2.09) and by using the Arlequin software to check the DR3-positive haplotypes (p = 0.004;OR = 1.93). The present results provide strong evidence of a second susceptibility region in the ancestral haplotype 18.2 in the Spanish population.
Moreover, we searched for T1D susceptibility factors in addition to the MHC classical ones, within the DR2-DQ6/DR3-DQ2/DR4-DQ8 negative population. Several genetic markers in both MHC class II (DQA1*0101-DQB1*0501 [p = 0.007;OR = 2.81], DQA1*0201-DQB1*0202 [p = 0.03; OR = 2.35]) and III (TNFa2b1 [p = 0.01 OR = 2.74], BAT-2*2 [p = 0.004; OR = 3.19]) were found. These different alleles associated with T1D were not independent and we observed linkage disequilibrium among them leading us to describe two new risk haplotypes (DQA1*0101-DQB1*0501-TNFa2b1 and DQA1*0201-DQB1*0202- BAT-2*2). Finally, we studied a T1D susceptibility/protection marker located in extended class I, D6S2223; however, no association was observed in our population.
Conclusion
Our results suggest that other associated MHC haplotypes might present susceptibility factors in loci different from HLA-class II and that the class II molecules are not necessarily the universal etiologic factor in every MHC haplotype.
==== Body
Background
Type 1 diabetes is a multifactorial autoimmune disease characterised by insulin deficiency, due to the T cell mediated destruction of pancreatic β-cells [1]. Among the genetic determinants of susceptibility, with more than 18 putative loci identified to date, a region in chromosome 6p21 (IDDM1) containing the Major Histocompatibility Complex (MHC) is the only one consistently associated with T1D in genome-wide screenings. The MHC spans approximately 4 Mb and consists of over 200 genes arranged in three subregions named class II, III and I. More than 90% of Caucasoid type 1 diabetic patients have at least one copy of the class II HLA-DR3 or DR4 allele, as compared to the 45% present in the general population and the genotype frequency of the DR3/DR4 heterozygote in T1D patients is 40% vs. 3% in controls [2]. In fact, the strongest susceptibility haplotypes described are DQB1*0201-DQA1*0501-DRB1*03 (DQ2-DR3) and DQB1*0302-DQA1*0301-DRB1*04 (DQ8-DR4), especially when both appear together in the genotype. However, not all HLA-DR3 or -DR4 positive haplotypes are equally predisposing [3,4]. This might suggest the role of other MHC loci responsible for modifying the susceptibility to diabetes conferred by class II genes, but their search has been difficult due to the strong linkage disequilibrium (LD) present in this chromosomal region. The term ancestral, extended haplotype was coined referring to continuous sequence derived with little, if any, change from an ancestor of all those now carrying all or part of the haplotype [5]. A published report stated that the DQ2-DR3-B18 AH 18.2 significantly increased risk compared to other haplotypes with the same class II alleles, but LD between markers hampered a more precise localisation of the presumed susceptibility gene [6]. Taking into account that this AH 18.2 is more frequent in Southern European populations, we decided to assess the possible role of other loci in the MHC region besides DQ-DR within this haplotype. We evaluated several genetic markers along the MHC in a case-control study with 302 Spanish T1D patients and 529 healthy controls. This study reproduced the especially strong association of the AH 18.2 with T1D in the Spanish population, supporting the existence of a second susceptibility locus within this haplotype. The present work extended the search to other T1D risk factors in the MHC besides DQ2-DR3/DQ8-DR4. Additional MHC haplotypes question the paradigm of class II genes as sole responsible for the association and open up the possibility that class III susceptibility factors would be also involved.
Results
We first aimed to corroborate in the Spanish population the published increased risk to autoimmune diabetes of the AH 18.2, as compared to other DR3-positive haplotypes [6]. Microsatellites TNFa and TNFb, located in the untranslated region upstream of the LTA gene, exhibit extensive polymorphism and are good haplospecific markers. In order to identify the extended haplotype AH 18.2, we selected both microsatellites TNFab, which are conveniently close to HLA-B and easier to genotype than the latter. The microsatellites TNFa1b5 tag the AH18.2 defined by the characteristic alleles of five markers (D6S273*2, BAT2*2, TNFa1b5, MICA*4) in our families with IgA deficiency [7] and celiac disease [8]. These AH 18.2 markers were also confirmed by the computer program Arlequin, by linkage disequilibrium studies and by analysis of homozygous individuals. As we are dealing with patients and controls with no families at our disposal, the DR3-TNFa1b5 phase was uncertain. However, due to heavy linkage disequilibrium between these alleles, most of the times they will probably be in cis. This proportion can be estimated based upon the presence of TNFa1b5 in DR3-negative individuals: 8 carriers out of 109 DR3-negative T1D patients, and 13 carriers out of 376 DR3-negative healthy controls. These numbers yield allelic frequencies of TNFa1b5 on DR3-negative haplotypes of 7 and 3%, respectively. The assumption can therefore be made that, most of the time; a DR3 heterozygous sample carrying TNFa1b5 will in fact carry these alleles in the same haplotype.
Once the AH 18.2 was defined, we looked for the already described different distribution of this haplotype between T1D patient and control cohorts. Initially, we compared the carrier rate of TNFa1b5 among DR3-positive individuals. In T1D patients, the ratio of DR3 homozygous subjects out of the total DR3-positive individuals is much higher than the observed in controls (47 out of 193 vs. 8 out of 153 in controls; p = 1 × 10-6; OR = 5.83). This makes necessary to calculate the TNFa1b5 carrier rate exclusively in DR3 heterozygous individuals to avoid the cumulative effect of those individuals. When the TNFa1b5 carrier rate was compared, it was significantly higher in DR3-positive heterozygous T1D patients than in DR3-positive heterozygous controls (66 out of 143 vs. 33 out of 120; p = 0.0019; OR (95% CI) = 2.26 [1.3–3.93]). If we do not want to discard the information corresponding to the DR3-homozygous individuals, the comparison should be made counting allelic rather than phenotypic frequency and then normalising these counts by the total number of DR3-positive haplotypes present in the DR3-positive individuals (112/123 vs. 40/92, p = 0.001, OR = 2.09) (Table 1).
Table 1 Comparison of the TNFa1b5 allelic frequency in the DR3-positive T1D patient and control cohorts.
DR3-positive Individuals Total haplotypes DR3-positive haplotypes TNFa1b5 alleles TNFa2b3 alleles
T1D 189 378 235 112 64
Controls 126 252 132 40 51
The results strongly suggest that the haplotype AH 18.2 is over represented among DR3 haplotypes. To formally elucidate whether this is the case, we observed that the allelic frequency of DR3-TNFa1b5 (AH 18.2) is higher than that of the DR3-positive abundant haplotype AH 8.1, which carries TNFa2b3, (112/40 vs. 64/51, p = 0.002, OR (95% CI) = 2.23 [1.29–3.86]). Moreover, when the allele frequencies of TNFa1b5 and TNFa2b3 were compared to the total of TNFab alleles excluding the two aforementioned, the AH18.2 showed significant difference (112/59 vs. 40/41, p = 0.01, OR = 1.95) whereas TNFa2b3, AH8.1, did not (64/59 vs. 51/41, p = 0.62, OR = 0.87). An alternative approach would consist of applying an expectation-maximisation algorithm, as the one implemented by the Arlequin program to the sample genotypes. Table 2 shows the DR3 three-loci haplotypes generated by this software sorted into TNFa1b5 positives, TNFa2b3 positives and other haplotypes. Again, TNFa1b5 is significantly different from other DR3 haplotypes, while TNFa2b3 behaves similarly to the rest.
Table 2 Frequency and number of DR3-positive haplotypes estimated by the Arlequin software.
T1D (2n = 596) Controls (2n = 1006)
Freq. Haplotypes Freq. Haplotypes
DR3-Total 0.39432 235 0.13120 132
I. DR3-a1b5 0.17276 103 0.03802 38
II. DR3-a2b3 0.10244 61 0.04866 49
III. DR3-other 0.11912 71 0.04452 45
I vs. II+III. OR = 1.93 (1.19–3.13); p = 0.004.
I vs. II. OR = 2.18 (1.24–3.83); p = 0.004.
I vs. III. OR = 1.72 (0.98–3.01); p = 0.043.
II vs. III. OR = 0.79 (0.45–1.39); p = 0.38.
In our Spanish sample this DR3-TNFa1b5 positive haplotype was present in a 33% T1D patients (98/298) vs. 7% controls (37/504). The haplotypes ascertained by the Arlequin software yielded an AH 18.2 almost five times more frequent in the Spanish patient than control cohorts (Table 2). Moreover, almost one half (46%, 66/142) of the DR3 heterozygous diabetic patients vs. one forth (27%, 33/120) of the controls displayed a DR3-TNFa1b5 positive phenotype. On the contrary, in this subpopulation the distribution of the other main DR3-positive haplotype, AH 8.1, was similar between patients (43/142, 30%) and controls (47/120, 39%).
Therefore, these results confirmed the reported specific diabetogenic role of the AH 18.2 in a Mediterranean population and are compatible with published observations that established the presence of a second susceptibility gene in the AH 18.2 [6,9] possibly located in either class III or class I.
The second aim of this work was searching for additional T1D susceptibility factors to the classical DQ2-DR3 and DQ8-DR4 in the MHC region. Obviously, the subjects studied should belong to DR3- and DR4-negative populations to eliminate the underlying genetic risk attributable to these known risk factors and they were also stratified by DR2, to avoid the effect of this allele negatively associated with the disease. We tested the predisposition determined by alleles of the five genetic markers in HLA class III already studied (Table 3), and we also compared the distribution of the DQA1-DQB1 alleles (Table 4). Then, alleles TNFa10b4, DQA1*0501-DQB1*0301 and DQA1*0103-DQB1*0603 for protection, and TNFa2b1, BAT-2*2, DQA1*0101-DQB1*0501 and DQA1*0201-DQB1*0202 for susceptibility, displayed statistical significant difference when T1D patients and healthy controls were compared in the DR3, DR4 and DR2- negative population. These different alleles associated with T1D were not mutually independent and linkage disequilibrium was observed between TNFa2b1 and DQA1*0101-DQB1*0501 (p = 9 × 10-5; D' = 0.34), BAT-2*2 and DQA1*0201-DQB1*0202 (p = 4 × 10-7; D' = 0.36) and weak LD between TNFa10b4 and DQA1*0103-DQB1*0603 (p = 0.09; D' = 0.16). These results would suggest the presence of extended haplotypes increasing the risk to autoimmune diabetes in these conditions, although the exact location of the susceptibility gene remains undetermined. The allele 2 of microsatellite BAT-2 within the haplotype DQA1*0201-DQB1*0202 is the same present in the AH 18.2, and therefore a common susceptibility region could potentially be shared by both haplotypes. To test this possibility two adjacent markers to BAT-2 were analysed, one telomeric, MN6S1879 and other centromeric, N3-2-3. The MN6S1879*14 allele present in the AH 18.2 is the same found in LD with BAT-2*2 in the DR2-DQ6/DR3-DQ2/DR4-DQ8 negative patients (p = 7 × 10-5; D' = 0.33). Curiously, provided the extensive LD displayed by the AH 18.2, there are two different alleles (7 and 9) in the N3-2-3 marker corroborated in our deficit IgA and celiac disease families, in homozygous individuals and by Arlequin software. The allele N3-2-3*7 did not show LD with BAT-2*2 in the triple negative population (p = 0.87, D' = 0.014), but N3-2-3*9 (p = 0.052, D' = 0.26) showed a trend for association. Although this latter allele seemed to mark higher risk than N3-2-3*7, the difference did not reach statistical significance. There was indirect evidence supporting the role of the allele N3-2-3*9 in T1D susceptibility. In two DR3-TNFa1b5 homozygous control individuals there were three alleles N3-2-3*7 and one N3-2-3*9, while in nine T1D patients, there were nine alleles each. Moreover, in DR3-DQ2/DR4-DQ8 individuals carrying TNFa1b5, the allele present in N3-2-3 is 9 (41 out of 83 vs. 0 from 14, p = 0.0012) instead of 7 (25 out of 83 vs. 7 out of 12, p = 0.06, OR = 0.31). Of notice the different trend shown for the allele 9 conferring susceptibility and for the allele 7 conferring protection. However, one must be cautious in the interpretation of these data, as allele 9 is much more haplospecific for AH 18.2 than allele 7, which appear in several other haplotypes. Any ulterior analysis was prevented given the lack of DR3-DQ2/DR4-DQ8 control subjects displaying the AH 18.2.
Table 3 Comparison of the carrier rate of HLA class III genetic markers in the DR2-DQ6/DR3-DQ2/DR4-DQ8 negative diabetic and control cohorts.
Marker Allele T1D Controls OR p
D6S273 n = 29 n = 251
1 6 30 1.92 0.15
2 2 24 0.70 0.48
3 18 121 1.76 0.16
4 13 134 0.71 0.38
5 10 114 0.63 0.26
BAT-2 n = 29 n = 263
2 20 108 3.19 0.004
3 13 98 1.37 0.43
7 10 127 0.56 0.16
8 3 20 1.40 0.40
TNFab n = 29 n = 234
1,5 1 10 0.80 0.65
2,1 12 48 2.74 0.01
2,3 1 6 1.36 0.56
4,5 3 25 0.96 0.63
5,5 4 21 1.62 0.29
6,5 8 61 1.08 0.86
7,4 9 51 1.61 0.26
10,4 3 73 0.25 0.02
11,4 2 21 0.75 0.52
MICA n = 29 n = 233
4 8 45 1.59 0.30
5 6 62 0.72 0.49
5.1 5 74 0.45 0.11
6 22 141 2.05 0.11
9 12 78 1.40 0.40
Table 4 DQA1-DQB1 haplotypes in DR2-DQ6/DR3-DQ2/DR4-DQ8 negative T1D patients and controls.
Haplotype T1D (n = 29) Controls (n = 282) OR p
0101–0501 16 86 2.81 0.007
0102–0604 2 20 0.96 0.66
0103–0603 1 50 0.17 0.03
0201–0202 16 97 2.35 0.03
0201–0303 1 27 0.34 0.24
0401–0402 5 21 2.59 0.08
0501-0301 6 120 0.35 0.02
Finally, we tried to ascertain the role as T1D susceptibility/protection marker of the microsatellite located 4.9 Mb telomeric of DQ in extended class I, D6S2223. In Northern European countries allele D6S2223*3 was associated with a reduction in risk independent of the class II effect [9,10] and allele D6S2223*1 accounted for susceptibility to T1D [6]. These associations were not observed in Sardinia [11]. In our DR3, DR4 and DR2-negative study group we did not find any association (allele D6S2223*3 for protection: 28/29 in T1D patients vs. 223/234 in healthy controls, OR = 1.38, p = 0.61; allele D6S2223*1 for susceptibility: 0/29 vs. 2/234, p = 0.79). However, the limited number of patients after stratification decreases the statistical power of detection for a positive association in our population.
Discussion
Our observations support the distinctive effect on type 1 diabetes susceptibility of the AH 18.2 among all the other DR3-positive haplotypes in the Spanish population. Both, allelic and phenotypic frequencies showed the increased risk conferred by the DR3-TNFa1b5 haplotype. Moreover, the different behaviour in terms of T1D predisposition of both DR3-positive main haplotypes, AH 8.1 and 18.2, argues in favour of a second susceptibility locus besides the standard class II molecules (DR and DQ) in the latter.
Previous studies reported the contribution to T1D susceptibility of a second region in the MHC besides DQ-DR genes. Among them, one concluded that this second region extended between HLA-B and BAT-3 in the highest risk DR3/DR4 genotype [12], while other mapped the critical region around the microsatellite D6S273 centromeric to TNF [13]. TNF-α has been identified as a diabetogenic effector molecule synergistic with IFN-γ in the induction of β-cell apoptosis [14-16] and therefore, it was of interest to consider the TNF region located in class III MHC as a candidate susceptibility locus. Transfected cells containing BAT1 promoter fragments from the 8.1 AH exhibited lower reporter activity compared to the 7.1 AH. Since the 7.1 and 8.1 AH are associated with resistance and susceptibility to T1D respectively, the observed effects are claimed to have significant implications for the pathogenesis of this disease [17]. We decided to study this region in our population. Our findings of BAT-2*2, TNFa2b1 and TNFa10b4 as genetic markers of the risk conferred by the DQ-DR locus agrees with those previous reports in terms of location. Most probably the etiological variant(s) will be in linkage disequilibrium with the BAT-2*2 allele. However, given the low recombination rate of the haplotype 18.2 (shown by the TNFa1b5 present in a DR3-negative context, see Table 3), and being BAT-2*2 a common allele, this microsatellite could mark a different susceptibility gene in the AH 18.2 and in other haplotypes. Clearly, the analysis of genetic markers in class II and class III provided an emerging pattern of susceptibility/protection haplotypes in addition to the classical DR3/DR4/DR2 determinants. In this sense, one could speculate that the traditional emphasis on class II alleles alone as responsible for increasing the T1D risk could be an oversimplification of a more complex reality where class III markers remained in the background.
Population studies provide invaluable information helping to map predisposition loci. A good example is the role of D6S2223 as a marker of autoimmune diabetes in several Nordic populations [9] which has been reproduced neither in Sardinia [11] nor in Spain (present data).
A large number of studies from different populations including the Spanish one [18] have indicated that common allelic variants at the class II HLA-DRB1, -DQA1 and -DQB1 loci are associated with T1D. These MHC class II molecules play an important role in the presentation of peptide antigens after intracellular processing to CD4 T- lymphocytes. The correlation between the relative T1D predisposition of class II alleles and the structure of their proteins has also been described [19]. However, the presence of HLA class II molecules alone does not by itself cause disease. The HLA transgenic mice develop diabetes when there is an islet "insult", even if the islet "insult" is, itself, not sufficient to precipitate disease [17].
Conclusion
Our data prove that there are additional susceptibility/protection factors besides DR3/DR4/DR2 in HLA in the Spanish population and particular combinations of these minor risk factors could have important effects on susceptibility. The results of the present study provide evidence of the importance of the whole HLA in T1D risk gradient after stratification for the DR3, DR4 and DR2 determinants. The accent placed on class II in the canonical DQ2-DR3/DQ8-DR4 haplotypes is not necessarily extensive to other associated haplotypes where susceptibility might map to different MHC loci.
Methods
302 T1D patients and 529 healthy unrelated controls, both groups composed of Caucasian individuals from the same Madrid area, were consecutively recruited after informed consent obtained from the indexed subjects or their parents when the patient was a child. The T1D patients were diagnosed at two hospitals in Madrid and the control subjects were collected among healthy blood donors. The age at onset for the T1D patients (150 men and 152 women) range 1–55 years old (median onset 15 years old). Diagnosis was based on patients' clinical features and laboratory data according to the criteria of the American Diabetes Association (ADA). All subjects were insulin-dependent at the time of the study. Ab-positivity was studied for GAD, IA-2, insulin and ICA. The protocol followed the principles expressed in the Declaration of Helsinki and was approved by the Hospital Ethics Committee.
Both patients and controls were genotyped for HLA-DQB1, DQA1 and DRB1 alleles [20] for 6 microsatellite markers (D6S273, BAT-2, TNFa, TNFb, MICA and D6S2223) as previously described [20-25]. Microsatellite alleles were ascertained using an ABI Prism™ 310 automatic sequencer (Applied Biosystems, Foster City, CA, USA). Each sample included an internal size standard (TAMRA 500, Applied Biosystems) to achieve a highly consistent measure.
Allelic distribution and frequencies of the haplotypes present in the control and diabetic cohorts were determined using a software for population genetics data analysis (Arlequin ver 2.000. Schneider S, Roessli D and Excoffier L. University of Geneva, Switzerland). Haplotypes were also corroborated by our previous findings in families affected by other immune related diseases, by LD studies in controls, and by the recent literature [6].
The frequencies of each marker allele in patients versus controls were compared by a standard chi-square test using a statistical software package (EPI-INFO v. 6.02; Center for Disease Control & Prevention CDC, U.S.A.) and a result was considered significant if p < 0.05. Statistics are performed upon the individuals with available data in the markers under comparison; therefore the final number could differ among different analyses. As our working hypothesis was to verify in the Spanish population the presumed special association with T1D of the characteristic alleles of the AH 18.2, there was no need for Bonferroni corrections in these initial comparisons.
Abbreviations
T1D type 1 diabetes
MHC Major Histocompatibility Complex
HLA human leukocyte antigen
AH ancestral haplotype
LD linkage disequilibrium
TNF tumor necrosis factor
OR odds ratio
LTA lymphotoxin alpha.
Freq. frequency
Authors' contributions
EU carried out the genotyping of some samples, participated in the statistical analysis and drafted the major part of the manuscript.
JLS carried out the genotyping of most of the patients and a great part of the controls, participated in the statistical analysis and drafted the manuscript. He participated in the design and coordination of the study.
HdlC made the diagnosis, and collaborated in the statistical analysis
AM carried out a great part of the genotyping of control samples and participated in the statistical analysis and drafted the manuscript
JM participated in the genotyping and recollection of samples, and in the statistical analysis.
JMI made the diagnosis, and collaborated in the statistical analysis
CM made the diagnosis, and collaborated in the statistical analysis
MFA participated in the design and coordination of the study and helped to collect the DNA samples.
EgdlC conceived of the study and helping its coordination, participated in the statistical analysis and drafted the manuscript
All authors have red and approved the final manuscript.
Acknowledgements
We thank Carmen Martínez for expert technical assistance. Elena Urcelay is recipient of a "Ramón y Cajal" contract of the Spanish Science and Technology Ministry. Alfonso Martínez is a recipient of a FIS contract (CP04/00175).
==== Refs
Mathis D Vence L Benoist C beta-Cell death during progression to diabetes Nature 2001 414 792 798 11742411 10.1038/414792a
Field LL Genetic linkage and association studies of Type I diabetes: challenges and rewards Diabetologia 2002 45 21 35 11845220 10.1007/s125-002-8241-7
Robinson WP Barbosa J Rich SS Thomson G Homozygous parent affected sib pair method for detecting disease predisposing variants: application to insulin dependent diabetes mellitus Genet Epidemiol 1993 10 273 288 8224807
Lie BA Todd JA Pociot F Nerup J Akselsen HE Joner G Dahl-Jorgensen K Ronningen KS Thorsby E Undlien DE The predisposition to type 1 diabetes linked to the human leukocyte antigen complex includes at least one non-class II gene Am J Hum Genet 1999 64 793 800 10053014 10.1086/302283
Degli-Esposti MA Leaver AL Christiansen FT Witt CS Abraham LJ Dawkins RL Ancestral haplotypes: conserved population MHC haplotypes Hum Immunol 1992 34 242 252 1464552 10.1016/0198-8859(92)90023-G
Johansson S Lie BA Todd JA Pociot F Nerup J Cambon-Thomsen A Kockum I Akselsen HE Thorsby E Undlien DE Evidence of at least two type 1 diabetes susceptibility genes in the HLA complex distinct from HLA-DQB1, -DQA1 and -DRB1 Genes Immun 2003 4 46 53 12595901 10.1038/sj.gene.6363917
De La Concha EG Fernandez-Arquero M Gual L Vigil P Martinez A Urcelay E Ferreira A Garcia-Rodriguez MC Fontan G MHC Susceptibility Genes to IgA Deficiency Are Located in Different Regions on Different HLA Haplotypes J Immunol 2002 169 4637 4643 12370403
Fernandez L Fernandez-Arquero M Gual L Lazaro F Maluenda C Polanco I Figueredo MA Gomez de la Concha E Triplet repeat polymorphism in the transmembrane region of the MICA gene in celiac disease Tissue Antigens 2002 59 219 222 12074713 10.1034/j.1399-0039.2002.590307.x
Lie BA Sollid LM Ascher H Ek J Akselsen HE Ronningen KS Thorsby E Undlien DE A gene telomeric of the HLA class I region is involved in predisposition to both type 1 diabetes and coeliac disease Tissue Antigens 1999 54 162 168 10488743 10.1034/j.1399-0039.1999.540207.x
Koeleman BV De Groot KN Van Der Slik AR Roep BO Giphart MJ Association between D6S2223 and type I diabetes independent of HLA class II in Dutch families Diabetologia 2002 45 598 599 12032641 10.1007/s00125-001-0725-1
Zavattari P Lampis R Motzo C Loddo M Mulargia A Whalen M Maioli M Angius E Todd JA Cucca F Conditional linkage disequilibrium analysis of a complex disease superlocus, IDDM1 in the HLA region, reveals the presence of independent modifying gene effects influencing the type 1 diabetes risk encoded by the major HLA-DQB1, -DRB1 disease loci Hum Mol Genet 2001 10 881 889 11285254 10.1093/hmg/10.8.881
Degli-Esposti MA Abraham LJ McCann V Spies T Christiansen FT Dawkins RL Ancestral haplotypes reveal the role of the central MHC in the immunogenetics of IDDM Immunogenetics 1992 36 345 356 1356098 10.1007/BF00218041
Moghaddam PH Zwinderman AH de Knijff P Roep BO Schipper RF Van der Auwera B Naipal A Gorus F Schuit F Giphart MJ TNFa microsatellite polymorphism modulates the risk of IDDM in Caucasians with the high-risk genotype HLA DQA1*0501- DQB1*0201/DQA1*0301-DQB1*0302. Belgian Diabetes Registry Diabetes 1997 46 1514 1515 9287056
Mandrup-Poulsen T Bendtzen K Dinarello CA Nerup J Human tumor necrosis factor potentiates human interleukin 1-mediated rat pancreatic beta-cell cytotoxicity J Immunol 1987 139 4077 4082 3320203
Campbell IL Iscaro A Harrison LC IFN-gamma and tumor necrosis factor-alpha. Cytotoxicity to murine islets of Langerhans J Immunol 1988 141 2325 2329 3139756
Suk K Kim S Kim YH Kim KA Chang I Yagita H Shong M Lee MS IFN-gamma/TNF-alpha synergism as the final effector in autoimmune diabetes: a key role for STAT1/IFN regulatory factor-1 pathway in pancreatic beta cell death J Immunol 2001 166 4481 4489 11254704
Wong FS Wen L The Study of HLA Class II and Autoimmune Diabetes Curr Top Med Chem 2003 3 1 15 12570684
Escribano-de-Diego J Sanchez-Velasco P Luzuriaga C Ocejo-Vinyals JG Paz-Miguel JE Leyva-Cobian F HLA class II immunogenetics and incidence of insulin-dependent diabetes mellitus in the population of Cantabria (Northern Spain) Hum Immunol 1999 60 990 1000 10566601 10.1016/S0198-8859(99)00077-4
Cucca F Lampis R Congia M Angius E Nutland S Bain SC Barnett AH Todd JA A correlation between the relative predisposition of MHC class II alleles to type 1 diabetes and the structure of their proteins Hum Mol Genet 2001 10 2025 2037 11590120 10.1093/hmg/10.19.2025
Kimura A Sasazuki T Tsuji K, Aizawa M, Sasazuki T Eleventh International Histocompability workshop reference protocol for the HLA DNA-typing technique HLA 1991 Proceedings of the Eleventh International Histocompatibility Workshop and Conference 1991 1 Oxford: Oxford University Press 397 419
Nedospasov SA Udalova IA Kuprash DV Turetskaya RL DNA sequence polymorphism at the human tumor necrosis factor (TNF) locus. Numerous TNF/lymphotoxin alleles tagged by two closely linked microsatellites in the upstream region of the lymphotoxin (TNF-beta) gene J Immunol 1991 147 1053 1059 1861069
Martin M Mann D Carrington M Recombination rates across the HLA complex: use of microsatellites as a rapid screen for recombinant chromosomes Hum Mol Genet 1995 4 423 428 7795597
Ota M Katsuyama Y Mizuki N Ando H Furihata K Ono S Pivetti-Pezzi P Tabbara KF Palimeris GD Nikbin B Trinucleotide repeat polymorphism within exon 5 of the MICA gene (MHC class I chain-related gene A): allele frequency data in the nine population groups Japanese, Northern Han, Hui, Uygur, Kazakhstan, Iranian, Saudi Arabian, Greek and Italian Tissue Antigens 1997 49 448 454 9174136
Vorechovsky I Cullen M Carrington M Hammarstrom L Webster AD Fine mapping of IGAD1 in IgA deficiency and common variable immunodeficiency: identification and characterization of haplotypes shared by affected members of 101 multiple-case families J Immunol 2000 164 4408 4416 10754342
Foissac A Salhi M Cambon-Thomsen A Microsatellites in the HLA region: 1999 update Tissue Antigens 2000 55 477 509 10902606 10.1034/j.1399-0039.2000.550601.x
| 15842729 | PMC1097726 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 Apr 20; 6:56 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-56 | oa_comm |
==== Front
BMC GeriatrBMC Geriatrics1471-2318BioMed Central London 1471-2318-5-61582632010.1186/1471-2318-5-6Research ArticleRisk factors for delirium in acutely admitted elderly patients: a prospective cohort study Korevaar Johanna C [email protected] Munster Barbara C [email protected] Rooij Sophia E [email protected] Department of Clinical Epidemiology and Biostatistics, Academic Medical Centre, Univ. of Amsterdam, The Netherlands2 Department of Internal Medicine, Academic Medical Centre, Univ. of Amsterdam, The Netherlands2005 13 4 2005 5 6 6 20 12 2004 13 4 2005 Copyright © 2005 Korevaar et al; licensee BioMed Central Ltd.2005Korevaar 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
Delirium is a neuropsychiatric syndrome frequently observed in elderly hospitalised patients and can be found in any medical condition. Due to the severe consequences, early recognition of delirium is important in order to start treatment in time. Despite the high incidence rate, the occurrence of delirium is not always identified as such. Knowledge of potential risk factors is important. The aim of the current study is to determine factors associated with the occurrence of a prevalent delirium among elderly patients acutely admitted to an internal medicine ward.
Methods
All consecutive patients of 65 years and over acutely admitted to the Department of Internal Medicine of the Academic Medical Centre, Amsterdam, a university hospital, were asked to participate. The presence of delirium was determined within 48 hrs after admission by an experienced geriatrician.
Results
In total, 126 patients were included, 29% had a prevalent delirium after acute admission. Compared to patients without delirium, patients with delirium were older, more often were cognitively and physically impaired, more often were admitted due to water and electrolyte disturbances, and were less often admitted due to malignancy or gastrointestinal bleeding. Independent risk factors for having a prevalent delirium after acute admission were premorbid cognitive impairment, functional impairment, an elevated urea nitrogen level, and the number of leucocytes.
Conclusions
In this study, the most important independent risk factors for a prevalent delirium after acute admission were cognitive and physical impairment, and a high serum urea nitrogen concentration. These observations might contribute to an earlier identification and treatment of delirium in acutely admitted elderly patients.
==== Body
Background
Delirium is a complex neuropsychiatric syndrome with an acute onset, characterized by disturbances of consciousness, attention, cognition, and perception. Delirium can be found in any medical condition and is the most common reason for acute cognitive dysfunction in hospitalised elderly patients [1-3]. This syndrome occurs in about 10 to 25% of all acute admissions to a general hospital. The frequency in older patients is higher, 20% to 40% [4]. Due to the aging of the population, the absolute number may increase in the future.
Delirium has been associated with a poor outcome: prolonged hospitalisation, decreased cognitive and physical functioning, increased nursing home admission, and with a threefold higher morbidity and mortality risk [5,6]. Due to the severe consequences, recognition of delirium is important to start treatment in time. Despite the high incidence rate, the occurrence of delirium is not always identified. Several studies reported that between 32% and 67% of delirious patients were not recognized by their physicians [7,8]. There are a number of hypotheses for this lack of recognition. First, delirium is not always regarded as an important clinical syndrome, because it is varied and the multiple aetiology defies the classic disease model to look for a single cause of disease. In addition, delirium is often believed to present with agitation, hallucination, and inappropriate behaviour, whereas it also often presents with lethargy and decreased activity. Finally, the fluctuating course of the delirium may confound the diagnosis [9].
The patho-physiology of delirium is still poorly understood, although a number of mechanisms have been hypothesized. Delirium might be the result of changes in neurotransmitter systems, in cytokines, and in the lipid metabolism [10-16]. Delirium and Alzheimer's disease (AD) may share several patho-physiological features as they share several symptoms [17]. Moreover, preliminary results support the assumption that genetic variation plays a role [18].
Several studies have examined risk factors that might predispose, or influence the development of delirium. A systematic review identified 27 articles studying 61 different risk factors [19]. Results of these 27 studies were not conclusive. Most studies found an increased risk for delirium in dementia, medical illness, and alcohol abuse. For risk factors like medication, male gender, or serum concentrations, of e.g. urea, creatinine, sodium, potassium or glucose, some studies did find an increased risk for delirium, whereas other studies did not find an increased risk of these factors [20-23]. Finally, most of these studies determined risk factor for new-onset delirium in hospitalised patients [24,25]. Yet, prevalent delirium can be a major problem as well, especially among acutely admitted patients. A prompt recognition of the syndrome is important to initiate appropriate treatment as soon as possible. Therefore, more knowledge of potential risk factors for acutely admitted patients is important.
The aim of the current study is to determine factors associated with a prevalent delirium among acutely admitted elderly patients to an internal medicine ward.
Methods
Patients
All consecutive patients of 65 years and over acutely admitted to the Department of Internal Medicine of the Academic Medical Centre, Amsterdam, a university teaching hospital, were invited. Patients admitted to our hospital are sometimes referred because of the university function, but about 60% are directly admitted to our hospital. Patients were excluded from the study if they were unable to speak or understand Dutch or English, if they or their relatives did not give permission for the study, if they came from or were transferred to another ward than Internal Medicine, or left the ward within 48 hours. Due to logistic limitations, a random sample of included patients was taken for the current study regarding risk factors and genetic variation. More detailed information regarding medication and DNA was collected of these selected patients. Inclusion period was between January 2003 and February 2004. Before enrolment, informed consent was obtained from the patient or substitute decision-maker. The hospital's Medical Ethics Committee approved the study.
Procedures
Members of the team completed an initial multidisciplinary evaluation for all study participants within 48 hrs after admission. The team was composed of a geriatric physician, a fellow in geriatric medicine, and two research nurses trained in geriatric medicine. Demographic and clinical data were collected. Severity and number of comorbidities were scored with the Charlson comorbidity index [26]. The final score was divided in 3 categories; mild (0 or 1 point), moderate (2 or 3 points) and severe (more than 3 points). Biochemistry values, i.e. urea nitrogen, creatinine, glucose, haemoglobin, natrium, and potassium concentrations, and leucocytes count were obtained from a serum blood taken within 48 hrs after admission.
Within 48 hrs after admission, the research nurses interviewed patients, medical and nursing staff. Cognitive impairment was recorded by two validated instruments (MMSE, IQCODE) at the time of hospital admission. The MMSE (Mini Mental State Examination) is the most widely used screening instrument for detection of cognitive impairment in the elderly [27]. Many studies have shown that the MMSE has got a high construct validity and test-retest reliability [28]. The MMSE measures cognitive functioning on a scale of 0 (poor) to 30 (excellent), with a score less than 24 indicating cognitive impairment. The IQCODE (Informant Questionnaire on COgnitive DEcline) assesses the possible presence of dementia before admission based on the response of an informant who had known the patient for at least ten years and could assess any decline in memory or cognition [29]. The informant was asked to recollect the situation 2 weeks before admission and to compare it with the situation 10 years before. The score is an average of the 16-item scores, each rated from 1 (much improved) to 5 (much worse). Patients with a mean score of 3.9 or more were considered to have dementia [30]. Final classification for having cognitive impairment was based on the MMSE score for patients without delirium, whereas for patients with delirium, the combination of both instruments (MMSE and IQCODE) was applied. In case of conflicting outcome, the score of the IQCODE was used.
The geriatric physician or the fellow scored the presence of delirium within 48 hrs after admission with the CAM (Confusion Assessment Method). The CAM is a structured interview of delirium symptoms based on the Diagnostic and Statistical Manual of Mental Disorders criteria (DSM-III-R). This instrument have been found reliable, sensitive and specific [31]. The delirium could have been present at admission, or developed within 48 hrs after admission.
The KATZ ADL scale is a 15-item scale for measuring functional status in a geriatric population. The KATZ-ADL consists of one scale for patients and one for their relative or informant [32]. KATZ-ADL as scored by the informant of a patient was taken as the final ADL score. In case this score was missing, the KATZ-ADL score of the patient was taken. Once more, the informant was asked to recall the situation 2 weeks before admission. The nurses that collected the risk scores were blinded to the presence or absence of delirium in the patient.
All medication before hospital admission was registered. The number of prescribed drugs was scored. Psychopharmaca included benzodiazepines, antidepressive medication such as selective serotonin reuptake inhibitors and tricyclic antidepressants, and antipsychotic medication.
Based on the literature and our clinical experience, we predefined five different types of drugs as suspected of being a high-risk drug for obtaining delirium [33-35]; namely anticholinergic medication, benzodiazepines, narcotic analgetics, corticosteroids and antihistaminics. Cholinergic drugs were classified according to the list of Han et al.[35], all drugs scored with 3 points were considered cholinergic drugs. We completed this list with parasympathicolytics. Benzodiazepines included sedative-hypnotics and anxiolytics. The antihistamincs group contains only H1 receptor blocking agents.
Statistical analysis
Standard descriptive statistics were used. Univariate and multivariate logistic regression analysis was performed to identify risk factors of delirium. All variables with a p-value <0.20 in the univariate analysis were used in the multivariate model. A backward elimination procedure was used to define the final risk factors. All variables with a p-value >0.10 were eliminated from the model. To confirm the final risk factors, the multiple logistic regression analysis was repeated with a forward selection procedure (all variables with a p-value <0.10 were included).
Results
During the inclusion period 576 patients were admitted to the internal ward. Of these patients, 88 patients came from another ward, resulting in 488 eligible patients. 182 patients were not included because no informed consent was given or they were unable to speak or understand Dutch or English. In total, 306 patients were included, a random sample of 126 patients was selected for the current study. Non-selected and selected patients were similar regarding mean age, the percentages male, and the frequency of patients with a prevalent delirium. Mean age of the non-selected patients was 78.1 (SD: 8.5), 45% were male and 28% of these patients had a prevalent delirium after acute hospital admission. For the 126 selected patients mean age was 79.1 (7.8), 41% were male and 36 patients (29%) had a prevalent delirium after acute hospital admission.
Baseline characteristics of the 126 selected patients with and without a prevalent delirium are presented in Table 1. Patients with delirium were significantly older, had more often cognitive impairment, and were more impaired in daily activities compared to patients without delirium. Reason for admission was less frequently a malignancy, or a gastrointestinal bleeding, and more frequently water or electrolyte disturbances for patients with delirium. On average, delirious patients had a significantly higher level of serum urea nitrogen at admission.
Table 1 Baseline characteristics of acutely admitted elderly patients with and without a prevalent delirium after acute admission.
Patients with delirium Patients without delirium
Number 36 (29%) 90 (71%)
Age (yrs.) * 82.1 (7.2) 77.8 (7.8)
65–70 (%) 8% 17%
70–75 (%) 14% 29%
75–80 (%) 14% 14%
80–85 (%) 25% 18%
≥ 85 (%) 39% 22%
Male (%) 50% 38%
Education (yrs.) 8.9 (2.1) 8.8 (3.2)
Comorbidity (%)
Mild 25% 19%
Moderate 31% 31%
Severe 44% 50%
Admission reason (%) *
Infectious disease 53% 56%
Malignancy 6% 17%
Gastrointestinal bleeding 3% 11%
Water and electrolyte disturbances 19% 2%
Other 19% 14%
MMSE: Cognitive impairment (%) * 89% 41%
IQCODE: Cognitive impairment (%) * 89% 24%
Katz ADL (%) *
0 0% 10%
1–3 9% 37%
4–6 15% 19%
≥ 7 77% 35%
Biochemistry
Urea nitrogen (mmol/L) * 15.9 (13.6) 10.6 (6.2)
Creatinine (μmol/L) 175 (223) 137 (193)
Glucose (mmol/L) 8.5 (5.1) 8.5 (5.3)
Haemoglobin (mmol/L) 7.6 (1.5) 7.2 (1.6)
Sodium (mmol/L) 134.6 (8.5) 133.9 (5.6)
Potassium (mmol/L) 4.1 (0.8) 4.0 (0.7)
Leucocytes (× 109/L) 11.2 (5.8) 13.5 (7.2)
CRP (mg/L) 119.3 (113.6) 99.0 (93.5)
Mean values (SD) are given for continuous variables
* p-value <0.05: patients with delirium versus without prevalent delirium after acute admission.
Table 2 presents the type of medication prescribed to more than 20% of the patients. The majority of the most frequently prescribed drugs are similar for patients with and without delirium, except for psychopharmaca. This type of drugs is prescribed to 31% of the patients with delirium, whereas 17% of the patients without delirium were taking these drugs. Patients without a prevalent delirium were on nearly 5 different prescribed drugs before admission, whereas delirious patients were taking on average 4.4 different drugs before admission. This difference was not statistically significant.
Table 2 Medication prescribed before admission to more than 20% of the acutely admitted elderly patients with and without a prevalent delirium after acute admission; descending order. Number of patients (percentages) are given per medication type.
Patients with delirium < 48 hours Number (%) Patients without delirium Number (%)
Glucose lowering medication 14 (39%) Diuretics 40 (44%)
Antithrombotics 13 (36%) Antithrombotics 28 (31%)
Gastrointestinal medication 13 (36%) Analgetics 27 (30%)
Diuretics 12 (33%) Antihypertensives 27 (30%)
Psychopharmaca 11 (31%) Glucose lowering medication 27 (30%)
Sympathicolytics 10 (28%) Gastrointestinal medication 27 (30%)
Analgetics 9 (25%) Sympathicolytics 26 (29%)
Spasmolytics and vasodilators 9 (25%) Spasmolytics and vasodilators 19 (21%)
Antihypertensives 8 (22%)
Number of drugs used before admission (mean (SD)) 4.4 (3.2) 4.9 (3.6)
Information on the number of patients taking one or more of the predefined high-risk medications is presented in Table 3. Delirious patients obtained more often drugs suspected of being a drug with a high risk, especially benzodiazepines, 17% versus 10%, and narcotic analgetics (14% versus 7%). Yet these differences were not significant.
Table 3 List of drugs suspected of being a high-risk drug for a delirium, presented for patients with and without a prevalent delirium after acute admission.
Patients with delirium Patients without delirium
(N = 36) (N = 90)
Suspected high-risk medication before admission (% patients)
Benzodiazepines (%) 17% 10%
Narcotic analgetics (%) 14% 7%
Corticosteroids (%) 14% 10%
Antihistaminics (%) 3% 2%
Cholinergic drugs (%) 9% 7%
Any of these 5 medications (%) 34% 27%
* p-value <0.05: patients with prevalent delirium versus without delirium after acute admission.
Based on univariate logistic regression analysis, risk factors for a prevalent delirium after acute admission are higher age, cognitive impairment, impaired physical functioning, admission reason, and urea nitrogen level (Table 4). No increased risk was observed for mean number of prescribed drugs, or the use of drugs suspected of being a high-risk drug. Results of the multivariate logistics regression analysis are presented in Table 5. Independent predictors for an increased risk for a prevalent delirium after acute admission were cognitive impairment, impaired physical functioning, increased urea nitrogen level, and the number of leucocytes. The backward and the forward selection procedure resulted in the same final risk factors (data not shown).
Table 4 Univariate logistic regression analysis of risk factors for having a prevalent delirium after acute admission
Variable Unadjusted Hazard ratio (95% CI) p-value
Age (yrs.) 1.08 (1.02–1.13) <0.01
Male 1.62 (0.74–3.53) 0.23
Comorbidity
Mild 1.00
Moderate 0.74 (0.26–2.16) 0.58
Severe 0.68 (0.25–1.81) 0.43
Cognitive impairment 10.98 (3.56–33.83) <0.01
Katz ADL
0–4 1.00
5–6 4.22 (0.90–19.92) 0.07
≥ 7 11.76 (3.23–42.77) <0.01
Admission reason
Infectious disease 0.71 (0.25–2.04) 0.52
Malignancy 0.25 (0.04–1.41) 0.12
Gastrointestinal bleeding 0.19 (0.02–1.77) 0.14
Water and electrolyte disturbances 6.50 (1.05–40.13) 0.04
Other 1.00
Urea (mmol/L) 1.06 (1.01–1.11) 0.01
Creatinine (μmol/L) 1.00 (1.00–1.00) 0.36
Leucocytes (* 109/L) 0.95 (0.88–1.01) 0.11
CRP 1.00 (1.00–1.01) 0.31
Number of medications 0.96 (0.86–1.08) 0.51
Benzodiazepines (%) 1.84 (0.60–5.62) 0.29
Narcotic analgetics (%) 2.31 (0.66–8.1) 0.19
Corticosteroids (%) 1.48 (0.46–4.78) 0.51
Antihistaminics (%) 1.28 (0.11–14.58) 0.84
Cholinergic drugs (%) 1.30 (0.31–5.50) 0.72
Any of these 5 medications (%) 1.41 (0.61–3.27) 0.42
Table 5 Multivariate logistic regression of risk factors for having a prevalent delirium after acute admission. Backward selection procedure of all variables with a p-value <0.20 in the univariate regression analysis
Variable Adjusted Hazard ratio (95% CI) p-value
Cognitive impaired 9.48 (2.27–39.54) <0.01
Katz ADL
0–4 1.00
6–5 8.14 (1.08–61.31) 0.04
≥ 7 14.13 (2.26–88.24) <0.01
Urea (mmol/L) 1.10 (1.02–1.18) <0.01
Leucocytes (* 109/L) 0.87 (0.79–0.97) 0.01
Discussion
In this prospective cohort study, nearly 30% of acutely admitted elderly patients had delirium within 48 hrs after admission. Increased risk for developing delirium was found for cognitive and physical impairment, elevated urea nitrogen level, and the number of leucocytes. No increased risk for developing delirium was noticed for comorbidity or type or number of medication.
A possible reason for not finding an increased risk of comorbidity could be the limited number of included patients. The use of a more liberal inclusion criteria in the multivariate logistic regression analysis, elimination of all variables with a p-value >0.15, resulted into the same final model. In the literature, inconclusive results can be found regarding the impact of the severity of the disease on the risk of delirium [20,23-25]. So, the limited number of included patients does not seem to be the sole reason for no association between comorbidity and the risk of delirium. Another possible explanation could be the balancing relationship of comorbidity with cognitive impairment. The more cognitive impaired, the less physical illness might be needed to become delirious.
No effect of type or number of drugs prescribed on the risk of delirium was observed. In the current literature, results on the association between anticholinergic drugs and delirium are conflicting. Both in patients with an incident as in patients with a prevalent delirium, some studies did find an increased risk [33,34,36], whereas other studies did not find an effect of anticholinergic drugs [22,23,37,38]. In addition, it could be possible that not just the medication type itself is associated with the risk of delirium, but rather the dosage of the medication, or a recent change in type or dosage of prescribed medication. We did not collect data on these detailed topics, nor did we collect the compliance of the patient to the prescribed medication.
Most studies report an increased risk associated with older age. A reason for not confirming this finding is the strong correlation between age and impaired physical functioning in our study. Indeed, if we excluded the KATZ ADL score, higher age is an independent risk factor for having delirium within 48 hrs after admission.
Our results could have been obscured due to selectively inclusion of patients, as not all patients gave informed consent. However, we found that nearly 30% of the acutely admitted patients had a prevalent delirium after acute admission. This finding is in concordance with the literature; a frequency between 20% and 40% is usually found. Moreover, the frequency in demented patients was much higher, 46% of these patients had delirium. This finding is supported by the literature as well.
Cognitive impairment was the strongest risk factor for a prevalent delirium; we found a ninefold increased risk. This effect is also reported in other studies for the development of delirium after admission. The amount of reported increased risk varies between 2.8 and 9.0 [20,23,24]. An increased urea nitrogen level was associated with an increased risk of a prevalent delirium. A most likely explanation for this finding is that an increased urea nitrogen concentration is, among others, an indication of dehydration. The effect of dehydration on risk of delirium was found in other studies as well [7,24,39]. Moreover, if we did not take urea nitrogen level in the multivariate analysis, we find a significant increased risk for the admission reason water and electrolyte disturbances (data not shown). Finally, we found that a higher number of leucocytes was associated with a lower risk of delirium. We cannot explain this finding, we speculate that this might be due to the higher percentages of non-delirious patients with infectious diseases and malignancies, both associated with a higher level of leucocytes. Indeed, the highest leucocytes counts were found in the non-delirious patients with reason of admission malignancy or infectious disease.
Conclusions
In this study the independent risk factors for a prevalent delirium after acute admission are premorbid cognitive and physical impairment, and a high urea nitrogen level. These observations might contribute to an earlier identification and treatment of delirium in acutely admitted elderly patients.
Competing interest
The author(s) declare that they have no competing interests.
Authors' contributions
SR is the principal investigator of the study; she designed the protocol, supervised its progress, and was involved in the acquisition of the data. BM was involved in the planning of the study and the acquisition of the data. JK was responsible for the statistical analysis. JK and BM drafted the manuscript. All authors contributed to the interpretation of the data, revisions of the paper, and read and approved the final manuscript.
Role of the funding source
Funding source was our own hospital, it was an unrestricted grant.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors wish to thank the research nurses of the department of internal medicine for their assistance in the logistics and data collection of this study.
==== Refs
Liptzin B Levkoff SE An empirical study of delirium subtypes Br J Psychiatry 1992 161 843 845 1483173
Sandberg O Gustafson Y Brannstrom B Bucht G Clinical profile of delirium in older patients J Am Geriatr Soc 1999 47 1300 1306 10573437
Meagher DJ O'Hanlon D O'Mahony E Casey PR Trzepacz PT Relationship between etiology and phenomenologic profile in delirium J Geriatr Psychiatry Neurol 1998 11 146 149 9894733
Fick DM Agostini JV Inouye SK Delirium superimposed on dementia: a systematic review J Am Geriatr Soc 2002 50 1723 1732 12366629 10.1046/j.1532-5415.2002.50468.x
Ely EW Gautam S Margolin R Francis J May L Speroff T Truman B Dittus R Bernard R Inouye SK The impact of delirium in the intensive care unit on hospital length of stay Intensive Care Med 2001 27 1892 1900 11797025 10.1007/s00134-001-1132-2
Francis J Kapoor WN Prognosis after hospital discharge of older medical patients with delirium J Am Geriatr Soc 1992 40 601 606 1587979
Inouye SK Bogardus ST JrCharpentier PA Leo-Summers L Acampora D Holford TR Cooney LM Jr A multicomponent intervention to prevent delirium in hospitalized older patients N Engl J Med 1999 340 669 676 10053175 10.1056/NEJM199903043400901
Armstrong SC Cozza KL Watanabe KS The misdiagnosis of delirium Psychosomatics 1997 38 433 439 9314712
Inouye SK The dilemma of delirium: clinical and research controversies regarding diagnosis and evaluation of delirium in hospitalized elderly medical patients Am J Med 1994 97 278 288 8092177 10.1016/0002-9343(94)90011-6
Flacker JM Lipsitz LA Neural mechanisms of delirium: current hypotheses and evolving concepts J Gerontol A Biol Sci Med Sci 1999 54 B239 B246 10411009
van der Mast RC Pathophysiology of delirium J Geriatr Psychiatry Neurol 1998 11 138 145 9894732
Roche V Southwestern Internal Medicine Conference. Etiology and management of delirium Am J Med Sci 2003 325 20 30 12544081
van der Mast RC van den Broek WW Fekkes D Pepplinkhuizen L Habbema JD Is delirium after cardiac surgery related to plasma amino acids and physical condition? J Neuropsychiatry Clin Neurosci 2000 12 57 63 10678514
van der Mast RC Fekkes D Moleman P Pepplinkhuizen L Is postoperative delirium related to reduced plasma tryptophan? Lancet 1991 338 851 852 1681217 10.1016/0140-6736(91)91504-N
Mortazavi S Thompson J Baghdoyan HA Lydic R Fentanyl and morphine, but not remifentanil, inhibit acetylcholine release in pontine regions modulating arousal Anesthesiology 1999 90 1070 1077 10201679 10.1097/00000542-199904000-00021
Wang JC Kwon JM Shah P Morris JC Goate A Effect of APOE genotype and promoter polymorphism on risk of Alzheimer's disease Neurology 2000 55 1644 1649 11113217
Eikelenboom P Hoogendijk WJ Do delirium and Alzheimer's dementia share specific pathogenetic mechanisms? Dement Geriatr Cogn Disord 1999 10 319 324 10473931 10.1159/000017162
Gorwood P Limosin F Batel P Hamon M Ades J Boni C The A9 allele of the dopamine transporter gene is associated with delirium tremens and alcohol-withdrawal seizure Biol Psychiatry 2003 53 85 92 12513948 10.1016/S0006-3223(02)01440-3
Elie M Cole MG Primeau FJ Bellavance F Delirium risk factors in elderly hospitalized patients J Gen Intern Med 1998 13 204 212 9541379 10.1046/j.1525-1497.1998.00047.x
Marcantonio ER Goldman L Mangione CM Ludwig LE Muraca B Haslauer CM Donaldson MC Whittemore AD Sugarbaker DJ Poss R A clinical prediction rule for delirium after elective noncardiac surgery JAMA 1994 271 134 139 8264068 10.1001/jama.271.2.134
Dyer CB Ashton CM Teasdale TA Postoperative delirium. A review of 80 primary data-collection studies Arch Intern Med 1995 155 461 465 7864702 10.1001/archinte.155.5.461
Francis J Martin D Kapoor WN A prospective study of delirium in hospitalized elderly JAMA 1990 263 1097 1101 2299782 10.1001/jama.263.8.1097
Schor JD Levkoff SE Lipsitz LA Reilly CH Cleary PD Rowe JW Evans DA Risk factors for delirium in hospitalized elderly JAMA 1992 267 827 831 1732655 10.1001/jama.267.6.827
Inouye SK Viscoli CM Horwitz RI Hurst LD Tinetti ME A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics Ann Intern Med 1993 119 474 481 8357112
Inouye SK Charpentier PA Precipitating factors for delirium in hospitalized elderly persons. Predictive model and interrelationship with baseline vulnerability JAMA 1996 275 852 857 8596223 10.1001/jama.275.11.852
Charlson ME Pompei P Ales KL MacKenzie CR A new method of classifying prognostic comorbidity in longitudinal studies: development and validation J Chron Dis 1987 40 373 383 3558716 10.1016/0021-9681(87)90171-8
Folstein MF Folstein SE McHugh PR "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician J Psychiatr Res 1975 12 189 198 1202204 10.1016/0022-3956(75)90026-6
Tombaugh TN McIntyre NJ The mini-mental state examination: a comprehensive review J Am Geriatr Soc 1992 40 922 935 1512391
Jorm AF Jacomb PA The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): socio-demographic correlates, reliability, validity and some norms Psychol Med 1989 19 1015 1022 2594878
de Jonghe JF Differentiating between demented and psychiatric patients with the Dutch version of the IQCODE Int J Geriatr Psychiatry 1997 12 462 465 9178050 10.1002/(SICI)1099-1166(199704)12:4<462::AID-GPS510>3.0.CO;2-Q
Inouye SK van Dyck CH Alessi CA Balkin S Siegal AP Horwitz RI Clarifying confusion: the confusion assessment method. A new method for detection of delirium Ann Intern Med 1990 113 941 948 2240918
Weinberger M Samsa GP Schmader K Greenberg SM Carr DB Wildman DS Comparing proxy and patients' perceptions of patients' functional status: results from an outpatient geriatric clinic J Am Geriatr Soc 1992 40 585 588 1587975
Tune LE Damlouji NF Holland A Gardner TJ Folstein MF Coyle JT Association of postoperative delirium with raised serum levels of anticholinergic drugs Lancet 1981 2 651 653 6116042 10.1016/S0140-6736(81)90994-6
Tune LE Bylsma FW Benzodiazepine-induced and anticholinergic-induced delirium in the elderly Int Psychogeriatr 1991 3 397 408 1687445 10.1017/S1041610291000832
Han L McCusker J Cole M Abrahamowicz M Primeau F Elie M Use of medications with anticholinergic effect predicts clinical severity of delirium symptoms in older medical inpatients Arch Intern Med 2001 161 1099 1105 11322844 10.1001/archinte.161.8.1099
Trzepacz PT Anticholinergic Model For Delirium Semin Clin Neuropsychiatry 1996 1 294 303 10320432
Marcantonio ER Juarez G Goldman L Mangione CM Ludwig LE Lind L Katz N Cook EF Orav EJ Lee TH. The relationship of postoperative delirium with psychoactive medications JAMA 1994 272 1518 1522 7966844 10.1001/jama.272.19.1518
Kolbeinsson H Jonsson A Delirium and dementia in acute medical admissions of elderly patients in Iceland Acta Psychiatr Scand 1993 87 123 127 8447239
Cook DJ Rooke GA Priorities in perioperative geriatrics Anesth Analg 2003 96 1823 1836 12761020 10.1213/01.ANE.0000063822.02757.41
| 15826320 | PMC1097727 | CC BY | 2021-01-04 16:30:32 | no | BMC Geriatr. 2005 Apr 13; 5:6 | utf-8 | BMC Geriatr | 2,005 | 10.1186/1471-2318-5-6 | oa_comm |
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-281582320410.1186/1472-6963-5-28Research ArticleThe effect of executive walk rounds on nurse safety climate attitudes: A randomized trial of clinical units Thomas Eric J [email protected] J Bryan [email protected] Torsten B [email protected] Allan [email protected] Robert L [email protected] Department of Internal Medicine, The University of Texas Medical School at Houston, Houston, TX, USA2 Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Quality and Safety Research Group, The Johns Hopkins University School of Medicine, Baltimore, USA3 Center for AIDS Prevention Studies (CAPS), University of California, San Francisco, San Francisco, CA, USA4 Partners Healthcare System, Prudential Tower, Boston, MA, USA5 Department of Psychology, The University of Texas at Austin, Austin, TX, USA2005 11 4 2005 5 28 28 29 9 2004 11 4 2005 Copyright © 2005 Thomas et al; licensee BioMed Central Ltd.2005Thomas 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
Executive walk rounds (EWRs) are a widely used but unstudied activity designed to improve safety culture in hospitals. Therefore, we measured the impact of EWRs on one important part of safety culture – provider attitudes about the safety climate in the institution.
Methods
Randomized study of EWRs for 23 clinical units in a tertiary care teaching hospital. All providers except physicians participated. EWRs were conducted at each unit by one of six hospital executives once every four weeks for three visits. Providers were asked about their concerns regarding patient safety and what could be done to improve patient safety. Suggestions were tabulated and when possible, changes were made. Provider attitudes about safety climate measured by the Safety Climate Survey before and after EWRs. We report mean scores, percent positive scores (percentage of providers who responded four or higher on a five point scale (agree slightly or agree strongly), and the odds of EWR participants agreeing with individual survey items when compared to non-participants.
Results
Before EWRs the mean safety climate scores for nurses were similar in the control units and EWR units (78.97 and 76.78, P = 0.458) as were percent positive scores (64.6% positive and 61.1% positive). After EWRs the mean safety climate scores were not significantly different for all providers nor for nurses in the control units and EWR units (77.93 and 78.33, P = 0.854) and (56.5% positive and 62.7% positive). However, when analyzed by exposure to EWRs, nurses in the control group who did not participate in EWRs (n = 198) had lower safety climate scores than nurses in the intervention group who did participate in an EWR session (n = 85) (74.88 versus 81.01, P = 0.02; 52.5% positive versus 72.9% positive). Compared to nurses who did not participate, nurses in the experimental group who reported participating in EWRs also responded more favorably to a majority of items on the survey.
Conclusion
EWRs have a positive effect on the safety climate attitudes of nurses who participate in the walk rounds sessions. EWRs are a promising tool to improve safety climate and the broader construct of safety culture.
==== Body
Background
Many hospitals are implementing executive walk rounds (EWRs), a widely used but unstudied activity to improve patient safety [1,2]. EWRs vary from hospital to hospital, but in general they consist of visits by hospital executives to patient care areas to discuss patient safety issues with providers. EWRs enlist leadership to break down the significant barriers to discuss human error in healthcare. The executive may ask providers to discuss specific events or general processes that could put patients at risk for harm, they ask for suggestions to improve safety, and verbalize their commitment to improving safety. Discussions are documented and lead to action which is followed by feedback to participants. EWRs help hospitals identify opportunities to improve care processes by utilizing the wisdom of frontline providers, they demonstrate the executives' and the organization's commitment to patient safety, and they may improve provider attitudes about safety-related issues. These attitudes are an important part of what is often called a hospital's safety culture [3].
Improving safety culture is The National Quality Forum's first of thirty safe practices for better healthcare [4]. The concept is derived from industries and organizations such as aerospace (NASA), nuclear power, aviation, and the military (naval aircraft carriers and nuclear submarines) that are known for their ability to reliably deal with risky processes [3]. One important driver of safety in these settings is a very explicit commitment to safety by leaders. Other components of a good safety culture include a focus on system improvement instead of blaming individuals for mistakes, reporting and learning from errors, and infrequent unsafe acts. A defective safety culture was highlighted as one of the organizational causes of the recent space shuttle Columbia accident [5] and has been cited as causes of high profile adverse events in hospitals.
Hospitals are feeling pressure to act to improve the safety-related attitudes that are part of safety culture [4]. The true prevalence of EWRs is not known, but at least 200 hospitals have used EWRs through participation in Institute for Healthcare Improvement collaboratives, and nine hospitals in Boston are participating in a 3 year study of EWRs. Over 100 more have requested a database to collect EWR data. However, despite their face validity in the eyes of these early adopters, no evaluation of EWRs has been reported so many questions exist about how EWRs should be conducted and whether or not they are effective. The purpose of this study was to measure the impact of EWRs on one important driver of safety culture – provider attitudes about safety climate as measured by The University of Texas Safety Climate Survey.
Methods
Overview
Safety culture has been be defined as "the product of individual and group values, attitudes, perceptions, competencies, and patterns of behavior that determine the commitment to, and the style and proficiency of, an organization's health and safety management [3]." Safety climate, the primary outcome measure in this study, is comprised of the provider attitudes about patient safety. Safety climate is a part of safety culture (attitudes are part of both definitions) and we measure it with the Safety Climate Survey (described in detail below). We theorized that EWRs would improve individual provider attitudes, which in turn would lead to improved safety climate (the composite of provider attitudes).
Our primary hypothesis was that EWRs would improve safety climate in clinical units. An underlying assumption was that attitudes would improve even among individual providers who did not directly participate in EWRs, due to a spill-over effect from the providers who participated in EWRs to those who did not. The existing model for EWRs has not tried to target every provider in a clinical area. We hypothesized that we would not have to expose all providers in a clinical unit to EWRs in order to improve the overall unit climate.
This was the first detailed study of EWRs so we had several secondary (but a-priori) exploratory hypotheses: 1) that the effect of EWRs would be stronger for providers who participated in EWRs; 2) that the effect of EWRs would vary by provider type; and 3) that some individual items on the Safety Climate Survey would be more responsive to EWRs than others. To test these hypotheses, we randomized units in a hospital to receive EWRs or usual care (the control group). The Safety Climate Survey was administered prior to, and after EWRs.
Participants and setting
The study was conducted at Memorial Hermann Hospital, a 711 bed urban tertiary care teaching hospital that contained adult and children's hospitals. We grouped into "units" the inpatient clinical areas (such as individual ICUs or wards) that cared for similar types of patients or provided similar services. The departments of radiology, pharmacy, respiratory therapy, and physical therapy/occupational therapy were each considered a unit and not grouped with clinical areas. The Emergency Department was not included. After this grouping process there were 23 units (the clinical care areas plus respiratory therapy, radiology, pharmacy, physical therapy/occupational therapy) eligible for randomization. A random number generator was used to allocate units to receive the executive walk round intervention or to be in the control group.
Intervention
Six executives (2 Vice Presidents and 4 Assistant Vice Presidents) participated in the intervention. Our rationale for choosing executives was to have executives visit units that they also supervised. There was no literature to inform the frequency of visits so the executives and investigators decided that the executive should visit their units during the daytime approximately once every four weeks for a total of 3 visits. Each executive was accompanied by the hospital's Patient Safety Officer and a staff member from the Performance Improvement Department. EWRs were scheduled in advance with managers for each unit. The executive met with providers either in a common work area on the unit such as the nurses station, or in a conference room. All providers present at that time were encouraged to attend (some providers had to continue patient care), but attendance was not mandatory. The executives had met with the Patient Safety Officer before the intervention began to review patient safety concepts, to educate the executives about EWRs, and to review the questions to ask during EWRs. These questions (Table 1) were generated from the literature [1] and the hospital's patient safety committee. Before asking the questions the executives stated that patient safety was a priority for the hospital, that the purpose of their visit was to foster a culture that encourages open communication and identifies ways to improve systems, that all comments would be kept confidential, and that no individual would be held accountable for system flaws, for revealing errors, or voicing their concerns. We did not enforce one standard script for all executives. Walk round sessions lasted 30–60 minutes. At the close of each session the providers were asked to tell two other providers in their unit who did not attend EWRs about the session. This was done to try and magnify the effect of the intervention and improve attitudes even among those who did not directly participate in an EWR.
Table 1 Examples of questions asked by executives during walk rounds.
Have there been any "near misses" that almost caused patient harm but didn't?
Have we harmed any patients recently?
What aspects of the environment are likely to lead to the next patient harm?
Is there anything we could do to prevent the next adverse event?
Can you think of any events in the past days that have resulted in prolonged hospitalization for a patient?
Can you think of a way in which the system or your environment fails you on a consistent basis?
What specific intervention from leadership would make the work you do safer for patients?
What would make this executive walk rounds more effective?
Outcome measure
The primary outcome measure was safety climate, measured using the Safety Climate Survey. The Safety Climate Survey is derived from similar surveys in commercial aviation [6] that measured safety-related attitudes of cockpit crew members. We adapted the aviation survey for healthcare by incorporating existing concepts from healthcare (especially research on safety [7] and organizational culture [8]), conducting focus groups with health care providers, consulting subject matter experts, and field testing items. The first healthcare version was for intensive care providers [9-11]. Our factor analysis of that survey identified a 7 item scale that we called the Safety Climate Scale. We then added items to the questionnaire which have been linked to safety and performance outcomes in prior aviation research [12,13], as well items that were identified through discussions with hospital executives, quality experts, and other end-users. This resulted in the 21 item Safety Climate Survey that uses a 5-point likert scale where 1 = disagree strongly and 5 = agree strongly. Excluding this study, over 8,000 healthcare providers in 251 clinical areas in 52 hospitals have completed the survey and the survey has been endorsed by the Institute for Healthcare Improvement.
We administered the baseline Safety Climate Survey between September 1st and October 15, 2002. EWRs occurred between late October 2002 and January 31, 2003. The survey was re-administered during March and April 2003. We surveyed Registered Nurses, Licensed Vocational Nurses, Nurse Managers, Pharmacists, technicians, Respiratory Therapists, Physical Therapists, Occupational Therapists, Speech Therapists, and Dieticians. Float pool staff were excluded although some non-float pool providers also worked in more than one unit. We did not survey physicians because many of them did not spend enough time on a specific unit to notice changes in safety climate over a short period of time, and their schedules made it likely that they would not be exposed to the intervention. Human Resources generated lists of eligible providers and unit managers reviewed the lists for accuracy. Managers distributed the baseline surveys at meetings or by placing them in mail boxes. To improve response rates they redistributed surveys every two weeks during each administration period. The post-EWRs surveys were administered as described for the baseline surveys and some providers completed the survey as part of previously scheduled in-service training. We surveyed day and evening shifts (even though walk rounds only occurred during the day) to measure if there was a spill-over effect that could influence the overall unit safety climate.
The study was approved by the University of Texas Health Science Center at Houston Committee for the Protection of Human Subjects.
Statistical analysis
We could find no research that used EWRs as an intervention to help us calculate sample size a priori so we used the entire hospital (except the emergency department) divided into 23 units as described above. We transformed the five point response scale on the safety climate survey to a 100 point scale and calculated means. The 100 point scale is better understood by hospital administrators and providers. We also calculated the percent positive safety climate score: the percent of respondents in a clinical unit who responded 4 or 5 (agree slightly or agree strongly) on the five point scale.
Our primary hypothesis was that units randomized to receive EWRs would have higher mean safety climate scores than control units. A secondary, a-priori, hypothesis was that some individual items on the SCS would be more responsive to EWRs than others. The five items were: 1) This institution is doing more for patient safety now, than it did one year ago (item 15); 2) The senior leaders in my hospital listen to me and care about my concerns (item 3); 3) Patient safety is constantly reinforced as the priority in this clinical area (item 19); 4) Leadership is driving us to be a safety-centered institution (item 5); and 5) I would feel safe being treated here as a patient (item 11).
Another a-priori hypothesis was that the effect of EWRs would vary by provider type and that the effect of EWRs would be stronger for providers who participated in EWRs. As part of the follow-up survey we asked providers if they recalled participating in an EWR session (executive names were listed by the question). Response options included yes, no, and not sure. This resulted in three groups in the intervention group (EWRs-participant, EWRs-not sure if participant, and EWRs-not a participant) and three groups in the control group (control EWRs-participant, control EWRs-not sure if participant, and control EWRs-not a participant). Some providers in the control units received EWRs because they temporarily worked in an intervention unit on the day EWRs occurred. We hypothesized that mean safety climate scores would be higher for providers in the intervention EWRs-participant group than in the control EWRs-not a participant group.
Mean safety climate scores
Differences in mean SCS scores were tested using random intercept linear models to account for the non-independence of research participants nested within clinical units. These models were fit using SAS PROC MIXED. Model assumptions were verified by examining univariate histograms and residual skewness and kurtosis values, and by generation of residual-by-predicted value plots.
Analyses of group differences on individual items
Individual survey item responses consist of the ordered categorical values that represent the scale response labels of "disagree strongly", "disagree slightly", "neutral", "agree slightly", and "agree strongly". Intervals between these responses need not be equal. For this reason, we employed a cumulative odds logistic regression model to compare groups' odds of agreement on each survey question. The performance of the cumulative odds model benefits from ten or more responses per survey category [14]. Response categories were collapsed as needed to attain this objective (described in the note following Table 4). To obtain proper standard errors, confidence intervals, and p-values adjusted for clustering of respondents within clinical areas, we employed the SURVEYLOGISTIC procedure in SAS 9.1.3 to fit the cumulative odds models.
Table 4 Effect of executive walk rounds on nurses responses to survey items: Distributions of responses by survey item and EWRs participation.
Survey Item EWRs Participant n (%) EWRs Non-participant n (%)
Disagree Strongly
Disagree Slightly
Neutral
Agree Slightly
Agree Strongly
Disagree Strongly
Disagree Slightly
Neutral
Agree Slightly
Agree Strongly
1. The culture of this clinical area makes it easy to learn from the mistakes of others. 0 2 (2.44) 15 (18.29) 22 (26.83) 43 (52.44) 6 (3.13) 21 (10.94) 50 (26.04) 53 (27.60) 62 (32.29)
2. Medical errors are handled appropriately in this clinical area. 3 (3.61) 2 (2.41) 12 (14.46) 12 (14.46) 54 (65.06) 3 (1.54) 17 (8.72) 38 (19.49) 48 (24.62) 89 (45.64)
3. The senior leaders in my hospital listen to me and care about my concerns. 3 (3.57) 6 (7.14) 15 (17.86) 26 (30.95) 34 (40.48) 29 (14.65) 23 (11.62) 40 (20.20) 54 (27.27) 52 (26.26)
4. The physician and nurse leaders in my area listen to me and care about my concerns. 0 6 (7.14) 11 (13.10) 27 (32.14) 40 (47.62) 10 (5.15) 20 (10.31) 40 (20.62) 56 (28.87) 68 (35.05)
5. Leadership is driving us to be a safety-centered institution. 1 (1.19) 0 15 (17.86) 21 (25.00) 47 (55.95) 4 (2.03) 18 (9.14) 48 (24.37) 59 (29.95) 68 (34.52)
6. My suggestions about safety would be acted upon if I expressed them to management. 1 (1.19) 4 (4.76) 14 (16.67) 25 (29.76) 40 (47.62) 11 (5.58) 26 (13.20) 39 (19.80) 50 (25.38) 71 (36.04)
7. Management/Leadership does not knowingly compromise safety concerns for productivity. 3 (3.61) 5 (6.02) 14 (16.87) 24 (28.92) 37 (44.58) 17 (8.76) 21 (10.82) 33 (17.01) 53 (27.32) 70 (36.08)
8. I am encouraged by my colleagues to report any patient safety concerns I may have. 2 (2.41) 3 (3.61) 4 (4.82) 20 (24.10) 54 (65.06) 1 (0.51) 13 (6.63) 25 (12.76) 55 (28.06) 102 (52.04)
9. I know the proper channels to direct questions regarding patient safety. 0 1 (1.23) 4 (4.94) 19 (23.46) 57 (70.37) 0 2 (1.03) 17 (8.72) 60 (30.77) 116 (59.49)
10. I receive appropriate feedback about my performance. 2 (2.38) 4 (4.76) 10 (11.90) 26 (30.95) 42 (50.00) 6 (3.06) 24 (12.24) 38 (19.39) 59 (30.10) 69 (35.20)
11. I would feel safe being treated here as a patient. 0 2 (2.53) 13 (16.46) 21 (26.58) 43 (54.43) 5 (2.56) 20 (10.26) 35 (17.95) 62 (31.79) 73 (37.44)
12. Briefing personnel before the start of a shift (i.e., to plan for possible contingencies) is an important part of patient safety. 0 0 5 (6.10) 19 (23.17) 58 (70.73) 2 (1.04) 4 (2.07) 19 (9.84) 34 (17.62) 134 (69.43)
13. Briefings are common here. 1 (1.30) 3 (3.90) 15 (19.48) 24 (31.17) 34 (44.16) 13 (7.30) 13 (7.30) 40 (22.47) 45 (25.28) 67 (37.64)
14a. I am satisfied with availability of clinical leadership (physician). 1 (1.25) 5 (6.25) 9 (11.25) 21 (26.25) 44 (55.00) 9 (4.62) 26 (13.33) 32 (16.41) 55 (28.21) 73 (37.44)
14b. I am satisfied with availability of clinical leadership (nursing). 2 (2.44) 3 (3.66) 8 (9.76) 31 (37.80) 38 (46.34) 12 (6.25) 13 (6.77) 33 (17.19) 60 (31.25) 74 (38.54)
14c. I am satisfied with availability of clinical leadership (pharmacy). 5 (6.33) 7 (8.86) 18 (22.78) 21 (26.58) 28 (35.44) 20 (10.53) 32 (16.84) 42 (22.11) 51 (26.84) 45 (23.68)
15. This institution is doing more for patient safety now than it did one year ago. 0 1 (1.32) 11 (14.47) 22 (28.95) 42 (55.26) 11 (6.11) 6 (3.33) 59 (32.78) 59 (32.78) 45 (25.00)
16. I believe that most adverse events occur as a result of multiple system failures, and are not attributable to one individual's actions. 0 4 (5.13) 8 (10.26) 24 (30.77) 42 (53.85) 8 (4.17) 16 (8.33) 23 (11.98) 65 (33.85) 80 (41.67)
17. The personnel in this clinical area take responsibility for patient safety. 0 0 8 (9.76) 23 (28.05) 51 (62.20) 2 (1.04) 9 (4.66) 26 (13.47) 75 (38.86) 81 (41.97)
18. Personnel frequently disregard rules or guidelines that are established for this clinical area. 30 (37.50) 10 (12.50) 14 (17.50) 12 (15.00) 14 (17.50) 47 (23.98) 43 (21.94) 39 (19.90) 40 (20.41) 27 (13.78)
19. Patient safety is constantly reinforced as the priority in this clinical area. 0 1 (1.27) 9 (11.39) 16 (20.25) 53 (67.09) 4 (2.02) 14 (7.07) 37 (18.69) 59 (29.80) 84 (42.42)
Results
A total of 1119 providers (67%) completed the baseline survey and 1,000 (55%) the post EWRs survey. There was no difference in safety climate between the EWR and control units when all providers were analyzed. We only report results for nurses (Licensed Vocational Nurses, Registered Nurses, and Nurse Managers) because we could not detect an effect of EWRs on safety climate scores for other providers. We randomized 23 units, 12 to the control group and 11 to the EWR group. Prior to EWRs, 547 nurses returned surveys; after EWRs, 598 nurses returned surveys. The types of nurses whose attitudes were measured post-EWR, their ages, and experience in the organization are shown in Table 2.
Table 2 Demographic characteristics of the post walk rounds nurse respondents.
Walk rounds n (%) Control n (%)
Total 260 338
Type of Nurse
LVN 30 (11.5) 30 (8.9)
RN 207 (79.6) 291 (86.1)
Nurse Manager 23 (8.8) 17 (5.0)
Age (years)
Less than 30 86 (33.1) 87 (25.7)
30–34 39 (15.0) 45 (13.3)
35–39 34 (13.1) 42 (12.4)
40–44 42 (16.1) 68 (20.8)
45 and over 54 (20.8) 89 (26.3)
Missing 5 (1.9) 7 (2.1)
Years in this hospital
Less than 1 52 (20.0) 44 (13.0)
1–2 41 (15.8) 46 (13.6)
3–7 61 (23.5) 96 (28.4)
8–12 49 (18.8) 90 (26.6)
13–20 24 (9.2) 34 (10.1)
21 and over 17 (6.5) 10 (3.0)
Missing 16 (6.2) 18 (5.3)
Providers and executives discussed specific issues during the Walk rounds and 12 themes were identified (Table 3). The hospital addressed 8 of the 12 themes prior to the follow up administration of the SCS. Some actions to address these themes were limited to units from which the theme arose (n = 4), other actions affected all units (n = 5). The follow up survey was administered before the hospital addressed: 1) difficulties in transitioning patients from Emergency Department to ICUs; 2) problems with TPN orders in neonatal intensive care; 3) Inconsistent application of the falls prevention program; and 4) difficulties caring for medical patients with significant psychiatric problems. Providers were notified of these changes through staff meetings.
Table 3 Themes identified during walk rounds.*
1. Medication ordering policy not followed (handwriting illegible, cannot identify ordering MD, etc.)
2. The medicaton administration record is not always reconciled with the most recent orders
3. Active interventions not maintained on the electronic medical record
4. Need to improve discharge education for patients on anticoagulants
5. House officers need better supervision when conducting procedures and nurses need a way to identify house officer training level and which procedures are appropriate for that level.
6. Difficulties in caring for medical patients with significant psychiatric problems
7. Management of overweight patients (inadequate equipment, difficulty turning, transporting)
8. Problems with TPN orders in neonatal intensive care
9. Inconsistent application of the falls prevention program
10. Difficulties in transitioning patients from Emergency Department to intensive care units (timing of transfer, use of different intravenous drug concentrations)
11. Improper use of oxygen tanks when patients transported
12. Beds not well maintained (wheel locks malfunction, frayed electrical cords)
*The hospital had not responded to items 6, 8, 9, and 10 prior to the follow-up safety climate survey.
Mean safety climate score comparisons
All 547 nurses measured pre-EWR returned useable safety climate data; by contrast, 570 participating nurses returned useable safety climate data following EWRs. Before EWRs the mean safety climate scores for nurses were similar in the control units and EWR units (78.97 and 76.78, P = 0.458) as were percent positive scores (64.6% positive and 61.1% positive). After EWRs the mean safety climate scores and percent positive scores were not significantly different in the control units and EWR units (77.93 and 78.33, P = 0.854) and (56.5% positive and 62.7% positive). However, nurses in the control group who did not participate in EWRs (n = 198) had lower safety climate scores than nurses in the intervention group who did participate in an EWR session (n = 85) (74.88 versus 81.01, P = 0.02; 52.5% positive versus 72.9% positive).
Comparisons of individual safety climate scale items
The distributions of responses by survey item and by EWRs participation are shown in Table 4. Compared to nurses who did not participate, nurses in the Experimental group who reported participating in EWRs responded more favorably on the items hypothesized a priori to be most sensitive to EWRs (Table 5). All five items showed statistically significant differences in odds of agreement with items for the EWR-participant compared to Control-Non-participant groups: This institution is doing more for patient safety now, than it did one year ago (item 15, OR = 3.82, p < .001); 2) The senior leaders in my hospital listen to me and care about my concerns (item 3, OR = 2.15, p = .012); 3) Patient safety is constantly reinforced as the priority in this clinical area (item 19, OR = 2.79, p = .001); 4) Leadership is driving us to be a safety-centered institution (item 5, OR = 2.48, p = .002); and 5) I would feel safe being treated here as a patient (item 11, OR = 2.05, p = .002). Examination of the odds ratios listed in Table 5 shows that nurses in the EWRs-participant group exhibited more favorable evaluations of safety climate through their responses to the individual safety climate items than did nurses in the control EWRs-not a participant group on 14 out of 21 items (Table 5; note that the items are labeled 1–19, but item 14 has three parts.
Table 5 Effect of executive walk rounds on survey items for nurses: Odds of agreement with an item for EWR Participants compared to EWR non-participants.
Survey Item OR 95% CI
1. The culture of this clinical area makes it easy to learn from the mistakes of others. 2.50 1.15 – 5.42
2. Medical errors are handled appropriately in this clinical area. 2.05 0.97 – 4.35
*3. The senior leaders in my hospital listen to me and care about my concerns. 2.15 1.18 – 3.92
4. The physician and nurse leaders in my area listen to me and care about my concerns. 1.89 1.13 – 3.16
*5. Leadership is driving us to be a safety-centered institution. 2.48 1.39 – 4.45
6. My suggestions about safety would be acted upon if I expressed them to management. 1.89 1.04 – 3.43
7. Management/Leadership does not knowingly compromise safety concerns for productivity. 1.56 0.90 – 2.70
8. I am encouraged by my colleagues to report any patient safety concerns I may have. 1.74 1.01 – 2.75
9. I know the proper channels to direct questions regarding patient safety. 1.62 0.87 – 3.03
10. I receive appropriate feedback about my performance. 1.98 1.23 – 3.20
*11. I would feel safe being treated here as a patient. 2.05 1.31 – 3.19
12. Briefing personnel before the start of a shift (i.e., to plan for possible contingencies) is an important part of patient safety. 1.16 0.65 – 2.07
13. Briefings are common here. 1.56 0.89 – 2.73
14a. I am satisfied with availability of clinical leadership (physician). 2.14 1.35 – 3.42
14b. I am satisfied with availability of clinical leadership (nursing). 1.62 0.90 – 2.93
14c. I am satisfied with availability of clinical leadership (pharmacy). 1.75 1.13 – 2.72
*15. This institution is doing more for patient safety now than it did one year ago. 3.82 1.87 – 7.81
16. I believe that most adverse events occur as a result of multiple system failures, and are not attributable to one individual's actions. 1.70 1.17 – 2.48
17. The personnel in this clinical area take responsibility for patient safety. 2.29 1.26 – 4.17
18. Personnel frequently disregard rules or guidelines that are established for this clinical area. 0.78 0.48 – 1.28
*19. Patient safety is constantly reinforced as the priority in this clinical area. 2.79 1.50 – 5.21
Notes:
1. N = 274 non-missing cases.
2. Odds ratios and 95% confidence intervals are derived via a five category cumulative odds logistic regression model fit using the SURVEYLOGISTIC procedure in SAS version 9.1.3. Confidence intervals are adjusted for clustering of participants within clinical areas. Each analysis is based on five ordered categories of response options: "strongly disagree", "disagree", "neither agree nor disagree", "agree", "strongly agree", with the following exceptions: Some items (1, 2, 5, 8, 10, 11, 15, 16, 17, 19) had fewer than ten respondents who endorsed "strongly disagree" or "disagree"; for these items the "strongly disagree" and "disagree" categories were pooled to yield n-per-category greater than or equal to ten cases. Similarly, items 9 and 12 required pooling of the "strongly disagree", "disagree", and "neither agree nor disagree" categories to yield ten or more cases per category.
3. The odds ratios indicate the odds of an EWR participant having more agreement with an item than a EWR non-participant
4. Item 18 is reversed scored.
Discussion
To our knowledge, this is the first published study to show that EWRs can improve safety climate among some providers in hospitals. Nurses in the intervention group who participated in EWRs had higher safety climate scores than non-participants in the control group. However, EWRs conducted in this manner did not result in higher safety climate scores for all nurses in the intervention units compared to the control units, suggesting a limited spill-over effect from nurses who participated in EWRs to those who did not participate. Furthermore, we did not detect an effect of EWRs on the attitudes of other provider types. This may be due to lack of power to detect a difference (there were relatively small numbers of some other provider types) or because EWRs or the safety climate survey may be less relevant to non-nursing providers.
Implications
Safety culture has been identified as a key safety practice for healthcare by the National Quality Forum and other experts because poor safety culture may contribute to unsafe practices. We found that EWRs can positively influence safety climate (one component of safety culture) among nurses who participated in EWRs. Our study should be encouraging and informative to hospitals and researchers who are experimenting with the use of EWRs.
The effect of EWRs on safety climate of participating nurses was detected after a short intervention (3 months) and after relatively minor changes in care processes. Future research should address the "dose-response" relationship between EWRs and safety climate. EWRs may need to be conducted more frequently, address a broader range of topics, address a larger audience during rounds, or occur for a longer duration than in our study (once a month for three months) to influence other providers and to have an impact on overall unit safety climate. A key goal should be to expose as many providers as possible (both day and night shifts) to the EWRs because the effect on attitudes may not diffuse throughout a clinical unit (although we had limited power to detect this effect). The effect of EWRs may also be modified by the actions taken to correct or address safety problems raised by providers during the EWRs. The actions taken after this set of EWRs were relatively minor in scope suggesting that the EWRs themselves may be important mediators of safety attitudes. More visible or more effective actions could result in greater improvements in safety climate, both among providers who participated in EWRs and those who did not.
Limitations
Our ability to detect differences between the control and EWR units was limited by sample size at the clinical unit level (23 units in total) and because some nurses in control units participated in EWRs because they were floating in another unit while a EWR occurred. Sample sizes of non-nursing groups also limited our ability to detect changes after EWRs.
Providers were not randomized so our provider level analysis that found a positive effect of EWRs may be biased. There may be factors that influenced participation in EWRs that also caused or were associated with more positive attitudes. The main determinant of participation was whether or not the nurse worked days or nights (EWRs were only done during the day). Day-shift nurses may respond differently than night-shift nurses to EWRs. Other reasons for non-participation for day-shift nurses were that EWRs occurred on days that they did not work, the nurse could have been too busy with patient care, or they could have chosen not to participate. The EWR schedule was determined by the executives, not nurse managers.
Generalizability is limited because we report findings from only one urban tertiary care hospital and because the EWR effect may be mediated by the individual executives of this hospital. Executives were advised about the purpose of EWRs and given questions to help guide their discussion (Table 1) but there was variability in the details of each EWR. This makes our findings more generalizable in that it mimics what most hospitals will do (use multiple executives), but also less generalizable in that others cannot exactly replicate our intervention. Future research should investigate executive effects to better establish what executives do well during EWRs, what level of executive is most effective, and whether a particular background (e.g., clinical/nonclinical) makes a difference. Randomized trials usually assume that the intervention (like a drug) is stable over time, and that the effect of the intervention is independent of other factors in the environment (although there are notable exceptions [15]). EWRs is a social intervention that varies depending upon the individual(s) conducting EWRs and the culture of the units receiving the intervention. Therefore, a formative evaluation that reports details of the interactions among executive characteristics (leadership style, gender, position in organization) and the units (size, recent safety issues, work routines) would be informative, as would a unit level analysis of safety climate scores. Future studies should also consider comparing EWRs to visits by executives who simply introduce themselves and ask how things are going. This may better isolate the effect of EWRs over and above the effect due to an executive appearing on the unit to talk.
Due to practical considerations for this particular hospital, we did not include physicians. Based upon our experience with other EWR efforts, inclusion of physicians in EWRs may lead to greater improvements in the unit-level safety climate because physicians may be positively influenced by the EWRs, and because they may be leaders in units so a positive attitude change could affect others on the unit.
Many readers will wonder if improved safety climate leads to reductions in errors and adverse events. We did not collect such data but experience at other institutions suggests that as error rates and lengths of stay decline because of interventions, safety climate scores increase [16]. Others may see safety climate attitudes as important outcomes regardless of their relationships with errors and adverse events, and research in aviation suggests a relationship among attitudes and performance [12,13].
Conclusion
EWRs have a positive effect on the safety climate attitudes of nurses who participate in the walk rounds sessions. EWRs may need to be performed more frequently or for a longer period of time than in this study in order to have a broader influence on provider attitudes. Future research should also look in detail at the interactions among executive and unit characteristics to better understand this safety intervention. By embracing the knowledge of front line providers, engaging them directly in improvement efforts, and aligning the concerns of providers and leaders [17,18], EWRs may improve safety climate and the broader construct of safety culture.
Competing interests
The author(s) declare that they have no competing interests.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors thank William Tudor, Patient Safety Officer at Memorial Hermann Hospital, and the executives and staff who participated in the study.
The study was funded by the Robert Wood Johnson Foundation and the Agency for Healthcare Research and Quality (grant # 1PO1HS1154401). The funding agencies had no role in determining the design, conduct, analysis, interpretation, or presentation of data.
==== Refs
Frankel A Graydon-Baker E Neppl C Simmonds T Gustafson M Gandhi TK Patient Safety Leadership Walk Rounds Jt Comm J Qual Saf 2003 29 16 26 12528570
Pronovost PJ Weast B Bishop K Paine L Griffith R Rosenstein BJ Kidwell RP Haller KB Davis R Senior executive adopt-a-work unit: a model for safety improvement Jt Comm J Qual Saf 2004 30 59 68 14986336
Nieva VF Sorra J Safety culture assessment: a tool for improving patient safety in healthcare organizations Qual Saf Health Care 2003 12 17 23
Safe Practices for Better Health Care: A Consensus Report National Quality Forum 2003
Columbia Accident and Investigation Board NASA Washington DC 2003
Helmreich RL Merritt AC Sherman PJ Gregorich SE Wiener EL The Flight management Attitudes Questionnaire (FMAQ) NASA/UT/FAA Technical Report 93-4 1993 Austin, TX, The University of Texas
Vincent CA Taylor-Adams S Stanhope N Framework for analyzing risk and safety in clinical medicine BMJ 1998 316 1154 1157 9552960
Shortell SM Denise M Rouseau DM Gillies RR Devers KJ Simons TL Organizational assessment in intensive care units (ICUs): construct development, reliability, and validity of the ICU Nurse-Physician Questionnaire Med Care 1991 29 709 723 1875739
Sexton JB Thomas EJ Helmreich RL Error, stress, and teamwork in medicine and aviation: cross sectional surveys BMJ 2000 320 745 749 10720356 10.1136/bmj.320.7237.745
Sexton JB A Matter of life or death: Social psychological and organizational factors related to patient outcomes in the intensive care unit PhD thesis 2002 The University of Texas at Austin
Thomas EJ Sexton JB Helmreich RL Discrepant attitudes about teamwork among critical care nurses and physicians Crit Care Med 2003 31 956 959 12627011 10.1097/01.CCM.0000056183.89175.76
Sexton JB Klinect JR The link between safety attitudes and observed performance in flight operations Proceedings of the Eleventh International Symposium on Aviation Psychology 2001 Columbus, Ohio, Ohio State University
Helmreich RL Foushee HC Benson R Russini W Cockpit resource management: Exploring the attitude-performance linkage Aviat Space Environ Med 1986 57 1198 1200 3800822
Allison P Logistic Regression Using the SAS System 1999 Cary, NC: SAS Institute 141
Tinetti ME Baker DI McAvay G Claus EB Garrett P Gottschalk M Koch ML Trainor K Horwitz RI A multifactorial intervention to reduce the risk of falling among elderly people living in the community N Engl J Med 1994 331 821 827 8078528 10.1056/NEJM199409293311301
Pronovost P Weast B Rosenstein B Sexton JB Holzmueller CG Paine L Davis R Rubin HR Implementing and Validating a Comprehensive Unit-based Safety Program Journal of Patient Safety
Edmondson A Roberto MA Tucker A Children's Hospitals and Clinics Case #N9-302-050 Harvard Business School, Allston, Massachusetts October 31, 2001
Weingart SN Making medication safety a strategic organizational priority Jt Comm J Qual Improv 2000 26 341 8 10840666
| 15823204 | PMC1097728 | CC BY | 2021-01-04 16:31:49 | no | BMC Health Serv Res. 2005 Apr 11; 5:28 | utf-8 | BMC Health Serv Res | 2,005 | 10.1186/1472-6963-5-28 | oa_comm |
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-291582901310.1186/1472-6963-5-29Research ArticleHealth insurance, neighborhood income, and emergency department usage by Utah children 1996–1998 Suruda Anthony [email protected] Thomas J [email protected] Stacey [email protected] J Michael [email protected] Intermountain Injury Control Research Center. University of Utah 615 Arapeen Drive #202 Salt Lake City, Utah 84108 USA2 Department of Sociology University of Oklahoma 331 Kaufman Hall Norman, Oklahoma 73019 USA2005 13 4 2005 5 29 29 5 10 2004 13 4 2005 Copyright © 2005 Suruda et al; licensee BioMed Central Ltd.2005Suruda 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
It is estimated that approximately half of emergency department (ED) usage in the U.S. and other developed countries is for non-urgent conditions and that this usage is related to availability, social, and economic factors. We examined pediatric ED usage in a U.S. state with respect to income, health insurance status, types of medical conditions, and whether introduction of managed care affected utilization by Medicaid children.
Methods
Emergency department usage rates were calculated from 1996 through 1998 using Utah ED data for children with commercial health insurance, Medicaid, for uninsured children, and by income group estimating neighborhood household income from Zip code of residence. We analyzed usage following the July 1996 transition of Utah Medicaid to managed care.
Results
Children with Medicaid had approximately 50% greater ED utilization rates than children with commercial health insurance or uninsured children. The majority of usage for Medicaid and uninsured children was for non-traumatic conditions. Only 35% of total ED usage was for non-emergent or non-urgent conditions and this was related to both Medicaid and low household income. Children lacking health insurance were more likely to be discharged against medical advice (OR = 2.36, 95% C.I. 1.88–2.96). There was no reduction in Medicaid ED usage following the transition to managed care.
Conclusion
Usage of ED services is related to both health insurance status and income. Children lacking health insurance and Medicaid children have excessive usage for conditions which could be treated in a primary care setting. That managed care does not reduce Medicaid ED usage is consistent with findings of other studies.
==== Body
Background
The increase in utilization of emergency medical services in the U.S. and other developed countries in recent decades is related to availability of ambulatory medical services and to the provision of health insurance [1]. Acute injury [2], longevity [3], and health insurance coverage across an array of conditions [4] are related to the absolute and relative distribution of income in a society. In the United States it is estimated that from 40% to 60% of ED visits are for non-urgent conditions and that such usage is related to availability, social and economic factors, availability of other medical services, and consumer choice [5].
The Emergency Medical Treatment and Active Labor Act of 1986 mandates that hospital emergency departments provide care to anyone who needs it regardless of ability to pay. Children without health insurance have been reported to be from 1.2 to 3.7 times more likely to use ED services than children with health insurance [6,7]. A study of the 1994 National Hospital Ambulatory Medical Care Survey found a relationship between lack of health insurance and increased ED utilization by adolescents [8]. Whether the increase in ED utilization is due to lack of health insurance alone or is related to other factors is unclear. A study of 1997 U.S. data reported no relationship between health insurance status and pediatric ED utilization after controlling for covariates such as age, number of parents present in the household, ethnicity, and family income [9]. Reduction of ED utilization for non-emergent conditions has been reported following enrollment in managed care plans [10] but this has been found only in children covered by commercial health insurance and not those covered by Medicaid [11].
We analyzed whether ED usage and charges in Utah children were related to health insurance status and income. Over 90% of Utah children have some form of health insurance [12] and ED usage rates are approximately 25% less than the U.S. average [13]. Utah Medicaid changed to a state-wide managed care plan in July 1996 and we also examined whether this affected ED utilization by Medicaid children. In the Utah Medicaid program families chose from a panel of provider plans; those not choosing were assigned to a plan closest to them. No preauthorization was required for ED usage. Some of the plans had co-pay requirements for ED usage but federal regulations prohibited co-pays for Medicaid children.
Methods
This study was approved by the University of Utah Institutional Review Board. Emergency department discharge data for all 443,040 visits for Utah children age 0–17 years for 1996 through 1998 were obtained from the Department of Health. These data excluded children admitted as inpatients through the emergency department, approximately 9% of ED visits [13] because we were unable to extract these from hospital inpatient data. Emergency department records were linked to ambulance run records for 1996–1998 using probabilistic linkage as previously described [14]. We assigned Abbreviated Injury Scale (AIS) and Injury Severity Score (ISS) scores and related information for children with injury as the principal diagnosis from the International Classification of Diseases, 9th revision (ICD-9) diagnoses in hospital inpatient and emergency department records using ICDMAP-90 software from TriAnalytics, Bel Air, MD.
Health insurance status for ED visits was determined from information in emergency department records from the coding for primary payor. The various codes were assigned to the categories "Medicaid, Commercial, Managed, and Uninsured." The percentage of children lacking any health insurance for 1996 was obtained from the 1996 Utah Health Status Survey; the percentages for 1997 and 1998 were extrapolated from the 1996 survey and the 2000 survey [12,15]. The number of Medicaid eligible children for each year was obtained from the annual report 416 for Utah submitted to the Health Care Financing Administration. The number of children with commercial health insurance was considered to be the total number minus the number of children with Medicaid and those lacking health insurance, adjusted for the small number with other payors. Usage rates were calculated for the seven most common pediatric diagnosis-related groups (DRG) found in the data.
We purchased marketing software from Claritas Corporation (Ithica, N.Y.) which provided median household income 1996–98 based upon zip code. We assigned the values for each of these to the zip code of residence from the hospital emergency department records. We excluded hospital zip codes of residence which were outside Utah. We did not have population estimates for health insurance by zip code.
When presenting findings related to distribution of median household income we chose to examine by 'tertiles' (lowest third, middle third, and upper third) according to the income estimates for the emergency department residence data. Lower income neighborhoods were a mixture of rural and urban with a mean population density of 12,153 per square mile (median 689 per square mile). Middle income neighborhoods were urban with a mean population density of 18,566 per square mile (median 12,697 per square mile) and upper income neighborhoods were also urban with a mean population density of 24,273 per square mile (median 23,321 per square mile).
Bivariate analysis was done using the Chi Square test. Multivariate analysis was done using logistic regression. In part because the analysis data set contained more than 400,000 observations, all of the differences shown in the tables are statistically significant at the p < .05 level.
Results
The ED usage rate per 100 children was 20.49 in 1996, 22.30 in 1997, 21.04 in 1998, and an average of 21.28 for the three year period. When analyzed by payor the majority of the increased usage in 1997 was by uninsured and Medicaid children. Children with commercial health insurance had an ED usage rate per 100 of 19.93 in 1996, 20.25 in 1997, 19.79 in 1998, and an average of 19.99. Medicaid children had an ED usage rate per 100 of 28.06 in 1996, 33.74 in 1997, 29.30 in 1998, and an average of 30.35. Uninsured children had an ED usage rate per 100 of 16.07 in 1996, 23.49 in 1997, 21.73 in 1998, and an average of 20.49.
The increased ED usage in 1997 for all groups was largely attributable to respiratory conditions such as otitis media, upper respiratory infection, and asthma. Usage for these conditions compared to 1996 rose 10% for children with commercial insurance, 35% for those with Medicaid, and 26% for uninsured children.
The seven most common DRG categories accounted for 55% of pediatric emergency visits during this period (Table 1).
Table 1 Emergency Department Usage for Children age 0–17 for Specific Conditions, Utah,1996–1998
Commercial Insurance Medicaid Uninsured All Children
Soft tissue trauma 3.77 3.17 3.12 3.49
Fracture or sprain of hand, forearm, or foot 1.55 0.87 0.95 1.33
Fracture or sprain of upper arm or lower leg 1.14 0.72 0.78 1.00
Concussion, < 1 hour 0.32 0.32 0.23 0.30
Otitis media or upper respiratory infecton 2.05 7.50 3.45 3.08
Gastroenteritis 1.34 2.89 2.46 1.68
Asthma or bronchitis 0.65 1.75 0.78 0.84
All conditions 19.99 30.35 20.49 21.28
Rates are per 100 children per year
There were differences with respect to age by payor and children with Medicaid had a mean age of 4.4 years, compared to children with commercial health insurance (8 years) and uninsured children (6.7 years).
Almost all of the children were discharged home (99%), 0.5% were transferred to another acute care hospital, and remainder discharged elsewhere. Children without health insurance were more likely to be discharged against medical advice (OR = 2.36, 95% CI 1.88–2.96) although the number of such discharges for the three year period was small (580, 0.1%).
For children with trauma, injury severity scores were similar for children with managed care, commercial health insurance, and no health insurance (mean, 1.8) and were somewhat less for children with Medicaid (mean, 1.6, p < .05)). Children with Medicaid had the lowest mean hospital emergency department charges of the various groups. For the common DRGs mean emergency department charges ranged from $127 to $461 (Table 2), with the average charge for all groups being $251. Overall 3.1% of emergency department records were linked to ambulance records and indicated transport by ambulance for children who were subsequently discharged from the emergency department. The majority of transports for all categories was for trauma.
Table 2 Emergency Department Charges for Children age 0–17, Utah, 1996–1998, for Seven Leading Diagnosis-Related Groups, All Conditions, and Payor Status
Commercial Insurance Mean Charge Annual Visits Average Annual Charges
Soft tissue trauma $245 18,360 $4,491,888
Fracture or sprain of hand, forearm, or foot $338 7,538 $2,544,492
Fracture or sprain of upper arm or lower leg $319 5,560 $1,775,007
Concussion, < 1 hour $461 1,578 $726,963
Otitis media or upper respiratory infection $148 9,993 $1,482,615
Gastroenteritis $268 6,958 $1,864,700
Asthma or bronchitis $220 3,173 $698,971
All conditions $277 99,621 $27,093,422
Medicaid Mean Charge Annual Visits Average Annual Charges
Soft tissue trauma $208 3,934 $818,272
Fracture or sprain of hand, forearm, or foot $307 1,088 $334,016
Fracture or sprain of upper arm or lower leg $277 890 $246,530
Concussion, < 1 hour $329 384 $126,336
Otitis media or upper respiratory infection $127 9,314 $1.182,878
Gastroenteritis $187 3,710 $693,770
Asthma or bronchitis $207 2,175 $450,225
All conditions $195 36,073 $7,376,833
Uninsured children Mean Charge Annual Visits Average Annual Charges
Soft tissue trauma $224 1,910 $427,840
Fracture or sprain of hand, forearm, or foot $335 585 $195,975
Fracture or sprain of upper arm or lower leg $293 481 $140,933
Concussion, < 1 hour $416 139 $57,824
Otitis media or upper respiratory infection $124 2,100 $260,400
Gastroenteritis $209 975 $203,775
Asthma or bronchitis $210 484 $101,640
All conditions $220 11,986 $2,767,264
Estimated household annual incomes ranged from $11,250 to $94,807 based upon zip code of residence. For children who were treated in emergency departments, 22% resided in neighborhoods with the lowest tertile of household income ($11,250–$35,936), 35% resided in neighborhoods with the middle tertile household income ($36,098–49,946), and 43% resided in neighborhoods within the upper tertile of household income ($50,000 -94,807). There was an approximately two-fold variation in ED usage rates by estimated household income (Table 3).
Table 3 ED usage rate by estimated neighborhood household income
1996 1997 1998
Lower income 28.4 32.6 32.1
Middle income 20.2 21.8 21.0
Upper income 16.7 17.4 15.4
Rates are per 100 children per year
Visits for emergency department care varied among the three income groups with respect to payor status (Table 4), with the proportion of uninsured children greatest in low income neighborhoods.
Table 4 Insurance coverage of children using emergency departments by estimated neighborhood income, Utah, 1996–1998
Low income Middle income Upper income
Indemnity insurance 30.2% 37.8% 39.9%
Managed care 18.1% 31.7% 41.3%
Medicaid 40.7% 22.3% 12.7%
Uninsured 11.0% 8.2% 6.1%
The most frequent diagnoses differed by both income and health insurance status. In the lower income neighborhood, both children with Medicaid and uninsured children had otitis media as the most frequent DRG (27% and 19% respectively) while children residing in upper income neighborhoods had usage patterns that more closely resembled children with commercial health insurance than children with Medicaid. For example, the most frequent DRG for uninsured children in upper income neighborhoods was soft tissue trauma (16% of visits), as it was for children with commercial health insurance (19%).
Overall, 65% of ED visits were coded as emergent or urgent. Children whose payor was managed care were coded as "emergent" 58% of the time, followed by other types of commercial insurance and Medicaid (41% each) and uninsured children (36%).
In a multivariate model having an ED visit coded as other than emergent was related to both low income (OR = 1.30, 95% CI 1.27–1.33) and to having Medicaid (OR = 1.69, 95% CI 1.64–1.75) when age was included in the model. Using an ambulance for transport and having the ED visit coded as non-emergent was related to low income (OR = 2.14, 95% CI 1.85–2.49) but not to Medicaid (OR = 1.27, 95% CI 0.91–1.39)
Discussion
In this study we have several important findings concerning pediatric ED visits, specific medical conditions, health insurance status, managed care, and neighborhood household income.
Emergency department usage rates during 1996–98 were approximately 50% greater for children with Medicaid than for children with commercial health insurance or uninsured children. We found a substantial proportion of pediatric ED usage was for conditions such as otitis media, upper respiratory infection, and gastroenteritis that could be expected to be treated in a lower cost primary care setting rather than in an ED (Tables 1, 2). Uninsured children and children with Medicaid had higher ED visit rates for these conditions than children with commercial health insurance and were less likely than children with commercial health insurance to be coded as "emergent." The data did not contain the time of ED visit, which together with date information would have allowed us to estimate whether ED care was sought during normal business hours when a primary care provider might ordinarily be available. The average ED charge per visit for otitis media for uninsured children, $124 (Table 2), probably exceeded the charge for initial and follow-up care from a primary care provider for the same condition. Both health insurance status and estimated neighborhood income were predictive of ED usage for these conditions which might be better treated in a primary care setting. Pediatric usage of ED services for non-urgent conditions contributes to ED overcrowding [9].
Although this study relied on secondary sources and no individual ED records were audited or reviewed, the proportion of records which were coded as non-urgent, 35%, is somewhat less than previous estimates of ED usage for non-urgent conditions [1,5] and could be due to greater participation in managed care plans than in other studies and also to exclusion of children who were admitted to the hospital from the ED.
The population under study was relatively prosperous; even the low income group's median neighborhood income was $28,000. Only 5% of neighborhoods had median estimated household incomes which were below the poverty line for a family of four, $16,000 [12]. This is consistent with reported findings that even in high per capita income countries there is an independent effect of income distribution on the health of individuals [17].
In a multivariate model we found that use of ED services for non-emergent conditions was related to both Medicaid and low income. This appears to be in contrast to the findings of Lou et al. [9] who found no relationship to income but we were unable to analyze for the additional covariates such as number of parents in the household and ethnicity which are related to family income.
Limitations of this study are the use of secondary data sources, the lack of a denominator for subcategories of children who had commercial health insurance such as managed care, and the lack of independent record review to verify the coding of ED records as emergent or urgent. The data which we examined did not contain information concerning children admitted to the hospital through the ED. The percentage of ED records which we linked to ambulance run reports was low (3.1%). The denominator used for Medicaid children was the total number of eligible children during the year rather than the average number of eligible children, and so the actual usage rates for Medicaid may be underestimated. Our analysis of income relied upon estimation of neighborhood income from zip code of residence, rather than the actual family income of any child.
Strengths of this study include the use of population-based data for an entire state and the ability to calculate separate rates for insured, uninsured, and Medicaid children. The fortuitous change of Medicaid to managed care in July 1996 allowed analysis of ED visits for Medicaid children for two succeeding years. We did not have the means to assess health outcomes following ED visits. The findings that children lacking health insurance had increasing rates of ED visits from 1996 through 1998, and that these children were more likely than others to be discharged from the ED against medical advice, suggests that lack of health insurance will be associated with adverse health outcomes.
Conclusion
This study reports condition-specific data on ED usage which provide evidence that both economic status and health insurance status are related to pediatric emergency department usage.
ED usage for all health conditions was related to health insurance status and to estimated neighborhood income. Children in upper income neighborhoods who were uninsured had usage patterns similar to insured children, while uninsured children in low income neighborhoods had usage patterns similar to Medicaid children. Both the middle and upper income neighborhoods in this study were in urban areas where walk-in urgent care centers provided an alternative source of care during weekends and evening hours. This is consistent with our finding that, regardless of insurance status, children living in high income areas were less likely to have ED visits coded as other than "emergent or urgent" than children living in low income areas.
We found that the July 1996 transition of Medicaid to managed care had little effect on Medicaid usage in 1997 and 1998. There was actually a small increase in following the transition. This is consistent with findings in other states [11]. The transition to Medicaid had no disincentives for ED usage and so we would not expect to find an effect of managed care unless there was increased access to primary care providers for children. Whether ED usage for minor health conditions is due to lack of timely access to a provider, consumer habits, or lack of requirement for a financial co-payment [16] for ED use is unknown. The lack of effect is unfortunate in that the largest single category of expenditure for annual ED charges in Medicaid children in this study was $1,182,878 for otitis media and upper respiratory infection, conditions which should be less expensive to treat in a primary care setting.
Abbreviations
DRG Diagnosis-Related Groups
ED Emergency Department
OR Odds Ratio
CI Confidence Interval
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in the study design, preliminary data analysis, and writing the manuscript. Anthony Suruda was the principal investigator and completed the data analysis. Thomas J. Burns contributed the analysis by estimated income. Stacey Knight provided statistical support. J. Michael Dean facilitated acquisition and analysis of the data.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by grant 5 H34 MC00020 from the Health Resources and Services Administration.
==== Refs
Beland F Lemay A Boucher M Patterns of visits to hospital emergency rooms Soc Sci Med 1998 47 165 179 9720636 10.1016/S0277-9536(98)00029-X
Laflamme L Social Inequality in Injury Risks 1998 Stockholm: National Institute of Public Health
Pappas G Queen S Hadden W Fisher G The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986 NEJM 1993 329 103 109 8510686 10.1056/NEJM199307083290207
Burns TJ Batavia AI DeJong G The health insurance work disincentive for persons with disabilities Res Sociol Health Care 1994 11 57 68
Lee A Lau FL Hazlett C Kam C Wong P Chow S Factors associated with non-urgent utilization of accident and emergency services Soc Sci Med 2000 51 1075 1085 11005394 10.1016/S0277-9536(00)00039-3
McCormick MC Kass B Elixhauser A Thompson J Simpson L Annual report on access to and utilization of health care for children and youth in the United States-1999 Pediatrics 2000 105 219 230 10617727 10.1542/peds.105.6.1375
Johnson WG Rimsza ME The effects of access to pediatric care and insurance coverage on emergency department utilization Pediatrics 2004 113 483 487 14993538 10.1542/peds.113.3.483
Ziv A Boulet JR Slap GB Emergency department utilization by adolescents in the United States Pediatrics 1998 101 987 994 9606224 10.1542/peds.101.6.987
Luo X Liu G Frush K Hey LA Children's health insurance status and emergency department utilization in the United States Pediatrics 2003 112 314 319 12897280 10.1542/peds.112.2.314
Simpson L Fraser I Children and managed care: What research can, can't, and should tell us about impact Med Car Res Rev 1999 56 13 36 10327822
Moody-Williams J Linzer J Stern A Wilkinson J Athey J Twenty-four-hour access to emergency care for children in managed care Ann Emerg Med 1999 34 761 766 10577407
Utah Department of Health 1996 Utah Health Survey Status Report: Health Insurance Coverage 1997 Salt Lake City, Utah
Utah Department of Health Utah Emergency Department Utilization and Charges Profile 1998 Salt Lake City, Utah
Dean JM Vernon DD Cook L Nechodom P Reading J Suruda A Probabilistic linkage of computerized ambulance and inpatient discharge records: A potential tool for the evaluation of emergency medical services Ann Emerg Med 2001 37 616 626 11385330 10.1067/mem.2001.115214
Utah Department of Health 2000 Utah Child Health Survey 2001 Salt Lake City, Utah
Wong MD Andersen R Sherbourne CD Hays R Shapiro M 2001. Effects of cost sharing on care seeking and health status: Results from the medical outcomes study Am J Public Health 2001 91 1889 94 11684621
Gravelle H Wildman J Sutton M Income, income inequality and health: What can we learn from aggregate data? Soc Sci Med 2002 54 577 589 11848275 10.1016/S0277-9536(01)00053-3
| 15829013 | PMC1097729 | CC BY | 2021-01-04 16:31:50 | no | BMC Health Serv Res. 2005 Apr 13; 5:29 | utf-8 | BMC Health Serv Res | 2,005 | 10.1186/1472-6963-5-29 | oa_comm |
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-91583310710.1186/1471-2199-6-9Research ArticleComparing protocols for preparation of DNA-free total yeast RNA suitable for RT-PCR Del Aguila Eduardo M [email protected] Marcio B [email protected] Joab T [email protected] Vânia MF [email protected] Departamento de Bioquímica, Instituto de Química, Universidade Federal do Rio de Janeiro, Centro de Tecnologia, Bloco A – Lab. 545 – CEP 21949-900 – Rio de Janeiro (RJ) – Brasil2 Centro de Ciências da Saúde, Universidade Salgado de Oliveira, Rua Lambari, 10- CEP 24456-570- São Gonçalo (RJ) – Brasil2005 15 4 2005 6 9 9 20 10 2004 15 4 2005 Copyright © 2005 Del Aguila et al; licensee BioMed Central Ltd.2005Del Aguila 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
Preparation of RNA free from DNA is a critical step before performing RT-PCR assay. Total RNA isolated from several sources, including those obtained from Saccharomyces cerevisiae, using routine methodologies are frequently contaminated with DNA, which can give rise to amplification products that mimic the amplicons expected from the RNA target.
Results
We investigated the efficiency of two DNase I based protocols for eliminating DNA contaminations from RNA samples obtained from yeast cells. Both procedures are very efficient in eliminating DNA contamination from RNA samples and entail three main steps, which involve treating of RNA samples with DNase I, inhibition of the enzyme by EDTA and its subsequent inactivation at 65°C. The DNase I treated samples were further purified with phenol: chloroform followed by precipitation with ice-cold ethanol (protocol I) or, alternatively, they were directly used in RT-PCR reactions (protocol II). Transcripts from ACT1, PDA1, CNA1, CNA2, TPS1 and TPS2 analyzed after each treatment showed that all mRNAs tested can be amplified if total RNA was extracted and purified after DNase I treatment, however, only TPS1, TPS2 and ACT1 mRNAs were amplified without extraction/purification step.
Conclusion
Although more laborious and requiring a higher initial amount of material, the inclusion of an extraction and purification step allows to prepare RNA samples that are free from DNA and from low molecular contaminants and can be applied to amplify any Saccharomyces cerevisiae mRNA by RT-PCR.
==== Body
Background
The adaptation of polymerase chain reaction (PCR) methodology to the investigation of RNA provides the researcher a method featuring speed, efficiency, specificity and sensitivity. Since RNA cannot serve as a template for DNA polymerase, a reverse transcription step was combined with PCR to transform RNA into a suitable complementary DNA (cDNA), a technique that is referred to as RT-PCR (reverse transcriptase – polymerase chain reaction) [1]. RT-PCR has enabled important experiments dealing with gene expression and its regulation. More sensitive and less laborious than Northern blotting hybridization and RNase protection assays, it has rapidly become a common procedure [2].
However, a frequent cause of concern among investigators performing quantitative RT-PCR is inaccurate data acquisition due to DNA contamination in RNA preparations, because PCR cannot discriminate between cDNA targets synthesized by reverse transcription and genomic DNA contamination. DNA contamination in RNA preparations is easily detected by performing a non-reverse transcriptase control. Furthermore, PCR primers can be designed for controlling the genomic DNA contamination. Primers that span intron-exon boundaries amplify a product from contaminating DNA that includes the intron, making it larger than the expected cDNA product. Alternatively, primers can be designed to anneal at a splice junction avoiding any signal based on DNA contamination [3]. Unfortunately, the genome of low eukaryotic or prokaryotic cells has few intron containing genes, which makes the above strategies useless. We have been faced with the problem of DNA contamination in our samples of RNA after initial attempts to study mRNA level in Sacchromyces cerevisiae by RT-PCR. Common methods used to remove DNA from RNA samples include poly (A) mRNA purification by oligo (dT) chromatography [4], selective RNA precipitation with lithium chloride [5] and selective DNA extraction with acid phenol: chloroform [6]. Oligo (dT) chromatography is expensive and requires extensive manipulation whereas LiCl precipitation and acid phenol: chloroform extraction could not be effective, mainly in the amplification of rare transcripts when an increasing number of cycles or amount of template RNA has to be used. An alternative method employs treatment of RNA samples with DNase I followed by heat inactivation of the enzyme [7]. Optionally, heat denatured DNase is extracted and RNA precipitated in order to avoid the presence of some compounds that could interfere with RT-PCR assay. The present study describes the comparison of two protocols for preparing yeast RNA free from DNA suitable for RT-PCR analysis based on DNase I removal of genomic DNA.
Here, we analyze the use in RT-PCR assay of yeast RNA samples isolated from routine techniques and treated by DNase I, in order to remove the DNA contamination. An additional step of RNA extraction followed by precipitation with ethanol must be included in order to guarantee that the unfractionated RNA samples are suitable for the analysis of expression of any Saccharomyces cerevisiae gene.
We hope that the re-evaluation of the methods for preparation of samples for RT-PCR showed here will encourage the use of this up to now under appreciated methodology for analyzing the mRNA levels in Saccharomyces cerevisiae.
Results and discussion
Figure 1 shows the results of RT-PCR experiments using DNase-treated or untreated total RNA samples amplified using ACT1 (fig 1A) or CNA1 primers (fig. 1B). Untreated RNA samples produced the expected 407 bp (ACT1) or 629 bp (CNA1) amplicons, respectively, in the presence of active (fig. 1A and 1B, lanes 2) or heat inactivated reverse transcriptase (fig. 1A and 1B, lanes 1). Since RNA cannot serve as a template for DNA polymerase, these results make evident that DNA was present as a contaminant in the crude RNA preparations. Treatment of RNA samples with DNase I followed by RNA precipitation eliminated DNA contamination as judged by RT-minus control for ACT1 (fig. 1A, lane 3) and CNA1 reactions (fig 1B, lane 3). At the same time, it did not interfere with the production of ACT1 or CNA1 amplicons (fig. 1A and 1B, lanes 4) from their respective mRNA. Similar results were found for ACT1 reactions when DNase I treated samples were used without following precipitation (protocol II) (fig. 1A, lanes 5 and 6). On the other hand, CNA1 amplicons were not detected in RT-minus control neither in RT-PCR reactions using these same RNA-treated samples as template (fig. 1B, lanes 5 and 6). Based on these results, it seems that the use of treated RNA samples without the subsequent extraction/purification step produced an inhibitory effect on RT-PCR. To further address this question, we used a RNA sample treated according to protocol II as the source of templates to amplify six different mRNAs. ACT1, TPS1 and TPS2 transcripts generated the expected amplicons (fig. 2, lanes 3, 4 and 7), whereas CNA1, PDA1 and CNA2 failed to be amplified by RT-PCR (fig. 2, lanes 2, 5 and 6). In contrast, all of the six transcripts have been successfully amplified by RT-PCR if RNA templates were treated with DNase I followed by extraction/precipitation step with ethanol [8,9]. The compound that interfered on RT-PCR reaction could be EDTA, which is added to DNase I reaction in order to inactivate the enzyme. When carried over to the RT-PCR assay, EDTA could affect the free Mg2+ concentration in the reaction mixture, changing the efficiency of amplification in a primer-template dependent way. To test if the excess of free EDTA carried over from DNase I treatment, indeed affects the efficiency of RT-PCR assay, DNase I activity was stopped by the addition of increasing amounts of EDTA (ranging from 1.25 to 2.25 mM), before the heat inactivation of the enzyme. The 629 bp CNA1 amplicons were detected when the reactions were stopped with 1.5 to 2.0 mM EDTA (resulting in a molar ratio from 0.7 to 1.2 of EDTA molecule to Mg2+ ions (Fig. 3, lanes 2, 3 and 4). The addition of EDTA in concentrations out of this range abolished the amplification (Fig. 3, lanes 1 and 5).
To emphasize the fact that a simple DNase I treatment on crude RNA preparation is sufficient to perform accurate estimative of mRNA level by RT-PCR, we carried out the study of the expression of ACT1 mRNA along of the growth of an yeast strain in YPAD medium, using untreated and treated RNA samples obtained from cells harvested at different growth stages (Fig. 4, A and 4B). The amplification by RT-PCR using ACT1 primers and DNase I treated samples produced less intense bands (Fig 4B) at all growth phases when compared to those where untreated RNA was used (Fig. 4A). At stationary phase, using total RNA devoid of DNA contamination (Fig. 4B, lane 6) no expression of ACT1 genes was observed, a result consistent with those obtained using northern-blotting analysis [10], that showed a 100-fold decrease in ACT1 mRNA level at stationary phase.
In conclusion, total yeast RNA preparations can be made suitable for RT-PCR analysis of gene expression if they are free of contaminant DNA. DNase I treatment was effective in reducing to an undetectable level the DNA originally present in RNA samples. Different DNase brands can be used for this purpose. Here, DNase I preparations supplied form different manufacturers were used and both were very efficient in eliminating contaminant DNA. An additional step of extraction of reminiscent RNA with phenol: chloroform: isoamyl alcohol and precipitation with ethanol, as described in Protocol I, although more laborious, can be applied to amplify any Saccharomyces cerevisiae RNA using RT-PCR methodology. The main problem with this procedure is the need of starting with a higher amount of RNA to minimize loss of the sample during the precipitation step. Protocol II, on the other hand, is simple, practical and rapid because it avoids the extraction/precipitation steps and allows the direct use of treated RNA samples for RT-PCR assays. However, as magnesium concentration is a critical parameter in both RT and PCR reactions, the amount of EDTA used for inhibiting DNase I activity must be carefully titrated.
Conclusion
RT-PCR can be a method for determining transcript level in total unfractionated yeast RNA, since the RNA samples are free from contaminant DNA.
DNA contamination from RNA samples can be efficiently eliminated by treatment with commercial DNase I preparations.
However, although simple and efficient, that treatment introduces chelant cation in RNA samples which makes it, in some cases, not suitable to be utilized in a subsequent enzymatic analysis like RT-PCR, because the enzymes used in that methodology, reverse transcriptase and DNA polimerase are Mg2+-dependent.
To ensure that no false positive result, or on the contrary, failure on mRNA amplification occur, we established a general procedure for any S. cerevisiae mRNA amplification, where an additional step of RNA extraction followed by precipitation with ethanol was included.
The re-evaluation of these methods described here will permit, a large scale or repetitive gene expression evaluation as those commonly performed during yeast utilization in industry, can be performed by RT-PCR, without additional cost, since even unfractionated RNA can be suitable for this methodology.
Methods
Yeast strain, growth conditions and extraction of total RNA
Total yeast RNA was isolated by a modification of the procedure described previously [11]. Strain W303-1A (Mata, ade2-1, trp1-1, leu2,3-112, his3-11,15, ura3, can1-100) was grown in YPD-supplemented medium (1% yeast extract, 2% bacto peptone, 2% glucose and 0,01% of adenine, uracil, tryptophan, leucine and histidine) in a rotatory shaker (160 rpm and 28°C). Cell growth was monitored by reading OD at 570 nm. Cells (10 mg of dry weight) were harvested by centrifugation and washed with DEPC (diethylpyrocarbonate) treated water at 2200 × g for 5 min before they were resuspended in 0.6 ml RNA extraction buffer (10 mM EDTA (ethylenediaminetetracetic acid), 50 mM Tris-HCl pH 7.5, 0.1M NaCl, 5% SDS) and 0.6 ml of phenol:chloroform:isoamyl alcohol (50:50:1) mixture. After 6 min at room temperature, 2 g of glass beads (0.45 mm diameter) were added and the cells were broken by vigorous agitation for 2 min on a vortex mix set at maximum speed. The extract was transferred to a microfuge tube (1.5 ml) and cell debris and organic phase were separated from upper aqueous phase by centrifugation at 2200 × g for 5 min (24°C). The upper phase was collected, extracted twice with 1 volume of phenol: chloroform: isoamyl alcohol (50:50:1) and once with 1 volume of chloroform: isoamyl alcohol (24:1). The RNA was precipitated from the last upper aqueous phase by the addition of 0.1 volume of 3M NaOAc, pH 5.2, plus 3 volumes of ice-cold absolute ethanol followed by incubation at -20°C for 1 h. The RNA was pelleted by centrifugation at 15000 × g for 15 min (4°C), washed once with ice-cold 70% ethanol and again pelleted at 15000 × g for 15 min. After the remaining alcohol was allowed to evaporate, the pellet was resuspended in 30 μl of DEPC treated water. Concentration of RNA in the sample was measured by reading OD at 260 nm in a Beckman DU-6 spectrophotometer (1 OD = 42 μg RNA/ml). All materials and solutions were previously treated with DEPC [12].
Removal of contaminating genomic DNA from unfractionated RNA
RNA samples were treated with RNase-free bovine pancreatic DNase I (E.C. 3.1.21.1) to eliminate DNA contamination using two different protocols. Protocol I was a modification of the procedure previously described for C. Botulinum [13]. Six micrograms of RNA, 6.25 mM MgCl2 and 10U of RNase-free DNase I (Sigma) in a 10 μl reaction mixture in water were incubated at 37°C for 30 min. The reaction was stopped by the addition of 2 mM EDTA, pH 8.0, followed by incubation at 37°C for 1 min before inactivation at 65°C for 10 min. RNA was extracted with 1 volume of phenol: chloroform (5:1) and after centrifugation at 2200 × g for 5 min, the RNA present in the upper aqueous phase was precipitated with ice-cold absolute ethanol and collected by centrifugation at 15000 × g for 15 min (4°C). The remaining alcohol was allowed to evaporate and the pellet was resuspended in 30 μl of DEPC treated water. Protocol II was performed using the DNase I amplification grade kit (Life Technologies, Inc.) following the recommended procedure. In brief, 1 μg of RNA was resuspended in 20 mM Tris-HCl pH 8.4; 2.0 mM MgCl2; 50 mM KCl and 1U of DNase I into a final volume of 10 μl. After incubation for 15 min at room temperature, DNase was inactivated by the addition of 2,2 mM EDTA, pH 8.0, and subsequent incubation at 65°C for 10 min. The product of this reaction was used directly for RT-PCR without further treatment.
RT-PCR assay
mRNAs were amplified by using the kit Ready-to-Go RT-PCR Beads (Amershan Pharmacia Biotech Inc.) in a GeneAmp PCR System 2400 (Perkin Elmer Inc.). Each reaction contained ~2.0U of TaqDNA polimerase, 10 mM Tris-HCl pH 9.0, 60 mM KCl, 1.5 mM MgCl2, 200 μM of each dNTP, MMuLV reverse transcriptase, RNA guard ribonuclease inhibitor, stabilizer including RNase/DNase free BSA, 1 μg of total RNA and 20 pmol of the appropriate primers (Del Aguila et al, 2003; Souza, 2001) (Table 1) into a final volume of 50 μl. Synthesis of cDNA was performed at 50°C for 20 min and stopped by inactivation of reverse transcriptase at 95°C for 10 min. Amplification of cDNA by PCR was performed for 30 sec at 95°C; 30 sec at 65°C and 1 min at 72° C for 30 cycles, chosen after testing amplification from 20 to 40 cycles. RT minus controls was incubated at 95°C for 10 min to inactivate M-MuLV (Moloney murine leukemia virus) reverse transcriptase before cDNA synthesis and PCR steps. RT-PCR products were run on 1.3% agarose gel in 1x TAE (40 mM Tris-acetate and 1 mM EDTA) buffer, pH.8.0, and stained with aqueous ethidium bromide at a concentration of 0.5 μg/mL [12]. Data represent a typical result obtained from three different experiments.
Abbreviations
DEPC, diethylpyrocarbonate; EDTA, ethylenediaminetetraacetic acid; EtBr, ethidium bromide; M-MulV-RT, moloney murineleukemia virus reverse transcriptase; RT-PCR, reverse transcriptase – polimerase chain reaction
Authors' contributions
EM performed all the experimental procedures and was the primary of the manuscript. MD participated in the study design. JT participated in the study design, data analysis and drafted the manuscript. VP conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by CNPq, FAPERJ, FUJB and CAPES.
Figures and Tables
Figure 1 DNA contamination in total RNA isolated from Saccharomyces cerevisiae. Total RNA was isolated from Saccharomyces cerevisiae cells harvested at cellular density of 1.2 mg (dry weight/mL). PCR and RT-PCR amplicons were generated by 30 reactions cycles using ACT1 or CNA1 primers and 1.0 μg of RNA. Ten micro liters of each reaction was electrophoresed on a 1.3% agarose gel and stained with ethidium bromide. A – ACT1 amplicons. B – CNA1 amplicons. Untreated total RNA amplified by PCR (lane 1) or RT-PCR (lane 2); treated samples according to protocol I, followed by PCR (lane 3) or RT-PCR amplifications (lane 4); treated samples according to protocol II followed by PCR (lane 5) or RT-PCR amplification (lane 6).
Figure 2 Primer-template dependent inhibitory effect on mRNA amplification. RNA samples extracted from cells harvested at 2,1 mg (dry weight)/mL were treated with DNase I according to protocol II. Different set of primers (as described in table I) was used to amplify CNA2 (lane 2), ACT1 (lane 3), TPS2 (lane 4), PDA1 (lane 5), CNA1 (lane 6), TPS1 (lane 7) mRNAs. 100 bp DNA ladder (lane P) and RT-minus control using ACT1 primers (lane 1).
Figure 3 Effect of EDTA concentration on CNA1 mRNA amplification by RT-PCR. RNA samples extracted from cells harvested at 1.2 mg (dry weight/mL) were treated with DNase I according to protocol II. The reactions were stopped by the addition of 1.25 mM (lane 1), 1.5 mM (lane 2), 1.75 mM (lane 3); 2.0 mM (lane 4) and 2.25 mM (lane 5) of EDTA and treated samples were amplified by RT-PCR using CNA1 primers. P, 100 bp DNA Ladder (BioLabs-Inc).
Figure 4 Abundance of ACT1 transcripts during growth of Saccharomyces cerevisiae in glucose. RT-PCR amplicons were generated by 30 reactions cycles using ACT1 primers and 1.0 μg of RNA. Ten micro liters of each reaction were electrophoresed on a 1.3% agarose gel and stained with EtBr. Total RNA was isolated from Saccharomyces cerevisiae cells grown in YPD-supplemented medium and harvested at 0.7 mg (dry weight)/ml (lane 1); 1.2 mg (dry weight)/ml (lane 2); 2.1 mg (dry weight)/ml (lane 3); 2.6 mg (dry weight)/ml (lane 4); 3.2 mg (dry weight)/ml (lane 5) and 7.0 mg (dry weight)/ml (lane 6). P, 123 bp DNA Ladder (BioLabs -Inc). A – Untreated RNA; B – RNA treated with Danes I by protocol I
Table 1 Oligonucleotides used as primers in RT-PCR analysis
Target mRNA Primer Sequence Amplicon Size (bp)
TPS1 Forward CGCTAAGGCGCAACTGACCTCGTCT
Reverse CGACGAGAATGCGTGGTTGGCATAC 379
TPS2 Forward CCCCCCAAACTATCAGATTGGAACAAC
Reverse ACCCAGCTGCAGCTATTCCATCGGC 612
PDA1 Forward GGTCAGGAGGCCATTGCTGT
Reverse GACCAGCAATTGGATCGTTCTTGG 673
CNA1 Forward CGAAAGACTTGAATTCTTCACGCATC
Reverse GAATGATCTGCAGCAAGCATCG 629
CNA2 Forward CCTTATATCTGTTCCCGCCC
Reverse GAGGAACCATGGTTTTGGAG 551
ACT1 Forward CCTACGTTGGTGATGAAGCT
Reverse GTCAGTCAAATCTCTACCGG 407
CNA1, CNA2, PDA1 and ACT1 primer sequences were obtained from Del Aguila et al, 2003 and TPS1 and TPS2, from Souza, 2001.
==== Refs
Kawasaki ES Clark SS Coyne MY Smith SD Champlin R Whitte ON McCormick FP Diagnosis of Chronic Myeloid and Acute Lymphocytic Leukemias by Detection of Leukemia-specific mRNA Sequences Amplified in vitro Proc Nat Acad Sci U S A 1988 85 5698 5702
Gause WC Adamovicz J The Use of the PCR to Quantitative Gene Expression. Methods Appl PCR 1994 3 S123 S135
Gibson UE Heid CA Williams PM A Novel Method for Real Time Quantitative RT-PCR Genome Research 1996 6 995 1001 8908519
Aviv H Leder P Purification of Biologically Active Globulin Messenger RNA by Chromatography on Oligothymidylic Acid-cellulose Proc Natl Acad Sci U S A 1972 69 1408 1412 4504350
Cathala G Savouret J Mendez B West BL Karin M Martial JA Baxter JD A Method for Isolation of Intact, Translationally Active Ribonucleic Acid DNA 1983 2 329 335 6198133
Brawerman G Mendecki J Lee SY A Procedure for the Isolation of Mammalian Messenger Ribonucleic Acid Biochemistry 1972 11 637 641
Grillo M Margolis F Use of Reverse Transcriptase Polymerase Chain Reaction to Monitor Expression of Intron less Genes Biotechniques 1990 9 262 268 1699561
Del Aguila EM Silva JT Paschoalin VMF Expression of the yeast calcineurin subunits CNA1 and CNA2 during growth and hyper-osmotic stress FEMS Microbiology Letters 2003 221 197 202 12725927 10.1016/S0378-1097(03)00181-2
Souza RC Trehalose metabolism in msn2 and msn4 mutant strains from Saccharomyces cerevisiae, PhD thesis 2001 Universidade Federal do Rio de Janeiro, Rio de Janeiro
Wenzel TJ Teunissen AWRH Steensma H PDA1 mRNA: a standard for quantitation of mRNA in Saccharomyces cerevisiae superior to ACT1 mRNA Nucleic Acids Res 1995 23 883 884 7708509
Piper W Johnston JR Measurement of Transcription Molecular Genetics of Yeast – A Practical Approach 1997 Oxford University Press, New York 135 138
Sambrook J Russell DW Molecular Cloning. A Laboratory Manual 2001 3 Cold Spring Harbor, New York, USA
McGrath S Dooley JSG Haylock RW Quantification of Clostridium botulinum Toxin Gene Expression by Competitive Reverse Transcription-PCR Appl Environ Microbiol 2000 66 1423 1428 10742222 10.1128/AEM.66.4.1423-1428.2000
| 15833107 | PMC1097730 | CC BY | 2021-01-04 16:22:25 | no | BMC Mol Biol. 2005 Apr 15; 6:9 | utf-8 | BMC Mol Biol | 2,005 | 10.1186/1471-2199-6-9 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-171585422710.1186/1471-2350-6-17Study ProtocolThe Familial Intracranial Aneurysm (FIA) study protocol Broderick Joseph P [email protected] Laura R [email protected] Tatiana [email protected] John [email protected] Nathan [email protected] Irene [email protected] Robert D [email protected] Department of Neurology, University of Cincinnati, 231 Albert Sabin Way, Cincinnati, OH 45267-0525, USA2 Medical & Molecular Genetics, Indiana University, 975 West Walnut St., IB 130, Indianapolis, IN 46202-5251, USA3 Department of Radiology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA4 Division of Cerebrovascular Disease and Department of Neurology, Mayo Clinic, 200, First Street SW, Rochester, MN 55905, USA2005 26 4 2005 6 17 17 31 3 2005 26 4 2005 Copyright © 2005 Broderick 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
Subarachnoid hemorrhage (SAH) due to ruptured intracranial aneurysms (IAs) occurs in about 20,000 people per year in the U.S. annually and nearly half of the affected persons are dead within the first 30 days. Survivors of ruptured IAs are often left with substantial disability. Thus, primary prevention of aneurysm formation and rupture is of paramount importance. Prior studies indicate that genetic factors are important in the formation and rupture of IAs. The long-term goal of the Familial Intracranial Aneurysm (FIA) Study is to identify genes that underlie the development and rupture of intracranial aneurysms (IA).
Methods/Design
The FIA Study includes 26 clinical centers which have extensive experience in the clinical management and imaging of intracerebral aneurysms. 475 families with affected sib pairs or with multiple affected relatives will be enrolled through retrospective and prospective screening of potential subjects with an IA. After giving informed consent, the proband or their spokesperson invites other family members to participate. Each participant is interviewed using a standardized questionnaire which covers medical history, social history and demographic information. In addition blood is drawn from each participant for DNA isolation and immortalization of lymphocytes. High- risk family members without a previously diagnosed IA undergo magnetic resonance angiography (MRA) to identify asymptomatic unruptured aneurysms. A 10 cM genome screen will be performed to identify FIA susceptibility loci. Due to the significant mortality of affected individuals, novel approaches are employed to reconstruct the genotype of critical deceased individuals. These include the intensive recruitment of the spouse and children of deceased, affected individuals.
Discussion
A successful, adequately-powered genetic linkage study of IA is challenging given the very high, early mortality of ruptured IA. Design features in the FIA Study that address this challenge include recruitment at a large number of highly active clinical centers, comprehensive screening and recruitment techniques, non-invasive vascular imaging of high-risk subjects, genome reconstruction of dead affected individuals using marker data from closely related family members, and inclusion of environmental covariates in the statistical analysis.
==== Body
Background
Stroke is the third leading cause of death and the leading cause of disability among adults. Subarachnoid hemorrhage (SAH) due to rupture of IAs is one of the three main subtypes of stroke. The incidence of SAH as well as its 30-day mortality has remained stable for over 3 decades despite advances in diagnosis and treatment [1-4]. Survivors are often left with substantial disability and a reduction of quality of life [5,6]. Most of the mortality after rupture of an IA is due to rapid and massive brain injury from the initial bleeding that is not correctable by medial and surgical intervention [7]. In addition, unruptured IAs are estimated to be present in at least 1.0% of the general population [4]. Unruptured IAs are associated with a variable risk of aneurysmal rupture that increases with size of the aneurysms [8,9]. Thus, the most effective ways to decrease morbidity and mortality associated with IAs are to prevent formation of IA in the population and to identify asymptomatic unruptured IA in affected individuals prior to rupture.
Non-modifiable risk factors for aneurysmal SAH include advancing age, female gender and African American race [2,3,10-12]. Smoking and hypertension are the modifiable environmental risk factors that have been very strongly linked to intracranial aneurysms and SAH whereas heavy alcohol use and other environmental factors have been less consistently linked [13-19].
Population-based and case-control studies suggest that genetic factors also play an important role in the formation and rupture of IA [18,20-22]. For example, the risk of unruptured IA, as determined by magnetic resonance angiography (MRA) screening of unaffected relatives in families with two or more members who have an IA, is about four times greater than the risk among the general population [23]. Several Mendelian disorders, such as polycystic kidney disease and Ehler's Danlos syndrome (particularly Type IV), are associated with an increased risk of IA formation. However, these disorders account for less than 1% of all IAs in the population and therefore cannot explain the familial aggregation of IA [10].
Several studies have sought to identify the genes contributing to IA susceptibility. Analyses of potential candidate genes for IA, such as α-1 antitrypsin, apolipoprotein E, and lipoprotein lipase, have yielded inconsistent results and appear unlikely to explain the substantial genetic risk for IA [24-30]. Linkage of FIA in two relatively small linkage studies to a region on chromosome seven containing the elastin gene has not been confirmed by other studies [31,32]. Another small linkage study in a Finnish population has identified a linked region on chromosome 19 that has yet to be confirmed by another study [33]. Most recently, a single, large family segregating an apparently autosomal dominant form of IA has been linked to chromosome 1p [34]. Thus, the genetic cause of IA remains largely unknown. The long-term aim of the Familial Intracranial Aneurysm Study is to identify genes that increase the risk of development and rupture of IAs.
A critical limitation in a number of family studies of IA is the significant early mortality. As a result, collection of a sufficient number of families with affected individuals can severely limit the success of potential genetic studies. One approach to improve the power to detect IA susceptibility genes is to recruit families with multiple members with IA, and to enroll the children and spouses of deceased affected in these families. Similar to the approach used in forensic studies, analysis of the DNA of closely related individuals of the deceased person allows for the reconstruction of the likely genotype of the deceased, affected person- thereby improving the power of genetic analyses. A second approach to improve the power of IA studies is to perform MRA studies among high-risk family members so as to identify individuals with unruptured aneurysms. When performing genetic analyses, those individuals with unruptured IA will be considered affected and thus improve the power of genetic analyses. A third approach to further increase the power of genetic studies is to perform analyses to detect loci contributing to IA susceptibility and/or rupture risk. In this way, unique and potentially overlapping loci contributing to each effect can be identified.
Methods and Research Design
The FIA collaborative group
The FIA Study consists of the Coordinating Center (University of Cincinnati), two genotyping centers (The Center for Inherited Disease Research and University of Cincinnati), the Imaging Center (Mayo), the Cell Repository (Coriell), two Statistical Genetics Centers (Indiana University and National Human Genome Research Institute) and 26 clinical centers. To maximize the ability to recruit families with IA, clinical centers with extensive experience in the clinical management and imaging of intracerebral aneurysms were selected as recruitment sites. These centers (41 recruitment sites) are located throughout North America, New Zealand, and Australia. The FIA study has been approved by the Institutional Review Boards/Ethics Committees at each of the study centers, recruitment sites and participating Centers.
Definition of phenotype
The primary phenotype is an intracranial aneurysm which is a berry-like defect in the wall of an intracranial artery at the base of the brain. Individuals with both ruptured and unruptured IAs are considered to have the phenotype.
Study population
Four hundred seventy-five families with multiple members diagnosed with IA will be enrolled to identify the chromosomal regions associated with an increased risk of IA and to determine the effects of environmental factors on the expression of genes within these regions. Eligible families for this study include: 1) Families with at least 2 living affected siblings. 2) Families with at least 2 affected siblings, one of whom is living and the other whose genotype can be reconstructed through the collection of closely related, living family members [i.e. Spouse and children] 3) Families with ≥ 3 affected family members (e.g. cousin, uncle, aunt), two of whom are alive and have living connecting relatives. 4) Families with ≥ 3 affected family members, with one living affected and at least one other affected relative whose genotype can be reconstructed through the collection of closely related, living family members.
Exclusion criteria include a fusiform-shaped unruptured IA of an intracranial artery; an IA which is part of an arteriovenous malformation; a family history of polycystic kidney disease, Ehlers Danlos Syndrome, Marfan's Syndrome, fibromuscular dysplasia or Moya-Moya syndrome; or failure to obtain informed consent from the patient or family members.
Probands
Probands are the first living person identified within a family, who have had a confirmed diagnosis of an IA and do not meet any of the exclusion criteria.
Other affected family members
The proband, or family members of the proband, contacts other potentially affected family members to determine their willingness to participate in the study. If they are agreeable, a study coordinator makes contact with the additional affected family members. For deceased and living relatives with a family history of IA, SAH or intracerebral hemorrhage (ICH), medical records (e.g. hospital records, imaging reports, autopsy, death certificate) are requested after the patient's or family's permission has been obtained. In addition, a phone screen is completed to document the symptoms, diagnostic testing, and management of a diagnosed IA.
Verification of phenotype
All medical records and the phone screen of probands and family members with a reported history of IA, SAH or ICH are reviewed by a Verification Committee. This committee consists of study neurologists at the University of Cincinnati and The Mayo Clinic. Two neurologists independently review the records and decide if the subject meets all the inclusion and exclusion criteria. In cases of disagreement, a third neurologist is used to resolve the case diagnosis as a tiebreaker. Each potential affected family member is ranked as:
Definite
Medical records document aneurysm on angiogram, operative report, autopsy, or a non-invasive imaging report (MRA, CTA) demonstrates an IA measuring greater than 7 mm.
Probable
Death certificate mentions probable intracranial aneurysm without supporting documentation or autopsy. Death certificate mentions subarachnoid hemorrhage without mention of aneurysm and a phone screen is consistent with ruptured IA (severe headache or altered level of consciousness [LOC]) rapidly leading to death. An MRA documents an IA that is less than 7 mm but greater than 3 mm.
Possible
Non-invasive imaging report documents an aneurysm measuring between 2 and 3 mm. Subarachnoid hemorrhage was noted on death certificate, without any supporting documentation, autopsy or recording of headache or altered LOC on phone screen. Death certificate lists 'aneurysm' without specifying cerebral location or accompanying SAH.
Not an affected Case
there is no supporting information for a possible IA.
Enrollment of non-affected family members
Once the eligibility of the family is established, all first degree relatives of the affected family members who have expressed interest in the study are contacted and enrolled. If the affected person is alive, the first-degree relatives, including the brothers and sisters of the affected person, are contacted for enrollment. If the affected person is deceased, then the brothers and sisters as well as the spouse and children of the affected person are recruited for the study. These additional family members, particularly the spouse and children, will be used to reconstruct the genotype of the deceased, affected individual. To further improve the power of genetic analyses, when the family consists of pairs of affected family members who are more distantly related (i.e. cousins, avuncular, etc), linking relatives such as aunts, uncles and parents are enrolled into the study.
Recruitment methodology
Every new case of IA or SAH is screened for a family history of IA or intracranial hemorrhage at the study clinical centers. To ensure adequate enrollment, each selected clinical site is a major referral center for the diagnosis and treatment of IAs. We utilize three recruitment strategies that include prospective monitoring of all IA cases at the clinical centers, retrospective review of IA cases, and advertisement on the internet and in scientific journals. To help with accessibility to the coordinating center, a study web site and toll free number (800-502-4-3427) has been established.
An unexpected source of potential FIA families has been direct contact of the FIA study personnel by potential probands who have heard about the study through print, radio or TV news media. This source of referrals has resulted in the identification of 77 (19%) of the first 405 families. Use of all four recruitment methods has lead to a four-fold increase of identified potential families over the expected rate of recruitment. At the end of the first 19 months, 405 families have been identified, as compared to an expected 164 families.
Enrollment of study subjects
Once a potential family is identified and the permission of the treating physician has been obtained, the proband or their proxy is given an information sheet describing the study and is asked to enroll into the study. Prior to giving informed consent, a subject must pass the short version of the Blessed Information-Memory-Concentration Scale (Short Blessed Test). This scale assesses the presence and severity of significant cognitive impairment. All six items are applicable to subjects whether in the home or in an institution, can be administered in person or over the phone, and can be applied to persons with physical handicaps (e.g. blindness). In accordance with local regulations, if the subject does not pass the Short Blessed Test, a proxy is chosen then a medical history questionnaire is administered and a blood sample is taken. A phone screen is completed and release of medical information is obtained for phenotype verification. The proband or their proxy representative is then asked to complete a family history questionnaire; study information sheets and recruitment letters are provided to be distributed to additional family members.
Interested family members who live at a distance from a study site are contacted by the Study Coordinating Center and informed consent is obtained via phone, mail or fax. The medical history questionnaire is obtained via phone. A home health agency, which has contracted with the study, then collects the blood samples and obtains the blood pressure readings.
MRA screening of FIA family members without known IA
MRA, a non-invasive imaging technique of cerebral vessels that images the movement of protons, has been demonstrated to be an effective method to screen for IA in asymptomatic relatives. Compared to the gold standard, cerebral angiography, the sensitivity of MRA in detecting IA ranges from 81 to 95 % (most studies greater than 90%) [35-43]. The interobserver consistency in identifying IA's by MRA, particularly for IA's >3 mm is good to excellent (kappa 0.59–0.82) [35-41,43].
Identification of asymptomatic IA cases by MRA provides additional living affected individuals to be used in linkage analysis. To maximize the ascertainment of asymptomatic IAs and to minimize the costs of imaging associated with the study, a study protocol is used to identify individuals at high risk for IA. This protocol calls for the imaging of first degree relatives of affected subjects who are 30 years of age or greater, have a history of hypertension or a mean blood pressure of > 140 mmHg systolic, or >90 mmHg diastolic and/or a 10 pack-year history of smoking.
Patients referred for MRA are imaged at a limited number of certified sites. To obtain certification, the site must demonstrate the ability to perform high quality MRA according to certain standards identified by the Imaging Center at the Mayo Clinic. All study MRAs are de-identified prior to shipment to the Imaging Center where they are reviewed by two neuroradiologists. Agreement between the 2 readers is declared when both detect no aneurysm or both identify an aneurysm at the same site with a measurement difference of no greater than 1 mm. When a disagreement between the readings of the two neuroradiologists occurs, a consensus reading is complete. If consensus cannot be reached, adjudication by a third, blinded reader is performed.
When an IA is identified by MRA, the Coordinating Center notifies the Principal Investigator (PI) at the site of subject enrollment. The local PI contacts the subject, advising them and their physician of the findings. Clinical management of those with unruptured IAs is not dictated by the study.
Yearly follow-up
Once a year, a medical information update and quality of life form is obtained from each participant. This form provides a means of maintaining contact with participants as well as collecting information concerning any deaths or new cases of IA that may have occurred since completion of the Family History Questionnaire.
DNA collection, abstraction, and initial genotyping
Each FIA site is provided with blood collection kits. Two tubes collected in EDTA are used by DNA extraction and a third tube is used for isolation and immortalization of lymphocytes at Coriell, the DNA repository for the National Institute of Neurologic Disorders and Stroke. The extracted DNA is batched and sent to The Center for Inherited Disease Research (CIDR). When it has been received by CIDR a 10 cM genome scan is performed using automated fluorescent microsatellite analysis
Reconstruction of genotypes
To improve the power of genetic analyses despite the high mortality of this condition, reconstruction methods are used to infer the genotypic data of a missing individual. Prior to initiating the study, three-generational pedigrees were simulated consisting of a pair of siblings, their deceased parents, one of the sibling's spouses, and three offspring. A marker with heterozygosity of 70% was simulated. Then, the genotypic data of the sibling with the spouse and offspring were removed. Analyses were then performed to estimate the accuracy with which the missing individual's genotype could be inferred using a sequentially smaller number of relatives. Six conditions were tested: a) three offspring, with and without DNA of the deceased individual's spouse; b) two offspring, with and without the deceased individual's spouse; and c) one offspring, with and without the DNA of the deceased individual's spouse.
The program Allegro was used to infer the genotype of the missing individual [44]. The haplotyping algorithm in Allegro was used to impute the most likely genotype of the missing individuals. The inferred genotype was tabulated for the missing individual and the percentage of families in which the correct genotype was inferred was determined. Results were independently obtained for each of the six family structures with varying numbers of individuals available for assisting in the reconstruction of the missing individual. The proportion of families in which the genotype was correctly inferred was clearly dependent on the number of family members available for genotyping. When three offspring and a spouse were available, the correct genotype was inferred as the most likely genotype in about 90% of the families. Even in the worst-case situation, in which only 1 child was available to assist in the reconstruction, the correct genotype was inferred in 73% of the pedigrees. These data demonstrate that the collection of family members of a critical, deceased individual is an effective way to infer the genotype of a missing affected subject.
Statistical methods
Our proposed collection of families will initially consist of families with at least one sib pair (3/4 of families) and families with three or more affected persons but no sib pair (1/4 of families) prior to MRA screening. However, through MRA screening to identify additional asymptomatic individuals, we anticipate that 1/3 of our sib pair families will now consist of a third affected individual. In addition, 1/3 of the other 100 families which initially did not include a sibling pair but had 3 or more affected relatives, will now also include a sibling pair.
To identify chromosomal regions linked to IA, a genome screen will be performed using polymorphic markers genotyped at regular intervals across the genome. Two complementary approaches will be initially employed to detect chromosomal regions linked to the risk for IA. First, parametric methods (i.e. Allegro) will be employed, with the risk for IA modeled as a reduced penetrance, autosomal dominant disorder. A conservative approach to the analysis will be implemented using only the affected individuals; unaffected individuals will be considered as having unknown phenotype. Second, nonparametric method (i.e. Merlin) will also be used to detect linkage. Allele sharing will only be estimated using affected individuals. In both the parametric and nonparametric analytic methods, the genotypes of the unaffected individual are used to infer the genotype of deceased, affected individuals. Both analytic approaches will be performed with both a narrow and broad disease definition. Under the narrow disease definition, individuals will be considered affected only if they are classified as having a definite aneurysm. The broader disease definition will include as affected those individuals with either definite or probable aneurysm.
Since the genotyping of the familial sample will be performed in stages, the initial goal of all linkage analyses is to detect chromosomal regions providing evidence of linkage to IA. Since these studies are an initial screen, modest genome-wide criteria for linkage will be employed. By simulating genotypes for the sampled pedigrees, genome-wide thresholds for linkage can be determined. Initial criteria will employ the 5% and 1% threshold for linkage.
Once linkage has been identified, further analyses will be performed to dissect the likely mechanism of action. Initially, the genetic analyses will be performed using as affected only those individuals whose IA had ruptured. This analysis is designed to determine if the linkage that has been detected is to a locus contributing to IA or to the risk of IA rupture. Subsequent analyses designed to further localize the chromosomal region linked to IA will utilize more effectively the 'unaffected' individuals. Specifically, data from important environmental covariates such as smoking and hypertension will be used to modify the risk an individual has inherited an IA susceptibility gene.
There are few analytic methods currently available that can perform genetic linkage analysis while simultaneously considering the risk of various non-genetic risk factors. Conditional logistic models that include covariates within the model allow the genetic relative risk to depend on the covariate [45-47]. Recent application of this technique suggest substantial improvement in the ability to detect apparently 'true' linkage signals across multiple datasets, which were undetected using conventional linkage analyses [46]. Another approach is a regression-based extension of the mixture likelihood, which includes pedigree features as covariates to determine the probability that a pedigree is linked [48]. For example, this approach was applied to prostate cancer and examined the family effects of the number of affected individuals in the pedigree, mean age at diagnosis and male-to-male transmission [47]. These investigators found that the covariate of male-to-male transmission identified a subset of families with evidence of a unique linkage. Importantly, they, as well as others, found that including insignificant covariates in the linkage analysis reduced the power to detect linked chromosomal regions [49]. A new approach recently applied to sibling pair data is a tree-based recursive partitioning method that produces more homogeneous subgroups of sibling pairs, which can lead to substantial increases in the power to detect linkage [48].
Discussion
A successful, adequately-powered genetic linkage study of IA is challenging given the very high early mortality of ruptured IA. Other diseases such as lung cancer also have a high associated mortality but patients survive months and even years after diagnosis whereas 40% of patients with a ruptured IA die within the first month. This clinical reality necessitates very aggressive recruitment of potential probands who have a ruptured IA within the first week or so after admission to the hospital. This aggressive approach to prospective recruitment of cases, retrospective case identification at very active clinical centers, and extensive coverage of the study in the media has greatly enhanced our recruitment of families over initial expectations.
Accurate and reproducible phenotyping is the key to any genetic linkage study. IA, in general, is a much more clearly defined phenotype than other neurologic diseases such as Parkinson's disease, Alzheimer's Disease, or even ischemic stroke where the signs and symptoms may be very subtle and poorly documented. IA is defined by a very clearly observed anatomic vascular defect visible on vascular imaging, at operation, or at autopsy. In addition, the large majority of subarachnoid hemorrhage is due to rupture of IA. The challenge of phenotyping for IA is the rapid mortality which precludes vascular imaging and may also result in incomplete information in the available medical records or death certificates. For example, a death certificate may be the only medical record and it may clearly indicate that a massive suabarachnoid hemorrhage was the cause of death, but does not mention an IA. To address these issues, we have designed a rigorous phenotyping procedure using medical records, family interviews, and death certificates with carefully defined levels of confidence (definite, probable, and possible). Thus, we will perform a primary statistical analysis using only those individuals who meet the criteria for definite or probable and a secondary analysis that also includes the possible cases.
Phenotyping is tied closely to imaging of cerebral arteries. Intra-arterial angiography, the gold-standard for identification of IAs, has an associated risk of stroke of 0.5–1.0% [50-55]. Major advances in non-invasive imaging of cerebral arteries over the past 10 years have dramatically changed the approach to IAs and have resulted in identification of additional affected individuals in genetic linkage studies of IA with an increase in study power [31,33,56]. However, inclusion of individuals as affected based upon MRA imaging alone, without confirmation by intra-arterial angiography, runs the potential risk of inclusion of some subjects in the analysis as "affected" who do not actually have an IA. Many physicians do not advocate invasive intra-arterial angiography to confirm IA detected by MRA or CT angiography unless there is serious consideration of clipping or coiling of the IA. Often, only those patients with aneurysms that are 5 mm or larger on MRA or CT angiography undergo subsequent intra-arterial angiography. To address this issue of phenotyping using MRA or CT angiography, we have required high-quality methods for MRA imaging at the clinical centers and have developed a rigorous protocol for centralized identification of MRA by highly experienced neuroradiologists. Our phenotyping of IA using MRA (definite, probable, and possible) is also defined by size of the IA since sensitivity and specificity of MRA for identification of IA clearly relates to the size of IA.
This study highlights several relatively novel approaches to increase the power of genetic analysis. The approach to genotype relatives of deceased, affected individuals is critical in genetic studies of IA as well as other high mortality disorders. Our modeling of various pedigree structures indicates that reconstruction of the most probable genotype of a dead affected can be determined by including the spouse and children. Another important aspect of IA genetic studies is the critical need to consider environmental covariates in disease risk. Specifically, inclusion of smoking as a risk factor for IA is essential since nearly 80% of patients with an IA have a history of smoking at some point during their lifetime and the attributable risk associated with smoking at any time for IA is over 50% [18]. However, not all individuals who smoke develop IA. Therefore, we hypothesize that smoking may substantially increase the risk of IA for individuals with particular genotypes at IA susceptibility loci. Therefore, to accurately model this interaction and localize the genes contributing to IA, it is critical to effectively incorporate key environmental covariates in genetic linkage and association analyses
In summary, the NINDS-funded FIA Study is by far the largest genetic linkage study of IA to date. The FIA Study has excellent power to detect genes associated with the development of rupture of IA and has a unique opportunity to understand the complex interrelationships between genes and environment that lead frequently to a deadly outcome.
List of Abbreviations
CIDR, The Center for Inherited Disease Research
FIA, Familial Intracranial Aneurysm
IAs, intracranial aneurysms
ICH, intracerebral hemorrhage
LOC, level of consciousness
MRA, magnetic resonance angiography
NINDS, National Institute of Neurological Disorders and Stroke
PI, Principal Investigator
SAH, subarachnoid hemorrhage
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
JPB conceived of the study and participated in its design and coordination and helped to draft the manuscript. LRS participated in the design and coordination of the study and drafted the manuscript. TF participated in the design of the study statistical methods and the reconstruction of genotypes and helped to draft the manuscript. JH participated in the design of the study imaging methods and helped to draft the manuscript. NP participated in the methods genotype reconstruction. IM participated in the design of the study. RDB participated in the design and coordination of the study and helped to draft the manuscript. All authors have read and approved the final manuscripts.
Appendix
Study Operational Centers
Coordinating Center – University of Cincinnati: J. Broderick, principal investigator ; D. Kleindorfer, co-principal investigator; L. Sauerbeck, study coordinator; S. Ewing, administrator; J. Sester, research assistant; Genotyping Center – University of Cincinnati: R. Deka, principal investigator; D. Smelser, research assistant;Linkage Analysis – Indiana University: T. Foroud, principal investigator; P. M. Conneally, co-principal investigator; M. Daubs, Data Manager; J. Gray, research coordinator; L. Flury, statistician; Imaging Center – Mayo Clinic: J. Huston III, co-principal investigator; D. Kallmes, study neuroradiologist, M. Maronie Smith, MRI study coordinator: Cell Storage Center – Camden, New Jersey: J. Beck, C. Royds,; National Institute of Neurological Disorders: J. Marler; K. Gwinn-Hardy.
Recruitment Centers
University of Alabama at Birmingham: W. Fisher, principal investigator, H. Forson, coordinator; Auckland New Zealand: C. Anderson, principal investigator, E. Mee, co-principal investigator, C. Howe, coordinator, S. Vos, coordinator: Australia: G. Hankey, principal investigator, P. DUrso, principal investigator, N. Knuckey, principal investigator, J. Laidlaw, principal investigator, P. Reilly, principal investigator, N. Dorsch, co-principal investigator, M. Morgan, principal investigator, M. Besser, principal investigator, K. Athanasiadis, coordinator, Claxton, coordinator, J. Davidson, coordinator, V. Dunne, coordinator, S. Ewen, coordinator, J. Griffith, coordinator, S. Pope, coordinator, J. Raftesath, coordinator, E. Ritson, coordinator; Brigham & Women's Hospital: A. Day, principal investigator, J. O'Hare, coordinator; University of Cincinnati: D. Woo, co-principal investigator, M. Zuccarello, co-principal, A. Ringer, co-principal investigator, H. Yeh, co-principal investigator, K. Franklin, coordinator; Cleveland Clinic Foundation: P. Ramussen, principal investigator, D. Andrews-Hinders, coordinator; Columbia University: E. S. Connolly, principal investigator, R. Sacco, co-principal investigator, R. Ellsasser, coordinator, P. Yung, coordinator; University of Florida: S. B. Lewis, principal investigator, R. Dettorre, coordinator, A. Royster, coordinator; Indianapolis Neurosurgical Group: T. Payner, principal investigator, N. Miracle, coordinator, K. Redelman, coordinator; London Health Science Center Research Inc.: G. Ferguson, principal investigator, C. Mayer, coordinator, J. Peacock, coordinator; John Hopkins University: K. Murphy, principal investigator, B. Kohler, coordinator; Massachusetts General Hospital: C. Ogilvy, principal investigator, D. Buckley, coordinator, T. Taytsel, coordinator;McGill University: G. Rouleau, principal investigator, A. Noreau, coordinator, N. Satge, coordinator; University of Maryland: E. F. Aldrich, principal investigator, C. Aldrich, coordinator; Mayo Clinic: R. D. Brown, principal investigator, I. Meissner, co-principal investigator; D. Weibers, co-principal investigator; L. Jaeger, coordinator; University of Michigan: L. Morgenstern, principal investigator, L. Lisabeth, co-principal investigator, A. Caveney, coordinator; New Jersey Medical School: A. I. Qureshi, principal investigator, P. Harris-Lane, coordinator; Northwestern University: H. Batjer, principal investigator, C. Concannon, coordinator, G. Joven, coordinator, K. Matijevich, coordinator; University of Ottawa: M. T. Richard, principal investigator, A. Hopper, coordinator; University of Pittsburgh: A. B, Kassam, principal investigator, G. Seever, coordinator, J. Genevro, coordinator; University of California, SF: C. Johnston, principal investigator, K. Katsura, coordinator; University of Southern California: S. Giannotta, principal investigator, V. Thomson, coordinator, D. Fishback, coordinator; Stanford University Medical Center: G. Steinberg, principal investigator, D. Luu, coordinator; University of Texas at Houston: M. Malkoff, principal investigator, A. Wojner, coordinator; University of Virginia: N. Kassel, principal investigator, B. Worrall, co-principal investigator, S. Cook, coordinator B. Stoutenger, coordinator; University of Washington: D. Tirschwell, principal investigator, P. Tanzi, coordinator; University of Manitoba (Winnipeg), A. Kaufmann, principal investigator, D. Gladish, coordinator; Washington Universit:C. Derdeyn, principal investigator, D. Rivet, co-principal investigator, M. Catanzare, coordinator, D. Gherardini, coordinator.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported in part by the National Institute of Neurological Disorders and Stroke of the National Institute of Health grant number RO1 NS 039512 (Dr. Joseph P. Broderick).
==== Refs
Ingall TJ Whisnant JP Wiebers DO O'Fallon WM Has there been a decline in subarachnoid hemorrhage mortality? Stroke 1989 20 718 724 2728035
Broderick JP Brott T Tomsick T Huster G Miller R The risk of subarachnoid and intracerebral hemorrhages in blacks as compared with whites N Engl J Med 1992 326 733 736 1738378
Broderick JP Brott T Tomsick T Miller R Huster G Intracerebral hemorrhage more than twice as common as subarachnoid hemorrhage J Neurosurg 1993 78 188 191 8421201
Menghini VV Brown RD JrSicks JD O'Fallon WM Wiebers DO Incidence and prevalence of intracranial aneurysms and hemorrhage in Olmsted County, Minnesota, 1965 to 1995 Neurology 1998 51 405 411 9710011
Hop JW Rinkel GJ Algra A van Gijn J Quality of life in patients and partners after aneurysmal subarachnoid hemorrhage Stroke 1998 29 798 804 9550514
Hackett ML Anderson CS Health outcomes 1 year after subarachnoid hemorrhage: An international population-based study. The Australian Cooperative Research on Subarachnoid Hemorrhage Study Group Neurology 2000 55 658 662 10980729
Broderick JP Brott TG Duldner JE Tomsick T Leach A Initial and recurrent bleeding are the major causes of death following subarachnoid hemorrhage Stroke 1994 25 1342 1347 8023347
The International Study of Unruptured Aneurysms Investigators Unruptured intracranial aneurysms – risk of rupture and risks of surgical intervention. International Study of Unruptured Intracranial Aneurysms Investigators N Engl J Med 1998 339 1725 1733 9867550 10.1056/NEJM199812103392401
Juvela S Porras M Poussa K Natural history of unruptured intracranial aneurysms: probability of and risk factors for aneurysm rupture J Neurosurg 2000 93 379 387 10969934
Longstreth W Gorelick P, Alter M Nontraumatic Subarachnoid Hemorrhage Handbook of Neuroepidemiology 1994 New York: Marckel Dekker, Inc 123
Ingall T Asplund K Mahonen M Bonita R A multinational comparison of subarachnoid hemorrhage epidemiology in the WHO MONICA stroke study Stroke 2000 31 1054 1061 10797165
The ACROSS Group Epidemiology of Aneurysmal Subarachnoid Hemorrhage in Australia and New Zealand Incidence and Case Fatality From the Australasian Cooperative Research on Subarachnoid Hemorrhage Study (ACROSS) Stroke 2000 31 1843 1850 10926945
Bonita R Cigarette smoking, hypertension and the risk of subarachnoid hemorrhage: a population-based case-control study Stroke 1986 17 831 835 3094199
Fogelholm R Murros K Cigarette smoking and subarachnoid haemorrhage: a population-based case-control study J Neurol Neurosurg Psychiatry 1987 50 78 80 3819759
Shinton R Beevers G Meta-analysis of relation between cigarette smoking and stroke BMJ 1989 298 789 794 2496858
Knekt P Reunanen A Aho K Heliovaara M Rissanen A Aromaa A Imprivaara O Risk factors for subarachnoid hemorrhage in a longitudinal population study J Clin Epidemiol 1991 44 933 939 1890435 10.1016/0895-4356(91)90056-F
Longstreth WT JrNelson LM Koepsell TD van Belle G Cigarette smoking, alcohol use, and subarachnoid hemorrhage Stroke 1992 23 1242 1249 1519278
Kissela BM Sauerbeck L Woo D Khoury J Carrozzella J Pancioli A Jauch E Moomaw CJ Shukla R Gebel J Fontaine R Broderick J Subarachnoid hemorrhage: a preventable disease with a heritable component Stroke 2002 33 1321 1326 11988610 10.1161/01.STR.0000014773.57733.3E
Anderson CS Feigin V Bennett D Ruey-Bin L Hankey G Jamrozik K Active and passive smoking and the risk of subarachnoid hemorrhage: an international population-based case-control study Stroke 2004 35 633 637 14752125 10.1161/01.STR.0000115751.45473.48
Wang PS Longstreth WT JrKoepsell TD Subarachnoid hemorrhage and family history. A population-based case-control study Arch Neurol 1995 52 202 204 7848133
Schievink WI Schaid DJ Michels VV Piepgras DG Familial aneurysmal subarachnoid hemorrhage: a community-based study J Neurosurg 1995 83 426 429 7666217
De Braekeleer M Perusse L Cantin L Bouchard JM Mathieu J A study of inbreeding and kinship in intracranial aneurysms in the Saguenay Lac-Saint-Jean region (Quebec, Canada) Ann Hum Genet 1996 60 99 104 8839125
Ronkainen A Hernesniemi J Puranen M Niemitukia L Vanninen R Ryynanen M Kuivaniemi H Tromp G Familial intracranial aneurysms Lancet 1997 349 380 384 9033463 10.1016/S0140-6736(97)80009-8
Kuivaniemi H Prockop DJ Wu Y Madhatheri SL Kleinert C Earley JJ Jokinen A Stolle C Majamaa K Myllyla VV Norrgard O Schievink WI Mokri B Fukawa O ter Berg JWM De Paepe A Lozano AM Leblanc R Ryynanen M Baxter BT Shikata H Ferrell RE Tromp G Exclusion of mutations in the gene for type III collagen (COL3A1) as a common cause of intracranial aneurysms or cervical artery dissections: results from sequence analysis of the coding sequences of type III collagen from 55 unrelated patients Neurology 1993 43 2652 2658 8255472
Bolger C Phillips J Gilligan S Zourob T Farrell M Croake D Collins P Bouchier-Hayes D Elevated levels of lipoprotein (a) in association with cerebrovascular saccular aneurysmal disease Neurosurgery 1995 37 241 245 7477775
Schievink WI Katzman JA Piepgras DG Schaid DJ Alpha 1-antitrypsin phenotypes among patients with intracranial aneurysms J Neurosurg 1996 84 781 784 8622151
Broderick J Sauerbeck L Khoury J Carrozzella J Kissela B Fontaine R Wenstrup R Woo D Shukla R Moomaw C S and Z mutations in the alpha-a antitrypsin gene are not associated with an increased risk of ruptured intracranial aneurysm Stroke 2000 31 285
Hofer A Hermans M Kubassek N Sitzer M Funke H Stogbauer F Ivaskevicius V Oldenburg J Burtsher J Knopp U Schoch B Wanke I Hubner F Deinsberger W Meyer B Boecher-Schwarz H Poewe W Raabe A Steinmetz H Auburger G Elastin polymorphism haplotype and intracranial aneurysms are not associated in Central Europe Stroke 2003 34 1207 1211 12690215 10.1161/01.STR.0000069013.83336.1C
Ruigrok YM Seitz U Wolterink S Rinkel GJ Wijmenga C Urban Z Association of polymorphisms and haplotypes in the elastin gene in dutch patients with sporadic aneurysmal subarachnoid hemorrhage Stroke 2004 35 2064 2068 15297630 10.1161/01.STR.0000139380.50649.5c
Woo D Kaushal R Pal R Haverbusch M Moomaw CJ Sauerbeck L Chakraborty R Deka R Broderick JP Aneurysmal Subarachnoid Hemorrhage: Associations With ApoE and Eln Gene Variants Stroke 2005 36 465
Onda H Kasuya H Yoneyama T Takakura K Hori T Takeda J Nakajima T Inoue I Genomewide-linkage and haplotype-association studies map intracranial aneurysm to chromosome 7q11 Am J Hum Genet 2001 69 804 819 11536080 10.1086/323614
Yoneyama T Kasuya H Onda H Akagawa H Jinnai N Nakajima T Hori T Inoue I Association of positional and functional candidate genes FGF1, FBN2, and LOX on 5q31 with intracranial aneurysm J Hum Genet 2003 48 309 314 12750963
Olson JM Vongpunsawad S Kuivaniemi H Ronkainen A Hernesniemi J Ryynanen M Kim L Tromp G Search for intracranial aneurysm susceptibility gene(s) using Finnish families BMC Med Genet 2002 3 7 12153705 10.1186/1471-2350-3-7
Nahed BV Seker A Guclu B Ozturk AK Finberg K Hawkins AA DiLuna ML State M Lifton RP Gunel M Mapping a Mendelian form of intracranial aneurysm to 1p34.3-p36.13 Am J Hum Genet 2005 76 172 179 15540160 10.1086/426953
Huston J IIINichols DA Luetmer PH Goodwin JT Meyer FB Wiebers DO Weaver AL Blinded prospective evaluation of sensitivity of MR angiography to known intracranial aneurysms: importance of aneurysm size AJNR Am J Neuroradiol 1994 15 1607 1614 7857409
Wilcock D Jaspan T Holland I Cherryman G Worthington B Comparison of magnetic resonance angiography with conventional angiography in the detection of intracranial aneurysms in patients presenting with subarachnoid haemorrhage Clin Radiol 1996 51 330 334 8641094
Korogi Y Takahashi M Mabuchi N Nakagawa T Fujiwara S Horikawa Y Miki H O'Uchi T Shiga H Shiokawa Y Watabe T Furuse M Intracranial aneurysms: diagnostic accuracy of MR angiography with evaluation of maximum intensity projection and source images Radiology 1996 199 199 207 8633146
Korogi Y Takahashi M Mabuchi N Watabe T Shiokawa Y Shiga H O'Uchi T Nakagawa T Miki H Horikawa Y Fujiwara S MR angiography of intracranial aneurysms: a comparison of 0.5 T and 1.5 T Comput Med Imaging Graph 1997 21 111 116 9152576 10.1016/S0895-6111(96)00064-X
Maeder PP Meuli RA de Tribolet N Three-dimensional volume rendering for magnetic resonance angiography in the screening and preoperative workup of intracranial aneurysms J Neurosurg 1996 85 1050 1055 8929494
Harrison MJ Johnson BA Gardner GM Welling BG Preliminary results on the management of unruptured intracranial aneurysms with magnetic resonance angiography and computed tomographic angiography Neurosurgery 1997 40 947 55 discussion 955-7 9149253 10.1097/00006123-199705000-00014
Grandin CB Mathurin P Duprez T Stroobandt G Hammer F Goffette P Cosnard G Diagnosis of intracranial aneurysms: accuracy of MR angiography at 0.5 T AJNR Am J Neuroradiol 1998 19 245 252 9504473
Cloft HJ Kallmes DF Kallmes MH Goldstein JH Jensen ME Dion JE Prevalence of cerebral aneurysms in patients with fibromuscular dysplasia: a reassessment J Neurosurg 1998 88 436 440 9488296
Brown BM Soldevilla F MR angiography and surgery for unruptured familial intracranial aneurysms in persons with a family history of cerebral aneurysms AJR Am J Roentgenol 1999 173 133 138 10397113
Gudbjartsson DF Jonasson K Frigge ML Kong A Allegro, a new computer program for multipoint linkage analysis Nat Genet 2000 25 12 13 10802644 10.1038/75514
Olson JM A general conditional-logistic model for affected-relative-pair linkage studies Am J Hum Genet 1999 65 1760 1769 10577930 10.1086/302662
Rice JP Rochberg N Neuman RJ Saccone NL Liu KY Zhang X Culverhouse R Covariates in linkage analysis Genet Epidemiol 1999 17 S691 5 10597515
Goddard KA Witte JS Suarez BK Catalona WJ Olson JM Model-free linkage analysis with covariates confirms linkage of prostate cancer to chromosomes 1 and 4 Am J Hum Genet 2001 68 1197 1206 11309685 10.1086/320103
Shannon WD Province MA Rao DC Tree-based recursive partitioning methods for subdividing sibpairs into relatively more homogeneous subgroups Genet Epidemiol 2001 20 293 306 11255239 10.1002/gepi.1
Leal SM Heath SC Searching for alcoholism susceptibility genes using Markov chain Monte Carlo methods Genet Epidemiol 1999 17 S217 22 10597439
Mani RL Eisenberg RL Complications of catheter cerebral arteriography: analysis of 5,000 procedures. II. Relation of complication rates to clinical and arteriographic diagnoses AJR Am J Roentgenol 1978 131 867 869 101046
Mani RL Eisenberg RL Complications of catheter cerebral arteriography: analysis of 5,000 procedures. III. Assessment of arteries injected, contrast medium used, duration of procedure, and age of patient AJR Am J Roentgenol 1978 131 871 874 101047
Dion JE Gates PC Fox AJ Barnett HJ Blom RJ Clinical events following neuroangiography: a prospective study Stroke 1987 18 997 1004 3686597
Earnest F 4Forbes G Sandok BA Piepgras DG Faust RJ Ilstrup DM Arndt LJ Complications of cerebral angiography: prospective assessment of risk AJR Am J Roentgenol 1984 142 247 253 6198889
Cloft HJ Jensen ME Kallmes DF Dion JE Arterial dissections complicating cerebral angiography and cerebrovascular interventions AJNR Am J Neuroradiol 2000 21 541 545 10730648
Willinsky RA Taylor SM TerBrugge K Farb RI Tomlinson G Montanera W Neurologic complications of cerebral angiography: prospective analysis of 2,899 procedures and review of the literature Radiology 2003 227 522 528 12637677
Wills S Ronkainen A van der Voet M Kuivaniemi H Helin K Leinonen E Frosen J Niemela M Jaaskelainen J Hernesniemi J Tromp G Familial intracranial aneurysms: an analysis of 346 multiplex Finnish families Stroke 2003 34 1370 1374 12750547 10.1161/01.STR.0000072822.35605.8B
| 15854227 | PMC1097731 | CC BY | 2021-01-04 16:03:33 | no | BMC Med Genet. 2005 Apr 26; 6:17 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-17 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-211585751410.1186/1471-2180-5-21Research ArticleStrain-specific differences in Neisseria gonorrhoeae associated with the phase variable gene repertoire Jordan Philip W [email protected] Lori AS [email protected] Nigel J [email protected] Bacterial Pathogenesis and Functional Genomics Group, The Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford, OX1 3RE, UK2005 27 4 2005 5 21 21 24 2 2005 27 4 2005 Copyright © 2005 Jordan 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
There are several differences associated with the behaviour of the four main experimental Neisseria gonorrhoeae strains, FA1090, FA19, MS11, and F62. Although there is data concerning the gene complements of these strains, the reasons for the behavioural differences are currently unknown. Phase variation is a mechanism that occurs commonly within the Neisseria spp. and leads to switching of genes ON and OFF. This mechanism may provide a means for strains to express different combinations of genes, and differences in the strain-specific repertoire of phase variable genes may underlie the strain differences.
Results
By genome comparison of the four publicly available neisserial genomes a revised list of 64 genes was created that have the potential to be phase variable in N. gonorrhoeae, excluding the opa and pilC genes. Amplification and sequencing of the repeat-containing regions of these genes allowed determination of the presence of the potentially unstable repeats and the ON/OFF expression state of these genes. 35 of the 64 genes show differences in the composition or length of the repeats, of which 28 are likely to be associated with phase variation. Two genes were expressed differentially between strains causing disseminated infection and uncomplicated gonorrhoea. Further study of one of these in a range of clinical isolates showed this association to be due to sample size and is not maintained in a larger sample.
Conclusion
The results provide us with more evidence as to which genes identified through comparative genomics are indeed phase variable. The study indicates that there are large differences between these four N. gonorrhoeae strains in terms of gene expression during in vitro growth. It does not, however, identify any clear patterns by which previously reported behavioural differences can be correlated with the phase variable gene repertoire.
==== Body
Background
Neisseria gonorrhoeae is the causative agent of the sexually transmitted disease gonorrhoea. In the male this is typically associated with a purulent discharge from the urethra. However, in women, infection of the cervix is often asymptomatic while gonococcal spread up the urinary tract or invasion across the epithelial layers can cause additional complications such as pelvic inflammatory disease (PID) and disseminated gonococcal infection (DGI). Depending on geographical location, between 0.1%–3% of cases of uncomplicated gonorrhoeae (UG) disseminate and cause DGI [1-3]. The mechanisms that allow some strains to invade and cause DGI are not well understood.
Many attempts have been made to correlate disease phenotype with the genetic characteristics of isolates from DGI and UG. Certain phenotypes have been associated with DGI, including the arginine, hypoxanthine, and uracil (AHU) auxotype in which AHU- strains tend to be more invasive [4,5]. Serum resistance, and therefore the ability to invade successfully, is associated with porin serotype 1A, via its interaction with C4 binding protein [6] as well as a role in the invasion of epithelial cells [7]. There is also a correlation between the presence of a combination of atlA and the traG variant, sac-4, and the DGI phenotype. These are thought to be a serum resistance locus and a cytotoxin that are located within the Gonococcal Genetic Island [8]. Addressing this issue clearly is difficult because it is possible that isolates capable of causing disseminated infection will be isolated from uncomplicated infections, and the converse may occur in hosts that are unable to produce a satisfactory immunological defence.
There are four experimental strains of N. gonorrhoeae commonly used in the laboratory. Strain FA1090, the only gonococcal strain for which a genome sequence is currently available , was isolated from the cervix of a patient with DGI. Strain FA19 was isolated from a patient with both UG and DGI. Strains F62 and MS11, both of which were isolated from patients with UG, are also frequently used. Several differences have been reported between these strains.
These include the ability to acquire iron from lactoferrin due to expression of lactoferrin binding protein (Lbp). A competitive advantage associated with this receptor in male urogenitary tract infection has been reported, however the absence of this gene from approximately half of gonococcal strains suggests that there may also be an advantage to its absence [9]. Other differences include an elevated ICAM-1 response generated in epithelial cell lines upon exposure to strain FA1090 as compared to strain MS11 [10]. One of the largest differences is the presence of the Gonococcal Genetic Island in strain MS11 [8] and strain FA19, but not the other two strains [11]. Microarray studies have revealed that there is little difference in the gene complements of these four strains. Indeed, these studies have shown that strain F62 contains all of the genes encoded by the FA1090 genome [11]. What comparative gene hybridization studies cannot tell us is whether there are small differences in the gene sequences, including frame-shifts or base substitutions, and whether there are differences in the phase variable gene repertoire.
Phase variation is a process employed by many bacteria to reversibly control gene expression. It is a switching mechanism that allows the gene to be reversibly turned ON or OFF. One difference between this and other regulatory mechanisms is that it does not occur in response to a stimulus but instead is mediated by changes to the DNA, which can occur during replication. There are many mechanisms for phase variation including inversions and transposon movement (for reviews see [12,13]), however, the most common mechanism within the Neisseria spp. is via a slippage mechanism in which the length of a simple sequence repeat changes during replication [14]. Within the coding sequence, changes in this repeat can cause frame-shifting and generation of truncated reading frames. Upstream of the coding sequences changes in repeat length can also alter the expression level of the genes by changing the efficiency with which RNA polymerase or other transcription factors bind to promoter components.
Due to the potential phase variation has to alter the expression profiles of a bacterium, this is one of the next logical areas to investigate to identify differences between the strains of N. gonorrhoeae that might account for differing behaviour. To achieve this a revised list of all of the potential phase variable genes, excluding opa and pilC genes, in N. gonorrhoeae was generated based upon the four publicly available neisserial genomes, Neisseria meningitidis serogroup A strain Z2491 [15], serogroup B strain MC58 [16], serogroup C strain FAM18 (available at ), and N. gonorrhoeae strain FA1090 (available at ). The presence of the potentially variable repeat and the expression status of each of these genes was then determined and compared between the four N. gonorrhoeae strains.
Results and Discussion
A revised list of potential phase variable genes in Neisseria gonorrhoeae
Potential phase variable genes were predicted through identification of simple sequence repeats within coding or promoter regions, the alteration in the length of which would alter the expression of the associated protein. Genes that are common to N. gonorrhoeae and different strains of N. meningitidis tend to be highly conserved, with typical sequence identities of greater than 90%. There is sufficient similarity between them for differences in the length of repeats, with the potential to disrupt reading frames, to increase the quality of phase variable gene predictions. In the Neisseria spp. the phase variable repertoire has been previously defined through three-way genome comparison to identify potential phase variation associated simple sequence repeats [17]. Based on the repeat tract length, the presence of the gene, the presence of variations in the repeat tract length, and the location of the repeat, each of the previously identified potential phase variable genes was scored for the probability of being phase variable. Using the additional genome sequence of N. meningitidis strain FAM18, the predicted phase variable gene repertoire was re-assessed through four-way genome sequence comparison (data not presented), using the same method as previously described [17]. In this way, those potential phase variable genes that are common between the four genome sequences could be evaluated in light of variation in the presence and length of the repeat tract associated with the genes. This analysis identified 64 repeats in N. gonorrhoeae that had the potential to mediate phase variation (Table 1), excluding the opa and pilC genes. These were investigated in N. gonorrhoeae strains FA1090, F62, FA19, and MS11. The opa and pilC genes were not included in this study because these genes are established as phase variable, to change very rapidly during culture, and do not have sufficiently distinct priming sites in the multiple alleles for amplification. Therefore analysis of these genes would not be possible, and it would not be possible to interpret the meaning of any observed changes.
Repeat length changes
Repeat length and composition were compared between the four gonococcal strains and differences were observed in 35 of the 64 (55%) repeats studied (Table 2), although no sequence was obtained for the repeat in XNG1511. The variation can be seen in two forms; either as variation of the repeat length, or sequence composition, between the four strains of N. gonorrhoeae as determined on the basis of the predominant population following sequencing, for example XNG0412 in which strain FA1090 contains a (G)7 repeat tract while strain FA19 contains a (G)6 repeat tract. In addition, although PCR can generate variation in repeat numbers leading to mixed populations in sequencing for templates of homopolymeric tracts of (C or G)11 or greater, when variation is seen in shorter homopolymeric tracts, or in repeats composed of longer repeated motifs, this can be taken to indicate that variation has occurred in vivo during culture. This has been established in a previous study in H. pylori using a similar amplification and sequencing strategy [18]. Observed changes in these repeat tracts provide additional evidence of phase variation, and is indicated by 'plate' in Table 2. For example, XNG0080 in strain FA19 is identified as having a (C)10 tract in Table 2, but this is the predominant population and in fact the population contains (C)8-(C)10 tracts in this gene. 12 of the 36 variable repeats show this form of instability during culture.
The threshold at which variation is observed in repeat length for N. gonorrhoeae appears to be in close agreement with that used to identify genes that had the potential to be phase variable in Neisseria spp. Variation is seen in re peats of poly-G/C tracts ≥ 7 nucleotides (nt), and poly-A/T tracts of ≥ 9 nt, while tetramers, pentamers, and heptamers show variation between strains even at low copy numbers of the repeat (differences between strains are observed in most of these repeats with only 2–4 copies of the repeat present e.g. the (CCCAA) repeat in XNG0520). There were no differences observed in either of the dinucleotide repeat tracts. G/C repeats ≥ 9 nt, tetramers, and pentamers (of all observed sizes) are referred to as variable copy number repeats subsequently because they tend to be unstable during in vitro culture. Therefore, when a variable copy number repeat shows no variation in the sequence data the associated ON or OFF phenotype is likely to be adaptive for the in vitro culture conditions. For example a (C)11 repeat in the gene XNG1942/XNG1941 in strain MS11 is relatively stable in vitro. This gene contains a (C)6 repeat in strain FA1090, which shows the value of investigating the shorter repeats and identifying variation based upon this in other strains.
Of the repeats previously experimentally determined to be phase variable (candidacy of 'K' in Table 1), nine of the ten genes show variation in the repeat including fetA (XNG1989) and lgtA (XNG2045). Variation in the hmbR repeat is not seen as the tract is interrupted by base substitutions and indeed the gene is degenerate (i.e. contains multiple frameshifts and/or in frame termination codons) in all four strains. hmbR was previously reported to be degenerate in N. gonorrhoeae strain MS11 [19]. In some genes predicted from sequence analysis (including many encoding hypothetical proteins), but not previously demonstrated to be phase variable, differences in repeat length provide new evidence for the phase variability of these genes. Furthermore, in some repeats mixed populations are observed in the sequence data indicating that these genes are indeed being phase varied during in vitro culture e.g. pglH (XNG0080). In Table 2 these are marked as 'plate' in the column marked 'Differences in Repeat'.
Repeat length stability in N. gonorrhoeae strain FA1090
The FA1090 strain used in this study has a separate passage history than that used for the genome sequence, yet the copy numbers of the repeats are highly consistent between those observed in this study and the genome sequence. The predominant populations of the variable copy number repeats are mostly the same as the genome sequence. Some variation in the copy number of tetramers and pentamers is observed, but these tend to be more variable in repeat length amongst all strains. There is only one low copy number repeat that differs in length compared to the genome sequence. This is in the coding sequence for methionine aminopeptidase (XNG1868) and the repeat is indicated as being (G)7 in the genome sequence but for all strains in this study it was observed as (G)6, which is unlikely to be phase variable. Variation in expression of this gene is probably not favourable for cell survival, as it codes for an important protein involved in removing the N-terminal methionine from nascent peptides. It is thought therefore that this variation is probably due to an error in the genome sequence rather than phase variation, and is consistent with independent sequencing that observed a (C)6 repeat in strain FA1090 (L Snyder, unpublished).
Repeat length differences between the unrelated strains
Amongst the four strains, there are differences in copy number or composition of the repeat in 55% of the genes studied. Not all of these differences are necessarily associated with changes in expression. Table 2 shows the repeat length and expression status of the gene and it can be seen that some repeats are variable but still maintain an OFF phenotype e.g. the type I restriction enzyme S (XNG0388). Other repeats are associated with genes that are in fact degenerate, e.g. hmbR (XNG1216), aspA (XNG0986), and porA (XNG0832), an association that had been previously described [14]. Not all of the observed strain differences are due to repeat variation as in genes that are irreversibly switched ON or OFF. For lbpA (XNG0245), a gene that is highly variable in N. meningitidis, there is no evidence of phase variation in the gonococcus and what is seen instead is expression or absence due to a deletion in the N-terminus of the gene in both strains FA1090 and F62. In approximately 45% of isolates this deletion prevents expression of the Lbp receptor. It is proposed that there may be a selective advantage for non-expression of this gene over strains with the ability to extract iron from lactoferrin [9].
The potential expression state of each gene was then examined with respect to variation in the repeat length and frameshifts in the gene. These core experimental strains have been passaged repeatedly under in vitro laboratory conditions. Therefore, the expression state of the phase variable repertoire in vitro may not reflect the phenotypes expressed during infection. The expression state of strain FA1090 has, however, shown considerable stability through different passage histories, and the phenotypes in the other strains may also be similarly stable and different. Indeed, theoretical models indicate that in the absence of selection that favours more fit phenotypes, phase varied phenotypes and their associated repeats will tend to be stable [20]. There will still be random changes between equally fit alternatives, but the net rate of change within the population will be substantially slower than the changes that occur in the presence of selective conditions. Observed differences may well account for at least some of the reported differences in strain behaviour, even if the phase variable repertoires are similar.
The genes that contain repeats associated with differences between the strains are not evenly distributed among the different functional groups of genes investigated, with the genes for surface sugar biosynthesis (lipopolysaccharide (LPS) and pilin glycosylation) and restriction modification systems being the most variable in expression and repeat copy number.
It is known that there are four phase variable genes involved in LPS biosynthesis in N. gonorrhoeae, lgtA, lgtC, lgtD, and lgtG, and repeat length variation is observed in all these genes in this study. When these genes switch they can produce a diverse range of surface structures even within a single strain. In vivo certain LPS structures have been shown to provide a mechanism for serum resistance because they can bind CMP-NANA and this allows resistance to the antibody-mediated arm of the complement cascade [21].
The pilin glycosylation enzymes attach sugars to the pilus of the Neisseria spp. and the four studied here all show evidence of phase variation, with in vitro instability. Phase variation has previously been seen in pgtA/pglA (XNG1644), pglE (XNG0189), pglG (XNG0081) and pglH (XNG0080) [22-24]. The role of glycosylation is unclear, but as the pilus is a surface-exposed structure glycosylation may have a role in masking immunogenic parts of the pilus. Indeed it has been reported that glycosylation may inhibit complement-mediated lysis due to its ability to bind anti-gal antibodies [25]. It has been seen that galactosidase treatment of pili markedly reduces attachment of the pilus to host cells suggesting an importance of the galactose residues for attachment [26]. The role of phase variation in the genes that produce the saccharide is unclear, although it is possible that different sugar forms may be associated with different properties with respect to attachment and colonisation.
N. meningitidis contains four uptake systems for iron that are phase variable. These are the lactoferrin uptake system LbpAB, the haemoglobin uptake systems HmbR and HpuAB, and the siderophore receptor FetA. In the N. gonorrhoeae strain FA1090 genome sequence hmbR is degenerate due to a premature frameshift, and this study shows that the hmbR gene is not intact in any of the four gonococcal strains. The lbpA gene contains a deletion in strains FA1090 and F62, but is present and intact in strains FA19 and MS11 but is not phase variable. fetA and hpuA are both present and contain phase variable tracts that exhibit variation during in vitro culture.
Genes with unassigned functions
13 of the 23 hypothetical genes show differences in their repeat lengths or composition. XNG0663 is a hypothetical protein that has limited homology to Curli-like repeat regions. These Curli-like repeats are found in Escherichia coli and are associated with adhesion to many different host cell proteins including MHC class I [27], and fibronectin [28]. They are also seen in Salmonella enterica [29] and Porphyromonas gingivalis. XNG0663 does not contain the same variable repeat as the homologous gene in the meningococcus as the poly-C repeat is interrupted by a T residue. It does, however, have a poly-T tract in which T(7) is associated with expression; but it is unlikely that this would be phase variable due to its length. The gene therefore appears to be fixed ON in strains F62 and MS11 and OFF in strains FA1090 and FA19.
Due to its apparent association with strains that cause DGI rather than UG the expression status of XNG0663 was determined in several clinical isolates of N. gonorrhoeae. In a study of 20 strains isolated from different types of infection it is apparent that the gene is present and turned ON in the majority of strains and that the suggested association with DGI is probably due to the initial small sample size. These results are shown in table 3.
Conclusion
This study extends the number of N. gonorrhoeae phase variable genes for which there is direct or indirect evidence of repeat length changes from ten to 28, in addition to the pilC and opa genes. Based upon the four most commonly used experimental strains no clear association between the presence of phase variable versions of genes or their phenotypes and invasive potential could be identified, other than the previously described association with pgtA [30]. An apparent association with XNG0663 was discounted following analysis of a larger collection of strains.
Phase varied phenotypes appear to be relatively stable in in vitro culture. The differences in the expressed genes (rather than their ability to phase vary), and combinations of phase varied genes may still play a role in determining the outcome of infection. A recent study shows that during infection the colonisation fitness of H. pylori correlates strongly with a need to phase vary fewer genes [31]. It may be that the early behaviour and subsequent outcome of neisserial infection is also influenced by the phenotypes of inoculating populations.
Methods
Whole genome analysis
The complete genome sequences of N. meningitidis serogroup A strain Z2491 [15], serogroup B strain MC58 [16], and serogroup C strain FAM18 (Sanger Institute; ) and N. gonorrhoeae strain FA1090 (available at ) were analysed using previously described whole-genome analysis methodology [14,17,32], using an ACEDB graphical interface [33]. Briefly, repeats composed of perfect repeats with motifs of 1–10 bases were identified using ARRAYFINDER [34]. All repeats were displayed in their sequence contexts with respect to ORFs and termination codons using the tools within ACEDB, and their neisserial protein and nucleotide homologies. These complete genome sequence databases were then analyzed for simple DNA repeats within their sequence contexts to determine the repertoire of putative phase variable genes. Homopolymeric tracts of greater than 6 Gs or Cs, and greater than 8 As or Ts, were each investigated and repeats below these thresholds when associated with a frameshift. Other repeats composed of ≥ 4 copies of dinucleotides and ≥ 3 copies of tetramer and longer motifs were also investigated. All repeats were analyzed to interpret the significance of the repeat on the basis of sequence context and the potential effect of length variation on the expression of the associated reading frames.
The selected genes addressed in this study are listed in Table 1 with an indication of their likelihood of phase variation based upon sequence analysis and previous publications. In this Table K indicates a gene known and reported to be phase variable from previous studies. Strong candidates (S) include those in which the tract length differs in different strains, but has not yet been shown to vary within a single strain. Strong candidates also include genes with tract lengths similar in composition to those that have been seen to vary in other genes, particularly if this is the source of a frame-shift. For example, the adhesin XNG1371 is not found in the meningococcal genome sequences, but contains a (CAAG)20 tract; similar tracts being seen in virG, vapA, and NMB1507, the lengths of which differ between strains. Further, the homopolymeric tracts in the gonococcal genome specific hypothetical genes XNG0470, XNG0473, XNG0503a, XNG1000, XNG1207a, XNG1511, and XNG1577 are each the sole source of a frame-shift mutation within the coding region and are of lengths that have been seen to mediate phase variation in other genes.
Moderate candidates (M) are those in which the tract length is typical of a phase variable gene in one or more sequences, but for which there is no evidence of tract length differences with other strains, largely due to the absence of the gene or the absence of the repeat tract in other strains. When present, the repeat is either not associated with a frame-shift, as are those strain-specific genes that have been identified as strong candidates for phase variation, or the frame-shift is present in all strains assessed and/or the frame-shift associated repeat length is < 7 bp in a homopolymeric tract.
Low candidates (L) are those in which the repeat tract has not been seen to change. Also included in this category are those genes in which a short homopolymeric repeat tract (< 7 bp) is different between strains and results in a frame-shift mutation that may not be readily reversible. For example, the reduction in the homopolymeric repeat from an in-frame (C)7 to a frame-shifted (C)5 may have then lead to subsequent additional mutations in the gene, rendering it degenerate (e.g. XNG0158).
Bacterial strains and growth conditions
The strains used are shown below in Table 4. Strains were grown on GC agar (Difco Laboratories) with the Kellogg supplement and ferric nitrate [35] at 37°C under 5% (v/v) CO2. Cultures were grown from frozen stocks, passaged once the following day to obtain an inoculum for semi-confluent growth, and DNA was prepared from young healthy colonies early the following morning.
PCR amplification and sequencing
Chromosomal DNA extractions were performed using the Aquapure genomic DNA kit (BioRad). PCR from chromosomal DNA was performed using Hotstar Taq DNA Polymerase (Qiagen) according to the manufacturers' instructions. Primer pairs used are shown in Table 5. If amplification was not successful with primer pair 1 it was repeated with primer pair 2. Automated sequencing was done using ABI Prism® BigDye™ Terminator Cycle Sequencing version 3.0 (Applied Biosystems), and was resolved on an ABI Prism® 3100 DNA Sequencer (Applied Biosystems).
Bioinformatics
ACEDB[33] was used to analyze the complete genome sequences of the Neisseria spp., as described previously[17]. Sequences were read, edited and aligned using Seqlab (Wisconsin Package, Version 10.2, Genetics Computer Group, Madison, Wisc., USA) through the Sir William Dunn School of Pathology/WIMM Computational Biology Research Group (CBRG). Homology searches were performed using BLAST[36] against the EMBL databases, accessed through the CBRG.
Authors' contributions
PWJ carried out the PCR, sequencing, sequence alignment, and drafted the manuscript. LASS generated the list of phase variable genes. NJS conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
LASS is supported by a Wellcome Trust Project Grant.
The N. gonorrhoeae sequence was obtained from the University of Oklahoma, the Gonococcal Genome Sequencing Project, which was supported by USPHS/NIH grant #AI-38399 (L. A. Lewis, A. F. Gillaspy, R. E. McLaughlin, M. Gipson, T. Ducey, T. Ownbey, K. Hartman, C. Nydick, M. Carson, J. Vaughn, C. Thomson and D. W. Dyer from the University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA, and L. Song, S. Lin, X. Yuan, F. Najar, M. Zhan, Q. Ren, H. Zhu, S. Qi, S. M. Kenton, H. Lai, J. D. White, S. Clifton and B. A. Roe from the University of Oklahoma, Norman, OK, USA).
The phase variable gene repertoire was determined using the completed N. meningitidis strain FAM18 genome sequence. These sequence data were produced by the N. meningitidis strain FAM18 Sequencing Group at the Sanger Institute and can be obtained from .
Figures and Tables
Table 1 List of genes identified as potentially being phase variable in N. gonorrhoeae – excluding the opa and pilC genes.
XNG# † TRACT ‡ CANDIDACY * FUNCTIONAL ROLE GENE ANNOTATED FUNCTION
XNG0023 GGGGCGG - L Other prolyl oligopeptidase family protein
XNG0060 C(G)6 - S Surface sugar biosynthesis wbpC / pglI putative LPS biosynthesis or pilin glycosylation protein
XNG0075 (C)7 - L Hypothetical protein hypothetical protein
XNG0080 (C)10 - S Hypothetical protein pglH pilin glycosylation protein
XNG0081 A(C)7 - S Surface sugar biosynthesis pglG pilin glycosylation protein
XNG0112 (A)6 p S Hypothetical protein hypothetical protein
XNG0115 (C)4 - L Other cvaA putative colicin V secretion protein
XNG0158 (C)5 degenerate L Hypothetical protein hypothetical protein
XNG0188 AA(C)5 - L Other potD-2 spermidine/putrescine ABC transporter,
XNG0189 (CAAACAC)4 - K Surface sugar biosynthesis pglE pilin glycosylation protein
XNG0245 degenerate S Iron acquisition lbpA lactoferrin-binding protein A
XNG0388 (G)7 + S R-M Systems hsdS|| type I restriction enzyme S protein
XNG0412 (G)7 + S Hypothetical protein hypothetical protein
XNG0470 (G)7 + S Hypothetical protein hypothetical protein ††
XNG0473 (T)9 + S Hypothetical protein hypothetical protein
XNG0503a (C)6A(C)9 + S Hypothetical protein hypothetical protein
XNG0508 AACCGGCAAACA - L / M Hypothetical protein hypothetical protein
XNG0520 (CCCAA)12 - S R-M Systems type III restriction system methylase
XNG0552 (C)5 degenerate L Hypothetical protein hypothetical protein
XNG0608 CCACACCC - M Other nifS aminotransferase
XNG0610 (G)7 - L Other lldD L-lactate dehydrogenase
XNG0612 (GCCA)37 - S R-M Systems putative restriction system methylase
XNG0644 (A)9 - L R-M Systems modification methylase NlaIV
XNG0663 CCCCTCC + L Hypothetical protein hypothetical protein
XNG0703 (C)7 - L Other dnaX DNA polymerase III, subunits gamma and tau
XNG0714 GGAAGG - L Other mobA molybdopterin-guanine dinucleotide biosynthesis
XNG0832 (T)11C(G)6 p K Surface associated proteins porA outer membrane protein
XNG0887 (AAGC)4 - S Hypothetical protein hypothetical protein
XNG0905 (AAGC)3 + S Surface associated proteins virG AIDA related protein
XNG0954 (C)7 - L Other ppx exopolyphosphatase
XNG0973 (C)6 - L Hypothetical protein hypothetical protein
XNG0986 (C)6 degenerate S Other aspA putative serotype-1-specific antigen
XNG1000 (G)7 + S Hypothetical protein hypothetical protein
XNG1041 piv-Nm1B inserted S Surface sugar biosynthesis putative glycosyl transferase
XNG1207a (G)8 + S Hypothetical protein hypothetical protein **
XNG1216 GGCGGCCAA K Iron acquisition hmbR hemoglobin receptor
XNG1268 (AT)5 - L Other fixP cytochrome c oxidase, subunit III
XNG1281 G(A)7 - S Hypothetical protein hypothetical protein
XNG1341 (CAAG)20 S Surface associated proteins adhesin
XNG1356 CCCTGCACC p M Other pntA NAD(P) transhydrogenase alpha subunit
XNG1384 TCCGCCC - L Other amiC N-acetylmuramoyl-L-alanine amidase
XNG1423 TGTGGGGG - S Other dca competence-associated protein
XNG1460 (C)7 - L Hypothetical protein hypothetical protein
XNG1511 (G)8 + S Hypothetical protein hypothetical protein
XNG1563 (AAGC)3 + S Surface associated proteins vapA virulence associated protein
XNG1577 (G)8 + S Hypothetical protein hypothetical protein **
XNG1589 (C)4TCC - L Other dinG probable ATP-dependent helicase
XNG1644 (G)11 - K Surface sugar biosynthesis pgtA pilin glycosylation protein
XNG1733 (G)8 - M Hypothetical protein hypothetical protein
XNG1740 (C)7 - L Other rplK 50S ribosomal protein L11
XNG1788 (G)8 + M Hypothetical protein hypothetical protein **
XNG1834 (C)8 - M Hypothetical protein hypothetical protein
XNG1856 (TA)5 - L Hypothetical protein hypothetical protein
XNG1858 (G)5 - L Surface associated proteins mafA-3 putative adhesin
XNG1868 (C)7 + L Other map methionine aminopeptidase
XNG1934 (A)9 p S Other putative lipoprotein
XNG1942/41 (C)6 + S Other putative hydrolase
XNG1968 (C)11 - K Surface sugar biosynthesis lgtG lipopolysaccharide glycosyl transferase
XNG1989 (C)13 p K Iron acquisition frpB / fetA iron-regulated outer membrane protein
XNG2004 (G)9 + K Iron acquisition hpuA hemoglobin-haptoglobin utilization protein
XNG2045 (G)11 + K Surface sugar biosynthesis lgtA acto-N-neotetraose biosynthesis glycosyl transferase
XNG2047 (G)14 + K Surface sugar biosynthesis lgtC LPS biosynthesis protein
XNG2048 (G)14 - K Surface sugar biosynthesis lgtD LPS biosynthesis protein
XNG2061 (TTCC)3 - L Other plsX fatty acid-phospholipid synthesis protein
† CDS designation numbers come from our own annotation of the N. gonorrhoeae strain FA1090 genome sequence.
‡ The known or potential phase variable repeat tract from N. gonorrhoeae strain FA1090. The equivalent sequence is identified for genes without a repeat tract that are phase variable candidates from the meningococci. Following the sequence a - or + signifies if the CDS is in-frame or frame-shifted, respectively. Repeats in promoter or predicted promoter locations are indicated with a p. Degenerate indicates a gene with multiple frame-shifts or in-frame termination conditions.
* The predicted likelihood of phase variability. K indicates a known phase variable gene; S indicates a strong candidate for phase variation; M indicates a moderate candidate; L indicates a low candidate. See text for details of these thresholds.
†† XNG0470 is a homologue of XNG1014, and XNG1513 which have a (G)6 and are not frame-shifted.
** XNG1207a, XNG1577, and XNG1788 are homologues of each other.
Table 2 Repeat lengths for each of the genes.
Functional Role Gene Annotated Function FA1090* FA19* F62* MS11* Differences in repeat‡
Surface Associated Proteins XNG0832 Outer membrane protein (T)11C(G)6 p (T)8C(G)8 p (T)11C(G)6 p (T)9C(G)6 p Plate
XNG0905 AIDA related protein (AAGC)3 (AAGC)2 (AAGC)3 (AAGC)3 Strains
XNG1341 Adhesin (CAAG)20 (CAAG)13 (CAAG)9 (CAAG)6 Plate
XNG1563 Virulence associated protein (AAGC)3 (AAGC)14 (AAGC)26 (AAGC)3 Plate
XNG1858 Putative adhesin (G)5 (G)5 (G)5 (G)5
Surface Sugar Biosynthesis XNG0060 LPS synthesis/pilin glycosylation C(G)6 C(G)6 C(G)6 C(G)6
XNG0080 Pilin glycosylation protein (C)10 (C)10 (C)10 Gene not present† Plate
XNG0081 Pilin glycosylation protein A(C)7 A(C) 9 A(C)7 - Gene not present† Plate
XNG0189 Pilin glycosylation protein (CAAACAC)4 (CAAACAC)6(CAAATAC)3 (CAAACAC)4 (CAAACAC)15(CAAATAC)2 Plate
XNG1041 Putative glycosyl transferase piv-Nm1B inserted‡‡ piv-Nm1B inserted‡‡ piv-Nm1B inserted‡‡ piv-Nm1B inserted‡‡
XNG1644 Pilin glycosylation protein (G)11 (G)17 GGGAGCGG GGGAGCGG Strains
XNG1968 LPS glycosyl transferase (C)11 (C)10 (C)10 (C)12 Strains
XNG2045 Glycosyl transferase (G)11 (G)12A (G)16 (G)12 Plate
XNG2047 LPS biosynthesis protein (G)14 (G)13 (G)10 (G)8 Strains
XNG2048 LPS biosynthesis protein (G)13 (G)13 (G)11 (G)13 Strains
Iron Acquisition XNG0245 Lactoferrin-binding protein A Gene dead†† (G)5CGG Gene dead†† TGAAACGG Strains
XNG1216 Hemoglobin receptor GGCGGCCAA GGCGGCCAA GGCGGCCAA GGCGGCCAA
XNG1989 Iron-regulated OMP (C)13 p (C)12 p T(C)10 p (C)15 p Plate
XNG2004 Hb-haptoglobin utilization (G)9 (G)9 (G)11 (G)9 Plate
Restriction Modification Systems XNG0388 Type I restriction enzyme S (G)7 (G)8 (G)8 (G)7 Strains
XNG0520 Type III restriction methylase (CCCAA)3 (CCCAA)2 (CCCAA)11 (CCCAA)8 Plate
XNG0612 Putative methylase (GCCA)11 No sequence** No sequence** (GCCA)22 Plate
XNG0644 Modification methylase NlaIV (A)9 (A)9 (A)9 (A)9
Other XNG0023 Prolyl oligopeptidase family GGGGCGG GGGGCGG GGGGCGG GGGGCGG
XNG0115 Putative colicin V secretion protein (C)4 (C)4 (C)4 (C)4
XNG0188 ABC transporter component AA(C)5 AA(C)5 AA(C)5 AA(C)5
XNG0608 Aminotransferase CCACACCC CCACACCC CCACACCC CCACACCC
XNG0610 L-lactate dehydrogenase (G)7 (G)7 (G)7 (G)7
XNG0703 DNA polymerase III subunits (C)7 (C)7 (C)7 (C)7
XNG0714 Biosynthesis protein GGAAGG GGAAGG GGAAGG GGAAGG
XNG0954 Exopolyphosphatase (C)7 (C)7 (C)7 (C)7
XNG0986 Putative serotype-1-specific antigen (C)6 (C)6 (C)6 (C)7 Strains
XNG1268 Cytochrome c oxidase, subunit III (AT)5 (AT)5 (AT)5 (AT)5
XNG1356 NAD(P) transhydrogenase subunit CCCTGCACC p CCCTGCACC p CCCTGCACC p CCCTGCACC p
XNG1384 N-acetylmuramoyl-L-alanine amidase TCCGCCC TCCGCCC TCCGCCC TCCGCCC
XNG1423 Competence-associated protein TGTGGGGG TGTGGGGG TGTGGGGG TGTGGGGG
XNG1589 Probable ATP-dependent helicase (C)4TCC (C)4TCC (C)4TCC (C)4TCC
XNG1740 50S ribosomal protein L11 (C)7 (C)7 (C)7 (C)7
XNG1868 Methionine aminopeptidase (C)6 (C)6 (C)6 (C)6 Genome
XNG1934 Putative lipoprotein (A)9 (A)8 p (A)8 p (A)8 p Strains
XNG1942/41 Putative hydrolase (C)6 CAAACCCC (C)6 (C)11 Strains
XNG2061 Fatty acid-phospholipid synthesis (TTCC)3 (TTCC)3 (TTCC)3 (TTCC)3
Hypothetical Proteins XNG0075 Hypothetical protein (C)7 (C)7 (C)7 (C)7
XNG0112 Hypothetical protein (A)6 p (A)6 p (A)6 p (A)6 p
XNG0158 Hypothetical protein (C)5 (C)5 (C)5 (C)5
XNG0412 Hypothetical protein (G)7 (G)6 (G)6 not the same gene*** Strains
XNG0470 Hypothetical protein (G)7T (G)6 (G)6CC (G)7T Strains
XNG0473 Hypothetical protein (T)9 (T)8 (T)9 Strains
XNG0503a Hypothetical protein (C)6A(C)9 (C)7A(C)4T(C)3 (C)6A(C)7 (C)6A(C)13 Plate
XNG0508 Hypothetical protein AACCGGCAAACA AACCGGCAAACA AACCGGCAAACA AACCGGCAAACA
XNG0552 Hypothetical protein (C)5 (C)5 (C)5 (C)4 Strains
XNG0663 Hypothetical protein CCCCTCC & T(7) CCCCTCC & T(7) CCCCTCC & T(6) CCCCTCC & T(6) Strains
XNG0887 Hypothetical protein (AAGC)4 (AAGC)15 (AAGC)20 (AAGC)7 Strains
XNG0973 Hypothetical protein (C)6 (C)6 (C)6 (C)6
XNG1000 Hypothetical protein (G)7(T)4 (G)7(T)6 (G)8(T)4 GAGG(T)6 Strains
XNG1207a Hypothetical protein (G)8TTGTT (G)8T (G)6TTGTT (G)10TT Strains
XNG1281 Hypothetical protein G(A)7 (A)8 G(A)7 G(A)7 Strains
XNG1460 Hypothetical protein (C)7 (C)7 (C)7 (C)7
XNG1511 Hypothetical protein No sequence** No sequence** No sequence** No sequence**
XNG1577 Hypothetical protein GGA(G)8TTGTT GGA(G)7T (G)6TTGTT GGA(G)4TTGTT Strains
XNG1733 Hypothetical protein (G)8 (G)7 (G)10 (G)7 Plate
XNG1788 Hypothetical protein (G)8TTGTT (G)6TTGTT (G)6TTGTT (G)12TT Plate
XNG1834 Hypothetical protein (C)8 (C)8 (C)9 (C)8 Plate
XNG1856 Hypothetical protein (TA)5 (TA)5 (TA)5 (TA)5
* A clear box indicates that the gene is expressed. A shaded box indicates that the gene is not expressed.
‡ Indicating variation in the repeat. 'Strains' indicates differences in the repeat between strains. 'Plate' indicates variation of the repeat during in vitro culture. 'Genome' indicates that there is variation from the genome sequence but no observed differences between strains.
† This gene is not present in this strain as seen by microarray comparative genome hybridisation [11].
‡‡ The site-specific recombinase piv-Nm1B has been inserted into this gene in all four strains
†† A large deletion is present in the N-terminus of this gene.
** Amplification and sequencing did not work.
*** Sequence analysis showed that the ORF present here had no homology to XNG0412 from strain FA1090.
Table 3 Clinical strains of N. gonorrhoeae and their expression of XNG0663.
Gonococcal Strain Site of Isolation Composition of poly-C tract Composition of poly-T tract Gene expression*
25527 Blood CCCCTCC (T)6 ON
25563 Blood CCCCTCC (T)6 ON
26034 Blood CCCCTCC (T)6 OFF
26241 Blood CCCCTCC (T)6 OFF
26593 Blood CCCCTCC (T)6 ON
25534 Joint fluid CCCCTCC (T)6 ON
25562 Joint fluid CCCCTCC (T)6 ON
26399 Joint fluid CCCCTCC (T)7 ON
26775 Joint fluid CCCCTCC (T)6 ON
27706 GC CCCCTCC (T)6 ON
27833 GC CCCCTCC (T)6 ON
27886 GC CCCCTCC (T)6 OFF
27921 GC CCCCTCC (T)6 ON
28197 GC CCCCTCC (T)6 ON
28252 GC CCCCTCC (T)6 ON
28386 GC CCCCTCC (T)7 ON
* Gene expression also depends on the presence or absence of an adenine residue between the two tracts. Therefore strains with the same poly-T repeat can be in the open reading frame or be frame-shifted.
Table 4 Gonococcal strains used in this study
STRAIN SITE OF ISOLATION ISOLATED FROM
25527 Blood DGI
25534 Joint fluid DGI
25562 Joint fluid DGI
25563 Blood DGI
26034 Blood DGI
26241 Blood DGI
26399 Joint fluid DGI
26593 Blood DGI
26775 Joint fluid DGI
27706 GC UG
27728 GC UG
27806 GC UG
27833 GC UG
27886 GC UG
27921 GC UG
28197 GC UG
28252 GC UG
28386 GC UG
28480 GC UG
28539 Joint fluid DGI
FA1090 Blood DGI
FA19 Blood DGI
F62 GC UG
MS11 GC UG
Table 5 Primers used for amplification in this study
Gene Forward Primer Reverse Primer
XNG0023-1 GTATTCGTCCGTCCAACTTGAACCG GTGGATTTGGAAGCAGGGGAATTGG
XNG0060-1 TGGAAGGCAAGGAATACGTC GCGGATTTCCACAAAAGAAA
XNG0075-1 CGATGGATACGACCCTGAAGC CGTCCTGCTCGTCGAAATCC
XNG0080-1 GACTGCATGGCGTAAGAGTGG GCCATCACCGCTTCGTCAAAC
XNG0081-1 CAAAAACGGCACGACAGAG GCGATTTGACCGGCTACTT
XNG0112-1 CAAACCCAAACCGTATGC CGTTCTTTGCCGAGAAAGTC
XNG0115-1 ATCCCGACAATTCCTGTCTG CTATTTGCGCTTTCGACCTC
XNG0158-1 AGTACGGCAGTGGTTTGTCC AACCTCACCCTCCACTTTCC
XNG0188-1 GCTTCAGACGGCATTTTCAC AAATAGGGGACGGCGTATTC
XNG0189-1 TTGTGAATGACGGGTCAAAA TCGACCCCGAATAATGTGAT
XNG0245-1 GTAACGGGCGGCTTTTACG TATTCGATTTCGTTGATTGCG
XNG0388-1 TTCGCATAAAAATGAAATGTGG CCCGGATAACCAAAACATCA
XNG0412-1 TATCTGTGGCACGAAGAACG CTATCCCCTTCCATGCAATC
XNG0470-1 AATGTACGCCGGTCTGAAAG ATATCCGCCACTTTTTGCAG
XNG0470-2 AAACCGCGCCTTCGAGTGCG TTCAGGGCGTTGTCCAAATCG
XNG0473-1 GAAATCGGCTTGAGATACGG AATCCCTGTTCCGGATTTTC
XNG0503a-1 CGCGTGTTGCCTAAAAACA TCAAAACGTACTCCCCCTTG
XNG0508-1 TGTTGTTGAAGGCTTCGTTG CTTCCGCATCCTGCAAAT
XNG0520-1 ATACGGCTACACCGTTTTCG TCTTCACGCTGTTTGCGTAG
XNG0552-1 ATTCGAGTTGGCAAAACACC GCGGGAATGTATCCTTGTTG
XNG0608-1 ATCAACATCGCCCAAATCAT ATCAAACCGTTTTCCTGCAC
XNG0610-1 CAAGCGCAAAATGCCGCGTATG TTGACCCAAAACCTGCAAATCGG
XNG0612-1 GTGGAAAAAACCGACCCGTCC CATTGAACAAAGATTAGCCCATCATC
XNG0612-2 CAAACCGAATTAGCCCAGCC TCGGTTTCTGTATTATACGG
XNG0644-1 CCCTGACCGTTATTTTGGAA AGGTGCTCCATATCCATTGC
XNG0663-1 CCCCCACAGCAACATAGAAT TTGGACTGGAGGCCATTTAC
XNG0703-1 TCTTGGAAAACGCCCAATAC GGATTTGTTCGCCGCTTAT
XNG0714-1 CGCCAAACCGACCCACAAG CGTCAAACTCGAATCTGACCG
XNG0832-1 GGCAAGGAAGAAGGGGATAC CGTATAGGCGGACTTGCTGT
XNG0832-2 CGAAAACGCCGAATACGAAGC GGCATCGCCAAACGGATTCGCA
XNG0887-1 ACCTGCAATACGCCTACACC TGTCGTCCTGTGTGTTGAAA
XNG0905-1 GGCCATAGTTTCCGAAGGAT TTAATCGGAAAAACCGAACG
XNG0905-2 CGATTTTTTCATTGAAGACCG GGAGTGGTTTTTCTTCTTCCC
XNG0954-1 CCGCCGGTCTGGACGAAC ACGCAGGATGAAGCCGCAGC
XNG0973-1 GCTTTCACACTCGTCAGCAC TTTCGTTTCGTTCTGGGTTT
XNG0973-2 GTTTTTCAGCTTGTCCGAACC ACGCTTCCGGCTTGACATGG
XNG0986-1 CGGCGTTTTAACAACAGTGA TGGTTCCATCATGCGTATTC
XNG1000-1 GGAAACCCTGTTCCAAGACA CGGGCGATGAGATGAAGTAT
XNG1041-1 GAAAAATACCTTCGCTGCTG TGATTCCGGACTGAGATACG
XNG1207a-1 GCTTCAAAGAAACCGACAGG CGCTCAAAGCCTGAAAAGTC
XNG1216-1 GACACGGTAGCCGGAAGTAA AGCAGGGGCATAATTACACG
XNG1268-1 TGTCGTGCCTGAGTCCAATA AACCAATCGCTATCGGTCAG
XNG1268-2 CATCCGTTCCAATGGGGTTCC TCTTCGCCCAAAGCAGGACC
XNG1281-1 CAAAACCGCTCAAAGCCTAC CTTTTGCTTCCCATGCTTGT
XNG1281-2 GACGCGGTCGTTAAAACTCCC GAAGTCAAACAAGTCAACGACC
XNG1341-1 TCACGTCCTGCATACTCTCG ATGCCAGATAGCCTTCTGCT
XNG1341-2 GCTCACCTCGGCATACTGCC TTTTCGATGCTTCAGAACTTCC
XNG1356-1 TCGAATCCATATCGCTGACA AAGCTGACGATGGTTTGACC
XNG1356-2 ATCCAACTCTCCGCGCATGG GGTAGCCGCTGATGTTTGCC
XNG1384-1 ACACCTACACCCGCCTGAC GCGTGGATGGAGACGAATAC
XNG1423-1 GCAATGGGAAACCGTAAACA GCTTTGGCTTCGGTTGTATT
XNG1423-2 ACGACATCATCAACGACATCC AACATTCCTTGATTAGACAGCC
XNG1460-1 TGCTGTGCAGTCAAAACCTC TGTACATTTGACGGGCATCT
XNG1460-2 GGACGGCGGCAAATGGCTGG TCAGTTCGGCGGTGCTGTGC
XNG1511-1 GGAGTACGCCATCGCCAGC GCGATGCCGCAGGCAAGG
XNG1511-2 TGCGACTGCGCCAACGACG GACGGCGTAGTTTTCAATCC
XNG1563-1 CCGCAAACTGTATTCCGAAC CCGCTGGAGAAATAATGGAA
XNG1577-1 AAAGTATCCCTTCCGCCTGT ATTGTCCGTTTTGAGCTTGC
XNG1589-1 GCCCGTATCGAACAAATCAG GGAGTTTGGCGATGATGACT
XNG1644-1 GGCATTTTGATTGCCCTGTA TAACTTGCATTCGCCATCAG
XNG1644-2 GTTTGGGATTCGCATTTACCC GGAAGCCGTTGACCTTGTCG
XNG1733-1 TCCTTCAATTGAGCCAGAGC GCAAATGTTGATGCGAAAGA
XNG1740-1 GTCAGCAGCAGTCAAATCAGGC TAACGGGGTGGTTGAGGAGG
XNG1788-1 CGCGAATTGCTGGTAAAAAT AAAGTATCCCTTCCGCCTGT
XNG1834-1 GATGCAGCCCTGATGAGTCGG CCATCTCCACCAAAATCGCACC
XNG1856-1 TCGTCAAATTCCCAGTCAAA CGGTAAAAGACGGCATCAAT
XNG1856-2 TGGGATTGGGTTAAAAATACC GATTATAATCGTCAAATTCCC
XNG1858-1 ATATCGTCGTCTTCGGCATC GCATTGGTTTTGGGTTTGAT
XNG1858-2 GGTCGAACAAGAACTTGTGG TCCTTTGCTTACTTTATACGG
XNG1868-1 CCAAGACTTCGTGTTCCCATTG CCGCAATCCAAGGAAGCATTATG
XNG1868RF-2 ACGGCATCATCATCAAAACC AGACTTCGTGTTCCCATTGG
XNG1934-1 GCCGTCTGAATACACAGCAA ACGCTGCTGATACGTTTGGT
XNG1934-2 GTCTGAAGGCTTGGGCTTGC TTTCCAATTCGTCAAACTCG
XNG1942/XNG1941-1 ACACCCTATGCCGTCTGAAC TTTGGTTTTTGCCAGTGTTG
XNG1968-1 GTGCAGAAGCCCAGTCCGAC CCCATCTTGTCGATCAACGC
XNG1989-1 GGAATTGGGACGTGCCGTTG CGATGACTTCGCCCATATCGG
XNG2004-1 ACGGTTATGGGGTCGGCGTAG ACAGCAGGAAATCCCCGTCG
XNG2045-1 GTGCAAAACCTGCTGAAAAA CCCAATTTGCTGACATCGTA
XNG2047-1 ACACATTCAAACACCGCCTG TGGTACGGTCGAGGTAAAGC
XNG2048-1 CGACAAAGCGTATGCTTCAA GACCAAGGCTTCGGGATAAT
XNG2061-1 ATGTCCGCCTGATTATGACC CCTTCGATGGTTTTGAGCAT
==== Refs
Tapsall JW Phillips EA Shultz TR Way B Withnall K Strain characteristics and antibiotic susceptibility of isolates of Neisseria gonorrhoeae causing disseminated gonococcal infection in Australia. Members of the Australian Gonococcal Surveillance Programme Int J STD AIDS 1992 3 273 277 1504159
Kerle KK Mascola JR Miller TA Disseminated gonococcal infection Am Fam Physician 1992 45 209 214 1728091
Meitzner TA Cohen MS Vaccines against gonococcal infection New Generation Vaccines 1997 Marcel Dekker, Inc., New York 817 842
Knapp JS Holmes KK Disseminated gonococcal infections caused by Neisseria gonorrhoeae with unique nutritional requirements J Infect Dis 1975 132 204 208 125773
Noble RC Reyes RR Parekh MC Haley JV Incidence of disseminated gonococcal infection correlated with the presence of AHU auxotype of Neisseria gonorrhoeae in a community Sex Transm Dis 1984 11 68 71 6431620
Ram S Cullinane M Blom AM Gulati S McQuillen DP Monks BG O'Connell C Boden R Elkins C Pangburn MK Dahlback B Rice PA Binding of C4b-binding Protein to Porin: A Molecular Mechanism of Serum Resistance of Neisseria gonorrhoeae J Exp Med 2001 193 281 296 11157049
Bauer FJ Rudel T Stein M Meyer TF Mutagenesis of the Neisseria gonorrhoeae porin reduces invasion in epithelial cells and enhances phagocyte responsiveness Mol Microbiol 1999 31 903 913 10048033
Dillard JP Seifert HS A variable genetic island specific for Neisseria gonorrhoeae is involved in providing DNA for natural transformation and is found more often in disseminated infection isolates Mol Microbiol 2001 41 263 277 11454218
Anderson JE Hobbs MM Biswas GD Sparling PF Opposing selective forces for expression of the gonococcal lactoferrin receptor Mol Microbiol 2003 48 1325 1337 12787359
Jarvis GA Li J Swanson VK Invasion of Human Mucosal Epithelial Cells by Neisseria gonorrhoeae Upregulates Expression of Intercellular Adhesion Molecule 1 (ICAM-1) Infect Immun 1999 67 1149 1156 10024555
Snyder LDJ Saunders N Microarray genomotyping of key experimental strains of Neisseria gonorrhoeae reveals gene complement diversity and five new neisserial genes associated with Minimal Mobile Elements BMC Genomics 2004 5 23 15084227
Henderson IR Owen P Nataro JP Molecular switches – the ON and OFF of bacterial phase variation Mol Microbiol 1999 33 919 932 10476027
Salaun L Snyder LA Saunders NJ Adaptation by phase variation in pathogenic bacteria Adv Appl Microbiol 2003 52 263 301 12964248
Saunders NJ Jeffries AC Peden JF Hood DW Tettelin H Rappuoli R Moxon ER Repeat-associated phase variable genes in the complete genome sequence of Neisseria meningitidis strain MC58 Mol Microbiol 2000 37 207 215 10931317
Parkhill J Achtman M James KD Bentley SD Churcher C Klee SR Morelli G Basham D Brown D Chillingworth T Davies RM Davis P Devlin K Feltwell T Hamlin N Holroyd S Jagels K Leather S Moule S Mungall K Quail MA Rajandream MA Rutherford KM Simmonds M Skelton J Whitehead S Spratt BG Barrell BG Complete DNA sequence of a serogroup A strain of Neisseria meningitidis Z2491 Nature 2000 404 502 506 10761919
Tettelin H Saunders NJ Heidelberg J Jeffries AC Nelson KE Eisen JA Ketchum KA Hood DW Peden JF Dodson RJ Nelson WC Gwinn ML DeBoy R Peterson JD Hickey EK Haft DH Salzberg SL White O Fleischmann RD Dougherty BA Mason T Ciecko A Parksey DS Blair E Cittone H Clark EB Cotton MD Utterback TR Khouri H Qin H Vamathevan J Gill J Scarlato V Masignani V Pizza M Grandi G Sun L Smith HO Fraser CM Moxon ER Rappuoli R Venter JC Complete Genome Sequence of Neisseria meningitidis Serogroup B Strain MC58 Science 2000 287 1809 1815 10710307
Snyder LAS Butcher SA Saunders NJ Comparative whole-genome analyses reveal over 100 putative phase-variable genes in the pathogenic Neisseria spp Microbiology 2001 147 2321 2332 11496009
Salaun L Linz B Suerbaum S Saunders NJ The diversity within an expanded and redefined repertoire of phase-variable genes in Helicobacter pylori Microbiology 2004 150 817 830 15073292
Stojiljkovic I Larson J Hwa V Anic S So M HmbR outer membrane receptors of pathogenic Neisseria spp.: iron-regulated, hemoglobin-binding proteins with a high level of primary structure conservation J Bacteriol 1996 178 4670 4678 8755899
Saunders NJ Moxon ER Gravenor MB Mutation rates: estimating phase variation rates when fitness differences are present and their impact on population structure Microbiology 2003 149 485 495 12624210
Parsons NJ Andrade JR Patel PV Cole JA Smith H Sialylation of lipopolysaccharide and loss of absorption of bactericidal antibody during conversion of gonococci to serum resistance by cytidine 5'-monophospho-N-acetyl neuraminic acid Microb Pathog 1989 7 63 72 2509853
Jennings MP Virji M Evans D Foster V Srikhanta YN Steeghs L van der Ley P Moxon ER Identification of a novel gene involved in pilin glycosylation in Neisseria meningitidis Mol Microbiol 1998 29 975 984 9767566
Banerjee A Wang R Supernavage SL Ghosh SK Parker J Ganesh NF Wang PG Gulati S Rice PA Implications of Phase Variation of a Gene (pgtA) Encoding a Pilin Galactosyl Transferase in Gonococcal Pathogenesis J Exp Med 2002 196 147 162 12119340
Power PM Roddam LF Rutter K Fitzpatrick SZ Srikhanta YN Jennings MP Genetic characterization of pilin glycosylation and phase variation in Neisseria meningitidis Mol Microbiol 2003 49 833 847 12864863
Hamadeh RM Estabrook MM Zhou P Jarvis GA Griffiss JM Anti-Gal binds to pili of Neisseria meningitidis : the immunoglobulin A isotype blocks complement-mediated killing Infect Immun 1995 63 4900 4906 7591153
Gubish ER JrChen KC Buchanan TM Attachment of gonococcal pili to lectin-resistant clones of Chinese hamster ovary cells Infect Immun 1982 37 189 194 6125476
Olsen A Wick MJ Morgelin M Bjorck L Curli, Fibrous Surface Proteins of Escherichia coli, Interact with Major Histocompatibility Complex Class I Molecules Infect Immun 1998 66 944 949 9488380
Olsen A Jonsson A Normark S Fibronectin binding mediated by a novel class of surface organelles on Escherichia coli Nature 1989 338 652 655 2649795
Collinson SK Emody L Muller KH Trust TJ Kay WW Purification and characterization of thin, aggregative fimbriae from Salmonella enteritidis J Bacteriol 1991 173 4773 4781 1677357
Banerjee A Ghosh SK The role of pilin glycan in neisserial pathogenesis Mol Cell Biochem 2003 253 179 190 14619968
Salaun L Ayraud S Saunders NJ Phase-variation-mediated niche adaptation during prolonged experimental murine infection with Helicobacter pylori Microbiology 2005 151 917 923 15758236
Saunders NJ Peden JF Hood DW Moxon ER Simple sequence repeats in the Helicobacter pylori genome Mol Microbiol 1998 27 1091 1098 9570395
Durbin R Thierry-Mieg JT AC. elegans DataBase. Documentation, code and data
Hancock JM Shaw PJ Bonneton F Dover GA High sequence turnover in the regulatory regions of the developmental gene hunchback in insects Mol Biol Evol 1999 16 253 265 10028291
Kellogg DS JrPeacock WL JrDeacon WE Brown L Pirkle CI Neisseria gonorrhoeae. I. Virulence genetically linked to clonal variation Journal of Bacteriology 1963 85 1274 1279 14047217
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712
| 15857514 | PMC1097732 | CC BY | 2021-01-04 16:03:40 | no | BMC Microbiol. 2005 Apr 27; 5:21 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-21 | oa_comm |
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-111586212510.1186/1472-6947-5-11Research ArticleOptimal search strategies for identifying sound clinical prediction studies in EMBASE Holland Jennifer L [email protected] Nancy L [email protected] R Brian [email protected] Hedges Team [email protected] Health Information Research Unit, Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, L8N 3Z5 Canada2 Department of Medicine, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, L8N 3Z5 Canada2005 29 4 2005 5 11 11 5 1 2005 29 4 2005 Copyright © 2005 Holland et al; licensee BioMed Central Ltd.2005Holland 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
Clinical prediction guides assist clinicians by pointing to specific elements of the patient's clinical presentation that should be considered when forming a diagnosis, prognosis or judgment regarding treatment outcome. The numbers of validated clinical prediction guides are growing in the medical literature, but their retrieval from large biomedical databases remains problematic and this presents a barrier to their uptake in medical practice. We undertook the systematic development of search strategies ("hedges") for retrieval of empirically tested clinical prediction guides from EMBASE.
Methods
An analytic survey was conducted, testing the retrieval performance of search strategies run in EMBASE against the gold standard of hand searching, using a sample of all 27,769 articles identified in 55 journals for the 2000 publishing year. All articles were categorized as original studies, review articles, general papers, or case reports. The original and review articles were then tagged as 'pass' or 'fail' for methodologic rigor in the areas of clinical prediction guides and other clinical topics. Search terms that depicted clinical prediction guides were selected from a pool of index terms and text words gathered in house and through request to clinicians, librarians and professional searchers. A total of 36,232 search strategies composed of single and multiple term phrases were trialed for retrieval of clinical prediction studies. The sensitivity, specificity, precision, and accuracy of search strategies were calculated to identify which were the best.
Results
163 clinical prediction studies were identified, of which 69 (42.3%) passed criteria for scientific merit. A 3-term strategy optimized sensitivity at 91.3% and specificity at 90.2%. Higher sensitivity (97.1%) was reached with a different 3-term strategy, but with a 16% drop in specificity. The best measure of specificity (98.8%) was found in a 2-term strategy, but with a considerable fall in sensitivity to 60.9%. All single term strategies performed less well than 2- and 3-term strategies.
Conclusion
The retrieval of sound clinical prediction studies from EMBASE is supported by several search strategies.
==== Body
Background
Clinical prediction guides (CPGs), also known as clinical prediction rules or clinical decision rules, are increasingly sought by frontline clinicians to assist in their decision making process. They provide an objective standard by which to gauge which elements in a patient's history, physical examination and laboratory tests are the most important in forming an accurate clinical assessment [1]. CPGs are created by way of deriving the rule, testing or validating the rule, and assessing the impact of the rule on clinical behaviour (impact analysis) [2-5]. CPGs vary in complexity, but those that require simple calculations on the part of the user are most recommended by CPG advocates [5,6].
CPGs can serve as decision aids for determination of causation, diagnosis, prognosis, or patient responsiveness to treatment [1-3]. Some CPGs have been tailored for online or personal digital assistant (PDA) tools to aid in bedside decision-making [6]. Currently available CPGs cover a wide range of topics. For example, guides help to establish the pretest probability of pulmonary embolus [7], to determine the treatment for pharyngitis [8] and to rule out the need for unnecessary radiography for knee injuries ("Ottawa Knee Rule") [9]. CPG advocates state that, when rigorously created and appropriately applied, CPGs have the potential to influence clinical opinion, change clinical behaviour and increase efficiency while preserving or improving quality patient care and satisfaction [5].
Retrieving CPG studies from the medical literature is problematic for several reasons. First they are relatively few in quantity in comparison to other types of studies and reports posted in major, online, clinical literature databases. For example, EMBASE contains more than 9 million records and is up dated by 6000 – 8000 records per week, spread over more than 4600 journal titles [10]. Second, only a fraction of CPG studies are of high quality [1,3-5,11]. A third problem interfering with retrieval is the plurality in terminology associated with CPG studies including test, rule, index, equation, scale, score, profile, prognosis, risk estimate, and model. Fourth, and coupled with varied terminology, is the lack of standardized controlled indexing vocabulary assigned to CPG studies, which precludes their easy extraction from large databases. Finally, studies show that clinicians lack searching skills and have little time to devote to the task of finding high quality studies on which to base their clinical practice [12,13].
Several research groups have identified search strategies ("hedges", in library parlance) for MEDLINE for topics such as etiology [14], diagnosis [15], prognosis [16,17], treatment [18] and review articles [19]. Our own studies of MEDLINE retrieval [16,17] led to the creation of the PUBMED Clinical Queries search tool .
For CPG study retrieval, 2 papers have reported filters that can be applied to MEDLINE. Ingui and Rogers [2] hand searched 4 to 6 selected journals for the years 1991 through 1998 for studies that described the development, validation or evaluation of a CPG, then developed and tested several search filters. Wong et al published CPG search strategies for MEDLINE developed by comparing hand searching 161 journals published in the year 2000 with several search terms and applying comparatively, more strict methodological quality criteria [20].
Despite the widespread use of EMBASE, especially in the UK and Europe, little is reported about methodological search filters. Bachmann et al [21] and Wilczynski and Haynes [22] reported high performance search strategies for diagnosis studies in EMBASE. In another study, Watson and Richardson tested searches against EMBASE, MEDLINE and PsycInfo for finding randomized controlled trials of cognitive therapy for depression [23]. No studies have yet been published on the retrieval properties of search terms and phrases for CPG studies in EMBASE. To fill this gap in the literature, we applied similar methodology to that used for the identification of CPG study search strategies in MEDLINE [20]. In this report, we describe the information retrieval properties of single terms and combinations of terms for identifying methodologically sound studies of clinical prediction guides in EMBASE.
Methods
We compared the retrieval performance of methodologic search terms and phrases in EMBASE with a manual review of every article for each issue of 55 journal titles for the year 2000. Originally, 170 journal titles were selected based on recommendations of clinicians and librarians, Science Citation Index Impact Factors provided by the Institute for Scientific Information, and ongoing assessment of their yield of studies and reviews of scientific merit and clinical relevance for the disciplines of internal medicine, general medical practice, mental health, and general nursing practice (list of journals provided by the authors upon request). Of these 170 journals, 135 were indexed in EMBASE. In previous work on search strategy development in MEDLINE, we determined that estimation of search term performance was not substantively affected by using smaller journal subsets, focusing on journals publishing at least some methodologically rigorous articles [unpublished data], and these smaller subsets greatly simplify data processing. Hence, 135 journals were further reduced to a 55 journals that were found to contain at least one study that met our criteria for scientific merit.
When previously developing search strategies for some categories of articles (e.g., therapy, prognosis) for MEDLINE, we split the database into 60% and 40% components to provide a development and validation database. We subsequently found that the comparison between development and validation database results was not statistically significant [24]. For CPG search strategy development for EMBASE, it was not feasible to split the database, as there were too few "pass" articles (e.g., 69 pass CPG articles in EMBASE). Thus, search strategies were developed using the entire database.
Six research staff were rigorously calibrated for hand searching before reviewing the 2000 literature and inter-rater agreement for application of all criteria exceeded 80% beyond chance [25]. Hand searching was performed across 55 journals titles for the year 2000, and methodologic criteria were applied to each item in each issue to determine if the article was methodologically sound for 7 purpose categories (two other types of articles, cost and qualitative studies, were also classified but had no rigor criteria). All purpose category definitions and corresponding methodologic rigor were outlined in a previous paper [25]. Clinical prediction studies were defined as having content that pertains directly to the prediction of some aspect of a disease or condition, and the following methodologic criteria were applied: 1) the guide is generated in one or more sets of real patients (training set); and 2) the guide is validated in another set of real patients (test set).
An initial list of index terms and textwords relating to studies of different purposes (clinical prediction guides, treatment, causation, diagnosis, prognosis, economics, reviews, costs, and studies of a qualitative nature) was compiled in house. The list grew with the addition of terms or phrases suggested by clinicians, librarians and known searchers in the United States and Canada, made upon our request. From here, we compiled a list of 5385 searching terms, of which 4843 were unique and 3524 returned results (terms available on request) for retrieval of studies across all of the purpose categories. Among the 3524 terms were 641 terms that depicted clinical prediction studies such as 'clinical prediction rule', 'derivation set', 'guide', and 'validation cohort', all as textwords; 'validation process', the index term, and the index term 'model', exploded.
The search strategies were treated as "diagnostic tests" for sound studies and the manual review of the literature was treated as the "gold standard." All CPG study search terms and phrases were run in EMBASE and an automated process determined their sensitivity, specificity, precision, and accuracy. Sensitivity for a given topic is defined as the proportion of high quality articles for that topic that are retrieved; specificity is the proportion of low quality articles not retrieved; precision is the proportion of retrieved articles that are of high quality; and accuracy is the proportion of all articles that are correctly classified.
The aim of testing was to identify the best single term, 2-term and multiple-term (greater than two terms) strategies that would optimize sensitivity or specificity or both sensitivity and specificity together. All combinations of terms used the Boolean OR, for example, "predict.tw. OR guide.tw.". (The Boolean AND was not used because this strategy invariably compromised sensitivity.) Next, we tested all 2-term search strategies with sensitivity at least 75% and specificity at least 50% to find multiple term strategies that were optimized for sensitivity. For optimizing accuracy in a multiple term strategy, all 2-term search strategies with accuracy >75% were tested. In total, 36,232 search strategies were tested in the development of clinical prediction guide hedges, which represents the second largest number of strategies (next to cost effectiveness studies) tested among EMBASE strategies investigated within our group.
A logistic regression approach to developing search strategies for MEDLINE did not improve performance [26], so it was not performed for this study.
Results
Indexing information was downloaded from EMBASE for 27,769 articles of various purpose categories, identified from hand-searching the 55 journals. Of these, 163 (0.58%) were classified as clinical prediction guides, of which 69 (42.3%) were methodologically sound.
Table 1 shows the best single term for high-sensitivity, high-specificity, and best balance of sensitivity and specificity. The single term, predict:.tw., produced both the best sensitivity (78.3%) and best balance of sensitivity (78.3%) and specificity (91.6%). High specificity (98.7%) was achieved using the single term, validat:tw., but with a concomitant drop in sensitivity to 60.9%. Precision ranged from 2.3% to 10.7% for these strategies, reflecting the low prevalence of high quality CPG studies combined with less than perfect specificity.
Table 1 Single Term with the Best Sensitivity (keeping Specificity ≥ 50%), Best Specificity (keeping Sensitivity ≥ 50%), and Best Optimization of Sensitivity and Specificity (based on the lowest possible absolute difference between sensitivity and specificity) for Detecting Studies of Clinical Prediction Guides in EMBASE in 2000. Values are percentages (95% confidence intervals).
Search term OVID search* Sensitivity (n = 69) Specificity (n = 27700) Precision† Accuracy (n = 27769)
Best sensitivity predict:.tw. 78.3 (68.5 to 88.0) 91.6 (91.2 to 91.9) 2.3 (1.7 to 2.9) 91.5 (91.2 to 91.7)
Best specificity validat:.tw. 60.9 (49.4 to 72.4) 98.7 (98.6 to 98.9) 10.7 (7.6 to 13.7) 98.6 (98.5 to 98.8)
Best Optimization of Sensitivity & Specificity predict:.tw. 78.3 (68.5 to 88.0) 91.6 (91.2 to 91.9) 2.3 (1.7 to 2.9) 91.5 (91.2 to 91.7)
*The search strategy is reported using Ovid's search engine syntax for EMBASE.
†Denominator varies by row. : = truncation; tw = textword (word or phrase appears in title or abstract).
Combination of terms with the best results for sensitivity, specificity and optimization of sensitivity and specificity are shown in Table 2. When multiple terms are run, nearly all measures for sensitivity and specificity improve over results for single term strategies. The 3-term strategy predict:.tw. OR exp methodology OR validat:.tw achieved the best sensitivity (97.1%), with a specificity of 74.2%. Specificity peaked at 98.8%, as did precision (10.8%) and accuracy (98.7%), with the 2-term search phrase validation.tw. OR prediction.tw., but with a significant drop in sensitivity to 60.9%. The strategy that performed best for optimization of sensitivity and specificity was validat:.mp. OR index.tw. OR model.tw, measured at 91.3% and 90.2%, respectively. This 3-term strategy outperformed the best single term search strategy by 13% in sensitivity, while maintaining a comparable value for specificity.
Table 2 Combination of Terms with the Best Sensitivity (keeping Specificity ≥ 50%), Best Specificity (keeping Sensitivity ≥ 50%), and Best Optimization of Sensitivity and Specificity (based on abs [sensitivity-specificity]<1%) for Detecting Studies of Clinical Prediction Guides in EMBASE in 2000. Values are percentages (95% confidence intervals).
Search Strategy OVID search* Sensitivity (n = 69) Specificity (n = 27700) Precision† Accuracy (n = 27769)
Best Sensitivity predict:.tw. OR exp methodology OR validat:.tw. 97.1 (93.1 to 100.0) 74.2 (73.7 to 74.7) 0.9 (0.7 to 1.2) 74.2 (73.7 to 74.7)
Best Specificity validation.tw. OR prediction.tw. 60.9 (49.4 to 72.4) 98.8 (98.6 to 98.9) 10.8 (7.7 to 13.9) 98.7 (98.5 to 98.8)
Best Optimization of Sensitivity & Specificity validat:.mp. OR index.tw. OR model.tw. 91.3 (84.7 to 98.0) 90.2 (89.9 to 90.6) 2.3 (1.7 to 2.8) 90.2 (89.9 to 90.6)
*Search strategies are reported using Ovid's search engine syntax for EMBASE.
†Denominator varies by row. : = truncation; tw = textword (word or phrase appears in title or abstract); exp = exploded subject heading; mp = multiple posting – term appears in title, abstract, or subject heading.
Discussion
In this study we report search filters found to be effective for the retrieval of clinical predication guide studies from EMBASE. These filters are optimized for sensitivity, specificity or best sensitivity and specificity combined, each lending the searcher unique results that can be geared to his/her needs. The strategy optimized for sensitivity should be applied in cases where retrieval of all relevant articles is key, and substantial weeding of irrelevant content is seen to be acceptable. The most specific search filter is effective when the aim of the search is to retrieve only highly relevant articles, where inclusion of all pertinent matter is less important. Where the intention is to uncover a balance of targeted hits with off topic material then the strategy that maximizes both sensitivity and specificity would be best.
When comparing the results of this study to that reported by Wong et al [20] for CPGs in MEDLINE, several similarities can be drawn. Both studies report low precision for most search strategies. Like the MEDLINE search strategies, low precision is attributed to the varying content of the EMBASE database, a small proportion of which are studies of clinical prediction guides. Precision may be improved with the application of the "AND" / "AND NOT" Boolean operators or the addition of clinical content terms or journal subsets using the Boolean AND along with the methodology search filters, but these tactics are likely to compromise sensitivity.
Other parallels occur between the 2 reports with respect to best identified search strategies. For single terms, predict (as a textword in EMBASE, predict:.tw., and as a "multiple posting" term in MEDLINE, predict:.mp.) achieved both highest sensitivity and best optimization for sensitivity and specificity in both EMBASE and MEDLINE. Similarly, validat:.tw. generated best results for specificity in both databases.
It is interesting to note that no indexing terms contributed to the optimized search strategies for CPG studies in EMBASE or MEDLINE; textwords were the composite for all winning strategies. This finding is consistent with indexing terminology not keeping pace with research methods, and suggests a means for improving indexing and retrieval in the future.
For practical purposes, we restricted our methods filter to just 2 criteria, the use of both a training and test set. Additional criteria that could have been applied include prospective validation, stating the mathematical technique used to develop the rule, clear definition and blinded assessment of predictor variables and outcomes, and prospectively testing the effect of the rule in clinical practice [1]. Further research would be needed to determine the performance of our search strategies for these additional criteria. However, it is predictable that adding more criteria would diminish the yield of our search strategies. For example, Laupacis et al [1] found that only 3% of studies of CPGs prospectively tested clinical use, whereas we found that 42% of CPG articles passed our filter. Thus, rather than incorporating additional criteria into the derivation of search strategies, we would suggest that the additional criteria be applied by end-users, to articles retrieved by our strategies, if relevant to their purposes.
For clinicians to be able to make optimal use of clinical prediction guides in their practice, their accessibility needs to be improved upon. This study highlights search terms that maximize the retrieval of CPG studies, as well as illustrating that there is room for improvement, especially in precision. The application of "AND" and "AND/NOT" combinations or multivariate statistical techniques may help, but this remains to be determined.
Conclusion
The retrieval of higher quality clinical prediction guides from EMBASE is facilitated by the application of search filters that have been optimized for sensitivity or specificity or both.
List of Abbreviations
Clinical Prediction Guide: CPG
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
NLW and RBH prepared grant submissions in relation to this project. JH, NLW and RBH drafted, commented on and approved the final manuscript. NLW and RBH supplied intellectual content to the collection and analysis of the data. NLW participated in the data collection and all authors were involved in data analysis.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This research was funded by the National Library of Medicine, USA. The Hedges Team includes Angela Eady, Brian Haynes, Susan Marks, Ann McKibbon, Doug Morgan, Cindy Walker-Dilks, Stephen Walter, Stephen Werre, Nancy Wilczynski, and Sharon Wong.
==== Refs
Laupacis A Sekar N Stiell IG Clinical prediction rules. A review and suggested modifications of methodological standards JAMA 1997 277 488 494 9020274 10.1001/jama.277.6.488
Ingui BJ Rogers MA Searching for clinical prediction rules in MEDLINE J Am Med Inform Assoc 2001 8 391 397 11418546
McGinn TG Guyatt GH Wyer PC Naylor CD Stiell IG Richardson WS Users' guides to the medical literature: XXII: how to use articles about clinical decision rules. Evidence-Based Medicine Working Group JAMA 2000 284 79 84 10872017 10.1001/jama.284.1.79
Randolph AG Guyatt GH Calvin JE Doig G Richardson WS Understanding articles describing clinical prediction tools. Evidence Based Medicine in Critical Care Group Crit Care Med 1998 26 1603 1612 9751601 10.1097/00003246-199809000-00036
Steiner JF Clinical prediction rules as guides to treatment Hosp Pract (Off Ed) 1999 34 17 20 9931573
McGinn T Practice corner: using clinical prediction rules ACP J Club 2002 137 A11 A12 12207456
Wells PS Ginsberg JS Anderson DR Kearon C Gent M Turpie AG Bormanis J Weitz J Chamberlain M Bowie D Barnes D Hirsh J Use of a clinical model for safe management of patients with suspected pulmonary embolism Ann Intern Med 1998 129 997 1005 9867786
Walsh BT Bookheim WW Johnson RC Tompkins RK Recognition of streptococcal pharyngitis in adults Arch Intern Med 1975 135 1493 1497 1103766 10.1001/archinte.135.11.1493
Emparanza JI Aginaga JR Estudio Multicentro en Urgencias de Osakidetza: Reglas de Ottawa (EMUORO) Group Validation of the Ottawa Knee Rules Ann Emerg Med 2001 38 364 8 11574791 10.1067/mem.2001.118011
STN Database Summary Sheet: EMBASE
Wasson JH Sox HC Neff RK Goldman L Clinical prediction rules. Applications and methodological standards N Engl J Med 1985 313 793 799 3897864
Ely JW Osheroff JA Ebell MH Chambliss ML Vinson DC Stevermer JJ Pifer EA Obstacles to answering doctors' questions about patient care with evidence: qualitative study BMJ 2002 324 710 11909789 10.1136/bmj.324.7339.710
McAlister FA Graham I Karr GW Laupacis A Evidence-based medicine and the practicing clinician J Gen Intern Med 1999 14 236 242 10203636 10.1046/j.1525-1497.1999.00323.x
Wilczynski NL Haynes RB Developing optimal search strategies for detecting clinically sound causation studies in MEDLINE AMIA Annu Symp Proc 2003 719 723 14728267
Bachmann LM Coray R Estermann P Ter Riet G Identifying diagnostic studies in MEDLINE: reducing the number needed to read J Am Med Inform Assoc 2002 9 653 658 12386115 10.1197/jamia.M1124
Haynes RB Wilczynski N McKibbon KA Walker CJ Sinclair JC Developing optimal search strategies for detecting clinically sound studies in MEDLINE J Am Med Inform Assoc 1994 1 447 458 7850570
Wilczynski NL Walker CJ McKibbon KA Haynes RB Assessment of methodologic search filters in MEDLINE Proc Annu Symp Comput Appl Med Care 1993 17 601 605 8130545
Adams CE Power A Frederick K Lefebvre C An investigation of the adequacy of MEDLINE searches for randomized controlled trials (RCTs) of the effects of mental health care Psychol Med 1994 24 741 748 7991756
Jadad AR McQuay HJ A high-yield strategy to identify randomized controlled trials for systematic reviews Online J Curr Clin Trials 1993 Doc No 33 3973
Wong SS Wilczynski NL Haynes RB Ramkissoonsingh R Developing optimal search strategies for detecting sound clinical prediction studies in MEDLINE AMIA Annu Symp Proc 2003 728 732 14728269
Bachmann LM Estermann P Kronenberg C Ter Riet G Identifying diagnostic accuracy studies in EMBASE J Med Libr Assoc 2003 91 341 346 12883560
Wilczynski NL Haynes RB Optimal search strategies for identifying diagnostic studies in EMBASE BMC Med
Watson RJ Richardson PH Identifying randomized controlled trials of cognitive therapy for depression: comparing the efficiency of Embase, Medline and PsycINFO bibliographic databases Br J Med Psychol 1999 72 535 542 10616135 10.1348/000711299160220
Haynes RB McKibbon KA Wilczynski NL Walter SD Werre SR for the Hedges Team Optimal search strategies for retrieving scientifically strong studies of treatment from MEDLINE: an analytic survey BMJ
Wilczynski NL McKibbon KA Haynes RB Enhancing retrieval of best evidence for health care from bibliographic databases: calibration of the hand search of the literature Medinfo 2001 10 390 393 11604770
Haynes RB Wilczynski NL Optimal search strategies for retrieving scientifically strong studies of diagnosis from Medline: analytical survey BMJ 2004 328 1040 15073027 10.1136/bmj.38068.557998.EE
| 15862125 | PMC1097733 | CC BY | 2021-01-04 23:52:12 | no | BMC Med Inform Decis Mak. 2005 Apr 29; 5:11 | utf-8 | BMC Med Inform Decis Mak | 2,005 | 10.1186/1472-6947-5-11 | oa_comm |
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-131584017710.1186/1471-2288-5-13Research ArticleEstimating the mean and variance from the median, range, and the size of a sample Hozo Stela Pudar [email protected] Benjamin [email protected] Iztok [email protected] Indiana University Northwest, Department of Mathematics, Gary, IN, 46408 USA2 Interdisciplinary Oncology Program, H. Lee Moffitt Cancer Center and Research Institute at the University of South Florida, Tampa, FL, USA2005 20 4 2005 5 13 13 26 9 2004 20 4 2005 Copyright © 2005 Hozo et al; licensee BioMed Central Ltd.2005Hozo 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
Usually the researchers performing meta-analysis of continuous outcomes from clinical trials need their mean value and the variance (or standard deviation) in order to pool data. However, sometimes the published reports of clinical trials only report the median, range and the size of the trial.
Methods
In this article we use simple and elementary inequalities and approximations in order to estimate the mean and the variance for such trials. Our estimation is distribution-free, i.e., it makes no assumption on the distribution of the underlying data.
Results
We found two simple formulas that estimate the mean using the values of the median (m), low and high end of the range (a and b, respectively), and n (the sample size). Using simulations, we show that median can be used to estimate mean when the sample size is larger than 25. For smaller samples our new formula, devised in this paper, should be used. We also estimated the variance of an unknown sample using the median, low and high end of the range, and the sample size. Our estimate is performing as the best estimate in our simulations for very small samples (n ≤ 15). For moderately sized samples (15 <n ≤ 70), our simulations show that the formula range/4 is the best estimator for the standard deviation (variance). For large samples (n > 70), the formula range/6 gives the best estimator for the standard deviation (variance).
We also include an illustrative example of the potential value of our method using reports from the Cochrane review on the role of erythropoietin in anemia due to malignancy.
Conclusion
Using these formulas, we hope to help meta-analysts use clinical trials in their analysis even when not all of the information is available and/or reported.
==== Body
Background
To perform meta-analysis of continuous data, the meta-analysts need the mean value and the variance (or standard deviation) in order to pool data. However, sometimes, the published reports of clinical trials only report the median, range and the size of the trial. In this article we use simple and elementary inequalities in order to estimate the mean and the variance for such trials. Our estimation is distribution-free, i.e., it makes no assumption on the distribution of the underlying data. In fact, the value of our approximation(s) is in giving a method for estimating the mean and the variance exactly when there is no indication of the underlying distribution of the data. In current practice, the median is often substituted for the mean, and the Range/4 or Range/6 for the standard deviation. However, it has not been shown that median can indeed be used to replace mean values, nor when the range-formulas are appropriate.
Methods
Assumptions
Suppose a clinical trial reports the following summary measures for a certain event:
m = Median
a = The smallest value (minimum)
b = The largest value (maximum)
n = The size of the sample.
In this article, we want to estimate the mean, and the standard deviation of this sample of size n. First we will order this sample by size:
a = x1 ≤ x2 ≤ x3 ≤ … xM-1 ≤ xM = m ≤ xM+1 ≤ … ≤ xn-1 ≤ xn = b,
where the Mth number is the median, and (for the sake of simplicity, we will assume that n is an odd number).
Results
Estimating the sample mean
We begin with several simple inequalities:
Adding up and diving by n, the middle column is exactly the sample mean, .
Adding up and diving by n for all three columns, we get the following inequality:
After replacing into the inequality above, we get:
Therefore, the lower bound for the sample mean is
The upper bound for the sample mean is
The sample mean can than be estimated as
, or
When the size of the sample is fairly large, the second fraction becomes negligible and the estimate can be written in a simplified form:
We can use this simple expression even if we do not know the size of the sample. The length of the interval which contains the sample mean (the interval [LB, UB]), is approximately
Estimating the sample variance
Even when the only information we have about a set of data is it's range: R = b - a, we can still estimate the standard deviation. If our data are normally distributed, then P[-2σ <X - μ < 2σ] = 0.95, and therefore, the range covers approximately 4σ, i.e., .
When the data we are dealing with are not normally distributed, we can still use the Chebyshev's inequality [1,2], and obtain the following for k = 3: . Therefore, the range covers approximately 6σ, i.e., .
On the other hand, if the summary results for a clinical trial include the median and the size of the sample, we can presumably do better than the two range approximations above. Next section deals with that situation.
The Variance S2 – distribution free inequalities
Using the inequalities (1) and taking in consideration that all the data are non-negative, we can multiply each row i with the value xi (i = 1, 2, 3, ..., n). We obtain the following inequalities:
Adding up by columns, we have the following:
Using the inequalities (1) again, we estimate the sums in LB and UB as
Therefore, the expressions in (7) can be estimated as
The sum of squares can be therefore estimated as
The sample variance can be evaluated from the computational formula
We can estimate using (10) and using (4). Therefore, after simplifying:
Note that if we let n grow without bound, the expression (12) becomes the well-known range formula .
The Variance S2 – equidistantly spaced data
The formula (4) can also be obtained by dividing the range [a, b] into two parts: [a, m), and [m, b]. We then subdivide each of these two parts into subintervals using equally spaced partition points. In other words, we are estimating each of the data points (except for a, m, and b) with uniformly spaced approximate points:
and
Therefore our sample is approximately given as
a = x1 ≤ x2 ≤ x3 ≤ … ≤ xM-1 ≤ y1 = m ≤ y2 ≤ … ≤ yM-1 ≤ yM = b.
We can use this partition to estimate the sample variance (and standard deviation, S). After a little algebra, the sample variance can be estimated by
If we let the number of estimation points increase without bounds, i.e., assume that n in the expression (15) is very large, we obtain a simplified version of the expression above:
Discussion
Analysis and performance of estimates
In order to verify the accuracy of these estimates, we ran several simulations using the computer package Maple where the data were variously distributed, and obtained the tables below.
We drew samples from five different distributions, Normal, Log-normal, Beta, Exponential and Weibull. The size of the sample ranged from 8 to about 100. In the first subsection we present the results of our estimation for a normal distribution, which is what meta-analysts would commonly assume. We also show the results of simulations where the data were selected from a skewed distributions. In each case we compared the relative error made by estimating the sample mean with the approximation given by formulas (4) and (5), as well as by the median, and the relative error made by estimating the sample variance by the formulas (12) and (16), as well as the well-known standard deviation estimators Range/4 and Range/6.
Normal distribution
We drew 200 random samples of sizes ranging from 8 to 100 from a Normal Distribution with a population mean 50 and standard deviation 17. Then we graphed the average relative error vs. the sample size. Both estimators for the mean, formulas (4) and (5), are very close to the sample mean (within 4%). For sample sizes smaller than 29, formula (5) is actually outperforming the median as a mean estimator. For larger sample sizes, however, the median is more consistent estimator for a normally distributed sample.
The variance estimators however show greater distinction. For a very small sample size (up to 15) the formula (16) is performing the best (within 10% of the real sample standard deviation). When the sample size is between 16 and 70, the formula Range/4 is the best estimator of the sample standard deviation, with a relative error between 10–15%. However, for larger sample sizes, the formula Range/6 performs the best for this distribution. To compare the precision of these estimates on average, we collected the results of our simulation in the Additional file 1.
Simulation with a skewed distribution (Log-Normal, Beta, Exponential and Weibull)
We also decided to run a simulation where the algorithm selects a sample from a skewed distribution. We decided to use Log-Normal distribution with parameters μ = 4, and σ = 0.3, Beta distribution with parameters a = 9 and b = 4, Exponential distribution with the parameter λ = 10 and Weibull distribution a = 2 and b = 35. These parameters were chosen arbitrarily, and the simulation results did not differ when we used different parameters (naturally, larger variance translates into larger relative error for mean estimators for any distribution). Just like in the case of Normal distribution, we ran our algorithm 200 times for each sample size ranging from 8 to 100. For each of the estimation formulas we then calculated the average relative error. We will summarize the best formula for estimation in Table 1.
Table 1 The best formula for estimation by distribution.
Best Formula for Sample size (n) Mean Estimation Standard Deviation Estimation
Formula (5) Median Formula (16) Range/4 Range/6
Log-Normal n ≤ 23 23 <n n ≤ 15 15 <n ≤ 64 64 <n
Beta n ≤ 30 30 <n n ≤ 15 15 <n ≤ 100 100 <n
Exponential n ≤ 21 21 <n n ≤ 15 15 <n ≤ 66 66 <n
Weibull n ≤ 25 25 <n n ≤ 16 16 <n ≤ 110 110 <n
Therefore, counter intuitively, even for the skewed distributions we tested, it seems like that for a larger sample size (usually more than 25) simply replacing sample mean with the reported median is the best estimate of the sample mean. This is an interesting result and we are not aware that it was previously demonstrated. It gives assurance to meta-analysts that simple replacement of mean with medians in meta-analysis is a viable option. Formula (5), even though taking more parameters into account (the range and the sample size), on average only outperforms the median for small sample sizes. However, a large number of trials used in meta-analyses do have very small number of patients for each arm (as small as 10–15). For these trials, formula (5) seems to give an alternative to just using the median.
When estimating the standard deviation, formula (16) is the best estimate for very small sample sizes (less than 16), after which the range formulas (Range/4 and Range/6) are better. Range/4 formula works best for samples of moderate size (between 16 and about 70), while for really large samples, Range/6 is the best estimator.
Detailed results of each simulation with a skewed distribution are given in the Additional file 2, Additional file 3, Additional file 4, and Additional file 5.
If the reader wants to try these formulas with a different set of data, we have provided an Excel spreadsheet file with the formulas at
Effect on the mean difference in meta-analysis
In this section we will discuss the use of these estimating formulas on the effect size for the meta-analysts. When pooling the means from various sources for a meta-analysis, the usual procedure is to calculate differences in the means between the experimental arm of a study and the control arm, m p = mc - me , and the combined variance for each study, (for example, see [3]). The pooled mean difference is then calculated by using weighted sum of these differences, where the weight is the reciprocal of the combined variance for each study.
To determine whether our estimates make a huge difference when compared to the actual mean difference and variance, we drew two samples of the same size from a same distribution. We applied our methods to the Log-Normal [4, 0.3] distribution since this skewed distribution is frequently encountered in biology and medicine.
First we ran a test-case meta-analysis. After drawing fifteen samples of random sizes (between 8 and 100) from our distribution, we used our estimation formulas to estimate the mean and the variance from the median and the range. Then we performed meta-analysis using STATA, treating the samples as one subgroup and their estimates as another subgroup to determine the pooled means and heterogeneity. Our results for the weighted mean difference, WMD (see Figure 1) are presented in Table 2.
Figure 1 Meta-Analysis of random data. After drawing fifteen samples of random sizes (between 8 and 100) from the Log-Normal [4, 0.3] distribution, we used our estimation formulas to estimate the mean and the variance from the median and the range. Then we performed meta-analysis using STATA, treating the real samples as one subgroup and their estimates as another subgroup to determine the results and heterogeneity.
Table 2 Results of our meta-analysis with the real sample data as one subgroup, and our estimates of the sample as the second subgroup.
Actual Sample Our Estimate
WMD [95% CI] % Weight WMD [95% CI] % Weight
Pooled WMD -0.37 [-37.17, 36.44] 42.00 0.41 [-30.92, 31.73] 58.00
Overall pooled WMD 0.08 100.00
Heterogeneity statistic degrees of freedom P I-squared
Sample 0.04 9 1.000 0.0%
Estimate 0.04 9 1.000 0.0%
Overall 0.08 19 1.000 0.0%
Overall Test for heterogeneity between sub-groups
0.00 1 0.975
Significance test(s) of WMD = 0 Sample z = 0.02 p = 0.984
Estimate z = 0.03 p = 0.980
Overall z = 0.01 p = 0.995
1I-squared: the variation in WMD attributable to heterogeneity
In order to capture a more consistent measure of the effect of our estimation on pooled mean difference, we repeated this process by varying the number of trials in the meta-analysis from 8 to 100. In particular we are interested in the difference between the real pooled weighted mean difference in the sample group and the pooled weighted mean difference from a meta-analysis using estimated means and variances.
The actual population mean from which we drew samples is 57.11 and the standard deviation is 17.53 (Log-Normal [4, 0.3]). The actual average pooled sample mean difference between two samples (one was control, the other experimental group) was 0.031. Using the medians and range, we estimated the means for each sample, and performed the meta-analysis using these estimates. The average pooled (estimated) mean difference was 0.002, making the difference between the two methods 0.029 (on average). Individually, the pooled means (both, the real sample pooled means, and the estimated pooled means) differed a little more. In Figure 2 the black diamonds represent the actual pooled mean difference using actual sample means. The red circles represent the same pooled mean differences using our estimation formulas (we connected the corresponding symbols for clarity). The horizontal axis represents the number of trials in the meta-analysis (from 8 to 100).
Figure 2 Actual pooled mean difference and estimated pooled mean difference. The black diamonds represent the actual pooled mean difference using sample means. The red circles represent the pooled mean differences for the same samples using our estimation formulas (we connected the corresponding symbols for clarity). The horizontal axis represents the number of trials in the meta-analysis (from 8 to 100).
As seen from the Figure 2, the estimates of the mean were fairly accurate and useful. On the other hand, the estimates for the variance were a lot less precise, missing the actual value of the variance by 10 % – 20% (see the Additional Files 1, 2, 3, 4, 5). However, in some situations, using these estimates might still be better than the alternative – excluding the trials which reported the wrong summary data (median instead of mean). Using our estimation method, we can see the effect of such trials on pooled summary measures. In the next section we will illustrate our method in an actual systematic review.
An illustrative example of the potential value of our methods
American Society of Hematology/ American Society of Oncology (ASH/ASCO) developed practice guidelines for the use of erythropoietin (Epo), a drug whose annuals sales exceed several billions of dollars in the US alone, based on the systematic review of the effects of Epo on various clinical outcomes of interest including improvement of anemia by increase of hemoglobin[4]. The results were expressed as the mean increase in hemoglobin in Epo arm compared with the control. However, a number of the papers reported median increase instead of mean increase and standard deviation. Due to lack of available methods to use median values, the authors of this important review, decided not to use these papers in their meta-analysis. Recently, the Cochrane review was published attempting to provide more updated analysis of the effects of Epo in anemia related to malignancy [5]. The Cochrane reviewers did meta-analyze data to calculate an average weighted mean increase in hemoglobin as the result of Epo treatment. However, the Cochrane investigators could not include the totality of evidence in relation to this outcome since a number of the trials reported data as medians instead of means. Therefore, published meta-analyses related to the effect of Epo in anemia due to malignancy suffer from the phenomena akin to the outcome reporting bias [6] simply due to fact that methods are not yet developed to allow researches to use data medians.
Here we illustrate that it is actually possible to use medians and pool, and improve inclusiveness of meta-analyses. For example, the Cochrane investigators were only able to pool 2 studies [7,8] to evaluate the effect of Epo on change in hemoglobin in the patients with the baseline level of hemoglobin >12 g/dl who underwent chemotherapy. Their results show that on average Epo increases hemoglobin by 2.05 g/dl. However, the Cochrane investigators could not pool data from other available studies in the literature with similar eligibility. ASH/ASCO guidelines listed two other studies that were eligible for the meta-analysis (and two that were not).
For the first of these studies [9], by Welch at al, the ASH/ASCO guidelines paper reported the mean hemoglobin change for each of the two arms, the experimental and the control. However, they did not report the data for the standard deviation of these means. Since the size of each arm is 15 patients, our formula (16) provides the best estimate of the standard deviation using the median and the range. We used Figure 1 on page 263 in Welch at al. [9] to estimate the range of the hemoglobin change for each arm and used formula (16) to determine the standard deviation. The ASH/ASCO guidelines paper also reported the difference in medians of hemoglobin response for the largest study eligible for the meta-analysis conducted by Thatcher at al [10]. Thatcher et al do report in their paper ranges of hemoglobin for patients treated by Epo and control. This trial was a three-arm study, in which two doses of Epo were compared against the control. For the purpose of this analysis, we separated the data from each of the Epo arms and compared them against one half of the control group (just like the rest of the studies in the Cochrane review). Using the methods described here, we were able to estimate mean increase (using formula (5)) and standard deviation (using Range/4 formula in both comparisons). When we incorporated these results into the Cochrane meta-analysis, we found that the effect of Epo on mean increase in hemoglobin significantly changed: the pooled estimate decreased from an average of 2.05 g/dl in hemoglobin increase to 1.22 g/dl, i.e., a decrease of approximately 40% (see Figure 3)!
Figure 3 An example: Meta Analysis with all eligible trials included. Cochrane investigators [5] were only able to pool two studies to evaluate the effect of Epo on change in hemoglobin in the patients with the baseline level of hemoglobin >12 g/dl who underwent chemotherapy. Their results show that on average Epo increases hemoglobin by 2.05 g/dl. Using our estimation formulas, we were able to include two other studies eligible for this meta-analysis ([9, 10]). The pooled estimate decreased to 1.22 g/dl, i.e., a decrease of approximately 40%.
Our estimates come with some uncertainty. To see what effect this uncertainty has on the outcome of our meta-analysis, we varied the estimated means in Thatcher at al by 4% and the estimated standard deviation in both, Thatcher at al and Welch at al, by 10% to 15% (according to sample sizes, as indicated in the Additional Files 1, 2, 3, 4, 5). The summary pooled estimate now ranged from the low of 1.09 to the high of 1.32, which represents a decrease between 36% and 47%.
This example outlines how our method can be potentially useful for meta-analysts. It is important to realize that this example is provided only to illustrate our method. Our goal here is not to challenge the Cochrane review or ASH/ASCO guidelines. Nevertheless, we believe that this example is a good illustration of the potential of our method. While it is common practice that the investigators simply pool what is available to them it is actually not known how often studies are excluded because of reporting a different summary statistic. In future we will attempt to systematically address this issue and evaluate, for example, how often the Cochrane reviews did not pool data from the available median values when they pooled data on continuous outcomes. We hope that availability of our methods to the wider meta-analytic audience may further improve the inclusiveness of all relevant studies for the Cochrane and other meta-analyses.
Conclusion
We found that a simple formula (5): can be used to estimate the mean using the values of the median (m), low and high end of the range (a and b , respectively).
Using simulation methods we were able to determine that formula (5) is a best estimator for the mean when dealing with a small sample size. As soon as sample size exceeds 25, the median itself is the best estimator.
The variance can be estimated using the formula (16)
Together with the well-known estimators (Range/4 for a normal distribution, and Range/6 for any random distribution) this formula provides a useful tool for meta-analysts. Using simulations, we determined that for very small samples (up to 15) the best estimator for the variance is the formula (16). When the sample size increases, Range/4 is the best estimator for the standard deviation (and variance) until the sample sizes reach about 70. For large samples (size more than 70) Range/6 is actually the best estimator for the standard deviation (and variance).
In summary, the best estimators for the mean and the standard deviation for different sample sizes are given in Table 3.
Table 3 The best estimating formula for an unknown distribution.
Sample Size: n ≤ 15 15 <n ≤ 25 25 <n ≤ 70 70 <n
Estimate Mean Formula (5) Median
Estimate Standard Deviation Formula (16) Range/4 Range/6
Using these formulas, we hope to enable meta-analysts use clinical trials even when not all of the information is available and/or reported.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SPH developed most of the formulas for estimation. IH conducted the simulations using Maple. BD conceived of the problem, participated in its framing and coordination, and helped SPH and IH draft the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Normal distribution. The top row of the table displays the results of estimating the mean, while the second row displays the results of estimating the standard deviation. Each number in this table represents the average relative error of 200 samples from a Normal distribution.
Click here for file
Additional File 2
Log-Normal distribution. The top row of the table displays the results of estimating the mean, while the second row displays the results of estimating the standard deviation. Each number in this table represents the average relative error of 200 samples from a Log-Normal distribution.
Click here for file
Additional File 3
Beta Distribution. The top row of the table displays the results of estimating the mean, while the second row displays the results of estimating the standard deviation. Each number in this table represents the average relative error of 200 samples from a Beta distribution.
Click here for file
Additional File 4
Exponential Distribution. The top row of the table displays the results of estimating the mean, while the second row displays the results of estimating the standard deviation. Each number in this table represents the average relative error of 200 samples from a Exponential distribution.
Click here for file
Additional File 5
Weibull Distribution. The top row of the table displays the results of estimating the mean, while the second row displays the results of estimating the standard deviation. Each number in this table represents the average relative error of 200 samples from a Weibull distribution.
Click here for file
==== Refs
Hogg RV Craig AT Introduction to mathematical statistics 1995 5th New York
Toronto , Macmillan College Pub. Co. ;
Maxwell Macmillan Canada ;
Maxwell Macmillan International xi, 564
Mood AMF Graybill FA Boes DC Introduction to the theory of statistics 1974 3d New York, , McGraw-Hill xvi, 564
Petiti DB Meta-analysis, decision analysis and cost-effectiveness analysis. Methods for quantitative synthesis in medicine. 2nd ed. 2000 New York , Oxford press
Rizzo JD Lichtin AE Woolf SH Seidenfeld J Bennett CL Cella D Djulbegovic B Goode MJ Jakubowski AA Lee SJ Miller CB Rarick MU Regan DH Browman GP Gordon MS Use of epoetin in patients with cancer: evidence-based clinical practice guidelines of the American Society of Clinical Oncology and the American Society of Hematology J Clin Oncol 2002 20 4083 4107 12351606 10.1200/JCO.2002.07.177
Bohlius J Langensiepen S Schwarzer G Seidenfeld J Piper M Bennet C Engert A Erythropoietin for patients with malignant disease. Cochrane Database Syst Rev 2004 CD003407. 15266483
Chan AW Hrobjartsson A Haahr MT Gotzsche PC Altman DG Empirical Evidence for Selective Reporting of Outcomes in Randomized Trials: Comparison of Protocols to Published Articles JAMA 2004 291 2457 2465 15161896 10.1001/jama.291.20.2457
Del Mastro L Venturini M Lionetto R Garrone O Melioli G Pasquetti W Sertoli MR Bertelli G Canavese G Costantini M Rosso R Randomized phase III trial evaluating the role of erythropoietin in the prevention of chemotherapy-induced anemia J Clin Oncol 1997 15 2715 2721 9215845
Kunikane H Watanabe K Fukuoka M Saijo N Furuse K Ikegami H Ariyoshi Y Kishimoto S Double-blind randomized control trial of the effect of recombinant human erythropoietin on chemotherapy-induced anemia in patients with non-small cell lung cancer Int J Clin Oncol 2001 6 296 301 11828949 10.1007/s10147-001-8031-y
Welch RS James RD Wilkinson PM Fb Recombinant Human Erythropoietin and Platinum-Based Chemotherapy In Advanced Ovarian Cancer Cancer J Sci Am 1995 1 261 9166486
Thatcher N De Campos ES Bell DR Steward WP Varghese G Morant R Vansteenkiste JF Rosso R Ewers SB Sundal E Schatzmann E H. S Epoetin alpha prevents anaemia and reduces transfusion requirements in patients undergoing primarily platinum-based chemotherapy for small cell lung cancer. Br J Cancer 1999 80 396 402 10408844 10.1038/sj.bjc.6990369
| 15840177 | PMC1097734 | CC BY | 2021-01-04 16:32:51 | no | BMC Med Res Methodol. 2005 Apr 20; 5:13 | utf-8 | BMC Med Res Methodol | 2,005 | 10.1186/1471-2288-5-13 | oa_comm |
==== Front
BMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 1471-2288-5-141585048510.1186/1471-2288-5-14Research ArticleThe strengths and limitations of meta-analyses based on aggregate data Lyman Gary H [email protected] Nicole M [email protected] Department of Medicine University of Rochester Medical Center Rochester, New York 14642 USA2005 25 4 2005 5 14 14 3 1 2005 25 4 2005 Copyright © 2005 Lyman and Kuderer; licensee BioMed Central Ltd.2005Lyman and Kuderer; 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
Properly performed systematic reviews and meta-analyses are thought by many to represent among the highest level of evidence addressing important clinical issues. Few would disagree that meta-analyses based on individual patient data (IPD) offer several advantages and represent the standard to which all other systematic reviews should be compared.
Methods
All cancer-related meta-analyses cited in Medline were classified as based on aggregate or individual patient data. A review was then undertaken of all reports comparing the comparative strengths and limitations of meta-analyses using either aggregate or individual patient data.
Results
The majority of published meta-analyses are based on summary or aggregate patient data (APD). Reasons suggested for this include the considerable resources, years of study and often, broad international cooperation required for IPD meta-analyses. Many of the most important features of systematic reviews including formal meta-analyses are addressed by both IPD and APD meta-analyses. The need for defining an explicit and relevant clinical question, exhaustively searching for the totality of evidence, meticulous and unbiased data transfer or extraction, assessment of between study heterogeneity and the use of appropriate statistical methods for estimating summary effect measures are essentially the same for the two approaches.
Conclusion
IPD offers advantages and, when feasible, should be considered the best opportunity to summarize the results of multiple studies. However, the resources, time and cooperation required for such studies will continue to limit their use in many important areas of clinical medicine which can be meaningfully and cost-effectively approached by properly performed APD meta-analyses. APD meta-analyses continue to be the mainstay of systematic reviews utilized by the US Preventive Services Task Force, the Cochrane Collaboration and many professional societies to support clinical practice guidelines.
==== Body
Background
A meta-analysis strives to obtain data on all patients in all relevant trials addressing an explicit and relevant clinical question. Meta-analyses most often attempt to achieve these goals by collecting aggregate patient data (APD) from completed studies that have been published in the medical literature, presented at professional meetings or directly provided by individual investigators. Few would argue that properly conducted meta-analyses based on individual patient data (IPD) have several advantages and represent the standard against which other meta-analyses should be considered. While the relative advantages and disadvantages of meta-analyses based on either APD or IPD have been debated in the medical literature, the majority of published meta-analyses continue to be based on APD. In addition, APD meta-analyses continue to be the mainstay of systematic reviews conducted by the US Preventive Services Task Force, the Cochrane Collaboration and many professional societies. A systematic search for IPD meta-analyses related to cancer in 2000 confirmed 38 of which 8 were unpublished [1]. A more recent search of Medline by the authors identified 1,595 reported meta-analyses related to cancer of which 76 (4.4%) were apparently based on IPD. The annual proportion of published cancer meta-analyses based on IPD has not changed significantly over the last several years (Figure 1). A recent review of IPD meta-analyses to evaluate diagnostic tests found no published reports of IPD meta-analyses in diagnostic research [2]. That the majority of published systematic reviews are based on APD meta-analyses suggests that they are not only more frequently completed but may also be considered relevant and valid to editors, reviewers and readers. Given the virtually infinite number of clinical questions that might be addressed in systematic reviews, it is important and timely to review the methodological differences and the relative advantages and disadvantages of these two types of meta-analysis.
Figure 1 Bar graph representing the number of published meta-analyses related to cancer and cited in MEDLINE over the past fifteen years. Shown are the total number of reported meta-analyses and the proportion of meta-analyses each year indicating an individual patient data meta-analyses. Note that data for 2004 is incomplete.
Methods
Many of the reasons why meta-analyses are generally considered high-level evidence pertain to both IPD and APD. In the case of both APD and IPD meta-analyses, a written study protocol should be generated pre-specifying the search process, inclusion and exclusion criteria and the hypotheses to be tested. Criteria for study inclusion and exclusion should be defined in advance and uniformly applied. In both situations, an exhaustive search for relevant studies should be conducted. In addition, careful data collection based on dual data extraction and entry should be employed. Both the primary and secondary outcomes of interest should be specified in advance of data extraction or analyses. The results from individual studies are then systematically analyzed first by assessing for heterogeneity and, if appropriate, combining of results providing summary estimates of the treatment effect. Secondary analyses then are often performed to explore the reasons for any heterogeneity. Publication bias represents an important limitation of any review and retrieval of data from all relevant studies should be the goal in order to avoid publication bias. Cochrane reviews attempt to identify all relevant studies, published and unpublished, and this should be the goal of any systematic review based on either IPD or APD. An acknowledged limitation of IPD is the need to exclude studies for which data are not available due to time, willingness or proprietary interest. Therefore, before abandoning APD meta-analyses, it is important to look more closely at the pros and cons of each approach.
Results
The strengths and limitations of APD meta-analyses
The purpose of a meta-analysis is to systematically review the results of previous research in order to derive valid conclusions concerning totality of evidence on a subject. Both IPD and APD meta-analyses attempt to avoid the potential bias of narrative literature reviews, which are selective in the studies included and subjective in the weighting of the studies included. Each is considered useful for summarizing the results of multiple individual studies that are each too small to provide valid results. Pooled analyses of APD is conceptually the same as meta-analyses of separate studies based on IPD including estimating study-specific treatment effects, assessing heterogeneity, estimating a summary effect size and evaluation of heterogeneity.
Strengths and limitations of IPD meta-analyses
Several authors have discussed the strengths and potential advantages of IPD meta-analyses [3,4]. Table 1 summarizes many of the suggested advantages of IPD meta-analyses. Authors have gone so far as to suggest that meta-analyses based on APD provide little reliable information and should be viewed as only exploratory [5]. Many of the stated advantages of IPD meta-analyses are not inherent to the analysis but rather represent epiphenomena that often accompany the enormous resources, time and energy devoted to IPD meta-analysis. As shown in Table 2, many of the specific steps provide such systematic reviews with their procedural superiority over narrative reviews are shared by IPD and APD analyses. It is also clear that in addition to considerable cost and years of effort, IPD meta-analyses often require the extraordinary cooperation of all original investigators, institutions, organizations, companies and investigational review boards to avoid publication bias that may challenge the validity of the analysis since the missing studies are seldom missing completely at random. Such studies also offer the strong temptation to conduct unplanned subgroup analyses for which the original studies were neither designed nor powered based on patient-specific data. It is essential to avoid the temptation to combine all patients as if they came from a single very large clinical trial. Most IPD meta-analyses use essentially the same summary and statistical measures employed for analysis of aggregate data. Clearly IPD meta-analyses often obtain updated data and perform checks on the original data when possible. However, again, these strategies are not inherent to the IPD process but reflect the additional resources and effort that is often employed in such analyses. Additional issues to consider in contrasting these different approaches are summarized below:
Table 1 Proposed advantages of individual patient data meta-analyses
Ability to use common definitions, coding and cutpoints
Address questions not addressed in original publication
Assess adequacy of randomization
Permits data checking
Permits data updating
Permits checking of analyses
Allows adjustment for the same variables across studies
Permits ready use of time-to-event data for estimating survival
Ability to address long-term outcomes
Facilitates exploration of heterogeneity at the patient level and subgroup analyses of patient level data
Table 2 Comparison of IPD and APD Meta-Analyses
Steps in Meta-Analysis IPD APD
Explicit and Relevant Clinical Question √ √
Exhaustive Search All published studies √ √
All presented studies √ √
All completed studies √ ±
Screening: inclusion/exclusion criteria √ √
Data Acquisition (extraction/transfer) Aggregate data √ √
Individual patient data √ -
Data Checking Source data ± -
Submitted data √ √
Published/presented √ √
Data Updating √ ±
Missing Studies/Data ± ±
Uniform outcomes √ -
Tests for heterogeneity √ √
Estimating Summary Effect Measures Binomial data √ √
Time-to-event data √ ±
Exploring Heterogeneity Subgroup analyses Study Level √ √
Patient Level √ -
a) Data access and checking
Gaining access to IPD in an era of increasing concern about confidentiality and greater oversight is not a trivial undertaking. In addition, access to IPD does not guarantee that the data collection was properly conducted, that the randomization process was appropriate or that the same data items were actually collected. Rarely does an IPD meta-analysis provide access to the source data such as patients, medical records, laboratory results etc. Rather, access is provided to data extracted at the point of care on each patient and the true accuracy of the APD is often not verified. It is argued that an advantage of IPD is the ability to check the data reported in the published trial. While data checking is sometimes considered, it is very costly and time consuming process and rarely undertaken. When checked data has been compared to unchecked data, differences in estimated outcomes are rare [6]. Even when all data are available rather than only summary data, analyses are generally based on meta-analysis estimators of treatment effect as in APD meta-analyses. Several studies have demonstrated that when the pool of studies is the same and similar measures are utilized, the effect size estimates for appropriate procedures are very similar for IPD and APD meta-analysis [7,8].
b) Updated data
Another possible advantage to IPD meta-analysis is the ability to update data from a previous publication providing longer follow-up with a greater number of events. Although updated data is more likely to become available with the extended resources and collaboration of individual investigators, it must be noted that updating of previously published data is not inherent nor confined to IPD meta-analysis. It has been pointed out, in fact, that unpublished data and late-appearing data may be different from early-appearing data [9]. Updated data available after the completion of the main study may be affected by crossover, missing information and unblinding. Using data from a study of the effect of high-dose acyclovir on the survival of patients with HIV, the authors found that APD and IPD lead to the same effect estimates. They conclude that discrepant results probably arise either by publication bias or retrieval bias in the IPD analyses or the inclusion of updated information that differentiates the databases used by the two methods but is not inherent to or exclusive of either.
c) Data accuracy and validity
Some authors have concluded that the results of IPD meta-analyses are more accurate and unbiased than those based on APD [11,12]. However, such reports often equate literature-based meta-analyses and those based on APD. Comparisons based on meta-analyses limited to published reports may be inherently biased if the IPD analysis includes the results of unpublished studies as was done in the above reports. Unpublished studies are often unpublished due to low observed treatment effect either leading the investigators not to pursue publication or editorial bias against negative study results. Stewart and Palmar [11] compared the results with the unpublished studies included and found a small but persistent difference between the effect estimates of the two approaches. It is likely that much of the remaining difference related to the ability of IPD investigators to obtain updated results compared to literature-based analyses where no such effort was made. It should be noted that there is no inherent obstacle to including aggregate data from unpublished results if they have been reported in abstract form, presented at major meetings or are willingly provided as summary measures by the investigators. Likewise, there is no barrier other than needed resources, time and cost to requesting updated summary measure from authors of published studies. Therefore, the favorable findings often reported for IPD meta-analyses may relate to the inclusion of additional, unpublished studies with low treatment effect and the addition of updated data from individual investigators. There are many unresolved issues related to the need for informed consent and HIPPA compliance for analyses beyond those planned in the original study particularly when based on IPD.
d) Analysis checking
While access to IPD may provide an opportunity to redo the actual analysis, there is little evidence that incorrect analyses of randomized controlled trials are frequent or that such reanalyses are likely to alter the conclusions of a systematic review. Of greater concern for potential bias is the design and conduct of the individual studies and with rare exception there is little opportunity to address these problems with IPD or aggregate meta-analyses. A poorly designed or conducted trial is just as likely to bias IPD as it is aggregate data. Divergent results in a meta-analysis often lead investigators to draw conclusions based on subgroups of subjects or studies. This problem may, in fact, represent a greater temptation in IPD meta-analyses due to ready access to individual patient characteristics. Oxman et al have argued that this is potentially dangerous due to the risk of being misled by both systematic error (bias) and random error (chance) arguing that it is far safer to base clinical decisions on a critical summary of all available evidence rather than on a subset of studies or patients [13].
e) Survival data
Time-to-event or survival data seems particularly suited to IPD meta-analyses as there is often access to the actual survival time for individual patients. Duchateau et al found significant differences between a IPD meta-analysis of chemotherapy for head and neck cancer and a literature-based APD when the later analysis is based on mortality at a specific time [14]. Time-to-event analyses or estimates of the actual survival function are more powerful than estimates based on the limited number of time points generally available with aggregate data. Parmar et al have proposed better methods for extracting summary statistics to perform meta-analyses of the published literature for survival endpoints [15]. They appropriately maintain that when reporting a randomized controlled trial with survival type data, that the most appropriate summary statistics are the log hazard ratio and its variance, which are particularly designed for comparing two survival curves by allowing for both censoring and time to an event. If the time to an event and censoring are ignored, the log hazard function becomes simply the log relative risk. The hazard ratio is a global summary of the difference between two survival curves and represents the total reduction in the risk of death with treatment compared to controls over the entire period of follow-up. Parmar et al point out that the hazard ratio is most easily interpreted when the hazards are proportional but is still valid and useful when they are not. These summary statistics can be used to perform a stratified analysis to combine results from each trial in a meta-analysis. The overall log hazard ratio is a weighted average of the log hazard ratios of each study where the weights are inversely proportion to the variance of the log hazard ratio for each trial. They note that the log hazard ratio and its variance can sometimes be estimated directly from reported trial results. The authors go on to discuss several indirect methods for estimating the log hazard ratio and its variance either from summary trial results or the published survival curves. The authors studied 209 randomized controlled trials comparing the survival of women treated for advanced breast cancer contrasting the estimates of the log hazard ratio directly or indirectly taken from the manuscript with those derived from survival curves. Among the three-fourths of the studies providing some summary data, the survival curve estimate of the log hazard ratio was nearly identical to that reported directly in the manuscript. There was no evidence of a systematic bias although the survival curve estimate tended to underestimate the treatment effect provided directly from the papers. Several additional techniques have been proposed for combining survival curves from APD. Earle et al examined the accuracy of these techniques from studies of patients treated with chemotherapy for advanced non-small cell lung cancer and compared each method's summary curve with that generated by the corresponding IPD meta-analysis [16]. The authors found that all methods were able to accurately reproduce summary survival curves statistically similar to the IPD-derived curves with maximum discrepancies ranging from 1.8% to 4.7%. The optimal method was found to depend upon the characteristics of the data and the purpose of the analysis. In addition to having a role in providing summary data when resources or time are limited or when IPD is not available, it has been proposed that APD meta-analyses of time-to-event studies should be performed to determine whether it would be worthwhile proceeding with the more resource-dependent IPD meta-analysis [17].
f) Exploring heterogeneity
The major advantage of IPD as opposed to APD meta-analysis is the ability to study the impact of individual patient level characteristics. It is important to note, however, that such analyses are often not prespecified and are therefore, by definition, secondary, exploratory or hypothesis generating in nature. There is little debate that the exploration of patient-level characteristics is best undertaken with patient-level data. The use of averages or proportions of patient characteristics in trials may lead to the common ecological bias, often underestimating the influence of such characteristics [18]. On the other hand, IPD offers no inherent advantage in the exploration of study level features such as study design characteristics. It must be remembered, in either case, such analyses are generally not pre-specified in either the meta-analysis or in the individual trials, which were generally underpowered to address subgroup evaluations. Such statistical limitations in subgroup or meta-regression analyses are equally applicable to IPD and APD meta-analyses.
Equivalence of meta-analysis using APD and IPD
Olkin and Sampson have shown that summary estimates obtained from a meta-analysis of APD are essentially equivalent to the least squares estimate of IPD computed from a two-way fixed-effects model without interaction where the effects in the model are those due to treatment and due to different studies, respectively [8]. Therefore, as long as the same set of studies are used for both, there appears to be no difference between a meta-analysis of the summary effect estimates obtained from each study and that obtained by pooling the original patient data. While Olkin and Sampson demonstrated this somewhat surprising result when the observations are independent within and across studies based on a common variance, Mathew and Nordstrom confirmed these findings in a much more general setting where the observations within a study are not necessarily independent and the observations across studies can have different covariance matrices [19]. Several investigators have attempted to compare the results of APD meta-analyses often based on published results to those of IPD. Needless to say, different investigators have reached different conclusions from these comparisons. A recent study by Angelillo and Villari contrasted a meta-analysis of APD based on published studies of the perinatal transmission rate with Cesarean section in HIV-positive women to a previously reported IPD based meta-analysis [20]. The two meta-analytic methods were found to yield very similar results although no formal comparison was made.
Discussion
Many of the most important and valued features of systematic reviews and formal meta-analyses in general are addressed by both IPD and APD meta-analyses. While IPD studies may more often obtain unpublished data and provide opportunity for data checking and updating, such features are not inherent to IPD meta-analyses but are largely attributable to the great resources and time devoted to such studies. Failure to obtain data on all patients and from all trials may lead to an acquisition bias since the missing studies or patients may not be missing completely at random. Clearly, IPD is advantageous when different outcomes or cutpoints are reported in the APD. Alternatively, when based on the same studies, summary effect measures based on IPD and APD meta-analyses are virtually identical. Survival data would appear to be one area where IPD meta-analyses have a clear advantage. However, several techniques have been developed and validated which provide estimates of survival outcomes with APD that are similar to those derived from IPD. APD meta-analyses of time-to-event studies may inform investigators as to whether it would be worthwhile proceeding with the more resource-intensive IPD meta-analysis. While both approaches permit exploration of study and summary patient sources of heterogeneity, only IPD permits full exploration of and adjustment for patient characteristics. It is important to remember, that such analyses are only exploratory and hypothesis generating. It is important to avoid the temptation to analyze IPD without consideration of the separate data sources and the secondary nature of such analyses. IPD offers advantages and, when feasible, should be considered the best opportunity for summarizing the results of multiple studies. However, the resources, time and cooperation required for such studies will continue to limit their use in many important areas of clinical medicine which can be meaningfully and cost-effectively approached by properly performed APD meta-analyses.
Abbreviations
IPD: individual patient data
APD: aggregate patient data
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
Both authors contributed to the study design, data collection and analysis and authorship of this report
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Tierney JF Clarke M Stewart LA Is there bias in the publication of individual patient data meta-analyses? Int J Technol Assess Health Care 2000 16 657 667 10932430 10.1017/S0266462300101217
Khan KS Bachmann LM Gerben ter Riet Systematic reviews with individual patient data meta-analysis to evaluate diagnostic tests Eur J Obstet Gynecol Repro Biol 2003 108 121 125
Stewart LA Tierney JF To IPD or Not to IPD? Advantages and Disadvantages of Systematic Reviews Using Individual Patient Data Eval Health Prof 2002 25 76 97 11868447 10.1177/0163278702025001006
Sylvester R Collette L Duchateau L The role of meta-analyses in assessing cancer treatments Eur J Cancer 2000 36 1351 1358 10899647 10.1016/S0959-8049(00)00125-8
Piedbois P Buyce M Meta-analyses based on abstracted data: a step in the right direction, but only a first step J Clin Oncol 2004 22 3839 3841 15326196 10.1200/JCO.2004.06.924
Burdett S Stewart LA A Comparison of the Results of Checked Versus Unchecked Individual Patient Data Meta-Analyses Int J Technol Assess Health Care 2002 18 619 624 12391954
Steinberg KK Smith SJ Stroup DF Olkin I Lee NC Williamson GD Thacker SB A comparison of effect estimates from a meta-analysis using individual patient data for ovarian cancer studies Am J Epidemiology 1997 145 917 925
Olkin I Sampson A Comparison of Meta-Analysis Versus Analysis of Variance of Individual Patient Data Biometrics 1998 54 317 322 9544524
Ioannidis JPA Contopoulos-Ioannidis DG Lau J Recursive Cumulative Meta-analysis: A Diagnostic for the Evolution of Total Randomized Evidence from Group and Individual Patient Data J Clin Epidemiol 1999 52 281 291 10235168 10.1016/S0895-4356(98)00159-0
Trikalinos TA Ioannidis JP Predictive modeling and heterogeneity of baseline risk in meta-analysis of individual patient data J Clin Epidemiol 2001 54 245 52 11223322 10.1016/S0895-4356(00)00311-5
Stewart LA Parmar MKB Meta-analysis of the literature or of individual patient data: is there a difference? The Lancet 1993 341 418 422 8094183 10.1016/0140-6736(93)93004-K
Jeng GT Scott JR Burmeister LF A comparison of meta-analytic results using literature versus individual patient data JAMA 1995 274 830 836 7650808 10.1001/jama.274.10.830
Oxman AD Clarke MJ Stewart LA From Science to Practice: Meta-Analyses Using Individual Patient Data are Needed JAMA 1995 274 845 846 7650811 10.1001/jama.274.10.845
Duchateau L Pignon JP Bijnens L Bertin S Bourhis J Sylvester R Individual Patient-versus Literature-Based Meta-analysis of Survival Data: Time to Event and Event Rate at a Particular Time Can Make a Difference, an Example Based on Head and Neck Cancer Controlled Clinical Trials 2001 22 538 547 11578787 10.1016/S0197-2456(01)00152-0
Parmar MKB Torri V Stewart L Extracting Summary Statistics to Perform Meta-Analyses of the Published Literature for Survival Endpoints Stat Med 1998 17 2815 2834 9921604 10.1002/(SICI)1097-0258(19981230)17:24<2815::AID-SIM110>3.0.CO;2-8
Earle CC Wells GA An Assessment of Methods to Combine Published Survival Curves Med Decis Making 2000 20 104 111 10638543
Tudor C Williamson PR Khan SA Best L The value of the aggregate data approach in meta-analysis with time-to-event outcomes J Royal Stat Soc, Series A 2001 164 357 370
Berlin JA Santanna J Schmid CH Szczech LA Feldman HI Individual patient-versus group-level data meta-regressions for the investigation of treatment effect modifiers: economical bias rears its ugly head Stat Med 2002 21 371 387 11813224 10.1002/sim.1023
Mathew T Nordstrom K On the Equivalence of Meta-Analysis Using Literature and Using Individual Patient Data Biometrics 1999 55 1221 1223 11315071 10.1111/j.0006-341X.1999.01221.x
Angelillo IF Villari P Meta-analysis of published studies or meta-analysis of individual patient data? Caesarean section in HIV-Positive women as a study case Public Health 2003 117 323 328 12909421 10.1016/S0033-3506(03)00105-7
| 15850485 | PMC1097735 | CC BY | 2021-01-04 16:32:51 | no | BMC Med Res Methodol. 2005 Apr 25; 5:14 | utf-8 | BMC Med Res Methodol | 2,005 | 10.1186/1471-2288-5-14 | oa_comm |
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-191584768910.1186/1471-2474-6-19Research ArticleThe effectiveness of extra corporeal shock wave therapy for plantar heel pain: a systematic review and meta-analysis Thomson Colin E [email protected] Fay [email protected] Gordon D [email protected] School of Health Sciences, Queen Margaret University College, Edinburgh, EH6 8HF, UK2 Tayside Centre for General Practice, The University of Dundee, The Mackenzie Building, Kirsty Semple Way, Dundee, DD1 4BF, UK3 Public Health Sciences, The University of Edinburgh Medical School, Teviot Place, Edinburgh, EH8 9AG, UK2005 22 4 2005 6 19 19 27 1 2005 22 4 2005 Copyright © 2005 Thomson et al; licensee BioMed Central Ltd.2005Thomson 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
There is considerable controversy regarding the effectiveness of extracorporeal shock wave therapy in the management of plantar heel pain. Our aim was to conduct a systematic review of randomised controlled trials to investigate the effectiveness of extracorporeal shock wave therapy and to produce a precise estimate of the likely benefits of this therapy.
Methods
We conducted a systematic review of all randomised controlled trials (RCTs) identified from the Cochrane Controlled trials register, MEDLINE, EMBASE and CINAHL from 1966 until September 2004. We included randomised trials which evaluated extracorporeal shock wave therapy used to treat plantar heel pain. Trials comparing extra corporeal shock wave therapy with placebo or different doses of extra corporeal shock wave therapy were considered for inclusion in the review. We independently applied the inclusion and exclusion criteria to each identified randomised controlled trial, extracted data and assessed the methodological quality of each trial.
Results
Six RCTs (n = 897) permitted a pooled estimate of effectiveness based on pain scores collected using 10 cm visual analogue scales for morning pain. The estimated weighted mean difference was 0.42 (95% confidence interval 0.02 to 0.83) representing less than 0.5 cm on a visual analogue scale. There was no evidence of heterogeneity and a fixed effects model was used.
Conclusion
A meta-analysis of data from six randomised-controlled trials that included a total of 897 patients was statistically significant in favour of extracorporeal shock wave therapy for the treatment of plantar heel pain but the effect size was very small. A sensitivity analysis including only high quality trials did not detect a statistically significant effect.
==== Body
Background
Plantar heel pain (plantar fasciitis) can be debilitating, often with severe limitations on activity. Typically, patients present with pain in the plantar aspect of the heel whilst walking, particularly after rest. Pain on first weight-bearing in the morning is a prominent diagnostic feature. The precise nature of the condition is poorly understood but literature suggests it is an enthesitis at the attachment of the plantar fascia to the plantar medial tubercle of the calcaneum.
A systematic review of the management of heel pain has highlighted the paucity of evidence for managing the condition. The review concluded that treatments used to reduce heel pain, including steroid injections, NSAIDs, night splints, orthoses and stretching regimes, seem to bring only marginal gains [1]. Extracorporeal shock wave therapy (ESWT) was originally used for lithotripsy, but within the last 10 years has become increasingly used to treat musculoskeletal injuries including calcific tendinitis of the shoulder [2], lateral epicondylitis (tennis elbow) [3-5], non-union or delayed osseous union [6] and plantar heel pain [1,7].
Non-systematic review articles, specific to the effectiveness of ESWT in the treatment of plantar heel pain, produce conflicting conclusions. One 'biometric' review [7] suggested that there is insufficient evidence on which to draw conclusions on the effectiveness of EWST and that more trials are required to detect any benefits from the intervention. Bodekker et al [7] incorporated all levels of evidence, including 4 randomised trials, that did not permit pooling of data or statistical synthesis. Study characteristics and quality assessments were provided in the form of lists. Ogden et al's review of ESWT [8] used a "vote counting" method to conclude that ESWT was a useful treatment for plantar heel pain. No quality assessment of the included trials was presented, but a quantitative data synthesis claims success rates ranging from 34% to 88%. Unfortunately, these estimates are not clearly attributed to any specific outcome. Heller and Niethard [9] identified poor trial methodological quality as a barrier to an assessment of the effectiveness of ESWT and were unable to demonstrate any benefit from the treatment in this narrative review article.
There is considerable controversy emerging regarding the use of ESWT for plantar heel pain. Three recent randomised controlled trials have failed to demonstrate a beneficial effect from the use of ESWT [10-12] and it has been suggested that no more clinical trials should be conducted to evaluate this therapy as a treatment for the painful heel [11]. A narrative review article [13] concluded that the available data do not provide substantive support for its use but this prompted correspondence which illustrates the defense for this electrophysical modality in the management of heel pain [14,15]
The purpose of this systematic review was to conduct a rigorous evaluation using a quantitative synthesis of evidence from randomised controlled trials to make a precise estimate of the effectiveness of ESWT. Our aim was to determine if ESWT is effective in the treatment of patients with plantar heel pain when compared with a control group.
Methods
Search strategy
Randomised controlled trials were identified by searching the following data sources: The Cochrane Musculoskeletal Injuries Group specialized register of trials (August 2003), the Cochrane Central Register of Controlled Trials (The Cochrane Library issue 3, 2003), MEDLINE (from 1966 to September 2004), EMBASE (from 1982 to September 2004), CINAHL (from 1982 to September 2004) and reference lists of articles and dissertations. In Medline (SilverPlatter), the first two levels of the optimum search strategy [16] were combined with the following subject-specific search terms:
1. HEEL* and SYNDROME*
2. (JOG* or TENNIS* or POLICE* or GONORREAL) near HEEL*
3. PLANTAR near FASCI*
4. explode "FASCIITIS"/ all subheadings
5. (PLANTAR or HEEL* or CALCAN* or FOOT*) near PAIN*
6. HEEL near SPUR
7. "CALCANEUS"/ all subheadings
8. #1 or #2 or #3 or#4 or #5 or #6 or #7
Further details of the search strategy and details of the hand search have been previously published [1], [see Additional file 1].
Study selection
We considered all randomised controlled trials of plantar heel pain treatments for inclusion in the review. Trials comparing ESWT with placebo or different doses of ESWT were considered. Participants with a clinically confirmed diagnosis of plantar heel pain were included. Adult participants in any trial whether they were part of the general population, athletes, or individuals with seronegative arthropathies and enthesopathies were also considered for inclusion. Any age group was admissible. It was our intention that trials involving children alone, or dealing specifically with young athletes, would be analysed separately. We excluded trials evaluating treatments for plantar heel pain arising from calcaneal fractures, calcaneal tumours, previous surgery for plantar heel pain, or posterior heel pain.
Outcome measures
We chose morning pain as our a priori primary outcome measure for this systematic review. We consider it to be the most important outcome as it is the single most consistent feature of plantar heel pain. Morning pain (pain on first rising, first step pain or start up pain) is universally reported by patients complaining of plantar heel pain and it is also strongly diagnostic for the condition[17]. The secondary outcome measures were walking pain, pressure pain, any measure of disability, quality of life measures and adverse events.
Data abstraction
Two of the authors (CT,FC) independently applied the inclusion and exclusion criteria to each trial and then extracted data regarding details of the patients (number, mean age and age range, inclusion and exclusion criteria), details of the interventions, nature and timing of outcome measures. Disagreements were resolved by discussion of the articles by the reviewers. We wrote to trialists for additional information on trial methodology (method of randomization) and results (usually requests for data not presented in the original reports such as standard deviations or some other measure of variance).
Validity assessment
A quality assessment tool[18] adapted for use in a related systematic review of interventions for the treatment of plantar heel pain for the Cochrane Library [1] was applied to each of the included trials. This addressed the following questions:
1. Was the generation of randomization sequence described?
2. Was the method of allocation concealment described?
3. Was an intention to treat analysis used?
We assessed intention to treat on the basis of whether patients were analyzed according to the allocated treatment irrespective of whether this treatment was delivered or not.
4. What number of patients were lost to follow-up?
In assessing loss to follow-up we considered whether authors had presented numbers lost and timing, and the reasons for the loss. We presented the numbers lost to follow up as percentages.
5. Was the outcome assessment blind?
6. Was the patient blind to treatment allocation?
This led to each trial being attributed a quality score out of a maximum of 6 points (Table 1.).
Table 1 Quality assessment of included trials
Author randomisation sequence allocation concealment Assessor blind Patient blind Loss to follow up at trial end Intention to treat Quality score
Abt et al [21] No No Yes Yes 11% No 3
Buch et al [27] Yes Yes Yes Yes 4% No 5
Buchbinder et al [10] Yes No Yes Yes 6% Yes 5
Cosentino et al [33] No No Yes Not stated Not stated No 1
Haake et al [11] Yes Yes Yes Yes 16% Yes 6
Krischek et al [22] No No No Not stated 6% No 1
Ogden et al [28] No No Yes Yes 1.5% No 3
Rompe et al [32] No No Yes Yes 16% No 3
Rompe et al [30] No No No Yes Not stated No 1
Rompe et al [31] No No Yes Yes 20% No 3
Speed et al [12] No No Yes Yes 14% Yes 4
Quantitative data synthesis
When measures of variance were not available from the original report, it was our intention to derive these from p-values. When data were available for a pooled estimate of the impact of intervention it was intended that meta-analyses would be conducted for direct comparisons. We intended to present weighted mean differences and 95% confidence intervals for outcomes for each randomised controlled trial and group them in relevant sub-groups according to the specific question they addressed. We intended to use a fixed effects model to estimate the pooled effect as our primary analysis where no evidence of heterogeneity was detected [19]. However, if evidence of heterogeneity was found to be present we intended to use a random effects model [20]. Meta-analyses were generated using RevMan software. We planned to perform subgroup analyses and sensitivity analyses, regarding any anomalies with the included trials, methodological scores and industry sponsorship. We proposed to perform a funnel plot to detect publication bias.
Results
Selection of trials
The search strategy identified a total of 205 studies, of which 15 were identified as RCTs that evaluated ESWT for plantar heel pain. Two of these were translated from German into English [21,22]. Four trials [23-26] were excluded from the review: in one, the intervention and control groups were treated at different time points making valid comparisons of patient outcomes in both groups impossible [24]. The second trial contained five year follow-up data from an RCT published in 1996 [23]. These trial data were confounded by placebo patients receiving additional therapies after 12 weeks. The third [25] and fourth [26]excluded trials were duplicated data previously reported by Buch [27] and by Ogden [28]respectively. The flow diagram in Figure 1 provides details of the included and excluded trials and those included in the final meta-analysis[29].
Figure 1 Progress through the stages of the meta-analysis [29].
Description of included studies
Eleven RCTs were included in this review and they reported data published between 1996–2003 from trials involving 1290 patients [10-12,21,22,27,28,30-33]. Table 1 shows the quality assessment scores and Table 2 and Table 3 the baseline data. The trials evaluated different doses of ESWT against either a placebo dose or a control dose so low as to be considered therapeutically ineffective [10] (Table 4). Only five of the trial reports contained summary statistics to permit pooling of data collected at 12 weeks in a forest plot [10-12,27,28]. Standard deviations were derived from the p value reported in one manuscript in order to incorporate a sixth trial in the meta-analysis, the timing of the outcomes varied between 17 and 20 weeks for this trial [21].
Table 2 Baseline characteristics of participants in respective trials. (N/a- data not available).
Author Age mean (SD and/or range) years Female:male (% female) Mean BMI (SD)
Treatment group Control group Treatment group Control group Treatment group Control group
Abt et al [21] 56.5 57.4 11:6 (64.7) 9:6 (60) 30.1 28.5
Buch et al [27] 50.4 (10.3, 26–69) 53.0 (9.7, 31–72) 61:14 (81.3) 46:26 (63.9) 28.9 28.5
Buchbinder et al [10] 52.2 (12.8) 54.2 (12.0) 46:3 (57.5) 47:3 (58.0) 29.5 28.9
Cosentino et al [33] 55.6 (45–68) 18:12 (60.0) 25:5 (83.3) N/a N/a
Haake et al [11] 53.1 (10.8) 52.9 (10.8) 98:37 (72.6) 106:30 (77.9) 29.4 (4.9) 29.7 (4.8)
Krischek et al [22] 54.0 55.0 (56.0) (72.0) N/a N/a
Ogden et al [28] 49.6 (20–79) 171 (65.9) N/a N/a
Rompe et al [32] 44.0 (26–61) 49.0 (31–63) 21:29 (42.0) 20:30 (40.) N/a N/a
Rompe et al [30] 47.0 (26–61) 51.0 (31–58) 5:10 (33.3) 6:9 (40) N/a N/a
Rompe et al [31] 43.0 (32–59) 40.0 (30–61) 10:12 (45.5) 13:10 (56.5) N/a N/a
Speed et al [12] 51.7 (25–76) 52.5 (30–73) 26:20 (56.5) 25:17 (59.5) N/a N/a
Table 3 Baseline characteristics of participants in respective trials continued (N/a- data not available).
Author Duration of heel pain Median (SD and/or range) months Base line morning pain VAS score (SD)
Treatment group Control group Treatment Group Control group
Abt et al [21] 19.0 19.0 5.7 5.3
Buch et al [27] 20.7 (21.1, 6–120) 24.0 (21.1,6–99) 7.7 (1.4) 7.7 (1.5)
Buchbinder et al [10] 9.0 (2–150) 10.8 (2–222.5) 7.3 (2.5) 6.8 (3.2)
Cosentino et al [33] 8.2(6–12) 8.2(1.2) N/a N/a
Haake et al [11] 13.0 (10–24) 13.0 (9–24) 7.8 (2.4) 7.7 (2.3)
Krischek et al [22] 22.0 23.0 N/a N/a
Ogden et al [28] 32.2 35.9 8.1 8.2
Rompe et al [32] 8.0 (6–19) 10.0 (6–20) N/a N/a
Rompe et al [30] 16.0 (12–36) 22.0 (12–38) N/a N/a
Rompe et al [31] 20.0 (12–60) 18.0 (12–72) 6.9 (1.3) 7.0 (1.3)
Speed et al [12] 16.7 (12–312) 13.5 (12–312) 7.4(2.0) 7.0(2.0)
Table 4 Details of studies included in the systematic review
Author Included in meta-analysis Local anesthetic to both groups Details of placebo/sub therapeutic dose Ultrasound guidance N Weighted mean difference – morning pain (95%CI) Timing of outcomes (weeks)
Abt et al [21] yes yes Absorbent block Not stated 32 2.00 (0.47 to 3.53) 19,32
Buch et al [27] yes yes Absorbent foil yes 150 0.70 (-0.26 to 1.66) 12
Buchbinder et al [10] yes no 6000 to 7500 vs 300 impulses yes 178 -0.50 (-1.55 to 0.55) 6,12
Cosentino et al [33] no no Not stated yes 60 Not available 4,12
Haake et al [11] yes yes Polythene foil barrier Not stated 272 0.50 (-0.31 to 1.31) 6,12, 52
Krischek et al [22] no no 1500 vs 300 impulses Not stated 50 Not available 6,12
Ogden et al [28] yes yes* Styrofoam block no 260 0.56 (-0.26 to 1.38) 4,8,12
Rompe et al [32] no no 3000 vs 30 impulses Not stated 119 Not available 12, 52
Rompe et al [30] no no 1 cm gap no 36 Not available 3,6,12,24
Rompe et al [31] no no Reflecting pad yes 45 2.60 (1.37 to 3.83) 26,52
Speed et al [12] yes no Focus outside patient yes 88 -0.36 (-1.66 to 0.94) 4,8,12,24
* Local anaesthetic was used for both groups but was different for the placebo group and the treatment group.
Table 2 and table 3 present details of the baseline pain scores, and demographic variables for participants from all eleven included trials. All included adult patients only. The duration of pain was greater than 6 months in ten trials [11,12,21,22,27,28,30-33]. In one trial [10] the duration of pain was shorter than six months for some patients but no patient had a duration of pain less than 8 weeks. The duration of pain ranged from 8–600 weeks and 8–980 weeks for the ESWT and placebo groups respectively. The median values for duration of pain were 36 weeks and 43 weeks. The demography of the patients in this systematic review of ESWT for plantar heel pain was similar to those patients who have participated in evaluations of other interventions for heel pain [1]. The effects of ESWT in people who had a calcaneal spur on x-ray [4,32], were running athletes [31], were being considered for surgical intervention [30,32,32], had failed to respond to conservative treatments [27,28,30,32], or were defined as recalcitrant cases [22], were all included in this systematic review.
There was diversity in the types of primary and secondary outcomes collected from patients in the 11 RCTs. Table 5. summarizes the most commonly reported outcomes measures indicating, where available, the outcomes provided. With the exception of three trials [22,30,32] all presented data for visual analogue scale scores of morning pain. Walking pain is a relevant outcome measure and was reported by eight trials [10,11,21,22,30,32,33]. Only two of these trials contained compatible data [30,32] and insufficient data are provided to permit pooling. The remaining trials described a wide variety of walking ability using incongruous scoring systems. Six of the trials [11,21,22,30,32,33], show a favourable outcome for walking pain after ESWT. Resting and night pain are not common symptoms of heel pain, in our experience, but data for these outcomes were collected in four trials [12,21,30,32]. Five trials reported the collection of pressure pain outcomes from the application of pressure from either a manual application or an electronic device [21,27,28,30,32]. Other outcomes reported were Roles and Maudsley scores [11,21,27], Maryland Foot score[10], SF12 [27], SF36 [10], problem elicitation technique [10] and The Ankle Hindfoot Scale [31].
Table 5 Summary of most commonly reported outcomes measures at 12 weeks (or nearest point to). P values relate to active treatment versus placebo or reduced dose. * Indicates a statistical significant difference in favour of EWST treatment. Figures in parentheses are 95% confidence intervals. Where p-values were not provided, the values for mean and standard deviations [SD] are given, I indicates EWST group, II indicates placebo group. "Favours ESWT" indicates a better outcome for ESWT where neither of the previous details are provided.
Author Morning/start up pain Overall pain Walking ability /activity related Foot specific score Pain at rest (100 mm VAS) Pain on pressure Night pain/evening pain End point
Abt et al [21] P = 0.016* - Favours ESWT - P = 0.01* P = 0.26 P = 0.01* 19 weeks
Buch et al [27] P = 0.0309 - P = 0.7377 I. (49.1–71.5)
II.(32.9–55.9) AOFAS - Not significant P < 0.4338 12 weeks
Buchbinder et al [10] P = 0.92 (-12.7 – 13.1) P = 0.99 (-10.3–11.5) P = 0.0.72 (0.6–1.9) P = 0.85 (-7.6–5.3) Maryland FS - - - 12 weeks
Cosentino et al [33] P < 0.0001* - P < 0.0001* - P < 0.0001* - - 12 weeks
Haake et al [11] I mean = 4.0, SD = 3.2
II mean = 4.5, SD = 3.0 - Favours ESWT - I mean = 2.4, SD = 2.6
II mean = 2.4, SD = 2.5 I mean = 4.0, SD = 3.2
II mean = 4.3, SD = 3.2 I mean = 1.5, SD = 2.4
II mean = 1.8, SD = 2.5 12 weeks
Krischek et al [22] - Favours ESWT Favours ESWT - - Favours ESWT - 12 weeks
Ogden et al [28] Favours ESWT - - - - Favours ESWT - 12 weeks
Rompe et al [32] - - Favours ESWT - P < 0.0001* P < 0.0001* P < 0.0001* 12 weeks
Rompe et al [30] Favours ESWT Favours ESWT P < 0.0001* - P < 0.05* P < 0.0001* P < 0.05* 12 weeks
Rompe et al [31] P = 0.0004* - - P = 0.0025* AOFAS - - - 26 weeks
Speed et al P = 0.664 (0.656–1.271) P = 0.246 (0.626–1.093) - - - - P = 0.378 (0.620–1.166) 12 weeks
Of the 11 RCTs that met our inclusion criteria, eight were placebo controlled trials [11,12,21,27,28,31-33]. Three trials used a low, sub-therapeutic dose as control [10,22,30]. The doses for the intervention groups and methods used to disable the equipment for the placebo group and the sub-therapeutic groups are provided in Table 2 and Table 3. The dose of ESWT varied between trials in both energy levels and the number of impulses administered. With the exception of two trials, [10,12], all excluded patients had the condition for less than six months. Only one trial [10] did not require patients to have exhausted conservative therapies for recalcitrant plantar heel pain before embarking on treatment with ESWT but information presented reveals that the majority of patients did receive a number of conservative therapies. Krischek et al [22] and Rompe et al [31] included only patients whose next management option was surgery.
Quantitative data synthesis
Figure 2. shows the pooled analysis of data from 6 trials which produce a weighted mean difference of 0.42 in favour of ESWT. This treatment effect is statistically significant (p = 0.04), but the effect is small (95% confidence interval of 0.02 to 0.83) with respect to morning pain (first step pain). All outcomes were taken at 12 weeks, except for one trial [21] which reported the first outcome measured at (on average) 19 weeks. There was no evidence of heterogeneity (p = 0.11) and a fixed effects model was used.
Figure 2 Pooled estimates of 10 cm VAS scores for morning pain at 12 weeks
We repeated the meta-analysis excluding the data from the trial by Abt et al [21], the only trial for which we had to impute measures of variance. The resultant weighted mean difference was 0.30 in favour of ESWT, with a 95% confidence interval of -0.12 to 0.72. This effect is no longer statistically significant.
We performed a sensitivity analysis for the quality of trial reports by dividing the six trials into two groups; those that received a quality assessment score of four or more [10-12,27] and those receiving a score of less than four [21,28] to perform meta-analyses using fixed effects models. The four better quality trials produced a non significant result (weighted mean difference 0.21, 95% confidence interval -0.29 to 0.70 cm, p = 0.41) whereas the two trials scoring less than three produced a significant result in favour of active treatment (weighted mean difference -0.90, 95% confidence interval -1.62 to -0.19, p = 0.01).
Adverse events
Two trials did not report adverse events [12,30]. Buchbinder et al [10] reported pain for one week by one patient in each arm of the trial; one patient in the active arm of the trial reported a sensation of heat and numbness, whilst another complained of bruising. One patient in the placebo arm complained of a burning sensation in the heel and ankle. Ogden et al [28] reported 38 procedure related complications, 18 of which occurred in the active treatment arm. The most common procedure related complications were mild neurological symptoms (numbness, tingling). One patient who suffered a plantar fascial rupture 4 weeks after active treatment had undergone multiple cortisone injections prior to embarking upon treatment with ESWT.
Haake et al [11] reported a statistically significant difference in the number of side effects in the active and placebo groups; OR 2.26 (95% confidence interval 1.02 to 5.18) [11]. These were; skin reddening, pain and local swelling. The same authors [11] also describe less frequent complaints of dizziness, sleep disturbance haematoma, nausea and hair loss as non-serious effects and discounted one report of a deep vein thrombosis in a placebo participant as a co-incidental event. In two trials, [31,32] the unpleasant nature of ESWT experienced by patients during treatment was reported. These sensations were regarded as less unpleasant than local cortisone infiltration. Krischek et al [22] reported that there were no adverse events noted in trial participants.
Industry sponsorship
Companies who produce ESWT equipment provided some sponsorship in three trials [11,27,28] (Table 6). One trial [28] was the basis for the first Food and Drug Administration (FDA) approval for ESWT. A financial interest with HealthTronics was declared in correspondence following the publication of the trial [34,35]. The trial by Buch et al [27] was sponsored by Dornier Med tech Inc and the data were also used to gain approval for the use of ESWT in the management of plantar heel pain from the FDA. Haake et al [11] stated no competing interests but did declare that a manufacturer of ESWT equipment had provided the machine used in the trial. Two trials [10,12] declared funding from sources other than industry. In the remaining trials there was no explicit declaration of competing interests [21,22,30-33] (Table 6).
Table 6 Details of ESWT devices, dose of impulses administered.
Author Device ESWT impulse dose × number of treatments Low energy/ high energy (energy level) Details of sponsorship
Abt et al [21] Ossatron High Medical Technology 1000 × 2 Low Energy (0.08 mJ/mm2 No declaration
Buch et al [27] Epos Ultra Dornier Medical Systems 3800 total High energy (0.03–0.36 mJ/mm2 -total 1300 mJ/mm2) Industry sponsored trial but this was not declared
Buchbinder et al [10] Epos Ultra Dornier Medical Systems 2000–2500 × 3 Low energy (0.02–0.33 mJ/mm2-total 1000 mJ/mm2) Declared funding – not from industry
Cosentino et al [33] Orthima Direx Med Sys Ltd 1200 × 6 Not stated (0.03–0.4 mJ/mm2) No declaration
Haake et al [11] Epos Ultra Dornier Medical Systems 4000 × 3 Low energy (0.08 mJ/mm2-total 0.96 J/mm2) Declared: industry provided machine
Krischek et al [22] Osteostar Siemans 500 × 3 Low energy (0.08 mJ/mm2) No declaration
Ogden et al [28] Ossatron High Medical Technology 1500 total High energy (0.22 mJ/mm2-total 324.25 J) Industry sponsored trial but this was not declared
Rompe et al [32] Osteostar Siemans 1000 × 3 Low energy (0.06 mJ/mm2) No declaration
Rompe et al [30] Osteostar Siemans 1000 × 3 Low energy (0.06 mJ/mm2) No declaration
Rompe et al [31] Sonocur Plus Siemens 2100 × 3 Low energy (0.16 mJ/mm2) No declaration
Speed et al [12] Sonocur Plus Siemens 1500 × 3 Low energy (0.06 mJ/mm2) Declared funding – not from industry
Discussion
The lack of convergence of findings from randomised evaluations of EWST for plantar heel pain has resulted in clinical uncertainty about its effectiveness. Within this systematic review, we have been able to evaluate the effectiveness of ESWT in a meta-analysis and used the pooled data to arrive at more precise conclusions about its usefulness in clinical practice.
The meta-analysis shows a statistically significant benefit with ESWT on plantar heel pain from outcomes of 897 patients' VAS scores of morning (first-step) pain assessed at or around 12 weeks but we do not consider this clinically significant since the observed benefit equates to less than one half centimeter on a 10 cm VAS. The 95% confidence interval is compatible with a mean treatment benefit of at most 0.83 cm. A sensitivity analysis including only those higher quality trials did not produce evidence of a statistically significant benefit. Only one trial included in the review discussed what might constitute a clinically meaningful reduction in plantar heel pain: Buchbinder et al [10], suggest that 0.7 cm reduction of heel pain may not be clinically relevant.
We included one trial in the meta-analysis which used sub-clinical doses as controls [10] and combined these patient outcomes with those from trials which used sham treatments as controls [11,12,21,27,28]. All six trials [10-12,21,27,28] also used different doses of ESWT but, despite the differences in the use of control interventions and doses, no evidence of heterogeneity in the patient outcomes was detected in the pooled estimate (figure 2). Nor does there appear to be a dose-response relationship for ESWT; trials using both high and low doses have reported similar effects as is evident from the estimates from the trials by Haake et al [11] and Abt et al [21] (Table 6, figure 2).
We were grateful to the authors of trials included in this review who provided supplementary data in response to our correspondence [10,11] but disappointed that data from all 11 trials were not available to us. Five trials were not included in the meta-analysis either because adequate data were not provided [22,33] the timing of the outcomes differed greatly from the other trials [31] or the outcomes were clinically irrelevant [30,32]. Consequently, information about the effects of ESWT in 310 patients with heel pain was effectively lost to re-analysis. Any future reporting of patient outcomes should include means of pain scores with measures of variance in order that new trials can be included in meta-analyses and weighted mean differences and confidence intervals calculated [36].
Rompe et al conducted a small trial (n = 40) which evaluated the benefits of ESWT in running athletes [31] and reported a mean difference of 2.60 (95% confidence interval 1.37 to 3.83) for morning pain at 6 months. This effect size is statistically significantly different from the combined outcomes presented in Figure 2 but not statistically different from the mean difference in outcomes reported in the small trial by Abt et al [21] 2.00 (95%confidence interval 0.47 to 3.53) at 19 weeks (n = 37). That the two smallest trials included in the review should produce between-group comparisons of pain in the morning that reach statistical significance when estimates from larger studies do not is surprising. Sample size is an important factor in experimental bias in clinical trials as effect size estimates from small studies can be highly variable [37]. The effect sizes from these small studies may be due to ESWT being beneficial in certain sub groups within the population (e.g. runners), or may be as a result of a failure to blind the participants successfully to their treatment allocation, as previously reported by one of the authors [30]. Alternatively, these data may be aberrant values that are more likely to occur by chance in small studies than larger ones [38].
ESWT was not considered a suitable therapy for the first-line management of heel pain by the majority of the investigators. This may be because of limited access to this relatively new and expensive equipment or, more likely, because of the favourable natural history of this condition.
In the absence of a validated heel pain specific outcome measure, our a priori choice of morning pain as the primary outcome measure was vindicated by eight of the of the eleven included trials collecting morning pain or first step/start up pain outcomes. One trialist [10] used a problem elicitation technique which confirmed "walking after getting out of bed in the morning" as the most frequently reported problem by patients with heel pain. We had planned to pool additional secondary outcome measures, such as walking pain, but this was not possible because of the diversity of the outcome measures used and differences in the data collected. Some of the outcomes that have been used to assess the effects of treatments were clinically irrelevant in our opinion [30-33]. Night pain and resting pain are not symptoms that we commonly encounter in patients seeking treatment for plantar heel pain. Three trials [11,21,27] incorporated the Roles Maudsley scale and one trial [10] used the Maryland Foot Score as measures of disability. It is commendable that two of the investigators [10,27] used generic health outcomes, SF36 and SF 12 respectively. Future trials should include outcomes of disability as well as the impact on health related quality of life and not just pain when assessing the effect of interventions for heel pain.
Of the eight outcomes listed in Table 5, only "pain at rest" is distinct with four of the five trials [11,21,30,32,33] favouring ESWT compared with placebo or reduced dose. As previously discussed, this outcome measure is not a key feature of plantar heel pain. All other outcome measures are equivocal.
Minimal side effects were reported by Abt et al [21] and Buchbinder et al [10]. The most frequently reported adverse event from the use of ESWT is pain [11,27,32,33] which appeared to affect some patients both during and after the procedure.
The quality of reporting varied amongst trials. The three most recent trials [10,11,31] all received above average quality scores for trial reporting. This is an encouraging development for those interested in improving the outcomes for patients who have heel pain and may reflect both the use of checklists such as the CONSORT statement [36] for trial reports now demanded by many journal editors as well as a greater awareness of good trial reporting practice by trialists themselves. There was however, a contrast in the results obtained from the four better quality trials, scoring three or above, when meta-analyzed separately from the two poorer quality trials. Better quality trials did not favour ESWT whilst the poorer quality ones did.
Industry sponsorship
At least two of the trials included in our meta-analysis, received some form of sponsorship from a company manufacturing ESWT [27,28] although this has not been made explicit within the published papers. Both these trials reported significant benefit from ESWT. One further trial Haake et al [11] declared being supplied with the ESWT equipment and reported no statistically significant effects between the two groups. Six of the trials [21,22,30-33] have not made it clear whether there is any conflict of interest or not. In a systematic review to investigate whether the funding of drug studies by the pharmaceutical industry is associated with bias, Lexchin et al [39] concluded that industry sponsorship was more likely to produce results favouring the sponsors' product than studies funded from other sources.
Publication bias
In view of concerns about publication bias, it is encouraging that three large, negative trials have been published in high impact journals. We were unable to recognize the existence of small, unpublished studies showing no statistically significant benefits. However, the existence of any such trials would only serve to endorse the findings of the meta-analysis in this systematic review.
Conclusion
It has been suggested that the poor outcomes reported by recent randomised controlled trials evaluating ESWT for plantar heel pain means no further trials should be conducted [11]. A meta-analysis of data from six randomized controlled trials that included a total of 897 patients was statistically significant in favour of extracorporeal shock wave therapy for the treatment of plantar heel pain but the effect size was very small. When the two poorest quality trials, and therefore the greatest source of bias, are removed from the meta-analysis, the result is not statistically significant. This systematic review does not support the use of ESWT for plantar heel pain in clinical practice.
Abbreviations
ESWT: Extracorporeal shock wave therapy
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FC and CT performed the literature search, extracted data, performed data analyses and compiled the manuscript. GM performed data analyses and compiled the manuscript. We can confirm that all authors have access to all data in the study and that they held final responsibility for the decision to submit for publication.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
"Details of EMBASE and MEDLINE search strategies". this additional file contains full details of the EMBASE and MEDLINE search strategies that were used for this systematic review. Only an abbreviated version was provided within the text.
Click here for file
==== Refs
Crawford F Thomson CE Interventions for treating plantar heel pain 2005 1 Wiley JW
Gerdesmeyer L Wagenpfeil S Haake M Maier M Loew M Wörtler K Lampe R Seil R Handle G Gassel S Rompe JD Extracorporeal shock wave therapy for the treatment of chronic calcifying tendonitis of the rotator cuff – a randomized controlled trial. Journal of the American Medical association 2003 290 2573 2580 14625334 10.1001/jama.290.19.2573
Speed CA Nichols D Humphreys H Wies JT Burnet S Extracorporeal shock wave therapy for lateral epicondylitis – a double blind randomised controlled trial. Journal of Orthopaedic Research 2002 20 895 898 12382950 10.1016/S0736-0266(02)00013-X
Buchbinder R Green S White M Barnsley L Smidt N Shockwave therapy for lateral elbow pain. 2003 3 Oxford. Update Software.
Haake M Konig IR Decker T Riedel C No effectiveness of extracorporeal shock wave therapy in the treatment of tennis elbow-results from a prospective randomized placebo-controlled multicenter trial. Journal of Bone and Joint Surgery 2004 84A 1982 1991
Schaden W Fischer A Sailler Extracorporeal Shock Wave Therapy of nonunion or delayed osseous union. Clin Orthop Relat Res 2001 387 90 94 11400900 10.1097/00003086-200106000-00012
Bodekker IR Schafer H Haake M Extracorporeal shock wave therapy (ESWT) in the treatment of plantar fasciitis - A biometrical review. Clinical Rheumatology 2001 20 324 330 11642513
Ogden JA Alverez RG Marlow M Shockwave therapy for chronic plantar fasciitis: a meta-analysis Foot Ankle Int 2002 23 301 308 11991474
Heller KD Niethard FU Der einsatz der extrakorporalen stosswellentherapie in der orthopadie-eine metaanalyse Z Orthop 1998 136 390 401 9823633
Buchbinder R Ptasznik R Gordon J Buchanan J Prabaharan V Forbes A Ultrasound guided Extracorporeal Shockwave Terapy for Plantar Fasciitis: A randomized controlled trial. Journal of the American Medical association 2002 288 1364 1372 12234230 10.1001/jama.288.11.1364
Haake M Buch M Goebel F Vogel M Mueller I Hausdorf J Zamzow K Schade-Brittinger C Mueller HH Extracorporeal shock wave therapy for plantar fasciitis: randomised controlled multicentre trial. B M J 2003 327
Speed CA Nichols D Wies J Humphreys H Richards C Burnet S Hazelman BL Extracorporeal shockwave therapy for plantar fasciitis. A double blind randomized controlled trial. Journal of Orthopaedic Research 2003 21 937 940 12919884 10.1016/S0736-0266(03)00048-2
Buchbinder R Plantar Fasciitis New England Journal of Medicine 2004 351 834 834 15317903 10.1056/NEJM200408193510824
Buchbinder R Plantar fasciitis New England Journal of Medicine 2004 350 2159 2166 15152061 10.1056/NEJMcp032745
Rompe JD Plantar fasciitis New England Journal of Medicine 2004 351 834 834 15317903 10.1056/NEJM200408193510824
Robinson KA Dickerson K Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. International Journal of Epidemiology 2002 31 150 153 11914311 10.1093/ije/31.1.150
Richardson EG Canale TS Disorders of tendons and fascia Campbell's Operative Orthopaedics 1998 43 9 St Louis Mosby 1913
Jadad AR Moore RA Carrol D Assssing the quality of reports of randomised clinical trials: is blinding necessary? Controlled Clinical Trials 1996 17 1 12 8721797 10.1016/0197-2456(95)00134-4
Whitehead A Whitehead J A general parametric approach to the meta-analysis of randomized clinical trials. Stat Med 2004 10 1665 1677 1792461
DerSimonian R Laird N Meta-analysis in clinical trials Controlled Clinical Trials 1986 7 177 188 3802833 10.1016/0197-2456(86)90046-2
Abt T Hopfenmuller W Mellerwicz H Stosswellentherapie bei therapieresistenter plantarfasziitis mi ferensporn: eine prospektiv randomised plazebokkontrollierte doppelblindstudie. Z Orthop 1998 136 169 174 9615981
Krischek O Rompe JD Herbsthofer B Nafe B Symptomatische niedrig-enegetische stosswellentherapie bei fersenschmerzen und radiologisch nachweisbarem plantaren fersensporn. Z Orthop 1998 136 169 174 9615981
Rompe JD Schoellner C Nafe B Evaluation of low-energy extra corporeal shock wave application for treatment of chronic plantar fasciitis. Journal of Bone and Joint Surgery 2002 84A 335 341 11886900
Hammer DS Rupp S Kreutz A Pape D Kohn D Seil R Extracorporeal shock wave therapy (ESWT) in patients with plantar fasciitis. Foot and Ankle International 2002 23 309 313 11991475
Theodore G HBM Amendola A Bachmann C Fleming L L Zingas C Extracorporeal shockwave therapy for the treatment of plantar fasciitis Foot and Ankle International 2004 25 290 297 15134608
Ogden JA Alvarez RG Levitt RL Johnson JE Marlow ME Electrohydraulic high-energy shock-wave treatment for chronic plantar fasciitis Journal of Bone and Joint Surgery 2004 86A 2216 2228 15466731
Buch M Knorr L Fleming TG Amenola A Bachman C Zingas C Siebert WE Extracorporeal shock wave therapy in the treatment of symptomatic heel spur - A review. Orthopade 2002 31 637 644 12219661 10.1007/s00132-002-0323-z
Ogden JA Alvarez R Levitt R Cross GL Marlow M Shock wave therapy for chronic proximal plantar fasciitis. Clinics in Orthopaedics and Related Research 2001 387 47 59 10.1097/00003086-200106000-00007
Moher D Cook DJ Eastwood S Olkin I Rennie D Stroup DF Improving the quality of reports of meta-anlyses of randomised controlled trials: the QUOROM statement. The Lancet 1999 354 1896 1900 10584742 10.1016/S0140-6736(99)04149-5
Rompe JD Hopf C Nafe B Burger R Low-energy extracorporeal shock wave therapy for painful heel: a prospective controlled single blind study. Archives Orthopaedic Trauma Surgery 1996 115 75 79 10.1007/BF00573445
Rompe JD Decking J Schoeliner C Nafe B Shockwave application for chronic plantar fasciitis in running athletes. A prospective randomized controlled trial. The American Journal of Sports Medicine 2003 31 268 275 12642264
Rompe JD Kullmer K Riehle HM Herbsthofer B Eckard A Burger R Effectiveness of low energy extracorporeal shock waves for chronic plantar fasciitis. Foot and Ankle Surgery 1996 2 215 221
Cosentino R Falsetti P Manca S De Stefano R Frati E Frediani B Baldi F Selvi E Marcolonga R Efficacy of extracorporeal shock wave treatment in calcaneal enthesophytosis. Annals of the Rheumatic Diseases 2001 60 1064 1067 11602481 10.1136/ard.60.11.1064
Beckman KD Letter to the editor Clin Orthop Relat Res 2002 398 267 268 11964660 10.1097/00003086-200205000-00037
Ogden JA Reply to Dr Kenneth Beckman Clin Orthop Relat Res 2002 398 268 269 10.1097/00003086-200205000-00038
Moher D Schulz KF Altman D The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised controlled trials. Journal of the American Medical association 2001 285 1992 1995 11308436 10.1001/jama.285.15.1992
NHS Centre for Reviews and Dissemination Undertaking systematic reviews of research on effectiveness. CRD guidelines for those carrying out or commissioning reviews 2001 4 York, The University of York
Sutton AJ Abrams KR Jones DR Sheldon TA Song F Systematic reviews of trials and other studies. Health Technol Assess 1998 2 1 276 10347832
Lexchin J Bero L Djulbecgovic BD Clark O Pharmaceutical industry sponsorship and research outcome and quality: systematic review. British Medical Journal 2003 326 1167 1170 12775614
| 15847689 | PMC1097736 | CC BY | 2021-01-04 16:32:04 | no | BMC Musculoskelet Disord. 2005 Apr 22; 6:19 | utf-8 | BMC Musculoskelet Disord | 2,005 | 10.1186/1471-2474-6-19 | oa_comm |
==== Front
BMC NeurolBMC Neurology1471-2377BioMed Central London 1471-2377-5-91583310810.1186/1471-2377-5-9Research ArticleCerebral infarction in diabetes: Clinical pattern, stroke subtypes, and predictors of in-hospital mortality Arboix Adrià [email protected] Antoni [email protected]ía-Eroles Luis [email protected] Marcos Lourdes [email protected] Joan [email protected] Montserrat [email protected] Cerebrovascular Division, Department of Neurology, Hospital del Sagrat Cor, Universitat de Barcelona, Viladomat 288, E-08029 Barcelona, Spain2 Clinical Information Systems, Hospital Universitari de Bellvitge, Feixa Llarga s/n, E-08907 L'Hospitalet de Llobregat, Barcelona, Spain3 Primary Health Care Center 'Sant Joan', Passeig Sant Joan 20, E-08010 Barcelona, Spain2005 15 4 2005 5 9 9 3 11 2004 15 4 2005 Copyright © 2005 Arboix et al; licensee BioMed Central Ltd.2005Arboix 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
To compare the characteristics and prognostic features of ischemic stroke in patients with diabetes and without diabetes, and to determine the independent predictors of in-hospital mortality in people with diabetes and ischemic stroke.
Methods
Diabetes was diagnosed in 393 (21.3%) of 1,840 consecutive patients with cerebral infarction included in a prospective stroke registry over a 12-year period. Demographic characteristics, cardiovascular risk factors, clinical events, stroke subtypes, neuroimaging data, and outcome in ischemic stroke patients with and without diabetes were compared. Predictors of in-hospital mortality in diabetic patients with ischemic stroke were assessed by multivariate analysis.
Results
People with diabetes compared to people without diabetes presented more frequently atherothrombotic stroke (41.2% vs 27%) and lacunar infarction (35.1% vs 23.9%) (P < 0.01). The in-hospital mortality in ischemic stroke patients with diabetes was 12.5% and 14.6% in those without (P = NS). Ischemic heart disease, hyperlipidemia, subacute onset, 85 years old or more, atherothrombotic and lacunar infarcts, and thalamic topography were independently associated with ischemic stroke in patients with diabetes, whereas predictors of in-hospital mortality included the patient's age, decreased consciousness, chronic nephropathy, congestive heart failure and atrial fibrillation
Conclusion
Ischemic stroke in people with diabetes showed a different clinical pattern from those without diabetes, with atherothrombotic stroke and lacunar infarcts being more frequent. Clinical factors indicative of the severity of ischemic stroke available at onset have a predominant influence upon in-hospital mortality and may help clinicians to assess prognosis more accurately.
==== Body
Background
Diabetes mellitus is a well-established independent risk factor for ischemic stroke [1]. Additionally, diabetes is the cerebrovascular risk factor associated with greater in-hospital mortality both in patients with ischemic stroke [2-5] and intracerebral hemorrhage [6]. However, little is known regarding the clinical pattern, outcome, and predictors of early mortality after an ischemic stroke in people with diabetes. To improve our knowledge of ischemic stroke in diabetes, we carried out a clinical study of patients with diabetes and cerebral infarction collected from an hospital-based stroke registry with the following objectives: 1) to compare demographic data, clinical variables, stroke subtypes, and prognostic features of ischemic stroke in patients with diabetes and without diabetes; and 2) to determine the independent predictors of in-hospital mortality in people with diabetes and ischemic stroke.
Methods
Study population
Between January 1986 and December 1997, data of 2,500 acute stroke patients admitted consecutively to the Department of Neurology of Sagrat Cor-Hospital of Barcelona, Spain were collected prospectively in a stroke registry.[7] Our institution is an acute care 350-bed university-affiliated hospital in the city of Barcelona and serves an urban population of approximately 250,000 people. The large majority of people are Caucasian. All patients with cerebrovascular disease are initially attended to in the emergency department and are then admitted to the Department of Neurology, which has 25 beds and acute stroke unit. Intensive care unit beds are also available. Patients are chosen for admission to the Department of Neurology if the reason for consultation is an acute cerebrovascular event occurring independently of the presence or absence of severe concomitant medical problems. Patients with transient ischemic attack (TIA) or reversible neurologic deficits who are evaluated on an outpatient basis are routinely referred to the emergency department for assessment and included in the registry. Thus, the proportion of patients experiencing minor strokes who are not treated at the hospital is negligible. Subtypes of stroke were classified according to the Cerebrovascular Study Group of the Spanish Neurological Society [8], which is similar to the National Institute of Neurological Disorders and Stroke Classification [9] and has been used by our group in previous studies [2,10,11]. Subtypes of cerebrovascular accident included transient ischemic attack (TIA), atherothrombotic stroke (n = 553), lacunar stroke (n = 484), cardioembolic infarction (n = 468), infarction of undetermined origin (n = 248), infarction of unusual cause (n = 87), intracerebral hemorrhage, subarachnoid hemorrhage, spontaneous subdural hematoma, and spontaneous epidural hematoma. For the purpose of this study, the group of 1,840 patients with cerebral infarction was selected.
All patients were admitted to the hospital within 48 hours of onset of symptoms. On admission, demographic characteristics, salient features of clinical and neurological examination and results of laboratory tests (blood cell count, biochemical profile, serum electrolytes, and urinalysis), chest radiography, and twelve-lead electrocardiography were recorded. Neurological examination was performed on a daily basis. In all diabetic patients, brain computed tomography scan was performed within this first week of hospital admission. Cardiac investigations included electrocardiogram in 100% (n = 393) of patients and echocardiography in 32% (n = 125). Carotid investigations consisted of Doppler and/or angio-magnetic resonance imaging in 63.9% (n = 251) of patients, arterial digital angiography in 7% (n = 28), and conventional angiography 6.5% (n = 25). Lumbar puncture was performed in 2% of cases. As used in previous studies [2,6], diabetic patients were those with known diabetes, treated with either insulin or oral hypoglycemic agents or not treated, whatever the plasma glucose level at stroke onset. Patients with post-stroke repeated fasting plasma glucose levels >7.8 mmol/L (140 mg/dL) were enrolled in accordance with the World Health Organization diagnostic criteria for diabetes used in 1993 [12]. Patients with reactive hyperglycemia were excluded. Serum determination of HbA1c was performed in doubtful cases to diagnose previous diabetes.
For each patient, demographic data (age and sex), vascular risk factors, clinical features, neuroimaging findings, and outcome were recorded. Anamnestic findings consisted of history of hypertension, myocardial infarction or angina, rheumatic heart disease, congestive heart failure, atrial fibrillation, smoking, alcohol abuse, intermittent claudication, TIA, previous cerebral infarction, hyperlipidemia, chronic nephropathy, cirrhosis or chronic liver disease, chronic obstructive pulmonary disease (COPD), and age of 85 years or older. Clinical variables, dichotomized as present versus absent, included sudden onset of symptoms (< 60 min), acute onset (1–24 hours), and subacute onset (> 24 hours); headache; dizziness; seizures; nausea or vomiting; altered consciousness; limb weakness; sensory symptoms; aphasia or dysarthria; ataxia; cranial nerve palsy; and presence of lacunar syndrome (pure motor hemiparesis, pure sensory stroke, sensorimotor stroke, ataxic hemiparesis, and dysarthria-clumsy hand) [13,14]. Topographic diagnoses based on data on neuroimaging variables also dichotomized as present versus absent, included middle cerebral, posterior cerebral, anterior cerebral artery involvement, and basilar and vertebral artery involvement, and frontal, parietal, temporal, occipital lobes, internal capsule, basal ganglia, thalamus, midbrain, and pons location. Outcome variables included infectious complications and cardiac, respiratory, and vascular events. Causes of death were analyzed according to criteria of Silver et al. [15]. The degree of clinical disability at hospital discharge was evaluated according to the scale recommended by the Ad Hoc Committee [16] and the modified Rankin Scale [17].
Statistical analysis
Univariate analysis for each variable in relation to vital status at discharge (alive, dead) as well as differences in the frequency of demographic characteristics, vascular risk factors, clinical events, neuroimaging data, and outcome between ischemic stroke patients with and without diabetes were assessed with the Student's t test and the chi-square test (χ2) (with Yate's correction when necessary), and the analysis of variance. Statistical significance was set at P < 0.05. Variables associated with ischemic stroke in patients with diabetes in the univariate analysis were subjected to multivariate analysis with a logistic regression procedure and forward stepwise selection if P < 0.10. Ischemic stroke in diabetic patients coded as absent = 0, present = 1, was the dependent variable. A first predictive model was based on demographic variables and vascular risk factors (total 8 variables). In addition to these variables, clinical features and ischemic stroke subtypes, --dichotomized as present versus absent--, were included in a second model (total 17 variables). In addition to these variables, neuroimaging data, --also dichotomized as present versus absent--, were included in a third model (total 22 variables). In all cases, the level of significance to remain in the model was 0.15. The tolerance level was established as 0.0001. The maximum likelihood approach was used to estimate weights of the logistic parameters. Odds ratio (OR) and 95% confidence intervals (CI) were calculated from the beta coefficients and standard errors. The hypothesis that the logistic model adequately fits the data was tested by means of the goodness-of-fit χ2 test [18]. The area under the receiver operating characteristics (ROC) curve for each predictive model was calculated [19]. Cox proportional-hazards models were used to estimate the relative risk of (RR) mortality (with discharge alive for hospital as the censoring event and death in hospital as the event of interest) after adjusting for age, sex, cerebrovascular risk factors and clinical findings, both in the group of diabetic and non-diabetic stroke patients. Data are expressed as RR and 95% confidence intervals (CIs). The SPSS-PC+ [20] and BMDP [21] computer programs were used for statistical analysis.
Results
Diabetes was identified in 393 (21.3%) of the 1,840 ischemic stroke patients. A total of 51.4% of patients were men. The mean age of diabetic patients with ischemic stroke was 73.6 ± 9.8 years. Cardiovascular risk factors included hypertension in 53.2% of cases atrial fibrillation in 26.4%, hyperlipidemia in 15.3%, ischemic heart disease in 18.6%, chronic nephropathy in 4.8%, and congestive heart failure in 2.7%. The frequency of stroke subtypes was as follows: atherothrombotic infarction in 41.2% of patients, lacunar infarction in 35.1%, presumed cardioembolic stroke in 18.6%, infarction of unknown etiology in 4.3%, and infarction of unusual cause in 0.8%. The in-hospital mortality was rate was 12.5% (n = 50). Causes of death included cerebral herniation in 17 patients, cardiac events in 11, respiratory events in 10, sepsis in 4, sudden death in 1, and unknown cause in 7. Symptom-free at hospital discharge was observed in 18.1% of patients. The median hospital stay was 13 days (interquartile range 8–21).
Differential features between people with and without diabetes with ischemic stroke patients are shown in Table 1. Patients with diabetes compared to patients without diabetes (n = 1,447) were more likely to have ischemic heart disease, previous cerebral infarction, peripheral vascular disease, hyperlipidemia, subacute stroke onset, atherothrombotic and lacunar infarctions, and thalamus, pons and cerebral posterior artery involvement. On the other hand, they were less likely to be 85 years or older and to have valvular heart disease, sudden stroke onset, seizures, cardioembolic infraction, stroke of unusual cause, stroke of unknown etiology, and parietal and temporal lobe involvement. There were no differences in medications used for the treatment of patients (patients with diabetes vs patients without diabetes) antiplatelets 84.5% vs 85.4%, anticoagulants at therapeutic doses (15.5% vs. 14.6%), antibiotic therapy (14.5% vs 12.6%), and other medical treatments (96% vs 94%). After multivariate analysis (Table 2), ischemic heart disease, hyperlipidemia, atherothrombotic and lacunar infarcts, subacute onset and thalamic infarcts were independently associated with diabetes in patients with ischemic stroke. Age of 85 years or older was inversely associated.
Table 1 Results of univariate analysis in diabetic patients with cerebral infarction as compared with patients with cerebral infarction without diabetes mellitus
Variable Patients with diabetes (n = 393) Patients without diabetes (n = 1,447) P value
Men 202 (51.4) 724 (50) 0.6313
Age, years, mean (SD) 73.6 (9.8) 74 (12.7) 0.4945
Age, 85 year or more 49 (12.5) 254 (17.6) 0.0159
Cardiovascular risk factors
Valvular heart disease 19 (4.8) 111 (7.7) 0.0517
Ischemic heart disease 73 (18.6) 193 (13.3) 0.0088
Previous cerebral infarction 77 (19.6) 223 (15.4) 0.0466
Peripheral vascular disease 39 (9.9) 103 (7.1) 0.0646
Hypelipidemia 89 (22.6) 241 (16.7) 0.0060
Clinical findings
Sudden onset 183 (46.3) 764 (52.8) 0.0225
Subacute onset (days) 40 (10.2) 90 (6.2) 0.0066
Seizures 3 (0.8) 31 (2.1) 0.0718
Lacunar syndrome 156 (39.7) 349 (27.2) <0.0001
Ischemic stroke subtypes
Atherothrombotic 162 (41.2) 391 (27) <0.0001
Cardioembolic 73 (18.6) 395 (27.3) 0.0004
Lacunar infarct 138 (35.1) 346 (23.9) <0.001
Unusual etiology 3 (0.8) 84 (5.8) <0.0001
Unknown cause 17 (4.3) 231 (16) <0.0001
Localization of cerebral infarction
Parietal 74 (18.8) 339 (23.4) 0.0527
Temporal 79 (20.1) 380 (26.3) 0.0123
Thalamus 44 (11.2) 88 (6.1) 0.0005
Pons 30 (7.6) 71 (4.9) 0.0353
Cerebral posterior involvement 52 (13.2) 128 (8.8) 0.0095
Outcome
Symptom free at discharge 71 (18.1) 272 (18.8) 0.4324
Hospital stay, days, median (interquartile range) 13 (8 to 21) 12 (8 to 21) 0.9752
In-hospital death 50 (12.5) 211 (14.6) 0.3488
Data as number and percentages in parenthesis except for the mean (SD) age and length of hospitalization.
Table 2 Independent factors associated with diabetes in patients with ischemic stroke.
Logistic regression models β SE (β) Odds ratio (95% CI)
Demographic, and vascular risk factors *
Ischemic heart disease 0.386 0.152 1.47 (1.09–1.98)
Hyperlipidemia 0.354 0.141 1.42 (1.08–1.88)
85 years old or more -0.351 0.169 0.70 (0.50–0.98)
Demographic, vascular risk factors, clinical variables and ischemic stroke subtypes†
Atherothrombotic infarct 1.910 0.250 6.75 (4.14–11.02)
Lacunar infarct 1.900 0.254 6.68 (4.07–10.99)
Subacute onset (days) 0.590 0.213 1.80 (1.19–2.74)
85 years old or more -0.391 0.172 0.68 (0.48–0.95)
Demographic, vascular risk factors, clinical variables and neuroimaging data‡
Atherothrombotic 1.929 0.250 6.88 (4.21–11.24)
Lacunar infarct 1.848 0.255 6.35 (3.85–10.46)
Subacute onset (days) 0.601 0.213 1.82 (1.20–2.77)
Thalamic topography 0.557 0.204 1.74 (1.17–2.61)
85 years old or more -0.381 0.172 0.683 (0.49–0.96)
* β = -1.381, SE (β) = 0.074, goodness of fit χ2 = 0.434, df = 3, P = 0.933. Area under the ROC curve = 0.563, sensitivity 28%, specificity 81%, correctly classified 67%.
† β = -2.802, SE (β) = 0.235, goodness of fit χ2 = 2.133, df = 4, P = 0.711. Area under the ROC curve = 0.663, sensitivity 29%, specificity 88%, correctly classified 55%.
‡ β = -2.847, SE (β) = 0.236, goodness of fit χ2 = 3.151, df = 5, P = 0.677. Area under the ROC curve = 0.671, sensitivity 30%, specificity 50%, correctly classified 42%.
The characteristics of ischemic stroke patients with diabetes according to vital status at discharge are shown in Table 3. Patients who died (n = 50) compared to those who were discharged alive from the hospital (n = 343) had a significantly higher occurrence of the following variables: age 85 years or older, atrial fibrillation, congestive heart disease, chronic kidney disease, sudden and acute stroke onset, seizures, decreased consciousness, limb weakness, sensory deficit, hemianopia, atherothrombotic stroke, cardioembolic infarction, parietal, temporal, internal capsule, mesencephalon and pons topography, basilar and middle cerebral artery involvement, and cardiac, respiratory, urinary, digestive and infectious complications.
Table 3 Results of univariate analysis in 393 diabetic patients with cerebral infarction according to vital status at discharge
Variable Dead (n = 50) Alive (n = 343) P value
Men 18 (36) 184 (53.6) 0.0197
Age, years, mean (SD) 78.4 (7.2) 72.9 (9.9 0.0002
Age, 85 years or more 13 (26) 36 (10.5) 0.0019
Cardiovascular risk factors
Atrial fibrillation 26 (52) 78 (22.7) <0.0001
Congestive heart disease 6 (12) 13 (3.8) 0.0296
Chronic nephropathy 4 (8) 7 (2) 0.0389
Clinical findings
Sudden onset 32 (64) 150 (43.7) 0.0072
Acute onset (hours) 11 (22) 122 (35.6) 0.0582
Seizures 2 (4) 1 (0.3) 0.0518
Decreased consciousness 31 (62) 29 (8.5) <0.0001
Limb weakness 48 (96) 249 (72.6) 0.0003
Sensory symptoms 33 (66) 130 (37.9) 0.0002
Hemianopia 17 (34) 54 (15.7) 0.0017
Lacunar syndrome 0 156 (45.5) <0.0001
Ischemic stroke subtypes
Atherothrombotic 29 (58) 1333 (38.8) 0.0988
Cardioembolic 19 (38) 54 (15.7) 0.0016
Lacunar infarct 0 138 (40.2) <0.0001
Unusual etiology 0 3 (0.9) 1.0000
Unknown cause 2 (4) 15 (4.4) 1.0000
Localization of cerebral infarction
Parietal 24 (48) 50 (14.6) <0.0001
Temporal 22 (44) 57 (16.6) 0.0001
Internal capsule 4 (8) 77 (22.4) 0.0183
Mesencephalon 4 (8) 1 (0.3) <0.0001
Pons 8 (16) 22 (6.4) 0.0171
Middle cerebral artery involvement 31 (62) 166 (48.4) 0.0723
Basilar artery involvement 9 (18) 23 /6.7) 0.0142
Outcome
Respiratory complications 19 (38) 19 (5.5) <0.0001
Digestive complications 4 (8) 5 (1.5) 0.0172
Urinary infections 9 (18) 30 (8.7) 0.0732
Cardiac complications 15 (30) 8 (2.3) <0.0001
Infectious complications 16 (32) 33 (9.6) <0.0001
Hospital stay, days, median (interquartile range) 13 (8 to 21) 12 (8 to 21) 0.9752
Data as number and percentages in parenthesis except for median (interquartile range) for hospitalization.
The relative risk for mortality in the groups of diabetic and non-diabetic ischemic stroke patients is shown in Table 4. Congestive heart disease, atrial fibrillation, decreased consciousness, and age were significantly adversely associated with outcome after acute ischemic stroke in both diabetic and non-diabetic ischemic stroke patients. Chronic nephropathy was a predictor of in-hospital mortality for the group of diabetic ischemic stroke patients, whereas limb weakness, nausea/vomiting, and seizures were predictors of in-hospital mortality for the group of non-diabetic stroke patients.
Table 4 Risk for in-hospital mortality in ischemic stroke patients with diabetes and without diabetes. Cox's proportional hazards model.
Predictors of death Hazard ratio P value 95% CI
Diabetic stroke patients
Demographic variables and vascular risk factors (model 1)
Chronic nephropathy 5.78 0.001 1.99–16.80
Congestive heart disease 2.52 0.045 1.02–6.23
Atrial fibrillation 2.39 0.003 1.36–4.20
Age 1.06 0.002 1.02–1.10
Demographic variables, vascular risk factors, clinical variables, and stroke subtypes (model 2)
Decreased consciousness 5.16 0.000 2.82–9.42
Congestive heart disease 4.08 0.002 1.65–10.09
Chronic nephropathy 3.27 0.036 1.08–9.89
Atrial fibrillation 2.41 0.002 1.38–4.23
Non-diabetic stroke patients
Demographic variables and vascular risk factors (model 1)
Congestive heart disease 2.00 0.001 1.35–2.96
Atrial fibrillation 1.83 0.000 1.38–2.43
Ischemic heart disease 1.74 0.002 1.23–2.58
Age 1.04 0.000 1.03–1.06
Demographic variables, vascular risk factors, clinical variables, and stroke subtypes (model 2)
Decreased consciousness 4.58 0.000 3.39–6.19
Limb weakness 1.93 0.038 1.04–3.59
Nausea/vomiting 1.79 0.009 1.16–2.76
Seizures 1.75 0.047 1.01–3.03
Congestive heart disease 1.74 0.006 1.17–2.59
Ischemic heart disease 1.72 0.003 1.02–1.80
Atrial fibrillation 1.35 0.039 1.02–1.80
Age 1.03 0.000 1.02–1.05
Discussion
In this hospital-based study of 1,840 consecutive patients with acute ischemic stroke, the prevalence of diabetes was 21%, a figure similar to that reported in the studies of Megherbi et al. [22] and Jorgensen et al. [4] and higher than the prevalence of diabetes in the Spanish population (6–7%) [23]. This may be explained by a stronger disposition to stroke in the diabetic patient, because diabetes mellitus is associated with accelerated atherogenesis [24] and also due to the fact that diabetic stroke patients have more often other cerebrovascular risk factors as, in our study, hyperlipidemia and ischemic heart disease, which in turn were independent predictors of ischemic stroke in the diabetic population as previously reported by others [25,26]. The present results are consistent with the study of Lehto et al. [25] in which hyperlipidemia was a strong predictor of stroke in middle-aged patients with non-insulin-dependent diabetes, probably because lipid abnormalities have been shown to be associated with cerebral atherosclerosis [1,24].
With regard to data of our study in comparison to previous works [3-5,22,26-31], Table 5 shows that there are only two previous studies [29,31] similar to ours. In the study of Kiers et al. [29], however, stroke subtypes are differentiated according to topography (cortical, lacunar, striatocapsular, brainstem/cerebellar) and in the study of Hamidon and Raymond [31] according to vascular topography (anterior, middle, and posterior artery involvement) and not according to cause (cardioembolic, atherothrombotic, lacunar, unusual etiology, undetermined cause) as in our report. In these two previous studies 50 and 90 patients were included as compared with 393 in our study. A strength of the present study is that predictors of in-hospital mortality in patients with diabetes and ischemic stroke were determined using logistic regression analysis, but a limitation is that a follow-up survival analysis is lacking. On the other hand, with a lower cut-point for the definition of hypertension (BP > 130/85 mm Hg), the prevalence of hypertension would have been probably substantially greater.
Table 5 Ischemic stroke and diabetes. Series reported in the literature
1st author, year No. patients with ischemic stroke and diabetes Type of study Ischemic stroke subtypes Predictors of in-hospital death Follow-up survival analysis
Asplund [27], 1980 53 Retrospective No No No
Oppenheimer [5], 1985 14 Retrospective No No No
Lithner [26], 1988 75 Retrospective No No No
Woo [28], 1990 47 Cross-sectional Yes No No
Olsson [3], 1990 121 Prospective No No Yes (0–10 yr)
Kiers [29], 1992 50 Prospective Yes Yes No
Jorgensen [4], 1994 233 Prospective, community-based stroke registry No Yes No
Weir [30], 1997 61 Prospective No No Yes (3 mo)
Megherbi [22], 2003 937 Prospective multicenter No No Yes (3 mo)
Hamidon [31], 2003 90 Prospective No Yes No
Present series, 2004 393 Prospective, hospital.-based stroke registry Yes Yes No
Ischemic heart disease was another independent predictor of ischemic stroke in diabetic patients. It could suggest that diabetic patients have concurrent vascular lesions in the heart and the brain due to widespread atherosclerotic disease. In the study of Manson et al. [32], maturity-onset clinical diabetes was a independent risk factor for coronary heart disease.
Subacute stroke onset (> 24 hours) was found to be a predictive clinical factor in diabetic ischemic stroke. This may be because thrombotic occlusion usually is gradual, and thrombotic infarcts show more frequently a fluctuating or progressive clinical course [33]. In contrast, sudden stroke onset is more characteristic of cardioembolic infarction and was observed in 83% of cases in the series of Bogousslavsky et al. [34] and in 79% of patients reported by Mohr et al. [35].
In our diabetic group, the distribution of pathological subtypes of ischemic stroke showed a higher occurrence of atherothrombotic and lacunar infarctions compared to nondiabetic ischemic stroke patients. Atherothrombotic infarcts were the most frequent stroke subtype (41.2%). This high frequency may be related to the increased susceptibility to atherosclerosis and the accelerated atherogenesis associated with diabetes mellitus.[1,24] Other authors found that diabetes was more frequently associated with angiographically demonstrated extracranial atherosclerotic carotid artery occlusion and atherosclerotic occlusive disease of the basilar artery, with a strong association between diabetes and carotid artery intimal-medial thickness [1,24]. The fact that we observed more lacunar infarcts in diabetic patients has been reported by our group in a previous study[2] and has been also observed in another recent series [22]. Diabetes can cause small vessel arteriolopathy, especially in the retina, kidney and brain (mainly in thalamus, internal capsule and pons topography) [36,37]. In a classical autopsy study of cerebrovascular accident in diabetes mellitus, Alex et al. [38] found that small vessel cerebral disease was present about 2.5 times more frequent in diabetes.
At discharge, the case fatality rate in the two groups of diabetic and nondiabetic ischemic stroke patients was comparable (12.5% and 14.6%, respectively). In our opinion, the increased occurrence of lacunar infarcts in diabetic patients, with a known good prognosis [36,37] (0% of in-hospital mortality in our study), may account for the lack of differences in early mortality. Functional recovery after lacunar infarcts as well as survival has been found to be favorable [38-40] and in a recent study [41], lacunar stroke had an odds ratio of 3.1 for an excellent outcome at 3 months.
In the multivariate analysis, independent clinical factors related to in-hospital mortality in diabetic patients with ischemic stroke were age, atrial fibrillation, congestive heart failure, chronic nephropathy and altered consciousness. Baseline plasma glucose level, which is a variable associated with poor outcome in stroke was not assessed. In the present study, like others [42], decreased consciousness and age were important clinical predictors of early mortality. In the study of Hamidon and Raymond [31], middle cerebral artery territory infarct and poor conscious level were independent predictors of mortality. Atrial fibrillation was another major aggravating factor in this population. Atrial fibrillation increases the risk of early recurrent stroke substantially and patients with atrial fibrillation may have larger infarcts [43]. In addition, atrial fibrillation may be a cause of more severe handicap through more severe motor or sensory deficits. Diabetic patients who died also tended to show a more severe neurological impairment at onset, characterized by a predominance of motor deficit and decreased consciousness, which is similar to results of the study of Olsson et al. [3]. The presence of congestive heart failure was another significant prognostic factor. This finding is similar to other studies [44,45] who reported a high mortality among patients with embolic stroke and ischemic heart diseases or congestive heart failure. Our results agree with the study of MacWalter et al. [46] who demonstrate that renal dysfunction had a higher mortality risk after acute stroke. Renal failure is a very rare primary cause of death in acute stroke [47]. However renal dysfunction represents the influence of generalized vascular disease in the kidney and is a potent predictor of in-hospital mortality in acute ischemic stroke diabetic people.
Conclusion
In the present series of ischemic stroke patients with diabetes collected from a prospective hospital-based stroke registry, the clinical picture of these patients was characterized by a more frequent concomitant ischemic heart disease and hyperlipidemia and a more frequent presence of atherothrombotic and lacunar infarcts as compared with ischemic stroke in people without diabetes. The in-hospital mortality is related to the presence of causal factors for stroke, including a more diffuse atherosclerotic disease and atrial fibrillation, and a higher frequency of other factors including a more advanced age, a strategic ischemic stroke in midbrain or large cortical topography, and the occurrence of cardiac or respiratory complications after stroke.
List of abbreviations
CI: confidence interval
COPD: chronic obstructive pulmonary disease
HbA1c: hemoglobin A, glycosylated
OR: odds ratio
ROC: receiver operating characteristics
RR: relative risk
TIA: transient ischemic attack
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
A. Arboix, was the principal investigator, chief of the Cerebrovascular Division, designed the study, diagnosed and took care of the patients, contributed to analyze the data, interpreted the results, wrote the paper, and prepared the final draft. He was also responsible for editorial decisions including the selection of the target journal.
A. Rivas participated in the collection of data, search and review of the literature, analysis of results, review of the manuscript, and approved the final draft.
L. García-Eroles was the statistician, participated in the study design, analysis and interpretation of data, wrote the part of the paper related to the statistical analysis, and approved the final draft.
L. de Marcos, J. Massons, and M. Oliveres diagnosed and took care of the patients, contributed in the review of the literature, interpretation of the results, review of the paper for intellectual content, and approved the final draft.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Drs. C. Fornós from the Service of Internal Medicine and M. Balcells, E. Comes, and C. Targa from the Service of Neurology for their assistance in the study, and Dr. Marta Pulido for editing the manuscript and editorial assistance.
==== Refs
Lukovits TG Mazzone T Gorelick PB Diabetes mellitus and cerebrovascular disease Neuroepidemiology 1999 18 1 14 9831810 10.1159/000026190
Arboix A Morcillo C García-Eroles L Massons J Oliveres M Targa C Different vascular risk factor profiles in ischemic stroke subtypes. The Sagrat Cor Hospital of Barcelona Stroke Registry Acta Neurol Scand 2000 102 264 270 11071113 10.1034/j.1600-0404.2000.102004264.x
Olsson T Viitanen M Asplund K Eriksson S Hägg E Prognosis after stroke in diabetic patients. A controlled prospective study Diabetologia 1990 33 244 249 2347437 10.1007/BF00404803
Jorgensen HS Nakayama H Raaschou HO Olsen TS Stroke in patients with diabetes. The Copenhagen Stroke Study Stroke 1994 25 1977 1984 8091441
Oppenheimer SM Hoffbrand BI Oswald GA Yudkin JS Diabetes mellitus and early mortality from stroke BMJ 1985 291 1014 1015 3931771
Arboix A García-Eroles L Massons J Oliveres M Targa C Diabetes mellitus is a risk factor for in-hospital mortality in intracerebral hemorrhage Diabetes Care 2000 23 1527 1532 11023147
Arboix A Vericat MC Pujades R Massons J García-Eroles L Oliveres M Cardioembolic infarction in The Sagrat Cor-Alianza Hospital of Barcelona Stroke Registry Acta Neurol Scand 1997 96 407 412 9449481
Arboix A Alvarez-Sabín J Soler L for the Cerebrovascular Study Group of the Spanish Society of Neurology Nomenclatura de las enfermedades vasculares cerebrales Neurologia 1998 1 10
Special Report from the National Institute of Neurological Disorders and Stroke Classification of cerebrovascular diseases. III Stroke 1990 21 637 676 2326846
Arboix A García-Eroles L Massons J Oliveres M Targa C Hemorrhagic lacunar stroke Cerebrovasc Dis 2000 10 229 234 10773650 10.1159/000016061
Arboix A Massons J García-Eroles L Oliveres M Balcells M Targa C Migrainous cerebral infarction in the Sagrat-Cor Hospital of Barcelona Stroke Registry Cephalalgia 2003 23 389 394 12780770 10.1046/j.1468-2982.2003.00534.x
World Health Organization WHO Study Group on Diabetes Mellitus 1985 Geneva, Switzerland: World Health Organization 103 108 Technical Report Series no 722
Arboix A Martí-Vilalta JL García JH Clinical study of 227 patients with lacunar infarcts Stroke 1990 21 842 847 2349585
Arboix A Padilla I García-Eroles Massons Comes E Targa C Clinical study of 222 patients with pure motor hemiparesis J Neurol Neurosurg Psychiatry 2001 71 239 242 11459902 10.1136/jnnp.71.2.239
Silver FL Norris JW Lewis AJ Hachinski VC Early mortality following stroke: a prospective review Stroke 1984 15 492 496 6729878
Ad Hoc Committee A classification and outline of cerebrovascular diseases Stroke 1975 6 565 616
Bamford JM Sandercock PAG Warlow CP Slattery J Interobserver agreement for the assessment of handicap in stroke patients Stroke 1989 20 828 2728057
Hosmer DW Lemershow S Goodness of fit tests for the multiple logistic regression model Commun Stat 1980 A9 1043 1069
Burgueño MJ García-Bastos JL González-Buitrago JM Las curves ROC en la evaluación de las pruebas diagnósticas Med Clin (Barc) 1995 104 661 670 7623495
Norusis MJ SPSS Advanced Statistical Student Guide 1990 Chicago: SPSS, Inc
Dixon WJ BMDP Statistical Software Manual 1990 Berkeley, CA: University of California Press 330 334
Megherbi SE Milan C Minier D Couvreur G Osseby GV Tilling K Di Carlo A Inzitari D Wolfe CD Moreau T Giroud M European BIOMED Study of Stroke Care Group Association between diabetes and stroke subtype on survival and functional outcome 3 months after stroke. Data from the European BIOMED Stroke Project Stroke 2003 34 688 694 12624292 10.1161/01.STR.0000057975.15221.40
Vázquez JA Gaztambide S Soto-Pedre E Estudio prospectivo de 10 años sobre la incidencia y factores de riesgo de diabetes mellitus tipo 2 Med Clin (Barc) 2000 115 534 539 11141379
Bell DSH Stroke in the diabetic patient Diabetes Care 1994 17 213 219 8174450
Lehto S Rönnemaa T Pyörälä K Laakso M Predictors of stroke in middle-aged patients with non-insulin-dependent diabetes Stroke 1996 27 63 68 8553405
Lithner F Asplund K Eriksson S Hägg E Strand T Wester PO Clinical characteristics in diabetic stroke patients Diabete Metab 1998 14 15 19 3391327
Asplund K Hagg E Helmers C Lithner F Strand T Wester PO The natural history of stroke in diabetic patients Acta Med Scand 1980 207 417 424 7386235
Woo J Lam CW Kay R Wong AH Teoh R Nicholls MG The influence of hyperglycemia and diabetes mellitus on immediate and 3-month morbidity and mortality after acute stroke Arch Neurol 1990 47 1174 1177 2241612
Kiers L Davis SM Larkins R Hopper J Tress B Rossiter SC Carlin J Ratnaike S Stroke topography and outcome in relation to hyperglycaemia and diabetes J Neurol Neurosurg Psychiatry 1992 55 263 270 1583510
Weir CJ Murray GD Dyker AG Lees KR Is hyperglycaemia an independent predictor of poor outcome after acute stroke? Results of a long-term follow up study BMJ 1997 314 1303 1306 9158464
Hamidon BB Raymond AA The impact of diabetes mellitus on in-hospital stroke mortality J Postgrad Med 2003 49 307 310 14699227
Manson JE Colditz GA Stampfer MJ Willett WC Krolewski AS Rosner B Arky RA Speizer FE Hennekens CH A prospective study of maturity-onset diabetes mellitus and risk of coronary heart disease and stroke in women Arch Intern Med 1991 151 1141 1147 2043016 10.1001/archinte.151.6.1141
Wilterdink JL Easton D Vascular event rates in patients with atherosclerotic cerebrovascular disease Arch Neurol 1992 49 857 863 1524519
Bogousslavsky J Cachin C Regli F Despland PA Van Melle G Kappenberger L Cardiac sources of embolism and cerebral infarction. Clinical consequences and vascular concomitants: The Lausanne Stroke registry Neurology 1991 41 855 859 2046930
Mohr JP Caplan LR Melski JW Goldstein RJ Duncan GW Kistler JP Pessin MS Bleich HL The Harvard Cooperative Stroke Registry: a prospective registry Neurology 1978 28 754 762 567291
Fisher CM Lacunar strokes and infarcts: a review Neurology 1982 32 871 876 7048128
Fisher CM Lacunar infarcts. A review Cerebrovasc Dis 1991 1 311 320
Alex M Baron EK Goldenberg S Blumenthal HT An autopsy study of cerebrovascular accident in diabetes mellitus Circulation 1962 25 663 673 13860317
Rothrock JF Clark Wm Lyden PD Spontaneous early improvement following ischemic stroke Stroke 1995 26 1358 1360 7631337
Libman RB Sacco RL Shi T Tatemichi TK Mohr JP Neurologic improvement in pure motor hemiparesis: implications for clinical trials Neurology 1992 42 1713 1716 1513458
Adams HP JrDavis PH Leira EC Chang KC Bendixen BH Clarke WR Woolson RF Hansen MD Baseline NIH Stroke Scale score strongly predicts outcome after stroke. A report of the trial of Org 10172 in acute stroke treatment (TOAST) Neurology 1999 53 126 131 10408548
Bogousslavsky J van Melle G Regli F The Lausanne Stroke Registry: an analysis of 1,000 consecutive patients with first stroke Stroke 1988 19 1083 1092 3413804
Candelise L Pinardi G Morabito A the Italian Acute Stroke Study Group Mortality in acute stroke with atrial fibrillation Stroke 1991 22 169 174 2003280
Broderick JP Phillips SJ O'Fallon WM Frye RL Whisnant JP Relation of cardiac disease to stroke occurrence, recurrence, and mortality Stroke 1992 23 1250 1256 1519279
Frithz G Werner I Studies on cerebrovascular strokes. II. Clinical findings and short-term prognosis in a stroke material Acta Med Scand 1976 199 133 140 1251767
MacWalter RS Wong SY Wong KY Stewart G Fraser CG Fraser HW Ersoy Y Ogston SA Chen R Does renal dysfunction predict mortality after acute stroke? A 7-year follow-up study Stroke 2002 33 1630 1635 12053003 10.1161/01.STR.0000016344.49819.F7
Wannamethee SG Shaper AG Perry IJ Serum creatinine concentration and risk of cardiovascular disease: a possible marker for increased risk of stroke Stroke 1997 28 557 563 9056611
| 15833108 | PMC1097737 | CC BY | 2021-01-04 16:28:53 | no | BMC Neurol. 2005 Apr 15; 5:9 | utf-8 | BMC Neurol | 2,005 | 10.1186/1471-2377-5-9 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-301585048910.1186/1471-2202-6-30Research ArticleGABAA receptor γ2 subunit knockdown mice have enhanced anxiety-like behavior but unaltered hypnotic response to benzodiazepines Chandra Dev [email protected] Esa R [email protected] Celia P [email protected] Blas Angel L [email protected] Gregg E [email protected] Department of Anesthesiology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA2 Department of Pharmacology, Institute of Biomedicine, Biomedicum Helsinki, University of Helsinki, Helsinki, Finland3 Department of Physiology and Neurobiology, University of Connecticut, Storrs, CT, USA4 Department of Pharmacology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA2005 25 4 2005 6 30 30 28 1 2005 25 4 2005 Copyright © 2005 Chandra 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
Gamma-aminobutyric acid type A receptors (GABAA-Rs) are the major inhibitory receptors in the mammalian brain and are modulated by a number of sedative/hypnotic drugs including benzodiazepines and anesthetics. The significance of specific GABAA-Rs subunits with respect to behavior and in vivo drug responses is incompletely understood. The γ2 subunit is highly expressed throughout the brain. Global γ2 knockout mice are insensitive to the hypnotic effects of diazepam and die perinatally. Heterozygous γ2 global knockout mice are viable and have increased anxiety-like behaviors. To further investigate the role of the γ2 subunit in behavior and whole animal drug action, we used gene targeting to create a novel mouse line with attenuated γ2 expression, i.e., γ2 knockdown mice.
Results
Knockdown mice were created by inserting a neomycin resistance cassette into intron 8 of the γ2 gene. Knockdown mice, on average, showed a 65% reduction of γ2 subunit mRNA compared to controls; however γ2 gene expression was highly variable in these mice, ranging from 10–95% of normal. Immunohistochemical studies demonstrated that γ2 protein levels were also variably reduced. Pharmacological studies using autoradiography on frozen brain sections demonstrated that binding of the benzodiazepine site ligand Ro15-4513 was decreased in mutant mice compared to controls. Behaviorally, knockdown mice displayed enhanced anxiety-like behaviors on the elevated plus maze and forced novelty exploration tests. Surprisingly, mutant mice had an unaltered response to hypnotic doses of the benzodiazepine site ligands diazepam, midazolam and zolpidem as well as ethanol and pentobarbital. Lastly, we demonstrated that the γ2 knockdown mouse line can be used to create γ2 global knockout mice by crossing to a general deleter cre-expressing mouse line.
Conclusion
We conclude that: 1) insertion of a neomycin resistance gene into intron 8 of the γ2 gene variably reduced the amount of γ2, and that 2) attenuated expression of γ2 increased anxiety-like behaviors but did not lead to differences in the hypnotic response to benzodiazepine site ligands. This suggests that reduced synaptic inhibition can lead to a phenotype of increased anxiety-like behavior. In contrast, normal drug effects can be maintained despite a dramatic reduction in GABAA-R targets.
==== Body
Background
GABAA-Rs are the major inhibitory receptors in the mammalian central nervous system and can be modulated by a number of sedative, hypnotic and anesthetic drugs [1]. GABAA-Rs are putatively pentameric complexes and are members of a 'superfamily' of related ligand-gated ion channels. There are a variety of subunit families that make up GABAA-Rs; a total of seventeen distinct subunits have been cloned, α1–6, β1–4, γ1–3, δ, ε, π, and θ. Much is known of the subunits that comprise GABAA-Rs, but the contribution of individual subunits to in vivo drug responses is only beginning to be understood, largely through the use of genetically engineered mice [2-4].
The γ2 subunit is highly expressed throughout developing and adult brain and spinal cord, is a component of almost 60% of all GABAA-Rs [5,6], and is required for synaptic clustering of GABAA-Rs [7]. In vitro electrophysiologic studies have demonstrated an obligatory role of the γ2 subunit for benzodiazepine responsiveness [8]. Gunther et al. created and studied mice that globally lacked all γ2 subunits of the GABAA-R [9]. Mice homozygous for the γ2 knockout died in the perinatal period with rare survivors reaching postnatal day 18. Knockouts appeared normal at birth, but developed sensorimotor abnormalities characterized by hyperactivity, impaired grasping and righting reflex, and abnormal gait. Pharmacologically, knockout of γ2 resulted in a severe deficit in benzodiazepine binding sites (~94% reduction compared to control) while GABA binding sites were only slightly reduced. Behaviorally, in the few mice that survived long enough to be tested, diazepam failed to induce sedation and loss of the righting reflex at doses that were fully effective in wild type mice [9]. γ2 heterozygous knockout mice were normally viable, had a ~50% reduction in γ2 protein levels, and displayed increased anxiety-like behaviors [10]. Drug responses of heterozygous γ2 global knockouts have not been reported, except for the high efficacy of low diazepam doses to induce anxiolysis [10].
We set out to make mice with markedly attenuated expression of the γ2 gene (i.e., γ2 knockdown mice) that would be viable and therefore useful for studies of basic behaviors and drug induced responses. We used gene targeting and embryonic stem cell technologies to create this mouse line by inserting a neomycin resistance gene (NEO) into intron 8 of the γ2 genomic locus. Insertion of NEO into intronic DNA has previously been used to create hypomorphic alleles of several genes [11-23]. In those studies, expression of the hypomorphic allele ranged from 5–25% of control; substantially lower than the 50% reduction typically observed in heterozygous knockouts. The γ2 knockdown mouse line described in the present report had a highly variable reduction in γ2 mRNA and protein levels. γ2 knockdown mice were normally viable, had increased anxiety-like behaviors, but did not differ in the hypnotic response to benzodiazepines. In addition, since exon 8 of the γ2 gene was flanked by loxP sites in this mouse line, these mice could also be converted into a global knockout mouse by crossing to a general deleter Cre-expressing transgenic mouse line.
Results
Production of gene targeted mice
Four out of 330 embryonic stem cell clones displayed the predicted restriction fragment length polymorphisms (data not shown) indicative of the targeting event presented in Fig 1A. Because the targeting event results in a γ2 locus in which exon 8 is flanked by loxP sites, we refer to this locus as being floxed (F) in contrast to the endogenous wild type allele (+). Two correctly targeted cell lines, 26L6 and 26C8 yielded germ-line competent chimeric males. Chimeras were mated to C57BL/6J females. Offspring that were +/F were interbred to produce +/+, +/F, and F/F mice. Mice were genotyped at weaning using Southern blot analysis as indicated in Fig 1B. In this analysis, Probe A hybridized to a 9.2 kB BglII fragment from the wild type γ2 allele and a 3.4 kB fragment from the targeted allele. Of 724 mice genotyped, 196 (27%) were +/+, 372 (51%) were +/F, and 156 (22%) were F/F. These values are in accord with the 1:2:1 genotype frequency as expected by Mendelian genetics. Thus, homozygous floxed mice were normally viable. They were also healthy and overtly indistinguishable from wild type mice.
Northern blot analysis
Semi-quantitative northern blots were used to compare the amount of γ2 mRNA in wild type and F/F animals. Hybridization of adult brain RNA with a γ2 cDNA probe yielded an intense 2.8 kB band in wild type mice. In contrast, RNA from F/F mice hybridized less intensely in most mice, however the attenuation within this group was highly variable. A sample northern blot is shown in Figure 2A. We semi-quantitatively determined γ2 mRNA amounts by comparing γ2 band densities between wild type (n = 23) and F/F (n = 36) mice following normalization with β-actin. To normalize results over four different blots, the ratio of wild type band densities on each blot was averaged and normalized to 100. All samples on each blot were then compared to the average of wild type band densities. Values are shown in Figure 2B. Mean wild type ratio was 100 whereas mean ratio in F/F samples was 35 with high variability including several mice with values that overlapped with wild type values. Thus, γ2 mRNA levels in knockdown mice were dramatically attenuated on average, but the levels varied significantly between mice.
Immunohistochemistry
Immunohistochemical staining of sagittal sections probed with a γ2-specific antibody was used to assess γ2 protein amount and distribution in the brain. A sampling of wild type (Fig 3A) and knockdown (Fig 3B) brain sections are displayed. In three of the four knockdown mouse sections shown, there was a marked overall reduction of γ2 immunoreactivity in all areas of the brain except the outer layer of the olfactory bulb. However, staining intensity was variable between knockdown mice, with one having near wild type intensity (Fig. 3B, d). A total of 10 F/F mice were studied. Seven out of 10 had a marked decrease in γ2 staining compared to wild type controls, and three out of 10 had near wild-type or intermediate levels of staining. Adjacent sections from these same mice showed no change in α1 or β2/3 immunoreactivity (CPM and ALD, unpublished observations).
Pharmacological characterization
Autoradiography on brain cryostat sections with various ligands binding to selective sites of the GABAA-R revealed clear differences in the homozygous γ2 knockdowns as compared to heterozygous and wild type mice (Table 1). Total flumazenil-sensitive [3H]Ro15-4513 binding to benzodiazepine sites was reduced in the knockdown in most brain regions. The reduced binding in the 5 mice analyzed ranged on the average from 25 to 57%. In several regions, the heterozygous mice also showed significant reduction of the binding, but clearly less than the homozygous knockdown samples. A small proportion of the reduction in the total binding was apparently due to reduction in diazepam-insensitive α4 and α6 subunit-containing receptors, this component being mostly affected in the cerebellar granule cell layer.
GABA-sensitive [3H]muscimol binding to GABA agonist sites was not altered in any brain region (Table 1).
Picrotoxinin-sensitive [35S]TBPS binding to the GABAA-R channels was little affected between the mouse lines: basal binding was increased in the hippocampus and decreased in the cerebellar granule cell layer of the homozygous γ2 knockdowns (Table 1). The GABA-insensitive [35S]TBPS binding was increased all over the brain, but this component still made up less than 10% of the basal binding, except for the cerebellar granule cell layer. There, the GABA-insensitive component was 20% as compared to 10% in the heterozygous and wild type animals.
Behavioral characterization
We examined the behavioral phenotype of the knockdown mice using the elevated plus maze and forced exploration in a novel open field. The elevated plus maze is widely used to assess anxiety-like behavior in response to the aversive stimulus of an elevated exposed space. On this test apparatus, knockdown mice had a 74% reduction in open arm entries (Fig. 4A) and an 83% reduction in time on open arms (Fig. 4B) compared to controls suggesting an increase in anxiety-like behavior. In contrast, knockdown mice had the same number of total arm entries as wild types (Fig. 4C) indicating that an alteration in locomotion does not account for the differences in anxiety-like behavioral measures on this assay.
Knockdown mice were also tested for behavioral response to an aversive, brightly lit open field test apparatus using the forced novelty exploration test. Knockdown mice had a 28% reduction in locomotor activity compared to wild type mice when placed into this test arena (Fig. 4D). This reduction in exploratory behavior in the knockdown animals is also indicative of an increase in anxiety-like behavior.
Knockdown of the γ2 subunit did not alter hypnotic responses to benzodiazepine site ligands or other sedative/hypnotic drugs tested. We measured the loss of righting reflex to assess the acute sensitivity to sedative/hypnotic drugs; no significant differences were found in response to injection of diazepam, midazolam, zolpidem, ethanol, or pentobarbital (Fig. 5).
Production of γ2 global knockout mice
The gene targeting strategy that we used to produce the γ2 knockdown mice also resulted in the insertion of loxP sites that flank exon 8 of the γ2 gene (see Fig. 1A). To determine if these loxP sites were functional in vivo, heterozygous floxed γ2 mice were mated to a general deleter cre-expressing transgenic mouse line (JAX Laboratory, stock #003376) where the cre transgene was driven by a ubiquitously expressed human beta actin promoter [24]. Restriction fragment polymorphism analysis by Southern blotting confirmed the deletion of genomic DNA between loxP sites. As predicted from Fig. 1A, Probe A hybridized to a 6.6 kb BglII fragment from the recombined knockout allele (data not shown).
To determine if cre-mediated recombination and deletion of exon 8 resulted in a null allele, mice hemizygous for the cre transgene and heterozygous for the recombined γ2 allele (+/f) were mated to mice heterozygous for the floxed, non-recombined γ2 allele (+/F). 10.2% (14 of 137) of offspring from this mating strategy were homozygous for the recombined locus (f/f). This ratio is in agreement with the 12.5% predicted from Mendalian genetics. 13 of the 14 f/f mice died within three days of birth. One f/f mouse survived until postnatal day 21. This neonatal lethality of γ2 f/f mice is consistent with those found in the global γ2 knockouts of Günther et al. [9] which also had a deletion of exon 8. Therefore, we conclude that cre-mediated recombination of the loxP sites and deletion of exon 8 inactivates the γ2 gene.
Discussion
Genetic dissection of the GABAA-R system is yielding remarkable insights into GABAA-R biology and the mechanisms of action of various drugs. Here we report the establishment of a new GABAA-R mutant mouse model, one with attenuated expression of the γ2 subunit that averages ~35% of normal levels. This novel γ2 knockdown model has several unique features that make it useful. First, compared to global knockouts which die as neonates [9], knockdown mice have normal viability. Secondly, the graded levels of expression of γ2 (range 10–95% of normal) in the mutant mice could be useful in studies correlating phenotype with the amount of γ2 protein present. Finally, global knockout mice can be created from these mice by crossing to a Cre recombinase expressing global deleter mouse line. Therefore, the mice reported here represent a useful addition to the rapidly expanding arsenal of mice with genetic alterations in the γ2 subunit of the GABAA-R. In addition to homozygous and heterozygous global γ2 knockouts [9], conditional γ2 knockout mice [25], knockin mice with a point mutation in γ2 that eliminates zolpidem and inverse agonist β-carboline sensitivity [26,27], global knockout of the long splice variant of γ2 [28], and transgenic mice that express either the long or short splice variants of the γ2 subunit [29] have been described.
An important unexpected finding from this study was the extremely variable degree to which the genetic alteration of the γ2 locus changed the amount of γ2 product produced. This γ2 knockdown model was produced by insertion of NEO into intronic DNA. This strategy has been used previously to create hypomorphic alleles [11-23]. Creation of hypomorphic alleles has been useful in determining function of the gene that has been targeted, especially in cases where complete disruption of the gene is lethal [12,14-16]. Intronic insertion of NEO creates hypomorphic alleles either through cryptic splicing signals in the NEO gene that lead to a frameshift or alterations in splicing, or by down regulating gene expression through an unknown mechanism [14-16]. In the majority of these studies, the authors' reported a 5–25% level of expression from the targeted gene. However, in none of these previously described hypomorphic alleles was extensive variability reported. The γ2 targeted homozygous mice in this study clearly display a wide range of γ2 mRNA and protein levels. Semi-quantitative northern blot analysis revealed that 4 out of 5 homozygous knockdown mice showed, on average, a 70% reduction of γ2 mRNA levels, whereas 1 out of 5 mice had γ2 mRNA at a level that was similar to wild type. Immunohistochemical data were consistent with northern blot data. The reason for this highly variable expression within homozygous knockdown mice has yet to be determined. The mixed background of Strain 129S1/X1 and C57BL6/J may not be the cause given that other studies have not observed substantial variability in expression using similar background strains [11,18,20]. Gender as a cause of variability also is not likely given that a set of three male knockdown siblings exhibited the full range of γ2 levels; one being near wild type levels, another being at an intermediate level, and the third exhibiting severely attenuated expression. Further study to determine the cause of this variability needs to be undertaken. Nevertheless, investigators who desire to create a hypomorphic allele by inserting NEO into intronic DNA should take into consideration the variability of gene expression observed in this study. The extent of variability may be locus dependent.
Knockdown of γ2 subunit containing GABAA-Rs resulted in behavioral changes that are indicative of increased anxiety-like behavior. These mutant knockdown mice showed reduced exploratory behavior for a novel open field apparatus and the open arms of the elevated plus maze. Similar changes have been observed in γ2 global heterozygous knockout mice [10]. Together, these results strengthen the possibility that GABAA-R dysfunction may be an underlying cause of anxiety disorders in humans. These mutant mouse models may be useful for dissecting the etiology of this pervasive condition and for developing effective therapeutic interventions.
As expected, brains from γ2 knockdown mice demonstrated significantly decreased binding of the benzodiazepine site ligand, Ro15-4513. This was expected since benzodiazepine binding sites are located at the interface of select alpha and γ2 subunits of the GABAA-R [30]. The reduction observed in Ro15-4513 binding agree with those from global γ2 knockout mice [9] in that deficiency of the γ2 subunit abolishes the binding of the benzodiazepine-site ligands. The global heterozygous γ2 knockout mice have a reduction in total [3H]Ro15-4513 binding in most brain regions [31], and the present data on the γ2 knockdown show a very similar pattern of reduced binding. However, the 5 knockdown brains that we examined had actually a greater reduction in benzodiazepine binding than the γ2 heterozygous global knockouts published earlier [31]. Similar to the γ2 heterozygous knockouts, [3H]muscimol binding was not affected, indicating that the reduction of the γ2 subunit levels is not compensated by increased δ subunit that is largely responsible for the [3H]muscimol binding signal in brain sections [32]. The basal binding of the ion channel ligand [35S]TBPS is usually reduced and its sensitivity to GABA increased by addition of γ2 subunits [31]. Both in the γ2 heterozygous global knockout and in our homozygous knockdown, TBPS binding was increased in some brain regions, and, more consistently, there emerged a "GABA-insensitive" receptor population with widespread distribution in the brain. This indicates production of GABAA-Rs with an αβ configuration. These receptors bind strongly the agonist and have reduced channel conductance [33], perhaps reflecting only partial agonist efficacy of GABA. This would explain the reduced allosteric efficacy of GABA in abolishing TBPS binding. In addition, these receptors have most likely non-synaptic, non-clustered localization as they lack the γ2 subunit [10,33]. Therefore, the partially γ2 depleted mice may have brain region-selective alterations in synaptic and extrasynaptic receptors, and the reduced synaptic GABAA-R function might correlate with the anxious phenotype. However, surprisingly, the high-dose hypnotic effects of the benzodiazepine site ligands diazepam, zolpidem and midazolam were unchanged in the knockdown mice.
Several hypotheses can be generated to explain the discrepancy between decreased benzodiazepine ligand binding and unchanged hypnotic effects. One possibility is that only a threshold level of γ2 containing GABAA-Rs are required for hypnotic responses to benzodiazepines. Mice that completely lack γ2 had only 6% of benzodiazepine binding compared to wild type mice and were insensitive to the hypnotic effects of diazepam [9]. Likewise, mice with a point mutation in γ2 that eliminated response to zolpidem eliminated the effects of this benzodiazepine site ligand [26]. It seems likely that the 20–40% of benzodiazepine sensitive receptors that remain in the knockdown mice are enough to mediate the hypnotic effect of benzodiazepine site ligands. This is supported by positron emission tomography studies in humans, in which only a small proportion of the benzodiazepine sites need to be occupied by lorazepam and zolpidem to induce clinical effects such as sedation [34,35]. There may thus be spare receptors that can be measured by biochemical techniques but are not needed for whole animal pharmacological effects. Another hypothesis is that knockdown of γ2 does not result in a proportional reduction in all γ2 containing GABAA-Rs. Recent studies conducted in knockin mice with a point mutation at the benzodiazepine binding site have attributed α1 and α2 containing GABAA-Rs to the mediation of benzodiazepine-induced sedation and anxiolysis, respectively [4,36-38]. Perhaps γ2 knockdown mice have disproportionate decreases in non-α1 containing receptors, e.g., α2 or α3 containing receptor isoforms. It may be possible that in our γ2 knockdown mice, the number of α2βxγ2 GABAA-Rs is greatly reduced while the number of α1βxγ2 receptors is not, which remains to be studied. This could be tested by immunohistochemical methods using antibodies against α2 and α3 or by testing for changes in zolpidem insensitive binding. Changes in subunit trafficking may also play a role. Essrich et al. [7] established strong evidence that GABAA-Rs are clustered at the synapse through indirect interactions between γ2 and gephyrin. Enough clustering may remain in the knockdown mice such that drug response does not change. Alternatively, synaptic clustering of GABAA-Rs may not be important for behavioral responses to benzodiazepines, and it is possible that benzodiazepines potentiate the extrasynaptic GABAA-R responses determined by electrophysiology in brain slices [39]. It is also possible that the high variability in γ2 levels in the brain of knockdown mice masked any change in behavioral response to these drugs. In hindsight, it would have been useful to quantify γ2 levels in the same mice that were used for the behavioral sensitivity studies. It would be interesting to see if sensitivity correlated with γ2 levels on an individual animal basis. This type of approach could be exploited in future studies of γ2 receptor function. Lastly, decreases in benzodiazepine binding may not directly translate into changes in in vivo insensitivity. For example, in α1 null mice, changes in benzodiazepine binding did not change whole animal sensitivity to midazolam but increased sensitivity to diazepam [40,41]. In contrast, in mice that selectively lack the long splice variant of γ2, affinity for benzodiazepine site ligands is increased and behavioral response to the sedative effects of midazolam and zolpidem is also increased [42]. Similarly, in mice lacking the β3 subunit, decreased benzodiazepine binding and decreased whole animal midazolam sensitivity were observed [43]. Further study is required to examine these possibilities.
Conclusion
We have produced a novel genetically engineered mouse line that exhibits a reduction in γ2 mRNA levels that is highly variable between mice (range 10–95% of control) but averages ~35% of control levels. These γ2 knockdown mice are viable and have enhanced anxiety-like behavioral abnormalities. Surprising, these mice are normally sensitive to the hypnotic effects of benzodiazepine site ligands. Lastly, this novel genetically engineered mouse line can also be easily converted to a γ2 global knockout mouse line by simply crossing to a cre-expressing global deleter transgenic mouse line.
Methods
Generation of genetically altered mice
A 13.6 kB BglII Strain 129/SvJ mouse genomic DNA fragment containing Exons 8–10 of the GABAA-R γ2 gene was subcloned from a P1 phage clone (Genome Systems, St. Louis, MO). An oligonucleotide containing a loxP site and an EcoRV site was inserted into a NcoI site 647 bp 5' to Exon 8. A blunted fragment containing a NEO resistance gene with PGK-1 promoter and polyadenylation signals (obtained from pPNT [44]) fused with a loxP site was inserted into a HincII site 855 bp 3' of Exon 8. The NEO gene was inserted in the opposite orientation to γ2 transcription. A PGK driven thymidine kinase expression cassette (obtained from pPNT [44]) was then cloned 3' to the γ2 genomic DNA. The targeting construct was linearized with SalI and electroporated into R1 embryonic stem cells [45] following previously described procedures [46]. G418 (270 μg/ml; Life Technologies, Gaithersburg, MD) and gancyclovir (2 μM; Sigma) resistant cells were screened for gene targeting by Southern blot analysis of BglII digested genomic DNA. Fragments were hybridized with a 5' probe that was external to the targeting construct (see Fig. 1). Proper targeting of the γ2 locus was confirmed by Southern blot analysis of EcoRV and SstI digested DNA. Additional Southern blot analysis with probes to the neomycin cassette or internal to the 3' arm of the targeting vector (with 1–3 restriction enzymes) were also conducted (results not shown). All results were consistent with a correctly targeted γ2 locus.
Two correctly targeted embryonic stem cell clones (26C8 and 26L6) were microinjected into C57BL/6J blastocysts to produce chimeric mice. Male chimeras were mated to C57BL/6J females. Agouti offspring heterozygous for the targeted allele (F/+) were interbred to produce mice that were wild type (+/+), heterozygous, or homozygous mutants (F/F). The mice used in studies reported here were derived from the 26L6 ES cell clone. All mice were housed under conditions of lights on at 07:00 and lights off at 19:00.
Northern blot analysis
Ten to fourteen week old mice were sacrificed by cervical dislocation. RNA was isolated from whole brain using Trizol reagent (Invitrogen). 20 μg of total RNA was electrophoresed in a 1.9% formaldehyde/1% agarose gel and blotted onto Hybond-N (Amersham Pharmacia). Blots were first probed with a 32P-labeled γ2 cDNA probe then re-hybrizided with a control human β-actin probe (Clontech). Autoradiographs were digitally photographed and band densities were measured using Kodak 1D Image Analysis Software. Ratio between density of γ2 hybridization versus actin hybridization was recorded.
Immunocytochemistry
The procedure has been described elsewhere [47]. Briefly, wild type and homozygous mutants were deeply anesthetized with Avertin and transcardially perfused with Dextran-Phosphate Buffer (PB) then fixative consisting of 0.01M periodate/0.075M lysine/4% paraformaldehyde in 0.1M PB (pH 7.4). The brains were frozen and sliced in parasaggital sections (25 μm) with a freezing microtome. Slices were incubated overnight at 4°C with affinity-purified rabbit anti-γ2 [48,49] in 0.1M PB (pH 7.4) with 0.3% Triton X-100, at a concentration of (1:100). The sections were processed using an avidin-biotin-peroxidase system (Vectastain Elite; Novocastra, Burlingame, CA). Peroxidase reaction was carried out with 3-3' diaminobenzidine tetrahydrochloride in the presence of cobalt chloride and nickel ammonium sulfate as chromogens and hydrogen peroxide as oxidant. Controls were performed either by omitting the primary antibody or by incubating the primary antibody with the corresponding peptide [50].
Pharmacology
Eight-week-old mice were sacrificed by decapitation and whole brains were rapidly dissected out and frozen on dry ice. For autoradiography, 14-μm horizontal serial cryostat sections were cut from 5 mouse brains of each genotype, thaw-mounted onto gelatin-coated object glasses, and stored frozen under desiccant at -20°C. All experiments were carried out in parallel fashion with respect to mouse lines, eliminating any day-to-day variation between the assays. The autoradiographic procedures for regional localization of [3H]Ro 15-4513, [3H]muscimol, and [35S]TBPS binding were as described in [51]. Briefly, sections were preincubated in an ice-water bath for 15 min in 50 mM Tris-HCl (pH 7.4) supplemented with 120 mM NaCl in the [3H]Ro 15-4513 and [35S]TBPS autoradiographic assays, and in 0.31 M Tris-citrate (pH 7.1) in the [3H]muscimol assay. All radioligands were purchased from Perkin Elmer Life Sciences, Inc. (Boston, MA, USA).
The final incubation in respective preincubation buffer was performed with 6 nM [35S]TBPS at room temperature for 90 min, assays with 10 nM [3H]muscimol at 0–4°C for 30 min, and assays with 10 nM [3H]Ro 15-4513 in the presence and absence of 100 μM diazepam (Orion Pharma, Espoo, Finland) at 0–4°C for 60 min. After incubation, sections were washed 3 × 15 s or 2 × 30 s in an ice-cold incubation buffer in [35S]TBPS and [3H]Ro 15-4513 or in [3H]muscimol assay, respectively. Sections were then dipped into distilled water, air-dried under a fan at room temperature, and exposed with plastic [3H]- or [14C]-methacrylate standards to Kodak Biomax MR films for 1 to 8 weeks. Nonspecific binding was determined with 10 μM flumazenil (Hoffmann-La Roche, Basel, Switzerland), 100 μM picrotoxinin (Sigma) and 100 μM GABA (Sigma) in [3H]Ro 15-4513, [35S]TBPS and [3H]muscimol assays, respectively. Binding densities in the relevant brain areas were quantitated with MCID M5-imaging software (Imaging Research Inc., Ontario, Canada) and converted to nCi/mg (for 3H) or nCi/g (for 35S) radioactivity values on the basis of the simultaneously exposed standards.
The concentrations of [3H]muscimol (10 nM) and [3H]Ro 15-4513 (10 nM) were greater than or equal to the dissociation constants for a range of recombinant and native GABAA receptors [8,52-54]. Therefore, the autoradiographic images should represent the density rather than affinity of binding sites.
Hypnotic sensitivity
Drug induced hypnosis was assessed by measuring the duration of the loss of righting reflex in response to various sedative/hypnotic drugs. Diazepam (25 mg/kg; Sigma), midazolam (45 mg/kg and 75 mg/kg; ESI Lederle, Philadelphia, PA), pentobarbital (45 mg/kg; Abbott, Chicago, IL), zolpidem (60 mg/kg; Searle, Malvern, PA) or ethanol (3.5 mg/kg; Pharmco, Brookfield, CT) were administered intraperitoneally to wild type and homozygous mutants. Drugs were diluted in saline such that the injection volume was 20 μL per gram body weight. After losing the righting reflex, mice were placed in a plastic V-shaped trough and the time was recorded. When the mouse was able to right itself three times in 30 s, the measure of hypnotic effect was over. Body temperature was maintained with aid of a heat lamp. Assays were performed in a blinded manner with respect to genotype. Effect of genotype on the duration of the loss of righting reflex was compared using Student's t-test.
Elevated plus-maze test
Basal anxiety-like behavior was tested using the elevated plus-maze. All mice were between 7 and 9 weeks of age and were tested between 09:00 and 11:00. Each mouse was placed on the central platform of the maze, facing an open arm and allowed to freely explore the maze for 5 min under ambient room light. Open-arm and closed-arm entries and the cumulative time spent on the open and closed arms was recorded. A mouse was considered to be on the central platform or on an arm when all four paws were within its perimeter. The percent open-arm entries, total number of entries, and percent time in open-arms was determined. Data were analyzed using one-way ANOVA.
Forced exploration
Basal motor activity of mice was determined using the forced exploration test. 8–10 week old mice were placed into a walled arena (43.2 cm × 43.2 cm × 30.5 cm) for 5 min. Distance traveled (cm) was measured automatically using an Activity Monitor (Med Associates, St. Albans, VT). All tests were performed between 12:00 and 15:00. Effect of genotype on basal motor activity was compared using Student's t-test.
Authors' contributions
CPM and ALD carried out the immunohistochemical studies. ERK carried out the ligand autoradiography studies. All other experiments were performed by DC and GEH. All authors contributed to composing and editing the manuscript and have read and approved the final version.
Acknowledgements
Carolyn Ferguson and Ed Mallick are gratefully acknowledged for superb technical support. This work was supported by NIH grants AA10422, GM52035, GM47818, NS039287, the Sigrid Juselius Foundation, and the Academy of Finland.
Figures and Tables
Figure 1 Gene targeting of the γ2 subunit of the GABAA-R and development of γ2 knockdown mice. (A) Targeting strategy used to produce γ2 knockdown mice. Relevant region of endogenous gene (+), targeting vector, correctly targeted knockdown γ2 (F), and cre-recombined knockout alleles (f) are shown. Relative locations of relevant restriction sites, exons (numbered orange boxes), loxP sites (blue triangles), plasmid backbone (wavy line), and positive (NEO) and negative (PGK-TK) selection cassettes are shown. (B) Southern Blot analysis of BglII digested tail DNAs. Probe A hybridizes to a 9.2 kb BglII fragment from the wild type endogenous allele and a 3.4 kb BglII fragment from the targeted allele.
Figure 2 Northern blot analysis. (A) Sample northern blot analysis of wild type (+/+) and homozygous knockdown (F/F) adult whole brain total RNA hybrizided with a γ2 cDNA or a human β-actin probe as a loading control. (B) Densitometric analysis of northern blots. The ratio of band density between γ2 and β-actin hybridization for wild type (n = 23) and homozygous knockdown (n = 36) mice was plotted. On each individual blot, the ratio of wild type band densities was averaged and normalized to 100. All samples on each blot were then compared to the average of wild type band densities on that blot. A total of 4 different blots were included in this analysis. The horizontal bar in each column represents the mean for that genotype.
Figure 3 Immunohistochemical distribution of γ2 subunit of the GABAA-R in sections from individual (A) wild type and (B) homozygous knockdown mice. Note the high level of abundance of γ2 immunoreactivity in olfactory bulb (OB), cerebral cortex (CC), hippocampus (HP), substantia nigra (SN), and cerebellum (CB) of wild type samples. The amount of γ2 is variably reduced in many brain regions of the knockdown samples.
Figure 4 Behavioral characterization of knockdown mice. (A-C) Elevated plus maze. Knockdown mice (A) entered open arms less often and (B) for less time, however, (C) total entries into arms did not differ from wild type mice. (D) Knockdown mice were also less active in the forced exploration test. The bars are means ± SEM. *p < .01, **p < .001.
Figure 5 Hypnotic sensitivity. No significant differences in the hypnotic effects of ethanol, pentobarbital, zolpidem, midazolam, or diazepam were observed. The bars are means ± SEM.
Table 1 Autoradiographic analysis of GABAA receptor binding sites in horizontal sections of wild type and mutant mice.
BRAIN REGION
Ligand/condition Ctx Th Hi CPu IC Gr Mol
[3H]Ro15-4513, basal
+/+ mice 102 ± 3 53 ± 3 102 ± 1 52 ± 4 63 ± 7 93 ± 4 84 ± 4
+/F mice 97 ± 2 42 ± 2c 91 ± 3c 44 ± 3b 52 ± 7 74 ± 7b 72 ± 7a
F/F mice 77 ± 6c 23 ± 2c 68 ± 5c 30 ± 3c 38 ± 13b 59 ± 7a 62 ± 6c
[3H]Ro15-4513 + Diazepam 100 μM
+/+ mice 4.4 ± 0.3 4.0 ± 0.3 7.0 ± 0.4 3.7 ± 0.3 3.3 ± 0.2 56 ± 5 5.4 ± 1.0
+/F mice 3.8 ± 0.3a 3.5 ± 0.4 5.7 ± 0.5c 3.3 ± 0.3 3.3 ± 0.3 41 ± 5c 4.7 ± 0.9
F/F mice 3.1 ± 0.2c 2.8 ± 0.2c 4.3 ± 0.3c 2.8 ± 0.2c 2.9 ± 0.1a 13 ± 4c 3.7 ± 1.5
[3H]Muscimol, basal
+/+ mice 16 ± 2 20 ± 4 9.3 ± 1.3 11 ± 2 4.2 ± 0.6 94 ± 7 8.7 ± 2.3
+/F mice 18 ± 2 23 ± 3 10.0 ± 2.1 11 ± 2 3.9 ± 0.8 101 ± 8 8.6 ± 1.0
F/F mice 15 ± 3 23 ± 7 9.7 ± 3.0 11 ± 4 3.8 ± 0.6 89 ± 16 8.1 ± 3.8
[35S]TBPS, basal
+/+ mice 210 ± 26 583 ± 143 211 ± 30 257 ± 55 914 ± 41 641 ± 114 80 ± 25
+/F mice 271 ± 47 527 ± 110 248 ± 34 277 ± 31 866 ± 51 722 ± 140 73 ± 16
F/F mice 277 ± 57 460 ± 145 299 ± 43b 292 ± 50 829 ± 98 425 ± 32b 100 ± 30
[35S]TBPS + GABA 1 mM
+/+ mice 3.0 ± 2.0 14 ± 9 1.7 ± 1.8 0.4 ± 2.6 4.6 ± 2.2 67 ± 6 1.7 ± 1.8
+/F mice 7.0 ± 1.3a 29 ± 13 5.9 ± 1.6b 4.9 ± 0.4a 18 ± 5a 75 ± 12 0.7 ± 2.4
F/F mice 16 ± 2c 42 ± 16a 13 ± 2.0c 18 ± 3c 76 ± 12c 86 ± 13a 5.0 ± 2.2
The data are means ± standard deviations for 5 mice in each genotype, and are expressed as nCi/mg for 3H-ligands and as nCi/g for 35S-ligand. Significance of the difference from wild type (Tukey-Kramer test after ANOVA): aP < 0.05, bP < 0.01, cP < 0.001. Ctx, cortex; Th, thalamus; Hi, hippocampus; CPu, caudate putamen; IC, inferior colliculus; Gr, cerebellar granule cell layer; Mol, cerebellar molecular layer.
==== Refs
Harris RA Mihic SJ Brozowski S Hadingham K Whiting PJ Ethanol, flunitrazepam, and pentobarbital modulation of GABAA receptors expressed in mammalian cells and xenopus oocytes Alcohol Clin Exp Res 1997 21 444 451 9161604
Homanics GE Accili D Gene targeting strategies in the analysis of alcohol and anesthetic mechanisms Genetic Manipulation of Receptor Expression and Function Receptor Biochemistry and Methodology Sibley D., Strader C. 2000 New York, NY , John Wiley & Sons, Inc. 93 110
Rudolph U Crestani F Mohler H GABA(A) receptor subtypes: dissecting their pharmacological functions Trends Pharmacol Sci 2001 22 188 194 11282419 10.1016/S0165-6147(00)01646-1
Rudolph U Mohler H Analysis of GABAa receptor function and dissection of the pharmacology of benzodiazepines and general anesthetics through mouse genetics Annu Rev Pharmacol Toxicol 2004 44 475 498 14744255 10.1146/annurev.pharmtox.44.101802.121429
DeBlas AL Brain GABAa receptors studied with subunit-specific antibodies Mol Neurobiol 1996 12 55 71 8732540
Whiting PJ The GABAA receptor gene family: new opportunities for drug development Curr Opin Drug Discov Devel 2003 6 648 657 14579514
Essrich C Lorez M Benson JA Fritschy JM Luscher B Postsynaptic clustering of major GABAA receptor subtypes requires the gamma 2 subunit and gephyrin Nat Neurosci 1998 1 563 571 10196563 10.1038/2798
Pritchett DB Sontheimer H Shivers BD Ymer S Kettenmann H Schofield PR Seeburg PH Importance of a novel GABA-A receptor subunit for benzodiazepine pharmacology Nature 1989 338 582 585 2538761 10.1038/338582a0
Gunther U Benson J Benke D Fritschy JM Reyes G Knoflach F Crestani F Aguzzi A Arigoni M Lang Y Bluethmann H Mohler H Luscher B Benzodiazepine-insensitive mice generated by targeted disruption of the gamma 2 subunit gene of gamma-aminobutyric acid type A receptors Proc Natl Acad Sci U S A 1995 92 7749 7753 7644489
Crestani F Lorez M Baer K Essrich C Benke D Laurent JP Belzung C Fritschy JM Luscher B Mohler H Decreased GABAA-receptor clustering results in enhanced anxiety and a bias for threat cues Nat Neurosci 1999 2 833 839 10461223 10.1038/12207
Saijoh Y Oki S Ohishi S Hamada H Left-right patterning of the mouse lateral plate requires nodal produced in the node Dev Biol 2003 256 160 172 12654299 10.1016/S0012-1606(02)00121-5
Ieraci A Forni PE Ponzetto C Viable hypomorphic signaling mutant of the Met receptor reveals a role for hepatocyte growth factor in postnatal cerebellar development Proc Natl Acad Sci U S A 2002 99 15200 15205 12397180 10.1073/pnas.222362099
Raffai RL Weisgraber KH Hypomorphic apolipoprotein E mice: a new model of conditional gene repair to examine apolipoprotein E-mediated metabolism J Biol Chem 2002 277 11064 11068 11792702 10.1074/jbc.M111222200
Nagy A Moens C Ivanyi E Pawling J Gertsenstein M Hadjantonakis AK Pirity M Rossant J Dissecting the role of N-myc in development using a single targeting vector to generate a series of alleles Curr Biol 1998 8 661 664 9635194 10.1016/S0960-9822(98)70254-4
Meyers EN Lewandoski M Martin GR An Fgf8 mutant allelic series generated by Cre- and Flp-mediated recombination Nat Genet 1998 18 136 141 9462741 10.1038/ng0298-136
Mohn AR Gainetdinov RR Caron MG Koller BH Mice with reduced NMDA receptor expression display behaviors related to schizophrenia Cell 1999 98 427 436 10481908 10.1016/S0092-8674(00)81972-8
Abu-Issa R Smyth G Smoak I Yamamura K Meyers EN Fgf8 is required for pharyngeal arch and cardiovascular development in the mouse Development 2002 129 4613 4625 12223417
Althini S Bengtsson H Usoskin D Soderstrom S Kylberg A Lindqvist E Chuva de Sousa Lopes S Olson L Lindeberg J Ebendal T Normal nigrostriatal innervation but dopamine dysfunction in mice carrying hypomorphic tyrosine hydroxylase alleles J Neurosci Res 2003 72 444 453 12704806 10.1002/jnr.10606
Liu Y Festing MH Hester M Thompson JC Weinstein M Generation of novel conditional and hypomorphic alleles of the Smad2 gene Genesis 2004 40 118 15452874 10.1002/gene.20072
Martin DM Skidmore JM Philips ST Vieira C Gage PJ Condie BG Raphael Y Martinez S Camper SA PITX2 is required for normal development of neurons in the mouse subthalamic nucleus and midbrain Dev Biol 2004 267 93 108 14975719 10.1016/j.ydbio.2003.10.035
Rucker EB Dierisseau P Wagner KU Garrett L Wynshaw-Boris A Flaws JA Hennighausen L Bcl-x and Bax regulate mouse primordial germ cell survival and apoptosis during embryogenesis Mol Endocrinol 2000 14 1038 1052 10894153 10.1210/me.14.7.1038
Tian E Kimura C Takeda N Aizawa S Matsuo I Otx2 is required to respond to signals from anterior neural ridge for forebrain specification Dev Biol 2002 242 204 223 11820816 10.1006/dbio.2001.0531
Wu L Gu J Cui H Zhang QY Behr M Fang C Weng Y Kluetzman K Swiatek PJ Yang W Kaminsky L Ding X Transgenic Mice with a Hypomorphic NADPH-Cytochrome P450 Reductase Gene: Effects on Development, Reproduction, and Microsomal Cytochrome P450 J Pharmacol Exp Ther 2005 312 35 43 15328377 10.1124/jpet.104.073353
Lewandoski M Meyers EN Martin GR Analysis of Fgf8 gene function in vertebrate development Cold Spring Harb Symp Quant Biol 1997 62 159 168 9598348
Schweizer C Balsiger S Bluethmann H Mansuy IM Fritschy JM Mohler H Luscher B The gamma 2 subunit of GABA(A) receptors is required for maintenance of receptors at mature synapses Mol Cell Neurosci 2003 24 442 450 14572465 10.1016/S1044-7431(03)00202-1
Cope DW Wulff P Oberto A Aller MI Capogna M Ferraguti F Halbsguth C Hoeger H Jolin HE Jones A McKenzie AN Ogris W Poeltl A Sinkkonen ST Vekovischeva OY Korpi ER Sieghart W Sigel E Somogyi P Wisden W Abolition of zolpidem sensitivity in mice with a point mutation in the GABA(A) receptor gamma2 subunit Neuropharmacology 2004 47 17 34 15165831 10.1016/j.neuropharm.2004.03.007
Leppa E Vekovischeva OY Linden AM Wulff P Oberto A Wisden W Korpi ER Agonistic effects of the beta-carboline DMCM revealed in GABA(A) receptor gamma2 subunit F77I point-mutated mice Neuropharmacology 2005 48 469 478 15755475 10.1016/j.neuropharm.2004.11.007
Homanics GE Harrison NL Quinlan JJ Krasowski MD Rick CEM deBlas AL Mehta A Kist F Mihalek RM Aul JJ Firestone LL Normal electrophysiological and behavioral responses to ethanol in mice lacking the long splice variant of the g2 subunit of the g-aminobutyrate type A receptor Neuropharmacology 1999 38 253 265 10218866 10.1016/S0028-3908(98)00177-4
Baer K Essrich C Balsiger S Wick MJ Harris RA Fritschy JM Luscher B Rescue of gamma2 subunit-deficient mice by transgenic overexpression of the GABAA receptor gamma2S or gamma2L subunit isoforms European Journal of Neuroscience 2000 12 2639 2643 10947838 10.1046/j.1460-9568.2000.00159.x
Sigel E Mapping of the benzodiazepine recognition site on GABA(A) receptors Curr Top Med Chem 2002 2 833 839 12171574 10.2174/1568026023393444
Sinkkonen ST Luscher B Luddens H Korpi ER Autoradiographic imaging of altered synaptic alphabetagamma2 and extrasynaptic alphabeta GABAA receptors in a genetic mouse model of anxiety Neurochem Int 2004 44 539 547 15209422 10.1016/j.neuint.2003.08.008
Korpi ER Mihalek RM Sinkkonen ST Hauer B Hevers W Homanics GE Sieghart W Lüddens H Altered receptor subtypes in the forebrain of GABAA receptor d subunit-deficient mice: recruitment of g2 subunits Neuroscience 2002 109 733 743 11927155 10.1016/S0306-4522(01)00527-9
Lorez M Benke D Luscher B Mohler H Benson JA Single-channel properties of neuronal GABAA receptors from mice lacking the gamma2 subunit J Physiol 2000 527 11 31 10944167 10.1111/j.1469-7793.2000.t01-1-00011.x
Abadie P Rioux P Scatton B Zarifian E Barre L Patat A Baron JC Central benzodiazepine receptor occupancy by zolpidem in the human brain as assessed by positron emission tomography Eur J Pharmacol 1996 295 35 44 8925872 10.1016/0014-2999(95)00633-8
Sybirska E Seibyl JP Bremner JD Baldwin RM al-Tikriti MS Bradberry C Malison RT Zea-Ponce Y Zoghbi S During M Goddard AW Woods SW Hoffer PB Charney DS Innis RB [123I]iomazenil SPECT imaging demonstrates significant benzodiazepine receptor reserve in human and nonhuman primate brain Neuropharmacology 1993 32 671 680 8395663 10.1016/0028-3908(93)90080-M
Low K Crestani F Keist R Benke D Brunig I Benson JA Fritschy JM Rulicke T Bluethmann H Mohler H Rudolph U Molecular and neuronal substrate for the selective attenuation of anxiety Science 2000 290 131 134 11021797 10.1126/science.290.5489.131
Rudolph U Crestani F Benke D Brunig I Benson JA Fritschy JM Martin JR Bluethmann H Mohler H Benzodiazepine actions mediated by specific gamma-aminobutyric acid(A) receptor subtypes Nature 1999 401 796 800 10548105 10.1038/44579
McKernan RM Rosahl TW Reynolds DS Sur C Wafford KA Atack JR Farrar S Myers J Cook G Ferris P Garrett L Bristow L Marshall G Macaulay A Brown N Howell O Moore KW Carling RW Street LJ Castro JL Ragan CI Dawson GR Whiting PJ Sedative but not anxiolytic properties of benzodiazepines are mediated by the GABA(A) receptor alpha1 subtype Nat Neurosci 2000 3 587 592 10816315 10.1038/75761
Lindquist CE Ebert B Birnir B Extrasynaptic GABA(A) channels activated by THIP are modulated by diazepam in CA1 pyramidal neurons in the rat brain hippocampal slice Mol Cell Neurosci 2003 24 250 257 14550784 10.1016/S1044-7431(03)00128-3
Kralic JE O'Buckley TK Khisti RT Hodge CW Homanics GE Morrow AL GABA(A) receptor alpha-1 subunit deletion alters receptor subtype assembly, pharmacological and behavioral responses to benzodiazepines and zolpidem Neuropharm 2002 43 685 694 10.1016/S0028-3908(02)00174-0
Kralic JE Korpi ER O'Buckley TK Homanics GE Morrow AL Molecular and pharmacological characterization of GABA(A) receptor alpha1 subunit knockout mice J Pharmacol Exp Ther 2002 302 1037 1045 12183661 10.1124/jpet.102.036665
Quinlan JJ Firestone LL Homanics GE Mice lacking the long splice variant of the g2 subunit of the GABAa receptor are more sensitive to benzodiazepines Pharmacol Biochem Behav 2000 66 371 374 10880692 10.1016/S0091-3057(00)00225-2
Quinlan JJ Homanics GE Firestone LL Anesthesia sensitivity in mice that lack the beta3 subunit of the gamma-aminobutyric acid type A receptor Anesthesiology 1998 88 775 780 9523823 10.1097/00000542-199803000-00030
Tybulewicz VLJ Crawford CE Jackson PK Bronson RT Mulligan RC Neonatal lethality and lymphopenia in mice with a homozygous disruption of the c-abl proto-oncogene Cell 1991 65 1153 1163 2065352 10.1016/0092-8674(91)90011-M
Nagy A Rossant J Nagy R Abramow-Newerly W Roder JC Derivation of completely cell culture-derived mice from early-passage embryonic stem cells Proc Natl Acad Sci, U S A 1993 90 8424 8428 8378314
Homanics GE Ferguson C Quinlan JJ Daggett J Snyder K Lagenaur C Mi ZP Wang XH Grayson DR Firestone LL Gene knockout of the alpha-6 subunit of the gamma-aminobutyric acid type A receptor: Lack of effect on responses to ethanol, pentobarbital, and general anesthetics Mol Pharmacol 1997 51 588 596 9106623
de Blas AL Vitorica J Friedrich P Localization of the GABAA receptor in the rat brain with a monoclonal antibody to the 57,000 Mr peptide of the GABAA receptor/benzodiazepine receptor/Cl- channel complex J Neurosci 1988 8 602 614 2828565
Christie SB Li RW Miralles CP Riquelme R Yang BY Charych E Wendou Y Daniels SB Cantino ME De Blas AL Synaptic and extrasynaptic GABAA receptor and gephyrin clusters Prog Brain Res 2002 136 157 180 12143379
Christie SB Miralles CP De Blas AL GABAergic innervation organizes synaptic and extrasynaptic GABAA receptor clustering in cultured hippocampal neurons J Neurosci 2002 22 684 697 11826098
Miralles CP Li M Mehta AK Khan ZU De Blas AL Immunocytochemical localization of the beta(3) subunit of the gamma-aminobutyric acid(A) receptor in the rat brain J Comp Neurol 1999 413 535 548 10495441 10.1002/(SICI)1096-9861(19991101)413:4<535::AID-CNE4>3.0.CO;2-T
Mäkelä R Uusioukari M Homanics GE Quinlan JJ Firestone LL Wisden W Korpi ER Cerebellar gamma-aminobutyric acid type a receptors: Pharmacological subtypes revealed by mutant mouse lines Mol Pharmacol 1997 52 380 388 9281599
Lüddens H Pritchett DB Kohler M Killisch I Keinanen K Monyer H Sprengel R Seeburg PH Cerebellar GABA-A receptor selective for a behavioural alcohol antagonist Nature 1990 346 648 651 2166916 10.1038/346648a0
Wisden W Herb A Wieland H Keinänen K Lüddens H Seeberg PH Cloning, pharmacological characteristics and expression pattern of the GABAA receptor a4 subunit FEBS Lett 1991 289 227 230 1655526 10.1016/0014-5793(91)81076-K
Pritchett DB Seeburg PH g-aminobutyric acid-A receptor a5-subunit creates novel type II benzodiazepine receptor pharmacology J Neurochem 1990 54 1802 1805 2157817
| 15850489 | PMC1097738 | CC BY | 2021-01-04 16:03:48 | no | BMC Neurosci. 2005 Apr 25; 6:30 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-30 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-411584016710.1186/1471-2458-5-41Research ArticleEthnic and age-related fat free mass loss in older Americans: The Third National Health and Nutrition Examination Survey (NHANES III) Obisesan Thomas O [email protected] Muktar H [email protected] Vernon [email protected] Richard G [email protected] Abimbola [email protected] Charles N [email protected] Section of Geriatrics, Department of Medicine, Howard University Hospital, Washington DC, USA2 Department of Epidemiology, University of Alabama at Birmingham, Birmingham, AL, USA3 Department of Physical Education, Howard University Washington DC, USA4 Pulmonary and Critical Care, Department of Internal Medicine, Howard University Hospital, Washington DC, USA5 Division of Geriatrics, Morehouse School of Medicine, Atlanta, GA, USA6 National Human Genome Center, Genetic Epidemiology Unit, Department of Microbiology, Howard University, Washington DC, USA2005 19 4 2005 5 41 41 13 11 2004 19 4 2005 Copyright © 2005 Obisesan et al; licensee BioMed Central Ltd.2005Obisesan 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
Although age-related loss of fat free mass (FFM) is well known, there is paucity of data on national estimates, and on the differential influence of ethnicity on the decline in FFM with increasing age. We determined whether age-related loss in FFM and fat free mass index (FFMI) vary by gender and or ethnicity, using representative data from the Third National Health and Nutrition Examination Survey (NHANES III).
Methods
Analyses were limited to 5,803 non-institutionalized, non-Hispanic Whites and African Americans (Blacks) over the age of 40 years. Body density was calculated from the sum of 3-skinfolds, and percent body fat estimated from body density. FFM was estimated by subtracting body fat from body weight, while FFMI was defined as FFM (kilograms) divided by the square of body height (meter2).
Results
Overall FFM and FFMI were significantly higher in black women than white women (P = 0.001; P = 0.001 respectively), but similar in black men compared to white men. Age-related decline in FFM reached significance level earlier in black men (at age 65–69) than white men (at age 70–74), and in black women (at age 70–74) than white women (at age 75–79). Similar decline in FFMI was noted in men and in women. In multivariate analyses, FFM significantly associated with ethnicity (p = 0.012) and with age (p < 0.001) in women, but only with age (p < 0.001) in men. In men and in women, FFMI significantly associated with ethnicity (p < 0.001; p = 0.003 respectively) and with age (p < 0.001; p = 0.004 respectively).
Conclusion
Age-related loss and decline in FFM and FFMI in older Americans is higher for black men and women, than for white men and women. The development of focused population-based preventive strategies is likely to improve functional independence in the aged.
==== Body
Background
Advancing age is associated with a number of changes in body composition. Notable among these changes is the reduction in FFM, that occurs primarily as a result of losses in skeletal muscle mass [1], a condition referred to as sarcopenia. For example, older men aged 80 years or older have 25 percent less leg muscle mass than young men between the ages of 20 and 29[2]. FFM has been closely linked with reduction in muscle strength in the aged. Further, a greater reduction in both muscle mass and strength have been reported in older nursing residents with history of falls. Thus reduction in muscular strength is associated with increased risk of falls and hip fracture, and prevalence of disabilities[3]. Independent of fall risk, low FFM has been linked to higher rates of all-cause mortality in women in the United States[4]. Therefore, losses in FFM with advancing age have important health implications.
Higher FFM and lean mass to height ratio in Blacks than Whites have been reported [5], but mostly in young non-representative samples. There is limited information on national estimates, and on the existence of differential influence of ethnicity on the decline in FFM with increasing age. Given the inverse association of FFM with all cause mortality[6] and the reported higher FFM in Blacks, lower mortality rates in Blacks would be expected. To the converse, both all cause mortality[7] and FFM are know to be higher in Blacks, suggesting that the relationship of FFM with ethnicity and all cause mortality may be due to reasons other than differences in absolute FFM. We hypothesized that black men and women will have higher degree of age-related decline in FFM and FFMI despite higher mean values in Blacks, using data from NHANES III. The characterization of changes in FFM across several decades of life is crucial to our ability to reduce the burden of sarcopenia, osteoporosis, falls and hip fracture, as well as associated mortality.
Methods
In order to characterize FFM in older populations in the U.S., we analyzed data from NHANES III. The data for NHANES III were collected by the National Center for Health Statistics of the Centers for Disease Control and Prevention between 1988 and 1994, using a national multistage, probability sample of the civilian non-institutionalized population of the United States. The initial NHANES III sample consisted of 33,994 individuals, 2 months to 99 years of age [8]. The survey design ensured over-sampling of Blacks, and those 60 years and older. Data were collected from each participant through face-to-face interview, physical examination, and laboratory analyses. The overall survey design and provisions for informed consent have been previously detailed [8].
Subjects in the present study included 8,679 non-institutionalized Whites and Blacks over the age of 40 years. Of this number, complete data on anthropometric measurements were only available for 5,803 persons. Age was categorized as decades of life up to the age of 79 years, while persons ages 80 and older were separately grouped to avoid over-stratification and loss of statistical power. We also categorized age by increments of 5 years for the purpose of graphical representations only.
Anthropometric measurements
Weight and standing height were determined according to standardized procedures [9]. Weight was recorded to the nearest 0.01 kg, while height was recorded to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight in kg divided by height in meter squared (m2). Except for a few measurements that were taken on the left side of the body because of examinees' limitations, the rest of the anthropometric measurements were taken on the right side of the body using standard procedures and were recorded to the nearest 0.1 mm [9]. All skinfolds were measured using Holtain skinfold calipers on the right side of the body. With the examinee standing erect and the arms hanging freely at the sides, the triceps skinfold was measured on the posterior surface at the midpoint of the right upper arm. The suprailiac skinfold was measured at a 45-degree angle at the marked iliac crest. Measurement of the thigh skinfold was taken from a vertical line over the quadriceps muscle at midline of the thigh, and half way between the top of the patella and the inguinal crease [9]. All examiners were trained in standard procedures for obtaining measurements. Out-of-range measurements were verified to differentiate errors in measurement from high values for very small or very large examinees.
Because relevant anthropometric measurements were available only for the triceps, suprailium and thigh, body density was estimated from the sum of 3 skinfolds, (triceps, suprailium, and thigh) and age, based on the equation of Jackson and Pollock (body density = 1.0994921-0.0009929 * (triceps + suprailium + thigh) + 0.0000023 * (triceps + suprailium + thigh)**2 - 0.0001392 * (age)) [10], and percent body fat was estimated from the Brozek equation (percent body fat=((4.570/body density) - 4.142) * 100) [11]. FFM was calculated by subtracting body fat from body weight. Based on the analogy of BMI and the high correlation between FFM and height, FFMI was derived by dividing FFM (kg) by the square of height (m2)[12].
Statistical analyses
Because our primary interest was in how ethnicity influenced age-related differences or decline in FFM by gender, initial analyses were conducted for each of the 4 demographic subgroups formed by the combination of sex and ethnicity. Separate analyses were conducted for each decade of life beginning from the 4th to the 8th decade. Mean values were calculated for FFM and FFMI. The χ2 and t-test statistics and 95 percent Confidence Interval were used to determine statistical significance. A General Linear Regression Model for continuous variable response, using the method of weighted least squares adapted to survey data [13] was used to test the association between the dependent variable (FFM) and independent variables (age, ethnicity and gender) each in a bivariate model. Similar models were constructed for FFMI as a second dependent variable. Estimates of individual regression coefficients (the vector of beta values) and standard errors were computed. Because of our interest in how ethnicity influenced age-related changes in FFM in men and in women, we tested for "ethnicity*gender" interaction (dependent variable = ethnicity, gender, ethnicity*gender); and "ethnicity*age" interaction (dependent variable = ethnicity, age ethnicity*age). Because of the observed gender- and ethnicity interaction, and the differences in FFM and FFMI in the initial analyses, final multiple regression models to test for the associations of ethnicity and age, with FFM and FFMI were conducted separately for each gender. The model for FFM included adjustment for BMI. However, because FFMI already adjusted for height, the regression model for FFMI did not include height. This precaution was taken to avoid co-linearity and the bias of structural relationship between FFMI and height. All initial analyses were performed using the Statistical Analysis System (SAS) package [14]. Given the multistage complex survey design of NHANES III, final analyses were conducted in SUDAAN (Proc Descript; Proc Regression) to produce a more robust variance estimate [15]. Sample weights were included in the final analysis to account for over sampling and non-response adjustment.
Results
Subject characteristics
A total of 4,072 Whites (2106 men and 1966 women) and 1731 Blacks (948 men and 783 women) were included in the final analyses (Table 1). Although Blacks had overall higher body weight, standing height and BMI compared to Whites, these differences did not reach statistical significance. Among women, Blacks weighed more, were taller, and had higher body mass index than white women. In contrast, white men had higher waist-hip-ratio and triceps skinfold than black men.
Table 1 Anthropometric measurements and other characteristics of the study sample. The Third National Health and Nutrition Examination Survey (NHANES III 1988–94).
Characteristics Non-Hispanic Whites African Americans
Men Women Men Women
Mean ± SE Mean ± SE Mean ± SE Mean ± SE
(N = 2,106) (N = 1,966) (N = 948) (N = 783)
Age (years)* 65.94 (14.2) 66.62 (14.5) 58.41 (12.9) 58.51 (13.6)
Body Weight (kg)* 81.20 (15.7) 68.28 (16.0) 81.40 (17.8) 78.64 (20.0)
Standing Height (cm) 174.00 (7.1) 159.54 (7.0) 174.48 (7.4) 161.91 (6.5)
Body Mass Index (kg/m2) 26.75 (4.5) 26.78 (5.8) 26.66 (5.2) 29.95 (7.2)
Triceps Skinfold (mm) 13.49 (5.9) 23.29 ± (8.1)* 12.34 (6.8) 26.08 (9.2)*
Suprailiac Skinfold (mm) 20.41 (8.9) 20.03 (9.6) 20.40 (10.6) 25.04 (10.3)*
Thigh Skinfold (mm) 15.09 (7.6) 30.40 (8.8) 12.48 (7.1) 29.40 (9.7)
Body Density (g/cm3) 1.05 (0.01) 1.03 (0.0) 1.05 (0.0) 1.03 (0.0)
Fat Free Mass (kg) 64.26 (0.3) 46.65 (0.2)* 63.87 (0.4) 49.53 (0.4)*
Fat Free Mass Index (kg/m2) 20.81 (0.1) 17.91 (0.1)* 20.82 (0.1) 18.82 (0.1)*
Note: Values are mean (SE); * Indicates statistical significant difference for within gender comparison of Non-Hispanic Whites with non-Hispanic African Americans at two-sided p < 0.05.
Ethnicity-based comparison
Collectively, mean FFM and FFMI were similar between Blacks and Whites. However, when stratified by gender-ethnicity, black women had higher FFM and FFMI than White women (FFM difference: 2.88 ± 0.28 kg; p < 0.01) and (FFMI difference: 0.91 ± 0.10 kg/m2; p < 0.01) (Table 1). The significantly higher FFM and FFMI in black women persisted through all decades of life, except in the 8th decade or higher. Overall, Blacks weighed higher than Whites (difference 5.53 ± 0.44 kg; p < 0.001) and this was explained by higher values in black women (difference 10.35 ± 0.59 kg; p < 0.001) (Table 1).
Age-based comparison
Using age group 40 – 45 as the reference, we observed exponential decline in FFM and body weight with advancing age in men and in women. This decline varied by gender and ethnicity (Figure 1). In Blacks and Whites combined, both FFM and FFMI peaked earlier in men (at age 51–54 yrs) than in women (at age 55–59 yrs). In black men, the decline in FFM started about 5 years earlier (at age 51–54 years) compared to white men (at age 55–59 years), and reached significance level at age 65–69 in black men and age 70–74 in white men.
Figure 1 Age-related changes in fat free mass in older white and black men and women. The Third National Health and Nutrition Examination Survey (NHANES III 1988–94)
Although FFM peaked earlier in white women compared to black women, statistically significant decline was reached earlier in black women at age 70–74 (p = 0.025) compared to age 75–79 in white women (p = 0.01). Initial onset of decline in FFM occurred almost a decade after that of men of similar ethnic background. In both white and black women, decline in FFMI was also noted, but did not attain significance level. Furthermore, decline in FFMI was noted in black men starting at age 60–64 and reaching significance at age 70–74 (p = 0.047). In white men, decline started at age 70–74, and reached significant level in the eight decade (of life (p < 0.01).
Multiple regression analyses
In multivariate analysis for the entire group including ethnicity, age and gender, while adjusting for BMI, FFM significantly associated with ethnicity (p < 0.043), age (p < 0.001), and gender (p < 0.001). In separate analyses for men and for women, FFM significantly associated with ethnicity (p < 0.012) and age (p < 0.001) in women, but only with age (p < 0.001) in men (Table 3). This confirmed our preliminary observation of higher FFM in black women compared to white women, and significant age-related changes in FFM. However, black and white men were similar in FFM. Because FFMI was already adjusted for height, the multivariate model for FFMI was adjusted for body weight instead of BMI. The association of FFMI with ethnicity, age and gender in the entire group was similar to that observed for FFM (all p < 0.001). In men and in women, FFMI significantly associated with ethnicity (p < 0.001; p = 0.003 respectively) and with age (p < 0.001; p = 0.004 respectively) (Table 3).
Table 3 Regression analysis of fat free mass, and fat free mass index by age in white, and black men and women. The Third National Health and Nutrition Examination Survey (NHANES III 1988–94).
Men Women
Characteristics Beta Coefficient Standard Error P Value Beta Coefficient Standard Error P Value
FFM (kg)
Ethnicity -0.338 0.276 0.227 -0.821 0.313 0.012
Age (yrs) -0.109 0.011 <0.001 -0.125 0.021 <0.001
BMI (kg/m2) 1.527 0.039 0.001 1.053 0.032 <0.001
FFMI (kg/m2)
Ethnicity -0.333 0.070 <0.001 -0.456 0.147 0.003
Age (yrs) 0.017 0.003 <0.001 -0.015 0.005 0.004
Body Weight (kg) 0.122 0.003 <0.001 0.479 0.006 <0.001
Note: Statistical significance was set at P ≤ 0.05; Fat free mass model adjusted for body mass index; and fat free mass index model adjusted for body weight. FFM indicates fat free mass; FFMI indicates fat free mass index; and BMI indicates body mass index.
Discussion
Based on a representative national epidemiological data, we observed significantly higher FFM and FFMI in black compared to white women, but not between black and white men. These differences were independent of body weight. Important differences in age-related decline in FFM and FFMI were also noted, starting earlier in black men than white men, but earlier in white women than black women despite a higher decline in black women. Despite the earlier onset of decline of FFM and FFMI in white women, the decline attained significance level earlier in black than white women. Overall, Blacks had greater decline in FFM than Whites.
Chumlea and colleagues reported higher mean FFM in black men and women compared to white men and women respectively [16]. An inverse association of FFM and FFMI with mortality has been reported [4]. It is not known whether this association is driven by differences in FFM and FFMI, or by dissimilarities in the onset of decline of FFM and FFMI with increasing age. Because mortality rates were reportedly higher for black men than for white men, and for black women than white women, [7] mean differences in FFM and FFMI are unlikely to significantly contribute to FFM-related mortality rates in Blacks. Although several causative factors are likely to be contributory to mortality rates in different populations, it is also possible that age-related changes in FFM and FFMI are important factors that must be considered in future studies on mortality and aging. Since clinical interventions to attenuate age-dependent decline in FFM are now available, longitudinal studies to examine the influence of age-related decline in FFM and FFMI on mortality patterns is imperative.
Because of the availability of more sophisticated measures of body composition, there has been a great deal of debate on the clinical utility of anthropometric measurements. Our findings of ethnicity-related differences in FFM in this study, is remarkably similar to recent estimates of FFM by Chumlea et al who used bioelectrical impedance (BIA) data from NHANES III [16]. Chumlea and colleagues provided the first national estimates of body composition as reference, to compare other studies. In fact, the age specific estimates of FFM in the Chumlea's study are concordant with our anthropometric estimates of FFM for white, and for black men and women. The BIA estimates of FFM, together with our anthropometric-derived FFM from the NHANES III, are analogous to DEXA-derived estimates of FFM by Gallagher et al [17]. The close approximation of our estimates with data from studies that used DEXA to estimate FFM supports the use of a generalized equation from anthropometric measurements to estimate FFM in a large representative sample, such as NHANES III, especially when adjusted for age. Despite the known methodological limitations associated with the use of generalized equations, our study provides valuable data on national estimates of FFM against which other future estimates from more robust measures can be compared.
Ethnicity-based differences
Men of both ethnic groups were similar in mean FFM and FFMI, whereas, black women had higher body weight, FFM and FFMI than white women. Marshall et al and others previously reported higher muscle mass in Blacks compared to Whites [18,19]. Comparable to our observation of -2.88 kg mean difference in FFM between white women and black women in this study, Gallagher et al reported -2.22 kg difference in FFM between these 2 groups of women [17]. Also, Gasperino et al observed greater appendicular muscle mass in black women relative to white women [20]. While black women in our study were significantly younger than white women, the presence of differences in FFM that were maintained across several decades of life, argue against differences in age as a likely explanation for the higher FFM in black women. Though Blacks in the study were slightly heavier, and anthropometric measurements have been reported to over-estimate body density at extremes of body fatness, we found no statistically significant difference in mean BMI (a measure of body fatness) between white and black women. This suggests that bias from differences in body fatness may not explain the observed differences in FFM between the two groups. Additionally, the final multiple regression model confirmed the association of FFMI with ethnicity in men and in women, and of FFM with ethnicity in women only (Table 3).
Although, knowledge of the exact mechanisms leading to ethnic differences in muscle mass remains rudimentary, differences in androgenic activity have been suggested to play an important role. For example, Dowling and Pi-Sunyer reported higher levels of testosterone and sex hormone binding globulin in obese black women compared to obese white women [21]. Moreover, a 3- to 4-fold higher allele frequency in the human myostatin gene (GDF8 – a gene which codes for muscle mass) was reported in Blacks compared to Whites [22]. Higher FFM and FFMI in black women compared to white women despite lower levels of physical activity (a major determinant of FFM) lend additional support to the presence of important contributing factors at the molecular level.
Age-related changes
FFM declined with advancing age starting earlier in black men compared to white men, but earlier in white women than black women (Figures 1). Using age as a continuous variable, Regression analysis showed significant decline in FFM and FFMI with increasing age in women (P < 0.001; P = 0.004 respectively) after accounting for the influence of body weight and ethnicity (Table 3). In a similarly adjusted model for men, decline in FFM and FFMI were associated with escalating age (P < 0.001; P < 0.001 respectively). This age-related decline in FFM and FFMI were greater for black compared to white women. In men, only FFMI (P < 0.001) but not FFM (P = 0.227) showed a greater decline in black men than white men. Although findings similar to some of our current observations have been reported, as far as we know, we are first to report ethnicity-related differential decline in FFM and FFMI with advancing age in the same study, using anthropometric measurements from a nationally representative sample (NHANES III). While our estimates of FFM for white men below age 70 are in agreement with reports from the Fels study by Guo and colleagues[23], our estimates for women were slightly higher, but within reported margins of error. Further, the onset of age-related decline in FFM at about age 50 in white men and women in our study are in concordance with the Fels study. The earlier onset of decline in FFM in black men is particularly interesting, given the similarity in mean body weight, standing height, BMI, suprailiac skinfold and body density to that of white men (Table 1). Similarly, white and black women did not differ significantly with respect to mean standing height, thigh skinfold and body density. Of note, is the observation of differential onset and decline in FFM and FFMI in this study that appears to parallel mortality patterns for the United States [7]. While underlying disease processes could potentially explain this observation, it is also possible that age-related decline in FFM and FFMI compared to absolute values may have more important implications for health outcomes. Moreover, low levels of physical activity in Blacks [24] may accelerate the process of sarcopenia and unmask several age-related sub-clinical illnesses. Alternatively, differential reduction in androgen and testosterone, reduction in responsiveness to trophic hormones [3], growth hormone [25], and insulin-like growth factors [3] with escalating age may play important roles. The relative increase in FFMI in white women is perhaps related to increased risk of osteoporosis-related compression fracture, reduction in standing height, and relative increase in FFMI. Independent reports from Schutz et al, and Hughes et al lend support to the observed age-related changes in FFMI in white women [26,27].
While decline in FFM in women mirrored changes in body weight, we found no consistent pattern of differential weight decline between white men and black men. The finding in women is particularly remarkable, since heavier body weight requires higher FFM for support (Table 2). Nonetheless, a greater reduction in body weight than can be explained by losses in FFM might suggest a greater reduction in absolute body fat with increasing age. Although it is possible that FFMI may overcompensate for standing height, the lack of significant correlation between FFMI and standing height in this sample makes this unlikely.
Table 2 Distributions of fat free mass, fat free mass index and body weight in white and black men, and in women. The Third National Health and Nutrition Examination Survey (NHANES III 1988–94).
Age Groups (yrs) Sample Size Whites Mean ± SE African Americans Mean ± SE P Value
FFMI (kg/m2)
Men 40–49 708 20.64 ± 0.12 20.91 ± 0.14 0.153
50–59 573 21.13 ± 0.11 20.92 ± 0.15 0.265
60–69 718 21.00 ± 0.12 20.92 ± 0.17 0.704
70–79 600 20.68 ± 0.14 20.33 ± 0.15 0.059
> 80 455 19.93 ± 0.12 19.89 ± 0.32 0.904
Women 40–49 607 17.49 ± 0.10 18.70 ± 0.20 <0.001
50–59 506 18.07 ± 0.17 18.71 ± 0.22 0.026
60–69 581 18.09 ± 0.12 19.26 ± 0.29 <0.001
70–79 605 18.21 ± 0.11 18.93 ± 0.25 0.014
> 80 450 18.09 ± 0.09 18.34 ± 0.32 0.432
FFM (kg)
Men 40–49 708 64.83 ± 0.52 65.30 ± 0.61 0.550
50–59 573 65.94 ± 0.38 64.29 ± 0.59 0.012
60–69 718 64.42 ± 0.42 63.23 ± 0.66 0.139
70–79 600 61.50 ± 0.43 60.36 ± 0.65 0.124
> 80 455 57.74 ± 0.46 57.23 ± 1.32 0.715
Women 40–49 607 46.90 ± 0.35 50.28 ± 0.66 <0.001
50–59 506 47.66 ± 0.46 49.59 ± 0.69 0.013
60–69 581 46.94 ± 0.40 50.15 ± 0.69 0.001
70–79 605 45.59 ± 0.39 48.01 ± 0.62 0.002
> 80 450 43.83 ± 0.27 45.04 ± 0.83 0.143
Body Weight (kg)
Men 40–49 708 83.76 ± 1.00 82.10 ± 1.00 0.252
50–59 573 85.35 ± 0.66 81.13 ± 1.03 <0.001
60–69 718 83.94 ± 0.73 80.41 ± 1.13 0.013
70–79 600 78.92 ± 0.70 76.91 ± 1.09 0.097
> 80 455 73.25 ± 0.63 72.77 ± 2.07 0.834
Women 40–49 607 65.71 ± 0.66 71.71 ± 1.41 <0.001
50–59 506 68.47 ± 0.93 71.83 ± 1.38 0.031
60–69 581 66.86 ± 0.68 72.33 ± 1.09 <0.001
70–79 605 64.11 ± 0.74 68.44 ± 1.07 0.002
> 80 450 60.46 ± 0.45 61.93 ± 1.62 0.365
Note: statistical significance was set at P ≤ 0.05; Whites and African Americans refer to respective ethnic groups of non-Hispanic origin. FFM indicates fat free mass; FFMI indicates fat free mass index; and BMI indicates body mass index.
Within each ethnic group, the decline in FFM and FFMI started earlier in men than women (Figure 1). This finding is consistent with published reports from others who showed linear decline in skeletal muscle mass with advancing age[28,29]. Unfortunately, many of the previous studies lacked the ethnic stratification, sample size and representativeness of the current study. Similar to our findings of gender-based differential decline in FFM, a Japanese study reported decreases in lean body mass in Japanese men but not in women[30]. While our current anthropometric estimates of FFM and FFMI from the NHANES III study may be at variance with estimates from a limited number of studies that showed loss of skeletal muscle mass in the 3rd decade of life[17], several other studies that used more robust methods to estimate skeletal muscle mass[31,32,23] support our current observations. Altogether, our study supports the presence of higher mean FFM in black women than white women, earlier onset of decline in FFM and FFMI in black men than white men respectively (Figures 1 and 2), and greater age-related decline in black men and women than in white men and women. Additionally, it provides support for the use of anthropometric measurements to estimate FFM in large epidemiological studies.
Figure 2 Age-related changes in fat free mass index in older white and black men and women. The Third National Health and Nutrition Examination Survey (NHANES III 1988–94)
Limitations
The observations from this study are subject to a number of limitations. First, at extremes of BMI and body fat, there is bias in the use of skinfold-thickness measurements to estimate body density from which percent body fat is determined. This concern is heightened in the aged because of increased centripetal fat distribution. Fortunately, Pollock's equation included adjustment for age and the final regression model in our study was additionally adjusted for BMI. While population-specific equations are appropriate alternatives and sometimes provide more accurate estimates, they are limited in generalizability. Though the equation used in the present study has greater applicability to women than men, its selection was based on the available anthropometric measurements in the NHANES III data.
Because the Pollock equation was not derived from the NHANES sample, it is possible that it may not predict body density with the same degree of precision compared to the original parent sample [33]. Further, the cross-sectional nature of the NHANES III data makes it possible that changes in FFM with advancing age may have been underestimated, given known estimates from longitudinal studies [34]. These concerns are minimized by the similarities of our findings with that of studies using Bioelectrical Impedance and DEXA to estimate FFM, and therefore, may not reduce the significance of the population-based observation from the current study. Inter-, intra-observer's and instrument bias appear unlikely in the NHANES given the uniform procedural standards for anthropometric measurements. Despite these limitations, the use of anthropometric measurements to estimate FFM and changes in FFM is simple, low cost, and has a predictive value of within 3 to 4 percent for 70 percent of the population [33,35] especially when adjusted for age. Finally, the unique strengths of the study include large sample size, standardized procedure for anthropometric measurements and its applicability to the entire United States population.
Conclusion
In conclusion, we provide additional support for the differential effect of ethnicity on the loss of FFM and FFMI as the population ages. Regardless of the explanation for these differences, losses in muscle mass have several important health consequences, and perhaps more so for FFM decline. Most notable among these is increased functional dependency and mortality. This underscores the public health imperative of an ongoing epidemiological surveillance of the variations in FFM. If properly implemented, public health strategies may facilitate the design of targeted optimization of functional independency as the United States population ages.
Abbreviations
Fat free mass (FFM); Fat free mass index (FFMI); Body mass index (BMI); The Third National Health and Nutrition Examination Survey (NHANES III)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TOO conceived the study, participated in the design, performed statistical analysis and writing of the manuscript; MHA participated in the design, performed statistical analysis and writing of the manuscript; VB participated in the analysis and writing of the manuscript; RGA participated in conceiving and drafting of the manuscript, AA participated in the drafting of the manuscript, CNR participated in study design, statistical analyses and drafting of the manuscript
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Supported by grant #AG00980 (NIA) to Obisesan TO, and RR10284 (NCRR). Charles Rotimi was supported by grants from NIDDK (DK-54001), MBRS (S06-GM008016) and RMCI program of the NCRR at NIH
==== Refs
Evans WJ Exercise and nutritional needs of elderly people: effects on muscle and bone Gerodontology 1998 15 15 24 10530167 10.1111/j.1741-2358.1998.00015.x
Lynch NA Metter EJ Lindle RS Fozard JL Tobin JD Roy TA Fleg JL Hurley BF Muscle quality. I. Age-associated differences between arm and leg muscle groups J Appl Physiol 1999 86 188 194 9887130
Baumgartner RN Waters DL Gallagher D Morley JE Garry PJ Predictors of skeletal muscle mass in elderly men and women Mech Ageing Dev 1999 107 123 136 10220041 10.1016/S0047-6374(98)00130-4
Zhu S Heo M Plankey M Faith MS Allison DB Associations of body mass index and anthropometric indicators of fat mass and fat free mass with all-cause mortality among women in the first and second National Health and Nutrition Examination Surveys follow-up studies Ann Epidemiol 2003 13 286 293 12684196 10.1016/S1047-2797(02)00417-9
Schutte JE Townsend EJ Hugg J Shoup RF Malina RM Blomqvist CG Density of lean body mass is greater in blacks than in whites J Appl Physiol 1984 56 1647 1649 6735823 10.1063/1.334152
Zhu S Heo O Plankey M Faith MS Allison DB Associations of Body Mass Index and Anthropometric Indicators of Fat Mass and Fat Free Mass with All-cause Mortality among Women in the First and Second National Health and Nutrition Examination Surveys Follow-up Studies Ann Epidemiol 2003 13 286 293 12684196 10.1016/S1047-2797(02)00417-9
Survey USG U.S. Mortality Database, 1988-1992 US Geological Survey 1999 Reston, VA
Plan and operation of the Third National Health and Nutrition Examination Survey, 1988-94. Series 1: programs and collection procedures Vital Health Stat 1 1994 1 407
Lohman TG Roche AF Martorell R Anthropometric standardization reference manual. Human Kinetic Books 1988 Champaign, IL
Jackson AS Pollock ML Ward A Generalized equations for predicting body density of women Med Sci Sports Exerc 1980 12 175 181 7402053
Brozek J Body composition: models and estimation equations Am J Phys Anthropol 1966 24 239 246 5938201
VanItallie TB Yang MU Heymsfield SB Funk RC Boileau RA Height-normalized indices of the body's fat-free mass and fat mass: potentially useful indicators of nutritional status Am J Clin Nutr 1990 52 953 959 2239792
Binder DA On the variances of asymptomatically normal estimates from complex surveys Stat Rev 1983 279 292
Software SASSTAT Changes and Enhancements, Release 8.0. SAS Institute, Inc 2001 Cary, NC
Shah BV Barnwell BG Bieler GS Software for Analysis of Correlated Data, Release 8.0 SUDAAN User's Manual: 2002 Research Triangle Park, NC: Reserach Triangle Institute
Chumlea WC Guo SS Kuczmarski RJ Flegal KM Johnson CL Heymsfield SB Lukaski HC Friedl K Hubbard VS Body composition estimates from NHANES III bioelectrical impedance data Int J Obes Relat Metab Disord 2002 26 1596 1609 12461676 10.1038/sj.ijo.0802167
Gallagher D Visser M De Meersman RE Sepulveda D Baumgartner RN Pierson RN Harris T Heymsfield SB Appendicular skeletal muscle mass: effects of age, gender, and ethnicity J Appl Physiol 1997 83 229 239 9216968
Marshall JA Grunwald GK Donahoo WT Scarbro S Shetterly SM Percent body fat and lean mass explain the gender difference in leptin: analysis and interpretation of leptin in Hispanic and non-Hispanic white adults Obes Res 2000 8 543 552 11156429
Roubenoff R Hughes VA Sarcopenia: current concepts J Gerontol A Biol Sci Med Sci 2000 55 M716 24 11129393
Gasperino JA Wang J Pierson RNJ Heymsfield SB Age-related changes in musculoskeletal mass between black and white women Metabolism 1995 44 30 34 7854161 10.1016/0026-0495(95)90285-6
Dowling HJ Pi-Sunyer FX Race-dependent health risks of upper body obesity Diabetes 1993 42 537 543 8454103
Ferrell RE Conte V Lawrence EC Roth SM Hagberg JM Hurley BF Frequent sequence variation in the human myostatin (GDF8) gene as a marker for analysis of muscle-related phenotypes Genomics 1999 62 203 207 10610713 10.1006/geno.1999.5984
Guo SS Zeller C Chumlea WC Siervogel RM Aging, body composition, and lifestyle: the Fels Longitudinal Study Am J Clin Nutr 1999 70 405 411 10479203
Crespo CJ Keteyian SJ Heath GW Sempos CT Leisure-time physical activity among US adults. Results from the Third National Health and Nutrition Examination Survey Arch Intern Med 1996 156 93 98 8526703 10.1001/archinte.156.1.93
Rudman D Feller AG Cohn L Shetty KR Rudman IW Draper MW Effects of human growth hormone on body composition in elderly men Horm Res 1991 36 Suppl 1 73 81 1806490
Schutz Y Kyle UU Pichard C Fat-free mass index and fat mass index percentiles in Caucasians aged 18-98 y Int J Obes Relat Metab Disord 2002 26 953 960 12080449 10.1038/sj.ijo.0802037
Hughes VA Frontera WR Wood M Evans WJ Dallal GE Roubenoff R Fiatarone Singh MA Longitudinal muscle strength changes in older adults: influence of muscle mass, physical activity, and health J Gerontol A Biol Sci Med Sci 2001 56 B209 17 11320101
Melton LJ Khosla S Crowson CS O'Connor MK O'Fallon WM Riggs BL Epidemiology of sarcopenia J Am Geriatr Soc 2000 48 625 630 10855597
Proctor DN Melton LJ Khosla S Crowson CS O'Connor MK Riggs BL Relative influence of physical activity, muscle mass and strength on bone density Osteoporos Int 2000 11 944 952 11193247 10.1007/s001980070033
Tsunenari T Tsutsumi M Ohno K Yamamoto Y Kawakatsu M Shimogaki K Negishi H Sugimoto T Fukase M Fujita T Age- and gender-related changes in body composition in Japanese subjects J Bone Miner Res 1993 8 397 402 8475789
Cohn SH Vartsky D Yasumura S Sawitsky A Zanzi I Vaswani A Ellis KJ Compartmental body composition based on total-body nitrogen, potassium, and calcium Am J Physiol 1980 239 E524 30 7446727
Gallagher D Heymsfield SB Muscle distribution: variations with body weight, gender, and age Appl Radiat Isot 1998 49 733 734 9569594 10.1016/S0969-8043(97)00096-1
Brodie D Moscrip V Hutcheon R Body composition measurement: a review of hydrodensitometry, anthropometry, and impedance methods Nutrition 1998 14 296 310 9583375 10.1016/S0899-9007(97)00474-7
Janssen I Heymsfield SB Wang ZM Ross R Skeletal muscle mass and distribution in 468 men and women aged 18-88 yr J Appl Physiol 2000 89 81 88 10904038
Brodie DA Techniques of measurement of body composition. Part I Sports Med 1988 5 11 40 3278354
| 15840167 | PMC1097739 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Apr 19; 5:41 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-41 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-431584769510.1186/1471-2458-5-43Study ProtocolRationale, design and conduct of a comprehensive evaluation of a school-based peer-led anti-smoking intervention in the UK: the ASSIST cluster randomised trial [ISRCTN55572965] Starkey Fenella [email protected] Laurence [email protected] Rona [email protected] Mark [email protected] Michael [email protected] (A Stop Smoking In Schools Trial) 1 Department of Social Medicine, University of Bristol, Bristol, England, UK2 Cardiff Institute of Society, Health and Ethics, Cardiff University, Cardiff, Wales, UK3 Faculty of Social Sciences, University of Glasgow, Glasgow, Scotland, UK2005 22 4 2005 5 43 43 31 3 2005 22 4 2005 Copyright © 2005 Starkey et al; licensee BioMed Central Ltd.2005Starkey 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
To date, no school-based intervention has been proven to be effective in preventing adolescent smoking, despite continuing concern about smoking levels amongst young people in the United Kingdom. Although formal teacher-led smoking prevention interventions are considered unlikely to be effective, peer-led approaches to reducing smoking have been proposed as potentially valuable.
Methods/design
ASSIST (A Stop Smoking in Schools Trial) is a comprehensive, large-scale evaluation to rigorously test whether peer supporters in Year 8 (age 11–12) can be recruited and trained to effect a reduction in smoking uptake among their fellow students. The evaluation is employing a cluster randomised controlled trial (RCT) design with secondary school as the unit of randomisation, and is being undertaken in 59 schools in South East Wales and the West of England. Embedded within the trial are an economic evaluation of the intervention costs, a process evaluation to provide detailed information on how the intervention was delivered and received, and an analysis of social networks to consider whether such a peer group intervention could work amongst schoolchildren in this age group.
Schools were randomised to either continue with normal smoking education (n = 29 schools, 5562 students), or to do so and additionally receive the ASSIST intervention (n = 30 schools, 5481 students). No schools withdrew once the trial had started, and the intervention was successfully implemented in all 30 schools, with excellent participation rates from the peer supporters. The primary outcome is regular (weekly) smoking, validated by salivary cotinine, and this outcome has been obtained for 94.4%, 91.0% and 95.6% of eligible students at baseline, immediate post-intervention, and one-year follow-up respectively.
Discussion
Comprehensive evaluations of complex public health interventions of this scale and nature are rare in the United Kingdom. This paper demonstrates the feasibility of conducting cluster RCTs of complex public health interventions in schools, and how the rigour of such designs can be maximised both by thorough implementation of the protocol and by broadening the scope of questions addressed in the trial by including additional evaluative components.
==== Body
Background
Smoking and young people in the United Kingdom
Smoking rates amongst 11 to 16 year-olds in the United Kingdom continue to be a cause for concern, with increases in these during the early to mid-1990s [1,2]. Since the late 1990s, rates have stabilised at around 10 per cent [3-6]. Nonetheless, there is a very strong policy initiative from the British Government to lower these rates to 9 per cent or less by the year 2010 [7]. The means by which such a downward trend could be achieved, however, are unclear as to date no intervention has been shown, via rigorous evaluation, to be effective in reducing smoking uptake or encouraging smoking cessation amongst young people [8-10].
Peer-led interventions
Many anti-smoking programmes for young people are school-based, due to the 'captive audience' at which such interventions can be targeted over several years [11-13]. However, evidence suggests that formal teacher-led interventions are unlikely to be effective in reducing smoking [14,15]. In contrast, peer-led approaches, particularly those which take an informal educational approach, have been identified as potentially promising for use in anti-smoking initiatives [16,17].
ASSIST (A Stop Smoking in Schools Trial)
This paper reports upon the design of ASSIST (A Stop Smoking in Schools Trial), a large-scale comprehensive evaluation of an intervention which used such an informal peer-led approach to smoking prevention in schools. The intervention approach used by ASSIST was developed in a feasibility study undertaken in Wales [18] and was based upon the 'diffusion of innovation' model [19]. This has as its exemplar the US 'gay hero' sexual health promotion programme [20], whereby the diffusion of new behavioural norms through social networks was effected by locally influential opinion-formers. The promising results of the feasibility study led the Medical Research Council (MRC) to fund a full-scale evaluation of the intervention. This paper describes and discusses the design of this evaluation and key issues in its successful implementation.
Methods/design
Study design
While a randomised controlled design is being used to test the intervention's effectiveness, it was recognised that this would not allow full and in-depth exploration of the complex processes operating in such an informal peer-led intervention. The incorporation of several other components within the randomised controlled trial, including a detailed process evaluation, an economic evaluation, and an analysis of social networks, has resulted in a comprehensive evaluation that is enabling the study team to answer questions beyond those of effectiveness (see Figure 1).
Figure 1 A Stop Smoking in Schools Trial (ASSIST) research design
Since the focus of the intervention was peer groups within a school year group, the unit of randomisation for evaluation of the intervention must be the school and a cluster randomised design has therefore been adopted.
The ASSIST intervention
The ASSIST intervention trained influential Year 8 students (age 11–12) to use informal contacts with peers in their school year group to encourage them not to smoke. All students in participating schools were asked to nominate students they viewed as influential in different respects, and those nominated students were invited to train to take on the role of 'peer supporter'. Training for these peer supporters was undertaken by health promotion trainers (who had themselves been trained to deliver the intervention in a standardised way). The two-day training course for peer supporters was provided for each school at a venue away from the school premises, and aimed to: increase knowledge about the health, economic, social and environmental risks of smoking; emphasise the benefits of remaining smoke-free; and encourage the development of skills to enable peer supporters to promote non-smoking among their peers. After this training, peer supporters were asked to intervene informally in everyday situations over a ten-week period to encourage their peers not to smoke, and were asked to keep a diary record of these informal conversations. Four follow-up visits to each school were made by the health promotion trainers over this ten-week period to provide further training to the peer supporters and to monitor their progress. Students who submitted a completed diary at the end of this ten-week period were given a £10 gift voucher in recognition of their contribution to the project. This intervention is described in more detail in Audrey et al [21].
The intervention design, formative evaluation and feasibility study for ASSIST were undertaken over a two-year period [18]. Further refinement of the intervention, and of data collection procedures and instruments, was undertaken by the ASSIST team during additional pre-trial piloting work in three schools outside the final trial's geographical areas. This additional piloting work was required to refine the peer supporter nomination process in order to recruit more male peer supporters, to update elements of the intervention, and to undertake a 'dress rehearsal' of both the intervention and evaluation with the new teams involved.
Selection, recruitment and randomisation of schools
In order to achieve representative and generalisable results, the design of ASSIST embraces as wide a range of secondary school settings as possible, including Welsh medium schools, private (fee-paying) schools (attended by a substantial fraction of students in the West of England area), religious schools, single-sex schools, and both urban and rural schools (serving populations with varying levels of social deprivation).
Letters and information sheets explaining the study's aims were sent to head teachers of 223 secondary schools in the former counties of Avon (England), Gwent, Mid Glamorgan, South Glamorgan and West Glamorgan (Wales). Schools expressing an interest (n = 127) were then visited individually by study team members. These meetings were conducted with a member of the school's senior management team, and included detail on the commitment required from schools. The importance of the randomised design was emphasised, and it was made clear to schools that it would be preferable for schools to decline participation than to join the study and then withdraw at a later point. Six schools withdrew before random selection and a further eight were excluded from the trial (since they had fewer than 60 students in Year 8, were special schools, or were involved in another anti-smoking initiative).
One hundred and thirteen schools remained interested in participating, from which the 66 schools required were randomly selected. The 113 schools were divided into four groups: (1) eight independent schools; (2) four Welsh medium schools; (3) four single-sex schools; and (4) 97 mixed-sex state schools. From each of the first three groups, 50 per cent of schools were randomly selected, giving eight schools. The 97 state schools were stratified by country (Wales or England) and whether the proportion of students entitled to free school meals was above or below the national averages (as an indicator of the socio-economic status of the school catchment area). Within these strata, schools were then listed according to Year 8 size. A systematic random sample was taken using a one-in-two sampling fraction, giving a further 50 schools. Two state schools were then randomly selected from each of the four strata defined by country and free school meal entitlement, to bring the total sampled to 66.
After this selection and before randomisation into trial arms, schools were sent a detailed memorandum of understanding to sign, to obtain head teachers' written commitment to the trial. Fifty-nine of these schools returned this document and thereby agreed to be randomised to either the intervention or control arm of the trial.
Stratified randomisation was then used to allocate these schools to either intervention or control group. Within each stratum, each school had a 50 per cent chance of being allocated to either group. Strata were defined using similar criteria to the random selection procedure described above, namely (1) independent schools, (2) Welsh medium schools, and (3) state schools. The two single-sex schools were not treated as a separate stratum since one was a boys' school and the other a girls' school. State schools were further grouped into eight strata, defined by (i) whether they were in Wales or England, (ii) whether they had a year size greater or less than the median (200 students), and (iii) whether they had greater or less than the median proportion of students entitled to free school meals (19%). To ensure concealment, the sequence in which schools were allocated was determined by one of the trial's principal investigators. Unaware of which school was the next to be allocated, a second principal investigator used a pocket calculator to generate a random number which determined the next school's group allocation.
Outcome measures and sample size calculations
The trial's primary outcome measure is smoking prevalence among the high-risk group, i.e. those who, at baseline, had experimented with cigarettes, were ex-smokers, or were occasional (less than weekly) smokers. These students have been chosen as the primary target group because they are at greatest risk of becoming regular smokers, and the feasibility study showed an effect amongst this group [18]. Smoking prevalence is defined as students smoking a cigarette in the previous seven days. A secondary outcome measure is smoking prevalence among the entire year group. These outcome measures are being validated by measurement of salivary cotinine (a metabolite of nicotine), as studies have found cotinine to be the most accurate biomarker of smoke exposure in the previous two to three days [22,23]. Other secondary measures include perceptions of norms regarding adolescent smoking, and intention to quit.
There were 59 schools at baseline data collection, with a mean year size of 187 students. On average, 38.6 per cent (n = 72) of the year group were in the high-risk group of students who, at baseline, had experimented with cigarettes, or were ex-smokers, or were occasional (less than weekly) smokers. The intra-cluster correlation of weekly smoking at baseline, and also of the proportion in the high-risk group, was 0.025. Based on these data, the trial has 80 per cent power to detect a 5.8 per cent difference in weekly smoking among the high-risk group, assuming that weekly smoking prevalence among the high-risk group in control schools at 12-month follow-up is 23 per cent. Among all students, the trial has 80 per cent power to detect a 4.3 per cent difference in weekly smoking, assuming control group prevalence at 12-month follow-up is 15 per cent.
Data collection
In order to obtain these outcome data, smoking behaviour was assessed at baseline with further assessments immediately post-intervention, and then at one year and two years post-intervention. At baseline, all students were asked to complete a questionnaire on smoking and a peer questionnaire, and to provide a saliva sample both for cotinine assay and to minimise reporting bias [24]. The smoking questionnaire drew upon questions about smoking used in the large-scale Office for National Statistics surveys of young people in England [25] and in the World Health Organisation's European Health Behaviour of School-Aged Children Study [26], to enable the results of ASSIST to be compared to national and international findings.
The peer questionnaire asked students to supply names of Year 8 students whom they viewed as influential in different ways. The results were used to identify students to invite to be peer supporters in intervention schools (although the purpose of this activity was not disclosed to students at the time of completion to avoid bias), and to identify a group of potential peer supporters in control schools for comparative purposes.
Subsequent data collection sweeps involve students completing a smoking questionnaire and a social network questionnaire (described later in this paper), and providing a saliva sample. The same core questions on smoking status are used at each follow-up to provide comparable outcome data, while additional questions are included at different follow-ups to explore secondary research questions.
All staff undertaking data collection have been provided with a data collection protocol and have been given training to maximise standardisation of data collection procedures across the trial sites and data collection sweeps.
Statistical analyses
Data analysis is being conducted following a pre-specified analysis plan, agreed by the trial's independent Data Monitoring and Ethics Committee. Confidence intervals for smoking prevalence among all students and among the high-risk group are being estimated using design-weighted survey estimators implemented in Stata, which account for the clustering of students in schools. Estimates of the intervention effect are being obtained at student level using random effects logistic regression models with school as random effect, and all models include the five school-level stratifying variables as covariates (these five variables being binary variables indicating [i] independent school, [ii] less than 19 per cent of students entitled to free school meals, [iii] in England, [iv] Welsh medium, and [v] fewer than 200 students in Year 8).
As well as the student-level analyses reported above, the analysis plan includes school-level analyses that treat the data as repeated cross-sections rather than as an individual student-level cohort. Although they may have less statistical power, these analyses are less likely to suffer from non-response bias, and mean that analysis and inference takes place at the same level as the unit of randomisation and implementation. School-level analysis is being undertaken using appropriately weighted multiple regression analysis of the logarithm of each school's smoking prevalence at baseline and follow-up. Multi-level modelling is also being used to identify any important interactions between school-level factors and student-level effects.
In addition to the design and implementation of this large-scale cluster randomised trial to rigorously test the intervention's effectiveness, further components are nested within the evaluation to enable a comprehensive assessment of the intervention's impact in schools.
Additional components of evaluation
Process evaluation
A detailed process evaluation is being undertaken by dedicated staff which aims to: examine the implementation and receipt of the intervention and its evaluation; explore the context and intensity of the intervention; and document factors external to the intervention which might impact upon both its implementation and its effectiveness. Both qualitative and quantitative methods are being used to collect data from key participants (students, teachers, health promotion trainers and researchers) at each stage of the intervention and evaluation. More details of the process evaluation design and implementation can be found in Parry-Langdon et al [27].
Economic evaluation
An economic evaluation is also being undertaken alongside the trial, relating costs to a range of outcomes in the form of a costs and consequences analysis. Although such analysis is not evaluative in the sense of informing technical or allocative efficiency, it is the most common form of economic study used in health care [28].
Costing is being undertaken using standard methods [29]. As the aim is to estimate the cost of replicating the intervention elsewhere, all costs associated with the evaluation are being excluded. Resources used have been recorded on a weekly basis by ASSIST staff using standardised forms, and include staff time, travel time and distance, consumables, accommodation, and vouchers for peer supporters.
Analysis of social networks
As the ASSIST intervention relies upon diffusing new behavioural norms through existing social networks within the school year group, an important component of the evaluation is to collect data on the nature of participating students' social networks. These data, obtained via questionnaires completed at the post-intervention data collection sweeps, are being used to explore the structural properties of teenage friendship groups (using social network analysis software). The social network data allow examination of the degree to which peer supporters were proximal in social space to those in the high-risk group. In combination with data from the immediate post-intervention follow-up, they also permit analysis of whether students' awareness of the intervention's existence was associated with their having a friendship link with a peer supporter and whether being a peer supporter had any impact on their friendship groups.
Consideration of variability in social networks amongst young people (for example the extent to which young people spend time with those in the same year group, or in other year groups) enhances our understanding of potential differences in the intervention's success in participating schools, therefore contributing to a more comprehensive picture of whether and under what conditions such peer-led diffusion-based interventions might succeed in affecting young people's behaviour.
Participation rates
Participation of schools
ASSIST attracted a high level of interest among the secondary schools initially contacted. No school that participated in baseline data collection withdrew from the project other than due to enforced closure (n = 2). The training and four follow-up sessions comprising the intervention were also successfully implemented in all intervention schools (n = 30).
Participation of students
Response rates for data collection sweeps are very high, as outlined in Figure 2.
Figure 2 Flow of students through trial up to one-year follow-up data collection
The slightly lower participation rate at immediate post-intervention follow-up was due to the postal method of absentee data collection used only at this sweep, which yielded lower returns.
With regard to the students invited to be peer supporters (n = 942), 92 per cent (n = 867) of those invited agreed to attend the training course. Eight hundred and forty eight young people attended this training and 835 (98%) of them then consented to undertake the intervention as peer supporters (with an equal gender balance of 418 boys and 417 girls). Eighty-two per cent (n = 687) of the peer supporters completed and submitted their diary as requested, suggesting that such an intervention can be successfully implemented in a school setting.
As can be seen, extremely high levels of participation are being maintained throughout the study, demonstrating that comprehensive evaluations of this kind can be successfully conducted in secondary school settings if adequate rigour is adopted and appropriate support provided to both schools and participating students.
Discussion
Key issues in conduct of evaluation
Particular issues arise in conducting such an evaluation in secondary school settings, which are likely to be of interest to researchers considering undertaking similar work. These issues include how to minimise contamination across trial arms, how to develop and maintain positive working relationships with schools, how to obtain consent from participants, and how to undertake data collection in schools sensitively and successfully.
Contamination
In order to minimise contamination of control schools and their students with information and awareness of the intervention, and to prevent declining commitment to the project among control schools, all schools are being asked not to publicise their involvement in the trial until after the one-year follow-up data collection. Neither schools nor local health promotion agencies have been informed which schools are participating in the trial. In this way, the study team aims to ensure that details of the intervention, for example training activities undertaken, are not provided to control schools and used by teachers as part of their health education activities.
Involving schools in a randomised controlled trial
Peterson et al [9] suggest key principles for developing and maintaining positive collaborative research relationships with schools, which include keeping schools informed, minimising the work burden upon schools, maintaining regular contact with schools, executing tasks as planned, demonstrating sensitivity and responsiveness to schools' needs, and expressing appreciation (p.154). These are all principles used by the ASSIST teams to maximise the recruitment and retention of schools.
For example, after randomisation each participating school received a planning visit from the research team, conducted with the head or deputy head teacher and a designated school contact. These meetings involved discussion of: information provision to school governors, staff and parents; planning for study activities; responsibilities of the school contact person; process evaluation activities; provision to schools of payment for teacher cover and administration; and media issues and confidentiality of study participants. This meeting was intended to emphasise the school's commitment to participating in a rigorously conducted trial, but also to focus upon ways in which the study teams aimed to place a minimal burden on schools involved. Schools' individual needs and circumstances have been taken into account as far as possible (for example if extra data collection staff would be required to provide support in schools with high numbers of students with learning difficulties).
This planning stage started the process of building good working relationships with school staff who are acting as key contacts for the study teams throughout the trial. Relationships with these key school staff are being maintained via regular contact, newsletters to schools on the study's progress, and the execution of all study activities as planned and agreed with schools (to avoid disruption to their timetabling processes).
Consent
Obtaining consent for participation in such a community-based trial with young people entails a different process to that which might be used in a drug trial with adult participants. The first phase of consent procedures implemented in ASSIST was to ask head teachers to give written consent for the school's participation in the trial, prior to randomisation. Subsequently (and at least two weeks before baseline data collection), standardised information letters were posted by schools (not distributed via students) to parents/carers of all Year 8 students, explaining the study and enclosing a reply slip to return if they did not want their child to participate. Parents/carers were also given the opportunity to contact the research team at any time to discuss the trial, and a total of fifteen did so.
This passive, 'opt-out' method of parental permission has been found to be an ethical and appropriate way of informing parents/carers of such 'low-risk' prevention research, and to avoid the problems of low response rate, significant sampling bias and under-reporting of illicit activities encountered in research which has used active consent procedures with parents/carers of young people involved in school-based research [9,30].
In order to ensure acceptability of such passive parental permission procedures, these should be combined both with an opportunity for young people to refuse to participate in research, and with detailed procedures to safeguard confidentiality of data [31]. The third phase of consent procedures therefore occurs at the data collection visits, when all students are given an information leaflet (the contents of which are explained to them) and provided with the opportunity to ask questions about the study. They are then asked to complete and sign an assent form that gives them the option to either take part in or refuse some or all of the activities. Confidentiality safeguards at data collection sweeps are described below. These procedures are administered individually for all new students at subsequent data collection sweeps and have been agreed with the Wales Multi-Centre Research Ethics Committee (MREC), which reviewed the trial protocol.
Undertaking data collection in schools
It has proved crucial to adopt 'user-friendly' data collection methods with participating students in order to build trust and assure confidentiality, thereby maximising participation and accuracy of the data collected.
As recommended by educational researchers as good practice [9,31], questionnaires and saliva samples are collected in schools by study staff, either in classrooms with individual groups or in halls with larger numbers of students. Teachers are asked to be present but not to become involved in the data collection process itself, to reassure students that their answers are not seen by school staff. At the start of each data collection session, study staff provide an assurance of confidentiality to students, explaining that their individual results are only seen by university staff and are not given either to the school or to their parents/carers. An assurance of anonymity is also provided, with an explanation that both their questionnaires and salivette are marked with an identification number so that their names are not required on these items, and that only university staff are able to link these identification numbers with participating students' names.
The questionnaire completion procedures are explained at the start of the session, and the saliva collection procedure is explained and demonstrated by study staff prior to the administration of this activity. Students are encouraged to ask questions throughout the session if they do not understand or are unclear about any questionnaire items. Teachers are asked to identify students with reading or concentration difficulties who might require extra support from study staff to participate. When students have completed their questionnaires, they seal these into individual envelopes and return them directly to study staff, to reinforce their confidential nature.
To maximise participation, study staff return to each school approximately two weeks after the data collection session to collect data from students who were absent. The data collection methods described above are also used at these absentee sessions. An alternative method of collecting data from absent students was tested during the immediate post-intervention data collection sweep, whereby school staff were given absent students' questionnaires and asked to arrange their completion and return (by 'freepost'). This method was found to be less effective in maximising participation rates and its use has therefore been discontinued.
Such carefully designed data collection procedures, which aim to maximise young people's confidence in the confidential and anonymous nature of the data provided, should be regarded as good practice when undertaking school-based research. As can be seen from the participation rates outlined above, these procedures are leading to very high response rates from students.
Evaluation results will be presented in future publications as they become available. However, it is hoped that the clear exposition of the study design and methodology in this paper, reinforced as good practice by the high recruitment and participation rates being achieved, will help to inform planning and implementation of future school-based cluster randomised trials, and facilitate the design of comprehensive, rigorous evaluations of complex public health interventions.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FS: drafted the paper, participant in the design of research tools, data collection, and analysis of the data.
LM: participant in the design, co-ordination, data collection and statistical analyses, and involved in revising the paper for important intellectual content.
RC: participant in the design, co-ordination, data collection and statistical analyses, and involved in revising the paper for important intellectual content.
MS: participant in the design of research tools and data collection, contributed to a draft of this paper.
MB: participant in the design, co-ordination and data collection, substantial contributor to the development of the experimental intervention, and involved in revising the paper for important intellectual content.
All authors read and approved the final manuscript.
Members of the ASSIST team
Lead Investigators: Laurence Moore (CISHE, Cardiff University), Rona Campbell (Department of Social Medicine, University of Bristol), Nina Parry-Langdon (Health Promotion Division, Welsh Assembly Government), Mick Bloor (Faculty of Social Sciences, University of Glasgow).
Bristol evaluation team: Fenella Starkey, Suzanne Audrey
Cardiff evaluation team: Mark Sidaway, Jo Holliday.
Health promotion trainers: Kathleen Cordall, Lin Cooper, Heather Anderson-Paine, Nicola Hewer, Lorna Coombes, Rob Sage.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We would like to thank all the students, teachers and schools for participating in this research so willingly. We are very grateful to members of the Trial Steering Committee and Data Monitoring and Ethics Committee, and of the project's Management Group, for their excellent advice. Pete Shiarly's input on database design and data management is invaluable. Valerie Karatzas, Lisa Baker, Linda Esprit and Zoë Macdonald have provided much valued clerical and administrative support to ASSIST. The project has been made possible by funding from the Medical Research Council (grant no. G9900538).
==== Refs
Jarvis L Smoking among Secondary School Children in 1996 1997 London: HMSO
National Assembly for Wales Health Statistics for Wales 1998 1999 Cardiff: National Assembly
Boreham R Shaw A Smoking, drinking and drug use among young people in England in 2000 2001 London: The Stationery Office
Department of Health Smoking, drinking and drug use among young people in England in 2002 2003 London: The Stationery Office
National Centre for Social Research/National Foundation for Educational Research Drug use, smoking and drinking among young people in England in 2003: Headline figures 2004 Department of Health website: National Centre for Social Research/National Foundation for Educational Research
Roberts C Kingdon A Parry-Langdon N Bunce J Young people in Wales: findings from the Health Behaviour in School-aged Children (HBSC) study 1986–2000 2002 Cardiff: Research and Evaluation Branch, Health Promotion Division, Welsh Assembly Government
Department of Health Smoking Kills: a White Paper on Tobacco 1998 London: The Stationery Office
Nutbeam D Macaskill P Smith C Simpson JM Catford J Evaluation of two school smoking education programmes under normal classroom conditions Brit Med J 1993 306 102 7 8435601
Peterson A Mann S Kealey K Marek P Experimental design and methods for school-based randomized trials: experience from the Hutchinson Smoking Prevention Project (HSPP) Control Clin Trials 2000 21 144 165 10715512 10.1016/S0197-2456(99)00050-1
Thomas R School-based programmes for preventing smoking (Cochrane Review) 2002 The Cochrane Library, 4. Oxford: Update Software
Department of Health Saving Lives: Our Healthier Nation 1999 London: The Stationery Office
National Assembly for Wales Promoting Health and Well-being: Implementing the National Health Promotion Strategy 2001 Cardiff: Health Promotion Division, The National Assembly for Wales
Scottish Office Towards a Healthier Scotland 1999 London: The Stationery Office
Aveyard P Cheng KK Almond J Sherratt E Lancashire R Lawrence T Griffin C Evans O Cluster randomised controlled trial of expert system based on the transtheoretical ('stages of change') model for smoking prevention and cessation in schools Brit Med J 1999 319 948 53 10514156
Black DR Tobler NS Sciacca JP Peer helping/involvement: an efficacious way to meet the challenge of reducing alcohol, tobacco, and other drug use among youth? J Sch Health 1998 68 87 93 9608448
Harden A Oakley A Oliver S Peer-delivered health promotion for young people: A systematic review of different study designs Health Educ J 2001 60 339 353
Wilton T Keeble S Doyal L Walsh A The Effectiveness of Peer Education in Health Promotion: Theory and practice 1995 Bristol: University of the West of England
Bloor M Frankland J Parry-Langdon N Robinson M Allerston S Catherine A Cooper L Gibbs L Gibbs N Hamilton-Kirkwood L Jones E Smith RW Spragg B A controlled evaluation of an intensive, peer-led, schools-based, anti-smoking programme Health Educ J 1999 58 17 25
Rogers E The Diffusion of Innovation 1983 3 New York: Free Press
Kelly JA Murphy DA Sikkema KJ McAuliffe TL Roffman RA Solomon LJ Winett RA Kalichman SC Community HIV Prevention Research Collaborative Randomised, controlled, community-level HIV-prevention intervention for sexual risk behaviour among homosexual men in US cities Lancet 1997 350 1500 1505 9388397 10.1016/S0140-6736(97)07439-4
Audrey S Cordall K Moore L Cohen D Campbell R The development and implementation of a peer-led intervention to prevent smoking among secondary school students using their established social networks Health Educ J 2004 63 266 284
Caraballo RS Giovino GA Pechacek TF Self-reported cigarette smoking vs. serum cotinine among U.S. adolescents Nicotine Tob Res 2004 6 19 25 14982684 10.1080/14622200310001656821
Rebagliato M Validation of self reported smoking J Epidemiol Community Health 2002 56 163 164 11854332 10.1136/jech.56.3.163
Murray DM Perry CL The measurement of substance use among adolescents: When is the 'Bogus Pipeline' method needed? Addictive Behaviors 1987 12 225 233 3661275 10.1016/0306-4603(87)90032-3
Goddard E Higgins V Smoking, Drinking and Drug Use among Young Teenagers in 1998: England 1999 1 London: The Stationery Office
Currie C Roberts C Morgan A Smith R Settertobulte W Samdal O Barnekow Rasmussen V Young people's health in context Health Behaviour in School-aged Children (HBSC) study: international report from the 2001/2002 survey 2004 Copenhagen: World Health Organization Regional Office
Parry-Langdon N Bloor M Audrey S Holliday J Process evaluation of health promotion interventions Policy Polit 2003 31 207 216 10.1332/030557303765371690
Pritchard C Trends in Economic Evaluation 1998 London: Office of Health Economics
Drummond MF O'Brien B Stoddart G Torrance G Methods for the Economic Evaluation of Healthcare Programmes 1997 2 Oxford: Oxford University Press
Ellickson PL Hawes JA An assessment of active versus passive methods for obtaining parental consent Eval Rev 1989 13 45 55 11659371
Severson H Biglan A Rationale for the use of passive consent in smoking prevention research: politics, policy, and pragmatics Prev Med 1989 18 267 279 2740296 10.1016/0091-7435(89)90074-1
| 15847695 | PMC1097740 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Apr 22; 5:43 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-43 | oa_comm |
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-201582630810.1186/1471-244X-5-20Research ArticleSubjective face recognition difficulties, aberrant sensibility, sleeping disturbances and aberrant eating habits in families with Asperger syndrome Nieminen-von Wendt Taina [email protected] Juulia E [email protected] Tero [email protected] Susan [email protected]ällman Tiia [email protected]ärvelä Irma [email protected] Wendt Lennart [email protected] Department of Child Neurology, Hospital for Children and Adolescents, Helsinki University Central Hospital, Helsinki, Finland2 Department of Child Psychiatry, Hospital for Children and Adolescents, Helsinki University Central Hospital, Helsinki, Finland3 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland4 Department of Medical Genetics, University of Helsinki, Helsinki, Finland5 Laboratory of Molecular Genetics, Helsinki University Hospital (Laboratory Services), Helsinki, Finland2005 12 4 2005 5 20 20 21 9 2004 12 4 2005 Copyright © 2005 Wendt et al; licensee BioMed Central Ltd.2005Wendt 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 present study was undertaken in order to determine whether a set of clinical features, which are not included in the DSM-IV or ICD-10 for Asperger Syndrome (AS), are associated with AS in particular or whether they are merely a familial trait that is not related to the diagnosis.
Methods
Ten large families, a total of 138 persons, of whom 58 individuals fulfilled the diagnostic criteria for AS and another 56 did not to fulfill these criteria, were studied using a structured interview focusing on the possible presence of face recognition difficulties, aberrant sensibility and eating habits and sleeping disturbances.
Results
The prevalence for face recognition difficulties was 46.6% in individuals with AS compared with 10.7% in the control group. The corresponding figures for subjectively reported presence of aberrant sensibilities were 91.4% and 46.6%, for sleeping disturbances 48.3% and 23.2% and for aberrant eating habits 60.3% and 14.3%, respectively.
Conclusion
An aberrant processing of sensory information appears to be a common feature in AS. The impact of these and other clinical features that are not incorporated in the ICD-10 and DSM-IV on our understanding of AS may hitherto have been underestimated. These associated clinical traits may well be reflected by the behavioural characteristics of these individuals.
==== Body
Background
Asperger Syndrome (AS), which belongs to the autism spectrum disorders, is a behavioural disorder defined by problems in social interaction, circumscribed unusual interests, repetitive routines and rituals, speech peculiarities and non-verbal communication problems usually without language delay. The overall cognitive level is generally within normal limits and motor clumsiness may also be seen [1-5].
The diagnosis of AS is a clinical one, based on a set of symptoms and signs, emphasizing development and behaviour in childhood, despite the fact that the diagnosis can also be made in adulthood. The definite diagnosis is based on the ICD-10 [3] and DSM-IV [4] (APA, 1994) classifications, which comprise the current official diagnostic criteria for AS.
In addition to the official diagnostic criteria for autism spectrum disorders listed in the ICD-10 [3], and DSM-IV [4], there appear to be some additional clinical features, which may be over-represented in these disorders. Difficulties of face recognition has been reported in a number of individuals with AS [6-9]. It has been suggested that face recognition problems may play a role as an essential feature in AS [6,9] and therefore could play a causal role for social dysfunction [6]. Njiokiktjien et al. [7] described the co-existence of face recognition difficulties and disordered recognition of emotions in individuals with AS thereby suggesting an impact of this feature on the social shortcomings characteristic for this disorder.
Barton et. al. [10] showed that there is heterogeneity in social developmental disorders (SDD), like AS and SEPD (social-emotional processing disorder) what it comes to face perception and recognition. They showed that a subgroup of persons with SDD performed normally on face perception and recognition, while those who did not could be divided into two subgroups. The other subgroup had poor perception of facial structure but relatively preserved imagery, while the other subgroup had better perception than imagery. The authors judged that the social disturbance in SDD does not invariably lead to impaired face recognition; meaning that the heterogeneity makes it less likely that poor social function was the cause of impaired face recognition [10].
The official diagnostic criteria (ICD-10, DSM-IV) [3,4] do not include sensory difficulties as a criterion for AS. However, already Hans Asperger [11] recognized sensory aberrations in children with AS. The children he described displayed a range of hyposensitivities and hypersensitivities to taste, tactile and auditory stimuli [11]. More recently, consideration has been given to the possibility that sensory processing is an underlying feature of AS, also reflected in their ability to conduct daily life successfully. Some evidence even suggest that difficulties in socialization can be linked to poor sensory processing as e.g Myles et al. [12] in their study concluded that temper tantrums in AS is due to an overload of sensory stimuli.
The sensory processing issue in AS has more extensively been addressed by Dunn et al. [13,14] using The Sensory Profile test which consists of a 125-item questionnaire, that describes responses to sensory events in daily life. The children with AS studied by them were normal on modulation on visual input affecting emotional responses. This implies that the children with AS can process visual-perceptual information in context, and perhaps can orient to social situations using their visual system in a similar way to other children, even though the other behavioural responses might be different [12-14]. Furthermore, it was shown that persons with AS were hyperresponsive, (detecting texture, taste and temperature differences), and that they had difficulty quieting down for sleep because they would have high detection of stimuli to grab their attention [13]. On the basis of these studies it seems that children with AS have difficulties in sensory modulation; and that the stereotypic behaviours serve as an organizing function [13]. Dunn et al. [13,14] suggested that sensory issues are important in AS and that including sensory processing status in diagnostic criteria might be warranted.
Concerning auditory processing Myles et al. [12] and Dunn et al. [13,14], showed that poor auditory processing in children with AS is associated with attention challenges. This was interpreted by the authors as showing children with AS having difficulties in capturing the auditory information in social conversations. The results of the study by Jansson-Verkasalo and co-workers [15] is in line with this as they demonstrated that children with AS exhibited difficulties in discriminating tone pitch. The findings suggest according to Jansson-Verkasalo et al. [15] that auditory sensory processing is deficient in children with AS.
Abnormal sleep patterns have been reported in up to 65 percent of children with autism spectrum disorders [16]. According to the study by Paavonen et al. [17] behavioural, attention and social problems in children with AS are aggravated when a sleep disorder is present. Interestingly, an open clinical trial on 15 children with AS aged 6–17 years showed that melatonin improved the sleep patterns in all individuals with AS, and half of them displayed excellent responses [17]. In a hospital-based series by Tani et al. [18] on twenty adults with AS, the sleep questionnaire revealed insomnia in 90 percent and the sleep diary in 75 percent. According to Tani et al. [18] the neuropsychiatric deficits inherent in AS predispose both to insomnia, anxiety and mood disorders. These two studies suggest that sleep difficulties inherent in AS are preserved from childhood to adulthood. This data might be useful diagnostically, and hence determining their prevalence might be of interest, even if they are not related to sensory processing.
According to Gillberg and Råstam [19] and Gillberg and Billstedt [20], special eating habits, such as particular food refusal, food fads, pica, hoarding, overeating and various degrees of anorectic behaviour, including complete food refusal and compulsive ordering of food on the plate, are involved in autism. In line with this are the results of the study by Sobanski et al. [21] who in a hospital-based series of children with AS (age-span 7–l9 years), demonstrated reduced weight gain in all males and anorectic-like eating behaviour in four out of 36. A similar result was reached by Bölte et. al. [22] in a hospital-based series of 103 individuals (74 males and 29 females; age-range 10.1–39.9 years), of whom 71 had been diagnosed with autism and 32 with AS. It was shown that 28 percent of the males with AS had a body mass index (BMI) in the fifth percentile or below. However, none fulfilled the ICD-10 criteria for anorexia nervosa. The only factor that significantly influenced BMI was hyperactivity and it was concluded that the co-occurrence of autism spectrum disorders and anorexia nervosa was due to chance [22].
As there is direct and indirect evidence that a number of clinical features related to sensibility, not used as official criteria in the DSM-IV [4] or ICD-10 [3] for AS occur fairly frequently in this syndrome, the present study was undertaken in order to determine whether these are associated with AS in particular or whether they are merely a familial trait that is not related to the diagnosis.
Methods
Subjects
The present study is part of larger study of autism spectrum disorders. The total clinical series comprises 250 families (1360 individuals). The Asperger family series in this large series was recruited through the Hospital for Children and Adolescents, Department of Child Neurology, Helsinki University Central Hospital (HUCH), and the Helsinki Asperger Center (HAC), Medical Center Dextra, Helsinki, Finland. The Department of Child Neurology at the Helsinki University Central Hospital primarily serves the catchment area of the Helsinki University Central Hospital (population 1.4 million), but it also serves the entire country as a tertiary referral unit. The HAC is a private clinic, freely available to anyone and serving the whole of Finland (population 5.2 million).
The Asperger clinical series consisted of 29 families, in which AS was present in at least two generations. From these families, 10 large families, in which the diagnosis was present in at least two generations exclusively on either the maternal or the paternal side and for which molecular genetic data had also been obtained, were recruited for the present analysis after informed consent had been obtained [26]. Large kindred were chosen, as including only small families would have compromised the statistical analyses. These ten families consisted of 138 persons, of whom 58 individuals were found to have a confirmed diagnosis of AS using the diagnostic procedure outlined above (range 4.5–78.2 years; mean 32.8 years; SEM 2.8; CI 27.1–38.5). Another 56 (range 3.5–77.9 years; mean 33.8 years; SEM 2.6; CI 28.2–38.5) also went through the diagnostic procedure and were found not to fulfill the criteria for AS. The remaining 24 family members could not be included in the study, as four lived abroad and the others did not agree to participate. This means that 83% of the family members participated.
Protocol
The individuals investigated in this study were diagnosed as representing AS if they fulfilled both the diagnostic criteria of the ICD-10 [3] and the DSM-IV [4]. For this purpose the following instruments were used:
The Asperger Syndrome Diagnostic Interview (ASDI) [23] is an investigator-based interview consisting of 20 items, all of which must be covered in detail. The interviews were conducted in accordance with the principles of this instrument.
The Autism Diagnostic Observation Schedule (ADOS) is a semi-structured, standardized assessment of communication, social interaction and play or imaginative use of materials for individuals who have been referred because of possible autism spectrum disorders [24].
The Autism Diagnostic Interview – Revised (ADI-R) [25] is an investigator-based interview containing structured coding for each behavioural item. It consists of six sections dealing with background orientation, developmental history and earlier and current behaviour.
Assessment of clinical features related to face processing, sensibility, sleeping and eating not used as official criteria in the DSM-IV or ICD-10
For this study, a particular set of questions the Asperger syndrome related additional features questionnaire (ASFQ) (Appendix 1) was developed in order to determine whether there were any defined data showing abnormalities in the domains of face recognition difficulties, presence of aberrant sensibilities, sleeping disturbances, and aberrant eating habits. The questions were presented during an interview and the answers were recorded by the interviewer as yes or no. The interviewers were two neuropediatricians and two nurses trained in autism spectrum disorders, all of them had a long experience in autism spectrum disorders. In order to increase the validity pre-constructed clarifying examples and additional instructions for each of the questions were used.
In addition to this available medical charts and other relevant documentation were scrutinized for previously set diagnoses, special examinations, and developmental history.
Statistical methods
First, researches conducted a frequency analysis, and secondly Mantel-Haentzel pooled risk estimates were calculated. Finally, logistic regression models using the generalized estimating equations procedure were developed (random effects model linear model). This method is similar to ordinary logistic regression, but it takes account of the possible dependences between the family members.
Results
The prevalence of face recognition difficulties emerged as being over-represented in the family members with AS in this study, as the percentage of occurrence was 46.6 percent compared with 10.7 percent in the family members without AS (Table 1). The prevalence of self-reported aberrant presence of sensibilities was significantly higher (OR, Mantel-Haentzel 38.27 (95% CI 6.29–232.77) in the individuals with AS than in their non-AS family members (Table 2). Among the subcomponents of aberrant sensibility; touch, light, sounds and smell emerged as being significantly altered in the group of family members with AS (Table 1). Other features which were significantly more frequent in family members who fulfilled the criteria for AS were sleeping disturbances and aberrant eating habits (Tables 1, 2).
Table 1 Mantel-Haentzel (MH) pooled risk estimates (risk of the symptom in Asperger individuals)
Overall prevalence % (n) Prevalence in controls % (n) Prevalence in cases % (n) ORMH (95% CI)
Face recognition difficulties 28.9 (33) 10.7 (6) 46.6 (27) 20.46 (4.51–92.88)
Presence of altered sensibilities 69.3 (79) 46.4 (26) 91.4 (53) 38.27 (6.29–232.77)
- of touch 31.6 (36) 8.9 (5) 53.4 (31) 12.16 (3.37–43.97)
- of light 25.4 (29) 7.1 (4) 43.1 (25) 7.57 (2.13–26.87)
- of sounds 33.3 (38) 16.1 (9) 50.0 (29) 5.59 (2.08–15.01)
- of smell 34.2 (39) 23.2 (13) 44.8 (26) 2.56 (1.01–6.47)
- of pain 11.4 (13) 7.1 (4) 15.5 (9) 2.54 (0.65–9.96)
Sleeping disturbances 36.0 (41) 23.2 (13) 48.3 (28) 3.11 (1.24–7.78)
Aberrant eating habits 37.7 (43) 14.3 (8) 60.3 (35) 8.41 (2.80–25.27)
OR = Odds Ratio
Table 2 Logistic regression models (risk of the symptom in Asperger individuals)
Model coefficient/ parameter estimate SE p-value
Face recognition difficulties 2,26 0,56 <0,001
Presence of altered sensibilities 2,70 0,58 <0,001
- of touch 2,42 0,54 <0,001
- of light 2,03 0,56 <0,001
- of sounds 1,61 0,44 <0,001
- of smell 1,00 0,42 0,017
Sleeping disturbances 1,11 0,41 0,007
Aberrant eating habits 2,24 0,48 <0,007
The results of the random effect models are presented in Table 3. They further confirm that face recognition difficulties (p < 0.001), aberrant sensibilities (p < 0.001), and aberrant eating habits (p = 0.007) are highly typical of AS. The risk of sleep problems (p < 0.007) is also slightly higher in persons with AS.
The sensitivity for the presence of aberrant sensibility in the family members with AS was 0.91, whereas the specificity was 0.54. The corresponding figures for face processing difficulties were 0.47 and 0.89, for sleeping disturbances 0.48 and 0.77 and for aberrant eating habits 0.60 and 0.86. The aberrant sensibility items had a higher sensitivity, they also had lower specificity, and that actually all the items had a very similar data value. Hence the differences between the items are due to a criterion shift, with all subjects having a greater bias to answer yes to the aberrant sensibility items.
Discussion
The diagnostic criteria for autistic disorders used in the ICD-10 and DSM-IV manuals apparently represent the experts' generally accepted views of the clinical delineation of AS at the time the latest revision was made. One evident drawback of these sets of diagnostic criteria is that the distribution of symptoms and signs underlying the criteria in an average population has not been satisfactorily outlined. The same also applies to many of the sets of screening questionnaires and other instruments used in the diagnostic process. It can even be questioned whether currently used instruments actually cover the core features of AS. Using several such instruments and including only individuals who fulfilled both ICD-10 and DSM-I0 criteria for AS, as was the case in this study should improve the detection rate of the features typical for AS.
The data was collected by a few experienced investigators using a standardized procedure that is expected to increase the reliability of the results. The same sets of diagnostic instruments were used for each age group to facilitate the comparison of the results. The chosen set of diagnostic instruments was considered to have several advantages including ADI-R [25] that covers the entire lifetime. Using different set of diagnostic instruments for each age group would have hampered the comparability of the results. The only new method used in this study was the Asperger syndrome related additional features questionnaire (ASFQ)(Appendix 1), which was designed specifically for the purpose of detecting additional clinical feature much in the same way as in ICD-10 and DSM-IV. This instrument was not validated, but in a pilot-like setting to find out whether family members with or without AS differed from each other it was considered as a suitable method. Another drawback of the ASFQ method is that the results are based on self-reported perceptions that are not objective. The relevance of ASFQ needs to be studied in a control group of average population before its applicability as a diagnostic instrument for AS can be assessed.
The families chosen for this analysis may not necessarily completely reflect the average AS population as we for analytical reasons focused on large pedigrees with several AS individuals. Nor is the series unequivocally a population-based one, as the families were recruited from originally clinical series in which the index case had been referred for a diagnostic evaluation. By using non-AS individuals from the same families as a comparison group, we think that the effects of the family social environment are eliminated at least to some extent. The methodological weakness need to be taken into account, but the results seem to be fairly straightforward, as face recognition difficulties and aberrant sensibility emerge as being fairly typical of AS.
The basic aim of the present research project on AS was to elucidate the pathogenesis, especially with respect to the molecular genetic issues [26]. As AS is a complex disorder it seems likely that the discovery of endophenotypes could be of help for the identification of both putative susceptibility genes and specific neuroimaging features. In line with this, it seems attractive to speculate that our findings indicate the existence of a general sensory processing abnormality in AS as these individuals frequently are sensitive to noise, bright lights, strong smells and touch. Thus, these findings may be useful as well for diagnostic purpose as for understanding of the pathogenesis. Our findings are supported by Dunn et al. [13,14] who considered that the contributions of sensory processing status in diagnostic criteria might be warranted. Another possibility is that individuals with AS can or should be sub-grouped on the basis of their pattern of sensory processing status, sleeping and eating problems.
In the case of face recognition difficulties, there are no data on its occurrence in the average population, but it is not unlikely that the non-AS family members may also represent an increased occurrence. In our study, 50 % of the persons with AS had difficulties in recognition of faces. This is in line with the study by Barton et al. [10], which showed that 2/3 of the persons with AS had heterogeneity in the perceptual or imagery processing of faces. Thus, these two lines of evidence support the assumption that individuals with AS have face recognition difficulties. Face recognition is one of the most complex visual tasks performed by the human brain [10]. According to the literature, persons with autism spectrum disorders process faces as if they were objects and appear to rely more on identifying individual pieces of the face rather than the overall configuration [27-29]. This is significant data, because recognition of faces is an important part of interpersonal interaction and successful functioning within a social group. These findings and the result of our study makes it attractive to speculate that the deficits in face perception may contribute to the impaired social communication in persons with AS. Therefore, including face recognition difficulties in the clinical criteria for AS may also be warranted, as it might represent a core dysfunction of AS. This is in line with the study by Barton et al. [10], which concludes the same point, that heterogeneity may be important to take into account in genetic investigations.
Aberrant sensory perception was a major finding in our study, especially in the domains of sound and touch. This is in agreement with the findings made by Dunn et al. [13,14] and Myles at al. [12], as it seems that the auditory and touch processing are the two most important aberrant sensory domains. The nature of aberrant auditory sensory processing in AS and autism is still open. According to Baranek [30], altered hyperreactions in auditory processing occur in autism in up to 100% of the cases. The auditory sensory processing was studied using mitch-match-negativity recording technique by Jansson-Verkasalo et al. [15] and clinically using the Sensory Profile by Dunn et al. [13,14]. The former study indicated that auditory sensory processing was deficient in children with AS. Dunn et al. [13,14] showed that children with AS had difficulties in auditory processing, but that these had more difficulty with auditory attention than children with autism. On the other hand, Čeponienė et al. [31], demonstrated auditory orienting deficits in autism, which could not be explained by sensory deficits and concluded that the orienting deficit in autism might be speech-sound specific. Auditory processing impairments in AS and autism deserve special attention, as they may constitute a part of the pathogenesis of the behavioral problems. On a general level, our observations of a change in sensory integration in the AS group are well in line with the results of corresponding studies of autism spectrum disorders [12-14,20,30,31].
Insomnia predisposes to emotional distress, daytime fatigue and a loss of productivity [18]. A careful assessment of sleep quality should therefore be integrated in the diagnostic interview for AS, as, according to our study, insomnia is a frequent finding in this group. The studies by Paavonen et al. [17] and Tani et al. [18] demonstrate an increased prevalence of disorders in initiating and maintaining sleep in both children and adults with AS, suggesting that sleep difficulties inherent to AS are preserved from childhood into adulthood. It is not known why persons with AS have insomnia or sleeping problems, but the reports by Godbout et al. [32,33] suggest that AS subjects may present the same difficulties in initiating and maintaining sleep as those previously described in autism. It has also been hypothesized that individuals with AS have a defective sleep control system which may be associated with the clinical picture of AS [33]. From a clinical point of view, sleep disturbances are typical among people with AS and should be taken into account when a person is referred for a diagnosis, especially as the problem can be alleviated by melatonin [17].
Our study recognized aberrant eating habits in 60 per cent of individuals with AS, which is much higher than the figure of approximately one per cent obtained by Sobanski et al. [21]. Eating problems [21,34,35] and also anorexia nervosa have been reported to be over-represented in AS [20,34-36]. As, there according to Gillberg et al. [19], is a 12 per cent co-morbidity of anorexia nervosa and AS and there frequently is a picky eating behavior it could be speculated that this is due to hypersensitivity, as low thresholds for taste in AS was shown by Dunn et al. [13,14]. This could be part of the explanation for a higher incidence rate of eating disorders among persons with AS.
Conclusion
Our results speak in favour of the fact that aberrant sensory processing is an extensive phenomenon in AS, as these individuals frequently are hypersensitive to noise, bright lights, strong smells and touch. Our results are preliminary and need to be substantiated also using objective assessment methods before they can be seen as definite. Provided that the present findings can be substantiated in this way it might be possible to use the aberrant sensory processing also as a diagnostic criteria. It seems obvious that such aberrant sensory processing is a key feature of autism spectrum disorders, and our results will therefore probably not help in the differentiation between autism and AS.
Authors' contributions
TN-vW was the main principal investigator collecting the data and the series. She was also principally responsible for preparing the manuscript. Together with author LvW and TY-O, she conceived the idea of this study
EJP participated in the design of the study and performed the statistical analysis
TY-0 was co-conceiver of the idea of this study and also in writing the manuscript
SS together with TK carried out all the ADI-R and/or ADOS investigations
TK together with SS carried out all the ADI-R and/or ADOS investigations
IJ participated in designing the study and in the writing of the manuscript
LvW helped to collect the data and the series. LvW also participated in designing the study and in the writing of the manuscript
Note
APPENDIX 1. The complementary questions regarding symptoms and signs not included in the ICD-10 and/or DSM-IV.
Asperger syndrome related additional features questionnaire (ASFQ)
The following domains were included
1. Presence of face recognition difficulties
Do you have difficulty recognizing familiar persons from their faces?
Standardized clarifying questions were used, such as:
- If the clothing, color of the hair of an individual familiar to you has changed or you meet unexpectedly in an unfamiliar place, would you be able to recognize him/her?
- Do you generally recognize people on the basis of their style of walking, clothing, voice-/ speech rather than by their face?
2. Presence of aberrant sensibilities
The aim was to find out whether the individual considered him/herself and-/or was considered to represent hyper sensibility in the following modes of sensation:
- touch (light or strong)
- pain
- light
- external temperature
- sound
- smell
Fixed sets of clarifying questions were used.
3. Sleeping/diurnal rhythm disturbances:
The clarifying questions dealt with: getting to sleep, awakenings during the night, amount of sleep, abnormality of diurnal rhythm, use of sleeping pills, restless legs, impact of possible difficulties related to diurnal rhythm of every day life
4. Aberrant eating habits
The fixed clarifying questions comprised the following items:
- unusual-/odd food preferences
- unusual routines regarding arranging food on the plate and eating
- presence of aversion due to the smell of the food
- ongoing or past anorectic behavior
5. Behavior
The clarifying questions pinpointed possible presence of selection difficulties in ordinary every day situations (ice-cream a or b).
It was also established whether making up one's mind in such situation lead to avoidance of such situations. It was also made clear whether the person in such situations used clearly excessive time for making the decisions or whether these caused anxiety?
Another set of questions aimed at clarifying whether the individual him/herself recognized traits of hyperactivity and/or attention difficulties or whether other people had made him/her aware of such traits.
6. External appearance
Possible presence of unusual physical features with respect to stature, flexibility of joints and minor stigmata were recorded.
In both genders the questions aimed at finding out whether the individual frequently was wrongly perceived with respect to chronological age. Also the interviewer's perception of a possible youngish appearance was recorded.
In males questions also were focused on the age for starting regular shaving and the maximum interval between shavings which was compatible with socially acceptable tidiness of the face. Answers reporting such shaving less frequently than twice a week were considered being out of average range for adult Finnish men.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by grants from the Academy of Finland, the Foundation for Pediatric Research in Finland, Helsinki University Hospital Research Funding, the Medical Society of Finland, the Päivikki and Sakari Sohlberg Foundation and the Perklen Foundation.
==== Refs
Wing L Asperger's syndrome: a clinical account Psychol Med 1981 11 115 130 7208735
Gillberg C Gillberg C Disorders of empathy: autism and autism spectrum disorders (including childhood onset schizophrenia) Clinical Child Neuropsychology 1995 1 1 Cambridge: Cambridge University Press 54 111
World Health Organization International classification of disease (10th ed, chap 5) Mental and behavioral disorders 1993 Diagnostic criteria for research. Geneva: Author
American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders (4th Ed) (DSM-IV) 1994 Washington DC: American Psychiatric Association
Szatmari P The classification of autism, Asperger's syndrome, and pervasive developmental disorder Can J Psychiatry 2000 45 731 738 11086556
Kracke I Developmental prosopagnosia in Asperger syndrome: presentation and discussion of an individual case Dev Med Child Neurol 1994 36 873 886 7926319
Njiokiktjien C Verschoor A de Sonneville L Huyser C Op het Veld V Toorenaar N Disordered recognition of facial identity and emotions in three Asperger type autists Eur Child Adolesc Psychiatry 2001 10 79 90 11315539 10.1007/s007870170050
Duchaine BC Nieminen-von Wendt T New J Kulomäki T Dissociations of visual recognition in a genetic prosopagnosic: Evidence for separate developmental processes Neurocase 2003 9 380 389 14972753 10.1076/neur.9.5.380.16556
Pietz J Ebinger F Rating D Prosopagnosia in a preschool child with Asperger syndrome Dev Med Child Neurol 2003 45 55 57 12549756
Barton JJ Cherkasova MV Hefter R Cox TA O'Connor M Manoach DS Are patients with social developmental disorders prosopagnosic? Perceptual heterogeneity in the Asperger and socio-emotional processing disorders Brain 2004 1706 1716 15215211 10.1093/brain/awh194
Asperger H Die "Autistischen Psychopathen" im Kindesalter Arch Psychiatr Nervenkr 1944 117 76 136
Myles B Hagiwara T Dunn W Rinner L Reese M Sensory issues in children with Asperger syndrome and autism ETDD 2004 39 283 290
Dunn W Myles BM Orr S Sensory processing issues associated with Asperger syndrome: a preliminary investigation Am J Occup Ther 2002 56 97 102 11833406
Dunn W Saiter J Rinner L Asperger syndrome and sensory processing: a conceptual model and guidance for intervention planning Focus on autism and other developmental disabilites 2002 17 172 185
Jansson-Verkasalo E Ceponiene R Kielinen Suominen K Jäntti Linna SL Moilanen I Näätänen R Deficient auditory processing in children with Asperger Syndrome, as indexed by event-related potentials Neurosci Lett 2003 338 197 200 12581830 10.1016/S0304-3940(02)01405-2
Taira M Takase M Sasaki H Sleep disorder in children with autism Psychiatry Clin Neurosci 1998 52 182 183 9628139
Paavonen EJ Nieminen-von Wendt T Vanhala R Aronen E von-Wendt L Effectiveness of melatonin in the treatment of sleep disturbances in children with Asperger Syndrome J Child Adol Psychop 2003 13 83 95 10.1089/104454603321666225
Tani P Lindberg N Nieminen-von Wendt T von Wendt L Alanko L Appelberg B Porkka-Heiskanen T Insomnia is a frequent finding in adults with Asperger syndrome BMC Psychiatry 2003 3 12 28 14563215 10.1186/1471-244X-3-12
Gillberg C Råstam M Do some cases of anorexia nervosa reflect underlying autistic-like condition? Behav Neurol 1992 5 27 32
Gillberg C Billstedt E Autism and Asperger syndrome: coexistence with other clinical disorders Acta Psychiatr Scand 2000 102 321 330 11098802 10.1034/j.1600-0447.2000.102005321.x
Sobanski E Marcus A Hennighausen K Hebebrand J Schmidt MH Further evidence for a low body weight in male children and adolescents with Asperger's disorder Eur Child Adolesc Psychiatry 1999 8 312 314 10654125 10.1007/s007870050106
Bölte S Ozkara N Poustka F Autism spectrum disorders and low body weight: is there really a systematic association? Int J Eating Disor 2002 31 349 351 10.1002/eat.10015
Gillberg C Gillberg C Råstam M Wentz E The Asperger Syndrome (and high-functioning autism) Diagnostic Interview (ASDI): a preliminary study of a new structured clinical interview Autism 2001 5 57 66 11708390 10.1177/1362361301005001006
Lord C Risi S Lambrecht L Cook EH JrLeventhal BL DiLavore PC Pickles A Rutter M The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits associated with the spectrum of autism J Aut Dev Disord 2002 30 205 223 10.1023/A:1005592401947
Lord C Rutter M Le Couteur A Autism Diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders J Autism Dev Disord 1994 24 659 685 7814313
Ylisaukko-oja T Nieminen-von Wendt T Kempas E Sarenius S Varilo T von Wendt L Peltone I Järvelä I A Genome-wide Scan for Loci of Asperger Syndrome Mol Psychiatry 2004 9 161 168 14966474 10.1038/sj.mp.4001385
Klin A Schultz R Cohen DJ Baron-Cohen S, Tager-Flusberg H, Cohen JD Theory of mind in action: developmental perspectives on social neuroscience Understanding Other Minds Perspectives from Developmental Cognitive Neuroscience 2000 2 Oxford University Press: Oxford 357 388
Schultz R Romanski L Tsatsanis K(a) Klin A, Volkmar FR, Sparrow SS Neurofunctional models of autistic disorders and Asperger syndrome. Clues from neuroimaging Asperger syndrome 2000 1 New York: The Guilford Press 172 209
Schultz R Gauthier I Klin A Fulbright R Anderson A Volkmar F Skudlarski P Lacadie C Cohen D Gore J(b) Abnormal ventral temporal cortical activity during face discrimination among individuals with autism and Asperger syndrome Arch Gen Psychiatry 2000 57 331 340 10768694 10.1001/archpsyc.57.4.331
Baranek GT Efficacy of sensory and motor interventions for children with autism J Aut Dev Disord 2002 32 397 422 10.1023/A:1020541906063
Čeponienė R Lepistö T Shestakova A Vanhala R Alku P Näätänen R Yaguchi K Speech-sound-selective auditory impairment in children with autism: They can perceive but do not attend Proc Natl Acad Sci U S A 2003 100 5567 5572 12702776 10.1073/pnas.0835631100
Godbout R Bergeron C Stip E Mottron A laboratory study of sleep and dreaming in a case of Asperger's syndrome Dreaming 1998 8 75 88 10.1023/B:DREM.0000005898.95212.58
Godbout R Bergeron C Limoges E Stip E Mottron L A laboratory study of sleep in Asperger's syndrome NeuroReport 2000 11 127 130 10683843
Råstam M Gillberg C Gillberg IC Johansson M Alexithymia in anorexia nervosa: a controlled study using the 20-item Toronto Alexithymia Scale Acta Psychiatr Scand 1997 95 385 388 9197902
Ahearn WH Castine T Nault K Green G An assessment of food acceptance in children with autism or pervasive developmental disorder-not otherwise specified J Autism Dev Disord 2001 31 505 511 11794415 10.1023/A:1012221026124
Hebebrand J Henninghausen K Nau S Himmelmann GW Schulz E Schafer H Remschmidt H Low body weight in male children and adolescents with schizoid personality disorder or Asperger's disorder Acta Psychiatr Scand 1997 96 64 67 9259226
| 15826308 | PMC1097741 | CC BY | 2021-01-04 16:33:02 | no | BMC Psychiatry. 2005 Apr 12; 5:20 | utf-8 | BMC Psychiatry | 2,005 | 10.1186/1471-244X-5-20 | oa_comm |
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-201582630810.1186/1471-244X-5-20Research ArticleSubjective face recognition difficulties, aberrant sensibility, sleeping disturbances and aberrant eating habits in families with Asperger syndrome Nieminen-von Wendt Taina [email protected] Juulia E [email protected] Tero [email protected] Susan [email protected]ällman Tiia [email protected]ärvelä Irma [email protected] Wendt Lennart [email protected] Department of Child Neurology, Hospital for Children and Adolescents, Helsinki University Central Hospital, Helsinki, Finland2 Department of Child Psychiatry, Hospital for Children and Adolescents, Helsinki University Central Hospital, Helsinki, Finland3 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland4 Department of Medical Genetics, University of Helsinki, Helsinki, Finland5 Laboratory of Molecular Genetics, Helsinki University Hospital (Laboratory Services), Helsinki, Finland2005 12 4 2005 5 20 20 21 9 2004 12 4 2005 Copyright © 2005 Wendt et al; licensee BioMed Central Ltd.2005Wendt 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 present study was undertaken in order to determine whether a set of clinical features, which are not included in the DSM-IV or ICD-10 for Asperger Syndrome (AS), are associated with AS in particular or whether they are merely a familial trait that is not related to the diagnosis.
Methods
Ten large families, a total of 138 persons, of whom 58 individuals fulfilled the diagnostic criteria for AS and another 56 did not to fulfill these criteria, were studied using a structured interview focusing on the possible presence of face recognition difficulties, aberrant sensibility and eating habits and sleeping disturbances.
Results
The prevalence for face recognition difficulties was 46.6% in individuals with AS compared with 10.7% in the control group. The corresponding figures for subjectively reported presence of aberrant sensibilities were 91.4% and 46.6%, for sleeping disturbances 48.3% and 23.2% and for aberrant eating habits 60.3% and 14.3%, respectively.
Conclusion
An aberrant processing of sensory information appears to be a common feature in AS. The impact of these and other clinical features that are not incorporated in the ICD-10 and DSM-IV on our understanding of AS may hitherto have been underestimated. These associated clinical traits may well be reflected by the behavioural characteristics of these individuals.
==== Body
Background
Asperger Syndrome (AS), which belongs to the autism spectrum disorders, is a behavioural disorder defined by problems in social interaction, circumscribed unusual interests, repetitive routines and rituals, speech peculiarities and non-verbal communication problems usually without language delay. The overall cognitive level is generally within normal limits and motor clumsiness may also be seen [1-5].
The diagnosis of AS is a clinical one, based on a set of symptoms and signs, emphasizing development and behaviour in childhood, despite the fact that the diagnosis can also be made in adulthood. The definite diagnosis is based on the ICD-10 [3] and DSM-IV [4] (APA, 1994) classifications, which comprise the current official diagnostic criteria for AS.
In addition to the official diagnostic criteria for autism spectrum disorders listed in the ICD-10 [3], and DSM-IV [4], there appear to be some additional clinical features, which may be over-represented in these disorders. Difficulties of face recognition has been reported in a number of individuals with AS [6-9]. It has been suggested that face recognition problems may play a role as an essential feature in AS [6,9] and therefore could play a causal role for social dysfunction [6]. Njiokiktjien et al. [7] described the co-existence of face recognition difficulties and disordered recognition of emotions in individuals with AS thereby suggesting an impact of this feature on the social shortcomings characteristic for this disorder.
Barton et. al. [10] showed that there is heterogeneity in social developmental disorders (SDD), like AS and SEPD (social-emotional processing disorder) what it comes to face perception and recognition. They showed that a subgroup of persons with SDD performed normally on face perception and recognition, while those who did not could be divided into two subgroups. The other subgroup had poor perception of facial structure but relatively preserved imagery, while the other subgroup had better perception than imagery. The authors judged that the social disturbance in SDD does not invariably lead to impaired face recognition; meaning that the heterogeneity makes it less likely that poor social function was the cause of impaired face recognition [10].
The official diagnostic criteria (ICD-10, DSM-IV) [3,4] do not include sensory difficulties as a criterion for AS. However, already Hans Asperger [11] recognized sensory aberrations in children with AS. The children he described displayed a range of hyposensitivities and hypersensitivities to taste, tactile and auditory stimuli [11]. More recently, consideration has been given to the possibility that sensory processing is an underlying feature of AS, also reflected in their ability to conduct daily life successfully. Some evidence even suggest that difficulties in socialization can be linked to poor sensory processing as e.g Myles et al. [12] in their study concluded that temper tantrums in AS is due to an overload of sensory stimuli.
The sensory processing issue in AS has more extensively been addressed by Dunn et al. [13,14] using The Sensory Profile test which consists of a 125-item questionnaire, that describes responses to sensory events in daily life. The children with AS studied by them were normal on modulation on visual input affecting emotional responses. This implies that the children with AS can process visual-perceptual information in context, and perhaps can orient to social situations using their visual system in a similar way to other children, even though the other behavioural responses might be different [12-14]. Furthermore, it was shown that persons with AS were hyperresponsive, (detecting texture, taste and temperature differences), and that they had difficulty quieting down for sleep because they would have high detection of stimuli to grab their attention [13]. On the basis of these studies it seems that children with AS have difficulties in sensory modulation; and that the stereotypic behaviours serve as an organizing function [13]. Dunn et al. [13,14] suggested that sensory issues are important in AS and that including sensory processing status in diagnostic criteria might be warranted.
Concerning auditory processing Myles et al. [12] and Dunn et al. [13,14], showed that poor auditory processing in children with AS is associated with attention challenges. This was interpreted by the authors as showing children with AS having difficulties in capturing the auditory information in social conversations. The results of the study by Jansson-Verkasalo and co-workers [15] is in line with this as they demonstrated that children with AS exhibited difficulties in discriminating tone pitch. The findings suggest according to Jansson-Verkasalo et al. [15] that auditory sensory processing is deficient in children with AS.
Abnormal sleep patterns have been reported in up to 65 percent of children with autism spectrum disorders [16]. According to the study by Paavonen et al. [17] behavioural, attention and social problems in children with AS are aggravated when a sleep disorder is present. Interestingly, an open clinical trial on 15 children with AS aged 6–17 years showed that melatonin improved the sleep patterns in all individuals with AS, and half of them displayed excellent responses [17]. In a hospital-based series by Tani et al. [18] on twenty adults with AS, the sleep questionnaire revealed insomnia in 90 percent and the sleep diary in 75 percent. According to Tani et al. [18] the neuropsychiatric deficits inherent in AS predispose both to insomnia, anxiety and mood disorders. These two studies suggest that sleep difficulties inherent in AS are preserved from childhood to adulthood. This data might be useful diagnostically, and hence determining their prevalence might be of interest, even if they are not related to sensory processing.
According to Gillberg and Råstam [19] and Gillberg and Billstedt [20], special eating habits, such as particular food refusal, food fads, pica, hoarding, overeating and various degrees of anorectic behaviour, including complete food refusal and compulsive ordering of food on the plate, are involved in autism. In line with this are the results of the study by Sobanski et al. [21] who in a hospital-based series of children with AS (age-span 7–l9 years), demonstrated reduced weight gain in all males and anorectic-like eating behaviour in four out of 36. A similar result was reached by Bölte et. al. [22] in a hospital-based series of 103 individuals (74 males and 29 females; age-range 10.1–39.9 years), of whom 71 had been diagnosed with autism and 32 with AS. It was shown that 28 percent of the males with AS had a body mass index (BMI) in the fifth percentile or below. However, none fulfilled the ICD-10 criteria for anorexia nervosa. The only factor that significantly influenced BMI was hyperactivity and it was concluded that the co-occurrence of autism spectrum disorders and anorexia nervosa was due to chance [22].
As there is direct and indirect evidence that a number of clinical features related to sensibility, not used as official criteria in the DSM-IV [4] or ICD-10 [3] for AS occur fairly frequently in this syndrome, the present study was undertaken in order to determine whether these are associated with AS in particular or whether they are merely a familial trait that is not related to the diagnosis.
Methods
Subjects
The present study is part of larger study of autism spectrum disorders. The total clinical series comprises 250 families (1360 individuals). The Asperger family series in this large series was recruited through the Hospital for Children and Adolescents, Department of Child Neurology, Helsinki University Central Hospital (HUCH), and the Helsinki Asperger Center (HAC), Medical Center Dextra, Helsinki, Finland. The Department of Child Neurology at the Helsinki University Central Hospital primarily serves the catchment area of the Helsinki University Central Hospital (population 1.4 million), but it also serves the entire country as a tertiary referral unit. The HAC is a private clinic, freely available to anyone and serving the whole of Finland (population 5.2 million).
The Asperger clinical series consisted of 29 families, in which AS was present in at least two generations. From these families, 10 large families, in which the diagnosis was present in at least two generations exclusively on either the maternal or the paternal side and for which molecular genetic data had also been obtained, were recruited for the present analysis after informed consent had been obtained [26]. Large kindred were chosen, as including only small families would have compromised the statistical analyses. These ten families consisted of 138 persons, of whom 58 individuals were found to have a confirmed diagnosis of AS using the diagnostic procedure outlined above (range 4.5–78.2 years; mean 32.8 years; SEM 2.8; CI 27.1–38.5). Another 56 (range 3.5–77.9 years; mean 33.8 years; SEM 2.6; CI 28.2–38.5) also went through the diagnostic procedure and were found not to fulfill the criteria for AS. The remaining 24 family members could not be included in the study, as four lived abroad and the others did not agree to participate. This means that 83% of the family members participated.
Protocol
The individuals investigated in this study were diagnosed as representing AS if they fulfilled both the diagnostic criteria of the ICD-10 [3] and the DSM-IV [4]. For this purpose the following instruments were used:
The Asperger Syndrome Diagnostic Interview (ASDI) [23] is an investigator-based interview consisting of 20 items, all of which must be covered in detail. The interviews were conducted in accordance with the principles of this instrument.
The Autism Diagnostic Observation Schedule (ADOS) is a semi-structured, standardized assessment of communication, social interaction and play or imaginative use of materials for individuals who have been referred because of possible autism spectrum disorders [24].
The Autism Diagnostic Interview – Revised (ADI-R) [25] is an investigator-based interview containing structured coding for each behavioural item. It consists of six sections dealing with background orientation, developmental history and earlier and current behaviour.
Assessment of clinical features related to face processing, sensibility, sleeping and eating not used as official criteria in the DSM-IV or ICD-10
For this study, a particular set of questions the Asperger syndrome related additional features questionnaire (ASFQ) (Appendix 1) was developed in order to determine whether there were any defined data showing abnormalities in the domains of face recognition difficulties, presence of aberrant sensibilities, sleeping disturbances, and aberrant eating habits. The questions were presented during an interview and the answers were recorded by the interviewer as yes or no. The interviewers were two neuropediatricians and two nurses trained in autism spectrum disorders, all of them had a long experience in autism spectrum disorders. In order to increase the validity pre-constructed clarifying examples and additional instructions for each of the questions were used.
In addition to this available medical charts and other relevant documentation were scrutinized for previously set diagnoses, special examinations, and developmental history.
Statistical methods
First, researches conducted a frequency analysis, and secondly Mantel-Haentzel pooled risk estimates were calculated. Finally, logistic regression models using the generalized estimating equations procedure were developed (random effects model linear model). This method is similar to ordinary logistic regression, but it takes account of the possible dependences between the family members.
Results
The prevalence of face recognition difficulties emerged as being over-represented in the family members with AS in this study, as the percentage of occurrence was 46.6 percent compared with 10.7 percent in the family members without AS (Table 1). The prevalence of self-reported aberrant presence of sensibilities was significantly higher (OR, Mantel-Haentzel 38.27 (95% CI 6.29–232.77) in the individuals with AS than in their non-AS family members (Table 2). Among the subcomponents of aberrant sensibility; touch, light, sounds and smell emerged as being significantly altered in the group of family members with AS (Table 1). Other features which were significantly more frequent in family members who fulfilled the criteria for AS were sleeping disturbances and aberrant eating habits (Tables 1, 2).
Table 1 Mantel-Haentzel (MH) pooled risk estimates (risk of the symptom in Asperger individuals)
Overall prevalence % (n) Prevalence in controls % (n) Prevalence in cases % (n) ORMH (95% CI)
Face recognition difficulties 28.9 (33) 10.7 (6) 46.6 (27) 20.46 (4.51–92.88)
Presence of altered sensibilities 69.3 (79) 46.4 (26) 91.4 (53) 38.27 (6.29–232.77)
- of touch 31.6 (36) 8.9 (5) 53.4 (31) 12.16 (3.37–43.97)
- of light 25.4 (29) 7.1 (4) 43.1 (25) 7.57 (2.13–26.87)
- of sounds 33.3 (38) 16.1 (9) 50.0 (29) 5.59 (2.08–15.01)
- of smell 34.2 (39) 23.2 (13) 44.8 (26) 2.56 (1.01–6.47)
- of pain 11.4 (13) 7.1 (4) 15.5 (9) 2.54 (0.65–9.96)
Sleeping disturbances 36.0 (41) 23.2 (13) 48.3 (28) 3.11 (1.24–7.78)
Aberrant eating habits 37.7 (43) 14.3 (8) 60.3 (35) 8.41 (2.80–25.27)
OR = Odds Ratio
Table 2 Logistic regression models (risk of the symptom in Asperger individuals)
Model coefficient/ parameter estimate SE p-value
Face recognition difficulties 2,26 0,56 <0,001
Presence of altered sensibilities 2,70 0,58 <0,001
- of touch 2,42 0,54 <0,001
- of light 2,03 0,56 <0,001
- of sounds 1,61 0,44 <0,001
- of smell 1,00 0,42 0,017
Sleeping disturbances 1,11 0,41 0,007
Aberrant eating habits 2,24 0,48 <0,007
The results of the random effect models are presented in Table 3. They further confirm that face recognition difficulties (p < 0.001), aberrant sensibilities (p < 0.001), and aberrant eating habits (p = 0.007) are highly typical of AS. The risk of sleep problems (p < 0.007) is also slightly higher in persons with AS.
The sensitivity for the presence of aberrant sensibility in the family members with AS was 0.91, whereas the specificity was 0.54. The corresponding figures for face processing difficulties were 0.47 and 0.89, for sleeping disturbances 0.48 and 0.77 and for aberrant eating habits 0.60 and 0.86. The aberrant sensibility items had a higher sensitivity, they also had lower specificity, and that actually all the items had a very similar data value. Hence the differences between the items are due to a criterion shift, with all subjects having a greater bias to answer yes to the aberrant sensibility items.
Discussion
The diagnostic criteria for autistic disorders used in the ICD-10 and DSM-IV manuals apparently represent the experts' generally accepted views of the clinical delineation of AS at the time the latest revision was made. One evident drawback of these sets of diagnostic criteria is that the distribution of symptoms and signs underlying the criteria in an average population has not been satisfactorily outlined. The same also applies to many of the sets of screening questionnaires and other instruments used in the diagnostic process. It can even be questioned whether currently used instruments actually cover the core features of AS. Using several such instruments and including only individuals who fulfilled both ICD-10 and DSM-I0 criteria for AS, as was the case in this study should improve the detection rate of the features typical for AS.
The data was collected by a few experienced investigators using a standardized procedure that is expected to increase the reliability of the results. The same sets of diagnostic instruments were used for each age group to facilitate the comparison of the results. The chosen set of diagnostic instruments was considered to have several advantages including ADI-R [25] that covers the entire lifetime. Using different set of diagnostic instruments for each age group would have hampered the comparability of the results. The only new method used in this study was the Asperger syndrome related additional features questionnaire (ASFQ)(Appendix 1), which was designed specifically for the purpose of detecting additional clinical feature much in the same way as in ICD-10 and DSM-IV. This instrument was not validated, but in a pilot-like setting to find out whether family members with or without AS differed from each other it was considered as a suitable method. Another drawback of the ASFQ method is that the results are based on self-reported perceptions that are not objective. The relevance of ASFQ needs to be studied in a control group of average population before its applicability as a diagnostic instrument for AS can be assessed.
The families chosen for this analysis may not necessarily completely reflect the average AS population as we for analytical reasons focused on large pedigrees with several AS individuals. Nor is the series unequivocally a population-based one, as the families were recruited from originally clinical series in which the index case had been referred for a diagnostic evaluation. By using non-AS individuals from the same families as a comparison group, we think that the effects of the family social environment are eliminated at least to some extent. The methodological weakness need to be taken into account, but the results seem to be fairly straightforward, as face recognition difficulties and aberrant sensibility emerge as being fairly typical of AS.
The basic aim of the present research project on AS was to elucidate the pathogenesis, especially with respect to the molecular genetic issues [26]. As AS is a complex disorder it seems likely that the discovery of endophenotypes could be of help for the identification of both putative susceptibility genes and specific neuroimaging features. In line with this, it seems attractive to speculate that our findings indicate the existence of a general sensory processing abnormality in AS as these individuals frequently are sensitive to noise, bright lights, strong smells and touch. Thus, these findings may be useful as well for diagnostic purpose as for understanding of the pathogenesis. Our findings are supported by Dunn et al. [13,14] who considered that the contributions of sensory processing status in diagnostic criteria might be warranted. Another possibility is that individuals with AS can or should be sub-grouped on the basis of their pattern of sensory processing status, sleeping and eating problems.
In the case of face recognition difficulties, there are no data on its occurrence in the average population, but it is not unlikely that the non-AS family members may also represent an increased occurrence. In our study, 50 % of the persons with AS had difficulties in recognition of faces. This is in line with the study by Barton et al. [10], which showed that 2/3 of the persons with AS had heterogeneity in the perceptual or imagery processing of faces. Thus, these two lines of evidence support the assumption that individuals with AS have face recognition difficulties. Face recognition is one of the most complex visual tasks performed by the human brain [10]. According to the literature, persons with autism spectrum disorders process faces as if they were objects and appear to rely more on identifying individual pieces of the face rather than the overall configuration [27-29]. This is significant data, because recognition of faces is an important part of interpersonal interaction and successful functioning within a social group. These findings and the result of our study makes it attractive to speculate that the deficits in face perception may contribute to the impaired social communication in persons with AS. Therefore, including face recognition difficulties in the clinical criteria for AS may also be warranted, as it might represent a core dysfunction of AS. This is in line with the study by Barton et al. [10], which concludes the same point, that heterogeneity may be important to take into account in genetic investigations.
Aberrant sensory perception was a major finding in our study, especially in the domains of sound and touch. This is in agreement with the findings made by Dunn et al. [13,14] and Myles at al. [12], as it seems that the auditory and touch processing are the two most important aberrant sensory domains. The nature of aberrant auditory sensory processing in AS and autism is still open. According to Baranek [30], altered hyperreactions in auditory processing occur in autism in up to 100% of the cases. The auditory sensory processing was studied using mitch-match-negativity recording technique by Jansson-Verkasalo et al. [15] and clinically using the Sensory Profile by Dunn et al. [13,14]. The former study indicated that auditory sensory processing was deficient in children with AS. Dunn et al. [13,14] showed that children with AS had difficulties in auditory processing, but that these had more difficulty with auditory attention than children with autism. On the other hand, Čeponienė et al. [31], demonstrated auditory orienting deficits in autism, which could not be explained by sensory deficits and concluded that the orienting deficit in autism might be speech-sound specific. Auditory processing impairments in AS and autism deserve special attention, as they may constitute a part of the pathogenesis of the behavioral problems. On a general level, our observations of a change in sensory integration in the AS group are well in line with the results of corresponding studies of autism spectrum disorders [12-14,20,30,31].
Insomnia predisposes to emotional distress, daytime fatigue and a loss of productivity [18]. A careful assessment of sleep quality should therefore be integrated in the diagnostic interview for AS, as, according to our study, insomnia is a frequent finding in this group. The studies by Paavonen et al. [17] and Tani et al. [18] demonstrate an increased prevalence of disorders in initiating and maintaining sleep in both children and adults with AS, suggesting that sleep difficulties inherent to AS are preserved from childhood into adulthood. It is not known why persons with AS have insomnia or sleeping problems, but the reports by Godbout et al. [32,33] suggest that AS subjects may present the same difficulties in initiating and maintaining sleep as those previously described in autism. It has also been hypothesized that individuals with AS have a defective sleep control system which may be associated with the clinical picture of AS [33]. From a clinical point of view, sleep disturbances are typical among people with AS and should be taken into account when a person is referred for a diagnosis, especially as the problem can be alleviated by melatonin [17].
Our study recognized aberrant eating habits in 60 per cent of individuals with AS, which is much higher than the figure of approximately one per cent obtained by Sobanski et al. [21]. Eating problems [21,34,35] and also anorexia nervosa have been reported to be over-represented in AS [20,34-36]. As, there according to Gillberg et al. [19], is a 12 per cent co-morbidity of anorexia nervosa and AS and there frequently is a picky eating behavior it could be speculated that this is due to hypersensitivity, as low thresholds for taste in AS was shown by Dunn et al. [13,14]. This could be part of the explanation for a higher incidence rate of eating disorders among persons with AS.
Conclusion
Our results speak in favour of the fact that aberrant sensory processing is an extensive phenomenon in AS, as these individuals frequently are hypersensitive to noise, bright lights, strong smells and touch. Our results are preliminary and need to be substantiated also using objective assessment methods before they can be seen as definite. Provided that the present findings can be substantiated in this way it might be possible to use the aberrant sensory processing also as a diagnostic criteria. It seems obvious that such aberrant sensory processing is a key feature of autism spectrum disorders, and our results will therefore probably not help in the differentiation between autism and AS.
Authors' contributions
TN-vW was the main principal investigator collecting the data and the series. She was also principally responsible for preparing the manuscript. Together with author LvW and TY-O, she conceived the idea of this study
EJP participated in the design of the study and performed the statistical analysis
TY-0 was co-conceiver of the idea of this study and also in writing the manuscript
SS together with TK carried out all the ADI-R and/or ADOS investigations
TK together with SS carried out all the ADI-R and/or ADOS investigations
IJ participated in designing the study and in the writing of the manuscript
LvW helped to collect the data and the series. LvW also participated in designing the study and in the writing of the manuscript
Note
APPENDIX 1. The complementary questions regarding symptoms and signs not included in the ICD-10 and/or DSM-IV.
Asperger syndrome related additional features questionnaire (ASFQ)
The following domains were included
1. Presence of face recognition difficulties
Do you have difficulty recognizing familiar persons from their faces?
Standardized clarifying questions were used, such as:
- If the clothing, color of the hair of an individual familiar to you has changed or you meet unexpectedly in an unfamiliar place, would you be able to recognize him/her?
- Do you generally recognize people on the basis of their style of walking, clothing, voice-/ speech rather than by their face?
2. Presence of aberrant sensibilities
The aim was to find out whether the individual considered him/herself and-/or was considered to represent hyper sensibility in the following modes of sensation:
- touch (light or strong)
- pain
- light
- external temperature
- sound
- smell
Fixed sets of clarifying questions were used.
3. Sleeping/diurnal rhythm disturbances:
The clarifying questions dealt with: getting to sleep, awakenings during the night, amount of sleep, abnormality of diurnal rhythm, use of sleeping pills, restless legs, impact of possible difficulties related to diurnal rhythm of every day life
4. Aberrant eating habits
The fixed clarifying questions comprised the following items:
- unusual-/odd food preferences
- unusual routines regarding arranging food on the plate and eating
- presence of aversion due to the smell of the food
- ongoing or past anorectic behavior
5. Behavior
The clarifying questions pinpointed possible presence of selection difficulties in ordinary every day situations (ice-cream a or b).
It was also established whether making up one's mind in such situation lead to avoidance of such situations. It was also made clear whether the person in such situations used clearly excessive time for making the decisions or whether these caused anxiety?
Another set of questions aimed at clarifying whether the individual him/herself recognized traits of hyperactivity and/or attention difficulties or whether other people had made him/her aware of such traits.
6. External appearance
Possible presence of unusual physical features with respect to stature, flexibility of joints and minor stigmata were recorded.
In both genders the questions aimed at finding out whether the individual frequently was wrongly perceived with respect to chronological age. Also the interviewer's perception of a possible youngish appearance was recorded.
In males questions also were focused on the age for starting regular shaving and the maximum interval between shavings which was compatible with socially acceptable tidiness of the face. Answers reporting such shaving less frequently than twice a week were considered being out of average range for adult Finnish men.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by grants from the Academy of Finland, the Foundation for Pediatric Research in Finland, Helsinki University Hospital Research Funding, the Medical Society of Finland, the Päivikki and Sakari Sohlberg Foundation and the Perklen Foundation.
==== Refs
Wing L Asperger's syndrome: a clinical account Psychol Med 1981 11 115 130 7208735
Gillberg C Gillberg C Disorders of empathy: autism and autism spectrum disorders (including childhood onset schizophrenia) Clinical Child Neuropsychology 1995 1 1 Cambridge: Cambridge University Press 54 111
World Health Organization International classification of disease (10th ed, chap 5) Mental and behavioral disorders 1993 Diagnostic criteria for research. Geneva: Author
American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders (4th Ed) (DSM-IV) 1994 Washington DC: American Psychiatric Association
Szatmari P The classification of autism, Asperger's syndrome, and pervasive developmental disorder Can J Psychiatry 2000 45 731 738 11086556
Kracke I Developmental prosopagnosia in Asperger syndrome: presentation and discussion of an individual case Dev Med Child Neurol 1994 36 873 886 7926319
Njiokiktjien C Verschoor A de Sonneville L Huyser C Op het Veld V Toorenaar N Disordered recognition of facial identity and emotions in three Asperger type autists Eur Child Adolesc Psychiatry 2001 10 79 90 11315539 10.1007/s007870170050
Duchaine BC Nieminen-von Wendt T New J Kulomäki T Dissociations of visual recognition in a genetic prosopagnosic: Evidence for separate developmental processes Neurocase 2003 9 380 389 14972753 10.1076/neur.9.5.380.16556
Pietz J Ebinger F Rating D Prosopagnosia in a preschool child with Asperger syndrome Dev Med Child Neurol 2003 45 55 57 12549756
Barton JJ Cherkasova MV Hefter R Cox TA O'Connor M Manoach DS Are patients with social developmental disorders prosopagnosic? Perceptual heterogeneity in the Asperger and socio-emotional processing disorders Brain 2004 1706 1716 15215211 10.1093/brain/awh194
Asperger H Die "Autistischen Psychopathen" im Kindesalter Arch Psychiatr Nervenkr 1944 117 76 136
Myles B Hagiwara T Dunn W Rinner L Reese M Sensory issues in children with Asperger syndrome and autism ETDD 2004 39 283 290
Dunn W Myles BM Orr S Sensory processing issues associated with Asperger syndrome: a preliminary investigation Am J Occup Ther 2002 56 97 102 11833406
Dunn W Saiter J Rinner L Asperger syndrome and sensory processing: a conceptual model and guidance for intervention planning Focus on autism and other developmental disabilites 2002 17 172 185
Jansson-Verkasalo E Ceponiene R Kielinen Suominen K Jäntti Linna SL Moilanen I Näätänen R Deficient auditory processing in children with Asperger Syndrome, as indexed by event-related potentials Neurosci Lett 2003 338 197 200 12581830 10.1016/S0304-3940(02)01405-2
Taira M Takase M Sasaki H Sleep disorder in children with autism Psychiatry Clin Neurosci 1998 52 182 183 9628139
Paavonen EJ Nieminen-von Wendt T Vanhala R Aronen E von-Wendt L Effectiveness of melatonin in the treatment of sleep disturbances in children with Asperger Syndrome J Child Adol Psychop 2003 13 83 95 10.1089/104454603321666225
Tani P Lindberg N Nieminen-von Wendt T von Wendt L Alanko L Appelberg B Porkka-Heiskanen T Insomnia is a frequent finding in adults with Asperger syndrome BMC Psychiatry 2003 3 12 28 14563215 10.1186/1471-244X-3-12
Gillberg C Råstam M Do some cases of anorexia nervosa reflect underlying autistic-like condition? Behav Neurol 1992 5 27 32
Gillberg C Billstedt E Autism and Asperger syndrome: coexistence with other clinical disorders Acta Psychiatr Scand 2000 102 321 330 11098802 10.1034/j.1600-0447.2000.102005321.x
Sobanski E Marcus A Hennighausen K Hebebrand J Schmidt MH Further evidence for a low body weight in male children and adolescents with Asperger's disorder Eur Child Adolesc Psychiatry 1999 8 312 314 10654125 10.1007/s007870050106
Bölte S Ozkara N Poustka F Autism spectrum disorders and low body weight: is there really a systematic association? Int J Eating Disor 2002 31 349 351 10.1002/eat.10015
Gillberg C Gillberg C Råstam M Wentz E The Asperger Syndrome (and high-functioning autism) Diagnostic Interview (ASDI): a preliminary study of a new structured clinical interview Autism 2001 5 57 66 11708390 10.1177/1362361301005001006
Lord C Risi S Lambrecht L Cook EH JrLeventhal BL DiLavore PC Pickles A Rutter M The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits associated with the spectrum of autism J Aut Dev Disord 2002 30 205 223 10.1023/A:1005592401947
Lord C Rutter M Le Couteur A Autism Diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders J Autism Dev Disord 1994 24 659 685 7814313
Ylisaukko-oja T Nieminen-von Wendt T Kempas E Sarenius S Varilo T von Wendt L Peltone I Järvelä I A Genome-wide Scan for Loci of Asperger Syndrome Mol Psychiatry 2004 9 161 168 14966474 10.1038/sj.mp.4001385
Klin A Schultz R Cohen DJ Baron-Cohen S, Tager-Flusberg H, Cohen JD Theory of mind in action: developmental perspectives on social neuroscience Understanding Other Minds Perspectives from Developmental Cognitive Neuroscience 2000 2 Oxford University Press: Oxford 357 388
Schultz R Romanski L Tsatsanis K(a) Klin A, Volkmar FR, Sparrow SS Neurofunctional models of autistic disorders and Asperger syndrome. Clues from neuroimaging Asperger syndrome 2000 1 New York: The Guilford Press 172 209
Schultz R Gauthier I Klin A Fulbright R Anderson A Volkmar F Skudlarski P Lacadie C Cohen D Gore J(b) Abnormal ventral temporal cortical activity during face discrimination among individuals with autism and Asperger syndrome Arch Gen Psychiatry 2000 57 331 340 10768694 10.1001/archpsyc.57.4.331
Baranek GT Efficacy of sensory and motor interventions for children with autism J Aut Dev Disord 2002 32 397 422 10.1023/A:1020541906063
Čeponienė R Lepistö T Shestakova A Vanhala R Alku P Näätänen R Yaguchi K Speech-sound-selective auditory impairment in children with autism: They can perceive but do not attend Proc Natl Acad Sci U S A 2003 100 5567 5572 12702776 10.1073/pnas.0835631100
Godbout R Bergeron C Stip E Mottron A laboratory study of sleep and dreaming in a case of Asperger's syndrome Dreaming 1998 8 75 88 10.1023/B:DREM.0000005898.95212.58
Godbout R Bergeron C Limoges E Stip E Mottron L A laboratory study of sleep in Asperger's syndrome NeuroReport 2000 11 127 130 10683843
Råstam M Gillberg C Gillberg IC Johansson M Alexithymia in anorexia nervosa: a controlled study using the 20-item Toronto Alexithymia Scale Acta Psychiatr Scand 1997 95 385 388 9197902
Ahearn WH Castine T Nault K Green G An assessment of food acceptance in children with autism or pervasive developmental disorder-not otherwise specified J Autism Dev Disord 2001 31 505 511 11794415 10.1023/A:1012221026124
Hebebrand J Henninghausen K Nau S Himmelmann GW Schulz E Schafer H Remschmidt H Low body weight in male children and adolescents with schizoid personality disorder or Asperger's disorder Acta Psychiatr Scand 1997 96 64 67 9259226
| 0 | PMC1097742 | CC BY | 2021-01-04 16:37:36 | no | Biomed Eng Online. 2005 Apr 15; 4:27 | latin-1 | Biomed Eng Online | 2,005 | 10.1186/1475-925X-4-27 | oa_comm |
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-121586012810.1186/1475-2867-5-12Primary ResearchDietary omega-3 fatty acids and ionizing irradiation on human breast cancer xenograft growth and angiogenesis Hardman W Elaine [email protected] LuZhe [email protected] Nicholas [email protected] Ivan L [email protected] Pennington Biomedical Research Center, Louisiana State University System, Baton Rouge, Louisiana 70808 USA2 University of Texas Health Science Center at San Antonio, Department of Cellular and Structural Biology, San Antonio, Texas 78229 USA2005 28 4 2005 5 12 12 7 1 2005 28 4 2005 Copyright © 2005 Hardman et al; licensee BioMed Central Ltd.2005Hardman 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 effects of an omega-3 (n-3) fatty acid enriched diet alone and in combination with gamma irradiation (IR) therapy in nude mice bearing a human MDA-MB231 breast cancer xenograft were tested. The cancer cells were injected into the mammary fat pad of young female mice. Six weeks later, mice were randomly divided into two diet groups: 1) mice with 10% corn oil (rich in omega 6 fatty acids) in their food, 2) mice consuming a 10% fat diet that was enriched in n-3 fatty acids. After two weeks on the diet, treatment with 200 cGy of IR every second day for four treatments (total 800 cGy) was initiated on half of the mice from each diet group. Some mice in each of the 4 groups were euthanized 24 hours after the end of IR while the remaining mice were followed for 3 additional weeks. Tumor sections were stained for endothelial cells with CD31 and PAS and for hypoxia inducible factor 1α (HIF-α).
Results
The tumor cortex within 100 microns of the well-vascularized capsule had little vascularization. Blood vessels, capillaries, and endothelial pseudopods were found at areas greater than 100 microns from the capsule (subcortex). Mice on the corn oil diet and treated with IR 24 hours previously or non-irradiated mice fed the n-3 diet had tumors with fewer blood vessels in the subcortex and more endothelial pseudopods projecting into hypoxic (HIF- α positive) areas than did mice from the non-irradiated corn oil fed group. The tumor growth rate of mice that received IR or that were fed the n-3 fatty acid enriched diet was significantly slower than in the mice fed the 10% corn oil diet. Harmful side effects were found only in the IR treated mice.
Conclusion
The omega-3 fatty acid enriched diet proved to be a safe means for retarding tumor growth and vascularization.
==== Body
Background
Animal and human epidemiological studies indicate that increasing the consumption of omega-3 (n-3) fatty acids versus intake of omega-6 (n-6) fatty acids decreased risk of cancer [1-10]. Animal studies indicate that increased consumption of n-3 fatty acids can slow the growth of cancer xenografts and can increase the efficacy and decrease the side effects of several chemotherapy agents [11-18]. Although consumption of an n-3 fatty acid enriched diet has been demonstrated to reduce tumor growth in animal xenografts, it is more likely to be used in combination with chemotherapy or with ionizing radiation (IR) therapy in the treatment of cancer patients. More than half of all cancer patients are given IR therapy during their course of treatment [19]. It therefore seemed important to determine the tumor growth retarding effects of IR and of an n-3 fatty acid enriched diet both alone and in combination. As both IR [20,21] and dietary n-3 fatty acids [22,23] are reported to suppress tumor angiogenesis, it was reasoned that the combined use of these two treatment modalities might have an additive effect on suppression of tumor growth and of angiogenesis. It was also decided to test the idea that continued consumption of n-3 fatty acids following a course of IR therapy might suppress regrowth of the IR treated tumor, perhaps by continued suppression of tumor angiogenesis.
The study reported here was therefore designed to investigate the potential of omega-3 fatty acids in the diet to inhibit growth and angiogenesis of a human breast cancer xenograft and to compare the effects of: 1) a commonly used course of IR involving exposure to 200 cGy every second day for a total of 800 cGy, 2) an n-3 fatty acid enriched diet, and 3) a combination of these two treatment regimens on tumor growth, on tumor angiogenesis, and on the side effects of each treatment regimen. Although this study used whole body IR therapy, most IR therapy of human patients is restricted to targeted regions of the body to minimize general side effects of IR treatment.
Our study results demonstrate the potential of the n-3 enriched fatty acid diet to inhibit tumor growth and angiogenesis without harmful side effects.
Results
The four groups of mice are: 1) corn oil diet, no IR treatment, 2) corn oil diet with IR treatment, 3) n-3 enriched diet, no IR treatment, 4) n-3 enriched diet with IR treatment.
Once the mice were divided into treatment groups, the body weight of each mouse was measured every 3 to 4 days for the remainder of the experiment. As illustrated in Fig. 1, the two groups that received IR therapy every second day for 4 doses lost body weight beginning during IR therapy and lasting for about 8 or 9 days. After completion of the IR therapy, the irradiated mice again began to regain their weight toward the mean weight of the two groups of mice not subjected to IR therapy. The mean body weights of the groups of mice that did not receive IR continuously increased until the end of the experiment and were not different due to the diet.
Figure 1 Mean body weight of each group during the experiment. The two groups of mice that received gamma irradiation lost body weight during and for a few days after the course of exposure, but the body weights later recovered towards the weights of the two groups of mice not exposed to gamma irradiation.
Fig. 2 illustrates mean tumor volume for each of the four groups of mice starting at the beginning of IR therapy. All tumors in each of the four groups were less than 35 mm3 and the mean tumor sizes of all four groups were similar at the start of IR treatment period. The mean growth rate of tumors from the non-irradiated mice that received the corn oil diet was significantly faster (p < 0.001) than that of the other three groups of mice (Fig. 3). The tumor growth rates of the remaining three groups of mice were not significantly different from each other.
Figure 2 The effect on tumor growth of consumption of a diet containing 10% corn oil (n-6) or an n-3 enriched diet (n-3) with or without gamma irradiation (IR). The tumors were all 35 mm3 or less at the start of treatment.
Figure 3 Tumor growth rates of each treatment group. The growth rate is the slope of the linear regression (Fig. 2) of the tumor volumes over time. Different letters demonstrate significant differences as shown by ANOVA. Compared to the group fed the n-6 diet and not irradiated, either consumption of the n-3 enriched diet or gamma irradiation significantly reduced the tumor growth.
Tumor vascularization patterns were determined from histological examination of midsections of PAS stained tumors. Tumors with a volume greater than 35 mm3 demonstrated a connective tissue capsule with blood vessels (Fig. 4A). The cortical area of the tumors within about 100 μm of the capsule revealed few blood vessels while the subcortical area greater than 100 μm from the capsule showed evidence of blood vessels and capillaries with many endothelial pseudopods extending away from the capillaries (Fig. 4B&C). Immunohistochemical localization of CD-31, used as a specific marker of endothelium, demonstrated a positive reaction to blood vessels, capillaries, and pseudopods (Fig. 4D). Areas of tumor necrosis were observed below the subcortical area (Fig. 4E&F). Immunohistochemical localization of hypoxia-inducible factor 1-α (HIF-α) revealed the subcortical area of the tumor to contain HIF positive cells while the tumor capsule, cortex and necrotic areas of the tumor demonstrate no evidence of HIF. Thus, the area found to be HIF positive was enriched in endothelial pseudopods.
Figure 4 Photomicrographs illustrating the pattern of the tumor vascular network. A, B and C illustrate endothelial pseudopods stained with PAS, D illustrates pseudopods stained with CD-31, a specific endothelial marker and E and F illustrate the localization of HIF. (A) The tumor capsule (left) reveals blood vessels. The cortex under the capsule reveals no blood vessels and few endothelial pseudopods while the subcortical area to the right has more pseudopods. (B) The subcortical area of the tumor reveals a small blood capillary with multiple endothelial pseudopods protruding at right angles into the tumor mass. (C) At higher magnification endothelial pseudopods are seen to branch. (D) The endothelial pseudopods react positively to the CD-31 specific endothelial marker. (E) Viable cell area can be seen beneath tumor capsule (left). Necrotic area can be seen to the right. (F) Enlarged subcortical area from E. In the subcortex, the brown stain localizes HIF in the hypoxic area between the viable and the necrotic tissue.
Ocular grid intercept counting was used to quantify tumor vascularization in 8 μm thick PAS stained histological sections of the tumors. The numbers of ocular grid intercepts were scored in the subcapsular regions of the tumor over blood vessels and capillaries, endothelial pseudopods and over the area with no indication of these structures. This method has been shown to be a usable measure of volume density occupied by recognizable structures. The results of the scoring of blood vessels and of endothelial pseudopods in the subcortical regions of the tumors are summarized in Fig. 5A&B. As illustrated in Fig. 5A&B, in n-6 fed mice, the volume density of tumor blood vessels was significantly decreased but the volume density of tumor pseudopods was significantly increased at one day after the last dose of irradiation compared to the n-6 fed mice that did not receive IR. The blood vessel and pseudopods volume densities of the tumors of IR treated corn oil fed mice returned to the level of the non-irradiated corn oil fed mice by 22 days after the last dose of irradiation. However, in non-irradiated mice fed the n-3 fatty acid enriched diet, the tumor blood vessel volume density was significantly lower and the pseudopod volume density was significantly higher than in the non-irradiated n-6 fed mice. IR treatment did not significantly change the volume densities of either blood vessels or pseudopods in the tumors of n-3 fed mice. Statistical analysis of the mean blood vessel volume density versus the pseudopod volume density revealed a significant correlation (r = 0.970, Fig. 6) to an exponential fit. Thus, as the blood vessel volume density increased, the endothelial pseudopod volume density decreased.
Figure 5 Quantification of vascularization in the tumor subcortex of each treatment group. The percent of areas (volume density) of blood vessels (A) and the percent of area of endothelial pseudopods (B) were determined using an ocular grid intercept counting method. The mean ± SEM of each treatment group is graphed. Columns that do not share a common letter within a graph are significantly different (p < 0.01). Data from tumors collected at the early and late times of euthanasia of non-irradiated, n-6 or n-3 fed mice were pooled because there were no significant differences between these groups due to time of euthanasia.
Figure 6 Correlation between blood vessel and pseudopod volume density (% area) of the subcortical areas of tumors. Data values are the mean values from Figure 5A&B. The correlation coefficient to a non-linear (logarithmic) equation is 0.970, indicating a significant inverse logarithmic fit. The pseudopod volume density decreased logarithmically as the blood vessel volume density increased.
Histological sections of viable areas of the tumor including both the cortex and subcortical regions were scored for number of metaphase figures. At least 1000 cancer cells were counted in each H&E stained tumor. The results are summarized in Fig. 7. The tumors of the corn oil fed group not treated with IR had the highest metaphase index. The tumors of groups of mice fed the n-3 fatty acids either with or without IR treatments had significantly fewer metaphase figures than the tumors of mice that were fed the corn oil diet but that were not treated with IR.
Figure 7 Metaphase index (mean % of cells ± SEM) in viable areas of the tumor. Mean values that do not share the same letter are significantly different.
To assess side effects, mice were euthanized at both one day and at 22 days after the last IR exposure. Data on liver and spleen weights are summarized in Table 1, and data on blood counts are summarized in Table 2. There were no statistically significant differences in mean liver weights among treatment groups at the earlier or later kills. On the other hand, at one day after their last IR exposure, the spleen weights of the two groups of mice that received IR were significantly less than the spleen weights of the groups that did not receive IR. The mean spleen weights of the IR treated mice recovered to at least the level of the non-irradiated corn oil fed mice by 22 days after their last IR exposure.
Table 1 Omega-3 Fatty Acid Enriched Diet (n-3) and Gamma Irradiation Therapy (IR) on Liver and Spleen Weights (Means ± SEM)
Therapy Group n Liver weight (grams) Spleen weight (grams)
Early Kill1
n-6 5 0.94 ± 0.04 0.107 ± 0.009
n-6/IR 5 0.87 ± 0.07 0.031 ± 0.004
n-3 10 1.22 ± 0.06 0.150 ± 0.010
n-3/IR 5 1.05 ± 0.07 0.033 ± 0.003
Late Kill2
n-6 9 1.26 ± 0.07 0.166 ± 0.008
n-6/IR 20 1.17 ± 0.03 0.146 ± 0.008
n-3 20 1.19 ± 0.04 0.166 ± 0.014
n-3/IR 13 1.20 ± 0.05 0.240 ± 0.027
1Mice euthanized one day after last irradiation treatment.
2Mice euthanized 22 days after last irradiation treatment.
Table 2 Omega-3 Fatty Acid Enriched Diet (n-3) and Gamma Irradiation Therapy (IR) on Blood Counts, (means ± SEM)
Therapy Group n WBC × 103/μL RBC × 106/μL Platelets × 103/μL Micronuclei (%RBC)
Early Kill1
n-6 4 2.91 ± 0.83 9.08 ± 0.07 584 ± 76 1.3 ± 0.04
n-6/IR 4 0.11 ± 0.01 7.50 ± 0.16 389 ± 17 1.2 ± 0.03
n-3 7 2.02 ± 0.30 8.45 ± 0.12 634 ± 51 1.2 ± 0.03
n-3/IR 4 0.16 ± 0.03 7.66 ± 0.07 358 ± 78 1.5 ± 0.04
Late Kill2
n-6 8 1.85 ± 0.37 8.30 ± 0.14 551 ± 83 1.6 ± 0.06
n-6/IR 13 1.48 ± 0.32 7.06 ± 0.16 973 ± 98 1.1 ± 0.01
n-3 16 2.85 ± 0.42 8.31 ± 0.20 582 ± 61 1.3 ± 0.04
n-3/IR 10 0.92 ± 0.18 6.10 ± 0.40 918 ± 131 0.9 ± 0.03
1Mice euthanized one day after last irradiation treatment.
2Mice euthanized 22 days after last irradiation treatment.
Results of Statistical Analyses:
1. At one day after the end of irradiation therapy, there were significant decreases in WBC, RBC, and platelet counts due to the irradiation. There were no other significant differences at p < 0.01.
2. At 22 days after the end of irradiation therapy, WBC and RBC counts were still significantly decreased and there was a significant increase in platelet counts due to irradiation. There were no other significant differences at p < 0.01.
There was a significant decrease in WBC, RBC, and platelets counts attributed to IR at 1 day after the end of IR treatment (Table 2). There were no significant differences in WBC, RBC, and platelets counts attributed to the n-3 fatty acid diet. At 22 days after the end of the IR treatment, the WBC and RBC counts in the two IR groups were not quite as low as they were one day post IR but remained significantly lower than that of the two non-irradiated groups. Both IR groups had significantly higher platelet counts than any of the other groups by 22 days after the end of IR therapy. Clearly, platelet counts have made a significant compensatory rebound by this time.
Micronuclei are pieces of broken chromosomes left over after the nucleus is extruded from a mature RBC and are usually an indication of genetic damage. Statistical analysis of micronuclei counts indicated no significant differences in the fraction of RBCs with micronuclei among any of the four groups at either time of euthanasia.
Another measure of possible side effects was the scoring of mitotic activity in the duodenal crypts. Dividing cells, such as those in duodenal crypts, are particularly sensitive to IR. Fig. 8 summarizes results of mitotic activity in the duodenal crypts of mice euthanized one day after the end of the IR therapy regimen. The metaphase index in the duodenal crypts of the IR treated corn oil fed mice was significantly lower than in the mice that did not receive IR, indicating that dividing cells were killed. However, the metaphase index in the duodenal crypts of the n-3 fed mice that were IR treated was not significantly different from that of the n-3 fed mice that were not IR treated.
Figure 8 Number of metaphase figures per midaxial histological section of duodenal crypts. Mice were euthanized one day after the last gamma irradiation exposure. Column height indicates mean ± SEM. The columns that do not share a common letter are significantly different. The number of metaphase figures in the duodenal crypts of mice that received IR or that consumed the n-3 diet was significantly less than in the non-irradiated corn oil fed mice.
Discussion
As expected, both IR and the n-3 fatty acid enriched diet worked to suppress tumor growth. Statistical analyses of the tumor growth rate data (Fig. 3) allow comparisons of the efficacy of the four treatment procedures. The n-6 fed mice not treated with IR had the fastest tumor growth rate. The IR decreased the tumor growth rate by 90% in n-6 fed mice. Compared to the non-irradiated n-6 fed mice, the n-3 diet alone decreased the tumor growth rate by 66% whereas the combination of the n-3 and IR decreased the tumor growth rate by 78%. The tumor growth rates of the three treated groups were not significantly different. Thus, in this study, an n-3 fatty acid enriched diet alone was as effective at decreasing the growth rate of MDA-MB 231 as was the n-6 fatty acid containing diet combined with radiation. Radiation in combination with the n-3 diet was not significantly more effective than the n-3 diet alone. Since the IR treatment resulted in such good suppression of tumor growth, we cannot determine if the n-3 fatty acid diet increased the efficacy of the IR treatment or not. A different model in which IR treatment resulted in partial suppression of tumor growth would be needed to make a conclusion about the ability of n-3 fatty acids to increase the efficacy of IR therapy.
What then are the possible factors involved in the observed decrease in tumor growth rate due to the different treatment protocols? The data on blood vessel volume density indicate that either IR alone or the n-3 diet alone were equally effective at reducing the density of blood vessels in the tumors compared to that of the non-irradiated corn oil fed group. The IR treatment alone in the corn oil fed group suppressed tumor blood vessel volume density as assessed one day after the end of IR therapy. With time, the volume density of blood vessels in the tumors of the IR treated, corn oil fed group recovered to the level of that of the tumors in the corn oil fed non-irradiated group. However, the combination of the n-3 fatty acid diet and the IR continued to suppress the density of blood vessels in the tumors for 22 days after IR treatment ended.
The rapid proliferation of cancer cells requires angiogenesis to maintain oxygen and nutrient supplies to the growing tumor. Tumor cells that are far enough away (more than 100 to 150 μm) from a blood vessel can become hypoxic, which causes these tumor cells to produce HIF-α and vascular endothelial growth factor (VEGF) [21]. The VEGF then stimulates endothelial cells to sprout and produce pseudopods that project into hypoxic areas of the tumor [21]. The endothelial sprouts and pseudopods form vacuoles that fuse with capillary lumens, allowing entry of blood cells. Other steps in angiogenesis include degradation of the basement membrane surrounding existing vessels, recruitment of smooth muscle and pericyte cells, endothelial cell proliferation, lumen formation, generation of a new basement membrane, fusion of newly formed vessels and initiation of blood flow. The observed inverse relationship between the volume densities of blood vessels and of pseudopods, combined with the fact that pseudopods are most abundant in the area of the tumor most enriched in HIF-α, indicates that pseudopod volume density is an indicator of hypoxic regions in the tumor. Given this relationship, the increased pseudopod volume density as observed in tumors suggests that the subcortical areas are hypoxic. Tumor pseudopods were significantly increased one day after IR therapy in the corn oil fed mice. In the corn oil fed mice, the pseudopod volume density then decreased as the volume density of blood vessels increased compared to the non-irradiated corn oil fed mice during the 22 days recovery following the end of IR therapy. The fact that the n-3 retarded recovery of the tumor blood vessels but did not retard formation of pseudopods suggests that the n-3 diet is working to retard angiogenesis at one or more of the later steps following pseudopod formation.
There are reports that an n-3 fatty acid enriched diet can suppress mitosis and growth of breast and colon tumors [1,6,9]. Thus, it was not surprising to find that the n-3 diet also suppressed the metaphase index in viable areas of the breast cancer tumors in this study.
Together the data reveal that consumption of the n-3 containing diet resulted in a decrease in tumor growth rate, cell proliferation (Fig. 7) and blood vessel volume density (Fig. 5A). Hardman [24] has reviewed some of the possible molecular mechanisms involved in suppression of tumor growth by addition of n-3 fatty acids to the diet. The mechanisms involved in suppression of tumor growth by an omega-3 fatty acid enriched diet include: 1) decreased expression of cyclooxygenase-2, reducing angiogenesis and decreasing cancer cell proliferation, 2) suppression of nuclear factor κB activation and bcl-2 expression, allowing apoptosis of cancer cells, 3) suppression of the oncogenes AP-1 and ras, 4) induced differentiation of the cancer cells, 5) reduction in aromatase activity that decreases estrogen levels, 6) inhibition of later steps in the tumor angiogenesis process.
There has been concern that adding n-3 fatty acids to the diet might increase radiation damage to the host. N-3 fatty acids are more unsaturated than an n-6 fatty acid with the same number of carbons. The double bonds are susceptible to radiation damage and when incorporated in cellular phospholipids could increase the susceptibility of a cell to radiation induced lipid peroxidation. However, we have shown that the production of endogenous antioxidative enzymes is increased in normal cells (but not tumor cells) when n-3 fatty acids are included in the diet [18]. This increase in endogenous antioxidative enzymes could protect normal cells from oxidative damage. Damage to the bone marrow due to the n-3 diet did not seem to be increased based on either blood cell counts or numbers of micronuclei. It appears that the lower basal mitotic index in the duodenal crypts of n-3 fed mice may also have contributed to protecting the intestine from radiation damage. Irradiation is most damaging to proliferating cells. A lower basal rate of proliferation might be expected to result in less IR induced damage to cells with proliferative capacity. The metaphase index of the duodenum crypts of n-3 fed mice after IR is much closer to that of the non-irradiated mice than in the n-6 fed mice after IR, indicating either faster recovery or less initial killing by IR of proliferative cells in the duodenum in the n-3 fed mice.
The whole body radiation of mice was expected to demonstrate significant side effects within a day following the eight-day course of IR therapy. The observed side effects included significant decreases in WBC, RBC, and platelet counts as well as loss of spleen weight. These measurable side effects serve as reference for comparison with any side effects of the n-3 enriched diet alone or in combination with IR therapy. In this study, there was no evidence of increased side effects due to the n-3 diet nor did it seem to significantly lessen the measured side effects of IR therapy. This failure of the n-3 diet to lessen the side effects of IR therapy differs from past reports that omega-3 fatty acid supplements ameliorated side effects of several chemotherapeutic agents [11-15]. Most of the observed side effects with whole body irradiation are limited in human patients by targeting of IR to tumorous regions.
The n-3 diet alone significantly suppressed tumor growth. However, in the short term, the efficacy of the IR therapy was not significantly greater in mice that consumed the n-3 enriched diet than in the mice that consumed the corn oil diet. Others have reported that the efficacy of radiation therapy was increased in the presence of an n-3 containing diet [25]. It is speculated that the n-3 oil diet reduced blood vessel volume density, causing increased areas of the tumor to become hypoxic, thus rendering these hypoxic areas of the tumor less sensitive to the oxidative damage of IR. Had our experiment continued longer than three weeks after the radiation, the results might well have been different. Angiogenesis suppression by the n-3 diet would have continued to suppress growth of residual tumors in the n-3 fed mice but residual tumors in the n-6 fed mice were stimulating angiogenesis and could have resumed growth. In support of this idea, analyses of tumor growth in the IR treated n-3 fed and n-6 fed mice beginning 7 days after the end of radiation indicates that there may be a resumption of growth of the residual tumor in the n-6 fed mice but not in the n-3 fed mice. The tumor growth rate (mean ± SD) from linear regression analyses of the last four available data points of the n-6 fed, IR treated mice is 2.3 ± 0.5 mm3/day, a significant (p 60; 0.05) positive growth rate. However, the tumor growth rate (mean ± SEM) of the n-3 fed, IR treated mice for the last four data points is -1.0 ± 1.0 mm3/day, a rate that is not significantly (p = 0.5) different from 0. Due to the few points, these tumor growth rates are not quite significantly different (p = 0.07) yet the data does suggest that in a longer term study, n-3 may provide a better treatment advantage than is indicated by the results of this study.
Conclusion
In conclusion, an omega-3 fatty acid enriched diet was found to significantly reduce the growth rate and angiogenesis of a human breast cancer xenograft without evidence of harmful side effects.
Methods
All animal use and care procedures were approved by the Institutional Animal Care and Use Committees of the University of Texas Health Science Center at San Antonio and of the Pennington Biomedical Research Center. The human breast cancer cell line MDA-MB-231 was obtained from the American Type Culture Collection. The cell lines were cultured in McCoys's 5A medium supplemented with pyruvate, vitamins, amino acids, antibiotics, and 10% fetal bovine serum [26]. The MDA-MB-231 cells were harvested from exponential cultures and inoculated at 2 × 106cells/inoculum in the inguinal mammary fat pad area of 160 female athymic nude mice, 6 weeks of age (purchased from the Harlan Sprague Dawley, Inc., Indianapolis, IN). The animals were housed under pathogen-free conditions and fed an AIN-76 semipurified diet slightly altered to contain 10% w/w corn oil during the initial tumor growth period. Growth of each xenograft was monitored three times per week by externally measuring tumors in three dimensions using digital calipers. Xenograft volume (V) was determined by the following equation: V = (L × W × D) × 0.5, where L is the length, W is the width and D is the depth of a xenograft. When the tumors reached an average diameter of about 3 mm after 5 weeks, the mice were divided into 2 groups such that the mean and median of tumor volume of the groups were closely matched.
Half of the mice fed the 10% w/w corn oil diet for the remainder of the study, half of the mice were placed on the experimental diet that contained 5% w/w n-3 fatty acids supplement (AAFA, Incell, Corp., LLC, San Antonio, TX) and 5% w/w canola oil. Corn oil is 61% linoleic acid (18:2n-6) and about 1% n-3 fatty acid thus it provides primarily n-6 FAs. The 5% w/w canola oil was used as the source of essential n-6 FAs in the n-3 diet. Canola oil is 20% linoleic acid and it also contains about 10% alpha-linolenic acid (18:3n-3). Five percent w/w n-3 FAs supplement (AAFA) containing 6% n-6 fatty acid and 61% total n-3 fatty acids was added to greatly increase the n-3 content of this diet. The total w/w composition of the n-6 diet was 0.1% n-3 FAs and 6.1% n-6 FAs with an n-3/n-6 ratio of 0.1. The total w/w composition of the n-3 diet was 3.65% n-3 FAs and 1.3% n-6 FAs with an n-3/n-6 FAs ratio of 2.8:1. These diets provided large differences in the fatty acids thought to be associated with cancer growth, linoleic acid and n-3 FAs.
Half of the tumor-bearing mice of each diet group received a cumulative dose of 800 cGy of IR (200 cGy each second day for 4 cycles). The 200 cGy/day dosage is based on the results from a preliminary dose response study using the same mouse model and on the fact that 180 to 250 cGy/day is the commonly used acute dose for radiotherapy of humans. About 815 cGy whole body radiation is the LD50/30 for mice (50% lethal within 30 days). Mice were transferred to a circular cage with individual compartments for each mouse during irradiation. Mice were irradiated in a 137Cs Gamma Cell-40 Irradiator (Atomic Energy of Canada) facility in our Department of Radiology. A dosimetric analysis for this instrument is performed monthly for calculation of a precise 200 cGy exposure. Mice were euthanized at one day or 22 days after the last radiation treatment.
Mice were deeply anesthetized using a ketamine/rompun mixture prepared by the UTHSCSA Laboratory Animal veterinarian, cervically dislocated, then were exsanguinated by cardiac puncture. Blood was collected into an EDTA containing microtube for complete blood counts.
The tumor, duodenum, spleen and liver were removed at the time of euthanasia. Samples of the duodenum and the tumor were fixed in Omni Fix II (Mt. Vernon N.Y.) and paraffin embedded. Embedded tissues were cut 4 μm or 8 μm thick, cut sections were placed on microscope slides then deparaffinized and stained with hematoxylin and eosin or with periodic acid-Schiff (PAS) for morphological analysis. Additional sections were prepared for immunohistochemistry.
To determine the effect of treatment on tumor angiogenesis, we measured the vascularity of excised tumors. Tumor tissues were fixed and embedded in paraffin. Mid tumor sections (8 μm thick) were cut from the embedded tissue and stained with periodic acid Schiff (PAS). Sections were examined by light microscopy. CD31 immunostaining for mouse blood vessels on 4 μm thick sections was performed by incubating tumor sections with a rat antimouse CD-31 (PECAM-1) monoclonal antibody (PharMingen) at 5 μg/ml for 30 min at 37°C. Sections were then incubated with a biotin-labeled goat anti-rat IgG (Zymed; 1:200 dilution) for 30 min at room temperature, followed by ABC reagent kit (Vector Laboratories) for 30 min at room temperature. Color reaction was performed with 3, 3'-diaminobenzidine (Vector Laboratories) and counterstained with hematoxylin. Hypoxia-inducible factor-1 alpha (HIF-1α) immunohistochemistry was done following the instruction for antigen retrieval (Biogenex protocol) and iso-IHC (inno-Genex Mouse-on-Mouse iso-IHC kit) with these changes: Dewaxing was with 3 changes of xylene, 10 minutes/change. Rehydration in 100%, 90%, and 70% ethanol for 10 minutes each. Initial dilution of the antibody was 1:200. Stressgen anti-HIF-1 alpha, product # OSA-601, was the antibody used. All sections were coded, treated as above and the extent of blood vessels, endothelial cell pseudopods and total area volume density was scored using an ocular grid. The number of grid line intercepts over blood vessels and endothelial pseudopods gave a measure of the total volume density of these structures.
A blood cell counter with veterinary pack was used for counts of red cells, white cells and platelets in EDTA anticoagulated blood of the mice. Our Laboratory Animal Resources division performed this test. Micronuclei counts were determined on thin smears of whole blood. Blood smears were stained with 0.1% acridine orange in phosphate buffered saline (pH 7.4) for 10 seconds. Slides were then examined for the presence of micronuclei using a fluorescent microscope and a 100X oil immersion objective. At least 1000 acridine orange stained red blood cells were counted and the percentage of erythrocytes containing micronuclei was determined [27].
Fixed specimens of duodenum were trimmed, processed and oriented for paraffin embedding. Four μm thick sections of the paraffin blocks were mounted on slides. Complete midaxially sectioned crypts on H&E stained slides were selected for analyses. Complete crypts were defined as those with: 1) the crypt base at the muscularis mucosa, 2) an open lumen from mouth to base and 3) a single column of epithelial cells up each side of the crypt. The numbers of metaphase figures per midaxial crypt section were counted for 10 crypts of each mouse.
SAS computer software was used for statistical analyses of numerical data. Tests for normality (basic statistics) were used on each data set. One-way and two-way analyses of variance followed by Student-Newman-Keuls multiple range tests, as appropriate, was used to determine if there were statistically significant (p ≤ 0.05) differences in any measured parameter due to the therapies. Prism™ (Graphpad Software, Inc.) was used for statistical analyses of tumor growth data. The mean growth rate for each group was determined using least squares linear regression analysis of mean tumor volume with time. Analyses of variance of the slopes of the linear regression (growth rate) were used to determine statistical differences in mean growth rates between treatment groups.
Competing interests
WEH and ILC are scientific advisors for Incell, Corp., LLC. They do not receive financial compensation from and no research support was received from Incell, Corp., LLC.
Authors' contributions
WEH and ILC contributed equally to writing this manuscript. LZS provided the tumor cells, NS assisted with daily animal care and graph preparation. All authors have read and approved this manuscript.
Acknowledgements
This work was supported by the American Institute for Cancer Research grant 01A010, NIH grant CA 75253 and funds from the Pennington Biomedical Research Foundation.
==== Refs
Rose DP Connolly JM Effects of dietary omega-3 fatty acids on human breast cancer growth and metastasis in nude mice J Natl Cancer Inst 1993 85 1743 1747 8411258
Reddy BS Sugie S Effect of different levels of omega-3 and omega-6 fatty acids on azoxymethane-induced colon carcinogenesis in F344 rats Cancer Res 1988 48 6642 6647 3180073
Gonzalez MJ Schemmel RA Gray JI Dugan LJ Sheffield LG Welsch CW Effect of dietary fat on growth of MCF-7 and MDA-MB231 human breast carcinomas in athymic nude mice: Relationship between carcinoma growth and lipid peroxidation product levels Carcinogenesis 1991 12 1231 1235 2070488
Lindner MA A fish oil diet inhibits colon cancer in mice Nutr Cancer 1991 15 1 11 2017394
Calviello G Paola P Maggiano N Franceschelli P Di Nicuolo F Marcocci ME Bartoli GM Effects of eicosapentaenoic and docosahexaenoic acids dietary supplementation on cell proliferation and apoptosis in rat colonic mucosa Lipids 1999 34 S111 10419112
Rose DP Connolly JM Liu XH Research AIC Fatty acid regulation of breast cancer cell growth and invasion Dietary Fat and Cancer 1997 New York, Plenum Press 47 56
Ziegler RG Hoover RL Pike MC Hildesheim A Nomura AMY West DW Wu-Williams AH Kolonel LN Horn-Ross PI Rosenthal JF Hyer MB Migration patterns and breast cancer risk in Asian-American women J Natl Cancer Inst 1993 85 1819 1827 8230262
Wynder EL Fujita Y Harris RE Hirayama T Hirayama T Comparative epidemiology of cancer between the United States and Japan: A second look. Cancer 1991 67 746 763 1985768
Abou-El-Ela SH Prasse KW Farrell RL Carroll RW Wade AE Bunce OR Effects of D,L-2-difluoromethylornithine and indomethicin on mammary tumor promotion in rats fed high n-3 and/or n-6 fat diets Cancer Res 1989 49 1434 1440 2538226
Maillard V Bougnoux P Ferrari P Jourdan ML Pinault M Lavillonnière F Body G Floch OL Chajès V N-3 and N-6 fatty acids in breast adipose tissue and relative risk of breast cancer in a case-control study in Tours, France Int J Cancer 2002 98 78 83 11857389 10.1002/ijc.10130
Hardman WE Moyer MP Cameron IL Dietary fish oil sensitizes A549 lung xenografts to doxorubicin chemotherapy Cancer Lett 2000 151 145 151 10738108 10.1016/S0304-3835(99)00396-1
Hardman WE Moyer MP Cameron IL Fish oil supplementation enhanced CPT-11 (Irinotecan) efficacy against MCF7 breast carcinoma xenografts and ameliorated intestinal side effects Br J Cancer 1999 81 440 448 10507768 10.1038/sj.bjc.6690713
Hardman WE Moyer MP Cameron IL Small amounts of a concentrated omega-3 fatty acid product, INCELL AAFA, in the diet reduces the side-effects of the cancer chemotherapy drug, CPT-11 (irinotecan). Br J Cancer 2002 86 983 988 11953833 10.1038/sj.bjc.6600175
Hardman WE Avula CPR Fernandes G Cameron IL Three percent dietary fish oil concentrate increased efficacy of doxorubicin against MDA-MB 231 human breast cancer xenografts Clin Cancer Res 2001 7 2041 2049 11448922
Barber MD Ross JA Voss AC Tisdale MJ Fearon KCH The effect of an oral nutritional supplement enriched with fish oil on weight-loss in patients with pancreatic cancer Br J Cancer 1999 81 80 86 10487616 10.1038/sj.bjc.6690654
Beck SA Smith KL Tisdale MJ Anticachectic and antitumor effect of eicosapentaenoic acid and its effect on protein turnover Cancer Res 1991 51 6089 6093 1657378
Tsujita-Kyutoku M Yuri T Danbara N Senzaki H Kiyozuka Y Uehara N Takada H Hada T Miyazawa T Ogawa Y Tsubara A Conjugated docosahexanoic acid suppresses KPL-1 human breast cancer cell growth in vitro and in vivo: potential mechanisms of action Breast Cancer Res 2004 6 R291 R299 15217495 10.1186/bcr789
Hardman WE Munoz J Cameron IL Role of lipid peroxidation and antioxidant enzymes in omega-3 fatty acids induced suppression of breast cancer xenograft growth in mice Cancer Cell Int 2002 2 17 10.1186/1475-2867-2-10
Owen JB Coia LR Hanks GE Recent patterns of growth in radiation therapy facilities in the United States: a pattern of care study report Int J Radiat Onc Bio Phys 1992 24 983 986
Gorski DH Beckett MA Jasowiak NT Calvin DP Mauceri HJ Salloum RM Seetharam S Koons A Hari DM Kufe DW Blockade of the vascular endothelial growth factor stress response increases antitumor effects of ionizing radiation Cancer Res 1999 59 3374 3378 10416597
Cameron IL Sun LZ Short N Hardman WE Williams CD Therapeutic electromagnetic field (TEMF) and gamma irradiation on human breast cancer xenograph growth, angiogenesis and metastasis Cancer Cell Int 2005 In Press
Form DM Auerbach R PGE2 and angiogenesis Proc Soc Exp Biol Med 1983 172 214 218 6572402
Connolly JM Rose DP Enhanced angiogenesis and growth of 12-lipoxygenase gene-transfected MCF-7 human breast cancer cells in athymic nude mice. Cancer Lett 1998 132 107 112 10397460 10.1016/S0304-3835(98)00171-2
Hardman WE Omega-3 fatty acids to augment cancer therapy J Nutr 2002 132 3508S 3512S 12421878
Colas S Paon L Denis F Prat M Louisot P Hoinard C Le Floch O Ogilvie G Bougnoux P Enhanced radiosensitivity of rat autochthonous mammary mamary tumor by dietary docosahexaenoic acid Int J Cancer 2004 109 449 454 14961586 10.1002/ijc.11725
Bandyopadhyay A López-Casillas F Malik SN Montiel JL Mendoz V Yang J Sun LZ Antitumor activity of a recombinant soluble betaglycan in human cancer xenograft Cancer Res 2002 62 4690 4695 12183427
Vijayalaxmi Frei MR Dusch SJ Guel V Meltz ML Jauchem JR Frequency of micronuclei in the peripheral blood and bone marrow of cancer-prone mice chronically exposed to 2450 MHz radiofrequency radiation Radiat Res 1997 147 495 500 9092931
| 15860128 | PMC1097743 | CC BY | 2021-01-04 16:40:10 | no | Cancer Cell Int. 2005 Apr 28; 5:12 | utf-8 | Cancer Cell Int | 2,005 | 10.1186/1475-2867-5-12 | oa_comm |
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-111585751810.1186/1476-7120-3-11Case ReportReduced coronary flow reserve in Anderson-Fabry disease measured by transthoracic Doppler echocardiography Dimitrow Paweł Petkow [email protected] Marek [email protected] Anetta [email protected] 2nd Department of Cardiology, Collegium Medicum Jagiellonian University, Cracow, Poland2 2nd Department of Internal Medicine, Collegium Medicum Jagiellonian University, Cracow, Poland2005 27 4 2005 3 11 11 13 4 2005 27 4 2005 Copyright © 2005 Dimitrow et al; licensee BioMed Central Ltd.2005Dimitrow 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.
Coronary flow reserve was assessed in a patient with Anderson-Fabry disease complicated by symmetric left ventricular hypertrophy. Coronary flow reserve was measurable in all three major coronary arteries providing an opportunity to compare regional coronary flow reserve from different vascular beds. In this patient all the three vascular beds supplied diffusely hypertrophied myocardium. Coronary flow disturbances in small intramyocardial perforating arteries were visible. The coronary flow reserve was reduced to a similar level (around to 2.0) in all three major arteries. In our patient with Anderson-Fabry disease, the coronary vasodilatation was blunted in a diffuse pattern corresponding to the myocardial hypertrophy distribution. In small intramyocardial arteries coronary flow was also disturbed. Accordingly, retrograde systolic flow and accelerated anterograde diastolic flow were documented.
==== Body
Background
Anderson-Fabry disease is an X-linked, multisystem, lysosomal storage disease (deficiency of enzyme α-galactosidase A), characterized by the accumulation of glycosphingolipids in various tissues and organs [1,2], including skin, vascular endothelium, heart, kidneys, liver, lungs, pancreas and ganglion cells of the peripheral nervous system. The incidence is 1:117000. However, the rate may be underestimated because a common cardiac manifestation is myocardial hypertrophy that mimics hypertrophic cardiomyopathy. Abnormal storage in the cardiovascular system may also involve cardiac conduction system, valvular apparatus and endothelial cells in coronary vessels [1-3].
To assess coronary flow abnormalities in a patient with Anderson-Fabry disease, we performed transthoracic Doppler echocardiography. Using this method, coronary flow reserve is effectively measurable [4] and all three major coronary arteries are accessible in some patients [5,6]. In a large series of 658 patients [6], coronary flow reserve was contemporarily recorded in left anterior descending (LAD) coronary artery (98% of patients), right coronary artery (RCA) (66% of patients) and circumflex (Cx) coronary artery (43% of patients). Additionally, flow disturbances in small intramyocardial perforating arteries were assessed as blood flow abnormalities at this level of coronary circulation were previously reported in left ventricular (LV) hypertrophy [7].
Case presentation
This 49-year-old male patient with Anderson-Fabry disease was referred to our hospital. He did not complain of anginal symptoms or dyspnea. As a part of an overall clinical evaluation, transthoracic echocardiography was performed, revealing a diffusely distributed myocardial hypertrophy, i.e. involving both (LV) free walls and the septum (the myocardial thickness at diastole was measured in the short parasternal-axis: the anterior segment of the septum- 19.6 mm; the posterior segment of the septum – 20.8 mm; the LV posterior wall – 20.7 mm; and the anterolateral wall 16.8 mm). Left ventricular systolic function was preserved (LV ejection fraction 68%). A precise assessment of myocardial hypertrophy by magnetic resonance imaging confirmed increased LV mass to 386 grams.
Methods
Using noninvasive, inexpensive and widely accessible method, B-mode and color Doppler transthoracic echocardiography, the segments of three major coronary arteries: LAD, Cx and right posterior descending (RPD) coronary artery or RCA were visualized (Figure 1,2). High quality recordings of flow velocity in all these coronary arteries were obtained using spectral Doppler.
Figure 1 Parasternal short axis view, color Doppler examination at the heart base level. Left main, left anterior descending (LAD) and origin of left circumflex (Cx) coronary arteries are seen. Blood flow within the left main coronary artery and proximal LAD is depicted in red, while in the Cx – in blue.
Figure 2 Modified apical four-chamber view, color Doppler examination. Flow detected in long segment of the right posterior descending (RPD) coronary artery is depicted in red.
Results
The coronary flow reserve in response to intravenous adenosine (140 μg/kg/min) was homogeneously reduced to a similar value in the major coronary arteries (LAD – 2.07; Cx- 2.18; RPD/RCA- 1.91). Small, intramyocardial branches of epicardial coronary arteries were visualized [intramyocardial perforators originated from LAD (Figure 3) and branches from RPD (Figure 4)]. Increased resistance to flow (probably due to myocardial hypertrophy and increased extravascular compressive forces) was demonstrated by the detection of flow with high velocity in spectral Doppler (figure 5, 6, 7) and color Doppler (Figures 3, 4, 5) in these penetrating vessels. The peak diastolic flow velocity was higher in the LAD perforator (41 cm/s – figure 6) than in the distal portion of LAD (25 cm/s – figure 7). The systolic flow in the LAD perforator was abnormally retrograde (figure 6).
Figure 3 Modified apical 2-chamber view, apical area, color Doppler examination. The LAD and the penetrating intramyocardial arteries (vertical perforators branching of the LAD) are seen.
Figure 4 Modified apical 2-chamber view, middle segment of the inferior wall on color Doppler examination. Small (arch-shaped) arteries branching of the right posterior descending (RPD) coronary artery are seen.
Figure 5 The orientation of Doppler gate within the LAD perforator.
Figure 6 The flow velocity spectrum in the LAD perforator visualized in figure 5. In diastole the flow velocity is negative (towards ventricular chamber) and during systole the retrograde flow (towards epicardium) results in positive value of velocity.
Figure 7 Diastolic flow velocity spectrum in distal portion of LAD.
Discussion
Prevalence of Anderson-Fabry disease in patients with late-onset hypertrophic cardiomyopathy is about 6.3% in males [8] and 12% in females [9]. In contrast, among males with hypertrophic cardiomyopathy diagnosed at <40 years of age, the rate of appropriate verification of the diagnosis to Anderson-Fabry disease was lower i.e. 1.4% [8]. If properly recognized, Anderson-Fabry disease is treatable by enzyme replacement therapy and both cardiac and non-cardiac abnormalities may be reversed/reduced by substitution of α-galactosidase (especially in the early stage) [10]. Therefore, it is important to consider Anderson-Fabry disease in the differential diagnosis of hypertrophic cardiomyopathy. Anderson-Fabry disease is hardly indistinguishable from hypertrophic cardiomyopathy by echocardiography [9,11,12], however we made an attempt to identify potential differences (mainly quantitative) in table 1. Recently [13], the findings of magnetic resonance imaging have appeared useful in differential diagnosis (table 1). Clinical findings may be more helpful in differential diagnosis and we recommend to evaluate the presence or absence of all non-cardiac manifestations (dermatological, nephrological, neurological, ophthamological) of Anderson-Fabry disease in patients diagnosed as having hypertrophic cardiomyopathy.
Table 1 Comparison of echocardiography and MRI findings between Anderson-Fabry and hypertrophic cardiomyopathy.
Echocardiography
Anderson-Fabry HCM
LVH pattern Concentric>Asymmetric Asymmetric>Concentric
Resting LVOT gradient Rare 25–30%
Massive LVH> 30 mm Rare 10–15%
LV diastolic dysfunction (tissue Doppler, strain rate)
With LVH Yes Yes
Without LVH (genotype +) Yes Yes
LV systolic dysfunction More frequent especially in older males
Magnetic resonance
Most common site of late-enhancement (if present) Infero-lateral segment Ventricular junction, Multi-focal
LVH- left ventricular hypertrophy; LVOT – left ventricular outflow tract
Conclusion
In our patient with Anderson-Fabry disease, the coronary vasodilatation was blunted in a diffuse pattern corresponding to the myocardial hypertrophy distribution. In small intramyocardial arteries coronary flow was disturbed. Accordingly, retrograde systolic flow and accelerated anterograde diastolic flow were documented. Transthoracic Doppler echocardiography is now the only method available to evaluate blood flow characteristics in small intramyocardial arteries.
List of Abbreviations
LAD – left anterior descending coronary artery
RCA – right coronary artery
RPD – right posterior descending coronary artery
Cx – circumflex coronary artery
Competing Interests
The author(s) declare that they have no competing interests.
==== Refs
Linhart A Palecek T Bultas J Ferguson JJ Hrudova J Karetova D Zeman J Ledvinova J Poupetova H Elleder M Aschermann M New insights in cardiac structural changes in patients with Fabry's disease Am Heart J 2000 139 1101 1108 10827394 10.1067/mhj.2000.105105
Kampmann C Wiethoff CM Perrot A Beck M Dietz R Osterziel KJ The heart in Anderson Fabry disease Z Kardiol 2002 91 786 795 12395219 10.1007/s00392-002-0848-5
Doi Y Toda G Yano K Sisters with atypical Fabry's disease with complete atrioventricular block Heart 2003 89 e2 12482812 10.1136/heart.89.1.e2
Dimitrow PP Transthoracic Doppler echocardiography- noninvasive diagnostic window for coronary flow reserve assessment Cardiovasc Ultrasound 2003 1 4 12740038 10.1186/1476-7120-1-4
Krzanowski M Bodzon W Dimitrow PP Imaging of all three coronary arteries by transthoracic echocardiography: an illustrated guide Cardiovasc Ultrasound 2003 1 16 14622441 10.1186/1476-7120-1-16
Rigo F Coronary flow reserve in stress-echo lab. From pathophysiologic toy to diagnostic tool Cardiovasc Ultrasound 2005 3 8 15792499 10.1186/1476-7120-3-8
Watanabe N Akasaka T Yamaura Y Akiyama M Kaji S Saito Y Yoshida K Intramyocardial coronary flow characteristics in patients with hypertrophic cardiomyopathy: non-invasive assessment by transthoracic Doppler echocardiography Heart 2003 89 657 658 12748226 10.1136/heart.89.6.657
Sachdev B Takenaka T Teraguci H Tei C Lee P McKenna WJ Elliott P Prevalence of Anderson-Fabry disease in male patients with late onset of hypertrophic cardiomyopathy Circulation 2002 105 1407 1411 11914245 10.1161/01.CIR.0000012626.81324.38
Chimenti C Pieroni M Morgante E Antuzzi D Russo A Russo MA Maseri A Frustaci A Prevalence of Fabry disease in female patients with late-onset hypertrophic cardiomyopathy Circulation 2004 110 1047 1053 15313943 10.1161/01.CIR.0000139847.74101.03
Weidemann F Breunig F Beer M Sandstede J Turschner O Voelker W Ertl G Knoll A Wanner C Strotmann JM Improvement of cardiac function during enzyme replacement therapy in patients with Fabry disease: a prospective strain rate imaging study Circulation 2003 108 1299 1301 12952834 10.1161/01.CIR.0000091253.71282.04
Nagueh SF Fabry disease Heart 2003 89 819 820 12860841 10.1136/heart.89.8.819
Pieroni M Chimenti C Ricci R Sale P Russo MA Frustaci A Early detection of Fabry cardiomyopathy by tissue Doppler imaging Circulation 2003 107 1978 1984 12668521 10.1161/01.CIR.0000061952.27445.A0
Moon JC Sachdev B Elkington AG McKenna WJ Mehta A Pennell DJ Leed PJ Elliott PM Gadolinium enhanced cardiovascular magnetic resonance in Anderson-Fabry disease. Evidence for a disease specific abnormality of the myocardial interstitium Eur Heart J 2003 24 2151 2155 14643276 10.1016/j.ehj.2003.09.017
| 15857518 | PMC1097744 | CC BY | 2021-01-04 16:38:32 | no | Cardiovasc Ultrasound. 2005 Apr 27; 3:11 | utf-8 | Cardiovasc Ultrasound | 2,005 | 10.1186/1476-7120-3-11 | oa_comm |
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-121585751910.1186/1476-7120-3-12ResearchLack of association between Chlamydia Pneumoniae serology and endothelial dysfunction of coronary arteries Ferrari Markus [email protected] Gerald S [email protected] Barbara M [email protected] Albrecht [email protected] Eberhard [email protected] Hans R [email protected] Clinic of Internal Medicine I, Friedrich-Schiller-University, D – 07740 Jena, Germany2 Institute of Microbiology, Friedrich-Schiller-University, D – 07740 Jena, Germany2005 27 4 2005 3 12 12 8 4 2005 27 4 2005 Copyright © 2005 Ferrari et al; licensee BioMed Central Ltd.2005Ferrari 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
Recent publications brought up the hypothesis that an infection with Chlamydia Pneumoniae (CP) might be a major cause of coronary artery disease (CAD). Therefore, we investigated whether endothelial dysfunction (ED) as a precursor of atherosclerosis might be detectable in patients with previous infection with CP but without angiographic evidence of CAD.
Methods
We included 16 patients (6 male / 10 female) of 52 consecutive patients with normal coronary angiography who had typical angina pectoris and pathologic findings in the stress test. Exclusion criteria were: active smoker, elevated cholesterol, hypertension, age > 65 years, diabetes mellitus, treatment with ACE-inhibitors, or known CAD. Blood sample analysis for serum titer against CP (aCP-IgG) was performed after coronary angiography. We looked for endothelial dysfunction analyzing the diameter of the left anterior descending coronary artery (LAD) before and after acetylcholine (ACh) i. c. Quantitative analysis of luminal diameter (LD) was performed in at least two planes during baseline conditions and after ACh for 2 minutes in dosages of 7.2 μg/min and 36 μg/min with an infusion speed of 2 ml/min. Using Doppler guide wire, the coronary flow velocity was measured continuously in the LAD. The coronary flow velocity reserve (CFVR) was measured after 20 μg adenosine i. c.
Results
10 patients had an elevated aCP-IgG (> 1:8). 6 patients with negative titers (aCP-IgG ≤ 1:8) served as control (CTRL). Both groups were comparable in age, gender, angina class, results of non-invasive stress-test and the baseline values of LD and flow. In the CP positive group 3 patients (30%) did not show an increase of LD after ACh as evidence of ED. In the CTRL group 4 patients (67 %) had ED. There was no association between aCP-IgG and changes of coronary blood flow after ACh. All patients showed normal CFVR (3.0 ± 0.27) irrespective of their aCP-IgG values.
Conclusion
In patients with typical symptoms of coronary ischemia but without angiographically visible CAD and absence of other factors affecting the endothelial function, a previous infection with CP is not associated with endothelial dysfunction.
==== Body
Introduction
If coronary angiography is carried out due to pathological stress-test or angina pectoris, 10% to 20% of the patients do not reveal any atherosclerotic alteration of the coronary vessel related to the clinical symptoms [1]. As a possible explanation of this phenomenon, an infectious mechanism was discussed, which leads to an endothelial dysfunction (ED) and thus functional impairment of the coronary circulation [2,3].
The intracellular bacterial pathogen Chlamydia pneumoniae (CP) causes respiratory tract infections of increasing incidence with age [4]. The proof of CP both in atherosclerotic coronary vessels at post-mortem examinations and also in tissue samples from coronary atherectomy brought up the hypothesis that an infection with CP is also an important promoter of atherosclerosis and CAD [5-7]. In addition, the successful treatment with antibiotics of patients suffering from unstable angina pectoris supported this hypothesis [8,9]. CP was accused of damaging the coronary endothelial cells and therefore causing a local inflammatory reaction and promoting the sub-endothelial storage of low density lipoprotein (LDL) cholesterol [10-12].
The ED can be regarded as an early form of CAD before the detection of angiographically visible alterations caused by storage of cholesterol in the vessel wall [13]. The lack of dilatation of the coronary vessels during infusion of acetylcholine (ACh) uncovers an ED in vivo. This method of inducing a paradoxical reaction of the artery was well evaluated in patients suffering from diabetes mellitus, hypercholesterolemia, obesity, hypertension, or CAD, and in smokers [14-17]. Under the medication of ACE- inhibitors an ED can be attenuated [18].
Therefore, we designed this study including only patients without any known factor which could influence the endothelial function except previous CP infection indicated by elevated antibodies. It was our aim to prove whether there is an association between an infection with CP and an ED in those patients who did not carry any of the known risk factors. A positive result would corroborate the hypothesis of a causal role of CP in atherogenesis.
Methods
Patient selection
All patients had to give written informed consent. The study was performed with approval of the local ethical committee of our university. Out of 1144 consecutive patients who were brought to an elective coronary angiography because of typical angina pectoris or a pathological stress test for the first time, 52 caucasian patients who fulfilled the following criteria were screened for this study during a period of 22 months. Exclusion criteria were myocardial infarction, unstable angina, ECG abnormalities at rest, disorders of wall motion or thickened left ventricular wall in echocardiography, vitiae of the valves, age ≥ 65 years, arterial hypertension (systolic blood pressure at rest ≥ 140 mmHg), any type of diabetes mellitus, obesity (body mass index ≥ 30), hypercholesterinemia (total cholesterol ≥ 5.0 mmol/l or LDL cholesterol ≥ 3.0 mmol/l), hypetriglyceridemia (triglycerides ≥ 5.0 mmol/l), and a history of smoking during the last 10 years. An acute infection represented by fever, elevated C-reactive protein (CRP > 5 mg/dl), or elevated white blood cell count (≥ 10,000 / μl) caused exclusion from the study before coronary angiography. We also excluded all patients who were on medication with ACE inhibitors, anti-hypertensive drugs, nitrates, hormones, or lipid lowering agents. The patients gave their written consent for the study 24 hours before they were transferred to the catheter laboratory.
Study protocol
Coronary angiography was performed in standard fashion via femoral approach with a 5F diagnostic catheter. We saw atherosclerotic lesions of at least one coronary artery in 34 patients (65.4%). They were excluded and an angioplasty was carried out where necessary. The following measurements were performed in the remaining 18 patients (34.6%) in whom we did not find any alterations of the coronary arteries (exclusion of atherosclerotic plaque in all projections).
After intra coronary administration of 10.000 IE heparin, we placed a 0.014 inches Doppler guide-wire (Flowire™, Cardiometrics, USA) in the mid LAD. The tip of the wire was positioned in a straight vessel segment at least 1 cm from a departure of any relevant side branch. In 2 patients we stopped the testing due to complications: one patient showed a coronary spasm after insertion of the Doppler wire, another patient developed a bradycardia of 35 beats / min and a pressure drop immediately after starting the ACh infusion. After removal of the wire and termination of the ACh infusion, these complications were completely reversible within a few seconds. However, these two patients were excluded from further examinations and from the study. In the remaining 16 patients (30.8 %) the measurements according to the study protocol were carried out without any complications.
The luminal diameter (LD) was measured 5 mm distal to the tip of the Doppler wire in two orthogonal planes by means of automatic quantitative coronary angiography (QCA). Automatic contour detection was performed with a geometric edge differentiation technique [19]. For prevention of a study bias, the automated QCA was performed by an independent cardiologist without knowledge of the antibody titer against Chlamydia pneumoniae. Each coronary angiography was performed in the same projection as the baseline procedure.
After recording a stable Doppler baseline signal, we infused acetylcholine (ACh) (Miochol™; CIBA Vision, Switzerland) at a dose of 7.2 μg/min through the catheter into the LAD for two minutes. Assuming a blood flow in the LAD of 80 ml/min, we estimated a final ACh concentration in the coronary bed of 1/2 × 10-6 mol/l (= dose 1, ACh D1). Careful attention was paid to the calculation of catheter dead space to ensure accurate delivery of acetylcholine to the coronary ostium. Subsequently, coronary angiography was carried out in the same projection as at the baseline. The ACh infusion was stopped for at least 3 minutes before proceeding. After reaching stable baseline conditions, i. c. infusion of 2 ml/min of ACh in a concentration of 36 μg/min aiming at a concentration of 1/4 × 10-5 mol/l (= dose 2, ACh D2) was started for another 2 minutes. Angiography was repeated immediately after each infusion.
After another recovery period of 3 minutes, we gave a continuous infusion of saline (NaCl 0.9 %) with an infusion rate of 2 ml/min. After restoring the baseline following the angiography, we injected 0.2 mg of nitrotriglycerin (NTG) i. c. and performed a final coronary angiography in the same projections as before.
The average coronary peak flow velocity (APV) was registered continuously by means of a Doppler wire during the complete examination. Attention was paid to insure that APV had returned to baseline values between each measurement. At the end of the examination we injected 20 μg adenosine i.c. recording the maximum rise of APV in comparison to the baseline value which yields the coronary flow velocity reserve (CFVR). The CFVR measurement was carried out at least twice, and the mean value was calculated for further analysis.
Calculation of coronary blood flow (CBF)
Coronary blood flow (CBF) was calculated out of the average peak flow velocity (APV) measured by Doppler guide-wire in the mid LAD. By offline analysis with QCA we calculated the average LD distal to the tip of the Doppler wire. Multiplying the square of 0.5 × LD with π we obtained the luminal area (LA) of the vessel at the site of measurement. LA × APV × 0.5 results in the CBF. This method of calculation of CBF has been evaluated in vitro and in vivo [20,21].
Serological analysis of antigenic titers against Chlamydia pneumoniae
Following the catheter examination we took 10 ml of blood for the antibody determination against CP. Testing for anti-CP antibody (IgG) we used a standard micro-immuno-fluorescence (MIF) test on consecutive dilution rows. We used the MIF serology with the antigens Chlamydia pneumoniae (TW183) antigen and Chlamydia trachomatis (type D) in order to exclude cross reacting antibodies. The serum tests were performed in a double-blinded setting. The researcher who performed the serological analysis did not receive any knowledge of the results of the coronary angiography and the flow measurements and vice versa. The patients were retrospectively allotted in two groups according to their aCP-IgG values: Those patients with an aCP-IgG ≤ 1:8 formed the group of CP negative patients, individuals with aCP-IgG > 1:8 were included in the CP positive group since they were supposed to suffer from a previous infection with CP. The results of the Chlamydia trachomatis MIF test were considered in this grouping.
Statistical analysis
All values are expressed as mean ± standard deviation (SD). P values were calculated using a two-tailed student's t test for statistical analysis of continuous variables by Excel™ (Mircosoft Co., USA). Multivariate logistic regression analysis including aCP-IgG, sex and age as risk factors of ED and two-tailed exact Fisher's test were performed using SPSS™ (SPSS Inc., USA). A p value ≤ 0.05 was considered statistically significant.
Results
The analysis of the serum probes showed increased aCP-IgG values (>1:8) in 10 patients (62.5%). They formed the subgroup of the CP positive individuals. The remaining 6 patients (37.5%) had negative aCP-IgG titers (≤ 1:8) and served as the control group. All 16 patients in the study were negative (≤ 1:8) against Chlamydia trachomatis (type D) in the MIF test.
The baseline data of all patients are summarized in table 1. The results of the testing in the catheter laboratory are listed in table 3. Coronary blood flow velocity was 18.9 ± 4.49 cm/sec in our study population. The APV increased to 35.0 ± 12.08 cm/sec during middle dosage and 42.1 ± 11.49 cm/sec during the infusion of the high dosage of ACh. After the bolus injection of 0.2 mg NTG, we measured an APV of 27.7 ± 8.22 cm/sec. There was no significant difference in the APV values between the 2 groups.
Table 1 Baseline characteristics of study patients, and number of patients with pathologic findings in bicycle or scintigraphy stress-test.
All patients CP positive CP negative
Number 16 10 (62.5 %) 6 (37.5 %)
Age (years) 50.7 ± 7.02 53.2 ± 6.48 47.0 ± 6.08
Sex (male) 6 male (37.5 %) 4 male (40 %) 2 male (33.3 %)
Angina class (CCS) 2.4 ± 0.48 2.4 ± 0.49 2.3 ± 0.47
Pathologic ECG in bicycle stress-test 12 (75 %) 7 (70 %) 5 (83.3 %)
Pathologic scintigraphy 5 (31.3 %) 3 (30 %) 2 (33.3 %)
Total Cholesterol levels 4.0 ± 0.58 mmol/l
154.1 ± 22.41 mg% 4.1 ± 0.68 mmol/l
157.5 ± 26.48 mg% 3.8 ± 0.28 mmol/l
148.4 ± 10.87 mg%
Systolic blood pressure 121 ± 12.0 mmHg 120 ± 11.7 mmHg 122 ± 12.4 mmHg
Smoking history (last 10 years) All negative All negative All negative
Total Cholesterol levels were below 200 mg% (5.2 mmol/l) in all patients, the systolic blood pressure did not exceed 139 mmHg in any case.
CP: Chlamydia pneumoniae {values are mean ± SD, % of each group}.
Table 3 Coronary blood flow and vessel diameter.
All patients CP positive p CP negative
CBF (ml / min) (+/- percentchange)
baseline 22.6 ± 14.39 22.0 ± 13.80 → 0.503 ← 23.4 ± 15.28
ACh D1 39.6 ± 21.21 (+75.2%) 40.4 ± 22.34 (+83.6%) → 0.172 ← 38.5 ± 19.33 (+64.5%)
ACh D2 48.7 ± 30.34 (+115.5%) 51.6 ± 27.82 (+134.5%) → 0.177 ← 41.3 ± 34.78 (+76.5%)
NaCl (control) 23.1 ± 12.15 (+2.2%) 22.1 ± 12.38 (+0.4%) → 0.697 ← 24.7 ± 11.57 (+5.5%)
NTG 40.3 ± 23.06 (+78.3%) 39.5 ± 20.74 (+79.5%) → 0.696 ← 41.0 ± 24.74 (+75.2%)
Vessel diameter (mm) (+/- percent change)
baseline 2.53 ± 0.50 2.55 ± 0.50 → 0.773 ← 2.51 ± 0.48
ACh D1 2.55 ± 0.50 (+0.8%) 2.56 ± 0.52 (+0.4%) → 0.658 ← 2.54 ± 0.46 (+1.2%)
ACh D2 2.54 ± 0.61 (+0.3%) 2.60 ± 0.62 (+2.0%) → 0.289 ← 2.46 ± 0.58 (-2.0%)
NaCl (control) 2.54 ± 0.45 (+0.1%) 2.52 ± 0.48 (-1.1%) → 0.543 ← 2.58 ± 0.39 (+2.7%)
NTG 2.88 ± 0.52 (+13.8%) 2.88 ± 0.47 (+13.0%) → 0.801 ← 2.89 ± 0.60 (+15.1%)
CFVR 3.0 ± 0.27 3.1 ± 0.31 → 0.652 ← 3.0 ± 0.10
Coronary blood flow (CBF) and vessel diameter before (baseline) and after i. c. infusion of acetylcholine (ACh) at medium dose (ACh D1: 7.2 μg/min → 1/2 × 10-6 mol/l) and high dose (ACh D2: 36 μg/min → 1/4 × 10-5 mol/l), 0.2 mg nitroglycerin (NTG), continuous infusion of saline (NaCl 0.9 % at 2 ml/min) served as control. The comparison of the CBF, and of the vessel diameter with the corresponding baseline value is expressed as percent change. CFVR: coronary flow velocity reserve (adenosine i. c. versus baseline), CP: Chlamydia pneumoniae. {values are mean ± SD, p values of two-tailed student's t-test}
Among the negative control group without previous contact with CP, four patients (66.7%) did not show any increase of luminal diameter (LD) during ACh infusion, what was judged as an ED (Figure 1). Three patients (30%) of the CP positive subgroup did not show an increase of the LD (Figure 2). The two-tailed exact Fisher's test did not unveil any significant difference between the two groups (p = 0.302) as shown in table 2. For the final measurement of the CFVR, we recorded a 2.5 to 3.5 fold increase of the APV (3.0 ± 0.27) after i.c. injection of 20 μg adenosine. There was no association between the occurrence of an ED and the height of the CFVR (p = 0.258) or between the aCP-IgG and the CFVR values (p = 0.735). We did not find any correlation between CFVR and the flow velocity increase during high dose ACh infusion (r = -0.19).
Figure 1 Decreased vessel diameter (62% of baseline value) during acetylcholine (36 μg/min) i. c. Continuous recording of flow velocity in this Chlamydia pneumoniae positive patient after i. c. ACh infusion (ACh D1: 7.2 μg/min, ACh D2: 36 μg/min).
Figure 2 Vessel diameter of all 16 patients at baseline, during i. c. infusion of acetylcholine (ACh) with 7.2 μg/min (ACh D1), and with 36 μg/min (ACh D2), during NaCl (0.9% 2 ml/min i. c.), and after 0.2 mg nitrotriglycerin (NTG) i. c. Patients without an elevated IgG serum titer against Chlamydia pneumoniae (CP negative) are presented by dotted lines.
Discussion
In this prospective study we investigated whether a previous infection with Chlamydia pneumonia is associated with an endothelial dysfunction of the coronary vessels before the presence of atherosclerosis. It was our intention to exclude all patients with known risk factors for atherosclerosis and endothelial dysfunction (e.g. hypertension, hyperlipidemia, smoking, and diabetes). In addition, those patients who took medication which could have influenced the endothelial function were excluded. Sixteen patients without any angiographically visible atherosclerotic alteration but typical signs of coronary ischemia at stress were included in the study. We evaluated the capability of the coronary endothelium to promote a vessel dilatation in this highly selected group of patients. A lack of flow mediated diameter increase of ≥ 5% was interpreted as an ED. Since we did not find a higher incidence of ED in those patients who had a previous CP infection than in the patients in the CP negative control group, we disproved our hypothesis that CP induces an endothelial damage or even disturbs the endothelial function before visible atherosclerotic changes take place.
We observed a relatively large proportion of so-called "exclusion of CAD" among the 52 patients who were screened according to the strict inclusion criteria. Previous angiography studies reported a prevalence of negative coronary angiograms of 10% to 20% in comparable populations [1]. In contrast to predominantly male patient populations in catheter laboratories, we viewed two thirds of female individuals in our study collective. Women show negative coronary angiograms more often than men [22,23]. Furthermore, our patients were relatively young which may also explain this relatively large proportion of negative coronary angiograms. On the other hand, all patients suffered from typical symptoms of coronary ischemia such as angina pectoris, ECG changes, or pathologic findings in non-invasive stress test. Another limitation of our study is the relatively low sample size due to difficulties in recruiting patients. We screened 1144 patients for the study, but only 18 patients fulfilled the strict inclusion criteria.
The CP positive group and the CP negative group were comparable in terms of age, gender, angina class, and non-invasive stress-test results. The tendency of a higher proportion of women to be CP positive was statistically insignificant (p: 0.172). However, there was no significant difference between the two groups regarding vessel diameter and flow velocity at the baseline and after administration of nitroglycerin. The average baseline APV of 18.9 ± 4.49 cm/s with an increase to 27.7 ± 8.22 cm/s is consistent with values reported in the literature [24]. During ACh infusion, we even recorded a higher increase of LD in patients with elevated aCP-IgG compared to the CP negative individuals, but there was no significant difference here either (Figure 3). We observed a lack of vasodilatation during ACh in 67 % of the CP negative and 30 % of the CP positive patients (p 0.302). Therefore, we were not able to verify the hypothesis that a previous infection with CP leads to a higher incidence of endothelial dysfunction among patients with normal coronary arteries.
The increase of APV to 300 % after adenosine was recorded among all patients irrespective of the presence of ED or the CP titer status. An average CFVR of 3.0 ± 0.27 indicates normal coronary blood flow regulation of the arteriolar coronary vessels of the study patients [24]. We conclude that a previous infection with CP does not influence the dilatory capacity of the arteriolar bed of the coronary circulation represented by the CFVR.
Non of our patients showed any signs of atherosclerotic lesions of a coronary vessel in the angiogram. Intra-vascular ultrasound would have been more sensitive to uncover changes of the integrity of the vessel wall in some patients despite a negative coronary angiogram [25]. However, the endothelial function tested with ACh is one of the most sensitive methods for detecting early damage of the endo-vascular structure on the basis of physiologic function [26].
In patients suffering from CAD, the presence of ED predicts an enlarged cardiovascular event rate [27]. Since atherosclerosis is considered to be a chronic inflammation of the artery vessel wall, it should be the interaction of CP with cells of the vasculature that can result in a local inflammatory response. It was reported that CP positive patients are in danger of having a higher cardiovascular event rate [8]. Furthermore, supporting the hypothesis of amplification of the progression of atherosclerosis by misbalance in the immune response, a correlation between high levels of aCP-IgG and acute coronary syndrome was demonstrated recently in a study on 830 patients [28]. Chlamydia pneumoniae infected patients with coronary artery sclerosis showed higher re-stenosis rates after a balloon angioplasty, but not after a coronary stent implantation [29]. However, we did not uncover an increased number of patients with ED among the individuals with CP antibodies. Therefore, we postulate that CP is not envolved in the early atherosclerotic cascade. The persistent CP infection of immune cells (T-cells, monocytes and macrophages) and non-immune cells (endothelial cells and smooth muscle cells) may contribute to a cascade of inflammatory mediators leading to an enhanced tissue remodeling of the arterial intima and instability of atherosclerotic plaques [30-33]. We would therefore conclude that a chronic infection with CP is not a major cause of coronary artery disease. The patho-physiologic link between CP infection and arteriosclerosis I still under discussion [34]. Recently published trials including more than 8,000 patients did not show any evidence for the hypothesis that coronary artery disease is caused by an infection with CP and antibiotic therapy is indicated [35,36]. The present model includes the involvement of CP in the secondary phase of atherogenesis including inflammation, fibrous plaque formation, plaque rupture, and thrombosis. Recent meta-analysis questioned any relevance of a CP infection in the early course of CAD [37].
Conclusion
To our understanding of the mechanism of CP involvement in atherogenesis, we interpret the results of our study in combination with the present knowledge about CP as favoring a possible contributory rather than a causal role. However, there is need for further studies to enlighten the early pathomechanisms of chlamydial infection of the endothelium, including signaling pathways in the genesis and the progression of coronary artery disease.
Figure 3 Relative changes in coronary blood flow (+ / - standard error of mean) The dotted line presents patients with previous infection with Chlamydia pneumoniae (CP). The bold line shows the results in the CP negative patients. The curves represent the relative changes compared to baseline during i. c. infusion of acetylcholine (ACh) with 7.2 μg/min (ACh D1), with 36 μg/min (ACh D2), and during NaCl (0.9% 2 ml/min i. c.)
Table 2 Changes of vessel diameter after i. c. infusion of acetylcholine (ACh D2: 36 μg/min) compared with baseline diameter.
n = 16 CP IgG < 1:16 (CP negative) CP IgG ≥ 1:16 (CP positive)
Increase of vessel diameter 2 (33 %) 7 (70 %)
No increase of vessel diameter 4 (67 %) 3 (30 %)
The two groups were allotted according to their Chlamydia pneumoniae (CP) IgG serum titer. Two-tailed exact Fisher's test did not show any significant difference between both groups (p = 0.302).
==== Refs
Cannon RO 3rdQuyyumi AA Mincemoyer R Stine AM Gracely RH Smith WB Geraci MF Black BC Uhde TW Waclawiw MA Maher K Benjamin SB Imipramine in patients with chest pain despite normal coronary angiograms N Engl J Med 1994 330 1411 1417 8159194 10.1056/NEJM199405193302003
Mehta JL Saldeen TG Rand K Interactive role of infection, inflammation and traditional risk factors in atherosclerosis and coronary artery disease J Am Coll Cardiol 1998 31 1217 1225 9581711 10.1016/S0735-1097(98)00093-X
Noll G Pathogenesis of atherosclerosis: a possible relation to infection Atherosclerosis 1998 140 3 9 10.1016/S0021-9150(98)00113-0
Persson K Epidemiological and clinical aspects on infections due to Chlamydia pneumoniae (strain TWAR) Scand J Infect Dis Suppl 1990 69 63 67 2263898
Juvonen J Laurila A Juvonen T Alakarppa H Surcel HM Lounatmaa K Kuusisto J Saikku P Detection of Chlamydia pneumoniae in human nonrheumatic stenotic aortic valves J Am Coll Cardiol 1997 29 1054 1059 9120159 10.1016/S0735-1097(97)00003-X
Thomas M Wong Y Thomas D Ajaz M Tsang V Gallagher PJ Ward ME Relation between direct detection of Chlamydia pneumoniae DNA in human coronary arteries at postmortem examination and histological severity (Stary grading) of associated atherosclerotic plaque Circulation 1999 99 2733 2736 10351965
Hayashida K Tanaka M Morita H Hayashi F Inada T Suzuki H Sakamoto T Katsuragawa M Hibino H Kambara H Chlamydia pneumoniae seropositivity predicts the risk of restenosis after percutaneous transluminal coronary angioplasty Heart Vessels 2002 16 137 145 12224784 10.1007/s003800200010
Gupta S Leatham EW Carrington D Mendall MA Kaski JC Camm AJ Elevated Chlamydia pneumoniae antibodies, cardiovascular events, and azithromycin in male survivors of myocardial infarction Circulation 1997 96 404 407 9244203
Gurfinkel E Bozovich G Daroca A Beck E Mautner B Randomised trial of roxithromycin in non-Q-wave coronary syndromes: ROXIS Pilot Study. ROXIS Study Group Lancet 1997 350 404 407 9259655 10.1016/S0140-6736(97)07201-2
Kontula K Vuorio A Turtola H Saikku P Association of seropositivity for Chlamydia pneumoniae and coronary artery disease in heterozygous familial hypercholesterolaemia Lancet 1999 354 46 47 10406370 10.1016/S0140-6736(99)01691-8
Hu H Pierce GN Zhong G The atherogenic effects of chlamydia are dependent on serum cholesterol and specific to Chlamydia pneumoniae J Clin Invest 1999 103 747 753 10074493
Bhakdi S Dorweiler B Kirchmann R Torzewski J Weise E Tranum-Jensen J Walev I Wieland E On the pathogenesis of atherosclerosis: enzymatic transformation of human low density lipoprotein to an atherogenic moiety J Exp Med 1995 182 1959 1971 7500042 10.1084/jem.182.6.1959
Cohen RA Zitnay KM Haudenschild CC Cunningham LD Loss of selective endothelial cell vasoactive functions caused by hypercholesterolemia in pig coronary arteries Circ Res 1988 63 903 910 2460269
Al Suwaidi J Higano ST Holmes R JrLennon R Lerman A Obesity is independently associated with coronary endothelial dysfunction in patients with normal or mildly diseased coronary arteries J Am Coll Cardiol 2001 37 1523 1528 11345360 10.1016/S0735-1097(01)01212-8
Ludmer PL Selwyn AP Shook TL Wayne RR Mudge GH Alexander RW Ganz P Paradoxical vasoconstriction induced by acetylcholine in atherosclerotic coronary arteries N Engl J Med 1986 315 1046 1051 3093861
Zeiher AM Drexler H Wollschlager H Just H Modulation of coronary vasomotor tone in humans. Progressive endothelial dysfunction with different early stages of coronary atherosclerosis Circulation 1991 83 391 401 1991363
Vita JA Yeung AC Winniford M Hodgson JM Treasure CB Klein JL Werns S Kern M Plotkin D Shih WJ Mitchel Y Ganz P Effect of cholesterol-lowering therapy on coronary endothelial vasomotor function in patients with coronary artery disease Circulation 2000 102 846 851 10952951
Mancini GB Henry GC Macaya C O'Neill BJ Pucillo AL Carere RG Wargovich TJ Mudra H Luscher TF Klibaner MI Haber HE Uprichard AC Pepine CJ Pitt B Angiotensin-converting enzyme inhibition with quinapril improves endothelial vasomotor dysfunction in patients with coronary artery disease. The TREND (Trial on Reversing ENdothelial Dysfunction) Study Circulation 1996 94 258 265 8759064
Reiber JH Serruys PW Kooijman CJ Wijns W Slager CJ Gerbrands JJ Schuurbiers JC den Boer A Hugenholtz PG Assessment of short-, medium-, and long-term variations in arterial dimensions from computer-assisted quantitation of coronary cineangiograms Circulation 1985 71 280 288 3965172
Labovitz AJ Anthonis DM Cravens TL Kern MJ Validation of volumetric flow measurements by means of a Doppler-tipped coronary angioplasty guide wire Am Heart J 1993 126 1456 1461 8249803 10.1016/0002-8703(93)90545-K
Ferrari M Andreas S Werner GS Wicke J Kreuzer H Figulla HR Evaluation of an active coronary perfusion balloon device using Doppler flow wire during PTCA Cathet Cardiovasc Diagn 1997 42 84 89 9286550 10.1002/(SICI)1097-0304(199709)42:1<84::AID-CCD24>3.0.CO;2-L
Pasternak RC Thibault GE Savoia M DeSanctis RW Hutter AM Jr Chest pain with angiographically insignificant coronary arterial obstruction. Clinical presentation and long-term follow-up Am J Med 1980 68 813 817 7386488 10.1016/0002-9343(80)90199-0
Rutledge T Reis SE Olson M Owens J Kelsey SF Pepine CJ Reichek N Rogers WJ Merz CN Sopko G Cornell CE Sharaf B Matthews KA Women's Ischemia Syndrome E History of anxiety disorders is associated with a decreased likelihood of angiographic coronary artery disease in women with chest pain: the WISE study J Am Coll Cardiol 2001 37 780 785 11693752 10.1016/S0735-1097(00)01163-3
Kern MJ Deligonul U Tatineni S Serota H Aguirre F Hilton TC Intravenous adenosine: continuous infusion and low dose bolus administration for determination of coronary vasodilator reserve in patients with and without coronary artery disease J Am Coll Cardiol 1991 18 718 729 1869735
Nissen SE Yock P Intravascular ultrasound: novel pathophysiological insights and current clinical applications Circulation 2001 103 604 616 11157729
Werns SW Walton JA Hsia HH Nabel EG Sanz ML Pitt B Evidence of endothelial dysfunction in angiographically normal coronary arteries of patients with coronary artery disease Circulation 1989 79 287 291 2914347
Al Suwaidi J Hamasaki S Higano ST Nishimura RA Holmes DR JrLerman A Long-term follow-up of patients with mild coronary artery disease and endothelial dysfunction Circulation 2000 101 948 954 10704159
Chandra HR Choudhary N O'Neill C Boura J Timmis GC O'Neill WW Chlamydia pneumoniae exposure and inflammatory markers in acute coronary syndrome (CIMACS) Am J Cardiol 2001 88 214 218 11472696 10.1016/S0002-9149(01)01628-9
Tanaka T Matsushita M Oka Y Sada T Kira Y Effect of Chlamydia pneumoniae infection on coronary flow reserve and intimal hyperplasia after stent implantation in patients with angina pectoris J Cardiol 2001 38 311 317 11806088
Kaul R Uphoff J Wiedeman J Yadlapalli S Wenman WM Detection of Chlamydia pneumoniae DNA in CD3+ lymphocytes from healthy blood donors and patients with coronary artery disease Circulation 2000 102 2341 2346 11067786
Roivainen M Viik-Kajander M Palosuo T Toivanen P Leinonen M Saikku P Tenkanen L Manninen V Hovi T Manttari M Infections, inflammation, and the risk of coronary heart disease Circulation 2000 101 252 257 10645920
Caligiuri G Paulsson G Nicoletti A Maseri A Hansson GK Evidence for antigen-driven T-cell response in unstable angina Circulation 2000 102 1114 1119 10973839
Dechend R Maass M Gieffers J Dietz R Scheidereit C Leutz A Gulba DC Chlamydia pneumoniae infection of vascular smooth muscle and endothelial cells activates NF-kappaB and induces tissue factor and PAI-1 expression: a potential link to accelerated arteriosclerosis Circulation 1999 100 1369 1373 10500035
Danesh J Antibiotics in the prevention of heart attacks Lancet 2005 365 365 367 15680439
Muhlestein JB Antibiotic treatment of atherosclerosis Curr Opin Lipidol 2003 14 605 614 14624138 10.1097/00041433-200312000-00009
Ridker PM Cannon CP Morrow D Rifai N Rose LM McCabe CH Pfeffer MA Braunwald E C-reactive protein levels and outcomes after statin therapy N Engl J Med 2005 352 20 28 15635109 10.1056/NEJMoa042378
Danesh J Whincup P Walker M Lennon L Thomson A Appleby P Wong Y Bernardes-Silva M Ward M Chlamydia pneumoniae IgG titres and coronary heart disease: prospective study and meta-analysis BMJ 2000 321 208 213 10903653 10.1136/bmj.321.7255.208
| 15857519 | PMC1097745 | CC BY | 2021-01-04 16:38:32 | no | Cardiovasc Ultrasound. 2005 Apr 27; 3:12 | utf-8 | Cardiovasc Ultrasound | 2,005 | 10.1186/1476-7120-3-12 | oa_comm |
==== Front
Cell Commun SignalCell communication and signaling : CCS1478-811XBioMed Central London 1478-811X-3-71583678510.1186/1478-811X-3-7CommentaryRecognizing scientific excellence in the biology of cell adhesion Wary Kishore K [email protected] Center for Extracellular Matrix Biology, Institute of Biosciences and Technology, Texas A & M University System-Health Science Center, Texas Medical Center, 2121 W. Holcombe Blvd. Houston, TX-77030, USA2005 18 4 2005 3 7 7 8 3 2005 18 4 2005 Copyright © 2005 Wary; licensee BioMed Central Ltd.2005Wary; 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 prestigious 2005 Japan Prize for Cell Biology has been awarded to Dr. Masatoshi Takeichi, Director of RIKEN Developmental Biology, Kobe, Japan, and Dr. Erkki Ruoslahti, Distinguished Professor, The Burnham Institute, La Jolla, USA for their "Fundamental contribution in elucidating the molecular mechanisms of cell adhesion". The award is scheduled to be presented to the scientists in ceremonies in Tokyo on April 20, 2005 as part of a week-long celebration of "Japan Prize Week".
==== Body
What is cell adhesion?
Well, why does our skin look so smooth on the surface? How do skin cells adhere to each other and the underlying connective tissue to resist wound and bruise? How do two 'unlike' or 'like' cells live side-by-side? How are muscles and tendons glued to the bones? How do endothelial and epithelial cells are separated from each other? What mechanisms divide astrocytes, neurons, and the endothelial cells that make up the neurovascular unit? The answer is "cell adhesion", which is because of the characteristic properties of proteins and molecules that act like 'glue' or 'sticky molecules'. If cells or tissues do not hold each other, like in blistering skin in which something as gentle as a human touch can cause the skin to blister and peel away, inviting fatal infection and wound that may never heal. Suffice to say, the chances of survival will be somewhat diminished.
What are cell adhesion molecules?
In the late 1970's two ideas were put forward. First, the chemoaffinity hypothesis proposed that cell-cell contacts are mediated by unique set of cell adhesion molecules presented by adjacent cells. Second, adhesion molecules are limited, but their affinity could switch from low to high and vice versa. Soon afterwards, several important cell adhesion molecules were discovered and described including the cadherins, neuronal cell adhesion molecules (NCAM), extracellular matrix (ECM) molecules, proteoglycans, the immunoglobulin cell adhesion molecules, junctional adhesion molecules (JAMs), connexins, and selectins. Those ideas are very much alive and many cell adhesion molecules discovered recently are being tested with stringent criteria with better technologies today.
How do these molecules promote cell adhesion?
There may not be a unifying answer to that question. In one of the landmark articles, Dr. Masatoshi Takeichi [Fig. 1A] described calcium-dependent and -independent mechanisms of cell adhesion [1]. Cell-aggregation assays of disaggregated tissue and cells provided indication that the cadherins promote 'homophilic' interactions, a process that requires presence of Calcium metal ions [1-3]. Cadherins are transmembrane proteins containing an extracellular, a transmembrane, and a cytoplasmic segment. The extracellular domains of cadherins mediate Calcium-dependent intercellular adhesion by homophilic interactions. The binding properties and specificities of the adhesive interactions are located in the N-terminal segment of the molecules. A total of 17 classical cadherins have been described in the literature. Cadherin superfamily is made of 85 members. The classical cadherins are mainly involved in the cell adhesion. The roles of the other members of cadherin superfamily remain to be elucidated. Cell adhesions mediated by cadherins are cell type specific. In epithelial and endothelial cells, cadherins mediate formation of adherens junctions. It is now clear that the intracellular signaling components of cadherin determine the epithelial morphogenesis and tissue architectures [2-5]. The loss of cadherin expression by neoplastic cells is a hallmark of tumor progression [6]. Dr. Erkki Ruoslahti [Fig 1B] provided evidence that most ECM molecules such as fibronectin [7] promote both cell-cell and cell-matrix interaction by interacting with a family of cell adhesion receptor called the integrins [8]. In contrast to static ECM, some of the soluble ECM molecules can serve as a 'bridge' between two like or unlike cells [Fig. 2]. Such interactions are both transient as well as static, for example, at the sites of injury and inflammation, and these interactions could be low or high affinity [9]. The development of specific monoclonal antibodies such as (Ligand-induced binding site, LIBS, and cation- and ligand-induced binding site, CLIBS) as well as fluorescence energy transfer experiments provided further clues to the nature of the molecular interactions of integrin with the ECM molecules [9]. Moreover, molecular genetic analyses have provided evidence that multicellular organisms are dependent on adhesion of cells to each other and the ECM molecules, without which many cells will fail to stick [10]. Accordingly, gene deletion studies in mouse embryos have provided evidence that both cadherin and fibronectin molecules are required for embryonic development. The studies of cultured cells have provided early evidence that both fibronectin and cadherins help organize the cytoskeleton. In short, the prize is all about elucidating the molecular mechanisms as to how cell adhesion works [Fig. 2, 3].
Figure 1 A Dr. Masatoshi Takeichi (left), and (B) Dr. Erkki Ruoslahti (right). Image (A) is provided by Dr. Takeichi, and (B) obtained from public domain.
Figure 2 Schematics of cell adhesion mediated by cadherin and by extracellular matrix (ECM) proteins. Cadherin molecule connects adjacent cells by homophilic interactions in a metal ion dependent manner. Integrin cell adhesion receptors can interact with both static as well as soluble ECM ligands. In addition, integrins can also bind cell-associated ligands (not shown).
Figure 3 (A) Cell-cell adhesion- Epithelial cells were stained with anti-E-cadherin monoclonal antibody and detected by fluorescent dye microcopy. Green fluorescent represents E-cadherin molecules connecting cells. Red color represents nucleus. (B) Cell-matrix adhesion- Adherent endothelial cell was stained with anti-fibronectin monoclonal antibody and detected by fluorescent microscopy. Images are not in scale, magnification, 200×
Cell-aggregation and cell adhesion assays
Cell biologists use proteolytic enzymes such trypsin and Ethylenediaminetetra-acetic acid (EDTA) to detach/disaggregate cells from the culture dishes and to prepare of primary cells from intact tissues. Trypsin is a proteolytic enzyme, while EDTA, a metal ion chelator. When used in right combination, they can disrupt both cell-cell and cell-matrix interactions, that is to say these two substances can disaggregate cells and tissues [1]. Cell-cell and cell-matrix interactions appear to go hand-in-hand [Fig. 2]. Upon attachment adherent cells sense presence of Calcium in the environment, calcium is required for both cell-cell and cell-matrix interactions. Dr. Takeichi described as to how cell-cell interactions between "like" cells and "unlike" cells can be induced by E-cadherin molecules [2,3]. In normal adherent cells such as endothelial (vascular endothelial cadherin, called VE-Cadherin) and epithelial cells (epithelial cadherin, called E-cadherin) cadherins connect two or more cells in a "zipper" like fashion in presence of calcium [1-3,11]. Importantly, calcium prevents the degradation of cadherin and promotes cell adhesion activity [1]. Cadherin may also be important for mediating "contact-inhibition", a property of normal adherent cells. In a nutshell, Dr. Takeichi observed and described calcium-dependent and -independent mechanisms of cell adhesion [1].
Short-term incubation of adherent cells with trypsin can digest most ECM molecules and addition of EDTA helps disrupt interactions of cell adhesion mediated by integrins [8]. Dr. Ruoslahti described purification and characterization of fibronectin from blood plasma, and went onto identify the Arg-Gly-Asp (RGD) tri-peptide cell adhesion motif, that remains as a conceptual breakthrough [7,12]. Structurally, the fibronectin represents a prototypic ECM molecule that displays highly modular structure [13]. Fibronectin possesses repeating structural motifs, classified as fibronectin repeats FN-I, FN-II, and FN-III that are grouped together into functional domains [12,13]. Many cell types secrete fibronectin polypeptide ranging between 220–250 kDa sizes. Functionally, fibronectin plays critical roles in cell adhesion-dependent cellular activities both in development and adult tissue homeostasis, proliferation, and migration [12,13]. Cell adhesion assays provided evidence that cell attachment mediated by specific subset of integrins onto fibronectin can be blocked by RGD tri-peptides. Subsequently, many laboratories around the world have also cloned, characterized and described ECM proteins that also may contain RGD or variant cell binding sites. Detergent solubilized plasma-membrane proteins passed through a column chromatography conjugated to RGD-peptides allowed purification of RGD-binding cell adhesion receptors, which is now known as integrins [14]. For detail, please see the RGD story by E. Ruoslahti [14]. Secreted fibronectin can organize and assemble large protein complex by interacting with many other ECM molecules in the extracellular space such as the fibrinogen and collagens, glycosaminoglycans, proteoglycans, tenascin, fibulin and thrombospondin. Many normal cells have been described as "anchorage-dependent" cells that require attachment factor such as fibronectin, without which cells die of apoptosis [15,16]. Apart from just being a structural support, Fibronectin can trap or sequester growth factors and cytokines, and induce signaling activities [16,17]. In contrast, neoplastic cells that accumulate mutant copies of genetic materials no longer require fibronectin to grow, divide or metastasize, and this phenomena is called "anchorage-independence". Fibronectin also interacts with bacterial adhesion molecules, a pathological process that helps bacteria to colonize and infect host tissues [18].
Conclusion
In addition to providing structural support, both cadherins and fibronectin molecules are also required for cell polarity, and informing the cells and tissues about their position in time and space, called positional cues [17]. A biological process that allow cells to sense their immediate physical and chemical environment correctly, for example, to help cells sense presence of glucose and insulin, cytokines and growth hormones, signaling molecules and metal ions. However, such regulatory mechanisms could be altered in many pathological states including tumor growth, angiogenesis and metastasis. Complete understanding of the mechanisms of regulation of cell-adhesion molecules and their signaling activities remains an active area of investigation in many disease settings including cardiovascular, cancer, neural networks, damage and repair mechanisms associated with traumatic injury, wound healing, host-pathogen interactions, nanotechnology, tissue engineering, and molecular therapeutics. The Japan Prize 2005 and cash award is slated to be given in presence of the Emperor of Japan in a week-long celebration beginning 20th April, 2005 .
Authors' contributions
K.K.W. wrote and edited the whole manuscript.
Acknowledgements
The field of cell adhesion is somewhat matured. I apologize to my colleagues and authors whose works could not be cited here due to space constrains. KKW is a recipient of an award from American Heart Association (National Council) and a member of Mission Connect, TIRR, Houston, Texas, and a member of Cardiovascular Research Institute (CVRI) of the Texas A & M University-System Health Science Center.
==== Refs
Takeichi M Functional correlation between cell adhesive properties and some cell surface proteins J Cell Biol 1977 75 464 474 264120 10.1083/jcb.75.2.464
Takeichi M The cadherins: cell-cell adhesion molecules controlling animal morphogenesis Development 1988 102 639 655 3048970
Takeichi M Cadherin cell adhesion receptors as a morphogenetic regulator Science 1991 251 1451 1455 2006419
Gumbiner BM Epithelial morphogenesis Cell 1992 69 385 387 1581959 10.1016/0092-8674(92)90440-N
Gumbiner BM Cell adhesion: the molecular basis of tissue architecture and morphogenesis Cell 1996 84 345 357 8608588 10.1016/S0092-8674(00)81279-9
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
Ruoslahti E Fibronectin in cell adhesion and invasion Cancer Metastasis Rev 1984 3 43 51 6324988 10.1007/BF00047692
Hynes RO Integrins: versatility, modulation, and signaling in cell adhesion Cell 1992 69 11 25 1555235 10.1016/0092-8674(92)90115-S
Springer TA Traffic signals for lymphocyte recirculation and leukocyte emigration: The multistep paradigm Cell 1994 76 301 314 7507411 10.1016/0092-8674(94)90337-9
Hynes RO Bader BL Targeted mutations in integrins and their ligands: their implications for vascular biology Thromb Haemost 1997 78 83 87 9198133
Vasioukhin V Bauer C Yin M Fuchs E Directed actin polymerization is the driving force for epithelial cell-cell adhesion Cell 2000 100 209 219 10660044 10.1016/S0092-8674(00)81559-7
Ruoslahti E Pierschbacher MD New perspectives in cell adhesion: RGD and integrins Science 1987 238 491 497 2821619
Yamada KM Fibronectins: structure, functions and receptors Curr Opin Cell Biol 1989 1 956 963 2534045 10.1016/0955-0674(89)90065-3
Ruoslahti E The RGD story: a personal account Matrix Biol 2003 22 459 465 14667838 10.1016/S0945-053X(03)00083-0
Ruoslahti E Reed JC Anchorage dependence, integrins, and apoptosis Cell 1994 77 477 478 8187171 10.1016/0092-8674(94)90209-7
Giancotti FG Ruoslahti E Integrin signaling Science 1999 285 1028 1032 10446041 10.1126/science.285.5430.1028
Giancotti FG Tarone G Positional control of cell fate through joint integrin/receptor protein kinase signaling Annu Rev Cell Dev Biol 2003 19 173 206 14570568 10.1146/annurev.cellbio.19.031103.133334
Patti JM Allen BL McGavin MJ Höök M MSCRAMM-mediated adherence of microorganisms to host tissues Annu Rev Microbiol 1994 48 585 617 7826020 10.1146/annurev.mi.48.100194.003101
| 15836785 | PMC1097746 | CC BY | 2021-01-04 16:39:15 | no | Cell Commun Signal. 2005 Apr 18; 3:7 | utf-8 | Cell Commun Signal | 2,005 | 10.1186/1478-811X-3-7 | oa_comm |
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-51583109710.1186/1476-069X-4-5ResearchImmune function biomarkers in children exposed to lead and organochlorine compounds: a cross-sectional study Karmaus Wilfried [email protected] Kevin R [email protected] Thomas [email protected] Jutta [email protected] Nadia [email protected] Hermann [email protected] Department of Epidemiology, Michigan State University, B601 West Fee Hall, East Lansing, MI 48824, USA2 Central Laboratory, University Hospital Mannheim, Germany3 Ministry of Social Welfare Hesse, Department of Health, Wiesbaden, Germany4 Epidemiological Working Group of the Ministry of Environment and Health and the Institute for Mathematics and Data Management in Medicine, University Hospital Hamburg-Eppendorf, Germany5 Institute of Toxicology, Christian-Albrecht University, Kiel, Germany2005 14 4 2005 4 5 5 30 12 2004 14 4 2005 Copyright © 2005 Karmaus et al; licensee BioMed Central Ltd.2005Karmaus 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
Different organochlorines and lead (Pb) have been shown to have immunomodulating properties. Children are at greater risk for exposure to these environmental toxicants, but very little data exist on simultaneous exposures to these substances.
Methods
We investigated whether the organochlorine compounds (OC) dichlorodiphenylethylene (DDE), hexachlorobenzene (HCB), hexachlorocyclohexane (γ-HCH), the sum of polychlorinated biphenyls (ΣPCBs) and Pb were associated with immune markers such as immunoglobulin (Ig) levels, white blood cell (WBC), counts of lymphocytes; eosinophils and their eosinophilic granula as well as IgE count on basophils. The investigation was part of a cross-sectional environmental study in Hesse, Germany. In 1995, exposure to OC and Pb were determined, questionnaire data collected and immune markers quantified in 331 children. For the analyses, exposure (OC and Pb) concentrations were grouped in quartiles (γ-HCH into tertiles). Using linear regression, controlling for age, gender, passive smoking, serum lipids, and infections in the previous 12 months, we assessed the association between exposures and immune markers. Adjusted geometric means are provided for the different exposure levels.
Results
Geometric means were: DDE 0.32 μg/L, ΣPCBs 0.50 μg/L, HCB 0.22 μg/L, γ-HCH 0.02 μg/L and Pb 26.8 μg/L. The ΣPCBs was significantly associated with increased IgM levels, whereas HCB was inversely related to IgM. There was a higher number of NK cells (CD56+) with increased γ-HCH concentrations. At higher lead concentrations we saw increased IgE levels. DDE showed the most associations with significant increases in WBC count, in IgE count on basophils, IgE, IgG, and IgA levels. DDE was also found to significantly decrease eosinophilic granula content.
Conclusion
Low-level exposures to OC and lead (Pb) in children may have immunomodulating effects. The increased IgE levels, IgE count on basophils, and the reduction of eosinophilic granula at higher DDE concentrations showed a most consistent pattern, which could be of clinical importance in the etiology of allergic diseases.
==== Body
Background
Environmental toxicants such as organochlorine compounds (OC) and lead (Pb) may alter immune responses. There is a paucity of studies reporting associations between organochlorine [1-4] and lead [5-8] exposures and immune function biomarkers in children.
We conducted a large-scale environmental study of second-grade school children in three regions south of the Federal State of Hesse, Germany in 1995. Two of the regions are situated in the Rhine Valley with low mountains on both sides. One of these areas with several municipalities is located within a 10 km radius around an industrial waste incinerator and other industries, such as chemical plants. One plant was associated with dichlorodiphenylethylene (DDE), hexachlorobenzene (HCB), and hexachlorocyclohexane (γ-HCH) pollution [9]. The other region, also industrial, is 15 km north (downwind) of the incinerator. Both Rhine valley regions are also intensively used for the production of vegetables. The third study region is located in low mountains (about 0.4 km above sea level) that separate it from the industrial area. Blood concentrations of PCBs were shown to be higher in children living close to the toxic waste incinerator [10]. Results on PCBs and thyroid hormones, chromium and lymphocytes, DDE and breastfeeding and asthma have been published elsewhere [4,11-15].
Considering infection and atopic disorder in children, we have previously shown an association between DDE blood levels; asthma and one immunoglobulin (Ig), namely IgE [4]. However, the potential effects of organochlorines on other Igs and cellular defense were not reported. Hence, the focus of this paper is to investigate the impact of organochlorine compounds and Pb on humoral immune markers and cell-mediated immune responses. Specifically, for immune responses we focus on leukocytes, lymphocytes, B-cell, T-cells and their subsets. Assuming a concurrent effect of OC on immune markers, we conducted cross-sectional analyses of the data from the first of three surveys conducted in 1994/1995, 1996, and 1997. Only the first investigation included an extensive clinical assessment of immune markers.
Methods
Study population
After obtaining approval from the Data Protection Agency of Hamburg, Germany, the Ministry of Cultural Affairs of Hesse, Germany, and the local school committees, we invited the parents of 1,091 second-grade school children in 18 townships to participate in our study. We obtained informed consent from all participating parents, according to the requirements of the Ethical Committee of the Board of Physicians, the Helsinki Declaration, and the Data Protection Agency of the State of Hamburg. We asked each parent to allow their child to participate in phlebotomy only when passive smoking in the private household did not exceeded 10 cigarettes per day during the previous 12 months.
Questionnaires
We used four self-administered parental questionnaires in the survey: one regarding the living condition and nutrition of the family, one each for the mother and the father, and one regarding information on the child. Duration of breastfeeding was recorded in weeks of total and in weeks of exclusive nursing. Environmental tobacco smoke (ETS) was graded as smoking in the child's home in the previous 12 months (no cigarettes, 1–10 cigarettes, 11–20 cigarettes, 21–30 cigarettes, more than 30 cigarettes per day). We recorded age, gender, and the number of infections, defined as cold, coughing, and sore throat with or without fever in the last 12 months (none, less than 5, 5–10, more than 10).
Laboratory analyses of blood samples
One parent accompanied each child in the medical examination. For blood sampling, we used the 'Vacutainer System' (Becton, Dickinson & Company, San José, California,). Approximately 25 mL were drawn and separated into different aliquots. Immunoglobulin (Ig) E in serum was quantified at the Medical, Alimentary and Veterinary Institute for Research Middle Hesse, Division of Human Medicine, Dillenburg, Germany, using a florescence-immunoassay (CAP, Pharmacia, Uppsala, Sweden). To determine levels of specific IgE against inhalant allergens (aeroallergens), we incubated serum with immunocaps containing a mixture of aeroallergens and determined the reactivity using a fluorescence measurement (UNICAP Pharmacia, Uppsala, Sweden). Results from this method were provided in semi-quantitative format. We also measured IgA, G, and M by laser immunonephelometry (Dade Behring, Liederbach, Germany). The results for IgA, G and M were provided in mg/dL and for IgE in kU/L serum. Triglycerides and cholesterol were measured on a clinical chemistry analyzer according to IFCC methods (Hitachi 717, Boehringer Mannheim).
Leukocyte subsets
We collected 8 mL of blood in tubes containing EDTA and mixed them to prevent clotting. This aliquot was transported to the Central Laboratory of the University Clinic of Mannheim and analyzed on the same day. We used 200 μL of blood for the automated differential (laser-based hematology analyzer CD3500, Abbott Diagnostics, Santa Clara, California), and 100 μL for each of the nine three-color test tubes analyzed by flow cytometry (FACScan, Becton, Dickinson, & Company, San José, California, equipped with a 488 nm air-cooled argon ion laser). Eosinophils were determined according to their specific depolarisation characteristics and their eosinophilic granula content by the intensity of light scatter by flow cytometry. Basophils were identified by their high IgE density on the cell surface using immunofluorescence with a Phycoerythrin (PE) labeled anti-IgE antibody.
We used monoclonal antibodies directed against specific cell surface antigens to differentiate cell populations by multicolour immunofluorescence. Three antibodies were simultaneously applied with the fluorochrome combination FITC/PE/PE-Cy5. CD4/CD8/CD3 was used to detect absolute number of lymphocytes, T-helper cells and cytotoxic T-cells; CD19/CD5/IGE was used to differentiate B-cell subsets and basophils; CD3/CD16 and CD56/CD57 were used for natural killer cells. CD45RO defines memory T-helper cells. The CD nomenclature assigns the antibodies to clusters of differentiation, according to the International Workshop on Human Leukocyte Differentiation Antigens [16].
Organochlorine compounds (OC) in blood
OC including eight PCB congeners (101, 118, 138, 153, 170, 180, 183, 187), DDE, HCB, and three HCH congeners (α-, β- and γ) were determined (in μg/L) at the Institute of Toxicology, University of Kiel, Germany. OC were analyzed in 5 mL samples of whole blood by high resolution gas chromatography (HRGC, Model 3400 by Varian Inc., Palo Alto, California) with a 63Ni-electron-capture-detector. The detection limit (DL) (two times the signal/low-noise ratio) was 0.02 μg/L for β- and γ-HCH, DDE and each PCB congener, and 0.01 μg/L for HCB and α-HCH. For extraction and clean-up procedures, we used florisil and n-hexane for elution (9 g florisil were deactivated with 3% H2O and placed in a chromatography column 22 mm in diameter and 48 mm in length). The capillary column amounted to 30 mm in length and 0.25 mm in diameter; nitrogen was used as a carrier gas. We determined the PCB congeners by retention times on the chromatograms and identified them by comparison with known standards. Additionally, we tested reliability with gas chromatography-mass spectroscopy (GC/MS). The laboratory successfully participated in nationwide quality assessments for the determination of these OC.
Lead in blood
Lead (Pb) analysis was done at the Institute of Toxicology, University of Kiel, Germany. The determination in whole blood samples was by flow injection atomic absorption spectroscopy (Perkin Elmer) after adding 0.1% Triton-X-1-solution and 15 mol nitric acid to from a solution. This solution was then centrifuged at 3000 rpm. The DL for Pb was 9 μg/L (48 nmol/l; atomic weight: 207.19).
Data analyses
Since the data for leukocytes (WBC) and their subsets (lymphocytes and eosinophils), immunoglobulins, DDE, PCB congeners, HCB, γ-HCH and Pb were not normally distributed, the geometric mean, 5-, 95-percentiles are provided. In order to obtain a multivariate normal distribution, we log-transformed the number of cells and immunoglobulins before testing associations with possible predictors by multiple linear regression models.
All statistical analyses were performed using SAS software [17]. We calculated the sum of the PCB congeners (ΣPCBs = sum of seven congeners, the congener PCB101 was not detected). For descriptive purposes, we substituted values of OC below detection limit with one half of the detection limit. The statistical procedure (PROC RANK) was used to group exposure variables into quartiles (DDE, PCBs, HCB and Pb) or tertiles (γ-HCH). All observations below the detection limit were part of the lowest level group (reference). To account for the influence of lipids on the concentration of OC, we controlled for the sum of triglycerides and cholesterol in the regression analyses. Further steps were taken to determine whether our results were different when lipids were represented as sum of triglycerides and cholesterol as opposed to triglycerides and cholesterol as individual variables.
We used linear regression models (PROC GLM) with immune markers as dependent variables and all organochlorine compounds and lead as independent variables in each model. We also controlled for potential confounders (age, gender, environmental tobacco smoke (ETS), number of infections during the last 12 months, and lipid concentration). Information on passive smoking (ETS) in the child's home in the previous 12 months was divided into four categories (no cigarettes, 1–10 cigarettes, 11–20 cigarettes, 21 cigarettes per day and more). For the number of infections we considered four categories (none, less than 5, 5–10, more than 10). Age of the child was divided into three groups; 7, 8 and 9–10 years.
From the results of the regression analyses, we calculated adjusted geometric means for leukocyte subsets and immunoglobulins for increasing categories of exposure. T-tests were used to compare the statistical effect of higher exposure group to the lowest (reference).
Since one major route of exposure to the pollutants analyzed is breast feeding [18-21] and breastfeeding provides passive immunity [22-24], immune markers and pollutants could be spuriously correlated if breast feeding is not controlled for. However, this triangle (Figure 1) cannot be tested with linear regression models, as intervening variables do not qualify as confounders [25]. Controlling will reduce the initial association between the risk factor and the marker, as one causal chain is split into two associations. Thus, we explored the relationship between childhood breastfeeding (total duration of breastfeeding in weeks), the concentration of OC, and immune response by path analysis [26], using the CALIS procedure SAS Institute [17].
Figure 1 Diagramatic representation of the breastfeeding, childhood exposures and immune markers associations
Results
The proportion of participation was 61.5 % (671 of 1091). We obtained blood samples from 350 children, conducted OC and Pb analyses on 343 samples, and quantified immunoglobulins in 340. Overall, information (i.e., questionnaires, exposure biomarkers, and immune markers) was available for 331 children. Fewer girls than boys participated in phlebotomy; and 96 % of the children were 7 to 8 years of age (Table 1). Due to the inclusion criterion for blood sampling (passive smoking of less than 10 cigarettes in the child's home), the prevalence of passive smoking was also lower in the group with phlebotomy than in the total group (Table 1). Nevertheless, the fact that parents were separated or divorced and shared cohabitation for their child, resulted in a re-assessment of the passive smoking status after phlebotomy. Eligibility was determined on the information provided by one parent (mother or father) for their household. In the case of separate dwellings, we re-assessed the exposure by taking the average number of cigarettes smoked in both homes. As a consequence, 26 (7.9%) children who were exposed to more than 10 cigarettes per day at home had a phlebotomy and were included in the analyses.
Table 1 Descriptive characteristics of the study cohort.
Total group Subgroup with OC and immune markers
(N = 671) (n = 331)
% %
Boys 53.1 56.8
Age 7 years 45.8 46.2
8 years 50.2 50.2
9–10 years 4.1 3.6
Passive smoking in the child's home during the last 12 months (cigarettes per day)
None 52.2 66.5
1–10 23.4 24.8
11–20 14.3 5.3
more than 30 10.1 2.4
Number of infections during the last 12 months
None 6.0 5.7
1 to < 5 74.7 74.8
5 to 10 17.2 17.4
more than 10 2.1 2.1
Duration of total breastfeeding (weeks)
0 19.1 15.1
1 to < 5 7.9 15.4
5 – 8 12.5 12.1
9–12 10.6 11.8
more than 12 34.7 41.1
Missing 5.2 4.5
Serum cholesterol concentration
(mean, 5–95%-value, mg/dL) 186 (143–235)
Triglyceride concentration
(mean, 5–95%-value, mg/dL) 130 (53–262)
For γ-HCH, 27.7 % of the observations were below the detection limit, 2.9% for Pb, whereas none for DDE and HCB. At least one of seven PCB congeners was detected in each sample. Whole blood concentration for the sum of PCB congeners (118, 138, 153, 170, 180, 183, 187), HCB and of Pb showed a decline with increasing age (Table 2). DDE, PCB, and HCB concentrations were lower in children with higher passive smoking exposure. Regarding infections, lead concentration was higher in children with more than 10 infections during the last 12 months, whereas DDE concentration was lower in this group (Table 2).
Table 2 Geometric mean and 5-, 95% values for whole blood OC and Pb by covariates.
Category (n) DDE (μg/L) Sum of PCBs (μg/L) HCB (μg/L) γ-HCH (μg/L) Pb (μg/L)
Gender Girls (143) 0.32 (0.13 – 1.07) 0.43 (0.16 – 1.39) 0.21 (0.11 – 0.48) 0.02 (0.01 – 0.06) 25.4 (11.0 – 4 3.8)
Boy (188) 0.31 (0.13 – 0.96) 0.54 (0.19 – 1.66) 0.23 (0.11 – 0.54) 0.02 (0.01 – 0.04) 27.8 (14.8 – 48.2)
Age-groups 7 years (153) 0.32 (0.13 – 0.97) 0.54 (0.18 – 1.90) 0.23 (0.11 – 0.56) 0.02 (0.01 – 0.06) 27.3 (13.9 – 48.2)
8 years (166) 0.31 (0.13 – 0.98) 0.47 (0.18 – 1.29) 0.21 (0.11 – 0.48) 0.02 (0.01 – 0.05) 26.4 (10.7 – 47.8)
9–10 years (12) 0.31 (0.20 – 0.84) 0.33 (0.10 – 0.99) 0.17 (0.10 – 0.46) 0.02 (0.01 – 0.06) 25.4 (16.6 – 39.4)
Passive smoking in the child's home during the last 12 months (cigarettes per day) None (220) 0.35 (0.14 – 1.08) 0.57 (0.21 – 1.70) 0.24 (0.11 – 0.55) 0.02 (0.01 – 0.06) 26.5 (10.1 – 47.4)
1 – 10 (84) 0.26 (0.12 – 0.88) 0.39 (0.17 – 1.02) 0.19 (0.11 – 0.45) 0.02 (0.01 – 0.05) 26.0 (16.0 – 43.0)
11 – 20 (18) 0.27 (0.09 – 0.69) 0.40 (0.13 – 1.29) 0.18 (0.10 – 0.49) 0.02 (0.01 – 0.04) 33.5 (18.9 – 113.7)
21 – 30 (8) 0.23 (0.13 – 1.11) 0.27 (0.18 – 0.34) 0.15 (0.11 – 0.21) 0.02 (0.01 – 0.04) 30.1 (19.4 – 47.3)
Number of infections during the last 12 months None (19) 0.60 (0.16 – 4.02) 0.49 (0.10 – 2.24) 0.21 (0.10 – 0.58) 0.02 (0.01 – 0.08) 28.8 (15.9 – 58.7)
1 to < 5 (247) 0.31 (0.13 – 0.94) 0.49 (0.18 – 1.39) 0.22 (0.11 – 0.48) 0.02 (0.01 – 0.06) 26.2 (10.7 – 46.7)
5–10 (57) 0.29 (0.13 – 0.79) 0.53 (0.19 – 2.21) 0.23 (0.11 – 0.70) 0.02 (0.01 – 0.04) 27.9 (16.0 – 47.8)
>10 (7) 0.25 (0.16 – 0.43) 0.56 (0.34 – 0.87) 0.21 (0.15 – 0.27) 0.02 (0.01 – 0.05) 33.4 (26.2 – 48.5)
The concentrations of DDE, ΣPCBs (sum of PCBs), and HCB were all correlated (Table 3). However, we used categorized levels of OC, which were then only marginally correlated; the highest rank correlation was for the PCB and HCB groups (rSpearman = 0.46). These correlations did not result in multicollinearity since the tolerance (variance of OC not explained by other predictors) was at least 53%. The volume-based organochlorine concentrations were only marginally correlated with the lipid serum levels. To adjust for lipid concentrations, we included lipid concentrations as a confounder in the explanatory models for leukocyte subsets and immunoglobulins. Results derived from models using the sum of triglycerides and cholesterol compared to triglycerides and cholesterol as individual variables did not reveal any substantial difference (data not shown). We therefore reported results from models using the sum of triglycerides and cholesterol.
Table 3 Spearman correlation coefficients between organochlorine compounds (wet-based and lipid-based, n = 331) and their geometric means.
ΣPCBs HCB γ-HCH Lipids ψ Lipids § DDE/lipid ΣPCBs/lipid HCB/lipid γ-HCH/lipid Geo-metric mean
DDE (μg/L) 0.61 p < 0.01 0.55 p < 0.01 0.16 p < 0.01 0.08 p = 0.09 0.06 p = 0.25 0.86 p < 0.01 0.51 p < 0.01 0.46 p < 0.01 0.14 p < 0.01 0.32
ΣPCBs (μg/L) 0.76 p < 0.01 0.04 p = 0.40 0.04 p = 0.65 0.05 p = 0.34 0.59 p < 0.01 0.90 p < 0.01 0.70 p < 0.01 0.09 p = 0.11 0.50
HCB (μg/L) 0.07 p = 0.19 0.03 p = 0.63 0.04 p = 0.47 0.54 p < 0.01 0.74 p < 0.01 0.83 p < 0.01 0.09 p = 0.11 0.22
γ-HCH (μg/L) 0.15 p = 0.01 0.11 p < 0.05 0.13 p = 0.02 0.05 p = 0.34 0.03 p = 0.64 0.88 p < 0.01 0.02
DDE, lipid-based Ψ (ng/g) 0.63 p < 0.01 0.63 p < 0.01 0.23 P < 0.01 103.03
ΣPCBs, lipid-based Ψ (ng/g) 0.81 p < 0.01 0.14 p = 0.01 164.99
HCB, lipid-based Ψ (ng/g) 0.16 p < 0.01 70.72
HCH, Lipid-based Ψ (ng/g) 6.65
Ψ total lipids calculated as sum of cholesterol and triglycerides
§total lipids calculated using formula 2 of Phillips et al.[28].
rank correlation between total lipids from both formulae was (rSpearman = 0.95).
ΣPCBs sum of PCB congeners 101, 118, 138, 153, 170, 180, 183, 187
Regarding lead in whole blood, we found weak correlations with whole blood levels of OC (DDE: r = 0.15, n = 331, p < 0.01; HCB: r = 0.14, n = 331, p < 0.01; γ-HCH: r = -0.02, n = 331, p < 0.70; ΣPCBs: r = 0.14, n = 331, p < 0.01)
Increased white blood cell count (WBC; total leukocytes) was evident in the group with highest DDE level, whereas Pb, at the second, along with PCB at the highest level was associated with a reduction in WBC count. An increase in the number of eosinophils – a leukocyte subset – was identified in the highest DDE category, but not statistically significant, (see Additional file 3). However, eosinophilic granula content was significantly reduced at the upper DDE levels. In addition, IgE count on basophils was increased at higher DDE exposure, being statistically significant for the 0.3–0.43 μg/L category.
Regarding lymphocytes and specific lymphocyte subsets (B-cells, T-cells), the number of T-cells (CD3+), cytotoxic T-cells (CD8+) and B-cells (CD19+) were all significantly reduced in the median Pb category (see Additional file 1). Both natural killer (NK) cells (CD56+) and a NK cells subset (CD57+) were significantly associated with γ-HCH. However, these associations did not reveal dose-dependency.
All four immunoglobulins were associated in a virtually dose dependent fashion to either DDE, HCB or PCBs (see Additional file 2). IgM serum levels increased with the concentration of PCBs (F-test, p < 0.01) but decreased with increasing concentration of HCB (F-test, p < 0.01). In the two upper quartiles of DDE exposures, IgA levels were significantly higher, but lower in the upper quartile of HCB. DDE was not associated in a dose-response mode with IgG (F-test, p = 0.14), however, compared to the reference, the highest DDE exposure group showed a significantly elevated IgG level (t-test, p = 0.04). IgE levels more than doubled as DDE concentration increased (F-test, p = 0.02). The Pb serum levels were related to a significant differences in IgE (F-test: p = 0.028), but not in a dose dependent fashion (see Additional file 2).
Figure 2 shows that both DDE and lead were associated with higher serum IgE levels in children. In groups with lower DDE blood concentrations, Pb concentrations above the median (28 μg/L) were related to increase IgE levels. In groups with higher DDE, there was no additional effect of Pb. Statistically, the combined effect of DDE and Pb on IgE was not significant.
Figure 2 The combined effect of increasing DDE and lead (Pb) on IgE serum levels
In order to determine whether breastfeeding confounded the associations identified in linear regression models (Figure 1), we repeated our analyses using structural model (path analysis) for exposures determined as significant in linear regressions. Inclusion of breast feeding did not substantially change our findings.
Discussion
In 331 school children, age 7–10 years, we demonstrated significant relationships between OC and Pb whole blood concentration and cellular and humoral immune markers. First, modest associations were found between NK cells (CD3-CD16+CD56+) and a subset of natural killer cells (CD3-CD16+CD56+CD57+) and γ-HCH (see Additional file 1). Second, HCB was inversely related to IgM. Third, ΣPCBs were directly related to IgM. Fourth, our data showed that Pb decreased the count of T-cells (CD3+), cytotoxic T-cells (CD3+CD8+), and B-cells (CD3+CD5+ CD19+). This reduction was most evident at the 22.1 – 28.3 μg/L Pb concentration, though not in a dose response fashion. Lastly, DDE was inversely related to all immunoglobulins, except IgM (see Additional file 2). However, DDE was not associated with total serum protein (data not shown). The DDE effect was strongest for IgE – more than twofold increase – which also corresponded to an increased count of IgE on basophils. We did not detect a significant relationship between DDE and eosinophils, nevertheless, the number of eosinophils was positively correlated with IgE (rSpearman = 0.4, p < 0.01). However, high DDE levels were found to be significantly associated with lower eosinophilic granula content. The granula contains basic proteins which are cytotoxic and part of the inflammatory response [27].
The cross-sectional nature of the study limits conclusions on whether exposure occurred before immune responses. We can assume that organochlorine concentrations do not vary substantially in childhood, post breastfeeding. There is a decline of PCBs and HCB with age (Table 2), however the assumption of the stability is supported by a follow-up of the same children and OC determined in 1997. The Spearman rank correlation between the two successive measurements were high, with the exception of γ-HCH: DDE: r = 0.86, n = 274, p < 0.01; HCB: r = 0.74, n = 274, p < 0.01; γ-HCH: r = 0.1, n = 270, p = 0.11; ΣPCBs: r = 0.82, n = 274, p < 0.01.
The reported concentrations for OC were not lipid-based. In this cohort, there is a high correlation between lipid- and non lipid-based concentrations for OC (Table 3). Thus, our findings are independent of lipid- or wet weight-based determinations. In our models we controlled for lipids instead of dividing the concentration of OC by the lipid concentration for three reasons. First, a simple division assumes a monotonous linear relation between lipids and organochlorines. Although Phillips and co-workers reported for 20 adults that division by lipids reduces the difference between fasting and non-fasting concentration of OC [28], there is no data to justify a linear relation. Our data in children showed only weak correlations between OC and the sum of cholesterol and triglycerides (Table 3). This correlation did not increase when the sum of lipids were derived by using the 2nd formula proposed by Phillips et al. [28]. Second, there is no standard approach to adjust concentrations below the limit of detection for lipids. In particular, the probability of detection may be influenced by the individual lipid concentration of a child. Third, division by lipids does not take into account that they may confound the organochlorine – immune response relationships. Confounding is likely since lipids and OC are correlated, plus lipids are, for example, associated with the count of lymphocytes [29,30].
There is evidence that breast milk is a significant source of OC, Pb [18-21], and passive immunity [22-24]. Path analytical techniques (Figure 1) were used to verify whether breastfeeding as an intervening variable confounded our associations. The inclusion of breastfeeding in the path analysis did not reveal results different from the linear regression models. Hence, the associations between pollutants and immune markers were independent of breastfeeding.
We found whole blood concentrations of OC in our cohort comparable to similar children in Germany [31]. Compared to children in the United States, age 12–19 years (NHANES – 1999–2000) [32], our DDE values were lower though still within the 95% confidence interval. However, when comparing our results (in whole blood) with those of NHANES (in serum), we have to consider differences between serum and whole blood concentrations. Mes et al. reported that DDE was higher in sera and plasma than in whole blood samples [33]. Conversely, PCBs were higher in whole blood samples. No other comparison with NHANES data was possible as the values for PCB congeners and other OC were below the limit of detection.
Regarding lead (Pb), the geometric mean of 27 μg/L in our investigation was similar to the 33 μg/L found in a study of 797 East-German children 5–14 years of age [34]. Against that, the 1999–2000 NHANES study showed a lower geometric mean (15.1 μg/L) in 905 children 6–11 years of age [32]. However other studies in areas of higher exposure, reported average concentrations above the NHANES value: 40 μg/L for children, 6 to 15 years of age in four communities with mining and smelting operations and two control groups in the United States [6], and 95 μg/L in Chinese children 3–6 years old [8].
We selected a subgroup for blood analyses due to budget constraints. The group having a lower ETS exposure in their homes was selected to reduce the potentially confounding effect of ETS. This group did not significantly differ from other participating children (Table 1). Parents did not know the individual results of the blood analyses, when they provided information on their children, thus reducing recall bias.
The inverse association between DDE and the number of infections 12 months prior to the interview is surprising (Table 2). However, in a logistic regression model the number of infections reported did not show a significant protective effect of DDE. Additionally, when infection was eliminated from the models, there were no major changes in the OC – immune markers association.
The few existing studies estimating the immunotoxicity of lead (Pb) in children, measured by immune markers, are inconsistent in their findings. Regarding immunoglobulins, our positive relation between Pb and IgE was consistent with that of Lutz et al. [7]. However, Sun and co-workers had different results [8]. Concerning lymphocytes, we found that the number of B-cells was significantly reduced with increased Pb concentration. Conversely, Sarasua et al. reported an increase in the number of B-cells for children less than 3 years old [6].
Studies assessing the relation between organochlorine and immune markers, determined in our study, also showed conflicting results and focused mostly on adults [35-37]. In comparison with these adult studies, Vine et al. reported similar modest findings for immunoglobulins and DDE. However, only results for IgA showed statistical significance [35]. Our findings regarding IgE and eosinophilic granula suggest that DDE shifts the immune response into a T helper (Th) 2 direction [38]. Mechanistically, immune responses have been polarized into Th1 and Th2 reactions. Th1 responses lead to the secretion of immunoglobulin G (IgG) and removal of the allergen. The allergic Th2 phenotype is characterized by secretion of cytokines that promote immunoglobulin E (IgE) production resulting in allergies. This suggestion is in agreement with findings of Daniel and co-workers, who reported an association between DDE and interleukin-4, a Th2 cytokine [39]. In addition, our interpretation that DDE may be associated with an allergy-like response is supported by the distribution of aeroallergen-specific IgE results over the four DDE exposure levels. In the lowest DDE exposure group 11.3% of the children showed a positive specific IgE, 10.9% and 12.2% in the two intermediate groups, but 23.0% in the highest DDE exposure group (p = 0.03).
Interestingly, the effects of lead (Pb) and DDE on IgE seems to be competitive. At lower DDE exposure, Pb seems to increase IgE concentrations (Figure 2). There was no additional effect of the other pollutant if one is high; therefore it is possible that both pollutants are involved in the same mechanism. Indeed, studies have surmised that Pb may also shift the immune responses in a Th2 direction [40-42].
There are only few studies on OC blood/serum concentration and immune responses in children. Weisglas-Kuperus et al. reported that prenatal PCB exposure was associated with an increase in the T-cell markers CD3CD8+ and CD4+CD45RO+ [2]. Our data did not support these findings. In another study with prenatal exposures to PCBs, HCB, and DDE, Dewailly et al. did not identify significant associations with immune markers including CD3+, CD4+, CD8+ lymphocytes nor with IgA, IgG, and IgM [3]. However, we found significant relationships between PCBs and HCB with IgM (see Additional file 2). Reichrtova et al. have shown that in utero exposure to DDE is positively correlated with cord serum IgE [43]. No other study of children has investigated the relationship between DDE determined postnatally and Th2 markers such as IgE and eosinophilic granula. This is the second publication showing an association between DDE and serum IgE [4] and the first to report associations between Pb, and DDE and IgE count on basophils and eosinophilic granula.
Conclusion
In conclusion, our study suggests a non-linear association between IgE and Pb concentration. Regarding OC, our data indicated an increase of IgE related to DDE serum concentrations. A parallel association between DDE, IgE count on basophils, and reduction of eosinophilic granula contents further supports a potential stimulation of a Th2 response related to DDE exposure.
Prospective studies should determine more than one OC in a scenario with multiple exposures in order to prevent spurious correlations and include repeated determinations of immune responses to determine changes in immune development during childhood. Furthermore, studies are warranted that determine allergic susceptibilities following DDE and Pb exposure in children.
List of abbreviations
DDE, dichlorodiphenyl dichloroethene
EDTA, ethylene diamine tetra acetate
ETS, environmental tobacco smoke
FACS, Fluorescence-activated Cell Sorter
FITC, fluorescein isothiocyanate
GC/MS, gas chromatography-mass spectroscopy
HCB, hexachlorobenzene
HCH, hexachlorocyclohexane
HRGC, high resolution gas chromatography
Ig, immunoglobulin
NK-cells, natural killer cells
OC, organochlorine compounds
PCB, polychlorinated biphenyls
PE, Phyoerythrin
PE-Cy5, tandem fluorochrome of PE and cyanine 5
Th1, T-helper 1 cells
Th2, T-helper 2 cells
WBC, white blood cells
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WK designed the study and developed the analytical approach. Data analyses and manuscript preparation were done by WK and KRB. TN conducted cell analyses and helped to interpret the findings. HK was responsible for the determinations of organochlorine and lead and revised the manuscript. NOO and JW helped develop the surveys, supported their implementation, and revised the manuscript. All authors approved the final manuscript.
Supplementary Material
Additional File 3
Table 4: White blood cell, eosinophilic characteristics, and basophilic surface IgE by OC and Pb (geometric mean)
Click here for file
Additional File 1
Lymphocyte phenotypes by whole blood DDE, PCBs, HCB, γ-HCH and Pb concentration (geometric mean). The geometric mean of lymphocytes for different levels of OC is presented. Both F- and significant t-tests are also shown.
Click here for file
Additional File 2
Immunoglobulins by whole blood DDE, PCBs, HCB, γ-HCH and Pb concentration in children (geometric mean). The geometric mean for different immunoglobulins at different levels of OC is presented. Both F- and significant t-tests are also shown.
Click here for file
Acknowledgements
This study was supported by the Ministry of Environment, Energy, Youth, Family and Health Hesse, Germany. We acknowledge Dr. Rauterberg for the analyses of immunoglobulins. The analyses were conducted in preparation of an EPA STAR grant (R830825). KRB was a graduate assistant in the EPA grant while preparing the final manuscript.
==== Refs
Marth E Sixl W Bencko V Medwed M Lapajne S Voncina E Brumen S People on the garbage dumps of Cairo: a toxicological in vivo model? Cent Eur J Public Health 1995 3 154 157 8535375
Weisglas-Kuperus N Patandin S Berbers GA Sas TC Mulder PG Sauer PJ Hooijkaas H Immunologic effects of background exposure to polychlorinated biphenyls and dioxins in Dutch preschool children Environ Health Perspect 2000 108 1203 1207 11133402
Dewailly E Ayotte P Bruneau S Gingras S Belles-Isles M Roy R Susceptibility to infections and immune status in Inuit infants exposed to organochlorines Environ Health Perspect 2000 108 205 211 10706525
Karmaus W Kuehr J Kruse H Infections and atopic disorders in childhood and organochlorine exposure Arch Environ Health 2001 56 485 492 11958547
Rabinowitz MB Allred EN Bellinger DC Leviton A Needleman HL Lead and childhood propensity to infectious and allergic disorders: is there an association? Bull Environ Contam Toxicol 1990 44 657 660 2344471
Sarasua SM Vogt RF Henderson LO Jones PA Lybarger JA Serum immunoglobulins and lymphocyte subset distributions in children and adults living in communities assessed for lead and cadmium exposure J Toxicol Environ Health A 2000 60 1 15 10832614 10.1080/009841000156556
Lutz PM Wilson TJ Ireland J Jones AL Gorman JS Gale NL Johnson JC Hewett JE Elevated immunoglobulin E (IgE) levels in children with exposure to environmental lead Toxicology 1999 134 63 78 10413189 10.1016/S0300-483X(99)00036-0
Sun L Hu J Zhao Z Li L Cheng H Influence of exposure to environmental lead on serum immunoglobulin in preschool children Environ Res 2003 92 124 128 12854692 10.1016/S0013-9351(02)00090-7
Forth W Hexachlorcyclohexan - Gift in den Lebensmitteln? Deutsches Aerzteblatt 1980 83 2169 2176
Holdke B Karmaus W Kruse H [Body burden of PCB in whole human blood of 7-10-year-old children in the vicinity of a hazardous waste incinerator] Gesundheitwesen 1998 60 505 512
Osius N Karmaus W Kruse H Witten J Exposure to polychlorinated biphenyls and levels of thyroid hormones in children Environ Health Perspect 1999 107 843 849 10504153
Karmaus W Huang S Osius N Nebe T Chromium urine concentration and effects on lymphocyte subpopulations in children J Environ Med 1999 1 153 161 10.1002/1099-1301(199907/09)1:3<153::AID-JEM29>3.0.CO;2-O
Karmaus W DeKoning EP Kruse H Witten J Osius N Early childhood determinants of organochlorine concentrations in school-aged children Pediatr Res 2001 50 331 336 11518819
Karmaus W Davis S Chen Q Kuehr J Kruse H Atopic manifestations, breast-feeding protection and the adverse effect of DDE Paediatr Perinat Epidemiol 2003 17 212 220 12675789 10.1046/j.1365-3016.2003.00488.x
Obi-Osius N Misselwitz B Karmaus W Witten J Twin frequency and industrial pollution in different regions of Hesse, Germany Occup Environ Med 2004 61 482 487 15150386 10.1136/oem.2003.008342
Lanza F Moretti S Papa S Malavasi F Castoldi G Report on the fifth international workshop on human-leukocyte differentiation antigens, Boston, November 3-7, 1993 Haematologica 1994 79 374 386 7806095
SAS Institute SAS/STAT Software 1999 Cary, NC, SAS Institute Inc.
Abraham K Papke O Gross A Kordonouri O Wiegand S Wahn U Helge H Time course of PCDD/PCDF/PCB concentrations in breast-feeding mothers and their infants Chemosphere 1998 37 1731 1741 9828301 10.1016/S0045-6535(98)00238-0
Lackmann GM Schaller KH Angerer J Organochlorine compounds in breast-fed vs. bottle-fed infants: preliminary results at six weeks of age Sci Total Environ 2004 329 289 293 15262173 10.1016/j.scitotenv.2004.03.014
Rabinowitz M Leviton A Needleman H Lead in milk and infant blood: a dose-response model Arch Environ Health 1985 40 283 286 4062363
Ettinger AS Tellez-Rojo MM Amarasiriwardena C Bellinger D Peterson K Schwartz J Hu H Hernandez-Avila M Effect of breast milk lead on infant blood lead levels at 1 month of age Environ Health Perspect 2004 112 1381 1385 15471729
Hanson LA Breastfeeding provides passive and likely long-lasting active immunity Ann Allergy Asthma Immunol 1998 81 523 33; quiz 533-4, 537 9892025
Hanson LA Korotkova M Lundin S Haversen L Silfverdal SA Mattsby-Baltzer I Strandvik B Telemo E The transfer of immunity from mother to child Ann N Y Acad Sci 2003 987 199 206 12727640
Van de Perre P Transfer of antibody via mother's milk Vaccine 2003 21 3374 3376 12850343 10.1016/S0264-410X(03)00336-0
Rothman KJ Modern Epidemiology 1986 Boston/Toronto, Little Brown and company
Hatcher L A step-by-step Approach to Using SAS for Factor Analysis and Structural Equation Modeling. 1994 6 Cary, NC, SAS Institute Inc.
Leung DY Molecular basis of allergic diseases Mol Genet Metab 1998 63 157 167 9608537 10.1006/mgme.1998.2682
Phillips DL Pirkle JL Burse VW Bernert JTJ Henderson LO Needham LL Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding Arch Environ Contam Toxicol 1989 18 495 500 2505694 10.1007/BF01055015
Moreno LA Sarria A Lazaro A Lasierra MP Larrad L Bueno M Lymphocyte T subset counts in children with hypercholesterolemia receiving dietary therapy Ann Nutr Metab 1998 42 261 265 9812016 10.1159/000012742
Ogawa Y Imaki M Yoshida Y Shibakawa M Tanada S An epidemiological study on the association between the total leukocyte and neutrophil counts, and risk factors of ischemic heart disease by smoking status in Japanese factory workers Appl Human Sci 1998 17 239 247 10052221 10.2114/jpa.17.239
Landesgesundheitsamt Baden-Wuerttemberg Projekt Beobachtungsgesundheitsamter. Belastungs- und Wirkungsmonitoring 2000 Stuttgart,
Department of Health and Human Services Centers for Disease Control and Prevention Second national report on human exposure to environmental chemicals
Mes J Marchand L Turton D Lau P Ganz P The determination of polychlorinated biphenyls congeners and other chlorinated hydrocarbon residues in human blood, serum and plasma. A comparative study Intern J Environ Anal Chem 1992 48 175 186
Jacob B Ritz B Heinrich J Hoelscher B Wichmann HE The effect of low-level blood lead on hematologic parameters in children Environ Res 2000 82 150 159 10662529 10.1006/enrs.1999.4011
Vine MF Stein L Weigle K Schroeder J Degnan D Tse CK Backer L Plasma 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE) levels and immune response Am J Epidemiol 2001 153 53 63 11159147 10.1093/aje/153.1.53
Daniel V Huber W Bauer K Suesal C Conradt C Opelz G Associations of dichlorodiphenyltrichloroethane (DDT) 4.4 and dichlorodiphenyldichloroethylene (DDE) 4.4 blood levels with plasma IL-4 Arch Environ Health 2002 57 541 547 12696651
Cooper GS Martin SA Longnecker MP Sandler DP Germolec DR Associations between plasma DDE levels and immunologic measures in African-American farmers in North Carolina Environ Health Perspect 2004 112 1080 1084 15238281
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
Daniel V Huber W Bauer K Suesal C Conradt C Opelz G Associations of blood levels of PCB, HCHs, and HCB with numbers of lymphocyte subpopulations, in vitro lymphocyte response, plasma cytokine levels, and immunoglobulin autoantibodies Environ Health Perspect 2001 109 173 178 11266329
McCabe MJJ Lawrence DA Lead, a major environmental pollutant, is immunomodulatory by its differential effects on CD4+ T cells subsets Toxicol Appl Pharmacol 1991 111 13 23 1719661 10.1016/0041-008X(91)90129-3
Heo Y Lee BK Ahn KD Lawrence DA Serum IgE elevation correlates with blood lead levels in battery manufacturing workers Hum Exp Toxicol 2004 23 209 213 15222397 10.1191/0960327104ht442oa
Heo Y Lee WT Lawrence DA In vivo the environmental pollutants lead and mercury induce oligoclonal T cell responses skewed toward type-2 reactivities Cell Immunol 1997 179 185 195 9268502 10.1006/cimm.1997.1160
Reichrtova E Ciznar P Prachar V Palkovicova L Veningerova M Cord serum immunoglobulin E related to the environmental contamination of human placentas with organochlorine compounds Environ Health Perspect 1999 107 895 899 10544157
| 15831097 | PMC1097747 | CC BY | 2021-01-04 16:36:33 | no | Environ Health. 2005 Apr 14; 4:5 | utf-8 | Environ Health | 2,005 | 10.1186/1476-069X-4-5 | oa_comm |
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-291586012610.1186/1477-7525-3-29ResearchQuality of life in children newly diagnosed with cancer and their mothers Eiser Christine [email protected] J Richard [email protected] Christopher B [email protected] CR-UK Child and Family Research Group, Department of Psychology, University of Sheffield S10 2TP, UK2 Institute of Work Psychology, Department of Psychology, University of Sheffield, S10 2TP, UK2005 28 4 2005 3 29 29 8 3 2005 28 4 2005 Copyright © 2005 Eiser et al; licensee BioMed Central Ltd.2005Eiser 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
With current treatments, approximately 75% of children diagnosed with cancer can expect to achieve disease-free survival. However, treatments are complex and aggressive, potentially compromising QOL for children and their parents. Although previous work has shown increased anxiety and depression among parents after diagnosis, the recent development of standardised measures of QOL enables us to look more directly at the impact of diagnosis on mothers' and children's QOL. The aims of this study are to i) describe QOL for children and their mothers after diagnosis by comparing their scores with population norms, ii) explore the relationship between mothers' worries about the illness and their QOL, and iii) determine the relationship between mothers ratings of their own QOL and their child.
Method
A total of 87 families took part, constituting 60% of those eligible. The children included 58 males and 29 females aged between 2 years 6 months to 16 years 3 months (mean = 7 years, median = 5 years 8 months). Diagnoses were acute lymphoblastic leukaemia (ALL, n = 57), brain tumours (n = 11), bone tumours (n = 17) and 2 rare cancers. Mothers completed questionnaires about their own and the child's QOL.
Results
Mothers' reported their own and the child's QOL to be significantly lower than population norms. There were significant correlations between mothers' worries and their own and their ratings of the child's QOL and mothers' ratings of their own QOL correlated with their ratings of the child's QOL.
Conclusion
Both children and their mothers experience significantly compromised QOL in the months following diagnosis. Mothers who rated their own QOL to be poor also rate their child's QOL to be low. These results suggest caution is required where mothers rate their child's QOL. Efforts must continue to be made to improve QOL of children especially in the period immediately following diagnosis.
ChildhoodCancerQuality of lifeproxy ratings.
==== Body
Background
Survival in childhood cancer has improved significantly in recent years [1,2]. This has been attributed to the organisation of care in key centres, and development of national and international clinical trials which facilitate more rapid knowledge and refinements of new protocols. Depending on the specific cancer, children are treated with a combination of radiotherapy, chemotherapy and surgery. Duration of chemotherapy also varies but can be up to 3 years for boys treated for the most common cancer, acute lymphoblastic leukaemia (ALL).
Despite the improved survival statistics, cancer remains a potentially life-threatening condition, and as such poses a major challenge to both child and family. During the course of treatment, most children experience unpleasant physical side-effects. Behavioural and emotional problems have also been identified. In the longer term, there is a considerable risk of late effects [3]. These include reduced linear growth, compromised endocrine and sensory functions, and damage to cardiac and reproductive systems.
In addition to the effects on children, adverse consequences for parents' immediate and longer-term physical and mental health have been reported [4-6]. Many parents report elevated levels of depression and anxiety especially in the months immediately after diagnosis [7], but for most this decreases over time [5,8,9].
Improvements in survival are the result of increasingly aggressive treatments, raising questions about the quality of life (QOL) as well as quantity or length of survival. For the child, QOL is likely to be compromised by the pain of illness and treatment, lack of energy to enjoy everyday activities, and fears about the future. Mothers themselves experience great changes to their lives, staying in hospital with their child, perhaps giving up or reducing the hours they spend at work, as well as learning how to manage the child's medical care at home. The relatively recent emergence of standardised QOL measures provides the opportunity to formalise the extent to which QOL is compromised for mothers and children following diagnosis, and to describe any differences in impact depending on characteristics of the child (age, gender) and illness.
In this study, we obtained ratings from mothers about their own, and their child's QOL, and compared these with population norms. We predicted that, in the months immediately after diagnosis, mothers would rate their own and their child's QOL significantly below norms for the normal population. Second, we explored relationships between mothers' QOL and more illness-specific worries, and third, predicted that mothers who rated their own QOL to be poor would also rate their child's QOL to be poor.
Methods
Procedure
Families of a newly diagnosed child were invited to participate in a study about coping with the child's illness. Families were approached approximately three months after diagnosis, as the child's condition is normally relatively stable, and treatment is on an out-patient basis. Families were approached by the clinic nurse at a routine clinic visit and given written and verbal information about the study. Those who agreed to participate gave signed consent and were subsequently contacted by the research team who visited the family at home. All aspects of the procedure were approved by the Hospital Ethics Committees.
Sample
English-speaking families of children aged 2–18 years diagnosed at five cancer centres in the UK over a two-year period were approached. Exclusions were children with advanced disease, cognitive or neurological impairment prior to diagnosis, or other complicating conditions (e.g. Down's syndrome).
A total of 87 families took part (60%) of those eligible. Families who refused cited lack of interest, too distressed or too busy as explanations. The sample was predominantly Caucasian; there were two families of Asian origin, but both had lived in the UK for more than 10 years. There were 58 males and 29 females. The ages of the children ranged from 2 years 6 months to 16 years 3 months (mean = 7 years, median = 5 years 8 months). Children were being treated for acute lymphoblastic leukaemia (ALL, n = 57), brain tumours (n = 11), bone tumours (n = 17) and 2 rare cancers.
Questionnaires
Child's QOL
The Pediatric Quality of Life Inventory (PedsQL ™ 4.0) [10] includes subscales to measure physical health, social, school and emotional functioning. There are separate age-appropriate versions including one for children aged 2.5–4 years. For each item, mothers are asked how much of a problem has been experienced over the last month. Items are rated on 5-point Likert scales, from 0 (never a problem) to 4 (almost always a problem). After aggregation and transformation, subscale scores range from 0–100, with higher scores representing better QOL.
The Maternal worry scale [11] is an 11-item scale that measures mothers' worries concerning the child's future. Examples of the items include worries about future reliance on medication, becoming worse in the future and feeling different from others as a result of the illness. The scale has shown adequate psychometric properties, an internal consistency (Cronbach's α) of 0.94 and a test-retest reliability of 0.84.
Mothers' well-being
The SF-36 scale [12] was included as a measure of mothers' own well-being. This includes a single-item measure of change in health, plus eight subscales, with varying numbers of items and response formats, defined as physical functioning, role limitation (physical), role limitation (social), social functioning, mental health, energy/vitality, pain and general health perception. The scale has been extensively used in health services research, has established psychometric properties and UK norms [13].
Medical data
Information about diagnosis and date of diagnosis were obtained from medical records.
Treatment of results
There was considerable missing data for the school functioning subscale of the PedsQL ™ 4.0, which mothers did not complete whenever children were below school age or had not yet returned to school after diagnosis. This subscale was therefore excluded from analyses. We then calculated reliabilities for all the remaining scales and subscales. Internal consistency was consistently good (see Table 1). Comparisons between sample means and population means were based on 1 sample t-tests and relationships between variables investigated suing simple correlations. All analyses were conducted using SPSS version 10 for Macintosh.
Table 1 Child and parental QOL measures: means, norms and reliabilities
Measure (n) mean Norm alpha t
Child's QOL
Physical Functioning 74 36.6* 84.9 0.86 -20.2
Emotional Functioning 74 48.2* 74.7 0.75 -10.76
Social Functioning 74 66.4* 86.6 0.80 -7.63
Parent's QOL
Physical Function 66 91.77 89.5 0.91 1.19
Social Functioning 68 59.2* 86.9 0.82 -7.37
Role Limitation – Physical 69 67.0* 84.6 0.85 -3.82
Role Limitation – Emotional 66 47.0* 80.6 0.77 -6.48
Mental Health 69 54.5* 72.0 0.76 -7.16
Energy/Vitality 68 44.6* 58.6 0.80 -5.01
Pain 68 84.2 79.9 0.77 1.48
General Health Perception 68 73.0 75.0 0.85 -0.86
Worry 65 2.07 N/A 0.87
* Sample mean indicating significantly lower QOL than norm (p < .05, 1-tailed test).
Results
Comparisons with population norms
Consistent with our first hypothesis, mothers reported significantly lower QOL for their child compared with population norms (14). They also reported their own QOL to be lower than expected across most subscales (13), notable exceptions being the SF-36 subscales relating to physical function and pain, where mothers reported functioning within the normal range. These data are shown in Table 1.
Relationship between mothers' own QOL and worries
Mothers who reported more worries rated their own (r = -.53) and the child's QOL (Physical r = -.31; Emotional r = -.36; Social r = -.42) to be lower.
Relationship between mothers' own QOL and their ratings of the child's QOL
There were significant correlations between mothers' ratings of their own QOL and their ratings of the child's QOL (Physical: r = 0.43; Emotional r = 0.43; and Social r = 0.52). Mothers' ratings of the child's QOL suggested no differences depending on diagnosis, child age or gender.
Discussion
Our results suggest that, in the three to five months following diagnosis of ALL, mothers report that children's QOL is significantly compromised compared with the normal population. Although this is to be expected, the extent to which QOL scores were below population means was considerable. On most subscales of the SF-36, mothers also reported their own QOL to be lower than population means. Our results attest to the huge burden experienced by children and their parents during the initial period of treatment for cancer.
Although correlations give no information about the direction of a relationship (whether worries compromised QOL or vice-vera), we found that mothers who were more worried reported lower QOL themselves. Finally, mothers' ratings of their own QOL were correlated with their ratings of the child's QOL. Although it is normally recommended that ratings of QOL should be made by the patient wherever possible, it has always been acknowledged that there are situations, especially where children are very young or ill, when it is necessary to rely on parents' ratings. Diagnosis of cancer in a child is likely to be one of those situations, in so far as children are often young, ill and distressed. Our data suggest that mothers can provide ratings of the child's QOL but these are related to their own QOL. Mothers who rated their own QOL to be lower and reported more worries, rated the child's QOL to be worse. There is a question about how well mothers are able to report their child's QOL. A number of reports suggest there are differences between mothers and children in their views of the child's QOL, although these have not been related in any simple way to variables such as age or gender [15].
Study limitations include reliance on mothers' ratings of their own and their child's QOL. Although we had wanted to measure children's QOL by asking them directly, there were a number of obstacles to achieving this. First, the average age on diagnosis of ALL is 4 years, meaning that many children were simply too young to respond for themselves. Second, on diagnosis, even the older children tended to be too ill to respond.
The reliance on mothers, rather than fathers, may also be a limitation. Mothers tend to be more involved with the care of the sick child, more responsible for medication and treatment decisions, and more likely to stay in hospital with the child. In contrast, fathers tend to work as much as possible and generally try to maintain normal family life for other children in the family [16]. This difference in experience, resulting in fathers remaining more involved in everyday life despite the child's illness, may have an impact on parenting and parents' perceptions of the child's QOL.
In addition, our sample included only 60% of those eligible. We have no way of knowing whether those who refused differed in any crucial way from families who agreed to take part, though typically they stated they were too busy or distressed. Given the suddenness of the diagnosis and the amount of hospitalisation and care typically needed, it is not surprising that some families were overwhelmed and did not wish to add an additional demand on their time. As Ethics committees do not allow investigators to probe families reasons for agreeing or not to take part in research, we are unable to be any more precise on this point.
A further limitation follows from use of PedsQL ™ 4.0 as a measure of QOL. The PedsQL ™ 4.0 was developed as a generic measure of children's QOL, and therefore does not tap specific implications associated with diagnosis and treatment. Mothers failed to complete ratings in the schools subscale, on the basis that many children were attending sporadically or experienced very lengthy absences. This suggests the measure may lack sensitivity for newly diagnosed children and challenges the adequacy of generic measures of QOL to provide a comprehensive assessment of QOL in children with serious illness. Use of cancer specific scales would be helpful [17], though currently lack UK norms.
It is important that work of this kind has clinical implications for parents and medical staff. There are concerns that the treatments currently used to treat cancer may unnecessarily compromise child QOL, both during treatment and in the longer term [18]. Although use of Hickman lines and anti-emetic drugs are now routine and have reduced children's experiences of pain and chemotherapy related sickness, our data suggest much still needs to be done to improve QOL in children on treatment. Fear of infection and children's fatigue mean that families lead very isolated lives after diagnosis. If QOL is to be further improved, we need to find ways to reduce the sense of loneliness and isolation experienced, as well as family fears for the future.
Authors' contributions
C. Eiser conceived the study. JR Eiser and CB Stride were responsible for analyses, and all contributed to the write-up.
Acknowledgements
This research was supported by a grant from the Cancer Research-UK (CP 1019/0101 & CP 1019/0104). We would like to thank Yvonne Vance, Veronica Greco, Sally-Ann Clarke and Linda Sheppard for help with data collection and families and children who took part in this study, and Prof J. Varni for permission to use the PedsQL ™ 4.0 scale.
==== Refs
Stiller CA Population based survival rates for childhood cancer in Britain, 1980–91 BMJ 1994 309 1612 1616 7819936
Stiller CA Bunch K Lewis I Ethnic group and survival from childhood cancer: Report from the UK Children's Cancer Study Group Br J Cancer 2000 82 1339 1343 10755411 10.1054/bjoc.1999.1101
Wallace H Green D Late effects of childhood cancer 2004 Arnold, London
Manne S Miller D Meyers P Wollner N Steinherz P Redd W Depressive symptoms among parents of newly diagnosed children with cancer: A 6-month follow-up study Child Health Care 1996 25 191 209
Nelson A Miles M Reed S Davis C Cooper H Depressive symptomatology in parents of children with chronic oncologic or hematologic disease J Psychosoc Oncol 1994 12 61 75
Vance YH Morse R Jenney M Eiser C Issues in measuring quality of life in childhood cancer: Measures, proxies, and parental mental health J Child Psychol Psychiatry 2001 42 661 667 11464970 10.1017/S0021963001007314
Grootenhuis MA Last BF Predictors of parental emotional adjustment to childhood cancer Psycho-Oncology 1997 6 115 128 9205969
Dahlquist LM Czyzewski Copeland KG Jones CL Taub E Vaughan JK Parents of children newly diagnosed with cancer: Anxiety, coping and marital distress J Pediatr Psychol 1993 365
Sawyer M Streiner D Antoniou G Toogood I Rice M Influence of parental and family adjustment on later psychological adjustment of children treated for cancer J Am Acad Child Adolesc Psychiatry 1998 37 73 112
Varni J Seid M Rode C The PedsQL: Measurement model for the pediatric quality of life inventory Med Care 1999 37 126 139 10024117 10.1097/00005650-199902000-00003
DeVet K Ireys H Psychometric properties of the maternal worries scale for children with chronic illness J Pediatr Psychol 1998 23 257 266 9718899
Ware JE Snow KK Kosinski M Gandek B SF-36 health survey manual and interpretation guide 1993 Boston, MA: The Health Institute, New England Medical Center
Jenkinson C Coulter A Wright L Short form 36 (SF-36) health survey questionnaire normative data for adults of working age BMJ 1993 306 1437 1440 8518639
Upton P Eiser C Cheung I Hutchings HA Jenney M Maddocks A Russell IT Williams JG Measurement properties of the UK-English version of the Pediatric Quality of Life Inventory 4.0 (PedsQL) generic core scales Health Qual Life Outcomes 2005 3 22 15804349 10.1186/1477-7525-3-22
Eiser C Morse R Quality of life in chronic diseases of childhood Health Technol Assess 2001 5 1 157 11262421
Wysocki T Gavin L Psychometric properties of a new measure of fathers' involvement in the management of pediatric chronic diseases J Pediatr Psychol 2004 29 231 240 15131140 10.1093/jpepsy/jsh024
Varni JW Rode CA Seid M Katz ER Friedman-Bender A Quiggins DJL The pediatric cancer quality of life inventory-32 (PCQL-32). II. Feasibility and range of measurement J Behav Med 1999 22 397 406 10495970 10.1023/A:1018730204210
Craft AW Childhood cancer – mainly curable so where next? Acta Paediatr 2000 89 386 392 10830445 10.1080/080352500750028041
| 15860126 | PMC1097748 | CC BY | 2021-01-04 16:38:14 | no | Health Qual Life Outcomes. 2005 Apr 28; 3:29 | utf-8 | Health Qual Life Outcomes | 2,005 | 10.1186/1477-7525-3-29 | oa_comm |
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-301586013210.1186/1477-7525-3-30ResearchEffects of chronic widespread pain on the health status and quality of life of women after breast cancer surgery Burckhardt Carol S [email protected] Kim D [email protected] Primary Care, School of Nursing, Oregon Health & Science University, Portland, Oregon, USA2005 28 4 2005 3 30 30 1 3 2005 28 4 2005 Copyright © 2005 Burckhardt and Jones; licensee BioMed Central Ltd.2005Burckhardt and Jones; 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
Most research and treatment of post-breast cancer chronic pain has focused on local or regional pain problems in the operated area. The purpose of this pilot study was to compare and contrast the pain characteristics, symptom impact, health status, and quality of life of post-breast cancer surgery women with regional chronic pain versus those with widespread chronic pain.
Methods
A cross-sectional, descriptive design compared two groups of women with chronic pain that began after surgery: regional pain (n = 11) and widespread pain (n = 12). Demographics, characteristics of the surgery, as well as standardized questionnaires that measured pain (Brief Pain Inventory (BPI), Short Form McGill Pain Questionnaire (MPQ-SF)), disease impact (Fibromyalgia Impact Questionnaire (FIQ), Functional Assessment of Cancer Therapy-Breast (FACT-B)), health status (Medical Outcomes Short Form (SF-36)) and quality of life (Quality of Life Scale (QOLS)) were gathered.
Results
There were no significant differences between the groups on any demographic or type of surgery variable. A majority of both groups described their pain as aching, tender, and sharp on the MPQ-SF. On the BPI, intensity of pain and pain interference were significantly higher in the widespread pain group. Differences between the two groups reached statistical significance on the FIQ total score as well as the FACT-B physical well-being, emotional well-being and breast concerns subscales. The SF-36 physical function, physical role, and body pain subscales were significantly lower in the widespread pain group. QOLS scores were lower in the widespread pain group, but did not reach statistical significance.
Conclusion
This preliminary work suggests that the women in this study who experienced widespread pain after breast cancer surgery had significantly more severity of pain, pain impact and lower physical health status than those with regional pain.
==== Body
Background
Breast cancer is the most common form of cancer among women in the United States, Canada and Europe [1,2]. A sharp increase in incidence has been seen over the past two decades due in large part to use of mammography and subsequent earlier detection of disease. Earlier detection and treatment has led to increased survival rates approaching 90% for noninvasive cancers [3,4] Thus, a large majority of women with breast cancer will survive for many years after the initial diagnosis and treatment. Because they are living longer, living well with a good quality of life has become a high priority [5,6].
A recent study of health-related quality of life (HRQOL) concluded that most women treated for early-stage breast cancer have generally high HRQOL when compared to norms for the general population [7]. Unfortunately, long-term disease and treatment-related symptoms, such as chronic pain, can have wide-ranging consequences for health, functioning, and life quality [8-10]. Several researchers have reported that a substantial number of long-term breast cancer survivors experience chronic pain that interferes with physical functioning, mood, work, relationships, sleep, and enjoyment of life [11-15].
Chronic localized or regional pain after breast cancer surgery is a common and well-recognized problem with prevalence rates ranging from 20 to 65% [16-18]. Much of this pain is believed to be neuropathic phenomena due to transsection of nerves during surgery, nerve entrapment, axillary hematoma, or development of a traumatic neuroma on the operated side [19-21]. In addition, significantly higher rates of chronic pain on the affected side have been reported in patients who have had a mastectomy with reconstruction versus those who have had a mastectomy alone [18] and those who have undergone extensive axillary node dissection [22,23].
However, growing evidence exists that neuropathic phenomena alone do not explain all of the chronic pain experienced by post-surgery patients. Many breast cancer patients suffer from widespread, diffuse persistent pain that may be due to the chronic activation of nociceptors [24,25]. Reports on mastectomy with or without breast reconstruction have found a significant risk for development of fibromyalgia (FMS) [26], a specific syndrome of widespread chronic pain that affects approximately 2–5% of the female adult populations in the United States and many other countries [27,28].
Recent evidence has documented central sensitivity due to chronic input from peripheral nerves as a major cause of chronic pain in FMS patients [29,30]. And further, that trauma, especially to the upper body, is more likely to eventually lead to diffuse, widespread pain [31]. In both breast cancer patients and FMS patients, upper body trauma may be a factor in the onset and persistence of chronic widespread pain [32,33]. These findings are important because dysfunction, disability and detriments to quality of life are substantial in the FMS population [34,35]. But as yet, diffuse, widespread chronic pain among post-surgery breast cancer patients has not received the same attention as localized neuropathic pain.
In summary, until now, most work in post-breast cancer chronic pain has focused on local or regional pain problems in the operated area. Yet, a substantial number of post-breast cancer surgery women may have chronic widely diffuse pain that limits their functioning and decreases their overall life quality. The purpose of this pilot study was to compare and contrast the pain characteristics, syndrome impact, health status, and quality of life of post-breast cancer surgery women with localized or regional chronic pain versus post-breast cancer surgery women with widespread chronic pain.
Methods
Design
This pilot project used a cross-sectional, descriptive design. Participants responded to a battery of self-report questionnaires. The project focused on gathering preliminary descriptive data using well-validated scales.
Participants
Participants were recruited through an advertisement placed in a local newspaper, and mailings containing the advertisement to a database of women with FMS at a university out patient FMS clinic and to a university breast cancer clinic. The target sample size was 30.
The advertisement invited women, who believed that they met the criteria listed below, to call and leave a message at a specified number. An investigator called the woman, answered any questions about the study and then invited the woman to come to the clinic at a specified time if it appeared that she met criteria. Written informed consent for the data collection was obtained at the clinic visit after eligibility requirements were confirmed by the investigator.
Selection of participants
Women were eligible if they had had a simple mastectomy, lumpectomy, or modified radical mastectomy for breast cancer and included those with expansion or breast reconstruction at the time of initial surgery or subsequent breast reconstruction. Other specific inclusion criteria included: (1) adults at least 18 years of age; (2) first time primary diagnosis of breast cancer; (3) post-mastectomy for at least 6 months; (4) at least 3 months post-primary adjuvant treatment (radiation, cytotoxic chemotherapy); (5) cancer-free by their own report and having seen their physician within the last year; and, (6) current chronic pain by their own report that began after the breast cancer surgery. Exclusion criteria included: (1) breast surgery for cosmetic reasons or prophylactic mastectomy; and, (2) other painful or disabling medical conditions, such as arthritis.
Data collection
Demographic and cancer history information was obtained using a checklist and short answer questionnaire. Demographic variables included age, ethnic background, marital status, education, employment status, and occupation. Cancer history variables included date of surgery, type of breast cancer surgery, type of breast reconstruction surgery, adjuvant therapy, present cancer status.
Instruments for measuring pain characteristics
1. Short Answer Questions related to location, onset, extent, and quality of the pain as well as any medications for pain.
2. Body Diagram developed for use in FMS patients (Chris Henriksson, unpublished data). It was used in this study to distinguish regional from widespread pain. The subject was asked to mark each of 36 segments with a number between 1 and 10 signifying intensity of pain if she had any pain in that segment. This enabled us to distinguish regional pain (limited to 1 or 2 quadrants – upper, lower, right or left) and widespread pain that involved 3 or 4 quadrants as well as obtain some measure of the intensity of the pain at the time of the data collection. The use of quadrants to distinguish regional from widespread pain follows the American College of Rheumatology criteria for the classification of FMS [36].
3. The Brief Pain Inventory (BPI) which provides information on intensity (sensory dimension) of pain and the degree to which pain interferes with function (reactive dimension). The BPI uses 0 to 10 numeric rating scales and asks patients to rate their pain at the time of responding to the questionnaire as well as at its worst, least, and average over the previous week. Using the same rating scales, it asks the patient to rate the degree of interference with general activity, mood, walking and other physical activity, work, social activity, relationships with others and sleep. Developed originally for use in cancer treatment and research, the BPI has been well validated in cancer patients [37,38].
4. The Short-form McGill Pain Questionnaire a 15-item form of the longer McGill Pain Questionnaire [39]. Each descriptor is rated on a 4-point scale. Evidence for validity in cancer and FMS has been established [40,41]. This questionnaire was used to characterize the similarities and differences in pain descriptions between the two groups.
Instruments for measuring syndrome impact, health status and quality of life
1. The Fibromyalgia Impact Questionnaire a 10-item instrument that measures difficulties with activities of daily living and symptoms of pain, fatigue, morning tiredness, stiffness, job difficulty, depression and anxiety along with amount of work missed and overall well-being during the past week [42]. The instrument has been validated for FMS patients and discriminates between FMS, rheumatoid arthritis and healthy people. It is scored as individual items and as a total score that indicates more FMS impact. This instrument was used to measure overall symptom impact of widespread pain. Scores on the individual items are standardized from 0 to 10 and a total score can range from 0 to 100 with a higher score indicating greater impact (Current scoring information can be obtained at ).
2. Functional Assessment of Cancer Therapy (FACT-B) a 44-item self-report questionnaire designed to measure multidimensional health-related quality of life in patients with breast cancer [43]. The FACT-B consists of the FACT-G plus additional items that make up the breast cancer subscale [44]. The FACT-G contains five subscales: physical, functional, social/family and emotional well being and satisfaction with doctors. Each item is rated on a 5-point scale with 0 equal to "not at all" and 4 equal to "very much." Items are reversed if necessary and summed so that a higher subscale score indicates higher well being or satisfaction. All ratings on the FACT-B are completed in terms of the past seven days.
3. The SF-36 a 36-item scale that measures 9 domains of health including physical functioning, physical role limitations, bodily pain, general health, vitality, social functioning, emotional role limitations, mental health and change in health [45]. Higher scores indicate better health. The SF-36 has been used in numerous studies of cancer patients and has evidence of validity in both cancer and widespread pain patient groups [46,47]. We used this instrument as the measure of health status.
4. The Quality of Life Scale (QOLS) a 16-item non-health focused scale that measures satisfaction with multiple domains of life [48,49]. It uses a 1 to 7 point rating scale anchored with the words terrible and delighted. A higher total score indicates higher quality of life. Use in chronic illness populations, including a small group of cancer patients with ostomies, has been validated.
Statistical analysis
All data were analyzed using the SPSS version 12 statistical software package. Descriptive statistics (frequencies, means, standard deviations, and percentages) were used to characterize the sample. Preliminary inferences using t-tests for two independent samples and Chi-square for proportions were made. Because normality and equality of variance could not be assumed in the small samples, Q-Q plots and Levene tests were carried out. Q-Q plots revealed no serious deviations from normality. Levene tests for equality of variances enabled us to adjust the significance levels of the t-tests if the assumption of equal variance was not met for a variable. Alpha level for a significant difference was set at 0.01 because of the number of variables. As a check on the validity of the parametric test, we also ran a nonparametric analysis for 2 independent groups (Mann Whitney U) and found the same variables to be significant as on the t-test. Formal directional hypotheses were not made a priori. However, we expected to find that the group with widespread pain would have more severe pain impact, poorer health status and lower quality of life than the group with regional pain only.
Results
In all 30 women replied to the advertisement, 27 expressed initial interest in the study after talking with an investigator by telephone, and 23 scheduled and kept a clinic appointment at the university that reviewed and approved the study, signed the written consent form and completed data collection. No data were collected from the 7 women who inquired about the study but did not sign a consent form. The regional pain group was comprised of 11 women who reported pain only in the upper body (1 or 2 quadrants) while the widespread pain group was comprised of 12 women who reported pain in either 3 or 4 quadrants. The regional pain group marked an average of 2 areas on the body diagram; whereas the widespread pain group marked an average of 12 areas. The demographic, surgery and treatment characteristics of the sample are shown in Table 1. All subjects were at least 1 year post-surgery, radiation and chemotherapy and 80% were within 5 years of initial treatment. Sixty-five percent were on anti-estrogen therapy with either tamoxifen or remedex at the time of the study.
Table 1 Demographic and Breast Cancer Surgery and Treatment Characteristics by Pain Extent Group
Variable Regional Pain n = 11 Widespread Pain n = 12
Age (years) 56.8 (5.5) 58.7 (8.6)
Education (years) 16.0 (2.4) 14.8 (3.2)
Ethnic (% white) 91 100
Marital Status (% married) 73 69
Employment (% employed) 63 67
Time since Surgery (years) 5.9 (2.9) 5.4 (3.6)
Type of Surgery
Lumpectomy (%) 36 33
Mastectomy (%) 64 67
Axillary Node Dissection (%) 54 67
Breast Reconstruction (%) 18 0
Tissue Expansion (%) 100 0
Radiation Therapy (%) 60 80
Time Since Radiation (years) 4.2 (3.2) 2.9 (3.0)
Chemotherapy (%) 55 64
Time Since Chemo (years) 5.5 (2.4) 2.8 (1.6)
Anti-estrogen Therapy (%) 73 75
Current Anti-estrogen Therapy (%) 36 58
Numbers are means and standard deviations except when noted as percentages
Pain characteristics summarized in Table 2 indicated that a majority in both groups experienced the onset of their chronic pain immediately after surgery with a lesser number noting the onset weeks to months later. Those with regional pain were much more likely to describe their pain as intermittent while those with widespread pain were evenly split in this regard. Those with widespread pain were more likely to rate their pain as worse when depressed or fatigued, and observed a worsening of pain with exercise, rest, cold or contact with clothing. All subjects in the widespread pain group rated medication as at least somewhat effective for relieving their pain. Only 56% of those with regional pain noted any pain relief from medication. Of the 23 subjects, 5 were taking no medications, 12 were on non-steroidal antiinflammatories, 3 were taking narcotic pain relievers, 2 were on neurontin and 1 took glucosamine.
Table 2 Pain Characteristics by Pain Extent Group
Variable Regional Pain n = 11 Widespread Pain n = 12
When did pain begin
Immediately 63 56
Weeks later 12 11
Months later 25 33
Pain
Constant 20 50
Intermittent 80 50
Worse when Depressed 22 50
Worse when Fatigued 67 80
Relationship to Activity (Worse)
Exercise 44 70
Clothing Contact 11 36
Rest 33 73
Cold 25 56
Relief with Medication
Yes 22 30
Somewhat 34 70
No 44 0
Numbers are percentages
As seen in Table 3, a majority of both groups described their pain as aching, tender, and sharp. A higher percentage of the regional pain group described their pain as throbbing and stabbing while a higher percentage of the widespread pain group described it as shooting, cramping, gnawing, hot-burning, heavy, and splitting. They were also much more likely to use the words that described the emotional components of pain, such as tiring-exhausting and sickening. Overall, those with widespread pain endorsed more items than those with regional pain.
Table 3 Short-Form McGill Pain Questionnaire by Pain Extent Group
Variable Regional Pain (n = 11) Widespread Pain (n = 12)
Throbbing 27 17
Shooting 46 50
Stabbing 64 42
Sharp 54 58
Cramping 18 33
Gnawing 0 58*
Hot-Burning 9 50
Aching 73 83
Heavy 45 50
Tender 55 67
Splitting 18 33
Tiring-Exhausting 9 67*
Sickening 0 42
Fearful 9 17
Punishing-Cruel 18 33
Total Score 4.4 (2.7) 7.0 (3.7)
Numbers are percent of group that endorsed the item
*<0.01; **<0.001
Intensity of pain as well as multiple measures of pain interference on the BPI were all higher in the widespread pain group with the differences reaching statistical significance on all but three items (Table 4). Results of the FIQ (Table 5) were similar. The widespread pain group had higher impact scores and the differences between the two groups reached statistical significance on 5 of the 10 items as well as the total FIQ score. The FACT-B and SF-36 results shown in Table 6 revealed a statistically significant difference in the two groups on subscales measuring physical health status and well-being. Table 7 lists the items on the QOLS along with the total score. Although most of the scores were lower in the widespread pain group, none reached statistical significance.
Table 4 Brief Pain Inventory by Pain Extent Subgroup
Variable Regional Pain Widespread Pain
Pain Worst 3.1 (1.9) 6.6 (2.3)**
Pain Least 1.0 (1.3) 3.2 (2.9)
Pain Average 2.7 (1.6) 4.9 (2.1)*
Pain Right Now 1.4 (1.4) 5.2 (2.4)**
Pain Interference
General Activity 0.6 (1.3) 4.5 (2.1)**
Mood 1.0 (1.8) 4.8 (2.2)**
Walking Ability 0.6 (1.5) 4.2 (2.0)**
Normal Work 1.1 (1.7) 4.9 (2.5)**
Relations with others 0.4 (0.9) 4.2 (2.6)**
Sleep 2.3 (3.2) 5.4 (2.9)
Enjoyment of Life 1.3 (1.0) 5.2 (2.3)**
Numbers are means and standard deviations.
*<0.01; **<0.001
Table 5 Fibromyalgia Impact Questionnaire by Pain Extent Subgroup
Variable Regional Pain Widespread Pain
Physical Activity 1.5 (2.4) 3.1 (2.4)
Felt Good 1.6 (2.3) 6.1 (3.1)**
Missed Work 0 (0) 1.9 (3.0)
Job Difficulty 1.2 (2.1) 5.2 (2.4)**
Pain 3.5 (2.9) 5.8 (1.5)
Fatigue 4.3 (3.4) 8.0 (1.2)*
Unrested 3.4 (2.6) 7.2 (1.7)**
Stiffness 3.0 (3.3) 5.8 (2.6)
Anxiety 1.1 (2.2) 4.6 (2.8)*
Depression 1.4 (2.4) 4.3 (3.0)
FIQ Total Score 20.9 (13.2) 52.0 (15.1)**
Number are means and standard deviations
*<0.01; **<0.001
Table 6 FACT-B and SF-36 by Pain Extent Subgroup
Variable Regional Pain Widespread Pain
FACT-B
Physical Well-Being 22.9 (2.9) 16.0 (4.7)**
Social/Family Well-being 19.4 (4.8) 15.7 (6.1)
Relationship with Doctor 6.1 (1.9) 4.8 (2.5)
Emotional Well-being 20.1 (2.8) 15.7 (4.0)**
Functional Well-being 22.1 (3.9) 16.7 (5.7)*
Additional Concerns 24.3 (6.7) 16.5 (5.5)*
SF-36
General Health 60.3 (19.5) 57.9 (21.4)
Physical Functioning 68.6 (23.9) 48.7 (17.1)*
Role-Physical 63.6 (39.3) 18.7 (24.1)*
Bodily Pain 60.4 (14.4) 41.4 (14.1)*
Vitality 41.4 (16.3) 27.1 (14.2)
Social Functioning 78.4 (20.2) 61.4 (22.3)
Role-Emotional 75.7 (30.1) 47.2 (38.8)
Mental Health 61.1 (13.9) 51.3 (14.3)
*<0.01; **<0.001
Numbers are means and standard deviations
Table 7 Quality of Life Scale (QOLS) by Pain Extent Subgroup
Variable Regional Pain Widespread Pain
Material Comforts 5.0 (1.7) 5.0 (1.6)
Health 4.2 (1.5) 3.2 (1.3)
Relationships with relative 4.8 (1.5) 4.7 (1.2)
Having and rearing children 5.0 (1.3) 5.4 (1.2)
Relationship with spouse or partner 5.5 (1.1) 5.0 (1.9)
Having Close friends 6.0 (0.9) 5.7 (1.4)
Helping Others 5.9 (0.8) 5.3 (1.3)
Civic Activities 5.1 (1.6) 4.8 (1.3)
Intellectual Development 5.7 (1.3) 4.8 (1.1)
Understanding Self 5.4 (0.8) 5.0 (1.3)
Occupational Role 4.9 (1.7) 4.2 (1.4)
Creative expression 5.4 (1.2) 4.2 (1.6)
Socializing 5.4 (1.3) 4.4 (1.6)
Passive recreation 6.2 (0.9) 5.7 (1.4)
Active Recreation 4.7 (1.9) 3.4 (1.1)
Independence 5.9 (1.8) 5.1 (1.1)
QOLS Total Score 85.2 (10.0) 76.0 (15.3)
Numbers are means and standard deviations
Discussion
The results of this pilot study describe two groups of post breast cancer surgery patients who were similar in demographic, surgery and pain characteristics but markedly different in the impact that pain has on their health status and functioning. Widespread pain, which not been studied as a separate entity in this population, interferes with most aspects of these women's lives to a much greater degree than does regional pain.
For many patients, the chronic pain began immediately after surgery or early in the post-surgical treatment period. This type of onset is well recognized by breast cancer surgeons [23,50] and a number of efforts are being made to treat this lingering effect [20]. However, chronic pain that begins later in the post-surgical period after weeks or months may be less likely to be recognized and aggressively treated. Notably, FMS tends to be diagnosed only after a long period of persistent pain and failed local treatment.
Although a majority of patients described their pain as worse when they were fatigued, a description often voiced by people with chronic pain, the higher percentage of those with widespread pain who experienced worse pain with after exercise, rest or exposure to cold is similar to those with the specific syndrome of FMS. Increased sensitivity to muscle activity or inactivity as well as sensory input, such as cold, noise, bright lights or smells, are common to people with FMS [29,30]. It appears that women in the widespread pain group resemble those with FMS in some significant ways.
It was interesting and unexpected that those with widespread pain described getting at least some relief from pain medications as many patients with chronic pain get little relief from medication. On the other hand, it is also discouraging to note that many patients with regional pain did not get relief from medication as major strides have been made to aggressively treat neuropathic pain with tricyclic antidepressants and anticonvulsants such as gabapentin [20].
There were few differences in pain descriptions between the two groups with the exception of the emotional impact of pain, which was experienced more, by the widespread pain group. One item, hot-burning, was unexpectedly endorsed more by the widespread pain group, when one might have expected that sensation to be more descriptive of neuropathic pain experienced by the regional pain group.
Both the BPI and the FIQ clearly differentiated the regional from the widespread pain group. As expected, the widespread group was much more impacted. On the FIQ, their overall scores were very close to those noted in FMS studies [42,51]. The regional pain group's scores were much lower and closer to those of rheumatoid arthritis patients who have been used as a comparison group in studies of the psychometric properties of the FIQ [52]. The degree of pain interference with activities and enjoyment of life measured by the BPI was three to four times greater in the widespread pain group when compared to the regional pain group.
Three of the questionnaires, SF-36, FIQ and FACT-B, contain a subscale that focuses on physical functioning or well-being and they are moderately correlated with each other (r = .58–.60). Yet, the FIQ physical activity subscale scores were not significantly different between the two groups while the other two subscales were. The FACT-B physical well-being subscale focuses on symptoms such as fatigue, pain and feeling ill, which would be expected to bother the widespread pain group more. Both the SF-36 and FIQ subscales focus on normal physical activities such as walking, climbing stairs, household activities. However, the SF-36 includes vigorous activities such as walking more than a mile and climbing several flights of stairs, activities that might be much more difficult for a person with widespread pain to accomplish while the FIQ subscale focuses mostly on every day household activities. Thus, it would seem important to clearly identify the contents of subscales such as the ones used in this pilot study before attempting to compare groups of patients on health status and functioning measures.
Although the widespread pain group' total QOLS score was lower than the score of the regional pain group, neither it nor the individual item scores reached statistical significance. The regional pain group's mean total score was as high as those seen in healthy groups while the widespread pain group's score was lower but not as low as the scores seen in untreated FMS patients [34]. The QOLS is an individual satisfaction instrument that focuses on a wide range of domains that are not directly health-related. Scores on the QOLS have been shown to vary more in response to mood and other psychological states than to physical functioning [49]. The widespread pain group had significantly more psychological distress on the measures contained within the FIQ, SF-36 and FACT-B. This may account for the lower QOLS scores that in a larger sample would have reached statistical significance.
There are a number of limitations to this study. First, the sample was one of convenience and cannot be construed as representative of the post-breast cancer surgery population. Second, the sample was small and the number of variables large. Thus, the findings must be viewed with caution as some of the significant differences could have occurred by chance. Third, it is possible that adjuvant therapy rather than the surgery could have caused the chronic pain since both radiation and chemotherapy are known to produce symptoms of fatigue, pain and poor sleep in many patients. However, all patients were at least 1 year post adjuvant therapy and there were no differences in pain severity or other symptoms based on whether or not the subject had gotten adjuvant therapy. In addition, those on current tamoxifen therapy did not differ from those who were not.
This pilot study was carried out primarily to characterize two groups of post-breast cancer surgery patients with different pain patterns. A secondary purpose was to test a set of instruments to see if they would differentiate the two groups. Several of the instruments yielded scores that were significantly different on pain, health status, and interference with multiple aspects of life. This finding points to the need to study chronic pain further after breast cancer surgery and especially attend to the group of women whose pain has become widespread.
Conclusion
This preliminary work suggests that the women in this study who experienced widespread pain after breast cancer surgery had significantly more pain impact and lower health status than those with regional pain. Further comparative work should be undertaken.
Authors' contributions
CSB took major responsibility for conceptualization of the project, wrote the proposal, obtained funding, analyzed the data and wrote the complete draft of the paper
KDJ assisted in the literature review and writing of the proposal, oversaw the data collection and data entry, assisted in the analysis of the data, and reviewed drafts of the manuscript. Both authors reviewed the final draft of the manuscript and approved its contents.
Acknowledgements
The authors gratefully acknowledge the work of Vicky Crosby who collected the data.
==== Refs
Garfinkel L Boring CC Health CWJ Changing trends: an overview of breast cancer incidence and mortality Cancer 1994 74 222 227 8004590
Ries LAG Wingo PA Miller DS Howe HL Weir HK Rosenberg HM Vernon SW Cronin K Edwards BK The annual report to the nation on the states of cancer, 1973–1997 Cancer 2000 88 2398 2424 10820364 10.1002/(SICI)1097-0142(20000515)88:10<2398::AID-CNCR26>3.0.CO;2-I
Peto R Boreham J Clarke M Davies C Beral V UK and USA breast cancer deaths down 25% in year 2000 at ages 20–69 Lancet 2000 355 1822 10832853 10.1016/S0140-6736(00)02277-7
Smart CR Byrne C Smith RA Garfinkel L Letton AH Dodd GD Beahrs OH Twenty-year follow-up of the breast cancers diagnosed during the Breast Cancer Detection Demonstration Project CA Cancer J Clin 1997 47 134 149 9152171
Bonneterre J Schraub S Lecomte S Mercier M Quality of life as an outcome in breast cancer: clinical application Pharmacoecon 1996 9 23 29
Hurny C Bernhard J Coates A Quality of life assessment in the International Breast Cancer Study Group: past, present and future Recent Result Cancer Res 1998 152 390 395
Bardwell WA Major JM Rock CL Newman VA Thomson CA Chilton JA Dimsdale JE Pierce JP Health-related quality of life in women previously treated for early-stage breast cancer Psych-oncol 2003 13 595 604
Schag CA Ganz PA Polinsky ML Fred C Hirji K Petersen L Characteristics of women at risk for psychosocial distress in the year after breast cancer J Clin Oncol 1993 11 783 793 8478672
Glover J Dibble SL Dodd MJ Miaskowski C Mood states of oncology outpatients: does pain make a difference? J Pain Symptom Manage 1995 10 120 128 7730684 10.1016/0885-3924(94)00073-T
Strang P Existential consequences of unrelieved cancer pain Palliative Med 1997 11 299 305
Amichetti M Caffo O Pain after quadranectomy and radiotherapy for early-stage breast cancer: Incidence, characteristics and influence on quality of life Oncology 2003 65 23 28 12837979 10.1159/000071201
Caffo O Amichetti M Ferro A Lucenti A Valduga F Galligioni E Pain and quality of life after surgery for breast cancer Breast Cancer Res Treat 2003 80 39 48 12889597 10.1023/A:1024435101619
Fortner BV Stepanski EJ Wang SC Kasprowicz S Durrence HH Sleep and quality of life in breast cancer patients J Pain Symptom Manage 2002 24 471 480 12547047 10.1016/S0885-3924(02)00500-6
Poole K Fallowfield LJ The psychological impact of post-operative arm morbidity following axillary surgery for breast cancer: a critical review Breast 2002 11 81 87 14965650 10.1054/brst.2001.0369
Rietman JS Dijkstra PU Debreczeni R Geertzen JHB Robinson DPH de Vries J Impairments, disabilities and health related quality of life after treatment for breast cancer: a follow-up study 2.7 years after surgery Disability Rehabil 2004 26 78 84 10.1080/09638280310001629642
Randal J Post-mastectomy pain found to common: treatment options sparse, but growing J Natl Cancer Inst 1998 90 731 732 9605640 10.1093/jnci/90.10.731
Stevens PE Dibble SL Miaskowski C Prevalence, characteristics and impact of postmastectomy pain syndrome: an investigation of women's experiences Pain 1995 61 61 68 7644250 10.1016/0304-3959(94)00162-8
Wallace MS Wallace AM Lee J Dobke MK Pain after breast surgery: a survey of 282 women Pain 1996 66 195 205 8880841 10.1016/0304-3959(96)03064-3
Blunt C Schmiedel A Some cases of severe post-mastectomy pain syndrome may be caused by an axillary haematoma Pain 2004 108 294 296 15030949 10.1016/j.pain.2003.12.011
Jung BF Ahrendt GM Oaklander AL Dworkin RH Neuropathic pain following breast cancer surgery; proposed classification and research update Pain 2003 104 1 13 12855309 10.1016/S0304-3959(03)00241-0
Tasmuth T Karaja M Blomqvist C von Smitten K Kalso E Treatment-related factors predisposing to chronic pain in patients with breast cancer Acta Oncologica 1997 36 625 630 9408154
Warmuth MA Bowen G Prosnitz LR Chu L Broadwater G Peterson B Leight G Winer EP Complications of axillary node dissection for carcinoma of the breast: a report based on a patient survey Cancer 1998 83 1362 1368 9762937 10.1002/(SICI)1097-0142(19981001)83:7<1362::AID-CNCR13>3.0.CO;2-2
Legeby M Segerdahl M Sandelin K Wickman M Ostman K Olofsson C Immediate reconstruction in breast cancer surgery requires intensive post-operative pain treatment but the effects of axillary dissection may be more predictive of chronic pain Breast 2002 11 156 162 14965663 10.1054/brst.2001.0386
Carpenter JS Andrykowski MA Sloan P Cunningham L Cordova MJ Studts JL McGrath PC Sloan D Kenady DE Postmastectomy/postlumpectomy pain in breast cancer survivors J Clin Epidemiol 1998 51 1285 1292 10086821 10.1016/S0895-4356(98)00121-8
Andrykowski MA Curran SL Carpenter JS Studts JL Cunningham L McGrath PC Sloan DA Kenady DE Rheumatoid symptoms following breast cancer treatment: a controlled comparison J Pain Symptom Manage 1999 18 85 94 10484855 10.1016/S0885-3924(99)00053-6
Nyren O Yin L Josefsson S McLaughlin JK Blot WJ Engqvist M Hakelius L Boice JD JrAdami H-O Risk of connective tissue disease and related disorders among women with breast implants: a nation-wide retrospective cohort study in Sweden BMJ 1998 316 417 422 9492663
Croft P Rigby AS Boswell R Schollum J Silman A The prevalence of chronic widespread pain in the general population J Rheumatol 1993 20 710 713 8496870
Wolfe F Ross K Anderson J Russell IJ Hebert L The prevalence and characteristics of fibromyalgia in the general population Arthritis Rheum 1995 38 19 28 7818567
Staud R Cannon RC Temporal summation of pain from mechanical stimulation of muscle tissue in normal controls and subjects with fibromyalgia Pain 2003 102 87 95 12620600 10.1016/s0304-3959(02)00344-5
Geisser ME Casey KL Brucksch CB Perception of noxious and innocuous heat stimulation among healthy women and women with fibromyalgia Pain 2003 102 243 250 12670665 10.1016/S0304-3959(02)00417-7
Buskila D Neumann L Vaisberg G Alkalay D Wolfe F Increased rates of fibromyalgia following cervical spine injury. A controlled study of 161 cases of traumatic injury Arthritis Rheum 1997 40 446 452 9082932
Sartariano WA Ragland DR DeLorenze GN Limitations in upper-body strength associated with breast cancer: a comparison of black and white women J Clin Epidemiol 1996 49 535 544 8636727 10.1016/0895-4356(95)00565-X
Waylonis GW Perkins RH Post-traumatic fibromyalgia: a long-term follow-up Am J Phys Med Rehabil 1994 73 403 412 7993614
Burckhardt CS Clark SR Bennett RM Fibromyalgia and quality of life: a comparative analysis J Rheumatol 1993 20 475 479 8478854
Wolfe F Anderson J Harkness D Work and disability status of persons with fibromyalgia J Rheumatol 1997 24 1171 1178 9195528
Wolfe F Smythe HA Yunus MB Bennett RM Bombardier C Goldenberg DL Tugwell P Campbell SM Abeles M Clark P Fam AG Fiechtner JJ Franklin CM Gatter RA Hamaty K Lessard J Lichtbroun AS Masi AT McCain GA Reynolds WJ Romano TJ Russell J Sheon RP The American College of Rheumatology 1990 criteria for the classification of fibromyalgia: Report of the Multicenter Criteria Committee Arthritis Rheum 1990 33 160 172 2306288
Daut RL Cleeland CS Flanery RC Development of the Wisconsin Brief Pain Questionnaire to assess pain in cancer and other diseases Pain 1983 17 197 210 6646795 10.1016/0304-3959(83)90143-4
Cleeland CS Nakamura Y Mendoze TR Eduwards KR Douglas J Serline RC Dimensions of the impact of cancer pain in a four country sample: new information from multidimensional scaling Pain 1996 67 267 273 8951920 10.1016/0304-3959(96)03131-4
Melzack R The short-form McGill Pain Questionnaire Pain 1987 30 191 197 3670870 10.1016/0304-3959(87)91074-8
Dudgeon D Raubertas RF Rosenthal SN The short-form McGill Pain Questionnaire in chronic cancer pain J Pain Symptom Manage 1993 8 191 195 7963759 10.1016/0885-3924(93)90126-G
Burckhardt CS Bjelle A A Swedish version of the short-form McGill Pain Questionnaire Scand J Rheumatol 1994 23 77 81 8165442
Burckhardt CS Clark SR Bennett RM The Fibromyalgia Impact Questionnaire: Development and validation J Rheumatol 1991 18 728 734 1865419
Cella DF Tulsky DS Gray G The Functional Assessment of Cancer Therapy scale: development and validation of the general measure J Clin Oncol 1993 11 570 579 8445433
Brady MJ Cella DF Mo F Bonomi AE Tulsky DS Lloyd SR Reliability and validity of the Functional Assessment of Cancer Therapy-Breast quality of life instrument J Clin Oncol 1997 15 974 986 9060536
Ware JE Sherbourne CD The MOS 26-item Short-Form Health Survey (SF-36). I. Conceptual famework and item selection Med Care 1992 30 473 481 1593914
Broeckel JA Jacobsen PB L B Horton J Lyman GH Quality of life after adjuvant chemotherapy for breast cancer Breast Cancer Res Treat 2000 62 141 150 11016752 10.1023/A:1006401914682
Neumann L Berzak A Buskila D Measuring health status in Israeli patients with fibromyalgia syndrome and widespread pain and healthy individuals: utility of the short form 36-item health survey (SF-36) Semin Arthritis Rheum 2000 29 400 408 10924026 10.1053/sarh.2000.7171
Burckhardt CS Woods SL Schultz AA Ziebarth DM Quality of life of adults with chronic illness: a psychometric study Res Nurs Health 1989 12 347 354 2602575
Burckhardt CS Anderson KA Archenholtz B Hagg O The Flanagan Quality of Life Scale: Further evidence of validity Health Quality Life Outcome 1 59 (23 October 2003). 10.1186/1477-7525-1-59
Schierle C Winograd JM Radiation-induced brachial plexopathy: review. Complication without a cure J Resconstr Microsurg 2004 20 149 152 10.1055/s-2004-820771
Bennett RM Burckhardt CS Clark SR O'Reilly CA Wiens AN Campbell SM Group treatment of fibromyalgia: description and results of a six-month out-patient program J Rheumatol 1996 23 521 528 8832996
Hedin P-J Hamne M Burckhardt CS Engstrom-Laurent A The Fibromyalgia Impact Questionnaire, a Swedish translation of a new tool for evaluationn of the fibromyalgia patient Scand J Rheumatol 1995 24 69 75 7747146
| 15860132 | PMC1097749 | CC BY | 2021-01-04 16:38:14 | no | Health Qual Life Outcomes. 2005 Apr 28; 3:30 | utf-8 | Health Qual Life Outcomes | 2,005 | 10.1186/1477-7525-3-30 | oa_comm |
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-91582631510.1186/1476-072X-4-9ResearchAssessing spatio-temporal variability of risk surfaces using residential history data in a case control study of breast cancer Han Daikwon [email protected] Peter A [email protected] Matthew R [email protected] Jing [email protected] John E [email protected] Paola [email protected] Maurizio [email protected] Jo L [email protected] Department of Social and Preventive Medicine, University at Buffalo, Buffalo, NY 14214 USA2 Department of Geography and National Center for Geographic Information and Analysis, University at Buffalo, Buffalo, NY 14261 USA3 Department of Biostatistics, University at Buffalo, Buffalo, NY 14214 USA4 Department of Epidemiology and Biostatistics, University of South Carolina, Columbia, SC 29208 USA2005 12 4 2005 4 9 9 28 2 2005 12 4 2005 Copyright © 2005 Han et al; licensee BioMed Central Ltd.2005Han 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
Most analyses of spatial clustering of disease have been based on either residence at the time of diagnosis or current residence. An underlying assumption in these analyses is that residence can be used as a proxy for environmental exposure. However, exposures earlier in life and not just those in the most recent period may be of significance. In breast cancer, there is accumulating evidence that early life exposures may contribute to risk. We explored spatio-temporal patterns of risk surfaces using data on lifetime residential history in a case control study of breast cancer, and identified elevated areas of risk and areas potentially having more exposure opportunities, defined as risk surfaces in this study. This approach may be more relevant in understanding the environmental etiology of breast cancer, since lifetime cumulative exposures or exposures at critical times may be more strongly associated with risk for breast cancer than exposures from the recent period.
Results
A GIS-based exploratory spatial analysis was applied, and spatio-temporal variability of those risk surfaces was evaluated using the standardized difference in density surfaces between cases and controls. The significance of the resulting risk surfaces was tested and reported as p-values. These surfaces were compared for premenopausal and postmenopausal women, and were obtained for each decade, from the 1940s to 1990s. We found strong evidence of clustering of lifetime residence for premenopausal women (for cases relative to controls), and a less strong suggestion of such clustering for postmenopausal women, and identified a substantial degree of temporal variability of the risk surfaces.
Conclusion
We were able to pinpoint geographic areas with higher risk through exploratory spatial analyses, and to assess temporal variability of the risk surfaces, thus providing a working hypothesis on breast cancer and environmental exposures. Geographic areas with higher case densities need further epidemiologic investigation for potential relationships between lifetime environmental exposures and breast cancer risk. Examination of lifetime residential history provided additional information on geographic areas associated with higher risk; limiting exploration of chronic disease clustering to current residence may neglect important relationships between location and disease.
==== Body
Background
In a recent analysis of breast cancer by New York State's Department of Health, a breast cancer cluster in the Western New York area was identified [1]. One of the objectives of such disease mapping is to generate hypotheses by identifying spatial patterns so that causal processes may be evaluated further by more rigorous epidemiologic study. Spatial analyses have played a valuable role in explaining different health outcomes and in uncovering environmental causes of disease [2,3]. Residential locations at the time of diagnosis have generally been used in these exploratory spatial analyses [4,5]. Disease mapping has been increasingly used to identify spatial patterns with the aid of Geographic Information Systems (GIS) and exploratory spatial analysis tools, and has been a valuable tool for studies of geographic and environmental epidemiology, especially when the causes of disease and their determinant processes are unknown [6-8]. In particular, there has been recent interest in the use of kernel density estimation methods in epidemiologic studies. Density estimation methods have been used to smooth out noise based on functions of the data in surrounding areas and to overcome problems associated with traditional disease mapping [9,10].
Previous studies using exploratory spatial analyses, however, have been based on either residence at the time of diagnosis or current residence, and only a few recent studies have examined disease risk using information on lifetime residence [11,12]. For chronic disease, there is increasing evidence that lifetime exposures may be more relevant in understanding disease etiology. For breast cancer in particular, several of the well established risk factors (age at menarche, age at first birth) are from early life. There is now evidence that childhood and even in utero exposures may affect risk [13]. To examine disease clustering, lifetime cumulative exposures or exposures at critical times in a life course may be more strongly associated with risk for breast cancer than exposures from any one time period, especially the recent period.
In this study, we explored spatio-temporal patterns of risk surfaces using data on lifetime residential history in a case control study of breast cancer. We had previously identified geographic clustering of residence at critical points in early life in relation to breast cancer risk [14]. Here we focused on lifetime cumulative exposure in relation to the disease risk. Risk surfaces were created based on the relative densities of cases and controls – this indicated areas with higher case density as being areas with higher breast cancer risk, thus identifying areas potentially having more exposure opportunities. We used residence as a proxy for potential environmental exposures, conducted exploratory spatial analyses of breast cancer, and produced risk surface maps using information on lifetime residence to identify areas with high breast cancer incidence. In particular, we assessed spatio-temporal variability of risk surfaces using the standardized difference in case and control density, and evaluated the potential use of different kernel density estimation methods in applying them to epidemiologic data.
Results
Descriptive characteristics of study participants by menopausal status are presented in Table 1, and characteristics of lifetime residential history for breast cancer cases and controls are summarized in Table 2. One-fourth of the study participants had at least one previous residence outside the study area, and these were excluded from the analysis. For those residences in the study area, we found that cases were somewhat more mobile, averaging 5.8 and 5.4 residences for pre- and postmenopausal participating cases, compared to 4.9 and 5 residences for pre- and postmenopausal participating controls, respectively.
Table 1 Descriptive characteristics of study participants (Mean ± Standard Deviation): WEB Study, 1996–2001.
Premenopausal Postmenopausal
Case (n = 325) Control (n = 610) Case (n = 841) Control (n = 1495)
Age (years) 44.9 ± 4.6 44.1 ± 4.6 63.0 ± 8.5 63.4 ± 8.9
Education (years) 14.0 ± 2.3 14.2 ± 2.2 13.3 ± 2.6 13.0 ± 2.3
Parity 1.9 ± 1.3 2.0 ± 1.3 3.0 ± 1.9 2.5 ± 1.8
Age at menarche (years) 12.5 ± 1.6 12.6 ± 1.6 12.6 ± 1.6 12.8 ± 1.7
Age at first birth (years) 25.0 ± 5.1 25.8 ± 4.8 23.8 ± 4.7 23.5 ± 4.3
Recent BMI (kg/m2) 26.7 ± 6.6 27.2 ± 6.8 28.9 ± 6.1 28.4 ± 6.4
Benign breast disease (yes) 37% 21% 33% 22%
Relative with breast cancer (yes) 21% 10% 20% 14%
Table 2 Descriptive characteristics of lifetime residential history for breast cancer cases and controls: WEB Study, 1996–2001;
Erie and Niagara (n = 15487) Outside (n = 4752) Total (n = 20240)
Case Control Case Control Case Control
Premenopausal
Total numbers of residences in lifetime 1661 2767 432 948 2093 3715
Average numbers of residences per participant 5.8 4.9 3.2 3.4 5.0 4.4
Average years in each residence* (Mean ± SD) 5.6 ± 6.0 6.2 ± 6.6 4.3 ± 5.4 3.9 ± 4.9 5.3 ± 5.9 5.5 ± 6.3
Postmenopausal
Total numbers of residences in lifetime 4217 6842 1290 2082 5508 8924
Average numbers of residences per participants 5.4 5.0 3.5 3.2 4.8 4.4
Average years in each residence* (Mean ± SD) 6.9 ± 7.3 7.3 ± 7.9 4.8 ± 5.6 5.2 ± 6.3 6.3 ± 7.0 6.7 ± 7.6
* Excludes residences with missing data for length of residence.
We constructed lifetime cumulative risk surfaces to represent exposure opportunities in lifetime because cumulative exposures may be a more accurate indicator of potential environmental exposures related to breast cancer risk. Figure 1 shows a boundary map of Erie and Niagara counties, while Figure 2 depicts geographic patterns of lifetime residential locations for breast cancer cases and controls by menopausal status. We used the rectangular region as an approximate boundary of the study area to protect individuals' confidentiality. In the study area, there are 4,812 lifetime residential locations for cases (1,328 pre-menopausal and 3,484 post-menopausal residences), and there are 7,886 lifetime residential locations for controls (2,270 pre-menopausal and 5,616 post-menopausal residences).
Figure 1 Map of study area: Erie and Niagara counties with zip-code boundaries.
Figure 2 Geographic distribution of breast cancer in Western New York; Shown are all residential locations of breast cancer cases and controls by menopausal status included in the analysis. One dot indicates each residential location. The rectangular region was used as an approximate boundary of the study area instead of actual county boundary in Figure 1. East (x) and north (y) coordinates in projected Universal Traverse Mercator (UTM) miles.
We evaluated spatial patterns of risk surfaces based on the geographic distribution of lifetime residences in Figure 2. Risk surfaces based on the standardized difference between case and control densities were obtained, and areas with relatively higher case density were identified by menopausal status in Figure 3. In the figure, areas with difference greater than 2 standard deviations (SD) were portrayed as contours and areas exceeding critical values were portrayed as red images. Testing for significance was performed and reported as p-values. Those areas with standardized difference greater than 2 SD were quite different between pre- and postmenopausal breast cancer, although ranges of standardized the difference were similar; -6.37 to 4.43 for pre-menopausal, and -6.57 to 3.39 for postmenopausal breast cancer, respectively. There were about 29 rectangular grids in those areas greater than 2 SD for premenopausal, while about 59 grids for postmenopausal breast cancer. Further, the statistical significance of areas must be assessed in light of the fact that multiple areas are tested; these are statistically significant if the difference in density exceeds the critical value of 3.56 at α = 0.05 (determined by simulation, where random labelling of cases and controls is carried out). There is one small geographic area of special interest for pre-menopausal residences in the central and upper region of the city. When these are compared with the geographic location of clusters of birth and menarche residences identified previously [14], we found that the size and location of these areas are about the same as clusters of menarche residences, but somewhat smaller than the clusters of birth residences. Thus, it is more likely that the same individuals are in both clusters. For post-menopausal residences, no area exceeding the critical value was detected.
Next, we evaluated effects of other risk factors on the risk surfaces. To create age-adjusted risk surfaces, the standardized difference between case and control densities stratified by menopausal status and age groups was examined. Table 3 presents the variability of risk surfaces when stratified by menopausal status and age groups. While there were similar numbers of geographic areas greater than 2 SD regardless of menopausal status and age groups, we found areas (about 8 rectangular grids) greater than critical values only for residences for premenopausal women, aged 35–44, and the geographic location of those areas was identical to the areas identified in Figure 3. In addition, we evaluated the effects of one known risk factor, nulliparity, on those spatial patterns of risk surfaces. Risk surfaces were examined in two groups, nulliparious women and those with at least one child. We observed no difference in spatial patterns of risk surfaces for these two groups.; for premenopausal women, there were seven and two geographic areas greater than critical values for nulliparious and parous women, respectively, but none was detected for either group of postmenopausal residences (data not shown).
Table 3 Standardized difference of residences for premenopausal and postmenopausal breast cancer cases and controls by age groups;
Premenopausal Case
(n = 1328) Control
(n = 2270) Standardized differences No. of areasa > 2SD No. of areasa > critical valuesb
Ages 35–44
(n = 1889) 647 1242 -5.12–5.88 34 8
Ages 45–56
(n = 1709) 681 1028 -4.72–3.14 30 0
Postmenopausal Case
(n = 3484) Control
(n = 5616) Standardized differences No. of areas > 2SD No. of areas > critical values
Ages 40–64
(n = 4916) 2115 2801 -3.52–3.77 49 0
Ages 65–79
(n = 4184) 1369 2815 -4.50–3.15 40 0
a Areas refer to the rectangular grid overlaid on to the study area, b Critical values of 3.88 for premenopausal and 3.71 for postmenopausal residences.
Lastly, we were interested in evaluating temporal variability of risk surfaces; the standardized difference was obtained for each decade, for both pre- and post-menopausal residences, from the 1940s to 1990s (Table 4). While there was not much difference in the number of areas greater than 2 SD for both menopausal groups, we were able to find geographic areas greater than critical values in the 1960s through 1990s only for residences of premenopausal women. This is consistent with results from the above spatial analysis.
Table 4 Standardized difference of residences for premenopausal and postmenopausal breast cancer cases and controls by decades;
Decades Case Control Standardized differences No. of areasa > 2SD No. of areasa > critical valuesb
Premenopausal 1940s 36 86 -2.35–2.12 2 0
1950s 222 461 -2.45–3.15 9 0
1960s 341 633 -4.02–3.78 21 2
1970s 514 1025 -3.89–3.96 35 2
1980s 552 1005 -4.26–4.50 24 3
1990s 457 723 -4.06–4.14 32 5
Postmenopausal 1930s 389 676 -2.51–2.62 24 0
1940s 809 1253 -3.01–3.04 23 0
1950s 1158 1968 -3.66–2.91 22 0
1960s 1247 2072 -4.18–2.99 18 0
1970s 1100 1839 -4.98–3.13 21 0
1980s 960 1651 -3.84–2.85 23 0
1990s 943 1542 -4.40–3.57 22 0
a Areas refer to the rectangular grid overlaid on to the study area, b Critical values of 3.88 for premenopausal and 3.71 for postmenopausal residences.
Discussion and conclusions
This study explored the use of kernel density estimation methods to identify spatio-temporal patterns of risk surfaces in a case-control study of breast cancer. We used standardized differences between case and control densities to produce risk surfaces. These risk surfaces were assessed for both pre- and postmenopausal breast cancer. We found a general tendency for spatial clustering of breast cancer cases, and observed stronger evidence of geographic clustering for pre-menopausal women than for postmenopausal women. Geographic areas greater than 2 SD of the standardized difference were identified among lifetime residences for pre- and postmenopausal women, but more rigorous testing showed such evidence only for premenopausal residences. We were able to pinpoint geographic areas with relatively higher case densities, and to assess temporal variability of risk surfaces.
This study focused on the investigation of breast cancer risk associated with lifetime residential history using GIS-based exploratory spatial analyses. Since environmental risk factors are of continuing interest in breast carcinogenesis, this approach may be more relevant in understanding the environmental etiology of breast cancer [15,16]. Residential location has often been used as a proxy for exposures, and the relationships between residential environment and breast cancer risk have been a focus in recent epidemiologic studies [17-19]. Although the role of clustering analyses remains controversial in scientific advances in our understanding of disease etiology [20,21], these GIS-based exploratory spatial analyses are well suited for environmental epidemiologic investigations; this study demonstrated that smoothed risk surfaces created by kernel methods are useful for large sets of data in space and time, but also when the form of cluster is not well defined. This method can be applied to other epidemiologic analyses. For example, this GIS-based spatial analysis can be effectively used in exposure analyses and assessment, as previously used in identifying people potentially exposed to environmental risk factors [22,23].
Given that these are exploratory methods, however, it is meaningful to compare the strengths and limitations of different approaches when applied to epidemiologic data. We have chosen the standardized difference approach to represent risk surfaces, as opposed to other methods (such as risk ratios) that can be used to create risk surfaces, because it is more easily able to handle the difficulties that arise with small densities. Unsmoothed risk surfaces are relatively easy to manipulate, but they are sensitive to geographic scales, while the ratio of case to control density results in unstable risk surfaces due to small number problems. We were able to reduce this problem in creating risk surfaces based on a standardized difference approach. The selection of optimal bandwidths in the application of kernel methods to cluster detection, and comparison of different types of bandwidths, such as adaptive kernel, will be a subject of future study [24].
It is important to note that current approaches to obtaining density surfaces of lifetime breast cancer risk are limited in several ways. First, we are using residential locations of breast cancer cases and controls; we obtained the difference in densities (risk surfaces) based on their residential location to identify areas with relatively higher case density. We do not know actual exposures and breast cancer risk associated with residential locations. However, this study provides evidence of differential exposure opportunities among cases and controls for further epidemiologic assessment, since spatio-temporal clustering of residential locations in a life course may be an indicator of differential exposure opportunities and the subsequent risk of breast cancer. Another limitation of this analysis is that we were unable to incorporate length of residence into the model. We were able to visualize risk surfaces with different weights based on the actual length of residence of each individual, but there were difficulties in incorporating this information into risk surfaces due to a substantial degree of variability in the length of residence. However, we observed spatial patterns consistent with those in Figure 3, despite the greater variability, when we visualized risk surfaces with length of residence information.
Figure 3 Risk surfaces of pre- and postmenopausal breast cancer using standardized difference; Areas with standardized difference greater than 2 SD are portrayed as contours of 2, and areas exceeding critical value of 3.56 as red images. The rectangular region was used as an approximate boundary of the study area instead of actual county boundary in Figure 1. East (x) and north (y) coordinates in projected UTM miles.
There is also need for cautious interpretation of these results due to the potential for selection bias inherent in the study design, including factors such as non-participation, and missing and excluded residential location. Although we had a relatively stable population and about 40% of study participants were lifetime residents in the study area, there may be different geographic patterns among those included and excluded groups. We excluded missing residential information and residential locations outside of the study area. In addition, we had a rate of non-participation of about 30% in the survey. However, in our earlier study [14], we found that characteristics of subjects in the study area were not different from characteristics of subjects with missing residential information and from subjects excluded due to residence outside of the study area, and that the geographic distribution of participants was not different from that of non-participants. Although we used the same methods for both cases and controls, and interviewers were blinded as to case and control status, there may be recall bias in the lifetime self-reported residential history. We validated the accuracy and reliability of lifetime residential history, especially earlier residences. Our finding was that reported residence information was generally correct. The greater problem was missing data; for this we conducted searches of historical records, as we have described earlier. We are now in the process of obtaining birth addresses from birth certificates for additional validation.
This study has significant implications for further studies on environmental exposure and breast cancer. We had previously found strong evidence of clustering of residence in early life, especially residence at birth and menarche [14]. In these analyses, we were able to show evidence of clustering of lifetime residence. In addition, we found breast cancer cases were more mobile than controls, and that premenopausal participants were more mobile than postmenopausal pariticipants. Average years at current residence was between 11 and 12 years for premenopausal participating cases and controls, compared to 22 and 23 years for postmenopausal participating cases and controls, respectively. Taking findings from this and our previous study together, it appears that examination of exposure opportunities in the past and across the lifespan may be critical for understanding environmental exposures related to breast cancer. Exploring spatio-temporal patterns of lifetime residential history may provide a link between this potential exposure and breast cancer risk. This study provides a more comprehensive analytical framework for the analysis of environmental exposures in relation to breast cancer by considering these components, such as migration and latency periods, and these spatio-temporal patterns of lifetime residential history may be a key to the understanding the actual relationships between environmental exposure and subsequent breast cancer risk.
To provide more accurate measures of personal cumulative exposures based on complete lifetime residential history, further studies should take into account the different effects of time periods or timing of exposures; construction of lifetime cumulative risk surfaces with different weights for exposure sensitivity at different points in time is a potential improvement and explanation of the methods employed to date. This approach may help to provide an answer to the question of where, when, and what kinds of exposures have influenced individuals' risk for a particular disease. Further, there has been recent development of a GIS-based framework to examine spatio-temporal patterns of lifetime residential history; the geospatial lifeline concept and space-time information system (STIS) approach is a good example of this [25-27]. We are currently testing the feasibility of similarity and difference measures of an individual's lifetime residential history using this case and control data, since it would be a powerful tool to analyze the personal environmental exposures associated with lifetime residential history.
In summary, we found evidence of clustering of lifetime residence for premenopausal cases relative to controls, and a substantial degree of spatio-temporal variability in the risk surfaces, thus providing a working hypothesis on breast cancer and environmental exposures. Geographic areas with higher case densities need further epidemiologic investigation for potential relationships between lifetime environmental exposures and breast cancer risk. Examination of lifetime residential history provided additional information on geographic areas associated with higher risk; limiting exploration of chronic disease clustering to current residence may neglect important relationships between location and disease. Further studies on the relationship between disease risk and environmental exposures associated with lifetime residential history should be replicated in other settings.
Methods
The Western New York Exposure and Breast Cancer Study (WEB Study)
Data from a population-based case-control study of breast cancer in western New York (the WEB Study) were used for our analyses. Participants were women, age 35–79 who were residents of Erie and Niagara counties, with no history of cancer other than non-melanoma skin cancer; cases were women with incident, primary, pathologically confirmed breast cancer, diagnosed during the period 1996–2001, and controls were randomly selected and frequency matched to cases on age, race, and county of current residence. Details of the WEB study, including selection, ascertainment, in-depth interview processes, have been described previously [14]. We collected lifetime residential histories for 1,166 cases and 2,105 controls, identified 20,240 lifetime addresses, an average of approximately 6 addresses for each individual, from participating cases and controls. Analyses were restricted to those residential locations within the two counties of study area.
Geocoding of residential location
Geocoding of residential locations enables us to record each individual's locational information as x and y coordinates to be used in further spatial analyses. Address geocoding is a process that creates a theme based on the address data in a tabular form (event theme) and a reference feature theme (street map) to add point locations defined by the street address to the map. Matching depends not only on the quality of the reference theme, but also on the quality of the tabular data to be mapped. We used GDT/Dynamap 2000, an enhanced version of Topographically Integrated Geographic Encoding and Referencing system (TIGER), as a reference theme. In a study validating the positional accuracy of TIGER for the use in epidemiologic study [28], we found positional accuracy to be extremely high.
The overall matching rate for the 15,487 Erie and Niagara County residential locations was 82% (12,698); 91% of the matches were matched with complete information (good match), while about 9% of were estimated with partial information (interactive match) as detailed below. Geocoding success rates were lower for earlier residences, mainly due to more missing and partial information of earlier residences and changes in streets names and zip codes. However, it is important to note that there were few changes in street structure for this region during the time period of interest. There was, of course, addition of new streets, but existing streets were unchanged for the most part, and thus we found that the process of geocoding using current information was not inappropriate. We utilized various resources, including historic city directories with address information for residents, historical maps, and commercial address databases, to find missing residence information, and we developed several strategies to improve matching rates [14]. In addition, for residences where we had a known street name but no known street number and if the total length of the street was one kilometre or less, we estimated the residence location as the midpoint of the street.
Kernel density estimation methods
Kernel density estimation methods have been used for disease mapping and for the detection of geographical location of clusters in epidemiologic settings [10,12,29]. A general form of the kernel k is defined as, , and by averaging over these individual kernel functions, we obtain the kernel density estimator, ;
where x is the location for density estimation, xi is the observed point location, n is the number of points, and h is a smoothing parameter that regulates the degree of smoothness. Kernel functions are symmetric around zero and integrate to one. Both kernel and kernel density estimator are density functions; thus ∫kx (x)dx = 1 and ∫fh(x)dx = 1. For the two-dimensional kernel density estimator:
While the estimation of intensity for one point pattern may show patterns of high and low risk areas, this is of limited utility in epidemiologic applications because it may largely reflect the pattern of population distribution [30]. The ratio of case to control density can be more effectively used in epidemiologic settings [9,31]. To estimate the ratio of the two density estimates, the ratio of case to control density, is used, where and are estimates of the intensity for cases and controls, respectively. Smoothed risk surfaces using the ratio or log ratio of case to control density can be less effective with small sample sizes; it is possible to have spurious, high risk areas in the application of the ratio of densities when the value of the control density is too small.
The relative difference between two densities provides an alternative way to assess the spatial variability of risk surfaces. Using the square root variance stabilizing transformation and the standardized difference, this measure allows us to identify areas with differences between case and control density exceeding two standard deviations [32]. The standardized difference between case and control densities is obtained by taking the square root of the case density minus the square root of the control density, and dividing by the standard deviation of the difference between the densities.
where
Analytical procedures
The first step in creating smoothed risk surfaces using kernel methods was to create reference grids and overlay the study area with them. We obtained the smoothed intensity for both cases and controls by calculating the distance between each point on the reference grid and the locations of breast cancer cases and controls. We used the quartic kernel to estimate the intensity of points at each grid point, although the choice of kernel type is not crucial as long as the kernel is symmetrical [33]. We applied equal, fixed bandwidths for both cases and controls using equation (1) because the objective is to describe overall patterns of the underlying spatial distribution [9]. The risk surface was obtained by forming the difference in densities based on equation (2).
The above analyses were repeated with varying bandwidths because the choice of appropriate bandwidth is one of the primary concerns of the kernel method. Although a subjective choice made from a range of values is commonly used [33], selections of bandwidth were made here on the basis of several factors. To avoid subjectivity, we first began with the commonly used optimal bandwidth designed to minimize the estimated mean square error [34]. Other established methods, such as cross validation, were tried and these resulted in small bandwidths because of the large sample size [34,35]. In addition, to take into account the spatial distribution of point patterns and to avoid problems associated with fixed bandwidths, we initially selected bandwidths based on the average distance among points. Since the size of the study area is approximately 30 miles in width and 60 miles in length, we selected a one-mile radius as an initial bandwidth of the kernel, with a range of 0.5 to 10 miles. In summary, we searched over a range of bandwidths and ultimately chose a two-mile bandwidth as a balance between over-smoothing and under-smoothing. Increasing bandwidth implies increasing the amount of smoothing in the estimate. A larger bandwidth results in very smooth density surfaces, while too small a bandwidth produces noisy density estimates. These issues are important in applications of our case-control data; the geographic distribution of cases and controls is dependent on the population distribution, which is greatly concentrated in urban areas and is sparser in rural areas. Thus, the use of a small bandwidth less than two miles resulted in an unrecognizable pattern of density.
Risk surfaces based on the ratio and difference in density surfaces between cases and controls were implemented in a GIS and S-Plus environment [36]. In addition to finding the risk surface associated with the standardized difference between densities, we tested the significance of the difference surfaces between cases and controls by Monte Carlo simulation. Under the null hypothesis of constant risk in the study area, we obtained critical values. We first randomly assigned case and control status to each of the case and control locations, based on the proportion of cases and controls among the set of cases and controls. Then we obtained the 95th percentile for the maximum difference between case and control densities from 999 simulations.
Authors' contributions
All authors have participated fully in the conduct of this study, the acquisition of data, the analysis and/or the preparation of the manuscript in a substantial manner. DH performed the statistical analysis and drafted the manuscript; PAR and JLF provided critical review and input. JLF, JEV, MAB, JN, PM and MT participated in the design of the study, participated in interpretation, as well as in data acquisition efforts. All authors read and approved this manuscript.
Acknowledgements
This work was supported in part by NIH Grants 1R01 ES09816-01, 1R21 CA87138-01, and U.S. Army Medical Research Grants DAMD17-03-1-0475, DAMD17-00-1-0417.
==== Refs
New York State Department of Health The New York State Cancer Surveillance Improvement Initiative 1998 Albany, NY,
Cromley EK McLafferty SL GIS and Public Health 2002 New York, The Guilford Press
Khan OA Skinner R Geographic Information Systems and Health Applications 2003 Hershey, PA, Idea Group Publishing
Rushton G Lolonis P Exploratory spatial analysis of birth defect rates in an urban population Stat Med 1996 15 717 726 9132899 10.1002/(SICI)1097-0258(19960415)15:7/9<717::AID-SIM243>3.0.CO;2-0
Timander LM McLafferty S Breast cancer in West Islip, NY: a spatial clustering analysis with covariates Soc Sci Med 1998 46 1623 1635 9672400 10.1016/S0277-9536(97)10131-9
Lawson A Biggeri A Bohning D Lesaffre E Viel J Bertollini R Disease Mapping and Risk Assessment for Public Health 1999 Chichester, John Wiley and Sons
Rushton G Gatrell, AC and Löytönen, M Improving the geographic basis of health surveillance using GIS GIS and Health 1998 London, Taylor and Francis 63 80
Elliott P Wakefield JC Best NG Briggs DJ Spatial Epidemiology: Methods and Applications 2000 Oxford, Oxford University Press
Kelsall JE Diggle PJ Non-parametric estimation of spatial variation in relative risk Stat Med 1995 14 2335 2342 8711273
Gatrell AC Bailey TC Diggle PJ Rowlingson BS Spatial point pattern analysis and its application in geographical epidemiology Transactions of the Institute of British Geographers 1996 21 256 274
Paulu C Aschengrau A Ozonoff D Exploring associations between residential location and breast cancer incidence in a case-control study Environ Health Perspect 2002 110 471 478 12003750
Sabel CE Gatrell AC Loytonen M Maasilta P Jokelainen M Modelling exposure opportunities: estimating relative risk for motor neurone disease in Finland Soc Sci Med 2000 50 1121 1137 10714932 10.1016/S0277-9536(99)00360-3
Potischman N Troisi R In-utero and early life exposures in relation to risk of breast cancer Cancer Causes Control 1999 10 561 573 10616825 10.1023/A:1008955110868
Han D Rogerson PA Nie J Bonner MR Vena JE Vito D Muti P Trevisan M Edge SB Freudenheim JL Geographic clustering of residence in early life and subsequent risk of breast cancer (United States) Cancer Causes Control 2004 15 921 929 15577294 10.1007/s10552-004-1675-y
Brody JG Rudel RA Environmental pollutants and breast cancer Environ Health Perspect 2003 111 1007 1019 12826474
DeBruin LS Josephy PD Perspectives on the chemical etiology of breast cancer Environ Health Perspect 2002 110 Suppl 1 119 128 11834470
Gammon MD Neugut AI Santella RM Teitelbaum SL Britton JA Terry MB Eng SM Wolff MS Stellman SD Kabat GC Levin B Bradlow HL Hatch M Beyea J Camann D Trent M Senie RT Garbowski GC Maffeo C Montalvan P Berkowitz GS Kemeny M Citron M Schnabe F Schuss A Hajdu S Vincguerra V Collman GW Obrams GI The Long Island Breast Cancer Study Project: description of a multi-institutional collaboration to identify environmental risk factors for breast cancer Breast Cancer Res Treat 2002 74 235 254 12206514 10.1023/A:1016387020854
Lewis-Michl EL Melius JM Kallenbach LR Ju CL Talbot TO Orr MF Lauridsen PE Breast cancer risk and residence near industry or traffic in Nassau and Suffolk Counties, Long Island, New York Arch Environ Health 1996 51 255 265 8757405
McKelvey W Brody JG Aschengrau A Swartz CH Association between residence on Cape Cod, Massachusetts, and breast cancer Ann Epidemiol 2004 14 89 94 15018880 10.1016/S1047-2797(03)00120-0
Rothman KJ A sobering start for the cluster busters' conference Am J Epidemiol 1990 132 S6 13 2356837
Wartenberg D Investigating disease clusters: why, when and how ? Journal of the Royal Statistical Society A 2001 164 13 22 10.1111/1467-985X.00181
Ward MH Nuckols JR Weigel SJ Maxwell SK Cantor KP Miller RS Identifying populations potentially exposed to agricultural pesticides using remote sensing and a Geographic Information System Environ Health Perspect 2000 108 5 12 10622770
Wartenberg D Greenberg M Lathrop R Identification and characterization of populations living near high-voltage transmission lines: a pilot study Environ Health Perspect 1993 101 626 632 8143596
Brunsdon C Estimating probability surfaces for geographical point data: an adaptive kernel algorithm Computers and Geosciences 1995 21 877 894 10.1016/0098-3004(95)00020-9
Jacquez GM Greiling DA Kaufmann AM Design and implementation of a space-time intelligence system for disease surveillance Journal of Geographical Systems 2005 7 7 23 10.1007/s10109-005-0147-6
Rushton G Elmes G McMaster R Considerations for improving geographic information system research in public health URISA Journal 2000 12 31 49
Mark DM Egenhofer MJ Bian L Hornsby KE Rogerson PA Vena JE Flahaut A, Toubiana L and Valleron AJ Spatiotemporal GIS analysis for environmental health: solutions using geospatial lifeline Geography and medicine: GEOMED '99 2000 Paris, Elsevier 65 78
Bonner MR Han D Nie J Rogerson P Vena JE Freudenheim JL Positional accuracy of geocoded addresses in epidemiologic research Epidemiology 2003 14 408 412 12843763
Bithell JF An application of density estimation to geographical epidemiology Stat Med 1990 9 691 701 2218172
Bailey TC Gatrell AC Interactive Spatial Data Analysis 1995 Harlow, England, Longman Scientific & Technical
Kelsall JE Diggle PJ Spatial variation in risk of disease: a nonparametric binary regression approach Applied Statistics 1998 47 559 573
Bowman AW Azzalini A Applied Smoothing Techniques for Data Analysis 1997 Oxford, Clarendon Press
Silverman BW Density Estimation for Statistics and Data Analysis 1986 London, Chapman and Hall
Diggle PJ Statistical Analysis of Spatial Point Patterns 1983 London, Academic Press
Bowman AW An alternative method of cross-validation for the smoothing of density estimates Biometrika 1984 71 353 360
Venables WN Ripley BD Modern Applied Statistics with S-Plus 1998 New York, Springer
| 15826315 | PMC1097750 | CC BY | 2021-01-04 16:39:06 | no | Int J Health Geogr. 2005 Apr 12; 4:9 | utf-8 | Int J Health Geogr | 2,005 | 10.1186/1476-072X-4-9 | oa_comm |
==== Front
J Autoimmune DisJournal of Autoimmune Diseases1740-2557BioMed Central London 1740-2557-2-41586971310.1186/1740-2557-2-4ResearchN-acetyl-L-cysteine ameliorates the inflammatory disease process in experimental autoimmune encephalomyelitis in Lewis rats Stanislaus Romesh [email protected] Anne G [email protected] Avtar K [email protected] Inderjit [email protected] Department of Biostatistics, Bioinformatics & Epidemiology, Medical University of South Carolina, Charleston, SC, USA2 Department of Pediatrics, Medical University of South Carolina, Charleston, SC, USA2005 3 5 2005 2 4 4 1 4 2005 3 5 2005 Copyright © 2005 Stanislaus et al; licensee BioMed Central Ltd.2005Stanislaus 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.
We report that N-acetyl-L-cysteine (NAC) treatment blocked induction of TNF-α, IL-1β, IFN-γ and iNOS in the CNS and attenuated clinical disease in the myelin basic protein induced model of experimental allergic encephalomyelitis (EAE) in Lewis rats. Infiltration of mononuclear cells into the CNS and induction of inflammatory cytokines and iNOS in multiple sclerosis (MS) and EAE have been implicated in subsequent disease progression and pathogenesis. To understand the mechanism of efficacy of NAC against EAE, we examined its effect on the production of cytokines and the infiltration of inflammatory cells into the CNS. NAC treatment attenuated the transmigration of mononuclear cells thereby lessening the neuroinflammatory disease. Splenocytes from NAC-treated EAE animals showed reduced IFN-γ production, a Th1 cytokine and increased IL-10 production, an anti-inflammatory cytokine. Further, splenocytes from NAC-treated EAE animals also showed decreased nitrite production when stimulated in vitro by LPS. These observations indicate that NAC treatment may be of therapeutic value in MS against the inflammatory disease process associated with the infiltration of activated mononuclear cells into the CNS.
EAEMacrophagesinfiltration N-acetyl-L-cysteineCNS
==== Body
1. Introduction
Multiple sclerosis (MS) is a chronic demyelinating disease marked by focal destruction of myelin, resulting in the loss of oligodendrocytes and axons accompanied by an inflammatory disease process [1-3]. Experimental autoimmune encephalomyelitis (EAE) is an animal model of MS. Both MS and EAE are initiated by a T-cell mediated autoimmune response (CD4+ and CD8+) against myelin components followed by induction of inflammatory mediators (chemokines and cytokines) that in turn define the pattern of perivascular migration of activated T-cells and mononuclear cells into the CNS [1-4].
The sequence of events associated with loss of oligodendrocytes and myelin in MS and EAE are not precisely understood. A complex interaction between the mediators released by infiltrating cells and brain endogenous activated glial cells (astrocytes and microglia) are believed to contribute towards the inflammatory disease process and tissue damage [1-3,5-7]. Numerous studies have documented the expression of proinflammatory cytokines (TNF-α, IL-1β, and IFN-γ) in EAE and MS tissue and increased levels of IFN-γ and TNF-α levels in CNS or plasma appear to predict relapse in MS [1-3,8]. On the other hand, enhanced expression of anti-inflammatory cytokines (IL-4, IL-10 and TGF-β) appears to mediate disease remission [1-3,9]. In MS brain, expression of iNOS by activated astrocytes, microglia and macrophages is associated with the demyelinating regions [10-13]. The NO derived from iNOS as ONOO- (a reaction product of NO and O2-) is thought to play a role in the pathobiology of MS and EAE. Peroxynitrite (ONOO-) is able to modify proteins, lipids and DNA resulting in damage to oligodendrocytes and myelin [1-3].
In spite of extensive research to develop pharmacotherapeutic agents to ameliorate or reduce the number of exacerbations and subsequent progression of neurological disability in MS, only a few therapies are available. Presently, IFN-β [14] and glatiramer acetate [15] are used in treatment of MS but the therapeutic efficacy of these compounds is limited by significant side effects. Recent studies from our laboratory [16,17] and others [18] report the potential of HMG-CoA reductase inhibitors (statins) in attenuating the disease process in EAE. The efficacy derives from a shift from an inflammatory Th1 response towards an anti-inflammatory Th2-biased response [16,18,19], blocked infiltration of mononuclear cells into CNS [20] and attenuation of the induction of proinflammatory cytokines (TNF-α, IFN-γ) and iNOS in the CNS of EAE animals [17,20].
Reactive oxygen species (ROS) and reactive nitrogen species (RNS), generated as a result of the inflammatory process, are believed to play a role in the pathobiology of EAE and MS [10,12,13]. Cell culture studies showed that NAC, a potent antioxidant, inhibited induction of TNF-α and iNOS and NO production in peritoneal macrophages, C6 glial cells and primary astrocytes, and blocked the activation of NFκB in peritoneal macrophages [21]. Accordingly, oral administration of the oxidant scavenger NAC was found to attenuate EAE clinical disease [22]. The present studies were designed to elucidate the mechanism of observed therapeutic efficacy of NAC against EAE. These studies document that NAC treatment inhibited the clinical disease by attenuating multiple events in EAE disease such as shifting the immune response from a Th1 bias, increasing IL-10 cytokine production by splenocytes, attenuating transmigration of mononuclear cells, and inhibiting induction of proinflammatory cytokines (TNF-α, IL-1β, IFN-γ) and iNOS in the CNS. Taken together these results suggest NAC may be of therapeutic value for cell-mediated autoimmune diseases such as multiple sclerosis.
2. Materials and methods
Chemicals
Myelin basic protein (MBP) isolated from guinea pig brain and complete Freund's adjuvant (CFA) and pertussis toxin were obtained from Sigma (St. Louis, MO). N-acetyl-L-cysteine (NAC) was obtained from Calbiochem (USA).
EAE induction and treatment with NAC in Lewis rats
Experiments were performed on female Lewis rats (Harlan Laboratory, USA) weighing 250–300 g. Animals were housed in the animal care facility of the Medical University of South Carolina, USA, throughout the experiment and provided with food and water ad libitum. All experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee. EAE was induced by subcutaneous injection of 50 μg of MBP (per animal) emulsified in complete Freund's adjuvant in the region of the footpad of the hind leg on day 1 followed by a booster injection of the same on day 7. Additionally, animals received 200 ng of pertussis toxin on days 0 and 1. Clinical signs in these rats manifest as ascending paralysis resulting in EAE in most animals. The clinical signs of EAE were scored by a masked investigator as 0 = normal; 1 = piloerection; 2 = loss in tail tonicity; 3 = hind leg paralysis; 4 = paraplegia; and 5 = moribund. NAC treatment was started on the first day of immunization (day 1) and continued daily for the duration of the experiment. One group of rats induced for EAE (n = 15) was given intraperitoneal injections of NAC (150 mg/kg body weight in PBS with pH adjusted to 7.2 with NaOH). The second group of rats (n = 15) was induced for EAE and treated with the vehicle (PBS). Animals receiving only CFA were used as the control group (n = 15). Untreated EAE animals were sacrificed at clinical stage 4 (paraplegia) or 5 (moribund) according to approved protocol. NAC treated animal group was sacrificed at their peak clinical disease, which was an average clinical score of 3, as determined from preliminary studies. Tissue for histology and immunohistochemistry and splenocytes were recovered for analysis.
Histopathology-Immunohistochemistry
The lumbar region of the spinal cord was dissected and carefully processed for histological and immunohistological examination (n = 12). Spinal cords were fixed in 10% buffered formalin (Stephens Scientific, Riverdale, NJ), embedded in paraffin and sectioned at 4 μm thickness. Sections were then stained for various cytokines and cell markers.
Immunohistochemistry for TNF-α, IFN-γ, IL-1β, iNOS and nitrotyrosine was done as previously described [17]. Sections were incubated with appropriate antibodies (1:100) overnight followed by fluorochrome conjugated secondary IgG antibody (1:100, Sigma, St. Louis, MO) and mounted with Fluoromount G (EMS, Fort Washington, PA). Non-immune IgG was used as control primary antibody. Sections were also incubated with TRITC or FITC conjugated IgG without the primary antibody as negative control. Nuclear staining was performed using DAPI (Sigma, St. Louis, MO) and hematoxylin and eosin (H&E) staining was performed as described by Kiernan, J.A (1990). All the sections were analyzed using an Olympus microscope (Olympus BX60, Opelco, Dulles, VA) and images were captured using a digital video camera (Olympus U-CMAD-2, Optronics, Galeta, CA) and Adobe Photoshop (Adobe Systems, CA).
Quantitative analysis of infiltrating cells
Infiltrating cells labeled with either ED1 or DAPI were analyzed using Image-Pro Plus 4.0 (Media Cybernetics, Maryland, USA) software. Individual sections were analyzed and the mean and SD were calculated for each group (n = 12). The group means were compared and the significance of difference was determined. A p value of <0.05 was considered significant. This analysis was done using the Regression Data Analysis tool of Microsoft Excel 4.0 (Microsoft, Redmount, WA).
Splenocyte Isolation and Cell Culture
Splenocytes were isolated from each animal group (Control, EAE, EAE+NAC) (n = 6) using Lympholyte®-Rat (Cedarlane Laboratories Ltd., Hornby, Canada) density separation medium according to manufacturer's instruction. The cell concentration in the suspension was adjusted to 2 × 107 nucleated cells per ml or less, layered on Lympholyte®-Rat density separation medium, and centrifuged for 20 min at 1000 g – 1500 g at room temperature. The interface formed after the centrifugation was then extracted using a Pasteur pipette, and transferred to a new centrifuge tube. The transferred cells were then diluted with medium, and centrifuged at 800 g for 10 min, washed twice with media, and cultured in 24-well plates at a concentration of 5 × 106 cells/ml. The cells were then stimulated in vitro with MBP (20 μg/mL), LPS (1 μg/mL), or PHA (10 μg/mL), (Sigma, St. Louis, MO, USA), or without any stimulants for 48 hrs. Each treatment was performed in triplicate. At the end of the 48 hr. incubation period, supernatants were collected and used for the measurement of cytokines and nitrite.
ELISA
Cytokines (IFN-γ and IL-10) were detected in culture supernatants using commercially available OptEIA™ kits from PharMingen (San Diego, CA, USA) according to manufacturer's instructions. The assay procedure is as follows: 96-well microplates were coated with capture antibody diluted in coating buffer overnight at 4°C. Plates were then washed and blocked with assay diluent (PharMingen, San Diego, CA, USA) for 1 hr at room temperature. Blocked plates were then washed, and the standards and samples added to the wells and incubated for 2 hr. at room temperature. At the end of incubation, plates were washed and working detector (detection antibody + Avidin-HRP) was added to the wells and incubated for 1 hr. at room temperature. Following incubation, plates were washed and TMB substrate reagent was added (PharMingen, San Diego, CA, USA) to the wells for 30 min. at room temperature in the dark. At the end of the incubation, stop solution (1 M H3PO4) was added, and absorbance read at 450 nm using a Spectramax® microplate spectrophotometer (Molecular Devices, Sunnyvale, CA, USA).
Nitrite measurement
Nitrite levels were determined on isolated splenocytes with Griess reagent as previously described [23] with minor modifications. One hundred μl of culture supernatant was allowed to react with 100 μl of Griess reagent and incubated at room temperature for 15 min. The optical density of the assay samples was measured at 570 nm using a 96-well plate Spectramax® microplate reader with SOFTMAX® software (Molecular Devices, Sunnyvale, CA, USA). Fresh culture media served as the blank in all experiments. Nitrite concentrations were calculated from a standard curve derived from the reaction of NaNO2 in the assay.
3. Results
Effect of NAC on the Clinical Signs of Rats
Our goal was to investigate the effect of NAC on rats induced for acute EAE. In the Lewis rat model MBP induces an acute monophasic disease progression. As shown in Fig. 1, clinical signs of EAE were evident in MBP-treated Lewis female rats from the 8th day after first immunization inducing an acute monophasic disease progression resulting in paraplegia (clinical score of 4) or moribund state (clinical score of 5) on or around the 12th day. However, the control animals receiving only complete Freund's adjuvant did not show any disease symptoms (Fig. 1). Animals induced for EAE but given only the vehicle closely followed the disease progression of MBP-treated rats. Treatment of MBP-injected rats with NAC, administered from the first day of immunization, protected the rats by attenuating the severity of disease progression (Fig. 1). NAC treated animals had milder clinical signs (average clinical score of 3 as compared to 5 for EAE).
Figure 1 The protective effect of NAC on the clinical signs of MBP induced EAE in female Lewis rats. EAE was induced as described in Materials and Methods. Data are given as average clinical disease score where 0 = normal; 1 piloerection; 2 = loss in tail tonicity; 3 = hind leg paralysis; 4 = paraplegia and 5 = moribund. Each group i.e., MBP (closed squares), MBP + NAC (open circles) treated and control group (closed diamonds) had 15 animals (n = 15). EAE animals reached a peak clinical score of 4 or 5 on or around the 11th day after first immunization and were sacrificed according to approved protocol. NAC treated animals had milder clinical signs (average clinical score of 3 as compared to 4 or 5 for EAE). Control group did not show any clinical symptoms (clinical score = 0).
Effect of NAC on the infiltration of inflammatory cells into the spinal cord
The neuropathological changes in EAE and MS are associated with the blood brain barrier breakdown and infiltration by mononuclear cells [24,25]. Clinical disease in EAE has been shown to correlate with the invasion of CNS by mononuclear cells. These studies demonstrate that MBP-induced EAE results in the induction of inflammatory disease, and treatment with NAC provides protection against the EAE disease process. Therefore, in order to understand the mechanism of therapeutic efficacy in EAE, we studied the effect of NAC on the invasion of mononuclear cells into the CNS in the EAE model.
The spinal cords of rats induced for EAE had heavy mononuclear inflammatory infiltrates on the meningeal surfaces, perivascular areas and interstitial areas as seen by H&E staining (Fig. 2a). EAE animals treated with NAC showed infiltration of the CNS by inflammatory cells but not to the extent as that seen in EAE animals. Further analysis of the cell infiltrates was performed to identify the major cell type infiltrating the CNS in addition to the T-cells. Immunohistochemical methods using ED1 (monocyte/ macrophage marker) and DAPI (for nucleated cells) were performed. As seen in Fig. 3, EAE animals showed the most infiltration by ED1 positive cells. In contrast the NAC-treated animals showed significantly less infiltration by ED1 positive cells (reduced by an average of 46 percent). Quantitative analysis of cell infiltrates into the CNS showed a significant amount of nucleated as well as ED1 positive cells in the CNS of EAE animals (Fig. 3b). In contrast, cell infiltration into the CNS of treated animals was significantly less than that seen in EAE animals (reduced by an average of 45 percent).
Figure 2 Inflammation and demyelination in sections of lumbar spinal cord from control, EAE and EAE + NAC (n = 12) treated Lewis rats. The spinal cords were isolated at peak manifestation of the disease (i.e. clinical score 5 in EAE and 3 in EAE + NAC treated animals). Photomicrographs represent regions from a) anterior cord b) central region and c) lateral cord. BV-denotes blood vessel. A. H&E staining of cross-sections of lumbar spinal cord. Compared to the control group, Lewis rats with EAE demonstrated gliosis (single arrow) and perivascular (double arrows), meningeal and interstitial chronic inflammatory infiltrates. These effects were attenuated in sections from EAE+NAC treated animals. B. LFB-PAS staining of cross sections of lumbar spinal cord from control, EAE, and EAE+NAC treated Lewis rats. Compared to the control animals, the interface of normal to demyelinating plaque (arrowhead) is notable in sections from the EAE group of animals. Myelin persists in the plaque as globules in the cytoplasm of macrophages. The EAE+NAC group showed demyelination, but to a lesser degree than that seen in the untreated group.
Figure 3 Quantification of the inflammatory infiltrates by immunostaining of Lewis rat spinal cord (n = 12). Top panel: The spinal cords were isolated when the animals were showing maximum clinical symptoms (i.e. for the EAE group clinical score of 5 and EAE+NAC clinical score of 3). The sections were stained for ED1 (macrophage/monocyte -green) and nuclei labeled with DAPI (blue) as described in materials and methods. Spinal cords of rats induced for EAE demonstrated increased numbers of ED1 positive cells (green) and other glial and inflammatory cells (blue) in the CNS. Original magnification 200×. Bottom panel: Quantification of the infiltrates showed significant numbers of glial and inflammatory cells (A: DAPI; nuclei stained blue) of which many also were positive for macrophage/monocyte (B: ED1; stained green) in the spinal cord of EAE animals as compared to control and EAE + NAC treated animals.
Effect of NAC on the expression of pro-inflammatory cytokines and iNOS in the spinal cord
Since the major source of IL-1β in EAE is monocytes/macrophages, as further evidence for macrophage infiltration we examined the expression of IL-1β in the CNS. As evidenced by Fig. 4c, expression of IL-1β was evident in the CNS of EAE induced animals and to a far lesser degree in the NAC treated animals. IL-1β expression was also co-localized to ED1 positive cells in EAE animal spinal cords (data not shown). We also examined the expression of proinflammatory cytokines (TNF-α and IFN-γ), iNOS and nitrotyrosine in the spinal cord sections from control, EAE, and NAC-treated EAE rats using immunohistochemistry. As seen in figure 4a–e, MBP-induced EAE resulted in the expression of TNF-α, IFN-γ, IL-1β, IFN-γ, iNOS and nitrotyrosine. NAC treatment of EAE blocked the induction of these cytokines, iNOS, and nitrotyrosine similar to control animals.
Figure 4 Immunofluorescent detection of IFN-γ, TNF-α, IL-1β, iNOS and nitrotyrosine in the CNS of female Lewis rats. The spinal cords were isolated when the animals were showing maximum clinical symptoms (i.e. for the EAE group clinical score of 5 and EAE+NAC clinical score of 3). Immunostaining was performed in spinal cord sections (n = 12) of female Lewis rats as described in Materials and Methods. EAE sections showed intense staining for IFN-γ, TNF-α, iNOS and nitrotyrosine with less intense staining for IL-1 β. Control and EAE + NAC sections showed minimal staining (original magnification 200×).
IFN-γ, IL-10 and nitrite production by splenocytes from EAE and treated animals
In vitro splenocytes assays were performed to elucidate whether NAC treatment could cause a shift to Th2-type T-cell activity. In order to examine this effect, we studied the effect of NAC on the major Th2 cytokine in the EAE disease process, IL-10. Splenocytes (8 × 105 cells per well) were obtained from Control, EAE, and EAE + NAC treated rats. Cells were stimulated in vitro with PHA (10 μg/ml, a & b), MBP (20 μg/ml, a & b) or LPS (1 μg/ml, c) for 48 hrs. The levels (pg/ml) of IFN-γ and IL-10 in culture supernatants were measured using ELISA kits. As seen in Fig. 5 there was a significant increase in IFN-γ (5a) and decrease in IL-10 (5b) in splenocytes from untreated EAE animals. NAC treatment reduced IFN-γ production by splenocytes (by 59% for PHA and 40 % for MBP) and up-regulated IL-10 production by EAE splenocytes (by 31% for PHA and 34% for MBP). Culture supernatants were collected and accumulated nitrite, a stable product of NO production, was measured using Griess reagent. NAC treatment also inhibited the production of nitrite by LPS-stimulated splenocytes by 71% as compared to splenocytes from EAE animals. These studies indicate that NAC treatment reduced IFN-γ, a proinflammatory Th1 cytokine and increased IL-10, an anti-inflammatory cytokine.
Figure 5 IFN-γ, IL-10 and nitrite production by splenocytes from control, EAE and treated animals. Splenocytes (8 × 105 cells per well) were obtained from Control, EAE, and EAE + NAC treated rats when they were showing maximum clinical signs, and were stimulated in vitro with PHA (10 μg/ml, a & b), MBP (20 μg/ml, a & b) or LPS (1 μg/ml, c) for 48 hrs. Each treatment was performed in triplicate for each animal group (n = 6). a and b: The levels (pg/ml) of IFN-γ and IL-10 in culture supernatants were measured using ELISA kits. There was a significant increase in IFN-γ and decrease in IL-10 in splenocytes from untreated EAE animals stimulated with PHA and MBP. EAE + NAC splenocytes showed reduced IFN-γ production whereas IL-10 production was increased. c: Culture supernatants were collected and accumulated nitrite, a stable product of NO production, was measured using Griess reagent. LPS-stimulated EAE splenocytes showed significantly higher levels of nitrite as compared to control and this was reduced with NAC treatment.
4. Discussion
The evidence presented in this paper demonstrates that NAC treatment reduced the inflammatory monocyte/macrophage cells in the CNS of Lewis rats with acute monophasic EAE. This in turn results in protection both in terms of clinical and histopathological changes. These conclusions are based on the following observations. 1) NAC treatment of EAE rats reduced the severity of EAE clinical symptoms, 2) attenuated the infiltration of mononuclear cells into the CNS of EAE rats, 3) blocked the induction of proinflammatory cytokines, iNOS and nitrotyrosine in the CNS, and 4) decreased proinflammatory Th1 cytokine responses (IFN-γ) from ex vivo splenocytes while increasing anti-inflammatory cytokine production (IL-10), and decreasing NO production in LPS-stimulated splenocytes.
The infiltration of activated mononuclear cells into the CNS of EAE is a critical event in the progression of the disease [26]. We have shown both qualitatively and quantitatively that ED1 positive leukocytes, namely macrophage/monocytes, were significantly decreased in animals treated with NAC as compared to the EAE animals. This decrease also correlated with the amelioration of clinical disease in female Lewis rats. As compared to our previous studies with lovastatin, NAC was not as effective in blocking the transmigration of inflammatory cells (NAC reduced by an average of 46%, while lovastatin reduced by 85%) and hence did not delay the onset of disease as was achieved with lovastatin treatment (EAE, EAE + NAC onset day 8 versus EAE + lovastatin onset day 11). However, NAC reduced the clinical scores to the same levels as those obtained with lovastatin (both had clinical scores maximum of 3). Other studies have also shown a correlation between macrophage infiltration and EAE clinical disease [27]. Inflammatory cytokine expression (IFN-γ, IL-1β, and TNF-α) was also inhibited in the CNS of EAE animals treated with NAC. As a consequence, inhibition of IFN-γ expression in NAC treated animals could in turn result in the reduced expression of MHC II molecules thereby inhibiting the proliferation of T-lymphocytes as has been shown with statins [28,29], copolymer 1 [30] and IFN-β [31].
Evidence indicates that iNOS while not a crucial factor for induction of EAE, plays a major role in the progression of the disease. The critical factors is the amount of NO produced that tips the balance in favor or against the pathogenesis of EAE [32]. The peroxynitrite (ONOO-) produced by reaction of NO and O2- can damage membranes and cells by nitrosylation of lipids, proteins and nucleic acids. The induction of IL-1β and activation of NFκB were shown to precede the induction of iNOS in ED1+ cells [33]. Here we report that NAC blocked the induction of IL-1β in the CNS of EAE animals. Ex vivo studies using splenocytes isolated from control, EAE and EAE+NAC treated animals showed that NAC inhibited IFN-γ production while increasing IL-10 production. These changes coincided with a decreased NO production in the cultured splenocytes. NAC treatment was not as effective as lovastatin in altering cytokine production, but the reduction in nitrite was identical. NAC treatment reduced IFN-γ production by splenocytes (NAC by 59% and 40%, LOV by 76% and 60% for PHA and MBP respectively) and up-regulated IL-10 production by EAE splenocytes (NAC by 31% and 34%, LOV by 350% and 490% for PHA and MBP respectively). NAC treatment also inhibited the production of nitrite by LPS-stimulated splenocytes by 71% as compared to splenocytes from EAE animals. These studies indicate that NAC treatment inhibited a proinflammatory Th1 biased cytokine response (IFN-γ) while promoting an increase in IL-10, an anti-inflammatory cytokine. Similar shifts from Th1 cytokine profile to Th2 have been correlated with disease recovery or improvement in both EAE and MS [16,18,19,34-37].
The brain is particularly vulnerable to oxidative stress due to its high consumption of oxygen and glucose, enrichment in unsaturated fatty acids that are subject to oxidation, and presence of regions enriched in iron and ascorbate that are potent pro-oxidants for brain membranes. Moreover, higher levels of glucose upregulate the neuroinflammatory process measured as induction of iNOS and NO production [38]. Coupled with the relatively reduced antioxidant defenses in the brain, exposure of brain cells to reactive oxygen or nitrogen species can be detrimental and is thought to contribute to the pathogenesis of many brain disorders [39]. Oxidative stress is important in the etiology of EAE and is thought to contribute directly to CNS damage [7,40]. N-acetyl-L-cysteine (NAC) as cysteine, a precursor of glutathione, is a potent anti-oxidant. By scavenging superoxide radicals, metallothionein and other antioxidants such as cysteine, N-acetyl-cysteine and glutathione offer neuroprotection [41]. In vivo NAC enhances hippocampal neuronal survival after transient forebrain ischemia in rats [42]. Partial protection of neurons from the dopaminergic neurotoxin N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine was achieved by four different antioxidants including NAC in the mouse [43]. NAC also has a protective effect in pneumococcal meningitis in the rat [44]. In vitro, NAC promotes oligodendrocyte survival in the presence of toxic stimuli or due to withdrawal of growth factors [45] and maturation of oligodendrocytes [46]. NAC inhibits Theiler's virus-induced NO and TNF-α production by murine SJL/J astrocyte cultures [47]. NAC treatment prevented cytokine-induced decrease in GSH and degradation of sphingomyelin to ceramide, also blocked cytokine-mediated ceramide production in rat primary oligodendrocytes, microglia, and C6 glial cells, thereby preventing cell death. These results suggest that intracellular levels of GSH may play a critical role in the regulation of cytokine-induced generation of ceramide leading to apoptosis of brain cells in demyelinating diseases. [48]
In summary, the ability of NAC to inhibit the induction of proinflammatory cytokines and inhibit the transmigration of inflammatory cells into the CNS of EAE-induced rats identifies it as a potential drug for the treatment of neuroinflammatory diseases and possibly other Th1-mediated autoimmune diseases. In addition, in vitro studies suggest that NAC may also promote survival of neurons and oligodendrocytes and thereby potentially facilitating remyelination. MS is a multifactorial disease and the etiology of the disease in unknown. Consequently, the targets for the prevention of the disease are currently unknown. However the disease signs and causes of these are known. For example an increase in pro-inflammatory cytokines and iNOS activity has been linked increase in clinical sign. As evidenced in the manuscript, NAC can inhibit the production of inflammatory cytokines and nitrotyrosine in the CNS during EAE pathogenesis. Thus, NAC holds out to be a promising therapeutic agent for the amelioration of MS/EAE.
Acknowledgements
We would like to thank Joyce Bryan and Carrie Barnes for laboratory assistance, and Hope Terry for secretarial assistance. This work was supported by the grants (NS-22576, NS-34741, NS-37766, and NS-40810) from National Institutes of Health.
==== Refs
Bar-Or A Oliveira EM Anderson DE Hafler DA Molecular pathogenesis of multiple sclerosis J Neuroimmunol 1999 100 252 259 10695735 10.1016/S0165-5728(99)00193-9
Markovic-Plese S McFarland HF Immunopathogenesis of the multiple sclerosis lesion Curr Neurol Neurosci Rep 2001 1 257 262 11898527
Hemmer B Archelos JJ Hartung HP New concepts in the immunopathogenesis of multiple sclerosis Nat Rev Neurosci 2002 3 291 301 11967559 10.1038/nrn784
Kivisakk P Trebst C Eckstein DJ Kerza-Kwiatecki AP Ransohoff RM Chemokine-based therapies for MS: how do we get there from here? Trends Immunol 2001 22 591 593 11698200 10.1016/S1471-4906(01)02056-7
Martiney JA Cuff C Litwak M Berman J Brosnan CF Cytokine-induced inflammation in the central nervous system revisited Neurochem Res 1998 23 349 359 9482247 10.1023/A:1022457500700
Navikas V Link H Review: cytokines and the pathogenesis of multiple sclerosis J Neurosci Res 1996 45 322 333 8872892 10.1002/(SICI)1097-4547(19960815)45:4<322::AID-JNR1>3.0.CO;2-B
Benveniste EN Role of macrophages/microglia in multiple sclerosis and experimental allergic encephalomyelitis J Mol Med 1997 75 165 173 9106073 10.1007/s001090050101
Beck J Rondot P Catinot L Falcoff E Kirchner H Wietzerbin J Increased production of interferon gamma and tumor necrosis factor precedes clinical manifestation in multiple sclerosis: do cytokines trigger off exacerbations? Acta Neurol Scand 1988 78 318 323 3146861
Bogdan C Paik J Vodovotz Y Nathan C Contrasting mechanisms for suppression of macrophage cytokine release by transforming growth factor-beta and interleukin-10 J Biol Chem 1992 267 23301 23308 1429677
Brosnan CF Cannella B Battistini L Raine CS Cytokine localization in multiple sclerosis lesions: correlation with adhesion molecule expression and reactive nitrogen species Neurology 1995 45 S16 21 7540265
Brosnan CF Lee SC Liu J Regulation of inducible nitric oxide synthase expression in human glia: implications for inflammatory central nervous system diseases Biochem Soc Trans 1997 25 679 683 9191181
Bo L Dawson TM Wesselingh S Mork S Choi S Kong PA Hanley D Trapp BD Induction of nitric oxide synthase in demyelinating regions of multiple sclerosis brains Ann Neurol 1994 36 778 786 7526776 10.1002/ana.410360515
Koprowski H Zheng YM Heber-Katz E Fraser N Rorke L Fu ZF Hanlon C Dietzschold B In vivo expression of inducible nitric oxide synthase in experimentally induced neurologic diseases Proc Natl Acad Sci U S A 1993 90 3024 3027 7681993
Interferon beta-1b is effective in relapsing-remitting multiple sclerosis. I. Clinical results of a multicenter, randomized, double-blind, placebo-controlled trial. The IFNB Multiple Sclerosis Study Group Neurology 1993 43 655 661 8469318
Johnson KP Brooks BR Cohen JA Ford CC Goldstein J Lisak RP Myers LW Panitch HS Rose JW Schiffer RB Copolymer 1 reduces relapse rate and improves disability in relapsing-remitting multiple sclerosis: results of a phase III multicenter, double-blind placebo-controlled trial. The Copolymer 1 Multiple Sclerosis Study Group Neurology 1995 45 1268 1276 7617181
Nath N Giri S Prasad R Singh AK Singh I Potential targets of 3-hydroxy-3-methylglutaryl coenzyme a reductase inhibitor for multiple sclerosis therapy J Immunol 2004 172 1273 1286 14707106
Stanislaus R Pahan K Singh AK Singh I Amelioration of experimental allergic encephalomyelitis in Lewis rats by lovastatin Neurosci Lett 1999 269 71 74 10430507 10.1016/S0304-3940(99)00414-0
Youssef S Stuve O Patarroyo JC Ruiz PJ Radosevich JL Hur EM Bravo M Mitchell DJ Sobel RA Steinman L Zamvil SS The HMG-CoA reductase inhibitor, atorvastatin, promotes a Th2 bias and reverses paralysis in central nervous system autoimmune disease Nature 2002 420 78 84 12422218 10.1038/nature01158
Stanislaus R Gilg AG Singh AK Singh I Immunomodulation of experimental autoimmune encephalomyelitis in the Lewis rats by Lovastatin Neurosci Lett 2002 333 167 170 12429374 10.1016/S0304-3940(02)00943-6
Stanislaus R Singh AK Singh I Lovastatin treatment decreases mononuclear cell infiltration into the CNS of Lewis rats with experimental allergic encephalomyelitis J Neurosci Res 2001 66 155 162 11592110 10.1002/jnr.1207
Pahan K Sheikh FG Namboodiri AM Singh I N-acetyl cysteine inhibits induction of no production by endotoxin or cytokine stimulated rat peritoneal macrophages, C6 glial cells and astrocytes Free Radic Biol Med 1998 24 39 48 9436612 10.1016/S0891-5849(97)00137-8
Lehmann D Karussis D Misrachi-Koll R Shezen E Ovadia H Abramsky O Oral administration of the oxidant-scavenger N-acetyl-L-cysteine inhibits acute experimental autoimmune encephalomyelitis J Neuroimmunol 1994 50 35 42 8300856 10.1016/0165-5728(94)90212-7
Pahan K Namboodiri AM Sheikh FG Smith BT Singh I Increasing cAMP attenuates induction of inducible nitric-oxide synthase in rat primary astrocytes J Biol Chem 1997 272 7786 7791 9065441 10.1074/jbc.272.12.7786
Prineas JW Wright RG Macrophages, lymphocytes, and plasma cells in the perivascular compartment in chronic multiple sclerosis Lab Invest 1978 38 409 421 205724
Adams CW Poston RN Buk SJ Pathology, histochemistry and immunocytochemistry of lesions in acute multiple sclerosis J Neurol Sci 1989 92 291 306 2809622 10.1016/0022-510X(89)90144-5
Eng LF Ghirnikar RS Lee YL Inflammation in EAE: role of chemokine/cytokine expression by resident and infiltrating cells Neurochem Res 1996 21 511 525 8734446
Nashold FE Miller DJ Hayes CE 1,25-dihydroxyvitamin D3 treatment decreases macrophage accumulation in the CNS of mice with experimental autoimmune encephalomyelitis J Neuroimmunol 2000 103 171 179 10696912 10.1016/S0165-5728(99)00247-7
Kwak B Mulhaupt F Myit S Mach F Statins as a newly recognized type of immunomodulator Nat Med 2000 6 1399 1402 11100127 10.1038/82219
Palinski W Immunomodulation: a new role for statins? Nat Med 2000 6 1311 1312 11100105 10.1038/82107
Aharoni R Teitelbaum D Arnon R Sela M Copolymer 1 acts against the immunodominant epitope 82-100 of myelin basic protein by T cell receptor antagonism in addition to major histocompatibility complex blocking Proc Natl Acad Sci U S A 1999 96 634 639 9892685 10.1073/pnas.96.2.634
Arnason BG Dayal A Qu ZX Jensen MA Genc K Reder AT Mechanisms of action of interferon-beta in multiple sclerosis Springer Semin Immunopathol 1996 18 125 148 8984676 10.1007/BF00792613
Pozza M Bettelli C Aloe L Giardino L Calza L Further evidence for a role of nitric oxide in experimental allergic encephalomyelitis: aminoguanidine treatment modifies its clinical evolution Brain Res 2000 855 39 46 10650128 10.1016/S0006-8993(99)02133-2
Hartlage-Rubsamen M Lemke R Schliebs R Interleukin-1beta, inducible nitric oxide synthase, and nuclear factor-kappaB are induced in morphologically distinct microglia after rat hippocampal lipopolysaccharide/interferon-gamma injection J Neurosci Res 1999 57 388 398 10412030 10.1002/(SICI)1097-4547(19990801)57:3<388::AID-JNR11>3.0.CO;2-2
Aharoni R Teitelbaum D Sela M Arnon R Copolymer 1 induces T cells of the T helper type 2 that crossreact with myelin basic protein and suppress experimental autoimmune encephalomyelitis Proc Natl Acad Sci U S A 1997 94 10821 10826 9380718 10.1073/pnas.94.20.10821
Zang Y Hong J Robinson R Li S Rivera VM Zhang JZ Immune regulatory properties and interactions of copolymer-I and beta-interferon 1a in multiple sclerosis J Neuroimmunol 2003 137 144 153 12667659 10.1016/S0165-5728(03)00049-3
Chen M Gran B Costello K Johnson K Martin R Dhib-Jalbut S Glatiramer acetate induces a Th2-biased response and crossreactivity with myelin basic protein in patients with MS Mult Scler 2001 7 209 219 11548979 10.1191/135245801680209303
Brenner T Arnon R Sela M Abramsky O Meiner Z Riven-Kreitman R Tarcik N Teitelbaum D Humoral and cellular immune responses to Copolymer 1 in multiple sclerosis patients treated with Copaxone J Neuroimmunol 2001 115 152 160 11282165 10.1016/S0165-5728(01)00250-8
Won JS Im YB Key L Singh I Singh AK The involvement of glucose metabolism in the regulation of inducible nitric oxide synthase gene expression in glial cells: possible role of glucose-6-phosphate dehydrogenase and CCAAT/enhancing binding protein J Neurosci 2003 23 7470 7478 12930785
Floyd RA Antioxidants, oxidative stress, and degenerative neurological disorders Proc Soc Exp Biol Med 1999 222 236 245 10601882 10.1046/j.1525-1373.1999.d01-140.x
Smith KJ Kapoor R Felts PA Demyelination: the role of reactive oxygen and nitrogen species Brain Pathol 1999 9 69 92 9989453
Hussain S Slikker WJ Ali SF Role of metallothionein and other antioxidants in scavenging superoxide radicals and their possible role in neuroprotection Neurochem Int 1996 29 145 152 8837043 10.1016/0197-0186(95)00114-X
Knuckey NW Palm D Primiano M Epstein MH Johanson CE N-acetylcysteine enhances hippocampal neuronal survival after transient forebrain ischemia in rats Stroke 1995 26 305 10; discussion 311 7831704
Perry TL Yong VW Clavier RM Jones K Wright JM Foulks JG Wall RA Partial protection from the dopaminergic neurotoxin N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine by four different antioxidants in the mouse Neurosci Lett 1985 60 109 114 3877260 10.1016/0304-3940(85)90229-0
Koedel U Pfister HW Protective effect of the antioxidant N-acetyl-L-cysteine in pneumococcal meningitis in the rat Neurosci Lett 1997 225 33 36 9143011 10.1016/S0304-3940(97)00177-8
Mayer M Noble M N-acetyl-L-cysteine is a pluripotent protector against cell death and enhancer of trophic factor-mediated cell survival in vitro Proc Natl Acad Sci U S A 1994 91 7496 7500 7914368
Cammer W Protection of cultured oligodendrocytes against tumor necrosis factor-alpha by the antioxidants coenzyme Q(10) and N-acetyl cysteine Brain Res Bull 2002 58 587 592 12372563 10.1016/S0361-9230(02)00830-4
Molina-Holgado F Hernanz A De la Fuente M Guaza C N-Acetyl-cysteine inhibition of encephalomyelitis Theiler's virus-induced nitric oxide and tumour necrosis factor-alpha production by murine astrocyte cultures Biofactors 1999 10 187 193 10609881
Singh I Pahan K Khan M Singh AK Cytokine-mediated induction of ceramide production is redox-sensitive. Implications to proinflammatory cytokine-mediated apoptosis in demyelinating diseases J Biol Chem 1998 273 20354 20362 9685387 10.1074/jbc.273.32.20354
| 15869713 | PMC1097751 | CC BY | 2021-01-04 16:38:03 | no | J Autoimmune Dis. 2005 May 3; 2:4 | utf-8 | J Autoimmune Dis | 2,005 | 10.1186/1740-2557-2-4 | oa_comm |
==== Front
J Exp Clin Assist ReprodJournal of Experimental & Clinical Assisted Reproduction1743-1050BioMed Central London 1743-1050-2-71584017110.1186/1743-1050-2-7ResearchA comparison of polarized and non-polarized human endometrial monolayer culture systems on murine embryo development Baghaban Eslami Nejad Mohamad Reza [email protected] Valojerdi Mojtaba [email protected] Ashtiani Saeed [email protected] Department of Embryology, Royan Institute, Tehran, Iran2 Department of Anatomy, School of Medical Sciences, Tarbiat Modarres University, Tehran, Iran2005 19 4 2005 2 7 7 1 2 2005 19 4 2005 Copyright © 2005 Nejad et al; licensee BioMed Central Ltd.2005Nejad 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
Co-culture of embryos with various somatic cells has been suggested as a promising approach to improve embryo development. Despite numerous reports regarding the beneficial effects of epithelial cells from the female genital tract on embryo development in a co-culture system, little is known about the effect of these cells when being cultured under a polarized condition on embryo growth. Our study evaluated the effects of in vitro polarized cells on pre-embryo development.
Methods
Human endometrial tissue was obtained from uterine specimens excised at total hysterectomy performed for benign indications. Epithelial cells were promptly isolated and cultured either on extra-cellular matrix gel (ECM-Gel) coated millipore filter inserts (polarized) or plastic surfaces (non-polarized). The epithelial nature of the cells cultured on plastic was confirmed through immunohistochemistry, and polarization of cells cultured on ECM-Gel was evaluated by transmission electron microscopy (TEM). One or two-cell stage embryos of a superovulated NMRI mouse were then flushed and placed in culture with either polarized or non-polarized cells and medium alone. Development rates were determined for all embryos daily and statistically compared. At the end of the cultivation period, trophectoderm (TE) and inner cell mass (ICM) of expanded blastocysts from each group were examined microscopically.
Results
Endometrial epithelial cells cultured on ECM-Gel had a highly polarized columnar shape as opposed to the flattened shape of the cells cultured on a plastic surface. The two-cell embryos cultured on a polarized monolayer had a higher developmental rate than those from the non-polarized cells. There was no statistically significant difference; still, the blastocysts from the polarized monolayer, in comparison with the non-polarized group, had a significantly higher mean cell number. The development of one-cell embryos in the polarized and non-polarized groups showed no statistically significant difference.
Conclusion
Polarized cells could improve in vitro embryo development from the two-cell stage more in terms of quality (increasing blastocyst cellularity) than in terms of developmental rate.
==== Body
Background
Embryos developing in vivo are nourished by the cells and fluid of the fallopian tube and the uterine endometrium [1,2]. In vitro embryo culture deprives the cells from the maternal environment, where they are bathed in an ever-changing milieu of fluid containing a range of protein and ions specific to the reproductive process [3,4]. Although mammalian embryos can be cultured in vitro, in vivo development is superior in terms of rate of development [5-7], cell number [6], various biochemical parameters and survival following embryo transfer [8,9]. Attempts to improve in vitro culture conditions by modifying electrolyte composition and energy substrates have met with limited success [1,10]. To solve this problem and in contrast to the use of a medium alone, a number of co-culture systems mimicking the embryotrophic cell environment of the genital tract have been devised in which embryos can develop to blastocyst and hatching blastocyst stages [11-16]. Even though the beneficial effects of co-culture systems have been suggested by a number of researchers, the mechanism of action of co-culture cells has not been fully elucidated. Several investigators have proposed that co-culture systems optimize in vitro culture conditions by secreting embryotrophic substance(s) into the culture medium [17] and removing embryotoxic materials from it [18]. Embryo-feeder cell contact is also necessary to allow cellular connections between the zona and feeder cells for cross-transfer of such embryotrophic factor(s) [19]. To prepare a co-culture system, epithelial cells are usually cultivated on a plastic surface in the form of a monolayer; embryos are then cultured on it. Several studies have demonstrated that during the culture procedure performed on plastic surfaces, epithelial cells lose polarity and differentiated function (such as polarized secretion) in several days [20-23]. Furthermore, the cultivated cells lose their cubical shape, become flat and fail to keep their lateral junctions, including tight junction and desmosomes. In contrast to the loss of polarity occurring during conventional epithelial cell culture on plastic surfaces, numerous studies have demonstrated the retention of structural polarity and differentiated function when the culturing condition mimics two aspects of the in situ epithelial cell environment: baso-lateral feeding and contact with an extra-cellular matrix [24-27].
To date, several reports have shown that polarity retention of epithelial cells during culture (compared to non- polarized cells) may improve motility and fertilizing capacity of the sperm co-cultured on it [28-30]. There is little information, however, regarding co-culture of polarized cells and embryos. The aim of our investigation was to observe polarized human endometrial cells on ECM-Gel and to evaluate their effects versus those of conventional non- polarized monolayers on murine pre-embryo development.
Methods
Cell culture
Human endometrial tissue was obtained from patients in the secretive phase (according to full thickness of the endometrium and histological appearance of the tissue) undergoing hysterectomy at the Department of Obstetrics and Gynecology, Arash Hospital (Tehran, Iran) for benign indications. Use of the endometrial tissue for research was approved by the Ethics Committee of Royan Institute (Tehran, Iran). Tissue was excised immediately after the hysterectomy and were rinsed in Hank's balanced salt solution to remove blood and debris before being minced into small pieces of about 1 mm3. A small portion of each specimen was fixed in Karnovsky solution for TEM. Tissue fragments were digested by 0.25% type I collagenase in a DMEM/HAM's-F12 (Sigma, USA), containing 100 IU/ml streptomycin (Sigma, USA), 100 IU/ml Penicillin(Sigma, USA), and 10% heat inactivated fetal calf serum (FCS, Gibco, UK). After 1.5 h, the majority of glands were dissociated from surrounding tissues and floated in the digestate produced by collagenase enzyme at 37°C. The digestate also contained free stromal cells and undigested fragments. The unit gravity sedimentation technique helped isolate the endometrial glands from the other components. First, the digestate was transferred into a 13-ml plastic tube (Falcon, USA), and after its volume was reconstituted to 12 ml with DMEM/HAM's-F12, the tube was turned upside down to resuspend the digestate. The endometrial fragments were then pelleted by being allowed to settle under normal gravity for ~1 min, and the pellets containing undigested materials were subsequently discarded. The supernatants having been transferred into another tube, the glands were separated from the free cells under normal gravity, washed by centrifugation (10 min 75 g, ×2) and resuspended in a medium containing 100 IU/ml streptomycin, 100 IU/ml Penicillin, 5% FCS, 10 μg/ml epidermal growth factor (Sigma, USA) and 10 μg/ml retinoic acid (Sigma, USA). Next, they were distributed on a 12 mm diameter filter in millicell CM inserts (Millipore, USA), previously coated with ECM-Gel (Sigma, USA). Coated filters were prepared by adding 0.1 ml of 1:4 diluted ECM-Gel in DMEM/HAM's-F12 to each insert, followed by drying under a laminar flow hood. Filters were stored under sterile conditions at 4°C. The ECM-Gel coated inserts were placed in a well on a standard 24-well tissue culture plate (polystyrene plastic, Falcon, USA), containing a DMEM HAM's-F12 medium. Optimal seeding density was achieved when the glands covered approximately 30% of the surface allowing them to explant in monolayer fashion. Cultures were usually confluent after 7d. Aliquots of the gland suspension were also cultured in parallel in a well of 24-well polystyrene dish before the plates were placed in an incubator. The medium was replaced every 3 days.
Transmission electron microscopy
The cultures on ECM-Gel and the tissue material were fixed for 1 h by Karnovsky fixative at room temperature, and then fixative was removed and the culture and tissue washed with 0.1 M OsO4 (Sigma, USA) in 0.1 M phosphate buffer × 1 h at 20°C. The ECM-Gel coated filter, containing glandular epithelial cells, was cut out of the millicell insert with a scalpel and subjected in parallel with the tissue material to graded ethanol dehydration before being embedded in Araldite (Araldite 2005, Sigma, USA).
Cells cultured on plastic were processed in the same manner as used for cells cultured on ECM-Gel and the tissue materials, but they were embedded differently. First, the embedding plastic blocks were filled with Araldite, and then the plastic containing the cultured cells were laid on the blocks and placed at 60°C × 18 hrs. Next, the plastic was removed and the cells remained within the resin surface in the blocks. Semi-thin sections (0.3 μm) were made and stained with toluidine blue, while the ultra-thin sections (70–100 nm) were contrasted with uranyl acetate and lead citrate and examined using a transmission electron microscope (Ziess TM 900, Germany).
Immunohistochemistry
An immunohistochemistry staining procedure utilizing Dako Envision+ system peroxidase (Dako, Denmark) was utilized in order to demonstrate the presence of cytokratin 7. The procedure was in accordance with the method proposed by the manufacturer. Briefly, the endogenous peroxidase was inhibited by incubating the cells cultured in Termanox coverslip with peroxidase block for 5 minutes, and then they were washed in distilled water and placed in wash solution. Next, the monolayer was incubated furthermore with the mouse antibody against cytokratin 7 at a concentration of 1:15 for 15 minutes. After that, it was overlaid with peroxidase labeled polymer conjugated to goat anti-mouse immunoglubins in Tris-Hcl buffer, containing carrier protein and an anti-microbial agent for 30 minutes before being washed. The last step was an incubation with DAB+ substrate- chromogen solution for 10 minutes, followed by Hematoxylin counterstaining. The negative control was monolayer undergoing stain procedure without antibody, and the positive control was endometrial frozen section. The majority (~95%) of cells isolated were epithelial.
Embryo and co-culture specifications
Female NMRI 6–8-week-old mice underwent ovulation induction by the injection of 10 IU human menopausal gonadotropin (HMG, Organon, Holland), followed 48 h later by the injection of 10 IU human chorionic gonadotropin (HCG, Organon, Holland). Females were mated with males from the same strain. Mice with vaginal plugs were considered pregnant and sacrificed by cervical dislocation 20–22 h and 44–48 h post-hCG for one- and two-cell embryos, respectively. Embryos were flushed from the oviduct with DMEM/HAM's-F12, supplemented with 5 mg/ml bovine serum albumin (BSA, Sigma, USA). Morphologically normal embryos were washed and pooled in fresh DMEM/HAM's F12 medium before use.
Parallel to mice superovulation and embryo collection, polarized and non-polarized epithelial monolayers were prepared in 24-well plates as described above. On day 7 after initiation of culture when monolayers became confluent, the medium was replaced with DMEM/HAM's-F12, containing 5 mg/ml BSA. 24 h later, one or two-cell embryos were cultured on the monolayers. In this study, the polarized and non- polarized groups were considered as experimental group I (exp I) and experimental group II (exp II), respectively; and the medium alone was designated as the control group (con). Each co-culture experiment was replicated 5 times. Embryo developmental rate was observed every 24 hrs and recorded for 96 or 120 h. Early cleavage stage embryos were classified as degenerate if >25% of each embryo contained cytoplasmic fragments or if the blastomeres appeared dark and granular. Morula and blastocyst stage embryos were considered as degenerate if they collapsed. Finally, the results of the development were statistically compared using χ2 test.
Blastocyst differential staining
At the end of cultivation (120 hrs for one-cell and 96 hrs for two-cell embryos), expanded blastocysts of each group were randomly selected, and ICM and TE were differentially stained according to Thouas et al. [31]. To stain the embryos, the blastocysts were first incubated in 500 μL of solution 1 (BSA-free Hepes buffered murine tubal fluid medium + 1% triton X-100 and 100 μg/ml propidium iodide, Sigma, USA) for up to 10 sec. Blastocysts were then immediately transferred into 500 μl of solution 2 (fixative solution of 100% ethanol + 25 μg/ml bisbenzimide, Sigma, USA) and stored at 4°C overnight. Stained blastocysts were transferred from solution 2 directly into glycerol, taking care to avoid carry-over of excessive amounts of the solutions. Blastocysts were mounted on glass slides in a drop of glycerol, and cell counting was performed from images obtained from an inverted microscope fitted with an ultraviolet lamp and excitation filters (460 nm for blue and red fluorescence and 500 nm for red only). Data were analyzed by one-way analysis of variance (ANOVA).
Results
Cell culture
After one day of being cultured on ECM-Gel, epithelial cells grew out of the glands, spread into the intervals among the glands, seeded on ECM-Gel coated inserts, and reached confluency after 7 days. These cells, columnar in shape and packed closely together, appeared as an epithelial plate on the ECM-Gel, whereas the cells on the plastic surfaces, appeared polygonal and not packed closely together, reaching confluency after 5 days of culture.
Polarized cell structure
The cells cultured on ECM-Gel surfaces had a highly polarized appearance with a basal nucleus and apical abundant microvilli (Figs. 1a and 1b). Lateral junctions (as desmosomes and tight junctions) were well established and basal lamina was formed under the epithelial cells (Fig. 1c). Nuclei of the in vitro cells were eukromatin with some invagination in the nuclear envelope (Fig. 1b). Their cytoplasm had abundant spherical or elongated mitochondria and cisterns of rER. In general, the cells cultured on ECM-Gel were similar to those observed in tissue.
Figure 1 Light and Electron Micrograph of Polarized and Non-Polarized Human Endometrial Epithelial Cells in Culture. a: Semi-thin Section of Polarized epithelial cells;tuluidine blue staining;bar:5 μm. b,c: Ultra-thin Section of Polarized epithelial cells; N:nucleous; arrow: microvilli; mitochondria; jc: junctional complex; Uranyl acetate and lead citrate staining; bar b:1.6 μm; bar e: 200 nm. d: Semi-thin Section of non-polarized epithelial cells; Tuluidine blue staining; bar: 5 μm. e,f: Ultra-thin Section of non-polarized epithelial cells; arrow: Thick process; M: mitochondria; V: Vacuoles; Dl: desmosome like junction; Uranyl acetate and lead citrate staining; bar d,f: 200 nm.
Non-polarized cell culture
Cells cultured on plastic surfaces (non-polarized cells) were short and spindly in section (Fig. 1d) as opposed to the columnar cells in the tissue fragment. These cells had only a few short apical processes (Fig. 1e). Extensions of adjacent cells were observed to lay upon each other and desmosome-like junctions fastened them together; there was no tight junction (Fig. 1f). The nuclei of these cells, in comparison with the epithelial cells in vivo, were relatively large. Their cytoplasm was relatively free of organelles; only a few mitochondria, cisterns of rER, and large vacuoles were seen (Fig. 1e).
Development and blastocyst cellularity
One-cell embryos
A total of 287 one-cell embryos were randomly allocated to the uterine polarized group (Exp 1; n = 97), and uterine non-polarized group (Exp 2; n = 77) and DMEM/HAM's-F12 (con; n = 113). The rate of development and embryo degeneration in the three culture systems are shown in Additional file 1, Table 1.
Compared to the control group, embryo development was enhanced significantly by all the feeder cell cultures throughout the cultivation period (120 h). In both experiment groups, a significantly higher number of embryos reached a more advanced developmental stage (blastocysts and hatching blastocysts), compared to the control group where no blastocyst hatching was noted. Embryo development rates in the polarized group were higher than that in the non-polarized group, but the difference was not statistically significant. During the cultivation period, embryo degeneration in the polarized group was lower (without statistical difference) than that observed in the non-polarized group. However, embryo degeneration in the feeder culture groups were significantly lower than those in the control group. At the end of the cultivation period, the mean total cell number of the blastocysts developed on the feeder layers was significantly higher than that in the control group (p < 0.005). For this parameter there was no statistically significant difference between the polarized and non-polarized groups, but mean total cell number was higher in the polarized group. There were similar relationships between our groups in terms of the mean percentage of TE and ICM (see Additional file 2, Table 3).
Two-cell embryos
A total of 339 two-cell embryos were randomly allocated to the uterine polarized group (n = 106), uterine non-polarized group (n = 109) and DMEM/ HAM's-F12 (n = 124). Both feeder layers significantly enhanced the embryonic development; as a result, more embryos reached blastocyst and hatching blastocyst stages in the co-culture groups in comparison with those in the control group. The rate of embryo degeneration in the co-culture groups was significantly lower than that in the control. The percentage of the blastocysts and hatching blastocysts produced in the polarized group was higher than that in the non-polarized group, but the difference was not statistically significant (see Additional file 1, Table 2). Cell count results showed that the blastocysts produced in the polarized and non-polarized groups had more blastomeres than those in the control group, which was statistically significant (p < 0.05). Mean percentages of TE and ICM of the blastocysts from the three groups had no significant difference (see Additional file 2, Table 4).
Discussion
Epithelial cells of the human endometrium have already been cultured on ECM-Gel for different purposes. Bentin-Ley et al. [32,33] used this culture system to study the changes occurring on the surface of human endometrial epithelial cells in the presence of an implanted blastocyst, while Arnold et al. [34] cultured these cells on ECM-Gel to describe the regulatory role of stromal cells cultured within the Gel on epithelial cell function and morphogenesis in vitro.
Park et al. [35] established a three-dimensional endometrial culture where human endometrial stromal cells were embedded in a mixture of collagen I and matrigel; epithelial cells were cultivated before they were replaced by KLE cells (endometrial cancer cells of epithelial origin). These investigators studied the invasion of KLE cells into the stromal fraction. Using ECM-Gel, Classen-Linke et al. [36] established an endometrial cell culture system to study progesterone and estrogen receptors as marker molecules for physiologically intact epithelial cells by immunohistochemistry and RT-PCR.
A number of researchers have hypothesized that polarized epithelial cells may hold promise for promoting growth and development of other cells. Pollard et al. suggested that maintaining cell polarity was necessary to optimize co-culture systems [37]. Having cultured epithelial cells of bovine and mouse oviducts as a polarized monolayer using Costar Transwell-Col and having also evaluated the embryo trophic potential of these cells in co-culture with mouse one-cell embryos, Ouhibi et al. [38] concluded that the maintenance of cell polarity of the feeder layers did not improve embryo development. It should be noted that the said authors did not use ECM-Gel for the cell polarity induction, nor did they evaluate polarity status of the cultured cells. Other investigators have cultured oviduct epithelial cells on ECM-Gel and shown that motility and fertilizing capacity of sperm is improved when co-cultured with polarized epithelial cells (compared to non-polarized cells cultured in plastic) [28-30]. Our work focused on human uterine epithelial cells that were cultured on ECM-Gel as a polarized monolayer; following polarity confirmation by TEM, their potential for improving the mouse embryo development was evaluated.
Ultra-structural evaluation of cell polarity has been the subject of previous research [22,39]. An important marker of polarity at the ultra-structural level is the presence of tight junctions at the boundary of the lateral and apical membrane domains. Vega Salas er al. [40] suggested that the interaction between ECM and epithelial cells generated differences in protein distributions between contacting and non-contacting surfaces and that these interactions also refined the apical-basal polarity, the useful sign of which is the location of tight junction at the apical-lateral membrane boundary. This junction prevents the mixing of specific proteins from different membrane domains. Other signs of cell polarity at the ultra-structural level include the reduction of distance between two adjacent cells, basal location of nuclei, and formation of apical microvilli. Our ultra-structural images showed that the cells cultured on ECM-Gel were polar and quite similar to cells observed in vivo.
Selecting a medium which could support both embryo and feeder cells presented the main challenge in our study. Frasor et al. [12] examined growth and development of mouse two-cell embryos on human oviduct epithelial cells using HTF (a simple medium able to support embryo development) and MEM-α (a complex medium capable of supporting somatic cell growth) and concluded that optimizing of growth characteristics of somatic cells in culture would improve embryo development. Several media were used in this research so that a polarized monolayer could be established; nonetheless, DMEM/HAM's-F12 was used as the co-culture medium as it was the only medium supporting the growth and maintenance of the polarized cells. DMEM/HAM's-F12 has also been used by other investigators as a co-culture medium. Specifically, Freeman et al. [41] cultured two cell mouse embryos on fibroblast, oviduct, and uterine epithelial cells as well as on follicular cells using DMEM/HAM's- F12 medium. It is noteworthy that their blastocyst rates were slightly higher than those in our study, although this could be associated with different mouse strains used in each experiment (B6C3F1 versus NMRI).
Our investigation showed that one-cell embryos of the polarized and non-polarized groups developed statistically well compared with those of the control, but the polarized groups showed no advantage over the non-polarized group when it came to improving the in vitro development of embryos from one-cell stage. This may be because of a developmental block occurring at the two cell stage, or perhaps the embryotrophic effects of the polarized monolayer manifest only when the co-culture system is initiated by using two-cell stage embryos. It seems that in co-culture systems, the activation mechanism is insufficient to overcome the embryo developmental block. This may cause the embryo not to respond properly even to positive influence of the polarized group. Such embryos would be expected to have a lower cleavage rate and reduced cellularity at the blastocyst stage, compared to those co-cultured from the 2-cell stage.
We also found a similar proportions of TE and ICM in the blastocysts that originated from both one- and two-cell embryos. It seems that irrespective of total cell number, cell allocation mechanisms put specified proportions of the cells into the TE and ICM component of each blastocyst. Indeed, simillar results were reported by Sherban et al [12].
One positive effect of a polarized culture system on two-cell embryo development would be reflected as efficiency of blastocyst production. Our results demonstrate that blastocysts developing from embryos cultured in a polarized state acquire significantly more blastomeres compared with those cultured among non-polarized cells. High mean cell numbers are also indicative of higher early cleavage rates and cell viability. In addition, enhanced embryo implantation may result from an increased rate of blastocoel fluid production and earlier attainment of a critical hatching diameter [42]; increased production of zona lysine by TE cells must also be considered [43].
Handyside and Hunter [44] demonstrated that the proportion of TE increased from ~60% in early mouse blastocysts to ~83% before implantation. In the first 36 h after blastocyst formation, TE components increase exponentially and then plateau as the increase in ICM cell numbers slows; this is coincident with the period of apoptosis in the ICM [45]. These influences result in an increase in the proportion of TE cells in the blastocyst during the period of blastocyst expansion and hatching. Our observations in this regard are consistent with those reported previously by Sherban et al. [46]. From this it may be hypothesized that blastocysts with a greater relative TE cell proportion are more advanced than those with a low TE cell proportion. In the present study, cell counting was performed at the end of a 120 h culture interval. Our results indicate that the mean TE percentage of embryos cultured on polarized cells is higher than that of embryos cultured on on-polarized cells, although the difference did not reach statistical significance.
It may be that the embryotrophic effects of experiment I (polarized group) are due both to the ECM and the polarized cells themselves, but it should be noted that cells cultured on ECM-Gel completely occupy all available surfaces of ECM-Gel and furthermore tight junctions retain ECM-Gel in the lower compartment during co-culture. This had the effect of rendering no exposed surfaces of ECM-Gel to influence the embryo development and therefore all embryotrophic effects would be due to highly polarized columnar epithelial cells. Positive influences of polarized cells on embryo growth and blastocyst cellularity produced from two-cell embryo development may be due to their differentiated state and morphology on matrigel. Given that feeder cells provide their effect by secreting an embryotrophic agent, it is logical to speculate that polarized cells by releasing different embryotrophic materials into the culture medium can improve embryo quality more than their non-polarized counterparts. Thomas et al. [47] found that oviduct epithelial cells polarized in vitro, had a different protein composition compared to those of the plastic cultured non-polarized cells. Also, Wolddesenbet and Newton [48] indicated that oviduct epithelial cells in polarized cultured had similar secretions to those of the cultured tissue fragment. The effect of the direct contact between feeder cells and embryos on embryo development must also be considered; preventing such contact reduces the positive effect of co-culture cells on the embryo. In polarized co-culture systems, feeder cells are columnar in shape and lie very close to one another, so each embryo establishes direct contact with more cells and probably gets more positive influences compared to those on the non-polarized cells. Determining the precise composition of the polarized cell secretions and the mechanism of action of direct contact on embryo development in co-culture systems requires more investigation.
Conclusion
Our results suggest that epithelial cells of human oviduct, being cultured in a polarized condition, could improve blastocyst quality when co-cultured with two-cell embryos (but not one-cell embryos). It seems that providing an appropriate environment capable of supporting feeder layer as well as embryo growth would further improve in vitro development of one and two-cell mouse embryos in terms of both blastocyst formation rate and quality.
Supplementary Material
Additional File 1
file containing table 1 and 2
Click here for file
Additional File 2
file containing table 3 and 4
Click here for file
Acknowledgements
This research was supported by Royan Institute. We thank the staff of the Department of Electron Microscopy of the Sahid Beheshty University for ultrastructural evaluation, and the clinicians in the Department of Obstetrics and Gynaecology of the Arash and Tehran Hospital for supplying endometrial samples.
==== Refs
Quinn p Kerin JF Warnes GM Improved pregnancy rate in human in vitro fertilization with the use of a medium based on the composition of human tubal fluid Fertil Steril 1985 44 493 498 3902512
Tervit HR Whittingham DG Rowson LEA Successful culture in vitro of sheep and cattle ova J Reprod Fertil 1972 30 493 497 4672493
Neider GL Macon GR Uterine and oviductal protein secretion during early pregnancy in the mouse J Reprod Fert 1987 81 287 294
Leese HJ The formation and function of oviduct fluid J Reprod Fert 1988 82 283 856
Bowman P McLaven A Cleavage rate of mouse embryos in vivo and in vitro J Embryol Exp Morphol 1970 24 203 207 5487156
Harlow GM Quinn P Development of preimplantation mouse embryo in vivo and in vitro Aust J Biol Sci 1982 35 187 193 7126059
Erbach GT Lawitts JA Papaioannou VE Biggers JD Differential growth of the mouse premplantation embryo in chemically defined media Biol Reprod 1994 50 1027 1033 8025158
Jung T Fisher B Correlation between diameter and DNA or protein synthetic activity in rabbit blastocyst Biol Reprod 1988 39 1111 1116 3219384
Carney EW Foote RH Effects of super ovulation, embryo recovery, culture system and embryo transfer on development of rabbit embryos in vivo and in vitro J Reprod Ferrtil 1990 89 543 551 10.1530/jrf.0.0890543
Fitzgerald L Di Mattina M Improved medium for long-term culture of human embryos overcomes the in vitro developmental block and increase blastocyst formation Fertil steril 1992 57 641 647 1740212
Barmat LI Worrilow KC Payton BV Growth factor expression by human oviduct and buffalo rat liver co-culture cells Fertil Steril 1997 67 775 779 9093210 10.1016/S0015-0282(97)81382-9
Frasor J Sherbahn R Soltes B Molo MW Binor Z Radwanska E Rawlins R Animal experimentation: optimizing tubal epithelial cell growth promotes mouse embryo hatching in co- culture J Assit Reprod Genet 1996 5 423 430
Hoshi K Kanno Y Katayose H Yanagida K Suzuki R Sato A Coculture of mouse embryo with cryopreserved human oviduct epithelial cells J Assist Reprod Genet 1994 11 367 372 7795370
Gandolfi F Moor RM Stimulation of early embryonic development in the sheep by co-culture with oviduct epithelial cells J Reprod Fert 1987 81 23 28
Sakkas D Trounson AO Kola I In vivo cleavage rates and viability obtained for early cleavage mouse embryo in co-culture with oviduct cells Roprod Fertil Dev 1989 1 127 136
Lavranos TC Seamark RF Addition of steroid to embryo-uterine monolayer co-culture enhances embryo survival and implantation in vitro Reprod Fertil Dev 1989 1 41 46 2710958
Thibodeaux J Godke R In vitro enhancement of early stage embryos with co-culture Arch Pathol Lab Med 1992 116 364 372 1558473
Fukui Y McGowan LT James RW Pugh PA Tervit HR Factors affecting the in vitro development of blastocysts of bovine oocytes matured and fertilized in vitro J Reprod Fertil 1991 92 125 131 1905351
Bongso A Soon-Chye NG Chui-Yee Fong Shan Ratnam Co-cultures: a new lead in embryo quality improvement for assisted reproduction Fertil Steril 1991 56 179 191 2070845
Glasser SR Julian J Decker GL Tang JP Carsen DD Development of morphological and functional polarity in primary cultures of immature rat uterine epithelial cells J Cell Biol 1988 107 2409 2423 3143736 10.1083/jcb.107.6.2409
Jacobs AL Decker GL Glasser SR Julian J Carsen DD Vectorial secretion of prostaglandins by polarized rodent uterine epithelial cells Endocrinology 1990 126 2125 2136 2318159
Schatz F Gordon RE Lanfer N Gurpide E Culture of human endometiral cells under polarizing conditions Differentiation 1990 42 184 190 2340973
Mani SK Decker GL Glasser SR Hormonal Responsiveness by immature rabbit uterine epithelial cells polarized in vitro Endocrinology 1991 128 1563 1573 1705507
Ailenberg M Fritz IB Control of levels of plasminogen activator activity secreted by sertoli cells maintained in a two-chamber assembly Endocrinol 1988 122 2613 2618
Carson DD Tang J-P Julian J Glasser SR Vectorial secretion of proteoglycans by polarized rat uterine epithelial cells J Cell Biol 1988 197 2425 2434 3198693 10.1083/jcb.107.6.2425
Chembard M Verrier B Gabrion J Mauchamp J Polarization of thyroid cells in culture: evidence for basolateral localization of the iodide "pump" and at the thyroid stimulating hormone receptor- adenly cyclase complex J Cell Biol 1983 96 1172 1177 6300145 10.1083/jcb.96.4.1172
Hardley MA Byers SW Snarez-Quarez-Quian CA Kleinman HA Dym M Extracellular matrix regulates sertoli cell differentiation, testicular cord formation, and germ cell development in vitro J Cell Biol 1985 101 1511 1522 4044644 10.1083/jcb.101.4.1511
Pollard JW Plante C King WT Hansen PJ Betteridge KJ Suarez SS Fertilizing capacity of bovine sperm may be maintained by binding of oviductal epithelial cells Biol Reprod 1991 44 102 107 2015341
Sidhu KS Mate KE Rodger JC Sperm – oviduct epithelial cell monolayer co-culture: an in vitro model of sperm female tract interactions in a marsupial, the tamer wallaby (macropus engenii) J Reprod Fertil 1988 114 55 61 9875155
Ellington JE Jones AE Davitt CM Schneider CS Brishois RS Hiss GA Wright RW Human sperm function in co-culture with human, macaque or bovine oviduct epithelial cells monolayers Hum Reprod 1998 13 2797 2804 9804233
Thouas GA Korfiatis NA French AJ Jones GM Trounson AO Simplified technique for differential staining of inner cell mass and trophoectoderm cells of mouse and bovine blastocysts Reprod Biomed Online 2001 3 25 29 12513888
Bentin-Ley U Horn T Sjogren A Sorensen S Falck Larsen J Hamberger L Ultra structure of human blastocyst – endometrial interactions in vitro J Reprod Fertil 2000 120 337 350 11058449 10.1530/reprod/120.2.337
Bentin-Ley U Sjogren Nilsson L Hamberger L Larsen JF Horn T Presence of uterine pinopodes at the embryo – endometrial interface during human implantation In vitro hum Reprod 1999 14 515 520 10.1093/humrep/14.2.515
Arnold JT Kaufman DG Seppala M Lessey BA Endometrial stromal cells regulate epithelial cell growth in vitro: a new co-culture model Hum Reprod 2001 16 836 845 11331626 10.1093/humrep/16.5.836
Park DW Choi DS Ryu HS Know HC Joo H Min CK A well – defined in vitro three dimensional culture of human endometrium and its applicability to endometrial cancer invasion Cancer Lett 2003 195 185 192 12767527
Classen-Linke I Kusche M Knauthe R Beier HM Establishment of a human endometrial cell culture system and characterization of its polarized hormone responsive epithelial cells Cell Tissue Res 1997 287 171 185 9011393 10.1007/s004410050743
Pollard JW Xu KP Porie R King WA Betteridge KJ Influence of various oviductal epithelial cell culture systems on the development of early cleavage stage bovine embryos in vitro Theriogenology 1989 31 239 10.1016/0093-691X(89)90647-X
Ouhibi N Hamidi J Guillad J Menezo Y Co-culture of one cell mouse embryo on different cell support Hum Reprod 1990 5 737 743 2254409
Bentin-Ley U Pederson B Linderberg S Larsen JB Hamberger L Horn T Isolation and culture of human endometrial cells in a three – dimensional culture system J Reprod Fertil 1994 101 327 332 7932366
Vega Salas DE salas PJ Gunderson D Rodriguez-Boulan E Formation of the apical pole of epithelial (madin – Darby Canine Kidney) cells : Polarity of an apical protein is independent of tight junctions while segregation of a basolateral marker requires cell – cell interactions J Cell Biol 1987 104 905 916 3558485 10.1083/jcb.104.4.905
Freeman MR Bastias MC Hill GA Osteen KG Co- culture of mouse embryos with cells isolated from the human ovarian follicle, oviduct and uterine endometrium Fertil Steril 1993 59 138 142 8419201
Boatman DE Bavister BD In vitro growth of non-human primate pre- and peri-implantation embryos:Mammalian Preimplantation Embryo 1987 New York: Plenum Press 273 308
Schiewe MC Hazeleger NI Sclimenti C Balmaceda JP Physiological characterization of blastocyst hatching mechanisms by use of a mouse antihatching model Fertil Steril 1995 63 288 294 7843433
Handyside AH Hunter S Cell division and death in the mouse blastocyst before implantation Roux's Arech Dev Biol 1986 195 519 526 10.1007/BF00375893
Hardy k Handyside AH Winston RM The human blastocyst:cell number, death and allocation during late preimplantation development in vitro Development 1989 107 597 604 2612378
Sherban R Frasor J Radwanska E Binor Z Wood-molo M Hibner M Mack S Rawlines RG Comparison of mouse embryo development in open and micro drop Co- Culture systems Hum Reprod 1996 11 2223 2229 8943534
Thomas PG Ignotz GG Ball BA Miller PG Brinsko SP Currie B Isolation, Culture and characterization of equine oviduct epithelial cells in vitro Mol Reprod Dev 1995 41 468 478 7576614 10.1002/mrd.1080410409
Woldesenbet S Newton GR Comparison of proteins synthesized by polarized caprine oviductal epithelial cells and oviductal explants in vitro Theriogenology 2003 60 533 543 12763166 10.1016/S0093-691X(03)00036-0
| 15840171 | PMC1097752 | CC BY | 2021-01-04 16:40:17 | no | J Exp Clin Assist Reprod. 2005 Apr 19; 2:7 | utf-8 | J Exp Clin Assist Reprod | 2,005 | 10.1186/1743-1050-2-7 | oa_comm |
==== Front
J Immune Based Ther VaccinesJournal of Immune Based Therapies and Vaccines1476-8518BioMed Central London 1476-8518-3-21584273010.1186/1476-8518-3-2Original ResearchKinetics and isotype profile of antibody responses in rhesus macaques induced following vaccination with HPV 6, 11, 16 and 18 L1-virus-like particles formulated with or without Merck aluminum adjuvant Ruiz Wanda [email protected] William L [email protected] Kathrin U [email protected] Mark T [email protected] Vaccine and Biologics Research Merck Research Laboratories 466 Devon Park Dr. Wayne, PA 19087-8630 USA2 Vaccine and Biologics Research Merck Research Laboratories West Point, PA 19486 USA2005 20 4 2005 3 2 2 3 3 2005 20 4 2005 Copyright © 2005 Ruiz et al; licensee BioMed Central Ltd.2005Ruiz 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
Human papillomaviruses (HPV) are the most common sexually transmitted viruses. Infection of the cervical epithelium by HPVs can lead to the development of cervical cancer. Recent advances in vaccine research have shown that immunization with papillomavirus-like particles (VLPs) containing the major structural viral protein, L1 from HPV 16 can provide protection from the establishment of a chronic HPV 16 infection and related cervical intraepithelial neoplasia (CIN) in baseline HPV 16 naïve women.
Methods
To better understand the quantitative and qualitative effects of aluminum adjuvant on the immunogenic properties of an HPV 6, 11, 16 and 18L1 VLP vaccine, we used an HPV-specific, antibody isotyping assay and a competitive immunoassay that measures antibodies to neutralizing epitopes to profile sera from rhesus macaques immunized with the HPV L1 VLP vaccine formulated with or without aluminum adjuvant.
Results
Immunization with VLPs formulated with the aluminum adjuvant elicited a significantly stronger immune response with higher peak antibody titers both at four weeks post vaccination (12.7 to 41.9-fold higher) as well as in the persistent phase at week 52 (4.3 to 26.7-fold higher) than that of VLPs alone. Furthermore, the aluminum adjuvant formulated HPV VLP vaccine elicited a predominantly T helper type 2 response, with high levels of IgG1 and IgG4 and low levels of IgG2. The vaccine also elicited high levels of serum IgA, which may be important in providing mucosal immunity to impart protection in the anogenital tract.
Conclusion
These results show that the HPV 6, 11, 16 and 18 L1-VLP vaccine formulated with Merck aluminum adjuvant elicits a robust and durable immune response and holds promise as a vaccine for preventing cervical cancer.
human papillomavirusvaccineneutralizing antibodyLuminex
==== Body
Background
Cervical cancer remains a leading cause of cancer-related deaths in women. HPV infection is the obligate first step in the development of cervical cancer [1]. Nearly a quarter of a million women die from cervical cancer and about half a million are diagnosed with this disease each year [2]. Cervical cancer accounts for 12% of all cancers in women and is the second most frequent gynecological malignancy in the world [3]. A large portion of this health burden is in the developing world, where women do not have access to good healthcare and Papanicolaou (Pap) screens. Although the widespread use of Pap screening in the developed world has increased the early detection of cervical dysplasia and cancer, thereby improving treatment outcomes for cervical cancer, it would be far more preferable to have a vaccine that blocks HPV infection, thereby preventing initiation of the disease process. Also, developing countries that usually do not have access to Pap screening and other preventive measures would further benefit from a vaccine that blocks HPV infection and its subsequent disease consequences. Therefore, there is a great need for an effective and generally well-tolerated HPV vaccine, having a low rate of occurrence of adverse events during administration.
Human papillomaviruses are small double-stranded DNA viruses. Infection with HPV is the most common viral sexually transmitted diseases worldwide [4]. HPV infects cutaneous and mucosal epithelial cells and causes benign and malignant hyperproliferative lesions, which includes genital warts and cervical cancer [5]. To date, more than 100 HPV types have been identified. Of these, 35 infect the genital tract [6,7]. Genital HPV types can be divided into two broad categories: low-risk types which cause genital warts, cervical dysplasia, but little cancer, and high-risk types, which cause dysplasia that can progress to cancer. HPV type 16 and 18 infection cause 70% of cancer cases and 25% of low grade cervical dysplasia [8]. HPV types 6 and 11, on the other hand, cause approximately 95% of genital warts (condylomata acuminata or venereal warts) and 25% of low grade cervical dysplasia (CIN1) [9,10]. Thus, a vaccine targeting HPV 6, 11, 16 and 18 will target the majority of HPV-related clinical disease.
Recently, we reported a proof-of-concept efficacy study for a prophylactic vaccine composed of HPV 16 L1 virus-like particles (VLP) formulated on Merck Aluminum Adjuvant (MAA) [11]. The double-blinded study randomized 2,391 women (16–23 years old) to receive either placebo or three doses of 40 μg of the HPV 16 VLPs in a 0, 2, and 6 month regimen. The primary efficacy endpoint was persistent HPV 16 infection, including HPV 16 related cervical dysplasia. Since the vaccine is being developed for prophylaxis against infection, the primary analysis was conducted in women who were naïve to HPV 16 at enrollment and remained free of HPV 16 infection through the completion of the vaccination series. Among placebo recipients within this cohort, 41 cases of persistent HPV 16 infection were detected. None of the women who received HPV 16 L1 VLP vaccine in this cohort developed an endpoint case [11].
Currently, we are investigating a quadrivalent vaccine, composed of HPV 6, 11, 16, and 18 L1 VLPs formulated with MAA. This vaccine has also been shown to be effective in preventing persistent infection and HPV 6, 11, 16 and 18 related cervical dysplasia [12]. To better understand the quantitative and qualitative effects of Merck aluminum adjuvant on the immunogenicity of the VLPs we used a novel HPV type-specific, antibody isotyping assay and a competitive Luminex immunoassay (cLIA) to measure HPV type-specific antibody responses in rhesus macaques. The results presented here show that formulation with Merck aluminum adjuvant increased the VLP's immunogenicity without affecting the isotype profile.
Methods
Vaccines
Virus-like particles were prepared as previously described (with modifications) from individual lysates of types- Saccharomyces cerevisiae expressing the L1 genes of HPV 6, 11, 16 and 18, respectively [13]. Equal concentrations of all four HPV VLPs were combined and used either directly, or adsorbed to MAA. A standard vaccine dose was composed of 2 μg each of the four VLP types, with or without 225 μg of MAA. The Merck Aluminum Adjuvant is a proprietary aluminum hydroxyphosphate sulfate based adjuvant used in other vaccines manufactured by Merck & Co., [14].
Vaccine study on Rhesus Macaques
Groups of male and female Rhesus macaque (n = 5) were immunized at weeks 0, 8 and 24 with the two experimental vaccines described above. Serum was collected at weeks 0, 2, 4, 8, 10, 12, 16, 20, 24, 26, 28 and 52. The animals were maintained in accordance with the Institutional Animal Care and Use Committees of Merck Research Laboratories (West Point, PA).
Antibodies
The monoclonal antibodies used in the HPV cLIA included H6.M48 [15] for HPV 6, K11.B2 for HPV 11, H16.V5 [16] for HPV 16 and H18.J4 [16] for HPV 18. These antibodies have been shown to be HPV type-specific and to bind to neutralizing epitopes [17]. The mAbs were conjugated to phycoerythrin (Chromaprobe, Aptos, CA) and used at a final concentration of 0.1 μg/mL each.
HPV-specific immunoglobulin isotyping assay
A multiplexed antibody isotyping Luminex assay was developed to characterize the qualitative aspects of the HPV-specific humoral immune response. The multiplexed, isotyping assay was used to classify HPV-specific antibodies as IgM, IgA, IgE, IgG1, IgG2, IgG4, or total IgG. IgG3 was not evaluated since rhesus macaques do not make the IgG3 subclass of antibodies [18,19]. A panel of isotype-specific monoclonal antibodies to IgE, IgG1, IgG2, and IgG4 were purchased from either Sigma (St. Louis, MO) or Southern Biotechnology (Birmingham, AL) and isotype-specific polyclonal antibodies to IgA, IgM, and total IgG were purchased from Rockland Inc. (Gilbertsville, PA). These antibodies (Table 1) were evaluated for reactivity to human, rhesus and African green monkey immunoglobulins. Specificity of the antibodies to their respective immunoglobulin isotypes was established by a Luminex assay using purified human IgM, IgA, IgE, IgG1, IgG2, IgG3 and IgG4 covalently conjugated to Luminex microspheres. Microspheres were incubated overnight with Antibody Depleted Human Serum (ADHS), washed and detected using the isotype specific antibodies at the concentrations shown in Table 1. Optimum antibody concentrations, shown in Table 1, were established by comparing differing concentrations of the detection Abs in titrations of sera prior to immunization and positive sera from HPV 6, 11, 16 and 18L1 VLP vaccinated African green monkeys. Ab concentrations were chosen that gave the most sensitivity and the greatest signal to noise ratio between pre-immune and sera from vaccinated animals. Cross-reactivity between isotype detection antibodies was examined by conjugating purified human Igs to Luminex microspheres and incubating these microspheres with each of the isotyping antibodies at their chosen assay concentrations. In these experiments we detected some cross-reactivity of the anti IgG1 MAb towards IgG2 (71%) and the anti IgG2 towards IgG1 (58%). This cross-reactivity was the rationale for decreasing the concentration of the anti-IgG1 and IgG2 MAbs to 10 μg/mL (Table 1).
Table 1 Isotype specific antibodies and streptavidin-PE used in the HPV antibody isotyping assays
Isotype Reagent Type or MAb Identity Specificity Source Supplier Concentration
Primary
IgM polyclonal Monkey IgM (mu chain specific) Goat Rockland, Gilbertsville, PA 50 μg/mL
IgA polyclonal Monkey IgA (alpha chain specific) Goat Rockland, Gilbertsville, PA 25 μg/mL
IgE HP-6029 Human IgE (epsilon chain specific) Mouse Southern Biotech, Birmingham, AL 25 μg/mL
IgG1 HP-6091or 8c/639 Human IgG1 Mouse Sigma, St. Louis, MO 10 μg/mL
IgG2 HP-6014 Human IgG2 Mouse Sigma, St. Louis, MO 10 μg/mL
IgG4 HP-6025 Human IgG4 Mouse Sigma, St. Louis, MO 25 μg/mL
Total IgG polyclonal Monkey IgG (gamma chain specific) Goat Rockland, Gilbertsville, PA 5 μg/mL
Secondary
PE- Streptavidin NA NA NA Rockland, Gilbertsville, PA 5 μg/mL
For the multiplexed, HPV-specific antibody isotyping assay, HPV 6, 11, 16, and 18 VLPs were covalently conjugated to four distinct Luminex microspheres identified as microsphere 6, 11, 16 and 18, respectively. Test sera from vaccinated rhesus macaques were serially diluted 5-fold starting from a dilution of 1:10 with ADHS and incubated overnight with the VLPs conjugated to microspheres (5,000 VLP-microspheres for each type in PBS + 1% Triton X-100 (PBST) in a total volume of 100 μl in a 1.2 μm hydrophilic, low protein binding, Durapore® membrane filter plate (Millipore, Bedford, MA). The plate contents were washed 3 times with 200 μl of PBST and resuspended in PBST. HPV type-specific antibodies bound to the VLPs were incubated for 2 hours with biotinylated, isotype-specific secondary detection antibodies at the concentrations shown in Table 1. The microspheres were washed three times, and resuspended in PBST. The biotinylated antibodies were detected by incubation with streptavidin conjugated to phycoerythrin (Strep-PE) for 30 min at a final concentration of 5 μg/mL. The plates were washed 3 times and read on a Bio-plex analyzer purchased from Bio-Rad Laboratories, Inc. (Hercules, CA). The Bio-plex analyzer, based on Luminex xMAP technology, is a modified flow cytometer that allows for simultaneous quantitation of up to 100 analytes in a single well and reports Median Fluorescence Intensity (MFI) signals from the Strep-PE detection reagent [20]. End-point dilution titers were defined by comparing signals to those of ADHS, the negative control. For positivity, an MFI value had to be above the average of ADHS + 10 standard deviations and the end-point dilution had to demonstrate positivity at the previous lower dilution.
HPV type-specific competitive Luminex Immunoassay (cLIA)
An optimized and previously validated HPV competitive Luminex Immunoassay (cLIA) first described by Opalka et al. [21] was used to quantify HPV 6, 11, 16, and 18 specific antibodies to known-neutralizing epitopes on VLPs. Using the same VLP-conjugated microspheres described above, antibody titers were determined using a competitive format, where HPV type-specific phycoerythrin (PE)-labeled monoclonal antibodies (MAbs) to known neutralizing epitopes compete with serum antibodies for binding to conformationally dependent, neutralizing epitopes. The fluorescent signals from bound HPV-specific MAbs are inversely proportional to the subject's neutralizing antibody titers. Relative inhibition of MAb binding in test serum was compared to a standard reference serum using a four-parameter logistic curve fit [22]. The reference sera for HPV 6, 11, 16 and 18 used for the standard curve were assigned arbitrary values expressed in milli-Merck Units per milliliter (mMU/mL). An antibody titer of >200 mMU/mL for HPV 11 has been shown to neutralize ~108 virions in the athymic mouse xenograft assay [23]. The titers for HPV 6 and 11 reference sera were previously determined in a pseudoneutralization assay [17]. The lower limit of quantitation for HPV6, 11 and 18 cLIAs is 8 mMU/mL and for HPV 16 is 12 mMU/mL.
Statistical Analysis
Antibody titers obtained from the HPV cLIA were log-transformed and analyzed using a one-tailed, non-paired, Student's t-Test. P-values were obtained by comparing the average log-transformed titers of each of the 5 monkeys from week 2 to 52 within the two groups. The existence of a statistically significant difference between antibody titers from rhesus macaques vaccinated with VLP + MAA versus VLP alone was shown by a P-value of less than 0.05. Error bars represent the standard error of log-transformed titers.
Results
Vaccination with HPV VLPs formulated with aluminum adjuvant induces significantly higher titers than VLPs alone
The purpose of this study was to characterize the quantitative and qualitative effects of including MAA to the HPV 6, 11, 16 and 18 VLP vaccine. First we wanted to measure antibody titers to neutralizing epitopes on HPV L1 induced by a quadrivalent vaccine formulated with or without aluminum adjuvant using the HPV cLIA. Antibody titers for all four HPV types in both vaccine groups developed after the first vaccination, declined, and then increased again following each boost (Fig. 1). The data are consistent with a typical prime boost response. Peak antibody titers four weeks post dose 3, at week 28, for all four HPV types were 12.7 to 41.9-fold higher for the group vaccinated with VLPs + MAA compared to the group vaccinated with VLPs alone, and between 4.3 to 26.7-fold higher in the persistence phase at week 52 (Table 2A). Student's t-Tests of log-transformed antibody titers at week 52 of the VLP + MAA versus the VLP alone group for HPV 6, 11, 16 and 18 also revealed statistically significant P-values of 0.045, 0.026, 0.039 and 0.008 respectively. Student's t-Tests of log-transformed antibody titers from weeks 2–52 of the VLP+MAA versus the VLP alone group for HPV 6, 11, 16 and 18 also revealed statistically significant P-values of 0.0022, 0.0026, 0.0021, and 0.0026 respectively (Fig. 1).
Table 2 Fold difference in geometric mean titers of VLP+MAA-immunized group over VLP alone-immunized group.
A. HPV cLIA time (weeks)
HPV type 0 2 4 8 10 12 16 20 24 26 28 52
HPV 6 1.1 3.2 2.4 4.3 10.0 12.6 36.3 10.9 6.3 22.7 41.9 4.3
HPV 11 0.8 2.9 2.2 5.5 23.5 18.1 16.2 8.2 6.1 18.4 36.8 5.3
HPV 16 1.0 5.2 8.8 8.4 21.4 15.7 16.1 11.4 17.7 4.8 12.7 6.4
HPV 18 1.0 4.8 13.7 19.4 70.5 18.6 27.1 35.4 26.3 6.7 26.8 26.7
B. HPV Total IgG antibodies time (weeks)
HPV type 0 2 4 8 10 12 16 20 24 26 28 52
HPV 6 1.9 9.5 9.5 13.1 125.0 328.3 125.0 65.7 34.4 34.5 65.7 18.1
HPV 11 1.9 13.1 25.0 25.0 125.0 90.6 125.0 47.6 34.5 34.5 90.6 18.1
HPV 16 0.9 5.0 9.5 6.9 65.7 65.7 90.6 65.7 13.1 13.1 65.7 25.0
HPV 18 0.7 9.5 13.1 25.0 125.0 47.6 65.7 65.7 25.0 47.6 125.0 90.6
Figure 1 HPV 6, 11, 16 and 18 type-specific antibody titers measured in the competitive Luminex Immunoassay (cLIA). Sera from rhesus macaques immunized with 2 μg each of HPV 6, 11, 16 and 18L1-VLPs formulated with Merck Aluminum Adjuvant (MAA) (-◆-) or 2 μg each of HPV 6, 11, 16 and 18L1-VLPs alone (-◊ -) were collected a t the indicated time points and tested for Abs to neutralizing epitopes on HPV 6, 11, 16 and 18 using the HPV cLIA. Responses are reported as GMTs in milli-Merck Units per milliliter (mMU/mL) (n = 5 animals per group) for HPV 6, 11, 16 and 18 (Fig 1A, 1B. 1C. and 1D respectively). Arrows indicate vaccination boosts at weeks 8 and 24. A one-tailed, un-paired t-Test analysis was conducted on log-transformed antibody titers obtained from the cLIA. Error bars represent the standard error of the titers within each group.
We also wanted to measure the total HPV-specific IgG levels to determine whether the formulation with MAA increased the total HPV-specific Ab titers. The total HPV-specific IgG peak titers, and titers seven months post dose three were consistently higher in the VLP+MAA group compared to the VLP group alone (Fig. 2). An anamnestic response was observed after each vaccine boost and total IgG levels were detectable through one year. The peak antibody titers four weeks post dose 3, at week 28, in the VLP+MAA group were between 65.7 to 125.0-fold higher than those in the VLP alone group throughout the vaccination series and between 18.1 to 90.6-fold higher in the persistence phase, at week 52 (Table 2B).
Figure 2 HPV 6, 11, 16 and 18 L1 VLP-specific total IgG antibody titers. Sera from rhesus macaques immunized at week 0, 8 and 24 with 2 μg each of HPV 6, 11, 16 and 18L1-VLP formulated with Merck Aluminum Adjuvant (+MAA) (-◆-) or 2 μg each of HPV 6, 11, 16 and 18L1-VLPs alone (-MAA) (-◊ -) were collected at the indicated time points and tested for HPV 6, 11, 16 and 18 L1 VLP specific total IgG (-◆-, -◊ -) titers in a multiplexed detection assay. Responses are reported as the geometric means of end point dilution titers (n = 5 animals per group) for HPV 6, 11, 16 and 18 (2A, B, C, and D respectively). Arrows indicate vaccination boosts at weeks 8 and 24. The starting dilution for each sample was 1:10 and responses above an end point dilution of 10 were considered positive.
HPV-specific antibody isotype responses
To characterize the antibody isotypes elicited by VLPs formulated with MAA or VLPs, alone we modified a multiplexed Ab isotyping assay first developed to measure Ab isotypes in humans [24]. These changes included using detection mAbs that react with human, rhesus and African green monkey Igs, where possible (Table 1). In addition we used biotinylated Abs so that we could use a common streptavidin-PE detection reagent. This novel HPV-specific antibody isotyping assay was found to be specific and sensitive for the detection of rhesus macaque IgM, IgA, IgE, IgG4 and total IgG. Low cross reactivity was seen with the anti-IgG1 and IgG2 antibodies to purified human IgG2 and IgG1 respectively. Responses to these antibody isotypes were measured separately for HPV 6, 11, 16 and 18 and were similar to each other. Results for HPV16 responses which are representative of all four types are shown in figure 3.
Figure 3 HPV L1-VLP specific IgM, IgA, IgG1 and IgG4 antibody titers. Sera from rhesus macaques immunized at week 0, 8 and 24 with 2 μg each of HPV 6, 11, 16 and 18 L1-VLPs formulated with Merck Aluminum Adjuvant (+MAA) (-◆-) or 2 μg each of HPV 6, 11, 16 and 18L1-VLPs alone (-MAA) (-◊ -) were collected at the indicated time points and tested for HPV L1 VLP specific (A) IgM, (B) IgA, (C) IgG1 and (D) IgG4 titers in a multiplexed detection assay. Responses are reported as GMTs (n = 5 monkeys per group) for HPV 16. Arrows indicate vaccination boosts at weeks 8 and 24. The starting dilution for each sample was 1:10 and responses above an end point dilution of 10 were considered above background. Graphs shown are for HPV 16 and are representative of HPV 6, 11, and 18.
We first examined the HPV-specific IgM responses because IgM antibodies are the first to be produced following vaccination and would indicate a primary immune response to the vaccine. As expected, an IgM response for all four types was measured after the first immunization at week 0 and is shown for the representative type HPV 16 (Fig. 3A). Interestingly, two additional primary IgM responses were observed after each of the booster vaccinations given at weeks 8 and 24 in both the VLP and VLP+MAA groups of the study (Fig. 3A). IgM responses in the VLP alone group returned to baseline by week 52, which was seven months post dose 3, while an IgM antibody response in the VLP + MAA group was detectable at week 52.
Since HPVs are sexually transmitted viruses that cross the mucosal barrier in the anogenital tract, we next examined whether the VLPs induced an IgA response and what effect formulation with MAA would have on an IgA response. In theory, the induction of IgA would greatly enhance the effectiveness of an HPV vaccine. Both vaccine groups demonstrated type-specific IgA responses to the VLPs that were maintained through one year (Fig. 3B). Each vaccine boost elicited secondary IgA responses in both vaccination study groups (Fig. 3B). Higher IgA end-point dilution titers were observed in the rhesus macaques vaccinated with VLP + MAA compared to those vaccinated with VLP alone.
We also examined whether there was an IgE response induced following vaccination. The geometric mean titers (GMTs) of IgE responses for the two vaccine groups were not above background, however one animal in the VLP group had a detectable IgE response at week 4 for HPV 6 and 16, and at week 24 for HPV 16 (data not shown). This response was just barely above the limit of detection and was not detected following the booster doses given at weeks 8 and 24.
We next measured the IgG subtypes to determine whether formulation with Merck aluminum adjuvant affected the Ig isotype profile. Characterizing the IgG isotype profile would also provide insight into whether formulation with MAA altered the TH1 or TH2 profile of the immune response. The main IgG subclass observed in both vaccination groups was IgG1 (Fig. 3C). Similar to responses observed for total IgG, secondary IgG1 responses were observed after each boost, and titers were maintained through one year in the VLP + MAA group (Fig. 3C). However, the group vaccinated with VLP alone had detectable titers only after the second immunization and these titers approached baseline values at one year (Fig 3C). We also observed high levels of IgG4 and Ab titers in the VLP + MAA group were detected early after each vaccination, but were undetectable by week 52 (Fig. 3D). The monkeys vaccinated with VLPs alone only demonstrated IgG4 responses after each vaccination boost, but the titers fell below the limit of detection soon after (Fig. 3D). We did not detect a strong IgG2 response in the VLP-alone group and IgG2 levels in the VLP + MAA group were observed only immediately following vaccination (data not shown). However, because we observed some cross-reactivity of the anti-IgG2 MAb to human IgG1, we cannot rule out the possibility that the IgG2 signals observed were due to cross-reactivity of the anti-IgG2 MAb to the high levels of rhesus IgG1 Ab (Fig. 3A). HPV-specific IgG3 responses were not measured because rhesus macaques do not make the IgG3 subclass of antibodies [18,19]. In summary, higher total IgG, IgG1, and IgG4 Ab titers were observed in the rhesus macaques vaccinated with VLP + MAA compared to those vaccinated with VLPs alone.
Discussion
This is the first study to directly examine the effects of the Merck aluminum adjuvant on immune responses to a quadrivalent HPV 6, 11, 16 and 18 L1 vaccine. The results here were generated using a novel HPV-type specific antibody isotyping assay to characterize the immunoglobulin subclasses to HPV 6, 11, 16 and 18 in rhesus macaques, and a competitive Luminex immunoassay (cLIA) that measures HPV-type specific antibodies to known neutralizing epitopes. The quadrivalent HPV L1 VLP vaccine formulated with MAA induced higher titers for total IgG (Fig. 2), as well as for all Ig isotypes measured (Fig. 3) compared to the vaccine without MAA. Also the MAA-formulated vaccine induced higher overall antibody titers to neutralizing epitopes compared to one without MAA (Fig. 1). Strong immune responses were elicited against all four HPV types and high antibody titers persisted through one year.
Results from the HPV cLIA showed that the vaccine induced antibodies that competed with MAbs to known neutralizing epitopes, and are, thus, theoretically able to neutralize HPV. It is worth noting that we have previously shown that a competitive radioimmunoassay (cRIA) [25] and the cLIA correlate well with an HPV 11 neutralization assay. Antibody titers in the cLIA also followed a "prime-boost" response wherein antibody titers were higher after every vaccine dose. Furthermore, cLIA titers show that formulation of the vaccine with MAA significantly increased peak neutralizing titers, and, more importantly, the differences in the responses were durable, persisting through one year.
Previously, we reported that the Merck HPV 16 L1 VLP vaccine induced a predominantly TH2 immune response with IL-4-producing T-helper cells, high levels of neutralizing Abs and low levels of IFN-γ secreting CD8+ or CD4+ cells [26]. Given that the HPV quadrivalent vaccine also contains the L1 capsid proteins, the response was also expected to be TH2-like [27]. Using the HPV isotyping assay, which is a direct binding assay detecting antibodies to conserved epitopes shared across multiple HPV types, we showed that the isotype profile of the vaccinated rhesus macaques was consistent with a TH2-like immune response. Interestingly, the formulation with MAA did not affect the overall isotype profile of the vaccinated monkeys.
Predictably, abundant amounts of IgG antibody isotype were found, and of all the IgG subclasses, IgG1 was most readily detected. The abundant presence of IgG1 elicited in both vaccination groups affirms that the vaccines did induce immune responses through a TH2 pathway. It is also interesting to note that the next abundant IgG subclass detected, primarily detected in the VLP+MAA group, was IgG4 (Fig. 3D). IgG4 is more characteristic of a TH1 response and, by inducing this IgG subclass, the vaccine is shown to be also effective in eliciting a cellular (TH1) immune response, albeit, predictably not long lasting. It is worth noting that we previously observed low levels of IFN-γ secreting CD8+ and CD4+ T-cells [26], suggesting the vaccine induces a low level cellular immune response. Similar to what we observed in rhesus macaques, isotype analysis of sera from women vaccinated with the HPV 11 VLP vaccine formulated with MAA also detected little to no levels of IgG2 [24]. Isotype analysis performed on an epidemiological study by Wang et al. also found no IgG2 antibody, which they attributed to the unfavorable cytokine profile of HPV infection for the induction of IgG2 [28]. Detection of the IgG3 subclass was not performed, because rhesus macaques do not make IgG3 [18,19].
We also measured high levels of serum IgA observed in vaccinated animals. IgA, the secretory immunoglobulin, is thought to be very important in blocking viral entry into mucosal tissues [29]. These findings are consistent with what we have observe in the sera from women vaccinated with the HPV 11 VLP vaccine [24]. The elicitation of IgA should be advantageous for a vaccine that prevents HPV infection in defending against entry of HPV, which commonly enters through the urogenital tract. The high levels of IgA elicited by the vaccine may also be one reason why the vaccine has been efficacious in human clinical trials [11]. IgE responses were measured because IgE antibodies are implicated in allergic symptoms. Consistent IgE antibody responses were not observed in either group. Only one animal in the VLP alone vaccine group had detectable IgE, but the response was barely above the limit of detection and transient as it was not detected following the booster immunizations. This is consistent with the observation of no allergic reactions or severe adverse events in the monkeys following vaccination. These observations are consistent with our clinical findings that a VLP + MAA vaccine was safe and well-tolerated [30]. In summary, the isotype analysis data are consistent with the VLP+MAA vaccine inducing primarily a TH2-like immune response with high titers of IgA and IgG1 antibodies.
In these isotyping studies, we used polyclonal goat anti-monkey IgM, IgA, and total IgG; because monkey-specific reagents are not available for the IgG subtypes and IgE, we used anti-human mouse MAbs to measure IgG1, IgG2, IgG4, and IgE responses. The polyclonal anti-IgA, IgM and total IgG reagents were selected for their specificity and their ability to broadly detect African Green monkey, rhesus macaque and human antibody isotypes. The various differences between human, rhesus macaque and African green monkey sera could account for changes in the sensitivity and specificity of the antibody-isotyping reagents and so the appropriate antibody reagent concentrations used in the HPV type-specific antibody isotyping assay were optimized for use with rhesus macaque serum.
Due to the unavailability of purified rhesus Igs, we used purified human Igs to determine the specificities of the type-specific antibody isotyping reagents. There was some observed reactivity of the anti- human IgG2 antibody to purified human IgG1, and vice versa, but these data were obtained using purified human Igs and the differences between human and rhesus antibodies may account for why we saw a strong IgG1 response with low IgG2 titers (Fig. 3). However, because we cannot rule out that the IgG2 titers were merely a reflection of the cross-reactivity to the high IgG1 titers, we cannot definitively say that there were any IgG2 responses observed.
In some cases, the test animals had low Ab titers to the VLPs prior to vaccination. This was not unexpected since more than 100 papillomaviruses have been identified [31] and many of these share conserved structural epitopes that would cross react in a direct binding assay. For this reason, we used Antibody Depleted Human Serum (ADHS) as a negative control rather than sera from rhesus macaques prior to vaccination.
Currently, the only FDA approved adjuvants for use with prophylactic vaccines are aluminum based [32]. The purpose of this study was to evaluate the effect of the Merck Aluminum Adjuvant on the immunogenicity of the HPV L1 VLP 6, 11, 16 and 18 vaccine and our results show that the addition of the MAA significantly increases the immunogenicity of the vaccine. In sharp contrast to our results, Harro et al. have reported that formulation with adjuvant did not significantly enhance the immunogenicity of an HPV VLP vaccine [33]. In that study, 72 human volunteers were randomized to receive placebo or an HPV 16 L1 VLP vaccine produced in insect cells. Vaccine recipients were given a dose of 10 μg or 50 μg of VLPs formulated without adjuvant, with an aluminum adjuvant, or with MF59 adjuvant. Since no significant differences in neutralizing antibody titers and ELISA titers were found in subjects in all three groups, the authors concluded adjuvant was not required. Our results were obtained using the quadrivalent Merck HPV-L1 VLP vaccine, which is formulated with Merck aluminum adjuvant. While the mechanisms by which aluminum adjuvants enhance immunogenicity are poorly understood, they may act to form an antigen depot, allowing the vaccine to linger at the site of administration and perhaps, more importantly, stimulate the immune system by inducing higher neutralizing Ab titers [34]. The inclusion of MAA to the vaccine did not induce a qualitatively different humoral immune response in that the isotype profile was similar between animals vaccinated with the two quadrivalent vaccines (Fig. 3). However, data from both the antibody isotyping assays and the cLIA showed that the inclusion of MAA to the VLP vaccine induced significantly higher Ab titers (Fig. 1, 2, 3). These data clearly show a benefit to the formulation of the quadrivalent HPV-L1 VLPs with MAA.
The differences observed in the immune responses elicited by the vaccine reported by Harro et al. and the Merck quadrivalent HPV vaccine may be attributed to several factors. First, would be the differences in the formulation and purification processes of the vaccines. The vaccine reported by Harro et al. was produced in insect cells as opposed to yeast for the Merck vaccine. Second, the differences in vaccine recipients, humans and rhesus macaques may respond differently to adjuvanted HPV VLP vaccines. Lastly, differences in formulation and composition of the two aluminum adjuvants used could have profound impacts on immunogenicity. Different aluminum salt adjuvants were used in these two studies; our study adsorbed VLPs to a proprietary formulation of aluminum hydroxyphosphate sulfate (MAA), while the Harro et al. study used aluminum potassium sulfate. Without a head to head comparison of the two vaccines, it is difficult to establish why the results are different. However, the magnitude and durability of antibody responses found in the quadrivalent VLP+MAA vaccinated rhesus macaques reported in this study are consistent with those found in human clinical trials of the HPV 11 or HPV 16 monovalent vaccines [30].
The isotyping and cLIA assays described in this study are useful in characterizing the immune responses to HPV virions after natural infection or following vaccination. The novel cLIA utilized in this study is a robust and sensitive method for detecting HPV-specific antibodies in sera to known neutralizing epitopes on HPV virions and has proven to be a valuable tool for monitoring HPV immune responses in human clinical trials of an HPV 6, 11, 16 and 18L1-VLP vaccine. These assays have potential use in future epidemiology studies and other vaccine clinical trials.
Conclusion
The results presented here show that an HPV 6, 11, 16 and 18 L1-VLP vaccine formulated with Merck aluminum adjuvant has increased immunogenicity without affecting the isotype profile. The VLPs formulated with Merck aluminum adjuvant elicited a robust and durable immune response that lasted up to 52 weeks. This vaccine holds promise as a vaccine for preventing cervical cancer.
List of abbreviations
MAb, monoclonal antibody; PBS, phosphate-buffered saline; cLIA, competitive Luminex immunoassay; ADHS, antibody depleted human serum; VLP, virus-like particles; HPV, human papillomavirus; GMT, geometric mean titers; Ig, immunoglobulin; Ab, antibody.
Competing interests
W. Ruiz, W. McClements, K. Jansen and M. Esser were all employees of Merck Research Laboratories, a division of Merck & Co., Inc. when this study was performed and potentially own stock and/or hold stock options in the Company. Merck is developing a quadrivalent HPV vaccine. Merck also funded this study in its entirety.
Authors' contributions
W. McClements and K. Jansen designed the immunization study. M. Esser and W. Ruiz developed the competitive Luminex immunoassay and the new isotyping assays. W. Ruiz performed all the laboratory assays. W. Ruiz and M. Esser wrote the manuscript and all authors read and approved the final manuscript.
Acknowledgements
We would like to acknowledge R. Skibbens and M. Kuchka (Lehigh University, Bethlehem, PA); L. Kierstead, A. Finnefrock, J. Drummond, N. Chirmule, E. Shaw, S. Schlottman, G. Wang, J. F. Smith, G. Page, the HPV and HIV groups (Merck, MRL), for their assistance and support; R. Charbonneau and S. Macmullen for previous isotyping development, P. Boerckel for her expert laboratory assistance, T. Green, for expert statistical advice and assistance, R. Marchese (Merck MRL Wayne), for her critical reading of the manuscript and expert guidance, and T. Fujiwara for his continued motivation and encouragement.
==== Refs
Walboomers JM Jacobs MV Manos MM Bosch FX Kummer JA Shah KV Snijders PJ Peto J Meijer CJ Munoz N Human papillomavirus is a necessary cause of invasive cervical cancer worldwide J Pathol 1999 189 12 19 10451482 10.1002/(SICI)1096-9896(199909)189:1<12::AID-PATH431>3.0.CO;2-F
Eiben GL da Silva DM Fausch SC Le Poole IC Nishimura MI Kast WM Cervical cancer vaccines: recent advances in HPV research Viral Immunol 2003 16 111 121 12828864 10.1089/088282403322017866
Pisani P Bray F Parkin DM Estimates of the world-wide prevalence of cancer for 25 sites in the adult population Int J Cancer 2002 97 72 81 11774246 10.1002/ijc.1571
Laukkanen P Koskela P Pukkala E Dillner J Laara E Knekt P Lehtinen M Time trends in incidence and prevalence of human papillomavirus type 6, 11 and 16 infections in Finland J Gen Virol 2003 84 2105 2109 12867641 10.1099/vir.0.18995-0
zur Hausen H Viruses in human cancers Science 1991 254 1167 1173 1659743
Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: baseline data from a randomized trial. The Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group J Natl Cancer Inst 2000 92 397 402 10700419 10.1093/jnci/92.5.397
Bosch FX Manos MM Munoz N Sherman M Jansen AM Peto J Schiffman MH Moreno V Kurman R Shah KV Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group J Natl Cancer Inst 1995 87 796 802 7791229
Schiffman MH Bauer HM Hoover RN Glass AG Cadell DM Rush BB Scott DR Sherman ME Kurman RJ Wacholder S Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia J Natl Cancer Inst 1993 85 958 964 8388478
Greer CE Wheeler CM Ladner MB Beutner K Coyne MY Liang H Langenberg A Yen TS Ralston R Human papillomavirus (HPV) type distribution and serological response to HPV type 6 virus-like particles in patients with genital warts J Clin Microbiol 1995 33 2058 2063 7559948
Brown DR Schroeder JM Bryan JT Stoler MH Fife KH Detection of multiple human papillomavirus types in Condylomata acuminata lesions from otherwise healthy and immunosuppressed patients J Clin Microbiol 1999 37 3316 3322 10488198
Koutsky LA Ault KA Wheeler CM Brown DR Barr E Alvarez FB Chiacchierini LM Jansen KU A controlled trial of a human papillomavirus type 16 vaccine N Engl J Med 2002 347 1645 1651 12444178 10.1056/NEJMoa020586
Villa LL Costa RLR Petta CA Andrade RP Ault KA Giuliano AR Wheeler CM Koutsky LA Malm C Lehtinen M Skjeldestad FE Olsson SE Steinwall M Brown DR Kurman RJ Ronnett BM Stoler MH Ferenczy A Harper DM Tamms GM Yu J Lupinacci L Railkar R Taddeo FJ Jansen KU Esser MT Sings HL Saah AJ Barr E A Randomized, Double-blind, Placebo-controlled Efficacy Trial of a Prophylactic Quadrivalent Human Papillomavirus (Types 6/11/16/18) L1 Virus-Like Particle Vaccine in Young Women Lancet Oncology 2005
Cook JC Joyce JG George HA Schultz LD Hurni WM Jansen KU Hepler RW Ip C Lowe RS Keller PM Lehman ED Purification of virus-like particles of recombinant human papillomavirus type 11 major capsid protein L1 from Saccharomyces cerevisiae Protein Expr Purif 1999 17 477 484 10600468 10.1006/prep.1999.1155
Baylor NW Egan W Richman P Aluminum salts in vaccines--US perspective Vaccine 2002 20 Suppl 3 S18 23 12184360 10.1016/S0264-410X(02)00166-4
Christensen ND Reed CA Cladel NM Hall K Leiserowitz GS Monoclonal antibodies to HPV-6 L1 virus-like particles identify conformational and linear neutralizing epitopes on HPV-11 in addition to type-specific epitopes on HPV-6 Virology 1996 224 477 486 8874508 10.1006/viro.1996.0554
Christensen ND Dillner J Eklund C Carter JJ Wipf GC Reed CA Cladel NM Galloway DA Surface conformational and linear epitopes on HPV-16 and HPV-18 L1 virus-like particles as defined by monoclonal antibodies Virology 1996 223 174 184 8806551 10.1006/viro.1996.0466
Yeager MD Aste-Amezaga M Brown DR Martin MM Shah MJ Cook JC Christensen ND Ackerson C Lowe RS Smith JF Keller P Jansen KU Neutralization of human papillomavirus (HPV) pseudovirions: a novel and efficient approach to detect and characterize HPV neutralizing antibodies Virology 2000 278 570 577 11118379 10.1006/viro.2000.0674
Martin LN Chromatographic fractionation of rhesus monkey (Macaca mulatta) IgG subclasses using deae cellulose and protein A-sepharose J Immunol Methods 1982 50 319 329 7097010 10.1016/0022-1759(82)90170-3
Calvas P Apoil P Fortenfant F Roubinet F Andris J Capra D Blancher A Characterization of the three immunoglobulin G subclasses of macaques Scand J Immunol 1999 49 595 610 10354371 10.1046/j.1365-3083.1999.00540.x
Fulton RJ McDade RL Smith PL Kienker LJ Kettman JRJ Advanced multiplexed analysis with the FlowMetrix system Clin Chem 1997 43 1749 1756 9299971
Opalka D Lachman CE MacMullen SA Jansen KU Smith JF Chirmule N Esser MT Simultaneous quantitation of antibodies to neutralizing epitopes on virus-like particles for human papillomavirus types 6, 11, 16, and 18 by a multiplexed Luminex assay Clin Diagn Lab Immunol 2003 10 108 115 12522048 10.1128/CDLI.10.1.108-115.2003
O'Connell MA Belanger BA Haaland PD The four parameter logistic model for calibration and assay development American Statistical Association Proceedings of the Biopharm Section 1992 180 185
Bryan JT Jansen KU Lowe RS Fife KH McClowry T Glass D Brown DR Human papillomavirus type 11 neutralization in the athymic mouse xenograft system: correlation with virus-like particle IgG concentration J Med Virol 1997 53 185 188 9365880 10.1002/(SICI)1096-9071(199711)53:3<185::AID-JMV1>3.0.CO;2-4
Emeny RT Wheeler CM Jansen KU Hunt WC Fu TM Smith JF MacMullen S Esser MT Paliard X Priming of human papillomavirus type 11-specific humoral and cellular immune responses in college-aged women with a virus-like particle vaccine J Virol 2002 76 7832 7842 12097595 10.1128/JVI.76.15.7832-7842.2002
Brown DR Bryan JT Schroeder JM Robinson TS Fife KH Wheeler CM Barr E Smith PR Chiacchierini L DiCello A Jansen KU Neutralization of human papillomavirus type 11 (HPV-11) by serum from women vaccinated with yeast-derived HPV-11 L1 virus-like particles: correlation with competitive radioimmunoassay titer J Infect Dis 2001 184 1183 1186 11598842 10.1086/323645
Tobery TW Smith JF Kuklin N Skulsky D Ackerson C Huang L Chen L Cook JC McClements WL Jansen KU Effect of vaccine delivery system on the induction of HPV16L1-specific humoral and cell-mediated immune responses in immunized rhesus macaques Vaccine 2003 21 1539 1547 12615451 10.1016/S0264-410X(02)00679-5
Chirmule N Current Advances in Molecular Immunology: Reference guide for reviews on molecular vaccines Frontiers in Bioscience 2004 9 2373 2387 15353293
Wang ZH Kjellberg L Abdalla H Wiklund F Eklund C Knekt P Lehtinen M Kallings I Lenner P Hallmans G Mahlck CG Wadell G Schiller J Dillner J Type specificity and significance of different isotypes of serum antibodies to human papillomavirus capsids J Infect Dis 2000 181 456 462 10669326 10.1086/315232
Brandtzaeg P Role of secretory antibodies in the defence against infections Int J Med Microbiol 2003 293 3 15 12755363
Fife KH Wheeler CM Koutsky LA Barr E Brown DR Schiff MA Kiviat NB Jansen KU Barber H Smith JF Tadesse A Giacoletti K Smith PR Suhr G Johnson DA Dose-ranging studies of the safety and immunogenicity of human papillomavirus Type 11 and Type 16 virus-like particle candidate vaccines in young healthy women Vaccine 2004 22 2943 2952 15246631 10.1016/j.vaccine.2003.11.058
de Villiers EM Fauquet C Broker TR Bernard HU zur Hausen H Classification of papillomaviruses Virology 2004 324 17 27 15183049 10.1016/j.virol.2004.03.033
Marciani DJ Vaccine adjuvants: role and mechanisms of action in vaccine immunogenicity Drug Discov Today 2003 8 934 943 14554157 10.1016/S1359-6446(03)02864-2
Harro CD Pang YY Roden RB Hildesheim A Wang Z Reynolds MJ Mast TC Robinson R Murphy BR Karron RA Dillner J Schiller JT Lowy DR Safety and immunogenicity trial in adult volunteers of a human papillomavirus 16 L1 virus-like particle vaccine J Natl Cancer Inst 2001 93 284 292 11181775 10.1093/jnci/93.4.284
HogenEsch H Mechanisms of stimulation of the immune response by aluminum adjuvants Vaccine 2002 20 Suppl 3 S34 9 12184362 10.1016/S0264-410X(02)00169-X
| 15842730 | PMC1097753 | CC BY | 2021-01-04 16:37:45 | no | J Immune Based Ther Vaccines. 2005 Apr 20; 3:2 | utf-8 | J Immune Based Ther Vaccines | 2,005 | 10.1186/1476-8518-3-2 | oa_comm |
==== Front
J NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 1742-2094-2-111583310910.1186/1742-2094-2-11ResearchPlatelet-activating factor enhancement of calcium influx and interleukin-6 expression, but not production, in human microglia Sattayaprasert Prasongchai [email protected] Hyun B [email protected] Sukumal [email protected] James G [email protected] Department of Pharmacology and Therapeutics, Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada2 Division of Neurology, Department of Medicine, University of British Columbia, Canada3 Department of Anatomy, Mahidol University, Bangkok, Thailand2005 15 4 2005 2 11 11 19 1 2005 15 4 2005 Copyright © 2005 Sattayaprasert et al; licensee BioMed Central Ltd.2005Sattayaprasert 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.
Calcium-sensitive fluorescence microscopy and molecular biology analysis have been used to study the effects of platelet-activating factor (PAF) on intracellular calcium [Ca2+]i and IL-6 expression in human microglia. PAF (applied acutely at 100 nM) elicited a biphasic response in [Ca2+]i consisting of an initial rapid increase of [Ca2+]i due to release from internal stores, followed by a sustained influx. The latter phase of the [Ca2+]i increase was blocked by SKF96365, a non-selective store-operated channel (SOC) inhibitor. RT-PCR analysis showed PAF treatment of microglia induced expression of the pro-inflammatory cytokine IL-6 in a time-dependent manner which was blocked in the presence of SKF96365. However, ELISA assay showed no production of IL-6 was elicited at any time point (1–24 h) for microglial exposures to PAF. These findings suggest that PAF stimulation of human microglia induces expression, but not production, of IL-6 and that SOC-mediated [Ca2+]i influx contributes to the enhanced expression of the cytokine.
Microgliaplatelet-activating factorinterleukin-6store-operated channels
==== Body
Background
Microglia are resident, immunocompetent cells in the brain. They show functional plasticity and can be activated by a diversity of inflammatory stimuli including ones associated with neurodegenerative diseases [9,18]. The functional responses of microglia following activation include proliferation, phagocytosis and secretion. In the latter case microglia can secrete pro- and anti-inflammatory cytokines, chemokines, neurotrophic factors and excitotoxins such as glutamate [20].
One important inflammatory agent is platelet-activating factor (PAF), an alkyl ether phospholipid compound, which both stimulates and is produced by microglia [13]. PAF contributes to inflammatory responses in the brain and is reported to be upregulated in CNS pathophysiology [2,17]. Acute application of PAF to human microglia induces a biphasic change in levels of intracellular Ca2+ ([Ca2+]i) with an initial rapid phase due to intracellular release from endoplasmic reticulum (ER) stores and a secondary phase due to influx through store operated channels (SOC) [15,31]. Importantly, SOC has been shown to exhibit sustained activation following stimulation of human [31] and rodent [29] microglia. Prolonged entry of Ca2+ through SOC in stimulated microglia could constitute a coupling signal between an activating stimulus and cellular functional response. Indeed, the involvement of sustained Ca2+ responses has been reported as a factor in the production of arachidonic acid by rat microglia [23].
The pro-inflammatory cytokine IL-6 is released from activated microglia and mediates inflammatory responses in brain. Levels of IL-6 in serum and cerebrospinal fluid have been found to be elevated in stroke patients [8,28] and the cytokine has also been implicated in the etiopathology of neurodegenerative disorders such as Alzheimer's disease (AD), Parkinson's disease (PD) and HIV encephalopathy [3,14,25]. Interestingly, some evidence is also available suggesting that under some conditions elevated levels of IL-6 in brain may actually be beneficial [27].
In this study we have examined a role for SOC mediated [Ca2+]i influx in mediating actions of the inflammatory stimulus PAF to induce IL-6 in human microglia.
Materials and methods
Preparation of cells
The procedures for the isolation of human microglia have been previously reported [24]. In brief, human embryonic brain tissues were dissected into small blocks, incubated in phosphate-buffered saline (PBS) containing 0.25% trypsin and 40 μg/ml DNase and then dissociated into single cells by repeated pipetting. Cells were plated in T75 flasks in a medium consisting of Dulbecco's modified Eagle's medium (DMEM) containing 5% horse serum, 5 mg/ml glucose, 25 μg/ml gentamicin, and 2.5 μg/ml amphotericin B. Freely floating microglia were harvested from a medium of mixed cell cultures after 7–10 days of growth in culture flasks and plated on aclar coverslips for identification, on poly-L-lysine-coated glass coverslips for calcium spectrofluorometry and plated on six-well multiplates for RT-PCR or ELISA. CD11b and ricinus communis agglutinin (RCA), specific markers for microglia, were used to confirm purity of the culture which was in excess of 98% [24,30].
Calcium spectrofluorometry
The procedures used for measurement of intracellular Ca2+ have been reported [6,31]. Microglia were incubated with 1 μM fura-2/AM (acetoxymethyl ester, Molecular Probes, Eugene, OR) plus 1 μM pluronic acid in normal physiological saline solution (PSS) for 30 min. PSS solution contained (in mM): NaCl (126), KCl (5), MgCl2 (1.2), HEPES (10), D-glucose (10) and CaCl2 (1); pH of 7.4. All reagents were obtained from Sigma (St. Louis, MO).
Following a 20 min wash in dye-free solution, coverslips were placed on the stage of a Zeiss Axiovert inverted microscope employing a ×40 quartz objective lens. Cells were exposed to alternating wavelengths of 340/380 nm at 6 s intervals and emission light passed through a 510 nm filter. An imaging system (Empix Imaging, Mississauga, ON) was used to record fluorescence ratios using a CCD camera (Retiga 1300i, Burnaby, BC). Fluorescence ratios were determined and converted to values of [Ca2+]i using published procedures [11]. All experiments were done at room temperature (20–22°C).
Reverse transcription-PCR and ELISA assay
IL-6 expression was detected with the reverse-transcriptase polymerase chain reaction (RT-PCR). Isolation of RNAs was performed using TRIzol (Gibco-BRL, Gaithersburg, MD, USA) and DNA contamination was eliminated using DNase. cDNA synthesis was done using M-MLV reverse transcriptase (Gibco-BRL). The sequences for the human specific primers for IL-6 as follows: sense primer: 5'-GTGTGAAAGCAGCAAAGAGGC-3'; antisense primer: 5'-CTGGAGGTACTCTAGGTATAC-3'. Human-specific IL-6 signals were generated with the GeneAmp thermal cycler and Amplitaq Gold DNA polymerase (Applied Biosystems, Foster City, CA). The conditions for PCR were as follows: initial denaturation at 95°C for 6 min followed by 28 cycles of denaturation at 95°C for 45 sec, annealing at 56°C for 1 min and extension at 72°C for 1 min. A final extension step at 72°C for 10 min was carried out. PCR products (159 bp) were identified using 1.5% agarose gels containing ethidium bromide and visualized under UV light. GAPDH was used as a reaction standard and human specific primer sequences were as follows: sense primer: 5'-CCATGTTCGTCATGGGTGTGAACCA-3'; antisense primer: 5'-GCCAGTAGAGGCAGGGATGATGTTC-3'. The intensities of each band were measured using NIH image J 1.24 software (National Institutes of Health, Bethesda, MD). Relative mRNA levels for each treatment were normalized to GAPDH.
Enzyme-linked immunosorbent assays (ELISA) were performed according to manufacturer instructions (R & D systems, Minneapolis, MN). Cells were plated on multi-well plates (≈105 cells/well) and treated with PAF (100 nM) in the absence or presence of SKF96365 (20 μM for 8 hr). The cell-free supernatants were used for analysis of IL-6 production (kit detects IL-6 as low as 0.7 pg/ml). Values were expressed as means ± SEM and statistical significance (p < 0.05) was determined using one-way ANOVA and Newman-Keuls multiple comparison post-test.
Results
Effects of SKF96365 on SOC-mediated [Ca2+]i influx by PAF
PAF-induced changes in [Ca2+]i from human microglia have previously been reported [15,21,31]. Initial study showed a transient increase in SOC [31] but more recent work has shown PAF application to evoke a sustained phase of SOC following an initial component due to depletion of Ca2+ from intracellular stores [15,21]. The differences in PAF responses is considered in the Discussion.
A representative response to acute application of PAF (applied at 100 nM) is presented in Fig 1A (n = 18 cells). A plateau level of [Ca2+]i was sustained for a duration exceeding 2 min after removal of PAF. Following establishment of a clearly defined plateau phase, the bath solution was replaced with Ca2+-free PSS. This procedure caused an immediate decline in [Ca2+]i to baseline levels (Fig 1A). Long durations of SOC-mediated influx of Ca2+ have also been documented in mouse microglial cells [29].
Figure 1 PAF-induced Ca2+ responses. A: Representative trace (n = 18 cells) showing change in [Ca2+]i induced by PAF (100 nM). Following a prolonged level of SOC-mediated influx of Ca2+, the perfusion of Ca2+-free PSS abolished the response. B: Results from a separate experiment showing effects of SKF96365 (20 μM) on a PAF-induced increase in [Ca2+]i (n = 21 cells). SKF96365 application, during a sustained entry of Ca2+ through SOC, effectively reduced [Ca2+] to baseline levels.
The results of application of the SOC inhibitor SKF96365 (at 20 μM) to the plateau phase of a PAF response is shown in the representative recording of Fig 1B (n = 21 cells). SOC-mediated entry of Ca2+ was reduced to baseline values by SKF96365. Amplitude of Ca2+ influx through SOC was measured as the difference between baseline and plateau levels and in five independent experiments (n = 107 cells) the amplitude prior to SKF96365 was 140 ± 21 nM and after SKF96365 was at baseline levels. Previous work has shown SKF96365 pretreatment of human microglia (50 μM for 5 min) abolished a transient SOC in the cells [31].
Effects of SKF96365 on microglial expression of IL-6
We next examined effects of PAF on expression of the pro-inflammatory cytokine IL-6 in the absence and presence of SOC inhibition. The time-dependence of PAF stimulation (100 nM) of human microglia on IL-6 are presented in Fig 2A. The representative RT-PCR showed no constitutive expression of IL-6 in unstimulated microglia (lane 1 of Fig 2A). IL-6 was maximally expressed at 1 h of exposure to PAF then declined to lower levels at longer treatment times (longest exposure of 6 h). A similar time-dependence for IL-6 expression was exhibited in a total of four experiments.
Figure 2 Expression of IL-6 in PAF treated human microglia. A: RT-PCR analysis for different exposure times of microglia to PAF (applied at 100 nM). B: Effects of PAF, PAF plus SKF96365, PAF plus Ca2+-free and SKF96365 applied alone (1 h treatments). Also shown are effects of LPS and LPS plus SKF96365 (6 hr treatments). GAPDH was used as a reaction standard. C: Semi-quantitative RT-PCR for effects of the different treatments. * P < 0.05 compared with unstimulated control; # P < 0.05 compared with PAF treated microglia.
A one hour exposure of human microglia to PAF was chosen for subsequent RT-PCR analysis. As shown in Fig 2B, constitutive expression of IL-6 was absent (lane 1). PAF treatment was effective in stimulating expression of the cytokine (Fig 2B, lane 2). The expression of IL-6 was abolished when SKF96365 was included with the PAF application (Fig 2B, lane 3). No evident IL-6 expression was observed for PAF application in Ca2+-free PSS (Fig 2B, lane 4). SKF96365, applied alone in PSS solution, did not cause any increase in IL-6 (Fig 2B, lane 5).
It was of interest to compare PAF as an inducer of microglial IL-6 to that of LPS (lipopolysaccharide) a potent inflammatory stimulus of cells. The results of exposure of human microglia to LPS (100 ng/ml for 6 h) is presented in Fig 2B (lane 6) showing LPS stimulation caused an intense band for IL-6. Altering the number of PCR cycles had no apparent effect on intensity (data not shown) suggesting IL-6 band saturation with LPS (Fig 2B, lane 6). Comparison of band intensity indicated LPS was a more effective inducer of IL-6 relative to PAF. Interestingly, a partial inhibition of LPS-induced IL-6 mRNA was observed when SKF96365 was applied with LPS (Fig 2B, lane 7).
Semi-quantitative RT-PCR analysis is presented in Fig 2C and shows PAF as an effective stimulator of IL-6 expression (n = 3). However, expression of IL-6 was considerably lower with PAF as a stimulus compared with LPS (Fig 2B,C). Inclusion of SKF96365 with PAF or application of PAF in Ca2+-free PSS eliminated expression of IL-6 (n = 3). Although LPS was not the subject of this study, the decrease in LPS induction of IL-6 with SKF96365 is of interest and is discussed below.
ELISA assay for effects of PAF on microglial production of IL-6
We next investigated production of IL-6 from PAF-treated human microglia using an exposure time of 8 h. No production of IL-6 was evident in four experiments (data not shown); levels of IL-6 were below the detection levels for ELISA assay (≤ 1 pg/ml). In order to determine if the treatment time was a limiting factor in IL-6 production, a series of experiments using different microglial times of exposure to PAF were undertaken (from 1–24 h). The results are presented in Fig 3; no significant production of IL-6 (n = 4) was found for any treatment time (PAF applied for 1,2,8 or 24 h).
Figure 3 ELISA assays for production of IL-6 in human microglia. PAF (at 100 nM) induced no significant production of IL-6 from microglia following exposures from 1–24 h (n = 4 for each time points). PAF (at 1 μM) induced no significant production of IL-6 (following exposures for 8 h and 24 h; n = 3 for both time points); these values are near the lower limits for sensitivity of the ELISA kits. LPS was used as a positive control in these experiments (n = 4); note the change of scale for the ordinate (from 10 to 400 pg/ml). * P < 0.05 compared with unstimulated control.
We also examined if a ten-fold increase in PAF concentration (to 1 μM) would be effective in producing IL-6. As shown in Fig 3, this higher concentration of PAF also had no effect to induce IL-6 production for treatment times of 8 or 24 h (n = 3 independent experiments). The effects of LPS stimulation were also determined in these experiments (using 100 ng/ml for 8 h). Microglia, treated with LPS, produced high concentrations of IL-6 to levels exceeding 400 pg/ml (n = 4 independent experiments).
Discussion
The results from this work indicate that PAF-mediated changes in [Ca2+]i are involved in the cellular expression of the pro-inflammatory agent, IL-6 in human microglia. In essence, activation of SOC acts as a transcriptional control for expression of IL-6. Our results show that inhibition of SOC with SKF96365 blocked both the influx of Ca2+ and microglial expression of IL-6. However, PAF-induced expression of IL-6 (Fig 2) did not translate into production of the cytokine (Fig 3). This result could suggest that an additional signal or factor may be required for microglial secretion of IL-6.
As found for other types of unexcitable cells, microglia do not normally express voltage-dependent Ca2+ channels [7]. The sustained entry of Ca2+ through SOC is likely an important pathway for microglial responses to specific inflammatory stimuli [15,22,26]. Although opening of SOC is required for re-filling of ER stores, other roles for this influx pathway have not been well established. Activation of SOC is necessary for expression of IL-6 but an additional signal is required to produce the pro-inflammatory cytokine in human microglia. The activation state of human microglia may influence the extent of Ca2+ influx through SOC. Microglia showing an ameboid morphology are considered representative of an activated state whereas cells with a ramified morphology are considered quiescent. We have found sustained SOC responses from PAF-stimulated microglia in cells demonstrating ameboid morphology [15,21] and also in the present work. However, an initial study using a mixture of ameboid and ramified shaped cells, showed a transient SOC response with stimulation by PAF [31]. Further work will be useful to correlate expression of SOC with cell activation.
A recent review has provided a detailed overview of ATP as an inducer of IL-6 expression and production in MG-5 microglial cell line [12]. ATP and the purinergic agonist BzATP were both effective in increasing expression of IL-6 with effects involving activation of the p38 MAPK pathway. However, ATP (activator of both metabotropic P2YR and ionotropic P2XR) but not BzATP (activator of the ionotropic subtype P2X7R), was found to induce production of the cytokine. The role of SOC in MG-5 cell responses is unclear since ATP evokes a monophasic change in [Ca2+]i due to P2YR dependent release from intracellular stores. In human microglia we have attributed the lack of a SOC phase of [Ca2+]i due to concomitant ATP binding to some P2XR (not P2X7R) causing cellular depolarization and block of Ca2+ influx [6].
PAF induction of IL-6 was found to be time-dependent (Fig 2A) in addition to the dependence on the presence of extracellular Ca2+ and SOC (Fig 2B). We observed no IL-6 expression at one-half hour and a maximal level at one hour of microglial exposure to PAF. Little or no IL-6 was expressed with longer PAF treatments of microglia. Inhibition of endoplasmic reticulum Ca2+ ATPase (SERCA) has been reported to increase IL-6 mRNA expression in rodent macrophages within 15 min [4,19]. Blockade of SERCA, by compounds such as thapsigargin, and subsequent depletion of intracellular stores is a stimulatory protocol for activation of SOC. However, SOC-mediated entry of Ca2+ was not determined in the rodent studies.
Although PAF was an effective stimulator of IL-6 expression in human microglia, LPS elicited a higher expression of the cytokine. Indeed, bands for IL-6 appeared saturated (Fig 2B) and showed no change in intensity with increased number of PCR cycles (data not shown). Saturation with LPS would prevent a quantitative comparison between PAF and LPS as activating stimuli for microglial expression of IL-6 (Fig 2C). An interesting observation was that SKF96365 partially inhibited the LPS-induced expression of IL-6 (Fig 2C). Although LPS has been reported to act in a Ca2+-independent manner on macrophages [19], several studies have found the bacterial compound evokes changes in [Ca2+]i in microglia/macrophages [1,5,16,32] suggesting possible involvement of SOC in LPS induction of cytokines.
The present results may have relevance to roles of IL-6 in aging. Several studies have provided evidence for age-dependent increases in levels of IL-6 in rodent brain [reviewed in [10]]. For example, one finding was that brains from older mice showed considerable elevations in expression and production of IL-6 compared with brains from younger animals [33]. This result was correlated with microglial production of the cytokine [33]. It will be of interest to determine if PAF-stimulated adult human microglia are more potent producers of IL-6 compared with fetal human cells.
List of abbreviations
PAF: platelet-activating factor; SOC: store-operated channels; IL-6; interleukin-6; PSS: physiological saline solution; PBS: phosphate-buffered saline; [Ca2+]i: intracellular calcium; DMEM: Dulbecco's modified Eagle's medium
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PS and HBC contributed equally to calcium imaging, RT-PCR and ELISA experiments. HBC also carried out isolation of microglia. SC participated in the design of experiments and reviewed and edited the manuscript. JGM designed and supervised all experiments, interpreted the data and finalized the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from the Heart and Stroke Foundation of British Columbia and Yukon and Alzheimer's Society of Canada (to JGM) and a doctoral research award from the Heart and Stroke Foundation of Canada (to HBC).
==== Refs
Bader MF Taupenot L Ulrich G Aunis D Ciesielski-Treska J Bacterial endotoxin induces [Ca2+]i transients and changes the organization of actin in microglia Glia 1994 11 336 344 7525480
Bielenberg GW Wagener G Beck T Infarct reduction by the platelet activating factor antagonist apafant in rats Stroke 1992 23 98 103 1731428
Blum-Degen D Muller T Kuhn W Gerlach M Przuntek H Riederer P Interleukin-1 beta and interleukin-6 are elevated in the cerebrospinal fluid of Alzheimer's and de novo Parkinson's disease patients Neurosci Lett 1995 202 17 20 8787820 10.1016/0304-3940(95)12192-7
Bost KL Mason MJ Thapsigargin and cyclopiazonic acid initiate rapid and dramatic increases of IL-6 mRNA expression and IL-6 secretion in murine peritoneal macrophages J Immunol 1995 155 285 296 7602106
Choi HB Khoo C Ryu JK van Breemen E Kim SU McLarnon JG Inhibition of lipopolysaccharide-induced cyclooxygenase-2, tumor necrosis factor-alpha and [Ca2+]i responses in human microglia by the peripheral benzodiazepine receptor ligand PK11195 J Neurochem 2002 83 546 555 12390516 10.1046/j.1471-4159.2002.01122.x
Choi HB Hong SH Ryu JK Kim SU McLarnon JG Differential activation of subtype purinergic receptors modulates Ca2+ mobilization and COX-2 in human microglia Glia 2003 43 95 103 12838502 10.1002/glia.10239
Eder C Ion channels in microglia (brain macrophages) Am J Physiol 1998 275 C327 C342 9688586
Ferrarese C Mascarucci P Zoia C Cavarretta R Frigo M Begni B Sarinella F Frattola L De Simoni MG Increased cytokine release from peripheral blood cells after acute stroke J Cereb Blood Flow Metab 1999 19 1004 1009 10478652 10.1097/00004647-199909000-00008
Gao MH Jiang J Wilson B Zhang W Hong JS Liu B Microglial activation-mediated delayed and progressive degeneration of rat nigral dopaminergic neurons: relevance to Parkinson's disease J Neurochem 2002 81 1285 1297 12068076 10.1046/j.1471-4159.2002.00928.x
Godbout JP Johnson RW Interleukin-6 in the aging brain J Neuroimmunol 2004 147 141 144 14741447 10.1016/j.jneuroim.2003.10.031
Grynkiewicz G Poenie M Ysien R A new generation of Ca2+ indicatiors with greatly improved fluorescence properties J Biol Chem 1985 260 3440 3450 3838314
Inoue K Microglial activation by purines and pyrimidines Glia 2002 40 156 163 12379903 10.1002/glia.10150
Jaranowska A Bussolino F Sogos V Arese M Lauro GM Gremo F Platelet-activating factor production by human fetal microglia. Effect of lipopolysaccharides and tumor necrosis factor-alpha Mol Chem Neuropathol 1995 24 95 106 7632332
Jones S Horiuchi S Topley N Yamamoto N Fuller G The soluble interleukin 6 receptor: mechanisms of production and implications in disease FASEB J 2001 15 43 58 11149892 10.1096/fj.99-1003rev
Khoo C Helm J Choi HB Kim SU McLarnon JG Inhibition of store-operated Ca2+ influx by acidic extracellular pH in cultured human microglia Glia 2001 36 22 30 11571781 10.1002/glia.1092
Letari O Nicosia S Chiavaroli C Vacher P Schlegel W Activation by bacterial lipopolysaccharide causes changes in the cytosolic free calcium concentration in single peritoneal macrophages J Immunol 1991 147 980 983 1907309
Lindsberg PJ Yue TL Frerichs KU Hallenbeck JM Feuerstein G Evidence for platelet-activating factor as a novel mediator in experimental stroke in rabbits Stroke 1990 21 1452 1457 2219210
Lue LF Rydel R Brigham EF Yang LB Hampel H Murphy GM JrBrachova L Yan SD Walker DG Shen Y Rogers J Inflammatory repertoire of Alzheimer's disease and nondemented elderly microglia in vitro Glia 2001 35 72 79 11424194 10.1002/glia.1072
Marriott I Bost KL Mason MJ Differential kinetics for induction of IL-6 mRNA expression in murine peritoneal macrophages: Evidence for calcium-dependent and independent-signalling pathways J Cell Physiol 1998 177 232 240 9766520 10.1002/(SICI)1097-4652(199811)177:2<232::AID-JCP5>3.0.CO;2-O
McGeer PL McGeer EG The inflammatory response system of brain: implications for therapy of Alzheimer and other neurodegenerative diseases Brain Res Rev 1995 21 195 218 8866675 10.1016/0165-0173(95)00011-9
McLarnon JG Helm J Goghari V Franciosi S Choi HB Nagai A Kim SU Anion channels modulate store-operated calcium influx in human microglia Cell Calcium 2000 28 261 268 11032781 10.1054/ceca.2000.0150
Moller T Calcium signaling in microglial cells Glia 2002 40 184 194 12379906 10.1002/glia.10152
Mori M Aihara M Kume K Hamanoue M Kohsaka S Shimizu T Predominant expression of platelet-activating factor receptor in the rat brain microglia J Neurosci 1996 16 3590 3600 8642404
Nagai A Nakagawa E Hatori K Choi HB McLarnon JG Lee MA Kim SU Generation and characterization of immortalized human microglial cell lines: expression of cytokines and chemokines Neurobiol Dis 2001 8 1057 1068 11741401 10.1006/nbdi.2001.0437
Neuroinflammation Working Group Inflammation and Alzheimer's disease Neurobiol Aging 2000 21 383 421 10858586 10.1016/S0197-4580(00)00124-X
Parekh AB Penner R Store depletion and calcium influx Physiol Rev 1997 77 901 930 9354808
Pavelko KD Howe CL Drescher KM Gamez JD Johnson AJ Wei T Ransohoff RM Rodriguez M Interleukin-6 protects anterior horn neurons from lethal virus-induced injury J Neurosci 2003 23 481 492 12533608
Tarkowski E Rosengren L Blomstrand C Wikkelso C Jensen C Ekholm S Tarkowski A Early intrathecal production of interleukin-6 predicts the size of brain lesion in stroke Stroke 1995 26 1393 1398 7631343
Toescu EC Moller T Kettenmann H Verkhratsky A Long-term activation of capacitative Ca2+ entry in mouse microglial cells Neuroscience 1998 86 925 935 9692728 10.1016/S0306-4522(98)00123-7
Walker DG Kim SU McGeer P Complement and cytokine gene expression in cultured microglia derived from postmortem human brains J Neurosci Res 1995 40 478 493 7616608 10.1002/jnr.490400407
Wang X Bae JH Kim SU McLarnon JG Platelet-activating factor induced Ca2+ signaling in human microglia Brain Res 1999 842 159 165 10526106 10.1016/S0006-8993(99)01849-1
Xie YC Schafer R Barnett JB Inhibitory effect of 3,4-dichloro-propionaniline on cytokine production by macrophages is associated with LPS-mediated signal transduction J Leukoc Biol 1997 61 745 752 9201266
Ye SM Johnson RW Increased interleukin-6 expression by microglia from brain of aged mice J Neuroimmunol 1999 93 139 148 10378877 10.1016/S0165-5728(98)00217-3
| 15833109 | PMC1097754 | CC BY | 2021-01-04 16:38:22 | no | J Neuroinflammation. 2005 Apr 15; 2:11 | utf-8 | J Neuroinflammation | 2,005 | 10.1186/1742-2094-2-11 | oa_comm |
==== Front
Kinetoplastid Biol DisKinetoplastid Biology and Disease1475-9292BioMed Central London 1475-9292-4-31586212810.1186/1475-9292-4-3Original ResearchStage-specific expression of the mitochondrial co-chaperonin of Leishmania donovani, CPN10 Zamora-Veyl Fanny Beatriz [email protected] Manfred [email protected] Dorothea [email protected] Joachim [email protected] Bernhard Nocht Institute for Tropical Medicine, Bernhard Nocht St. 74, D-20359 Hamburg, Germany2005 29 4 2005 4 3 3 27 1 2005 29 4 2005 Copyright © 2005 Zamora-Veyl et al; licensee BioMed Central Ltd.2005Zamora-Veyl 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
Leishmania spp., in the course of their parasitic life cycle, encounter two vastly different environments: the gut of sandflies and the phagosomes of mammalian macrophages. During transmission into a mammal, the parasites are exposed to increased ambient temperature as well as to different carbon sources. Molecular chaperones or heat shock proteins are implicated in the necessary adaptations which involve the ordered differentiation from the flagellated, extracellular promastigote to the intracellular amastigote stage.
Results
Here, we show that the Leishmania donovani co-chaperonin, CPN10, is synthesised to a significantly increased concentration during in vitro differentiation to the amastigote stage. We show by fluorescence microscopy and by immunogold electron microscopy that, like its putative complex partner CPN60.2, CPN10 is localised to the single, tubular mitochondrion of the parasites and, moreover, that it co-precipitates with CPN60.2, the major mitochondrial chaperonin of Leishmania spp..
Conclusion
Our data indicate an increased requirement for CPN10 in the context of mitochondrial protein folding during or early in the mammalian stage of this pathogen. Moreover, they confirm the CPN60.2 as bona fide mitochondrial GroEL homologue in L. donovani and the postulated interaction of eukaryotic chaperonins, CPN60 and CPN10.
==== Body
Background
The euglenozoa which, beside the euglenophyta, also include the pathogenic protozoa of the Order Kinetoplastida, have branched off early in eukaryotic evolution. The kinetoplastida have developed, or retained, molecular mechanisms not shared by most Crown Group Eukaryota, such as trans-splicing of messenger RNA, RNA editing, and polycistronic transcription. The genetic information of their single mitochondrion, the so-called kDNA, is localised to the kinetoplast, an organelle unique to this order. The kinetoplast is part of a tubular, single mitochondrion that shows little morphological likeness to the mitochondria of Crown Group Eukaryota. The kinetoplastida encompass two genera of parasitic pathogens, Trypanosoma and Leishmania, both of which responsible for considerable human morbidity and mortality.
The parasites of the genus Leishmania exist in two morphologically distinct life cycle stages. The slender, flagellated, promastigote proliferates in the alimentary tract of female sandflies at ambient temperature. The amastigote, a rounded, intracellular form, with no protruding flagellum, persists in the phagosomes of mammalian macrophages and at temperatures of up to 39°C. The difference of temperatures in the two environments not only triggers the respective stage development [1], but also poses a challenge to the system of protein chaperones. At least two heat shock proteins, Hsp100 and CPN60.2, show significantly induced abundance during the amastigote stage [2-4].
Apart from the temperature difference, the two environments, insect gut and macrophage phagosome, also offer different qualitative and quantitative composition of carbon sources. The life cycle stage differentiation, therefore, must include an adaptation of the parasites' metabolic pathways. Subcellular compartments, e.g. the single tubular mitochondrion, also have to undergo adaptation. The folding of a different set of metabolic enzymes after their translocation across the outer and inner mitochondrial membranes can be expected to challenge the mitochondrial chaperone system.
Import of nuclear encoded proteins is vital for the function of mitochondria. A complex machinery of chaperone proteins is required for the unfolding of mitochondrial proteins, their trans-membrane transport, and their refolding inside the mitochondrial matrix [5]. The latter is thought to be brought about by the 60 kDa chaperonin (CPN60). This protein is a functional and structural homologue to the GroEL chaperonin of the eubacteriae. The function of GroEL, being the subject of considerable research efforts, is well understood [6,7]. In bacteria, GroEL forms barrel-like structures of two rings of 7 subunits each, with two central cavities. 7 subunits of the GroES co-chaperonin form a lid that covers one of the cavities. Folding intermediates of client proteins are bound inside one cavity where they may attain energetically favoured conformations without interference from other polypeptides.
Less is known, by comparison, about the function of eukaryotic GroEL homologues, the CPN60 proteins. They are nuclear-encoded mitochondrial proteins. Rather than forming two rings of seven GroEL subunits, the chaperonin complex in mitochondria consists of a single ring of seven CPN60 and one ring of CPN10 subunits, and it was shown that the function of CPN60/CPN10 complexes inside the mitochondria is similar to the chaperonin function of bacterial, cytosolic GroEL/GroES complexes [8,9]. GroES may also play a role in modulating the substrate specificity of GroEL. For instance it was reported that GroES binding shifts the preference of GroEL for hydrophobic amino acids towards amino acids with hydrophilic side chains [10]. In yeast, not all CPN60 substrates need CPN10 for proper folding. Other proteins require both CPN60 and CPN10, while newly imported CPN60 depends on CPN10 for correct folding [11].
We had previously identified the Leishmania donovani CPN60.2 gene which encodes the mitochondrial 60 kD chaperonin [4]. This protein shows increased expression during in vitro promastigote to amastigote differentiation, indicating an altered requirement for CPN60 in the amastigote stage. A second, diverged gene copy, CPN60.1, is not expressed to detectable levels in any in vitro culture stage of L. donovani or L. major [4]. To better understand mitochondrial protein chaperoning during the leishmanial life cycle, we identified the CPN10 gene of Leishmania donovani, determined its abundance in the two life cycle stages, and analysed the subcellular localisation of the gene product and its association with CPN60.2.
Results
Amplification of L. donovani CPN10 DNA
An alignment of the known eukaryotic CPN10 amino acid sequences revealed four limited stretches of sequence conservation (not shown). Oligonucleotides CPN10.1 to CPN10.4 were designed according to the codon bias of Leishmania spp.. Using L. donovani genomic DNA as template, combinations of the aforementioned primers were tested in a polymerase chain reaction. Only the combination of primers CPN10.1 and CPN10.4 yielded an amplification product of the expected size (not shown). The amplification product was then subcloned into the vector pBluescript and subjected to sequence analysis using standard pBluescript sequencing primers. The sequence analysis confirmed that the amplification product harboured probable CPN10 sequences (not shown).
Isolation of a CPN10 genomic DNA clone
The amplified putative CPN10 gene fragment was used to screen a L. donovani genomic DNA cosmid library [3]. Positive cosmids were subjected to a restriction digest and Southern Blot analysis to identify common subfragments which harboured the CPN10 gene(s). A 1.5 kb XhoI I fragment which hybridised with the CPN10 probe was subcloned into pBluescript and subjected to sequence analysis. The sequence (GenBank AF394959) contains an uninterrupted open reading frame of 303 bp encoding a putative 10.7 kDa polypeptide. A sequence comparison by BLAST search confirmed this putative polypeptide as related to eukaryotic and prokaryotic co-chaperonins. Figure 1A shows the position of the L. donovani CPN10 in the lineage of eukaryotic CPN10 proteins. L. donovani CPN10 groups within the kinetoplastid homologs. Figure 1B shows an alignment with selected eukaryotic CPN10 polypeptides. The high degree of sequence conservation between LdCPN10 and its counterparts in higher eukaryotes indicates a conserved function.
Figure 1 (A) Phylogenetic analysis (clustalW) of eukaryotic CPN10 family members. The bar shows the scale for sequence divergence. (B) Amino acid sequence alignment of the LdCPN10 and selected eukaryotic CPN10 family members. The alignment was perfomed using the clustalW algorithm.
Southern Analysis
Within the haploid set of chromosomes, L. donovani CPN10 appears to be single copy. This was determined by Southern Blot analysis of L. donovani genomic DNA, digested with 7 restriction endonucleases (Figure 2A). The pattern of restriction fragments does not indicate the presence of additional gene copies. We also performed pulsed field gel electrophoresis (Figure 2B) to determine the chromosomal localisation of the CPN10 gene. The CPN10 specific probe hybridised to a chromosomal band of approximately 1,100 kb. This size corresponds to chromosomes 26 or 27b. This was confirmed when sequencing of the L. major genome was finished; two copies of CPN10 are found on chromosome 26 .
Figure 2 (A) Southern analysis of L. donovani genomic DNA. The DNA (2 μg) was digested with restriction enzymes as indicated and separated on a 1 % agarose gel. After Southern transfer the membrane was probed with a CPN10 specific probe. The positions of size markers (1 kb ladder, Fermentas) are indicated. (B) Pulsed field gel electrophoresis of L. donovani chromosomes. The gel was stained with ethidium bromide and photographed (right). After denaturation the DNA was blotted on a Nylon membrane and probed with a CPN10 specific probe (left). The sizes of S. cerevisiae chromosomes which were run alongside as size markers are shown to the right. The chromosome band which hybridised with the CPN10 probe is marked by an arrow. (C) RT-PCR of L. donovani RNA. Whole cell RNA was isolated from L. donovani cells and subjected to reverse transcription at 37°C for 60 min using an oligo-dT primer. An aliquot of the cDNA reaction mix was inactivated at 65°C for 10 min prior to incubation at 37°C. Either RNA (lane 1), the cDNA (lane 2), the heat inactivated cDNA mix (lane 3) or genomic DNA (lane 4) were added to a PCR reaction mix using CPN10 specific primers. The sizes of selected marker DNA molecules is shown to the right.
RT-PCR analysis
Northern analysis of L. donovani RNA using a CPN10-specific probe did not yield any bands in the expected size range. To ascertain that the putative CPN10 gene encodes a stable RNA, we performed RT-PCR. Whole cell RNA was reversely transcribed using an oligo-dT primer. The first strand cDNA was used as template in a PCR amplification using primers CPN10-1 and CPN10-4. Negative controls included an amplification of material from a cDNA synthesis mix that had been inactivated at 65°C for 10 min (Figure 2C, lane 1), and an amplification with an equivalent amount of whole cell RNA as template (lane 3), to exclude DNA contaminations in the RNA preparation. Both reactions did not give rise to an amplification product. The reaction with cDNA as template (lane 2), however, produced a ~350 bp PCR product, much the same as the reaction with genomic DNA as template (lane 4). We conclude that stable, polyadenylated RNA is produced from the CPN10 gene.
Production of anti-CPN10 antibodies
The CPN10 open reading frame was subcloned into pJC45, a derivative of pJC40 [4,12], for expression in E. coli strain Bl21 (DE3) [13]. His-tagged CPN10 was purified by metal chelate chromatography (Figure 3) and used to immunise laying hens. Antibodies were prepared from egg yolk before and after immunisation to yield pre-immune and CPN10-specific antibody preparations. In Western Blot analyses, these antibodies faithfully recognised recombinant rCPN10 (not shown).
Figure 3 Expression of L. donovani CPN10 in E. coli. The plasmids pJC45 or pCPN10-45 was used to transform E. coli strain BL21 (DE3) [pAPlacIQ]. Expression was induced by addition of IPTG, and the bacteria were harvested. After lysis, aliquots of the cultures before and after IPTG induction were analysed by SDS-PAGE and Coomassie Blue staining. The lysate containing the rCPN10 was subjected to metal chelate chromatography to purify the rCPN10 by virtue of its histidine tag. The purified rCPN10 was also subjected to SDS-PAGE analysis.
LdCPN10 expression is induced in the amastigote
Using the anti-CPN10 antibodies, we performed immunoblot analysis on lysates of L. donovani promastigotes before and after heat shock, and of axenically cultured amastigotes (Figure 4B). In promastigotes cultured at 25°C, we observe only a faint band in the expected size range. After a 24 h treatment of the promastigotes at 37°C, a signal is observed. The CPN10 signal increases strongly during in vitro amastigote differentiation. Figure 4A shows identical samples separated by SDS-PAGE and Coomassie Blue staining. We conclude that Leishmania donovani CPN10 is expressed preferentially in the amastigote stage, induced at least in part by the elevated temperature.
Figure 4 (A) SDS-PAGE of various in vitro culture forms of L. donovani. Equivalent protein aliquots from L. donovani promastigotes, cultivated at either 25°C (promastigote -25°), at 37°C (promastigote -37°), and from axenic amastigotes after 3 days of in vitro differentiaton were subjected to Tricine Gel Electrophoresis and to SDS-PAGE. Protein bands were visualised by Coomassie Brilliant Blue staining. (B) Immunoblot analysis. A parallel gel was subjected to Western transfer and probed with an anti-CPN10 antibody. (C) Co-immune precipitation. Leishmania donovani lysate was precipitated using anti-CPN10 antibodies, separated by SDS-PAGE and subjected to WesternBlot analysis with anti-CPN60.2 antibodies. Lanes 1 and 2 are negative controls representing immune precipitations without anti-CPN10 antibody and with a preimmune antibody preparation, respectively. Lane 3 represents the CPN10 immune precipitation proper. The position of CPN60.2 is indicated.
CPN10 is the co-chaperonin in L. donovani
By analogy, CPN10 is expected to act as co-chaperonin to the 60 kDa chaperonin, CPN60. However, there are two diverged members of the CPN60 gene family in L. donovani, CPN60.1 and CPN60.2. Of these, only CPN60.2 could be shown to encode a protein which localised to the mitochondrion [4]. We therefore tested whether CPN10 interacts with CPN60.2 protein in L. donovani and performed co-immune precipitation. L. donovani cell lysates were incubated with anti-CPN10 antibodies, anti-chicken IgG (rabbit) and Protein-A agarose. The precipitated material was then analysed by immunoblot using anti-CPN60.2 antibody. As shown in Figure 4C, the anti-CPN10 precipitate contains a protein which reacts with the anti-CPN60.2 antibodies. We conclude that CPN60.2 is the bona fide chaperonin and that CPN10 serves as its co-chaperonin.
Subcellular localisation of LdCPN10
Since CPN10 serves as co-chaperone to CPN60 we expected CPN10 to localise to the kinetoplast/mitochondrion complex inside the leishmaniae. To test this prediction, the CPN10 coding sequence was fused to a GFP (Green Fluorescent Protein) coding DNA in the expression vector pIRmcs3-. The linearised expression construct was recombined into the rRNA locus of L. donovani to yield stable expression of the CPN10::GFP chimera. The recombinant parasites were cultivated as promastigotes and analysed by fluorescence microscopy. The results are shown in Figure 5. We observe brightly fluorescent tubular structures in the promastigotes. No fluorescence is observed in the cytoplasm. This indicates that the CPN10 sequence can quantitatively direct GFP into a subcellular compartment or organelle.
Figure 5 Fluorescence microscopy. L. donovani promastigotes transfected with a CPN10/GFP gene chimera on plasmid pIRCPN10::GFP were subjected to bright field microscopy at 63× magnification (panels A-C). The same microscopic fields were viewed under UV excitation at XXX nm (panels D-F). Panels G-I show the overlays.
To identify the subcellular compartment that harbours CPN10, we performed immune electron microscopy using sections of heat stressed promastigotes and of axenic amastigotes. Figure 6 shows the result. We find the gold particles associated both with tubular structures and with the kinetoplast. We conclude that CPN10 is a mitochondrial protein.
Figure 6 Immunogold electron microscopy. Heat-shocked promastigotes (a,d) or axenic amastigotes (b,c) were embedded in LR-White. Microsections were stained with pre-immune antibodies (a) or anti-CPN10 (b-c) antibodies. Antibody binding was detected using rabbit anti-chicken IgG and Protein A Gold (10 nm). Annotations: n = nucleus; k = kinetoplast; m = tubular mitochondrium.
When we analysed the subcellular localisation of CPN60.2 in L. donovani, this chaperonin, too, localised to a tubular subcellular compartment. However, we did not find it associated with the kinetoplast. Therefore, the exact subcellular localisation was still unclear. Using the strain which overexpressed the CPN10::GFP chimera, we performed a co-immune electron microscopy. We used anti-GFP mAB in conjunction with anti-mouse immunogold particles (5 nm). Then the sections were co-stained using anti-CPN60 antibody, anti-chicken IgG (rabbit), and protein A gold (10 nm). As shown in Figure 7, CPN10::GFP and CPN60 colocalise to the same tubular compartment. This shows that CPN60.2 is indeed a mitochondrial protein in Leishmania.
Figure 7 Double immune electron microscopy. Leishmania cells that express CPN10::GFP chimera were embedded and microsections were prepared. After staining with anti-GFP Mab and anti-mouse immunogold (5 nm), sections were treated with anti-CPN60 antibodies, rabbit anti-chicken IgG, and Protein A gold (10 nm). Thus, 10 nm gold particles represent CPN60 and 5 nm gold particles represent CPN10::GFP. k = kinetoplast.
Discussion
We have cloned and analysed the CPN10 gene of the protozoan parasite Leishmania donovani. This is, to our knowledge, the first characterisation of a protozoan member of the conserved GroES family of co-chaperonins. Since CPN10 is a very small protein with an impact on mitochondrial function, the restraints on sequence variation should be rather tight. Accordingly, Leishmania CPN10 is >60 % similar to mammalian CPN10 sequences. A comparison of the L. donovani CPN10 gene with other eukaryotic GroES homologs further confirms that the leishmaniae belong to a remote entity among the eukaryota, clearly separated from the Eukaryotic Crown Group. A similarly diverged position was observed for the GroEL homologue of L. donovani, CPN60.2 [4].
The gene for CPN10 is single copy in the L. donovani genome. This is in contrast to the CPN60.2 gene which is present in two copies, in addition to a, as far as known silent, CPN60.1 gene [4]. It is also in contrast with the situation in L. major. According to the L. major genome database, the L. major CPN10 is encoded by two gene copies located in tandem. If confirmed, this difference could serve to distinguish both Leishmania species which have overlapping endemic regions.
Eukaryotic CPN10, like its bacterial counterpart, GroES, is assumed to act as co-chaperonin to CPN60 in the mitochondrial matrix. Immune electron microscopy reveals that CPN10 localises both to the kinetoplast and to the tubular part of mitochondrion. More importantly, co-immunoprecipitation experiments establish a direct interaction between CPN10 and CPN60.2. This not only indicates a function of CPN10 as bona fide co-chaperonin, but also establishes CPN60.2 as the Leishmania GroEL homolog. In addition, the mitochondrial localisation of both proteins is now well established.
The mitochondrial transport signal sequence of CPN10 can direct the import of the much larger CPN10::GFP chimera into the mitochondria and the kinetoplast. The chimera is expressed stably from the integration site within the rRNA gene repeats. The chimera may serve for future experimental designs as mitochondrial localiasation marker and as marker for the integrity of the mitochondrion.
The CPN10 gene is expressed preferentially under heat stress and in axenic amastigotes of L. donovani. The putative complex partner, CPN60.2, also shows increased abundance under heat stress [4]. The induction of CPN10, however, is more pronounced, and comparable to the induction of Hsp100 [3]. In this regard, three heat shock proteins, Hsp100, Cpn60.2, and CPN10, differ from the major chaperones, Hsp70 and Hsp90 (Hsp83). The latter show high constitutive abundance and only a marginal increase during heat stress and in the amastigote stage [3,14]. It is likely that these different expression kinetics reflect different roles during the life cycle. Hsp100 is required only in intracellular amastigotes of L. donovani and has a role in amastigote specific gene expression [15]. CPN60.2, and even more so, CPN10, are obviously also in increased demand during the amastigote stage.
Is there a specific function for CPN10 during the mammalian stage of the parasite's life cycle? It has been suggested that GroES can alter the affinity of GroEL for amino acid side chains [10]. In the yeast, substrate proteins may fold either independently of CPN10 or in a CPN10-dependent manner [11]. This raises the possibility that the preferential expression of CPN10 in amastigotes of L. donovani may reflect a change in the substrate specificity of the mitochondrial chaperonin complex, necessitated by the change of environment and the concomitant changes in nutrient availability. Since CPN10 is a single copy gene that is hardly expressed in the highly proliferative promastigote stage, it may be a suitable target for a gene replacement approach. This should allow to assess the role of CPN10 for the virulence and pathogenicity of Leishmania parasites.
Conclusion
We have identified the Leishmania donovani homologue of the bacterial co-chaperonin, GroES, and assigned the name CPN10 to the gene and the protein. CPN10 is a heat shock protein and shows increased intracellular concentration at an elevated temperature and in axenically cultivated amastigotes of L. donovani. CPN10 localises to the mitochodrial compartment and co-localises with the GroES homologue, CPN60.2. A direct interaction of both proteins was demonstrated using co-immuno precipitation. Both CPN10 and CPN60.2 are therefore suitable markers for the mitochondria of Leishmania spp. and related species. Moreover, the increased concentration in the amastigote stage suggests a role of CPN10 in this stage.
Methods
Parasite strains and culture
For our analyses we used the Lo8 strain of L. donovani, a gift from D. Zilberstein. Promastigotes were routinely cultivated at 25°C in supplemented M199 medium [3]. Cell density was monitored using a Schaerfe Systems CASY Cell Counter. In vitro promastigote to amastigote differentiation was performed as described [3].
PCR amplification of CPN10 DNA
Four primers were delineated from an amino acid sequence comparison of known GroES and CPN10 proteins. The extreme bias in Leishmania towards G and C residues in the wobble positions of codons was used to create primers with minimal degeneracy:
primer CPN10.1: CCGCTGTTCG ACCGCGTGCT GG
primer CPN10.2: GGCGGCATCR TGCTGCC
primer CPN10.3: CGGCAGCAYG ATGCCGCC
primer CPN10.4: CAGCACCTTG TCGCCCACCT TCAC
100 ng of L. donovani genomic DNA was mixed with 40 ng of oligonucleotide primer pairs, 25 nmoles of each dNTP, 10 × reaction buffer, and 1 unit of Taq DNA polymerase (Beckman) in a 50 μl reaction. The reaction mix was incubated for 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C. The cycle was repeated 35 times. 10% of the reaction was analysed on a 1% agarose gel. Negative controls included each primer by itself and omission of template DNA. Bands absent from the negative controls were recovered by preparative agarose gel electrophoresis and by using glass beads adsorption (PureGen Kit, Biozyme). PCR products were subcloned using the TA cloning kit (Invitrogen) and subjected to sequence analysis..
Reverse transcriptase (RT-) PCR
Total RNA from promastigotes stage parasites strain Lo8 was prepared using a Ribolyser Kit (Hybaid). cDNA synthesis was carried out using the First Strand synthesis kit (Pharmacia) with the enclosed (dT)18 primer and 5 μg of total RNA. The cDNA was then amplified enzymatically using the primers CPN10.1 and CPN10.4 The amplification was carried out as described above.
Library screening
We used a cosmid library of L. donovani previously described in [16]. The library was screened by hybridisation with digoxigenin-labeled DNA probes derived from the subcloned amplification products. Hybridisation was performed at 65°C in HYB 9 solution (Biozyme) for 16 h. Washes were performed at decreasing SSC concentrations (6 × to 0.2 × SSC, 0.5 % SDS) at 65°C. Filters were developed using anti-digoxigenin FAB/AP conjugate (Boehringer Mannheim) followed by colorimetric staining with NBT/BCIP. Positive cosmid clones were picked, rescreened, and verified in an amplification reaction using CPN10-specific primers.
Subcloning and sequence analysis
Cosmid DNA from positive clones was subjected to restriction endonucleased digest and Southern Blot to determine suitable restriction endonucleases. Cosmid DNA was then digested at a preparative scale and subcloned into an appropriately cleaved pBluescript KS+ vector (Stratagene) or pJC45 vector [4]. Plasmid subclones were then subjected to bidirectional primer walking sequence analysis, starting from the known sequences of the PCR amplificates. Sequencing was performed on an Applied Biosystems Model 370 sequencer using the DyeTerminator Kit.
Sequence evaluation
Contig alignment, in silico translation, and sequence alignment were performed using the MacMolly Tetris and the MacVector software packages. Phylogenetic analyses were performed using the clustalW algorithm included with MacVector at default settings.
Recombinant Expression of CPN10
The putative open reading frame of CPN10 was amplified using specific primers that modify the start codon sequence into a NdeI site and the stop codon sequence into an EcoRI site. The amplification products were cleaved with NdeI and EcoRI and ligated into the expression vector pJC45 [4], between the NdeI site and the EcoRI site. The expression plasmids were transformed into the bacterial strain BL21 (DE3) [pAPlacIQ], a gift from Olivier Payet, CNRS Toulouse. Bacteria were grown in CircleGrow medium (BIO 101) supplemented with 50 μg/ml Ampicillin and 10 μg/ml Kanamycin to OD600 = 0.5. IPTG was added to 1 mM and incubation was continued for 1 h. The purification of His-tagged proteins from bacterial lysates by metal chelate chromatography has been described [12,14].
To express CPN10::GFP chimera, the CPN10 coding sequence, minus the stop codon, was enzymatically amplified to create BamHI and HindIII sites at the 5' and 3' ends, respectively. Digested with BamHI and HindIII, the amplification products were ligated between the BamHI and HindIII sites of the yeast plasmid p426/PQ25 [17]. By this procedure, CPN10 and GFP coding sequences were fused in frame. The chimeric CPN10::GFP coding sequence was then excised with BamHI and XhoI and ligated into pIRmcs3- [18]. The plasmid pIRCPN10::GFP and the parent plasmid pIRmcs3- were linearised using SwaI and transfected into L. donovani by electroporation. Recombinant parasites were placed under ClonNAT (Werner Bioreagents) selection (100 μg/ml) and analysed by fluorescence microscopy.
Immunisation and antibody preparation
The immunisation of laying hens and the preparation of antibodies from egg yolk has been described [19].
Immunoblot analysis
SDS-PAGE and Western Transfer were performed as described [14,19]. Briefly, membranes were treated with blocking buffer (4% skim milk powder and 0.1% Tween 20 in DPBS), with antibodies at appropriate dilutions in blocking buffer, and with secondary antibody/alkaline phosphatase conjugate (Dianova) in blocking buffer. Blots were stained with nitrobluetetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate.
Immune precipitation
1 × 108 promastigotes were harvested by centrifugation and lysed in 500 μl solubilising buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, and 2.5 μg × ml-1 each of aprotinin, leupeptin, pepstatin, and antipain). Lysates were centrifuged at 13000 × g, 4°C. Soluble proteins were incubated with 20 μl anti-CPN10 antibody for 1 h at 4°C. 350 μl of protein A-sepharose slurry were preincubated with anti-chicken IgG from rabbit and added. The slurry was further incubated at 4°C for 2 h. The immunoabsorbent was centrifuged and washed three times in washing buffer A (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0,2% NP-40, 2 mM EDTA), twice in washing buffer B (as above, except NaCl is 500 mM) and once in washing buffer C (10 mM Tris-HCl pH 7.5). For analysis on SDS-PAGE 250 μl of SDS sample buffer was added and the samples were heated for 5 minutes at 95°C. After centrifugation, 50 μl of the supernatant was loaded onto a 12% PA gel.
Immune electron microscopy
Immune electron microscopy was performed essentially as described [3]. LR-White embedded microsections were treated with chicken antibody at 1:500 (anti-CPN10) or 1:2000 (anti-CPN60.2) in PBS (0.1 % Tween 20, 1% BSA). Anti-chicken IgG (rabbit) and Protein-A immunogold particles (10 nm) were used for detection. For detection of CPN10::GFP chimera, anti-GFP monoclonal antibody and 5 nm anti-mouse immunogold conjugate (Sigma) were used.
Fluorescence microscopy
L. donovani promastigotes expressing CPN10:GFP chimera were grown to 2 × 107 cells ml-1. 1 ml of parasite culture was subjected to centrifugation for 10 min at 1500 × g. The supernatant was discarded and the cell pellet was resuspended in 200 μl of Dulbecco's PBS. 2 μl was applied to a multiwell microscope slide and subjected to fluorescence microscopy at 63-fold magnification. Samples were analysed in bright field and in the FITC channel using a Hitachi Model monochrome CCD camera. Images were taken and merged using the Improvision Open Lab® software package and exported in TIFF format. Cropping and juxtapositioning were done using Adobe Photoshop® software.
Pulsed Field Gel Electrophoresis
Leishmania cells were harvested by centrifugation and washed twice in PBS. Following centrifugation, parasites were resuspended in PBS and mixed with an equal volume of prewarmed 1,5 InCert agarose (FMC BioProducts, Rockland). This mixture was aliquoted in block formers (2 × 107 parasites per block, Pharmacia). To lyse parasites, the agarose blocks were incubated in 2 mg/ml proteinase K (1% Laurylsarkosyl, 0,5 m EDTA pH 9.0) at 37°C for 48 h, and stored in 0.5 EDTA at 4°C.
PFGE was performed at 13°C in 0,25 × TBE buffer under the following conditions: intervall in sec: 100-10 log; switching angle: -110 lin; voltage: 200-150 log (Rotaphor, Biometra).
Saccharomyces cerevisiae strain YPH80 chromosomes (New England Biolabs) were used as size standards.
After staining with ethidium bromide (1 μg ml-1) and strand breakage by a 5 min exposure to UV radiation (254 nm), the gel was blotted onto a positively charged nylon membrane (Qiagen) by alkaline transfer [20]. The membrane was hybridised with a digoxigenin-labeled CPN10 probe. Blots were stained as described above.
Digital imaging
Primary experimental data were digitalised either on a flatbed scanner (Microtek Scanmaker 8700) or on a 35 mm film scanner (NIKON LS-4000). Phosphor imager (Molecular Dynamics) data were imported directly as TIFF images. Images were cropped and optimised for colour saturation using Adobe Photoshop® software. No filters or other image altering functions were employed on the images or parts thereof. Images were combined with vector graphics and text using ClarisDraw® software, version 1.0d.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
F.B.Z-V. performed the cloning of CPN10, the pulsed field gel electrophoresis, RT-PCR, the expression of recombinant protein and production of polyclonal antibodies, the construction, transfection and analysis of CPN10::GFP chimera, and the initial immunoblot analyses.
M.K. performed the immune electron microscopy and additional immunoblot experiments.
D.Z. amplified and characterised CPN10 DNA from L. donovani genomic DNA.
J.C. conceived the study, supervised the execution and prepared the final draft of the manuscript.
All authors participated in the drafting of the manuscript, and read and approved the final manuscript.
Acknowledgements
We wish to thank C. Hoyer and K. Mellenthin for the use of cosmid DNA libraries and for practical advice, T. Jacobs for help with fluorescence microscopy, S. Krobitsch for a gift of plasmid p426/PQ25, and A. Macdonald for further technical assistance. F. Zamora-Veyl was a post-doctoral fellow of the Deutscher Akademischer Austauschdienst.
==== Refs
Zilberstein D Shapira M The role of pH and temperature in the development of Leishmania parasites Annu Rev Microbiol 1994 48 449 470 7826014 10.1146/annurev.mi.48.100194.002313
Hubel A Krobitsch S Horauf A Clos J Leishmania major Hsp100 is required chiefly in the mammalian stage of the parasite Mol Cell Biol 1997 17 5987 95 9315657
Krobitsch S Brandau S Hoyer C Schmetz C Hübel A Clos J Leishmania donovani heat shock protein 100: characterization and function in amastigote stage differentiation J Biol Chem 1998 273 6488 6494 9497383 10.1074/jbc.273.11.6488
Schlueter A Wiesgigl M Hoyer C Fleischer S Klaholz L Schmetz C Clos J Expression and Subcellular Localization of Cpn60 Protein Family Members in Leishmania donovani Biochim Biophys Acta 2000 1491 65 74 10760571
Neupert W Pfanner N Roles of molecular chaperones in protein targeting to mitochondria Philos Trans R Soc Lond B Biol Sci 1993 339 355 61 discussion 361-2 8098540
Bukau B Horwich AL The Hsp70 and Hsp60 chaperone machines Cell 1998 92 351 66 9476895 10.1016/S0092-8674(00)80928-9
Richardson A Landry SJ Georgopoulos C The ins and outs of a molecular chaperone machine Trends Biochem Sci 1998 23 138 43 9584617 10.1016/S0968-0004(98)01193-1
Nielsen KL Cowan NJ A single ring is sufficient for productive chaperonin-mediated folding in vivo Mol Cell 1998 2 93 9 9702195 10.1016/S1097-2765(00)80117-3
Nielsen KL McLennan N Masters M Cowan NJ A single-ring mitochondrial chaperonin (Hsp60-Hsp10) can substitute for GroEL-GroES in vivo J Bacteriol 1999 181 5871 5 10482535
de Crouy-Chanel A el Yaagoubi A Kohiyama M Richarme G Reversal by GroES of the GroEL preference from hydrophobic amino acids toward hydrophilic amino acids J Biol Chem 1995 270 10571 5 7737993 10.1074/jbc.270.18.10571
Dubaquie Y Looser R Funfschilling U Jeno P Rospert S Identification of in vivo substrates of the yeast mitochondrial chaperonins reveals overlapping but non-identical requirement for hsp60 and hsp10 Embo J 1998 17 5868 76 9774331 10.1093/emboj/17.20.5868
Clos J Brandau S pJC20 and pJC40 – two high-copy-number vectors for T7 RNA polymerase-dependent expression of recombinant genes in Escherichia coli Prot Expression Purif 1994 5 133 137 10.1006/prep.1994.1020
Studier FW Rosenberg AH Dunn JJ Dubendorff JW Use of the T7 RNA polymerase to direct expression of cloned genes Methods Enzymol 1990 185 60 89 2199796
Brandau S Dresel A Clos J High constitutive levels of heat-shock proteins in human-pathogenic parasites of the genus Leishmania Biochem J 1995 310 225 32 7646449
Krobitsch S Clos J A novel role for 100 kD heat shock proteins in the parasite Leishmania donovani Cell Stress Chaperones 1999 4 191 198 10547068 10.1379/1466-1268(1999)004<0191:ANRFKH>2.3.CO;2
Hoyer C Mellenthin K Schilhabel M Platzer M Clos J Leishmania and the Leishmaniases: Use of genetic complementation to identify gene(s) which specify species-specific organ tropism of Leishmania Med Microbiol Immunol 2001 190 53 56
Krobitsch S Lindquist S Aggregation of huntingtin in yeast varies with the length of polyglutamine expansion and the expression of chaperone proteins Proc Natl Acad Sci USA 2000 97 1589 1594 10677504 10.1073/pnas.97.4.1589
Hoyer C Zander D Fleischer S Schilhabel M Kroener M Platzer M Clos J A Leishmania donovani gene that confers accelerated recovery from stationary phase growth arrest Int J Parasitol 2004 34 803 11 15157763 10.1016/j.ijpara.2004.02.006
Hubel A Brandau S Dresel A Clos J A member of the ClpB family of stress proteins is expressed during heat shock in Leishmania spp Mol Biochem Parasitol 1995 70 107 18 7637691 10.1016/0166-6851(95)00012-P
Sambrook J Fritsch EF Maniatis T Molecular Cloning Plainview 1989 NY.: Cold Spring Harbor Laboratory Press
| 15862128 | PMC1097755 | CC BY | 2021-01-04 16:38:32 | no | Kinetoplastid Biol Dis. 2005 Apr 29; 4:3 | utf-8 | Kinetoplastid Biol Dis | 2,005 | 10.1186/1475-9292-4-3 | oa_comm |
==== Front
Kinetoplastid Biol DisKinetoplastid Biology and Disease1475-9292BioMed Central London 1475-9292-4-31586212810.1186/1475-9292-4-3Original ResearchStage-specific expression of the mitochondrial co-chaperonin of Leishmania donovani, CPN10 Zamora-Veyl Fanny Beatriz [email protected] Manfred [email protected] Dorothea [email protected] Joachim [email protected] Bernhard Nocht Institute for Tropical Medicine, Bernhard Nocht St. 74, D-20359 Hamburg, Germany2005 29 4 2005 4 3 3 27 1 2005 29 4 2005 Copyright © 2005 Zamora-Veyl et al; licensee BioMed Central Ltd.2005Zamora-Veyl 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
Leishmania spp., in the course of their parasitic life cycle, encounter two vastly different environments: the gut of sandflies and the phagosomes of mammalian macrophages. During transmission into a mammal, the parasites are exposed to increased ambient temperature as well as to different carbon sources. Molecular chaperones or heat shock proteins are implicated in the necessary adaptations which involve the ordered differentiation from the flagellated, extracellular promastigote to the intracellular amastigote stage.
Results
Here, we show that the Leishmania donovani co-chaperonin, CPN10, is synthesised to a significantly increased concentration during in vitro differentiation to the amastigote stage. We show by fluorescence microscopy and by immunogold electron microscopy that, like its putative complex partner CPN60.2, CPN10 is localised to the single, tubular mitochondrion of the parasites and, moreover, that it co-precipitates with CPN60.2, the major mitochondrial chaperonin of Leishmania spp..
Conclusion
Our data indicate an increased requirement for CPN10 in the context of mitochondrial protein folding during or early in the mammalian stage of this pathogen. Moreover, they confirm the CPN60.2 as bona fide mitochondrial GroEL homologue in L. donovani and the postulated interaction of eukaryotic chaperonins, CPN60 and CPN10.
==== Body
Background
The euglenozoa which, beside the euglenophyta, also include the pathogenic protozoa of the Order Kinetoplastida, have branched off early in eukaryotic evolution. The kinetoplastida have developed, or retained, molecular mechanisms not shared by most Crown Group Eukaryota, such as trans-splicing of messenger RNA, RNA editing, and polycistronic transcription. The genetic information of their single mitochondrion, the so-called kDNA, is localised to the kinetoplast, an organelle unique to this order. The kinetoplast is part of a tubular, single mitochondrion that shows little morphological likeness to the mitochondria of Crown Group Eukaryota. The kinetoplastida encompass two genera of parasitic pathogens, Trypanosoma and Leishmania, both of which responsible for considerable human morbidity and mortality.
The parasites of the genus Leishmania exist in two morphologically distinct life cycle stages. The slender, flagellated, promastigote proliferates in the alimentary tract of female sandflies at ambient temperature. The amastigote, a rounded, intracellular form, with no protruding flagellum, persists in the phagosomes of mammalian macrophages and at temperatures of up to 39°C. The difference of temperatures in the two environments not only triggers the respective stage development [1], but also poses a challenge to the system of protein chaperones. At least two heat shock proteins, Hsp100 and CPN60.2, show significantly induced abundance during the amastigote stage [2-4].
Apart from the temperature difference, the two environments, insect gut and macrophage phagosome, also offer different qualitative and quantitative composition of carbon sources. The life cycle stage differentiation, therefore, must include an adaptation of the parasites' metabolic pathways. Subcellular compartments, e.g. the single tubular mitochondrion, also have to undergo adaptation. The folding of a different set of metabolic enzymes after their translocation across the outer and inner mitochondrial membranes can be expected to challenge the mitochondrial chaperone system.
Import of nuclear encoded proteins is vital for the function of mitochondria. A complex machinery of chaperone proteins is required for the unfolding of mitochondrial proteins, their trans-membrane transport, and their refolding inside the mitochondrial matrix [5]. The latter is thought to be brought about by the 60 kDa chaperonin (CPN60). This protein is a functional and structural homologue to the GroEL chaperonin of the eubacteriae. The function of GroEL, being the subject of considerable research efforts, is well understood [6,7]. In bacteria, GroEL forms barrel-like structures of two rings of 7 subunits each, with two central cavities. 7 subunits of the GroES co-chaperonin form a lid that covers one of the cavities. Folding intermediates of client proteins are bound inside one cavity where they may attain energetically favoured conformations without interference from other polypeptides.
Less is known, by comparison, about the function of eukaryotic GroEL homologues, the CPN60 proteins. They are nuclear-encoded mitochondrial proteins. Rather than forming two rings of seven GroEL subunits, the chaperonin complex in mitochondria consists of a single ring of seven CPN60 and one ring of CPN10 subunits, and it was shown that the function of CPN60/CPN10 complexes inside the mitochondria is similar to the chaperonin function of bacterial, cytosolic GroEL/GroES complexes [8,9]. GroES may also play a role in modulating the substrate specificity of GroEL. For instance it was reported that GroES binding shifts the preference of GroEL for hydrophobic amino acids towards amino acids with hydrophilic side chains [10]. In yeast, not all CPN60 substrates need CPN10 for proper folding. Other proteins require both CPN60 and CPN10, while newly imported CPN60 depends on CPN10 for correct folding [11].
We had previously identified the Leishmania donovani CPN60.2 gene which encodes the mitochondrial 60 kD chaperonin [4]. This protein shows increased expression during in vitro promastigote to amastigote differentiation, indicating an altered requirement for CPN60 in the amastigote stage. A second, diverged gene copy, CPN60.1, is not expressed to detectable levels in any in vitro culture stage of L. donovani or L. major [4]. To better understand mitochondrial protein chaperoning during the leishmanial life cycle, we identified the CPN10 gene of Leishmania donovani, determined its abundance in the two life cycle stages, and analysed the subcellular localisation of the gene product and its association with CPN60.2.
Results
Amplification of L. donovani CPN10 DNA
An alignment of the known eukaryotic CPN10 amino acid sequences revealed four limited stretches of sequence conservation (not shown). Oligonucleotides CPN10.1 to CPN10.4 were designed according to the codon bias of Leishmania spp.. Using L. donovani genomic DNA as template, combinations of the aforementioned primers were tested in a polymerase chain reaction. Only the combination of primers CPN10.1 and CPN10.4 yielded an amplification product of the expected size (not shown). The amplification product was then subcloned into the vector pBluescript and subjected to sequence analysis using standard pBluescript sequencing primers. The sequence analysis confirmed that the amplification product harboured probable CPN10 sequences (not shown).
Isolation of a CPN10 genomic DNA clone
The amplified putative CPN10 gene fragment was used to screen a L. donovani genomic DNA cosmid library [3]. Positive cosmids were subjected to a restriction digest and Southern Blot analysis to identify common subfragments which harboured the CPN10 gene(s). A 1.5 kb XhoI I fragment which hybridised with the CPN10 probe was subcloned into pBluescript and subjected to sequence analysis. The sequence (GenBank AF394959) contains an uninterrupted open reading frame of 303 bp encoding a putative 10.7 kDa polypeptide. A sequence comparison by BLAST search confirmed this putative polypeptide as related to eukaryotic and prokaryotic co-chaperonins. Figure 1A shows the position of the L. donovani CPN10 in the lineage of eukaryotic CPN10 proteins. L. donovani CPN10 groups within the kinetoplastid homologs. Figure 1B shows an alignment with selected eukaryotic CPN10 polypeptides. The high degree of sequence conservation between LdCPN10 and its counterparts in higher eukaryotes indicates a conserved function.
Figure 1 (A) Phylogenetic analysis (clustalW) of eukaryotic CPN10 family members. The bar shows the scale for sequence divergence. (B) Amino acid sequence alignment of the LdCPN10 and selected eukaryotic CPN10 family members. The alignment was perfomed using the clustalW algorithm.
Southern Analysis
Within the haploid set of chromosomes, L. donovani CPN10 appears to be single copy. This was determined by Southern Blot analysis of L. donovani genomic DNA, digested with 7 restriction endonucleases (Figure 2A). The pattern of restriction fragments does not indicate the presence of additional gene copies. We also performed pulsed field gel electrophoresis (Figure 2B) to determine the chromosomal localisation of the CPN10 gene. The CPN10 specific probe hybridised to a chromosomal band of approximately 1,100 kb. This size corresponds to chromosomes 26 or 27b. This was confirmed when sequencing of the L. major genome was finished; two copies of CPN10 are found on chromosome 26 .
Figure 2 (A) Southern analysis of L. donovani genomic DNA. The DNA (2 μg) was digested with restriction enzymes as indicated and separated on a 1 % agarose gel. After Southern transfer the membrane was probed with a CPN10 specific probe. The positions of size markers (1 kb ladder, Fermentas) are indicated. (B) Pulsed field gel electrophoresis of L. donovani chromosomes. The gel was stained with ethidium bromide and photographed (right). After denaturation the DNA was blotted on a Nylon membrane and probed with a CPN10 specific probe (left). The sizes of S. cerevisiae chromosomes which were run alongside as size markers are shown to the right. The chromosome band which hybridised with the CPN10 probe is marked by an arrow. (C) RT-PCR of L. donovani RNA. Whole cell RNA was isolated from L. donovani cells and subjected to reverse transcription at 37°C for 60 min using an oligo-dT primer. An aliquot of the cDNA reaction mix was inactivated at 65°C for 10 min prior to incubation at 37°C. Either RNA (lane 1), the cDNA (lane 2), the heat inactivated cDNA mix (lane 3) or genomic DNA (lane 4) were added to a PCR reaction mix using CPN10 specific primers. The sizes of selected marker DNA molecules is shown to the right.
RT-PCR analysis
Northern analysis of L. donovani RNA using a CPN10-specific probe did not yield any bands in the expected size range. To ascertain that the putative CPN10 gene encodes a stable RNA, we performed RT-PCR. Whole cell RNA was reversely transcribed using an oligo-dT primer. The first strand cDNA was used as template in a PCR amplification using primers CPN10-1 and CPN10-4. Negative controls included an amplification of material from a cDNA synthesis mix that had been inactivated at 65°C for 10 min (Figure 2C, lane 1), and an amplification with an equivalent amount of whole cell RNA as template (lane 3), to exclude DNA contaminations in the RNA preparation. Both reactions did not give rise to an amplification product. The reaction with cDNA as template (lane 2), however, produced a ~350 bp PCR product, much the same as the reaction with genomic DNA as template (lane 4). We conclude that stable, polyadenylated RNA is produced from the CPN10 gene.
Production of anti-CPN10 antibodies
The CPN10 open reading frame was subcloned into pJC45, a derivative of pJC40 [4,12], for expression in E. coli strain Bl21 (DE3) [13]. His-tagged CPN10 was purified by metal chelate chromatography (Figure 3) and used to immunise laying hens. Antibodies were prepared from egg yolk before and after immunisation to yield pre-immune and CPN10-specific antibody preparations. In Western Blot analyses, these antibodies faithfully recognised recombinant rCPN10 (not shown).
Figure 3 Expression of L. donovani CPN10 in E. coli. The plasmids pJC45 or pCPN10-45 was used to transform E. coli strain BL21 (DE3) [pAPlacIQ]. Expression was induced by addition of IPTG, and the bacteria were harvested. After lysis, aliquots of the cultures before and after IPTG induction were analysed by SDS-PAGE and Coomassie Blue staining. The lysate containing the rCPN10 was subjected to metal chelate chromatography to purify the rCPN10 by virtue of its histidine tag. The purified rCPN10 was also subjected to SDS-PAGE analysis.
LdCPN10 expression is induced in the amastigote
Using the anti-CPN10 antibodies, we performed immunoblot analysis on lysates of L. donovani promastigotes before and after heat shock, and of axenically cultured amastigotes (Figure 4B). In promastigotes cultured at 25°C, we observe only a faint band in the expected size range. After a 24 h treatment of the promastigotes at 37°C, a signal is observed. The CPN10 signal increases strongly during in vitro amastigote differentiation. Figure 4A shows identical samples separated by SDS-PAGE and Coomassie Blue staining. We conclude that Leishmania donovani CPN10 is expressed preferentially in the amastigote stage, induced at least in part by the elevated temperature.
Figure 4 (A) SDS-PAGE of various in vitro culture forms of L. donovani. Equivalent protein aliquots from L. donovani promastigotes, cultivated at either 25°C (promastigote -25°), at 37°C (promastigote -37°), and from axenic amastigotes after 3 days of in vitro differentiaton were subjected to Tricine Gel Electrophoresis and to SDS-PAGE. Protein bands were visualised by Coomassie Brilliant Blue staining. (B) Immunoblot analysis. A parallel gel was subjected to Western transfer and probed with an anti-CPN10 antibody. (C) Co-immune precipitation. Leishmania donovani lysate was precipitated using anti-CPN10 antibodies, separated by SDS-PAGE and subjected to WesternBlot analysis with anti-CPN60.2 antibodies. Lanes 1 and 2 are negative controls representing immune precipitations without anti-CPN10 antibody and with a preimmune antibody preparation, respectively. Lane 3 represents the CPN10 immune precipitation proper. The position of CPN60.2 is indicated.
CPN10 is the co-chaperonin in L. donovani
By analogy, CPN10 is expected to act as co-chaperonin to the 60 kDa chaperonin, CPN60. However, there are two diverged members of the CPN60 gene family in L. donovani, CPN60.1 and CPN60.2. Of these, only CPN60.2 could be shown to encode a protein which localised to the mitochondrion [4]. We therefore tested whether CPN10 interacts with CPN60.2 protein in L. donovani and performed co-immune precipitation. L. donovani cell lysates were incubated with anti-CPN10 antibodies, anti-chicken IgG (rabbit) and Protein-A agarose. The precipitated material was then analysed by immunoblot using anti-CPN60.2 antibody. As shown in Figure 4C, the anti-CPN10 precipitate contains a protein which reacts with the anti-CPN60.2 antibodies. We conclude that CPN60.2 is the bona fide chaperonin and that CPN10 serves as its co-chaperonin.
Subcellular localisation of LdCPN10
Since CPN10 serves as co-chaperone to CPN60 we expected CPN10 to localise to the kinetoplast/mitochondrion complex inside the leishmaniae. To test this prediction, the CPN10 coding sequence was fused to a GFP (Green Fluorescent Protein) coding DNA in the expression vector pIRmcs3-. The linearised expression construct was recombined into the rRNA locus of L. donovani to yield stable expression of the CPN10::GFP chimera. The recombinant parasites were cultivated as promastigotes and analysed by fluorescence microscopy. The results are shown in Figure 5. We observe brightly fluorescent tubular structures in the promastigotes. No fluorescence is observed in the cytoplasm. This indicates that the CPN10 sequence can quantitatively direct GFP into a subcellular compartment or organelle.
Figure 5 Fluorescence microscopy. L. donovani promastigotes transfected with a CPN10/GFP gene chimera on plasmid pIRCPN10::GFP were subjected to bright field microscopy at 63× magnification (panels A-C). The same microscopic fields were viewed under UV excitation at XXX nm (panels D-F). Panels G-I show the overlays.
To identify the subcellular compartment that harbours CPN10, we performed immune electron microscopy using sections of heat stressed promastigotes and of axenic amastigotes. Figure 6 shows the result. We find the gold particles associated both with tubular structures and with the kinetoplast. We conclude that CPN10 is a mitochondrial protein.
Figure 6 Immunogold electron microscopy. Heat-shocked promastigotes (a,d) or axenic amastigotes (b,c) were embedded in LR-White. Microsections were stained with pre-immune antibodies (a) or anti-CPN10 (b-c) antibodies. Antibody binding was detected using rabbit anti-chicken IgG and Protein A Gold (10 nm). Annotations: n = nucleus; k = kinetoplast; m = tubular mitochondrium.
When we analysed the subcellular localisation of CPN60.2 in L. donovani, this chaperonin, too, localised to a tubular subcellular compartment. However, we did not find it associated with the kinetoplast. Therefore, the exact subcellular localisation was still unclear. Using the strain which overexpressed the CPN10::GFP chimera, we performed a co-immune electron microscopy. We used anti-GFP mAB in conjunction with anti-mouse immunogold particles (5 nm). Then the sections were co-stained using anti-CPN60 antibody, anti-chicken IgG (rabbit), and protein A gold (10 nm). As shown in Figure 7, CPN10::GFP and CPN60 colocalise to the same tubular compartment. This shows that CPN60.2 is indeed a mitochondrial protein in Leishmania.
Figure 7 Double immune electron microscopy. Leishmania cells that express CPN10::GFP chimera were embedded and microsections were prepared. After staining with anti-GFP Mab and anti-mouse immunogold (5 nm), sections were treated with anti-CPN60 antibodies, rabbit anti-chicken IgG, and Protein A gold (10 nm). Thus, 10 nm gold particles represent CPN60 and 5 nm gold particles represent CPN10::GFP. k = kinetoplast.
Discussion
We have cloned and analysed the CPN10 gene of the protozoan parasite Leishmania donovani. This is, to our knowledge, the first characterisation of a protozoan member of the conserved GroES family of co-chaperonins. Since CPN10 is a very small protein with an impact on mitochondrial function, the restraints on sequence variation should be rather tight. Accordingly, Leishmania CPN10 is >60 % similar to mammalian CPN10 sequences. A comparison of the L. donovani CPN10 gene with other eukaryotic GroES homologs further confirms that the leishmaniae belong to a remote entity among the eukaryota, clearly separated from the Eukaryotic Crown Group. A similarly diverged position was observed for the GroEL homologue of L. donovani, CPN60.2 [4].
The gene for CPN10 is single copy in the L. donovani genome. This is in contrast to the CPN60.2 gene which is present in two copies, in addition to a, as far as known silent, CPN60.1 gene [4]. It is also in contrast with the situation in L. major. According to the L. major genome database, the L. major CPN10 is encoded by two gene copies located in tandem. If confirmed, this difference could serve to distinguish both Leishmania species which have overlapping endemic regions.
Eukaryotic CPN10, like its bacterial counterpart, GroES, is assumed to act as co-chaperonin to CPN60 in the mitochondrial matrix. Immune electron microscopy reveals that CPN10 localises both to the kinetoplast and to the tubular part of mitochondrion. More importantly, co-immunoprecipitation experiments establish a direct interaction between CPN10 and CPN60.2. This not only indicates a function of CPN10 as bona fide co-chaperonin, but also establishes CPN60.2 as the Leishmania GroEL homolog. In addition, the mitochondrial localisation of both proteins is now well established.
The mitochondrial transport signal sequence of CPN10 can direct the import of the much larger CPN10::GFP chimera into the mitochondria and the kinetoplast. The chimera is expressed stably from the integration site within the rRNA gene repeats. The chimera may serve for future experimental designs as mitochondrial localiasation marker and as marker for the integrity of the mitochondrion.
The CPN10 gene is expressed preferentially under heat stress and in axenic amastigotes of L. donovani. The putative complex partner, CPN60.2, also shows increased abundance under heat stress [4]. The induction of CPN10, however, is more pronounced, and comparable to the induction of Hsp100 [3]. In this regard, three heat shock proteins, Hsp100, Cpn60.2, and CPN10, differ from the major chaperones, Hsp70 and Hsp90 (Hsp83). The latter show high constitutive abundance and only a marginal increase during heat stress and in the amastigote stage [3,14]. It is likely that these different expression kinetics reflect different roles during the life cycle. Hsp100 is required only in intracellular amastigotes of L. donovani and has a role in amastigote specific gene expression [15]. CPN60.2, and even more so, CPN10, are obviously also in increased demand during the amastigote stage.
Is there a specific function for CPN10 during the mammalian stage of the parasite's life cycle? It has been suggested that GroES can alter the affinity of GroEL for amino acid side chains [10]. In the yeast, substrate proteins may fold either independently of CPN10 or in a CPN10-dependent manner [11]. This raises the possibility that the preferential expression of CPN10 in amastigotes of L. donovani may reflect a change in the substrate specificity of the mitochondrial chaperonin complex, necessitated by the change of environment and the concomitant changes in nutrient availability. Since CPN10 is a single copy gene that is hardly expressed in the highly proliferative promastigote stage, it may be a suitable target for a gene replacement approach. This should allow to assess the role of CPN10 for the virulence and pathogenicity of Leishmania parasites.
Conclusion
We have identified the Leishmania donovani homologue of the bacterial co-chaperonin, GroES, and assigned the name CPN10 to the gene and the protein. CPN10 is a heat shock protein and shows increased intracellular concentration at an elevated temperature and in axenically cultivated amastigotes of L. donovani. CPN10 localises to the mitochodrial compartment and co-localises with the GroES homologue, CPN60.2. A direct interaction of both proteins was demonstrated using co-immuno precipitation. Both CPN10 and CPN60.2 are therefore suitable markers for the mitochondria of Leishmania spp. and related species. Moreover, the increased concentration in the amastigote stage suggests a role of CPN10 in this stage.
Methods
Parasite strains and culture
For our analyses we used the Lo8 strain of L. donovani, a gift from D. Zilberstein. Promastigotes were routinely cultivated at 25°C in supplemented M199 medium [3]. Cell density was monitored using a Schaerfe Systems CASY Cell Counter. In vitro promastigote to amastigote differentiation was performed as described [3].
PCR amplification of CPN10 DNA
Four primers were delineated from an amino acid sequence comparison of known GroES and CPN10 proteins. The extreme bias in Leishmania towards G and C residues in the wobble positions of codons was used to create primers with minimal degeneracy:
primer CPN10.1: CCGCTGTTCG ACCGCGTGCT GG
primer CPN10.2: GGCGGCATCR TGCTGCC
primer CPN10.3: CGGCAGCAYG ATGCCGCC
primer CPN10.4: CAGCACCTTG TCGCCCACCT TCAC
100 ng of L. donovani genomic DNA was mixed with 40 ng of oligonucleotide primer pairs, 25 nmoles of each dNTP, 10 × reaction buffer, and 1 unit of Taq DNA polymerase (Beckman) in a 50 μl reaction. The reaction mix was incubated for 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C. The cycle was repeated 35 times. 10% of the reaction was analysed on a 1% agarose gel. Negative controls included each primer by itself and omission of template DNA. Bands absent from the negative controls were recovered by preparative agarose gel electrophoresis and by using glass beads adsorption (PureGen Kit, Biozyme). PCR products were subcloned using the TA cloning kit (Invitrogen) and subjected to sequence analysis..
Reverse transcriptase (RT-) PCR
Total RNA from promastigotes stage parasites strain Lo8 was prepared using a Ribolyser Kit (Hybaid). cDNA synthesis was carried out using the First Strand synthesis kit (Pharmacia) with the enclosed (dT)18 primer and 5 μg of total RNA. The cDNA was then amplified enzymatically using the primers CPN10.1 and CPN10.4 The amplification was carried out as described above.
Library screening
We used a cosmid library of L. donovani previously described in [16]. The library was screened by hybridisation with digoxigenin-labeled DNA probes derived from the subcloned amplification products. Hybridisation was performed at 65°C in HYB 9 solution (Biozyme) for 16 h. Washes were performed at decreasing SSC concentrations (6 × to 0.2 × SSC, 0.5 % SDS) at 65°C. Filters were developed using anti-digoxigenin FAB/AP conjugate (Boehringer Mannheim) followed by colorimetric staining with NBT/BCIP. Positive cosmid clones were picked, rescreened, and verified in an amplification reaction using CPN10-specific primers.
Subcloning and sequence analysis
Cosmid DNA from positive clones was subjected to restriction endonucleased digest and Southern Blot to determine suitable restriction endonucleases. Cosmid DNA was then digested at a preparative scale and subcloned into an appropriately cleaved pBluescript KS+ vector (Stratagene) or pJC45 vector [4]. Plasmid subclones were then subjected to bidirectional primer walking sequence analysis, starting from the known sequences of the PCR amplificates. Sequencing was performed on an Applied Biosystems Model 370 sequencer using the DyeTerminator Kit.
Sequence evaluation
Contig alignment, in silico translation, and sequence alignment were performed using the MacMolly Tetris and the MacVector software packages. Phylogenetic analyses were performed using the clustalW algorithm included with MacVector at default settings.
Recombinant Expression of CPN10
The putative open reading frame of CPN10 was amplified using specific primers that modify the start codon sequence into a NdeI site and the stop codon sequence into an EcoRI site. The amplification products were cleaved with NdeI and EcoRI and ligated into the expression vector pJC45 [4], between the NdeI site and the EcoRI site. The expression plasmids were transformed into the bacterial strain BL21 (DE3) [pAPlacIQ], a gift from Olivier Payet, CNRS Toulouse. Bacteria were grown in CircleGrow medium (BIO 101) supplemented with 50 μg/ml Ampicillin and 10 μg/ml Kanamycin to OD600 = 0.5. IPTG was added to 1 mM and incubation was continued for 1 h. The purification of His-tagged proteins from bacterial lysates by metal chelate chromatography has been described [12,14].
To express CPN10::GFP chimera, the CPN10 coding sequence, minus the stop codon, was enzymatically amplified to create BamHI and HindIII sites at the 5' and 3' ends, respectively. Digested with BamHI and HindIII, the amplification products were ligated between the BamHI and HindIII sites of the yeast plasmid p426/PQ25 [17]. By this procedure, CPN10 and GFP coding sequences were fused in frame. The chimeric CPN10::GFP coding sequence was then excised with BamHI and XhoI and ligated into pIRmcs3- [18]. The plasmid pIRCPN10::GFP and the parent plasmid pIRmcs3- were linearised using SwaI and transfected into L. donovani by electroporation. Recombinant parasites were placed under ClonNAT (Werner Bioreagents) selection (100 μg/ml) and analysed by fluorescence microscopy.
Immunisation and antibody preparation
The immunisation of laying hens and the preparation of antibodies from egg yolk has been described [19].
Immunoblot analysis
SDS-PAGE and Western Transfer were performed as described [14,19]. Briefly, membranes were treated with blocking buffer (4% skim milk powder and 0.1% Tween 20 in DPBS), with antibodies at appropriate dilutions in blocking buffer, and with secondary antibody/alkaline phosphatase conjugate (Dianova) in blocking buffer. Blots were stained with nitrobluetetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate.
Immune precipitation
1 × 108 promastigotes were harvested by centrifugation and lysed in 500 μl solubilising buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, and 2.5 μg × ml-1 each of aprotinin, leupeptin, pepstatin, and antipain). Lysates were centrifuged at 13000 × g, 4°C. Soluble proteins were incubated with 20 μl anti-CPN10 antibody for 1 h at 4°C. 350 μl of protein A-sepharose slurry were preincubated with anti-chicken IgG from rabbit and added. The slurry was further incubated at 4°C for 2 h. The immunoabsorbent was centrifuged and washed three times in washing buffer A (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0,2% NP-40, 2 mM EDTA), twice in washing buffer B (as above, except NaCl is 500 mM) and once in washing buffer C (10 mM Tris-HCl pH 7.5). For analysis on SDS-PAGE 250 μl of SDS sample buffer was added and the samples were heated for 5 minutes at 95°C. After centrifugation, 50 μl of the supernatant was loaded onto a 12% PA gel.
Immune electron microscopy
Immune electron microscopy was performed essentially as described [3]. LR-White embedded microsections were treated with chicken antibody at 1:500 (anti-CPN10) or 1:2000 (anti-CPN60.2) in PBS (0.1 % Tween 20, 1% BSA). Anti-chicken IgG (rabbit) and Protein-A immunogold particles (10 nm) were used for detection. For detection of CPN10::GFP chimera, anti-GFP monoclonal antibody and 5 nm anti-mouse immunogold conjugate (Sigma) were used.
Fluorescence microscopy
L. donovani promastigotes expressing CPN10:GFP chimera were grown to 2 × 107 cells ml-1. 1 ml of parasite culture was subjected to centrifugation for 10 min at 1500 × g. The supernatant was discarded and the cell pellet was resuspended in 200 μl of Dulbecco's PBS. 2 μl was applied to a multiwell microscope slide and subjected to fluorescence microscopy at 63-fold magnification. Samples were analysed in bright field and in the FITC channel using a Hitachi Model monochrome CCD camera. Images were taken and merged using the Improvision Open Lab® software package and exported in TIFF format. Cropping and juxtapositioning were done using Adobe Photoshop® software.
Pulsed Field Gel Electrophoresis
Leishmania cells were harvested by centrifugation and washed twice in PBS. Following centrifugation, parasites were resuspended in PBS and mixed with an equal volume of prewarmed 1,5 InCert agarose (FMC BioProducts, Rockland). This mixture was aliquoted in block formers (2 × 107 parasites per block, Pharmacia). To lyse parasites, the agarose blocks were incubated in 2 mg/ml proteinase K (1% Laurylsarkosyl, 0,5 m EDTA pH 9.0) at 37°C for 48 h, and stored in 0.5 EDTA at 4°C.
PFGE was performed at 13°C in 0,25 × TBE buffer under the following conditions: intervall in sec: 100-10 log; switching angle: -110 lin; voltage: 200-150 log (Rotaphor, Biometra).
Saccharomyces cerevisiae strain YPH80 chromosomes (New England Biolabs) were used as size standards.
After staining with ethidium bromide (1 μg ml-1) and strand breakage by a 5 min exposure to UV radiation (254 nm), the gel was blotted onto a positively charged nylon membrane (Qiagen) by alkaline transfer [20]. The membrane was hybridised with a digoxigenin-labeled CPN10 probe. Blots were stained as described above.
Digital imaging
Primary experimental data were digitalised either on a flatbed scanner (Microtek Scanmaker 8700) or on a 35 mm film scanner (NIKON LS-4000). Phosphor imager (Molecular Dynamics) data were imported directly as TIFF images. Images were cropped and optimised for colour saturation using Adobe Photoshop® software. No filters or other image altering functions were employed on the images or parts thereof. Images were combined with vector graphics and text using ClarisDraw® software, version 1.0d.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
F.B.Z-V. performed the cloning of CPN10, the pulsed field gel electrophoresis, RT-PCR, the expression of recombinant protein and production of polyclonal antibodies, the construction, transfection and analysis of CPN10::GFP chimera, and the initial immunoblot analyses.
M.K. performed the immune electron microscopy and additional immunoblot experiments.
D.Z. amplified and characterised CPN10 DNA from L. donovani genomic DNA.
J.C. conceived the study, supervised the execution and prepared the final draft of the manuscript.
All authors participated in the drafting of the manuscript, and read and approved the final manuscript.
Acknowledgements
We wish to thank C. Hoyer and K. Mellenthin for the use of cosmid DNA libraries and for practical advice, T. Jacobs for help with fluorescence microscopy, S. Krobitsch for a gift of plasmid p426/PQ25, and A. Macdonald for further technical assistance. F. Zamora-Veyl was a post-doctoral fellow of the Deutscher Akademischer Austauschdienst.
==== Refs
Zilberstein D Shapira M The role of pH and temperature in the development of Leishmania parasites Annu Rev Microbiol 1994 48 449 470 7826014 10.1146/annurev.mi.48.100194.002313
Hubel A Krobitsch S Horauf A Clos J Leishmania major Hsp100 is required chiefly in the mammalian stage of the parasite Mol Cell Biol 1997 17 5987 95 9315657
Krobitsch S Brandau S Hoyer C Schmetz C Hübel A Clos J Leishmania donovani heat shock protein 100: characterization and function in amastigote stage differentiation J Biol Chem 1998 273 6488 6494 9497383 10.1074/jbc.273.11.6488
Schlueter A Wiesgigl M Hoyer C Fleischer S Klaholz L Schmetz C Clos J Expression and Subcellular Localization of Cpn60 Protein Family Members in Leishmania donovani Biochim Biophys Acta 2000 1491 65 74 10760571
Neupert W Pfanner N Roles of molecular chaperones in protein targeting to mitochondria Philos Trans R Soc Lond B Biol Sci 1993 339 355 61 discussion 361-2 8098540
Bukau B Horwich AL The Hsp70 and Hsp60 chaperone machines Cell 1998 92 351 66 9476895 10.1016/S0092-8674(00)80928-9
Richardson A Landry SJ Georgopoulos C The ins and outs of a molecular chaperone machine Trends Biochem Sci 1998 23 138 43 9584617 10.1016/S0968-0004(98)01193-1
Nielsen KL Cowan NJ A single ring is sufficient for productive chaperonin-mediated folding in vivo Mol Cell 1998 2 93 9 9702195 10.1016/S1097-2765(00)80117-3
Nielsen KL McLennan N Masters M Cowan NJ A single-ring mitochondrial chaperonin (Hsp60-Hsp10) can substitute for GroEL-GroES in vivo J Bacteriol 1999 181 5871 5 10482535
de Crouy-Chanel A el Yaagoubi A Kohiyama M Richarme G Reversal by GroES of the GroEL preference from hydrophobic amino acids toward hydrophilic amino acids J Biol Chem 1995 270 10571 5 7737993 10.1074/jbc.270.18.10571
Dubaquie Y Looser R Funfschilling U Jeno P Rospert S Identification of in vivo substrates of the yeast mitochondrial chaperonins reveals overlapping but non-identical requirement for hsp60 and hsp10 Embo J 1998 17 5868 76 9774331 10.1093/emboj/17.20.5868
Clos J Brandau S pJC20 and pJC40 – two high-copy-number vectors for T7 RNA polymerase-dependent expression of recombinant genes in Escherichia coli Prot Expression Purif 1994 5 133 137 10.1006/prep.1994.1020
Studier FW Rosenberg AH Dunn JJ Dubendorff JW Use of the T7 RNA polymerase to direct expression of cloned genes Methods Enzymol 1990 185 60 89 2199796
Brandau S Dresel A Clos J High constitutive levels of heat-shock proteins in human-pathogenic parasites of the genus Leishmania Biochem J 1995 310 225 32 7646449
Krobitsch S Clos J A novel role for 100 kD heat shock proteins in the parasite Leishmania donovani Cell Stress Chaperones 1999 4 191 198 10547068 10.1379/1466-1268(1999)004<0191:ANRFKH>2.3.CO;2
Hoyer C Mellenthin K Schilhabel M Platzer M Clos J Leishmania and the Leishmaniases: Use of genetic complementation to identify gene(s) which specify species-specific organ tropism of Leishmania Med Microbiol Immunol 2001 190 53 56
Krobitsch S Lindquist S Aggregation of huntingtin in yeast varies with the length of polyglutamine expansion and the expression of chaperone proteins Proc Natl Acad Sci USA 2000 97 1589 1594 10677504 10.1073/pnas.97.4.1589
Hoyer C Zander D Fleischer S Schilhabel M Kroener M Platzer M Clos J A Leishmania donovani gene that confers accelerated recovery from stationary phase growth arrest Int J Parasitol 2004 34 803 11 15157763 10.1016/j.ijpara.2004.02.006
Hubel A Brandau S Dresel A Clos J A member of the ClpB family of stress proteins is expressed during heat shock in Leishmania spp Mol Biochem Parasitol 1995 70 107 18 7637691 10.1016/0166-6851(95)00012-P
Sambrook J Fritsch EF Maniatis T Molecular Cloning Plainview 1989 NY.: Cold Spring Harbor Laboratory Press
| 15877819 | PMC1097756 | CC BY | 2021-01-04 16:36:36 | no | Mol Cancer. 2005 May 6; 4:17 | latin-1 | Mol Cancer | 2,005 | 10.1186/1476-4598-4-17 | oa_comm |
==== Front
Mol PainMolecular Pain1744-8069BioMed Central London 1744-8069-1-151584017210.1186/1744-8069-1-15ResearchPhysical interaction and functional coupling between ACDP4 and the intracellular ion chaperone COX11, an implication of the role of ACDP4 in essential metal ion transport and homeostasis Guo Dehuang [email protected] Jennifer [email protected] Mong-Heng [email protected] Jin-Xiong [email protected] Jianguo [email protected] Cong-Yi [email protected] Center for Biotechnology and Genomic Medicine, Medical College of Georgia, 1120 15th Street, CA4098, Augusta, GA 30912, USA2 Department of Oral and Maxillofacial Surgery, Mcknight Brain Institute and College of Dentistry, University of Florida, Gainesville, Florida, 32610, USA3 Department of Physiology, Medical College of Georgia, 1120 15th Street, Augusta, GA 30912, USA2005 19 4 2005 1 15 15 17 3 2005 19 4 2005 Copyright © 2005 Guo et al; licensee BioMed Central Ltd.2005Guo 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.
Divalent metal ions such as copper, manganese, and cobalt are essential for cell development, differentiation, function and survival. These essential metal ions are delivered into intracellular domains as cofactors for enzymes involved in neuropeptide and neurotransmitter synthesis, superoxide metabolism, and other biological functions in a target specific fashion. Altering the homeostasis of these essential metal ions is known to connect to a number of human diseases including Alzheimer disease, amyotrophic lateral sclerosis, and pain. It remains unclear how these essential metal ions are delivered to intracellular targets in mammalian cells. Here we report that rat spinal cord dorsal horn neurons express ACDP4, a member of Ancient Conserved Domain Protein family. By screening a pretransformed human fetal brain cDNA library in a yeast two-hybrid system, we have identified that ACDP4 specifically interacts with COX11, an intracellular metal ion chaperone. Ectopic expression of ACDP4 in HEK293 cells resulted in enhanced toxicity to metal ions including copper, manganese, and cobalt. The metal ion toxicity became more pronounced when ACDP4 and COX11 were co-expressed ectopically in HEK293 cells, suggesting a functional coupling between them. Our results indicate a role of ACDP4 in metal ion homeostasis and toxicity. This is the first report revealing a functional aspect of this ancient conserved domain protein family. We propose that ACDP is a family of transporter protein or chaperone proteins for delivering essential metal ions in different mammalian tissues. The expression of ACDP4 on spinal cord dorsal horn neurons may have implications in sensory neuron functions under physiological and pathological conditions.
==== Body
Background
Essential metal ions such as copper, manganese and cobalt are vital elements involved in functions of numerous enzymes and proteins in mammalian cells. Mammalian cells not only possess efficient uptake mechanisms to obtain these ions from their extracellular environment, but also have intracellular delivery system to translocate essential metal ions to specific enzymes and proteins. Deficiency in these essential metal ions affects normal cell functions, but they are toxic when present in excess. For example, they can damage DNA and proteins to induce cell death. Therefore, proper delivery of these essential metal ions into intracellular functional domains is vital. It is known that alteration of essential metal ion homeostasis is associated with diseases including Alzheimer's disease, amyotrophic lateral sclerosis, prion diseases, cataracts, mitochondrial disorders and Parkinson's disease [1-8]. Essential metal ion homeostasis also has important implications in sensory physiology, pathology and pain [9-11]. For example, copper is a cofactor for peptidylglycine a-amidating monooxygenase, an enzyme catalyzes the formation of a number of biologically active peptides including the pronociceptive peptide substance P [9]. Copper and manganese are cofactors for copper/zinc superoxide dismutase (Cu/ZnSOD) and Mn-superoxide dismutase (MnSOD), respectively [10,11]. These metal ion-dependent enzymes are involved in superoxide metabolism. Studies have shown that activity of these SOD enzymes can become abnormal during inflammation, which is an important underlying mechanism of pathological pain conditions and other neurological disorders [10,11].
Metal ion homeostasis is maintained through highly regulated processes, including transport, translocation, storage and secretion. Essential metal ions are transported into cells and then translocated to intracellular organelles to function as catalytic and structural cofactors for compartmentalized enzymes [12]. Although a number of ion transporters have been identified and characterized biochemically over the past several decades [13-18], the molecular identities of transporters for many metal ions are still elusive in mammalian cells. Unlike membrane ion channels permeable for ions such as Ca2+ [19], essential metal ion transporters are coupled with intracellular metal ion chaperones [12,20-22]. Metal ion chaperones interact with transporters to receive metal ions, and then carry metal ions to target enzymes or proteins and donate metal ions to the targets. Several chaperones for copper ions have been identified in mammalian cells including COX17 and COX11 and they are shown to be essential for cell survival [20-22].
We have recently cloned and characterized a novel gene family named Ancient Conserved Domain Protein (ACDP) [23]. ACDP encodes four protein members (ACDP1-4) in both human and mouse [23,24]. The most prominent feature of ACDP gene family is the ancient conserved domain (ACD) found in evolutionarily divergent species ranging from bacteria, yeast, C. elegans, and D. melanogaster to mammals. ACDP proteins showed high amino acid homology to bacteria CorC protein (e.g., 35% AA identity with 55% homology), a protein believed to be involved in metal ion toxicity in bacteria [25]. Based on TopPred analysis it appears that ACDP can be a family of membrane proteins [24]. Consistent with this analysis, a recent study from our group has suggested that ACDP1, the first member of ACDP family, is localized close to or on the plasma membranes of hippocampus neurons [24]. Because ACDP4 was shown to have broader tissue distributions, including in both neuronal and non-neuronal tissues, we took ACDP4 as a start for exploring functions of ACDP family.
Results
Expression of ACDP4 proteins on spinal cord dorsal horn neurons
We performed immunostaining of ACDP4 on the spinal cord dorsal horn neurons. ACDP1 immunostaining was also performed as additional information of ACDP family expression on dorsal horn neurons. Spinal cord dorsal horn neurons are CNS (central nervous system) neurons involved in somatosensory functions, including transmitting signals of nociceptive, mechanical, and thermal stimuli. Immunostaining was performed using polyclonal antibodies specific for ACDP1 and ACDP4, respectively. Both antibodies have been previously shown to have specific interactions with the corresponding target proteins [24]. We used cultured neurons grown on a monolayer of astracyte bedding for immunostaining. Therefore, cellular and subcellular distribution of ACDP1 and ACDP4 can be viewed clearly under a microscope. ACDP1 was expressed on both soma and dendrites of neurons (Fig. 1A &1B). Interestingly, ACDP1-ir on dendrites was shown to be punctuated (Fig. 1A &1B). ACDP4-ir was also found on the soma and dendrites of neurons, its distribution on dendrites was not punctuated (Fig. 1C). Neither ACDP1-ir nor ACDP4-ir was observed positively on astrocytes, the bedding underneath neurons. Confocal images showed that ACDP1-ir (Fig. 1D) and ACDP4-ir (Fig. 1E) were strongest at the edge of cells, suggesting that most ACDP1 and ACDP4 proteins were located close to and some of them maybe on the plasma membranes of spinal cord dorsal horn neurons.
Figure 1 Immunoreactivity of ACDP1 and ACDP4 on rat spinal cord dorsal horn neurons A. The Micrograph shows a cultured dorsal horn neuron with its processes. The image was taken under a DIC (Nomarski differential interference contrast) microscope. B. ACDP1 immunoreactivity (ACDP1-ir) on the same neuron in A. Arrows indicate several punctuated sites of ACDP1-ir along dendrites. C. ACDP4-ir on another dorsal horn neuron. Scale bars: 10 μm. Two-week neuron cultures were used for the immunostaining. D. A confocal image of ACDP1-ir on a dorsal horn neuron. E. A confocal image of ACDP4-ir on anther dorsal horn neuron. The images were taken at the meddle sections of the cells. Scale bars were 20 μm for A, B and C and 10 μm for D and E.
Interactions between ACDP4 and intracellular metal ion chaperone COX11
To explore the potential functions of ACDP proteins, we first searched whether any known proteins interact with them by screening a pretransformed human fetal brain cDNA library in a yeast two-hybrid system. This Matchmaker system is an advanced GAL4-based two-hybrid system that provides a transcriptional assay for detecting protein interactions in vivo in yeast. We used ACDP4 as a model in this study because it has broader tissue distribution than ACDP1. In the Matchmaker system ACDP4 (bait gene) was expressed as a fusion to the GAL4 DNA-binding domain (DNA-BD), which was used to screen a pretransformed human fetal Matchmaker cDNA library. When the bait and library fusion proteins interact, the DNA-BD and activation domain (AD) are brought into proximity, thus activating transcription of the reporter gene. In order to prevent non-specific interactions, we cultured the cells in the highest stringency condition. Among 2 × 106 transformants, only one clone was identified to strongly interact with ACDP4. The DNA for this clone was recovered from yeast and then transformed into XL-1 Blue competent cells (Stratagene). Individual clones were screened first by PCR and then sequenced using the primers from vector. We found that the yeast clone actually contains two plasmids, one plasmid is corresponding to the intracellular metal ion chaperone COX11, and the other one is an RNA binding motif protein 30 (RBM30) with unknown function.
To confirm the above results, we established full-length constructs for the pGADT7-COX11 and pGADT7-RBM30 (as targets). Subsequently, the pGBKT7-ACDP4 plasmid was co-transformed into an AH109 yeast strain either with the pGADT7-COX11 or the pGADT7-RBM30 plasmid. An empty pGADT7 vector was used as a control. The cultures were assayed for β-galactosidase to verify two-hybrid interactions (Fig. 2). In response to GAL4 activation, yeast containing the lacZ reporter gene secretes β-galactosidase, which can be detected in the presence of the chromogenic substrate ONPG (ortho-nitrophenyl β-D-galactopyranoside). When hydrolyzed by β-galactosidase, colorless ONPG turns to ONP (ortho-nitrophenyl), a yellow product. Therefore, enzyme activity of β-galactosidase can be measured by the rate of appearance of yellow color using a spectrophotometer. We found that ACDP4 strongly interacts with both COX11 and RBM30 (Fig. 2). On the other hand, there was no β-galactosidase activity detected in the controls, suggesting that the interactions detected are true. The interactions between ACDP4 and the metal ion chaperone COX11 provided an initial clue for a potential role of ACDP4 in metal ion homeostasis.
Figure 2 Interactions of ACDP4 with COX11 or with RBM30 in a yeast two-hybrid system. Cultures were assayed for β-galactosidase to verify two-hybrid interactions. β-galactosidase secreted from yeast containing the lacZ reporter gene after GAL4 activation were detected in the presence of chromogenic substrate ONPG (ortho-nitrophenyl β-D-galactopyranoside). Vector, empty pGADT7 vector.
The effects of ectopic expression of ACDP4 on metal ion toxicity in HEK293 cells
Essential metal ions can produce oxidative toxicity when excessive amounts are accumulated inside cells. We first determined metal ion toxicity using HEK293 cells that were not transfected with ACDP4 plasmids. This serves as baseline of metal ion toxicity. Five divalent ions including Cu2+, Mn2+, Co2+, Mg2+ and Zn2+ were selected for the study. To establish a killing curve for each of the selected metal ions, HEK293 cells were plated at 3 × 105 per well in 6-well cultural plates and then cultured in medium containing different concentrations of CuCl2, MnCl2, CoCl2, MgCl2 and ZnCl2 for 48 hrs, respectively. The cells were then examined for viability by trypan blue staining. As shown in Fig. 3, the concentrations responsible for 50% of cell death (EC50) for CuCl2, MnCl2, CoCl2 and ZnCl2 were 0.9 mM, 1.4 mM, 0.8 mM and 0.6 mM, respectively (Fig. 3A). On the other hand, MgCl2, showed much less cell toxicity at low minimolar concentrations (Fig. 3B).
Figure 3 Killing curve of HEK293 cells for selected divalent ions A. Killing curves for Cu2+, Mn2+, Co2+ and Zn2+. B. Killing curve for Mg2+. Cell death was estimated by counting a total of 300 cells in each field with 2–3 fields under microscope. Apparent EC50 values were 0.9 mM for Cu2+, 1.4 mM for Mn2+, 0.8 mM for Co2+, 0.6 mM for Zn2+ and 64 mM for Mg2+.
Next, we tested whether ectopic ACDP4 expression may enhance metal ion toxicity in HEK293 cells. pcDNA3.1-ACDP4 plasmid was transfected into HEK293 cells using the Effectene Transfection Reagent. Transfection of an empty vector (pcDNA3.1) was used as a control. The cells were then cultured in medium containing the above metal ions each at its apparent EC50 identified above, and metal toxicity was measured at 48 hrs after the addition of metal ions. Ectopic ACDP4 expression significantly increased metal ion toxicity for cells cultured with Cu2+, Mn2+ and Co2+ (Fig. 4). The cell viabilities with ectopic ACDP4 in medium containing Cu2+, Mn2+ and Co2+ was 12% (P < 0.04), 14% (P < 0.02) and 8% (P = 0.05) lower than that of the cells transfected with empty vectors, respectively (Fig. 4). We did not observe any significant difference of cell viabilities between cells with ectopic ACDP4 and control cells when Mg2+ or Zn2+ were tested (Fig. 4). These results suggest that ACDP4 is probably preferential for Mn2+, Cu2+ and Co2+ over Mg2+ and Zn2+.
Figure 4 The effects of ectopic expression of ACDP4 on metal ion toxicity in HEK293 cells. ACDP4 is ectopically expressed in HEK293 cells using pcDNA3.1-ACDP4 plasmid transfection. Transfection of an empty vector (pcDNA3.1) was used as a control. Cells were cultured for 48 hrs in medium containing the 5 type of metal ions, and each type of metal ions was at its concentration of the apparent EC50 established in Figure 3. For normalization, the viability of control cells was scaled to 100%, and cell viabilities for cells transfected with ACDP4 were normalized in reference to the control cells. Data represent mean ± SEM, * P < 0.05, Student-t test.
Enhanced metal ion toxicity by ectopic co-expression of ACDP4 and COX11
While we showed physical interaction between ACDP4 and COX11 in a yeast two-hybrid system, it is more important to know whether ACDP4 may be functionally coupled with COX11 in mammalian cells. To address this question, we investigated metal ion toxicity in cells ectopically co-expressing ACDP4 and COX11. For this purpose, HEK293 cells were co-transfected with ACDP4-pcDNA3.1 plasmid along with COX11-pIRES-eGFP plasmid. The transfected cells were cultured in medium with the five metal ions at their apparent EC50 concentrations identified in Figure 3, and metal ion toxicity was examined after 48 hrs. Co-transfection of ACDP4 and COX11 significantly increased HEK293 cell toxicity to Cu2+, Mn2+ and Co2+ (Fig. 5A) but not to Zn2+ and Mg2+ (not shown). The cell viabilities in medium with Cu2+, Mn2+ and Co2+ were 32% (P < 0.006), 34% (P < 0.001) and 23% (P < 0.002) lower than the control cells, respectively (Fig. 5A). The cell death for Cu2+, Mn2+ and Co2+ were increased by 2.8 fold, 2.5 fold 2.9 fold, respectively, compared to transfection of ACDP4 alone (Fig. 4).
Figure 5 Functional coupling of ACDP4 with COX11 in metal ion toxicity A. The effects of ectopic co-expression of ACDP4 and COX11 on metal ion toxicity in HEK293 cells. Co-expression of ACDP4 and COX11 significantly enhanced metal ion toxicity. B. Metal ion toxicity for HEK293 cells ectopically expressed COX11 alone. There were no significant differences in cell viability between control cells and cells ectopically expressed COX11 alone.
While tranfection of ACDP4 increased cell toxicity to Cu2+, Mn2+ and Co2+ (Fig. 4) and co-tranfection of ACDP4 with COX11 further enhanced cell toxicity to these ions (Figure 5A), transfection of COX11 alone had no significant effect on baseline metal ion toxicity (Figure 5B). These results together suggest a functional coupling between ACDP4 and COX11 for metal ion toxicity in HEK239 cells.
Discussion
In the present study, we have demonstrated that ACDP4 physically interacts with COX11 and functionally coupled with this metal ion chaperone. We have shown that ectopic expression of ACDP4 enhances cell toxicity to several essential metal ions including Cu2+, Mn2+ and Co2+, and that co-expression of ACDP4 with COX11 produces more pronounced increases of metal ion toxicity. This is the first study to explore potential functions of a member of ACDP protein family. The physical interaction and functional coupling with a metal ion chaperone indicates that ACDP4 is involved in the homeostasis and toxicity of essential metal ions.
The ACDP gene family is highly conserved in both human and mouse [23,24]. However, it has been complete unknown for their potential functions before the present study. Nevertheless, the sequence conservation and the presence of multiple members within a species imply a functional importance associated with this gene family. While we were completing this work, Yang et al. reported the involvement of a yeast ACDP homolog, Mam3p, in manganese homeostasis and toxicity in yeast [26]. Their results demonstrated that Mam3p operates independently to the well-established manganese trafficking pathways in yeast involving the manganese transporters, Pmrlp, Smf2p and pho84p [26]. In our present study, we have demonstrated the involvement of ACDP4 in copper and cobalt toxicity in addition to manganese. Furthermore, our functional results were obtained from mammalian cells.
From our functional study using metal ion toxicity as a measure, it appears that ACDP4 functions at a site upstream to COX11. This idea is supported by three lines of evidence. First, ectopic expression of ACDP4 alone could enhance metal ion toxicity. Second, ectopic expression of COX11 alone did not affect metal ion toxicity. Third, co-expression of ACDP4 with COX11 produced more pronounced metal ion toxicity than ectopic expression of ACDP4 alone. Interactions between ACDP4 and COX11 suggest that ACDP4 may be a metal ion transporter in mammalian cells (Fig. 6). This is because a metal ion chaperone normally interacts with upstream metal ion transporters to receive metal ions and interacts with downstream target proteins to deliver metal ions. However, a direct study on ACDP4-mediated metal ion accumulation in cells is needed to confirm transport functions of ACDP4. Alternatively, ACDP4 may be an intermediate metal ion chaperone upstream to COX11. Detailed studies in the future on ACDP4 membrane localization and its interactions with metal ion chaperones in mammalian cells will provide more insights into the mechanisms by which ACDP4 is involved in metal ion homeostasis and toxicity. It would also be interesting to study whether other members of ACDP family may be involved in essential metal ion homeostasis and toxicity in mammalian cells. A particular interesting issue is the potential neuronal functions of ACDP1 since it is almost exclusively expressed in CNS neurons in the brain (24) and the spinal cord neurons (Fig. 1).
Figure 6 A postulated model of ACDP-mediated metal ion delivery in mammalian cells. An essential metal ion is first transferred through plasma membrane via an ACDP. An intracellular chaperone specifically interacts with the intracellular parts of ACDP. The chaperone receives the metal ion from ACDP and then carries the ion to a specific target protein. Overload of cells with essential metals results in oxidative cell death.
Interactions between ACDP4 and COX11 provide a structural and functional linkage to COX11. COX11 is an intracellular copper chaperone originally identified in yeast as essential for cytochrome c oxidase activity and heme α stability in subunit I. [20]. A recent study suggested a role for this protein in formation of the binuclear copper-heme center [21]. Spectroscopic and mutagenesis studies indicate that the copper ion in COX11 is ligated by three conserved cysteine residues [22]. The C-terminal domain of this protein forms a dimer that coordinates a single CuC per monomer. While COX11 was found to be a dimer, it remains possible that copper transfer to and from COX11 occurs through heterodimeric interaction with other proteins [22]. If ACDP4 is confirmed to be a metal ion transporter or an upstream metal ion chaperone, it would be interesting to know how ACDP4 delivers metal ions to COX11, whether ACDP4 and COX11 are co-localized on or near the plasma membranes, whether ACDP4 and COX11 are also co-localized on the membranes of other intracellular organelles such as mitochondria. We have shown that ACDP4 also interacts with RBM30, an RNA binding motif protein 30. RBM30 contains a zinc knuckle, raising a possibility that it may serve as an intracellular zinc chaperone. However, we did not observe an enhanced toxicity to zinc in cells transfected with RBM30 or co-transfected with both ACDP4 and RBM30. Therefore, it remains to determine whether interaction of ACDP4 with RBM30 may have any biological significance.
The expression of ACDP4 and ACDP1 on dorsal horn neurons of the spinal cord shown in the present study may have implications in sensory physiology and pathology. Both ACDP4-ir and ACDP1-ir were found to be localized close to the plasma membranes of dorsal horn neurons. If ACDP4 and ACDP1 are metal ion transporters or chaperones in these neurons, they should be involved in regulating homeostasis of essential metal ions in the spinal cord dorsal horn neurons. Essential metal ions are essential for activity of many enzymes, including those for synthesis of neuropeptides (e.g. substance P) and neurotransmitters (e.g. monoamine) in the spinal cord [9], and those for superoxide metabolism in peripheral sensory nerves and CNS sensory neurons in the spinal cord [10,11]. Abnormal superoxide metabolism has been shown to be a critical factor in inflammation and pathological pain conditions [10,11]. It is conceivable that through regulating essential metal ion homeostasis, ACDP4 and perhaps also other ACDP members can affect sensory process including nociception. It is also predictable that the functions of ACDP family will go beyond somatosensory system.
Methods
Neuronal cell preparation and immuostaining
Sprague-Dawley rats were used according to the Institutional Animal Care and Use Committee guideline of the University of Florida. Dorsal horn neuron cultures were prepared as described previously. [27]. In brief, spinal cord dorsal horns were dissected out from rat embryos at the age of 16 days in utero (E16). Dorsal horns were incubated separately for 25 min at 37°C in S-MEM medium (Gibco, Grand Island, NY) with 2.5% trypsin (Gibco) and then triturated to dissociate neurons. The neurons were plated on glass coverslips previously prepared with a monolayer of rat cortical astrocytes. Neurons were maintained in MEM (Gibco) culture medium that contained 5% heat-inactivated horse serum (JRH Biosciences, Lenexa, KS), uridine/5-flouro-2'-deoxyuridine (10 μM; Sigma, St. Louis, MO), 8 mg/ml glucose and 1% vitamin solution (Gibco). The cultures were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO2, and were fed weekly with fresh culture medium.
Neurons were used for immunostaining of ACDP1 and ACDP4 at two weeks in culture. For immunostaining, neuronal cells on the coverslips were first fixed in PBS containing 4% paraformaldehyde (PFA) for 12 hrs at 4°C and then incubated in a solution containing 4% PFA and 0.4% Triton X-100 at 4°C for 1 hr. After washing with PBS three times, the cells were incubated with a blocking solution containing 1:30 normal goat serum, and subsequently incubated with a rabbit polyclonal anti-ACDP antibody (1:3000) overnight at 4°C. After extensive washing with 1% goat serum PBS solution, the cells were incubated with an Alex 488 conjugated secondary antibody (1:100 in 1% goat serum PBS solution, Molecular Probes) for 3 hrs at room temperature. Following final washes with 1% goat serum PBS solution, the neuronal cells on the coverslips were cover-slipped with a glycerol-based anti-photobleach medium. The cells were viewed under a fluorescence microscope (Olympus) with a 40X oil-immersion objective or a confocal fluorescence microscope (Carl Zeiss) with a 60X objective.
Establishment of killing curve
HEK293 cells were plated at 3 × 105 per well of 6 well plates. The cells were cultured for 48 hrs in medium containing different concentrations of MgCl2, CuCl2, ZnCl2, MnCl2 and ECoCl2, respectively. Dead cells were then stained with trypan blue solution at a ratio of 4:6 (Gibco). The percentage of cell death was determined by counting a total of 300 cells in each field with 2–3 fields under microscope.
Plasmid construction
Full-length ACDP4 gene was cloned into the pGBKT7 and pcDNA3.1 vectors using the EcoR I cutting site. COX11 and RBM30 were amplified from fetal brain cDNA and then cloned into pIRES-eGFP vector with EcoR I and BamH I, and Xho I and BamH I cutting sites, respectively.
Yeast two-hybrid analysis
The Matchmaker Galt4 two-hybrid system 3 kit (Clontech) was used for two-hybrid analyses. The ACDP4 coding sequence was PCR engineered and cloned into the pGBKT7 vector, which was used as a bait to screen a pretransformed human fetal brain Matchmaker cDNA library at high stringency culture condition. To confirm the interaction between ACDP4 and COX11 or RBM30, the full-length COX11 and RBM30 cDNA were cloned into the pGADT7 vector (as targets), which was then co-transfected into an AH109 yeast strain along with the pGBKT7-ACDP4 plasmid, respectively. An empty pGADT7 vector was used as a control. The cultures were assayed for β-galactosidase to verify two-hybrid interaction according to the manufacturer's instruction.
Cell culture and transfection
HEK293 cells were cultured in Dulbecco's modified Eagles's medium (Mediatech, Inc.) supplemented with 10% fetal bovine serum and 100 units/ml antibiotic-antimycotic (Invitrogen). Transfections were carried out using 2 ug of plasmid DNA/100-mm dish and Effectene Transfection Reagent according to the instructions of the manufacturer (Qiagen).
Metal toxicity assay
We used an in vitro toxicology assay kit (Sigma) for measurement of metal toxicity according to the manufacturer's instruction. Cell viability estimated by this in vitro toxicology assay determines cell number spectrophotometrically as a function of mitochondrial activity in living cells. Briefly, the cells were cultured with indicated metal ion for 44 hrs and then supplemented with 20 μl MTT per well. The cells were cultured for additional 4 hrs. Subsequently, the cultural plates were spun for 5 min at 1000 rpm to remove the supernatant. The plates were dried in an incubator for about 4 hrs. 200 μl of MTT solubilization solution was then used to dissolve the resulting formazan crystals. The results were read at 570 nm after 6 hrs incubation with a Synergy HT plate reader (Bio-Tek).
Data analysis and statistics
Unless otherwise indicated, data represent mean ± SEM, p < 0.05, student-t test.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
We thank Drs. Junyan Han and Jinlei Xi and Daniel Eisenman for their help for preparation of figures of the manuscript. This work was supported by the Juvenile Diabetes Research Foundation International (1-2004-235), the American Diabetes Association (1-05-JF-47) to CYW and a National Institute of Health Grant NS38254 to JGG.
==== Refs
Nelson N Metal ion transporters and homeostasis EMBO J 1999 18 4361 4371 10449402 10.1093/emboj/18.16.4361
Bush AI Metals and neuroscience Curr Opin Chem Biol 2000 4 184 191 10742195 10.1016/S1367-5931(99)00073-3
HaMai D Bondy SC Oxidative basis of manganese neurotoxicity Ann NY Acad Sci 2004 1012 129 141 15105260 10.1196/annals.1306.010
Mattson MP Metal-catalyzed disruption of membrane protein and lipid signaling in the pathogenesis of neurodegenerative disorders Ann NY Acad Sci 2004 1012 37 50 15105254 10.1196/annals.1306.004
Normandin L Ann BL Salehi F Manganese distribution in the brain and neurobehavioral changes following inhalation exposure of rats to three chemical forms of manganese Neurotoxicology 2004 25 433 441 15019306 10.1016/j.neuro.2003.10.001
Uversky VN Li J Fink AL Metal-triggered structural transformations, aggregation, and fibrillation of human alpha-synuclein. A possible molecular NK between Parkinson's disease and heavy metal exposure J Biol Chem 2001 276 44284 44296 11553618 10.1074/jbc.M105343200
Yoshida S Ektessabi A Fujisawa S XANES spectroscopy of a single neuron from a patient with Parkinson's disease J Synchrotron Radiat 2001 8 998 1000 11513007 10.1107/S0909049500017726
Sziraki I Mohanakumar KP Rauhala P Kim HG Yeh KJ Chiueh CC Manganese: a transition metal protects nigrostriatal neurons from oxidative stress in the iron-induced animal model of parkinsonism Neuroscience 1998 85 1101 1111 9681949 10.1016/S0306-4522(97)00660-X
Marchand JE Hershman K Kumar MS Thompson ML Kream RM Disulfiram administration affects substance P-like immunoreactive and monoaminergic neural systems in rodent brain J Biol Chem 1990 265 264 73 1688429
Wang ZQ Porreca F Cuzzocrea S Galen K Lightfoot R Masini E Muscoli C Mollace V Ndengele M Ischiropoulos H Salvemini D A newly identified role for superoxide in inflammatory pain J Pharmacol Exp Ther 2004 309 869 78 14988418 10.1124/jpet.103.064154
Muscoli C Mollace V Wheatley J Masini E Ndengele M Wang ZQ Salvemini D Superoxide-mediated nitration of spinal manganese superoxide dismutase: a novel pathway in N-methyl-D-aspartate-mediated hyperalgesia Pain 2004 111 96 103 15327813 10.1016/j.pain.2004.06.004
Eide DJ Metal ion transport in eukaryotic microorganisms: insights from Saccharomyces cerevisiae Adv Microb Physiol 2000 43 1 38 10907553
Li H Li F Kwan M He QY Sun H NMR structures and orientation of the fourth transmembrane domain of the rat divalent metal transporter (DMT1) with G185D mutation in SDS micelles Biopolymers 2005 77 173 183 15660380 10.1002/bip.20204
Agranoff D Collins L Kehres D Harrison T Maguire M Krishna S The Nramp orthologue of Cryptococcus neoformans is a pH-dependent transporter of manganese, iron, cobalt and nickel Biochem J 2005 385 225 232 15350193 10.1042/BJ20040836
Wang F Kim BE Petris MJ Eide DJ The mammalian Zip5 protein is a zinc transporter that localizes to the basolateral surface of polarized cells J Biol Chem 2004 279 51433 51441 15322118 10.1074/jbc.M408361200
Lopez-Millan AF Ellis DR Grusak MA Identification and characterization of several new members of the ZIP family of metal ion transporters in Medicago truncatula Plant Mol Biol 2004 54 583 596 15316291 10.1023/B:PLAN.0000038271.96019.aa
Eide DJ The SLC39 family of metal ion transporters Pflugers Arch 2004 447 796 800 12748861 10.1007/s00424-003-1074-3
Paulsen IT Saier MH Jr A novel family of ubiquitous heavy metal ion transport proteins J Membr Biol 1997 156 99 103 9075641 10.1007/s002329900192
Aarts M Iihara K Wei WL Xiong ZG Arundine M Cerwinski W MacDonald JF Tymianski M A key role for TRPM7 channels in anoxic neuronal death Cell 2003 115 863 77 14697204 10.1016/S0092-8674(03)01017-1
Tzagoloff A Capitanio N Nobrega MP Gatti D Cytochrome oxidase assembly in yeast requires the product of COX11, a homolog of the P. denitrificans protein encoded by ORF3 EMBO J 1990 9 2759 2764 2167832
Hiser L Di Valentin M Hamer AG Hosler JP Cox11p is required for stable formation of the Cu(B) and magnesium centers of cytochrome c oxidase J Biol Chem 2000 275 619 623 10617659 10.1074/jbc.275.1.619
Hamza I Gitlin JD Copper chaperones for cytochrome c oxidase and human disease J Bioenerg Biomembr 2002 34 381 388 12539965 10.1023/A:1021254104012
Wang CY Shi JD Yang P Molecular cloning and characterization of a novel gene family of four ancient conserved domain proteins (ACDP) Gene 2003 306 37 44 12657465 10.1016/S0378-1119(02)01210-6
Wang CY Yang P Shi JD Molecular cloning and characterization of the mouse Acdp gene family BMC Genomics 2004 5 7 14723793 10.1186/1471-2164-5-7
Kehres DG Lawyer CH Maguire ME The CorA magnesium transporter gene family Microb Comp Genomics 1998 3 151 169 9775386
Yang M Jensen LT Gardner AJ Culotta VC Manganese toxicity and Saccharomyces cerevisiae Mam3p, a member of the ACDP (Ancient Conserved Domain Protein) family of proteins Biochem J 2004 386 479 487 15498024
Gu JG Albuquerque C Lee CJ MacDermott AB Synaptic strengthening through activation of Ca2+-permeable AMPA receptors Nature 1996 381 793 796 8657283 10.1038/381793a0
| 15840172 | PMC1097757 | CC BY | 2021-01-04 16:40:08 | no | Mol Pain. 2005 Apr 19; 1:15 | utf-8 | Mol Pain | 2,005 | 10.1186/1744-8069-1-15 | oa_comm |
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-151585751610.1186/1477-7827-3-15ResearchDiabetes (db/db) mutation-induced endometrial epithelial lipoapoptosis: Ultrastructural and cytochemical analysis of reproductive tract atrophy Garris David R [email protected] Division of Cell Biology and Biophysics, School of Biological Sciences, University of Missouri-Kansas City, Kansas City, Missouri 64110 USA2005 27 4 2005 3 15 15 8 3 2005 27 4 2005 Copyright © 2005 Garris; licensee BioMed Central Ltd.2005Garris; 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 diabetes (db/db) mutation in C57BL/KsJ mice promotes a progressive cytolipidemia within the endometrial epithelial (EE) layer of the female reproductive tract which results in premature cellular and organ atrophy. The current studies focus on the ultrastructural and cytochemical changes which promote nuclear apoptosis and cytostructural disruption following the expression of endometrial hypercytolipidemia which promotes diabetes-associated organoinvolution and manifest infertility.
Methods
Control (normal:+/+) and diabetes (db/db) genotype groups were prepared for high resolution light microscopic analysis of cytolipidemia and nuclear apoptosis (TUNEL-labeled 3'-DNA fragmentation) indices and compared to the transmission electron (TEM) microscopic analysis of endometrial tissue samples collected from 8–16 week-old groups.
Results
Compared to controls, db/db mutation expression induced a dramatic increase in EE cytolipid vacuole volume and density within the epithelial endometrial layer. TEM analysis revealed that cytolipid vacuole accumulations initially aggregated at the baso-polar regions of UEE cells in response to the systemic hyperglycemic/hypertriglyceridemic conditions which characterized the (db/db) groups. Progressive cytoplasmic movement of the lipid pools into perinuclear compartments of affected EE cells induced nuclear isolation from organelles that were displaced towards peripheral cytoplasmic compartments. Cytochemical analysis of lipid vacuole accumulations indicated attraction towards, and incorporation within, the nuclear envelope of hyperlipidemic cells. Co-localization of nuclear apoptotic 3'-DNA fragments within identified hyperlipidemic EE cells was coincident with the cytochemical and ultrastructural identification of lipid penetration through the nuclear envelope in db/db mutants.
Conclusion
These results are the first cytochemical indication that the metabolic disturbances in db/db mutants which promote hypercytolipidemia are coincident with lipoapoptosis-induced nuclear dissolution, as denoted by DNA fragmentation analysis. The lipidemia-induced alterations in intracellular organelle and nuclear architectures suggests that the metabolic disturbances in glucose and lipid metabolic cascades in diabetes (db/db) mutants disrupts cytointegrity, culminating in nuclear disregulation (as indicated by lipoapoptosis) and eventual premature reproductive tract organoinvolution and resultant, manifest, reproductive sterility.
==== Body
Background
The cytoarchitecture of the female reproductive tract is severely compromised by the deleterious influences of diabetes-induced alterations in utero-ovarian cellular glucometabolism [1-3]. In humans [4-6] and experimental models [7-11], diabetes-associated alterations in uterine endometrial metabolism and structure have been associated with pronounced hypercytolipidemia, a hyper-caloric metabolic response that induces cellular hyperlipidemia and subsequent promotion of premature reproductive tract involution [2,3,12-16]. The resulting reproductive incompetence is characterized by reproductive acyclicity [13,14], compromised ovarian follicular development [14], depressed ovarian steroid hormone synthesis [17], depressed sensitivity and responsivity to endocrine stimulated cellular metabolism [18-20] and enhanced utero-epithelial atrophy [2]. The affected endometrial architecture is characterized by an enormous increase in intra-and inter-cellular lipid depositions [2,13], resulting from the interstitial perivascular escape and imbibition of elevated systemic triglyceride and free fatty acid moities [21,22] which characterize the overt diabetes (Type 2) metabolic (X) syndrome [23,24]. Ultimately, exposure to the chronic influences of the non-homeostatic metabolic condition induces a lipoatrophy syndrome [12-14], characterized by the progressive accumulation of cytolipid inclusions [2], organelle dissolution [2], nuclear compartment isolation [15], suppressed cellular oxidative metabolism [13], and cyto-atrophy [9,13,14,17]. Recent reports have indicated that the expression of the diabetes (db/db) mutation in C57BL/KsJ mice compromises reproductive tract maturation by promoting hypercytolipidemia within the endometrial epithelial (EE) layer [2] that is characterized by a progressive lipid-isolation of the cell nuclei from surrounding cytoplasmic organelle compartments [15]. The expanding endometrial cytolipid volume in db/db mutants has been associated with disrupted nuclear chromatin (DNA) structural integrity and pycnosis-associated degeneration [24]. However, the co-incident expression of metabolic hypercytolipidemia and structural nuclear dissolution, as indexed by 3'-DNA fragmentation [24] within apoptotic nuclei, remains to be demonstrated. The present studies were designed to evaluate the co-incident cytochemical and ultrastructural alterations which promote premature, progressive lipoapoptotic nuclear degeneration within the endometrial epithelial tissue layer of the obese, hyperglycemic, hyperlipidemic and hypogonadal (infertile) db/db-mutant reproductive tract.
Materials and methods
Animals
Adult, female C57BL/KsJ mice (Jackson Laboratory, Bar Harbor, ME), between 8 and 16 weeks of age, denoting the overt and chronic phases of the Type 2 diabetes syndrome [14], were used in these studies and maintained in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals (NIH publication no. 80-23). Littermate controls (+/+) and diabetes (db/db)-mutant genotypes, were pair matched for phenotype, tissue sampling and blood glucose concentration comparisons during the course of these studies. All mice were housed five per cage, grouped according to genotype, under controlled environmental conditions (23°C), with an established photoperiod of 12 hr light/day (lights on: 0600 h) [13,14]. Blood glucose levels (Ames Glucometer method), serum triglyceride concentrations (Sigma, St. Louis) and body weights were monitored for each of the 8 to 16-week-old age groups as previously described [13,14]. Animals exhibiting both obesity (≥25 grams) and pronounced hyperglycemia (≥200 mg/dl) and hyperlipidemia (≥ 200 mg/ml serum triglycerides) relative to controls (≤150 mg/dl and mg/ml, respectively) were considered as overt, obese-diabetics [16,23], with the comparative expression of these indices noted relative to control or genotypic mutation (Table I) groups throughout the experimental period.
Table 1 Phenotype, Uterine Tissue Biomass and Glycemia Indicators of Diabetes (db/db) Mutation-Induced Alterations in C57BL/KsJ Mice.
Index N Groups P < :
Control (+/+) Diabetes (db/db)
Body Weight (g) 5 21 ± 3 47 ± 5 0.001
Blood Glucose (mg/dl) 5 103 ± 8 428 ± 16 0.001
Uterine Weight (mg) 5 44 ± 3 13 ± 4 0.01
Serum Triglycerides (mg/ml) 5 143 ± 8 312 ± 24 0.001
All values for the indicated parameters are represented as group (N) means (± SEM) for control (+/+) and diabetes (db/db)-mutant C57BL/KsJ mice, with statistical intergroup differences (P ≤) indicated.
Tissue Collection and Preparation
Uterine endometrial tissue samples from each group of control (+/+) and diabetic (db/db) matched-paired genotypes were collected, weighed and prepared for high resolution light microscopy (HRLM), cytochemical analysis of cytoplasmic lipid depositions and transmission electron microscopic (TEM) examination as previously described [2]. In brief, mice were anesthetized at 8 (i.e. overt phase) or 16 weeks (i.e., the chronic phase of Type 2 syndrome expression and reproductive tract compromise) [14] with sodium pentobarbital and systemically perfused with 50 ml of physiological saline and 100 ml of Karnovsky's fixative solution. Collected mid-cornua uterine tissue samples were cleaned, blotted, blocked and embedded in either paraffin or plastic using conventional techniques [2]. All tissue samples were subsequently sectioned and stained with a toluidine blue-basic fuchsin mixture for polychromatic identification [24,25] of cellular lipid pools by HRLM or with osmium tetroxide [2] prior to examination by TEM.
High Resolution, Digital and TEM, Hypercytolipidemia Analysis
Tissue sections prepared for light microscopic analysis were used for polychromatic organelle differentiation, the localization of intracellular lipid inclusion accumulations, and the determination of cytoplasmic changes associated with the progressive expression of the db/db mutation as previously described [25]. Photographic images of uterine tissue compartments and epithelial cell populations from the prepared tissue samples were captured with an Olympus (Olympus Optical, Tokyo, Japan) digital graphics camera and microscope unit, with lipid vacuole pools digitally enhanced utilizing polychromatic stain identification, and digital-color scale conversion for chemical-specific triglyceride localization analysis [25]. Tissue sections prepared for TEM analysis from the same groups were analyzed for structural variations in cytoplasmic changes in organelle and lipid inclusion density, as well as for uterine basal lamina and peri-nuclear changes induced by the expressed hypercytolipidemia associated with the expression (db/db) mutation [2].
Localization and Analysis of db/db-Associated Nuclear Lipoapoptosis
Uterine samples from the designated groups were collected and rapidly frozen for cryostat (-20°C) sectioning then subsequently prepared for TUNEL (FD NeuroTechnologies; Ellicott City, MD) labeled apoptotic 3'-DNA fragmentation analysis, a recognized chemical marker associated with nuclear chromatin dissolution, as previously described [20]. Slides were placed in 0.1 M phosphate buffer (PBS: pH 7.4) containing 4% (v/v) paraformaldehyde for 30 minutes. Tissue sections were subsequently washed (x2 rinses @ 5 minutes each) in 0.01M PBS, then fixed in pre-cooled (-20°C) ethanol:acetic acid (2:1 v/v) for 5 minutes, washed (x2) in PBS for 10 minutes to assure proper tissue preservation prior to preparation for TUNEL labeling. Detection of free 3'-hydroxyl terminus DNA fragments was performed using the in situ TdT-mediated dUTP-biotin nick end labeling (TUNEL) technique. Terminal deoxynucleotidyl transferase was utilized for the catalyzed incorporation of biotinylated deoxyuridines onto the exposed 3'-hydroxyl termini of DNA fragments which labeled apoptotic cells. The integrated biotins were enhanced and visualized as dense, localized avidin-biotin-complexes identifiable by HRLM. All TUNEL labeled endometrial samples were subsequently counterstained for polychromatic cytostructural analysis [25], allowing for the identification of hypercytolipidemia within the same cells labeled for nuclear apoptosis. Intracellular nuclear or cytoplasmic organelle (mitochondrial) TUNEL-label specificity was evaluated prior to cytochemical analysis of co-localized perinuclear lipid vacuole density profiles and 3'-DNA fragments, indicative of apoptotic cytodegenerative alterations, in affected endometrial epithelial cells exhibiting hypercytolipidemia.
Statistical Analysis
Values for body weights and blood glucose concentrations were expressed as group means (± SEM) for the designated genotype groups. Intergroup differences were determined using the Student's T-test and Analysis of Variance exams, with a p ≤ 0.05 accepted as representing statistical intergroup measurement differences.
Results
Cytochemical and Ultrastructural Analysis of Endometrial Epithelial Hypercytolipidemia
The changes in body weights, uterine weights and blood glucose concentrations in C57BL/KsJ mice resulting from the expression of the diabetes (db/db) mutation are indicated in Table I. Dramatic increases in phenotypic obesity and associated hyperglycemic conditions characterized (db/db) groups relative to (+/+) indices. In contrast, uterine weights in the db/db-mutant group decreased (Table I) in association with the progressive endometrial hypercytolipidemia and atrophy which characterized the uterine samples prepared for HRLM and TEM analysis.
The accumulation and retention of EE cell cytolipid stores by (+/+) tissues was found to be restricted to the baso-polar regions of all control samples examined by HRLM (Figure 1A, C) and TEM (Figure 2A) analysis between 8 and 16 weeks of age. The chemical-specific localization of basal cytoplasmic triglyceride deposits characterized +/+ endometrial epithelial cells, in which cytoplasmic organelle distribution and organization were indicative of a viable cytoarchitecture (Figure 2A). In contrast, the polychromatic identification of hypercytolipidemic depositions in (db/db) tissue samples was demonstrated by the cytoarchitectural alterations noted by both HRLM cytochemical (Figure 1B, D) and TEM (Figure 2B) analysis. Characteristic of all 8 – 16 week old db/db-mutant tissue samples (Figures 2B; 3) was the enormous increase in cytoplasmic triglyceride pools which were distributed in a prominent gradient pattern between basal, perinuclear and apical cytoplasmic regions (Figures 2B,3). TEM analysis of endometrial epithelial samples from db/db groups indicated that the basal pole cytoplasmic compartments contained dense, expanded lipid vacuole accumulations which progressively migrated to surround and occupy the perinuclear space of affected cells (Figure 3). The progressive changes in nuclear envelope configurations (i.e., convoluted membrane, pycnotic distortions) induced by the perinuclear lipid pool expansions included dramatic invaginations of the external nuclear membrane, chromatin clumping along the inner nuclear membrane and physical disruption of the envelope integrity in regions where lipid moieties were observed to contact and penetrate the outer nuclear lamina (Figures 2B, 3, 4A) as demonstrated by TEM analysis. The nuclear membrane and chromatin disturbances associated with the pronounced perinuclear lipid accumulations were further characterized by a prominent expansion of the perinuclear space surrounding each affected cell (Figure 3), promoting a physical isolation of the nuclear compartment from the cytoplasmic organelles which were peripherally displaced by the expanding cytoplasmic lipid pools.
Figure 1 Photomicrographic (x400) comparisons of representative control (+/+:A) and diabetes (db/db)-mutant (B) uterine endometrial epithelial tissue layers by HRLM analysis, depicting the normal, sub-nuclear (n: arrow line) basal pole lipid (Lb) pools typical of +/+ groups located adjacent to the underlying stromal basal lamina (bl), as compared with the dramatic expansions of Lb and apical pole lipid (La) pools in db/db (B) groups. By digital enhancement of cytochemical lipid (triglyceride) pools in +/+ (C) and db/db (D) groups, the sub- and supra-nuclear cytoplasmic lipid pool expansions in db/db groups (D) were accentuated relative to the cytoplasmic volume and distribution patterns of lipid vacuoles in control specimens (C). (Technical Note: Enzymatic incubation of tissue sections, as described in the Materials & Methods section, for cytochemical and TUNEL-labeling is responsible for a moderate reduction in image resolution presented in Figures 1, 4, and 5, represents an intrinsic technical limitation associated with co-localization analysis of hyperlipidemia and apoptosis indices within tissue preparations.)
Figure 2 Ultrastructural (× 7350) analysis of control (A: +/+) endometrial epithelial tissue cells indicated prominent central nuclei (n) supported by a well-defined smooth stromal basal lamina (bl) and a cytoplasm possessing a rich organelle population including abundant mitochondrial (m) and prominent rough endoplasmic reticulum (er) compartments in the basal pole regions, as well as a well-developed Golgi vesicular apparatus (gv) and surface ciliary (cl) arrangement in the apical zones. In contrast, the endometrial epithelial cell layers of diabetes (db/db) mutant groups (B) were characterized by a convoluted bl membrane and a sub-nuclear (n) region dominated by enormous concentrations of basal pole lipid (Lb) vacuoles. Intracytoplasmic migration of the lipid into the perinuclear compartment (Ln), as well as a prominent reduction in basal cytoplasmic organelle populations, occurred in conjunction with nuclear envelope convolutions, with internal nuclear membrane invaginations associated with Ln depositions (arrows). Apical pole lipid (La) vacuole densities were prominent in db/db epithelial tissue samples, and were associated with expanded Golgi vesicular (gv) cisterns and blunted apical ciliary (cl) arrays as compared with +/+ structural (A) indices.
Figure 3 Transmission electron microscopic (× 6600) analysis of the progressive alterations induced within the endometrial epithelial cell layers of db/db mutants relative to expanding intracytoplasmic lipid pools in the basal (Lb) and perinuclear (Ln) compartments of affected cells. The prominent convolutions of the basal lamina (bl) of cells demonstrating enormous Lb accumulations were coincident with the recognized expansion of the perinuclear space (pns) in regions associated with Ln contact, or approximation with, the external nuclear envelope (arrows). The prominent nuclear membrane contact by Ln vacuoles occurred along all nuclear membrane planes, often in association with membrane invagination into the nuclear (n) compartment.
Figure 4 Ultrastructural (A: ×3750) and cytochemical (B-C: ×400) analysis of endometrial hypercytolipidemia depositions within epithelial cells of db/db mice localized within the basal (Lb) and perinuclear (Ln) compartments of affected cells. Digital enhancement (B) of cytoplasmic lipid vacuole accumulations (l) below the nuclear (n) layer, and within the perinuclear space (arrows), indicated intense TUNEL-indicated (black stain moieties) fragmentation of 3'-DNA components localized within the mitochondria-rich basal pole region as well as the nuclei (nT) of cells exhibiting lipoapoptosis. The cytochemical co-localization (C) of cytoplasmic lipid pools (yellow fluorescence) with TUNEL-labeled nuclear apoptotic DNA fragments (nT) within the same endometrial epithelial cell populations of db/db mutants indicated the coincident relationship between metabolic hypercytolipidemia and nuclear apoptosis-induced cytodissolution which resulted in uterine involution (Table I) and reproductive sterility.
Cytochemical and TUNEL-Label Analysis of Nuclear Lipoapoptosis
The influence of db/db-induced hypercytolipidemia on nuclear disruption was evaluated by the combined cytochemical localization of intracellular triglyceride depositions by computer-assisted, digital cytochemical analysis and the co-localization of 3'-DNA fragmentation by TUNEL-labeled counterstaining as an index of nuclear chromatin dissolution (apoptosis) events (Figure 4B–C). TUNEL-indexed nuclear apoptosis was localized in db/db UEE cells with co-incident hypercytolipidemic vacuolar expansion (Figure 4B). When subjected to triglyceride cytochemical analysis (Figures 4C, 5), nuclear TUNEL-label was co-localized within the cytolipid depositions present in both the mitochondrial-rich basal pole cytoplasmic compartment, as well as within the perinuclear and nuclear compartments of affected db/db cells (Figure 4C). Dense nuclear TUNEL-label was located within both the defined nuclear compartments of db/db cells (Figure 4C), and was prominent within the basal cell layer of proliferating epithelial tissue in which both TUNEL and hypercytolipidemic vacuole depositions were co-localized (Figure 5). In db/db EE samples demonstrating perinuclear lipid pool penetration of the affected nuclear envelope, TUNEL-indicated lipoapoptosis was a consistent index of cellular compromise promoted by triglyceride penetration and disruption of epithelial nuclear organization (Figure 5). The lipometabolic disruption of cytostructural organization was coincident with the co-localization of nuclear apoptosis (TUNEL-label) and hyperlipidemic, cytochemical indices, which characterized the premature EE cytoatrophic involution of the female reproductive tract associated with the overt expression of the type 2 (NIDDM) diabetes syndrome.
Figure 5 Photomicrograph (× 2000) of the db/db uterine endometrial epithelial (EE Layer) cell layer and underlying endometrial stroma (UT Endomet) demonstrating the coincident events of hypercytolipidemia (Lp: yellow fluorescent depositions) and TUNEL-labeled (T: black stain indicator) nuclear (n) co-localized within affected cells. The infiltration of the nuclear envelope by hyperlipidemic vacuole pools was denoted in epithelial cells that demonstrated ultrastructural (Figures 3-4) or cytochemical (Figure 4) indicators of apoptotic cytoatrophy.
Discussion & Conclusions
The current results demonstrate that the hyperlipidemic metabolic microenvironment induced by the expression of the diabetes (db/db)-mutation promotes a progressive lipoapoptotic cytoatrophy of EE tissue, events which contribute to premature organoinvolution of the female reproductive tract and manifest sterility in the C57BL/KsJ murine model of a gene-mutation linked, inherited, dysregulated metabolic syndrome [12,23]. The unique co-localization of dense cytolipid vacuole pools in affected cells experiencing TUNEL-indexed nuclear apoptosis and chromatin dissolution indicates that the hypercytoplasmic sequestration of extracellular lipids into the cytoplasmic perinuclear compartment compromises nuclear organization by the promotion of DNA fragmentation and subsequent nuclear degradation by lipo-infiltration (Figure 6). Ultimately, the progressive cytometabolic disruption of the endometrial layers compromises reproductive tract cytostructural and tissue integrity [2,3], as indicated by the reduced uterine biomass in db/db groups. Similar to diabetes- and obesity-associated reproductive complications in human clinical studies [4,6], the recognized alterations in both phenotypic and cytolipid (metabolic) indices induced an adipose-like cellular organization within the EE layer [2]. The resulting changes in intracellular organelle displacement towards peripheral cytoplasmic compartments, the blunting of apical EE ciliary and microvillus expressions [2], the altered chemical (i.e., hyperlipidemia) composition and the coincident nuclear apoptotic dissolution (Figure 6), correlated with the recognized functional compromise and premature organo-involution of the female reproductive tract [2] in db/db genotype mutants. The progressive intracytoplasmic trafficking of intracellular lipid pools from basal-to-perinuclear-to-apical cytoplasmic loci, has been recognized to be associated with both the duration and severity of the systemic metabolic aberrations in db/db-mutants [2,15]. These collective data suggest that the progressive lipid infiltration of the EE layer promotes the indicated structural, metabolic and lipoapoptotic disruption of nucleus (DNA)-directed transcriptional metabolic cascades that ultimately induce non-homeostatic cytoarchitectural changes in affected db/db cells which become incapable of supporting normal reproductive tract function (Figure 6). The progressive disruption of these structural indices and interdependent metabolic cascades culminates in the resultant, cumulative, cytoatrophic premature organoinvolution of the female reproductive tract and manifest sterility.
Figure 6 Diagrammatic representation of the cumulative influences of chronic diabetes-obesity syndrome influences on the progressive cytotransformation into hyperlipidemic cell types, eventually results in lipoidal infiltration into, and dissolution of, the nuclear chromatin (DNA) matrix continuity culminating in lipoapoptosis and premature cytoatrophy.
Of particular interest was the progressive perinuclear accumulation and infiltration of the nuclear compartment by db/db-mutation associated expansion of cytoplasmic lipid pools. By both HRLM and TEM analysis, the identification of lipid vacuole contact with, or infiltration through, the nuclear envelope was evidenced in cells exhibiting TUNEL-labeled lipoapoptosis. Progressively, the lipid migrations accumulated as low density lipid vacuole pools in the perinuclear space, but progressively migrated into contact with the external nuclear envelope and expanded extranuclear cytoplasmic space. Subsequent isolation of centric nuclei by lipid infiltration into the perinuclear space (Figures 3,4) was accompanied by the induction of prominent nuclear envelope pycnotic convolutions that were associated with lipid vacuole migration into contact with, or through, the external nuclear lamina. The translaminal migration and intranuclear lipid depositions (Figure 6) occurred in association with coincident TUNEL-indexed DNA fragmentation. These observations suggest that the hypercytolipidemic metabolic condition promotes a lipid-induced dissolution or chemical disruption [26] of intrinsic nuclear DNA (chromatin) organization, altering normal metabolic (transcriptional) cascade responses from being activated in response to the hypercaloric microenvironment [14,15]. The ensuing nuclear isolation, chemical disruption and structural dissolution collectively promote the subsequent apoptotic, autolytic demise of cellular organization and structural viability [15], which results in premature cytoatrophy within the affected tissues. The previously noted therapeutic effectiveness of various lipolytic agents [26-28] towards the restoration and maintenance of reproductive tract cytoarchitecture in hypogonadal genotype mutants supports the concept that lipoapoptosis, representing a lipometabolic disruption of cytointegrity within affected db/db cells, effectively compromises reproductive efficiency in experimental models or humans which suffer from Type 2 (NIDDM) diabetes- and obesity-related, hyperlipidemic metabolic (X) syndrome-induced, reproductive dysfunction [24].
In summary, the results of the present studies are the first cyto-chemical and ultrastructural evidence that apoptotic disruption of EE tissue layers in diabetes (db/db) mutant C57BL/KsJ mice occurs coincident with nuclear lipid-infiltration and DNA fragmentation, events that are linked to the hypercaloric metabolic disturbances resulting from progressive cytolipidemia within the female reproductive tract compartments [2,12]. The severity of the cytolipidemia-induced apoptosis was structurally associated with the co-localization of lipid infiltrates into the nuclear compartment of affected cells. Trans-nuclear lipid migration was progressive, migrating from an expanded perinuclear locus, through external nuclear membrane contacts, and ultimate transmembrane diffusion, into the nucleoplasm (Figure 6). The gradual, progressive accumulation of nucleo-lipid depositions eventually disrupted chromatin patterning and distribution, as evidenced by TUNEL-labeled 3'-DNA fragmentation. The gradual lipoapoptotic dissolution of nuclear continuity, and resulting separation from cytoplasmic organelle compartments, promoted pronounced EE cytoatrophy and uterine involution. The hyperlipidemia-induced, apoptotic degradation of intracellular structural integrity and metabolic homeostatic signal cascades [15] compromised reproductive competency, representing common cellular events [5,26], and shared fertility complications [28], experienced by humans and experimental models expressing obesity and Type II (NIDDM) diabetes metabolic syndromes.
Acknowledgements
The author wishes to express his sincere appreciation for the excellent technical and experimental assistance provided by Dr. Bryan L. Garris, and the histochemical and photographic support provided by Dr. Lesya Novikova, Jessica Kueker and Matthew J. Garris, during various phases of these studies.
==== Refs
Chieri RA Pivetta OH Foglia VG Altered ovulation pattern in experimental diabetes Fertil Steril 1969 20 661 668 5795045
Garris DR Ultrastructural analysis of progressive endometrial hypercytolipidemia induced by obese (ob/ob) and diabetes (db/db) genotype mutations: structural basis of female reproductive tract involution I Tissue & Cell 2004 36 19 28 14729450 10.1016/j.tice.2003.08.002
Garris DR Ovarian hypercytolipidemia induced by obese (ob/ob) and diabetes (db/db) mutations: basis of female reproductive tract involution II Tissue & Cell 2004 36 157 169 15140593 10.1016/j.tice.2004.01.001
Anderson B Mattsson LL Hahn L Marin P Lapidus L Holm G Bengtsson BA Bjorntorp P Estrogen replacement therapy decreases hyperandrogenicity and improves glucose homeostasis and plasma lipids in postmenopausal women with non-insulin dependent diabetes mellitus J Clin Endocr Metabol 1997 82 638 643 10.1210/jc.82.2.638
Cefalu WT Insulin resistance: cellular and clinical concepts Exp Biol Med 2001 226 13 26
Berg G Mesch V Boero L Sayegh F Prada M Royer M Muzzio ML Schreier L Siseles N Benencia H Lipid and lipoprotein profile in menopausal transition. effects of hormones, age and fat distribution Horm Metab Res 2004 36 215 220 15114519 10.1055/s-2004-814450
Johnson LM Sidman RL A reproductive endocrine profile in the diabetes (db) mutant mouse Biol Reprod 1979 20 552 559 378276
Garris DR Effects of diabetes on uterine condition, decidualization, vascularization and corpus luteum function in the pseudopregnant rat Endocrinology 1988 122 665 672 3276500
Garris DR Effects of estradiol and progesterone on diabetes-associated utero-ovarian atrophy in C57BL/KsJ (db/db) mutant mice Anat Rec 1989 225 310 317 2686492 10.1002/ar.1092250407
Patti ME Kahn CR Transgenic animal models: insights into the pathophysiology of NIDDM Diabetes Rev 1997 5 149 164
Morley JE Sex hormones and diabetes Diabetes Rev 1998 6 6 15
Garris DR Variable onset determinants and consequences of diabetes (db/db) obesity mutation expression: adrenergic promotion of utero-ovarian dysfunction Horm Metab Res 2004 36 312 318 15156412 10.1055/s-2004-814488
Garris DR Garris BL Diabetes-induced, progressive, endometrial involution: characterization of periluminal epithelial lipoatrophy Diabetes 2003 52 51 58 12502493
Garris DR Garris BL Diabetes (db/db) mutation-induced ovarian involution: progressive hypercytolipidemia Exp Biol Med 2003 228 1040 1050
Garris DR Garris BL Cytolipotoxicity-induced involution of the female reproductive tract following expression of obese (ob/ob) and diabetes (db/db) genotype mutations: progressive, hyperlipidemic transformation into adipocytic tissues Reprod Toxicol 2004 18 81 91 15013067 10.1016/j.reprotox.2003.10.001
Garris DR Garris BL Genomic modulation of diabetes (db/db) and obese (ob/ob) mutation-induced hypercytolipidemia: cytochemical basis of female reproductive tract involution Cell Tissue Res 2004 316 233 241 15024641 10.1007/s00441-004-0863-0
Garris DR Effects of estradiol and progesterone on reproductive tract atrophy and tissue adrenergic indices in diabetic C57BL/KsJ mice Proc Soc Exptl Biol Med 1990 193 39 45 2152978
Foreman D Kolettis E Garris DR Diabetes prevents the normal responses of the ovary to FSH Endocr Res 1983 19 187 205 8287834
Garris DR Garris BL Lipotrophic diabetes-associated utero-ovarian dysfunction: influence of cellular lipid deposition on norepinephrine indices Horm Res 2002 58 120 127 12218377 10.1159/000063579
Garris BL Novikova L Lau YS Garris DR Hypophyseal lipoapoptosis: diabetes (db/db) mutation-associated cytolipidemia promotes pituitary cellular disruption and dysfunction Pituitary 2004 7 5 14 15638292 10.1023/B:PITU.0000044628.84041.99
Garris DR Garris BL Hypercytolipidemia promotes diabetes (db/db) mutation-associated utero-ovarian involution: counter-regulatory influences of progesterone Pathophysiology 2004 11 41 50 15177515 10.1016/j.pathophys.2004.02.001
Garris DR Garris BL Diabetes (db/db) mutation-induced reproductive tract hypercytolipidemia: estrogenic restoration of utero-ovarian indices Reprod Toxicol 2004 18 641 651 15219626 10.1016/j.reprotox.2004.04.007
Coleman DL Obese and diabetes: two mutant genes causing diabetes-obesity syndromes in mice Diabetologia 1978 14 141 148 350680 10.1007/BF00429772
Garris DR Novikova L Garris BL Lau YS Hypercytolipidemia-induced nuclear lipoapoptosis: cytochemical analysis and integrated review of hypogonadal, diabetes-obesity syndrome-induced female reproductive axis disruption Metab Syndr Rel Disord 2004 2 198 209
Garris BL Garris DR Digital chemospectrophotographic identification of intracellular hyperlipidemia in diabetic endometrial epithelial cells: structural and metabolic basis of organoatrophy Pathobiology 2004 71 77 83 14707442 10.1159/000074420
Unger RH Zhou YT Lipotoxicity of B-cells in obesity and in other causes of fatty acid spillover Diabetes 2001 50 S118 S121 11272168
Coleman DL Leiter EH Appleweig N Therapeutic effects of dehydroepiandrosterone metabolites in diabetes mutant mice (C57BL/KsJ-db/db) Endocrinology 1984 115 239 243 6234160
Garris DR Estrogenic stimulation of ovarian follicular maturation in diabetes (db/db) mutant mice: restoration of euglycemia prevents hyperlipidemic cytoatrophy Cell Tissue Res 2004 318 365 373 15503160 10.1007/s00441-004-0967-6
| 15857516 | PMC1097758 | CC BY | 2021-01-04 16:37:12 | no | Reprod Biol Endocrinol. 2005 Apr 27; 3:15 | utf-8 | Reprod Biol Endocrinol | 2,005 | 10.1186/1477-7827-3-15 | oa_comm |
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-281588821010.1186/1742-4690-2-28ResearchInactivation of HIV-1 in breast milk by treatment with the alkyl sulfate microbicide sodium dodecyl sulfate (SDS) Urdaneta Sandra [email protected] Brian [email protected] Elizabeth B [email protected] Cheston M [email protected] Cara-Lynne [email protected] Hung-Mo [email protected] Mary K [email protected] Department of Microbiology and Immunology, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA2 Department of Microbiology and Immunology, Institute for Molecular Medicine and Infectious Diseases, Drexel University, College of Medicine, Philadelphia, Pennsylvania 19104 USA3 Department of Neural and Behavioral Sciences, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA4 Department of Pediatrics, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA5 Department of Pharmacology, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA6 Department of Biochemistry, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA7 Department of Health Evaluation Sciences, Penn State College of Medicine, Hershey, Pennsylvania 17033 USA8 Department of Bioscience and Biotechnology, Drexel University, College of Medicine, Philadelphia, Pennsylvania 19104 USA2005 29 4 2005 2 28 28 14 2 2005 29 4 2005 Copyright © 2005 Urdaneta et al; licensee BioMed Central Ltd.2005Urdaneta 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
Reducing transmission of HIV-1 through breast milk is needed to help decrease the burden of pediatric HIV/AIDS in society. We have previously reported that alkyl sulfates (i.e., sodium dodecyl sulfate, SDS) are microbicidal against HIV-1 at low concentrations, are biodegradable, have little/no toxicity and are inexpensive. Therefore, they may be used for treatment of HIV-1 infected breast milk. In this report, human milk was artificially infected by adding to it HIV-1 (cell-free or cell-associated) and treated with ≤1% SDS (≤10 mg/ml). Microbicidal treatment was at 37°C or room temperature for 10 min. SDS removal was performed with a commercially available resin. Infectivity of HIV-1 and HIV-1 load in breast milk were determined after treatment.
Results
SDS (≥0.1%) was virucidal against cell-free and cell-associated HIV-1 in breast milk. SDS could be substantially removed from breast milk, without recovery of viral infectivity. Viral load in artificially infected milk was reduced to undetectable levels after treatment with 0.1% SDS. SDS was virucidal against HIV-1 in human milk and could be removed from breast milk if necessary. Milk was not infectious after SDS removal.
Conclusion
The proposed treatment concentrations are within reported safe limits for ingestion of SDS by children of 1 g/kg/day. Therefore, use of alkyl sulfate microbicides, such as SDS, to treat HIV1-infected breast milk may be a novel alternative to help prevent/reduce transmission of HIV-1 through breastfeeding.
==== Body
Background
As proven in developed countries, MTCT of HIV-1 is preventable with highly active antiretroviral therapy combined with total avoidance of breastfeeding. The most widely promoted mode of replacement feeding is the use of infant formula. However, thus far, it has not been applicable in resource-constrained countries, the epicenter of the HIV/AIDS epidemic. In this setting, lack of clean water, absence of financial resources to purchase formula, and cultural stigma represent stumbling blocks for a generalized implementation of this prevention plan. Alternatives to reduce, if not prevent, the risk of transmission of HIV-1 through breast milk are in demand to act in synergy with antiretroviral regimens that prevent peripartum transmission of HIV-1. Here we introduce the novel concept of using microbicides to treat HIV-1 infected breast milk to prevent MTCT of HIV-1.
The alkyl sulfate family of microbicides are agents with both surfactant and protein denaturant properties. The prototypic alkyl sulfate, sodium dodecyl sulfate (SDS, C12H26O4SNa, CAS No. 151-21-3), is an anionic surfactant and detergent. SDS is a common ingredient used in the cosmetic and personal care products industry (e.g., toothpastes, shampoos, bubble baths, dishwashing formulations, moisturizing lotions, baby wipes, etc.), and in the laboratory environment as a denaturing agent in gel electrophoresis and other protein solubilization techniques[1,2]. SDS is listed in the Generally Recognized As Safe (GRAS) list of chemicals of the United States Food and Drug Administration (FDA)[3]. Also, the United Nations Environment Programme (UNEP) has classified SDS as "readily biodegradable" and, after extensive toxicological analysis, UNEP concluded that "sodium dodecyl sulfate is of no concern with respect to human health"[2]. According to this report, the Estimated Human Exposure (EHE) level of SDS on a daily basis is 0.158 mg/kg/day and 0.034 mg/kg/day, in children (15 kg of weight) and babies (5 kg) respectively. This includes exposure by means of body lotions and oral intake by means of contaminated water or food and toothpaste. The maximum safe ingested dose for children is estimated to be up to 1.0 g/kg/day[4].
We have previously reported that SDS and related compounds inactivate sexually transmitted viruses including HIV-1, herpes simplex virus type 2 (HSV-2) and human papillomaviruses [5-9]. SDS can inactivate cell-free macrophage-tropic (i.e., CCR5 receptor-using), T-cell tropic (i.e., CXCR4 receptor-using) or dual receptor tropic HIV-1 (i.e., strain 89.6) with concentrations as low as 0.025%[5,6]. There is an urgent need to develop safer methods to provide infants of HIV-1-infected women the benefits of human milk without the risk of the disease. To this end, the possible use of treatment with alkyl sulfates (i.e., SDS) of breast milk infected with HIV-1 has been examined. We hypothesize that treatment of expressed breast milk with this microbicide will effectively inactivate HIV-1 in breast milk. Efficiency of viral inactivation in breast milk is hereon reported. The effects of microbicidal treatment on breast milk components have also been studied (i.e., gross protein content, immunoglobulins, lipids and energy content, cellular fraction, electrolytes) and no significant changes were observed[10,11]. The results of the biochemical analysis of breast milk treated with SDS will be published elsewhere.
Results
Virucidal activity of SDS against HIV-1 in breast milk
The virucidal activity of SDS against cell-free HIV-1 in breast milk was assessed by adding high titer HIV-1 IIIB to breast milk obtained from apparently healthy donors of unknown HIV serostatus. Within 1 min of incubation of breast milk containing cell-free HIV-1 with 0.1% SDS, HIV-1 infectivity was decreased to uninfected control levels (Figure 1A). The minimum concentration of 0.05% was required to observe inactivation of HIV-1 (Figure 1B). Infectivity of cell-associated HIV-1 (i.e., HIV-1-infected Sup-T1 cells) was abolished with treatment with 0.1% SDS. This inactivation was due to induced lysis of Sup-T1 cells at this concentration (data not shown). Cell-associated HIV-1 was partially susceptible to 0.01% SDS (Figure 2A). Nonetheless, even when cell lysis is absent is not an issue low SDS concentrations abolished cell-associated HIV-1 infectivity. With 0.01% SDS, maximum inactivation of infectious cell-associated HIV-1 was achieved within 7 min of treatment (Figure 2B). Using branched DNA technology to determine HIV-1 load in spiked breast milk samples treated with ≥1% SDS, it was determined that viral RNA titers were reduced to undetectable levels (Figure 3).
Figure 1 Irreversible inactivation of cell-free HIV-1 in breast milk treated with SDS. A. Breast milk from a healthy donor was artificially infected with cell-free HIV-1 IIIB and treated with 0.1% SDS for up to15 min at 37°C prior to plating on P4-R5 MAGI indicator cells (see methods section for details). Two days later, β-gal expression was measured in relative luminescent units per second (RLU/s) in triplicate samples. Results shown are representative of three experiments. B. Infectivity of cell-free HIV-1 in breast milk treated with SDS (0.05% and 0.1%) was assessed before and after removal of SDS with SDS-300 Detergent-Out™ (see methods section for details). Results are representative of two experiments, each with triplicate samples.
Figure 2 Inactivation of cell-associated HIV-1 in breast milk with SDS. A. Supt-T1 cells infected with HIV-1 IIIB were mixed into breast milk from a healthy donor and treated with 1% or 0.1% SDS for 10 min at 37°C prior to plating on P4-R5 MAGI indicator cells (see methods section for details). Two days later, β-gal expression was measured in relative luminescent units per second (RLU/s) in triplicate samples. Levels of β-gal expression by P4-R5 cells correlates with infectivity of cell-associated HIV-1 (i.e., infected Sup-T1 cells). Results are representative of four experiments. B. Representative results of the time-course of inactivation of cell-associated HIV-1. Sup-T1 cells in media infected with HIV-1 IIIB were treated for up to 15 min with 0.01% SDS and assayed for infectivity using P4-R5 indicator cells. Samples were assayed in triplicate.
Figure 3 Reduction of HIV-1 RNA levels in breast milk treated with SDS. Cell-free HIV-1 IIIB was added to breast milk and treated with ≤1% SDS for 10 min prior to viral load determination using branched DNA technology. Shown are results of 2 independent experiments. Assay sensitivity range: 75–500,000 RNA copies/ml.
Removal of SDS from breast milk
Despite the overall benign nature of SDS, the possibility of removing SDS from breast milk in case it was deemed necessary or desirable prior to feeding was still examined. Several methods were assessed with respect to their efficiency of removing SDS from the breast milk preparations (i.e. potassium salts, Microcon® YM-10 [Amicon®, Inc.], SDS 300-Detergent-Out® [Geno Technology, Inc.]). Of these, the SDS-300 Detergent-Out® Medi kit was as efficient as potassium salts[12] with respect to the removal of the surfactant from breast milk (data not shown). The mechanism of action of this resin is proprietary information. However, >90% of the SDS initially present was removed from all samples, with a remaining concentration of SDS of 0.1% or less, as determined with reagents provided in the kit (Figure 4). Differences among treatment groups were not statistically significant (p > 0.05). If removal of the microbicide would be necessary or desirable prior to feeding the mother's milk, it is relevant to determine the potential reversal of the antiviral effect after removal of SDS. To this end, the effect of SDS removal with Detergent-Out™ on infectivity of HIV-1 was also assessed. HIV-1 infectivity was not recovered either after removal of SDS (Figure 1B). Passage of virus solutions through the resin itself decreased infectivity by 40% – 60% (Figure 1B). Paired t-test of HIV-1-infected media and milk samples before and after being passed through the column showed this difference to be statistically significant (Media p = 0.0056, Milk p = 0.0002).
Figure 4 Efficiency of SDS removal from breast milk, whole bovine milk and bovine serum albumin. Mixtures of human milk, cow's milk or bovine serum albumin (BSA) containing SDS (0.1%-1%) were subject to SDS removal with SDS-300 Detergent-Out®, as per manufacturer's instructions. SDS remaining in solution was quantified spectrophotometrically with the reagents included in the SDS-300 Detergent-Out® kit.
Discussion
We have previously shown that SDS, has broad-spectrum microbicidal activity, including anti-HIV-1 activity with concentrations as low as 0.025% [5-9]. The positive impact of feeding mother's own milk on infant health and survival are well known and promoted, even in the context of HIV-1 infection [13-15]. Here we report that, with concentrations as low as 0.1% SDS (1 mg/ml), we can inactivate in vitro high titers of HIV-1 added to breast milk. This is evidenced by the irreversible loss of infectivity of cell-free and cell-associated HIV-1, and by significant decrease in HIV-1 RNA titers. At treatment concentrations of 0.1% SDS, Sup-T1 cells were lysed contributing to the lack of infectivity observed. This result is congruent with our previously reported findings[16]. However, T cells, as well as macrophages, in colostrum were conserved after treatment with this concentration (data not shown). This discrepancy is possibly due to differences in membrane lipid and protein composition among these cell populations[17]. At this time, we do not understand why the efficiency of treatment with 0.01% SDS in inactivating cell-associated HIV-1 in breast milk is lower at 10 and 15 min of treatment. However, this should not be confused with increased infectivity because infectivity at these time points was still significantly reduced relative to the untreated milk sample (Figure 2B).
Adequate methods of milk storage were put in place to minimize the effects of freeze-thaw cycles on milk components[18,19]. Surprisingly, P4-R5 cells exposed to infected breast milk had higher expression of β-gal than those exposed to infected media (Figures 1A and 2A), and the opposite would have been expected considering the anti-HIV-1 properties inherent to breast milk. However, because the results are expressed in relative luminescent units per seconds (RLU/s), any change in β-gal expression is relative to its matched controlled. Any interference in the milk control would be the same across all milk samples in that experiment because the milk from the same donor was used for all test samples in a single experiment. In addition, we did not pool donors' milk. Therefore, the results and their interpretation should not be affected. When comparing media with breast milk, we are comparing the overall efficacy of SDS in each milieu, and we can observe that efficacy is comparable.
The decrease in HIV-1 RNA titers after microbicidal treatment (Figure 3) has also been observed by other researchers using microbicidal compounds (e.g., Nonoxynol-9) in cervico-vaginal fluids, and may be due to exposure of the viral RNA to RNases in the milk after dissolution of the viral envelope (Deborah J. Anderson, Ph.D., personal communication 12/19/03). If deemed necessary or desirable, a commercially available resin resuspended in water that can remove SDS from milk has been identified. The effects of SDS-removal with this method on human milk nutrients are data presented in a separate manuscript to be published elsewhere, where we report conservation of total milk protein species, conservation of milk immunoglobulins (number and function), and conservation of milk's energy value[10,11].
To date, we have only tested this method on very small volumes (up to 1 ml) using a column device to filter the SDS out of milk. On a greater scale, we envision a model in which breast milk could be expressed manually or mechanically (depending on the living conditions of the nursing mother) into a recipient container or bottle containing SDS. Due to the fast acting effect of SDS against HIV-1 and other pathogens, milk decontamination would occur as warm milk gets expressed into the container. The broad-spectrum action of SDS could also clear milk of other pathogens (e.g., secondary bacterial contamination) that could potentially contaminate it during expression and handling. If removal of SDS prior to feeding would be required, a filtering device comprised by the ion-exchange resin could be located within the nipple manifold in such a way that milk would be filtered through the resin as it is suctioned out of the bottle. If an infant (assuming 5 kg of weight) ingests about 700 ml of breast milk a day[18], at a treatment concentration of 0.1% this would represent an intake of SDS 0.7 g. If 90% of SDS is removed through filtration of treated milk, the final SDS concentration ingested at the end of the day would be 0.07 g; or 0.7 g if milk is instead treated with 1% SDS. Because the toxicological properties of SDS have been broadly studied in animals and humans without toxic effects even at enormous doses (e.g., 258 g in 38 days to an adult human)[2,20-23], the need for removal of SDS still requires further assessment. The metabolism and degradation pathway of SDS and other alkyl sulfates has also been elucidated in Pseudomonas, rats, dogs and humans [24-26]. Sulfatase is known to remove the sulfate, and the carbon chain is then metabolized as a fatty acid. We are currently in the process of identifying other candidate microbicides for potential use to decontaminate breast milk with respect to HIV-1 (unpublished observations). Use of edible compounds that can inactivate HIV-1 in breast milk would circumvent the issue of removing the microbicide prior to feeding treated milk[10,27-30].
Among the advantages of microbicidal treatment of expressed HIV-1-infected milk are that it is rapid, discreet (i.e., can be performed in private, minutes to hours before feeding), of low cost, and able to preserve breast milk's nutritional and protective functions. In light of the susceptibility of HIV-1 to heat[31,32], other research groups have looked into the use of heat treatment of milk to inactivate HIV-1 [33-38]. However, heat can be detrimental to important breast milk constituents[39]. In addition, lack of a readily available source of heat in some areas prevents practical application of this option[40]. Refrigeration of expressed milk would not be a sine qua non requirement as milk can sit at room temperature for up to 6–8 hours and still be considered bacteriologically safe[18,34], and SDS also has microbicidal activity at room temperature (~23°C) (data not shown). Limitations of our proposed method may be the need for bottle-feeding in settings where cup feeding may be the norm, and milk expression may represent a two-fold stumbling block for a wide spread use of this method because: (1) of the time it may require to express milk, and (2) of the added cost of the final device if a mechanical milk pumping device would be required. An economic assessment of this milk treatment option has not yet been performed. Feasibility of this preventative option also needs to be determined because we, as others, face one of the worst aspects of this epidemic: stigma of not breastfeeding.
Conclusion
Here we have introduced the novel concept of using microbicides (e.g., SDS) to treat HIV-1 infected breast milk to prevent MTCT of HIV-1. Characteristics of an ideal microbicide for treatment of breast milk include: (1) efficacy at low doses; (2) low level of toxicity; (3) broad-spectrum microbicidal activity; (4) tasteless and odorless; 5) practical to use; and (6) conservation of milk's nutritional and immunoprotective functions. SDS meets most of these requirements. However, we still need to determine the effects of SDS treatment on milk's physical properties (e.g., taste, smell). We anticipate SDS will have similar efficacy to that here reported in naturally HIV-1 infected milk. It remains to be determined, though, whether conservation of milk cells (infected and non-infected) with elimination of cell-free HIV-1 is sufficient to significantly decrease transmission. It is possible that this may be a simple way to prevent milk-borne transmission of HIV-1, while allowing HIV-1-infected mothers to continue providing the nutritional and immunological benefits of breast milk to their children.
Methods
Human milk
Breast milk was obtained, from anonymous healthy donors, of unknown HIV serostatus, and regardless of age or parity. The subjects who donated milk were either mothers of children followed in our Outpatient Clinic or nurses that work in our Pediatric Outpatient Clinic. The study was explained to them, and they signed the consent form. The milk samples used were all mature milk (>2 weeks postpartum) unless otherwise stated. Aliquots of unpooled milk were stored at -70°C in polypropylene tubes, and thawed as needed. Because milk samples were not pooled, at least two different donors were used for each experiment to control for outcomes that could be due to individual differences of each donor. This study was performed under approval of the Institutional Review Board of the M. S. Hershey Medical Center (Protocol# 0628EP).
Microbicidal treatment with sodium dodecyl sulfate (SDS)
Stock solutions of 10% (100 mg/ml) SDS (Bio-Rad Laboratories) were prepared in sterile water and kept at room temperature for up to two weeks. Volume/volume dilutions in media or breast milk were prepared fresh to obtain concentrations of ≤1%. Treatment of human milk was for 10 min at 37°C with final SDS concentrations of 1%, 0.5% or 0.1%. After treatment, SDS was removed with SDS-300 Detergent-Out™ Medi (Geno Technology, Inc.) as described below. In all experiments untreated, uninfected samples were used as controls.
Removal of SDS and SDS Detection
SDS removal was accomplished by centrifugation of 1 ml of each sample through ion exchange matrix columns (SDS-300 Detergent-Out™ Medi [Geno Technology, Inc.], Extract Clean™ IC-Ba and Extract Clean™ IC-OH [Alltech Associates, Inc.]). Reagents provided in the SDS-300 Detergent Out kit were used to colorimetrically quantify SDS remaining in solution after removal, in addition to an assay using chloroform and methylene blue as previously described[41]. Results were compared to a standard curve of SDS in deionized water. Standard curves of SDS diluted in water were compared to breast milk and whole bovine milk. At concentrations ≤0.1% SDS, there was no significant difference between absorbance measured in milk samples (human or bovine) or water samples using the SDS-300 Detegent Out™ reagents (data not shown). The chloroform-methylene blue assay has the advantage that milk (bovine or human) does not interfere with the absorbance of the sample at any SDS concentration in the standards (≤2%) and, therefore, was used for the later experiments. Optical density of the samples was measured using a visible light spectrophotometer (Spectronic 20®, Bausch & Lomb®).
HIV-1 inactivation in vitro
Inactivation of infectious cell-free HIV-1 in human milk was studied by a rapid in vitro system that quantifies remaining viral infectivity after microbicidal treatment. This system, designated MAGI (Multinuclear Activation of Galactosidase Indicator) assay[42], is based on the use of indicator P4-R5 MAGI cells. These cells are HeLa cells (immortalized cervical cancer cell line) stably expressing the HIV-1 receptor (CD4) and co-receptors (CXCR4 and CCR5) on the surface, and stably transformed with β-galactosidase (β-gal) under the control of the HIV-1 long terminal repeat (LTR). Thus, as a result of HIV-1 Tat activation of the LTR, cells infected with HIV should express β-gal. P4-R5 MAGI cells (8 × 104; obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH: P4-R5 MAGI from Dr. Nathaniel Landau) were seeded overnight in 12-well plates. Concentrated HIV-1 IIIB (5 ml; Advanced Biotechnologies, Inc.; Titer: 107.67 TCID50/ml) was treated with SDS (≤0.1% diluted in media or breast milk) for 10 min at 37°C. Media was then added to each reaction tube (1:100 dilution) and plated in triplicate. After 2 h incubation at 37°C, cells were washed and fresh media (2 ml) was added to each well. β-gal expression was measured 46 h later using a chemiluminescent reporter gene assay system (Galacto-Star™ System, Applied Biosystems). All samples were tested in triplicate.
Inactivation of cell-associated HIV-1 was achieved by treating infected Sup-T1 cells (CD4+ human T cells) with SDS (≤1%) for 10 min at 37°C prior to overlaying on P4-R5 cells. In brief, 3 × 106 Sup-T1 cells were infected with a 1:10,000 dilution of stock HIV-1 IIIB. Infected cells were subject to centrifugation, resuspended in fresh media, and incubated in the presence or absence of SDS (≤0.1%, 10 min at 37°C), three days later. Infected Sup-T1 cells (1 × 106; incubated in the presence or absence of SDS) were co-incubated with indicator P4-R5 cells (1:100 dilution of the inactivation mixture). After 2h, P4-R5 cells were washed and fed with new media. Chemiluminescent expression of β-gal was measured 46 h later. Inactivation of cell-associated HIV-1 in the breast milk was performed in a similar manner, except that infected Sup-T1 cells were resuspended in breast milk instead of media. All samples were tested in triplicate.
All chemiluminescent data was collected with a Fluorosckan® Ascent FL from Thermolab® Systems, except for data in figure 1B, which was collected with a Zylux Corporation® FB15 luminometer. We have determined that the final concentrations of SDS to which P4-R5 cells are exposed to in these assays are not toxic[6].
HIV-1 RNA load assay
In 10 μl reactions, HIV-1 IIIB (1 μl of virus stock previously diluted 1:100 in media) was added to breast milk or media, and treated with 1%, 0.5% or 0.1% SDS at 37°C. After 10 min, treatment was blocked by adding 990 μl of cold media. Samples were then immediately processed in the Clinical Laboratories of the M. S. Hershey Medical Center for viral load determination using the branched DNA (bDNA) VERSANT® HIV-1 RNA 3.0 Assay (Bayer Corporation, Inc.). This in vitro assay is clinically used to directly quantify HIV-1 RNA in plasma of HIV-1-infected individuals.
Statistical Analysis
Where indicated, samples were tested in duplicate or triplicates. All experiments were repeated two to four times to ensure reproducibility of results. All results are presented here in the form of averages ± standard deviations or as representative results, as applicable to each case. Paired t-test was used to compare samples before and after removal of SDS. ANOVA was used to compare treatment groups.
List of Abbreviations
SDS – Sodium Dodecyl Sulfate
HIV-1 – Human Immunodeficiency Virus type 1
AIDS – Acquired Immune Deficiency Syndrome
MTCT – Mother-to-Child Transmission
GRAS – Generally Recognized As Safe
FDA – Food and Drug Administration
UNEP – United Nations Environment Programme
EHE – Estimated Human Exposure
HSV-2 – Herpes Simplex Virus type 2
MAGI – Multinuclear Activation of Galactosidase Indicator
LTR – Long Terminal Repeat
bDNA – Branched DNA
Competing interests
The funding sources, NIH/NIAID No. PO1 AI37829 (MKH), NRSA Fellowship NIH/NICHD No. F32 HD41346 (SU), and Lancaster First United Methodist Church Scholarship Fund (SU), had no role in the study design; in the collection, analysis and interpretation of the data; in the writing of the report; or in the decision to submit the paper for publication. MKH is inventor in and part owner of the U.S. Patent No. 20030129588 that protects the intellectual property surrounding the use of sodium dodecyl sulfate and related alkyl sulfate compounds as microbicidal agents. MKH also serves as President of Renaissance Scientific, LLC, a virtual biotechnology company founded for the purpose of developing licenses related to this patent and other patents. To date, MKH has not received any remuneration in conjunction with alkyl sulfate-related patents. All other authors have no actual or potential, neither personal nor financial conflict of interest that may inappropriately bias their work and/or statements here presented.
Authors' contributions
SU contributed to the design of the study, acquisition of data, analysis and interpretation of the data, and drafted the manuscript. BW contributed to the design of the study and in the interpretation of the data. EBN participated in the acquisition of data. CMB obtained the IRB approval for this study and coordinated the collection of breast milk samples. CLS participated in the study design, and supervised some of the technical work. HML contributed with the statistical analysis of the data. MKH conceived the study, supervised the technical work, and contributed to the analysis and interpretation of the data. All authors critically revised the manuscript for intellectual content. All authors approved of the final version of the manuscript to be published.
Acknowledgements
The author(s) declare that they have no competing interests.
==== Refs
Tanford C Anfinsen CB NMLEJTRFM Protein Denaturation Advances in protein chemistry 1968 23 New York, Academic Press 122 127; 211-222
United Nations Environment Programme (UNEP) UNEP/OECD/UN/IRPTC SIDS initial assessment report. Sodium dodecyl sulfate (CAS No. 151-21-3) Screening Information Data Sheet (SIDS) for High Volume Chemicals Volume 4, Part 2 1997 4 Geneva, United Nations 1 39
U.S. Food and Drug Administration (FDA) EAFUS: A Food Additive Database 2002 2003 , Center for Food Safety & Applied Nutrition. Office of Food Additive Safety
Dreisbach RH Robertson WO Miscellaneous Chemicals Handbook of poisoning: Prevention, diagnosis & treatment 1987 12 Norwalk, Appleton & Lange 286 291
Howett MK Neely EB Christensen ND Wigdahl B Krebs FC Malamud D Patrick SD Pickel MD Welsh PA Reed CA Ward MG Budgeon LR Kreider JW A broad-spectrum microbicide with virucidal activity against sexually transmitted viruses Antimicrob Agents Chemother 1999 43 314 321 9925525
Krebs FC Miller SR Malamud D Howett MK Wigdahl B Inactivation of human immunodeficiency virus type 1 by nonoxynol-9, C31G, or an alkyl sulfate, sodium dodecyl sulfate Antiviral Res 1999 43 157 173 10551374 10.1016/S0166-3542(99)00044-3
Krebs FC Miller SR Catalone BJ Welsh PA Malamud D Howett MK Wigdahl B Sodium dodecyl sulfate and C31G as microbicidal alternatives to nonoxynol 9: comparative sensitivity of primary human vaginal keratinocytes Antimicrob Agents Chemother 2000 44 1954 1960 10858360 10.1128/AAC.44.7.1954-1960.2000
Howett MK Welsh PA Budgeon LR Ward MG Neely EB Patrick SD Weisz J Kreider JW Transformation of human vaginal xenografts by human papillomavirus type 11: Prevention of infection with a microbicide from the alkyl sulfate chemical family Pathogenesis 2000 1 265 276
Howett MK Wigdahl B Malamud D Christensen ND Wyrick PB Krebs FC Catalone BJ Editore M Alkyl sulfates: a new family of broad spectrum microbicides (Abstract No. ThOrC663) 2000 Monduzzi Editore 707 712
Urdaneta SL Neely EB Berlin CMJ Wigdahl B Schengrund CL Lin HM Howett MK Microbicidal compounds for treatment of HIV-1 infected breast milk to prevent postnatal transmission of HIV-1 (Abstract No. 553): ; October 4-8, Mexico City, Mexico. 2002
Urdaneta SL Berlin CM Wigdahl B Howett MK Prevention of mother-to-child transmission of HIV-1 by treatment of HIV-1-infected breast milk with alkyl sulfate microbicides (Abstract No. TuPeF5420): ; July 7-12, 2002, Barcelona, Spain. 2002 Monduzzi Editore S.p.A. Medimond Inc., pp.135-139
Takeda K Wada A Sato Y Hamada S Removal of Dodecyl Sulfate Ions Bound to Bovine Serum Albumin and Chymotrypsinogen from the Peteins. Effects of Reduction of Disulfide Bridges and Cleavage of Peptide Bonds on the Removal Extent. Journal of Colloidal Interface Science 1991 147 51 56 10.1016/0021-9797(91)90133-S
Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: a pooled analysis. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality Lancet 2000 355 451 455 10841125
Coutsoudis A Pillay K Spooner E Kuhn L Coovadia HM Influence of infant-feeding patterns on early mother-to-child transmission of HIV-1 in Durban, South Africa: a prospective cohort study. South African Vitamin A Study Group Lancet 1999 354 471 476 10465172 10.1016/S0140-6736(99)01101-0
Coutsoudis A Pillay K Kuhn L Spooner E Tsai WY Coovadia HM Method of feeding and transmission of HIV-1 from mothers to children by 15 months of age: prospective cohort study from Durban, South Africa AIDS 2001 15 379 387 11273218 10.1097/00002030-200102160-00011
Krebs FC Miller SR Catalone BJ Fichorova R Anderson D Malamud D Howett MK Wigdahl B Comparative in vitro sensitivities of human immune cell lines, vaginal and cervical epithelial cell lines, and primary cells to candidate microbicides nonoxynol 9, C31G, and sodium dodecyl sulfate Antimicrob Agents Chemother 2002 46 2292 2298 12069993 10.1128/AAC.46.7.2292-2298.2002
le Maire M Champeil P Moller JV Interaction of membrane proteins and lipids with solubilizing detergents Biochim Biophys Acta 2000 1508 86 111 11090820
Lawrence RA Lawrence RM Breastfeeding. A guide for the medical profession 1999 5th St. Louis, Mosby
Clinical Protocol #8: Human milk storage information for home use for healthy full term infants ABM Protocols (http://wwwbfmedorg/protoindxhtml) ABM Protocols (http://www.bfmed.org/protoindx.html) 2004 , Academy of Breastfeeding Medicine (ABM)
Kirsner B Wolff RA The effect of sodium alkyl sulfate on the peptic activity of the gastric contents and on the healing of gastric ulcer in man Gastroenterology 1944 2 93 101
Walker AI Brown VK Ferrigan LW Pickering RG Williams DA Toxicity of sodium lauryl sulphate, sodium lauryl ethoxysulphate and corresponding surfactants derived from synthetic alcohols Food Cosmet Toxicol 1967 5 763 769 5596421
Gloxhuber C Toxicological properties of surfactants Arch Toxicol 1974 32 245 270 4614759 10.1007/BF00330108
Gloxhuber C Kunstler K Anionic Surfactants: biochemistry, toxicology, dermatology Surfactant Science Series 1992 43 2 New York, Marceld Dekker, Inc.
Thomas OR White GF Metabolic pathway for the biodegradation of sodium dodecyl sulfate by Pseudomonas sp. C12B Biotechnol Appl Biochem 1989 11 318 327 2503014
Yao G Dodecyl Sulfate Degradation Pathway Map (http://www.labmed.umn.edu/umbbd.dds/dds_map.html)
Denner WH Olavesen AH Powell GM Dodgson KS The metabolism of potassium dodecyl [35-S]sulphate in the rat Biochem J 1969 111 43 51 5773748
Kristmundsdottir T Arnadottir SG Bergsson G Thormar H Development and evaluation of microbicidal hydrogels containing monoglyceride as the active ingredient J Pharm Sci 1999 88 1011 1015 10514348 10.1021/js9900396
Thormar H Bergsson G Gunnarsson E Georgsson G Witvrouw M Steingrimsson O De Clercq E Kristmundsdottir T Hydrogels containing monocaprin have potent microbicidal activities against sexually transmitted viruses and bacteria in vitro Sex Transm Infect 1999 75 181 185 10448397
Neyts J Kristmundsdottir T De Clercq E Thormar H Hydrogels containing monocaprin prevent intravaginal and intracutaneous infections with HSV-2 in mice: impact on the search for vaginal microbicides J Med Virol 2000 61 107 110 10745241 10.1002/(SICI)1096-9071(200005)61:1<107::AID-JMV17>3.0.CO;2-W
Maguire RA Bergman N Phillips DM Comparison of microbicides for efficacy in protecting mice against vaginal challenge with herpes simplex virus type 2, cytotoxicity, antibacterial properties, and sperm immobilization Sex Transm Dis 2001 28 259 265 11354263 10.1097/00007435-200105000-00003
Pennypacker C Perelson AS Nys N Nelson G Sessler DI Localized or systemic in vivo heat inactivation of human immunodeficiency virus (HIV): a mathematical analysis J Acquir Immune Defic Syndr Hum Retrovirol 1995 8 321 329 7882096
McDougal JS Martin LS Cort SP Mozen M Heldebrant CM Evatt BL Thermal inactivation of the acquired immunodeficiency syndrome virus, human T lymphotropic virus-III/lymphadenopathy-associated virus, with special reference to antihemophilic factor J Clin Invest 1985 76 875 877 2993366
Orloff SL Wallingford JC McDougal JS Inactivation of human immunodeficiency virus type 1 in human milk: Effects of intrinsic factors in human milk and pasteurization J Hum Lact 1993 9 13 17 8489717
Chantry CJ Morrison P Panchula J Rivera C Hillyer G Zorrilla C Diaz C Effects of lipolysis or heat treatment on HIV-1 provirus in breast milk JAIDS 2000 24 325 329 11015148
Jeffery BS Mercer KG Pretoria Pasteurisation: A potential method for the reduction of postnatal mother to child transmission of the human immunodeficiency virus Journal of Tropical Pediatrics 2000 46 219 223 10996983 10.1093/tropej/46.4.219
Jeffery BS Webber L Mokhondo KR Erasmus D Determination of the effectiveness of inactivation of human immunodeficiency virus by Pretoria pasteurization J Trop Pediatr 2001 47 345 349 11827302 10.1093/tropej/47.6.345
Boisen F Jørgensen AF Pasteurization of HIV contaminated breast milk (Abstract No. LbPp122): ; July 7-12, 2000, Durban, South Africa. 2000
Israel-Ballard K Donovan R Enge B Gesner M Scott M Sheppard H Sage A Abrams B Chantry C A novel approach for evaluating pasteurization methods to inactivate HIV in breast milk, Poster #899: February 8-11 2004. 2004
Ford JE Law BA Marshall VM Reiter B Influence of the heat treatment of human milk on some of its protective constituents J Pediatr 1977 90 29 35 318682
Papathakis PC Rollins NC Adequacy of feeding recommendations for infants of HIV-infected mothers: assessment of nutrient content, cost and preparation time of breastmilk replacements in the South African context (Abstract No. MoOrF1030).: ; Barcelona, Spain. 2002 Prous Science, S.A.
Arand M Friedberg T Oesch F Colorimetric quantitation of trace amounts of sodium lauryl sulfate in the presence of nucleic acids and proteins Anal Biochem 1992 207 73 75 1283291 10.1016/0003-2697(92)90502-X
Kimpton J Emerman M Detection of replication-competent and pseudotyped human immunodeficiency virus with a sensitive cell line on the basis of activation of an integrated beta-galactosidase gene J Virol 1992 66 2232 2239 1548759
| 15888210 | PMC1097759 | CC BY | 2021-01-04 16:36:40 | no | Retrovirology. 2005 Apr 29; 2:28 | utf-8 | Retrovirology | 2,005 | 10.1186/1742-4690-2-28 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-341583311510.1186/1743-422X-2-34ResearchBarriers to coliphage infection of commensal intestinal flora of laboratory mice Kasman Laura M [email protected] Department of Microbiology and Immunology, Medical University of South Carolina, BSB-201, P.O. Box 250504, 173 Ashley Avenue, Charleston, SC 29403, USA2005 15 4 2005 2 34 34 8 4 2005 15 4 2005 Copyright © 2005 Kasman; licensee BioMed Central Ltd.2005Kasman; 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
Growth characteristics of coliphage viruses indicate that they are adapted to live with their Eschericia coli hosts in the intestinal tract. However, coliphage experimentally introduced by ingestion persist only transiently if at all in the gut of humans and other animals. This study attempted to identify the barriers to long term establishment of exogenous coliphage in the gastrointestinal (GI) tracts of laboratory mice. Intestinal contents were screened for the presence of coliphage and host bacteria, and strains of E. coli bacteria from different segments of the GI tract were tested for susceptibility to six common laboratory coliphages.
Results
Contrary to expectations, coliphage were not evident in the GI tracts of laboratory mice, although they were occasionally detected in feces. Commensal flora showed extreme variability within groups of mice despite identical handling and diet. Less than 20% of 48 mice tested carried E. coli in their gut, and of 22 commensal E. coli strains isolated and tested, 59% were completely resistant to infection by lambda, M13, P1, T4, T7, and PhiX174 coliphage. Lysogeny could not be demonstrated in the commensal strains as mitomycin C failed to induce detectable phage. Pre-existing immunity to phages was not evident as sera and fecal washes did not contain significant antibody titers to six laboratory phage types.
Conclusion
Lack of sufficient susceptible host bacteria seems to be the most likely barrier to establishment of new coliphage infections in the mouse gut.
==== Body
Background
Coliphage have traditionally been isolated from sewage, where they arrived, presumably, after passing through the GI tracts of animals inhabited by commensal coliform bacteria. However, information on the interaction of natural coliphage with the commensal flora of the GI tract in situ is sparse. The results of ingestion of defined high titer bacteriophage preparations by laboratory animals or by humans have been described in many previous studies (reviewed in [1]). Although there is evidence in some reports of coliphage replication in the gut, the phage infections are consistently transient, becoming undetectable in 3–10 days [2-4]. Notable exceptions are cases of gnotobiotic mice inoculated with defined phage-host systems, in which phage and host populations in feces were detectable for up to 98 days [4]. However, in animals with complex, established gastrointestinal microflora, observations concur that exogenous phage do not establish sustained productive infections of the commensal bacteria. The nature of this apparent barrier to persistent bacteriophage infection of the normal GI tract is of practical interest since as parasites of commensal bacteria, bacteriophage have the potential to impact health by altering the GI flora. Conversely, it may be possible to engineer phage specifically to alter the commensal flora in situ for therapeutic benefit.
Previous work has already shown that potential physical obstacles to bacteriophage infection of bacteria in the gut are not significant. The acid environment of the stomach would at first appear to be an obvious barrier, but coliphage have been shown to maintain their infectivity when passed through the stomach or exposed to gastric fluids [4-7]. Attachment of phage to the host bacteria could be inhibited by secretions or deficiencies of cofactors in the GI environment, however infection of bacteria in the gut has also been demonstrated for a variety of phage [5,8,9], including the temperate phage, CTXphi, of Vibrio cholerae. CTXphi in fact infects its host more efficiently in vivo than in vitro, because the phage has adapted to use as its receptor V. cholerae surface molecules expressed only when the bacteria colonizes the gut [10]. However, V. cholerae is not normal flora, and CTXphi lysogens and phage are cleared in the usual time frame. In addition, antimicrobial phage therapy trials have demonstrated phage infection of bacteria in the peritoneal cavity, blood, muscle [1,11] and embryonated hen eggs (L. Kasman, unpublished observations). Therefore, the in vivo environment itself does not directly prevent infection of bacteria by phage.
Phage are immunogenic, even when introduced into the GI tract [5,9] and so anti-phage antibodies may play a role in the barrier to new phage infections in commensal coliforms. Alternatively, susceptible host bacteria may not be present at high enough densities to allow persistent infection by coliphage. When phage concentration are low, very high densities of host cells can be required to enable a detectable number of infections to occur[12]. Commensal intestinal bacteria are a climax community in which an estimated 400 different species of bacteria fill all niches with regard to locations and energy sources, so that newcomers find it difficult or impossible to become established [13,14]. A new bacteriophage could in theory face a similar obstacle, although not in terms of metabolic needs, as most phage lysogens inhibit superinfection of their host by similar phage types.
In this study, these potential barriers phage colonization of commensal bacteria were assessed for E. coli, by testing the gut contents and feces of several strains of laboratory mice for the existence of free coliphage and E. coli capable of supporting phage infection. In addition, six well characterized coliphage types were used to test commensal E coli bacteria for susceptibility to phage infection, and sera and mucosal secretions were examined for evidence of anti-coliphage antibodies.
Results
Screening for endogenous coliphage in the feces of laboratory mice
Feces of two independent groups of 40 and 48 ICR mice, were screened every three days for a total of four samplings for the presence of naturally occurring (endogenous) coliphage. Fresh fecal pellets were collected for each mouse individually, resuspended in LB, then centrifuged and the supernatants spotted onto indicator E. coli in top agar within 6 h of collection. Feces of small groups of C57BL/6 and nude mice were also collected such that over 400 samples were screened on indicator E. coli strains CN13 (F-) and ER2738 (F+) for somatic and male-specific phages respectively. A total of three samples from two mice were positive for phage on the somatic phage host only. All other feces samples were negative, indicating that even in the mice positive for phage, the phage could only be detected transiently.
Screening for coliphage and lysogens in the murine intestinal tract
In order to investigate coliphage prevalence in the GI tract as well as feces, 48 ICR and 15 C57BL/6 mice were euthanized after a final feces collection, and GI tracts removed aseptically to individual sterile petri dishes. Each GI tract was divided into four segments: (1) stomach and duodenum, (2) 10 cm of small intestine midway between the duodenum and cecum, (3) cecum, and (4) colon and rectum. Intestinal contents were collected by flushing each segment with 1 ml L-Broth containing CaCl2 and MgSO4, into a microfuge tube. Samples were gently vortexed, subjected to low speed centrifugation to pellet large particulates, and the supernatants retained for phage detection as described below. Known amounts of T4 and M13 phage were added to rinses of one mouse as a control, which demonstrated that the collection procedure and centrifugation did not remove or destroy phage (data not shown).
Due to the near absence of detectable phage in ICR mouse feces initially collected, two strategies to enhance coliphage detection were applied to the supernatants from the murine GI tracts and final feces collection from ICR, C57BL/6 and nude mice. First, in order to amplify phage which might be present at levels too low to detect, permissive E. coli host strains ER2738 or CN-13 were added to liquid cultures containing aliquots of gut or fecal supernatants. Second, mitomycin C was added since coliphage lysogens, which are otherwise undetectable by plaque assay, can usually be reactivated to the free state by treatment of the bacteria with the chemical mitomycin C. Combinations of permissive cells and mitomycin C produced five different culture conditions for the enrichment of phage recovery from feces and gut contents (Table 1). Supernatants from overnight cultures under these conditions were assayed for the presence of phage by spotting on ER2738 and CN13 agar overlays. Table 1 summarizes the results of several experiments. No male-specific bacteriophage were detected in gut contents or feces of any laboratory mice from two suppliers and three animal strains (nude mice were analyzed by feces only). Somatic bacteriophage were detected in the feces but not the gut contents of one ICR mouse after enrichment, regardless of mitomycin C treatment. Mitomycin C failed to produce any more detectable coliphage.
Table 1 Titers of bacteriophage after enrichment cultures from murine feces and contents of the GI tract
Enrichment method
1 2 3 4 5
Mouse strain source No additions CN-13 (F-) bacteria CN-13 (F-) bacteria With mitomycin C ER2738(F+) bacteria ER2738(F+) bacteria With mitomycin C
ICR feces -2 7 × 105 PFU3 6 × 105 PFU3 - -
C57BL/6 feces - - - - -
BALB/c nude feces - - - - -
ICR GI1 - - - - -
C57BL/6 GI1 - - - - -
1 All four compartments were tested separately
2 – represents none detected. Minimum detectable titer was 102/ml before enrichment.
3 Average of two mice, out of 88 tested. All others were negative.
Prevalence of commensal E. coli in laboratory mice
Prevalence of commensal E. coli was studied in 48 six-week old female ICR mice, shipped together and housed in 12 groups of 4 mice for two weeks in the MUSC animal facility, maintained on a diet of tap water and sterilized pellets. Aliquots of fecal suspensions and intestinal segment rinses for all mice were plated on individual MaConkey agar plates. Although large numbers of bacteria grew from all of the samples except some duodenum segments, less than half of the plates had colonies with morphology and color consistent with E. coli. Sixty-four individual colonies were chosen from 64 different plates and typed using the BioMerieux api20E system. It was found that only 20 of the 64 selected colonies were E. coli, 36 were Klebsiella teragina or Enterobacter aerogenes, 3 were Klebsiella ornithinolytica, 1 was Klebsiella pneumoniae pneumoniae, 1 was Myroides chryseobacterium indologenes, and 3 could not be identified. The 20 E. coli isolates were from 8 mice, so that only 8 of 48, or 17%, of the mice were found to carry E. coli. E. coli prevalence and diversity of commensal flora was similar in subsequent smaller groups of mice surveyed.
Susceptibility of commensal E. coli to laboratory-adapted coliphage
In total, 31 commensal E. coli strains recovered from four different gastrointestinal compartments and verified by BioMerieux species identification were collected from both ICR and C57BL/6 mice. These strains were tested for susceptibility to six well-characterized laboratory adapted coliphage (Lambda, M13, P1, øX174, T4, and T7) by spotting of two dilutions of each phage onto an agar overlay inoculated with the strain to be tested. Divalent cations CaCl2 and MgSO4 were supplied in the overlay as they are required for infection by P1. Susceptibility was observed as confluent lysis or individual plaques in the bacterial lawn after 18 h incubation. Isolates from different compartments of the same mouse that had identical phage susceptibility profiles were considered to be the same strain, reducing the number of unique isolates from 31 to 22. Less than 50% of these commensal strains were susceptible to any of the six coliphage (Table 2). None of the 22 strains were susceptible to M13mp18 or lambda phage. P1 was the broadest host range phage, and was able to lyse 8 of 22 strains (36%) although plaques were considerably smaller in all cases compared to T4, T7 or øX174. Similar resistance was found in a panel of 15 human clinical E. coli isolates tested for susceptibility in the same manner (Table 2).
Table 2 Susceptibility of commensal E. coli strains to lysis by laboratory coliphage (percent susceptible strains)
Source of E. coli Lambda M13 P1 øX174 T4 T7 none
Mouse feces and GI tracts (n = 22) 0 (0%) 0 (0%) 8 (36%) 3 (14%) 4 (18%) 3 (14%) 13 (59%)
Human clinical specimens (n=15) 1 (7%) 1 (7%) 6 (40%) 1 (7%) 1 (7%) 2 (13%) 7 (47%)
Screening of mouse serum and mucosal secretions for anti-coliphage antibodies
Pre-existing immunity in the form of mucosal or serum antibodies might account for the absence of coliphage in the gut. Fecal washes from 22 mice positive for lactose fermenting commensals were screened for IgA antibodies against M13, Lambda, P1, PhiX174, T4 or T7 phages by ELISA. There was variability in IgA reactivity between mice, but each individual mouse had similar reactivity to all 6 phages (Figure 1A). Based on a cut-off of two standard deviations from the mean IgA response for any given mouse, there was no evidence of phage specific mucosal IgA antibodies in this group of animals. Sera from 9 mice verified to be carriers of E. coli were likewise assayed for the presence of IgG and IgM antibodies. No evidence of phage-specific serum antibodies were found in any of the mice (Fig 1B)
Figure 1 Absence of specific antibodies to commensal coliphage by ELISA. (A) Fecal washes from 22 ICR mice found to be colonized with lactose fermenting commensal bacteria were assayed in duplicate for IgA antibodies against M13mp18, Lambda, P1, øX174, T4 and T7 coliphages in parallel ELISA plates. Mice 2B, 2R, 2G, 2N, 3G, 3R, 4N, and 6G were eventually shown to carry E. coli. Each symbol represents the mean absorbance reading for one phage type. (B) Sera from 9 mice verified as E. coli carriers were assayed in duplicate for IgG and IgM antibodies against the same six coliphage by ELISA. Each symbol represents the mean absorbance reading for one phage type.
Discussion
Bacteriophages are assumed to be a normal component of mammalian gastrointestinal microbial flora since they are commonly isolated from feces and raw sewage and grow most efficiently at temperatures approximating mammalian body temperature. This study focused on the bacteriophages of E. coli, since it is the commensal bacterial species for which the most characterized phages are known and it is easily grown in vitro. We also primarily used outbred mice (strain ICR) to more closely approximate a human population, although diet and environment were kept homogeneous. Contrary to expectations, with the exception of transient somatic phage infections in the feces of 2 of 88 mice, coliphage were not evident in the GI tracts of laboratory mice examined in this study. Our detection limit was 100 PFU/ml for GI tract rinses and approximately 100 PFU/gram feces before enrichment. However, in theory, a single infectious phage particle or single lysogenic cell should be detectable after enrichment. Coliphage densities in feces of domestic animals have been reported to vary widely by species [15]. Our results suggest that laboratory mice have some of the lowest fecal phage densities tested for domestic animals, but comparable to that reported for humans [15]. This was true for mice from two suppliers, and could have been a result of intentional association of the laboratory animals with a defined flora at the breeding facility. However, a defined flora also was not evident in these animals. At least with regard to aerobes, the commensal flora of the mice was very diverse despite the fact that the animals were all maintained in a barrier facility and fed the identical sterilized chow and water. E. coli, in particular, were detected in less than 20% of the animals, and 54% of animals produced no lactose fermenters at all on MaConkey agar.
The finding that murine commensal E. coli strains isolated were mostly resistant to infection by lambda, M13, P1, T4, T7, and PhiX174 coliphage was at first suggestive of a climax community of phage. However, superinfection exclusion became an unlikely mechanism for this resistance given the lack of phage production in response to mitomycin C. In addition, no phage-specific antibodies to known coliphage types were found in the sera or fecal washes of mice known to be colonized with E. coli. Phage are well known to be highly immunogenic in mice when injected and to pass into the blood when ingested in large numbers [7]. Taken together, these results indicate that the inability to establish persistent coliphage infections in commensal flora was not due to a climax community of phage already existing in the mouse.
Given the small percentage of commensal E. coli strains found to be susceptible to known coliphage types, the main barrier to establishment of new coliphage infections in the gut was most likely insufficient host cell density. It is estimated that in human E. coli carriers, 1 in 10,000 bacterial cells in the gut is an E. coli cell, and one gram of normal human feces contains roughly 106 live E. coli. If ideal conditions were present, such that there were 106 susceptible cells per ml of gut content, and the gut is assumed to approximate a liquid environment, then it can be calculated that a given phage present at a concentration of only 10 PFU/ml would be expected to infect on average two E. coli cells per ml of gut per hour[12]. However, conversely, in order to infect every available E. coli cell in such ideal conditions would require a phage concentration exceeding 1.5 × 106 PFU/ml. Based on our experience, the mouse intestinal volume is less than 3 mls, and the susceptible E. coli population is much lower. In addition, the majority of intestinal bacteria are located in the cecum and colon, where they will be eliminated from the body the most rapidly. It should be noted that the majority of microorganisms in the mammalian gut are anaerobes. Very little is known about bacteriophage of these organisms, but is possible that bacteriophages of anaerobes or bacteria other than E. coli would produce a different outcome.
Conclusion
The purpose of this investigation was to identify the barriers to long term establishment of coliphage in the GI tracts of laboratory mice. It was found that (1) laboratory animals living under identical conditions and with an identical diet can have very different populations of commensal intestinal bacteria, (2) E. coli, while common in human intestinal tracts were relatively rare in the guts of laboratory mice, (3) commensal E. coli found in the GI tract of normal laboratory mice were resistant to several types of coliphage. In conclusion, lack of sufficient susceptible host bacteria seems to be the most likely barrier to establishment of new coliphage infections in the mammalian gut. The results suggest that altering commensal E. coli with coliphage in situ will be very difficult due to the high variability of commensal flora between individuals, and the low probability of a coliphage particle making contact with a susceptible cell in the gut.
Methods
Bacterial strains and coliphage
Male specific phage host strain E. coli ER2738 (F'proA+B+ lacIq delta(lacZ)M15zzf::Tn10(TetR) and M13mp18 coliphage were obtained from New England Biolabs (Beverly MA). Somatic phage host E. coli strain CN13, and coliphages T4 and T7 were purchased from the American Type Culture Collection, Rockland MD. Coliphage P1 and PhiX174 were from generously provided by Dr. Caroline Westwater (Medical University of South Carolina). Lambda gt11 (Promega, Madison, WI). was generously provided by Phillip Werner, Medical University of South Carolina. E. coli bacterial strains grown from de-identified human clinical specimens were provided by Dr. Lisa Steed, Dept. of Pathology and Laboratory Medicine, MUSC. Bacteria were grown in Luria Broth with or without agar (Difco) or on Maconkey agar (Difco) as noted. Mitomycin C was purchased from Sigma and used at a concentration of 2 μg/ml.
Mice
Female ICR and C57BL/6 mice were obtained from Harlan Laboratories, C57BL/6 mice were also obtained from Jackson Laboratories. Feces from nude mice were provided by Dr. Mark Hyer (Medical University of South Carolina). All mice were housed and cared for in the MUSC animal facility and handled according to protocols approved by the MUSC Institutional Animal Care and Use Committee.
Survey for lysogeny and isolation of E. Coli strains
Gastrointestinal tracts from adult ICR and C57/BL6 mice were collected from fresh cadavers and divided into four segments: (1) stomach and duodenum, (2) 10 cm of small intestine midway between the duodenum and cecum, (3) cecum, and (4) colon and rectum. Intestinal contents were collected by flushing each segment with 1 ml L-Broth containing 5 mM CaCl2 and 10 mM MgSO4, vortexed briefly, and centrifuged at 50 RCF to remove large particulates. Aliquots of supernatant were plated onto MaConkey agar to isolate commensal E. coli. Remaining supernatant was divided and incubated under 5 different conditions to enrich for coliphage: (1) No added host bacteria, no mitomycin C, (2) CN-13 host bacteria and 2 μg mitomycin C/ml, (3) CN-13 host bacteria and no mitomycin C, (4) ER2738 host bacteria and 2 μg mitomycin C/ml, (5) ER2738 host bacteria and no mitomycin C. After 18 h shaking at 37°C, enriched cultures were pelleted by centrifugation at 20,800 RCF. The presence of male-specific phage and somatic phage in the supernatant was assayed by spotting 10 μl onto host cells according to Method 1601 of the United States Environmental Protection Agency for detection of coliphage in fresh water samples[16], except that New England Biolabs strain ER2738 was substituted for strain Famp E. coli. Fecal washes were prepared by resuspending fecal pellets in PBS containing 0.2% sodium azide at a ratio of 100 mg feces per ml. Particulates were pelleted at 14 krpm in a microfuge and the supernatants frozen until analysis.
Spotting Assay for presence of phage and analysis of phage susceptibility
The EPA protocol for detection of coliphage in drinking water[16] was followed except that E. coli strain ER2738 was substituted for strain Famp. Fecal pellets were resuspended as a 10% slurry (weight to volume) in LB containing 10 mM CaCl2 and MgSO4. For detection of phage in feces, slurries were centrifuged 10 minutes at 14 Krpm in a microcentrifuge to remove particulates before spotting 10 μl onto E. coli inoculated top agar overlay. For determination of phage susceptibility, verified commensal E. coli strains or ER2738 and CN13 control strains were grown to logarithmic phase in LB and 0.1 ml of culture used to inoculate 3 mls molten LB top agar (0.6% agar, 42°C) which was poured onto an LB agar base. When top agars were dry, 10 μl intestinal wash or 5 μl of two dilutions of lab adapted phage (Lambda, M13mp18, P1, PhiX174, T4, and T7) were spotted onto the top agar. For phage susceptibility testing, the highest dilution producing complete clearing on the control strain, and the dilution ten-fold more concentrated were used. Plates were read after incubation 18–20 h at 37°C.
Bacterial species verification
For isolation and verification of commensal E. coli strains, single bacterial colonies with suspected E. coli morphology were picked from MaConkey agar plates inoculated with gut contents or feces. Each colony was restreaked on a separate MaConkey agar plates and single colonies analyzed with the BioMerieux api20e typing kit for species identification according to manufacturer's instructions (BioMerieux, Inc. Durham, NC).
ELISA
ELISAs were performed in Reacti-Bind™ maleic anhydride activated clear 96-well plates (Pierce, Rockford IL) coated overnight with one of 6 phages (M13mp18, Lambda, P1, øX174, T4 or T7) in sterile filtered bacterial culture supernatants diluted in PBS. Fecal washes were tested for the presence of phage-specific IgA antibodies, at a dilution of 1:3 in duplicate wells for 2 h at room temperature, using a goat anti-mouse IgA-alkaline phosphatase conjugated secondary antibody (Southern Biotechnology Assoc., Birmingham AL) and Super Signal chemiluminescent substrate (Pierce). Mouse sera were incubated in the same manner except at a 1:6 dilution and utilizing secondary HRP-conjugated antibodies recognizing mouse IgG or IgM (Sigma) to detect phage-specific antibodies. Sera from mice immunized with the laboratory phages by intraperitoneal injection served as positive controls. Horseradish peroxidase conjugated secondary antibodies were detected by incubation with 1-Step ABTS (Pierce).
Competing interests
The author received salary and support from September 1999 to May 2001 under a research contract, which ended in 2001, between Hexal AG, (Industriestraβe 25, 83607 Holzkirchen, Germany) and the Medical University of South Carolina to develop bacteriophage-based antimicrobial therapeutics. No products or patents based on this work or any bacteriophage therapeutic are being pursued by either Hexal AG or the Medical University of South Carolina. The author declares no non-financial competing interests.
Authors' contributions
LK conceived of and designed the study, carried out the experiments, analyzed the data and prepared the manuscript.
Acknowledgements
This work was supported by a grant from the Medical University of South Carolina University Research Council to LK. Laboratory space and access to equipment were generously provided by James S. Norris, and the MUSC Department of Microbiology and Immunology. Phillip A Werner provided helpful assistance with bacteria typing assays.
==== Refs
Sulakvelidze A Alavidze Z Morris JGJ Bacteriophage therapy Antimicrob Agents Chemother 2001 45 649 659 11181338 10.1128/AAC.45.3.649-659.2001
Weber-Darbrowska B Dabrowski S Studies on bacteriophage penetration in patients subjected to phage therapy Arch Immunol Ther Exper 1987 35 563 568
Berchieri AJ Lovell MA Barrow PA The activity in the chicken alimentary tract of bacteriophages lytic for Salmonella typhimurium Res Microbiol 1991 142 541 549 1947426 10.1016/0923-2508(91)90187-F
Boyd JSK Portnoy B Bacteriophage therapy in bacillary dysentery Trans R Soc Trop Med Hyg 1944 37 243 262 10.1016/S0035-9203(44)90037-4
Delmastro P Meola A Monaci P Cortese R Galfre G Immunogenicity of filamentous phage displaying peptide mimotopes after oral administration Vaccine 1997 15 1276 1285 9286056 10.1016/S0264-410X(97)00072-8
Smith HW Huggins MB Shaw KM The control of experimental Escherichia coli diarrhoea in calves by means of bacteriophages J Gen Microbiol 1987 133 1111 1126 3309177
Bogovazova GG Voroshilova NN Bondarenko VM [The efficacy of Klebsiella pneumoniae bacteriophage in the therapy of experimental Klebsiella infection] Zh Mikrobiol Epidemiol Immunobiol 1991 5 8 1882608
Ando A Furuse K Watanabe I Propagation of ribonucleic acid coliphages in gnotobiotic mice Appl Environ Microbiol 1979 37 1157 1165 384907
Zuercher AW S. M. Miescher M. Vogel M. P. Rudolf M. B. Stadler Stadler. BM Oral anti-IgE immunization with epitopr displaying phage. Eur J Immunol 2000 30 128 135 10602034 10.1002/1521-4141(200001)30:1<128::AID-IMMU128>3.3.CO;2-O
Waldor MK Mekalanos JJ Lysogenic conversion by a filamentous phage encoding cholera toxin Science 1996 272 1910 1914 8658163
Westwater C Kasman LM Schofield DA Werner PA Dolan JW Schmidt MG Norris JS Use of genetically engineered phage to deliver antimicrobial agents to bacteria: an alternative therapy for treatment of bacterial infections Antimicrob Agents Chemother 2003 47 1301 1307 12654662 10.1128/AAC.47.4.1301-1307.2003
Kasman LM Kasman A Westwater C Dolan J Schmidt MG Norris JS Overcoming the phage replication threshold: a mathematical model with implications for phage therapy J Virol 2002 76 5557 5564 11991984 10.1128/JVI.76.11.5557-5564.2002
Savage DC Microbial ecology of the gastrointestinal tract Annu Rev Microbiol 1977 31 107 133 334036 10.1146/annurev.mi.31.100177.000543
Smith HW Survival of orally administered E. coli K12 in alimentary tract of man Nature 1975 255 500 502 1094297
Leclerc H Edberg S Pierzo V Delattre JM Bacteriophages as indicators of enteric viruses and public health risk in groundwaters J Appl Microbiol 2000 88 5 21 10735238 10.1046/j.1365-2672.2000.00949.x
EPA Document 821-R-00-009 (April 2000 Draft) Method 1601: Male Specific (F+) and Somatic Coliphage in Water by Two-step Enrichment Procedure. 2000
| 15833115 | PMC1097760 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 Apr 15; 2:34 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-34 | oa_comm |
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-211583678610.1186/1477-7819-3-21ReviewNeck dissections: radical to conservative Harish K [email protected] Professor & Head, Department of Surgical Oncology, M. S. Ramaiah Medical College & Hospital, Bangalore – 560054, India2005 18 4 2005 3 21 21 13 11 2004 18 4 2005 Copyright © 2005 Harish; licensee BioMed Central Ltd.2005Harish; 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
Neck dissection is an important surgical procedure for the management of metastatic nodal disease in the neck. The gold standard of neck nodal management has been the radical neck dissection. Any modification in the neck dissection is always compared with this standard. Over the last few decades, in order to alleviate the morbidity of radical neck dissection, several modifications and conservative procedures have been advocated. These procedures retain certain lymphatic or non-lymphatic structures and have been shown not to compromise oncological safety.
Methods
A literature search of the Medline was carried out for all articles on neck dissections. The articles were systematically reviewed to analyze and trace the evolution of neck dissection. These were then categorized to address the nomenclature, management of node positive and node negative neck including those who had received chemoradiation.
Results
The present article discusses the neck nodal nomenclature, the radical neck dissection, its modifications and migration to more conservative procedures and possible advances in the near future.
Conclusion
Radical neck dissection is now replaced with modified radical neck dissections in most situations. Attempts are being made to replace modified radical neck dissections with selective neck dissections for early node positivity. Sentinel node biopsy is being studied to address the issue of node negative neck. More conservative surgeries are likely to replace the 'radical' surgeries of bygone era. This process is facilitated by earlier detection of the disease and better understanding of cancer biology.
==== Body
Background
Gasparo Aselli, in 1622, first described the lacteal vessels indicating the presence of a lymphatic system. Paolo Mascagni, in 1787, and Sappey, in 1875, published atlases on lymphatics. Most of the later knowledge was derived from autopsy or post-surgical studies on nodes. The standard anatomical depictions of nodes and their drainage were based on Rouviere's descriptions. Surgical removal of nodes has paralleled the anatomical understanding. From these in vitro autopsy or post surgical studies on lymph drainage, we have progressed to in vivo sentinel node evaluation.
In the early 1800s, complete removal of disease was considered impossible once cancer had spread to 'cervical glands', although Warren in 1847 described an operation for removal of metastatic neck nodes. Butlin advised removal of nodes through a Köcher incision. A systematic operative procedure for removal of cervical lymphatics and nodes, and the first description of radical neck dissection (RND) based on anatomical principles, was described by George Crile Sr., in 1906 [1]. Hayes Martin (the father of modern head and neck cancer surgery) and his associates standardized the technique of RND [2]. Both of these surgeons followed Halsted's principles which were popular at that time. The RND is still considered by some to be the 'gold standard' treatment for neck nodes. However, modifications of the procedure are accepted and are being performed with increasing frequency. Research into breast cancer has contributed a great deal to the understanding of cancer as a systemic disease in contrast to the earlier Halstedian principles that cancer followed an anatomical step-wise progress. Suarez proposed that muscles, vessels and nerves could be preserved without adversely affecting regional control and described 'functional neck dissection' [3]. With increasing knowledge of tumor biology, the next stage in the surgery of metastatic neck disease was the promotion of the functional neck dissection by Bocca [4]. A comparable recurrence rate compared with that of RND was also claimed [5]. Later studies have shown that there is no compromise on oncological effectiveness even though there is deviation from the previous standard en bloc resection.
The regional lymphatics
Regional lymphatic drains from the scalp and skin of head and neck region, the mucosa of the upper aero-digestive tract, salivary glands, and the thyroid gland to specific regional lymph nodal groups. In addition, tumor dissemination via regional lymphatics to these lymph node groups occurs in a predictable and sequential fashion. Hence, specific regional lymph node groups should be appropriately addressed in treatment planning for a given primary tumor site. The anatomy of the nodal involvement and the biology of the disease have to be considered in planning treatment.
At this stage it is essential that we understand the division of neck nodes into 'nodal groups' described as 'levels' rather than 'anatomical regional groups'. This serves the dual purpose of understanding the natural anatomical spread of disease and the principles of various neck dissections. The nomenclature has been standardized for uniformity in documentation and cross-communication. Retrospective studies have documented the patterns of spread of cancers from various primary sites in the head and neck area to cervical nodes [6]. The terminology for categorizing the lymph node groups was originally described at Memorial Sloan-Kettering Cancer Centre, New York, and is widely accepted. It divides the hemi-neck into five regions. A sixth region was added later [7]. The nodal levels are diagrammatically shown in figure 1 and a brief description of neck nodal levels [8] is described in Table 1. The regional lymph node staging [9] is described in Table 2. Thus, all regional nodes are not at risk of metastases initially from any primary site in the absence of grossly palpable nodes. For example, in oral tumors, if neck is clinically negative, level IV and V nodes are generally not at risk of harboring micro-metastases. A similar situation applies to other organ and node metastases. The first nodal basin varies for each site and is described in the Table 3. Skip metastases in the absence of first echelon nodes being involved are unusual. On the other hand, when clinically palpable nodes are present, comprehensive clearance of all regional nodal groups is warranted (see also discussion under selective dissection for node positive neck). Several well documented studies have confirmed this to be the case [7].
Table 1 Cervical nodal levels
Levels Nodes Boundaries
Superior Inferior Anterior (medial) Posterior (lateral)
IA Submental group Submental nodes Symphysis of mandible Body of hyoid Anterior belly of contralateral digastric muscle Anterior belly of ipsilateral digastric muscle
IB Submandibular group Submandibular nodes Body of mandible Posterior belly of digastric muscle Anterior belly of digastric muscle Stylohyoid muscle
IIA Upper Jugular group Nodes Nodes around upper portions of Internal jugular vein and Accessory Nerve Skull base Horizontal plane defined by the inferior border of hyoid bone Stylohyoid muscle Vertical plane defined by the spinal accessory nerve
IIB Upper Jugular group Nodes Skull base Horizontal plane defined by the inferior border of hyoid bone Vertical plane defined by the spinal accessory nerve Lateral border of the sternocleidomastoid muscle
III Mid Jugular group Nodes around mid portions of Internal jugular vein Horizontal plane defined by the inferior border of hyoid bone Horizontal plane defined by the inferior border of cricoid cartilage Lateral border of the sternohyoid muscle Lateral border of the sternocleidomastoid muscle or sensory branches of the cervical plexus
IV Lower Jugular group Nodes around lower third of IJV Horizontal plane defined by the inferior border of cricoid cartilage Clavicle Lateral border of the sternohyoid muscle Lateral border of the sternocleidomastoid muscle or sensory branches of the cervical plexus
VA Posterior triangle group Nodes around lower part of Accessory nerve and transverse cervical vessels Apex of the convergence of sternocleidomastoid and trapezius muscles Horizontal plane defined by the inferior border of cricoid cartilage Lateral border of the sternocleidomastoid muscle or sensory branches of the cervical plexus Anterior border of the trapezius muscle
VB Posterior triangle group Horizontal plane defined by the inferior border of cricoid cartilage Clavicle Lateral border of the sternocleidomastoid muscle or sensory branches of the cervical plexus Anterior border of the trapezius muscle
VI Central or anterior group Nodes surrounding midline visceral structures of neck Hyoid bone Suprasternal Common carotid artery Common carotid artery
Table 2 Regional lymph node staging
Nx Regional lymph nodes cannot be assessed
N0 No regional lymph node metastasis
N1 Metastasis in a single ipsilateral lymph node, 3 cm or less in greatest dimension
N2a Metastasis in a single ipsilateral lymph node, more than 3 cm but less than 6 cm in greatest dimension
N2b Metastasis in a multiple ipsilateral lymph nodes, none more than 6 cm in greatest dimension
N2c Metastasis in bilateral or contralateral lymph nodes, none more than 6 cm in greatest dimension
N3 Metastasis in a lymph node more than 6 cm in greatest dimension (Midline nodes are considered ipsilateral nodes)
(Adapted from UICC / AJCC TNM Classification of Malignant Tumors, 6th edition, 2002)
Table 3 Patterns of neck nodal metastasis
Primary site First echelon nodes
Oral cavity Levels I, II, III
Larynx, Pharynx Levels II, III, IV
Thyroid Levels IV, VI, superior mediastinal
Parotid Levels II, III, Pre-auricular, Peri & intra parotid, Upper accessory chain
Submandibular, sublingual glands Level I, II, III
Figure 1 Diagrammatic representation of the neck showing various nodal levels and sublevels
The risk of nodal metastases is dependent on various factors related to the primary tumor. These include the site, size (large or small), T stage, location of the primary tumor (within an organ such as the vocal cord compared with supraglottis) and histomorphology of the primary tumor. The risk of metastases increases as one progresses from the anterior to posterior part of the upper aero-digestive tract; from lip (10%) progressing along the tongue (25%), gum (30%), floor of mouth (40%), oropharynx (55%) to hypopharynx (65%). Endophytic tumors, poorly differentiated tumors, and tumors with a greater thickness (tongue and floor of mouth) are more likely to have metastases [10].
The goal of treatment of cervical nodal metastases is regional control of disease. Micrometastases and minimal gross metastases may be controlled by radiotherapy but surgery remains the mainstay of treatment as it provides comprehensive clearance of all grossly enlarged nodes and offers accurate histological information regarding the presence of micrometastases. While the indications for comprehensive surgical clearance of regional nodes in patients with clinically metastatic disease are obvious, the indications for elective nodal dissection (END) in patients with N0 neck status are more challenging and are discussed separately. The present discussion is aimed at answering why, when, and how to treat neck nodes in patients with primary squamous head and neck cancers?
Neck dissection nomenclature
The term "neck dissection" refers to a surgical procedure in which the fibro-fatty soft tissue content of the neck is excised to remove the lymph nodes that are contained therein [11]. Over the years, a number of variations and modifications of RND have been described. In addition, different terminologies have been used to represent the same surgery as well as different surgeries. The classification by the American Academy of Head and Neck Surgery, and which is also endorsed by American Society of Head and Neck Surgery, has become accepted widely [7]. The principles are as follows:
1. RND is the standard basic procedure for cervical lymphadenectomy, and all other procedures represent one or more modifications of this procedure.
2. When the modification involves preservation of one or more non-lymphatic structures, the procedure is termed a modified RND (MRND)
3. When the modification involves removal of additional nodal groups or non-lymphatic structures relative to RND, the procedure is termed extended RND.
4. When the modification involves preservation of one or more lymph node groups that are routinely removed in RND, the procedure is termed as selective neck dissection (SND).
The first three procedures are also grouped together under the name of "comprehensive neck dissections". A modified classification is outlined in Table 4. An example of extended neck dissection includes removal of parapharyngeal or buccinator or superior mediastinal nodes. It could also involve removal of non-lymphatic structures such as the hypoglossal or vagus nerves, paraspinal muscles or carotid artery.
Table 4 Classification of neck dissections*
Comprehensive neck dissection Radical Neck Dissection
Extended Radical Neck Dissection
Modified Radical Neck Dissection type I
Modified Radical Neck Dissection type II
Modified Radical Neck Dissection type III
Selective neck dissection Supraomohyoid neck dissection
Jugular (antero-lateral) neck dissection
Central compartment (anterior) neck dissection
Posterolateral neck dissection
*The nomenclature of various selective neck dissections are not recommended for usage. They are replaced with SND followed by the nodal levels or sublevels removed. However these named procedures are retained here as this is a transition phase and this nomenclature was being followed not very long ago.
Other terminologies used are "elective" and "therapeutic" neck dissections. If a neck dissection is carried out when there is no evidence of neck disease it is termed an "elective" neck dissection (END). Some authors use the word "prophylactic" instead of "elective" to denote the same procedure. However, the word "prophylactic" is not recommended as it gives a false impression. If the neck dissection is undertaken for metastatic disease in the neck it is called a "therapeutic" neck dissection. An END is usually, but not always, a SND. Hence END and SND are not synonymous terms
Management of positive neck nodes
Palpable cervical nodes in patients with carcinoma of the upper aero-digestive tract are usually assumed to be due to metastases. Further evaluation with FNAC is needed only in patients where the nodes do not appear to be due to metastatic diseases as assessed clinically. At the other extreme, computerized tomography (CT) scan or magnetic resonance imaging (MRI) or a Doppler study may be required depending on what vital structure appears to be involved by nodal disease, or when there is a doubt about the ability to surgically remove the disease.
The presence of nodal metastases reduces survival by 50% [12,13]. The prognosis of patients with cervical metastases is affected by the number of metastases, level of metastases, tumor burden, presence of extra capsular spread, nodal resectability and previous treatment with surgery or irradiation [12]. With the exception of patients with nasopharyngeal carcinomas, clinically apparent cervical lymph node metastases should be treated with a neck dissection [14].
A classical RND has been regarded as the gold standard for surgical treatment of clinically apparent metastatic cervical lymph node disease. When regional metastases are clinically palpable, comprehensive clearance of all the regional nodes which are at risk is mandatory. The underlying principle is that once metastases have occurred to one node, than there is likely to be sub-clinical metastases in other nodes. However, the operation has a significant morbidity which includes shoulder syndrome (shoulder weakness, deformity, pain, and adhesive capsulitis), anesthesia over the cutaneous distribution of the cervical plexus, and cosmetic deformity due to loss of the sternocleidomastoid muscle. The procedure involves removal of all ipsilateral cervical lymph node groups extending from the body of the mandible superiorly, to the clavicle inferiorly, anteriorly from the lateral border of sternohyoid, hyoid, contralateral anterior belly of digastric muscle belly, to the anterior border of trapezius posteriorly. Nodal groups from level I through to V are included in this dissection. Important extra-nodal structures such as the spinal accessory nerve, internal jugular vein and sternocleidomastoid muscle are sacrificed. Post- auricular, periparotid, suboccipital, perifacial, buccinator, retropharyngeal and paratracheal nodes are not removed in this nodal dissection.
The current indications for classical RND [15] are:
• N3 neck disease, especially in the upper neck
• Bulky metastatic disease near the accessory nerve
• Tumor directly involving the accessory nerve
• Clinically palpable multiple nodes, especially if near the accessory nerve (N2b, N2c)
• Recurrent metastatic tumor after previous irradiation therapy
• Recurrent disease in the neck after previous neck dissection
• Salvage surgery in patients after chemo-irradiation therapy, especially in those who presented with bulky or level II nodal disease
• Involvement of the platysma or skin, requiring sacrifice of a portion of skin in the upper neck
• Clinical or radiological signs of obvious extra-nodal disease e.g. fixity
On the other hand, when appropriate indications exist, a function-preserving comprehensive neck dissection, sparing one or more vital structures, should be considered as long as it does not compromise the performance of a satisfactory clearance of metastatic disease. Preservation of the accessory nerve alone significantly reduces the morbidity of neck dissection. Thus, if the spinal accessory nerve is not involved by metastatic disease, it should be routinely preserved even in patients with clinically palpable metastatic nodes as the accessory nerve is rarely invaded by metastatic disease, and its preservation does not affect local recurrence of disease or overall patient survival. The cervical plexus contributes in part to the motor supply of trapezius through the accessory nerve [16-18]. However, it has also been pointed out that such a contribution may be insignificant and inconsistent [19]. Spinal accessory nerve dysfunction occurs even after selective dissection and is usually attributed to stretching of the nerve due to retraction during clearance of level IIb lymph nodes and/or ischemia [20,21]. Recent anatomical findings have shown that, functionally, the most important descending part of the trapezius muscle is innervated by a fine single branch arising from the spinal accessory nerve in the posterior triangle of the neck. Preservation of this branch may help to prevent more morbidity in patients undergoing modified radical neck dissections [22]. Accessory nerve sparing should not result in less than optimum nodal disease clearance and the nerve should be sacrificed if warranted in order to achieve this. Preservation of the sternocleidomastoid muscle, or internal jugular vein, is not recommended in patients with palpable nodes from an upper aero-digestive tract primary tumor. However, the same can be preserved, in addition to the accessory nerve, in nodal metastases from a well differentiated thyroid carcinoma. In addition, the sternomastoid can be preserved for cosmesis in young individuals and in those with minimal nodal disease. When bilateral neck dissection is undertaken simultaneously, it is important to preserve at least one internal jugular vein, preferably on the less involved side. It has also been realized that neck dissection is inadequate when employed as a sole modality in disease with adverse histological features such as perineural invasion and extracapsular spread. This has led to the usage of combination therapy of surgery with irradiation in selected patients. The common indications for RND, MRND and structures removed or spared are listed in Table 5. The overall results comparing RND and MRND [10,23-31] irrespective of the primary tumor are listed in Table 6. From the above discussion it is apparent that spinal accessory nerve preservation lessens the morbidity to a great extent and preservation of sternocleidomastoid muscle and internal jugular vein is of lesser importance. However, this issue becomes of significance where bilateral neck dissections are required.
Table 5 Comprehensive neck dissection*
Neck dissection Indication Comments
RND Clinically metastatic neck nodes of upper aero-digestive tract Sternocleidomastoid, accessory nerve, Internal Jugular vein and submandibular gland removed
MRND TYPE I As above, when accessory nerve away from nodal disease Accessory nerve spared. Rest as RND
MRND TYPE II Thyroid well differentiated cancer, nodes selectively involving IJV Sternocleidomastoid, accessory nerve spared. Internal Jugular vein sacrificed. Rest as RND
MRND TYPE III Thyroid well differentiated cancer Sternocleidomastoid, accessory nerve, Internal Jugular vein spared
EXTENDED RND Extensive involvement of nodes beyond usual levels or involvement of contiguous organs Additional lymph nodes** or other non lymphatic structures removed
IJV: internal jugular vein; RND: radical neck dissection
* All procedures involve removal of nodes from level I through V
** Nodal areas include retropharyngeal, parapharyngeal, mediastinal or axillary, Structures include cranial nerves, carotid artery, muscles, skin etc.
Note: It is recommended to uniformly use the term modified radical neck dissection (MRND) and not classify as I or II or III but instead name the structures preserved.
Table 6 Overall results (irrespective of site of primary tumor)
Nodal status Neck dissection Neck failure
N+ RND 10% – 22%
MRND type I 4.8% – 26%
N1 RND 8% – 15%
MRND type I 0 – 16%
N2 RND 12% – 26%
MRND type I 15% – 25%
N3 RND 21%
RND: radical neck dissection; MRND: modified radical neck dissection
The outlook for patients with "fixed" nodes is poor [12,32]. The poor prognosis is probably due to an incomplete resection and there being extracapsular spread of disease. The term "fixed node" is vague in the sense it conveys different meaning to different people unless it is known to what structure is the node "fixed". True fixation occurs in 23 to 50% of patients [32]. Tumors fixed to mandible, larynx, sternocleidomastoid, prevertebral muscles and mastoid process can be resected with a margin of normal tissue. Invasion of skin can similarly be addressed by a wide margin of resection with resurfacing then carried out. Adjuvant irradiation can then be employed for microscopic disease. However, the 5-year survivals are poor and are reported to be around 15% [32]. Irradiation followed by salvage surgery did not improve survival and the concept of preoperative irradiation 'to make an inoperable tumor' "operable" is questionable [32]. Invasion of the carotid sheath does not preclude surgery and can be present in 5% of patients [32,33]. Carotid artery invasion is also associated with a poor prognosis. When the external carotid artery alone is involved, it can be sacrificed. When the common or internal carotid is involved, the options are either no treatment, debulking with adjuvant therapy, mixed beam therapy or resection with or without revascularization. Patients should undergo cerebral blood flow studies prior to surgery and elective carotid artery resection can be undertaken. Earlier, carotid resection carried a mortality of 12% and hemiplegia rate of 33% [34]. Recently 22% 2 year disease-free survivals, with acceptable morbidity, have been described [35]. A median disease-specific survival of 12 months, with 24% of patients dying from distant metastases has been reported. The high-risk of complications, loss of quality of life and mortality, must be balanced against the natural history of the disease if it is left untreated [36]. When cervical spine or brachial plexus is involved, sacrifice of these with their attendant morbidities should be weighed against the local control of disease that could be achieved and the benefits that it will offer to patients.
Bilateral neck disease at initial presentation is a bad prognostic sign with a 5-year survival of only 5% then being likely [13]. Survival is not affected by laterality but by nodes larger than 6 cm in size [37]. Therefore, the presence of bilateral mobile nodes should not preclude treatment. At least one internal jugular vein should be preserved or reconstructed with a saphenous graft [38] or polytetrafluoroethelyene graft. Simultaneous ligation of both internal jugular veins can cause venous congestion, edema of the head and neck, raised intracranial pressure and the syndrome of inappropriate secretion of anti-diuretic hormone. Salvage surgery following a previous neck dissection also carries a dismal outlook, with a long-term survival of less than 5% [39].
As an extension of conservation surgery, some authorities advocate SND for metastatic neck nodes [40,41]. These have been only applied in cases without evidence of massive lymphadenopathy, node fixation or gross extra-capsular spread. In addition, when a suspicious node is encountered in the lower limit of the dissection, a formal RND or its modifications can be undertaken. Although the primary site in these cases varies, it has been reportedly performed in low grade buccal or buccoalveolar lesions by an experienced surgeon. SND, when used in combination with postoperative irradiation therapy, has been shown to achieve comparable results to those reported for RND [42]. However, as a routine, a SND is not recommended for patients whose neck nodes are positive for tumor.
The role of neck dissection after definitive chemoradiotherapy for squamous cell head and neck cancer is not clearly defined [43]. After chemoradiotherapy (5 Fluorouracil and cisplatin), clinical parameters do not identify those patients with residual neck node disease or those at risk for regional failure, suggesting that neck dissection should be considered for all patients with N2-N3 disease [43]. A retrospective non- randomized study showed that the 5-year survival rate was significantly higher for those who had postoperative radiotherapy (38.9%) compared with patients who had pre-operative radiotherapy (9.1%) and those having surgery alone (0%). They were unable to evaluate the role of definitive chemotherapy and/or radiotherapy and salvage surgery as the results were inconsistent and the available data was limited [44].
Surgical resection is an essential component of aggressive chemoradiation protocols to ensure tumor control at the primary site and also in the neck. Neck dissections in patients with initial node positive disease reveal residual nodal metastases in 22% [45]. A German multicenter randomized trial showed that accelerated chemoradiation (Mitomycin and 5 Fluorouracil) to 70.6 Gy was more effective than accelerated radiation to 77.6 Gy alone. Selective lymph node dissection of residual neck masses after completion of hyperfractionated accelerated radio-(chemo-)therapy is likely to contribute to loco-regional tumor control in patients with advanced head and neck cancers [46]. On the contrary, another study did not recommend neck dissection for patients who respond completely after chemoradiation but did recommend it for patients with either no or partial radiographic responses [47]. Another study showed improved neck control with neck dissection for patients with clinically residual disease or N3 neck cancer but no significant impact on the outcome of patients with N2 stage disease who were rendered clinically disease-free with intensive concurrent chemoradiation [48]. There is also a suggestion that neck dissection followed by chemoradiotherapy has a higher, and statistically significant, disease-specific survival rate compared to chemoradiation alone [49]. Consensus regarding the use of neck dissections for complete responders and incomplete responders has yet to be achieved and the data are controversial. A possible benefit from neck dissection after a complete response of the primary tumor after combined chemoradiation or definitive external radiation for advanced squamous cell carcinoma of the head and neck may only be anticipated in patients with persisting sub-clinical neck disease who have no other sites of disease. Some clinicians have even argued that the salvage rate for clinically detectable residual neck disease does not justify neck dissection being performed. Randomized trials addressing these questions, and a trial addressing the accuracy of new imaging modalities such as post chemotherapy and post radiation positron emission tomography scanning, across multiple institutions would be appropriate [50,51].
Pathological evaluation of the neck dissection specimens is important. A RND or MRND specimen should comprise of at least 10 nodes on pathological examination. Apart from the number of involved nodes it is important to identify adverse prognostic factors resulting in a predisposition to local recurrence and a reduced survival. Perineural invasion, extracapsular spread and vascular invasion indicate a bad prognosis [52,53]. Desmoplastic stromal reaction was associated with a seven fold increase in the risk of recurrent disease. Tumor has been identified up to 12 cm along the perineural space in patients with perineural invasion, and should be considered as a positive tumor margin [2]. Patients with poor prognostic factors listed above should receive adjuvant therapy such as irradiation, with or without chemotherapy, as required.
Management of patients with N0 neck status
As discussed earlier, comprehensive neck dissection is the standard treatment for obvious nodal disease in the neck. With emphasis on function preservation and cosmesis in addition to achieving adequate local disease control, conservative neck dissections have gained popularity. Growing historical evidence suggests that modified and selective neck dissections offer disease control comparable to radical neck dissection but with less morbidity [6,40,54,55].
The problem
It is well known that 20% to 40% of patients with head and neck cancer who have no palpable disease in their necks will harbor occult disease in their necks. It is, therefore, therapeutically tempting to treat the neck in these patients on the basis that this will avoid subsequent treatment, which may be not as effective. It means that 60%-80% of patients are "over-treated". The pros and cons of this [56-61] are addressed in Table 7.
Table 7 Pros and cons for elective neck dissection (END)
For END Against END
Neck dissection has low morbidity & mortality END results in a large number of unnecessary surgical procedures and is associated with inevitable morbidity
Cure rate for neck dissection is decreased if gland enlargement occurs or multiple nodes appear Cure rates are no lower if the surgeon waits for the neck to convert from N0 to N1
It is impossible to provide follow-up necessary to detect the earlier conversion of a neck from N0 to N1 Careful clinical follow-up will allow detection of the earliest conversion from N0 to N1
Allowing the neck metastases to develop increases the incidence of distant metastasis END removes the barrier to the spread of disease and also has a detrimental immunological effect
If neck has been entered to remove the primary it is better to perform an in-continuity resection Radiation is as effective as neck dissection in N0 neck
High incidence of occult metastatic disease
Evaluation of patients with N0 neck status
Cervical nodal staging is a major challenge. Clinical examination is influenced by the skill of the examiner, the patient's body habitus and whether the patient has had previous surgical or irradiation therapy. As a result of these factors, the false negative rate in clinical assessment ranges from 20%-51% [62]. Imaging techniques such as CT scanning, MRI and ultrasound have been used for evaluation. The sensitivity of CT scanning is 81% [63]. Size of nodes and central lucency of nodes have been used as criteria for identifying nodes containing metastatic tumor but there is a low specificity [64]. Routine scanning of the neck in a patient with a head and neck primary tumor is not justifiable at present. With ultrasound and ultrasound guided aspiration, the accuracy has improved to 89% but it is highly operator- dependent [65]. No pre-treatment study can accurately assess or replace the requirement to stage the neck histopathologically [63]. Hence the goal of identifying 'sub-clinical disease' without surgical intervention remains elusive.
Surgical management
A standard RND has no role in the management of patients with N0 neck status. A SND, or rarely a MRND, would be required. A recent study has shown comparable recurrence rates for RND compared with SND with there being no statistically significant difference [66]. It has to be emphasized that an END procedure must be "individualized". The decision regarding the type of neck dissection depends on the site of primary tumor, other tumor factors (e.g. location, depth, size, differentiation, vascular or perineural invasion), the patient and the surgeon undertaking the procedure. Tumors of the tongue, floor of mouth, nasopharynx, oropharynx, hypopharynx and supraglotic larynx have a high incidence of nodal metastases. In contrast, tumors of the buccal mucosa, lip, paranasal sinuses and glottic larynx metastasize less frequently. For a patient who stays some distance from the surgical centre and who is not expected to come for regular follow-up (and for unreliable patients) an END would be a better option. Documentation of differences in regional control rates or survival between patients undergoing END for N0 disease or patients undergoing therapeutic neck dissection for N1 disease remains controversial. Some authors claim no difference [10] while other studies have shown that END significantly improved regional control of the disease [67,68]. Unfortunately, many patients do not present later with just N1 disease but with disease greater than that of N1 staging [10]. Hence END must always be considered based on the primary tumor characteristics as prognosis depends on the extent of metastatic neck nodal disease. An END accurately stages the disease pathologically, detects and treats micro-metastases, thus controlling regional disease and helping in the planning of adjuvant therapy.
When level IV nodes are removed along with supraomohyoid neck dissection (SOND), as in primary tongue, terminologies like extended SOND have been used. This adds to the complexity of the classification. Further, the classification of types of SND does not have nomenclature options if the surgeon preserves certain nodal sublevels. Hence, it is recommended to use the nodal levels or sublevels which are dissected instead of the named SND [8].
Radiation in patients with N0 neck status
External radiation in doses of 40 to 50 Gy will control occult metastases in 90 to 95% of cases [69]. Although prospective evidence is lacking, retrospective data suggest that for most sites and for early lesions, elective nodal irradiation (ENI) and END offers equivalent local control [54,70]. Proponents of ENI assert that the morbidity is low with limited soft tissue changes and does not have systemic ramifications. However, considerable acute adverse effects such as mucositis and xerostomia, together with late effects like endarteritis, radionecrosis etc .can occur. Systemic effects include suppression of humoral and cell mediated immunity.
The issue of the use of elective surgery versus elective radiation ends not at which treatment modality is more beneficial, but which one is less harmful. The patient's age, general health, family support, reliability and patient's own wishes are important. It is impossible to compare elective neck dissection and elective nodal irradiation efficacy because the status of neck disease is unknown when elective irradiation is used. The accurate histological information on micrometastases in neck nodes in patients with clinically negative neck nodes is probably one of the prime factors that tilts the argument towards nodal dissection, apart from the lesser associated morbidity. If the treatment planning for the primary tumor involves surgical excision through a neck approach, END is opted for when indicated. If the primary tumor is being treated with irradiation, an elective irradiation of nodal area should be planned.
Most authors recommend that END be performed when the expected incidence of microscopic or subclinical disease exceeds 20%. The National Cancer Comprehensive Network's adopted practice guidelines recommend END for cancers of the oral cavity, oropharynx, hypopharynx and supraglottic larynx when the nodal status is N0.
Tumors of the nasopharynx are primarily treated with radiotherapy, which includes nodal management. Surgery plays a limited role with this being restricted to salvage of residual neck disease following failure of primary radiotherapy.
For oral tumors, the high incidence of occult nodal metastases, and notoriously poor salvage rates, make expectant treatment ill-advised in most cases [54,55,71]. The lower alveolar ridge carcinomas have a low potential for neck metastases [72]. With the exception of this, treatment of sub-clinical neck disease is indicated for all oral cavity lesions. It is also important to note that carcinomas of the oral cavity rarely spread to lower jugular or posterior cervical nodes in the absence of involvement of level I-III. An elective SND (levels I – III) provides staging information and may be the only therapy necessary for occult disease. It has minimal morbidity and it reduces the risk of occult disease evolving into clinically evident metastases. In addition, the undesirable effects of radiotherapy are avoided. A tongue tumor with a depth of invasion of more than 4 mm is likely to produce occult nodal metastases [73]. As a sub-site of oral tumors, the tongue is notorious for nodal metastases. For a well differentiated T1 tumor, with less than 4 mm of muscle invasion and a clinical N0 status in the neck, a model predicts a 14% chance of nodal metastases. Sub-digastric and mid jugular are the most frequently involved nodal groups in primary oral cavity tumors [65]. Evidence also indicates that level IV nodes may also be at risk in a tongue primary cancer [74]. An ipsilateral SND (levels I – IV) is indicated for tongue primaries. Both sides of the neck have to be addressed in midline, or near the midline, tongue primaries.
In tumors of the oropharynx, the risk of occult nodal metastases is 30%-35%. There is a high incidence of level II, III and IV involvement [55,74]. In addition, there is an increased risk of bilateral occult nodal disease in tumors near the midline and treatment should include both sides of the neck. Tumors of the hypopharynx are known for extensive sub-mucosal spread. The retropharyngeal and paratracheal nodes are also at risk of harboring metastases. Occult nodal metastases are present in approximately 30%-55% of patients with primary hypopharyngeal tumors and occur frequently in the contralateral side of the neck [55,75]. As stated earlier, the modality used in the treatment of the primary is an important consideration whilst planning treatment of the neck. Larger tumors (T3 or greater) usually require combined therapy. In cases where surgery is chosen to treat an advanced primary cancer, a bilateral neck dissection aiming at removal of all the nodes at risk is performed. If combined therapy is planned, a unilateral neck dissection is undertaken and irradiation may be used to treat the un-dissected neck. The choice of treatment is dependent on various factors as discussed earlier.
In tumors of the supraglottic larynx, the location of the tumor has a prognostic significance. Marginal lesions like those on aryepiglottic fold and suprahyoid margin of the epiglottis behave more aggressively with an increased risk of nodal metastases. There is also a high risk of contralateral nodal metastasis for lateral (aryepiglottic fold) as well as midline (epiglottic) lesions. This makes bilateral neck dissections a necessity when electively performed [76]. Carcinoma of the supraglottic larynx has a high incidence of occult metastases in the range of around 30% [55]. Based on current evidence, if surgery is planned for primary therapy, SND of levels II through IV is acceptable [77]. Dissections need to be bilateral to prevent contralateral neck failure and recurrence of disease. As stated earlier, if irradiation is planned, the contralateral side of the neck may be irradiated.
In contrast, glottic cancers remain localized due to sparse lymphatics. T1 and T2 carcinomas have less than 7% nodal metastases. If such patients are available for regular follow-up, these patients may be closely observed. But, recurrent cancers have a significantly higher occult metastases rate reaching 20%-22%. An END incorporating levels II through IV may be planned [55,77]. T3 and T4 glottic carcinomas have occult nodal metastasis in the range of 20%-40%. In addition, salvage rates for recurrences are dismal. Hence, T3 and T4 glottic carcinomas should receive ipsilateral SND of levels II through IV and level VI.
The reported incidence of subglottic tumors is low. These have a high incidence of extralaryngeal spread due to the proximity of the cricothryoid membrane and the rich post-cricoid lymphatics. They drain to pre- and para-tracheal, pre-laryngeal and superior mediastinal nodes. The sparse data makes conclusive recommendations difficult. However, a wide field laryngectomy, bilateral para-tracheal dissection, pre-laryngeal nodal dissection with an ipsilateral hemithyroidectomy should be planned with postoperative radiation to stoma, mediastinum and both sides of neck. An organ based general plan of SND is outlined in Table 8.
Table 8 Selective neck dissection (N0 neck)
Organ Nodal clearance
Oral cavity I, II, III
Tongue I, II, III, !V
Hypopharynx, larynx, oropharynx II, III, IV
Some laryngeal and hypopharyngeal lesions where IIB is not removed IIA, III, IV
Laryngeal, hypopharyngeal extending below glottis II, III, IV, VI
Thyroid, hypopharynx, cervical trachea, cervical esophagus, sub-glottic larynx VI,
Cutaneous carcinoma of posterior scalp and upper neck II – V, Post auricular, Suboccipital
Cutaneous malignancy from pre-auricular, anterior scalp and temporal region II, III, VA, parotid, facial, external jugular nodes
Cutaneous malignancy of anterior or lateral face I, II, III
A SND should include at least 6 nodes on pathological examination. If nodes appear involved near the lower end of the dissection, the SND should be converted to a formal MRND. Alternatively, if facilities are available, frozen section can be employed.
SND is traditionally looked upon as a diagnostic or staging procedure. Increasing evidence indicates that the procedure could definitely be therapeutic, at least in N0 and early N+ disease, with results comparable to RND [40]. The therapeutic efficacy of SND is not diminished in the presence of occult nodal disease. However, other evidence indicates that recurrence rates are higher if post-operative irradiation is not used when one or more nodes are pathologically positive after SND [78]. SND followed by irradiation if nodal metastases are detected by pathology later is one area under investigation.
Trends in management of patients with N0 neck status
Sentinel node biopsy (SNB) This new concept is an in vivo assessment of tumoral spread to lymph nodes and could be useful in patients with N0 neck status where the problem is whether to treat, or wait and watch. The procedure is similar to that followed for SNB in patients with breast cancer or melanoma, the drawbacks are also similar. The SNB is performed after radiocolloid and blue dye injection (either alone or in combination). Preoperative lymphoscintigraphy and peri-operative use of a gamma probe identifies radioactive sentinel nodes and visualization of blue stained lymphatics and identify the blue sentinel node(s). No further nodal dissection is done if the SNB is negative, whilst therapeutic neck dissection is performed when the SNB is found to harbor malignancy. Various studies have illustrated the feasibility of the procedure in patients with head and neck cancers [79,80]. The sensitivity of the procedure is reported to be 94%. The argument for undertaking such a procedure is that it does not result in an excessive form of treatment whilst at the same time it does not neglect the oncological safety aspects.
A suggestion has been made to divide the occult metastases seen on microscopy in these nodes into isolated tumoral cells, micrometastases and macrometastases. Preliminary results of one trial shows that SNB is of value in patients with T1 and T2 tumors of the oral cavity and oropharynx [81]. Although the procedure holds promise for the future, more randomized studies and long-term follow-up are needed before this procedure is accepted as the standard of care [81-84]. Unresolved issues include identification of patient and tumor characteristics appropriate for this methodology, for example, stage, site or subsite [85]. Another important issue which has to be addressed is whether immunohistochemical or molecular studies are required on SNB and the clinical implication of such micrometastases.
Endoscopic neck dissections
With the advent of minimally invasive surgery, and after its success in certain abdominal conditions, the same has been tried in other areas of cancer management. Early experiments on animals and cadavers have been fairly successful [86]. Although the feasibility of the procedure has been shown, it's oncological safety and useful in practice are other issues that need to be addressed. The questions to be answered include whether such a procedure benefits the patient to the extent other endoscopic procedures have [87]. It remains to be seen if the procedure gains widespread acceptance in clinical practice.
Positron emission tomography (PET) and single photon emission computed tomography (SPECT)
PET may be able to detect nodal metastases in nodes that are judged to be negative by size criteria. SPECT imaging with deoxy-18F-fluoro-D-glucose or thallium is able to detect nodal disease. SPECT using Tc-99 m labeled monoclonal antibodies can visualize most lymph node metastases. However, the accuracy is superior when combined with CT scanning [87]. PET is at present prohibitively expensive but in future it may have a role in detecting occult nodal disease. Further studies are required to correlate the scan results with histological results of node examinations.
Conclusion
Neck dissection is the standard therapy for patients with metastatic neck nodes. RND has now been replaced with MRND for selected cases of patients with node positive neck disease. The improvements in understanding the biology of cancer and techniques in surgery have led to this change with its associated reduction in morbidity. Imaging studies to predict histological nodal metastases have so far not been successful. Evaluation is underway to replace MRND with SND for early nodal disease. SNB may replace SND for patients with N0 disease in future but certain issues still need to be addressed. Finally, whether endoscopic neck dissections will replace the existing procedures remains to be evaluated.
List of abbreviations
RND: Radical Neck Dissection
MRND: Modified Radical Neck Dissection
END: Elective Nodal Dissection
SND: Selective Neck Dissection
SNB: Sentinel Node Biopsy
ENI: Elective Nodal Irradiation
SOND: Supra Omohyoid Neck Dissection
PET: Positron Emission Tomography
SPECT: Single Photon Emission Computerized Tomography
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KH Conceptualized organized the textual material after the literature search and has written the manuscript.
==== Refs
Crile G Excision of cancer of the head and neck: with special reference to the plan of dissection based on 132 operations JAMA 1906 47 1780 1785
Martin H Del Valle B Ehrlich H Cahan WG Neck dissection Cancer 1951 4 441 499 14839606
Gavilan J Gavilan C Herranz J Functional neck dissection: three decades of controversy Ann Otol Rhinol Laryngol 1992 101 339 341 1562139
Bocca E Conservative neck dissection Laryngoscope 1975 85 1511 1515 1177643
Bocca E Pignataro O Oldini C Cappa C Functional neck dissection: an evaluation and review of 843 cases Laryngoscope 1984 94 942 945 6738274
Shah JP Patterns of cervical lymph node metastasis from squamous carcinomas of the upper aerodigestive tract Am J Surg 1990 160 405 409 2221244
Robbins KT Medina JE Wolfe GT Levine PA Sessions RB Pruet CW Standardizing neck dissection terminology. Official report of the Academy's Committee for Head and Neck Surgery and Oncology Arch Otolaryngol Head Neck Surg 1991 117 601 605 2036180
Robbins KT Clayman G Levine PA Medina J Sessions R Shaha A Som P Wolf GT Neck dissection classification update: revisions proposed by the American Head and Neck Society and the American Academy of Otolaryngology-Head and Neck Surgery Arch Otolaryngol Head Neck Surg 2002 128 751 758 12117328
Greene FL Page DL Fleming ID Fritz AG Balch CM Haller DG Morrow M AJCC Cancer Staging Manual 2002 6th New York , Springer Verlag
Shah JP Head and Neck Surgery 1996 2nd Barcelona , Msoby-Wolfe 355 392 8821817
Samant S Robbins KT Evolution of neck dissection for improved functional outcome World J Surg 2003 27 805 810 14509511 10.1007/s00268-003-7113-6
Grandi C Alloisio M Moglia D Podrecca S Sala L Salvatori P Molinari R Prognostic significance of lymphatic spread in head and neck carcinomas: therapeutic implications Head Neck Surg 1985 8 67 73 4077553
Spiro RH Alfonso AE Farr HW Strong EW Cervical node metastasis from epidermoid carcinoma of the oral cavity and oropharynx. A critical assessment of current staging Am J Surg 1974 128 562 567 4421032 10.1016/0002-9610(74)90276-1
Snow GB Patel P Leemans CR Tiwari R Management of cervical lymph nodes in patients with head and neck cancer Eur Arch Otorhinolaryngol 1992 249 187 194 1642874 10.1007/BF00178467
Shaha AR Radical neck dissection Operative Techniques in General Surgery 2004 6 72 82 10.1053/j.optechgensurg.2004.05.008
Soo KC Guiloff RJ Oh A Della Rovere GQ Westbury G Innervation of the trapezius muscle: a study in patients undergoing neck dissections Head Neck 1990 12 488 495 2258287
Guo CB Zhang Y Zhang L Zou LD [Surgical anatomy and preservation of the accessory nerve in radical functional neck dissection] Zhonghua Kou Qiang Yi Xue Za Zhi 2003 38 12 15 12760768
Karuman PM Soo KC Motor innervation of the trapezius muscle: a histochemical study Head Neck 1996 18 254 258 8860767
Nori S Soo KC Green RF Strong EW Miodownik S Utilization of intraoperative electroneurography to understand the innervation of the trapezius muscle Muscle Nerve 1997 20 279 285 9052805 10.1002/(SICI)1097-4598(199703)20:3<279::AID-MUS3>3.0.CO;2-8
Coskun HH Erisen L Basut O Selective neck dissection for clinically N0 neck in laryngeal cancer: is dissection of level IIb necessary? Otolaryngol Head Neck Surg 2004 131 655 659 15523444 10.1016/j.otohns.2004.04.014
Brown H Hidden G Ledroux M Poitevan L Anatomy and blood supply of the lower four cranial and cervical nerves: relevance to surgical neck dissection Proc Soc Exp Biol Med 2000 223 352 361 10721004 10.1046/j.1525-1373.2000.22350.x
Kierner AC Burian M Bentzien S Gstoettner W Intraoperative electromyography for identification of the trapezius muscle innervation: clinical proof of a new anatomical concept Laryngoscope 2002 112 1853 1856 12368629 10.1097/00005537-200210000-00028
Brandenburg JH Lee CY The eleventh nerve in radical neck surgery Laryngoscope 1981 91 1851 1859 7300535
Andre P Laccourreye H Haguet JF Beauvillain de Montreuil C Henin JM Aubert JD [Comparison of long-term carcinologic results between radical and conservative cervical surgery] Ann Otolaryngol Chir Cervicofac 1975 92 113 126 1217804
Andersen PE Shah JP Cambronero E Spiro RH The role of comprehensive neck dissection with preservation of the spinal accessory nerve in the clinically positive neck Am J Surg 1994 168 499 502 7977984
Bocca E Pignataro O A conservation technique in radical neck dissection Ann Otol Rhinol Laryngol 1967 76 975 987 6074244
Chu W Strawitz JG Results in suprahyoid, modified radical, and standard radical neck dissections for metastatic squamous cell carcinoma: recurrence and survival Am J Surg 1978 136 512 515 707733 10.1016/0002-9610(78)90272-6
Jesse RH Ballantyne AJ Larson D Radical or modified neck dissection: a therapeutic dilemma Am J Surg 1978 136 516 519 707734 10.1016/0002-9610(78)90273-8
Lingeman RE Helmus C Stephens R Ulm J Neck dissection: radical or conservative Ann Otol Rhinol Laryngol 1977 86 737 744 596770
Molinari R Cantu G Chiesa F Grandi C Retrospective comparison of conservative and radical neck dissection in laryngeal cancer Ann Otol Rhinol Laryngol 1980 89 578 581 7458153
Skolnik EM Tenta LT Wineinger DM Tardy MEJ Preservation of XI cranial nerve in neck dissections Laryngoscope 1967 77 1304 1314 6034855
Stell PM Dalby JE Singh SD Taylor W The fixed cervical lymph node Cancer 1984 53 336 341 6690015
Kennedy JT Krause CJ Loevy S The importance of tumor attachment to the carotid artery Arch Otolaryngol 1977 103 70 73 836232
Kornblut AD Shumrick DA Complications of head and neck surgery Arch Otolaryngol 1971 94 246 254 5094537
Snyderman CH D'Amico F Outcome of carotid artery resection for neoplastic disease: a meta-analysis Am J Otolaryngol 1992 13 373 380 1343570 10.1016/0196-0709(92)90079-9
Freeman SB Hamaker RC Borrowdale RB Huntley TC Management of neck metastasis with carotid artery involvement Laryngoscope 2004 114 20 24 14709989 10.1097/00005537-200401000-00003
Jones AS Stell PM Is laterality important in neck node metastases in head and neck cancer? Clin Otolaryngol 1991 16 261 265 1879068
Dulguerov P Soulier C Maurice J Faidutti B Allal AS Lehmann W Bilateral radical neck dissection with unilateral internal jugular vein reconstruction Laryngoscope 1998 108 1692 1696 9818828 10.1097/00005537-199811000-00019
Ridge JA Squamous cancer of the head and neck: surgical treatment of local and regional recurrence Semin Oncol 1993 20 419 429 7692603
Medina JE Byers RM Supraomohyoid neck dissection: rationale, indications, and surgical technique Head Neck 1989 11 111 122 2722487
Traynor SJ Cohen JI Gray J Andersen PE Everts EC Selective neck dissection and the management of the node-positive neck Am J Surg 1996 172 654 657 8988670 10.1016/S0002-9610(96)00296-6
Muzaffar K Therapeutic selective neck dissection: a 25-year review Laryngoscope 2003 113 1460 1465 12972914 10.1097/00005537-200309000-00005
McHam SA Adelstein DJ Rybicki LA Lavertu P Esclamado RM Wood BG Strome M Carroll MA Who merits a neck dissection after definitive chemoradiotherapy for N2-N3 squamous cell head and neck cancer? Head Neck 2003 25 791 798 12966502 10.1002/hed.10293
Chan SW Mukesh BN Sizeland A Treatment outcome of N3 nodal head and neck squamous cell carcinoma Otolaryngol Head Neck Surg 2003 129 55 60 12869917 10.1016/S0194-5998(03)00477-7
Wanebo H Chougule P Ready N Safran H Ackerley W Koness RJ McRae R Nigri P Leone L Radie-Keane K Reiss P Kennedy T Surgical resection is necessary to maximize tumor control in function-preserving, aggressive chemoradiation protocols for advanced squamous cancer of the head and neck (stage III and IV) Ann Surg Oncol 2001 8 644 650 11569779
Hehr T Classen J Schreck U Glocker S Koitschev A Bamberg M Budach W Selective lymph node dissection following hyperfractionated accelerated radio-(chemo-)therapy for advanced head and neck cancer Strahlenther Onkol 2002 178 363 368 12163990 10.1007/s00066-002-0965-0
Clayman GL Johnson CJ Morrison W Ginsberg L Lippman SM The role of neck dissection after chemoradiotherapy for oropharyngeal cancer with advanced nodal disease Arch Otolaryngol Head Neck Surg 2001 127 135 139 11177029
Argiris A Stenson KM Brockstein BE Mittal BB Pelzer H Kies MS Jayaram P Portugal L Wenig BL Rosen FR Haraf DJ Vokes EE Neck dissection in the combined-modality therapy of patients with locoregionally advanced head and neck cancer Head Neck 2004 26 447 455 15122662 10.1002/hed.10394
Carinci F Cassano L Farina A Pelucchi S Calearo C Modugno V Nielsen I Api P Pastore A Unresectable primary tumor of head and neck: does neck dissection combined with chemoradiotherapy improve survival? J Craniofac Surg 2001 12 438 443 11572248 10.1097/00001665-200109000-00007
Yao M Graham MM Hoffman HT Smith RB Funk GF Graham SM Dornfeld KJ Skwarchuk M Menda Y Buatti JM The role of post-radiation therapy FDG PET in prediction of necessity for post-radiation therapy neck dissection in locally advanced head-and-neck squamous cell carcinoma Int J Radiat Oncol Biol Phys 2004 59 1001 1010 15234033 10.1016/j.ijrobp.2004.01.040
Garg M Beitler JJ Controversies in management of the neck in head and neck cancer Curr Treat Options Oncol 2004 5 35 40 14697155
Fagan JJ Collins B Barnes L D'Amico F Myers EN Johnson JT Perineural invasion in squamous cell carcinoma of the head and neck Arch Otolaryngol Head Neck Surg 1998 124 637 640 9639472
Olsen KD Caruso M Foote RL Stanley RJ Lewis JE Buskirk SJ Frassica DA DeSanto LW O'Fallon WM Hoverman VR Primary head and neck cancer. Histopathologic predictors of recurrence after neck dissection in patients with lymph node involvement Arch Otolaryngol Head Neck Surg 1994 120 1370 1374 7980903
Byers RM Modified neck dissection. A study of 967 cases from 1970 to 1980 Am J Surg 1985 150 414 421 4051103 10.1016/0002-9610(85)90146-1
Byers RM Wolf PF Ballantyne AJ Rationale for elective modified neck dissection Head Neck Surg 1988 10 160 167 3235344
Collins SL Thawley SE, Panje WR, Batsakis JG, Lindberg RD Controversies in the management of cancer of the neck Comprehensive management of head and neck tumors 1999 2 2nd Philadelphia , WB Saunders Company 1479 1563
Pendergrast WJJ Soloway MS Myers GH Futrell JW Regional lymphadenectomy and tumor immunity Surg Gynecol Obstet 1976 142 385 390 1251321
Crile GJ The effect on metastasis of removing or irradiating regional nodes of mice Surg Gynecol Obstet 1968 126 1270 1272 5652665
Nahum AM Bone RC Davidson TM The case for elective prophylactic neck dissection Laryngoscope 1977 87 588 599 839943
Reed GF Miller WA Elective neck dissection--1970 Laryngoscope 1970 80 1292 1304 5469232
Reiss CK Volenec FJ Humphrey M Singla O Humphrey LJ The role of the regional lymph node in breast cancer: a comparison between nodal and systemic reactivity J Surg Oncol 1983 22 249 253 6834845
Sako K Pradier RN Marchetta FC Pickren JW Fallibility Of Palpation In The Diagnosis Of Metastases To Cervical Nodes Surg Gynecol Obstet 1964 118 989 990 14143467
Haberal I Celik H Gocmen H Akmansu H Yoruk M Ozeri C Which is important in the evaluation of metastatic lymph nodes in head and neck cancer: palpation, ultrasonography, or computed tomography? Otolaryngol Head Neck Surg 2004 130 197 201 14990916 10.1016/j.otohns.2003.08.025
Stevens MH Harnsberger HR Mancuso AA Davis RK Johnson LP Parkin JL Computed tomography of cervical lymph nodes. Staging and management of head and neck cancer Arch Otolaryngol 1985 111 735 739 4051864
van den Brekel MW Stel HV Castelijns JA Croll GJ Snow GB Lymph node staging in patients with clinically negative neck examinations by ultrasound and ultrasound-guided aspiration cytology Am J Surg 1991 162 362 366 1951890 10.1016/0002-9610(91)90149-8
Pitman KT Johnson JT Myers EN Effectiveness of selective neck dissection for management of the clinically negative neck Arch Otolaryngol Head Neck Surg 1997 123 917 922 9305240
Dias FL Kligerman J Matos de Sa G Arcuri RA Freitas EQ Farias T Matos F Lima RA Elective neck dissection versus observation in stage I squamous cell carcinomas of the tongue and floor of the mouth Otolaryngol Head Neck Surg 2001 125 23 29 11458209 10.1067/mhn.2001.116188
Kligerman J Lima RA Soares JR Prado L Dias FL Freitas EQ Olivatto LO Supraomohyoid neck dissection in the treatment of T1/T2 squamous cell carcinoma of oral cavity Am J Surg 1994 168 391 394 7977957
Fletcher GH Lucy Wortham James Lecture. Subclinical disease Cancer 1984 53 1274 1284 6692318
Chow JM Levin BC Krivit JS Applebaum EL Radiotherapy or surgery for subclinical cervical node metastases Arch Otolaryngol Head Neck Surg 1989 115 981 984 2751859
McGuirt WFJ Johnson JT Myers EN Rothfield R Wagner R Floor of mouth carcinoma. The management of the clinically negative neck Arch Otolaryngol Head Neck Surg 1995 121 278 282 7873143
Byers RM Newman R Russell N Yue A Results of treatment for squamous carcinoma of the lower gum Cancer 1981 47 2236 2238 7226119
Spiro RH Huvos AG Wong GY Spiro JD Gnecco CA Strong EW Predictive value of tumor thickness in squamous carcinoma confined to the tongue and floor of the mouth Am J Surg 1986 152 345 350 3766861 10.1016/0002-9610(86)90302-8
Byers RM Weber RS Andrews T McGill D Kare R Wolf P Frequency and therapeutic implications of "skip metastases" in the neck from squamous carcinoma of the oral tongue Head Neck 1997 19 14 19 9030939 10.1002/(SICI)1097-0347(199701)19:1<14::AID-HED3>3.0.CO;2-Y
Candela FC Kothari K Shah JP Patterns of cervical node metastases from squamous carcinoma of the oropharynx and hypopharynx Head Neck 1990 12 197 203 2358329
Johnson JT Bacon GW Myers EN Wagner RL Medial vs lateral wall pyriform sinus carcinoma: implications for management of regional lymphatics Head Neck 1994 16 401 405 7960736
Shah JP Tollefsen HR Epidermoid carcinoma of the supraglottic larynx. Role of neck dissection in initial surgical treatment Am J Surg 1974 128 494 499 4420265 10.1016/0002-9610(74)90262-1
Byers RM Clayman GL McGill D Andrews T Kare RP Roberts DB Goepfert H Selective neck dissections for squamous carcinoma of the upper aerodigestive tract: patterns of regional failure Head Neck 1999 21 499 505 10449664 10.1002/(SICI)1097-0347(199909)21:6<499::AID-HED1>3.0.CO;2-A
Stoeckli SJ Steinert H Pfaltz M Schmid S Sentinel lymph node evaluation in squamous cell carcinoma of the head and neck Otolaryngol Head Neck Surg 2001 125 221 226 11555757 10.1067/mhn.2001.118074
Chikamatsu K Kamada H Ninomiya H Takahashi K Sakurai T Oriuchi N Furuya N A preliminary study on sentinel lymph node biopsy: feasibility and predictive ability in oral cavity cancer Ann Nucl Med 2004 18 257 262 15233288
Ross GL Soutar DS Gordon MacDonald D Shoaib T Camilleri I Roberton AG Sorensen JA Thomsen J Grupe P Alvarez J Barbier L Santamaria J Poli T Massarelli O Sesenna E Kovacs AF Grunwald F Barzan L Sulfaro S Alberti F Sentinel node biopsy in head and neck cancer: preliminary results of a multicenter trial Ann Surg Oncol 2004 11 690 696 15197011 10.1245/ASO.2004.09.001
Ross GL Soutar DS MacDonald DG Shoaib T Camilleri IG Robertson AG Improved staging of cervical metastases in clinically node-negative patients with head and neck squamous cell carcinoma Ann Surg Oncol 2004 11 213 218 14761927 10.1245/ASO.2004.03.057
Mozzillo N Chiesa F Caraco C Botti G Lastoria S Longo F Ionna F Therapeutic implications of sentinel lymph node biopsy in the staging of oral cancer Ann Surg Oncol 2004 11 263S 6S 15023765 10.1245/ASO.2004.12.910
Rigual NR Wiseman SM Neck dissection: current concepts and future directions Surg Oncol Clin N Am 2004 13 151 166 15062367 10.1016/S1055-3207(03)00119-4
Dulguerov P Leuchter I Szalay-Quinodoz I Allal AS Marchal F Lehmann W Fasel JH Endoscopic neck dissection in human cadavers Laryngoscope 2001 111 2135 2139 11802011 10.1097/00005537-200112000-00010
Richtsmeier WJ Dissecting the "endoscopic neck" Arch Otolaryngol Head Neck Surg 2003 129 612 12810461 10.1001/archotol.129.6.612
Beyer T Townsend DW Brun T Kinahan PE Charron M Roddy R Jerin J Young J Byars L Nutt R A combined PET/CT scanner for clinical oncology J Nucl Med 2000 41 1369 1379 10945530
| 15836786 | PMC1097761 | CC BY | 2021-01-04 16:39:04 | no | World J Surg Oncol. 2005 Apr 18; 3:21 | utf-8 | World J Surg Oncol | 2,005 | 10.1186/1477-7819-3-21 | oa_comm |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.