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
CoughCough (London, England)1745-9974BioMed Central London 1745-9974-1-81627093510.1186/1745-9974-1-8Case ReportA rare cause of specific cough in a child: the importance of following-up children with chronic cough Barr Richard Lloyd [email protected] David John [email protected] Christopher Francis [email protected] Anne B [email protected] Senior Resident, Royal Children's Hospital, Herston Rd, Brisbane, Qld 4029, Australia2 ENT Registrar, Royal Children's Hospital, Herston Rd, Brisbane, Qld 4029, Australia3 Consultant in ENT Surgery, Royal Children's Hospital, Brisbane; Herston Rd, Brisbane, Qld 4029, Australia4 Consultant Respiratory Physician, Dept of Respiratory Medicine, Royal Children's Hospital, Brisbane; Herston Rd, Brisbane, Qld 4029, Australia; and A/Professor of Paediatrics, University of Queensland, Herston Rd, Brisbane, Australia2005 21 9 2005 1 8 8 13 7 2005 21 9 2005 Copyright © 2005 Barr et al; licensee BioMed Central Ltd.2005Barr 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.
For many years, the term 'specific cough' has been used as a clinical cough descriptor in children to signify the likelihood of an underlying disease causing the cough. In this case study, we describe a child with specific cough caused by a rare carcinoma, a mucoepidermoid carcinoma of the bronchus. The cough only totally resolved after the primary cause was successfully treated. This report highlights the importance of following up children with cough, especially those with specific cough.
==== Body
Clinical Record
An 8-year-old girl from a remote Aboriginal community approximately 2500 km from Brisbane was transferred to our hospital for management of a bronchial lesion. She had received 7-days of intravenous amoxicillin prior to transfer. She had a 4-year history of daily wet and sometimes productive cough, which was worse on exertion. There was no history of exertional dyspnoea, haemoptysis or weight loss. She also had a history of recurrent admissions for pneumonia at the local hospital (3 in the past 6 months). In the child's community, two adults were recently diagnosed with active pulmonary tuberculosis.
On arrival, the child was thin (weight 5th percentile, height 25th), appeared well and had a wet cough, reduced air entry over the right side and inspiratory crepitations. Spirometry values were invalid as she could not adequately perform maximum expiratory manoeuvres. Chest x-ray (CXR) showed right upper lobe (RUL) collapse, tram-tracks signs and increased peribronchial and interstitial markings of the right lower lobe. These CXR changes were documented at least 4-months ago (figures 1 and 2). Chest high resolution computerised tomography (CT) scan revealed RUL collapse and severe cystic bronchiectasis and cylindrical bronchiectasis of the right middle and lower lobes (figures 3 and 4). Sputum cultures grew Moraxella catarrhalis, and the microscopy was negative for acid-fast bacilli. Mantoux tests (M. tuberculum, M. Avium) were negative, sweat test and immunological workup were normal. Flexible bronchoscopy revealed a large lesion at the carina (Figure 5). Rigid bronchoscopy was then immediately performed during which the lesion was only partially removed piecemeal because of the presumed diagnosis of tuberculosis and length of time required to remove the bulk of the lesion (2-hours). Given the significant tuberculosis contact, anti-tuberculous medications were commenced and later ceased when cultures and Quantiferon test were negative. Histology showed a subepithelial neoplasm comprising glandular and solid areas with no evidence of significant mitotic activity or atypia, consistent with a low-grade muco-epidermoid carcinoma (MEC). Cytogenetic investigation on the tumour was not performed. Chest and abdomen CT scans revealed no metastases. Bronchoscopy was repeated and the remaining small lesions were biopsied. Right upper lobectomy and lymph node sampling was then performed and histological examination of the operative specimen demonstrated a small amount of residual tumour (with clear resection margins) and bronchiectasis. No metastases were found in the sampled lymph nodes. Postoperative progress was uneventful and the child was discharged 9-days later and was cough free. When reviewed 4 months post-discharge, she remained cough free and a repeat flexible bronchoscopy then confirmed the absence of any bronchial lesion or secretions.
Figure 1 Chest x-ray of the child 4 months before referral. The CXR shows collapse and tram tracks of the right upper lobe and increased peribronchial and interstitial markings of the right lower lobe.
Figure 2 CXR of child from referral hospital showing minimal increased changes from CXR taken 4 months ago.
Figure 3 Representative high resolution CT chest slices demonstrating collapse and severe bronchiectasis of the right upper lobe.
Figure 4 Representative high resolution CT chest slices demonstrating 'mild' bronchiectasis of the right lower lobe with partial collapse of right middle lobe. Bronchiectasis also present in the right middle lobe is not clearly demonstrated here.
Figure 5 Bronchoscopic picture of the carina prior to bronchoscopic partial removal of the tumour. The mucoepidermoid carcinoma that arose from the right upper lobe bronchus was so large it protruded into and obstructed the entire right main stem and is clearly visible at the carina (large arrow). The left main bronchus (small thick arrow) is partially occluded by secretions.
Discussion
We have described a child with several features of chronic specific cough caused by suppurative lung disease secondary to a rare life threatening lesion, a mucoepidermoid carcinoma obstructing a major bronchus. The child's cough only totally resolved upon removal of the tumour; i.e. after the primary cause was successfully treated. This report illustrates the importance of following-up children with chronic cough. Cough was this child's only symptom that was consistently present between the child's recurrent hospitalisations.
Paediatric cough, unlike cough in adults, is generally classified for practical purposes into cough descriptors of 'non-specific' and 'specific' cough [1,2]. In children with wet cough, airway secretions are always present [3]. Wet cough is a feature of specific cough as children (especially young children), unlike adults, do not often expectorate sputum. Several features of specific cough were present in this child; specifically, daily moist or productive cough, recurrent pneumonia and abnormal auscultatory findings [1] were present. Thus she had specific cough pointers and, in ideal circumstances, clinicians would be cognisant that the cough is likely associated with an underlying respiratory problem and hence requires further workup and follow-up to define the aetiology. Also, in children, the recommended minimum investigations for any child with a chronic cough are a CXR and spirometry [4]. In this child, the CXR was clearly abnormal – another indicator that further follow-up and investigations are usually required. This child had clinical features of bronchiectasis for at least several months and most likely a few years before eventual diagnosis of the underlying cause of her cough and respiratory illness. Also, radiological evidence of bronchiectasis was present and was secondary to a low-grade MEC that caused obstructive bronchiectasis (hence chronic wet cough from suppurative lung disease) and recurrent pneumonia. Unfortunately, the bronchiectasis was not restricted to the RUL; the delay in diagnosis allowed growth of the tumour that was so large it obstructed the entire right main bronchus and lead to obstructive bronchiectasis of the right lung.
Lung carcinoma remains the most common cancer in adults but is very rare in children [5]. Pulmonary MEC are even more rare (only 53 paediatric reports) [6-8] and represent approximately 10% of paediatric pulmonary tumours [7]. Macroscopically, MEC appear as a polypoid mass extending into the lumen [6-9] which may appear similar to bronchial mycobacteria lesions (Figure 6). Definitive diagnosis requires tissue biopsy, usually taken at bronchoscopy [6,7]. Because MEC are covered by normal respiratory epithelium bronchial brushings are usually not diagnostic [7,10]. MEC is thought to arise from mucous glands in the submucosal layer of respiratory walls [8,11] and is phylogenetically similar to salivary gland tumours [10]. Cytogenetic analysis of MEC tumours have described the presence of translocation t(11;19) (q14-21;p12-13) [12]. MEC has an 'iceberg-like' tendency to extend partially into the airway lumen but may extend into surrounding lung parenchyma [7]. Histologically, these tumours consist of a mixture of epidermoid, mucous and intermediate cells and may be classified as low, intermediate or high grade, reflecting differing compositions of cell types, extent of mitosis, anaplasia, and morphological variance ranging from cystic through to solid in nature [7,8,10]. Low grade tumours, more common in children, predominantly consist of mucous cells with occasional intermediate cells, tend to be locally invasive and, are associated with long term survival [9]. Intermediate grade tumours are more solid with predominance of intermediate cells and occasional mucous cells [8]. High grade tumours, more common in adults have a poorer prognosis [6-8,10,11]. with metastatic spread via blood or lymphatics to skin, bone and pericardium [8]. In all but two of the reported paediatric cases including ours, MEC was found to be low grade, and these tumours were successfully resected with no recurrence on follow up [7,8]. Children with high grade tumour succumb early, with one report of a child with a high grade tumour who succumbed eight months after diagnosis [7].
Figure 6 Figure showing bronchial non-tuberculous mycobacterium lesion of right upper lobe subsegment from another child. Macroscopically MEC appear similar to bronchial tuberculosis and can only be confidently differentiated by histopathology. This non-indigenous child presented with a few months history of chronic cough.
Presentation of patients with MEC is unusual until some obstruction of the involved airway occurs [6-9]. Common presenting symptoms include cough, recurrent pneumonia, haemoptysis, wheeze, dyspnoea, fever, and chest pain [7,8,13]. The rarity of these tumours contributes to delays in diagnosis [7,8]. While a diagnostic delay of up to 20-months has been reported [8], the likely several years interval in this child seemed particularly noteworthy. Deficiencies in health resources available in remote regions are well documented [14]. Indigenous Australians comprise a significant subset of this population and are particularly afflicted by respiratory illness [15,16]. As many of the presenting respiratory symptoms have an infective cause, the diagnostic suspicion of carcinoma in this setting is potentially further reduced. While adverse outcomes may be minimal, delays in diagnosis could lead to increased and prolonged morbidity. This report highlights the need to clinically follow-up all children with chronic cough especially those with chronic specific cough. After successful treatment of the underlying cause, cough almost always resolves in children. In patients with chronic specific cough and/or other respiratory symptoms not responsive to standard medical therapy, further investigations that include radiology and, in selected children, bronchoscopy should be promptly initiated [4].
Acknowledgements
The authors are grateful to Dr. Peter Borzi and Dr. Morgan Windsor who expertly performed the lobectomy. We also thank Barry Dean who provided the digital images.
==== Refs
Chang AB Cough: are children really different to adults? Cough 2005 1 7 16270937 10.1186/1745-9974-1-7
Chang AB Chung FK, Widdicombe JG, Boushey HA Causes, assessment and measurement in children Cough: Causes, Mechanisms and Therapy 2003 London: Blackwell Science 57 73
Chang AB Eastburn MM Gaffney J Faoagali J Cox NC Masters IB Cough quality in children: a comparison of subjective vs. bronchoscopic findings Respir Res 2005 6 3 15638942 10.1186/1465-9921-6-3
Chang AB Asher MI A review of cough in children J Asthma 2001 38 299 309 11456383 10.1081/JAS-100002296
Parkin DM Bray F Ferlay J Pisani P Global Cancer Statistics, 2002 CA Cancer J Clin 2005 55 74 108 15761078
Anton-Pacheco J Jimenez MA Rodriguez-Peralto JL Cuadros J Berchi FJ Bronchial mucoepidermoid tumor in a 3-year-old child Pediatr Surg Int 1998 13 524 525 9716686 10.1007/s003830050390
Granata C Battistini E Toma P Balducci T Mattioli G Fregonese B Mucoepidermoid carcinoma of the bronchus: a case report and review of the literature Pediatr Pulmonol 1997 23 226 232 9094733 10.1002/(SICI)1099-0496(199703)23:3<226::AID-PPUL10>3.0.CO;2-9
Welsh JH Maxson T Jaksic T Shahab I Hicks J Tracheobronchial mucoepidermoid carcinoma in childhood and adolescence: case report and review of the literature Int J Pediatr Otorhinolaryngol 1998 45 265 273 9865445 10.1016/S0165-5876(98)00120-7
Torres AM Ryckman FC Childhood tracheobronchial mucoepidermoid carcinoma: a case report and review of the literature J Pediatr Surg 1988 23 367 370 3290425
Vadasz P Egervary M Mucoepidermoid bronchial tumors: a review of 34 operated cases Eur J Cardiothorac Surg 2000 17 566 569 10814920 10.1016/S1010-7940(00)00386-9
Yousem SA Hochholzer L Mucoepidermoid tumors of the lung Cancer 1987 60 1346 1352 3040215
Spence SH Barrett PM Turner CM Psychometric properties of the Spence Children's Anxiety Scale with young adolescents J Anxiety Disord 2003 17 605 625 14624814 10.1016/S0887-6185(02)00236-0
Vogelberg C Mohr B Fitze G Friedrich K Hahn G Roesner D Mucoepidermoid carcinoma as an unusual cause for recurrent respiratory infections in a child J Pediatr Hematol Oncol 2005 27 162 165 15750450 10.1097/01.mph.0000155120.50936.73
Cunningham J Diagnostic and therapeutic procedures among Australian hospital patients identified as Indigenous Med J Aust 2002 176 62
Chang AB Masel JP Boyce NC Torzillo PJ Respiratory morbidity in central Australian Aboriginal children with alveolar lobar abnormalities Med J Aust 2003 178 490 494 12741934
Chang AB Masel JP Boyce NC Wheaton G Torzillo PJ Non-CF bronchiectasis-clinical and HRCT evaluation Pediatr Pulmonol 2003 35 477 483 12746947 10.1002/ppul.10289
|
16270935
|
PMC1277010
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Cough. 2005 Sep 21; 1:8
|
utf-8
|
Cough
| 2,005 |
10.1186/1745-9974-1-8
|
oa_comm
|
==== Front
CoughCough (London, England)1745-9974BioMed Central London 1745-9974-1-91627093910.1186/1745-9974-1-9ReviewIdiopathic chronic cough: a real disease or a failure of diagnosis? McGarvey LPA [email protected] Department of Medicine, The Queen's University of Belfast, Grosvenor Road, Belfast BT126BJ, N Ireland, UK2005 23 9 2005 1 9 9 24 3 2005 23 9 2005 Copyright © 2005 McGarvey; licensee BioMed Central Ltd.2005McGarvey; 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.
Despite extensive diagnostic evaluation and numerous treatment trials, a number of patients remain troubled by a chronic and uncontrollable cough. Eosinophilic bronchitis, atopic cough and non-acid reflux have been recently added to the diagnostic spectrum for chronic cough. In some cases, failure to consider these conditions may explain treatment failure. However, a subset of patients with persisting symptoms may be regarded as having an idiopathic cough. These individuals are most commonly female, of postmenopausal age and frequently report viral upper respiratory tract infections as an initiating event. This paper seeks to explore the validity of idiopathic cough as a distinct clinical entity.
Coughidiopathicdiagnostic protocol
==== Body
Introduction
Despite considerable advance in the understanding of cough, the effective management of patients with a chronic cough can be difficult. For the patient, a cough which persists can be associated with considerable distress and impaired quality of life [1]. For the physician, failure to obtain a treatment response may lead to the mistaken belief that the cough is functional or psychogenic. There are a number of reasons why the cough may be difficult to treat. In some cases it may reflect an inadequate approach to diagnostic evaluation and failure to appreciate both pulmonary and extra pulmonary causes for chronic cough [2,3]. In other cases, trials of therapy may be of inadequate dose and of insufficient duration. However, an alternative explanation is that a distinct diagnostic entity exists, namely idiopathic cough [4]. If this is the case then almost nothing is known about the underlying pathophysiological processes responsible for this condition and at present there are no effective treatment options. This article seeks to examine the evidence for idiopathic cough as either a distinct diagnosis or simply the result of incomplete evaluation and inadequate courses of therapy.
Diagnostic protocols for chronic cough
The term 'idiopathic' comes from the Greek word idiopatheia and is defined in the Oxford English Dictionary as a 'disease not preceded or occasioned by another, or by any known cause' [5]. In the original description of cough evaluation and management by Irwin and colleagues, idiopathic cough was not described and indeed treatment failure was extremely rare [6]. Using a stepwise approach known as the anatomic diagnostic protocol, Irwin and colleagues reported that a cause for cough could be determined successfully in up to 98% of cases and was due to either cough variant asthma (CVA), rhinosinusitis associated with postnasal drip syndrome (PNDS) or gastro-oesophageal reflux disease (GORD) [6]. The subsequent experience from this group [7,8] and a number of others in hospital-based settings [9,10] has remained the same and the diagnostic protocol has been recommended by the American College of Chest Physicians in their clinical guidelines for the management of cough [11].
Although the systematic evaluation of both extrapulmonary and pulmonary causes for cough is widely held to be effective, doubt has been cast on the perception that the diagnostic triad of CVA, PNDS and GORD accounts for the almost all causes of chronic cough [12,13]. Despite adopting a comprehensive evaluation of patients referred with cough, many groups have reported diagnostic and treatment failure in anything from 12 – 42% of patients [14-16]. For some, this represents a population with idiopathic cough [16] but others suggest it reflects failed management [17]. Specifically, the failure to prescribe sedating antihistamines for postnasal drip syndromes [17] and the inadequate treatment of gastro-oesophageal reflux disease have been highlighted [18].
There are a number of possible explanations for the impressive treatment response described by Irwin and others. Firstly, it is probable that the original referral populations included patients with cough following a viral upper respiratory infection. It is now recognised that cough following an upper respiratory tract infection may persist beyond three weeks and only resolve spontaneously some weeks or months later. Therefore some of the 'treatment success' may merely have reflected the natural resolution of a prolonged post-viral cough. Secondly, many patients were prescribed older generation antihistamines, which have an imprecise pharmacological action but presumably exert most of their antitussive effect by crossing the blood-brain barrier and acting directly on the cough control centre within the brain. Crucially, response to such therapy tells us little about the aetiology of the cough. Finally, these original studies reported on short-term treatment outcomes and provided little information on the long-term treatment response. Initial treatment benefit may well diminish over time and the timing of patient follow-up may explain some of the variance in outcome described by different centres [19].
Failure to adequately treat cough
Current guidelines have recommended a combination of diagnostic testing and empirical trials in the management chronic cough [20]. Some authors have reported that the characteristics of a cough confer little diagnostic information [21] but in practice, prominent symptoms of an upper airway disorder or indigestion should prompt a treatment trial of anti-rhinitic therapy or anti-reflux therapy [20]. The question of how much and for how long of a specific treatment has yet to be unequivocally answered. This point is perhaps best illustrated in the management of GORD associated cough. Although lacking a strong evidence base, it may be necessary to embark on intensive courses of anti-reflux therapy, because in contrast to the symptoms of heartburn, which usually resolve after a few days treatment, improvement in cough seems to take much longer [18,22]. In one study, mean duration to treatment success was 179 days [18]. As a consequence, failure to comply with prolonged therapy and lifestyle changes may result in relapse and explain poor treatment success even in patients with a high suspicion of GORD associated cough [19].
Alternatively, some individuals on relatively high doses of acid suppression may exhibit relative proton pump therapy resistance. This is particularly the case with attempts to suppress proximal and laryngophayngeal reflux where despite single and higher dose treatment regimes, 44% of patients demonstrated abnormal levels of acid exposure on simultaneous oesophagel and laryngeal pH testing [23]. Finally, a minority of patients who fail adequate courses of acid suppressive therapy may ultimately require anti-reflux surgery [24]. This final observation has contributed to the growing appreciation that acid may not be the sole aggravating factor in gastric refluxate. Until recently, this concept of 'non-acid reflux' as a cause for cough had been infrequently considered. It will be discussed together with a number of other 'new causes for cough' in the subsequent section of this review.
New causes for cough
Given the extent of associated literature, it is barely conceivable that any respiratory physician is unaware of the most common associations with chronic cough, namely asthma, GORD and rhinosinusitis, more recently termed upper airway cough syndrome. In the last decade, a series of important observations have led to the appreciation of new diagnostic possibilities. Most importantly, the application of induced sputum in the evaluation of cough has led to the identification of eosinophilic airway syndromes [25]. These conditions are characterized by the presence of eosinophilic airway inflammation but crucially the absence of the airway dysfunction (airflow variability or bronchial hyperreactivity) normally attributed to asthma. The best-described condition is eosinophilic bronchitis, which may account for up to 15% of patients referred to hospital with chronic cough [26]. It frequently responds to inhaled corticosteroids, and as these are often prescribed empirically in the community the exact prevalence of this condition is unknown. More recently, a number of Japanese groups have described a syndrome of "Atopic Cough" [27]. These patients are atopic, have an isolated bronchodilator resistant cough and an eosinophilic tracheobronchitis. Like eosinophilic bronchitis, there is no evidence of airway hyperreactivity but in contrast, the cough does not respond to inhaled corticosteroids. Without adequate attention to the inflammatory characteristics of the airway, and reluctance to prescribe inhaled steroids to patients with normal airway function then either of these syndromes may be incorrectly labeled as having an idiopathic cough.
The concept of 'Non-acid reflux' has recently gained attention. Irwin and colleagues [24] reported on a group of 8 patients that had persistent cough despite total or near total acid suppression utilizing proton pump inhibitors, prokinetic agents and antireflux diet (omeprazole 20–80 mg p.o. daily and cisapride 40–80 mg p.o. daily). These 8 patients had 24 hour ambulatory oesophageal pH monitoring while on medical therapy, and in all patients the % of 24 hours spent at pH < 4.0 was zero or near zero. Despite this, all 8 patients underwent antireflux surgery with marked reduction in cough scores after surgery, which were maintained after 12 months of follow up. This study suggests antireflux surgery may improve cough that is resistant to medical therapy, and that the improvement is sustained. Acid reflux disease in patients with cough and GORD may be a misnomer since non-acid reflux may be responsible for cough in some patients (volume reflux with gastric enzymes, bile salts etc.) [28]. Thus failure to respond to antireflux therapy may not indicate an idiopathic chronic cough.
Finally an association between cough, GORD and a familial sensory neuropathy has recently been reported [29]. The locus for the particular gene appears to be located on chromosone 3. In a series of personal communications with other cough specialists, it would appear similar associations have been encountered suggesting such clinical features may represent a new cough syndrome.
The common and less common associations with cough must be rigorously excluded before a diagnosis of idiopathic cough can be assigned. None-the-less, this author firmly believes such a condition exists and it will be addressed in some detail in the following section.
Idiopathic cough as a distinct clinical entity
The accumulation of experience and information regarding idiopathic cough suggests a fairly well defined population of patients. The over-representation of women in the specialist cough clinic referral population is widely acknowledged, and the preponderance of females among idiopathic coughers is particularly striking. Some centers have reported female prevalence rates of more than 80% [14-16,30-33] (See table 1). Gender differences in health-related quality of life and as a consequence differences in health seeking behaviour is one explanation [34] but others have suggested a distinct clinical phenotype [4]. Typically the female patients are of peri- or post menopausal age, report a preceding upper respiratory tract infection (URTI) and have a heightened cough reflex to tussive stimuli [16]. These observations raise the possibility that sex hormones and viral URTIs in some way contribute to the development of an idiopathic cough in susceptible individuals.
Table 1 Characteristics of idiopathic cough patients attending specialist cough clinics
Number (% female) Mean age (SD) (years) Median cough duration (range) (months)
O'Connell F et al [14] 16(81%) 51(31–70)* 72 (12–240)
McGarvey L et al [15] 8(75%) 46(8) 19 (6–36)
Forsythe P et al [30] 6(66%) 47(13) 72(2–240)
Jatakanon A et al [31] 10(50%) 60(4) 60 (18)^
Birring SS et al [32] 25(72%) 55(3) 12 (7–360)
Chaudhuri R et al [33] 6(60%) 58(9) 14(19)^
Haque R et al [16] 31(76%) 57(32–81)* 72 (8–324)*
*Data given as median (range), ^Data given as mean (SD)
Possible mechanisms for idiopathic cough
The human cough reflex consists of an afferent arm comprising cough receptors, afferent pathways, central processing and an efferent pathway. The cough reflex can be modified at any point along this reflex and unraveling the mechanisms responsible is key to a more complete understanding of cough pathophysiology and its successful treatment. Afferent sensory nerves are not static entities and the term 'plasticity' has been used to describe changes in function contributing to the sensitization that occurs in response to various stimuli, in particular those associated with airway inflammatory processes [35]. Although viral infections are a major cause of cough and appear to be frequently reported in patients with idiopathic cough, little is known regarding the effects of viruses on cough sensitivity. Following respiratory syncytial virus infection in rats, tachykinin content within the lung is increased [36] along with an upregulation in the substance P receptor, neurokinin (NK) 1 [37]. These changes appear to persist for some time after the virus is cleared. In guinea pigs, inoculation with the Sendai virus has been associated with a qualitative change in airway sensory nerves whereby nonnociceptive neurons express tachykinins [38]. This 'phenotypic switch' is one plausible mechanism whereby viral infection causes increased tachykinergic content in airway nerves which possibly contribute to persistent reflex hypersensitivity and cough. It is unknown if such processes occur in man, but abnormal intraepithelial nerves containing increased neuropeptide content have been reported in bronchial biopsies from patients with idiopathic cough [39].
Only a few studies have specifically commented on findings in the airways of patients with idiopathic cough. Birring et al. observed a mild chronic lymphocytic airway inflammation in a predominately female population of idiopathic coughers and highlighted the striking association with organ specific autoimmune disease in particular hypothyroidism [40]. They suggested that the presence of increased lymphocytes within the airway reflected either an aberrant homing of lymphocytes from the primary site of autoimmune inflammation to the lung or a distinct autoimmune process within the lungs [40]. A more recent study has confirmed the dominance of lymphocytes in the airways of females with idiopathic cough. In this study, significantly elevated numbers of activated CD4+ lymphocytes were noted in bronchoalveolar lavage fluid from menopausal women with isolated dry cough compared to matched controls. This group hypothesized that menopausal effects on lymphocyte activation within the airway may lead to disordered responses to airway insults such as infection [41].
Gender and sex hormones may have important effects on neuro-immune events within the airway. A number of studies have demonstrated a heightened cough reflex sensitivity in females compared to males both in healthy individuals [42,43] and cough subjects [44]. This gender difference has not been observed in children, raising the possibility that sex hormones may influence the reflex [45]. Women of post-menopausal age appear to have a heightened cough reflex although this has not been consistently demonstrated [46]. None-the-less, oestrogen levels begin to decrease around the time of the menopause, which may exert an effect on cough reflex sensitivity. Danazol, a synthetic androgen that decreases oestrogen levels, has been shown to inhibit the upregulation of the cough reflex observed in female guinea pigs following treatment with an ACE-inhibitor [47].
Conclusion
Although inadequate management will continue to explain a significant number of patients with a chronic and uncontrollable cough, an attempt has been made in this article to highlight idiopathic cough as a distinct clinical entity. Although without firm evidence, idiopathic cough may arise as a consequence of the persisting effects of viral infection or other noxious aggravants in susceptible individuals. The excess of middle-aged females with idiopathic cough raises the possibility of some sex hormonal influence. Precision in this area will be greatly hampered unless further research is undertaken.
==== Refs
French CL Irwin RS Curley FJ Krikorian CJ Impact of chronic cough on quality of life Arch Intern Med 1998 158 1657 1661 9701100 10.1001/archinte.158.15.1657
Al-Mobeireek AF Al-Sarhani A Al-Amri S Bamgboye E Ahmed S Chronic cough at a non-teaching hospital: Are extrapulmonary causes overlooked? Respirology 2002 7 141 146 11985737 10.1046/j.1440-1843.2002.00378.x
McGarvey LPA Heaney LG MacMahon J A retrospective survey of diagnosis and management of patients presenting with chronic cough to a general chest clinic Int J Clin Pract 1997 52 158 161 9684430
McGarvey LPA Ing AJ Idiopathic cough, prevalence and underlying mechanisms Pulm Pharmacol Ther 2004 17 435 439 15564088 10.1016/j.pupt.2004.09.012
The Oxford English Dictionary 2 New York: Oxford University Press
Irwin RS Corrao WM Pratter MR Chronic persistent cough in the adult: the spectrum and frequency of cases and successful outcome of specific therapy Am Rev Respir Dis 1981 123 414 417
Irwin RS Curley FJ French CL Chronic cough: the spectrum and frequency of causes, key components of the diagnostic evaluation and outcome of specific therapy Am Rev Resp Dis 1990 141 640 647 2178528
Smyrnios NA Irwin RS Curley FJ Chronic cough with a history of excessive sputum production: The spectrum and frequency of causes key components of the diagnostic evaluation, and outcome of specific therapy: Chest 1995 108 991 997 7555175
Pratter MR Bartter T Akers S Dubois J An algorithmic approach to chronic cough Ann Intern Med 1993 119 977 83 8214994
Palombini BC Villanova CA Araujo E Gastal OL Alt DC Stolz DP Palombini CO A pathogenic triad in chronic cough: asthma, postnasal drip syndrome and gastrooesophageal reflux disease Chest 1999 116 279 8 10453852 10.1378/chest.116.2.279
Irwin RS Boulet LP Cloutier MM Fuller R Gold PM Hoffstein V Ing AJ McCool FD O'Byrne P Poe PH Prakash UB Pratter MR Rubin BK Managing cough as a defence mechanism and as a symptom. A consensus panel report of the American College of Chest Physicians Chest 1998 114 133S 181S 9725800
Poe HR Harder RV Israel RH Chronic persistent cough: experience in diagnosis and outcome using an anatomic diagnostic protocol Chest 1989 95 723 27 2924600
Morice AH Kastelik JA Cough. 1: Chronic cough in adults Thorax 2003 58 901 7 14514949 10.1136/thorax.58.10.901
O'Connell F Thomas VE Pride NB Fuller RW Cough sensitivity to inhaled capsaicin decreases with successful treatment of chronic cough Am J Respir Crit Care Med 1993 150 374 80 8049818
McGarvey LPA Heaney LG Lawson JT Johnston BT Scally CM Ennis M Shepherd DRT MacMahon J Evaluation and outcome of patients with chronic non-productive cough using a comprehensive diagnostic protocol Thorax 1998 53 738 743 10319055
Haque RA Usmani OS Barnes PJ Chronic Idiopathic cough: a discrete clinical entity? Chest 2005 127 1710 1713 15888850 10.1378/chest.127.5.1710
Irwin RS Madison JM Diagnosis and treatment of chronic cough due to gastro-esophageal reflux disease and postnasal drip syndrome Pulm Pharmacol Ther 2002 15 293 4 12099781 10.1006/pupt.2002.0345
Irwin RS Madison JM Anatomical diagnostic protocol in evaluating chronic cough with specific reference to gastro-oesophageal reflux disease Am J Med 2000 108 126S 130S 10718465 10.1016/S0002-9343(99)00351-4
Patterson RN Johnston BT MacMahon J Heaney LG McGarvey LPA Oesophageal pH monitoring is of limited value in the diagnosis of 'reflux-cough' Eur Respir J 2004 24 724 7 15516662 10.1183/09031936.04.00007404
Morice AH Fontana GA Sovijarvi ARA Pistolesi M Chung KF Widdicombe J ERS Task Force The diagnosis and management of cough Eur Respir J 2004 24 481 492 15358710 10.1183/09031936.04.00027804
Mello CJ Irwin RS Curley FJ The predictive values of the character, timing and complications of chronic cough in diagnosing its cause Arch Int Med 1993 119 997 983
Benini L Ferrari M Sembenini C Olivieri M Micciolo R Zuccali V Bulighin GM Fiorino F Ederle A Cascio VL Vantini I Cough threshold in reflux oesophagitis: influence of acid and of laryngeal and oesophageal damage Gut 2000 46 762 7 10807885 10.1136/gut.46.6.762
Amin MR Postma GN Johnston P Digges N Koufman JA Proton pump inhibitor resistance in the treatment of laryngopharyngeal reflux Otolaryngol Head Neck Surg 2001 125 374 8 11593175 10.1067/mhn.2001.118691
Irwin RS Zawacki JK Wilson MM French CT Callery MP Chronic cough due to gastro-oesophageal reflux disease: Failure to resolve despite total/near total elimination of oesophageal acid Chest 2002 121 1132 1140 11948043 10.1378/chest.121.4.1132
Gibson PG Dolovich J Denburg J Ramsdale EH Hargreave FE Chronic cough: eosinophilic bronchitis without asthma Lancet 1989 1 1346 8 2567371 10.1016/S0140-6736(89)92801-8
Brightling C Ward R Goh KL Wardlaw AJ Pavord ID Eosinophilic bronchitis is an important cause of chronic cough Am J Respir Crit Care Med 1999 160 406 10 10430705
Fujimura M Ogawa H Nishizawa Y Nishi K Comparison of atopic cough with cough variant asthma: Is atopic cough a precursor of asthma? Thorax 2003 58 14 18 12511712 10.1136/thorax.58.1.14
Sifrim D Dupont L Blondeau K Zhang X Tack J Janssens J Weakly acidic reflux in patients with chronic unexplained cough during 24 hour pressure, pH, and impedance monitoring Gut 2005 54 449 54 15753524 10.1136/gut.2004.055418
Kok C Kennerson ML Spring PJ Ing AJ Pollard JD Nicholson GA A locus for hereditary sensory neuropathy with cough and gastroesophageal reflux on chromosone 3p22-p24 Am J Hum Genet 2003 73 632 7 12870133 10.1086/377591
Forsythe P McGarvey L Heaney LG MacMahon J Ennis M Sensory neuropeptides induce histamine release from bronchoalveolar lavage cells in both non asthmatic coughers and cough variant asthmatics Clin Exper Allergy 2000 30 225 32 10651775 10.1046/j.1365-2222.2000.00770.x
Jatakanon A Lalloo UG Lim S Chung KF Barnes PJ Increased neutrophils and cytokines, TNF-alpha and IL-8, in induced sputum of non-asthmatic patients with chronic dry cough Thorax 1999 54 234 7 10325899
Birring SS Murphy AC Scullion JE Brightling CE Browning M Pavord ID Idiopathic chronic cough and organ specific autoimmune diseases: a case control study Respir Med 2004 98 242 6 15002760 10.1016/j.rmed.2003.10.005
Chaudhuri R McMahon AD Thomson LJ MacLeod KJ McSharry CP Livingston E McKay A Thomson NC Effect of inhaled corticosteroids on symptom severity and sputum mediator levels in chronic persistent cough J Allergy Clin Immunol 2004 113 1063 70 15208586 10.1016/j.jaci.2004.03.019
French CT Fletcher KE Irwin RS Gender differences in health-related quality of life in patients complaining of chronic cough Chest 2004 125 482 8 14769728 10.1378/chest.125.2.482
Carr MJ Plasticity of vagal afferent fibres mediating cough Pulm Pharmacol Ther 2004 17 447 51 15564090 10.1016/j.pupt.2004.09.020
Piedimonte G Hegele RG Auais A Persistent airway inflammation after resolution of respiratory syncytial virus infection in rats Pediatr Res 2004 55 657 65 14711892 10.1203/01.PDR.0000112244.72924.26
Hu C Wedde-Beer K Auais A Rodriguez MM Piedimonte G Nerve growth factor and nerve growth factor receptors in respiratory syncytial virus-infected lungs Am J Physiol Lung Cell Mol Physiol 2002 283 L494 502 12114213
Carr MJ Hunter DD Jacoby DB Undem BJ Expression of tachykinins in nonnociceptive vagal afferent neurons during respiratory viral infection in guinea pigs Am J Respir Crit Care Med 2002 161 1985 90
O'Connell F Springall DR Moradoghi-Haftvani A Krausz T Price D Fuller RW Polak JM Pride NB Abnormal intraepithelial airway nerves in persistent unexplained cough Am J Respir Crit Care Med 1995 152 2068 75 8520777
Birring SS Brightling DE Symon FA Barlow SG Wardlaw AJ Pavord ID Idiopathic chronic cough: association with organ specific autoimmune disease and bronchoalveolar lymphocytosis Thorax 2003 58 1066 70 14645977 10.1136/thorax.58.12.1066
Mund E Christensson B Gronneberg R Larsson K Noneosinophilic CD4 lymphocytic airway inflammation in menopausal women with chronic dry cough Chest 2005 127 1714 1721 15888851 10.1378/chest.127.5.1714
Fujimura M Sakamoto S Kamio Y Matsuda T Sex difference in the inhaled tartaric acid cough threshold in non-atopic healthy subjects Thorax 1990 45 633 4 2402729
Dicpinigaitis PV Rauf K The influence of gender on cough reflex sensitivity Chest 1998 113 1319 21 9596313
Kastelik JA Thompson RH Aziz I Ojoo JC Redington AE Morice AH Sex-related differences in cough reflex sensitivity in patients with chronic cough Am J Respir Crit Care Med 2002 166 961 4 12359654 10.1164/rccm.2109061
Chang AB Phelan PD Sawyer SM Del Brocco S Robertson CF Cough sensitivity in children with asthma, recurrent cough and cystic fibrosis Arch Dis Child 1997 77 331 4 9389238
Prudon B Birring SS Vara DD Hall AP Thompson JP Thompson JP Pavord ID Cough and glottic-stop reflex sensitivity in health and disease Chest 2005 127 550 7 15705995 10.1378/chest.127.2.550
Ebihara T Sekizawa K Ohtui T Nakzawa H Sasaki H Angiotensin-converting enzyme inhibitor and danazol increase sensitivity of cough reflex in female guinea pigs Am J Respir Crit Care Med 1996 153 812 9 8564137
|
16270939
|
PMC1277011
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Cough. 2005 Sep 23; 1:9
|
utf-8
|
Cough
| 2,005 |
10.1186/1745-9974-1-9
|
oa_comm
|
==== Front
CoughCough (London, England)1745-9974BioMed Central London 1745-9974-1-91627093910.1186/1745-9974-1-9ReviewIdiopathic chronic cough: a real disease or a failure of diagnosis? McGarvey LPA [email protected] Department of Medicine, The Queen's University of Belfast, Grosvenor Road, Belfast BT126BJ, N Ireland, UK2005 23 9 2005 1 9 9 24 3 2005 23 9 2005 Copyright © 2005 McGarvey; licensee BioMed Central Ltd.2005McGarvey; 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.
Despite extensive diagnostic evaluation and numerous treatment trials, a number of patients remain troubled by a chronic and uncontrollable cough. Eosinophilic bronchitis, atopic cough and non-acid reflux have been recently added to the diagnostic spectrum for chronic cough. In some cases, failure to consider these conditions may explain treatment failure. However, a subset of patients with persisting symptoms may be regarded as having an idiopathic cough. These individuals are most commonly female, of postmenopausal age and frequently report viral upper respiratory tract infections as an initiating event. This paper seeks to explore the validity of idiopathic cough as a distinct clinical entity.
Coughidiopathicdiagnostic protocol
==== Body
Introduction
Despite considerable advance in the understanding of cough, the effective management of patients with a chronic cough can be difficult. For the patient, a cough which persists can be associated with considerable distress and impaired quality of life [1]. For the physician, failure to obtain a treatment response may lead to the mistaken belief that the cough is functional or psychogenic. There are a number of reasons why the cough may be difficult to treat. In some cases it may reflect an inadequate approach to diagnostic evaluation and failure to appreciate both pulmonary and extra pulmonary causes for chronic cough [2,3]. In other cases, trials of therapy may be of inadequate dose and of insufficient duration. However, an alternative explanation is that a distinct diagnostic entity exists, namely idiopathic cough [4]. If this is the case then almost nothing is known about the underlying pathophysiological processes responsible for this condition and at present there are no effective treatment options. This article seeks to examine the evidence for idiopathic cough as either a distinct diagnosis or simply the result of incomplete evaluation and inadequate courses of therapy.
Diagnostic protocols for chronic cough
The term 'idiopathic' comes from the Greek word idiopatheia and is defined in the Oxford English Dictionary as a 'disease not preceded or occasioned by another, or by any known cause' [5]. In the original description of cough evaluation and management by Irwin and colleagues, idiopathic cough was not described and indeed treatment failure was extremely rare [6]. Using a stepwise approach known as the anatomic diagnostic protocol, Irwin and colleagues reported that a cause for cough could be determined successfully in up to 98% of cases and was due to either cough variant asthma (CVA), rhinosinusitis associated with postnasal drip syndrome (PNDS) or gastro-oesophageal reflux disease (GORD) [6]. The subsequent experience from this group [7,8] and a number of others in hospital-based settings [9,10] has remained the same and the diagnostic protocol has been recommended by the American College of Chest Physicians in their clinical guidelines for the management of cough [11].
Although the systematic evaluation of both extrapulmonary and pulmonary causes for cough is widely held to be effective, doubt has been cast on the perception that the diagnostic triad of CVA, PNDS and GORD accounts for the almost all causes of chronic cough [12,13]. Despite adopting a comprehensive evaluation of patients referred with cough, many groups have reported diagnostic and treatment failure in anything from 12 – 42% of patients [14-16]. For some, this represents a population with idiopathic cough [16] but others suggest it reflects failed management [17]. Specifically, the failure to prescribe sedating antihistamines for postnasal drip syndromes [17] and the inadequate treatment of gastro-oesophageal reflux disease have been highlighted [18].
There are a number of possible explanations for the impressive treatment response described by Irwin and others. Firstly, it is probable that the original referral populations included patients with cough following a viral upper respiratory infection. It is now recognised that cough following an upper respiratory tract infection may persist beyond three weeks and only resolve spontaneously some weeks or months later. Therefore some of the 'treatment success' may merely have reflected the natural resolution of a prolonged post-viral cough. Secondly, many patients were prescribed older generation antihistamines, which have an imprecise pharmacological action but presumably exert most of their antitussive effect by crossing the blood-brain barrier and acting directly on the cough control centre within the brain. Crucially, response to such therapy tells us little about the aetiology of the cough. Finally, these original studies reported on short-term treatment outcomes and provided little information on the long-term treatment response. Initial treatment benefit may well diminish over time and the timing of patient follow-up may explain some of the variance in outcome described by different centres [19].
Failure to adequately treat cough
Current guidelines have recommended a combination of diagnostic testing and empirical trials in the management chronic cough [20]. Some authors have reported that the characteristics of a cough confer little diagnostic information [21] but in practice, prominent symptoms of an upper airway disorder or indigestion should prompt a treatment trial of anti-rhinitic therapy or anti-reflux therapy [20]. The question of how much and for how long of a specific treatment has yet to be unequivocally answered. This point is perhaps best illustrated in the management of GORD associated cough. Although lacking a strong evidence base, it may be necessary to embark on intensive courses of anti-reflux therapy, because in contrast to the symptoms of heartburn, which usually resolve after a few days treatment, improvement in cough seems to take much longer [18,22]. In one study, mean duration to treatment success was 179 days [18]. As a consequence, failure to comply with prolonged therapy and lifestyle changes may result in relapse and explain poor treatment success even in patients with a high suspicion of GORD associated cough [19].
Alternatively, some individuals on relatively high doses of acid suppression may exhibit relative proton pump therapy resistance. This is particularly the case with attempts to suppress proximal and laryngophayngeal reflux where despite single and higher dose treatment regimes, 44% of patients demonstrated abnormal levels of acid exposure on simultaneous oesophagel and laryngeal pH testing [23]. Finally, a minority of patients who fail adequate courses of acid suppressive therapy may ultimately require anti-reflux surgery [24]. This final observation has contributed to the growing appreciation that acid may not be the sole aggravating factor in gastric refluxate. Until recently, this concept of 'non-acid reflux' as a cause for cough had been infrequently considered. It will be discussed together with a number of other 'new causes for cough' in the subsequent section of this review.
New causes for cough
Given the extent of associated literature, it is barely conceivable that any respiratory physician is unaware of the most common associations with chronic cough, namely asthma, GORD and rhinosinusitis, more recently termed upper airway cough syndrome. In the last decade, a series of important observations have led to the appreciation of new diagnostic possibilities. Most importantly, the application of induced sputum in the evaluation of cough has led to the identification of eosinophilic airway syndromes [25]. These conditions are characterized by the presence of eosinophilic airway inflammation but crucially the absence of the airway dysfunction (airflow variability or bronchial hyperreactivity) normally attributed to asthma. The best-described condition is eosinophilic bronchitis, which may account for up to 15% of patients referred to hospital with chronic cough [26]. It frequently responds to inhaled corticosteroids, and as these are often prescribed empirically in the community the exact prevalence of this condition is unknown. More recently, a number of Japanese groups have described a syndrome of "Atopic Cough" [27]. These patients are atopic, have an isolated bronchodilator resistant cough and an eosinophilic tracheobronchitis. Like eosinophilic bronchitis, there is no evidence of airway hyperreactivity but in contrast, the cough does not respond to inhaled corticosteroids. Without adequate attention to the inflammatory characteristics of the airway, and reluctance to prescribe inhaled steroids to patients with normal airway function then either of these syndromes may be incorrectly labeled as having an idiopathic cough.
The concept of 'Non-acid reflux' has recently gained attention. Irwin and colleagues [24] reported on a group of 8 patients that had persistent cough despite total or near total acid suppression utilizing proton pump inhibitors, prokinetic agents and antireflux diet (omeprazole 20–80 mg p.o. daily and cisapride 40–80 mg p.o. daily). These 8 patients had 24 hour ambulatory oesophageal pH monitoring while on medical therapy, and in all patients the % of 24 hours spent at pH < 4.0 was zero or near zero. Despite this, all 8 patients underwent antireflux surgery with marked reduction in cough scores after surgery, which were maintained after 12 months of follow up. This study suggests antireflux surgery may improve cough that is resistant to medical therapy, and that the improvement is sustained. Acid reflux disease in patients with cough and GORD may be a misnomer since non-acid reflux may be responsible for cough in some patients (volume reflux with gastric enzymes, bile salts etc.) [28]. Thus failure to respond to antireflux therapy may not indicate an idiopathic chronic cough.
Finally an association between cough, GORD and a familial sensory neuropathy has recently been reported [29]. The locus for the particular gene appears to be located on chromosone 3. In a series of personal communications with other cough specialists, it would appear similar associations have been encountered suggesting such clinical features may represent a new cough syndrome.
The common and less common associations with cough must be rigorously excluded before a diagnosis of idiopathic cough can be assigned. None-the-less, this author firmly believes such a condition exists and it will be addressed in some detail in the following section.
Idiopathic cough as a distinct clinical entity
The accumulation of experience and information regarding idiopathic cough suggests a fairly well defined population of patients. The over-representation of women in the specialist cough clinic referral population is widely acknowledged, and the preponderance of females among idiopathic coughers is particularly striking. Some centers have reported female prevalence rates of more than 80% [14-16,30-33] (See table 1). Gender differences in health-related quality of life and as a consequence differences in health seeking behaviour is one explanation [34] but others have suggested a distinct clinical phenotype [4]. Typically the female patients are of peri- or post menopausal age, report a preceding upper respiratory tract infection (URTI) and have a heightened cough reflex to tussive stimuli [16]. These observations raise the possibility that sex hormones and viral URTIs in some way contribute to the development of an idiopathic cough in susceptible individuals.
Table 1 Characteristics of idiopathic cough patients attending specialist cough clinics
Number (% female) Mean age (SD) (years) Median cough duration (range) (months)
O'Connell F et al [14] 16(81%) 51(31–70)* 72 (12–240)
McGarvey L et al [15] 8(75%) 46(8) 19 (6–36)
Forsythe P et al [30] 6(66%) 47(13) 72(2–240)
Jatakanon A et al [31] 10(50%) 60(4) 60 (18)^
Birring SS et al [32] 25(72%) 55(3) 12 (7–360)
Chaudhuri R et al [33] 6(60%) 58(9) 14(19)^
Haque R et al [16] 31(76%) 57(32–81)* 72 (8–324)*
*Data given as median (range), ^Data given as mean (SD)
Possible mechanisms for idiopathic cough
The human cough reflex consists of an afferent arm comprising cough receptors, afferent pathways, central processing and an efferent pathway. The cough reflex can be modified at any point along this reflex and unraveling the mechanisms responsible is key to a more complete understanding of cough pathophysiology and its successful treatment. Afferent sensory nerves are not static entities and the term 'plasticity' has been used to describe changes in function contributing to the sensitization that occurs in response to various stimuli, in particular those associated with airway inflammatory processes [35]. Although viral infections are a major cause of cough and appear to be frequently reported in patients with idiopathic cough, little is known regarding the effects of viruses on cough sensitivity. Following respiratory syncytial virus infection in rats, tachykinin content within the lung is increased [36] along with an upregulation in the substance P receptor, neurokinin (NK) 1 [37]. These changes appear to persist for some time after the virus is cleared. In guinea pigs, inoculation with the Sendai virus has been associated with a qualitative change in airway sensory nerves whereby nonnociceptive neurons express tachykinins [38]. This 'phenotypic switch' is one plausible mechanism whereby viral infection causes increased tachykinergic content in airway nerves which possibly contribute to persistent reflex hypersensitivity and cough. It is unknown if such processes occur in man, but abnormal intraepithelial nerves containing increased neuropeptide content have been reported in bronchial biopsies from patients with idiopathic cough [39].
Only a few studies have specifically commented on findings in the airways of patients with idiopathic cough. Birring et al. observed a mild chronic lymphocytic airway inflammation in a predominately female population of idiopathic coughers and highlighted the striking association with organ specific autoimmune disease in particular hypothyroidism [40]. They suggested that the presence of increased lymphocytes within the airway reflected either an aberrant homing of lymphocytes from the primary site of autoimmune inflammation to the lung or a distinct autoimmune process within the lungs [40]. A more recent study has confirmed the dominance of lymphocytes in the airways of females with idiopathic cough. In this study, significantly elevated numbers of activated CD4+ lymphocytes were noted in bronchoalveolar lavage fluid from menopausal women with isolated dry cough compared to matched controls. This group hypothesized that menopausal effects on lymphocyte activation within the airway may lead to disordered responses to airway insults such as infection [41].
Gender and sex hormones may have important effects on neuro-immune events within the airway. A number of studies have demonstrated a heightened cough reflex sensitivity in females compared to males both in healthy individuals [42,43] and cough subjects [44]. This gender difference has not been observed in children, raising the possibility that sex hormones may influence the reflex [45]. Women of post-menopausal age appear to have a heightened cough reflex although this has not been consistently demonstrated [46]. None-the-less, oestrogen levels begin to decrease around the time of the menopause, which may exert an effect on cough reflex sensitivity. Danazol, a synthetic androgen that decreases oestrogen levels, has been shown to inhibit the upregulation of the cough reflex observed in female guinea pigs following treatment with an ACE-inhibitor [47].
Conclusion
Although inadequate management will continue to explain a significant number of patients with a chronic and uncontrollable cough, an attempt has been made in this article to highlight idiopathic cough as a distinct clinical entity. Although without firm evidence, idiopathic cough may arise as a consequence of the persisting effects of viral infection or other noxious aggravants in susceptible individuals. The excess of middle-aged females with idiopathic cough raises the possibility of some sex hormonal influence. Precision in this area will be greatly hampered unless further research is undertaken.
==== Refs
French CL Irwin RS Curley FJ Krikorian CJ Impact of chronic cough on quality of life Arch Intern Med 1998 158 1657 1661 9701100 10.1001/archinte.158.15.1657
Al-Mobeireek AF Al-Sarhani A Al-Amri S Bamgboye E Ahmed S Chronic cough at a non-teaching hospital: Are extrapulmonary causes overlooked? Respirology 2002 7 141 146 11985737 10.1046/j.1440-1843.2002.00378.x
McGarvey LPA Heaney LG MacMahon J A retrospective survey of diagnosis and management of patients presenting with chronic cough to a general chest clinic Int J Clin Pract 1997 52 158 161 9684430
McGarvey LPA Ing AJ Idiopathic cough, prevalence and underlying mechanisms Pulm Pharmacol Ther 2004 17 435 439 15564088 10.1016/j.pupt.2004.09.012
The Oxford English Dictionary 2 New York: Oxford University Press
Irwin RS Corrao WM Pratter MR Chronic persistent cough in the adult: the spectrum and frequency of cases and successful outcome of specific therapy Am Rev Respir Dis 1981 123 414 417
Irwin RS Curley FJ French CL Chronic cough: the spectrum and frequency of causes, key components of the diagnostic evaluation and outcome of specific therapy Am Rev Resp Dis 1990 141 640 647 2178528
Smyrnios NA Irwin RS Curley FJ Chronic cough with a history of excessive sputum production: The spectrum and frequency of causes key components of the diagnostic evaluation, and outcome of specific therapy: Chest 1995 108 991 997 7555175
Pratter MR Bartter T Akers S Dubois J An algorithmic approach to chronic cough Ann Intern Med 1993 119 977 83 8214994
Palombini BC Villanova CA Araujo E Gastal OL Alt DC Stolz DP Palombini CO A pathogenic triad in chronic cough: asthma, postnasal drip syndrome and gastrooesophageal reflux disease Chest 1999 116 279 8 10453852 10.1378/chest.116.2.279
Irwin RS Boulet LP Cloutier MM Fuller R Gold PM Hoffstein V Ing AJ McCool FD O'Byrne P Poe PH Prakash UB Pratter MR Rubin BK Managing cough as a defence mechanism and as a symptom. A consensus panel report of the American College of Chest Physicians Chest 1998 114 133S 181S 9725800
Poe HR Harder RV Israel RH Chronic persistent cough: experience in diagnosis and outcome using an anatomic diagnostic protocol Chest 1989 95 723 27 2924600
Morice AH Kastelik JA Cough. 1: Chronic cough in adults Thorax 2003 58 901 7 14514949 10.1136/thorax.58.10.901
O'Connell F Thomas VE Pride NB Fuller RW Cough sensitivity to inhaled capsaicin decreases with successful treatment of chronic cough Am J Respir Crit Care Med 1993 150 374 80 8049818
McGarvey LPA Heaney LG Lawson JT Johnston BT Scally CM Ennis M Shepherd DRT MacMahon J Evaluation and outcome of patients with chronic non-productive cough using a comprehensive diagnostic protocol Thorax 1998 53 738 743 10319055
Haque RA Usmani OS Barnes PJ Chronic Idiopathic cough: a discrete clinical entity? Chest 2005 127 1710 1713 15888850 10.1378/chest.127.5.1710
Irwin RS Madison JM Diagnosis and treatment of chronic cough due to gastro-esophageal reflux disease and postnasal drip syndrome Pulm Pharmacol Ther 2002 15 293 4 12099781 10.1006/pupt.2002.0345
Irwin RS Madison JM Anatomical diagnostic protocol in evaluating chronic cough with specific reference to gastro-oesophageal reflux disease Am J Med 2000 108 126S 130S 10718465 10.1016/S0002-9343(99)00351-4
Patterson RN Johnston BT MacMahon J Heaney LG McGarvey LPA Oesophageal pH monitoring is of limited value in the diagnosis of 'reflux-cough' Eur Respir J 2004 24 724 7 15516662 10.1183/09031936.04.00007404
Morice AH Fontana GA Sovijarvi ARA Pistolesi M Chung KF Widdicombe J ERS Task Force The diagnosis and management of cough Eur Respir J 2004 24 481 492 15358710 10.1183/09031936.04.00027804
Mello CJ Irwin RS Curley FJ The predictive values of the character, timing and complications of chronic cough in diagnosing its cause Arch Int Med 1993 119 997 983
Benini L Ferrari M Sembenini C Olivieri M Micciolo R Zuccali V Bulighin GM Fiorino F Ederle A Cascio VL Vantini I Cough threshold in reflux oesophagitis: influence of acid and of laryngeal and oesophageal damage Gut 2000 46 762 7 10807885 10.1136/gut.46.6.762
Amin MR Postma GN Johnston P Digges N Koufman JA Proton pump inhibitor resistance in the treatment of laryngopharyngeal reflux Otolaryngol Head Neck Surg 2001 125 374 8 11593175 10.1067/mhn.2001.118691
Irwin RS Zawacki JK Wilson MM French CT Callery MP Chronic cough due to gastro-oesophageal reflux disease: Failure to resolve despite total/near total elimination of oesophageal acid Chest 2002 121 1132 1140 11948043 10.1378/chest.121.4.1132
Gibson PG Dolovich J Denburg J Ramsdale EH Hargreave FE Chronic cough: eosinophilic bronchitis without asthma Lancet 1989 1 1346 8 2567371 10.1016/S0140-6736(89)92801-8
Brightling C Ward R Goh KL Wardlaw AJ Pavord ID Eosinophilic bronchitis is an important cause of chronic cough Am J Respir Crit Care Med 1999 160 406 10 10430705
Fujimura M Ogawa H Nishizawa Y Nishi K Comparison of atopic cough with cough variant asthma: Is atopic cough a precursor of asthma? Thorax 2003 58 14 18 12511712 10.1136/thorax.58.1.14
Sifrim D Dupont L Blondeau K Zhang X Tack J Janssens J Weakly acidic reflux in patients with chronic unexplained cough during 24 hour pressure, pH, and impedance monitoring Gut 2005 54 449 54 15753524 10.1136/gut.2004.055418
Kok C Kennerson ML Spring PJ Ing AJ Pollard JD Nicholson GA A locus for hereditary sensory neuropathy with cough and gastroesophageal reflux on chromosone 3p22-p24 Am J Hum Genet 2003 73 632 7 12870133 10.1086/377591
Forsythe P McGarvey L Heaney LG MacMahon J Ennis M Sensory neuropeptides induce histamine release from bronchoalveolar lavage cells in both non asthmatic coughers and cough variant asthmatics Clin Exper Allergy 2000 30 225 32 10651775 10.1046/j.1365-2222.2000.00770.x
Jatakanon A Lalloo UG Lim S Chung KF Barnes PJ Increased neutrophils and cytokines, TNF-alpha and IL-8, in induced sputum of non-asthmatic patients with chronic dry cough Thorax 1999 54 234 7 10325899
Birring SS Murphy AC Scullion JE Brightling CE Browning M Pavord ID Idiopathic chronic cough and organ specific autoimmune diseases: a case control study Respir Med 2004 98 242 6 15002760 10.1016/j.rmed.2003.10.005
Chaudhuri R McMahon AD Thomson LJ MacLeod KJ McSharry CP Livingston E McKay A Thomson NC Effect of inhaled corticosteroids on symptom severity and sputum mediator levels in chronic persistent cough J Allergy Clin Immunol 2004 113 1063 70 15208586 10.1016/j.jaci.2004.03.019
French CT Fletcher KE Irwin RS Gender differences in health-related quality of life in patients complaining of chronic cough Chest 2004 125 482 8 14769728 10.1378/chest.125.2.482
Carr MJ Plasticity of vagal afferent fibres mediating cough Pulm Pharmacol Ther 2004 17 447 51 15564090 10.1016/j.pupt.2004.09.020
Piedimonte G Hegele RG Auais A Persistent airway inflammation after resolution of respiratory syncytial virus infection in rats Pediatr Res 2004 55 657 65 14711892 10.1203/01.PDR.0000112244.72924.26
Hu C Wedde-Beer K Auais A Rodriguez MM Piedimonte G Nerve growth factor and nerve growth factor receptors in respiratory syncytial virus-infected lungs Am J Physiol Lung Cell Mol Physiol 2002 283 L494 502 12114213
Carr MJ Hunter DD Jacoby DB Undem BJ Expression of tachykinins in nonnociceptive vagal afferent neurons during respiratory viral infection in guinea pigs Am J Respir Crit Care Med 2002 161 1985 90
O'Connell F Springall DR Moradoghi-Haftvani A Krausz T Price D Fuller RW Polak JM Pride NB Abnormal intraepithelial airway nerves in persistent unexplained cough Am J Respir Crit Care Med 1995 152 2068 75 8520777
Birring SS Brightling DE Symon FA Barlow SG Wardlaw AJ Pavord ID Idiopathic chronic cough: association with organ specific autoimmune disease and bronchoalveolar lymphocytosis Thorax 2003 58 1066 70 14645977 10.1136/thorax.58.12.1066
Mund E Christensson B Gronneberg R Larsson K Noneosinophilic CD4 lymphocytic airway inflammation in menopausal women with chronic dry cough Chest 2005 127 1714 1721 15888851 10.1378/chest.127.5.1714
Fujimura M Sakamoto S Kamio Y Matsuda T Sex difference in the inhaled tartaric acid cough threshold in non-atopic healthy subjects Thorax 1990 45 633 4 2402729
Dicpinigaitis PV Rauf K The influence of gender on cough reflex sensitivity Chest 1998 113 1319 21 9596313
Kastelik JA Thompson RH Aziz I Ojoo JC Redington AE Morice AH Sex-related differences in cough reflex sensitivity in patients with chronic cough Am J Respir Crit Care Med 2002 166 961 4 12359654 10.1164/rccm.2109061
Chang AB Phelan PD Sawyer SM Del Brocco S Robertson CF Cough sensitivity in children with asthma, recurrent cough and cystic fibrosis Arch Dis Child 1997 77 331 4 9389238
Prudon B Birring SS Vara DD Hall AP Thompson JP Thompson JP Pavord ID Cough and glottic-stop reflex sensitivity in health and disease Chest 2005 127 550 7 15705995 10.1378/chest.127.2.550
Ebihara T Sekizawa K Ohtui T Nakzawa H Sasaki H Angiotensin-converting enzyme inhibitor and danazol increase sensitivity of cough reflex in female guinea pigs Am J Respir Crit Care Med 1996 153 812 9 8564137
|
16270929
|
PMC1277012
|
CC BY
|
2021-01-04 16:37:14
|
no
|
Head Face Med. 2005 Aug 24; 1:4
|
latin-1
|
Head Face Med
| 2,005 |
10.1186/1746-160X-1-4
|
oa_comm
|
==== Front
Head Face MedHead & Face Medicine1746-160XBioMed Central London 1746-160X-1-51627092610.1186/1746-160X-1-5ResearchCephalometric norms for the Saudi children living in the western region of Saudi Arabia: a research report Hassan Ali H [email protected] P.O. Box 80209, Jeddah 21589, Preventive Dental Sciences Department, Faculty of Dentistry, King Abdulaziz University, Jeddah, Saudi Arabia2005 24 8 2005 1 5 5 30 4 2005 24 8 2005 Copyright © 2005 Hassan; licensee BioMed Central Ltd.2005Hassan; 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
Previous studies have established specific cephalometric norms for children with different ethnic backgrounds, showing different facial features for each group. Up till now, there is a paucity of information about the cephalometric features of Saudi children living in the western region of Saudi Arabia, who have distinct social and climatic characteristics. The aim of the present study was to establish cephalometric norms for children living in the western region of Saudi Arabia.
Methods
A total of 62 lateral cephalometric radiographs of Saudis (33 females and 29 males; aged 9–12 years) having good facial proportions and Class I dental occlusion, were traced and analyzed. Using the t-test, the mean value, standard deviation and the range of 20 angular and linear variables were calculated and compared to norms of adult Saudis living in the Western region of Saudi Arabia using the t-test. Male and female groups were also compared using the t- test.
Results
Saudi children tend to have a significantly shorter and lower face height, a larger angle of convexity, and more proclined and protruded incisors when compared with adult Saudis (P < 0.05). There were no statistically significant differences between male and female groups.
Conclusion
Saudi children have distinct cephalometric features, which should be used as a reference in the orthodontic treatment of young Saudi patients.
==== Body
Background
In orthodontic diagnosis and treatment planning, a cephalometric radiograph is an essential tool to relate patients with different malocclusions to their associated norms. Previous studies have established cephalometric norms for children in different countries who are descendants of special racial backgrounds [1-9]. Saudis were found to have distinct craniofacial features as compared with European-Americans [6-9]. Unfortunately, all the previously mentioned studies were performed in the central region of Saudi Arabia and there was only one study conducted in the western region, in which cephalometric norms were established for Saudi adults and then represented graphically on a wiggle to count for the variability of the readings among the Saudi population [9]. Results showed that Saudis, in general, have an increased ANB angle and bimaxillary protrusion when compared with European-American norms. It was concluded that the established norms should be used as a reference in the orthodontic treatment of Saudi adults. In addition, cephalometric norms should be presented on a polygon to count for the high variability that was observed in the Saudis because of their multiethnicity. A wiggle, as described by Vorhies and Adams [10], is a graph in which all average norms are plotted on a central vertical line. The maximum and the minimum readings of each norm are plotted on either side of the central line in a way that all the Class II readings are placed on the left side and the Class III readings are placed on the right side of the central line [10]. Unlike Vorhies and Admas [5], Hassan [9] used one standard deviation instead of the maximum and minimum readings of each reading.
The objectives of the present study were to establish norms for Saudi children living in the western region of Saudi Arabia and to present them graphically in the form of a polygon to count for any possible variation due to age, gender and multiracial background of the representing sample.
Methods
The present study was approved by the Ethical Committee of the Faculty of Dentistry, King Abdulaziz University (KAAU), in which a total of 62 lateral cephalometric radiographs of Saudi children (33 females and 29 males; aged 9–12 years) having acceptable profiles with competent lips, Class I dental and skeletal relationships, minimum overbite and overjet, minimum or no crowding, and no previous orthodontic treatment were selected to be included in the study group. The selected subjects were Saudis (by nationality) born and living in the western region of Saudi Arabia and of Arab descent. They were selected through the public health program that was conducted by the Department of Preventive Dental Sciences at KAAU, in which primary and intermediate public schools were visited for caries assessment and patient oral health education.
The radiographs were traced and analyzed manually by a single examiner. Twenty angular and linear measurements were calculated (Table 1 and figure. 1). The mean value, standard deviation and range of each variable was calculated and compared with the norms established for Saudi adults living in the western region of Saudi Arabia [4]. In addition, measurements were compared between male and female children. An independent sample t-test was used in the comparison between children and adults and as well as between male and female groups. To assess tracing errors, a second tracing was prepared for every 10 tracings. The mean error in linear measurements was ± 0.35 mm. The mean error in angular measurements was ± 0.92°. A set of cephalometric values was established for Saudi children. The resulting data (means and standard deviation) were represented diagrammatically in the form of a polygon (Wiggle) (Figure 2) using the mean value plus or minus one standard deviation.
Table 1 Different linear and angular measurements used
NPog-FH Intersection between NPog plane and Frankfort horizontal plane
NPog-FH Intersection between NPog plane and Frankfort plane
SNA Maxillary apical base relationship to anterior cranial base
SNB Mandibular apical base relationship to anterior cranial base
ANB Apical base relationship
NA-APog Angle of convexity
MP-FH Inclination of mandibular plane to FH
MP-SN Inclination of mandibular plane angle to anterior cranial base
OC-PL-SN Inclination of occlusal plane to anterior cranial base
Y-axis Angle made between SN and NGn line
L-FC. Ht Lower face height (Anterior nasal spine-Menton)
U1-SN Inclination of maxillary incisors to anterior cranial base
U1-NAz Inclination of maxillary incisors to NA
U1-NAmm Protrusion of maxillary incisors to NA
U1-L1 Inclination of maxillary incisors to mandibular incisors
L1-MP Inclination of mandibular incisors to mandibular plane
L1-NBz Inclination of mandibular incisors to NB
L1-NBmm Protrusion of maxillary incisors to NB
L1-APogz Inclination of mandibular incisors to APog plane
L1-APogmm Protrusion of mandibular incisors to APog plane
Figure 1 Cephalometric reference points. Different reference points used in the present study and their abbreviations.
Figure 2 Graphical presentation of cephalometric norms for Saudi children. A graph (Wiggle) in which cephalometric norms for Saudi children (mean age: 12.2 years) are plotted on a central vertical line. The readings of plus or minus one standard deviation of each norm are plotted on either side of the central line in a way that all the Class II readings were placed on the left side and the Class III readings were placed on the right side of the central line.
Results
Table 2 and figure 2 show the mean and standard deviation of the 20 angular and linear measurements selected, which represent the norms established in the present study. As compared with adult Saudis, children have a significantly increased angle of convexity which indicates more convex profiles in the Saudi children (P < 0.05). In addition, the lower face height was significantly shorter in the children's group (P < 0.05). The ANB angle was insignificantly increased in the children's group (P < 0.05). Dentally, upper and lower incisors were significantly more proclined and more protruded in the children's group (P < 0.05). The other readings were generally similar between Saudi children and adults. In addition there were no statistically significant differences between Saudi males and females (P < 0.05) (Table 3).
Table 2 Cephalometric Standards for Saudi children
Cephalometric Variables Saudis (Children) n = 62 Saudis (Adults) n = 68 t p
Mean SD Mean SD
NPog_FH 86.6 3.2 86.9 3.64 .054 0.957
NPog- SN 76.4 3.7 78.7 4.61 2.887 0.005*
SNA 79.6 5.4 80.8 4.06 .816 0.416
SNB 75.5 3.6 77.5 4.48 1.984 0.050
ANB 4.1 1.7 3.3 1.52 1.717 0.089
NA-APog 7.7 4.5 5.01 3.05 4.238 0.000*
MP/FH 27.3 4 28.5 4.79 1.103 0.272
MP/SN 37.2 5 35.9 5.96 1.940 0.055
Occl. Pl -SN 20.4 4.8 18.7681 6.21 .257 0.79
Y_AXIS 71.1 3.6 69.6 4.19 .460 0.647
L-F.HT 54.7% 1.198 56.0% 2.7 2.503 0.014*
U1_SN 105.5 5.6 106.8 8.07 2.922 0.004*
U1_NAZ 26.3 8.3 27.3 7.55 2.718 0.008*
U1_NAMM 6 1.8 6.8 2.92 4.491 0.000*
U1_L1 120.8 7.4 120.6 11.97 .277 0.782
L1_MP 94.7 6 93.9 7.7 2.276 0.025*
L1_NBZ 28.8 5.8 29.3 6.89 1.219 0.225
L1_NBMM 6.8 2.3 7.5 2.63 2.975 0.004*
L1_APogZ 26.5 5 27.23 6.07 0.564 0.574
L1_APogMM 4.5 1.97 4.89 2.89 2.064 0.042*
* P < 0.05
Table 3 Comparison of cephalometric measurements of the Saudi male and female children using t-test
Cephalometric Variable Male (n = 29) Female (n= 33) P
Mean SD Mean SD
NPog_FH 87.9 3.3 86.1 2.96 .64
NPog- SN 75.2 4.4 76.9 3.2 .42
SNA 74 5.1 77.8 4.98 .06
SNB 74.75 4.3 76.2 3.1 .38
ANB 3.8 1.7 3.9 1.8 .105
NA-APog 7 4.8 7.3 4.5 .058
MP/FH 25.4 4.5 26 3.8 .12
MP/SN 38 5.7 35 4.4 .418
Occ.P -SN 21.2 5.9 20 4.14 .174
Y_AXIS 70.4 4.1 69.5 3.4 .224
U1_SN 104.4 5.96 105.3 5.4 .095
U1_NAZ 28.6 8.2 25.4 5.5 .155
U1_NAMM 5.96 1.6 5.9 1.9 .062
U1_L1 122.2 7.6 120.95 7.4 .074
L1_MP 95.97 5.9 98.4 6 .224
L1_NBZ 28.7 5.3 29.3 6 .072
L1_NBMM 7.6 2.3 6.8 2.4 .471
L1_APogZ 25.5 4.5 26.96 5.3 .188
L1_APogMM 4.875 1.86 4.3 2 .496
Pg to NBMM 0.5 1.6 1.5 1.6 1.26
* P < 0.05
Discussion
Considering the ethnic background of patients in setting treatment objectives is an important requirement for successful orthodontic treatment. This can be achieved by establishing cephalometric and facial norms for the different racial groups. Unfortunately, there are no cephalometric norms for Saudi children living in the western region of Saudi Arabia. This study is considered as the first trial to establish norms for Saudi children living in that region.
The Saudi race in the western region of Saudi Arabia is unique in composition, which is multiracial in nature and has been established through interbreeding among the different communities who migrated there to be close to the Holy mosques of the Islamic world [9]. The sample used in the present study was carefully selected to include Arab-Saudis born and living in the western region of the Kingdom of Saudi Arabia. Hassan (2006) found that Saudi adults have an increased facial convexity, a more convex profile, a steeper mandibular plane, more protruded upper and lower incisors and shorter lower face height as compared with European-Americans [9].
Results of the present study have shown that Saudi children have statistically different skeletal and dental features than Saudi adults living in the western region of Saudi Arabia. These differences were noticed in the facial plane angle, angle of convexity, lower face height and the inclination of incisors. Most of the differences go with the general growth pattern of the human face, in which chins are more retrognathic and profiles are more convex during childhood and tend to straighten by age [11] (Table 2). In addition, lower face height tends to increase with age during the transition from childhood to adulthood which could be attributed to the cephalo-caudal gradient of growth of the facial bones [12]. Dentally incisors tend to protrude and procline with age in the Saudis, which could be attributed to environmental factors such as the clinically observed high incidence of mouth breathing and tongue thrusting habits among Saudi youngsters.
Although the ANB angle is insignificantly increased in children as compared with adults, it is still considered as an important result to emphasize, for proper orthodontic diagnosis. An ANB angle of four degrees and an angle of convexity of seven degrees should be considered normal in children. In addition, an ANB angle of two degrees, which is considered normal in adults, should be investigated more in children to exclude the tendency for Skeletal Class III relationship.
Conclusion
Saudi children living in the western region, have distinct facial and skeletal features which are different than Saudi adults. Therefore distinction should be made between young and adult patients. This can be achieved by using specific cephalometric norms for each age group.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
The author would like to thank Dr. B. Kuzonoto from the University of Illinois at Chicago and Dr. T. El Bialy from King Abdulaziz University at Jeddah, Saudi Arabia for their contributions
==== Refs
Bhat M Sudha P Tandon S Cephalometric norms for Bunt and Brahmin children of Dakshina Kannada based on McNamara's analysis J Indian Soc Pedod Prev Dent 2001 19 41 51 11692821
Abraham KK Tandon S Paul U Selected cephalometric norms in south Kanara children J Indian Soc Pedod Prev Dent 2000 18 95 102 11324204
Anderson AA Anderson AC Hornbuckle AC Hornbuckle K Biological derivation of a range of cephalometric norms for children of African American descent (after Steiner) Am J Orthod Dentofacial Orthop 2000 118 90 100 10893478 10.1067/mod.2000.103258
Hajighadimi M Dougherty HL Garakani F Cephalometric evaluation of Iranian children and its comparison with Tweed's and Steiner's standards Am J Orthod 1981 79 192 197 6937142 10.1016/0002-9416(81)90317-1
Sahin Saglam AM Gazilerli U Analysis of Holdaway soft-tissue measurements in children between 9 and 12 years of age Eur J Orthod 2001 23 287 294 11471271 10.1093/ejo/23.3.287
Shalhoub SY Sarhan OA Shaikh HS Adult cephalometric norms for Saudi Arabians with a comparison of values for Saudi and North American Caucasians Br J Orthod 1987 14 273 279 3481279
Sarhan OA Nashashibi IA A comparative study between two randomly selected samples from which to derive standards for craniofacial measurements J Oral Rehabil 1988 15 251 255 3164364
AL-Jasser NM Cephaloetric evaluation of craniofacial variations in normal Saudi population according to Steiner analysis Saudi Med J 2000 21 746 750 11423887
Hassan AH Cephaloetric Norms for Saudi Adults living in the Western Region of Saudi Arabia Angle Orthod 2006
Vorhies JM Adams JW Polygonic interpretation of cephalometric findings Angle Orthod 1951 21 194 197 14894865
Bishara SE Jakobsen JR Hession TJ Treder JE Soft tissue profile changes from 5 to 45 years of age Am J Orthod Dentofacial Orthop 1998 114 698 706 9844211 10.1053/od.1998.v114.a84780
Proffit WR Fields HW Concepts of Growth and Development In Contemporary Orthodontics 2000 St. Louis, Missouri: Mosby
|
16270926
|
PMC1277013
|
CC BY
|
2021-01-04 16:37:13
|
no
|
Head Face Med. 2005 Aug 24; 1:5
|
utf-8
|
Head Face Med
| 2,005 |
10.1186/1746-160X-1-5
|
oa_comm
|
==== Front
Head Face MedHead & Face Medicine1746-160XBioMed Central London 1746-160X-1-61627092710.1186/1746-160X-1-6Short ReportStrain driven fast osseointegration of implants Joos Ulrich [email protected]üchter Andre [email protected] Hans-Peter [email protected] Ulrich [email protected] Department of Cranio-Maxillofacial Surgery, University of Münster, Waldeyerstraße 30, D-48129 Münster, Germany2005 1 9 2005 1 6 6 3 5 2005 1 9 2005 Copyright © 2005 Joos et al; licensee BioMed Central Ltd.2005Joos 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 the bone's capability of dental implant osseointegration has clinically been utilised as early as in the Gallo-Roman population, the specific mechanisms for the emergence and maintenance of peri-implant bone under functional load have not been identified. Here we show that under immediate loading of specially designed dental implants with masticatory loads, osseointegration is rapidly achieved.
Methods
We examined the bone reaction around non- and immediately loaded dental implants inserted in the mandible of mature minipigs during the presently assumed time for osseointegration. We used threaded conical titanium implants containing a titanium2+ oxide surface, allowing direct bone contact after insertion. The external geometry was designed according to finite element analysis: the calculation showed that physiological amplitudes of strain (500–3,000 ustrain) generated through mastication were homogenously distributed in peri-implant bone. The strain-energy density (SED) rate under assessment of a 1 Hz loading cycle was 150 Jm-3 s-1, peak dislocations were lower then nm.
Results
Bone was in direct contact to the implant surface (bone/implant contact rate 90%) from day one of implant insertion, as quantified by undecalcified histological sections. This effect was substantiated by ultrastructural analysis of intimate osteoblast attachment and mature collagen mineralisation at the titanium surface. We detected no loss in the intimate bone/implant bond during the experimental period of either control or experimental animals, indicating that immediate load had no adverse effect on bone structure in peri-implant bone.
Conclusion
In terms of clinical relevance, the load related bone reaction at the implant interface may in combination with substrate effects be responsible for an immediate osseointegration state.
==== Body
Findings
Although the bone's capability of dental implant osseointegration has clinically been utilised as early as in the Gallo-Roman population [1], the specific mechanisms for the emergence and maintenance of peri-implant bone under functional load have not been identified. Here we show that under immediate loading of specially designed dental implants with masticatory loads, osseointegration is rapidly achieved. As the osseointegration state is much faster reached than commonly assumed, osseointegration is a strain dependant highly dynamic process.
Osseointegration is defined as a direct and stable anchorage of an implant by the formation of bony tissue without growth of fibrous tissue at the bone-implant interface [2]. A defining feature of osseointegration is that osteoblasts and mineralized matrix contacts the implant surface even when loads are applied. A common perception is that several weeks must be given to achieve implant osseointegration.
We have departed from this time related hypothesis by proposing that only minimal time (for example few hours, the time that is necessary for osteoblast adhesion on artificial substrates [3]) is required for osseointegration when the peri-implant tissue receives an optimal mechanical environment. We examined the bone reaction around non- and immediately loaded dental implants inserted in the mandible of mature minipigs during the presently assumed time for osseointegration (approved by the Animal Ethics Committee of the University of Münster under the reference number G 90/99). We used threaded conical titanium implants containing a titanium2+ oxide surface, allowing direct bone contact after insertion. The external geometry was designed according to finite element analysis: the calculation showed that physiological amplitudes of strain (500–3,000μstrain) generated through mastication were homogenously distributed in peri-implant bone (Figure 1). The strain-energy density (SED) rate [4] under assessment of a 1 Hz loading cycle was 150 Jm-3 s-1, peak dislocations were lower then nm. Eigth male Göttinger minipigs, 14 to 16 months of age with an average body weight of 35 kg, were used in this study. At day 3, day 7 and 28 animals were sacrificed with an overdose of T61 given intravenously.
Figure 1 Biomechanics and biology of implant osseointegration.
Bone was in direct contact to the implant surface (bone/implant contact rate 90%) from day one of implant insertion, as quantified by undecalcified histological sections (Figure 2). This effect was substantiated by ultrastructural analysis of intimate osteoblast attachment (Figure 3) and mature collagen mineralisation at the titanium surface. We detected no loss in the intimate bone/implant bond during the experimental period of either control or experimental animals, indicating that immediate load had no adverse effect on bone structure in peri-implant bone (Figure 4).
Figure 2 Finite element model of strain distribution in peri-implant bone. Bone strains do not exceed physiological values, bone dislocations are between 0 and 50 nm.
Figure 3 Histological picture of implant containing bone one day after insertion. Direct contact between bone and the implant is visible in the scanning electron micrographs.
Figure 4 Immuno-scanning electron microscopy of intimate osteoblast adhesion at the titanium surface by fibronectin mediated focal adhesions (fractured specimens, one day under loading).
Bone response on an implant surface depends on the reaction of cells and matrix towards the material surface as well as to the mechanical constraints in the vincinity of the implant. The maintenance of bone and its adaptation to external loads is based on a complex strain driven regulatory process of cells and matrix components [5,6]. Outside-in mechanical tension exert direct effects on cell behaviour by activating biochemical signalling pathways and regulating gene expression through focal adhesions [7]. Frost [8] provided a paradigm for the mechanical control of cellular bone modelling, the process whereby bone is laid down onto surfaces without necessarily preceded by resorption. Recent investigations have indicated that the strain related bone modelling process is also regulative for bone tissue formation in healing tissue [9].
Using an atomic force microscope, a molecular mechanistic origin for the remarkably fast recovery of toughness after bone deformation was found, when deformation of less then 50 nm at the surface of multivalent ions (as in the case of Ti-oxide) is present [10]. Our understanding of osseointegration theorises that bone strengthening responds to a highly specific mechanical design. Even if long-term osseointegrated implants show what seems to be similar bone tissue reactions, osseointegration might be able to be achieved more rapidly than otherwise observed. Screw type titanium implants, as used in dental implantology, have in contrast to orthopaedic implants not only been convincingly shown very good clinical long-term success [11], but were also successful when load transfer is immediately present as seen in traumatology. In terms of clinical relevance, the load related bone reaction at the implant interface may in combination with substrate effects be responsible for an immediate osseointegration state.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
UJ designed the study, searched the database, extracted the data. AB helped with the study design and analysis. HPW had analysis the histological probes and UJ developed the implant design.
==== Refs
Crubezy E Murail P Girard L Bernadou JP False teeth of the Roman world Nature 1998 391 29 9422504 10.1038/34067
Branemark PI Adell R Breine U Hansson BO Lindstrom J Ohlsson A Intra-osseous anchorage of dental prostheses. I. Experimental studies Scand J Plast Reconstr Surg 1969 3 81 100 4924041
Okumura A Goto M Goto T Yoshinari M Masuko S Katsuki T Tanaka T Substrate affects the initial attachment and subsequent behavior of human osteoblastic cells (Saos-2) Biomaterials 2001 22 2263 2271 11456066 10.1016/S0142-9612(00)00415-4
Turner CH Takano Y Owan I Aging changes mechanical loading thresholds for bone formation in rats J Bone Miner Res 1995 10 1544 1549 8686511
Rubin CT Pratt GW JrPorter AL Lanyon LE Poss R Ultrasonic measurement of immobilization-induced osteopenia: an experimental study in sheep Calcif Tissue Int 1988 42 309 312 3135099
Huiskes R Ruimerman R van Lenthe GH Janssen JD Effects of mechanical forces on maintenance and adaptation of form in trabecular bone Nature 2000 405 704 706 10864330 10.1038/35015116
Chicurel ME Singer RH Meyer CJ Ingber DE Integrin binding and mechanical tension induce movement of mRNA and ribosomes to focal adhesions Nature 1998 392 730 733 9565036 10.1038/33719
Frost HM The mechanostat: a proposed pathogenic mechanism of osteoporoses and the bone mass effects of mechanical and nonmechanical agents Bone Miner 1987 2 73 85 3333019
Frost HM A brief review for orthopaedic surgeons: Fatigue damage (microdamage in bone (its determinants and clinical implications) J Orthop Sci 1998 3 272 281 9732562 10.1007/s007760050053
Thompson JB Kindt JH Drake B Hansma HG Morse DE Hansma PK Bone indentation recovery time correlates with bond reforming time Nature 2001 13 773 776 11742405 10.1038/414773a
Albrektsson T Johansson Osteoinduction, osteoconduction and osseointegration C Eur Spine J 2001 10 96 101 10.1007/s005860100282
|
16270927
|
PMC1277014
|
CC BY
|
2021-01-04 16:37:14
|
no
|
Head Face Med. 2005 Sep 1; 1:6
|
utf-8
|
Head Face Med
| 2,005 |
10.1186/1746-160X-1-6
|
oa_comm
|
==== Front
Head Face MedHead & Face Medicine1746-160XBioMed Central London 1746-160X-1-71627094210.1186/1746-160X-1-7ReviewTrends and characteristics of oral and maxillofacial injuries in Nigeria: a review of the literature Adeyemo Wasiu Lanre [email protected] Akinola Ladipo [email protected] Mobolanle Olugbemiga [email protected] Olutayo [email protected] Department of Oral and Maxillofacial Surgery, Lagos University Teaching Hospital, P.M.B. 12003, Lagos, Nigeria2 Department of Oral and Maxillofacial Surgery, College of Medicine, University of Lagos, P.M.B. 12003, Lagos, Nigeria2005 4 10 2005 1 7 7 9 6 2005 4 10 2005 Copyright © 2005 Adeyemo et al; licensee BioMed Central Ltd.2005Adeyemo 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 etiology of maxillofacial injuries varies from one country to another and even within the same country depending on the prevailing socioeconomic, cultural and environmental factors. Periodic verification of the etiology of maxillofacial injuries helps to recommend ways in which maxillofacial injuries can be averted. The aim of the present study is therefore to analyse the characteristics and trends of maxillofacial injuries in Nigeria based on a systematic review of the literature.
Methods
A literature search using MEDLINE was conducted for publications on maxillofacial injuries in Nigeria. The relevant references in these publications were manually searched for additional non-Medline articles or abstracts. Forty-two studies met the inclusion criteria and the full-texts of these articles were thoroughly examined. Due to lack of uniformity and consistency in assessment and measurement variables, and treatment modalities in most of the studies, it was impossible to apply the traditional methods of a systematic review. Therefore, a narrative approach was conducted to report the findings of the included studies.
Results
Although, other causes like assaults, sport injuries, and industrial accidents increased in numbers, throughout the period between 1965 and 2003, road traffic crashes remained the major etiological factor of maxillofacial injuries in all regions, except northeastern region where assault was the major cause. A significant increase in motorcycles related maxillofacial injuries was observed in most urban and suburban centres of the country. Animal attacks were not an unusual cause of maxillofacial injuries in most parts of northern Nigeria. Patients in the age group of 21–30 years were mostly involved. A strong tendency toward an equal male-to-female ratio was observed between earlier and later periods.
Conclusion
Road traffic crashes remain the major cause of maxillofacial injuries in Nigeria, unlike in most developed countries where assaults/interpersonal violence has replaced road traffic crashes as the major cause of the injuries. There is a need to reinforce legislation aimed to prevent road traffic crashes and the total enforcement of existing laws to reduce maxillofacial injuries among children and adults. Special attention should also be paid by the authority to improve the socioeconomic conditions of Nigerian populace.
==== Body
Background
Skeletal and soft tissue injuries of the face constitute quite a significant portion of the workload of the oral and maxillofacial surgeons in Nigeria [1]. Being the most exposed part of the body, the face is particularly vulnerable to such injuries, 20–60% of all those involved in automobile accidents having some level of facial fractures [2,3]. Surveys of facial injuries have shown that the etiology varies from one country to another and even within the same country depending on the prevailing socioeconomic, cultural and environmental factors [4-6]. Earlier studies from Europe and America revealed that road traffic crashes (RTC) were the most frequent cause of facial injuries [7,8]. However, more recent studies have shown that assault is now the most common cause of maxillofacial injuries in developed countries [9-11], whereas traffic accidents remain the most frequent cause in many developing countries [12-19].
Periodic verification of the etiology of maxillofacial injuries helps to assess the proficiency of road safety measures such as speed limit, drunk driving, and seat beat belt laws and the behavioural patterns of the people in different countries and helps to recommend other ways in which injuries to the face can be averted [20].
The aim of the present study is therefore to analyse the characteristics and trends of maxillofacial injuries in Nigeria based on a systematic review of the literature.
Methods
A computerized literature search using MEDLINE was conducted for publications on maxillofacial injuries in Nigeria published between 1970 and 2005. For this search, the medical subject headings "maxillofacial injuries" or "maxillofacial fractures" or "mandible fractures" or "middle-third fractures" or "facial fractures" or "zygoma fractures" were combined with "Nigeria" or "Africa". The Boolean operator 'AND' was used to combine and narrow the searches. We manually searched the references in these articles to look for additional relevant non-Medline articles or abstracts. The full-texts of all these articles were thoroughly examined. Personal contacts were also made with institutions and investigators of previous studies for missing data and also for the provision of articles found suitable for the review, but not readily available to us. One author (WLA) conducted the literature search. All the authors agreed upon inclusion and exclusion criteria.
Inclusion criteria
1. Availability of the full-text article
2. Retrospective or prospective studies
3. All age groups (Children and adults)
4. Civilian-type injuries
Publications on maxillofacial injuries sustained during Nigerian civil war were excluded from the review.
Assessment of the studies
A total of 44 full-text articles and abstracts were identified. Two articles on maxillofacial injuries sustained during the Nigerian civil war were excluded. A total of 42 publications published between January 1977 and April 2005, which satisfied the inclusion criteria were, therefore included in the review. These included 34 Medline and 8 non-Medline articles. These publications were based on patients seen and treated between 1965 and 2003 from different centers of the six geopolitical zones of the country (Figure 1) including: Ibadan, south west (SW) [18,19,21-35], Lagos (SW) [36-40], Ife (SW) [1,41-46], Kaduna, north central (NC) [47-50], Sokoto, north west (NW) [51,52], Maiduguri, north east (NE) [4,53,54], Enugu, south east (SE) [15,16,55], and Benin city, south south (SS) [56] (Table 1).
Figure 1 Map of Nigeria (source: CIA's The World Factbook). Ife (not shown) lies in the north east of Ibadan below Oshogbo.
Table 1 Location of investigations, aetiology of injury and gender distribution.
Author (Ref.-No.) Locationa Major cause of injury 2nd major cause of injury % motorcycle related Male/female ratio (%) Tissue affected
Ajagbe et al. (21) Ibadan (SW) RTC (63) Falls (19) 10 2.1:1 Hard Tissue
Nwoku et al. (36) Lagos (SW) RTC ## ### 3:1 Hard/Soft Tissues
Ajagbe et al. (34) Ibadan (SW) RTC Falls 15.7 3:1 Hard Tissue
Adekeye (47) Kaduna (NC) RTC (76) Assaults (13) 22 16.9:1 Hard/Soft Tissues
Adekeye (48) Kaduna (NC) RTC (82) Falls ### 24:1 Hard Tissue
Adekeye (49) Kaduna (NC) RTC ## ### 2:1 Hard Tissue
Nyako (23) Ibadan (SW) RTC (77) Assaults (9) 10.6 6.4:1 Hard Tissue
Odusanya (41) Ife (SW) RTC (53) Falls 22.1 5.4:1 Hard Tissue
Abiose (22) Ibadan (SW) RTC (81) Assaults (9) ### 5.5:1 Hard Tissue
Akinwande (37) Lagos (SW) RTC (65) Assaults (12) 18.5 4.2:1 Hard/Soft Tissues
Abiose (32) Ibadan (SW) RTC (81) Assaults (7) ### 14:1 Hard Tissue
Arotiba (38) Lagos (SW) RTC (100) # 6.3 2.3:1 Hard Tissue
Arotiba (39) Lagos (SW) RTC (63) Assaults (20) 4 2.1:1 Hard Tissue
Oji (15) Enugu (SE) RTC (83) Assaults (8) 21 3:1 Hard Tissue
Ogunbodede (52) Sokoto (NW) bCamel bite # # # Hard/Soft Tissues
Denloye et al. (33) Ibadan (SW) RTC (47) Falls (41) ### 1.8:1 Hard/Soft Tissues
Ugboko et al. (1) Ife (SW) RTC (72) Falls (11) 14.5 4.1:1 Hard/Soft Tissues
Akinwande et al. (40) Lagos (SW) bGunshots (100) # # 5.1:1 Hard/Soft Tissues
Oji (55) Enugu (SE) RTC (28) Assaults (25) 5 2.6:1 Hard Tissue
Ugboko et al. (46) Ife (SW) RTC (50) Falls (31) 1.9 6.4:1 Hard/Soft Tissues
Oji (16) Enugu (SE) RTC (83) Assaults (8) 21 3:1 Hard Tissue
Olasoji (53) Maiduguri (NE) bAssaults (100) # # 2.5:1 Hard/Soft Tissues
Ugboko et al. (42) Ife (SW) bGunshots (100) # # 21:1 Hard /Soft Tissues
Fasola et al. (28) Ibadan (SW) bSports (100) # # 4.1:1 Hard Tissue
Fasola et al. (30) Ibadan (SW) RTC (38) Falls (25) ### 2.6:1 Soft Tissue
Fasola et al. (19) Ibadan (SW) RTC (79) Assaults (9) ### 7.6:1 Hard/Soft Tissues
Fasola et al. (25) Ibadan (SW) RTC (52) Falls (24) 3.2 2.6:1 Hard/Soft Tissues
Fasola et al. (26) Ibadan (SW) RTC (53) Falls (24) ### 2.8:1 Hard Tissue
Olasoji et al. (4) Maiduguri (NE) Assaults (48) RTC (36) 9 2.2:1 Hard /Soft Tissues
Olasoji et al. (54) Maiduguri (NE) RTC (54) Falls (25) 2 7.5:1 Hard/Soft Tissues
Ugboko et al. (51) North (NE, NW, NC) bAnimal attacks # # 4:1 Hard/Soft Tissues
Oginni et al. (43) Ife (SW) bDog bites # # 5:3 Soft Tissue
Oginni et al. (44) Ife (SW) Falls (38) RTC (33) # 1.4:1 Soft Tissue
Fasola et al. (29) Ibadan (SW) RTC (76) Assaults (9) ### 2.7:1 Hard/Soft Tissues
Fasola et al. (31) Ibadan (SW) RTC (82) Sports (8) ### 5.3:1 Hard Tissue
Fasola et al. (18) Ibadan (SW) RTC (69) Assaults (12) 20.6 3.3:1 Hard Tissue
Fasola et al. (24) Ibadan (SW) RTC (69) # 15.1 2.9:1 Hard Tissue
Fasola et al. (27) Ibadan (SW) RTC (59) Falls (21) 1.9 1:1 Hard Tissue
Saheeb et al. (56) Benin (SS) RTC (66) Assaults (10) 26.5 2.7:1 Hard /Soft Tissues
Adebayo et al. (50) Kaduna (NC) RTC (56) Falls (24) ### 4.7:1 Hard Tissue
Bankole et al. (35) Ibadan (SW) Falls (66) RTC (18) ### 2.3:1 Soft Tissue
Ugboko et al. (45) Ife (SW) RTC (74) Falls/Assaults (14) 9.4 6:1 Hard Tissue
a(SW) South-west, (SE) South-east (SS) South-south, (NW) North-west, (NE) North-east, (NC) North-central
bPublications on a single specific etiology
RTC = road traffic crash
# = not applicable
## = not specified
### = not separately classified
A protocol was prepared to identify the following features of each study: type of participants (i.e. adults or children or both groups), number of injuries analyzed, etiology of injury, peak age of incidence, gender predilection, site of injury, target population, as well as period and location of the study (Table 1, 2). Treatment modalities were also assessed.
Table 2 Type of included study, number of patients analyzed, target population and peak age of incidence.
Author (Ref.-No.) Type of study n of patients Target population Bone mostly affected (%) Peak age of incidence, years (%)
Ajagbe et al (21) retrospective 203 total mandible (60.5) 21–30 (32)
Nwoku et al (36) retrospective 84 total mandible (90) ##
Ajagbe et al (34) retrospective 324 total mandible (60) 21–30
Adekeye (47) prospective 1447 total mandible (62.5) 21–30 (56)
Adekeye (48) retrospective 337 total # 21–40 (80)
Adekeye (49) retrospective 85 Children mandible >10
Nyako (23) retrospective 341 total mandible (73) 21–30 (46)
Odusanya (41) retrospective 231 total mandible (67) 21–30
Abiose (22) retrospective 104 total mandible (75) 21–30 (43)
Akinwande (37) prospective 208 total mandible 21–30 (51)
Abiose (32) retrospective 59 total # 21–30
Arotiba (38) prospective 128 total mandible (62) 20–29 (>40)
Arotiba (39) prospective 202 total mandible (64) 20–29 (40)
Oji (15) retrospective 900 total mandible (42) 21–30 (36)
Ogunbodede (52) case report 1 # # #
Denloye et al (33) retrospective 106 Children mandible 0–8 (62)
Ugboko et al (1) retrospective 442 total mandible (64) 21–30 (39)
Akinwande et al (40) prospective 35 total mandible 20–34 (66)
Oji (55) retrospective 40 Children mandible (89) 9–11 (40)
Ugboko et al (46) retrospective 52 Children mandible (62) 12–14 (50)
Oji (16) retrospective 900 total mandible (53) 21–30 (36)
Olasoji (53) retrospective 105 total mandible (43) 20–29 (42)
Ugboko et al (42) retrospective 22 total Zygoma (27) 21–40
Fasola et al (28) retrospective 77 total mandible (54.4) 21–30 (52)
Fasola et al (30) retrospective 831 total # 21–30 (33)
Fasola et al (19) prospective 103 total # 21–30 (47)
Fasola et al (25) retrospective 93 children mandible (86) 11–16 (54)
Fasola et al (26) retrospective 72 children # 12–16 (57)
Olasoji et al (4) prospective 306 total mandible (66) 21–30 (41)
Olasoji et al (54) retrospective 102 Children mandible (73) 12–15 (54)
Ugboko et al (51) retrospective 34 total mandible (56) 11–30 (74)
Oginni et al (43) retrospective 8 children # ##
Oginni et al (44) retrospective 174 children # ##
Fasola et al (29) aretrospective 531 total # 21–30 (39)
Fasola et al (31) prospective 76 total # 21–30 (51)
Fasola et al (18) bpro/retrospective 824 total mandible (75) 21–30 (36)
Fasola et al (24) prospective 159 total mandible 21–30 (36)
Fasola et al (27) retrospective 53 adults mandible (96) 60–70 (77)
Saheeb et al (56) retrospective 250 total mandible (65) 20–30 (32)
Adebayo et al (50) retrospective 443 total mandible (64) 20–39 (65)
Bankole et al (35) retrospective 64 children # 0–5
Ugboko et al (45) retrospective 128 total # 21–30 (38)
a analysis of concomitant injuries in patients with maxillofacial fractures
bcomparative study
total = all age groups
# = not applicable
## = not specified
Most of the studies lack uniformity and consistency in assessment and measurement variables (information bias) and treatment modalities. The age bracket of patients considered as "children" by several investigators varied considerably (Fasola et al [26], 16 years and below; Oji [55], under 11 years; Olasoji, under 15 years; Ugboko et al, 14 years and below [46]; Denloye et al [33], less than 17 years; Oginni et al [44], 15 years and below). Repetition of the same data in different studies was also observed. While most of the published articles focused only on hard tissue injuries, few others reported on either hard and soft tissue injuries or soft tissue only (Table 1). Although, the majority of the patients in the studies were treated by closed reduction and fixation methods, uniformity in treatment was lacking. Due to the heterogeneity of the study methodologies in this review it was not possible to apply the traditional methods of a systematic review. A meta-analysis is only suitable if there is sufficient similarity in the populations studied and the measurements used. This was not the case with the studies identified in this review. Therefore, a narrative approach was taken to report the findings of the included studies.
Data was analysed using the SPSS for window (version 11.5; SPSS Inc, Chicago, IL) statistical software package. Descriptive statistics and the non-parametric chi square test were used to analyse the incidence of injuries in different time periods. The critical level of significance was set at p < .05.
Results
Of the 42 articles reviewed, 31 were retrospective studies, 9 prospective, 1 article was a case report and 1 article was a comparative study of a prospective and a retrospective data. Road traffic crash (RTC) was the major cause of maxillofacial injuries in both children and adults in all the zones of the country with the exception of north eastern states where assault was the major cause of injuries (Table 1). Although, motor vehicles were responsible for most cases of RTC, motorcycle related injuries increased significantly between 1965 and 2003. Between 1965 and 1999 in Ibadan, the number of motorcycle-related maxillofacial injuries increased by a factor of 2.6, and more significant cases (p = .02) of motorcycle related injuries were recorded in 1978–1982 period compared to 1995–1999 (Table 3). In Enugu (SE) Nigeria, between 1985 and 1995, the number of motorcycle related maxillofacial injuries increased by a factor of 1.6 (Table 3). An increase in the number of motorcycle related maxillofacial injuries was also observed between 1973 and 2000, and between 1976 and 1995 in Kaduna (NC) [48,51] and Ife (SW) [1,41] respectively. In Benin (SS) [56] and Lagos (SW) [37], 26.5% and 19.0% of cases with maxillofacial injuries were involved in motorcycle related crashes respectively, and motorcycle passengers sustained more severe injuries than other vehicle users [37,56].
Table 3 Analysis of road traffic injuries due to motor vehicles and motorcycles between 1965 and 1999 in Ibadana and between 1985 and 1995 in Enugub.
IBADAN (South-west, Nigeria)
Types of automobile involved Study period
1965–1975 1978–1982 1982–1984 1995–1999
Motor vehicles 46.3% 84.9% 80% 63.4%
Motorcycles 7.8% 10.6% # 20.6%*
ENUGU (South-east, Nigeria)
Study period
1985–1990 1991–1995
Motor vehicles 59% 59%
Motorcycles 16% 25%
a adapted from Abiose [22], Ajagbe et al [21], Fasola et al [18] and Nyako [23]
b adapted from Oji [16]
# = not specified
* statistically significant (p = 0.02)
Pedestrian related maxillofacial fractures also increased in major cities across the country. In Ibadan (SW), an increase by a factor of 3.2 was reported between 1978 and 1999 [18,23] and in Lagos (SW), 35.6% (1983–1986) and 28.1% (1989–1992) of patients involved in RTC were pedestrians hit by vehicles [37,38].
Assaults were the second most common cause of injuries in most centres followed by falls (Table 1). Falls were important causes of injuries in children. Increase in the number of patients who sustained injuries as a result of assaults, falls, sports injuries and industrial accidents was observed in most centers over the years [1,18,35,37,46,48,51,53]. Animal attacks were also a frequent cause of maxillofacial injuries especially in northern part of the country [4,43,45,52].
The peak age of incidence of maxillofacial injuries was 21–30 years in most centers followed by 31–40 years. In children, injuries occurred mostly in children aged > 10 years. More males were affected than females in all age groups. A tendency towards an equal male-to-female ratio was observed between earlier and later studies in most urban centers. A significant reduction in male-to-female ratio from 16.9:1 (1973–1978) to 3:1 (1991–2000) was reported from Kaduna (NC) (Table 1). Another significant reduction in male-to-female ratio from 6.4:1 (1978–1982) to 3.3:1 (1995–1999) was reported from Ibadan (SW) (Table 1).
The Mandible was the most frequently involved bone in maxillofacial fractures in all the centers across the country, and the most frequently involved middle-third bone was the zygoma [1,18,22,32,50]. The LeFort I fracture was the most common of the LeFort fracture types [1,4,22]. Analysis of fracture of the mandible revealed mandibular body as the most frequently involved part, followed by symphyseal/ parasymphyseal region [1,4,18,23,25,37-39,46,47,50]. Dentoalveolar and condylar fractures were less frequently reported. Another remarkable feature of maxillofacial injuries in most reports was extensive soft tissue injuries [30,18,37,38,56].
Closed reduction and dental wiring with arch bars, direct wires and eyelet wires combine with intermaxillary fixation were the most common form of treatment [1,21,34,47,50] for mandibular fractures. Wire osteosynthesis is employed for open reduction and internal fixation of mandibular fractures in few cases [1,21,22,50]. Fractures of the maxillae/LeFort fractures were reduced and fixed by eyelets/arch bars combined with suspension wires and intermaxillary fixation [1,32,34,47,50]. Zygomatic complex fractures were treated either conservatively or by either closed or open reduction with Gillies' temporal approach, lateral coronoid approach or transosseous wiring [21,42,45,47].
Discussion
The large variations in assessment and measurement variables, as well repetition of data employed by previous investigators of maxillofacial injuries in Nigeria made a systematic review impossible. However, analysis of the previous studies on maxillofacial injuries in Nigeria showed a noticeable trend and characteristic.
Although, road traffic crashes remained the major etiological factor of maxillofacial injuries other causes like assaults, sport injuries and industrial accidents have increased in numbers between 1965 and 2003 in Nigeria. This finding is in agreement with reports from other developing countries where RTC remains the major etiologic factor of maxillofacial injuries [12,13,17], but contrasts reports from developed countries where assaults and interpersonal violence has replaced RTC as the major cause of maxillofacial injuries [6,10,11,18]. Civilian-type maxillofacial injuries were rare prior to Nigerian independence in 1960 [21]. Immediate post independence period witnessed a significant increase in the numbers of motor vehicles imported into the country. It is worthwhile to note that the period from 1965 up to the present time has witnessed a steady increase in the number of second-hand vehicles into Nigeria. Also, lack of enforcement of reshipment inspection rules and regulations has encouraged the importation of vehicles whose road worthiness leaves much to be desired [1]. In addition, the roads are badly maintained, and there is general lack of enforcement of traffic rules and regulations, especially the use of seat belts. Non-usage of protective elements was also thought to be responsible for extensive soft tissue injuries seen in maxillofacial injured patients [18,37,38,56].
Over the last 40 years, there has been a significant increase in the number of maxillofacial injuries that resulted from motorcycle accidents in Nigeria (Table 3). These findings contrast that of others [57] who reported a decrease in the number of motorcyclists involved in maxillofacial injuries. However, Konto et al [58] reported that bicycle related maxillofacial fractures increased by 19.3% between 1981 and 1997 in Finland. The increase in the present study is due to a significant increase in the number of motorcycles plying Nigeria roads. Even in Abuja, the nation's capital, anecdotal evidence has shown that motorcyclists and their passengers are involved in more than 55% of cases of road traffic crashes. In the United States of America (USA), the number of registered motorcycles increased from 600,000 units in 1961 to 3.3 million units in 1971; a 450% increase within a decade [59,60]. This pattern was also recognised in Nigeria when the number increased from 144,480 units to 284,124 units between 1976 and 1981, an increase of almost 200% within 5 years [61]. Motorcycles have become a prominent mode of transportation in both urban and suburban cities in Nigeria. Frequent traffic jams as a result of poor road network in the country have made motorcycles attractive to commuters because motorcycles can pass through narrow ways [18]. However, most of the motorcyclists are unlicensed and often do not follow traffic rules and regulation. Fasola et al [24] reported that only one (3.8%) of the motorcyclists who sustained maxillofacial injuries within Ibadan city (SW) wore a crash helmet while Saheeb and Etetafia [56] reported that none of the motorcyclists and their passengers involved in RTC in Benin city (SS) wore protective helmet.
The number of pedestrians involved in maxillofacial injuries has also been on the increase especially in urban centres unlike reported elsewhere [57]. This is peculiar to the overpopulated cities with few subways and overhead bridges. Therefore, it is relatively common for pedestrians to have to run in oncoming vehicular traffic [18,38]. Konto et al [58] also reported an increase in pedestrian related maxillofacial fractures in their study.
While RTC have been steadily falling in the developed countries, they continue to rise with horrifying speed in the low and middle-income (LMIC) countries of Africa and Asia [62]. The World Health Organisation (WHO) has estimated that nearly 25% of all injury fatalities worldwide are a result of road traffic crashes, with 90% of the fatalities occurring in LMIC [62]. The reductions in RTC in developed countries are largely attributed to a wide range of road safety measures such as seat belt use, traffic calming measures and traffic law enforcement. Therefore, there is an urgent need to get down to what the developed nations have done to reduce/prevent road traffic crashes.
Assaults and falls were the second most common cause of maxillofacial injuries in adults and children respectively in all centres except the north eastern part of the country, where assaults remained a major cause (Table II). Other common causes were sport injuries, industrial accidents and animal attacks. Fasola et al [18] in Ibadan (SW) reported an increase in number of maxillofacial injuries due to assaults, falls, sporting injuries and industrial accidents between 1978 and 1999 by a factor of 1.4, 1.5, 3.5, and 1.5 respectively. Increase in number of assaults related maxillofacial injuries could be attributed to the poor socioeconomic conditions of the country leading to stress and propensity to crime. In fact, the employment rate among college and university graduates has increased from 4% during the early 1970s to 45% currently [4]. Furthermore, the poor capital income of an average Nigerian has decreased by 75% during the past 20–25 years [4,63]. The prevalence of assaults related injury in north eastern Nigeria could be attributed to nomadic form of life style in this region, where animals are moved over several kilometres of land grazing without strict laws guiding their movement thereby destroying cash crops [53]. This frequently led to fights between farmers and cattle men, and various objects such as cutlasses/machetes, arrows and wooden objects are used in inflicting injuries during fight [4,53]. This is unlike European and American studies where most of the fights occurred in the streets, clubs and pubs [6,7,10,11].
Also, the increase in maxillofacial injuries due to sports injuries and industrial accidents could be attributed to increase involvement of Nigerians in recreational and professional sport activities, and increase in the numbers of industries over the years without corresponding increase in protective measures. Onyeaso and Adegbesan [64] in a survey among Nigerian sport persons reported that only one-third of them ever used protective elements during sporting activities, whereas about 60% of them have had one form of orofacial injury or the other before.
Maxillofacial skeletal and soft injuries due to animal attacks were not infrequent, especially in northern part of the country [4,51,52]. While dogs remain the animals most commonly implicated in other reports [65,66], cows, camels and donkeys were mostly involved in Nigeria, because cattle rearing and use of animals as "beasts of burden" are still prevalent practices in northern part of Nigeria [4,51,52].
The peak age of incidence of maxillofacial injuries of 21–30 years among Nigerians is not different from reports from other parts of the world [5-10,12-14,57]. The possible explanation for this is that people in this age group take part in dangerous exercises and sports, drive motor vehicles carelessly, and are most likely to be involved in violence [16].
More males were involved in maxillofacial injuries than females in agreement with previous reports [5-10,13,14]. However, a tendency towards an equal male-to female ratio was observed between earlier and later studies in most centres across the country. This can be attributed to a changing workforce. Women, who are used to stay at home, now work in outdoor and high-risk occupations, thus becoming exposed to RTC and other causes of maxillofacial injuries [18,50].
Most of the fractures of maxillofacial skeleton in Nigerian patients were of the mandible, the findings comparable to other reports [9,12-14]. The mobility of the mandible and the fact that it has less bony support than the maxilla have been implicated [16,67]. Dentoalveolar and condylar fractures were reported to be less in Nigerian patients [1,9,12-14,54]. Dental/dentoalveolar injury is frequently overlooked in surveys that review maxillofacial injury [68-70]. Only the analysis of a large number of injuries reveals the risk of suffering from dentoalveolar trauma [68-70]. Gassner et al [69] in a large series of 9,543 patients with 21,067 maxillofacial injuries reported an incidence of 49.9% of dentoalveolar injuries among their patients. Gassner et al [70] in another large series of craniomaxillofacial trauma in 3,385 children younger than 15 years of age reported an incidence of 76.3% cases of dentoalveolar injuries. Midfacial bone fractures especially LeFort types and orbital floor fractures were reported to be commoner than mandibular fractures [69,70] in contrast to Nigerian reports. A low utilization of technological advances in the imaging of maxillofacial fractures (e.g. CT Scan) in Nigeria may be partially responsible for the observed difference. The midfacial skeleton is much more difficult to assess using plain films than is the mandible [71]. The presence of thin bones, fluid-filled spaces (e.g. congested sinuses), and soft tissues (e.g. orbital contents) make accurate assessment difficult with images that do not offer a high degree of contrast [71]. The difference in the incidence of middle-third fractures has also been related to the refusal of Nigerian motorists to use safety devices, which has reduced their survival after severe middle-third fractures [50].
Although, open reduction and internal fixation remains the "gold standard" of treatment of maxillofacial fractures [72,73], this form of treatment however, has not become popular in our environment [1,50]. Presently, the full compliment of equipment and materials for rigid fixation is not readily available in all parts of the country; and where available, the cost of treatment is usually quite prohibitive [45]. Previous Nigerian reports have, however attested to the satisfactory results obtained using simple conservative methods of closed reduction and mandibulo-maxillary fixation [1,4,16,19,21,25,32,36,45,50,54].
Conclusion
No apparent shift from road traffic crashes as the leading cause of maxillofacial injuries in Nigeria over a period of 40 years was observed, unlike in most developed countries where assaults/interpersonal violence has replaced road traffic crashes as the major cause of the injuries. Injuries have causes, they do not simply befall us from fate or bad luck. Since no magic pill is envisaged for the prevention of road traffic crashes, we need to take good stock of all the tools at our disposal, and to get down to what the developed nations have done to reduce/prevent road traffic crashes. Therefore, an awareness campaign to educate the public about the importance of restraints and protective headgear in cars and motorcycles should be championed. These findings should also alert the authorities, particularly the government and the Road Safety Commission to the need for the provision of good roads, enforcement of existing traffic laws, and general improvement of socioeconomic condition of the populace.
Competing interests
The author(s) declare they have no competing interest.
Authors' contributions
WLA conceived the study and did the literature search, coordinated the write-up and submission of the article. WLA, ALL, MOO and OJ participated in the writing of the manuscript. All the authors read and approved the final manuscript.
Acknowledgements
The authors are grateful to the followings: Prof. J.A. Akinwande, Dr. A.O.Fasola, Prof. V.I. Ugboko, Dr. H.O. Olasoji, Dr. G.T. Arotiba and Dr. J.T. Arotiba, for their assistance during the preparation of this manuscript.
==== Refs
Ugboko VI Odusanya SA Fagade OO Maxillofacial fractures in a semi- urban Nigerian teaching hospital. A review of 442 cases Int J Oral Maxillofac Surg 1998 27 286 289 9698176
Kihlbert JK Head injury in automobile accidents Automobile Injury Research Report 1965 No VJ-1823-R17.
Nahum AM Siegel AW Brooks S The reduction of collision injuries. Past, present and future Proc 14th STAPP Car Conference 1970 New York: Society of Automobile Engineers 1 43
Olasoji HO Tahir A Arotiba GT Changing picture of facial fractures in northern Nigeria Br J Oral Maxillofac Surg 2002 40 140 143 12180207 10.1054/bjom.2001.0716
Khan AA A retrospective study of injuries to the maxillofacial skeleton in Harare, Zimbabwe Br J Oral Maxillofac Surg 1988 26 435 439 3191094 10.1016/0266-4356(88)90098-8
Magennis P Shepherd J Hutchison I Brown A Trends in facial injuries: increasing violence more than compensate for decreasing road trauma BMJ 1998 316 325 332 9487161
Van Hoof RF Mrérlex CA Stekelenberg EC The different pattens of fractures of the facial skeleton in four european countries Int J Oral Surg 1977 6 3 11 402321
Afzeilus LE Rosen C Facial fractures: A review of 368 cases Int J Oral Surg 1980 9 25 33 6769827
Brown RD Cowpe JG Patterns of maxillofacial trauma in two different cultures J R Coll Surg Edinb 1985 30 299 302 4078777
King RE Scianna JM Petruzzelli GJ Mandible fracture patterns: a suburban trauma center experience Am J Otolaryngol 2004 25 301 307 15334392 10.1016/j.amjoto.2004.03.001
Laski R Ziccardi VB Broder HL Janal M Facial trauma: a recurrent disease? The potential role of disease prevention J Oral Maxillofac Surg 2004 62 685 688 15170278 10.1016/j.joms.2003.12.008
Ansari MH Maxillofacial fractures in Hamedan province, Iran: a retrospective study (1987–2001) J Craniomaxillofac Surg 2004 32 28 34 14729047
Al Ahmed HE Jaber MA Abu Fana SH Karas M The pattern of maxillofacial fractures in Sharjah, United Arab Emirates: a review of 230 cases Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2004 98 166 170 15316543 10.1016/j.tripleo.2004.01.020
Erol B Tanrikulu R Gorgun B Maxillofacial fractures: Analysis of demographic distribution and treatment in 2901 pateints (25-year expereience) J Craniomaxillofac Surg 2004 32 308 313 15458673
Oji C Maxillofacial injuries Plast Reconstr Surg 1996 97 866 868 8628787
Oji C Jaw fractures in Enugu, Nigeria, 1985–1995 Br J Oral Maxillofac Surg 1999 37 106 109 10371311 10.1054/bjom.1997.0083
Nwoku AL Oluyadi BA Retrospective analysis of 1206 maxillofacial fractures in an urban Saudi hospital: 8 year review Pakistan Oral Dent J 2004 24 13 16
Fasola AO Nyako EA Obiechina AE Arotiba JT Trends in the Characteristics of Maxillofacial Fractures in Nigeria J Oral Maxillofac Surg 2003 61 1140 1143 14586847 10.1016/S0278-2391(03)00671-2
Fasola AO Obiechina AE Arotiba JT An audit of midfacial fractures in Ibadan, Nigeria Afr J Med Med Sci 2001 30 183 186 14510125
Telfer MR Jones GM Shepherd JP Trends in the aetiology of maxillofacial fracture in the United Kingdom (1977–1987) Br J Oral Maxillofacial Surg 1991 29 250 255
Ajagbe HA Daramola JO Oluwasanmi JO Civilian-type facial injuries- A retrospective study of cases seen at the University College Hospital, Ibadan Nigerian Med J 1977 7 432 436
Abiose BO Maxillofacial skeleton injuries in the Western states of Nigeria Br J Oral Maxillofac Surg 1986 24 31 39 2936381 10.1016/0266-4356(86)90037-9
Nyako AE Review of facial fractures at the University College Hospital Ibadan (1978–1982) 1985 Thesis, West African Postgraduate Medical College, Lagos, Nigeria
Fasola AO Lawoyin JO Obiechina AE Arotiba JT Inner city maxillofacial fractures due to road traffic accidents Dent Traumatol 2003 19 2 5 12656847 10.1034/j.1600-9657.2003.00073.x
Fasola AO Denloye OO Obiechina AE Arotiba JT Facial bone fractures in Nigerian children Afr J Med Med Sci 2001 30 67 70 14510154
Fasola AO Obiechina AE Arotiba JT Fractures of the mandible in children East Afr Med J 2001 78 61 3
Fasola AO Obiechina AE Arotiba JT Incidence and pattern of maxillofacial fractures in the elderly Int J Oral Maxillofac Surg 2003 32 206 208 12729784 10.1054/ijom.2002.0323
Fasola AO Obiechina AE Arotiba JT Sports related maxillofacial fractures in 77 Nigerian patients Afr J Med Med Sci 2000 29 215 217 11713992
Fasola AO Obiechina AE Arotiba JT Concomitant injuries in 531 patients with maxillofacial fractures Afr J Med Med Sci 2002 31 101 105 12518901
Fasola AO Obiechina AE Arotiba JT Soft tissue injuries of the face: a 10 year review Afr J Med Med Sci 2000 29 59 62 11379471
Fasola AO Obiechina AE Arotiba JT Zygomatic complex fractures at the University College Hospital, Ibadan, Nigeria East Afr Med J 2002 79 137 139 12389959
Abiose BO The incidence and management of middle third facial fractures at the University Teaching Hospital, Ibadan East Afr Med J 1991 68 164 173 2070751
Denloye OO Fasola AO Arotiba JT Dental emergencies in children seen at University College Hospital (UCH), Ibadan, Nigeria- 5 year review Afr J Med Med Sci 1998 27 197 199 10497648
Ajagbe HA Daramola JO Pattern of facial bone fractures seen at the University Teaching Hospital, Ibadan, Nigeria East African Med J 1980 57 267 273
Bankole OO Fasola AO Denloye OO Oro-facial soft tissue injuries in Nigerian children: a five-year review Afr J Med Med Sci 2004 33 93 97 15565923
Nwoku AL Akinosi JO Solarin EO Obisesan BA Results of out-patient treatment of mandibulo-maxillary fractures in LUTH Niger Med J 1979 9 100 106 463320
Akinwande JA Fracture of the mandible due to road traffic accidents in Lagos, Nigeria Nig Dent J 1990 9 15 22
Arotiba GT Facial fractures due to road traffic accidents in Lagos Niger Med J 1996 30 66 69
Arotiba GT The aetiology of facial fractures in Lagos Niger Postgrad Med J 1996 3 37 42
Akinwande JA Ogunbanjo B Ladeinde A Ogunlewe O Obisesan B Oluseye T Armed robbery gunshot injuries to the maxillofacial region Journal clinical Practice 1998 1 9 14
Odusanya SA Maxillofacial fractures in South Western Nigeria (1976–1981) Odontostomatologie Tropicale 1985 3 153 156
Ugboko VI Owotade FJ Oginni FO Odusanya SA Gunshot injuries of the orofacial region in Nigerian civilians SADJ 1999 54 418 422 10860051
Oginni FO Akinwande JA Fagade OO Arole GF Odusanya SA Facial dog bites in Southwestern Nigerian children. An analysis of eight cases Trop Doc 2002 32 239 240
Oginni FO Fagade OO Akinwande JA Arole GF Odusanya SA Pattern of soft tissue injuries to the oro-facial region in Nigerian children attending a teaching hospital Int J Paediatr Dent 2002 12 201 206 12028312 10.1046/j.1365-263X.2002.00353.x
Ugboko V Udoye C Ndukwe K Amole A Aregbesola S Zygomatic complex fractures in a suburban Nigerian Population Dent Traumatol 2005 21 70 75 15773885 10.1111/j.1600-9657.2004.00275.x
Ugboko V Odusanya S Ogunbodede E Maxillofacial fractures in children: An analysis of 52 Nigerian cases Ped Dent J 1998 8 31 35
Adekeye EO The pattern of fractures of the facial skeleton in Kaduna, Nigeria. A survey of 1,447 cases Oral Surg 1980 6 491 495 6930068 10.1016/0030-4220(80)90068-7
Adekeye EO Fractures of the zygomatic complex in Nigerian patients J Oral Surg 1980 38 596 9 6930461
Adekeye EO Pediatric fractures of the facial skeleton: a survey of 85 cases from Kaduna, Nigeria J Oral Surg 1980 38 355 358 6928934
Adebayo ET Ajike OS Adekeye EO Analysis of the pattern of maxillofacial fractures in Kaduna, Nigeria Br J Oral Maxillofac Surg 2003 41 396 400 14614869 10.1016/S0266-4356(03)00165-7
Ugboko VI Olasoji HO Ajike SO Amole AOD Ogundipe OT Facial injuries caused by animals in northern Nigeria Br J Oral Maxillofac Surg 2002 40 433 437 12379192 10.1016/S026643560200181X
Ogunbodede EO Arotiba JT Camel bite injuries of the orofacial region: report of a case J Oral Maxillofac Surg 1997 55 1174 1176 9331246 10.1016/S0278-2391(97)90303-7
Olasoji HO Maxillofacial injuries due to assault in Maiduguri, Nigeria Trop Doc 1999 29 106 108
Olasoji HO Tahir A Bukar A Jaw fractures in Nigerian children: an analysis of 102 cases Cent Afr J Med 2002 48 109 112 14562532
Oji C Fractures of the facial skeleton in children: a survey of patients under the age of 11 years J Craniomaxillofac Surg 1998 26 322 325 9819684
Saheeb BDO Etetafia MO Influence of positions on the incidence and severity of maxillofacial injuries in vehicular crashes West Afr J Med 2003 22 146 149 14529225
van Beek GJ Merkx CA Changes in the pattern of fractures of the maxillofacial skeleton Int J Oral Maxillofac Surg 1999 28 424 428 10609743 10.1034/j.1399-0020.1999.280605.x
Kontio R Suuronen R Ponkkonen H Lindqvist C Laine P Have the causes of maxillofacial fractures changed over the last 16 years in Finland? An epidemiological study of 725 fractures Dent Traumatol 2005 21 14 19 15660750 10.1111/j.1600-9657.2004.00262.x
Kraus JF Riggins RS Franti CE Some epidemiological features of motorcycle collision injuries Am J Epid 1975 102 74 113
Deaner RM Fitchett VH Motorcycle trauma J Trauma 1975 15 678 681 1152089
Adegbehingbe BO Oluwadiya KS Adegbehingbe OO Motorcycle associated ocular injuries in Ile-Ife, Nigeria African Journal of Trauma 2004 2 35 39
Kobusingye OC Why poor countries cannot afford to ignore road safety African Journal of Trauma 2004 2 6
Robinson S Hope for Nigeria Newsweek International 1999 7 23
Onyeaso CO Adegbesan OA Orofacial injury and mouthguard usage by athletes in Nigeria Int Dent J 2003 53 231 236 12953891
Wolff KD Management of animal bite injuries of the face: expereience with 94 patients J Oral Maxillofacial Surg 1998 56 838 843 10.1016/S0278-2391(98)90009-X
Fourie L Cartilidge D Fracture of the maxilla following dog bite to the face Injury 1995 26 61 62 7868216 10.1016/0020-1383(95)90557-E
Kelly DE Harrigan WE A survey of facial fractures: Bellevue Hospital, 1948–1974 J Oral Surg 1975 33 146 149 1054389
Tuli T Hachl O Rasse M Kloss F Gassner R Dentoalveolar trauma analysis of 4763 patients with 6237 injuries in 10 years Mund Kiefer Gesichtschir 2005 July 2
Gassner R Tuli T Hachl O Rudisch A Ulmer H Cranio-maxillofacial trauma: a 10 year review of 9,543 cases with 21,067 injuries J Craniomaxillofac Surg 2003 31 51 61 12553928
Gassner R Tuli T Hachl O Moreira R Ulmer H Craniomaxillofacial trauma in children: a review of 3,385 cases with 6,060 injuries in 10 years J Oral Maxillofac Surg 2004 62 399 407 15085503 10.1016/j.joms.2003.05.013
Ellis E 3rdScott K Assessment of patients with facial fractures Emerg Med Clin North Am 2000 18 411 448 10967733 10.1016/S0733-8627(05)70137-1
Aziz SR Ziccardi VB Borah G Current therapy: complications associated with rigid internal fixation of facial fractures Compend Contin Educ Dent 2005 26 565 571 16101097
Van Sickels JE A review and update of new methods for immobilisation of the mandible Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2005 100 S11 S16 16037787 10.1016/j.tripleo.2005.04.003
|
16270942
|
PMC1277015
|
CC BY
|
2021-01-04 16:37:14
|
no
|
Head Face Med. 2005 Oct 4; 1:7
|
utf-8
|
Head Face Med
| 2,005 |
10.1186/1746-160X-1-7
|
oa_comm
|
==== Front
Plant MethodsPlant Methods1746-4811BioMed Central London 1746-4811-1-31627092410.1186/1746-4811-1-3MethodologyHigh throughput, high resolution selection of polymorphic microsatellite loci for multiplex analysis Cryer Nicholas C [email protected] David R [email protected] Mike J [email protected] School of Biological Sciences, University of Reading, Reading, Berkshire, RG6 6AS, UK2 Cocoa Research Unit, The University of West Indies, St. Augustine, Trinidad and Tobago2005 18 8 2005 1 3 3 25 5 2005 18 8 2005 Copyright © 2005 Cryer et al; licensee BioMed Central Ltd.2005Cryer 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
Large-scale genetic profiling, mapping and genetic association studies require access to a series of well-characterised and polymorphic microsatellite markers with distinct and broad allele ranges. Selection of complementary microsatellite markers with non-overlapping allele ranges has historically proved to be a bottleneck in the development of multiplex microsatellite assays. The characterisation process for each microsatellite locus can be laborious and costly given the need for numerous, locus-specific fluorescent primers.
Results
Here, we describe a simple and inexpensive approach to select useful microsatellite markers. The system is based on the pooling of multiple unlabelled PCR amplicons and their subsequent ligation into a standard cloning vector. A second round of amplification utilising generic labelled primers targeting the vector and unlabelled locus-specific primers targeting the microsatellite flanking region yield allelic profiles that are representative of all individuals contained within the pool. Suitability of various DNA pool sizes was then tested for this purpose. DNA template pools containing between 8 and 96 individuals were assessed for the determination of allele ranges of individual microsatellite markers across a broad population. This helped resolve the balance between using pools that are large enough to allow the detection of many alleles against the risk of including too many individuals in a pool such that rare alleles are over-diluted and so do not appear in the pooled microsatellite profile. Pools of DNA from 12 individuals allowed the reliable detection of all alleles present in the pool.
Conclusion
The use of generic vector-specific fluorescent primers and unlabelled locus-specific primers provides a high resolution, rapid and inexpensive approach for the selection of highly polymorphic microsatellite loci that possess non-overlapping allele ranges for use in large-scale multiplex assays.
MultiplexMicrosatelliteHigh ThroughputFluorescentDinucleotideHigh-ResolutionAllelic Ladder
==== Body
Background
Microsatellite analysis using fluorescently labelled primers and capillary fractionation is the pre-eminent method for the genetic analysis of eukaryotic organisms. The approach is routinely used for many applications including forensic analysis [1], linkage mapping and association genetics [2], population genetics [3,4] and genetic analysis of diversity [5]. The need to screen microsatellite loci for polymorphism between genotypes within the target organism, and for their suitability in multiplex analysis, is an inevitable part of such efforts. The high cost of fluorescently labelled primers has meant that selection of microsatellite markers has typically relied on initial, low-resolution screens of unlabelled primers prior to high-resolution marker selection using labelled primers. However, the preliminary screen is inevitably crude and inefficient, making it either prone to error or reliant upon the more expensive high-resolution selection. There is therefore a need for a high throughput and high-resolution single-step method of selecting appropriate microsatellite markers for genetic studies [6].
For multiplex analysis, greatest efficiency is achieved when utilising many polymorphic loci possessing closely spaced, non-overlapping allelic ranges. Unexpected allelic range overlap between multiplexed microsatellite loci yields ambiguous alleles that may be misassigned to an inappropriate locus, compromising the integrity of the data set. One inevitable problem lies in the possibility that the screen does not encompass all alleles present in the population under study. Confidence in the definition of allelic ranges is invariably a function of the number and diversity of genotypes screened. There is therefore a balance between the desire to examine many individuals and the cost of doing so using fluorescently labelled primers. Thus, screening invariably becomes expensive as the number of genotypes tested grows, and as the number of discarded markers increases. Common approaches to selecting microsatellite markers for multiplex use include assembling panels from previously fluorescently characterised individual markers [7,8], and pre-screening markers on polyacrylamide gels utilising radioactivite labelling of PCR products [9]. Several authors have proposed low cost alternatives for preliminary screens using direct DNA staining following polyacrylamide gel electrophoresis [10-13]. Such strategies have merit, but are labour-intensive, cannot assign actual size ranges and generally lack the resolution required to accurately predict polymorphism in dinucleotide markers [10-13]. One methodology that is able to generate high resolution allelic ladders in a similar fashion to the method reported here is that of Oetting [14]. This method employs the use of locus specific primers tailed with generic sequence allowing a second round of labelled PCR and subsequent capillary fractionation. This method however suffers from a number of potential disadvantages relative to our method. The use of long oligonucleotides for PCR of genomic templates at below optimal annealing temperature allows for an increased frequency in the production non specific amplification products. The PCR amplification conditions required for locus specific amplification using tailed oligonucleotides are often different to those conditions optimal for amplification with equal length 20 mer oligonucleotides. The method of Oetting is not suited to the genotypic analysis of dinucleotide repeat markers due the possibility of extensive stutter profiles generated by the second round of PCR complicating the allelic profiles and so is only considered of merit for marker selection.
Here, we propose a simple but novel approach in which microsatellite amplicons generated from pooled genomic DNA templates are ligated into a standard cloning vector, re-amplified using a labelled universal primer targeting the plasmid insert flanking region, and an unlabelled locus-specific primer. The resultant profiles represents allelic ladders derived from the component alleles contained by the pooled DNA. This procedure thereby offers a single assay, high resolution and inexpensive means of screening microsatellite loci for polymorphism and allelic size range. The profile also offers a qualitative indication of the locus with regard to stutter, a problem often associated with the use of dinucleotide repeat markers for genetic analysis, but also of interest to laboratories utilising tri- and tetranucletide repeat markers.
Results
When employing a pooling strategy, there is a balance between sampling extensively to encompass the full range of variation, and dilution of individuals within the pool such that rare alleles are not detected. DNA pools were created by combining equal amounts of individual DNA samples before dilution with nano-pure water to 5 ng·μL-1. To select the most appropriate pool size, while allowing the detection of rare alleles, DNA pools of 8, 12, 16, 24, 32, and 48 individuals were compared. PCR amplification of pooled DNA utilising unlabelled primer pairs specific to single microsatellite loci [15] were performed incorporating 5 ng template DNA with AccuPrime Taq DNA Polymerase in supermix I, using half recommended volumes (Invitrogen Ltd). The thermal cycling protocol was 96°C for 2 min; 35 cycles of 96°C for 30 s, 51°C or 46°C for 30 s dependent on primer annealing characteristics, 72°C for 2 min; followed by 72°C for 10 min, in a MJ Research PTC-100 thermal cycler (Genetic Research Instrumentation Ltd). Successful PCR was confirmed by 1.5% (w/v) agarose gel electrophoresis [16]. PCR products for multiple individual microsatellite loci, amplified from aliquots of the same template DNA pool, were combined then purified using NucleoFast 96 PCR cleanup plates (Macherey-Nagel GmbH & Co. KG), before ligation into pDrive vector (Qiagen Ltd). Ligation products were diluted 1/10 with HPLC grade water and used as template for a second round of PCR using the 'reverse' microsatellite specific primer and a generic fluorescently labelled primer, M13 (-40), targeting the plasmid. Labelled amplicons were diluted 1/100 in HPLC grade water and fractionated by capillary electrophoresis on an ABI 3100 and viewed with genotyper 3.7 software (Applied Biosystems UK Ltd). The allele size reported by this method is that expected from the microsatellite primers plus an additional 150 bases of vector sequence. Comparison of profiles from pooled amplifications and those of constituent members of the pools demonstrated that homozygous individuals possessing a rare allele could be detected reliably in pools of 12 individuals or less. Thus, one strategy would be to assemble several small pools, allowing variance in allelic limits to be described, and continue screening until the addition of more pools no longer increases the allele range. In practice, however, it may be preferable to use much larger pools and accommodate for uncertainty over rare alleles by imposing buffer zones around detected allelic ranges prior to multiplexing. We empirically tested this approach. Template DNA from 96 diverse genotypes of Theobroma cacao was adjusted to 5 ng·uL-1, pooled and individually amplified by PCR for 84 dinucleotide cocoa microsatellite markers described by Pugh et al [15]. The complex profiles generated (Figure 1) broadly represent the array of alleles present when genotypes were assayed individually. We therefore selected 36 markers generating the widest range of homogeneous peaks for further study. The allelic range of peak sizes is taken to be indicative of the allelic range in the unsampled gene pool. In general, loci generating large numbers of peaks with approximately even height (Figure 1B,C) were highly informative whereas those producing few peaks (Figure 1A) or profiles dominated by one peak (Figure 1D) have less utility for genetic analysis.
Figure 1 Representative capillary electrophoresis traces for allelic screening of dinucleotide microsatellite loci against pooled DNA samples. Microsatellite loci are A) mTcCIR080; B) mTcCIR131; C) mTcCIR155; D) mTcCIR190.
We then examined the relationship between predicted allele range in pooled profiles and that observed after wider genotype sampling. For this, we employed multiplex PCR microsatellite analyses performed on 672 individual cocoa genotypes (Table 1). Two loci predicted to yield few alleles on the basis of the sample pool (mTcCIR080 and mTcCIR155) produced the same number of alleles when the sample range was expanded to include 672 individual cocoa genotypes. However, the number of alleles in more variable loci (mTcCIR131 and mTcCIR190) increased from 9 to 12 when the sample range was expanded, with the size range increasing by 67% and 62% respectively. Given a modest increase in allelic range when sample size was increased six fold, one approach would be to accommodate undetected alleles by imposing a buffer between the ranges of neighbouring loci prior to multiplexing. In this case, a spacing of 1× predicted range either side of the mean allele size would appear adequate. Overall, adoption of this protocol allows for improved selection of compatible polymorphic microsatellite markers, with reduced likelihood of producing overlapping profiles in multiplexed microsatellite reactions.
Table 1 Comparison of predicted and observed microsatellite allele frequency. The number of alleles each locus was predicted to generate as described in Figure 1 was compared to the actual alleles observed when screened over 672 genotypes of wild, uncultivated cocoa. Predictions were based on the height and number of peaks reported from the pooled samples and took account of the extra DNA amplified from the pDrive vector when predicting the size of the DNA fragments.
PREDICTED OBSERVED
Locus min max mode alleles min max mode alleles
mTcCIR080 97 103 101 4 97 105 99 4
mTcCIR131 198 212 212 9 185 214 210 12
mTcCIR155 264 274 272 5 265 275 267 5
mTcCIR190 156 172 162 9 148 174 161 12
mTcCIR065 230 250 240 6 231 255 237 11
mTcCIR066 280 310 287 8 280 308 284 9
mTcCIR069 185 205 202 11 175 206 202 13
mTcCIR088 182 197 189 7 180 200 187 8
mTcCIR092 277 286 282 3 269 284 279 6
mTcCIR103 90 116 112 6 88 128 110 16
mTcCIR113 130 150 142 8 124 153 133 15
mTcCIR158 210 220 212 4 205 227 212 7
mTcCIR172 125 135 126 9 115 140 124 14
mTcCIR195 335 351 348 5 319 349 349 10
mTcCIR203 210 220 216 5 212 218 216 5
mTcCIR266 170 200 177 9 165 206 200 13
Conclusion
Adoption of this methodology allows for both a qualitative and semi quantitative characterisation of polymorphism at individual microsatellite loci. When using pooled samples, combining DNA from up to 12 individuals allowed for the reliable detection of single copy alleles within that sample. If characterising microsatellite loci using DNA pools of greater than 12 individuals the incorporation of a buffer zone into the final genotyping assay, based on the observed range of allele sizes, can allow for the variability likely to be encountered in a larger sample size. The methodology is suitable for high throughput applications by the combination of differing fluorescent dyes in association with convenient liquid handling formats. This protocol benefits from initially utilising unlabelled primers identical to those used in the final genotyping assay, reducing the possibilities of unexpected banding patterns due to changes in primer sequence or assay conditions. The high resolution DNA size measurement makes this protocol suitable for characterising dinucleotide microsatellite loci.
List of abbreviations
PCR, Polymerase chain reaction; DNA, Deoxyribonucleic acid; ng, 10-9 gram
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
NCC conceived and developed the microsatellite selection protocol and drafted the manuscript. The overall project was conceived by DRB. MJW aided the experimental design and played a major role in developing the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
We thank Olivier Sounigo, and Claire Lanaud for providing DNA, Didier Clement for providing as then unpublished microsatellite information (Pugh et al., 2004) and Steve Brown for providing a linkage map of cocoa microsatellite markers. This work forms part of a collaboration between The University of Reading and the Cocoa Research Unit, The University of West Indies, Trinidad and Tobago, funded by The Biscuit, Cake, Chocolate and Confectionery Association, United Kingdom.
==== Refs
Carey L Mitnik L Trends in DNA forensic analysis Electrophoresis 2002 23 1386 1397 12116148 10.1002/1522-2683(200205)23:10<1386::AID-ELPS1386>3.0.CO;2-M
Laan M Paabo S Demographic history and linkage disequilibrium in human populations Nat Genet 1997 17 435 438 9398845 10.1038/ng1297-435
PerezLezaun A Calafell F Seielstad M Mateu E Comas D Bosch E Bertranpetit J Population genetics of y-chromosome short tandem repeats in humans J Mol Evol 1997 45 265 270 9302320
PerezLezaun A Calafell F Mateu E Comas D Ruiz-Pacheco R Bertranpetit J Microsatellite variation and the differentiation of modern humans Hum Genet 1997 1 1 7 9003483
Tsutsui ND Suarez AV Holway DA Case TJ Reduced genetic variation and the success of an invasive species Proc Natl Acad Sci USA 2000 97 5948 5953 10811892 10.1073/pnas.100110397
Frasier TR Wilson PJ White BN Rapid screening of microsatellite markers for polymorphisms using SYBR® green 1 and a DNA sequencer BioTechniques 2004 36 408 409 15038155
Narvel JM Chu WC Fehr WR Cregan PB Shoemaker RC Development of multiplex sets of simple sequence repeat DNA markers covering the soybean genome Molecular Breeding 2000 6 175 183 10.1023/A:1009637119947
Tang S Kishore VK Knapp SJ PCR-multiplexes for a genome-wide framework of simple sequence repeat marker loci in cultivated sunflower Theor Appl Genet 2003 107 6 19 12835928
Tommasini L Batley J Arnold GM Cooke RJ Donini P Law JR Lowe C Moule C Trick M Edwards KJ The development of multiplex simple sequence repeat (SSR) markers to complement distinctness, uniformity and stability testing of rape (Brassica napus L.) varieties Theor Apl Genet 2003 106 1091 1101 12671758
Morin PA Smith DG Non-radioactive detection of hypervariable simple sequence repeats in short polyacrylamide gels BioTechniques 1995 19 223 228 8527143
Scrimshaw BJ Non-radioactive detection of hypervariable simple sequence repeats in short polyacrylamide gels BioTechniques 1992 13 188 1382464
White HW Kusukawa N Agarose-based system for separation of short tandem repeat loci BioTechniques 1997 22 976 980 9149885
Houriham RN O'Sullivan GC Morgan JG High-resolution detection of loss of heterozygosity of dinucleotide microsatellite markers BioTechniques 2001 30 342 346 11233603
Oetting WS Lee HK Flanders DJ Wiesner GL Sellers TA King RA Linkage analysis with multiplexed short tandem repeat polymorphisms using infrared fluorescence and M13 tailed primers Genomics 1995 30 450 458 8825630 10.1006/geno.1995.1264
Pugh T Fouet O Risterucci AM Brottier P Abouladze M Deletrez C Courtois B Clement D Larmande P N'Goran JAK Lanaud C A new cacao linkage map based on codominant markers: development and integration of 201 new microsatellite markers Theor App Genet 2004 108 1151 1161 14760486 10.1007/s00122-003-1533-4
Sambrook J Fitsch EF Maniatis T Molecular Cloning: A Laboratory Manual 1989 Cold Spring Harbor, Cold Spring Harbor Press
|
16270924
|
PMC1277016
|
CC BY
|
2021-01-04 16:37:28
|
no
|
Plant Methods. 2005 Aug 18; 1:3
|
utf-8
|
Plant Methods
| 2,005 |
10.1186/1746-4811-1-3
|
oa_comm
|
==== Front
Plant MethodsPlant Methods1746-4811BioMed Central London 1746-4811-1-41627093810.1186/1746-4811-1-4MethodologyA rapid and versatile combined DNA/RNA extraction protocol and its application to the analysis of a novel DNA marker set polymorphic between Arabidopsis thaliana ecotypes Col-0 and Landsberg erecta Berendzen Kenneth [email protected]_tuebingen.deSearle Iain [email protected] Dean [email protected] Csaba [email protected] Alfred [email protected] George [email protected] Imre E [email protected]Ülker Bekir [email protected] Max-Planck-Institute for Plant Breeding Research, Department of Developmental Biology, Carl-von-Linné Weg 10, D-50829 Köln, Germany2 Philipps-Universität, Biology-Plant Physiology/Photobiology, Karl-von-Frisch-Str. 8, D-35032 Marburg, Germany3 Max-Planck-Institute for Plant Breeding Research, Department of Plant-Microbe Interactions, Carl-von-Linné Weg 10, D-50829 Köln, Germany2005 23 8 2005 1 4 4 25 5 2005 23 8 2005 Copyright © 2005 Berendzen et al; licensee BioMed Central Ltd.2005Berendzen 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 established PCR-based approaches in plant molecular biology rely on lengthy and expensive methods for isolation of nucleic acids. Although several rapid DNA isolation protocols are available, they have not been tested for simultaneous RNA isolation for RT-PCR applications. In addition, traditional map-based cloning technologies often use ill-proportioned marker regions even when working with the model plant Arabidopsis thaliana, where the availability of the full genome sequence can now be exploited for the creation of a high-density marker systems.
Results
We designed a high-density polymorphic marker set between two frequently used ecotypes. This new polymorphic marker set allows size separation of PCR products on agarose gels and provides an initial resolution of 10 cM in linkage mapping experiments, facilitated by a rapid plant nucleic acid extraction protocol using minimal amounts of A. thaliana tissue. Using this extraction protocol, we have also characterized segregating T-DNA insertion mutations. In addition, we have shown that our rapid nucleic acid extraction protocol can also be used for monitoring transcript levels by RT-PCR amplification. Finally we have demonstrated that our nucleic acid isolation method is also suitable for other plant species, such as tobacco and barley.
Conclusion
To facilitate high-throughput linkage mapping and other genomic applications, our nucleic acid isolation protocol yields sufficient quality of DNA and RNA templates for PCR and RT-PCR reactions, respectively. This new technique requires considerably less time compared to other purification methods, and in combination with a new polymorphic PCR marker set dramatically reduces the workload required for linkage mapping of mutations in A. thaliana utilizing crosses between Col-0 and Landsberg erecta (Ler) ecotypes.
==== Body
Background
PCR and RT-PCR (reverse transcriptase-PCR) are the most widely used analytical methods in plant genetics and molecular biology, providing simple tools for studying the segregation of mutations and monitoring the transcription of wild type and mutant alleles of genes in different plant tissues. Like several other classical methods (e.g. Southern and northern hybridization analysis of nucleic acids), PCR and RT-PCR applications also require a sufficient amount and quality of nucleic acids suitable for these assays. Because most well-established protocols include procedures based on the use of either potentially toxic chemicals or expensive commercial kits, numerous quick DNA isolation methods have been developed to promote large-scale genomic applications during the past years [1]. One of the major disadvantages of these quick isolation methods is that they are not suitable for applications requiring amplification of DNA fragments greater than 2 kb in size. Additionally, the available DNA purification methods have not been combined with rapid isolation of RNA from plant tissue.
Upon completion of the A. thaliana genome sequence, a major goal of post-genomic research is to understand the function and regulation of over 26000 genes in this model species. Several EMS (ethylmethansulfonate) and T-DNA mutagenized populations offer valuable genetic resources for wide-scale functional genomics studies in A. thaliana [2-6]. These studies require high-throughput DNA and RNA isolation from tens to thousands of plants. Progress in the molecular and genetic characterization of EMS and T-DNA insertion mutants is thus largely dependent on the speed, simplicity and quality of nucleic acid isolation methods. Map-based cloning of mutant alleles generated by EMS or radiation mutagenesis has been simplified by developing various pooling strategies, which are aided by well-characterized molecular markers. Many linkage mapping techniques are based either on enzymatic digestion of PCR products [7], or on the use of SNaP shot assays (SNP polymorphism markers; [8]), which may require separation on large, labor intensive acrylamide gels and detection by silver staining or radiography. Linkage mapping strategies in A. thaliana are still restricted by the number of known polymorphisms available between various ecotypes, such as Col-0 and Ler. The frequent use of segregating populations derived from Col-0 × Ler crosses, especially in the study of flowering time and plant development, would thus significantly benefit from a larger set of well-characterized and tested markers. Therefore, to develop a facile map based cloning approach, we refined the current design of polymorphic markers such that all polymorphic markers can now be resolved on 3% (w/v) agarose gels and detected by ethidium bromide staining.
To facilitate high-throughput application of our new polymorphic markers, as well as PCR-based identification and characterization of insertion mutants, we have developed a simple technique for rapid nucleic acid isolation from minimal quantity of A. thaliana tissue. This technique isolates both DNA and RNA templates, the quantity and quality of which are sufficient for PCR and RT-PCR analyses, respectively. Our data illustrate that this new technique can considerably accelerate PCR screening for T-DNA knockout mutations and greatly facilitates tracking segregating progeny of Col-0 × Ler crosses, which we used for identification of an EMS-induced mutation affecting the regulation of flowering time in A. thaliana.
Results
The Sucrose Prep method for rapid isolation of nucleic acid templates for PCR analysis
We have systematically tested various protocols optimized for DNA isolation at room temperature for the efficiency of PCR amplification while omitting the inhibitory component EDTA. Combinations of DNA isolation protocols from Edwards et al. [9] and Walbot and Warren [10] were thus compared using ~2.5 mg (~3 mm2) of A. thaliana leaf tissue in 50 μl of extraction mixture, from which 1 μl was used as template in a routine PCR application. We have found that one of the variant extraction buffers, hereafter called Sucrose Solution, exhibited no change in the efficiency of PCR amplification judged by ethidium bromide-stained agarose gels in response to varying the pH from 7.0 to 8.0, or changing the salt and sucrose concentrations between 200 to 400 mM NaCl and 300 to 440 mM sucrose, respectively (data not shown). Due to the presence of sucrose in the extraction buffer, we named this nucleic acid extraction method the Sucrose Prep. In agreement with Thomson and Henry [11], we found that heating the crude extract for 10 min followed by centrifugation for 5 sec at 6000 g eliminated nearly all debris that interferes with PCR amplification.
The Sucrose Prep protocol was further optimized by controlling sampling size. Optimal results were obtained by harvesting of leaf tissue with a 500 μl Eppendorf cap punch (yielding about 10 mg fresh tissue) in 200 μl of Sucrose Solution. Multiple samples were stored on ice prior to extraction and were ground using either sterile pipette tips or plastic pestles. The extracts could be stored at 4°C overnight or at -20°C for longer term storage, but were normally used immediately. We could store samples at -20°C for up to 4 weeks with no detectable decrease in PCR amplification efficiency. Following long term storage however, samples required re-heating before PCR amplification. Successful PCR amplification was obtained with A. thaliana DNA samples extracted from 2 to 4 week-old leaf material, petals, sepals, stigmas, styles, anthers, and embryos. The maximum size of products obtained in routine PCR amplification reactions was around 3 kb. Using optimized thermal cycling conditions, no difference was observed for products up to 4 kb when compared to DNA templates prepared according to Edwards et al. [9]. By limiting the amount of plant debris carried over into the mixture, PCR product sizes of up to 5.5 kb were obtained (Figure 1).
Figure 1 Efficiency of PCR amplification using DNA templates prepared by the Sucrose Prep. (A) Extracts were prepared from various tissues using 10 mg sample/200 μl Sucrose Solution. (1) 4 week-old leaf, (2) senescent leaf, (3) anthers, (4) petals, (5) sepals, (6) gynoecium, (7) petiole, (8) 8–12 days old embryos. (B) Comparison of DNA templates prepared by the methods of Edwards (ED; Edwards et al., [9]) and Sucrose Prep (SP) in long-range PCR amplification. Optimized PCR conditions were used to amplify 2.2, 3.7, and 4 kb size products. (C) Long-range PCR amplification on DNA template isolated by CTAB (CT; see methods) and Sucrose Prep using LA TaKaRa polymerase. Sucrose Prep replicates show the inhibition of PCR amplification by inclusion of plant debris in the reaction (SP1) in comparison to samples avoiding contamination with debris (SP2).
Novel marker set for mapping between A. thaliana ecotypes Landsberg erecta and Col-0
The availability of the entire genome sequence of A. thaliana ecotype Col-0 and the high abundance of genomic sequence from Landsberg erecta (Ler) allowed us to design a new set of markers that show polymorphism between these ecotypes [12,13]. A novel set of 51 DNA markers were identified at regular intervals across all five chromosomes (Table 1; Figure 2A). This marker set allows mapping at a 10 cM (~2 Mbp) resolution.
Table 1 List of chromosomal positions of BSA markers detecting polymorphism between A. thaliana ecotypes Col-0 and Ler.
Chr Marker Forward primer (5'-3') Reverse primer (5'-3') Col-0 (bp) Ler (bp)
I nga63 GCCTAAACCAAGGCACAGAAG TCATCAGTATTCGACCCAAG 87 99
I F3F19 CCACAAAACAATTTGGTTCACTC TCCCGTTGGGGATATTAAAG 100, 143 243
I F20D23 TTATGCCAACTCATGTGGAAAG TGTCAAAGCGTCTGGTTCTG 233 254
I F12K8 ACCAACACCACAACAAACGAC CTTTTTCTGTTCTTCCGCTATTC 171 192
I F3I6 AGATGGAAGAGGAGGAGATGG TGCATGTATATGATGAGCGAGAG 251 311
I SO392 GTTGTTGATCGCAGCTTGATAAG TTTGGAGTTAGACACGGATCTG 142 156
I F7P12 TCGAGGATATGTTTCGTGTTTG ACAGTTTTGATGCATTGTGTGAG 315 100, 215
I F1I21 TCGTAAATTGTGACTGGGAGA CCCTGTAGATCTGTTGTTTTAG 308 113, 195
I ciw1 ACATTTTCTCAATCCTTACTC GAGAGCTTCTTTATTTGTGAT 159 135
I nga280 CCTGATCTCACGGACAATAGTG GGCTCCATAAAAAGTGCACC 106 86
I F23H11 GATATGGGAGTAAGTATGAAATCGG TTCGTCCGGGTAAAAGTCAAG 300 250
I NF514a GTTGAGTCTTGGCATCACAGTTC CTGCCTGAAATTGTCGAAAC 221 240
I F20P5 GATACGTTCAAAATTAGGGACTTC TGTATTTTGCTAATTGAGGTTATGG 218 186
I AthATPase CCTGGGAACGGTTCGATTCGAG GTTCACAGAGAGACTCATAAACCA 86 70
II T20F6 CGTTCGAAACTGAATTAGCTG ACCATCTTTGTTGAGCCCTTC 297 347
II F18P14 ATTCCCGCAATTTATTTTGTTC GTTTGATGGCAGATTTGTTTTC 123 144
II ciw3 GGAAACTCAATGAAATCCACTT GTGAACTTGTTGTGAGCTTTGA 230 200
II F26B6 CTCTATCTGCCCACGAACAAG GCCATTGCAAAAGAACATCAG 233 274
II F16P2 CAGCAATCAAATAACGTGGTG CTCTCTTCTTTCTTCGCCATTAG 237 167
II F4P9 TGGTCCATACCCATTTCATAAC ATGAATTTTCATTCTACTGTTTTG 299 262
II T2H17 ATTGCATACCACGCAGTTCAC CCATTTTGCCCTTTCCTTCTAC 250 274
II AthBIO2b TGACCTCCTCTTCCATGGAG TTAACAGAAACCCAAAGCTTTC 141 209
III nga172 AGCTGCTTCCTTATAGCGTCC CCATCCGAATGCCATTGTTC 162 136
III nga162 CATGCAATTTGCATCTGAGG CTCTGTCACTCTTTTCCTCTGG 107 89
III MSA6 TTGGAGGTGCTCTTAGGTTC GGGCTTTTCACATACGCTTTC 175 225
III ciw11a GTTTTTTCTAATCCCCGAGTTGAG GAAGAAATTCCTAAAGCATTC 192 242
III N7N14 CAATACACTTTATCCAGATGCTG GGGATTTGTTGATTGAAAAAGGAC 150 143
III T6H20 CGGCTGAAACTTGGAAGGGAC AGGAAGAACGTGTGATTGTG 273 293
III ciw4 GTTCATTAAACTTGCGTGTGT TACGGTCAGATTGAGTGATTC 189 215
III K27K19 TGCTTTTGAAGAGATGGTTATTAGG CCCCATTTCACTTATCATTGG 216 198
III nga6a AGCGAATCCGAAAATAATGGAG TGGATTTCTTCCTCTCTTCAC 159 137
IV ciw5 GGTTAAAAATTAGGGTTACGA AGATTTACGTGGAAGCAAT 164 144
IV F14G16 ACAAACCGATCAGCATTCAAG GCCTTTGTCACGGATTCAAC 250 198
IV T3H13 TTTGGTGGGTCAAGAGTCAAG GCAAAAGTCATTACGGACAATAC 275 229
IV T26M18 CAATTAGCGGAGGCCACTTC GGGCAAAAGCTTCCAGTAC 330 271
IV FCALL CCACCGTCAACATCCCTAAC GCTCTTATACTTCTCAGCTCTTGTC 170 180
IV F28A21 GCATCATCATTCATCACCAAC TGTGAAGTGTTTGTCTTTGTG 198 169
IV F16G20 TGTCAACCAATCGCCTTAGTC TTAATGTCCATTATTGGAACGC 113 79
IV F26K10 AGAGAGCACGATGCCTGATAG AATGCTTCAGCGATTGAGAAC 180 205
IV F6E21 TTCTTTGTTCAAGTTCCATGTCTC CGGTGATTGTCTCAAGTGTTTG 199 225
IV F23E13 TGACCGTTGAAAGTGTTGTTG GCCCGAGAAGCCTGATAG 264 246
V MHF15 CTCCTCCTTTAATTTTCTCTCTGTG AGTTCCAGCTTTGGACTTCTTC 295 268
V nga151a ATCTCATACTGACCCATATGTTCC ATTGTACAGTCTAAAAGCGAGAG 198 170
V ciw8a TACTAGTGAAACCTTTCTCAG TTTTATGTTTTCTTCAATCAGTTAG 100 135
V nga139 AGGGTTTCGTTTCACTATCCAG TGAGAGCTACCAGATCCGATG 174 132
V T1N24 CCGATGGCATAACAAGTAGAG GGGAAAGGTACACATATACAAAAGG 383 356
V nga76 GGAGAAAATGTCACTCTCCAC AGGCATGGGAGACATTTACG 231 250
V ciw9 CAGACGTATCAAATGACAAATG GACTACTGCTCAAACTATTCGG 165 145
V MPL12 GTCCCCAAAACCAATCATAAG TCCGAGTGAGAAGAGAGTTTG 319 293
V K9P8 TTATGGGTTTCTCAGAGTTTCTCAC TTGTATGCGTTTGCTTTTTCC 284 251
V MNC17 GTACCGGATCTGTGTTGTGAAG GTGCTCAAGGAAATGGGATAG 168 187
V MQB2 CTTTGATAGTAACCTTTTTCAAACCA TGCCATTTATTTGGTCAACAC 252 231
DNA oligonucleotide primer pairs are listed according to their positions from the top to the bottom of each chromosome. The size (bp) of PCR amplification products from Col-0 and Ler are depicted in the last two columns to the right. F, forward primer and R, reverse primer.
Figure 2 Chromosomal location of polymorphic markers and their use for genotyping. (A) Chromosomal location of the 51 polymorphic Col-0/Ler markers is depicted graphically and the sequence of each primer pair and fragment polymorphism is listed in Table 1. The region on chromosome II that is linked to the early flowering phenotype is boxed. (B) PCR amplification of DNA markers F14G16 on chromosome IV and F26B6 on chromosome II from the parents and early and normal flowering F2 DNA bulks. For DNA marker F14G16 both the Ler and Col-0 alleles were amplified from the early flowering bulk demonstrating that the marker is not linked to the early flowering mutation. Only the Ler allele of DNA marker F26B6 was amplified from the early flowering DNA bulk indicating the marker is linked to the early flowering mutation. (C) PCR amplification of DNA marker F26B6 from the parents and progeny.
Using the new marker set to map a novel early flowering mutant of Ler
To test the efficiency of linkage mapping with the new polymorphic marker set, a Ler line carrying T-DNA constructs overexpressing CONSTANS (CO) and FLOWERING LOCUS C (FLC), as well as a fusion between the promoter of SUPRESSOR of OVEREXPRESSION OF CONSTANS 1 (SOC1) and a beta-glucuronidase (GUS) gene (35S::CO 35S::FLC 1 kb::SOC1:GUS; [14]), was subjected to EMS mutagenesis in order to screen for mutations affecting flowering time. An early flowering mutant displaying an elongated hypocotyl when grown in white light was identified in the M2 generation. After demonstrating that the early flowering mutant phenotype was stably inherited to the M4 generation, the mutant was back-crossed with the progenitor transgenic Ler parent. F1 progeny showed wild type flowering time (hereafter referred as normal flowering time), indicating that the mutation was recessive. Recessive inheritance was confirmed by the analysis of 96 F2 progeny, of which about one quarter showed early flowering. Subsequently, a Col-0 mapping parent was generated by crossing the 35S::CO/35S::FLC/1 kb::SOC1:GUS transgenes from the Ler progenitor to Col-0 four times. The early flowering Ler mutant line was subsequently crossed with the Col-0 mapping parent, and a segregating F2 population was generated to map the mutation. Six early flowering plants were chosen from the F2 generation, and one leaf from each plant was bulked together and DNA purified using the protocol of Edwards et al. [9]. Similarly 6 plants showing a normal flowering phenotype were identified, their leaves were also bulked together and DNA purified.
DNA fragments were amplified from each of the DNA bulks by PCR using the entire new set of 51 poylmorphic markers. Two markers, ciw3 and F26B6 on chromosome II were identified to be linked to the early flowering mutation as both markers were homozygous for the Ler allele from the early flowering bulk and heterozygous for the normal flowering bulk. The other DNA markers were heterozygous in both bulks indicating that they were not linked to the early flowering mutation, with the exception of three markers that most likely were linked to the loci of the T-DNA carrying the CO, FLC and SOC1 gene constructs (data not shown). Figure 2B shows the linked marker F26B6 and an unlinked marker F14G16 amplified from the Col-0 and Ler parents and the DNA bulks of early and normal flowering F2 progeny. DNA marker F26B6 was confirmed to be linked to the mutation by PCR amplification of the marker from five early flowering and four normal flowering plants from the F2 mapping population (Figure 2C).
We then used the Sucrose Prep to rapidly screen 96 early flowering plants from the F2 population, confirming that the mutation was located between the markers ciw3 and F26B6 (data not shown). Analysis of the annotated DNA sequence for candidate genes within this region revealed PHYB as one of the most likely candidates responsible for the early flowering mutant phenotype. Previously, a phyB mutant has been demonstrated to flower earlier under long day conditions and has an elongated hypocotyl under white light conditions [15,16]. Therefore, we sequenced the PHYB gene from wild type Ler and our early flowering mutant. A base substitution of cytosine to thymine was detected at position 1660 bp downstream of the PHYB translational start site, resulting in a premature stop codon in the first exon. This nonsense mutation is predicted to result in a truncated protein that is unlikely to be functional as it contains neither the PAS repeat domain nor the histidine kinase related domain essential for the known function of the protein.
The Sucrose Prep as a method for identification of T-DNA insertion mutations
To determine if the Sucrose Prep method is suitable for screening of homozygous T-DNA mutants, we screened segregating T2 progeny from the SALK_098205 line, in which a T-DNA was inserted in exon 3 of the AtWRKY22 gene (At4g01250; Figure 3). In parallel, we isolated RNA from the same plants using a commercial kit. As illustrated in Figure 3, the Sucrose Prep (DNA; upper panel) produced results that were consistent with the observed expression of the gene (cDNA; lower panel) thereby identifying line 3 as a homozygous loss-of-function mutant of AtWRKY22.
Figure 3 Use of the Sucrose Prep to identify T-DNA insertion knock-out mutations. Putative mutant plants homozygous for the T-DNA insertion were identified with the Sucrose Prep using gene specific primers for AtWRKY22 (Table 2) and amplification of DNA by PCR. The lines were also tested by RT-PCR for the loss of AtWRKY22 transcript using a commercial RNA isolation kit (Qiagen). Positions of the primers used for amplification are indicated below the schematic diagram.
The simplicity of the Sucrose Prep also facilitated the screening for T-DNA insertions in larger populations. For example, the identification of double knockouts carrying T-DNA insertions with identical selectable markers is readily feasible using the Sucrose Prep procedure. To illustrate this point, we performed an experiment to create an atwrky46 (At2g46400), atwrky53 (At4g23810) double mutant. These WRKY transcription factors belong to the same sub-group (group III) and show very similar transcription induction kinetics in response to pathogen and elicitor treatments (data not shown). After crossing the T-DNA insertion lines (Figure 4), F2 progeny were screened by PCR for the loss of gene specific products. Seven candidates were immediately identified as potentially being homozygous for both insertion mutations within 90 plants assayed (Figure 4B). Two of these 7 candidates were subsequently confirmed to be true double knockouts (not shown).
Figure 4 Isolation of homozygous T-DNA insertion mutant lines carrying the atwrky46, atwrky53 double mutation. (A) Schematic structure of atwrky46 and atwrky53 mutant alleles carrying T-DNA insertions. (B) PCR amplification using gene specific primers shown in panel A. Putative lines homozygous for the atwrky46, atwrky53 double knockout mutations fail to amplify the wild-type allele (marked by circles).
The Sucrose Prep can also effectively be used to detect the presence or absence of specific transcripts by RT-PCR
To demonstrate that the Sucrose Prep method is also suitable for facile detection of specific transcripts in small amounts of plant tissue, we tested the expression of SOC1 (At2g45660) and AGC1-10 (At2g26700) genes in leaf samples by RT-PCR amplification of RNA templates prepared by the Sucrose Prep. Since RNA might be destroyed during heating, the extracts from leaf samples were subjected to various heating times of 1 to 5 min prior to RT-PCR amplification. The length of the heating step did not appear to influence the efficiency of RT-PCR amplification since the SOC1 transcript was detected in all samples (Figure 5). Although the amplification of the cDNA product was weaker in comparison to the efficiency obtained with the commercial RNA purification kit, the cDNA product was specific and its amplification was reproducible.
Figure 5 RT-PCR analysis of RNA templates isolated by the Sucrose Prep. (A) Amplification of a specific SOC1 cDNA product from leaf tissue. Leaf material was harvested and frozen in liquid nitrogen then ground in Sucrose Solution and subjected to heating at 99°C for 1, 3 or 5 min. A control PCR was performed with DNA isolated according to Edwards et al. [9]. A second control was performed with cDNA made from fresh leaf material prepared from RNA isolated by extraction using a commercial RNA isolation kit (Qiagen). (B) Isolation of RNA was performed as in (A) and heat treated at 99°C for 5 min. The samples were subjected to RT-PCR analysis using AGC1-10 specific primers, which flank the two introns depicted in the schematic diagram (expected sizes, genomic: 1300 bp, Intron I: 520 bp; Intron II: 215 bp). The cDNA product from the mature mRNA (600 bp) was detectable in shoot meristem tissues (s-m) and weakly in floral tissue, whereas only the first splicing product was observed in leaf tissues due to variations in the RNA yield (i.e., indicated by the amount of DNA product in the reactions).
Next, we tested homozygous T-DNA knockout lines for loss of WRKY transcripts. A T-DNA in the AtWRKY36 (At1g69810) gene was localized within the first intron. This insertion resulted in a loss of detectable transcript by RT-PCR analysis compared to wild type (data not shown). RT-PCR analysis of wild-type and homozygous atwrky36 knockout mutant plants was performed by using the Sucrose Prep. Since the T-DNA was located in the first intron, no amplification of cDNA product was expected in the knockout mutant (Figure 6, panel A). Nonetheless, a faint cDNA product was detected by the RT-PCR assay suggesting that the T-DNA insertion was spliced out from a small fraction of primary transcripts.
Figure 6 Transcript analysis in A. thaliana knockout mutants. (A) Leaf material was harvested from wild type or homozygous atwrky36 knockout mutant plants and the Sucrose Prep was used for RT-PCR analysis with gene specific primers. The T-DNA is located in the first intron of AtWRKY36 and was occasionally transcribed and spliced out since a cDNA product corresponding to a segment of mature RNA is detected in the knockout mutant. The second column is digitally enhanced to highlight the cDNA product detected in the knockout line. (B) Leaf material from wild-type and a homozygous atwrky70 mutant were sampled using the 'Touch-and-Go' method (see text for details). The T-DNA insertion in the gene prevented PCR amplification of both DNA and cDNA products (larger than 4 kb).
The 'Touch-and-Go' approach for PCR and RT-PCR applications
We also developed an alternative method for isolating nucleic acids from very small sample sizes designated 'Touch-and-Go'. This method eliminates the preparation steps required by the Sucrose Prep, since the extraction of DNA/RNA templates is made simply by capturing plant tissue with a pipette tip which is immediately available for the amplification of nucleic acids by PCR and RT-PCR methods respectively. In practice, leaf tissue was punctured using an RNAse-free 20 μl pipette-tip against a firm surface, i.e. a finger covered with a latex glove, and then the pipette-tip was immediately touched into 50 μl PCR solution mix in prepared reaction wells. Due to the very small amount of leaf material taken by pipette tip puncturing, the number of PCR cycles should be 35 to 40 when using this method. For plant material located in the greenhouse or the field, 10 μl water was first delivered to PCR tubes or plates kept on ice, and 'Touch-and-Go' sampling was performed by touching the water in the tubes or wells with the pipette tip containing the plant tissue. After sampling, 40 μl PCR solution mix was added to the tubes or wells in the laboratory and DNA/RNA was amplified by PCR for 40 cycles using a thermocycler. Figure 6, panel B illustrates that the 'Touch-and-Go' method can be used to PCR amplify DNA products a maximum of 1.5 kb in size. This mini-preparation method was also tested in combination with RT-PCR by monitoring for the loss of detectable expression of AtWRKY70 (At3g56400) in a T-DNA insertion mutant line. The atwrky70 insertion mutant carries the T-DNA insertion in the last exon of the gene. Figure 6, panel B illustrates that no specific cDNA signal was detectable in the homozygous atwrky70 knockout line in comparison to wild type extracts from which both DNA and cDNA products were well amplified. To verify that the observed lower band of the correct predicted cDNA size of is indeed AtWRKY70, this fragment was gel isolated and subsequently sequenced. The sequencing data clearly confirmed that this fragment represents the full-length AtWRKY70 cDNA fragment and that the two known introns were spliced out.
The 'Touch-and-Go' extraction method was also tested in screening for an atwrky40, atwrky18 double mutant. Figure 7 shows a screening of F2 progeny homozygous for the atwrky18 T-DNA insertion mutation. Whereas a DNA fragment of 644 bp specific for the AtWRKY18 locus was amplified from extracts prepared from wild type Col-0 and homozygous atwrky40 mutant plants, this was not the case when the assay was applied to extracts that were prepared from homozygous atwrky18 lines. By screening 72 segregating F2 progeny from an atwrky40 × atwrky18 cross, we identified 19 individuals to be homozygous for the atwrky18 mutation (as judged by the absence of the 644 bp PCR product). These results are in agreement with the expected Mendelian segregation ratio (i.e. 18/72; Figure 7)
Figure 7 Screening for homozygous atwrky18 T-DNA knockouts using the 'Touch-and-Go' method. Seventy-two F2 progeny obtained by crossing of homozygous atwrky40 and atwrky18 mutants, were screened with primers flanking the T-DNA insertion site in the AtWRKY18 gene allowing detection of the wild-type AtWRKY18 allele. Amplification of a 644 bp size fragment indicates that F2 progeny are either wild type or heterozygous for AtWRKY18. No amplification suggests that the progeny are homozygous for the atwrky18 mutant allele. Controls, including DNA from the homozygous parental lines (knockout; KO), were replicated twice in pairs, giving a total of four independent control reactions. Primer control, lanes that contain no DNA template.
The Sucrose Prep and the 'Touch-and-Go' methods can successfully be used in other plants
We isolated DNA from the dicotyledonous crop species tobacco (Nicotiana tabacum) and its close relative N. benthamiana, as well as from the monocotyledonous crop species barley (Hordeum vulgare) using Sucrose Prep and the 'Touch-and-Go' methods to demonstrate their applicability for plants other than A. thaliana. As illustrated in Figure 8 (upper two panels), both methods produced sufficient quality and quantity of DNA that can be amplified using primers for two tobacco genes NtCDPK2 (calcium dependent protein kinase2) and NtRBCS that yield 2 kb and 0.8 kb PCR amplified DNA products respectively. Amplification failed in only one out of eight reactions using the Sucrose Prep while none failed using the 'Touch-and-Go' method. A similar isolation and PCR amplification was performed with barley tissue using four barley specific primer pairs producing varying sizes of amplified DNA fragments (Figure 8; lower two panels and Table 2). The sizes and patterns of the observed amplified DNA fragments are identical to those observed with other conventional DNA isolation methods (personal communication T. Zhao, M. Böhmer, and G. Freymark, MPIZ Köln). All of the primer pairs produced the expected size fragments (1.7, 0.7, 0.6 and 0.4 kb) in PCR analysis using DNA isolated by the Sucrose Prep. Using the Touch-and-Go' method, three primer pairs successfully amplified the expected smaller size fragments (0.7, 0.6 and 0.4 kb) but failed to amplify the largest size fragment of 1.7 kb.
Figure 8 Sucrose Prep and the 'Touch-and-Go' methods works well in other plant species. DNA was isolated in duplicate using the Sucrose Prep or the 'Touch-and-Go' methods (indicated on the left) from leaves of one month old N. tabacum and N. benthamiana plants (upper two panels) or from leaves of 10 day-old barley plants (lower two panels). PCR was carried out for 40 cycles in a 50 μl reaction volume using primers for NtCDPK2 and NtRBCS in tobacco, and four different pairs of primers for barley. Only 10 μl of the final reaction was resolved by agarose (1%) gel electrophoresis for analysis.
Table 2 List of primers and T-DNA lines used
Species/gene T-DNA knockout Primer No Primer (5'-3')
AtSOC1 GGATCGAGTCAGCACCAAACC
CTTGAAGAACAAGGTAACCCA
AtAGC1-10 CGTTTCACTATCTCCTCCACAAG
GGTGCTTTCAGAATGTTTACTAACGT
AtWRKY22 SALK_098205 2003 AAGAAAGTGTGCCATGTAGCAG
2004 CCGGAGACGATGAATAAGTAGC
AtWRKY46 GABI-Kat 038C07 182 ATGGAGGAGGTTCTAGCGAGAGTC
712 AAACGTCTTTACCATCATCAAGC
AtWRKY53 SALK_034157 713 ACGAATTGGAACTAGGGAAAGAG
714 CCATCATCAATAGAGCCATTTTC
AtWRKY36 GABI-Kat 258B10 72 CCTGCCTACAAAGATCATCTAGTTTCG
136 ATGATCAAAGAGGAGACCGTTTC
AtWRKY70 GABI-Kat 752F08 ATGGATACTAATAAAGCAAAAAAGC
AGATAGATTCGAACATGAACTGAAG
AtWRKY18 GABI-Kat 328G03 CATGGGTTCATTTCAAATTTTCG
CGATCTGCTCATGTTGCTGATGATG
NtCDPK2 ATGGGCAACGCATGCGGCGG
GATGACTCTCAAAGCCATTTTC
NtRBCS CCTCTGCAGCAGTTGCCACC
CCTGTGGGTATGCCTTCTTC
Hordeum vulgare 28-12A ATACCTGCACAGCCACAAGTC
GCAACTTCGCCTCTACGTTC
Hordeum vulgare 29-19A ACATGTGAGCTTGCTGGTTG
TGGGGGATGGTTAATGGTAG
Hordeum vulgare 33-22A CCTGCCGATGTAATCTGGTT
GATCTTTGCCATGTCTGTTTCG
Hordeum vulgare 41-30A AACATGCAAGCACACGTCAT
CATGATTGCTGTGGCTGACT
Flow diagrams for the Sucrose Prep and the 'Touch-and-Go' methods for PCR and RT-PCR applications are shown in Figure 9.
Figure 9 Flow diagram of the Sucrose Prep and the 'Touch-and-Go' methods for PCR and RT-PCR applications.
Discussion
The Sucrose Prep is not unique in being a quick DNA isolation protocol. Kasajima et al. [18] have exploited the method of Edwards et al. [9] to develop a rapid method for marker and transgene detection, and have demonstrated amplification of fragments up to 1.4 kb in size. As Langridge et al. [19] and Petersen et al. [20], we have also observed that DNA can be extracted by grinding the plant tissue in pure water and transferring a sample aliquot of the extract to a PCR reaction or that DNA templates can be delivered by adding a small amount of tissue to a PCR reaction mixture (data not shown). Therefore, we tried using extremely small amounts of tissue sampled with pipette tips by puncturing leaf tissue and immediately touching the tips into prepared PCR reaction mixtures. This 'Touch-and-Go' method is comparable in PCR amplification efficiency of DNA fragments up to about 1 kb with the Sucrose Prep and other rapid nucleic acid isolation methods. The stability of the Sucrose Prep in providing templates for numerous PCR or RT-PCR reactions however encouraged us to continue optimization of the method in connection with various high-throughput applications.
Alkaline lysis with NaOH [21-23] has also been successfully used in rapid isolation of DNA, however PCR amplification of DNA fragments are typically smaller than 2 kb in size. A common step between the majority of rapid procedures and the Sucrose Prep is the inclusion of a 'boiling' step. Burr et al. [24] used a thermal cycling protocol from 65°C to 96°C for a total time of 11.5 min, whereas Thomas and Henry [11] had optimized their protocol for DNA extraction from dried tissue by heating for 10 min at 95°C. Many quick DNA preparation methods dilute out contaminants from the harvested tissue, which interfere with the PCR reaction, by raising the extraction volume [22-26]. Sucrose Prep also employs such a dilution step by using about 50 μl extraction buffer for 2.5 mg tissue. Another component of the Sucrose Solution is the use of high salt, which is also employed in the protocol of Wang et al. [22] and is a principal component of DNA extraction buffers described by Edwards et al. [9] and Walbot and Warren [10].
The Sucrose Prep has been optimized for A. thaliana tissue, but is also suitable for DNA isolation from other species, including tobacco and barley (Figure 8). Therefore, Sucrose Prep should also work for species such as maize, wheat, rice, potato, tomato and other plants that have low to moderate concentrations of phenolics and starch
The combination of bulk segregant analysis with our new polymorphic DNA marker set proved very effective in rapidly locating a mutation of interest within a 10 cM interval. Tracking F2 segregation with the Sucrose Prep dramatically eased the analysis, as we employed a DNA shaker and thermal cycling blocks for heating, making future optimisation with robots possible. DNA sequencing of a candidate gene identified a mutation in the first exon of the PHYB gene, causing a premature stop codon at amino acid 554. The resulting truncated protein is predicted not to contain the PAS domain, which are important in phytochrome function and mediate interaction with putative signalling partners [27-29].
The Sucrose Prep also proved useful in screening for homozygous T-DNA insertion mutants and it was particularly useful for rapid identification of double homozygous knock-out lines that carried T-DNA insertions bearing identical selectable markers (Figure 3 and Figure 5). Occasionally, certain PCR primers did behave differently between conventional DNA preparations and templates obtained by the Sucrose Prep (Figure 1 and Figure 4). Nonetheless, in most cases where a primer combination did not work with Sucrose Prep, they also failed to produce PCR amplification with conventional methods under the aforementioned size limits (data not shown).
Pre-screening of segregating F2 or T2 EMS- and T-DNA-induced mutation populations with the Sucrose Prep greatly reduced the number of lines requiring further characterization. Upon fast screening with Sucrose Prep, detailed analysis always led to the identification of homozygous mutant lines that were confirmed by other DNA and RNA isolation methods.
RNA isolation from plants is often a lengthy process requiring toxic chemicals or expensive kits, and requires a very clean practice due to vast contamination of RNAses. A quick RNA isolation method is therefore highly desirable. We have demonstrated above that plant extracts prepared by the Sucrose Prep are also suitable for RT-PCR assays (Figures 5 and 6). Due to high concentration of DNA in the extracts, the primers must be designed in exons separated by introns in order to distinguish DNA from cDNA (Figure 6). DNA contamination is not a unique problem to our approach but is common to several other RNA isolation methods. The only major disadvantage of our quick RNA isolation method is that the RNA to DNA ratio is very low as compared to other methods where RNA is concentrated through several steps. However, our method is very useful and even superior to other methods in certain applications requiring speed and use of limited amounts of plant tissue. Thus, the method is particularly useful if cells expressing the gene of interest are restricted to certain tissue, such as hydathodes, major and minor veins, emerging young leaves, flower organs such as nectaries, flower abscission zones, sepals petals, anthers, gynoecium, root tips, as well as local pathogen infected tissues and islands of cells generated by transposon jumping. One of the limitations of our method is that it is not suitable for genes that are expressed at very low levels. The 'Touch-and-Go' method is not suitable for quantitative RT-PCR applications, due to the varying amount of RNA isolated during sampling.
Using the Sucrose Prep and the 'Touch-and-Go' methods, we have identified homozygous T-DNA knockouts for the AtWRKY36 and AtWRKY70 genes (Figure 6). With respect to PCR amplification of DNA samples, we found the 'Touch-and-Go' method extremely useful and time saving, especially when screening for the presence or absence of PCR products of less than 500 bp in size. We had however variable success rates if the size of the expected PCR products was larger than 1 kb. As with the Sucrose Prep, the 'Touch-and-Go' method worked very well for tobacco in amplifying DNA fragments up to 2 kb, however, in barley it failed to amplify the largest predicted fragment of 1.7 kb and the amount of amplification product varied for fragments less than 1 kb in length. This difference observed in the reproducibility of DNA isolation between tobacco and barley is possibly due to the lower number of cells that are disrupted in barley plants upon puncturing with the pipette tip. Nevertheless, the simplicity, speed and reproducibility of the 'Touch-and-Go' approach and the robustness and relative speed of the Sucrose Prep method makes them ideal for high-throughput PCR based screens in alternative transgenic approaches replacing the use antibiotic resistance selectable markers [31].
Conclusion
In comparison to other rapid nucleic acid isolation protocols described for plant samples, the Sucrose Prep is thus far the only extraction protocol, which is shown to be compatible with simultaneous isolation of DNA and RNA templates. This minimal-step nucleic acid isolation method can be combined with the use of our high resolution marker set that can be resolved on agarose gel after amplification with PCR to perform fast and precise mapping of mutations using DNA polymorphisms between A. thaliana ecotypes Col-0 and Ler. We anticipate that the utilization of Sucrose Prep as well the 'Touch-and-Go' method will facilitate the improvement of automated high-throughput genomic techniques used in functional genomics studies of the model plant A. thaliana, as well as in other plants species, including important crops.
Methods
The Sucrose Prep
Sucrose Solution: 50 mM Tris-HCl pH 7.5, 300 mM NaCl and 300 mM sucrose.
(A) Individual samples: approximately 10 mg of leaf tissue was placed directly in 200 μl Sucrose Solution and ground at room temperature or on ice using a pipette tip or pestle. The samples were then heated to 99–100°C for 10 min and then briefly spun at 2000–6000 g for 5 sec. The samples were placed on ice until PCR. One μl of the supernatant was used for PCR, avoiding debris.
(B) Multiple sample 96-well format: between 10–20 mg leaf tissue from 14 d-old plants was harvested as leaf discs into 96 well plates. Metal balls (3 mm, tungsten carbide beads) were added and shaken in a Retsch MM300 shaker for 10 sec, then 300 μl of Sucrose Solution was added, the plate heated at 99°C for 10 min and placed on ice until use. Following storage at 4°C or -20°C, samples were reheated at 99–100°C for 10 min and then immediately placed on ice.
The 'Touch-and-Go' method
Leaves were punctured against a firm surface, like a finger covered with a latex glove using a fine pipette tip (TipOne from Starlab GMBH, catalog no. S1111-3000 or S1110-3000). The DNA/RNA on the tip of the pipette was transferred to the pre-prepared PCR solution in the PCR tubes by touching the tip of the pipette to the solution and pipetting up and down a few times. For plants in the field or greenhouse, 10 μl water was aliquoted into PCR tubes or microtiter plates. The tubes/plates were kept on ice while puncturing the leaves with a fine pipette tip against a firm surface and DNA/RNA the tip was transferred into water by pipeting up and down. After returning the samples to the laboratory, 40 μl PCR master mix was added to each well. Thermocycling with the 'Touch-and-Go' method requires 40 cycles of PCR amplification.
CTAB method
The CTAB protocol used was developed by Murray and Thompson [32], modified from Rogers and Bendich [33] and adapted by Rios et al. [2].
Plant Growth Conditions
A. thaliana plants were grown under a 16 h photoperiod at 20°C in a greenhouse.
EMS Mutagenesis
Approximately 6,500 35S::CO 35S::FLC 1 kb:SOC1::GUS co-2 seeds were mutagenized by imbibition in 0.3% ethyl methanesulfonate (EMS; Sigma) for 8 to 9 h, followed by washing with 0.1 M Na2SO4 and distilled water. The M1 mutagenized seed was planted into about 260 pools each containing 25 M1 seeds. About 300 M2 seed from each pool were sown under long day conditions and scored for flowering time.
PCR and RT-PCR
Routine PCR: 3 min 94°C, 35–40 cycles of: (30 sec 94°C, 45 sec 55°C, 1 min 72°C), 10 min 72°C, 4°C until analysis. 2.5 μM each gene specific primers, 2.5 mM dNTPs, 5–10 U (0.5–1 μl) Taq polymerase (Invitrogen), 1 × Taq Buffer (commercially supplied). For products larger than 2 kb, 0.5 U of enzyme LA Taq polymerase (TaKaRa) was substituted for Invitrogen Taq polymerase and the PCR protocol from Rios et al. [2] was used.
Qiagen OneStep RT-PCR Kit (catalog no. 210210) was used following the manufacturer's recommendations. For isolation of cDNA, RNA was extracted with RNeasy Plant Mini Kit from Qiagen (catalog no. 74904) following the manufacturer's instructions.
Primers and T-DNA knockout lines are listed in Table 2.
Gel documentation
All of the agarose gel pictures are ethidium bromide stained gels and the images are inverted in Adobe Photoshop.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KB developed the Sucrose Prep, initiated collaboration with IS to use the Sucrose Prep for mapping purposes and together with BU was involved in the design, coordination, and drafting of the manuscript. IS designed the novel marker set between Col-0 and Ler, IS and DR mapped the mutation in the PHYB gene and assisted in drafting the manuscript. BU developed the 'Touch-and-Go' method and demonstrated that the nucleic acids isolated by Sucrose Prep or the 'Touch-and-Go' methods are suitable for RT-PCR assays, tested all of these methods in screening T-DNA knockouts in Arabidopsis and demonstrated that these methods are also suitable for tobacco and barley. CK, GC and IES provided the laboratory facilities, gave valuable experimental advises and extensively helped in drafting the manuscript. AB and CK are the Ph.D supervisors of KB. All authors were involved in reading, correcting and approving the final version of the manuscript.
Acknowledgements
We would like to acknowledge the skilful assistance of Nicole Kamphaus, Sandra Kröber and Anja Reinstädler and thanks to Marcel Lafos for critical reading of the manuscript. Many thanks to Tiehan Zhao, Maik Böhmer, Gerald Freymark for sharing their primers for barley, tobacco and N. benthamiana respectively.
==== Refs
Zhang J Stewart J Economical and rapid method for extracting cotton genomic DNA The Journal of Cotton Science 2000 4 193 201
Ríos G Lossow A Hertel B Breuer F Schaefer S Broich M Kleinow T Jásik J Winter J Ferrando A Farrás R Panicot M Henriques R Mariaux J Oberschall A Molnár G Berendzen K Shukla V Lafos M Koncz Z Rédei G Schell J Koncz C Rapid identification of Arabidopsis insertion mutants by non-radioactive detection of T-DNA tagged genes Plant J 2002 32 243 253 12383089 10.1046/j.1365-313X.2002.01416.x
Krysan PJ Young JC Sussman MR T-DNA as an insertional mutagen in Arabidopsis Plant Cell 1999 11 2283 2290 10590158 10.1105/tpc.11.12.2283
Samson F Brunaud V Balzergue S Dubreucq B Lepiniec L Pelletier G Caboche M Lecharny A FLAGdb/FST: a database of mapped flanking insertion sites (FSTs) of Arabidopsis thaliana T-DNA transformants Nucleic Acids Res 2002 30 94 97 11752264 10.1093/nar/30.1.94
Alonso JM Stepanova AN Leisse TJ Kim CJ Chen H Shinn P Stevenson DK Zimmerman J Barajas P Cheuk R Genome-wide insertional mutagenesis of Arabidopsis thaliana Science 2003 301 653 657 12893945 10.1126/science.1086391
Rosso MG Li Y Strizhov N Reiss B Dekker K Weisshaar B An Arabidopsis thaliana T-DNA mutagenized population (GABI-Kat) for flanking sequence tag-based reverse genetics Plant Mol Biol 2003 53 247 259 14756321 10.1023/B:PLAN.0000009297.37235.4a
Konieczny A Ausubel FM A procedure for mapping Arabidopsis mutations using co-dominant ecotype-specific PCR-based markers Plant J 1993 4 403 410 8106085 10.1046/j.1365-313X.1993.04020403.x
Torjek O Berger D Meyer R Mussig C Schmid K Rosleff Sorensen T Weisshaar B Mitchell-Olds T Altmann T Establishment of a high-efficiency SNP-based framework marker set for Arabidopsis Plant J 2003 36 122 140 12974817 10.1046/j.1365-313X.2003.01861.x
Edwards K Johnstone C Thompson C A simple and rapid method for the preparation of plant genomic DNA for PCR analysis Nucleic Acids Res 1991 19 1349 2030957
Walbot V Warren C Regulation of Mu element copy number in maize lines with an active or inactive Mutator transposable element system Mol Gen Gen 1988 211 27 34 10.1007/BF00338389
Thomson D Henry R Single-Step protocol for preparation of plant tissue for analysis by PCR BioTechniques 1995 19 394 400 7495552
Arabidopsis Genome Initiative Analysis of the genome sequence of the flowering plant Arabidopsis thaliana Nature 2000 408 796 813 11130711 10.1038/35048692
Jander G Norris SR Rounsley SD Bush DF Levin IM Last RL Arabidopsis map-based cloning in the post-genome era Plant Physiol 2002 129 440 450 12068090 10.1104/pp.003533
Hepworth S Valverde F Ravenscroft D Mouradov A Coupland G Antagonistic regulation of flowering-time gene SOC1 by CONSTANS and FLC via separate promoter motifs EMBO J 2002 21 4327 37 12169635 10.1093/emboj/cdf432
Reed J Nagpal P Poole D Furuya M Chory J Mutations in the gene for the red/far-red light receptor phytochrome B alter cell elongation and physiological responses throughout Arabidopsis development Plant Cell 1993 5 147 157 8453299 10.1105/tpc.5.2.147
Valverde F Mouradov A Soppe W Ravenscroft D Samach A Coupland G Photoreceptor regulation of CONSTANS protein in photoperiodic flowering Science 2004 303 1003 1006 14963328 10.1126/science.1091761
Eulgem T Rushton PJ Robatzek S Somssich IE The WRKY superfamily of plant transcription factors Trends Plant Sci 2000 5 199 206 10785665 10.1016/S1360-1385(00)01600-9
Kasajima I Ide Y Ohkama-Ohtsu N Hayashi H Yoneyama T Fujiwara T A protocol for rapid DNA extraction from Arabidopsis thaliana for PCR analysis Plant Mol Bio Reporter 2004 22 49 52
Langridge Ursula Schwall M Langridge P Squashes of plant tissue as subtrate for PCR Nucleic Acids Res 1991 19 6954 1762926
Petersen G Johansen B Seberg O PCR and sequencing from a single pollen grain Plant Mol Biol 1996 31 189 191 8704154 10.1007/BF00020620
Klimyuk V Carroll B Thomas C Jones J Alkali treatment for rapid preparation of plant material for reliable PCR analysis Plant J 1993 3 493 494 8220456 10.1046/j.1365-313X.1993.t01-26-00999.x
Wang H Qi M Cutler A A simple method of preparing plant samples for PCR Nucleic Acids Res 1993 21 4153 4154 8371994
Paris M Carter M Cereal DNA: A rapid high-throughput extraction method for marker assisted selection Plant Mol Bio Reporter 2000 18 357 360
Burr K Harper R Linacre A One-Step isolation of plant DNA suitable for PCR amplification Plant Mol Bio Reporter 2001 19 367 371
Dilworth E Frey J A rapid method for high throughput DNA extraction from plant material for PCR amplification Plant Mol Bio Reporter 2000 18 61 64
Kreader C Weber S Song K One-Tube preparation and PCE amplification of DNA from plant lead tissue with Extract-N-Amp™; Molecular Biology Application Notes Life Science Quartely 2001 2
Furuya M Phytochromes: their molecular species, gene family and functions Annu Rev Plant Physiol Plant Mol Biol 1993 44 617 645 10.1146/annurev.pp.44.060193.003153
Quail PH Boylan MT Parks BM Short TM Xu X Wagner TM Phytochromes: photosensory perception and signal transduction Science 1995 268 675 680 7732376
Krall L Reed JW The histidine kinase-related domain participates in phytochrome B function but is dispensable Proc Natl Acad Sci U S A 2000 97 8169 8174 10869441 10.1073/pnas.140520097
Mockler TC Guo H Yang H Duong H Lin C Antagonistic actions of Arabidopsis cryptochromes and phytochrome B in the regulation of floral induction Development 1999 126 2073 2082 10207133
Francis KE Spiker S Identification of Arabidopsis thaliana transformants without selection reveals a high occurrence of silenced T-DNA integrations Plant J 2005 41 464 477 15659104
Murray M Thompson W Rapid isolation of high-molecular-weight plant DNA Nucleic Acids Res 1980 8 4321 4325 7433111
Rogers S Bendich A Extraction of DNA from milligrams of fresh, herbarium and mummified plant tissues Plant Mol Biol 1985 5 69 76 10.1007/BF00020088
|
16270938
|
PMC1277017
|
CC BY
|
2021-01-04 16:37:28
|
no
|
Plant Methods. 2005 Aug 23; 1:4
|
utf-8
|
Plant Methods
| 2,005 |
10.1186/1746-4811-1-4
|
oa_comm
|
==== Front
Plant MethodsPlant Methods1746-4811BioMed Central London 1746-4811-1-51627093410.1186/1746-4811-1-5ReviewReview of methodologies and a protocol for the Agrobacterium-mediated transformation of wheat Jones Huw D [email protected] Angela [email protected] Huixia [email protected] CPI Division, Rothamsted Research, Harpenden, AL5 2JQ, UK2005 5 9 2005 1 5 5 8 7 2005 5 9 2005 Copyright © 2005 Jones et al; licensee BioMed Central Ltd.2005Jones 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.
Since the first report of wheat transformation by Agrobacterium tumefaciens in 1997, various factors that influence T-DNA delivery and regeneration in tissue culture have been further investigated and modified. This paper reviews the current methodology literature describing Agrobacterium transformation of wheat and provides a complete protocol that we have developed and used to produce over one hundred transgenic lines in both spring and winter wheat varieties.
==== Body
Introduction
Transformation of cereal crops is a powerful research tool for gene discovery and function to investigate genetically controlled traits and is fast becoming a key element in the process of varietal improvement. It provides key underpinning knowledge to inform and short-cut conventional breeding strategies. For specific crops, it also enables the introduction of novel genes directly into locally-adapted germplasm and the creation of new genetically modified varieties. As testament to this, a total of 81 million Ha of approved GM crops, mainly for herbicide tolerance or insect resistance, were planted in 2004 [1], although wheat does not currently form part of this portfolio.
Wheat was among the last of the major crops to be transformed with the first fertile transgenic plants being reported using particle bombardment little over a decade ago [2-6]. Advances in the design of micro-projectile devices, choice of explant, media composition and selection systems has enabled the application of wheat transformation to study the role specific genes in a wide range of agronomically important traits (reviewed by [7-9]). Particle bombardment remains a robust, relatively efficient method for the genetic manipulation of wheat [10], however at the molecular level, the DNA integration sites are often unnecessarily complex. There are several significant advantages to transferring DNA via Agrobacterium, including a reduction in transgene copy number, the stable integration with fewer rearrangements of long molecules of DNA with defined ends and the ability to generate lines free from selectable marker genes [7,11-14]. This has been a driving force in the development of methods using Agrobacterium tumefaciens to deliver DNA although the ability to routinely transform wheat in this way is currently restricted to a few, well-resourced public and commercial laboratories worldwide. This is partly due to the need for experienced personnel and expensive laboratory and plant growth infrastructure but also through a lack of clearly-written, complete, publicly-available protocols. There are several research papers and patents describing specific improvements to methodologies but these fail to provide a step-by-step guide to the transformation process as a whole.
We have compared the published literature under headings that describe the main variables in the transformation process. First, we consider the relatively narrow range of wheat genotypes that have been successfully transformed, the choice of explant and the pre-treatments that were carried out. Second, we compare the Agrobacterium strains, resident Ti plasmids and binary vectors used and consider the importance of additional virulence genes. The various inoculation and co-cultivation conditions are discussed and finally the key steps to control the overgrowth of Agrobacterium cells and the selection of regenerating transformed plants are described. We then provide a detailed protocol for the transformation of freshly isolated immature embryos and regeneration of fertile plants in 9–12 weeks.
Genotype and explant pre-treatments
Immature embryos of Bobwhite, pre-cultured for between 1 and 6 days on CM4C medium, are the most commonly used explant [15-18], although the use of 9 day pre-cultured immature embryos of cv. Fielder [19] and callus derived from immature embryos of Bobwhite [17] and cv. Veery 5 [20] has also been reported (see Table 1 for summary). Although immature embryos of Bobwhite are commonly pre-cultured prior to inoculation, Cheng et al. [17] report no significant difference in transformation efficiencies between immature embryos, pre-cultured ones or embryogenic callus. In an alternative approach, freshly isolated immature embryos of the winter and spring wheat cultivars Florida and Cadenza were found preferable to pre-cultured ones [21] and it is this explant type that is described in the accompanying protocol as it has potential to be applied to other varieties. Precocious zygotic germination is a significant problem when using immature embryo explants but can be suppressed by the addition of hormones such as dicamba, abscisic acid or high levels of 2,4-D to the culture medium. Some authors specifically state that the embryo axis was removed or damaged to prevent zygotic germination [19-21]. A marked effect of embryo size/age on T-DNA delivery and regeneration has been demonstrated, with large embryos (>2 mm) giving significantly higher transient expression levels but lower regeneration frequencies [21] than smaller ones (<1.5 mm). We emphasise the need to use embryos of 0.8–1.5 mm in the accompanying protocol.
Table 1 Summary of main parameters reported for Agrobacterium-mediated transformation of wheat.
Wheat variety (S – spring) (W – winter) Explant type Embryo Axis removed Agrobacterium strain (binary vector) Inoculation (Co-culture) *rt – room temp Control of Agrobacterium cells Plant selective agent Transformation Freq. (%) No of plants reported Refs
Bobwhite (S) IE (age NS*); 1–6 d PCIE; 10–25 d EC NS* C58-ABI (pMON18365) 3 h, 23–25°C (2–3 d, 24–26°C) Carbenicillin (250 mg/l) G418 1.4–4.3 >100 [17]
Bobwhite (S) 4 d PCIE NS* C58-ABI (pMON30139 and others) 15–30 min, 23–25°C (2–3 d, 23–25°C) Carbenicillin (250–500 mg/l) Glyphosate 4.4 3354 [16]
Bobwhite (S) 1–6 d PCIE; 8–30 d EC NS* C58-ABI (pMON18365) 5–60 min, 23–26°C (2–3 d, 24–26°C) Carbenicillin (250 mg/l) G418
Paromomycin Glyphosate 4.8–19 154 [18]
Bobwhite (S) 3–6 PCIE NS* C58C1 (pPTN155) 45 min – 3 h, 25°C (1–3 d, 25°C) Ticarcillium; Vancomycin Cefatoxin; (50 mg/l) G418 0.5–1.5 13 [15]
Cadenza (S) Florida (W) 0–72 h IE Yes AGL1 (pAL154/156) 15 min-5 h, rt* (1–5 d, 24–25°C rt*) Timentin (160 mg/l) PPT (L-Phosphinothricin) 0.3–3.3 44 [21]
Fielder (S) 6–9 d PCIE Yes AGL0 (pBGX1) 30–60 min rt* (2–3 d, 23–24°C) Timentin (150 mg/l) GFP, Bialaphos 1.8 4 [19]
Veery-5 (S) 14 d EC Yes LBA4404 (pHK21) 15 min at rt* (1 d 27°C, 2 d 22°C) Timentin (150 mg/l) Glufosinate ammonium 1.2–3.9 17 [20]
Vesna (S) IE (age NS*) NS* LBA4404 (pTOK233) AGL1 (pDM805) 15–30 min, (3 d, 27°C) Cefotaxime (300 mg/l) PPT (L-Phosphinothricin) 0.13–0.41 6 [45]
Various Chinese varieties (NS*) EC (age NS*) NS* AGL1 (pUNN-2) 30–60 min (2 d, 28°C) Timentin (150 mg/l) Paromomycin 3.7–5.9 44 [46]
IE – freshly isolated immature embryos; PCIE – pre-cultured immature embryos; EC – embryogenic callus; *NS not specified.
Various explant pre-treatment steps have been evaluated in attempts to improve T-DNA delivery or tissue-culture response in particular varieties. Osmotic and desiccation treatments have been evaluated and incorporated into protocols based on particle bombardment [22-26], and have also been tested for Agrobacterium transformation of wheat. Air-drying pre-cultured immature embryos and embryogenic callus explants during Agrobacterium co-cultivation increased T-DNA-delivery and suppressed Agrobacterium cell growth which in turn facilitated better plant cell recovery [18]. The same authors found no such advantage when explants were desiccated prior to inoculation or when osmotic conditioning was used, however other reports indicate a beneficial effect on transformation of air-drying prior to co-culture for rice suspension cell cultures [27] and sugarcane callus [28]. Osmotic conditioning on 10% sucrose prior to Agrobacterium inoculation caused a marked increase of GUS transient expression in pre-cultured rice calli [29] but a plasmolysis step using 20% maltose failed to improve T-DNA delivery in 10 day pre-cultured wheat embryos [30].
Agrobacterium tumefaciens strains and binary vectors
The ability of particular Agrobacterium strains to transform plant cells is defined by their chromosomal and plasmid genomes which between them must encode all the machinery necessary for attachment and DNA-transfer. The Agrobacterium strains that have been successfully used for wheat transformation are based on only two chromosomal backgrounds, LBA4404 (Ach5) and C58 but these have been used with a wide range of Ti and binary plasmids. Some strains, notably AGL0 and AGL1 have been engineered to contain the so-called hypervirulent Ti plasmid, pTiBo542 harbouring additional vir genes originating from the Agrobacterium strain A281 which in its oncogenic form possesses a broad host range and a induces large, rapidly appearing tumours [31]. The strains used in the papers reviewed (see Table 2), also contain a binary and sometimes helper plasmids, often conferring yet more copies of virulence genes. A comparison of different Agrobacterium strains demonstrated that AGL0, a hypervirulent strain containing a disarmed pTiBo542 plasmid [32], was better at generating wheat transformants than other strains tested [19]. The ability of the Ti plasmid pTiBo542 to confer higher transformation efficiencies was first observed in dicots [33-35] and the vir genes from this plasmid have been widely adopted for monocot transformation vectors (reviewed by [11]). The weakly virulent Agrobacterium strain LBA4404, was successful in transforming wheat only when augmented by the superbinary plasmid pHK21 which possessed extra copies of vir B, C and G genes from pTiBo542 but not when carrying a standard binary plasmid [20]. Further evidence of the positive effect of additional vir genes was provided by the demonstration that a 15 Kb fragment of pTiBo542 on a pSOUP helper plasmid [36] enhanced T-DNA delivery and the production of transgenic wheat plants, even when in a hypervirulent AGL1 background already containing pTiBo542 as a resident Ti plasmid [21,37]. Although there has been a tendency to incorporate additional vir genes, particularly virG, into binary vectors this is not always necessary, at least for cv Bobwhite, in which a large number of transgenic lines have been reported using apparently standard Agrobacterium strains and binary vectors [16-18]. There is also one report [15] of transformation with a normal binary in the Agrobacterium strain C58C1 which the authors describe as disarmed, however it is our understanding that the C58C1 strain is actually cured of its pTiC58 plasmid [38,39]. There is currently insufficient data to define precisely which vir genes are necessary and where they should reside for optimal wheat transformation in different genotypes. There is also scope for further research into the effect on wheat transformation of specific Agrobacterium mutants that have shown beneficial effects for other plant species. For example, strains containing mutations in the vir gene regulator virG resulting in constitutive expression of this gene and presumably the other vir genes it regulates, gave significant increases in efficiency of transformation in tobacco and cotton [40], Catharanthus roseus [41] and Norway spruce [42]. This virG mutant was also combined with a high copy number plasmid to further improve transformation rates in rice and soybean [43].
Table 2 Summary of Agrobacterium strains and vectors used to investigate wheat transformation.
Agrobacterium strain (binary vector) Chromosomal background Ti plasmid Opine classification Additional vir genes on binary or helper plasmids Binary type Selectable and scorable marker on T-DNA. (Promoter shown in parentheses)
ABI (pMON18365) [17, 18] C58 Disarmed pTiC58 Nopaline pMON18365, none reported normal-binary nptII (E35S) GUS (E35S)
C58C1 (pPTN155) [15] C58 Cured/disarmed? Nopaline pPTN155, none reported normal-binary nptII (35S) GUS (E35S)
AGL1 (pAL154/156) [21] C58, RecA pTiBo542 ΔT-DNA Succinamopine pAL154, 15.2 Kb fragment from pTiBo542 [47], pAL156, none super-binary bar (Ubi1) GUS (Ubi1)
AGL0 (pBGX1 and pTO134) [19] C58 pTiBo542 ΔT-DNA Succinamopine pBGX1, none reported pTO134, none reported normal-binary hpt (35S) gfp (Ubi1)
bar (35S) sgfpS65T (35S)
AGL1 (pDM805) C58, RecA pTiBo542 ΔT Succinamopine pDM805, none reported normal-binary bar (Ubi1) GUS (Act1)
LBA4404 (pTOK233) [45] Ach5 DNA Disarmed pAL4404 Octopine pTOK233, extra copy of virB, virC and virG from pTiBo542 47, [48] super-binary hpt (35S) GUS (35S)
LBA4404 (pHK21) [20] RecA Ach5 Disarmed pAL4404 Octopine pHK21, extra copy of virB, virC and virG from pTiBo542 [47] super-binary bar (Ubi1) GUS (Ubi1)
AGL1 (pUNN-2) [46] C58, RecA pTiBo542 ΔT-DNA Succinamopine pUNN-2, none reported normal-binary nptII (Ubi1))
ABI (pMON30139 and others) [16] C58 Disarmed pTiC58 Nopaline pMON30139, none reported normal-binary aroA:CP4 (Act1) aroA:CP4 (e35S+ hsp intron)
Inoculation and co-cultivation
The Agrobacterium infection process is divided into two stages: first, a short period, typically a few minutes to a few hours (see Table 1), of inoculation by complete or partial immersion of explants in an Agrobacterium suspension. Then, after the majority of Agrobacterium cells are removed by pouring or pipetting, the explants are co-cultivated for a further 1–3 days. One or both these steps are carried out in darkness at approximately 25°C, although a two temperature co-cultivation step has also been tried with one day at 27°C then two days to 25°C [20]. During the co-cultivation period, phenolic inducers such as acetosyringone work alongside other signalling factors such as temperature and an acid environment to promote the expression of vir genes. The presence of 200 μM acetosyringone in the Agrobacterium or co-cultivation medium markedly increased T-DNA delivery [21]. Enhanced transient GFP expression was observed in wheat cell clusters with acetosyrigone at 400 μM in the co-cultivation but not the inoculation media [19]. The need for acetosyringone been reported for a variety of wheat explants types [17,37,44] but not for wheat cell suspension cultures where exogenous induction agents were not necessary for stable transformation [17].
The use of surfactants during inoculation and co-cultivation significantly increases T-DNA delivery. Increasing concentrations of Silwet L-77 up to 0.04% had positive effects on T-DNA delivery as measured by the number of immature embryos with GUS foci and the number of GUS foci per embryo [21]. However, concentrations higher than 0.05% reduced survival and callus formation in freshly isolated immature embryos and an optimal concentration of 0.01% was chosen [21]. Positive effects of surfactants were also reported in study [17] which used Silwet and pluronic acid F68 at 0.02%. Silwet has been used at concentrations as high as 0.05% for pre-cultured embryos and calli [15]. The protocol presented here uses Silwet L-77 at 0.015% but no pre-culture or special inoculation treatments.
Control of Agrobacterium, regeneration and selection
After the co-cultivation period, infected explants progress in a series of tissue culture steps on media designed to inhibit the growth of Agrobacterium cells and promote regeneration and selection of transformants. The antibiotics used to control the growth of Agrobacterium are added immediately after co-cultivation during the callus induction phase and are maintained in all subsequent media. Timentin or carbenicillin are commonly used but other compounds such as cefatoxin, cefotaxime, ticarcillium and vancomycin have also been reported (see Table 1). Plant selection agents complementary to the marker gene on the T-DNA are introduced to kill or compromise the growth of untransformed material. Selection for plant transformation is often initiated a few days after co-cultivation during the callus-induction phase and maintained during the latter regeneration steps. Delayed selection, started at the later plant regeneration phase was preferred by [21] and is the method described in the accompanying protocol. Three selectable marker gene systems have been reported for Agrobacterium transformation of wheat. The first is based on antibiotic selection using either hpt (aph4-Ib) or nptII (aph3'II) which encode phosphotransferase enzymes that confer tolerance to the aminoglycoside antibiotics such as kanamycin, neomycin, paromomycin, G418 and hygromycin. A second system utilizes the bar gene which confers tolerance to glufosinate ammonium-based herbicides such as PPT, Basta, Bialaphos etc. A third system is based on the aroA:CP4 genes conferring tolerance to glyphosate-based herbicides such as Roundup. The use of 0.02 mM glyphosate on regenerating meristems has been reported to reduce the number of plants escaping selection to zero [16]. NptII, bar and aroA:CP4 have been successfully used by different groups to produce transgenic wheat plants but it is not possible to draw direct comparisons between selection systems because often a visual marker was also used in conjunction with chemical selection. For example, in several studies, the GUS reporter gene has been used in addition to the conventional selectable marker to help optimise the identification of transformants [15,17,18,20,21,45]. Also, a T-DNA containing both hpt and GFP, along with hygromycin selection, has been used to identify early events in the transformation process [19].
In wheat transformation via Agrobacterium, the total length of time, from isolation of the original explant to the transfer of young plants to soil, is typically 12–16 weeks depending on the length of pre-culture and the number of selection steps. A shortened protocol taking only 7–11 weeks, achieved mainly by reducing the selection step to one week, has also been reported [16]. The protocol described in the present paper was optimised for bar/glyphosate selection with a GUS assay to confirm T-DNA integration and expression and takes approximately 12 weeks.
Concluding remarks
The advantages arising from simple molecular integrations of single copy DNA fragments with defined ends have driven research into Agrobacterium-mediated plant transformation. Compared to rice and maize, progress with wheat has been slower but as described here, robust methods for the transformation of wheat using Agrobacterium now exist. There is scope to further optimise the media components and pH and to investigate the ideal virulence gene complement. Current bottlenecks limiting throughput include the labour-intensive steps of embryo isolation and transfers between media. Unlike biolistics, Agrobacterium suspensions can be manipulated by liquid handing robots and this combined with the use of callus cultures and the automation of transfer steps would enable a higher throughput which even at low efficiency would allow significantly more transgenic lines to be produced
A protocol for wheat (Triticum aestivum L.) transformation mediated by Agrobacterium tumefaciens
Scope and limitations
This method was developed for the winter wheat cultivar Florida but with minor modifications has also been used to successfully transform the spring wheat varieties Fielder and Cadenza. It utilises the super-virulent Agrobacterium tumefaciens strain AGL1 [32] containing the plasmids pAL154/pAL156 which are based on the plasmid pSoup/pGreen [36], . The binary vector pAL156 contains a single T-DNA incorporating the bar gene conferring Basta resistance and a modified uidA (GUS) gene with an intron within the open reading frame to prevent its expression in Agrobacteium itself. Both the bar and uidA genes are driven by the maize ubiquitin1 promoter plus ubiquitin1 intron [49]. The bar gene is located next to the left border, and uidA is adjacent to the right border. A helper plasmid pAL154 provides replication functions for pAL156 in trans and also contains the 15 kb Komari fragment [35,47] supplying extra vir genes. Other Agrobacteium strains and plasmid combinations may also be appropriate in our protocol but have not yet been tested.
There are three main steps in the method: 1. incubation of freshly-isolated immature embryos with Agrobacterium tumefaciens; 2. induction of embryogenic callus and regeneration of shoots and roots; 3. application of a herbicide selection system to allow only the transgenic plantlets to survive. The average efficiency of transformation (number of independent transgenic lines/total number of immature embryos inoculated) is approximately 1%. The protocol takes 9–12 weeks from the isolation of immature embryos to the potting of putative transgenic plantlets to soil (Figure 1).
Figure 1 Main steps in the protocol for Agrobacterium transformation of wheat from inoculation to the transfer of transgenic wheat plants to soil.
Protocol
Growth of donor plants
1.1 Sow seeds, 4–5 per 21 cm diameter pot, in compost which contains 75% fine-grade peat, 12% screened sterilised loam, 10% 6 mm screened lime-free grit, 3% medium vermiculite, 2 kg Osmocote Plus/m3 (slow-release fertiliser, 15N/11P/13K plus micronutrients), 0.5 kg PG mix/m3 (14N/16P/18K granular fertiliser plus micronutrients (Petersfield Products, Leicestershire, UK). Although other soil formulations may also be suitable.
1.2 Grow wheat plants in environmentally controlled growth rooms for approximately 11 weeks to provide immature seeds.
1.3 Growth rooms are maintained at 18–20°C day and 14–15°C night temperatures with a relative air humidity of 50–70% under a 16 h photo-period provided by banks of 400 W High Temperature Quartz Iodine lamps (Osram Ltd., Berkshire, UK) which give light intensity ~700 μmolm-2s-1 photosynthetically active radiation (PAR).
1.4 Before transferring to these conditions, winter wheat varieties are vernalised from seed for 8 weeks at 4–5°C with a 12 hour photoperiod provided by 70 W fluorescent lamps giving approximately 150 μmolm-2s-1 PAR at 300 mm from the lights.
1.5 The water is supplied by an automated flooding system, but seedling-stage plants are initially top watered individually for a few weeks [50].
2 Growth and preparation of Agrobacterium cells for inoculation
2.1 Initiate Agrobacterium liquid cultures by adding ~200 μl of a standard glycerol inoculum to 10 ml MG/L [51] (Table 3) plus antibiotics. Prepare as many 10 ml cultures as plates to be treated.
Table 3 Composition of medium MG/L
Component /litre
Mannitol 5 g
L-Glutamic acid 1 g
KH2PO4 250 mg
NaCl 100 mg
MgSO4·7H2O 100 mg
Tryptone 5 g
Yeast extract 2.5 g
pH 7.0
Biotin (added after autoclaving from stock at 1 mg/100 ml (add 100 μl to 1 litre MG/L) 1 μg
2.2 Incubate at 27–29°C, shaking (250 rpm) for 12–24 hours (to reach an OD >1 (Abs = 600 nm)).
2.3 Pellet the Agrobacterium culture at 4500 g for 10 minutes and resuspend in 4 ml single-strength inoculation medium (see 6.2.2) supplemented with 200 μM acetosyringone for each 10 ml culture.
2.4 Replace the cultures back on the shaker until required, but they should be used within 3 hours.
Note, The antibiotics used depend on the selectable markers in the Agrobacterium strain and binary vectors used. For the AGL1 strain used in this protocol, carbenicillin (200 mg/l) is used and pAL154/156 combinations are selected with kanamycin (100 mg/l) which is the selectable marker on pAL156.
3 Preparation of explants
3.1 Ear collection and surface sterilization
3.1.1 Collect ears at approximately 12–16 days post-anthesis, a few seeds can be opened at the time of collection to determine the size and texture of the embryos, which should be 0.8 – 1.5 mm in length and translucent in appearance.
3.1.2 Surface sterilise by rinsing in 70% (v/v) aqueous ethanol for 1 minute then 15 minutes in 10% (v/v) Domestos bleach solution (Lever) with gentle shaking. Rinse with sterile distilled water at least three times.
Note, due to asynchronous development, only half or two thirds of the seeds on any one ear will be suitable, the seeds nearest to the peduncle are generally younger and smaller.
3.2 Isolation of immature embryos
3.2.1 Isolate the embryos from the seed under a stereo microscope in a sterile environment using a sharp scalpel.
3.2.2 Remove and discard the embryo axis first then isolate the remaining portion of the embryo which is now referred to as the scutellum.
3.2.3 Plate scutella with the axis side (now removed) down onto semi-solid inoculation medium in 55 mm Petri dishes, about 50 scutella per plate.
3.2.4 It is important to inoculate each plate of 50 scutella with Agrobacterium tumefaciens, as described below, before isolating embryos for the next plate.
4 Inoculation of scutella with Agrobacterium tumefaciens
4.1 Take the resuspended Agrobacterium suspension from the shaker, add 60 μl 1% Silwet to make a final concentration of 0.015% and pour the whole 4 ml over a batch of 50 plated scutella.
4.2 Incubate for 1–3 hours at room temperature while preparing more scutella for inoculation as described in 3.2.
4.3 Transfer the scutella without blotting, keeping the ex-axis side down, onto fresh inoculation medium in 55 mm dishes. Allow to co-cultivate in the dark at 22–23°C for 2–3 days.
5 Control of Agrobacterium and induction of embryogenic calli, regeneration and selection
5.1 After 2–3 days co-cultivation, transfer all scutella to induction medium (Table 4) and continue to incubate in the dark at 24–25°C.
Table 4 Composition of double-strength culture media. All concentrations are shown double-strength except for the supplements added after pH adjustment and sterilisation which are shown at their final concentrations.
Component Inoculation (/L) Induction (/L) RDZ (/L) RPPT (/L) R (/L)
MS Macro salts (×10) 200 ml 200 ml 200 ml 200 ml 200 ml
L7 Micro salts (×1000) 2 ml 2 ml 2 ml 2 ml 2 ml
FeNaEDTA (×100) 20 ml 20 ml 20 ml 20 ml 20 ml
MS vitamins (×1000) 2 ml 2 ml - - -
Vitamins/Inositol (×200) - - 10 ml 10 ml 10 ml
Inositol 200 mg 200 mg 200 mg 200 mg 200 mg
Glutamine 1 g 1 g - - -
Casein hydrolysate 200 mg 200 mg - - -
MES 3.9 g 3.9 g - - -
Glucose 20 g - - - -
Maltose 80 g 80 g 60 g 60 g 60 g
pH adjusted to 5.8 then autoclaved pH adjusted to 5.7 then filter sterilised
2,4-D 2 mg 0.5 mg 0.1 mg - -
Picloram 2.0 mg 2.0 mg - - -
Acetosyringone 200 μM - - - -
Timentin - 160 mg 160 mg 160 mg 160 mg
Zeatin - - 5 mg - -
PPT - - - 2–4 mg 3–4 mg
5.2 After 18 days, transfer embryogenic calli to RDZ medium (Table 4), and incubate at 24–25°C but in the light. Embryogenic calli derived from the same immature embryo should be kept intact without breaking up.
5.3 After 3 weeks, transfer embryogenic calli to selection medium RPPT (or appropriate selection agent, Table 4). At this point, the calli can be broken into defined shoots/roots, but it is important to keep these together, or mark them clearly as there is possibility that these may be clones.
5.4 Continue transferring to fresh RPPT every 3 weeks until PPT tolerant plantlets are ready to be potted to soil.
Note, at the end of the first round of selection, some of the transgenic plants may be identified by GUS assay on leaf fragments. If they have good strong roots, they may be transferred to soil or put into the vernalisation room immediately, otherwise, transfer them to R medium without PPT for root strengthening (Table 4).
6 Materials
6.1 Media for growing Agrobacterium tumefaciens
See Table 3.
6.2 Media for plant tissue culture
6.2.1 Plant tissue culture media are prepared from stock solutions at double strength to allow the addition of an equal volume of gelling agent; Phytagel for inoculation and induction media, agargel for RDZ, RPPT, and R media. Gelling agents are also prepared at double strength (Phytagel at 4 g/l and agargel at 10 g/l) and autoclaved at 121°C for 20 min (see Table 4).
6.2.2 To make single-strength liquid inoculation media for resuspending Agrobacterium cells in section 2.3, simply mix double-strength medium with autoclaved, distilled water.
Stock solutions for basal culture media
Detailed below are the recipes for stock solutions of basal culture media components adapted from [50].
6.2.3 MS Macrosalts (×10):
16.5 g/l NH4NO3 (Fisher Scientific, Leicestershire, UK),
19.0 g/l KNO3 (Sigma-Aldrich, Dorset, UK),
1.7 g/l KH2PO4 (Fisher Scientific UK),
3.7 g/l MgSO4·7H2O (Fisher Scientific UK),
4.4 g/l CaCl2·2H2O (Fisher Scientific UK).
Note, Dissolve each component in distilled water separately before mixing. Autoclave at 121°C for 20 min and store at 4°C.
6.2.4 L7 Microsalts (×1000):
15.0 g/l MnSO4 (Fisher Scientific UK),
5.0 g/l H3BO3 (Fisher Scientific UK),
7.5 g/l ZnSO4·7H2O (Fisher Scientific UK),
0.75 g/l KI (Fisher Scientific UK),
0.25 g/l Na2MoO4·2H2O (VWR International Ltd., Leicestershire, UK),
0.025 g/l CuSO4·5H2O (Fisher Scientific, UK),
0.025 g/l CoCl2·6H2O (Sigma-Aldrich).
Note, MnSO4 may have various hydrated states which will alter the required weight. For MnSO4·H2O, add 17.05 g/l, for MnSO4·4H2O, add 23.22 g/l, for MnSO4·7H2O, add 27.95 g/l. Prepare 100 ml microsalt stock solution at a time. Filter sterilise, and store at 4°C.
6.2.5 MS Vitamins (-Glycine) (×1000):
0.1 g/l Thiamine HCl (Sigma-Aldrich),
0.5 g/l Pyridoxine HCl (Sigma-Aldrich),
0.5 g/l Nicotinic acid (Sigma-Aldrich).
Prepare 100 ml at a time. Filter sterilise, and store at 4°C.
6.2.6 Vitamins/Inositol (×200):
40.0 g/l Myo-Inositol (Sigma-Aldrich),
2.0 g/l Thiamine HCl (Sigma-Aldrich),
0.2 g/l Pyridoxine HCl (Sigma-Aldrich),
0.2 g/l Nicotinic acid (Sigma-Aldrich),
0.2 g/l Ca-Pantothenate (Sigma-Aldrich),
0.2 g/l Ascorbic acid (Sigma-Aldrich).
Filter sterilize and store at -20°C in 10 ml aliquots.
6.2.7 Supplements
• Acetosyringone (3',5'-Dimethoxy-4'-hydroxyacetophenone) (Aldrich D12,440-6: MW-196.20), Dissolve in 70% ethanol to give 10 mg/ml or 50 mM stock solution. Filter sterilise, aliquot and store at -20°C.
• 2,4-Dichlorophenoxyacetic acid (2,4-D) (Sigma-Aldrich), 1 mg/ml in ethanol/water (dissolve powder in ethanol then add water to volume). Filter sterilise, and store at -20°C in 1 ml aliquots.
• Zeatin mixed isomers (10 mg/ml) (Sigma-Aldrich), Dissolve powder in small volume 1 M HCl and make up to volume with water, mix well/vortex. Filter sterilise, and store at -20°C in 1 ml aliquots.
• Picloram (1 mg/ml) (Sigma-Aidrich), Dissolve picloram in water, filter sterilise and store at -20°C in 2 ml aliquots.
• Timentin (300 mg/ml) (Melford, UK), Dissolve Timentin (Ticarcillin/Clavulanic (15:1)) in water, filter sterilise and store at -20°C in 1 ml aliquots.
• PPT (10 mg/ml)(Glufosinate Ammonium) (Melford, UK), Dissolve in water, mix well/vortex, filter sterilize, and store at -20°C in 1 ml aliquots.
• Silwet L-77 (1% v/v) (Lehle seeds, USA), Dissolve in water, filter sterilize, and store at 4°C in 0.5 ml aliquots.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
Rothamsted Research receives grant-aided support from the Biotechnological and Biological Sciences Research Council UK. HW and AD were funded by the Department of the Environment, Food and Rural Affairs UK.
==== Refs
James C Preview: Global status of commercial biotech/gm crops: 2004 ISAAA Briefs No 32 www.ISAAA.org 2004
Vasil V Castillo AM Fromm ME Vasil IK Herbicide resistant fertile transgenic wheat plants obtained by microprojectile bombardment of regenerable embryogenic callus Bio-Technology 1992 10 667 674
Vasil V Srivastava V Castillo AM Fromm ME Vasil IK Rapid production of transgenic wheat plants by direct bombardment of cultured immature embryos Bio-Technology 1993 11 1553 1558
Weeks JT Anderson OD Blechl AE Rapid production of multiple independent lines of fertile transgenic wheat (Triticum aestivum) Plant Physiol 1993 102 1077 1084 12231889
Nehra NS Chibbar RN Leung N Caswell K Mallard C Steinhauer L Baga M Kartha KK Self-fertile transgenic wheat plants regenerated from isolated scutellar tissues following microprojectile bombardment with 2 distinct gene constructs Plant J 1994 5 285 297 10.1046/j.1365-313X.1994.05020285.x
Becker D Brettschneider R Lorz H Fertile transgenic wheat from microprojectile bombardment of scutellar tissue Plant J 1994 5 299 307 8148881 10.1046/j.1365-313X.1994.05020299.x
Jones HD Wheat transformation: current technology and applications to grain development and composition J Cereal Sci 2005 41 137 147 10.1016/j.jcs.2004.08.009
Barcelo P Rasco-Gaunt S Thorpe C Lazzeri PA Transformation and gene expression Adv Bot Res 2001 34 59 126
Sahrawat AK Becker D Lutticke S Lorz H Genetic improvement of wheat via alien gene transfer, an assessment Plant Sci 2003 165 1147 1168 10.1016/S0168-9452(03)00323-6
Altpeter F Baisakh N Beachy R Bock R Capell T Christou P Daniell H Datta K Datta S Dix PJ Fauquet C Huang N Kohli A Mooibroek H Nicholson L Nguyen TT Nugent G Raemakers K Romano A Nomers DA Stoger E Taylor N Visser R Particle bombardment and the genetic enhancement of crops: Myths and realities Mol Breed 2005 15 305 327 10.1007/s11032-004-8001-y
Cheng M Lowe BA Spencer TM Ye XD Armstrong CL Factors influencing Agrobacterium-mediated transformation of monocotyledonous species In Vitro Cell Dev Biol-Plant 2004 40 31 45 10.1079/IVP2003501
Smith RH Hood EE Agrobacterium-tumefaciens transformation of monocotyledons Crop Sci 1995 35 301 309
Dai SH Zheng P Marmey P Zhang SP Tian WZ Chen SY Beachy RN Fauquet C Comparative analysis of transgenic rice plants obtained by Agrobacterium-mediated transformation and particle bombardment Mol Breed 2001 7 25 33 10.1023/A:1009687511633
Travella S Ross SM Harden J Everett C Snape JW Harwood WA A comparison of transgenic barley lines produced by particle bombardment and Agrobacterium-mediated techniques Plant Cell Reports 2005 23 780 789 15761662 10.1007/s00299-004-0892-x
Haliloglu K Baenziger PS Agrobacterium tumefaciens-mediated wheat transformation Cereal Res Commun 2003 31 9 16
Hu T Metz S Chay C Zhou HP Biest N Chen G Cheng M Feng X Radionenko M Lu F Fry J Agrobacterium-mediated large-scale transformation of wheat (Triticum aestivum L.) using glyphosate selection Plant Cell Reports 2003 21 1010 1019 12835912 10.1007/s00299-003-0617-6
Cheng M Fry JE Pang SZ Zhou HP Hironaka CM Duncan DR Conner TW Wan YC Genetic transformation of wheat mediated by Agrobacterium tumefaciens Plant Physiol 1997 115 971 980 12223854
Cheng M Hu TC Layton J Liu CN Fry JE Desiccation of plant tissues post-Agrobacterium infection enhances t-DNA delivery and increases stable transformation efficiency in wheat In Vitro Cell Dev Biol-Plant 2003 39 595 604 10.1079/IVP2003471
Weir B Gu X Wang MB Upadhyaya N Elliott AR Brettell RIS Agrobacterium tumefaciens-mediated transformation of wheat using suspension cells as a model system and green fluorescent protein as a visual marker Aust J Plant Physiol 2001 28 807 818
Khanna HK Daggard GE Agrobacterium tumefaciens-mediated transformation of wheat using a superbinary vector and a polyamine-supplemented regeneration medium Plant Cell Reports 2003 21 429 436 12789445
Wu H Sparks C Amoah B Jones HD Factors influencing successful Agrobacterium-mediated genetic transformation of wheat Plant Cell Reports 2003 21 659 668 12789416
Zhou H Arrowsmith JW Fromm ME Hironaka CM Taylor ML Rodriguez D Pajeau ME Brown SM Santino CG Fry JE Glyphosate-tolerant cp4 and gox genes as a selectable marker in wheat transformation Plant Cell Reports 1995 15 159 163
Vain P McMullen MD Finer JJ Osmotic treatment enhances particle bombardment-mediated transient and stable transformation of maize Plant Cell Reports 1993 12 84 88 10.1007/BF00241940
Brettschneider R Becker D Lorz H Efficient transformation of scutellar tissue of immature maize embryos Theor Appl Genet 1997 94 737 748 10.1007/s001220050473
Chen L Zhang S Beachy RN Fauquet CM A protocol for consistent, large scale production of fertile transgenic rice plants Plant Cell Reports 1998 18 25 31 10.1007/s002990050526
Rasco-Gaunt S Riley A Cannell M Barcelo P Lazzeri PA Procedures allowing the transformation of a range of european elite wheat (Triticum aestivum L.) varieties via particle bombardment J Exp Bot 2001 52 865 874 11413224
Urushibara S Tozawa Y Kawagishi-Kobayashi M Wakasa K Efficient transformation of suspension-cultured rice cells mediated by Agrobacterium tumefaciens Breed Sci 2001 51 33 38 10.1270/jsbbs.51.33
Arencibia AD Carmona ER Tellez P Chan MT Yu SM Trujillo LE Oramas P An efficient protocol for sugarcane (Saccharum spp. L.) transformation mediated by Agrobacterium tumefaciens Transgenic Res 1998 7 213 222 10.1023/A:1008845114531
Uze M Wunn J Puonti-Kaerlas J Potrykus I Sautter C Plasmolysis of precultured immature embryos improves Agrobacterium-mediated gene transfer to rice (Oryza sativa L.) Plant Sci 1997 130 87 95 10.1016/S0168-9452(97)00211-2
Uze M Potrykus I Sautter C Factors influencing t-DNA transfer from agrobacterium to precultured immature wheat embryos (Triticum aestivum L.) Cereal Res Commun 2000 28 17 23
Jin SG Komari T Gordon MP Nester EW Genes responsible for the supervirulence phenotype of Agrobacterium-tumefaciens A281 J Bacteriol 1987 169 4417 4425 2443480
Lazo GR Stein PA Ludwig RA A DNA transformation-competent Arabidopsis genomic library in Agrobacterium Bio-Technology 1991 9 963 967 1368724
Hood EE Fraley RT Chilton MD Virulence of Agrobacterium-tumefaciens strain-A281 on legumes Plant Physiol 1987 83 529 534 16665283
Komari T Transformation of callus-cultures of nine plant-species mediated by Agrobacterium Plant Sci 1989 60 223 229 10.1016/0168-9452(89)90170-2
Komari T Transformation of cultured-cells of Chenopodium quinoa by binary vectors that carry a fragment of DNA from the virulence region of pTiBo542 Plant Cell Reports 1990 9 303 306 10.1007/BF00232856
Hellens RP Edwards EA Leyland NR Bean S Mullineaux PM Pgreen: A versatile and flexible binary Ti vector for Agrobacterium-mediated plant transformation Plant Mol Biol 2000 42 819 832 10890530 10.1023/A:1006496308160
Amoah BK Wu H Sparks C Jones HD Factors influencing Agrobacterium-mediated transient expression of uidA in wheat inflorescence tissue J Exp Bot 2001 52 1135 1142 11432931 10.1093/jexbot/52.358.1135
Van Larebeke N Engler G Holsters M Van den Elsacker S Zaenen I Schilperoort RA Schell J Large plasmid in Agrobacterium tumefaciens essential for crown gall-inducing ability Nature 1974 252 169 170 4419109 10.1038/252169a0
Ashby AM Watson MD Loake GJ Shaw CH Ti plasmid-specified chemotaxis of Agrobacterium tumefaciens C58C1 towards vir-inducing phenolic compounds and soluble factors from monocolyledonous and dicotyledonous plants J Bacteriol 1988 170 4181 4187 3410827
Hansen G Das A Chilton MD Constitutive expression of the virulence genes improves the efficiency of plant transformation by Agrobacterium Proc Natl Acad Sci U S A 1994 91 7603 7607 8052627
van der Fits L Deakin EA Hoge JHC Memelink J The ternary transformation system: Constitutive virG on a compatible plasmid dramatically increases Agrobacterium-mediated plant transformation Plant Mol Biol 2000 43 495 502 11052201 10.1023/A:1006440221718
Wenck AR Quinn M Whetten RW Pullman G Sederoff R High-efficiency Agrobacterium-mediated transformation of Norway spruce (Picea abies) and loblolly pine (Pinus taeda) Plant Mol Biol 1999 39 407 416 10092170 10.1023/A:1006126609534
Ke J Khan R Johnson T Somers DA Das A High-efficiency gene transfer to recalcitrant plants by Agrobacterium tumefaciens Plant Cell Reports 2001 20 150 156 10.1007/s002990000271
McCormac AC Wu HX Bao MZ Wang YB Xu RJ Elliott MC Chen DF The use of visual marker genes as cell-specific reporters of Agrobacterium-mediated t-DNA delivery to wheat (Triticum aestivum L.) and barley (Hordeum vulgare L.) Euphytica 1998 99 17 25 10.1023/A:1018303102488
Mitic N Nikolic R Ninkovic S Miljus-Djukic J Neskovic M Agrobacterium-mediated transformation and plant regeneration of Triticum aestivum L Biol Plant 2004 48 179 184 10.1023/B:BIOP.0000033442.15611.7d
Xia G-ML He Z-Y Chen H-M C-X Brettell R Transgenic plant regeneration from wheat (Triticum aestivum L.) mediated by Agrobacterium tumefaciens Acta Phytophysiologica Sinica 1999 25 22 28
Komari T Hiei Y Saito Y Murai N Kumashiro T Vectors carrying two separate T-DNAs for co-transformation of higher plants mediated by Agrobacterium tumefaciens and segregation of transformants free from selection markers Plant J 1996 10 165 174 8758986 10.1046/j.1365-313X.1996.10010165.x
Hiei Y Ohta S Komari T Kumashiro T Efficient transformation of rice (Oryza-sativa L.) mediated by Agrobacterium and sequence-analysis of the boundaries of the T-DNA Plant J 1994 6 271 282 7920717 10.1046/j.1365-313X.1994.6020271.x
Christensen AH Quail PH Ubiquitin promoter-based vectors for high-level expression of selectable and/or screenable marker genes in monocotyledonous plants Transgenic Res 1996 5 213 218 8673150 10.1007/BF01969712
Sparks CA Jones HD Curtis IP Transformation of wheat by biolistics Tansgenic crops of the world – essential protocols Springer 2004 454
Garfinkel DJ Nester EW Agrobacterium tumefaciens mutants affected in crown gall tumorigenesis and octopine catabolism J Bacteriol 1980 144 732 743 6253441
|
16270934
|
PMC1277018
|
CC BY
|
2021-01-04 16:37:29
|
no
|
Plant Methods. 2005 Sep 5; 1:5
|
utf-8
|
Plant Methods
| 2,005 |
10.1186/1746-4811-1-5
|
oa_comm
|
==== Front
Plant MethodsPlant Methods1746-4811BioMed Central London 1746-4811-1-61627094410.1186/1746-4811-1-6MethodologyProtocol: Precision engineering of plant gene loci by homologous recombination cloning in Escherichia coli Roden Laura C [email protected]öttgens Berthold [email protected]öttgens Effie S [email protected] Broom's Barn Research Station, Higham, Bury St Edmunds, Suffolk IP28 6NP, UK2 Cambridge Institute for Medical Research, Hills Road, Cambridge, CB2 2XY, UK3 Dept. Mol. & Cell Biol., UCT, Private Bag Rondebosch, 7701, Cape Town, South Africa2005 29 9 2005 1 6 6 14 7 2005 29 9 2005 Copyright © 2005 Roden et al; licensee BioMed Central Ltd.2005Roden 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.
Plant genome sequence data now provide opportunities to conduct molecular genetic studies at the level of the whole gene locus and above. Such studies will be greatly facilitated by adopting and developing further the new generation of genetic engineering tools, based on homologous recombination cloning in Escherichia coli, which are free from the constraints imposed by the availability of suitably positioned restriction sites. Here we describe the basis for homologous recombination cloning in E. coli, the available tools and resources, together with a protocol for long range cloning and manipulation of an Arabidopsis thaliana gene locus, to create constructs co-ordinately driven by locus-specific regulatory elements.
==== Body
Introduction
Plant bacterial artificial chromosome (BAC) resources are being generated for ever increasing numbers of species, providing scientists with long-range physical maps and associated sequence data for both model and crop plants. This provides opportunities for reverse genetics and functional studies at the level of the gene locus and above. The latter requires methods for the cloning and manipulation of large DNA fragments, without the limitations imposed by the need for suitably positioned restriction enzyme sites. Significant advances in this respect arose from the development of homologous recombination (HR) cloning in Escherichia coli, based on RecE/RecT (ET) [1,2] and λ RED operon gene products [3,4]. Essentially, in ET-based strategies, PCR-amplified linear DNA fragments with short regions of homology (~50 bp to 60 bp) are precisely targeted into any DNA sequence including high copy number plasmids, the E. coli chromosome and BACs. RED-based protocols rely on a defective λ prophage to provide functions that protect and recombine the linear DNA fragments, under the control of a temperature sensitive λ cl-repressor, with recombinogenic functions switched on at 42°C and off at 32°C. This fixed induction window helps to reduce unwanted rearrangements, allowing DNA to be stably cloned.
HR-cloning in E. coli is widely used in the biomedical research field and is becoming an established tool for BAC engineering in functional genomic studies [5]. Its applications include recombinogenic targeting for gene disruption or replacement and subcloning of BAC DNA by direct isolation of specific genomic regions. A general schematic of HR cloning is given in Fig. 1. Thus, the construction of transgenes for plant functional genomics or the next generation of genetically modified crop plants may benefit from the level of precision engineering offered by HR-cloning.
Figure 1 Schematic representation of the basic applications of homologous recombination cloning in E. coli for genetic engineering. Homologous recombination cloning in E. coli can be used for gene replacement (A), insertion (B) or sub-cloning of target sequences into alternative plasmid vectors. The recombination is mediated by linear DNA fragments (usually generated by PCR), including target site-specific homology arms and a counter selectable antibiotic resistance gene marker.
Our interest in long-range HR cloning was driven by a desire to create plant-specific tools and transgene constructs that target expression to the shoot apical meristem. We wanted to express the bean (Phaseolus coccineus) GAPc2ox1 (encoding GA 2-OXIDASE 1, which degrades bioactive gibberellin) in the shoot apex of sugar beet (Beta vulgaris) plants and study the effect on flowering. We present details of our constructs and the molecular tools (plasmids) developed to create these constructs by RED cloning.
Materials
Reagents
• E. coli strain EL250 (genotype DH10B [λcI857(cro-bioA)<> araC-PBADflpe] where <> indicates that cro-bioA has been substituted with araC-PBADflpe) available from the authors of [3] who have developed a number of different strains including EL350 (with inducible araC-PBADcre). These strains carry a defective λ prophage with red and gam recombination genes under the control of the λPL promoter and exo and bet tightly controlled by the temperature sensitive cI857 repressor. Exo and Beta provide recombinogenic function while Gam inhibits the E. coli RecBCD nuclease from degrading electroprated linear DNA fragments. The promoter of the araBAD operon (PBAD) is induced by L-arabinose for flpe and cre expression enabling removal of sequences between FRT and LoxP sites respectively. We used EL250 to enable removal of the kanamycin gene in our FRT-mPGK-Tn5-neo-FRT cassette. OUR RESULTS: The marker gene was removed as described [3] and worked with 90%–100% efficiency and we were able to recover 100 s of colonies which had become kanamycin sensitive.
• Luria Bertani (LB) broth and plates supplemented with antibiotics as required
• Fully sequenced BAC, PAC or other clones with desired gene locus. Plant BAC and PAC clones are widely available from a number of different sources, including individual labs and organisations e.g. The Arabidopsis Biological Resource Centre (ABRC) or the Nottingham Arabidopsis Stock Centre (NASC) at and, for rice genes and others. The AtSTM locus used in our experiments is cloned in BAC F24o1, sourced from the Arabidopsis Biological Resource Centre, Columbus, Ohio.
• pUC-based vectors to be used for making (i) the locus rescue gap-repair construct (must be counter selectable to the BAC/PAC), and (ii) the gene of interest (GOI) targeting cassette construct (must contain a counter selectable marker to the gap-repair construct).
• High fidelity Taq DNA polymerase. Preferably one which retains A-tails for TA cloning, e.g. the Expand High Fidelity PCR system (Roche Diagnostics).
• PCR primers – four locus-specific primers to amplify DNA fragments at the locus border flanks for the gap-repair rescue construct and two target site-specific primers (minimum 70 bp long) to generate GOI targeting products with destination site specific 5' and 3' homology arms.
• PCR product and gel purification kits e.g. the Qiagen QIAquick™ range and DpnI restriction enzyme – used to remove plasmid templates from PCR reactions because it only cleaves methylated sites.
• General reagents for standard gene cloning and gel electrophoresis
Equipment
• Orbital shaking incubator
• Orbital shaking water bath e.g. Grant OLS 200 – essential for induction of recombination functions in bacterial cells.
• Electroporator e.g. Bio-Rad E. coli Pulser
• PCR Machine
• Long wave UV transilluminator – long wave ultra violet light is less damaging to DNA during excision of bands from gels. UV-damaged DNA will not recombine efficiently.
• Electrophoresis equipment capable of field inversion gel electrophoresis (FIGE) or pulsed field gel electrophoresis (PFGE) e.g. BioRad CHEF DR-II, DR-III or Mapper™ XA, for efficient resolution of large DNA fragments
• Spectrophotometer for cell density quantification
• Temperature controlled centrifuge able to run at 4°C
Protocol
The protocols outlined below describe the development of (i) an AtSTM-locus specific gap-repair rescue vector, (ii) a plant gene targeting construct with a removable kanamycin resistance marker cassette from pGK-FRT [6], under the control of both the bacterial Tn5 promoter and the mouse phosphoglycerate kinase (mPGK) promoter for selection in prokaryotes and eukaryotes respectively. This provides templates for PCR amplification of selectable gene fragments that can be precisely targeted into any desired gene locus; and (iii) a bean (Phaseolus coccineus) GAPc2ox1 transformation construct, co-ordinately driven by "all" AtSTM locus elements, designated pSTM17::GAPc2ox1. We have also constructed a pENTR4-based AtSTM gap-repair rescue vector for the production of a Gateway™ (Invitrogen) compatible entry clone and generic T-DNA transformation constructs as well as an mgfp5-ER targeting cassette. The pSTM17::GAPc2ox1 was successfully transformed into sugar beet, demonstrating for the first time that the mouse PGK promoter is fully functional in transgenic plants, thus enabling the direct exploitation of existing mammalian tools.
Key steps in the EL250 RED-HR locus rescue and engineering procedure
1. Design of PCR primers for amplification of locus rescue (retrieval) homology arms and also for GOI targeting
2. Construction of a gap-repair locus rescue vector.
3. Construction of a targeting vector containing the GOI upstream of a counter selectable marker (different from that in the gap-repair construct).
4. Electroporation of EL250 cells with the BAC or clone containing the desired gene locus and preparation of electrocompetent BAC/EL250 cells induced for Exo, Beta and Gam functions.
5. Performance of gap-repair locus rescue, in cells treated as above; selection of recombinants and confirmation by restriction digestion analysis and sequencing. Transformation of the rescued locus plasmid into fresh EL250 cells.
6. PCR amplification, purification and quantification of the GOI targeting cassette and its site-specific recombination into the rescued locus plasmid in EL250 cells. Selection and confirmation of recombinants as above.
The recombineered plasmid is now ready for application in functional analyses as desired.
Primer design and plasmid constructs
Primers
Primer sequences for the AtSTM HR rescue protocol described here are given in Additional file 1. Careful attention must be paid to the design of primers for generating locus rescue (LR) homology arms (HA) to ensure that their orientation in the resultant gap-repair vector is correct for DNA double stranded break repair homologus recombination. A total of four short (18 to 20 bp) primers will be required and can if necessary, include restriction sites to enable cloning into the gap-repair vector so that the gap-repair construct can be linearised between the LR-HAs. Fig. 2A shows our AtSTM gap-repair construct.
Figure 2 Maps of the Arabidopsis BAC F24o1 and the AtSTM gap-repair construct. A: A physical map of BAC F24o1, showing the relative positions of the 21 coding sequences (CDS), including STM (CDS 9) and its immediate neighbours (CDS 7, 8 and 10) which, are illustrated in different colours and greater detail to show the exon blocks making up each open reading frame. Grey arrows show the orientations of the serine protease, predicted pentatricopeptide repeat protein (PPR), SHOOTMERISTEMLESS (STM) and Zinc Finger Protein (ZFP) open reading frames. B: The gap-repair construct and a schematic representation of the basic protocol used to generate the AtSTM downstream (yellow vertical dashed line: W - X) and upstream (blue diagonal dashed line: Y - Z) homology arms to create it in the pBluescriptII KS+ vector backbone. W, X, Y and Z are the PCR primers used to generate each homology arm fragment The SalI, SphI, and HindIII cloning sites were incorporated into the PCR primers, the sequences of which are given in Additional file 1.
For the GOI targeting cassette, primers must include at least 30 to 50 bp at the 5' end, to provide homology arms for site-specific recombination. The target site sequence must not have any mismatches as this will inhibit recombination. It is therefore essential to source primers from suppliers able to guarantee sequences of long primers. Our primers were custom made by Sigma Genosys.
Plasmid constructs
IN OUR HANDS: Creation of these basic plasmids was the key limiting step as it is dependent on conventional cloning and therefore, on the availability of suitably placed restriction sites. However, once constructed, the gene targeting constructs can be used to target the expression cassette into any desired site whereas the gap-repair construct is suitable only for subcloning the specific gene fragment/locus. Our targeting construct backbone has therefore been designed to include a plant-specific polyA signal (Nopaline synthase (nos) termination sequence), for generic use with any plant cDNA sequence.
AtSTM gap-repair construct
Using BAC F24o1 DNA template (represented in Fig. 2A), and the Expand High Fidelity DNA polymerase PCR system (Roche Diagnostics) we amplified 564 bp (incorporating 5'SalI and 3'SphI sites) and 479 bp (incorporating 5'SphI and 3'HindIII sites) homology arms respectively at the downstream and upstream flanks of the AtSTM locus (Fig. 2B). At the start of our project, the pentatricopeptide repeat protein (PPR) downstream of the STM coding sequence was annotated as a predicted ORF and we therefore opted to include it in the STM locus fragment. Now, it would be excluded as the locus boundary. PCR reactions included primers HindIII 3' hyp/SphI 5' hyp or Sprot H1 SalI/Sprot H1 SphI at 0.3 μM each and were incubated for 1 cycle at 94°C for 2 min. followed by 30 cycles of 94°C for 15 sec; 64°C for 30 sec; 72°C for 1.5 min; and 1 cycle of 72°C for 5 min. The individual products were then sub-cloned into pGEM-T Easy (Promega), recovered and cloned into pBluescript II SK+ (Stratagene) in a three-way ligation reaction, to create the gap-repair construct.
NOTE: In our hands, cloning of PCR products is more efficient if we shuttle them via a PCR cloning vector. Any TA cloning vector is suitable. In this case, it is important to ensure that the proof reading activity of the Taq Polymerase used does not remove A-tails.
Targeting construct backbone
The 1.8 kb FRT-mPGK-Tn5-neo-FRT cassette was PCR amplified with SacII in the 5' primer (PGK-FRT upper) and ApaI in the 3' primer (PGK-FRT lower) from pPGK-FRT (obtained under a Material Transfer Agreement from Dr Francis Stewart, EMBL, Heidelberg – now at University of Technology, Dresden) and cloned into SacII/ApaI cloning sites downstream of the Green Fluorescent Protein (GFP) gene of polyGFP3 (a kind gift from Dr E. Amaya, Gurdon Institute, Cambridge). The GFP gene was then replaced with the Nopaline synthase Terminator (NosTer) from pAL69 (pFC6 with NosTer in the multiple cloning site. A kind gift from Dr Dave Lonesdale at the John Innes Centre, Norwich, UK), to create the basic plant gene targeting vector pNosTerFRT-neo (Fig. 3A), which can be used to receive any GOI.
Figure 3 Map of the targeting plasmid backbone pNosTerFRT-neo and GAPc2ox1/mgfp5-ER targeting cassettes. A: Structure of the targeting plasmid backbone pNosTerFRT-neo, showing the relative positions of the Nopaline synthase polyA signal sequence (NosTer), the FRT-mPGK-Tn5-neo-FRT selection cassette and the restriction sites mapped up and downstream of the NosTer sequence. Upstream sites may be used for cloning any GOI cDNAs to create targeting constructs. Sites shown in green were introduced with the NosTer fragment; all other sites originate from PolyGFP3. Unique sites are underlined. B: Representations of GOI sequences with available upstream and downstream restriction sites as appropriate, together with the cloning strategies employed to clone them into pNosTerFRT-neo to make the targeting cassette constructs.
GAPc2ox targeting construct
The full length GAPc2ox1 cDNA in plasmid pST33 (a kind gift from Drs Andy Philips and Peter Hedden, Rothamsted Research) was excised with SpeI/XhoI and cloned into the NheI/SalI site of pNosTerFRT-neo (Fig. 3Bi).
mgfp5-ER targeting cassette
The mgfp5-ER gene (GenBank U87974) was isolated from pBIN35S-mgfp5-ER (a kind gift from Dr Jim Haseloff, Department of Plant Sciences, University of Cambridge, UK) as a BamHI/SacI fragment and cloned into the SmaI site of pNosTerFRT-neo (Fig. 3Bii)
WARNING! GUS reporter cassettes are not suitable as they are able to recombine with the endogenous (chromosomal) E. coli gene during HR cloning.
RED Cloning Protocol
Original methods and information on how to obtain host cells can be found at the recombineering website
A: Preparation of electrocompetent EL250 cells
These cells will be periodically used to receive new plasmids as they are constructed and required for recombineering. It is therefore advisable to make and store a sizable batch.
1. Streak out cells on LB plates and grow at 32°C. The cells are temperature sensitive and will die at 37°C
2. Inoculate a single colony into 5 ml LB and grow overnight.
3. Inoculate 1 ml of the overnight culture into 50 ml LB in a 500 ml flask and grow at 32°C with shaking at 200 revolutions per minute (rpm) until the cells density has reached OD600 = 0.5 – 0.8.
4. Spin cells at 4°C (rotor must be pre-cooled) and wash with 5 ml ice cold sterile distilled water. Spin and discard supernatant. Repeat wash with 5 ml aliquots of ice cold SDW two more times.
5. Finally resuspend cells in 500 μl of ice cold sterile distilled water and aliquot 100 μl lots into cooled 1.5 ml microfuge tubes.
NOTE: Cells can be used immediately or re-suspended in ice cold sterile 10% (v/v) glycerol and stored at -80°C until required.
B: Transformation of BAC F24o1 and induction of recombinogenic function in EL250
1. On ice, add 10 ng-100 ng of F24o1 DNA to 50 μl of competent EL250 cells. Mix by gentle pipetting and transfer to a pre-cooled 0.1 cm electroporation cuvette.
2. Pulse at 1.75 kV in a Bio-Rad E. coli Gene Pulser. Immediately add 1 ml LB broth and incubate at 32°C for 1–1.5 h in a shaking incubator set at 200 rpm
3. Plate cells on LB kanamycin and select for transformants. NOTE: It is advisable at this stage to check the integrity of the BAC clone by restriction digestion analysis, to ensure that there have been no rearrangements.
4. Grow F24o1/EL250 cells as described in A: steps 1 – 3 except that all LB media must be supplemented with kanamycin (or relevant antibiotic) to select for the BAC. NOTE: before the next step, ensure that the shaking water bath is switched on early and stabilised at 42°C ready for use and pre-warm the conical flask. An ice slurry bath must also be made ready – DO NOT USE JUST ICE – it will not cool cells fast enough.
5. For induction, transfer 10 ml of the growing culture into a pre-warmed 250 ml conical flask and incubate in the water bath at 42 °C with shaking at 200 rpm for a total of 15 min. NOTE: Keep the remaining 40 ml of culture at 32 °C to act as a non-induced control.
6. Immediately after 15 min. place the flask in the ice slurry bath and swirl by hand to quickly cool down the cells. Include a similar flask with 10 ml of non-induced control cells – this will be cooled down and treated in the same way as the test cells from now on. NOTE: (i) Induced cells must be used immediately as they will lose activity above 0°C. Therefore it is important to work quickly from now on. However, cells may be kept on ice for a total of 40 min. without significant loss of activity. (ii) Ensure that the centrifuge and rotor are pre-cooled to 4 °C before the next step.
7. Centrifuge the 10 ml aliquots of induced and control cells for 8 min at 5500 g and at 4 °C. Retrieve pellets and wash three times in 1 ml ice cold sterile distilled water and centrifuge as above. NOTE: To save time, washing steps can be carried out in 1.5 ml microfuge tubes keeping everything ice cold and centrifuging at 4 °C for 20 seconds each time.
8. After final wash, re-suspend the cell pellet in 100 μl of ice cold sterile distilled water. This is enough for two electroporation transformation reactions.
C: AtSTM locus rescue from BAC F24o1 by gap-repair HR
Before starting: Ensure that purified and linearised gap-repair vector is available at concentrations suitable to deliver 10 to 100 ng in volumes up to 10 μl. We strongly recommend gel quantification with known standards as we find this more accurate than OD260 nm measurements.
N.B. All HR experiments should be carried out with the induced and un-induced control cells in parallel.
1. Using linearised gap-repair construct DNA, electroporate induced competent F24o1/EL250 cells as described in B: steps 1 – 2.
2. Select recombinants on LB supplemented with antibiotic marker for the gap-repair vector. We used pBluescriptII KS+ and therefore selected on LB ampicillin. The use of pBluescript also limits the size of insert which can be rescued and 17 kb was the largest fragment we were able to retrieve by gap repair HR cloning.
3. Recover recombinant plasmids and confirm correct recombination events by restriction digestion analysis and sequencing. This is important since incorrect events may still be selectable with the antibiotic marker. OUR RESULTS: The number of colonies recovered was typically small (2 – 4) but of these, 50% were correct. The remainder were the result of illegitimate recombination events. See Fig. 4 for our HR strategy and the result of our AtSTM gap-repair rescue to give the plasmid pBlueAtSTM17. NOTE: Because of the large DNA fragments involved, it is advisable to use Field Inversion Gel Electrophoresis or Pulsed Field Gel Electrophoresis as appropriate for clarity in resolution.
Figure 4 gap repair HR rescue of the AtSTMlocus. Schematic representation of the HR rescue of the AtSTM locus into the pBluescript gap repair vector (A) and the resultant pBlueAtSTM17 construct as expected from the correct recombination event. Diagnostic EcoRV sites are shown in blue, PmeI site in red, together with their co-ordinates within pBlueAtSTM17. SalI and HindIII sites originally engineered in the homology arms of the gap repair vector are shown in green. (B). The results of EcoRV and PmeI restriction digestion of three plasmids resulting from independent recombination events from a single experiment, showing that only the plasmid in lane 3 resulted from the correct recombination. M1 = 1 kb ladder (New England Biolabs); M2 = High Molecular Weight Marker (Invitrogen).
TROUBLE SHOOTING: Growth of un-induced colonies on selective plates suggests incomplete digestion of the gap-repair construct during linearization. However, the number of colonies should be low (we typically recovered 5 – 10 colonies from un-induced cells). Otherwise repeat with improved digestion and/or gel purification of the linearised gap repair construct.
4. Transform the rescued plasmid into fresh EL250 cells and prepare induced competent cells as described, ready for the locus targeting experiment. We designated these cells STM17/EL250 because they contained rescued ~17.5 kb of the AtSTM locus. NOTE: For our application, we use a protoplast based direct transformation method and therefore opted to use pBluescript as the backbone for our gap repair and eventual transformation construct. However, for Agrobacterium-based systems, we recommend using a Gateway compatible Entry vector (available from Invtrogen: ) as this will enable subsequent transfer of the captured, manipulated locus into a T-DNA binary destination vector for example the ones available from Plant Systems Biology (VIB-Gent University: ) or the pEarlyGates vectors, details of which can be found at the website: .
Recently, we have successfully created an AtSTM locus rescue vector based on the Invitrogen pENTR4 Gateway™ compatible vector in which we plan to capture/manipulate the locus as described and determine the success rate of transfer into a promoterless T-DNA destination vector pB7WG2Δ35S (based on pB7WG2 from VIB-Gent University). These are newly available resources that should enable the creation of constructs for the more generic Agrobacterium-mediated plant transformation systems.
WARNING!: Direct use of T-DNA vectors as gap-repair constructs in RED cloning although attractive, may prove problematic because of the common use of a limited number of identical or very similar promoter and polyA signal sequences, which if also present in the targeting cassette will result in illegitimate recombination events. For this reason, we did not attempt any experiments with T-DNA vectors, opting instead to go via the Gateway™ system as detailed above.
D: Replacement of AtSTM exon1 by in-frame fusion of the promotorless GAPc2ox-FRT-neo-FRT targeting cassette
Before starting: The purified, DpnI treated and quantified PCR amplified GOI targeting cassette should be made ready for this experiment.
1. Using up to 100 ng of the PCR amplified GAPc2ox1-FRT-neo-FRT targeting cassette, electroporate induced STM17/EL250 cells as described in B: steps 1 – 2. The targeting cassette was amplified with HotStar Taq DNA polymerase (Qiagen) and primers 28001 frtlow and 2oxexon1 (0.3 μM each) in a 50 μl reaction volume. Incubation conditions were 1 cycle 95°C for 1 min followed by 20 cycles of 94°C for 15 sec; 68°C for 4.5 min. (with a 5 sec. time increment in each cycle); followed by 1 cycle of 68°C for 10 min.
2. Select recombinants on LB kanamycin. Recover plasmids and confirm by restriction digestion analysis and sequencing. Plasmids are now ready for application in functional assays. OUR RESULTS: We recovered many colonies at this stage (100 s), of which ~6 % were correct by restriction digestion analysis and sequencing. For example, we typically screened between 30 and 35 colonies from which two were correct. See Fig. 5 for the results of our GOI targeting experiment. NOTE: It is advisable at this stage to at least sequence across the recombination site into the GOI cassette to confirm the integrity of the gene cassette before proceeding with functional assays in transgenic plants.
Figure 5 Replacement of AtSTM exon 1 by in-frame fusion of GAPc2ox1. A: Schematic representation of the strategy used to generate the GAPc2ox1 cassette from pNosTerFRT-neo (plus GAPc2ox1) plasmid and target it into the AtSTM locus (captured in pBlueAtSTM17) by in-frame substitution of exon 1 sequences. B: FIGE gel results of PmeI and EcoRV digested pSTM17::GAPc2ox1 from two independent recombination events (lanes 2 and 3), clearly showing the increased size of the linearised construct (PmeI digest) and the additional fragment (EcoRV digest) due to the recombination of the GAPc2ox1 cassette. Lane 1 shows the control result with the original pBlueAtSTM17. M1 = 1 kb ladder (New Endgland Biolabs); M2 = High Molecular Weight Marker (Invitrogen).
TROUBLE SHOOTING: High un-induced colony numbers on selective plates suggest targeting cassette template contamination instead of recombination. Check DpnI digests and use this in combination with gel purification to remove template DNA from the target cassette PCR product prior to electoporating for HR. In our experience, if the number of colonies from un-induced cells is at least 50 – 100 fold less than from the induced cells, then it was worth screening colonies from induced cells.
Our final recombineered construct was designated AtSTM17:: GAPc2ox1 and was transformed into sugar beet guard cell protoplasts [7] from which we successfully selected transgenic callus and shoots under kanamycin selection driven by the mouse PGK promoter in the FRT-mPGK-Tn5-neo-FRT cassette (Fig. 6). We are now in the process of conducting phenotypic analyses of our AtSTM17:: GAPc2ox1 plants.
Figure 6 Kanamycin selection and regeneration of transgenic sugar beet guard cells transformed with the pSTM17::GAPc2ox1 constructs containing the neomycin gene driven by the mouse PGK promoter. A: Transgenic sugar beet callus grown from transformed sugar beet guard cell protoplasts cultured on medium supplemented with kanamycin at 100 μg ml-1, clearly showing a difference between the healthy (green) callus and the dead (brown) non-transgenic callus. B: Shoots regenerated from transgenic sugar beet callus as above. C: Resultant transgenic seedlings in compost. D: Agarose gel showing the results of neomycin gene-specific PCR conducted on genomic DNA template extracted from sugar beets generated from two independent transformation events with pSTM17::GAPc2ox1 including the kanamycin driven by the mouse PGK promoter (arrows) compared with individuals from lines transformed with constructs containing a kanamycin cassette driven the CaMV35S promoter (the remaining PCR product bands). Sequences from other regions of the pSTM17::GAPc2ox1 construct were also detected by PCR and Southern blot hybridisation (results not shown). +ve = DNA plasmid template control, -ve = DNA template-free control; M = New England Biolabs 100 bp ladder. Diagnostic PCR for the neomycin gene sequences was performed using the primers Neo-For 5' CAG GAT GAT CTG GAC GAA GA 3' (Tm = 57.3 °C) and Neo-Rev 5' AAG AAG GCG ATA GAA GGC GA 3' (Tm = 57.3°C). The reactions (Qiagen Master Mix) contained 10 μM of each primer, and 50–100 ng genomic DNA template and were incubated at 94°C for 3 min, followed by 30 cycles of 94°C for 15 sec.; 55°C for 30 sec.; 72°C for 1 min and one cycle of 72 °C for 2 min.
Comments
Manipulation of large DNA fragments to make complex constructs for functional genomics or genetic engineering for crop improvement is possible using HR cloning in E. coli. We have successfully used HR cloning in E. coli to sub-clone the Arabidopsis thaliana SHOOTMERISTEMLESS (STM) gene locus from a BAC clone into pBluescript and to replace exon 1 sequences with a Gibberellin 2-oxidase cDNA gene-of-interest cassette tightly linked to an FRT-flanked kanamycin selection marker gene. This cassette is of generic use because firstly, it can be targeted/recombineered into any locus/destination site. Secondly, the kanamycin resistance gene is under the control of both the bacterial Tn5 promoter and the mouse phosphoglycerate kinase promoter (mPGK), which respectively allow for selection in prokaryotes and eukaryotes. We have now demonstrated the utility of the mPGK promoter for driving expression in transgenic plants and this suggests that there may well be increased scope for plant scientists to directly benefit from existing molecular genetic tools developed for application in the biomedical field.
E. coli ET- and RED-HR cloning are well established technologies within the biomedical field and they have many uses besides the creation of transformation constructs with long-range regulatory elements. The identification of regulatory elements or locus control regions located at a distance from the gene sequence can be assisted by this strategy. Point mutations, deletions or insertions, gene fusions and antisense constructs can be engineered on any BAC for functional genomics studies. The scope for plant science is further enhanced by the recently reported application of HR to convert BACs into binary vectors [8] together with (i) the availability of a BAC-based physical map of A. thaliana, (ii) freely available genome sequence information through the Arabidopsis Genome Initiative, (iii) access to rice sequence data and BAC resources through The Institute for Genomic Research (TIGR) and the Rice Genome Resource Center (RGP).
Resources for RED/ET cloning are available from Neal Copeland and Nancy Jenkins for both profit and non-profit organisations. Details can be found at the following website: . The commercial company GeneBridges also offers reagents and a DNA engineering service.
Abbreviations
BAC = Bacterial artificial chromosome; bp = base pairs; FIGE = field inversion gel electrophoresis; GA 2ox = gibberellin 2-oxidase; GFP = gree fluorescent protein; GOI = gene of interest; HA = homology arm(s); HR = homologous recombination; LB = Luria Bertani medium; LR = Locus rescue; mPGK = mouse phosphoglycerate kinase promoter; OD = optical density; PCR = polymerase chain reaction; ORF = open reading frame; rpm = revolutions per minute, UV = ultraviolet.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ESM-G and BG conceived of the study and participated in its design. ESM-G and LCR drafted the manuscript. LCR participated in the design of the study and carried out most of the experimental. ESM-G directed the work, participated in experimental work. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
PCR Primer Sequences. Details of the primers used isolate and manipulate the AtSTM gene locus by homologous recombination in E. coli EL250
Click here for file
Acknowledgements
Ann Mathews, Roz Williamson and Sarah Yallop for technical assistance molecular analyses and sugar beet transformation.
The project was funded by the Biotechnology and Biological Sciences Research Council of the UK as part of the ROPA scheme.
==== Refs
Zhang Y Buchholz F Muyrers JP Stewart AF A new logic for DNA engineering using recombination in Escherichia coli Nat Genet 1998 20 123 128 9771703 10.1038/2417
Zhang Y Muyrers JP Testa G Stewart AF DNA cloning by homologous recombination in Escherichia coli Nat Biotechnol 2000 18 1314 1317 11101815 10.1038/78475
Lee EC Yu D Martinez de Velasco J Tessarollo L Swing DA Court DL Jenkins NA Copeland NG A highly efficient Escherichia coli-based chromosome engineering system adapted for recombinogenic targeting and subcloning of BAC DNA Genomics 2001 73 56 65 11352566 10.1006/geno.2000.6451
Yu D Ellis HM Lee EC Jenkins NA Copeland NG Court DL An efficient recombination system for chromosome engineering in Escherichia coli Proc Natl Acad Sci U S A 2000 97 5978 5983 10811905 10.1073/pnas.100127597
Copeland NG Jenkins NA Court DL Recombineering: a powerful new tool for mouse functional genomics Nat Rev Genet 2001 2 769 779 11584293 10.1038/35093556
Angrand PO Daigle N van der Hoeven F Scholer HR Stewart AF Simplified generation of targeting constructs using ET recombination Nucleic Acids Res 1999 27 e16 10446259 10.1093/nar/27.17.e16
Hall RD Riksen-Bruinsma T Weyens GJ Rosquin IJ Denys PN Evans IJ Lathouwers JE Lefebvre MP Dunwell JM van Tunen A Krens FA A high efficiency technique for the generation of transgenic sugar beets from stomatal guard cells Nat Biotechnol 1996 14 1133 1138 9631066 10.1038/nbt0996-1133
Takken FL Van Wijk R Michielse CB Houterman PM Ram AF Cornelissen BJ A one-step method to convert vectors into binary vectors suited for Agrobacterium-mediated transformation Curr Genet 2004 45 242 248 14745506 10.1007/s00294-003-0481-5
|
16270944
|
PMC1277019
|
CC BY
|
2021-01-04 16:37:28
|
no
|
Plant Methods. 2005 Sep 29; 1:6
|
utf-8
|
Plant Methods
| 2,005 |
10.1186/1746-4811-1-6
|
oa_comm
|
==== Front
Plant MethodsPlant Methods1746-4811BioMed Central London 1746-4811-1-71627094310.1186/1746-4811-1-7MethodologyCell type-specific characterization of nuclear DNA contents within complex tissues and organs Zhang Changqing [email protected] Fang Cheng [email protected] Georgina M [email protected] David W [email protected] Department of Plant Sciences, The University of Arizona, Tucson, Arizona, 85721, USA2 Operon Biotechnologies, Inc., 2705 Artie Street Bldg. 400, Ste. 27, Huntsville, AL 35805, USA2005 4 10 2005 1 7 7 24 8 2005 4 10 2005 Copyright © 2005 Zhang et al; licensee BioMed Central Ltd.2005Zhang 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
Eukaryotic organisms are defined by the presence of a nucleus, which encloses the chromosomal DNA, and is characterized by its DNA content (C-value). Complex eukaryotic organisms contain organs and tissues that comprise interspersions of different cell types, within which polysomaty, endoreduplication, and cell cycle arrest is frequently observed. Little is known about the distribution of C-values across different cell types within these organs and tissues.
Results
We have developed, and describe here, a method to precisely define the C-value status within any specific cell type within complex organs and tissues of plants. We illustrate the application of this method to Arabidopsis thaliana, specifically focusing on the different cell types found within the root.
Conclusion
The method accurately and conveniently charts C-value within specific cell types, and provides novel insight into developmental processes. The method is, in principle, applicable to any transformable organism, including mammals, within which cell type specificity of regulation of endoreduplication, of polysomaty, and of cell cycle arrest is suspected.
==== Body
Background
The amount of DNA contained within a haploid nucleus of eukaryotic organisms is termed the C (constant) value [1]. For many eukaryotes, the nuclei of somatic cells contain a 2C DNA amount, and the growing cells participate in a simple mitotic cell cycle in which four temporally-linked phases, G1, S, G2 and M, serve to separate the processes of DNA duplication (S-phase) from chromosomal segregation (M-phase). Monosomatic tissues containing mitotically active cells therefore are characterized by cells having nuclear DNA contents ranging from 2C to 4C, depending on the position of the cells within the cell cycle. The proportions of cells within these phases is a function of the proportions of cells that are actively cycling and the degree of cycle synchrony, and evidently reflects also whether or not the cells are arrested at particular points within the cell cycle, most commonly G0/G1 (2C) or G2 (4C). In polysomatic tissues, the situation is complicated by the occurrence of an alternative cell cycle, termed endoreduplication, in which successive S-phases are not followed by M-phases. This produces uninucleate cells having multiplicative DNA contents (2n C, where n = 1,2,3..., for most sources of somatic cells, and 3 × 2n-1 C for the endoreduplicated endosperm derived from triploid progenitor cells). Polysomaty is particularly common in higher plants [2]; for some species, such as A. thaliana, it is encountered throughout the mature tissues of the organism [3], while in others it is restricted to specific tissues [4].
The functional significance of the state of the nuclear C-value at which DNA synthesis arrests remains obscure, in part due to a lack of facile and precise methods for identifying its occurrence as a function of specific cell types. It is clear that, in the analysis of developmental gene expression and the cell biology underlying its regulation, the nuclear C-value represents an important parameter reflecting both the cell cycle status of the cell within which the nucleus is located, as well as the participation of the cells of polysomatic tissues within cycles of endoreduplication. Conversely, the regulated arrest of the cell at specific nuclear C-values reflects the activities of regulatory mechanisms about which we know very little.
We wondered if our flow cytometric methods for analysis of nuclear C-values [5] might be combined with transgenic expression of a nuclear-targeted version of the Green Fluorescent Protein (GFP) placed under the regulation of cell type-specific promoters, thereby permitting analysis of the C-value status of specific cell types. We reasoned that the labeling of nuclei of specific cell types with GFP would allow their detection using flow cytometry and, via simultaneous biparametric analysis of DNA content, lead to their assignment to various C-value classes. In this report, we validate this experimental approach, describing recombinant DNA constructions that encode Fluorescent Protein (FP)-fusions that are appropriately targeted to the nucleus, and which are quantitatively retained within the nuclei following cell homogenization. We go on to describe conditions for confocal examination of transgenic plants exhibiting a number of different cell-type specific patterns of expression, and for flow cytometric analysis of homogenates prepared from these plants. We finally employ the method to uncover evidence of cell type-specific arrest of particular cell types within different C-value states. The significance of these observations is discussed.
Results
The proposed experimental concept requires that it be possible to target GFP, or other Fluorescent Proteins, to the nuclei of transgenic plants under the control of cell type-specific promoters, that the nuclei display sufficient fluorescent signal to be detectable by microscopy and flow cytometry, that the GFP-based signal not interfere with counterstaining and flow analysis of nuclear DNA content, and that the GFP-based fluorescence be retained within the nuclei following homogenization and during flow analysis. To be maximally useful, the concept and the procedures should be applicable to plants having small (cf. A. thaliana) as well as larger genomes.
To test this concept, we employed A. thaliana, a model plant species for which a uniquely comprehensive amount of molecular information is available. A. thaliana also comprises one of the smallest nuclear genomes within the flowering plants [6], thereby providing an excellent test of the lower limit of resolution of the methods. For nuclear labeling, we evaluated the performance of a number of different translational fusions of nuclear proteins with GFP. Optimal for our purpose was a fusion of GFP with the coding region of a histone 2A gene (HTA6; At5g59870). Under the transcriptional control of the Cauliflower Mosaic Virus (CaMV) 35S promoter, transgenic A. thaliana plants expressing HTA6-GFP were phenotypically normal, and displayed brightly fluorescent nuclei within all parts of the plant (Figure 1). Nuclei of similar brightness were seen for transgenic plants expressing HTA6-YFP. No effects of transgenic GFP expression were detected on plant fresh weights or root growth rates (Figure 2), or by using whole genome long oligonucleotide microarrays to monitor alterations in gene expression (unpublished data).
Figure 1 Confocal and bright-field images of wild-type plants, and plants transgenically expressing HTA6-GFP and HTA6-YFP under the transcriptional control of the CaMV35S promoter. For the bright-field picture, seeds of the three genotypes were germinated on MS agar plates. 3-day-old, similar-sized seedlings were transferred onto fresh MS agar plates, and were grown in a vertical position for two weeks.
Figure 2 (A) Fresh weights and (B) root growth amounts of plants that are transgenic or non-transgenic for HTA6-GFP expression. Measurements were made on three groups of 20 pooled plants, and the error bars indicate standard deviations. Abbreviations: WT, wild-type; HTA6-GFP, transgenic plant expressing HTA6-GFP.
We concomitantly chose to focus on plant roots: roots of many important crops and model species either are polysomatic or comprise a large proportion of cells arrested at a 4C nuclear DNA content (Figure 3). For example, uniparametric flow analysis of root homogenates prepared from the apical 1 cm regions of the primary root of ten day-old A. thaliana plants identifies four populations of nuclei (Figure 3A), equally spaced along the DNA content axis (logarithmic scale) corresponding to nuclei respectively having 2C, 4C, 8C, and 16C DNA contents. Polysomaty was also observed for root tips of cucumber, pea, and tomato. For the other species examined (tobacco, Vinca, maize, rice, and carrot), polysomaty was absent, but for maize, tobacco, petunia and Vinca, a large minority of the nuclear populations represented cells having 4C nuclei (see also [5]). Our observations are consistent with other compilations [4].
Figure 3 Flow cytometric analysis of the nuclear DNA contents of plant root homogenates. Homogenates, prepared as described by Galbraith et al. [5], stained with DAPI, were analyzed using a Partec CCAII Flow Cytometer (Partec GmbH, Munster, Germany). The instrument PMT and amplification settings were adjusted between samples to provide distributions conveniently distributed across the abscissae. This means that the numerical positions of the nuclear peaks should not be used for comparative analysis of DNA content between species. (A) Arabidopsis thaliana. (B) Pisum sativum. (C) Cucumis sativus. (D) Daucus carota. (E) Nicotiana tabacum. (F) Lycopersicon esculentum. (G) Petunia hybrida. (H) Vinca rosea. (I) Zea mays. (J) Oryza sativa.
Examination of the roots of wild-type and transgenic A. thaliana plants expressing HTA6-GFP was done via confocal microscopy. The confocal images and the corresponding biparametric flow cytometric analyses of the GFP and DNA contents of their nuclei are presented in Figure 4. Wild-type plants display no nuclear GFP fluorescence and, in the flow analysis (Figure 4A), the four populations of nuclei, corresponding to the 2C, 4C, 8C, and 16C nuclei, are located close to the abscissa. In comparison, the roots of plants constitutively expressing HTA6-GFP under the control of the CaMV 35S promoter contained green-fluorescent nuclei, and the flow histograms display clusters of nuclei corresponding in DNA content to 2C, 4C, 8C, and 16C but also producing a GFP signal that increases with DNA content (Figure 4B). The intranuclear GFP fluorescence was stable over the period of time following homogenization required for the flow analyses (Figure 5A), and the amounts of intranuclear GFP fluorescence scaled linearly with nuclear DNA content (Figure 5B). Finally, the proportions of nuclei within the different C-value classes were not significantly different when wild-type and transgenic plants were compared (Figure 6). Within the apical 10 mm of the A. thaliana primary root, therefore, exist 2C, 4C, 8C, and 16C cells, and the nuclei of these cells appear equally capable of accumulating GFP-labelled histone H2A.
Figure 4 Confocal and biparametric flow cytometric analyses of wild-type and transgenic plants expressing nuclear GFP. Flow cytometry was done using a Cytomation MoFlo flow cytometer with laser excitation at 365/488 nm, and biparametric detection of DAPI fluorescence (418–482 nm; FL4; log units), and GFP fluorescence (505–530 nm; FL1; log units), with triggering based on 90°-light scatter [59]. For confocal microscopy, roots were counterstained by dipping in propidium iodide (1 μg/mL in water) for 2 minutes. Abbreviations: p35S: CaMV 35S promoter; pSCR: SCARECROW promoter; pSHR: SHORTROOT promoter; pRPL16B: ribosomal protein large subunit 16B promoter; pSultr2-1: sulfate transporter 2-1 promoter.
Figure 5 Analysis of the stability and amounts of targeted GFP fluorescence within nuclei following homogenization. Biparametric flow cytometric analyses were done of transgenic plants constitutively expressing nuclear GFP at various times following homogenization. A. The modes of the GFP fluorescence distributions for the four classes of nuclei (2C, 4C, 8C, 16C) are plotted as a function of time after homogenization. B. The modes obtained from the GFP fluorescence distributions were averaged across the time course for the different nuclear classes, and these mean values are plotted against C-value. Error bars indicate standard deviations.
Figure 6 Comparison of the distributions of nuclei within the various C-value classes found in the roots of wild-type plants and plants transgenic for expression of HTA6-GFP. Measurements were made on three replicate samples (wild-type and transgenic) each sample comprising ~100 pooled seedlings. The error bars indicate standard deviations.
We next wondered whether the presence of nuclei of different C-value classes might be associated with specific cell types or root sub-regions. To address this question, we produced transgenic plants expressing HTA6-GFP under the control of both cell type-specific and region-specific regulatory sequences. Transgenic plants expressing HTA6-GFP under the control of the Sultr2-1 promoter [7] exhibited nuclear GFP fluorescence restricted to the phloem companion cells (PCC; Figure 4F). Regulation of HTA6-GFP expression by the promoters of the SCARECROW (SCR), and SHORTROOT (SHR) genes resulted in a restriction of GFP fluorescence (Figures 4D and 4E) respectively to nuclei of the endodermis, the cortex/endodermal initials, and the quiescent center, and to nuclei of the stele (the pericycle and internal vascular tissue) [8,9]. Regulation of expression by the promoter of a gene encoding protein 16B of the large ribosomal subunit resulted in nuclear fluorescence more generally localized to the meristematic region (Figure 4C). Flow cytometric analysis of homogenates indicated that PCC and the stele exclusively contained 2C and 4C nuclei, as did the cells within the meristem. In contrast, endodermal cells predominantly contained 4C and 8C nuclei (Figure 7).
Figure 7 Quantification of the proportions of nuclei within the various C-value classes for transgenic plants expressing HTA6-GFP under the control of constitutive and cell type-specific promoters. The proportions were determined by integrating the total number of events within the biparametric plots of Figure 4 falling within a rectangular box containing the specific C-value class and having a GFP-intensity value of 18 or greater. Abbreviations: See Legend to Figure 4.
To explore whether the occurrence of 4C and 8C nuclei was directly correlated with SCR expression, we examined the distribution of C-values of GFP-positive nuclei within transgenic plants producing supernumerary endodermal cell layers, as a consequence of ectopic expression of SHR under the control of the SCR promoter [10]. These transgenic plants contained various proportions of such supernumerary cells, clearly identified by the presence of nuclear GFP (Figure 8A). The proportion of GFP-positive 4C nuclei was dramatically elevated as compared to the wild-type control and as compared to the total distribution of nuclei within the transgenic plants (Figure 8B, 8C). In contrast, no differences were seen in the proportions of all nuclei within the various C-value classes when transgenic and wild-type plants were compared.
Figure 8 Production of supernumerary endodermal cell layers is associated with accumulation of 4C and depletion of 2C cells. Supernumerary endodermal cell layers were produced by transgenic expression of SHR under the control of the SCR promoter, and these layers were identified via expression of HTA6-GFP under the control of the SCR promoter. A. Confocal analysis of the transgenic plants. B. C-value distributions of GFP-positive and -negative nuclei as determined through biparametric flow analysis (axis designations as described in the Legend to Figure 4). C. Classification of the proportions of nuclei within the various C-value classes for non-transgenic controls (WT), for all nuclei within transgenic plants producing supernumerary endodermal cells (SN-SCR), and for GFP-positive nuclei (GFP+) within these same transgenic plants.
Discussion
The described method relies on the targeting of GFP to the nucleus, and its retention within the nucleus during cellular homogenization and flow cytometric analysis. In previous work, we described the use of a tobacco nuclear localization signal to target a chimeric protein comprising the complete coding region of β-glucuronidase fused to GFP [11,12]. Although such a molecule is effectively targeted to the nucleoplasm in vivo, it appears to slowly leak out of nuclei following homogenization. This is not an issue for the flow cytometric analysis of nuclei having large DNA contents, such as tobacco, since the nuclear GFP signal remains well above the background detection level of the flow cytometer for reasonable periods of time following homogenization [12]. In contrast, for plants having small nuclear genome sizes, such as A. thaliana, the small size of the nuclei and the low amplitude of the GFP-signal, coupled to leakage of the targeted molecules, means that nucleoplasmic targeting is unsuitable for flow cytometric analysis of isolated nuclei. This problem can be avoided by employing as the targeting signal a nuclear protein that represents a structural component of the nucleus, in this case histone HTA6. Constitutive transgenic expression of the HTA6-GFP fusion protein has no detectable effect upon plant growth or development. Interestingly, constitutive expression of the GFP-HTA6 (i.e. in a reversed orientation) fusion protein also had no perceptible effects on growth and development (data not shown). This is consistent with our understanding of the three-dimensional structures of histones [13]. For HTA6, both the N- and C-termini are exposed at the nucleosomal surface and, of the 13 predicted A. thaliana H2A proteins (HTA; ), HTA6 has the second longest N-terminus (14 aa), and the longest C-terminus (21 aa).
The patterns observed within the two-dimensional frequency distributions produced by flow cytometry indicate that a majority of the root cells of transgenic plants contain green fluorescent nuclei, which confirms that the CaMV 35S promoter is active during the development of the different cell types present within the region analyzed [14]. The fact that HTA6-GFP fluorescence scales linearly and very precisely with DNA content implies the accumulation of nuclear histone 2A is tightly correlated with DNA content. This observation, coupled to the lack of leakage of HTA6-GFP from the nuclei in homogenates, is consistent with the hypothesis that most of the HTA6-GFP is complexed within chromatin rather than being free within the nucleoplasm.
Although the two-dimensional frequency distributions provide unambiguous identification of the DNA content values of GFP-labelled nuclei, for these to be meaningful, it is crucial that expression of HTA6-GFP not perturb the system under study. As far as we can tell, this appears to be the case: no phenotypical differences were seen between transgenic and wild-type plants, nor were differences seen in the proportions of nuclei within the various C-value categories. It should be noted that the same flow cytometric strategy should be technically applicable to transgenic plants expressing any GFP (or other FP) fusion that is targeted to and retained within the nucleus, with the same caveat that such expression not perturb the system under study.
The observation of cell type-specific patterns of C-value suggests that increasing nuclear DNA content represents one strategy evolved by multicellular organisms to specify cell types. As far as we are aware, this is the most precise experimental evidence supporting this rather simple idea in higher plants, and application of this method to other situations in which increased DNA content is associated with cell type differentiation should rapidly provide a valuable body of data. For example, in other work, it has been hypothesized that basal cells within the xylem pericycle cells arrest in G2, thereby becoming susceptible to auxin-mediated signals that trigger the first formative divisions leading to lateral root initiation [15-18]. Consistent with this hypothesis, genes characteristic of the G2/M boundary are coordinately induced in A. thaliana shortly after imposition of conditions leading to synchronized induction of lateral roots [18].
In both of these situations, cell type specification appears associated with an increase in the proportion of cells containing 4C nuclei. At the cytological level of analysis, such a situation can arise through G2 arrest of cells within a monosomatic diploid cell cycle, or, equally-well, through G1 arrest of cells that have entered the first endoreduplicative cell cycle (i.e. having become tetraploid). Complicating cytological analysis is the potential for formation of polytene chromosomes. Further experiments will evidently be required to clarify the situation, and methods of in situ hybridization utilizing endogenous [19-22] and transgenic chromosomal markers [23] should prove invaluable.
At the molecular level, many candidates for cell cycle regulators responsible for accumulation of 4C cells can be identified from the existing body of knowledge for other eukaryotic organisms [24,25], and that emerging for plants, particularly A. thaliana. These include cyclins that are specifically active during the G2/M transition [26], mitotic cyclin-dependent kinase and its activators and inhibitors [27-34], the anaphase promoting complex (APC) and components that interact with the APC [24,35,36], and molecules associated with check-points relating to cell size [37,38], radiation damage [36,39], and spindle assembly [36].
The mechanisms regulating cell cycle status and nuclear DNA content within the endodermal and cortical cells may also reflect the nature of the SCR and SHR genes, which encode members of the GRAS-STAT family of transcription factors [40,41]. Other members of this family have been shown to interact with regulators of the cell cycle [42-44]. One role of SHR and/or SCR may be to arrest the diploid cell cycle within endodermal cells at the G2/M boundary, and perhaps also to regulate an endoreduplicative event converting these nuclei from a G2 to a G1 state without an intervening M-phase. Of the 30 genes found to be most strongly up- or down-regulated within the endodermis [45], one candidate for a regulatory role is At5g26900, since it exhibits homology to fizzy1 of X. laevis which is required for APC activation in that system [46].
Interestingly, recent reports implicate expression of additional members of the GRAS-STAT family in the establishment of nodulation in Medicago truncatula [47,48]. We note that the flow cytometric method described here should be appropriate for unambiguous determination of the C-value status of root initials responsive to nodulation signals, and of the different cell types that subsequently develop during root nodule formation. If nodule development can be shown to specifically involve G2-arrested cells within the root cortex, this would imply co-option of cellular mechanisms that normally regulate lateral root formation.
In general, the method outlined in this report should be applicable to any transformable plant species within which the regulated expression of HTA6-GFP results in fluorescent nuclei. For promoters of low activity, coupling of nuclear GFP expression to amplification systems (for example provided by GAL4/VP16 [49]) may be required. Orthologues of histones 2A should be readily accessible for most species, and we have established that a GFP fusion to the rice HTA6 orthologue is targeted to the nucleus in transgenic rice plants (CQZ, C. Santhosh Kumar, V. Sundaresan, and DWG, unpublished results). Plant cell types for which a determination of nuclear DNA content should be of particular interest include those undergoing regulated endoreduplicative cycles, such as found within developing seed storage tissues [50], within developing trichomes [51], and in the establishment of symbioses [52], since the method is not restricted to cells operating within a conventional diploid cell cycle. The method should also be helpful in clarifying reports of the unexpected onset of reductive mitoses within endoreplicated cells [27]. It should also be possible to develop multiparametric flow cytometric methods combining the identification of the C-values of nuclei of specific cell types with a determination of the occurrence of S-phase (relying on antibody-based detection of bromodeoxyuridine incorporation [53]). This would allow direct determination of the onset of DNA replication particularly within endoreduplicating cells at various C-value levels, thereby providing a greatly increased degree of sophistication in the analysis of processes of this type. The method should also be applicable to lower plants, and could be readily tested using Physcomitrella, which can be transformed and for which the specific G2-arrest within the chloronema has been recently described [54].
Finally, it should also be noted the flow cytometric method should be equally applicable to transformable non-plant species, including mammals. The relevance of endoreplication to mammalian development, both under normal and abnormal circumstances, is increasingly evident [55,56], and the ability to accurately chart its occurrence within specific cell types should prove important in the analysis of development as well as of specific disease states.
Methods
Recombinant constructions
All general molecular manipulations were done according to standard procedures [57]. PfuUltra™ high-fidelity DNA polymerase (Stratagene, La Jolla, CA, USA) was used for PCR-based amplification of fragments for cloning.
To construct a T-DNA binary vector for expressing GFP in plants, a 2445 bp fragment covering the sGFP expression cassette was released with HincII and SspI from pGFP-JS (Jen Sheen, Massachusetts General Hospital, Boston MA), and inserted into pCAMBIA1302 between the SmaI (9755) and PmlI (752) sites. For convenience of discussion, we call this reassembled vector pCsGFPB. The A. thaliana core histone HTA6 coding sequence (450 bp) was PCR amplified from a cDNA first strand preparation, using forward primer 5'- CATGCCATGGAATCCACCGGAAAAGTG-3' and reverse primer 5'- CATGCCATGGCAGCTTTCTTTGGAGACTTGACTG-3'. The cDNA first strand was prepared using reverse transcriptase SuperScript II according to the manufacturer's recommendations (Invitrogen, Carlsbad, CA, USA). The amplified fragment was inserted into the NcoI site of pCsGFPB. This resulted in in-frame fusion of HTA6 to the N-terminus of GFP, which is downstream of an enhanced CaMV 35S promoter. In the coding region of HTA6, a single nucleotide change at position 118 (G to A) was confirmed by sequencing analysis. This single nucleotide change leads to a point mutation (I39V), and this mutation is retained in all derivative constructs.
Vector pCsGFPB carrying the HTA6 coding sequence was further modified by removing the stop codon (TAA) of the GFP open reading frame as well as the following 14 bp. This modification shifts the contiguous BamHI and XbaI sites to the sGFP open reading frame, and leads to three amino acids (Gly-Ile-Leu) being added to the original sGFP, with the "TAG" within the XbaI site becoming the new stop codon. Then a 70 bp computer generated random sequence (CGAATGTAGTACGTATTCTCCGAACTGAAGCACCTGAGACGTGTAATGTCGGGCCATCTCATACGTACGG) was inserted immediately after the new stop codon, to serve as a transcriptional tag for monitoring the sGFP mRNA level using microarrays printed with the appropriate complementary sequence. Finally, the CaMV 35S promoter (780 bp) upstream of sGFP was excised using EcoRI; and an attR cassette (Frame C, Invitrogen, Carlsbad, CA, USA) was installed. This Gateway-adapted vector was named pCGTAG, and it was used in making the remaining constructs in this study.
Genomic sequences immediately upstream of the start codons of the SCARECROW (SCR), Sulphate Transporter 2-1 (Sultr2-1), SHORTROOT (SHR), and RPL16B coding sequences were PCR amplified. A 2131 bp fragment for SCR was amplified with forward primer 5' CACCGGATAAGGGATAGAGGAAGAGG 3' and reverse primer 5' GGAGATTGAAGGGTTGTTGGTCG 3'. A 2048 bp fragment for Sultr2-1 was amplified with forward primer 5' CACCGCTGACAACTAACACTCCTC-3' and reverse primer 5' CTTCTTCTCGAGTTTTGACGTTGTG 3'. A 2495 bp fragment for SHR was amplified with forward primer 5' CACCGGACAAAGAAGCAGAGCGTGG 3' and reverse primer 5' TTAATGAATAAGAAAATGAATAGAAGAAAGGGAGACCCAC 3'. A 1074 bp fragment for RPL16B was amplified with forward primer 5' CACCTTTCCCACCTCTCTTCAACTTC 3' and reverse primer 5' CGTAAATAGTAAGTTAAATCCCCAAAACGAGGAACG 3'. The five fragments were then cloned into the Gateway entry vector pENTR/D-TOPO (Invitrogen, Carlsbad, CA, USA). The plasmids carrying the regulatory sequences of SCR, Sultr2-1, SHR, and RPL16B were linearized with restriction enzymes EcoRV, EcoRV, EcoNI, DrdI, and EcoRV respectively. These linearized plasmids were employed in the Gateway LR reaction, to insert the individual regulatory sequences into the Gateway-adapted T-DNA binary vector pCGTAG.
Plant transformation
Plasmids carrying the above constructs were introduced into Agrobacterium tumefaciens strain GV3101. A. thaliana 'Columbia' (Col-0) was transformed using the floral dip method [58]. Seeds (T1) were harvested and selected on MS agar plates supplemented with 40 mg L-1 hygromycin and 75 mg L-1 carbenicillin. Roots from hygromycin-resistant seedlings were examined for GFP fluorescence using confocal microscopy. Confirmed transformants were transferred to soil.
Introduction of pSCR-HTA6-GFP into plants carrying supernumerary endodermal cell layers
Transgenic A. thaliana seeds carrying SCRpro::SHR were kindly provided by Philip Benfey (Department of Biology, Duke University). The roots of this transgenic line have an increased number of cell layers which display characteristics of the endodermis [10]. We crossed this transgenic line with transgenic plants carrying pSCR-HTA6-GFP. F1 seeds were germinated on MS plates lacking antibiotics. Roots of three-week old seedlings were subjected to confocal and biparametric flow analyses.
Confocal microscopy
Roots from T1 or T2 seedlings were counterstained with 1 μg ml-1 propidium iodide, PI (Sigma, St Louis, MO, USA) for 1 to 2 minutes, and were placed on slides carrying a drop of water for observation. GFP fluorescence was imaged by confocal microscopy using a MRC 1024 MP (Bio-Rad, Hercules CA) confocal scanner attached to an Olympus BX-50 upright microscope, equipped with UPlanFl 4X/0.13, UPlanFl 10X/0.30, and UPlanApo 20X/0.70 objective lenses. LaserSharp2000 (Bio-Rad) was employed for image acquisition and Photoshop 5.0 (Adobe Systems Inc., San Jose, CA) for image processing.
Growth comparisons
Seeds of Col-0 and of a homozygous transgenic line carrying CaMV35S-HTA6-GFP were sterilized and planted on MS agar plates (2% sucrose, 1.2% agar). The plates were kept at 4°C for three days before transfer to a Conviron growth chamber under 16 hour days / 8 hour nights, with an incident light flux of 150-175 μmol m-2 sec-1, and temperatures of 22°C (day) and 20°C (night). The plates were placed vertically for 3 days. Seedlings of similar sizes were then transferred onto fresh plates, each plate containing ten wild-type and ten transgenic seedlings. The position of the root tip of each seedling was marked on the bottom lid of the plate using a marker pen. Root elongation was measured five days after transfer, and the seedlings were then collected for measurement of fresh weights.
Flow cytometry
Flow cytometry was done using a MoFlo flow cytometer (Dako Cytomation, Fort Collins, CO) equipped with a Coherent Enterprise II laser providing separate beam paths comprising 50 mW at 365 nm and 200 mW at 488 nm. Homogenates of A. thaliana roots were prepared by chopping [5], and nuclei were stained by addition of 2 μg/mL 4',6-diamidino-2-phenylindole (DAPI). Samples were analyzed at an event rate of 200 nuclei/s. GFP fluorescence, excited by the 488 nm beam, was detected following reflection by a 555DCLP dichroic beam splitter through a 530/40 band pass filter. DAPI fluorescence, excited by the 365 nm beam, was routed directly through a 450/65 band pass filter. Biparametric histograms of log DAPI versus log GFP signals were triggered on side scatter and collected to a total count of at least 100,000 events.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CZ produced the constructions and the transgenic plants, characterized the transgenic plants using confocal microscopy, participated in the flow cytometric experiments, and contributed to the preparation of the manuscript. GL assisted in the confocal microscopy, performed the flow cytometric measurements, and contributed to the preparation of the manuscript. FG participated in the flow cytometric measurements, and contributed to the preparation of the manuscript. DG conceived of the study, participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by a grant to DWG from the Plant Genome Program of the National Science Foundation (grant DBI 0211857).
==== Refs
Swift H The constancy of desoxyribose nucleic acid in plant nuclei Proc Natl Acad Sci USA 1950 36 643 654 14808154
Joubès J Chevalier C Endoreduplication in higher plants Plant Molec Biol 2000 43 735 745 11089873 10.1023/A:1006446417196
Galbraith DW Harkins KR Knapp S Systemic endopolyploidy in Arabidopsis thaliana Plant Physiol 1991 96 985 989 16668285
Barow M Meister A Endopolyploidy in seed plants is differently correlated to systematics, organ, life strategy and genome size Plant Cell Envir 2003 26 571 584 10.1046/j.1365-3040.2003.00988.x
Galbraith DW Harkins KR Maddox JR Ayres NM Sharma DP Firoozabady E Rapid flow cytometric analysis of the cell cycle in intact plant tissues Science 1983 220 1049 1051 17754551
Bennett MD Leitch IJ Price HJ Johnston JS Comparisons with Caenorhabditis (similar to 100 Mb) and Drosophila (similar to 175 Mb) using flow cytometry show genome size in Arabidopsis to be similar to 157 Mb and thus similar to 25% larger than the Arabidopsis genome initiative estimate of similar to 125 Mb Ann Bot 2003 91 547 557 12646499 10.1093/aob/mcg057
Takahashi H Watanabe-Takahashi A Smith FW Blake-Kalff M Hawkesford MJ Saito K The roles of three functional sulphate transporters involved in uptake and translocation of sulphate in Arabidopsis thaliana Plant J 2000 23 171 182 10929111 10.1046/j.1365-313x.2000.00768.x
DiLaurenzio L Wysocka-Diller J Malamy JE Pysh L Helariutta Y Freshour G Hahn MG Feldmann KA Benfey PN The SCARECROW gene regulates an asymmetric cell division that is essential for generating the radial organization of the Arabidopsis root Cell 1996 86 423 433 8756724 10.1016/S0092-8674(00)80115-4
Helariutta Y Fukaki H Wysocka-Diller J Nakajima K Jung J Sena G Hauser MT Benfey PN The SHORT-ROOT gene controls radial patterning of the Arabidopsis root through radial signaling Cell 2000 101 555 567 10850497 10.1016/S0092-8674(00)80865-X
Nakajima K Sena G Nawy T Benfey PN Intercellular movement of the putative transcription factor SHR in root patterning Nature 2001 413 307 311 11565032 10.1038/35095061
Grebenok RJ Pierson EA Lambert GM Gong F-C Afonso CL Haldeman-Cahill R Carrington JC Galbraith DW Green-Fluorescent Protein fusions for efficient characterization of nuclear localization signals Plant J 1997 11 573 586 9107043 10.1046/j.1365-313X.1997.11030573.x
Grebenok RJ Lambert GM Galbraith DW Characterization of the targeted nuclear accumulation of GFP within the cells of transgenic plants Plant J 1997 12 685 696 10.1046/j.1365-313X.1997.00685.x
Wolffe AP Hayes JJ Chromatin disruption and modification Nucleic Acids Res 1999 27 711 720 9889264 10.1093/nar/27.3.711
Benfey PN Chua NH The Cauliflower Mosaic Virus-35S promoter – combinatorial regulation of transcription in plants Science 1990 250 959 966 17746920
Beeckman T Burssens S Inze D The peri-cell-cycle in Arabidopsis J Exp Bot 2001 52 403 411 11326046
Bhalerao R Eklöf J Ljung K Marchant A Bennett M Sandberg G Shoot-derived auxin is essential for early lateral root emergence in Arabidopsis seedlings Plant J 2002 29 325 332 11844109 10.1046/j.0960-7412.2001.01217.x
Himanen K Boucheron E Vanneste S de Almeida Engler J Inzé D Beeckman T Auxin-mediated cell cycle activation during early lateral root initiation Plant Cell 2002 14 2339 2351 12368490 10.1105/tpc.004960
Himanen K Vuylsteke M Vanneste S Vercruysse S Boucheron E Alard P Chriqui D Van Montagu M Inzé D Beeckman T Transcript profiling of early lateral root initiation Proc Natl Acad Sci USA 2004 101 5146 5151 15051881 10.1073/pnas.0308702101
Kato A Vega JM Han FP Lamb JC Birchler JA Advances in plant chromosome identification and cytogenetic techniques Curr Opin Plant Biol 2005 8 148 154 15752994 10.1016/j.pbi.2005.01.014
Lysak MA Fransz PF Ali HBM Schubert I Chromosome painting in Arabidopsis thaliana Plant J 2001 28 689 697 11851915 10.1046/j.1365-313x.2001.01194.x
Lysak MA Pecinka A Schubert I Recent progress in chromosome painting of Arabidopsis and related species Chromosome Res 2003 11 195 204 12769287 10.1023/A:1022879608152
Koorneef M Fransz P de Jong H Cytogenetic tools for Arabidopsis thaliana Chromosome Res 2003 11 183 194 12769286 10.1023/A:1022827624082
Kato N Lam E Detection of chromosome tagged with green fluorescent protein in live Arabidopsis thaliana plants Genome Biology 2001 2 1 10 Online ISSN 165–6914 10.1186/gb-2001-2-11-research0045
Murray AW Recycling the cell cycle: cyclins revisited Cell 2004 116 221 234 14744433 10.1016/S0092-8674(03)01080-8
O'Connell MJ Cimprich KA G2 damage checkpoints: what is the turn-on? J Cell Sci 2005 118 1 6 15615778 10.1242/jcs.01626
Inze D Green light for the cell cycle EMBO J 2005 24 657 662 15678103 10.1038/sj.emboj.7600561
Weinl C Marquardt S Kuijt SJH Nowack MK Jakoby MJ Hülskamp M Schnittger A Novel functions of plant cyclin-dependent kinase inhibitors, ICK1/KRP1, can act non-cell-autonomously and inhibit entry into mitosis Plant Cell 2005 17 1704 1722 15749764 10.1105/tpc.104.030486
Sherr CJ Roberts JM CDK inhibitors: positive and negative regulators of G1-phase progression Genes Dev 1999 13 1501 1512 10385618
Verkest A Manes CLD Vercruysse S Maes S Van der Schueren E Beeckman T Genschik P Kuiper M Inzé D De Veylder L The cyclin-dependent kinase inhibitor KRP2 controls the onset of the endoreduplication cycle during Arabidopsis leaf development through inhibition of mitotic CDKA;1 kinase complexes Plant Cell 2005 17 1723 1736 15863515 10.1105/tpc.105.032383
Wang H Fowke LC Crosby WL A plant cyclin-dependent kinase inhibitor gene Nature 1997 386 451 452 9087400 10.1038/386451a0
Wang H Qi Q Schorr P Cutler AJ Crosby WL Fowke LC ICK1, a cyclin-dependent protein kinase inhibitor from Arabidopsis thaliana interacts with both Cdc2a and CycD3, and its expression is induced by abscisic acid Plant J 1998 15 501 510 9753775 10.1046/j.1365-313X.1998.00231.x
Wang H Zhou Y Gilmer S Whitwill S Fowke LC Expression of the plant cyclin-dependent kinase inhibitor ICK1 affects cell division, plant growth and morphology Plant J 2000 24 613 623 11123800 10.1046/j.1365-313x.2000.00899.x
De Veylder L Beeckman T Beemster GT Krols L Terras F Landrieu I van der Schueren E Maes S Naudts M Inzé D Functional analysis of cyclin-dependent kinase inhibitors of Arabidopsis Plant Cell 2001 13 1653 1668 11449057 10.1105/tpc.13.7.1653
Jasinski S Riou-Khamlichi C Roche O Perennes C Bergounioux C Glab N The CDK inhibitor NtKIS1a is involved in plant development, endoreduplication and restores normal development of cyclin D3;1-overexpressing plants J Cell Sci 2002 115 973 982 11870216
Harper JW Burton JL Solomon MJ The anaphase-promoting complex: it's not just for mitosis any more Genes Dev 2002 16 2179 2206 12208841 10.1101/gad.1013102
Vodermaier HC APC/C and SCF: controlling each other and the cell cycle Curr Biol 2004 14 R787 R796 15380093 10.1016/j.cub.2004.09.020
Kellogg DR Wee1-dependent mechanisms required for coordination of cell growth and cell division J Cell Sci 2003 116 4883 4890 14625382 10.1242/jcs.00908
Sun YJ Dilkes BP Zhang CS Dante RA Carneiro NP Lowe KS Jung R Gordon-Kamm WJ Larkins BA Characterization of maize (Zea mays L.) Wee1 and its activity in developing endosperm Proc Natl Acad Sci USA 1999 96 4180 4185 10097184 10.1073/pnas.96.7.4180
Culligan K Tissier A Britt A ATR regulates a G2-phase cell-cycle checkpoint in Arabidopsis thaliana Plant Cell 2004 16 1091 1104 15075397 10.1105/tpc.018903
Richards DE Peng J Harberd NP Plant GRAS and metazoan STATs: one family? Bioessays 2000 22 573 577 10842311 10.1002/(SICI)1521-1878(200006)22:6<573::AID-BIES10>3.0.CO;2-H
Bolle C The role of GRAS proteins in plant signal transduction and development Planta 2004 218 683 692 14760535 10.1007/s00425-004-1203-z
Coqueret O Gascan H Functional interaction of STAT3 transcription factor with the cell cycle inhibitor p21(WAF1/CIP1/SD11) J Biol Chem 2000 275 18794 18800 10764767 10.1074/jbc.M001601200
Bromberg J Darnell JE Jr The role of STATs in transcriptional control and their impact on cellular function Oncogene 2000 19 2468 2473 10851045 10.1038/sj.onc.1203476
Calo V Migliavacca M Bazan V Macaluso M Buscemi M Gebbia N Russo A STAT proteins: from normal control of cellular events to tumorigenesis J Cell Physiol 2003 197 157 168 14502555 10.1002/jcp.10364
Birnbaum K Shasha DE Wang JY Jung JW Lambert GM Galbraith DW Benfey PN A gene expression map of the Arabidopsis root Science 2003 302 1956 1960 14671301 10.1126/science.1090022
Lorca T Castro A Martinez AM Vigneron S Morin N Sigrist S Lehner C Doree M Labbe JC Fizzy is required for activation of the APC/cyclosome in Xenopus egg extracts EMBO J 1998 17 3565 3575 9649427 10.1093/emboj/17.13.3565
Kaló P Gleason C Edwards A Marsh J Mitra RM Hirsch S Jakab J Sims S Long SR Rogers J Kiss GB Downie JA Oldroyd GED Nodulation signaling in legumes requires NSP2, a member of the GRAS family of transcriptional regulators Science 2005 308 1786 1789 15961668 10.1126/science.1110951
Smit P Raedts J Portyanko V Debellé F Gough C Bisseling T Geurts R NSP1 of the GRAS protein family is essential for rhizobial Nod Factor-induced transcription Science 2005 308 1789 1791 15961669 10.1126/science.1111025
Johnson AA Hibberd JM Gay C Essah PA Haseloff J Tester M Guiderdoni E Spatial control of transgene expression in rice (Oryza sativa L.) using the GAL4 enhancer trapping system Plant J 2005 41 779 789 15703064 10.1111/j.1365-313X.2005.02339.x
Grafi G Larkins BA Endoreduplication in maize endosperm: Involvement of M phase-promoting factor inhibition and induction of S phase-related kinases Science 1995 269 1262 1264 17732113
Marks MD Molecular genetic analysis of trichome development in Arabidopsis Ann Rev Plant Physiol Plant Molec Biol 1997 48 137 163 15012260 10.1146/annurev.arplant.48.1.137
Patriarca EJ Tatè R Ferraioli S Iaccarino M Organogenesis of legume root nodules Intl Rev Cytol 2004 234 201 262
Lucretti S Nardi L Nisini PT Moretti F Gualberti G Doležel J Bivariate flow cytometry DNA/BrdUrd analysis of plant cell cycle Methods Cell Sci 1999 21 155 166 10728648 10.1023/A:1009893008892
Schween G Gorr G Hohe A Reski R Unique tissue-specific cell cycle in Physcomitrella Plant Biol 2003 5 50 58 10.1055/s-2003-37984
Keighren MA Macfadyen LP Hill AS Patek CE Telfer EE West JD Polyploid cells in the mouse ovary J Anat 2003 202 563 571 12846477 10.1046/j.1469-7580.2003.00189.x
Lilly MA Duronio RJ New insights into cell cycle control from Drosophila endocycle Oncogene 2005 24 2765 2775 15838513 10.1038/sj.onc.1208610
Sambrook J Fritsch EF Maniatis T Molecular Cloning: a Laboratory Manual 1989 2 Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press
Clough SJ Bent AF Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana Plant J 1998 16 735 743 10069079 10.1046/j.1365-313x.1998.00343.x
Macas J Lambert GM Doležel D Galbraith DW Nuclear expressed sequence tag (NEST) analysis: A novel means to study transcription through amplification of nuclear RNA Cytometry 1998 33 460 468 9845441 10.1002/(SICI)1097-0320(19981201)33:4<460::AID-CYTO10>3.0.CO;2-#
|
16270943
|
PMC1277020
|
CC BY
|
2021-01-04 16:37:28
|
no
|
Plant Methods. 2005 Oct 4; 1:7
|
utf-8
|
Plant Methods
| 2,005 |
10.1186/1746-4811-1-7
|
oa_comm
|
==== Front
J Ethnobiol EthnomedJournal of Ethnobiology and Ethnomedicine1746-4269BioMed Central London 1746-4269-1-41627093010.1186/1746-4269-1-4ResearchUrinary diseases and ethnobotany among pastoral nomads in the Middle East Abu-Rabia Aref [email protected] Department of Middle East Studies, Ben-Gurion University, Beer-Sheva, 84105, Israel2005 2 8 2005 1 4 4 9 7 2005 2 8 2005 Copyright © 2005 Abu-Rabia; licensee BioMed Central Ltd.2005Abu-Rabia; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This article is derived from a broad, twenty-year study of ethnobotany and folk medicine among pastoral nomads in the Middle East which took place from 1984 to 2004. The article presents examples of different treatments of diseases and disorders of the urinary tract carried out by healer herbalists. The preparation of remedies includes boiling infusions, extraction of dry or fresh leaves, flowers, seeds or whole plants. Some of these plants were used both as food and as medicine, by ingesting different parts of the plants, such as leaves, flowers, fruits, and so on, either while soft, cooked or dried. Data were collected by using unstructured interviews and by observation. These plants were identified by healers, patients, and university botanists. This paper identified eighty-five plant species, which belong to thirty-six families. The most representative families are: Asteraceae (8), Brassicaceae (6), Poaceae (6), Umbelliferae (6).
==== Body
Introduction
People have been using traditional medicine including ethno-botany for several thousand years. Ancient Arabic medicine was influenced by the ancient medicinal practices of Mesopotamia, Greece, Rome, Persia and India. The Greco-Roman system of medicine was developed based primarily on the writings of Hippocrates (460-360 B.C.), Dioscorides (circa 54 to 68 AD) and Galen (130–201 AD). A combination of political and religious factors caused many Greek and Syriac-speaking scholars to move eastward to Persia and to establish centers of learning there. The city of Gundishapur in southwest Iran also became a center of learning, with a well-known medical school, in the sixth century AD [1,2]. One of the Arab physicians during the time of the Prophet Muhammad (571–632 AD) was al-Harith ibn Kalada (d. 634), one of the most prominent physicians of his time, who traveled to Gundishapur in Persia and studied medicine prior to the establishment of Islam. Another renowned Arab physician was Ibn Abi Rimtha. The sayings (Hadith) of the Prophet Muhammad on health and illness were systemized and became known as The Medicine of the Prophet (al-Tibb al-Nabawi) [3,2]. During the Umayyad rule (from 661–750 in the East, based in Damascus), many ancient medical works began to be translated. For five centuries (750–1258) the Abbasids, based in Baghdad, dominated the socio-political life of the greater part of the Muslim world. Countless manuscripts, particularly those written in Greek, were collected and stored in Bayt al-hikmah (The House of Wisdom, established in 830, by the Caliph al-Ma'mun), where scholars worked to translate them into Arabic [4,5].
Within a century, Muslim physicians and scientists were writing original contributions to medical and botanical knowledge. One of the greatest and most famous Islamic doctors was Ibn Sina (Avicenna 980–1037), author of The Canon of Medicine (Kitab al-Qanun fi al-Tibb), the epitome of Islamic medicine. This work is the culmination and masterpiece of the Arab systematization of medical science, and includes many descriptions of the uses of medicinal plants [6]. Other Arabic philosopher-physicians were al-Razi (Rhazes 865–923) who wrote The Comprehensive Book on Medicine (Kitab al-Hawi fi al-Tibb). The material written by al-Hawi is arranged under headings of different diseases, with separate sections on pharmacological topics. Ibn Sina's and al-Razi's works were later translated into Latin, and continued to influence medical science well into the nineteenth century [7-9].
In the western part of the Islamic empire, the Umayyads of Andalus (Islamic Spain) made their capital at Cordoba. Areas of Cordoba and Granada became centers of learning. The richness and diversity of the flora of Spain were major contributing factors to the development of medical botany. The majority of physicians were herbalists and vice versa. The physician Ibn al-Baytar (1197–1248), authored The Compendium of Simple Drugs and Food (al-jami' li-mufradat al-adwiya wa'l-aghdhiya), in which he described more than 1400 medicinal drugs, 300 of which had not previously been described, recording them alphabetically and discussing them with great clarity and detail. The work specified the names of herbs and remedies in various languages, thus providing a first class tool for the comparative research of medicinal plants. Other well-known physicians who wrote on plant uses were: Ibn Juljul, al-Ghafiqi, Ibn Bajjah, Ibn Samajun, and Abu'l-Hassan al-Andalusi [10,7]. Traditional medical information grounded in the Arab medicine of the Middle Ages was gradually transferred to traditional healers and to the general public [11]. The use of herbal medicine is still widespread throughout the populations of the Middle East, including the pastoral nomadic tribes [12-19].
Among the pastoral Bedouin, hundreds of species of trees and shrubs are employed as analgesics, astringent, diuretics, emetics, purgatives, poultices, salves, and tonics. Some of these herbs are aimed at cleansing the pastoralist's body of polluting influences, bad spirits, jinns, and the negative effects of sorcery and/or witchcraft. The pastoral nomadic tribes depend on their local healers and traditional medicine as recorded in Table 1 (see Additional file 1).
Methodology
The data for this paper are derived from a broad twenty-year study of ethnobotany and folk medicine among the pastoral nomadic Bedouin tribes in the Negev, Jordan and Sinai deserts, carried out from (1984–2004). The paper is based on interviews with healers and patients. All the material was recorded in field logs, and some was tape-recorded. Unstructured interviews and the observation of participants were carried out in the informants' homes (120 men and 120 women), as well as in the homes of traditional healers (15 men and 10 women). Most of the healers were in the age range of forty to eighty. All the informants were married and over thirty. There were five males from each desert, and four female healers from the Negev, three from Sinai and three from Jordan. The informants were divided into two groups of forty men and forty women from each desert. The collected information was used to construct a list of the indigenous ethnobotanic medicine. Samples from all the plants were collected and identified by healers, patients and university botanists.
Results and Discussion
This paper describes the treatment of diseases and disorders of the urinary tract by traditional herbalists among the pastoral nomadic Bedouin tribes in the Middle East. In this study, we identified eighty-five plant species, which belong to thirty-six families.
The use of traditional medicine by the pastoral nomads, and the appeal to traditional healers over the course of many centuries established a psychological-therapeutic dependence of the pastoral nomadic tribes upon these healers. The rich variety of approaches employed by pastoral nomadic healers to treat disorders and diseases of the urinary tract is indicative of the depth and breadth of indigenous medicine practiced among the pastoral nomads in the twentieth century. The analysis of my collected data, together with the information extracted from the literature on herbal and ethnobotanic medicine of countries in the Middle East [14,20,15,21,18], yielded Table 1 (see Additional file 1). This table includes eighty-five plants with medicinal potential which have been used among the pastoral nomadic Bedouin tribes in the Middle East from generation to generation as reported by my informants.
Table 1 presents information on which parts of the plants are used and in what manner. It should be noted that for some plants, the uses in different countries of the Middle East are similar [22,16]. However, dissimilar uses were also observed for certain plants in different countries/tribes in the Middle East [23,15,19,24]. The important information gathered in this study will help to preserve the heritage and knowledge of ethnobotanic and folk medicine of the indigenous pastoral tribes of the Middle East. This study will generate awareness in the region concerning the potential for conserving plant resources in medicine, food, nutrition and folk heritage. It is of the utmost importance to preserve this heritage, which relates to the traditional, economic and medicinal uses of available plant resources in the countries of the Middle East.
The many medicinal substances which we were able to identify as used in traditional medicine included various plants species. The analysis of the findings shows that the three deserts where I conducted my research served as the geographic origin of the medicinal substances. These plants were available because they grew as wild and cultivated plants and were part of the natural flora of these deserts. The pastoral nomads used these plants as food and as medicine, by eating different parts of the plants including the following: leaves, flowers, barks, stems, stalks, roots, rhizomes, bulbs, pith, fruit, corms, inflorescenes, shells, berries, seeds, stones/pits (in fruit), soft seed pods, buds, and shoots.
It should be noted that wild desert plants also contain a host of other biologically active compounds besides nutrients. The physiological effects of these other compounds in relation to plant nutrients are not well known, but could affect nutrient and medical utilization or other functions. These topics are of relevance for future research in terms of improving our understanding of human nutritional and medical requirements of the pastoral nomads in the Middle East.
Supplementary Material
Additional file 1
Table 1: Urinary Diseases and Ethnobotany among Pastoral Nomads in the middle East
Click here for file
Acknowledgements
Part of my research was supported by a grant from the British Council in Tel-Aviv; The Faculty of Humanities and Social Sciences at the Ben-Gurion University of the Negev; The Natural Medicine Research Unit at Hadassah Jerusalem; The National Council for Research and Development in the Ministry of Science. To all of these I owe a debt of gratitude. I am indebted to Prof. Clinton Bailey and Dr. Lauren Basson for their reading, advice and comments. Special thanks to Prof. Allan Witztum, Nissim Krispil and Yusef Or, for identifying the plants. Finally, my great thanks go to the pastoral nomadic healers and the many people I have not mentioned individually by name, who kindly provided me with invaluable information, without whose help and kindness I would not have managed to conduct this research.
==== Refs
Mursi 'Arab: Dirasat fi'l-Shi'un al-tibbiyah al-'arabiyah = Studies on Arab Medical Affairs Al-Iskandariya: al-Ma'arif (Arabic) 1966 7 12
Savage-Smith Emilie Roshdi Rashid, Regis Morelon Medicine Encyclopedia of the History of Arab Science 1996 3 London and New York: Routledge 907 908
Hawting GR Bosworth CE, Charles Issawi, Roger Savory, Udovitch AL Princeton The Development of the Biography of al-Harith ibn Kalada and the Relationship between Medicine and Islam The Islamic World, from Classical to Modern Times 1989 New Jersey: The Darwin Press, Inc 127 137
Hitti Philip History of the Arabs, From the Earliest Times to the Present 1952 London: Macmillan & Co 311 312
Ullmann Manfred Islamic Surveys: Islamic Medicine 1978 Edinburgh: University Press 7 40
Al-Said Mansour Solyman Helaine Selin Medicine in Islam Encyclopedia of the History of Science, Technology, and Medicine in Non-Western Cultures 1997 Dordrect/Boston/ London: Klewer Academic Publishers 695 698
Johnstone Penelope Penelope Johnstone Prophet Medicine Ibn Qayyim al-Jawziyya Medicine of the Prophet 1998 Cambridge: The Islamic Texts Society xxx xxxii
Murad Amnah Lamhat min tarikh al-tibb al-qadim = Glimpses From the History of Early Medicine. al-Qahira Maktabat al-Nasr al-Haditha (Arabic) 1966 1 32
Al-Shatti Ahmad al-'Arab wa-al-tibb = The Arabs and Medicine Dimashq: Manshurat Wazarat al-Thaqafa (Arabic) 1970 12 65
Al-Najjar Amer Fi Tarikh al-Ttib fi al-Dawlah al-Islamyya: History of Medicine in the Islamic Empire Al-Qahira: Dar al-Ma'aref (Arabic) 1994 159 229
Lev Efraim Zohar Amar Ethnopharmacological survey of traditional drugs sold in Israel at the end of the 20th century Journal of Ethopharmacology 2000 72 191 205 10.1016/S0378-8741(00)00230-0
Abu-Khusa Ahmad Beer al-Saba' wal-Hayah al-Badawiya Amman: Matabi' al-Mu'assasa al-Sahafiya al-Urdunniya (Arabic) 1976 70–71 106 113
Abu-Rabia Aref Traditional Bedouin Medicine 1999 Tel-Aviv: Mod Publishing, (Hebrew) 35 92
Bailey Clinton Danin A Bedouin plant utilization in Sinai and the Negev Economic Botany 1981 35 145 162
Jabbur Jibrail The Bedouins and the Desert: Aspects of Nomadic Life in the Arab East 1995 Albany: State University of New York Press 64 81
Palevitch Dan Zohara Yaniv Medicinal Plants of the Holy Land 2000 Tel-Aviv: Modan Publishing House 9 88
Pillsbury Barbara Traditional Health Care in the Near East 1978 A report prepared for the U.S. Agency for International Development, Washington D.C 1 25
Shuqayr N Tarikh Sina Cairo: Matba'at al-Ma'arif (Arabic) 1916 340 375
Tal Pnina Medicinal Plants 1981 Tel-Aviv: Reshafim (Hebrew) 15 17
Burckhardt John Lewis Travels in Syria and the Holy Land 1992 London: Darf Publishers Ltd 12 97
Musil Alois The Manners and Customs of the Rwala Bedouins 1928 New York: American Geographical Society 90 98
Boulos Loutfy Medicinal Plants of North Africa 1983 Algonac: Reference Publications, Inc 23 67
Ghazanfar Shahina Handbook of Arabian Medicinal Plants 1994 London: CRC Press 44 87
Abu-Rabia Aref Bedouin Health Services in Mandated Palestine Middle Eastern Studies 2005 41 421 429
|
16270930
|
PMC1277084
|
CC BY
|
2021-01-04 16:40:12
|
no
|
J Ethnobiol Ethnomed. 2005 Aug 2; 1:4
|
utf-8
|
J Ethnobiol Ethnomed
| 2,005 |
10.1186/1746-4269-1-4
|
oa_comm
|
==== Front
J Ethnobiol EthnomedJournal of Ethnobiology and Ethnomedicine1746-4269BioMed Central London 1746-4269-1-51627093110.1186/1746-4269-1-5ReviewWhy study the use of animal products in traditional medicines? Alves Rômulo RN [email protected] Ierecê L [email protected] Departamento de Biologia, Universidade Estadual da Paraíba and Programa de Pós-Graduação em Ciências Biológicas (Zoologia), Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba, 58059-970 João Pessoa, PB, Brasil2 Departamento de Sistemática e Ecologia, CCEN, Universidade Federal da Paraíba, 58059-900 João Pessoa, PB, Brazil2005 30 8 2005 1 5 5 10 7 2005 30 8 2005 Copyright © 2005 Alves and Rosa; licensee BioMed Central Ltd.2005Alves and Rosa; 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 World Health Organization (WHO) estimates that as many as 80% of the world's more than six billion people rely primarily on animal and plant-based medicines. The healing of human ailments by using therapeutics based on medicines obtained from animals or ultimately derived from them is known as zootherapy. The phenomenon of zootherapy is marked both by a broad geographical distribution and very deep historical origins. Despite their importance, studies on the therapeutic use of animals and animal parts have been neglected, when compared to plants. This paper discusses some related aspects of the use of animals or parts thereof as medicines, and their implications for ecology, culture (the traditional knowledge), economy, and public health.
==== Body
Introduction
The World Health Organization (WHO) estimates that as many as 80% of the world's more than six billion people rely primarily on animal and plant-based medicines [1]. Traditional human populations have a broad natural pharmacopoeia consisting of wild plant and animal species. Ingredients sourced from wild plants and animals are not only used in traditional medicines, but are also increasingly valued as raw materials in the preparation of modern medicines and herbal preparations [2].
Animals and products derived from different organs of their bodies have constituted part of the inventory of medicinal substances used in various cultures since ancient times [3,6,7]; such uses still exist in traditional medicine. The healing of human ailments by using therapeutics based on medicines obtained from animals or ultimately derived from them is known as zootherapy [4]. As Marques [5] states, "all human culture which presents a structured medical system will utilize animals as medicines". The phenomenon of zootherapy is marked both by a broad geographical distribution and very deep historical origins.
In modern societies, zootherapy constitutes an important alternative among many other known therapies practiced worldwide. Wild and domestic animals and their by-products (e.g., hooves, skins, bones, feathers, tusks) form important ingredients in the preparation of curative, protective and preventive medicine [6,7]. For example, in Traditional Chinese Medicine (TCM), more than 1500 animal species have been recorded to be of some medicinal use [8]. In India nearly 15–20 percent of the Ayurvedic medicine is based on animal-derived substances [9]. In Bahia State, in the northeast of Brazil, over 180 medicinal animals have been recorded [10].
There are many reasons why studies on the use of animals, integrally or in parts, as medicines and their implications should be carried out and recorded. Among several approaches to be considered, this paper briefly discusses those concerning with the ecological, cultural (traditional knowledge), economical, and sanitary aspects of zootherapy.
Ecological Approach
The world is facing potentially massive loss of wildlife due to over-hunting [11-13] and overfishing [14-17]. Transformation of ecosystems wrought through economic activities has been putting severe constraints on the availability and accessibility of specific types of plant and animal species used for medicinal purposes [18].
Regrettably, the demand created by traditional medicine is one of the causes of the overexploitation of the wild population of numerous animal species [19]. The use of animals in popular medicine certainly provokes pressure on natural resources exploited through traditional forms of collection, mainly due to general acceptance of popular medicine [20]. Medically speaking, the one major negative consequence of this trend is that there will be essentially less choice for the future development of medicines [19]. At present, about 40% of all prescription drugs are substances originally extracted from plants, animals, fungi and microorganisms [21].
Natural resource users may be the first to observe depletion [22]. However, as traditional peoples are integrated into the global economy, and come under trade, acculturation and population pressures, they lose their attachment to their own restricted resource catchments. This may lead to a loss of motivation in sustainable uses of a diversity of local resources, along with the pertinent indigenous knowledge [23].
Traditional ecological knowledge is of significance from a conservation perspective and an attribute of societies with continuity in resource use practice [23], therefore the dissociation of traditional knowledge from managerial ecology may result in the adoption of inadequate management options. Holders of traditional knowledge not only have a role as natural resource managers, but can also provide a model for biodiversity policies.
There is a need to shift the focus from how to obtain the greatest amount of zootherapeutical resources to how to ensure future uses. There is also a need for a transdisciplinary approach to integrate the various aspects of zootherapy in such a way that frameworks or methods to amalgamate ecological and social components of that practice can be increasingly tested. In this context, it is important not only to document the traditional uses of animal species, but also to integrate the cultural and biological aspects of such pratices into a broader discourse encompassing conservation, cooperative management, and sustainability.
Cultural Approach
Human communities with historical practices of using resources acquire information of the ecosystem, process and local fauna and flora properties called Ecological Knowledge, which may be traditional, local or recently acquired [24-26]. Medicinal animals are important resources linking people to the environment and their use promotes the traditional lore related to them.
There is an increased interest in the knowledge that traditional populations possess on the use of animals for medicinal purposes, partly because the empirical basis developed throughout centuries may have, in many cases, scientific corroboration; but above all due to the historical, economical, sociological, anthropological, and environmental aspects of such a practice [3].
For centuries, healers and indigenous people have been collecting medicines from local plants and animals without threatening the population dynamics of the species because of the low level of harvesting. Loss of traditional knowledge has impact on the development of modern medicine. Medicinal folklore over the years has proved to be an invaluable guide in present day to the screening of important modern drugs (e. g., digitoxin, reserpine, tubocurarine, ephedrine, to name a few) that have been discovered by following leads from folk uses [18]. In view of this, it is evident the need to document the traditional knowledge of human communities, mainly because the majority of such communities are rapidly losing their socioeconomic and cultural characteristics.
The importance of protecting traditional knowledge and its cultural environmental resources is crucial, particularly in the context of globalisation and increased demand on natural resources worldwide. Traditional knowledge is valuable not only to those directly involved with it, but also to modern medicine and agriculture, among others. Moreover, protection of traditional knowledge can be used to raise the profile of the knowledge and its custodians. This not only has implications for the continuation of traditional practices within communities, but also for the interactions (e.g., economic, ecological) established outside the communities.
Economical Approach
The value of biodiversity to human health has been highlighted in literature [27]. The most obvious benefit is the large proportion of the pharmaceutical armamentarium that is derived from the natural world. Over 50% of commercially available drugs are based on bioactive compounds extracted (or patterned) from non-human species [28]. Almost every class of drug includes a model structure derived from nature, exhibiting the classical effects of the specific pharmacological category. A great number of these natural products have come to us from the scientific study of remedies traditionally employed by various cultures [29]. In addition to plants and microbes, there has been increasing attention paid to animals, both vertebrates and invertebrates, as sources for new medicines. Animals have been methodically tested by pharmaceutical companies as sources of drugs for modern medical science [30], and the current percentage of animal sources for producing essential medicines is quite significant. Of the 252 essential chemicals that have been selected by the World Health Organization, 11.1% come from plants, and 8.7% from animals [31]. And of the 150 prescription drugs currently in use in the United States of America, 27 have animal origin [32].
Underlying the debate over traditional knowledge may be a much bigger issue such as the position of indigenous communities within the wider economy and society of the country in which they reside, and their access to or ownership of land they have traditionally inhabited. In that sense, concerns about the preservation of traditional knowledge, and the continued way of life of those holding such knowledge, may be symptomatic of the underlying problems that face these communities in the face of external pressures [33]
The trade in wildlife body parts and products includes traditional medicine, and it is well known that the annual global trade in animal-based medicinal products accounts for billions of dollars per year [31]. Nevertheless, in countries such as Brazil, the trade of animals for medicinal purposes has had little impact on the socioeconomic conditions of collectors, who generally are illiterate, underpaid, and perceive their activity as clandestine or semi-clandestine. The monetary value of animals sold for medicinal purposes in the country increases at each level of trade, and the socioeconomic profile of traders varies accordingly (I. L. Rosa and R. R. N. Alves, unpublished data).
Additionally, there is a need to assure that custodians of traditional knowledge receive fair compensation if the traditional knowledge leads to commercial gain, and to prevent appropriation of traditional knowledge by unauthorized parties [33].
Sanitary Approach
Traditional drugs and traditional medicine in general represent a still poorly explored field of research in terms of therapeutic potential or clinical evaluation. There is a current preoccupation about this, since it is well-established that all sorts of vegetable, animal and mineral remedies used in a traditional setting are capable of producing serious adverse reactions. It is essential, however, that traditional drug therapies be submitted to an appropriate benefit/risk analysis [34]. Unfortunately, little research has been done so far to prove the claimed clinical efficacy of animal products for medicinal purposes [19].
Numerous infectious diseases can be transmitted from animals to humans (i.e. zoonoses). In this context, the possibility of transmitting infections or ailments from animal preparations to the patient should be seriously considered [19]. Several organs and tissues including bones and bile can be a source of Salmonella infection causing chronic diarrhoea and endotoxic shock. The possibility of transmission of other serious and widespread zoonoses such as tuberculosis or rabies should be considered whenever animal tissues from unknown sources are handled and used as remedies [35]. The possibility of toxic or allergic reactions to animal products should also be considered [36].
Broad categories of sanitary and phytosanitary regulatory measures are recognized for the food trade: 1) information measures which restrict the behaviour of suppliers only to the extent that they are required to disclose specified facts about their products; 2) measures that impose prior approval certifying that their products have met some pre-specified safety criteria before they can be released onto the market and 3) measures that allow suppliers to sell products without any prior official approval but imply that an offence is committed if the products fail to meet certain minimum safety standards [37]
The implementation of equivalent sanitary measures to the trade of animal or their parts for medicinal purposes poses considerable challenges, among them ensuring adequate participation of all stakeholders involved, combating illegal, unreported and unregulated trade, and monitoring of the activity.
Final Considerations
Despite their importance, studies on the therapeutic uses of animals and their body parts have been neglected, when compared to plants [38]. Scholarly investigation of studies on medicinal uses of animals and their products, as well as of inorganic materials, should not be neglected and should be considered as an important complementary body of knowledge [3].
The extensive practice of traditional medicine in developing countries and the rapidly growing demand for alternative and basic therapeutic means (also in industrialized countries) constitute the international relevancy of research and development in the field of traditional drugs [39]. An additional motivation for such activities is found in the practical necessity to integrate the potential of traditional medicine into current practices of modern health care [39]. It is important to emphasize that some traditional medicinal systems, like the Chinese Traditional Medicine, is recognized by the WHO – World Health Organization [40] and accepted by one-fourth of the world human population.
There are numerous reasons to urgently re-think the medicinal use of animal products in traditional medicine both in humans and animals. In doing this, we should particularly take into account the rarity of some species, the unnecessary suffering involved in the harvesting (e.g., hunting, fishing) process, and the possible health risks linked to the administration of the animal-based remedies.
It is important to consider that human health is dependent on biodiversity and on the natural functioning of healthy ecosystems [41]. In this aspect, the use of animals for medicinal purposes is not simply a matter of the pharmaceutical and medical sciences; joint-research programmes should be undertaken with experts in the fields of ecology, linguistics, sociology, anthropology, etc. Thus, discussing zootherapy within the multidimensionality of sustainable development turns out to be one the key elements in achieving the sustenance of medicinal faunistic resources [10]. The use of endangered species in all forms of traditional medicine is a cause of growing concern.
Simultaneously, there is increasing dialogue between the conservation communities and traditional medicine communities globally. Showing respect and communicating in a language understood by all sides are not profound concepts. However, they demand time, money and goodwill [42]. Indigenous peoples have a storehouse of knowledge with regard to raw materials used in a range of products and processes, e.g., in agriculture, medicines, cosmetics, and foodstuffs, their knowledge of ecosystems being crucial to the care and management of biological diversity [43].
A growing respect for traditional knowledge has led modern science to adapt its procedures for assessing the impact of development projects on biological diversity; for monitoring of ecosystems, species, particular genetic resources and species at risk; for controlling alien species; and for promoting the in-situ conservation and sustainable management of biological diversity generally to identify but a few examples [44]. The use of animals for medicinal purposes is part of a body of traditional knowledge which is increasingly becoming more relevant to discussions on conservation biology, public health policies, sustainable management of natural resources, biological prospection, and patents.
==== Refs
WHO/IUCN/WWF Guidelines on Conservation of MedicinalPlants Switzerland 1993
Kang S Phipps M A question of attitude: South Korea's Traditional Medicine Practitioners and Wildlife Conservation 2003 TRAFFIC East Asia, Hong Kong
Lev E Traditional healing with animals (zootherapy): medieval to present-day Levantine practice J Ethno pharmacol 2003 86 107 118
Costa-Neto EM Animal-based medicines: biological prospection and the sustainable use of zootherapeutic resources An Acad Bras Cienc 2005 77 33 43 15692677
Marques JGW A fauna medicinal dos índios Kuna de San Blas (Panamá) e a hipótese da universalidade zooterápica [abstract] Anais da 46a Reunião Anual da SBPC 1994 324
Adeola MO Importance of wild Animals and their parts in the culture, religious festivals, and traditional medicine, of Nigeria Environmental Conservation 1992 19 125 134
Anageletti LR Agrimi U Curia C French D Mariani-Costantini R Healing rituals and sacred serpents Lancet 1992 340 223 225 1353146 10.1016/0140-6736(92)90480-Q
China National Corporation of Traditional and Herbal Medicine Materia Medica Commonly Used in China 1995 Science Press. Beijing
Unnikrishnan PM Animals in Ayurveda Amruth 1998 1 15
Costa-Neto EM Implications and applications of folk zootherapy in the State of Bahia, Northeastern Brazil Sust Dev 2004 12 161 174 10.1002/sd.234
Robinson JG Bennett EL Robinson JG, Bennett EL Carrying capacity limits to sustainable hunting in tropical forests Hunting for Sustainability in Tropical Forests 2000 New York: Columbia University Press 13 30
Robinson JG Bennett EL Will alleviating poverty solve the bushmeat crisis? Oryx 2002 36 332 10.1017/S0030605302000662
Bennett EL Milner-Gulland EJ Bakarr M Eves HE Robinson JG Wilkie DS Hunting the world's wildlife to extinction Oryx 2002 36 328 329 10.1017/S0030605302000637
Boehlert GW Biodiversity and the sustainability of marine fisheries Oceanography 1996 9 28 35
Steneck RS Human influences on coastal ecosystems; does overfishing create trophic cascades? Trends Ecol Evol 1998 13 429 430 10.1016/S0169-5347(98)01494-3
Jennings S Kaiser MJ The effects of fishing on marine ecosystems Adv Mar Biol 1998 34 203 351
Jennings S Kaiser MJ Reynolds JD Marine Fisheries Ecology 2001 Blackwell Science, Oxford
Anyinam C Ecology and Ethnomedicine: Exploring Links Between Current Environmental Crisis and Indigenous Medical Practices Soc Sci Med 1995 40 321 329 7899944 10.1016/0277-9536(94)E0098-D
Still J Use of animal products in traditional Chinese medicine: environmental impact and health hazards Complement Ther Med 2003 11 118 122 12801499 10.1016/S0965-2299(03)00055-4
Almeida CFCBR Albuquerque UP Uso de plantas e animais medicinais no estado de Pernambuco (Nordeste do Brasil): Um estudo de caso Interciencia 2002 27 276 284
Wilson EO Wildlife: legions of the doomed Time October 1995 77 79
Johannes RE Freeman MM Hamilton RJ Ignore fishers' knowledge and miss the boat Fish and Fisheries 2000 1 257 271 10.1046/j.1467-2979.2000.00019.x
Gadgil M Berkes F Folke C Indigenous knowledge for biodiversity conservation Ambio 1993 22 151 156
Schultes RE Johannes RE Reasons for ethnobotanical conservation' Traditional Ecological Knowledge: a collection of essays 1989 IUCN Publications Services 31 37
Kurien J Traditional ecological knowledge and ecosystem sustainability: new meaning to Asian coastal proverbs Ecol Appl 1998 52 55
Hipwell MAB Integrating local/traditional ecological knowledge into fisheries management in Canada 1998 Final Report submitted to Marine Ecosystem Conservation Branch. Department of Fisheries and Oceans Canada
Grifo F Rosenthal J Biodiversity and Human Health 1997 Island Press, Washington, D.C
Grifo F Newman D Fairfield AS Grifo F, Rosenthal J The origins of prescription drugs Biodiversity and human health 1997 Washington, DC: Island Press 131 163
Holmstedt B Bruhn JG Ethnopharmacology – a challenge J Ethno pharmacol 1983 8 251 6
Kunin WE Lawton JH Gaston KJ Does biodiversity matter? Evaluating the case for conserving species Biodiversity: a biology of numbers and differences 1996 Oxford: Blackwell Science 283 308
Marques JGW Fauna medicinal: Recurso do ambiente ou ameaça à biodiversidade? Mutum 1997 1 4
World Resources Institute World Resources Report 2000–2001 People and ecosystems the fraying web of life 2000 Washington D.C.: World Resources Institute 389
Commission on Intellectual Property Rights Traditional knowledge and geographical indications Report of the Commission on Intellectual Property Rights 2002
De Smet PAGM Is there any danger in using traditional remedies? J Ethno pharmacol 1991 32 43 50
Schnurrenberger PR Hubbert WT An outline of the zoonoses 1981 Ames, IA: Iowa State University Press
Gang L Jianqin X Application of traditional Chinese patent drugs in pet dog and cat 1996 Beijing, PR China: College of Veterinary Medicine, China Agricultural University
Henson S Heasman M "Food Safety Regulation and the Firm: Understanding the Compliance Process" Food Policy 1998 23 9 24 10.1016/S0306-9192(98)00015-3
Solovan A Paulmurugan R Wilsanand V Ranjith Sing AJA Traditional therapeutic uses of animals among tribal population of Tamil Nadu Indian Journal of Traditional Knowledge 2004 3 206 207
Labadie RP Problems and possibilities in the use of traditional drugs J Ethno pharmacol 1986 15 221 230
CITES – Convention on International Trade in Endangered Species of Wild Fauna and Flora Resolutions of CITES 10 th meeting of conference of the Parties (Conf 1019) Harare, Zimbabwe 1997
Chivian E Grifo F, Rosenthal J Global environmental degradation and biodiversity loss: implications for human health Biodiversity and human health 1997 Washington DC: Island Press 7 38
Lee SKH Trade in traditional medicine using endangered species: an international context Proceedings of the second Australian Symposium on traditional medicine and wildlife conservation, Melbourne Australia 1999
Puri K "Is Traditional or Cultural Knowledge a Form of Intellectual Property?" Oxford Electronic Journal of Intellectual Property Rights 2000
Article 8(j): Traditional Knowledge, Innovations and Practices Introduction
|
16270931
|
PMC1277085
|
CC BY
|
2021-01-04 16:40:12
|
no
|
J Ethnobiol Ethnomed. 2005 Aug 30; 1:5
|
utf-8
|
J Ethnobiol Ethnomed
| 2,005 |
10.1186/1746-4269-1-5
|
oa_comm
|
==== Front
J Ethnobiol EthnomedJournal of Ethnobiology and Ethnomedicine1746-4269BioMed Central London 1746-4269-1-61627093610.1186/1746-4269-1-6ResearchFocusing on the ethnobotanical uses of plants in Mersin and Adana provinces (Turkey) Everest Ayse [email protected] Ersin [email protected] Mersin University, Science &Art Faculty, Biology Department, Ciftlikkoy-Mersin, Turkey2005 6 9 2005 1 6 6 6 8 2005 6 9 2005 Copyright © 2005 Everest and Ozturk; licensee BioMed Central Ltd.2005Everest and Ozturk; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This paper presents the result of a study on the herbal drugs in the herbal markets in Mersin and Adana. The data were collected through direct interviews with herbalists and customers between 2002–2005 and the popular medicinal plants were investigated. A total of 107 species belonging to 56 families were investigated and the samples were listedwith their local and Latin names. The investigation includes cross-checking the disorders and their herbal cures and their recommended use stated by the local herbalists, by the parts used, and by the preparations. The cultivated species and their ethno botanical uses, are documented and extensive inventory is presented.
As a result, we observed that these plants are used especially for intestinal digestive disorders of the gastrointestinal tract, (21.68%), respiratory tract system disorders (10.43%), heart-blood circulatory system disorders (8.48%), urinary tract system disorders (7.70%), skin disorders (6.48%) and others.
Ethno-botanyMedicinal plants, Mersin, Adana, TURKEY.
==== Body
Background
In recent years, the increase in the residential and agricultural areas, and the decrease in medical plants have triggered the interest in ethno-botanical studies throughout the world [1-4]. The interest in herbal medicine in Turkey has progressed parallel to the increased interest in other developed countries. Recently, various studies have been conducted to prevent the folk medicine from disappearing [5-12].
For centuries, Turkish people have been using herbal medicine for the treatment of some daily diseases. The Taurus Mountains are one of the centers of the Mediterranean Region with a rich plant diversity. Accordingly, the traditional herbal medicines are important for the life of people. In this area, contagious diseases, cardiovascular disorders and cancer were investigated [13-18]. The world health report that provide us with the global rates and causes of mortality of cancer, contagious/parasitical diseases, circulatory system disorders, respiratory tract system and nerve disorders is presented in this study as a main source of analogy [19].
The aim of this, research is to focus on the kinds of medical diversity found in the herbal markets, on the frequency of usage of the plants, and thus, to show the different treatment types that are applied in the region.
The study is to be the first survey stating the herbal drugs in the herbal markets in the south of Turkey (Adana and Mersin provinces).
Methods
Our chosen study area, The Taurus Mountains' hills are located in the south of Turkey with C4–5–6 grid squares. The population densities of the two central cities (Mersin and Adana) are 537, 842 and 807, 934 [20].
There are about 70 herbal markets in the centers of Mersin and Adana. The plants presented in the herbal markets are collected from the villages in the Taurus Mountains (Figure 1). In the villages, dominant forest species are as follows: Pinus brutia, P. nigra, Quercus coccifera, Q. cerris, Q. ithaburensis, Q. infectoria, Juniperus oxycedrus, J. excelsa, J. drupacea and Abies cilicica [21,22]. The medical plants are harvested from places such as open areas, steppes, scrubs and roadsides. The plant materials are sold as dried bunches in open or pre-packed mixtures or as fresh preparations. The purchases depend on the request of the patient or on recommendation of the herbalists. Consumers generally boil these plants, make them into ointments or mix them with other plants depending on their intended use. The information about herbal medicine is gathered from at least two sources. The first source is the oral folklore that is passed on from one generation to the next and the second source is Ottoman, Arabic and Turkish herbal books which are sold in the bookstores.
Figure 1 The Research Area.
During a two year-long survey, almost all herbal markets in the research area are investigated. The fresh plants and dried samples that are purchased from various herbal markets have been cross-examined with reference books [21-25]. The data were collected through direct interviews with herbalists and customers. Thirty herbalist and 10 customers were interviewed. The number of the recommended uses are 1732 in total. Among them 26 species are recommended by Mersin' s customers. The popular medicinal herbs used for treatments are shown to be monthly by low: 3 *, medium: 5 ** and high: 7 ***. A total 107 species (22 of them are cultivated) that belong to 56 families were investigated and the samples were listedby their local and Latin names by the treatment rates and by the medical health data given by WHO [26].
Some bio-reactive components are indicated by Evans [27] and Baytop [28].
Voucher specimens, in duplicates, are saved both in the Herbarium of Biology Department in Mersin and the Chemistry Department in Cukurova University.
Results
While, Origanum vulgare L., Micromeria myrtifolia Boiss et. Hohen, Teucrium polium L., Saturea hortensis L., Nepeta italica L. and Sideritis spp. are used for making tea in the villages of the research area, nowadays the popular usage of tea plants such as Rosa canina, Helichrysum stoechas, Myrtus communis, Sideritis congesta and Anthemis spp. are used against obesity. The local people of Mersin ('yoruk') use Helichyrsum stoechas and Hypericum perforatum for stomachalgia, Halimione portulacoides for asthma, Arum maculatum for colitis, Prunus avium, Myrtus communis for obesity, Mandragora officinarum, Ferula communis for aphrodisiac, Tussilago farfara for cough, and Capparis spinosa, Portulaca oleraceae, Crocus sativus, Juglans regia for salad, pickle and jam.
According to the information obtained from some of the herbalists, the patients should start to accept the alternative therapies in health and medicine. Since diseases emerge due to the collapse of the immune system, the body should be cleaned up primarily. To achive this, it is suggested that a 3- month long therapy with Urtica spp., Equisetum arvense and Achillea spp. and a subsequent medical therapy would give good results.
Discussion
A range from 7–33 species were from our list, have also been documented by several other researchers in a number of other countries [29-34]. In addition, about 30 plants in the list are declared in the synopsis of ESCOP and WHO Monographs on medicinal plants [35]. Except the diuretic plants, the uses of other plants are different in Turkey and Italy (such as; Teucrium chamaedrys for anti-malaria, Laurus nobilis for anti-stress, Plantago lanceolata for supporting and Ocimum basilicum for anti-headache in Italy [34]). Also, similar 23 species, 9 taxa, 41 taxa and 22 taxa in our list were reported by Turkish researchers [9,10,36,37].
From our list, taxa containing endemic species (Thymus sipyleus var.sipyleus, T. cilicicus, Sideritis congesta and Liquidambar orientalis)- are the only ones planted in the Tarsus by the East Mediterranean Forest Research Institute. These include Mediterranean elements (21.05 %), Euro-Sib. Elements (8.42 %), Ir.-Tur. Elements (5.26 %) and some widespread, widely cultivated plants (62.10 %) and several species like Coriandrum sativum and Petroselinum crispum which are unknown in origin according to Davis [21].
In Table 1 (see additional file 1) medicinal plants in the south of Turkey that belong to 56 families are listed below: Lamiaceae (13 sp.), Asteraceae (11 sp.) and Rosaceae and Apiaceae (8 sp.), Fabaceae and Brassicaceae (5 sp.), Poaceae and Urticaceae (3 sp.), Anacardiaceae, Iridaceae, Malvaceae, Moraceae, Polygonaceae, Solanaceae, Zygophyllaceae and Liliaceae (2 sp.). Other families are, Adianthaceae, Amaranthaceae, Araceae, Aspleniaceae and Berberidaceae (1 sp.), Boraginaceae, Buxaceae, Capparidaceae, Cupressaceae, Chenopodiaceae, Ericaceae, Equisetaceae, Fagaceae, Gentianaceae, Hamamelidaceae, Hypericaceae, Juglandaceae, Lauraceae, Linaceae, Loranthaceae and Lycopodiaceae (1 sp.), Myrtaceae (1 sp.), Onagraceae, Oleaceae, Orchidaceae, Papaveraceae, Pinaceae, Paeoniaceae, Plantaginaceae, Primulaceae, Plumbaginaceae, Portulacaceae, Rhamnaceae, Ranunculaceae, Resedaceae, Salicaceae and Tiliaceae and (1 sp.).
Helichrysum stoechas and Anthemis spp. (sigir papatyasi), Reseda officinalis and Fumaria asephala (sahdere), Paliurus spina- christii and Lamium album (ballibaba), Hypnum cupressiforme and Lycopodium clavatum (kurtpencesi), Sideritis congesta and Salvia officinalis (adacayi), Cichorium inthybus and Taraxacum officinale (hindiba), Anthemis spp. and Matricaria chamomilla (papatya), Cupressus sempervirens and Quercus, Anagallis arvensis and Origanum majorona (mercankosk) may be used for the same ailments with the same local names in some herbal markets of Adana. Locally, Arbutus andrachne is preferred instead of Liquidamber orientalis and some Hypericum spp. are preferred instead of Centaurium erythrae in Mersin. As various herbs can be advised to cure the same disease under one common name, to reach to the exact and right herbtype, and to prevent any misunderstanding or misusage of the herbal plants, herbalists and medical firms need to know the original Latin names of these herbs and ask accordinly before any purchase.
In the old herbal books Gentiana lutea, Arum maculatum, Rumex acetosella, Nargissus and Opopanax spp., were claimed to be abortive [38,39]. However, today, Nasturtium officinale, Salvia officinalis, Nargissus, Cyclamen spp, Cinnamomum zeylanicum (from India; tarcın) and Cinchona officinalis (from India; kına kına, kinin) are belived to be abortive and dangerous in herbal markets. In the research area these therapies were conducted either together with the medical therapy under physician control or in the cases that the medical therapy has failed.
In table 1, gunluk, dari, anason, ardic, papatya, sogan, meyan koku, mese, cemen, servi kozalagi, yabani roka, menengic, corek otu, defne, gul, semiz otu, kimyon, hindiba, gebere, keten, egrelti, sakiz, kisnis, mercankosk, nane, anason, ravent, sinirli ot, kekik, susen, sahdere, tere, yarpuz, safran was used in Central Asia according to Turk medicine books [40-45]. Today, these plants can be seen in Asian medicine treatments such as Capparis spinosa, Apium graveolens, Equisetum arvense, Berberis vulgaris, Zea mays, Urtica dioica, Sambucus nigra, Capsella bursa-pastoris, Salix caprea in Azerbaijan [46] and Pteridium aquilinum, Myrtus communis, Plantago lanceolata, Portulaca oleraceae, Coriandrum sativum, Malva neglecta, Buxus sempervirens, Adianthum capillus-veneris, Mentha longifolia, Origanum vulgare, Nasturtium officinale in Pakistan [47].
If we look at the XVI. Century medicinal plant list of the Ottaman Empire epoch, we can find the same list such as, adam otu, ada sogani, ahlat, adacayi, anason, andiz otu, ayva, bakla, baldıran, bogurtlen, biberiye, burcak, cakal erıgı, ceviz, corek otu, defne, dere otu, egrelti otu, feslegen, findik, funda, gul, guzel avrat otu, hardal, hindiba, incir, isirgan otu, karadut, keten, kekik, kedi otu, kıraz, kımyon, kuzukulagı, labada, lavanta, maydanoz, mercankosk, mersin, nane, zeytin, melisa, kantaron, kuskonmaz, misir puskulu, mese kabugu, ogul otu, pelin and safran, sakiz, sandal, sumak, susen [41,48-50]. Species that have been indicated above had been applied also in Central Asia and in the Ottoman-Turk Medical science. Even today they are still included on our list [44,45,51-53].
A great many of the vernacular names and common families and 15–24 species were shared in Anatolia and Central Asia [6,8,11,12,54]. For example, andiz/andiz for Inula sp., Quirkbog' um/kırkkilit otu for Equisetum sp., yarpuz/yarpuz for Mentha sp., Qoratut/dut for Morus sp., itburnu/kus burnu for Rosa sp., kılıchak/kılıc otu for Plantago sp., Asteraceae, Apiaceae, Lamiaceae and Rosaceae families (Table 1) [54].
In the light of these data, some plants that are presented in our list, can be said to be based on the Central Asia Turk medical science and also their prescriptions that had been used between the XIV-XVI Th. centuries.
It would not be inappropriate to say that, the similar climate and environmental conditions (especially in the regions of Mediterranean cultures) have been the cause of using the similar plants in that region [2,29,31,33,34], [54].
Table 2 shows clearly that in the south of Turkey (Mersin and Adana), the plants were used mainly for pathologies of the digestive, respiratory, heart-blood-liver, intestinal, urinary, skin system disorders, inflammatory and related ailments, nerve and related ailments and rheumatism, sprains and related ailments. These are followed by the others.
Table 2 The comparison with the other provinces and villages (Marmara region) in Turkey
Pathologies In Mersin, Adana (%) (Our study) In Istanbul: Sile (%) [9] In Balikesir: Gonen (%) [37] In Sakarya (%) [36]
Intestinal-digestive disorders 21.68 9.47 45.63 8.6
Respiratory system disorders 10.43 16.71 10.77 10.1
Heart-blood disorders 8.48 5.07 1.54 13
Urinary system disorders 7.70 7.99 9.23 9.4
Skin disorders 6.48 20.37 12.31 12.2
Antiinflammatory, antiseptic 6.20 3.63 4.61 -
Nerve disorders 5.73 - - 13.7
Liver-spleen disorders 4.67 1.44 - -
Gynecological disorders 4.42 0.72 - 2.9
Arthritis 3.16 7.26 7.69 1.4
Analgesic, anodyne, emollient 2.42 - 1.03 -
Sedative 1.90 - 1.53 2.9
Conclusion
In conclusion, the comparison of the treatments between provinces and villages, shows us a decreased incidence of heart-blood, liver-spleen and gynecological disorders in the villages and increased incidence of the arthritis in the villages (Table 2) [9,36,37]. Contrary to the Mediterranean Region, there is an increased incidence of skin problems that are related to a humid climate and other different environmental conditions of the Marmara Region.
The percentage of skin disorders of Uzbekistan, Italy, Turkey and Greece implies that the increase may also be related to the technological developments, environmental pollution and humidity (Table 3).
Table 3 The comparison of the incidence of remedies between the other countries
Pathologies In Turkey (%) (Our study) In Uzbekistan (%) [54] In Italy (%) [33] In Greece (%) [29]
Intestinal-digestive disorders 21.68 31.20 10.60 16.93
Respiratory system disorders 10.43 13.30 13.00 3.06
Heart-blood disorders 8.48 5.10 1.35 7.93
Urinary system disorders 7.70 4.10 3.80 7.66
Skin disorders 6.48 16.60 11.50 4.32
Antiinflammatory, antiseptic 6.20 4.50 5.35 1.35
Nerve disorders 5.73 10.40 5.65 5.87
Liver-spleen disorders 4.67 8.10 2.56 5.22
Gynecological disorders 4.42 2.30 3.16 2.97
Arthritis 3.16 3.50 2.18 6.31
Analgesic, anodyne, emollient 2.42 - 1.27 -
Sedative 1.90 10.40 2.71 4.96
In Table 3, these results are interpreted with regional medical data; the first place in the list is taken by intestinal- digestive pathologies [13,14,17,55]. These results can be related to the fact that received immigration from the less developed cities and that they have rather poor hygienic conditions with regard to food and water [13,14]. However we have observed the same rates in Uzbekıstan [54], in Italy [32,33] and in Greece [29]. The World Health Report indicates that this problem appears in less developed countries of the world [19].
Nerve disorders are similarly seen in the Mediterranean countries as given in Table 3. Also, similar results are obtained -such as the percentages of rheumatism, digestive and respiratory system disorders- in Turkey and Uzbekistan.
In less developed countries, during childhood and adolescence, there is a factor of risk of contagious (infectious) diseases. The global mortality rate (over 52 million) depends on contagious and parasitic diseases (over 17 million), heart-blood system disorders (over 15 million), cancer (over 6 million) and chronic respiratory tract disorders (over 3 million) [19].
Following the listed results, we came up with the fact that the causes of mortality were mostly respiratory tract and circulatory system disorders [17,18], and that there is a low ratio of cancer (1.28 %) [16,55]. However, except cancer that is ranked as the third disease in the list of clinical world diseases, other rankings in our research findings seem identical/parallel with this list. This fact led us think that the herbal/cheap cures for cancer might have been deliberately exchanged with the chemical/expensive ones, or just carelessly overlooked.
In the end, the close analogy was discovered between the respiratory tract disorders, circulation system disorders, the intestinal-digestive diseases- which are related to the malnutrition, undeveloped hygiene habits, and the use of unnecessary antibiotics [56,57], and the medical health rates that are stated at the top three list of Turkey and the compared countries, drives our attention to the following herbs that can be used for all three health problems: Origanum majorana, Equisetum arvense, Glycyrrhiza glabra, Matricaria chamomila, Nigella arvensis, Paliurus spina-christii, Armeniaca vulgaris, Linum catharticum, Orchis anatolica, Rosmarinus officinalis, Myrtus communis, Lavandula stoechas and Mentha pulegium.
Supplementary Material
Additional File 1
Table 1. The list of Medicinal plants of research area
Click here for file
Acknowledgements
I would like to thank to Bilgehan Cetinkaya (Sifa Naturel Urunler), Nurs' i Lokman Hekim, Cerci Yusuf 1–4, Kor Yusuf, Her Derde Deva 1, 2 and the others who have provided me with the various herbal samples and information.
==== Refs
Liu Y Pharmaceutical composition for treating cancer PGPUB Production Database 20040072790
Harsha VH Hebbar SS Hedge GR Shripathi V Ethnobotanical knowledge of plants used by Kunabi Tribe of Karnataka in India Fitoterapia 2002 73 281 287 12234569 10.1016/S0367-326X(02)00078-3
Naguib YMA Herbal compositions and methods for diabetes and weight loss management PGPUB Production Database 20030143291
You HJ Natural cell control carrier modular compounded Inula helenium L. extract and water-soluble chitosan PGPUB Production Database 20020155175
Baytop A Farmasotik Botanik 1983 I U Ecz Fak Yayinlari
Bingol F Some Drug samples sold in the herbal markets of Ankara OT Sistematik Botanik Dergisi 1995 2 83 110
Yesilada E Honda G Sezik E Tabata M Fujita T Tanaka T Takeda Y Yakaishi Y Traditional Medicine in Turkey V, Folk Medicine in the inner Taurus Mountains Journal of Ethnopharmocology 1995 46 133 152 10.1016/0378-8741(95)01241-5
Bagci Y Aladaglar (Yahyali-Kayseri) ve Cevresinin Etnobotanik Ozellikleri The Herb Journal of Systematic Botany 2000 7 89 94
Tuzlacı E Tolon E Turkish folk medicinal plants, part III: Sile (İstanbul) Fitoterapia 2000 71 673 685 11077175 10.1016/S0367-326X(00)00234-3
Sezik e Yesilada E Honda G Takaishi Y Takeda Y Tanaka T Traditional Medicine in Turkey X. Folk Medicine in Central Anatolia Journal of Ethnopharmacology 2001 75 95 115 11297840 10.1016/S0378-8741(00)00399-8
Keskin M Alpinar K Kislak (yayladagi-Hatay) hakkinda etnobotanik bir arastirma OT Sistematik Botanik Dergisi 2002 9 91 100
Akcicek E Vural M Kumalar Dagi (Afyon) ve Cevresindeki Bazi Bitkilerin Yoresel Adlari ve Etnobotanik ozellikleri The Herb Journal of Systematic Botany 2003 10 151 168
Ergoz G Uguz K Kaya Z Kaya A Mersın Universitesi Tıp Fakultesi Hastanesinde Hastane Infeksiyonları ME U Tıp Fak Derg 2003 4 226 230 128638
Borekcı G Ates A Polat G Emen FM Sakallı S Mersın Ilı Icme ve Kullanma Sularının Mıkrobiyolojık ve Kımyasal Analızı ME U Tıp Fak Derg 2002 3 417 422 128638
Caglıkurekcı M Akca T Dırlık M Ogetman Z Ocal K Colak T Aydın S Kolon ve Rektum Kanser Olgularımızın Analizi ve Tedavi Yaklasımlarımız ME U Tıp Fak Derg 2002 3 373 381 128638
Deveci F Turgut T Altuntas E Muz MH Gogus Hastalıkları Konsultasyonu Yapılan Olguların Degerlendırılmesi MEU Tıp Fak Derg 2002 3 170 177 9830266
Baykara N Hosten T Tutan H Salbes A Solak M Toker K Yogun Bakım Unitesinde Yeniden Yatıs ve Sonrasında Gorulen Servıs Olumlerı MEU Tıp Fak Derg 2003 4 130 139 9830266
Avkarogulları M Kalp yetersizliginde Anemi ve Eritropoietin Tıpta uzmanlık 2004 5161 61
World Health Organisation Oz V The World Health Report, 28–30 October, Geneva, Switzerland TC Saglık Bakanlıgı, Dısiliskiler Dairesi Baskanlıgı 1997 50 Dunya Saglık Assemblesi Raporu
Adana Bolge Istatistik Mudurlugu Mersın ve Adana merkez ve ilceleri nufus sayımı sonucları 2004
Davis PH Flora of Turkey and the East Aegean Islands 1965 Edinburgh University Press
Guner A Ozhatay N Ekim T Baser KHC Flora of Turkey 2000 Edinburgh University press
Baytop T Turkce Bitki Adlari Sozlugu 1994 Ataturk Kultur Dil ve Tarih Yuksek Kurumu, Turk Dil Kurumu Yayinlari
Yildirimli S The Chorology of the Turkish species of Acanthaceae, Aceraceae, Aizoiaceae, Amaranthaceae, Anacardiaceae families The Herb Journal of Systematic Botany 1997 4 125 133
Yildirimli S The chorology of the Turkish species of Crassulaceae, Cucurbitaceae, Cuscutaceae and Cynocrambaceae families The Herb Journal of Systematic Botany 2003 10 255 260
WHO Traditional Medicine Strategy; 2002
Evans WC Trease and Evans' Pharmacognosy 2002 Edinburgh: WB Saunders
Baytop T Türkiye' de Bitkiler ile Tedavi, Geçmişte ve Bugün 1999 Nobel Tıp Kitabevleri
Handlidou E Karousou R Kleftoyanni V Kokkini S The herbal market of Thessaloniki (N Greece) and its relation to the ethnobotanical tradition Journal of Ethnopharmacology 2004 91 281 299 15120452 10.1016/j.jep.2004.01.007
Katewa SS Chaudhary BL Jain A Folk herbal medicines from tribal area of Rajasthan, India Journal of Ethnopharmacology 2004 90 171 177 14698527
Abu-Irmaileh EB Afifi UF Herbal medicine in Jordan with special emphasis on commonly used herbs Journal of Ethnopharmacology 2003 89 193 197 14611882 10.1016/S0378-8741(03)00283-6
Ballero M Poli F Sacchetti G Loi MC Ethno botanical research in the territory of Fluminimaggiore (south-western Sardinia) Fitoterapia 2001 72 788 801 11677018 10.1016/S0367-326X(01)00334-3
Loi MC Poli F Sacchetti G Selenu MB Ballero M Ethnopharmacology of Ogliastra (Villagrande Strisaili, Sardinia, Italy) Fitoterapia 2004 75 277 295 15158984 10.1016/j.fitote.2004.01.008
Pieroni A Cassandra LQ Rocco FS Folk Pharmaceutical knowledge in the territory of the Dolomiti Lucane, inland southern Italy Journal of Ethnopharmacology 2004 9 373 384 15507363 10.1016/j.jep.2004.08.012
Silano M Vincenzi M Vincenzi A Silano V The new European Legislation on Traditional Herbal Medicines: main features and perspectives Fitoterapia 2004 75 107 116 15030913 10.1016/j.fitote.2004.01.001
Uzun E Sarıyar G Adsersen A Karakoc B Otuk G Oktayoglu E Pırıldar S Traditional medicine in Sakarya province (Turkey) and antimicrobial activities of selected species Journal of Ethnopharmacology 2004 95 287 296 15507351 10.1016/j.jep.2004.07.013
Tuzlacı E Aymaz EP Turkish folk medicinal plants, part IV: Gonen (Balıkesir) Fitoterapia 2001 72 323 343 11395255 10.1016/S0367-326X(00)00277-X
Gurbuz A Sifali Nebatlar 1982 Bayrak Yayimcilik-Matbaacilik Ltd
Palaz T Hastaliklarin Bitkilerle Tedavi Receteleri 1983 Cevik Matbaasi
Onler Z XV. Yüzyıl Türk Hekimlerinden Sabuncu Serafettin' in "Mücerrebname" adlı Eseri Fırat Universitesi Dergisi 1987 169 190
Onler Z Celaluddin Hizir (Haci pasa) Muntahab-ı Sifa I, Giris-Metin 1990 Atatürk Kültür, Dil ve Tarih Yüksek Kurumu Türk Dil Kurumu yayınları
Onler Z Muntahab-ı Sifa II, Sozluk Simurg yayınları 1999
Uzel I Ilk Turkce Kitap Yazmalarının Agız-Dıs Hastalıkları Yonunden Incelenmesı Doktora Tezi 1979
Uzel I Anadolu Uygarlıklarında Dishekimligi 2000
Yakupoglu HD Adısoglu DI Azerbaycan' da İlaç Nebatlarının Öğrenilmesinin Bazı Sonuçları Doga Tr J Of Pharmacy 1991 15 178 181
Iqbal I Hamayun M Studies on the traditional uses of plants of Malam Jabba valley, District Swat, Pakistan Ethnobotanical Leaflets 2004 1 30
Eyuboglu IZ Anadolu Halk Ilacları 1980 Hur Yayın
Unver S Fatih Devri Yemekleri 1960 Edb Fak Yayınları
Esin K Ibn-i Sina El-Kanun fi't Tıbb II. Kitap, Atatürk Kültür Merkezi Baskanligi Yayinlari, Türk Tarih Kurumu Basımevi; 2005
Ilhan B Sifalı Otlar Kitabı Karacan Yayınları 1982 Ozal Matbaası
Yalcın S Nedim Divanının Tedkiki, Bitkiler, Cicekler ve Hayvanlar H U Yüksek lisans tezi 1979
Demırhan E Mısır Carsısı Drogları Doktora Tezi 1971
Sezik E Yesilada E Shadidoyatov H Kuliyev Z Nigmatullaev AM Aripov HN Takaishi Y Takeda Y Honda G Folk medicine in Uzbekistan Journal of Ethnopharmocology 2004 92 197 207 10.1016/j.jep.2004.02.016
Akca S Ustun H Bozcuk H Balcı M Yardımsever M Ersoy F Twoyears experience in Medical Intensive Care Unit in Akdeniz University Faculty of Medicine, CU Tıp Fak Der 1998 23 82 88
|
16270936
|
PMC1277086
|
CC BY
|
2021-01-04 16:40:12
|
no
|
J Ethnobiol Ethnomed. 2005 Sep 6; 1:6
|
utf-8
|
J Ethnobiol Ethnomed
| 2,005 |
10.1186/1746-4269-1-6
|
oa_comm
|
==== Front
J Ethnobiol EthnomedJournal of Ethnobiology and Ethnomedicine1746-4269BioMed Central London 1746-4269-1-71627094010.1186/1746-4269-1-7ResearchEthnopharmacological survey of different uses of seven medicinal plants from Mali, (West Africa) in the regions Doila, Kolokani and Siby Togola Adiaratou [email protected] Drissa [email protected]élé Seydou [email protected] Hilde [email protected] Berit Smestad [email protected] Section of Pharmacognosy, Department of Pharmaceutical Chemistry, University of Oslo PO Box 1068 Blindern, 0316, Norway2 Department of Traditional Medicine, BP 1746, Bamako, Mali2005 27 9 2005 1 7 7 23 6 2005 27 9 2005 Copyright © 2005 Togola et al; licensee BioMed Central Ltd.2005Togola 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.
An ethnopharmacological survey was carried out to collect information on the use of seven medicinal plants in rural areas in the nearby regions of Bamako, Mali. The plants were Opilia celtidifolia, Anthocleista djalonensis, Erythrina senegalensis, Heliotropium indicum, Trichilia emetica, Piliostigma thonningii and Cochlospermum tinctorium
About 50 medical indications were reported for the use of these plants in traditional medicine. The most frequent ailments reported were malaria, abdominal pain and dermatitis. The highest number of usages was reported for the treatment of malaria (22%). The majority of the remedies were prepared from freshly collected plant material from the wild and from a single species only. They were mainly taken orally, but some applications were prepared with a mixture of plants or ingredients such as honey, sugar, salt, ginger and pepper. Decoction of the leaves was the main form of preparation (65%) and leaf powder was mostly used for the preparation of infusions (13%). The part of the plants most frequently used was the leaves. There was a high degree of informant consensus for the species and their medicinal indications between the healers interviewed.
The results of this study showed that people are still dependent on medicinal plants in these rural areas of Mali.
EthnopharmacologyMaliOpilia celtidifoliaAnthocleista djalonensisErythrina senegalensisHeliotropium indicumTrichilia emeticaPiliostigma thonningiiCochlospermum tinctorium
==== Body
Background
Located in West Africa, Mali is a landlocked country with an area of approximately 1, 246,000 Km2 for an estimated population of 13 million inhabitants. Mali is one of the poorest countries in the world with a GPD of 725$ (2002) per capita [1]. The economy is essentially based on agriculture; the health sector policy promotes community-based, self-supported health care services and the administration of essential medicines including traditional medicines [2]. Like in many other developing countries, people in Mali use medicinal plants to improve their state of health. Traditional medicine is a significant element in the cultural patrimony. Its use has increased with the increase in price of conventional medicine in the local currency. Traditional medicine still remains the main recourse for a large majority of people for treating health problems. Approximately 80% of the population in Mali use traditional medicine as their only type of medicine [2]. Official medical attention is usually based on commercial drugs that have to be purchased with money, while a traditional medical consultancy has a much lower cost, including the consumption of the medicinal plants required [3]. Most of the plants used in this traditional medicine have never been investigated for their chemical composition and pharmacological activities. It is therefore important to study these plants to substantiate the traditional medical knowledge. People, and especially the traditional healers, should be informed of the benefits, risk and limitation of the plants they use for medical purposes.
A traditional healer is defined as a person with competence to practice traditional medicine. From 1968 to 1978 registration of traditional healers and medicinal plants was carried out in all the administrative regions in Mali by an interdisciplinary team. The competence of a healer is evaluated on the person's achievements on curing diseases and the results are essential for consideration of registering the person as a traditional healer. After being registered, the DMT sets up a principle of collaborating with the traditional healers. The collaborating healer is not obliged to deliver samples of his medications to DMT, but if he wishes to do so the plants will be subjected to toxicological, pharmacological and phytochemical analyses, the results of which are given to the healer. As a result of this collaboration, the healer is granted official recognition as a practitioner in traditional medicine and is provided with an identity card for traditional practitioners. Other traditional healers are also allowed to practice with no restriction, but they do not have a registration card. In some localities of Mali, the healers are grouped in association and have created gardens of medicinal plants [2]. All studies being undertaken between DMT and the traditional healers follow ethical aspects and rules set down by the local government as both DMT and the traditional healers are part of the health care system of Mali.
The Department of Traditional Medicinal (DMT), the first research establishment for the study of medicinal plants in Mali, and a collaborating centre of World Health Organisation (WHO) on traditional medicine, has as the main objectives: the registration of traditional healers, traditional knowledge and medicinal plants, in addition to perform research, and to develop Improved Traditional Medicines (ITMs) from local plants.
Several medicinal plants have been studied in the laboratory of DMT using classical methodology for phytochemistry, pharmacology and toxicology. According to the results of these studies, pharmaceutical formulae have been developed from plants in their natural form or in the form of infusion in ointments and syrups. These phytomedicines are called Improved Traditional Medicines. DMT has so far developed 12 ITMs that have been standardised according to traditional administration regimes. The doses have been investigated for lack of toxicity and the expiration dates for the products have been determined. These products are now being sold in drugstores in Mali. Seven of these are acknowledged as essential medicines by the Health department in Mali. These ITMs are: Balembo against cough, Dysenterial against dysentery, Gastrosedal against ulcers and gastritis, Hepatisane against hepatitis, Laxia-cassia against constipation, Malarial against malaria and Psorospermine against dermatitis [2].
In this willpower to develop new traditional medicines, several ethnobotanical studies were carried out on medicinal plants from Mali [4-7]. Our study is placed within this framework.
The aim of this study was to identify different uses of certain medicinal plants which were retained for the investigations as new ITMs, but a lack of substantiated information on their use in traditional medicine made this survey necessary. Seven plants were chosen for this study and their uses in traditional medicine were investigated. In February 2005 an ethnopharmacological survey was carried out in Siby, Doila and Kolokani in nearby areas of Bamako, the capital of Mali. The village and healers in the areas to be interviewed were selected randomly and no appointment was made prior to the visits.
The result will give an overview on the cure potency of these plants according to the traditional medicinal healers in these areas. The plants will later be investigated for chemical pharmacological and toxicological aspects by DMT in order to be developed into new Improved Traditional Medicines that can be registered by the Health department in Mali.
Methods
Description of the study area
The rural district of Siby is situated 50 km south of Bamako, Kolokani 140 km north of Bamako and Doila approximately 130 km east of Bamako. Five of the visited villages belong to the Siby region, these are: Dioulafondo, Guena, Kalassa, Djissoumala and Kakan; Five belong to the Dioila region: Dioila, Falakono, Diana, Finianan and Wolome. Didieni, Kolokani and Niamabougou belong to the Kolokani region. The location of the main areas can be seen from the map (figure 1). These areas are part of Koulikoro, one of the administrative divisions of Mali. The main income sources in these areas are agriculture and the commerce of agricultural products. Doila is one of the main areas for production of cotton in Mali. Public health services are used, but home medication is practiced primarily with medicinal plants. Traditional medicine is the first choice for the population for health problems, and healers in these areas are reputed to have good knowledge on medicinal plants and disease treatment.
Figure 1 Map of Mali with focus on the survey areas.
Interviews with the traditional healers
Conversations with the healers were used to obtain information on the use of the seven plants being the object for this survey. The healers that consented were asked to give their knowledge about the diseases they used those plants against, the method of preparation of the remedy, details of administration, including the approximate amounts and number of doses per day or week, the adverse effects of the remedy and how to treat these adverse effects. Traditional gifts of cola nuts and money were bestowed upon the traditional healers. The conversations were built on trust with the common goal to improve the health situation in the country and to preserve and increase the knowledge on medicinal plants.
The following seven plants were the focus for the ethnopharmacological survey: Opilia celtidifolia Guill. & Perr. (Opiliaceae), Anthocleista djalonensis A Chev. (Loganiaceae), Erythrina senegalensis DC (Leguminosae: Papilionoideae), Heliotropium indicum Linn (Boraginaceae), Trichilia emetica Vahl, (Meliaceae) Piliostigma thonningii Schum (Leguminosae: Caesalpinioideae) and Cochlospermum tinctorium A. rich. (Cochlospermaceae). The results of the survey are given in Table 2 (see additional file 1).
Fidelity level
The fidelity level (Fl) [8] among the healers from the same district was calculated according to a following formula:
Fl (%) = (Np/N) × 100
Np is the number of healers from one given district that claim a use of a plant species to treat a particular disease, and N is the number of healers from the same district that use the plants as a medicine to treat any given disease. The formula was applied in order to compare data from different district where the survey was performed.
Below we will discuss the information given on the uses of these plants as well as to give a survey on the scientific knowledge available in the literature. The literature research was carried out for all the plants within the databases available via the library of the University of Oslo, Norway. These are: Scifinder, BIBYS, Biological abstract/WebSPIRS and OVID web.
Results and discussion
In our study, totally 94 healers from 13 villages from the three areas were interviewed (Table 1). Most of them were members of the traditional healers association in their region and also registered by the Department of Traditional Medicine. 34% of the interviewed healers were from Siby, 35% were from Doila and 31% were from Kolokani. Men dominate the practice of traditional medicine, 76 of the interviewed healers (81%) were men over 40 years; the oldest being 104 years old. Few healers were below 40 years, amongst these, the youngest was 27. 18 women were interviewed in this study; they seem to have less knowledge than men about traditional medicine. This is because they mainly treat children and typical child diseases, while men treat both children and adults. The average age of the women is not known, most of the women have an approximate idea about when they were born, but not the exact year.
Table 1 An overview of the traditional healers interviewed.
Districts Villages Numbers of healers Sex Age range
Males Females
Siby Djissoumala 7 6 1 28–76
Dioulafondo 12 8 4 30–79
Guena 7 7 0 42–85
Kalassa 2 2 0 35–55
Kakan 4 3 1 65–85
Doila Doila 6 5 1 47–70
Falakono 12 8 4 38–80
Diana 5 4 1 27–70
Finianan 8 7 1 49–104
Wòlòmè 2 1 1 56–80
Kolokani Kolokani 15 12 3 45–73
Didieni 9 9 0 55–62
Niamabougou 5 4 1 50–61
The results of the survey are presented in Table 2 (see additional file 1) and discussed below.
Medical applications of the plants
The present ethnopharmacological survey has gathered information on about 50 different diseases treated by these 7 plants (Table 2, see additional file 1). The most frequent reported ailments were malaria, abdominal pain and dermatitis. Opilia celtidifolia and Trichilia emetica were the most used plants based on the high number of uses reported. The healers consensus based on the fidelity level (Fl) index [8] was calculated for the most frequently reported diseases or ailments (Table 3). The fidelity level of the healers that use O. celtidifolia against malaria and abdominal pain is high in Doila (61 and 54% respectively). In Siby O. celtidifolia is mainly used against dermatitis (Fl = 75%) while the main use of the plant in Kolokani is against abdominal pain (Fl = 23%). T. emetica is mainly used against malaria in Doila and Siby (Fl = 47 and 41% respectively), while the use against abdominal pain is the most reported use in Kolokani (Fl = 32%). FI were not determined for the other plants and other uses due to the low number of reports.
Table 3 Comparison of the use of Opilia celtidifolia (Oc) and Trichilia emetica (Te) in the Doila, Siby and Kolokani regions based on the fidelity level.
Fidelity level (%)
Siby Doila Kolokani
Main reported diseases Oc Te Oc Te Oc Te
Malaria 17 41 61 47 22 13
Abdominal pain 23 36 54 32 23 32
Dermatitis 75 40 21 40 4 20
Below the most frequent uses of each plan is coupled with information from scientific literature.
Opilia celtidifolia Guill. & Perr. (Opiliaceae)
Local Name: Korôgué
Traditional use
Opilia celtidifolia is well known to the traditional healers in our study region as a remedy to cure several diseases. The main reported disease is dermatitis (dermatitis is by the healers used as a common terminology for all kinds of skin disorders); the frequency of citations is 19.2% of 125 total citations. This use is common in Siby were the plant has a reputation for curing skin lesions and wounds; followed by malaria (14.4%), another frequent disease reported by the healers in the southern part of Mali where the survey was carried out. O. celtidifolia is known in the district of Doila as an appetizer (10.4% of frequency of citation), as an abdominal pain killer (10.4%) and an intestinal worm cure (7.2%); this last use was reported several times in Kolokani as well. Various other ailments were also reported (Table 2, see additional file 1).
According to the literature, decoction of the leaves is used as febrifuge in Ivory Coast. In Senegal it is used as a gargle, against dental abscesses, to treat oedema leprosy, acting as a purgative, and used against headache. A macerate left to stand overnight and strongly salted, taken on an empty stomach is meant to be particularly effective in expelling oxyuris worms from children [9].
Biological activities
Little is found on the biological activities of O. celtidifolia in the literature. Shihata et al. [10] isolated saponins from the methanol extract and found antispasmodic and anthelmintic activities for these compounds. The authors also signalled the lack of information regarding the biological properties and possible therapeutic value of the plan. The effects shown above may explain both the use as an abdominal pain killer and against intestinal worms as found to be common uses in our survey.
Anthocleista djalonensis A Chev. (Loganiaceae)
Local Name: Fartanlafla
Traditional use
The uses of this plant were reported by traditional healers in Siby only. Malaria and abdominal pain are the most frequent reported ailments with 32% of citation frequency for each. The other uses of this plant were only reported once (Table 2, see additional file 1).
According to the literature, the decoction of the leaves is in Sierra Leone drunk as a treatment against jaundice. In Ivory Coast the root is used as a diuretic and a vigorous purgative, and also as a poison-antidote, against leprosy, as an emmenagogue and in the treatment of oedemas and elephantiasis of the scrotum. The root decoction is taken against chest pains, for constipation and against gonococci [11].
Biological activities
In an evaluation of an extract of A. djalonensis for activity against bacterial isolates from cases of non gonococcal urethritis performed by Okoli et al. [12], the cold water and ethanol extract of the roots showed a remarkable broad spectrum activity against Staphylococcus aureus and Escherichia coli. Thus, the antibacterial activity exhibited by the extracts against theses organisms justifies their general use of the plant in the treatment of sexually transmitted diseases and the folklore use of aqueous decoctions of the plant in the treatment of dysentery and other gastrointestinal diseases.
Erythrina senegalensis DC (Leguminosae: Papilionoideae)
Local Name: N'tékissè
Traditional use
This plant is not well known to the healers in Siby. Only a few indications have been reported among which the use against amenorrhoea, urinary bilharzias and sterility are the most frequent mentioned (Table 2, see additional file 1).
According to the literature, in Gambia and Senegal the sap from the crushed leaves is applied to wounds for two or three days to promote healing. In Ghana and Nigeria the pounded bark and leaves are taken by women in a soup against barrenness. They are also used as enemas. In Mali the decoction of leaves is used to provoke diuretic activity and is taken against urinary bilharzia. In Senegal a macerate of the trunk-bark is taken internally for amenorrhoea and externally against headaches and eye-troubles. In Ivory Coast the wood is chewed as an aphrodisiac [9].
Biological activities
The suppressive activity of E. senegalensis water extract of the bark against Plasmodium berghei in mice has been evaluated by Saidu et al. [13]. Doses of 50 and 100 mg/kg reduced the mean parasitemia, but the effects were not significant. The suppressive effect were 16.5 and 23.2 % respectively, while the reference standard (chloroquine) produced a profound suppression effect of 95.8%
The analgesic effect of E. senegalensis was examined by the same author [13]. The water extract of the bark (50–100 mg/kg) significantly inhibited acetic acid induced abdominal constriction in mice in a dose dependant manner.
In a screening for antibacterial activity by Kone et al. [14], E. senegalensis root ethanol extract was found to contain active bactericides with IC50 values of 12 μg/ml against Staphylococcus aureus, Enterococcus faecalis, Bacillus subtilis and Streptococcus pyogenes. These studies have no relevance to the uses of the plant in Mali and such studies are needed.
Heliotropium indicum Linn (Boraginaceae)
Local Name: Nonsikou
Traditional use
In Siby the use against vomiting is the most frequently reported. Other reported uses are against amenorrhoea, baby thinness, ocular infections and high blood pressure (Table 2, see additional file 1).
According to the literature, the use of a decoction of the leaves is recorded in Sierra Leone for washing new-born babies. In Nigeria and Ghana the sap is applied to gumboils, to clean ulcers and to cure eye infections. In Guinea the decoction of the whole plant is taken as a febrifuge. In Senegal the leaf powder is applied to dermatitis and especially to suppurating eczema and impetigo in children [15]. The leaf decoction is used in Indonesia for thrush and in poultices for herpes and rheumatism. In Ivory Coast the dried leaf powder is taken up by the nose as decongestant in colds and sinusitis [16]
Biological activities
Wound healing activity has been reported by Reddy et al. [17]. They showed that topical application of 10% w/v of H. indicum increased the percentage of wound contraction and completed wound healing by 14th day indicating rapid epithelization and collagenization. The control used healed a similar wound in 23 days. An increase of the tensile strength indicated the increase in collagen facilitating wound healing.
Kugelman et al. [18] isolated the N-oxide of the alkaloid indicine from H. indicum and observed significant anti-tumour activity of the compound in W-256 carcinosarcoma, L-1210 leukemia, P-388 leukemia, P-1534 leukemia and melanoma B-16 tumour systems. On the basis of these results the compound was selected for human clinical trials. Studies related to the uses in Mali have not been performed.
Trichilia emetica Vahl, (Meliaceae)
Synonym: Trichilia roka Chiov.
Local Name: Soulafinzan
Traditional use
Several uses were reported for T. emetica, the main according to the citation frequency is against malaria 23.8%. This use is a common knowledge wherever the survey was carried out. The next most frequent use is against abdominal pain (19.2%), against dermatitis (7.7%), haemorrhoids (6.2%), and jaundice and chest pain (5.4%). Several other diseases were also reported for the use of T. emetica (Table 2, see additional file 1).
According to the literature, in eastern Africa the root bark decoction is used as an emetic and a purgative, against fever, epilepsy, leprosy and makes women fecund. In Senegal the leaf decoction is used against blennorrhoea; the infusion against headache and as lotion on burns [9]. The leaf decoction is used against malaria and scabies; the stem and leaf decoction is used against intestinal, coetaneous or mouth infections. The fruit is in eastern Africa used as a diuretic. T. emetica is also used against poisoning, hepatitis, ulcer, dysmenorrhoea, asthma, cirrhosis and internal worms [19].
Biological activities
In the exploration of biological activities, the family Meliaceae has attracted extensive attention. T. emetica particularly has been largely investigated. El Tahir et al. [20] investigated the anti-plasmodia activity of the methanol extract of the leaves and found an IC50 of 2.5 μg/ml against Plasmodium falciparum sensitive strain Dd2 and 17.5 μg/ml against the resistant strain 3D7. Anti-inflammatory activity was demonstrated by McGaw et al. [21]; they found an inhibition of 89% of prostaglandin synthesis by the ethanol extract of the plant. Sanogo et al. [22] demonstrated the antipyretic activity, which confirmed the traditional use of this plant as an antipyretic agent. The complement activating effect was investigated by Diallo et al. [23] that found the leaf water extract to have an effect on the complement system with an IC50 of 45 μg/ml which may be related to the healing of burns and wounds.
Piliostigma thonningii Schum (Leguminosae: Caesalpinioideae)
Local Name: Niama
Traditional use
According to the traditional healers in Doila this plant is call "child remedy'' as it is mainly used as a remedy for children. The different indications of this plant are almost all related to children, except its use against arthritis, headache, haemorrhoids and backache. The most frequent of them according to the citation frequency is the use against malaria (40%) followed by the use against the children digestive disorder called abdominal flatulence (16%) and child malnutrition (8%). According to the local traditional beliefs,"preparation from P. thonningii is the first plant remedy given to a child in his life".
In the literature, the most frequent use of the bark of P. thonningii is in treating cough, usually as an infusion or by chewing of the bark. A common use in Uganda is to stop diarrhoea, dysentery and intestinal upsets. The bark infusion or maceration also enters into the treatment of malaria and leprosy; analgesic properties are described to the bark; preparations are also used for sore throat, toothache, stomach-ache and earache [24].
The leaf decoction is a laxative and is given to children [25]. The infusion is given to new born babies and used as a tonic embrocating to massage the mother's abdomen; it serves also as a lotion for lumbago. The leaves, after soaking in hot water, are applied topically as wound-dressing and a leaf decoction is applied to the excision wounds in the South-West Africa region [11].
Biological activities
Asuzu et al. [26] found that the D-3-O-Methylchiroinositol, the anthelmintic component of P. thonningii stem bark extract, induced approximately 60% larval paralysis within 24 h of contact with Haemonchus contortus larvae at 4.4 mg/ml. This level of activity confirms the use of P. thonningii stem bark extract to treat helminthiasis in African traditional medicine.
Akinpelu et al. [27] founded that P. thonningii stem bark extract, at a concentration of 20 mg/ml, exhibited an antibacterial activity against Bacillus subtilis, Staphylococcus aureus, Shigella dysenteries, Escherichia coli and Proteus ulgaris.
Cochlospermum tinctorium A. Rich. (Cochlospermaceae)
Local Name: N'Tiribara
Traditional use
According to our ethnopharmacological survey, this plant is mainly used against jaundice (42.4%) based on its citation frequency. It is also used against malaria (27.3%). The uses against abdominal pain, in wound healing, haemorrhoids, intestinal worms, bilharzias and hepatitis were also reported (Table 2, see additional file 1).
These results are quite similar to those found by Nergård et al. [7] during another survey performed on C. tinctorium in another area in Mali. In addition to the ailments mentioned above it was also reported to be used against gastrointestinal diseases such as ulcer and stomach ache, flatulence and constipation.
In Ivory Coast the root is used for oedematous conditions, for orchites, schistosomiasis, jaundice, fevers, epilepsy, pneumonia, intercostal pains, and bronchial affections, in eye instillations for conjunctivitis and for indigestion and stomach pain [28].
Biological activities
The ethanol extract of the root of C. tinctorium showed a pronounced activity against a chloroquine-sensitive (3D7) and -resistant (Dd2) Plasmodium falciparum strain with an IC50 of 2.3 and 3.8 μg/ml respectively [29].
The aqueous extract, ethanol and hydro-ethanol extract of the root of C. tinctorium significantly inhibited the toxic effect of tert-butyl hydro peroxide-induced malonaldehyde formation in isolated rat hepatocytes at doses equivalent to 1 mg of dried plant material per ml of cell suspension. They exhibited marked effect against induction of lipid peroxidation and hepatocyte lysis. This result showed the hepatoprotective effects of these extracts [30].
In the anti ulcer test Nergård et al. [7] found that the oral administration of 25, 50 and 100 mg/kg of the crude extract of C. tinctorium 1 h before the HCl/Ethanol treatment significantly reduced the occurrence of mucosal gastric lesions in mice. This result confirms the reported use against gastro-intestinal diseases by the traditional healers.
Plant parts used and mode of preparation
The leaves are the most frequently used plant part (56.3%), the root and fruits are used about 30% and 8.5% respectively, and the less used plant part is the bark (5.3%). In another study of plants used for wound healing in Dogonland (Mali), Inngjerdingen et al. [5] found that the leaves and the roots were the most frequently plant parts used, constituting about 22 and 24% of the preparations, respectively, followed by the stem bark and fruits (12% each). In our study the exception were for Cochlospermum tinctorium, almost all traditional use reported for this plant was related to the root (about 94% of the number of citation of the plant). The leaves of this plant were used for the treatment of malaria and jaundice, a result that is similar to those obtained by Nergård et al. [7] on the use of Cochlospermun tinctorium in Mali. The root was in that report also the part of the plant most frequently used (95%), while the leaves were used by a minority of the healers for the treatment of malaria, ulcer and flatulence, the flowers were used by only one healer in the treatment of constipation. In our study no use of flower was reported for Cochlospermun tinctorium.
According to one of the traditional healers, the need for the use of stem bark will increase when the leaves are not available. The study area has a wet season from June to September, the main rainy season, and a dry and windy season from February to May. During this period most of the plants do not have leaves and the use of stem bark and dry plant parts is for this reason frequent. The fruits are also used, but not often due to their short time of availability.
The majority of the remedies are prepared in the form of decoction of fresh leaves. In our study area people do usually not store remedies for prolonged period of time. When needed they go out and collect the plant and prepare the remedy from fresh or sun dried material.
The powders are prepared by pounding the fresh plant part or the crushed plant material after sun drying, in a wooden mortar.
Water is the most frequent liquid used in preparations; powders are sometimes suspended in milk or consumed in food such as porridge and sauce.
Decoction is the most frequent way of preparation of the remedies (65%) followed by infusion (13%), which is used for the powders; the maceration (11%) is mostly used for the root preparation. Some remedies are prepared from a single plant species, but in a few cases mixtures of plants or other substances are added as can be seen from Table 2 (see additional file 1).
Route of administration and dosage
Most of the remedies are taken orally and by external application as body bath, steam bath, and as ointment in the case of dermatitis. Some remedies used for the treatment of haemorrhoids and genital infection are used as enema. Fumigation is mainly used in the treatment of headache and chest pain.
According to some healers certain additives are frequently used to improve the acceptability of some remedies that are taken orally. This can be salt, which is often mixed with powder; sugar or honey is added to decoctions and macerations to reduce the bitterness of the remedies in order to make them easer to drink. Lack of data on the biological role of these additives has been notified in the literature. The plants cited in the remedies are generally evaluated by biological screening, but the additives mentioned by the healers are normally not investigated.
Some healers reported that restrictions are imposed when certain types of remedies are taken by patients. For example, food is not given from the morning until noon to a patient who is taking a remedy against intestinal worms. This is the estimated time for getting diarrhoea which expulses the worm from the intestine. It is believed that food will reduce the efficacy of the remedy.
For most of the remedies, the dose given to the patient depends on age, physical and health condition of the patient, and the duration of the illness. The doses vary from a teacup (70 ml) for adults to a handful (25 ml) for a child; a lack of agreement among the healers on doses of remedies was sometimes noted. The variation of the doses from one healer to another may show that the plants have a low degree of toxicity. For pharmacological investigation the active doses of these plants may not be high since they appear to treat the patients with low doses. The duration of treatment is not given for all remedies. According to the healers duration of treatment is difficult to determine and depends on how long the patient has been ill. The patient is supposed to take the remedy until healed, and then, the only one able to determine the end of a treatment is the patient himself since the remedy is taken at home in the absence of the healers.
Adverse effects
The reported adverse effects for the use of these seven medicinal plants are diarrhoea, vomiting and dizziness. According to the healers these effects are generally due to an overdose of the remedy. Sometimes the expected effect of the remedy is diarrhoea, such as in the cases of constipation and intestinal worms. For intoxication treatment, the patient is supposed to eliminate the poisons by vomiting. In a survey of toxic plants on the market in the district of Bamako [31], Trichilia emetica was reported to be a toxic plant. The toxic effects reported were diarrhoea and vomiting. In our study no toxic effect was reported, but the same symptoms (diarrhoea and vomiting) were qualified as adverse effects for some remedies. According to the healers the adverse effects are generally moderate, and disappear at the end of the treatment. The commonly used remedy against diarrhoea and vomiting is a cold bath; another plant part can be used when the adverse effect is violent and stopping of the treatment is recommended.
Conclusion
Our ethnopharmacology survey showed that medicinal plants are still widely used by the population in the area where the study was conducted. It allowed us to report 50 different diseases or ailments treated by the seven medicinal plants included in this survey. Several types of preparations of these plants were used. The plants grow over an extended area and are used by healers separated by long distances. This may explain the many different types of uses observed. The healers' consensus in the treatment of the main reported diseases is fairly high, giving an additional validity to the plants as a traditional remedy.
This study complements the on-going activities of evaluation of different uses of medicinal plants and the development of new Improved Traditional Medicine by the Department of Traditional Medicine in Mali. Pharmacological, toxicological and phytochemical studies will be carried on these plants in order to ascertain the effectiveness as well as the possible toxicity of the remedies followed by designing therapeutic strategies based on the most effective and least toxic plants.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AT, DD and SD performed the interviews with the healers and identified all plant material described.
AT, HB and BSP drafted and finalised the manuscript.
Supplementary Material
Additional file 1
Click here for file
Acknowledgements
We are very grateful to all the local healers who shared their knowledge on the use of plants with us. Without their contribution this study would have been impossible.
This study is a part of the NUFU project PRO 22/2002.
==== Refs
WHO website
Diallo D Paulsen BS Svarstad H, Dhillion S Pharmaceutical research and traditional practitioners in Mali: Experiences with benefit sharing Responding to Bioprospecting from biodiversity in the South to medicines in the North 2000 Printed in Norway 133 144
Macia Manuel J Garcia E Vidaurre Prem J An ethnobotanical survey of medicinal plants commercialized in the markets of La Paz and El Alto, Bolivia Journal of ethnopharmacology 2005 97 337 350 15707774 10.1016/j.jep.2004.11.022
Diallo D Sogn C Samake FB Paulsen BS Michaelsen TE Keita A Wound healing plants in Mali, The Bamako region. An Ethnobotanical survey and complement fixation of water extracts from selected plants Pharmaceutical Biology 2002 40 117 128
Inngjerdingen K Nergard CS Diallo D Mounkoro PP Paulsen BS An Ethnopharmacological survey of plants used for wound healing in Dogonland, Mali, West Africa Journal of ethnopharmacology 2004 92 233 244 15138006 10.1016/j.jep.2004.02.021
Nergard CS Diallo D Michaelsen TE Malterud KE Kiyohara H Matsumoto T Yamada H Paulsen BS Isolation, partial characterisation and immunomodulating activities of polysaccharides from Vernonia kotschyana Sch. Bip. ex Walp Journal of Ethnopharmacology 2004 91 141 152 15036481 10.1016/j.jep.2003.12.007
Nergard Cecilie S Diallo D Inngjerdingen K Michaelsen Terje E Matsumoto T Kiyohara H Yamada H Paulsen BS Medicinal use of Cochlospermum tinctorium in Mali Anti-ulcer-, radical scavenging- and immunomodulating activities of polymers in the aqueous extract of the roots Journal of ethnopharmacology 2005 96 255 269 15588678 10.1016/j.jep.2004.09.018
Alexiades MN ed Selected Guidelines for Ethnobotanical Research: A Field Manual Advances in Economic Botany 10 The New York Botanical Garden, Bronx
Burkill HM The Useful plants of West Tropical Africa 1997 4 2 Royal Botanic Gardens, Kew
Shihata IM El-Gendi AYI Abd El-Malik MM Pharmacochemical studies on saponin fraction of Opilia celtidifolia Planta Medica 1977 31 60 67 557210
Burkill HM The useful plants of West Tropical Africa 1995 3 2 Royal Botanic Gardens, Kew
Okoli AS Iroegbu CU Evaluation of extracts of Anthocleista djalonensis, Nauclea latifolia and Uvaria afzalii for activity against bacterial isolates from cases of non-gonococcal urethritis Journal of ethnopharmacology 2004 92 135 144 15099860 10.1016/j.jep.2003.12.015
Saidu K Onah J Orisadipe A Olusola A Wambebe C Gamaniel K Antiplasmodial, analgesic, and anti-inflammatory activities of the aqueous extract of the stem bark of Erythrina senegalensis Journal of ethnopharmacology 2000 71 275 280 10904174 10.1016/S0378-8741(00)00188-4
Kone WM Atindehou KK Terreaux C Hostettmann K Traore D Dosso M Traditional medicine in north Cote-d'Ivoire: screening of 50 medicinal plants for antibacterial activity Journal of ethnopharmacology 2004 93 43 49 15182903 10.1016/j.jep.2004.03.006
Kerharo J Adams J La Pharmacopée Senegalaise Traditionnelle. Plantes médicales et Toxiques 1974 Edition Vigot et Frères, Paris
Burkill HM The useful plants of West Tropical Africa 1985 1 2 Royal Botanic Gardens, Kew
Reddy JS Rao PR Reddy Mada S Wound healing effects of Heliotropium indicum, Plumbago zeylanicum and Acalypha indica in rats Journal of ethnopharmacology 2002 79 249 251 11801388 10.1016/S0378-8741(01)00388-9
Kugelman M Lui W-C Axelrod M McBride TJ Rao KV Indicine-N-oxide: the antitumor principle of Heliotropium indicum Lloydia 1976 39 125 128 948236
Malgras D Arbres et arbustes guérisseurs des savanes maliennes 1992 Edits. Karthala, Paris
El Tahir A Satti GM Khalid SA Antiplasmodial activity of selected Sudanese medicinal plants with emphasis on Maytenus senegalensis (Lam.) Exell Journal of ethnopharmacology 1999 64 227 233 10363837 10.1016/S0378-8741(98)00129-9
McGaw LJ Anna KJ van Staden J Prostaglandin synthesis inhibitory activity in Zulu, Xhosa and Sotho medicinal plants Phytotherapy Research 1997 11 113 117 10.1002/(SICI)1099-1573(199703)11:2<113::AID-PTR27>3.0.CO;2-S
Sanogo R Germano MP D'Angelo V Forestieri AM Ragusa S Rapisarda A Trichilia roka Chiov. (Meliaceae): Pharmacognostic researches Farmaco 2001 56 357 360 11482758 10.1016/S0014-827X(01)01051-5
Diallo D Paulsen BS Liljeback THA Michaelsen TE The malian medicinal plant Trichilia emetica ; studies on polysaccharides with complement fixing ability Journal of Ethnopharmacology 2003 84 279 287 12648827 10.1016/S0378-8741(02)00330-6
Aubreville A Flore Forestiere Soudano-guinéenne, Afrique Occidental Française -Cameroun- Afrique Equatorial Française 1950 Societé d'Edition Geographiques Maritimes et Coloniales, Paris
Bouquets A Plantes medicinales du Congo-Brazzaville Uvariopsis, Pauridiatha, Diospyrsos, Trav Doc ORSTOM 1972
Asuzu IU Gray AI Waterman PG The anthelmintic activity of D-3-O-methylchiroinositol isolated from Piliostigma thonningii stem bark Fitoterapia 1999 70 77 79 10.1016/S0367-326X(98)00006-9
Akinpelu DA Obuotor EM Antibacterial activity of Piliostigma thonningii stem bark Fitoterapia 2000 71 442 443 10925021 10.1016/S0367-326X(00)00136-2
Kerharo J Bouquets A Plantes medicinales et toxiques de la Côte d'Ivoire-Haute Volta 1950 Vigot Frères, Paris
Zederkopff Ballin N Traore M Tinto H Sittie A Molgaard P Olsen CE Kharazmi A Christensen SB Antiplasmodial compounds from Cochlospermum tinctorium Journal of Natural Products 2002 65 1325 13257 12350157 10.1021/np020008h
Diallo B Vanhaelen M Kiso Y Hikino H Antihepatotoxic actions of Cochlospermum tinctorium rhizomes Journal of ethnopharmacology 1987 20 239 243 3682848 10.1016/0378-8741(87)90051-1
Maiga A Diallo D Fane S Sanogo R Paulsen BS Cisse B A survey of toxic plants on the market in the district of Bamako, Mali: traditional knowledge compared with a literature search of modern pharmacology and toxicology Journal of ethnopharmacology 2005 96 183 193 15588669 10.1016/j.jep.2004.09.005
|
16270940
|
PMC1277087
|
CC BY
|
2021-01-04 16:40:12
|
no
|
J Ethnobiol Ethnomed. 2005 Sep 27; 1:7
|
utf-8
|
J Ethnobiol Ethnomed
| 2,005 |
10.1186/1746-4269-1-7
|
oa_comm
|
==== Front
J Ethnobiol EthnomedJournal of Ethnobiology and Ethnomedicine1746-4269BioMed Central London 1746-4269-1-81627094110.1186/1746-4269-1-8ResearchThe ethnobotany of Christ's Thorn Jujube (Ziziphus spina-christi) in Israel Dafni Amots [email protected] Shay [email protected] Efraim [email protected] Institute of Evolution, University of Haifa, Haifa, 31905, Israel2 Dep. of Erets Israel Studies, University of Haifa, Haifa, 31905, Israel2005 28 9 2005 1 8 8 28 8 2005 28 9 2005 Copyright © 2005 Dafni et al; licensee BioMed Central Ltd.2005Dafni et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This article surveys the ethnobotany of Ziziphus spina-christi (L.) Desf. in the Middle East from various aspects: historical, religious, philological, literary, linguistic, as well as pharmacological, among Muslims, Jews, and Christians. It is suggested that this is the only tree species considered "holy" by Muslims (all the individuals of the species are sanctified by religion) in addition to its status as "sacred tree " (particular trees which are venerated due to historical or magical events related to them, regardless of their botanical identity) in the Middle East. It has also a special status as "blessed tree" among the Druze.
IsraelethnobotanyChrist's Thorn JujubeZiziphus spina-christiholy treesacred tree
==== Body
Introduction
Christ's Thorn Jujube (Ziziphus spina-christi (L.) Desf. [Rhamnaceae] is a tropical evergreen tree of Sudanese origin. It grows in Israel in all valleys and lowlands, and usually is confined to low elevations below a.s.l. 500 m [1].
The tree and its parts appear to have been in use in Pharaonic industry (carpentry), diet, and in medicine. The fruits were sometimes made into bread, which may also have been used for dressings when warm. Egyptian peasants made similar bread as late as the beginning of the 20 th century [2].
The Christ's Thorn Jujube has been mentioned in classical sources. The Greek botanist. Theophrastus (4th–3rd centuries BC) wrote, "The (Egyptian) 'Christ Thorn is more shrubby than the lotos (might be Ziziphus lotus (l.) Lam.); it has a leaf like the tree of the same name of our country, but the fruit is different, for it is not flat, but round and red, and in size as large as the fruit of the prickly cedar or a little larger; it has a stone which is not eaten with the fruit, as in the case of the pomegranate, but the fruit is sweet, and, if one pours wine over it, they say that it becomes sweeter and that it makes the wine sweeter" [3]. Pliny (1st century AD) mentions the tree in comparison with related species: "The region of Cryonic ranks the lotus below its own Christ-thorn" [4].
This common species is frequently mentioned in Christian as well as Muslim traditions, and was also recorded by pilgrims who visited the Holy Land during generations. We may therefore say that this species is "well soaked" in the local folklore as well as the ethno medicine of almost all the ethnic groups living in the Land of Israel.
Botanists expert in the Bible are constantly engaged in a great debate about what constitutes the "bramble" or "thorns" (Judges 9; 14–15), "thorns" (Matthew 27:27–29) and the "crown of thorn" (John 19:5). Based on local traditions and old sources, today these citations are commonly deemed to refer to Z. spina-christi [5-7].
The Quran mentioned the tree twice (LIII: 13–18; LVI: 28–32); the lote-tree is commonly identified as Z. spina-christi [8] and accordingly this species is highly respected by the Muslims through the Middle East. This tree has been widely used as a fruit plant and as a medicinal plant since antiquity and is still in use at present.
The aim of this paper is to review the current ethnobotanical status of Z. spina-christi in Israel, based on our field studies, in relation to historical and current literature.
Materials and methods
The field study (1999–2004) centered on Arabic villages in Galilee. Informants were asked about the ritual importance of the plant, why it is respected, which parts are used, and for what purposes. The survey covered 92 informants, consisting of 38 Druze, 54 Muslims (36 Arabs and 18 Bedouins). The informants were mainly chosen according to their knowledge of common traditions and/or religious status. The average age of the informant was 56 years (+/-14 years). The respondents were 90 males and two females (the women were interviewed in the presence of other family members). The question asked was: "What are the significance, uses, traditions, and stories you know about the Christ's Thorn Jujube?" Complementary questions on other trees were introduced only after the informant had expressed his or her view.
The list of medicinal uses during the medieval period was compiled from a survey of written medieval sources [9,10]; the list of medicinal uses of present-day ethnic groups in Israel is based on an ethnobotanical survey [11], an ethnopharmacological survey [12], and other surveys that have been conducted in the Middle East. Medicinal uses mentioned by Palevitch et al. [11], which were recorded only from one or two informants, were validated in this survey.
The plant names
The plant is named sheisaf in Hebrew, and a few Bible commentators have identified the tree with the "atad" (Job 40:21–22), identified otherwise with Lycium sp. [bramble, thorn bush, boxthorn], "n'atsuts", and even the "tse'elym" [5,13].
In rabbinical literature, the plant is called rimin" (Mishna, Demai, 1:1; Kilayim, 1:4), and in the Talmud it is called "kanari" (Bab. Talmud, Baba Bathra, 48b). It may be that it was so named because it is widespread around Lake Kinneret – the Sea of Galilee (Bab. Talmud, Mgillah, 6a).
Several common Arabic names are still in use today: "nabq, dum, sidr, tsal, sadr [[14-16] and [17]]." Sidr" serves as the common name for lotus jujube Z. lotus, which is also named "rubeida" after its crouch-shaped treetop. The names are used interchancheably in various geographical areas such as Lower Galilee.
In Christian tradition the tree was identified with the thorn bush with which Jesus was crowned before his crucifixion (Matthew 27:28–29; John 19:5; Mark 15:17). This is also the source for the scientific name (spina-christi).
The tree is rare in the vicinity of Jerusalem (A. Shmida, personal communication 10 May 2004). But Henry Baker Tristram wrote that he saw a tree in the Kidron valley, outside the city, albeit in the form of a small bush [18]. Tristram gave both the Arabic and the scientific name; so presumably, he was closely familiar with the species. The debate over the identity of the "crown of thorns" in the New Testament is long-lived, and various plants have been suggested as candidates [6,7,19].
Islamic sources
The Qur'an says, "And verily he saw him yet another time. By the lote-tree of the utmost boundary. Nigh unto which is the Garden of Abode. When that which shroudeth did enshroud the lote-tree. The eye turned not aside nor yet was overbold. Verily he saw one of the Greater Signs revelations of his Lord" (LIII: 13–18, Pickthall edition).
The only other reference to the lote-tree is in the sura of the Event, namely the Day of Judgment, when those at Allah's right hand, that is, the faithful, will dwell "among thorn less lote-trees and clustered plantains, and spreading shade, and water gushing, and fruit is plenty" (Qur'an, LVI: 28–32, Pickthall edition).
Farooqi, in his book "Plants of the Qur'an" discusses at length the different names of the Qur'an's lotus tree: he suggests Z. spina-christi as an option, but on the other hand Z. lotus and Z. spina-christi are wild plants in Arabia. Another possibility he mentions is the Lebanon cedar (Cedrus libani L.), which is also called "sidr in Arabic. Farooqi concludes that the lotus tree of the Quran was indeed the Lebanon cedar, and the historical misunderstanding has perpetuated the mistaken name until the present day [8].
The continuing debate regarding the lotus tree in the Qur'an, not withstanding, the sacred tree found in the Middle East called sidr, the name given in the Qu'ran, is the Christ's Thorn Jujube (Ziziphus spina-christi).
Christ's Thorn Jujube in the medieval Levantine literature
Muslim as well as Christian pilgrims and travelers have described Z. spina-christi as a large tree that grew in the Land of Israel. The tree was usually recorded for its uses and as a symbol of holiness [13]. The pilgrims took branches of the tree back to their homeland as souvenirs in the belief that the Jesus's crown of thorns was made from such branches [20,21]. Estori ha-Parhi, for example, who visited the Land of Israel during the Mamluk period (13th–16th centuries), wrote that the "rimin" is the "nabaq" in the language of Egypt and "dum" in the land of Canaan, and is also the tree named "sidar" [22].
In medieval medical literature the jujube appears frequently under various names, such as "sidar" or "tsal", while the fruit is called "nabaq" or dum" [14,23].
Clear-cut evidence of the medicinal uses and the economic value of the tree in the Land of Israel during the medieval period are found in Temple Mount documents: sidar features in a list of medicinal substances sold by the "atarin" (medicine vendors) in the markets of Jerusalem during the Mamluk period [24].
Al-Qazwini cites earlier authorities in stating that if the seeds are soaked in rose water and than planted, the fruits of the future tree will smell like roses. Similarly, if soaked in honey and milk, the future fruits will be sweeter and better [25].
In the past, various species of jujube grew in the Land of Israel that bore excellent fruit [26]. Al-Muqaddasi lists the tree among the widespread crops cultivated in the district of Falastin [27], and it was a popular food among the inhabitants of Tiberias [27]. Another source notes that the women of Egypt and the al-Sham region [Levant] used to comb their hair with the "sidar" [28].
Christ's Thorn Jujube as a useful plant in present day Levant
The fresh and the dried fruits of the plant are edible and highly valued locally by Arabs as well as Bedouins [17,29-32]. Bedouins collect and dry the fruits for future use in the winter, making a thick paste to be used as bread [17], a practice known already in ancient Egypt [2]. Z. lotus L. is similarly used in Cyprus [33] and in Arabia [34].
The wood is heavy and durable, and serves for artistic woodwork, while the branches and trunk are used as firewood and high quality charcoal [17,30-32].
Ethnopharmacology of Christ's Thorn Jujube
The tree and its various parts have been an important source for pharmaceuticals since antiquity. Data on the medicinal uses of the plant are presented here in Table 1.
Table 1 Medicinal uses of Christ's Thorn Jujube
Illness/uses Parts and preparation Region/ethnic group and reference Historical references World references
Toothache, gum problems Rubbing the teeth/gums with root powder (or bark) Arabs, Bedouins [11]. Iraq [35]; Arabian peninsula [37].
Arthritis Paste of crushed roots, leaves or branches; branches and leaves – inhale steam Arabs, Bedouins [11, 36]. Arabian peninsula [37]. Dhofar [38].
General painkiller Paste of crushed roots or branches and flour is applied Arabs, Bedouins [11].
Anodyne Bark, leaves Iraq [35].
Muscle pains Branches and leaves – inhale steam Sinai and Negev Bedouins [36, 39].
Soothe pains Leaves Yemenite Jews [40].
Bruises Fruit, leaves, seeds Arabian peninsula [37]. Dhofar [38].
Chest pains, asthma Fruit, leaves, seeds X Century [41]. Arabian peninsula [37].
Headache Fruit, leaves, seeds Arabian peninsula [37]. Dhofar [38].
Heart pains Branches Sinai and Negev Bedouins [36, 39].
Eye inflammations Powder of seeds, green leaves or roots as cataplasm. Arabs, Bedouins [11]; Iraqi Jews [43]. Egypt (Bedouins) [42].
Stomach disorders: aches, constipation, heartburn. Decoction of seeds leaves or fruit is drunk. Arabs, Bedouins [11]. Ancient Egypt [2], (X Century [41, 45]. Iraq [35]; Morocco, [42]. Iberian Peninsula, 13th century [46]. Dhofar [38].
Anthelmintic Fruit, seed or leaf infusion Arabs, Bedouins [11], Iraqi Jews [43]. Morocco, [42].
Hemorrhoids Leaves Yemenite Jews [40]; Iraqi Jews [43]. XIII Century [25].
Diarrhea Fruit or leaf infusion Sinai and Negev Bedouins [36, 39], Yemenite Jews [40]; Iraqi Jews [43]. Ancient Egypt [2], X Century [41]; XIII [25]. Morocco, [42].
Increase milk production Leaves boiled and liquid drunk Sinai and Negev Bedouins [36, 39].
Promoting pregnancy Fruit – tea Sinai and Negev Bedouins [36, 39].
To ease prolonged labor Leaves boiled and liquid drunk Arabian peninsula [37]. Dhofar [38].
Wounds Application of fruit juice Iraqi Jews [43]; Arabs [11]. Ancient Egypt [2]. Dhofar [38].
Blisters Fruit, leaves, seeds Arabian peninsula [37].
Burns Fruit – crushed and boiled Iraqi Jews [43].
Skin diseases and disorders Boiled or crushed leaves, resin Iraqi Jews [43]. X Century [41]. Arabian peninsula [37]. Dhofar [38].
Abscesses and furuncles Cataplasm of boiled leaves Morocco, [42].
Lung, chest and pectoral problems Leaves or fruit Iraqi Jews [43]. Iraq [35]; Arabian peninsula [37]; Iberian Peninsula, 13th century [46].
Blood purifier and tonic Leaves or fruit Yemenite Jews [40]. Ancient Egypt [2, 44]. Iraq [35]; Arabian peninsula [37]; Dhofar [38];
High blood pressure Leaves Israel [47, 48]. Jordan [12].
Fractures Cataplasm of boiled leaves Arabian peninsula [37].
Emollient Fruit or leaf infusion Iraq [35]; Arabian peninsula [37]; Morocco, [42].
Depurative Fruit Iraq [35].
Cooling Bark, leaves, fruit Ancient Egypt [2, 44]. Iraq [35].
Tonic Bark, leaves Ancient Egypt [2]. Iraq [35].
Stomachic Bark, leaves, fruit Ancient Egypt [2, 44]. Iraq [35].
Measles Fruit infusion Morocco [42].
Febrifuge Fruit infusion, resin X Century [41]. Morocco [42].
Snake bites Wood ash in vinegar Leaves for bee or wasp stings, XIIIth century [49]. Morocco [42].
Astringent Leaf infusion Morocco [42]; Iraq [35].
Hair problems Liquid from branches, fruit, leaves, seeds, resin. Arabs [11]; Iraqi Jews [43]. X Century [41]; XIII Century [25]. Southwestern Saudi Arabia [50]; Arabian peninsula [37]. Dhofar [38].
Infant's powder Powdered leaves Yemen [37].
Colds Fruit Israel [47, 48]. Jordan [12].
Weight reduction Fruit Israel [47, 48]. Jordan [12].
Nervousness Branches and leaves Negev [36].
Swollen organs Fruit Ancient Egypt [2].
Diuretic Wood Ancient Egypt [2].
Liver problems Fruits Ancient Egypt [2].
Anus problems Fruits Coptic Egyptian Medicine [2].
Christ's Thorn Jujube as a sacred tree
An old Muslim legend tells about a Christ's Thorn Jujube that grows in Paradise and has leaves as many as there are human beings. Each leaf bears the names of a particular person and his or her parents. Every year, one day in the middle of the month of Ramadan, just after sunset, the tree is shaken. The names on the leaves that fall are of those who face death in the coming year. The process of the leaves' decay intimates the timing of their death; some leaves dry up and fall immediately while others wither slowly, signifying the time the person has left [51-53].
This legend reflects the respect in which Muslims hold all Christ's Thorn Jujube trees, wherever they are. No wonder that the tree has received so much attention in Arab folklore in the Land of Israel in the past, and continues to do so to the present day.
The Christ's Thorn Jujube is considered a sacred tree in Israel. When the tree reaches its 40 th year, the saints sit under it; therefore, the saints will destroy anyone who dares to cut down the tree or one of its branches. One story tells that "every Thursday evening the music of some instrument could be heard coming from some Christ's Thorn Jujube trees. Another story told and recorded in the Holy Land relates that lights were seen every Thursday night among the branches of few trees near "N'an'a" (Na'an) and 'Aqir" ('Aqron) [54,55].
The presence of saints under Christ's Thorn Jujube trees imparted their holiness to the tree, as happens with other species of trees. However, no other tree species are mentioned in the Holy Land as preferred by the saints.
Goldziher [56] sheds light on the importance and the great honor the Christ's Thorn Jujube has among the Arab population in the Holy Land. He cites the Abbé Barges [57] who described a large tree that grew in the garden of an Arab's house in Jaffa. This tree was treated with a special reverence by the local Muslims, who would hang colored cloths and lamps on it. The owner of the tree explained this kind of worship as the belief that the seeds from which the tree grew had fallen from the sky so the tree was sacred to the Prophet, who visited it at night. He added that all "good Muslims honored the tree". We assume that it was a Christ's Thorn Jujube since cloths and lamps were hung on its branches (Z. lotus is too low for this purpose), furthermore, Jaffa is far from the natural habitat of the lotus jujube.
We recorded a similar story in Kabul, a village in Western Galilee: "Lamps were lit every Thursday evening among the branches of the big Christ's Thorn Jujube tree which was in the village. Then the Sufi dervishes held their zikr ceremony (a special Sufic dance meeting) under this tree. They gathered there from various villages thereabouts, and the tree had its own sheikhs, who were not known in our village. The villagers were afraid to approach the tree, except for one old lady who would bring food and meat for the dervishes, as a gift from the local folk. The tree stood in the village until the 1950s (AĦmad Ţaha Yāsīn, Kabūl, 6 June 2004).
At times Christ's Thorn Jujube trees were used to mark the borders between estates of neighboring villages according to the common belief that the fence around Paradise was build of the wood of Christ's Thorn Jujube [55].
The special attitude to the Christ's Thorn Jujube in the Holy Land can be summarized and explained as the traditional belief that the tree should be esteemed and respected since it was probably the host of certain saints or other spirits [17]. In the words of Ţāhir 'Abu 'Antar (Ţamra, 14 June 2004), "The sidr tree is like a sheikh", and you have to pay it respect as you would elderly people.
In modern Islam, sitting under a Christ's Thorn Christ's thorn Jujube tree is considered lucky, since the Prophet saw such tree in Paradise [58]. This idea might underlie the traditional belief that a potion made of Christ's Thorn Jujube leafs is the best supernatural remedy to expel demons ('Ădil Abu Hamīd and Ibrāhīm QadaĦ, Kufr Manda, 3 June 2004; Abu Amīn Xāldi, Xawālid, 26 August 2004; Sāmya Hādi, Mazra'a, 24 August 2004, MaĦmūd Zir'ēni, 22 November 2004, Tur'ān).
In the village of Mġār several informants were recorded commenting on the fruit of two sacred Christ's Thorn Jujube trees, named after Sheikh Rabīs and Nabi Shu'eib: "You will never find worms [in the fruit of these two trees], as you do in the fruit of other similar trees" (i.e., of this species). This was due to the sacredness of these two trees, which were blessed (Muhra Bahajāt, Qāsim Fādhil, Şalah Fādhil, Mġār, 19 May 2001). According to another informant (Ġāsam MuĦammad 'uqabi, Ţūba-Zanġariyya, 16 June 2004), "The fruit of the Christ's Thorn Jujube was eaten by Muslim fighters of the early Islamic period; therefore the tree is honored and may not be uprooted".
It is common, then, to find Christ's Thorn Jujube trees serving as sacred trees in many villages and at sheikhs' graves all over the Holy Land, but mainly in Upper Galilee (Sheikh Rabīs and Nabi Shu'eib, Mġār, Rabbi Avdimi grave in Haifa (cut at 2003), Sakhnīn, Sheikh Radwān near Nahariya, Sajarāt al-'Arūsa near Kābri).
In the Negev, barren women had to make a pilgrimage to a sacred Christ's Thorn Jujube [36].
Christ's Thorn Jujube proverbs
Demons (genies) avoid the Christ's Thorn Jujube tree because of its sanctity, and this precisely is why it is "good" to sleep under it (MuSţafa Kamirat, Ibtin, 13 January 2003; Ibrāhīm QadaĦ, Kufr Manda, and 3 June 2004; 'Ali Sulaymān Xuţba, 'Arrābe, 6. June 2004; Yusuf Nimmr Masar, Sakhnīn 1 January 2005). The following Arab proverb signifies that Christ's Thorn is blessed while the Carob tree (Ceratonia siliqua L.) is considered cursed: "innōm biĦlu taĦt iddōm, innōma taĦt ilxarrūbe mush marghūbe", which means "the sleep under Christ's Thorn Jujube tree is sweet and the sleep under Carob tree is not desirable" [59] and thirteen informants in our survey. The Carob tree is associated with a bad luck, and sitting under it is considered dangerous, especially at night since the tree is a dwelling place for bad spirits. The red color of the leaf petioles, which resembles blood, is a sign of bad luck as well [17].
One explanation for why the Carob is cursed is this: "There is always a chance of finding a snake in the trunk and therefore you may not sleep under it; bees dwell in its trunk and branches as well" ('Ali Khalil Musa Kna'ane, 'Arrābe, 6 June 2004; 'Ali Sulaymān Xuţba, 'Arrābe, 6 June 2004: Yusuf Nimmr Miser, Sakhnīn 1 January 2005). Another informant, Ibrāhīm QadaĦ (Kufr Manda, 3 June 2004), added more information regarding the dangers sleeping under the Carob tree: "The demons gather under Carob trees and every once in a while God punishes them by striking them with lightning. These trees are big, and attract lightning anyway". Other informants reported that snakes and demons like the Carob tree, and therefore it is cursed; people who slept under this tree went mad (Atef Mansur, Kaukab Abu-el-Heija, 13 May 2003; Hassan Jadir, Bir El Maksur, 30 December 2003; Nassr Khalil, Sakhnīn, 1 January 2005). One informant recounted "The Carob is a bad tree, the snakes that dwell in its branches hurt people. However, snakes that dwell in the Christ's Thorn Jujube will never harm a human being" (Sāmya Hadi (Mazra'a, 24 August 2004). Because of their black color, the Fig tree and the Carob tree are both considered cursed and the cause of misfortune [54]. The following proverb expresses this feeling: " Xarrūbe wa ttīn – maskanat iššayaţīn", means "Carob and Fig tree – (are) Satan's residency" (Abu-Amin Xāldi, Xawālid, 26 August 2004; 'Ali Khalil Musa Kna'ane, 'Arrābe 6 June 2004). Canaan likewise explains that both trees (Carob and Fig) bear black fruits, and they are the preferred dwelling places of demons. The genies gather under them for their meetings and nightly parades [59].
Additional explanations of why it is "good" under the Christ's Thorn Jujube are these: "The Christ's Thorn Jujube is a tree from heaven and therefore it is good to sleep under it" (MuHammad Ţāher, Rummāna, 22 November 2004; Abu-Razz, Bu'eina-Nujeidāt, 16 August 2004; Yusuf Nimr Masar, Sakhnīn 1 January 2005); "In arid zones the Christ's Thorn Jujube is the only shady tree and therefore it deserves a special attitude" (Raja Xaţīb, Deir Ħanna, 2 August 2004).
Christ's Thorn Jujube and animism
Special honor is given to the Christ's Thorn Jujube in Iraq, even more than to the Palm. Uprooting of a tree that has already fallen is considered a sign of impending disaster; should a man cut down a Christ's Thorn Jujube tree, he will soon fall ill and die. The tree is thought to groan when it cut and its sap, red as blood, gushes out of the slashed trunk, justifying the idea that the tree has a life similar to a human's [60].
The belief that blood flows from trees organs has ancient roots. Ovid [61] tells of Erysichthon, king of Thrace, who commanded that a sacred oak tree dedicated to Demeter be cut down. The appeal of the Dryad that lived in the tree was in vain. The tree was chopped down, and she was doomed to die with the demise of her abode; Demeter's revenge was immediate and singularly cruel. The king was condemned to an eternal unsatisfied hunger.
The tradition of regarding the punishment of whoever touches a sacred tree universal, and it has been one of the main characteristics of tree worship throughout history. Stories about groaning or bleeding trees are common [62-64].
Among the Bedouins of the Negev (southern Israel), a similar tradition is known: the red sap drops that flow out of Christ's Thorn Jujube trees gives them a human essence because of its resemblance to blood. It is a short step from these phenomena to tree worship, animism and the worship of the saint's spirit that dwells in the tree [65].
The old people of Mġār used to tell about the Christ's Thorn Jujube tree of Nabi Shu'eib (see above) that bled when it was cut; the children of the village, being skeptical, would make small cuts in the tree to see if it really would bleed ('Issa Sakrān, Mġār, 16 June 2004).
A similar story explaining the sacredness of Christ's Thorn Jujube tree was told in the Negev in 1977:
"In a place where we used to live a long time ago, in the southern part of Israel, there was one Christ's Thorn Jujube tree (sidr) under which stains of red sap, similar to blood, were found every morning... The women would hang white pieces of cloths on [the tree], the men did not cut it, and the tree grew to be a very big, even bigger than the tent we sit in". The sacredness of the tree gushes from the presence of the spirits of the dead people that dwell there. The tradition is that Christ's Thorn Jujube trees are sacred wherever they are... The Bedouins' explanation for this phenomenon usually concerns the holy man or holy people that dwell in or under the tree [65].
In Morocco a story was recorded about a tree named after Sides Bumhadi. A man climbed up it to cut some branches, and a copious flow of blood, as if fifty bulls had been slaughtered, came out of the tree. The terrified man jumped down and stayed there stunned [66].
In Iraq women occasionally visited Christ's Thorn Jujube trees. They lit straw torches under the tree and put incense on charcoal. Before departing they would leave four lighted candles on the ashes. This ceremony was to heal sick family members [60].
Lighting candles and other ceremonies for healing people are typical to tree worship in the Muslim and the Christian world alike, among the different ethnic groups in Israel (our observations), and all over the world [67,68].
As part of the ceremonies in Iraq, green clothing were hung on the branches of the Christ's Thorn Jujube tree, and sometimes even offerings of food were left under it [60].
The custom of hanging clothes on trees in general and sacred trees in particular is world-wide. The idea is that the disease is transferred from the sick person's clothes to the tree [69]. Green clothes are used especially in the Muslim world, because the color green is sacred in Islam [70].
Other evidence of respect for Christ's Thorn Jujube trees in Islam is the following custom. In Iran [52], Iraq [60], India [71], and southwestern Saudi Arabia [50] the bodies of dead Muslims were washed with water in which Christ's Thorn Jujube leaves had been soaked. The purpose was to preserve the body and satisfy the angels [52]. A similar usage was recorded in Israel by one informant [11], and others described this as custom that was practiced in Israel in the past, and although rarely, even today ('Ădil Abu Ħamīd and Ibrāhīm QadaĦ, Kufr Manda, 3 June 2004; Abu Rāzi, Bu'eina-Nujeidāt, 16 August 2004; Sheikh AĦmad Abu 'Umar, Mjd ilKrūm, 28 June 2004).
Customs related to the tree
The grave of Sheikh ŞalāĦ, which is in the old Muslim cemetery 400 meters east of the al-Jazzār Mosque in Acre (Jehoshafat Street), is prominent mainly because of its green domed roof. Ten meters south of the building a huge Christ's Thorn Jujube tree once grew. Its trunk was a meter in diameter and the tree shaded many graves. But its roots caused damage to a few graves so it was cut down at April 2000. Many iron nails were found on the cut trunk, arranged in groups of three. Māhir Zahra (Acre, 12 April 2002), a researcher on the history of Acre, described and explained this phenomena:
"Women who felt the need to break an evil eye spell would conduct the following ceremony. The woman would come purified, a few days after her period, on Friday early morning before the call of the muezzin, and approach the tree. She was not allowed to talk to anybody from the time she had awoken that day. She had to wait near the Christ's Thorn Jujube tree with an hammer and nails; after each call of 'Allahu 'akbar, she had to drive a nail into the tree – three times in all".
In this cemetery the sounds from the Mosque were heard very well. The nails were hammered into exorcise the woman or one of her family members of the evil eye. She had to remain quiet and was not allowed to talk all the way back home. Then she had to wash again and purify herself and go to sleep.
This tradition has to do with the Suffis who arrived in Acre at the beginning of the 18th century from North Africa. The citizens of Acre have known of the tree and the nails tradition since the middle of the 18th century. According to 'Araf [72], "Such nails can be seen in many other old trees found near saints' graves, or sacred trees, such as the trunk of the tree in the Druze village Jat in the upper Galilee. This tree is named Sajarat Abu 'Arus. Yet when we visited this specific tree (April 2003) we found no nails in it at all! Occasionally nails were seen in the trunks of sacred trees, but these were usually species of oak and the nails did not appear in triplets.
Similar customs have been recorded in other countries in the Middle East, India, and Europe [62,64,73-75].
Old women in the village of Ţamra, Western Galilee, used to tell of a custom of hammering nails into big old olive trees to protect against the evil eye. They called them "nails in the eye of the devil" (AĦmad Ţaha Yāsīn, Kabūl, 14 June 2004).
The purpose of this custom was twofold: to "stick" someone to the sacredness by attaching him or her to the powers of the tree and to expel demons with the tree's help ('Ădil Abu Ħamīd. Kufr Manda, 3 June 2004).
An interesting piece of medieval evidence of a similar custom appears in the Cairo Genizah (14th century): "Beautiful trees around the grave, they smell good and nails are stuck in them" [76].
Hammering nails and hanging cloths are "tying" rituals, whereby the person seeks healing or a solution to problems by transferring his or her illness or problems to the tree, or to whatever object the cloths are hung on or nails hammered into. Such "tying" is one of the best known and commonest belief practices all over the world among Christians, as well as among Muslims and their predecessors in the Middle East [53,64,70,77]. This tradition still exists today in Europe (Belgium) [78]. In Egypt, nails driven into tree trunks signify the prayers of the believers. People come to sheikhs' trees to be cured of headache or other ailments. In asking the sheikh for help they hammer nails into the trunk and wind some of their hair around the nails [79]. A ceremony of this kind was recorded at sacred graves in Turkey [64].
In England (Cornwall) and Germany (Oldenburg), believers used to hammer nails into tree trunks "where the sun cannot reach" to heal toothache [80]. Variations of this were seen in Nepal and India (Samir Ţafiš, Beit Jan, 18 March 2002) and in Turkey [64,81].
In India the Emetic nut tree (Strychnos nux-vomica L.) is considered the prison of all demons. Occasionally such trees can be seen with trunks full of nails as a precaution against demons. If a demon or bad spirit dares to attack a human, the exorcist forces it back into the tree with a nail. With each driven nail the demon declares that it will not attack again. Nailing the demon into the tree trunk is the best way to give it a life sentence [82,83].
Sacred trees in the West Himalayan region are the object of a similar custom: travelers hammer nails into the trunk when passing by as a protective step against diseases, death, and any damage to their sheep, cattle, or crops. The explanation for this act, according to traditional beliefs, is that it dispels evil powers [84].
A square in central Vienna is named "Stock am Eisen", which means literally "iron on the stick". A glass case stands on one of the corners of the square containing a replica of a piece of wood in which some nails are stuck. A known tradition from the 16th century tells that any apprentice who completed his duties in the town would hammer a nail into a tree that grew in the square for good luck [73,85]. Some are of the opinion that this was a gypsy tradition introduced from India [74].
Returning to Acre, cutting down the "nails tree" in the city apparently entailed the sad result of the extinguishing a unique and rare custom in Israel. This ritual in Israel and many others concern sacred trees, were evidently a link in the cultural chain between India and Europe.
"Holy" vs. "Sacred" trees
Simmons [86] distinguishes tree rituals wherein a certain species of tree is considered holy, as in the case of the sacred fig [Bo tree] (Ficus religiosa L.), from ritual in which individual trees are sacred because of special characteristics or have won respect through their location in a holy place or their association with a holy person [86].
The veneration of trees in Israel mainly stems from events with saints that occurred near them, so different tree species, were sanctified. In Iraq, by contrast, the Christ's Thorn Jujube tree is worshiped because this tree is mentioned in the Qur'an.
Most trees belong to the first group above, so their botanical species is not relevant. We apply the adjective "sacred" to these trees; we use the epithet "holy" for trees that have gained special respect because their botanical species is part of a religious ritual.
As far as we know, in the Land of Israel, all known admired trees are "sacred", having won honor because of their physical location near saint's graves or their connection to the deeds of religious, military, or other admired figures [64,79,87]. We are not aware of any "holy" trees.
One citation we recorded from an informant sums up the matter: "The sacred trees are the tombstones – the memento of a saint or holy man, so their importance is relative to the man holiness. The holier the saint, the more sacred the tree" (Sa'īd MaĦamūd, YānuĦ, 25 May 2003).
Another explanation we recorded has this variation: "Blessed trees are memorials to unique figures in the Druze history and religion; because the Druze tradition forbids any tomb sign or offerings, special people are remembered by big sacred trees" (Sheikh Šāhīn Ħusayn, Beit Jan, 12 September 03).
Christ's Thorn Jujube tree in the Druze tradition
Among the Druze, "sacred" trees are called "blessed" because according to their tradition only humans can be sacred. However, sacred figures can transfer some of their special powers to a tree. God's blessing is thought pass to the holy man and is then transferred to the blessed tree [70].
The holiness of the Christ's Thorn Jujube tree among the Druze is a tradition beginning in Islam (it is the tree of the seventh heaven). When the prophet ascended to Paradise he reached the seventh heaven and the last tree, named "Sidrat al-Muntaha" (the sidr tree of the last frontier).
The Druze argues that the last Christ's Thorn Jujube tree symbolizes a certain figure and the story is about an imaginary spiritual trip in which the "sidr signifies an important and admired figure (ŞalāĦ Xaţīb, Mġār, 4 October 2003).
On the main road to Nabi Shu'eib, one of the holiest places for Druze in Israel (believed to be the grave of the prophet Jethro); there is a huge Christ's Thorn Jujube tree. In the past this important tree served as a meeting point for pilgrims before approaching the holy place. Whoever arrived first waited for the others under that tree. Over the years the tradition of the first meeting point took root, and the tree became a station for praying as well. When the pilgrims reached the tree they became very excited, and this is how the tree was named "Sidrat Nebi Shu'eib" (Abu Jalāl Qizl, ŞalāĦ Xatib and 'Ali 'Araydi, Mġār, 19 May 2001). The tale says that this tree grew out of a runner of a Christ's Thorn Jujube tree named "Sheikh Rabis (a huge "blessed" Christ's Thorn Jujube tree that grows in the middle of Mġār, two kilometers away from this tree). The person Sheikh Rabis (to whom the tree is dedicated) was a Muslim, but Druze's and Christians admire him as well. One informant explained: "As the Druzes came out of Islam, the Druzes' blessed tree came out of the Muslim one" (ŞalāĦ Xaţīb, Mġār, 4 October 2003).
Conclusion
The Christ's Thorn Jujube tree is widespread in various parts of the Land of Israel. In fact, its distribution is very broad and goes beyond desert areas. It also grows along the coastal plain and around Jerusalem. Throughout history, but mainly in the Middle Ages, the tree was seen as sacred [74] and was used for food [27]. Evidence of medicinal use of the Christ's Thorn Jujube in our region is very rare; we were able to trace one source, a list of substances sold in Jerusalem by medicine vendors in the Mamluk period (14th–15th century) [24]. The current medicinal uses in Israel are the same as in other parts of the Middle East (Table 1). We are unable to show any indication that some uses are endemic or more prevalent in one or another ethnic group.
Tree worship is deeply rooted in human culture and attitudes to the Christ's Thorn Jujube are a good example of these traditions. Eliade has taught us about the important religious role these trees play in different cultures. In almost every culture trees are presented as symbols of life, ongoing fertility and eternity, as well as resurrection. Trees are not admired for themselves, but because of what was discovered through them, what they symbolized and expressed [89]. Canaan puts a similar argument: "The villagers (of the Holy Land) do not admire the trees, but the divine powers acting through them and derived from the imaginary figure of God, which his soul is probably still in a temple, cave or spring to which the trees are connected. Occasionally the holy man is revealed in a tree or around it" [59].
Frese and Gray [90] emphasize that trees are an embodiment of nature, which represents the holy continuity of spiritual worlds, cosmic as well as physical. The tree sometimes symbolizes divinity or another holy entity or even a holy object.
Tree worship was prevalent in pre-Islamic pagan Arabia. Trees were frequently considered the abode of genies. The pagan Arabs associated divine characteristics with certain trees and worshiped them through special rituals, which included the hanging of colored cloths, ornaments, and weapons on their branches [58,91,92].
When the new religion of Islam evolved a war of annihilation was waged between the new Muslims and the pagans. This war involved the cutting down of sacred trees by the Muslims and strict prohibition against any form of tree worship [52,56].
This short review shows that tree worship prevailed in Arabia in the pre-Islamic period. Despite the battle against the pagan practice, some of the ancient rituals survived and still prevail in the Middle East [70].
The Christ's Thorn Jujube tree won unique honor because of a mention in the Qur'an; however, this species enjoys no special significance, botanical uniqueness, or exceptional size (compared with the great oak trees in Europe) which form the basis for tree worship in pagan Europe, including the Greek and Roman cultures) [93,94].
Most of the known traditions and rituals that were recorded in relation to the Christ's Thorn Jujube in the Middle East, such as candle lighting, incense burning, cloth hanging, and nail hammering, are known to be part of worshiping sacred trees in Israel and around the world. A similar case is the supernatural powers of the sacred tree and the fear of punishment as a result of cutting down the tree or dishonoring it. These customs do not relate to a particular tree species and are not found in Islamic lands alone.
The supernatural qualities attributed to the Christ's Thorn Jujube, such as blood flowing in its veins, sounds it makes when it is cut, and its being the abode of a saint's spirit or ancestors, are well documented in the literature of sacred trees.
The custom of washing the dead with leaves of the Christ's Thorn Jujube seems unique to this species and to Islam; we could not trace any similar custom elsewhere. We would like to suggest that this might be evidence of the uniqueness of the Christ's Thorn Jujube tree.
In light of the distinction between "holy" and "sacred" trees, we maintain that the Christ's Thorn Jujube is the closest species to a holy tree in Israel and the Middle East, as well as being a sacred tree. The holiness originated with the citation in the Qur'an, and the sacredness arose from events in the lives of saints, heroes, and holy men that happened near the trees.
Acknowledgements
The authors express their deepest thanks to Prof. Simcha Lev-Yadoon, Prof. Peter Bernhardt and Dr. Idit Pintel-Ginsburg for their critical comments, to Moris Zemach for translations and Dr. Aharon Geva for the Arabic transcriptions.
==== Refs
Zohary M Flora Palaestina 1972 II Jerusalem: The Israel Academy of Science and Humanities 307 308
Manniche L An Ancient Egyptian Herbal 1989 London: British Museum Press 157, 158
Theophrastus Hort A, Loeb Classical Library Inquiry into plants 1916 IV London and Cambridge (Massachusetts): Harvard University Press 3
Pliny Rackham H, Loeb Classical Library Natural History 1945 XIII London and Cambridge (Massachusetts): Harvard University Press 111
Löw I Die Flora der Juden 1922 III Vienna-Leipzig: Kohut Foundation 134 136
Zohary M Plants of the Bible 1982 Cambridge: Cambridge University Press 154, 155
Moldenke HN Moldelke AL Plants of the Bible 1952 New York: Ronald Press 248, 249
Farooqi A Plants of the Qur'an 1997 Lucknow (India): Sidrah Publishers 65 74
Lev E Reconstructed materia medica of the Medieval and Ottoman al-Sham Journal of Ethnopharmacology 2000 80 167 179 12007707 10.1016/S0378-8741(02)00029-6
Lev E Medicinal Substances of the Medieval Levant 2002 Tel Aviv: Eretz, (Hebrew)
Palevitch D Yaniv Z Dafni A Friedman J Ethnobotanical Survey of the Flora of Israel as a Source for Drugs 1985 Jerusalem: Ministry of Science (Hebrew) 159
Lev E Amar Z Ethnopharmacological survey of traditional drugs sold in the Kingdom of Jordan Journal of Ethnopharmacology 2002 82 131 145 12241988 10.1016/S0378-8741(02)00182-4
Amar Z Agricultural Production in the Land of Israel in the Middle Ages: An Agro-historical Study of Its Transformation PhD thesis 1996 Bar-Ilan University, Department of Land of Israel Studies (Hebrew) 287, 288 289
Ben Maimon Moshe (Maimonides) Meyerhof M Un Glossaire de Matiere Medicale Compose par Maimonide Memoires Presentes a l'Institut d'Egypte 1940 41
Dinsmore JE Dalman G Die Pflanzen Palastinas Zeitschrift des Deutschen Palastina – Vereins 1911 34 1 38 147–172, 185–241
Issa BA Dictionnaire des Noms des Plantes 1930 Le Caire: Imprimrie nationale 192
Crowfoot L Baldensperger GM From Cedar to Hyssop 1932 London: Sheldon Press 107, 112, 113
Tristram HB The Natural History of the Bible 1867 London: Society for Promotion of Christian Knowledge 429
Folkard R Plant Lore, Legend and Lyrics 1892 London: Samson Bow, Marston and Co 46
Amar Z Topics on the Settlement, Agriculture and Vegetation Landscape of Palestine during the Middle Ages MA thesis 1991 Bar Ilan University, Department of Land of Israel Studies (Hebrew) 136
Bellorini T Hoade B Bagatti B Visit to the Holy Places of Egypt, Sinai, Palestine and Syria in 1384 by Frescobaldi, Gucci, and Sigoli 1948 Jerusalem: Franciscan Press 121
Estori ha-Parhi (Isaac Ben Moses) Luncz AM Sepher Kaftor va-Ferah 1897 II Jerusalem: Luncz 526
Kohen al-Attar al-Israili (Abu al-Muna Dawud b. Abi Nasr) Minhaj al-Dukhan wa-dastur al-a'ydan fi a 'mal wa-tarakib al-adwiya lil-insan 1940 Cairo: Bulaq (Arabic) 247
Lutfi H Al-Quds al-Mamlukiyya 1985 Berlin: Klaus Schwarz Verlag 292
Al-Qazwini 'Ajab-al-Makhluqat 1981 Beirut: Dar al-Sharq al-'Arabi (Arabic) 235
Bar-Droma H The Negev 1935 Jerusalem (Hebrew) 464
Al-Muqaddas De Goeje L Ahsan al-Taqasim fi Marifat al-Aqalim 1906 Leiden: Bibliotheca Geographorum Aarabicum (Arabic) 181, 161, 204
Ibn Sīda Kitāb al-Muxşşaş 1902 XI Cairo: Būlāq (Arabic) 195
Bailey C Danin A Bedouin plant utilization in Sinai and the Negev Economic Botany 1981 35 1 95
Uphof JC Dictionary of Economic Plants 1968 New York: Cramer 563
Dafni A Edible Plants of the Land of Israel 1985 Tel Aviv: The Society for the Protection of Nature in Israel (Hebrew) 116
Chizik B Otsar ha-Tsmahim (A Treasury of Plants) 1952 Herzliya (Hebrew) 721
Georgiades CCH Flowers of Cyprus 1987 II Nicosia 97
Grieve M A Modern Herbal 1931 New York: Hafner 451
Al-Rawi A Chakravarty HL Medicinal Plants of Iraq 1964 Baghdad: Ministry of Agriculture Technology, Bulletin No. 146 99
Abu-Rabia A Traditional Bedouin Medicine 1999 Tel Aviv: Ministry of Defense (Hebrew) 82, 83, 90
Ghazanfar AS Arabian Medicinal Plants 1994 Boca Raton, Ann Arbor, London, Tokyo: CRC Press 182
Miller AG Strut-Smith S Plants of Dhofar: The Southern Region of Oman, Traditional, Economic and Medicinal Uses 1988 Oman: The Office of the Adviser for Conservation of the Environment, Diwan of Royal Court Sultanate of Oman 242
Levey S Medicine, Hygiene and Health among the Sinai Bedouins 1978 Tel Aviv: The Society for the Protection of Nature in Israel (Hebrew) 87
Reiani Y Medicinal substances of the Yemenite Jews MSc Thesis 1963 Hebrew University Jerusalem, School of Pharmacology (Hebrew) 55
Ibn Sina, Kitāb al-Qanun fi 'l-Tibb 1877 Baghdad, Beirut: Al-Munthanna Library (Arabic) 277, 278
Naury J Jouve Contribución a etude de la pharmacaine tradionnelle (drogues vegetals) These 1954 Univ Paris
Ben-Ya'akov A The Traditional Medicine of the Babylonian Jews 1992 Jerusalem: Yerid Hasfarim (Hebrew) 164, 336
Muntner S Assaph (Harofe) the Physician, 'Sefer Refuoth' Korot 1969 4 1967 38 Hebrew
Levey M al-Khaledy N The Medical Formulary of al-Samarqandi and the Relation of Early Arabic Simples to Those Found in the Indigenous Medicine of the Near East and India 1967 Philadelphia: University of Pennsylvania Press 123, 124
Amar Z Buchman Y (Eds) Nathan ben Yoel Falquera R Sori ha-Guf 2004 Ramat Gan: Division of the History of Medicine, Departments of Land of Israel Studies and Archaeology, Bar-Ilan University (Hebrew) 136
Lev E Amar A Ethnopharmacological survey of traditional drugs sold in Israel at the end of the 20th Century Journal of Ethnopharmacology 2000 72 191 205 10967472 10.1016/S0378-8741(00)00230-0
Lev E Medicinal Substances in Jerusalem from Early Times to the Present Day 2003 Oxford: Archeopress 124
Ben Maimon Moshe (Maimonides) Muntner S Poisons and Their Antidotes 1942 Jerusalem: Rubin Mass (Hebrew) 125
Abulafatih HA Medicinal plants in Southwestern Saudi Arabia Economic Botany 1987 41 354 360
Rodwell JM Koran 1937 London: J.M. Dent Trans
Donaldson BA The Wild Rue 1973 New York: Arno Press 141, 243, 244
Zwemer SM The Influence of Animism on Islam 1920 Macmillan, London 69, 213
Baldensperger PJ Peasant Folk-lore of Palestine Palestine Exploration Fund, Quarterly Statement 1893 203 219
Hanauer JE Folk-lore of the Holy Land: Moslem, Christian and Jews 1935 The London: Sheldon Press 216
Goldziher I Stern SM Muslim Studies 1971 London: George Allen and Unwin Ltd 317, 318
Barges JJ Vie du célèbre marabout Cidi Abou Médien 1884 44 Paris (cited by Goldziher 1971:370)
Reat R The tree symbol in Islam Studies in Comparative Religion 1975 64 182
Canaan T Plant lore in Palestine superstition Journal of Palestine and Oriental Studies 1928 8 129 168
Drower ES The Mandaens of Iraq and Iran: Their Cults, Customs, Magic, Legends and Folklore 1962 Leiden: Brill 348
Ovid Melville AD Metamorphosis 1986 8 Oxford: Oxford University Press 738 878 3491406
Porteous A Forest Folklore, Mythology and Romance 1928 London: George Allen and Unwin Ltd 63
Frazer JG The Golden Bough: The Roots of Religion and Folklore 1981 New York: Avenel Books 61, 62
Hasluck FW Hasluck MM Christianity and Islam under the Sultans 1973 New York: Octagon Books 80, 166, 172, 175, 176 178, 198, 206, 262
Ben-Zvi S Abu Rabia A Kressel GM The Charm of Graves: Mourning Rituals and Tomb Worshiping among the Negev Bedouin 1998 Tel Aviv: Ministry of Defense Press (Hebrew) 89, 90
Westermarck E Ritual and Belief in Morocco 1968 University Press Books, New Hyde Park, New York 78
Hornblower GD A Sacred Grove in Egypt Man 1930 30 17 19
Legey F The Folklore of Morocco 1935 London: George Allen and Unwin Ltd 80
Frazer JG Magic and Religion 1944 London: Watts and Co 18
Dafni A Why are rags tied to the sacred trees of the Holy Land? Economic Botany 2002 56 315 327
Rodwell JM Koran 1937 London: J.M. Dent 70 Trans
'Aref S Shiller E Graves of Muslim Saints in Galilee and Northern Israel Religion, Rituals and Graves of Muslim Saints in the Land of Israel Ariel 1996 117–118 Jerusalem (Hebrew) 99 106
Fergusson J Tree and Serpent Worship 1971 Delhi, Varanasi: Indological Book House 21
Walhouse MJ Rag-bushes and kindred observances Journal of the Anthropological Institute of Great Britain and Ireland 1880 9 97 106
Bechmann R Tree and Man: The Forest in the Middle Ages 1990 New York: Paragon House 279
Ilan Z Tombs of the Righteous in the Land of Israel 1997 Jerusalem: Canna (Hebrew) 99
Texier CFM Voyage de L'asie Mineure 1825 II 198 Paris: Firmin Didot Freres (cited by Hasluck 1973:198)
De Cleene M Lejeun MC Compendium of Symbolic and Ritual Plants in Europe 2003 Ghent: Man and Culture
Blackman AM Some Egyptian and Nubian notes Man 1910 10 25 29 56–57
Quantz JD Dendro-psychosis American Journal of Psychology 1989 9 447 506
Yassin A Akgul OGH Ugur MS Solak MH The Monument Trees of Manisa (Turkish) 1989 128
Thankappan-Nair P Gupta SN Tree symbols worship among the Nairs of Kerala Tree Worship in India A New Survey of a Pattern of Folk-Religion 1965 Calcutta: Indian Publications 97, 93 103
Majupuria TC Johshi DP Religious and Useful Plants of Nepal and India 1997 India: Gupta Lashkar 183
Gupta SM Plant Myth and Traditions in India 2001 Calcutta: Munshiram Manoharlal xxi
Leland CG Gypsy Sorcery and Fortune Telling 1891 London: T. Fisher Unwin 147
Simmons FJ Plants of Life, Plants of Death 1988 Madison: University of Wisconsin Press 41 100, 293
Canaan T Mohammedan Saints and Sanctuaries in Palestine Journal of the Palestine Oriental Society 1927 7 1 331
Fabri F Palestine Pilgrims Text Society 1887 London: A.M.S Press 9 10, 493
Eliade M Patterns in Comparative Religion 1958 London: Sheed and Ward 266 268
Frese PR Gray SJM Eliade M Trees The Encyclopaedia of Religion 1995 15 Macmillan Library Reference USA. New York: Simon, Schuster and Macmillan 26 33
Robertson-Smith W Lectures on the Religion of the Semites 1889 Edinburgh: Adam and Charles 185
Nöldeke T Geschichte der Preser und Araber zur Zeit der Sasaniden- Aus der Arabischen Chronik des Tabari, Übersetz ausfurlichen Ergänzungen Erläuterungen und Ergänzungen 1973 Leiden: Brill 181 (First published 1879)
Grimm J Stalybrass JS Teutonic Mythology 1966 II New York: Dover Publications, Inc 651 (Original German edition)
Munro-Chadwick H The Oak and the Thunder God Journal of the Anthropological Institute of Great Britain and Ireland 1900 30 22 44
|
16270941
|
PMC1277088
|
CC BY
|
2021-01-04 16:40:12
|
no
|
J Ethnobiol Ethnomed. 2005 Sep 28; 1:8
|
utf-8
|
J Ethnobiol Ethnomed
| 2,005 |
10.1186/1746-4269-1-8
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7020ehp0112-00121715345368ToxicogenomicsArticlesGene Interaction Network Suggests Dioxin Induces a Significant Linkage between Aryl Hydrocarbon Receptor and Retinoic Acid Receptor Beta Toyoshiba Hiroyoshi Yamanaka Takeharu Sone Hideko Parham Frederick M. Walker Nigel J. Martinez Jeanelle Portier Christopher J. Laboratory of Computational Biology and Risk Analysis, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USAAddress correspondence to C.J. Portier, National Institute of Environmental Health Sciences, PO Box 12233, Research Triangle Park, NC 27709. Telephone: (919) 541-3802. Fax: (919) 541-3647. E-mail:
[email protected] authors declare they have no competing financial interests.
8 2004 23 6 2004 112 12 1217 1224 10 2 2004 23 6 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Gene expression arrays (gene chips) have enabled researchers to roughly quantify the level of mRNA expression for a large number of genes in a single sample. Several methods have been developed for the analysis of gene array data including clustering, outlier detection, and correlation studies. Most of these analyses are aimed at a qualitative identification of what is different between two samples and/or the relationship between two genes. We propose a quantitative, statistically sound methodology for the analysis of gene regulatory networks using gene expression data sets. The method is based on Bayesian networks for direct quantification of gene expression networks. Using the gene expression changes in HPL1A lung airway epithelial cells after exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin at levels of 0.1, 1.0, and 10.0 nM for 24 hr, a gene expression network was hypothesized and analyzed. The method clearly demonstrates support for the assumed network and the hypothesis linking the usual dioxin expression changes to the retinoic acid receptor system. Simulation studies demonstrated the method works well, even for small samples.
Bayesian networksdioxingene regulatory networksMarkov chain Monte Carloretinoic acid receptorrisk assessmentsystems biologytoxicogenomics
==== Body
Gene expression arrays (gene chips) have enabled researchers to simultaneously monitor the approximate level of mRNA expression for a large number of genes. These mRNA expression levels are one component of the machinery that controls the function and survival of cells; the other components constitute the other major biochemical constituents of a cell such as the actual DNA sequence, protein levels, and cellular substructures. Signal transduction pathways have long been used to describe the sequence of biochemical events that control cellular function and generally include all aspects of the biochemistry of a cell. In the absence of full proteomic data (both primary proteins and modified proteins), it is valuable to understand the quantitative relationship between genes, which we will refer to as gene expression networks. The rates derived from the quantification of gene expression networks provide crude estimates for the overall rates linking genes through complicated signaling pathways. In addition, hypothesized linkages between genes will aid in focusing research efforts in other areas such as proteomics, metabolomics, and toxicologic assays.
We recently used toxicogenomic analysis to examine the response of human peripheral lung epithelial cells to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, dioxin) in vitro (Martinez et al. 2002). Exposure to this persistent environmental pollutant has been associated in human populations with increased risk of lung cancer and chronic obstructive pulmonary disease; therefore, understanding its mechanism of action may provide insights into the risk of persistent human exposure not only to TCDD but to other ligands of the aryl hydrocarbon receptor (AhR). In this study we showed a variety of cell-signaling pathways that exhibited a dose-dependent alteration by TCDD. One observation in this study was an alteration in retinoic acid (RA)-responsive genes. Alterations in RA homeostasis have been observed previously in rodents, leading to a retinoid-deficient state. In addition TCDD exposure in rats has been associated with increased incidence of squamous neoplastic and nonneoplastic lesions including squamous cell carcinoma of the lung and hard palate region the oral mucosa (Kociba et al. 1978). Given that alterations in retinoid signaling can affect the differentiation of squamous epithelia, it is possible that the increase in these squamous lesions may be due to a retinoid-deficient state induced by the alteration in retinoid homeostasis.
Identification of the retinoid-responsive genes in the TCDD microarray analyses suggested a functional relationship between AhR activation and retinoid homeostasis and/or signaling in the human lung epithelial cells. Although such relationships can be tested empirically, invariably a large number of functional relationships are possible within a given microarray data set; therefore, priority setting for functional validation studies is often a challenge. In this article we develop a computational approach for evaluating the likelihood that observed changes in gene expression are due to hypothesized functional relationships. We then test the AhR–retinoid interaction using this method.
Several methods have already been proposed for the analysis of gene expression data. The most commonly used methods rely on description of simple fold increases in expression, phylogenetic tree analyses, clustering methods, classification methods, or combinations of these. Methods have also been proposed to develop gene expression networks using dynamical systems defined by ordinary differential equations (Chen et al. 1999), modified linear regression methods (Gardner et al. 2003), Boolean networks (Akutsu et al. 2000) where gene expression data are converted to two states (ON and OFF), discrete networks (Hartemink et al. 2002), and many others. Bayesian networks (Friedman et al. 2000; Pe’er et al. 2001) have been proposed as a means of identifying gene interaction networks (Imoto et al. 2002; Tamayo et al. 1999) and for predicting protein–protein interactions using a combination of different types of genomic data (Jansen 2003). Many of the available methods are discussed in a recent review article (Lockhart and Winzeler 2000). Few methods exist that combine careful statistical estimation and hypothesis testing with quantitative gene interaction models to provide a systems biology–based approach for the analysis of microarray data.
In this article, a Bayesian network approach (Friedman et al. 2000; Imoto et al. 2002) previously suggested is modified to provide direct quantification of gene expression networks using microarray data for a known network. This analytical approach provides a model that can be used for mechanism-based mathematical models and for formal analyses of biological hypotheses.
Materials and Methods
Definition of Gene Expression Network
The basic concept for Bayesian networks in the analysis of gene expression data has been described previously (Friedman et al. 2000; Imoto et al. 2002; Tamada et al. 2003). A gene expression network consists of a collection of P genes, denoted by X1, X2,…XP, linked by weighting functions, wi(θi) ( i = 1,2,…P ), where the subscript i denotes that this weighting function pertains to the control of gene Xi by all genes linked to it and θi denotes the vector of parameters defining the functional relationship. In cases where the relationship between individual genes is monotonic (i.e., Xi either stimulates or inhibits Xj but cannot have mixed effect), such a network can be easily represented graphically as in Figure 1. Figure 1 is a simple gene expression network consisting of four genes (squares) and four weighting functions (circles), with lines linking the genes and the weighting functions. Two kinds of lines appear in the model. A line with a bar implies inhibition (e.g., gene X3 inhibits gene X4 in Figure 1), and a line with no bar implies stimulation (e.g., gene X1 stimulates gene X4). No line between genes implies these genes have no direct relationship to each other (X2 and X4 are not directly linked). The weighting function combining the effects of genes X1 and X3 on gene X4 is denoted by w4(θ4) in Figure 1.
The vector W(θ) = [w1(θ1) w2(θ2)… wP(θP)] fully characterizes the functional relationships between genes in a gene expression network and is the target of any estimation effort to identify and quantify a network. The functional form that can be used for any individual wi(θi) is not restricted. One example is the log-linear gene expression network.
Log-Linear Gene Expression Network
One of the simplest types of weighting function used to describe a gene expression network is the log-linear weighting function given by the following form:
where xj is the observed level of expression (or ratios of expression) of gene Xj, βji is the magnitude by which a change in one log unit of gene Xj will affect the level of expression of gene Xi, and Iji is an indicator variable describing the direction of the change denoted by βji, where Iji = 1 for stimulation, Iji = −1 for inhibition, and Iji = 0 for no effect. For simplicity of notation, we define B = [βji]j
= 1,2…p,
i=1,2…p, T = [Iji]j
=1,2…p,
i=1,2…p, and A = [log(x1), log(x2),…log(xp)], where we refer to T as the transition matrix and B as the parameter matrix. It is then possible to rewrite Equation 1 in its matrix form given by
where θ = [β11 β12…βpp], and the dot represents element-by-element multiplication of B and T. In the example given by Figure 1, the matrices B and T are 4 × 4 matrices and have only 4 nonzero elements each [(1,3), (1,4), (2,3), and (3,4)], so the vector of parameters is θ = [β13 β14 β23 β34]. Tamada et al. (2003) used a non-parametric B-spline for wi (θi). Such a method could be used in this context as well, where the breakpoints in the splines are at individual doses or times used for an experimental design.
The transition matrix provides the qualitative structure of the gene expression network, and the parameter matrix quantifies the strength of the relationship between the genes. In the following we use Np(θ) to represent a general gene expression network with P genes and NTP(θ) to specifically represent a log-linear gene expression network with P genes.
Bayesian Network Estimation Procedure
Like any other biological measurement, it can be presumed that two observations taken from seemingly identical examples may differ because of uncontrolled variables or simple random fluctuation; this difference is traditionally defined as random variation about the mean behavior in a model. With random variation, x = [x1, x2…xp] is an observation from a random matrix X = [X1, X2…XP]. The simplest method by which random variation can be included in a gene interaction network is to assume that Xi is conditional on knowledge of the other X’s and θ follows a prescribed probability density function. Define Xi
= [X1, X2,…Xi−1, Xi+1,…XP] and define fi (Xi|X̄i, θ) to be the conditional density of Xi. If a gene has a regulatory effect on gene Xi, that gene is referred to as a “parent of gene Xi”; in other words, it belongs to the set referred to by Pa(Xi). Hence, for example, in the model depicted by Figure 1, Pa(X3) = [X1, X2]. This notation has been used in other cases and in the context of this modeling, the distribution could then be written as fi (Xi |Pa(Xi), θ). A greater level of statistical complexity is possible by also presuming that the parameters have probability density functions; hi(θi) is referred to as the prior distribution of θi. This formulation places the network defined by NP(θ) and the data into the context of classical Bayesian networks (Jensen 1996).
Suppose that we have m sets of microarray data [x1j, x2j,…xPj]j
= 1,2,…m from gene expression network NP(θ), where individual arrays are independent random samples from the joint density function for the genes. The joint density function for the parameters given the gene expression data, denoted g (θ|
X ), is referred to as the posterior distribution and can be estimated using the Markov chain Monte Carlo (MCMC) method (Hastings 1970). In the examples given in this article, the Metropolis algorithm (Andrec and Prestegard 1998) is used to sample from the MCMC to generate samples from the joint density.
Specific Cases Used in This Analysis
In all analyses that follow, the gene expression network is presumed to be a log-linear network defined by NTP(θ) in Equations 1 and 2. It is assumed that data arise from microarrays using a relative comparison between two samples (no change results in a value of 1, increased expression > 1, decreased expression < 1), and the distributions for the log of the individual relative gene expression levels conditional on knowledge of T, θ and the other X’s, fi (X̄i|Xi, θ), are assumed to be normal, with mean defined as the exponent of e in Equation 1 and with standard deviation (SD)σ. All parameters in θ = (B,S), where S = [σ1, σ2,… σP] are assumed to have prior distributions (normal for the elements of B and uniform for the elements of S).
Assume that the structure of T (transition matrix) is known without error. In this situation, the qualitative relationship between genes in the gene expression network is known. Taking Figure 1 as an example, expectation of each log (Xi) (i = 1,2,3) becomes E [log(X1)|T,B,— X̄1] = 0, E[log(X2)|T,B,
X̄2] = 0, E[log(X3)|T,B,
X̄3] = β13 log(X1) + β23 log(X2), and E[log(X4)|T,B,
X̄4] = β14 log(X1) − β34 log(X3). The ultimate goal of defining a Bayesian network is to derive the posterior distribution for the parameters of interest. To derive the posterior, we must first calculate the conditional likelihood of the data, denoted LN[X |NTP(θ)]. The likelihood is the product of the individual conditional densities and is written
In the MCMC analysis, we must assume a mean and variance for the prior normal distributions for the β’s and bounds on the prior uniform distributions for the σ’s. Several options were chosen for the prior means of the β’s and an uninformative SD (10) was chosen for the prior variance. To develop bounds on the prior uniform distributions for the σ’s, SDs were calculated for each gene across replicates, and the maximum SD observed was multiplied by 2 to set the upper bound, with 0 set as the lower bound. Given these priors and the data, MCMC iterations for each data set analyzed are run until the estimates for the posterior distributions for the β’s and the σ’s are stabilized.
Other distributions and methods could be used to define the priors and generate the posterior distributions for the likelihood and the parameters in the model. In considering a more complicated functional relationship between genes, Michaelis-Menten–type equations could be used to develop networks with restricted maximum and minimum linkages. Such networks would require substantially more data.
A user-friendly software package for these analyses is available from the corresponding author.
Application to Microarray Gene Expression Data
Martinez et al. (2002) evaluated the change in expression of 2,091 genes in triplicate samples of HPL1A and A549 cells exposed to differing levels (0, 0.1, 1.0, and 10 nM) TCDD for 24 hr. Total RNA was extracted and, using methods described by Martinez et al., hybridized to NIEHS Human ToxChip, version 1.0 (http://dir.niehs.nih.gov/microarray/chips.htm) to obtain changes in gene expression in dioxin-treated cells (one channel) relative to the controls (second channel). They identified 68 genes that were altered in at least one cell line and 15 genes that were altered in both cell lines. Of these, they identified 11 genes that appear to be involved in the effects of TCDD on the retinoid-signaling pathway. In this article we hypothesize a gene interaction network defining the quantitative role of TCDD in altering retinoid signaling based on the current available literature. The data for these 11 genes from the HPL1A cells and the hypothesized network are analyzed using the methods described above.
Results
Dioxin Analysis
2,3,7,8-Tetrachlorodibenzo-p-dioxin is a known human carcinogen, a suspected teratogen, and highly toxic in most mammalian species. There has been considerable speculation that TCDD alters the retinoic acid receptor (RAR)–dependent signaling pathway via alteration of the synthesis and metabolism of RA. Microarray data (Martinez et al. 2002) on changes in gene expression in HPL1A lung airway epithelial cells after exposure to TCDD at levels of 0.1, 1.0, and 10.0 nM for 24 hr identified 11 genes with significant changes at the 99% confidence level. The gene identifiers and data are given in Tables 1 and 2. Figure 2 hypothesizes a gene interaction network linking the traditional TCDD-induced genes and genes in the RAR-dependent signaling pathway.
Vitamin A (retinol) is taken up from blood and binds to the CRBP in the cytoplasm. Retinol and alcohol dehydrogenases convert the sequestered retinol to retinal, which is then converted to RA by retinal dehydrogenases such as ALDH6 (Rexer et al. 2001). It is also possible that cytochrome P450s such as CYP1A1 may also convert retinal to RA (Zhang et al. 2000). Once RA is synthesized, it binds to cytosolic RA binding proteins (such as CRABP). RA enters the nucleus, where it binds to two types of ligand-activated nuclear transcription factors, the RA receptors (e.g., RARB) and the retinoid X receptors. Several groups have hypothesized that changes observed in RA levels from dioxin exposures are mediated through increased metabolism of retinal to RA through retinal dehydrogenases or cytochrome P450s or both (Schmidt et al. 2003). Using these data together with the known AhR gene battery, we developed a hypothetical gene interaction network (Figure 2).
The predominant linkage to RAR is through upregulation of ALDH6 and CYP1A1, which synthesize RA. TCDD alters the metabolism of all-trans-RA (Schmidt et al. 2003), suggesting the linkage between ALDH6 and RARB in Figure 2. RARB has been shown to play a role in the inhibition of cellular replication (Sun et al. 2000). RARB is assumed to modify the expression levels of four genes: ELF3, NCOA2, ZNF42, and CDKN1A. These genes have been shown to be parts of the differentiation pathways of various cell types and are hypothesized to be modified by changes in the RA-signaling pathway. ELF3 is an epithelial-specific transcriptional regulator that may play a role in lung carcinogenesis (Tymms et al. 1997). NCOA2, also known as GRIP1, interacts with the five steroid hormone receptor types (Hong et al. 1997; Schmidt et al. 1998). ZNF42, also known as MZF-1, is a putative transcriptional regulator induced by RA in human myeloid cells (Hromas et al. 1991). CDKN1A is induced by RA through RARB in human neuroblastoma tumors (Cheung et al. 1998; Liu et al. 1996). Both ALDH6 and RARB affect the regulation of NCOA2, which in turn alters the regulation of ZNF42 and CDKN1A. ACOX1, the human peroxisomal acylcoenzyme A oxidase, is hypothesized in the model to upregulate both RARB and NCOA2. ACOX1 is the first enzyme of the fatty acid beta-oxidation pathway (Varanasi et al. 1994), and changes in this gene are likely to affect endogenous levels of fatty acids known to activate the retinoic X receptor, thereby modulating gene expression (Issemann et al. 1993). The second major linkage occurs between cytochrome P4501A1, CYP1A1, and RARB. An inducer of CYP1A1 (β-naphthoflavone) induced the metabolism of all-trans-RA in human intestinal epithelial cells (Lampen et al. 2000). CYP1A1 is upregulated by TCDD (Portier et al. 1993), suggesting the linkage between TCDD and genes in the RAR-signaling pathway such as RARB, NCOA2, and CRABP, a specific carrier protein for vitamin A that influences metabolism of RA and increases the sensitivity of a cell to vitamin A signaling (Boylan and Gudas 1992; Ong 1987). The CRABP promoter contains an enhancer region through which RA inhibits CRABP transcription (Means et al. 2000).
The slope parameters for all the linkages between genes (βij in Equation 1) in Figure 2 were estimated using the Bayesian gene interaction network approach as described above. Prior probability distributions for the log gene expression values were assumed to be normally distributed with a mean of zero and a variance of 1. The SDs (σ1, σ2, …σP) were assumed to have uniform priors ranging from zero to two times the largest SD observed for any one gene (an uninformative prior). A total of 100,000 MCMC samples were obtained, and the last 80% (80,000) were used to estimate the posterior density of the parameters. Although no formal MCMC stopping rule was used, analyses of the last 80,000 MCMC samples clearly supported convergence.
Figure 2 also illustrates the type of results routinely obtained in Bayesian analyses. The four histograms shown in Figure 2 are the empirical posterior densities for 4 of the 19 linkages. The linkage between TCDD and CYP1A1 (A) has a distribution for which there are virtually no values below zero, indicating a strong statistical relationship in these data. The mean value, 0.269, indicates the degree of change in CYP1A1 expression as a function of the change in TCDD concentration. The means, SDs, and percentages of values below zero are summarized for all 19 linkages in Table 3. The distributions for the variances are presented in Table 4. Other uninformative priors were tried with no significant alteration in the results presented in Table 3. In our Bayesian analysis we estimate the posterior distributions for each parameter given the data. If a distribution for a given parameter has a small probability of being < 0 (such as ≤ 0.1), that parameter supports a linkage between genes.
The network depicted in Figure 2 was developed to test the hypothesis of a linkage between dioxin-responsive genes CYP1A1 and ALDH6, and the RAR-signaling gene RARB. The distribution for βCYP1A1
→
RARB had a substantial mass less than zero (26% < 0, Table 3), suggesting a lack of support for the linkage between changes in message for these two genes. Similar results were seen for βALDH6
→
RARB (20% < 0, Table 3). Examination of the joint density for βCYP1A1
→
RARB and βALDH6
→
RARB suggested a negative correlation, indicating that the data may not support both linkages simultaneously. This is not surprising, as they are both acting upon the same component of RA synthesis. By forcing βCYP1A1
→
RARB = 0 and again estimating the remaining parameters, we can examine the distribution of βALDH6
→
RARB under the condition that the other linkage is not present; in this case, βALDH6
→
RARB had no estimates less than zero (0%) and there was no change in the posterior distribution for the log-likelihood, suggesting almost no change in the fit of the network to the data, even though we dropped the linkage between CYP1A1 and RARB. Conversely, we can set βALDH6
→
RARB = 0 and examine the distribution of βCYP1A1
→
RARB; here also we see 0% < 0 and no change in the log-likelihood. These two analyses support the hypothesized linkage between TCDD-responsive genes and RARB-responsive genes, but only through either ALDH6 or CYP1A1, not both. Finally, setting both βCYP1A1
→
RARB = 0 and βALDH6
→
RARB = 0 significantly shifts the distribution of the posterior log-likelihood to smaller values (10% reduction overall), suggesting that at least one of these linkages is needed to explain these data.
The only other linkage that did not appear to be supported by these data was the hypothesized linkage between NCOA2 and ZNF42. The distribution for βNCOA2
→
ZNF42 had a mean estimate of zero, with 48.5% of the estimates less than zero. Assuming βNCOA2
→
ZNF42 = 0 had no impact on the log-likelihood, suggesting this linkage was not needed in the model and that there was no correlation offset with other parameters. Given the sample size and the number of genes in the network, it is surprising that all other linkages appeared to be supported by these data, with the percentage of β values less than zero ranging from 0% for several pairs (βTCDD →
CYP1A1, βRARB
→
CDKN1A, βRARB
→
ELF3, βRARB-
→
NCOA2, βRARB
→
ZNF42, and βACOX1
→
NCOA2) to 9.4% (βACOX1
→
RARB).
Simulation Studies
Although the TCDD example is illustrative of the method, it does not address how well this method works under diverse conditions; this is best addressed by Monte Carlo simulations. One thousand (1,000) simulated experiments from the simple four-gene network in Figure 1 were generated by the computer using sample sizes of 50, 25, and 10 gene chips in each experiment. Twenty-two combinations of the model parameters (θ = [β13, β14, β23, β34, σ1σ2, σ3, σ4]) were considered. For each simulation, posterior distributions were calculated and summarized by their means, medians, and SDs. The MCMC process used was identical to that used for the dioxin example, with the exception that only 8,000 iterations of the Metropolis algorithm were performed, and the last 20% (1,600) values were used to calculate the summary statistics. Multiple runs with different starting points were used, with no difference in the final results (not shown).
Table 5 provides representative results from two of the simulation studies. The results indicate that, when sample sizes are sufficiently large, Bayes estimates of the model parameters appear to be close to the assumed value. When sample size is reduced, SDs of the β’s become larger, going from 0.2 to 0.45 as the sample size drops from 50 to 10. However, estimation itself seems to be unbiased, even in the case of only 10 replicates. In the second example in Table 5, one parameter was set to zero, providing a case where there is no linkage from gene 1 to gene 3. In this case, we see that the estimation for a nonexistent link is approximately zero, and one could discard this link. Similar results were seen for all the cases studied.
To further challenge the estimation procedure, an eight-gene network was simulated and estimated. This network had 18 parameters and was a greater challenge to the Bayesian method. Because of the increase in the number of parameters, SDs were substantially larger than in the four-gene model, but the estimation was still effectively unbiased.
Discussion
Many methods have been developed for the analysis of gene expression microarray data, but few methods exist for using these data to quantify the interrelated behavior of genes within gene interaction networks. Most network-based methods are focused on network identification, not quantification. Given a hypothesized gene interaction network, this article develops and demonstrates the use of Bayesian network models as a tool for the analysis of a network using microarray data. The method allows for evaluating the strength of relationships within a hypothesized network and could also be used to test for additional linkages within the network.
There were two key points raised by these analyses. First, the application of this quantitative approach to the experimental data on TCDD effects in human lung epithelial cells clearly identified two subnetworks as significantly related to the AhR battery and the retinoid signaling. This indicates that the observed gene expression changes are consistent with the underlying hypothesized mechanism of action. In one sense this represents an alternate validation step in a tiered approach to evaluating microarray analyses. For example, ZNF42, although annotated as a retinoid-responsive gene, has not been previously validated as a retinoid-responsive gene in this cell system. The quantitative modeling suggests a highly significant relationship between ZNF42 expression and other genes in the retinoid-signaling subnetwork, which provides confidence that its alteration was indeed due to activation of the retinoid-signaling pathway. The testing of other subnetworks within a given data set can further serve to increase confidence that inferences on relationships between genes obtained from other types of analyses (evaluation of gene annotation, clustering, pathway analysis, informatic-based network mapping, literature searches) are real.
The second point from these analyses is that we were able to test the interaction between the two subnetworks (AhR and retinoid) and illustrated that a functional relationship was likely real. Such an analysis is useful in that it supports further testing of this mechanism experimentally. It may be that some fraction of the toxicity associated with chronic exposure to TCDD could be the direct result of TCDD-induced increases in RA in the cells. The hypothesized network clearly supports a significant change in gene expression associated with signaling through the RARB pathway. The quantitative linkages observed in this experiment are unlikely to hold for an in vivo system but suggest that an experiment exposing laboratory animals to TCDD, which includes both TCDD and RA measurements with gene expression measurements, would be useful. Two recent experiments address these issues to a limited extent. Schmidt et al. (2003) examined RA levels and changes in expression of CRBP1 in male Sprague-Dawley rats and saw significant changes in RA levels in kidney, liver, and serum, and a marginal change in liver CRBP1 after 28 days. They did not examine any of the genes in the network shown in Figure 2, so it is difficult to compare directly with our results. Johnson et al. (2004) used in vitro data from three experiments with AhR ligands activating genes in the heart, kidney, and thoracic aorta of mouse embryos. They used an exhaustive search of three linkages for each gene to identify the most likely gene–gene interactions. They also identified linkages to genes in the RA-signaling pathway (IGFBP-3 and IGFBP-6), but again, not the specific genes used in Figure 2.
The simulation experiments were different from the analysis of the TCDD study. In the TCDD study, the network linkages were perturbed to cause significant quantitative changes in expression, which then could be used to quantify the linkages between genes. In contrast, the simulation study used only the random variation in expression levels to quantify the network. The simulation studies indicate that the proposed method appears to be unbiased and, on average, produces the correct results. However, sample size could be a problem for small experiments with minor changes in gene expression. When the sample size is only 10 microarrays, the SD can be large relative to the expected value of the linkage between two genes, suggesting one might misinterpret a linkage as having little statistical support. This problem gets worse as the number of genes in the network increases. In contrast, large sample sizes of 50 microarrays are unlikely to have this problem.
Directed changes in the network, as in the dioxin experiment, can help overcome this problem and allow the quantification of significant linkages by as few as nine microarrays. To address this question, two additional simulations were conducted. Using the network shown in Figure 2 and the parameters estimated for the TCDD network shown in Table 3, we simulated 500 data sets consisting of nine microarrays—three for each dioxin dose; that is, we replicated the experiment 500 times using the predicted model. On average the resulting parameter estimates were identical to those observed from fitting the original data but appeared to have a slightly smaller SD than that estimated in the model. This decrease in SD could indicate a degree of model misspecification, as the simulated data appear to fit better than the observed data. In addition, whereas the observed data showed a nonsignificant linkage between CYP1A1 to RARB, 48% of the simulated data sets found this linkage to be significant. Similarly, 54% found the linkage between ALDH6 and RARB to be significant. In contrast, the simulations found a significant linkage between NCOA2 and ZNF42 in only 6% of the cases (hence the Type I error appears to be good) and between TCDD and CYP1A1 in 100% of the cases (power is high).
In a second simulation, the network shown in Figure 2 was again simulated, this time without TCDD included in the experimental design and using just random variation in the genes to produce the data. Again, the results were unbiased, but the SDs more than doubled. In addition, the probability of observing a significant linkage was reduced by about 20% for most linkages. This illustrates the value of stimulating the system when trying to identify gene interaction networks.
Clearly, this type of modeling approach is limited in terms of interpretation. First, the model cannot be cyclic; hence, increases in CRABP as a function of RARB that might then result in greater binding of RA in the cytosol, reducing RARB expression, could not be included. Given time-course data, it could be possible to explore this linkage using a more complicated modeling form or some other method of analysis such as semicyclic Bayesian networks. Second, the method is dependent on a parametric model, and the choice of this model could impact the overall findings from the analysis. For example, if certain genes reached their maximal expression at lower doses of TCDD, the use of a log-linear model could underestimate low-dose changes while overestimating high-dose changes. This, in turn, could lead one to accept or reject a given model incorrectly. It should be noted that this type of criticism applies to all the other network analysis methods as well. Finally, although not seen in this analysis, it is possible that the resulting distributions for the linkages between the genes could be sensitive to the choice of prior distributions, and one should be careful to evaluate if such an impact might exist with the data.
Although the approach presented here involves only gene expression data, it can easily be expanded to include other data relevant to the linkages between genes and the quantification of signal transduction pathways in cells. Data quantifying protein levels in cells could easily be folded into a general likelihood, linked via a similar model, and analyzed to quantify the entire network. Such an approach leads to rational, mechanism-driven simultaneous analyses of genomics, proteomics, and metabolomics data. In addition, the networks identified through this type of analysis can easily be combined with other mechanism-based mathematical models such as physiologically based pharmaco-kinetic and pharmacodynamic models to present a true, systems-biology approach for the quantification of risks from exposures to xenobiotics like dioxin. This analysis would form one module of an overall model for TCDD toxicity. For example, if microarray data were available in rats exposed to TCDD, existing models like that of Kohn et al. (2001) could easily be linked to the gene interaction network discussed above. These, in turn, could be linked to cancer data using a mechanistic model to test hypotheses regarding cancer incidence and the mechanisms involved, as shown by Brooks et al. (1999).
The method proposed here is not restricted to the log-linear model used in this analysis, nor is it linked to the statistical likelihood chosen for the analysis. Other models such as dynamic models (Chen et al. 1999) and other statistical likelihoods (Wolfinger et al. 2001) could easily be incorporated into the analysis methods.
Bayesian networks have been used in a number of settings to provide insight into the complicated linkage between variables that interact. Quantifying the distributions linking genes into networks and expanding this to include proteins and protein modifications will make it possible to quantify the impact of a given chemical agent on the signal transduction pathways in a cell. Although many different methods could be used for this, Bayesian networks have the advantage of flexibility, which will make it possible to build on existing knowledge while bringing new data into the analysis. For the dioxin study presented here, the limitations of the sample size preclude an overall conclusion concerning the validity of the final model for predictions about the role of dioxin in changes to the RAR-signaling pathway. However, this analysis has strengthened the underlying hypothesis that changes in RAR signaling may play an important role in dioxin-mediated toxicity and suggest a number of experiments that could lead to a better-characterized network; this is left for future work.
In this article we used known scientific inferences and gene annotation to develop the initial tested network. This approach can also be applied to evaluating the likelihood of any hypothesized network developed by other approaches. As such, it can be applied to networks developed using other types of analyses including Bayesian, Boolean, and informatics-based approaches, as well as other known networks in the scientific literature. The ability to test hypotheses in the context of the network and to build modules that can be quantitatively linked to toxicity are first steps in a true systems-biology approach to mechanism-based use of genomics in risk assessment. This analysis is unique in that it directly addresses these uses.
Figure 1 A simple gene expression network consisting of four genes and four nonzero functional relationships.
Figure 2 Hypothesized network describing the linkage between AhR-responsive genes and RARB-responsive genes, where the numbers represent the mean estimate for the linkage (β) between the two genes on any given line, and the four distributions (A–D) are the posterior distributions for the linkages between (A) TCDD and CYP1A1, (B) CYP1A1 and NCOA2, (C) NCOA2 and CDKN1A, and (D) RARB and CDKN1A.
Table 1 Description of genes included in the gene interaction network shown in Figure 2.
Gene symbol (alternate symbols)a Accession no.a Gene namea Biological role
ALDH3 (ALDH3A1) AA069024 Aldehyde dehydrogenase 3 family, memberA1 May play a role in the oxidation of lipid aldehydes, especially those generated by lipid peroxidation (Vasiliou et al. 2000); is induced in rat liver by TCDD (Unkila et al. 1993)
ALDH6 (ALDH1A3) AA054748 Aldehyde dehydrogenase 1 family, member A3 Has ability to synthesize retinoic acid from both retinol and retinal (Rexer et al. 2001)
ALDH10 (ALDH3A2) H63779 Aldehyde dehydrogenase 3 family, member A2 Oxidizes long-chain aliphatic aldehydes to fatty acid
CYP1A1 AA418907 Cytochrome P450, subfamily I, polypeptide 1A Phase I enzyme; its expression is controlled by the AhR. Metabolically activates procarcinogens to genotoxic electrophilic intermediates (Nebert et al. 1996)
CRABP N23941 Cellular retinoic acid binding protein 1 Small intracellular protein that is a carrier for RA (vitamin A)
NCOA2 (SRC-2, TIF2, GRIP1) R77770 Nuclear receptor coactivator 2 Transcription coactivator of retinoid/thyroid receptors; a histone acetyltransferase that plays an important role in lipid metabolism and energy balance (Picard et al. 2002; Xu and Li 2003)
RARB W93713 Retinoic acid receptor, beta Hetero/homodimers associated with oncogenicity (Lin and Evans 2000); overexpression in oral squamous carcinoma cell lines; leads to growth arrest and apoptosis (Hayashi et al. 2001)
CDKN1A (p21, Cip1) N23941 Cyclin-dependent kinase inhibitor 1A Functions as a regulator of cell-cycle progression; overexpression linked to carcinogenesis (Biankin et al. 2001)
ZNF42 (MZF1, MZF-1, MZF1B) R83364 Zinc finger protein 42 Transcription factor that belongs to the Kruppel family of zinc finger proteins; RA-responsive; plays a role in cell proliferation (Hromas et al. 1991)
ELF3 (ESX, ESE1) H27939 E74-like factor 3 (ets domain transcription factor, epithelial-specific) Transcription factor that transactivates genes involved in epithelial differentiation and host defense and mediators of proinflammatory responses (e.g., Socs3, Cebp/delta, Bcl3, and CC/CXC chemokines) (Mysorekar et al. 2002; Yoshida et al. 2000)
ACOX1 (ACOX, PALMCOX) AA040205 Human peroxisomal acyl-CoA oxidase First enzyme of the fatty acid β-oxidation pathway (Varanasi et al. 1994); changes in this gene are likely to affect endogenous levels of fatty acids known to activate the retinoic X receptor, thereby modulating gene expression (Issemann et al. 1993)
a From the NCBI (National Center for Biotechnology Information) Unigene database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene).
Table 2 Relative expression level (to control) of genes in the HPL1A cells exposed in replicate to three different concentrations of TCDD.a
Genes
ALDH10 1.56 1.33 1.24 1.42 1.56 1.40 1.69 1.47 1.25
ALDH3 2.10 2.09 2.34 3.88 2.94 4.09 3.11 3.91 3.76
ALDH6 2.42 2.00 1.77 3.40 4.12 3.37 3.76 4.60 3.66
CRABP 0.63 0.69 0.74 0.51 0.48 0.47 0.29 0.46 0.41
CDKN1A 1.56 1.16 1.49 1.30 1.34 1.58 1.51 1.49 1.63
CYP1A1 3.07 2.63 1.31 14.45 6.85 6.09 15.35 14.91 8.08
ELF3 1.56 1.37 1.18 2.19 1.70 1.91 3.15 2.00 2.02
NCOA2 1.42 1.41 0.82 1.34 1.07 0.92 1.42 1.22 0.82
RARB 1.64 1.42 0.93 1.77 1.56 1.21 1.48 1.63 1.15
ZNF42 1.88 1.47 1.11 1.62 1.43 1.32 1.60 1.45 1.22
ACOX1 1.94 1.50 0.78 1.93 1.03 0.84 10.87 1.22 0.59
TCDDb 0.10 0.10 0.10 1.00 1.00 1.00 10.0 10.0 10.0
a Data from Martinez et al. (2002).
b TCDD dose unit is measured in nanomolars. Actual doses are used for TCDD in the analysis.
Table 3 Type of linkage, mean, SD, and percentage of the posterior distribution below zero for all gene–gene relationships in Figure 2.
From To Type Mean SD % < 0
TCDD ALDH3 A 0.140 0.037 0.03
ALDH6 A 0.150 0.035 0.01
ALDH10 A 0.041 0.013 0.23
CYP1A1 A 0.269 0.056 0.003
ALDH6 CRABP R 0.348 0.152 1.27
NCOA2 R 0.132 0.062 2.10
RARB A 0.149 0.191 19.81
CYP1A1 CRABP R 0.235 0.099 1.03
NCOA2 R 0.046 0.038 11.40
RARB A 0.072 0.120 26.34
NCOA2 CDKN1A R 0.847 0.298 0.50
ZNF42 A 0.000 0.168 48.53
RARB CRABP A 0.418 0.234 3.13
CDKN1A A 1.186 0.199 0.00
ELF3 A 1.423 0.220 0.00
NCOA2 A 0.912 0.085 0.00
ZNF42 A 0.975 0.113 0.00
ACOX1 NCOA2 A 0.125 0.021 0.00
RARB A 0.081 0.065 9.38
Abbreviations: A, activate; R, repress.
Table 4 Estimated mean and median SD (σ) for genes included in the gene interaction network shown in Figure 2.
Posterior distribution for σ
Genes Mean (median) SD
ALDH10 0.22 (0.22) 0.04
ALDH3 0.63 (0.61) 0.12
ALDH6 0.62 (0.61) 0.12
CRABP 0.11 (0.11) 0.02
CDKN1A 0.15 (0.14) 0.03
CYP1A1 0.94 (0.92) 0.17
ELF3 0.25 (0.24) 0.05
NCOA2 0.04 (0.04) 0.01
RARB 0.13 (0.12) 0.03
ZNF42 0.08 (0.08) 0.02
ACOX1 0.86 (0.82) 0.20
Table 5 Mean, median, and SD from two simulation studies of the simple four-gene model (Figure 1).
Model Sample size Estimation β14 β13 β23 β34 σ1 σ2 σ3 σ4
β14 = −2 50 Mean (SD) −1.98 (0.22) 0.81 (0.21) 0.83 (0.19) −1.32 (0.13) 1.01 (0.12) 1.00 (0.12) 1.03 (0.12) 1.03 (0.13)
β13 = 0.8 Median (SD) −1.97 (0.22) 0.81 (0.21) 0.82 (0.19) −1.33 (0.13) 1.01 (0.13) 1.00 (0.12) 1.02 (0.13) 1.02 (0.14)
β23 = 0.8 25 Mean (SD) −2.00 (0.29) 0.80 (0.25) 0.81 (0.23) −1.29 (0.19) 1.05 (0.15) 1.03 (0.15) 1.05 (0.16) 1.06 (0.16)
β34 = −1.3 Median (SD) −1.98 (0.29) 0.80 (0.26) 0.78 (0.24) −1.30 (0.19) 1.02 (0.15) 1.01 (0.15) 1.02 (0.16) 1.03 (0.17)
σi = 1, i = 1,2,3,4 10 Mean (SD) −1.97 (0.45) 0.79 (0.40) 0.80 (0.38) −1.29 (0.29) 1.13 (0.26) 1.10 (0.27) 1.15 (0.32) 1.19 (0.29)
Median (SD) −1.95 (0.45) 0.79 (0.40) 0.81 (0.37) −1.31 (0.29) 1.08 (0.24) 1.04 (0.26) 1.04 (0.31) 1.11 (0.28)
β14 = −2 50 Mean (SD) 2.00 (0.17) 0.01 (0.17) 0.80 (0.17) −1.30 (0.12) 1.02 (0.11) 1.03 (0.12) 1.04 (0.12) 1.01 (0.12)
β13 = 0 Median (SD) 2.0 (0.18) 0.01 (0.18) 0.81 (0.18) −1.31 (0.13) 1.01 (0.12) 1.03 (0.12) 1.04 (0.14) 1.00 (0.13)
β23 = 0.8 25 Mean (SD) 2.0 (0.22) 0.01 (0.22) 0.79 (0.21) −1.31 (0.16) 1.04 (0.15) 1.05 (0.15) 1.06 (0.16) 1.04 (0.17)
β34 = −1.3 Median (SD) 2.01 (0.23) 0.00 (0.23) 0.77 (0.21) −1.30 (0.16) 1.02 (0.15) 1.03 (0.15) 1.03 (0.16) 1.03 (0.17)
σi = 1, i = 1,2,3,4 10 Mean (SD) 2.02 (0.40) −0.02 (0.40) 0.83 (0.40) −1.30 (0.32) 1.14 (0.25) 1.13 (0.27) 1.16 (0.30) 1.18 (0.26)
Median (SD) 1.99 (0.40) −0.02 (0.39) 0.85 (0.39) −1.29 (0.31) 1.08 (0.24) 1.06 (0.27) 1.10 (0.30) 1.10 (0.25)
==== Refs
References
Akutsu T Miyano S Kuhara S 2000 Algorithms for identifying Boolean networks and related biological networks based on matrix multiplication and fingerprint function J Comput Biol 7 3-4 331 343 11108466
Andrec M Prestegard JH 1998 A Metropolis Monte Carlo implementation of Bayesian time-domain parameter estimation: application to coupling constant estimation from antiphase multiplets J Magn Reson 130 2 217 232 9500892
Biankin AV Kench JG Morey AL Lee CS Biankin SA Head DR 2001 Overexpression of p21(WAF1/CIP1) is an early event in the development of pancreatic intraepithelial neoplasia Cancer Res 61 24 8830 8837 11751405
Boylan JF Gudas LJ 1992 The level of CRABP-I expression influences the amounts and types of all-trans -retinoic acid metabolites in F9 terato-carcinoma stem cells J Biol Chem 267 30 21486 21491 1328234
Brooks EA Kohn M van Birgelen A Lucier G Portier C 1999 Stochastic models for papilloma formation following exposure to TCDD Organohalogen Compd 41 521 524
Chen T He H Church G 1999 Modeling gene expression with differential equations Pac Symp Biocomput 4 29 40 10380183
Cheung B Hocker JE Smith SA Norris MD Haber M Marshall GM 1998 Favorable prognostic significance of high-level retinoic acid receptor beta expression in neuroblastoma mediated by effects on cell cycle regulation Oncogene 17 6 751 759 9715277
Friedman N Linial M Nachman I Pe’er D 2000 Using Bayesian networks to analyze expression data J Comput Biol 7 3-4 601 620 11108481
Gardner TS di Bernardo D Lorenz D Collins JJ 2003 Inferring genetic networks and identifying compound mode of action via expression profiling Science 301 5629 102 105 12843395
Hartemink AJ Gifford DK Jaakkola TS Young RA 2002 Bayesian methods for elucidating genetic regulatory networks IEEE Intell Syst 17 2 37 43
Hastings WK 1970 Monte Carlo sampling methods using Markov chains and their applications Biometrika 57 1 97 109
Hayashi K Yokozaki H Naka K Yasui W Lotan R Tahara E 2001 Overexpression of retinoic acid receptor beta induces growth arrest and apoptosis in oral cancer cell lines Jpn J Cancer Res 92 1 42 50 11173543
Hong H Kohli K Garabedian MJ Stallcup MR 1997 GRIP1, a transcriptional coactivator for the AF-2 transactivation domain of steroid, thyroid, retinoid, and vitamin D receptors Mol Cell Biol 17 5 2735 2744 9111344
Hromas R Collins SJ Hickstein D Raskind W Deaven LL O’Hara P 1991 A retinoic acid-responsive human zinc finger gene, MZF-1 , preferentially expressed in myeloid cells J Biol Chem 266 22 14183 14187 1860835
Imoto S Goto T Miyano S 2002 Estimation of genetic networks and functional structures between genes by using Bayesian networks and nonparametric regression Pac Symp Biocomput 175 186 11928473
Issemann I Prince RA Tugwood JD Green S 1993 The peroxisome proliferator-activated receptor: retinoid X receptor heterodimer is activated by fatty acids and fibrate hypolipidaemic drugs J Mol Endocrinol 11 1 37 47 8240670
Jansen R Yu H Greenbaum D Kluger Y Krogan N Chung S 2003 A Bayesian networks approach for predicting protein-protein interactions from genomic data Science 302 449 453 14564010
Jensen FV 1996. An Introduction to Bayesian Networks. London:UCL Press.
Johnson CD Balagurunathan Y Tadesse MG Falahatpisheh MH Brun M Walker MK 2004 Unraveling gene-gene interactions regulated by ligands of the aryl hydrocarbon receptor Environ Health Perspect 112 4 403 412 15033587
Kociba RJ Keyes DG Beyer JE Carreon RM Wade CE Dittenber DA 1978 Results of a two-year chronic toxicity and oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p -dioxin in rats Toxicol Appl Pharmacol 46 2 279 303 734660
Kohn MC Walker NJ Kim AH Portier CJ 2001 Physiological modeling of a proposed mechanism of enzyme induction by TCDD Toxicology 162 3 193 208 11369115
Lampen A Meyer S Arnhold T Nau H 2000 Metabolism of vitamin A and its active metabolite all-trans -retinoic acid in small intestinal enterocytes J Pharmacol Exp Ther 295 3 979 985 11082432
Lin RJ Evans RM 2000 Acquisition of oncogenic potential by RAR chimeras in acute promyelocytic leukemia through formation of homodimers Mol Cell 5 5 821 830 10882118
Liu M Lee MH Cohen M Bommakanti M Freedman LP 1996 Transcriptional activation of the Cdk inhibitor p21 by vitamin D3 leads to the induced differentiation of the myelomonocytic cell line U937 Genes Dev 10 2 142 153 8566748
Lockhart DJ Winzeler EA 2000 Genomics, gene expression and DNA arrays Nature 405 6788 827 836 10866209
Martinez JM Afshari CA Bushel PR Masuda A Takahashi T Walker NJ 2002 Differential toxicogenomic responses to 2,3,7,8-tetrachlorodibenzo-p -dioxin in malignant and nonmalignant human airway epithelial cells Toxicol Sci 69 2 409 423 12377990
Means AL Thompson JR Gudas LJ 2000 Transcriptional regulation of the cellular retinoic acid binding protein I gene in F9 teratocarcinoma cells Cell Growth Differ 11 2 71 82 10714763
Mysorekar IU Mulvey MA Hultgren SJ Gordon JI 2002 Molecular regulation of urothelial renewal and host defenses during infection with uropathogenic Escherichia coli J Biol Chem 277 9 7412 7419 11744708
Nebert DW McKinnon RA Puga A 1996 Human drug-metabolizing enzyme polymorphisms: effects on risk of toxicity and cancer DNA Cell Biol 15 4 273 280 8639263
Ong DE 1987 Cellular retinoid-binding proteins Arch Dermatol 123 12 1693 1695a 2825608
Pe’er D Regev A Elidan G Friedman N 2001 Inferring subnetworks from perturbed expression profiles Bioinformatics 17 suppl 1 S215 S224 11473012
Picard F Gehin M Annicotte J Rocchi S Champy MF O’Malley BW 2002 SRC-1 and TIF2 control energy balance between white and brown adipose tissues Cell 111 7 931 941 12507421
Portier C Tritscher A Kohn M Sewall C Clark G Edler L 1993 Ligand/receptor binding for 2,3,7,8-TCDD: implications for risk assessment Fundam Appl Toxicol 20 1 48 56 8381755
Rexer BN Zheng WL Ong DE 2001 Retinoic acid biosynthesis by normal human breast epithelium is via aldehyde dehydrogenase 6, absent in MCF-7 cells Cancer Res 61 19 7065 7070 11585737
Schmidt CK Hoegberg P Fletcher N Nilsson CB Trossvik C Hakansson H 2003 2,3,7,8-Tetrachlorodibenzo-p -dioxin (TCDD) alters the endogenous metabolism of all-trans -retinoic acid in the rat Arch Toxicol 77 7 371 383 12851740
Schmidt S Baniahmad A Eggert M Schneider S Renkawitz R 1998 Multiple receptor interaction domains of GRIP1 function in synergy Nucleic Acids Res 26 5 1191 1197 9469826
Sun SY Wan H Yue P Hong WK Lotan R 2000 Evidence that retinoic acid receptor beta induction by retinoids is important for tumor cell growth inhibition J Biol Chem 275 22 17149 17153 10747926
Tamada Y Kim S Bannai H Imoto S Tashiro K Kuhara S 2003 Estimating gene networks from gene expression data by combining Bayesian network model with promoter element detection Bioinformatics 19 suppl 2 II227 II236 14534194
Tamayo P Slonim D Mesirov J Zhu Q Kitareewan S Dmitrovsky E 1999 Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation Proc Natl Acad Sci USA 96 6 2907 2912 10077610
Tymms MJ Ng AY Thomas RS Schutte BC Zhou J Eyre HJ 1997 A novel epithelial-expressed ETS gene, ELF3 : human and murine cDNA sequences, murine genomic organization, human mapping to 1q32.2 and expression in tissues and cancer Oncogene 15 20 2449 2462 9395241
Unkila M Pohjanvirta R Honkakoski P Torronen R Tuomisto J 1993 2,3,7,8-Tetrachlorodibenzo-p -dioxin (TCDD) induced ethoxyresorufin-O -deethylase (EROD) and aldehyde dehydrogenase (ALDH3) activities in the brain and liver. A comparison between the most TCDD-susceptible and the most TCDD-resistant rat strain Biochem Pharmacol 46 4 651 659 8363638
Varanasi U Chu R Chu S Espinosa R LeBeau MM Reddy JK 1994 Isolation of the human peroxisomal acyl-CoA oxidase gene: organization, promoter analysis, and chromosomal localization Proc Natl Acad Sci USA 91 8 3107 3111 8159712
Vasiliou V Pappa A Petersen DR 2000 Role of aldehyde dehydrogenases in endogenous and xenobiotic metabolism Chem Biol Interact 129 1-2 1 19 11154732
Wolfinger RD Gibson G Wolfinger ED Bennett L Hamadeh H Bushel P 2001 Assessing gene significance from cDNA microarray expression data via mixed models J Comput Biol 8 6 625 637 11747616
Xu J Li Q 2003 Review of the in vivo functions of the p160 steroid receptor coactivator family Mol Endocrinol 17 9 1681 1692 12805412
Yoshida N Yoshida S Araie M Handa H Nabeshima Y 2000 Ets family transcription factor ESE-1 is expressed in corneal epithelial cells and is involved in their differentiation Mech Dev 97 1-2 27 34 11025204
Zhang QY Dunbar D Kaminsky L 2000 Human cytochrome P-450 metabolism of retinals to retinoic acids Drug Metab Dispos 28 3 292 297 10681373
|
15345368
|
PMC1277115
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug 23; 112(12):1217-1224
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7020
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7034ehp0112-00122515345369ToxicogenomicsArticlesValproic Acid Teratogenicity: A Toxicogenomics Approach Kultima Kim Nyström Anna-Maja Scholz Birger Gustafson Anne-Lee Dencker Lennart Stigson Michael Department of Pharmaceutical Biosciences, Division of Toxicology, The Biomedical Center, Uppsala University, Uppsala, SwedenAddress correspondence to M. Stigson, Department of Pharmaceutical Biosciences, Division of Toxicology, Uppsala University, BMC, Box 594, SE-75124 Uppsala, Sweden. Telephone: 46 18 4714208 or 46 730 216605. Fax: 46 18 4714253. E-mail:
[email protected] material is available online (http://ehp.niehs.nih.gov/txg/members/2004/7034/supplemental.pdf).
Microarray data are available in the ArrayExpress database (http://www.ebi.ac.uk/arrayexpress/) at accession numbers A-MEXP-32 and E-MEXP-45.
We thank R. Engdahl and L. Norgren for excellent technical assistance, and J. Hansson, M. Norrman, E. Lopez Fernandez de Villaverde, and K. Zemiran for help. We also thank I. Lönnstedt and D. von Rosen for valuable discussions about statistical issues and data analysis.
This work was supported by grant 02/04-36 from the Swedish Animal Welfare Agency (http://www.djurskyddsmyndigheten.se). M.S was supported by a grant from AstraZeneca, and A-L.G. by the Swedish Medical Research Council (MFR).
The authors declare they have no competing financial interests.
8 2004 3 6 2004 112 12 1225 1235 16 2 2004 3 6 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Embryonic development is a highly coordinated set of processes that depend on hierarchies of signaling and gene regulatory networks, and the disruption of such networks may underlie many cases of chemically induced birth defects. The antiepileptic drug valproic acid (VPA) is a potent inducer of neural tube defects (NTDs) in human and mouse embryos. As with many other developmental toxicants however, the mechanism of VPA teratogenicity is unknown. Using microarray analysis, we compared the global gene expression responses to VPA in mouse embryos during the critical stages of teratogen action in vivo with those in cultured P19 embryocarcinoma cells in vitro. Among the identified VPA-responsive genes, some have been associated previously with NTDs or VPA effects [vinculin, metallothioneins 1 and 2 (Mt1, Mt2), keratin 1-18 (Krt1-18)], whereas others provide novel putative VPA targets, some of which are associated with processes relevant to neural tube formation and closure [transgelin 2 (Tagln2), thyroid hormone receptor interacting protein 6, galectin-1 (Lgals1), inhibitor of DNA binding 1 (Idb1), fatty acid synthase (Fasn), annexins A5 and A11 (Anxa5, Anxa11)], or with VPA effects or known molecular actions of VPA (Lgals1, Mt1, Mt2, Id1, Fasn, Anxa5, Anxa11, Krt1-18). A subset of genes with a transcriptional response to VPA that is similar in embryos and the cell model can be evaluated as potential biomarkers for VPA-induced teratogenicity that could be exploited directly in P19 cell–based in vitro assays. As several of the identified genes may be activated or repressed through a pathway of histone deacetylase (HDAC) inhibition and specificity protein 1 activation, our data support a role of HDAC as an important molecular target of VPA action in vivo.
biomarkerembryocarcinomagalectin-1histone deacetylasein vitro toxicologymetallothioneinmicroarraymouse embryoneural tube defectSp1teratogenvalproic acidvinculin
==== Body
Exposure during pregnancy to pharmaceuticals and environmental chemicals remains a worldwide problem. Assessing risk for human developmental toxicity is a major obstacle in drug development, as it relies on data from animal experiments, with associated concordance problems. A common understanding of basal mechanisms of developmental toxicity could assist risk assessment, but such mechanisms have unfortunately remained elusive. How individual teratogenic agents induce early developmental errors, and how widely different teratogens induce apparently similar defects by common or distinct mechanisms are still largely unknown. Compared with most established adult organs, the mammalian embryo comprises a moving target of highly dynamic cell interactions. This inherent complexity impedes the mechanistic interpretation of a chemical insult and may ultimately preclude what appear as more desirable in vitro methods from completely replacing whole-animal experiments in developmental toxicology. Nevertheless, cell-based screening methods could be devised based on knowledge of molecular mechanisms, pathways, and biomarkers of toxicity.
Recently, toxicogenomics has emerged as an attractive approach to uncover critical molecular events altered by toxicants (Aardema and MacGregor 2002; Iannaccone 2001; Nuwaysir et al. 1999). Using microarrays and profiling techniques, investigators can determine how gene expression responses to toxic exposure are linked to toxic outcome (phenotypic anchoring) (Paules 2003) and identify molecular targets and biomarkers of chemically induced toxicity. However, few microarray studies so far have addressed developmental toxicity (Docterman and Smith 2002) or embryonic development (Ko 2001; Smith and Greenfield 2003). We predict that disruption of the hierarchies of signaling and gene regulatory networks that control embryonic development may underlie many cases of chemically induced birth defects. Teratogenic chemicals are therefore likely to affect downstream gene expression as a cause or consequence, or both, of their adverse developmental effects. Hence, compound-specific gene expression responses should be possible to detect.
In this study we used spotted cDNA microarrays to monitor global gene expression changes in response to the antiepileptic drug valproic acid (VPA), a potent teratogen that most notably induces neural tube defects (NTDs) in human, mouse, and other vertebrate embryos (Lammer et al. 1987; Nau et al. 1991; Oberemm and Kirschbaum 1992; Whitsel et al. 2002). NTDs with varying penetrance can be induced in the mouse embryo by many chemical treatments (Copp et al. 1990) and by the functional disruption of a plethora of genes (Copp et al. 2003; Juriloff and Harris 2000). Induction and development of NTDs in the mouse embryo is thus a relevant model for studying chemically induced teratogenicity. In this context, we believe that VPA is a good model substance to be addressed by a toxicogenomics approach. Although the molecular mechanism by which VPA causes NTDs remains obscure, several genes and molecular targets have been associated with VPA action, both in embryos (Craig et al. 2000; Faiella et al. 2000; Wlodarczyk et al. 1996) and various cell lines (Blaheta and Cinatl 2002; Phiel et al. 2001; Walmod et al. 1999; Werling et al. 2001; Yuan et al. 2001) and therapeutically in epilepsy and bipolar disorders (Gurvich and Klein 2002; Johannessen 2000). We report here the altered expression of multiple genes in mouse embryos after treatment with VPA, and discuss some of these genes in the light of neural tube development and previously known VPA actions. Employing the mouse embryocarcinoma cell line P19 as an in vitro model of early pluripotent embryonic cells, we identify further a subset of VPA-responsive genes that may be particularly relevant to evaluate as potential biomarkers of VPA teratogenicity.
Materials and Methods
Embryos
NMRI mice (B&K Universal AB, Sollentuna, Sweden) were kept on a 12-hr light cycle (1100–2200 hr) in the Laboratory Animal Facility at The Biomedical Center. Females were mated with males for 2 hr at the end of the dark period (0800–1000 hr). Females were then checked for vaginal plugs, and the midpoint of the mating period (0900 hr) was taken as 0 days postcoitum (dpc) Pregnant dams were treated 8.0 dpc by ip injection of 600 mg/kg body weight sodium valproate (Sigma Chemical Co., St. Louis, MO, USA) in approximately 100–200 μL 0.9% saline; control mice received saline only. Dams were sacrificed by cervical dislocation at 1.5 hr [RNA for quantitative reverse transcription-polymerase chain reaction (RT-PCR)], 6 hr (RNA for microarrays and quantitative RT-PCR), or 48 hr [morphological examination and detection of programmed cell death (PCD) by terminal deoxynucleotidyl transferase-mediated (dUTP) biotinylated nick end labeling (TUNEL) staining] posttreatment. The uterus was quickly transferred to phosphate-buffered saline (PBS), pH 7.4, and the embryos were removed. For RNA preparation, embryos were lysed in Trizol reagent (Invitrogen, Carlsbad, CA, USA), and stored at −80°C until further use. Because of some within-litter and between-litter variation in the size and developmental stage of 8.25-dpc embryos (6 hr posttreatment), each embryo was quickly evaluated morphologically before lysis, and embryos that appeared younger than the late bud stage, as defined by Downs and Davies (1993) were excluded. Three pools each of treated and control embryos were created containing embryos from 2, 3, and 12 VPA-treated litters or 3, 4, and 12 control litters, respectively.
Detection of Programmed Cell Death
Embryos removed from control and treated animals were fixed overnight in 4% paraformaldehyde (PF) in PBT (0.1% Tween-20 in PBS), then processed into 100% methanol and stored at −20°C until use. PCD detection was performed using the In Situ Cell Death Detection Kit, AP (Roche Diagnostics, Indianapolis, IN, USA) according to manufacturer instructions, with the following minor changes: permeabilization of the embryos was performed for 10 min in 10 μg/mL proteinase K, followed by 4% PF for 10 min, before inactivation of the endogenous peroxidase. All washing steps were performed with PBT.
Cell Culture
P19 mouse embryocarcinoma cells (ATCC CRL-1825; American Type Culture Collection, Manassas, VA, USA) were cultured at 37°C and 5% CO2 in Dulbecco’s modified Eagle’s medium (National Veterinary Institute, Uppsala, Sweden) supplemented with 10% fetal bovine serum (Seromed, Berlin, Germany), 1% l-glutamine, and 1% penicillin/streptavidin. Cells from a subconfluent T75 flask were split 1:20 onto 10-cm plates (Nunc, Roskilde, Denmark) in 10 mL medium; the next day half the plates were treated with 1 μM sodium valproate by adding 10 μL from a 1-mM stock solution; control plates received 10 μL water. After incubation for 24 hr, the plates were washed twice with PBS, and the cells were lysed with 3 mL Trizol reagent per plate for 5 min at room temperature. Genomic DNA was sheared by drawing the lysate several times through a pipette until it appeared nonviscous. Subsequently, the lysates from all 10 treated and control plates were pooled and stored at −80°C until further use.
RNA Isolation
Total RNA was isolated from frozen embryos and cells using Trizol reagent (Invitrogen) according to manufacturer instructions. RNA concentration was determined spectrophotometrically, and RNA quality was checked using the Agilent 2100 Bioanalyzer and the RNA 6000 LabChip kit (Agilent Technologies, Palo Alto, CA, USA). The yield was approximately 1 μg per 8.25-dpc embryo and 300 μg per confluent plate of P19 cells. The A260/A280 ratios ranged from 1.8–2.0, and the 28S/18S ribosomal RNA ratios were approximately 1.8.
cDNA Synthesis
Equal amounts of RNA from control and treated samples were separately converted to fluorescently labeled cDNA by incorporation of the dye-conjugated nucleotides Cyanine 3 (Cy3)-dCTP or Cy5-dCTP (or vice versa) during first-strand cDNA synthesis. Briefly, 30 μg embryonic RNA or 50 μg cellular RNA was mixed with anchored dT17 primer, heated to 70°C and reverse transcribed with Superscript II reverse transcriptase (Invitrogen) for 2 hr at 42°C in the presence of 100 μM fluorescent nucleotide. After hydrolysis of the RNA template, unincorporated fluorescent nucleotides were removed by ethanol precipitation.
Microarray Hybridization
Spotted cDNA microarrays [mouse NIA clone set arrays, slide 1 (http://www.hgmp.mrc.ac.uk/Research/Microarray/HGMP-RC_Microarrays/array_description_files.jsp)], containing a subset of 6,144 clones from approximately 15,000 developmentally expressed mouse genes in the National Institute on Aging (NIA) 15 K mouse cDNA clone set (http://lgsun.grc.nia.nih.gov/cDNA/15k.html (Tanaka et al. 2000) and spotted in duplicate were purchased from the Human Genome Mapping Project Resource Centre (Hinxton, United Kingdom; http://www.hgmp.mrc.ac.uk/). For each hybridization, equal amounts of fluorescently labeled cDNA were mixed with 4 μg polyA DNA carrier and 6 μg mouse Cot1 DNA (Invitrogen) and subsequently denatured by boiling in hybridization buffer [5× sodium chloride/sodium citrate (SSC); 6× Denhard’s solution; 60 mM Tris, pH 7.6, 0.12% sarcosyl; 48% formamide]. After cooling, 40 μL hybridization mix was applied to the microarray slide, and hybridization was carried out at 50°C for 12–24 hr in a humid hybridization chamber. After hybridization, slides were washed on a shaker at 55°C for 10 min with 2× SSC, 0.2% sodium dodecyl sulfate; 10 min with 2× SSC; 10 min with 0.2× SSC; 1 min with ultrapure water; and 1 min with isopropanol and subsequently dried in a centrifuge for 5 min at 500 × g. For embryos, four separate hybridizations [embryo microarray (EM) 1 through 4] addressed both biological and technical variation, using independent samples for EM1, EM2, and EM3/4, and dye reversal of the same samples (based on 12 litters) for EM3 and EM4. For cells, four separate hybridizations [cell microarray (CM) 1 through 4] intentionally addressed only technical variation with duplicate dye reversal. Control samples were labeled with Cy3 for the odd-numbered hybridizations (i.e., EM1, EM3, CM1, CM3) and with Cy5 for the even-numbered ones.
Microarray Data Analysis
We acquired fluorescent images of microarray slides using a ScanArray confocal laser microarray scanner (Packard BioChip Technologies, Billerica, MA, USA). We quantified fluorescence intensities for the Cy3 and Cy5 channels using the Spot 2.0 software package (Jain et al. 2002). We did not perform background subtraction, as it did not improve the data set. For each spot the base 2 logarithm (log2) ratios between the two channels were used to quantify the fold change in relative gene expression levels between experimental and control samples. To remove systematic sources of variation, we used a within-print group scaled normalization method (Yang et al. 2002). A mean value for the duplicate spots was calculated for each array. A parametric empirical Bayes approach (Lönnstedt and Speed 2002) was used to identify differentially expressed genes. The p-value was fixed at 0.01, and differentially expressed genes were defined as genes with an absolute log of odds score value above 1. Because no spots were excluded in the analysis (e.g., flagged for morphologic or other defects), and the Bayes approach penalizes for both low absolute expression ratio and high variance between duplicate spots on the same or replicate slides, false negatives may result from a bad spot on one of the four slides. To recover such clones, we repeated the analysis, omitting the four arrays one by one. Genes with an absolute log odds > 1 in any analysis were included in the total list of genes. Hierarchical clustering with complete linkage and Euclidian distance as the distance metric was computed using J-Express 2.1 (Dysvik and Jonassen 2001). In the hierarchical clustering, the outlying value for any gene identified by the leave-one-out procedure (above) was replaced in the omitted array by the median based on the three remaining arrays.
Quantitative Real-Time RT-PCR
Six genes identified by microarray analysis were selected for reanalysis by quantitative real-time RT-PCR (qPCR). We selected the supposed housekeeping gene peptidylpropyl isomerase A (Ppia), also known as cyclophilin, as the endogenous reference because our microarray analysis indicated that its expression is unlikely to be altered by VPA in either mouse embryos or P19 cells. The average log2 fold change for Ppia (represented by five clones in duplicate in the NIA 1 array) was −0.08 for embryos and −0.04 for cells (data not shown). Moreover, this gene was previously used as endogenous reference for gene expression analysis at the RNA level in the context of VPA and NTDs (Wlodarczyk et al. 1996). Primers were designed with Primer Express software (Applied Biosystems, Palo Alto, CA, USA), using default setting for the TaqMan mode, and synthesized by Applied Biosystem. Primer sequences are given in Table 1. For qPCR, 2 μg total RNA was reverse transcribed in a final volume of 100 μL using TaqMan Reverse Transcription Reagents (Applied Biosystems) with random hexamer primers according to manufacturer instructions. Reactions excluding MultiScribe Reverse Transcriptase (Applied Biosystems) were performed as negative controls. cDNA targets at a 100-fold final dilution were amplified in replicate wells (four for target genes and six for the endogenous reference), using optimized primer concentrations (Table 1) in 1× qPCR Mastermix Plus for SYBR Green I (Eurogentec, Seraing, Belgium) in an ABI Prism 7700 Sequence Detector System (Applied Biosystems) with the following thermal profile: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C. Standard curves for each gene were obtained by amplifying (in quadruplicate) 10-fold serial dilutions of a reference mixture containing 25% each of cDNA derived from VPA-treated and control embryos and cells. Outlying cycle to threshold (CT) values were detected using median absolute deviation (Young et al. 2003) with an arbitrary threshold > 10, leading to the removal of one data point. Using the standard curves, CT values for target genes were converted to relative input amounts and normalized to the corresponding values for Ppia. Differences in the mean of normalized relative input amounts between VPA-treated and control samples were tested for statistical significance using a two-tailed t-test.
Results
To monitor gene expression changes associated with VPA teratogenicity, we adopted conditions of early exposure previously reported to induce NTDs in approximately 60% of live fetuses in the NMRI strain, as observed 18 dpc (Nau and Löscher 1986). By administering a single dose of sodium valproate (600 mg/kg body weight) ip to 8.0-dpc pregnant NMRI dams and examining embryos for developmental defects 48 hr posttreatment (10.0 dpc), we found that 22 of 42 embryos (52%) from VPA-treated dams had different degrees of NTDs, mostly coupled with growth retardation; 15 (36%) were growth retarded but appeared otherwise morphologically normal; and 5 (12%) had diverse abnormalities such as absence of caudal structures, cardiac dysfunction (no heartbeats), and edema. In contrast, we found that no control embryos from saline-treated dams had NTDs or other apparent developmental anomalies. Unlike several other NTD-inducing teratogens (Mirkes 2002), we found that VPA induced no apparent increase in apoptosis along the tips of the neural folds (Copp et al. 2003), as detected by TUNEL staining (Figure 1). Instead, we found a transversal band of apoptotic cells in the forebrain neuroepithelium of VPA-treated embryos (Figure 1D), which to our knowledge has not been reported previously and is the subject for further investigation.
Valproic Acid–Associated Gene Expression Changes in Mouse Embryos
To study the gene expression response to VPA during the susceptible stages, that is, when VPA exerts most of its teratogenic effect on the developing neural tube, we extracted total RNA from pools of whole embryos removed from control and VPA-treated 8.25-dpc NMRI mice (6 hr posttreatment) and subjected them to replicated microarray analysis. To identify differentially expressed genes, we used an empirical Bayes model (Lönnstedt and Speed 2002) to rank genes by their log posterior odds of differential expression (Figure 2A). We found that 81 clones of the 6,144 cDNAs from the NIA 15 K mouse cDNA clone set (Tanaka et al. 2000) represented in the NIA array 1 were expected (log odds > 1) to be upregulated (51 of 81) or downregulated (30 of 81) in response to VPA [Supplemental Material, Table 1 (http://ehp.niehs.nih.gov/txg/members/2004/7034/supplemental.pdf)]. An additional 14 clones were selected after the leave-one-out procedure described in “Materials and Methods” [Figure 2A; Supplemental Material, Table 2 (http://ehp.niehs.nih.gov/txg/members/2004/7034/supplemental.pdf)]. Among the wide variety of putative VPA-responsive genes thus listed, we found that metallothionein 2 (Mt2) was represented by both top-ranked clones (Supplemental Material, Table 1). Similarly, galectin-1 (Lgals1) appeared to be represented by three clones; karyopherin β1 (Kpnb1) and H3 histone 3B (H3f3b) were represented by two clones each. Approximately one-third of the selected clones appeared to represent uncharacterized or unknown genes (Supplemental Material, Tables 1 and 2). Although the identified candidate genes belong to several functional categories, those encoding matrix/structural proteins appeared to be slightly overrepresented, comprising 35% of the functionally annotated clones (Supplemental Material, Tables 1 and 2).
Valproic Acid–Associated Gene Expression Changes in P19 Embryocarcinoma Cells
Genes that respond transcriptionally to VPA in embryos (Supplemental Material, Tables 1 and 2) may provide not only important clues about mechanisms of VPA action but also potential biomarkers of VPA teratogenicity that could be exploited in a cell-based screening system. Toward this goal, we employed the pluripotent P19 mouse embryocarcinoma cell line (McBurney 1993) as a possibly relevant cell model for early embryos. To identify general VPA-responsive genes at a dose level close to the range of therapeutic and teratogenic concentrations while attaining convenient and supposedly robust bioassay conditions, total RNA was extracted from P19 cells cultured in the presence or absence of 1 mM VPA for 24 hr and subjected to replicated microarray analysis. Ranking the genes by their log posterior odds of differential expression (Figure 2B), we found 168 clones expected (log odds > 1) to be upregulated (114 of 168) or downregulated (54 of 168) in response to VPA [Supplemental Material, Table 3 (http://ehp.niehs.nih.gov/txg/members/2004/7034/supplemental.pdf)], with 16 additional clones selected by the leave-one-out procedure [Figure 2B; Supplemental Material, Table 4 (http://ehp.niehs.nih.gov/txg/members/2004/7034/supplemental.pdf)]). Again, approximately one-third of the selected clones appeared to represent uncharacterized or unknown genes (Supplemental Material, Tables 3 and 4). Although less apparent than for embryos (Supplemental Material, Tables 1 and 2), genes encoding matrix/structural proteins again represented the largest functional category, comprising 28% of the functionally annotated clones (Supplemental Material, Tables 3 and 4). The similar magnitude of change in cells (Figure 2B; Supplemental Material, Tables 3 and 4) and embryos (Figure 2A; Supplemental Material, Tables 1 and 2) for VPA-responsive genes may support the relevance of the VPA dose (1 mM) used in vitro.
Confirmation of Valproic Acid–Responsive Genes
To independently assess the altered expression of genes identified by microarray analysis, we arbitrarily selected six genes that appeared biologically relevant while showing diverse responses to VPA in the two-model systems (Figure 2). Using qPCR, we found that the selected genes expected to be upregulated by VPA in embryos [Lgals1, Mt2, and vinculin (Vcl)] or cells [keratin 1-18 (Krt1-18), Lgals1, and Mt2] were significantly (p < 0.05) induced (Figure 3). Similarly, the gene expected to be downregulated by VPA in cells [uridine phosphorylase (Upp)] was significantly (p < 0.05) repressed (Figure 3). Unlike the microarray analysis, using qPCR we could detect a weak but significant (p < 0.05) induction of Krt1-18 and Vcl in embryos and cells, respectively (Figure 3). The microarray analysis appeared to underestimate the fold change of expression compared with qPCR (Figure 3), with the only exception being Kpnb1, for which a downregulation in embryos was not supported by qPCR (Figure 3).
Identification of Potential Biomarkers of Valproic Acid Teratogenicity
Genes that respond similarly to a teratogen in a cultured cell model as in intact embryos might be directly exploited as biomarkers in an in vitro test system, using the same cell line. To identify candidates for such potential biomarkers, we compared the results presented in Supplemental Material, Tables 1–4, and found 29 clones (three of which were recovered in embryos by the leave-one-out procedure) likely to be VPA responsive in both embryos and P19 cells (Table 2). These clones probably represent no more than 25 genes, of which 16 currently have known identity (Table 2). Among these genes are several that were top ranked in embryos (Supplemental Material, Table 1), such as Mt2, metallothionein 1 (Mt1), Lgals1, H3f3b, creatine kinase–brain (Ckb), and transgelin 2 (Tagln2). The similar transcriptional response to VPA in the cell model as in embryos strengthens the case not only for these genes as VPA targets in the embryo but also for a number of other genes (Table 2; Supplemental Material, Tables 1 and 2) such as cytochrome c oxidase subunit VIIa polypeptide 2-like (Cox7a2l), ubiquitin carboxy-terminal hydrolase L1 (Uchl1), eukaryotic translation initiation factor 4 gamma 2 (Eif4g2), bromodomain containing protein 4 (Brd4), annexin A11 (Anxa11), leukotriene B4 12-hydroxydehydrogenase (Ltb4dh), inhibitor of DNA binding 1 (Idb1), and fatty acid synthase (Fasn).
As we used a cutoff for differential expression intended to minimize the number of false positives, the number of genes with a similar transcriptional response to VPA in the cell model as in embryos might be underestimated in our analysis. By clustering the mean log2 fold changes measured in the eight individual microarray slides for all 220 selected clones (i.e., clones with log odds > 1 in either embryos or cells), we found that most of the clones displayed in Table 2 form two well-defined clusters of commonly upregulated genes (clusters C1 and C2), whereas the rest of the clones are found within two indistinct clusters of commonly upregulated (cluster C3) or downregulated (cluster C4) genes (Figure 4). Across all 6,144 clones, we found that the highest log odds score for which the transcriptional change in embryos was not in the same direction as in cells (disregarding the magnitude of change) was −0.57. Among the 122 clones above this level in embryos, the highest and lowest log2 fold changes detected were 0.33 and −0.31, respectively. Applying log2 fold change > 0.3 and < −0.3 (corresponding to > 27% fold change) as the cutoff, we could identify 41 additional clones as putative candidates for genes with a similar transcriptional response to VPA in the cell model as in embryos (Figure 4).
Rapid Valproic Acid–Induced Transcriptional Activation of the Genes Encoding Galectin-1 and Vinculin
Toward understanding the mechanism of VPA action and validating candidate genes as potentially useful biomarkers of VPA responses, we reinvestigated the VPA-induced transcriptional response of selected genes (Figure 2) in mouse embryos. VPA reaches peak levels in mouse serum about 30 min after a single ip injection, and the half-time of VPA clearance from mouse serum is about 1 hr (Nau et al. 1991). To investigate whether VPA at such a peak concentration can induce a rapid transcriptional response of selected genes, we extracted total RNA from pools of whole embryos removed from control and VPA-treated pregnant NMRI dams 1.5 hr posttreatment (dose and gestational day of treatment as before) and subjected to qPCR. Among the six tested genes, the expression of Lgals1 and Vcl were significantly (p < 0.05) induced after 1.5 hr (Figure 5), which may support independently the VPA responsiveness of these two genes.
Discussion
Although several modes of VPA action have been proposed [e.g., histone deacetylase (HDAC) and protein kinase (PKC) inhibition, extracellular signal-regulated kinase (ERK) and activator protein-1 (AP-1) activation, and effects on the actin cytoskeleton] (Blaheta and Cinatl 2002; Gurvich and Klein 2002; and Walmod et al. 1999), the mechanism of VPA teratogenicity remains poorly understood. In this study we used microarrays to monitor gene expression changes in response to VPA during stages critical to VPA-induced NTDs in the mouse embryo. Some of the more than 70 putative VPA target genes thus identified (Supplemental Material, Table 1) have previously been directly or indirectly linked to VPA effects or to NTDs or processes relevant to neural tube formation and closure, but most appear to be novel candidates. Moreover, we propose that many of these genes by virtue of their similar expression changes in embryos and cultured P19 mouse embryocarcinoma cells (Table 2; Figure 4) could be directly exploited as potential biomarkers of VPA action in cell-based assays.
To date, mouse embryos have not been extensively studied by microarray analysis (Carter et al. 2003; Smith and Greenfield 2003), partly because of their small size and limited RNA content. Here we overcome this limitation by pooling whole embryos from several similarly treated mice. Despite our primary goal to identify candidate genes that may be VPA targets in disturbed neural tube closure, the use of whole embryos rather than isolated neural tubes may be warranted for at least three reasons. First, VPA accumulates in the neuroepithelium (Dencker et al. 1990), which constitutes a major region of the mouse embryo during the stages investigated. Second, neural tube closure can be influenced by genes that are mostly or only expressed outside the neural tube [e.g., cartilage homeoprotein 1 (Cart1), Twist, and sonic hedgehog (Shh)] (Copp et al. 1990). Third, potential dissection artifacts (Diaz et al. 2003) are minimized. In addition, bulk approaches such as the pooling strategy we used in this study may be warranted to allow the study of gene expression changes during early stages of neural tube development without the need to know which individual embryos will subsequently develop NTDs in response to the VPA treatment (50–60%). An unfavorable consequence of pooling whole embryos from multiple litters, however, is that we dilute the expression changes for those genes expressed only in certain defined regions of the embryo, such as distinct areas of the neural tube, as well as for any genes that may be responding to VPA mostly or only in those embryos that will become malformed. To some degree, this may account for why we are able to identify fewer differentially expressed genes in the embryo (Supplemental Material, Tables 1 and 2) than in the P19 cell line (Supplemental Material, Tables 3 and 4).
The formation and closure of the neural tube is a highly coordinated set of events that involves a multitude of morphogenetic movements and regulated cell behavior (Copp et al. 2003; DeSesso et al. 1999; Smith and Schoenwolf 1997). Several null mutations for actin-regulating genes have been reported to be associated with NTDs (Copp et al. 2003), illustrating that proper regulation of cell shape and cell movements is crucial for neurulation processes to occur normally. In this study, we found that the expression of the gene encoding Vcl is rapidly induced in embryos after VPA treatment (Figure 5). Vinculin, which is essential for neural tube closure (Xu et al. 1998), is an actin-binding protein associated with focal adhesions (De Arcangelis and Georges-Labouesse 2000) and has previously been reported to be increased in such points of integrin-mediated cell–matrix interactions after VPA treatment in vitro (Walmod et al. 1999). Some integrins and extracellular matrix components, along with components downstream of integrin signaling, also appear to be essential for neural tube closure (Brouns et al. 2000; De Arcangelis and Georges-Labouesse 2000; Juriloff and Harris 2000). Adverse effects of VPA on the dynamics of the actin cytoskeleton may therefore contribute to VPA teratogenicity, as has been previously suggested (Walmod et al. 1999). The supposed actin dependency of anterior but not posterior neuropore closure in mouse embryos (Ybot-Gonzalez and Copp 1999) could thus explain why exencephaly is the dominant NTD observed in VPA-exposed mice (Nau et al. 1991). Although emerging as a conceivable VPA target from a mechanistic point of view, vinculin might be less straightforward to exploit as a biomarker of VPA effects given the weak response in our cell model (Figure 3). Conversely, we found that the expression of the gene encoding Tagln2, an actin-binding protein with the ability to cross-link actin filaments (Shapland et al. 1993), was induced, and the genes encoding spermidine synthase (Srm), an actin-regulating protein (Caruso et al. 1994), and thyroid hormone receptor interactor 6 (Trip6 ), a focal adhesion-binding protein with nuclear shuttling activity, were repressed in response to VPA in both embryos and P19 cells (Table 2; Figure 4). Our findings may support a role of integrin-mediated actin regulation in VPA teratogenicity, even though we were unable to observe any apparent reorganization of the actin cytoskeleton, as visualized by phalloidin staining (data not shown), in P19 cells treated with VPA at the present concentration (1 mM).
Along these lines, we found the expression of the gene Lgals1, which encodes the β-galactoside–binding protein galectin-1 (Barondes et al. 1994), was induced by VPA in both embryos and cells (Table 2; Figure 4). Galectin-1 is a multifunctional homodimeric lectin whose extracellular and intracellular activities are thought to regulate cellular processes as diverse as cell–matrix interactions, signal transduction, migration, differentiation, proliferation, apoptosis, and RNA splicing (Hughes 2001; Liu et al. 2002; Perillo et al. 1998). Binding of galectin 1 to the extracellular portion of β1 integrin (Moiseeva et al. 2003) may modulate cell adhesion to extracellular matrix components such as fibronectin and laminin and activate downstream events of integrin signaling (Giancotti and Ruoslahti 1999). Intracellularly, galectin-1 appears to be recruited by the G-protein H-Ras, a membrane-associated transducer of integrin and receptor tyrosine kinase signaling (Giancotti and Ruoslahti 1999), to stabilize its active guanosine triphosphate (GTP)-bound state and membrane anchorage in microdomains segregated from lipid rafts (Hancock 2003). Overexpression of galectin-1 and its binding to H-Ras-GTP may thus enhance signaling through Raf1 and the ERK pathway (Hancock 2003), which along with the downstream effector AP-1 is activated by VPA (Blaheta and Cinatl 2002; Yuan et al. 2001).
The rapid activation of the Lgals1 gene (Figure 5) indicates that VPA may act at the immediate level of the Lgals1 promoter. Because VPA is a direct HDAC inhibitor (Göttlicher et al. 2001; Phiel et al. 2001) and because other HDAC inhibitors such as butyrate and trichostatin A (TSA) induce galectin-1 expression at the transcriptional level (Lu and Lotan 1999), it is likely that VPA induces expression at the Lgals1 promoter by virtue of its activity as an HDAC inhibitor. As a short-chain carboxylic acid structurally related to VPA, butyric acid has been reported to be teratogenic in whole-embryo culture (Coakley et al. 1986), and the structurally unrelated TSA has been reported to induce NTDs in mouse embryos in vitro (Svensson et al. 1998). Hdac1−/−mouse embryos die severely growth retarded before 10.5 dpc, and Hdac1−/− embryonic stem cells proliferate poorly, indicating that the silencing of gene expression by Hdac1 is essential for cell proliferation during embryonic development (Lagger et al. 2002). As shown for other HDAC inhibitors, the effects on gene expression by VPA may depend on the transcription factor specificity protein 1 (Sp1) (Arinze and Kawai 2003), which binds to GC-rich promoter elements (Suske 1999). The DNA-binding activity of Sp1 is modulated by direct interaction with HDAC (Doetzlhofer et al. 1999), and it has been suggested that VPA and TSA inhibit the activity of HDAC by interfering with its catalytic site (Göttlicher et al. 2001). Butyrate may affect Sp1-mediated induction of Lgals1 transcription by a mechanism, as yet unknown, distinct from that of TSA (Lu and Lotan 1999). It is therefore tempting to speculate that VPA, by virtue of its known blocking of TSA binding to HDAC (Göttlicher et al. 2001) and structural similarity to butyric acid, could act by either or both of these mechanisms to induce gene expression at the Lgals1 promoter. Hence, the activation of the ERK/AP-1 pathway by VPA could occur downstream of HDAC inhibition and Sp1-induced galectin-1 overexpression.
The transcriptional induction we found in embryos and P19 cells (Table 2; Figure 4) of the genes encoding the metal-regulating proteins Mt1 and Mt2 may also be attributed to the HDAC-inhibitory activity of VPA (Marks et al. 2003), as these genes may be activated through Sp1 derepression (Ogra et al. 2001) or chromatin-opening histone acetylation (Ghoshal et al. 2002). Butyrate increases mRNA levels of both these genes in embryocarcinoma cell lines (Andrews and Adamson 1987), and VPA increases the level of metallothionein protein in the liver of mice (Kaji and Mikawa 1991) and pregnant rats (Bui et al. 1998; Keen et al. 1989), causing zinc depletion in the embryos. In the present study we found induced metallothionein expression also in the embryo after maternal VPA administration (Table 2), possibly exacerbating the depletion of Zn2+ available for developmental processes. It is conceivable that disturbed Zn2+ availability could affect, among other processes, the activity of zinc finger–containing transcription factors essential for normal neural tube development, such as Yin Yang 1 (YY1) (Donohoe et al. 1999) and Zic1 through Zic3 (Klootwijk et al. 2000; Nagai et al. 1997, 2000). Because of this evidence and the association between Zn deficiency and human NTDs (Zimmerman 1984), disrupted Zn homeostasis may appear as an attractive mechanism for VPA teratogenesis despite some evidence against Zn deficiency as the cause of VPA-induced exencephaly in the mouse embryo (Wegner et al. 1990). Copper depletion could also be the culprit because metallothionein binds Cu2+ more strongly than Zn2+ (Holt et al. 1980), VPA treatment enhances copper excretion (Kuzuya et al. 2002), and knockout of the copper transporter Ctr1 results in NTDs (Lee et al. 2001). An obvious weakness with the concept of VPA teratogenesis being mediated by metallothionein induction is that Mt1 and Mt2 are coordinately induced by such a wide variety of stressors (Andrews 2000) that they could reasonably be dismissed as being part of a general stress response (Brady 1981) rather than being linked to a specific compound such as VPA. However, exposure to an HDAC inhibitor may not be automatically stressful in this regard. TSA alone, for example, activates the MT1 promoter in some cell types (Dressel et al. 2000) but not in others (Ghoshal et al. 2002).
Metallothioneins are also potent antioxidants (Andrews 2000). The role of reactive oxygen species (ROS) in developmental toxicity is well documented (Fantel 1996). Their role in VPA teratogenicity, however, remains unclear, although ROS production has been detected in response to VPA in vitro (Na et al. 2003), and VPA-induced NTDs are prevented by the antioxidant vitamin E in vivo (Al Deeb et al. 2000). It is therefore interesting that we found the gene Ltb4dh, which encodes a protein with antioxidant properties (Dick et al. 2001), induced by VPA in both embryos and P19 cells (Table 2).
In addition to the known developmental importance of a gene, as determined by the phenotypes of knockout mice, a gene’s putative involvement in neural tube development may also be inferred from its expression domain. For example, we find a reduced expression of the gene Fasn in both embryos and P19 cells (Table 2). This gene is normally expressed in developmentally active regions such as the dorsal neural folds of the closing neural tube (Chirala et al. 2003). If Fasn protein is crucial for these processes, VPA-induced downregulation of Fasn gene expression could evidently disturb them. Considering that homozygous Fasn knockout mice die before implantation and also that most of the Fasn+/− mice die or develop abnormally (Chirala et al. 2003), Fasn emerges as a putative target for teratogen action. Intriguingly, the Fasn gene is transcriptionally regulated by Sp1 (Fukuda et al. 1999), again pointing toward HDAC inhibition as a potential cause for the VPA responsiveness.
Another way by which a gene may be associated with NTDs is if its product acts in a pathway essential for neural tube development. Bone morphogenetic protein (BMP) signaling is likely to represent such a pathway, given that knockout of both the BMP signaling transducer Smad5 (Chang et al. 1999) and the BMP inhibitor noggin (McMahon et al. 1998) results in NTDs (mice homozygous for null alleles of the genes encoding BMP-2, BMP-4, and BMP receptor type IA die before or around neurulation). Members of the inhibitor of DNA binding family, which are dominant negative regulators of basic helix–loop–helix transcription factors with diverse cellular effects, are among the most important downstream targets and effectors of BMP signaling (Miyazono and Miyazawa 2002). The enhanced expression of the Idb1 gene we observed in both VPA-treated embryos and cells (Table 2) could therefore reflect or mimic disturbed BMP signaling. Idb1 gene activation by Smads is inhibited by YY1 (Kurisaki et al. 2003), the repressive activity of which may be mediated by interactions with HDAC1 (Yang et al. 1996) and Sp1 (Lee et al. 1993) through a GC-rich Sp1/YY1-binding enhancer site in the Idb1 gene (Lopez-Rovira et al. 2002). HDAC inhibition, by relieving both Sp1 and YY1 repression, may therefore cause dysregulation of Idb1 gene activity.
It is striking that so many of the identified VPA-responsive genes encode for proteins that are multifunctional or have integrative activities, galectin-1 being an obvious example. Similar to galectin-1, members of the annexin family have intracellular, extracellular, and membrane-bound functions and are involved in cell–matrix interactions, cell growth, and differentiation (Seaton and Dedman 1998). The VPA-induced transcriptional induction of annexin A5 (Anxa5) and Anxa11 we detect in both embryos and cells (Table 2) may depend on HDAC inhibition, as the mouse Anxa5 and the human ANXA11 genes have Sp1 sites in their promoters (Bances et al. 2000; Rodriguez-Garcia et al. 1999). Interestingly, annexins may be substrates for and negatively regulate phosphatidylinostiol-dependent PKC activity, the inhibition of which is an established VPA effect (Blaheta and Cinatl 2002). Kpnb1 (also known as importin-β) may also be categorized as multifunctional, as it is a nuclear transport factor with additional roles as a chaperone in the cytoplasm and during mitosis (Jäkel et al. 2002). Although Kpnb1 is a gene whose expression has been reported to be downregulated by HDAC inhibition (Marks et al. 2003), our present data do not provide conclusive evidence for its repression by VPA in mouse embryos (Supplemental Material, Table 1; Figure 3). Trip6 is another good example of a protein with dual localization and function. In addition to its association with focal adhesions, it functions as an intracellular signaling molecule that shuttles between the cell surface and the nucleus (Wang et al. 1999), where it acts as a transcription factor.
Recently, the expression of Trip6, along with Eif4g2, and particularly Upp, was reported to be stem cell specific (Ramalho-Santos et al. 2002). Essentially, these genes were identified as part of a set of genes defining “stemness,” that is, promotion of cell self-renewal and suppression of differentiation. It is therefore interesting that we found the expression of Eif4g2 and Trip 6 downregulated in response to VPA in P19 cells and embryos (Table 2; Figure 4) and the expression of Upp downregulated in P19 cells (Figure 3; Supplemental Material, Table 4). Similarly, Krt1-18 has been reported as a marker of stem cell differentiation (Kelly and Rizzino 2000). Increased expression of the Krt1-18 gene has recently been reported to be a marker for VPA-induced differentiation in F9 mouse embryocarcinoma cells (Werling et al. 2001), a cell line similar to P19 cells. An Sp1 site in the Krt1-18 promoter is important for the expression of this gene (Gunther et al. 1995), which is also activated by butyrate and TSA in F9 cells (Miyashita et al. 1994), suggesting that VPA may activate Krt1-18 expression through HDAC inhibition. Our data support that Krt1-18 and the co-regulated Krt2-8 gene may be VPA inducible in cultured embryocarcinoma cells (Figure 3; Supplemental Material, Table 2) and that Krt1-18 may respond weakly in the embryo (Figure 3).
We conclude that microarray-based toxicogenomics approaches may be useful for identifying target genes and biomarkers of developmental toxicity. By linking gene expression changes to toxic outcome, we detected alterations in gene expression at the level of whole embryos that may be further investigated in terms of hypotheses about mechanisms underlying defective neural tube development (Vcl, Tagln2, Trip6, Mt1, Mt2, Fasn, Id1). By comparing gene expression changes in whole embryos with those in a cultured cell model, we defined a subset of VPA-responsive genes that may be evaluated as potential biomarkers of VPA teratogenicity (Lgals1, Id1, Fasn, Anxa5). A recurrent theme among these genes, as well as for others (Mt1, Anxa11, Krt1-18), is that they may be activated or repressed through HDAC inhibition and Sp1 activation, indicating that HDAC may be a primary molecular target of VPA action in vivo. It remains to be determined to what extent the disruptive effect of VPA on neural tube development may be compounded from the deregulation of a wide variety of target genes acting downstream of HDAC inhibition. Our toxicogenomics approach provides a framework for further studies of developmental toxicity induced by VPA and other chemicals, addressing parameters such as dose, time, and duration of exposure, and genetic susceptibility. In summary, the parallel use of in vivo and in vitro models in conjunction with global expression profiling emerges as a relevant approach toward the identification of biomarkers associated with toxicity after exposure to a wide variety of environmental teratogens.
Supplementary Material
Supplemental Tables Figure 1 NTDs and apoptosis in VPA-exposed mouse embryos. Whole 10-dpc embryos, stained with the TUNEL technique, viewed from the right (A,B) and front (C,D). Abbreviations: ba, first branchial arch; fb, forebrain; mb, midbrain; nfs, neural folds. Control (A,C) and VPA-treated embryos (B,D) were removed (48 hr posttreatment) from the uteri of NMRI dams after ip administration of sodium valproate (600 mg/kg body weight) on 8.0 dpc. Note that VPA-exposed embryos exhibit unfused neural folds, resulting in apparent signs of failed anterior neural tube closure (black arrowheads in B). Apoptotic cells (dark) are seen along the line of neural fold fusion in control embryos (white arrowheads in C) but not in VPA-exposed embryos (D), where instead a transversal band of apoptotic cells can be seen in the neuroepithelium of the forebrain (D). Angles of views in C and D are indicated by white arrows in A and B, respectively.
Figure 2 Microarray analysis of transcriptional response to VPA in (A) mouse embryos and (B) P19 mouse embryocarcinoma cells. The log posterior odds for each clone to be differentially expressed are plotted against the log2 fold change of expression for all cDNA clones in the NIA array 1 (see “Materials and Methods”), based on analysis including four replicate microarray slides. The horizontal line marks the threshold (log odds > 1) for selection of a clone as differentially expressed. Upregulated clones are labeled red, and downregulated clones are labeled blue. The clones under the threshold line labeled red or blue were selected (log odds > 1) by leaving out either one of the four replicate slides from the analysis (see “Materials and Methods”). Arrows indicate clones representing the six genes Kpnb1, Krt1-18, Lgals1, Mt2, Upp, and Vcl selected for reanalysis by qPCR (Figure 3). (A) Transcriptional response in 8.25-dpc embryos (6 hr posttreatment) from pregnant NMRI mice after ip administration of sodium valproate (600 mg/kg body weight) on 8.0 dpc. (B) Transcriptional response in P19 cells cultured in the presence of 1 mM sodium valproate for 24 hr.
Figure 3 Comparison of log2-transformed expression ratios in embryos and P19 cells as determined by microarray analysis and qPCR.
*Significant difference between VPA-treated and control, p < 0.05.
Figure 4 Hierarchical two-way clustering of all 220 genes expected to be transcriptionally responsive to VPA (log odds > 1) in either embryos (Supplemental Material, Tables 1 and 2) or cells (Supplemental Material, Tables 3 and 4), using the mean log2 fold change of expression (represented by a blue–red color scale; bottom) on both sets of four replicate microarray slides (embryos: EM1–EM4 and cells: CM1–CM4; see “Materials and Methods”). The thick vertical lines to the right of the heat map mark the 29 clones with log odds > 1 that respond similarly to VPA in both embryos and cells (black), and the additional 41 clones with log2 fold changes > 0.3 or < −0.3 in both embryos and cells (gray). At the right, four discernible clusters (C1–C4) are marked with thin vertical lines.
Figure 5 Transcriptional response in embryos 1.5 and 6 hr posttreatment after ip administration of sodium valproate (600 mg/kg body weight) on 8.0 dpc, as determined by qPCR.
*Significant difference between VPA-treated and control, p < 0.05.
Table 1 PCR primers.
Gene symbol Forward primer sequence Forward primer concentration (nM) Reverse primer sequence Reverse primer concentration (nM) Amplicon length (nt)
Kpnb1 5′-GGGAATCGTCCAGGGATTG- 3′ 900 5′-AAATAAATTCTACTCTGGCTGTACCA- 3′ 900 83
Krt1-18 5′-AATCGAGGCACTCAAGGAAGAA- 3′ 300 5′-GGCATCCACTTCCACGTCA- 3′ 300 112
Lgals1 5′-GAATCTCTTCGCTTCAGCTTCA- 3′ 50 5′-CAGGTTTGAGATTCAGGTTGCT- 3′ 50 68
Mt2 5′-CGCCATGGACCCAACT- 3′ 50 5′-AGGAAGTACATTTGCATTGTTTGC-3′ 50 89
Upp 5′-TCACCATCATCCGCATTGG- 3′ 300 5′-GCCTGCTGCGTGATGACA- 3′ 900 73
Vcl 5′-TGCCAAGCAGTGCACAGATAA- 3′ 50 5′-GGTCCGGCCCAGCATAGT- 3′ 50 124
Ppia 5′-TTCCTCCTTTCACAGAATTATTCCA- 3′ 50 5′-CCGCCAGTGCCATTATGG -3′ 50 75
Table 2 Genes responding transcriptionally to VPA in embryos and P19 cells.
Embryos
Cells
NIA EST ID Gene symbola Gene namea Log odds Log2 fold change Log odds Log2 fold change Functionb
H3014B09 Anxa11 Annexin A11 2.2 0.58 7.1 0.97 Matrix/structural proteins
H3056G04 Brd4 Bromodomain containing 4 2.5 0.76 8.6 0.81 Signal transduction
H3055H05 Carhsp1 Calcium-regulated heat stable protein 1 1.1 0.42 3.6 0.59 ?
H3007E10 Ckb Creatine kinase, brain 8.2 1.45 10.0 1.01 Matrix/structural proteins
H3024F03 Cox7a2l Cytochrome c oxidase subunit VIIa polypeptide 2-like 3.0 0.50 4.0 0.52 ?
H3027A06 Eif4g2 Eukaryotic translation initiation factor 4 gamma 2 2.7 −0.48 2.3 −0.44 Protein synthesis/translational control
H3011D03 Fasn Fatty acid synthase 1.7 −0.68 1.7 −0.46 Matrix/structural proteins
H3054D02 H3f3b H3 histone, family 3B 5.9 0.71 8.4 0.81 Transcription/chromatin
H3013D08 H3f3b H3 histone, family 3B 5.7 0.66 6.9 0.92 Transcription/chromatin
H3003F10 Idb1 Inhibitor of DNA binding 1 1.8 0.51 5.2 0.65 Transcription/chromatin
H3009D05 Lgals1 Lectin, galactose binding, soluble 1 8.7 1.02 6.4 0.86 Matrix/structural proteins
H3003A03 Lgals1 Lectin, galactose binding, soluble 1 4.2 0.88 7.4 0.84 ?
H3022G07 Lgals1 Lectin, galactose binding, soluble 1 3.1 0.67 3.1 0.58 Matrix/structural proteins
H3040C04 Ltb4dh Leukotriene B4 12-hydroxydehydrogenase 1.8 0.69 9.0 1.34 ?
H3020C02 Mt1 Metallothionein 1 10.3 1.26 5.4 0.71 Heat shock/stress
H3013D11 Mt2 Metallothionein 2 11.6 1.58 2.2 0.90 Heat shock/stress
H3010E09 Mt2* Metallothionein 2* 11.4 1.60 4.0 0.67 Heat shock/stress**
H3012D03 Pmf1 Polyamine-modulated factor 1 1.6 −0.53 4.1 −0.66 ?
H3031E04 Rpo1-4 RNA polymerase 1–4 1.8 −0.46 3.5 −0.64 Protein synthesis/translational control
H3017H12 Tagln2 Transgelin 2 4.4 0.64 10.3 1.09 ?
H3059F01 Uchl1 Ubiquitin carboxy-terminal hydrolase L1 2.9 0.93 1.2 0.59 ?
H3010C05 1110007A10Rik RIKEN cDNA 1110007A10 gene 1.3 0.38 1.5 0.41 ?
H3031A04 2410016F19Rik RIKEN cDNA 2410016F19 gene 3.0 −0.74 2.7 −0.46 ?
H3009E04 2410043F08Rik RIKEN cDNA 2410043F08 gene 1.2 0.64 2.6 0.56 ?
H3003F11 9130023P14Rik RIKEN cDNA 9130023P14 gene 1.6 0.63 3.3 0.85 Matrix/structural proteins
H3004B11 A930014C21Rik RIKEN cDNA A930014C21 gene 2.8 0.51 5.0 0.81 ?
H3022F11 ? ? 1.9 −0.49 1.2 −0.46 ?
H3005E03 ? ? 1.6 0.73 10.6 1.29 ?
H3013C04 ? ? 1.3 0.44 5.1 0.63 ?**
Abbreviations: ?, unknown genes and/or functions; EST, expressed sequence tag; ID, identifier.
a Gene symbols are from the NIA web site (http://lgsun.grc.nia.nih.gov/cDNA/15k.html) as of 30 May 2003, except where indicated by asterisk.
b Gene names are from SOURCE (http://source.stanford.edu/cgi-bin/source/SourceSearch) as of 19 September 2003, except where indicated by asterisk.
* Annotation by tBLASTx (http://www.ncbi.nlm.nih.gov/BLAST/).
** Annotation performed manually, i.e., computer-assigned functions of unknown genes were removed, or known functions for genes identified by tBLASTx were added.
==== Refs
References
Aardema MJ MacGregor JT 2002 Toxicology and genetic toxicology in the new era of “toxicogenomics”: impact of “-omics” technologies Mutat Res 499 1 13 25 11804602
Al Deeb S Al Moutaery K Arshaduddin M Tariq M 2000 Vitamin E decreases valproic acid induced neural tube defects in mice Neurosci Lett 292 3 179 182 11018306
Andrews GK 2000 Regulation of metallothionein gene expression by oxidative stress and metal ions Biochem Pharmacol 59 1 95 104 10605938
Andrews GK Adamson ED 1987 Butyrate selectively activates the metallothionein gene in teratocarcinoma cells and induces hypersensitivity to metal induction Nucleic Acids Res 15 13 5461 5475 3601676
Arinze IJ Kawai Y 2003 Sp family of transcription factors is involved in valproic acid-induced expression of galphai2 J Biol Chem 278 20 17785 17791 12624107
Bances P Fernandez MR Rodriguez-Garcia MI Morgan RO Fernandez MP 2000 Annexin A11 (ANXA11) gene structure as the progenitor of paralogous annexins and source of orthologous cDNA isoforms Genomics 69 1 95 103 11013079
Barondes SH Cooper DN Gitt MA Leffler H 1994 Galectins. Structure and function of a large family of animal lectins J Biol Chem 269 33 20807 20810 8063692
Blaheta RA Cinatl J Jr 2002 Anti-tumor mechanisms of valproate: a novel role for an old drug Med Res Rev 22 5 492 511 12210556
Brady FO 1981 Synthesis of rat hepatic zinc thionein in response to the stress of sham operation Life Sci 28 14 1647 1654 7242251
Brouns MR Matheson SF Hu KQ Delalle I Caviness VS Silver J 2000 The adhesion signaling molecule p190 RhoGAP is required for morphogenetic processes in neural development Development 127 22 4891 4903 11044403
Bui LM Taubeneck MW Commisso JF Uriu-Hare JY Faber WD Keen CL 1998 Altered zinc metabolism contributes to the developmental toxicity of 2-ethylhexanoic acid, 2-ethylhexanol and valproic acid Toxicology 126 1 9 21 9585088
Carter MG Hamatani T Sharov AA Carmack CE Qian Y Aiba K 2003 In situ -synthesized novel microarray optimized for mouse stem cell and early developmental expression profiling Genome Res 13 5 1011 1021 12727912
Caruso A Pellati A Bosi G Arena N Stabellini G 1994 Effects of spermidine synthase inhibition on cytoskeletal organization in cultured chick embryo fibroblasts Eur J Histochem 38 3 245 252 7530518
Chang H Huylebroeck D Verschueren K Guo Q Matzuk MM Zwijsen A 1999 Smad5 knockout mice die at mid-gestation due to multiple embryonic and extraembryonic defects Development 126 8 1631 1642 10079226
Chirala SS Chang H Matzuk M Abu-Elheiga L Mao J Mahon K 2003 Fatty acid synthesis is essential in embryonic development: fatty acid synthase null mutants and most of the heterozygotes die in utero Proc Natl Acad Sci USA 100 11 6358 6363 12738878
Coakley ME Rawlings SJ Brown NA 1986 Short-chain carboxylic acids, a new class of teratogens: studies of potential biochemical mechanisms Environ Health Perspect 70 105 111 3830097
Copp A Brook FA Estibeiro JP Shum AS Cockroft DL 1990 The embryonic development of mammalian neural tube defects Progr Neurobiol 35 363 403 2263736
Copp AJ Greene ND Murdoch JN 2003 The genetic basis of mammalian neurulation Nat Rev Genet 4 10 784 793 13679871
Craig JC Bennett GD Miranda RC Mackler SA Finnell RH 2000 Ribonucleotide reductase subunit R1: a gene conferring sensitivity to valproic acid-induced neural tube defects in mice Teratology 61 4 305 313 10716750
De Arcangelis A Georges-Labouesse E 2000 Integrin and ECM functions: roles in vertebrate development Trends Genet 16 9 389 395 10973067
Dencker L Nau H D’Argy R 1990 Marked accumulation of valproic acid in embryonic neuroepithelium of the mouse during early organogenesis Teratology 41 6 699 706 2353317
DeSesso JM Scialli AR Holson JF 1999 Apparent lability of neural tube closure in laboratory animals and humans Am J Med Genet 87 2 143 162 10533029
Diaz E Yang YH Ferreira T Loh KC Okazaki Y Hayashizaki Y 2003 Analysis of gene expression in the developing mouse retina Proc Natl Acad Sci USA 100 9 5491 5496 12702772
Dick RA Kwak MK Sutter TR Kensler TW 2001 Antioxidative function and substrate specificity of NAD(P)H-dependent alkenal/one oxidoreductase. A new role for leukotriene B4 12-hydroxydehydrogenase/15-oxoprostaglandin 13-reductase J Biol Chem 276 44 40803 40810 11524419
Docterman KE Smith SM 2002 Of meis and men: lessons from a microarray study of teratogen action Teratology 66 5 217 223 12397629
Doetzlhofer A Rotheneder H Lagger G Koranda M Kurtev V Brosch G 1999 Histone deacetylase 1 can repress transcription by binding to Sp1 Mol Cell Biol 19 8 5504 5511 10409740
Donohoe ME Zhang X McGinnis L Biggers J Li E Shi Y 1999 Targeted disruption of mouse Yin Yang 1 transcription factor results in peri-implantation lethality Mol Cell Biol 19 10 7237 7244 10490658
Downs KM Davies T 1993 Staging of gastrulating mouse embryos by morphological landmarks in the dissecting microscope Development 118 4 1255 1266 8269852
Dressel U Renkawitz R Baniahmad A 2000 Promoter specific sensitivity to inhibition of histone deacetylases: implications for hormonal gene control, cellular differentiation and cancer Anticancer Res 20 2A 1017 1022 10810390
Dysvik B Jonassen I 2001 J-Express: exploring gene expression data using Java Bioinformatics 17 4 369 370 11301307
Faiella A Wernig M Consalez GG Hostick U Hofmann C Hustert E 2000 A mouse model for valproate teratogenicity: parental effects, homeotic transformations, and altered HOX expression Hum Mol Genet 9 2 227 236 10607833
Fantel AG 1996 Reactive oxygen species in developmental toxicity: review and hypothesis Teratology 53 3 196 217 8761887
Fukuda H Noguchi T Iritani N 1999 Transcriptional regulation of fatty acid synthase gene and ATP citrate-lyase gene by Sp1 and Sp3 in rat hepatocytes(1) FEBS Lett 464 3 113 117 10618488
Ghoshal K Datta J Majumder S Bai S Dong X Parthun M 2002 Inhibitors of histone deacetylase and DNA methyltransferase synergistically activate the methylated metallothionein I promoter by activating the transcription factor MTF-1 and forming an open chromatin structure Mol Cell Biol 22 23 8302 8319 12417732
Giancotti FG Ruoslahti E 1999 Integrin signaling Science 285 5430 1028 1032 10446041
Göttlicher M Minucci S Zhu P Kramer OH Schimpf A Giavara S 2001 Valproic acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells EMBO J 20 24 6969 6978 11742974
Gunther M Frebourg T Laithier M Fossar N Bouziane-Ouartini M Lavialle C 1995 An Sp1 binding site and the minimal promoter contribute to overexpression of the cytokeratin 18 gene in tumorigenic clones relative to that in nontumorigenic clones of a human carcinoma cell line Mol Cell Biol 15 5 2490 2499 7537848
Gurvich N Klein PS 2002 Lithium and valproic acid: parallels and contrasts in diverse signaling contexts Pharmacol Ther 96 1 45 66 12441177
Hancock JF 2003 Ras proteins: different signals from different locations Nat Rev Mol Cell Biol 4 5 373 384 12728271
Holt D Magos L Webb M 1980 The interaction of cadium-induced rat renal metallothionein with bivalent mercury in vitro Chem Biol Interact 32 1–2 125 135 7428106
Hughes RC 2001 Galectins as modulators of cell adhesion Biochimie 83 7 667 676 11522396
Iannaccone PM 2001 Toxicogenomics: “the call of the wild chip Environ Health Perspect 109 1 A8 A11 11277097
Jain AN Tokuyasu TA Snijders AM Segraves R Albertson DG Pinkel D 2002 Fully automatic quantification of microarray image data Genome Res 12 2 325 332 11827952
Jäkel S Mingot JM Schwarzmaier P Hartmann E Görlich D 2002 Importins fulfils a dual function as nuclear import receptors and cytoplasmic chaperones for exposed basic domains EMBO J 21 3 377 386 11823430
Johannessen CU 2000 Mechanisms of action of valproate: a commentatory Neurochem Int 37 2–3 103 110 10812195
Juriloff DM Harris MJ 2000 Mouse models for neural tube closure defects Hum Mol Genet 9 6 993 1000 10767323
Kaji M Mikawa H 1991 Induction of metallothionein in mouse liver by valproic acid Toxicology 69 2 143 149 1949043
Keen CL Peters JM Hurley LS 1989 The effect of valproic acid on 65 Zn distribution in the pregnant rat J Nutr 119 4 607 611 2495343
Kelly DL Rizzino A 2000 DNA microarray analyses of genes regulated during the differentiation of embryonic stem cells Mol Reprod Dev 56 2 113 123 10813842
Klootwijk R Franke B van der Zee CE de Boer RT Wilms W Hol FA 2000 A deletion encompassing Zic3 in bent tail, a mouse model for X-linked neural tube defects Hum Mol Genet 9 11 1615 1622 10861288
Ko MS 2001 Embryogenomics: developmental biology meets genomics Trends Biotechnol 19 12 511 518 11711195
Kurisaki K Kurisaki A Valcourt U Terentiev AA Pardali K Ten Dijke P 2003 Nuclear factor YY1 inhibits transforming growth factor beta- and bone morphogenetic protein-induced cell differentiation Mol Cell Biol 23 13 4494 4510 12808092
Kuzuya T Amioka K Nabeshima T 2002 Valproic acid increases biliary copper excretion in the rat Epilepsy Res 51 3 279 285 12399078
Lagger G O’Carroll D Rembold M Khier H Tischler J Weitzer G 2002 Essential function of histone deacetylase 1 in proliferation control and CDK inhibitor repression EMBO J 21 11 2672 2681 12032080
Lammer EJ Sever LE Oakley GP Jr 1987 Teratogen update: valproic acid Teratology 35 3 465 473 3114906
Lee J Prohaska JR Thiele DJ 2001 Essential role for mammalian copper transporter Ctr1 in copper homeostasis and embryonic development Proc Natl Acad Sci USA 98 12 6842 6847 11391005
Lee JS Galvin KM Shi Y 1993 Evidence for physical interaction between the zinc-finger transcription factors YY1 and Sp1 Proc Natl Acad Sci USA 90 13 6145 6149 8327494
Liu FT Patterson RJ Wang JL 2002 Intracellular functions of galectins Biochim Biophys Acta 1572 2–3 263 273 12223274
Lönnstedt I Speed TP 2002 Replicated microarray data Statist Sinica 12 31 46
Lopez-Rovira T Chalaux E Massague J Rosa JL Ventura F 2002 Direct binding of Smad1 and Smad4 to two distinct motifs mediates bone morphogenetic protein-specific transcriptional activation of Id1 gene J Biol Chem 277 5 3176 3185 11700304
Lu Y Lotan R 1999 Transcriptional regulation by butyrate of mouse galectin-1 gene in embryonal carcinoma cells Biochim Biophys Acta 1444 1 85 91 9931450
Marks PA Miller T Richon VM 2003 Histone deacetylases Curr Opin Pharmacol 3 4 344 351 12901942
McBurney MW 1993 P19 embryonal carcinoma cells Int J Dev Biol 37 1 135 140 8507558
McMahon JA Takada S Zimmerman LB Fan CM Harland RM McMahon AP 1998 Noggin-mediated antagonism of BMP signaling is required for growth and patterning of the neural tube and somite Genes Dev 12 10 1438 1452 9585504
Mirkes PE 2002 2001 Warkany lecture: to die or not to die, the role of apoptosis in normal and abnormal mammalian development Teratology 65 5 228 239 11967922
Miyashita T Yamamoto H Nishimune Y Nozaki M Morita T Matsushiro A 1994 Activation of the mouse cytokeratin A (endo A) gene in teratocarcinoma F9 cells by the histone deacetylase inhibitor Trichostatin A FEBS Lett 353 2 225 229 7523197
Miyazono K Miyazawa K 2002. Id: a target of BMP signaling. Sci STKE 2002(151):PE40.
Moiseeva EP Williams B Goodall AH Samani NJ 2003 Galectin-1 interacts with beta-1 subunit of integrin Biochem Biophys Res Commun 310 3 1010 1016 14550305
Na L Wartenberg M Nau H Hescheler J Sauer H 2003 Anticonvulsant valproic acid inhibits cardiomyocyte differentiation of embryonic stem cells by increasing intracellular levels of reactive oxygen species Birth Defects Res Part A Clin Mol Teratol 67 3 174 180 12797459
Nagai T Aruga J Minowa O Sugimoto T Ohno Y Noda T 2000 Zic2 regulates the kinetics of neurulation Proc Natl Acad Sci USA 97 4 1618 1623 10677508
Nagai T Aruga J Takada S Gunther T Sporle R Schughart K 1997 The expression of the mouse Zic1 , Zic2 , and Zic3 gene suggests an essential role for Zic genes in body pattern formation Dev Biol 182 2 299 313 9070329
Nau H Hauck RS Ehlers K 1991 Valproic acid-induced neural tube defects in mouse and human: aspects of chirality, alternative drug development, pharmacokinetics and possible mechanisms Pharmacol Toxicol 69 5 310 321 1803343
Nau H Löscher W 1986 Pharmacologic evaluation of various metabolites and analogs of valproic acid: teratogenic potencies in mice Fundam Appl Toxicol 6 4 669 676 3086174
Nuwaysir EF Bittner M Trent J Barrett JC Afshari CA 1999 Microarrays and toxicology: the advent of toxicogenomics Mol Carcinog 24 3 153 159 10204799
Oberemm A Kirschbaum F 1992 Valproic acid induced abnormal development of the central nervous system of three species of amphibians: implications for neural tube defects and alternative experimental systems Teratog Carcinog Mutagen 12 6 251 262 1363963
Ogra Y Suzuki K Gong P Otsuka F Koizumi S 2001 Negative regulatory role of Sp1 in metal responsive element-mediated transcriptional activation J Biol Chem 276 19 16534 16539 11279094
Paules R 2003 Phenotypic anchoring: linking cause and effect Environ Health Perspect 111 6 A338 339 12760838
Perillo NL Marcus ME Baum LG 1998 Galectins: versatile modulators of cell adhesion, cell proliferation, and cell death J Mol Med 76 6 402 412 9625297
Phiel CJ Zhang F Huang EY Guenther MG Lazar MA Klein PS 2001 Histone deacetylase is a direct target of valproic acid, a potent anticonvulsant, mood stabilizer, and teratogen J Biol Chem 276 39 36734 36741 11473107
Ramalho-Santos M Yoon S Matsuzaki Y Mulligan RC Melton DA 2002 “Stemness”: transcriptional profiling of embryonic and adult stem cells Science 298 5593 597 600 12228720
Rodriguez-Garcia MI Morgan RO Fernandez MR Bances P Fernandez MP 1999 Mouse annexin V genomic organization includes an endogenous retrovirus Biochem J 337 pt 1 125 131 9854034
Seaton BA Dedman JR 1998 Annexins Biometals 11 4 399 404 10191502
Shapland C Hsuan JJ Totty NF Lawson D 1993 Purification and properties of transgelin: a transformation and shape change sensitive actin-gelling protein J Cell Biol 121 5 1065 1073 8501116
Smith JL Schoenwolf GC 1997 Neurulation: coming to closure Trends Neurosci 20 11 510 517 9364665
Smith L Greenfield A 2003 DNA microarrays and development Hum Mol Genet 12 spec no 1 R1 R8 12668591
Suske G 1999 The Sp-family of transcription factors Gene 238 2 291 300 10570957
Svensson K Mattsson R James TC Wentzel P Pilartz M MacLaughlin J 1998 The paternal allele of the H19 gene is progressively silenced during early mouse development: the acetylation status of histones may be involved in the generation of variegated expression patterns Development 125 1 61 69 9389664
Tanaka TS Jaradat SA Lim MK Kargul GJ Wang X Grahovac MJ 2000 Genome-wide expression profiling of mid-gestation placenta and embryo using a 15,000 mouse developmental cDNA microarray Proc Natl Acad Sci USA 97 16 9127 9132 10922068
Walmod PS Skladchikova G Kawa A Berezin V Bock E 1999 Antiepileptic teratogen valproic acid (VPA) modulates organisation and dynamics of the actin cytoskeleton Cell Motil Cytoskeleton 42 3 241 255 10098937
Wang Y Dooher JE Koedood Zhao M Gilmore TD 1999 Characterization of mouse Trip6: a putative intracellular signaling protein Gene 234 2 403 409 10395914
Wegner C Drews E Nau H 1990 Zinc concentrations in mouse embryo and maternal plasma. Effect of valproic acid and nonteratogenic metabolite Biol Trace Elem Res 25 3 211 217 1698419
Werling U Siehler S Litfin M Nau H Gottlicher M 2001 Induction of differentiation in F9 cells and activation of peroxisome proliferator-activated receptor delta by valproic acid and its teratogenic derivatives Mol Pharmacol 59 5 1269 1276 11306712
Whitsel AI Johnson CB Forehand CJ 2002 An in ovo chicken model to study the systemic and localized teratogenic effects of valproic acid Teratology 66 4 153 163 12353211
Wlodarczyk BC Craig JC Bennett GD Calvin JA Finnell RH 1996 Valproic acid-induced changes in gene expression during neurulation in a mouse model Teratology 54 6 284 297 9098922
Xu W Baribault H Adamson ED 1998 Vinculin knockout results in heart and brain defects during embryonic development Development 125 2 327 337 9486805
Yang WM Inouye C Zeng Y Bearss D Seto E 1996 Transcriptional repression by YY1 is mediated by interaction with a mammalian homolog of the yeast global regulator RPD3 Proc Natl Acad Sci USA 93 23 12845 12850 8917507
Yang YH Dudoit S Luu P Lin DM Peng V Ngai J 2002. Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30(4):e15 [http://nar.oupjournals.org/cgi/content/full/30/4/e15].
Ybot-Gonzalez P Copp AJ 1999 Bending of the neural plate during mouse spinal neurulation is independent of actin microfilaments Dev Dyn 215 3 273 283 10398537
Young MB DiSilvestro MR Sendera TJ Freund J Kriete A Magnuson SR 2003 Analysis of gene expression in carbon tetrachloride-treated rat livers using a novel bioarray technology Pharmacogenomics J 3 1 41 52 12629582
Yuan PX Huang LD Jiang YM Gutkind JS Manji HK Chen G 2001 The mood stabilizer valproic acid activates mitogen-activated protein kinases and promotes neurite growth J Biol Chem 276 34 31674 31683 11418608
Zimmerman AW 1984 Hyperzincemia in anencephaly and spina bifida: a clue to the pathogenesis of neural tube defects? Neurology 34 4 443 450 6366610
|
15345369
|
PMC1277116
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug 3; 112(12):1225-1235
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7034
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7036ehp0112-00123615345370ToxicogenomicsArticlesDiscriminating Different Classes of Toxicants by Transcript Profiling Steiner Guido 12Suter Laura 1Boess Franziska 1Gasser Rodolfo 1de Vera Maria Cristina 1Albertini Silvio 1Ruepp Stefan 11Non-Clinical Drug Safety and2Bioinformatics, F. Hoffmann-La Roche Ltd., Basel, SwitzerlandAddress correspondence to S. Ruepp, F. Hoffmann-La Roche Ltd., PRBN-S (90/5.18), CH-4070 Basel, Switzerland. Telephone: 41 61 688 3315. Fax: 41 61 688 8101. E-mail:
[email protected] data is available online (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf).
We thank M. Haiker, N. Flint, S. Romer, K. Rupp, K. Schad, and C. Zihlmann for their excellent technical support and the General Toxicology group for their support. We are also deeply indebted to C. Broger, M. Neeb, B. Gaisser, and D. Wolf from the Bioinformatics group for their excellent support.
The authors declare they have no competing financial interests.
8 2004 1 7 2004 112 12 1236 1248 17 2 2004 1 7 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Male rats were treated with various model compounds or the appropriate vehicle controls. Most substances were either well-known hepatotoxicants or showed hepatotoxicity during preclinical testing. The aim of the present study was to determine if biological samples from rats treated with various compounds can be classified based on gene expression profiles. In addition to gene expression analysis using microarrays, a complete serum chemistry profile and liver and kidney histopathology were performed. We analyzed hepatic gene expression profiles using a supervised learning method (support vector machines; SVMs) to generate classification rules and combined this with recursive feature elimination to improve classification performance and to identify a compact subset of probe sets with potential use as biomarkers. Two different SVM algorithms were tested, and the models obtained were validated with a compound-based external cross-validation approach. Our predictive models were able to discriminate between hepatotoxic and nonhepatotoxic compounds. Furthermore, they predicted the correct class of hepatotoxicant in most cases. We provide an example showing that a predictive model built on transcript profiles from one rat strain can successfully classify profiles from another rat strain. In addition, we demonstrate that the predictive models identify nonresponders and are able to discriminate between gene changes related to pharmacology and toxicity. This work confirms the hypothesis that compound classification based on gene expression data is feasible.
livermicroarraypredictive toxicologyratsupport vector machinestoxicogenomics
==== Body
Microarray technology is a powerful tool allowing simultaneous investigation of gene expression changes of thousands of genes in response to various stimuli. Large-scale and even whole transcriptome analyses have successfully been applied in various fields including variation in budding yeast (Brem et al. 2002), development of Drosophila melanogaster (Arbeitman et al. 2002), variation in primates (Enard et al. 2002), and human cancer (Ramaswamy et al. 2003). Class identification and prediction of defined end points using gene expression arrays have shown promising results in oncology (Alizadeh et al. 2001; Ramaswamy et al. 2001; Van de Vijver et al. 2002).
The application of gene expression analysis in toxicology has led to the emergence of the discipline of toxicogenomics. We anticipate that toxicogenomics will greatly improve the sensitivity, accuracy, and speed of toxicologic investigations. Toxicogenomics assumes that toxicity is accompanied by changes in gene expression that are either causally linked or represent a response to toxicity. Indeed, researchers have been able to link toxicity with expression changes of single genes or whole groups of genes (Hamadeh et al. 2002c; Ruepp et al. 2002; Suter et al. 2003).
A transcriptome-wide overview of altered expression patterns can assist the mechanistic understanding of underlying changes induced by chemicals (Hamadeh et al. 2002b). This requires a comprehensive knowledge of the biological system under investigation, and only known genes are considered for analysis. This functional approach is also promising for the generation and testing of toxicity hypotheses (Donald et al. 2002; Zhang et al. 2002) or the identification of perturbed pathways (Wang et al. 1999; Zimmermann et al. 2003). Furthermore, identification of toxic mechanisms is valuable for risk assessment because it allows extrapolation of the hazard in humans.
Predictive toxicology is based on the hypothesis that similar treatments leading to the same end point will share comparable changes in gene expression. Several investigators have used gene expression profiling for the classification of toxicants in rodents (Bulera et al. 2001; Hamadeh et al. 2002a; Thomas et al. 2001; Waring et al. 2001b). These studies varied in design and number of compounds investigated, but all indicated the potential of toxicogenomics in predictive risk assessment.
A major challenge in predicting toxicologic end points based on transcriptional data lies in discriminating changes due to interanimal variation or experimental background noise from treatment-related changes. Compounds may directly affect expression of certain well-characterized, compound-specific genes. These compound-specific genes are not suited for discrimination between different classes of compounds. Drugs, in contrast to other toxic substances, have pharmacologic as well as toxicologic effects that might affect gene expression. These two effects can, but need not, be related. Despite these confounding factors, gene expression analysis after treatment with various compounds that result in the same toxicologic end point should enable identification of a toxic fingerprint.
Various methods are used to analyze large-scale gene expression data. Unsupervised methods widely reported in the literature include agglomerative clustering (Eisen et al. 1998), divisive clustering (Alon et al. 1999), K-means clustering (Everitt 1974), self-organizing maps (Kohonen 1995), and principal component analysis (Joliffe 1986). Support vector machines (SVMs), on the other hand, belong to the class of supervised learning algorithms. Originally introduced by Vapnik and co-workers (Boser et al. 1992; Vapnik 1998), they perform well in different areas of biological analysis (Schölkopf and Smola 2002). Given a set of training examples, SVMs are able to recognize informative patterns in input data and make generalizations on previously unseen samples. Like other supervised methods, SVMs require prior knowledge of the classification problem, which has to be provided in the form of labeled training data. Used in a growing number of applications, SVMs are particularly well suited for the analysis of microarray expression data because of their ability to handle situations where the number of features (genes) is very large compared with the number of training patterns (microarray replicates). Several studies have shown that SVMs typically tend to outperform other classification techniques in this area (Brown et al. 2000; Furey et al. 2000; Yeang et al. 2001). In addition, the method proved effective in discovering informative features such as genes that are especially relevant for the classification and therefore might be critically important for the biological processes under investigation. A significant reduction of the gene number used for classification is also crucial if reliable classifiers are to be obtained from microarray data. A proposed method to discriminate the most relevant gene changes from background biological and experimental variation is gene shaving (Hastie et al. 2000). However, we chose another method, recursive feature elimination (RFE) (Guyon et al. 2002), to create sets of informative genes.
The liver is a primary site for drug metabolism and is frequently involved in adverse drug reactions. Thus, hepatotoxic compounds were chosen for our toxicogenomic studies. In this study 28 hepatotoxic compounds and 3 nonhepatotoxic compounds were investigated. Time-matched controls dosed with the corresponding vehicles were used to allow discrimination between temporal and compound-induced changes. This is essential for large-scale transcriptome analysis, as extensive circadian gene expression patterns have recently been reported in the liver and heart of the mouse (Kita et al. 2002; Panda et al. 2002; Storch et al. 2002).
Depending on the substance and category of toxicity, different time points were chosen for classification, as manifestation of toxicity was observed earlier for certain compounds than for others. Clinical chemistry, hematology, and histopathology were used to assess toxicity of each individual animal.
Models for discrimination of toxic and nontoxic substances as well as models specifying the category of toxicity were built using data from a variety of toxicity studies. The hypothesis that unknown blinded compounds could accurately be classified based solely on gene expression profiles was subsequently tested. In the majority of cases, SVMs were able to predict toxicity as well as the mode of toxicity. The potential for obtaining the same level of predictivity with only a small number of carefully selected genes was investigated. This subset of genes includes potential biomarkers for hepatotoxicity.
Materials and Methods
Animal Treatment
Permission for animal studies was obtained from the local regulatory agencies, and all study protocols were in compliance with animal welfare guidelines. Male HanBrl:Wistar rats approximately 12 weeks of age (300 g ± 20%) were obtained from BRL (Füllinsdorf, Switzerland). The animals were housed individually in Macrolone (Tecniplast GmbH, Hohenpeissenberg, Germany) cages with wood shavings as bedding at 20°C and 50% relative humidity in a 12-hr light/dark rhythm with free access to water and Kliba 3433 rodent pellets (Provimi Kliba AG, Kaiseraugst, Switzerland). For the WY14643 study, male Sprague-Dawley Crl:CD(SD)IGS.BR rats approximately 6 weeks of age (200 g ± 20%) were obtained from Charles River Ltd. (Margate, U.K.)
Animals were dosed with test compounds or the corresponding vehicles orally or by ip, iv, or sc injections and sacrificed at specified times by CO2 inhalation (Table 1). Immediately preceding sacrifice, terminal blood samples for clinical chemistry investigations were collected from the retroorbital sinus. Liver samples from the left medial lobe were removed immediately and placed into RNALater (Ambion, Austin, TX, USA) for RNA extraction and gene expression analysis (Table 1). The exposure period for each compound was based on reports in the literature and results from pilot studies using histopathology and clinical chemistry anchoring to assess toxicity. Thus, for unknown compounds best results are expected if several time points (e.g., 6 hr, 1 day, and 1 week) are tested.
Clinical Chemistry
The following determinations were made from the serum: blood urea nitrogen (BUN), alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-glutamyltransferase (GGT), lactate dehydrogenase (LDH), sorbitol dehydrogenase (SDH), alkaline phosphatase (ALP), 5′-nucleotidase (5′-NT), glutamate dehydrogenase (GLD), urea, glucose, creatinine, bilirubin, total protein, albumin, globulins, total cholesterol, triglycerides, phospholipids, fatty acids, bile acids, sodium, potassium, chloride, calcium, and phosphorus.
Histology
Representative liver samples were fixed in 10% neutral-buffered formalin. One additional liver sample from the cranial half of the left lateral lobe was placed in Carnoy fixative for glycogen staining. All samples were processed using routine procedures and embedded in Paraplast (Sherwood Medical Ltd., Tullamore, Ireland). Tissue sections approximately 2–3 μ were cut and stained with hematoxylin and eosin or periodic acid-Schiff for glycogen. Fat Red 7B stain (Fluka, Buchs, Switzerland) was performed on frozen formalin-fixed sections to visualize lipid deposits.
Sample Preparation and Hybridization
RNA isolation, processing, and hybridization were essentially carried out as recommended by Affymetrix (Affymetrix, Santa Clara, CA, USA) with minor modifications [Supplemental data (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)].
Data Acquisition and Preprocessing
Primary data were obtained by laser scanning (Hewlett Packard, Palo Alto, CA, USA) and collated using the Affymetrix Microarray Suite Version 5.0 software (Affymetrix). Before performing any downstream analysis, data were preprocessed in a standardized way. First, the gene expression values of every single microarray experiment were rescaled to a mean value of zero and a standard deviation of 1 to establish comparability across all samples. Because single outlying expression values occur rather frequently and are likely to affect any analysis method, a modified version of the Nalimov outlier test (Kaiser and Gottschalk 1972) was applied to identify these potential artifacts. Expression values reported as outliers were replaced by the respective mean values. The test was performed separately for each classification group (i.e., class of toxicity). In contrast to the published method, our modified version does only one round of outlier removal rather than multiple iterations. A normal distribution model is calculated for the expression levels to be tested, and outliers are removed at a 99% confidence level. As a final preprocessing step, the expression values were rescaled so that the expression of each single gene across multiple arrays has a mean value of zero and a standard deviation of 1. This transformation increases the numerical stability of the SVM algorithm and facilitates the assessment of the relative importance (weight) of single genes within a reduced feature set. Again, this was performed separately for each classification group.
Support Vector Machines
A detailed introduction into theory and application of SVMs is beyond of the scope of this article. We refer the interested reader to the available literature (Cristianini and Shawe-Taylor 2000; Schölkopf and Smola 2002) and the Supplemental Data(http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)].
All SVM classifications were based on the free available software package LIBSVM, 2.36, which was downloaded from the World Wide Web (Chang and Lin 2001). The source code was modified according to our needs and compiled to run on the operating system IRIX, version 6.5, (Silicon Graphics, Inc., Mountain View, CA, USA). Extensions such as parameter optimization, feature selection, enhanced cross-validation (CV) options, the one-versus-all training scheme, and report generation were implemented in a C library on top of LIBSVM.
Choice of Parameters
A linear kernel k(xi,xj) = 〈xi,xj〉 was chosen for the SVM, as higher order correlation functions could easily lead to overinterpretation of the data, given the unfavorable ratio of features and replicates. LIBSVM offers two different SVM formulations for classification: C-SVM and υ-SVM. These formulations use different parameters for adjusting the accuracy versus margin tradeoff but should produce comparable solutions. We tried both formulations and tuned their respective parameters for optimal CV performance.
To handle the multiple class situation, we applied the one-versus-all training paradigm. Using this approach, a set of binary SVMs is created, each of which separates the samples of one class (positive examples) from all remaining training data (negative examples). Because the number of negative examples usually outweighs the number of positive examples in this scheme, there is always a risk of losing sensitivity for the smaller class. However, practice showed that no additional class bias had to be introduced after appropriate values for the C or υ parameter had been determined for each single SVM. Optimization of these crucial parameters was done in an iterative manner. We typically started with either a C value set to 1.0 or a υ value of 0.5 and performed a complete gene selection run. Optimization of SVMs using different feature numbers suggested improvements in the initial settings as well as a sensible range for the parameters. Feature selection was then repeated with the new settings, and individual SVMs were again tuned to determine good parameters for different gene numbers. This process led to a noticeably improved classification performance.
Classifier Validation
The predictive power of individual SVMs was primarily rated by their CV performance. However, as our main interest was to estimate the generalization properties of classifiers with respect to new compounds, we did not select the frequently applied leave-one-out or randomization-based schemes. Instead, all microarrays that resulted from the treatment with a certain compound (regardless of dose and time point) were left out as a whole group in one CV cycle. Whenever CV is combined with feature selection, special care must be taken to avoid any bias leading to over-optimistic performance estimates. Therefore, we applied external CV exclusively where feature selection was done separately for each group of left-out examples, thereby avoiding the use of information from the excluded examples in the feature selection process. Although the final classifier is built on all available training examples, the described method was used to determine the optimal number of genes as well as the parameter settings. As a consequence, we expect that the resulting classifiers are less influenced by the given selection of compounds and that CV provides a more realistic estimate for the generalization on new compounds.
Quantitative measures for training and CV performance were sensitivity and specificity values as well as the Matthews correlation coefficient (MCC),
where TP = number of true positives, TN = number of true negatives, FP = number of false positives, and FN = percentage of false negatives. The MCC is commonly used as a measure of the predictive power of a system that gives categoric variables as output (Matthews 1975). It was our main performance indicator.
When several SVMs showed exactly the same CV result, performance on the training set was also taken into account. If this still yielded equal results, we finally selected the simplest model (i.e., smallest number of support vectors and smallest number of features).
Gene Selection
Although SVMs can easily tolerate the high-dimensional gene space typical of microarray studies, most of the features are usually irrelevant for the classification task and only introduce noise. To obtain a meaningful decision function that generalizes well, the number of variables must be reduced as much as possible.
Various methods exist for selecting discriminating features for classification purposes; most deal with variables individually. RFE overcomes some deficiencies of this univariate approach (Guyon et al. 2002). Basically, RFE is a greedy backward elimination method. Starting with all features (except for Affymetrix control genes), a ranking is produced based on the relative importance of a particular feature in the SVM decision function. A certain fraction of the least important variables is then removed, and the process is repeated iteratively until the feature list is empty. The precise order of features might change from iteration to iteration. Because of the multivariate properties of the SVM algorithm, each feature ranking takes into account (at least to some extent) correlations between single variables. Evaluating the classification performance at each step makes it possible not only to identify a suitable subset of descriptors but also to determine how many of them are actually needed for a reliable classification. Redundant features also tend to be eliminated during RFE, typically resulting in very compact feature sets (Guyon et al. 2002).
We implemented RFE on top of the libsvm software. In the beginning, a user-definable fraction of the least important genes is removed in each iteration. After reaching a certain threshold number, only one more gene is eliminated in each step. We experimented with several values for the fraction and lower threshold values to further improve the classification performance of our classifiers.
Presentation of Support Vector Machine Results
A binary SVM discriminating between two classes is trained by presenting the training samples of one class (A) as positive examples while samples belonging to the other class (B) act as negative examples. An SVM prediction (i.e., the value of the decision function when a new data example is tested) is simply a real number called the discriminant. If the discriminant is positive, the example is considered belonging to class A. Similarly, a negative number would indicate membership in class B. The absolute value of the discriminant can be regarded as a measure of confidence for the classification.
If there are more than two distinct classes, several binary classifiers must be combined to obtain a prediction for a new sample. When applying the one-versus-all scheme (see above), n classifiers have to be created for n classes. A new data example is then tested with all these SVMs and therefore the result consists of n real values from which the most probable class assignment must be inferred. Classification results can be presented as plots of discriminant values that were obtained from a set of SVMs (Figure 1). A unique assignment is possible if only one SVM produces a positive output for a certain sample. If a treatment group is not classified uniformly, we assign the corresponding compound to a category by majority vote, with 60% as the cutoff.
Sometimes two- or three-dimensional scatter plots were produced for visualizing the class separation of one model (Figure 2). These diagrams map all training and test examples into one coordinate system and often reveal some (expected or unexpected) internal structure of the data such as subclusters or single outliers. The dimensionality reduction is achieved by plotting linear combinations of features against each other. Coefficients are obtained from the SVM decision function [Supplemental data (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)].
Results
Histopathology and Clinical Chemistry—Profiles Used for Training Support Vector Machine Models
We used SVMs as a supervised learning method to generate classification rules. It was of crucial importance to provide training labels on the basis of solid evidence. Therefore, a complete serum chemistry profile and liver histopathology were performed on virtually all rats treated with various model compounds or the appropriate vehicle controls. This information in conjunction with published data provided the basis to allocate gene expression profiles to a specific training class (Table 1).
Gene Expression Analysis
Gene expression profiles from individual rat livers treated with vehicle or test compounds were analyzed using the Affymetrix U34A GeneChip. All microarrays included in the analysis fulfilled our established quality parameters [Supplemental data (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)]. All treatments caused transcriptional changes with respect to their corresponding time-matched controls. In all studies, > 150 genes were expressed above background and showed at least a 2-fold modulation with a p-value < 0.05 (two-tailed, unpaired t test).
Assessment of Time Effects in Vehicle Control–Treated Rat Livers (Early versus Late)
Supervised analysis of gene expression data suffers if parameters other than the investigated effects correlate with the classes for which one tries to identify typical finger-prints. The studies evaluated in this article differed in vehicle, application route, and time point (Table 1). We assumed that time-dependent effects could be the confounding factor with the most noticeable impact on the results. Thus, we analyzed gene expression patterns from vehicle-treated animals (i.e., controls) at various time points. A classification attempt was made using the same time points used for the toxicity classifications in this article (early class is 6 hr, late class is 24 hr up to several days). We obtained a prediction accuracy of 70% and an MCC of 0.41, whereas random shuffling of the analyzed microarrays gave MCC values close to zero, indicating that the observed variations can be attributed to time effects. Using this approach, 14 genes were selected as the best set of discriminative features. These results confirmed that there are indeed some observable time-related effects [Supplemental data, Table 1 (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)]. However, because time points of the control microarrays typically vary within one class of toxicity, we expected the SVM to identify the time-dependent genes as not relevant for the toxicity predictions. Actually, only one gene of this subset appeared in one toxicity classifier (rc_AA799616_at) but with a low weight.
Effect of the Rat Strain
Because different rat strains are widely used in toxicology, we investigated the effect of strain differences of Wistar and Sprague-Dawley rats for classification based on transcript profiles. Our database consisting of Wistar rat data was used to generate an SVM. Subsequently, gene expression profiles from vehicle control and WY14643-treated Sprague-Dawley livers were used to assess whether the model would correctly classify individual animals from another rat strain. All five controls were clearly identified as controls (Figure 1). Their transcript profiles yielded negative discriminants for all SVMs except the control SVM, where positive values marked those profiles as controls. Animals treated with 250 mg/kg WY14643 were unanimously assigned to the peroxisomal proliferator class. Here, discriminant values were positive only for the peroxisomal SVM and negative for all other categories. As those results indicated that gene expression profiles from Sprague-Dawley and Wistar rats are comparable, transcript profiles from WY14643 were included in further models.
Generation of Toxic/Nontoxic and Multitoxicity Models
We generated a binary classifier for the discrimination of vehicle controls and animals treated with a toxic compound. In addition, to predict the mode of action, multitoxicity models were also created. For both the binary and the multiclass case, we used the same data set for training the SVMs, applying either the C-formulation or the υ-formulation of the algorithm. To extract smaller sets of truly discriminative genes, we integrated RFE in the process of model building. This enabled us to study performance parameters such as sensitivity and specificity under CV in relation to the number of genes used.
All the models were evaluated with an external CV scheme that omits a whole treatment group (typically five animals) per cycle. Therefore, a complete RFE run had to be carried out for each of the 60 groups (see “Materials and Methods”). Furthermore, all SVMs were validated with an independent test set that contained different doses and time points of the same substances used for training as well as some new compounds.
The results obtained were similar for C-SVMs and υ-SVMs, although the number of used genes at the point of optimal performance seemed to be smaller for the υ-SVM. However, C-SVM most often outperformed υ-SVM in terms of classification accuracy. Of all 26 toxic substances, C-SVM could not detect 4, whereas υ-SVM missed 6 compounds. A clear majority of the control groups were correctly classified under CV. In this respect there was no pronounced difference between the two formulations.
Toxic/Nontoxic Model
Results for the binary toxic/nontoxic classification are summarized in Table 2. The test set of 63 vehicle-control groups demonstrates how well those models generalize on previously unseen data [details in Supplemental data (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)]. Not every single microarray, but all groups were correctly identified as controls using the described voting procedure. Almost 90% of the toxic test groups were correctly classified as toxic using C-SVM. The model did not produce any false-positive predictions. However, there were some false-negatives, as not all toxic treatments could be recognized as toxic.
Multiclass Model
As the next step we aimed at predicting the mode of toxicity. For this purpose, a control class and three categories of toxicity were initially defined: cholestasis, steatosis, and direct acting. Subsequently, we added peroxisome proliferator-activated receptor α (PPAR-α) agonists as a separate class without any loss in prediction accuracy (Table 3). We refer to this as the 4-modes-of-toxicity (4MOT) model. This is an imperfect simplification of the classification task, as some of the compounds show more than one form of hepatotoxicity, depending on dose and time. Therefore, time points at which a specific toxicity was most apparent were selected for the analysis.
We generated five different SVMs following the one-versus-all approach, that is, each of the models was trained to discriminate between a certain class of toxicity and the set union of all other expression profiles. In a first step, the individual classifiers were built and optimized separately using the same CV procedure described before. Subsequently, a class assignment for each single microarray in the training or test set was done by combining the output of the five models (Tables 4 and 5). In most cases the prediction was unanimous, that is, just one SVM delivered a positive discriminant and the others returned negative values (e.g., Figure 3A–C). In cases where a profile obtained more than one positive discriminant value or only negative numbers, the biggest value determined the classification (e.g., Figure 4). The optimal gene number for classification depends on the category of toxicity. For example, peroxisomal proliferation/PPAR agonists could be recognized with one single marker gene. Nevertheless, the final classifier used four features because of our strategy of simplifying the model by also minimizing the number of support vectors (see “Materials and Methods”). The four top probe sets represent only two distinct genes, acetyl-Coenzyme A acyltransferase 1 (peroxisomal 3-oxoacyl-Coenzyme A thiolase) and cytochrome P450 4A1. Both genes are well-known PPAR-α–responsive genes, and the corresponding upregulation has been described extensively in the literature (Hansmannel et al. 2003; Lee et al. 2003).
The model for the control group required the most features (122). Performance again was rather similar for υ-SVM and C-SVM. In the case of υ-SVM, 274 distinct features were used altogether for discriminating among the five classes of toxicity. However, a reduction to 86 features did not lead to a significant loss in predictivity, indicating that this set could be used for an assay in a 96-well format (data not shown).
Categories of toxicity differ not only in optimal feature number but also in prediction accuracy. Under CV as well as in the test set, all toxicant categories are recognized with a very high specificity, whereas the controls are identified with a high sensitivity. This means that our model produces virtually no false-positive outcomes but at the cost of some false-negative results. All treatment groups within the direct-acting category are either correctly classified or, in the case of aflatoxin, at least recognized as treated with a toxic substance (Table 4). Phalloidin is another example that was identified as toxic, but profiles were classified into two toxicity categories—cholestatic and direct acting. Amiodarone, glibenclamide, and chlorpromazine 1 were not recognized as toxic. Classification of our test set again confirmed the good performance of the model. The classification of 332 test control microarrays with an error rate of 0.6% is remarkable. Using the criteria described above, the success rate in classifying the corresponding 63 control groups is 100% (as seen before in the binary classification).
Identification of Nonresponding Animals
Galactosamine treatment of rats usually leads to hepatitis associated with necrosis and inflammation. Animals were treated once with 400 mg/kg galactosamine or vehicle only and sacrificed after 24 hr. In four of five galactosamine-treated animals, there was clear evidence of toxicity assessed by hematology, clinical chemistry, and histopathology; one animal was a nonresponder. Gene expression profiles of individual animals were tested using the 4MOT model described previously. Classification results are in perfect agreement with the assessment using conventional end points. However, gene expression profiling seems to be more sensitive than clinical chemistry and histopathology, as the data point corresponding to the nonresponding rat is clearly shifted toward the direct-acting group (Figure 2).
Pharmacologic Effects Are Differentiated from Toxicologic Effects
Pharmacologically active substances can alter gene expression, but a predictive model for hepatotoxicity should not confuse a substance with a desired pharmacologic effect with an unwanted toxic outcome. Three nonhepatotoxic but pharmacoactive substances were tested with the SVM models. Although 100 mg/kg gentamicin (sc) led to nephrotoxicity at 24 hr, no hepatotoxicity was associated with it, nor was hepatotoxicity detected with deprenyl or lazabemide. All three nonhepatotoxic substances were correctly classified as nontoxic using both, the toxic/nontoxic as well as the 4MOT model (Figure 3). These results show that our toxicity classifiers can distinguish well between pharmacologic effects without toxicity and toxicologically relevant transcriptional changes.
Classification of Hepatotoxic Compounds with Mechanisms of Toxicity not Represented in Our Database—Lipopolysaccharide, Phenobarbital, and Indomethacin
Compounds with mechanisms of toxicity (MOTs) not represented in our training set were used to investigate how they would be classified by our models. The toxic/nontoxic model had the easier task, as dissimilarity with control profiles would already indicate some toxicity-related abnormality in gene expression. The 4MOT model had to classify an unrepresented profile to one of the five available classes.
Lipopolysaccharide (LPS) (4 mg/kg iv) was investigated 6 and 24 hr after dosing and identified as toxic by the toxic/nontoxic model. The 4MOT model classified four animals as steatotic and one as cholestatic after 6 hr (Figure 4A). After 24 hr four animals were also classified as steatotic and one as direct acting. No sample was misclassified as a control. Phenobarbital (80 mg/kg ip) was also investigated at 6 and 24 hr. At 24 hr all animals fit into the steatotic category (Figure 4B). At 6 hr, four profiles were most similar to the steatotic group and one to the cholestatic group. In this case most discriminant values were very low, indicating differences with respect to the existing classes. Another interesting example was indomethacin, which was administered either as a single high dose (20 mg/kg po; sample collection at 6 or 24 hr) or as a repeated low dose (5 mg/kg po; daily dosing during 1 week). In the liver, minimal to slight hepatocellular hypertrophy and decreased glycogen deposition were observed in animals treated with 20 mg/kg at 24 hr and in animals treated for 1 week with 5 mg/kg/day. The repeated dosing also caused tubular dilation in the kidney and erosive and/or ulcerative inflammations in the gastrointestinal tract. At 6 hr the substance was classified as predominately cholestatic and at 24 hr clearly as steatotic. After 7 days of dosing, three animals were classified as steatotic and two as cholestatic. These profiles had positive discriminants for three toxicity categories (cholestatic, steatotic, direct acting). This indicates that indomethacin is different from our predefined toxicity categories and displays mixed toxicity (Figure 4C). Most important, a very clear dissimilarity from the control group indicated that the indomethacin-treated animals had been exposed to a toxic compound, although the mode of toxicity could not be unequivocally defined.
Discussion
Gene Expression Profiling
The present work aimed to provide evidence that transcript profiles can be used to distinguish compound-treated rat livers from controls and to discriminate between different MOTs. Rats were treated with a variety of vehicles, and hepatotoxic or non-hepatotoxic but pharmacologically active compounds. We focused on hepatotoxicity, as the liver is a main target for toxic reactions. Various questions were addressed in the context of predictive toxicity modeling, including sanimal variability, rat strain differences, effect of time, and discrimination of pharmacologically from toxicologically induced gene changes.
Several authors have described the use of gene expression profiling to classify toxicants in rodent liver and thereby demonstrated the potential of toxicogenomics in predictive risk assessment (Bulera et al. 2001; Hamadeh et al. 2002a; Thomas et al. 2001; Waring et al. 2001b). We used a larger number of compounds and selected a different bioinformatics approach to analyze the data. New in this study is the modeling of different categories of toxicity in conjunction with numeric measures for the classification confidence. Our results demonstrate that for different compounds with similar MOT, the likely toxicologic end point can be inferred from gene expression profiles using a database of model compounds as a training set. Moreover, we found good correlation of gene expression changes with histopathologic findings. These results are consistent with those of a previously reported study where methapyrilene toxicity correlated with the severity of pathologic changes (Hamadeh et al. 2002c).
Feature Selection
SVMs can handle very high-dimensional feature spaces, so there is no pressing need to filter out a small number of genes in a first step. In contrast to many published microarray studies, we did not apply strict cutoffs like 2-fold changes, p-value thresholds, or similar criteria.. These approaches could easily spoil one of the main advantages of a multivariate classification method such as SVMs, as prefiltering of features by common univariate methods (such as the t test) might remove genes that do not reach significance when tested individually but provide useful information when taken together with other, correlated variables. In contrast, RFE allowed us to combine feature selection and model building in a consistent framework, making use of the mutual information between genes (Guyon et al. 2002). We leave it to the method to eliminate noisy, irrelevant variables in the process of forming smaller and smaller subsets of genes with discriminatory power. The approach also helped to avoid the introduction of a feature selection bias, which occurs if information from all experiments is used to reduce the number of genes before any CV is done. However, it is important to remember that the gene lists we obtained are in no way a complete picture of the cellular response but a redundancy-reduced selection of markers that together allow a maximum predictivity.
The relationship between gene number and classification performance was studied using RFE, and subsequently the optimal iteration was chosen. Our results indicate that accurate prediction of toxicity (including the category of toxicity) can be achieved using a small set consisting of a few up to some dozens of features (Table 3). In the case of the 4MOT model, the feature number can be reduced from 274 to 86 without major performance impairment. The observation that more genes do not necessarily translate into higher predictive accuracy is consistent with previous findings (Ramaswamy et al. 2003; Thomas et al. 2001), indicating that it is not necessary to measure the whole transcriptome or thousands of genes to predict toxicity. Once initial experiments have led to an optimized set of relevant informative features, a potentially faster and cheaper assay could be developed providing essentially the same classification performance. Interestingly, using only the selected features for hierarchical clustering also resulted in a toxicologically meaningful result, whereas unsupervised clustering with all genes often failed at classifying the animals according to the criteria of interest (data not shown). However, it is worth mentioning that none of the genes in the final set is guaranteed to act as a good toxicity marker on its own because we do not rank features according to their suitability as single markers (univariate approach) but rather optimize whole subsets of features (multivariate approach). In this setting it is possible that a gene that does not appear differentially expressed in two groups can still contribute useful information by combination with other genes. Therefore, it is often the signature taken as a whole that provides the decisive discriminatory power. Marker gene sets identified with the described method are especially prone to show this effect because of the multivariate nature of SVMs and the tendency of the RFE algorithm to eliminate redundant features from the set (Guyon et al. 2002).
As gene expression analysis can also be applied in vitro (Burczynski et al. 2000; Waring et al. 2001a), the question arises whether the list of features obtained could be used in a cell-based assay. This seems questionable, as significant differences in gene expression in vitro compared with in vivo were reported (Boess et al. 2003). Therefore, we expect that results concerning discriminative features and their weights cannot be directly transferred to in vitro classification systems. In addition, the evaluation of the compound effects in vivo is especially important when multiple cell types and possibly multiple organs are involved in the toxicologic response.
Confounding Effects
A crucial issue when using supervised classification methods is that there must be solid evidence for the initial assignment of gene expression profiles to each category. Therefore, we included only microarrays from animals where independent evidence justified allocation to a specific class. In most cases, histopathologic anchoring was used, but clinical chemistry and occasionally additional biochemical assays (triglyceride assays, data not shown) were also considered. Anchoring to conventional end points was the reason for the heterogeneity of time points used in the training procedure. This kind of heterogeneity might act as a confounding factor, introducing signatures not related to the toxicity classification problem itself. Special care must be taken to ensure that these confounding factors do not exhibit decisive influence on the model. The potentially confounding effect of time was addressed first, as several authors have highlighted extensive circadian gene expression changes (Kita et al. 2002; Panda et al. 2002; Storch et al. 2002). For this purpose, the same time points (6 hr, 24 hr, and several days) used within our toxicity models were used to train a two-class SVM model for classification of early or late time points. A classifier based on 14 genes was obtained, but predictivity was far from perfect and resulted in a relatively low MCC of 0.41. (Test MCC values for the toxicity classifiers were all > 0.80.) Although these results confirm some time dependency in our experiments, we have no reason to assume that this strongly affects our toxicity models, as we always combined control profiles from all time points in the same group for training. Together with the fact that none of the genes from the time classifiers appeared at a prominent position (with significant weight) in the toxicity models, these results suggest that there is no distinct time bias. In fact, classification of vehicle controls from the test set (originating from independent studies and including various time points) was correct in more than 99% of the cases, which confirms the absence of time bias for the control component of the classifier.
Wistar, Sprague-Dawley, and Fischer rats are all frequently used in risk assessment. There is ample evidence that those strains vary in their susceptibilities to various toxicants or mutagens (Asamoto et al. 1989; Kulkarni et al. 1996). Therefore, we investigated whether a model built with Wistar rat expression profiles would be predictive for treatment effects in Sprague-Dawley rats. PPAR agonists were chosen for this comparison for pragmatic reasons. At the time we studied proprietary PPAR agonists, we were also involved in the Consortium for Metabonomic Toxicology (COMET), where liver tissue collection of WY14643-treated Sprague-Dawley rats could be included. [COMET has been formed by Imperial College (London) and six major pharmaceutical companies. The objective is to apply metabonomics to the toxicologic assessment of compounds (Lindon et al. 2003).] Treated rats as well as controls fit perfectly into the anticipated classes. The classification was successful despite the additional confounding factor introduced by the fact that the Sprague-Dawley rats were approximately 6 weeks younger than the Wistar rats. This successful class prediction was the rationale for including those expression profiles in our predictive models. As the results suggest, the discriminative transcriptional changes are largely conserved across strains, although the doses required to produce comparable toxicity may vary.
Another confounding factor for the classification task is that pharmaceuticals not only show a toxic effect on gene regulation but might also influence gene expression according to their pharmacologic action. A crucial test for the classification of toxicants based on gene expression profiles is certainly the ability to separate pharmacologic from true toxic effects. Our models succeeded at classifying three pharmacologically active, nonhepatotoxic compounds. In the case of gentamicin, not even the observed nephrotoxicity led to a false prediction of hepatotoxicity. The classification of these three compounds as nonhepato-toxic was not due to a general lack of effects on hepatic gene expression; more than 100 genes were differentially expressed for these compounds, as assessed by fold change together with t test (at least 2-fold change and p-value < 0.05).
Mixed Toxicities
All transcript profiles were assigned to a specific category, implying that they fit exactly into one class. However, in reality, substances often cause mixed toxicities. We aimed to allocate substances to the best-fit-ting class, knowing the limitations due to the potential overlap of effects. Our results indicate that characteristic gene expression changes are indeed associated with distinct classes of toxicants. However, as compounds cannot be put into exclusive bins in a strict sense, some substances (aflatoxin, indomethacin, and phalloidin) were predicted to be associated with multiple toxicities. Aflatoxin, for example, needs metabolic activation to exert its toxic effect. It causes generation of reactive oxygen species, lipid peroxidation, glutathione depletion, and necrosis and therefore has a direct effect on cells (Liu et al. 1999). On the other hand, it is a well-known carcinogen (Smela et al. 2001) and is reported to induce both cholestasis (Unger et al. 1977) and steatosis (Amaya-Farfan 1999). Based on classical end points, we decided to allocate aflatoxin to the direct-acting group. The SVM classification of gene expression profiles, however, indicated a greater similarity to cholestatic than to direct-acting compounds. One possible way to address this problem might be to generate several one-versus-control categories and include the aflatoxin samples in both the direct-acting and the cholestatic classes. Another option would be to exclude all compounds from training that do not unambiguously fit into one single category. Reported effects of indomethacin in rats are immediately direct, like adenosine triphosphate depletion in hepatocytes (Masubuchi et al. 1998) and a marked decrease in the hepatic monooxygenase system (Fracasso et al. 1990). Gene expression profiles of rats dosed with indomethacin were classified as cholestatic and steatotic but also matched the direct-acting group. Clinical chemistry supported this mixed toxicity prediction to some extent, as ALP, GGT, AST, and LDH were increased. Histopathology revealed hypertrophy and minimal to slight necrosis, but changes were considered to be adaptive rather than reflecting an adverse effect. In patients, however, cases of cholestasis and steatosis have been reported (Farrel 1994). It remains to be confirmed whether the genomics approach is more sensitive than histopathology in detecting liabilities.
Results classifying galactosamine-treated rats using the multitoxicity model support this hypothesis. Galactosamine treatment leads to hepatitis associated with necrosis and inflammation, but a high degree of interanimal variation is well known (Vomel and Platt 1986). In our study four of five expression profiles were identified as toxic while the fifth was classified as control. This classification as nonresponder was in agreement with absence of findings using conventional end points. However, a three-dimensional plot of the SVM results revealed a shift of the expression profile of the nonresponder toward the direct-acting group (Figure 2 ), suggesting increased sensitivity of the toxicogenomics approach.
Model Assessment
We used a compound-based external CV scheme (see “Materials and Methods”) to obtain more realistic estimates for the classification performance and to select a model from which we can expect a good generalization power. It has to be kept in mind that the compound database is still limited in size, and we do not know whether our set of substances is a representative sampling of the complete toxicology space. Therefore, we cannot completely rule out some sampling bias, which would render our performance estimates too optimistic. Conversely, our CV procedure intrinsically tends to deliver a rather conservative assessment of the performance, as at least some of the compounds provide vital information that is lost as soon as a whole treatment group is withheld from training. For example, glibenclamide (dosed at 25 mg/kg) was not recognized as a cholestatic compound under CV; the final SVMs correctly classified two of five animals in the test set, despite the fact that these had been treated with a lower dose (2.5 mg/kg) and histopathology was only evident in animals that received the high dose.
Because of the partial overlap of compounds in the training and test set, one would expect a smaller fraction of misclassifications under test conditions than with the more rigorous CV method. This was indeed observed in most cases (Tables 2 and 3) and emphasizes the extent to which interpretation of results depends on the details of the applied evaluation method. Although in this case the number of CV errors can provide information about the generalization ability of a model, the test performance should be regarded as a measure for its consistency with respect to a certain selection of compounds. For our application we clearly wanted to optimize the former; therefore, we used the described CV scheme to select the best SVMs. When we tentatively switched to a more standard, leave-one-out procedure, not a single CV error occurred. However, the test accuracy was significantly decreased, indicating that a classifier with less generalization ability had been generated by this standard CV method.
The current model was based on histopathologic and clinical chemistry data and performs best on data comparable to the training data. If there is no evidence of toxicity (see deprenyl, lazabemide, gentamicin), the gene expression profiles are not wrongly assigned to a toxicity category. Classification of lower but still somewhat toxic doses was successful with dichlorobenzene (1,500 mmol/kg), amineptine (0.25 mmol/kg), acetaminophen (1,000 mg/kg), bromobenzene (1 mmol/kg), Rx50 (1 mg/kg/day), Rx51 (0.13 mg/kg/day), and Rx60 (0.38 mg/ kg/day). Nevertheless, detection of toxic substances applied at subtoxic doses can be successful; examples include Rx10 (125 mg/kg/day) and some animals treated with glibenclamide 2.5 mg/kg. However, it is important to remember that the current model was based on solid pathology and therefore optimized for specificity. If borderline or doses just below detectable pathology were used to generate the model, correct classification at subtoxic doses could be expected in more cases. However, an increase in sensitivity is expected to be paid for by a reduced specificity (i.e., greater number of false positives).
If there were evidence for toxicity, although with a lower histopathologic score and less-pronounced clinical chemistry changes, the model generally performed well, as indicated by the relatively high test MCC values. Examples for successful classification of earlier time points are Rx99 (24 hr) and methylene dianiline (3 hr). However, very early times can often affect a different set of genes than those noted at later times (Heijne et al. 2004; Ruepp et al. 2002).
Although the test set contained some of the same compounds as the training set, the experiments in the test set used lower doses or samples collected at earlier time points. Therefore, the high classification accuracy observed with the test data indicates good sensitivity of our models.
An interesting observation was that the two experiments using chlorpromazine were not equally well classified. Chlorpromazine was expected to have a cholestatic effect at the tested doses (Knodell 1975), but the animals were classified predominantly as nontoxic in the first experiment and as cholestatic in the second experiment. However, these differences in gene expression profiles are in agreement with differences in conventional end points, probably because of biological and/or experimental differences, as both experiments were performed at different sites and with slightly different sample processing protocols.
Summarizing, we demonstrated that classification problems in toxicogenomics can be effectively addressed by a supervised learning approach. We applied SVMs on microarray data from a set of model hepatotoxicants. Combining SVM parameter optimization with a compound-based external CV scheme and (RFE), we were able to obtain accurate classification (i.e., high sensitivity and specificity) of the compounds included in the training set as well as for previously unseen compounds. In addition, RFE allowed us to select a relatively compact subset of probe sets with potential use as biomarkers. Thus, our results show that toxicogenomics is a very powerful tool for classification of compounds according to their toxicity mechanism when a well-designed database is combined with appropriate bioinformatics tools.
Despite these promising results, further investigations must be performed to increase the usefulness of transcript profiling in toxicology. A larger database and refined analysis methods are anticipated to further improve prediction accuracy.
We focused mainly on high doses that led to clear toxicity as assessed by conventional end points. However, it has been reported that a compound affects different genes and pathways depending on the administered dose (Andrew et al. 2003). Thus, a next step will be to include expression profiles from lower doses in the model-building process. Earlier time points should also be considered. This will allow us to assess whether gene expression changes are already indicative of toxic liabilities when standard parameters do not yet detect toxicity. In addition, for classification purposes it is irrelevant whether the gene expression changes considered good discriminants for a toxic response are causally linked to the toxicity. Nevertheless, to gain further insight into a specific MOT, it is valuable to interpret results in a biological context, analyzing the altered pathways and their relationship to observed pathology or phenotype. These investigations could help separate transcriptional changes that are relevant for the mode of toxicity from mere bystander effects.
Supplementary Material
Supplemental Tables Figure 1 Classification of five vehicle control or five WY14643-treated rats. Gene expression profiles of Sprague-Dawley rat livers treated either with vehicle or WY14643 were assessed with a model built exclusively on data from Wistar rats. Results of the SVM for peroxisomal proliferation are shown. All profiles from treated rats yield clearly positive discriminants, indicating that the transcriptional changes identify the substance to cause peroxisomal proliferation. Controls have clearly negative values, indicating no match with the fingerprint of the peroxisomal proliferation class.
Figure 2 Identification of nonresponding animal. Gene expression profiles from galactosamine-treated rats and vehicle controls were tested using the 4MOT model. Results from the direct-acting SVM (based on 104 genes) are projected onto a three-dimensional coordinate system for better visualization [Supplemental data (http://ehp.niehs.nih.gov/txg/members/2004/7036/7036supplement.pdf)]. Top left: gene expression profiles from direct-acting compounds. Bottom right: profiles from the remaining categories cluster together. Classification results are in line with histopathology and clinical chemistry data. The shift of the nonresponding animal toward the direct-acting group is a hint that gene expression profiling could be more sensitive than classical end points used in this study.
Figure 3 Assessment of gentamicin, deprenyl, and lazabemide. Animals were treated with a high dose of (A) gentamicin (GEN; 100 mg/kg sc, 24 hr), (B) deprenyl (DPR; 20 mg/kg/day, 4 days), or (C) lazabemide (LAZ; 1,000 mg/kg/day, 4 days). No hepatotoxicity was detected with any of the three treatments. However, nephrotoxicity was evident in GEN-treated animals. Gene expression signatures in liver tissue were related to pharmacology without association to hepatotoxicity. Thus, GEN, DPR, and LAZ were correctly identified as nontoxic. Classification of those animals with the controls is indicated by the positive discriminant values for the control SVM.
Figure 4 Classification of lipopolysaccharide, phenobarbital, and indomethacin. Abbreviations: IND, indomethacin; LPS, lipopolysaccharide; PHB, phenobarbital. (A) Animals were treated with an acute dose of LPS (4 mg/kg iv) and assessed after 6 hr. Four animals were classified as steatotic and one animal as cholestatic (animal 5). (B) Rats were dosed with PHB (80 mg/kg po) and assessed 24 hr thereafter. All five animals were considered steatotic. Rats treated with LPS or PHB had very low positive discriminant values for the toxicity categories, indicating no good fit with the representative data in the predictive model. However, the large negative discriminant values of the control SVMs in A, B, and C clearly indicate that all animals were treated with a toxicant. (C) Animals were treated with a high dose of IND (5 mg/kg po) and assessed after 1 week. Positive scores were obtained for three different toxic categories. Obviously, the profiles match some characteristics of the finger-prints of all three classes at the same time.
Table 1 Histopathology and clinical chemistry results of rats used included in the SVM training set.
Substance/dose/CAS no./supplier Vehicle/route of administration Expected binary class/4-MOT class Liver histopathology Serum clinical chemistry
Aflatoxin B1 Saline + 0.5% Toxic/direct Hepatocellular hypertrophy, apoptosis, inflammation, glycogen depletion, bile duct proliferation Increased bile acids, bilirubin, AST, ALT, LDH, ALP, 5′-NT
4 mg/kg, 24 hr DMSO/ip
1162-65-8
Sigma
Bromobenzene Corn oil/ip Toxic/direct Centrilobular to midzonal hepatocellular hydropic swelling, necrosis with mixed inflammation Increased bilirubin, 5′-NT, albumin; decreased triglycerides
3 mmol/kg, 24 hr
108-86-1
Aldrich
Carbon tetrachloride (CCl4) Corn oil/po Toxic/direct Hepatocellular degeneration, single-cell necrosis, inflammation, microvesicular steatosis Increased GGT, liver triglycerides; decreased glucose, albumin
2 mg/kg, 24 hr
56-23-5
Fluka
Hydrazine Saline/ip Toxic/direct Hepatocellular necrosis with inflammation, mild microvesicular steatosis Increased 5′-NT
60 mg/kg, 24 hr
302-01-2
Sigma
Thioacetamide Saline/ip Toxic/direct Hepatocellular vacuolation and necrosis Increased GGT, AST, ALT, ALP,5′-NT; decreased glucose, triglycerides, cholesterol, protein
50 mg/kg, 24 hr
62-55-5
Sigma-Aldrich
1,2-Dichlorobenzene Corn oil/ip Toxic/direct Centrilobular to midzonal hepatocellular hydropic swelling, necrosis with mixed inflammation Increased ALP, albumin; decreased triglycerides
4,500 mmol/kg, 24 hr
95-50-1
Fluka
Coumarin Corn oil/po Toxic/direct Hepatocellular hypertrophy, single-cell necrosis, lymphocytic infiltration Increased total protein, GLD
200 mg/kg, 24 hr
91-64-5,
Sigma
Acetaminophen Saline + 0.5% Toxic/direct Centrilobular hepatocellular vacuolation, single-cell necrosis, inflammation Increased albumin; decreased triglycerides
2 g/kg, 24 hr DMSO/po
103-90-2
Fluka
Amineptine Saline/ip Toxic/steatosis Hepatocellular microvesicular steatosis, glycogen depletion Increased GGT, ALP, cholesterol; decreased triglycerides
0.5 mmol/kg/day, 2 days
57574-09-1
Servier Laboratories
Amiodarone 7.5% gelatine/ip Toxic/steatosis Hepatocellular microvesicular steatosis, glycogen depletion Increased GGT, 5′-NT; decreased serum and increased liver triglycerides
100 mg/kg/day, 4 days
1951-25-3
Sigma
Rx74 (Antidiabetic) Klucel/po Toxic/steatosis NDa NDa
250 mg/kg/day, 5 days
Not available
Roche
Rx75 (Antidiabetic) Klucel/po Toxic/steatosis NDa NDa
100 mg/kg/day, 5 days
Not available
Roche
Rx10 (Antidiabetic) Klucel/po Toxic/steatosis NDb NDb
500 mg/kg/day, 5 days
Not available
Roche
Rx99 (5-HT6 antagonist) H2O/po Toxic/steatosis Hepatocellular microvesicular steatosis Increased ALT, GGT; increased liver lipids and phospholipids
400 mg/kg/day, 14 days
Not available
Roche
Chlorpromazine 1 Saline/iv Toxic/cholestasis ND Increased bilirubin, glucose; decreased triglycerides
15 mg/kg, 6 hr
69-09-0
Sigma
Chlorpromazine 2 Saline/iv Toxic/cholestasis Hepatocellular microvesicular steatosis, glycogen depletion Increased glucose; decreased triglycerides, protein
15 mg/kg, 6 hr
69-09-0
Sigma
Cyclosporin A 10% intralipid/iv Toxic/cholestasis NSF Increased bile acids, bilirubin, GGT
30 mg/kg, 6 hr
59865-13-3
Alexis
Glibenclamide 7.5% gelatine/iv Toxic/cholestasis Hepatocellular hypertrophy Increased ALT; decreased glucose
25 mg/kg, 6 hr
10238-21-8
Roche
Phalloidin Saline/iv Toxic/cholestasis Hepatocellular necrosis, hemorrhage, glycogen depletion Increased bilirubin, bile acids, 5′-NT, ALP, AST, ALT, LDH, SDH; decreased cholesterol, phospholipids
0.8 mg/kg, 6 hr
17466-45-4
Sigma
Methylene dianiline Corn oil/po Toxic/cholestasis Single-cell necrosis of bile duct epithelium, inflammation Increased bilirubin, bile acids, GGT, 5′-NT, glucose, phospholipids
100 mg/kg, 6 hr
101-77-9
Fluka
WY14643 Corn oil/po Toxic/PP Increased hepatocellular mitoses, slight glycogen depletion, increased liver weight (7 days) Increased ALP, glucose, SDH
250 mg/kg, 14 days
50892-23-4
Sigma-Aldrich
Rx90 (PPAR-δ agonist) PBS/po Toxic/PP Liver enlargement, diffuse hepatocellular hypertrophy Increased AST, ALT
180 mg/kg/day, 14 days
Not available
Roche
Rx53 (PPAR-α,γ co-agonist) PBS/po Toxic/PP Increased liver weight, hepatocellular hypertrophy and cytoplasmic granulation Decreased cholesterol, protein
0.9 mg/kg/day, 14 days
Not available
Roche
Rx60 (PPAR-α,γ co-agonist) PBS/po Toxic/PP Increased liver weight, hepatocellular hypertrophy and cytoplasmic granulation, increased mitoses, single-cell necrosis with mixed inflammation Increased serum ALP; decreased protein, bilirubin
1.5 mg/kg/day, 14 days
Not available
Roche
Rx51 (PPAR-α,γ co-agonist) PBS/po Toxic/PP Increased liver weight, hepatocellular hypertrophy and cytoplasmic granulation Increased ALP; decreased cholesterol, bilirubin, protein
0.5 mg/kg/day, 14 days
Not available
Roche
Rx50 (PPAR-α,γ co-agonist) PBS/po Toxic/PP Increased liver weight, hepatocellular hypertrophy and cytoplasmic granulation Increased ALP, glucose; decreased protein, bilirubin, cholesterol
4 mg/kg/day, 14 days
Not available
Roche
Abbreviations: DMSO, dimethylsulfoxide; ND, not done; NSF, no significant findings; PBS, phosphate-buffered saline; PP, peroxisome proliferator.
a No clinical chemistry or histopathology data were available from animals used for gene profiling, but repeated dosing with this compound in animals used for other measurements resulted in microvesicular steatosis.
b No clinical chemistry or histopathology data were available from animals used for gene profiling. Microvesicular steatosis was not detected in rats with this treatment schedule. However, in vitro treatment of primary rat hepatocytes inhibited β-oxidation and resulted in fat accumulation.
Table 2 Performance of the toxic/nontoxic models and summarized results of the binary (toxic/nontoxic) classification.a
Arrays/groups for classification ν-SVM C-SVM
Classification under external CV
26 treatment groups 20 of 26 treatments correct 22 of 26 treatments correct
116 arrays 89 of 116 arrays correct 90 of 116 arrays correct
34 control groups 32 of 34 groups correct 32 of 34 groups correct
163 arrays 154 of 163 arrays correct 154 of 163 arrays correct
Classification of test set
19 treatment groups 16 of 19 treatments correct 17 of 19 treatments correct
91 arrays 74 of 91 arrays correct 74 of 91 arrays correct
63 control groups 63 of 63 (all groups correct) 63 of 63 (all groups correct)
332 arrays 322 of 332 arrays correct 327 of 332 arrays correct
a During RFE, the least informative 5% of genes were removed in each iteration starting with all features (genes) down to 64 genes. After that, only a single gene was removed in one step. The number of features finally selected was 63 for the ν-SVM and 228 for the C-SVM. In the case of ν-SVM, RFE was carried out with ν = 0.1. The optimized ν of the selected (using 63 genes) is 0.203. For C-SVM we set C to 0.008 during RFE and ended up with C = 0.00429 for the selected iteration. Both SVMs were equally successful in classifying vehicle controls, but the C-SVM was slightly better in identifying toxic treatments.
Table 3 Performance assessment of the five SVMs that form the 4MOT model.a
Class Features CV specificity CV sensitivity CV MCC Optimized Test specificity Test sensitivity Test MCC
Classification with υ-SVM
Direct 101 1 0.86 0.92 0.0377 1 0.75 0.83
PP 4 1 1 1 0.01 1 1 1
Cholestasis 19 0.99 0.6 0.71 0.0193 0.99 0.83 0.82
Steatosis 28 1 0.54 0.72 0.0744 1 0.91 0.95
Control 122 0.78 0.94 0.75 0.111 0.84 0.98 0.86
Classification with C-SVM
Direct 38 1 0.84 0.9 0.0176 1 0.75 0.83
PP 16 1 1 1 0.0222 1 1 1
Cholestasis 32 0.98 0.57 0.61 0.1 0.98 0.83 0.81
Steatosis 50 0.99 0.67 0.75 0.00869 1 0.91 0.95
Control 228 0.78 0.94 0.74 0.00429 0.8 0.98 0.83
a Results are shown for υ-SVM and C-SVM. The RFE procedure was identical to that described in Table 2. The number of features selected was typically smaller for υ-SVM than for C-SVM. Both types of SVM were comparably successful in classification.
Table 4 Classification of individual microarrays and treatment groups in training set and overview of CV and test results for a υ-SVM–based model discriminating between different MOTs.a
Treatment Expected toxicity category CV accuracy (binary) CV accuracy (4MOT) Misclassification in 4MOT
Chlorpromazine 1 Cholestatic 1/5b 1/5b 4 controlsb
Chlorpromazine 2 Cholestatic 4/5 4/5 1 control
Cyclosporin A Cholestatic 4/5 4/5 1 control
Glibenclamide Cholestatic 0/5b 0/5b 5 controlsb
Methylene dianiline Cholestatic 5/5 5/5 –
Phalloidin Cholestatic 3/5 2/5 1 direct acting, 2 controls
Aflatoxin B1 Direct acting 2/3 1/3 1 cholestatic, 1 control
1,2-Dichlorobenzene Direct acting 5/5 5/5 –
APAP Direct acting 3/5 3/5 2 controls
Bromobenzene Direct acting 5/5 5/5 –
CCl4 Direct acting 5/5 5/5 –
Coumarin Direct acting 5/5 5/5 –
Hydrazine Direct acting 5/5 5/5 –
Thioacetamide 1 Direct acting 3/5 3/5 2 controls
Rx50 (PPAR-α, γ) PP 5/5 5/5 –
Rx53 (PPAR-α, γ) PP 2/4b 2/4b 2 controlsb
Rx51 (PPAR-α, γ) PP 5/5 5/5 –
Rx60 (PPAR-α, γ) PP 5/5 5/5 –
WY14643 PP 5/5 5/5 –
Rx90 (PPAR-δ) PP 5/5 5/5 –
Rx99 (5HT6) Steatotic 3/5 3/5 2 controls
Amineptine Steatotic 4/5 4/5 1 control
Amiodarone Steatotic 0/5b 0/5b 5 controlsb
Rx74 (anitdiabetic) Steatotic 3/3 3/3 –
Rx75 (anitdiabetic) Steatotic 2/3 2/3 1 control
Rx10 (anitdiabetic) Steatotic 3/3 3/3 –
a Predictions for individual microarrays and treatment groups as a whole were obtained using different voting schemes described in the text. A compound-based external CV method was used for the assessment of model quality. The upper part of the table reports the number of microarrays correctly classified under CV conditions, either with correct mechanism of action predicted (column 4) or with at least a toxic effect recognized (column 3).
b Misclassifications.
Table 5 Performance summary of the υ-SVM–based model discriminating between different MOTs.
Arrays/groups for classification Summary
26 treatment groups 20 of 26 treatment groups correct MOT identified
22 of 26 treatment groups correctly identified as toxic
116 microarrays 85 of 116 microarrays correctly classified
34 control groups 33 of 34 groups correctly identified as vehicle controls
163 microarrays 160 of 163 microarrays correctly classified
Classification of independent test set
19 treatment groups 15 of 19 treatment groups correct MOT identified
15 of 19 treatment groups correctly identified as toxic
91 microarrays 74 of 91 microarrays correctly classified
63 treatment groups 63 of 63 (all groups correctly identified)
332 microarrays 330 of 332 microarrays correctly classified
==== Refs
References
Alizadeh AA Ross DT Perou CM Van De Rijn M 2001 Towards a novel classification of human malignancies based on gene expression patterns J Pathol 195 41 52 11568890
Alon U Barkai N Notterman DA Gish K Ybarra S Mack D 1999 Broad patterns of gene expression revealed by clustering analysis of tumor and normal colon tissues probed by oligonucleotide arrays Proc Natl Acad Sci USA 96 6745 6750 10359783
Amaya-Farfan J 1999 Aflatoxin B1 -induced hepatic steatosis: role of carbonyl compounds and active diols on steatogenesis Lancet 353 747 748 10073532
Andrew AS Warren AJ Barchowsky A Temple KA Klei L Soucy NV 2003 Genomic and proteomic profiling of responses to toxic metals in human lung cells Environ Health Perspect 111 825 838 12760830
Arbeitman MN Furlong EE Imam F Johnson E Null BH Baker BS 2002 Gene expression during the life cycle of Drosophila melanogaster Science 297 2270 2275 12351791
Asamoto M Tsuda H Kato T Ito N Masuko T Hashimoto Y 1989 Strain differences in susceptibility to 2-acetylaminofluorene and phenobarbital promotion of hepatocarcinogenesis: immunohistochemical analysis of cytochrome P-450 isozyme induction by 2-acetyl-aminofluorene and phenobarbital Jpn J Cancer Res 80 1041 1046 2514165
Boess F Kamber M Romer S Gasser R Muller D Albertini S 2003 Gene expression in two hepatic cell lines, cultured primary hepatocytes, and liver slices compared to the in vivo liver gene expression in rats: possible implications for toxicogenomics use of in vitro systems Toxicol Sci 73 386 402 12657743
Boser BE Guyon IM Vapnik VN 1992. A training algorithm for optimal margin classifiers. In: Proceedings of the 5th Annual International Conference on Computational Learning Theory, 27–29 July 1992, Pittsburgh, Pennsylvania. Pittsburgh, PA:ACM Press, 144–152.
Brem RB Yvert G Clinton R Kruglyak L 2002 Genetic dissection of transcriptional regulation in budding yeast Science 296 752 755 11923494
Brown MPS Grundy WN Lin D Cristianini N Sugnet CW Furey TS 2000 Knowledge-based analysis of microarray gene expression data by using support vector machines Proc Natl Acad Sci USA 97 262 267 10618406
Bulera SJ Eddy SM Ferguson E Jatkoe TA Reindel JF Bleavins MR 2001 RNA expression in the early characterization of hepatotoxicants in Wistar rats by high-density DNA microarrays Hepatology 33 1239 1258 11343254
Burczynski ME McMillian M Ciervo J Li L Parker JB Dunn RT II 2000 Toxicogenomics-based discrimination of toxic mechanism in HepG2 human hepatoma cells Toxicol Sci 58 399 415 11099651
Chang CC Lin CJ 2001. LIBSVM: a library for support vector machines. Available: http://www.csie.ntu.edu.tw/~cjlin/libsvm [1 January 2003].
Cristianini N Shawe-Taylor J 2000. An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge, UK:Cambridge University Press.
Donald S Verschoyle RD Edwards R Judah DJ Davies R Riley J 2002 Hepatobiliary damage and changes in hepatic gene expression caused by the antitumor drug ecteinascidin-743 (ET-743) in the female rat Cancer Res 62 4256 4262 12154027
Eisen MB Spellman PT Brown PO Botstein D 1998 Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci USA 95 14863 14868 9843981
Enard W Khaitovich P Klose J Zollner S Heissig F Giavalisco P 2002 Intra- and interspecific variation in primate gene expression patterns Science 296 340 343 11951044
Everitt B 1974. Cluster Analysis. London:Heinemann.
Farrell GC 1994. Drug-Induced Liver Disease. London:Churchill Livingstone.
Fracasso ME Leone R Cuzzolin L Del Soldato P Velo GP Benoni G 1990 Indomethacin induced hepatic alterations in mono-oxygenase system and faecal Clostridium perfringens enterotoxin in the rat Agents Actions 31 313 316 2128167
Furey TS Cristianini N Duffy N Bednarski DW Schummer M Haussler D 2000 Support vector machine classification and validation of cancer tissue samples using microarray expression data Bioinformatics 16 906 914 11120680
Guyon IM Weston J Barnhill S Vapnik VN 2002 Gene selection for cancer classification using support vector machines Mach Learn 46 389 442
Hamadeh HK Bushel PR Jayadev S DiSorbo O Bennett L Li L 2002a Prediction of compound signature using high density gene expression profiling Toxicol Sci 67 232 240 12011482
Hamadeh HK Bushel PR Jayadev S Martin K Di Sorbo O Sieber S 2002b Gene expression analysis reveals chemical-specific profiles Toxicol Sci 67 219 231 12011481
Hamadeh HK Knight BL Haugen AC Sieber S Amin RP Bushel PR 2002c Methapyrilene toxicity: anchorage of pathologic observations to gene expression alterations Toxicol Pathol 30 470 482 12187938
Hansmannel F Clemencet MC Le Jossic-Corcos C Osumi T Latruffe N Nicolas-Frances V 2003 Functional characterization of a peroxisome proliferator response-element located in the intron 3 of rat peroxisomal thiolase B gene Biochem Biophys Res Commun 311 149 155 14575706
Hastie T Tibshirani R Eisen MB Alizadeh A Levy R Staudt L 2000 ‘Gene shaving’ as a method for identifying distinct sets of genes with similar expression patterns Genome Biol 1 3 21
Heijne WH Slitt AL van Bladeren PJ Groten JP Klaassen CD Stierum RH 2004 Bromobenzene-induced hepatotoxicity at the transcriptome level Toxicol Sci 79 411 422 15056800
Joliffe IT 1986. Principal Component Analysis. New York:Springer.
Kaiser R Gottschalk G 1972. Elementare Tests zur Beurteilung von Messdaten [in German]. Mannheim/Wien/Zuerich:Bibliographisches Institut 18–21.
Kita Y Shiozawa N Jin WH Majewski RR Besharse JC Greene AS 2002 Implications of circadian gene expression in kidney, liver and the effects of fasting on pharmacogenomic studies Pharmacogenetics 12 55 65 11773865
Knodell RG 1975 Effects of chlorpromazine on bilirubin metabolism and biliary secretion in the rat Gastroenterology 69 965 972 1175890
Kohonen T 1995. Self Organizing Maps. Berlin:Springer.
Kulkarni SG Duong H Gomila R Mehendale HM 1996 Strain differences in tissue repair response to 1,2-dichlorobenzene Arch Toxicol 70 714 723 8896717
Lee CH Olson P Evans RM 2003 Minireview: lipid metabolism, metabolic diseases, and peroxisome proliferator-activated receptors Endocrinology 144 2201 2207 12746275
Lindon JC Nicholson JK Holmes E Antti H Bollard ME Keun H 2003 Contemporary issues in toxicology: the role of metabonomics in toxicology and its evaluation by the COMET project Toxicol Appl Pharmacol 187 137 146 12662897
Liu J Yang CF Lee BL Shen HM Ang SG Ong CN 1999 Effect of Salvia miltiorrhiza on aflatoxin B1 -induced oxidative stress in cultured rat hepatocytes Free Radic Res 31 559 568 10630680
Masubuchi Y Saito H Horie T 1998 Structural requirements for the hepatotoxicity of nonsteroidal anti-inflammatory drugs in isolated rat hepatocytes J Pharmacol Exp Ther 287 208 213 9765339
Matthews BW 1975 Comparison of the predicted and observed secondary structure of t4 phage lysozyme Biochim Biophys Acta 405 442 451 1180967
Panda S Antoch MP Miller BH Su AI Schook AB Straume M 2002 Coordinated transcription of key pathways in the mouse by the circadian clock Cell 109 307 320 12015981
Ramaswamy S Ross KN Lander ES Golub TR 2003 A molecular signature of metastasis in primary solid tumors Nat Genet 33 49 54 12469122
Ramaswamy S Tamayo P Rifkin R Mukherjee S Yeang CH Angelo M 2001 Multiclass cancer diagnosis using tumor gene expression signatures Proc Natl Acad Sci USA 98 15149 15154 11742071
Ruepp S Tonge RP Shaw J Wallis N Pognan F 2002 Genomics and proteomics analysis of acetaminophen toxicity in mouse liver Toxicol Sci 65 135 150 11752693
Schölkopf B Guyon IM Weston J 2003. Statistical learning and kernel methods in bioinformatics. In: Artifical Intelligence and Heuristic Methods in Bioinformatics (Frasconi P, Shamir R, eds). Amsterdam:IOS Press, 1–21.
Schölkopf B Smola A 2002. Learning with Kernels. Cambridge, MA:MIT Press.
Smela ME Currier SS Bailey EA Essigmann JM 2001 The chemistry and biology of aflatoxin B(1): from mutational spectrometry to carcinogenesis Carcinogenesis 22 535 545 11285186
Storch KF Lipan O Leykin I Viswanathan N Davis FC Wong WH 2002 Extensive and divergent circadian gene expression in liver and heart Nature 417 78 83 11967526
Suter L Haiker M De Vera MC Albertini S 2003 Effect of two 5-ht(6) receptor antagonists on the rat liver: a molecular approach Pharmacogenomics J 3 320 334 14610520
Thomas RS Rank DR Penn SG Zastrow GM Hayes KR Pande K 2001 Identification of toxicologically predictive gene sets using cDNA microarrays Mol Pharmacol 60 1189 1194 11723225
Unger PD Mehendale HM Hayes AW 1977 Hepatic uptake and disposition of aflatoxin B1 in isolated perfused rat liver Toxicol Appl Pharmacol 41 523 534 918986
Van de Vijver MJ He YD van’t Veer LJ Dai H Hart AA Voskuil DW 2002 A gene-expression signature as a predictor of survival in breast cancer N Engl J Med 347 1999 2009 12490681
Vapnik VN 1998. Statistical Learning Theory. New York:Wiley.
Vomel T Platt D 1986 Age-dependent phagocytosis of erythrocytes by the isolated perfused rat liver after galactosamine hepatitis and alpha-naphthylisothiocyanate cholestasis Arch Gerontol Geriatr 5 351 539 3827405
Wang Y Rea T Bian J Gray S Sun Y 1999 Identification of the genes responsive to etoposide-induced apoptosis: application of DNA microarray technology FEBS Lett 445 269 273 10094470
Waring JF Ciurlionis R Jolly RA Heindel M Ulrich RG 2001a Microarray analysis of hepatotoxins in vitro reveals a correlation between gene expression profiles and mechanisms of toxicity Toxicol Lett 120 359 368 11323195
Waring JF Jolly RA Ciurlionis R Lum PY Praestgaard JT Morfitt DC 2001b Clustering of hepatotoxins based on mechanism of toxicity using gene expression profiles Toxicol Appl Pharmacol 175 28 42 11509024
Yeang C Ramaswamy S Tamayo P Mukherjee S Rifkin RM Angelo M 2001 Molecular classification of multiple tumor types Bioinformatics 17 316 322
Zhang W Wang H Song SW Fuller GN 2002 Insulin-like growth factor binding protein. 2: Gene expression microarrays and the hypothesis-generation paradigm Brain Pathol 12 87 94 11770904
Zimmermann N King NE Laporte J Yang M Mishra A Pope SM 2003 Dissection of experimental asthma with DNA microarray analysis identifies arginase in asthma pathogenesis J Clin Invest 111 1863 1874 12813022
|
15345370
|
PMC1277117
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug 1; 112(12):1236-1248
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7036
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7125ehp0112-00124915345371ToxicogenomicsArticlesAssessment of Prediction Confidence and Domain Extrapolation of Two Structure–Activity Relationship Models for Predicting Estrogen Receptor Binding Activity Tong Weida 1Xie Qian 2Hong Huixiao 2Shi Leming 1Fang Hong 2Perkins Roger 21Center for Toxicoinformatics, and2Bioinformatics Laboratory, National Center for Toxicological Research, Food and Drug Administration, Jefferson, Arkansas, USAAddress correspondence to W. Tong, Center for Toxicoinformatics, Division of Biometry and Risk Assessment, National Center for Toxicological Research, 3900 NCTR Rd., HFT 020, Jefferson, AR 72079, USA. Telephone: (870) 543-7142. Fax: (870) 543-7662. E-mail:
[email protected] authors declare they have no competing financial interests.
8 2004 16 7 2004 112 12 1249 1254 26 3 2001 15 7 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Quantitative structure–activity relationship (QSAR) methods have been widely applied in drug discovery, lead optimization, toxicity prediction, and regulatory decisions. Despite major advances in algorithms and software, QSAR models have inherent limitations associated with a size and chemical-structure diversity of the training set, experimental error, and many characteristics of structure representation and correlation algorithms. Whereas excellent fit to the training data may be readily attainable, often models fail to predict accurately chemicals that are outside their domain of applicability. A QSAR’s utility and, in the case of regulatory decisions, justification for usage increasingly depend on the ability to quantify a model’s potential for predicting unknown chemicals with some known degree of certainty. It is never possible to predict an unknown chemical with absolute certainty. Here we report on two QSAR models based on different data sets for classification of chemicals according to their ability to bind to the estrogen receptor. The models were developed by using a novel QSAR method, Decision Forest, which combines the results of multiple heterogeneous but comparable Decision Tree models to produce a consensus prediction. We used an extensive cross-validation process to define an applicability domain for model predictions based on two quantitative measures: prediction confidence and domain extrapolation. Together, these measures quantify the accuracy of each prediction within and outside of the training domain. Despite being based on large and diverse training sets, both QSAR models had poor accuracy for chemicals within the domain of low confidence, whereas good accuracy was obtained for those within the domain of high confidence. For prediction in the high confidence domain, accuracy was inversely proportional to the degree of domain extrapolation. The model with a larger training set of 1,092, compared with 232 for the other, was more accurate in predicting chemicals at larger domain extrapolation, and could be particularly useful for rapidly prioritizing potential endocrine disruptors from large chemical universe.
applicability domainDecision Forestdomain extrapolationEDCsendocrine-disrupting chemicalsestrogen receptor bindingQSARprediction confidenceregulatory application
==== Body
Quantitative structure–activity relationships (QSARs) have been extensively applied in a broad range of scientific areas, including chemistry, biology and toxicology (Hansch et al. 1995a ,1995b). QSAR is now an inexorably imbedded tool in drug development, from lead discovery to lead optimization (Hopfinger and Tokarski 1997; Kubinyi et al. 1998). There is increasing use of QSAR early in the drug discovery process as a screening and enrichment tool to eliminate from further development those chemicals lacking drug-like properties (Lipinski et al. 1997) or those chemicals predicted to elicit a toxic response. The availability of powerful new algorithms and scientists trained in their usage suggests the eventual common use of QSAR beyond the pharmaceutical industry to human and environmental regulatory authorities (Benigni and Richard 1998; Bradbury 1994; Hansch et al. 1995a, 1995b; Russom et al. 1995; Schultz and Seward 2000; Tong et al. 2002 ,2003a).
Any QSAR model produces some degree of error. This is partially due to the inherent limitation in predicting a biological activity based solely on the chemical structure. One can argue from the principles of chemistry that molecular structure of a chemical is the key to understanding its physicochemical properties and ultimately its biological activity and the influence on organisms (Johnson and Maggiora 1990). However, biological activity of a chemical is an induced response that is influenced by numerous factors dictated by the levels of biological complexity of the system under investigation. The relationship between structure and activity is thus more implicit and thereby requires a more thorough investigation and rigorous validation (Tong et al., 2004).
Application of QSARs in regulation has proven to be cost effective for prioritizing untested chemicals for more extensive and costly experimental evaluation. However, for QSARs to be accepted by the regulatory communities, their limitation for use needs to be identified. This is important because a QSAR model’s ability to predict unknown chemicals depends largely on the nature of the training set and the algorithm used to establish the structure–activity relationship (Eriksson et al. 2003). A model’s predictive accuracy and confidence for different unknown chemicals varies according to how well the training set represents the unknown chemicals and how robust the model is in extrapolating beyond the chemistry space defined by the training set (i.e., training domain). Therefore, assessing a model’s “prediction confidence,” defined as the certainty for a prediction, and “domain extrapolation,” defined as the prediction accuracy outside the training domain, is a vital step toward defining the application domain of a model for the regulatory acceptance of QSARs.
A large number of environmental chemicals known as endocrine-disrupting chemicals (EDCs) are suspected of disrupting endocrine functions by mimicking or antagonizing natural hormones in experimental animals, wildlife, and humans (Hileman 1997). EDCs may exert adverse effects through a variety of mechanisms, including estrogen receptor (ER)–mediated mechanisms of toxicity (Fang et al. 2003b). Accordingly, the U.S. Congress in 1996 mandated that the U.S. Environmental Protection Agency (EPA) develop a strategy for screening and testing a large number of chemicals found in drinking water (Safe Drinking Water Act 1996), and food additives (Food Quality Protection Act 1996) for their endocrine disruption potential. Consequently, more than 58,000 environmental and industrial chemicals have been identified as candidates for possible experimental testing. QSARs could be used as an inexpensive prescreening tool to prioritize the chemicals for further testing (Tong et al. 2002).
In this article, we applied a novel consensus QSAR method, called Decision Forest (DF) (Tong et al. 2003b), to classify chemicals into active and inactive categories of ER binding as a priority-setting tool for EDCs. We assessed the applicability domain of the DF models through characterizing the prediction confidence and domain extrapolation for predicting unknown chemicals.
Material and Methods
Estrogen Receptor Data Sets and Structural Descriptors
Two data sets were used, and the ER binding activity for both data sets was obtained from the competitive ER binding assay (Blair et al. 2000; Branham et al. 2002). The first data set, designated ER232, contained 232 chemicals, 131 active, and 101 inactive that were tested in our lab (Fang et al. 2003a). This data set has been extensively used by others and us to develop SAR/QSAR models for predicting ER binding activity (Hong et al. 2002; Shi et al. 2001, 2002; Tong et al. 2002, 2003c). The second data set, designated ER1092, is an aggregation of data from the literature containing 1,092 chemicals, of which 350 are active and 736 are inactive. Inactive means that no activity was detectable in the assay. Both data sets span a wide range of structural diversity and activity.
Because a previous study indicated no significant difference in results between two-dimensional (2D) descriptors and 3D descriptors in DF (Tong et al. 2003b), only 2D descriptors were used in this study, and these were computed using Molconn-Z, version 4.07 (http://www.eslc.vabiotech.com/molconn/). After removing descriptors that were constant across all chemicals in a data set, more than 270 descriptors remained and were used in model development.
The structural diversity of both data sets was compared in the chemistry space defined by the 2D descriptors on the first three principle components plot (Figure 1). Not surprisingly, ER1092 was found to span much greater structural diversity than ER232.
Decision Forest
DF is a consensus modeling technique (Tong et al. 2003b) that combines multiple Decision Tree models (hereafter called trees) in a manner that results in more accurate predictions. Because combining several identical trees produces no gain, the rationale behind DF is use of individual trees that are different (i.e., heterogeneous) yet comparable in their prediction accuracy to represent the association of structure and biological activity. The heterogeneity requirement assures that each tree uniquely contributes to the combined prediction, whereas the quality comparability requirement assures that each tree contributes equally to the combined prediction. Because a certain degree of noise is always present in biological data, optimizing a tree inherently risks overfitting the noise. DF attempts to minimize overfitting by maximizing the difference among individual trees to cancel some random noise through combining the trees. The maximum difference was achieved by constructing each individual tree using a distinct set of descriptors.
Details of the DF algorithm have been reported by Tong et al. (2003b). Briefly, developing a DF model (called forest hereafter) comprises four steps: a) construct and prune a tree; b) develop the next tree based on only the descriptors that have not been used in the previous tree(s); c) repeat steps 1 and 2 until no more trees can be developed; d) classify (i.e., predict) a chemical based on the results of all trees.
Each tree in a forest is developed using a variant of the Classification and Regression Tree (CART) method (Breiman et al. 1995) that has two steps: a) tree construction and b) tree pruning. During tree construction, the algorithm identifies the descriptors that best divide the chemicals in the parent node into two child nodes. The split maximizes the homogeneity of the activity population in each child node (e.g., one node predominately contains active chemicals, whereas the other predominately contains inactive chemicals). Then, the child nodes become parent nodes for further splits and splitting continues until chemicals in each node are either in one classification category or cannot be split further to improve the quality of the tree. To avoid overfitting the training data, the tree is then cut down to a desired size using tree cost-complexity pruning (Clark and Pregibon 1997). At the end, the terminal node of each tree generally is populated by different ratios of active versus inactive chemicals.
In each tree, the probability (0–1) for an “unknown” chemical to be active is taken to be the percentage of active chemicals in the terminal node to which the chemical belongs. The mean probability value for a chemical in all trees in the forest is calculated to assign the classification of the chemical. Chemicals that have a mean probability > 0.5 are designated active, whereas those that have a mean probability < 0.5 are designated inactive.
Prediction Confidence
Past results have shown that DF predictions are of high confidence for active chemicals with a large probability value (approaching 1) and for inactive chemicals with low probability value (approaching zero), whereas the low confidence predictions are mostly found for chemicals with probability approaching 0.5 (Tong et al. 2003b). Based on this observation, the following equation was used to calculate the confidence level of a prediction:
where Pi is the probability value for chemical i. In this equation, the confidence associated with active and inactive prediction is scaled in parallel to the range between zero and 1. If we assume that a high confidence prediction is defined as confidence level > 0.4, both probability ranges of 0.0–0.3 and 0.7–1.0 will be considered the high confidence (HC) region, and 0.3–0.7 is the low confidence (LC) region. In other words, a high prediction certainty is expected when a chemical with predicted probability in the range 0.0–0.3 is classified as inactive, or when a chemical with probability in the range 0.7–1.0 is predicted as active. In contrast, prediction confidence is lower for chemicals with probabilities in the range 0.3–0.7.
Domain Extrapolation
Suppose there is a forest that contains n trees (i =1,…n). For the ith tree, the classification of an unknown chemical is determined by only one terminal node that is descendent from the root node through a set of “IF-THEN” rules based on k descriptors xij (j=1,…k) (Figure 2). Let xij(max) and xij(min) denote the maximum and minimum values for xij across the entire data set and yij denote the descriptor values of the unknown chemicals corresponding to xij. If yij is either > xij(max) or <xij(min), then it is outside the range of the training domain defined by xij in the “IF-THEN” rule in the path from the root to the terminal node in the ith tree. Thus, the distance beyond the training domain for the unknown chemical in the tree i can be calculated by dij = |yij − xij(max)| if yij > xij(max), dij = |yij − xij(min)| if yij < xij(min), or dij = 0 if yij> xij(min) or yij < xij(max) (within the training domain). For the forest, the total percentage of extrapolation outside the training domain is:
The prediction accuracy within domain d is calculated by dividing correct predictions by total number of chemicals in this extrapolated domain.
Cross-Validation for Assessing Prediction Confidence and Domain Extrapolation
We used 10-fold cross-validation to assess a forest’s prediction accuracy for unknown chemicals in different domains of prediction confidence and extrapolation. In this procedure the data set is randomly divided into 10 equal portions, and each portion is excluded once and predicted by the forest produced using the remaining nine portions to train the model. Because the 10-fold cross-validation results vary for each run due to random partitioning of the data set, we repeated the process 2,000 times. The average result of the multiple cross-validation runs provides an unbiased assessment of a forest for predicting unknown chemicals with respect to prediction confidence and extrapolation sensitivity.
Results
Table 1 summarizes the fitting results of the forests for both the ER232 and ER1092 data sets. The forests had concordances around 95% with high specificity and sensitivity.
Since a statistically sound fitted model provides limited indication of its capability for predicting chemicals that are not included in training, we applied 2,000 runs of 10-fold cross-validation to assess the prediction confidence and extrapolation sensitivity of the model for predicting unknown chemicals.
Figure 3 plots forest prediction accuracy versus prediction confidence for ER232 (Figure 3A) and ER1092 (Figure 3B), respectively. For comparison, the results of the first tree in each forest are also plotted in Figure 3. It is readily apparent that the forests have substantially higher prediction accuracy than the tree across the entire range of confidence levels. Importantly, there is a strong trend of higher accuracy with increasing confidence level. We arbitrarily defined two confidence regions, HC and LC corresponding to confidence levels > 0.4 and < 0.4, respectively. Table 2 compares the HC, LC, and overall prediction accuracy. The HC prediction accuracy is approximately 86%, about 22% higher than the prediction accuracy for the LC regions (~ 64%). There is about 5–7% higher prediction accuracy for the HC regions than for the overall prediction accuracy (Table 2). The HC predictions account for approximately 80% of chemicals for ER232 and approximately 70% for ER1092.
On the basis of the same cross-validation results, we also assessed the prediction accuracy for the chemicals as a function of extrapolation outside the training domain. Figure 4 compares for both ER232 and ER1092 the overall prediction accuracy for chemicals within the domain defined by the training set chemicals with accuracy for chemicals falling several degrees of extrapolation outside the training domain, as defined by Equation 2. Generally, the farther away the chemicals were from the training domain, the more loss in prediction accuracy was observed. For ER232 the prediction accuracy was reduced by some 10% for chemicals with a 10% extrapolation. In contrast for ER1092, a major decrease in accuracy only occurred beyond a 30% extrapolation.
Table 3 further breaks down the overall prediction accuracy shown in Figure 4 into the accuracies for the HC and LC regions and also gives the distribution of predictions within the extrapolated domains. For the HC prediction region the trend of decreasing prediction accuracy with increasing extrapolation is consistent with Figure 4 for both ER232 and ER1092. In the HC region for both data sets, prediction accuracy is comparable when extrapolation does not exceed 10%. Prediction accuracy declines more notably for chemicals with > 10% extrapolation for ER232 (some with > 16%), and for chemicals with > 30% extrapolation for ER1092. In contrast the LC region prediction accuracy is consistently lower, as expected, and exhibits no discernable trend with extent of extrapolation.
Discussion
We used the novel QSAR method DF to develop two classification models to predict ER binding. Such models could be important in prioritizing chemicals for testing based on likelihood of activity. We furthermore objectively and quantitatively assessed the applicability domains of the models by computing prediction confidence and domain extrapolation for predicting unknown chemicals with an extensive cross-validation. We found that accuracy in classifying unknown chemicals is dependent on both prediction confidence and domain extrapolation, with the dependence most pronounced for prediction confidence. The prediction accuracy is notably higher for the chemicals in the HC domain than for those in the LC domain. In the HC domain, the forest model based on the large data set ER1092 is much better able to extrapolate outside the structural domain defined by the training data than is the forest model based on the small data set ER232 and specifically by some 30% compared with 10%. We propose that the ER1092 model is most suitable for aiding in prioritizing chemicals for testing as possible EDCs.
The consistently lower prediction accuracy in the LC domain compared to that of the HC domain seems minimally affected by the extent of extrapolation. For many repeated runs of cross-validation with random partitioning, chemicals in the HC domain average 70–80% of the total for both data sets. It should be noted that the distribution of the chemicals between the high and low confidence regions could vary when applying the model to a test set. Actual distribution depends largely on how well the training set represents the test set chemicals. In the cross-validation, however, the proportion of chemicals in the HC domain is sensitive to the structural diversity and quality of the training set.
The ability to quantify confidence greatly enhances the utility of any classification or QSAR method. The ability to accurately gauge confidence of predictions may also determine how best to apply the model. For example, considering the forest models presented here for use in screening and testing for potential EDCs, the HC and LC domain predictions could be used in separate ways. Chemicals in the HC domain are candidates for applying more rigorous quantitative models (Shi et al. 2001) to calculate binding affinities that are, in turn, used to rank-order chemicals for experimental evaluation. However, for chemicals in the LC domain, more thorough evaluation based on other types of models (Hong et al. 2002; Shi et al. 2002; Tong et al. 2003a) and/or assays should be required.
Validation is an important step in developing a useful QSAR model. There are two common validation methods— cross-validation and external validation (Tong et al. 2003c). For most classification methods, descriptor selection is normally executed prior to model training. Without preselection of the descriptor variables, the computational expense of cross-validation could be prohibitive. However, preselection of descriptors also constitutes a bias, suggesting that cross-validation may overestimate a model’s true predictive accuracy for unknown chemicals. For such cases preselecting an external test set not used in training becomes critical to estimating predictive accuracy. But, setting aside an external test set detrimentally reduces the size of the available training set, resulting in the loss of data that would likely improve the model. Ideally, the external test set would be rationally selected to represent the chemicals to which the model would be applied. In reality, however, because of the difficulty of such a task, we are unaware of any model development and test set selection in the literature that incorporates a systematic selection of a representative test set.
Bias in descriptor selection is not a factor in DF, where in each step of the cross-validation a new set of descriptors is selected that forms the best forest to represent each random spilt between training and testing data. The full integration of variable selection with forest construction means that the cross-validation accuracy is more likely to represent the true predictivity. Of course, a prediction test on external data is always desirable because it is a real-world application of the model, but very rarely is sufficient data available to warrant complete exclusion of some data from the training data. In a sense, cross-validation closely resembles the conduct of multiple tests on external data. Thus, we choose a rigorous and extensive cross-validation method to validate the models’ predictivities in this study, which is able to assess many possible partitions of the training and test sets and then can provide an unbiased and objective means for assessing a model’s quality.
A large number of QSAR models for ER binding are reported in the literature (Bradbury et al. 1996; Sadler et al. 1998; Waller et al. 1996; Wiese et al. 1997; Zheng and Tropsha 2000), including our own (Tong et al. 1997a, 1997b, 1998; Xing et al. 1999). Although these models yield good statistical results, none explicitly address and assess the confidence in predicting unknown chemicals. We demonstrated in this study that there could be more than a 22% difference in prediction accuracy for the chemicals with high confidence compared with those with low confidence. Thus, for practical applications, having prediction confidence together with the actual predictions greatly extends the usefulness of QSAR and classification models. In regulatory application, the justification for using such models may very well depend on having measures of confidence in the predictions.
Four types of uncertainty are generally recognized as affecting the prediction confidence of a QSAR model (Tong et al. 2003c), and all generally are dependent on either the nature of the data set or the choice of the statistical algorithm. First, predictions from any model are intrinsically no better than the experimental data employed to develop the model. Any limitations of the assay used to generate the training data equally extends to the model’s predictions. Second, commonly employed statistical methods vary in their abilities to appropriately capture the functional relationship of structural descriptors and activity. Third, for classification models specifically, class assignment is sensitive to a defined cutoff value to distinguish active from inactive. As the cutoff value is lowered, it is likely that the error will increase, even for a well-designed and well-executed assay. The increased experimental error in close proximity to the cutoff value will be transferred to the classification model, which in turn will increase false prediction rate for chemicals with activity in this region. Fourth, a chemical can be represented by different types of descriptors. We often find that, even for a simple mechanism such as ER-binding, some descriptors may well represent binding dependencies for one structural class, whereas other features will better represent binding dependencies for a different structural class. In such cases, regardless of how rigorously the validation procedure is employed, the model may give incorrect predictions for some chemicals, as the entire chemistry space of active chemicals is unknown. These four types of uncertainty determine the applicability domain of a QSAR model, and adequate assessment of this domain that bounds and guides the model’s usage, especially in regulatory application, is paramount. The assessment procedure proposed in this study should be equally applicable to other QSAR methods.
Figure 1 Comparison of structural diversity of ER232 and ER1092 in a chemistry space defined by three principal components of over 270 2D structural descriptors.
Figure 2 Schematic illustration for defining the training domain of a tree. For an unknown chemical predicted by the tree, its classification is determined by the terminal node (dark circle) to which it belongs. There are three descriptors used in the path (bold line) from the root to the terminal node and the range of these three descriptors across all chemicals in the training set determines the training domain.
Figure 3 Prediction accuracy versus confidence level. Data were calculated from 2,000 runs of 10-fold cross-validation for (A) ER232 and (B ) ER1092.
Figure 4 Prediction accuracy in different domains of extrapolation for ER232 and ER1092 from 2,000 runs of 10-fold cross-validation.
Table 1 Statistics of the forest models based on ER232 and ER1092.
ER232 ER1092
Number of chemicals 232 1092
Number (%) of misclassifications 5 (2.16%) 50 (4.58%)
Number of trees combined 6 4
Number of descriptors used 79 138
Accuracy 96.6% 95.4%
Specificity 96.0% 91.0%
Sensitivity 96.9% 97.6%
Table 2 The HC and LC predictions from 2,000 runs of 10-fold cross-validation for ER232 and ER1092.
ER232
ER1092
Confidence regions Accuracy (%) Percentage of chemicals Accuracy (%) Percentage of chemicals
HC 86.6 79.2 86.3 69.9
LC 63.8 20.8 64.7 30.1
All 81.9 100 79.7 100
Abbreviations: HC , high confidence; LC, low confidence.
Table 3 Prediction accuracy in different regions of confidence and extrapolation derived from 2,000 runs of 10-fold cross-validation for ER232 and ER1092.
HC region
LC region
Data set Extrapolation [(d) %] Accuracy (%) No. of predictions Accuracy (%) No. of predictions
ER232 0 87.8 349,595 64.7 83,393
0–10 79.7 10,442 55.8 6,216
10–20 61.1 1,325 50.5 2,651
20–30 65.3 853 63.9 1,086
> 30 39.7 5,614 65.9 2,825
ER1092 0 86.4 1,511,180 64.4 645,177
0–10 89.4 6,896 61.5 5,135
10–20 88.5 3,914 68.1 3,453
20–30 96.8 1,209 75.3 959
> 30 48.9 3,560 54.1 2,517
==== Refs
References
Benigni R Richard AM 1998 Quantitative structure-based modeling applied to characterization and prediction of chemical toxicity Methods 14 264 276 9571083
Blair R Fang H Branham WS Hass B Dial SL Moland CL 2000 Estrogen receptor relative binding affinities of 188 natural and xenochemicals: structural diversity of ligands Toxicol Sci 54 138 153 10746941
Bradbury S Mekenyan O Ankley G 1996 Quantitative structure-activity relationships for polychlorinated hydroxybiphenyl estrogen receptor binding affinity—an assessment of conformer flexibility Environ Toxicol Chem 15 1945 1954
Bradbury SP 1994 Predicting modes of toxic action from chemical structure: an overview SAR QSAR Environ Res 2 89 104 8790641
Branham WS Dial SL Moland CL Hass BS Blair RM Fang H 2002 Phytoestrogens and mycoestrogens bind to the rat uterine estrogen receptor J Nutr 132 658 664 11925457
Breiman L Friedman J Olshen R Stone C Steinberg D Colla P 1995. CART: Classification and Regression Trees. Stanford, CA:Salford System.
Clark LA Pregibon D 1997. Tree-based models. In: Modern Applied Statistics with S-Plus (Venables WN, Ripley BD, eds). New York:Chambers and Hasties, 413–430.
Eriksson L Jaworska J Worth A Cronin M McDowell RM Gramatica P 2003 Methods for reliability and uncertainty assessment and for applicability evaluations of classification- and regression-based QSARs Environ Health Perspect 111 1361 1375 12896860
Fang H Tong W Branham WS Moland CL Dial SL Hong H 2003a Study of 202 natural, synthetic, and environmental chemicals for binding to the androgen receptor Chem Res Toxicol 16 1338 1358 14565775
Fang H Tong W Sheehan D 2003b QSARs in receptor-mediated effects: the nuclear receptor superfamily J Mol Struct (THEOCHEM) 622 113 125
Food Quality Protection Act of 1996 1996. Public Law 104–170.
Hansch C Hoekman D Leo A Zhang L Li P 1995a The expanding role of quantitative structure-activity relationships (QSAR) in toxicology Toxicol Lett 79 45 53 7570673
Hansch C Telzer BR Zhang L 1995b Comparative QSAR in toxicology: examples from teratology and cancer chemotherapy of aniline mustards Crit Rev Toxicol 25 67 89 7734060
Hileman B 1997 Hormone disrupter research expands Chem Eng News 75 24 25
Hong H Tong W Fang H Shi L Xie Q Wu J 2002 Prediction of estrogen receptor binding for 58,000 chemicals using an integrated system of a tree-based model with structural alerts Environ Health Perspect 110 29 36 11781162
Hopfinger AJ Tokarski JS 1997. Practical applications of computer-aided drug design. In: Practical Applications of Computer-Aided Design (Charifson PS, ed). New York:Marcel Dekker, 105–164.
Johnson M Maggiora GM 1990. Concepts and Applications of Molecular Similarity. New York:Wiley.
Kubinyi H Folkers G Martin YC 1998 3D QSAR in drug design—recent advances Perspect Drug Disc Design 12 R5 R7
Lipinski CA Lombardo F Dominy BW Feeney PJ 1997 Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings Adv Drug Delivery Rev 23 3 25
Russom CL Bradbury SP Carlson AR 1995 Use of knowledge bases and QSARs to estimate the relative ecological risk of agrichemicals: a problem formulation exercise SAR QSAR Environ Res 4 83 95 8765904
Sadler BR Cho SJ Ishaq KS Chae K Korach KS 1998 Three-dimensional quantitative structure-activity relationship study of nonsteroidal estrogen receptor ligands using the comparative molecular field analysis/cross-validated r 2 -guided region selection approach J Med Chem 41 2261 2267 9632359
Safe Drinking Water Act of 1996 1996. Public Law 104–182.
Schultz TW Seward JR 2000 Health-effects related structure-toxicity relationships: a paradigm for the first decade of the new millennium Sci Total Environ 249 73 84 10813448
Shi L Tong W Fang H Xie Q Hong H Perkins R 2002 An integrated “4-phase” approach for setting endocrine disruption screening priorities—Phase I and II predictions of estrogen receptor binding affinity SAR QSAR Environ Res 13 69 88 12074393
Shi LM Fang H Tong W Wu J Perkins R Blair RM 2001 QSAR models using a large diverse set of estrogens J Chem Inf Comput Sci 41 186 195 11206373
Tong W Fang H Hong H Xie Q Perkins R Anson JF 2003a Regulatory application of SAR/QSAR for priority setting of endocrine disruptors—a perspective Pure and Appl Chem 75 2375 2388
Tong W Fang H Hong H Xie Q Perkins R Sheehan D 2004. Receptor-mediated toxicity: QSARs for oestrogen receptor binding and priority setting of potential oestrogenic endocrine disruptors. In: Predicting Chemical Toxicity and Fate (Cronin MTD, Livingstone D, eds). Boca Raton, FL:CRC Press, 285–314.
Tong W Hong H Fang H Xie Q Perkins R 2003b Decision forest: Combining the predictions of multiple independent decision tree models J Chem Inf Comput Sci 43 525 531 12653517
Tong W Lowis DR Perkins R Chen Y Welsh WJ Goddette DW 1998 Evaluation of quantitative structure-activity relationship methods for large-scale prediction of chemicals binding to the estrogen receptor J Chem Inf Comput Sci 38 669 677 9722424
Tong W Perkins R Fang H Hong H Xie Q Branham SW 2002 Development of quantitative structure-activity relationships (QSARs) and their use for priority setting in the testing strategy of endocrine disruptors Regul Res Perspect 1 1 16
Tong W Perkins R Strelitz R Collantes ER Keenan S Welsh WJ 1997a Quantitative structure-activity relationships (QSARs) for estrogen binding to the estrogen receptor: predictions across species Environ Health Perspect 105 1116 1124 9353176
Tong W Perkins R Xing L Welsh WJ Sheehan DM 1997b QSAR models for binding of estrogenic compounds to estrogen receptor alpha and beta subtypes Endocrin 138 4022 4025
Tong W Welsh WJ Shi L Fang H Perkins R 2003c Structure-activity relationship approaches and applications Environ Toxicol Chem 22 1680 1695 12924570
Waller CL Oprea TI Chae K Park HK Korach KS Laws SC 1996 Ligand-based identification of environmental estrogens Chem Res Toxicol 9 1240 1248 8951225
Wiese TE Polin LA Palomino E Brooks SC 1997 Induction of the estrogen specific mitogenic response of MCF-7 cells by selected analogues of estradiol-17 beta: a 3D QSAR study J Med Chem 40 3659 3669 9357533
Xing L Welsh WJ Tong W Perkins R Sheehan DM 1999 Comparison of estrogen receptor alpha and beta subtypes based on comparative molecular field analysis (CoMFA) SAR QSAR Environ Res 10 215 237 10491851
Zheng W Tropsha A 2000 A novel variable selection QSAR approach based on the k-nearest neighbor principle J Chem Inf Comput Sci 40 185 194 10661566
|
15345371
|
PMC1277118
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug 16; 112(12):1249-1254
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7125
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7152ehp0112-00125515345372ToxicogenomicsArticlesBiokinetics and Subchronic Toxic Effects of Oral Arsenite, Arsenate, Monomethylarsonic Acid, and Dimethylarsinic Acid in v-Ha-ras Transgenic (Tg.AC) Mice Xie Yaxiong 1Trouba Kevin J. 2Liu Jie 1Waalkes Michael P. 1Germolec Dori R. 21Inorganic Carcinogenesis Section, Laboratory of Comparative Carcinogenesis, National Cancer Institute at the National Institute of Environmental Health Sciences and2Environmental Immunology, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USAAddress correspondence to D.R. Germolec, Environmental Immunology, National Institute of Environmental Health Sciences, P.O. Box 12233, Mail Drop C1-03, Research Triangle Park, NC 27709 USA. Telephone: (919) 541-3230. Fax: (919) 541-0870. E-mail:
[email protected] authors thank J. Pi, L. Benbrahim-Tallaa, and L. Keefer for their critical reviews and help during the preparation of this manuscript.
The authors declare they have no competing financial interests.
8 2004 18 6 2004 112 12 1255 1263 2 4 2004 17 6 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Previous research demonstrated that 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment increased the number of skin papillomas in v-Ha-ras transgenic (Tg.AC) mice that had received sodium arsenite [(As(III)] in drinking water, indicating that this model is useful for studying the toxic effects of arsenic in vivo. Because the liver is a known target of arsenic, we examined the pathophysiologic and molecular effects of inorganic and organic arsenical exposure on Tg.AC mouse liver in this study. Tg.AC mice were provided drinking water containing As(III), sodium arsenate [As(V)], monomethylarsonic acid [(MMA(V)], and 1,000 ppm dimethylarsinic acid [DMA(V)] at dosages of 150, 200, 1,500, or 1,000 ppm as arsenic, respectively, for 17 weeks. Control mice received unaltered water. Four weeks after initiation of arsenic treatment, TPA at a dose of 1.25 μg/200 μL acetone was applied twice a week for 2 weeks to the shaved dorsal skin of all mice, including the controls not receiving arsenic. In some cases arsenic exposure reduced body weight gain and caused mortality (including moribundity). Arsenical exposure resulted in a dose-dependent accumulation of arsenic in the liver that was unexpectedly independent of chemical species and produced hepatic global DNA hypomethylation. cDNA microarray and reverse transcriptase–polymerase chain reaction analysis revealed that all arsenicals altered the expression of numerous genes associated with toxicity and cancer. However, organic arsenicals [MMA(V) and DMA(V)] induced a pattern of gene expression dissimilar to that of inorganic arsenicals. In summary, subchronic exposure of Tg.AC mice to inorganic or organic arsenicals resulted in toxic manifestations, hepatic arsenic accumulation, global DNA hypomethylation, and numerous gene expression changes. These effects may play a role in arsenic-induced hepatotoxicity and carcinogenesis and may be of particular toxicologic relevance.
arsenicals (arsenic forms)gene expressionmouse liversubchronic toxicitytoxicokinetics
==== Body
Arsenic is an important environmental toxicant and carcinogen [International Agency for Research on Cancer (IARC) 1987; National Research Council (NRC) 1999]. Chronic exposure to arsenic via drinking water is a major health concern throughout the world (Gebel 2000; NRC 1999). The carcinogenic effects of environmental arsenic exposure in human populations are well documented (IARC 1987; NRC 1999), and exposure can lead to tumors in and toxicity of the skin, lung, urinary bladder, liver, and other sites.
The adverse effects of arsenic are dependent, in part, on its chemical form and metabolism (Aposhian 1997; Vahter 2002). Humans are exposed primarily to trivalent [arsenite, As(III)] and pentavalent [arsenate, As(V)] inorganic arsenicals present in the environment, as well as to organic arsenic [e.g., dimethylarsinic acid, DMA(V)] in some situations (Kenyon and Hughes 2001; Shen et al. 2003b). In mammals, As(V) is first reduced to As(III), whereas As(III), produced by this reduction or from direct ingestion, is methylated primarily to pentavalent organic arsenicals including monomethylarsonic acid [MMA(V)] and DMA(V)]. MMA and DMA are the predominant metabolites of inorganic arsenic (Vahter 2002), although DMA may be further methylated to trimethylarsine oxide (TMAO) (Hughes 2002; Yoshida et al. 1998). The forms of arsenic to which humans are exposed, either directly or via metabolism, further complicate the elucidation of their toxic and carcinogenic mechanisms of action. Previously, inorganic arsenicals were thought to be more acutely toxic than organic species, as the methylation of inorganic arsenic was proposed to be a detoxification process. However, recent studies indicate that trivalent organic arsenicals [e.g., MMA(III) and DMA(III)] that are metabolic products of inorganic arsenic can be more toxic than the parent compound (Petrick et al. 2001; Styblo et al. 2000). Furthermore, DMA can act as a tumor promoter at various sites and as a complete carcinogen for the urinary bladder in rats (Salim et al. 2003; Wei et al. 2002; Yamamoto et al. 1997). MMA produces preneoplastic changes in liver and urinary bladder but does not produce overt neoplasia (Shen et al. 2003a), whereas TMAO can induce hepatocellular adenomas (Shen et al. 2003b). Therefore, it is important to compare and evaluate the toxicity of As(III), As(V), MMA(V), and DMA(V) under similar experimental conditions.
Recent studies demonstrated that arsenic acts as a co-promoter with 12-O-tetradecanoylphorbol-13-acetate (TPA) because together they enhance skin tumor development in transgenic (Tg.AC) mice, which overexpress the v-Ha-ras oncogene (Germolec et al. 1997, 1998; Trouba et al. 2003). Because hepatic metabolism in Tg.AC mice is not compromised by over-expression of the v-Ha-ras oncogene (Sanders et al. 2001), we hypothesized that organic and inorganic arsenicals produce similar yet distinct changes in Tg.AC liver gene expression that may be predictive of hepatotoxicity. The latter is important because the liver is an important target organ of arsenic toxicity in animals (Waalkes et al. 2000b) and humans (Lu et al. 2001). The liver is also a major target organ of arsenic carcinogenicity after in utero exposure in mice (Waalkes et al. 2003, 2004b) and in humans exposed to environmental arsenic (Centeno et al. 2002; Zhou et al. 2002). To address the above hypothesis, we examined the effects of subchronic inorganic and organic arsenical exposure on the Tg.AC mouse liver. Our results indicate that in Tg.AC mice, a) hepatic arsenic [e.g., As(III), As(V), MMA(V), and DMA(V))] accumulation, based on biokinetic analyses, was dose dependent; b) global DNA hypomethylation occurred after exposure to As(III) As(V), MMA(V), and DMA(V); c) pathological changes were present in the liver after exposure to As(III), MMA(V), and DMA(V); and d) arsenic-induced gene expression changes, determined using cDNA microarray and real-time reverse transcriptase–polymerase chain reaction (RT–PCR) analysis, occurred in the liver of animals treated with As(III), As(V), MMA(V), and DMA(V).
Materials and Methods
Chemicals
As(III) and As(V) were purchased from Aldrich Chemical Co. (Milwaukee, WI) and Fluka Chemical Corp. (Milwaukee, WI), respectively. MMA(V) was obtained from AccuStandard, Inc. (New Haven, CT). DMA(V) and TPA were purchased from Sigma Chemical Co. (St. Louis, MO). Customer-designed cDNA micro-arrays (600 genes) were purchased from BD Biosciences Clontech, Inc. (Palo Alto, CA). [α-32P]-deoxyadenosine 5′-triphosphate was purchased from PerkinElmer, Inc. (Boston, MA), and 3H-labeled S-adenosylmethionine ([3H]-SAM) was from Amersham (Arlington Heights, IL).
Animal Treatment
All animals were handled and treated in compliance with the Guide for the Care and Use of Laboratory Animals (NRC 1996). Female, homozygous Tg.AC mice containing the fetal zeta-globin promoter fused to the v-Ha-ras structural gene (with mutations at codons 12 and 59) and linked to a simian virus 40 polyadenylylation/ splice sequence were obtained from Taconic Farms (Germantown, NY) (Leder et al. 1990). Mice were maintained in an animal facility at a temperature of 20–22°C, a relative humidity of 50%, and a 12-hr light/dark cycle. Mice were randomly assigned to five groups (n = 15 in each group) and were provided unaltered drinking water (control) and drinking water containing As(III) (150 ppm as arsenic), As(V) (200 ppm as arsenic), MMA(V) (1,500 ppm as arsenic), and DMA(V) (1,000 ppm as arsenic), respectively, for 17 weeks. The doses of arsenicals used were based on our previous studies (Germolec et al. 1997, 1998). Multiple doses of each arsenical were originally used to examine papilloma development. However, to detect gene expression changes in the liver that may be related to arsenic hepatotoxicity and hepatocellular carcinogenesis, animals treated with the maximal dose of each arsenical were selected for analysis.
Four weeks after initiation of arsenic treatment, TPA at a dose of 1.25 μg/200 μL acetone was applied twice weekly for 2 weeks to the shaved dorsal skin of all mice, including the mice not receiving arsenic (control). At 17 weeks the mice were sacrificed by CO2 asphyxiation and necropsied. Liver tissue was excised and stored at −70°C until analysis or fixed for histology as described below.
During the exposure to arsenic, mortality, moribundity, clinical symptoms, body weight, and water intake of the mice were monitored. All mice, including those found deceased or sacrificed as moribund, underwent complete necropsy.
Pathological Examination
Liver samples were fixed with neutral-buffered formalin, processed by standard procedures, embedded in paraffin, sectioned, and stained with hematoxylin and eosin for light microscopy examination. All pathological assessments were performed in a blind fashion.
Hepatic Arsenic Levels
A portion of the frozen liver (120–150 mg) was digested in nitric acid. Total arsenic, which would include inorganic and organic forms, was determined using graphic furnace atomic absorption spectrometry (Perkin-Elmer AAnalyst100; PerkinElmer, Inc., Norwalk, CT). Results were expressed as micrograms arsenic per gram wet weight liver, as reported in our recent publications (Liu et al. 2001a; Xie et al. 2004).
Global DNA Methylation Assay
Genomic DNA was extracted from liver tissue and purified using DNeasy Kits (Qiagen, Valencia, CA). Global DNA methylation status was assessed by methyl acceptance assay (Chen et al. 2004). Briefly, DNA (1 μg) was incubated at 37°C for 2 hr in a 30-μL mixture containing 1.25 μM (3 μCi) [3H]-SAM, 4 units CpG methylase (M. Sss I) (New England Biolabs, Inc., Beverly, MA), 10 mM DDT, Tris-EDTA buffer (100 mM Tris, 10 mM EDTA, pH 8.0), and 100 mM NaCl. The reaction was terminated on ice and transferred onto a Whatman DE81 filter (Whatman International Ltd., Maidstone, U.K.). The filter was washed on a vacuum filtration apparatus with 2 mL 0.5 M phosphate buffer (pH 7.0) 5 times, followed by a wash with 2 mL 70% ethanol and 2 mL absolute ethanol. After the filter was dried, the bound radioactivity was measured by scintillation (Beckman LS 6500 Scintillation Counter; Beckman Coulter, Inc., Fullerton, CA).
cDNA Microarray Analysis
Microarray analysis was performed as previously described (Xie et al. 2004). Briefly, total RNA was extracted from liver tissues with Trizol reagent and purified with RNeasy columns (Qiagen). Five micrograms pooled RNA (n = 5) was converted to [α-32P]-dATP–labeled cDNA probe with Atlas specific cDNA synthesis primers (BD Biosciences Clontech Inc.). The probe was purified with a NucleoSpin column (BD Biosciences Clontech), denatured at 100°C for 2–3 min, and hybridized to the membrane in triplicate with Expresshyb buffer (BD Biosciences Clontech) at 68°C overnight. The membranes were washed at 68°C four times (30 min each) in 2 × sodium chloride/sodium citrate (SSC)/1% sodium dodecyl sulfate (SDS), twice in 0.1 × SSC/0.5% SDS, and exposed to a phosphoimage screen. Images were acquired by PhosphorImager Scanner (Model Storm 860; Molecular Dynamics, Sunnyvale, CA) and analyzed densito-metrically using AtlasImage software (version 2.01; Clontech).
Real-time RT–PCR Analysis
Total RNA was reverse transcribed with MMLV reverse transcriptase and oligodT primers (PerkinElmer Inc.). The PCR primers were designed with Primer Express software and the SYBR Green DNA PCR kit (Applied Biosystems, Foster City, CA) was used for real-time RT–PCR analysis. Differences in gene expression between groups were calculated using cycle time (Ct) values, which were normalized against β-actin and expressed as relative increases/ decreases, setting control as 1.0. Assuming that the Ct value is reflective of the initial template amount (copy number) and that there is 100% efficiency, a difference of one cycle is equivalent to a 2-fold difference in initial copy number (Walker 2001).
Statistics
Data are expressed as mean ± SEM or as incidence (for mortality). For comparisons of gene expression between two groups, the Student t test was used. For comparisons among three or more groups, data were analyzed using a one-way analysis of variance, followed by Duncan’s multiple range test. The p-value was calculated by Fisher’s exact test for incidence data. The level of significance was set at p < 0.05 in all cases. Two-dimensional hierarchical cluster analysis of microarray data (hybrid intensity ratios to control values) was performed. The results from clustered analysis were examined by interactive graphical analysis using TreeView software (http://rana.lbl.gov/EisenSoftware.htm).
Results
Clinical Symptoms
During the 17 weeks of arsenical exposure, several arsenic-treated mice were found deceased or were euthanized because of moribundity. Exposure to As(III) (150 ppm) and DMA(V) (1,000 ppm) resulted in 20% mortality, and exposure to MMA(V) (1,500 ppm) resulted in 40% mortality (Table 1). MMA(V) exposure alone produced significant toxicity when compared with control. In general, the body weight in arsenic-treated groups was lower than that in control groups. At the end of arsenic exposure (17 weeks), body weight was decreased by approximately 15, 8, 10, and 8% in mice treated with As(III) (150 ppm), As(V) (200 ppm), MMA(V) (1,500 ppm), or DMA(V) (1,000 ppm), respectively (Figure 1). Our findings suggest that exposure of Tg.AC mice to these arsenicals produced mild to moderate [for the MMA(V) group] toxicity.
Pathology
The treatment of As(III) plus TPA did not induce liver tumor formation in Tg.AC mice treated with arsenicals for 17 weeks (unpublished data). However, morphologic changes including inflammation, foci of apoptosis and necrosis, and hepatocellular degeneration were observed in arsenic-treated mice (Figure 2). Foci of apoptosis and necrosis were observed in animals treated with As(III) (150 ppm); however, no apparent histologic alterations were present in animals that received As(V) (200 ppm). MMA(V) (1,500 ppm) produced inflammatory cell infiltration, degeneration, and swelling; DMA(V) (1,000 ppm) produced foci of inflammation and hepatocellular degeneration. These findings indicate that subchronic arsenical exposure produces pathological alterations in the liver.
Hepatic Arsenic Content
Although not detectable in livers of controls, arsenic was found in the livers of all treatment groups (Figure 3). Particularly high levels of arsenic were present in the livers of the MMA-treated group (Figure 3A). When hepatic arsenic content was plotted against arsenical dose, a strong linear correlation was observed (r = 0.98) (Figure 3B), suggesting that subchronic arsenic exposure results in arsenic accumulation in the liver that is dose dependent.
Global DNA Methylation Status
Global DNA methylation was assessed by methyl acceptance assay (Figure 4). This assay uses a bacterial DNA methyltransferase that indiscriminately methylates all unmethylated cytosines using [3H]-SAM. Thus, higher [3H]-SAM incorporation corresponds to a lower degree of methylation (i.e., hypomethylation) of cellular DNA. The amount of unmethylated DNA from all the arsenical-treated groups was significantly higher (p < 0.05) than control, indicating that DNA hypomethylation occurs in the Tg.AC mouse liver after subchronic exposure to arsenic, regardless of the chemical form. When this is correlated with actual arsenic dose, As(III) is the most potent hypomethylating agent; MMA(V) is the least.
Genomic Analysis by cDNA Microarray
Among the 600 genes examined via microarray analysis, 70 displayed increased or decreased expression after subchronic arsenic exposure. The hybrid intensity (ratio to control value) for these 70 genes was calculated for comparison then subjected to cluster analysis to compare alterations in gene expression patterns related to the type of arsenical exposure. TreeView revealed both similar and dissimilar changes in gene expression patterns among the four arsenicals (Figure 5).
The most significant arsenic-induced changes in gene expression are listed in Table 2. Genes associated with glutathione S-transferase (GST) function/metabolism, stress, apoptosis, cell proliferation, and early neoplasia are thought to be related to arsenic toxicity (Liu et al. 2004; Trouba et al. 2002, 2003; Xie et al. 2004) and thus are included for comparison. For example, all arsenicals produced increases in GSTs (alpha, mu, pi, and theta) and fibroblast growth factor 2, a gene related to cell proliferation. A significant increase in the expression of insulin-like growth factor binding protein 1 (IGFBP-1) also was found in MMA-treated mice. In general, all of the arsenicals produced similar effects (i.e., increase/decrease) on gene expression; however, the degree of change was different in some cases.
Real-Time RT–PCR Analysis
Real-time RT–PCR analysis was performed for selected genes in each cluster. Figure 6 shows data for some of the genes of interest. GST-π, early growth response protein 1 (EGR-1), heme oxygenase 1 (HO-1), c-myc, and α-fetoprotein gene expression was enhanced after arsenical exposure. Generally, real-time RT–PCR analysis confirmed our microarray results.
Discussion
This study demonstrated that subchronic exposure of transgenic (Tg.AC) mice to both inorganic and organic arsenicals through drinking water produced various effects on the liver, a major target organ of arsenic toxicity and carcinogenesis (Centeno et al. 2002; NRC 1999; Waalkes et al. 2003, 2004b). Arsenic-induced toxicity was evidenced by an increase in moribundity and death, a depression in body weight, hepatic pathological changes, and significant changes in gene expression.
An original goal of our research was to examine the effects of inorganic and organic arsenic on TPA-promoted skin papilloma development in Tg.AC mice. Although TPA was administered to all mice (including controls that received no arsenic), the effects of this skin tumor promoter were not deemed critical to our analyses of liver pathology, DNA methylation, and gene expression. Interestingly, topical application of TPA in some experimental models has systemic effects; we recently found that it promoted liver tumors initiated by transplacental arsenic exposure in female mice (Waalkes et al. 2004b). In this study epidermal TPA treatment resulted in no mortality and did not affect hepatic pathology, indicating that the biological end points/ changes measured are most likely dependent on arsenical treatment alone.
Because the liver is a major target organ of arsenic toxicity and carcinogenesis (Waalkes et al. 2000b, 2003, 2004b), we examined gene expression as well as pathological changes in the livers of Tg.AC mice to further explore the usefulness of this system as an in vivo model of arsenic carcinogenesis and toxicity. To detect gene expression changes that may be related to arsenic toxicity, animals treated with the maximal dose of each arsenical were selected for analysis. Generally, 150 ppm As(III) produced more toxicity and more dramatic changes in gene expression than 200 ppm As(V). Organic arsenicals at doses [1,500 and 1,000 ppm as arsenic for MMA(V) and DMA(V), respectively] 5-to 10-fold higher produced toxic effects comparable to those produced by As(III). Although rats are tolerant to 200 ppm MMA(V) in drinking water for 104 weeks (Shen et al. 2003a), the mice in our study did not tolerate MMA(V) at 1,500 ppm, as 40% mortality (i.e., moribundity and death) occurred in these mice over the 17-week exposure period. The dose of DMA(V) in this study was also higher than the doses (50 and 200 ppm) used to induce urinary bladder tumors in rats (Wei et al. 2002) and also exceeded the maximum tolerated dose, as it produced 20% mortality.
In our study, promoted and nonpromoted, arsenic-treated Tg.AC mice did not display direct evidence of liver tumor formation. However, preneoplastic lesions (e.g., cell proliferation) occur in the liver after chronic oral arsenic exposures in several strains of mice (Chen et al. 2004; Shen et al. 2003a; Waalkes et al. 2000b) and were also observed in the liver of Tg.AC mice exposed to arsenic in this study. Exposure to arsenic in the drinking water resulted in a dose-dependent accumulation of arsenic in the liver that was independent of chemical form. The highest hepatic content, which was observed in the high-dose (1,500 ppm) MMA(V) group, might contribute to the high degree of mortality (40%) in this group. The hepatic arsenic contents in the Tg.AC mice receiving 150 ppm As(III) and 200 ppm As(V) in this study were 1.2 and 2.0 μg/g tissue, respectively. This was less than the arsenic content in the skin (8.3 μg/g tissue) and much less than that in the hair (170.2 μg/g tissue) of Tg.AC mice exposed to 200 ppm As(III) in the drinking water for 14 weeks in our previous study (Germolec et al. 1998), indicating that arsenic accumulation in the liver is lower than that in the hair or skin. This may be because liver is the major target organ for arsenic metabolism, and arsenic elimination generally occurs through the bile (Gregus et al. 2000) or urine.
DNA hypomethylation occurs after chronic arsenic exposure in cells (Zhao et al. 1997) and also in intact animals (Chen et al. 2004; Okoji et al. 2002). In the present study, all arsenicals produced significant DNA hypomethylation in the liver, regardless of dose. Although the doses of MMA(V) (1,500 ppm) and DMA(V) (1,000 ppm) used in our study were much higher than those of As(III) (150 ppm) and As(V) (200 ppm), MMA(V) and DMA(V) induced less hypomethylation of hepatic DNA than As(III) and As(V). This suggests that inorganic arsenicals are more potent stimulators of DNA hypomethylation compared with MMA(V) and DMA(V). It should be noted that global DNA hypomethylation could co-exist with regional or individual gene hypermethylation, as arsenic-induced p53 hypermethylation has been reported (Mass and Wang 1997). In our recent study, we proposed that arsenic-induced hypomethylation of the estrogen receptor-α gene plays an important role in hepatocellular proliferation (Chen et al. 2004; Waalkes et al. 2004a). Efforts are currently being undertaken to examine the methylation status of individual genes after arsenic exposure.
DNA hypomethylation is an important mechanism involved in aberrant gene expression and carcinogenesis (Baylin et al. 1998; Goodman and Watson 2002). In particular, it is thought that aberrant DNA methylation is central to the development of liver cancers (Goodman and Watson 2002) and is an epigenetic mechanism that underlines the aberrant expression of genes involved in mouse liver carcinogenesis (Counts et al. 1997). In the present study, As(III), As(V), MMA(V), and DMA(V) produced variable gene expression changes, accounting for approximately 10% of genes on the array. We focused primarily on a few categories, for example, glutathione (GSH)-, apoptosis-, and cell proliferation–related genes, and genes important for tumor development, as previous studies have shown these to be related to aberrant cell growth and neoplasia.
Glutathione systems play important roles in arsenic toxicity and carcinogenesis (NRC 1999; Trouba et al. 2002, 2003; Xie et al. 2004). In the present study, the expression of GST-μ, GST-π, GST-α, and GST-τ was increased by all arsenicals, although to a variable extent. GSTs are a group of enzymes catalyzing the conjugation and oxidation of GSH with arsenic (Xie et al. 2004). An increase in GST expression/activity (particularly GST-π) has been reported to play an important role in cellular efflux of arsenic–GSH conjugates and to be a mechanism of arsenic tolerance (Brambila et al. 2002; Liu et al. 2001a; Wang et al. 1996). Increases in GST-π positive foci have been proposed to be a hepatic preneoplastic biomarker in chronic arsenic-exposed populations (Nishikawa et al. 2002; Shen et al. 2003a). Changes in GST activity in humans also are associated with altered arsenic metabolism (Chiou et al. 1997; Marnell et al. 2003), and GST polymorphisms are thought to be a susceptibility factor for arsenic toxicity in humans (Marnell et al. 2003). Together, these data indicate that increases in GST gene expression and/or function are consistent events associated with arsenic carcinogenicity and toxicity.
Oxidative stress is proposed to play an important role in arsenic toxicity and carcinogenesis (Kitchin and Ahmad 2003; Liu et al. 2001b; NRC 1999; Trouba et al. 2002; Xie et al. 2004). In addition to GSTs, other biomarkers for arsenic-induced oxidative stress such as HO-1 (Del Razo et al. 2001; Liu et al. 2001b), EGR-1 (Liu et al. 2004; Simeonova et al. 2000), DT- diaphorase (Pi et al. 2003), and cytochrome P450 3A25 (Liu et al. 2001b) were all increased in Tg.AC mice after exposure to arsenicals. Evidence is accumulating regarding the ability of arsenicals to produce reactive oxygen species and free radicals as measured using electron spin response. This includes inorganic As(III) and As(V) (Barchowsky et al. 1999) and organic MMA(III), DMA(III) (Nesnow et al. 2002), and DMA(V) (Yamanaka et al. 1990). Our data (e.g., gene expression changes and pathology) lend further evidence for the presence of oxidative stress during subchronic exposure to arsenicals.
Arsenic induces apoptosis involved in its mechanism of acute toxicity (NRC 1999). However, after chronic arsenic exposure and the induction of malignant transformation, the development of apoptosis resistance occurs (Brambila et al. 2002; Qu et al. 2002) and is associated with the downregulation of apoptosis-related genes (Brambila et al. 2002; Chen et al. 2001a). In the present study, arsenical exposure resulted in downregulation of apoptosis-associated genes such as FasL, tumor necrosis factor receptor–associated factor 3, Bad, and granzyme A, and also increased the expression of cell proliferation–related genes including c-myc, proliferating cell nuclear antigen, and fibroblast growth factor 2. These data are interesting in light of evidence that apoptosis tolerance and cell proliferation are important mechanisms involved in chemical carcinogenesis (Waalkes et al. 2000a), including arsenic. Apoptosis tolerance also is accompanied by cell proliferation, as seen in arsenic-transformed cells (Brambila et al. 2002; Chen et al. 2001b; Qu et al. 2002) in chronic arsenic-exposed animals (Chen et al. 2004; Xie et al. 2004), and in liver tumor and nontumor tissues from mice exposed to arsenic in utero (Liu et al. 2004; Waalkes et al. 2003). Thus, the depression of apoptosis genes and the overexpression of cell proliferation genes could be important in arsenic toxicity and carcinogenesis.
Liver is a major target of arsenic carcinogenesis in transplacentally exposed animal models (Waalkes et al., 2003) and in arsenic-exposed humans (Centeno et al. 2002; Chen et al. 1997; Zhou et al. 2002). The expression of α-fetoprotein (AFP), a biomarker for hepatocellular carcinogenesis, was increased in transplacental arsenic-induced hepatocellular carcinoma (HCC) and tumor-surrounding tissues (Liu et al. 2004). In the present study, all the arsenicals tested increased AFP expression up to 3-fold in MMA(V)-treated mice. The enhanced expression of AFP lends further support that preneoplastic alterations occur after subchronic arsenic exposure. Other notable alterations in gene expression were the overexpression of IGFBP-1 and suppression of insulin-like growth factor 1 (IGF-1). Chronic exposure to nongenotoxic chemicals such as oxazepam and Wyeth-14,643 increased the expression of IGFBP-1 in a time-dependent manner (Iida et al. 2003), and overexpression of IGFBP-1 was also seen in transplacental arsenic-induced HCC and tumor-surrounding tissues (Liu et al. 2004). Dysregulation of the IGF axis has been implicated in liver tumor formation and progression (Scharf et al. 2001). Thus, sub-chronic exposure to arsenicals can produce aberrant gene expression related to hepatocarcinogenesis, some of which were confirmed in the present study.
In summary, this study demonstrated that subchronic exposure to As(III), As(V), MMA(V), and DMA(V) in the drinking water resulted in variable toxic effects, accumulation of arsenic in the liver, hepatic global DNA hypomethylation, and alterations in gene expression in Tg.AC mice. These findings indicate that liver is a target organ of subchronic arsenical exposure in this model and support the idea that altered DNA methylation and its effects on gene expression may contribute in an epigenetic manner to arsenic carcinogenesis.
Figure 1 Change in animal body weight during exposure to arsenicals.
Figure 2 Hepatic pathological changes induced by arsenicals. (A) Control; (B) As(III), 150 ppm; (C) As(V), 200 ppm; (D) MMA(V), 1,500 ppm; (E) DMA(V), 1,000 ppm. Original magnification, ×200; bars = 50 μm.
Figure 3 Hepatic arsenic content expressed as micrograms per gram wet weight. ND, not detected. (A) Arsenical species; (B) dose relationship.
Figure 4 Methyl acceptance capacity of hepatic genomic DNA of mice exposed to arsenicals.
*Statistically significant (p < 0.05) compared with vehicle alone treatment (control).
Figure 5 Cluster analysis of cDNA microarray data of selected genes. Data are ratio of control values: [red] ratio > 1.0; [black] ratio = 1.0; [green] ratio < 1.0. Relative changes in gene expression compared with those of control are presented as increased (↑), decreased (↓), or no change (NC). Double arrows highlighted in gray indicate ratio ≥2. Gene names and accession numbers are from GenBank (http://www.ncbi.nlm.nih.gov/entrez/query.fci?db=nucleotide).
Figure 6 Real-time RT–PCR analysis of selected genes. Data are mean ± SEM (n = 5). Dosage (ppm as arsenic) of the arsenicals: As(III), 150 ppm; As(V), 200 ppm; MMA(V), 1,500 ppm; DMA(V), 1,000 ppm.
* Statistically significant (p < 0.05) compared with control.
Table 1 Mortality due to arsenical exposure.a
Arsenical No. Found dead Sacrificed moribund Total (%)b
Control 15 0 0 0 (0)
AS(III) (150 ppm) 15 1 2 3 (20)
As(V) (200 ppm) 15 0 0 0 (0)
MMA(V) (1,500 ppm) 15 3 3 6 (40)*
DMA(V) (1,000 ppm) 15 2 1 3 (20)
a Arsenicals (ppm as arsenic) were administered for 17 weeks in drinking water; animal health was monitored twice daily.
b Total dead or euthanized before end of the experiment.
* Statistically significant (p < 0.05) compared with control.
Table 2 Effect of arsenicals on selected gene expression.a
Intensity relative to that of control
Protein/gene Accession no.b Hybrid intensity of control As (III) (150 ppm) As (V) (200 ppm) MMA (V) (1,500 ppm) DMA (V) (1,000 ppm)
GST gene
GST-alpha J03958 5,306 2.38* 2.25* 1.23* 1.17
GST-mu U24428 3,824 1.52 1.83* 1.43 1.82*
GST-pi D30687 25,368 1.40* 1.25 0.96 1.40
GST-theta-1 X98055 3,892 1.70* 1.42 1.03 1.12
Stress-related genes
HO-1 M33203 13,480 1.19 1.31* 1.31* 1.26*
EGR-1 M20157 6,356 0.88 1.59* 1.76* 1.61*
DT diaphorase U12961 2,926 1.81* 1.79* 1.28 1.79*
Cytochrome P450 IIIA25 (CYP3A25) Y11995 19,041 1.22 1.21 1.48* 1.47*
Genes related to apoptosis and cell proliferation
FasL U06948 2,006 0.60 0.65 0.63 0.33*
TNF receptor–associated factor 3 U21050 3,452 0.81 0.82 0.58* 0.49*
Bad L37296 3,002 0.70* 0.67* 0.81 0.92
Granzyme A M13226 2,802 0.58* 0.53* 0.42* 0.90
Proliferating cell nuclear antigen X53068 2,768 1.33* 1.21 0.99 0.96
Fibroblast growth factor 2 M30644 1,582 2.00* 2.19* 2.16* 1.80*
c-myc protooncogene X00195 1,981 1.18 1.11 0.93 1.71
Tumor-related genes
Alpha fetoprotein V00743 1,631 1.72* 1.63 0.88 1.28
Insulin-like growth factor binding protein 1 X81579 7,605 0.94 1.20 2.93* 1.54*
Insulin-like growth factor 1 AF056187 1,216 0.65 0.95 0.23 0.47
a Data are based on the average value of arrays run in triplicate.
b Accession numbers are from GenBank (http://www.ncbi.nlm.nih.gov/entrez/query.fci?db=nucleotide).
* Original hybrid intensity is significantly different from that of control (p < 0.05).
==== Refs
References
Aposhian HV 1997 Enzymatic methylation of arsenic species and other new approaches to arsenic toxicity Annu Rev Pharmacol Toxicol 37 397 419 9131259
Barchowsky A Klei LR Dudek EJ Swartz HM James PE 1999 Stimulation of reactive oxygen, but not reactive nitrogen species, in vascular endothelial cells exposed to low levels of arsenite Free Radic Biol Med 27 1405 1412 10641735
Baylin SB Herman JG Graff JR Vertino PM Issa JP 1998 Alterations in DNA methylation: a fundamental aspect of neoplasia Adv Cancer Res 72 141 196 9338076
Brambila EM Achanzar WE Qu W Webber MM Waalkes MP 2002 Chronic arsenic-exposed human prostate epithelial cells exhibit stable arsenic tolerance: mechanistic implications of altered cellular glutathione and glutathione S -transferase Toxicol Appl Pharmacol 183 99 107 12387749
Centeno JA Mullick FG Martinez L Page NP Gibb H Longfellow D 2002 Pathology related to chronic arsenic exposure Environ Health Perspect 110 suppl 5 883 886 12426152
Chen CJ Yu MW Liaw YF 1997 Epidemiological characteristics and risk factors of hepatocellular carcinoma J Gastroenterol Hepatol 12 S294 308 9407350
Chen H Li S Liu J Diwan BA Barrett JC Waalkes MP 2004 Chronic inorganic arsenic exposure induces hepatic global and individual gene hypomethylation: implications for arsenic hepatocarcinogenesis Carcinogenesis 25 1 8 15217905
Chen H Liu J Merrick BA Waalkes MP 2001a Genetic events associated with arsenic-induced malignant transformation: applications of cDNA microarray technology Mol Carcinog 30 79 87 11241755
Chen H Liu J Zhao CQ Diwan BA Merrick BA Waalkes MP 2001b Association of c-myc over-expression and hyperproliferation with arsenite-induced malignant transformation Toxicol Appl Pharmacol 175 260 268 11559025
Chiou HY Hsueh YM Hsieh LL Hsu LI Hsu YH Hsieh FI 1997 Arsenic methylation capacity, body retention, and null genotypes of glutathione S -transferase M1 and T1 among current arsenic-exposed residents in Taiwan Mutat Res 386 197 207 9219558
Counts JL McClain RM Goodman JI 1997 Comparison of effect of tumor promoter treatments on DNA methylation status and gene expression in B6C3F1 and C57BL/6 mouse liver and in B6C3F1 mouse liver tumors Mol Carcinog 18 97 106 9049185
Del Razo LM Quintanilla-Vega B Brambila-Colombres E Calderon-Aranda ES Manno M Albores A 2001 Stress proteins induced by arsenic Toxicol Appl Pharmacol 177 132 148 11740912
Gebel T 2000 Confounding variables in the environmental toxicology of arsenic Toxicology 144 155 162 10781883
Germolec DR Spalding J Boorman GA Wilmer JL Yoshida T Simeonova PP 1997 Arsenic can mediate skin neoplasia by chronic stimulation of keratinocyte-derived growth factors Mutat Res 386 209 218 9219559
Germolec DR Spalding J Yu HS Chen GS Simeonova PP Humble MC 1998 Arsenic enhancement of skin neoplasia by chronic stimulation of growth factors Am J Pathol 153 1775 1785 9846968
Goodman JI Watson RE 2002 Altered DNA methylation: a secondary mechanism involved in carcinogenesis Annu Rev Pharmacol Toxicol 42 501 525 11807181
Gregus Z Gyurasics A Csanaky I 2000 Biliary and urinary excretion of inorganic arsenic: monomethy-larsonous acid as a major biliary metabolite in rats Toxicol Sci 56 18 25 10869450
Hughes MF 2002 Arsenic toxicity and potential mechanisms of action Toxicol Lett 133 1 16 12076506
IARC 1987 Arsenic and arsenic compounds (Group 1) IARC Monogr Eval Carcinog Risks Hum Suppl 7 100 103
Iida M Anna CH Hartis J Bruno M Wetmore B Dubin JR 2003 Changes in global gene and protein expression during early mouse liver carcinogenesis induced by non-genotoxic model carcinogens oxazepam and Wyeth-14, 643 Carcinogenesis 24 757 770 12727805
Kenyon EM Hughes MF 2001 A concise review of the toxicity and carcinogenicity of dimethylarsinic acid Toxicology 160 227 236 11246143
Kitchin KT Ahmad S 2003 Oxidative stress as a possible mode of action for arsenic carcinogenesis Toxicol Lett 137 3 13 12505428
Leder A Kuo A Cardiff RD Sinn E Leder P 1990 v-Ha-ras transgene abrogates the initiation step in mouse skin tumorigenesis: effects of phorbol esters and retinoic acid Proc Natl Acad Sci USA 87 9178 9182 2251261
Liu J Chen H Miller DS Saavedra JE Keefer LK Johnson DR 2001a Overexpression of glutathione S -transferase II and multidrug resistance transport proteins is associated with acquired tolerance to inorganic arsenic Mol Pharmacol 60 302 309 11455017
Liu J Kadiiska MB Liu Y Lu T Qu W Waalkes MP 2001b Stress-related gene expression in mice treated with inorganic arsenicals Toxicol Sci 61 314 320 11353140
Liu J Xie Y Ward JM Diwan BA Waalkes MP 2004 Toxicogenomic analysis of aberrant gene expression in liver tumors and non-tumorous livers of adult mice exposed in utero to inorganic arsenic Toxicol Sci 77 249 257 14691202
Lu T Liu J LeCluyse EL Zhou YS Cheng ML Waalkes MP 2001 Application of cDNA micro-array to the study of arsenic-induced liver diseases in the population of Guizhou, China Toxicol Sci 59 185 192 11134558
Marnell LL Garcia-Vargas GG Chowdhury UK Zakharyan RA Walsh B Avram MD 2003 Polymorphisms in the human monomethylarsonic acid (MMA(V)) reductase/hGSTO1 gene and changes in urinary arsenic profiles Chem Res Toxicol 16 1507 1513 14680363
Mass MJ Wang L 1997 Arsenic alters cytosine methylation patterns of the promoter of the tumor suppressor gene p53 in human lung cells: a model for a mechanism of carcinogenesis Mutat Res 386 263 277 9219564
Nesnow S Roop BC Lambert G Kadiiska M Mason RP Cullen WR 2002 DNA damage induced by methylated trivalent arsenicals is mediated by reactive oxygen species Chem Res Toxicol 15 1627 1634 12482246
Nishikawa T Wanibuchi H Ogawa M Kinoshita A Morimura K Hiroi T 2002 Promoting effects of monomethylarsonic acid, dimethylarsinic acid and trimethylarsine oxide on induction of rat liver preneoplastic glutathione S -transferase placental form positive foci: a possible reactive oxygen species mechanism Int J Cancer 100 136 139 12115560
NRC 1999. Arsenic in the drinking water. In: National Research Council Report. Washington, DC:National Academy Press, 1–310.
NRC 1996. Guide for the Care and Use of Laboratory Animals. 7th ed. Institute of Laboratory Animal Research. Washington, DC:National Academy Press.
Okoji RS Yu RC Maronpot RR Froines JR 2002 Sodium arsenite administration via drinking water increases genome-wide and Ha-ras DNA hypomethylation in methyl-deficient C57BL/6J mice Carcinogenesis 23 777 785 12016150
Petrick JS Jagadish B Mash EA Aposhian HV 2001 Monomethylarsonous acid (MMA(III)) and arsen-ite: LD(50) in hamsters and in vitro inhibition of pyruvate dehydrogenase Chem Res Toxicol 14 651 656 11409934
Pi J Qu W Reece JM Kumagai Y Waalkes MP 2003 Transcription factor Nrf2 activation by inorganic arsenic in cultured keratinocytes: involvement of hydrogen peroxide Exp Cell Res 290 234 245 14567983
Qu W Bortner CD Sakurai T Hobson MJ Waalkes MP 2002 Acquisition of apoptotic resistance in arsenic-induced malignant transformation: role of the JNK signal transduction pathway Carcinogenesis 23 151 159 11756236
Salim EI Wanibuchi H Morimura K Wei M Mitsuhashi M Yoshida K 2003 Carcinogenicity of dimethylarsinic acid in p53 heterozygous knockout and wild-type C57BL/6J mice Carcinogenesis 24 335 342 12584185
Sanders JM Burka LT Chanas B Matthews HB 2001 Comparative xenobiotic metabolism between Tg.AC and p53+/− genetically altered mice and their respective wild types Toxicol Sci 61 54 61 11294974
Scharf JG Dombrowski F Ramadori G 2001 The IGF axis and hepatocarcinogenesis Mol Pathol 54 138 144 11376124
Shen J Wanibuchi H Salim EI Wei M Doi K Yoshida K 2003a Induction of glutathione S -transferase placental form positive foci in liver and epithelial hyperplasia in urinary bladder, but no tumor development in male Fischer 344 rats treated with monomethylarsonic acid for 104 weeks Toxicol Appl Pharmacol 193 335 345 14678743
Shen J Wanibuchi H Salim EI Wei M Kinoshita A Yoshida K 2003b Liver tumorigenicity of trimethylarsine oxide in male Fischer 344 rats—association with oxidative DNA damage and enhanced cell proliferation Carcinogenesis 24 1827 1835 12919961
Simeonova PP Wang S Toriuma W Kommineni V Matheson J Unimye N 2000 Arsenic mediates cell proliferation and gene expression in the bladder epithelium: association with activating protein-1 transactivation Cancer Res 60 3445 3453 10910055
Styblo M Del Razo LM Vega L Germolec DR LeCluyse EL Hamilton GA 2000 Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in rat and human cells Arch Toxicol 74 289 299 11005674
Trouba KJ Hamadeh HK Amin RP Germolec DR 2002 Oxidative stress and its role in skin disease Antioxid Redox Signal 4 665 673 12230879
Trouba KJ Nyska A Styblo M Dunson D Lomnitski L Grossman S 2003. Effect of antioxidants on the papilloma response and liver glutathione modulation mediated by arsenic in Tg.AC transgenic mice. In: Arsenic Exposure and Health Effects V (Chappell WR, Abernathy CO, Calderon RL, eds). Amsterdam:Elsevier Science Ltd., 281–291.
Vahter M 2002 Mechanisms of arsenic biotransformation Toxicology 181–182 211 217
Waalkes MP Fox DA States JC Patierno SR McCabe MJ Jr 2000a Metals and disorders of cell accumulation: modulation of apoptosis and cell proliferation Toxicol Sci 56 255 261 10910982
Waalkes MP Keefer LK Diwan BA 2000b Induction of proliferative lesions of the uterus, testes, and liver in Swiss mice given repeated injections of sodium arsenate: possible estrogenic mode of action Toxicol Appl Pharmacol 166 24 35 10873715
Waalkes MP Liu J Chen H Xie Y Achanzar WE Zhou YS 2004a Estrogen signaling in livers of male mice with hepatocellular carcinoma induced by exposure to arsenic in utero J Natl Cancer Inst 96 466 474 15026472
Waalkes MP Ward JM Diwan BA 2004b Induction of tumors of the liver, lung, ovary and adrenal in adult mice after brief maternal gestational exposure to inorganic arsenic: promotional effects of postnatal phorbol ester exposure on hepatic and pulmonary, but not dermal cancers Carcinogenesis 25 133 141 14514661
Waalkes MP Ward JM Liu J Diwan BA 2003 Transplacental carcinogenicity of inorganic arsenic in the drinking water: induction of hepatic, ovarian, pulmonary, and adrenal tumors in mice Toxicol Appl Pharmacol 186 7 17 12583988
Walker NJ 2001 Real-time and quantitative PCR: applications to mechanism-based toxicology J Biochem Mol Toxicol 15 121 127 11424221
Wang Z Dey S Rosen BP Rossman TG 1996 Efflux-mediated resistance to arsenicals in arsenic-resistant and -hypersensitive Chinese hamster cells Toxicol Appl Pharmacol 137 112 119 8607136
Wei M Wanibuchi H Morimura K Iwai S Yoshida K Endo G 2002 Carcinogenicity of dimethylarsinic acid in male F344 rats and genetic alterations in induced urinary bladder tumors Carcinogenesis 23 1387 1397 12151359
Xie Y Liu J Liu Y Klaassen CD Waalkes MP 2004 Toxicokinetic and genomic analysis of chronic arsenic exposure in multidrug-resistance mdr1a/1b(−/−) double knockout mice Mol Cell Biochem 255 11 18 14971641
Yamamoto S Wanibuchi H Hori T Yano Y Matsui-Yuasa I Otani S 1997 Possible carcinogenic potential of dimethylarsinic acid as assessed in rat in vivo models: a review Mutat Res 386 353 361 9219572
Yamanaka K Hoshino M Okamoto M Sawamura R Hasegawa A Okada S 1990 Induction of DNA damage by dimethylarsine, a metabolite of inorganic arsenics, is for the major part likely due to its peroxyl radical Biochem Biophys Res Commun 168 58 64 2158319
Yoshida K Inoue Y Kuroda K Chen H Wanibuchi H Fukushima S 1998 Urinary excretion of arsenic metabolites after long-term oral administration of various arsenic compounds to rats J Toxicol Environ Health A 54 179 192 9643871
Zhao CQ Young MR Diwan BA Coogan TP Waalkes MP 1997 Association of arsenic-induced malignant transformation with DNA hypomethylation and aberrant gene expression Proc Natl Acad Sci USA 94 10907 10912 9380733
Zhou YS Du H Cheng ML Liu J Zhang XJ 2002 The investigation of death from diseases caused by coal-burning type of arsenic poisoning [in Chinese] Chin J Endemiol 21 484 486
|
15345372
|
PMC1277119
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug 18; 112(12):1255-1263
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7152
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0066215345373PerspectivesEditorialGuest Editorial: Toxicogenomics in Risk Assessment: Communicating the Challenges Pettit Syril D. Health and Environmental Sciences Institute, International Life Sciences Institute, Washington, DC, E-mail:
[email protected] D. Pettit is a senior scientific program manager at the ILSI Health and Environmental Sciences Institute where she has managed collaborative science programs since 2000.
8 2004 112 12 A662 A662 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
In 1999 the membership of the International Life Sciences Institute (ILSI) Health and Environmental Sciences Institute (HESI) formed a multisector consortium to address challenges associated with the integration of genomics data into risk assessment (Pennie et al. 2004). Following its formation, the HESI Committee on the Application of Genomics to Mechanism-Based Risk Assessment identified several key hurdles. These included a lack of publicly available toxicogenomics databases, a lack of validation of available technologies, questions concerning the comparability of different technical platforms and how transcription products relate to toxicity, and uncertain regulatory applications.
In 2004 we have seen considerable progress in many of the areas mentioned above, particularly in our technical ability to execute microarrays and to analyze and interpret the resultant data. The experimental program of the HESI Genomics Committee clearly demonstrated that it is possible to replicate data on biologic pathways across laboratories and technical platforms (Kramer et al. 2004; Ulrich et al. 2004). The committee’s work also revealed the need to interpret modulations in gene expression on microarrays in the context of a broader biologic data set (e.g., clinical chemistry, histopathology). Additionally, within the United States, the recent release of draft regulatory guidance from the U.S. Food and Drug Administration (FDA 2003) on the use of pharmacogenomics data in risk assessment and the release of a white paper from the U.S. Environmental Protection Agency (U.S. EPA 2004) on potential regulatory applications of genomics data have further focused potential applications. However, the routine application of genomics to preclinical risk assessment has not yet been accepted universally. Why?
The efforts of the HESI Committee on Genomics regarding experimental collaboration and toxicogenomics database development (Mattes et al. 2004) suggest that some of the greatest outstanding challenges relate to effective communication across key user groups. It is critical that regulated industries share with the regulatory community the focus of their current approaches to the use of genomics. For example, are microarray data used primarily for early screening or for researching mechanisms of toxicity and and under what circumstances? Additionally, open information exchange regarding typical means of data analysis and presentation is needed. This exchange, which has been initiated via several multisector forums including HESI, will help ensure that a common understanding of the technology’s practical strengths and limitations is reached and allow genomics to be applied more effectively to safety assessment. Discussions concerning the interpretation of patterns of change in gene expression in relation to other biologic end points will provide critical context for determining the suitability of a data set for risk evaluation. Consensus as to when changes in gene expression via microarray represent definitive biomarkers of effect is also needed. Until these conditions are clarified, the utility of genomics for classifying effects of concern will remain debatable.
The risk assessment community is also striving both to harness the collective power of publicly available data sets and to facilitate exchange of single data sets for safety evaluation. As such, numerous formats for the capture and exchange of microarray and toxicology data have become available and/or are under development (Mattes et al. 2004). Diversity of approach is not in itself problematic and clearly has its benefits. However, the development of flexible and comprehensible data exchange platforms that meet the needs of multiple user groups is essential for routine exchange of toxicogenomics data.
The HESI Committee on Genomics looks forward to an ongoing role as a multi-stakeholder consortium committed to facilitating discussion on the scientifically sound use of genomics for risk assessment.
==== Refs
References
FDA 2003. Draft. Guidance for Industry. Pharmacogenomic Data Submissions. Draft Guidance. Washington, DC:Food and Drug Administration. Available: www.fda.gov/cder/guidance/5900dft.pdf [accessed 26 July 2004].
Kramer JA Pettit SD Amin RP Bertram TA Car B Cunningham M 2004 Overview of the application of transcription profiling using selected nephrotoxicants for toxicology assessment Environ Health Perspect 112 460 464 15033596
Mattes WB Pettit SD Sansone S-A Bushel PR Waters MD 2004 Database development in toxicogenomics: issues and efforts Environ Health Perspect 112 495 505 15033600
Pennie W Pettit SD Lord PG 2004 Toxicogenomics in risk assessment: an overview of an HESI collaborative research program Environ Health Perspect 112 417 419 15033589
Ulrich RG Rockett JC Gibson GG Pettit S 2004 Overview of an interlaboratory collaboration on evaluating the effects of model hepatotoxicants on hepatic gene expression Environ Health Perspect 112 423 427 15033591
U.S. EPA 2004. Draft. Potential Implication of Genomics for Regulatory and Risk Assessment Applications at EPA. Washington, DC:U.S. Environmental Protection Agency, Science Policy Council.
|
15345373
|
PMC1277120
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A662
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277120
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0066315345374PerspectivesEditorialGuest Editorial: Regulatory Acceptance of Toxicogenomics Data Frueh Felix W. Huang Shiew-Mei Lesko Lawrence J. Office of Clinical Pharmacology and Biopharmaceutics, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Rockville, Maryland E-mail:
[email protected] W. Frueh is Associate Director for Genomics, OCPB, CDER, FDA. He directs the Interdisciplinary Pharmacogenomics Review Group (IPRG) responsible for the review of genomic data submissions to the FDA. He is a member of the FDA Pharmacogenomics Working Group and chairs the Pharmacogenomics Focus Group of the American Association for Pharmaceutical Scientists (AAPS).
Shiew-Mei Huang is Deputy Office Director for Science, OCPB, CDER, FDA. She currently chairs an FDA drug interaction working group and an OCPB good review practices working group and is a member of the FDA Pharmacogenomics, FDA Ethnicity/Race, and CDER QT Working Groups, and the CDER Research Coordinating Committee.
Lawrence J. Lesko is director, OCPB, CDER, FDA. He chairs the FDA Pharmacogenomics Working Group and is a key member of the FDA Critical Path Initiatives. Note: Photograph for L.J. Lesko was unavailable.
8 2004 112 12 A663 A664 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Early identification of toxicologic side effects of a drug candidate is critical to an efficient drug discovery and development process. Toxicogenomics, the marriage of data-rich genomics approaches with traditional toxicologic end point evaluation combined with increasingly powerful in silico modeling approaches, promises to accelerate this process. The advent of parallel experimental platforms, for example, DNA microarrays, has enabled us to gain insight into complex biologic responses to drugs. The challenge is to analyze and correctly interpret these large data sets. Currently, no common standards exist for such data even though attempts are being made to streamline and standardize the presentation of the information. These efforts include ArrayExpress infrastructure for microarray data (http://www.ebi.ac.uk/arrayexpress), Minimum Information About a Microarray Experiment (http://www.mged.org/Workgroups/MIAME/miame.html), and MicroArray Gene Expression (MAGE) markup language (http://www.mged.org; http://www.omg.org/technology/documents/formal/gene_expression.htm).
The creation of vast amounts of genomics and toxicogenomics data has sparked the development of novel systems to handle this type of information. Ultimately, the success of a toxicogenomics approach in drug development depends on our ability to interpret the data in relation to existing information (e.g., screening of a drug-induced gene expression fingerprint against a database containing drug-related gene expression toxicity profiles). It is critical that interdisciplinary information (chemistry, biochemistry, genetic, genomics, clinical) be integrated into the same data warehouse. Incorporating toxicogenomics data into this approach, which is often referred to as systems biology, will help us understand in much more depth how cells maintain homeostasis and how organisms respond to drug exposure at the molecular level.
“. . . the identification, verification, and validation of biomarkers are critical components of every pharmacogenomics, as well as toxicogenomics, study of cases in regulatory decision making.”
The mission of the U.S. Food and Drug Adminstration (FDA) states that the agency “. . . is responsible for advancing the public health by helping to speed innovations that make medicines and foods more effective, safer, and more affordable. . . .” (FDA 2004). Former agency commissioner Mark McClellan stated that “the FDA priority is facilitating the use of pharmacogenetics-driven treatments” (Salerno and Lesko 2004). The FDA has recently issued a draft, “Guidance for Industry: Pharmacogenomic Data Submissions” (FDA 2003), and has held workshops to discuss issues related to pharmacogenomics data submissions (Salerno and Lesko 2004a, 2004b; Leighton et al. 2004; Ruaño et al. 2004; Trepicchio et al. 2004). This guidance is being revised on the basis of public comments, and a final guidance should be issued later this year. Many principles found in this guidance apply to toxicogenomics studies. In particular, the identification, evaluation, and validation of biomarkers are critical components of every pharmacogenomics, as well as toxicogenomics, study of cases in regulatory decision making. The guidance is general and includes examples of genetic and genomic biomarkers: a CYP2D6 (cytochrome P450 2D6) mutation versus an increase in HER2 (human epidermal growth factor receptor 2) expression can be viewed as genetic and genomic biomarkers, respectively. However, it is anticipated that future data submissions will contain many more complex gene expression profiles and large-scale single nucleotide polymorphism maps (e.g., from whole genome scans), which will present new challenges to define the analytical and clinical validity of such new and highly complex bio-marker sets. The guidance represents the FDA’s current view on pharmacogenomics and what the agency believes are the scientific grounds for evaluating such information as it relates to voluntary versus required submission of data.
What are the next steps? Regulators have been criticized for the lack of guidance in the new era of genomics-based drug development. In addition to the guidance on pharmacogenomics data submissions (FDA 2003), the FDA is embarking on a new guidance initiative for the co-development of pharmacogenomics-based drugs and biologic products and the diagnostic tests necessary for therapeutic decision making. Recently, the FDA and the Drug Information Association (DIA) sponsored a pharmacogenomics workshop (FDA/DIA 2004). The purpose of the workshop was to identify issues in the development of pharmacogenomics-based combination products. We hope to see the base of pharmacogenomics knowledge grow and expand, and we look forward to the use of this information in the drug discovery and regulatory evaluation processes. We expect that not only the novel scientific but also the newly created regulatory tools such as voluntary submissions of genomics data will provide the means by which genomics-based research can excel in advancing public health and drug development.
==== Refs
References
FDA 2003. Draft. Guidance for Industry. Pharmacogenomic Data Submissions. Draft Guidance. Washington, DC:Food and Drug Administration. Available: www.fda.gov/cder/guidance/5900dft.pdf [accessed 26 July 2004].
FDA 2004. FDA Mission Statement. 2004. Washington, DC:U.S. Food and Drug Administration. Available: http://www.fda.gov/opacom/morechoices/mission.html [accessed 26 July 2004].
FDA/DIA 2004. Co-Development of Drug, Biological and Device Products, 29 July 2004, Arlington, VA. Washington, DC:U.S. Food and Drug Administration/Horsham, PA:Drug Information Association. Available: http://www.diahome.org/Content/Events/04040.pdf [accessed 26 July 2004].
Leighton JK DeGeorge J Jacobson-Kram D MacGregor J Mendrick D Worobec A 2004 Pharmacogenomic data submissions to the FDA: non-clinical case studies Pharmacogenomics 5 5 507 511 15212587
Ruaño G Collins JM Dorner AJ Wang S-J Guerciolini R Huang S-M 2004 Pharmacogenomic data submissions to the FDA: clinical pharmacology case studies Pharmacogenomics 5 5 513 517 15212588
Salerno RA Lesko LJ 2004a Pharmacogenomic data: FDA voluntary and required submission guidance Pharmacogenomics 5 5 503 505 15212586
Salerno RA Lesko LJ 2004b Pharmacogenomics in drug development and regulatory decision-making: the Genomic Data Submission (GDS) proposal Pharmacogenomics 5 1 25 30 14683418
Trepicchio WL Williams GA Essayan D Hall ST Harty LC Shaw PM 2004 Pharmacogenomic data submissions to the FDA: clinical case studies Pharmacogenomics 5 5 519 524 15212589
|
15345374
|
PMC1277121
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A663-A664
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277121
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a00664PerspectivesEditorialNote from the Editor: Addressing Applications of Genomics Data Goehl Thomas J. Editor-in-Chief, EHP, NIEHS, National Institutes of Health, Department of Health and Human Services, Research Triangle Park, North Carolina, E-mail:
[email protected] 2004 112 12 A664 A664 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
The two editorials in this issue of EHP address the application of toxicogenomics data in the risk assessment and regulatory processes. Also addressing these issues is the National Research Council/National Academy of Sciences Committee on Emerging Issues and Data on Environmental Contaminants. (http://dels.nas.edu/emergingissues/index.asp). This committee, formed at the request of the National Institute of Environmental Health Sciences (NIEHS), provides a public forum for discussing emerging issues in environmental toxicology.
The committee comprises experts from academia, industry, and public interest groups whose specialties include toxicology, toxicogenomics, genetics, bioinformatics, risk assessment, medical ethics, epidemiology, communications, public health. In addition, a U.S. federal government liaison group has been created to work with the committee with representatives from the NIEHS, the Centers for Disease Control and Prevention, the Environmental Protection Agency, the Food and Drug Administration, the Department of Agriculture, the Department of Energy, the Occupational Safety and Health Administration, the Department of Defense, and the Department of Transportation.
A subcommittee is being formed to write a “Consensus Report on the Applications of Toxicogenomic Technologies to Predictive Toxicology.” This report will highlight how the study of gene and protein activities and other biological processes can improve the characterization of toxic substances and their potential risks. Ultimately, this report should show how major new or anticipated uses of these technologies could improve the protection of public health and the environment.
|
0
|
PMC1277122
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A664
|
utf-8
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0066615345376PerspectivesCorrespondenceMicroarray Data Standards: An Open Letter On behalf of the MGED Society Ball Catherine 1Brazma Alvis 2Causton Helen 3Chervitz Steve 4Edgar Ron 5Hingamp Pascal 6Matese John C. 7Parkinson Helen 2Quackenbush John 8Ringwald Martin 9Sansone Susanna-Assunta 2Sherlock Gavin 1Spellman Paul 10Stoeckert Christian 11Tateno Yoshio 12Taylor Ronald 13White Joseph 8Winegarden Neil 141Stanford University, Stanford, CA, USA;2EMBL-The European Bioinformatics Institute, Cambridge, UK;3MRC Clinical Sciences Centre/Imperial College, London, UK;4Affymetrix, Inc., Emeryville, CA, USA;5The National Center for Biotechnology Information, Bethesda, MD, USA;6INSERM ERM 206, Marseille, France;7Carl Icahn Laboratory, Princeton University, Princeton, NJ, USA;8The Institute for Genomic Research, Rockville, MD, USA;9The Jackson Laboratory, Bar Harbor, ME, USA;10Lawrence Berkeley National Laboratory, Berkeley, CA, USA;11University of Pennsylvania, Philadelphia, PA, USA;12DNA Data Bank of Japan, Mishima, Shizuoka, Japan;13Pacific Northwest National Laboratory, Richland, WA, USA;14University Health Network, University of Toronto, Toronto, Ontario, Canada.Address correspondence to R. Taylor, Biological Sciences Division, Pacific Northwest National Laboratory, PO Box 999, MS K1-92, Richland, WA 99352 USA.The authors declare they have no competing financial interest.
Editor’s Note: We are publishing this open letter to encourage discussion on standards for handling of microarray data. All responses will be published in the November 2004 Toxicogenomics section of EHP.
8 2004 112 12 A666 A667 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
A fundamental principle guiding the publication of scientific results is that the data supporting any scholarly work must be made fully available to the research community in a form that allows the basic conclusions to be evaluated independently. In the context of molecular biology, this has typically meant that authors of papers describing a newly sequenced genome, gene, or protein must deposit the primary data into a permanent, public data repository such as the sequence databases maintained by the DNA Data Bank of Japan (DDBJ), the EMBL-European Bioinformatics Institute (EBI), and the National Center for Biotechnology Information (NCBI). Similarly, we, the members of the Microarray Gene Expression Data (MGED) Society (http://www.mged.org), believe that all scholarly scientific journals should now require the submission of microarray data to public repositories as part of the process of publication. While some journals have already made this a condition of acceptance, we feel that submission requirements should be applied consistently and that journals recognize ArrayExpress (Brazma et al. 2003), Gene Expression Omnibus (GEO; Edgar et al. 2002), and the Center for Information Biology gene Expression Database (CIBEX; Ikeo et al. 2003) as acceptable public repositories. To this end, the members of the MGED Society propose the following as a new paradigm for the publication of microarray-based studies:
Authors should continue to take primary responsibility for ensuring that all data collected and analyzed in their experiments adhere to the MIAME (Minimum Information About a Microarray Experiment; http://www.mged.org/Workgroups/MIAME/miame.html) guidelines and continue to use the MIAME checklist (http://www.mged.org/Workgroups/MIAME/miame_checklist.html) as a means of achieving this goal.
The scientific journals should require that all primary microarray data be submitted to one of the public repositories—ArrayExpress, GEO, or CIBEX—in a format that complies with the MIAME guidelines.
The public databases should work with authors and the scientific journals to establish data submission and release protocols to ensure compliance with MIAME.
To assist with the review process, the databases should continue to work in collaboration with publishers to provide qualified referees with secure means of access to prepublication data. Authors should be strongly encouraged to submit data to the databases during review.
Naturally, data should be protected from general release prior to either publication or authorization from the data submitters, whichever comes first. At a minimum, the journals should require valid accession numbers for microarray data as a requirement for publication, and these accession numbers should be included in the text of the manuscript to allow members of the community to find and access the underlying data.
Since its inception in 1999, the MGED Society has been working with the broader scientific community to establish standards for the exchange and annotation of microarray data. In December 2001, we proposed the MIAME guidelines (Brazma et al. 2001) and requested that interested parties provide feedback on its relevance and utility. The feedback from both researchers and scientific journals was overwhelmingly positive, yet almost everyone who responded also asked for help in implementing these guidelines.
Subsequently, in the summer of 2002, we submitted an open letter to various journals (e.g., Ball et al. 2002a, 2002b) urging the community to adopt the MIAME requirements for microarray data publication. We provided a checklist so that authors could ensure that sufficient information would be available to allow their data to be re-analyzed by others. Again, the response from the community was extremely positive, and most of the major scientific journals now require publications describing microarray experiments to comply with the MIAME standards. While the adoption of these standards has greatly improved the accessibility of microarray data, much of these data remain on individual authors’ websites in a variety of formats; consequently, obtaining and comparing data sets remains a significant challenge. Clearly we need additional requirements for publication that include submission of expression data to public data repositories.
Though one might ask why this requirement was not part of the original MIAME recommendation, the answer is quite simple—MIAME was ahead of its time. While the major public DNA sequence database groups at the NCBI and EMBL-EBI had developed nascent microarray data repositories, and work was under way to create a similar database at the DDBJ, submitting data to these databases was a considerable burden for authors. However, since that time, improvements in the data-entry utilities available for GEO (http://www.ncbi.nlm.nih.gov/geo), ArrayExpress (http://www.ebi.ac.uk/arrayexpress), and CIBEX (http://cibex.nig.ac.jp) databases, as well as a growing number of commercial and academic software packages capable of writing MAGE-ML documents (Spellman et al. 2002) that can be directly submitted to these public databases, have lowered the barriers for data submission to the point where we as a community must now reconsider that submission to one of these databases be a requirement.
Requiring authors to submit microarray data to the public databases will provide a number of distinct advantages to the entire research community:
These established repositories have a commitment to continued community service and to providing some level of assurance that published gene expression data sets will continue to be available into the future.
Having the data available in these public repositories in a standardized format will not only make it more accessible, but it will allow expression data to be integrated with other relevant data, including the available genome sequences, single nucleotide polymorphism (SNP) and haplotype mapping information, the literature, and other resources that can aid in further interpretation of expression patterns. Although many authors now provide some or all of this information, the established databases are much more likely to ensure that these links are maintained and current.
Curation of data submitted to public data repositories will assist authors, reviewers, and publishers in ensuring that the data comply with the MIAME requirements, further enhancing its utility.
The standardization of microarray data formats will enable the development of additional data analysis and integration tools and makes it easier for scientists to access, query, and share data.
Finally, submission prior to publication will make it easier for referees to access the data confidentially, facilitating the review and publication process.
In the same way that availability of sequence data had a profound impact on a wide range of disciplines, we believe that requiring that microarray data be deposited into public repositories as a necessity for publication will accelerate the rate of scientific discovery.
What this proposal requires is a change in the way in which we approach the publication of microarray-based studies. Both authors and journals have a responsibility to ensure that the requisite data are available, and because submitting MIAME-compliant data can take considerable time and effort, this process should be factored into review and publication timelines. However, while this process may be time consuming and painful at first, we believe that the benefits of building an open repository of micro-array data will far outweigh any initial disadvantages. As always, it is our sincere hope that these suggestions stimulate discussion within the community and that together we can arrive at a consensus that ensures that microarray data are widely and easily accessible. Finally, we would like to urge the DDBJ, EMBL-EBI, and NCBI to work together toward exchanging all MIAME-compliant microarray data.
==== Refs
References
Ball CA Sherlock G Parkinson H Rocca-Sera P Brooksbank C Causton HC 2002a Standards for microarray data Science 298 5593 539 12387284
Ball CA Sherlock G Parkinson H Rocca-Sera P Brooksbank C Causton HC 2002b The underlying principles of scientific publication Bioinformatics 18 11 1409 12424109
Brazma A Hingamp P Quackenbush J Sherlock G Spellman P Stoeckert C 2001 Minimum information about a microarray experiment (MIAME)—toward standards for microarray data Nat Genet 29 4 365 371 11726920
Brazma A Parkinson H Sarkans U Shojatalab M Vilo J Abeygunawardena N 2003 ArrayExpress—a public repository for microarray gene expression data at the EBI Nucleic Acids Res 31 1 68 71 12519949
Edgar R Domrachev M Lash AE 2002 Gene Expression Omnibus: NCBI gene expression and hybridization array data repository Nucleic Acids Res 30 1 207 210 11752295
Ikeo K Ishi-i J Tamura T Gojobori T Tateno Y 2003 CIBEX: Center for Information Biology gene EXpression database C R Biol 326 10–11 1079 1082 14744116
Spellman PT Miller M 2002 Design and implementation of microarray gene expression markup language (MAGE-ML) Genome Biol 3 9 RESEARCH0046 12225585
|
15345376
|
PMC1277123
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A666-A667
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277123
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067015345377EnvironewsForumGenomics: Encyclopedia of DNA Dahl Richard 8 2004 112 12 A670 A670 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
The completion of the Human Genome Project in April 2003 was a landmark accomplishment, but much remains to be learned before scientists fully understand the true functionality of the DNA sequences in our genetic matter. To that end, the National Human Genome Research Institute (NHGRI) has initiated a variety of research projects to better understand the sequence. One of the most intriguing and potentially far-reaching of these efforts is the Encyclopedia of DNA Elements, or ENCODE, project, which aspires to create a complete catalog of all the functional elements of the human genome.
“The ultimate goal of the ENCODE project is to create a reference work that will help researchers fully utilize the human sequence to gain a deeper understanding of human biology, as well as to develop new strategies for preventing and treating disease,” says Elise A. Feingold, one of the NHGRI program directors in charge of the ENCODE project.
The goal of the Human Genome Project was simply to sequence the human genome; no distinctions were made between protein coding and noncoding regions. The ENCODE project is intended to pick up where the Human Genome Project left off, by providing answers about the roles that are played by the different genetic elements in the sequence. In addition to studying the human genome, the ENCODE project is also looking at genomic sequences from a variety of animals to provide multispecies comparisons. This will help to identify conserved sequences, which are thought to be strong indicators of functionally important regions in the human genome.
NHGRI launched ENCODE last year with the first round of a total $36 million in grants that will be awarded over a three-year period. The first round of awards went to 14 recipients in the United States and abroad. In addition to the grantees, several other academic and scientific groups are providing specific technical expertise, such as database coordination, to assist the project.
According to Feingold, the grantees and other contributors are working as a consortium to analyze about 1% of the genome. Their goal is to determine the most effective set of methodologies, which will then be applied to the remaining 99%.
One of the grantees is Anindya Dutta, a professor of biochemistry and molecular genetics at the University of Virginia. He and his colleagues are studying ways to map replication elements on human chromosomes. The completion of the Human Genome Project created what Dutta considers an obvious opportunity to embark on such a replication study.
“There are very few origins of replication mapped in human cells—five to ten if you’re generous, but I would say three or four,” Dutta says, referring to genetic elements that are necessary to initiate DNA synthesis. “It was pretty clear when the sequence of the human genome came out that this is a great tool for us to find hundreds of origins and how they are controlled by chromatin structure, gene density, promoter activity, and, of course, sequence.”
Following the model established by the Human Genome Project, NHGRI is calling for the data generated by the ENCODE project to be stored in databases and made freely available to the scientific community. The Center for Biomolecular Science and Engineering at the University of California, Santa Cruz—which is one of the institutions involved in providing support work for the ENCODE project and which also developed the computer programs that ran the sequencing of the human genome—is in charge of maintaining the database for sequence-related ENCODE data. In June 2004 the center added an ENCODE page (http://genome.ucsc.edu/encode/) to its existing genome browser, which gets 5,000 visits a day.
It’s not yet clear what might happen further with the data after the initial pilot project ends in 2006. Feingold says that when the first period ends, “we will evaluate what we have learned and determine the best path for moving forward.”
Encyclopedia genomica. The goal of the ENCODE project of the National Human Genome Research Institute is to create a complete catalog of all of the functional elements of the human genome.
|
15345377
|
PMC1277124
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A670
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277124
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067115366185EnvironewsForumSystems Biology: BAC to the Future Potera Carol 8 2004 112 12 A671 A671 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
In a step forward into the future of gene expression research, molecular biologists and neurobiologists have joined forces to map the genes that control brain structure and neural circuits. The project, called the Gene Expression Nervous System Atlas, or GEN-SAT, maps mouse genes that are also present in the human genome as expressed in the central nervous system. According to project director Nathaniel Heintz, head of the Laboratory of Molecular Biology at The Rockefeller University, New York, GENSAT means that researchers studying degenerative conditions such as Parkinson disease can now have access to gene expression within the brain without having to do their own molecular genetics from scratch. Some unexpected insights have already come to light, giving neuroscientists new places to search for the roots of cognitive impairment.
GENSAT is sponsored by the National Institute of Neurological Disorders and Stroke (NINDS) and is based at The Rockefeller University, although prescreening of candidate genes is conducted by Tom Curran, chair of developmental neurobiology at St. Jude Children’s Research Hospital in Memphis, Tennessee. In situ hybridization is used to screen thousands of candidates to find genes that are active in the central nervous system. Of these, an advisory committee selects 250 genes each year for in-depth analysis by the Rockefeller group. Says Heintz, “Having an advisory committee means this research is done with consensus from many parts of the neuroscience community.”
Information gathered through the project is posted in a public database at http://www.gensat.org/. Started in 2003, the GENSAT database contains detailed information for 300 genes and is updated regularly. With the goal of analyzing 250 genes yearly, the project is planned to run for at least several more years, according to Heintz.
The main tools of GENSAT are bacterial artificial chromosomes (BACs), which are simple loops of bacterial DNA that reproduce outside the cell. BACs adeptly incorporate chunks of introduced DNA from other species, which are preserved and duplicated along with the BACs. The Human Genome Project relied on BACs to help map the human genome.
To measure gene activity and patterns of gene expression in the brain, the GEN-SAT team inserts a reporter gene for enhanced green fluorescent protein into each BAC. When genes are active, the enhanced green fluorescent protein glows bright green. Each BAC is then inserted into eggs harvested from mice, and the eggs are implanted into foster mothers.
The resulting offspring carry the BAC throughout their bodies in all of the cells that express the corresponding gene. Groups of mice are sacrificed at three time points—two of which correspond to critical periods of human central nervous system development—and their brains and spinal cords are analyzed. Mapping gene activity at three different points reveals how the cells migrate and interact.
The first samples are taken when the mouse embryos are 15 days old, which corresponds to the sixth to seventh month of human gestation. “During this period the cortex forms, and defects that lead to malformations occur,” explains project codirector Mary Beth Hatten, head of the Laboratory of Developmental Neurobiology at Rockefeller. The second time point, at 7 days after birth, is equivalent to 6–8 months of age in humans. At this age, interconnections form in the cerebellum, which controls movement, and in the hippocampus, which controls short-term memory. The final observations are made on adult mouse brains at age 7 months, which are similar to those of 30-year-old humans.
Findings published in the 30 October 2003 issue of Nature reveal some of the surprising connections the GENSAT project is uncovering. For example, people with DiGeorge syndrome, a congenital condition marked by heart defects and learning disorders, lack a gene called Gscl. Heintz, Hatten, and other GENSAT researchers discovered that Gscl is produced by neurons in the interpeduncular nucleus, the brain region that also regulates rapid-eye-movement sleep. Another finding reported in this paper relates to the striatum, which degenerates in patients with Parkinson disease. In end-stage Parkinson disease, up to 95% of so-called spiny neurons are lost. Until recently, the striatum had been the only place where spiny neurons were found, says Hatten. Yet, the BAC method identified vectors that can be used to separately analyze spiny neurons that project to the substantia nigra and the globus pallidus.
The GENSAT methods can also monitor the effects of environmental toxicants, such as lead, on brain development. “You can expose the BAC mice to any environmental condition you want, to see how the migration and maturation of neurons changes,” says Hatten.
“The tools and mouse lines provided by this project allow the neuroscience community to perform detailed studies of each gene,” says Laura Mamounas, the GEN-SAT project officer at the NINDS. “GEN-SAT also may serve as a model for future gene expression projects.”
Indeed, BAC mice can be used to screen gene activity in other organs. The BAC mice are made available to other researchers who are interested in performing systematic studies of gene expression. Scientists in other specialties are “just starting to bootstrap our efforts to get their particular information,” says Heintz.
Mr. Greengenes. The GENSAT project uses enhanced green fluorescent protein to map mouse genes that are also present in humans and expressed in the central nervous system.
|
15366185
|
PMC1277125
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A671
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277125
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067215366186EnvironewsForumBioinformatics: The Path to Species Comparison Holton W. Conard 8 2004 112 12 A672 A672 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Systems biology relies on integrating genetic, proteomics, and metabolic data, and on understanding interdependent cellular and intercellular events that are constantly in flux. To accomplish this feat, researchers have relied on DNA and protein sequence databases and high-throughput expression analysis techniques such as microarrays to produce ever-growing libraries of expression data. DNA and protein sequences can be quickly such as BLAST (Basic Local Alignment Search Tool), a program that identifies similar genes in different organisms. Now scientists are applying this computational approach to protein interaction networks, which are the means by which proteins communicate.
“As we move from a focus on sequences to one on networks, we need a tool similar to BLAST,” says Trey Ideker, an assistant professor of bioengineering at the University of California, San Diego. The software program PathBLAST was developed to fill this need by a group consisting of researchers from Ideker’s lab and the lab of Brent Stockwell, now an assistant professor of biological sciences at Columbia University. At the time, both Ideker and Stockwell were fellows at the Whitehead Institute for Biomedical Research in Cambridge, Massachusetts, and worked on the program development with Richard Karp, a professor of bioengineering and mathematics at the University of California, Berkeley, known for his work in combinatorial algorithms and bioinformatics.
The PathBLAST program rapidly compares protein interaction networks across two different organisms using fast-executing algorithms. The program searches for high-scoring alignments involving one path from each network. The proteins of the first path are paired with putative homologs—or proteins presumed to have a common origin and function—from the other species and occurring in the same order in the second path. PathBLAST is built as a plug-in to Cytoscape, a widely used software platform. Scientists use Cytoscape to visualize molecular interaction networks and integrate these interactions with gene expression profiles and other data.
“The important stuff in biology is revealed by comparing things,” says Ideker. “By comparing protein interaction networks of two different species or even within species, we can identify pathways and complexes that have been conserved over evolution.” These evolutionarily conserved pathways allow interpretation of the network of a poorly understood organism based on its similarity to that of a well-known species. This comparison could provide a model of signaling and regulatory pathways that are related to a response to an environmental toxicant. It could also help target drugs to pathways that are present in a pathogenic organism but absent from its human host. Such a model could furthermore help identify drugs that would repair damaged pathways or even cause new ones to be formed.
The PathBLAST development group published a paper in the 30 September 2003 issue of Proceedings of the National Academy of Sciences in which they identified the conserved pathways within the yeast Saccharomyces cerevisiae and the bacterium Helicobacter pylori. For example, the authors found that one pathway that was critical in catalyzing DNA replication and another in protein degradation were conserved in both organisms as a single network. Within seconds, the program had determined that the bacterium contained 1,465 interactions among 732 proteins, and the yeast contained 14,489 interactions among 4,688 proteins.
This report proved that the method works for matching conserved networks from among all the networks in two species, according to software engineer Brian Kelley, a member of Stockwell’s lab. Kelley says, “The next step is to prove the software in a novel application where you start with a given disease network and see if it is conserved in other species. Once you prove this utility, then the use of PathBLAST will skyrocket.” Kelley adds that research into the mTOR cell growth–triggering protein pathway may prove to be that application. This pathway is composed of a complex of proteins that respond to nutrient cues; understanding it will clarify the role that nutrients and metabolism play in disease.
Other researchers have taken a complementary approach by comparing what’s known about a disease to a known network. At Beyond Genomics in Waltham, Massachusetts, researchers measure quantitative differences between transcripts, proteins, and metabolites across a given disease model, determine correlations within the data set, and then compare the experimentally derived network with a known biological network or pathway.
“As the protein interaction databases become more heavily populated with interactions among higher eukaryotes, PathBLAST and related approaches will start to shine as they can help elucidate the set of core biological networks for a given genome,” says Tom Plasterer, the principal scientist for bioinformatics at Beyond Genomics. “These networks—when coupled with a tightly defined experimental context—will be invaluable in understanding mechanisms of disease, where one expects compensatory and subtly differing biological networks to emerge.”
The PathBLAST website is hosted by the Whitehead Institute and available at http://www.pathblast.org/; it will soon be mirrored at the San Diego Supercomputer Center at the University of California, San Diego. And as for whether industry will embrace PathBLAST, Ideker says, “It’s still early. Speculating too far about these technologies is like asking industry in 1980, ‘Is genome sequencing going to revolutionize your drug discovery pipeline?’ Even in 2004, the verdict is still out on that one!”
Conservation comparison. PathBLAST software allows researchers to identify protein interaction networks that are conserved across multiple species.
|
15366186
|
PMC1277126
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A672
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277126
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067315366187EnvironewsForumPharmacogenomics: Activating Cancer Drug Discovery Mead M. Nathaniel 8 2004 112 12 A673 A673 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Normal cells that transform into cancer cells undergo various metabolic changes, including shifts in activities of enzymes that mediate macromolecule synthesis and growth-signaling pathways. Proteomics technology now provides an elegant way to identify the enzymes that are active in processes linked with tumor progression. As demonstrated by a recent study conducted at The Burnham Institute’s Cancer Center in La Jolla, California, this approach is beginning to unveil some novel high-efficacy targets for cancer control and treatment.
In work published 15 March 2004 in Cancer Research, the Burnham team used a novel proteomics screen based on probes that bind to the active site of the enzyme target. By competing with such probes for the active site, one can simultaneously identify protein targets and screen for their inhibitors. Activity-based proteomics screening is fast emerging as “the wave of the future,” says coauthor Steven J. Kridel, a postdoctoral fellow at the time of the research and now an assistant professor of cancer biology at Wake Forest University of Winston-Salem, North Carolina—it enables the generation of hypotheses that can lead to meaningful clinical applications. The chemical strategy for activity-based proteomics was pioneered in the laboratories of cell biologist Ben Cravatt of The Scripps Research Institute and pathologist Matthew Bogyo of Stanford University. Kridel and colleague Jeffrey Smith, associate scientific director for technology at The Burnham Institute, are among the first to use the approach to identify a therapeutic lead.
The activity-based strategy may mark a major improvement over the usual proteomics approaches, which are based on the relative abundance of a particular protein target. “Measuring the abundance of a protein only provides a static picture of a potential target enzyme,” says Kridel. “There are several levels of regulation between protein abundance and protein activity. With activity-based proteomics, you also can tell whether there is a specific physiologic state that turns off the enzyme’s activity and whether an inhibitor of that particular enzyme exists.”
Kridel and Smith applied the activity-based strategy to identify proteins that exhibit different activities in cancer cells as compared to normal cells. They screened a group of enzymes known as serine hydrolases by measuring the activity levels of these enzymes in normal prostate epithelial cells and in three standard prostate cancer cell lines. They found that serine hydrolase expression was generally similar among all cell lines, with two key exceptions: one of the hydrolases was active in normal prostate cells but virtually inactive in all the tumor cells, while another was expressed in all of the tumor lines but absent in the normal cells. The latter enzyme was shown to be fatty acid synthase (FAS), which had earlier been strongly linked to tumor progression, making it an attractive therapeutic target.
Having identified their molecular target of choice, the investigators then screened possible inhibitor drugs, hoping to find unforeseen side benefits in drugs already approved for human use. “Our goal from the outset was to find an anticancer drug that might not have been considered before,” says Kridel. “We wanted a drug that inhibits a protein that is only expressed in cancer cells, not in normal cells, in part because we believed this would minimize toxic side effects.” Among the many agents reviewed was the anti-obesity drug orlistat (trade name Xenical). Kridel says orlistat had not previously been shown to inhibit FAS, and FAS inhibition is not believed to be relevant to orlistat’s mode of action in weight loss.
In cell culture studies, the Burnham team found that orlistat inhibited proliferation and induced apoptosis in at least two lines of prostate cancer cells. The antiproliferative effects were reversed by the addition of palmitate, the precursor for the majority of nonessential fatty acids, which cancer cells use primarily for energy and growth. This strongly implicated FAS inhibition, as FAS is the only eukaryotic enzyme capable of synthesizing palmitate. In rodent experiments, orlistat blocked tumor growth significantly, and the animals showed no outward signs of toxicity or adverse changes in blood chemistry.
By revealing some of the unanticipated effects of a drug, activity-based proteomics could markedly reduce the cost of drug development. “Orlistat just happens to be an approved drug with relatively minor toxicity that could be utilized quickly once its effectiveness in human prostate cancer is validated,” says Massimo Loda, an associate professor of pathology at Harvard Medical School and the Dana Farber Cancer Institute in Boston, Massachusetts. “The implications of this study are dual: this activity-based proteomics approach can now be applied to the screening of diverse families of enzymes that sustain tumor survival, and it may reveal unsuspected activity of known drugs utilized in diseases other than cancer.”
Such research may eventually pave the way for construction of a proteomics profile of susceptibility to cancer progression. “If a man presents with prostate cancer and has a biopsy, it is entirely possible that the proteomics screening approach can be used to assess whether his tumor has upregulated FAS,”
Smith says. “If it does, you can then prescribe a specific treatment regimen: to reduce dietary fat and block FAS activity using orlistat. This is moving toward personalized medicine.” Smith believes a low-fat diet could reinforce orlistat’s cancer-fighting effects in humans. “We know that tumor cells have a unique requirement for fat,” he says. “If you restrict dietary fat and knock out the tumor’s ability to synthesize its own fat from carbohydrates, then the antitumor effect should be even greater.”
Dual-purpose drug? A novel activity-based proteomics screen of the weight-loss drug orlistat revealed its surprising potential as a cancer treatment.
|
15366187
|
PMC1277127
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A673
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277127
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067415366188EnvironewsForumTXGnet: National Center for Biotechnology Information Dooley Erin E. 8 2004 112 12 A674 A674 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
The National Center for Biotechnology Information (NCBI) of the National Library of Medicine was created as a means of developing new information technologies to help advance the field of molecular biology. Composed of a group of scientists from disciplines including mathematics, research medicine, and structural biology, the center’s programs and activities are centered on both basic and applied research in computational molecular biology. Since the center’s inception in 1988, NCBI scientists have developed important novel algorithms and research approaches in the fields of computational biology and gene sequencing, among others. Today, researchers and others can find an impressive array of “omics” resources and tools collected into one central resource on the NCBI website at http://www.ncbi.nlm.nih.gov/.
General information on the center’s work can be found within the top center section of the homepage as well as by clicking the About NCBI link on the left side of the homepage. A large portion of the NCBI web-site is devoted to the number of free, publicly accessible databases and software programs that the center hosts. Included among the databases on offer are GenBank, the Online Mendelian Inheritance in Man, the Cancer Genome Anatomy Project, and numerous organism-specific genome databases. The site classifies these databases as literature search databases, molecular databases, or genomic biology databases to make finding them easier; GenBank, NIH’s annotated genetic sequence database, stands under its own heading. The Molecular Databases page—which is based on Entrez, the integrated search and retrieval system developed by the NCBI—has more than 25 additional database classifications such as nucleotide, protein, and expression. Entrez can also be quickly accessed through a pull-down menu at the top of the homepage.
The Genomic Biology section of the site provides a brief overview of this relatively new area of science. Visitors to this section will find a wealth of human genome resources, including a map viewer that can be used to browse the human genome. By selecting the Human Genome Resources page, users can access PDF background documents on the databases available on the site, two of which provide instructions on how to use the map viewer to explore genomes.
The main portion of this Genomic Biology area is divided into sections on genes and human health and contains links to Online Mendelian Inheritance in Man, RefSeq, dbSNP, and Gene Database. Visitors can also access BLAST for comparing genomic sequences and gene products, and a centralized registry of genomic clones, end sequences, mapping data, and distributor information. Other tools are available as well, and are classified under Maps and Markers, Transcribed Sequences, Cytogenetics, and Comparative Genomics.
The right-hand toolbar for the Genomic Biology section contains resources for 20 specific organisms. Selecting any one of these organisms takes the user to a guide outlining the many different search tools and other resources offered by the NCBI and other groups. These resources include gene maps, sequences, and annotation projects.
Educators and others using the site can access an online guide to GenBank and NCBI resources through the Education link on the homepage. Also on the Education page are tutorials for BLAST, Entrez, and other tools available on the site, as well as access to NCBI newsletters and map viewer exercises. Also within this section is an online science primer developed by the NCBI as a way to familiarize nonspecialists with terms and concepts such as “genome mapping,” “expressed sequence tags,” and “phylogenetics.”
|
15366188
|
PMC1277128
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A674
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277128
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067615345378EnvironewsNCT UpdateCooperation Achieves Results at UNC-CH Eubanks Mary 8 2004 112 12 A676 A676 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Now that scientists have solved the code for the human genome and are moving forward to sequence the genomes of other organisms, a significant challenge lies in how to harness this vast amount of new information to benefit society. Recognizing that cross-communication and cooperation among different research labs is essential to realizing the potential of genomic science for toxicology, the NIEHS earmarked $37 million to establish the Toxicogenomics Research Consortium (TRC) in November 2001.
A major objective of the TRC is to link investigator-initiated research projects at five academic centers—the Massachusetts Institute of Technology, Duke University, the University of North Carolina at Chapel Hill (UNC-CH), Oregon Health & Science University, and the Fred Hutchinson Cancer Research Center/University of Washington—in a collaborative effort to understand how organisms respond to chemical exposures and other kinds of environmental stress. At UNC-CH, investigator-initiated research projects focus on different model systems—from mice and cell lines to humans—all with a common, unifying focus on environmental challenges that cause DNA damage or oxidative stress, leading to cancer.
With the goal of finding molecular signatures that can be used for early risk assessment, these projects employ microarray tools to assay for gene expression responses to cancer-related chemicals. Some use the same compounds that cancer patients are already receiving, to gather information on how different populations may react to certain drugs.
Dampening the Background Noise
In one TRC projects, David W. Threadgill, an assistant professor of environmental sciences and engineering, uses mouse models to investigate differences in response to various tumor-inducing and chemotherapeutic chemicals from one mouse strain to another. In his model, all the mice of one strain are essentially identical, providing an unlimited number of individuals that have the same genotype. This experimental approach enriches the opportunity to separate out what is “noise” in the data from what is truly meaningful. Threadgill’s lab is currently profiling about a dozen mouse strains, some of which are susceptible to certain types of cancer while others are resistant.
Among organisms, there is a wide range of susceptibility to the effects of different toxicants. The classes of chemical compounds that cause cancer—including alkylating agents, DNA-damaging agents, and toxicants that affect the cell cycle—vary in the mechanisms by which they cause the disease. There is also extensive variation in the way different cancer patients respond to treatment and how they handle the toxic effects of different chemical therapies.
To better understand why this is so, Threadgill exposes different mouse strains to various chemicals over a variety of dose regimens and time spans, then compares the results to a control group. After exposure, tissue is harvested from the liver, colon, and breast for extraction of RNA that is converted into cDNA. The cDNA is hybridized to microarray chips containing 17,000 mouse genes. If a gene is turned on or up as a result of exposure, it will fluoresce more brightly than the same gene on the corresponding chip for the control group. Analysis of the signals reveals which genes increase their expression level as a result of exposure to a particular toxicant and which genes are turned down.
Threadgill explains, “Lots of genes are turned on, up, or down. Comparison of the signals from different strains reveals which genes are unique to certain classes of compounds, at what dose levels and time treatments.” His lab also examines direct causes of the associated physical changes and their correlation with prior knowledge of how certain classes of compounds affect tissue. He has detected significant strain-dependent differences in gene expression in both baseline and treated colon tissue. He says, “The characterization of these differences in the context of disease and treatment end points should reveal new insights into the mechanistic causes of the cancer end points.”
Screening thousands of genes and expression changes presents a huge computational challenge to discern what the critical signatures are. Biostatisticians have begun to mine the data collected by Threadgill over the first two and a half years of his project to evaluate thousands of signals and identify similarities and differences between treatment time courses, exposure levels, and individual strains. Knowing which genes are significant is vital to finding prognostic markers for whether an individual is susceptible to effects from certain compounds and for evaluating what makes certain individuals more susceptible than others.
Layers of Response
In a coordinated experimental design, Charles M. Perou, an assistant professor of genetics, is screening some of the same compounds as Threadgill in his investigations of the transcriptional response in breast cancer using human cell lines from breast tumors and immortalized breast epithelial cell lines (that is, publicly available cell lines that have been maintained over time for experimental use). This may allow for comparative analyses of results across different model systems that could lead to new insights into the resulting biological insults from exposure to specific toxicants. Cell lines are treated with different regimens, and then examined for gene expression changes using microarrays.
Perou is also studying the effects of these drugs on real breast tumors of patients before, during, and after chemotherapy for comparison of similarities or differences in experimental cell lines with tumors in patients. In a paper in the June 2004 issue of Cancer Research, Perou’s lab reported that a small but significant set of genes were identified as being changed in both in vitro cells and living tumors, and that these are likely to be very important in cellular response to toxicants. In this study, 20–30 changes observed in the expression profiles identified potential candidate genes for a general response to these compounds.
Perou and colleagues have discovered that although the compounds tested have different mechanisms and targets in cells, the overall cellular response to the different drugs is the same—a sort of dominant generic stress response that is similar across genes. The secondary drug-specific response pattern is layered on top of this generic stress response pattern. Perou says, “Arrays do a great job of implicating genes in certain processes in response to different drugs, but the genes involved in the response are the same. . . . This leads to the hypothesis that some people with reduced ability to respond might be more susceptible because of variants in these generic stress response genes.”
Perou’s work has also uncovered unique molecular signatures that distinguish two subtypes of breast cancer. Patients with one type will do well under the standard drug therapies and survive. However, for those with the other type, prognosis for survival if given standard treatment is not good; more aggressive treatment such as radical mastectomy and radiation would be needed, and even then the chances of survival are not nearly as good. This complex overlapping pattern of drug-specific response genes embedded within the common response will be reported in a second paper submitted for publication.
Perou is particularly interested in the generic stress phenomenon because individuals who have susceptible variants of these genes could be unable to respond to many different toxicants in their environment. He says, “Microarrays provide a big list of candidate genes. The challenge is to figure out [which genes] are the passengers and [which] are the drivers, then apply that information to population studies to identify genetic variants that are hyperactive and correspond to susceptibility.” Through the TRC, he says, it may ultimately be possible to link findings and design experiments such that scientists can identify a common set of genes in mice and humans for independent validation for susceptibility. “As we learn more about evolutionarily conserved genes across species that are involved in reactions to environmental chemicals,” Perou says, “we will have more confidence in the scientific findings.”
Confidence Building
William K. Kaufmann, director of the UNC-CH Program in Toxicogenomics and director of the Genetic Susceptibility Research Core in the UNC Center for Environmental Health and Susceptibility, studies expression of tumor suppressor genes in the cell cycle. Kaufmann is interested in how changes in the DNA at the ends of chromosomes provide life span checkpoint monitors that signal the time for aging cells to die. Experimental evidence indicates that responses similar to those involved in cell senescence occur in breast, liver, and colon cells exposed to carcinogenic compounds. Such changes are accompanied by reproducible gene expression patterns.
Kaufmann says, “When the same unexpected patterns of response are seen in lines from several different individuals, our confidence that the response to stress is biologically meaningful increases.” Knowledge gained from the characterization of modulations in expression of genes as a result of exposure to different carcinogens could one day provide a fingerprint to enable design of precise treatment strategies for specific individuals exposed to different toxicants.
As the robustness of microarray technology for gene expression profiling improves and becomes more standardized, scientists will be able to employ molecular signatures to tailor therapies as well as conduct exposure risk assessment. Kaufmann concludes that in spite of sometimes feeling overwhelmed by too many genes and the need for a bigger computer, the UNC-CH microarray research projects are producing a data-rich picture, and are spawning whole new areas of discipline to look for patterns of change. Such individual scientific discoveries and applications of the principles of toxicogenomics are resulting in a serendipitous research crossover that will continue to spur insights, leading to faster, broader, and more effective applications for public health.
A perfect likeness. UNC-CH researcher David W. Threadgill uses strains of mice in which all the animals have the identical genotype to study cancer susceptibility from exposure to a variety of environmental toxicants. Some of the mouse strains are resistant to cancer, while others are more susceptible to the disease.
Molecular mysteries. UNC-CH researchers are elucidating how susceptibility to breast cancer is impacted by a “generic” stress response that is similar across genes. Such a generic response underlies characteristic drug-related responses, which molecular signatures indicate vary from person to person.
|
15345378
|
PMC1277129
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A676
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-1277129
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0067815345379EnvironewsFocusToxicogenomics Data: The Road to Acceptance Freeman Kris 8 2004 112 12 A678 A685 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Questions abound regarding the use by regulatory agencies of data from microarray experiments. How does a regulator deal with risk assessment data that scientists are often unable to interpret—data that some companies are anxious to submit and others to withhold? How does this same regulator evaluate information that is produced without universally recognized standards for laboratory protocols or data formats? Pharmaceutical and chemical companies have their own questions. Do you submit all your data voluntarily, without knowing whether regulators will be able to understand it, and if and exactly how they will use it? Will submission of such complex data slow down approvals? What if data that cannot be interpreted now are later show to indicate toxicity, perhaps at levels that couldn’t be detected in animal testing? Could lawsuits follow? Regulatory penalties?
These and other questions have been discussed at numerous meetings between industry, environmental groups, the Food and Drug Administration (FDA), which regulates medications, and the Environmental Protection Agency (EPA), which regulates pesticides significantly in the last two years, according to John Leighton, supervisory pharmacologist in the Division of Oncologic Drug Products of the FDA Center for Drug Evaluation and Research (CDER). At an FDA-sponsored workshop held in May 2002, Leighton says, the take-home question was whether microarray technology was sufficiently developed for scientific purposes. Within 18 months, at a follow-up workshop in November 2003, he says, “the issues had shifted from ‘Is the technology useful?’ to How is it useful?”
Proving that microarray data, or “expression signatures,” can be valid measures of environmental exposure is a major accomplishment of a 1999–2003 research program by the Health and environmental Sciences Institute. The HESI experiments, published in the March 2004 toxicogenomics issue of EHP, showed the patterns of gene expression detected in microarray experiments can give insight into biologic mechanisms, and can even distinguish between damage found in different cell types. Although most expression signatures still can’t be interpreted or linked to biologic effects, officials at both the FDA and the EPA express optimism that the use of microarray data could help them better protect public health. Industry however, taken as a whole, may not be quite so sure.
Efforts to encourage communication between regulators and industry have included workshops held by the National Research Council (NRC) Committee on Emerging Issues and Data on Environmental Contaminants, which is funded by the NIEHS. Since its establishment in April 2002, the committee has held seven workshops to discuss issues related to the future use of toxicogenomics data in government risk assessment and regulatory decision and policy making. These issues include the many challenges that remain to be resolved before these tools find direct application in chemical risk assessment, says David Eaton, chair of the NRC committee and director of the NIEHS Center for Ecogenetics and Environmental Health at the University of Washington.
Because of the expense involved in running microarray experiments, including the costs of analyzing data, microarrays are generally not used in detailed dose and time-course studies, says Eaton. As a result, current microarray data often provide a limited snapshot of information that Eaton says can be very useful in terms of generating hypotheses about mechanisms of exposure, although the application of such information for regulatory purposes is fraught with uncertainty.
We think there are powerful uses for genetic data, including microarray data, in the real world of drug safety, to both test products and do “forensic” studies—that is, go back and investigate safety problems after marketing. –Janet Woodcock, CDER
Government’s Take on Microarray Data
“We think there are powerful uses for genetic data, including microarray data, in the real world of drug safety, to both test products and do ‘forensic’ studies—that is, go back and investigate safety problems after marketing,” says CDER director Janet Woodcock. “In cases of some adverse drug effects, companies may be able to go and look for specific genotypes that are distinctive and at risk for an adverse event.” Agency regulators, she adds, hope that intractable drug toxicity problems, such as hepatotoxicity, could be solved through microarray or gene expression technologies.
EPA representatives hope that microarray technology, along with proteomics and metabolomics experiments, will help the agency better screen the vast number of chemicals it is mandated to regulate. Using traditional tests, it can easily take 3–4 years and $20 million to test the toxicity of a pesticide, says Robert Kavlock, director of the Reproductive Toxicology Division in the EPA Office of Research and Development. “In the long run, we expect that the use of ‘omics’ technologies can be applied to a variety of bioassays, some in vitro, some in vivo, that will help us prioritize chemicals for testing in the more lengthy, expensive, and animal-intensive testing batteries, and perhaps even to guide selection of which tests should be done within those batteries,” says Kavlock. “By doing so, we will become more efficient and effective in our utilization of animal tests.”
“Genomics won’t replace animal testing, not yet,” adds William Benson, director of the Gulf Ecology Division of the EPA National Health and Environmental Effects Research Laboratory. “But we hope it will allow us to use animals more wisely.”
Microarrays and other “omics” technologies could also be used in environmental monitoring, such as water testing. As the technology decreases in cost, local regulators may be able take a microarray chip into the field, apply water samples, and get an answer right there regarding the presence of bacteria, viruses, and other pathogens, according to Kerry Dearfield, senior scientist for science policy in the EPA Office of the Science Advisor. Dearfield, along with Benson and Kathryn Gallagher, science policy council staff in the Office of the Science Advisor, wrote a March 2004 EPA draft white paper on the impact of genomics technologies on EPA regulatory activities.
Although both the EPA and the FDA have discussed possible uses of microarray data, only the FDA has issued requirements for the submission of such data. At press time the agency was working on a final version of its “Guidance for Industry: Pharmacogenomic Data Submission,” released in draft form in November 2003. Under the draft guidance, companies may be required to submit microarray data used to determine differential dosing of a medication by genotype during development (a requirement that applies to animal testing as well as human clinical trials). The guidance also encourages, but does not require, companies to develop suitable genetic tests for such medications to allow physicians to determine if a drug is appropriate for a given patient.
“The centers for drugs and devices are working together for the development of drug–device combinations,” says Atiqur Rahman, acting deputy director for the CDER Division of Pharmaceutical Evaluation 1. “If a drug’s approval becomes based upon a specific test, you can’t approve the drug unless the test is available.”
The draft guidance also encourages, but again does not require, voluntary submission of microarray data from exploratory studies such as experiments to screen multiple compounds for possible toxicity or efficacy. Companies are also asked to supply research data resulting from general gene expression analyses in cells, animals, and humans, as well as analysis of single-nucleotide polymorphisms in trial participants.
In addition, all data on “known valid biomarkers,” including those collected during exploratory studies, must be submitted to the FDA. Although the guidance does not specify the types of biomarkers that must be submitted, Woodcock clarifies that the agency is mainly interested in so-called safety biomarkers, those that indicate toxicity. “Companies don’t have to submit any data on nonclinical efficacy biomarkers,” she says.
Currently, the EPA’s official dictum on the regulatory use of microarray data is limited to a four-page “Interim Policy on Genomics” issued in June 2002. The interim policy states that microarray data are expected to be valuable, and that they “may be received as supporting information for various assessment and regulatory purposes, e.g., identifying an environmental stressor’s mode or mechanism of action.” But the interim policy does not provide any details on potential required submissions of gene expression data. There is no current effort at the EPA to expand or update the interim policy.
What It Means for Industry
Industry response to these regulatory efforts ranges from enthusiasm to extreme caution. “Some companies will not test a drug with a microarray experiment that has any chance of becoming part of a regulatory package,” says Kurt Jarnagin, vice president for biological sciences and chemical genomics at Iconix Pharmaceuticals. “And then there are companies who view [submission of microarray data] as a positive, who say the FDA gets more information, we get more information, and we might find a positive aspect to our drug that we didn’t know about.”
There are already a number of drugs approved for people with specific genetic variations. Most are powerful cancer drugs for which the boundaries between efficacy and toxicity are narrow. One example is imatinib mesylate (trade name Gleevec), which is approved for patients with a specific type of leukemia characterized by a chromosomal rearrangement in the cancerous cells. Another example is trastuzumab (trade name Herceptin), an intravenous treatment for advanced metastatic breast cancer. Trastuzumab is effective in treating tumors that produce excess amounts of the HER2 protein, a tyrosine kinase receptor.
In addition to determining who is most likely to respond to a drug, genetic studies could also be used to screen out those most susceptible to toxic side effects. One example is the case of the lung cancer drug gef-tinib (trade name Iressa), which inhibits a tyrosine kinase that is overexpressed in non–small cell lung cancer, the leading cause of cancer deaths in the United States. After the drug was approved, the FDA received reports of severe, sometimes fatal, toxicity in 0.3–2.0% of patients receiving the drug. In addition, during clinical trials, the drug was effective in only 10–19% of persons with non–small cell lung cancer. Preliminary results published 20 May 2004 in the New England Journal of Medicine indicate that the drug is effective only in people who have heterozygous mutations in the tyrosine kinase epidermal growth factor receptor, coded by the gene EGFR.
Microarray data were not submitted during the approval process for any of these drugs, but could be used in the future to help develop population-specific treatments, according to Rahman. “A certain type of gene expression constituting a gene signature may help determine if a person is a candidate for treatment with a particular drug and is likely to respond to the therapy,” he says.
In contrast to the pharmaceutical industry, “the chemical industry is not chomping at the bit to use toxicogenomics data,” claims Linda Greer, director of the Natural Resources Defense Council public health program and a member of the NRC committee. “The status quo works better for them rather than a system where chemicals can be screened systematically,” she adds. More than 90% of the industrial chemicals in commerce have not been tested for their toxicity, Greer says, and better screening might cause increased scrutiny of such compounds under the Toxic Substances Control Act (TSCA) and the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA).
In the long run, we expect that the use of “omics” technologies can be applied to a variety of bioassays, some in vitro, some in vivo, that will help us prioritize chemicals for testing in the more lengthy, expensive, and animal-intensive testing batteries, and perhaps even to guide selection of which tests should be done within those batteries. –Robert Kavlock, EPA
TSCA gives the EPA authority to require reporting or testing of industrial chemicals that may pose an environmental or human health hazard, and to ban the manufacture and import of chemicals that pose high risks. However, the EPA is not required—nor does it have the resources—to perform extensive toxicity testing on every industrial chemical available for sale in the United States. Nor can chemical companies increase their profits by determining genetically based differences in responses to their general-use products. “What we sell is going to be out there for the general population to use, so we’re compelled to protect the most sensitive individual,” says George Daston, a senior toxicologist in the Central Product Safety group of Procter and Gamble.
Risk exposure testing for industrial chemicals can also be less straightforward than pharmaceutical testing, increasing challenges for both industry and the EPA. In contrast to pharmaceuticals, which people generally are exposed to at known doses for intended biologic effects, environmental exposures to industrial chemicals among the general public are often quite low. In addition, people are often exposed to mixtures of compounds—for example, to several pesticides from a piece of fruit, or to hundreds of chemicals from swimming near a storm sewer outfall. As a result, singling out the effects of a single industrial compound can be extremely difficult.
Nevertheless, some chemical companies are conducting microarray experiments to better understand mechanisms of toxicity, which could lead to better risk assessment information regarding susceptible populations, co-mixtures of chemicals, and low levels of exposure, according to Greer and Dearfield. For example, Greer says, microarray studies could build on research published in the August 2004 issue of EHP linking exposure to the complex mixtures of disinfection by-products in drinking water and low birth weight in children of women with polymorphisms in the CYP2E1 and C677T genes.
Research such as this raises tough regulatory issues for the EPA, Greer adds—is the EPA going to lower the standard of disinfection by-products to protect what might turn out to be a substantial group, or are they going to warn people and tell them to get tested for genetic susceptibility? “Our answer,” she says, “is that regulators need to protect the most susceptible individuals. You can’t tell people not to drink water or to buy bottled water.”
Microarray data may also be able to detect cellular activity in whole animals at levels far lower than those that cause discernible changes such as tumors or weight loss. In recent studies, researchers at the Microarray Center of the NIEHS National Center for Toxicogenomics (NCT) detected early indicators of mitochondrial damage before any adverse effect could be detected through traditional toxicity tests, says Microarray Center director Richard Paules. Such research doesn’t necessarily indicate that such low doses are toxic, says Paules, “but these gene expression changes could be an indication that higher or longer exposures have the potential to cause adverse effects and should be studied more closely.”
Some companies will not test a drug with a microarray experiment that has any chance of becoming part of a regulatory package. And then there are companies who view [submission of microarray data] as a positive, who say the FDA gets more information, we get more information, and we might find a positive aspect to our drug that we didn’t know about. –Kurt Jarnagin, Iconix Pharmaceuticals
Signatures of mechanisms of toxicity in different species may also improve researchers’ ability to compare the results of animal studies and human health outcomes. A chemical that causes, say, cancer in a rat may not have the same effect in people if the two species process the compound differently. Such research is especially important for the manufacturers of industrial chemicals, who do not test their products on humans, according to Jim Bus, director of external technology at The Dow Chemical Company and a member of the NRC committee.
Concerns Voiced
Although the use of microarray data and acceptance by regulators can be beneficial for pharmaceutical and other chemical manufacturers, many industry representatives still express concerns about the use of such data by the FDA and the EPA. One of these concerns—which runs contrary to the optimism expressed by government—is that submission of complex expression data will slow processing and approval of applications rather than streamline the process. Current FDA approval times for regular drug applications are about 18 months, down from 30 months several years ago. Approval for priority drugs with a public health benefit, such as AIDS medications, can take as little as 6 months, according to Woodcock.
Pharmaceutical and chemical companies have also expressed concern that the FDA and the EPA might overreact to microarray data. The primary issue, according to Daston, is determining what type of signature constitutes an adverse effect; the challenge is to distinguish adverse responses to a chemical exposure from homeostatic responses—that is, normal changes that may indicate a cell is disposing of a toxicant in a way that will not lead to lasting damage or that may not be related to the exposure at all.
Dearfield explains further: “Gene expression changes all the time. You can walk from a dark room into the sunlight, and you’re going to get all kinds of genomic signature changes. Is that bad? No—it’s the way the body normally operates. You need to sort out that kind of change from a change caused by an adverse stressor.”
Similarly, there is concern that the agencies will be “excessively reactive” to single gene changes, says Jarnagin. “Hypothetically you do a microarray expression on a potential drug’s effect, and lo and behold, the oncogene RAS is elevated five- or tenfold. Using a [toxicogenomics] database, you can see that there are many approved drugs that elevate RAS. Every drug in our database elevates at least one known oncogene. None of these drugs are known to cause cancer at therapeutic doses.”
Although the FDA draft guidance states that voluntary submissions of data will not be used for regulatory purposes, some companies still are reluctant to part with the results of exploratory microarray experiments. Some companies fear that proprietary data from one application will be used to judge data in another. In response, Woodcock says, “We cannot apply proprietary data to another application; we can’t make it public.” However, she says, regulators do learn from the reviews they conduct. And although they can’t directly compare data from one application to another, problems they see in one application might cause them to more carefully scrutinize another application with similar results.
Legal concerns include the potential for being sued if microarray data that couldn’t be interpreted at the time of submission later turn out to indicate toxicity in some people or under some conditions. The fear of lawsuits is such that some companies haven’t gone to the next stage in using microarrays for evaluating the effects of drugs under development, for either good or bad effects, says Roger Ulrich, president of Rosetta Inpharmatics, a subsidiary of Merck and Company.
Manufacturers of industrial chemicals are also concerned about EPA penalties. If a company discovers a previously unknown adverse effect for a given chemical, the company is required under TSCA to submit a report to the EPA within a few days, says Bus. The same is true if toxicity is detected at concentrations lower than previously found. “Say you’re dosing animals with a chemical where, historically, an effect has not been seen below a dose of ten milligrams per kilogram,” Bus explains. “You do another study and suddenly, you find a unique effect at one milligram per kilogram. Under TSCA, you’re required to report that.” If such a report were delayed because the significance of the microarray data wasn’t understood at the time of testing, manufacturers could conceivably face retroactive fines and penalties. If penalties are levied per day and a significant amount of time has passed, fines can be substantial, says Bus. Under TSCA, the EPA has the authority to levy fines of up to $27,500 per day for nondisclosure of required information.
Fine-tuning the Process
The FDA is working to alleviate some of industry’s concerns. To facilitate its ability to handle microarray data and keep approvals moving, the agency has collaborated with private firms on training exercises. Through a material transfer agreement, Iconix has given the CDER access to its proprietary relational toxicogenomics database, DrugMatrix, for evaluative and educational purposes, says Karol Thompson, molecular toxicology team leader in the CDER Division of Applied Pharmacology Research. The DrugMatrix database contains expression information related to more than 600 substances, including many approved medications. Iconix also led two workshops on microarray technology in February 2003 and January 2004 for members of the Nonclinical Pharmacogenomics Subcommittee of the CDER Pharmacology/Toxicology Coordinating Committee. The pharmacogenomics firm Gene Logic also has provided the FDA with expression data from its proprietary GeneExpress system database as part of a collaborative project with CDER research scientists and statisticians to identify endogenous genes that can serve as indicators of microarray sample quality.
In addition, the FDA worked with the company Expression Analysis on a mock submission using toxicology data developed by Schering-Plough Corporation for a candidate drug that did not go on to clinical trials. The submission included microarray data, histology data, clinical chemistry data, and phenotype data. The exercise served as a practice run to help the FDA understand the format and content of future drug submissions containing microarray data.
“I think the FDA, Expression Analysis, and Schering-Plough gained a tremendous amount from this collaboration,” says Steve McPhail, CEO of Expression Analysis, which provides commercial microarray testing, analysis, and data management services. “We gained a great perspective in working with the FDA and in beginning to understand their thinking on how this type of data should be formatted for future regulatory submissions. And I think the FDA gained value from the submission from our experience with lots of clients and users of data and the way that they need to become prepared for submission.”
Although regulators and industry are working hard to hammer out the issues around the submission of gene expression data, such submissions are still somewhat premature, says William Mattes, a researcher on the HESI effort and senior scientific director of toxicogenomics at Gene Logic. For example, researchers and regulators have not yet even decided how to report data. The ultimate goal is to “submit data in some tabular format that is computer-friendly and will allow regulators to crunch the data, analyze it with software,” he says. “We have not seen the FDA truly, openly discuss what data standards would be. . . . The issue is hugely in flux.”
Mattes serves on a committee on pharmacogenomics standards sponsored by the Interoperable Informatics Infrastructure Consortium, Health Level Seven, and the Clinical Data Interchange Standards Consortium, nonprofit organizations developing data standards for health care and clinical trials. According to Mattes, the joint committee is discussing high-level questions regarding the kind of data that should be included in microarray submissions. Other groups are promoting the use of specific data formats such as the MIAME (Minimum Information About a Microarray Experiment) standards for content, as well as the accompanying MAGE (MicroArray and Gene Expression) data format standards developed by the Microarray Gene Expression Data Society. The European Bioinformatics Institute, the NCT, and HESI have proposed definitions for MIAME/Tox, which would add toxicogenomics annotations to the basic MIAME content framework.
The chemical industry is not chomping at the bit to use toxicogenomics data. The status quo works better for them rather than a system where chemicals can be screened systematically. –Linda Greer, Natural Resources Defense Council
Agreement on data formats will do industry and regulators little good if experimental protocols are weak or inconsistent. The HESI studies found significant variation among results of microarray experiments that were caused by differences in procedures among participating laboratories, including different operating procedures for isolating and labeling mRNA samples, nonstandard settings on hardware and software, and differences in gene coverage and annotation across different technology platforms. Jarnagin and others say the standardization problems found in the HESI experiments, some of which were conducted 3–5 years ago, are not as serious now. “There’s been substantial advancement in the field in the last few years,” says Jarnagin.
The quality and consistency of microarray chips has improved since the HESI experiments were conducted, agrees Brenda Weis, who along with William Suk administers the Toxicogenomics Research Consortium (TRC), a component of the NCT. “The commercial products are particularly good,” she says. “The manufacturing is at a very high level.”
Testing different microarray types was an important part of initial standardization experiments by the TRC, which involves researchers at five academic centers across the country, as well as the NIEHS Microarray Center. The consortium’s work builds on the HESI studies by systematically addressing different steps of the microarray experiment to see where variability is most likely to be introduced, says Weis.
In the consortium’s first set of experiments, reported in the March 2004 toxicogenomics issue of EHP, the centers used a total of 12 different microarray platforms. In the multifaceted experiments, all six consortium centers used two common platforms: an oligo microarray manufactured at one of the centers and the commercial Agilent mouse microarray platform, developed by TRC investigators working collaboratively with Agilent and the NCT microarray resource contractor, Paradigm Genetics. There were also 10 other “resident” cDNA- or oligo-based platforms that were manufactured at and used by the individual centers.
Other variables addressed in the experiments have included the use of spike-in RNA and RNA reference samples, known sequences of RNA used as controls in microarray experiments (spike-in RNA is added to samples at a known concentration whereas reference RNA is kept separate from the samples but run through the same microarray experiment). The goal was “to see if they provided utility in helping us understand how the different platforms performed,” says Paules.
There are still other aspects of microarray analysis that can introduce variability into results, according to Weis. During the TRC studies, as during the HESI studies, researchers found that the way each individual center handled the RNA—including the labeling of the samples, the hybridization and wash conditions, and variables in the scanning and analysis—all had an impact on the eventual outcomes, says Paules. Results and recommendations for improving standardization have been submitted for publication.
Now that studies have addressed the technology, the consortium has begun another series of experiments focusing on the replication of genomic signatures. Each center will receive common reference RNA samples, Agilent microarray chips, and compounds (acetaminophen and its nontoxic isomer) to test using experimental animals. All of the centers will use standardized protocols for the microarray analyses. The hope, says Weis, is “to standardize the technical aspects of the experiment in order to address the issue of reproducibility of the biological response across multiple research groups. Whether or not we can do this successfully is important information for the regulatory community.”
Other groups that are studying method standardization include the External RNA Controls Consortium, a volunteer group sponsored by the National Institute of Standards and Technology. The group is working to develop methods to evaluate the performance of gene expression assays based on the measurement of external RNA controls, such as spike-in controls.
Standardization of animal models is another concern in microarray experiments. Researchers with the National Toxicology Program (NTP), an interagency organization based at the NIEHS, are studying changes in microarray results caused by homeostatic responses in Fisher 344 rats, one of the primary animal models used by the NTP. Results thus far, currently in press at Toxicologic Pathology, show differences in microarray signatures in samples taken from the left lobe of the liver compared to those from the median liver lobe of the same animal.
“You may get the same overall story from the two samples, but not the same number of genes or the same intensity of expression,” says Gary Boorman, a research scientist with the NTP and the NIEHS Environmental Toxicology Program, and a coauthor of the forthcoming paper. These results indicate that when labs coordinate their efforts, they should not only look at the technical issues, such as the microarray platforms each group is using, but also make sure that their methods for sampling animal models are uniform, says Boorman.
The NTP group is also studying variables including the time of day that tissue is collected, and the life stage and sex of the animal. The goal is to describe how normal variability in an animal strain can affect the interpretation of studies using microarray technology, says Nigel Walker, chair of the NTP’s toxicogenomics faculty and a staff scientist with the NIEHS Environmental Toxicology Program. “We’re trying to define ‘normal,’” says Walker, “so we know when the change in a gene is beyond the range of normal physiological variability.”
The Burden of Interpretation
Once results of microarray experiments are reproduced, scientists and regulators are still faced with the difficulty of interpreting genomic signatures. “There seems to be a lack of consensus on how data should be analyzed,” says Timothy Zacharewski, an assistant professor of biochemistry and molecular biology at Michigan State University and a member of the NRC committee.
For example, although the FDA draft guidance requires the submission of all data on “known valid biomarkers,” the agency currently does not recognize any genomic signatures as valid biomarkers, according to Leighton. “A lot of stuff has been published, but not all of it is of the same quality, even though it’s in the peer-reviewed literature, either because of a small population study or inadequate controls,” he says.
The FDA draft guidance does not specify how genomic signatures are to be validated as biomarkers. “It’s an important issue, and we’re discussing that,” says Leighton. “In the near future, we may need to come out with guidance on how to validate a genomic signature.” However, he says he doesn’t anticipate such guidance being issued soon; the agency doesn’t want to act in haste lest a less-than-optimal procedure be institutionalized prematurely.
There are logistical questions to consider. “How do you validate a safety biomarker, say for liver injury? You can’t run a clinical trial where you cause liver injury,” says Ulrich. Instead, scientists must compare genomic signatures to traditional toxicity tests using animals. But some traditional biomarkers can be subjective and often equivocal, says Ulrich. He cites the examples of alanine aminotransferase and aspartate aminotransferase, biomarkers of liver injury that also are occasionally associated with muscle injury. “Weight lifters and long-distance runners express [these enzymes],” he says. “They’re subjective biomarkers because they’re not liver-specific.” (Leighton notes, however, that the biomarkers industry and academia rely upon the most are less subjective. “Every lab uses the same core set,” he says, “and they’ve been in use for many years.”)
Private companies have little motivation to validate safety biomarkers, in part because they don’t know how they’ll be used, says one pharmaceutical representative who asked to remain anonymous. “It’s expensive to validate a genomic signature. And if we make the investment and develop a better biomarker for toxicity, all it will do is make it tougher to get approvals.” As a result, the bulk of validation efforts probably will be conducted and disseminated by nonprofit groups, academia, and government labs.
One leader in this area is the NCT. In addition to studies of experimental protocols and replicability, the NCT is also studying signatures generated by specific exposures. The NCT’s Microarray Center is focusing on liver toxicants such as acetaminophen [see “Phenotypic Anchoring: Linking Cause and Effect,” EHP 111: A338–A339 (2003)]. Other government facilities are contributing as well. EPA research into gene expression includes studies of sentinel species such as amphibians, fish, and aquatic microbes. The Department of Energy is using microarray experiments and other techniques to study microbial communities used in the remediation of toxic waste. In the nonprofit realm, the public ArrayExpress database of expression data, managed by the European Bioinformatics Institute, contains all the results from the HESI experiments as well as expression data from other studies.
There is also at least one commercial company that is contributing to public domain information on safety biomarkers. In March 2004, Iconix announced plans to publish five expression signatures of drug-induced toxicity in the liver, kidney, and heart. Information on one of the signatures, for injury to renal tubules, was presented that month at the annual meeting of the Society of Toxicology.
As industry and regulators wrestle with the intricacies of microarray data formats and submission, even more complex challenges loom: the data produced by proteomics and metabolomics research. “We’re well aware that metabolomic and proteomic data might be more important in the long term than the genomic data,” says Leighton. Among other reasons, samples are more readily available; it’s easier to collect blood and urine than to take a liver biopsy. “Then again,” says Zacharewski, “with proteomics and metabolomics you still have the problem of large, complex data sets of which only a fraction can be interpreted as being linked to any biological effect.”
Issues of standardization and validation will be similar for all of the “omics” technologies. So will tensions between concerns of industry and statutory obligations of regulators. That means many more meetings between industry and agencies. “We’re absolutely committed to not setting standards in isolation,” says Benson. “It is essential for the agencies to work together and with industry and academia when developing this regulatory framework.”
Regulatory Resources
GROUPS
FDA Center for Drug Evaluation and Research (CDER)
The CDER reviews applications for new prescription and over-the-counter drugs to ensure they been adequately tested and are safe for human use. The CDER also monitors drugs that are already on the market for unexpected health risks. http://www.fda.gov/cder/
National Research Council (NRC) Committee on Emerging Issues and Data on Environmental Contaminants
This committee provides a public forum for government, industry, environmental groups, and the academic community to discuss emerging evidence and issues in toxicogenomics, environmental toxicology, risk assessment, exposure assessment, and other related fields. http://dels.nas.edu/emergingissues/index.asp
Microarray Gene Expression Data (MGED) Society
This international group of biologists, computer scientists, and data analysts aims to facilitate microarray data sharing by establishing standards for data annotation and exchange, fostering the creation of microarray databases and related software implementing these standards, and promoting the sharing of high-quality, well-annotated data within the life sciences community. The group hopes to extend this mission to other “omics” technologies. http://www.mged.org/
DOCUMENTS
Guidance for Industry: Pharmacogenomic Data Submission
This guidance, issued by the FDA in draft form in November 2003, contains nonbinding recommendations on the submission of pharmacogenomics data during the drug application process. http://www.fda.gov/cder/guidance/index.htm [select “Pharmacogenomic Data Submissions” under the heading “Procedural (Draft)”]
Interim Policy on Genomics
This four-page policy paper outlines the EPA’s standing position on the relevance and use of genomics technologies in risk assessment. http://epa.gov/osa/spc/htm/genomics.pdf
Mini-Monograph: Genomics and Risk Assessment
This mini-monograph published in the March 2004 toxicogenomics issue of EHP includes recommendations for conducting studies and handling data based on studies by the Health and Environmental Sciences Institute/International Life Sciences Institute. http://ehp.niehs.nih.gov/txg/docs/2004/112-4/toc.html?section=toxicogenomics
Potential Implications of Genomics for Regulatory and Risk Assessment Applications at EPA
This March 2004 draft white paper by the EPA examines the possible impact of genomics technologies on agency regulatory activities. http://www.epa.gov/osa/genomics-external-review-draft.pdf
OTHER RESOURCES
ArrayExpress
This public database of expression data is managed by the European Bio-informatics Institute. http://www.ebi.ac.uk/arrayexpress/
External RNA Controls Consortium Workshop: Specifications for Universal External RNA Spike-In Controls
The External RNA Controls Consortium, sponsored by the National Institute of Standards and Technology, has posted presentations from this bioinformatics workshop on its website. The consortium is working to develop methods to evaluate the performance of gene expression assays based on the measurement of external RNA controls. http://www.cstl.nist.gov/biotech/workshops/ERCC2003/
MAGE (MicroArray and Gene Expression)
This page of the MGED Society website provides MAGE-related links, tools, and other resources. http://www.mged.org/Workgroups/MAGE/mage.html
MIAME (Minimum Information About a Microarray Experiment)
This page of the MGED Society website describes MIAME principles and requirements, and lists links to relevant tools and news. Also includes a document by the European Bioinformatics Institute, the NIEHS National Center for Toxicogenomics, and the Health and Environmental Sciences Institute/International Life Sciences Institute in which these groups propose definitions for MIAME/Tox, which would add toxicogenomics annotations to the basic MIAME content framework. http://www.mged.org/Workgroups/MIAME/miame.html
Drug data dilemma. Researchers and regulators alike are struggling with the complexities—and uncertainties—of toxicogenomics data.
Chemical conundrum. The EPA is moving cautiously toward considering toxicogenomics data in chemical regulation.
|
15345379
|
PMC1277130
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A678-A685
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.112-a678
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a00686EnvironewsScience SelectionsSupervised Sorting: Training Computers to Classify Toxicants McGovern Victoria 8 2004 112 12 A686 A686 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
A new application of an existing computer learning system can improve the use of gene expression profiles to classify toxicants to which an animal has been exposed, according to work published this month by Guido Steiner and colleagues at the pharmaceutical company F. Hoffmann–La Roche [EHP 112:1236–1248]. The authors write that these investigations could help separate transcriptional changes that are of relevance for the mode of toxicity from mere bystander effects—coincidences that have little predictive value and that can be amplified by other computational approaches.
Predicting how any given toxicant will affect an organism is possible if similar compounds produce comparable changes in gene expression. Identifying useful panels of genes whose expression profiles change predictably in response to toxicants is a major goal for predictive toxicology. Finding effective ways to sort, interpret, and anticipate changes in expression is critical for moving these observations into a practical understanding of toxic effects. Background “noise” and the natural variation between experimental animals as well as compound-related characteristics in pharmacologic and toxic action complicate matters by creating variability in the analyzed data.
Other approaches to sorting gene expression data for toxicogenomics have included several “unsupervised” methods, in which modeling programs search for patterns within data and generate models of toxicity without being given any hint as to what kinds of expression patterns the researchers expect to find. The strength of this approach is that it allows for unbiased data exploration. On the other hand, it is not guaranteed to primarily retrieve information that is relevant for addressing the scientific problem at hand.
The approach reported by Steiner and colleagues is different. This group has applied a “supervised” method to toxicogenomics using what are called support vector machines (SVM). SVMs—which are computational tools, not physical machines—take advantage of additional data in the form of pathology and serum measurements that are fed into the algorithm. These data are used to assign gene expression profiles to specific modes of toxicity. The SVM identifies the most relevant information for discriminating among the given samples. After learning from a “training set” of biological samples, the model should be able to correctly classify new samples exposed to compounds that the SVM has not encountered before. Therefore, the method has to construct classification rules that still work with data different from the initial training data.
Steiner and colleagues used an SVM to find classification rules connecting patterns of gene expression in response to a series of known or suspected hepatotoxicants. The predictive genes were picked using another computational tool, recursive feature elimination (RFE), which is an integral part of SVM creation. In RFE, the computer produces a ranking of all the features that it uses to define a fingerprint—in this case, the expression profiles of each gene on a microarray. Then it calculates how much each feature contributed to that fingerprint. Uninformative or redundant features tend to be eliminated in an iterative process as less relevant, allowing refinement of the fingerprint’s definition to include only its most reliable features. These compact signatures can then be used to identify the class of toxicant to which an animal has been exposed.
In testing the system, the authors looked at 28 hepatotoxicants and 3 nonhepatotoxic compounds. Looking in rats, they laid out gene expression profiles, clinical chemistry, hematology, and histopathology for the different chemicals at various time points following exposure. In addition to discriminating between compounds that are hepatotoxic and ones that are not, their predictive models were in most cases also able to predict what kind of toxicant the animals had been exposed to—a direct-acting one that causes damage itself, a cholestatic one that interferes with bile, or a steatotic one that drives buildup of fat in the liver. By the same strategy, the SVM was able to recognize animals that had been exposed to the hepatotoxicant galactosamine but failed to respond with the typical necrosis and inflammation of the liver.
Pharmacologic activity can alter gene expression in the liver without necessarily signaling hepatotoxicity. The SVM correctly identified 3 tested pharmacoactive compounds as nonhepatotoxic and also correctly identified 3 hepatotoxic compounds whose mechanisms of toxicity were not included in the data sets used to train the machines. In two out of three cases, the SVM was able to correctly identify the general mode by which these compounds were toxic.
The models were extended to unknown rat strains, as well. After identifying expression profiles in Wistar rats induced by several peroxisome proliferator–activated receptor (PPAR) agonists, the SVM was used to look at data for the livers of Sprague-Dawley rats exposed to another PPAR agonist, WY14643. The SVM was able to correctly recognize both Sprague-Dawley rats exposed to WY14643 and the control animals, and could predict that treatment with WY14643 would stimulate peroxisome proliferation.
So far, the work has dealt predominantly with toxicant concentrations that yield substantial and largely unambiguous effects. To optimize the system to more accurately predict subtle changes, toxicologists and bioinformaticians will need both further improvements in computational methods and a larger database linking compounds and their effects on gene expression.
Old computers learn new tricks. Computational tools known as support vector machines discern relevant gene expression data from samples and apply what they learn in order to classify new compounds.
|
0
|
PMC1277131
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A686
|
utf-8
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a00687EnvironewsScience SelectionsIt’s All in the Interaction: Quantitating Gene Networks Potera Carol 8 2004 112 12 A687 A687 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Toxicologists who use microarrays hope to uncover relationships that link gene expression data to signal transduction pathways, gene networks that are often used to describe the sequence of biochemical events controlling cellular function. The large quantities of data generated by microarray studies generally are examined qualitatively—for example, by comparing whether one gene is turned on relative to another. These qualitative relationships, however, fail to describe how genes in a network influence each other. Still in their infancy are tools that quantitate the complex relationships within gene networks more comprehensively than simple correlations between pairs of genes. Now, for the first time, researchers describe a new quantitative statistical technique that assesses the interactions of genes in a network [EHP 112:1217–1224].
The team, led by Hiroyoshi Toyoshiba of the NIEHS Laboratory of Computational Biology and Risk Assessment, created a statistical software program that verifies concurrently that the expression of one gene is linked to the expression of several others. The first proof-of-concept demonstration evaluated genes that are directly responsive to tetrachlorodibenzo-p-dioxin (TCDD; a ubiquitous environmental pollutant and known human carcinogen) and their effect on the retinoic acid signal transduction pathway.
Signal transduction pathways respond to different environmental conditions; they are like molecular circuits that detect and integrate diverse external signals to alter gene transcription. This results in changes in enzyme activities as well as the production of abnormal levels of proteins, which further results in changes in biochemical processes. Alterations in signal transduction pathways can lead to cancer and other disorders.
Toyoshiba and colleagues had earlier identified genes that are altered in lung airway epithelial cells after exposure to TCDD. Starting with microarrays composed of 2,000 genes that are known to be expressed in response to environmental toxicants, the researchers had identified 11 genes that responded significantly to TCDD in two different lung cell lines. These genes appeared to be involved in the effects of TCDD on the retinoic acid signal transduction pathway.
The researchers constructed a hypothetical model of the retinoic acid signal transduction pathway that describes how the 11 genes interrelate. Based on published reports on retinoic acid metabolism, the model postulated that dietary vitamin A (retinol) is converted first to retinal and then to retinoic acid by alcohol dehydrogenases and, possibly, by cytochrome P450 enzymes. Once synthesized, retinoic acid enters the cell nucleus. There, it binds retinoic acid receptor beta, which, in turn, alters the expression of genes that may play a role in tumor formation. The hypothetical model included genes that produce three alcohol dehydrogenases, a cytochrome P450 enzyme, retinoic acid binding proteins and receptors, and four nuclear proteins.
Following exposure to three concentrations of TCDD, the expression levels of the 11 genes were calculated relative to unexposed controls. Statistical methods were applied to these data to test the hypothetical linkages between TCDD-responsive genes and the retinoic acid signal transduction pathway. These tests confirmed strong linkages between the genes included in the hypothetical model.
Epidemiological studies show a strong association between TCDD and lung cancer; the model offers a potential explanation for how TCDD damages the lungs. TCDD appears to activate genes associated with the synthesis of retinoic acid, which—through the retinoic acid signal transduction pathway—turns on nuclear genes that promote cell proliferation and carcinogenesis. Scientists can focus future experiments on particular genes directly related to TCDD-induced tumor progression.
The new statistical tool makes it possible to understand biological pathways in cells, tissues, organs, and whole organisms. It can be expanded to include other relevant data, such as protein levels in cells. These data can be combined with pharmacological models to present a true systems biology approach to quantifying risks from exposure to xenobiotics such as TCDD, suggest the authors. Other researchers can obtain the statistical software by contacting laboratory director Christopher Portier at [email protected].
Notating networks. A new statistical package goes beyond qualifying interactions between a single pair of genes to describe how multiple genes within a network influence expression.
|
0
|
PMC1277132
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A687
|
utf-8
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a00690AnnouncementsFellowships, Grants, & AwardsFellowships, Grants, & Awards 8 2004 112 12 A690 A695 2005Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
NHLBI Clinical Proteomics Programs
This Request for Application (RFA) will establish Clinical Proteomics Programs to promote systematic, comprehensive, large-scale validation of existing and new candidate protein markers that are appropriate for routine use in the diagnosis and management of heart, lung, blood, and sleep diseases. These programs will facilitate validation of protein panels that may be used to predict disease susceptibility or to assist in differential diagnosis, disease staging, selection of individualized therapies, or monitoring of treatment responses. In addition, this RFA seeks to establish a high quality education and skills development program to encourage and ensure that scientists develop competencies and expertise needed to address the complex, multifaceted challenges in clinical proteomics.
Heart, lung, blood, and sleep diseases are major causes of morbidity and mortality. Cardiovascular disease is the number one killer in the United States in both men and women, across all major racial groups and totals nearly one million deaths a year. Lung diseases such as chronic bronchitis, emphysema, asthma and other obstructive or interstitial conditions account for more than 230,000 deaths annually, placing an enormous burden on our healthcare system. Blood diseases such as venous thrombosis and pulmonary embolisms are causes of significant public health concern, as well. Sleep disorders and insufficient sleep represent severe health concerns for tens of millions of Americans.
Improving patient care through the use of protein markers is well established clinically. For example, the definition of heart attack, as well as the determination of the benefit derived from antithrombotic treatments, rests on serum troponin measurement. The detection of extremely small quantities of this protein identifies patients at high risk for adverse outcomes as well those that will derive greater benefit from antithrombotic and other interventional strategies. Assay of the B-type natriuretic peptide also contributes to standard clinical information in the diagnosis of congestive heart failure. Myeloperoxidase was recently shown to help in the diagnosis of atherosclerosis and acute coronary syndromes.
The predictive values, sensitivity, and specificity of many of the individual protein markers, currently in clinical use, could potentially be enhanced if analyzed and measured in a panel. Observational studies have shown that combining protein markers troponin I, C-reactive protein and B-type natriuretic peptide into panels can provide valuable information on stratifying risk for acute coronary syndromes. Panels of protein markers, appropriately validated, could facilitate better and earlier diagnosis, improve disease staging and selection of individual therapies and lead to more reliable monitoring of treatment responses, leading to substantial improvements in public health.
The application of proteomics in the clinical environment is limited due to a lack of knowledge regarding which proteins are most useful for analysis and how data are interpreted and represented. Important research needs include the identification of panels of protein markers that are likely to provide useful clinical information, design of practical assays for these panels, and validation of these panels and assays in well characterized populations of human subjects. The emergence of clinical proteomics promises major advances in disease management, provided that a continuous channel exists for translating protein discoveries into tangible clinical benefits.
The purpose of this RFA is to establish an infrastructure for research teams to validate protein panels and to measure multiple candidate markers accurately, for heart, lung, blood, and sleep diseases. The Clinical Proteomics Programs established for this purpose will design panels of candidate proteins for disease areas, develop high throughput analytic methods, assess the predictive value of these proteomic measurements using biological specimens and clinical data from existing study populations, and establish procedures and standards for quality control.
A major shortfall of clinical proteomics is the lack of a robust infrastructure for clinical candidate panel validation. Validation is necessary to confirm the relationship to the target disease in large numbers of patient samples and requires highly standardized protein measurement systems. The samples must be derived from well characterized sample sets with associated high-quality clinical information. The validation process provides the critical evidence necessary for translating protein knowledge into practices impacting public health. A significant opportunity now exists to enhance the validation stage and help translate protein discoveries into clinical practice. Many completed and ongoing clinical trials and epidemiologic studies have disease associated biological samples in addition to detailed clinical data. Leveraging this investment will enhance validation efforts.
Panels of protein markers in the following areas would represent appropriate topics for proposed projects. This list is not intended to be all-inclusive, and other topics should be considered. 1) Predict susceptibility to coronary artery disease or acute and chronic pulmonary disease; 2) Assess the severity and rate of progression of atherosclerosis or pulmonary disease; 3) Differential diagnosis for patients presenting with shortness of breath, chest pain or elevated blood pressure; 4) Detect occult myocardial infarction and subclinical cardiac disease and/or damage; 5) Select optimal, individualized medical management strategies; 6) Monitor therapeutic and adverse responses to antihypertensive drugs or drugs for asthma and other lung diseases such as inhaled corticosteroids, bronchodilators, and leukotrienes; 7) Identify early stages of pulmonary disease before significant pathogenesis has occurred; 8) Evaluate risk of thrombosis in individuals with a predisposition to cardiovascular disease or stroke; 9) Evaluate risk of bleeding and appropriate management strategies in patients with bleeding disorders - hemophilia, autoimmune blood disorders, von Willebrand disease; 10) Manage anticoagulation therapy in patients with thromboembolic disorders; 11) Identify markers for early diagnosis and prognosis of heart, lung, blood, and sleep disorders; 12) Develop tests to rapidly and accurately distinguish thromboembolic stroke from hemorrhagic stroke.
Projects outside the scope of this RFA will not be considered responsive and include: 1) Studies that do not address heart, lung, blood, or sleep disorders; 2) Studies that are focused on developing new proteomic technologies to identify protein markers 3) Proteomic discovery efforts.
We encourage inquiries concerning this RFA and welcome the opportunity to answer questions from potential applicants. It is highly recommended that prospective applicants contact program staff (please see the “Contact” section below) about proposed projects.
A Clinical Proteomics Program should be an identifiable organizational unit formed by a single institution or a consortium of cooperating institutions. Each Clinical Proteomics Program must provide a multidisciplinary team structure, ensuring effective coordination and integration between the selection and validation components of the Program. The team should encompass multi-disciplinary expertise and should include proteomic researchers, bio-engineers, clinical chemists, protein chemists, experts in biostatistics and bioinformatics, clinical investigators, and epidemiologists.
The marker selection process should focus on the design of protein marker panels that are most useful in clinical situations with under met needs. The team should primarily be responsible for prioritizing candidate protein markers and panels for validation. The selection component should actively develop candidate protein marker panels from a wide range of sources, such as proteomic discovery efforts, published reports, differential expression based research studies and in silico sequence-based predictions.
The use of biological samples obtained by minimally invasive methods (e.g., blood, sputum, and urine) is encouraged. Samples from ongoing studies can also be used provided appropriate Institutional Review Board (IRB) amendments to existing protocols have been obtained.
Since quantitative measurements of candidate markers in large and well defined clinical samples is central to the validation effort, criteria for the selection of the source material as well as the criteria for validation of the candidate markers must be specified. Where possible, existing technology platforms should be explored as multiplexing tools during panel development. Efforts to minimize sample consumption are encouraged to ensure the maximum number of assays. Emphasis will be placed on development of panels with high predictive value, specificity, and sensitivity; development of flexible assay protocols to accommodate the inclusion of newly identified proteins into ongoing validation efforts; refinement and development of innovative biostatistical tools and methods for selection of protein marker panels and for increasing the diagnostic sensitivity and specificity; and assay development applicable to clinical settings.
The multidisciplinary team will also evaluate pre-analytic issues, (including those relating to sample collection, storage, processing, and handling), and set criteria, standardize, and implement preanalytic protocols prior to validation. Each Clinical Proteomics Program should have access to characterized samples with well defined clinical data and the appropriate IRB approvals before funding. Furthermore, they should operate on an ‘open source’ model system, making the data, statistical and bioinformatic tools that are generated and developed in the programs, accessible to the public domain within a time period to be determined at the first meeting of the Inter-Program Steering Committee.
An Inter-Program Steering Committee (with membership from all the programs) will be appointed and will have scientific management oversight and responsibility for developing communication, coordination and collaboration among the Programs. In addition, there will be an External Scientific Panel, advisory to the National Heart, Lung and Blood Institute (NHLBI) that will evaluate the progress of the Clinical Proteomics Programs.
In order to facilitate the functions that are common to each program, one of the programs will be selected to function as an Administrative Coordinating Center (ACC) for all the programs. Therefore, applicants must include as a separate section in their proposal, a description of an Administrative Coordinating Center that will be reviewed separately, independent of the scientific application. Specification for the ACC application can be found under the section, “Packaging the Clinical Proteomics Program Application”.
Each program is expected to develop mechanisms towards education of skills necessary for clinical proteomics. Full implementation of a nationwide effort in translational research for clinical proteomics requires availability of trained M.D., M.D. /Ph.D., and Ph.D. scientists. These individuals must be knowledgeable about the diverse aspects of clinical proteomics and able to integrate the translational and clinical concepts necessary for application to heart, lung, blood, and sleep diseases. One unique feature of the Clinical Proteomics Program is to function as a spring board for advancing education, at the National level, by establishing various mechanisms, such as specialized short courses, and ‘hands on’ programs that will focus on guiding graduate students, trainees, technical personnel, M.D./Ph.D. and Ph.D. scientists in translation research for clinical proteomics. Both Clinical Proteomics Program and NHLBI-supported investigators would be eligible for these educational opportunities.
This RFA will use the National Institutes of Health (NIH) cooperative agreement (U01) award mechanism. In the cooperative agreement mechanism, the Principal Investigator retains the primary responsibility and dominant role for planning, directing, and executing the proposed project, with NIH staff being substantially involved as a partner with the Principal Investigator, as described under the section "Cooperative Agreement Terms and Conditions of Award"
Prospective applicants are asked to submit a letter of intent that includes the following information: descriptive title of the proposed research; name, address, and telephone number of the Principal Investigator; names of other key personnel; participating institutions; number and title of this RFA. Although a letter of intent is not required, is not binding, and does not enter into the review of a subsequent application, the information that it contains allows NHLBI staff to estimate the potential review workload and plan the review.
Applications must be prepared using the PHS 398 research grant application instructions and forms (rev. 5/2001). Applications must have a Dun and Bradstreet (D&B) Data Universal Numbering System (DUNS) number as the Universal Identifier when applying for federal grants or cooperative agreements. The DUNS number can be obtained by calling 866-705-5711 or through the web site at http://www.dunandbradstreet.com/. The DUNS number should be entered on line 11 of the face page of the PHS 398 form. The PHS 398 is available at http://grants.nih.gov/grants/funding/phs398/phs398.html in an interactive format. For further assistance contact GrantsInfo, 301-435-0714, email: [email protected].
Each application to establish a Clinical Proteomics Program must be submitted as one application by a Clinical Proteomics Program Director, who will be responsible for organizing and maintaining effective integration and interaction of the program. A clear description of interaction among the various components, plans for communication, collaboration and sharing among investigators in the Clinical Proteomics Program should be included. The Clinical Proteomics Program Director should also indicate the mechanism for handling day-to-day administrative details, program, coordination, planning and evaluation. The director will be required to have a minimum of 25 percent level of effort, and the responsibility of oversight and coordination of all projects or components of the Program, whether or not they are at his/her institution. Each program should clearly outline its administrative and organizational structure.
Applications should include appropriate budget forms providing adequate budget justification with all applicable direct and facilities and administrative costs. Estimating of staffing needs, including principal investigator, other professional and support staff must be included. During the course of the project period, it is anticipated that technologies will improve and the proposed studies may change. Accordingly, it is expected that the principal investigators will be allowed adjustments in their scientific projects to accommodate such things. Budgets should include travel costs for Awardees Meetings and Inter-Program Steering Committee Meetings, as detailed under the section titled, “Special Requirements” along with statements indicating willingness to participate in these meetings and abide by its governance.
An educational component is another integral part of each Clinical Proteomics Program. A clear description of the efforts to educate and cross train across disciplines of clinical proteomics must be outlined, including the plans for developing short courses and ‘hands on’ programs. The process of selection and monitoring of candidates for these educational activities must be portrayed as well.
A separate section not exceeding 5 pages, detailing plans for an Administrative Coordination Center, should be included in each Clinical Proteomics Program application. This section should be placed following the section on the Research Plan. The Center will facilitate functions common to all the Clinical Proteomics Programs, coordinate meetings of the awardees, the Inter-Program Steering Committee and the External Advisory Panel, and manage a Clinical Proteomics Program intranet website. The Center will also be responsible for setting up the monthly conference calls of the Steering Committee. This section should also include separate budget justification pages for the operation of the Administrative Coordination Center not to exceed 100,000 direct costs in any year. Applications should provide adequate budget justification with all applicable direct and facilities and administrative costs, including estimated costs associated with the travel of the External Advisory Panel (6–8 members). Estimation of staffing needs and communication costs must be included. The award will be subject to administrative review annually.
Applications not conforming to these guidelines will be considered unresponsive to this RFA and will be returned without further review.
The RFA label available in the PHS 398 (rev. 5/2001) application form must be affixed to the bottom of the face page of the application. Type the RFA number on the label. Failure to use this label could result in delayed processing of the application such that it may not reach the review committee in time for review. In addition, the RFA title and number must be typed on line 2 of the face page of the application form and the YES box must be marked. The RFA label is also available at: http://grants.nih.gov/grants/funding/phs398/labels.pdf.
The Center for Scientific Review (CSR) will not accept any application in response to this RFA that is essentially the same as one currently pending initial review, unless the applicant withdraws the pending application. However, when a previously unfunded application, originally submitted as an investigator-initiated application, is to be submitted in response to an RFA, it is to be prepared as a new application. That is, the application for the RFA must not include an Introduction describing the changes and improvements made, and the text must not be marked to indicate the changes from the previous unfunded version of the application.
Letters of intent are due 17 September 2004. Applications are due 14 October 2004. The earliest anticipated start date is July 2005.
Contact: Pothur R Srinivas, Division of Heart and Vascular Diseases, NHLBI, 6701 Rockledge Drive, Rm 10188, Bethesda, MD 20892-0001 USA, 301-435-0550, fax: 301-480-2858, email: [email protected]; Anne P. Clark, Chief, Review Branch, Division of Extramural Affairs, NHLBI, 6701 Rockledge Drive, Rm 7214, MSC 7924, Bethesda, MD 20892-7924 USA, Bethesda, MD 20817 (for express/courier service), 301-435-0270, fax: 301-480-0730, e-mail: [email protected].
Reference: RFA No. RFA-HL-04-019
New Technology for Proteomics and Glycomics (SBIR/STTR)
Notice: this program announcement (PA) must be read in conjunction with the current Omnibus Solicitation of the National Institutes of Health (NIH), Centers for Disease Control and Prevention (CDC), and Food and Drug Administration (FDA) for Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) Grant Applications. The solicitation (see http://grants.nih.gov/grants/funding/sbirsttr1/index.pdf or http://grants.nih.gov/grants/funding/sbirsttr1/index.doc) contains information about the SBIR and STTR programs, regulations governing the programs, and instructional information for submission. All of the instructions within the current SBIR/STTR Omnibus Solicitation apply.
The principal limitations in the field of proteomics are technological in nature. Proteomics, and the sub-discipline of glycomics, are rapidly developing, technology-intensive fields. Separations, mass spectrometry, microarray, bioinformatics, and other tools have advanced rapidly to support the explosive growth of biomedical applications in this area. However, technologies and methods remain largely inadequate to address the majority of meaningful biological problems, particularly with respect to quantitative and real time measurements. Continued intensive development of advanced tools is essential to meet two needs. First, improvements in basic bioanalytical technologies are essential to these endeavors. This includes but is not restricted to robotics, sample preparation and pre-fractionation, analytical separations, gel and array imaging, quantitation, mass spectrometry, intelligent automated data acquisition, and database searching. Second, improved informatics technologies are essential for the conversion of data into meaningful results and interaction models. Improved informatics tools will also facilitate the integration and synergistic development of the basic analytical tools mentioned above. Additionally, the translation of advances in proteomics to a clinical setting should be a priority.
Proteomics is a rapidly expanding field. Many of the potential scientific and medical rewards of proteomics’ successful application to complex systems seem deceptively near. A broad range of technologies is evolving rapidly to meet the needs of the field. However, despite explosive growth in both academic and commercial efforts, concrete technical capabilities are far from adequate to realize this promise. Proteomics technologies and methods in the three broad, interacting domains of biology, analytical chemistry, and informatics are still largely inadequate to address the bulk of challenging biological problems. This is the case with respect to both core capabilities and scale.
The broad scope of proteomics might perhaps be broken down into six types of questions that are addressed in some form: (1) identification of individual proteins, (2) recognition of protein interactions, (3) relative quantitation to distinguish differential expression of proteins, (4) characterization of post-translational modifications, (5) qualitative or quantitative measurements at high spatial and/or temporal resolution to address the dynamics of protein interactions, and (6) formulation of models based on results from components 1–5.
The categories above define the type of information being sought, and imply the need for technologies capable of addressing the challenges inherent in each type of experiment. Those specific technologies may reside within any of the three domains that define proteomics, or may function as a bridge between them. For example, tools for tissue or subcellular fractionation may reside squarely in the biological domain, but could also be designed in such a way as to maximize synergy with widely used analytical separations methods.
It is important that in a field as complex and interdisciplinary as proteomics, technology development be pursued with a sound understanding of context. One area of particular interest is the development of technologies that will permit observations to be quantitative and made in real time, whether for clinical studies or experimental systems.
In addition to the development of broadly applicable research tools that address the core technical challenges in proteomics, unique constraints in two subordinate areas merit special attention. We especially encourage applications in response to this announcement that address the unique needs of glycomics and clinical proteomics, described below.
The application of proteomics tools in the clinical setting lags far behind their use in basic science and drug discovery. Though this is not due solely to technological constraints, the unique challenges associated with development of simple, rapid, and robust technologies for the clinic demand a somewhat different perspective than might be taken in consideration of a purely research-driven project. Likewise, this difference in perspective and priorities should open the possibility of approaches that might be wholly inadequate from a research perspective but may be appropriate in the clinic. Finally, the exploitation of insights previously developed in research-oriented proteomics to develop more specific, robust tools for clinical applications is also an appropriate goal.
The complexity and diversity of glycosylation significantly complicates the linkage between genetic sequence and mature, active proteins. Glycobiology-focused proteomics, or glycomics, requires the development of novel approaches and tools directed at the special challenges of glycobiology. Among post-translational modifications, glycosylation is the only one that requires structural characterization of the modifying moiety beyond noting its presence. Strategies for separation, profiling, quantitation, and detailed characterization of carbohydrate structures are central challenges. Informatics tools are needed for data handling and reduction, correlation of carbohydrate and protein information, and a variety of other purposes. Discovery-based analytical tools that can survey the complexities of glycosylation on a system-wide basis may have significant biological impact.
The goals of this PA are deliberately discussed with respect to fundamental challenges, rather than in relation to specific technologies, in order to emphasize the overriding importance of surmounting obstacles, irrespective of the analytical strategy adopted to pursue those solutions. This solicitation is open to unconventional or alternative approaches.
This PA uses the SBIR and STTR mechanisms, which are set-aside programs. As an applicant, you will be solely responsible for planning, directing, and executing the proposed project. Future unsolicited, competing- continuation applications based on this project will compete with all SBIR/STTR applications and will be reviewed according to the customary peer review procedures.
This PA uses just-in-time concepts. It also uses the modular budgeting format. Specifically, if you are submitting an application budget of $100,000 total costs (direct, F&A and fee) or less, use the modular format and instructions as described in the current SBIR/STTR Omnibus Solicitation. Otherwise follow the instructions for non-modular budget research grant applications. This program does not require cost sharing as defined in the current NIH Grants Policy Statement at http://grants.nih.gov/grants/policy/nihgps_2003/NIHGPS_Part2.htm#matching_or_cost_sharing.
Applications may be submitted for support as Phase I STTR (R41) or Phase I SBIR (R43) grants; Phase II STTR (R42) or Phase II SBIR (R44) grants; or the SBIR/STTR FAST-TRACK option as described in the SBIR/STTR Omnibus Solicitation. Phase II applications in response to this PA will only be accepted as competing continuations of previously funded NIH Phase I SBIR/STTR awards. The Phase II application must be a logical extension of the Phase I research but not necessarily a Phase I project supported in response to this PA.
The PHS 398 research grant application must be used for all SBIR/STTR Phase I, Phase II and Fast-Track applications (new and revised.) Effective October 1, 2003, applications must have a DUN and Bradstreet (D&B) Data Universal Numbering System (DUNS) number as the Universal Identifier when applying for federal grants or cooperative agreements. The DUNS number can be obtained by calling 866-705-5711 or through the website at http://www.dunandbradstreet.com/. The DUNS number should be entered on line 11 of the face page of the PHS 398 form. The PHS 398 is available at http://grants.nih.gov/grants/funding/phs398/phs398.html. Prepare your application in accordance with the SBIR/STTR Omnibus Solicitation and the PHS 398. Helpful information for advice and preparation of the application can be obtained at: http://grants.nih.gov/grants/funding/sbir-grantsmanship.pdf. The NIH will return applications that are not submitted on the 5/2001 version of the PHS 398. For further assistance contact GrantsInfo, 301-435-0714, e-mail: [email protected]. The title and number of this PA must be typed on line 2 of the face page of the application.
The CSR will not accept any application in response to this PA that is essentially the same as one currently pending initial review unless the applicant withdraws the pending application. The CSR will not accept any application that is essentially the same as one already reviewed. This does not preclude the submission of a substantial revision of an unfunded version of an application already reviewed, but such application must include an Introduction addressing the previous critique.
Receipt and review schedule: see http://grants.nih.gov/grants/funding/sbirsttr_receipt_dates.htm.
Contact: Douglas M. Sheeley, Division of Biomedical Technology, National Center for Research Resources, 6701 Democracy Blvd, MSC 4874, Bethesda, MD 20892-4874 USA, 301-435-0755, fax: 301-480-3659, e-mail: [email protected]; Pamela A. Marino, NIGMS, Rm 2As.43k, Natcher Building, Bethesda, MD 20892-6200 USA, 301-594-3827, fax: 301-480-2802, e-mail: [email protected]; Susan E. Old, Division of Heart and Vascular Disease, NHLBI, 6701 Rockledge Dr, MSC 7940, Bethesda, MD 20892-7940 USA, 301-435-1802, fax: 301 480-1335, e-mail: [email protected]; Danilo A. Tagle, Neuroscience Center, NINDS, Rm 2133, 6001 Executive Blvd, Bethesda, MD 20892-0001 USA, 301-496-5745, fax: 301-402-1501, e-mail: [email protected].
Reference: PA No. PA-04-089
Intellectual Property Rights in Genetics and Genomics
The purpose of this RFA is to encourage the study of the role of laws and policies regarding intellectual property rights in genetics and genomics research and development, and the effect of such laws and policies on progress in these fields and on commercialization, drug development, health care delivery, and the public health.
Since its inception, the Human Genome Project has attempted to follow a policy of free and open access to genetic and genomic data e.g., National Human Genome Research Institute (NHGRI) Policy Regarding Intellectual Property of Human Genomic Sequence (April 9, 1996), http://www.genome.gov/10000926; NHGRI Policy on Human Genomic Sequence Data (Dec. 21, 2000), http://www.genome.gov/10000910. The National Institutes of Health (NIH) policy recognizes the appropriateness of intellectual property protections for discoveries that are associated with useful products, but promotes the free dissemination of research tools whenever possible, especially when the prospect of commercial gain is remote (Report of the National Institutes of Health (NIH) Working Group on Research Tools, http://www.nih.gov/news/researchtools/).
Over the past three decades, however, many patents have been granted on gene sequences and other types of basic information derived from genetic sequence. For some, this has generated apprehension that gene patents are being granted too broadly or freely, especially for foundational tools. The concern is that the too-liberal issuance of such patent rights, especially when coupled with exclusive licensing practices, will result in the imposition of reach-through restrictions or excessive fees, and inhibit investigators from conducting additional research with these tools. This, it is feared, will ultimately be to the detriment of advances in medical research and to public health.
In January 2001, partly in response to a letter from the NIH urging the implementation of stricter criteria for the issuance of biotechnology patents, the U.S. Patent and Trademark Office revised its guidelines to patent examiners regarding patents on DNA sequence and sequence-derived intellectual property, effectively “raising the bar” on utility standards in this area [U.S. Patent and Trademark Office, Utility Examination Guidelines, Fed Reg 66(4) (January 5, 2001)]. However, questions remain about whether this revision raised the “bar” high enough to serve the public interest. An example of the potential problem is the recent acquisition and aggressive pursuit by Genetic Technologies Limited (GTG), an Australian company, of exceptionally broad global patent protection covering the use of information to derive risks of disease in all non-coding regions of the genome [see Nature (2003) 423:105]. While this is perhaps an extreme example (and the validity of GTG’s patents has not yet been tested in the courts), other controversial cases can also be cited (e.g., the Myriad Genetics BRCA1 patent, the University of Miami Canavan disease patent, the CCR5 HIV co-receptor gene patent). Such cases are increasingly leading genetics and genomics researchers, business entities, health care providers, and consumers to question how the balance between providing intellectual property protection and fostering biomedical innovation can best be attained.
Issues regarding the appropriate scope of protection for intellectual property rights in genetics and genomics research and development will only increase in complexity as progress in these fields continues. For example, large-scale proteomics efforts [such as protein biomarker discovery projects, the NIGMS Protein Structure Initiative (http://www.nigms.nih.gov/psi/) and initiatives to characterize protein-protein interactions] will generate new types of potentially patentable information, and with this information, new intellectual property challenges. Such challenges will also arise in several areas of research being emphasized under the new NIH Roadmap Initiative (http://nihroadmap.nih.gov/). For example, in the “chemical genomics” area, questions will arise about whether patents should be filed on the compounds that will be discovered or whether to place such compounds in the public domain, and about how pricing should be determined should a compound discovered through this process end up as a drug. In the bioinformatics and computational biology area, questions will arise about how best to promote the widespread distribution of new software to be developed (e.g., using an open source model of licensing or some other model).
Anticipating the growing need to confront questions of this type, the NHGRI has identified addressing intellectual property issues as one of the “Grand Challenges” for the future of genomics. Specifically, the Institute’s document “A Vision for the Future of Genomics Research,” [Nature (2003) 422:835–847], also available at: http://www.genome.gov/11006873), called for “the development of policy options in the area of intellectual property that will facilitate the widespread use of genetic and genomic information in both research and clinical settings.” To be maximally informed and effective, however, the development of such policy options must be based on a solid and broad-based body of theoretic and empiric data. While a number of studies already conducted or now underway provide a good preliminary foundation on which to build, there is a clear need for additional research and scholarship in this area.
In 2004, the Board on Science Technology and Economic Policy (STEP Board) and the Science, Technology, and Law Program of the National Academies of Sciences convened a committee on Intellectual Property in Genomic and Protein Research and Innovation (the “NAS Committee”). The NAS Committee’s charge is to review the patenting and licensing of human genetic material and proteins and their implications for biomedical research, therapeutic and diagnostic products, and medical practice. The NAS Committee is expected to release its report in the Summer of 2005, but there will clearly be a need for other, more in depth, examinations and analyses of these issues, by investigators from a broad range of disciplines.
To assist in addressing this need, the NHGRI proposes a new initiative to encourage the study of the role of laws and policies regarding intellectual property rights in genetics and genomics research and development, and the effect of such laws and policies on progress in these fields and on commercialization, drug development, health care delivery, and public health. The initiative is designed to support rigorous, carefully focused legal, statistical, economic, political science, historical, and other social scientific investigations, both theoretical and empirical.
As used in this RFA, the term “genetics and genomics” includes genomics (broadly defined to include both nucleic acid and protein products of large-scale analyses of the human and other genomes and methods for identifying and analyzing them) and human molecular genetics. The term is not, however, meant to include all of biotechnology, although the line between genomics and biotechnology is frequently hard to define. For example, the term “genetics and genomics subject matter” includes the following: (1) Both individual elements of data and comprehensive databases or other resources regarding genes and gene fragments; gene regulatory sequences; ESTs; SNPs; haplotypes; proteins and protein structures; protein-protein interactions; cellular pathways; computational models of the cell; gene expression profiling (microarrays); small molecules; and mouse (or other animal) knockouts. (2) The relationships between diseases or traits and genes, SNPs, haplotypes, or proteins; the relationships among genotype, environment, and phenotype (e.g., in large databases); and the use of such information in diagnostics. (3) Fundamental tools or methods for the production or analysis of data or databases of the types listed above, the bioinformatics software to probe the databases, and the algorithms that the software elaborates. The term “genetics and genomics subject matter” as used in this initiative does not, however, include such subject matter as biomedical devices, engineered tissues, stem cells, large-scale cell culture, whole organism cloning, or individual treatment applications.
Some examples of appropriate topic areas, with examples of specific research questions for each area, are listed below. Investigators are welcome to propose research in one or more of these topic areas, or in similar areas. Investigators should not be constrained by the specific research questions included on this list. The focus of the research, however, should remain on intellectual property rights to genetics and genomics-related subject matter, and should not be so broad as to encompass other major areas of biotechnology. (1) Types of Intellectual Property Rights and Related Policy Implications. What types of intellectual property rights to genetics and genomics-related subject matter are being, or should be, sought, obtained, or refused? What types of entities are seeking, obtaining, or being refused, intellectual property rights in this field? What are, or should be, the standards for novelty, non-obviousness, and utility in this field? What is, or should be, the breadth of the claims in this field? Do intellectual property rights to genetics and genomics-related subject matter benefit the public when there is no identifiable product? What has been the effect of intellectual property rights in this field on research in the private sector? What are the mechanisms, existing or proposed as well as legal or business custom, for protecting information contained in databases generally, and what are the policy implications of allowing or refusing protection for genomic and genetic databases, whether through intellectual property or sui generis protection? What is, or should be, the role of patents, copyrights, trade secrets, and sui generis intellectual property rights for various data types? How do the laws governing patents, copyrights, trade secrets, and sui generis intellectual property rights act as an incentive, a disincentive, or a neutral factor in determining the planning, content, and progress of genetics and genomics research and development programs? (2) Ownership and Assignment of Intellectual Property Rights and Related Policy Implications. What are, or should be, the mechanisms for exploiting intellectual property rights to genetics and genomics-related subject matter? How frequently are, or should, such rights be assigned (e.g., sold, or licensed exclusively or non-exclusively to third parties)? What are, or should be, the usual mechanisms of such assignments? Who are, or should be, the usual parties to such assignments? To what extent would genetics and genomics subject matter be treated differently if the corresponding intellectual property rights were not assigned? What are, or should be, the practices of biotechnology and pharmaceutical companies regarding the sharing of commercially valuable data? What are, or should be, the practices of universities regarding the sharing of commercially valuable data (government funded and non-government funded)? How have universities interpreted the Bayh-Dole Act, and what has been the impact of Bayh-Dole on genetics and genomics research? Are, or should, assignments in this field under Bayh-Dole typically be pursuant to employment contract or policies, or the result of arms-length negotiations? What is the practical impact of restrictions or limitations on the ownership of intellectual property rights imposed by government funding agencies (such as “Declaration of Exceptional Circumstances”)? How will the mechanisms of assignment of intellectual property rights, and restrictions on such assignment, likely affect genetics and genomics subject matter in the future? (3) Licensing Practices and Related Policy Implications. What are the categories of genetics and genomics subject matter for which intellectual property rights are licensed or may be licensed in the future? What are the relative numbers of intellectual property rights involving genetics and genomics subject matter that are subject to licensing arrangements? What are the terms of such licenses (including exclusivity versus non-exclusivity, royalty rates, fields of use restrictions, etc.), and who are the parties to such agreements? What are the structures for such licensing arrangements (e.g., cross-licensing, block or blanket licenses, compulsory licenses, etc.)? What are the structures and operation of patent pools? How are end user license agreements (EULAs) attached to the sale of research tools being used, and how broad are their “reach-through” provisions? To what extent might the genetics and genomics subject matter be differently treated if the corresponding intellectual property rights were not licensed or were not disclosed and treated as a trade secret? How are the planning, content, and progress of genetics and genomics research and development programs affected by refusals to license or offers to license on unacceptable terms? How does the way in which genetics and genomics subject matter is licensed affect the prospects for commercialization? What is the effect of being required to obtain multiple licenses to conduct some types of research or clinical tests? Does an open source model of licensing genomic software tools increase the usefulness of the tools and improve their acceptance in the research community? Are intellectual property rights involving genetics and genomics subject matter to which licensing arrangements pertain more or less likely to be involved in infringement litigation? What would be the policy implications of limiting exclusive licenses in the field of genetics and genomics to therapeutics and vaccines (i.e., excluding diagnostics)? (4) Enforcement and Related Policy Implications. What are the categories of genetics and genomics subject matter for which the intellectual property rights have been involved in administrative or judicial action? What legal issues have been raised in such lawsuits, and who have been the parties to such lawsuits? What has been the resolution of such cases (e.g., dismissal, settlement, administrative action, trial verdict or judgment, appellate judgment, remedies and relief awarded, etc.)? What are the relative numbers of intellectual property rights involving genetics and genomics subject matter that have been filed in various forums? What are the numbers of intellectual property rights involving genetics and genomics subject matter that have been challenged but that do not actually reach litigation? How frequently are cease and desist letters issued, and how do universities or companies respond to them? How have the planning, content, and progress of genetics and genomics research and development programs been affected by threat, actual or perceived, of infringement litigation? What strategies are employed to allocate the risk of, to prepare for, or to defend against, infringement litigation? What impact has Madey v. Duke, 64 USPQ2d 1737, 307 F.3d 1351 (Fed Cir 2002), cert. Denied, 156 L.3d. 656 (2003), interpreting the experimental use (research) exemption to patent infringement in the context, had in the context of academic research? What would be the policy implications of formalizing a research exemption in the patent law? (5) International Issues and Related Policy Implications. What are the categories of genetics and genomics subject matter for which intellectual property rights have been or may be sought both in the United States and abroad? How does the operation of intellectual property rights involving genetics and genomics subject matter differ in the United States from other countries (e.g., what are the differences in the criteria for patentability applied in the U.S. and by other major patent offices, such as in Europe and Japan)? What mechanisms of procurement, ownership, licensing, and enforcement (or restrictions on these activities) exist only in other countries, and what are the advantages and disadvantages of such? How are international treaty obligations likely to affect the laws and customs in the United States governing intellectual property rights to genetics and genomics related subject matter? How do territorial and jurisdictional limitations on intellectual property rights affect the planning, content, and progress of genetics and genomics research and development programs? (6) Overarching Issues. Has the planning, content, and progress of genetics and genomics research and development programs been enhanced, or conversely chilled, by intellectual property rights? Have intellectual property rights positively or negatively affected the quantity and quality of the publication of scientific advances involving genetics and genomics, or the timing of data release and publication? What are the legal and practical implications for unfettered research activities (e.g., the significance of a bona fide research use exemption to patent infringement, a fair use defense to copyright infringement, a reverse engineering exception to trade secret misappropriation, etc.)? Are existing mechanisms of protection of intellectual property rights to genetics and genomics related subject matter adequate or inadequate to the task of striking the proper balance between intellectual property rights and open access to devices, methods, products and data involved in genetics and genomics research and development? How have intellectual property rights to genetics and genomics-related subject matter positively or negatively affected public access to health care (e.g., accelerated or delayed the commercial availability of diagnostics or treatments, increased or decreased their cost, etc.)?
A major goal of this initiative is to help expand the research base necessary to inform the future development of policy options regarding intellectual property in the contexts of genetics and genomics research and development. In this sense, the proposed development of policy options by applicants to this initiative is not required, but is encouraged when feasible. Investigators may propose to examine existing databases related to biotechnology and intellectual property rights or to gather new empirical data. However, proposals that are primarily dependent on data mining efforts should identify and incorporate innovative analytical methodologies to interpret the data.
Although applications for proposals to examine issues regarding intellectual property, genetics, and genomics in the specific context of differing cultures and belief systems are beyond the scope of this initiative, the NHGRI encourages research on these topics as part of its regular research program in the area of Ethical, Legal, and Social Implications (ELSI). Applicants interested in conducting research on such topics are strongly encouraged to consider submitting R01 or R03 applications under one of the appropriate standing NHGRI PAs for the ELSI Program. See http://grants.nih.gov/grants/guide/pa-files/PA-04-050.html (R01 Program Announcement); http://grants.nih.gov/grants/guide/pa-files/PA-04-051.html (R03 Program Announcement).
This RFA will use NIH R01 and R03 award mechanisms. Applicants are solely responsible for planning, directing, and executing the proposed project. This RFA is a one-time solicitation. Future unsolicited, competing-continuation applications based on this project will compete with all investigator-initiated applications and will be reviewed according to the customary peer review procedures. The earliest anticipated award date is 15 July 2005. Applications that are not funded in the competition described in this RFA may be resubmitted as new investigator-initiated applications using the standard receipt dates for new applications described in the instructions to the PHS 398 application.
This RFA uses just-in-time concepts. It also uses the modular as well as the non-modular budgeting formats (see http://grants.nih.gov/grants/funding/modular/modular.htm).
Annual meetings of investigators will be held. This will facilitate the sharing of information, encourage collaboration, reduce possible duplication of effort, and promote more rapid dissemination of research findings. The initial meeting will take place shortly after the awards are made. Funds for travel to these meetings for up to two investigators per year should be included in the requested budget.
Prospective applicants are asked to submit a letter of intent that includes the following information: descriptive title of the proposed research; name, address, and telephone number of the Principal Investigator; names of other key personnel; participating institutions; number and title of this RFA. Although a letter of intent is not required, is not binding, and does not enter into the review of a subsequent application, the information that it contains allows Institute Center (IC) staff to estimate the potential review workload and plan the review.
Applications must be prepared using the PHS 398 research grant application instructions and forms (rev. 5/2001). Applications must have a DUN and Bradstreet (D&B) Data Universal Numbering System (DUNS) number as the Universal Identifier when applying for federal grants or cooperative agreements. The DUNS number can be obtained by calling 866-705-5711 or through the website at http://www.dunandbradstreet.com/. The DUNS number should be entered on line 11 of the face page of the PHS 398 form. The PHS 398 document is available at http://grants.nih.gov/grants/funding/phs398/phs398.html in an interactive format. For further assistance contact GrantsInfo, 301-435-0714, e-mail: [email protected].
The RFA label available in the PHS 398 (rev. 5/2001) application form must be affixed to the bottom of the face page of the application. Type the RFA number on the label. Failure to use this label could result in delayed processing of the application such that it may not reach the review committee in time for review. In addition, the RFA title and number must be typed on line 2 of the face page of the application form and the YES box must be marked. The RFA label is also available at: http://grants.nih.gov/grants/funding/phs398/label-bk.pdf.
The Center for Scientific Review (CSR) will not accept any application in response to this RFA that is essentially the same as one currently pending initial review, unless the applicant withdraws the pending application. However, when a previously unfunded application, originally submitted as an investigator-initiated application, is to be submitted in response to an RFA, it is to be prepared as a new application. That is, the application for the RFA must not include an introduction describing the changes and improvements made, and the text must not be marked to indicate the changes from the previous unfunded version of the application.
Letters of intent are due 21 October 2004, with applications due 18 November 2004. The earliest anticipated start date is 15 July 2005.
Contact: Jean E. McEwen, NHGRI, Division of Extramural Research, Ethical, Legal, and Social Implications Program, 5635 Fishers Lane, Suite 4076, MSC 9305, Bethesda, MD 20892-9305 USA, until 28 June 2004: 301-402-4997, after 28 June 2004: 301-496-7531, fax: 301-402-1950, e-mail: [email protected]; Rudy O. Pozzatti, NHGRI, Scientific Review Branch, 5635 Fishers Lane, Suite 4076, MSC 9306, Bethesda, MD 20892-9306 USA, 301-402-0838, fax: 301-435-1580, e-mail: [email protected].
Reference: RFA No. RFA-HG-04-004
SBIR/STTR: Circulating Cells and DNA in Cancer Detection
Notice: This Request for Application (RFA) must be read in conjunction with the current Omnibus Solicitation of the National Institutes of Health (NIH), Centers for Disease Control and Prevention (CDC), and Food and Drug Administration (FDA) for Small Business Innovation Research (SBIR) Small Business Technology Trandfer (STTR) Grant Applications. The solicitation (see http://grants.nih.gov/grants/funding/sbirsttr1/index.pdf or http://grants.nih.gov/grants/funding/sbirsttr1/index.doc) contains information about the SBIR and STTR programs, regulations governing the programs, and instructional information for submission. All of the instructions within the SBIR/STTR Omnibus Solicitation apply with the exception of the following: special receipt dates, and initial review convened by the National Cancer Institute (NCI) Division of Extramural Activities.
The Division of Cancer Prevention of the NCI invites small business applications for research projects to develop novel technologies for capturing, enriching, and preserving exfoliated abnormal cells and circulating DNA from body fluids or effusions and to develop methods to concentrate these cells and DNA for cancer biomarker detection.
In body fluids, such as sputum, the number of exfoliated tumor cells is often low compared to the number of normal cells, making it difficult to detect these abnormal cells by routine cytopathology. Separation of dysplastic cells from degenerating cells and cells undergoing non-specific reactive changes is problematic. Moreover, exfoliated cells are frequently contaminated with normal cells, bacteria, and cellular debris. Therefore, enrichment methods are needed to allow for routine detection and molecular analysis of small numbers of exfoliated cells.
Circulating extracellular DNA was first reported in 1948. It has been shown that the circulating DNA in the blood of cancer patients has genetic characteristics identical to those of the primary tumors. Thus, circulating DNA is an important material that may be useful for cancer detection. Currently available methods for isolating undegraded circulating DNA are limited, and there is a need to develop novel methods which improve the yield of undegraded DNA and to adapt detection assays so that this DNA can be used to detect mutations, microsatellite instabilities, loss of heterozygosity, epigenetic changes, and other molecular genetic changes.
This RFA will utilize the SBIR and STTR mechanisms, but will be run in parallel with a program announcement of identical scientific scope (PA-04-035) that will utilize the exploratory/developmental (R21) grant mechanism.
Cellular and molecular changes that ensue during tumor progression occur over a number of years and in an apparently stochastic manner. For example, it takes an average of 15 to 20 years for a small adenomatous polyp to become malignant. Prior to the appearance of a morphologically identified precancerous lesion, numerous genetic and molecular alterations have occurred. During the early stages of cancer development, there is a window of opportunity to detect precancerous cells with genetic or molecular biomarkers that identify and characterize their progression towards cancer. Finding molecular and genetic biomarkers of malignancy is an extraordinary opportunity for the NCI and is particularly important in detecting the emergence of precancerous cell populations. In these earliest stages of neoplasia, lesions are more likely to be amenable to eradication. This principle has been well-demonstrated in cervical neoplasia, where screening for dysplastic exfoliated cells can result in a 70 percent or greater reduction in mortality due to cervical cancer. Detection of genetic abnormalities in preneoplastic lesions poses challenges because of the small size of lesions, the heterogeneity of precancerous cells, and the relatively low number of abnormal cells compared to normal cells.
More than 80 percent of human tumors (e.g. colon, lung, prostate, oral cavity, esophagus, stomach, uterine cervix, and bladder) originate from epithelial cells, often at a mucosal surface, and are clonal in origin.Cells from these tumors exfoliate spontaneously into blood, sputum, urine, and various effusions. Abnormalities within these exfoliated cells could be used to detect and identify precancerous lesions or very early stage cancers if highly sensitive technologies were available to identify the presence of a few abnormal cells among millions of normal cells. For example, PCR has been used to detect mutant DNAs in neoplastic exfoliated cells; mutations have been detected in ras genes present in stool samples obtained from patients with colorectal cancer, and in p53 from the urine of patients with bladder cancer and in the sputa of patients with lung cancer. Assays to detect genetic mutations, microsatellite instability, or hypermethylation may be adapted for use with exfoliated cells. As these assays are complex and technically challenging, their general use will require the development of novel technologies for isolating and enriching abnormal exfoliated cells.
Studies performed in the early 1970s showed that increased quantities of DNA are found in the plasma of patients suffering from different malignancies, but it was not until the 1990s that this circulating DNA was shown to exhibit tumor-related alterations. Mutant DNA has been found in the plasma of patients with colorectal, pancreatic, biliary tree, skin, head-and-neck, lung, breast, kidney, ovarian, nasopharyngeal, liver, bladder, gastric, prostate, and cervical cancers as well as in haematologic malignancies. Allelic imbalance (AI), which involves the loss or gain of chromosomal regions, is found in many cancers. AI can be detected in genomic tumor DNA released into the blood after cellular necrosis or apoptosis. These observations indicate that plasma/serum may be a suitable specimen source for noninvasive diagnostic, prognostic, and follow-up tests for cancer.
Precancerous exfoliated cells can be identified by cytologic examination of washings or brushings from bronchi, oral cavity, esophagus, stomach, bile and pancreatic ducts, as well as of sputum and urine specimens. However, the detection of these exfoliated cancer cells by routine cytopathological examination is very difficult because the number of abnormal cells in the specimens is usually very low compared to the number of normal cells. It is also difficult to distinguish low grade dysplasia from non-specific reactive or inflammatory changes due to the low sensitivity and specificity of current diagnostic methodologies. This is particularly true of urine cytology, where most low-grade papillary lesions are missed by cytologic examination. Even with new PCR-based technologies with enhanced sensitivity, current technologies for isolating exfoliated cells are too inefficient to be of practical utility. Therefore, the development of novel, high-throughput, sensitive technologies for sample preparations is a prerequisite for the successful detection of the small number of exfoliated cells or of the small amounts of DNA, RNA and proteins in these cells.
There are a variety of approaches to detect and analyze precancerous and cancerous cells in body fluids [e.g., cytopathological analysis, morphometric analysis, molecular biomarkers for specific receptors or genetic changes, Fluorescence in Situ Hybridization (FISH) analysis, or PCR-based analysis].The selection of approach, in many instances, depends on the type of biological specimens (sputum, bronchial washing, cervical brushing, voided urine, etc.). Given that the concentration of the atypical epithelial cells can be very low compared to that of normal cells, all of these approaches require between 1 to 10,000 and 1 million enrichments of the atypical cells. Currently, there are two broad categories of enrichment methods: mechanical (centrifugation, cytospin, sucrose gradients, etc.) and antibody-based selection with mechanical separation (FACS – flow-assisted cell sorting, MACS - magnetic assisted cell sorting, etc.). While these two types of enrichment processes can be used in series to improve the yield, none of the currently available methods achieve sufficient enrichment of atypical cells to allow them to be routinely used for cancer detection.
The single largest barrier to using circulating DNA for cancer detection is the amount of circulating undegraded DNA that can be isolated is low, making it unsuitable for currently available assay technologies. Several factors affect the yield and purity of circulating DNA. Intracellular nuclease activity in both apoptotic and necrotic cells in a particular organ affect the degree of DNA degradation found in body fluids. Also, the degree to which a particular tissue is represented in the total circulating DNA is dependent on the mechanism and efficiency by which apoptotic cells are eliminated from the tissue.
As with any other diagnostic technique, practical application of circulating DNA technology is dependent on concurrent increase in the sensitivity and reproducibility of molecular based-assays. The potential use of circulating DNA for cancer detection could be greatly enhanced by developing isolation methods that result in less degradation and by adapting assay methods to use the low amounts that can be isolated. Because of the limitations of “conventional” markers, there has been a search for additional sources of specificity so as to expand the target pool of cancer-associated molecules. Circulating cells and DNA offer such opportunity for detection molecular aberrations in plasma/serum, or other body fluids, that accurately reflect the situation in primary tumor. This will, however, require the development of methodological consistencies so as to allow valid comparisons between various assays based on circulating cells or DNA.
The primary purpose of this initiative is to encourage the development of technologies for isolating and characterizing exfoliated cells, circulating cells, and plasma/serum DNA. A secondary purpose is the analytical validation of existing and/or newly developed technologies for their usefulness in cancer detection. Analytical validation refers to the measurement of sensitivity and reproducibility of the proposed assay/technology. The long-term goal of the technology development is to identify a panel of well-characterized biomarkers derived from exfoliated cells and/or circulating DNA that can be sampled in a clinical setting. These methodologies will be tested and validated in future population-based clinical trials, and integrated into a comprehensive information system that will be developed under the Early Detection Research Network (www.cancer.gov/edrn). In pursuit of these goals, the NCI invites applications which address the following areas: 1) Development of high-throughput, high-yield technologies for isolating exfoliated cells, circulating cells and DNA in body fluids; 2) Development of methods for enrichment and preservation of exfoliated cells, circulating cells and DNA isolated from body fluids; 3) Development of sensitive, high-throughput molecular, cytomorphometric, immunologic, and other relevant technologies to isolate and characterize tumor cells in malignant effusions for detection of low tumor burden, to help distinguish reactive cells from tumor cells, and to perform accurate assays on circulating DNA; 4) Validation of the sensitivity and reproducibility of current technologies for isolating and characterizing exfoliated cells, circulating cells and DNA isolated from body fluids.
This RFA uses the SBIR and STTR mechanisms, which are set-aside programs. As an applicant, you will be solely responsible for planning, directing, and executing the proposed project. Future unsolicited, competing-continuation applications based on this project will compete with all SBIR/STTR applications and will be reviewed according to the customary peer review procedures. The anticipated award date is approximately 9–11 months from the respective receipt date. Applications that are not funded in the competition described in this RFA may be resubmitted as new SBIR/STTR applications using the standard receipt dates for new applications described in the current SBIR/STTR Omnibus Solicitation. As there are multiple receipt dates, it is possible that an unfunded application can be resubmitted under this RFA as a revised application.
This RFA uses just-in-time concepts. It also uses the modular budgeting as well as the non-modular budgeting formats. Specifically, if you are submitting an application budget of $100,000 total costs (direct, F&A and fee) or less, use the modular budget format. For applications requesting more than $100,000, use the non-modular budget format. Instructions for both are described in the current SBIR/STTR Omnibus Solicitation. This program does not require cost sharing as defined in the current NIH Grants Policy Statement at http://grants.nih.gov/grants/policy/nihgps_2003/NIHGPS_Part2.htm.
Except as otherwise stated in this RFA, awards will be administered under NIH grants policy as stated in the NIH Grants Policy Statement, December 2003, available at http://grants.nih.gov/grants/policy/nihgps_2003/.
Applications may be submitted for support as Phase I STTR (R41) or Phase I SBIR (R43) grants; Phase II STTR (R42) or Phase II SBIR (R44) grants; or the SBIR/STTR FAST-TRACK option as described in the SBIR/STTR Omnibus Solicitation. Phase II applications in response to this RFA will only be accepted as competing continuations of previously funded NIH Phase I SBIR/STTR awards. A Phase II application must be a logical extension of the Phase I research but not necessarily a Phase I project supported in response to this RFA. Fast Track applications will benefit from expedited evaluation of progress following the Phase I feasibility study for transition to Phase II funding for expanded developmental work.
Prospective applicants are asked to submit a letter of intent that includes the following information: descriptive title of the proposed research; name, address, and telephone number of the Principal Investigator; names of other key personnel; participating institutions, number and title of this RFA. Although a letter of intent is not required, is not binding, and does not enter into the review of a subsequent application, the information that it contains allows IC staff to estimate the potential review workload and plan the review.
The PHS 398 research grant application must be used for all SBIR/STTR Phase I, Phase II, and Fast-Track applications (new and revised). Effective 1 October 2003, applications must have a Dun and Bradstreet (D&B) Data Universal Numbering System (DUNS) number as the Universal Identifier when applying for federal grants or cooperative agreements. The DUNS number can be obtained by calling 866-705-5711 or through the website at http://www.dunandbradstreet.com/. The DUNS number should be entered on line 11 of the face page of the PHS 398 form. The PHS 398 is available at http://grants.nih.gov/grants/funding/phs398/phs398.html. Prepare your application in accordance with the SBIR/STTR Omnibus Solicitation and the PHS 398. Helpful information for advice and preparation of the application can be obtained at http://grants.nih.gov/grants/funding/sbirgrantsmanship.pdf. The NIH will return applications that are not submitted on the 5/2001 version of the PHS 398. For further assistance, contact GrantsInfo 301-435-0714; e-mail: [email protected].
Applications hand delivered by individuals to the NCI will no longer be accepted. This policy does not apply to courier deliveries (i.e., FEDEX, UPS, DHL, etc.) (see http://grants.nih.gov/grants/guide/notice-files/NOT-CA-02-002.html). This policy is similar to and consistent with the policy for applications addressed to Centers for Scientific Review as published in the NIH Guide Notice at http://grants.nih.gov/grants/guide/notice-files/NOT-OD-02-012.html.
The Center for Scientific Research (CSR) will not accept any application in response to this RFA that is essentially the same as one currently pending initial review unless the applicant withdraws the pending application. The CSR will not accept any application that is essentially the same as one already reviewed. However, when a previously unfunded application, originally submitted as an investigator-initiated application, is to be submitted in response to an RFA, it is to be prepared as a new application. That is, the application for the RFA must not include an introduction describing the changes and improvements made, and the text must not be marked to indicate the changes from the previous unfunded version of the application.
Letters of Intent are due 17 January 2005, 16 May 2005, and 14 September 2005. Applications are due 14 February 2005, 13 June 2005, and 12 October 2005. The earliest anticipated start dates are January 2006, April 2006, and July 2006.
Contact: Sudhir Srivastava, Division of Cancer Prevention, NCI, 6130 Executive Blvd, EPN Rm 3144, Bethesda, MD 20892-0001 USA, Rockville, MD 20852 (for express/courier service), 301-496-3983, fax: 301-402-8990, e-mail: [email protected]; (for peer review issues) Referral Officer, NCI, Division of Extramural Activities, 6116 Executive Blvd, Rm 8041, MSC 8329, Bethesda, MD 20892-8329 USA, Rockville, MD 20852 (for express/courier service), 301-496-3428, fax: 301-402-0275, e-mail: [email protected].
Reference: RFA No. RFA-CA-06-001
|
0
|
PMC1277133
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A690-A695
|
utf-8
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0706aAnnouncementsBook ReviewAdvances in Protein Chemistry, Volume 65: Proteome Characterization and Proteomics Wetmore Barbara A. Merrick B. Alex Barbara A. Wetmore is a senior postdoctoral fellow in the Proteomics Group at NIEHS. Her research involves site-specific phosphorylation of p53 in cancer research. B. Alex Merrick heads the Proteomics Group in the National Center for Toxicogenomics with the mission of identifying key proteins and pathways involved in toxicant exposure.8 2004 112 12 A706 A706 2005Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Edited by Richard D. Smith and Timothy D. Veenstra
San Diego, CA:Academic Press, 2003. 413 pp. ISBN: 0-12-034265-0, $139.95 cloth
Proteins, like characters in a novel, can be described by their appearance, behavior, interactions, peculiarities in specific situations, and movement in time and space. Also like characters in a novel, proteins are created and undergo maturation but eventually cease to function and are eliminated. By analogy, proteomics has the formidable task of describing extremely large sets of proteins, or “proteomes,” within organisms. Proteome Characterization and Proteomics delivers an excellent balance of state-of-the-art technologies, chemistries, and instrumentation designed to measure proteomes and is supplemented by applications of proteomics to biologic problems in species ranging from single-cell to complex organisms.
The beginning chapter, “Proteomics in the Postgenomic Age,” traces completion of the human genome project to the development of transcriptomics and proteomics. The authors highlight the comparative higher complexity of functioning proteins, the difficulty of predicting post-translational processing from mRNA sequence alone, and the frequent disparity between mRNA levels and protein expression, suggesting that only direct analysis of the proteome itself can characterize proteins with certainty. “The Tools of Proteomics” lucidly explains the types of mass spectrometers and ionization methods available and describes their use and impact in proteomics. Subsequent chapters describe long-standing protein separation techniques including two-dimensional (2D)-gel electrophoresis, liquid chromatography (LC), and capillary electrophoresis.
A recent instrumental refinement for greater sensitivity is discussed in a chapter on the use of accurate mass tags generated during Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) to determine protein identity. FTICR MS, assisted by the DREAMS algorithm (dynamic range enhancement applied to mass spectrometry) makes possible very high mass measurement accuracy (MMA). The following chapter provides an excellent discussion of quantitative proteomic techniques, covering 2D-PAGE techniques, multiplexing, metabolic and postextraction labeling, and isolation and quantitation of phosphopeptides. A chapter covering post-translational modifications describes detection of phosphorylated and glycosylated proteins as well as immunoaffinity chromatography, phosphopeptide mapping, and isotopic labeling and collision-induced dissociation strategies. An approach to mapping post-translational modifications at the amino acid level using LC-MS-MS is nicely described in a section discussing a scoring algorithm for spectra analysis (SALSA).
Advances in structural and functional proteomics are described in a chapter where electrospray ionization mass spectrometry is showcased in studies of noncovalent protein complexes. In a section devoted to proteomic strategies in drug discovery, the authors stress the need for higher-throughput, parallel-analysis platforms to fulfill the pharmaceutical industry’s pressing needs for mass screening. The final section discusses proteomics and bioinformatics, reviewing the DNA and protein sequence databases, 2D-gel annotated databases, and the impending need for genome and proteome database integration.
Topics that the authors do not explore are the explosive growth in proteomics in disease diagnosis, biomarker discovery, and drug-toxicant profiling using retentate chromatography mass spectrometry (RC-MS), and also the growing number of protein/antibody arrays. In one of the most significant and controversial success stories of proteomics, the use of RC-MS has already given cancer researchers a proteomic serum signature that can detect ovarian cancer at an earlier stage than previously possible. Further, microarrays of antibodies or other affinity ligands hold great promise for large parallel analysis at an economy of sample volume and expense.
Overall, this is a well-written volume on the current state of proteomics technologies. The editors convey an understanding of the latest developments in mass spectrometry, protein fractionation and applications. This volume is essential for use of proteomic tools in global protein characterization and discovery research. Like any good novel, researchers are finding that each proteome has a cast of thousands of proteins requiring careful study, characterization, and a means for sophisticated interpretation.
|
0
|
PMC1277134
|
CC0
|
2021-01-04 23:40:10
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A706a
|
utf-8
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0706aAnnouncementsBook ReviewAdvances in Protein Chemistry, Volume 65: Proteome Characterization and Proteomics Wetmore Barbara A. Merrick B. Alex Barbara A. Wetmore is a senior postdoctoral fellow in the Proteomics Group at NIEHS. Her research involves site-specific phosphorylation of p53 in cancer research. B. Alex Merrick heads the Proteomics Group in the National Center for Toxicogenomics with the mission of identifying key proteins and pathways involved in toxicant exposure.8 2004 112 12 A706 A706 2005Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Edited by Richard D. Smith and Timothy D. Veenstra
San Diego, CA:Academic Press, 2003. 413 pp. ISBN: 0-12-034265-0, $139.95 cloth
Proteins, like characters in a novel, can be described by their appearance, behavior, interactions, peculiarities in specific situations, and movement in time and space. Also like characters in a novel, proteins are created and undergo maturation but eventually cease to function and are eliminated. By analogy, proteomics has the formidable task of describing extremely large sets of proteins, or “proteomes,” within organisms. Proteome Characterization and Proteomics delivers an excellent balance of state-of-the-art technologies, chemistries, and instrumentation designed to measure proteomes and is supplemented by applications of proteomics to biologic problems in species ranging from single-cell to complex organisms.
The beginning chapter, “Proteomics in the Postgenomic Age,” traces completion of the human genome project to the development of transcriptomics and proteomics. The authors highlight the comparative higher complexity of functioning proteins, the difficulty of predicting post-translational processing from mRNA sequence alone, and the frequent disparity between mRNA levels and protein expression, suggesting that only direct analysis of the proteome itself can characterize proteins with certainty. “The Tools of Proteomics” lucidly explains the types of mass spectrometers and ionization methods available and describes their use and impact in proteomics. Subsequent chapters describe long-standing protein separation techniques including two-dimensional (2D)-gel electrophoresis, liquid chromatography (LC), and capillary electrophoresis.
A recent instrumental refinement for greater sensitivity is discussed in a chapter on the use of accurate mass tags generated during Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) to determine protein identity. FTICR MS, assisted by the DREAMS algorithm (dynamic range enhancement applied to mass spectrometry) makes possible very high mass measurement accuracy (MMA). The following chapter provides an excellent discussion of quantitative proteomic techniques, covering 2D-PAGE techniques, multiplexing, metabolic and postextraction labeling, and isolation and quantitation of phosphopeptides. A chapter covering post-translational modifications describes detection of phosphorylated and glycosylated proteins as well as immunoaffinity chromatography, phosphopeptide mapping, and isotopic labeling and collision-induced dissociation strategies. An approach to mapping post-translational modifications at the amino acid level using LC-MS-MS is nicely described in a section discussing a scoring algorithm for spectra analysis (SALSA).
Advances in structural and functional proteomics are described in a chapter where electrospray ionization mass spectrometry is showcased in studies of noncovalent protein complexes. In a section devoted to proteomic strategies in drug discovery, the authors stress the need for higher-throughput, parallel-analysis platforms to fulfill the pharmaceutical industry’s pressing needs for mass screening. The final section discusses proteomics and bioinformatics, reviewing the DNA and protein sequence databases, 2D-gel annotated databases, and the impending need for genome and proteome database integration.
Topics that the authors do not explore are the explosive growth in proteomics in disease diagnosis, biomarker discovery, and drug-toxicant profiling using retentate chromatography mass spectrometry (RC-MS), and also the growing number of protein/antibody arrays. In one of the most significant and controversial success stories of proteomics, the use of RC-MS has already given cancer researchers a proteomic serum signature that can detect ovarian cancer at an earlier stage than previously possible. Further, microarrays of antibodies or other affinity ligands hold great promise for large parallel analysis at an economy of sample volume and expense.
Overall, this is a well-written volume on the current state of proteomics technologies. The editors convey an understanding of the latest developments in mass spectrometry, protein fractionation and applications. This volume is essential for use of proteomic tools in global protein characterization and discovery research. Like any good novel, researchers are finding that each proteome has a cast of thousands of proteins requiring careful study, characterization, and a means for sophisticated interpretation.
|
0
|
PMC1277135
|
CC0
|
2021-01-04 23:40:11
|
no
|
Environ Health Perspect. 2004 Aug; 112(12):A706b
|
latin-1
|
Environ Health Perspect
| 2,004 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0112-a0706aAnnouncementsBook ReviewAdvances in Protein Chemistry, Volume 65: Proteome Characterization and Proteomics Wetmore Barbara A. Merrick B. Alex Barbara A. Wetmore is a senior postdoctoral fellow in the Proteomics Group at NIEHS. Her research involves site-specific phosphorylation of p53 in cancer research. B. Alex Merrick heads the Proteomics Group in the National Center for Toxicogenomics with the mission of identifying key proteins and pathways involved in toxicant exposure.8 2004 112 12 A706 A706 2005Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Edited by Richard D. Smith and Timothy D. Veenstra
San Diego, CA:Academic Press, 2003. 413 pp. ISBN: 0-12-034265-0, $139.95 cloth
Proteins, like characters in a novel, can be described by their appearance, behavior, interactions, peculiarities in specific situations, and movement in time and space. Also like characters in a novel, proteins are created and undergo maturation but eventually cease to function and are eliminated. By analogy, proteomics has the formidable task of describing extremely large sets of proteins, or “proteomes,” within organisms. Proteome Characterization and Proteomics delivers an excellent balance of state-of-the-art technologies, chemistries, and instrumentation designed to measure proteomes and is supplemented by applications of proteomics to biologic problems in species ranging from single-cell to complex organisms.
The beginning chapter, “Proteomics in the Postgenomic Age,” traces completion of the human genome project to the development of transcriptomics and proteomics. The authors highlight the comparative higher complexity of functioning proteins, the difficulty of predicting post-translational processing from mRNA sequence alone, and the frequent disparity between mRNA levels and protein expression, suggesting that only direct analysis of the proteome itself can characterize proteins with certainty. “The Tools of Proteomics” lucidly explains the types of mass spectrometers and ionization methods available and describes their use and impact in proteomics. Subsequent chapters describe long-standing protein separation techniques including two-dimensional (2D)-gel electrophoresis, liquid chromatography (LC), and capillary electrophoresis.
A recent instrumental refinement for greater sensitivity is discussed in a chapter on the use of accurate mass tags generated during Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) to determine protein identity. FTICR MS, assisted by the DREAMS algorithm (dynamic range enhancement applied to mass spectrometry) makes possible very high mass measurement accuracy (MMA). The following chapter provides an excellent discussion of quantitative proteomic techniques, covering 2D-PAGE techniques, multiplexing, metabolic and postextraction labeling, and isolation and quantitation of phosphopeptides. A chapter covering post-translational modifications describes detection of phosphorylated and glycosylated proteins as well as immunoaffinity chromatography, phosphopeptide mapping, and isotopic labeling and collision-induced dissociation strategies. An approach to mapping post-translational modifications at the amino acid level using LC-MS-MS is nicely described in a section discussing a scoring algorithm for spectra analysis (SALSA).
Advances in structural and functional proteomics are described in a chapter where electrospray ionization mass spectrometry is showcased in studies of noncovalent protein complexes. In a section devoted to proteomic strategies in drug discovery, the authors stress the need for higher-throughput, parallel-analysis platforms to fulfill the pharmaceutical industry’s pressing needs for mass screening. The final section discusses proteomics and bioinformatics, reviewing the DNA and protein sequence databases, 2D-gel annotated databases, and the impending need for genome and proteome database integration.
Topics that the authors do not explore are the explosive growth in proteomics in disease diagnosis, biomarker discovery, and drug-toxicant profiling using retentate chromatography mass spectrometry (RC-MS), and also the growing number of protein/antibody arrays. In one of the most significant and controversial success stories of proteomics, the use of RC-MS has already given cancer researchers a proteomic serum signature that can detect ovarian cancer at an earlier stage than previously possible. Further, microarrays of antibodies or other affinity ligands hold great promise for large parallel analysis at an economy of sample volume and expense.
Overall, this is a well-written volume on the current state of proteomics technologies. The editors convey an understanding of the latest developments in mass spectrometry, protein fractionation and applications. This volume is essential for use of proteomic tools in global protein characterization and discovery research. Like any good novel, researchers are finding that each proteome has a cast of thousands of proteins requiring careful study, characterization, and a means for sophisticated interpretation.
|
16255776
|
PMC1277814
|
CC BY
|
2021-01-04 16:38:33
|
no
|
AIDS Res Ther. 2005 Oct 28; 2:10
|
latin-1
|
AIDS Res Ther
| 2,005 |
10.1186/1742-6405-2-10
|
oa_comm
|
==== Front
AIDS Res TherAIDS Research and Therapy1742-6405BioMed Central London 1742-6405-2-91621910110.1186/1742-6405-2-9ResearchHuman immunodeficiency virus type-1 episomal cDNA in semen Xu Chong [email protected] Joseph A [email protected] Kenneth H [email protected] Deborah J [email protected] Division of Reproductive Biology, Department of Obstetrics and Gynecology, Boston University School of Medicine, Boston, MA 02118, USA2 Fenway Community Health Center, Boston, MA 02115, USA3 Department of Medicine, Brown University Medical School, Providence, RI 02912, USA2005 11 10 2005 2 9 9 20 6 2005 11 10 2005 Copyright © 2005 Xu et al; licensee BioMed Central Ltd.2005Xu 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
Episomal 2-long terminal repeat (LTR) HIV-1 cDNA, a by-product of HIV-1 infection, is used in clinical trials as a marker for ongoing viral replication. It would be useful to employ 2-LTR cDNA to monitor cryptic HIV-1 infection in the genital tract of men on antiretroviral therapy (ART) to predict the evolution of sexually transmissible drug-resistant HIV-1, but studies thus far have failed to detect this marker in semen. The objectives of this study were: 1) to use a technique that maximizes DNA recovery from HIV-1 infected white blood cells in semen to determine if episomal 2-LTR cDNA is detectable in semen of ART-naïve men with other evidence of genital tract HIV-1 infection, and 2) to compare levels of HIV-1 2-LTR cDNA, RNA, and proviral DNA in semen from HIV-1+ men on ART.
Results
Using a somatic cell DNA extraction technique, 2-LTR cDNA was detected by PCR/ELISA in 4 out of 8 semen samples from ART-naïve men selected for other signs of seminal HIV-1 infection (positive controls). Southern blot and DNA sequencing confirmed that the amplified sequences were HIV-1 2-LTR cDNA; copy numbers ranged from 55 to 504 copies/sample. Two semen samples from a cohort of 22 HIV-1-infected men on dual nucleoside therapy, one with and one without detectable seminal HIV-1 RNA, were 2-LTR cDNA positive (336 and 8,560 copies/sample). Following addition of indinavir to the therapy regimen, no semen samples from 21 men with controlled peripheral and seminal viral loads were 2-LTR cDNA positive at 1 and 6 month time points, despite the persistence of HIV-1 proviral DNA+ semen cells and seminal cytomegalovirus (CMV) shedding in some cases. However, one individual who failed indinavir therapy and later developed distinct protease inhibitor (PI) drug resistance mutations in semen, maintained elevated levels of HIV-1 RNA and 2-LTR cDNA in semen.
Conclusion
2-LTR HIV-1 cDNA is detectable in semen of HIV-1-infected men. Two men on ART had 2-LTR HIV-1 cDNA in semen, suggesting that this marker may prove to be useful to monitor HIV-1 infection in the genital tract of men on ART to predict the evolution of drug resistance mutations in semen.
==== Body
Background
The global AIDS epidemic is primarily attributed to the sexual transmission of Human Immunodeficiency Virus type-1 (HIV-1), a retrovirus that infects CD4+ T cells and macrophages and causes severe immunosuppression in most untreated infected individuals. In HIV-1-infected men, sexual transmission rates are high during acute infection and late-stage disease when HIV-1 viral loads are elevated in both semen and peripheral blood [1,2]. However, a number of studies have also described elevated HIV-1 levels in semen from men with low or undetectable HIV-1 concentrations in peripheral blood [3-5], and others have documented genotypic differences between seminal and peripheral blood HIV-1 [6-8]. These data indicate that the male genital tract is a compartment, like the central nervous system, in which HIV-1 replication and divergent evolution can occur under the influence of local factors. The male genital tract has several distinctive features that can affect local HIV-1 infection processes: 1) the distal portion of the male genital tract, especially the penile urethra, is capable of mounting a mucosal immune response including production of secretory IgA [9] and epithelial cell-derived mediators of innate immunity which can suppress HIV-1 infection [10]; 2) co-infection with other sexually transmitted pathogens can create inflammatory conditions that lead to increased levels of infectious HIV-1 in semen [11]; 3) the blood-testis barrier and locally produced immunosuppressive factors may protect HIV-1 from local immune responses in certain regions of the male genital tract [12], and 4) some antiretroviral drugs may not adequately penetrate regions of the male genital tract [13]. It is essential to better understand HIV-1 infection and persistence in the male genital tract in order to develop better intervention strategies to prevent the sexual transmission of HIV-1.
A circular episomal HIV-1 DNA fragment, 2-long terminal repeat (LTR) complementary DNA (cDNA), has been used as a marker for HIV-1 infection in vivo. 2-LTR cDNA is a by-product of HIV-1 infection, formed after reverse transcription of HIV-1 RNA and prior to integration of the HIV-1 cDNA genome into host cell DNA [14]. Nonintegrated HIV-1 DNA has gained clinical importance because its accumulation within infected cells is thought to indicate recent HIV-1 replication [14-16], and because its presence has been associated with cytopathic effects [17,18]. Although the stability and half-life of 2-LTR extrachromosomal HIV-1 DNA is a matter of some debate [19,20], a recent clinical study, showing that patients on suboptimal therapy acquire drug resistance mutations in episomal viral cDNA while provirus remains unchanged, reaffirms the clinical importance of this marker [21]. 2-LTR cDNA has been detected in blood and tissues of individuals on highly active antiretroviral therapy (HAART), and may be an indication of cryptic HIV-1 replication [16]. This marker is used to monitor residual viral replication and potential evolution of drug resistance mutations in viral reservoir sites in individuals on HAART [22].
Whereas effective antiretroviral therapy (ART) suppresses HIV-1 viral load in both blood and semen, HIV-1 proviral DNA often persists in semen cells [23,24]. 2-LTR cDNA could provide a useful marker for HIV-1 infection in the male genital tract, including cryptic HIV-1 replication in the genital tract in men on HAART. A recent study measured 2-LTR cDNA in blood and semen of men on HAART, but failed to detect any positive semen samples [25], possibly due to a cell selection step that can reduce the sensitivity of HIV-1 DNA detection in semen. We have applied an extraction approach that does not require cell separation and results in selective recovery of DNA from semen somatic cells including all HIV-1 infected white blood cells (WBC). The first goal of our study was to quantify 2-LTR cDNA in semen from ART-naïve men with other signs of HIV-1 infection in the genital tract (positive controls). Our second goal was to assess levels of 2-LTR cDNA in semen from men on ART.
Materials and methods
Semen processing
The semen samples used in this study were archived from previously published studies [1,23]. Positive control samples used for validation of 2-LTR cDNA detection/quantification were from ART-naïve men with evidence of seminal HIV-1 viremia (Group I). HIV-1 proviral DNA in the cellular fraction had been assessed by quantitative PCR; infectious HIV-1 was detected in seminal plasma and cellular fractions by p24 assay following up to 21 days of coculture with phytohaemagglutinin A (PHA)-stimulated peripheral blood mononuclear cells (PBMC) [1]. These samples were collected before RT-PCR was widely available, so HIV-1 RNA copy numbers were not assessed. Samples used to test for seminal 2-LTR cDNA (cryptic HIV-1 replication) in men on antiretroviral therapy were from 22 HIV-1-positive men before and at 1 and 6 months after addition of indinavir to dual nucleoside analog therapy (Group II). Clinical information for this cohort has been presented in an earlier report (23). Seminal white blood cell counts were obtained at the time of semen collection by microscopic analysis of peroxidase-stained polymorphonuclear neutrophils (PMN) and differential counts of lymphocytes and macrophages by immunohistology [26]. Semen from this group was tested for HIV-1 proviral DNA and RNA by quantitative polymerase chain reaction (PCR)/reverse transcriptase (RT)-PCR and blood viral loads were monitored by the Chiron branched DNA RNA-1 assay [23]. For 2-LTR cDNA assessment, aliquots containing cell pellets from 250 μL of semen that had been stored at -80°C were thawed. Nonspermatozoal semen cells were lysed in buffer containing 1% SDS and 100 μg/mL proteinase K. DNA was extracted with the standard phenol/chloroform/ethanol procedure, and stored in distilled water at 4°C until use.
2-LTR cDNA detection and quantification
The 2-LTR cDNA sequence was amplified by PCR in the presence of 10 nM digoxigenin (DIG)-labeled dUTP (Roche Diagnostics, Indianapolis, IN), to label the PCR-amplified DNA products with DIG for their subsequent ELISA analysis. The PCR cycling conditions were: 95°C for 10 min, 40 cycles at 95°C for 30 sec, 57°C for 30 sec and 72°C for 30 sec. The sense and antisense primers were: 5'-AACTA GGGAA CCCAC TGCTT AAG-3' and 5'-TCCAC AGATC AAGGA TATCT TGTC-3'[27]. Positive controls (PBMC from HIV-1-infected men) and negative controls (PBMC and semen from uninfected men) were included in each experiment and underwent the same processing as the test samples.
For the ELISA and Southern blot assays, a 29-base biotinylated probe was constructed (5'-GGAAA ATCTC TAGCA GTACT GGAAG GGCT-3') containing, successively, the last 15 bases of the HIV-1 genome (bases 9705–9719), an additional 4 bases, GTAC, specifically found at HIV-1 2-LTR circle junction, and the first 10 bases of the HIV-1 genome (bases 1–10). In ELISA, DIG-labeled PCR products were bound to the streptavidin-coated microtiterplate (Roche Diagnostics, Indianapolis, IN) via the biotinated probe during a 3 hour-incubation at 56°C, and reacted with anti-DIG antibody-HRP conjugate. Optical density (OD) was measured at 405 nm after adding the HRP substrate ABTS. Samples were considered positive if the OD405 was higher than the average of the negative PCR controls plus 3 times standard deviation (ODneg + 3 SD). Copy numbers of 2-LTR cDNA were determined from a standard curve obtained by amplification of serially-diluted 2-LTR cDNA positive control amplicons. As few as 10 copies of 2-LTR c DNA could be consistently detected in repeated runs. For each sample, a housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), was also amplified and quantified by ELISA to verify the presence of amplifiable DNA.
The PCR products were also analyzed by agarose gel electrophoresis, Southern blot and DNA sequencing. Fifteen microliters of PCR product were separated on a 1% agarose gel and transferred to a Hybond N+ membrane (Schleicher & Schuell, Keene, NH); Southern blot was performed using the 29-base biotinylated probe. After a high-stringency wash, streptavidin-HRP conjugate and its substrate in Luminol/enhancer solution (Pierce, Rockford, IL) were added to the membrane. The image was acquired with Fluorchem SP (Alpha Innotech, San Leandro, CA). Direct sequencing of the PCR product was performed by the Dana-Farber/Harvard Cancer Center core facility.
Cytomegalovirus (CMV) detection
Because CMV infection is common in HIV-1 infected men, and seminal shedding of CMV has been associated with elevated levels of HIV-1 in semen [28], we assayed semen cell extracts for CMV DNA by PCR/ELISA. The sense and antisense primers for CMV were: AGGCG TGTAC GGTGG GAGGT CT and CCGCG TTCCA ATGCA CCGTT CC [29]. The biotin-labeled probe was CCATA GAAGA CACCG GGACC GATCC AGCCT. The PCR parameters were 95°C 30 sec and 60°C 1 min for 40 cycles.
Results
2-LTR cDNA was detected in semen from 4/8 men in Group I (positive controls). Copy numbers ranged from 55 to 504 copies/sample (Table 1). Semen from 2/22 men in Group II on dual nucleoside analog ART were positive for 2-LTR cDNA (336 and 8,560 copies/sample); one of the 2-LTR cDNA positive samples had undetectable HIV-1 RNA, the other had elevated HIV RNA (4 × 105 copies, Table 1). Following the addition of the protease inhibitor (PI), indinavir, to the ART regimen, 0/40 semen samples from 21 men with peripheral and seminal HIV-1 RNA suppression were positive for 2-LTR cDNA (all 21 tested at one-month, and 19 retested at 6 months). The remaining subject in Group II, patient A, was intolerant of the indinavir regimen (nausea) and missed several doses. By one month, blood HIV-1 RNA remained undetectable for this individual, but seminal HIV-1 RNA copy numbers increased from 4 × 105 (baseline) to 2 × 106. He was positive for seminal 2-LTR cDNA while on dual nucleoside therapy before indinavir therapy (8,560 copies/sample), and levels increased at the 1-month post indinavir treatment time point to 20,944/sample (Table 1). At 5-months after initiation of indinavir, his treatment regimen was changed (stavudine, sequinavir and continued lamivudine); by the 6-month time point blood viremia remained undetectable, semen HIV-1 RNA decreased to 3 × 105, and 2-LTR cDNA was undetectable. Of interest, we previously reported that distinct PI resistance mutations were detectable in semen and blood from Patient A 18 months after the failed indinavir therapy (23).
Table 1 2-LTR cDNA Positive Semen Samples
Group ID# Antiretroviral Therapy Seminal HIV-1
2-LTR cDNAa Seminal proviral
HIV-1 DNA Seminal HIV-1 RNAa
or Infectious HIV-1 Seminal CMV DNA Seminal HIV
Host Cellsb
I 1 - 504 + - (Inf) + ND
I 2 - 55 + + (Inf) + 1.1 × 106
I 3 - 430 + + (Inf) - ND
I 4 - 70 + ? (Inf) - ND
II A-0c ZVD, 3TC 8,560 + 400,000 + 5.1 × 106
II A-1d ZVD, 3TC, IDV 20,944 + 2,000,000 + 3.6 × 105
II B-0 ZVD, 3TC 336 + 0 + 2.6 × 106
a Expressed as copy number per sample
b Sum of macrophages and CD4+ lymphocytes/sample
c Baseline
d After 1 month on protease inhibitor therapy
2-LTR cDNA from the four positive samples with the strongest signals in PCR/ELISA were also detected as 190 bp bands on agarose gels, confirming the correct size of the amplified products (Figure 1). The binding of the biotinylated probe to the PCR product in Southern blot assay verified the correct orientation of the PCR amplicon. (Figure 2). To confirm that HIV-1 2-LTR cDNA had been amplified, PCR products from four 2-LTR cDNA positive semen samples and positive PBMC controls were purified and directly sequenced. The sequence results showed that the PCR products were the expected fragment containing the exact sequence spanning the HIV-1 genome base numbers 9585 to 9719 and 1 to 51.
Figure 1 Agarose gel electrophoresis of HIV-1 2-LTR cDNA amplified from human semen. The arrow indicates the 2-LTR cDNA bands. Lanes 5 and 11 are positive controls. Lane M is a DNA ladder. Lanes 4, 8, 10, 12, 13 are positive semen samples.
Figure 2 Southern blot (right panel) confirmation of seminal HIV-1 2-LTR cDNA transferred from agarose gel (left panel, lane 3)
All seven of the 2-LTR cDNA positive semen samples showed other indications of HIV-1 infection. Four of the 2-LTR cDNA+ semen samples from Group I were from PI-naïve men with leukocytospermia (LCS, i.e., ≥1,000,000 WBC/mL semen) and high seminal HIV-1 proviral DNA copy numbers. Two out of three of the Group I 2-LTR cDNA+ samples also contained cell-free infectious HIV-1; the fourth 2-LTR cDNA+ was indeterminate in the culture assay due to bacterial contamination. Seminal HIV-1 RNA levels were not available for Group I. All three of the 2-LTR cDNA positive semen samples from Group II men contained proviral DNA; one was leukocytospermic, and the two samples from Patient A at baseline and one-month also contained high concentrations of HIV-1 RNA. Five out of seven of the 2-LTR cDNA+ semen samples (Groups I and II) were also positive for CMV DNA (Table 1).
On the other hand, many semen samples with indicators of HIV-1 infection were negative in the 2-LTR cDNA assay. Of the four negative Group I samples, all were from ART-naïve men with seminal HIV-1 proviral DNA (inclusion criteria for Group I); one was culture-positive for HIV-1, and 2 were CMV DNA+. Of the 20 2-LTR cDNA-negative Group II baseline samples, one was leukocytospermic, 4 had >100 copies of HIV-1 proviral DNA, 5 contained >1,000 copies of cell-free HIV-1 RNA and 6 were CMV DNA-positive. Of the 40 2-LTR cDNA-negative post-indinavir semen samples, four were leukocytospermic, 12 had proviral HIV-1 DNA, and 5 were from men with seminal CMV DNA.
Discussion
This study demonstrates for the first time that HIV-1 2-LTR cDNA, a clinical marker of HIV-1 infection, can be detected in semen from HIV-1-infected men. Because seminal HIV-1 RNA can potentially originate from outside the genital tract, and seminal HIV-1 proviral DNA may represent latent infection, 2-LTR cDNA may prove to be a valuable marker for monitoring HIV-1 infection within genital tract tissues. The highest copy numbers of 2-LTR cDNA in this study were detected in semen of a man failing indinavir therapy who later developed unique PI resistance mutations in semen. Data from this study suggest that seminal 2-LTR cDNA may provide a useful marker to predict the evolution of sexually transmissible HIV-1 drug resistance mutations in men on ART.
Since the number of HIV-1-infected cells in semen is restricted by small sample size, steps must be taken to maximize HIV-1 DNA extraction. In the only other published study to investigate 2-LTR cDNA in semen, Nunnari et al. [25] isolated seminal WBC on Ficoll before DNA extraction and 2-LTR c-DNA assay. This procedure eliminates HIV-1-infected macrophages, which are an important source of infectious virus, and also reduces the T cell yield [30]. Therefore, the sensitivity of 2-LTR cDNA detection in their study may have been compromised. In our study, we used a direct lysis technique to extract DNA from nonspermatozoal cells in semen cell pellets. Sperm DNA is tightly bound by disulphide bonds in histones, and its extraction requires the addition of a reducing agent such as dithiothreitol (DTT) [31]. Our lysis technique leaves the sperm heads intact and provides extracts enriched for WBC DNA. In this study, numbers of semen WBC ranged from 1.8 × 105 to 2.7 × 106/mL for 2-LTR cDNA-positive samples, and 2-LTR cDNA copy numbers were relatively low (range 10–630 per 106 cells). The sensitivity of 2-LTR cDNA detection in semen could be further increased by using a larger fraction of the semen sample, instead of a small aliquot as was used in this study, and by centrifugation procedures that enrich episomal HIV-1 DNA prior to the PCR reaction. Cell fractionation procedures could be applied to samples with high copy numbers of 2-LTR cDNA to identify cellular targets of HIV-1 infection in the male genital tract.
All of the 2-LTR cDNA+ semen samples contained HIV-1 proviral DNA, and all but two were positive for HIV-1 RNA or infectious virions. Five out of 7 of the 2-LTR cDNA+ semen samples were also positive for seminal CMV DNA, but this association was not statistically significant because CMV infection was common overall (4/8 men in Group I and 9/22 men in Group II were positive for seminal CMV).
Two men on dual nucleoside analog treatment had 2-LTR cDNA in semen; one of these individuals failed subsequent indinavir therapy and continued to have high levels of HIV-1 RNA and 2-LTR cDNA in semen, although peripheral viral load was undetectable. These data provide further evidence that HIV-1 replication can occur in the genital tract of men on ART, potentially leading to the evolution of drug resistance mutations in semen that cannot be monitored in peripheral blood. On the other hand, none of 40 post-indinavir semen samples from men controlling HIV-1 RNA levels in blood and semen was positive for 2-LTR cDNA, suggesting that cryptic HIV-1 replication in the genital tract (without the appearance of HIV-1 RNA in blood or semen) may be uncommon in men on effective ART.
The origin of HIV-1 and infected cells in semen is poorly understood. Some infected cells and HIV-1 virions in semen may originate from infection outside the genital tract, but genetic evidence indicates that at least some of the HIV-1 detected in semen is distinct from HIV-1 in peripheral blood [6-8]. Although germ cells and other cell types have been studied as potential HIV-1 reservoirs in the male genital tract, recent attention has focused on genital tract WBC. Semen contains variable numbers of CD4+ T cells and macrophages [32], and recent studies using immunobeads or gradient/sperm-wash methods have demonstrated that these semen cell types carry HIV-1 proviral DNA and are highly infectious [30,33]. HIV-1-infected T cells and macrophages have been detected throughout the male genital tract [34]. Inflammatory conditions, such as those occurring during infection with other sexually transmitted pathogens, attract large numbers of lymphocytes and macrophages into the seminal compartment, and are associated with higher numbers of infected cells and infectious virus in semen [1,11,35]. It is unknown to what extent HIV-1 infected WBC migrate to the genital tract from peripheral blood or other tissues, or are infected in situ by HIV-1 in the genital tract. A recent study entailing phylogenetic analysis of nucleotide sequences of the C2-V5 region of HIV-1 gp120 from HIV-1 RNA isolated from semen cells and seminal plasma from 5 HIV-1-infected men provided evidence that cell-free HIV-1 in semen has distinct sequences and may not be derived from cells represented in semen [36]. Future studies on the presence of HIV 2-LTR cDNA in genital tract and semen cell populations should provide improved insight into sources of HIV-1 infection in the male genital tract.
Conclusion
This study, which shows that HIV-1 2-LTR c-DNA is detectable in semen cells, provides evidence that HIV-1 infection occurs in the genital tract. This molecular marker of HIV infection could provide an important research tool for better understanding sites of HIV-1 infection in the male genital tract, and as a clinical tool for monitoring genital tract HIV-1 infection in men on suppressive ART.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CX performed all molecular assays. JAP processed samples, performed semen and WBC analyses and maintained databases. KHM coordinated the procurement of semen and clinical information. DJA was responsible for the overall experimental design and implementation of the project. All authors contributed to the writing of the manuscript.
Acknowledgements
Research for this study was supported by NIH grants AI035564 and DK072933 awarded to DJA.
==== Refs
Anderson DJ O'Brien TR Politch JA Martinez A Seage GR Padian N Horsburgh CRJ Mayer KH Effects of disease stage and zidovudine therapy on the detection of human immunodeficiency virus type 1 in semen Jama 1992 267 2769 2774 1349654 10.1001/jama.267.20.2769
Pilcher CD Tien HC Eron JJJ Vernazza PL Leu SY Stewart PW Goh LE Cohen MS Brief but efficient: acute HIV infection and the sexual transmission of HIV J Infect Dis 2004 189 1785 1792 15122514 10.1086/386333
Coombs RW Reichelderfer PS Landay AL Recent observations on HIV type-1 infection in the genital tract of men and women Aids 2003 17 455 480 12598766 10.1097/00002030-200303070-00001
Sadiq ST Taylor S Kaye S Bennett J Johnstone R Byrne P Copas AJ Drake SM Pillay D Weller I The effects of antiretroviral therapy on HIV-1 RNA loads in seminal plasma in HIV-positive patients with and without urethritis Aids 2002 16 219 225 11807306 10.1097/00002030-200201250-00011
Vernazza PL Gilliam BL Dyer J Fiscus SA Eron JJ Frank AC Cohen MS Quantification of HIV in semen: correlation with antiviral treatment and immune status Aids 1997 11 987 993 9223732 10.1097/00002030-199708000-00006
Coombs RW Speck CE Hughes JP Lee W Sampoleo R Ross SO Dragavon J Peterson G Hooton TM Collier AC Corey L Koutsky L Krieger JN Association between culturable human immunodeficiency virus type 1 (HIV-1) in semen and HIV-1 RNA levels in semen and blood: evidence for compartmentalization of HIV-1 between semen and blood J Infect Dis 1998 177 320 330 9466517
Pillai SK Good B Pond SK Wong JK Strain MC Richman DD Smith DM Semen-specific genetic characteristics of human immunodeficiency virus type 1 env J Virol 2005 79 1734 1742 15650198 10.1128/JVI.79.3.1734-1742.2005
Ping LH Cohen MS Hoffman I Vernazza P Seillier-Moiseiwitsch F Chakraborty H Kazembe P Zimba D Maida M Fiscus SA Eron JJ Swanstrom R Nelson JA Effects of genital tract inflammation on human immunodeficiency virus type 1 V3 populations in blood and semen J Virol 2000 74 8946 8952 10982338 10.1128/JVI.74.19.8946-8952.2000
Pudney J Anderson DJ Immunobiology of the human penile urethra Am J Pathol 1995 147 155 165 7604877
Porter E Yang H Yavagal S Preza GC Murillo O Lima H Greene S Mahoozi L Klein-Patel M Diamond G Gulati S Ganz T Rice PA Quayle AJ Distinct defensin profiles in Neisseria gonorrhoeae and Chlamydia trachomatis urethritis reveal novel epithelial cell-neutrophil interactions Infect Immun 2005 73 4823 4833 16040996 10.1128/IAI.73.8.4823-4833.2005
Cohen MS Sexually transmitted diseases enhance HIV transmission: no longer a hypothesis Lancet 1998 351 Suppl 3 5 7 9652712 10.1016/S0140-6736(98)90002-2
Anderson DJ Pudney J Mestecky J, Lamm ME, Strober W, Bienenstock J, McGhee JR and Mayer L Human male genital tract immunity and experimental models Mucosal Immunology 2005 2 3rd New York, N.Y., Elsevier Academic Press 1647 1659
Solas C Lafeuillade A Halfon P Chadapaud S Hittinger G Lacarelle B Discrepancies between protease inhibitor concentrations and viral load in reservoirs and sanctuary sites in human immunodeficiency virus-infected patients Antimicrob Agents Chemother 2003 47 238 243 12499197 10.1128/AAC.47.1.238-243.2003
Pauza CD Trivedi P McKechnie TS Richman DD Graziano FM 2-LTR circular viral DNA as a marker for human immunodeficiency virus type 1 infection in vivo Virology 1994 205 470 478 7975248 10.1006/viro.1994.1667
Morlese J Teo IA Choi JW Gazzard B Shaunak S Identification of two mutually exclusive groups after long-term monitoring of HIV DNA 2-LTR circle copy number in patients on HAART Aids 2003 17 679 683 12646790 10.1097/00002030-200303280-00005
Sharkey ME Teo I Greenough T Sharova N Luzuriaga K Sullivan JL Bucy RP Kostrikis LG Haase A Veryard C Davaro RE Cheeseman SH Daly JS Bova C Ellison RT Mady B Lai KK Moyle G Nelson M Gazzard B Shaunak S Stevenson M Persistence of episomal HIV-1 infection intermediates in patients on highly active anti-retroviral therapy Nat Med 2000 6 76 81 10613828 10.1038/71569
Panther LA Coombs RW Zeh JE Collier AC Corey L Unintegrated circular HIV-1 DNA in the peripheral mononuclear cells of HIV-1-infected subjects: association with high levels of plasma HIV-1 RNA, rapid decline in CD4 count, and clinical progression to AIDS J Acquir Immune Defic Syndr Hum Retrovirol 1998 17 303 313 9525430
Zazzi M Romano L Catucci M Venturi G De Milito A Almi P Gonnelli A Rubino M Occhini U Valensin PE Evaluation of the presence of 2-LTR HIV-1 unintegrated DNA as a simple molecular predictor of disease progression J Med Virol 1997 52 20 25 9131453 10.1002/(SICI)1096-9071(199705)52:1<20::AID-JMV4>3.0.CO;2-T
Bushman F Measuring covert HIV replication during HAART: the abundance of 2-LTR circles is not a reliable marker Aids 2003 17 749 750 12646799 10.1097/00002030-200303280-00014
Butler SL Johnson EP Bushman FD Human immunodeficiency virus cDNA metabolism: notable stability of two-long terminal repeat circles J Virol 2002 76 3739 3747 11907213 10.1128/JVI.76.8.3739-3747.2002
Sharkey M Triques K Kuritzkes DR Stevenson M In vivo evidence for instability of episomal human immunodeficiency virus type 1 cDNA J Virol 2005 79 5203 5210 15795303 10.1128/JVI.79.8.5203-5210.2005
Dornadula G Nunnari G Vanella M Roman J Babinchak T DeSimone J Stern J Braffman M Zhang H Pomerantz RJ Human immunodeficiency virus type 1-infected persons with residual disease and virus reservoirs on suppressive highly active antiretroviral therapy can be stratified into relevant virologic and immunologic subgroups J Infect Dis 2001 183 1682 1687 11343220 10.1086/320715
Mayer KH Boswell S Goldstein R Lo W Xu C Tucker L DePasquale MP D'Aquila R Anderson DJ Persistence of human immunodeficiency virus in semen after adding indinavir to combination antiretroviral therapy Clin Infect Dis 1999 28 1252 1259 10451162
Zhang H Dornadula G Beumont M Livornese LJ Van Uitert B Henning K Pomerantz RJ Human immunodeficiency virus type 1 in the semen of men receiving highly active antiretroviral therapy N Engl J Med 1998 339 1803 1809 9854115 10.1056/NEJM199812173392502
Nunnari G Otero M Dornadula G Vanella M Zhang H Frank I Pomerantz RJ Residual HIV-1 disease in seminal cells of HIV-1-infected men on suppressive HAART: latency without on-going cellular infections Aids 2002 16 39 45 11741161 10.1097/00002030-200201040-00006
Politch JA Wolff H Hill JA Anderson DJ Comparison of methods to enumerate white blood cells in semen Fertil Steril 1993 60 372 375 8393411
Butler SL Hansen MS Bushman FD A quantitative assay for HIV DNA integration in vivo Nat Med 2001 7 631 634 11329067 10.1038/87979
Speck CE Coombs RW Koutsky LA Zeh J Ross SO Hooton TM Collier AC Corey L Cent A Dragavon J Lee W Johnson EJ Sampoleo RR Krieger JN Risk factors for HIV-1 shedding in semen Am J Epidemiol 1999 150 622 631 10490002
Mostad SB Kreiss JK Ryncarz A Chohan B Mandaliya K Ndinya-Achola J Bwayo JJ Corey L Cervical shedding of herpes simplex virus and cytomegalovirus throughout the menstrual cycle in women infected with human immunodeficiency virus type 1 Am J Obstet Gynecol 2000 183 948 955 11035345 10.1067/mob.2000.106589
Quayle AJ Xu C Mayer KH Anderson DJ T lymphocytes and macrophages, but not motile spermatozoa, are a significant source of human immunodeficiency virus in semen J Infect Dis 1997 176 960 968 9333154
Li HH Gyllensten UB Cui XF Saiki RK Erlich HA Arnheim N Amplification and analysis of DNA sequences in single human sperm and diploid cells Nature 1988 335 414 417 3419517 10.1038/335414a0
Wolff H Anderson DJ Immunohistologic characterization and quantitation of leukocyte subpopulations in human semen Fertil Steril 1988 49 497 504 3342902
Politch JA Xu C Tucker L Anderson DJ Separation of human immunodeficiency virus type 1 from motile sperm by the double tube gradient method versus other methods Fertil Steril 2004 81 440 447 14967387 10.1016/j.fertnstert.2003.06.028
Pudney J Anderson D Orchitis and human immunodeficiency virus type 1 infected cells in reproductive tissues from men with the acquired immune deficiency syndrome Am J Pathol 1991 139 149 160 1853930
Xu C Politch JA Tucker L Mayer KH Seage GR Anderson DJ Factors associated with increased levels of human immunodeficiency virus type 1 DNA in semen J Infect Dis 1997 176 941 947 9333152
Paranjpe S Craigo J Patterson B Ding M Barroso P Harrison L Montelaro R Gupta P Subcompartmentalization of HIV-1 quasispecies between seminal cells and seminal plasma indicates their origin in distinct genital tissues AIDS Res Hum Retroviruses 2002 18 1271 1280 12487815 10.1089/088922202320886316
|
16219101
|
PMC1277815
|
CC BY
|
2021-01-04 16:38:33
|
no
|
AIDS Res Ther. 2005 Oct 11; 2:9
|
utf-8
|
AIDS Res Ther
| 2,005 |
10.1186/1742-6405-2-9
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-181621612110.1186/1744-9081-1-18ResearchStimulus-dependent effects on tactile spatial acuity Tannan V [email protected] RG [email protected] M [email protected] Department of Biomedical Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA2005 10 10 2005 1 18 18 15 7 2005 10 10 2005 Copyright © 2005 Tannan et al; licensee BioMed Central Ltd.2005Tannan 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
Previous studies have shown that spatio-tactile acuity is influenced by the clarity of the cortical response in primary somatosensory cortex (SI). Stimulus characteristics such as frequency, amplitude, and location of tactile stimuli presented to the skin have been shown to have a significant effect on the response in SI. The present study observes the effect of changing stimulus parameters of 25 Hz sinusoidal vertical skin displacement stimulation ("flutter") on a human subject's ability to discriminate between two adjacent or near-adjacent skin sites. Based on results obtained from recent neurophysiological studies of the SI response to different conditions of vibrotactile stimulation, we predicted that the addition of 200 Hz vibration to the same site that a two-point flutter stimulus was delivered on the skin would improve a subject's spatio-tactile acuity over that measured with flutter alone. Additionally, similar neurophysiological studies predict that the presence of either a 25 Hz flutter or 200 Hz vibration stimulus on the unattended hand (on the opposite side of the body from the site of two-point limen testing – the condition of bilateral stimulation – which has been shown to evoke less SI cortical activity than the contralateral-only stimulus condition) would decrease a subject's ability to discriminate between two points on the skin.
Results
A Bekesy tracking method was employed to track a subject's ability to discriminate between two-point stimuli delivered to the skin. The distance between the two points of stimulation was varied on a trial-by-trial basis, and several different stimulus conditions were examined: (1) The "control" condition, in which 25 Hz flutter stimuli were delivered simultaneously to the two points on the skin of the attended hand, (2) the "complex" condition, in which a combination of 25 Hz flutter and 200 Hz vibration stimuli were delivered to the two points on the attended hand, and (3) a "bilateral" condition, in which 25 Hz flutter was delivered to the two points on the attended hand and a second stimulus (either flutter or vibration) was delivered to the unattended hand. The two-point limen was reduced (i.e., spatial acuity was improved) under the complex stimulus condition when compared to the control stimulus condition. Specifically, whereas adding vibration to the unilateral two-point flutter stimulus improved spatial acuity by 20 to 25%, the two-point limen was not significantly affected by substantial changes in stimulus amplitude (between 100 – 200 μm). In contrast, simultaneous stimulation of the unattended hand (contralateral to the attended site), impaired spatial acuity by 20% with flutter stimulation and by 30% with vibration stimulation.
Conclusion
It was found that the addition of 200 Hz vibration to a two-point 25 Hz flutter stimulus significantly improved a subject's ability to discriminate between two points on the skin. Since previous studies showed that 200 Hz vibration preferentially evokes activity in cortical area SII and reduces or inhibits the spatial extent of activity in SI in the same hemisphere, the findings in this paper raise the possibility that although SI activity plays a major role in two-point discrimination on the skin, influences relayed to SI from SII in the same hemisphere may contribute importantly to SI's ability to differentially respond to stimuli applied to closely spaced skin points on the same side of the body midline.
==== Body
Introduction
Recently, we reported the development of a semi-automated method for measuring a human subject's ability to discriminate between two points on the skin [1]. In that study, a Two-Point Stimulator (TPS) was employed to deliver tactile stimuli simultaneously to two separate skin sites. Since distance between the two points of the TPS can be adjusted on a trial-by-trial basis, it was possible to employ a Bekesy tracking method to determine a subject's two-point limen under several different conditions of two-point stimulation. Two-point stimuli were presented to the skin under static conditions (two probes simply pressed into the skin), in the presence of flutter stimulation (probes oscillated at 25 Hz as they were pressed into the skin), or in the presence of vibration (probes oscillated at 200 Hz). The results duplicated the finding of Vierck and Jones [2] that demonstrated that oscillating the two probes improved a subject's spatial acuity (as measured by the two-point limen). Furthermore, both our study and the Vierck and Jones report showed that spatial acuity is better in the 25 Hz stimulus condition than in the 200 Hz stimulus condition.
Mountcastle and Darian-Smith [3] proposed that a subject's ability to spatially discriminate between two points on the skin would be dependent on the lateral inhibition that enables the formation of the peaks of neuronal activity in SI cortex. Additionally, LaMotte and Mountcastle [4,5] asserted that the capacity of a subject to accurately localize a flutter stimulus on the skin is determined by the locus and clarity of the flutter-evoked neuron population response within the topographically organized SI network. If this is the case, then the ability of a subject to discriminate between two points would improve if the locus of the responses in SI to the stimuli at the two corresponding skin sites were more clearly defined – i.e., if the spatial extent of the response in SI to a point stimulus waslimited or reduced. Observations by Tommerdahl and colleagues demonstrated that the SI response to a complex stimulus (one comprised of both flutter and vibration) is spatially constrained when compared to the response to flutter alone [6-9]. In other words, the SI response evoked by a complex stimulus is smaller in spatial extent than that evoked by 25 Hz flutter alone. Thus, based on the effect that same-site vibration has on the SI response to flutter, we were led to the prediction that vibration, if presented simultaneously at the same sites as two-point flutter stimuli (i.e., as a complex stimulus comprised of 25 Hz and 200 Hz components), would improve a subject's ability to discriminate between two points. Alternatively, recent findings comparing the SI activity evoked by different conditions of contralateral, ipsilateral and bilateral stimulation in the cat show that the magnitude of response in SI evoked by contralateral stimulation is reduced in the presence of an ipsilateral stimulus [10]. Similar results have been found in the non-human primate (unpublished observations). This led us to the prediction that, because of the decrease in prominence of the two peaks of neuronal activity in SI evoked and consequently, the reduction in the spatial clarity between those peaks of cortical activity, a subject's two-point limen would increase (indicating reduced spatial acuity) with the addition of a stimulus to the unattended hand.
Results
Bekesy tracking algorithms were used to find a subject's two-point limen at the dorsal surface of the right hand under four different stimulus conditions. Exemplary results for a single session (four runs) of a subject are shown in Figure 1. The two-point limen of the subject was tracked for two points delivered simultaneously and oscillated at 25 Hz on the attended hand (AH). The data presented indicate that under this condition the subject was able to detect the presence of two points at a separation of approximately 19 mm (average response for the last five trials). In a second run (the "complex" stimulus condition), the two-point limen was tracked under identical conditions as the first run, with the exception that the 25 Hz stimulus waveform was delivered with an additional 200 Hz vibration on the attended hand (see Methods). The addition of the 200 Hz vibration to the 25 Hz flutter resulted in a decrease in the two-point limen to approximately 16.4 mm. In the two other conditions, the two-point limen was tracked to a two-point 25 Hz flutter stimulus on the attended hand, under identical conditions as the first run, but with the addition of a simultaneous 25 Hz flutter or 200 Hz vibration stimulus to the opposite, unattended hand (UH). Interestingly, in both cases, stimulation of the unattended hand impaired the subjects' ability to discriminate between two points on the attended hand, and thus, the two-point limen actually increased to values of approximately 22 mm and 24 mm for 25 Hz and 200 Hz unattended conditions, respectively. To summarize, the detection of two points presented simultaneously with flutter was improved with same-site vibration and degraded with the addition of either a flutter or vibration stimulus on the opposite, unattended hand.
Figure 1 A tracking protocol was used to conduct a two-point limen threshold test. Separation between the two probe tips on the attended hand (AH) versus time was observed under four conditions of stimulation. One condition consisted of 25 Hz flutter applied by the TPS on the AH. In a second condition, the two tips were applied to the AH by a complex stimulus (25 Hz+200 Hz). For the other two conditions, 25 Hz flutter was applied to the AH with either a 25 Hz flutter or 200 Hz vibration stimulus applied simultaneously to the unattended hand (UH). A single trial consisted of stimuli presented to the skin for 1 sec, and then completely removed from the skin for an inter-stimulus interval of 2 sec. Each run consisted of 30 trials, or a duration of 90 sec total.
To determine subject consistency of the above findings, the tracking data collected under each condition for an individual subject were averaged. The data were normalized to the flutter condition since the primary objective of this study was to determine the effect of vibration on the response normally evoked by two-point flutter stimulation. Thus, the two-point limen for the flutter condition was defined as the value "1" and all other distances are plotted as a proportion of the values obtained under the flutter condition [1]. The normalized average two-point limen plot for one subject is displayed in Figure 2. Note that the two-point limen was reduced (i.e., spatial acuity was improved) for the complex condition – the two-point limen tracks at approximately 80% of the values measured under the flutter condition. In contrast, the two-point limen was larger (i.e., spatial acuity is worse) for both bilateral conditions. In the case in which the opposite or unattended hand was presented with a simultaneous 25 Hz flutter stimulus, the two-point limen tracks approximately 20% higher than the control (attended hand only) condition. Similarly, applying a 200 Hz vibration stimulus simultaneously to the unattended hand resulted in two-point limen values that were approximately 30% higher than the control condition.
Figure 2 Average of two-point limen tracking, under all conditions, for one exemplary subject. All distances are normalized to the two-point distance recorded under the attended hand (AH) 25 Hz flutter condition. Standard error bars demonstrate that across-session variability for the two-point limen tracking method is fairly consistent.
To determine the across-subject consistency of the above findings, the data normalization process applied to the single subject case, as shown in Figure 2 and described above, was repeated for data collected under each condition across all subjects. Normalized and averaged data are shown in Figure 3. Similar to the data presented in Figure 2, the two-point limen for the complex condition tracked between 75 and 80% of that measured under the flutter condition, indicating a 20–25% improvement in spatial acuity resulting from the presence of vibration during the flutter stimulus driving the TPS on the attended hand. Alternatively, tracking of the two-point limen showed an increase of approximately 20% and 30% for the conditions in which the unattended hand was stimulated with flutter and vibration, respectively.
Figure 3 Average of two-point limen tracking across all subjects. All distances are normalized to the two-point distance recorded under the attended hand (AH) 25 Hz flutter condition. Standard error bars demonstrate that across-subject variability for the two-point limen tracking method is fairly consistent.
In order to more directly compare the responses measured under each of the stimulus conditions, the tracking values obtained from the last five trials across all subjects was averaged and normalized to the flutter condition (Figure 4). Again, it is apparent that the two-point limen values decreased by approximately 20% under the complex condition, or when vibration was presented with flutter, by dual-site stimuli on the attended hand. Alternatively, the two-point limen increased when a second, simultaneous stimulus was added to the unattended hand – approximately 20% for the 25 Hz flutter condition and 30% for 200 Hz vibration condition. Standard error bars demonstrate that across-subject variability for the two-point limen tracking method is fairly consistent. ANOVA testing was conducted on this data with the null hypothesis that the mean under the control flutter condition is significantly different than the means obtained under the three test conditions. The means for the bilateral conditions of unattended hand 25 Hz (F = 47.7; p < 0.00000001) and unattended hand 200 Hz (F = 76.3; p < 0.00000001), as well as the complex condition (F = 27.6; p < 0.00001) are significantly different from the mean under the flutter condition.
Figure 4 Average of last five trials of two-point limen tracking across all subjects, with standard error bars. All distances are normalized to the two-point distance recorded under the attended hand (AH) flutter condition.
To ensure that the enhanced acuity of a subject under the complex stimulus condition was not due simply to the increased amplitude that resulted from adding vibration to a flutter stimulus (which resulted in a stimulus amplitude of 120 μm), the two-point limen was tracked on the attended hand at 25 Hz flutter of varying amplitudes. Specifically, a separate series of sessions were conducted to track and compare the two-point limen for the amplitudes of 100, 150, and 200 μm in the flutter-only condition. The results were normalized to the distances observed under the 100 μm condition and were plotted in the same manner as the previous results (see Figure 5). In both the 150 and 200 μm conditions, the two-point limen oscillated approximately within 10% of that observed at the 100 μm condition, suggesting that there was no consistent effect on the two-point limen due to the increased amplitude of the complex stimulus and that the effect seen under the complex condition was most likely attributable to the additional high-frequency component.
Figure 5 Average of two-point limen tracking across all subjects for control conditions: 25 Hz flutter stimulus applied by the TPS to the attended hand at amplitudes of 100 μm, 150 μm, or 200 μm. All distances are normalized to the two-point distance recorded under the 25 Hz-100 μm condition.
Discussion
In the present study, we observed stimulus-dependent effects on two-point tracking of a flutter stimulus at the dorsal surface of the attended hand. The two-point limen was reduced (spatial acuity was improved) with a complex stimulus that consisted of 25 Hz flutter and 200 Hz vibration components. Specifically, it was found that adding vibration to the unilateral two-point flutter stimulus improved spatial acuity by 20 to 25%. When the amplitude of the unilateral two-point flutter stimulus was significantly varied (between 100 – 200 μm), the two-point limen was not affected. Simultaneous stimulation of the hand contralateral to the attended site, however, impaired or reduced spatial acuity by 20% with a flutter stimulus and 30% with a vibratory stimulus.
Vega-Bermudez and Johnson [11], using grating orientation studies, cited the importance of skin deformation as a factor affecting spatial acuity. For this reason, we considered the possibility that enhanced spatial acuity with a complex stimulus may be due to the fact that adding vibration to the flutter stimulus introduces another amplitude component, thereby increasing the overall magnitude of the stimulus. Results from our study showed that, within the amplitude range used, there were no significant differences in the two-point limen. This is also consistent with the idea that increasing amplitude of a stimulus does not increase the spatial extent of its cortical response – a finding recently reported [12]. In that report, observations obtained from imaging the optical intrinsic signal in non-human primates showed that higher amplitudes of stimulation with a 25 Hz flutter stimulus in the amplitude range studied (50–400 μm at a frequency of 25 Hz) did not produce larger areas of cortical activation in primary somatosensory cortex (SI). Rather, the spatial extent of the cortical patterns of activation evoked by the flutter stimulus was limited. Simons et al. postulated that the cortical response is sculpted or refined by lateral inhibition, thereby limiting changes in spatial extent [12]. These findings are consistent with the idea that the spatial extent of the SI response evoked by each of the point stimuli plays a role in a subject's ability to discriminate between two stimulus sites on the skin, since changing the amplitude of a flutter stimulus has little effect on either the spatial extent of the SI response in primates or on the two-point limen observations made in this report.
Summers and Chanter [13] reported results on tactile acuity in the fingertip in response to stimuli presented by a broadband tactile array. They found that localization of a 40 Hz target stimulus was improved with the addition of a 320 Hz background stimulus (which surrounded the target) compared to that with a 40 Hz background stimulus. However, Summers and Chanter also stated that this type of interpretation (that the addition of high-frequency vibration to a lower-frequency stimulus results in improvement in perception of that stimulus) was problematic because of the known differences between mechanoreceptors [13]. Previous studies had established the fact that spatial acuity was worse at high frequencies (in the Pacinian range) than at low frequencies (RA/SA range) [1,2,14]. However, if spatial acuity can be attributed to the spatial clarity between regions of cortical activity as LaMotte and Mountcastle [4] proposed, then the cortico-cortical interactions that result from the condition of simultaneous flutter and vibration [8] would undoubtedly have an effect on measures of spatial acuity. Flutter stimuli, such as the ones presented in this study, are known to evoke significant and sustained activity in SI cortex. Skin stimulation at 200 Hz, on the other hand, has been shown to reduce the spatial extent of SI response normally evoked by a 25 Hz flutter stimulus [6-9].
Tommerdahl et al. [6] compared the intrinsic signal evoked in areas 3b/1 by 25 Hz skin stimulation to the intrinsic signal evoked by a same-site skin stimulus containing both 25 and 200 Hz sinusoidal components (a "complex waveform stimulus"). Such experiments revealed that the increase in absorbance evoked in areas 3b/1 by a stimulus having both 25 and 200 Hz components was substantially smaller than the increase in absorbance evoked by "pure" 25 Hz stimulation of the same skin site. It was concluded that within a brief time after stimulus onset, 200 Hz skin stimulation evokes a powerful inhibitory action on area 3b/1 QA neurons. Inhibition due to same-site 200 Hz vibration may play a role in limiting the spatial extent of the cortical activity due to flutter stimulation, creating a sharper and more finely tuned response, suggesting improved spatial acuity.
The finding in previous OIS imaging experiments in cats that high-frequency skin stimulation is accompanied by a contralateral absorbance increase in area SII and, simultaneously, by a decline in absorbance in SI in the same hemisphere led Tommerdahl et al. [7] to consider the possibility that activity in the corticocortical connections that link SII with SI in the same hemisphere [15,16] leads to suppression/inhibition of SI during high-frequency skin stimulation. Insofar as the detailed mechanism by which SII might suppress/inhibit SI, the most straightforward possibility (first suggested by Hirsch and Gilbert) [17] is that long-range corticocortical (i.e., SII→SI) inhibition results from the distinct axonal termination patterns of the local inhibitory neurons in SI. That is, because the two major types of local inhibitory cells in the upper layers of somatosensory cortex (basket and chandelier cells) [18] terminate on cell bodies and initial segments of pyramidal cells, and either do not establish synaptic contacts with other inhibitory cells (this is the case for chandelier cells), or terminate only on the dendrites of inhibitory neurons (characteristic of basket cells), a strong excitatory input from another cortical area (e.g., the input that SI presumably receives from SII at only a brief delay after the onset of high frequency skin stimulation) should evoke an inhibitory process in the SI region that receives the upper layer input, and the inhibition should be selectively expressed on pyramidal cells.
A recent report described that the ability to localize a stimulus on the fingertips of one hand may be impaired with the interference of a similar stimulus on a fingertip of the opposite hand [19], suggesting that spatial acuity may be worse with bilateral stimulation than with unilateral stimulation under certain conditions. In a separate study, we reported that the SI cortical response to contralateral skin stimulation was reduced when an identical stimulus was presented simultaneously to the ipsilateral (mirror image) skin site [10]. Specifically, Tommerdahl and colleagues found that the magnitude of response in SI to bilateral stimulation was 30–35% smaller than the response evoked by a contralateral flutter stimulus. This finding led us to postulate that, since contralateral SI is recognized as the cortical region most responsible for spatial localization [4,5], a reduction in the magnitude of the contralateral SI response – via ipsilateral stimulation – could cause a reduction in spatial acuity. Results from the present study support this hypothesis, suggesting that bilateral stimulation of two homologous body parts leads to a decrease in the percept of spatial acuity.
In a previously published report, Vierck and Jones [2] found that two-point discrimination is improved when the stimuli applied to the skin are oscillated versus static (not oscillated). Consequently, they proposed a model of how spatial acuity improved with oscillating versus static probes. In their report, Vierck and Jones [2] postulated that receptive fields in SI were smaller as a result of the oscillating stimulus condition, and that smaller receptive fields were less likely to overlap with one another, and thus, spatial acuity could improve as a result of changing stimulus conditions. We propose to extend that model by suggesting that the two-point limen is highly correlated with improvements in contrast between peaks of neuronal activity in SI that are evoked by stimulation of two adjacent or near-adjacent points on the skin. Figure 6 summarizes the effect that modification of the stimulus conditions, as reported in this paper, has on our proposed model of SI activity. It should be noted that this conceptual model has been influenced by recent findings about the SI cortical response to skin stimulation [8,10,12].
Figure 6 Model of predicted SI cortical activity in response to specific conditions of tactile stimulation. This model is an extension of the Vierck and Jones model (1970) on two-point receptive fields. a. When stimuli consisting of two points are oscillated on the skin at low-frequency 25 Hz flutter at distant sites, the peaks of SI response are distinct and non-overlapping, and therefore the subject is easily able to discriminate between the two points. b. As the points are positioned at stimulus sites that are closer together, the peaks begin to overlap. Because the peaks are no longer easily distinguishable, discriminability is reduced. c. Adding a same-site high-frequency 200 Hz vibration to the flutter stimuli ("complex" stimuli) has been shown to reduce the spatial extent of the peaks of response in SI and, as found in the present study, would make it easier to distinguish between two points on the skin. d. Presentation of a stimulus at the same skin site on the unattended hand would reduce the magnitude of SI response by flutter stimulation. This reduction in magnitude of SI response would consequently lead to a reduction in the clarity (or contrast) between the activity evoked by the adjacent, or near-adjacent, cortical regions activated by the two stimuli, and as a result, lead to a decrease in spatial acuity.
When stimuli consisting of two points are oscillated on the skin at low-frequency 25 Hz flutter at distant sites, the peaks of SI response are distinct and non-overlapping (Figure 6a). Thus, the subject is easily able to discriminate between the two points. As the points are positioned at stimulus sites that are closer together, the peaks of response begin to overlap (Figure 6b), and because the peaks of activity are no longer easily distinguishable, the two-point limen is increased (i.e., spatial acuity is worse). Adding a same-site high-frequency 200 Hz vibration to the flutter stimuli ("complex" stimuli) has been shown to reduce the spatial extent of the peaks of response in SI [6,8] (Figure 6c) and, as found in the present study, would make it easier to distinguish between two points on the skin. Presentation of a stimulus at the same skin site on the unattended hand would, predictably, reduce the magnitude of SI response by flutter stimulation [10]. This reduction in magnitude of cortical response would consequently lead to a reduction in the clarity (or contrast) between the activity evoked by the adjacent, or near-adjacent, cortical regions activated by the two stimuli (Figure 6d), and as a result, lead to a decrease in spatial acuity.
Conclusion
In this paper, we propose a model that predicts a correlation between SI cortical activity and spatial acuity. Spatial acuity, as measured by the two-point limen, can be modified by changing stimulus conditions that would be predicted to have an impact on the SI cortical response. In particular, while vibration has the effect of reducing the spatial extent of SI cortical response normally evoked by flutter, such as when a vibrotactile stimulus comprised of both flutter and vibration is delivered to the skin, it also has the effect of improving a subject's ability to discriminate between two points on the skin. Presumably, this occurs as a result of vibration decreasing the spatial extent of the SI cortical response. Alternatively, stimulus conditions that are known to reduce the magnitude of the SI cortical response without changing the shape of response, such as when a second and simultaneous stimulus is delivered to a homotopic skin site on the opposite unattended hand, result in a reduction in spatial discrimination. While SI is regarded as playing a major role in two-point discrimination, this study provides evidence that other cortical areas that are connected to SI (such as SII) contribute importantly to SI's ability to differentially respond to closely spaced tactile stimuli.
Materials & methods
Five naïve subjects (21–32 years in age) participated in this psychophysical study. All procedures were reviewed and approved in advance by an institutional review board.
Sinusoidal vertical skin displacement stimuli were delivered using the Cantek Metatron CS-525 vertical displacement stimulator (Cantek Metatron Corp., Canonsburg, PA). The stimulator made contact with the skin via the two tips of the Two-Point Stimulator (TPS) attachment (2.5 cm long, diameter 2 mm) fitted to the terminal end of the moving shaft of the stimulator transducer. The TPS is described in detail in a separate report [1]. An adjustable mechanical arm with lockable joints mounted to a free-standing, rigid platform (fabricated locally) enabled convenient adjustment and maintenance of stimulus position. A second identical Cantek stimulator, implemented in trials that required bilateral stimulation, was fitted with a single 2 mm diameter probe tip and positioned on the hand opposite the TPS in a similar fashion.
The subject was seated in a chair with arms placed comfortably on a table surface. Both arms were placed on X-ray bags filled with glass beads. The investigators molded the bags to fit the contours of the subject's arms, and when the subject was comfortable and the arms positioned appropriately to allow unimpeded access of the stimulator to the center of the dorsal surfaces of each hand, the bags were made rigid by evacuating them of air (achieved by connecting the bag to a vacuum line). In this way the arms were maintained in a comfortable but stable position for the full duration of the experimental session. The subject was unable to see either the experimenter or the stimulator and stimulus-control instrumentation. White noise presented via headphones eliminated potential auditory cues. A micrometer permitted the stimulator transducers and probe assembly to be lowered towards the predefined skin sites. The micrometer position at which the digital display on the stimulator controllers registered a 0.1–0.2 g change in resistive force was interpreted as the point at which the stimulator probes made initial contact with the skin.
A tracking protocol was used to conduct a two-point limen test, which determines the "least two-point separation at which the subject feels (has the subjective impression of) two points," [20] at the dorsal surface of the right hand. The hand dorsum was chosen because the innervation density at this site coincided with optimal resolution and separation capabilities of the TPS, and also because the surface is relatively flat, reducing confounds of skin curvature present at other potential sites of stimulation. Previous studies indicate that response to tactile acuity tests on the hand dorsum is similar to that on the fingertip, suggesting the dorsum to be a suitable site for such tests as well [21]. The subject was instructed to attend to the two-point stimulus presented by the TPS on the tested hand throughout experimentation. For each run, the two probe tips were initially spaced 30 mm apart. The stimuli were presented to the skin simultaneously for 1 sec at an indentation of 500 μm and then completely removed from the skin for 1 sec at an offset of -500 μm. The subject was given these two seconds to report feeling one or two points using a footswitch – no press for one point; a single press for two points. When two points were detected, the two probe tips moved closer together by a step (1 step = 1 mm); when only one point was detected, the two points moved farther apart by a step. The probe tips remained off the skin for the tip movement duration of 1 sec, thus the inter-stimulus interval lasted for a total of 2 sec. This process was repeated until a threshold could be determined, usually around 30 trials, hence a single run took approximately 90 sec. The inter-run interval was 60 sec in duration. The two-point limen was measured under four conditions of frequency and amplitude: unilateral 25 Hz-100 μm, unilateral 25 Hz-100 μm + 200 Hz-20 μm ("complex"), bilateral stimulation of 25 Hz-100 μm on both hands, and bilateral stimulation of 25 Hz-100 μm on the attended hand and 200 Hz-20 μm on the opposite unattended hand. In a session, four runs were conducted, each with one of the aforementioned stimulus conditions. In the bilateral conditions, stimuli were applied by a single timing mechanism and thus were presented to the skin in phase and synchrony. Order of stimulus conditions within a session was randomized and varied for each subject.
Authors' contributions
VT conducted the experiments, analyzed the data and drafted the manuscript. RD had a role in the conduct and design of the experiments. MT was involved with the design of the experiments and the preparation of the manuscript.
Acknowledgements
Supported, in part, by NIH R01 grant NS043375 (M. Tommerdahl, P.I.).
==== Refs
Tannan V Dennis RG Tommerdahl M A novel device for delivering two-site vibrotactile stimuli to the skin J Neurosci Methods 2005 147 75 81 15970331 10.1016/j.jneumeth.2005.05.001
Vierck CJ Jones MB Influences of low and high frequency oscillation upon spatio-tactile resolution Physiol Behav 1970 5 1431 1435 5524531 10.1016/0031-9384(70)90132-0
Mountcastle VB Darian-Smith I Mountcastle VB Neural mechanisms in somesthesia Medical Physiology 1968 2 12 St. Louis: Mosby 1372 1423
LaMotte RH Mountcastle VB Capacities of humans and monkeys to discriminate between vibratory stimuli of different frequency and amplitude: a correlation between neural events and psychophysical measurements J Neurophysiol 1975 38 539 559 1127456
LaMotte RH Mountcastle VB Disorders in somesthesis following lesions of parietal lobe J Neurophysiol 1979 42 400 419 106093
Tommerdahl M Delemos KA Whitsel BL Favorov OV Metz CB Response of anterior parietal cortex to cutaneous flutter and vibration J Neurophysiol 1999 82 16 33 10400931
Tommerdahl M Whitsel BL Favorov OV Metz CB O'Quinn BL Responses of contralateral SI and SII in cat to same site cutaneous flutter versus vibration J Neurophysiol 1999 82 1982 1992 10515988
Tommerdahl M Favorov OV Whitsel BL Effects of high-frequency skin stimulation on SI cortex: mechanisms and functional implications Somatosens Mot Res
Whitsel BL Kelly EF Xu M Tommerdahl M Quibrera M Frequency-dependent response of SI RA-class neurons to vibrotactile stimulation of the receptive field Somatosens Mot Res 2001 18 263 285 11794729 10.1080/01421590120089659
Tommerdahl M Simons SB Chiu JS Favorov OV Whitsel BL Response of SI cortex to ipsilateral, contralateral and bilateral flutter stimulation in the cat BMC Neurosci 2005 6 29 15847693 10.1186/1471-2202-6-29
Vega-Bermudez F Johnson KO Fingertip skin conformance accounts, in part, for differences in tactile spatial acuity in young subjects, but not for the decline in spatial acuity with aging Percept Psychophys 2004 66 60 67 15095940
Simons SB Tannan V Chiu J Favorov OV Whitsel BL Tommerdahl M Amplitude-dependency of response of SI cortex to vibrotactile stimulation BMC Neurosci 2005 6 43 15969752 10.1186/1471-2202-6-43
Summers IR Chanter CM A broadband tactile array on the fingertip J Acoust Soc Am 2002 112 2118 2126 12430823 10.1121/1.1510140
Sherrick CE Cholewiak RW Collins AA The localization of low- and high frequency vibrotactile stimuli J Acoust Soc Am 1990 88 169 179 2380445 10.1121/1.399937
Burton H Fabri M Ipsilateral intracortical connections of physiologically defined cutaneous representations in areas 3b and 1 of macaque monkeys: projections in the vicinity of the central sulcus J Comp Neurol 1995 355 508 538 7636029 10.1002/cne.903550404
Alloway KD Burton H Homotypical ipsilateral cortical projections between somatosensory areas I and II in the cat Neuroscience 1985 14 15 35 3974877 10.1016/0306-4522(85)90161-7
Hirsch JA Gilbert CD Synaptic physiology of horizontal connections in the cat's visual cortex J Neurosci 1991 11 1800 1809 1675266
Jones EG Varieties and distribution of non-pyramidal cells in the somatic sensory cortex of the squirrel monkey J Comp Neurol 1975 160 205 267 803518 10.1002/cne.901600204
Braun C Hess H Burkhardt M Wuhle A Preissl H The right hand knows what the left hand is feeling Exp Brain Res 2005 162 366 373 15827739 10.1007/s00221-004-2187-4
Johnson KO Phillips JR Tactile spatial resolution. I. Two-point discrimination, gap detection, grating resolution, and letter recognition J Neurophysiol 1981 46 1177 1192 7320742
Schlereth T Magerl W Treede R Spatial discrimination thresholds for pain and touch in human hairy skin Pain 2001 92 187 194 11323139 10.1016/S0304-3959(00)00484-X
|
16216121
|
PMC1277816
|
CC BY
|
2021-01-04 16:39:20
|
no
|
Behav Brain Funct. 2005 Oct 10; 1:18
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-18
|
oa_comm
|
==== Front
BMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 1472-6882-5-191622569310.1186/1472-6882-5-19Study ProtocolAcupuncture and rehabilitation of the painful shoulder: study protocol of an ongoing multicentre randomised controlled clinical trial [ISRCTN28687220] Vas Jorge [email protected] Emilio [email protected] Camila [email protected] Antonia Herrera [email protected] Fernando [email protected] Ivan [email protected] Caridad [email protected] Victoria [email protected] Francisco Perez [email protected] Luz [email protected] Jose Maria [email protected] Mauricio [email protected] Francisco [email protected] Isabel [email protected] Ana Maria [email protected] Carmen [email protected] Manuel Anselmo [email protected] Joaquin [email protected] Alonso [email protected] Rosa [email protected] Pablo [email protected] Antonio [email protected] Juan Vicente [email protected] Unidad de Tratamiento del Dolor, Centro de Salud Dos Hermanas "A", Segovia s/n, 41700 Dos Hermanas, Spain2 Unidad de Apoyo a la Investigación (Red IRYSS), Hospital Costa del Sol, Ctra Nacional 340, km 187, 29600 Marbella. Spain3 Servicio de Coordinación de Procesos Asistenciales, Subdirección Asistencial-Servicio Andaluz de Salud, Avenida de la Constitución 18, 41001 Sevilla, Spain4 Servicio de Rehabilitación, Complejo Hospitalario Carlos Haya, Avda Dr Galves Ginachero s/n, 29009 Málaga. Spain5 Servicio de Rehabilitación, Hospital Valme, Carretera de Cádiz s/n, 41014 Sevilla. Spain6 Servicio de Rehabilitación, Hospital Infanta Elena. Ctra. Sevilla-Huelva s/n, 21080 Huelva. Spain7 Servicio de Rehabilitación, Hospital Infanta Margarita. Avda. de Góngora s/n, 14940 Cabra. Spain8 Servicio de Rehabilitación, Hospital General Básico de la Defensa, Carretera de Tentegorra s/n, 30071 Cartagena. Spain9 Servicio de Anestesia y Reanimación, Hospital Serranía, Ctra. El Burgo km 1, 29400 Ronda. Spain10 Servicio de Rehabilitación, Hospital Serranía, Ctra. El Burgo km 1, 29400 Ronda. Spain11 Servicio de Rehabilitación, Hospital Morales Meseguer, Marqués de los Vélez s/n, 30008 Murcia. Spain2005 14 10 2005 5 19 19 19 8 2005 14 10 2005 Copyright © 2005 Vas et al; licensee BioMed Central Ltd.2005Vas 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 the painful shoulder is one of the most common dysfunctions of the locomotor apparatus, and is frequently treated both at primary healthcare centres and by specialists, little evidence has been reported to support or refute the effectiveness of the treatments most commonly applied. According to the bibliography reviewed, physiotherapy, which is the most common action taken to alleviate this problem, has not yet been proven to be effective, because of the small size of sample groups and the lack of methodological rigor in the papers published on the subject. No reviews have been made to assess the effectiveness of acupuncture in treating this complaint, but in recent years controlled randomised studies have been made and these demonstrate an increasing use of acupuncture to treat pathologies of the soft tissues of the shoulder. In this study, we seek to evaluate the effectiveness of physiotherapy applied jointly with acupuncture, compared with physiotherapy applied with a TENS-placebo, in the treatment of painful shoulder caused by subacromial syndrome (rotator cuff tendinitis and subacromial bursitis).
Methods/design
Randomised controlled multicentre study with blind evaluation by an independent observer and blind, independent analysis. A study will be made of 465 patients referred to the rehabilitation services at participating healthcare centres, belonging to the regional public health systems of Andalusia and Murcia, these patients presenting symptoms of painful shoulder and a diagnosis of subacromial syndrome (rotator cuff tendinitis and subacromial bursitis). The patients will be randomised into two groups: 1) experimental (acupuncture + physiotherapy); 2) control (TENS-placebo + physiotherapy); the administration of rescue medication will also be allowed. The treatment period will have a duration of three weeks. The main result variable will be the change produced on Constant's Shoulder Function Assessment (SFA) Scale; as secondary variables, we will record the changes in diurnal pain intensity on a visual analogue scale (VAS), nocturnal pain intensity on the VAS, doses of non-steroid anti-inflammatory drugs (NSAIDs) taken during the study period, credibility scale for the treatment, degree of improvement perceived by the patient and degree of improvement perceived by the evaluator. A follow up examination will be made at 3, 6 and 12 months after the study period has ended. Two types of population will be considered for analysis: per protocol and per intention to treat.
Discussion
The discussion will take into account the limitations of the study, together with considerations such as the choice of a simple, safe method to treat this shoulder complaint, the choice of the control group, and the blinding of the patients, evaluators and those responsible for carrying out the final analysis.
==== Body
Background
Painful shoulder is one of the most common complaints affecting the locomotor apparatus, and is frequently attended both at primary healthcare centres and by specialists. The annual incidence at primary healthcare centres is 1.2% [1,2]. This pathology, which becomes more common with age [3] and with the practice of certain occupations and sports, is evidenced mainly by pain, restricted movement and strength and by the loss of shoulder functionality. The incidence on occupational invalidity, though mentioned by most authors, remains unknown; only approximate data are available, such as those provided by Instituto de Biomecánica at Valencia, which has estimated that 50% of sick leave is accounted for by muscle or bone injuries in the shoulder or neck.
The most common actions taken at present to relieve the symptoms of painful shoulder include steroid injections, physiotherapy, oral NSAIDs and "wait and see"; hardly any of these measures have been tested scientifically to demonstrate their effectiveness in this situation [4]. According to the bibliography reviewed, physiotherapy, which is the most common treatment applied in this case, has not been conclusively shown to be effective; sample sizes have been small and methodological rigor lacking from the papers published in this respect. On the other hand, solid evidence has been produced that ultrasound therapy is ineffective in treating the painful shoulder [1,5,6].
The most relevant bibliography we have examined comprises reviews of the question carried out by the Cochrane group in 2000, 2003 and 2004, together with certain systematic reviews by other authors. In general, the conclusions of the analysts in the Cochrane studies, and of other reviewers working for professional organisations, is that there is very little evidence to support or refute the effectiveness of the treatment commonly applied for the painful shoulder. Therefore, well-designed studies are necessary, using uniform methods of diagnosis and measurement to ensure the validity of the results obtained. We have also analysed other studies, with less backing than the above-mentioned opinions of expert reviewers, but with an acceptable methodology, that have concluded in favour of one or other mode of treatment, such as physiotherapy [7,8] or the injection of corticoids [9,10]. We have also examined the recommendations of panels of experts and members of associations of physiotherapists and GPs.
Acupuncture has been used to treat this type of medical problem in China for over 3000 years. At present, it is under consideration as a technique to be applied in Western medical practice for a great many complaints, especially for cases in which modern techniques are either of limited effectiveness or are unsuitable [11]. Acupuncture is now widely used in the treatment of chronic pain [12-14]. The systematic review carried out by Lewith & Machin on the effectiveness of acupuncture in treating chronic pain concluded that treatment with "real" acupuncture was significantly more effective than that with "false" acupuncture and with a placebo [15]. It has also been shown that acupuncture provokes fewer adverse side effects than does the use of NSAIDs or opiates [16]. In recent years, randomised controlled studies have provided further evidence supporting the use of acupuncture in the treatment of pathologies of the soft tissues of the shoulder. For example, Kleinhenz et al. [17] observed an improvement of 19.2 points on Constant's scale for an experimental group versus one of 8.37 points among a control group, using a placebo featuring retractable needles, but these authors seem to have focused more on demonstrating the effectiveness of the technique used with the control group than on the specificity of the selection of the acupuncture points; moreover, with respect to the analysis of the main result variable (the absolute improvement achieved), the fit to the baseline measurement was not made. On the other hand, Sun et al. [18] (did aim to locate the specificity of the points, by selecting a point that was distal from the affected area, using a randomised controlled test with two groups; 13 patients were treated with acupuncture and specific exercises for the shoulder, and another 22 patients were treated solely with exercise. The design of the latter study is similar to that presented in this project although we believe the sample size is too small and a significant degree of bias was introduced by the fact that the control group (which received only exercises) did not receive the same medical attention as the experimental group. Nevertheless, the results obtained with the experimental group were significantly better than those of the control group. A further problem is that the point proposed by the authors (Zhongpin of the leg) is difficult to locate, as its situation is not constant, on the contrary to that of Tiaokou (ST38), which is situated exactly 8 cun below the articular line of the knee and 1 cun lateral from the tibial crest. Another clinical trial was reported by Gilbertson et al. [19], who described a case in which, after an arthroscopic intervention on the shoulder, traditional and sham acupuncture were compared. It was concluded that real acupuncture provides a significant improvement, concerning the degree of analgesia achieved, the reduction in the quantity of analgesics required, increased mobility and patient satisfaction; however, the acupuncture points selected are not described and so the trial is not reproducible.
Since the introduction of acupuncture techniques into primary healthcare at the Dos Hermanas "A" Health Centre, with the establishment of the Pain Treatment Unit, data have been compiled to obtain an initial evaluation of the reactions of patients who are given acupuncture treatment [20]. Moreover, a pilot study has been carried out to assess the immediate effects of this technique when applied to cases of supraspinal tendinitis [21] as a previous step to the development of the present study.
In this article, we describe a randomised, blinded, multicentre study carried out with a sufficiently-large sample group, with systematised, uniform diagnostic criteria, homogeneous therapeutic interventions including the use of a placebo for the control group, follow up periods exceeding three months and validated measurement of results. We believe such a systematic approach is necessary to clearly describe the current context of treatment for the painful shoulder. In the study, we work on the hypothesis that the acupuncture of Tiaokou ST38, together with physiotherapy, can reduce pain and improve functionality in situations of subacromial syndrome (rotator cuff tendinitis and subacromial bursitis) to a greater extent than does physiotherapy associated with a TENS-placebo treatment. The study began in March 2005 and the recruitment phase remains open.
Methods/design
Design
Randomised controlled multicentre study with blind evaluation by an independent observer and blind, independent analysis.
Study subjects
Patients referred to the Rehabilitation services of the health centres participating in the study. These centres are part of the public health systems of the regions of Andalusia and Murcia (Spain). The patients presented chronic symptoms of subacromial syndrome (rotator cuff tendinitis and subacromial bursitis) and were offered treatment with physiotherapy together with acupuncture or transcutaneous stimulus. They were informed of the characteristics of the study and of the techniques to be used, as well as of the possible risks (infection, lipothymia, hematomas). They were told they could leave the study at any moment, with no type of penalisation or loss of benefits to which they were entitled.
Selection criteria
Inclusion criteria
• Patients with a clinical diagnosis of subacromial syndrome (rotator cuff tendinitis and subacromial bursitis) with a case history > 3 months
• Prior radiography, with normal results
• Informed consent
• Unilateral injury
Exclusion criteria: surgery, luxations or fractures in the proximity of the shoulder; other severe direct or indirect traumas (in traction) observed in the anamnesis and clearly related to the onset of the current episode; hypocoagulates, generalised disorders of the musculoskeletal system or neurologic disorders, vascular trophic disorders in the lower limbs, lymphedema.
Ethical criteria
The ethical validity of this study has been analysed and approved by the corresponding ethical and research committees at the healthcare centres involved. The study design takes into account the Principalism criteria of Beauchamp & Childress (beneficence, non-maleficence, autonomy and justice) and expressly guarantees the patient's right to privacy and informed decision-making. The study also complies with the norms for Good Clinical Practice and the Edinburgh 2000 revision of the Helsinki Declaration. All the patients who participate give their written, informed consent to the clinical research methods applied. During the development of the study, audits will be performed, according to the criteria of the Research and Ethics Committee and the healthcare centre's Quality Committee, independently of the external audits (research funding provider) that may be required.
Criteria and procedures for withdrawal from the study
A patient may be withdrawn from the study at any time, either at will or by decision of the researcher. The reasons for interrupting participation in the study will be recorded on the summary page of the Digital Data Record (DDR). The following procedure is to be followed when a patient withdraws from the study:
• Assess the relevant study variables
• Record any adverse events
• Evaluate the taking of rescue medication
• Indicate the possible co-interventions carried out
• Complete the DDR, record the date and reason for withdrawal.
Randomisation
The patients will be assigned on a random basis to the two study groups: 1) experimental group treated with acupuncture plus physiotherapy; 2) control group, to be given the TENS placebo plus physiotherapy. Randomisation will be carried out in each digital data recording system. Every healthcare centre participating in the study has a specially-designed DDR based on a Dell Axim × 30 PDA; once a new patient's data are entered, the patient is randomly assigned a treatment code (A = experimental; B = control). This code is concealed from the evaluator. Each PDA system has three access codes, one for the study controller, one for the evaluator and one for the doctor carrying out the treatment, and only the latter has access to the treatment code. The research team will take the necessary measures to ensure the confidentiality of the patients taking part, including their deidentification within the databases constructed for the analysis.
Interventions
(See Figure 1)
Acupuncture (experimental group): 3 sessions (once weekly)
Once a week, before the physiotherapy session, the doctors responsible for the treatment (specialists who are well-acquainted with the technique) will apply acupuncture at the Tiaokou ST38 point, following the tiao-shan homolateral technique. This consists of the perpendicular insertion of a single-use sterile filiform acupuncture needle, 7.5 cm long, 30 gauge body diameter, using a guide-tube. The insertion is to be made, after sterilising the skin and with the patient in a prone position, at the Tiaokou point (located equidistally from the flexion fold of the knee and the vertex of the lateral malleolus and 1 inch laterally from the tibial crest, to a depth of 4.5 – 5.0 cm, towards the Chengshan UB57 point (located on the rear surface of the leg, half way between the popliteo fold and the heel, in an inverted-V shaped crease separating the cords of the external calf. Insertion of the needle is followed by vigorous stimulation by means of broad bidirectional rotation movements of the body of the needle, intended to produce the sensation known as Deqi, often described as one of irradiance. The needle is maintained in place for 20 minutes, and manipulated for 1 minute every 5 minutes (i.e. 4 manipulations per session). While the needle is being manipulated, the patient should perform abduction and external and internal rotation exercises.
The single-use sterile needles are manufactured by Cloud & Dragaon Radical Device Co., Ltd (Wujiang, China), according to EU norms, and are imported by Acupuncture Shop, Storegade 58, 6800 Vade (Denmark).
TENS placebo (control group): three sessions (once weekly)
Once a week, before the physiotherapy session, the doctors responsible for the treatment will apply the TENS placebo, which consists of placing two adhesive electrodes, one each on the front and rear surfaces of the leg that is homolateral to the affected shoulder. The electrodes are connected to a deactivated TENS apparatus, model 8016 M. The stimulation unit remains placed in front of the patient, such that the flashing of the diode simulating the stimulus is visible at all times. The patient remains in the same position for 20 minutes, after which the TENS unit is disconnected and the electrodes removed from the patient's body. The frequency of the sessions is the same as that for acupuncture.
Physiotherapy (experimental and control groups): 15 sessions (3 weeks)
The physiotherapy sessions last 40 minutes each and consist of the following (see Additional file 1 1 – Physiotherapy protocol [ISRCTN28687220]):
• Superficial heat therapy (graduated according to the patient's sensations; 5 minutes)
• Recentering of the humeral head (active manoeuvres: 5 minutes; passive manoeuvres: 5 minutes)
• Diadynamic currents, diphase attached with positive pole at the point of greatest pain (graduated according to the patient's sensations; 5 minutes)
• Post-session cryotherapy (10 minutes).
The patient is recommended to avoid any activity that may cause pain in the affected arm for the duration of the study. Ultrasound therapy was not included as an applicable technique because it has been shown to be ineffective in treating pathologies in this region [4].
The daily physiotherapy sessions will follow those of acupuncture and the TENS placebo in the following way: first, a session of acupuncture or TENS placebo followed by the first one of physiotherapy; the next four sessions of physiotherapy to be applied on following consecutive days. The second acupuncture or TENS placebo session is applied prior to the sixth one of physiotherapy, and is followed by another four physiotherapy sessions during the next four consecutive days. The third and final acupuncture or TENS placebo session is applied prior to the eleventh one of physiotherapy, and is followed by another four physiotherapy sessions during the next four consecutive days, after which the final assessment is made. The acupuncture and the TENS placebo sessions take place in identical rooms.
Rescue medication
The patients are allowed to take analgesics and/or NSAIDs if they wish. If they do, this fact, and the daily dose taken, should be recorded in the digital data record (DDR).
• Anti-inflammatory medication
• If taken, the maximum dose allowed is 1 diclophenac pill (50 mg) three times a day for a maximum of 4 weeks.
• Administration instructions: to be taken with meals, to alleviate possible gastric irritation.
• Gastroprotective medication
• Specific indications:
• No past record of ulcers or risk factors such as anticoagulant therapy, association with corticoids, age > 60 years or severe baseline illness (kidney insufficiency, cirrhosis, COPD): NO GASTROPROTECTION
• No past record of ulcers and risk factors such as anticoagulant therapy, association with corticoids, age > 60 years or severe baseline illness (kidney insufficiency, cirrhosis, COPD): GASTROPROTECTION with 200 μg misoprostol every 6 or 12 hours or 20 mg famitidine every 12 hours.
• Past record of ulcers: GASTROPROTECTION with 20 mg omeprazole every 24 hours.
• Contraindications: diclophenac should not be taken by patients allergic to it, to acetylsalicylic acid or other NSAIDs, by patients with a history of asthma, angiodema or rhinitis provoked by NSAIDs, by patients affected by porphyria or with a history of ulcers, coagulation pathologies or haemorrhages. Caution is advised for patients with kidney insufficiency, cardiac insufficiency, hypertension or liver insufficiency.
Study variables
Baseline assessment (T-0)
For the sake of consistency, criteria of selective tension and mobility patterns [22]that we believe are clear enough for homogeneous diagnoses to be achieved (see Additional file 2) will be applied. When the patient has been diagnosed, he/she will be invited to give informed consent to take part in the study. If this is received, the initial assessment will be made by an external rehabilitation evaluator, who will record the following data:
• Sociodemographic data
• Age
• Sex
• Influence on the shoulder of the type of work performed (no effect, moderate effect, severe effect)
• Background
• Previous episodes of shoulder pain (number of episodes)
• If previous episodes occurred, did they provoke sick leave? (No. of such events and their duration)
• Current episode
• Concomitant neck pain(yes/no)
• Duration of the present episode (in months)
• Previous treatment received for the same episode (corticoid injections, analgesic or anti-inflammatory medication)
• Acute onset of the present episode (yes/no)
• Direct cause (excessive tension or otherwise, slight injury or unknown cause)
• Is the affected shoulder the dominant one? (yes/no)
• Constant's Shoulder Function Assessment (SFA) Scale
• Pain intensity in the shoulder during the day, on a 10 cm Visual Analogue Scale (VAS)
• Pain intensity in the shoulder during the night, on a 10 cm Visual Analogue Scale (VAS)
• Analgesic and NSAID medication taken during the previous 2 weeks (4 point Likert scale: 0 no medication; 1 less than the usual daily dose; 2 the normal dose; 3 more than the normal dose).
Main result variable (endpoint)
At 3 weeks (T-1) after the start of the study, the results of the intervention will be assessed by an independent expert evaluator. The 3-month follow up will be performed by the same evaluator. An additional follow up will be carried out at 6 and 12 months after the final evaluation (by independent telephone interviewer), including the main result measurements. These are the change in the SFA score, with respect to baseline values, after the 15th physiotherapy session. This assessment will be made by an external rehabilitation specialist with no information regarding the treatment received by the patient. Constant's Shoulder Function Assessment (SFA) Scale has a maximum score of 100 points, including subjective and objective elements in a proportion of 35/65, respectively. The subjective parameters describe the degree of pain felt by the patient and his/her ability to carry out normal daily activities, as regards both the level of activity and the position of the arm. The objective parameters are based on the range of active compound movements that enable the arm to be moved to functionally relevant positions, using a goniometer to measure the rear and lateral elevation and the positioning of the hand in relation to the head and the trunk in order to assess the degree of rotation achieved. The score for the power exerted by the shoulder is based on the weight (in kg) the patient can raise in abduction, to a maximum of 11 kg. A total SFA score of 100 indicates a shoulder with perfect freedom of movement, no pain and normal functioning.
Secondary variables
After the first week of treatment, the level of confidence in the treatment is measured on a Treatment Credibility Scale (TCS) [23]. This was first proposed by Borkovec and Nau [24] and comprised four items that are assessed on a continuous VAS from 0 to 10 (0 totally disagree; 10 totally agree). The following elements are included, but at the baseline interview, only the first two questions are asked:
1. Are you confident this treatment will alleviate the pain you feel?
2. Does the treatment seem a logical one?
3. Would you recommend this treatment to a friend or relative who had the same problem?
4. Do you think this treatment could be applied to other problems?
After three weeks of treatment (T-1) (15 sessions of physiotherapy), the following secondary variables will be assessed by an independent evaluator:
• Pain intensity in the shoulder during the day, on a 10 cm VAS (DPI-VAS)
• Pain intensity in the shoulder during the night, on a 10 cm VAS (NPI-VAS)
• Degree of improvement perceived by the patient (IPP) (7 point Likert categoric scale: 0 much worse; 1 worse; 2 slightly worse; 3 no change; 4 slightly better; 5 better; 6 much better) [25]
• Degree of improvement perceived by the evaluator (IPE) (7 point Likert categoric scale: 0 much worse; 1 worse; 2 slightly worse; 3 no change; 4 slightly better; 5 better; 6 much better)
• NSAIDs taken (NT)
• Treatment Credibility Scale items 3 and 4
• Adverse events (ADEV) of a traumatic nature that might distort the results of the evaluation.
Follow up at 3 months after completing the treatment (T-2)
An independent evaluator and rehabilitation specialist will make the following assessments at 3 months after the treatment ends:
• Score on Constant's Shoulder Function Assessment (SFA) Scale
• Pain intensity in the shoulder during the day, on a VAS (DPI-VAS)
• Pain intensity in the shoulder during the night, on a VAS (NPI-VAS)
• Degree of improvement perceived by the patient (IPP)
• Degree of improvement perceived by the evaluator (IPE)
• NSAIDs taken (NT)
• Adverse events (ADEV)
• Subsequent episodes of shoulder pain, their duration and treatment applied.
Follow up at 6 months after completing the treatment, carried out by telephone interview (T-3)
An independent evaluator will make the following assessments by telephone interview at 6 months after the treatment ends:
• Score on Constant's Shoulder Function Assessment (SFA) Scale (only the subjective parameters)
• Pain intensity in the shoulder during the day, on a VAS (DPI-VAS)
• Pain intensity in the shoulder during the night, on a VAS (NPI-VAS)
• Degree of improvement perceived by the patient (IPP)
• NSAIDs taken (NT)
• Adverse events (ADEV)
• Subsequent episodes of shoulder pain, their duration and treatment applied.
Follow up at 12 months after completing the treatment, carried out by telephone interview (T-4)
An independent evaluator will make the following assessments by telephone interview at 12 months after the treatment ends:
• Score on Constant's Shoulder Function Assessment (SFA) Scale (only the subjective parameters)
• Pain intensity in the shoulder during the day, on a VAS (DPI-VAS)
• Pain intensity in the shoulder during the night, on a VAS (NPI-VAS)
• Degree of improvement perceived by the patient (IPP)
• NSAIDs taken (NT)
• Adverse events (ADEV)
• Subsequent episodes of shoulder pain, their duration and treatment applied.
Data collection and analysis
Data collection and recording will be carried out using an Dell Axim × 30 PDA, fitted with the Windows® operating system, and specially designed for the purposes of this study, with obligatory fields, validation rules and quality control for error avoidance; a 512 bit encoding system is incorporated to ensure the confidentiality of all the records. Furthermore, as additional functions, it will include a security analysis system with stored passwords, and functions for the export/import and synchronisation of the database content. Access to this PDA will be limited to THREE users: 1) the healthcare centre evaluator; 2) the doctor applying the treatments; 3) the study controller. Each will have a different access code.
The data tables describing the characteristics of the study subjects and the results of each of the evaluations will be entered separately. Each record corresponding to an evaluation, after approval by the researcher responsible, will be stored such that further editing cannot be performed. A backup copy of the database will be encrypted in a 512 bit system and uploaded weekly by safe transmission system to a web server. The study controller will store on the central computer the data obtained from each of the participating healthcare centres and will include all the records in the central database. Similarly, the study controller will make a weekly printout of the updated tables from each of the healthcare centres; these documents will be stored securely and kept as an original record for purposes of inspection or auditing or in case of loss of digital data. Each PDA will be the responsibility of the corresponding researcher at each health centre.
Sample size and associated power
The sample size was predetermined for a level of significance of 0.05, a power of 0.80 and a final average score on Constant's SFA scale of 70 among the experimental group (standard deviation 17) and 65 among the control group (standard deviation 18), with a two-tailed test, according to data taken from Kleinhenz [17]and Sun [18]. Thus, 188 patients are required for the experimental group and 199 for the control group. Assuming a 20% dropout rate, the final sample size was set at 226 patients for the experimental group and 239 for the control group [26].
Population
For the purposes of the analysis, two types of population are to be considered:
• Per intention to treat (ITT): this population will consist of all the randomised patients. Those taking part in the selection phase but not subsequently randomised for treatment will be excluded from this population. This will comprise the main population for analysis of the parameters of effectiveness.
• Per protocol: this population will consist of all the patients in the ITT population with no serious deviations from the protocol. This will comprise the secondary population for analysis.
Treatment comparisons
All the tests will be bilateral and carried out at a level of significance of α = 0.05. So as not to interrupt the study, no intermediate analyses are foreseen.
Data analysis
To describe the different baseline characteristics of the patients, we shall use measures both of position (the mean or the median, according to the asymmetry) and of variability (standard deviation or interquartile range) for the continuous variables, while the others will be described via their frequency distributions, comparing the experimental and control groups. Moreover, a graphic analysis will be made, using smoothed versions of the histograms, boxes and other representations.
For the principal result variable, the difference between the final value and the baseline Constant SFA score, measured in absolute terms, the experimental and control groups will be compared after adjusting by the baseline value and subsequently by other possible confounders. This analysis will be repeated in the follow ups at 3, 6 and 12 months. We will evaluate the effect of possible points of influence by repeating the model after exclusion of the observations with a large Cook-distance value. The assumptions of the model will be tested and reanalysed using a Q-Q plot diagram for the assumption of normality, and by comparing the studentised residuals versus the values of the independent variable for the assumption of constant variances.
For the secondary result variables, the bivariate analysis will be performed using contingency tables analysed by ji-squared tests (for tables larger than 2 × 2), by ji-squared tests corrected for continuity, or Fisher's exact test (for 2 × 2 tables), for the categoric variables, and by simple linear regression for all other cases. Two-dimensional graphs will also be used.
The multivariate analysis will depend on the nature of the response variable. When this is dichotomic, a binary logistic regression analysis will be made, and for this purpose we will establish the dummy variables needed for the categoric predictors. The inclusion of the "experimental group" or "control group" variable will be forced. The "forward procedure" method will be used to select the possible confounding variables in the model, and the entry criterion used will be the change in the likelihood ratio. The functional form of the continuous predictors will be studied via the corresponding statistical criteria, these being mainly graphic procedures.
For continuous variables, a multiple linear regression model will be used, and the same model-selection criteria and diagnostic procedures as described above will be applied.
Current status of the trial
Recruiting of the patients began in March 2005 and will continue until December 2005. Follow up is planned to end in March 2007.
Discussion
We believe that one of the most interesting features of this study is the fact that it introduces a neurostimulative technique, and one that is relatively simple to apply, into rehabilitative medicine. Nevertheless, we are also aware that this comprises one of its greatest difficulties, as this alternative treatment has been very little used, to date, in this field. The possibility of extending this clinical practice into rehabilitation services is one of the challenges we seek to overcome.
It is very difficult to assess all the aspects that go to make up chronic pain, as it is such a subjective experience. This poses problems as regards finding theoretical models and appropriate measuring instruments, although the correct application of experts' recommendations will make it possible to standardise results for later analysis.
There is little evidence to support or refute the effectiveness of the most common interventions to treat the painful shoulder. As well as the necessity to carry out further, well-designed clinical studies, we need to establish a uniform means of defining shoulder pathologies and of developing valid result measurements that are consistent and adaptable to changes in the population. The choice of a single type of pathology prevents us from extrapolating the possible results of research to all the situations in which painful shoulder is diagnosed. However, the fact of selecting a single diagnosis will ensure the specificity of the intervention and will enable us to evaluate particular aspects of the process.
The clinical studies carried out to examine non conventional medical treatments such as acupuncture pose serious problems regarding the study design and possible bias. In selecting the variables to be examined in this study, we have taken into consideration the previous experience of the Pain Treatment Unit at the Dos Hermanas "A" Health Centre and on the comments made in the systematic reviews of the field made by the Cochrane Group.
The placebo selected differs considerably from the acupuncture technique under study, but the intervention in itself is, in fact, difficult to disguise [27]. The depth of needle insertion and the manipulations subsequently effected prevent the use of placebos that are totally credible [28]. This is why we decided to apply an inactive control treatment (the TENS placebo), in order to subject the two groups of patients to the same rhythm of intervention. Nevertheless, the degree of credibility of the intervention and of the placebo will be assessed at the start and at the end of the study. This type of control was first used experimentally by Macdonald [29]. The placebo effect is related to the patient's expectation, to the observer and to the doctor's attention, in combination with classical Pavlovian-type response conditioning activated by the positive or negative expectation of a cure. Petrie and Hazleman [30]used a scale to measure the credibility of acupuncture treatment and that of the TENS placebo, and reached the conclusion that both methods were equally credible, thus justifying the use of the TENS placebo as a control.
Although the characteristics of the intervention prevent us from establishing a double blind design, we intend to blind the patients from the evaluator, and believe this simple blinding will make it possible to achieve sufficient control of possible bias.
As with any other multicentre study, we are confronted with certain disadvantages, such as the loss of unity of judgement, with the participation of various researchers, or the reduced homogeneity of the sample group and the parallel increase in data dispersion. In order to overcome these problems, we have had to create very simple criteria for patient selection, to avoid possible subjectivity as concerns the different researchers. Additionally, we have laid great stress on the need to obtain a good level of communication between the different work groups, a condition that is sometimes essential to standardise criteria and to guarantee the internal validity of the study. The best way to achieve this communication is by means of frequent meetings, the coordination of which will be the responsibility of the study controller.
The limited number of patients and the restrictions imposed on the study concerning the criteria for inclusion are counteracted by its extension to healthcare centres serving populations living in a different cultural situation, and to centres that are organised in different ways. Thus, the external validity of the study is safeguarded.
Abbreviations
COPD: Chronic Obstructive Pulmonary Disease
DDR: Digital data record
GP: General Practitioner
ITT: Intention to treat
NSAIDs: Non-steroid anti-inflammatory drugs
PDA: Personal Digital Assistant
SFA: Shoulder Function Assessment Scale
TCS: Treatment Credibility Scale
TENS: Transcutaneus Electric Nerve Stimulation
VAS: Visual analogue scale
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Conception and design: J. Vas. Revision of the different versions of the study protocol: J. Vas, E. Perea-Milla, C. Mendez, A. Herrera Galante, F. Madrazo, I. Medina, C. Ortega, V. Olmo, F. Perez Fernandez, L. Hernandez, J. M. Seminario, M. Brioso, F. Luna. Substantial contributions to the conception and design of the digital data record: I. Gordo, A. M. Godoy, C. Jiménez, M. A. Ruiz, J. Montes, A. Hidalgo, R. Gonzalez-Quevedo, P. Bosch, A. Vazquez, and J. V. Lozano. All authors have read and approved the final manuscript.
The study protocol was developed in 2004 and is co-funded by Fundación Progreso y Salud (File No. 82045) from the recruitment phase until the short term evaluation, and by Consejeria de Salud de la Junta de Andalucia (File No. 136/04) for the long term follow up phases. The publishing of this article is funded by the IRYSS Network.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
HD Physiotherapy protocol ISRCTN28687220.doc Additional explicative document of the physiotherapy protocol for the patients included in the study.
Click here for file
Additional File 2
HD Table 1 ISRCTN28687220.doc Table 1. Examination protocol
Click here for file
Acknowledgements
We thank all the physiotherapists who have agreed to take part in this study and the patients whose collaboration makes the project possible.
==== Refs
Green S Buchbinder R Hetrick S Physiotherapy interventions for shoulder pain Cochrane Database Syst Rev 2003 CD004258 12804509
van der Windt DA Koes BW de Jong BA Bouter LM Shoulder disorders in general practice: incidence, patient characteristics, and management Ann Rheum Dis 1995 54 959 964 8546527
Cunningham LS Kelsey JL Epidemiology of musculoskeletal impairments and associated disability Am J Public Health 1984 74 574 579 6232862
van der Heijden GJ van der Windt DA de Winter AF Physiotherapy for patients with soft tissue shoulder disorders: a systematic review of randomised clinical trials BMJ 1997 315 25 30 9233322
Green S Buchbinder R Glazier R Forbes A Interventions for shoulder pain Cochrane Database of Systematic Reviews 2000 CD001156
Buchbinder R Green S Youd JM Corticosteroid injections for shoulder pain Cochrane Database Syst Rev 2003 CD004016 12535501
Ginn KA Herbert RD Khouw W Lee R A randomized, controlled clinical trial of a treatment for shoulder pain Phys Ther 1997 77 802 809 9256868
Philadelphia P Philadelphia Panel evidence-based clinical practice guidelines on selected rehabilitation interventions for shoulder pain Physical Therapy 2001 81 1719 1730 11589645
Hay EM Thomas E Paterson SM Dziedzic K Croft PR A pragmatic randomised controlled trial of local corticosteroid injection and physiotherapy for the treatment of new episodes of unilateral shoulder pain in primary care Ann Rheum Dis 2003 62 394 399 12695148 10.1136/ard.62.5.394
van der Windt DA Koes BW Deville W Boeke AJ de Jong BA Bouter LM Effectiveness of corticosteroid injections versus physiotherapy for treatment of painful stiff shoulder in primary care: randomised trial BMJ 1998 317 1292 1296 9804720
Lewith GT Vincent CA Filshie J and White A The clinical evaluation of acupuncture Medical Acupuncture A Western Scientific Approach 1998 Edinburgh, Churchill Livingstone 205 224
Sodipo JO Therapeutic acupuncture for chronic pain Pain 1979 7 359 365 530742 10.1016/0304-3959(79)90092-7
Junnila SY Long-term treatment of chronic pain with acupuncture. Part I Acupunct Electrother Res 1987 12 23 36 2883834
Patel M Gutzwiller F Paccaud F Marazzi A A meta-analysis of acupuncture for chronic pain Int J Epidemiol 1989 18 900 906 2695475
Lewith GT Machin D On the evaluation of the clinical effects of acupuncture Pain 1983 16 111 127 6348651 10.1016/0304-3959(83)90202-6
Lewith GT How effective is acupuncture in the management of pain? J R Coll Gen Pract 1984 34 275 278 6376797
Kleinhenz J Streitberger K Windeler J Gussbacher A Mavridis G Martin E Randomised clinical trial comparing the effects of acupuncture and a newly designed placebo needle in rotator cuff tendinitis Pain 1999 83 235 241 10534595 10.1016/S0304-3959(99)00107-4
Sun KO Chan KC Lo SL Fong DY Acupuncture for frozen shoulder Hong Kong Med J 2001 7 381 391 11773673
Gilbertson B Wenner K Russell LC Acupuncture and arthroscopic acromioplasty J Orthop Res 2003 21 752 758 12798078 10.1016/S0736-0266(03)00028-7
Vas J Mendez C Perea-Milla E Vega E Panadero MD Leon JM Borge MA Gaspar O Sanchez-Rodriguez F Aguilar I Jurado R Acupuncture as a complementary therapy to the pharmacological treatment of osteoarthritis of the knee: randomised controlled trial BMJ 2004 329 1216 1219 15494348 10.1136/bmj.38238.601447.3A
Vas J Perea-Milla E Les effets immédiats de la puncture du tiaokou ES38 dans l'épaule douloureuse Acupuncture & moxibustion 2004 3 167 174
Hernandez Conesa S Exploracion clinica del aparato locomotor en medicina ortopedica 1980 1 Madrid, Editorial Cientifico Medica
Vincent C Lewith G Placebo controls for acupuncture studies J R Soc Med 1995 88 199 202 7745565
Borkovec TD Nau SD Credibility of analogue therapy rationales J Beh Ther Exp Pschiat 1972 3 257 260 10.1016/0005-7916(72)90045-6
Villanueva I Guzman MM Toyos FJ Ariza R Navarro F Sensibilidad y especificidad de los criterios OARSI de mejoria para artrosis: el efecto de la utilización de tres diferentes medidas de dolor Rev Esp Reumatol 2003 30 105 109
Normal Power Calculations 2005
Vickers AJ Placebo controls in randomized trials of acupuncture Eval Health Prof 2002 25 421 435 12449085 10.1177/0163278702238055
Streitberger K Kleinhenz J Introducing a placebo needle into acupuncture research Lancet 1998 352 364 365 9717924 10.1016/S0140-6736(97)10471-8
Macdonald AJ Macrae KD Master BR Rubin AP Superficial acupuncture in the relief of chronic low back pain Ann R Coll Surg Engl 1983 65 44 46 6218776
Petrie J Hazleman B Credibility of placebo transcutaneous nerve stimulation and acupuncture Clin Exp Rheumatol 1985 3 151 153 4017314
|
16225693
|
PMC1277817
|
CC BY
|
2021-01-04 16:31:46
|
no
|
BMC Complement Altern Med. 2005 Oct 14; 5:19
|
utf-8
|
BMC Complement Altern Med
| 2,005 |
10.1186/1472-6882-5-19
|
oa_comm
|
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-221624203710.1186/1471-2091-6-22Research ArticleSubstrate specificity analysis of protein kinase complex Dbf2-Mob1 by peptide library and proteome array screening Mah Angie S [email protected] Andrew EH [email protected] Geeta [email protected] Jason [email protected] Mike [email protected] Michael [email protected] Michael B [email protected] Raymond J [email protected] Department of Biology, California Institute of Technology, Pasadena, CA 91125, USA2 Howard Hughes Medical Institute, California Institute of Technology, Pasadena, CA 91125, USA3 Center for Cancer Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA4 Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA5 Division of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA6 Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT 06520, USA7 Department of Molecular Biophysics & Biochemistry, Yale University, New Haven, CT 06520, USA8 JPT Peptide Technologies GmbH, Invalidenstrasse 130, 10115 Berlin, Germany, USA2005 21 10 2005 6 22 22 11 6 2005 21 10 2005 Copyright © 2005 Mah et al; licensee BioMed Central Ltd.2005Mah 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 mitotic exit network (MEN) is a group of proteins that form a signaling cascade that is essential for cells to exit mitosis in Saccharomyces cerevisiae. The MEN has also been implicated in playing a role in cytokinesis. Two components of this signaling pathway are the protein kinase Dbf2 and its binding partner essential for its kinase activity, Mob1. The components of MEN that act upstream of Dbf2-Mob1 have been characterized, but physiological substrates for Dbf2-Mob1 have yet to be identified.
Results
Using a combination of peptide library selection, phosphorylation of opitmal peptide variants, and screening of a phosphosite array, we found that Dbf2-Mob1 preferentially phosphorylated serine over threonine and required an arginine three residues upstream of the phosphorylated serine in its substrate. This requirement for arginine in peptide substrates could not be substituted with the similarly charged lysine. This specificity determined for peptide substrates was also evident in many of the proteins phosphorylated by Dbf2-Mob1 in a proteome chip analysis.
Conclusion
We have determined by peptide library selection and phosphosite array screening that the protein kinase Dbf2-Mob1 preferentially phosphorylated substrates that contain an RXXS motif. A subsequent proteome microarray screen revealed proteins that can be phosphorylated by Dbf2-Mob1 in vitro. These proteins are enriched for RXXS motifs, and may include substrates that mediate the function of Dbf2-Mob1 in mitotic exit and cytokinesis. The relatively low degree of sequence restriction at the site of phosphorylation suggests that Dbf2 achieves specificity by docking its substrates at a site that is distinct from the phosphorylation site
==== Body
Background
In the budding yeast Saccharomyces cerevisiae the protein phosphatase Cdc14 must be activated to turn off mitotic Cdk activity for cells to exit mitosis. There are two groups of proteins that regulate Cdc14 activity, the Cdc14 early anaphase release (FEAR) network and the mitotic exit network (MEN) (reviewed in [1]).
There is one cyclin-dependent kinase (Cdk), Cdc28, that controls cell cycle progression in S. cerevisiae By associating with mitotic cyclins, Cdc28 promotes entry into mitosis. Cdc14 plays a critical role in inactivating mitotic Cdk activity, thereby promoting exit from mitosis. Cdc14 dephosphorylates Hct1/Cdh1 which can then target Clb2, the main mitosis-specific cyclin, for degradation [2-5]. Cdc14 also promotes the accumulation of Sic1, a Cdk inhibitor, by acting on Sic1 as well as transcriptional activator Swi5, which promotes transcription of SIC1 [3,6]. Together, the degradation of mitotic cyclins and the production of active Sic1 conspire to down-regulate Cdc28 activity and return the cell cycle to an interphase state.
Cdc14 is held in an inactive state in the nucleolus by its inhibitor, Net1 [7,8]. Cdc14 is tethered to Net1 throughout the cell cycle until the onset of anaphase, at which time it is released. The FEAR network and MEN both regulate Cdc14 release and therefore its activity (reviewed in [1,9]). The FEAR network initiates early anaphase release of Cdc14 from the nucleolus by promoting Net1 phosphorylation by mitotic Cdks, weakening the interaction between Cdc14 and Net1 [10]. The FEAR network consists of Esp1 (also known as separase), polo-like kinase Cdc5, the kinetochore protein Slk19, the nuclear protein Spo12, and its homologue Bns1. The action of these proteins is opposed by Pds1 (also known as securin) and the nucleolar replication fork block protein Fob1 [11-13]. However, it is still unclear how the factors that promote and restrain FEAR interact. Although the release of Cdc14 in early anaphase by the FEAR network is transient and insufficient for mitotic exit, exit is delayed when the FEAR network is compromised by mutation. One possible explanation is that the FEAR network weakens the Cdc14-Net1 interaction, enabling the MEN to more rapidly cause a sustained release of Cdc14. The FEAR network also has other mitotic functions that may play an important role in coordinating events during exit from mitosis (reviewed in [1,9]).
In contrast to the FEAR network, the MEN is essential for mitotic exit (reviewed in [1,9,14]). This pathway consists of the GTPase Tem1, the putative guanine-nucleotide exchange factor (GEF) Lte1, the two-component GTPase activating protein (GAP) Bub2-Bfa1, the protein kinases Cdc5, Cdc15, Dbf2, the Dbf2 binding protein Mob1, and the scaffolding protein, Nud1. Genetic and biochemical data have provided significant insight into how this signaling cascade is activated by the localization of its components. Bub2-Bfal binds and inhibits Tem1 at the spindle pole body (SPB) that enters the daughter cell during nuclear division. As the spindle elongates into the bud, Tem1 is presumably activated by Lte1, which is localized in the bud [15,16]. The activated GTP-bound form of Tem1 then somehow activates bound Cdc15, which then phosphorylates and activates the Dbf2-Mob1 kinase complex [17]. However, how activated Dbf2-Mob1 affects Cdc14 release from Net1 and mitotic exit is unknown.
In addition to their role in mitotic exit, Dbf2-Mob1 and the other MEN proteins function in cytokinesis. Dbf2 localizes to the SPB in anaphase as do Tem1, Cdc5, Cdc15, and Mob1 [15,16,18-21]. During late mitosis, Dbf2 and Mob1 migrate to the bud neck. Bud neck localization of Dbf2 and Mob1 are dependent on each other as well as the MEN proteins Cdc5, Cdc14, Cdc15, Nud1, and the septins Cdc12 and Cdc3 [19,22,23]. Several lines of evidence suggest that localization of MEN proteins to the bud neck is crucial for cytokinesis. Mutant mob1ts cells, as well as tem1Δ net1-1 and cdc15Δ net1-1 cells whose mitotic exit defects are bypassed by the net1-1 allele fail to undergo cytokinesis [7,23,24]. Interestingly, localization of Dbf2-Mob1 to the bud neck depends upon Cdc14 [23,25]. MEN-dependent release and activation of Cdc14 may help to ensure that mitotic exit occurs prior to cytokinesis.
The function of Dbf2-Mob1 in cytokinesis is unclear. Also unknown is how Dbf2-Mob1 ultimately leads to release of Cdc14 from the nucleolus during mitotic exit. To give us insight into these two key cell cycle processes, we sought to identify potential substrates and phosphorylation sites of Dbf2-Mob1. Here, we report the substrate specificity of Dbf2-Mob1 and a number of putative substrates that contain a Dbf2 phosphorylation motif and are phosphorylated by Dbf2-Mob1 in vitro.
Results
Determination of optimal peptide sequence motif phosphorylated by Dbf2-Mob1
To identify potential physiological substrates of Dbf2-Mob1, we first proceeded to determine an optimal substrate sequence by using an oriented degenerate peptide library technique [26]. Dbf2, a Ser/Thr kinase, was initially tested to determine whether there was a preference for phosphorylation on Ser or Thr residues. Degenerate peptide libraries containing either a fixed Ser residue, XXXXSXXXX, or a fixed Thr residue, XXXXTXXXX, were incubated with [γ-32P]ATP and Dbf2-Mob1 that was expressed in insect cells, purified, and activated by recombinant Cdc15. All amino acids except Cys, Ser, Thr, and Tyr are represented by X, where the last 3 residues were omitted to limit phosphorylation to the fixed Ser or Thr. The level of phosphorylation was determined by the amount of radioactive phosphate incorporated in the peptides. This analysis suggested that Dbf2-Mob1 had a preference for Ser phosphorylation over Thr (Table 1).
Table 1 Relative phosphate incorporation into peptide libraries by Dbf2-Mob1 kinase complex
Peptide Library Relative Phosphate Incorporation
XXXXTXXXX 1
XXXXSXXXX 2
XXXXRXXSXXXX 16
XXXXSPXXXX 2
Recombinant baculovirus-derived FHHDbf2-H6Mob1TM9 activated by baculovirus-derived Cdc15His6 was used to screen different peptide libraries.
To determine the optimal peptide substrate for Dbf2-Mob1, we screened secondary libraries that contain a fixed residue in addition to the serine phosphorylation site. The rationale for this is articulated in Songyang et. al. [27]. We chose to examine a library with an R fixed at -3 (X4RX2SX4), because Dbf2 resides on the 'AGC' branch of the protein kinase family tree [28]. Other kinases on this branch, including AKT [26] and PKA [27] preferentially phosphorylate substrates with an R at -3. As a control, we also examined a library with the sequence X4SPX4. The X4RX2SX4 library was found to incorporate 8 times more phosphate than the X4SPX4 library (Table 1). As a result, the X4RX2SX4 library was used to determine an optimal Dbf2-Mob1 substrate motif. Sequencing a pool of Dbf2-phosphorylated peptides enriched from the RS library revealed a strong preference for Ile and Phe at the -2 and -1 positions, respectively (Table 2). There was a moderate selection for Met at both the -4 and +1 positions (Table 2). The predicted optimal consensus motif for Dbf2-Mob1 substrates was determined to be MRIFSM.
Table 2 Amino acids selected in a peptide library screen for Dbf2-Mob1 substrates
-7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4
X X X M(1.4) R I(2.8) F(2.8) S M(1.4) X X X
F(1.1) V(1.7) V(1.6) I(1.1)
H(1.7) I(1.3) L(1.1)
M(1.2) M(1.1) V(1.1)
Activated FHHDbf2-H6Mob1TM9 was used to screen the peptide library with the sequence X-X-X-X-R-X-X-S-X-X-X-X. Relative selectivities for amino acids are indicated in parentheses. Bold letters indicate amino acids that are strongly selected; X indicates no selectivity. The one-letter amino acid code is used.
Optimal sequence phosphorylation efficiency
To evaluate the contribution of each residue in the predicted optimal consensus motif, we synthesized a peptide based on the consensus motif, as well as a set of variants in which each position was substituted by an alanine residue (Figure 1A). Peptides with Tyr at -1 (F-1Y) and Lys at -3 (R-3K) were also synthesized to determine whether the selection of Phe at position -1 could be replaced by another bulky residue like Tyr, or if the Arg at position -3 could be substituted with the similarly charged Lys. The NT-Control peptide contains a substitution in the Arg residue that lies outside of the predicted consensus motif to determine whether there is a selection at that position. Finally, we generated a negative control peptide that contains the consensus except that the Ser phosphorylation site was replaced with Ala.
Figure 1 Dbf2-Mob1 peptide substrate requires arginine at position -3. (A) Synthetic peptides based on the predicted optimal substrate of Dbf2-Mob1. The underlined residues represent the predicted preferred amino acids for Dbf2-Mob1 substrate specificity; the asterik denotes the single amino acid substitution. (B) The various peptides denoted in (A) at a concentration of 250 μM were treated with ~13 ng of FHHDbf2 bound to H6Mob1TM9. Aliquots of the kinase reaction were quenched at the indicated timepoints to determine the amount of phosphorylation by liquid scintillation. The Optimal peptide was also treated with the kinase-inactive FHHDbf2(N305A)-H6Mob1TM9 complex, as denoted by D2M. Similar results were obtained in 4 independent experiments. (C) Using the conditions in (B), Km and Vmax was determined for each peptide with the exception of R-3K, R-3A, negative control, and the Optimal peptide treated with kinase-inactive FHHDbf2(N305A)-H6Mob1TM9, due to low phosphorylation.
The synthetic peptides were treated with Dbf2-Mob1 in vitro for 0, 4, 8 or 12 minutes and the amount of phosphorylation was determined (Figure 1B). As a negative control, the optimal consensus peptide was treated in parallel with the kinase-inactive Dbf2(N305A)-Mob1 that had undergone the same treatment with Cdc15 as the wild type Dbf2-Mob1 complex (Opt D2M). As expected, the negative control peptide, as well as the R-3A peptide both had negligible levels of phosphorylation (Figure 1B; Inh, R-3A). The optimal peptide treated with kinase inactive Dbf2(N305A)-Mob1 was also negligibly phosphorylated (Figure 1B; Inh). For the remaining peptides in which measurable levels of phosphorylation were detected, the Vmax and kcat for Dbf2-Mob1 were calculated (Figure 1C).
There did not appear to be selection for Ile at the -2 position nor selection for the Arg in the -7 position outside of the consensus motif, as both mutant peptides were phosphorylated to a similar degree as the optimal peptide (Figure 1B; I-2A, NT, Opt), with similar Vmaxand kcat values for Dbf2-Mob1 (Figure 1C). Mutations of the bulkier groups at positions -1, -4, and +1 to Ala actually increased the amount of phosphorylation of the peptides (Figure 1B; F-1A, M-4A, M+1A), increasing the Vmax and kcat values for Dbf2-Mob1 (Figure 1C). This was surprising because the peptide library screen had indicated a strong selection for Phe at position -1. However, the FLAG peptide used to elute Dbf2-Mob1 may have influenced the selection of phosphopeptides during the screening procedure. Interestingly, the substitution of Arg by Lys in position -3 decreased the amount of peptide phosphorylation to a level comparable to the negative controls (Figure 1B; R-3K). Taken together, these results revealed a preference for non-bulky residues proximal to the Ser phosphorylation site and that the critical Arg required for substrate phosphorylation cannot be substituted with Lys – at least in the context of a peptide substrate.
To confirm the results of the peptide library screen by an independent method, we screened a phosphosite array with active and kinase-dead Dbf2-Mob1 in the presence of [32P]-ATP. The phosphosite array contains 2296 peptides that correspond to annotated sites of phosphorylation in the human proteome [29]. Replicate experiments were performed, and the 'hits' that were obtained with kinase-dead Dbf2-Mob1 were subtracted. Analysis of the top 20 peptide substrates for active Dbf2 (based on intensity of incorporated label) from the first experiment revealed that 18/20 contained a serine phosphorylation site (the remaining two contained threonine), whereas all twenty possessed an R at -3. In the replicate experiment, all eight candidates had an RXXS motif. The only other bias revealed in this analysis – albeit a modest one – was a general preference for R in positions N-terminal to the phosphorylation site. Thus, two different methods – a library-based selection for phosphopeptides and a screen of a phosphopeptide array – suggest that Dbf2 phosphorylates serines spaced three amino acid downstream of an arginine.
Proteome array screen identifies in vitro Dbf2-Mob1 substrates
The relatively low sequence complexity of the Dbf2 phosphorylation motif diminished the power of using genome-wide bioinformatics screens to identify putative substrates. Accordingly, we carried out a proteome array screen to identify putative yeast substrates. Proteome chips spotted with ~4,400 different glutathione-S-transferase (GST) fusion proteins purified from yeast were probed with activated Dbf2-Mob1, kinase-inactive Dbf2(N305A)-Mob1, and buffer alone. After taking into account the proteins phosphorylated in the negative controls, 67 proteins were determined to be putative Dbf2-Mob1 substrates (Table 3).
Table 3 Putative Dbf2-Mob1 substrates from proteome chip screen
ORF Name Common Name Slide Signal (after normalization)
YJR060W CBF1 118.4224
YAL051W OAF1 81.8589
YBR138C HDR1 8.9063
YOL012C HTZ1 4.0738
YMR165C SMP2 3.2182
YPL150W 3.1545
YDR226W ADK1 3.1016
YMR229C RRP5 2.6954
YBR118W TEF2 2.619
YJL108C PRM10 2.5802
YPR091C 2.5784
YKL168C KKQ8 2.5013
YNL101W AVT4 2.1486
YBR285W 1.9355
YMR239C RNT1 1.593
YNR047W 1.4015
YJL076W NET1 1.3521
YNL155W 1.2912
YNR006W VPS27 1.1429
YIL135C VHS2 1.116
YMR184W 1.1144
YDL220C CDC13 1.0274
YBR108W 1.0216
YDL019C OSH2 0.9962
YOR362C PRE10 0.9895
YNL284CA MRPL10 0.9442
YKL077W 0.8905
YDR134C 0.8135
YDL002C NHP10 0.8102
YMR072W ABF2 0.7585
YGR038CA 0.7292
YOR228C 0.6681
YKL140W TGL1 0.6321
YDL070W BDF2 0.6162
YCR105W ADH7 0.6146
YBL106C SRO77 0.5988
YNL125C ESBP6 0.5694
YHR182W 0.4773
YJL213W 0.4755
YDR299W BFR2 0.47
YPL211W NIP7 0.4521
YML037C 0.4503
YDR171W HSP42 0.442
YOL104C NDJ1 0.3611
YKL146W AVT3 0.3592
YGL245W 0.2863
YIL010W DOT5 0.2606
YNL007C SIS1 0.2239
YHL021C 0.2094
YMR196W 0.2089
YJR142W 0.2005
YGR220C MRPL9 0.1976
YLR177W 0.1626
YJL211C 0.1594
YML035C AMD1 0.1384
YGL105W ARC1 0.1332
YGR264C MES1 0.1328
YPL257WA 0.1302
YBL024W NCL1 0.1294
YJR094WA 0.1141
YLR007W NSE1 0.0988
YLR303W MET17 0.0985
YGR223C 0.0837
YKR022C 0.0806
YLL008W DRS1 0.0702
YFR033C QCR6 0.042
YLR004C 0.0228
Activated FHHDbf2-H6Mob1TM9 was used to screen the yeast proteome chip. Of 86 proteins phosphorylated by FHHDbf2-H6Mob1TM9, 67 were determined to be putative substrates after taking into account proteins that were phosphorylated in the control slides treated with the kinase-inactive FHHDbf2(N305A)-H6Mob1TM9.
To confirm that the proteins identified in the proteome array screen could be phosphorylated by Dbf2-Mob1 as opposed to being bound to Dbf2 substrates, we performed further analyses on the 25 proteins with the highest incorporation signal relative to protein amount (relative protein amounts were determined by anti-GST immunoblot of the proteome chip). Interestingly, three or more copies of this motif were found in 16, or 64%, of the proteins in the list (Table 4), compared to only 29% of the proteins encoded in the yeast genome. This enrichment for proteins with 3 or more copies of the RXXS motif is highly significant (p = 1.2 × 10-4; G. Kleiger, unpublished data).
Table 4 Proteins with highest phosphorylation signal from proteome chip screen for Dbf2-Mob1 substrates
ORF Name Common Name Slide Signal (after normalization) MW (kDa) # of R-3 Sites
YJR060W CBF1 118.4224 39 3
YAL051W OAF1 81.8589 121 3
YBR138C HDR1 8.9063 61 2
YOL012C HTZ1 4.0738 14 1
YMR165C SMP2 3.2182 95 7
YPL150W 3.1545 100 16
YDR226W ADK1 3.1016 24 0
YMR229C RRP5 2.6954 193 5
YBR118W TEF2 2.619 50 0
YJL108C PRM10 2.5802 41 0
YPR091C 2.5784 87 3
YKL168C KKQ8 2.5013 84 14
YNL101W AVT4 2.1486 80 7
YBR285W 1.9355 17 0
YMR239C RNT1 1.593 54 2
YNR047W 1.4015 100 21
YJL076W NET1 1.3521 128 11
YNL155W 1.2912 31 1
YNR006W VPS27 1.1429 71 4
YIL135C VHS2 1.116 48 11
YMR184W 1.1144 22 3
YDL220C CDC13 1.0274 105 3
YBR108W 1.0216 93 7
YDL019C OSH2 0.9962 146 6
YOR362C PRE10 0.9895 32 0
Of the 67 putative substrates, the 25 putative substrates with the highest relative amount of phosphorylation signal (amount of signal relative to protein expression) as listed were chosen for further study. MW: predicted molecular weight, # or R-3 Sites: number of RXXS motifs
TAP-tagged strains were obtained for 22 of the 25 top candidates. TAP-tagged alleles for three of the genes (YJL108C, YBR285W, and YBR108W) were not available, and therefore these candidates were not subjected to further analysis. Asynchronous cultures of the 22 TAP-tagged strains were grown and the TAP-tagged proteins immunoprecipitated with IgG sepharose and analyzed by Western blotting (Figure 2A). Candidates were determined to be phosphorylated if a radioactive signal was detected at the molecular weight predicted for the tagged protein. Of the 22 strains in which immunoprecipitations were performed, YBR138C (HDR1), YAL051W (OAF1), YNL101W (AVT4), YIL135C (VHS2), and YMR184W did not have detectable protein expression. The 17 TAP-tagged proteins that were expressed and purified were used in Dbf2-Mob1 kinase assays (Figure 2B). Of the 17, 10 were determined to be phosphorylated. To determine whether the phosphorylation of these proteins was specific to Dbf2-Mob1, rather than due to a co-precipitating protein kinase or residual Cdc15 used to activate Dbf2-Mob1, kinase assays were performed using kinase-inactive Dbf2(N305A-Mob1) as a negative control (Figure 2C). In all cases, there was a strong decrease in incorporation when kinase-inactive Dbf2-Mob1 was used. These results suggest that the proteins identified by the proteome chip screen were indeed in vitro substrates of Dbf2-Mob1.
Figure 2 Yeast proteins phosphorylated by Dbf2-Mob1. (A) Of the 25 proteins with the highest phosphorylation signal as shown in Table 4, 22 of these genes were TAP-tagged in the Open Biosystems TAP-tagged yeast library. The TAP-tagged proteins were immunoprecipitated with IgG sepharose beads from asynchronous cultures, fractionated on SDS-PAGE and immunoblotted with anti-TAP. Of the 22 strains, 5 did not have detectable protein expression, such as VHS2 as shown. (B) The TAP-tagged proteins expressed in (A) were treated with FHHDbf2-H6Mob1TM9 in the presence of [γ-32P]ATP, fractionated on SDS-PAGE and detected by autoradiography. (C) The TAP-tagged proteins phosphorylated by FHHDbf2-H6Mob1TM9 in (B) were treated with either FHHDbf2-H6Mob1TM9 or the kinase inactive FHHDbf2(N305A)-H6Mob1TM9 in the presence of [γ-32P]ATP, fractionated on SDS-PAGE and detected by autoradiography.
In attempts to confirm whether any of the Dbf2-Mob1 substrates identified by our analyses are true physiological substrates of this complex, we took four approaches. First, we examined whether any of the substrates undergo a molecular weight shift upon their phosphorylation by Dbf2-Mob1, which might serve as a simple diagnostic to evaluate phosphorylation in vivo. Next, we immunoprecipitated each protein from yeast cells via the TAP tag, and immunoblotted for Mob1 to determine if the putative substrates were associated with the Dbf2 complex. Third, we queried the yeast GFP localization database, to see if any of the candidates display SPB or bud neck localization characteristic of Dbf2-Mob1. Finally, we searched a database of 700 mapped yeast phosphorylation sites to see if any of them reside in our candidate substrates. Unfortunately, none of these efforts yielded a positive result (A. M., unpublished data). This experience highlights that validation of putative protein kinase substrates identified by proteome chip analysis may require considerable investment in the mapping of in vivo phosphorylation sites.
Discussion
The substrate used to test Dbf2-Mob1 kinase activity, histone H1, is a commonly used artificial substrate for many protein kinases. We wanted to find physiological protein substrates of Dbf2-Mob1. To do so, we first sought to define the optimal phosphorylation site motif for Dbf2-Mob1 substrates. Peptide library screening revealed the putative optimal substrate motif to be MRIFSM (Figure 1A). However, when we tested each residue in in vitro Dbf2-Mob1 peptide kinase assays, the only substitutions that diminished incorporation were swapping the Arg in position -3 for Ala or Lys (Figure 2B). The latter result was unexpected, because Ndr1, a human homologue of Dbf2, was proposed to phosphorylate sequences with either Lys or Arg in the -3 position [30]. Although Dbf2 exhibits a strong requirement for an R at -3, it displays remarkably little bias for any other position adjacent to the phosphorylation site. A similar result was obtained by performing a screen of a phosphosite array composed of peptides that correspond to annotated sites of phosphorylation in the human proteome.
The proteins identified by proteome chip screening gave further evidence that the RXXS motif serves as a substrate for Dbf2-Mob1 phosphorylation, as 80% of the top 25 proteins that were identified in the proteome chip analysis contained this motif. Of these proteins, 17 were tested further and 10 of these were identified as in vitro substrates, 8 of which have the RXXS motif. Pre10 and Adk1 were phosphorylated by Dbf2-Mob1 but not its kinase inactive form despite not having the RXXS motif (Figure 2C; Table 4). One reason may be that both proteins are immunoprecipitated at high levels and therefore may serve as non-specific substrates of Dbf2-Mob1.
Recently, the results of a systematic proteome array analysis of yeast protein kinases have been posted online (J. Ptacek et al., submitted). There were a few minor discrepancies between the results posted online for Dbf2-Mob1 and those of Table 3, with the exception of Cbf1, Oaf1, and Htz1. These 3 proteins were within the top 4 proteins that gave the highest signal in our original data set (Table 3) yet were not in the data set posted online. The discrepancies were based on the methodology for identifying positive signals. In our original data set, the results were obtained by computer analysis. Further visual analysis to confirm positives was performed for the data available online. Upon visual confirmation, Cbf1, Oaf1, and Htz1 were determined to be incorrectly identified by computer analysis (J. Ptacek, personal communication). Our in vitro analysis confirmed that two of these proteins (Oaf1 was not tested because a TAP-tagged strain was not available) were indeed negatives as neither Cbf1 nor Htz1 were found to be phosphorylated by Dbf2-Mob1 (Figure 2B).
Conclusion
We have determined that protein kinase Dbf2-Mob1 has a preference for phosphorylating peptides and proteins that bear one or more RXXS motifs. Although there is strong selection for the R at -3, there is remarkably little selection at any other position, suggesting that Dbf2 achieves specificity by docking its substrates at a site that is distinct from the phosphorylation site.
Methods
Purification and activation of Dbf2-Mob1 kinase complex
FlagHis6HA3Dbf2 (FHHDbf2) was co-immunoprecipitated with His6Mob1TEVmyc9 (H6Mob1TM9) from Hi5 insect cells as previously described [17]. To activate FHHDbf2-H6Mob1TM9, the protein complex bound to anti-FLAG M2 beads (Sigma) was incubated with baculovirus-expressed Cdc15His6 in the presence of kinase buffer containing 50 mM HEPES (pH 7.5), 5 mM MgCl2, 2.5 mM MnCl2, 5 mM β-glycerophosphate, 1 mM DTT, and 1 mM ATP for 30 min at room temperature. The beads were washed three times with buffer containing 50 mM Tris (pH 7.6), 150 mM NaCl, 0.2% Triton X-100 to remove ATP and Cdc15His6. FHHDbf2-H6Mob1TM9 was then eluted from the beads with 1 μg/ml FLAG peptide (Sigma) in Dbf2 kinase buffer (DKB) containing 50 mM Tris (pH 7.4), 60 mM potassium acetate, 10 mM MgCl2, 1 mM DTT, and 10 μM ATP for four hours at 4°C. The same procedure was used to produce FHHDbf2(N305A)-H6Mob1TM9, the kinase inactive point mutant of Dbf2. The eluted active or inactive FHHDbf2-H6Mob1TM9 was used for subsequent assays.
Peptide library screening
Baculovirus-derived FHHDbf2-H6Mob1TM9 was used for peptide library screening. Peptide library screening and data analysis were performed as previously described [26,31,32]. Briefly, the X4RX2SX4 peptide library was used for screening. This library consists of peptides with the general sequence MAXXXXRXXSXXXXAKKK, where X represents all amino acids except Cys, Ser, Thr, and Tyr. The total library is predicted to contain 1.1 × 1012 distinct sequences. Peptides (~1 mg) were incubated with FHHDbf2-H6Mob1TM9, 100 μM unlabeled ATP, and tracer amounts of [γ-32P]ATP at 30°C for 2 h until ~0.5–1% of peptides were phosphorylated. ATP was removed by DEAE-dextran column and a ferric-iminodiacetic acid (IDA) column was then used to separate the phosphopeptides from non-phosphorylated peptides. The phosphopeptides were then sequenced in batch by automated Edman degradation. Data analysis was as described [32]. Selectivity values refer to the relative preference for an amino acid at a given degenerate position, based on the amount of each amino acid recovered at that position compared to the amount in the starting library [31].
Peptide kinase assays
Synthetic peptides (Abgent) were used as substrates for FHHDbf2-H6Mob1TM9 kinase assays. Reactions containing 250 μM of peptide substrate and ~13 ng of FHHDbf2 bound to H6Mob1TM9 were incubated in the presence of 30 μl of DKB and 2 μCi [γ-32P]ATP at room temperature. Reaction aliquots were terminated at indicated timepoints by addition of 10 μl of stop solution (8 N HCl, 1 mM ATP). Phosphate incorporation was determined by spotting reactions on P81-phosphocellulose paper (Whatman), washing with 0.5% phosphoric acid, air-drying the filters, and then quantifying the bound radioactivity by scintillation counting. For each individual peptide, values were normalized to time zero.
Phosphosite array screening
Baculovirus-derived FHHDbf2-H6Mob1TM9 was used for phosphosite array screening. Peptide microarrays displaying phosphosite derived peptides and data analysis were performed as previously described [29,33,34]. Briefly, 2296 peptides derived from human phosphosites (annotated phosphosite in the middle position of 13mer peptide) were printed in triplicates onto aldehyde-modified glass slides [29]. FHHDbf2 bound to H6Mob1TM9 was incubated in the presence of 300 μl of DKB and 20 μCi [γ-32P]ATP at room temperature for 4 hours. Slide was washed 5 times with 25 mL 0.1 M phosphoric acid for 3 minutes followed by washings with 25 mL deionised water. Finally, microarrays were washed with 25 mL methanol and dried at room temperature. Detection of incorporated radioactivity was performed by exposition of the microarrays for 8 hours to a BAS-MS imaging plate (Fuji Photo Film Co., Ltd., Japan) followed by readout with a FLA-3000 Phosphor Imager (Fuji, Japan). Data evaluation was carried out using ArrayPro software (Media Cybernetics, Silver Spring, MD, USA). Similar experiments with the kinase dead mutant were performed. Signals with high intensity in the active kinase experiment but low intensity in the kinase dead control were considered as peptide substrates specific for the Dbf2-Mob1 complex.
Proteome chip assays
Yeast proteome microarrays were prepared as previously described [35]. Overexpressed GST-tagged yeast proteins were purified from ~4,400 yeast strains and spotted on slides. To determine the optimal amount of kinase to use for probing proteome chips, we performed trial assays as described (J. Ptacek et al., submitted). Multiple dilutions of FHHDbf2-H6Mob1TM9 and kinase inactive FHHDbf2(N305A)-H6Mob1TM9 (~20 ng/μl of Dbf2) in DKB buffer containing 2 μl [γ-33P]ATP were used on trial proteome chips before using on the full proteome array. The full proteome array was probed with 4 μl FHHDbf2-H6Mob1TM9 in 200 μl of DKB supplemented with [γ-33P]ATP. As a control, FHHDbf2(N305A)-H6Mob1TM9 was used to probe the proteome array. To normalize the background signal, 2 μl of the kinase-inactive complex in 200 μl DKB supplemented with [γ-33P]ATP was used. To control for autophosphorylated proteins, the proteome array was probed with 200 μl of DKB supplemented with [γ-33P]ATP. Proteome chips were assayed in duplicate in each case. Data analysis was performed as described (J. Ptacek et al., submitted). Briefly, signals were analyzed by a computer algorithm designed to normalize background and identify signals as positive if 3 of 4 spots (each protein is spotted twice on each slide and each kinase or control was tested on 2 slides) were 2 standard deviations above background and the fourth spot was 1.5 standard deviations above background.
Immunoprecipitations and kinase assays of TAP-tagged proteins
Yeast strains containing TAP-tagged genes (Open Biosystems) were grown to OD600 ~2.0 in 25 mL of YPD. Cells were harvested by centrifugation and washed in buffer containing 150 mM NaCl and 50 mM Tris (pH 7.4). Cells were then resuspended in 600 μl lysis buffer containing 150 mM NaCl, 50 mM Tris (pH 7.4), 2 mM EDTA (pH 8.0), 1% Triton-X 100, 10% glycerol, 2 mM DTT, 5 μg/ml aprotinin, 5 μg/ml pepstatin, 5 μg/ml chymostatin, 5 μg/ml leupeptin, 0.5 mM AEBSF, 1 mM PMSF, 10 mM NaF, 60 mM β-glycerophosphate, 10 mM sodium pyrophosphate, 2 mM sodium vanadate). An equal volume of glass beads was added. The cells were then lysed by 4 cycles of vortexing (ThermoSavant FastPrep) at 4°C for 45 s at setting 5.5 alternating with cycles of icing samples for 1 min. Lysates were clarified by centrifugation then added to 60 μl IgG sepharose beads (Amersham) for 1 h at 4°C on a rotator. Beads were then washed 3 times with lysis buffer, twice with buffer containing 150 mM NaCl, 50 mM Tris (pH 7.4), 2 mM EDTA (pH 8.0), 1% Triton-X 100, 10% glycerol, and 2 mM DTT, and a final wash with buffer containing 150 mM NaCl and 50 mM Tris (pH 7.4). Immunoprecipitated TAP-tagged proteins were analyzed by SDS-PAGE and detected by Western blotting using the primary antibody anti-TAP (Open Biosystems) followed by goat anti-rabbit horseradish peroxidase (HRP)-conjugate (Bio-Rad), and ECL. For kinase assays, TAP-tagged proteins bound to 20 μl beads were washed with DKB then incubated with FHHDbf2-H6Mob1TM9 or FHHDbf2(N305A)-H6Mob1TM9 (~13 ng of Dbf2) with 2 μCi [γ-32P]ATP for 30 min at room temperature. Kinase reactions were stopped by addition of 2X SDS-PAGE sample buffer, fractionated on SDS-PAGE and detected by autoradiography.
List of abbreviations
MEN, mitotic exit network; FEAR, Cdc14 early anaphase release; GEF, guanine-nucleotide exchange factor; GAP, GTPase activating protein; SPB, spindle pole body; GST, gluathione-S-transferase; IDA, iminodiacetic acid; FHH, FlagHis6HA3; H6, His6; TM9, TEV-myc9; DKB, Dbf2 kinase buffer
Authors' contributions
ASM performed the peptide kinase assays and analysis, the immunoprecipitations and kinase assays of the TAP-tagged strains, preparation of recombinant Dbf2-Mob1 complexes used throughout the work, and preparation of the manuscript. AEHE carried out the peptide library screening and analysis. GD and JP carried out the proteome chip studies. MS performed the phosphosite array screening. MS, MBY, and RJD contributed to the experimental design, analysis, and interpretation.
Acknowledgements
We thank Ramzi Azzam for his invaluable insight and enthusiasm for initiating this project. We also thank Dane Mohl and William Ja for their thoughts and comments on this work. We are also grateful to Heng Zhu for performing initial proteome chip experiments and Gary Kleiger for providing bioinformatics expertise. This research was supported by an NIH grant to RJD (GM059940).
==== Refs
D'Amours D Amon A At the interface between signaling and executing anaphase – Cdc14 and the FEAR network Genes Dev 2004 18 2581 2595 15520278 10.1101/gad.1247304
Schwab M Lutum AS Seufert W Yeast Hct1 is a regulator of Clb2 cyclin proteolysis Cell 1997 90 683 693 9288748 10.1016/S0092-8674(00)80529-2
Visintin R Craig K Hwang ES Prinz S Tyers M Amon A The phosphatase Cdc14 triggers mitotic exit by reversal of Cdk-dependent phosphorylation Mol Cell 1998 2 709 718 9885559 10.1016/S1097-2765(00)80286-5
Zachariae W Schwab M Nasmyth K Seufert W Control of cyclin ubiquitination by CDK-regulated binding of Hct1 to the anaphase promoting complex Science 1998 282 1721 1724 9831566 10.1126/science.282.5394.1721
Jaspersen SL Charles JF Morgan DO Inhibitory phosphorylation of the APC regulator Hct1 is controlled by the kinase Cdc28 and the phosphatase Cdc14 Curr Biol 1999 9 227 236 10074450 10.1016/S0960-9822(99)80111-0
Moll T Tebb G Surana U Robitsch H Nasmyth K The role of phosphorylation and the CDC28 protein kinase in cell cycle-regulated nuclear import of the S. cerevisiae transcription factor SWI5 Cell 1991 66 743 758 1652372 10.1016/0092-8674(91)90118-I
Shou W Seol JH Shevchenko A Baskerville C Moazed D Chen ZW Jang J Charbonneau H Deshaies RJ Exit from mitosis is triggered by Tem1- dependent release of the protein phosphatase Cdc14 from nucleolar RENT complex Cell 1999 97 233 244 10219244 10.1016/S0092-8674(00)80733-3
Visintin R Hwang ES Amon A Cfi1 prevents premature exit from mitosis by anchoring Cdc14 phosphatase in the nucleolus Nature 1999 398 818 823 10235265 10.1038/19775
Stegmeier F Amon A Closing Mitosis: The Functions of the Cdc14 Phosphatase and Its Regulation Annu Rev Genet 2004 38 203 232 15568976 10.1146/annurev.genet.38.072902.093051
Azzam R Chen SL Shou W Mah AS Alexandru G Nasmyth K Annan RS Carr SA Deshaies RJ Phosphorylation by cyclin B-Cdk underlies release of mitotic exit activator Cdc14 from the nucleolus Science 2004 305 516 519 15273393 10.1126/science.1099402
Stegmeier F Visintin R Amon A Separase, polo kinase, the kinetochore protein Slk19, and Spo12 function in a network that controls Cdc14 localization during early anaphase Cell 2002 108 207 220 11832211 10.1016/S0092-8674(02)00618-9
Sullivan M Uhlmann F non-proteolytic function of separase links the onset of anaphase to mitotic exit Nat Cell Biol 2003 5 249 254 12598903 10.1038/ncb940
Stegmeier F Huang J Rahal R Zmolik J Moazed D Amon A The replication fork block protein Fob1 functions as a negative regulator of the FEAR network Curr Biol 2004 14 467 480 15043811 10.1016/j.cub.2004.03.009
Bardin AJ Amon A Men and sin: what's the difference? Nat Rev Mol Cell Biol 2001 2 815 826 11715048 10.1038/35099020
Pereira G Hofken T Grindlay J Manson C Schiebel E The Bub2p spindle checkpoint links nuclear migration with mitotic exit Mol Cell 2000 6 1 10 10949022 10.1016/S1097-2765(00)00002-2
Bardin AJ Visintin R Amon A A mechanism for coupling exit from mitosis to partitioning of the nucleus Cell 2000 102 21 31 10929710 10.1016/S0092-8674(00)00007-6
Mah AS Jang J Deshaies RJ Protein kinase Cdc15 activates the Dbf2-Mob1 kinase complex Proc Natl Acad Sci USA 2001 98 7325 7330 11404483 10.1073/pnas.141098998
Xu S Huang HK Kaiser P Latterich M Hunter T Phosphorylation and spindle pole body localization of the Cdc15p mitotic regulatory protein kinase in budding yeast Curr Biol 2000 10 329 332 10744974 10.1016/S0960-9822(00)00382-1
Frenz LM Lee SE Fesquet D Johnston LH The budding yeast Dbf2 protein kinase localises to the centrosome and moves to the bud neck in late mitosis J Cell Sci 2000 113 3399 3408 10984431
Visintin R Amon A Regulation of the mitotic exit protein kinases Cdc15 and Dbf2 Mol Biol Cell 2001 12 2961 2974 11598184
Hwa Lim H Yeong FM Surana U Inactivation of mitotic kinase triggers translocation of MEN components to mother-daughter neck in yeast Mol Biol Cell 2003 14 4734 4743 12937277 10.1091/mbc.E03-04-0238
Yoshida S Toh-e A Regulation of the localization of Dbf2 and Mob1 during cell division of Saccharomyces cerevisiae Genes Genet Syst 2001 76 141 147 11434459 10.1266/ggs.76.141
Luca FC Mody M Kurischko C Roof DM Giddings TH Winey M Saccharomyces cerevisiae Mob1p is required for cytokinesis and mitotic exit Mol Cell Biol 2001 21 6972 6983 11564880 10.1128/MCB.21.20.6972-6983.2001
Lippincott J Shannon KB Shou W Deshaies RJ Li R The Tem1 small GTPase controls actomyosin and septin dynamics during cytokinesis J Cell Sci 2001 114 1379 1386 11257003
Frenz LM Lee SE Fesquet D Johnston LH The budding yeast Dbf2 protein kinase localises to the centrosome and moves to the bud neck in late mitosis J Cell Sci 2000 113 3399 3408 10984431
Obata T Yaffe MB Leparc GG Piro ET Maegawa H Kashiwagi A Kikkawa R Cantley LC Peptide and protein library screening defines optimal substrate motifs for AKT/PKB J Biol Chem 2000 275 36108 36115 10945990 10.1074/jbc.M005497200
Songyang Z Blechner S Hoagland N Hoekstra MF Piwnica-Worms H Cantley LC Use of an oriented peptide library to determine the optimal substrates of protein kinases Curr Biol 1994 4 973 982 7874496 10.1016/S0960-9822(00)00221-9
Hunter T Plowman GD The protein kinases of budding yeast: six score and more Trends Biochem Sci 1997 22 18 22 9020587 10.1016/S0968-0004(96)10068-2
Schutkowski M Reimer U Panse S Dong L Lizcano JM Alessi DR Schneider-Mergener J High-Content Peptide Microarrays for Deciphering Kinase Specificity and Biology Angew Chem 2004 116 2725 2728 10.1002/ange.200453900
Millward TA Heizmann CW Schafer BW Hemmings BA Calcium regulation of Ndr protein kinase mediated by S100 calcium-binding proteins EMBO J 1998 17 5913 5922 9774336 10.1093/emboj/17.20.5913
Songyang Z Cantley LC The use of peptide library for the determination of kinase peptide substrates Methods Mol Biol 1998 87 87 98 9523263
Yaffe MB Schutkowski M Shen M Zhou XZ Stukenberg PT Rahfeld JU Xu J Kuang J Kirschner MW Fischer G Sequence-specific and phosphorylation-dependent proline isomerization: a potential mitotic regulatory mechanism Science 1997 278 1957 1960 9395400 10.1126/science.278.5345.1957
Panse S Dong L Carus R Schutkowski M Reimer U Schneider-Mergener J Profiling of generic and anti-phosphopeptide antibodies and kinases with peptide microarrays using radioactive and fluorescence-based assays Molecular Diversity 2004 8 291 299 15384422 10.1023/B:MODI.0000036240.39384.eb
Rychlewski L Kschischo M Dong L Schutkowski M Reimer U Target specificity analysis of the Abl kinase using peptide microarray data J Mol Biol 2004 336 307 311 14757045 10.1016/j.jmb.2003.12.052
Zhu H Snyder M Protein arrays and microarrays Curr Opin Chem Biol 2001 5 40 45 11166646 10.1016/S1367-5931(00)00170-8
|
16242037
|
PMC1277818
|
CC BY
|
2021-01-04 16:26:24
|
no
|
BMC Biochem. 2005 Oct 21; 6:22
|
utf-8
|
BMC Biochem
| 2,005 |
10.1186/1471-2091-6-22
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2481622130910.1186/1471-2105-6-248Research ArticleBetter prediction of protein contact number using a support vector regression analysis of amino acid sequence Yuan Zheng [email protected] Institute for Molecular Bioscience and ARC Centre in Bioinformatics, The University of Queensland, St. Lucia, 4072, Australia2005 13 10 2005 6 248 248 4 7 2005 13 10 2005 Copyright © 2005 Yuan; licensee BioMed Central Ltd.2005Yuan; 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
Protein tertiary structure can be partly characterized via each amino acid's contact number measuring how residues are spatially arranged. The contact number of a residue in a folded protein is a measure of its exposure to the local environment, and is defined as the number of Cβ atoms in other residues within a sphere around the Cβ atom of the residue of interest. Contact number is partly conserved between protein folds and thus is useful for protein fold and structure prediction. In turn, each residue's contact number can be partially predicted from primary amino acid sequence, assisting tertiary fold analysis from sequence data. In this study, we provide a more accurate contact number prediction method from protein primary sequence.
Results
We predict contact number from protein sequence using a novel support vector regression algorithm. Using protein local sequences with multiple sequence alignments (PSI-BLAST profiles), we demonstrate a correlation coefficient between predicted and observed contact numbers of 0.70, which outperforms previously achieved accuracies. Including additional information about sequence weight and amino acid composition further improves prediction accuracies significantly with the correlation coefficient reaching 0.73. If residues are classified as being either "contacted" or "non-contacted", the prediction accuracies are all greater than 77%, regardless of the choice of classification thresholds.
Conclusion
The successful application of support vector regression to the prediction of protein contact number reported here, together with previous applications of this approach to the prediction of protein accessible surface area and B-factor profile, suggests that a support vector regression approach may be very useful for determining the structure-function relation between primary protein sequence and higher order consecutive protein structural and functional properties.
==== Body
Background
Prediction of protein three-dimensional structure from primary sequence is the central problem in structural bioinformatics. One approach is to use known structur∈homolog proteins as templates to determine the tertiary structures of novel proteins of unknown structure. Approaches include comparative modelling, threading and fold recognition methods. One protein structural feature is of particular interest here, namely, residue contact number (CN) which can be used to enhance protein fold recognition [1]. This measure has also been regarded as the conserved solvent exposure descriptor of similar folds without a common evolutionary origin [2]. Furthermore, contact number may be used to determine the energy function allowing molecular dynamics simulations of protein structures [3]. Here, we seek to use protein contact number to assist with the tertiary fold prediction of novel proteins for which an accurate functional relationship between a protein's primary sequence and its residues' contact numbers must be determined. To fulfil the task, we use a new method, the so-called support vector regression, to approximate the sequence-contact number relationship. We demonstrate that, as a result, we achieve more accurate predicted contact numbers than have been achieved to date.
The contact number, or coordination number, of a given residue of a folded protein is defined as the number of Cβ (or Cα) atoms in other residues within a sphere around the Cβ (or Cα) atom of that given residue. Previous approaches to the prediction of protein contact number fall into two categories: classification and regression. In the classification approach, residue contact numbers were first classified into two populations allowing a subsequent use of machine learning methods such as recurrent neural networks to perform predictions [4,5]. Unfortunately, decomposing contact numbers into two states via an arbitrary threshold oversimplifies the problem and much original information is lost. In contrast, the regression approach provides a direct and more accurate way to determine a functional relationship matching contact numbers and protein sequence and thus to provide more accurate contact number predictions. A recent study of Kinjo et al. [3], followed this approach but used a simple linear regression scheme to determine the functional relationship. They reported that the predicted and observe contact numbers had a correlation coefficient (CC) of 0.627. However, most functions in nature are non-linear and cannot be accurately approximated by linear formulas. Under the reasonable expectation that the sequence-contact number is indeed nonlinear, we use a more complicated machine learning method to determine the relationship and expect thereby to obtain more accurate predictions. In particular, we adopt a support vector regression (SVR) algorithm fully capable of determining a non-linear sequence-contact number relationship.
In our former work, we studied the dependence of protein accessible surface area (ASA) [6,7] and B-factor [8] on primary sequence. These works established that ASAs can be predicted and match observed values with a correlation coefficient of 0.69, while B-factors can be predicted and match observed values with a correlation coefficient of 0.53. These approaches established that multiple sequence inputs outperform single sequence inputs significantly. The importance of using multiple sequences was also observed in prior predictions of contact numbers [3]. Thus, in this present work, we focus on multiple sequence inputs. For completeness, we examine a range of different definitions of contact number ("consecutive" and "discrete"), and also examine whether including further information such as protein molecular weight and amino acid composition allows improved predictions. As a result, we are able to make predictions which match observed values with a correlation coefficient of 0.73, a significant improvement on earlier studies.
Results
Contact numbers according to different rd values
We give 8 definitions of contact number and show their CN distributions in Fig. 1. For each definition, we compute the mean and standard deviation (Table 1). For the same radius cutoff rd, the discrete and consecutive definitions have very similar distributions with nearly the same mean and standard deviation. Their correlations are greater than 0.99 for all values of rd (8 Å, 10 Å, 12 Å and 14 Å). The contact numbers defined by different radius cutoffs have CCs greater than 0.83. Distributions with larger radius cutoffs more closely approximate normal distributions as their left-hand tails are almost all present. Since absolute contact numbers are normalized by a linear transformation (Equation 3), the general characteristics of their distributions will still be kept even after the normalization.
Figure 1 Contact number distributions according to different definitions. The radius cutoffs are selected as 8 Å, 10 Å, 12 Å and 14 Å, represented by dotted, slashed, solid and dot-and-slashed lines, respectively. A is for discrete contact number while B is for consecutive contact number.
Table 1 The mean (N¯
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaqdaaqaaiabd6eaobaaaaa@2DE2@) and standard deviation (SD) of contact numbers according to different radius (rd) cutoffs. All results are expressed as (N¯
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaqdaaqaaiabd6eaobaaaaa@2DE2@, SD).
rd = 8 Å rd = 10 Å rd = 12 Å rd = 14 Å
Discrete 6.14, 3.29 12.90, 6.19 23.50, 10.46 35.39,15.39
Consecutive 6.27, 3.25 13.07, 6.14 23.56, 10.41 35.53, 15.39
To study the relationship between CN and ASA of a residue, we obtained the ASA for each residue in the 945 proteins using the DSSP program [9]. Using discrete definition of CN with a radius cutoff of 12 Å, we calculate the mean and standard deviation of ASA for each contact number and show the results in Fig. 2. A strong negative correlation between the two solvent exposure descriptors can be observed as indicated by a correlation coefficient of -0.734.
Figure 2 The accessible surface area as a function of contact number. Discrete contact numbers are used with a radius cutoff of 12 Å. Error bars represent the standard deviations.
Estimating the sequence-contact number relationship
When we train the SVR algorithm, normalized CNs are used instead of absolute CNs because the normalized values are always located between -3.0 and 3.0 for all radius cutoffs. Therefore, in all cases, the same set of SVR learning control parameters can be applied. Among the three groups of proteins (each has 315 chains), we estimate the sequence-contact number function in turn using one group and examine the estimated function using the remaining two groups. The correlation coefficient and root mean square error (RMSE) are computed for each test group, and their averages are shown in Table 2.
Table 2 Correlation coefficient (CC) and root mean square error (RMSE) for different contact number predictions. All results are expressed as mean ± standard deviation.
rd = 8 Å rd = 10 Å rd = 12 Å rd = 14 Å
Discrete CC 0.64 ± 0.01 0.66 ± 0.01 0.69 ± 0.01 0.69 ± 0.01
RMSE 0.77 ± 0.01 0.75 ± 0.01 0.72 ± 0.01 0.72 ± 0.02
Consecutive CC 0.66 ± 0.01 0.67 ± 0.01 0.70 ± 0.01 0.70 ± 0.01
RMSE 0.75 ± 0.01 0.74 ± 0.01 0.72 ± 0.01 0.72 ± 0.02
For all radius cutoffs, predictions using consecutive contact numbers are slightly better than predictions using their discrete counterparts. However, when the larger thresholds (e.g. 12 Å and 14 Å) are used, the accuracy difference decreases to insignificance. Previous work has shown that CNs with a radius cutoff of 12 Å or 14 Å are more useful for protein fold recognition [1]. Likewise, in this work, we also find that these cutoffs give better predictions. But, compared with the discrete contact numbers, the consecutive contact numbers give only a very slight improvement in predictions. The best accuracies are for consecutive contact numbers with thresholds of 12 Å and 14 Å, in which case the correlation coefficient between predicted and observed values can reach 0.70. If we convert the normalized contact numbers to their original absolute ones, the RMSE of 0.72 is equal to an actual error of 7.5 for a threshold of 12 Å, and an actual error of 11.1 for a threshold of 14 Å.
We also calculate the CC and RMSE for each individual protein using discrete contact numbers with a radius cutoff of 12 Å. The average CC and RMSE are then 0.67 and 7.31, respectively. More than half of the proteins are predicted with CCs greater than 0.70, and more than half are predicted with RMSEs less than 6.93. To illustrate the meaning of the CC and RMSE measures in this study, two comparisons of predicted and observed values are given in Figure 3. This figure shows the better agreement between the predicted and observed values in GP130 (PDB: 1bj8) as the CC is 0.75 and the RMSE is 6.07. In contrast, the prediction for human chorionic gonadotropin (PDB: 1dz7, chain A) yields a correlation coefficient of 0.58 and a RMSE of 9.73 with the region between position 40 and 50 being worst predicted.
Figure 3 The predicted and observed contact numbers for proteins GP130 (PDB: 1bj8) and Human chorionic gonadotropin (PDB: 1dz7, chain A). Discrete contact numbers are used with a radius cutoff of 12 Å. Observed and predicted contact numbers are represented by solid and dashed lines, respectively. A) GP 130 is predicted with a correlation coefficient of 0.75 and a root-mean-squar∈error of 6.07; B) Human chorionic gonadotropin is predicted with a correlation coefficient of 0.58 and a root mean square error of 9.73.
Sequence weight, as a feature, can improve prediction accuracy significantly
Prediction accuracy can be improved by taking account of protein size as measured by its weight. Given a sequence, the weight is the sum of individual weights of consisting amino acids. We use discrete contact numbers (radius cutoff = 12 Å) and divide all proteins into three groups with equal number of proteins, according to their weights. For the three groups, the weights (mean ± standard deviation) are 10611 ± 1637, 17421 ± 2666 and 40073 ± 5952 Daltons. Their average correlation coefficient between predicted and observed contact numbers are 0.61, 0.67 and 0.72, while their average RMSEs are 7.49, 7.09 and 7.35, respectively. These results suggest that smaller molecules are worst predicted.
To consider this effect, we calculate weight for each protein sequence and include this data as an additional input to the machine learning algorithm and re-run the training and testing procedures. Additional information, that of protein amino acid composition, was also included as an input, either individually or together with sequence weight data. For the separate groups of small chains, median chains, large chains and all chains, we calculate the mean and median values of the correlation coefficient between predicted and observed values of contact number, and their RMSEs, according to each set of different input information: local sequence ("LS"), local sequence plus sequence weight ("LS+W"), local sequence plus amino acid composition ("LS+AA") and local sequence plus sequence weight and amino acid composition ("LS+W+AA"). All results are shown in Table 3.
Table 3 Correlation coefficients (CCs) and root mean square errors (RMSEs) for individual proteins in different weight groups. The results are given as (mean, median).
LS LS+W LS+AA LS+W+AA
W≤3485 CC 0.61, 0.65 0.64, 0.67 0.62, 0.66 0.64, 0.67
RMSE 7.49, 6.95 6.68, 6.41 7.24, 6.82 6.71, 6.45
13485<W≤22750 CC 0.67, 0.70 0.68, 0.71 0.68, 0.71 0.69, 0.72
RMSE 7.09, 6.79 6.76, 6.54 7.05, 6.79 6.76, 6.60
W>22750 CC 0.72, 0.73 0.72, 0.74 0.72, 0.73 0.73, 0.74
RMSE 7.35, 7.07 7.12, 6.95 7.24, 6.94 7.10, 6.90
All CC 0.67, 0.70 0.68, 0.71 0.68, 0.71 0.68, 0.71
RMSE 7.31, 6.94 6.86, 6.66 7.18, 6.86 6.86, 6.66
For all cases, it was determined that amino acid composition information can improve prediction performance. However, sequence weight can give yet more significant improvements. For example, in the group of small molecules, data about amino acid composition can increase the CC mean to 0.62 and decrease the RMSE mean to 7.24, while sequence weight data can increase the CC mean to 0.64 and decrease the RMSE mean to 6.68. When information about both sequence weight and amino acid composition is used together, we find still further improvement compared with using each data individually, although this may not be reflected by all measures. Particularly, the difference between "LS+W" and "LS+W+AA" is very minor. However, all the results clearly show that sequence weight is more important than amino acid composition in the prediction of contact numbers.
Fig. 4 gives the overall distributions of CCs and RMSEs for 945 proteins. Compared with the CC, the RMSE is more sensitive in that its distribution more clearly reflects the improvement, while the distributions of "LS+W" and "LS+W+AA" are nearly identical. For all cases, the peak values of CC and RMSE are around 0.70 and 6.0, respectively.
Figure 4 Distributions of correlation coefficients and root mean square errors given different input information. Discrete definition is used with a radius cutoff of 12 Å. The four inputs "LS", "LS+W", "LS+AA" and "LS+W+AA" are represented by dotted, slashed, dot-and-slashed and solid lines, respectively. A is for correlation coefficients while B is for root mean square errors.
In addition to the above analyses based on individual proteins, we also measure the accuracies on protein residues in the whole data set. We calculate CCs and normalized RMSEs for six testing groups, and express them as mean ± standard deviation, as given in Table 4. The CCs for "LS", "LS+W", "LS+AA" and "LS+W+AA" are 0.69, 0.733, 0.708 and 0.734, respectively. The normalized RMSEs are 0.72, 0.68, 0.71 and 0.68, respectively. Therefore, an obvious improvement can be found even if we measure it at the residue level.
Table 4 Correlation coefficients (CCs) and root mean square errors (RMSEs) are calculated for all residues when performing support vector regression algorithm using different input information.
CC RMSE
LS 0.69 ± 0.01 0.72 ± 0.01
LS+W 0.733 ± 0.005 0.68 ± 0.01
LS+AA 0.708 ± 0.009 0.71 ± 0.01
LS+W+AA 0.734 ± 0.006 0.68 ± 0.01
To measure the prediction performance for residues with different contact numbers, we compute the absolute errors for the residue with contact numbers from 0 to 60. The mean absolute errors for certain contact numbers are shown in Fig. 5, partitioned according to the four information inputs. Clearly, "LS+W+AA" gives the least absolute errors and therefore perform the best. Residues with about 20 contacting Cβ atoms are predicted with the least mean absolute error (4.1) and are the best predicted. This is because they have the largest number of samples in the dataset. Greater errors are found at each tail-end of the distribution corresponding to the residues with smaller or greater contact numbers. This is due to the small number of data points in each tail fed into support vector machines, and their representation is not adequate. Furthermore, this can be used to explain why the region from residue 40 to residue 50 of protein 1dz7 in Fig 3B is the worst predicted. This region mostly contains exposed residues with smaller contact numbers and thus, the residues cannot be well predicted by our method. An improvement on this part may be achieved by using other predicted features such as accessible surface area.
Figure 5 The mean absolute errors for residues of different contact numbers. The four inputs "LS", "LS+W", "LS+AA" and "LS+W+AA" are represented by dotted, slashed, dot-and-slashed and solid lines, respectively.
Examining performance by formulating regression as a two-class problem
CN prediction has previously been examined as a two-class classification problem through use of a threshold with the accuracy being defined as the percentage of the correctly predicted residues on the overall residues [3-5]. However, this is not a good measure because the accuracy is susceptible to changes in the selected threshold that splits the data set. If the data set is heavily unbalanced, accuracy is always very high [10]. In this study, we use a different measure, and adopt the least (worst) prediction accuracy for each case to reflect its performance. Table 5 gives the least two-class prediction accuracies for eight definitions of contact number when using only local sequence information. Note that all the accuracies are the average of six tests. All accuracies are found to be greater than 74%, and in particular, when rd = 12 Å and a consecutive contact number definition is adopted, the least prediction accuracy is around 76.1%, which is comparable with the accuracy 76.3% recently reported based on choosing a particular threshold in linear models also using the same consecutive contact number definition [3].
Table 5 The least prediction accuracy (%) for two-class problems according to different contact number definitions. Only local sequence information is used.
rd = 8 Å rd = 10 Å rd = 12 Å rd = 14 Å
Discrete 74.1 74.5 75.8 75.2
Consecutive 75.6 75.7 76.1 75.6
We use discrete definitions of contact number and let rd = 12 Å. Using all the SVM outputs from six tests, we choose a number of thresholds to classify the data points as being either "contacted" or "non-contacted" and calculate their accuracies. All accuracies are plotted in Fig. 6, according to different information input. The least accuracies for "LS", "LS+W", "LS+AA" and "LS+W+AA" are 75.8%, 77.2%, 76.1% and 77.1%, respectively. Using sequence weight is much better than using amino acid composition.
Figure 6 Prediction accuracies when predictions are formulated as two-class problems using different contact number thresholds. The four inputs "LS", "LS+W", "LS+AA" and "LS+W+AA" are represented by dotted, slashed, dot-and-slashed and solid lines, respectively.
Discussion
Protein structural properties such as secondary structure, solvent accessibility and contact number provide valuable information for prediction of protein tertiary structures. How to improve the prediction accuracy of these parameters is still a challenging problem. Following Rost and Sander's pioneering work [11] on how to find a conserved and useful prediction index, Hamelryck [2] examined the conservation of nine solvent-exposure measures and found that contact number is the most conserved (correlation coefficient 0.72). His study suggested that CN is more suitable for fold recognition than other descriptors such as ASA. However, difficulties in accurately expressing the prediction problem (for example, it was previously framed as a two class problem using an arbitrary threshold) limited its further application. Recent work on contact number [3] formulating the problem for a regression analysis has enhanced studies in this area. From our work here, we confirm the utility of a regression analysis, and more specifically, establish that allowing for non-linearity via support vector regression allows a more accurate determination of the sequence-contact number relation which further illuminates relationships between protein structural and functional properties and their primary sequence and other features.
Conclusion
We provide a new method for the prediction of protein contact number. Using protein local sequence information generated by multiple sequence alignments, the correlation coefficient between predicted and observed contact numbers can reach 0.70, with normalized root mean square error less than 0.72. The addition of information about sequence weight and amino acid composition as input features can increase the correlation coefficient to 0.734 and decrease the root mean square error to 0.68. This improvement is mainly attributed to the information about sequence weight while the information about amino acid composition only contributes slightly. Moreover, more than half of the proteins are predicted with correlation coefficients greater than 0.71. The prediction accuracies in the two-class problems, regardless of the cutoff thresholds, are greater than 77.0%. The successful application of SVR approach in this study suggests that it can more accurately describe the relationship between protein contact numbers and primary sequence.
Methods
Residue contact number
We take two definitions of contact number in this study, namely, that of "discrete" and "consecutive" contact number. The "discrete" contact number, Nd, is defined by the number of Cβ atoms on other residues located within a sphere of radius rd centred on the Cβ atom of the residue of interest. The discrete contact number for i-th residue in a sequence with M residues is given by
Ndi=∑j:|j−i|>2Mσ(ri,j){σ(ri,j)=1ifri,j<rdσ(ri,j)=0ifri,j≥rd, (1)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGobGtdaqhaaqaaiabdsgaKbqaaiabdMgaPbaacqGH9aqpdaaeWbqaaGGaciab=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@8A4A@
where ri,j is the distance between the Cβ atoms of the ith and jth residues which are understood to be separated in sequence by at least two amino acids. Note that Ndi
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGobGtdaqhaaWcbaGaemizaqgabaGaemyAaKgaaaaa@30AA@ is a discrete integer. By replacing the step function σ(ri,j) with a sigmoid function, Ndi
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGobGtdaqhaaWcbaGaemizaqgabaGaemyAaKgaaaaa@30AA@ becomes a real number. This procedure was previously adopted by Kinjo et al. [3] to smooth the discrete contact numbers. A particular sigmoid function is given by
σ(ri,j) = 1/{1 + exp [3(ri,j - rd)]}. (2)
We have tried four values of rd (8 Å, 10 Å, 12 Å and 14 Å) with discrete and consecutive definitions and thus have 8 combinations all of which will be used in our SVR approach.
Normalization of contact number
The distributions of contact numbers can be approximated by normal distributions, as shown in Fig. 1. With respect to a certain rd, we calculate the mean (N¯
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaqdaaqaaiabd6eaobaaaaa@2DE2@) and standard deviation (SD). So, the normalized contact number Nnorm is determined by the following formula:
Nnorm=N−N¯SD. (3)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGobGtdaWgaaWcbaGaemOBa4Maem4Ba8MaemOCaiNaemyBa0gabeaakiabg2da9maalaaabaGaemOta4KaeyOeI0Yaa0aaaeaacqWGobGtaaaabaGaem4uamLaemiraqeaaiabc6caUiaaxMaacaWLjaWaaeWaaeaacqaIZaWmaiaawIcacaGLPaaaaaa@3EE6@
At the first step, we predict the normalized contact number because 1) it is easy to handle the data, and 2) it is easy to compare the results for different rd thresholds. At the second step, we recover the absolute contact numbers from their predicted normalized values using this equation.
Sequence coding
We predict contact number from protein local sequence. For a given residue, the local sequence contains its N-terminal and C-terminal seven nearest-neighbour residues. Thus, the local sequence makes a window of fifteen amino acids. We code each residue in the window using the PSI-BLAST position-specific scoring matrix [12]. The matrices are obtained by querying the input sequence using PSI-BLAST against the NCBI non-redundant protein sequence database with three rounds, masking coil-coiled and low-complexity regions [13]. The elements in the row of the matrix reflect the probabilities for 20 amino acids occurring at this position. All the elements are divided by 10 for normalization and thus each residue is represented by a 20-dimesional vector. Since the residues in coil-coil and low-complexity regions do not have meaningful scores, we encode the residue with an orthogonal scheme. In the 20-dimensional vector coding a given residue, only the entry representing this type of amino acid is assigned as 0.5 with all other entries set as zeros. To consider the terminal residues, we expend the 20-dimensional vector to being 21-dimensional for all residues. When the last entry is set as 0.5 and other entries have zeros, it represents a blank residue added to the N-terminal or the C-terminal to make a local sequence of 15-residue length. For all other residues, the 21-st entries are set to zero. In summery, a residue is coded by a 315-dimensional vector.
Support vector regression
To find the function between protein local sequence and normalized contact number, we use ∈-insensitive support vector regression (∈-SVR) [14,15]. The expected function can be formulated as
f(Xi) = 〈W, Φ(Xi)〉 + b, (4)
where W is the weight and b is the bias. Φ(Xi) is a non-linear function mapping a data point from the input space to the feature space, so consequently, SVR is able to perform non-linear regression. The goal of the regression is to find the optimal W and b using some optimisation criteria. In ε-SVR, errors greater than ε are penalized, where two positive variables ξ and ξ* are used to measure the deviation of samples outside the ε-insensitive tube. The optimisation problem can be expressed as
Minimize12‖W‖2+C∑i=1M(ξi+ξi*),
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeGabaa6imXvP5wqSXMqHnxAJn0BKvguHDwzZbqegyvzYrwyUfgaiqaacaWFnbGaa8xAaiaa=5gacaWFPbGaa8xBaiaa=LgacaWF6bGaa8xzaiaaxMaadaWcaaqaaiabigdaXaqaaiabikdaYaaadaqbdaqaaiabdEfaxbGaayzcSlaawQa7amaaCaaaleqabaGaeGOmaidaaOGaey4kaSIaem4qam0aaabCaeaacqGGOaakiiGacqGF+oaEdaWgaaWcbaGaemyAaKgabeaakiabgUcaRiab+57a4naaDaaaleaacqWGPbqAaeaacqGGQaGkaaGccqGGPaqkaSqaaiabdMgaPjabg2da9iabigdaXaqaaiabd2eanbqdcqGHris5aOGaeiilaWcaaa@5B39@
subject to{f(Xi)−yi≤ε+ξiyi−f(Xi)≤ε+ξi*ξi,ξi*≥0 for i=1,…,M (5)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeGabaa6imXvP5wqSXMqHnxAJn0BKvguHDwzZbqegyvzYrwyUfgaiqaacaWFZbGaa8xDaiaa=jgacaWFQbGaa8xzaiaa=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@8590@
where C is the regularization constant that determines the trad∈off between the norm and the error penalty.
The solution of the above problem was given by the authors of ∈SVR [14,15] as follows,
f(X)=∑i=1M(αi−αi*)〈Φ(Xi),Φ(X)〉+b, (6)
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGMbGzcqGGOaakcqWGybawcqGGPaqkcqGH9aqpdaaeWbqaaiabcIcaOGGaciab=f7aHnaaBaaaleaacqWGPbqAaeqaaOGaeyOeI0Iae8xSde2aa0baaSqaaiabdMgaPbqaaiabcQcaQaaakiabcMcaPmaaamaabaGaeuOPdyKaeiikaGIaemiwaG1aaSbaaSqaaiabdMgaPbqabaGccqGGPaqkcqGGSaalcqqHMoGrcqGGOaakcqWGybawcqGGPaqkaiaawMYicaGLQmcacqGHRaWkcqWGIbGycqGGSaalcaWLjaGaaCzcamaabmaabaGaeGOnaydacaGLOaGaayzkaaaaleaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGnbqta0GaeyyeIuoaaaa@5667@
where αi and αi*
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaaiiGacqWFXoqydaqhaaWcbaGaemyAaKgabaGaeiOkaOcaaaaa@30B6@ are Lagrange multipliers. We can replace 〈Φ(Xi), Φ(X)〉, the inner product of Φ(Xi) and Φ(X), by a kernel function K(Xi, X), if K(Xi, X) = 〈Φ(Xi), Φ(X)〉. The radial basis function are used in our study, as given by
K(Xi, X) = exp(-γ||Xi - X||2), (7)
where γ is a parameter to be tuned by the user.
We constantly set ε as 0.01, γ as 0.01 and C as 5.0, because this set of parameters yielded the best performance in our previous work [6,8]. A number of software packages can be used to find the solution such as SVMlight [16].
Dataset preparation and prediction evaluation
To test our approach, we selected 945 unique protein chains, which were previously used for prediction of protein ASA, and were prepared by PDB-REPRDB [17]. The structures solved by X-ray crystallography were with resolution less than 2.0 Å and with an R-factor less than 0.2. All chains are at least 60 amino acids or longer, and the pair-wise identity is less than 25%. The protein names can be found in the additional file 1 (supplementary material).
The proteins are randomly divided into three groups with each group having 315 chains. Each group is in turn used for training with the remaining two groups used for testing. Therefore, each group is tested twice by the two functions derived from the other groups, and as a result we have six groups of examination results.
Pearson's correlation coefficients and root mean square errors are calculated with respects to all residues and individual proteins. In addition, the absolute errors are calculated for the residues with different contact numbers. In order to compare with previous classification methods, we use different thresholds to classify contact numbers as "contacted" or "non-contacted" and compute the overall accuracy. The accuracy is defined as the ratio between the number of correctly predicted residues and the total number.
Supplementary Material
Additional File 1
The names of 945 protein chains. The first four characters are their PDB names. The fifth is the chain name and "_" means single chain.
Click here for file
Acknowledgements
The author thanks Dr. Rohan Teasdale for kind assistance and Dr. Michael Gagen for critical reading of this article and some helpful suggestions. The work was supported by funds from the Australian Research Council (ARC) and The University of Queensland Early Research Grant. The computer simulations were performed at the High Performance Computing Facility at The University of Queensland.
==== Refs
Karchin R Cline M Karplus K Evaluation of local structure alphabets based on residue burial Proteins 2004 55 508 518 15103615 10.1002/prot.20008
Hamelryck T An amino acid has two sides: A new 2D measure provides a different view of solvent exposure Proteins 2005 59 38 48 15688434 10.1002/prot.20379
Kinjo AR Horimoto K Nishikawa K Predicting absolute contact numbers of native protein structure from amino acid sequence Proteins 2005 58 158 165 15523668 10.1002/prot.20300
Pollastri G Baldi P Fariselli P Casadio R Prediction of coordination number and relative solvent accessibility in proteins Proteins 2002 47 142 153 11933061 10.1002/prot.10069
Pollastri G Baldi P Fariselli P Casadio R Improved prediction of the number of residue contacts in proteins by recurrent neural networks Bioinformatics 2001 17 Suppl 1 S234 42 11473014
Yuan Z Huang B Prediction of protein accessible surface areas by support vector regression Proteins 2004 57 558 564 15382233 10.1002/prot.20234
Yuan Z Bailey TL Prediction of protein solvent profile using SVR Proceedings of the 26th Annual International Conference of the IEEE EMBS 2004 2889 2892
Yuan Z Bailey TL Teasdale RD Prediction of protein B-factor profiles Proteins 2005 58 905 912 15645415 10.1002/prot.20375
Kabsch W Sander C Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features Biopolymers 1983 22 2577 2637 6667333 10.1002/bip.360221211
Thompson MJ Goldstein RA Predicting solvent accessibility: higher accuracy using Bayesian statistics and optimized residue substitution classes Proteins 1996 25 38 47 8727318 10.1002/(SICI)1097-0134(199605)25:1<38::AID-PROT4>3.3.CO;2-H
Rost B Sander C Conservation and prediction of solvent accessibility in protein families. Proteins 1994 20 216 226 7892171 10.1002/prot.340200303
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Jones DT Protein secondary structure prediction based on position-specific scoring matrices J Mol Biol 1999 292 195 202 10493868 10.1006/jmbi.1999.3091
Vapnik V The Nature of Statistical Learning Theory 2000 Springer, New York,
Smola A Scholkopf B A tutorial on support vector regression NeuroCOLT Technical Report Series, NC-TR-1998-030 1998
Joachims T Schölkopf B, Burges C and Smola A Making large-Scale SVM Learning Practical. Advances in Kernel Methods - Support Vector Learning 1999 , MIT Press
Noguchi T Akiyama Y PDB-REPRDB: a database of representative protein chains from the Protein Data Bank (PDB) in 2003 Nucleic Acids Res 2003 31 492 493 12520060 10.1093/nar/gkg022
|
16221309
|
PMC1277819
|
CC BY
|
2021-01-04 16:27:23
|
no
|
BMC Bioinformatics. 2005 Oct 13; 6:248
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-248
|
oa_comm
|
==== Front
BMC BiolBMC Biology1741-7007BioMed Central London 1741-7007-3-221623617810.1186/1741-7007-3-22Research ArticleThe complete chloroplast DNA sequences of the charophycean green algae Staurastrum and Zygnema reveal that the chloroplast genome underwent extensive changes during the evolution of the Zygnematales Turmel Monique [email protected] Christian [email protected] Claude [email protected] Département de Biochimie et de Microbiologie, Université Laval, Québec, Québec, G1K 7P4, Canada2005 20 10 2005 3 22 22 8 7 2005 20 10 2005 Copyright © 2005 Turmel 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 Streptophyta comprise all land plants and six monophyletic groups of charophycean green algae. Phylogenetic analyses of four genes from three cellular compartments support the following branching order for these algal lineages: Mesostigmatales, Chlorokybales, Klebsormidiales, Zygnematales, Coleochaetales and Charales, with the last lineage being sister to land plants. Comparative analyses of the Mesostigma viride (Mesostigmatales) and land plant chloroplast genome sequences revealed that this genome experienced many gene losses, intron insertions and gene rearrangements during the evolution of charophyceans. On the other hand, the chloroplast genome of Chaetosphaeridium globosum (Coleochaetales) is highly similar to its land plant counterparts in terms of gene content, intron composition and gene order, indicating that most of the features characteristic of land plant chloroplast DNA (cpDNA) were acquired from charophycean green algae. To gain further insight into when the highly conservative pattern displayed by land plant cpDNAs originated in the Streptophyta, we have determined the cpDNA sequences of the distantly related zygnematalean algae Staurastrum punctulatum and Zygnema circumcarinatum.
Results
The 157,089 bp Staurastrum and 165,372 bp Zygnema cpDNAs encode 121 and 125 genes, respectively. Although both cpDNAs lack an rRNA-encoding inverted repeat (IR), they are substantially larger than Chaetosphaeridium and land plant cpDNAs. This increased size is explained by the expansion of intergenic spacers and introns. The Staurastrum and Zygnema genomes differ extensively from one another and from their streptophyte counterparts at the level of gene order, with the Staurastrum genome more closely resembling its land plant counterparts than does Zygnema cpDNA. Many intergenic regions in Zygnema cpDNA harbor tandem repeats. The introns in both Staurastrum (8 introns) and Zygnema (13 introns) cpDNAs represent subsets of those found in land plant cpDNAs. They represent 16 distinct insertion sites, only five of which are shared by the two zygnematalean genomes. Three of these insertions sites have not been identified in Chaetosphaeridium cpDNA.
Conclusion
The chloroplast genome experienced substantial changes in overall structure, gene order, and intron content during the evolution of the Zygnematales. Most of the features considered earlier as typical of land plant cpDNAs probably originated before the emergence of the Zygnematales and Coleochaetales.
==== Body
Background
About 450 million years ago, green algae belonging to the class Charophyceae emerged from their aquatic habitat to colonize the land [1-3]. This important event in the history of life gave rise to all the land plant species that make up the flora of our planet. The few thousand species of charophycean green algae that are alive today exhibit great variability in cellular organization and reproduction [4]. With the land plants, they form the green plant lineage Streptophyta [5], whereas all other green algae (more than 10,000 species), with perhaps the exception of Mesostigma viride, belong to the sister lineage Chlorophyta [4]. Five monophyletic groups of charophycean green algae have been recognized: the Chlorokybales, Klebsormidiales, Zygnematales, Coleochaetales and Charales [6], given here in order of increasing cellular complexity. Mesostigma may represent an additional lineage of the Charophyceae, the Mesostigmatales, as indicated by phylogenetic studies that placed this unicellular green alga at the base of the Streptophyta [7-10]. This lineage, however, remains controversial, considering that separate analyses based on a large number of chloroplast- or mitochondrial-encoded proteins [11-13] and on the chloroplast small and large subunit rRNA genes [14] identified Mesostigma before the divergence of the Chlorophyta and Streptophyta.
On the basis of morphological characters alone, the two charophycean groups that exhibit the greatest cellular complexity, i.e. the Charales and Coleochaetales, have been proposed to be the closest relatives of land plants [15,16]. Recent analyses of the combined sequences of four genes from the nucleus (small subunit rRNA gene), chloroplast (atpB and rbcL) and mitochondria (nad5) of 25 charophycean green algae and eight green plants revealed that the Charales and land plants form a highly supported clade; however, moderate bootstrap support was observed for the positions of the other charophycean groups [8]. The best trees inferred by Bayesian and maximum likelihood methods in this four-gene analysis support an evolutionary trend toward increasing cellular complexity [17]. In contrast, all phylogenies of charophycean green algae previously inferred from a smaller number of genes failed to provide any conclusive results concerning the branching order of the charophycean green algae and their relationships with land plants [15,16].
We have recently undertaken the sequencing of complete chloroplast genomes from representatives of the various charophycean lineages in order to elucidate the branching order of these lineages and also to understand the evolution of chloroplast DNA (cpDNA) within the Streptophyta. We have reported thus far the cpDNA sequences of Mesostigma (Mesostigmatales) [11] and Chaetosphaeridium globosum (Coleochaetales) [18]. Comparative analyses of the Mesostigma cpDNA sequence (136 genes, no introns) with its land plant counterparts (110–120 genes, about 20 introns) revealed that the chloroplast genome underwent substantial changes in its architecture during the evolution of streptophytes (namely gene losses, intron insertions and scrambling of gene order). At the levels of gene content (125 genes), intron composition (18 introns) and gene order, Chaetosphaeridium cpDNA is remarkably similar to land plant cpDNAs, implying that most of the features characteristic of land plant lineages were acquired from charophycean green algae. Like the cpDNAs of many chlorophytes, those of Mesostigma, Chaetosphaeridium and most land plant species exhibit a quadripartite structure that is characterized by the presence of two copies of a rDNA-containing inverted repeat (IR) separated by large and small single-copy regions. All the genes they have in common, with a few exceptions, reside in corresponding genomic regions.
In this study, we report the complete cpDNA sequences of two members of the Zygnematales that belong to distinct lineages, Staurastrum punctulatum and Zygnema circumcarinatum. Although the chloroplast genomes of these charophycean green algae closely resemble their Chaetosphaeridium and bryophyte counterparts at the primary sequence and gene content levels, they feature substantial differences at the levels of structure, gene order and intron content. Like the cpDNA of the zygnematalean alga Spirogyra maxima [19], both Staurastrum and Zygnema cpDNAs lack a large IR. Clearly, loss of the IR appears to be a major event that shaped the architecture of the chloroplast genome in the Zygnematales, an event that apparently occurred early during the evolution of this group of charophycean green algae.
Results
Selection of taxa
The Zygnematales as circumscribed by Bold and Wynne [20] comprise the green algae whose mode of sexual reproduction is conjugation. This is the most important charophycean lineage in terms of diversity and number of species (~50 genera and ~6,000 species) [16]. Classification schemes based on cell wall organization have recognized two groups of conjugating green algae: first, the unicellular or multicellular green algae with an ornamented and segmented cell wall, also called placoderm desmids and often treated as members of the order Desmidiales, and second, the green algae that bear a smooth cell wall, which are often classified separately in the Zygnematales [21]. Among the latter group are found filamentous forms and the saccoderm desmids that consist either of unicells or loosely joined cells. Phylogenies inferred using rbcL [21] or the combined rbcL and nuclear small subunit rRNA genes [22] support the monophyly of placoderm desmids and place the filamentous and saccoderm desmids together in a distinct monophyletic group. For our study, we have selected a representative of each of these two monophyletic groups: Staurastrum is a unicellular, placoderm desmid, whereas Zygnema is a filamentous green alga with a non-ornamented cell wall.
General features
The 157,089-bp Staurastrum [GenBank:AY958085] and 165,372-bp Zygnema [GenBank:AY958086] cpDNAs map as circular molecules containing 121 and 125 genes, respectively (Fig. 1). Both genomes lack a rDNA-containing IR and no remnant of such a sequence could be detected during our analysis of repeated elements. All genes are present in single copy, with the exception of the duplicated Zygnema trnE(uuc) gene, the sequences of which differ at two positions. Note that the matK gene was not included in the total number of genes calculated for Zygnema cpDNA, because this gene occurs as an intron ORF in all other streptophytes where it has been identified. Aside from the absence of the IR, the most prominent differences displayed by the two zygnematalean cpDNAs relative to their counterparts in Chaetosphaeridium [18] and land plants (here represented by the bryophyte Marchantia polymorpha [23]) are their larger size (taking into consideration the absence of the IR from these genomes) and their smaller number of cis-spliced group II introns (Table 1). The larger size of zygnematalean cpDNAs is mainly explained by the expansion of intergenic spacers (Table 2). The latter sequences represent 42% of the genome in both Staurastrum and Zygnema cpDNAs compared to about 20% in Chaetosphaeridium and land plant cpDNAs. Introns have also expanded in size in both zygnematalean cpDNAs compared to their Chaetosphaeridium and land plant homologues (Table 2).
Gene content
Table 3 compares the gene contents of Staurastrum, Zygnema, Chaetosphaeridium and Marchantia cpDNAs. The two zygnematalean cpDNAs share 120 genes, 116 of which are present in both Chaetosphaeridium and Marchantia cpDNAs. Five genes in Zygnema cpDNA are missing from Staurastrum cpDNA; they encode the tRNAPro(GGG), tRNASer(CGA), ribosomal protein L5, and the proteins CysA and CysT that are involved in sulfate transport. Although there is no functional trnS(cga) in Staurastrum cpDNA, a trnS(cga) pseudogene was identified in this genome. A standard acceptor stem could not be modelled from the RNA sequence derived from this pseudogene; the 5' region of this sequence diverges considerably from homologous tRNA sequences in other streptophytes and cannot base pair with the 3' region. Staurastrum exhibits only one chloroplast gene (rpl22) that is missing from Zygnema. To our knowledge, this is the first time that the loss of rpl22 together with that of rpl32 (a gene absent from both zygnematalean cpDNAs) has been reported in the Streptophyta. As in land plant cpDNAs, but in contrast to Chaetosphaeridium cpDNA, no tufA-like sequence was detected in the two zygnematalean cpDNAs. It appears that only the chlI, odpB and ycf62 genes were specifically lost just before or concurrently with the emergence of land plants (Table 3). Note that the rps16 gene cannot be included in this category, as it is present in the majority of land plant cpDNAs sequenced to date.
Gene order
Staurastrum and Zygnema cpDNAs differ substantially from one another and from their Chaetosphaeridium and land plant counterparts at the level of gene organization (Table 4). Eighty-two genes in the two zygnematalean cpDNAs form 22 blocks of colinear sequences, which are highly scrambled in order (Fig. 1). A minimum of 59 inversions would be required to convert the gene order of Staurastrum cpDNA into that of Zygnema cpDNA (Table 4).
Of the two zygnematalean cpDNAs, that showing the most similar gene arrangement with its Chaetosphaeridium and land plant counterparts is Staurastrum cpDNA (Table 4). In both Staurastrum and Zygnema cpDNAs, the gene organization more closely resembles that of Marchantia than that of Chaetosphaeridium (Table 4). Staurastrum cpDNA shares with its Marchantia counterpart 22 blocks of colinear sequences that contain a total of 101 genes, whereas Zygnema cpDNA shares 20 blocks featuring 81 genes (Fig. 1). Close inspection of these blocks relative to those conserved between Mesostigma and Marchantia cpDNAs [11] reveals that 13 ancestral gene clusters, including those containing the rDNA, atpA, psbB and rpoB operons, were fragmented at 27 sites during the evolution of the Zygnematales (Fig. 2). Eleven of these rearrangement breakpoints are common to the two green algal cpDNAs, whereas 2 and 14 breakpoints are unique to Staurastrum and Zygnema cpDNAs, respectively. Assuming that these unique rearrangement breakpoints appeared after the divergence of the two zygnematalean species, we infer that the chloroplast genome of the common ancestor of Staurastrum and Zygnema shared a number of derived gene clusters with Chaetosphaeridium and land plants. For example, the cluster of 29 genes extending from petL to trnI(cau) in Marchantia cpDNA and that of 13 genes delimited by rps12b and atpI were likely present in the common ancestor of Staurastrum and Zygnema. Only four gene clusters are shared specifically between zygnematalean and Marchantia cpDNAs: rps4-trnS(gga)-ycf3 (cluster 9 in Fig. 1), atpB-atpE-trnV(uac)-trnMe(cau)-ndhC-ndhK-ndhJ (cluster 15), trnH(gug)-ftsH-trnD(guc) (in Staurastrum only), and trnE(uuc)-cysA-trnT(ggu) (in Zygnema only).
The higher degree of ancestral characters displayed by Staurastrum cpDNA compared to its Zygnema homologue at the gene organizational level is also evident when one examines the genomic region in which each gene locus would be expected to map if the IR had been retained (Fig. 3). In Staurastrum cpDNA, the 15 genes predicted to have been present in the small single-copy region occupy a discrete region just beside five of the eight genes that usually make up the IR; in Zygnema cpDNA, however, the genes usually located in the small single-copy region and the IR are more widely dispersed in the genome.
Intron composition
As in Chaetosphaeridium cpDNA, the introns in Staurastrum and Zygnema cpDNAs represent subsets of those found in land plant cpDNAs (Fig. 4). Both zygnematalean cpDNAs share with their Chaetosphaeridium and land plant counterparts one group I intron in trnL(uaa), two cis-spliced group II introns in rpl16 and trnG(ucc), and one trans-spliced group II intron in rps12. Only three group II introns in Staurastrum and/or Zygnema cpDNAs (in atpF, rps12 at site 346 and ycf3) have no homologues in Chaetosphaeridium cpDNA. Evidence for a charophycean green algal origin of land plant group II introns is lacking for only the clpP intron at site 363. The Staurastrum trans-spliced rps12 intron resembles its Chaetosphaeridium homologue in exhibiting a large ORF in domain IV. The putative protein of 404 amino acids encoded by the Staurastrum ORF is related to reverse transcriptases, whereas the smaller protein (247 amino acids) specified by the Chaetosphaeridium ORF lacks similarity with such proteins.
Like its Chaetosphaeridium and land plant counterparts, the cis-spliced group II intron in Staurastrum trnK(uuu) encodes the maturase MatK. As mentioned earlier, a freestanding matK gene was identified in Zygnema cpDNA even though an intron is absent from trnK(uuu) in this charophycean green alga. Close inspection of the regions immediately flanking the Zygnema matK gene for the presence of sequences conserved in domains V and VI of group II introns failed to reveal any evidence that this gene had once been an integral part of a group II intron. The Zygnema matK is most probably a functional gene because its predicted protein features the vast majority of the conserved amino acids that the trnK intron-encoded MatK of Staurastrum shares with its Chaetosphaeridium, Chara, Nitella and land plant homologues (Fig. 5).
Repeated sequences
Comparison of each zygnematalean cpDNA sequence against itself using PipMaker [24] indicated the presence of repeats in many intergenic regions of Zygnema cpDNA and the virtual absence of such sequences from Staurastrum cpDNA. Analysis of the Zygnema genome sequence with REPuter [25] revealed that the great majority of the repeat regions larger than 30 bp are composed of short tandem repeats. Each of the 35 repeat regions identified consists of 4 to 16 bp units that are repeated in tandem 4 to 50 times (Table 5). Most regions (29/35) feature repeat units of 4 or 5 bp, and the regions with GTAT, ATAC, TAGAA, TTCTA and CTTA units occur at more than one location on the chloroplast genome (Fig. 1). All three regions carrying the CTTA units feature sequences that are in direct orientation relative to one another; however, the 13 regions with the GTAT and complementary ATAC units and the four regions with the TAGAA and complementary TTCTA units form a population of dispersed repeats that are in direct or inverted orientation relative to one another. Eighty percent of the repeat regions (28/35) reside outside the blocks of sequences that are colinear with Staurastrum cpDNA. We estimate that at least 2,245 bp of Zygnema cpDNA, i.e. about 60% of the increased size of the Zygnema intergenic regions compared to their Staurastrum homologues, are accounted for by short tandem repeats.
Only two loci of the Staurastrum chloroplast genome contain short tandem repeats: a region composed of four units of the GAATAAATA sequence in the infA-rpl36 spacer and a region containing nine units of the GTATTT sequence in the rps16-odpB spacer. Aside from two copies of 45-bp sequence (in the atpF-atpH and atpH-rps14 spacers) that are in direct orientation, no dispersed repeats larger than 30 bp were detected in Staurastrum cpDNA.
Discussion
Although Staurastrum and Zygnema cpDNAs bear high similarity in primary sequence and gene content to their Chaetosphaeridium and land plant counterparts, they differ substantially from one another and from the latter genomes in overall structure, gene order and intron content. From our comparative analysis of streptophyte cpDNAs, we infer that the chloroplast genome of the last common ancestor of Staurastrum and Zygnema probably lacked a large IR encoding the rRNA genes, had a low gene density, and more closely resembled Chaetosphaeridium and land plant cpDNAs at the gene organizational and intron levels than do Zygnema and Staurastrum cpDNAs. At least 16 of the 22 intron positions commonly found in land plant cpDNAs, including three sites that have not been identified in Chaetosphaeridium, were probably present in the common ancestor of Staurastrum and Zygnema.
Considering the absence of an rDNA-encoding IR region in both Staurastrum and Zygnema cpDNAs, it is not surprising that these genomes are considerably rearranged relative to their coleochaetalean and land plants counterparts that have retained the quadripartite structure. All green plant cpDNAs that have lost the IR tend to be highly scrambled in gene order [26,27]. It has been hypothesized that the loss of the IR enhances opportunities for intramolecular recombination between small dispersed repeats [28]. In agreement with the idea that there is a direct link between the frequency of intramolecular recombination events and the abundance of small dispersed repeats [28], we identified more rearrangements in the repeat-rich genome of Zygnema than in the repeat-poor genome of Staurastrum. As in the cpDNAs of the nonphotosynthetic, parasitic flowering plant Epifagus virginiana [29] and the evening primrose Oenothera [30], the repeated sequences in Zygnema cpDNA consist essentially of tandem repeats that probably arose by replication slippage.
A single event of IR loss likely accounts for the absence of a quadripartite structure from both Staurastrum and Zygnema cpDNAs. This hypothesis is more parsimonious than the alternative scenario involving two independent losses, and is consistent with previous evidence that the cpDNA of Spirogyra (a distant relative of Zygnema) has no IR [19]. It is also supported by our finding that Staurastrum and Zygnema cpDNAs share 11 rearrangement breakpoints within ancestral gene clusters. Given the close connection between IR loss and gene rearrangements, several of these shared breakpoints might have appeared following the loss of the IR in the lineage leading to the last common ancestor of Staurastrum and Zygnema. Considering that this ancestor occupies a basal position in the tree describing the relationships among zygnematalean green algae [21,22], then most, if not all, of the algae belonging to the Zygnematales are expected to lack an IR in their chloroplast genome.
As introns appear to be generally stable in land plant cpDNAs [28], the important difference in intron content displayed by Staurastrum and Zygnema cpDNAs is unexpected. The two zygnematalean cpDNAs share only five of the 16 intron insertion sites they exhibit in total.Staurastrum cpDNA lacks seven of the 13 introns that are present in Zygnema cpDNA, whereas the latter cpDNA lacks five of the eight introns found in the former genome. The intron distributions in these cpDNAs are best explained by assuming that all 16 insertion sites were populated with introns in the common ancestor of Staurastrum and Zygnema and that subsequently, several introns were specifically lost in each of the lineages leading to these green algae. Obviously, we cannot exclude the possibility that chloroplast introns occupying common insertion sites were lost independently in the Staurastrum and Zygnema lineages; thus, the predicted number of introns in the common ancestor of these algae may represent a minimal estimate. Given that intron losses are thought to result from insertions, through homologous recombination, of intron-less cDNA copies generated by reverse transcription [31], the frequency of homologous recombination events or the level of reverse transcriptase activity might be higher in the chloroplasts of conjugating green algae than in land plant chloroplasts. In this respect, it is interesting to note that the Staurastrum trans-spliced rps12 intron specifies a reverse transcriptase and is the only known streptophyte chloroplast intron encoding such an activity.
Our finding that matK is free-standing in Zygnema cpDNA together with the absence of the trnK(uuu) intron in which it usually resides strongly suggests that its putative maturase product is essential for the splicing of group II introns other than the trnK(uuu) intron. Circumstantial evidence that MatK functions in splicing of multiple introns has previously been reported for land plant chloroplasts. The matK gene is located within the group II intron of trnK(uuu) in all photosynthetic land plants, but occurs as a free-standing gene in Epifagus cpDNA [29]. In vivo splicing analyses of the complete set of chloroplast group II introns in land plant mutants lacking chloroplast ribosomes disclosed specific splicing defects involving mainly group IIA introns (in atpF, rpl2, rps12, trnA, trnI, trnK), thus implying that cpDNA-encoded protein(s) act as splicing factors [32-35]. It has been proposed that MatK evolved from a trnK(uuu) intron-specific maturase to a more versatile maturase that assists the splicing of most or all group IIA introns of land plants [32-35].
Conclusion
Our structural analyses of the Staurastrum and Zygnema chloroplast genomes have revealed that many of the features considered earlier as typical of land plant cpDNAs originated before the emergence of the Coleochaetales and Zygnematales. While the chloroplast genome appears to have remained relatively stable in the coleochaetalean lineage, it has lost the IR and has undergone many changes in gene order and intron content during the evolution of the Zygnematales.
Methods
DNA isolation and cloning
Chloroplast DNA fractions from Staurastrum punctulatum de Brébisson (SAG 679-1) and Zygnema circumcarinatum Czurda (SAG 698-1a) were obtained by isopycnic centrifugation of total cellular DNAs in CsCl-bisbenzimide gradients [36]. A random clone library was prepared from each algal cpDNA fraction as follows. DNA was sheared by nebulization and 1,500–2,000-bp fragments were recovered by electroelution after agarose gel electrophoresis. These fragments were treated with E. coli Klenow fragment and T7 DNA polymerase and cloned into the SmaI site of Bluescript II KS+ or into ligation-ready pSMART-HCKan (Lucigen Corporation, Middleton). After filter hybridization of the clones with the original DNA used for cloning as a probe, DNA templates from positive clones were prepared with the QIAprep 96 Miniprep kit (Qiagen Inc., Canada).
Sequence analyses
Nucleotide sequences were determined with the PRISM BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA), the PRISM dGTP BigDye terminator ready reaction kit (Applied Biosystems), and the DYEnamic ET terminator cycle sequencing kit (Amersham Pharmacia Biotech, Canada) on ABI model 373 or 377 DNA sequencers (Applied Biosystems), using T3 and T7 primers as well as oligonucleotides complementary to internal regions of the plasmid DNA inserts. Genomic regions not represented in the clones analyzed were sequenced from PCR-amplified fragments. Sequences were assembled using SEQUENCHER 4.1.1 (Gene Codes Corporation, Ann Arbor, MI) and analyzed using the Genetics Computer Group (Madison, WI) software (version 10.3) package. Protein-coding and rRNA genes were identified by BLAST searches [37] of the nonredundant database at the National Center for Biotechnology Information, and tRNA genes were found using tRNAscan-SE [38]. Repeated sequence elements were searched using REPuter [25]. The GRIMM web server [39] was used to infer the number of gene permutations by inversions. Genes within copy A of the Chaetosphaeridium and Marchantia IRs were excluded in these gene order analyses. Pairwise comparisons of genome sequences were carried out using PipMaker [24].
Abbreviations
cpDNA, chloroplast DNA; IR, inverted repeat; ORF, open reading frame; rRNA, ribosomal RNA.
Authors' contributions
MT conceived and designed the study, contributed to the analysis and interpretation of the data and wrote the manuscript. CO carried out the sequencing of the Staurastrum and Zygnema chloroplast genomes. CO and CL participated in the assembly of the genome sequences. CL performed all sequence analyses and generated the figures. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to Jonathan Gagnon and Mélanie Bourassa for their assistance in determining the Zygnema cpDNA sequence, to Jean-François Rochette for his assistance in sequencing Staurastrum cpDNA, and to Jules Gagnon and Patrick Charlebois for their help with the bioinformatics analyses. This work was supported by the Natural Sciences and Engineering Research Council of Canada.
Figures and Tables
Figure 1 Gene maps of Staurastrum and Zygnema cpDNAs. Genes (filled boxes) shown on the outside of each map are transcribed in a clockwise direction, whereas those on the inside of each map are transcribed counterclockwise. Genes absent from Marchantia cpDNA are represented in beige. Gene clusters shared with Marchantia cpDNA [GenBank:NC_001319] are shown as alternating series of green and red boxes. Genes present in Marchantia cpDNA but located outside conserved clusters are shown in grey. Gene clusters shared by the two zygnematalean cpDNAs are represented by labelled bars outside each map. Genes containing introns (open boxes) are denoted by asterisks. Dispersed repeat regions in Zygnema cpDNA that contain short tandem repeats are denoted by symbols. The repeat units in these regions are as follows: filled squares, TAGAA; open squares, TTCTA; filled circles, GTAT; open circles, ATAC; filled triangles, CTTA. Note that filled and open symbols with the same geometric shape represent the repeat regions of which the sequences are in inverted orientation relative to one another. The intron sequences bordering the rps12 exons (rps12a and rps12b) are spliced in trans at the RNA level. tRNA genes are indicated by the one-letter amino acid code (Me, elongator methionine; Mf, initiator methionine) followed by the anticodon in parentheses. The ORFs unique to Staurastrum or Zygnema cpDNA are not indicated (see [GenBank:AY958085] and [GenBank:AY958086] for more details).
Figure 2 Fragmentation of ancestral chloroplast gene clusters during the evolution of the Zygnematales. The ancestral clusters shown are found in both Mesostigma [GenBank:NC_002186] and Marchantia [GenBank:NC_001319] cpDNAs. The top and bottom arrows denote the sites where they are broken in Staurastrum and Zygnema cpDNAs, respectively. For the polarities of the genes relative to one another, the reader should consult the gene map of Mesostigma cpDNA [11].
Figure 3 Compared patterns of gene partitioning in zygnematalean and Marchantia cpDNAs. Each gene in Staurastrum and Zygnema cpDNAs is colour-coded according to the region of Marchantia cpDNA [GenBank:NC_001319] carrying its homologue; cyan, large single-copy region; magenta, small single-copy region; and yellow, IR. Genes shown in grey are absent from Marchantia cpDNA.
Figure 4 Distributions of introns in streptophyte cpDNAs. Circles denote the presence of group I introns, and squares denote the presence of group II introns. Divided squares represent trans-spliced group II introns. Open symbols denote the absence of intron ORFs, whereas filled symbols denote their presence. Intron insertion sites in protein-coding and tRNA genes are given relative to the corresponding genes in Mesostigma cpDNA; the insertion site in rrl is given relative to the Escherichia coli 23S rRNA. For each insertion site, the position corresponding to the nucleotide immediately preceding the intron is reported. Note that rps16 is lacking in Marchantia cpDNA and that the rrl intron at position 2593 is absent from all completely sequenced land plant cpDNAs, with the exception of Anthoceros cpDNA. The intron data were taken from the following accession numbers: Staurastrum, [GenBank:AY958085]; Zygnema, [GenBank:AY958086]; Chaetosphaeridium, [GenBank:NC_004115]; Marchantia, [GenBank:NC_001319]; and Anthoceros formosae [GenBank:NC_004543].
Figure 5 Sequence conservation among streptophyte MatK proteins. The MatK sequences of selected green algae and land plants were aligned with T-COFFEE [40] and arranged into two separate groups. Identical amino acids in all the sequences examined are displayed on a black background, whereas identical amino acids in all the green algal or land plant sequences are shown on a dark grey background. In each group, sets of residues sharing eight of the 10 features in the property matrix of AMAS [41] are shown on a light grey background. The accession numbers for the MatK sequences analyzed are as follows: Zygnema, [GenBank:AY958086]; Staurastrum, [GenBank:AY958085]; Chaetosphaeridium, [GenBank:NC_004115]; Chara connivens, [GenBank:AY170442]; Nitella opaca, [GenBank:AY170449]; Marchantia, [GenBank:NC_001319]; and Physcomitrella patens [GenBank:NC_005087].
Table 1 General features of cpDNAs from Staurastrum, Zygnema, other streptophytes and Mesostigma
Feature
Mesostigma
Staurastrum
Zygnema
Chaetosphaeridium
Marchantia
Sizea (bp)
IR 6,057 - - 12,431 10,058
SSC 22,619 - - 17,639 19,813
LSC 83,627 - - 88,682 81,095
Genome 118,360 157,089 165,372 131,183 121,024
A+T content (%) 69.9 67.5 68.9 70.4 71.2
Gene contentb 136 121 125 125 120
Introns
Group I 0 1 1 1 1
Group II
Cis-spliced 0 6 11 16 18
Trans-spliced 0 1 1 1 1
a Because Staurastrum and Zygnema cpDNAs lack an IR, only the genome size is given for each of these cpDNAs. SSC, small single-copy region; LSC, large single-copy region.
b Unique ORFs, intron ORFs and pseudogenes were not taken into account. Note that Chaetosphaeridium tufA was considered to be a functional gene.
Table 2 Proportion and base composition of coding sequences, intergenic spacers and introns in Staurastrum, Zygnema, Chaetosphaeridium and Marchantia cpDNAs
Sequences
Staurastrum
Zygnema
Chaetosphaeridium
Marchantia
Coding sequencesa
Fraction of genome (%) 51.4 50.8 67.5 69.9
A+T content (%) 65.1 63.0 66.1 67.7
Intergenic spacers
Fraction of genome (%) 42.0 42.2 23.1 19.3
A+T content (%) 70.0 75.7 79.8 80.6
Average size (bp) 536 546 223 178
Introns
Fraction of genome (%) 6.6 7.0 9.4 10.7
A+T content (%) 70.8 71.1 77.6 76.8
Average size (bp) 1,298 892 686 650
a Unique ORFs and intron ORFs were not considered to be coding sequences.
Table 3 Differences between the gene repertoires of Staurastrum, Zygnema, Chaetosphaeridium and Marchantia cpDNAs
Genea
Staurastrum
Zygnema
Chaetosphaeridium
Marchantia
chlI + + + -
cysA - + - +
cysT - + - +
odpB + + + -
rpl5 - + + -
rpl22 + - + +
rpl32 - - + +
rps16 + + + -
tufA - - +c -
ycf62 + + + -
trnP(ggg) - + + -b
trnS(cga) -b + - -
a Only the conserved genes that are missing in one or more chloroplast genomes are indicated. Plus and minus signs denote the presence and absence of genes, respectively.
b Pseudogenes.
c Chaetosphaeridium tufA could be a pseudogene because its sequence is highly divergent from those of other green plants.
Table 4 Number of inversions accounting for the gene rearrangements between Staurastrum, Zygnema, Chaetosphaeridium and Marchantia cpDNAs
Compared cpDNA Number of inversions
Staurastrum Zygnema Chaetosphaeridium Marchantia
Staurastrum - 59 45 35
Zygnema - - 59 54
Chaetosphaeridium - - - 13
Marchantia - - - -
Table 5 Zygnema cpDNA regions containing tandem repeats
Repeat regiona Repeat unit Number of unitsb
3276 – 3360 CTTAA 17
11203 – 11242 GTAT 10
11535 – 11602 GTAT 17
14272 – 14319 CTTA 12
15765 – 15807 AGAAAG 7
17944 – 18047 GTAT 26
18110 – 18179 TAGAA 14
18184 – 18263 CTTTT 16
24490 – 24565 ATAC 19
30556 – 30597 AAGTAC 7
32429 – 32533 GTAAA 21
34907 – 35018 ATAC 28
49994 – 50038 TAGAA 9
51521 – 51580 GTAT 15
51618 – 51817 CAAA 50
55388 – 55442 CTTTA 11
59550 – 59613 TGTGTTTGTATATTTA 4
60129 – 60183 TTCTA 11
68724 – 68763 TTCT 10
73516 – 73571 CTTA 14
73876 – 73915 ATAC 10
73919 – 73954 GTAT 9
88870 – 88925 ATAC 13
90538 – 90581 GAAT 11
92651 – 92730 TATATTACAT 8
102484 – 102531 TTTTAAAT 6
103132 – 103183 AATT 13
103629 – 103676 ATAC 12
104932 – 105090 GTAT 33
106702 – 106737 GTAT 9
132449 – 132496 GTAT 12
134893 – 134932 CTTA 10
140237 – 140308 TTACAATAGATT 6
143451 – 143485 TAATA 7
161662 – 161696 TTCTA 7
a Only the repeat regions larger than 30 bp are indicated; their coordinates refer to [GenBank:AY958086].
b The number of units was estimated by allowing one substitution per repeat unit.
==== Refs
Graham LE Cook ME Busse JS The origin of plants: body plan changes contributing to a major evolutionary radiation Proc Natl Acad Sci USA 2000 97 4535 4540 10781058 10.1073/pnas.97.9.4535
Kenrick P Crane PR The origin and early evolution of plants on land Nature 1997 389 33 39 10.1038/37918
Sanderson MJ Thorne JL Wikstrom N Bremer K Molecular evidence on plant divergence times Am J Bot 2004 91 1656 1665
Lewis LA McCourt RM Green algae and the origin of land plants Am J Bot 2004 91 1535 1556
Bremer K Humphries CJ Mishler BD Churchill SP On cladistic relationships in green plants Taxon 1987 36 339 349
Mattox KR Stewart KD Irvine DEG, John DM Classification of the green algae: a concept based on comparative cytology The Systematics of the Green Algae 1984 London: Academic Press 29 72
Bhattacharya D Weber K An SS Berning-Koch W Actin phylogeny identifies Mesostigma viride as a flagellate ancestor of the land plants J Mol Evol 1998 47 544 550 9797404
Karol KG McCourt RM Cimino MT Delwiche CF The closest living relatives of land plants Science 2001 294 2351 2353 11743201 10.1126/science.1065156
Marin B Melkonian M Mesostigmatophyceae, a new class of streptophyte green algae revealed by SSU rRNA sequence comparisons Protist 1999 150 399 417 10714774
Martin W Rujan T Richly E Hansen A Cornelsen S Lins T Leister D Stoebe B Hasegawa M Penny D Evolutionary analysis of Arabidopsis, cyanobacterial, and chloroplast genomes reveals plastid phylogeny and thousands of cyanobacterial genes in the nucleus Proc Natl Acad Sci USA 2002 99 12246 12251 12218172 10.1073/pnas.182432999
Lemieux C Otis C Turmel M Ancestral chloroplast genome in Mesostigma viride reveals an early branch of green plant evolution Nature 2000 403 649 652 10688199 10.1038/35001059
Turmel M Otis C Lemieux C The complete mitochondrial DNA sequence of Mesostigma viride identifies this green alga as the earliest green plant divergence and predicts a highly compact mitochondrial genome in the ancestor of all green plants Mol Biol Evol 2002 19 24 38 11752187
Martin W Deusch O Stawski N Grunheit N Goremykin V Chloroplast genome phylogenetics: why we need independent approaches to plant molecular evolution Trends Plant Sci 2005 10 203 209 15882651 10.1016/j.tplants.2005.03.007
Turmel M Ehara M Otis C Lemieux C Phylogenetic relationships among streptophytes as inferred from chloroplast small and large subunit rRNA gene sequences J Phycol 2002 38 364 375 10.1046/j.1529-8817.2002.01163.x
Chapman RL Waters DA Green algae and land plants – an answer at last? J Phycol 2002 38 237 240 10.1046/j.1529-8817.2002.03821.x
Qiu YL Palmer JD Phylogeny of early land plants: insights from genes and genomes Trends Plant Sci 1999 4 26 30 10234267 10.1016/S1360-1385(98)01361-2
McCourt RM Delwiche CF Karol KG Charophyte algae and land plant origins Trends Ecol Evol 2004 19 661 666 10.1016/j.tree.2004.09.013
Turmel M Otis C Lemieux C The chloroplast and mitochondrial genome sequences of the charophyte Chaetosphaeridium globosum : insights into the timing of the events that restructured organelle DNAs within the green algal lineage that led to land plants Proc Natl Acad Sci USA 2002 99 11275 11280 12161560 10.1073/pnas.162203299
Manhart JR Hoshaw RW Palmer JD Unique chloroplast genome in Spirogyra maxima (Chlorophyta) revealed by physical and gene mapping J Phycol 1990 26 490 494 10.1111/j.0022-3646.1990.00490.x
Bold H Wynne MJ Introduction to the Algae 1985 2 Englewood Cliffs, New Jersey: Prentice-Hall, Inc
McCourt RM Karol KG Bell J Helm-Bychowski KM Grajewka A Wojciechowski MF Hoshaw R Phylogeny of the conjugating green algae (Zygnemophyceae) based on rbcL sequences J Phycol 2000 36 747 758 10.1046/j.1529-8817.2000.99106.x
Gontcharov AA Marin B Melkonian M Are combined analyses better than single gene phylogenies? A case study using SSU rDNA and rbcL sequence comparisons in the Zygnematophyceae (Streptophyta) Mol Biol Evol 2004 21 612 624 14739253 10.1093/molbev/msh052
Ohyama K Fukuzawa H Kohchi T Shirai H Sano T Sano S Umesono K Shiki Y Takeuchi M Chang Z Chloroplast gene organization deduced from complete sequence of liverwort Marchantia polymorpha chloroplast DNA Nature 1986 322 572 574 10.1038/322572a0
Schwartz S Zhang Z Frazer K Smit A Riemer C Bouck J Gibbs R Hardison R Miller W PipMaker: a web server for aligning two genomic DNA sequences Genome Res 2000 10 577 586 10779500 10.1101/gr.10.4.577
Kurtz S Choudhuri JV Ohlebusch E Schleiermacher C Stoye J Giegerich R REPuter: the manifold applications of repeat analysis on a genomic scale Nucleic Acids Res 2001 29 4633 4642 11713313 10.1093/nar/29.22.4633
Palmer JD Osorio B Aldrich J Thompson WF Chloroplast DNA evolution among legumes: Loss of a large inverted repeat occurred prior to other sequence rearrangements Curr Genet 1987 11 275 286 10.1007/BF00355401
Strauss SH Palmer JD Howe GT Doerksen AH Chloroplast genomes of two conifers lack a large inverted repeat and are extensively rearranged Proc Natl Acad Sci USA 1988 85 3898 3902 2836862
Palmer JD Bogorad L, Vasil K Plastid chromosomes: structure and evolution The Molecular Biology of Plastids 1991 San Diego: Academic Press 5 53
Wolfe KH Morden CW Palmer JD Function and evolution of a minimal plastid genome from a nonphotosynthetic parasitic plant Proc Natl Acad Sci USA 1992 89 10648 10652 1332054
Sears BB Chiu W-L Wolfson R Tsenewaki K Replication slippage as a molecular mechanism for evolutionary variation in chloroplast DNA due to deletions and insertions Plant genome and plastome: their structure and evolution 1995 Tokyo: Kodansha Scientific Ltd 139 146
Dujon B Group I introns as mobile genetic elements: Facts and mechanistic speculations – A review Gene 1989 82 91 114 2555264 10.1016/0378-1119(89)90034-6
Hess WR Hoch B Zeltz P Hubschmann T Kossel H Borner T Inefficient rpl2 splicing in barley mutants with ribosome-deficient plastids Plant Cell 1994 6 1455 1465 7994178 10.1105/tpc.6.10.1455
Hubschmann T Hess WR Borner T Impaired splicing of the rps12 transcript in ribosome-deficient plastids Plant Mol Biol 1996 30 109 123 8616228 10.1007/BF00017806
Jenkins BD Kulhanek DJ Barkan A Nuclear mutations that block group II RNA splicing in maize chloroplasts reveal several intron classes with distinct requirements for splicing factors Plant Cell 1997 9 283 296 9090875 10.1105/tpc.9.3.283
Vogel J Borner T Hess WR Comparative analysis of splicing of the complete set of chloroplast group II introns in three higher plant mutants Nucleic Acids Res 1999 27 3866 3874 10481026 10.1093/nar/27.19.3866
Turmel M Lemieux C Burger G Lang BF Otis C Plante I Gray MW The complete mitochondrial DNA sequences of Nephroselmis olivacea and Pedinomonas minor: two radically different evolutionary patterns within green algae Plant Cell 1999 11 1717 1729 10488238 10.1105/tpc.11.9.1717
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Lowe TM Eddy SR tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence Nucleic Acids Res 1997 25 955 964 9023104 10.1093/nar/25.5.955
Tesler G GRIMM: genome rearrangements web server Bioinformatics 2002 18 492 493 11934753 10.1093/bioinformatics/18.3.492
Notredame C Higgins DG Heringa J T-Coffee: a novel method for fast and accurate multiple sequence alignment J Mol Biol 2000 302 205 217 10964570 10.1006/jmbi.2000.4042
Livingstone CD Barton GJ Protein sequence alignments: a strategy for the hierarchical analysis of residue conservation Comput Appl Biosci 1993 9 745 756 8143162
|
16236178
|
PMC1277820
|
CC BY
|
2021-01-04 16:02:56
|
no
|
BMC Biol. 2005 Oct 20; 3:22
|
utf-8
|
BMC Biol
| 2,005 |
10.1186/1741-7007-3-22
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1301621267010.1186/1471-2407-5-130Research ArticleModeling the effect of age in T1-2 breast cancer using the SEER database Tai Patricia [email protected] Gábor [email protected] De Steene Jan [email protected] Georges [email protected] Mia [email protected] Melanie [email protected] Sang-Joon [email protected] Vincent [email protected] Guy [email protected] University of Saskatchewan, Faculty of Medicine, Department of Radiation Oncology, Regina, Canada2 Bács-Kiskun County Teaching Hospital, Surgical Pathology, Kecskemét, Hungary3 AZ-VUB, Oncologisch Centrum, and BISI-VUB, Computer Science and Medical Informatics, Jette, Belgium4 Geneva University Hospitals, Department of Gynecology and Obstetrics, Senology and Gynecologic Oncology Unit, Geneva, Switzerland5 University of New Mexico, Cancer Research and Treatment Center, Albuquerque, New Mexico, USA6 University of New Mexico, Department of Internal Medicine, Division of Epidemiology and Biostatistics, Albuquerque, New Mexico, USA2005 8 10 2005 5 130 130 25 2 2005 8 10 2005 Copyright © 2005 Tai 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
Modeling the relationship between age and mortality for breast cancer patients may have important prognostic and therapeutic implications.
Methods
Data from 9 registries of the Surveillance, Epidemiology, and End Results Program (SEER) of the United States were used. This study employed proportional hazards to model mortality in women with T1-2 breast cancers. The residuals of the model were used to examine the effect of age on mortality. This procedure was applied to node-negative (N0) and node-positive (N+) patients. All causes mortality and breast cancer specific mortality were evaluated.
Results
The relationship between age and mortality is biphasic. For both N0 and N+ patients among the T1-2 group, the analysis suggested two age components. One component is linear and corresponds to a natural increase of mortality with each year of age. The other component is quasi-quadratic and is centered around age 50. This component contributes to an increased risk of mortality as age increases beyond 50. It suggests a hormonally related process: the farther from menopause in either direction, the more prognosis is adversely influenced by the quasi-quadratic component. There is a complex relationship between hormone receptor status and other prognostic factors, like age.
Conclusion
The present analysis confirms the findings of many epidemiological and clinical trials that the relationship between age and mortality is biphasic. Compared with older patients, young women experience an abnormally high risk of death. Among elderly patients, the risk of death from breast cancer does not decrease with increasing age. These facts are important in the discussion of options for adjuvant treatment with breast cancer patients.
==== Body
Background
In many clinical situations, age is an important determinant of treatment decision in breast cancer. For example, after mastectomy, patients with T2 tumors and one to three positive nodes are at high risk of isolated loco-regional recurrences. Authors have advocated the routine use of postmastectomy radiotherapy in those patients who have T2 tumors and who are younger than 45 years [1]. In another study about close margins at mastectomy, the subgroup of patients aged 50 or younger with clinical T1-2 tumors and 0–3 positive nodes who have close (5 mm or less) or positive margins were at high risk (28% at 8 years) for chest wall recurrence regardless of adjuvant systemic therapy. Therefore, such patients should be considered for postmastectomy radiation [2]. Young women aged less than 45 should be regarded as high-risk patients, on the basis of age alone, and should be given adjuvant cytotoxic treatment [3]. The latter study showed a non-linear relationship between age and relative risk of dying.
At the other end of the age spectrum, breast cancers in elderly patients have been considered by some authors to exhibit a less aggressive behavior than in younger patients [4,5]. Other authors have argued that breast cancer does not become more indolent as age increases [6].
There are still controversial issues about the relationship between age and prognosis in breast cancer. Detailed analysis would be useful in order to provide more insight into this relationship. In the present study, we used proportional hazards to model the survival of T1-2, node-negative (N0) and node-positive (N+) breast cancer patients. Outcomes which we considered included all-cause mortality and cancer specific mortality from breast cancer. The primary aim of the study is to present how age relates with the risk of death. The secondary objective is to search for a simple algebraic representation of this relationship.
Methods
The Surveillance, Epidemiology, and End Results Program (SEER) of the United States collected data about the incidence of cancer and related matters from 11 population-based registries [7]. The data extracted in this study was from 9 registries: San Francisco-Oakland, Connecticut, Metropolitan Detroit, Hawaii, Iowa, New Mexico, Seattle (Puget Sound), Utah, and Metropolitan Atlanta. Selected patients were women who were without previous history of cancer and presented with non-inflammatory invasive breast carcinoma, diagnosed and histologically confirmed pT1-2 pM0 between 1988 and 1997, and for whom curative surgery and axillary lymph node dissections were performed. In 1987, the American Joint Committee on Cancer (AJCC) staging defined pT1 tumors as 2 cm or less in greatest dimension, and pT2 tumor as more than 2 cm but not more than 5 cm in greatest dimension. These definitions did not change until 1997. Some records were rejected because of concerns about the quality of data: non-hospital based data records, uncertain sequence of treatment, unknown month of diagnosis and unknown race. Records with missing histological grade and receptor status were not excluded. Examination of statistical outliers excluded one case with 75 nodes involved. Events for the study were death from all causes and death from breast cancer. Follow-up cutoff date was December 31, 1999 as provided by the database.
In order to verify the linearity of the continuous variables, the martingale residuals (differences between observed and expected numbers of events) were used. The martingale residuals were examined by a non-parametric smoothing (fitting the scatter-plots of residuals) against the quantitative covariates of interest. The smoothing used a Poisson regression implementation of generalized additive model (GAM) [8]. The GAM procedure provided two outputs. One was the non-parametric smoothed curves approximating the residuals. The other was a significance test of the non-linearity of the curves. For the covariates that significantly departed from linearity, an iterative search was performed to identify parametric families of functions that approximated the curves. The criteria used to end the search were: [a] simple parametric expression, [b] the corresponding function introduced as a transform in the Cox model satisfying the GAM linearity test, and [c] without deteriorating the model fit as assessed by the sum of squares of "deviance residuals" [8]. If the transforms were valid, the graphical displays should be linear shapes, and the non-linearity test results should be non-significant. Finally, scaled Schoenfeld residuals were used to verify that the relative hazards were constant over time [9]. The hypothesis underlying this dual modeling approach was as follows. If the algebraic functions are valid, their use as plug-in transforms should appropriately linearize the functional forms of the covariates of interest. Other information about the implementation of these procedures have been described earlier [10-12].
The analysis was applied first to node-negative cases ("training set") in order to find a simple expression of the functional form which relates age to mortality. The functional form obtained from node-negative cases was then applied to node-positive cases ("validation set"). In addition to the validation with the same transformation which was obtained for node-negative patients, a further iterative search was performed in order to improve the fit for node-positive patients.
This analysis was applied also to a European dataset, the German Breast Cancer Study Group (GBSG-2), in which the outcome studied was disease-free survival [14]. From a data analysis perspective, this GBSG-2 dataset is a very different database of 686 patients containing some extreme observations. One case had 51 involved nodes (range for other patients 1–38), and another case had a tumor size of 120 mm (range for other patients 3–100). There were 299 events (either recurrence of disease or death) in this German database.
The statistical analyses were performed with Splus (Insightful Corporation, Seattle, WA, USA) statistical software. Parametric fitting of curves used TableCurve 2D (Systat Software Inc, Richmond, CA, USA).
Results
There are 83,804 T1-2 cases (58,139 N0 and 25,665 N+, mean: 4 nodes involved, range: 1–48) available for analysis from the SEER database. Table 1 shows the characteristics of the patients. This table has been presented elsewhere [13]. Except for 28 additional cases (because of updated registration), there are no noticeable differences in the distribution of the characteristics. Table 2 shows the results of proportional hazards models in N0 and N+ groups, without using transforms for covariates. The supplemental Table 2b (Additional file 1) shows results of the check for Cox proportional hazards for all covariates. Note that some P-values are very small because of the very large size of the data. The rho-values (slope) indicate very small departures from the assumption of proportional hazards.
Figures 1 and 2 show graphically the effect of age on the log hazard ratio for death from all causes, for N0 and N+ patients, respectively. Both curves have similar U-shapes. The mortality is lowest for patients about 50 years of age at diagnosis. The mortality increases the farther away from 50 years of age at diagnosis, for both younger and older patients.
The shape of the smoothed curve for age suggests the use of a quadratic function. A fractional polynomial analogous to Sauerbrei and Royston [14], but with different exponents, combining a linear term (age) and a quasi-quadratic term |age-50|1.5, i.e. age+ |age-50|1.5, provides a good fit and passes the test of linearity (Chi-square = 6.530, P = 0.089) in N0 patients (Table 3).
We note that the age transform derived from node-negative cases does not provide a perfect linearization in N+ patients (Table 3). A better linearization in N+ patients was obtained by replacing the 1.5 exponent with 1.8, though without improving global model fit (Table 3).
The proportional hazard check for age shows a deviation from the assumption of constant hazard (Table 3). The "rho" values are positive when considering overall mortality, i.e. an increasing risk of death with longer follow-up. The values are negative when considering breast cancer specific mortality, i.e. a decreasing risk of breast cancer death with longer follow-up.
The age transforms suggest two components in the effect of age. One component is linear (linear for the log hazard ratio, i.e. exponential for the hazard ratio) and corresponds to a natural increase in mortality with each year of age. The other component is quasi-quadratic and is centered around age 50. It contributes to an increased risk of mortality as age increases beyond 50. It suggests a hormonally related process, not pre- versus post-menopausal, but perimenopausal versus non-perimenopausal (premenopausal + postmenopausal). The further age at diagnosis is from the age at menopause, the more prognosis is influenced by the quasi-quadratic component.
The results display a complex functional form of the effect of age on mortality. The curves clearly highlight the biological anomaly that younger patients experience the same relative mortality risk from all causes as do older patients. Figures 1 and 2 show that a 30-year old patient has a risk of death almost equal to a 60-year old patient.
The marked increase in mortality risk at older ages is attributable to the increased risk of death from causes other than breast cancer (co-morbidity). It should be noted that breast cancer does not become less virulent in older patients. An increase in the risk of death from breast cancer associated with older age was observed both in N0 and in N+ patients (Figures 3 and 4).
The German Breast Cancer Study Group GBSG-2 dataset [14] is a separate database of 686 patients. Using the GAM procedure on the GBSG-2 data, age was significantly non-linear (Chi2 = 31.744, 3 degrees of freedom, P < 0.000001). The age transforms improved the linearity for the age variable, and also improved the proportional hazards model (Table 4).
Discussion
In studies addressing the effect of age on breast cancer, several authors have reported a biphasic mortality [15-19]. This large study concurs with others in the literature. As in any modeling, the validity and the utility of the model may be questioned. Data from the GBSG-2 study were considered for verification of the model. The GBSG-2 study differs from the present SEER study in several respects. This German study was a prospective controlled clinical trial about the adjuvant treatment of node-positive breast cancer patients. Inclusion of patients was not restricted by tumor size. Histopathological classification and grading were performed centrally by one reference pathologist. The GBSG-2 data have been extensively investigated for the effect of age on the prognosis of breast cancer [20]. The GBSG-2 data thus provide an indication of the capability of our results to be extrapolated to a different population. It is also complementary, since the SEER has no data on recurrence and can provide no information on disease-free survival.
Applying different methods to estimate the effect of age on event-free survival of breast cancer (linear, categorization based on cutpoints, classification and regression trees, quadratic, fractional polynomial, cubic splines), Hollaender found that all methods showed a decrease in risk with increasing age up to 45–50 years [20]. A slight increase in risk was observed for older patients in the GBSG-2 data. Taking into account the wide confidence intervals for ages older than 80 years, our Figure 4 for node-positive breast cancer specific survival shows a good concordance with the node-positive GBSG-2 event-free survival.
Regarding the proportional hazards assumption, Hollaender noted that assuming a linear risk function, a small correlation value rho of 0.147 was obtained [20]. Our result for the GBSG-2 data shows the value of rho to be 0.131 (Table 4). The small difference is attributable to the incorporation of different covariates to our proportional hazards model (additional file 2 "outputgbsg2.doc"). For the SEER data, the rho values are smaller (Table 3).
Our results are also in keeping with a closely related investigation of the SEER data in which a group of 4,616 patients 35 years old or younger was compared to a group of 20,319 patients aged 50–55 years [21]. The authors observed that younger breast cancer patients had poorer survival explained in part by presentation with later stage disease and more aggressive tumors, in terms of grade and receptor status. But the known factors could not account for the remaining unexplained difference in survival. In contradiction, recently Rapiti et al have argued that age is not an independent prognostic factor when accounting for breast tumor characteristics and treatment [22]. However, this latter study included only 82 patients who were 35 years old or younger.
In order to try to understand the biphasic mortality, we looked at hormonal status and treatments of the patients. The age of 50 corresponds to the menopause. A large proportion of younger women were estrogen receptor (ER) negative (Figure 5). The proportion of ER-negative patients decreases with increasing age without any inflection. On the other hand, the proportion of progesterone receptor (PR) negative patients increases at age 50 then slowly decreases again. The reporting of hormonal receptor status is incomplete in SEER (~33–35% missing data).
Data on systemic treatment were not available from the SEER database, but the types of surgery and radiotherapy were provided. Mastectomy was performed less frequently on younger patients, but increased markedly among older patients. Post-operative radiotherapy was given less frequently at both ends of the age spectrum; somewhat less frequently in the young and considerably less frequently in the elderly patients (Figure 6). Researchers have reported under-treatment of elderly patients and this fact may account in part for the poor prognosis in the elderly [23-25]. Whether hormonal status or type of treatment or other factors may explain the biphasic mortality will need to be researched.
There are several limitations in the present analysis. The data are retrospective. Several orders of statistically significant interactions have not been incorporated in the models. Receiving systemic treatment is a particularly important prognostic factor in younger patients [3], but data on systemic treatment were not available for analysis.
Despite the limitations and regardless of the modeling, our major finding is that the relationship of age and mortality is biphasic. Such a finding has been described by many other authors [16,17,20,26]. It is important to remember this biphasic relationship when analyzing the effect of age on patients with breast cancer. Otherwise, there is a substantial risk of misinterpreting results when age is inappropriately categorized [26] or inappropriately modeled. (Table 2 would suggest erroneously almost no effect of age on mortality). Taking into account the full shape of the relationship between age and breast cancer specific mortality, we conclude that: 1) young women experience a much higher risk of death than do older patients; 2) among elderly patients, the risk of death from breast cancer does not decrease with increasing age. These are two facts that should be remembered by those when discussing adjuvant treatment with breast cancer patients.
Conclusion
The present analysis confirms that the relationship between age and mortality is biphasic. It is important that clinical research takes this relationship into account.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PT, VVH: writing the manuscript; SJL, VVH: data analysis;
GCs, JVDS, GV, MR, MV, GS: concept, design, and drafting 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
Check of the Cox proportional hazards assumption.
Click here for file
Additional File 3
Output of Cox, GAM and cox zph.
Click here for file
Additional File 2
GBSG-2 data set.
Click here for file
Acknowledgements
GCs: János Bolyai Research Fellowship from the Hungarian Academy of Sciences.
Figures and Tables
Figure 1 Mortality for all causes as a function of age for N0 patients. Dotted lines: twice-standard-error.
Figure 2 Mortality for all causes as a function of age for N+ patients. Dotted lines: twice-standard-error.
Figure 3 Breast cancer mortality as a function of age for N0 patients. Dotted lines: twice-standard-error.
Figure 4 Breast cancer mortality as a function of age for N+ patients. Dotted lines: twice-standard-error.
Figure 5 Distribution of hormone receptor status by age.
Figure 6 Distribution of treatment by age.
Table 1 Characteristics of the patients.
Characteristics of patients Total No. % of all cases Total N(0) % of all cases Total N(+) % of all cases
Patient number 83804 100.0% 58139 69.4% 25665 30.6%
SEER area:
East (Connecticut, Detroit, Atlanta) 31795 37.9% 21683 25.9% 10112 12.1%
Central (Iowa, New Mexico, Utah) 20459 24.4% 14202 16.9% 6257 7.5%
West (San Francisco-Oakland, Hawaii, Seattle) 31550 37.6% 22254 26.6% 9296 11.1%
Year of diagnosis:
1988–89 15390 18.4% 10424 12.4% 4966 5.9%
1990–91 16660 19.9% 11589 13.8% 5071 6.1%
1992–93 17042 20.3% 11894 14.2% 5148 6.1%
1994–95 17072 20.4% 11897 14.2% 5175 6.2%
1996–97 17640 21.0% 12335 14.7% 5305 6.3%
Age at diagnosis (years):
20–39 5869 7.0% 3396 4.1% 2473 3.0%
40–49 16231 19.4% 10262 12.2% 5969 7.1%
50–59 18051 21.5% 12362 14.8% 5689 6.8%
60–69 20703 24.7% 15110 18.0% 5593 6.7%
70–79 17161 20.5% 12802 15.3% 4359 5.2%
80+ 5789 6.9% 4207 5.0% 1582 1.9%
Race:
White and Other 78025 93.1% 54572 65.1% 23453 28.0%
Black 5779 6.9% 3567 4.3% 2212 2.6%
Marital status at diagnosis:
Single, widowed, other 33224 39.6% 23171 27.6% 10053 12.0%
Married 50580 60.4% 34968 41.7% 15612 18.6%
Topography:
Inner quadrant 12537 15.0% 9726 11.6% 2811 3.4%
Others 71267 85.0% 48413 57.8% 22854 27.3%
Histology:
Ductal 64370 76.8% 43924 52.4% 20446 24.4%
Others 19434 23.2% 14215 17.0% 5219 6.2%
Grade:
Poor/undifferentiated 25161 30.0% 15265 18.2% 9896 11.8%
Others 58643 70.0% 42874 51.2% 15769 18.8%
ER/PR status (+ includes unspecified):
ER+ and PR+ 63306 75.5% 44332 52.9% 18974 22.6%
ER+ and PR- 7001 8.4% 4852 5.8% 2149 2.6%
ER- and PR+ 2554 3.0% 1696 2.0% 858 1.0%
ER- and PR- 10943 13.1% 7259 8.7% 3684 4.4%
Tumor size (mm) (T stage):
< = 5 (T1a) 3349 4.0% 3069 3.7% 280 0.3%
> 5 and < = 10 (T1b) 16997 20.3% 14691 17.5% 2306 2.8%
> 10 and < = 20 (T1c) 36754 43.9% 26443 31.6% 10311 12.3%
> 20 and < = 50 (T2) 26704 31.9% 13936 16.6% 12768 15.2%
Number of nodes examined:
1–9 14225 17.0% 10525 12.6% 3700 4.4%
10–14 27004 32.2% 19210 22.9% 7794 9.3%
15–19 23102 27.6% 15820 18.9% 7282 8.7%
20+ 19473 23.2% 12584 15.0% 6889 8.2%
Number of nodes involved:
0 58139 69.4% 58139 69.4% 0 0.0%
1–3 16778 20.0% 0 0.0% 16778 20.0%
4+ 8887 10.6% 0 0.0% 8887 10.6%
Breast conserving surgery/Radiotherapy
no/no 46862 55.9% 31752 37.9% 15110 18.0%
no/yes 3972 4.7% 1024 1.2% 2948 3.5%
yes/no 4438 5.3% 2946 3.5% 1492 1.8%
yes/yes 28532 34.0% 22417 26.7% 6115 7.3%
ER = estrogen receptor, N(0) = node negative, N(+) = node positive, PR = progesterone receptor, SEER = Surveillance, Epidemiology, and End Results Program.
Table 2 Results of Cox proportional hazards models without transform. Hazard ratios (95% confidence intervals), values > 1 indicate increased risk of death.
Overall mortality Breast cancer specific mortality
N0 N+ N0 N+
SEER central area 1.00 (0.95–1.05) 0.93 (0.87–0.98) 1.03 (0.94–1.13) 0.92 (0.86–0.99)
SEER western area 0.92 (0.88–0.97) 0.87 (0.82–0.92) 0.82 (0.75–0.90) 0.86 (0.80–0.91)
Race black 1.38 (1.27–1.49) 1.42 (1.31–1.52) 1.37 (1.21–1.55) 1.44 (1.31–1.57)
Married 0.75 (0.72–0.78) 0.82 (0.78–0.86) 0.88 (0.81–0.95) 0.91 (0.86–0.97)
Inner quadrant 1.09 (1.04–1.15) 1.16 (1.08–1.24) 1.32 (1.21–1.44) 1.3 (1.19–1.41)
Ductal histology 1.14 (1.09–1.2) 1.11 (1.04–1.17) 1.44 (1.31–1.58) 1.12 (1.05–1.21)
ER negative 1.39 (1.28–1.5) 1.52 (1.41–1.64) 1.58 (1.4–1.78) 1.58 (1.44–1.73)
PR negative 1.11 (1.03–1.18) 1.27 (1.18–1.36) 1.36 (1.21–1.52) 1.39 (1.28–1.52)
Grade 3–4 1.23 (1.17–1.29) 1.34 (1.28–1.41) 1.6 (1.48–1.73) 1.48 (1.39–1.57)
BCS 1.12 (1.02–1.24) 0.96 (0.86–1.07) 1.01 (0.84–1.20) 0.98 (0.86–1.12)
Radiotherapy 1.12 (0.96–1.29) 0.89 (0.83–0.95) 1.31 (1.08–1.6) 0.91 (0.84–0.99)
BCSxRT 0.61 (0.51–0.73) 0.92 (0.8–1.06) 0.63 (0.48–0.82) 0.88 (0.74–1.04)
Year Diagnosis (continuous, year) 0.98 (0.97–0.99) 0.96 (0.95–0.97) 0.92 (0.91–0.94) 0.94 (0.93–0.95)
Age at Diagnosis (continuous, year) 1.05 (1.05–1.05) 1.02 (1.02–1.03) 1.00 (1.00–1.01) 1.00 (1.00–1.01)
Tumor size (continuous, mm) 1.03 (1.03–1.03) 1.02 (1.02–1.02) 1.05 (1.04–1.05) 1.02 (1.02–1.03)
Number positive nodes (continuous, n) (-) 1.07 (1.07–1.08) (-) 1.09 (1.08–1.09)
Number nodes examined (continuous, n) 0.99 (0.99–0.99) 0.98 (0.97–0.98) 0.99 (0.99–1.00) 0.97 (0.96–0.97)
BCS = breast conserving surgery, BCSxRT = breast conserving surgery and radiotherapy, ER = estrogen receptor, N0 = node negative, N+ = node positive, PR = progesterone receptor, SEER = Surveillance, Epidemiology, and End Results Program.
Table 3 Results of GAM and proportional hazards tests for SEER data
Overall mortality Node-negative Node-positive
no transform transform age+|age-50|1.5 transform age+|age-50|1.8 no transform transform age+|age-50|1.5 transform age+|age-50|1.8
GAM Chisq for Age (smaller value better) 582.4 6.53 52.4 439.2 18.4 1.83
GAM p-value for Age (larger better) <.0001 0.089 <.0001 <.0001 0.0003 0.557
Test PH for Age: rho (smaller absolute value better) 0.069 0.071 0.075 0.048 0.044 0.045
Test PH for Age: chisq (smaller value better) 53.6 46.3 48.2 21.1 15.8 16.0
Full PH model Rsquare (larger value better) 0.088 0.094 0.093 0.126 0.135 0.135
Full PH model Likelihood ratio test (larger value better) 5364 5745 5656 3443 3713 3709
Breast cancer specific mortality Node-negative Node-positive
no transform transform age+|age-50|1.5 transform age+|age-50|1.8 no transform transform age+|age-50|1.5 transform age+|age-50|1.8
GAM Chisq for Age (smaller value better) 52.7 11.1 8.10 63.7 9.20 4.92
GAM p-value for Age (larger better) <.0001 0.010 0.040 <.0001 0.027 0.179
Test PH for Age: rho (smaller absolute value better) -0.027 -0.038 -0.042 -0.007 -0.023 -0.027
Test PH for Age: chisq (smaller value better) 2.5 4.5 5.5 0.3 2.8 3.6
Full PH model Rsquare (larger value better) 0.03 0.03 0.031 0.1 0.1 0.1
Full PH model Likelihood ratio test (larger value better) 1779 1799 1803 2698 2715 2716
Chisq = chi-square value, GAM = generalized additive model, PH = proportional hazards. The additional file 3 "output4rev.doc" lists the output of the full PH models.
Table 4 Results of GAM and proportional hazards tests for GBSG-2 data
GBSG-2 data set no transform transform age+|age-50|1.5 transform age+|age-50|1.8
Hazard ratio for Age 1.001 1.004 1.002
p-value of Hazard ratio for Age 0.92 0.019 0.010
GAM Chisq for Age 31.74 9.29 8.70
GAM p-value for Age 0.0000006 0.0256 0.0336
Test PH for Age: rho 0.131 0.0459 0.0240
Test PH for Age: chisq 6.118 0.581 0.169
Full PH model Rsquare 0.135 0.142 0.143
Full PH model Likelihood ratio test 99.7 105 106
Chisq = chi-square value, GAM = generalized additive model, GBSG-2 = German Breast Cancer Study Group, PH = proportional hazards. The Additional file 2 "outputgbsg2.doc" lists the other covariates included in the PH models.
==== Refs
Fodor J Polgar C Major T Nemeth G Locoregional failure 15 years after mastectomy in women with one to three positive axillary nodes with or without irradiation the significance of tumor size Strahlenther Onkol 2003 179 197 202 12627264 10.1007/s00066-003-1010-7
Freedman GM Fowble BL Hanlon AL Myint MA Hoffman JP Sigurdson ER Eisenberg BL Goldstein LJ Fein DA A close or positive margin after mastectomy is not an indication for chest wall irradiation except in women aged fifty or younger Int J Radiat Oncol Biol Phys 1998 41 599 605 9635708 10.1016/S0360-3016(98)00103-5
Kroman N Jensen MB Wohlfahrt J Mouridsen HT Andersen PK Melbye M Factors influencing the effect of age on prognosis in breast cancer: population based study BMJ 2000 320 474 478 10678859 10.1136/bmj.320.7233.474
Hughes KS Schnaper LA Berry D Cirrincione C McCormick B Shank B Wheeler J Champion LA Smith TJ Smith BL Shapiro C Muss HB Winer E Hudis C Wood W Sugarbaker D Henderson IC Norton L Cancer and Leukemia Group B; Radiation Therapy Oncology Group; Eastern Cooperative Oncology Group Lumpectomy plus tamoxifen with or without irradiation in women 70 years of age or older with early breast cancer N Engl J Med 2004 351 971 977 15342805 10.1056/NEJMoa040587
Kunkler I Williams L Prescott R King C Re: Breast-conserving surgery with or without radiotherapy: pooled-analysis for risks of ipsilateral breast tumor recurrence and mortality J Natl Cancer Inst 2004 96 1255 15316061
Vinh-Hung V Verschraegen C Royce M Van de Steene J Tai P Cserni G Storme G Vlastos G Response: Breast-conserving surgery with or without radiotherapy: pooled-analysis for risks of ipsilateral breast tumor recurrence and mortality J Natl Cancer Inst 2004 96 1255 1257 15316061
National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) Program Public-Use Data (1973–1999), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April based on the November 2001 submission 2002 National Cancer Institute, Bethesda, MD
Therneau TM Grambsch PM Modeling survival data: extending the Cox model 2000 New York, NY, Springer-Verlag 87 152
Grambsch PM Therneau TM Proportional hazards tests and diagnostics based on weighted residuals Biometrika 1994 81 515 526
Vinh-Hung V Burzykowski T Cserni G Voordeckers M Van De Steene J Storme G Functional form of the effect of the numbers of axillary nodes on survival in early breast cancer Int J Oncol 2003 22 697 704 12579326
Verschraegen C Vinh-Hung V Cserni G Gordon R Royce ME Vlastos G Tai P Storme G Modeling the effect of tumor size in early breast cancer Ann Surg 2005 241 309 318 15650642 10.1097/01.sla.0000150245.45558.a9
Vinh-Hung V Gordon R Quantitative target sizes for breast tumor detection prior to metastasis: a prerequisite to rational design of 4D scanners for breast screening Technol Cancer Res Treat 2005 4 11 21 15649083
Vinh-Hung V Burzykowski T Van de Steene J Storme G Soete G Post-surgery radiation in early breast cancer: survival analysis of registry data Radiother Oncol 2002 64 281 290 12242116 10.1016/S0167-8140(02)00105-6
Sauerbrei W Royston P Building multivariable prognostic and diagnostic models: transformation of the predictors by using fractional polynomials J Roy Stat Soc A Sta 1999 162 71 94 10.1111/1467-985X.00122
Bryant J Fisher B Gunduz N Costantino JP Emir B S-phase fraction combined with other patient and tumor characteristics for the prognosis of node-negative, estrogen-receptor-positive breast cancer Breast Cancer Res Treat 1998 51 239 253 10068082 10.1023/A:1006184428857
Sauerbrei W Royston P Bojar H Schmoor C Schumacher M Modelling the effects of standard prognostic factors in node-positive breast cancer. German Breast Cancer Study Group (GBSG) Br J Cancer 1999 79 1752 1760 10206288 10.1038/sj.bjc.6690279
Adami HO Malker B Holmberg L Persson I Stone B The relation between survival and age at diagnosis in breast cancer N Engl J Med 1986 315 559 563 3736639
Fisher ER Anderson S Tan-Chiu E Fisher B Eaton L Wolmark N Fifteen-year prognostic discriminants for invasive breast carcinoma: National Surgical Adjuvant Breast and Bowel Project Protocol-06 Cancer 2001 91 1679 1687 11309768 10.1002/1097-0142(20010415)91:8+<1679::AID-CNCR1183>3.0.CO;2-8
Aebi S Gelber S Castiglione-Gertsch M Gelber RD Collins J Thurlimann B Rudenstam CM Lindtner J Crivellari D Cortes-Funes H Simoncini E Werner ID Coates AS Goldhirsch A Is chemotherapy alone adequate for young women with oestrogen-receptor-positive breast cancer? Lancet 2000 355 1869 1874 10866443 10.1016/S0140-6736(00)02292-3
Holländer N Estimating the functional form of the effect of a continuous covariate on survival time Dissertation 2002 Dortmund University
Maggard MA O'Connell JB Lane KE Liu JH Etzioni DA Ko CY Do young breast cancer patients have worse outcomes? J Surg Res 2003 113 109 113 12943818 10.1016/S0022-4804(03)00179-3
Rapiti E Fioretta G Verkooijen HM Vlastos G Schafer P Sappino AP Kurtz J Neyroud-Caspar I Bouchardy C Survival of young and older breast cancer patients in Geneva from 1990 to 2001 Eur J Cancer 2005 41 1446 1452 15919199 10.1016/j.ejca.2005.02.029
Bouchardy C Rapiti E Fioretta G Laissue P Neyroud-Caspar I Schafer P Kurtz J Sappino AP Vlastos G Undertreatment strongly decreases prognosis of breast cancer in elderly women J Clin Oncol 2003 21 3580 3587 12913099 10.1200/JCO.2003.02.046
Lash TL Silliman RA Guadagnoli E Mor V The effect of less than definitive care on breast carcinoma recurrence and mortality Cancer 2000 89 1739 1747 11042569 10.1002/1097-0142(20001015)89:8<1739::AID-CNCR14>3.0.CO;2-F
Truong PT Lee J Kader HA Speers CH Olivotto IA Locoregional recurrence risks in elderly breast cancer patients treated with mastectomy without adjuvant radiotherapy Eur J Cancer 2005 41 1267 1277 15939262 10.1016/j.ejca.2005.02.027
Vinh-Hung V Verschraegen C Royce M Van de Steene J Tai P Cserni G Storme G Vlastos G Re: Breast-conserving surgery with or without radiotherapy: pooled-analysis for risks of ipsilateral breast tumor recurrence and mortality J Natl Cancer Inst 2004 96 1255 1257 15316061
|
16212670
|
PMC1277821
|
CC BY
|
2021-01-04 16:03:05
|
no
|
BMC Cancer. 2005 Oct 8; 5:130
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-130
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1361624202110.1186/1471-2407-5-136Research ArticleLong-term all-sites cancer mortality time trends in Ohio, USA, 1970–2001: differences by race, gender and age Tyczynski Jerzy E [email protected] Hans J [email protected] Cancer Prevention Institute, Dayton, Ohio, USA2005 20 10 2005 5 136 136 5 5 2005 20 10 2005 Copyright © 2005 Tyczynski and Berkel; 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 were significant changes in cancer mortality in the USA over the last several decades, in the whole country and in particular states. However, no in depth analysis has been published so far, dealing with changes in mortality time trends in the state of Ohio. Since the state of Ohio belongs to the states of relatively high level of all-sites mortality in both males and females, it is of interest to analyze recent changes in mortality rates, as well as to compare them with the situation in the rest of the USA. The main aim of this study was to analyze, describe and interpret all-sites cancer mortality time trends in the population of the State of Ohio.
Methods
Cancer mortality data by age, sex, race and year for the period 1970–2001 were obtained from the Surveillance Research Program of the National Cancer Institute SEER*Stat software. A joinpoint regression methodology was used to provide estimated annual percentage changes (EAPCs) and to detect points in time where significant changes in the trends occurred.
Results
In both, males and females mortality rates were higher in blacks compared with whites. The difference was bigger in males (39.9%) than in women (23.3%). Mortality rates in Ohio are generally higher than average USA rates – an overall difference was 7.5% in men in 1997–2001, and 6.1% in women. All-sites mortality trends in Ohio and in the whole USA are similar. However, in general, mortality rates in Ohio remained elevated compared with the USA rates throughout the entire analyzed period. The exceptions are the rates in young and middle-aged African Americans.
Conclusion
Although direction of time trends in Ohio are similar in Ohio and the whole US, Ohio still have cancer mortality rates higher than the US average. In addition, there is a significant discrepancy between white and black population of Ohio in all-sites mortality level, with disadvantage for Blacks. To diminish disparities in cancer mortality between African Americans and white inhabitants of Ohio efforts should be focused on increasing knowledge of black people regarding healthy lifestyle and behavioral risk factors, but also on diminishing socioeconomic differences, and last but not least, on better access to medical care.
==== Body
Background
Mortality from cancer is an indicator of the effectiveness of cancer control efforts in a given population, and, thus, is an important measure from the public health point of view. Declining cancer mortality rates may indicate that cancer prevention activities have been successfully designed and implemented, may also indicate favorable results of the implementation of improved therapeutic procedures, as well as successful early detection programs. On the other hand, increases in mortality rates can indicate failures in controlling cancer risk factors and/or an appearance of new ones. Hence, the analysis of cancer mortality time trends and patterns in a given population can be helpful in assessing successes, failures and future need in cancer control programs.
There were significant changes in cancer mortality in the USA over the last several decades. These changes were described many times, however, dealing mainly with mortality in the USA as a whole [1-5]. Much less has been published on cancer mortality time trends in particular sub-populations, at the state or county levels [6-8]. In particular, no in depth analysis has been published so far, dealing with changes in mortality time trends in the state of Ohio. Since the state of Ohio belongs to the states with relatively high level of total cancer mortality in both males and females, it is of interest to analyze recent changes in mortality rates, as well as to compare them with the situation in the rest of the USA.
There are several methods of assessing changes in cancer mortality trends over time. The simplest way is based on visual inspection of rates, while more sophisticated methods are based on statistical modeling of observed data (e.g. age-period-cohort modeling). For the purpose of this study we decided to employ a joinpoint regression analysis (a non-linear regression modeling known as piece-wise or segmented regression). This approach was chosen to allow for detecting points in time where significant changes in the direction of trend occurred, as well as to assess average percentage changes in mortality rates.
The main goal of this paper is to analyze and discuss changes in total cancer mortality in the state of Ohio, and to compare this to the national USA patterns, and discuss possible reasons for existing differences both within Ohio, and between Ohio and the USA as a whole.
Methods
Cancer mortality data for all sites (ICD-9 codes 140–208) for the state of Ohio and for all states combined were obtained from the NCHS via Surveillance Research Program, National Cancer Institute SEER*Stat software – version 5.2.2 [9]. The data were available for the period 1970–2001. Corresponding population data, by age, sex, race and year, were extracted from the same source.
Age-standardized mortality rates (ASRs) were calculated for each calendar year, for all ages combined and for the following age groups 20–44, 45–64, and 65 and over (for each gender separately, for all races combined and for Whites and African Americans separately). The World Standard Population was used for age-adjustment [10]. To assess the most recent differences between sub-populations (by gender, age and race) average age-standardized mortality rates were also calculated for the last 5-year period (years 1997–2001 combined). Percentage differences for the period 1997–2001 between rates were calculated for blacks and whites, for particular age categories, and between Ohio and the USA.
A joinpoint regression was fitted to provide estimated annual percentage change (EAPC) and to detect points in time where significant changes in the trends occur [11,12]. A Joinpoint software version 2.6 was used [11]. For each EAPC estimate we also calculated the corresponding 95% confidence interval (95%CI). A maximum number of 3 joinpoints was allowed for estimations.
The joinpoint regression model describes continuous changes in rates and uses the grid-search method to fit the regression function with unknown joinpoints. In this model, the annual age-adjusted rates over a given period of time are examined and the points in time when the direction of the trends changes significantly are detected [12]. Thus, joinpoint is a useful way to summarize trends in cancer rates, and it allows one to assess recent changes in trend.
Results
General pattern of cancer mortality in Ohio
There were an average of 12,800 cancer deaths per year in men in Ohio in the period 1997–2001 (11,380 in white men and 1,400 in African Americans). In women, the average number of deaths per year in that period was 12,060 (10,800 and 1,260 in white and black populations respectively). An average age-adjusted mortality rate was 156.4/100,000 in males and 109.0/100,000 in females. In both, males and females mortality rates are higher in black populations compared with whites (Figure 1). The difference is bigger in males (39.9%) than in females (the difference of 23.3%). The biggest difference in rates between black and white populations was observed in males in the age group 45–64 year (53.7%), whilst the lowest difference was observed in elderly women (aged 65 and over) – 19.8% (Table 1). Mortality rates in Ohio are generally higher than average USA rates – an overall difference in men was 7.5% in 1997–2001, and 6.1% in women (Table 1). Also in particular age categories overall (in all races combined) mortality rates in Ohio were higher compared to the USA average – the biggest difference was noted in elderly men (aged 65 and more) – 8.7% (Table 1). The most frequently recorded cancer site was lung – in both males and females, and in both whites and blacks (Table 2). In white males the second most frequent site was colorectal followed by prostate, while in black males the second most frequent site was prostate (Table 2). In women, in both whites and blacks, lung was followed by breast and colorectal cancers (Table 2).
Changes in mortality in time – overall
Results of the joinpoint analysis of mortality time trends are shown in Table 1 and in Figure 2. For all ages combined in males, after a moderate increase of 0.7% per year until 1982, a significant decline in mortality occurred later on, reaching -1.5% per year after 1993 (Table 1). In women, mortality rates were going up until the end of 1980s, a significant, but moderate increase of 0.4% per year, after which a decline occurred with -0.8% per year (Table 1).
In young adults (aged 20–44 years) a permanent decline in mortality was observed in both males and females. Also, the EAPCs for both genders were similar (-1.4% and -1.5% per year in men and women respectively) (Table 1). In middle-aged men and women (45–64 years of age) a pattern of changes in time was also similar: after small increase by 0.3% per year in 1970–1985, a decline of rates occurred since 1986 (the EAPC for males was -1.9% and for females -1.5%) (Table 1). In older men (aged 65 and more) an increase of mortality was observed in the 1970s (by 1.2% per year), followed by plateau in the 1980s and beginning of the 1990s, and then followed by decrease by -1.3% since 1994 (Table 1). In elderly women an increase of mortality by 1.1% per year was observed until 1994, and then followed by significant decline in mortality (EAPC -0.9%) (Table 1).
Changes in mortality by race
There are some differences in time trends development between white and black populations. In men, all-ages mortality was increasing faster in blacks than in whites, and decline in African Americans commenced later than in whites. In women, the situation was opposite – mortality started to decline earlier in African American women (in 1984) than in white women (1990) (Table 1). In young adults, permanent decline in mortality was noted in both genders, regardless the race. Also the EAPCs were similar in both races in young people. In middle-aged white men plateau in mortality was observed until the mid of 1980s, while in black middle-aged men an increase by 0.8% per year was observed in the same time. Since the mid 1980s a decline in mortality was observed in both races, however higher EAPC was noted in African Americans (-2.8%) compared with whites (-1.6%) (Table 1). In middle-aged women, decline in mortality commenced earlier in blacks (1985) than in whites (1987) and was more pronounced in the former (-2.0% vs -1.5%). In the older men (65+) decline in mortality began in 1989 in African Americans and in 1994 in white men, and was similar in both races (-1.0% and -1.2%) (Table 1). In women aged 65 years and more decline in mortality has been observed since the mid of the 1990s and approximately two-fold faster in African American women (-1.7% per year) compared with white women (-0.9%) (Table 1).
Time trends in Ohio vs. whole USA
All-sites mortality trends in Ohio and in the whole USA are similar, however, in general mortality rates in Ohio remained elevated compared with the USA rates throughout the entire analyzed period. The exceptions are the rates in young and middle-aged African Americans. In middle-aged African Americans rates were at the similar level in Ohio and in the USA in the second half of the 1990s and the beginning of the 2000s (Table 1, Figure 3), in both males and females. In young African Americans (20–44 years of age) mortality rates in 1997–2001 were lower in Ohio than in the whole USA (Table 1).
Despite the differences in the absolute level of mortality rates in Ohio and the USA, trends are, in general, similar in directions. However, in the most recent periods the overall rates (for all-ages and all races combined) have been declining faster in the USA than in Ohio (Table 1), and this difference applied mainly to black populations. This phenomenon is also visible in young adults (20–44) where mortality seems to decline faster in the USA than in Ohio, especially in African Americans of both genders (Table 1). In contrast, mortality rates are declining faster in Ohio in the oldest age category, especially in women (in both races).
Discussion
The main aim of this study was to analyze, describe and interpret all-sites cancer mortality time trends in the population of Ohio State. We also attempted to compare the results for Ohio with those obtained for the whole USA. Some brief analyses of mortality time trends in Ohio were published before [13,14]. However, they dealt with all-ages time trends only, and did not use any statistical modeling to be applied to the observed data. In our analysis we have applied Joinpoint regression approach (called also stepwise approach) to examine data and to quantify observed changes. It is, to our knowledge, the first analysis of this type done for the Ohio population.
The population of Ohio is heterogeneous – according to the 2000 US Census, 85% of the population was white, 11% were African Americans, and 4% were other race. The heterogeneity is also visible in cancer mortality rates in Ohio. In all age categories and in both genders mortality rates in African Americans were elevated compared with whites. A similar phenomenon has been observed in the whole USA population. However, these differences are not identical in Ohio and in the USA, especially in young adults. In young people the difference in mortality between blacks and whites in Ohio is approximately 25%, while in the whole USA it reaches nearly 50%. Also in middle-aged males and females, as well as in elderly men differences between races are smaller in Ohio compared with the whole country. The only exception is the oldest age category in women, where mortality in Ohio exceeds that of the USA. The phenomenon of elevated cancer mortality in African Americans compared to non-Hispanic whites is well known in the literature [15]. It is also well recognized that people in lower socio-economic groups of the society tend to have higher cancer mortality rates than wealthy people. It was shown, among others, for breast cancer in black and white American women [16]. It is, thus, not surprising that also in Ohio cancer mortality is higher among African Americans compared with whites.
The question is why the difference between the two races is generally lower in Ohio than in the USA. It has been suggested by Bach and colleagues that the differences in survival (and consequently in mortality) are not because of biological factors, but more likely caused by other factors such as differences in treatment, stage of disease at the diagnosis, and co-morbidity (influence of other diseases) [17]. One of the possible explanations for the difference between Ohio and the USA as a whole is the access to the health care system for African Americans, measured by the percentage of people without health insurance. It has been pointed out by Prothrow-Stith and colleagues that lack of health insurance contributes to increased morbidity and mortality from cancer [18]. For example, in the year 2001 in the whole USA 19.0% of blacks had no health insurance, while in Ohio 16.2%. In the same year the proportion of uninsured whites was in the USA and in Ohio virtually the same, 10.0% and 9.7% respectively [19,20]. Our analysis showed that in young and middle-aged individuals, although overall cancer mortality (for all races combined) was higher in Ohio than in the USA, this phenomenon was not present in the black population of that age (Table 1). Moreover, in young adult African Americans (aged 20–44) all-sites cancer mortality was lower in Ohio than in the whole country. Also in all-ages mortality in men the difference between Ohio and the USA was two-fold higher for all races combined and for whites compared with blacks.
The difference between African Americans and white population in health insurance may also influence the difference in cancer mortality between the two sub-populations. In Ohio, black people constitute more than one fifth (22.2%) of all uninsured individuals, while they constitute only 11% of the whole population [21].
Similarly to the USA as a whole, cancer mortality rates have been decreasing also in Ohio. Favorable trends has been observed in the 1990s and the beginning of the 2000s in both males and females, for all ages combined, as well as in all analyzed age categories. However, while mortality was declining throughout the whole period among young people, decrease appeared in the mid of the 1980s in middle-aged individuals, and only in the mid of the 1990s in the oldest age group.
Although cancer mortality rates in Ohio are higher in African Americans compared with whites, recent trends in rates are favorable in both groups. All-sites cancer mortality is mostly influenced by those sites, which make the biggest proportion of cancer deaths. Hence, changes in mortality from the most common sites determine changes in all-sites mortality trends. In Ohio males, in both whites and blacks, three most common cancer sites (lung, prostate, and large bowel) constitute more than 50% of all cancer deaths. An analysis of lung cancer mortality in Ohio showed declining trends in both races [22]. In females, similarly to men, three the most frequent sites (lung, breast and large bowel) form over 50% of all cancers. Lung cancer mortality has been plateauing in women in both races in the 1990s [21].
There is a need for more detailed analysis of time trends for particular cancer sites and/or groups of sites, to be able to define these areas where intervention would be most desired. The most striking phenomenon observed in all-sites cancer mortality among Ohio inhabitants is a gap between black and white populations, with strongly unfavorable patterns among Blacks. It seems that in order to diminish disparities in cancer mortality between African Americans and white inhabitants of Ohio efforts should be focused on increasing knowledge of black people regarding healthy lifestyle and behavioral risk factors, but also on diminishing socioeconomic differences, and last but not least, on better access to medical care [23].
Pre-publication history
The pre-publication history for this paper can be accessed here:
Figures and Tables
Figure 1 Cancer mortality in the USA, 1997–2001, by state, gender and race # – Age-adjusted mortality rates (World Standard Population).
Figure 2 Cancer mortality time trends, all sites, Ohio, 1970–2001, by race # – Age-adjusted mortality rates (World Standard Population).
Figure 3 Cancer mortality time trends, all sites, Ohio vs USA, 1970–2001, by race # – Age-adjusted mortality rates (World Standard Population).
Table 1 All-sites cancer mortality time trends, 1970–2001, Ohio and USA, by gender, age and race
Sex, age, race Rate 1997–2001## Difference (%) Ohio vs. USA Difference (%) Blacks vs Whites Trend 1 Trend 2 Trend 3 Trend 4
Years EAPC# Years EAPC# Years EAPC# Years EAPC#
All ages
Males, OHIO 156.4 7.5 1970–1982 0.7a 1983–1992 -0.4a 1993–2001 -1.5a
Males, OHIO, white 152.1 7.3 39.9 1970–1982 0.5a 1983–1992 -0.4a 1993–2001 -1.4a
Males, OHIO, black 212.8 3.1 1970–1986 1.0a 1987–2001 -1.7a
Males, USA 145.5 - 1970–1979 0.5a 1980–1989 0.1 1990–1992 -0.8a 1993–2001 -1.8a
Males, USA, white 141.7 - 45.7 1970–1979 0.4a 1980–1990 0.1 1991–2001 -1.5a
Males, USA, black 206.4 - 1970–1981 1.6a 1982–1989 0.6a 1990–1993 -1.2a 1994–2001 -2.5a
Females, OHIO 109.0 6.1 1970–1989 0.4a 1990–2001 -0.8a
Females, OHIO, white 107.2 5.4 23.3 1970–1989 0.3a 1990–2001 -0.8a
Females, OHIO, black 132.2 5.9 1970–1983 1.1a 1984–2001 -0.6a
Females, USA 102.7 - 1970–1974 -0.2 1975–1990 0.4a 1991–2001 -1.0a
Females, USA, white 101.7 - 22.7 1970–1974 -0.2 1975–1990 0.4a 1991–2001 -1.0a
Females, USA, black 124.8 - 1970–1975 -0.1 1976–1991 0.7a 1992–2001 -1.3a
Age 20–44
Males, OHIO 18.4 7.0 1970–2001 -1.4a
Males, OHIO, white 18.2 10.3 24.2 1970–2001 -1.4a
Males, OHIO, black 22.6 -4.6 1970–2001 -1.6a
Males, USA 17.2 - 1970–1975 -2.3a 1976–1993 -1.2a 1994–2001 -2.4a
Males, USA, white 16.5 - 43.6 1970–1975 -2.6a 1976–1993 -1.3a 1994–2001 -2.1a
Males, USA, black 23.7 - 1970–1984 -0.7a 1985–1994 -1.7a 1995–2001 -4.5a
Females, OHIO 20.7 4.0 1970–2001 -1.5a
Females, OHIO, white 20.2 6.9 29.2 1970–2001 -1.6a
Females, OHIO, black 26.1 -7.8 1970–2001 -1.3a
Females, USA 19.9 - 1970–1977 -2.2a 1978–1989 -1.1a 1990–2001 -1.7a
Females, USA, white 18.9 - 49.7 1970–1980 -2.0a 1981–1983 0.0 1984–2001 -1.6a
Females, USA, black 28.3 - 1970–1975 -3.4a 1976–1990 -0.7a 1991–2001 -1.9a
Age 45–64
Males, OHIO 274.3 5.5 1970–1985 0.3a 1986–2001 -1.9a
Males, OHIO, white 263.9 6.2 53.7 1970–1984 0.2 1985–2001 -1.6a
Males, OHIO, black 405.7 -0.7 1970–1985 0.8a 1986–2001 -2.8a
Males, USA 259.9 - 1970–1977 0.4a 1978–1989 -0.2a 1990–2001 -2.3
Males, USA, white 248.5 - 64.4 1970–1988 0.0 1989–2001 -2.1a
Males, USA, black 408.6 - 1970–1977 1.7a 1978–1989 0.0 1990–2001 -3.0
Females, OHIO 221.9 5.0 1970–1985 0.3a 1986–2001 -1.5a
Females, OHIO, white 217.4 4.8 28.1 1970–1986 0.2 1987–2001 -1.5a
Females, OHIO, black 278.4 1.2 1970–1984 0.9a 1985–2001 -2.0a
Females, USA 211.3 - 1970–1986 0.1 1987–1992 -0.9a 1993–2001 -2.0a
Females, USA, white 207.5 - 32.6 1970–1987 0.1 1988–1993 -1.2a 1994–2001 -2.0
Females, USA, black 275.1 - 1970–1989 0.2a 1990–2001 -1.8
Age 65+
Males, OHIO 1,382.7 8.7 1970–1981 1.2a 1982–1993 0.2 1994–2001 -1.3a
Males, OHIO, white 1,350.4 7.9 34.0 1970–1981 1.1a 1982–1993 0.2 1994–2001 -1.2a
Males, OHIO, black 1,809.6 6.0 1970–1988 1.4a 1989–2001 -1.0a
Males, USA 1,272.2 - 1970–1979 1.1a 1980–1991 0.5a 1992–2001 -1.3a
Males, USA, white 1,251.7 - 36.4 1970–1979 1.0a 1980–1991 0.4a 1992–2001 -1.2a
Males, USA, black 1,707.2 - 1970–1981 2.3a 1982–1988 1.7a 1989–1992 0.5 1993–2001 -1.9a
Females, OHIO 840.8 7.3 1970–1994 1.1a 1995–2001 -0.9a
Females, OHIO, white 829.2 5.9 19.8 1970–1994 1.0a 1995–2001 -0.9a
Females, OHIO, black 993.7 12.4 1970–1995 1.4a 1996–2001 -1.7a
Females, USA 783.4 - 1970–1974 0.0 1975–1983 1.3a 1984–1992 1.0a 1993–2001 -0.4a
Females, USA, white 783.3 - 12.9 1970–1974 0.0 1975–1983 1.3a 1984–1992 0.9a 1993–2001 -0.4a
Females, USA, black 884.2 - 1970–1974 0.3 1975–1992 1.7a 1993–2001 -0.6a
(a) – p < 0.05
# – EAPC – Estimated Annual Percentage Change, ## – Age-adjusted standardized rate (World standard population)
Table 2 Cancer mortality in Ohio, 1997–2001, by gender and race
Cancer site Rate# No of deaths % Cancer site Rate# No of deaths %
Male – White Male – Black
All sites 152.1 56,924 100 All sites 212.8 7,057 100
Lung 52.3 18,996 33.4 Lung 70.9 2,306 32.7
Colon and Rectum 15.4 5,897 10.4 Prostate 31.1 1,130 16.0
Prostate 12.4 5,628 9.9 Colon and Rectum 21.0 705 10.0
Non-Hodgkin Lymphoma 7.0 2,642 4.6 Pancreas 10.0 328 4.6
Pancreas 7.0 2,569 4.5 Esophagus 7.7 233 3.3
Leukemia 6.3 2,302 4.0 Stomach 6.8 230 3.3
Esophagus 5.4 1,913 3.4 Liver 6.8 218 3.1
Urinary Bladder 4.4 1,810 3.2 Leukemia 6.1 203 2.9
Kidney 4.2 1,495 2.6 Myeloma 5.9 199 2.8
Brain 4.4 1,395 2.5 Non-Hodgkin Lymphoma 5.0 168 2.4
Female – White Female – Black
All sites 107.2 53,996 100 All sites 132.2 6,325 100
Lung 28.1 13,274 24.6 Lung 33.3 1,524 24.1
Breast 18.2 8,520 15.8 Breast 25.3 1,134 17.9
Colon and Rectum 10.7 6,236 11.5 Colon and Rectum 14.5 759 12.0
Ovary 5.8 2,747 5.1 Pancreas 7.8 401 6.3
Pancreas 4.8 2,706 5.0 Ovary 4.0 195 3.1
Non-Hodgkin Lymphoma 4.5 2,541 4.7 Myeloma 3.6 191 3.0
Leukemia 3.7 1,977 3.7 Corpus and Uterus, NOS 3.7 179 2.8
Corpus and Uterus, NOS 2.7 1,379 2.6 Leukemia 3.8 178 2.8
Brain 2.8 1,093 2.0 Stomach 3.0 173 2.7
Kidney 2.0 1,008 1.9 Cervix Uteri 3.2 152 2.4
# – Age-adjusted mortality rate (World Standard Population)
==== Refs
Wingo PA Ries LA Rosenberg HM Miller DS Edwards BK Cancer incidence and mortality, 1973-1995: a report card for the U.S Cancer 1998 82 1197 1207 9506368 10.1002/(SICI)1097-0142(19980315)82:6<1197::AID-CNCR26>3.0.CO;2-0
Ries LA Wingo PA Miller DS Howe HL Weir HK Rosenberg HM Vernon SW Cronin K Edwards BK The annual report to the nation on the status of cancer, 1973-1997, with a special section on colorectal cancer Cancer 2000 88 2398 2424 10820364 10.1002/(SICI)1097-0142(20000515)88:10<2398::AID-CNCR26>3.0.CO;2-I
Howe HL Wingo PA Thun MJ Ries LA Rosenberg HM Feigal EG Edwards BK Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends J Natl Cancer Inst 2001 93 824 842 11390532 10.1093/jnci/93.11.824
Edwards BK Howe HL Ries LA Thun MJ Rosenberg HM Yancik R Wingo PA Jemal A Feigal EG Annual report to the nation on the status of cancer, 1973-1999, featuring implications of age and aging on U.S. cancer burden Cancer 2002 94 2766 2792 12173348 10.1002/cncr.10593
Weir HK Thun MJ Hankey BF Ries LA Howe HL Wingo PA Jemal A Ward E Anderson RN Edwards BK Annual report to the nation on the status of cancer, 1975-2000, featuring the uses of surveillance data for cancer prevention and control J Natl Cancer Inst 2003 95 1276 1299 12953083
German RR Slemp C Keener B Jubelirer SJ Thompson GL Uhler RJ Long-term trends in cancer mortality rates for West Virginia W V Med J 1997 93 362 367 9123939
Wilkinson JD Wohler-Torres B Trapido E Fleming LE MacKinnon J Voti L Peace S Cancer trends among Hispanic men in South Florida, 1981-1998 Cancer 2002 94 1183 1190 11920490 10.1002/cncr.10321
Nasseri K Secular trends in cancer mortality, California 1970-1998 Cancer Detect Prev 2004 28 143 154 15068838 10.1016/j.cdp.2003.12.007
National Cancer Institute SEER*Stat Software Surveillance Research Program 2004 5.2.2 , National Cancer Institute
Doll R Cook P Summarizing indices for comparison of cancer incidence data Int J Cancer 1967 2 269 279 6041994
National Cancer Institute Joinpoint Regression Program Version 2.6 2002
Kim HJ Fay MP Feuer EJ Midthune DN Permutation tests for joinpoint regression with applications to cancer rates Statist Med 2000 19 335 351 10.1002/(SICI)1097-0258(20000215)19:3<335::AID-SIM336>3.0.CO;2-Z
American Cancer Society Ohio Cancer Facts & Figures 2003 2003 Dublin, OH, American Cancer Society, Ohio Division
American Cancer Society Ohio Cancer Facts & Figures for African Americans 2002 2002 Dublin, American Cancer Society, Ohio Division
Ward E Jemal A Cokkinides V Singh GK Cardinez C Ghafoor A Thun M Cancer disparities by race/ethnicity and socioeconomic status CA Cancer J Clin 2004 54 78 93 15061598
Gordon NH Socioeconomic factors and breast cancer in black and white Americans Cancer Metastasis Rev 2003 22 55 65 12716037 10.1023/A:1022212018158
Bach PB Schrag D Brawley OW Galaznik A Yakren S Begg CB Survival of blacks and whites after a cancer diagnosis Jama 2002 287 2106 2113 11966385 10.1001/jama.287.16.2106
Prothrow-Stith D Gibbs B Allen A Reducing health disparities: from theory to practice Cancer Epidemiol Biomarkers Prev 2003 12 256s 260s 12646523
Ohio Department of Health Health Insurance Coverage in Ohio, 1999-2001 2003 Columbus, OH, Ohio Department of Health
Freeman HP Poverty, culture, and social injustice: determinants of cancer disparities CA Cancer J Clin 2004 54 72 77 15061597
Ohio Department of Health Health Insurance Coverage in Ohio, Adults and Children, 1999-2001 2003 Columbus, OH, Ohio Department of Health
Tyczynski JE Berkel HJ Mortality from lung cancer and tobacco smoking in Ohio (U.S.): will increasing smoking prevalence reverse current decreases in mortality? Cancer Epidemiol Biomarkers Prev 2005 14 1182 1187 15894669 10.1158/1055-9965.EPI-04-0699
Underwood SM Reducing the burden of cancer borne by African Americans: if not now, when? Cancer Epidemiology Biomarkers Prev 2003 12(Suppl.) 270s 276s
|
16242021
|
PMC1277822
|
CC BY
|
2021-01-04 16:03:04
|
no
|
BMC Cancer. 2005 Oct 20; 5:136
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-136
|
oa_comm
|
==== Front
BMC Dev BiolBMC Developmental Biology1471-213XBioMed Central London 1471-213X-5-241624201910.1186/1471-213X-5-24Research Articlenfi-1 affects behavior and life-span in C. elegans but is not essential for DNA replication or survival Lazakovitch Elena [email protected] John M [email protected] Reiko [email protected] Keiko [email protected] Yuji [email protected] Richard M [email protected] Dept. of Biochemistry, SUNY at Buffalo, 140 Farber Hall, 3435 Main St., Buffalo, NY, 14214, USA2 Dept. of Biology, Canisius College, Buffalo, NY, USA3 CREST and Gene Network Lab, National Institute of Genetics, Mishima, Japan2005 20 10 2005 5 24 24 20 7 2005 20 10 2005 Copyright © 2005 Lazakovitch et al; licensee BioMed Central Ltd.2005Lazakovitch 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 Nuclear Factor I (one) (NFI) family of transcription/replication factors plays essential roles in mammalian gene expression and development and in adenovirus DNA replication. Because of its role in viral DNA replication NFI has long been suspected to function in host DNA synthesis. Determining the requirement for NFI proteins in mammalian DNA replication is complicated by the presence of 4 NFI genes in mice and humans. Loss of individual NFI genes in mice cause defects in brain, lung and tooth development, but the presence of 4 homologous NFI genes raises the issue of redundant roles for NFI genes in DNA replication. No NFI genes are present in bacteria, fungi or plants. However single NFI genes are present in several simple animals including Drosophila and C. elegans, making it possible to test for a requirement for NFI in multicellular eukaryotic DNA replication and development. Here we assess the functions of the single nfi-1 gene in C. elegans.
Results
C. elegans NFI protein (CeNFI) binds specifically to the same NFI-binding site recognized by vertebrate NFIs. nfi-1 encodes alternatively-spliced, maternally-inherited transcripts that are expressed at the single cell stage, during embryogenesis, and in adult muscles, neurons and gut cells. Worms lacking nfi-1 survive but have defects in movement, pharyngeal pumping and egg-laying and have a reduced life-span. Expression of the muscle gene Ce titin is decreased in nfi-1 mutant worms.
Conclusion
NFI gene function is not needed for survival in C. elegans and thus NFI is likely not essential for DNA replication in multi-cellular eukaryotes. The multiple defects in motility, egg-laying, pharyngeal pumping, and reduced lifespan indicate that NFI is important for these processes. Reduction in Ce titin expression could affect muscle function in multiple tissues. The phenotype of nfi-1 null worms indicates that NFI functions in multiple developmental and behavioral systems in C. elegans, likely regulating genes that function in motility, egg-laying, pharyngeal pumping and lifespan maintenance.
==== Body
Background
We are studying the role of the highly conserved Nuclear Factor I (NFI) family of site-specific DNA-binding proteins in metazoan development. NFI was first identified as a protein from human HeLa cells required for efficient adenovirus (Ad) DNA synthesis in vitro [1]. A binding site for NFI proteins in the Ad origin of replication is essential for viral replication both in vitro and in vivo [1-5]. These and other studies suggested that NFI proteins may function in the replication of host cell DNA [6,7]. However there is no direct evidence to support or refute a role for NFI proteins in host DNA synthesis. In contrast, many studies have identified NFI binding sites in the promoter and distal control regions of cellular genes, and deletion analysis of sites shows that NFI proteins are important for gene expression in a variety of cell types [8,9].
Four conserved NFI genes are present in vertebrates (NFIA, NFIB, NFIC and NFIX in humans; Nfia, Nfib, Nfic and Nfix in mice) [10-13]. Single NFI genes have been identified in the simple metazoans Amphioxus, C. elegans, and Drosophila [14,15], but no NFI genes are present in fungi, Arabidopsis or any sequenced prokaryotic genome. Thus the NFI gene family arose during early metazoan evolution and appears to be present only in multicellular animals. NFI proteins bind as either homo- or heterodimers [16,17] to the symmetric consensus sequence TTGGC(N5)GCCAA in duplex DNA [18,19]. NFI proteins also bind with lower affinity to sites containing a single TTGGC motif [20,21]. NFI homo- and heterodimers bind to the same sites with similar affinities, making it difficult to determine which family members play essential roles at specific cellular promoters. In addition, the 4 NFI genes in vertebrates are alternatively spliced [16,22] and are expressed in specific but widely overlapping patterns during embryogenesis and adult life [13] making it difficult to assess the role of specific NFI genes in development. The role of the NFI genes in development is of particular interest because binding sites for NFI proteins have been identified in genes expressed in virtually every tissue and organ system of vertebrates including brain [23,24], muscle [25] and other tissues. NFI proteins have also been implicated in the control of gene expression by a number of hormones and physiological modulators including glucocorticoids [26,27], insulin [28,29], TGFβ [30,31] and others.
To assess the roles of NFI genes in development we began a genetic analysis of the NFI genes in mice and C. elegans. Disruption of Nfia results in neurological defects including agenesis of the corpus callosum, loss of midline glial cells [32], hydrocephalus and perinatal lethality [33]. Disruption of Nfic causes defects in tooth development [34] while loss of Nfib results in perinatal lethality due to defects in lung maturation [35,36]. In each knockout defects are seen in the presence of the other three vertebrate NFI genes, suggesting that the 4 mouse NFI genes each have essential roles in development. However the presence of 4 NFI genes in mice has made it impossible to test whether NFI activity per se is essential for survival.
Since C. elegans has only one NFI gene (nfi-1), it provides an ideal system to assess the role of NFI in DNA replication and simple animal development. We show here that C. elegans nfi-1 and its products share many properties with their vertebrate homologs including similar DNA-binding activity, alternative splicing and expression during embryogenesis and in adult tissues. Loss of nfi-1 results in viable worms with multiple defects in locomotion, pharyngeal pumping and egg-laying, and a shortened life-span. Thus, nfi-1 is non-essential for survival but plays an important role in C. elegans physiology.
Results
Alternative splicing and promoter of the nfi-1 gene
The C. elegans nuclear factor I gene (nfi-1) was first identified by the C. elegans sequencing consortium by its homology to mammalian NFI genes [14]. Only a single NFI gene is present in the C. elegans genome (Fig. 1A) while vertebrates possess 4 NFI genes. To confirm the structure of the nfi-1 gene, we obtained cDNAs from the Kohara and TIGR libraries and made primers within predicted exons for production of cDNAs from RNA of adult worms and purified eggs. These cDNAs confirmed the presence of 14 exons and showed alternative splicing of exons 2 and 13 (Fig. 1A). The differential splicing pattern of nfi-1 indicates that different isoforms may be expressed in different cell types. Primary transcripts from most C. elegans genes are trans-spliced to SL1 leader sequences and detection of SL1-linked transcripts is frequently used to assess the initially transcribed exons of genes [37]. We confirmed the start site of the gene using PCR with an SL1 primer for SL1-containing transcripts and a primer in exon 2.
Figure 1 A) Alternatively spliced products of the nfi-1 gene. The nfi-1 gene is shown as a line with exons as numbered solid boxes and alternatively spliced exons as gray boxes. Arrows show the direction of transcription. Below the line are cDNAs from the Kohara and TIGR libraries; the vertical bar indicates the 5' end of the cDNA. In yK213C10 the letters a and b above exon 2 denote that it is alternatively spliced generating 2a and 2b and the arrow on the right denote undetermined sequence in the cDNA. The asterisk (*) on yk42f10 denotes an alternative 3'splice acceptor site in exon 13 used in yk42f10 but not in CEESQ09. Below are depictions of cDNAs obtained by RT-PCR from total and polyA+ RNA of whole worms or isolated eggs using nfi-1 exon-specific primers (RT-PCR nfi-1 primers). Lastly we show cDNA obtained by PCR using SL1 primer and nfi-1 exon-specific primers (RT-PCR SL1-linked products). Some of the alternatively spliced cDNAs have been described previously [14]. B) Comparison of CeNFI-DBD with the consensus mouse NFI DBD.The sequence of the CeNFI-DBD (top line) is aligned to a consensus sequence from the 4 mouse NFI-DBDs (bottom line). Gaps in the mouse consensus indicate residues that are not identical between the 4 mouse NFI-DBDs. The dash in the 6th aligned row indicates a single insertion in the CeNFI-DBD sequence needed to align it with the mouse consensus. Dark gray boxes show identical residues in CeNFI-DBD and mouse NFI-DBDs, light gray boxes show residues not identical but similar between CeNFI and mouse NFI-DBDs, unboxed residues with black letters show residues that are not similar between the CeNFI and mouse DBDs, and gray letters in the CeNFI-DBD above gaps in the mouse consensus indicate positions where the 4 mouse genes are not identical. 151 of 190 residues are identical in all 4 mouse NFI-DBDs while 60 of these 151 residues are different in the CeNFI-DBD and 27 of these 60 differences are non-conservative substitutions. The alignment was done in Macvector 6.5.3 using the ClustalW similarity matrix.
DNA binding by CeNFI is indistinguishable from that of human NFI-C
The predicted DNA-binding domain of nfi-1 (CeNFI-DBD) shares homology with the vertebrate NFI proteins, however 60 of 151 residues that are completely conserved among the 4 mouse NFI proteins are changed in CeNFI (Fig. 1B). To assess the DNA-binding activity and specificity of CeNFI, the DNA-binding domain (DBD) of CeNFI was cloned in frame with a 6 histidine-tag, expressed in E. coli, partially purified by nickel-affinity chromatography, and was used for in vitro DNA binding assays. The DNA-binding activity of CeNFI-H6 was indistinguishable from that of the human hNFI-C220H6, with both proteins binding the NFI-site oligo 2.6 (Fig. 2A) but not the C2 oligo containing a point mutation that prevents vertebrate NFI binding [19,38]. To ask if wild-type C.elegans contains a protein with similar DNA binding properties as CeNFI-H6, extracts were prepared from a mixed population of worms and tested for binding to the 2.6 and C2 sites. As expected, proteins were detected that bound to the 2.6 but not the C2 site, confirming the binding specificity of CeNFI (Fig. 2B). The specificity of the native and recombinant CeNFI proteins was also measured by competition of binding of the 2.6 oligo with various unlabeled oligonucleotides and was indistinguishable from the specificity of hNFI-C220-H6 (data not shown). Thus despite the differences between CeNFI and vertebrate NFIs in conserved residues of their DNA-binding domains, their DNA-binding activities are indistinguishable.
Figure 2 A) Specific NFI DNA-binding activity of recombinant CeNFI-H6. Partially purified recombinant H6-tagged CeNFI (containing parts of exons 2–8) and human NFI-C220 were incubated with a duplex oligonucleotide containing an NFI binding site (2.6, even lanes) or an oligo with a single point mutation that abolishes NFI binding (C2, odd lanes) and analyzed on a 6.5% non-denaturing polyacrylamide gel. Lanes 1 and 2, crude E. coli extract (neg. control); lanes 3 and 4, ~5 ng partially purified CeNFI-H6; lanes 5 and 6, ~40 ng partially purified CeNFI-H6; lanes 7 and 8, ~5 ng purified human NFI-C220H6. See bottom of panel for sequences of oligonucleotides. B) Specific NFI DNA-binding activity in worm extracts. Nuclear extracts of a mixed population of C. elegans were prepared and used in a gel mobility shift assay with an oligonucleotide that contains an NFI-binding site (lanes 1 & 2, 2.6) or the same oligo with a single point mutation that abolishes NFI binding (lane 3, C2). Lane 1, no extract; Lanes 2 and 3, C. elegans extract (~10 μg). See A for sequences of oligonucleotides.
Expression of nfi-1 in embryo and adults
In the mouse, the 4 NFI genes are expressed in a complex overlapping pattern during embryogenesis and in adult tissues [13]. To assess the expression pattern of nfi-1 transcripts in C. elegans, a digoxin-labeled antisense probe from the 3' end of the nfi-1 transcript was used for in-situ hybridization to fixed embryos and whole worms [39]. nfi-1 transcripts are present in the one-cell (not shown) and two-cell stages (Fig. 3a) prior to the onset of zygotic transcription [40,41], indicating that the transcript is maternally inherited. Expression continues in most cells of the embryo throughout early and mid-embryogenesis (Fig. 3b–d) but decreases after gastrulation and no expression is seen in L1 larvae (Fig. 3f–i and data not shown). As expected for a maternally inherited transcript, expression of nfi-1 reappears in the adult gonad (Fig. 3j–m). Expression of nfi-1 transcripts are also seen in the cytoplasm of gut cells (Fig. 3j–m). No signal is seen using control sense probes (Fig. 3n–o and data not shown). This in situ expression pattern indicates that nfi-1 could function both early in embryogenesis and in adult worms.
Figure 3 Expression of endogenous nfi-1 mRNA in embryo and adults. N2 worms were fixed and hybridized with digoxigenin (DIG)-labeled antisense nfi-1 probe (from plasmid yk42f10) and bound probe was detected using alkaline phosphatase-conjugated anti-DIG antibodies. Panels: a, 2 cell embryo; b, 4 cell embryo; c, 24 cell embryo; d, beginning gastrulation; e, mid-gastrulation; f, late gastrulation; g, comma stage; h, 1.5-fold stage and i, 2-fold stage. Embryos in a-d and f-i are positive for staining. We are currently investigating the apparent loss of signal at mid-gastrulation (panel e). Panels j-m, antisense probe, adults with exposed internal organs: mature oocytes; intestine; gonadal germ cells; n-o, control sense probe, no specific staining is seen. Right panels k, m, o are two-fold magnifications of those on the left.
Isolation of a null allele of nfi-1
To assess the role of nfi-1 in worms, we isolated a nfi-1 mutant using a reverse genetic approach based on the insertion-excision of the transposon Tc1 [42]. Worms carrying an ~2 kb deletion in nfi-1 were isolated by PCR screening and sib-selection (Fig. 4A,B). Sequence analysis of the nfi-1 transposon excision allele (designated nfi-1(qa524)) showed loss of the genomic region corresponding to nucleotides 14754–16715 of cosmid ZK1290. This eliminates the first 6 exons of nfi-1 including sequences encoding the DNA-binding domain. RT-PCR shows the absence of nfi-1 transcripts in mutant worms (Fig. 4C). A gel mobility shift assay, using nuclear extracts prepared from mix-stage populations of C. elegans wild type and nfi-1 mutant worms confirmed the loss of CeNFI DNA-binding activity in the mutant worms (Fig 4D). Thus, this mutation in the nfi-1 gene is a null allele. The survival of worms homozygous for the null allele shows that nfi-1 is not essential for worm survival. The nfi-1 mutant allele was backcrossed 12 times to the wild-type N2 strain to remove unwanted mutations prior to assessment of the phenotype.
Figure 4 A) Deletion in C. elegans nfi-1 gene. The relative locations of confirmed exons (boxes) are shown. Square brackets indicate the location of the NFI-1 DNA-binding domain and the region deleted in nfi-1 mutant. The arrows indicate the position of the Tc1 insertion in an intron of the nfi-1 gene in strain NL747 pk240 and locations of PCR primers used in the screening for the deletion. B) Single-worm PCR reactions on N2 worm and nfi-1 homozygous mutant isolated by sib-selection. The arrows indicate a 1007 bp and 2969 bp PCR products corresponding to the nfi-1 mutant (qa524) and wild type alleles respectively. Lane 2 is 1 kb DNA ladder. C) RT-PCR on N2 worms and nfi-1 homozygous mutants. The arrows indicate a 480 bp and 320 bp RT-PCR products amplified using total RNA obtained from N2 worms (lane 1). nfi-1 mutants show loss of nfi-1 transcripts (lane 2). Lane (3) is 100 bp DNA ladder. D) Loss of NFI DNA-binding activity in extracts of nfi-1 mutants. Nuclear extracts of a mixed population of N2 worms and nfi-1 mutants were prepared and used in a gel mobility shift assay with an oligonucleotide (2.6) that contains an NFI-binding site (lanes 3, 5) or the same oligo with a single point mutation that abolishes NFI binding (C2) (lane 4, 6). See Fig. 2A for sequences of oligonucleotides. Extract of nfi-1 mutants show loss of NFI DNA-binding activity (lanes 5, 6). Lanes 1 & 2, no extract.
Phenotype of nfi-1 mutant worms
Locomotion defect
Loss of nfi-1 results in a body movement defect (Unc, uncoordinated). nfi-1 mutant animals are fairly active and healthy, but are sluggish and flaccid at rest and slightly longer and thinner than N2 worms (Fig. 5A,D). While wild type worms usually move in straight long lines, nfi-1 mutant worms often change direction abruptly (Fig. 5B,E). The nfi-1mutant worms produce less regular tracks on the bacterial lawn with higher amplitude, and sometimes are slightly coiled when compared to wild-type worms (Fig. 5C,F). This phenotype is more severe in older adults.
Figure 5 Locomotion in nfi-1 mutants. Single young adults were spotted in the center of fresh plates and left for 10 min. (A, D) Photographs of N2 worms and nfi-1 mutants; (B, E) Track patterns of N2 worms and nfi-1 mutants; (C, F) Track patterns of N2 worms and nfi-1 mutants with higher magnification. Note less regular tracks in nfi-1 mutant vs. N2 worms.
Egg-laying defect (Egl)
17–35% of older nfi-1 mutants have a "bag of worms" phenotype, where the mother is unable to lay fertilized eggs and fills with hatched progeny (Fig. 6A). Appearance and severity of this egg-laying phenotype correlates with the progressive locomotion defect, as young adult nfi-1 mutant worms do not bag. In young adults, serotonin stimulated egg-laying in both the wild type and nfi-1 mutant animals to a similar extent (data not shown), indicating that the postsynaptic response to serotonin is normal and the contractile apparatus for egg-laying is intact in nfi-1 mutant worms.
Figure 6 A) Egg-laying defect in nfi-1 mutants and transgenic rescue. Bagging was measured in wild-type N2, nfi-1, N2 worms carrying the transgenic array qaEx507(N2 qaEx507) and nfi-1 worms carrying this array (nfi-1 qaEx507). Bars represent % of bagging as the mean of 3–4 independent experiments and error bars show the standard deviation. 30–75 worms of each genotype were scored in each independent experiment. N2 and N2 qaEx507showed <5% bagging. The nfi-1 mutant worms showed ~30% bagging while the rescued nfi-1 qaEx507showed <5% bagging. B) Shortened life span in nfi-1 mutants and transgenic rescue. Survival curves for the strains described above N2 (n = 57), nfi-1 (n = 58), N2 qaEx507 (n = 52) and nfi-1 qaEx507 (n = 31) are shown. Kaplan-Meier analysis (SPSS11 software) was use to determine median, percentile and p values (log rank test) and Excel was used to construct survival curves. The array generated ~50% rescue of the life-span. The experiment was repeated twice with similar results.
Life-span reduction
The nfi-1 mutant worms have a median life span of 10.00 ± 0.69 days as compared to 14.00 ± 0.76 days for wild type worms (p < 0.001) (Fig. 6B). Mutant worms become progressively more sluggish and flaccid as they age. We are currently investigating whether this lifespan reduction is due to the apparent progressive muscle weakness and partial paralysis seen or to a direct effect on known ageing pathways [43-45].
Pharyngeal pumping rate defect
Since the Unc and Egl phenotypes of nfi-1 mutant worms could reflect aberrant muscle function, we examined another process that can influenced by muscle defects, pharyngeal pumping rate. Pharyngeal pumping rates are reduced in nfi-1 mutant worms (Table 1, Days 1–4), with more severe reductions in older vs. younger adult animals (Table 1, Days 3&4 vs. Day 1).
Table 1 Rescue of pharyngeal pumping defect by nfi-1 transgene
Daya N2 nfi-1 N2 qaEx507b nfi-1 qaEx507b
1 (n = 10) 241 ± 5 228 ± 9c 243 ± 5 241 ± 8e
2 (n = 12) 242 ± 10 187 ± 64d 237 ± 8 195 ± 78d
3 (n = 12) 214 ± 18 125 ± 79c 183 ± 78 200 ± 72
4 (n = 12) 160 ± 62 92.5 ± 86d 183 ± 79 133 ± 90
a Pharyngeal pumping was counted for one minute starting on day one, when animals first reached adulthood, for four consecutive days.
b N2 and nfi-1 mutant worms were transformed with a transgene containing the nfi-1 gene and upstream promoter and rol-6 (qaEx507).
c This value is statistically significantly different from N2 (P < 0.01)
d This value is statistically significantly different from N2 (P < 0.05)
e This value is statistically significantly different from nfi-1 (P < 0.01)
Rescue of nfi-1 mutant with nfi-1 transgene
Transgenic rescue was performed to test whether loss of the nfi-1 gene was responsible for the observed phenotypes. Transgenic strain XA512 qaEx507 was made by injecting a plasmid containing a 10 kb region of genomic DNA including the nfi-1 coding region and ~4 kb of upstream promoter region together with a rol-6(gf) expressing plasmid into the gonads of N2 worms. We crossed the resulting transgenic array into nfi-1 mutant worms to produce strain XA550 nfi-1(qa524) qaEx507. Egg-laying was completely rescued in nfi-1 qaEx507 worms when compared to the nfi-1 mutant worms (Fig. 6A). In addition, the median life span in nfi-1 qaEx507 worms of 13.0 ± 0.7 days was significantly longer than the 10.0 ± 0.7 day life span of nfi-1 mutant worms (p < 0.05), but slightly less than the 14.0 ± 0.8 day life span of N2 worms (Fig. 6B). The N2 qaEx507 stain used as a control has a 14.00 ± 0.49 day median life span, identical to that in non-transgenic N2 worms. The pharyngeal pumping defect was also partially rescued by transgenic expression of nfi-1 (Table 1, nfi-1 qaEx507 vs. nfi-1). The nfi-1 transgene had little or no effect on pumping rates in N2 worms (N2 qaEx507 vs. N2). These data provide a well-defined developmental system affected by loss of nfi-1 that can be examined for cell-autonomous or inductive roles of nfi-1. The presence of the rol-6 marker gene prevented scoring of rescue of the locomotion phenotype in nfi-1 qaEx507 worms.
Our in situ hybridization data indicate that nfi-1 transcripts are provided maternally. To test whether maternal nfi-1 transcripts could rescue the nfi-1 Egl phenotype, nfi-1 qa524/qa524, nfi-1 qa524/+ and +/+ progeny of heterozygous nfi-1 qa524/+ parents were scored for the egg-laying defect (Fig. 7). Bagging was seen in 41% of the resulting nfi-1 qa524/ qa524 worms, in 20.7% of nfi-1 qa524/+ worms but in <3% of +/+ animals. These data indicate the absence of maternal transcript rescue and possible haploinsufficiency at the nfi-1 locus. Thus, transgenic replacement of nfi-1 yields either partial (pumping rate and lifespan) or complete (egg-laying) rescue of the phenotypes seen in the nfi-1 mutant worms whereas maternal nfi-1 transcripts are insufficient to rescue the egg-laying defect.
Figure 7 Haploinsufficiency of nfi-1 locus. Egg-laying defect in progeny of nfi-1 heterozygous mutant animals are shown. Bagging was scored in all progeny of two nfi-1(qa524/+) heterozygous worms (n = 146) derived from eggs laid over 7 hours. All worms were genotyped by single-worm PCR. Bars represent % of bagging in wild type (+/+), nfi-1 heterozygous (qa524/+) and nfi-1 homozygous worms (qa524/qa524). The number of worms scored of each genotype are shown above the bars.
Ce titin expression is reduced in nfi-1 mutant worms
We used cDNA microarrays to identify genes whose expression is affected by loss of nfi-1. Such genes could be either direct or indirect targets of nfi-1. Poly A+ RNA was purified from wild type and nfi-1 mutant synchronized gravid adults, labeled, and used to probe DNA microarrays containing ~17,000 C. elegans genes (Stanford Microarray Database). We analyzed RNA from gravid adults because the phenotype differences between nfi-1 mutant and wild type animals are clearer in adults than at earlier stages. The nfi-1 gene was scored as the most down-regulated gene in nfi-1 mutant worms in all experiments (data not shown). Several dozen genes showed small apparent reductions or increases in levels (2–3 fold) in mutant adults (data not shown) but only one gene, C. elegans titin (Ce titin) showed larger changes.
Ce titin (also known as tag-58, temporarily assigned gene 58) was predicted to be 5.7-fold lower in mutant worms by microarray analysis. Quantitative PCR confirmed that Ce titin is reduced 8–11 fold in adult nfi-1mutant worms (Fig. 8). To date this is the gene that shows the largest decrease in expression in nfi-1 mutant worms. A search of the Ce titin gene reveals no overabundance of NFI binding sites (data not shown). Also, a CeTPro transgene expressing a translational fusion of GFP to the 5'-end of the Ce titin gene [46] appears to be expressed at similar levels in WT and nfi-1 mutant worms (data not shown). It will be important in future studies to determine whether Ce titin is a direct or indirect target of nfi-1 and the possible role of Ce titin in the phenotypes observed.
Figure 8 Down regulation of Ce titin expression in nfi-1 mutants assessed by QPCR. Bars represent fold changes in Ce titin transcript level in wild type N2 vs. nfi-1 mutant worms. RNA samples were obtained from 3 independent synchronized adult worm populations for each genotype.
Expression of nfi-1::GFP reporter transgenes
One limitation of in situ hybridization in C. elegans is that in older embryos and postembryonic stages it is sometimes not sensitive enough to unambiguously identify individual cells. Since we saw little or no nfi-1 expression in muscle by in situ hybridization, in an effort to develop a more sensitive assay for nfi-1 expression we constructed two GFP-reporter transgenes (Fig. 9A). In Pro1CeNFI::GFP, 4 kb of genomic DNA sequence upstream of the nfi-1 open reading frame and the sequence encoding the first four residues of the CeNFI protein was fused in frame to GFP. In Pro2CeNFI::GFP, the 4 kb promoter region and the sequence encoding the first 94 residues of CeNFI has been fused to GFP. GFP expression for both transgenes was detected in embryos (Fig. 9B). Faint GFP expression is first detected at the late gastrulation stage of embryogenesis (>300 cells) as a diffuse green glow throughout the embryo. By the comma stage expression is detected in many cells along the outer edge of the embryo and expression continues through embryogenesis and is detected in L1-L4 larvae in many of the same cells as in adults. Adult transgenic animals show GFP expression in muscles, neurons and intestinal cells (Fig. 9B–I). Among the muscles, fluorescence was strongest in the pharynx and head muscles, was observed with less frequency in other body wall muscles and was seen occasionally in vulva muscles. Expression was also seen in two pairs of neurons located near the posterior bulb of the pharynx, and in several as yet unidentified tail neurons. Expression patterns in multiple transgenic lines from each reporter were similar with the exception that Pro2CeNFI::GFP expression was detected more consistently in head neurons and Pro1CeNFI::GFP was seen with higher frequency in body-wall muscles. However expression of both transgenes was mosaic, showing expression in only subsets of cells and animals in each population. Since mosaic expression of GFP was seen in transgenic strains from both arrays, transgenic array Pro1CeNFI::GFP was integrated by γ-irradiation. However similar mosaic expression was seen with the integrated array (data not shown). These data may indicate that additional elements are needed for stable regulation of nfi-1 expression and that such elements may be located further downstream in the nfi-1 genomic sequence.
Figure 9 Expression pattern of the nfi-1::GFP reporter transgenes. The nfi-1 locus and structure of the nfi-1-GFP fusion constructs are shown (A). Nfi-1 coding regions are shown in black, gray boxes indicate alternatively spliced exons, untranslated regions are in white. GFP is shown as a hatched box. Expression of nfi-1-GFP reporter constructs was observed in embryos (B), intestinal cells (C), body wall muscles (D, E), pharynx (F), egg-laying muscles (G), several head (H) and tail neurons (I). Expression was assessed using a FXA Nikon microscope (B-D, F-I) and a Bio-Rad confocal microscope (E).
Discussion
nfi-1 gene structure and DNA binding properties
These data show that nfi-1 is not essential for embryogenesis and thus is not essential for DNA replication. However the multiple defects seen in nfi-1-deficient worms, abnormal locomotion, pharyngeal pumping and egg-laying defects, and a reduced lifespan indicate that nfi-1 is essential for these developmental and behavioral processes. nfi-1 shares a number of properties with the vertebrate NFI genes, including alternative splicing (Fig. 1) and the DNA-binding properties of its protein product CeNFI (Fig. 2B,C). The alternative splicing seen in nfi-1 is reminiscent of the complex splicing pattern seen in the vertebrate NFI genes [22]. While the relevance of this alternative splicing in C. elegans is unclear, the finding that alternatively spliced forms of vertebrate NFIs have different transcriptional modulation properties [16,47,48] suggests that the multiple C. elegans isoforms may also have distinct functions. It will be of particular interest to determine whether alternatively spliced isoforms of nfi-1 are differentially expressed in worm cells during development and in adults. In addition, when specific downstream targets of nfi-1 are identified it will be important to determine whether the isoforms differ in their ability to modulate gene expression in C. elegans.
The sequence conservation of the predicted DNA-binding domain of nfi-1 with the vertebrate NFI proteins was the initial indication that nfi-1 encodes the single C. elegans NFI gene. However among the 4 mouse (and human) NFI genes, 151 of 190 residues of their DNA-binding domains are completely conserved, while 60 of these 151 residues are changed in CeNFI (Fig. 1B). We show that CeNFI binds to the same DNA sequence as hNFI-C and that binding is abolished by a point mutation known to abolish binding by products of the 4 vertebrate NFI genes (Fig. 2) [38]. This raises the question of why there has been so little divergence of the 4 vertebrate NFI DNA-binding domains during evolution. While we have shown similar DNA binding specificities of CeNFI and hNFI-C to a matched set of WT and mutant NFI binding sites, it is possible that subtle differences exist in their DNA-binding specificity or affinity or that the conserved vertebrate residues are more important for binding within the more complex vertebrate genome. Since the 4 vertebrate NFI genes are sometimes expressed in the same cells in vivo [13], it is possible that there has been selection for conservation of residues involved in DNA binding due to competition between the 4 NFI gene products, and that this has led to the observed high degree of sequence conservation. Similar high levels of sequence conservation are seen in the DNA-binding domains of all the homologous vertebrate NFI genes from Xenopus [49,50] to humans [12,51].
It is possible that the additional conserved residues in the vertebrate NFI DBDs compared to nfi-1 confer additional functions to the vertebrate NFI proteins, such as interacting with proteins required for transcriptional modulation specifically in vertebrates, or for the stimulation of Ad DNA replication. Indeed, 2 sets of mutations in the NFI-C DBD that abolish Ad replication without affecting DNA-binding or dimerization are in residues not conserved between NFI-C and CeNFI [52]. However, each of the 4 cysteine residues shown previously to be essential for DNA-binding activity and redox-regulation of DNA-binding activity of vertebrate NFIs are conserved in CeNFI (Fig. 1B, labeled C) [53,54]. The conservation of these cysteine residues is consistent with our observation that the DNA-binding activity of CeNFI is sensitive to reversible inactivation by chemical oxidizing agents such as diamide in vitro (data not shown), as is the activity of the vertebrate NFIs [53,54]. Given the divergence of the CeNFI and vertebrate NFI DBDs it will be important to determine whether the nfi-1 DBD can substitute for the vertebrate NFI DBDs in vivo, and whether the vertebrate NFI DBDs possess equivalent activities in C. elegans in vivo.
Defects in nfi-1 mutant worms
Deletion of the nfi-1 gene results in a number of behavioral defects, each of which may be related to defects in muscle or neural function. The C. elegans hermaphrodite has four major muscles types, body wall, pharyngeal, vulval and enteric [55]. Body-wall muscles and pharyngeal muscles, used in locomotion and pumping food, respectively, function in the processes affected in nfi-1 mutants. The structures most often affected in Unc mutants are the body-wall muscles, the nervous system and the hypodermis [56]. Polarized light microscopy showed no obvious deformities in the body wall-muscles of nfi-1 mutant animals, such as defects in muscle attachment and overall sarcomere structure. Thus, the muscles appear normal at a gross level. Likewise, we analyzed the expression pattern of the body wall muscle specific hlh-1::GFP [57] and muscle specific CeTPro transgenes [46] in the nfi-1 deletion mutant background by confocal microscopy and saw no differences between mutant and wild type animals (data not shown). Thus any muscle defects are not due to gross structural changes within the muscle. Many older nfi-1 deficient animals display egg-laying defects, suggesting that vulval or/and uterine muscles or their controlling neurons may be affected. The normal stimulation of egg-laying by serotonin suggests that there is no major loss of vulval or uterine muscle function in young adult nfi-1 deficient worms [58,59]. We have visually examined the functions of the enteric muscles involved in food movement through the gut and defecation and found no obvious defects in young adults. In older adults, contractions of the enteric muscles were less regular, but it is unclear whether this is an intrinsic defect or related to the defects in pharyngeal pumping seen in older adults.
The absence of obvious muscle defects in nfi-1 null worms raises the possibility that neural defects could underlie the phenotypes seen in these animals. For example, nfi-1 mutants show a reduced pharyngeal pumping rate, a process regulated primarily by the M3, M4 and MC pharyngeal neurons [60]. Some or all of defects seen in nfi-1 mutant worms, mild Unc with flaccidity, pharyngeal pumping and egg-laying are also seen in worms containing single mutations in genes expressed in neurons. For example, some mutations in egl-30, a heterotrimeric Gqα subunit expressed in muscles and neurons, have weak Unc, egg-laying and pharyngeal defects [61]. In addition, some loss of function mutations in the G protein signaling molecule egl-10 and the Gβ5 ortholog eat-11 cause sluggish movement, egg-laying, and pharyngeal pumping defects [62,63]. Since a number of eat genes are expressed specifically in neurons, these data indicate that neural dysfunction caused by loss of nfi-1 could generate this range of phenotypes. It will be important in future studies to determine whether nfi-1 expression is needed neurons, muscles, both cell types or other cell types to alleviate the phenotypes seen in nfi-1 mutant worms. While our preliminary QPCR data indicate no changes in the transcript levels of egl-30, egl-10 or eat-11 (N. Butz, unpublished data), it is possible that the activity of the G protein signaling system is affected in nfi-1 mutant worms. Thus, it would be useful in future studies to directly test whether nfi-1 is involved in G protein signaling in C. elegans.
The only gene whose expression has been shown to change in response to loss of nfi-1 is the muscle-specific gene Ce titin. This reduction in a muscle-specific gene indicates that muscle defects could contribute to the observed locomotion and other phenotypes. Ce titin is a massive protein, with isoforms in C. elegans of 2.2MDa, 1.2MDa and 301KDa [46]. Ce titin is found in the I-bands of larval and adult muscle in worms. Mutations in mouse titin genes cause muscular dystrophy [64], cardiac development defects and muscle weakness [65]. Thus, down regulation of Ce titin expression could contribute to the motility and egg-laying defects observed in nfi-1 mutants. However, the large size of the Ce titin gene has made it difficult to generate transgenic lines that conditionally express Ce titin. Recent studies have reported that neither existing mutations in Ce titin, nor RNAi experiments, have provided clues as to the function of Ce titin in worms [46]. Thus it will be important in the future to test directly the role of Ce titin in the defects seen in nfi-1 mutant worms.
Finally, it is unknown whether the observed motility, egg-laying and pharyngeal pumping defects contribute to the shortened life-span of nfi-1 null worms. For example, as pharyngeal pumping becomes more defective the animals could starve, leading to progressive muscle wasting, paralysis and death. However, most severe eat mutants have extended lifespans rather than reduced lifespans, most likely due to caloric restriction [66]. These data indicate that the reduced lifespan seen in nfi-1 deficient worms could be independent from the pharyngeal pumping defect. In addition, while bagging could reduce apparent lifespan, we eliminated bagged animals from our lifespan analysis so as to exclude bagging as a direct cause of the shortened lifespan. Lastly, the motility defect is seen in young adults a few days prior to bagging and is seen in all adults, even those that do not bag. Thus it appears unlikely that bagging contributes directly to the movement defect. It will be important to use directed expression of nfi-1 in different cell types to test the dependence or independence of the observed phenotypes from each other. We are also currently testing whether nfi-1 is part of the genetic pathways known to be important in worm aging including the daf pathway [67].
Roles of NFI transcription factors in development
Since NFI genes have been found in all metazoa sequenced to date and multiple NFI genes are present in complex animals [68], it was our initial hypothesis that nfi-1 would be essential in worms and that deletion of the gene would be lethal. Lethality would be predicted if nfi-1 has an essential role in DNA replication. However, all of the defects observed are late physiological defects and are most severe in older adults. This set of late defects can be compared to the late gestational defects seen with deletion of single NFI genes in mice. Disruption of the mouse Nfia gene causes agenesis of the corpus callosum, the loss of specific midline glial populations and perinatal lethality [32,33]. Loss of Nfic produces specific defects in tooth development including aberrant incisor formation and failure of root formation in molar teeth [34]. Targeted insertion into the Nfib locus causes perinatal lethality due to a failure of late fetal lung maturation [35,36]. One common feature of the defects seen in NFI-deficient mice is that they occur either late in fetal development (Nfia and Nfib), or early in postnatal development (Nfic). While some of the developmental systems affected by the loss of NFI genes in mice are not present in worms (e.g. lungs and teeth), it is possible that the underlying molecular mechanisms disrupted in NFI-deficient mice are also affected by loss of nfi-1 in worms. Thus it will be important to test genes identified as important in the phenotypes of NFI-deficient mice for potential roles in the motility, egg-laying and pharyngeal-pumping defects seen in nfi-1mutant worms.
Conclusion
These data show that nfi-1 is not essential for worm survival but plays important roles in locomotion, egg-laying, pharyngeal pumping and maintenance of a normal life-span. These are the first data to show that while NFI proteins are essential for adenovirus DNA replication, they are not essential for DNA replication in simple animals.
We show that while many residues of the DNA binding domain of worm NFI differ from those found in vertebrate NFIs, the DNA binding activity and specificity of the worm protein is indistinguishable from that of the vertebrate NFIs. These data suggest that the very strong conservation of many residues in the DNA-binding domains of vertebrate NFIs is not needed for DNA-binding specificity but may serve another function, perhaps for interactions with specific vertebrate proteins. In addition, like the vertebrate NFI genes nfi-1 is alternatively spliced, raising the possibility that different CeNFI isoforms may have different biological functions.
The phenotypes of nfi-1 null worms are rescued by zygotic expression of nfi-1 from its natural promoter but not by maternal transcripts. We propose that nfi-1 regulates gene expression directly in cells in which it is expressed and affect the function of these cells. However, the precise cell types in which loss of nfi-1 causes these defects in motility, egg-laying, pharyngeal pumping and lifespan is unknown. Directed expression of nfi-1 in different cells of the worm should allow us to determine in which cells nfi-1 is needed for normal lifespan, motility, egg-laying and pharyngeal pumping.
Future studies will assess whether C. elegans can be used as an experimental system to assess the potentially distinct physiological functions of the vertebrate NFI gene products. For example, if the 4 vertebrate NFI genes have segregated the functions of the single worm nfi-1 gene into 4 distinct entities, then each vertebrate NFI gene might have distinct properties when expressed in nfi-1mutant worms. It will also be important to determine the specific nfi-1 target genes in C. elegans whose altered expression is responsible for the phenotypes observed, the specific cell types in which nfi-1 must be expressed to prevent each phenotype, and to use C. elegans genetics to identify genes essential for NFI-regulated transcription.
Methods
Strains and transgene plasmids
The Bristol strain N2 was used as wild type and worms were grown at 20°C using standard techniques [69]. Strains containing hlh-1:GFP, CeTPro and NL747 pk240 were kindly provided by Michael Krause (NIH/NIDDK/LMB, Bethesda, MD), Guy M. Benian (Emory University School of Medicine, Atlanta, GA) and Ronald Plasterk (Hubrecht Laboratory, Utrecht, The Netherlands), respectively.
For rescue experiments the nfi-1 locus from cosmid ZK1290 was cloned into pBluescriptKS+ to generate pCeNFIG. pCeNFIG (25 ng/ul) was injected in N2 wild type worms along with the rol-6(gf) marker plasmid pRF4 (125 ng/ul) using standard microinjection procedures to produce the transgenic strain XA512 qaEx507 [69]. Transgenic arrays were crossed onto the nfi-1 deletion strain (XA549 nfi-1(qa524), see below) screening F1 and F2 progeny by PCR for the wild-type and nfi-1(qa524) deletion alleles.
Two GFP-reporter constructs were made to assess nfi-1 expression. Pro1CeNFI::GFP was made by cloning a PCR-generated HindIII-XmaI fragment of ZK1290 into pPD95.69 (kindly provided by A. Fire) to generate a translational fusion of nfi-1 and GFP. The 5'end of the construct was extended by cloning the HindIII-BglII fragment of pCeNFIG into the vector. Pro2CeNFI::GFP was made by cloning a HindIII-PstI fragment of pCeNFIG into corresponding sites of pPD95.69. The constructs were coinjected with rol-6 marker into N2 worms as described above. Three independent transgenic lines were analyzed for Pro1CeNFI::GFP expression and two lines for Pro2CeNFI::GFP expression.
A vector expressing the 6his-tagged DNA-binding domain of CeNFI (pCeNFI-H6) in E. coli was produced by cloning a fragment of nfi-1 cDNA encompassing the predicted DNA-binding domain of nfi-1 downstream of a 6 histidine tag into the pET8C vector [70].
Generation of the nfi-1 deletion mutant
C. elegans NL747 pk240 contains a Tc1 transposon insertion in the 6 th intron of nfi-1. The location of the Tc1 element was mapped by genomic PCR and sequencing. 100 populations of worms were established at 15°C on 3 cm NGM/OP-50 plates starting with 10–20 worms each, and the worms were collected in M9 buffer when nearing starvation. One third of the worms were put onto fresh plates and the rest were used to prepare two separate lysates for DNA. Thirteen of the 100 populations showed a strong excision bands by PCR screening (Platinum Taq Polymerase, Invitrogen) with primers 1 (5'GTATTTGTACGACCCTCTGCG) and 2 (5'TGCTGTTGAACGGAATGCACC). To distinguish between germ line and somatic mutations the second lysate for each positive population was screened by PCR. Only two populations showed deletions with both lysates indicating that more than one worm carried the deletion and therefore the deletions were in the germ line. The locations of the deletions were determined by subcloning PCR products into pCRII-TOPO (Invitrogen) and sequencing. A clonal strain of worms carrying one of these deletions was isolated by multiple rounds of PCR screening and sib-selection (strain XA549 nfi-1(qa524)). The mutant worms were backcrossed 12 times onto the wild-type N2 strain to remove unwanted mutations. For genotyping, PCR with primers 1 and 3 (5'TCGGAGGAGGTGGTAGACAT) amplified a 623 bp product from the deletion allele and primers 1 and 4 (5'GTGAGTCTTGAGGTGCTTCTG) amplified a 968 bp product from the wild type allele.
RT-PCR cloning and cDNA isolation
Total RNA was prepared from adult worms or eggs (Trizol, Life Technologies) and was reverse transcribed using oligo dT (Superscript, Life Technologies) according to the manufacturers instructions. Poly A+ mRNA was isolated from total RNA using standard techniques (Oligotex, Qiagen). PCR was performed as described previously [14] using primers from within predicted exons 1–14 of nfi-1 or SL1 primers (GGTTTAATTACCCAAGTTTGAG) (nfi-1 primer sequences available upon request) and PCR products were cloned into the pCRII-TOPO vector (Invitrogen). DNA from individual clones was isolated (Wizard, Promega) and sequenced (Roswell Park Cancer Institute DNA Sequencing Core) and the sequence obtained compared with that of the predicted exons of nfi-1 (C. elegans cosmid ZK1290, Genbank Acc. #U21308). The cDNA clones yk42f10 and yk213C10 were from the C. elegans cDNA sequencing/expression project (CREST and Gene Network Lab, National Institute of Genetics, Mishima) and CEESQ09 was from The Institute for Genomic Research (TIGR).
In situ hybridization
Endogenous nfi-1 transcripts were analyzed as described previously [39], using digoxin-labeled antisense probe for nfi-1. Embryos were either freeze-fractured or were treated with chitinase to digest the eggshell prior to hybridization.
DNA-binding assays
Electrophoretic mobility shift assays for the analysis of NFI protein binding to oligonucleotides was performed as described previously [38,54]. Worm extracts were prepared from dounce homogenized mixed-age worms using the NP40-based extraction buffer described previously [38]. CeNFI-H6 (332aa) and hNFI-C220-H6 (230aa) proteins were produced and partially purified from extracts of E. coli as described previously [54]. The labeled oligonucleotides used contained a wild-type NFI binding site (2.6) or a site with a single point mutation (C2) shown previously to abolish the binding of vertebrate NFI proteins [19,38,54,71].
Behavioral and functional assays
For locomotion assays wild type N2 and nfi-1 mutants were raised at 20°C on NGM/OP50 plates. To obtain age-synchronized worms young adults were placed on a bacterial lawn to allow them to lay eggs for 3 hours and then removed. Worms were observed at all larval stages. To photograph worms tracks, single young adult worms were moved to new plates containing a 1 day old bacterial lawn and were left undisturbed for 10 min.
"Bag of worms" phenotype was scored for each strain using age-synchronized worms. Worms were observed each day from the start of egg laying until one day after cessation of egg laying. They were moved to new plates every second day to prevent overcrowding and starvation. Other aspects of egg-laying behavior such as the brood size, stage of newly laid eggs, and response to serotonin were assessed as described previously [58,59]. Brood size was the same in WT and nfi-1 null worms.
For life-span assays wild type N2 and nfi-1 mutants were raised and synchronized as described above except that 250 μg/ml of fungizone was included to reduce fungal contamination. Worms were maintained at 20°C and adult worms were moved to new plates every second day until progeny production ceased. Worms were observed every day and were scored as dead when they failed to respond to touch. Animals that crawled off the plate or died from internally hatched progeny were censored, but incorporated into the data until the day of disqualification.
Pharyngeal pumping was counted for one minute starting on the first day of adulthood for four consecutive days [60]. Worms were placed on NGM/OP50 plates and left undisturbed for 1 hour before measuring. All animals remained on food during the period of observation.
Microarray assays and QPCR
Synchronized populations were generated by hatching eggs after alkaline hypochlorite-treatment and the worms were collected as gravid adults. Total RNA was prepared with Trizol (Life Techologies) and poly A RNA was purified using Poly(A)Purist™ kit (Ambion). DNA microarray assays were performed and analyzed in the Stanford Microarray Database (Stanford). cDNA for QPCR was made using random primers and Super-Script (Invitrogen). QPCR was performed using AmpliTaq GoldR with the Gene AmpR SYBR green kit (Applied Biosystems) on a Bio-Rad iCycler. The inf-1 gene was used as an internal control for RT-PCR reactions and for normalized quantification in QPCR reactions. Primers used in QPCR are available on request.
Authors' contributions
EL isolated the nfi-1 deletion strain and most transgene plasmids, performed most transgenic generation and phenotype characterization, and helped draft the manuscript. JMK performed microinjections for transgenic rescue and some phenotype characterization and contributed to manuscript preparation. RM performed QPCR. KH performed in situ hybridizations and β gal staining. YK identified specific cell types in which nfi-1 is expressed. RMG initiated the project and designed some of the experiments, generated some transgenics, produced vectors and proteins used in the study, and finalized preparation of the manuscript.
Acknowledgements
The authors thank Ann Billetz and Emily Harville for initial studies on the nfi-1 gene. We also thank the Plasterk lab for isolation of the Tc1 insertion in the nfi-1 gene, the Kim lab and the Stanford Microarray Database facility for cDNA array screening, Drs. Michael Krause (NIH/NIDDK/LMB, Bethesda, MD), Wayne Materi and Dave Pilgrim (Univ. of Alberta, Edmonton, Alberta, CA), Jim McGhee (Univ. of Calgary, Calgary, Alberta, CA) and the C. elegans Genetic Center (Univ. of Minnesota, St. Paul, MN) for transgenic strains, the C. elegans Genomic Sequencing Consortium (Sanger Center, Cambridge UK for cosmid ZK1290 and TIGR for cDNA clones. Drs. Phil Morgan and Marge Sedensky (Case Western Reserve Univ., Cleveland, OH) and their lab members were instrumental in helping RMG. learn C. elegans biology. The authors also thank Drs. Christine Campbell (UB) and Paul Mains (U. Calgary) for helpful discussions and reading the manuscript. These studies were supported in part by the National Institutes of Health grant HD34908 to RMG.
==== Refs
Nagata K Guggenheimer RA Hurwitz J Specific binding of a cellular DNA replication protein to the origin of replication of adenovirus DNA Proc Natl Acad Sci USA 1983 80 6177 6181 6336326
Nagata K Guggenheimer RA Hurwitz J Adenovirus DNA replication in vitro: Synthesis of full-length DNA with purified proteins Proc Natl Acad Sci USA 1983 80 4266 4270 6308611
Hay RT Origin of adenovirus DNA replication: Role of the nuclear factor I binding site in vivo J Mol Biol 1985 186 129 136 4078897 10.1016/0022-2836(85)90263-3
Leegwater PAJ van Driel W van der Vliet PC Recognition site of nuclear factor I, a sequence-specific DNA-binding protein from HeLa cells that stimulates adenovirus DNA replication EMBO J 1985 4 1515 1521 4040852
de Jong RN van der Vliet PC Mechanism of DNA replication in eukaryotic cells: cellular host factors stimulating adenovirus DNA replication Gene 1999 236 1 12 10433960 10.1016/S0378-1119(99)00249-8
Santoro C Mermod N Andrews PC Tjian R A family of human CCAAT-box-binding proteins active in transcription and DNA replication: cloning and expression of multiple cDNAs Nature 1988 334 218 224 3398920 10.1038/334218a0
Guo ZS DePamphilis ML Specific transcription factors stimulate simian virus 40 and polyomavirus origins of DNA replication Mol Cell Biol 1992 12 2514 2524 1317005
Gronostajski RM Knox J Berry D Miyamoto NG Stimulation of transcription in vitro by binding sites for nuclear factor I Nucl Acids Res 1988 16 2087 2098 3357767
Cooke DW Lane MD The transcription factor nuclear factor I mediates repression of the GLUT4 promoter by insulin J Biol Chem 1999 274 12917 12924 10212282 10.1074/jbc.274.18.12917
Gil G Smith JR Goldstein JL Slaughter CA Orth K Brown MS Osborne TF Multiple genes encode nuclear factor 1-like proteins that bind to the promoter for 3-hydroxy-3-methylglutaryl-coenzyme A reductase Proc Natl Acad Sci USA 1988 85 8963 8967 3194401
Rupp RAW Kruse U Multhaup G Gobel U Beyreuther K Sippel AE Chicken NFI/TGGCA proteins are encoded by at least three independent genes: NFI-A, NFI-B and NFI-C with homologues in mammalian genomes Nucl Acids Res 1990 18 2607 2616 2339052
Qian F Kruse U Lichter P Sippel AE Chromosomal localization of the four genes NFIA, B, C, and X for the human transcription factor Nuclear Factor I by FISH Genomics 1995 28 66 73 7590749 10.1006/geno.1995.1107
Chaudhry AZ Lyons CE Gronostajski RM Expression patterns of the four Nuclear Factor I genes during mouse embryogenesis indicate a potential role in development Dev Dyn 1997 208 313 325 9056636 10.1002/(SICI)1097-0177(199703)208:3<313::AID-AJA3>3.0.CO;2-L
Fletcher CF Jenkins NA Copeland NG Chaudhry AZ Gronostajski RM Exon structure of the nuclear factor I DNA-binding domain from C. elegans to mammals Mamm Genome 1999 10 390 396 10087299 10.1007/s003359901008
Gronostajski RM Role of NFI/CTF gene family in transcription and development Gene 2000 249 31 45 10831836 10.1016/S0378-1119(00)00140-2
Mermod N O'Neill EA Kelly TJ Tjian R The proline-rich transcriptional activator of CTF/NF-1 is distinct from the replication and DNA binding domain Cell 1989 58 741 753 2504497 10.1016/0092-8674(89)90108-6
Kruse U Sippel AE Transcription factor nuclear factor I proteins form stable homo- and heterodimers FEBS Lett 1994 348 46 50 8026582 10.1016/0014-5793(94)00585-0
Gronostajski RM Adhya S Nagata K Guggenheimer RA Hurwitz J Site-specific DNA binding of nuclear factor I: Analyses of cellular binding sites Mol Cell Biol 1985 5 964 971 4039788
Gronostajski RM Analysis of nuclear factor I binding to DNA using degenerate oligonucleotides Nucl Acids Res 1986 14 9117 9132 3786147
Meisterernst M Gander I Rogge L Winnacker EL A quantitative analysis of nuclear factor I/DNA interactions Nucl Acids Res 1988 16 4419 4435 3380685
Roulet E Bucher P Schneider R Wingender E Dusserre Y Werner T Mermod N Experimental analysis and computer prediction of CTF/NFI transcription factor DNA binding sites J Mol Biol 2000 297 833 848 10736221 10.1006/jmbi.2000.3614
Kruse U Sippel AE The genes for transcription factor nuclear factor I give rise to corresponding splice variants between vertebrate species J Mol Biol 1994 238 860 865 8182757 10.1006/jmbi.1994.1343
Bedford FK Julius D Ingraham HA Neuronal expression of the 5HT3 serotonin receptor gene requires nuclear factor 1 complexes J Neurosci 1998 18 6186 6194 9698312
Baumeister H Gronostajski RM Lyons GE Margolis FL Identification of NFI-binding sites and cloning of NFI-cDNAs suggest a regulatory role for NFI transcription factors in olfactory neuron gene expression Mol Brain Res 1999 72 65 79 10521600 10.1016/S0169-328X(99)00210-7
Spitz F Salminen M Demignon J Kahn A Daegelen D Maire P A combination of MEF3 and NFI proteins activates transcription in a subset of fast-twitch muscles Mol Cell Biol 1997 17 656 666 9001219
Archer TK Lefebvre P Wolford RG Hager GL Transcription factor loading on the MMTV promoter - A bimodal mechanism for promoter activation Science 1992 255 1573 1576 1347958
Spangenberg C Eisfeld K Stunkel W Luger K Flaus A Richmond TJ Truss M Beato M The mouse mammary tumour virus promoter positioned on a tetramer of histones H3 and H4 binds nuclear factor 1 and OTF1 J Mol Biol 1998 278 725 739 9614938 10.1006/jmbi.1998.1718
Ortiz L Aza-Blanc P Zannini M Cato AC Santisteban P The interaction between the forkhead thyroid transcription factor TTF-2 and the constitutive factor CTF/NF-1 is required for efficient hormonal regulation of the thyroperoxidase gene transcription J Biol Chem 1999 274 15213 15221 10329730 10.1074/jbc.274.21.15213
Cooke DW Lane MD Transcription Factor NF1 Mediates Repression of the GLUT4 Promoter by Cyclic-AMP Biochem Biophys Res Commun 1999 260 600 604 10403812 10.1006/bbrc.1999.0959
Rossi P Karsenty G Roberts AB Roche NS Sporn MB de Crombrugghe B A nuclear factor 1 binding site mediates the transcriptional activation of a type I collagen promoter by transforming growth factor-b Cell 1988 52 405 414 2830985 10.1016/S0092-8674(88)80033-3
Alevizopoulos A Dusserre Y Tsai-Pflugfelder M von der Weid T Wahli W Mermod N A proline-rich TGF-beta-responsive transcriptional activator interacts with histone H3 Genes Dev 1995 9 3051 3066 8543151
Shu T Butz KG Plachez C Gronostajski RM Richards LJ Abnormal development of forebrain midline glia and commissural projections in Nfia knock-out mice J Neurosci 2003 23 203 212 12514217
das Neves L Duchala CS Godinho F Haxhiu MA Colmenares C Macklin WB Campbell CE Butz KG Gronostajski RM Disruption of the murine Nuclear Factor I-A gene (Nfia) results in perinatal lethality, hydrocephalus and agenesis of the corpus callosum. Proc Natl Acad Sci USA 1999 96 11946 11951 10518556 10.1073/pnas.96.21.11946
Steele-Perkins G Butz KG Lyons GE Zeichner-David M Kim HJ Cho MI Gronostajski RM Essential role for NFI-C/CTF transcription-replication factor in tooth root development Mol Cell Biol 2003 23 1075 1084 12529411 10.1128/MCB.23.3.1075-1084.2003
Grunder A Ebel TT Mallo M Schwarzkopf G Shimizu T Sippel AE Schrewe H Nuclear factor I-B (Nfib) deficient mice have severe lung hypoplasia Mech Dev 2002 112 69 77 11850179 10.1016/S0925-4773(01)00640-2
Steele-Perkins G Plachez C Butz KG Yang GH Bachurski CJ Kinsman SL Litwack ED Richards LJ Gronostajski RM The transcription factor Nfib is essential for both lung maturation and brain development Mol Cell Biol 2005 25 685 698 15632069 10.1128/MCB.25.2.685-698.2005
Blumenthal T Steward K Riddle DL, Blumenthal T, Meyer BJ and Priess JR RNA processing and gene structure C elegans II 1999 33 Cold Spring Harbor, Cold Spring Harbor Laboratory Press 117 145
Goyal N Knox J Gronostajski RM Analysis of multiple forms of nuclear factor I in human and murine cell lines Mol Cell Biol 1990 10 1041 1048 2304457
Tabara H Motohashi T Kohara Y A multi-well version of in situ hybridization on whole mount embryos of Caenorhabditis elegans Nucl Acids Res 1996 24 2119 2124 8668544 10.1093/nar/24.11.2119
Edgar LG Wolf N Wood WB Early transcription of Caenorhabditis elegans embryos Development 1994 120 443 451 7512022
Seydoux G Fire A Soma-germline asymmetry in the distributions of embryonic RNAs in Caenorhabditis elegans Development 1994 120 2823 2834 7607073
Zwaal RR Broeks A van Meurs J Groenen JT Plasterk RH Target-selected gene inactivation in Caenorhabditis elegans by using a frozen transposon insertion mutant bank Proc Natl Acad Sci USA 1993 90 7431 7435 8395047
Murphy CT McCarroll SA Bargmann CI Fraser A Kamath RS Ahringer J Li H Kenyon C Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans Nature 2003 424 277 283 12845331 10.1038/nature01789
Kimura KD Tissenbaum HA Liu Y Ruvkun G daf-2, an insulin receptor-like gene that regulates longevity and diapause in Caenorhabditis elegans Science 1997 277 942 946 9252323 10.1126/science.277.5328.942
Holzenberger M Dupont J Ducos B Leneuve P Geloen A Even PC Cervera P Le Bouc Y IGF-1 receptor regulates lifespan and resistance to oxidative stress in mice Nature 2003 421 182 187 12483226 10.1038/nature01298
Flaherty DB Gernert KM Shmeleva N Tang X Mercer KB Borodovsky M Benian GM Titins in C.elegans with unusual features: coiled-coil domains, novel regulation of kinase activity and two new possible elastic regions J Mol Biol 2002 323 533 549 12381307 10.1016/S0022-2836(02)00970-1
Chaudhry AZ Vitullo AD Gronostajski RM Nuclear factor I (NFI) isoforms differentially activate simple versus complex NFI-responsive promoters J Biol Chem 1998 273 18538 18546 9660824 10.1074/jbc.273.29.18538
Chaudhry AZ Vitullo AD Gronostajski RM Nuclear factor I-mediated repression of the mouse mammary tumor virus promoter is abrogated by the coactivators p300/CBP and SRC-1 J Biol Chem 1999 274 7072 7081 10066764 10.1074/jbc.274.11.7072
Roulet E Armentero MT Krey G Corthesy B Dreyer C Mermod N Wahli W Regulation of the DNA-binding and transcriptional activities of Xenopus laevis NFI-X by a novel C-terminal domain Mol Cell Biol 1995 15 5552 5562 7565707
Puzianowska-Kuznicka M Shi YB Nuclear factor I as a potential regulator during postembryonic organ development J Biol Chem 1996 271 6273 6282 8626421 10.1074/jbc.271.11.6273
Kruse U Qian F Sippel AE Identification of a fourth Nuclear Factor I gene in chicken by cDNA cloning: NFIX Nucl Acids Res 1991 19 6641 1754401
Armentero MT Horwitz M Mermod N Targeting of DNA polymerase to the adenovirus origin of DNA replication by interaction with nuclear factor I Proc Natl Acad Sci USA 1994 91 11537 11541 7972097
Bandyopadhyay S Starke DW Mieyal JJ Gronostajski RM Thioltransferase (Glutaredoxin) reactivates the DNA-binding activity of oxidation-inactivated Nuclear Factor I J Biol Chem 1998 273 392 397 9417094 10.1074/jbc.273.1.392
Bandyopadhyay S Gronostajski RM Identification of a conserved oxidation-sensitive cysteine residue in the NFI family of DNA-binding proteins J Biol Chem 1994 269 29949 29955 7961993
Moerman DG Fire A Riddle DL, Blumenthal T, Meyer BJ and Priess JR Muscle: Structure, Function, and Development C elegans II 1997 33 Cold Spring Harbor, Cold Spring Harbor Laboratory Press 417 470
Harris TW Chen N Cunningham F Tello-Ruiz M Antoshechkin I Bastiani C Bieri T Blasiar D Bradnam K Chan J Chen CK Chen WJ Davis P Kenny E Kishore R Lawson D Lee R Muller HM Nakamura C Ozersky P Petcherski A Rogers A Sabo A Schwarz EM Van Auken K Wang Q Durbin R Spieth J Sternberg PW Stein LD WormBase: a multi-species resource for nematode biology and genomics Nucleic Acids Res 2004 32 Database issue D411 7 14681445 10.1093/nar/gkh066
Krause M Harrison SW Xu SQ Chen L Fire A Elements regulating cell- and stage-specific expression of the C. elegans MyoD family homolog hlh-1 Dev Biol 1994 166 133 148 7958441 10.1006/dbio.1994.1302
Trent C Tsuing N Horvitz HR Egg-laying defective mutants of the nematode Caenorhabditis elegans Genetics 1983 104 619 647 11813735
Shyn SI Kerr R Schafer WR Serotonin and Go modulate functional states of neurons and muscles controlling C. elegans egg-laying behavior Curr Biol 2003 13 1910 1915 14588249 10.1016/j.cub.2003.10.025
Avery L Horvitz HR Pharyngeal pumping continues after laser killing of the pharyngeal nervous system of C. elegans Neuron 1989 3 473 485 2642006 10.1016/0896-6273(89)90206-7
Brundage L Avery L Katz A Kim UJ Mendel JE Sternberg PW Simon MI Mutations in a C. elegans Gqalpha gene disrupt movement, egg laying, and viability Neuron 1996 16 999 1009 8630258 10.1016/S0896-6273(00)80123-3
Koelle MR Horvitz HR EGL-10 regulates G protein signaling in the C. elegans nervous system and shares a conserved domain with many mammalian proteins Cell 1996 84 115 125 8548815 10.1016/S0092-8674(00)80998-8
Robatzek M Niacaris T Steger K Avery L Thomas JH eat-11 encodes GPB-2, a Gbeta(5) ortholog that interacts with G(o)alpha and G(q)alpha to regulate C. elegans behavior Curr Biol 2001 11 288 293 11250160 10.1016/S0960-9822(01)00074-4
Witt CC Ono Y Puschmann E McNabb M Wu Y Gotthardt M Witt SH Haak M Labeit D Gregorio CC Sorimachi H Granzier H Labeit S Induction and Myofibrillar Targeting of CARP, and Suppression of the Nkx2.5 Pathway in the MDM Mouse with Impaired Titin-based Signaling J Mol Biol 2004 336 145 154 14741210 10.1016/j.jmb.2003.12.021
Gotthardt M Hammer RE Hubner N Monti J Witt CC McNabb M Richardson JA Granzier H Labeit S Herz J Conditional expression of mutant M-line titins results in cardiomyopathy with altered sarcomere structure J Biol Chem 2003 278 6059 6065 12464612 10.1074/jbc.M211723200
Lakowski B Hekimi S The genetics of caloric restriction in Caenorhabditis elegans Proc Natl Acad Sci USA 1998 95 13091 13096 9789046 10.1073/pnas.95.22.13091
Nasrin N Ogg S Cahill CM Biggs W Nui S Dore J Calvo D Shi Y Ruvkun G Alexander-Bridges MC DAF-16 recruits the CREB-binding protein coactivator complex to the insulin-like growth factor binding protein 1 promoter in HepG2 cells Proc Natl Acad Sci USA 2000 97 10412 10417 10973497 10.1073/pnas.190326997
Murtagh J Martin F Gronostajski RM The Nuclear Factor I (NFI) gene family in mammary gland development and function J Mamm Gland Biol Neopl 2003 8 241 254 10.1023/A:1025909109843
Wood WB Wood WB The Nematode Caenorhabditis elegans 1988 Cold Spring Harbor, New York, Cold Spring Harbor Laboratory
Studier FW Rosenberg AH Dunn JJ Dubenforff JW Use of T7 RNA polymerase to direct expression of cloned genes Meth Enzymol 1990 185 60 89 2199796
Gronostajski RM Site-specific DNA binding of nuclear factor I: effect of the spacer region Nucl Acids Res 1987 15 5545 5559 3039460
|
16242019
|
PMC1277823
|
CC BY
|
2021-01-04 16:30:32
|
no
|
BMC Dev Biol. 2005 Oct 20; 5:24
|
utf-8
|
BMC Dev Biol
| 2,005 |
10.1186/1471-213X-5-24
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1441623617210.1186/1471-2164-6-144Research ArticleGeneration, annotation, analysis and database integration of 16,500 white spruce EST clusters Pavy Nathalie [email protected] Charles [email protected] Lee [email protected] John A [email protected] Marie-Josee [email protected] Janice [email protected] James E [email protected] Etienne [email protected] Carine [email protected] Yaron [email protected] Sarah [email protected] George [email protected] Jerry [email protected] Jeff [email protected] Robert [email protected] Asim [email protected] Robert [email protected] Marco [email protected] Armand [email protected] Ernest [email protected] Jean [email protected] John [email protected] ARBOREA and Canada Research Chair in Forest Genomics, Pavillon Charles-Eugène-Marchand, Université Laval, Ste.Foy, Québec G1K 7P4, Canada2 Center for Computational Genomics and Bioinformatics, University of Minnesota, 420 Delaware St. S.E., MMC 43, Minneapolis, MN 55455, USA3 Laurentian Forestry Center (Canadian Forestry Service), Natural Resources Canada, 1055 rue du PEPS, Québec, Québec, G1V 4C7, Canada4 Genome Sciences Center, BC Cancer Agency, 675 West 10 th Avenue, Vancouver, BC, V5Z 1L3, Canada5 Department of Biological Sciences, University of Alberta, Edmonton, Alberta, T6G 2E9, Canada2005 19 10 2005 6 144 144 17 8 2005 19 10 2005 Copyright © 2005 Pavy et al; licensee BioMed Central Ltd.2005Pavy 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 sequencing and analysis of ESTs is for now the only practical approach for large-scale gene discovery and annotation in conifers because their very large genomes are unlikely to be sequenced in the near future. Our objective was to produce extensive collections of ESTs and cDNA clones to support manufacture of cDNA microarrays and gene discovery in white spruce (Picea glauca [Moench] Voss).
Results
We produced 16 cDNA libraries from different tissues and a variety of treatments, and partially sequenced 50,000 cDNA clones. High quality 3' and 5' reads were assembled into 16,578 consensus sequences, 45% of which represented full length inserts. Consensus sequences derived from 5' and 3' reads of the same cDNA clone were linked to define 14,471 transcripts. A large proportion (84%) of the spruce sequences matched a pine sequence, but only 68% of the spruce transcripts had homologs in Arabidopsis or rice. Nearly all the sequences that matched the Populus trichocarpa genome (the only sequenced tree genome) also matched rice or Arabidopsis genomes. We used several sequence similarity search approaches for assignment of putative functions, including blast searches against general and specialized databases (transcription factors, cell wall related proteins), Gene Ontology term assignation and Hidden Markov Model searches against PFAM protein families and domains. In total, 70% of the spruce transcripts displayed matches to proteins of known or unknown function in the Uniref100 database (blastx e-value < 1e-10). We identified multigenic families that appeared larger in spruce than in the Arabidopsis or rice genomes. Detailed analysis of translationally controlled tumour proteins and S-adenosylmethionine synthetase families confirmed a twofold size difference. Sequences and annotations were organized in a dedicated database, SpruceDB. Several search tools were developed to mine the data either based on their occurrence in the cDNA libraries or on functional annotations.
Conclusion
This report illustrates specific approaches for large-scale gene discovery and annotation in an organism that is very distantly related to any of the fully sequenced genomes. The ArboreaSet sequences and cDNA clones represent a valuable resource for investigations ranging from plant comparative genomics to applied conifer genetics.
==== Body
Background
Genomics projects have been initiated in several pine and spruce species to identify genes involved in traits of economic interest and of ecological significance in conifers. It is unlikely, however, that conifer genomes will be completely sequenced in the near future because of their shear size [1]. For example, estimates of the haploid DNA content of Pinus taeda ranged from 11 pg [2] to 23.2 pg [3] and that of Picea glauca ranged between 4.5 pg [4] to 20.2 pg [PGI5.0; [5]]. With around 10–20,000 Mb [6], conifer genomes are more than 100 times larger than that of Arabidopsis and three times larger than the human genome. Such a large genome suggests that strategies that aim at characterizing the coding component of the genome will be more cost efficient for the recovery of information, in the short term.
The large-scale sequencing and analysis of ESTs remain a fundamental part of genomics research to enable gene discovery and annotation in most forest tree species, but especially in conifers. Several EST sequencing projects have been initiated in pines; 191,229 ESTs from several species were assembled to produce 35,053 consensus sequences in the Pinus Gene Index [7]. A large majority of conifer sequences were shown to have sequence similarity to Angiosperm genes or genome sequences like Arabidopsis, however the identification of homologous sequences depends largely on the length of sequences available to conduct similarity searches [8,9]. In loblolly pine, for example, the majority of contigged sequences which had no sequence similarity to other genomes were very short and more than 90% of sequences above 1 kb in length gave strong matches to Arabidopsis [8]. Therefore, effective annotation of conifer coding sequences through comparative approaches is best achieved with complete information, which may be obtained by combining 3' and 5' sequences or by full length sequencing strategies. A recent investigation of the knox gene family in conifers showed that gene evolution and conifer protein family structure may diverge quite significantly from those of Angiosperm genomes [10]. It is unknown how widespread this phenomenon may be; however, the finding suggests that although conserved protein motifs may be unambiguously identified, the biological role of genes belonging to conifer protein families may not be readily inferred from their Angiosperm homologs. These data would support the argument in favour of thorough cDNA sequencing projects in conifers because they are distantly related to model Angiosperms like Arabidopsis, in order to fully characterize protein families.
Many conifer EST sequencing projects have focused on wood formation and secondary xylem in pines (e.g. due to the ecological significance of the genus and the economic importance of wood [8,11]). More recently, programs have emerged that involve other species including Douglas-fir [12] and spruce [13], and address other important aspects of tree physiology like the response to abiotic stresses or biotic stresses [12,14]. Macroarrays and microarrays ranging in scope from a few hundred to a few thousand genes have been developed to help identify genes involved in wood formation and to characterize their putative roles in determining wood quality (e.g. in maritime pine [15], and in loblolly pine [16]). The relatively high level of sequence similarity between genera within the Pinaceae family has lead to the use of loblolly pine arrays for expression profiling experiments in scots pine, norway spruce [17] and white spruce [18]. Transcript profiling has also been integrated into investigations of xylem differentiation in poplar [19], different questions related to wood formation have also been investigated by transcript profiling in Angiosperm trees, including heartwood of black locust trees [20], tissue differentiation in poplar [19] and tension wood formation in Eucalyptus [21].
Spruce is the most widely used genus for forest tree plantations in Canada, with hundreds of million seedlings planted each year [22]. It is also widely divergent from pine [23,24]. Genetic improvement of spruce species, mainly white and black spruces, has been ongoing in Canada since the 1950s and extensive information has been accumulated on the genetic control of commercially important traits. Genome mapping of spruces is underway to enable molecular breeding applications (e.g. [25]). Association mapping approaches have been proposed as most promising to identify genes underlying phenotypic variation in quantitative traits, and thus, to support the development of molecular breeding strategies in conifers [26]. Large-scale EST sequencing and analysis are expected to enable association studies and gene mapping research as they are prerequisite steps to identifying SNPs to use in high throughput genotyping assays.
The objective of this study was to produce extensive collections of EST sequences and cDNA clones to support manufacture of cDNA microarrays and gene discovery efforts in white spruce (Picea glauca [Moench] Voss). This collection of ESTs constitutes an important new resource for the genomics of white spruce and related species. In this paper, we report the sequence analysis of around 71,000 sequence reads obtained through 3' and 5' sequencing of cDNAs. Comparative analyses were conducted to assign a functional annotation based upon similarities. Spruce contigs were also correlated with terms derived from the Gene Ontology [27], and similarity searches were conducted against specialized databases to identify putative transcription factors, cell wall related proteins and protein domains available in PFAM. To mine this new sequence resource, a database called SpruceDB has been developed at the Center of Computational Genomics and Bioinformatics (CCGB, University of Minnesota) [28], which supports multiple queries on the occurrence of the ESTs in the libraries and on the functional annotations.
Results and discusion
Library development and resulting sequences
Tissue sampling and EST sequencing strategies
The cDNA libraries were developed with the goal of augmenting the representation of conifer transcripts available in public databases, and to support experimental goals related to vascular development. We sequenced ESTs from 16 non-normalized cDNA libraries, synthesized from diverse spruce organs and tissues, and representing various stages of development from immature embryos to 30 year-old trees in diverse growth conditions (Table 1) [see also Additional file 1].
Table 1 Sequencing and quality parameters of white spruce cDNA libraries. Quality reads had a Phred score above 20 over at least 100 bp after vector trimming.
Libraries, treatments and tissues Number of reads Library quality Sequence quality
3' 5' % Empty % >1.6 Kb Nb of quality reads % Quality reads Average length of quality reads (nt)
Male strobili development sequence 1,536 1,536 4 19 2,589 84 527
Female cones development sequence 1,536 1,536 15 9 2,324 76 500
Vegetative buds development sequence 1,536 0 5 15 1,062 69 560
Secondary xylem – mature trees 4,608 4,608 10 27 7,735 84 600
Cambium, phloem – mature trees 4,608 3,072 2 8 6,705 87 635
Secondary xylem – girdled seedlings 3,072 0 9 24 1,053 69 556
Cambium to bark – girdled seedlings 1,536 1,536 NA NA 937 31 577
Elongating root tips – saplings 1,536 1,536 6 19 1,053 69 395
Primary, secondary shoots-N treatments 3,072 1,536 16 50 3,031 66 736
Immature somatic embryos 3,072 0 4 44 2,220 72 692
Clean roots systems – N treatments 1,536 0 7 37 858 56 659
Clean roots systems – P treatments 3,072 1,536 15 19 3,776 82 705
Clean roots systems – Diurnal cycle 6,144 4,608 16 33 8,601 80 757
Root secondary xylem – mature trees 3,072 0 7 8 1,532 50 598
Annual flush shoots diurnal cycle – trees 4,608 3,072 11 10 5,164 67 658
Needles – N fertilization treatments 1,536 0 15 20 461 30 686
Total 46,848 24,576 49,101
We sequenced close to 50,000 cDNA clones, sampling between 1,536 and 6,144 clones from each cDNA library. The library quality assessment data, the number of sequencing reactions, and the number of high quality reads for each library are presented in Table 1. All clones were sequenced from the 3' end ; in addition, 5' sequencing was carried on many clones from the libraries of highest quality or most relevant to our research goals. In total, 71,424 reads were obtained and processed to remove vector and sequences of low quality (Phred score below 20). We thus retained 49,101 quality reads (QR) comprised of at least 100 contiguous nucleotides with a Phred score above 20 (Table 1). Among the quality reads 33.5% were from secondary vascular tissues, 32.2% were from roots, 16.7% from young shoots (all tissues), and the remaining 17.6% were from various organs including male strobili, female cones, buds, somatic embryos, and needles (Table 1).
EST assembly into contigs
The assembly of the 49,101 quality reads resulted in 9,354 contigs and 7,224 singletons, representing a total of 16,578 consensus sequences named ArboreaSet in the following. As a result of our sequencing strategy, 46% of the consensus sequences were derived from overlapping 3' and 5' reads of one or more cDNA clones (Figure 1). We considered that non-overlapping 3'and 5' reads derived from the same cDNA clone (i.e. a spanning clone) belonged to the same transcript. We thus used spanning clones to link several consensus sequences and obtained a reduced set of 14,471 sequences that we defined as "transcripts".
Figure 1 Composition of white spruce consensus sequences (contigs and singletons) according to orientation of direction of the reads (3' or 5') and according to their redundancy in the database (number of clones).
The proportion of consensus sequences represented by more than one cDNA clone was only 39%, which provides an estimate of the sequencing redundancy. The bidirectional sequencing strategy and the average length of quality reads (Table 1) also impacted upon the length distribution of consensus sequences. The most striking feature of the set of spruce consensus sequences is the small proportion of sequences under 600 nucleotides compared to the PGI5.0 pine sequence assembly (mainly derived from 5' reads) despite its much larger number of sequences (Figure 2). The average consensus lengths were 797 and 690 nucleotides, for the Arborea and PGI5.0 sets, respectively; the median lengths were 784 and 612 nucleotides for these same datasets. The deepest contigs in the ArboreaSet included sequences homologous to genes coding for a DNA methylase (202 clones), the translation elongation factor-1 alpha (111 clones), a polyubiquitin (68 clones), an homocysteine methyltransferase (68 clones), a S-adenosylmethionine synthetase (65 clones).
Figure 2 Sequence sizes. Size distribution of the consensus sequences derived from the pine (PGI5.0) and white spruce (ArboreaSet) assemblies.
We also estimated the level of redundancy among the 16,578 consensus sequences by comparing the entire set of sequences to itself with the blastn program (Table 2). High scoring pairs revealing more than 98% of identity over more than 100 bp were used to define 13,686 contig groups, indicating a level of 21.1% of redundancy among the consensus sequences (Table 2). The very large majority of the 13,686 contig groups were comprised of one or two consensus sequences; however a few groups (3) were made up of more than ten distinct sequences. In a collection of 43,141 consensus sequences derived from 260,000 sugarcane ESTs, the redundancy was estimated at 22%, based upon 98% over 100 bp [29]. In a Citrus EST sequence assembly, the level of redundancy was estimated at 25% [30]. Overall the redundancy is in the same range as observed in other projects conducted in mouse [31] or honey bee [32].
Table 2 Contig groups according to several levels of sequence identity based on 100 nt of overlap
Number of contigs per group 90% 96% 98% 99%
1 10,036 10,997 11,767 13,295
2 1,576 1,422 1,377 1,083
3 443 386 341 210
4 175 153 103 61
5 93 72 48 22
6 52 40 26 6
7 15 17 10 2
8 13 8 7 1
9 10 1 3 1
≥10 21 12 3 3
Total number of groups 12,435 13,109 13,686 14,685
Spruce and pine EST datasets are populated with allelic variants for many loci because conifers are outbred and highly heterozygous. As a consequence, the number of genes sampled may be estimated more or less accurately from the number of contigs or contig groups, depending upon the parameters that are used for their assembly and clustering. To our knowledge, the impact of assembly parameters has not been directly assessed in conifers or other Gymnosperms. On the other hand, the average nucleotide diversity was reported to be low for conifers [24,26]; for example, sequence variation was estimated in pines with the average mutation population parameter ? = 0.00407 in Pinus taeda [33], ? = 0.00241 in P. pinaster, ? = 0.00186 in P. radiata [34] and ? = 0.0013 in P. sylvestris [35]. These data suggest that the use of stringent criteria were appropriate for the assembly (into contigs) of the spruce sequence dataset comprised in part of allelic sequences. We also defined contig groups with less stringent criteria aiming to evaluate sequence redundancy. We recognize, however, that some contigs may contain paralogs, especially for slow-evolving gene families as discussed in other reports on plant EST clustering [29,30]. For these reasons, the contig groups are thought to provide a conservative estimate of the number of genes, i.e. the minimum number of genes sequenced.
Sequence comparisons with other species
We performed sequence similarity searches using tblastx and blastx to compare the ArboreaSet to several sequence datasets from Angiosperms (Arabidopsis, rice, poplar) and Gymnosperms (Cycas and pine), and to the Uniref100 protein database for several e-value cutoffs (Figure 3). In the following sections, the data were obtained with an e-value cutoff of 1e-10 unless specified otherwise.
Figure 3 Sequence similarities. Number of white spruce transcript sequences similar to Uniref100 proteins, Arabidopsis, pine, Cycas according to the blast e-value cutoff.
Sequence comparisons with the pine database and Angiosperm genomes
We found that 84.4% the Arborea transcript set (12,108 transcripts) showed sequence similarity with a contig of the Pine Gene Index (PGI5.0) which contains the largest assembly of publicly available pine ESTs (Figure 3). All of the tblastx searches detected a greater number of matches with PGI5.0 than with the Uniref100 protein database, in which the PGI5.0 consensus sequences are not represented. We examined whether the lack of similarity of the remaining 15.6% spruce transcripts (with no counterpart in the pine database) could be attributed to the non overlap of pine and spruce contigs derived from 5' and 3' sequences, respectively. More than half of the non matching spruce transcripts (9.8% of the total transcripts) were indeed derived only from 3' reads. Therefore, the lack of similarity of many of the sequences is not sufficient to conclude whether a pine homolog is absent from the database. Nonetheless, 6.6% of the spruce transcripts were derived from 5' reads alone (predominantly) or both 5' and 3' reads and, did not match a pine contig. For these sequences, there is a high likelihood that a similar pine transcript has not been sequenced thus far.
As might be expected, the overall sequence similarity was lower with Angiosperms than pine sequences. There were fewer matches and the number of matches decayed more rapidly as we used more stringent e-value cutoffs with tblastx against Angiosperms sequences. At the protein level, 68.4% (9,898) of spruce transcripts matched a sequence from Arabidopsis or rice with a tblastx e-value < 1e-10 and the proportion dropped to 37.6% for highly conserved sequences (e-value < 1e-50) compared to 65.3% with pine. A similar trend was observed with the poplar genome sequence which gave slightly lower similarities than Arabidopsis and rice sets, i.e. 64.3% and 21.6% matches with e-values below 1e-10 and 1e-50, respectively (Figure 3).
Complementarity of the sequencing projects in several species
We analyzed and compared the overlap of sequence datasets derived from spruce, pine, Arabidopsis, rice and poplar, to develop an overall understanding of the complementarity between the sequencing projects in these diverse species. The extent of the overlap based upon tblastx matches is shown in Figure 4. In total, 77.5% transcripts found both in pine and spruce databases (9,384 of 12,108) gave a match with Arabidopsis or rice. However, only 514 (3.6%) spruce transcripts without any homolog in the pine database had a homolog in Arabidopsis or rice. In contrast, 41.7% out of the 26,616 consensus sequences from PGI5.0 that had no match in the ArboreaSet, gave a hit in Arabidopsis or rice. These sequence results appear consistent with the extent of divergence that might be expected between the genomes of Gymnosperms and Angiosperms. In a previous study, pine consensus sequences gave 61.5%, 59.4% and 55% matches against Arabidopsis, rice and poplar, respectively [9]. With the same similarity search parameters, the spruce transcripts – which contains longer sequences on average – gave slightly more matches against Arabidopsis or rice (68.4%), and as well as against poplar (64.3%).
Figure 4 Hierarchical presentation of the number of spruce transcripts with or without similarities with pine, Arabidopsis, rice and poplar. The numbers were derived by the filtering of tblastx searches with an e-value < 1e-10.
Comparisons to the poplar genome sequence gave fewer matches and only a small number of matches not identified with Arabidopsis or rice (Figure 4). Only 89.1% of the spruce transcripts (8,823 out of 9,898) that matched an Arabidopsis or rice sequence also had a similarity to a sequence in the poplar genome. Furthermore, 3.5% of the spruce transcripts which lacked similarity to Arabidopsis or rice gave a match against the poplar genome. In the end, sequence similarity searches against the poplar genome only allowed us to annotate 162 additional sequences (0.7% of the spruce transcripts) compared to data derived from comparisons with Arabidopsis or rice. Such a trend is expected given the relatively close proximity between poplar (Salicaceae) and Arabidopsis (Brassicaceae).
The results indicate that data derived from Angiosperms species alone are insufficient for annotating sequences in conifers and that computational tools specifically developed for Gymnosperms are needed to help recognize functional regions in sequences like coding sequences or motifs in around 30% of conifer sequences with no obvious counterpart in Angiosperms. For example, the software Diogenes for predicting open reading frames in sequences was trained based on Pinaceae derived sequences for this purpose [36].
Functional annotation
In total, 10,130 (70%) of the spruce transcripts displayed matches to proteins of known or unknown function, based on the blastx analysis against the Uniref100 database. We conducted Hidden Markov Model (HMM) searches against the PFAM protein family database [37,38] to evaluate the proportion of the spruce transcripts homologous to families with an assigned function. Overall, we found that 52% of the 14,471 spruce "transcripts" showed similarity with 1,655 PFAM protein families (p-score below 1e-10). There were 157 of these PFAM families annotated as "DUF, Domain of Unknown Function", which showed similarities with 488 transcripts, and 20 families annotated as "UPF, Uncharacterized Protein Family" showing similarities with 45 transcripts. In the end, a total 48% of the spruce transcripts were similar to 1,478 PFAM families when DUFs and UPFs were excluded.
A separate approach using the Gene Ontology scheme [27] categorized 39% of the ArboreaSet contigs into 16 molecular functions based on similarity with functionally annotated genes in other organisms (Table 3). Functional categories were assigned by using the GO terms correlated to similar proteins from Uniref100 [39] or from the Arabidopsis databases [40]. In the molecular function category, 39% of the contigs were correlated to a GO term. When the less reliable electronically inferred functional annotations were excluded, 30% of the ArboreaSet contigs were assigned molecular function annotations. The catalytic activity category included the largest number of sequences, followed by the proteins of unknown function. The classification we obtained was similar to that in the PGI5.0 database [41]. A significantly larger proportion of contigs were annotated in spruce than in pine, since we considered all of the blastx hits that met the alignment criteria, while the PGI5.0 annotations used only the top hit. Due to the restricted number of well-characterized conifer genes, correlating conifer sequences to Gene Ontology terms relies primarily on conservation with Angiosperms sequences (mainly Arabidopsis and rice). Therefore, the GO annotated contigs in spruce and in pine are the ones conserved with Angiosperms.
Table 3 Consensus sequences correlated to terms belonging to the "molecular function" categories of the Gene Ontology
Molecular functions Annotations including electronic annotations Annotations excluding electronic annotations
Number of consensus sequences % of the number of annotated consensus sequences % of the total number of consensus sequences Number of consensus sequences % of the number of annotated consensus sequences % of the total number of consensus sequences
Triplet codon-amino acid adaptor activity 0 0 0 0 0 0
Chaperone regulator activity 0 0 0 0 0 0
Motor activity 23 0.35 0.14 3 0.06 0,02
Enzyme regulator activity 47 0.71 0.28 27 0.53 0.16
Nutrient reservoir activity 50 0.76 0.30 4 0.08 0.02
Translation regulator activity 70 1.06 0.42 59 1.16 0.36
Antioxidant activity 73 1.10 0.44 52 1.02 0.31
Signal transducer activity 77 1.16 0.46 33 0.65 0.2
Obsolete molecular function 113 1.71 0.68 76 1.5 0.46
Transcription regulator activity 118 1.78 0.71 73 1.44 0.44
Chaperone activity 166 2.51 1 142 2.79 0.86
Structural molecule activity 283 4.28 1.70 240 4.72 1.45
Transporter activity 503 7.60 3.03 335 6.59 2.02
Binding 1,248 18.87 7.52 741 14.6 4.46
Molecular function unknown 1,340 20.26 8.07 1,340 26.4 8.07
Catalytic activity 2,504 37.85 15.08 1,956 38.5 11.8
Total 6,615 100 39.84 5,081 100 30.6
HMM searches against the PFAM database showed that the most abundant sequences in plant genomes were also among the most represented in the ArboreaSet (Figure 5). Highly comparable findings were made with the pine dataset (PGI5.0). A similar analysis conducted with the sugarcane SUCEST database indicated that the most abundantly represented molecular functions were largely overlapping between conifers and sugarcane [29].
Figure 5 Protein families. Occurrence of the 30 most abundant protein families in the white spruce dataset identified by HMM searches with an e-value < 1e-10 against the PFAM database.
Families of putative transcription factors
We identified putative transcription factors based upon the assignment of GO terms, as well as sequence comparison to PFAM domains and families [37,38]. The GO based annotation "transcription regulator activity" was assigned to 113 spruce sequences (including 40 assignments based upon automatic annotations) and the annotation "transcription factor activity" (GO:0003700) was assigned to 90 of the same consensus sequences. We also conducted HMM searches with the 41 PFAM profiles representing the plant transcription factors described in the Arabidopsis thaliana Transcription Factor Database (AtTFDB, from the Arabidopsis Gene Regulatory Information Server, AGRIS) [42] and identified 304 spruce transcripts (Table 4). Only 43 of these putative transcription factors were identified by both approaches. The combined total represented 388 putative transcription factors sequences. The most frequent sequence similarities were with C3HC4 zinc finger domain, WD, and AP2, respectively.
Table 4 Identification of transcripts encoding putative regulatory proteins. Sequences were identified based on HMM searches suported by p-score < 1e-10 with PFAM profiles available for families of regulatory proteins. The PFAM accessions for which no homology was found in SpruceDB through HMM search were not reported.
Protein family PFAM accession Number of spruce transcripts
Zinc finger, C3HC4 type (RING finger) PF00097 66
WD, G-beta repeat PF00400 44
AP2 domain-B3 DNA binding domain PF00847 19
HMG (high mobility group) box PF00505 16
MADS Family – SRF-type transcription factor – K-box region PF00319 14
MYB DNA-binding PF00249 13
AUX/IAA PF02309 12
Histone-like transcription factor (CBF/NF-Y) and archaeal histone PF00808 11
PHD finger – CW-type Zinc Finger PF00628 10
No apical meristem (NAM) protein PF02365 10
GRAS Family PF03514 10
WRKY DNA-binding domain PF03106 9
NAC domain PF01849 9
Homeobox domain PF00046 8
bZIP transcription factor – bZIP Maf transcription factor-G-box binding protein MFMR PF00170 8
B-box zinc finger PF00643 6
TUB Family PF01167 6
Helix-loop-helix DNA-binding domain – Myc amino-terminal region PF00010 5
KNOX2 domain PF03791 3
LIM domain family – PET Domain PF00412 5
Dof domain, zinc finger PF02701 4
GATA zinc finger PF00320 3
TCP family transcription factor PF03634 2
CCAAT-HAP2 Family CCAAT-binding transcription factor (CBF-B/NF-YA) subunit B PF02045 2
SBP (Sqamosa-promoter binding protein) floral development PF03110 1
HSF Family (Heat shock protein promoter binding) PF00447 1
EIL Family ethylene insensitive 3 PF04873 1
B3 DNA binding domain PF02362 1
ARID/BRIGHT DNA binding domain – ELM2 domain PF01388 1
Cell wall related genes
Many of the libraries that we constructed were derived from secondary vascular tissues from stems or roots, or from whole stems or roots containing primary as well as secondary vascular regions. Therefore, we aimed to classify genes which encode proteins potentially involved in cell wall assembly. As a first step toward this goal, our collection of spruce transcripts was blasted against the sequences from the Cell Wall Navigator Database [43] [see Additional file 2], comprised of proteins involved in primary cell wall structure and assembly. In total, we found that 708 spruce contigs were similar to sequences of cell wall related proteins, with nearly all of the subclasses represented. We also searched for genes encoding enzymes involved in the biosynthesis monolignol precursors based upon sequence similarity with the set identified in Arabidopsis by Reas et al. [44], and identified 47 additional contigs (Supplemental data 2).
Redundancy analysis suggests larger size of selected protein families in spruce compared to Angiosperms
It is not expected that Gymnosperm genomes will be sequenced in the foreseeable future, therefore we undertook a preliminary comparative analysis of protein families using the 14,471 spruce transcripts, to assess whether insights may be gained into the relative size of protein families in Gymnosperms and Angiosperms. We compared the occurrence of proteins in the ArboreaSet to that observed in the Arabidopsis genome (1,611 families) as well as in the rice genome (1,601 families) identified with HMM searches against the PFAM database. As might be expected from the partial coverage of the spruce genome, the vast majority of the protein families were represented by a larger number of sequences in the Arabidopsis and rice genomes than in ArboreaSet (Figure 6). However, several families gave twice as many hits in the ArboreaSet (67 and 58 families compared to Arabidopsis and rice, respectively) and a few families had at least 4 times more sequences (6 for Arabidopsis and 10 for rice, including 3 families for both). Some of these families encoded proteins that can be linked to the cell wall catabolism (PF01476), single carbon metabolism (S-adenosylmethionine synthetase PF02773), the cytoskeleton (Translationally Controlled Tumour Protein, TCTP family PF00838) or the cellular membrane (AWPM-19-like family PF05512). We verified that the size of the 4-fold larger families of spruce sequences was not inflated due to incomplete assembly of 3' and 5' reads. For two of the putatively larger families, we examined the protein and nucleic acid sequence diversity between the different consensus sequences in order to estimate the number of family members (results presented below).
Figure 6 Number of spruce consensus sequences (identified by HMM searches against PFAM) relative to the size of the gene families in Arabidopsis (a) and rice (b). Each point represents a protein family detected by the HMM searches with p-score < 1e-10. Point coordinates are the number of genes found in the analysed Angiosperm genome (x axis) and the number of contigs found in the spruce database (y axis), after a log transformation. The red, blue and green lines represent the ratios 1:1, 1:2, and 1:4, respectively. Red points represent sequences found 4 times more in white spruce than in Arabidopsis: 1. AWPM-19-like family [PF05512], 2. Chalcone and stilbene synthases, C-terminal domain [PF02797], 3. Phosphoenolpyruvate carboxykinase [PF01293]. Blue points represent sequences found 4 times more in spruce than in rice : 4. Ribosomal protein S28e [PF01200], 5. Cyclin-dependent kinase regulatory subunit [PF01111], 6. TIR domain [PF01582], 7. Splicing factor 3B subunit 10 [PF07189], 8. Ribosomal Proteins L2, C-terminal domain [PF03947]. Green points represent sequences found 4 times more in spruce compared to both Arabidopsis and rice: 9. Translationally controlled tumour protein [PF00838], 10. S-adenosyl-L-homocysteine hydrolase [PF05221], 11. S-adenosylmethionine synthetase, C-terminal domain [PF02773].
The cytoskeleton related TCTP family
The translationally controlled tumour proteins (TCTPs) are anti-apoptotic proteins, named for their preferential synthesis in the early phase of some tumours [45]. They are implicated in both cell growth and division and have been shown to bind to tubulin in the cytoskeleton. In plants, similar proteins were identified in alfafa [46] and Pharbitis mil [47].
The TCTP domain (accession : PF00838) was found in only two Arabidopsis sequences (At3g16640.1 and At3g05540.1) and one rice sequence (location in Gramene: LOC_Os11g43900.1). In contrast, there were 11 transcripts in the ArboreaSet that encompassed a highly conserved region of TCTPs and showed a high level of sequence conservation with Arabidopsis TCTPs (e.g. 70% a.a. identity for predicted sequence of Contig9531 and the Arabidopsis sequence gb|AAM66134.1). In total, 8 of the 11 spruce TCTP transcripts encompassed a putative complete coding sequence that overlapped with the Arabidopsis proteins. Pairwise nucleic acid sequence comparisons of the 11 spruce transcripts were used to identify 5 distinct sequence groups, likely representing a minimum of 5 different genes (Table 5). Based upon these data, the TCTP family provided an example of putative differential protein family size between spruce and the Angiosperms represented by rice and Arabidopsis.
Table 5 Pairwise comparison of white spruce consensus sequences related to the translationally controlled tumour proteins (TCTP). Nucleic acid identities were determined using the Smith-Waterman algorithm (water) available in the EMBOSS suite [71] in a 138 bp region of the 5' UTR immediately upstream of the first codon (ATG), (above the diagonal); and, along the complete sequence of the consensus sequences (under the diagonal). The diagonal shows the contig length.
Sequence10076 Sequence10707 Sequence9531 Sequence7749 Sequence1882
Sequence10076 805 88/162 (54.3%) 54/84 (64.3%) 70/159 (44.0%) 83/144 (57.6%)
Sequence10707 761/890 (85.5%) 977 111/157 (70.7%) 71/147 (48.3%) 99/154 (64.3%)
Sequence9531 759/889 (85.4%) 925/1034 (89.5%) 1124 65/133 (48.9%) 101/159 (63.5%)
Sequence7749 515/659 (78.1%) 548/736 (74.5%) 596/938 (63.5%) 945 73/147 (49.7%)
Sequence1882 719/815 (88.2%) 742/823 (90.2%) 750/906 (82.8%) 523/687 (76.1%) 796
The SAMS family
Sequences encoding S-adenosylmethionine synthetases (SAMS), a family of enzymes involved in single carbon metabolism and in lignin precursor biosynthesis [48] were represented by 24 consensus sequences encompassing at least seven spruce genes (Table 6). It has been reported that sams genes belong to small gene families in other plant species [49-53]. In Arabidopsis, four sams genes were identified. In rice, three sequences encoding complete proteins of 396 amino acids were found, as well as two sequences encoding truncated proteins of 164 amino acids [54].
Table 6 Pairwise comparison of white spruce consensus sequences related to the S-adenosylmethionine synthetase (SAMS). Nucleic acid identities were determined using the Smith-Waterman algorithm (water) available in the EMBOSS suite [71] in a 99 bp region of the 3' UTR immediately downstream the stop codon (above the diagonal) and along the complete sequence of the consensus sequences (under the diagonal). The diagonal shows the contig length.
Sequence 10446 Sequence 10482 Sequence 10630 Sequence 10683 Sequence 10828 Sequence 8600 Sequence 9676
Sequence10446 1677 46/97 (47.4%) 48/113 (42.5%) 51/117 (43.6%) 45/85 (52.9%) 85/106 (80.2%) 44/98 (44.9%)
Sequence10482 1096/1607 (68.2%) 1467 54/78 (69.2%) 65/92 (70.7%) 50/84 (59.5%) 49/114 (43%) 45/96 (46.9%)
Sequence10630 1126/1641 (68.6%) 1343/1557 (86.3%) 1540 69/113 (61.1%) 48/78 (61.5%) 47/95 (49.5%) 55/111 (49.5%)
Sequence10683 1143/1711 (66.8%) 1342/1521 (88.2%) 1357/1582 (85.8%) 1531 49/103 (47.6%) 58/116 (50%) 46/117 (39.3%)
Sequence10828 1202/1814 (66.3%) 1262/1534 (82.3%) 1343/1714 (78.4%) 1306/1604 (81.4%) 1679 49/95 (51.6%) 49/109 (45%)
Sequence8600 1349/1691 (79.8%) 1058/1536 (68.9%) 1089/1532 (71.1%) 1092/1583 (69%) 1120/1656 (67.6%) 1476 41/71 (57.7%)
Sequence9676 1025/1418 (72.3%) 1314/1459 (90.1%) 1276/1397 (91.3%) 1261/1381 (91.3%) 1179/1369 (86.1%) 1026/1462 (70.2%) 1356
We analyzed the 8 spruce sams transcripts that encompassed complete protein coding sequences averaging 393 amino acids in length. The predicted proteins were very highly conserved with Angiosperm SAMS. For example, the Arabidopsis SAMS2 protein (locus At4g01850) had a similarity of 88% (345/390a.a) and 90% (354/390 a.a) with the predicted proteins from the spruce contigs 10446 and 10482, respectively. Pairwise comparisons of the spruce coding sequences showed they are highly conserved, yet they could be divided into seven groups of sequences with 66.3% to 91.3% identity (Table 6). We also analyzed the nucleic acid sequence of their 96 bp 3' UTR and found significant variability between groups, with sequence identities varying from 42.5% to 70.7% (Table 6). These results provided a strong indication that these putative sams transcripts represented 7 distinct genes. Protein and nucleic acid sequence comparisons supported the hypothesis that the SAMS proteins form a larger family in the spruce genome than in Arabidopsis and rice. In rice, the presence of two pseudogenes indicated that protein family expansions through duplication events have been followed by gene loss during the evolution. Two sams genes were described in Pinus contorta [55]; however, large-scale EST sequencing in Pinus taeda [56] identified 16 consensus sequences, suggesting that the relatively large family size of SAMS in spruce may also apply to pine and other Gymnosperm genomes.
Development of Spruce DB
The relational database, SpruceDB, was created to allow complex queries into the spruce ESTs, assembled consensus sequences and results of similarity analyses. The database can be accessed via web browser [28]. Web-based tools provide facilities for exploration of this information resource. The ESTs or contigs can be retrieved based on library composition and sequence similarities. Web links from the database query pages retrieve the actual EST and contig sequences from the Biodata web pages [56].
Structure and data sources
The database schema for SpruceDB is identical to the one successfully used by the MtDB2.0 database for Medicago truncatula EST data [57]. SpruceDB is hosted on a Sun V880 server running the Oracle 8i Database Management System. The data sources and core tables for the database are illustrated in Figure 7. Sequence trimming methods and assembly parameters for Phrap are described in the Methods section. Information about ESTs and consensus sequences assembled with Phrap is extracted from flat files and loaded into the core tables Read, Contig and Contig_Element. These tables store the sequences and lengths, base qualities, EST libraries, clone names, and assembly name. Several tables store pre-computed blast hit information from blast similarity analyses against several target databases: UniRef100 peptides, Arabidopsis proteins, SpruceDB itself. Data fields include analysis program, target name, hit identifier, e-value, identities and taxonomy identifier for each hit. All blast hits with e-values less than 0.01 are loaded into the database.
Figure 7 SpruceDB core tables and data sources. Data from flat files on ESTs, Assemblies and blast hits is loaded into the core tables Read, Contig, Contig_Element and Blast_Hsp. Additional information on taxonomy identifiers and Uniref100 peptides is obtained from shared databases.
Interface
The web pages used to query the database allow retrieval of ESTs or contigs based on the cDNA libraries and blast hits (Figure 8). Since the nine query pages consist of check boxes and pull-down menus, no programming or knowledge of SQL is required, yet users can generate complex queries. Query 1 retrieves consensus sequences that have blast hits containing user-specified keywords or accession numbers. Queries 2–7 are library filter queries which retrieve ESTs or consensus sequences containing "any of", "all of", or "only" ESTs from user-specified cDNA libraries. Queries 3–7 contain taxonomy and e-value filters which retrieve sequences that have blast hits to organisms from specified taxa such as "all pines", "all poplars", or Arabidopsis. Query 5 combines the library, taxonomy and e-value filters in a single web page. Query 8 retrieves EST sequences using different names (aliases). Query 9 compares consensus sequences between different assemblies.
Figure 8 Examples of the interface of the SpruceDB database. A) Use of Query 1 to search for contigs matching "cinnamoyl alcohol dehydrogenase" among the blastx results loaded in the database. B) Display of the results indicating alignment parameters (alignment length, similarity and identity level). C) BioDATA page linked to by clicking on MNC5693153 in Query 1 results. The upper figure illustrates the alignment of the members of the contigs in a color coded manner. Read names written in blue and white color refer to 5'and 3'reads, respectively. D) Query 8 allowing to retrieve sequence aliases and library names for specified MN_Ids. E) Query 8 results showing libraries GQ004 and GQ006.
Conclusion
In this report, we described a new conifer EST resource derived from 49,101 high quality 5' and 3' reads that were assembled to produce 16,578 consensus sequences averaging 797 nucleotides in length, and representing 14,471 different "transcripts". We estimated the sequencing redundancy at 39% based on the number of consensus sequences represented by more than one cDNA clone. Comparison of the spruce sequences to public sequence datasets from Angiosperms and pine showed that approximately 70% of the sequences had similarity with Arabidopsis, rice or poplar sequences, but 84% matched a pine sequence. The majority of the sequences that did not give a match in pine did not produce a match with any of the Angiosperms either. We used a variety of approaches based on sequence similarity searches to assigned putative functions to the ArboreaSet sequences, including blast searches against general and specialized datasets, GO term assignation, HMM searches against PFAM protein families and domains. These analyses were used for the systematic identification of diverse putative transcription factors, cell wall related enzymes and structural proteins, and revealed a few protein families that are thought to be larger in spruce than in the well-characterized genomes of Arabidopsis and rice.
These comprehensive analyses to enable the annotation of spruce sequences provide critical information to help identify target genes for functional analysis and association studies. Studies are now being planned based on these data to search for DNA polymorphisms underlying the extensive phenotypic variation which occurs in natural and breeding populations. These studies will focus on sequences encoding proteins relevant for adaptation, growth and wood formation for large-scale SNP discovery and genotyping required for association studies and gene mapping. It is therefore essential that we develop databases of annotated coding sequences so that we may rapidly identify and screen the most suitable targets. As a first step toward this goal, the relational database SpruceDB was created to allow complex queries into the spruce ESTs, assembled consensus sequences and results of similarity analyses. By using this EST resource, we have also developed a low redundancy cDNA microarray comprised of 9,690 sequences, which, in combination with multiple sequence annotations, will be a powerful tool to investigate transcriptome modulation in spruces and conifers.
Methods
Plant material
All of the libraries were comprised of a single organ or tissue, and the majority of libraries were developed by pooling samples collected at different points along a time course, along the diurnal cycle, at several stages of differentiation or from different treatments (Supplemental data 2 and [58]). Treatments known to affect plant physiology were applied to saplings (young trees) aiming to stimulate different transcript profiles. These treatments included N and P fertilization as well as stem girdling. Three libraries were made from whole root systems of very young spruce seedlings, produced through tissue culture, grown in sterile growth media. Most of the libraries were derived from one genotype (pg-653), however four libraries were comprised of two or more genotypes. The secondary xylem collected from saplings (library GQ007) was comprised of the entire sampling of woody tissues collected from seedlings; however, only the differentiating partly-lignified secondary xylem was collected from mature trees as previously described [16]. The secondary xylem tissues were collected by first gently separating the bark from the underlying wood and scraping the soft tissues inward of the cambial area. The secondary phloem of mature trees was collected by gently scrapping the inner surface of the bark with a scalpel blade. All tissue samples were frozen in liquid nitrogen and then stored at -80°C until RNA extraction immediately upon removal from the tree, seedling or tissue culture vessel.
cDNA library construction, quality controls and high-throughput sequencing
We began the construction of each library with 1000 micrograms of total RNA or more, isolated using the method of Chang et al. [59]. Poly A+ RNA was isolated using the PolyATtract mRNA Isolation System (Promega, San Luis Obispo, CA, USA). The polyA+ RNA was treated with methylmercury hydroxide according to the manufacturer's instruction (Stratagene, La Jolla, CA, USA) to relax its secondary structure. Double-stranded cDNA was synthesized from 5 micrograms (µg) of poly A+ selected RNA using a pBluescript II SK (+) XR cDNA Library Construction kit (Stratagene, La Jolla, CA, USA). The reverse transcription step was carried out with either Superscript II or Superscript III and StrataScript (Invitrogen, Burlington, ON, Canada; Stratagene, La Jolla, CA, USA) as described in the manufacturer's instructions. The double stranded cDNAs were fractionated using the Drip column method (Stratagene, La Jolla, CA, USA) or by agarose gel electrophoresis on NuSieve GTG Agarose (Mendel, Guelph, ON, Canada) followed by selective elution of particular-sized cDNA molecules by ß-Agarase I digests according to the manufacturer's instruction (NEB, Pickering, ON, Canada). The size distribution of the resulting double cDNA synthesized in second-strand fractions was visualized by electrophoresis on a 1.4% alkaline agarose gel [60]. The fractions of 600 pb to 1.2 kb and above 1.2 kb were selected, pooled, directionally ligated into the EcoRI and XhoI restriction sites of the pBluescript II SK (+) XR vector (Stratagene, La Jolla, CA, USA), and transformed into E.coli DH10B competent cells (Invitrogen, Burlington, ON, Canada) by electroporation. The library quality assessment used test ligations to determine library titer. We also estimated the proportion of empty vectors as based upon the proportion of blue to white colonies grown on LB agar supplemented with X-GAL/IPTG (Table 1). The average cDNA insert size was determined by PCR screening of 48 to 96 random white colonies (assumed contain plasmids with inserts) per test ligation, followed by determination of the PCR product size by gel electrophoresis. The highest quality libraries were those estimated to have the highest proportion of inserts above 1.6 Kb (Table 1). High-throughput sequencing of libraries was completed using standard methodsas described by Yang et al. [61].
EST processing and assembly
Sequence traces from the spruce EST libraries were analyzed with the Phred base calling software (version 0.980904) to generate raw sequences [62]. Peaks with Phred quality values less than 20 were considered to be ambiguous and were assigned base N. Quality trimming and vector filtering (with polyA/polyT removal, as appropriate) were done. Processed sequences were then assembled using the base quality files and Phrap (version 0.990329) [63]. Phrap contigs were evaluated for chimeric sequences, and reassembled after removing chimeric reads. The Phrap assembly parameters used were minmatch 50 and minscore 100. Only reads with at least 100 nt of sequence with a quality score above 20 were assembled. EST sequences were submitted to dbEST at the National Center for Biotechnology Information [64] under accession numbers : [Genbank:CK434215-CK445169] and [Genbank:CO472624-CO490610].
Quality control of consensus sequences
The quality control of resulting consensus sequences used a system developed at the CCGB. This system uses information that is included in the contig ace file generated by Phrap. From the ace file, several important characteristics of a consensus sequence and its member sequences can be determined. The first characteristic used in this process is the "shape" of the consensus sequence, or how the assembled reads overlap each other. This can be thought of as the profile of the consensus sequence member distribution. Consensus sequences are classified as being of block, staircase, or dumbell shape. Contigs with a dumbell shape are candidates for additional evaluation.
Reads within a dumbell shaped contig are evaluated for their similarity to the consensus sequence of the contig. Phrap provides information on the quality regions of assembled sequences, which is used for this step. If the high quality region of the read (as defined by the Phrap ace file) has less than 95% consistency with the consensus sequence of the contig, or has more than 5 mismatched bases relative to the consensus, the read is flagged as a suspected chimera, provided it also shows evidence of either a polyA or polyT region.
The final step of the quality control process is to examine the flagged reads visually to find chimeric qualities. Chimeric reads are selected and removed based on their similarity to the consensus sequence and to the individual reads in the contig. A chimeric read may also be indicated if blast hits to different proteins are found to be adjacent in the read. The process of chimera detection and removal is often repeated numerous times before arriving at a finished assembly.
Sequence comparison and assignment into functional categories
Similarity searches were performed with the tblastx or blastx programs [65] against the TIGR Gene Indices available for Arabidopsis (AGI11), rice (OGI16) and pine (PGI5.0), retrieved from the TIGR web site [66] and against one Cycas EST assembly [67], retrieved from Sputnik web site [68]. Blast searches were conducted against several databases: the NCBI non redundant database (nr), the Uniref100 peptides set [39], and the Cell Wall Navigator Database [43]. HMM searches were conducted with the PFAM profiles (PFAM release16.0) with the local alignment setting since the spruce consensus sequences are fragmentary sequences. The Arabidopsis and rice coding sequences were downloaded from the TAIR web site [69] and the Rice Genome Annotation Database from TIGR [70], respectively.
To correlate the spruce consensus sequences to a Gene Ontology (GO) molecular function term, the annotations of homologous Uniref100 and Arabidopsis proteins were analysed. For each spruce consensus sequence, the blastx hits with a minimum similarity value of 0.75 and a minimum coverage of 0.5 were used in the GO assignment procedure. Similarity was defined as hsp positive/hsp alignment length (hsp : high scoring pair). Coverage was defined as the high scoring pair alignment length × 3/ query length. Among the retained hits, whenever a spruce sequence matched a protein with an associated GO term, this term was transferred to the spruce consensus sequence. Two GO annotation lists were completed: one including evidence codes Inferred from Electronic Annotation (IEA) evidence codes and one excluding IEA evidence.
Authors' contributions
NP, coordination of bioinformatics activities, data analysis, preparation of the manuscript; CP, LP, JC, JEJ, ER, sequence processing, assembly and annotation, web publishing and database development; MJM, JC, ASé, plant material production, library synthesis, and evaluation; EN, CGC, protein family sequence analyses; YB, SB, GY, JS, ASi, RH, MM, high-throughput EST sequencing and quality assurance; CP, JB, preparation of manuscript; JM, overall project supervision, preparation of manuscript.
Supplementary Material
Additional File 1
Description of tissues used for cDNA library synthesis: genotype, treatments (type, level and duration), organ, tissue and developmental stage.
Click here for file
Additional File 2
Annotation of proteins related to the cell wall based on similarities with sequences from the Cell Wall Navigator Database [44]and lignin biosynthesis enzymes [45]. Spruce homologs were identified by tblastx searches with e-value < 1e-10.
Click here for file
Acknowledgements
Funding for this work was provided by Genome Canada and Genome Québec to J.M., A. Sé. and J.B. for the project Arborea. We acknowledge the "Centre de Bioinformatique de l'Université Laval" for bioinformatics support. We also acknowledge H. Bérubé, S. Blais, C. Delisle, S. Forest, V. Roy for their technical assistance and B. Pelgas for her reading of the draft.
==== Refs
Ahuja MR Recent advances in molecular genetics of forest trees Euphytica 2001 121 173 195 10.1023/A:1012226319449
Dhillon SS Bonga JM, Durzan DJ DNA in tree species Cell and Tissue Culture in Forestry 1987 1 Martinus Nijhoff Publishers, Dordrecht 298 313
Wakamiya I Newton RJ Price JS Genome size and environmental factors in the genus Pinus Am J Bot 1993 80 1235 1241
Rake AW Miksche JP Hall RB Hanson KM DNA reassociation kinetics for four conifers Can J Genet Cytol 1980 22 69 79
Ohri D Khoshoo TN Genome size in gymnosperms Plant Syst Evol 1986 153 119 132 10.1007/BF00989421
Murray BG Nuclear DNA amounts in gymnosperms Ann Bot 1998 3 15 10.1006/anbo.1998.0764
Quackenbush J Cho J Lee D Liang F Holt I Karamycheva S Parvizi B Pertea G Sultana R White J The TIGR Gene Indices: analysis of gene transcript sequences in highly sampled eukaryotic species Nucleic Acids Res 2001 29 159 164 11125077 10.1093/nar/29.1.159
Kirst M Johnson AF Baucom C Ulrich E Hubbard K Staggs R Paule C Retzel E Whetten R Sederoff R Apparent homology of expressed genes from wood-forming tissues of loblolly pine (Pinus taeda L.) with Arabidopsis thaliana Proc Natl Acad Sci USA 2003 100 7383 7388 12771380 10.1073/pnas.1132171100
Pavy N Laroche J Bousquet J Mackay J Large-scale statistical analysis of secondary xylem ESTs in pine Plant Mol Biol 2005 57 203 224 15821878 10.1007/s11103-004-6969-7
Guillet-Claude C Isabel N Pelgas B Bousquet J The evolutionary implications of knox-I gene duplications in conifers: correlated evidence from phylogeny, gene mapping, and analysis of functional divergence Mol Biol Evol 2004 21 2232 2245 15317878 10.1093/molbev/msh235
Allona I Quinn M Shoop E Swope K St Cyr S Carlis J Riedl J Retzel E Campbell M Sederoff R Whetten RW Analysis of xylem formation in pine by cDNA sequencing Proc Natl Acad Sci USA 1998 95 9693 9698 9689143 10.1073/pnas.95.16.9693
Dendrome project
Treenomix project
Dubos C Plomion C Identification of water-deficit responsive genes in maritime pine (Pinus pinaster Ait.) roots Plant Mol Biol 2003 51 249 262 12602883 10.1023/A:1021168811590
Le Provost G Paiva J Pot D Brach J Plomion C Seasonal variation in transcript accumulation in wood-forming tissues of maritime pine (Pinus pinaster Ait.) with emphasis on a cell wall glycine-rich protein Planta 2003 217 820 830 12768425 10.1007/s00425-003-1051-2
Egertsdotter U van Zyl LM MacKay J Peter G Kirst M Clark C Whetten R Sederoff R Gene expression during formation of earlywood and latewood in loblolly pine: expression profiles of 350 genes Plant Biol 2004 6 654 663 15570469 10.1055/s-2004-830383
van Zyl L von Arnold S Bozhkov P Chen Y Egertsdotter U MacKay J Sederoff R Shen J Zelena L Clapham D Heterologous array analysis in Pinaceae: Hybridization of high density arrays of Pinus taeda cDNA with cDNA from needles and embryogenic cultures of P. taeda, P. sylvestris or Picea abies Function Compar Genomics 2002 3 306 318 10.1002/cfg.199
Stasolla C Belmonte MF van Zyl L Craig DL Liu W Yeung EC Sederoff R The effect of reduced glutathione on morphology and gene expression of white spruce (Picea glauca) somatic embryos J Exp Bot 2004 55 695 709 14966213 10.1093/jxb/erh074
Hertzberg M Aspeborg H Schrader J Andersson A Erlandsson R Blomqvist K Bhalerao R Uhlen M Teeri TT Lundeberg J Sundberg B Nilsson P Sandberg G A transcriptional roadmap to wood formation Proc Natl Acad Sci USA 2001 98 14732 14737 11724959 10.1073/pnas.261293398
Yang J Park S Kamdem DP Keathley DE Retzel E Paule C Kapur V Han KH Novel gene expression profiles define the metabolic and physiological processes characteristic of wood and its extractive formation in a hardwood tree species, Robinia pseudoacacia Plant Mol Biol 2003 52 935 956 14558656 10.1023/A:1025445427284
Paux E Tamasloukht M Ladouce N Sivadon P Grima-Pettenati J Identification of genes preferentially expressed during wood formation in Eucalyptus Plant Mol Biol 2004 55 263 280 15604680 10.1007/s11103-004-0621-4
Canadian Council of Forest Ministers
Florin R The distribution of conifer and taxad genera in time and space Acta Horti Bergiani 1963 20 121 312
Bouillé M Bousquet J Trans-species shared polymorphisms at orthologous nuclear gene loci among distant species in the conifer Picea (Pinaceae): Implications for the long-term maintenance of genetic diversity in trees Am J Bot 2005 92 63 73
Pelgas B Bousquet J Beauseigle S Isabel N A composite linkage map from two crosses for the species complex Picea mariana [Mill.] B.S.P x Picea rubens (Sarg.) and analysis of synteny with other Pinaceae Theor Applied Genetics 2005
Neale DB Savolainen O Association genetics of complex traits in conifers Trends Plant Sci 2004 9 325 330 15231277 10.1016/j.tplants.2004.05.006
Gene Ontology Consortium Creating the gene ontology resource: design and implementation Genome Res 2001 11 1425 1433 11483584 10.1101/gr.180801
SpruceDB
Vettore AL da Silva FR Kemper EL Souza GM da Silva AM Ferro MI Henrique-Silva F Giglioti EA Lemos MV Coutinho LL Nobrega MP Carrer H Franca SC Bacci Junior M Goldman MH Gomes SL Nunes LR Camargo LE Siqueira WJ Van Sluys MA Thiemann OH Kuramae EE Santelli RV Marino CL Targon ML Ferro JA Silveira HC Marini DC Lemos EG Monteiro-Vitorello CB Tambor JH Carraro DM Roberto PG Martins VG Goldman GH de Oliveira RC Truffi D Colombo CA Rossi M de Araujo PG Sculaccio SA Angella A Lima MM de Rosa Junior VE Siviero F Coscrato VE Machado MA Grivet L Di Mauro SM Nobrega FG Menck CF Braga MD Telles GP Cara FA Pedrosa G Meidanis J Arruda P Analysis and functional annotation of an expressed sequence tag collection for tropical crop sugarcane Genome Res 2003 13 2725 2735 14613979 10.1101/gr.1532103
Forment J Gadea J Huerta L Abizanda L Agusti J Alamar S Alos E Andres F Arribas R Beltran JP Berbel A Blazquez MA Brumos J Canas LA Cercos M Colmenero-Flores JM Conesa A Estables B Gandia M Garcia-Martinez JL Gimeno J Gisbert A Gomez G Gonzalez-Candelas L Granell A Guerri J Lafuente MT Madueno F Marcos JF Marques MC Martinez F Martinez-Godoy MA Miralles S Moreno P Navarro L Pallas V Perez-Amador MA Perez-Valle J Pons C Rodrigo I Rodriguez PL Royo C Serrano R Soler G Tadeo F Talon M Terol J Trenor M Vaello L Vicente O Vidal Ch Zacarias L Conejero V Development of a citrus genome-wide EST collection and cDNA microarray as resources for genomic studies Plant Mol Biol 2005 57 375 391 15830128 10.1007/s11103-004-7926-1
Kawai J Shinagawa A Shibata K Yoshino M Itoh M Ishii Y Arakawa T Hara A Fukunishi Y Konno H Adachi J Fukuda S Aizawa K Izawa M Nishi K Kiyosawa H Kondo S Yamanaka I Saito T Okazaki Y Gojobori T Bono H Kasukawa T Saito R Kadota K Matsuda H Ashburner M Batalov S Casavant T Fleischmann W Gaasterland T Gissi C King B Kochiwa H Kuehl P Lewis S Matsuo Y Nikaido I Pesole G Quackenbush J Schriml LM Staubli F Suzuki R Tomita M Wagner L Washio T Sakai K Okido T Furuno M Aono H Baldarelli R Barsh G Blake J Boffelli D Bojunga N Carninci P de Bonaldo MF Brownstein MJ Bult C Fletcher C Fujita M Gariboldi M Gustincich S Hill D Hofmann M Hume DA Kamiya M Lee NH Lyons P Marchionni L Mashima J Mazzarelli J Mombaerts P Nordone P Ring B Ringwald M Rodriguez I Sakamoto N Sasaki H Sato K Schonbach C Seya T Shibata Y Storch KF Suzuki H Toyo-oka K Wang KH Weitz C Whittaker C Wilming L Wynshaw-Boris A Yoshida K Hasegawa Y Kawaji H Kohtsuki S Hayashizaki Y RIKEN Genome Exploration Research Group Phase II Team and the FANTOM Consortium Functional annotation of a full-length mouse cDNA collection Nature 2001 409 685 690 11217851 10.1038/35055500
Whitfield Ch.W Band MR Bonaldo MF Kumar ChG Liu L Pardinas JR Robertson HM Soares MB Robinson GE Annotated expressed sequence tags and cDNA microarrays for studies of brain and behavior in the honey bee Genome Res 2002 12 555 566 11932240 10.1101/gr.5302
Brown GR Gill GP Kuntz RJ Langley CH Neale DB Nucleotide diversity and linkage disequilibrium in loblolly pine Proc Natl Acad Sci USA 2004 101 15255 15260 15477602 10.1073/pnas.0404231101
Pot D McMillan L Echt C Le Provost G Garnier-Géré P Cato S Plomion C Nucleotide variation in genes involved in wood formation in two pine species New Phytologist 2005 167 101 112 15948834 10.1111/j.1469-8137.2005.01417.x
Garcia-Gil MR Mikkonen M Savolainen O Nucleotide diversity at two phytochrome loci along a latitudinal cline in Pinus sylvestris Mol Ecol 2003 12 1195 1206 12694283 10.1046/j.1365-294X.2003.01826.x
Crow JA Diogenes – Reliable prediction of protein-encoding regions in short genomic sequences 2005
PFAM database
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 D138 D141 14681378 10.1093/nar/gkh121
Bairoch A Apweiler R Wu CH Barker WC Boeckmann B Ferro S Gasteiger E Huang H Lopez R Magrane M Martin MJ Natale DA O'Donovan C Redaschi N Yeh LL The Universal Protein Resource (UniProt) Nucleic Acids Res 2005 D154 D159 15608167
Rhee SY Beavis W Berardini TZ Chen G Dixon D Doyle A Garcia-Hernandez M Huala E Lander G Montoya M Miller N Mueller LA Mundodi S Reiser L Tacklind J Weems DC Wu Y Xu I Yoo D Yoon J Zhang P The Arabidopsis Information Resource (TAIR): a model organism database providing a centralized, curated gateway to Arabidopsis biology, research materials and community Nucleic Acids Res 2003 31 224 228 12519987 10.1093/nar/gkg076
Pine Gene Index PGI5.0 database
Davuluri RV Sun H Palaniswamy SK Matthews N Molina C Kurtz M Grotewold E AGRIS: Arabidopsis gene regulatory information server, an information resource of Arabidopsis cis-regulatory elements and transcription factors BMC Bioinformatics 2003 4 25 12820902 10.1186/1471-2105-4-25
Girke T Lauricha J Tran H Keegstra K Raikhel N The Cell Wall Navigator database. A systems-based approach to organism-unrestricted mining of protein families involved in cell wall metabolism Plant Physiol 2004 136 3003 3008 15489283 10.1104/pp.104.049965
Raes J Rohde A Christensen JH Van de Peer Y Boerjan W Genome-wide characterization of the lignification toolbox in Arabidopsis Plant Physiol 2003 133 1051 1071 14612585 10.1104/pp.103.026484
Li F Zhang D Fujise K Characterization of fortilin, a novel antiapoptotic protein J Biol Chem 2001 276 47542 47549 11598139 10.1074/jbc.M108954200
Pay A Heberle-Bors E Hirt H An alfalfa cDNA encodes a protein with homology to translationally controlled human tumor protein Plant Mol Biol 1992 19 501 503 1623194 10.1007/BF00023399
Sage-Ono K Ono M Harada H Kamada H Dark-induced accumulation of mRNA for a homolog of translationally controlled tumor protein (TCTP) in Pharbitis Plant Cell Physiol 1998 39 357 360 9588028
Campbell M Sederoff RR Variation in lignin content and composition Plant Physiol 1996 110 3 13 12226169
Peleman J Saito K Cottyn B Engler G Seurinck J Van Montagu M Inze D Structure and expression analyses of the S-adenosylmethionine synthetase gene family in Arabidopsis thaliana Gene 1989 84 359 369 2482229 10.1016/0378-1119(89)90510-6
Shen B Li C Tarczynski MC High free-methionine and decreased lignin content result from a mutation in the Arabidopsis S-adenosyl-l-methionine synthetase 3 gene Plant J 2002 29 371 380 11844113 10.1046/j.1365-313X.2002.01221.x
Sanchez-Aguayo I Rodriguez-Galan JM Garcia R Torreblanca J Pardo JM Salt stress enhances xylem development and expression of S-adenosyl-L-methionine synthase in lignifying tissues of tomato plants Planta 2004 220 278 285 15322882 10.1007/s00425-004-1350-2
Espartero J Pintor-Toro JA Pardo JM Differential accumulation of S-adenosylmethionine synthetase transcripts in response to salt stress Plant Mol Biol 1994 25 217 227 8018871 10.1007/BF00023239
Schröder G Eichel J Breining S Schröder J Three differentially expressed S-adenosylmethionine synthetase from Catharantus roseus : molecular and functional characterization Plant Mol Biol 1997 33 211 222 9037140 10.1023/A:1005711720930
Gramene database
Lindroth AM Saarikoski P Flygh G Clapham D Gronroos R Thelander M Ronne H von Arnold S Two S-adenosylmethionine synthetase-encoding genes differentially expressed during adventitious root development in Pinus contorta Plant Mol Biol 2001 46 335 346 11488480 10.1023/A:1010637012528
CCGB biodata database
Lamblin AF Crow JA Johnson JE Silverstein KA Kunau TM Kilian A Benz D Stromvik M Endre G VandenBosch KA Cook DR Young ND Retzel EF MtDB: a database for personalized data mining of the model legume Medicago truncatula transcriptome Nucleic Acids Res 2003 31 196 201 12519981 10.1093/nar/gkg119
Arborea project
Chang S Puryear J Cairney J A simple and efficient method for isolating RNA from pine trees Plant Mol Biol Rep 1993 11 113 116
Sambrook J Fritsch EF Maniatis T Molecular Cloning: A Laboratory Manual 1989 2 Plainview, NY: Cold Spring Harbor Laboratory Press
Yang GS Stott JM Smailus D Barber SA Balasundaram M Marra MA Holt RA High-throughput sequencing: a failure mode analysis BMC Genomics 2005 6 2 15631628 10.1186/1471-2164-6-2
Ewing B Hillier L Wendl MC Green P Base-calling of automated sequencer traces using phred. I. Accuracy assessment Genome Res 1998 8 175 185 9521921
Phrap software
National Center for Biotechnology Information
Altschul SF Madden TL Schäffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
The Institute for Genomic Research
Brenner ED Stevenson DW McCombie RW Katari MS Rudd SA Mayer KF Palenchar PM Runko SJ Twigg RW Dai G Martienssen RA Benfey PN Coruzzi GM Expressed sequence tag analysis in Cycas, the most primitive living seed plant Genome Biol 2003 4 R78 14659015 10.1186/gb-2003-4-12-r78
Sputnik database
The Arabidopsis Information Resource
Rice Genome Annotation Database
EMBOSS
|
16236172
|
PMC1277824
|
CC BY
|
2021-01-04 16:32:49
|
no
|
BMC Genomics. 2005 Oct 19; 6:144
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-144
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1481625577010.1186/1471-2164-6-148Research ArticleSpecificity and overlap in gene segment-defined antibody repertoires Arnaout Ramy A [email protected] Department of Pathology, Brigham and Women's Hospital, Boston, MA 02115 USA2 The Broad Institute, Cambridge, MA 02141 USA3 Program for Evolutionary Dynamics, Harvard University, Cambridge, MA 02138 USA2005 28 10 2005 6 148 148 15 4 2005 28 10 2005 Copyright © 2005 Arnaout; licensee BioMed Central Ltd.2005Arnaout; 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 several studies have sought to catalog the full suite of antibodies that humans naturally produce against single antigens or other specificities (repertoire). Here we analyze the properties of all sequenced repertoires in order to better understand the specificity of antibody responses. Specifically, we ask whether the large-scale sequencing of antibody repertoires might provide a diagnostic tool for detecting antigen exposure. We do this by examining the overlap in VH-, D-, and JH- segment usage among sequenced repertoires.
Results
We find that repertoire overlap in VH-, D-, and JH-segment use is least for VH segments and greatest for JH segments, consistent with there being more VH than JH segments in the human genome. We find that for any two antigens chosen at random, chances are 90 percent that their repertoires' VH segments will overlap by less than half, and 98 percent that their VDJH combinations will overlap by ≤10 percent. We ran computer simulations to test whether enrichment for specific VDJH combinations could be detected in "antigen-exposed" populations, and found that enrichment is detectable with moderate-to-high sensitivity and high specificity, even when some VDJH combinations are not represented at all in some test sets.
Conclusion
Thus, as large-scale sequencing becomes cost-effective for clinical testing, we suggest that sequencing an individual's expressed antibody repertoire has the potential to become a useful diagnostic modality.
==== Body
Background
The antigen-binding variable regions of antibody molecules draw combinatorially from a set of somatically encoded V, D, and J gene segments [1]. Mathematically, this strategy allows for ~6,000 possible heavy chain (subscript H) and ~300 possible light chain (subscript L) V(D)J combinations, for a total of ~1.8 million possible heavy-and-light chain pairings [2,3].
Much work in immunology and structural biology has gone into studying how antibody sequence and structure affect antigen specificity [1]. In each antibody, contact with the antigen is made by six short regions, three on each heavy and light chain. These are known as the complementarity-determining regions (CDRs). CDR1 and CDR2 lie entirely within the V segment, while CDR3 spans the D segment and flanking parts of V and J (in heavy chain; in light chain, which lacks a D segment, CDR3 spans the V-J junction). In general, heavy chain contributes more than light chain to antigen binding and specificity, and CDR3 contributes more than CDR1 and CDR2 [4]. Hence heavy chain VDJ (VDJH) segment usage is a major determinant of antigen specificity.
There are other determinants. The part of an antigen that an antibody binds is called an epitope; the part of an antibody that an epitope binds is called a paratope. Single antigens may have multiple epitopes, and single antibodies may have multiple paratopes [5,6]. Moreover, nontemplated nucleotide insertions and deletions at gene segment junctions, together with CDR hypermutation, expand antibody diversity and antigen binding possibilities far beyond what is available through V(D)J combinatorics alone [1]. Hence V(D)J segment choice and sequence-level modification provide coarse- and fine-tuning, respectively, for antigen specificity, but different V(D)J and sequence combinations may well bind the same antigen.
These considerations and substantial experimental data (summarized in [4]) argue against a strict one-to-one relationship between antibody sequence and antigen specificity. However, they do suggest the possibility that antigens may have signature antibody repertoires. Here a repertoire is defined as a set of antibodies, defined by gene segment usage, that is produced in a population of people against a given specificity. A specificity comprises a single epitope, a set of epitopes on a single antigen, or a set of antigens.
To date several studies have addressed this idea in particular instances by sequencing antibodies specific for particular antigens. In one such study, circulating B cells from seven infants vaccinated against Hemophilus influenzae type b (Hib) were affinity enriched aganst Hib capsular polysaccharide (PS); rearranged V(D)J heavy and light chain gene libraries were then constructed and screened for Hib PS-specific antibodies [7]. The antibodies recovered all used the same VH segment (VH3–23) and only two JH and two VL and JL segments, consistent with previous studies [8,9]. This is consistent with the pattern seen in natural antibody populations, allowing consideration of data from this in vitro "scrambling" approach.
Repertoires against other antigens have also been shown to have restricted segment usage, although the degree and pattern of restriction vary. For example, using a technique similar to that described for Hib PS, the repertoire against Streptococcus pneumoniae serotype 23F PS was found to be dominated by four VH segments, which account for 90 percent of the repertoire's observed VH diversity; four JH segments (93% of JH diversity); and two VL-kappa segments (93%) [10]. For comparison, the repertoire against S. pneumoniae serotype 6B PS was found to be dominated by three VH segments (93%) and three JH segments (98%), but was found to lack strong VL-kappa restriction (90% in six segments) [11]. Association patterns among segments and chains were also found to vary.
In all, repertoires for over a dozen antigens have been studied individually, with various aims and to various extents, mainly through enrichment and cloning or through screening of phage-display libraries [7,10-14]. The aim of the present study is to analyze these repertoires as a group in order to better understand the specificity of antibody responses. The practical goal is to explore the possibility that in the future, large-scale sequencing of antibodies in an individual may be used as a fingerprint, or "pan-scan," of that person's antigen exposure.
Results
We analyzed VDJH segment usage for the 16 best-represented natural human repertoires in the IMGT database (see Methods). These comprised 292 antibody sequences (mean, 18 sequences per repertoire; range, 8–41). Six repertoires were directed against infectious agents, while 10 were directed against autoimmune agents (Table 1 and Additional File 1).
Table 1 Repertoire composition
specificity sequences VH genes D genes JH genes VDJH combos
E. histolytica 9 7 5 2 7
HBsAg (HBV) 12 9 8 3 11
PS (S. pneumo 23F) 23 7 10 4 15
gp120 (HIV) 26 10 16 6 24
PS (S. pneumo 6B) 41 5 10 3 11
dsDNA (human) 8 7 6 4 8
MAG (human) 9 7 7 4 9
PL (human) 9 8 8 4 9
Fab (human) 11 9 7 3 10
factor VIII (human) 19 3 5 3 6
cardiolipin (human) 12 7 7 3 10
gpIIb/IIIa (human) 14 12 10 3 14
myosin (human) 14 12 10 5 14
RhD (human) 22 9 14 5 20
DNA (human) 22 12 13 4 20
TPO (human) 41 6 11 4 16
total 292 36 21 6 192
Number of sequences, gene segments, and gene segment combinations in the repertoires of all specificities in the data set. Specificities associated with infectious agents listed first; species names within parentheses (where necessary). Abbreviations: E. histolytica, Entameba histolytica; HBV, hepatitis B virus; PS, polysaccharide; S. pneumo, Streptococcus pneumoniae; ds, double-stranded; MAG, myelin-associated glycoprotein; PL, phospholipid; TPO, thyroid peroxidase.
Gene segment usage patterns
Genome-level diversity was well represented among the repertoires as a group. All but one (VH7) of the VH and D gene segment families were represented, and the majority of individual VH (78%), D (91%), and JH (100%) gene segments appeared in at least one sequence. VH and D gene families were represented about as often as in a previous study of healthy individuals [15] (p = 0.01 and 0.13, R2 = 0.78 and 0.96 for VH and D families, respectively), as were individual JH gene segments (p = 0.004, R2 = 0.94). However, individual VH gene segments were used more variably (p = 0.90, R2 = 0.25).
These observations are consistent with there being more than one VDJH combination used in antibodies with a given specificity (see below). They also suggest either that our set of repertoires is a good representation of at least the kinds [16] of antigen or antigen patterns encountered naturally, or conversely that B cell populations of the healthy individuals sampled in the previous study [15] comprise clones expanded against specificities similar to the ones included in our present analysis. These possibilities are not mutually exclusive.
Figure 1 shows VH, D, and JH segment usage and VDJH combination usage patterns for the repertoires of representative specificities. Some repertoires were peaked and narrow, suggesting few epitopes or immunodominance among the epitopes in their specificities, or little diversity among individuals for these specificities ("public" or "semi-public" repertoires; see Discussion). Other repertoires were flat and broad, suggesting many epitopes or codominance among the epitopes in their specificities, or greater diversity among individuals for these specificities. Details of VH, D, and JH segment usage for particular specificities have been discussed elsewhere (see references for specific sequences in IMGT, Table 1, and Additional File 1).
Figure 1 Gene segment use for representative repertoires. Repertoires for three specificities are shown: human coagulation factor VIII, Streptococcus pneumoniae serotype 6B capsular polysaccharide (PS), and S. pneumoniae ser. 23F PS. Each histogram shows the frequency distribution of VH gene segments, D segments, JH segments, and VDJH combinations. More peaked distributions indicate that the repertoire is VH, D, JH, or VDJH restricted. For example, the S. pneumoniae ser. 6B repertoire is 80% restricted to JH gene segment JH4.
The data did not allow conclusive generalization about whether or not, for a given repertoire, VH, D, and JH segments are combined randomly or with some bias. This is because the number of antibodies sequenced in a given repertoire was small (8–41 sequences) relative to the number of VDJH combinations that could in principle be constructed from the VH, D, and JH segments that appeared in that repertoire (~50–1,000 possibilities).
For only one of the 16 repertoires – the repertoire for thyroid peroxidase – was there a tight, statistically significant correlation between the observed frequencies of VDJH combinations and the frequencies that would be expected if segments were combined at random (p < 0.01; R2 = 0.85). The repertoire for S. pneumoniae strain 6B polysaccharide also showed a tight correlation, but this correlation fell short of statistical significance (p = 0.10; R2 = 0.95). No tight, statistically significant correlation was observed for any other repertoire. These findings are consistent with the conclusion that VH, D, and JH segments are not joined at random in at least 14 of these 16 repertoires, but more sequencing is needed to settle this issue.
Overlap in gene segment usage
From a practical perspective, for repertoires to serve as signatures for particular specificities, the overlap in gene segments or in V(D)J combinations among different repertoires must be low. To estimate this overlap quantitatively, we calculated the percent overlap between each pair of specificities in the data set (Fig. 2).
Figure 2 Overlap in segment use among repertoires. Overlap in (a) VH gene segment and (b) VDJH use among all repertoires. The percent overlap is grayscale-coded according to the key below each plot. In (b), the range at the lower end of the scale is expanded in order to show the four pairs with 11–20 percent overlap (see text). Abbreviations: ds-DNA, double-stranded DNA; gp120, HIV-1 gp120; Sp, Streptococcus pneumoniae serotype; fVIII, clotting factor VIII; HBsAg, HBV surface antigen; IIb/IIIa, glycoprotein IIb/IIIa; MAG, myelin-associated glycoprotein; PL, phospholipid; TPO, thyroid peroxidase. Species of origin are as in Table 1.
We found that for any two specificities picked at random from our set, the probability was 90 percent that their repertoires' VH gene segment usage overlapped by half or less (Fig. 2a, red tones). Adding D and JH segment information decreased the overlap markedly: of the 240 pairwise comparisons between different specificities in our data set, only four (1.7%) showed more than 10 percent overlap: between dsDNA and RhD (12%), thyroid peroxidase (TPO) and factor VIII (16%), TPO and phospholipid (11%), and phospholipid and integrin gpIIb/IIIa (11%) – all autoimmune specificities. Although not random in segment usage, autoimmune antibodies may share common features that result from impaired negative selection. Overall, for any two specificities chosen at random, the probability was 98.3 percent that their repertoires' VDJH combinations overlapped by 10 percent or less (Fig. 2b).
Given the large number of possible VDJH combinations (~6,000) and the relatively small size of the data set (292 sequences), it is reasonable to ask whether or not such a small amount of overlap is likely to occur by chance. Probability calculations show that it is not. The two most common human haplotypes allow a maximum of 5,244 and 6,348 possible functional VDJH combinations, respectively; the probability that the small amount of overlap observed in our data should arise by chance is p = 0.004 (0.4%) and 0.011 (1.1%) for these two haplotypes, respectively (see Methods). Note that nonrandom association among VH, D, and JH segments means that only a fraction of these 5,244 or 6,348 possible combinations are actually observed. The smaller the number of combinations, the higher the probability that repertoires will overlap by chance. Hence the small amount of overlap observed in the data is even less likely to be the result of chance than these calculations suggest. The probabilities are therefore upper limits.
If the specificities analyzed in this study are indeed representative of the specificities to which human beings are exposed (see above), this finding suggests that VDJH-defined sequences may be able to distinguish dependably among a wide variety of specificities.
Simulating detection
For repertoires to be of practical use, it must be possible to detect when certain VDJH combinations are present at a higher-than-background frequency. This may indicate, for example, prior or ongoing exposure to an infectious agent or the presence of a response to a vaccine [17]. Ideally detection should be possible even when this frequency is barely above background – that is, when the signal-to-noise ratio is low.
To test whether enrichment might be detectable, we ran computer simulations for each specificity. These were done briefly as follows (for details, see Methods). For each specificity, we assembled several sets of sequences that were each enriched for sequences of that specificity's repertoire. (The analogy is that each set of sequences corresponds to what might be obtained from a blood sample of an individual known to have a clinical history of that specificity.) The collection of these sets was our "reference collection" for the test (medically, the gold standard). The strategy was to see if test sets could be assigned as exposed or unexposed by comparing their patterns of VDJH combinations to the ones from the reference collection. If antibodies in a test set had a similar pattern and prevalence of VDJH combinations as those in the reference collection, the test set was assigned as "exposed." If the patterns were dissimilar, the test set was assigned as "unexposed." Assignment was performed with the aid of a computerized algorithm (see Methods).
We tested this approach for each specificity by seeing how well exposed and unexposed sets could be assigned. In clinical infections, B cells specific for an infectious agent rarely exceed 5–10 percent of the total B cell population. Therefore, as a conservative test, the sets in the reference collection had only 1–2.5 percent of their VDJH combinations purposely drawn from the repertoire for the given specificity. For example, in testing for exposure to HIV gp120, of 1,000 VDJH combinations determined for a set in the reference collection, only 10–25 would be guaranteed to be combinations that appeared in the HIV gp120 repertoire; the rest would be from the repertoires of S. pneumoniae serotype 6B PS, double-stranded (ds) DNA, and the other 14 specificities. Note that in this approach not all combinations are guaranteed to appear in any one set; however, the more frequently a combination appears in the repertoire – the higher its prevalence – the more likely (and more often) it is to appear in a given set. Also, the larger the reference collection, the more likely that less prevalent combinations will also appear in at least one set.
A training collection for each specificity was assembled comprising 10 exposed and 10 unexposed sets. An additional 50 test sets, whose exposed/unexposed status was known to us but not to the algorithm, were presented for assignment. Performance was measured by sensitivity and specificity (see Methods). Figure 3 shows results for two typical simulations. Sensitivity generally reached between 0.7 and 0.8 when exposure-specific antibodies/sequences were five percent of the total; specificity was higher (most likely due to false negatives in the sensitivity because of the small size of the reference sets). Sensitivity was improved by increasing the size of and enrichment in the sets in the reference collection. (Here the terms "sensitivity" and "specificity" are used in the epidemiological sense; see Methods.)
Figure 3 Detection of exposure to representative specificities. Representative plots of sensitivity and specificity for detecting exposure at various levels (f; see Methods for details): (a) Streptococcus pneumoniae serotype 6B and (b) S. pneumoniae ser. 26F. In general sensitivities reached between 0.7 and 0.8 when repertoires were enriched to at least 5 percent (f = 0.05), and specificities reached between 0.95 and 1.
Discussion
The majority of modern clinical tests assay for just one analyte at a time [18]. They determine the presence or absence of the analyte, and sometimes its quantity, but provide no information about other analytes. For example, a nucleic acid test for HIV-1 determines whether or not HIV-1 RNA is present in blood, and how much, but provides no information about, for example, the presence of antibodies to CMV. Although such tests are the mainstay of modern medicine, conceptually, they are limited to providing a "20 questions," yes-or-no approach to diagnosis.
The major exception is the standard culture-based method for diagnosing bacterial infections. In this method, the first step is to apply a clinical sample to standard culture media to see what grows [19]. This method is powerful in that it presupposes little about the identity of the bacteria: it can distinguish among many bacteria with a single test, and often reveals the presence of species that were clinically unexpected. Conceptually, this is an open-ended, "what-is-there" approach to diagnosis. It is of general interest in medicine to develop more diagnostic techniques that use this approach.
Antibodies play a crucial role in protective immunity and immunopathology, and also are important in surveillance against cancer [1]. The relationship between antibody gene sequence and epitope specificity is complex, but several studies have shown that certain gene segments and gene segment combinations are used preferentially against specific epitopes, antigens, or sets of antigens – what we here call "specificities" [7,10-14]. The identity and frequency of gene segments or combinations define antibody repertoires.
In this paper we have analyzed the growing, albeit limited, data that exists on VDJH combination defined repertoires to see whether they might one day provide an open-ended diagnostic for antigens to which a person has been exposed. For statistical confidence, we analyzed only those specificities for which at least eight antibodies have been sequenced and annotated for VH, D, and JH gene segment use. A similar amount of systematic data for immunoglobulin light chains and T cell receptors is still unavailable, and so the present analysis was limited to immunoglobulin heavy chains.
Our data set represented nearly every gene segment family, and at frequencies similar to those seen in two healthy individuals in a previous study [15]. One interpretation is that this reflects an intrinsic bias in the frequency with which different VDJH combinations are formed or expressed. Another interpretation is that the specificities in our data set are representative of the exposures that shape repertoires in healthy individuals, since certain types of antigens – bacterial polysaccharides, for instance – select for certain canonical structures in antibodies, and segments of the same gene family are more likely to produce similar structures [16]. These two interpretations are not mutually exclusive.
The narrowness or breadth of the repertoires for individual specificities (Fig. 2) could simply reflect the number of epitopes per specificity. For example, the antibodies against factor VIII, which formed a narrow repertoire, are known to have been raised against relatively well defined domains of factor VIII that comprise few epitopes [20], while antibodies against dsDNA, which formed a broad repertoire, were not raised this way [21].
The fact that the same VDJH combinations were recovered from multiple individuals in many repertoires (e.g., the S. pneumoniae PS repertoires [10,11]) suggests that despite genetic differences, different individuals may often use the same or at least overlapping sets of VDJH combinations in the antibodies they make against a given epitope. These could be called "public" or "semi-public" combinations [7,3]. Such commonalities might shed light on the evolutionary forces – repeat exposure to particular infectious agents, for example [4] – that may have shaped and maintained germline gene segment diversity. Further sequencing experiments using specificities defined at the epitope level would be useful to determine how often and to what epitopes public and semi-public combinations occur. The more frequent public combinations turn out to be, the more narrowly defined specificities can be and remain detectable, and vice versa.
Repertoires' VDJH combinations overlapped rarely (Fig. 3b), and less often than would be predicted by chance (p ≤ 0.011). Specifically, for any two specificities chosen at random, chances were 98.3 percent that they overlapped by 10 percent or less. This suggests that determining VDJH usage for a sampling of antibodies can be used to identify exposure to a particular antigen or set of antigens with reasonable specificity.
To further explore this idea, we conducted a set of simulation experiments to see whether individuals could one day be diagnosed as being exposed or not exposed to a given specificity (relative to a normal baseline) by assaying for enrichment of certain VDJH combinations. We show that even at modest levels of enrichment, which represents an increased frequency of B cells specific to a certain exposure, and using just 10 reference sets as the "gold standard" for exposure, assignment of unknown sets as either exposed or unexposed was possible with a high degree of sensitivity and specificity. In principle, such a sequence-based method has the advantage of being able to detect patterns of exposure even when the specificity of the antibodies or the identity of the offending agent is completely unknown. This "open-ended" approach is most useful for the early detection of emerging diseases, and will become practicable as improvements in sequencing technology make it possible to use in the clinic [22]. Data on antibody titers and functionality will doubtless add to the utility of this approach.
Conclusion
In sum, this study is the first to our knowledge that investigates the relationship between antibody specificity and VDJH segment usage for a large number of sequenced antibodies. Further sequencing studies should make it possible to refine the conclusions presented here, and also to assess the contribution of light chain in antibodies and of alpha and beta chains in T cell receptors to antigen specificity in human immune responses. Whether or not large-scale sequencing will prove useful as a future diagnostic tool will depend on these further studies.
Methods
Antibody repertoire data
The ImMunoGeneTics database (IMGT; ) is a publically available curated online repository of ~88,000 sequenced immunoglobulin and T cell receptor genes from a number of species [2]. We extracted all ~531 entries that contained recombined human immunoglobulin genes annotated with VH, D, and JH gene segment use and antigen specificity.
To approximate only natural repertoires, we limited our analysis to sequences isolated from B cells of individuals, and excluded all sequences that had been designed or modified in vitro. Allowed sequences included ones obtained from Epstein-Barr Virus (EBV)-immortalized B cells, through combinatorial cloning or phage-display libraries constructed from B cells of antigen-exposed patients, and from single sequenced B cells. For statistical power we considered only those specificities that had at least eight sequences in IMGT. There were 16 such specificities, comprising a total of 292 individual antibody sequences (mean, 18 sequences per specificity; range, 8–41).
Frequency distributions and overlap
We calculated and tabulated VH, D, and JH frequency distributions from all specificities and calculated their pairwise overlap computationally. Because specificities generally differed in the number of unique VDJH combinations in their repertoires, overlap was not symmetric: for example, if one specificity's repertoire comprised five different VDJH combinations, and another specificity's repertoire had those same five combinations as well as an additional 15, the overlap would be 100 percent in one direction, but only 25 percent in the other.
Student's t-test for two independent samples was used to obtain p-values for calculated vs. observed frequencies of VDJH combinations for each repertoire. Heatmap plots were made using R .
Humans most commonly encode 38 functional VH genes, 23 functional D genes, and 6 functional JH genes, as well as a number of pseudogenes [2]. These allow for a theoretical maximum of 38 × 23 × 6 = 5,244 possible VDJH combinations. In addition, many Caucasians contain a partial duplication of the VH region that results in 46 functional VH genes [2]; this partial duplication allows for a theoretical maximum of 46 × 23 × 6 = 6,348 VDJH combinations. For a person with a maximum of 5,244 possible VDJH combinations, the probability that two sets of 10 randomly chosen combinations will not overlap at all is approximately [(5,244 - 10)/5,244]10 = 0.98, or 98 percent. The probability that a third set of, for example, seven combinations will not overlap at all with either of these two sets is approximately [(5,244 - 10)/5,244]10 × [(5,244 - 10 - 10)/5,244]7 = 0.96, or 96 percent. The probability of overlap among any group of sets may be approximated by extending this method.
Simulations
We built a pattern-detecting computer algorithm for detecting enrichment of antibody sequences that correspond to particular specificities [23] (single-hidden layer, feed-forward neural networks with backpropagation; Brainstem v1.4; ).
For each specificity, the algorithm was trained on a reference collection representing 10 exposed and 10 unexposed sequence sets as follows. Each set constituted a list of the frequency of each of 100 VDJH combinations drawn equally from all the specificities in the data set. For each specificity, we sampled a fraction (f) of combinations from that specificity's repertoire according to their frequency distribution, allowing resampling. The remainder were sampled from all sequences in the data set, including those of the chosen repertoire, again allowing resampling. This remainder represents a background of noise against which a signal – enrichment of specific sequences – might be detected. 0 <f ≤ 1 for exposed sets and f = 0 for unexposed sets. Hence, a set is "exposed" if it is statistically enriched for sequences a particular specificity, and "unexposed" otherwise. Note that unexposed sets will contain some sequences from the chosen exposed repertoire by chance, just at lower frequency than in exposed sets. Our question was, how well can we assign, or "diagnose," exposure: i.e., how well can we detect enrichment.
The algorithm was used to evaluate test sets, each comprising an additional 25 exposed and 25 unexposed patients. For each specificity, the algorithm was trained at 0.01 ≤ f ≤ 0.025 (for the exposed patients) and tested over the range 0.01 ≤ f ≤ 1. To quantify the results, we calculated the sensitivity [(true positives)/(true positives + false negatives)] and specificity [(true negatives)/(true negatives + false positives)] of the algorithm for each test set. These are standard metrics for diagnostic tests in the clinical setting [18].
List of abbreviations
dsDNA, double-stranded DNA; HIV-1, human immunodeficiency virus type 1; Sp, Streptococcus pneumoniae serotype; fVIII, clotting factor VIII; HBsAg, hepatitis B virus surface antigen; IIb/IIIa, glycoprotein IIb/IIIa; MAG, myelin-associated glycoprotein; PL, phospholipid; TPO, thyroid per-oxidase; CMV, cytomegalovirus.
Authors' contributions
R.A.A. performed all the work presented in this paper.
Supplementary Material
Additional File 1
VDJtable.pdf, is a PDF file that contains a table listing VDJ combinations for all specificities analyzed in this paper.
Click here for file
Acknowledgements
The author would like to thank Miguel N. Rivera and Eric S. Lander for helpful conversations.
==== Refs
Abbas AK Pober JS Lichtman AH Cellular and Molecular Immunology Saunders 2001
Lefranc MP IMGT, the international ImMunoGeneTics information system Novartis Found Symp 2003 254 126 36 discussion 136–42, 216–22, 250–2 14712935
Ohlin M Zouali M The human antibody repertoire to infectious agents: implications for disease pathogenesis Mol Immunol 2003 40 1 11 12909126 10.1016/S0161-5890(03)00099-3
Frank SA Immunology and Evolution of Infectious Disease Princeton 2002
James LC Roversi P Tawfik DS Antibody multispecificity mediated by conformational diversity Science 2003 299 1362 7 12610298 10.1126/science.1079731
Richards FF Konigsberg WH Rosenstein RW Varga JM On the specificity of antibodies Science 1975 187 130 7 46122
Lucas AH McLean GR Reason DC O'Connor AP Felton MC Moulton KD Molecular ontogeny of the human antibody repertoire to the Haemophilus influenzae type B polysaccharide: expression of canonical variable regions and their variants in vaccinated infants Clin Immunol 2003 108 119 27 12921758 10.1016/S1521-6616(03)00094-9
Pinchuk GV Nottenburg C Milner EC Predominant V-region gene configurations in the human antibody response to Haemophilus influenzae capsule polysaccharide Scand J Immunol 1995 41 324 30 7899819
Adderson EE Shackelford PG Quinn A Wilson PM Cunningham MW Insel RA Carroll WL Restricted immunoglobulin VH usage and VDJ combinations in the human response to Haemophilus influenzae type b capsular polysaccharide. Nucleotide sequences of monospecific anti-Haemophilus antibodies and polyspecific antibodies cross-reacting with self antigens J Clin Invest 1993 91 2734 43 8514881
Zhou J Lottenbach KR Barenkamp SJ Lucas AH Reason DC Recurrent variable region gene usage and somatic mutation in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae type 23F Infect Immun 2002 70 4083 91 12117915 10.1128/IAI.70.8.4083-4091.2002
Zhou J Lottenbach KR Barenkamp SJ Reason DC Somatic hypermutation and diverse immunoglobulin gene usage in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae Type 6B Infect Immun 2004 72 3505 14 15155658 10.1128/IAI.72.6.3505-3514.2004
Chardes T Chapal N Bresson D Bes C Giudicelli V Lefranc MP Peraldi-Roux S The human anti-thyroid peroxidase autoantibody repertoire in Graves' and Hashimoto's autoimmune thyroid diseases Immunogenetics 2002 54 141 57 12073143 10.1007/s00251-002-0453-9
Sun Y Park MK Kim J Diamond B Solomon A Nahm MH Repertoire of human antibodies against the polysaccharide capsule of Streptococcus pneumoniae serotype 6B Infect Immun 1999 67 1172 9 10024557
Uray K Hudecz F Fust G Prohaszka Z Comparative analysis of linear antibody epitopes on human and mycobacterial 60-kDa heat shock proteins using samples of healthy blood donors Int Immunol 2003 15 1229 36 13679392 10.1093/intimm/dxg122
Brezinschek HP Foster SJ Brezinschek RI Dorner T Domiati-Saad R Lipsky PE Analysis of the human VH gene repertoire. Differential effects of selection and somatic hypermutation on human peripheral CD5(+)/IgM+ and CD5(-)/IgM+ B cells J Clin Invest 1997 99 2488 501 9153293
Vargas-Madrazo E Lara-Ochoa F Almagro JC Canonical structure repertoire of the antigen-binding site of immunoglobulins suggests strong geometrical restrictions associated to the mechanism of immune recognition J Mol Biol 1995 254 497 504 7490765 10.1006/jmbi.1995.0633
Plotkin SA Orenstein WA Offit PA Eds Vaccines Saunders 2003
McClatchey KD Clinical Laboratory Medicine 2002 2 Lippincott Williams & Wilkins
Murray PR Baron EJ Jorgensen JH Pfaller MA Yolken RH Eds Manual of Clinical Microbiology 2003 American Society of Microbiology
van den Brink EN Bril WS Turenhout EAM Zuurveld M Bovenschen N Peters M Yee TT Mertens K Lewis DA Ortel TL Lollar P Scandella D Voorberg J Two classes of germline genes both derived from the V(H)1 family direct the formation of human antibodies that recognize distinct antigenic sites in the C2 domain of factor VIII Blood 2002 99 2828 34 11929772 10.1182/blood.V99.8.2828
Behrendt M Partridge LJ Griffiths B Goodfield M Snaith M Lindsey NJ The role of somatic mutation in determining the affinity of anti-DNA antibodies Clin Exp Immunol 2003 131 182 9 12519403 10.1046/j.1365-2249.2003.02026.x
Shendure J Mitra RD Varma C Church GM Advanced sequencing technologies: methods and goals Nat Rev Genet 2004 5 335 44 15143316 10.1038/nrg1325
Fausett LV Fundamentals of Neural Networks 1994 Prentice-Hall
|
16255770
|
PMC1277825
|
CC BY
|
2021-01-04 16:32:49
|
no
|
BMC Genomics. 2005 Oct 28; 6:148
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-148
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-881623617710.1186/1471-2334-5-88Research ArticleA multicenter trial of the efficacy and safety of tigecycline versus imipenem/cilastatin in patients with complicated intra-abdominal infections [Study ID Numbers: 3074A1-301-WW; ClinicalTrials.gov Identifier: NCT00081744] Oliva María E [email protected] Arcot [email protected] Albert [email protected] Jacyr [email protected] Maria [email protected] Gilbert M [email protected] Timothy [email protected] Evelyn J [email protected] Evan [email protected] 301 Study Group 1 Hospital San Martin, Provincia de Entre Rios, Argentina2 Sri Ramachandra Medical College and Research Institute, Tamil Nadu, India3 LAC-USC Medical Center, Los Angeles, California. USA4 Real E Benemérita Sociedade Portuguesa de Beneficéncia, Hospital Säo Jaoquim, São Paulo/SP, Brazil5 Hospital de Urgencia Asistencia Publica, Santiago, Chile6 Clinical Research Group, Wyeth Research, Collegeville, Pennsylvania, USA2005 19 10 2005 5 88 88 12 4 2005 19 10 2005 Copyright © 2005 Oliva et al; licensee BioMed Central Ltd.2005Oliva 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
Complicated intra-abdominal infections (cIAI) remain challenging to treat because of their polymicrobial etiology including multi-drug resistant bacteria. The efficacy and safety of tigecycline, an expanded broad-spectrum glycylcycline antibiotic, was compared with imipenem/cilastatin (IMI/CIS) in patients with cIAI.
Methods
A prospective, double-blind, multinational trial was conducted in which patients with cIAI randomly received intravenous (IV) tigecycline (100 mg initial dose, then 50 mg every 12 hours [q12h]) or IV IMI/CIS (500/500 mg q6h or adjusted for renal dysfunction) for 5 to14 days. Clinical response at the test-of-cure (TOC) visit (14–35 days after therapy) for microbiologically evaluable (ME) and microbiological modified intent-to-treat (m-mITT) populations were the co-primary efficacy endpoint populations.
Results
A total of 825 patients received ≥ 1 dose of study drug. The primary diagnoses for the ME group were complicated appendicitis (59%), and intestinal (8.8%) and gastric/duodenal perforations (4.6%). For the ME group, clinical cure rates at TOC were 80.6% (199/247) for tigecycline versus 82.4% (210/255) for IMI/CIS (95% CI -8.4, 5.1 for non-inferiority tigecycline versus IMI/CIS). Corresponding clinical cure rates within the m-mITT population were 73.5% (227/309) for tigecycline versus 78.2% (244/312) for IMI/CIS (95% CI -11.0, 2.5). Nausea (31.0% tigecycline, 24.8% IMI/CIS [P = 0.052]), vomiting (25.7% tigecycline, 19.4% IMI/CIS [P = 0.037]), and diarrhea (21.3% tigecycline, 18.9% IMI/CIS [P = 0.435]) were the most frequently reported adverse events.
Conclusion
This study demonstrates that tigecycline is as efficacious as imipenem/cilastatin in the treatment of patients with cIAI.
==== Body
Background
Complicated intra-abdominal infections are characterized as local or systemic infections secondary to a physical perforation in the gastrointestinal tract or via a necrotic gut wall into the peritoneal space, leading to abscess formation or peritonitis [1]. These infections require a combination of appropriate and timely surgical source control and broad spectrum antimicrobial therapy for optimal outcome. Nearly all intra-abdominal infections are caused by multiple microorganisms resident in the gastrointestinal tract; these include aerobes and facultative and obligate anaerobes [2], with Enterobacteriaceae (eg, Escherichia coli) isolated most frequently [1,3]. Although isolation of enterococci from an intra-abdominal source were once suggestive of normal flora, these bacteria are now recognized as true pathogens, with upwards of one third of intra-abdominal cultures yielding enterococci [2]. In fact, the isolation of Enterococcus spp. from an intra-abdominal focus of infection has been linked with treatment failure [4].
Treatment of complicated intra-abdominal infections remains a challenge, primarily because of their polymicrobial etiology coupled with the high risk of complications and death. Because frequently recovered isolates may possess multiple resistance factors (eg, extended spectrum beta-lactamases [ESBLs]) that express antimicrobial resistance, empiric antimicrobial therapy should have anticipated activity against these difficult-to-treat isolates [1,5]. As such, combination antibiotic therapy has often been a standard of care for treatment of these infections [1]. The recent 2003 guidelines of the Infectious Diseases Society of America (IDSA) advocates broad-spectrum single or combination therapy (eg, carbapenem or piperacillin/tazobactam monotherapy, third- or fourth-generation cephalosporins or fluoroquinolones plus metronidazole) for high-risk patients with severe or postoperative nosocomial intra-abdominal infections wherein polymicrobial infection and/or resistant flora are more prevalent [1]. When very resistant bacteria are suspected (eg, vancomycin-resistant Enterococcus, methicillin-resistant Staphylococcus aureus, Pseudomonas aeruginosa), however, a complex multidrug regimen is recommended [1]. The initial selection of antimicrobial therapy for treatment of intra-abdominal infections is extremely important because inappropriate empiric antimicrobial therapy has been associated with delayed clinical resolution, increased length of hospital stay, and an increased risk of mortality [6,7]. Adequate surgical source control is also an important determinant of outcome; insufficient drainage and repair may compromise the effectiveness of antibiotic therapy [1].
Tigecycline is a novel, first-in-class, glycylcycline antibiotic with expanded broad-spectrum wide in vitro activity against the microorganisms commonly encountered in intra-abdominal infections. Specifically, tigecycline's spectrum of in vitro activity includes aerobic and facultative gram-positive and gram-negative bacteria and anaerobic bacteria [8-11]. Tigecycline also provides in vitro activity against antibiotic-resistant bacteria such as vancomycin-resistant Enterococcus faecalis and E. faecium, ESBL-producing enteric gram-negative bacteria, and methicillin-resistant S. aureus [8-16]. The primary objective of this multicenter trial was to evaluate the efficacy and safety of tigecycline monotherapy compared with imipenem/cilastatin in the treatment of hospitalized adult patients with complicated intra-abdominal infections. A second goal of the study was to evaluate the in vitro susceptibility of tigecycline against common bacteria implicated as causes of intra-abdominal infection.
Methods
Study design and enrollment criteria
This was a phase 3, multicenter, double-blind (third-party unblinded) trial of adult patients who were candidates for or had undergone a laparotomy, laparoscopy, or percutaneous drainage of an intra-abdominal abscess and had a known or suspected diagnosis of complicated intra-abdominal infection. All patients were hospitalized at the time of study entry. Before screening of the first patient, the protocol was reviewed and approved by the institutional review board or ethical review committee at each participating center. Written informed consent was obtained from each patient or his or her legal representative before the start of any study procedures. The trial was conducted in accordance with the Declaration of Helsinki.
Inclusion criteria
Men and women were eligible for inclusion if they were 18 years of age or older and required a surgical procedure for a complicated intra-abdominal infection. Complicated intra-abdominal infections included conditions such as an intra-abdominal abscess (including liver and spleen) that developed in a postsurgical patient after receiving standard antibacterial therapy (ie, at least 48 hours, but note more than 5 days of antibiotics); appendicitis complicated by perforation and/or a periappendiceal abscess; perforated diverticulitis complicated by abscess formation or fecal contamination; complicated cholecystitis with evidence of perforation, empyema, or gangrene; perforation of a gastric or duodenal ulcer with symptoms exceeding 24 hours; purulent peritonitis or peritonitis associated with fecal contamination; or perforation of the large or small intestine with abscess or fecal contamination. In addition, patients could not have received more than 1 dose of an antibiotic (single broad-spectrum agent or 1 dose of each antibiotic in a combination regimen such as metronidazole, ampicillin, gentamicin) after the baseline intra-abdominal culture was obtained from the infected site.
Exclusion criteria
Patients were not allowed to participate if they had any concomitant condition that precluded evaluation of a response or made it unlikely that the planned course of therapy could be completed. Other primary reasons for ineligibility included the following: preoperative suspicion of a diagnosis of spontaneous bacterial peritonitis, simple cholecystitis, gangrenous cholecystitis without rupture, simple appendicitis, acute suppurative cholangitis, pancreatic abscess, or infected necrotizing pancreatitis; Acute Physiologic and Chronic Health Evaluation (APACHE) II score greater than 30; active or treated leukemia or systemic malignancy within the prior 3 months or metastatic malignancy to the abdomen within the prior 6 months; known acquired immunodeficiency syndrome (AIDS); presence of any uncontrolled central nervous system disease; pregnant or breastfeeding women; known or suspected hypersensitivity to either study drug or to related compounds; concomitant ganciclovir therapy; significant hepatic disease (ie, aspartate aminotransferase [AST] or alanine aminotransferase [ALT] level > 10 times the upper limit of normal [ULN] or total bilirubin value > 3 times the ULN) or acute hepatic failure or acute decompensation of chronic hepatic failure; significant renal disease (ie, calculated creatinine clearance < 41 mL/min/1.73 m2 after adequate hydration); neutropenia with absolute neutrophil count < 1000/mm3, with counts as low as 500/mm3 permitted if due to the acute infectious process; current intra-abdominal infection known to be caused by one or more bacterial isolates not susceptible to either of the study drugs (eg, P. aeruginosa, Proteus mirabilis); surgical procedure requiring that fascia or deep muscular layers be left open or expectation of planned abdominal re-exploration either in or out of the operating room; and administration of intraoperative antibacterial irrigants or peritoneal antibacterial agents (eg, irrigants, antibiotic-impregnated sponges). Any patient requiring additional systemic antibacterial therapy, for any reason, was not allowed to participate in the trial.
Antimicrobial regimens
Patients were stratified at randomization into 2 groups based on their scores on APACHE II: ≤15, or >15 but <31. Using a 1:1 ratio, patients were randomly assigned to receive either tigecycline (initial 100-mg dose given by intravenous [IV] infusion over a 30-minute period, followed by 50 mg IV every 12 hours) or IV imipenem/cilastatin (500 mg/500 mg every 6 hours or dose-adjusted based on weight and creatinine clearance). Patients randomized to tigecycline received a 100 mL normal saline intravenous infusion 6 hours after active drug each day in order to maintain the blind. Unless the patient was a clinical failure (see definition below), the duration of study drug therapy ranged from 5 to 14 days.
Study drug was administered only when there was a strong suspicion (ie, elevated white blood cell count, elevated band cell counts [ie, evidence of a "shift to the left"], fever, or highly suggestive radiographic findings) or a confirmed diagnosis of an intra-abdominal infection (presence of pus within the abdominal cavity), and a baseline intra-abdominal culture was obtained from the site of infection. Patients could be enrolled before drainage of the intra-abdominal infection and may have received up to 2 doses of study drug before the baseline cultures were obtained. Patients did not receive more than 1 dose (or combination) of parenteral nonstudy antibacterial drugs after the baseline intra-abdominal cultures were obtained. However, wound irrigation solutions of sterile water or normal saline and topical antiseptics were permitted throughout the course of the study.
Clinical evaluations
The clinical status of the intra-abdominal infection was assessed at serial visits throughout the study by the presence or absence of the following signs and symptoms: fever; localized or diffuse abdominal wall rigidity or involuntary guarding; abdominal tenderness or pain; ileus or hypoactive bowel sounds; nausea or vomiting. The clinical response to study drug was determined by the investigator. At the test-of-cure visit (14–35 days after therapy), each patient's response was categorized as one of the following: Cure – the course of study drug and the initial intervention (operative and/or radiologically guided drainage procedure) resolved the intra-abdominal infectious process; Failure – the patient required additional antibacterial therapy other than the study drug, the patient required additional surgical or radiologic intervention to cure the infection, death due to infection occurred after 48 hours of therapy, the patient received an extended course of study drug (ie, >120% of the planned number of doses), or the patient was prematurely discontinued from study drug due to an adverse event (after receiving at least 8 doses in 5 days) and required additional antibiotic therapy or surgical intervention; and Indeterminate – the patients was lost to follow-up, or died within 48 hours after the first dose of study drug for any reason, or died after 48 hours because of noninfectious-related reasons (as judged by the investigator).
Microbiologic evaluations
Baseline aerobic and anaerobic cultures from the primary intra-abdominal site of infection and two sets of blood cultures were obtained within 24 hours of the first dose of study drug. All aerobic and anaerobic bacterial isolates, regardless of the source of cultured material, were identified and tested at a central laboratory (Covance Central Laboratory Services, Inc., Indianapolis, IN, or Geneva, Switzerland) by using a standard procedure approved by the National Committee of Clinical Laboratory Standards (NCCLS) Subcommittee on Antimicrobial Susceptibility Testing. For tigecycline, provisional minimum inhibitory concentration (MIC) breakpoints were used (susceptible ≤2 mg/L; intermediate 4 mg/L; resistant ≥8 mg/L).
Based on the results of the baseline intra-abdominal culture, the susceptibilities of identified organisms, and the clinical outcome of the patient, the investigator also determined the microbiologic response at the patient level and at the isolate level. Microbiologic response by patient was categorized at the test-of-cure visit as eradication, persistence, superinfection (ie, the emergence of a new isolate was documented at the site of infection with worsening signs and symptoms of infection). The microbiologic response for each baseline isolate at the test-of-cure visit was described according to the following definitions: eradication, persistence, or indeterminate. Because many patients did not have follow-up cultures, many microbiologic responses both at the patient and isolate level were categorized as either presumed eradication or presumed persistence.
Safety/tolerability assessments
All patients who received at least one dose of study drug were evaluated for safety (modified intent-to-treat [mITT] population). Safety was assessed from serial medical history and physical examinations, reports of clinical adverse events, and findings from routine electrocardiograms (ECGs), and serum chemistry, hematology, coagulation, and urinalysis tests. Adverse events were recorded throughout the study period, up to and including the test-of-cure visit. Before unblinding, the investigator categorized the severity of each adverse event and the potential for relationship to study drug. Serious adverse events (ie, those that were life-threatening, led to prolongation of the existing hospitalization, caused persistent or significant disability or incapacity, or death) were also recorded.
Analysis populations
Several subpopulations of patients were assessed for safety, clinical, and bacteriologic outcomes. Patients who satisfied the inclusion/exclusion criteria were included in the intent-to-treat (ITT) population, whereas the subset of patients who received at least 1 dose of study drug made up the mITT population. Those patients in the mITT population who had clinical evidence of a complicated intra-abdominal infection, by meeting the minimal disease criteria, and had a confirmed baseline isolate made up the microbiological-modified (m-mITT) population. From this latter group, the microbiologically evaluable (ME) population was defined as those who met all inclusion/exclusion criteria; had at least 5 days of therapy; did not receive concomitant antibiotics after the baseline intra-abdominal culture was obtained through the test-of-cure visit; had a test-of-cure visit 14 to 35 days after the first dose of study drug; and had a baseline intra-abdominal culture containing at least one causative isolate that was susceptible to both study drugs. If these criteria were not met at any time during the study, the patient was declared non-evaluable and the outcome of cure/failure/indeterminate was analyzed within the m-mITT population. Patients were considered nonevaluable for inclusion in the ME population if death occurred or if they withdrew from the study <48 hours after the first dose of study drug.
Statistical analysis
The primary endpoints of the study were clinical response at the test-of-cure visit (14–35 days after therapy) for the m-mITT and ME populations. Secondary analyses included bacteriologic response at the test-of cure visit by patient and isolate, as well as clinical response rates stratified as monomicrobial versus polymicrobial, and by isolate.
Statistical analysis was performed by the Clinical Biostatistics department of Wyeth Research, Collegeville, PA. Categorical baseline demographic and medical variables were analyzed using the Fisher exact test. Continuous variables were compared using a one-way analysis of variance (ANOVA) model with treatment as a factor. Between-group comparisons of adverse events were analyzed by using the Fisher exact test. For laboratory tests, vital signs, and ECG results, within-group changes from baseline were analyzed by using a paired t-test and between-group comparisons were made by using the analysis of covariance, adjusting for baseline value. The difference between treatment groups in the percentage of premature withdrawal from study drug was evaluated by using a 2-sided Fisher exact test.
The noninferiority efficacy of tigecycline compared with imipenem/cilastatin was evaluated for clinical and microbiologic responses by using a 2-sided 95% confidence interval (CI) for the true difference in efficacy (tigecycline minus imipenem/cilastatin) adjusted for the stratification variable APACHE II score and corrected for continuity. Noninferiority was concluded if the lower limit of the 2-sided 95% CI was greater than or equal to -15%. For all subpopulation analyses (eg, monomicrobial versus polymicrobial infection), an adjusted difference between treatment groups with its 95% CI was calculated from a generalized linear model with a binomial probability function and an identity link (SAS® Proc GENMOD). Interaction effects were tested at the 0.10 level of significance. With the planned sample size (n = 788) and an evaluability rate of 50%, the trial had a power of at least 90% to determine the noninferiority of tigecycline compared with imipenem/cilastatin.
Results
Eight hundred ninety-eight (898) patients were screened for study participation at 96 sites in 17 countries in the United States, Canada, Europe, Latin America, India, and Asia from November 2002 to August 2004. Of these, 64 patients did not meet protocol requirements (Figure 1). The remaining 834 patients were randomized in a 1:1 ratio to one of the two treatment regimens and represented the ITT population; however, 9 patients never received study drug. Accordingly, 825 patients (413 tigecycline, 412 imipenem/cilastatin) comprised the mITT (safety) population. The majority of the mITT population (98%; 807 of 825) had clinical evidence of a complicated intra-abdominal infection (clinical mITT population). Within this latter cohort, 692 patients were clinically evaluable (clinically evaluable [CE] population). One hundred thirty three (133; 16.1%) mITT patients (72 tigecycline, 61 imipenem/cilastatin) were not included in the CE population for the following primary reasons (patients could have been excluded for more than one reason): no clinical evaluation at the test-of-cure visit (n = 47); entry criteria not met (n = 28); blind broken (n = 22); and received more than 1 dose of a nonstudy antibiotic after pretherapy culture (n = 12). From the mITT population, 621 of 825 (75%) patients had a pretherapy isolate isolated and comprised the m-mITT population. A total of 502 m-mITT patients (247 tigecycline, 255 imipenem/cilastatin) met both clinical evaluability criteria and had a pretherapy isolate isolated from an intra-abdominal source (ME population).
Figure 1 Patient disposition and analysis population.
Demographic/baseline medical characteristics
The demographic characteristics for the 502 ME patients were comparable between the two treatment groups (Table 1). The study population was of mixed racial/ethnic background with whites (41.8%) and Hispanics (19.5%) represented most often. There was a predominance of men (67.5%) and the mean age of enrolled patients was 43 years old. Complicated appendicitis (59%) was the most common intra-abdominal infection diagnosis, followed by perforated intestine (8.8%) and gastric/duodenal ulcer (4.6%). No significant differences between the treatment groups were observed in the number or types of infections diagnosed at baseline. The severity of intra-abdominal illness was similar in each treatment group (mean APACHE II score was ~5.7).
Table 1 Demographic and baseline medical characteristics (ME population)
Tigecycline N = 247 Imipenem/Cilastatin N = 255
Mean ± SD age, years 42.9 ± 18.0 43.1 ± 17.6
Sex, n (%) male 173 (70.0) 166 (65.1)
Ethnic origin, n (%)
White 104 (42.1) 106 (41.6)
Black 16 (6.5) 25 (9.8)
Asian 30 (12.1) 30 (11.8)
Hispanic 54 (21.9) 44 (17.3)
Other 43 (17.4) 50 (19.6)
Mean ± SD weight, kg 70.3 ± 15.7 69.3 ± 15.9
Mean ± SD creatinine clearance, mL/min 94.2 ± 35.3 94.3 ± 34.1
Mean ± SD therapy duration, days 8.1 ± 2.8 7.9 ± 2.7
Mean APACHE II score 5.6 5.5
Primary intra-abdominal diagnosis, n (%)
Complicated appendicitis 152 (61.5) 145 (56.9)
Perforation of intestine 21 (8.5) 23 (9.0)
Complicated diverticulitis 17 (6.9) 25 (9.8)
Intra-abdominal abscess 17 (6.9) 17 (6.7)
Peritonitis 14 (5.7) 16 (6.3)
Gastric/duodenal perforation 13 (5.3) 10 (3.9)
Complicated cholecystitis 12 (4.9) 16 (6.3)
Other* 1 (0.4) 3 (1.2)
*Other diagnoses included infected hematoma, pelvic inflammatory disease, acute abdomen subocclusion, acute inflammatory abdomen, disease pelvic infectious, tubo-ovarian abscess, right tubal abscess, infected left subphrenic hematoma.
Clinical efficacy
For the ME population, clinical cure rates were 80.6% for tigecycline and 82.4% for imipenem/cilastatin (95% CI -9.0, 5.4; Table 2). Corresponding clinical cure rates for the m-mITT population were 73.5% and 78.2% (95% CI -11.8, 2.3), respectively. For both the ME and m-mITT populations, tigecycline was efficacious and statistically noninferior to imipenem/cilastatin. Multiple subgroup analyses of clinical responses (eg, age, sex, race, geographic location) found consistently efficacious clinical responses between the treatment groups. No significant treatment differences in clinical response were observed between the two treatment groups when patients were stratified by the number of isolated baseline isolates (Table 2). For the ME population, tigecycline had a 89.8% clinical cure rate at the test-of-cure visit for monomicrobial infections and a 75.3% clinical cure rate for polymicrobial infections. Similar rates were observed for recipients of imipenem/cilastatin (88.5% and 78.1%, respectively).
Table 2 Clinical cure rates at test-of-cure visit
Tigecycline Imipenem/cilastatin Difference Tigecycline-Imipenem/cilastatin Test for Noninferiority Test for Differences
Population N % (95% CI) N % (95% CI) % (95% CI) P value
CE 282/341 82.7 (78.3, 86.6) 295/351 84.0 (79.8, 87.7) -1 (-7.2, 4.5) <0.0001 0.70
Overall -1 (-6.9, 4.2)*
c-mITT 303/408 74.3 (69.7, 78.4) 317/399 79.4 (75.1, 83.3) -5 (-11.2, 0.0) <0.0001 0.00
Overall -5 (-11.0, 0.0)
ME 199/247 80.6 (75.1, 85.3) 210/255 82.4 (77.1, 86.8) -1.8 (-9.0, 5.4) 0.0001 0.6892
Monomicrobial 80/89 89.9 (81.7, 95.3) 92/104 88.5 (80.7, 93.9) 1.4 (-8.7, 11.0)
Polymicrobial 119/158 75.3 (67.8, 81.8) 118/151 78.1 (70.7, 84.5) -2.8 (-12.6, 7.1)
Overall -1.7 (-8.4, 5.1)*
m-mITT 227/309 73.5 (68.2, 78.3) 244/312 78.2 (73.2, 82.7) -4.7 (-11.8, 2.3) 0.0019 0.1976
Monomicrobial 96/121 79.3 (71.0, 86.2) 109/128 85.2 (77.8, 90.8) -5.8 (-15.9, 4.3)
Polymicrobial 131/188 69.7 (62.6, 76.2) 135/184 73.4 (66.4, 79.6) -3.7 (-13.1, 5.9)
Overall -4.3 (-11.0, 2.5)*
*Adjusted difference and its 95%CI are calculated from a generalized linear model with a binomial probability function and an identity link.
For complicated appendicitis, the most frequent diagnosis, clinical cure rates at the test-of-cure visit was 84.2% for tigecycline and 86.2% for imipenem/cilastatin (Table 3). In both treatment groups, lower clinical cure rates (≤72%) were observed in patients who had intra-abdominal abscess, complicated diverticulitis, or intestinal perforation (Table 3). Overall, there were no significant differences in clinical cure rates between tigecycline and imipenem/cilastatin based on primary intra-abdominal diagnosis. A total of 14 tigecycline- and 27 imipenem/cilastatin-treated patients in the ME population had a positive pretherapy blood culture. Clinical cure in patients with bacteremia was reported for 71.4% of tigecycline and 74.1% of imipenen/cilastatin recipients.
Table 3 Clinical cure rate by baseline diagnosis (ME population) at test-of-cure visit
Tigecycline Imipenem/cilastatin Difference Tigecycline-Imipenem/cilastatin
Clinical Diagnosis N % (95% CI) N % (95% CI) % (95% CI)
Complicated appendicitis 128/152 84.2 (77.4, 89.6) 125/145 86.2 (79.5, 91.4) -2.0 (-10.6, 6.7)
Perforation of the intestines 13/21 61.9 (38.4, 81.9) 15/23 65.2 (42.7, 83.6) -3.3 (-32.4, 26.2)
Complicated diverticulitis 12/17 70.6 (44.0, 89.7) 18/25 72.0 (50.6, 87.9) -1.4 (-32.0, 26.7)
Intra-abdominal abscess 11/17 64.7 (38.3, 85.8) 12/17 70.6 (44.0, 89.7) -5.9 (-37.6, 27.4)
Peritonitis 12/14 85.7 (57.2, 98.2) 15/16 93.8 (69.8, 99.8) -8.0 (-38.2, 20.5)
Complicated cholecystitis 11/12 91.7 (61.5, 99.8) 14/16 87.5 (61.7, 98.4) 4.2 (-29.4, 32.4)
Gastric and abdominal perforations 11/13 84.6 (54.6, 98.1) 10/10 100.0 (69.2, 100.0) -15.4 (-46.3, 21.3)
Other 1/1 100.0 (2.5, 100.0) 1/3 33.3 (0.8, 90.6) 66.7 (-42.3, 98.2)
Concomitant bacteremia 10/14 71.4 (41.9, 91.6) 20/27 74.1 (53.7, 88.9) -2.6 (-35.3, 25.4)
Microbiologic efficacy
For the ME population, eradication of intra-abdominal isolates at the patient level was reported for 80.6% of tigecycline- and 82.4% of imipenem/cilastatin-treated patients (95% CI -9.0, 5.4), indicating that tigecycline was efficacious and statistically noninferior to imipenem/cilastatin (Table 4). No significant differences between the treatment groups were found when eradication rates were stratified by monomicrobial versus polymicrobial infection (Table 4).
Table 4 Microbiologic response at the patient level (ME Population) at test-of-cure visit
Tigecycline Imipenem/cilastatin Difference Tigecycline-Imipenem/cilastatin Test for Noninferiority Test for Differences
Response N % (95% CI) N % (95% CI) % (95% CI) P value
Eradication 199/247 80.6 (75.1, 85.3) 210/255 82.4 (77.1, 86.8) -1.8 (-9.0, 5.4) 0.0001 0.6892
Persistence 39/247 15.8 (11.5, 20.9) 42/255 16.5 (12.1, 21.6)
Documented 4/39 10.3 (2.9, 24.2) 1/42 2.4 (0.1, 12.6)
Presumed 35/39 89.7 (75.8, 97.1) 41/42 97.6 (87.4, 99.9)
Superinfection 9/247 3.6 (1.7, 6.8) 3/255 1.2 (0.2, 3.4)
Overall -1.7 (-8.4, 5.1)*
*Adjusted difference and its 95%CI are calculated from a generalized linear model with a binomial probability function and an identity link.
Generally, eradication rates at the test-of-cure visit for the most commonly isolated intra-abdominal isolates were similar between the two treatment groups (Table 5). For E. coli, the most commonly isolated aerobe, eradication rates were 80.4% for tigecycline versus 83.5% for imipenem/cilastatin. Corresponding eradication rates for Klebsiella spp, the second most frequently isolated gram-negative aerobe, were 87.1% and 85.7%, respectively. A total of 6 ESBL-producing E. coli and 7 ESBL-producing K. pneumoniae isolates were identified pretherapy. The majority of these isolates were eradicated by tigecycline: 83% (5/6) and 71% (5/7), respectively. Eradication rates for Bacteroides fragilis were 69.8% for tigecycline and 72.5% for imipenem/cilastatin.
Table 5 Microbiologic response at the isolate level: selected baseline isolates at test-of-cure visit (ME population)
Tigecycline Imipenem/cilastatin
Isolate N MIC90 % (95% CI) N MIC90 % (95% CI)
Bacteroides fragilis 30/43 2.0 69.8 (53.9, 82.8) 29/40 0.5 72.5 (56.1, 85.4)
Citrobacter spp. 13/15 1.0 86.7 (59.5, 98.3) 5/7 0.5 71.4 (29.0, 96.3)
Clostridium spp. 16/19 1.0 84.2 (60.4, 96.6) 14/18 2.0 77.8 (52.4, 93.6)
Enterobacter spp. 6/8 1.0 75.0 (34.9, 96.8) 5/10 1.0 50.0 (18.7, 81.3)
Enterococcus faecalis (non-VRE) 10/16 0.25 62.5 (35.4, 84.8) 9/18 4.0 50.0 (26.0, 74.0)
Escherichia coli 135/168 0.5 80.4 (73.5, 86.1) 152/182 0.25 83.5 (77.3, 88.6)
Fusobacterium spp. 3/5 0.25 60.0 (14.7, 94.7) 6/7 0.25 85.7 (42.1, 99.6)
Klebsiella spp. 27/31 1.0 87.1 (70.2, 96.4) 36/42 0.25 85.7 (71.5, 94.6)
Peptostreptococcus spp. 6/10 0.12 60.0 (26.2, 87.8) 5/8 0.25 62.5 (24.5, 91.5)
Proteus spp. 5/10 4.0 50.0 (18.7, 81.3) 3/3 4.0 100.0 (29.2, 100.0)
Pseudomonas aeruginosa 13/18 32.0 72.2 (46.5, 90.3) 19/21 2.0 90.5 (69.6, 98.8)
Staphylococcus aureus (MRSA) 1/2 NA 50.0 (1.3, 98.7) 0/1 NA 0.0 (0.0, 97.5)
S. aureus (non-MRSA) 7/8 0.25 87.5 (47.3, 99.7) 3/4 0.12 75.0 (19.4, 99.4)
Streptococcus spp. 63/81 0.12 77.8 (67.2, 86.3) 46/67 0.12 68.7 (56.2, 79.4)
MRSA = methicillin-resistant Staphylococcus aureus; VRE = vancomycin-resistant enterococci.
NA = MIC90 values are not valid if the number of isolates is less than 10.
Pretherapy in vitro activity against baseline isolates for tigecycline and imipenem/cilastatin are shown in Table 6. The mean MIC90 for tigecycline against the most commonly isolated aerobes and anaerobes was ≤2.0 mg/L. No pretherapy isolates displayed resistance to tigecycline based on the provisional breakpoints used. Bacterial susceptibilities to tigecycline appeared to be consistent with clinical responses.
Table 6 MIC range, and MIC50 and MIC90 values of selected primary baseline isolates (ME population)
Tigecycline Imipenem/Cilastatin
Isolate n MIC range MIC50 MIC90 MIC range MIC50 MIC90
Bacteroides fragilis 83 0.06–16.0 1.0 2.0 0.12–4.0 0.25 0.5
Clostridium perfringens 12 0.06–2.0 1.0 2.0 0.12–0.25 0.12 0.25
Enterococcus faecalis (non-VRE) 32 0.06–0.25 0.12 0.25 1.0–4.0 1.0 4.0
Escherichia coli 350 0.06–1.0 0.25 0.50 0.12–1.0 0.12 0.25
Klebsiella pneumoniae 58 0.25–2.0 0.50 1.00 0.12–0.50 0.25 0.25
Pseudomonas aeruginosa 39 8.0–32.0 16.0 32.0 0.25–4.0 1.0 2.0
Staphylococcus aureus (MRSA) 3 0.12–0.25 NA NA 0.12–32.0 NA NA
S. aureus (non-MRSA) 12 0.12–0.50 0.25 0.25 0.12–0.12 0.12 0.12
NA = MIC50 and MIC90 values are not valid if the number of isolates is less than 10.
MRSA = methicillin-resistant Staphylococcus aureus.
VRE = vancomycin-resistant enterococci.
Safety and tolerability
Data from all patients in the mITT population (n = 825) were analyzed for safety. The m-ITT population received a median of 7 days of tigecycline or imipenem/cilastatin treatment. Regardless of study drug causality, the frequency and distribution of treatment-emergent adverse events occurring in at least 3% of patients in either treatment group were similar to those observed in the imipenem/cilastatin treatment group. The majority of these adverse events were related to study medication (56%) and were mild to moderate in intensity (94%). Digestive system (56.9% vs 49.8%, P = 0.043), nausea (31.0% tigecycline, 24.8% imipenem/cilastatin; P = 0.052), vomiting (25.7% tigecycline, 19.4% imipenem/cilastatin; P = 0.037), and diarrhea (21.3% tigecycline, 18.9% imipenem/cilastatin; P = 0.435) were the most frequently reported adverse events in both treatment groups. The majority of patients in both treatment groups experienced mild to moderate nausea and/or vomiting (94%). There was no significant difference between the treatment groups in the number of patients who required antiemetic therapy for nausea and/or vomiting. No tigecycline-treated patient has a positive Clostridium difficile toxin assay, nor developed C. difficile associated diarrhea.
In the tigecycline group, infections (13.6% vs 7.5%, P = 0.006), hypoproteinemia (8.0% vs 4.1%, P = 0.028), and dyspnea (6.8% vs 2.9%, P = 0.014) were statistically higher than in the imipenem/cilastatin treatment group. The difference in infection rates between the treatment groups was primarily due to the development of secondary wound infections. No apparent trends or risk factors were identified in the development of secondary wound infections in either treatment group.
One hundred forty-five (145) patients had one or more serious adverse events during the study (81 [19.6%] tigecycline, 64 [15.5%] imipenem/cilastatin) (P = 0.143). The most frequently reported serious adverse events were abnormal healing (14 tigecycline, 6 imipenem/cilastatin), abscess (10 tigecycline, 8 imipenem/cilastatin), and infection (10 tigecycline, 9 imipenem/cilastatin). Significantly more patients treated with tigecycline (6 [1.5%]) versus none treated with imipenem/cilastatin reported pneumonia as a serious adverse event (P = 0.031).
Adverse events were the primary reason for early withdrawal of study drug. A total of 27 (6.5%) tigecycline- and 15 (3.6%) imipenem/cilastatin-treated patients discontinued treatment prematurely because of an adverse event (P = 0.080). A total of 10 (2.5%) tigecycline- and 4 (1.0%) imipenem/cilastatin-treated patients stopped therapy prematurely secondary to either nausea (6 tigecycline, 2 imipenem/cilastatin) and/or vomiting (4 tigecycline, 2 imipenem/cilastatin). There were no significant differences between treatment groups in any single adverse event leading to the discontinuation of study drug.
Twenty-nine (29) patients died during the study: 17 patients in the tigecycline group and 12 patients in the imipenem/cilastatin treatment group. Only two of the deaths, both in the tigecycline group, were considered by the investigators to be possibly related to study drug secondary to treatment failure. The first patient was a 78 year old female who received tigecycline for one week. Two days following discontinuation of therapy the patient developed septic shock; she died one day later. The second patient, a 23 year old female, presented with sepsis and received 3 days of tigecycline therapy. On day 3 she was found to have pneumonia and progressed to multiple organ failure with sepsis and died the same day.
Few clinically important or unexpected changes in any routine hematologic or serum chemistry tests, vital signs, or ECG data were associated with the use of tigecycline or imipenem/cilastatin. However, significantly more patients treated with imipenem/cilastatin (312/410, 76.1%) than those treated with tigecycline (275/408, 67.4%) had 1 or more laboratory findings of potential clinical importance (P = 0.007). Imipenem/cilastatin-treated patients had significantly lower serum potassium (≤3 mmol/L; P = 0.004), phosphorus values (≤ 0.8 mmol/L; P < 0.001), and lymphocytes values (≤0.6 cells × 109/L; P < 0.001). Yet, significantly more patients treated with tigecycline than those treated with imipenem/cilastatin had clinically significant hypoproteinemia (≤35 g/L; P = 0.001). No significant changes in QTc interval were observed in either treatment group at any time point.
Discussion
This large trial demonstrated that tigecycline (100 mg initial dose, followed by 50 mg q12 hours) is effective for the treatment of hospitalized adult patients with complicated intra-abdominal infections. For patients with proven bacterial infections, clinical cure rates were 80.6% for tigecycline versus 82.4% for imipenem/cilastatin at the test-of-cure visit, demonstrating that tigecycline met the statistical criteria for noninferiority compared with the carbapenem regimen. We also observed that tigecycline's clinical efficacy was similarly effective in patients who had either monomicrobial versus polymicrobial infection, as well as across the variety of anatomical infections encountered. While many previous studies have reported a higher percentage of polymicrobial infection from intra-abdominal sites, the lower rates seen with tigecycline may be explained by a larger proportion of patients with appendicitis as the source of infection. Overall, the efficacy of tigecycline was consistent among all predefined populations analyzed (m-mITT, c-mITT, CE) and consistent across different species of infecting bacteria.
This large study extends the findings of two other studies that evaluated tigecycline's efficacy in the treatment of complicated intra-abdominal infections. In a small, open-label, phase 2 tigecycline trial of 66 hospitalized patients with primarily perforated appendicitis, cure rates at the test-of-cure visit and end-of-treatment visit were 67% and 76%, respectively [17]. A similarly designed phase 3 trial reported comparable clinical cure rates for the m-mITT cohort of 86.6% (279/322) for tigecycline compared with 84.6% (270/319) for imipenem/cilastatin therapy [18].
The current trial demonstrated that tigecycline was effective at eradicating commonly encountered aerobic and anaerobic intestinal bacteria. Overall eradication rates were nearly identical in the two treatment groups: 80.6% after tigecycline therapy compared with 82.4% in the imipenem/cilastatin group. More than 80% of E. coli and Klebsiella spp. (the two most frequently isolated gram-negative aerobes) were eradicated by tigecycline, followed by 78% of Streptococcus spp, and 70% of B. fragilis. Comparable eradication rates were observed following imipenem/cilastatin therapy, further establishing that tigecycline was at least as effective as the standard carbapenem regimen. These data support in vitro observations that tigecycline has broad-spectrum activity against common isolates found in intra-abdominal infections [8-16]. While the etiologic role of P. aeruginosa remains unclear in patients with community-acquired intra-abdominal infections, tigecycline lacks reliable in vitro activity against this organism [8,9,11,13] despite a 72% eradication rate in this study.
Because few resistant isolates were isolated in the current trial, we could not conclusively establish the in vivo effectiveness of tigecycline against organisms that typically convey resistance (eg, E. faecalis, methicillin-sensitive and -resistant S. aureus, ESBL-producing Enterobacter spp.). However, tigecycline successfully eradicated the majority (77%) of the 13 ESBL-producing E. coli and K. pneumoniae that were recovered from patients with cIAI. These limited data confirm tigecycline's documented in vitro activity against many gram-positive and gram-negative bacterial isolates that typically are resistant [11,16,19].
Both tigecycline and imipenem/cilastatin were well tolerated in the current trial, with a similar frequency and distribution of treatment-emergent adverse events. Nausea, vomiting, and diarrhea were the most frequently reported adverse events in both the tigecycline and imipenem/cilastatin treatment groups. Although the individual adverse events of nausea and vomiting occurred at higher rates after tigecycline compared with imipenem/cilastatin therapy, only the incidence of vomiting was found to be significantly higher in tigecycline recipients. Furthermore, the majority of nausea/vomiting events in both treatment groups were of mild to moderate intensity (94%). Supporting this fact, these gastrointestinal adverse events rarely led to early discontinuation of therapy in either treatment group (<2%) and there was no difference in the number of patients requiring interventional antiemetic therapy between the tigecycline and imipenem/cilastatin groups. It is also noteworthy that tigecycline monotherapy was not associated with the development of C. difficile diarrhea. These findings support previous safety data from phase 2 and 3 studies [20-25].
Conclusion
Tigecycline is an effective and well-tolerated monotherapy option for the treatment of patients with complicated intra-abdominal infections, with comparable efficacy to imipenem/cilastatin. Because of the rising rates of antibiotic-resistant bacteria, both in the community and hospital settings, there remains a need for new antibiotic options. According, tigecycline is a promising new monotherapy when empiric coverage is needed against both gram-positive and non-pseudomonal gram-negative bacteria, including improved in vitro activity against certain resistant isolates.
List of abbreviations
AIDS – acquired immunodeficiency syndrome
ALT – alanine aminotransferase
ANOVA – analysis of variance
APACHE – Acute Physiologic and Chronic Health Evaluation
AST – aspartate aminotransferase
CE – clinically evaluable
cIAI – complicated intra-abdominal infections
CI – confidence interval
c-mITT – clinical modified intent-to-treat
ECG – electrocardiograms
ESBL – extended spectrum beta-lactamases
IDSA – Infectious Diseases Society of America
IMI/CIS – imipenem/cilastatin
ITT – intent-to-treat
IV – intravenous
ME – microbiologically evaluable
MIC – minimum inhibitory concentration
m-mITT – microbiologically modified intent-to-treat
MRSA – methicillin-resistant Staphylococcus aureus
NCCLS – National Committee of Clinical Laboratory Standards
q12h – every 12 hours
SD – standard deviation
TOC – test-of-cure
ULN – upper limit of normal
VRE – Vancomycin-Resistant Enterococci
Competing interests
Financial competing interests: Drs. Oliva, Rekha, Yellin, Paternak, and Campos are investigators for this tigecycline study sponsored by Wyeth. Dr. Oliva is an investigator for a clinical trial sponsored by Roche. Drs. Rose, Babinchak, Ellis-Grosse, and Loh are employees of Wyeth. None of the authors have non-financial competing interests to disclose.
Authors' contributions
The first five authors were investigators in the clinical trial and enrolled the highest number of evaluable patients (MEO, AR, AY, JP, MC). Authors GMR, TB, EE-G, EL made substantial contributions to the conception and design of the study. All authors made substantial contributions acquisition of data and analysis and interpretation of data. Each author (MEO, AR, AY, JP, MC, GMR, TB, EE-G, EL) was involved in critically revising the paper for intellectual content and has given final approval of this version to be published.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by Wyeth Research. We thank the tigecycline 301 study group investigators for their valuable involvement in this study: Fathi Abuzgaya, Louis H. Alarcon, Marc Alpert, Eduardo G. Arathoon, Rebeca Georgina Northland Areyuna, Annadan C. Ashok, Jeffrey A. Bailey, Ian McNicoll Baird, Philip S. Barie, M.Y. Bapaye, Robert W. Beart, Jr, Jean-François Bellemare, Guillermo Alberto Benchetrit, German Berbel, Carlos Enrique Bergallo, Joaquin Bermejo, Thomas B. Berne, Marcela Alicia Vera Blanch, John MA Bohnen, Patricia Brown, Maria Isabel Campos, Iris Lorena Cazali (Leal), Nicolas V. Christou, Daniel Jorge Curcio, Alexey Datsenko, Mario Del Castillo, E. Patchen Dellinger, Sanjay P. Desmukh, Puneet Dhar, Julia Garcia-Diaz, John W. Drover, John M.A. Embil, Zilvinas Endzinas, David Evans, Peter Fomin, Joseph Fraiz, Amalia Rodriquez French, Gary E. Garber, Doria Grimard, Gene Grindlinger, Virsing Punabhai Hathila, Ernesto Julio Jakob, Abel Jasovich, A. Mark Joffe, Ashok Tarachandji Kamble, Ricardo Eiji Kawamoto, Paul Kearney, Min-Ja Kim, Yang Soo Kim, Robert G. Kingman, Stanley R. Klein, Wen-Je Ko, William K.K, Lau, Patrick C. Lee, Dawei Liu, Carlos Lovesio, John Mazuski, Charles Morrow, Chau Nguyen, Maria Eugenia Oliva, Maria Costa Orlando, Guilermo M. Ruiz-Palacios (Santos), Eduardo Parra-Davila, Jacyr Pasternak, Andrejs Pavars, André Poirer, Germain Poirer, Guntars Pupelis, K. Ramachandra Pai, Hariharan Ramesh, M.K. Ramesh, Arturas Razbadauskas, Arcot Rekha, Ronald D. Robertson, Ori D. Rotstein, Rajkumar Janavicularm Sankaran, Ragulagedda Adikesava Sastry, Yan-Shen Shan, Rabih Salloum, Stephen D. Shafran, Jae-Hoon Song, Yaoqing Tang, Osvaldo Teglia, Jüri Teras, Shirin Towfigh, Tiit Vaasna, Walter Vasen, Carlos Rodolfo, Mejia Villatoro, Ramses Wassef, Junmin Wei, John Weigelt, Samuel E. Wilson, Yonghong Xiao, Lunan Yan, Albert E. Yellin, Dah-Shyong Yu, Yingyuan Zhang, and Juan Carlos Zlocowski.
We thank Wyeth Research employee Patricia Bradford for microbiological analysis and Upside Endeavors for professional medical writing services. Wyeth Research, Collegeville, PA, supported and funded this study.
==== Refs
Solomkin JS Mazuski JE Baron EJ Sawyer RG Nathens AB DiPiro JT Buchman T Dellinger EP Jernigan J Gorbach S Chow AW Bartlett J Guidelines for the selection of anti-infective agents for complicated intra-abdominal infections Clin Infect Dis 2003 37 997 1005 14523762 10.1086/378702
Gorbach SL Intraabdominal infections Clin Infect Dis 1993 17 961 5 8110953
Lorber B Swenson RM The bacteriology of intra-abdominal infections Surg Clin North Am 1975 55 1349 54 1198292
Burnett RJ Haverstock DC Dellinger EP Reinhart HH Bohnen JM Rotstein OD Vogel SB Solomkin JS Definition of the role of enterococcus in intraabdominal infection: analysis of a prospective randomized trial Surgery 1995 118 716 21 7570327
de Vera ME Simmons RL Antibiotic-resistant enterococci and the changing face of surgical infections Arch Surg 1996 131 338 42 8611101
Marshall JC Innes M Intensive care unit management of intra-abdominal infection Crit Care Med 2003 31 2228 37 12973184 10.1097/01.CCM.0000087326.59341.51
Krobot K Yin D Zhang Q Sen S Altendorf-Hofmann A Scheele J Sendt W Effect of inappropriate initial empiric antibiotic therapy on outcome of patients with community-acquired intra-abdominal infections requiring surgery Eur J Clin Microbiol Infect Dis 2004 23 682 7 15322931 10.1007/s10096-004-1199-0
Petersen PJ Jacobus NV Weiss WJ Sum PE Testa RT In vitro and in vivo antibacterial activities of a novel glycylcycline, the 9-t-butylglycylamido derivative of minocycline (GAR-936) Antimicrob Agents Chemother 1999 43 738 44 10103174
Gales AC Jones RN Antimicrobial activity and spectrum of the new glycylcycline, GAR-936 tested against 1,203 recent clinical bacterial isolates Diagn Microbiol Infect Dis 2000 36 19 36 10744364 10.1016/S0732-8893(99)00092-9
Petersen PJ Bradford PA Weiss WJ Murphy TM Sum PE Projan SJ In vitro and in vivo activities of tigecycline (GAR-936), daptomycin, and comparative antimicrobial agents against glycopeptide-intermediate Staphylococcus aureus and other resistant gram-positive pathogens Antimicrob Agents Chemother 2002 46 2595 601 12121938 10.1128/AAC.46.8.2595-2601.2002
Milatovic D Schmitz FJ Verhoef J Fluit AC Activities of the glycylcycline tigecycline (GAR-936) against 1,924 recent European clinical bacterial isolates Antimicrob Agents Chemother 2003 47 400 4 12499224 10.1128/AAC.47.1.400-404.2003
Patel R Rouse MS Piper KE Steckelberg JM In vitro activity of GAR-936 against vancomycin-resistant enterococci, methicillin-resistant Staphylococcus aureus and penicillin-resistant Streptococcus pneumoniae Diagn Microbiol Infect Dis 2000 38 177 9 11109018 10.1016/S0732-8893(00)00193-0
Boucher HW Wennersten CB Eliopoulos GM In vitro activities of the glycylcycline GAR-936 against gram-positive bacteria Antimicrob Agents Chemother 2000 44 2225 9 10898710 10.1128/AAC.44.8.2225-2229.2000
Cercenado E Cercenado S Gomez JA Bouza E In vitro activity of tigecycline (GAR-936), a novel glycylcycline, against vancomycin-resistant enterococci and staphylococci with diminished susceptibility to glycopeptides J Antimicrob Chemother 2003 52 138 9 12805264 10.1093/jac/dkg289
Low DE Kreiswirth BN Weiss K Willey BM Activity of GAR-936 and other antimicrobial agents against North American isolates of Staphylococcus aureus Int J Antimicrob Agents 2002 20 220 2 12385702 10.1016/S0924-8579(02)00132-2
Biedenbach DJ Beach ML Jones RN In vitro antimicrobial activity of GAR-936 tested against antibiotic-resistant gram-positive blood stream infection isolates and strains producing extended-spectrum beta-lactamases Diagn Microbiol Infect Dis 2001 40 173 7 11576790 10.1016/S0732-8893(01)00269-3
Murray JA Wilson S Klein S Yellin A Loh E The clinical response to tigecycline in the treatment of complicated intra-abdominal infections in hospitalized patients. A phase 2, clinical trial [abstract] Presented at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy 2003 43 416
Dartois N Gioud-Paquet M Ellis-Grosse EJ Loh E Tigecycline vs imipenem/cilastatin for treatment of complicated intra-abdominal infections [abstract] Presented at the 44th Annual Meeting of the Interscience Conference on Antimicrobial Agents and Chemotherapy 2004 Session #97, Control #3841, Presentation #L-992c
Henwood CJ Gatward T Warner M Antibiotic resistance among clinical isolates of Acinetobacter in the UK, and in vitro evaluation of tigecycline (GAR-936) J Antimicrob Chemother 2002 49 479 87 11864948 10.1093/jac/49.3.479
Muralidharan G Getsy J Mayer P Paty I Micalizzi M Speth J Wester B Mojaverian P Pharmacokinetics (PK), safety and tolerability of GAR-936, a novel glycylcycline antibiotic, in healthy subjects [abstract] Presented at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy 1999 39 303
Muralidharan G Mojaverian P Micalizzi M Raible DG Troy SM Effects of age and sex on single-dose pharmacokinetics of tigecycline in healthy subjects Antimicrobial Agents and Chemotherapy 2005 49 1656 15793165 10.1128/AAC.49.4.1656-1659.2005
Sesoko S Umemura K Nakashima M Pharmacokinetics (PK), safety and tolerability of tigcycline (GAR-936) in healthy Japanese males [abstract] Presented at the 42nd Interscience Conference on Antimicrobial Agents and Chemotherapy 2002 42 22
Postier RG Green SL Klein SR Ellis-Grosse EJ Loh E Tigecycline 200 Study Group Results of a multicenter, randomized, open-label efficacy and safety study of two doses of tigecycline for complicated skin and skin-structure infections in hospitalized patients Clin Ther 2004 26 704 14 15220014 10.1016/S0149-2918(04)90070-7
Troy SM Mauralidharan G Micalizzi M Mojavarian P Salacinski L Raible D The effects of renal disease on the pharmacokinetics of tigecycline (GAR-936) [abstract] Presented at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy 2003 Poster A-22
Zhanel GG Homenuik K Nichol K Noreddin A Vercaigne L Embil J Gin A Karlowsky JA Hoban DJ The glycylcyclines: a comparative review with the tetracyclines Drugs 2004 64 63 88 14723559
|
16236177
|
PMC1277826
|
CC BY
|
2021-01-04 16:28:16
|
no
|
BMC Infect Dis. 2005 Oct 19; 5:88
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-88
|
oa_comm
|
==== Front
BMC Palliat CareBMC Palliative Care1472-684XBioMed Central London 1472-684X-4-61621267310.1186/1472-684X-4-6Research ArticlePatterns and predictors of place of cancer death for the oldest old Lock Anna [email protected] Irene [email protected] Coventry Community Palliative Care Team, 25 Warwick Road, C/O Christchurch House, Grey Friars Lane, Coventry, CV1 2GQ, UK2 Department of Palliative Care and Policy, King's College London, Weston Education Centre, Cutcombe Road, Denmark Hill, London, SE5 9RJ, UK2005 8 10 2005 4 6 6 15 3 2005 8 10 2005 Copyright © 2005 Lock and Higginson; licensee BioMed Central Ltd.2005Lock and Higginson; 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
Cancer patients increasingly are among older age groups, but to date little work has examined the trends in cancer among older people, particularly in relation to end of life care and death. This study describes the older population who die of cancer and the factors which may affect their place of death.
Methods
A Cross-sectional analysis of national data was performed. The study included all people aged 75 and over dying of cancer in England and Wales between 1995 and 1999. The population was divided into exclusive 5 year age cohorts, up to 100 years and over. Descriptive analysis explored demographic characteristics, cancer type and place of death.
Results
Between 1995 and 1999, 315,462 people aged 75 and over were registered as dying from cancer. The number who died increased each year slightly over the 5 year period (1.2%). In the 75–79 age group, 55 % were men, in those aged 100 and over this fell to 16%. On reaching their hundreds, the most common cause of death for men was malignancies of the genital organs; and for women it was breast cancer.
The most frequent place of death for women in their hundreds was the care home; for men it was hospitals. Those dying from lymphatic and haematopoietic malignancies were most likely to die in hospitals, those with head and neck malignancies in hospices and breast cancer patients in a care home.
Conclusion
The finding of rising proportions of cancer deaths in institutions with increasing age suggests a need to ensure that appropriate high quality care is available to this growing section of the population.
==== Body
Background
Between 1991 and 2001 [1], the UK the population aged over 85 rose by 29.6%. Further since the 1950's the number of people aged over 100 has doubled every decade [2]. A shrinking wage earning group and the growing number of dependent older people is creating a society in which taxable income will fall, whilst demands on health and social services are already rising. This growing section of older people are often socially excluded and at risk of experiencing a combination of linked problems such as unemployment, poor skills, low incomes, poor housing, high crime environments, bad health and family breakdown [3]. Age-based discrimination can be experienced with poorer access to and availability of health care services [4,5].
Cancer is more common with increasing age [1] and the experience of having cancer in old age has been shown to be different to that of younger people [4]. Older people with cancer have multiple co-existent pathologies such as cardiac failure and chronic respiratory disease, prolonged recovery phase and rapid deterioration if the illness is left untreated [6]. Higher levels of functional dependence or co-morbidities[7] have been found in all ages to be associated with increased institutional death.
Meeting a person's choice regarding place of care and death at the end of life is seen as a useful outcome marker of the quality of palliative care. In younger age groups 50% to 80% of people wish to die at home. Increasing age in the general population [8] has been found to be related to preference for not dying at home, where as in Israel when home care patients were asked older people (50–59 yrs) were more likely to prefer death at home (95.2%) than the youngest (21–29: 50%)[9]. Little data is available on the preferences of the oldest old cancer patients or those who lack informal care givers [10]. Despite the critical position of choice of place of death in an older person's expression of their autonomy [11] older people have been found to recognise both practical and moral difficulties in achieving home death [12]. Expressing a preference for home death has been identified in the general population as a factor which may increase home deaths [13].
Previous data gathered on preference for place of death has often focused on home or 'elsewhere', very rarely has care home been reported as an option. An Australian interview survey of general population members found that 2.5% would prefer to die in a nursing home if terminally ill [8]. When patients referred to a palliative care hospital support team were asked, 3% stated that nursing home was their preferred place of terminal care [14].
Increasing age has been found to be associated with reducing the probability of dying at home or hospice and increasing death in care homes [7,9,13,15-19]. None of the studies specifically described the changes in place of death with increasing age beyond 84+.
The association of gender with place of death has been found in a number of studies. In general women are more likely to die in care homes[19], whilst men are generally more likely to die at home[15] or in hospital [18,20]. Italian data conflicts with this, in a region with no care homes, female home care patients were more likely to die at home than men[21]
The effect of cancer group on place of death has been found to be most pronounced for haematological malignancies. These have been repeatedly shown to be associated with death in hospitals [19,20,22]. This is unlike other tumour groups which have shown little consistency between studies.
The availability of caregivers has been highlighted as a factor which may influence place of death at any age of cancer patients. Having a spouse has been found to be associated with a reduced probability of death in a care home [7] and an increased number of care givers was also shown to increase the probability of home death [23-25].
However, no studies have analysed place of death in the oldest old. This study sought to examine the epidemiology of the oldest old dying with cancer and factors which may influence their place of death.
Methods
Cross-sectional analysis was performed on routinely collected death registration data for cancer deaths (ICD 9 codes 140–239) in England and Wales, for the years 1995–1999, of all people aged 75 and upward under licence from the Office of National Statistics. Data included: age, gender, social class, country of birth, cause (ICD-9) and place of death.
Using SPSS for Windows v11.5, age was examined in 5 year exclusive cohorts ranging from those aged 75–79 years (the 'younger old') to those over 100 years and over (the 'oldest old'). Because of small numbers in the 105–110 cohort (n = 16), all those aged 100 years and over were analysed together. Trends in cause of death and country of birth were described for the whole population and separately by gender with increasing 5 year age cohorts.
Results
A total of 315,462 of deaths from cancer of people aged 75 and over were registered in England and Wales. This accounted for 46% of all cancer deaths. There was an increase of 1.2% (783) between 1995 (62,266) and 1999 (63,049). In contrast the total number of cancer deaths fell by 5% between 1995 (140,791) and 1999 (133,749). Complete data was available regarding age, gender, cause of death and place of death. 99. 7% of social class information was missing.
Of cancer deaths, 119,852 were aged from 75 to 79, and 454 were aged 100 and over (Figure 1). The oldest person registered as dying in this period was 108, 49% were men. The proportion of women rose from 45% of the 75–79 cohort to 84% of the 100 and over cohort (Figure 2).
Figure 1 Pie chart of number of deaths by age cohort (N = 315,462).
Figure 2 Graph of percentage deaths for men and women by age cohort (N = 315,462).
Most people (92.8%, n = 292,795) had been born in England and Wales, the remainder comprised people from 148 other countries. Of those born outside of England and Wales, 21% came from the Irish Republic and 21% from Continental Europe. The proportion of those born outside of England and Wales fell towards the extreme of old age when compared to the younger old.
For the population as a whole, the most common cause of death was lung and other intra-thoracic malignancies, which was recorded as the cause of death in 19.9% (n = 62,604) of deaths. In the 75–79 cohort, lung and other intra-thoracic malignancies comprised 25.1% of deaths – this fell to 3.5% in those over 100. In this oldest group the most common cause of death was female breast cancer, which contributed 26.4% of deaths, contrasting with 6.3% in the 75–79 cohort. The proportion dying of cancer described as 'other' rose from 14.5% in the 75–79 cohort to 18.1% in the over 100's (Figure 3).
Figure 3 Percentage cause of cancer death by age cohort (N = 315,462).
When the population was described separately for men and women, the most common cause of death for men was lung cancer contributing 25.2% of all deaths and 29. 4% of those aged 75–79. On reaching their hundreds the most common cause of death for men were malignancies of the male genital organs (24.7%) with lung and other intra-thoracic malignancies contributing 11% of deaths.
For women, upper gastrointestinal malignancies were the most common cause of death overall, with 18.2% of total deaths. In the youngest cohort, 19.9% of deaths were due to lung and intra-thoracic malignancies. By the 80–84 cohort, gastrointestinal malignancies were the most common cause of death and remained so until the 94–99 cohort when breast cancer was the most common cause of death (22.2%), and reaching 31.5% in women over 100.
Most people died in hospitals (n = 156,334/50%), and fewest in hospices (n = 39,576/13%). Home deaths accounted for 19% of deaths, and care homes for the remaining 16% of deaths. There was a small increase in the percentage of deaths in hospices over time, from 11.5% in 1995, to 13.7% in 1999.
There were marked differences between the 'younger old' (aged 75–79 years) and the 'oldest' old (the over 100's) in place of death. Those dying in hospital reduced with increasing age, from 51.8% to 27.5%. Hospice deaths also fell, from 15.7% to 2%, as did home deaths, from 23.8% to 14.2%. Care home deaths increased from 8.7% in the 'younger old' to 56.3% of the 'oldest old' (Figure 4).
Figure 4 Percentage deaths in each place of death by age cohort (N = 307,613).
In the 'younger old', hospital was the most common place of death for both men and women. However, among the 'oldest old', care homes became the most common place of death (41.1%) for women. Although smaller in number, men in their 100's still died mainly in hospitals (49.4%). Care home deaths were the second most common place of death (29.2%).
The proportion of those dying in different settings varied with cause of death. Of those dying from malignancies of the lymphatic haematopoietic tissues, 66% died in general hospitals or multifunction site, compared to 45% for those with head and neck malignancies. Death in care homes was highest for those with breast cancer (31%), compared with 11% of those with lung and intra-thoracic malignancies. Hospice death was most common among those dying of head and neck tumours (19%), and least common among those with lymphatic and haematopoietic tissue malignancies (8%). Home was the most common place of death amongst those with upper GI malignancies (24%) closely followed by lung and intra-thoracic malignancies (23%).
For men and women within individual cancer types place of death was similar – although clearly breast cancer patients could not be compared. The highest frequency of death in care homes for men was cancer of the genital organs (19%) and, for women, was those dying of breast cancer (31%).
Discussion
This study highlights the rich potential of an easily available, economic data source that has data for a whole population. This enabled a picture to be drawn of an often neglected population without encumbering individuals with interviews. As with all routine data analysis, the study was limited to variables which were previously collected and their completeness. Due to the cross-sectional study design, we were unable to control for confounding factors which may have influenced the results.
Nevertheless, our study demonstrates the increasing number of cancer deaths per year in the oldest old, contrasted with the overall fall of cancer deaths. The greater number of older women dying with cancer, is a reflection of the general population, which has disproportionately more women than men in extreme old age [2]. The low proportion of deaths of people born outside of England and Wales was unsurprising as these groups in the UK have a younger age profile than that of the white population[26], although this situation will change [27].
Our analysis of cause of death is likely to be limited by the accuracy of diagnosis and physician certification [28]. The older population are likely to have multiple pathologies and as a consequence it may be unclear as to the actual cause of death, demonstrated by the rising proportion of cause of cancer deaths described as 'other'. The lower levels of investigation in older people may mean in fact that their cancer is never diagnosed, leading to cases being missed from this study.
The finding that the most common cancer groups in the very old population were of breast and male genital organ origin is significant for service planners. Both diseases can have a protracted course, local and distant spread with associated complications, and multiple treatment options. It is likely that these tumour groups will become an increasing part of the cancer work load as our population ages. Therefore plans for longer term supportive and palliative care are needed[29].
Our finding of variations in place of death by diagnosis supports clinical experiences. Tumours which can lead to complex symptoms such as those of head and neck cancers can be especially difficult to palliate. These patients may be therefore more likely to be perceived to require specialist palliative care and be referred to hospices. The high proportion of deaths in hospitals for those dying of haematological malignancies has been previously documented in the general population[20] and may be linked to their health team having fewer links with specialist palliative care services and therefore reduced access to hospice beds and supportive services at home. It may also be related to the clinical course, which can be rapid with a very short period from diagnosis to death. Both of these factors can reduce the likelihood of home death[16].
Our data on place of death is limited by the data collection methods, in particular completion of the death certificate and coding variations. It only gives a snap shot of the dying pathway, and importantly does not tell us how long a person has been at that place before they died. Our findings support and build on earlier studies, showing a trend away from home death and towards care homes, with increasing age[7,9,18,30]. We found that this trend continues even to the oldest old. On a global scale there is wide range in the reported difference in place of death for the older population. In Italy 33% of those aged 75–84 died at home rising to 42.4% of those aged 85 and above[31]. In USA[20] and Australia[19] the picture seems to be similar to the UK, although this needs more investigation with direct comparative analysis.
The reasons for the variation in place of death as the population reaches the extremes of old age are complex. For many older people life in a care home is their reality. As by aged 65 and above 4% of people live in care homes, rising to 20% of people over 85 years old. Of all residents 75% are women [32]. So for a large number of older people who die in a care home, this is their de facto place of residence.
It is not possible using the data available to ascertain how long the older people recorded as dying in care homes had been resident there, in some cases care home deaths may be true home deaths. With hospices and hospitals increasingly discharging people for ongoing care to care homes the duration of admission may be as short as days, this will have implication for service planning achieved. However, the type of care needed will vary very much through a course of a palliative illness with the intensity of input required by carers increasing in the terminal phase.
Increasing in the UK, care home death may be in part due to the influence of General Practitioners who act as gate-keepers to services [33] and may perceive older people to have needs related to physical and social difficulties and be amenable to care homes [34] rather than having complex psychological and symptom control needs which may suggest referral to a specialist palliative care unit. It may be that the course of illness is less predictable, and not suitable for the relatively short length of hospice admission. The high proportion of care home and hospital deaths has implications for these organisations whose staff may not be trained to care for dying cancer patients. In many cases staff have a variable understanding of the concepts of safe amount of analgesia and definitions of euthanasia [34].
The previously noted reduction in proportion of patients preferring home as a place of death as death approaches[35] may be especially pertinent in the oldest dying. Increasing disability towards their life may be exacerbated by co-morbidities and lack of carers may make death at home less desirable. For stretched community health and social care services, which are often augmented by informal caregivers, it may not be possible to input enough care for older people who are more likely to be living alone. Previous data suggests that Hospice team involvement [16,21] and having special equipment [13] may increase the probability of home death. So for older people who are known to experience age-based discrimination with access to and availability of services [36] this may be another factor which reduces their probability of home death.
This highlights the issue of whether there can ever be a true choice. That realism and a desire to spare others the burden of care often clouds the issue of choice of place of care and death for a section of society that are already disadvantaged.
In the UK, there is a general trend towards increasing provision of care at home for people who require palliative care. This have been encouraged by studies that have shown that in general, people prefer care at home [10]. However, home death does not occur for the majority of older people, and so it is essential that appropriate care services are available in all settings, home, hospital and care home.
The death of a large proportion of the population in care homes and acute hospitals emphasises the need for all health care workers caring for older people in these settings to be adequately trained in generic palliative care in addition to training in complex needs assessment. Models for collaboration between primary and secondary care practitioners with palliative care specialists should be developed and implemented.
Conclusion
46% of all cancer deaths occur in the 75 years and over population. The causes of cancer deaths changes with progression into extreme old age. Lung cancer becomes less common and breast and prostate cancer becomes the most frequent. Care homes and hospitals dominate the place of death in the oldest old, suggesting a need to target training and resources in these organizations as the general population ages.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AL designed the study and performed the data analysis and interpretation as part of the MSc in Palliative Care at King's College London. IH acquired the data, conceived of the study idea, participated in its design, gave editorial advice and was the MSc supervisor. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Office of National Statistics mortality data 2003
Thatcher R The Demography of Centenarians in England and Wales Office of National Statistics Population Trends 1999
Social Exclusion Unit 2003
Fentiman IS How should we approach comprehensive cancer care for elderly people? Progress in Palliative care 1997 1 14 16
Lawton The quality of life among elderly care receivers Journal of Applied Gerentology 1995 14 150 171
Corner J Some reflections on frailty in elderly patients with cancer European Journal of Cancer Care 1993 2 5 9
Jordhoy MS Saltvedt I Fayers P Loge JH Ahlner-Elmqvist M Kaasa S Which cancer patients die in nursing homes? Quality of life, medical and sociodemographic characteristics Palliative Medicine 2003 17 433 444 12882262 10.1191/0269216303pm774oa
Ashby M Wakefield M Attitudes to some aspects of death and dying, living wills and substituted health care decision-making in South Australia: public opinion survey for a parliamentary select committee Palliative Medicine 1993 7 273 282 8261193
Gilbar O Steiner M When death comes: where should patients die? Hospice Journal 1996 11 31 48 8920310
Higginson I Sen-Gupta G Place of care in advanced cancer: A qualitative systematic literature review of patient preferences Journal of Palliative Medicine 2000 3 287 300 15859670 10.1089/jpm.2000.3.287
Debate of the age health and care study group the future of health and care of older people: the best is yet to come Age Concern 1999
Merryn Gott Jane Seymour Gary Bellamy David Clark Sam Ahmedzai Older people's views about home as a place of care at the end of life Palliative Medicine 2004 18 460 467 15332424 10.1191/0269216304pm889oa
Karlsen S Addington-Hall J How do cancer patients who die at home differ from those who die elsewhere? Palliative Medicine 1998 12 279 286 9743826 10.1191/026921698673427657
Dunlop R Preferred versus actual place of death: A hospital palliative care support team study Palliative Medicine 1989 3 197 201
Higginson IJ Jarmen B Astin P Dolan S Do social factors affect where patients die: an analysis of 10 years of cancer deaths in England Journal of Public Health Medicine 1999 21 22 28 10321855 10.1093/pubmed/21.1.22
Moinpour CM Polissar L Factors affecting place of death of hospice and non-hospice cancer patients American Journal of Public Health 1989 79 1549 1551 2817170
Burge F Lawson B Johnston G Trends in the place of death of cancer patients, 1992–1997 Canadian Medical Association Journal 2003 26 637 643
Gatrell A Harman J Francis B Thomas C Morris S Mclllmurray Place of Death: Analysis of cancer deaths in part of North West England Journal of Public Health Medicine 2003 25 53 58 12669919 10.1093/pubmed/fdg011
Hunt R Bonett A Roder D Trends in the terminal care of cancer patients: South Australia 1981–1990 Australian and New Zealand Journal of Medicine 1993 23 245 251 7688953
McCusker J Factors affecting place of death of hospice and non-hospice cancer patients Public Health Reports 1983 98 170 176 6856741
Costantini M Camoirano E Madeddu L Bruzzi P Verganelli E Henriquet F Palliative home care and place of death among cancer patients: a population-based study Palliative Medicine 1993 7 323 331 7505188
Bruera E Sweeney C Russell N Willey J Palmer J Place of Death of Houston Area residents with cancer over a two year period Journal of Pain and Symptom control 2002 26 637 643 10.1016/S0885-3924(03)00204-5
Fukui S Kawagoe H Masako S Noriko N Hiroko N Toshie M Determinants of the place of death among terminally ill cancer patients under home hospice care in Japan Palliative Medicine 2003 17 445 453 12882263 10.1191/0269216303pm782oa
Maida V Factors that promote success in home palliative care: a study of a large suburban palliative care practice Journal of Palliative Care 2002 18 282 286 12611319
Cantwell P Turco S Brenneis C Hanson J Neumann CM Bruera E Predictors of home death in palliative care cancer patients Journal of Palliative Care 2000 16 23 28 10802960
Schuman J The ethnic minority populations of Great Britain - latest estimates 2003
Smaje C Field D Absent Minorities? Ethnicity and the use of palliative care services Death, gender and ethnicity 1997 London: Routledge
Heasman LA Lipworth L Accuracy of certification of causes of death General Register Officer 20, 84–118 London, HMSO Studies on Medical and Population Subjects 1966
National institute for clinical excellence Improving supportive and palliative care for adults with cancer – the manual 1-3-2004.
Malden L Sutherland C Tattersall M Morgan J Forsyth S Levi J Dying of cancer. Factors influencing the place of death of patients Medical Journal of Australia 1984 141 147 150 6749028
De Conno F Caraceni A Goff L Effects of home care on the place of death of advanced cancer patients European Journal of Cancer 1996 32 1142 1147 8758244 10.1016/0959-8049(96)00036-6
Laing W Health Care market review 2001–2002 2002 London, Laing and Buisson
Katz J Sidell M Komaromy Understanding palliative care in residential and nursing homes International Journal Of Palliative Nursing 1999 5 58 64
Gibbs G Nurses in private nursing homes: a study of their knowledge and attitudes to pain management in palliative care Palliative Medicine 1995 9 245 253 7582181
Hinton J Can Home Care maintain an acceptable quality of life for patients with terminal cancer and their relatives? Palliative Medicine 1994 8 189 196
Black D Inequalities in Health Penguin 1992
|
16212673
|
PMC1277827
|
CC BY
|
2021-01-04 16:30:52
|
no
|
BMC Palliat Care. 2005 Oct 8; 4:6
|
utf-8
|
BMC Palliat Care
| 2,005 |
10.1186/1472-684X-4-6
|
oa_comm
|
==== Front
BMC PharmacolBMC Pharmacology1471-2210BioMed Central London 1471-2210-5-151622569510.1186/1471-2210-5-15Research ArticleCurcumin, a diferuloylmethane, attenuates cyclosporine-induced renal dysfunction and oxidative stress in rat kidneys Tirkey Naveen [email protected] Gaganjit [email protected] Garima [email protected] Kanwaljit [email protected] Pharmacology division, University Institute of Pharmaceutical Sciences, Panjab University, Chandigarh-160014, India2005 15 10 2005 5 15 15 19 6 2005 15 10 2005 Copyright © 2005 Tirkey et al; licensee BioMed Central Ltd.2005Tirkey 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 India, Curcumin (CMN) is popularly known as "Haldi", and has been well studied due to its economic importance. Traditional Indian medicine claims the use of its powder against biliary disorders, anorexia, coryza, cough, diabetic wounds, hepatic disorder, rheumatism and sinusitis. This study was designed to examine the possible beneficial effect of CMN in preventing the acute renal failure and related oxidative stress caused by chronic administration of cyclosporine (CsA) in rats. CMN was administered concurrently with CsA (20 mg/kg/day s.c) for 21 days. Oxidative stress in kidney tissue homogenates was estimated using thiobarbituric acid reactive substances (TBARS), reduced glutathione (GSH) content, superoxide dismutase (SOD), and Catalase (CAT). Nitrite levels were estimated in serum and tissue homogenates.
Results
CsA administration for 21 days produced elevated levels of TBARS and marked depletion of renal endogenous antioxidant enzymes and deteriorated the renal function as assessed by increased serum creatinine, Blood Urea Nitrogen (BUN) and decreased creatinine and urea clearance as compared to vehicle treated rats. CMN markedly reduced elevated levels of TBARS, significantly attenuated renal dysfunction increased the levels of antioxidant enzymes in CsA treated rats and normalized the altered renal morphology.
Conclusion
In conclusion our study showed that CMN through its antioxidant activity effectively salvaged CsA nephrotoxicity.
==== Body
Background
Cyclosporine (CsA) (formerly called cyclosporine A), a hydrophobic cyclic undecapeptide produced by the fungus Tolypocladium inflatum, can be considered the prototype of immunosuppressant that has revolutionized the management of allotransplantation. This drug specifically and reversibly inhibits immunocompetent T-helper lymphocytes by suppressing the interleukin-2 driven proliferation of activated T-cells [1]. CsA combines low myelotoxicity with effectiveness in preventing allograft rejection and graft versus host disease as well as in the treatment of various autoimmune and ocular inflammatory diseases [2]. Nephrotoxicity and hypertension are the major adverse effects that often limit CsA treatment following solid organ transplantation and autoimmune diseases [3]. The functional changes caused by CsA are dose dependant and are usually reversible after short-term CsA treatment [4].
Cumulative data suggest a role for reactive oxygen metabolites as one of the postulated mechanisms in the pathogenesis of CsA nephrotoxicity. CsA results in enhanced generation of hydrogen peroxide in cultured hepatocytes [5] and mesangial cells [6,7]. In vitro and in vivo studies indicate that CsA enhances lipid peroxidation, reduces renal microsomal NADPH cytochrome P450, and renal reduced/oxidized glutathione ratio (GSH/GSSG) in kidney cortex as well as renal microsomes and mitochondria [8-11]. Antioxidants such as α-tocopherol, ascorbate, silibinin, lazaroid, propionyl carnitine and superoxide dismutase/catalase, have been shown to ameliorate cyclosporine-induced renal toxicity [5,12].
Current traditional Indian medicine claims the use of Curcuma longa L. (Zingiberaceae) powder against biliary disorders, anorexia, coryza, cough, diabetic wounds, hepatic disorder, rheumatism and sinusitis [13]. Curcumin (CMN) is a major component in curcuma/turmeric, being responsible for its biological actions. More and more studies now show that CMN exhibit anti-inflammatory[14,15], anti-human immunodeficiency virus [16,17], anti-bacterial [18] and nematocidal activities [19]. Various in-vitro and in-vivo studies increasingly establish the antioxidant properties of CMN [20-22]. It is well documented that CMN scavenges superoxide anions [23], peroxynitrite radicals [24,25], and quenches singlet oxygen [26]. CMN has also been shown to inhibit hydrogen-peroxide-induced cell damage [20].
Thus the present study was designed to examine the possible beneficial effect of CMN in preventing the acute renal failure and related oxidative stress caused by chronic administration of CsA in rats.
Results
Effect of CMN on renal function
CsA treatment for 21 days significantly increased the serum creatinine and blood urea nitrogen (BUN) as compared with the control group. Chronic CMN treatment significantly and dose-dependently prevented this rise in BUN and serum creatinine (Table-1). Moreover, the creatinine and urea clearance, which was markedly reduced by CsA-administration, was significantly and dose-dependently improved by CMN treatment (Table-1). However, CMN (15 mg/kg) per se had no effect on serum creatinine, BUN, creatinine and urea clearance.
Table 1 Effect of CMN on cyclosporine-induced nephrotoxicity
Variables Control CsA (20) CMN(15) CsA (20)+ CMN(5) CsA (20)+ CMN(10) CsA (20)+ CMN(15)
Serum creatinine (mg/dl) 0.95 ± 0.01 3.12 ± 0.17a 0.87 ± 0.01b 2.00 ± 0.11a,b 1.5 ± 0.06a,b 1.00 ± 0.01a,b
Creatinine clearance (ml/min) 0.76 ± 0.06 0.078 ± 0.05a 0.87 ± 0.05b 0.44 ± 0.03a,b 0.65 ± 0.04a,b 0.80 ± 0.05b
BUN (mg/dl) 24.55 ± 0.77 87.44 ± 4.37a 26.87 ± 0.64b 73.65 ± 1.32a,b 53.21 ± 0.9a,b 35.89 ± 0.64 a,b
Urea clearance (ml/min) 0.58 ± 0.04 0.19 ± 0.05a 0.61 ± 0.03b 0.49 ± 0.02a,b 0.53 ± 0.03a,b 0.59 ± 0.03b
Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA)
Effect of CMN on CsA-induced nitrosative stress
Serum and tissue nitrite levels were significantly elevated by CsA-administration. Curcumin treatment significantly and dose dependently improved this increase in nitrite levels both in serum and tissue (Table-2). However, CMN (15 mg/kg) per se had no effect on serum nitrite levels.
Table 2 Effect of CMN on cyclosporine-induced Nitrite levels
Variables Control CsA (20) CMN(15) CsA (20)+ CMN(5) CsA (20)+ CMN(10) CsA (20)+ CMN(15)
Serum Nitrite(μmol/ml) 62 ± 3.72 91.9 ± 50.6a 60 ± 3.15b 77 ± 4.55a,b 69 ± 8.75b 61 ± 3.05b
Tissue nitrite(μmol/mg) 103.518 ± 2.73 190.656 ± 7.97a 101.814 ± 2.27b 174.704 ± 4.01a,b 144.79 ± 3.01a,b 116.912 ± 2.27a,b
Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA)
Effect of CMN on CsA-induced lipid peroxidation
Renal TBARS levels were markedly increased by CsA administration as compared to control group. Treatment with curcumin produced a significant and dose-dependent reduction in TBARS in CsA-treated rats, however curcumin per se did not alter TBARS (Fig. 1).
Figure 1 Effect of curcumin (CMN) on Cyclosporine-induced lipid peroxidation in rat kidney. Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA).
Effect of CMN on CsA-induced changes in the antioxidant profile
Treatment with CsA significantly decreased the reduced glutathione (GSH) levels (Fig. 2) and activities of superoxide dismutase (SOD) (Fig. 3) and catalase (CAT) (Fig. 4). This reduction was significantly and dose-dependetly improved by the treatment with curcumin. However curcumin per se did not alter the endogenous antioxidant profile.
Figure 2 Effect of curcumin (CMN) on Cyclosporine-induced Reduced Glutathione in rat kidney. Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA).
Figure 3 Effect of curcumin (CMN) on Cyclosporine-induced SOD levels in rat kidney. Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA).
Figure 4 Effect of curcumin (CMN) on Cyclosporine-induced catalase levels in rat kidney. Values are expressed mean ± mean. a = Statistical significant at P < 0.05 as compared to control, b = Statistical significant at P < 0.05 as compared to Cyclosporine (CsA).
Effect of CMN on CsA-induced changes on renal morphology
The histopathological changes were graded and summarized in (Table 3). The sections of the control group showed normal glomeruli, afferent arterioles, and tubule cells. By contrast, the kidneys of rats treated with CsA showed marked histological changes in the cortex and outer medulla. The renal sections showed marked tubulointerstital fibrosis, severe epical blebbing and hyaline casts and glomerular basement thickening. Treatment with CMN preserved the normal morphology of the kidney and shows normal glomeruli, no cast formation and slight oedema of the tubular cells.
Table 3 Effect of curcumin (15 mg/kg) treatment on morphological changes as assessed by histopathological examination of kidney in Cyclosporine treated rats
Group Tubular brush border loss Interstitial oedema Tubular dilatation Necrosis of epithelium Hyaline casts
Control - - - - -
CsA +++ +++ +++ +++ +++
CMN+ CsA +/- +/- +/- +/- +/-
CMN - - - - -
Discussion
The exact mechanism of CsA-induced hypertension and nephrotoxicity remain obscure but several studies suggest that a defect in intracellular calcium handling [27], magnesium deficiency [28], oxidative stress [29,30], and nitric oxide (NO) system [31] are involved. Acute renal failure due to CsA is widely attributed to the generation of reactive oxygen species (ROS) by CsA.
It has been reported that binding of pimonidazole, a hypoxia marker in the kidneys, was increased nearly threefold by CsA, indicating marked tissue hypoxia [32]. Moreover, free radicals in the urine were increased dramatically after CsA treatment. [7]. It is also known that CsA increases renal nerve activity resulting in vasoconstricton in the kidney [33]. In addition, CsA causes vasoconstriction directly in isolated renal arterioles [34,35]. It has been demonstrated that CsA blocks mitochondrial Calcium (Ca+2) release, inducing a drastic enhancement in intracellular free Ca+2, which could account for the vasoconstrictive effect of CsA [36,37]. These alterations could theoretically lead to a classical hypoxia-reoxygenation injury involving oxygen free radicals. In addition, ROS could be derived directly from CsA or during its metabolism by the CYP450 system [6]. It has been demonstrated that cyclosporine increased level of superoxide (O2-) in endothelial and mesangial cells [9]. Studies show that CsA-induced local production of hydroxyl radical, a highly active and detrimental radical, plays an important role in CsA nephrotoxicity [38].
Couple of studies suggested that CsA induces apoptosis characterized by internucleosomal DNA cleavage due to endonuclease activation, chromatin condensation, and apoptotic bodies in hematopoietic cells [39,40]. Because oxidants are capable of inducing apoptosis in various types of cells [41], including renal tubular epithelial cells [42]. It is conceivable that reactive oxygen metabolites may play a role in apoptotic mechanism of CsA-induced nephrotoxicity.
The present study revealed that chronic administration of CsA for 21 days caused a marked impairment of renal function alongwith significant oxidative stress in the kidneys. Curcumin significantly and dose-dependently improved creatinine and urea clearance, and decreased the elevated levels of serum creatinine and BUN. Earlier studies have also shown that CMN pretreatment decreases ischemia-reperfusion induced rise in serum creatinine levels in kidney [43]. Chronic administration of CsA also produced oxidative stress and increased the lipid peroxidation in kidneys as is seen by the renal TBARS levels. This effect of CsA was again ameliorated by CMN treatment and is in line with various previous reports, which show that CMN decreases lipid peroxidation possibly by its antioxidant mechanism [44]. Oxidative stress can promote the formation of a variety of vasoactive mediators that can affect renal function directly by causing renal vasoconstriction or decreasing the glomerular capillary ultrafiltration coefficient; and thus reducing glomerular filtration rate [45]. Thus the attenuation of lipid peroxidation in CsA-treated rats by CMN provides a convincing evidence for the involvement of ROS in CsA-induced lipid peroxidation. Rukkumani et al. [46] reported protective effect of CMN on circulating lipids in plasma and lipid peroxidation products in alcohol and polyunsaturated fatty acid-induced toxicity. In-vitro findings support the hypothesis that CMN inhibits free radical induced apoptosis in cell lines [47]. Sreejayan et al claimed that the CMN inhibit iron-catalyzed lipid peroxidation in rat brain tissue homogenates by chelation of iron[48].
More and more studies now established the ability of CMN to mainly eliminate the hydroxyl radical [49], superoxide radical [50], singlet oxygen[51], nitrogen dioxide[52] and NO[53]. It has also been demonstrated that CMN inhibits the generation of the superoxide radical[54]. In our study, CsA administration caused marked deterioration of endogenous antioxidant profile as evidenced by decrease in SOD and CAT activities, an effect which was effectively reversed by CMN treatment. Vajragupta et al., [23] have reported that CMN manganese complex and acetylcurcumin manganese complex, low molecular weight synthetic compounds, showed much greater SOD activity and an inhibitory effect on lipid peroxidation. Priyadarsini et al. [55] have shown, by DPPH scavenging in vitro, that origin of the antioxidant activity of CMN is mainly from the phenolic OH group, although a small fraction may be due to the >CH2 site.
Further GSH, a major nonprotein thiol in living organisms plays a crucial role in coordinating the body's antioxidant defense processes. Results in the present study indicate that CsA administration drastically lowered the levels of GSH in the kidney. Improvement of renal GSH levels in CMN treated rats in comparison to CsA administered rats further demonstrates the anti-antioxidative effect of CMN. CMN has been shown to increase the levels of glutathione reductase in ischemic brains of rats as well as alveolar and human leukemia cell [20,56,57]. Chronic treatment of CMN also improved the levels of two key antioxidant enzymes SOD and catalase in CsA administered rats.
Peroxynitrite anions have been generated by the reaction of nitric oxide with superoxide anion. These peroxynitrite anions oxidize biomolecules, which finally leads to lipid peroxidation and tubular cell damage [58]. Large amounts of nitric oxide can lead to the depletion of cellular ATP which can inactivate enzymes that contain iron-sulfur clusters, such enzymes involved in mitochondrial electron transport [59]. Nitrosylation of sulfhydryl groups or tyrosine residues in proteins may impair the functional properties of these proteins. Nitric oxide damages DNA, and this in turn, stimulates the DNA repair enzyme poly-ADP-ribose synthetase [60]. Studies done by Amore and colleagues demonstrate that CsA induces apoptosis in various renal cell lines, and this effect is mediated by the induction of iNOS [61]. In line with studies where CMN is reported to inhibit iNOS gene expression in isolated BALB/c mouse peritoneal macrophages and also in the livers of lipopolysaccharide injected mice [62], our study shows that CsA-induced nitrosative stress was significantly and dose dependently attenuated by CMN. Very recently, Sumanont [24] have studied the effect of CMN and its analogues on peroxynitrite anions scavenging activity in vitro using sodium nitroprusside (SNP) generating nitric oxide system. All compounds effectively reduced the generation of NO radicals in a dose dependent manner. They exhibited strong NO radical scavenging activity with low IC(50) values. It is also known that ROS mediates peroxidation of lipid structures of the tissue, resulting in subcellular damage, as observed in histopathological examination. In our study, the kidney of CsA treated rats has shown characteristic morphological findings such as interstial fibrosis and arteriolar hyalinosis. The vasoconstriction induced by CsA produces an ischemic local environment, which leads to a number of cellular changes such as deterioration in membrane integrity the marked histological changes are prominent in the outer cortex and medullary region of the kidney. Because limited oxygen availabity these structures are particularly vulnerable to ischemia. These changes were not observed in the group treated CMN (15 mg/kg) suggesting the protective effect of CMN in attenuating CsA-induced morphological changes.
Conclusion
In conclusion this study demonstrates that CMN through its marked antioxidant activity coupled with favorable haemodynamic effects salvages CsA nephrotoxicity.
Methods
Animals
Wistar albino rats of either sex (150–200 g) were housed in 3 per cage, with food and water ad libitum for several days before the beginning of the experiment. The animals were kept on straw bedding in animal quarters with a natural light: dark cycle. The animals had free access to standard rodent food pellets and water. Animals were acclimatized to the laboratory conditions one day before the start of experiment and daily at least for one hour before the experiment. All the experiments were conducted between 09.00 and 17.00 hrs. The experimental protocols were approved by the Panjab University Animal Ethical Committee.
Drugs
Curcumin (Sigma Chemicals USA) was suspended in 0.5% Carboxy methyl cellulose (CMC) and administered orally. CsA was a gift from Panacea Biotech India.
Study design
Rats were divided into six groups each consisting of 5 to 6 animal. Group I received vehicle of CsA i.e. olive oil, subcutaneously (s.c.) and 0.5% Carboxy methyl cellulose (CMC, vehicle for CMN) orally for 21 days. Group II received CsA (20 mg/kg/day, s.c.) dissolved in olive oil for 21 days. This group served as positive control. Three different doses of CMN were tested in Group III, IV, V in which animals received both CsA (20 mg/kg/day s.c) and CMN 5,10,15 mgkg-1 respectively for 21 days. A VI group received only CMN 15 mgkg-1 for 21 days so as to see its per se effect. CsA dose was selected from previous studies done in our laboratory. On 21st day of CsA treatment, animals were immediately kept in individual metabolic cages after drug administration for collection of urine. The animals were sacrificed after 24 hr and all the estimations were done as described later.
Assessment of renal functions
Before sacrifice, rats were kept individually in metabolic cages for 24 h to collect urine for estimation of renal function. A midline abdominal incision was performed and both the kidneys were isolated, the left kidney was deep frozen till further enzymatic analysis, whereas, the right kidney was stored in 10% formalin for the histological studies. Plasma samples were assayed for blood urea nitrogen (BUN), urea clearance, serum creatinine & creatinine clearance by using standard diagnostic kits (Span Diagnostics, Gujarat, India).
Assessment of oxidative stress
Post mitochondrial supernatant preparation (PMS)
Kidneys were, perfused with ice cold saline (0.9% sodium chloride) and homogenized in chilled potassium chloride (1.17%) using a homogenizer. The homogenates were centrifuged at 800 g for 5 minutes at 4°C to separate the nuclear debris. The supernatant so obtained was centrifuged at 10,500 g for 20 minutes at 4°C to get the post mitochondrial supernatant which was used to assay catalase and superoxide dismutase (SOD) activity.
Estimation of lipid peroxidation
The malondialdehyde (MDA) content, a measure of lipid peroxidation, was assayed in the form of thiobarbituric acid reacting substances (TBARS) by the method of Okhawa et al. [63]. Briefly, the reaction mixture consisted of 0.2 ml of 8.1% sodium lauryl sulphate, 1.5 ml of 20% acetic acid solution adjusted to pH 3.5 with sodium hydroxide and 1.5 ml of 0.8% aqueous solution of thiobarbituric acid was added to 0.2 ml of 10%(w/v) of PMS. The mixture was brought up to 4.0 ml with distilled water and heated at 95°C for 60 minutes. After cooling with tap water, 1.0 ml distilled water and 5.0 ml of the mixture of n-butanol & pyridine (15:1 v/v) was added and centrifuged. The organic layer was taken out and its absorbance was measured at 532 nm. TBARS were quantified using an extinction coefficient of 1.56 × 105 M-1/cm-1 and expressed as nmol of TBARS per mg protein. Tissue protein was estimated using Biuret method of protein assay and the renal MDA content expressed as nanomoles of malondialdehyde per milligram of protein.
Estimation of reduced glutathione
Reduced glutathione (GSH) in the kidneys was assayed by the method of Jollow et al [64]. Briefly, 1.0 ml of PMS (10%) was precipitated with 1.0 ml of sulphosalicylic acid (4%). The samples were kept at 4°C for at least 1 hour and then subjected to centrifugation at 1200 g for 15 minutes at 4°C. The assay mixture contained 0.1 ml filtered aliquot and 2.7 ml phosphate buffer (0.1 M, pH 7.4) in a total volume of 3.0 ml. The yellow colour developed was read immediately at 412 nm on a spectrophotometer.
Estimation of superoxide desmutase(SOD)
SOD activity was assayed by the method of Kono et al[65] The assay system consisted of EDTA 0.1 mM, sodium carbonate 50 mM and 96 mM of nitro blue tetrazolium (NBT). In the cuvette, 2 ml of above mixture, 0.05 ml hydroxylamine and 0.05 ml of PMS were taken and the auto-oxidation of hydroxylamine was observed by measuring the absorbance at 560 nm.
Estimation of catalase
Catalase activity was assayed by the method of Claiborne et al [66]. Briefly, the assay mixture consisted of 1.95 ml phosphate buffer (0.05 M, pH 7.0), 1.0 ml hydrogen peroxide (0.019 M) and 0.05 ml PMS (10%) in a final volume of 3.0 ml. Changes in absorbance were recorded at 240 nm. Catalase activity was calculated in terms of k minutes-1.
Assessment of serum/tissue nitrite concentration
Serum and tissue nitrite was estimated using Greiss reagent and served as an indicator of NO production. 500 μl of Greiss reagent (1:1 solution of 1% sulphanilamide in 5% phosphoric acid and 0.1% napthaylamine diamine dihydrochloric acid in water) was added to suitably diluted 100 μl of plasma and absorbance was measured at 546 nm [67]. Nitrite concentration was calculated using a standard curve for sodium nitrite. Nitrite levels were expressed as μmol/ml in serum and as μmol/mg protein in homogenate.
Histopathological examination
For microscopic evaluation kidney were fixed in 10% neutral phosphatebuffered formalin solution. Following dehydration in ascending series of ethanol (70, 80, 96, 100%), tissue samples were cleared in xylene and embedded in paraffin. Tissue sections of 5 μm were stained with hematoxylin-eosin (H-E). A minimum of 10 fields for each kidney slide were examined and assigned for severity of changes by an observer blinded to the treatments of the animals and assigned for severity of changes using Scores of none (-), mild (+), Moderate (++) and Severe (+++)
Statistical analysis
Results were expressed as mean± SEM. The intergroup variation was measured by one way analysis of variance (ANOVA) followed by Fischer's LSD test. Statistical significance was considered at p < 0.05. The statistical analysis was done using the Jandel Sigma Stat Statistical Software version 2.0.
Authors' contributions
Naveen Tirkey, Gangandeep kaur and Garima Vij did all the biochemical estimations in kidney and did the data interpretation after statistical analysis. Kanwaljit Chopra contributed in manuscript preparation.
Figure 5 (A) Hematoxylin and Eosin-stained sections of Normal rat kidneys. (B) Kidney section of CsA treated rats showing tubular brush-border loss, interstitial oedema, Necrosis of epithelium and Hyaline Casts. (C) Kidney Section of CMN (15 mg/kg p.o) + CsA treated rats showing prevention of CsA induced alterations. (D) Kidney section of CMN (15 mg/kg p.o.) treated rats showing almost normal morphology.
Acknowledgements
The grants from University Grants commission for conducting the study is gratefully acknowledged. The authors like to express their thanks to Ms Saraswati Gupta, Senior Technical officer, University Institute of Pharmaceutical sciences Panjab University Chandigarh for her help in conducting the spectrophotometric analysis. The authors also like to thank Dr (Mrs) Anju Bhandari(MBBS, MD (pathology)), of Navjeevan Clinical Laboratory Chandigarh for helping in performing the histological studies.
==== Refs
Buurman WA Daemen AJ van der Linden CJ Kootstra G Clinically used concentrations of cyclosporine A only partially inhibit interferon-gamma production by activated T lymphocytes Transplant Proc 1987 19 1193 3152624
Kahan BD Cyclosporine N Engl J Med 1989 321 1725 1738 2687689
Mason J Pharmacology of cyclosporine (sandimmune). VII. Pathophysiology and toxicology of cyclosporine in humans and animals Pharmacol Rev 1990 41 423 434 2185493
Remuzzi G Perico N Cyclosporine-induced renal dysfunction in experimental animals and humans Kidney Int Suppl 1995 52 S70 4 8587288
Wolf A Clemann N Frieauff W Ryffel B Cordier A Role of reactive oxygen formation in the cyclosporin-A-mediated impairment of renal functions Transplant Proc 1994 26 2902 2907 7940917
Ahmed SS Napoli KL Strobel HW Oxygen radical formation during cytochrome P450-catalyzed cyclosporine metabolism in rat and human liver microsomes at varying hydrogen ion concentrations Mol Cell Biochem 1995 151 131 140 8569758 10.1007/BF01322335
Knight JA Cheung AK Pieper RK Servilla K Increased urinary lipoperoxide levels in renal transplant patients Ann Clin Lab Sci 1989 19 238 241 2667454
Inselmann G Hannemann J Baumann K Cyclosporine A induced lipid peroxidation and influence on glucose-6-phosphatase in rat hepatic and renal microsomes Res Commun Chem Pathol Pharmacol 1990 68 189 203 2162073
Perez de Lema G Arribas-Gomez I Ruiz-Gines JA de Arriba G Prieto A Rodriguez-Puyol D Rodriguez-Puyol M Reactive oxygen species mediate the effects of cyclosporine A on human cultured mesangial cells Transplant Proc 1997 29 1241 1243 9123290 10.1016/S0041-1345(96)00482-4
Wang C Salahudeen AK Lipid peroxidation accompanies cyclosporine nephrotoxicity: effects of vitamin E Kidney Int 1995 47 927 934 7752594
Walker G Kunz D Pignat W Wiesenberg I Van den Bosch H Pfeilschifter J Tetranactin inhibits interleukin 1 beta and cAMP induction of group II phospholipase A2 in rat renal mesangial cells Eur J Pharmacol 1996 306 265 270 8813640 10.1016/0014-2999(96)00203-8
Longoni B Giovannini L Migliori M Bertelli AA Bertelli A Cyclosporine-induced lipid peroxidation and propionyl carnitine protective effect Int J Tissue React 1999 21 7 11 10463135
Ammon HP Wahl MA Pharmacology of Curcuma longa Planta Med 1991 57 1 7 2062949
Joe B Lokesh BR Effect of curcumin and capsaicin on arachidonic acid metabolism and lysosomal enzyme secretion by rat peritoneal macrophages Lipids 1997 32 1173 1180 9397403
Satoskar RR Shah SJ Shenoy SG Evaluation of anti-inflammatory property of curcumin (diferuloyl methane) in patients with postoperative inflammation Int J Clin Pharmacol Ther Toxicol 1986 24 651 654 3546166
Taher MM Lammering G Hershey C Valerie K Curcumin inhibits ultraviolet light induced human immunodeficiency virus gene expression Mol Cell Biochem 2003 254 289 297 14674708 10.1023/A:1027393719610
De Clercq E Current lead natural products for the chemotherapy of human immunodeficiency virus (HIV) infection Med Res Rev 2000 20 323 349 10934347 10.1002/1098-1128(200009)20:5<323::AID-MED1>3.0.CO;2-A
Pal A Pal AK Studies on the genotoxicity of turmeric extracts in bacterial system Int J Antimicrob Agents 2000 16 415 417 11118850 10.1016/S0924-8579(00)00268-5
Jurgens TM Frazier EG Schaeffer JM Jones TE Zink DL Borris RP Nanakorn W Beck HT Balick MJ Novel nematocidal agents from Curcuma comosa J Nat Prod 1994 57 230 235 8176399 10.1021/np50104a006
Biswas SK McClure D Jimenez LA Megson IL Rahman I Curcumin induces glutathione biosynthesis and inhibits NF-kappaB activation and interleukin-8 release in alveolar epithelial cells: mechanism of free radical scavenging activity Antioxid Redox Signal 2005 7 32 41 15650394 10.1089/ars.2005.7.32
Iqbal M Sharma SD Okazaki Y Fujisawa M Okada S Dietary supplementation of curcumin enhances antioxidant and phase II metabolizing enzymes in ddY male mice: possible role in protection against chemical carcinogenesis and toxicity Pharmacol Toxicol 2003 92 33 38 12710595 10.1034/j.1600-0773.2003.920106.x
Balasubramanyam M Koteswari AA Kumar RS Monickaraj SF Maheswari JU Mohan V Curcumin-induced inhibition of cellular reactive oxygen species generation: novel therapeutic implications J Biosci 2003 28 715 721 14660871
Vajragupta O Boonchoong P Watanabe H Tohda M Kummasud N Sumanont Y Manganese complexes of curcumin and its derivatives: evaluation for the radical scavenging ability and neuroprotective activity Free Radic Biol Med 2003 35 1632 1644 14680686 10.1016/j.freeradbiomed.2003.09.011
Sumanont Y Murakami Y Tohda M Vajragupta O Matsumoto K Watanabe H Evaluation of the nitric oxide radical scavenging activity of manganese complexes of curcumin and its derivative Biol Pharm Bull 2004 27 170 173 14758027 10.1248/bpb.27.170
Kim JE Kim AR Chung HY Han SY Kim BS Choi JS In vitro peroxynitrite scavenging activity of diarylheptanoids from Curcuma longa Phytother Res 2003 17 481 484 12748983 10.1002/ptr.1179
Das KC Das CK Curcumin (diferuloylmethane), a singlet oxygen ((1)O(2)) quencher Biochem Biophys Res Commun 2002 295 62 66 12083767 10.1016/S0006-291X(02)00633-2
Cheng CH Hsieh CL Shu KH Chen YL Chen HC Effect of calcium channel antagonist diltiazem and calcium ionophore A23187 on cyclosporine A-induced apoptosis of renal tubular cells FEBS Lett 2002 516 191 196 11959131 10.1016/S0014-5793(02)02563-2
Mervaala EM Pere AK Lindgren L Laakso J Teravainen TL Karjala K Vapaatalo H Ahonen J Karppanen H Effects of dietary sodium and magnesium on cyclosporin A-induced hypertension and nephrotoxicity in spontaneously hypertensive rats Hypertension 1997 29 822 827 9052902
McGrath LT Treacy R McClean E Brown JH Oxidative stress in cyclosporin and azathioprine treated renal transplant patients Clin Chim Acta 1997 264 1 12 9267698 10.1016/S0009-8981(97)00088-0
Satyanarayana PS Chopra K Oxidative stress-mediated renal dysfunction by cyclosporine A in rats: attenuation by trimetazidine Ren Fail 2002 24 259 274 12166693 10.1081/JDI-120005360
De Nicola L Thomson SC Wead LM Brown MR Gabbai FB Arginine feeding modifies cyclosporine nephrotoxicity in rats J Clin Invest 1993 92 1859 1865 8408638
Zhong Z Arteel GE Connor HD Yin M Frankenberg MV Stachlewitz RF Raleigh JA Mason RP Thurman RG Cyclosporin A increases hypoxia and free radical production in rat kidneys: prevention by dietary glycine Am J Physiol 1998 275 F595 604 9755131
Moss NG Powell SL Falk RJ Intravenous cyclosporine activates afferent and efferent renal nerves and causes sodium retention in innervated kidneys in rats Proc Natl Acad Sci U S A 1985 82 8222 8226 3865224
Lanese DM Falk SA Conger JD Sequential agonist activation and site-specific mediation of acute cyclosporine constriction in rat renal arterioles Transplantation 1994 58 1371 1378 7809931
Lanese DM Conger JD Effects of endothelin receptor antagonist on cyclosporine-induced vasoconstriction in isolated rat renal arterioles J Clin Invest 1993 91 2144 2149 8486781
Avdonin PV Cottet-Maire F Afanasjeva GV Loktionova SA Lhote P Ruegg UT Cyclosporine A up-regulates angiotensin II receptors and calcium responses in human vascular smooth muscle cells Kidney Int 1999 55 2407 2414 10354289 10.1046/j.1523-1755.1999.00481.x
Lo Russo A Passaquin AC Andre P Skutella M Ruegg UT Effect of cyclosporin A and analogues on cytosolic calcium and vasoconstriction: possible lack of relationship to immunosuppressive activity Br J Pharmacol 1996 118 885 892 8799558
Zhong Z Connor HD Yin M Moss N Mason RP Bunzendahl H Forman DT Thurman RG Dietary glycine and renal denervation prevents cyclosporin A-induced hydroxyl radical production in rat kidney Mol Pharmacol 1999 56 455 463 10462532
Gottschalk AR Boise LH Thompson CB Quintans J Identification of immunosuppressant-induced apoptosis in a murine B-cell line and its prevention by bcl-x but not bcl-2 Proc Natl Acad Sci USA 1994 91 7350–7354 7518929
Kitagaki K Niwa S Hoshiko K Nagai H Hayashi S Totsuka T Augmentation of apoptosis in bronchial exuded rat eosinophils by cyclosporin A Biochem Biophys Res Commun 1996 222 71 77 8630077 10.1006/bbrc.1996.0715
Buttke TM Sandstrom PA Oxidative stress as a mediator of apoptosis Immunol Today 1994 15 7–10 8136014 10.1016/0167-5699(94)90018-3
Ueda N Shah SV Endonuclease-induced DNA damage and cell death in oxidant injury to renal tubular epithelial cells J Clin Invest 1992 90 2593–2597 1334979
Shoskes DA Effect of bioflavonoids quercetin and curcumin on ischemic renal injury: a new class of renoprotective agents Transplantation 1998 66 147 152 9701255 10.1097/00007890-199807270-00001
Skrzydlewska E Ostrowska J Farbiszewski R Michalak K Protective effect of green tea against lipid peroxidation in the rat liver, blood serum and the brain Phytomedicine 2002 9 232 238 12046864 10.1078/0944-7113-00119
Garcia-Cohen EC Marin J Diez-Picazo LD Baena AB Salaices M Rodriguez-Martinez MA Oxidative stress induced by tert-butyl hydroperoxide causes vasoconstriction in the aorta from hypertensive and aged rats: role of cyclooxygenase-2 isoform J Pharmacol Exp Ther 2000 293 75 81 10734155
Rukkumani R Sri Balasubashini M Menon VP Protective effects of curcumin and photo-irradiated curcumin on circulatory lipids and lipid peroxidation products in alcohol and polyunsaturated fatty acid-induced toxicity Phytother Res 2003 17 925 929 13680826 10.1002/ptr.1254
Somasundaram S Edmund NA Moore DT Small GW Shi YY Orlowski RZ Dietary curcumin inhibits chemotherapy-induced apoptosis in models of human breast cancer Cancer Res 2002 62 3868 3875 12097302
Sreejayan Rao MN Curcuminoids as potent inhibitors of lipid peroxidation J Pharm Pharmacol 1994 46 1013 1016 7714712
Reddy AC Lokesh BR Studies on the inhibitory effects of curcumin and eugenol on the formation of reactive oxygen species and the oxidation of ferrous iron Mol Cell Biochem 1994 137 1 8 7845373 10.1007/BF00926033
Sreejayan N Rao MN Free radical scavenging activity of curcuminoids Arzneimittelforschung 1996 46 169 171 8720307
Rao CV Rivenson A Simi B Reddy BS Chemoprevention of colon carcinogenesis by dietary curcumin, a naturally occurring plant phenolic compound Cancer Res 1995 55 259 266 7812955
Unnikrishnan MK Rao MN Curcumin inhibits nitrogen dioxide induced oxidation of hemoglobin Mol Cell Biochem 1995 146 35 37 7651374 10.1007/BF00926878
Sreejayan Rao MN Nitric oxide scavenging by curcuminoids J Pharm Pharmacol 1997 49 105 107 9120760
Ruby AJ Kuttan G Babu KD Rajasekharan KN Kuttan R Anti-tumour and antioxidant activity of natural curcuminoids Cancer Lett 1995 94 79 83 7621448 10.1016/0304-3835(95)03827-J
Priyadarsini KI Maity DK Naik GH Kumar MS Unnikrishnan MK Satav JG Mohan H Role of phenolic O-H and methylene hydrogen on the free radical reactions and antioxidant activity of curcumin Free Radic Biol Med 2003 35 475 484 12927597 10.1016/S0891-5849(03)00325-3
Chen J Wanming D Zhang D Liu Q Kang J Water-soluble antioxidants improve the antioxidant and anticancer activity of low concentrations of curcumin in human leukemia cells Pharmazie 2005 60 57 61 15700780
Thiyagarajan M Sharma SS Neuroprotective effect of curcumin in middle cerebral artery occlusion induced focal cerebral ischemia in rats Life Sci 2004 74 969 985 14672754 10.1016/j.lfs.2003.06.042
Stamler JS Singel DJ Loscalzo J Biochemistry of nitric oxide and its redox-activated forms Science 1992 258 1898 1902 1281928
Rivas-Cabanero L Rodriguez-Lopez AM Martinez-Salgado C Saura M Lamas S Lopez-Novoa JM Gentamicin treatment increases mesangial cell nitric oxide production Exp Nephrol 1997 5 23 30 9052845
Dawson TM Dawson VL Snyder SH A novel neuronal messenger molecule in brain: the free radical, nitric oxide Ann Neurol 1992 32 297 311 1384420 10.1002/ana.410320302
Amore A Emancipator SN Cirina P Conti G Ricotti E Bagheri N Coppo R Nitric oxide mediates cyclosporine-induced apoptosis in cultured renal cells Kidney Int 2000 57 1549 1559 10760090 10.1046/j.1523-1755.2000.00999.x
Chan MM Huang HI Fenton MR Fong D In vivo inhibition of nitric oxide synthase gene expression by curcumin, a cancer preventive natural product with anti-inflammatory properties Biochem Pharmacol 1998 55 1955 1962 9714315 10.1016/S0006-2952(98)00114-2
Ohkawa H Ohishi N Yagi K Assay for lipid peroxides in animal tissues by thiobarbituric acid reaction Anal Biochem 1979 95 351 358 36810 10.1016/0003-2697(79)90738-3
Jollow D Mitchell L Zampaglione N Gillete J Bromobenze induced liver necrosis: protective role of glutathione and evidence for 3,4-bromobenzenoxide as the hepatotoxic intermediate Pharmacol 1974 11 151 169
Kono Y Generation of superoxide radical during autoxidation of hydroxylamine and an assay for superoxide dismutase. Arch Biochem Biophys 1978 186 189 195 24422 10.1016/0003-9861(78)90479-4
Claiborne A Boca Raton FL Handbook of Methods for Oxygen Radical Reaserch 1985 CRC Press 283 284
Green LC Wagner DA Glogowski J Skipper PL Wishnok JS Tannenbaum SR Analysis of nitrate, nitrite, and [15N]nitrate in biological fluids Anal Biochem 1982 126 131 138 7181105 10.1016/0003-2697(82)90118-X
|
16225695
|
PMC1277828
|
CC BY
|
2021-01-04 16:33:00
|
no
|
BMC Pharmacol. 2005 Oct 15; 5:15
|
utf-8
|
BMC Pharmacol
| 2,005 |
10.1186/1471-2210-5-15
|
oa_comm
|
==== Front
BMC SurgBMC Surgery1471-2482BioMed Central London 1471-2482-5-221624202410.1186/1471-2482-5-22Research ArticleMissing effects of zinc in a porcine model of recurrent endotoxemia Krones Carsten J [email protected] Bernd [email protected] Michael [email protected] Michael [email protected] Uwe [email protected] Alexander P [email protected] Volker [email protected] Department of Surgery, Technical University of Aachen, Pauwelsstr. 30, 52074 Aachen, Germany2 Institute of Pathology, Technical University of Aachen, Pauwelsstr. 30, 52074 Aachen, Germany3 Joint Institute for Surgical Research, Leninskie Gory, Moscow 119992, Russian Federation2005 20 10 2005 5 22 22 17 5 2005 20 10 2005 Copyright © 2005 Krones et al; licensee BioMed Central Ltd.2005Krones 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
Chronic human sepsis often is characterised by the compensatory anti-inflammatory response syndrome (CARS). During CARS, anti-inflammatory cytokines depress the inflammatory response leading to secondary and opportunistic infections. Proved in vitro as well as in vivo, zinc's pro-inflammatory effect might overcome this depression.
Methods
We used the model of porcine LPS-induced endotoxemia established by Klosterhalfen et al. 10 pigs were divided into two groups (n = 5). Endotoxemia was induced by recurrent intravenous LPS-application (1.0 μg/kg E. coli WO 111:B4) at hours 0, 5, and 12. At hour 10, each group received an intravenous treatment (group I = saline, group II = 5.0 mg/kg elementary zinc). Monitoring included hemodynamics, blood gas analysis, and the thermal dilution technique for the measurement of extravascular lung water and intrapulmonary shunt. Plasma concentrations of IL-6 and TNF-alpha were measured by ELISA. Morphology included weight of the lungs, width of the alveolar septae, and rate of paracentral liver necrosis.
Results
Zinc's application only trended to partly improve the pulmonary function. Compared to saline, significant differences were very rare. IL-6 and TNF-alpha were predominately measured higher in the zinc group. Again, significance was only reached sporadically. Hemodynamics and morphology revealed no significant differences at all.
Conclusion
The application of zinc in this model of recurrent endotoxemia is feasible and without harmful effects. However, a protection or restoration of clinical relevance is not evident in our setting. The pulmonary function just trends to improve, cytokine liberation is only partly activated, hemodynamics and morphology were not influenced. Further pre-clinical studies have to define zinc's role as a therapeutic tool during CARS.
==== Body
Background
Although most patients survive the initial insult of a hyper-inflammatory sepsis syndrome, they remain at an increased risk of secondary or opportunistic infections. Basic course is the compensatory anti-inflammatory response syndrome (CARS) which derives from sustained or chronic stages of sepsis [1,2]. During CARS, anti-inflammatory members of the cytokine network as IL-10, TGF-β, and IL-1RA depress the inflammatory response, thus reversing the effects of pro-inflammatory cytokines as TNFα, IL-1, and IL-6 [3-10]. Accompanied by a macrophage deactivation, a T-cell anergy, and an apoptotic loss of lymphoid tissues, the human clinical syndrome of CARS is characterised by uncontrolled infections followed by organ damages associated with high rates of morbidity and mortality [2].
Zinc's pro-inflammatory impact is reliably proved in various in vitro as well as in vivo models. In separated human mononuclear cells and monocytes, zinc induces the expression of cytokines [11]. In pigs, zinc upregulates the release of inflammatory mediators by the induction the heat shock response [12]. And in a rat model of endotoxemia, zinc thus reduces the rates of apoptosis and mortality [13]. During the acute phase of sepsis, this pro-inflammatory capacity can even be deleterious due to complementary properties [14]. As CARS represents an overbalanced anti-inflammatory stage of sepsis, zinc's well-known pro-inflammatory potential could be an interesting tool to recover immune capability.
To verify a rehabilitative effect of zinc on the immune system, this study analyses the impact of zinc on hemodynamics, pulmonary function, and cytokine liberation during a recurrent endotoxemia in a porcine model.
Methods
Animal model
The study was performed in adherence to the National Institutes of Health guidelines for the use of experimental animals observing the Interdisciplinary Principles and Guidelines for the Use of Animals in Research, Testing, and Education. Basically, we used the animal model for endotoxemia in domestic pigs as described by Klosterhalfen et al. [15]. Experiments were performed on female farm pigs (n = 10, Deutsche Landrasse) weighing 28–32 kg. The animals were randomised and divided into 2 groups (n = 5). After premedication (azaperon 3 mg/kg) and intubation (pentobarbital 6 mg/kg) anaesthesia was maintained by pentobarbital 0.1 mg/kg/min and ketamine 0.1 mg/kg/min. Arterial and venous catheters were placed in the right carotid artery and the right jugular vein. In addition, a measurement catheter of the Pulsion Cold System® (PULSION Medical systems, Munich, Germany) was introduced into the femoral artery in the right groin. After the completion of all surgical manipulations, hemodynamics and respiration were stabilised over a period of 60 min. The pigs were fully anaesthetised for the whole study period of 18 hours. Fluid balance was maintained by infusion of Ringer solution (3–6 ml/kg/h). Continuous measuring of the central venous pressure served as control.
The first infusion of LPS defined hour 0 of the examination. For each application of LPS, 1.0 μg/kg E. coli endotoxin WO111:B4 (Difco Laboratories, Detroit, USA) was intravenously infused over 30 min. This LPS-infusion was repeated in each group at hours 5 and 12. To mimic an interim treatment during recurrent septicaemia, each group received an intravenous treatment (group I = saline, group II = zinc) at hour 10. As a result of former dose-finding studies [12,15], 25 mg/kg zinc-bis-(DL-hydrogenaspartate) = 5 mg/kg elementary Zn2+ (Unizink®, Köhler Pharma, Alsbach-Hähnlein, Germany) were infused in group II. According to the suppliers instructions, the aspartate component in zinc-bis-(DL-hydrogenaspartate) has no established impact during endotoxemia. To avoid acute toxic effects, zinc as well as saline infusions were administered over 2 hours [12,15].
Hemodynamics and arterial oxygen saturation were continuously measured by arterial monitoring and pulse oxymetry (Oxyshuttle®, Criticaon, Hamburg, Germany). Due to former studies [12], blood samples for the measurement of respiratory parameters and cytokines (TNFα and IL-6) were registered at defined points of time (600, 615, 630, 645, 660, 720, 735, 750, 765, 780, 840, 930, 1020 min). The extravascular lung water (EVLW) and the arterial-venous intrapulmonary shunt were registered using the thermal dye dilution technique of the Pulsion Cold Z-021® system (PULSION medical systems, München, Germany) [16,17]. In brief, the method uses ice-cold water and indocyanin green as indicators. Whereas cold distributes to the intra- and extravascular volumes, indocyanin green remains intravascular at 99.9%. Both indicators are injected into the right atrium, and concentration changes in time are recorded. Using the different dilution curves of both indicators, rates of the intrapulmonary shunt (Qs/Qt), the extravascular lung water (EVLW index (ml/kg) = intrathoracic thermal volume index – intrathoracic blood volume index), the cardiac output, and the systemic vessel resistance (SVRI = (systemic arterial pressure – central venous pressure) / (cardiac output / body mass)) can be calculated. The measuring points of the system corresponded with the blood gas samples.
At the end of the study period (18 hours), all animals were killed by an intravenous infusion of potassium chloride. A necropsy followed with special regard to the macroscopic findings in the lungs and the liver.
Morphology
After removal, both lungs were weighed to indirectly measure the pulmonary endothelial leakage respectively the capillary leak syndrome. The width of the alveolar septums was defined as another indirect parameter of the interstitial edema. Therefore, multiple tissue samples from all pulmonary lobes were fixed in 10 % buffered formalin, embedded in paraffin, and sections of 4 μm were stained with hematoxylin and eosin. All measurements were quantified by morphometry at 500 different randomized locations per animal (Quantimed 500®, Leica Microsystems, Lübeck, Germany). In addition, the paracentral necrosis rate in the liver was quantified. After the histologic preparation of two representative samples of the right and left liver lobes, measurements were again analyzed by morphometry (Quantimed 500®, Leica Microsystems, Lübeck, Germany).
Enzyme-linked immunosorbent assay (ELISA) of TNFα and IL-6
The plasma concentrations of TNFα and IL-6 were measured by cross-reactive (human) ELISA kits according to the supplier's instructions (Biozol, Hamburg, Germany). All procedures were sandwich ELISAs. The molecules of interest were first bound by immobilized primary monoclonal antibodies, afterwards washed free and finally bound to polyclonal antibodies. The visualization followed by production of color after a peroxidase reaction. The color intensity was proportional to the amount of bound conjugation and thus, to the amount of present cytokine. The absorbance was measured using a Tito-Tek-Multiscan MK" ELISA reader (Flow ICN, Meckenheim, Germany) at 490 or 450 nm comparing the samples with pooled plasma from controls with increasing amounts of recombinant cytokine. The pooled plasma did not contain detectable concentrations of endogenous cytokines.
Statistics
For statistical analysis, the Statistical Package for Social Sciences (SPSS®) software was used. The significance of the pulmonary, hemodynamical, and biochemical parameters was tested by a corrected variance analysis. In case of significant differences, an independent t test followed. P-values < 0.05 were considered significant. All data was expressed as mean ± standard deviation (SD).
Results
Respiratory parameters
After infusion of zinc (t = 600–720 min), the pigs of group II merely trended to an only partly improved pulmonary function compared to group I (saline) (Fig 1A–F). The arterial oxygen pressure was continuously measured higher reaching significance only at two times (t = 630 min and t = 1020 min). The venous oxygen pressure and the arterial oxygen saturation continuously showed higher values in the zinc-group, too. However, only once, the arterial oxygen saturation reached significance (t = 1020 min). The venous oxygen saturation and the arterial and venous carbon dioxide pressure demonstrated no significant differences. After the last infusion of LPS (t = 720 min), group II (zinc) showed steady measurements. In contrast, group I (saline) demonstrated a further slight decrease of the arterial oxygen pressure and the arterial oxygen saturation.
Figure 1 A-F Respiratory parameters: (A) arterial oxygen pressure, (B) venous oxygen pressure, (C) arterial carbon dioxide pressure, (D) venous carbon dioxide pressure, (E) arterial oxygen saturation, and (F) venous oxygen saturation; group I (saline) = triangle and group II (zinc) = circle; *p < 0,05; x-axis nonlinear
Extravascular lung water (EVLW) and intrapulmonary shunt
Analysing the thermal dye dilution data, EVLW and intrapulmonary shunt revealed no significant differences in both groups. In detail, the results were controversial. Group II (zinc) tended to result in higher rates of EVLW compared to group I (saline). In contrast, the intrapulmonary shunt consistently showed lower measurements (Fig 2A–B). However, all differences were not significant.
Figure 2 A-B Course of (A) extravascular lung water (EVLW, ml/kg body weight) and (B) intrapulmonary shunt (Qs/Qt, %); group I (saline) = triangle and group II (zinc) = circle; *p < 0.05; x-axis nonlinear
Hemodynamics
In general, the hemodynamics revealed no impressive differences. Compared to group I (saline), the mean systemic arterial blood pressure (SAP) slightly decreased in group II during the infusion of zinc. Later on, both groups showed almost similar levels without significant differences (Fig 3A). Though zinc obviously induced no gageable effect, the systemic vessel resistance index (SVRI) initially showed lower measurements in group II (zinc). Compared to group I (saline), the difference reached significance at two times (615 and 630 min). Induced by a decline of SVRI in the saline group, both groups finally showed a parallel course (Fig. 3B). Measurements of the cardiac output revealed no remarkable differences at any time (Fig. 3C). Regarding all parameters, the last application of LPS (t = 720 min) induced no further changes in both groups.
Figure 3 A-C Hemodynamics: (A) mean systemic arterial pressure (SAP), (B) systemic vessel resistance index (SVRI), (C) cardiac output; group I (saline) = triangle and group II (zinc) = circle; *p < 0.05; x-axis nonlinear
Inflammatory mediators TNFα and IL-6
After zinc-infusion, TNFα in group II steadily increased compared to group I (saline). Boosted by the last LPS-infusion, this effect even reached significance but persisted only temporary. At the end, both groups showed almost similar results. In group I (saline), the level of TNFα showed no remarkable changes (Fig. 4A).
Figure 4 A, B Course of the cytokines TNFα (A) und IL-6 (B); group I (saline) = triangle and group II (zinc) = circle; *p < 0.05; x-axis nonlinear
The expression of IL-6 was not influenced by zinc. However, the last application of LPS induced a drastic increase of IL-6 in group II (zinc). In parallel to TNFα, this increase temporary reached significance compared to group I (saline). Later on, IL-6 increased to a smaller amount in group I (saline) as well. Caused by a parallel decrease in group II (zinc), both groups finally revealed similar results (Fig. 4B).
The significant increases of TNFα and IL-6 did not correlate with the measured differences in hemodynamics or pulmonary parameters.
Morphology
With an increased alveolar width (group I: 11.3 μm, group II: 12.2 μm) and an elevated rate of the paracentral liver necrosis (group I: 4.3 %, group II: 7.9 %), both groups showed the typical signs of a septic organ damage. Both measurements revealed a broad range and the differences were not significant. The weight of the lungs (group I: 320.8 g, group II: 338.6 g) was measured higher in group II (zinc). Again, the results included high standard errors and did not reach significance.
Discussion
Administered as a prophylaxis, zinc's pro-inflammatory impact protects from septic cellular damages in vivo as well as in vitro [11-13,18-20]. In a porcine model of acute endotoxemia, this pro-inflammatory effect even leads to deleterious results [14]. Thus, zinc could be predestined to restore the depressed immune capability during CARS. To our knowledge, the application of zinc during a recurrent endotoxemia in vivo is not yet described in literature.
To mimic a prolonged and repeated endotoxemia comparable to the human CARS-syndrome, we used the porcine model introduced by Klosterhalfen et al. [15]. The chronic form of endotoxemia was induced by recurrent LPS-infusions. In contrast to high dosage or bolus regimens, this application form more reliably imitates the clinical course of human sepsis [15]. Firstly, the typical septic organ damages found during necropsy confirmed the usability of this model. In both groups, the alveolar width and the rate of the paracentral liver necrosis were increased. Nevertheless, we found no significant differences comparing both study groups. Obviously, the application of zinc did not protect from cellular damages in our setting. Due to former dose finding studies in recurrent endotoxemia, this should rather not be a question of dosage [12,15]. However, kinetics in recurrent and acute endotoxemia are different [12,14,15,18], and zinc's optimal application rate still needs to be defined. In our opinion, the lack of impact more likely could derive from a too short registration time. Possibly, observations limited to 18 hours are insufficient to draw conclusions about the late state of LPS-induced endotoxemia.
In general, our data confirm, that the application of zinc in between a recurrent endotoxemia is well tolerated. With an almost parallel course, hemodynamic changes were rare in both groups. Besides a slight and temporary decrease of the systemic arterial pressure in the zinc-group, we did not find a relevant hemodynamic depression neither induced by zinc nor by LPS. These results contrast the deleterious pro-inflammatory impact of zinc during the acute phase of sepsis in-vivo, where complementing effects severely degraded hemodynamics and lead to an early death after 3 hours [14]. However, we observed no positive effect on the hemodynamic variables as well. Thus, the application of zinc in our model of recurrent endotoxemia in vivo is feasible but without measurable beneficial effects in hemodynamics.
The respiratory parameters reproduced the hemodynamic findings, as remarkable differences in both groups were very rare. Again, the application of zinc was well tolerated. In detail, the animals of group II (zinc) even tended to a slightly improved pulmonary function. However, significant differences of clinical relevance were missing. In our setting, zinc apparently not achieved its protective effect on pulmonary function and cell integrity reported on various studies [18-20]. Besides a longer observation time, this could be clarified by measuring the heat shock response (HSR) representing a possible pathway [13,18,20-23]. Regarding extravascular lung water (EVLW) and intrapulmonary shunt, our results were even conflicting. In conclusion, they could prove neither any protective nor any negative impact associated to zinc. Referring to their prognostic value, these parameters currently rank even higher than simple reflections of the gas exchange [24]. Thus, our findings on this field underline the lack of a clinically relevant pulmonary protection supposed above.
Regarding the liberation of the cytokines, our results remain disappointing. The changes of TNFα and IL-6 after the last LPS-application in the zinc-group revealed an increased liberation of cytokines, but the effect remained short-termed. Significant differences were only sporadically found. Except for a questionable temporary boosting, this phenomenon cannot represent a reliable reactivation of the pro-inflammatory response. Obviously, the inductive potential of zinc on the expression of pro-inflammatory cytokines [11,12,14,19,25] has no long-lasting impact in our model of recurrent endotoxemia. As the infusion of LPS in the saline-group additionally induced no measurable decrease in cytokine liberation, this may be again a question of observation time.
Conclusion
In conclusion, the application of zinc in our experimental setting of recurrent endotoxemia is feasible and induces no harmful effects. However, we could not derive any protective or rehabilitative effect of clinical relevance on respiration, hemodynamics, or cytokine liberation. Though the rate of septic cellular damages in this animal setting is quite similar to those seen in patients with chronic sepsis, it cannot completely reflect the situation in humans. Time course, dosages, and perhaps even the LPS-model must be challenged [26]. As multiple in vitro and in vivo studies as well as clinical trials point out that depressed plasma levels of pro-inflammatory cytokines are accompanied by high rates of lethality [6-8,27-30], the restoration of the immune capability remains a challenging therapeutic aim during prolonged endotoxemia. Despite our deflating results, further pre-clinical studies including long-term observations and measurements of the heat shock response have to be established to define zinc's role in this context.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
C.J. Krones conceived of the study and design, carried out measurements, co-ordination, draft manuscript
B. Klosterhalfen conceived of the study, carried out measurements, co-ordination, reviewed manuscript
M. Anurov participated in study design, carried out measurements, co-ordination
M. Stumpf participated in study design, reviewed results and manuscript
U. Klinge conceived of the study, reviewed manuscript
A. Oettinger conceived of the study, participated in study design, reviewed results and manuscript
V. Schumpelick participated in study design, reviewed manuscript
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Bone RC Sir Isaac Newton, sepsis, SIRS, and CARS Crit Care Med 1996 24 1125 28 8674323 10.1097/00003246-199607000-00010
Oberholzer A Oberholzer C Moldawer LL Sepsis syndromes: Understanding the role of innate and acquired immunity Shock 2001 16 83 96 11508871
Dinarello CA Gelfand JA Wolff SM Anticytokine strategies in the treatment of the systemic inflammatory response syndrome JAMA 1993 269 1829 35 8459516 10.1001/jama.269.14.1829
Doughty L Kaplan S Carcillo J The IL-10 response in pediatric sepsis and organ failure Crit Care Med 1996 24 1137 43 8674325 10.1097/00003246-199607000-00012
Platzer C Meisel CH Vogt K Patzer M Volk HD Upregulation of monocytic IL-10 by tumor necrosis factor-α and cAMP elevating drugs Int Immunol 1995 7 517 23 7547677
Gerard C Bruyns C Marchant A Abramowicz D Vandenbeele P Delvaux A Fiers W Goldman M Velu T Interleukin 10 reduces the release of tumor necrosis factor and prevents lethality in experimental endotoxemia J Exp Med 1993 177 547 50 8426124 10.1084/jem.177.2.547
Howard M Muchamel T Andrade S Menon S Interleukin 10 protects mice from lethal endotoxemia J Exp Med 1993 177 1205 8 8459215 10.1084/jem.177.4.1205
Gomez Jimenez J Martin MC Sauri R Segura RM Esteban F Ruiz JC Nuvials X Boveda JL Peracaula R Salgado A Interleukin-10 and the monocyte/macrophage-induced inflammatory response in septic shock J Infect Dis 1995 171 472 5 7844393
Junger WG Hoyt DB Liu FC Loomis WH Coimbra R Immunosuppression after endotoxin shock: the results of multiple anti-inflammatory factors J Trauma 1996 40 702 9 8614067
Muchamel T Menon S Pisacane P Howard MC Cockayne DA IL-13 protects mice from lipopolysaccharide-induced lethal endotoxemia: correlation with down-modulation of TNF-alpha, IFN-gamma, and IL-12 production J Immunol 1997 158 2898 903 9058827
Driessen C Hirv K Rink L Kirchner H Induction of cytokines by zinc ions in human peripheral blood mononuclear cells and separated monocytes Lymphokine Cytokine Res 1994 13 15 20 8186320
Klosterhalfen B Toens C Hauptmann S Tietze L Offner FA Kuepper W Kirkpatrick CJ Influence of heat shock protein 70 and metallothionein induction by zinc-bis-(DL-hydrogenaspartate) on the release of inflammatory mediators in a porcine model of recurrent endotoxemia Biochem Pharmacol 1996 52 1201 10 8937427 10.1016/0006-2952(96)00469-8
Klosterhalfen B Hauptmann S Offner FA Amo-Takyi B Toens C Winkeltau G Affify M Kuepper W Kirkpatrick CJ Mittermayer C Induction of heat shock protein 70 by zinc-bis-(DL-hydrogenaspartate) reduces cytokine liberation, apoptosis, and mortality rate in a rat model of LD100 endotoxemia Shock 1997 7 254 62 9110410
Krones CJ Klosterhalfen B Fackeldey V Junge K Rosch R Schwab R Stumpf M Klinge U Schumpelick V Deleterious effect of zinc in a pig model of acute endotoxemia J Invest Surg 2004 17 249 56 15385258 10.1080/08941930490502817
Klosterhalfen B Hörstmann-Jungemann K Vogel P Dufhues G Simon B Kalff G Kirkpatrick CJ Mittermayer C Heinrich PC Hemodynamic variables and plasma levels of PGI2, TxA2 and IL-6 in a porcine model of recurrent endotoxemia Circ Shock 1991 35 237 44 1777960
Lewis FR Elings VB Bedside measurement of lung water J Surg Res 1979 27 250 61 384088 10.1016/0022-4804(79)90138-0
Pfeiffer UJ Backus G Blümel G Lewis FR, Pfeiffer UJ A fiberoptics based system for integrated monitoring of cardiac output, intravascular blood volume, extravascular lung water, O2-saturation, and a-v differences Practical applications of fiberoptics in critical care monitoring 1990 Berlin: Springer 114 225
Krones CJ Klosterhalfen B Butz N Hoelzl F Junge K Stumpf M Peiper C Klinge U Schumpelick V Effect of zinc pretreatment on pulmonary endothelial cells in vitro and pulmonary function in a porcine model of endotoxemia J Surg Res 2005 123 251 256 15680386 10.1016/j.jss.2004.08.024
Unoshima M Nishizono A Takita-Sonoda Y Iwasaka H Noguchi T Effects of zinc acetate on splenocytes of endotoxemic mice: enhanced immune response, reduced apoptosis, and increased expression of heat shock protein 70 J Lab Clin Med 2001 137 28 37 11150021 10.1067/mlc.2001.111514
Hennig B Wang Y Ramasamy S McClain CJ Zinc protects against tumor necrosis factor-induced disruption of porcine endothelial cell monolayer integrity J Nutr 1993 123 1003 9 8389399
Villar J Ribeiro SP Mullen BM Kuliszewski M Post M Slutsky AS Induction of heat shock response reduces mortality rate and organ damage in a sepsis-induced acute lung injury Crit Care Med 1994 22 914 22 8205824
Wong HR Ryan M Menendez IY Deneberg A Wispe JR. Heat shock protein induction protects human respiratory epithelium against nitric oxide-mediated cytotoxicity Shock 1997 8 213 8 9377169
Cheng Y Liu Y Liang J Zinc is a potent heat shock protein inducer during liver cold preservation of rats Chin Med J (Engl) 2002 115 1777 9 12622922
Sakka SG Klein M Reinhart K Meier-Hellmann A Prognostic value of extravascular lung water in critically ill patients Chest 2002 122 2080 6 12475851 10.1378/chest.122.6.2080
Klosterhalfen B Hauptmann S Tietze L Toens C Winkeltau G Kuepper W Kirkpatrick CJ The influence of heat shock protein 70 induction on hemodynamic variables in a porcine model of recurrent endotoxemia Shock 1997 7 358 63 9165671
Remick DG Newcomb DE Bolgos GL Call DR Comparison of the mortality and inflammatory response of two models of sepsis: lipopolysaccharide vs. Cecal ligation and puncture Shock 2000 13 110 6 10670840
Florquin S Amraoui Z Abramowicz D Goldman M Systemic release and protective role of IL-10 in staphylococcal enterotoxin B-induced shock in mice J Immunol 1994 153 2618 23 7915740
Arai T Hiromatsu K Nishimura H Kimura Y Kobayashi N Ishida H Nimura Y Yoshikai Y Endogenous interleukin 10 prevents apoptosis in macrophages during salmonella infection Biochem Biophys Res Commun 1995 213 600 7 7646518 10.1006/bbrc.1995.2174
Drazan KE Wu L Bullington D Shaked A Viral IL-10 gene therapy inhibits TNF-alpha and IL-1 beta, not Il-6, in the endotoxemic mouse J Pediatr Surg 1996 31 411 5 8708914 10.1016/S0022-3468(96)90749-6
Cabioglu N Bilgic S Deniz G Aktas E Seyhun Y Turna A Gunay K Esen F Decreased cytokine expression in peripheral blood leukocytes of patients with severe sepsis Arch Surg 2002 137 1037 43 12215155 10.1001/archsurg.137.9.1037
|
16242024
|
PMC1277829
|
CC BY
|
2021-01-04 16:28:04
|
no
|
BMC Surg. 2005 Oct 20; 5:22
|
utf-8
|
BMC Surg
| 2,005 |
10.1186/1471-2482-5-22
|
oa_comm
|
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-301620216810.1186/1475-2867-5-30ReviewABC transporters as multidrug resistance mechanisms and the development of chemosensitizers for their reversal Choi Cheol-Hee [email protected] Research Center for Resistant Cells, Chosun University Medical School, 375 Seosuk-dong, Dong-gu, Gwangju 501-759, South Korea2005 4 10 2005 5 30 30 20 3 2004 4 10 2005 Copyright © 2005 Choi; licensee BioMed Central Ltd.2005Choi; 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.
One of the major problems related with anticancer chemotherapy is resistance against anticancer drugs. The ATP-binding cassette (ABC) transporters are a family of transporter proteins that are responsible for drug resistance and a low bioavailability of drugs by pumping a variety of drugs out cells at the expense of ATP hydrolysis. One strategy for reversal of the resistance of tumor cells expressing ABC transporters is combined use of anticancer drugs with chemosensitizers. In this review, the physiological functions and structures of ABC transporters, and the development of chemosensitizers are described focusing on well-known proteins including P-glycoprotein, multidrug resistance associated protein, and breast cancer resistance protein.
ABC transporterbioavailabilitychemosensitizerdrug resistanceP-glycoproteinmultidrug resistance associated proteinbreast cancer resistance protein.
==== Body
Background
One of the major problems related with anticancer chemotherapy is resistance against anticancer drugs. Some cancers such as non-small cancer, lung cancer, and rectal cancer show what is called primary resistance or natural resistance in which they do not respond to standard chemotherapy drugs from the beginning. On the other hand, many types of sensitive tumors respond well to chemotherapy drugs in the beginning but show acquired resistance later. Experimentally, drug resistance could be very specific to the drug used due to abnormal genetic machinery such as gene amplification within tumor cells in many cases. Multidrug resistance (MDR) is especially problematic in acquired drug resistance. MDR is the phenomenon in which cancer cells exposed to one anticancer drug show resistance to various anticancer drugs that are structurally and functionally different from the initial anticancer drug. The most investigated mechanisms with known clinical significance are: a) activation of transmembrane proteins effluxing different chemical substances from the cells; b) activation of the enzymes of the glutathione detoxification system; c) alterations of the genes and the proteins involved into the control of apoptosis (especially p53 and Bcl-2). The cell membrane, cytoplasm, and nuclear protein participate in these resistance mechanisms [1]. The resistance mechanism is called typical MDR or classical MDR when overexpression of the membrane efflux pumps is involved in MDR. The classical MDR is due mostly to increased efflux pumps in the cell membrane of cells pumping anticancer drugs out of cells. The most typical efflux pumps in the cell membrane is P-glycoprotein (Pgp) [2] having the molecular weight of 170 KD, due to the gene amplification of the normal human gene, MDR1. The efflux pump Pgp is responsible for transporting various xenobiotics (not limited to anticancer drugs) out of cells by using ATP (Fig. 1) [3]. Pgp is one of the membrane transporter superfamily having the ATP-binding cassette (ABC) with well-preserved homology of the site where ATP binds. There are more than 100 ABC transporters distributed from prokaryotes to humans. Forty-eight ABC genes have been reported in humans, among which the functions of 16 genes have been determined and 14 genes are related with diseases present in humans (cystic fibrosis, adrenoleukodystrophy, Stargardt's disease, drug-resistant tumors, Dubin-Johnson syndrome, Byler's disease, progressive familiar intrahepatic cholestasis, X-linked sideroblastic anemia, ataxia, and persistent and hyperinsulimenic hypoglycemia in children) [4,5].
Figure 1 Schematic structural organization of P-glycoprotein. Each half contains a highly hydrophobic domain with 6 transmembrane α-helices involved in chemotherapeutic drug efflux, and a hydrophilic domain located at the cytoplasmic face of the membrane, nucleotide binding domain 1(NBD1) or NMD 2, containing an ATP-binding site with cheracteristic Walker motifs A and B and the S signature of ABC transporters. The two half molecules are separated by a highly charged "linker region which is phosphorylated at several sites by protein kinase C and the first extracellular loop is heavily N-glycosylated [3].
Other efflux pumps of the mammalian cell membrane in ABC superfamily include multidrug resistance-associated proteins (MRP) [6] and breast cancer resistance proteins (BCRP; mitoxantrone resistance proteins, MXR) [7,8]. Other than the fact that these resistant proteins belong to the ABC superfamily, they are quite different with respect to gene locus, amino acid sequence, structure and substrate (Table 1 and 2). In this review, the physiological functions and structures of ABC transporters, and development of chemosensitizers are described focusing on well-known proteins including Pgp, MRP, and BCRP.
Table 1 Gene locus and tissue distribution of ABC transporters
Name Alternate name Gene locus Tissue distribution
MDR1 ABCB1, P-GP 7q36 [9] Gut (apical membrane), liver (canalicular membrane), kindey (apical membrane of epithelial cells of proximal tubule), blood brain barrier (luminal membrane of endothelial cells), testis (endothelial cells of capillary), placenta (trophoblast)
MRP1 ABCC1 16p13.1 [6] Many tissues (brain etc)
MRP2 ABCC2, cMOAT 10q24 [10] Liver, gut, kidney, placenta
MRP3 ABCC3 17q21.3 [11] Liver, gut, adrenal cortex, placenta
MRP4 ABCC4 13q32 [11] Many tissues
MRP5 ABCC5 3q27 [11] Many tissues(brain etc)
MRP6 ABCC6 16p13.1 [12] Liver, kidney
MRP7 ABCC10 6p12-21 [13] Many tissues
MRP8 ABCC11 16q12.1 [14] Breast, testes
BCRP ABCG2, MXR1, ABCP 4q22 [15] Placenta (syncytiotrophoblasts), intestine (epithelium), liver (canalicular membrane), breast (ducts and lobules), endometrium (vein and capillary but not artery), gut
Table 2 Endogenous and exogenous substrates for ABC transporters
Name Endogenous substrate Exogenous cytotoxic substance
MDR1 Estrogen glucuronide conjugates (estradiol, estriol), endorphin, glutamate, steroids (cortisol, aldosterone, corticosterone), beta-amyloid, 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine (generically platelet-activating factor, PAF) Anthracyclines (doxorubucin, daunorubicin, epirubicin), actinomycin D, colchicine, podophyllotoxin (etoposide, teniposide), methotrexate (only in carrier-deficient cells), mitomycin C, mitoxantrone, taxenes (paclitaxel, docetaxel), vinca alkaloids (vincristine, vinblastine)
MRP1 Estradiol-17beta(beta-D-glucuronide) glutathione, glutathione S-conjugate leukoetriene C4, glucuronosyl bilirubin Anthracyclines, cochicine, etoposide, heavy metals (arsenite, arsenate, antimonials), vincristine, vinblastine, paclitaxel
MRP2 Estradiol-17beta(beta-D-glucuronide), glutathione, glutathione S-conjugate Leukoetriene C4, glucuronosyl bilirubin, Cisplatin, CPT-11, doxorubicin, etoposide, methotrexate, SN-38, vincristine, vinblastine
MRP3 S-(2,4-dinitrophenyl)glutathione Cisplatin, doxorubicin, etoposide, methotrexate, teniopside, vincristine,
MRP4 Glucuronide and glutathione conjugates Methotrexate, nucleotide analogs, PMEA*
MRP5 Glutamate and phosphate conjugates Doxorubicin, methotrexate, nucleotide analogs, topotecan,
MRP6 Cyclic nucleotides (cAMP, cGMP), glutathione conjugate Doxorubicin, etoposide, teniposide
MRP7 ? ?
MRP8 17beta-estradiol-(17-beta-D-glucuronide), leukotriene C4, cyclic nucleotides 5'-Fluorouracil, 5'-fluoro-2'-deoxyuridine, 5'-fluoro-5'-deoxyuridine, PMEA*
BCRP Heme or porphyrin Anthracyclines, bisantrene, camptothecin, epirubicin, flavopiridol, mitoxantrone, S-38, topotecan
* PMEA, 2',3'-dideoxycytidine 9'-(2'-hosphonylmethoxynyl)adenine
Functions of ABC transporters
Although the physiologic functions of ABC transporters are not well known, they are expressed constitutively in not only tumor cells but also normal cells in the digestive system including the small intestine, large intestine, liver, and pancreas; epithelial cells in the kidneys, adrenals, brain, and testes; and endothelial cells (Table 1). From the aspect of the tissue distribution, ABC transporters are thought to participate in the absorption and secretion of endogenous and exogenous substances. Endogenous and exogenous substrates for ABC transporters reported so far are summarized in Table 2. Especially, the ABC transporters have shown to function as an efflux pump for lipid, multiple drugs, natural products and peptides. It is proposed to operate as a hydrophobic vacuum cleaner, expelling non-polar compounds from the membrane bilayer to the exterior, driven by the energy of ATP hydrolysis [143]. ATP-dependent transbilayer lipid transporters are classified into cytofacially-directed flippases and exofacially-directed floppases. Floppase activity has been associated with the ABC transporters although not all ABC transporters are floppases [144]. Endogenous substrates for Pgp include corticosterone [145], beta-estradiol 17beta-D-glucuronide, an endogenous cholestatic metabolite of estradiol [146], 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine (generically platelet-activating factor, PAF) [147], glutamate [148] and endorphin [149]. It was also recently reported that Pgp has the function of removing beta-amyloid, which was reported as the causal substance of Alzheimer's disease [150,151]. MRP1 effluxes various conjugated substrates such as leukotriene C4 conjugates [152], steroid conjugates [153] and the GSH conjugate of aflatoxin B1, which is a mycotoxin [154]. Cells can, upon hypoxic demand, use BCRP to reduce heme or porphyrin accumulation, which can be detrimental to cells [155]. When cancer originates not only from cells normally expressing efflux pump but also cells having genes but not expressing, gene expression is initiated due to the exposure to anticancer drugs, resulting in resistance to anticancer drugs, eventually interfering with chemotherapy.
Pgp is mainly present in the apical membrane of intestinal mucosal membrane and lowers bioavailability of drugs by preventing the absorption of the drugs. Digoxin, which shows a low bioavailability and is mainly excreted through stool in normal mice due to poor absorption in the mouse intestine, shows a high bioavailability and mainly excreted through urine in mice with mdr1 knocked out[156]. The bioavailability of the substrate of Pgp, paclitaxel, also increased significantly in mice with mdr1 knocked out and in mice administered with the Pgp inhibitor, PSC-833 [157].
Recently, multiple MDR1 polymorphisms including more than 20 single nucleotide polymorphism (SNP) have been identified. The mutations at positions 2677(G→T) and 2995(G →A) of MDR1 in normal cells were firstly reported [158]. MDR1 polymorphism could be not only associated with alteration of Pgp expression and/or function, drug disposition and treatment outcome but also increase the risk of diseases such as Parkinson's disease and ulcerative colitis [159]. The influence of MDR1 SNP(C3435T and G2677T) on disposition of Pgp substrates or treatment outcome has been examplified in digoxin, phenytoin, fexofenadine, nelfinarvir, cyclosporine, talinolol and loperamide [159]. Polymorphisms of other ABC transporters have been reported [160-163].
If a substance in food affects Pgp, this substance also could affect the bioavailability of substrate drugs for Pgp. It was reported that substances present in grape juice or orange juice could increase the bioavailability of a drug being the substrate of Pgp by inhibiting it. [164]. These substances could also affect pharmacokinetics of other drugs [160,165]. On the contrary, some drugs could increase the expression of Pgp. St John's Wort used as an antidepressant increases the expression of Pgp, so it could significantly lower the serum concentration of indinavir or cyclosporin [166]. Digoxin is the substrate of Pgp and induces paclitaxel resistance by increasing Pgp [166]. Not only Pgp but also MRP and BCRP could affect the bioavailability of drugs.
One of the important physiological functions of efflux pump present in the cell membrane is to provide a pharmacological sanctuary for tissues present in the blood-tissue barriers such as in the case of blood-brain barrier (BBB), blood-placental barrier and blood-testes barrier. Hydrophilic substances present in blood could not go into tissues when they are not small enough to pass through the tight junction with simple diffusion. Nonetheless, various hydrophobic substances could not enter these tissues because they are effluxed out by efflux pumps. Actually, Pgp effluxes neurotransmitters or neuromodulators such as glutamate [148] and opioids [149,167] into blood from the brain. Compared with wild-type mice, drugs beings the substrate of Pgp were significantly increased in the brain of fetus when the mdr1 gene is knocked out in mice [168-171]. When the BCRP inhibitor, GF120918, was introduced to pregnant mice, the topotecan level was increased by two-folds in mouse fetus, suggesting that BCRP would function as the maternal-fetal barrier in the placenta [172]. Thus, quantitative and qualitative changes of transporters present in the membrane could affect pharmacokinetics such as the distribution of endogenous and exogenous substances.
Structure of ABC transporters
Pgp is a 170-kDa membrane protein glycosylated at the first extracellular loop (Fig. 1). Pgp is composed of 12 hydrophobic transmembrane domains (TMDs) and 2 nucleotide-binding domain (NBD). One NBD connects two TMDs with a hydrophilic NBD loop. TDMs form channels for substrate drugs, determine the characteristics of substrate, and efflux substrate drugs whereas NBDs are located in the interior of cytoplasm, and participate in ATP binding and hydrolysis [173]. Pgp undergoes conformational changes upon binding of nucleotide to the NBDs [174]. Rosenberg et al. have analyzed the three-dimensional structures of Pgp and its conformational change in the presence and absence of nucleotide [175-177]. The projection of the protein perpendicular to the membrane is roughly rectangular with a maximum depth of 8 nm, a pore size of 2.5 nm and two 3-nm lobes exposed at the cytoplasmic face of the membrane. The conformational change revealed a major reorganization of the TMDs throughout the entire depth of the membrane upon binding of nucleotide (Fig. 2A). In the absence of nucleotide, the two TMDs form a single barrel 5–6 nm in diameter and about 5 nm deep with a central pore that is open to the extracellular surface and spans much of the membrane depth. Upon binding nucleotide, the TMDs reorganize into three compact domains that are each 2–3 nm in diameter and 5–6 nm deep (Fig. 2B). This reorganization opens the central pore along its length in a manner that could allow access of hydrophobic drugs (transport substrates) directly from the lipid bilayer to the central pore of the transporter [176].
Figure 2 Comparison of nucleotide free-Pgp (nf-Pgp) and Pgp-AMP-PNP (Pgp-AMP-PNP) three-dimensional structures. A, stereo pair of the nf-Pgp three-dimensional structure, displayed using netting at 1.0 σ (red) and 1.5 σ (yellow) above the mean density level and viewed perpendicular to the crystal plane from the more heavily stained side (corresponding to the extracellular surface). B, equivalent views of the Pgp-AMP-PNP structure. The arrow indicates the gap along one side of the central pore. The locations of the three discrete densities A, B, and C are indicated. C, stereo pair of a side view of Pgp-AMP-PNP with the same color scheme as above. The directions of the principle crystallographic axes a and b are shown. Scale bar = 2.2 nm. AMP-PNP, non-hydralizable ATP analogue [176].
When one of two NBDs of Pgp is inactivated, not only drug transport but also ATP hydrolysis of normal NBD is inhibited. This result indicates that two NBDs would function cooperatively and they could not hydrolyze ATP independently [178]. It was recently reported that structural changes of NBDs are brought about when a drug binds to TMD so that the distance between NBDs is changed to affect the activity of ATPase as shown in Fig. 3[179]. Unlike in Pgp, however, the substrate leukotriene C4 could not be transported once NBD2 is inactivated but the substrate transport could not be inhibited when NBD1 is inactivated in MRP1 [180]. This result suggests that among ABC transporters, interactions of NBDs are not simple but function differently for every transport. Although the exact site and number of Pgp binding with drugs have not yet been determined, the important binding sites such as TMD 4, 5, 6, 10, 11 and 12 have been determined [181] whereas substrate drugs do not bind to NBDs [182].
Figure 3 Model of the NBD conformational change by the drug binding to TDM. [178].
Development of chemosensitizers to overcome resistance
One of the causes for the failure of chemotherapy in the treatment of cancer is the emergence of MDR. Once MDR appears, chemotherapy is not effective even when using high doses of drugs enough to overcome resistance, toxic effects are brought about and the resistance mechanism could be further stimulated. These problems could be resolved by the use of the anticancer drugs that could bypass the resistance mechanism. For example, we could use other anticancer drugs such as alkylating drugs (cyclophosphamide), antimetabolites (5-fluorouracil), and the anthracycline modified drugs (annamycin and doxorubicin-peptide) that would not function as the substrates of ABC transporters [183-185]. The final method of overcoming resistance is to administer substances inhibiting ABC transporters with anticancer drugs at the same time. These substances would reverse resistance against anticancer drugs to eventually being sensitized for anticancer drugs so they are called chemosensitizers. They are also called MDR modulators and MDR reverters. Chemosensitizers against each transporter are summarized according to the publishing years (Table 3). Of these, some are chemosensitizers against one transporter and some others against more than two transporters.
Many drugs such as the calcium channel blocker verapamil and the immunosuppressant cyclosporin A would inhibit resistance by functioning as competitive substrates of Pgp regardless of their innate pharmacological functions. Different clinical studies also showed that these drugs could reverse resistance to anticancer drugs. Verapamil is the first reported chemosensitizer inhibiting MDR [119] and its effect was also proven in the recent clinical study [186]. However, verapamil brings about cardiac toxicity at the concentration inhibiting resistance; thus, in order to resolve this problem, the attempts were made to develop (R)-verapmil [187] and verapamil analogues having lower cardiac toxicity compared with (S)-verapamil [188,189]. The immunosuppressant cyclosporin A was first reported to reverse resistance by acute leukemia against vincristine and daunorubicin [190]. Following cyclosporin A, researchers found that other immunosuppressants including FK506 and rapamycin could inhibit MDR [191]. However, when cyclosporin A is applied clinically, researchers placed efforts to develop cyclosporin analogues having few side effects due to their innate immuno suppressant effects and hepatic and renal toxicity with excellent chemosensitizing effects. As a result, PSC-833 (Valspodar), which is the non-immune suppressant analogue of cyclosporin, was developed [110]. In addition to non-immunosuppressant effect, its chemosensitivity is about 10 times higher than that by cyclosporin in Pgp-mediated MDR, so clinical studies are being performed on this drug [192]. Among those drugs having their innate pharmacological activities such as verapamil and cyclosporin A, those having chemosensitizing effect is called the first-generation chemosensitizers. The problems related with the first-generation chemosensitizers are that they generally show low effects and high toxicity at resistance-inhibiting doses. In order to supplement these problems, the chemosensitizers developed only for chemosensitizing effects are called the second-generation chemosensitizers, which include PSC-833, VX-710, LY335979, XR9051 and XR9576 [193]. Multi-national companies are pursuing the development of second-generation chemosensitizers by overcoming the problems of existing chemosensitizers (low effects, side effects, and drug-drug interaction), and some of these chemosensitizers are in the process of being tested clinically.
Most chemosensitizers bind with TMD in transporter, but steroid and flavonoid are new recently introduced chemosensitizers, which inhibit transporters by binding with NBD. The binding site of steroid is different from the binding site of ATP but is probably in the vicinity of the ATP binding site [194]. On the other hand, the flavonoid, kaempferide, is bifunctional in that it would partially block the binding of the antiprogestin RU-486 in the cytoplasm domain of Pgp and block ATP binding [195] (Fig. 4). Recently, flavonoid chemosensitizers reversing Pgp-mediated MDR have been screening. It is believed that flavonoid chemosensitizers have a significant advantage with respect with a therapeutic index (Table 4). These may be second-generation flavonoid chemosensitizers [33].
Figure 4 Proposed schematic model of NBDs showing the relative positions of different nucleotide- and effector-binding sites. MANT-ATP binding is prevented by preincubation with antiprogestin RU-486 and bound MANT-ATP is displaced by Ru-486, suggesting the existence of a cytosolic steroidal-interacting region adjacent to the ATP-binding site. Since the flavonoid binding is prevented by preincubation with ATP and RU-486, bound flavonoids most likely cover both ATP site and the vicinal steroid site. MANT, 2'(3')-N-methylanthraniloyl [3].
Table 3 Chemosensitizers inhibiting Pgp, MRP and BCRP
Name Year Chemosensitizer
Pgp 2004 Benzyl-, phenethyl-, and alpha-naphthyl isothiocyanates [16], diallyl sulfide [17], PK11195 [18], small scFv recombinant Pgp antibody fragment [19]
2003 Amooranin [20], etrandrine, fangchinoline [21], ginsenoside Rg(3) [22], KR30031 [23], methylenedioxyethylamphetamine [24], protopanaxatriol ginsenosides [25], saquinavir [26], siRNA of mdr1 gene [27, 28], tRA 98006* [29]
2002 3,5-dibenzoyl-1,4-dihydropyridines[30], PKC412 [31], pyronaridine [32], sinensetin [33]
2001 Agosterol A [34], haloperidol and dihydrohaloperidol [35], SB203580 [36], tropane alkaloid esters [37], SNF4435C and D [38], tea polyphenol [39], trans-N,N'-bis(3,4-dimethoxybenzyl)-N-solanesyl-1,2-diaminocyclo hexane (N-5228) [40]
2000 Astemizole [41], atorvastatin [42], 7-O-benzoylpyripyropene A [43], 5-O-benzoylated taxinine k [44], clarithromycin and YM17K (3,4'-dideoxy mycaminosyl tylonolide hydrochloride) [45], cyclopamine and tomatidine[46], 3,5-diacetyl-1,4-dihydropyridines [47], 7, 8-dihydroxy-3-benzazepine [48], doxorubicin-gallium-transferrin conjugate [49], macrolide antibiotics (josamycin, tamolarizine) [50], nelfinavir [51] norverapamil [52], ontogen (ONT-093, formerly OC-144-093) [53], R101933 [54], taxuspine C, 2'-desacetoxyaustrospicatine and 2-desacetoxytaxinine [55], V-104 [56]
1999 D-alpha-tocopheryl polyethylene glycol 1000 succinate [57], anti-MDR1 ribozymes [58], AR-2 [59], carvedilol [60], erythromycin [61], ketoconazole [62], kopsiflorine [63], nomegestrol [64], PAK-200S [65], pluronic block copolymer [66], reversin [67], ritonarvir [68], rosemary extract [69], TTD [70], XR9576(2) [71]
1998 Ardeemins [72], AV200 [73], 5-O-benzoylated taxuspine C [74], bromocriptine [75], dipyridamole [76], droloxifene [77], imidazothiazole derivatives (N276-12, N276-14, N276-17) [78], oxalyl bis(N-phenyl)hydroxamic acid [79], tetrandine and fangchinoline [21], tiamulin [80], XR9051 [81]
1997 Biricodar (VX-710; Incel) [82, 83], cyproheptadine [84]
1996 CL 329,753 [85], indole-3-carbinol [86], itraconazole [87], LY335979 [88], medroxyprogesterone [89], mefloquine [90], mifepristone (RU-486) [91], reserpine [92]
1995 Azelastine and flezelastine [93], B9209-005 [94], dexniguldipine (B8509-035) [95], dexverapamil [96], epidermal growth factor (EGF), insulin-like growth factor I (IGF-I) [97], quercetin [98]
1994 MS-209 [99], pentoxifylline [100], Ro11-2933 (DMDP) [101], RU486 [102]
1993 Dilantin [103], GF120918[104], meperidine, pentazocine, and methadone [105], Pgp monoclonal antibodies and antisense oligonucleotide [106], tamoxifen and toremifene [107]
1992 Staurosporine and NA-382 [108]
1991 Biperidil [109], SDZ PSC 833[110]
1990 Cremophor EL [111]
1989 Cefoperazone, cetriaxone [112], phenothiazine [113], YM534 [114]
1987 Diltiazem[115], cyclosporine A [116]
1986 Aamiodarone [117]
1984 Quinidine [118]
1981 Verapamil [119],
MRP 2004 benzyl-, phenethyl-, and alpha-naphthyl isothiocyanates [16]
2003 tRA 98006 [29]
2001 Agosterol A [34]
2000 5-O-benzoylated taxinine k [44], 4-deacetoxyagosterol A [120], doxorubicin-gallium-transferrin conjugate [49], V-104 [56], pluronic block copolymer [66], quinoline-based drugs (chloroquine, quinine, quinidine, and primaquine) [121],
1999 dipyridamole [122], erythromycin and ofloxacin [123], mifepristone (RU-486) [124], MS-209 [125], rifampicin [126]
1998 Biricodar (VX-710; Incel) [83], imidazothiazole derivatives (N276-12, N276-14, N276-17) [78], NSAIDs (indomethacin, sulindac, tolmetin, acemetacin, zomepirac and mefenamic acid) [127], ONO-1078 [128], quercetin [98]
1997 Indomethacin [129], probenecid [130]
1996 Acrolein and chloroacetaldehyde [131], d,l-buthionine-(S,R)-sulfoximine [132], itraconazol [87], PAK-104P [133]
1995 Difloxacin [134], MK571 [135]
BCRP 2004 Chrysin and biochanin A [136], genistein and naligenin [137], Imatinib mesylate (Gleevec, STI571) [138]
2003 Estrone, diethylstilbestrol and TAG-139 [139], tRA 98006 [29]
2002 Ko143 [140]
1999 GF120918 [141]
1998 fumitremorgin C [142]
* Boldface compounds indicate chemosensitizers inhibiting more than two transporters
Table 4 Comparison of chemosensitizing effects of flavonoids and verapamil against Pgp
Chemosensitizer IC50a (μM) CIc
(VCRb-) (VCR+)
5,7,3',4',5' – pentamethoxyflavone > 400 0.4 >1000
7,3',4' – trimethoxyflavone > 400 1.2 >333.3
3',4' – dimethoxyflavone 386 1.2 321.7
3,6,3',4' – tetramethoxyflavone > 400 1.9 >210.5
Verapamil 61 0.4 152.5
5,6,7,3',4' – pentamethoxyflavone > 400 3.2 >125
Cytotoxic and chemosensitizing effects of chemosensitizers in the presence or absence of vincristine in Pgp-overexpressing AML-2/D100 cells.
a, Drug concentrations with inhibit 50% growth of the cells.
b, Vincristine (100 ng/ml)
c, Chemosensitizing index = IC50 (VCR-)/IC50(VCR+)
The fungal toxin, fumitremorgin C (FTC), is a strong inhibitor of BCRP but its use in vivo has been limited due to its neurotoxicity [196]. It was recently reported that the tetracyclic analogue of FTC, Ko143, is the most strong chemosensitizer aginst BCRP having little toxicity [140].
Since ABC transporters can be coexpressed in some types of cancer cells, the development of chemosensitizers against MRP and/or BCRP as well as Pgp has been highly demanding. These include VX-710 against Pgp and MRP [82,83], GF120918 against Pgp and BCRP [104,141] and tRA98006 against all three transporters [29].
Conclusion
One of the major causes of failure in anticancer chemotherapy is resistance against anticancer drugs. Overexpression of ABC transporters such as Pgp, MRP and BCRP has been shown to be responsible for the major portion of MDR. Therefore elucidation of the structure and the function for each ABC transporter is prerequisite for understanding how these transporters work and for reversing MDR. One strategy for reversal of MDR cells expressing ABC transporters is combined use of anticancer drugs with chemosensitizers. Second-generation chemosensitizers have been developed for the purpose of obtaining higher efficacy and lower toxicity than first-generation chemosensitizers. Inhibitors of ABC transporters can be exploited to enhance the oral bioavailablilty or the brain penetration of various drugs.
Combination of a conventional anticancer chemotherapy with new strategies such as chemosensitizers, receptor-mediated targeting and nanotechnology will shed light on cancer patients in the near future.
Acknowledgements
This study was supported by grants from Ministry of Science and Technology, Korea, and from Korea Science and Engineering Foundation through Research Center for Resistant Cells (R13-2003-009). I gratefully thank Dr. Bum-Chae Choi of the CL hospital located in Gwangju (Korea) for critical reading of the manuscript.
==== Refs
Stavrovskaya AA Cellular mechanisms of multidrug resistance of tumor cells Biochemistry (Mosc) 2000 65 95 106 10702644
Riordan JR Ling V Purification of P-glycoprotein from plasma membrane vesicles of Chinese hamster ovary cell mutants with reduced colchicine permeability J Biol Chem 1979 254 12701 12705 500733
Di Pietro A Dayan G Conseil G Steinfels E Krell T Trompier D Baubichon-Cortay H Jault J P-glycoprotein-mediated resistance to chemotherapy in cancer cells: using recombinant cytosolic domains to establish structure-function relationships Braz J Med Biol Res 1999 32 925 939 10454753 10.1590/S0100-879X1999000800001
Dean M Hamon Y Chimini G The human ATP-binding cassette (ABC) transporter superfamily J Lipid Res 2001 42 1007 1017 11441126
Efferth T The human ATP-binding cassette transporter genes: from the bench to the bedside Curr Mol Med 2001 1 45 65 11899242 10.2174/1566524013364194
Cole SP Bhardwaj G Gerlach JH Mackie JE Grant CE Almquist KC Stewart AJ Kurz EU Duncan AM Deeley RG Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line Science 1992 258 1650 1654 1360704
Doyle LA Yang W Abruzzo LV Krogmann T Gao Y Rishi AK Ross DD A multidrug resistance transporter from human MCF-7 breast cancer cells Proc Natl Acad Sci U S A 1998 95 15665 15670 9861027 10.1073/pnas.95.26.15665
Maliepaard M van Gastelen MA de Jong LA Pluim D van Waardenburg RC Ruevekamp-Helmers MC Floot BG Schellens JH Overexpression of the BCRP/MXR/ABCP gene in a topotecan-selected ovarian tumor cell line Cancer Res 1999 59 4559 4563 10493507
Bell DR Trent JM Willard HF Riordan JR Ling V Chromosomal location of human P-glycoprotein gene sequences Cancer Genet Cytogenet 1987 25 141 148 2879621 10.1016/0165-4608(87)90169-5
van Kuijck MA Kool M Merkx GF Geurts van Kessel A Bindels RJ Deen PM van Os CH Assignment of the canalicular multispecific organic anion transporter gene (CMOAT) to human chromosome 10q24 and mouse chromosome 19D2 by fluorescent in situ hybridization Cytogenet Cell Genet 1997 77 285 287 9284939
Kool M de Haas M Scheffer GL Scheper RJ van Eijk MJ Juijn JA Baas F Borst P Analysis of expression of cMOAT (MRP2), MRP3, MRP4, and MRP5, homologues of the multidrug resistance-associated protein gene (MRP1), in human cancer cell lines Cancer Res 1997 57 3537 3547 9270026
Kool M van der Linden M de Haas M Baas F Borst P Expression of human MRP6, a homologue of the multidrug resistance protein gene MRP1, in tissues and cancer cells Cancer Res 1999 59 175 182 9892204
Hopper E Belinsky MG Zeng H Tosolini A Testa JR Kruh GD Analysis of the structure and expression pattern of MRP7 (ABCC10), a new member of the MRP subfamily Cancer Lett 2001 162 181 191 11146224 10.1016/S0304-3835(00)00646-7
Bera TK Lee S Salvatore G Lee B Pastan I MRP8, a new member of ABC transporter superfamily, identified by EST database mining and gene prediction program, is highly expressed in breast cancer Mol Med 2001 7 509 516 11591886
Knutsen T Rao VK Ried T Mickley L Schneider E Miyake K Ghadimi BM Padilla-Nash H Pack S Greenberger L Cowan K Dean M Fojo T Bates S Amplification of 4q21-q22 and the MXR gene in independently derived mitoxantrone-resistant cell lines Genes Chromosomes Cancer 2000 27 110 116 10564593 10.1002/(SICI)1098-2264(200001)27:1<110::AID-GCC14>3.0.CO;2-4
Hu K Morris ME Effects of benzyl-, phenethyl-, and alpha-naphthyl isothiocyanates on P-glycoprotein- and MRP1-mediated transport J Pharm Sci 2004 93 1901 1911 15176077 10.1002/jps.20101
Arora A Seth K Shukla Y Reversal of P-glycoprotein-mediated multidrug resistance by diallyl sulfide in K562 leukemic cells and in mouse liver Carcinogenesis 2004 25 941 949 14729595 10.1093/carcin/bgh060
Walter RB Raden BW Cronk MR Bernstein ID Appelbaum FR Banker DE The peripheral benzodiazepine receptor ligand PK11195 overcomes different resistance mechanisms to sensitize AML cells to gemtuzumab ozogamicin Blood 2004 103 4276 4284 14962898 10.1182/blood-2003-11-3825
Haus-Cohen M Assaraf YG Binyamin L Benhar I Reiter Y Disruption of P-glycoprotein anticancer drug efflux activity by a small recombinant single-chain Fv antibody fragment targeted to an extracellular epitope Int J Cancer 2004 109 750 758 14999785 10.1002/ijc.20037
Ramachandran C Rabi T Fonseca HB Melnick SJ Escalon EA Novel plant triterpenoid drug amooranin overcomes multidrug resistance in human leukemia and colon carcinoma cell lines Int J Cancer 2003 105 784 789 12767063 10.1002/ijc.11180
Choi SU Park SH Kim KH Choi EJ Kim S Park WK Zhang YH Kim HS Jung NP Lee CO The bisbenzylisoquinoline alkaloids, tetrandine and fangchinoline, enhance the cytotoxicity of multidrug resistance-related drugs via modulation of P-glycoprotein Anticancer Drugs 1998 9 255 261 9625436
Kim SW Kwon HY Chi DW Shim JH Park JD Lee YH Pyo S Rhee DK Reversal of P-glycoprotein-mediated multidrug resistance by ginsenoside Rg(3) Biochem Pharmacol 2003 65 75 82 12473381 10.1016/S0006-2952(02)01446-6
Lee BH Lee CO Kwon MJ Yi KY Yoo SE Choi SU Differential effects of the optical isomers of KR30031 on cardiotoxicity and on multidrug resistance reversal activity Anticancer Drugs 2003 14 175 181 12569305 10.1097/00001813-200302000-00012
Ketabi-Kiyanvash N Weiss J Haefeli WE Mikus G P-glycoprotein modulation by the designer drugs methylenedioxymethamphetamine, methylenedioxyethylamphetamine and paramethoxyamphetamine Addict Biol 2003 8 413 418 14690877 10.1080/13556210310001646475
Choi CH Kang G Min YD Reversal of P-glycoprotein-mediated multidrug resistance by protopanaxatriol ginsenosides from Korean red ginseng Planta Med 2003 69 235 240 12677527 10.1055/s-2003-38483
Dupuis ML Tombesi M Sabatini M Cianfriglia M Differential effect of HIV-1 protease inhibitors on P-glycoprotein function in multidrug-resistant variants of the human CD4+ T lymphoblastoid CEM cell line Chemotherapy 2003 49 8 16 12714803 10.1159/000069782
Nieth C Priebsch A Stege A Lage H Modulation of the classical multidrug resistance (MDR) phenotype by RNA interference (RNAi) FEBS Lett 2003 545 144 150 12804765 10.1016/S0014-5793(03)00523-4
Wu H Hait WN Yang JM Small interfering RNA-induced suppression of MDR1 (P-glycoprotein) restores sensitivity to multidrug-resistant cancer cells Cancer Res 2003 63 1515 1519 12670898
Brooks T Minderman H O'Loughlin KL Pera P Ojima I Baer MR Bernacki RJ Taxane-based reversal agents modulate drug resistance mediated by P-glycoprotein, multidrug resistance protein, and breast cancer resistance protein Mol Cancer Ther 2003 2 1195 1205 14617793
Kawase M Shah A Gaveriya H Motohashi N Sakagami H Varga A Molnar J 3,5-dibenzoyl-1,4-dihydropyridines: synthesis and MDR reversal in tumor cells Bioorg Med Chem 2002 10 1051 1055 11836114 10.1016/S0968-0896(01)00363-7
Ganeshaguru K Wickremasinghe RG Jones DT Gordon M Hart SM Virchis AE Prentice HG Hoffbrand AV Man A Champain K Csermak K Mehta AB Actions of the selective protein kinase C inhibitor PKC412 on B-chronic lymphocytic leukemia cells in vitro Haematologica 2002 87 167 176 11836167
Qi J Yang CZ Wang CY Wang SB Yang M Wang JH Function and mechanism of pyronaridine: a new inhibitor of P-glycoprotein-mediated multidrug resistance Acta Pharmacol Sin 2002 23 544 550 12060530
Choi CH Sun KH An CS Yoo JC Hahm KS Lee IH Sohng JK Kim YC Reversal of P-glycoprotein-mediated multidrug resistance by 5,6,7,3',4'-pentamethoxyflavone (Sinensetin) Biochem Biophys Res Commun 2002 295 832 840 12127970 10.1016/S0006-291X(02)00755-6
Murakami N Sugimoto M Morita M Kobayashi M Total synthesis of agosterol A: an MDR-modulator from a marine sponge Chemistry 2001 7 2663 2670 11465457 10.1002/1521-3765(20010618)7:12<2663::AID-CHEM26630>3.0.CO;2-U
Kataoka Y Ishikawa M Miura M Takeshita M Fujita R Furusawa S Takayanagi M Takayanagi Y Sasaki K Reversal of vinblastine resistance in human leukemic cells by haloperidol and dihydrohaloperidol Biol Pharm Bull 2001 24 612 617 11411546 10.1248/bpb.24.612
Barancik M Bohacova V Kvackajova J Hudecova S Krizanova O Breier A SB203580, a specific inhibitor of p38-MAPK pathway, is a new reversal agent of P-glycoprotein-mediated multidrug resistance Eur J Pharm Sci 2001 14 29 36 11457647 10.1016/S0928-0987(01)00139-7
Silva GL Cui B Chavez D You M Chai HB Rasoanaivo P Lynn SM O'Neill MJ Lewis JA Besterman JM Monks A Farnsworth NR Cordell GA Pezzuto JM Kinghorn AD Modulation of the multidrug-resistance phenotype by new tropane alkaloid aromatic esters from Erythroxylum pervillei J Nat Prod 2001 64 1514 1520 11754602 10.1021/np010295+
Kurosawa K Takahashi K Tsuda E Tomida A Tsuruo T Reversal of multidrug resistance by novel nitrophenyl pyrones, SNF4435C and D Jpn J Cancer Res 2001 92 1235 1241 11714449
Zhu A Wang X Guo Z Study of tea polyphenol as a reversal agent for carcinoma cell lines' multidrug resistance (study of TP as a MDR reversal agent) Nucl Med Biol 2001 28 735 740 11518657 10.1016/S0969-8051(00)90202-6
Hayashi M Koike K Rho MC Kuwano M Kishiye T Komiyama K Reversal of P-glycoprotein associated multidrug resistance by new isoprenoid derivatives Anticancer Drug Des 2001 16 255 260 12049484
Ishikawa M Fujita R Takayanagi M Takayanagi Y Sasaki K Reversal of acquired resistance to doxorubicin in K562 human leukemia cells by astemizole Biol Pharm Bull 2000 23 112 115 10706423
Boyd RA Stern RH Stewart BH Wu X Reyner EL Zegarac EA Randinitis EJ Whitfield L Atorvastatin coadministration may increase digoxin concentrations by inhibition of intestinal P-glycoprotein-mediated secretion J Clin Pharmacol 2000 40 91 98 10631627 10.1177/00912700022008612
Rho MC Hayashi M Fukami A Obata R Sunazuka T Tomoda H Komiyama K Omura S Reversal of multidrug resistance by 7-O-benzoylpyripyropene A in multidrug-resistant tumor cells J Antibiot (Tokyo) 2000 53 1201 1206 11132967
Okumura H Chen ZS Sakou M Sumizawa T Furukawa T Komatsu M Ikeda R Suzuki H Hirota K Aikou T Akiyama SI Reversal of P-glycoprotein and multidrug-resistance protein-mediated drug resistance in KB cells by 5-O-benzoylated taxinine K Mol Pharmacol 2000 58 1563 1569 11093797
Wang L Kitaichi K Hui CS Takagi K Sakai M Yokogawa K Miyamoto KI Hasegawa T Reversal of anticancer drug resistance by macrolide antibiotics in vitro and in vivo Clin Exp Pharmacol Physiol 2000 27 587 593 10901387 10.1046/j.1440-1681.2000.03308.x
Lavie Y Harel-Orbital T Gaffield W Liscovitch M Inhibitory effect of steroidal alkaloids on drug transport and multidrug resistance in human cancer cells Anticancer Res 2001 21 1189 1194 11396162
Shah A Gaveriya H Motohashi N Kawase M Saito S Sakagami H Satoh K Tada Y Solymosi A Walfard K Molnar J 3,5-diacetyl-1,4-dihydropyridines: synthesis and MDR reversal in tumor cells Anticancer Res 2000 20 373 377 10769682
Kawase M Saito S Motohashi N Chemistry and biological activity of new 3-benzazepines Int J Antimicrob Agents 2000 14 193 201 10773487 10.1016/S0924-8579(99)00155-7
Wang F Jiang X Yang DC Elliott RL Head JF Doxorubicin-gallium-transferrin conjugate overcomes multidrug resistance: evidence for drug accumulation in the nucleus of drug resistant MCF-7/ADR cells Anticancer Res 2000 20 799 808 10810357
Miyake N Fujita R Ishikawa M Takayanagi M Takayanagi Y Sasaki K Reversal of multidrug resistance in human leukemia K562 by tamolarizine, a novel calcium antagonist Jpn J Pharmacol 2000 82 265 268 10887958 10.1254/jjp.82.265
Shiraki N Hamada A Yasuda K Fujii J Arimori K Nakano M Inhibitory effect of human immunodeficiency virus protease inhibitors on multidrug resistance transporter P-glycoproteins Biol Pharm Bull 2000 23 1528 1531 11145192
Pauli-Magnus C von Richter O Burk O Ziegler A Mettang T Eichelbaum M Fromm MF Characterization of the major metabolites of verapamil as substrates and inhibitors of P-glycoprotein J Pharmacol Exp Ther 2000 293 376 382 10773005
Newman MJ Rodarte JC Benbatoul KD Romano SJ Zhang C Krane S Moran EJ Uyeda RT Dixon R Guns ES Mayer LD Discovery and characterization of OC144-093, a novel inhibitor of P-glycoprotein-mediated multidrug resistance Cancer Res 2000 60 2964 2972 10850444
van Zuylen L Sparreboom A van der Gaast A van der Burg ME van Beurden V Bol CJ Woestenborghs R Palmer PA Verweij J The orally administered P-glycoprotein inhibitor R101933 does not alter the plasma pharmacokinetics of docetaxel Clin Cancer Res 2000 6 1365 1371 10778964
Kobayashi J Shigemori H Hosoyama H Chen Z Akiyama S Naito M Tsuruo T Multidrug resistance reversal activity of taxoids from Taxus cuspidata in KB-C2 and 2780AD cells Jpn J Cancer Res 2000 91 638 642 10874217
Evers R Kool M Smith AJ van Deemter L de Haas M Borst P Inhibitory effect of the reversal agents V-104, GF120918 and Pluronic L61 on MDR1 Pgp-, MRP1- and MRP2-mediated transport Br J Cancer 2000 83 366 374 10917553 10.1054/bjoc.2000.1260
Dintaman JM Silverman JA Inhibition of P-glycoprotein by D-alpha-tocopheryl polyethylene glycol 1000 succinate (TPGS) Pharm Res 1999 16 1550 1556 10554096 10.1023/A:1015000503629
Wang FS Kobayashi H Liang KW Holland JF Ohnuma T Retrovirus-mediated transfer of anti-MDR1 ribozymes fully restores chemosensitivity of P-glycoprotein-expressing human lymphoma cells Hum Gene Ther 1999 10 1185 1195 10340550 10.1089/10430349950018175
Alexanian AR Arutyunian NS Reversal of drug resistance in sarcoma-45 by the new calmodulin antagonist--trihydrochloride of [1,2,5-trimethyl-4-phenyl-4-beta-[N-(beta-ethylamino)-N-4'-methoxybe nzy l]-ethylamino] piperidine (AR-2) Invest New Drugs 1999 17 105 110 10638481 10.1023/A:1006397014409
Jonsson O Behnam-Motlagh P Persson M Henriksson R Grankvist K Increase in doxorubicin cytotoxicity by carvedilol inhibition of P-glycoprotein activity Biochem Pharmacol 1999 58 1801 1806 10571255 10.1016/S0006-2952(99)00262-2
Siedlik PH Olson SC Yang BB Stern RH Erythromycin coadministration increases plasma atorvastatin concentrations J Clin Pharmacol 1999 39 501 504 10234598
Kim RB Wandel C Leake B Cvetkovic M Fromm MF Dempsey PJ Roden MM Belas F Chaudhary AK Roden DM Wood AJ Wilkinson GR Interrelationship between substrates and inhibitors of human CYP3A and P-glycoprotein Pharm Res 1999 16 408 414 10213372 10.1023/A:1018877803319
Rho MC Toyoshima M Hayashi M Koyano T Subramaniam G Kam TS Komiyama K Reversal of multidrug resistance by kopsiflorine isolated from Kopsia dasyrachis Planta Med 1999 65 307 310 10364833 10.1055/s-1999-13991
Li J Xu L He K [Modulation of multiple drug resistance by nomegestrol acetate and droloxifene in K562/A02] Zhonghua Xue Ye Xue Za Zhi 1999 20 288 291 11721384
Vanhoefer U Muller MR Hilger RA Lindtner B Klaassen U Schleucher N Rustum YM Seeber S Harstrick A Reversal of MDR1-associated resistance to topotecan by PAK-200S, a new dihydropyridine analogue, in human cancer cell lines Br J Cancer 1999 81 1304 1310 10604726 10.1038/sj.bjc.6694384
Miller DW Batrakova EV Kabanov AV Inhibition of multidrug resistance-associated protein (MRP) functional activity with pluronic block copolymers Pharm Res 1999 16 396 401 10213370 10.1023/A:1018873702411
Sharom FJ Yu X Lu P Liu R Chu JW Szabo K Muller M Hose CD Monks A Varadi A Seprodi J Sarkadi B Interaction of the P-glycoprotein multidrug transporter (MDR1) with high affinity peptide chemosensitizers in isolated membranes, reconstituted systems, and intact cells Biochem Pharmacol 1999 58 571 586 10413294 10.1016/S0006-2952(99)00139-2
Drewe J Gutmann H Fricker G Torok M Beglinger C Huwyler J HIV protease inhibitor ritonavir: a more potent inhibitor of P-glycoprotein than the cyclosporine analog SDZ PSC 833 Biochem Pharmacol 1999 57 1147 1152 11230802 10.1016/S0006-2952(99)00026-X
Plouzek CA Ciolino HP Clarke R Yeh GC Inhibition of P-glycoprotein activity and reversal of multidrug resistance in vitro by rosemary extract Eur J Cancer 1999 35 1541 1545 10673984 10.1016/S0959-8049(99)00180-X
Xu JY Zhou Q Shen P Tang W Reversal effect of TTD on human multidrug resistant KBV200 cell line J Exp Clin Cancer Res 1999 18 549 552 10746983
Roe M Folkes A Ashworth P Brumwell J Chima L Hunjan S Pretswell I Dangerfield W Ryder H Charlton P Reversal of P-glycoprotein mediated multidrug resistance by novel anthranilamide derivatives Bioorg Med Chem Lett 1999 9 595 600 10098671 10.1016/S0960-894X(99)00030-X
Chou TC Depew KM Zheng YH Safer ML Chan D Helfrich B Zatorska D Zatorski A Bornmann W Danishefsky SJ Reversal of anticancer multidrug resistance by the ardeemins Proc Natl Acad Sci U S A 1998 95 8369 8374 9653193 10.1073/pnas.95.14.8369
Mendez-Vidal C Quesada AR Reversal of P-glycoprotein-mediated multidrug resistance in vitro by AV200, a new ardeemin derivative Cancer Lett 1998 132 45 50 10397452 10.1016/S0304-3835(98)00156-6
Sako M Suzuki H Hirota K Syntheses of taxuspine C derivatives as functional inhibitors of P-glycoprotein, an ATP-associated cell-membrane transporter Chem Pharm Bull (Tokyo) 1998 46 1135 1139 9692220
Orlowski S Valente D Garrigos M Ezan E Bromocriptine modulates P-glycoprotein function Biochem Biophys Res Commun 1998 244 481 488 9514944 10.1006/bbrc.1998.8289
Hosoi E Hirose M Hamano S Morimoto M Kuroda Y Effect of MDR antagonists on the cidal activity of vincristine for cells expressing MDR-1 is superior to those expressing MRP Int J Oncol 1998 13 343 348 9664131
Nussler V Pelka-Fleisc R Gieseler F Hasmann M Loser R Gullis E Stotzer O Zwierzina H Wilmanns W In vitro efficacy of known P-glycoprotein modulators compared to droloxifene E and Z: studies on a human T-cell leukemia cell line and their resistant variants Leuk Lymphoma 1998 31 589 597 9922050
Naito S Koike K Ono M Machida T Tasaka S Kiue A Koga H Kumazawa J Development of novel reversal agents, imidazothiazole derivatives, targeting MDR1- and MRP-mediated multidrug resistance Oncol Res 1998 10 123 132 9700723
Choudhuri SK Chatterjee A Reversal of resistance against doxorubicin by a newly developed compound, oxalyl bis(N-phenyl)hydroxamic acid in vitro Anticancer Drugs 1998 9 825 832 9840730
Baggetto LG Dong M Bernaud J Espinosa L Rigal D Bonvallet R Marthinet E In vitro and in vivo reversal of cancer cell multidrug resistance by the semi-synthetic antibiotic tiamulin Biochem Pharmacol 1998 56 1219 1228 9802334 10.1016/S0006-2952(98)00229-9
Dale IL Tuffley W Callaghan R Holmes JA Martin K Luscombe M Mistry P Ryder H Stewart AJ Charlton P Twentyman PR Bevan P Reversal of P-glycoprotein-mediated multidrug resistance by XR9051, a novel diketopiperazine derivative Br J Cancer 1998 78 885 892 9764579
Germann UA Shlyakhter D Mason VS Zelle RE Duffy JP Galullo V Armistead DM Saunders JO Boger J Harding MW Cellular and biochemical characterization of VX-710 as a chemosensitizer: reversal of P-glycoprotein-mediated multidrug resistance in vitro Anticancer Drugs 1997 8 125 140 9073309
Rowinsky EK Smith L Wang YM Chaturvedi P Villalona M Campbell E Aylesworth C Eckhardt SG Hammond L Kraynak M Drengler R Stephenson J JrHarding MW Von Hoff DD Phase I and pharmacokinetic study of paclitaxel in combination with biricodar, a novel agent that reverses multidrug resistance conferred by overexpression of both MDR1 and MRP J Clin Oncol 1998 16 2964 2976 9738565
Miao ZH Li XT Zhu YJ [Reversal of multidrug resistance by cyproheptadine in KBV200 cells] Yao Xue Xue Bao 1997 32 321 325 11498864
Greenberger LM Collins KI Annable T Boni JP May MK Lai FM Kramer R Citeralla RV Hallett WA Powell D alpha-(3,4-dimethyoxyphenyl)-3,4-dihydro-6,7-dimethoxy-alpha- [(4-methylphenyl)thio]-2(1H)-isoquinolineheptanenitrile (CL 329,753): a novel chemosensitizing agent for P-glycoprotein-mediated resistance with improved biological properties compared with verapamil and cyclosporine A Oncol Res 1996 8 207 218 8884813
Christensen JG LeBlanc GA Reversal of multidrug resistance in vivo by dietary administration of the phytochemical indole-3-carbinol Cancer Res 1996 56 574 581 8564974
Kurosawa M Okabe M Hara N Kawamura K Suzuki S Sakurada K Asaka M Reversal effect of itraconazole on adriamycin and etoposide resistance in human leukemia cells Ann Hematol 1996 72 17 21 8605275
Dantzig AH Shepard RL Cao J Law KL Ehlhardt WJ Baughman TM Bumol TF Starling JJ Reversal of P-glycoprotein-mediated multidrug resistance by a potent cyclopropyldibenzosuberane modulator, LY335979 Cancer Res 1996 56 4171 4179 8797588
Claudio JA Emerman JT The effects of cyclosporin A, tamoxifen, and medroxyprogesterone acetate on the enhancement of adriamycin cytotoxicity in primary cultures of human breast epithelial cells Breast Cancer Res Treat 1996 41 111 122 8944329 10.1007/BF01807156
Riffkin CD Chung R Wall DM Zalcberg JR Cowman AF Foley M Tilley L Modulation of the function of human MDR1 P-glycoprotein by the antimalarial drug mefloquine Biochem Pharmacol 1996 52 1545 1552 8937469 10.1016/S0006-2952(96)00556-4
Fardel O Courtois A Drenou B Lamy T Lecureur V le Prise PY Fauchet R Inhibition of P-glycoprotein activity in human leukemic cells by mifepristone Anticancer Drugs 1996 7 671 677 8913436
Schuetz EG Beck WT Schuetz JD Modulators and substrates of P-glycoprotein and cytochrome P4503A coordinately up-regulate these proteins in human colon carcinoma cells Mol Pharmacol 1996 49 311 318 8632764
Hu YP Robert J Azelastine and flezelastine as reversing agents of multidrug resistance: pharmacological and molecular studies Biochem Pharmacol 1995 50 169 175 7632160 10.1016/0006-2952(95)02055-1
Borchers C Ulrich WR Klemm K Ise W Gekeler V Haas S Schodl A Conrad J Przybylski M Boer R B9209-005, an azido derivative of the chemosensitizer dexniguldipine-HCl, photolabels P-glycoprotein Mol Pharmacol 1995 48 21 29 7623771
Hofmann J Gekeler V Ise W Noller A Mitterdorfer J Hofer S Utz I Gotwald M Boer R Glossmann H Mechanism of action of dexniguldipine-HCl (B8509-035), a new potent modulator of multidrug resistance Biochem Pharmacol 1995 49 603 609 7887974 10.1016/0006-2952(94)00479-6
Thurlimann B Kroger N Greiner J Mross K Schuller J Schernhammer E Schumacher K Gastl G Hartlapp J Kupper H Dexverapamil to overcome epirubicin resistance in advanced breast cancer J Cancer Res Clin Oncol 1995 121 R3 6 8698740
Hirsch-Ernst KI Ziemann C Schmitz-Salue C Foth H Kahl GF Modulation of P-glycoprotein and mdr1b mRNA expression by growth factors in primary rat hepatocyte culture Biochem Biophys Res Commun 1995 215 179 185 7575588 10.1006/bbrc.1995.2450
Kim SH Yeo GS Lim YS Kang CD Kim CM Chung BS Suppression of multidrug resistance via inhibition of heat shock factor by quercetin in MDR cells Exp Mol Med 1998 30 87 92 9873828
Tsuruo T [MDR reversing drugs for clinical development] Gan To Kagaku Ryoho 1994 21 962 967 8002634
Breier A Barancik M Stefankova Z Uhrik B Tribulova N Effect of pentoxifylline on P-glycoprotein mediated vincristine resistance of L1210 mouse leukemic cell line Neoplasma 1994 41 297 303 7854500
Wigler PW Patterson FK Reversal agent inhibition of the multidrug resistance pump in human leukemic lymphoblasts Biochim Biophys Acta 1994 1189 1 6 7905747
Gruol DJ Zee MC Trotter J Bourgeois S Reversal of multidrug resistance by RU 486 Cancer Res 1994 54 3088 3091 7911394
Ganapathi R Hercbergs A Grabowski D Ford J Selective enhancement of vincristine cytotoxicity in multidrug-resistant tumor cells by dilantin (phenytoin) Cancer Res 1993 53 3262 3265 8100737
Hyafil F Vergely C Du Vignaud P Grand-Perret T In vitro and in vivo reversal of multidrug resistance by GF120918, an acridonecarboxamide derivative Cancer Res 1993 53 4595 4602 8402633
Callaghan R Riordan JR Synthetic and natural opiates interact with P-glycoprotein in multidrug-resistant cells J Biol Chem 1993 268 16059 16064 8101846
Efferth T Volm M Modulation of P-glycoprotein-mediated multidrug resistance by monoclonal antibodies, immunotoxins or antisense oligodeoxynucleotides in kidney carcinoma and normal kidney cells Oncology 1993 50 303 308 8098865
Kirk J Houlbrook S Stuart NS Stratford IJ Harris AL Carmichael J Selective reversal of vinblastine resistance in multidrug-resistant cell lines by tamoxifen, toremifene and their metabolites Eur J Cancer 1993 29A 1152 1157 8518026
Miyamoto K Wakusawa S Inoko K Takagi K Koyama M Reversal of vinblastine resistance by a new staurosporine derivative, NA-382, in P388/ADR cells Cancer Lett 1992 64 177 183 1351792 10.1016/0304-3835(92)90079-B
van Kalken CK van der Hoeven JJ de Jong J Giaccone G Schuurhuis GJ Maessen PA Blokhuis WM van der Vijgh WJ Pinedo HM Bepridil in combination with anthracyclines to reverse anthracycline resistance in cancer patients Eur J Cancer 1991 27 739 744 1829915
Boesch D Gaveriaux C Jachez B Pourtier-Manzanedo A Bollinger P Loor F In vivo circumvention of P-glycoprotein-mediated multidrug resistance of tumor cells with SDZ PSC 833 Cancer Res 1991 51 4226 4233 1678313
Friche E Jensen PB Sehested M Demant EJ Nissen NN The solvents cremophor EL and Tween 80 modulate daunorubicin resistance in the multidrug resistant Ehrlich ascites tumor Cancer Commun 1990 2 297 303 1976341
Gosland MP Lum BL Sikic BI Reversal by cefoperazone of resistance to etoposide, doxorubicin, and vinblastine in multidrug resistant human sarcoma cells Cancer Res 1989 49 6901 6905 2582432
Ford JM Prozialeck WC Hait WN Structural features determining activity of phenothiazines and related drugs for inhibition of cell growth and reversal of multidrug resistance Mol Pharmacol 1989 35 105 115 2563302
Sato S Watanabe Y Okamura K Hamanaka R Mori T Kohno K Kuwano M Potentiation of vincristine and actinomycin D by a new synthetic imidazole anti-tumor agent YM534 active against human cancer cells and multidrug-resistant cells Anticancer Drug Des 1989 4 125 135 2803461
Cornwell MM Pastan I Gottesman MM Certain calcium channel blockers bind specifically to multidrug-resistant human KB carcinoma membrane vesicles and inhibit drug binding to P-glycoprotein J Biol Chem 1987 262 2166 2170 2434476
Twentyman PR Fox NE White DJ Cyclosporin A and its analogues as modifiers of adriamycin and vincristine resistance in a multi-drug resistant human lung cancer cell line Br J Cancer 1987 56 55 57 3040060
Chauffert B Martin M Hammann A Michel MF Martin F Amiodarone-induced enhancement of doxorubicin and 4'-deoxydoxorubicin cytotoxicity to rat colon cancer cells in vitro and in vivo Cancer Res 1986 46 825 830 3940645
Tsuruo T Iida H Kitatani Y Yokota K Tsukagoshi S Sakurai Y Effects of quinidine and related compounds on cytotoxicity and cellular accumulation of vincristine and adriamycin in drug-resistant tumor cells Cancer Res 1984 44 4303 4307 6467192
Tsuruo T Iida H Tsukagoshi S Sakurai Y Overcoming of vincristine resistance in P388 leukemia in vivo and in vitro through enhanced cytotoxicity of vincristine and vinblastine by verapamil Cancer Res 1981 41 1967 1972 7214365
Murakami N Sugimoto M Morita M Akiyama S Kobayashi M Synthesis and evaluation of 4-deacetoxyagosterol A as an MDR-modulator Bioorg Med Chem Lett 2000 10 2521 2524 11086720 10.1016/S0960-894X(00)00502-3
Vezmar M Georges E Reversal of MRP-mediated doxorubicin resistance with quinoline-based drugs Biochem Pharmacol 2000 59 1245 1252 10736425 10.1016/S0006-2952(00)00270-7
Curtin NJ Turner DP Dipyridamole-mediated reversal of multidrug resistance in MRP over-expressing human lung carcinoma cells in vitro Eur J Cancer 1999 35 1020 1026 10533488 10.1016/S0959-8049(99)00038-6
Terashi K Oka M Soda H Fukuda M Kawabata S Nakatomi K Shiozawa K Nakamura T Tsukamoto K Noguchi Y Suenaga M Tei C Kohno S Interactions of ofloxacin and erythromycin with the multidrug resistance protein (MRP) in MRP-overexpressing human leukemia cells Antimicrob Agents Chemother 2000 44 1697 1700 10817732 10.1128/AAC.44.6.1697-1700.2000
Payen L Delugin L Courtois A Trinquart Y Guillouzo A Fardel O Reversal of MRP-mediated multidrug resistance in human lung cancer cells by the antiprogestatin drug RU486 Biochem Biophys Res Commun 1999 258 513 518 10329417 10.1006/bbrc.1999.0671
Nakamura T Oka M Aizawa K Soda H Fukuda M Terashi K Ikeda K Mizuta Y Noguchi Y Kimura Y Tsuruo T Kohno S Direct interaction between a quinoline derivative, MS-209, and multidrug resistance protein (MRP) in human gastric cancer cells Biochem Biophys Res Commun 1999 255 618 624 10049760 10.1006/bbrc.1999.0245
Courtois A Payen L Vernhet L de Vries EG Guillouzo A Fardel O Inhibition of multidrug resistance-associated protein (MRP) activity by rifampicin in human multidrug-resistant lung tumor cells Cancer Lett 1999 139 97 104 10408915 10.1016/S0304-3835(99)00024-5
Duffy CP Elliott CJ O'Connor RA Heenan MM Coyle S Cleary IM Kavanagh K Verhaegen S O'Loughlin CM NicAmhlaoibh R Clynes M Enhancement of chemotherapeutic drug toxicity to human tumour cells in vitro by a subset of non-steroidal anti-inflammatory drugs (NSAIDs) Eur J Cancer 1998 34 1250 1259 9849488 10.1016/S0959-8049(98)00045-8
Nagayama S Chen ZS Kitazono M Takebayashi Y Niwa K Yamada K Tani A Haraguchi M Sumizawa T Furukawa T Aikou T Akiyama S Increased sensitivity to vincristine of MDR cells by the leukotriene D4 receptor antagonist, ONO-1078 Cancer Lett 1998 130 175 182 9751271 10.1016/S0304-3835(98)00132-3
Draper MP Martell RL Levy SB Indomethacin-mediated reversal of multidrug resistance and drug efflux in human and murine cell lines overexpressing MRP, but not P-glycoprotein Br J Cancer 1997 75 810 815 9062400
Gollapudi S Kim CH Tran BN Sangha S Gupta S Probenecid reverses multidrug resistance in multidrug resistance-associated protein-overexpressing HL60/AR and H69/AR cells but not in P-glycoprotein-overexpressing HL60/Tax and P388/ADR cells Cancer Chemother Pharmacol 1997 40 150 158 9182837 10.1007/s002800050640
Manzano RG Wright KA Twentyman PR Modulation by acrolein and chloroacetaldehyde of multidrug resistance mediated by the multidrug resistance-associated protein (MRP) Clin Cancer Res 1996 2 1321 1326 9816303
Vanhoefer U Cao S Minderman H Toth K Skenderis BS 2ndSlovak ML Rustum YM d,l-buthionine-(S,R)-sulfoximine potentiates in vivo the therapeutic efficacy of doxorubicin against multidrug resistance protein-expressing tumors Clin Cancer Res 1996 2 1961 1968 9816155
Vanhoefer U Cao S Minderman H Toth K Scheper RJ Slovak ML Rustum YM PAK-104P, a pyridine analogue, reverses paclitaxel and doxorubicin resistance in cell lines and nude mice bearing xenografts that overexpress the multidrug resistance protein Clin Cancer Res 1996 2 369 377 9816180
Gollapudi S Thadepalli F Kim CH Gupta S Difloxacin reverses multidrug resistance in HL-60/AR cells that overexpress the multidrug resistance-related protein (MRP) gene Oncol Res 1995 7 213 225 8534927
Gekeler V Ise W Sanders KH Ulrich WR Beck J The leukotriene LTD4 receptor antagonist MK571 specifically modulates MRP associated multidrug resistance Biochem Biophys Res Commun 1995 208 345 352 7887949 10.1006/bbrc.1995.1344
Zhang S Yang X Morris ME Flavonoids are inhibitors of breast cancer resistance protein (ABCG2)-mediated transport Mol Pharmacol 2004 65 1208 1216 15102949 10.1124/mol.65.5.1208
Imai Y Tsukahara S Asada S Sugimoto Y Phytoestrogens/Flavonoids Reverse Breast Cancer Resistance Protein/ABCG2-Mediated Multidrug Resistance Cancer Res 2004 64 4346 4352 15205350 10.1158/0008-5472.CAN-04-0078
Houghton PJ Germain GS Harwood FC Schuetz JD Stewart CF Buchdunger E Traxler P Imatinib mesylate is a potent inhibitor of the ABCG2 (BCRP) transporter and reverses resistance to topotecan and SN-38 in vitro Cancer Res 2004 64 2333 2337 15059881 10.1158/0008-5472.CAN-03-3344
Sugimoto Y Tsukahara S Imai Y Ueda K Tsuruo T Reversal of breast cancer resistance protein-mediated drug resistance by estrogen antagonists and agonists Mol Cancer Ther 2003 2 105 112 12533678
Allen JD van Loevezijn A Lakhai JM van der Valk M van Tellingen O Reid G Schellens JH Koomen GJ Schinkel AH Potent and specific inhibition of the breast cancer resistance protein multidrug transporter in vitro and in mouse intestine by a novel analogue of fumitremorgin C Mol Cancer Ther 2002 1 417 425 12477054
de Bruin M Miyake K Litman T Robey R Bates SE Reversal of resistance by GF120918 in cell lines expressing the ABC half-transporter, MXR Cancer Lett 1999 146 117 126 10656616 10.1016/S0304-3835(99)00182-2
Rabindran SK He H Singh M Brown E Collins KI Annable T Greenberger LM Reversal of a novel multidrug resistance mechanism in human colon carcinoma cells by fumitremorgin C Cancer Res 1998 58 5850 5858 9865745
Sharom FJ Liu R Qu Q Romsicki Y Exploring the structure and function of the P-glycoprotein multidrug transporter using fluorescence spectroscopic tools Semin Cell Dev Biol 2001 12 257 265 11428918 10.1006/scdb.2000.0251
Daleke DL Regulation of transbilayer plasma membrane phospholipid asymmetry J Lipid Res 2003 44 233 242 12576505 10.1194/jlr.R200019-JLR200
Wolf DC Horwitz SB P-glycoprotein transports corticosterone and is photoaffinity-labeled by the steroid Int J Cancer 1992 52 141 146 1354202
Liu Y Huang L Hoffman T Gosland M Vore M MDR1 substrates/modulators protect against beta-estradiol-17beta-D-glucuronide cholestasis in rat liver Cancer Res 1996 56 4992 4997 8895755
Ernest S Bello-Reuss E Secretion of platelet-activating factor is mediated by MDR1 P-glycoprotein in cultured human mesangial cells J Am Soc Nephrol 1999 10 2306 2313 10541289
Liu XD Liu GQ P glycoprotein regulated transport of glutamate at blood brain barrier Acta Pharmacol Sin 2001 22 111 116 11741514
King M Su W Chang A Zuckerman A Pasternak GW Transport of opioids from the brain to the periphery by P-glycoprotein: peripheral actions of central drugs Nat Neurosci 2001 4 268 274 11224543 10.1038/85115
Vogelgesang S Cascorbi I Schroeder E Pahnke J Kroemer HK Siegmund W Kunert-Keil C Walker LC Warzok RW Deposition of Alzheimer's beta-amyloid is inversely correlated with P-glycoprotein expression in the brains of elderly non-demented humans Pharmacogenetics 2002 12 535 541 12360104 10.1097/00008571-200210000-00005
Lam FC Liu R Lu P Shapiro AB Renoir JM Sharom FJ Reiner PB beta-Amyloid efflux mediated by p-glycoprotein J Neurochem 2001 76 1121 1128 11181832 10.1046/j.1471-4159.2001.00113.x
Leier I Jedlitschky G Buchholz U Cole SP Deeley RG Keppler D The MRP gene encodes an ATP-dependent export pump for leukotriene C4 and structurally related conjugates J Biol Chem 1994 269 27807 27810 7961706
Jedlitschky G Leier I Buchholz U Barnouin K Kurz G Keppler D Transport of glutathione, glucuronate, and sulfate conjugates by the MRP gene-encoded conjugate export pump Cancer Res 1996 56 988 994 8640791
Loe DW Stewart RK Massey TE Deeley RG Cole SP ATP-dependent transport of aflatoxin B1 and its glutathione conjugates by the product of the multidrug resistance protein (MRP) gene Mol Pharmacol 1997 51 1034 1041 9187270
Krishnamurthy P Ross DD Nakanishi T Bailey-Dell K Zhou S Mercer KE Sarkadi B Sorrentino BP Schuetz JD The stem cell marker Bcrp/ABCG2 enhances hypoxic cell survival through interactions with heme J Biol Chem 2004 279 24218 24225 15044468 10.1074/jbc.M313599200
Tsuruoka S Sugimoto KI Fujimura A Imai M Asano Y Muto S P-glycoprotein-mediated drug secretion in mouse proximal tubule perfused in vitro J Am Soc Nephrol 2001 12 177 181 11134265 10.1159/000051255
Malingre MM Beijnen JH Schellens JH Oral delivery of taxanes Invest New Drugs 2001 19 155 162 11392449 10.1023/A:1010635000879
Mickley LA Lee JS Weng Z Zhan Z Alvarez M Wilson W Bates SE Fojo T Genetic polymorphism in MDR-1: a tool for examining allelic expression in normal cells, unselected and drug-selected cell lines, and human tumors Blood 1998 91 1749 1756 9473242
Fromm MF The influence of MDR1 polymorphisms on P-glycoprotein expression and function in humans Adv Drug Deliv Rev 2002 54 1295 1310 12406646 10.1016/S0169-409X(02)00064-9
Brinkmann U Eichelbaum M Polymorphisms in the ABC drug transporter gene MDR1 Pharmacogenomics J 2001 1 59 64 11913728
Ito S Ieiri I Tanabe M Suzuki A Higuchi S Otsubo K Polymorphism of the ABC transporter genes, MDR1, MRP1 and MRP2/cMOAT, in healthy Japanese subjects Pharmacogenetics 2001 11 175 184 11266082 10.1097/00008571-200103000-00008
Kerb R Hoffmeyer S Brinkmann U ABC drug transporters: hereditary polymorphisms and pharmacological impact in MDR1, MRP1 and MRP2 Pharmacogenomics 2001 2 51 64 11258197 10.1517/14622416.2.1.51
Moriya Y Nakamura T Horinouchi M Sakaeda T Tamura T Aoyama N Shirakawa T Gotoh A Fujimoto S Matsuo M Kasuga M Okumura K Effects of polymorphisms of MDR1, MRP1, and MRP2 genes on their mRNA expression levels in duodenal enterocytes of healthy Japanese subjects Biol Pharm Bull 2002 25 1356 1359 12392094 10.1248/bpb.25.1356
Tian R Koyabu N Takanaga H Matsuo H Ohtani H Sawada Y Effects of grapefruit juice and orange juice on the intestinal efflux of P-glycoprotein substrates Pharm Res 2002 19 802 809 12134950 10.1023/A:1016100715125
Kurata Y Ieiri I Kimura M Morita T Irie S Urae A Ohdo S Ohtani H Sawada Y Higuchi S Otsubo K Role of human MDR1 gene polymorphism in bioavailability and interaction of digoxin, a substrate of P-glycoprotein Clin Pharmacol Ther 2002 72 209 219 12189368 10.1067/mcp.2002.126177
Hennessy M Kelleher D Spiers JP Barry M Kavanagh P Back D Mulcahy F Feely J St Johns wort increases expression of P-glycoprotein: implications for drug interactions Br J Clin Pharmacol 2002 53 75 82 11849198 10.1046/j.0306-5251.2001.01516.x
Kastin AJ Fasold MB Zadina JE Endomorphins, Met-enkephalin, Tyr-MIF-1, and the P-glycoprotein efflux system Drug Metab Dispos 2002 30 231 234 11854138 10.1124/dmd.30.3.231
Kawahara M Sakata A Miyashita T Tamai I Tsuji A Physiologically based pharmacokinetics of digoxin in mdr1a knockout mice J Pharm Sci 1999 88 1281 1287 10585223 10.1021/js9901763
Miyama T Takanaga H Matsuo H Yamano K Yamamoto K Iga T Naito M Tsuruo T Ishizuka H Kawahara Y Sawada Y P-glycoprotein-mediated transport of itraconazole across the blood-brain barrier Antimicrob Agents Chemother 1998 42 1738 1744 9661014
Mayer U Wagenaar E Dorobek B Beijnen JH Borst P Schinkel AH Full blockade of intestinal P-glycoprotein and extensive inhibition of blood-brain barrier P-glycoprotein by oral treatment of mice with PSC833 J Clin Invest 1997 100 2430 2436 9366556
Schinkel AH Smit JJ van Tellingen O Beijnen JH Wagenaar E van Deemter L Mol CA van der Valk MA Robanus-Maandag EC te Riele HP Disruption of the mouse mdr1a P-glycoprotein gene leads to a deficiency in the blood-brain barrier and to increased sensitivity to drugs Cell 1994 77 491 502 7910522 10.1016/0092-8674(94)90212-7
Jonker JW Smit JW Brinkhuis RF Maliepaard M Beijnen JH Schellens JH Schinkel AH Role of breast cancer resistance protein in the bioavailability and fetal penetration of topotecan J Natl Cancer Inst 2000 92 1651 1656 11036110 10.1093/jnci/92.20.1651
Hyde SC Emsley P Hartshorn MJ Mimmack MM Gileadi U Pearce SR Gallagher MP Gill DR Hubbard RE Higgins CF Structural model of ATP-binding proteins associated with cystic fibrosis, multidrug resistance and bacterial transport Nature 1990 346 362 365 1973824 10.1038/346362a0
Sonveaux N Shapiro AB Goormaghtigh E Ling V Ruysschaert JM Secondary and tertiary structure changes of reconstituted P-glycoprotein. A Fourier transform attenuated total reflection infrared spectroscopy analysis J Biol Chem 1996 271 24617 24624 8798727 10.1074/jbc.271.40.24617
Rosenberg MF Velarde G Ford RC Martin C Berridge G Kerr ID Callaghan R Schmidlin A Wooding C Linton KJ Higgins CF Repacking of the transmembrane domains of P-glycoprotein during the transport ATPase cycle Embo J 2001 20 5615 5625 11598005 10.1093/emboj/20.20.5615
Rosenberg MF Kamis AB Callaghan R Higgins CF Ford RC Three-dimensional structures of the mammalian multidrug resistance P-glycoprotein demonstrate major conformational changes in the transmembrane domains upon nucleotide binding J Biol Chem 2003 278 8294 8299 12501241 10.1074/jbc.M211758200
Rosenberg MF Callaghan R Ford RC Higgins CF Structure of the multidrug resistance P-glycoprotein to 2.5 nm resolution determined by electron microscopy and image analysis J Biol Chem 1997 272 10685 10694 9099718 10.1074/jbc.272.3.1665
Senior AE Bhagat S P-glycoprotein shows strong catalytic cooperativity between the two nucleotide sites Biochemistry 1998 37 831 836 9454572 10.1021/bi9719962
Loo TW Bartlett MC Clarke DM Drug binding in human P-glycoprotein causes conformational changes in both nucleotide-binding domains J Biol Chem 2003 278 1575 1578 12421806 10.1074/jbc.M211307200
Gao M Cui HR Loe DW Grant CE Almquist KC Cole SP Deeley RG Comparison of the functional characteristics of the nucleotide binding domains of multidrug resistance protein 1 J Biol Chem 2000 275 13098 13108 10777615 10.1074/jbc.275.17.13098
Loo TW Clarke DM Identification of residues within the drug-binding domain of the human multidrug resistance P-glycoprotein by cysteine-scanning mutagenesis and reaction with dibromobimane J Biol Chem 2000 275 39272 39278 11013259 10.1074/jbc.M007741200
Loo TW Clarke DM The transmembrane domains of the human multidrug resistance P-glycoprotein are sufficient to mediate drug binding and trafficking to the cell surface J Biol Chem 1999 274 24759 24765 10455147 10.1074/jbc.274.35.24759
Consoli U Priebe W Ling YH Mahadevia R Griffin M Zhao S Perez-Soler R Andreeff M The novel anthracycline annamycin is not affected by P-glycoprotein-related multidrug resistance: comparison with idarubicin and doxorubicin in HL-60 leukemia cell lines Blood 1996 88 633 644 8695811
Perez-Soler R Neamati N Zou Y Schneider E Doyle LA Andreeff M Priebe W Ling YH Annamycin circumvents resistance mediated by the multidrug resistance-associated protein (MRP) in breast MCF-7 and small-cell lung UMCC-1 cancer cell lines selected for resistance to etoposide Int J Cancer 1997 71 35 41 9096663 10.1002/(SICI)1097-0215(19970328)71:1<35::AID-IJC8>3.0.CO;2-4
Mazel M Clair P Rousselle C Vidal P Scherrmann JM Mathieu D Temsamani J Doxorubicin-peptide conjugates overcome multidrug resistance Anticancer Drugs 2001 12 107 116 11261883 10.1097/00001813-200102000-00003
Belpomme D Gauthier S Pujade-Lauraine E Facchini T Goudier MJ Krakowski I Netter-Pinon G Frenay M Gousset C Marie FN Benmiloud M Sturtz F Verapamil increases the survival of patients with anthracycline-resistant metastatic breast carcinoma Ann Oncol 2000 11 1471 1476 11142488 10.1023/A:1026556119020
Haussermann K Benz B Gekeler V Schumacher K Eichelbaum M Effects of verapamil enantiomers and major metabolites on the cytotoxicity of vincristine and daunomycin in human lymphoma cell lines Eur J Clin Pharmacol 1991 40 53 59 2060546 10.1007/BF00315139
Teodori E Dei S Scapecchi S Gualtieri F The medicinal chemistry of multidrug resistance (MDR) reversing drugs Farmaco 2002 57 385 415 12058813 10.1016/S0014-827X(02)01229-6
Toffoli G Simone F Corona G Raschack M Cappelletto B Gigante M Boiocchi M Structure-activity relationship of verapamil analogs and reversal of multidrug resistance Biochem Pharmacol 1995 50 1245 1255 7488241 10.1016/0006-2952(95)02003-U
Slater LM Sweet P Stupecky M Gupta S Cyclosporin A reverses vincristine and daunorubicin resistance in acute lymphatic leukemia in vitro J Clin Invest 1986 77 1405 1408 3457021
Arceci RJ Stieglitz K Bierer BE Immunosuppressants FK506 and rapamycin function as reversal agents of the multidrug resistance phenotype Blood 1992 80 1528 1536 1381629
Fracasso PM Brady MF Moore DH Walker JL Rose PG Letvak L Grogan TM McGuire WP Phase II study of paclitaxel and valspodar (PSC 833) in refractory ovarian carcinoma: a gynecologic oncology group study J Clin Oncol 2001 19 2975 2982 11408492
Mistry P Stewart AJ Dangerfield W Okiji S Liddle C Bootle D Plumb JA Templeton D Charlton P In vitro and in vivo reversal of P-glycoprotein-mediated multidrug resistance by a novel potent modulator, XR9576 Cancer Res 2001 61 749 758 11212278
Dayan G Jault JM Baubichon-Cortay H Baggetto LG Renoir JM Baulieu EE Gros P Di Pietro A Binding of steroid modulators to recombinant cytosolic domain from mouse P-glycoprotein in close proximity to the ATP site Biochemistry 1997 36 15208 15215 9398248 10.1021/bi9718696
Conseil G Baubichon-Cortay H Dayan G Jault JM Barron D Di Pietro A Flavonoids: a class of modulators with bifunctional interactions at vicinal ATP- and steroid-binding sites on mouse P-glycoprotein Proc Natl Acad Sci U S A 1998 95 9831 9836 9707561 10.1073/pnas.95.17.9831
Rabindran SK Ross DD Doyle LA Yang W Greenberger LM Fumitremorgin C reverses multidrug resistance in cells transfected with the breast cancer resistance protein Cancer Res 2000 60 47 50 10646850
|
16202168
|
PMC1277830
|
CC BY
|
2021-01-04 16:40:10
|
no
|
Cancer Cell Int. 2005 Oct 4; 5:30
|
utf-8
|
Cancer Cell Int
| 2,005 |
10.1186/1475-2867-5-30
|
oa_comm
|
==== Front
Cardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 1475-2840-4-161620737810.1186/1475-2840-4-16Original InvestigationDistinct effects of glucose and glucosamine on vascular endothelial and smooth muscle cells: Evidence for a protective role for glucosamine in atherosclerosis Duan Wenlan [email protected] Latha [email protected] Sivaram [email protected] Reddy US therapeutics, 3065 Northwoods Circle, Norcross, GA 30071, USA2 Angion Biomedica, 350 Community Dr, Manhasset, NY 110303 Department of Radiation Oncology, North Shore-Long Island Jewish Health System, 350 Community Dr, Manhasset, NY 11030, USA2005 5 10 2005 4 16 16 18 8 2005 5 10 2005 Copyright © 2005 Duan et al; licensee BioMed Central Ltd.2005Duan 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.
Accelerated atherosclerosis is one of the major vascular complications of diabetes. Factors including hyperglycemia and hyperinsulinemia may contribute to accelerated vascular disease. Among the several mechanisms proposed to explain the link between hyperglycemia and vascular dysfunction is the hexosamine pathway, where glucose is converted to glucosamine. Although some animal experiments suggest that glucosamine may mediate insulin resistance, it is not clear whether glucosamine is the mediator of vascular complications associated with hyperglycemia. Several processes may contribute to diabetic atherosclerosis including decreased vascular heparin sulfate proteoglycans (HSPG), increased endothelial permeability and increased smooth muscle cell (SMC) proliferation. In this study, we determined the effects of glucose and glucosamine on endothelial cells and SMCs in vitro and on atherosclerosis in apoE null mice. Incubation of endothelial cells with glucosamine, but not glucose, significantly increased matrix HSPG (perlecan) containing heparin-like sequences. Increased HSPG in endothelial cells was associated with decreased protein transport across endothelial cell monolayers and decreased monocyte binding to subendothelial matrix. Glucose increased SMC proliferation, whereas glucosamine significantly inhibited SMC growth. The antiproliferative effect of glucosamine was mediated via induction of perlecan HSPG. We tested if glucosamine affects atherosclerosis development in apoE-null mice. Glucosamine significantly reduced the atherosclerotic lesion in aortic root. (P < 0.05) These data suggest that macrovascular disease associated with hyperglycemia is unlikely due to glucosamine. In fact, glucosamine by increasing HSPG showed atheroprotective effects.
==== Body
Introduction
Generalized vascular dysfunction including microvascular (nephropathy) and macrovascular (accelerated atherosclerosis) is a characteristic of diabetic complications [1-3]. Hyperglycemia by several mechanisms may contribute to increased atherosclerosis. Glucose can increase intracellular oxidative stress and generation of reactive oxygen species in endothelial cells [3]. This can result in activation of Redox sensitive transcription factors such as nuclear factor kappa B and inflammatory genes. Glucose can form adducts with proteins by non-enzymatic mechanisms leading to generation of glycated proteins and advanced glycation end products (AGE) [4,5]. In different animal models blocking AGE and its receptor RAGE reduced vascular disease [6]. Another pathway that has been postulated to play an important role in insulin resistance and potentially vascular complications is the hexosamine pathway [7,8]. In this pathway glucose is converted to glucosamine by the enzymatic actions of glutamine:fructose-6-phosphate amidotransferase (GFAT). In vitro in certain cell types glucosamine was shown to increase expression of growth factor TGFβ and PAI-1 [9,10]. Glucosamine is also the precursor for proteoglycan biosynthesis and increases proteoglycan production including heparan sulfate proteoglycan (HSPG) production in different cell types including vascular cells.
In vessel wall HSPG are produced by all cell types either as components of cell membrane (syndecan and glypican) or extracellular matrix (perlecan) [11]. Perlecan is the major HSPG of endothelial cells and SMC [12,13]. In atherosclerotic lesions the content of HSPG is reduced and studies show an inverse correlation between the amount of cholesterol in the lesion and the concentration of HS [14,15]. Thus, unlike chondroitin sulfate proteoglycans, HSPG is negatively correlated with atherosclerosis.
In this study we show that glucosamine and glucose have distinct effects on vascular HSPG and cell growth and glucosamine by virtue of its beneficial effects in vascular cells also reduces atherosclerosis in mice.
Methods
Materials
D (+) Glucosamine and D(+) glucose were purchased from Sigma Chemical Co. (St. Louis, MO). L- (4,5 3H) Leucine, 3H-Thymidine, (35S) as sulfate in aqueous solutions and (125I) were from Amersham Life Science Corp (Arlington Heights, IL). Heparinase, heparitinase and chondroitinase ABC were purchased from Seikagaku America Inc (Bethesda, MD).
Lipoproteins
LDL (d < 1.063), Lp(a) (d = 1.11) and HDL (d = 1.1–1.23) were isolated from fresh plasma by sequential ultracentrifugation. For some experiments 125I-LDL radio-iodinated using iodine monochloride [5] was used.
Endothelial cells
Bovine aortic endothelial cells were isolated and cultured as described [20]. The cells (5–20 passages) were grown in minimal essential medium containing 10% FBS (Life-Technologies, Gaithersburg, MD).
125I-LDL transport
For 125I-LDL transport experiments, endothelial cells were grown in tissue culture inserts (Falcon-0.3 μm pore size) to facilitate its access to the upper (luminal) and lower (subendothelial) surface of endothelial cells. The barrier function of the endothelial cell monolayer was examined as previously reported [16], using [3H]dextran (average mol wt 150,000) and [14C]albumin. Transport of these molecules from the apical to the basolateral side of the monolayers was < 5%/h, a rate similar to that reported by others. At the conclusion of each LDL transport experiment, the monolayers were stained with 2% toluidene blue to verify the uniformity of the monolayer. On the day of the experiment, 125I-labeled LDL was added to the upper chamber and following incubation radioactivity in the lower-chamber medium and associated with the extracellular matrix was determined. The total amount of each lipoprotein transported across the monolayers (net transport) is the sum of these two measurements. To determine 125I-labeled LDL associated with the subendothelial cell matrix, the cells were incubated for 10 min with medium containing 50 U/ml of heparin (Elkins-Sinn, Inc., Cherry Hill, NJ)
Metabolic labeling and preparation of subendothelial matrix (SEM)
Endothelial PG was radiolabeled with (35S) sulfate along with the indicated doses of glucosamine for 12–16 h. Cellular PG were assessed by removing the cells with Triton X-100/NH4OH. Endothelial cells were grown in 24 or 48 well plates (Falcon: Beckton Dickinson, Lincoln Park, NJ). Subendothelial matrix (SEM) was prepared as described [17]. Briefly, confluent monolayers of endothelial cells were washed three times with PBS and incubated for 10 min in a solution containing 0.1% TritonX-100 and 20 mM NH4OH at room temperature. Detached cells were removed by washing four times with PBS. This procedure leaves the intact matrix attached to the surface of the well. Matrix PG was extracted with 6 M guanidine HCl for 4 h at 4°C. For enzyme treatments, SEM was incubated with a mixture of heparinase and heparitinase (1 U/ml each) or chondroitinase ABC (1–2 U/ml) for 3 h at 37°C.
Glycosaminoglycan (GAG) size estimation
To prepare GAG chains, 1-ml aliquots of purified cell surface PG were treated with 100 μl of 10 N NaOH for 18 h at 26°C with constant shaking and then neutralized with 10 N HCl [18]. To remove core proteins, the samples were adjusted to 7 M urea and loaded on 1-ml DEAE mini column that was previously equilibrated with 3 bed volumes of 7 M urea, 0.1% Triton X-100, 0.2 M NaCl in 0.05 M Tris, pH7.2. Peak fractions were pooled and dialyzed against 10 mM Tris and 0.1% Triton X-100 to remove urea. To determine GAG size, 0.6 ml of the above protein free GAG were chromatographed on Sephacryl-200 (Pharmacia-Biotech) gel-filtration column that was previously calibrated with known molecular weight standards using 0.2 M NaCl as an eluate.
Monocytes and adhesion assay
THP-1 cells were purchased from the American Type Culture Collection (Rockville, MD) and were grown in RPMI 1640 (Gibco-Laboratories, Grand Island, NY) containing 10% fetal bovine serum (Gemini Bioproducts Inc., Calabasas, CA).
Adhesion assay was done as described previously [19]. Monocytes were incubated in leucine-free DMEM-BSA medium before labeling. 100 μCi of (3H) leucine was added to 1 × 107 cells and incubated for another 2 h under cell culture conditions. The labeled cells were centrifuged at 800 rpm for 5 min to remove the unincorporated label. The cells were washed three times and re-suspended in DMEM-BSA and then added to endothelial cell monolayers or to SEM in 24 well plates (24 × 105 cells/well). Binding was performed for 2 h at 37°C. Unbound monocytes were removed by washing four times with DMEM-BSA and the bound radioactivity was extracted with 0.5 N NaOH for 1 h at 37°C and then counted.
SMC proliferation
Rat aortic SMCs were cultured as previously described [20] in basal medium supplemented with growth factors, bFGF, hEGF and 5% fetal bovine serum (Clonetics Corp, San Diego, CA). SMCs were plated in low density (9 × 104 cells/ well) in 6 or 12 well plates and cultured in the presence or absence of 30 mM glucose or 2.5 mM glucosamine for three days. The cells were then trypsinized and an aliquot of trypsinate was counted for the final cell number with hemacytometer. Net growth was assessed by subtracting the final cell number from the initial cell number.
In other experiments, cells grown under above conditions were labeled with (3H)-thymidine (5 μCi/ml) for 6 h and the cells were washed four times with DMEM-BSA to remove unincorporated label. The cells were then lysed by 0.5 N NaOH and the thymidine incorporation into the DNA was assessed.
Animal studies
Sulfate incorporation in tissues of mice
To determine if glucosamine increases HS in vivo (determined by 35SO4 incorporation), C57Bl/6J mice from Jackson Laboratories (Bar Harbor, Maine) (8 weeks old, three controls and three glucosamine) were given saline or saline containing 5 mg/kg of glucosamine intraperitoneally for 3 days. On the day of experiment, mice were given 100 μCi of 35SO4 in 100 μl of saline. Mice were sacrificed after 4 h, tissues were perfused with PBS and liver and heart together with proximal aorta, were removed. Tissues were homogenized with polytron for 30 sec in ice cold HEPES buffer containing 4 M urea, 0.5% CHAPS, 0.5 M NaCl, 1 mM each of PMSF, benzamidine-hydrochloride and 5 μg/ml of leupeptin. Tissue homogenates were centrifuged (14000 rpm, 20 min) and the supernatants were dialyzed extensively against PBS to remove low molecular weight free sulfate. Aliquots of dialyzed supernatants were precipitated with 3 volumes of alcohol and counted and the radioactivity was expressed per mg of tissue protein.
Glucosamine effects on atherosclerosis
Male apoE-/- mice on C57BL/6J background were purchased from Jackson Laboratories (Bar Harbor, Maine). Mice were housed at 25°C on a 12 h light-dark cycle and were fed on a chow diet and water ad libitum throughout the study. At four weeks of age, they were randomly divided into vehicle (n = 7) and glucosamine treated (n = 8) groups. Glucosamine were administered intraperitoneally once a day (5 mg/kg). The study was terminated at 12 weeks of age. Mice were euthanized by CO2 and exanguination. Blood samples were collected for glucose and lipid assay by Colorimetric assays (Sigma Diagnostics). Aortic roots were snap frozen in (optimal cutting temperature) OCT for lesion evaluation by Oil-Red-O staining.
Results
Glucosamine but not glucose increases 35SO4 incorporation into endothelial HSPG
To determine whether glucosamine increases endothelial HS production, aortic endothelial cell proteoglycans were labeled with 35S sulfate. Glucosamine treatment increased 35SO4 incorporation into the media PG by 2 fold and into matrix PG by 3 fold (Figure 1A). Addition of glucose (30 mM) to the medium did not affect PG production. Enzymatic analysis showed that the increase was found to be exclusively in HSPG and glucosamine treatment did not affect CS/DS PG in endothelial cells (Figure 1B). Perlecan is the major HSPG secreted by endothelial cells, we therefore tested if the increase was in perlecan. Real time PCR analysis showed a 1.9 fold increase in perlecan mRNA (Figure 1C bars) and consistent with this immunoprecipitation analysis showed that media from glucosamine treated cells contain 2.3 fold perlecan in medium (Figure 1C line). Thus these data suggest that glucosamine primarily increased endothelial cell perlecan. Highly sulfated blocks of HS are referred to as heparin-like HS, which confer several biological properties to HS. To identify heparin-like HS GAG, matrix HSPG were subjected to heparitinase digestion, followed by low pH nitrous acid digestion [18,21]. Heparitinase-resistant and nitrous acid-sensitive HS was increased by two fold (2300 cpm in control versus 4550 cpm in glucosamine treated) suggesting that glucosamine treatment of endothelial cells increased heparin-like HS. Glucose or glucosamine did not significantly increase macrophage PG but glucosamine showed a moderate increase in SMC HSPG (Figure 2).
Figure 1 A. Glucosamine treatment increases 35SO4 incorporation into PG. Endothelial cells were labeled with 35S sulfate (25 μCi/ml) in medium alone (Control) or medium containing 30 mM glucose or 2.5 mM glucosamine 16 h. Radioactivity associated with proteoglycans (PG) from cells, media and extracellular matrix was determined. Values are expressed as mean ± SD B. Glucosamine mediated increase in PG is mostly in heparan sulfate proteoglycans (HSPG). Media and matrix PG prepared from control and glucosamine treated endothelial cells were treated with chondroitinase ABC for 4 h at 37°C and undigested PG were precipitated and determined as HSPG. To determine if the undigested glycosaminoglycan (GAG) is HS, aliquots were subjected to low pH nitrous acid digestion. 1C. Glucose and glucosamine effects on perlecan mRNA (Real time PCR – 1C bar) and protein (immunoprecipitated protein in media 1C line) p < 0.01.
Figure 2 Glucosamine effects on macrophage (A) and smooth muscle cells (SMC) (B) PG. THP-1 human monocyte-macrophages or confluent monolayers of rat aortic SMC were labeled with 35SO4 in medium or medium containing 2.5 mM glucosamine or medium containing 30 mM glucose (for SMC) for 16 hours. Total PG and CS/DS PG and HSPG were determined as described in Figure 1B. (data not mentioned in the Result)
Glucosamine treatment improved endothelial barrier function
Subendothelial matrix HSPG is thought to play a key role in endothelial barrier function, however, whether decreased HSPG increases lipoprotein transport across endothelial monolayers is not known. We first tested whether removal of HSPG increases LDL transport across endothelium. Removal of HSPG by heparinase treatment led to a 2.1 fold increase in 125I-LDL transport at 10 min. (Figure 3A). At 15 min and 30 min the increase in LDL transport was 57% and 36% higher than controls. Conversely, after glucosamine treatment, 125I-LDL transport across the EC monolayers decreased by 15–22% at different time points in glucosamine treated cells (Figure 3B).
Figure 3 A. HSPG modulate LDL transport across EC monolayers. Endothelial cells were grown to confluence in tissue culture inserts (Falcon, 0.3 μm pore size) in 24 well plates to facilitate its access to the upper (luminal) and lower (subendothelial) surface of endothelial cells. The cells were incubated with medium alone (control) or medium containing 1 unit/ml each of heparinase and heparitinase in the bottom chamber for 2 h at 37°C. 125I-LDL was then added to the cells in the upper chamber and the 125I-LDL appeared in the media from the lower chamber was counted. Values represent Mean ± SD of triplicate measurements. B. Glucosamine treatment decreases LDL transport. Endothelial cells on tissue culture inserts were incubated with medium alone or medium containing 2.5 mM glucosamine for 16 h. 125I-LDL transport was then determined as described above. Figure 3C. Monocyte adhesion to glucosamine treated endothelial cells decreases. Endothelial cells were grown to confluence in 24 well tissue culture plates. Cells were then incubated in medium or medium with glucosamine for 16 h. Subendothelial matrix was prepared from control and glucosamine treated endothelial cells and incubated with (3H)leucine labeled THP-1 monocytes for 2 h. Unbound monocytes were washed four times with DMEM-BSA and the bound radioactivity was determined.
Monocyte binding to matrix is decreased in glucosamine treated endothelial cells
We previously showed that removal of HSPG increases monocyte retention in the subendothelial matrix [19]. We determined THP-1 monocyte to the subendothelial matrix prepared from control, glucose and glucosamine-treated cells. Monocyte binding to the glucosamine stimulated endothelial cells was decreased by 52% compared to non-stimulated cells (Fig. 3C). In contrast monocyte binding to glucose-treated endothelial cells was slightly increased similar to previous observations (20%, not shown ref [22]).
Glucose and Glucosamine effects on PG in SMC- Glucosamine but not glucose treatment decreases SMC proliferation
We next determined the effects of glucose and glucosamine on SMC proliferation. Sub-confluent SMC were incubated in media containing 30 mM glucose or 2.5 mM glucosamine for 48 h and net growth was determined. Glucose treatment did not alter SMC growth as assessed by cell number (Figure 4A) or thymidine incorporation (Figure 4B). In contrast, the cell number and 3H thymidine incorporation into the DNA was decreased by 60% by glucosamine treatment (4A and B). The major HSPG secreted by SMC is perlecan, which was known to negatively correlate with SMC growth [11]. Immunoprecipitation analysis showed a 2.5 fold increase in perlecan protein in SMC treated with glucosamine (not shown). We next tested whether perlecan mediates the antiproliferative effect of glucosamine on SMC. Addition of an anti-perlecan antibody completely abolished the antiproliferative effect of glucosamine on SMC proliferation (Figure 5). These results suggest that glucosamine treatment increases perlecan production, which in turn modulates SMC growth in proliferating cells. Consistent with a perlecan-based mechanism, glucosamine did not inhibit endothelial cell growth (Figure 4C)
Figure 4 Glucosamine treatment decreases the growth of SMC (A and B) but not endothelial cell (C) proliferation. Sub-confluent SMC (9 × 104/well) were incubated in growth medium or growth medium containing 2.5 mM glucosamine or 30 mM glucose for 24–48 h and cell growth was determined. (A) Initial and final cell (SMC) number was counted and net growth was determined. (B) SMC were labeled with 3H-Thymidine and its incorporation into the DNA was assessed. (C) Endothelial cell growth was determined by thymidine incorporation into DNA.
Figure 5 The antiproliferative effect of glucosamine requires perlecan: SMC proliferation was carried out as described in Figure 4 with or without anti-perlecan antibody (10 μg/ml) in the presence or absence of glucosamine.
Glucosamine increases HS and reduces atherosclerotic lesion of aortic root in vivo
We tested glucosamine effects on HSPG and atherosclerosis. When given intraperitoneally (5 mg/kg once a day for 3 days) glucosamine increased 35S sulfate incorporation into HSPG in liver (by 61%) and heart (by 82%) (Figure 6). To determine the effects of glucosamine on atherosclerosis, apoE null mice were treated with glucosamine at 5 mg/kg for 2 months. Plasma glucose and total cholesterol was not affected by glucosamine (not shown). Oil Red-O staining revealed 30% reduction in lesions in glucosamine treated group (p < 0.05) (Figure 7). These data suggest that macrovascular disease associated with hyperglycemia is unlikely due to glucosamine. In fact, glucosamine by inducing HSPG showed atheroprotective effects.
Figure 6 Glucosamine increases 35SO4 incorporation in vivo. To determine if glucosamine can increase tissue PG, mice were injected with glucosamine (intraperitoneal, 100 ul containing 5 mg, every other day for 3 days). On the day of the experiment 35SO4 was injected intravenously and mice were sacrificed after 4 h. Following perfusion with saline, liver, and heart were removed and homogenized in phosphate buffered saline containing 1 mM PMSF, 1 mM benzamidine and 0.5% CHAPS. PG were precipitated with 3 volumes of 100% alcohol. 35S-cpm in the precipitate was determined.
Figure 7 Glucosamine reduces atherosclerotic lesion of aortic root in apoE null mice. apoE null mice were treated with glucosamine intraperitoneally at 5 mg/kg for 2 months from 4 weeks to 12 weeks of age. At the end of the study, mice were euthanized. Aortic roots were collected and snap frozen in OCT for lesion evaluation by Oil-Red-O staining.
Discussion
Glucosamine is a precursor of GAG biosynthesis. In cells glucosamine is produced from glucose by the hexosamine pathway in a reaction requiring fructose 6-phosphate and glutamine and catalyzed by the enzyme glutamine:fructose-6 phosphate amidotransferase [7,8]. About 1–2% of the incoming glucose enters this pathway. Glucosamine is primarily used for protein glycosylation and GAG synthesis [23]. Although cells can synthesize glucosamine, exogenous glucosamine can also be taken up and converted to its 6-phosphate derivative, which can then be utilized for HSPG synthesis. Current studies show that glucose and glucosamine have distinct effects on vascular HSPG and cell proliferation. Unlike glucose, glucosamine increased matrix HSPG (perlecan) production both in endothelial cells and SMC. Glucosamine increased sulfate incorporation specifically into HSPG.
Several previous studies suggested a role for glucosamine in the development of insulin resistance [8,24]. Based on several in vitro studies glucosamine was also suggested to be a player in mediating the vascular complications [7,9]. These observation received great attention because glucosamine is frequently used by patients with osteoarthritis [25]. However, recent studies show that in humans at doses used by arthritis patients glucosamine does not appear to induce insulin resistance [26].
Loss of endothelial HS has been postulated to lead to several pathological events, in particular to events related to atherosclerosis [13,27]. Agents that decrease endothelial HSPG include, lipopolysaccharide, TNF-alpha [28], homocysteine [29], lysolecithin and oxidized LDL [30]. Thus, decrease in HS may be a general inflammatory reaction.
In atherosclerosis, CS/DS PG positively correlated with lesion progression [14,15]. Thus, glucosamine treatment not only increased athero-protective HSPG but also decreased (in SMC, figure 2B) or did not affect (in endothelial cell, figure 1B) atherogenic CS/DS PG. Glucosamine also increased the amount of heparin-like HS (oligosaccharide sequences that are resistant to heparitinase digestion but sensitive to heparinase and low pH nitrous acid) in endothelial cells. Subendothelial HSPG (perlecan) contains substantial amounts of these sequences [31]. In our experiments, glucosamine primarily increased extracellular HSPG (in media and in the matrix).
Glucosamine treatment also inhibited SMC proliferation. The extent of this inhibition was much greater than that of other antiproliferative substances such as apoE, nitric oxide and TGF-β (not shown). The antiproliferative effect of glucosamine is most likely due to increased HSPG production in media. It has been well documented that while cell surface HSPG are required for growth factor activity (as co-receptors) exogenous HSPG are a potent inhibitors of SMC proliferation [32,33]. Perlecan is the major HSPG secreted by SMC and it negatively correlates with SMC proliferation [34]. In the present studies an anti-perlecan antibody completely blocked glucosamine antiproliferative effect. These data suggest that glucosamine inhibits SMC proliferation by increasing media perlecan. These data also show that the SMC growth inhibition by glucosamine is not due to general cell toxicity. Among the vascular cells only SMC, but not endothelial cells and macrophages, are sensitive to HSPG inhibition. Consistent with this glucosamine did not inhibit growth of endothelial cells.
How glucosamine increased perlecan production is not clear. Our data suggest that glucosamine increased HS GAG content but not chain length. Because perlecan contains only three HS chains per core protein, it is conceivable that more perlecan core protein is associated with HS chains facilitating its secretion. Alternatively, glucosamine may have induced perlecan expression. Glucosamine is also utilized for glycosylation of proteins including certain transcription factors. Glycosylation state of the transcription factors affects their activity [35]. Thus, it is conceivable that glucosamine treatment increased the glycosylation of transcription factors involved in perlecan expression. Glucosamine was also shown to induce growth factor expression [36-38], such as TGFβ, which can stimulate perlecan [39]. However, it should be noted that in these studies, in contrast to the present experiments, high concentrations of glucose had the same effect as that of glucosamine. Nevertheless, if this occurs, since perlecan antibody blocked the antiproliferative effect of glucosamine, these data raise the possibility that the antiproliferative effects of TGFβ are mediated by perlecan.
Our data also showed that glucosamine administration to mice increased 35SO4 incorporation in tissues. Glucosamine is taken by cells via the glucose transporters and is generally absent in circulating plasma [40]. A dose of 5 mg/kg of glucosamine for three days increased 35SO4 incorporation into the liver by 82% and into hearts (containing proximal aorta) by 61%. It remains to be determined whether this increase is specifically in HSPG. The in vivo effect of glucosamine was tested in one other study. Because CS and DS PG are thought to mediate lipoprotein retention and therefore, atherogenic. Recent in vitro studies showed that glucosamine induced proteoglycans have reduced binding to LDL therefore less atherogenic [41]. An earlier study looked at the effect of glucosamine on plasma and aortic cholesterol in rabbits [42]. Surprisingly, this study found a two fold decrease in cholesterol/unit-wet weight of aorta. However, our data showing that glucosamine increases only HSPG but not CS/DS may explain why lipoprotein accumulation is decreased. The ability of glucosamine to inhibit atherogenesis has recently been postulated [43] and has been demonstrated in our apoE null mice study. This occurred without changes in plasma glucose and lipids. Taken together with our present data, studying the effects of increased HSPG on atherosclerosis appears to be feasible in mice.
In summary, our data show that glucosamine increases HSPG production in vascular cells and 35SO4 incorporation into tissue. Atherogenic PG, like CS/DS PG were not increased. By increasing HSPG, glucosamine reduced atherogenic events including lipoprotein transport, monocyte retention and SMC proliferation, combined with it protective effect on atherosclerotic lesion in apoE null mice, raising the possibility that it is a potential anti-atherogenic agent.
Abbreviations
HSPG – heparan sulfate proteoglycans, PG-proteoglycan, LDL-low density lipoprotein, SMC – smooth muscle cell, GAG – glycosaminoglycan
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WD performed in vivo studies and LP performed in vitro studies. SP designed, supervised the study and provided support.
Acknowledgements
These studies were in part supported by a grant in aid and Investigatorship from the American Heart Association, New York City Affiliate.
==== Refs
Hsueh WA Law RE Diabetes is a vascular disease J Investig Med 1998 46 387 90 9805424
Sowers JR Stump CS Insights into the biology of diabetic vascular disease: what's new? Am J Hypertens 2004 17 2S 6S 15539107 10.1016/j.amjhyper.2004.08.007
Nishikawa T Edelstein D Du XL Yamagishi S Matsumura T Kaneda Y Yorek MA Beebe D Oates PJ Hammes HP Giardino I Brownlee M Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage Nature 2000 404 787 90 10783895 10.1038/35008121
Ahmed N Advanced glycation endproducts–role in pathology of diabetic complications Diabetes Res Clin Pract 2005 67 3 21 15620429 10.1016/j.diabres.2004.09.004
Jakus V Rietbrock N Advanced glycation end-products and the progress of diabetic vascular complications Physiol Res 2004 53 131 42 15046548
Hudson BI Bucciarelli LG Wendt T Sakaguchi T Lalla E Qu W Lu Y Lee L Stern DM Naka Y Ramasamy R Yan SD Yan SF D'Agati V Schmidt AM Blockade of receptor for advanced glycation endproducts: a new target for therapeutic intervention in diabetic complications and inflammatory disorders Arch Biochem Biophys 2003 1;419 80 8 14568011 10.1016/j.abb.2003.08.030
Schleicher ED Weigert C Role of the hexosamine biosynthetic pathway in diabetic nephropathy Kidney Int Suppl 2000 77 S13 8 10997685 10.1046/j.1523-1755.2000.07703.x
Marshall S Bacote V Traxinger RR Discovery of a metabolic pathway mediating glucose-induced desensitization of the glucose transport system. Role of hexosamine biosynthesis in the induction of insulin resistance J Biol Chem 1991 266 4706 12 2002019
Kolm-Litty V Sauer U Nerlich A Lehmann R Schleicher ED High glucose-induced transforming growth factor beta1 production is mediated by the hexosamine pathway in porcine glomerular mesangial cells J Clin Invest 1998 101 160 9 9421478
Goldberg HJ Scholey J Fantus IG Glucosamine activates the plasminogen activator inhibitor 1 gene promoter through Sp1 DNA binding sites in glomerular mesangial cells Diabetes 2000 49 863 71 10905498
Rosenberg RD Shworak NW Liu J Schwartz JJ Zhang L Heparan sulfate proteoglycans of the cardiovascular system. Specific structures emerge but how is synthesis regulated? J Clin Invest 1997 99 2062 70 9151776
Iozzo RV Cohen IR Grassel S Murdoch AD The biology of perlecan: the multi faceted HSPG of basement membranes and pericellular matrices Biochem J 1994 302 625 639 7945186
Pillarisetti S Lipoprotein modulation of subendothelial heparan sulfate proteoglycans (perlecan) and atherogenicity Trends Cardiovasc Med 2000 10 60 5 11150731 10.1016/S1050-1738(00)00048-7
Hollmann J Schmidt A von Bassewitz DB Buddecke E Relationship of sulfated glycosaminoglycans and cholesterol content in normal and atherosclerotic human aorta Arteiosclerosis 1989 9 154 158
Volker W Schmidt A Oortmann W Broszey T Faber V Buddecke E Mapping of proteoglycans an atherosclerotic lesions Eur Heart J 1990 11 29 40 2226532
Saxena U Klein MG Vanni TM Goldberg IJ Lipoprotein lipase increases low density lipoprotein retention by subendothelial cell matrix J Clin Invest 1992 89 373 80 1737833
Stins MF Maxfield FR Goldberg IJ Polarized binding of lipoprotein lipase to endothelial cells Arterioscler Thromb 1992 12 1437 1446 1450176
Edwards IJ Xu H Obunike JC Goldberg IJ Wagner WD Differentiated macrophages synthesize a heparan sulfate proteoglycan and an oversulfated chondroitin sulfate proteoglycan that bind lipoprotein lipase Arteriosclero Thromb 1995 15 400 409
Pillarisetti S Obunike JC Goldberg IJ Lysolecithin induced alterations of subendothelial heparan sulfate proteoglycans increases monocyte binding to matrix J Biol Chem 1995 270 29760 29765 8530367 10.1074/jbc.270.52.31225
Paka S Goldberg IJ Obunike JC Choi SY Saxena U Goldberg ID Pillarisetti S Perlecan mediates the antiproliferative effect of apolipoprotein E on smooth muscle cells. An underlying mechanism for the modulation of smooth muscle cell growth J Biol Chem 1999 274 36403 36408 10593935 10.1074/jbc.274.51.36403
Paka L Kako Y Obunike JC Pillarisetti S Apo E stimulates endothelial production of heparan sulfate rich in biologically active heparin-like domains J Biol Chem 1999 274 4816 4823 9988721 10.1074/jbc.274.8.4816
Vogl-Willis CA Edwards IJ High glucose-induced alterations in subendothelial matrix perlecan leads to increased monocyte binding Arterioscler Thromb Vasc Biol 2004 24 858 63 15031130 10.1161/01.ATV.0000126375.60073.74
Kjellen L Lindahl U Proteoglycans: structures and interactions Annu Rev Biochem 1991 60 443 75 1883201 10.1146/annurev.bi.60.070191.002303
Hawkins M Barzilai N Liu R Hu M Chen W Rossetti L Role of the glucosamine pathway in fat-induced insulin resistance J Clin Invest 1997 99 2173 82 9151789
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 27;357 251 6 11214126 10.1016/S0140-6736(00)03610-2
Pouwels MJ Jacobs JR Span PN Lutterman JA Smits P Tack CJ Short-term glucosamine infusion does not affect insulin sensitivity in humans J Clin Endocrinol Metab 2001 86 2099 103 11344213 10.1210/jc.86.5.2099
Engelberg H Endogenous heparin activity deficiency: the 'missing link' in atherogenesis? Atherosclerosis 2001 159 253 60 11730804 10.1016/S0021-9150(01)00650-5
Ramasamy S Lipke DW McClain CJ Hennig B Tumor necrosis factor reduces proteoglycan synthesis in cultured endothelial cells J Cell Physiol 1995 162 119 126 7814444 10.1002/jcp.1041620114
Nishinaga M Ozawa T Shimada K Homocysteine, a thrombogenic agent suppresses anticoagulant heparan sulfate expression in cultured porcine aortic endothelial cells J Clin Invest 1993 92 1381 1386 8376590
Pillarisetti S Paka L Obunike JC Berglund L Goldberg IJ Subendothelial retention of lipoprotein (a). Evidence that reduced heparan sulfate promotes lipoprotein (a) retention by subendothelial matrix J Clin Invest 1997 100 867 874 9259586
Nader HB Dietrich CP Buonassisi V Colburn P Heparin sequences in the heparan sulfate chains of an endothelial cell proteoglycan Proc Natl Acad Sci USA 1987 84 3565 3569 2954157
Ruoslahti E Yamaguchi Y Proteoglycans as modulators of growth factor activities Cell 1991 64 867 869 2001586 10.1016/0092-8674(91)90308-L
Castellot JJ Addonizio ML Rosenberg R Karnovsky MJ Cultured endothelial cells produce heparinlike inhibitor of smooth muscle cell growth J Cell Biol 1981 90 372 379 7287812 10.1083/jcb.90.2.372
Weiser MC Belknap JK Grieshaber SS Kinsella MG Majack RA Developmental regulation of perlecan gene expression in aortic smooth muscle cells Matrix Biol 1996 15 331 340 8981329 10.1016/S0945-053X(96)90136-5
Han I Kudlow JE Reduced O glycosylation of Sp1 is associated with increased proteasome susceptibility Mol Cell Biol 1997 17 2550 2556 9111324
McClain DA Paterson AJ Roos MD Wei X Kudlow JE Glucose and glucosamine regulate growth factor gene expression in vascular smooth muscle cells Proc Natl Acad Sci USA 1992 89 8150 8154 1518840
Sayeski PP Kudlow JE Glucose metabolism to glucosamine is necessary for glucose stimulation of transforming growth factor-alpha gene transcription J Biol Chem 1998 271 15237 15243 8663078
Kolm-Litty V Sauer U Nerlich A Lehmann R Schieicher ED High glucose induced transforming growth factor beta1 production is mediated by the hexosamine pathway in porcine glomerular mesangial cells J Clin Invest 1998 101 160 169 9421478
Iozzo RV Pillarisetti J Sharma B Murdoch AD Danielson KG Uitto J Mauviel A Structural and functional characterization of human perlecan gene promoter. Transcriptional activation of transforming growth factor-beta via a nuclear factor 1-binding element J Biol Chem 1997 272 5219 5228 9030592 10.1074/jbc.272.8.5219
Tesoriere G Dones F Magistro D Castagnetta L Intestinal absorption of glucosamine and N-acetylglucosamine Experentia 1972 28 770 771 10.1007/BF01923118
Tannock LR Little PJ Wight TN Chait A Arterial smooth muscle cell proteoglycans synthesized in the presence of glucosamine demonstrate reduced binding to LDL J Lipid Res 2002 43 149 57 11792734
Stender S Astrup P Glucosamine and experimental atherosclerosis: Increased wet weight and changed composition of cholesterol fatty acids in aorta of rabbits fed a cholesterol enriched diet with added glucosamine Atherosclerosis 1977 26 205 213 836356 10.1016/0021-9150(77)90103-4
McCarty MF Glucosamine may retard atherogenesis by promoting endothelial production of heparan sulfate proteoglycans Med Hypotheses 1997 48 245 251 9140889 10.1016/S0306-9877(97)90328-5
|
16207378
|
PMC1277831
|
CC BY
|
2021-01-04 16:38:10
|
no
|
Cardiovasc Diabetol. 2005 Oct 5; 4:16
|
utf-8
|
Cardiovasc Diabetol
| 2,005 |
10.1186/1475-2840-4-16
|
oa_comm
|
==== Front
Chiropr OsteopatChiropractic & Osteopathy1746-1340BioMed Central London 1746-1340-13-211619755510.1186/1746-1340-13-21ResearchLow back pain risk factors in a large rural Australian Aboriginal community. An opportunity for managing co-morbidities? Vindigni Dein [email protected] Bruce F [email protected] Jennifer R [email protected] Costa Cliff [email protected] Lynne [email protected] Steve [email protected] Private practice of chiropractic, 12 David Street, Lalor, Victoria, 3075, Australia2 School of Medicine, James Cook University, Townsville, Queensland, Australia3 School of Chiropractic, Murdoch University, Western Australia4 School of Mathematical & Geospatial Sciences, RMIT University, Melbourne, Australia5 Centre for Research and Education in Ageing, Faculty of Health, The University of Newcastle, New South Wales, Australia6 Chief Executive Officer, Durri Aboriginal Corporation Medical Service, Kempsey, New South Wales, Australia2005 30 9 2005 13 21 21 20 5 2005 30 9 2005 Copyright © 2005 Vindigni et al; licensee BioMed Central Ltd.2005Vindigni 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 back pain (LBP) is the most prevalent musculo-skeletal condition in rural and remote Australian Aboriginal communities. Smoking, physical inactivity and obesity are also prevalent amongst Indigenous people contributing to lifestyle diseases and concurrently to the high burden of low back pain.
Objectives
This paper aims to examine the association between LBP and modifiable risk factors in a large rural Indigenous community as a basis for informing a musculo-skeletal and related health promotion program.
Methods
A community Advisory Group (CAG) comprising Elders, Aboriginal Health Workers, academics, nurses, a general practitioner and chiropractors assisted in the development of measures to assess self-reported musculo-skeletal conditions including LBP risk factors. The Kempsey survey included a community-based survey administered by Aboriginal Health Workers followed by a clinical assessment conducted by chiropractors.
Results
Age and gender characteristics of this Indigenous sample (n = 189) were comparable to those reported in previous Australian Bureau of Statistics (ABS) studies of the broader Indigenous population. A history of traumatic events was highly prevalent in the community, as were occupational risk factors. Thirty-four percent of participants reported a previous history of LBP. Sporting injuries were associated with multiple musculo-skeletal conditions, including LBP. Those reporting high levels of pain were often overweight or obese and obesity was associated with self-reported low back strain. Common barriers to medical management of LBP included an attitude of being able to cope with pain, poor health, and the lack of affordable and appropriate health care services.
Though many of the modifiable risk factors known to be associated with LBP were highly prevalent in this study, none of these were statistically associated with LBP.
Conclusion
Addressing particular modifiable risk factors associated with LBP such as smoking, physical inactivity and obesity may also present a wider opportunity to prevent and manage the high burden of illness imposed by co-morbidities such as heart disease and type-2 diabetes.
Low back painrisk factorschiropracticgeneral healthAustralianAboriginalIndigenous
==== Body
Background
Low back pain (LBP) is the most prevalent musculo-skeletal condition in rural and remote communities [1-3]. Indigenous people in these communities are over-represented in low-skilled, manual jobs and the community-service sector [4]. As such they are more likely to be exposed to greater manual handling of loads, repetitive strains and risk of musculo-skeletal conditions. Formal reporting of such conditions in the Australian Indigenous community is infrequent [1]. These occupational factors and resulting LBP may be compounded by lifestyle risk factors including smoking, physical inactivity, and obesity [5].
There is an abundance of literature reporting on the risk factors associated with LBP in the general population [6]. Known modifiable risk factors for low back pain are lack of fitness, poor health, obesity, smoking, drug dependence, and occupational factors including heavy lifting, twisting, bending, stooping, awkward posture at work and prolonged sitting. Those that are non-modifiable are increasing age, number of children, a previous episode of LBP and major scoliosis [6]. Within the public health context it is important to prevent injuries and painful conditions by addressing modifiable risk factors [7-9].
Australian Indigenous communities experience sub-optimal mortality and morbidity rates. As such it has been argued that by adopting a holistic approach and addressing modifiable risk factors associated with LBP, such as smoking, physical inactivity and obesity, the clinical management of co-morbidities such as heart disease and diabetes may also be partially addressed [10]. Exercise, for example, has been reported as the single most important lifestyle factor for preventing and managing insulin resistance especially among those who are obese [11,12] It is also known that once their presenting musculoskeletal condition has been effectively managed, patients are more likely to comply with their practitioner's advice to promote other aspects of their health including weight loss and increased physical activity [10].
Modifiable risk factors for LBP mentioned above have been further classified as lifestyle (physical inactivity, poor muscle strength, obesity, smoking), and occupational (heavy lifting, twisting, bending, stooping, prolonged sitting, awkward posture at work, previous history of injury to the area) [6]. These are summarised in Table 1. Where high levels of evidence (Level I evidence) such as meta-analyses or systematic reviews were not available, less rigorous studies (Level II, III and IV evidence) were reported to represent the current levels of knowledge.
Table 1 Individual modifiable risk factors associated with low back pain
Factors strongly associated with LBP (OR > 1.2-) Factors moderately associated with LBP (OR ≥ 1–1.2-)
Lack of fitness/Physical inactivity Balague, 1999 [44]*
Feuerstein, 1999[45] ****
Smoking Balague, 1999[44] *
Feldmann, 1999[47] ***
Levangie, 1999[48] ***
Power, 2001[49] ** Leboeuf-Yde, 1995[46] *
Obesity Koda, 1991[50] ****
Alcouffe, 1999[51] ****
Walker, 1999[52] **
Fransen, 2002[53] ****
Webb, 2003[55] **** Leboeuf-Yde, 1999[46] *
Balague, 1999[44] *
Levangie, 1999[48] ****
Lecerf, 2003[54] ****
Mirtz, 2005[56] **
Psychosocial stress Balague, 1995[57] ***
Hagg, 1997[58] ****
Josephson, 1998[60] ****
Adams, 1999[61] ***
Krause, 1998[62] ***
Feuerstein, 1999[45] ****
Bildt, 2000[63] ***
Thorbjornsson, 2000[64] ***
Vingard, 2000[65] ****
Yip, 2001[66] ****
Power, 2001[67] **
Harkness, 2003[68] ***
Van den Heuvel, 2004[69] *** Balague, 1999[44] *
Hoogendorm, 2000[59] ***
Physical trauma Harkness, 2003[68] ***
Balague, 1999[44] *
Factors strongly associated with LBP (OR > 1.2-) Factors moderately associated with LBP (OR > 1–1.2-)
Awkward posture (at work) Koda, 1991[50] ****
Alcouffe, 1999[51] ****
Jin, 2000[53] ** Picavet, 2000[70] ***
Frequent bending and twisting Alcouffe, 1999[51] ****
Hoogendoorm, 2000[59] ***
Vingard, 2000[65] ****
Jin, 2000[71] **
Van den Heuvel, 2004[69] *** Picavet, 2000[70] ***
Heavy lifting, repetitive lifting Suadicani, 1994[72] ****
Marras, 1995[73] ****
Magnusson, 1996[74] ****
Sturmer, 1997[75] ****
Krause, 1998[62] ***
Josephson, 1998[60] ****
Alcouffe, 1999[51] ****
Thorbjornsson, 2000[64] ***
Vingard, 2000[65] ****
Hartvigsen, 2001[76] ***
Nahit, 2001[77] ****
Fransen, 2002[53] ****
Harkness, 2003[68] ***
Jarring, Gripping, vibration, repetitive actions Bongers, 1993[78] *
Magnusson, 1996[74] ****
Levangie, 1999[48] ***
Pope, 1999[79] ***
Jin, 2000[71] *
Prolonged sitting & prolonged standing Burdorf, 1994[80] ****
Bongers, 1993[78] **
Thorbjornsson, 2000[64] *** Hartvigsen, 2000[81] ***
NB: Only first authors included.
Legend: + OR: Odds ratio
Level I evidence *, Level II evidence **, Level III evidence ***, Level IV evidence****
• Level I – based on studies such as meta-analyses or systematic reviews of all relevant randomised controlled trials (RCTs);
• Level II – based on well-designed RCTs;
• Level III – based on well-designed prospective or case-control analytical studies; and
• Level IV – based on opinions of respected authorities, clinical experience, descriptive studies and case reports or reports of expert committees.
As part of a study investigating the prevalence of LBP in this community [3], the risk factors known to be associated with LBP and other serious causes of morbidity and mortality were measured. This paper aims to describe the most commonly reported risk factors for LBP in a large rural Indigenous community; and examine their association with reported LBP as a basis for informing the development of a broad health promotion intervention in this community.
Methods and materials
Design
A cross-sectional self-report survey (Kempsey survey) was conducted to determine the extent of risk factors (Table 1) and their association with LBP in the study community.
Ethics: consent and approval
Participating community members completed a consent form that explained the purpose of the survey. Ethics approval was obtained from the Durri Aboriginal Corporation Medical Service (ACMS) Board of Directors and the Human Research Ethics Committee of the University of Newcastle.
Community consultation, collaboration and ownership of the program
The Durri Community of Kempsey, NSW, Australia, comprises one of Australia's largest rural Aboriginal communities. The Durri (ACMS) is at the forefront of providing culturally appropriate care, largely via its Aboriginal Health Workers (AHWs). Durri ACMS aims to:
'make primary health care and education accessible to all members of the community in a culturally appropriate and spiritually sensitive manner, endeavouring to improve not only the health status but also the well-being of the Durri Aboriginal community' [13].
Discussions with a cross section of community members led to the formation of a Community Advisory Group (CAG) (which included representatives from the Durri ACMS, Booroongen Djugun Aboriginal Health Worker College, Hands On Health Australia and the University of Newcastle). The CAG aimed to advise on the development and implementation of the musculo-skeletal prevalence study [14]. Aboriginal Health Workers were chosen as the study agents because they are recognised as essential in providing culturally appropriate and effective health-care for their communities [15-22].
Community consultation occurred throughout the study. This process involved regular discussions with key-informants from the community including AHWs, elders and health professionals. The community was informed of developments via information sheets and the publication of a summary report during the process and at the completion of the study.
Sample
Our goal was to select a representative cross-sectional sample of the local Aboriginal community of sufficient size to generalise our major findings to the whole local community (population 550). A random sampling procedure stratifying for age and sex was used to derive a representative sample of the local community. The sample size was generated using Epi-Info 6 [23]. With a population size of 550, the expected frequency of the main variable of interest (low back pain) was estimated at 50%. The value chosen as the farthest acceptable from the real population was 44%. Using these values and a 95% confidence interval, the ideal random sample size calculated was 180. However, we expected that logistically this was unlikely to be achieved, as many of the sample selected were likely to be uncontactable given the transient nature of community residents [24]. Accordingly, where randomly selected community members were unable to participate, they were replaced using a convenience sampling approach to achieve the required sample size. Although this strategy was not ideal, all attempts were made to attain a representative sample. Participants within the community were selected from persons aged 15-years or older who had been previously identified as Aboriginal (according to the definition of Aboriginal adopted by the Department of Aboriginal Affairs Constitutional Section) [25]. These participants were recruited by distributing letters inviting them to contact the assisting AHWs at the ACMS. If no response was received within a week, an attempt to contact the person via telephone was made by the assisting AHW.
Procedure
The Kempsey survey included a screening survey administered by Aboriginal Health Workers immediately followed by a clinical conducted by chiropractors blinded to the findings of the screening survey.
Those who consented to participate were asked to attend the Durri ACMS. If participants found transport to the ACMS difficult, either the research team (including the researcher, the AHW and volunteer chiropractors/chiropractic students) would travel to the participants' homes, or the assisting AHW would arrange for the Durri ACMS bus to provide transportation at no charge.
Screening survey
Participants completed a screening survey previously found to be culturally acceptable and sensitive in measuring musculo-skeletal conditions and associated risk factors in this community. The survey achieved satisfactory measurement agreement (Kappa scores) when compared to a clinical assessment performed by chiropractors (a proxy "Gold Standard") [22]. Although some authors argue that a 'Gold standard' does not exist in many areas of musculo-skeletal practice [26], standard clinical assessments performed by musculo-skeletal health professionals provide the best available tools for measuring painful and limited ranges of motion and a provisional diagnosis [27]. The purpose of the screening survey was to identify those who had experienced a musculo-skeletal condition including ache, pain or discomfort. The questionnaire also assessed self-reported limitations to Activities of Daily Living (ADL) imposed by pain.
Participants screened by the AHW-administered survey subsequently underwent a clinical examination conducted by four chiropractors previously trained and assessed in standard, clinical assessment procedures according to a procedural manual which outlined the cultural considerations and logistical processes required by researchers. The content of the procedural manual was revised in a two-hour workshop for participating researchers to clarify and standardise study requirements. The exam was based on accepted clinical parameters for conducting musculoskeletal conditions and included the domains of assessment used by teaching institutions [28]. Thus attempts were made to fulfil content and face validity.
Assessment
Participants attended a clinical assessment immediately following the screening survey to confirm the presence of musculo-skeletal conditions [22]. Chiropractors and 5th year chiropractic students performed a follow-up clinical assessment (based on clinical assessment parameters used in 1999 at the School of Chiropractic, RMIT University, Victoria, Australia) [28] to validate the findings reported in the screening questionnaire.
A positive pain finding in the clinical assessment was derived by practitioner-based examination, including the patient's history of involved site(s) followed by standard orthopaedic and range of motion tests to localise sites of pain and restricted movement. A negative pain finding was indicated by the absence of reported pain and/or restricted orthopaedic and range of motion findings as examined by the practitioner. Trivial LBP was differentiated from important LBP using a Likert scale. High levels of pain were interpreted as those ranging between 6–10 on a Likert scale of 0–10. Only those reporting "High" levels of pain were analysed in this study. Further questions related to any musculo-skeletal condition(s) experienced in the last seven days. In particular, probable causes of symptoms, past history, initial episode(s) of symptoms, duration of symptom(s), 'average' severity of symptoms and any associated limitation of daily activities. Also examined were, social routine and work activities, the type of treatment received and any barriers to receiving treatment were sought.
In the history component of the clinical assessment, chiropractors once again questioned participants about the presence of musculo-skeletal risk factors (according to the criteria reported in Table 1). Risk factor data were derived in the history component of the clinical assessment by asking questions from a list of modifiable occupational and lifestyle factors. Results for LBP as measured in the clinical assessment were used in the analysis. Clinical findings requiring follow-up treatment, management or referral was also identified.
Health workers using a laptop computer entered data on-site into a specifically designed, Microsoft Access database.
Screening and assessment agreement
The questionnaire results were compared to the data from the clinical examination and published in a previous study (Table 2). Eighty-three percent of all participants reporting LBP in the screening survey also tested positive for LBP via the clinical assessment. Sensitivity of the screening survey for LBP was 0.83, specificity 0.63 and Kappa 0.46. Thus the screening survey achieved an adequate level of agreement with the clinical assessment [29].
Table 2 Sensitivity, specificity and Kappa for LBP screening survey compared to clinical assessment (n = 189)
Survey results Clinical Assessment
Negative Positive Total Sensitivity Specificity Kappa coefficient
Negative 43 21 64 0.83 0.63 0.46
Positive 25 100 125
Total 68 121 189
Measures
The main variables of interest from the survey and clinical assessment were:
• Demographic and other sample characteristics-age, sex, number of children, occupation, weight, and Body Mass Index (BMI).
• Prevalence of LBP (within the last seven days, according to self report).
• Pain levels were recorded using a Likert scale where a score of 0 corresponded to no pain and 10 to severe pain.
• Duration of LBP was categorised as less than/equal to or more than seven weeks.
• Disability levels were recorded using a Likert scale where a score of 0 corresponded to no disability and 10 to severe disability. Disability was defined as "how much the condition (ache, pain or discomfort) had affected the participants ability to carry out daily activities (e.g., housework, washing, dressing, lifting, walking, driving, climbing stairs, getting in and out of bed or a chair, sleeping, working, social activities and sport)".
• Self-reported modifiable risk factors as described in Table 1 (according to a standardised clinical history).
• Other musculo-skeletal conditions.
Analyses
Frequencies and confidence intervals were reported for characteristics of the sample, prevalence of LBP and reported risk factors for low back pain. Chi-square analyses were performed to test for factors associated with low back pain. Given the number of variables, only significant associations were reported.
Results
Sample
The study was conducted between January 2001 and July 2002. The sample comprised 189 Indigenous people: 80 were selected randomly and the remainder were convenience sampled as described above.
Sample characteristics
Age and sex
The mean age of participants was 44 years ( ± 14.8) and the median age 43 years. The sample comprised 87 males (46%) and 102 females (53%) ranging in age from 15 to 80 years. There were no significant differences in the distribution of males and females in the various age categories (p = 0.35). Gender was comparable with previous ABS census data for Indigenous people in Australia [26]. Age categories were also similar in breakdown to those described in census data for the entire Indigenous community (Table 3) [30].
Table 3 Age and sex of study participants
Age category (years) Male Female Total % Male % Female % Total
15 – 25 20 20 40 23.0 19.6 21.2
26 – 35 14 16 30 16.1 15.7 15.9
36 – 45 25 29 54 28.7 28.4 28.6
46 – 55 13 10 23 14.9 9.8 12.2
56 + 12 24 36 13.8 23.5 19.0
Unknown 3 3 6 3.4 2.9 3.2
Total 87 102 189 100 100 100
Despite a high consent rate (85% of the randomly recruited sample), the response rate was low (40%) because many members of this highly mobile community were unable to be contacted.
Number of children
Approximately one third (31%) of participants had between two or three children. Thirty percent of participants had no dependent children and 17% had 4–5 children. Of note, 15% had six or more children. These findings are comparable to those of other Indigenous studies [5]. An Australian Bureau of Statistics (ABS) study reported that Indigenous families tend to be larger than Australian families overall. According to the 1996 Census, approximately 13% of Indigenous families had four or more children compared with less than 5% of other Australian families [5].
Occupation
Occupational demographics of the participants in the study are summarised in Table 4. Approximately one third of the community surveyed were students or unemployed. A significant number of people surveyed were associate professionals, retired workers, involved in home duties or labourers. These data were generally comparable with those reported for Indigenous people by the ABS (2000). However, for males in the Kempsey survey, there were significantly less professionals, managers, tradespersons and transport workers, and more intermediate clerical, sales and service persons, compared to the ABS population. For females there were significantly more professional, and associates professionals (such as Aboriginal Health Workers), and less tradespersons or transport workers as well as many less intermediate clerical, sales and service persons, compared to the ABS population [5].
Table 4 Occupation of study participants according to sex
Occupation Male Female Total % Male % Female % Total
Managers and Administrators 5 3 8 5.7 2.9 4.2
Professionals 7 9 16 8.0 8.8 8.5
Associate professionals* 5 16 21 5.7 15.7 11.1
Tradespersons and related workers 1 2 3 1.1 2.0 1.6
Advanced clerical and service workers 3 2 5 3.4 2.0 2.6
Intermediate clerical, Sales and service workers 3 2 5 3.4 2.0 2.6
Elementary Clerical, Sales and Service workers 2 6 8 2.3 5.9 4.2
Labourers and Related workers 13 3 16 14.9 2.9 8.5
Unemployed/Student 38 28 66 43.7 27.5 34.9
Home duties 1 16 17 1.1 15.7 9.0
Retired 4 15 19 4.6 14.7 10.1
Unknown 5 0 5 0.0 2.6
Total 87 102 189 100 100 100
* Associate Professionals
BMI
Table 5 shows that 32% of participants were overweight and 39% were obese. Using Body Mass Index (BMI) estimates, 26% (95% CI: 20%–32%) of participants were overweight (BMI = 25.0–29.9) and 45% (95% CI: 38%–52%) were obese (BMI = 30.00). The high prevalence of obesity in this study agrees with national figures demonstrating a greater prevalence of obesity among Indigenous people than non-Indigenous Australians [5].
Table 5 Body Mass Index (BMI) of participants, according to age and sex (n = 189)
BMI classification
Age (yrs) Sex Normal (%) Overweight (%) Obese (%) Unknown (%) Total (%)
15 – 25 Male 10 23% 7 14% 2 .02% 0 0% 19 10%
Female 7 16% 5 10% 9 12% 0 0% 21 12%
Total 17 39.5% 12 24% 11 14% 0 0% 40 22%
26 – 45 Male 5 12% 13 26% 18 23% 4 33% 40 22%
Female 14 33% 9 18% 18 23% 5 42% 46% 25%
Total 19 44% 22 44% 36 47% 9 75% 86 47%
> 45 Male 4 9% 6 12% 13 17% 1 8% 24 13%
Female 3 7% 10 20% 17 22% 2 17% 32 18%
Total 7 16% 16 32% 30 39% 3 25% 56 31%
TOTAL 43 100% 50 100% 77 100% 12 100% 182 100%
Note: BMI = Weight (kg) divided by square of height (m)
Self-Report of LBP within the last seven days
The prevalence of all LBP (i.e. including all levels of pain) within the last seven days was 72% (95% CI: 63%–80%) and all LBP lasting seven weeks or longer was 34 % (95% CI: 27%–40%).
Previous history of LBP
Previous history of LBP was present in 34% (95% CI: 27%–40%) of respondents. A previous history of LBP is known to predispose individuals to recurrent episodes of back pain [31].
Other modifiable risk factors for LBP
Smoking
Smoking was highly prevalent 46% (95% CI: 38%–53%) in the community, with equal numbers of males and females smoking. Thirty eight per cent (95% CI: 31%–45%) of people smoked between 10–20 cigarettes daily and 8% (95% CI: .04%–11%) smoked more than 20 cigarettes per day. This is consistent with the 2001 National Health Survey (NHS), which found that 51% of Indigenous people aged 18 years or older were current smokers, compared with 24% of non-Indigenous people [32].
Physical inactivity
Sixteen percent (95%CI: 10%–21%) of participants spent no time actively exercising and 35.9% (95% CI: 26%–45%) exercised less than 30 minutes per week. There are no other detailed data available on the levels of physical activity among Indigenous people. However, the 2001 NHS reported that 43% of Indigenous people aged 18 years or older living in non-remote areas were sedentary, compared with 30% of non-Indigenous people [32].
Psychosocial stress
For those reporting LBP 72% (CI: 65%–78%), the most commonly reported traumatic events included sporting injuries 26.5% (95% CI: 20%–38%), motor vehicle accidents 18% (95% CI: 12%–23%) and work-related trauma 17.5% (95% CI: 12%–22%). There was, however, no association between LBP and physical trauma.
Physical trauma
For those reporting LBP (66.1% CI: 54%–68%), the most commonly reported traumatic events included sporting injuries 26.5% (95% CI: 20%–38%), motor vehicle accidents 18% (95% CI: 12%–23%) and work-related trauma 17.5% (95% CI: 12%–22%). There was, however, no association between LBP and physical trauma.
Occupational risk factors
Figure 1, Modifiable occupational risk factors for musculo-skeletal conditions details reported occupational risk factors for LBP. Common risk factors were adopting awkward postures at work 32% (95% CI: 25%–39%), frequent bending and twisting 29% (95%: CI: 22%–35%) and heavy lifting 26% (95% CI: 20% – 32%). However, there was no association between LBP and occupational risk factors.
Figure 1 Figure 1
Factors associated with reported LBP
Even though a trend was evident, no statistical association between LBP and the lifestyle factors detailed above. However, more participants reporting high levels of LBP were overweight or obese and obesity was statistically associated with self-reported strain causing reported LBP (χ2 = 9.02, df = 2 10, p = 0.01). While sporting injuries were not statistically associated with report of LBP in particular, participants reporting sporting injuries experienced between two and four musculo-skeletal conditions (χ2 = 7.90, df = 2, p = 0.02).
Discussion
The 72% seven day prevalence of LBP found in the Kempsey survey is greater than similar prevalence levels reported in other rural Indigenous Communities [1,2,33,34]. In their study, Honeyman and Jacobs [2] reported a 1-day LBP prevalence for the majority of community members, 68% (95% CI: 61%–74%). The majority of participants in the Kempsey survey also experienced their presenting LBP for seven weeks or more. Thus according to accepted definitions of chronicity [35], the majority of Indigenous people in this Community were suffering from chronic pain and were therefore, likely to be at greater risk of enduring prolonged disability [31]. Thirty-four percent of participants also reported a previous history of LBP, which was likely to predispose them to recurrent, future episodes [31]. Furthermore, trauma particularly that incurred in sporting injuries was associated with multiple musculo-skeletal conditions. Past studies have reported that Indigenous people are more likely to experience transport accidents, intentional self-harm and assault than other Australians with rates approximating three times those of the rest of the Australian population [32].
The findings in this study of higher levels of smoking, physical inactivity and obesity are consistent with those reported by other studies of Indigenous Australians [9]. Though many of the modifiable risk factors known to be associated with LBP were highly prevalent in this study, none of these were statistically associated with LBP. One explanation for this finding is that the size of the sample, though sufficiently large to demonstrate comparability with ABS findings for demographic categories, may not have been sufficiently large to achieve the statistical power to detect any association between LBP and associated study factors.
Obesity and physical inactivity are the two most important modifiable factors contributing to the development of type 2 diabetes mellitus. These factors were highly prevalent in the community with 26% of subjects overweight, 45% obese and 16% spending no time actively exercising plus a further 35.9% exercising less than 30 minutes per day. Exercising was assessed by self-report according to total time spent exercising ranging from 'No time' to 'More than 10 hours per week'. Obesity in this study was associated with self-reported low back strain. The prevalence of obesity in this community is of concern, first because obesity is an independent predictor of back pain [36], but more importantly as obesity has a global health impact.
Health providers including chiropractors and osteopaths commonly counsel LBP sufferers to lose weight to unload their spines. Weight loss also offers other musculo-skeletal benefits. Females with a BMI of over 25 kg/m2, can, by losing 5 kg (2 BMI units) reduce future onset of knee osteoarthritis by 50% and males by 25% [37]. Obesity has also been associated with a higher prevalence of work limitations, hypertension, dyslipidemia, type 2 diabetes and the metabolic syndrome in adults of working age [38]. Furthermore, Australia-wide some 50% of cases of type 2 diabetes are asymptomatic, undiagnosed and persons subclinically undergo progressive macro and micro-vascular changes [39]. The current findings suggest that screening this population group for evidence of glucose intolerance when reviewing musculo-skeletal conditions such as LBP may be valuable.
Of those reporting LBP, 72% of participants (CI: 65%–78%) were frequently exposed to "stressful situations" in their occupation. However, psychosocial stress outside of the work place was not measured given the cultural sensitivity of this factor according to the CAG. Psychosocial stress in general is a strong predictor of LBP [40,41]. If conducted in a culturally appropriate manner, future studies assessing LBP in Indigenous Communities should ideally attempt to also measure psychosocial stress as a potential contributing study factor.
Another concurrent health hazard is the high prevalence of cigarette smoking. In addition to the well documented risks of smoking it has been found that compared with matched groups of non-smokers, chronic cigarette smokers are more likely to be insulin resistant, hyperinsulinemic, and dyslipidemic [39].
Exercise is the most common method of treating LBP in Australia [42]. In addition it may be the single most important lifestyle factor for both preventing and reversing insulin resistance, particularly among obese individuals [12,13]. This suggests a good case for concentrating on general exercise health promotion for Indigenous communities.
Lifestyle interventions incorporated into a culturally sensitive health promotion program could potentially benefit the health and modify the morbidity and mortality of this population group. These results suggest an opportunity to review and address risk factors associated with LBP along with more serious diseases affecting Indigenous people. Addressing modifiable risk factors associated with LBP, such as smoking, physical inactivity, and obesity could significantly contribute to the management of co-morbidities including diabetes and heart disease which so commonly affect Indigenous Australians.
An understanding of the modifiable risk factors for LBP identified in this paper also formed the basis for a culturally acceptable musculo-skeletal intervention designed to address the high prevalence of LBP. This involved using a pilot training program for Aboriginal Health Workers (AHWs). The intervention was designed to promote the musculo-skeletal and general health of Indigenous people living in this rural community [12]. Culturally sensitive approaches to managing musculoskeletal conditions have been successfully implemented in other Indigenous Communities [43].
The Community Oriented Program for the Control of the Rheumatic Diseases (COPCORD) represents the largest, ongoing collaborative attempt to measure the prevalence of musculo-skeletal conditions and risk factors in rural populations throughout the world [43]. COPCORD has also developed implemented and evaluated culturally sensitive approaches for managing these conditions and their associated risk factors through community-based initiatives with applicability in other Indigenous Communities.
We propose that any future musculo-skeletal study or intervention in an Indigenous community be accompanied by a review of the modifiable risk factors associated with LBP and counselling about those factors. This may have a beneficial effect on the overall well being of indigenous communities. Further research should test such a program for efficacy and effectiveness.
Conclusion
The disturbingly high prevalence of LBP experienced in this community necessitates a serious response. Managing LBP through health services and addressing the modifiable risk factors through culturally sensitive, health promotion programs will be an important step in addressing the high burden of illness imposed by LBP and other more serious conditions suffered in this community.
Competing interests
Dr. Bruce Walker is Editor-in-Chief of Chiropractic & Osteopathy.
Acknowledgements
The authors would like to acknowledge the assistance of the Durri ACMS, NSW, as well as the Booroongen Djugun College, NSW, The Murray School of Health Education, NSW, and volunteers from the RMIT University, Victoria, Australia. The authors also thank Hands on Health Australia for funding the program. Also, Dr Janice Perkins for introducing the authors to the community and assisting in the design of the original program from which this study was drawn, Mrs Karen Woulfe for kindly proof-reading the text, Michael Dalton for data and statistical consultancy and Julie Bateman for formatting the paper.
==== Refs
Mayhew C Aboriginal and Torres Strait Islander occupational health and safety: a pilot study in Queensland Australian Aboriginal Studies 1996 2 61 8
Honeyman PT Jacobs EA Effects of culture on back pain in Australian aboriginals Spine 1996 21 841 3 8779015 10.1097/00007632-199604010-00013
Vindigni D Griffen D Parkinson J Da Costa C Parkinson L The prevalence of musculo-skeletal conditions, associated pain and disability and the barriers to managing these conditions in a rural, Australian Aboriginal Community Rural and Remote Health 2004
Boreham P Whitehouse G Harley B The labour force status of Aboriginal people: a regional comparison Labour and Industry 1993 5 16 32
Australian Institute of Health and Welfare Australia's Health 2000: the seventh biennial health report of the Australian Institute of Health and Welfare 2000 Canberra: AIHW
Haldeman S Hartvigsen J Risk factors for low back and neck pain: An introduction to clinical epidemiology and review of commonly suspected risk factors Principles and Practice of Chiropractic 2005 3 The McGraw-Hill Companies 465 483
Dwyer AP Backache and its prevention Clin Orthop 1987 222 35 43 2957137
Farrelly P Ergonomic Applications in the Workplace for Clients with Low Back Pain Proceedings of Commonwealth Rehabilitation Conference 1991 Melbourne
Mathers C Penm RV Health system costs of injury, poisoning and musculoskeletal disorders in Australia 1993–94 AIHW Australian Institute of Health and Welfare, AIHW Cat No HWE 12: Health and Welfare Expenditure Series No 6 1999 Canberra: AIHW
Pill RM Jones-Elwyn G Stott NC Opportunistic health promotion: quantity or quality? JR Coll Gen Prac 1989 39 196 200
Lehmann R Vokac A Niedermann K Loss of abdominal fat and improvement of the cardiovascular risk profile by regular moderate exercise training in patients with NIDDM Diabetologia 1995 38 1313 1319 8582541
Kelley DE Goodpaster B Wing RR Simoneau JA Skeletal muscle fatty acid metabolism in association with insulin resistance, obesity and weight loss Am J Physiol 1999 277 E1130 E114 10600804
Durri Aboriginal Corporation Medical Service Annual Report. 2000–2001 Kempsey, NSW, Durri Aboriginal Corporation Medical Service 2001
Vindigni D Parkinson L Blunden S Perkins J Rivett D Walker B Aboriginal health in Aboriginal hands: Development, delivery and evaluation of a training program for Aboriginal Health Workers to promote the musculoskeletal health of indigenous people living in a rural community Rural and Remote Health 2004 281 15887988
Saggers S Gray D In Aboriginal Health and Society: the traditional and contemporary Aboriginal struggle for Better Health 1991 Sydney, Allen and Unwin
Ezzy J Competency based nutrition education and training for trainee Aboriginal health workers at Batchelor College Nutrition Networks '95 Conference Proceedings of the 4th National Conference 1995 'Waranyjarrigun Yagarrama Yagarrayirr' Gathering and Sharing'; Broome, Western Australia
Doyle MJ Training Koori Health Workers for Effective Community-based Eye care Services Ninth National Health Promotion Conference People, Places and Programs Darwin 18–20 May, 1997
Keefe J McCarthy C Doyle M Taylor H The Rumbalara Eye Care Project: initial results of vision screening by Aboriginal health workers Proceedings, 28th Annual Conference of the Public Health Association of Australia, Threats to Public Health Challenges and Strategies Perth, Western Australia 29 September-2 October 1996
Doyle M Keefe J McCarthy C Harper C Taylor H Training Koori Health Workers for effective community-based eye care services AAHP and PHA [HPSIG] Ninth National health promotion Conference People Places and Programs: exploring better practice and health promotion for diverse populations Darwin 18–20 May, 1997
Pacza T Steele L Tennant M Dental train-the-trainer program for Aboriginal health workers in the Kimberley Region Report prepared by the School of Oral Health Sciences 2000 The University of Western Australia, Nedlands 6907
Clifford A Atkinson D Aboriginal and Torres Strait Islander Health Worker Training in Western Australia Conference Proceeding 2000 Perth 31 39
Vindigni D Perkins J Blunden S The prevalence of musculoskeletal conditions among Indigenous people living in rural Australia: an opportunity for health promotion? Aboriginal & Torres Strait Islander Health Worker Journal 2003 27 4 8
Epi-Info Version 6.04 1997 Center for Disease Control and Prevention (CDC), Atlanta
Donovan RJ Spark R Towards guidelines for survey research in remote Aboriginal Communities Aust NZJ Public Health 1997 21 89 95
Department of Aboriginal Affairs Constitutional Section Report on a Review of the Administration of the Working Definition of Aboriginal and Torres Strait Islander 2000 Canberra; Department of Aboriginal Affairs
Beruskens AJ de Vet HC Koke AJ van der Heijden GJ Knipschild PG Measuring the functional status of patients with low back pain. Assessment of the quality of disease-specific questionnaires Spine 1995 20 1017 1028 7631231
Haldeman S, Chapman-Smith D, Petersen DM Guidelines for Chiropractic Quality Assurance and Practice Parameters (Mercy Guidelines) 1992 Consensus Conference Commissioned by the Congress of Chiropractic State Associations. Mercy Conference Center, Burlingame, California, USA
Protocol, School of Chiropractic, RMIT 1999 History and clinical assessment proforma for 4th and 5th year student preceptorship; Bundoora, Victoria; RMIT School of Chiropractic
Jekel J Elmore J Katz D Epidemiology, Biostatistics and Preventive Medicine 1996 Philadelphia: W.B. Saunders Company
Australian Bureau of Statistics Experimental estimates of the Aboriginal and Torres Strait Islander population, 30 June 1991 – 30 June 1996 1998 Canberra; Australian Bureau of Statistics
Frymoyer JW Back pain and sciatica N Engl J Med 1988 318 291 300 2961994
Australian Bureau of Statistics National Aboriginal and Torres Strait Islander survey detailed findings: 1995 1994 Canberra, Australian Bureau of Statistics
Lee TG Indigenous Health. A Needs Assessment Study of the Outer Eastern Metropolitan Region of Melbourne 1998 Detailed Report, Yarra Ranges Health Service, Victoria, a member of the inner and Eastern Health Care Network
McLennan W Madden R The Health and Welfare of Australia's Aboriginal and Torres Strait Islander Peoples ABS Catalogue no 470400 AIHW Catalogue no IHW 3 1999 Canberra, Australian Bureau of Statistics & the Australian Institute of Health and Welfare
National Advisory Committee on Health and Disability New Zealand Acute Low Back Pain Guide 1999 Wellington, New Zealand, Ministry of Health, Manatu Hauora. National Health Committee
Webb R Brammah T Lunt M Urwin M Allison T Symmons D Opportunities for prevention of 'clinically significant' knee pain: results from a population-based cross sectional survey J Public Health (Oxf) 2004 26 277 84 15454597 10.1093/pubmed/fdh162
Meisler JG St Jeor S ADA Position: Weight Management Journal of the American Dietetic Association 1997 97 1996
Hertz RP Unger AN McDonald M Lustik MB Biddulph-Krentar J The Impact of Obesity on Work Limitations and Cardiovascular Risk Factors in the U.S. Workforce J Occup Environ Med 2004 46 1196 1203 15591970
Facchini FS Hollenbeck CB Jeppersen J Insulin resistance and cigarette smoking Lancet 1992 339 1128 1130 1349365 10.1016/0140-6736(92)90730-Q
Papageorgiou AC Macfarlane GJ Thomas E Psychosocial factors in the workplace-do they predict new episodes of low back pain? Evidence from the South Manchester Back Pain Study Spine 1997 22 1137 42 9160473 10.1097/00007632-199705150-00014
Adams MA Mannion AF Dolan P Personal risk factors for first-time low back pain Spine 1999 24 2497 505 10626313 10.1097/00007632-199912010-00012
Walker BF Muller R Grant WD Low back pain in Australian adults. health provider utilization and care seeking J Manipulative Physiol Ther 2004 27 327 35 15195040 10.1016/j.jmpt.2004.04.006
Muirden KD The Origins Evolution and Future of COPCORD. APLAR Journal of Rheumatology 1997 1 44 8
Balague F Troussier B Salminen JJ Non-specific low back pain in children and adolescents: risk factors Eur Spine 1999 J8 429 38 10.1007/s005860050201
Feuerstein M Berkowitz SM Huang GD Predictors of occupational low back disability: implications for secondary prevention J Occup Environ Med 1999 41 1024 31 10609220
Leboeuf-Yde C Yashin A Smoking and low back pain: is the association real? J Manipulative Physiol Ther 1995 18 457 63 8568428
Feldman DE Rossignol M Shrier I Abenhaim L Smoking. A risk factor for development of low back pain in adolescents Spine 1999 24 2492 6 10626312 10.1097/00007632-199912010-00011
Levangie PK Association of Low Back Pain with Self-Reported Risk Factors Among Patients Seeking Physical Therapy Services Physical Therapy 1999 79 Research Report
Power C Frank J Hertzman C Schierhout G Li L Predictors of low back pain onset in a prospective British study Am J Public Health 2001 91 1671 8 11574334
Koda S Hisashige A Ogawa T Kurumatani N Dejima M Miyakita T Kodera R Hamada H Nakagiri S Aoyama H An epidemiological study on low back pain and occupational risk factors among clinical nurses Sangyo Igaku 1991 33 410 22 [Article in Japanese] 1836503
Alcouffe J Manillier P Brehier M Fabin C Faupin F Analysis by sex of low back pain among workers from small companies in the Paris area: severity and occupational consequences Occup Environ Med 1999 56 696 701 10658550
Walker BF The prevalence of low back pain in Australian adults. A systematic review of the literature from 1966–1998 Asia-Pacific Journal of Public Health 1999 11 45 51 10829828
Fransen M Woodward M Norton R Coggan C Dawe M Sheridan N Risk factors associated with the transition from acute to chronic occupational back pain Spine 2002 27 92 8 11805644 10.1097/00007632-200201010-00022
Lecerf JM Reitz C Chasteigner AD Evaluation of discomfort and complications in a population of 18102 overweight or obese patients Presse Medicale 2003 32 689 95
Webb R Brammah T Lunt M Prevalence and Predictors of Intense, Chronic, and Disabling Neck and Back Pain in the UK General Population Spine 2003 28 1195 202 12782992 10.1097/00007632-200306010-00021
Mirtz TA Greene L Is obesity a risk factor for low back pain? An example using the evidence to answer a clinical question Chiropr Osteopat 2005 13 Published online 2005 April 11. doi: 10.1186/1746-1340-13-2 15967048
Balague F Skovron ML Nordin M Low back pain in schoolchildren. A study of familial and psychological factors Spine 1995 20 1265 70 7660235
Hagg GM Astrom A Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders Int Arch Occup Environ Health 1997 69 423 32 9215929 10.1007/s004200050170
Hoogendoorm WE Bongers PM de Vet HC Flexion and rotation of the trunk and lifting at work are risk factors for low back pain: results of a prospective cohort study Spine 2000 25 3087 92 11145822 10.1097/00007632-200012010-00018
Josephson M Vingard E Workplace factors and care seeking for low-back pain among female nursing personnel. MUSIC-Norrtalje Study Group Scand J Work Environ Health 1998 24 465 72 9988088
Adams MA Mannion AF Dolan P Personal risk factors for first-time low back pain Spine 1999 24 2497 505 10626313 10.1097/00007632-199912010-00012
Krause N Ragland DR Fisher JM Syme SL Psychosocial job factors, physical workload, and incidence of work-related spinal injury: a 5-year prospective study of urban transit operators Spine 1998 23 2507 16 9854749 10.1097/00007632-199812010-00005
Bildt C Alfredsson L Michelsen H Occupational and nonoccupational risk indicators for incident and chronic low back pain in a sample of the Swedish general population during a 4-year period: an influence of depression? International Journal Behavioral Medicine 2000 7 372 92 10.1207/S15327558IJBM0704_07
Thorbjornsson CB Alfredsson L Fredriksson K Michelsen H Punnett L Vingard E Torgen M Kilbom A Physical and psychosocial factors related to low back pain during a 24-year period. A nested case-control analysis Spine 25 369 74 2000 Feb 1, discussion 375 10703112 10.1097/00007632-200002010-00019
Vingard E Alfredsson L Hagberg M To what extent do current and past physical and psychosocial occupational factors explain care-seeking for low back pain in a working population? Results from the Musculoskeletal Intervention Center-Norrtalje Study Spine 2000 25 493 500 10707397 10.1097/00007632-200002150-00017
Yip Y A study of work stress, patient handling activities and the risk of low back pain among nurses in Hong Kong J Adv Nurs 2001 36 794 804 11903709 10.1046/j.1365-2648.2001.02037.x
Power C Frank J Hertzman C Schierhout G Li L Predictors of low back pain onset in a prospective British study Am J Public Health 2001 91 1671 8 11574334
Harkness EF Macfarlane GJ Nahit ES Silman AJ McBeth J Risk factors for new-onset low back pain among cohorts of newly employed workers Rheumatology (Oxford) 2003 42 959 68 12730508 10.1093/rheumatology/keg265
van den Heuvel SG Ariens GA Boshuizen HC Hoogendoorn WE Bongers PM Prognostic factors related to recurrent low-back pain and sickness absence Scand J Work Environ Health 2004 30 459 67 15633597
Picavet HS Schouten JS Physical load in daily life and low back problems in the general population. The MORGEN study Prev Med 2000 31 506 12 11071830 10.1006/pmed.2000.0737
Jin K Sorock GS Courtney T Risk factors for work-related low back pain in the People's Republic of China Int J Occup Environ Health 2000 6 26 33 10637534
Suadicani P Hansen K Fenger AM Gyntelberg F. Low back pain in steel plant workers Occup Med (Lond) 1994 44 217 21 7949066
Marras WS Lavender SA Leurgans SE Fathallah FA Ferguson SA Allread WG Rajulu SL Ergonomics: Biomechanical risk factors for occupationally related low back disorders 1995 38 377 410 7895740
Magnusson ML Pope MH Wilder DG Areskoug B Are occupational drivers at an increased risk for developing musculoskeletal disorders? Spine 1996 21 710 7 8882693 10.1097/00007632-199603150-00010
Sturmer T Luessenhoop S Neth A Soyka M Karmaus W Toussaint R Liebs TR Rehder U Construction Work and Low Back Disorder: Preliminary Findings of the Hamburg Construction Worker Study Spine 22 2558 2563 1997 November 1 9383865 10.1097/00007632-199711010-00018
Hartvigsen J Bakketeig LS Leboeuf-Yde C Engberg M Lauritzen T The association between physical workload and low back pain clouded by the "healthy worker" effect: population-based cross-sectional and 5-year prospective questionnaire study Spine 26 1788 92 2001 Aug 15, discussion 1792-3 11493851 10.1097/00007632-200108150-00011
Nahit ES Macfarlane GJ Pritchard CM Cherry NM Silman AJ Short term influence of mechanical factors on regional musculoskeletal pain: a study of new workers from 12 occupational groups Occup Environ Med 2001 58 374 81 11351052 10.1136/oem.58.6.374
Bongers PM de Winter CR Kompier MA Hildebrandt VH Psychosocial factors at work and musculoskeletal disease Scand J Work Environ Health 1993 19 297 312 8296178
Pope MH Wilder DG Magnusson ML A review of studies on seated whole body vibration and low back pain Proc Inst Mech Eng [H] 1999 213 435 46 10635692 10.1243/0954411991535040
Burdorf A Naaktgeboren B de Groot HC. Occupational risk factors for low back pain among sedentary workers J Occup Med 1993 35 1213 20 8113925
Hartvigsen J Leboeuf-Yde C Lings S Corder EH Is sitting-while-at-work associated with low back pain? A systematic, critical literature review Scand J Public Health 2000 28 230 9 11045756 10.1080/140349400444940
|
16197555
|
PMC1277832
|
CC BY
|
2021-01-04 16:38:23
|
no
|
Chiropr Osteopat. 2005 Sep 30; 13:21
|
utf-8
|
Chiropr Osteopat
| 2,005 |
10.1186/1746-1340-13-21
|
oa_comm
|
==== Front
CoughCough (London, England)1745-9974BioMed Central London 1745-9974-1-101627090710.1186/1745-9974-1-10ResearchNo effect of omeprazole on pH of exhaled breath condensate in cough associated with gastro-oesophageal reflux Torrego Alfonso [email protected] Stefan [email protected] Mark [email protected] Kian Fan [email protected] Department of Thoracic Medicine, National Heart & Lung Institute, Imperial College and Royal Brompton Hospital, London, UK2005 19 10 2005 1 10 10 21 6 2005 19 10 2005 Copyright © 2005 Torrego et al; licensee BioMed Central Ltd.2005Torrego 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
Endogenous airway acidification evaluated as pH in exhaled breath condensate (EBC) has been described in patients with chronic cough. Proton pump inhibitors improve gastro-oesophageal reflux (GOR)-associated cough.
Methods
We examined pH levels in EBC and capsaicin cough response in 13 patients with chronic cough (mean age 41 years, SD 9) associated with GOR before and after omeprazole treatment (40 mg/day for 14 days) and its relationship with clinical response.
Results
Omeprazole abolished symptoms associated with GOR. Patients with chronic cough had an EBC pH of 8.28 (SD 0.13) prior to treatment but this did not change with omeprazole treatment. There was a significant improvement in the Leicester Cough Questionnaire symptom scores from 80.8 points (SD 13.2) to 95.1 (SD 17) (p = 0.02) and in a 6-point scale of cough scores, but there was no change in capsaicin cough response.
Conclusion
An improvement in GOR-associated cough was not associated with changes in EBC pH or capsaicin cough response. These parameters are not useful markers of therapeutic response.
==== Body
Introduction
Chronic cough, conventionally defined as a cough persisting for more than 8 weeks, is a common respiratory problem and, at times, presents as a difficult management issue. Asthma, rhino-sinusitis and gastro-oesophageal reflux (GOR) have been identified as the most common diagnoses associated with chronic cough [1]. GOR alone or in combination with other factors is the cause of chronic cough in 10–40% of adult patients [2,3]. Two main pathogenic mechanisms in GOR related cough have been described: micro-aspiration of gastric contents and a vagally-mediated oesophageal-tracheobronchial reflex [4]. The acid content of the refluxate may be an important component of the cough trigger associated with GOR, and this is supported by the fact that the chronic cough in some patients associated with GOR is improved or controlled by proton pump inhibitors that suppress gastric acid output [3,5,6]. Therefore, reflux of the gastric acid could directly activate cough receptors in the upper airways or indirectly through an oesophageal-tracheobronchial reflex [7].
Exhaled breath condensate (EBC) is a simple non-invasive technique for the monitoring of airway inflammation, since it may be representative of the epithelial lining fluid. Endogenous airway acidification, as assessed by the pH of exhaled breath condensate, has been reported in patients with non-asthmatic chronic cough, including GOR [8]. The fall in pH represented a doubling in the amount of H+ ions and this could contribute to the sensitised cough reflex measured with capsaicin since an acid environment has been shown to activate Aδ and C fibres in the airways of rodents [7,9].
In order to examine further the significance of acid pH in the pathogenesis of GOR-associated cough, we measured pH of exhaled breath condensate in patients with chronic cough associated with abnormal lower oesophageal pH. We determined whether the improvement in cough associated with treatment with proton pump inhibitors was associated with changes in capsaicin responsiveness and in EBC pH.
Methods
Subjects
We recruited 13 patients with chronic cough (age 41 ± 9, 5 males) defined as a cough persisting for more than 2 months, associated solely with GOR as defined by an abnormal 24-hour oesophageal pH measurement from our Cough Clinic. In these patients, we had excluded the presence of asthma and rhino-sinusitis. FEV1 (predicted value: 99.8 ± 8.0%) and FVC (103 ± 8.0%) were within the normal range. The chest radiograph was normal and histamine responsiveness measured as PC20 (the concentration of histamine causing a 20% fall in FEV1) as greater than 16 mg/ml. Skin prick to common allergens were negative and they had no nasal symptoms. Eight of 13 patients reported symptoms of heartburn, regurgitation or dyspepsia; the rest were asymptomatic. All participants were non-smokers. All subjects gave informed consent to participate in the study which was approved by the Royal Brompton and Harefield NHS Trust Ethics Committee.
Oesophageal pH study
An ambulatory 24 hour pH study was performed with the Synectics Digitrapper Mk III (Synectics Medical A/B, Sweden). An Antimony pH electrode was placed just above the upper border of the lower oesophageal sphincter. An acid reflux episode was defined as a drop in pH below 4.0. Significant reflux was defined as the total duration of reflux episodes exceeding 3.4% of the total study time.
Symptom questionnaire
Cough severity was assessed using the Leicester Cough Questionnaire [10]. This consist of 19 questions (scored from 1 to 7 points each) relating to quality of life issues associated with chronic cough. A higher score indicates better health status and the range of the scale is from 19 to 133. Additionally, we used a 6-scale incremental cough symptom score with 0 as being no cough and 5 being the worst score for distressing cough most of the time [11].
Capsaicin cough challenge
Capsaicin (8-methyl-N-vanillyl-6-nonenamide, 98%) obtained from Sigma-Aldrich, Gillingham, UK, was dispensed from a nebuliser chamber attached to a breath-activated dosimeter (PK Morgan Ltd, Gillingham, Kent, UK) set at driving pressure of 22 lbs/sq inch and a dosing period of 1 second. As described previously by Lalloo [12], the procedure started with the inhalation of 0.9% sodium chloride, followed with doubling doses of capsaicin from 0.976 μM (dose number 1) until 500 μM (dose number 10). The test was terminated when the subject coughed 5 times or more. The concentration of capsaicin causing 5 coughs or more (C5) was recorded.
Exhaled breath condensate collection
Exhaled breath condensate (EBC) was obtained non-invasively by using a condenser (EcoScreen; Jaeger; Wurzburg, Germany) that collected the nongaseous components of the expiratory air. Subjects breathed tidally through a mouthpiece and a two-way non-rebreathing valve, which also served as a saliva trap. They were asked to breathe at a normal frequency and tidal volume, wearing a nose clip, for a period of 10 min. If subjects felt saliva in their mouth, they were instructed to swallow it. The condensate (at least 1 ml) was collected on ice at -20°C, and was transferred to 15 ml Corning tubes. Measurement of pH was performed following de-aeration with argon (350 ml/min for 10 min), using a pH meter (Jenway 350 pH meter, Spectronic Instruments, Leeds, UK).
Study design
EBC collection, spirometry and capsaicin challenge were performed on the same day in this order. These measurements were performed before and after treatment with omeprazole (40 mg/day for 14 days)
Statistical analysis
Data were analysed using Graph-Prism version 3.0 (Graph-Pad Software, San Diego, CA, US). Data are expressed as the mean ± SD. Differences between groups were determined using the Mann-Whitney U test. Capsaicin C5 values were analysed as log10C5. All reported p values are two-tailed. A p value of less than 0.05 was considered statistically significant.
Results
The 8 patients with symptoms of gastro-oesophageal reflux reported disappearance of these symptoms. Using the Leicester cough questionnaire, in which the patients assessed their cough and related symptoms on a scale from 19 to 133 points, the patients reported a partial but significant symptomatic improvement after two weeks of omeprazole treatment (80.8 ± 13.2 vs. 95.1 ± 17 points, p = 0.02; Figure 1). Using the 6-point symptom score scale, we also found a reduction in cough score from 3.3 ± 0.7 to 2.6 ± 0.8 (p = 0.01). However, cough reflex sensitivity to capsaicin was not altered by omeprazole (log C5: 0.753 ± 0.23 vs. 0.707 ± 0.2; NS; Figure 2). There was no significant correlation between changes in the cough sensitivity reflex to inhaled capsaicin and the Leicester cough score. The log C5 was significantly lower than that measured in a cohort of 80 non-coughing normal volunteers (log C5: 1.83 ± 0.89; p < 0.0001), indicating that the coughers had a sensitised cough reflex. The pH of EBC was 8.28 ± 0.1 and did not change after 2 weeks of omeprazole treatment 8.25 ± 0.1 (Fig 3). EBC pH did not correlate with symptoms or with log C5.
Figure 1 Cough scores measured by the Leicester Cough Questionnaire before and after 2 weeks of omeprazole treatment. * p = 0.02.
Figure 2 Cough reflex sensitivity to inhaled capsaicin measured as the concentration of capsaicin causing 5 or more coughs (C5) before and after omeprazole treatment.
Figure 3 pH values in exhaled breath condensate in patients before and after omeprazole treatment. There was no effect of omeprazole.
Discussion
After 2 weeks' treatment with omeprazole, we found a partial but significant clinical improvement in cough severity as assessed using the validated Leicester cough questionnaire. This was not accompanied by changes in capsaicin cough response or by changes in pH of the exhaled breath condensate. We conclude that these measurements do not reflect the clinical response. Additionally, omeprazole does not change the pH of exhaled breath condensate, most likely a reflection of the lack of change in pH of the epithelial lining fluid. This may also indicate that direct reflux of gastric acid into the upper airway is an unlikely explanation of GOR-associated cough.
GOR is a common associated cause of chronic cough and treatment with gastric acid suppressing proton pump inhibitors is often effective in controlling cough [3,5,6]. We ascertained the presence of GOR by performing 24-hour lower oesophageal pH monitoring in 13 patients, in whom only 8 had symptoms of GOR. Although the main purpose of the study was to determine any change in pH of the exhaled breath condensate, we did find a significant improvement in cough severity after 14 days of treatment. This indicates that the therapeutic response resulting from suppression of GOR by proton pump inhibitors occurs rapidly. In a recent open study by Poe and Kallay, improvement in cough was observed in 16 of 42 patients at 2 weeks and in 38 at 4 weeks [3]. Therefore, we might have seen further improvement with prolonged treatment. The short duration of treatment might be a limitation of our study.
The baseline EBC pH value in our patients was not lower than that previously published for healthy controls [13,14]. However, in a previous study performed in our department, Niimi et al [8] found that the mean EBC pH of patients with cough due to GOR was significantly lower (7.90) than in our present study. A possible explanation for this discrepancy may due to the small number of patients with GOR-associated cough included in Niimi's work (n = 5) and the fact that one of the patients had an uncharacteristically low pH. If this outlier were to be excluded, the other 4 values would be in a similar range to ours.
The pathophysiological mechanisms underlying GOR-associated cough are not fully understood. Micro-aspiration of oesophageal contents into the larynx and tracheobronchial tree is one of the possible explanations [15]. Our study indicates that this is unlikely since suppression of gastric acid by omeprazole did not alter the pH of exhaled breath condensate. Yorulmaz et al, in a recently published work, could not demonstrate a significant relationship between acid reflux episodes, pH variations in the upper oesophageal segments and symptoms of laryngeal irritation such as cough [16].
We found that cough sensitivity to capsaicin was increased when compared to a group of historical non-coughing normal volunteers [17,18]. However, there was no effect of omeprazole on capsaicin sensitivity, despite a significant symptomatic improvement in cough, a finding that has been previously reported [18]. In one report, where capsaicin cough reflex improved after omeprazole, the patients had more severe GOR symptoms including posterior laryngitis and acid flooding of the oesophagus [17].
In conclusion, our results indicate that EBC pH measurement is not a good tool in the follow-up of GOR-associated chronic cough during treatment with a proton pump inhibitors. In GOR, episodes of micro-aspiration are short and do not produce a persisting level of airway acidification, which may be the reason why changes in EBC pH are not detected.
Acknowledgements
We thank the Lung Function laboratory of the Royal Brompton Hospital for the measurement of oesophageal pH.
==== Refs
Irwin RS Curley FJ French CL Chronic cough: the spectrum and frequency of causes, key components of the diagnostic evaluation, and outcome of specific therapy Amer Rev Respir Dis 1990 141 640 7 2178528
Fontana GA Pistolesi M Cough. 3: chronic cough and gastro-oesophageal reflux Thorax 2003 58 1092 5 14645983 10.1136/thorax.58.12.1092
Poe RH Kallay MC Chronic cough and gastroesophageal reflux disease: experience with specific therapy for diagnosis and treatment Chest 2003 123 679 84 12628862 10.1378/chest.123.3.679
Ing AJ Ngu MC Cough and gastro-oesophageal reflux Lancet 1999 353 944 6 10459900 10.1016/S0140-6736(98)00354-7
Ours TM Kavuru MS Schilz RJ Richter JE A prospective evaluation of esophageal testing and a double-blind, randomized study of omeprazole in a diagnostic and therapeutic algorithm for chronic cough Am J Gastroenterol 1999 94 3131 8 10566703 10.1111/j.1572-0241.1999.01504.x
Kiljander TO Salomaa ER Hietanen EK Terho EO Chronic cough and gastro-oesophageal reflux: a double-blind placebo-controlled study with omeprazole Eur Respir J 2000 16 633 8 11106204 10.1034/j.1399-3003.2000.16d11.x
Kollarik M Undem BJ Mechanisms of acid-induced activation of airway afferent nerve fibres in guinea-pig J Physiol 2002 543 591 600 12205192 10.1113/jphysiol.2002.022848
Niimi A Nguyen LT Usmani O Mann B Chung KF Reduced pH and chloride levels in exhaled breath condensate of patients with chronic cough Thorax 2004 59 608 12 15223872 10.1136/thx.2003.012906
Fox AJ Urban L Barnes PJ Dray A Effects of casazepine against capsaicin- and proton-evoked excitation of single airway c-fibres and vagus nerve from the guinea-pig Neurosci Lett 1995 15 421 28
Birring SS Prudon B Carr AJ Singh SJ Morgan MD Pavord ID Development of a symptom specific health status measure for patients with chronic cough: Leicester Cough Questionnaire (LCQ) Thorax 2003 58 339 43 12668799 10.1136/thorax.58.4.339
Hsu JY Stone RA Logan Sinclair RB Worsdell M Busst CM Chung KF Coughing frequency in patients with persistent cough: assessment using a 24 hour ambulatory recorder Eur Respir J 1994 7 1246 53 7925902 10.1183/09031936.94.07071246
Lalloo UG Fox AJ Belvisi MB Chung KF Barnes PJ Capsazepine inhibits cough induced by capsaicin and citric acid but not by hypertonic saline in guinea pigs J Appl Physiol 1995 79 1082 7 8567546
Vaughan J Ngamtrakulpanit L Pajewski TN Turner R Nguyen TA Smith A Urban P Hom S Gaston B Hunt J Exhaled breath condensate pH is a robust and reproducible assay of airway acidity Eur Respir J 2003 22 889 94 14680074
Kostikas K Papatheodorou G Ganas K Psathakis K Panagou P Loukides S pH in expired breath condensate of patients with inflammatory airway diseases Am J Respir Crit Care Med 2002 15 1364 70 12016097 10.1164/rccm.200111-068OC
Jack CI Calverley PM Donnelly RJ Tran J Russell G Hind CR Evans CC Simultaneous tracheal and oesophageal pH measurements in asthmatic patients with gastroesophageal reflux Thorax 1995 50 201 4 7701464
Yorulmaz I Ozlugedik S Kucuk B Gastroesophageal reflux disease: symptoms versus pH monitoring results Otolaryngol Head Neck Surg 2003 129 582 6 14595283 10.1016/S0194-5998(03)01585-7
Benini L Ferrari M Sembenini C Olivieri M Micciolo R Zuccali V Bulighin GM Fiorino F Ederle A Cascio VL Vantini I Cough threshold in reflux oesophagitis: influence of acid and of laryngeal and oesophageal damage Gut 2000 46 762 7 10807885 10.1136/gut.46.6.762
Nieto L de Diego A Perpina M Compte L Garrigues V Martinez E Ponce J Cough reflex testing with inhaled capsaicin in the study of chronic cough Respir Med 2003 97 393 400 12693800 10.1053/rmed.2002.1460
|
16270907
|
PMC1277833
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Cough. 2005 Oct 19; 1:10
|
utf-8
|
Cough
| 2,005 |
10.1186/1745-9974-1-10
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-221623616610.1186/1476-069X-4-22ResearchNeuropsychological effects of chronic low-dose exposure to polychlorinated biphenyls (PCBs): A cross-sectional study Peper Martin [email protected] Martin [email protected] Rudolf [email protected] Charité Universitätsmedizin Berlin, Institute of Pharmacology and Toxicology (CCM), Dorotheenstr. 94, 10117 Berlin, Germany2 Public Health Unit Rhein-Neckar-Kreis, State of Baden-Württemberg, Kurfürstenanlage 38-40, 69115 Heidelberg, Germany3 Present address: University of Freiburg, Department of Psychology, Belfortstr. 20, 79085 Freiburg, Germany2005 19 10 2005 4 22 22 18 3 2005 19 10 2005 Copyright © 2005 Peper et al; licensee BioMed Central Ltd.2005Peper 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
Exposure to indoor air of private or public buildings contaminated with polychlorinated biphenyls (PCBs) has raised health concerns in long-term users. This exploratory neuropsychological group study investigated the potential adverse effects of chronic low-dose exposure to specific air-borne low chlorinated PCBs on well-being and behavioral measures in adult humans.
Methods
Thirty employees exposed to indoor air contaminated with PCBs from elastic sealants in a school building were compared to 30 non-exposed controls matched for education and age, controlling for gender (age range 37–61 years). PCB exposure was verified by external exposure data and biological monitoring (PCB 28, 101, 138, 153, 180). Subjective complaints, learning and memory, executive function, and visual-spatial function was assessed by standardized neuropsychological testing. Since exposure status depended on the use of contaminated rooms, an objectively exposed subgroup (N = 16; PCB 28 = 0.20 μg/l; weighted exposure duration 17.9 ± 7 years) was identified and compared with 16 paired controls.
Results
Blood analyses indicated a moderate exposure effect size (d) relative to expected background exposure for total PCB (4.45 ± 2.44 μg/l; d = 0.4). A significant exposure effect was found for the low chlorinated PCBs 28 (0.28 ± 0.25 μg/l; d = 1.5) and 101 (0.07 ± 0.09 μg/l; d = 0.7). Although no neuropsychological effects exceeded the adjusted significance level, estimation statistics showed elevated effect sizes for several variables. The objectively exposed subgroup showed a trend towards increased subjective attentional and emotional complaints (tiredness and slowing of practical activities, emotional state) as well as attenuated attentional performance (response shifting and alertness in a cued reaction task).
Conclusion
Chronic inhalation of low chlorinated PCBs that involved elevated blood levels was associated with a subtle attenuation of emotional well-being and attentional function. Extended research is needed to replicate the potential long-term low PCB effects in a larger sample.
==== Body
Background
The neurobehavioral effects of polychlorinated biphenyls (PCBs) have been extensively studied in neonates and children [39,62,77]. However, no conclusive evidence is available on chronic nervous system effects in adult humans. The present neuropsychological group study explored the potential cognitive and affective consequences of long-term exposure to air-borne PCBs that were characterized by specific low chlorinated, ortho-substituted congeners. Effects sizes of behavioral and self-report measures were estimated to provide information that could be relevant for preparing extended epidemiological studies.
PCBs have been used as a component of insulation fluids, paints, and softening agents in lacquer, glues and sealing compounds. Low-level presence of PCBs has been discovered in many industrial settings in the USA and worldwide [14,57,58]. Due to the ubiquitous presence and poor degradation of PCBs, public health concerns continue to exist. Major exposure routes in humans include food intake, inhalation, and skin contact [59,72]. In particular, the indoor air of contaminated private or public buildings has been identified as a significant exposure source [29,65].
PCBs represent mixtures of up to 209 structurally related congeners differing by degree of chlorination which can be classified with respect to their similarity to 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) [38,40,58,59]. Several of the congeners most frequently detected in the US population such as IUPAC-Nos. 138, 153, 180, have been categorized as nondioxin-like. Certain mono-ortho-substituted congeners, among them PCB 105, 118, and 156, represent the most frequently detected congeners with aryl hydrocarbon receptor activity (weakly dioxin-like). Several non-ortho-substituted congeners such as PCB 77, 126, 169 have been characterized as dioxin-like [14,38].
Animal experiments have stressed the neurotoxic potency of PCBs [25,72]. The mechanisms of PCB neurotoxicity appear to include direct cerebral effects as well as indirect steroid- and thyroid-agonistic modulation [30]. Changes in several neurotransmitter systems involving dopamine- and serotonin-antagonistic effects have been reported [45,46,48]. Perinatal exposure to nonplanar PCBs was associated with dopamine-antagonistic effects, whereas exposure to coplanar PCBs showed dopamine-agonistic results [10,66].
There has been a growing interest in the neurodevelopmental toxicity of PCBs [e.g., [39,71]]. Among the brain regions that have been studied for perinatal exposure in rats, the striatum, prefrontal cortex and cerebellum showed neurodevelopmental effects which also depended upon age and sex [25,48]. Prenatal exposure to low concentrations of mono-ortho substituted or coplanar congeners showed reduced LTP in the hippocampus [50]. Despite the fact that mono-ortho substituted and nonplanar PCBs have lower TCDD-toxicity equivalents (TEQ), some studies ascribed a greater neurotoxic potency to these substances [27,67].
In human subjects, a considerable body of research has reported negative associations between prenatal PCB exposure and cognitive functioning and motor development in childhood [62,77]. However, the information available on long-term neurobehavioral consequences in adults is sparse and conclusive results are not yet available. For example, acute PCB intoxication by contaminated food was associated with subjective complaints such as fatigue, headache, dizziness, muscle weakness and memory and concentration problems [15,16,56]. Consumption of PCB contaminated fish was associated with memory and learning impairment [61].
Contaminated indoor air has also been identified as a significant source of chronic PCB exposure. Potential long-term health effects in school and office buildings where elastic sealants containing technical PCB mixtures were used have raised public health concerns [7,12,29,42,49,65]. The potential neuropsychological effects of such long-term inhalation remain unknown.
The present study was initiated after low chlorinated PCBs were detected in indoor air of three school buildings and verified by biological monitoring in employees of one of these schools [29]. The latter group was subjected to a health-screening program including neurobehavioral testing. Due to the lack of conclusive neurobehavioral results in adult humans, the purpose of this study is exploratory. Behavioral effects were expected for executive, that is, frontostriatal function being modulated by potential dopamine-antagonistic effects. Firstly, we tested the global hypothesis that there is a difference between exposed subjects and controls. Secondly, estimation statistics were computed to obtain effect size information that might be useful for risk assessment and for evaluating the reproducibility independent of sample size.
Methods
The present study was initiated after PCB-contaminated elastic sealant material was detected in a school building and indoor air concentrations of up to 10.655 ng/m3 were measured. The school was closed for renovation and employees were immediately submitted to a surveillance procedure that also included neuropsychological testing. All subjects underwent a medical examination including history of medical and psychosocial life events, environmental risk factors and dietary habits. An identical procedure was carried out in matched controls employed by an uncontaminated secondary school.
Study population
60 teachers and employees of two secondary modern schools were investigated. Thirty subjects were chronically exposed to air-borne PCBs in a school located in a rural region close to Heidelberg, Germany. This PCB group represents the total staff of this school. Thirty controls with no PCB exposure at work were drawn from another secondary school located in the city area of Heidelberg. The latter subjects were matched with the PCB group for education, age and professional status (see Table 1). The mean age was 49.2 years (SD = 7 years, range 37–61 years), with no differences between the PCB-group (48.2 years, SD = 7 years, range 39–60 years) and controls (49.9 years; SD = 7 years, range 37–61 years). The exposure group included 12 women whereas controls encompassed 18 women (χ2(1) = 1.67; p = .20). Gender differences of PCB levels that might be, for example, the result of excretion of PCBs during breast-feeding could not be confirmed [29]. Moreover, no substantial interactions of gender with exposure group were found for the present neurobehavioral variables. Nevertheless, the reported statistics given in the Tables were adjusted for the gender main effect to obtain unbiased information. Moreover, in a re-analysis of the data, an objectively exposed subgroup (> PCB 28 median 0.20 μg/l) was identified and compared with gender-matched controls.
Table 1 Demographic and exposure data of subjects exposed to PCB and of control subjects
PCB Controls
Mean SD Mean SD F[1,56] p
Gender [N; m/f] 18/12 12/18 1.671 0.20
Age [a] 48.2 7 49.9 7 0.89 0.35
Education [a] 2 12.5 2 12.4 2 0.03 0.86
Vocational index 2 5.9 0 5.8 0 0.39 0.54
Estimated intelligence [IQ] 2 117.3 5 117.3 4 0.00 0.99
Height [cm] 174 8 169 7 1.913 0.17
Weight [kg] 76 14 66 12 3.003 0.09
BMI [kg/m2] 24.7 3 23.5 3 2.273 0.14
Self-reported data
Alcohol consumption [g/week] 94.4 76 72.4 75 1.26 0.27
Nikotin consumption [cig./d] 6.4 10 5.7 10 0.07 0.79
Q16-Score 4 4.4 4 3.6 4 0.67 0.42
EQ Euroquest complaint score 4 146.4 32 131.6 31 3.29 0.07
Memory and attention 27.0 6 23.9 6 3.62 0.06
Drive and motivation 23.7 6 19.7 6 6.17 0.02 *
Tiredness 25.2 6 23.0 6 1.98 0.16
Emotional reactivity 20.0 6 17.7 6 2.41 0.13
Sensory complaints 8.1 3 7.4 3 0.65 0.42
Motor complaints 5.2 2 4.8 2 0.96 0.33
Cardiovascular complaints 14.8 5 12.0 5 4.76 0.03 *
Bowel and stomach 9.7 3 10.5 3 0.96 0.33
Head and neck 11.0 4 12.3 4 1.90 0.17
Stress inventory score 4 25.6 44 23.0 43 0.05 0.82
Exposure indices
PCB-28 [μg/l] 0.28 0.25 0.016 5 0.02 0.0001 ***
PCB-101 [μg/l] 0.07 0.09 0.01 5 0 0.0003 ***
PCB-138 [μg/l] 1.29 0.69 1.13 0.46 1.44 0.16
PCB-153 [μg/l] 1.68 0.96 1.56 0.58 1.07 0.29
PCB-180 [μg/l] 1.14 0.65 0.94 0.39 1.76 0.08
Total PCBs [μg/l] 4.45 2.44 3.65 1.40 1.87 0.067
Total occupational time [a] 20.9 6 22.0 9 1.09 0.28
Duration of exposure [a] 6 16.7 9 n.a.
Weighted duration of exposure [a] 6,7 10.5 6 n.a.
1χ2[1]; *: p < 0.05; ***: p < 0.001.
2 According to [76]
3 Statistically adjusted for gender, F[1,55]
4 A greater score corresponds to a greater number of complaints
5 Median and IQR/2 is given, 95 percent < measurement threshold
6 n.a.: not applicable
7 Considering mean time of presence at school.
The profile of vocational activities of the two populations of employees was comparable (number of occupational years: PCB: 20.9 ± 6 years, controls: 22.0 ± 9 years; weekly working hours at school: PCB: 24.6 ± 6 h; controls: 24.3 ± 9 h). The exposed group spent 4.2 ± 4 years of their vocational life outside and 16.7 ± 7 years within the contaminated school building (range 1–25 years). Assuming 40 weeks working time per year, the mean weighted exposure duration was 10.5 ± 6 years.
According to their history, laboratory tests and a medical examination, pathological conditions of the nervous system could be ruled out. The clinical interview and self-report questionnaire yielded no history of neurological or psychiatric disorders. This lack of psychiatric diagnoses may be accounted for by the fact that persons with active disease are not permitted to remain in employment as teachers.
77% of all subjects did not take any medication; 5% of the total group reported taking drugs for allergies, hypertonia or hypothyreosis, but no drugs with substantial cerebral side effects. Physical measures were within normal limits (mean normative T-values for the body mass index (BMI): PCB: 54.9 ± 3.5; controls: 52.0 ± 5.0; n.s.) and an increase or decrease of body weight was not reported. Alcohol and nicotine consumption was moderate and did not differ between groups (Table 1).
External exposure
Contamination by PCBs was determined by chemical analysis of indoor air and of elastic sealant materials. Air samples were collected during 24 h periods with closed doors and windows at a temperature of 20–22°C. External exposure was done by commercial institutes, analyzed according to standard procedures, and collected by the state Public Health Authority [for detailed data on indoor air PCB-concentrations of highly contaminated rooms, see [28,29]]. These analyses showed that the sealant material contained up to 50 percent of PCB. Indoor measurements revealed total airborne PCB concentrations of up to 17.460 ng/m3. Air concentrations in unrenovated rooms were between 2.870 ng/m3 and 10.655 ng/m3.
In order to exclude other possible sources of exposure to chlorinated pollutants, subjects were interviewed concerning nutritional factors and life style. No differences were evident concerning wood interiors potentially treated with preservatives (PCB 50%; controls 40%; χ2(1) = 0.27; n.s.), leather wear PCB (60%; 47%; χ2(1) = 0.67; n.s.), or daily consumption of meat products (53%; 30%, χ2(1) = 2.47; p = .12). Consumption of fish was more frequent in controls (20%; 57%, χ2(1) = 10.15; p = 0.001). No previous occupations were mentioned that might indicate exposure to other toxic substances. In the PCB group, three persons might occasionally have had contact with chlorinated compounds (lab technician, joiner, plumber). One control subject had worked in a brewery, another as a lab technician. The number of chemistry teachers was comparable in both groups (PCB 33%, controls 20%). The total frequency of these potential vocational risk factors was not significantly different (PCB: 37%; controls 20%; χ2(1) = 1.31; n.s.). The confounding effects of additional exposure sources appear to be irrelevant (Table 2).
Table 2 Correlations of external exposure indices and internal PCB-values in all subjects
PCB 28 PCB 101 PCB 138 PCB 153 PCB 180 Total PCB
Age1 0.17 0.17 0.27* 0.27* 0.26* 0.28*
BMI1 0.39** 0.26*
Hours of work/week1 0.63*** 0.51*** 0.19 0.21 0.25x
PCB years of exposure1 0.71*** 0.54*** 0.37** 0.32* 0.39** 0.43***
Weighted total exposure index1 0.71*** 0.59*** 0.39** 0.33** 0.40** 0.45***
Alcohol/week1 0.31* 0.27* 0.24x 0.22x
Cigarettes/day1 -0.20 -0.26* -0.17
Alternative vocational sources2 0.28* 0.18
Leather clothing2 0.24 0.24 0.33* 0.33*
Consumption of fish2 0.42*** 0.27*
Consumption of poultry2 0.19
Note. *: p < 0.05, **: p < 0.01, ***: p < 0.001 nominal α; only correlations with at least a small effect size are given; correlations with potential exposure sources such as indoor installations of chemically treated wood, frequent meat or milk consumption were trivial and are not presented; correlations with large effect sizes are printed in bold.
1 Spearman rank correlations
2 Cramer-V.
Biological Monitoring
Venous blood samples were drawn by the local Public Health Unit during a medical examination. Blood samples were analyzed by the state Public Health Authority using mass spectrometric gas chromatography (GC-MS) with standard protocols [see [29,65]]. Air-borne PCBs have previously been assessed by GC analyses of representative congeners such as PCB 28, 52, 101, 138, 153, and 180. These compounds have been used as markers of the specific exposure effects that can be traced back to polymer plasticizers used in Germany in the 1970's [3,29,37].
These congeners were analyzed according to the routine methods established by national authorities [37]. Although some authors recommended lipid standardization for the measurement of persistent lipophilic chemicals [e.g., [11]], a recent simulation study showed that PCB lipid standardization or the division of serum concentrations by serum lipids is potentially prone to bias [63]. Since group differences of serum lipids were not evident and because lipid adjustment is likely to produce spurious associations and biased results [63], unadjusted values were used.
A PCB sum value was computed except PCB 52 because quality assurance requirements failed for this congener [29]. A total toxicity index was not estimated because the TEQ concept is based on TCDD-toxicity equivalents mainly involving dioxin-like effects of coplanar PCBs. However, these congeners were not in the focus of the present study.
Blood sampling was carried out on average 4 weeks after the last exposure to contaminated air. The interval between first air sampling and blood sampling was 3 months. The interval between blood sampling and neuropsychological testing was 1 to 3 days. In addition to internal indicators, a weighted cumulative index was computed which indicated the total duration of exposure taking into account full- or part-time occupation and working days per year.
Neurobehavioral assessment
Standardized neuropsychological testing [33,44] was used to assess subtle subjective and behavioral changes. Tests selection was motivated by previous findings in humans [61] and experimental animals [25,48,50]. Most of the tests used have been recommended by the WHO due to their known sensitivity to neurotoxic compounds [2] and have been integrated into current neurotoxicity batteries [13]. The battery is only briefly summarized here because it has been described in previous work [54,55].
In addition, computerized testing of attention was implemented with the Test battery for Attentional Performance (TAP) [78]. The TAP has been established in the context of an EU Biomed project for the standardized assessment of attention disorders in brain damaged patients [80]; its subtests are equivalent to the reaction tasks of current computerized neurotoxicity batteries such as the Milan Automated Neurobehavioral System (MANS) [13].
All neuropsychological investigations were performed in the morning using an uncontaminated environment. The administration of the neuropsychological battery took about 90 min including a 10 min break. The order of tests was randomized across subjects except for memory tests that required a fixed retention interval.
Since neuropsychological measures are partly intercorrelated, explorative factor analyses (principal component analyses with Varimax rotation, using data from a pool of available control subjects, N = 72) were computed separately for behavioral and self-report variables. The Scree-test suggested 8 factors for behavioral measures (each explaining 8 to 15 percent of the variance, total 80 percent) and 5 factors for self-report measures (each explaining 15 to 30 percent of the variance, total 70 percent). The obtained factor structure was used to group the scores and to derive factor descriptions (using variables with loads >.50). Moreover, median effect sizes were computed for each factor and presented in the Tables.
Self-report measures
Subjective complaints and personality trait measures were organized in five clusters: current mood/emotional state, attention and motivation state, trait emotionality and health complaints, introversion, and sociability. Aggregated scores were also computed for each factor. In addition, psychosocial life stress was assessed [69] and weighted yielding a sum score for stressful events during the previous two years [36].
State descriptions of general physical well-being and mood
The questionnaire Q16 [35] is a well-known instrument for assessing neurotoxicity related symptom descriptions in solvent-exposed workers. Furthermore, a German version of a neurotoxicity symptoms questionnaire [17] was used to assess current complaints that are potentially related to neurotoxicity (items were aggregated according to the factorial structure of the Freiburger Beschwerdenliste (FBL-R) [23]).
Experience of attention and motivation state
State descriptions of attention and motivation as experienced in daily life were assessed by a 27-item Questionnaire of experienced deficits of attention (FEDA) [79], yielding scores for the factors motivation and drive, fatigue and slowing of practical activities, and distractibility of mental processes.
Trait measures of general physical well-being and emotional instability
The Freiburg Personality Inventory (FPI-R) [24] includes the scales impaired well-being (FPI Factor SI), aggressive arousability (FPI Factor SII), poor satisfaction with life (FPI 1), arousability (FPI 5), emotional stress (FPI 7), physical complaints (FPI 8), health worry (FPI 9), and emotional instability (FPI N). Depressed affect during the previous week was assessed by the Center for Epidemiological Studies Depression Scale (CES-D), German version (ADS) [34]. Other scales loading on this factor were General health complaints as assessed with the FBL [23] which contains 10 complaint item clusters such as the sum of bodily complaints (FBL11), general well-being and physical complaints (FBL1), emotional reactivity (FBL2), cardiovascular complaints (FBL3), bowels and stomach (FBL4), tension and strain (FBL6), sensory sensitiveness (FBL7), pain (FBL8), and skin problems and cold hands (FBL10).
Introversion
This factor included the scales introversion (FPIE), low aggression (FPI6), reserve and low openness (FPI10), as well as low achievement and work motivation (FPI3).
Sociability
This sociability/psychoticism factor included inhibition (FPI4), low social orientation (FPI2), motor restlessness (FBL9), and head-neck irritation (FBL5).
Behavioral tests
General intelligence
An estimation of present intelligence (IQ) as an overall measure of intellectual functioning was derived from the information, similarities, block design, and picture completion subtests of the Wechsler Adult Intelligence Scale (WAIS) [20,70].
Fluid intelligence
This factor included fluid intelligence measures related to verbal concept formation and reasoning processes (WAIS similarities, picture completion, and digit span forward).
Visuo-motor performance
Visuo-motor performance was assessed by the WAIS Block design subtest.
Concentration, alertness and speed
Selective attention and exploration speed was assessed with the Trail Making Test parts A and B (seconds) [44]. Alertness was measured with a simple and a cued reaction time task from the TAP [78]. Subjects were requested to respond whenever a cross appeared on the screen. In one condition, 40 visual stimuli were presented, each preceded by an acoustic warning stimulus. In the other condition, the cross appeared without warning. The difference between simple and cued reaction time was used as a measure of phasic alertness.
Working memory
This factor included the visual span forward and backward, verbal span backward subtests taken from the Wechsler Memory Scale-Revised (WMS-R) [32,74]. The digit symbol subtest from the WAIS was used to assess working memory, flexibility and speed [70]. Moreover, the TAP-subtest error scores for working memory, response shifting, and divided attention were associated with this factor [78]. The working memory subtest required a continuous control of the information flow through short-term memory. One-digit consecutively presented numbers had to be compared continuously with the preceding-but-one number (N-back task). In the response flexibility task, shifting of focused attention was tested by alternations between two sets of targets (letters or numbers) that were presented simultaneously and randomly, one on the left, the other on the right side of the fixation point. From one presentation to the next the target changed from letter to number and vice versa. The subject was requested to press the key on the side of the target (left or right). Divided attention was investigated with a dual task paradigm which was realized by a simultaneous visual/acoustic choice condition. A series of 75 matrices was presented on the screen, each for a duration of 3 s, with an inter-stimulus interval of 500 ms. A matrix consisted of a regular array of 4 × 4 dots with seven small 'x's superimposed randomly upon them. The subject was required to react whenever four 'x's formed a square. Simultaneously, the subjects listened to high and low pitched tones in regular alternation for a period of 5 min. Occasionally, a tone was followed by a tone of the same frequency that had to be detected.
Learning and memory
Verbal memory tests included the WMS-R subtest logical memory (immediate and delayed recall of stories). Visual memory scores were derived from the WMS-R visual reproductions (immediate and delayed recall of designs) [32]. Additionally, the Auditory Verbal Learning Test (AVLT) was used to assess free recall from verbal short-term and long-term memory [44].
Specific frontal lobe functions
Word fluency measures were obtained from the Regensburg Word Fluency Test (RWT) [1] and a design fluency task (production of non-recurrent figures) was added.
Psychomotor speed and attention
The alertness, working memory, response shifting, and divided attention subtests from the TAP [78] were used to assess simple and complex choice reaction time.
Statistical analysis
A traditional strategy for risk assessment in populations is to apply distribution-based statistics. However, it could also be useful to express exposure-related performance differences in a metric-free form [6]. Since the distribution-based null hypothesis testing approach (NHT) critically depends upon sample size, this does not provide information as to whether an effect is potentially replicable in larger study groups [31]. An estimation statistics approach is suitable to quantify group differences by means of the effect size d [18,31]. Therefore, a dual approach was applied in the present study [52,53]:
First, the question whether a behavioral PCB effect can be demonstrated was answered by NHT. ANOVAs with the factors exposure group and gender were computed to test the global hypothesis of μPCB<μCON. The results of the univariate F-tests for the exposure effect (with means adjusted for gender) are provided in the Tables. The α-level was set to p = .10 in order to control for the β-error (since it is inappropriate not to detect subtle differences at this stage of research [75]). Nominal α's are reported in the Tables. Since MANOVAs could not be computed due to insufficient samplesize [8], a revised Bonferoni α-adjustment was used for correction of dependencies [19]. Moreover, subjects with blood values above the PCB 28 median were compared with matched control subjects in a reanalysis of the data.
Second, estimation statistics were computed to evaluate which behavioral effects might be replicable independent of sample size [18]. Effect sizes of these differences were derived from η2 of the gender adjusted group effect (d1, corresponding to the above NHT approach). An additional effect size estimate was computed as the deviation of the empirical T-score of a behavioral variable from the distribution of the normative sample (d2, corresponding to the specific one-tailed hypothesis of μT(PCB)<50).
All reported d- and T-values were uniformly scaled so that elevated values of self-report-variables (d≥0.20, T>50) indicated elevated scores or complaints, whereas lower values in behavioral tests (d≤-0.20, T<50) indicated attenuated performance. An effect size for a behavioral measure was classified as meaningful in terms of potential reproducibility and marked with Δ in the Tables if both d1 as well as d2 showed the predicted attenuation at least to a moderate degree (e.g., d≤-0.20) [18].
To determine dose-response relationships, we computed correlations between exposure variables (total PCB, the marker congener PCB 28 [2,4,4'-trichlorobiphenyl], and weighted exposure duration) and age-adjusted variables in the exposed group. Positive correlations with exposure were expected for self-report variables and inverse associations were expected for behavioral variables. Depending on distribution characteristics, the results were verified using Spearman rank correlations. To control for spurious correlations and potential suppressor effects, the correlation analyses were done with and without partialling out specific confounders. Since self-reported complaints may be influenced by the subjects' openness (PCB subjects were more reserved, see Table 3) and by alcohol consumption (PCB subjects showed slightly elevated values compared to controls, see Table 1), these variables were considered in the analyses. Since behavioral test performance may be confounded by exposure-independent intelligence level, an estimate of this variable as well as alcohol consumption was considered. Analyses were done with MS-Excel, SPSS [68] and SAS for Windows [60].
Table 3 Mood, physical complaints and personality trait measures: primary and aggregated secondary factors in 30 PCB-exposed and 30 control subjects
PCB Controls Effect sizes
Mean SD Mean SD F[1,56]2 p2 d13 d23
Current emotional mood state4 0.415 0.365
EQ Tiredness/deactivation 54.3 13 50 10 2.06 0.16 0.38 0.36 Δ
EQ Emotional reactivity 53.6 11 50 10 2.41 0.13 0.41 0.33 Δ
EQ Low well-being 55.5 13 50 10 3.29 0.07 0.49 0.47 Δ
Attentional and motivational state 0.25 0.24
FEDA Poor motivation and drive 50.4 12 48.2 11 0.55 0.46
FEDA Fatigue and slowing 52.5 11 49.7 10 1.07 0.30 0.28 0.24 Δ
FEDA Distractability 55.0 11 48.8 10 5.26 0.03 0.61 0.48 Δ
General physical well being and emotional instability (trait) 0.02 0.21
FPI Factor SI/impaired well-being 48.1 6 50.0 6 1.40 0.24 -0.32 -0.23 Δ
FPI Factor SII/aggr. arousability 48.3 7 52.3 7 4.44 0.04 -0.57
FPI 1 Poor satisfaction with life 46.6 10 49.6 10 1.35 0.25 -0.31 -0.34 Δ
FPI 5 Arousability 50.5 12 51.8 12 0.16 0.69
FPI 7 Emotional stress 51.9 10 51.6 10 0.01 0.92
FPI 8 Physical complaints 46.2 9 47.3 9 0.23 0.64 -0.40
FPI 9 Health worry 47.7 9 51.5 9 2.39 0.13 -0.42 -0.23 Δ
FPI N Emotional instability 48.4 11 47.2 11 0.16 0.69
FBL11 Sum of bodily complaints 54.6 9 53.4 9 0.25 0.62 0.48
FBL 1 Well-being/phys. complaints 53.9 9 49.8 9 3.22 0.08 0.48 0.41 Δ
FBL 2 Emotional reactivity 55.0 10 51.9 9 1.65 0.20 0.34 0.51 Δ
FBL 3 Cardiovascular complaints 53.1 9 52.3 9 0.12 0.73 0.32
FBL 4 Bowels and stomach 53.8 8 53.9 8 0.00 0.96 0.41
FBL 6 Tension, strain 53.2 10 52.2 10 0.17 0.68 0.32
FBL 7 Sensory sensitiveness 56.0 9 55.2 9 0.13 0.72 0.63
FBL 8 Pain 52.3 10 51.9 9 0.03 0.87 0.23
FBL 10 Skin and cold hands 54.2 10 55.9 10 0.44 0.51 0.42
ADS Depressed affect 48.0 9 47.8 12 0.01 0.94
Introversion 0.45 0.35
FPI E Introversion 54.6 11 50.4 11 2.12 0.15 0.39 0.43 Δ
FPI 6 Low Aggressivity 54.7 8 49.2 8 6.42 0.01 0.68 0.51 Δ
FPI 10 Reserve/low openness 50.9 11 45.7 10 3.63 0.06 0.51
FPI 3 Low achievement/work motiv. 52.6 9 52.5 9 0.00 0.97 0.27
Sociability 0.38 0.34
FPI4 Inhibition 52.0 11 47.9 11 2.22 0.14 0.40
FPI 2 Low social orientation 42.9 8 47.5 8 4.45 0.04 -0.57 -0.77 Δ
FBL 9 Motor restlessness 54.7 9 51.4 9 1.84 0.18 0.36 0.48 Δ
FBL 5 Head-neck irritation 56.1 9 51.5 9 3.73 0.06 0.52 0.63 Δ
Aggregated secondary factors 0.30 0.26
Current mood/emotional state 53.6 8 50.4 8 2.25 0.14 0.40 0.40 Δ
Reduced attention and motivation 52.6 10 49.7 10 1.25 0.26 0.30 0.26 Δ
Low well being/trait emotionality 49.3 7 50.4 7 0.42 0.52
Introversion 53.2 6 49.5 6 5.08 0.03 0.61 0.38 Δ
Low Sociability 47.4 6 47.7 6 0.02 0.88 -0.31
1 T-Score; greater values correspond to elevated feature score
2 F-values for group with control of gender, nominal α
3 Effect sizes in comparison with controls (d1; with control of gender) and in comparison with normative sample (d2); Δ: at least moderate effect size of d1 and d2 suggesting potential reproducibility
4 T-values relative to controls
5 Effect size median.
Results
External exposure
External exposure measurements indicated that 5 rooms were contaminated with indoor air PCB values ranging from 1.587 to 10.655 ng/m3 (mean 7.749 ng/m3) [28,29]. The elastic sealant material was the primary source of exposure but walls and floors showed a similar PCB pattern. The lower chlorinated congeners 28 and 52 were responsible for about 90% of measured PCB marker congeners. The higher chlorinated and non-ortho-substituted or mono-ortho-substituted PCBs were of minor importance [29]. Figure 1 shows aggregated exposure measures for contaminated rooms indicating increased PCB values for the congeners 28, 52 and 101. The school was closed and renovated; follow up measurements in renovated rooms indicated that PCB contamination had fallen below 3.000 ng/m3.
Figure 1 PCB-concentrations of indoor air. Means of PCB measurements from three highly contaminated rooms (room 303, 407 and teacher's room) [from 29, p. 1058, Table 3, modified].
Internal exposure
Overall PCB exposure had a low to moderate effect size (d = 0.4 – 0.5) relative to expected background exposure values as derived from individual, age-group related median plasma PCB levels taken from national exposure data [5,37]. This was mainly due to low chlorinated PCBs (PCB 28: 0.28 ± 0.25 μg/l; PCB 101: 0.07 ± 0.09 μg/l), which are known to accumulate from respiratory rather than from nutritional sources (Figure 2). In contrast, exposure to the congeners PCB 138, 153 and 180 was not markedly elevated and most likely associated with general background exposure including food. More than 90 percent of control subjects showed PCB 28 and PCB 101 levels below detection threshold, whereas most of the exposed subjects showed detectable blood levels of these congeners (p < 0.001, Fisher's exact test; see Table 1). PCB28 values were above the range of controls in 53 percent of the PCB-exposed subjects corresponding to a large effect size (d = 1.5).
Figure 2 Blood values of PCB-exposed and control subjects. Blood values (means and standard deviations) of the PCB-exposed subjects and the control group (median PCB-28 and -101 levels were below 0.01 μg/l in controls). Expected values (background exposure) were estimated for each exposed subject from age-group related reference values [median PCB-plasma values for PCB-138, -153 and -180 according to 37].
Table 2 shows the relationship of internal and external indices. Blood PCB 28 and 101 were correlated with the cumulative index on an adjusted significance level. For PCB 138, 153 and 180, these correlations were not significant. Above-average fish consumption was also correlated with PCB 28. Therefore, we examined whether PCB 28 could have been modified by alternative sources of exposure. An explorative logistic regression model with stepwise variable selection was used which included demographic information as well as potential sources of exposure. A significant model for predicting blood values (which were dichotomized at the median of controls) could only be generated for PCB 28 (χ2(2) = 59.5; p < 0.0001; R2 = .84). The weighted exposure index significantly contributed to the explanation of variance (p = 0.005). Age, alcohol and nicotine, as well as fish consumption had no strong predictive value. A similar but weaker model was found for PCB 101 (χ2(1) = 19.6; p < 0.0001; R2 = .42).
Thus, nutritional and other exposure sources appeared to be of minor importance for predicting the present PCB 28 blood values. In contrast, all other congener values were associated with additional risk factors such as leatherwear, additional vocational sources or potentially contaminated food, etc. This confirms the decision to focus dose-response analyses on PCB 28 which was causally connected with the present elastic sealants material.
General health
General health of all subjects was satisfactory and no symptoms indicated acute or chronic PCB intoxication. The extent of complaints as assessed by self-report inventories did not exceed expected values. Diseases reported most frequently by both groups were allergies (42%), asthma and bronchitis (29%), and hypertonia (16%). No significant group or gender related differences were evident for complaints associated with the cardiovascular system (PCB: 10%; controls: 20%), skeletal motor-system (10%; 17%), respiratory system (17%; 17%), allergies (27%; 23%), thyroid dysfunction (7%; 3%), hepatitis (7%; 3%), diabetes (0%; 7%), or headache (0%; 7%). Neurological or psychiatric disorders could not be identified.
Neuropsychological results
Whereas the Q16 did not reveal elevated complaints, the Euroquest indicated a trend towards low motivation and cardiovascular problems (Table 1) as well as reduced well-being and distractibility of mental processes (Table 3). Moreover, estimation statistics yielded moderate effect sizes for distractibility, low level of well-being, head-neck pain syndrome, and for the personality trait variables of introversion, social orientation and low aggression.
The behavioral results indicated comparable intellectual functioning in both groups (Table 4). Both teachers groups showed normative values and effects sizes suggesting a high level of verbal intelligence (d2>1). Small, yet non-significant effects corresponding to hypothesis were found only for the TAP divided attention subtest. Moderate effect sizes were also observed for phasic alertness and Trails A. A general trend towards slightly increased reaction times in all of the computerized attention tasks in the exposed group must be noted which, however, were within the range of the normative population. Learning and memory performance of both groups was average. Inconsistent with hypothesis, the exposed group showed better immediate and delayed visual memory performance; however, this effect could not be replicated by the re-analyses summarized below.
Table 4 Neuropsychological results in 30 PCB-exposed and 30 control subjects
Raw Values T-Scores1
PCB Contr. PCB Contr. Effect sizes
Mean SD Mean SD Mean SD Mean SD F[1,56]2 p2 d13 d23
General intelligence -0.034 1.094
WAIS IQ 72.2 9 72.2 9 113.3 10 115.8 10 0.86 0.36 -0.25 1.09
WAIS General knowledge 18.2 3 17.7 3 58.2 9 58.8 9 0.07 0.79 0.97
WMS Digit span forward 9.7 2 9.2 2 60.0 10 57.7 10 0.79 0.38 0.24 0.99
WAIS Similarities 19.7 3 19.3 3 63.0 9 62.8 9 0.01 0.94 1.38
WAIS Picture completion 12.4 2 12.7 2 61.1 10 63.2 10 0.74 0.39 -0.23 1.15
Visuo-motor performance -0.21 0.47
WAIS Block design 21.9 6 22.5 6 53.6 7 55.0 7 0.63 0.43 -0.21 0.47
Concentration, alertness and speed -0.41 0.05
Trail making Test, part A 37.5 11 34.4 11 50.1 7 53.5 7 3.67 0.06 -0.51
Trail making Test, part B 79.3 27 74.6 27 53.7 7 55.6 7 1.11 0.30 -0.28 0.46
TAP Phasic alertness 0.024 0.1 0.058 0.1 45.9 12 50.8 12 2.34 0.13 -0.41 -0.38 Δ
Working memory -0.08 0.08
WMS Visual span forward 8.5 2 8.2 2 50.7 13 50.5 13 0.00 0.95
WMS Visual span backward 8.0 2 7.3 2 50.6 10 47.6 10 1.31 0.26 0.31
WMS Verbal span backward 7.2 2 7.4 2 50.2 11 52.8 11 0.84 0.36 -0.25
TAP Working memory/errors 4.9 4 6.3 4 52.35 11 505 10 2.08 0.16 0.39 0.23
TAP Response shifting/errors 6.3 4 6.0 4 47.55 14 505 10 0.09 0.76 -0.20
TAP Divided attention/errors 2.0 1 1.2 1 43.55 10 505 10 4.79 0.03 -0.58 -0.65 Δ
WAIS Digit symbol 53.3 11 54.0 11 56.3 9 57.1 9 0.12 0.73 0.61
Verbal memory 0.40 -0.83
WMS Logical memory recall 22.5 7 21.7 7 42.2 9 41.5 9 0.09 0.76 -0.85
WMS Logical memory delay 18.9 7 16.3 6 42.8 9 39.4 9 2.24 0.14 0.40 -0.83
AVLT Word list learning 75.8 11 70.8 11 53.8 9 49.6 9 3.08 0.08 0.47 0.29
Visual memory 0.93 0.83
WMS Visual memory recall 39.5 5 34.2 5 61.7 10 52.4 10 13.26 0.001 0.97 1.24
WMS Visual memory delay 34.2 9 27.1 9 56.3 12 46.2 12 11.09 0.002 0.89 0.43
Frontal lobe functions 0.24 0.42
RWT Word fluency 17.2 3 16.0 3 55.3 6 53.7 6 1.19 0.28 0.29 0.69
Design fluency 30.6 7 29.3 7 51.5 9 50.2 10 0.53 0.47
Psychomotor speed [TAP, 78] -0.39 -0.24
Simple reaction time (RT) 281.1 125 253.8 123 48.7 14 52.3 13 1.06 0.31 -0.27
RT with warning stimulus 275.5 120 236.2 118 46.6 13 51.2 12 2.08 0.15 -0.39 -0.24 Δ
RT working memory task 687.7 229 556.5 226 46.2 12 53.5 12 5.91 0.02 -0.65 -0.36 Δ
RT response shifting task 869.9 275 847.3 270 48.0 12 52.3 12 2.06 0.16 -0.38
RT divided attention task 702.5 96 665.8 94 42.8 10 46.8 10 2.56 0.12 -0.43 -0.69 Δ
1 Lower values indicate poorer performance
2 F-value for group effect adjusted for gender, nominal α
3 Effect sizes (>0.2): d1 relative to controls, adjusted for gender; d2 relative to the normative population; Δ: at least moderate attenuation of d1 and d2 suggesting potential reproducibility
4 Effect size median
5 Relative to controls.
Dose-response-relationships
Significant relationships of dose indicators (total PCB, PCB 28 and cumulative index) and response measures (self report or behavior) could not be demonstrated on an adjusted significance level. Self-reported complaints and mood state showed no substantial positive association with PCB or the cumulative index. For behavioral variables, however, several correlations were found for PCB 28. These correlations with figural fluency (r = -0.54; p < 0.01), simple reaction time (r = 0.31; p < 0.05), TAP response shifting errors (r = -0.31; p < 0.05), AVLT word list learning (r = -0.38; p < 0.05), and digit symbol (r = -0.32; p < 0.05) remained when rank correlations were computed or when estimated intelligence and alcohol were partialled out. Mood and personality variables showed no clear association with the behavioral data.
Re-analysis of subgroup with elevated PCB 28 blood levels
Since exposure status was variable due to different working habits in contaminated rooms, a re-analysis was done with objectively exposed subjects with PCB 28 levels ≥ 0.20 μg/l. This congener was chosen because it was significantly elevated in the present sample and correlated with the indoor air PCB burden. Two persons from the former exposure group were assigned to the control group because they had been working in the contaminated school only for a short while and showed no elevated blood values. This objectively exposed group (12 males, 4 females, 49.8 ± 6 years, weighted exposure duration 17.9 ± 7 years, range 4–25 years) and controls (12 males, 4 females, 48.6 ± 8 years) were matched for sex, age and education and were comparable with respect to physical characteristics, alcohol and smoking. Estimated intelligence (IQ 119.5 ± 5 and 118.0 ± 7; d = 0.24) was partialled out in the analyses of behavioral data.
PCB 28 levels of exposed subjects (median = 0.30 μg/l; range 0.20–1.05 μg/l) were above the distribution of controls (> 0.01 μg/l). Significant differences were also found for PCB 138 (1.453 ± 0.59 μg/l and 0.953 ± 0.37 μg/l; p = 0.01), PCB 153 (1.906 ± 0.85 μg/l and 1.272 ± 0.47 μg/l; p = 0.01) as well as PCB 180 (1.316 ± 0.69 μg/l and 0.725 ± 0.25 μg/l; p = 0.003).
The comparison of neuropsychological data did not show differences on an adjusted significance level. Nevertheless, when the effect sizes for the five self-report factors were inspected, the aggregated value for attention/motivation showed a medium effect (mean T = 54.6 ± 10 and T = 47.8 ± 11; d = 0.58) that was due to greater report of tiredness and slowing (T = 54.6 ± 10 and T = 45.9 ± 7; d = 0.70). The exposed subgroup also showed a trend towards more frequent reports of emotional reactions (mean T = 53.1 ± 11 and T = 46.8 ± 10; d = 0.46). Openness to answer questionnaires correctly was similar in both groups. Relative to the total PCB group, the PCB exposure subgroup described greater inattention, tiredness, distractibility as well as emotional and aggressive reactions. For behavioral measures, the subgroup comparisons showed relevant effect sizes only for attentional functions as indicated by TAP phasic alertness (T = 45.3 ± 12 and T = 48.1 ± 14; d = -0.32) and response shifting (T = 45.2 ± 15 and T = 50.7 ± 9; d = -0.40).
Re-analysis of subgroup with low PCB 28 blood levels
The subjects were aware of the fact that they had been exposed. The examiner was blind to the objective exposure status but not blind to the exposure site. Thus, the current study could only be performed in an open fashion. To test the hypothesis that prior information might have induced additional complaints or behavioral changes, 10 subjects with low objective PCB exposure (PCB 28≤0.1 μg/l) were identified and compared with matched controls. Although results were not significant on an adjusted level, there was a trend towards greater emotionality (T = 53.0 ± 9 and T = 43.5 ± 7; d = 1.1) and subjective distractibility (T = 59.9 ± 13 and T = 47.4 ± 5; d = 1.3) in the low PCB group. However, behavioral data showed no difference except of superior short term retention of visual designs (T = 62.5 ± 6 and T = 54.9 ± 8; d = 1).
Discussion
Chronic low-dose inhalation of polychlorinated hydrocarbons has repeatedly raised health concerns in subjects exposed in their everyday work environment or at home. This study contributes to our knowledge of potential neurobehavioral effects in the area of chronic exposure to low chlorinated PCB congeners in adult humans.
External and internal PCB exposure
Elevated PCB values were confirmed for low chlorinated congeners such as PCB 28, thus corresponding to previous reports [29,65]. This additional PCB burden relative to background exposure has been estimated to 2.8% [29]. This corresponds to an approximate PCB 28 exposure effect size of d = 0.34 or of d = 0.54 if PCB 52 is included. The present analysis yielded a large exposure effect size for PCB 28 (d>1) suggesting an acceptable discrimination from background exposure for this congener.
Although a significant increase was not observed for PCB 138, 153 and 180 relative to controls and to background exposure (Figure 2), significantly elevated levels were found for these congeners in the highly exposed subgroup. The fact that, for example, PCB 180 was found in the contaminated air (Figure 1) and was correlated with exposure duration but not with age could be a result of local low incorporation or of a slower rate of metabolization. The fact that PCB 180 was below background level in controls might have further added to this subgroup difference.
However, the measurement of low chlorinated compounds could be compromised by several factors. PCB 28, 52 and 101 show a relatively fast decomposition and half-life of about 60 days [29,64]. Due to the interval of 4 weeks between termination of exposure and blood sampling, these levels might have been underestimated. A decline of PCB blood values was observed in single subjects eight months after termination of exposure (data not presented).
Moreover, fluctuations of internal exposure due to factors such as individual differences in metabolism, room use and ventilation conditions may have further obscured dose-response relationships. Blood PCB values may also misrepresent the concentration in the brain where only about 10% of the blood PCB burden can be found [4]. Nevertheless, the fact that the cumulative exposure index successfully predicted internal PCB 28 suggests that the latter congener represents an appropriate marker of the present PCB burden.
PCB 28 is planar and mono-ortho-substituted but it can be regarded as nondioxin-like due to a low number of chlorine atoms. In contrast, neurobehavioral changes could also be mediated by dioxin-related toxicity effects of high-chlorinated and dioxin-like PCBs or by polychlorinated dibenzodioxins and -furans (PCDD/Fs) [10] which were not assessed in the present study. However, a pooled analysis of the current blood samples showed no significant increase of the planar PCB 77, 126 and 169 as well as dioxins [65] suggesting that our results are unlikely to be biased by dioxin-related toxicity.
Thus, a generalization of the present findings is limited to conditions of air-borne exposure to and confirmed incorporation of low to medium chlorinated PCBs. Since employees of other contaminated schools showed no elevated blood values for both low and high chlorinated congeners [12,22], an extrapolation to conditions without confirmed internal exposure does not seem warranted.
Subjective and behavioral effects of chronic exposure to low chlorinated PCBs
The unadjusted statistics given in the Tables must be cautiously interpreted due to alpha inflation and multiple testing. On an adjusted significance level of p = 0.004, the global null hypothesis of at least one group difference in self report or neurobehavioral variables could not be rejected because differences between exposed and non-exposed subjects were relatively small. Additional estimation statistics were used to detect subtle effects and to evaluate the reproducibility independent of sample size. Moderate effect sizes were found for distractibility, well-being, as well as for trait measures indicating low aggression and greater social orientation. The results of subgroup analyses confirmed a trend towards increased self-reported tiredness and slowing, and emotional reactions.
However, awareness of the current PCB exposure condition might have induced stress in terms of fear of being intoxicated, attention towards complaints, or general behavioral activation. It has been shown, for example, that perceived olfactory stimuli may influence well-being, emotional reactivity and cause health concerns [21,51]. Certain personality traits may sensitize for the consequences of exposure events [43]. Self-report data could thus be confounded by prior information about exposure and health concerns. In our subjects, however, personality disorders or chemical sensitivities could not be ascertained. Nevertheless, the degree of complaints was not correlated with PCB exposure level and subjects with low PCB exposure also reported elevated complaints. Therefore, the present self-reports are likely to have been biased by prior information about exposure.
Behavioral measures, in contrast, are less likely to be influenced by such information. Weak to moderate effect sizes were found for attention measures such as alertness and response shifting. Furthermore, moderate correlations between PCB and behavioral variables were found for figural fluency, response shifting and digit symbol. These findings deviate from prior reports of reduced learning and memory but not executive function [61]. A possible reason might be a greater exposure of fish eaters to high chlorinated, mono-ortho or coplanar congeners. Such an elevation of nutritional PCBs could not be confirmed for PCB or control subjects reporting greater consumption of fish.
Certain chlorinated hydrocarbons appear to alter frontostriatal function by dopamine depletion in animals. For example, nonplanar PCBs produced dopamine-antagonistic effects in the striatum and prefrontal cortex [10,46,48,66]. In human subjects, emotional regulation, motivation as well as alertness and response shifting rest upon and require contributions of the ventromedial and superior prefrontal cortex [26,41]. It is well accepted that mental flexibility, shifting the attentional focus and other executive functions are associated with dopamine-related frontostriatal activity in healthy individuals [73]. Conversely, disorders of frontal brain regions may affect attentional functions, alertness and working memory [47] and mood [9]. The present subjective and behavioral effects are therefore compatible with the hypothesis of a subtle attenuation of frontostriatal functions.
The fact that the low exposure subgroup showed a trend of greater visual memory (in addition to elevated emotionality discussed above) suggests that additional confounders independent of exposure might have moderated behavioral performance. Given that schools tend to differ with respect to educational programs, teachers may therefore show similar performance differences. Nevertheless, group differences in teaching preferences, medical and psychosocial events, additional environmental risk factors or dietary habits could not be substantiated.
Taken together, the present estimation statistics provided continuous indicators of exposure [6] yielding low response effects in a relatively small sample. Despite the limitations of this exploratory approach, these results suggest that findings may be replicable and should be replicated in the context of a more comprehensive epidemiological study. A sample size of approximately 200 to 500 exposed persons would be necessary to produce "significant" results according to conventional statistics. Moreover, we have argued elsewhere that even low behavioral effects may provide soft endpoints of neurotoxicity that could incur considerable economical costs (loss of work efficiency or motivation) when a large population is affected across an extended time span [53].
Conclusion
This exploratory neuropsychological group study showed that a discriminative, low chlorinated PCB marker congener typical for the present indoor air exposure condition could be identified. Although neurobehavioral effects could not be demonstrated by traditional significance testing, estimation statistics showed group differences with moderate effect sizes indicative of subjective attentional and emotional complaints as well as attenuated attention performance. Extended epidemiological research is needed to replicate and further substantiate the hypothesis of a subtle frontostriatal dysfunction in PCB exposed adults.
Abbreviations
AVLT Auditory Verbal Learning Test
BMI Body Mass Index
CES-D Center for Epidemiological Studies Depression Scale (Allgemeine Depressions-Skala, ADS)
d Effect size measure (weak: d<0.2; moderate: d = 0.5; strong: d>0.8)
Δ Potentially relevant effect sizes d>0.2
EQ Euroquest Symptom Questionnaire
ETA2 (η2); f Effect sizes derived from analysis of variance
FBL-R Freiburg Complaint Questionnaire Revised (Freiburger Beschwerdenliste)
FEDA Questionnaire of experienced deficits of attention (Fragebogen erlebter Defizite der Aufmerksamkeit)
FPI-R Freiburg Personality Inventory Revised (Freiburger Persönlichkeitsinventar)
MANS Milan Automated Neurobehavioral System
NHT Null hypothesis testing
PCBs Polychlorinated biphenyls
PCB 28 2,4,4'-Trichlorobiphenyl
PCB 52 2,2',5,5'-Tetrachlorobiphenyl
PCB 101 2,2'4,5,5'-Pentachlorobiphenyl
PCB 138 2,2',3,4,4',5'-Hexachlorobiphenyl
PCB 153 2,2',4,4',5,5'-Hexachlorobiphenyl
PCB 180 2,2',3,4,4',5,5'-Heptachlorobiphenyl
PCDD/F Polychlorinated dibenzodioxin/furan
Q16 Neurotoxicity Symptom Questionnaire
RT Reaction time
RWT Regensburg Word Fluency Test
T T-Score (normative values with mean = 50, standard deviation = 10)
TAP Test Battery for Attentional Performance
TCDD 2,3,7,8-Tetrachlorodibenzo-p-dioxin
TEQ Toxicity Equivalents
IQ Intelligence Score (values with mean = 100, standard deviation = 15)
WAIS Wechsler Adult Intelligence Scale
WMS-R Wechsler Memory Scale Revised
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MP provided all basic contributions to conception, design and methodology, performed the statistical analysis, wrote the draft version of the article and revised it critically for content. MK participated in the design and coordination, collected blood samples, provided internal monitoring and additional medical data, and corrected the final manuscript. RM participated in the maintenance of the study, revising the article critically for important intellectual content, and providing final approval. All authors read and approved the manuscript.
Acknowledgements
We acknowledge the co-operation of the Public Health Authority Stuttgart, State of Baden-Württemberg, Germany, for providing biological monitoring data. The present work was partially supported by the Public Health Unit, Heidelberg, and the Research Program in Neuropsychology and Neurolinguistics, University of Freiburg, funded by the Ministry of Science, State of Baden-Württemberg. Astrid Henke and Claudine Lautenschläger contributed to the present study by coordinating sessions and collecting neuropsychological data. This research was conducted in accordance with national and institutional guidelines for the protection of human subjects.
==== Refs
Aschenbrenner S Tucha O Lange KW Regensburger Wortflüssigkeits-Test (RWT) Göttingen, Hogrefe 2000
Baker EL Letz R Neurobehavioral testing in monitoring hazardous workplace exposure J Occ Med 1986 28 987 990 3534164
Ballschmiter K Zell M Analysis of polychlorinated biphenyls (PCBs) by glass capillary gas chromatography Fresenius' Zschr Analyt Chemie 1980 302 20 31 10.1007/BF00469758
Beck H Mathar W Palavinskas R Polychlorierte Dibenzodioxine, Dibenzofurane und Biphenyle in Organproben von Säuglingen und Kleinkindern Tätigkeitsbericht 1996 des Bundesinstituts für gesundheitlichen Verbraucherschutz und Veterinärmedizin München, MMV Medizin 1997 104 110
Becker K Kaus S Krause C Lepom P Schulz C Seiwert M Seifert B Umwelt-Survey Band III, Human-Biomonitoring, WaBoLu-Heft 01/02 1998 Berlin, Umweltbundesamt
Bellinger DC What is an adverse effect? A possible resolution of clinical and epidemiological perspectives on neurobehavioral toxicity Env Res 2004 95 394 405 15220073 10.1016/j.envres.2003.07.013
Benthe C Heinzow B Jessen H Mohr S Rotard W Polychlorinated biphenyls indoor air contamination due to thiokol-rubber sealants in an office building Chemosphere 1992 25 1481 1486 10.1016/0045-6535(92)90173-O
Bortz J Lienert GA Boehnke K Verteilungsfreie Methoden in der Biostatistik 1990 Heidelberg, Springer
Brody AL Barsom MW Bota RG Saxena S Prefrontal-subcortical and limbic circuit mediation of major depressive disorder Semin Clin Neuropsychiatry 2001 6 102 112 11296310 10.1053/scnp.2001.21837
Brouwer A Ahlborg UG Van den Berg M Birnbaum LS Boersma ER Bosveld B Denison MS Gray LE Hagmar L Holene E Huisman M Jacobson SW Jacobson JL Koopman-Esseboom C Koppe JG Kulig BM Morse DC Muckle G Peterson RE Sauer PJJ Seegal RF Prooije AESV Touwen BCL Weisglas-Kuperus N Winneke G Functional aspects of developmental toxicity of polyhalogenated aromatic hydrocarbons in experimental animals and human infants Europ J Pharmacol 1995 293 1 40 10.1016/0014-2999(95)90111-6
Brown JF JrLawton RW Polychlorinated biphenyl (PCB) partitioning between adipose tissue and serum Bull Environ Contam Toxicol 1984 33 277 280 6434008 10.1007/BF01625543
Burkhardt U Bork M Balfanz E Leidel J Innenraumbelastung durch polychlorierte Biphenyle (PCB) in dauerelastischen Dichtungsmassen Öff Gesundheitswes 1990 52 567 574
Cassitto MG Gilioli R Camerino D Experiences with the Milan Automated Neurobehavioral System (MANS) in occupational neurotoxic exposure Neurotoxicol Teratol 1989 11 571 574 2696877 10.1016/0892-0362(89)90039-1
Centers for Disease Control and Prevention (CDC) Second national report on human exposure to environmental chemicals 2003 Atlanta, Department of Health and Human Services, Centers for Disease Control and Prevention
Chia LG Chu FL Kuratsune M, Shapiro R Neurological studies on polychlorinated biphenyl (PCB)-poisoned patients PCB Poisoning in Japan and Taiwan 1984 New York, Liss 117 126
Chia LG Chu FL A clinical and electrophysiological study of patients with polychlorinated biphenyl poisoning J Neurol Neurosurg Psychiat 1985 48 894 901 2995592
Chouanière D Cassitto MG Spurgeon A Verdier A Gilioli R An international questionnaire to explore neurotoxic symptoms Env Res 1997 73 70 72 9311532 10.1006/enrs.1997.3701
Cohen J Statistical power analysis for the behavioral sciences 1988 New York, Academic Press
Cross EM Chaffin WW Use of the binomial theorem in interpreting results of multiple tests of significance Educ Psychol Meas 1982 42 25 34
Dahl G WIP – Reduzierter Wechsler Intelligenztest Meisenheim am Glan, Hain 1972, 1986
Dalton P Wysocki CJ Brody MJ Lawley HJ The influence of cognitive bias on the perceived odor, irritation and health symptoms from chemical exposure Int Arch Occ Env Health 1997 69 407 417 10.1007/s004200050168
Ewers U Wittsiepe J Barth G Bork M Kaesler C Leidel J Strobel J Blutuntersuchungen auf PCB bei Lehrerinnen und Lehrern einer stark PCB-belasteten Schule Öff Gesundheitswes 1998 60 357 362 9697359
Fahrenberg J Die Freiburger Beschwerdenliste (FBL-R) Göttingen, Hogrefe 1994
Fahrenberg J Hampel R Selg H Das Freiburger Persönlichkeitsinventar-Revised (7 ed) Göttingen, Hogrefe 2001
Faroon O Jones D de Rosa C Effects of polychlorinated biphenyls on the nervous system Toxicol Ind Health 2001 16 305 333 11693948
Fellows LK Farah MJ Ventromedial frontal cortex mediates affective shifting in humans: evidence from a reversal learning paradigm Brain 2003 126 1830 1837 12821528 10.1093/brain/awg180
Fischer LJ Seegal RF Ganey PE Pessah IN Kodavanti PR Symposium overview: toxicity of non-coplanar PCBs Toxicol Sci 1998 41 49 61 9520341 10.1006/toxs.1997.2386
Gabrio T Piechotowski I Wallenhorst T Klett M Cott L Friebel P Schwenk M PCB-blood levels in teachers, working in PCB-contaminated schools Organohalog Comp 1998 38 9 12
Gabrio T Piechotowski I Wallenhorst T Klett M Cott L Friebel P Link B Schwenk M PCB-blood levels in teachers, working in PCB-contaminated schools Chemosphere 2000 40 1055 1062 10739046 10.1016/S0045-6535(99)00353-7
Golden RJ Noller KL Titus-Ernstoff L Kaufman RH Mittendorf R Stillman R Reese EA Environmental endocrine modulators and human health: an assessment of the biological evidence Crit Rev Toxicol 1998 28 109 227 9557209 10.1080/10408449891344191
Greenwald AG Gonzalez R Harris RJ Guthrie D Effect sizes and p values: what should be reported and what should be replicated? Psychophysiology 1996 33 175 183 8851245
Härting C Markowitsch HJ Neufeld H Calabrese P Deisinger K Kessler J Wechsler Gedächtnis Test – Revidierte Fassung (WMS-R) Göttingen, Hogrefe 2000
Hartman DE Neuropsychological toxicology 1995 2 New York, Plenum
Hautzinger M Bailer M Allgemeine Depressions Skala ADS 1993 Weinheim, Beltz
Hogstedt C Andersson K Hane M Aitio A, Riihimäki V, Vainio H A questionnaire approach to the monitoring of early disturbances in central nervous functions Biological monitoring and surveillance of workers exposed to chemicals 1984 Washington DC, Hemisphere Publishers
Holmes TH Rahe RH The social readjustment scale J Psychosom Res 1967 11 213 218 6059863 10.1016/0022-3999(67)90010-4
Human Biomonitoring Commission of the National Environmental Authority (Ed.) Stoffmonographie PCB – Referenzwerte für Blut. Umweltbundesamt Berlin Bundesgesundheitsbl Gesundheitsforsch Gesundheitsschutz 1999 42 511 521 10.1007/s001030050147
Hutzinger O Safe S Zitko V The chemistry of PCBs 1974 Cleveland Ohio, CRC-Press
Jacobson JL Jacobson SW Evidence for PCBs as neurodevelopmental toxicants in humans Neurotoxicol 1997 18 415 424 9291491
Kafafi SA Afeefy HY Ali AH Said HK Kafafi AG Binding of polychlorinated biphenyls to the aryl hydrocarbon receptor Environ Health Perspect 1993 101 422 428 8119253
Konishi S Jimura K Asari T Miyashita Y Transient activation of superior prefrontal cortex during inhibition of cognitive set J Neurosci 2003 23 7776 7782 12944506
Köppl B Piloty M PCB in Dichtungsmassen – Erfahrungen und Ergebnisse bei Vorgehen in Berlin und Sanierung einer Schule (Teil II) Öff Gesundheitswes 1993 55 629 634 8111159
Labarge AS McCaffrey RJ Multiple chemical sensitivity: a review of the theoretical and research literature Neuropsychol Rev 2000 10 183 211 11132100 10.1023/A:1026460726965
Lezak MD Neuropsychological assessment 1995 3 Oxford, Oxford University Press
Mariussen E Morch Andersen J Fonnum F The effect of polychlorinated biphenyls on the uptake of dopamine and other neurotransmitters into rat brain synaptic vesicles Toxicol Appl Pharmacol 1999 161 274 282 10620485 10.1006/taap.1999.8806
Mariussen E Andersson PL Tysklind M Fonnum F Effect of polychlorinated biphenyls on the uptake of dopamine into rat brain synaptic vesicles: a structure-activity study Toxicol Appl Pharmacol 2001 175 176 183 11543650 10.1006/taap.2001.9231
McDonald BC Flashman LA Saykin AJ Executive dysfunction following traumatic brain injury: neural substrates and treatment strategies Neurorehabilitation 2002 17 333 344 12547981
Morse DC Seegal RF Borsch KO Brouwer A Long-term alterations in regional brain serotonin metabolism following maternal polychlorinated biphenyl exposure in the rat Neurotoxicol 1996 17 631 638 9086483
Neisel F von Manikowsky S Schumann M Feindt W Hoppe HW Melchiors U Humanes Biomonitoring auf Polychlorierte Biphenyle bei 130 in einer Grundschule exponierten Personen Gesundheitswesen 1999 61 137 149 10226386
Niemi WD Audi J Bush B Carpenter DO PCBs reduce long-term potentiation in the CA1 region of rat hippocampus Exp Neurol 1998 151 26 34 9582252 10.1006/exnr.1998.6793
Orbaek P Österberg K Akesson B Bergendorf U Karlson B Seger L Suprathreshold intensity and annoyance reactions in experimental challenge to toluene and n-butyl acetate among subjects with long-term solvent exposure Scand J Work Env Health 1998 24 432 438 9869316
Peper M Untersuchungen zur neurobehavioralen Toxizität polychlorierter Kohlenwasserstoffe nach chronischer Exposition im Niedrigdosisbereich 2004 Medical Faculty of the Charité Universitätsmedizin Berlin, Dissertation
Peper M Umweltbezogene Verhaltensstörungen 2004 Landsberg: Ecomed Medizin
Peper M Ertl M Gerhard I Long-term exposure to wood-preserving chemicals containing pentachlorophenol and lindane is related to neurobehavioral performance in women Am J Ind Med 1999 35 632 642 10332516 10.1002/(SICI)1097-0274(199906)35:6<632::AID-AJIM10>3.0.CO;2-R
Peper M Klett M Frentzel-Beyme R Heller WD Neuropsychological effects of chronic exposure to environmental dioxins and furans Env Res 1993 60 124 135 8432263 10.1006/enrs.1993.1021
Rogan WJ Gladen BC Neurotoxicology of PCBs and related compounds Neurotoxicol 1992 13 27 35 1508429
Ross G The public health implications of polychlorinated biphenyls (PCBs) in the environment Ecotoxicol Environ Saf 2004 59 275 291 15388267 10.1016/j.ecoenv.2004.06.003
Safe S Polychlorinated biphenyls (PCBs): environmental impact, biochemical and toxic responses, and implications for risk assessment Crit Rev Toxicol 1994 24 87 149 8037844
Safe S Goldstein JA Kimbrough RD, Jensen A Mechanisms of action and structure-activity relationships for the chlorinated dibenzo-p-dioxins and related compounds Halogenated biphenyls, terphenyls, naphtalenes, dibenzodioxins and related products 1989 2 Amsterdam, Elsevier 239 294
SAS Institute Inc The SAS System for Windows Software Release 612 1999 Cary, North Carolina, SAS Institute Inc
Schantz SL Gasior DM Polverejan E McCaffrey RJ Sweeney AM Humphrey HEB Gardiner JC Impairments of memory and learning in older adults exposed to polychlorinated biphenyls via consumption of Great Lakes fish Environ Health Perspect 2001 109 605 611 11445515
Schantz SL Widholm JJ Rice DC Effects of PCB exposure on neuropsychological function in children Environ Health Perspect 2003 111 357 376 12611666
Schisterman EF Whitcomb BW Louis GM Louis TA Lipid adjustment in the analysis of environmental contaminants and human health risks Environ Health Perspect 2005 113 853 857 16002372
Schrenk D Beyer A, Eis D Polychlorierte Biphenyle (PCB) Praktische Umweltmedizin 1994 Heidelberg, Springer
Schwenk M Gabrio T Päpke O Wallenhorst T Human biomonitoring of polychlorinated biphenyls and polychlorinated dibenzodioxins and dibenzofuranes in teachers working in a PCB-contaminated school Chemosphere 2002 47 229 233 11993638 10.1016/S0045-6535(01)00307-1
Seegal RF Brosch KO Okoniewski RJ Effects of in utero and lactational exposure of the laboratory rat to 2,4,2',4'- and 3,4,3',4'-tetrachlorobiphenyl on dopamine function Toxicol Appl Pharmacol 1997 146 95 103 9299601 10.1006/taap.1997.8226
Shain W Bush B Seegal R Neurotoxicity of polychlorinated biphenyls: structure-activity relationship of individual congeners Toxicol Appl Pharmacol 1991 111 33 42 1949034 10.1016/0041-008X(91)90131-W
SPSS Inc Statistical Package for the Social Sciences 110 2000 Chicago, Illinois, SPSS Inc
Sulz SKD Das Verhaltensdiagnostiksystem VDS 1991 München, CIP-Medien
Tewes U Hamburg-Wechsler-Intelligenztest für Erwachsene – Revision Göttingen, Hogrefe 1991
Tilson HA Jacobson JL Rogan WJ Polychlorinated biphenyls and the developmental nervous system: cross-species comparison Neurotoxicol Teratol 1990 12 239 248 2115098 10.1016/0892-0362(90)90095-T
Tilson HA Kodavanti PR The neurotoxicity of polychlorinated biphenyls Neurotoxicol 1998 19 517 525 9745906
Volkow ND Gur RC Wang GJ Fowler JS Moberg PJ Ding YS Hitzemann R Smith G Logan J Association between decline in brain dopamine activity with age and cognitive and motor impairment in healthy individuals Am J Psychiat 1998 155 344 349 9501743
Wechsler D A standardized memory scale for clinical use J Psychol 1945 19 37 95
Willmes K Sturm W, Herrmann M, Wallesch CW Statistische und psychometrische Aspekte in der Neuropsychologie Lehrbuch der Klinischen Neuropsychologie 2000 Lisse NL, Swets & Zeitlinger 229 249
Wilson RS Rosenbaum G Brown G The problem of premorbid intelligence in neuropsychological assessment J Clin Neuropsychol 1979 1 49 93
Winneke G Walkowiak J Lilienthal H PCB-induced neurodevelopmental toxicity in human infants and its potential mediation by endocrine dysfunction Toxicol 2002 181–182 161 165 10.1016/S0300-483X(02)00274-3
Zimmermann P Fimm B Testbatterie zur Aufmerksamkeitsprüfung (TAP) Herzogenrath, Psytest 1993, 2000
Zimmermann P Poser U Ein Fragebogen erlebter Defizite der Aufmerksamkeit 1996 Universität Freiburg: Institut für Psychologie
Zoccolotti P Matano A Deloche G Cantagallo A Passadori A Leclercq M Braga L Cremel N Pittau P Renom M Rousseaux M Truche A Fimm B Zimmermann P Patterns of attentional impairment following closed head injury: A collaborative European study Cortex 2000 36 93 107 10728900
|
16236166
|
PMC1277834
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Oct 19; 4:22
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-22
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-231624204110.1186/1476-069X-4-23ResearchSmall business owners' health and safety intentions: A cross-sectional survey Brosseau Lisa M [email protected] Shelby Yahui [email protected] Division of Environmental Health Sciences, School of Public Health, University of Minnesota, 420 Delaware St SE, Minneapolis, Minnesota, 55455, USA2 Cardiac Rhythm Management, Medtronic, Inc., 1015 Gramsie Road, Shoreview, Minnesota, 55126, USA2005 21 10 2005 4 23 23 8 6 2005 21 10 2005 Copyright © 2005 Brosseau and Li; licensee BioMed Central Ltd.2005Brosseau and Li; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Little is known about the variables underlying small business owners' behavioural intentions toward workplace health and safety. This project explores the relationship between three mediating variables (Attitude Toward Safety, Subjective Norm and Perceived Behavioural Control) and owners' Intentions Toward Safety, following the Theory of Planned Behaviour. We also investigate the role of beliefs underlying each mediating variable.
Methods
Seven hundred businesses (5–50 employees) were randomly selected from 4084 eligible companies in a manufacturing business database (SIC codes 24 to 39). The 348 respondents are on average 51 yrs of age, 86% male, 96% white and have 2 to 4 years of post-secondary school.
Results
All three mediator variables are significantly correlated with Intentions Toward Safety; Attitude Toward Safety shows the strongest correlation, which is confirmed by path analysis. Owners with higher attitudes toward safety have a higher probability of believing that improving workplace health and safety will make employees' healthier and happier, show that they care, increase employee productivity, lower workers' compensation costs, increase product quality and lower costs.
Conclusion
These results suggest that interventions aimed at increasing owners' health and safety intentions (and thus, behaviours) should focus on demonstrating positive employee health and product quality outcomes.
==== Body
Background
Small businesses are an important sector of the United States economy. Nearly 98% of the 5.7 million U.S. businesses have fewer than 100 employees and account for 36% of all employment [1]. Ninety-three percent of the approximately 305,000 U.S. manufacturing firms have fewer than 100 employees [2].
Employees in small and medium-sized manufacturing businesses experience higher levels of work-related injuries and illnesses than employees in large businesses. For example, the 2003 incidence rates for non-fatal injuries and illnesses in the United States were highest in businesses with 50 to 249 employees in all economic sectors. In the manufacturing sector, the highest rates of non-fatal injuries and illnesses are experienced in durable goods manufacturing businesses with 11 to 249 employees [3].
Data on owners' behaviours toward health and safety in small businesses are limited and conflicting. In interviews, owners describe numerous barriers including limited resources, lack of in-house expertise, and production pressures [4,5]. One study found that insurers, quality assurance programs and regulatory agencies are important incentives to improve health and safety [5], while another found that owners do not trust government agencies or consultants and seek input on environmental improvements only from suppliers, other owners and customers [6]. Champoux and Brun found that most small business owners do not think that resources are a significant barrier to their improving health and safety. Only 37% of 223 owners of small businesses (fewer than 50 employees) thought cost was an important barrier to health and safety improvements [7].
Eakin et al. describe additional barriers, including owners' limited perspective about what can be accomplished, limited unionization and informal management structures [4]. Numerous investigators have shown that the incidence of injuries is lower in businesses that actively engage employees in decision making and joint labour-management safety committees [8-11]. However, Champoux and Brun found that fewer than 5% of small businesses had participatory safety programs [7].
A few investigators have studied the effectiveness of health and safety interventions in large and small businesses using randomized, controlled trials [5,12-14], but have been largely unsuccessful at bringing about significant changes in workplace health and safety. It has been conjectured that intervention activities may require more focused efforts targeted at motivating business owners to make improvements, in addition to changing the behaviour of employees [12,14].
This project was prompted by the need for more information about the factors that influence small business owners' intentions toward workplace health and safety. We selected the Theory of Planned Behaviour [15] to guide the development of a survey of owners' intentions. Behavioural intentions are measured as a surrogate for actual behaviour, which must be defined in terms of action, target, context and time. Three mediating variables – Attitude Toward the Behaviour, Subjective Norm and Perceived Behavioural Control – combine to influence a person's intentions toward the target behaviour. Three types of beliefs may indirectly contribute to Behavioural Intention: Outcome Beliefs, Normative Beliefs and Control Beliefs.
We used a modification (Figure 1) of the original model, eliminating the measurement of belief strength for each of the three belief constructs, because initial testing indicated that respondents did not understand the difference between these constructs. The specific behaviour is defined as: "In the next six months [time], how likely is it you will improve [action] workplace health and safety [target] in your business [context]."
Figure 1 Theory of planned behaviour.
The goals of this project are to 1) evaluate the association of the three mediating variables with owners' health and safety intentions and 2) investigate the role various beliefs play in formulating intentions. Insights from these relationships are expected to lead to the development of more effective interventions.
Methods
Survey Design
Following a protocol developed by Ajzen and Fishbein [16], we conducted open-ended telephone interviews with 16 small business owners in a variety of industries. Owners were asked to identify outcomes of their behaviour (what happens when you work on health and safety?), who influences the behaviour (who encourages or discourages you to work on health and safety?), and barriers and supports for the behaviour (what helps or hinders you to work on health and safety?). Answers to these three questions were categorized and the most frequent responses were used to develop survey items for specific outcome, normative and control beliefs, respectively.
A draft survey was developed and reviewed for face validity by experts in health and safety and small business assistance. Two pilot studies were conducted, each of which involved initial and follow-up (with a $2 incentive) mailings to 120 independent manufacturing businesses with 5 to 50 employees in Minnesota. The data from these surveys received a preliminary analysis to determine if any changes in survey design were necessary. Only minor changes in wording and organization were made before the main study.
In the main study, the survey was mailed to 700 owners of small manufacturing businesses in Minnesota, with a second mailing, including a $2 incentive, three weeks later.
Study Population
Eligible businesses were drawn from the Manufacturers' News, Inc. manufacturing businesses database (2001) for Minnesota, which included all businesses in SIC codes 24 to 39. The final set included 4084 businesses with 5 to 50 employees established after 1998 (in the database at least 3 years) without parent companies (independent businesses). Seven hundred businesses were drawn randomly from the set; respondents to the pilot study surveys were not included in the main trial.
In all, there were 348 respondents, a 49.7% response rate. Respondents are on average 51 yrs of age (standard deviation = 10.22), 86% male, and 96% white. They have a mean maximum education level between 2 and 4 years of post-secondary school. On average, businesses have been in operation 30 years; respondents have worked in the industry for 23 years and owned their business 16 years. Companies have a mean of 14 production and 22 total employees. Sixty-six percent indicate they are the president, 49% are owners, and 22% are managers (respondents could indicate one or more positions).
Survey Content
Each variable is measured by one or more items in the survey (Table 1). The score for a question is the sum of scores for each item in the question. Questions about intentions, attitudes, subjective norm, outcome beliefs and normative beliefs were operationalised as described by Ajzen and Fishbein [16]. Questions about behavioural control and control beliefs used a format described by Fishbein [15].
Table 1 Description of variables and basic statistics
Variable Name Number of Items Mean* (Standard Deviation) Range* N
(Y) Intentions Toward Safety 9 19 (8.2) 9–45 330
(Z1) Attitude Toward Safety 3 6.4 (2.5) 3–15 326
(Z2) Subjective Norm 1 1.9 (0.8) 1–5 345
(Z3) Perceived Behavioural Control 1 2.6 (1.0) 1–5 344
(X1) Outcome Beliefs 11 32 (5.4) 15–52 339
(X2) Normative Beliefs 5 16 (4.5) 5–25 343
(X3) Control Beliefs 4 9.5 (3.0) 4–17 344
* Lower scores correspond to more positive responses.
Intentions
Intentions Toward Safety is measured by the sum of nine questions asking "In the next six months, how likely is it you will [take a specific action]?" (measured on a 5-pt scale ranging from 1 = "very likely" to 5 = "very unlikely"). The specific actions were:
• Talk to employees about health and safety rules
• Reward employees for following safe work rules
• Wear safety equipment when enter the work area
• Walk through business and identify safety hazards
• Check that employees are wearing safety equipment
• Make sure access is clear to exits and extinguishers
• Talk to employees about the hazards of their job
• Train employees to handle emergencies
• Ask employees for recommendations on safer ways to do their work
In many cases, complex behaviours cannot be measured by a single action (e.g. weight loss involves both dieting and exercise behaviours). In such cases, a behavioural index comprised of several individual behaviours will provide a better measure [16]. Improving workplace health and safety is certainly a complex behaviour, but there is no accepted set of actions that define such behaviour for a business owner. Some health and safety behaviours are one-time, programmatic actions, while others are activities that must take place on a regular basis.
We focused on health and safety activities that owners should conduct on a regular basis, but did not specifically define "regularly," because it will vary by activity, industry, number of employees, etc. We did not ask about specific programs or hazards, because owners were located in a broad range of manufacturing industries.
The topic of health and safety behaviour has received very little formal research or validation. We drew our list of actions from those developed by practitioners and regulators to describe "good" health and safety. The best examples of these are found in recognition programs, such as the United States Occupational Safety and Health Administration's Voluntary Protection Program (VPP) [17] and in health and safety management systems proposed by individuals and organizations [18-20]. Several occupational health professionals with small business experience and a range of perspectives (e.g. consulting, business assistance, regulatory) were asked to review this set of actions for face validity.
Attitude
Attitude Toward Safety is measured by the sum of three questions about the importance, necessity and convenience of improving health and safety in the next six months (scored on a 5-pt scale; for example: 1 = "very convenient"; "somewhat convenient"; "neither"; "somewhat inconvenient"; 5 = "very inconvenient").
With input from small business owners and health and safety professionals, we selected these three as the most relevant to health and safety behaviour using guidelines developed by Osgood et al. [21].
Subjective Norm
Subjective Norm is measured by agreement (1 = strongly agree; 5 = strongly disagree) with "Most people important to me think I should improve health and safety in my business in the next six months."
Perceived Behavioural Control
Perceived Behavioural Control is measured by how easy it will be for owners to improve health and safety in their business in the next six months (1 = very easy; 5 = very difficult).
Outcome Beliefs
Outcome Beliefs are measured by summing answers to questions about the likelihood of eleven outcomes (5-pt scale; 1 = "very likely" to 5 = "very unlikely" scale):
1. Make employees happier
2. Make employees healthier
3. Increase costs
4. Increase employees' productivity
5. Cause employees to complain
6. Show that I care about employees
7. Cut into profits
8. Lower workers' compensation costs
9. Take too much time
10. Increase the quality of products
11. Lower the business' productivity
These eleven outcome beliefs were identified from the most frequent responses in our open-ended interviews with small business owners. Ajzen and Fishbein discourage the selection of outcome beliefs not derived from interviews with the target population [16].
Normative Beliefs
Normative Beliefs are measured by summing the responses to questions about the influence of five groups on owners' behaviour (5-pt scale, 1 = "strongly agree" to 5 = "strongly disagree"):
1. My employees think I should improve safety in my business.
2. My workers' compensation company...
3. Government agencies...
4. My customers...
5. My vendors or suppliers...
As with outcome beliefs, these five normative beliefs were selected from the most frequent responses of interviewed owners.
Control Beliefs
Control Beliefs are measured by summing responses to four questions (5-pt scale; 1 = "strongly agree" to 5 = "strongly disagree"):
1. I have enough resources available for improving safety in my business.
2. I am well-informed about how to improve safety in my business.
3. My employees are supportive of my efforts to improve safety in my business.
4. I have enough time to improve safety in my shop.
These four beliefs were derived from the most frequent responses obtained in owner interviews.
Results
Correlation analysis, linear and logistic regression and structural equation modelling techniques were used to explore relationships among the variables of interest. Negative outcome beliefs (#3, 5, 7, 9 and 11 shown above) were re-coded to ensure that all outcome beliefs were evaluated in the positive direction.
Pearson's correlation analysis was first applied to examine the relationship between the response variable, Intentions Toward Safety, and the three main covariates: Attitude toward Safety, Subjective Norm, and Perceived Behavioural Control (Table 2). The response variable was transformed to the log scale to satisfy the normality assumption. All three variables were significantly correlated with Intentions Toward Safety; correlation with Attitude Toward Safety was the strongest, ρ = 0.56 (p < 0.0001).
Table 2 Correlation between the response variable (Log(Intentions Toward Safety)) and covariates*
Variable Correlation Coefficient p-value N
Attitude Toward Safety 0.559 <0.0001 326
Subjective Norm 0.370 <0.0001 345
Perceived Behavioural Control 0.297 <0.0001 344
* The variable "Intentions Toward Safety" is in the log scale to satisfy the normality assumption.
We also explored correlations between the covariates (Table 3). The correlations between Attitude Toward Safety vs. Outcome Beliefs and between Perceived Behavioural Control vs. Control Beliefs are moderate (ρ = 0.46 and 0.42, respectively) (p < 0.05). The correlation of Subjective Norm with Normative Beliefs is weak (ρ = 0.10; p = 0.06).
Table 3 Correlation between covariates
Covariates Correlation Coefficient p-value N
Attitude Toward Safety vs. Outcome Beliefs 0.459 <0.0001 326
Perceived Behavioural Control vs. Control Beliefs 0.425 <0.0001 344
Subjective Norm vs. Normative Beliefs 0.101 0.06 345
Attitude Toward Safety vs. Perceived Behavioural Control 0.449 <0.0001 326
Attitude Toward Safety vs. Subjective Norm 0.384 <0.0001 326
Subjective Norm vs. Perceived Behavioural Control 0.271 <0.0001 345
Finally, six of the eleven outcome beliefs are moderately correlated with owners' Attitude Toward Safety (p < 0.05) (Table 4):
Table 4 Correlations between outcome beliefs and Attitude Toward Safety
Outcome Belief Correlation Coefficient p-value
Make employees happier 0.489 <0.0001
Make employees healthier 0.538 <0.0001
Decrease costs* -0.092 0.10
Increase employee productivity 0.385 <0.0001
Not cause employee complaints* -0.112 0.04
Show that I care about employees 0.422 <0.0001
Not cut into profits* -0.048 0.38
Lower workers' compensation costs 0.282 <0.0001
Not take too much time* 0.165 0.003
Increase product quality 0.331 <0.0001
Raise business productivity* 0.006 0.91
* Wording has been changed from the original to reflect re-coding.
• Make employees healthier
• Make employees happier
• Increase employee productivity
• Show that I care about employees
• Lower workers' compensation costs
• Increase product quality
A total number of 300 surveys were available for structural equation modelling (i.e. no missing responses). Path analysis using AMOS 5.0.1 (SPSS Inc., 2003) was used to fit the following multiple regression equations simultaneously:
Y = β1Z1 + β2Z2 + β3Z3 + D
Z1 = γ11X1 + D1
Z2 = γ21X2 + D2
Z3 = γ31X3 + D3
where
Y = log (Intentions Toward Safety)
Z1 = Attitude toward Safety
Z2 = Subjective Norm
Z3 = Perceived Behavioural Control
X1 = Outcome Beliefs
X2 = Normative Beliefs
X3 = Control Beliefs
D1, D2, D3 and D are disturbances and β's and γ's are regression coefficients.
Although sample size does not contribute to the identifiability of a path model, it does contribute to the precision of path analysis estimates. There were 7 main variables observed for this study, therefore a maximum of 28 parameters could be estimated. Using the rule that N/P should be greater than 10, where N = number of subjects, P = number of parameters, we found that the statistical stability of the results from this model should be acceptable [22].
A fitted path model is shown in Figure 2. Arrows symbolize direct effects. The values associated with each path are standardized regression coefficients (weights), which represent the amount of standard deviation change in a dependent or mediating variable given a standard deviation change in the corresponding predicting variable while holding other variables constant. Since we observed significant correlations between the belief variables and Attitude Toward Safety during the model modification process, the direct effects of Normative Beliefs and Control Beliefs on Attitude Toward Safety were added to the path model.
Figure 2 Path analysis results.
The path analysis confirmed the significant effect of Attitude Toward Safety on Intentions Toward Safety (weight = 0.50). Attitude Toward Safety was equally moderately affected by Outcome Beliefs, Normative Beliefs and Control Beliefs (weights of 0.27, 0.28 and 0.28, respectively). The direct effects of Subjective Norm and Perceived Behavioural Control on Intentions Toward Safety were not strong (weights of 0.16 and 0.05, respectively). The link between Normative Beliefs and Subjective Norm was also weak (weight of 0.10). Outcome and Control Beliefs were also strongly associated (weight = 0.49).
We also used logistic regression to assess various relationships between the predictor variables and Intentions Toward Safety. Using a technique proposed by Ajzen and Fishbein [16], Intentions Toward Safety was dichotomized at its median (a better approximation of the center of a skewed distribution) into two groups: owners with high intentions and those with low intentions. The probability of having high intentions was regressed on Attitude Toward Safety, Perceived Behavioural Control and Subjective Norm. Results showed that Attitude Toward Safety (p < 0.0001) and Perceived Behavioural Control (p = 0.0153) were both significantly associated with Intentions Toward Safety (Table 5). Age was also considered, but did not have a significant effect on the outcome.
Table 5 Odds ratio estimates for covariates on dichotomized response variable (Intentions Toward Safety)
Covariate Odds Ratio 95% Confidence Limits
Attitude Toward Safety 1.4 1.2, 1.6
Perceived Behavioural Control 1.3 1.0, 1.7
Subjective Norm 1.7 1.2, 2.5
The largest effect in the path model on log(Intentions Toward Safety) comes from Attitude Toward Safety. Therefore, we further explored which specific beliefs might be associated with business owners' safety attitudes. Attitude toward Safety was dichotomized into "high" vs. "low" at its median. Logistic regression was used to model the probability of having high Attitude Toward Safety. We first examined the association between the eleven outcome beliefs and Attitude Toward Safety (Table 6). Three outcome beliefs showed a high probability of being associated with higher attitudes toward improving safety:
Table 6 Odds ratio estimates for covariates on dichotomized response variable (Attitude Toward Safety)*
Outcome Belief Odds Ratio 95% Confidence Limits
Make employees healthier 2.02 1.51, 2.70
Lower costs 1.50 1.06, 2.11
Show that I care 1.39 1.08, 1.80
* Stepwise selection of variables. Only variables with significant effects are shown.
1. Make employees healthier
2. Lower costs
3. Show that I care
All of the specific beliefs (Outcome, Normative and Control) were then added to the model. One outcome belief (make employees happy), one normative belief (a workers' compensation company that thinks owners should improve health and safety) and two control beliefs (being well-informed and having supportive employees) had a high and significant probability of being associated with higher attitudes toward improving safety.
Discussion
Results from the correlation and path analyses suggest that business owners' intentions toward improving safety are most strongly associated with their attitude toward safety. While neither subjective norm nor behavioural control are predictors of owners' intentions, they are moderately correlated with owners' attitudes toward safety. In addition, outcome and control beliefs are both moderately correlated with attitudes.
These results suggest that it is owners' attitudes that most strongly influence their intentions (and thus behaviours) to improve employee health and safety. A small set of outcome beliefs is associated with higher attitudes. Correlation and logistic regression analyses suggest that this set includes beliefs that improving health and safety will make employees healthier and show that I care. Other outcome beliefs (occurring in at least one of the analyses) that may influence health and safety attitudes include:
• Make employees happier
• Increase employee productivity
• Lower workers' compensation costs
• Increase product quality
• Lower costs
Knowledge, supportive employees, and an influential workers' compensation company are also associated with a more positive attitude toward workplace health and safety.
These results suggest that interventions should be aimed at increasing owners' expectations about the positive outcomes of improving health and safety and building more positive interactions between employees and owners. Demonstrating that business productivity and employee well-being can be enhanced by improvements in health and safety may also lead to higher attitudes. And workers' compensation companies may play an important role in raising attitudes (and thus, intentions) toward workplace health and safety.
There is general consensus that high levels of workplace health and safety require both management commitment and employee involvement [8,9], which are measured by observing the presence of specific activities, programs, systems, and policies. We found no systematic study of owners' or managers' beliefs and attitudes or their relation to intentions and behaviours. The majority of safety behaviour research has focused on employees' opinions about workplace safety, usually termed "safety climate," compared to safety behaviours or injury rates.
A few investigators have explored the relationship between managers' and employees' beliefs about workplace health and safety. For example, Rabin et al. found that employees are more likely to report receiving information about workplace hazards when managers hold more positive outcome beliefs and have more confidence about helping others [23]. Parker et al. found that employee self-reported safe behaviour is associated longitudinally with supportive supervision, job autonomy and the quality of communication about their job [24]. Both suggest that interventions be aimed at improving supervisors' communication with employees about safety. These corroborate our findings that the relationship between owners and employees is important to the development of high intentions toward health and safety among owners.
Study Limitations
Given a single study, it is seldom appropriate to infer causality from the results of statistical analyses, including path analysis. When the variables are concurrently measured, researchers have to make a very clear rationale for specifying the direction of causal effects since ultimately path analysis deals with correlation, not causation [25]. However, the path analysis shows that the data we collected were mostly consistent with the hypothesized model. Cause and effect can be established through intervention trials in which subjects undergo the same experience except for the single facet of interest [25].
Non-response bias is always a concern with survey studies, since owners not responding may differ in some important way from the respondents. However, a 50% response rate is much higher than normally encountered in this population [26]. Government agencies and business associations generally encounter very low response rates (20–30%) unless much more intensive survey methods (e.g. multiple telephone calls) are used. Certainly, non-response should be considered when interpreting these data. However, we believe that these results are still important and relevant to designing interventions in small businesses.
This study relies on self-reported intentions to improve health and safety, which are not validated by observations of behaviour. In many cases intentions have been shown to be correlated with actual behaviour [16]. Resources were not available in this study to observe or measure owners' behaviour.
Conclusion
The results of this study suggest that small business owners' intentions toward improving workplace health and safety are primarily influenced by their attitudes. Owners' outcome, normative and control beliefs all contribute to their attitudes toward workplace health and safety. Subjective norm and perceived behavioural control do not have any significant impact on small business owners' behavioural intentions toward workplace health.
Interventions aimed at these underlying beliefs, particularly those shown to be most highly associated with high-intentioned owners, may be successful in bringing about improvements in attitudes, intentions and behaviour. Raising owners' expectations about positive employee health and business productivity outcomes may lead to long-term improvements in their attitudes, intentions and behaviour toward workplace health and safety.
Competing interests
The author(s) declare they have no competing interests.
Authors' contributions
LMB designed and carried out the survey studies, participated in the data analysis and drafted the manuscript. SYL performed the statistical analysis and helped draft the manuscript. Both authors read and approved the final manuscript.
Acknowledgements
This work was supported NIOSH grant 5 K01 OH00171-03.
==== Refs
Statistics of United States businesses (2002)
Statistics of United States businesses, manufacturing by employment size of enterprise (2001)
Injury and illness data, quartile data, Table Q1. Total recordable cases – injuries and illnesses – detailed industry by establishment size (2003)
Eakin JM Lamm P Limborg HJ Frick K, Quinlan M, Wilthagen T International perspectives on the promotion of health and safety in small workplaces Systematic Occupational Health and Safety Management: Perspectives on an International Development 2000 Amsterdam, The Netherlands: Elsevier 227 247
Barbeau E Roelofs C Youngstrom R Sorensen G Stoddard A LaMontagne AD Assessment of occupational safety and health programs in small businesses Am J Ind Med 2004 45 371 379 15029570 10.1002/ajim.10336
Bierma TJ Waterstraat FL Overcoming Barriers to Pollution Prevention in Small Businesses, Application in the Metal Parts Fabrication Industry 1995 Champaign, Illinois: Illinois Hazardous Waste Research and Information Center
Champoux D Brun J-P Occupational health and safety management in small size enterprises: An overview of the situation and avenues for intervention and research Safety Sci 2003 41 301 318 10.1016/S0925-7535(02)00043-7
Shannon HS Walters V Lewchuck W Richardson J Moran LA Haines T Verma D Workplace organizational correlates of lost-time accident rates in manufacturing Am J Ind Med 1996 29 258 268 8833778 10.1002/(SICI)1097-0274(199603)29:3<258::AID-AJIM5>3.0.CO;2-M
Shannon HS Mayr J Haines T Overview of the relationship between organizational and workplace factors and injury rates Safety Sci 1997 26 201 217 10.1016/S0925-7535(97)00043-X
Habeck RV Leahy MJ Hunt HA Employer factors related to workers' compensation claims and disability management Rehab Counselling Bull 1991 32 210 226
Smith MJ Cohen H Cleveland RJ Characteristics of successful safety programs J Safety Res 1978 10 5 15
Lazovich D Parker DL Brosseau LM Milton FT Dugan SK Pan W Hock L Effectiveness of a worksite intervention to reduce an occupational exposure: The Minnesota wood dust study Am J Public Health 2002 92 1498 1505 12197983
Brosseau LM Parker DL Lazovich D Milton T Dugan S Designing intervention effectiveness studies for occupational health and safety: The Minnesota wood dust study Am J Ind Med 2002 41 54 61 11757055 10.1002/ajim.10029
LaMontagne AD Barbeau E Youngstrom RA Lewiton M Stoddard AM McLellan D Wallace LM Sorensen G Assessing and intervening on OSH programmes: Effectiveness evaluation of the Wellworks-2 intervention in 15 manufacturing worksites Occup Environ Med 2004 61 651 660 15258270 10.1136/oem.2003.011718
Ajzen I Kuhl J, Beckman J From intentions to actions: A theory of planned behavior Action-Control: From Cognition to Behavior 1985 Heidelberg, Germany: Springer 11 39
Ajzen I Fishbein M Understanding Attitudes and Predicting Social Behavior 1980 Englewood Cliffs, N.J.: Prentice-Hall
Occupational Health and Safety Administration Safety and health program management guidelines; issuance of voluntary guidelines Fed Reg 1989 54 3904 3916
British Standards Institute Guide to Health and Safety Management Systems 1996 London, England: British Standards Institute
Dyjack DT Levine SP Holtshouser JL Schork MA Comparison of AIHA ISO 9001-based occupational health and safety management system guidance document with a manufacturer's occupational health and safety assessment instrument Am Ind Hyg Assoc J 1998 59 419 429 9670471
Redinger CF Levine SP Development and evaluation of the Michigan occupational health and safety management system assessment instrument: A universal OHSMS performance measurement tool Am Ind Hyg Assoc J 1998 59 572 581
Osgood CE Suci GJ Tannenbaum RH The Measurement of Meaning 1957 Urbana, IL: University of Illinois Press
Kline RB Principles and Practice of Structural Equation Modeling 1998 New York, New York: The Guilford Press
Rabin S Feiner M Shaham J Yekutieli D Ribak J Impact of managers' personal determinants in notifying workplace hazards Am J Ind Med 1998 33 493 500 9557173 10.1002/(SICI)1097-0274(199805)33:5<493::AID-AJIM9>3.0.CO;2-T
Parker SK Axtell CM Turner N Designing a safer workplace: Importance of job autonomy, communication quality, and supportive supervisors J Occup Health Psychol 2001 6 211 228 11482633 10.1037/1076-8998.6.3.211
Everitt BS Dunn G Applied Multivariate Data Analysis 1991 New York, NY: Halsted Press 257 275
Dennis WJ Raising response rates in mail surveys of small business owners: Results of an experiment Journal of Small Business Management 2003 41 278 295 10.1111/1540-627X.00082
|
16242041
|
PMC1277835
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Oct 21; 4:23
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-23
|
oa_comm
|
==== Front
Genet Vaccines TherGenetic Vaccines and Therapy1479-0556BioMed Central London 1479-0556-3-81621910810.1186/1479-0556-3-8ResearchA trial of somatic gene targeting in vivo with an adenovirus vector Ino Asami [email protected] Yasuhiro [email protected] Hiroyuki [email protected] Naofumi [email protected] Takao [email protected] Ichizo [email protected] Department of Medical Genome Sciences, Graduate School of Frontier Science, University of Tokyo & Institute of Medical Science, University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan2 Graduate Program in Biophysics and Biochemistry, Graduate School of Science the University of Tokyo3 Department of Environmental Information, Keio University, 5322 Endo, Fujisawa, Kanagawa 252-8520, Japan4 Laboratory of Gene Transfer and Regulation, National Institute of Biomedical Innovation, Asagi 7-6-8, Saito, Ibaraki, Osaka 567-0085, Japan5 Pharmaceuticals and Medical Devices Agency, Shin-Kasumigaseki Bldg. 3-3-2, Kasumigaseki, Chiyoda-ku, Tokyo 100-0013, Japan2005 12 10 2005 3 8 8 1 7 2005 12 10 2005 Copyright © 2005 Ino et al; licensee BioMed Central Ltd.2005Ino 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
Gene targeting in vivo provides a potentially powerful method for gene analysis and gene therapy. In order to sensitively detect and accurately measure designed sequence changes, we have used a transgenic mouse system, MutaMouse, which has been developed for detection of mutation in vivo. It carries bacteriophage lambda genome with lacZ+ gene, whose change to lacZ-negative allele is detected after in vitro packaging into bacteriophage particles. We have also demonstrated that gene transfer with a replication-defective adenovirus vector can achieve efficient and accurate gene targeting in vitro.
Methods
An 8 kb long DNA corresponding to the bacteriophage lambda transgene with one of two lacZ-negative single-base-pair-substitution mutant allele was inserted into a replication-defective adenovirus vector. This recombinant adenovirus was injected to the transgenic mice via tail-vein. Twenty-four hours later, genomic DNA was extracted from the liver tissue and the lambda::lacZ were recovered by in vitro packaging. The lacZ-negative phage was detected as a plaque former on agar with phenyl-beta-D-galactoside.
Results
The mutant frequency of the lacZ-negative recombinant adenovirus injected mice was at the same level with the control mouse (~1/10000). Our further restriction analysis did not detect any designed recombinant.
Conclusion
The frequency of gene targeting in the mouse liver by these recombinant adenoviruses was shown to be less than 1/20000 in our assay. However, these results will aid the development of a sensitive, reliable and PCR-independent assay for gene targeting in vivo mediated by virus vectors and other means.
==== Body
Background
Gene targeting, which is the precise alteration of genomic information by homologous recombination, has provided a powerful means of genetic analysis in microorganisms and mammalian systems [1]. In mouse systems, embryonic stem-cell lines modified in vitro can be used to generate mice that are altered at the germ-line level. If the gene targeting of somatic cells is made possible by gene transfer in vivo, it will facilitate the analysis of gene function, and provide a means of gene therapy for genetic and other diseases [2].
There are two major inherent problems with the use of gene targeting in vivo. First, its low efficiency makes it difficult to detect and analyze. A sensitive and accurate measurement system is therefore needed to detect such low-frequency events. Although there have been several reports of gene targeting in the rat liver with specifically designed oligonucleotides [3,4], their reproducibility remains controversial [5]. PCR-based detection methods might thus be inaccurate and prone to various artifacts. In order to detect and measure gene targeting in mice with sufficient sensitivity, we used a bacteriophage transgenic-mouse system, MutaMouse, which has been developed for the detection of mutagenesis in vivo (Figure 1) [6]. The MutaMouse carries tandem repeats of the bacteriophage lambda genome with the lacZ+ gene, in which the change to a lacZ-negative allele is detected after its in vitro packaging into viable bacteriophage particles.
The second major problem with gene targeting in vivo is that non-homologous recombination is much more frequent than homologous recombination in mammalian cells. Rare accurately modified cells are selected and purified in the case of embryonic stem cells that are treated in vitro. For gene targeting in vivo, imprecise modification would be detrimental for analytical uses and therapeutic purposes. Accurate gene modification has been achieved efficiently using replication-defective adenovirus vectors for gene delivery in vitro [7,8]. Fujita and colleagues used a mammalian plasmid as a model target [7]. The gene targeting was frequent (~10-4 per cell) and analysis of the products revealed that homologous recombination was more frequent than non-homologous recombination. One possible reason for this high accuracy was protection of the viral DNA by the terminal protein, which is covalently attached to the ends of the viral DNA and to other viral proteins during its transfer to the nucleus and target DNA. Breaks in unprotected DNA would lead to non-homologous recombination.
Figure 1 Experimental steps to detect gene targeting in vivo. Gene targeting in vivo in liver cells was attempted after the delivery of donor DNA with an adenovirus vector. The gene with the required sequence change (lacZ-) on the lambda transgene in the mouse will be detected after its recovery in bacteriophage particles. Only lacZ-negative mutants can form plaques under the selective conditions.
The adenovirus is useful for gene delivery in vivo because it has a broad host-range, is easy to prepare to a high titer and only rarely integrates into the host genome by non-homologous recombination [9,10]. To date, more than 170 clinical studies have used recombinant adenovirus vectors to express cDNA in humans [11]. Numerous adenovirus-infection experiments have been carried out with mice, and have established that the injection of adenovirus recombinants into the mouse tail-vein leads to the expression of their genes in approximately one-half of the liver cells [12,13].
In the present study, we investigated gene targeting in the mouse liver using a replication-defective adenovirus vector and a transgenic mouse system (Figure 1). Although our initial attempts did not detect the predicted gene targeting (the frequency of the expected recombinants was less than 1/20,000 per lambda genome), the strategy and methods detailed here will aid the development of virus-mediated gene targeting in vivo.
Materials and methods
Bacteria, bacteriophages and plasmids
The bacteria, bacteriophages and plasmids used in this study are listed together with details of their construction in Additional file 1.
BIK12001 was used for the titration of bacteriophage lambda and the measurement of lacZ-negative bacteriophage lambda by phenyl beta-D-galactoside (p-gal) selection (see below). BIK1564 was used for the growth of all bacteriophage lambda strains in this study. BIK2206 was used for confirmation of the LacZ-negative phenotype of the bacteriophage selected with p-gal using 5-bromo-4-chloro-3-indlyl-beta-D-galactose (X-gal).
The construction of the plasmids used in this study is detailed in additional file 1. The construction of pAdNY58 is also illustrated in Figure 2. The construction of pAdNY57 was as follows. The SmaI(1)-SacI fragment of LIA7 within the lacZ gene (Figure 2) was used to replace the shorter SmaI-SacI fragment of pUC18. The Glu461Gly mutation (Figure 3) was introduced into the resulting plasmid (pNY15) by site-directed mutagenesis using PCR [14] as follows. The PCR products generated with the primer pair LZG-U (5'-ACCGGCGATGAGCGAA-3') and LZG-MA (5'-GCCTGATCCATTCCCCAGCGACCA-3'), and the primer pair LZG-MS (5'-GGGAATGGATCAGGCCACGGCCGC-3') and LZG-D (5'-GGGCTGGTCTTCATCC-3'), were mixed and used as templates for the second round of PCR with the primer pair LZG-U and LZG-D. The MluI-BssHII fragment of the wild-type lacZ gene of pNY15 was replaced by the MluI-BssHII fragment of the PCR product. The targeted change in the resulting plasmid (pNY15G3.11) was confirmed by sequencing. pNY20 was produced by replacing the smaller SmaI-SacI fragment of pNY19 with the homologous SmaI-SacI fragment of pNY15G3.11, which carries the mutant sequence.
Figure 2 Construction of the recombinant adenovirus AdNY58. The bacteriophage lambda LIA7 was recovered from the MutaMouse by in vitro packaging. An SmaI-SacI fragment of LIA7 within its lacZ gene was inserted into pIK153. The Tyr105Stop mutation (Figure 3) was introduced into the resulting plasmid (pIK153LZS.6) using site-directed mutagenesis by PCR as follows. The PCR products generated with the primer pair LZT-U (5'-CGAAGAGGCCCGCAC-3') and LZT-MA (5'-TAATGGGCTAGGTTACGTTGGTGTAG-3'), and the primer pair LZT-MS (5'-TAACCTAGCCCATTACGGTCAATCC-3') and LZT-D (5'-GGCAACATGGAAATCGC-3') were mixed and used as templates for the second PCR with the primer pair LTZ-U and LZT-D. Replacement of an FspI-AatII fragment of pIK153LZS.6 by the FspI-AatII fragment of the resulting PCR product resulted in pIK153 T10.1. A BamHI-SmaI fragment covering the lacZ gene of LIA7 was inserted into the BamHI site of pIK153 (resulting in pNY19). pNY21 was made by replacing the smaller SmaI-SacI fragment of pNY19 with the homologous SmaI-SacI fragment of pIK153T10.1, which carries the mutant sequence. An XbaI-BglII fragment of pNY21 was used to replace the smaller XbaI-BamHI fragment of pHM5 (resulting in pNY58). pAdNY58 was made by replacement of the smaller I-CeuI-PI-SceI fragment of pAdHM4 with an I-CeuI-PI-SceI fragment of pNY58. The longer PacI fragment of pAdNY58 was transfected into 293 cells. The recombinant adenovirus AdNY58 was prepared and purified from the cell culture.
These two lacZ mutations were transferred back to lambda by homologous recombination in vivo [15] in order to generate LIA15 and LIA11, respectively. The recombinational transfer was carried out as follows. Cells of BIK12015 or BIK12018 were grown to OD600 = ~0.3 in LB (10 g bactotrypton, 5 g yeast extract and 10 g NaCl per liter) containing 20 μg/ml chloramphenicol, 0.2% maltose and 10 mM MgSO4. LIA7 was adsorbed onto the cells at a multiplicity of 1.0 at 37°C for 15 minutes. The mixture was shaken at 37°C until the OD600 dropped below 0.3. One drop of CHCl3 was added to the mixture, which was then shaken for 30 seconds. The mixture was centrifuged and the supernatant was recovered. The supernatant was assayed for BIK12001 on agar plates containing p-gal as detailed below. The plaques on the p-gal plates were isolated and analyzed for the designed sequence change by restriction of the PCR products (see Analysis of the mutant bacteriophage DNA).
Selection of lacZ-negative bacteriophage with p-gal
The lacZ-negative bacteriophage particles were detected using positive selection [15,16]. BIK12001 cells were grown with shaking at 37°C to OD600 = 1.0 in LB containing ampicillin (50 μg/ml), kanamycin (20 μg/ml) and 0.2% maltose. The culture was centrifuged at 3,500 rpm for 15 minutes at 4°C. The pellets were dissolved into one-half the volume of LB containing 10 mM MgSO4. The bacteriophage was adsorbed onto these cells at room temperature for 20 minutes. To estimate the total number of bacteriophages, 2.5 ml molten 1/4 LB top agar (5 g LB broth base (Gibco BRL, Rockville, MD, USA), 6.4 g NaCl and 7.5 g Bactoagar per liter) was added to 0.25 ml of the mixture of cells and bacteriophages, and the entire content was poured onto a 1/4 LB plate (5 g LB broth base, 6.4 g NaCl and 15 g Bactoagar per liter). To estimate the number of lacZ-negative bacteriophages, 2 ml of the mixture of cells and bacteriophages, and 22 ml of molten 1/4 LB top agar containing 0.3% p-gal (Sigma Chemical Co., MO, USA), were mixed and poured onto four 1/4 LB plates. The plates were incubated at 37°C for 12 hours.
Construction of recombinant adenoviruses
pNY56 was constructed by replacing the shorter XbaI-BamHI fragment of pHM5 by the XbaI-BglII fragment of pNY19 (Figure 2). pAdHM4 includes the entire genome of the recombinant adenovirus vector. The plasmid pAdNY56 was constructed by replacing the shorter I-CeuI-PI-SceI fragment of pAdHM4 by an I-CeuI-PI-SceI fragment of pNY56. The PacI fragment of pAdNY56 was transfected into cells of cell-line 293, which allows replication of the replication-defective adenoviruses. The recombinant adenovirus AdNY56 was prepared and purified as described previously [18]. Similarly, AdNY57 was constructed from pNY20 via pNY57 (Additional file 1), and AdNY58 was constructed from pNY21 via pNY58 (Figure 2, Additional file 1).
Adenovirus infection
Female MutaMice (7 weeks old) were obtained from Covance Research Products Inc. (Denver, PA, USA). The MutaMice were maintained under specific pathogen-free conditions in the animal faculty of the Institute of Medical Science at the University of Tokyo, Japan. After the animals were anesthetized with Nembutal (Dainippon Pharmaceutical Co., Osaka, Japan), 3 × 109 plaque-forming units (PFU) of the recombinant adenovirus in 200 μl of PBS (137 mM NaCl, 8.10 mM Na2HPO4, 2.68 mM KCl, 1.47 mM KH2PO4, 0.9 mM CaCl2, 0.33 mM MgCl2) was injected into the tail-vein of each mouse using a 30-gauge needle. AdNY56 was injected into one mouse, AdNY57 was injected into two mice and AdNY58 was injected into two mice.
Isolation of genomic DNA, recovery of lambda bacteriophage and measurement of mutant frequency
Twenty-four hours after injection, the mice were sacrificed. A lobe of the liver of each animal was excised, frozen by submersion in liquid nitrogen and stored in a 1.5-ml plastic tube at -80°C. Genomic DNA was isolated from the liver tissue with phenol-chloroform and precipitated by ethanol/sodium as described in the manual for MutaMouse. Lambda bacteriophage particles were recovered from the isolated DNA by incubation with packaging extracts (Mutaplax, Epicentre, WI, USA). The lacZ-negative mutants were detected by p-gal selection as described above. Each plaque on the selective agar was recovered in 100 μl of SM buffer (50 mM Tris-HCl (pH 7.5), 10 mM MgSO4, 100 mM NaCl and 0.01% gelatin). In order to verify the lacZ-negative phenotype, each isolate was assayed on agar with X-gal using a spot assay as follows. BIK2206 was grown in LB containing ampicillin (50 μg/ml) and tetracycline (10 μg/ml). Twice-concentrated culture (1.25 ml) was mixed with 6 ml molten LB/MM agar (100 ml LB medium, 0.75 g Bactoagar, 10 mM MgSO4, 0.2% maltose and 0.35 mg/ml X-gal) and spread on agar. A 10-μl aliquot of each bacteriophage sample was spotted onto these cells. The plates were incubated overnight at 37°C. The mutant frequency was estimated by dividing the number of PFU on the selective plate (as verified with X-gal) by the number of total PFU on 1/4 LB agar.
Analysis of the mutant bacteriophage DNA
The lacZ-negative lambda bacteriophage DNA from the mice was analyzed using restriction enzymes following PCR. For the lacZ-negative lambda DNA from the AdNY57-treated mouse, PCR was carried out with the primer pair LG-1 (5'-TACCGGCGATGAGCGAAC-3') and LG-2 (5'-CTCCAGGTAGCGAAAGCC-3'). The 288-bp product was purified by ethanol/sodium precipitation, digested with TfiI (New England Biolabs, Beverly, MA, USA) (recognition site, 5'-G|AWTC-3' (W = A or T)) at 65°C and analyzed using agarose electrophoresis. The mutant sequence was resistant to TfiI, while the wild-type sequence was sensitive, yielding 204 and 84 bp fragments. The primer pair Lam-1 (5'-TACTGTCGTCGTCCCCTC-3') and Lam-2 (5'-CGCAGATGAAACGCCGAGT-3') was used for the lacZ-negative lambda DNA from the AdNY58-treated mouse. The 213-bp PCR product was digested with XspI (Takara Bio Inc., Shiga, Japan) (recognition site, 5'-C|TAG-3') at 37°C and analyzed using agarose electrophoresis. The wild-type sequence was resistant to XspI, while the mutant sequence was sensitive, yielding 146 and 67 bp fragments.
Results
Experimental design for the detection of gene targeting in vivo
Figure 1 illustrates our experimental design for the sensitive detection of gene targeting in vivo. The MutaMouse carries approximately 40 copies of bacteriophage lambda gt10lacZ on a chromosome [6,19]. The single integration site is located in band C on chromosome 3 [20]. Our target sequence was the wild-type lacZ gene. The donor DNA was delivered to the liver cell nuclei by tail-vein injection of the recombinant adenovirus. Genomic DNA was isolated from the liver and its in vitro packaging allowed the recovery of the lambda genome in viable bacteriophage particles. A lacZ-negative mutant bacteriophage was selected as a plaque-former in an Escherichia coli mutant defective in the galE gene on an agar plate containing p-gal. This chemical is converted by the lacZ gene product (beta-galactosidase) into UDP-galactose, which accumulates in the absence of the GalE protein to induce cell death. The ratio of the mutant plaque-formers to the total plaque-formers was used to estimate the fraction of the mutated gene. The mutant gene was further analyzed using restriction enzymes.
Replication-defective recombinant adenoviruses constructed by an in vitro-ligation method were used to deliver the donor DNA [18,21]. Figure 3 shows the structure of the recombinant adenoviruses used in the present study (see Figure 2, Additional file 1, and Materials and methods for further details). An 8077-bp fragment of lambda gt10lacZ was inserted into the E1 deletion site of the mutant adenovirus [18,21]. AdNY56 had wild-type lacZ, while AdNY57 and AdNY58 had a point mutation in lacZ (Figure 3B).
Figure 3 Design for gene targeting and its detection. (A) The donor carrying the mutant lacZ gene is inserted into an adenovirus vector. The lacZ mutation will be transferred to the lacZ gene of the lambda transgene in the mouse genome. (B) Expected sequence changes and their detection using restriction analysis.
AdNY57 was constructed so as to introduce a point mutation at the active site of LacZ. The target sequence was the 5' GAA that codes for Glu461, which is essential for the activity of LacZ [22,23]. AdNY57 was expected to change its second base (that is, the 1437 th base) from A to G, thereby generating the Glu461Gly mutant, which shows a 76-fold decrease in activity [23]. The mutant and wild-type sequences can be distinguished using the restriction enzyme TfiI (Figure 3B).
AdNY58 was constructed so as to introduce a point mutation at the 5' TAT that codes for Tyr105. AdNY58 was expected to change its third base (that is, the 369th base) from T to G, thereby generating the Tyr105Stop mutant. The mutant and wild-type sequences can be distinguished using the restriction enzyme XspI (Figure 3B).
Control experiments
We demonstrated that lacZ mutants that were predicted to be generated by the recombinant adenovirus could be selected with p-gal as follows. Bacteriophage lambda strains carrying the mutations were produced by transferring each mutation on a plasmid back to lambda through homologous recombination in E. coli (as detailed in Materials and methods). The two bacteriophage strains, lambda gt10lacZ- Tyr105Stop (LIA11) and lambda gt10lacZ- Glu461Gly (LIA15), were then used in the p-gal selection. As shown in Table 1, lambda with wild-type lacZ showed a plaque-formation efficiency of less than 1/10,000 on the selective agar relative to that on the non-selective agar. By contrast, each of the mutant lambda strains showed similar or slightly decreased plaque-formation efficiency on the selective agar. We concluded that the expected targeted product with AdNY57 and AdNY58, if it was produced, should be selected and measured using the p-gal-selection procedure.
Table 1 Selection efficiency of lambda lacZ-negative mutants
Lambda Genotype Titer Titer on p-gal selective plate Relative plaque formation
LIA7 lacZ+ 2.2 × 1010 1.9 × 106 8.6 × 10-5
LIA11 lacZ- (Tyr105Stop) 9.6 × 1010 8.8 × 1010 9.2 × 10-1
LIA15 lacZ- (Glu461Gly) 1.4 × 1010 1.3 × 1010 9.1 × 10-1
Delivery of donor DNA and measurement of mutant frequency
The recombinant adenovirus particles (3 × 109 PFU in 200 μl of PBS) were injected into the tail-vein of a MutaMouse. It is well established that the adenovirus genome accumulates in the liver cell nuclei after tail-vein injection [12,13]. Most of the hepatocyte nuclei are expected to receive several copies of the adenovirus genome under these conditions (see Discussion). After 24 hours, the liver was excised from the MutaMouse, genomic DNA was isolated from the liver tissue and the lambda genome was recovered as a bacteriophage particle by in vitro packaging. The lacZ-negative phage was detected selectively on agar with p-gal. The plaques on these selective plates were isolated and the LacZ-negative phenotype was confirmed on agar plates containing X-gal. The mutant frequency was estimated as the fraction of the lacZ-negative phage (Table 2). The control mouse (animal number 0) received no injections.
Table 2 Detection of lacZ- phage
Packaging exp. RAd Genotype Animal number Packaging Total number of plaque formers lacZ- plaques Mutant Frequency Expected genotype
1 None Not relevant #0 Tube 1 4.5 × 104 4 8.9 × 10-5 n.t.
Tube 2 3.2 × 104 3 9.4 × 10-5 n.t.
Tube 3 8.5 × 104 8 9.4 × 10-5 n.t.
average 9.2 × 10-5
AdNY56 lacZ+ #1 Tube 4 8.5 × 104 8 9.4 × 10-5 n.t.
Tube 5 6.4 × 104 3 4.7 × 10-5 n.t.
Tube 6 8.8 × 104 1 1.1 × 10-5 n.t.
average 5.1 × 10-5
2 None Not relevant #0 Tube 7 4.9 × 104 5 10 × 10-5 n.t.
Tube 8 6.6 × 104 4 6.1 × 10-5 n.t.
Tube 9 5.1 × 104 10 20 × 10-5 n.t.
average 12 × 10-5
AdNY57 lacZ- (Glu461Gly) #2 Tube 10 3.8 × 104 6 16 × 10-5 0/6
Tube 11 3.0 × 104 7 23 × 10-5 0/7
Tube 12 4.5 × 104 9 20 × 10-5 0/9
average 20 × 10-5 total 0/22
3 None Not relevant #0 Tube 13 3.7 × 104 6 16 × 10-5 n.t.
Tube 14 6.0 × 104 5 8.3 × 10-5 n.t.
Tube 15 4.4 × 104 4 9.1 × 10-5 n.t.
average 11 × 10-5
AdNY57 lacZ- (Glu461Gly) #2 Tube 16 1.3 × 104 9 69 × 10-5 0/9
Tube 17 3.9 × 104 19 49 × 10-5 0/19
Tube 18 6.5 × 104 26 40 × 10-5 0/26
average 53 × 10-5 total 0/54
4 None Not relevant #0 Tube 19 2.6 × 105 8 8.5 × 10-5 n.t.
AdNY57 lacZ- (Glu461Gly) #3 Tube 20 1.6 × 105 5 6.3 × 10-5 0/5
Tube 21 4.1 × 105 9 8.6 × 10-5 0/9
average 1.5 × 10-5 total 0/14
5 None Not relevant #0 Tube 22 3.3 × 104 3 9.1 × 10-5 n.t.
AdNY58 lacZ- (Tyr105 Stop) #4 Tube 23 8.6 × 104 4 4.7 × 10-5 0/4
Tube 24 3.1 × 104 3 9.7 × 10-5 0/3
average 7.2 × 10-5 total 0/7
RAd: Recombinant adenovirus
n.t.: Not tested.
The mutant frequencies of the AdNY56-injected and control mice were similar (Table 2, Experiment 1), and did not differ significantly from those reported previously using this method (see [15] and the references cited therein). No significant increase in the mutant form of the gene was induced by injection of the recombinant adenovirus: the mutant frequency of the AdNY57- and AdNY58-injected mice was similar to that of the control mouse, which was approximately 1/10,000 (Table 2).
All of the lacZ-negative bacteriophages were purified and their lacZ genes were analyzed using restriction-enzyme treatment of the PCR products (Figure 4). As shown in Figures 3B and 4A, the PCR product of the Glu461Gly mutant, as predicted from the AdNY57 injection, could not be cut with TfiI. By contrast, the wild-type and most of the other possible mutants could be cut with TfiI. In fact, all of the lacZ-negative bacteriophages from the AdNY57-injected mouse were cleavable with this restriction enzyme. As shown in Figure 3B and 4B, the PCR product of the Tyr105Stop mutant, as predicted from the AdNY58 injection, could be cut with XspI. By contrast, the wild-type and most of the other mutants could not be cut with XspI. None of the lacZ-negative bacteriophages from the AdNY58-injected mice were cleavable with this restriction enzyme.
Figure 4 Restriction analysis of the lacZ-negative gene from mice treated with a recombinant adenovirus. (A) AdNY57-injected mouse. The PCR product of the lambda bacteriophage DNA with primers that flank the target site is 288 bp long. The wild-type PCR product is cut with TfiI into 84 and 204 bp fragments, whereas the Glu461Ala mutant PCR product is not cut. Lane M: Marker DNA prepared by HinfI digestion of the plasmid pUC19; 1–12, lacZ-negative bacteriophages from animal number 2; lacZ+: Lambda bacteriophage recovered from control mouse; lacZ-Glu461Gly: lambda bacteriophage LIA15. (B) AdNY58-injected mouse. The PCR product of the lambda bacteriophage DNA with primers that flank the target site is 213 bp long. The Tyr105Stop mutant PCR product is cut with XspI into 146 and 67 bp fragments, whereas the wild-type product is not. Lane M: Marker DNA prepared by HinfI digestion of plasmid pUC19; 1–4, lacZ-negative bacteriophages from animal number 3; lacZ+: Lambda bacteriophage recovered from control mouse; lacZ-Tyr105Stop: lambda bacteriophage LIA11.
We did not detect the expected gene replacement in any of the isolates. Moreover, the gene-correction frequency by these adenovirus constructs was shown to be less than 1/20,000 in the present system.
Discussion
Here we attempted to perform gene targeting in a transgenic mouse system that allowed the sensitive detection of mutagenesis by various agents, such as those directly interacting with DNA in the liver and other organs [24,25]. The limit of sensitivity in this system was 1/20,000 (see also [15]). This procedure might provide an alternative to the PCR-based assay for gene targeting in vivo, although our initial trials did not detect any of the expected recombinants.
In the present system, the sensitivity appeared to be limited by the high level of spontaneous mutagenesis in the target gene. The MutaMouse system was produced to detect mutagenesis at numerous sites within a gene, rather than to study gene targeting. Experimental designs involving the specific selection of homologous recombination events, such as those used in the previous work in vitro [7], would therefore be preferred.
Also, in the present system, a successful gene-targeting event would not be distinguishable in the phenotype of the mouse cell. In transgenic mice with a single copy of the mutant lacZ gene [26], correction to the wild-type gene would result in a direct positive readout in the mouse body (for example, through staining with dye). However, as the authors admit, it would be difficult to detect the targeting events with a high sensitivity. The presence of multiple copies of the target gene would improve the sensitivity because the lacZ+ allele is dominant over, and epistatic to, the lacZ- alleles with respect to the above phenotype. The MutaMouse carries multiple (approximately 40) copies of the target gene, which amount to 0.4% of the genome. This should be able to improve the sensitivity of detection of gene targeting, although the sensitivity is limited by spontaneous mutagenesis. In addition, the presence of tandem repeats might have other types of negative effect on gene targeting, as detailed below.
How efficient is adenovirus infection and delivery to the hepatocyte nucleus? Tail-vein injection is an established method for the delivery of adenovirus to liver cells. The average copy number of a replication-defective recombinant adenovirus genome per liver cell has been estimated as 14–28 copies using Southern hybridization after tail-vein injection of 5 × 109 PFU of the virus [12]. This corresponds to 40% of the injected adenovirus. Fluorescence in situ hybridization revealed that, after tail-vein injection of 2 × 109 PFU, all of the hepatocyte nuclei had 1–100 copies of a recombinant adenovirus genome, with an average of 20 copies [27]. After tail-vein injection of 2 × 108 PFU of a recombinant adenovirus with the lacZ expression cassette, 40% of the hepatocytes expressed beta-galactosidase [13]. We assumed that the majority of the liver cells received several copies of the adenovirus genome, at least sufficient for gene expression, after injecting 3 × 109 PFU in our experiment. (We cannot raise the titer any more because of the toxicity of the virus.) This type of information can be confirmed by Southern hybridization and fluorescence in situ hybridization.
The gene-targeting frequency with recombinant adenoviruses in vitro varies from ~10-7–10-4 per cell [7,8,28]. We did not detect any signal using recombinant adenovirus for gene delivery in the mouse liver. In order to achieve gene targeting in vivo using an adenovirus vector or any other means, it will be necessary to increase the frequency of gene targeting. So how can we achieve this goal?
The efficiency of gene targeting in vitro varies from one locus to another [29,30]. Such locus-dependence might reflect drastic effects of the chromatin structure on the frequency of homologous recombination [30,31]. Thus, the target transgene could be placed at a different locus that is known to be a hot spot in gene targeting in embryonic stem (ES) cells.
Repetitive sequences are methylated in the mouse genome [32]. Ikehata and colleagues suggested that the whole coding region of the MutaMouse lacZ transgene is methylated to a high degree at every CpG site [33]. One possible reason for this phenomenon is that the CpG content of the lacZ gene (9%) [34] is much higher than the average CpG content of the mouse genome (~1%) [35]. Methyl-CpG binding protein 2 (MeCP2) might bind to methylated CpG and somehow compact chromatin [36]. Furthermore, Manuelidis analyzed the structure of a mouse chromosome bearing a huge (~11 Mb) insert of a tandem-repeated transgene (~1,000 copies) [37]. This transgene was localized on an arm of chromosome 3 at a distance from the centromere. According to Manuelidis, the transgene is heterochromatic and highly condensed. Therefore, the MutaMouse transgene might be heterochromatic. The accessibility of nucleases to the heterochromatic structure is lower than that of euchromatin [38,39]. Reducing the copy number of the transgene and/or using another transgene that is lower in CpG content might increase gene targeting, although the decrease in copy number might affect the sensitivity of detection. An important experiment that can be done is to test whether the coding region of the MutaMouse lacZ transgene is really heterochromatic, using, for example, CHIP assay with the antibody against the methylated histones and PCR primers on the lacZ genes.
Chromosome replication is known to stimulate homologous recombination. Partial hepatectomies in mice might stimulate liver cell proliferation and DNA replication, which in turn might stimulate recombination. Hara et al. (1999) reported that partial hepatectomies increased mutagenesis with N-ethyl-N-nitrosourea, which is a direct-acting DNA-ethylation agent, in the MutaMouse [40].
It might be easier to modify the donor DNA than the recipient DNA. One can generate recombinogenic damage on the donor DNA. Irradiating adenovirus particles with ultraviolet light of 1500 J/m2 resulted in an approximately three-fold increase in their mutual homologous recombination [41]. Recombinogenic cross-links are induced by some mutagens, such as psoralens, cisplatin (cis-diamminedichloroplatinum) and mitomycin C [42]. Such agents, both mutagenic and recombinogenic, might be suitable for gene targeting in vivo if they are shown to be active in mutagenesis in a transgenic-reporter mouse system. The effect of such recombinogenic damage might be much larger with replication-defective adenovirus recombinants than with replication-competent adenoviruses, because their replication-intermediates are responsible for their high recombination frequency [41,43-45].
The gene-targeting frequency is strongly dependent on the length of homology; the frequency increases as the homology length increases up to 10 kb [46-48]. If the deviation from this rule above 10 kb is due to the shearing and/or degradation of longer DNA after electroporation in embryonic stem cells, donor DNAs that are protected by the DNA binding proteins in the adenovirus particle might show greater length dependence over a wider range of values. Adenoviral vectors with a larger capacity for inserts, which are known as high-capacity 'gutless' vectors [49-51] might therefore be suitable for use in this approach.
Conclusion
Here we attempted to perform gene targeting in a transgenic mouse system that allowed the sensitive detection of mutagenesis. The frequency of gene targeting in the mouse liver by these recombinant adenoviruses was shown to be less than 1/20000 with the sensitive and PCR-independent detection system.
List of abbreviations
PCR, polymerase chain reaction; PFU, plaque-forming unit; RFLP, restriction fragment length polymorphism; p-gal, phenyl-beta-D-galactoside; X-gal, 5-bromo-4-chloro-3-indlyl-beta-D-galactose
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AI carried out the injection of the recombinant adenovirus and the analysis of the mouse DNA. YN and HM constructed the recombinant adenovirus. NH injected the recombinant adenovirus to the mouse. YN constructed the experimental design as well as cloning of the part of lambda DNA from the MutaMouse genomic DNA. IK provided the original experimental idea and coordinated the experimental design. All authors read and approved the final manuscript.
Supplementary Material
Additional file 1
Bacterial strains, plasmids, bacteriophage strains and recombinant adenovirus constructs.
Click here for file
Acknowledgements
Ms. Kuniko Iwasaki and Dr. Ryuichi Miura from the Laboratory Animal Research Center of the Institute of Medical Science, Japan, guided us in our manipulation of the mice. Dr. Noriko Takahashi from our laboratory helped with the maintenance of the mice. Dr. Yoichiro Iwakura of the Institute of Medical Science provided critical comments on an early version of the manuscript. This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (No.0828102: General Mechanisms of DNA Recombination Repair. 1996–1999) and the Japan Owners Association (JOA) (1999–2002) as arranged by the Japan Society for Gene Therapy.
==== Refs
Capecchi MR Altering the genome by homologous recombination Science 1989 244 1288 1292 2660260
Yanez RJ Porter AC Therapeutic gene targeting Gene Ther 1998 5 149 159 9578833 10.1038/sj.gt.3300601
Kren BT Bandyopadhyay P Steer CJ In vivo site-directed mutagenesis of the factor IX gene by chimeric RNA/DNA oligonucleotides Nat Med 1998 4 285 290 9500600 10.1038/nm0398-285
Kren BT Parashar B Bandyopadhyay P Chowdhury NR Chowdhury JR Steer CJ Correction of the UDP-glucuronosyltransferase gene defect in the gunn rat model of crigler-najjar syndrome type I with a chimeric oligonucleotide Proc Natl Acad Sci USA 1999 96 10349 10354 10468611 10.1073/pnas.96.18.10349
Taubes G Gene therapy. The strange case of chimeraplasty Science 2002 298 2116 2120 12481116 10.1126/science.298.5601.2116
Gossen JA de Leeuw WJ Tan CH Zwarthoff EC Berends F Lohman PH Knook DL Vijg J Efficient rescue of integrated shuttle vectors from transgenic mice: a model for studying mutations in vivo Proc Natl Acad Sci USA 1989 86 7971 7975 2530578
Fujita A Sakagami K Kanegae Y Saito I Kobayashi I Gene targeting with a replication-defective adenovirus vector J Virol 1995 69 6180 6190 7666520
Mitani K Wakamiya M Hasty P Graham FL Bradley A Caskey CT Gene targeting in mouse embryonic stem cells with an adenoviral vector Somat Cell Mol Genet 1995 21 221 231 8525428 10.1007/BF02255777
Doerfler W The fate of the DNA of adenovirus type 12 in baby hamster kidney cells Proc Natl Acad Sci USA 1968 60 636 643 5248820
Harui A Suzuki S Kochanek S Mitani K Frequency and stability of chromosomal integration of adenovirus vectors J Virol 1999 73 6141 6146 10364373
Volpers C Kochanek S Adenoviral vectors for gene transfer and therapy J Gene Med 2004 6 S164 171 14978759 10.1002/jgm.496
Vrancken Peeters MJ Perkins AL Kay MA Method for multiple portal vein infusions in mice: quantitation of adenovirus-mediated hepatic gene transfer Biotechniques 1996 20 278 285 8825158
Nakatani T Kuriyama S Tominaga K Tsujimoto T Mitoro A Yamazaki M Tsujinoue H Yoshiji H Nagao S Fukui H Assessment of efficiency and safety of adenovirus mediated gene transfer into normal and damaged murine livers Gut 2000 47 563 570 10986218 10.1136/gut.47.4.563
Cormack B Ausubel MF, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struho K Mutagenesis by the polymerase chain reaction Current protocols in moecular biology 1991 1 John Wiley and Sons, Inc 8.5.1 8.5.9
Ino A Yamamoto S Kaneda Y Kobayashi I Somatic gene targeting with RNA/DNA chimeric oligonucleotides: an analysis with a sensitive reporter mouse system J Gene Med 2004 6 1272 1280 15459966 10.1002/jgm.617
Dean SW Myhr B Measurement of gene mutation in vivo using Muta Mouse and positive selection for lacZ- phage Mutagenesis 1994 9 183 185 7934957
Gossen JA Vijg J A selective system for lacZ-phage using a galactose-sensitive E.coli host Biotechniqnics 1993 14 326 330
Mizuguchi H Kay MA Efficient construction of a recombinant adenovirus vector by an improved in vitro ligation method Hum Gene Ther 1998 9 2577 2583 9853524 10.1089/10430349850019418
Blakey DH Douglas GR Huang KC Winter HJ Cytogenetic mapping of lambda gt10 lacZ sequences in the transgenic mouse strain 40.6 (Muta Mouse) Mutagenesis 1995 10 145 148 7603331
Swiger RR Myhr B Tucker JD The LacZ transgene in MutaMouse maps to chromosome 3 Mutat Res 1994 325 145 148 7527906 10.1016/0165-7992(94)90077-9
Mizuguchi H Kay MA A simple method for constructing E1- and E1/E4-deleted recombinant adenoviral vectors Hum Gene Ther 1999 10 2013 2017 10466635 10.1089/10430349950017374
Jacobson RH Zhang XJ DuBose RF Matthews BW Three-dimensional structure of beta-galactosidase from E. coli Nature 1994 369 761 766 8008071 10.1038/369761a0
Cupples CG Miller JH Effects of amino acid substitutions at the active site in Escherichia coli beta-galactosidase Genetics 1988 120 637 644 2906303
Cosentino L Heddle JA A comparison of the effects of diverse mutagens at the lacZ transgene and Dlb-1 locus in vivo Mutagenesis 1999 14 113 119 10474832 10.1093/mutage/14.1.113
Krebs O Favor J Somatic and germ cell mutagenesis in lambda lacZ transgenic mice treated with acrylamide or ethylnitrosourea Mutat Res 1997 388 239 248 9057886
Nickerson HD Colledge WH A LacZ-based transgenic mouse for detection of somatic gene repair events in vivo Gene Ther 2004 11 1351 1357 15229632 10.1038/sj.gt.3302311
Schowalter DB Tubb JC Liu M Wilson CB Kay MA Heterologous expression of adenovirus E3-gp19K in an E1a-deleted adenovirus vector inhibits MHC I expression in vitro, but does not prolong transgene expression in vivo Gene Ther 1997 4 351 360 9176522 10.1038/sj.gt.3300398
Wang Q Taylor MW Correction of a deletion mutant by gene targeting with an adenovirus vector Mol Cell Biol 1993 13 918 927 8423811
Bird D Bradshaw R Gene targeting is locus dependent in the filamentous fungus Aspergillus nidulans Mol Gen Genet 1997 255 219 225 9236780 10.1007/s004380050492
Petes TD Meiotic recombination hot spots and cold spots Nat Rev Genet 2001 2 360 369 11331902 10.1038/35072078
Ohta K Shibata T Nicolas A Changes in chromatin structure at recombination initiation sites during yeast meiosis Embo J 1994 13 5754 5763 7988571
Lau S Jardine K McBurney MW DNA methylation pattern of a tandemly repeated LacZ transgene indicates that most copies are silent Dev Dyn 1999 215 126 138 10373017 10.1002/(SICI)1097-0177(199906)215:2<126::AID-DVDY5>3.0.CO;2-8
Ikehata H Takatsu M Saito Y Ono T Distribution of spontaneous CpG-associated G:C --> A:T mutations in the lacZ gene of Muta mice: effects of CpG methylation, the sequence context of CpG sites, and severity of mutations on the activity of the lacZ gene product Environ Mol Mutagen 2000 36 301 311 11152563 10.1002/1098-2280(2000)36:4<301::AID-EM6>3.0.CO;2-R
Chevalier-Mariette C Henry I Montfort L Capgras S Forlani S Muschler J Nicolas JF CpG content affects gene silencing in mice: evidence from novel transgenes Genome Biol 2003 4 R53 12952532 10.1186/gb-2003-4-9-r53
Takai D Jones PA Comprehensive analysis of CpG islands in human chromosomes 21 and 22 Proc Natl Acad Sci USA 2002 99 3740 3745 11891299 10.1073/pnas.052410099
Bowen NJ Palmer MB Wade PA Chromosomal regulation by MeCP2: structural and enzymatic considerations Cell Mol Life Sci 2004 61 2163 2167 15338046 10.1007/s00018-004-4177-x
Manuelidis L Heterochromatic features of an 11-megabase transgene in brain cells Proc Natl Acad Sci USA 1991 88 1049 1053 1992455
Vega-Palas MA Venditti S Di Mauro E Heterochromatin organization of a natural yeast telomere. Changes of nucleosome distribution driven by the absence of Sir3p J Biol Chem 1998 273 9388 9392 9545262 10.1074/jbc.273.16.9388
Ye F Signer ER RIGS (repeat-induced gene silencing) in Arabidopsis is transcriptional and alters chromatin configuration Proc Natl Acad Sci USA 1996 93 10881 10886 8855276 10.1073/pnas.93.20.10881
Hara T Sui H Kawa Kami K Shimada Y Shibuya T Partial hepatectomy strongly increased the mutagenicity of N-ethyl-N-nitrosourea in Mutamouse liver Environ Mol Mutagen 1999 34 121 123 10529735 10.1002/(SICI)1098-2280(1999)34:2/3<121::AID-EM10>3.0.CO;2-R
Rainbow AJ Castillo JE Homologous recombination of adenovirus DNA in mammalian cells: enhanced recombination following UV-irradiation of the virus Mutat Res 1992 274 201 210 1380653
Wang YY Maher VM Liskay RM McCormick JJ Carcinogens can induce homologous recombination between duplicated chromosomal sequences in mouse L cells Mol Cell Biol 1988 8 196 202 3122022
Flint SJ Berget SM Sharp PA Characterization of single-stranded viral DNA sequences present during replication of adenovirus types 2 and 5 Cell 1976 9 559 571 1009576 10.1016/0092-8674(76)90038-6
Epstein LH Young CS Adenovirus homologous recombination does not require expression of the immediate-early E1a gene J Virol 1991 65 4475 4479 1830115
Young CS Cachianes G Munz P Silverstein S Replication and recombination in adenovirus-infected cells are temporally and functionally related J Virol 1984 51 571 577 6471165
Deng C Capecchi MR Reexamination of gene targeting frequency as a function of the extent of homology between the targeting vector and the target locus Mol Cell Biol 1992 12 3365 3371 1321331
Scheerer JB Adair GM Homology dependence of targeted recombination at the Chinese hamster APRT locus Mol Cell Biol 1994 14 6663 6673 7935385
Fujitani Y Yamamoto K Kobayashi I Dependence of frequency of homologous recombination on the homology length Genetics 1995 140 797 809 7498755
Wang L Hernandez-Alcoceba R Shankar V Zabala M Kochanek S Sangro B Kramer MG Prieto J Qian C Prolonged and inducible transgene expression in the liver using gutless adenovirus: a potential therapy for liver cancer Gastroenterology 2004 126 278 289 14699506 10.1053/j.gastro.2003.10.075
Kreppel F Kochanek S Long-term transgene expression in proliferating cells mediated by episomally maintained high-capacity adenovirus vectors J Virol 2004 78 9 22 14671083 10.1128/JVI.78.1.9-22.2004
Stilwell JL McCarty DM Negishi A Superfine R Samulski RJ Development and characterization of novel empty adenovirus capsids and their impact on cellular gene expression J Virol 2003 77 12881 12885 14610209 10.1128/JVI.77.23.12881-12885.2003
Casadaban MJ Cohen SN Analysis of gene control signals by DNA fusion and cloning in Escherichia coli J Mol Biol 1980 138 179 207 6997493 10.1016/0022-2836(80)90283-1
Takahashi NK Kusano K Yokochi T Kitamura Y Yoshikura H Kobayashi I Genetic analysis of double-strand break repair in Escherichia coli J Bacteriol 1993 175 5176 5185 8349557
Huynh T Young R Davis R Glover D Constructing and Screening cDNA Libraries in λgt10 and λgt11 DNA cloning Volume I a practical approach 1984 Washinton DC: IRL PRESS 49 78
Naito Y Naito T Kobayashi I Selfish restriction modification genes: resistance of a resident R/M plasmid to displacement by an incompatible plasmid mediated by host killing Biol Chem 1998 379 429 436 9628334
Yanisch-Perron C Vieira J Messing J Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors Gene 1985 33 103 119 2985470 10.1016/0378-1119(85)90120-9
|
16219108
|
PMC1277836
|
CC BY
|
2021-01-04 16:39:08
|
no
|
Genet Vaccines Ther. 2005 Oct 12; 3:8
|
utf-8
|
Genet Vaccines Ther
| 2,005 |
10.1186/1479-0556-3-8
|
oa_comm
|
==== Front
Genet Vaccines TherGenetic Vaccines and Therapy1479-0556BioMed Central London 1479-0556-3-81621910810.1186/1479-0556-3-8ResearchA trial of somatic gene targeting in vivo with an adenovirus vector Ino Asami [email protected] Yasuhiro [email protected] Hiroyuki [email protected] Naofumi [email protected] Takao [email protected] Ichizo [email protected] Department of Medical Genome Sciences, Graduate School of Frontier Science, University of Tokyo & Institute of Medical Science, University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan2 Graduate Program in Biophysics and Biochemistry, Graduate School of Science the University of Tokyo3 Department of Environmental Information, Keio University, 5322 Endo, Fujisawa, Kanagawa 252-8520, Japan4 Laboratory of Gene Transfer and Regulation, National Institute of Biomedical Innovation, Asagi 7-6-8, Saito, Ibaraki, Osaka 567-0085, Japan5 Pharmaceuticals and Medical Devices Agency, Shin-Kasumigaseki Bldg. 3-3-2, Kasumigaseki, Chiyoda-ku, Tokyo 100-0013, Japan2005 12 10 2005 3 8 8 1 7 2005 12 10 2005 Copyright © 2005 Ino et al; licensee BioMed Central Ltd.2005Ino 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
Gene targeting in vivo provides a potentially powerful method for gene analysis and gene therapy. In order to sensitively detect and accurately measure designed sequence changes, we have used a transgenic mouse system, MutaMouse, which has been developed for detection of mutation in vivo. It carries bacteriophage lambda genome with lacZ+ gene, whose change to lacZ-negative allele is detected after in vitro packaging into bacteriophage particles. We have also demonstrated that gene transfer with a replication-defective adenovirus vector can achieve efficient and accurate gene targeting in vitro.
Methods
An 8 kb long DNA corresponding to the bacteriophage lambda transgene with one of two lacZ-negative single-base-pair-substitution mutant allele was inserted into a replication-defective adenovirus vector. This recombinant adenovirus was injected to the transgenic mice via tail-vein. Twenty-four hours later, genomic DNA was extracted from the liver tissue and the lambda::lacZ were recovered by in vitro packaging. The lacZ-negative phage was detected as a plaque former on agar with phenyl-beta-D-galactoside.
Results
The mutant frequency of the lacZ-negative recombinant adenovirus injected mice was at the same level with the control mouse (~1/10000). Our further restriction analysis did not detect any designed recombinant.
Conclusion
The frequency of gene targeting in the mouse liver by these recombinant adenoviruses was shown to be less than 1/20000 in our assay. However, these results will aid the development of a sensitive, reliable and PCR-independent assay for gene targeting in vivo mediated by virus vectors and other means.
==== Body
Background
Gene targeting, which is the precise alteration of genomic information by homologous recombination, has provided a powerful means of genetic analysis in microorganisms and mammalian systems [1]. In mouse systems, embryonic stem-cell lines modified in vitro can be used to generate mice that are altered at the germ-line level. If the gene targeting of somatic cells is made possible by gene transfer in vivo, it will facilitate the analysis of gene function, and provide a means of gene therapy for genetic and other diseases [2].
There are two major inherent problems with the use of gene targeting in vivo. First, its low efficiency makes it difficult to detect and analyze. A sensitive and accurate measurement system is therefore needed to detect such low-frequency events. Although there have been several reports of gene targeting in the rat liver with specifically designed oligonucleotides [3,4], their reproducibility remains controversial [5]. PCR-based detection methods might thus be inaccurate and prone to various artifacts. In order to detect and measure gene targeting in mice with sufficient sensitivity, we used a bacteriophage transgenic-mouse system, MutaMouse, which has been developed for the detection of mutagenesis in vivo (Figure 1) [6]. The MutaMouse carries tandem repeats of the bacteriophage lambda genome with the lacZ+ gene, in which the change to a lacZ-negative allele is detected after its in vitro packaging into viable bacteriophage particles.
The second major problem with gene targeting in vivo is that non-homologous recombination is much more frequent than homologous recombination in mammalian cells. Rare accurately modified cells are selected and purified in the case of embryonic stem cells that are treated in vitro. For gene targeting in vivo, imprecise modification would be detrimental for analytical uses and therapeutic purposes. Accurate gene modification has been achieved efficiently using replication-defective adenovirus vectors for gene delivery in vitro [7,8]. Fujita and colleagues used a mammalian plasmid as a model target [7]. The gene targeting was frequent (~10-4 per cell) and analysis of the products revealed that homologous recombination was more frequent than non-homologous recombination. One possible reason for this high accuracy was protection of the viral DNA by the terminal protein, which is covalently attached to the ends of the viral DNA and to other viral proteins during its transfer to the nucleus and target DNA. Breaks in unprotected DNA would lead to non-homologous recombination.
Figure 1 Experimental steps to detect gene targeting in vivo. Gene targeting in vivo in liver cells was attempted after the delivery of donor DNA with an adenovirus vector. The gene with the required sequence change (lacZ-) on the lambda transgene in the mouse will be detected after its recovery in bacteriophage particles. Only lacZ-negative mutants can form plaques under the selective conditions.
The adenovirus is useful for gene delivery in vivo because it has a broad host-range, is easy to prepare to a high titer and only rarely integrates into the host genome by non-homologous recombination [9,10]. To date, more than 170 clinical studies have used recombinant adenovirus vectors to express cDNA in humans [11]. Numerous adenovirus-infection experiments have been carried out with mice, and have established that the injection of adenovirus recombinants into the mouse tail-vein leads to the expression of their genes in approximately one-half of the liver cells [12,13].
In the present study, we investigated gene targeting in the mouse liver using a replication-defective adenovirus vector and a transgenic mouse system (Figure 1). Although our initial attempts did not detect the predicted gene targeting (the frequency of the expected recombinants was less than 1/20,000 per lambda genome), the strategy and methods detailed here will aid the development of virus-mediated gene targeting in vivo.
Materials and methods
Bacteria, bacteriophages and plasmids
The bacteria, bacteriophages and plasmids used in this study are listed together with details of their construction in Additional file 1.
BIK12001 was used for the titration of bacteriophage lambda and the measurement of lacZ-negative bacteriophage lambda by phenyl beta-D-galactoside (p-gal) selection (see below). BIK1564 was used for the growth of all bacteriophage lambda strains in this study. BIK2206 was used for confirmation of the LacZ-negative phenotype of the bacteriophage selected with p-gal using 5-bromo-4-chloro-3-indlyl-beta-D-galactose (X-gal).
The construction of the plasmids used in this study is detailed in additional file 1. The construction of pAdNY58 is also illustrated in Figure 2. The construction of pAdNY57 was as follows. The SmaI(1)-SacI fragment of LIA7 within the lacZ gene (Figure 2) was used to replace the shorter SmaI-SacI fragment of pUC18. The Glu461Gly mutation (Figure 3) was introduced into the resulting plasmid (pNY15) by site-directed mutagenesis using PCR [14] as follows. The PCR products generated with the primer pair LZG-U (5'-ACCGGCGATGAGCGAA-3') and LZG-MA (5'-GCCTGATCCATTCCCCAGCGACCA-3'), and the primer pair LZG-MS (5'-GGGAATGGATCAGGCCACGGCCGC-3') and LZG-D (5'-GGGCTGGTCTTCATCC-3'), were mixed and used as templates for the second round of PCR with the primer pair LZG-U and LZG-D. The MluI-BssHII fragment of the wild-type lacZ gene of pNY15 was replaced by the MluI-BssHII fragment of the PCR product. The targeted change in the resulting plasmid (pNY15G3.11) was confirmed by sequencing. pNY20 was produced by replacing the smaller SmaI-SacI fragment of pNY19 with the homologous SmaI-SacI fragment of pNY15G3.11, which carries the mutant sequence.
Figure 2 Construction of the recombinant adenovirus AdNY58. The bacteriophage lambda LIA7 was recovered from the MutaMouse by in vitro packaging. An SmaI-SacI fragment of LIA7 within its lacZ gene was inserted into pIK153. The Tyr105Stop mutation (Figure 3) was introduced into the resulting plasmid (pIK153LZS.6) using site-directed mutagenesis by PCR as follows. The PCR products generated with the primer pair LZT-U (5'-CGAAGAGGCCCGCAC-3') and LZT-MA (5'-TAATGGGCTAGGTTACGTTGGTGTAG-3'), and the primer pair LZT-MS (5'-TAACCTAGCCCATTACGGTCAATCC-3') and LZT-D (5'-GGCAACATGGAAATCGC-3') were mixed and used as templates for the second PCR with the primer pair LTZ-U and LZT-D. Replacement of an FspI-AatII fragment of pIK153LZS.6 by the FspI-AatII fragment of the resulting PCR product resulted in pIK153 T10.1. A BamHI-SmaI fragment covering the lacZ gene of LIA7 was inserted into the BamHI site of pIK153 (resulting in pNY19). pNY21 was made by replacing the smaller SmaI-SacI fragment of pNY19 with the homologous SmaI-SacI fragment of pIK153T10.1, which carries the mutant sequence. An XbaI-BglII fragment of pNY21 was used to replace the smaller XbaI-BamHI fragment of pHM5 (resulting in pNY58). pAdNY58 was made by replacement of the smaller I-CeuI-PI-SceI fragment of pAdHM4 with an I-CeuI-PI-SceI fragment of pNY58. The longer PacI fragment of pAdNY58 was transfected into 293 cells. The recombinant adenovirus AdNY58 was prepared and purified from the cell culture.
These two lacZ mutations were transferred back to lambda by homologous recombination in vivo [15] in order to generate LIA15 and LIA11, respectively. The recombinational transfer was carried out as follows. Cells of BIK12015 or BIK12018 were grown to OD600 = ~0.3 in LB (10 g bactotrypton, 5 g yeast extract and 10 g NaCl per liter) containing 20 μg/ml chloramphenicol, 0.2% maltose and 10 mM MgSO4. LIA7 was adsorbed onto the cells at a multiplicity of 1.0 at 37°C for 15 minutes. The mixture was shaken at 37°C until the OD600 dropped below 0.3. One drop of CHCl3 was added to the mixture, which was then shaken for 30 seconds. The mixture was centrifuged and the supernatant was recovered. The supernatant was assayed for BIK12001 on agar plates containing p-gal as detailed below. The plaques on the p-gal plates were isolated and analyzed for the designed sequence change by restriction of the PCR products (see Analysis of the mutant bacteriophage DNA).
Selection of lacZ-negative bacteriophage with p-gal
The lacZ-negative bacteriophage particles were detected using positive selection [15,16]. BIK12001 cells were grown with shaking at 37°C to OD600 = 1.0 in LB containing ampicillin (50 μg/ml), kanamycin (20 μg/ml) and 0.2% maltose. The culture was centrifuged at 3,500 rpm for 15 minutes at 4°C. The pellets were dissolved into one-half the volume of LB containing 10 mM MgSO4. The bacteriophage was adsorbed onto these cells at room temperature for 20 minutes. To estimate the total number of bacteriophages, 2.5 ml molten 1/4 LB top agar (5 g LB broth base (Gibco BRL, Rockville, MD, USA), 6.4 g NaCl and 7.5 g Bactoagar per liter) was added to 0.25 ml of the mixture of cells and bacteriophages, and the entire content was poured onto a 1/4 LB plate (5 g LB broth base, 6.4 g NaCl and 15 g Bactoagar per liter). To estimate the number of lacZ-negative bacteriophages, 2 ml of the mixture of cells and bacteriophages, and 22 ml of molten 1/4 LB top agar containing 0.3% p-gal (Sigma Chemical Co., MO, USA), were mixed and poured onto four 1/4 LB plates. The plates were incubated at 37°C for 12 hours.
Construction of recombinant adenoviruses
pNY56 was constructed by replacing the shorter XbaI-BamHI fragment of pHM5 by the XbaI-BglII fragment of pNY19 (Figure 2). pAdHM4 includes the entire genome of the recombinant adenovirus vector. The plasmid pAdNY56 was constructed by replacing the shorter I-CeuI-PI-SceI fragment of pAdHM4 by an I-CeuI-PI-SceI fragment of pNY56. The PacI fragment of pAdNY56 was transfected into cells of cell-line 293, which allows replication of the replication-defective adenoviruses. The recombinant adenovirus AdNY56 was prepared and purified as described previously [18]. Similarly, AdNY57 was constructed from pNY20 via pNY57 (Additional file 1), and AdNY58 was constructed from pNY21 via pNY58 (Figure 2, Additional file 1).
Adenovirus infection
Female MutaMice (7 weeks old) were obtained from Covance Research Products Inc. (Denver, PA, USA). The MutaMice were maintained under specific pathogen-free conditions in the animal faculty of the Institute of Medical Science at the University of Tokyo, Japan. After the animals were anesthetized with Nembutal (Dainippon Pharmaceutical Co., Osaka, Japan), 3 × 109 plaque-forming units (PFU) of the recombinant adenovirus in 200 μl of PBS (137 mM NaCl, 8.10 mM Na2HPO4, 2.68 mM KCl, 1.47 mM KH2PO4, 0.9 mM CaCl2, 0.33 mM MgCl2) was injected into the tail-vein of each mouse using a 30-gauge needle. AdNY56 was injected into one mouse, AdNY57 was injected into two mice and AdNY58 was injected into two mice.
Isolation of genomic DNA, recovery of lambda bacteriophage and measurement of mutant frequency
Twenty-four hours after injection, the mice were sacrificed. A lobe of the liver of each animal was excised, frozen by submersion in liquid nitrogen and stored in a 1.5-ml plastic tube at -80°C. Genomic DNA was isolated from the liver tissue with phenol-chloroform and precipitated by ethanol/sodium as described in the manual for MutaMouse. Lambda bacteriophage particles were recovered from the isolated DNA by incubation with packaging extracts (Mutaplax, Epicentre, WI, USA). The lacZ-negative mutants were detected by p-gal selection as described above. Each plaque on the selective agar was recovered in 100 μl of SM buffer (50 mM Tris-HCl (pH 7.5), 10 mM MgSO4, 100 mM NaCl and 0.01% gelatin). In order to verify the lacZ-negative phenotype, each isolate was assayed on agar with X-gal using a spot assay as follows. BIK2206 was grown in LB containing ampicillin (50 μg/ml) and tetracycline (10 μg/ml). Twice-concentrated culture (1.25 ml) was mixed with 6 ml molten LB/MM agar (100 ml LB medium, 0.75 g Bactoagar, 10 mM MgSO4, 0.2% maltose and 0.35 mg/ml X-gal) and spread on agar. A 10-μl aliquot of each bacteriophage sample was spotted onto these cells. The plates were incubated overnight at 37°C. The mutant frequency was estimated by dividing the number of PFU on the selective plate (as verified with X-gal) by the number of total PFU on 1/4 LB agar.
Analysis of the mutant bacteriophage DNA
The lacZ-negative lambda bacteriophage DNA from the mice was analyzed using restriction enzymes following PCR. For the lacZ-negative lambda DNA from the AdNY57-treated mouse, PCR was carried out with the primer pair LG-1 (5'-TACCGGCGATGAGCGAAC-3') and LG-2 (5'-CTCCAGGTAGCGAAAGCC-3'). The 288-bp product was purified by ethanol/sodium precipitation, digested with TfiI (New England Biolabs, Beverly, MA, USA) (recognition site, 5'-G|AWTC-3' (W = A or T)) at 65°C and analyzed using agarose electrophoresis. The mutant sequence was resistant to TfiI, while the wild-type sequence was sensitive, yielding 204 and 84 bp fragments. The primer pair Lam-1 (5'-TACTGTCGTCGTCCCCTC-3') and Lam-2 (5'-CGCAGATGAAACGCCGAGT-3') was used for the lacZ-negative lambda DNA from the AdNY58-treated mouse. The 213-bp PCR product was digested with XspI (Takara Bio Inc., Shiga, Japan) (recognition site, 5'-C|TAG-3') at 37°C and analyzed using agarose electrophoresis. The wild-type sequence was resistant to XspI, while the mutant sequence was sensitive, yielding 146 and 67 bp fragments.
Results
Experimental design for the detection of gene targeting in vivo
Figure 1 illustrates our experimental design for the sensitive detection of gene targeting in vivo. The MutaMouse carries approximately 40 copies of bacteriophage lambda gt10lacZ on a chromosome [6,19]. The single integration site is located in band C on chromosome 3 [20]. Our target sequence was the wild-type lacZ gene. The donor DNA was delivered to the liver cell nuclei by tail-vein injection of the recombinant adenovirus. Genomic DNA was isolated from the liver and its in vitro packaging allowed the recovery of the lambda genome in viable bacteriophage particles. A lacZ-negative mutant bacteriophage was selected as a plaque-former in an Escherichia coli mutant defective in the galE gene on an agar plate containing p-gal. This chemical is converted by the lacZ gene product (beta-galactosidase) into UDP-galactose, which accumulates in the absence of the GalE protein to induce cell death. The ratio of the mutant plaque-formers to the total plaque-formers was used to estimate the fraction of the mutated gene. The mutant gene was further analyzed using restriction enzymes.
Replication-defective recombinant adenoviruses constructed by an in vitro-ligation method were used to deliver the donor DNA [18,21]. Figure 3 shows the structure of the recombinant adenoviruses used in the present study (see Figure 2, Additional file 1, and Materials and methods for further details). An 8077-bp fragment of lambda gt10lacZ was inserted into the E1 deletion site of the mutant adenovirus [18,21]. AdNY56 had wild-type lacZ, while AdNY57 and AdNY58 had a point mutation in lacZ (Figure 3B).
Figure 3 Design for gene targeting and its detection. (A) The donor carrying the mutant lacZ gene is inserted into an adenovirus vector. The lacZ mutation will be transferred to the lacZ gene of the lambda transgene in the mouse genome. (B) Expected sequence changes and their detection using restriction analysis.
AdNY57 was constructed so as to introduce a point mutation at the active site of LacZ. The target sequence was the 5' GAA that codes for Glu461, which is essential for the activity of LacZ [22,23]. AdNY57 was expected to change its second base (that is, the 1437 th base) from A to G, thereby generating the Glu461Gly mutant, which shows a 76-fold decrease in activity [23]. The mutant and wild-type sequences can be distinguished using the restriction enzyme TfiI (Figure 3B).
AdNY58 was constructed so as to introduce a point mutation at the 5' TAT that codes for Tyr105. AdNY58 was expected to change its third base (that is, the 369th base) from T to G, thereby generating the Tyr105Stop mutant. The mutant and wild-type sequences can be distinguished using the restriction enzyme XspI (Figure 3B).
Control experiments
We demonstrated that lacZ mutants that were predicted to be generated by the recombinant adenovirus could be selected with p-gal as follows. Bacteriophage lambda strains carrying the mutations were produced by transferring each mutation on a plasmid back to lambda through homologous recombination in E. coli (as detailed in Materials and methods). The two bacteriophage strains, lambda gt10lacZ- Tyr105Stop (LIA11) and lambda gt10lacZ- Glu461Gly (LIA15), were then used in the p-gal selection. As shown in Table 1, lambda with wild-type lacZ showed a plaque-formation efficiency of less than 1/10,000 on the selective agar relative to that on the non-selective agar. By contrast, each of the mutant lambda strains showed similar or slightly decreased plaque-formation efficiency on the selective agar. We concluded that the expected targeted product with AdNY57 and AdNY58, if it was produced, should be selected and measured using the p-gal-selection procedure.
Table 1 Selection efficiency of lambda lacZ-negative mutants
Lambda Genotype Titer Titer on p-gal selective plate Relative plaque formation
LIA7 lacZ+ 2.2 × 1010 1.9 × 106 8.6 × 10-5
LIA11 lacZ- (Tyr105Stop) 9.6 × 1010 8.8 × 1010 9.2 × 10-1
LIA15 lacZ- (Glu461Gly) 1.4 × 1010 1.3 × 1010 9.1 × 10-1
Delivery of donor DNA and measurement of mutant frequency
The recombinant adenovirus particles (3 × 109 PFU in 200 μl of PBS) were injected into the tail-vein of a MutaMouse. It is well established that the adenovirus genome accumulates in the liver cell nuclei after tail-vein injection [12,13]. Most of the hepatocyte nuclei are expected to receive several copies of the adenovirus genome under these conditions (see Discussion). After 24 hours, the liver was excised from the MutaMouse, genomic DNA was isolated from the liver tissue and the lambda genome was recovered as a bacteriophage particle by in vitro packaging. The lacZ-negative phage was detected selectively on agar with p-gal. The plaques on these selective plates were isolated and the LacZ-negative phenotype was confirmed on agar plates containing X-gal. The mutant frequency was estimated as the fraction of the lacZ-negative phage (Table 2). The control mouse (animal number 0) received no injections.
Table 2 Detection of lacZ- phage
Packaging exp. RAd Genotype Animal number Packaging Total number of plaque formers lacZ- plaques Mutant Frequency Expected genotype
1 None Not relevant #0 Tube 1 4.5 × 104 4 8.9 × 10-5 n.t.
Tube 2 3.2 × 104 3 9.4 × 10-5 n.t.
Tube 3 8.5 × 104 8 9.4 × 10-5 n.t.
average 9.2 × 10-5
AdNY56 lacZ+ #1 Tube 4 8.5 × 104 8 9.4 × 10-5 n.t.
Tube 5 6.4 × 104 3 4.7 × 10-5 n.t.
Tube 6 8.8 × 104 1 1.1 × 10-5 n.t.
average 5.1 × 10-5
2 None Not relevant #0 Tube 7 4.9 × 104 5 10 × 10-5 n.t.
Tube 8 6.6 × 104 4 6.1 × 10-5 n.t.
Tube 9 5.1 × 104 10 20 × 10-5 n.t.
average 12 × 10-5
AdNY57 lacZ- (Glu461Gly) #2 Tube 10 3.8 × 104 6 16 × 10-5 0/6
Tube 11 3.0 × 104 7 23 × 10-5 0/7
Tube 12 4.5 × 104 9 20 × 10-5 0/9
average 20 × 10-5 total 0/22
3 None Not relevant #0 Tube 13 3.7 × 104 6 16 × 10-5 n.t.
Tube 14 6.0 × 104 5 8.3 × 10-5 n.t.
Tube 15 4.4 × 104 4 9.1 × 10-5 n.t.
average 11 × 10-5
AdNY57 lacZ- (Glu461Gly) #2 Tube 16 1.3 × 104 9 69 × 10-5 0/9
Tube 17 3.9 × 104 19 49 × 10-5 0/19
Tube 18 6.5 × 104 26 40 × 10-5 0/26
average 53 × 10-5 total 0/54
4 None Not relevant #0 Tube 19 2.6 × 105 8 8.5 × 10-5 n.t.
AdNY57 lacZ- (Glu461Gly) #3 Tube 20 1.6 × 105 5 6.3 × 10-5 0/5
Tube 21 4.1 × 105 9 8.6 × 10-5 0/9
average 1.5 × 10-5 total 0/14
5 None Not relevant #0 Tube 22 3.3 × 104 3 9.1 × 10-5 n.t.
AdNY58 lacZ- (Tyr105 Stop) #4 Tube 23 8.6 × 104 4 4.7 × 10-5 0/4
Tube 24 3.1 × 104 3 9.7 × 10-5 0/3
average 7.2 × 10-5 total 0/7
RAd: Recombinant adenovirus
n.t.: Not tested.
The mutant frequencies of the AdNY56-injected and control mice were similar (Table 2, Experiment 1), and did not differ significantly from those reported previously using this method (see [15] and the references cited therein). No significant increase in the mutant form of the gene was induced by injection of the recombinant adenovirus: the mutant frequency of the AdNY57- and AdNY58-injected mice was similar to that of the control mouse, which was approximately 1/10,000 (Table 2).
All of the lacZ-negative bacteriophages were purified and their lacZ genes were analyzed using restriction-enzyme treatment of the PCR products (Figure 4). As shown in Figures 3B and 4A, the PCR product of the Glu461Gly mutant, as predicted from the AdNY57 injection, could not be cut with TfiI. By contrast, the wild-type and most of the other possible mutants could be cut with TfiI. In fact, all of the lacZ-negative bacteriophages from the AdNY57-injected mouse were cleavable with this restriction enzyme. As shown in Figure 3B and 4B, the PCR product of the Tyr105Stop mutant, as predicted from the AdNY58 injection, could be cut with XspI. By contrast, the wild-type and most of the other mutants could not be cut with XspI. None of the lacZ-negative bacteriophages from the AdNY58-injected mice were cleavable with this restriction enzyme.
Figure 4 Restriction analysis of the lacZ-negative gene from mice treated with a recombinant adenovirus. (A) AdNY57-injected mouse. The PCR product of the lambda bacteriophage DNA with primers that flank the target site is 288 bp long. The wild-type PCR product is cut with TfiI into 84 and 204 bp fragments, whereas the Glu461Ala mutant PCR product is not cut. Lane M: Marker DNA prepared by HinfI digestion of the plasmid pUC19; 1–12, lacZ-negative bacteriophages from animal number 2; lacZ+: Lambda bacteriophage recovered from control mouse; lacZ-Glu461Gly: lambda bacteriophage LIA15. (B) AdNY58-injected mouse. The PCR product of the lambda bacteriophage DNA with primers that flank the target site is 213 bp long. The Tyr105Stop mutant PCR product is cut with XspI into 146 and 67 bp fragments, whereas the wild-type product is not. Lane M: Marker DNA prepared by HinfI digestion of plasmid pUC19; 1–4, lacZ-negative bacteriophages from animal number 3; lacZ+: Lambda bacteriophage recovered from control mouse; lacZ-Tyr105Stop: lambda bacteriophage LIA11.
We did not detect the expected gene replacement in any of the isolates. Moreover, the gene-correction frequency by these adenovirus constructs was shown to be less than 1/20,000 in the present system.
Discussion
Here we attempted to perform gene targeting in a transgenic mouse system that allowed the sensitive detection of mutagenesis by various agents, such as those directly interacting with DNA in the liver and other organs [24,25]. The limit of sensitivity in this system was 1/20,000 (see also [15]). This procedure might provide an alternative to the PCR-based assay for gene targeting in vivo, although our initial trials did not detect any of the expected recombinants.
In the present system, the sensitivity appeared to be limited by the high level of spontaneous mutagenesis in the target gene. The MutaMouse system was produced to detect mutagenesis at numerous sites within a gene, rather than to study gene targeting. Experimental designs involving the specific selection of homologous recombination events, such as those used in the previous work in vitro [7], would therefore be preferred.
Also, in the present system, a successful gene-targeting event would not be distinguishable in the phenotype of the mouse cell. In transgenic mice with a single copy of the mutant lacZ gene [26], correction to the wild-type gene would result in a direct positive readout in the mouse body (for example, through staining with dye). However, as the authors admit, it would be difficult to detect the targeting events with a high sensitivity. The presence of multiple copies of the target gene would improve the sensitivity because the lacZ+ allele is dominant over, and epistatic to, the lacZ- alleles with respect to the above phenotype. The MutaMouse carries multiple (approximately 40) copies of the target gene, which amount to 0.4% of the genome. This should be able to improve the sensitivity of detection of gene targeting, although the sensitivity is limited by spontaneous mutagenesis. In addition, the presence of tandem repeats might have other types of negative effect on gene targeting, as detailed below.
How efficient is adenovirus infection and delivery to the hepatocyte nucleus? Tail-vein injection is an established method for the delivery of adenovirus to liver cells. The average copy number of a replication-defective recombinant adenovirus genome per liver cell has been estimated as 14–28 copies using Southern hybridization after tail-vein injection of 5 × 109 PFU of the virus [12]. This corresponds to 40% of the injected adenovirus. Fluorescence in situ hybridization revealed that, after tail-vein injection of 2 × 109 PFU, all of the hepatocyte nuclei had 1–100 copies of a recombinant adenovirus genome, with an average of 20 copies [27]. After tail-vein injection of 2 × 108 PFU of a recombinant adenovirus with the lacZ expression cassette, 40% of the hepatocytes expressed beta-galactosidase [13]. We assumed that the majority of the liver cells received several copies of the adenovirus genome, at least sufficient for gene expression, after injecting 3 × 109 PFU in our experiment. (We cannot raise the titer any more because of the toxicity of the virus.) This type of information can be confirmed by Southern hybridization and fluorescence in situ hybridization.
The gene-targeting frequency with recombinant adenoviruses in vitro varies from ~10-7–10-4 per cell [7,8,28]. We did not detect any signal using recombinant adenovirus for gene delivery in the mouse liver. In order to achieve gene targeting in vivo using an adenovirus vector or any other means, it will be necessary to increase the frequency of gene targeting. So how can we achieve this goal?
The efficiency of gene targeting in vitro varies from one locus to another [29,30]. Such locus-dependence might reflect drastic effects of the chromatin structure on the frequency of homologous recombination [30,31]. Thus, the target transgene could be placed at a different locus that is known to be a hot spot in gene targeting in embryonic stem (ES) cells.
Repetitive sequences are methylated in the mouse genome [32]. Ikehata and colleagues suggested that the whole coding region of the MutaMouse lacZ transgene is methylated to a high degree at every CpG site [33]. One possible reason for this phenomenon is that the CpG content of the lacZ gene (9%) [34] is much higher than the average CpG content of the mouse genome (~1%) [35]. Methyl-CpG binding protein 2 (MeCP2) might bind to methylated CpG and somehow compact chromatin [36]. Furthermore, Manuelidis analyzed the structure of a mouse chromosome bearing a huge (~11 Mb) insert of a tandem-repeated transgene (~1,000 copies) [37]. This transgene was localized on an arm of chromosome 3 at a distance from the centromere. According to Manuelidis, the transgene is heterochromatic and highly condensed. Therefore, the MutaMouse transgene might be heterochromatic. The accessibility of nucleases to the heterochromatic structure is lower than that of euchromatin [38,39]. Reducing the copy number of the transgene and/or using another transgene that is lower in CpG content might increase gene targeting, although the decrease in copy number might affect the sensitivity of detection. An important experiment that can be done is to test whether the coding region of the MutaMouse lacZ transgene is really heterochromatic, using, for example, CHIP assay with the antibody against the methylated histones and PCR primers on the lacZ genes.
Chromosome replication is known to stimulate homologous recombination. Partial hepatectomies in mice might stimulate liver cell proliferation and DNA replication, which in turn might stimulate recombination. Hara et al. (1999) reported that partial hepatectomies increased mutagenesis with N-ethyl-N-nitrosourea, which is a direct-acting DNA-ethylation agent, in the MutaMouse [40].
It might be easier to modify the donor DNA than the recipient DNA. One can generate recombinogenic damage on the donor DNA. Irradiating adenovirus particles with ultraviolet light of 1500 J/m2 resulted in an approximately three-fold increase in their mutual homologous recombination [41]. Recombinogenic cross-links are induced by some mutagens, such as psoralens, cisplatin (cis-diamminedichloroplatinum) and mitomycin C [42]. Such agents, both mutagenic and recombinogenic, might be suitable for gene targeting in vivo if they are shown to be active in mutagenesis in a transgenic-reporter mouse system. The effect of such recombinogenic damage might be much larger with replication-defective adenovirus recombinants than with replication-competent adenoviruses, because their replication-intermediates are responsible for their high recombination frequency [41,43-45].
The gene-targeting frequency is strongly dependent on the length of homology; the frequency increases as the homology length increases up to 10 kb [46-48]. If the deviation from this rule above 10 kb is due to the shearing and/or degradation of longer DNA after electroporation in embryonic stem cells, donor DNAs that are protected by the DNA binding proteins in the adenovirus particle might show greater length dependence over a wider range of values. Adenoviral vectors with a larger capacity for inserts, which are known as high-capacity 'gutless' vectors [49-51] might therefore be suitable for use in this approach.
Conclusion
Here we attempted to perform gene targeting in a transgenic mouse system that allowed the sensitive detection of mutagenesis. The frequency of gene targeting in the mouse liver by these recombinant adenoviruses was shown to be less than 1/20000 with the sensitive and PCR-independent detection system.
List of abbreviations
PCR, polymerase chain reaction; PFU, plaque-forming unit; RFLP, restriction fragment length polymorphism; p-gal, phenyl-beta-D-galactoside; X-gal, 5-bromo-4-chloro-3-indlyl-beta-D-galactose
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AI carried out the injection of the recombinant adenovirus and the analysis of the mouse DNA. YN and HM constructed the recombinant adenovirus. NH injected the recombinant adenovirus to the mouse. YN constructed the experimental design as well as cloning of the part of lambda DNA from the MutaMouse genomic DNA. IK provided the original experimental idea and coordinated the experimental design. All authors read and approved the final manuscript.
Supplementary Material
Additional file 1
Bacterial strains, plasmids, bacteriophage strains and recombinant adenovirus constructs.
Click here for file
Acknowledgements
Ms. Kuniko Iwasaki and Dr. Ryuichi Miura from the Laboratory Animal Research Center of the Institute of Medical Science, Japan, guided us in our manipulation of the mice. Dr. Noriko Takahashi from our laboratory helped with the maintenance of the mice. Dr. Yoichiro Iwakura of the Institute of Medical Science provided critical comments on an early version of the manuscript. This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (No.0828102: General Mechanisms of DNA Recombination Repair. 1996–1999) and the Japan Owners Association (JOA) (1999–2002) as arranged by the Japan Society for Gene Therapy.
==== Refs
Capecchi MR Altering the genome by homologous recombination Science 1989 244 1288 1292 2660260
Yanez RJ Porter AC Therapeutic gene targeting Gene Ther 1998 5 149 159 9578833 10.1038/sj.gt.3300601
Kren BT Bandyopadhyay P Steer CJ In vivo site-directed mutagenesis of the factor IX gene by chimeric RNA/DNA oligonucleotides Nat Med 1998 4 285 290 9500600 10.1038/nm0398-285
Kren BT Parashar B Bandyopadhyay P Chowdhury NR Chowdhury JR Steer CJ Correction of the UDP-glucuronosyltransferase gene defect in the gunn rat model of crigler-najjar syndrome type I with a chimeric oligonucleotide Proc Natl Acad Sci USA 1999 96 10349 10354 10468611 10.1073/pnas.96.18.10349
Taubes G Gene therapy. The strange case of chimeraplasty Science 2002 298 2116 2120 12481116 10.1126/science.298.5601.2116
Gossen JA de Leeuw WJ Tan CH Zwarthoff EC Berends F Lohman PH Knook DL Vijg J Efficient rescue of integrated shuttle vectors from transgenic mice: a model for studying mutations in vivo Proc Natl Acad Sci USA 1989 86 7971 7975 2530578
Fujita A Sakagami K Kanegae Y Saito I Kobayashi I Gene targeting with a replication-defective adenovirus vector J Virol 1995 69 6180 6190 7666520
Mitani K Wakamiya M Hasty P Graham FL Bradley A Caskey CT Gene targeting in mouse embryonic stem cells with an adenoviral vector Somat Cell Mol Genet 1995 21 221 231 8525428 10.1007/BF02255777
Doerfler W The fate of the DNA of adenovirus type 12 in baby hamster kidney cells Proc Natl Acad Sci USA 1968 60 636 643 5248820
Harui A Suzuki S Kochanek S Mitani K Frequency and stability of chromosomal integration of adenovirus vectors J Virol 1999 73 6141 6146 10364373
Volpers C Kochanek S Adenoviral vectors for gene transfer and therapy J Gene Med 2004 6 S164 171 14978759 10.1002/jgm.496
Vrancken Peeters MJ Perkins AL Kay MA Method for multiple portal vein infusions in mice: quantitation of adenovirus-mediated hepatic gene transfer Biotechniques 1996 20 278 285 8825158
Nakatani T Kuriyama S Tominaga K Tsujimoto T Mitoro A Yamazaki M Tsujinoue H Yoshiji H Nagao S Fukui H Assessment of efficiency and safety of adenovirus mediated gene transfer into normal and damaged murine livers Gut 2000 47 563 570 10986218 10.1136/gut.47.4.563
Cormack B Ausubel MF, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struho K Mutagenesis by the polymerase chain reaction Current protocols in moecular biology 1991 1 John Wiley and Sons, Inc 8.5.1 8.5.9
Ino A Yamamoto S Kaneda Y Kobayashi I Somatic gene targeting with RNA/DNA chimeric oligonucleotides: an analysis with a sensitive reporter mouse system J Gene Med 2004 6 1272 1280 15459966 10.1002/jgm.617
Dean SW Myhr B Measurement of gene mutation in vivo using Muta Mouse and positive selection for lacZ- phage Mutagenesis 1994 9 183 185 7934957
Gossen JA Vijg J A selective system for lacZ-phage using a galactose-sensitive E.coli host Biotechniqnics 1993 14 326 330
Mizuguchi H Kay MA Efficient construction of a recombinant adenovirus vector by an improved in vitro ligation method Hum Gene Ther 1998 9 2577 2583 9853524 10.1089/10430349850019418
Blakey DH Douglas GR Huang KC Winter HJ Cytogenetic mapping of lambda gt10 lacZ sequences in the transgenic mouse strain 40.6 (Muta Mouse) Mutagenesis 1995 10 145 148 7603331
Swiger RR Myhr B Tucker JD The LacZ transgene in MutaMouse maps to chromosome 3 Mutat Res 1994 325 145 148 7527906 10.1016/0165-7992(94)90077-9
Mizuguchi H Kay MA A simple method for constructing E1- and E1/E4-deleted recombinant adenoviral vectors Hum Gene Ther 1999 10 2013 2017 10466635 10.1089/10430349950017374
Jacobson RH Zhang XJ DuBose RF Matthews BW Three-dimensional structure of beta-galactosidase from E. coli Nature 1994 369 761 766 8008071 10.1038/369761a0
Cupples CG Miller JH Effects of amino acid substitutions at the active site in Escherichia coli beta-galactosidase Genetics 1988 120 637 644 2906303
Cosentino L Heddle JA A comparison of the effects of diverse mutagens at the lacZ transgene and Dlb-1 locus in vivo Mutagenesis 1999 14 113 119 10474832 10.1093/mutage/14.1.113
Krebs O Favor J Somatic and germ cell mutagenesis in lambda lacZ transgenic mice treated with acrylamide or ethylnitrosourea Mutat Res 1997 388 239 248 9057886
Nickerson HD Colledge WH A LacZ-based transgenic mouse for detection of somatic gene repair events in vivo Gene Ther 2004 11 1351 1357 15229632 10.1038/sj.gt.3302311
Schowalter DB Tubb JC Liu M Wilson CB Kay MA Heterologous expression of adenovirus E3-gp19K in an E1a-deleted adenovirus vector inhibits MHC I expression in vitro, but does not prolong transgene expression in vivo Gene Ther 1997 4 351 360 9176522 10.1038/sj.gt.3300398
Wang Q Taylor MW Correction of a deletion mutant by gene targeting with an adenovirus vector Mol Cell Biol 1993 13 918 927 8423811
Bird D Bradshaw R Gene targeting is locus dependent in the filamentous fungus Aspergillus nidulans Mol Gen Genet 1997 255 219 225 9236780 10.1007/s004380050492
Petes TD Meiotic recombination hot spots and cold spots Nat Rev Genet 2001 2 360 369 11331902 10.1038/35072078
Ohta K Shibata T Nicolas A Changes in chromatin structure at recombination initiation sites during yeast meiosis Embo J 1994 13 5754 5763 7988571
Lau S Jardine K McBurney MW DNA methylation pattern of a tandemly repeated LacZ transgene indicates that most copies are silent Dev Dyn 1999 215 126 138 10373017 10.1002/(SICI)1097-0177(199906)215:2<126::AID-DVDY5>3.0.CO;2-8
Ikehata H Takatsu M Saito Y Ono T Distribution of spontaneous CpG-associated G:C --> A:T mutations in the lacZ gene of Muta mice: effects of CpG methylation, the sequence context of CpG sites, and severity of mutations on the activity of the lacZ gene product Environ Mol Mutagen 2000 36 301 311 11152563 10.1002/1098-2280(2000)36:4<301::AID-EM6>3.0.CO;2-R
Chevalier-Mariette C Henry I Montfort L Capgras S Forlani S Muschler J Nicolas JF CpG content affects gene silencing in mice: evidence from novel transgenes Genome Biol 2003 4 R53 12952532 10.1186/gb-2003-4-9-r53
Takai D Jones PA Comprehensive analysis of CpG islands in human chromosomes 21 and 22 Proc Natl Acad Sci USA 2002 99 3740 3745 11891299 10.1073/pnas.052410099
Bowen NJ Palmer MB Wade PA Chromosomal regulation by MeCP2: structural and enzymatic considerations Cell Mol Life Sci 2004 61 2163 2167 15338046 10.1007/s00018-004-4177-x
Manuelidis L Heterochromatic features of an 11-megabase transgene in brain cells Proc Natl Acad Sci USA 1991 88 1049 1053 1992455
Vega-Palas MA Venditti S Di Mauro E Heterochromatin organization of a natural yeast telomere. Changes of nucleosome distribution driven by the absence of Sir3p J Biol Chem 1998 273 9388 9392 9545262 10.1074/jbc.273.16.9388
Ye F Signer ER RIGS (repeat-induced gene silencing) in Arabidopsis is transcriptional and alters chromatin configuration Proc Natl Acad Sci USA 1996 93 10881 10886 8855276 10.1073/pnas.93.20.10881
Hara T Sui H Kawa Kami K Shimada Y Shibuya T Partial hepatectomy strongly increased the mutagenicity of N-ethyl-N-nitrosourea in Mutamouse liver Environ Mol Mutagen 1999 34 121 123 10529735 10.1002/(SICI)1098-2280(1999)34:2/3<121::AID-EM10>3.0.CO;2-R
Rainbow AJ Castillo JE Homologous recombination of adenovirus DNA in mammalian cells: enhanced recombination following UV-irradiation of the virus Mutat Res 1992 274 201 210 1380653
Wang YY Maher VM Liskay RM McCormick JJ Carcinogens can induce homologous recombination between duplicated chromosomal sequences in mouse L cells Mol Cell Biol 1988 8 196 202 3122022
Flint SJ Berget SM Sharp PA Characterization of single-stranded viral DNA sequences present during replication of adenovirus types 2 and 5 Cell 1976 9 559 571 1009576 10.1016/0092-8674(76)90038-6
Epstein LH Young CS Adenovirus homologous recombination does not require expression of the immediate-early E1a gene J Virol 1991 65 4475 4479 1830115
Young CS Cachianes G Munz P Silverstein S Replication and recombination in adenovirus-infected cells are temporally and functionally related J Virol 1984 51 571 577 6471165
Deng C Capecchi MR Reexamination of gene targeting frequency as a function of the extent of homology between the targeting vector and the target locus Mol Cell Biol 1992 12 3365 3371 1321331
Scheerer JB Adair GM Homology dependence of targeted recombination at the Chinese hamster APRT locus Mol Cell Biol 1994 14 6663 6673 7935385
Fujitani Y Yamamoto K Kobayashi I Dependence of frequency of homologous recombination on the homology length Genetics 1995 140 797 809 7498755
Wang L Hernandez-Alcoceba R Shankar V Zabala M Kochanek S Sangro B Kramer MG Prieto J Qian C Prolonged and inducible transgene expression in the liver using gutless adenovirus: a potential therapy for liver cancer Gastroenterology 2004 126 278 289 14699506 10.1053/j.gastro.2003.10.075
Kreppel F Kochanek S Long-term transgene expression in proliferating cells mediated by episomally maintained high-capacity adenovirus vectors J Virol 2004 78 9 22 14671083 10.1128/JVI.78.1.9-22.2004
Stilwell JL McCarty DM Negishi A Superfine R Samulski RJ Development and characterization of novel empty adenovirus capsids and their impact on cellular gene expression J Virol 2003 77 12881 12885 14610209 10.1128/JVI.77.23.12881-12885.2003
Casadaban MJ Cohen SN Analysis of gene control signals by DNA fusion and cloning in Escherichia coli J Mol Biol 1980 138 179 207 6997493 10.1016/0022-2836(80)90283-1
Takahashi NK Kusano K Yokochi T Kitamura Y Yoshikura H Kobayashi I Genetic analysis of double-strand break repair in Escherichia coli J Bacteriol 1993 175 5176 5185 8349557
Huynh T Young R Davis R Glover D Constructing and Screening cDNA Libraries in λgt10 and λgt11 DNA cloning Volume I a practical approach 1984 Washinton DC: IRL PRESS 49 78
Naito Y Naito T Kobayashi I Selfish restriction modification genes: resistance of a resident R/M plasmid to displacement by an incompatible plasmid mediated by host killing Biol Chem 1998 379 429 436 9628334
Yanisch-Perron C Vieira J Messing J Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors Gene 1985 33 103 119 2985470 10.1016/0378-1119(85)90120-9
|
16232311
|
PMC1277837
|
CC BY
|
2021-01-04 16:36:49
|
no
|
Harm Reduct J. 2005 Oct 18; 2:21
|
latin-1
|
Harm Reduct J
| 2,005 |
10.1186/1477-7517-2-21
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-601620972010.1186/1477-7525-3-60ResearchEthnic differential item functioning in the assessment of quality of life in cancer patients Pagano Ian S [email protected] Carolyn C [email protected] Cancer Research Center of Hawaii, Honolulu, HI 96813, USA2005 7 10 2005 3 60 60 14 6 2005 7 10 2005 Copyright © 2005 Pagano and Gotay; licensee BioMed Central Ltd.2005Pagano and Gotay; 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
Past research has shown that Filipino cancer patients report lower levels of quality of life (QoL) than other ethnic groups. One possible explanation for this is that Filipinos do not define QoL in the same manner as others, resulting in bias in their assessments. Hence, Filipinos would not necessarily have lower QoL.
Methods
Item response theory methods were used to assess differential item functioning (DIF) in the quality of life (measured by the EORTC QLQ-C30) of cancer patients across four ethnic groups (Caucasian, Filipino, Hawaiian, and Japanese). The sample consisted of 359 cancer patients.
Results
Results showed the presence of DIF on several items, indicating ethnic differences in the assessment of quality of life. Relative to the Caucasian and Japanese groups, items related to physical functioning, cognitive functioning, social functioning, nausea and vomiting, and financial difficulties exhibited DIF for Filipinos. On these items Filipinos exhibited either higher or lower QoL scores, even though their overall QoL was the same.
Conclusion
This evidence may explain why Filipinos have previously been found to have lower overall QoL. Although Filipinos score lower on QoL than other groups, this may not reflect lower QoL, but rather differences in how QoL is defined. The presence of DIF did not appear, however, to alter the psychometric properties of the QLQ-C30.
differential item functioningethnicityitem response theoryquality of life
==== Body
Background
In recent years medical researchers have shown increasing interest in the physical, psychological, and social health of individuals suffering from disease and treatment-related toxicity [1-3]. These broad characteristics are generally grouped under the inclusive heading quality of life (QoL), and offer a contrast to the more traditional biomedical markers, such as survival time or disease remission. A general definition of QoL is patients' perspectives on their ability to live useful and fulfilling lives, as influenced by, but not completely dependent on disease and treatment [1]. As an instrument of measurement in the clinical setting, QoL is defined functionally by patients' own perceptions of their performance in physical, occupational, psychological, social, financial, and somatic (i.e., physical symptomatology) areas [4,5].
The QoL construct is an important one with respect to measuring disease progress and treatment effectiveness. Because treatments for health conditions often have both short and long-term sequelae, including pain, fatigue, and depression, biomedical markers often fail to give a complete picture of patient status. Even when a person is disease free, the individual may still suffer from debilitating physical and mental anguish. Similarly, a shorter survival prognosis may be less onerous if the time remaining can be lived with enjoyment. Numerous areas of medical research, including heart disease, diabetes, arthritis, pharmacology, mental disorders, aging, and trauma are now examining QoL.
Cancer research is one area in particular that has shown an increased amount of QoL studies. The assessment of sequelae resulting from cancer therapies is an essential part of the cancer treatment process. One reason for this is that cancer treatments often involve therapies such as chemotherapy and radiation, which have high toxicities. Studies have shown that patients often experience fatigue [6], pain, sleep difficulties [7], depression [8,9], and sexual dysfunction [10], both during and after cancer treatment. Clearly, the impact that cancer has on QoL, both from the disease itself as well as its treatment, is significant. As a result, many prominent and important groups have expressed the need for QoL measures. Among these are international cancer institutes and societies [11], clinical trial groups [12], regulatory agencies [13], and the pharmaceutical industry [14].
While few would deny the importance of QoL outcomes in addition to biomedical markers, QoL assessment is more challenging. Because QoL is subjective [15], its assessment almost invariably requires patient ratings for measurement, and the outcome measures of interest are certain to contain measurement error. This measurement error can lead to invalid assessments of QoL. Therefore, questionnaire development and continuing validation are essential for the evaluation of QoL.
Although research suggests that that reliable data for QoL exists, potential differences between ethnic groups have not been fully examined. These can affect people's responses and thus estimates of reliability and validity [16-19]. Assessment of QoL is primarily reflective of Western concepts of illness, and it is not known whether these concepts are consistent in other communities (e.g., African, Asian, or Pacific Islander). In the Western view, illness is perceived primarily as an external disturbance that prevents an otherwise self-determined life course. However, many cultures do not share this perspective. Fatalism, karma, and cultural predeterminism play vital roles in their belief systems and illness is considered to be an integral part of one's life journey [5]. These contrasting views of illness are likely to create different perceptions of QoL in relation to illness. Therefore, cross-cultural studies are needed to provide a more complete picture of the multifaceted QoL construct [20].
In one multiethnic study, Gotay et al. [21] found that Filipinos have lower QoL, as measure by the QLQ-C30, when compared to Japanese and Caucasians. This finding persisted even after the effects of cancer stage, comorbidity, treatment, age, education, marital status, and place of birth had been controlled. Ethnicity explained as much as 8% of the variance in QoL after controlling for these factors. There are several possible causes for this ethnic difference. One is that there is a characteristic of either Filipino culture or Filipino genetics that results in lower QoL. If so, it is important to establish what it is, so that improvements could hopefully be made. However, another potential cause has to do with how QoL is measured. It is possible that the definition of QoL is not the same for Filipinos as it is for the Caucasian and Japanese groups, and that measured ethnic differences are not reflective of true differences in QoL.
An important but often overlooked aspect of questionnaire validation is the evaluation of differential item functioning (DIF, formerly called item bias). DIF occurs when one group of individuals responds differently from another group on a given questionnaire item, even though both groups are equivalent on the underlying construct that is assessed. For example, assume that the underlying construct assessed by the Scholastic Aptitude Test (SAT) is equivalent across males and females, but on the instrument there exists a question (an item) that is answered correctly more often by women than by men. This item is exhibiting DIF with respect to gender: It is biased against males in favor of females.
Confusion often exists regarding the use of the term DIF as opposed to item bias. When all of the items on a questionnaire are measuring the same global construct, and two groups have the same overall ability (i.e., the same average level of the construct of interest), DIF and item bias are equivalent. However, if the groups being compared are not equal in the underlying construct being measured, items exhibiting DIF are not necessarily biased and biased items will not necessarily show DIF. For example, suppose that women are of higher ability than men with respect to the construct assessed by the SAT. If no attempt were made to statistically control for overall ability, all items would be expected to show DIF with women outperforming men. Therefore, items reflecting DIF with women showing greater ability would not be biased, and items showing men and women as equal (no DIF) would be biased in favor of men. For this reason, it is important to statistically control (see below) for overall ability before assessing DIF.
When DIF is present, a questionnaire's validity becomes questionable and its generalizability is reduced. It may be more valid for one group of individuals than another. With SAT scores, this would imply that two people of equal ability, but from different groups, would not receive the same score. Although traditionally most research in DIF has been conducted involving achievement data such as SAT scores (the Educational Testing Service, ETS, began examining and removing items based on DIF with respect to ethnicity in 1986), the same principles apply to medical data involving QoL. When assessing QoL, DIF implies that two people with the same underlying QoL, who are from different groups, will not receive equivalent scores.
Research has shown that ethnicity does lead to potential DIF problems in quality of life [22,23], and on cognitive screening tests [24]. In comparing Danish and U.S. samples, Bjorner et al. [23] found DIF on 12 out of 35 items on the SF-36 Health Survey. Johnson et al. [22] looked at three QoL-related measures, the Sickness Impact Profile, the Ferrans and Powers' Quality of Life Index, and the Adult Self-Image Scales, and found that African-Americans had lower functional and affective scores when compared to Caucasians. Teresi et al. [24] found that an item related to remembering was less likely to be endorsed by Latinos than by Caucasians or African-Americans.
If ethnic DIF is present in quality of life, it might imply that the existence of ethnic differences in QoL is a result of biased items, and not real differences in the population. This would have implications for the assessment of QoL when working with multi-ethnic samples. In addition, reliability and validity measures obtained from previous studies may be suspect.
The primary goal of the present study was to establish whether or not DIF existed across ethnic groups for any of the items on the QLQ-C30. If DIF was found to exist, a second goal was to establish how this might have affected assessments of the instrument's psychometric properties. In this study conducted with patients living in Hawai'i, it was hypothesized that there would be a tendency for items to be biased with respect to Filipinos when compared to Japanese and Caucasians. That is, on certain items Filipinos would respond with either higher or lower QoL scores, even though their overall QoL was the same. If supported, this would suggest that ethnic differences in QoL, as measured by the QLQ-C30, do not necessarily reflect real QoL differences, but rather differences in how ethnic groups define QoL.
Methods
Participants
The study sample consisted of 366 cancer patients, 56% of the 646 eligible patients who were invited to participate. The most frequent reasons for nonparticipation were not feeling well enough to take part and being "not interested." Of the participants, 56% were women, 70% were married, 40% had a high school education or less, and the mean age was 62 years (standard deviation = 12.7). Ethnic breakdowns were 129 (35%) Japanese, 124 (34%) Caucasian, 61 (17%) Filipino, 42 (11%) Hawaiian, and 10 (3%) unknown. The most common cancer site was the breast (34%), followed by the prostate (28%). Most patients had received surgical treatment (83%), with several receiving radiation (42%), chemotherapy (20%), or hormonal treatment (25%).
Participants were identified through registrations on the Hawai'i Tumor Registry (HTR), a member of the National Cancer Institute-supported Surveillance, Epidemiology, and End Results Registry, which maintains records for all cancers diagnosed in the state. Eligibility criteria were histologic confirmation of any kind of cancer diagnosed between four and six months previously; ability to understand English; permission of primary physician; Oahu residency; Caucasian, Filipino, Hawaiian, or Japanese ethnic origin; and 18 years of age or older. Participation was not limited by stage or site of disease, but not all cancer sites were represented (e.g., no patients had colorectal, head and neck, lung, or ovarian cancer).
Procedures
Permission to approach patients for this study was obtained from the attending physician before patients were contacted. Patients received a letter introducing the study's intent, followed by a telephone call to set appointments. Data were collected by interviews, most often at the patient's home. (In some circumstances the patients preferred to be interviewed at the Cancer Research Center of Hawaii.) Data were also abstracted from the chart for age, ethnicity, sex, marital status, and site and stage of cancer. Asking patients to indicate the ethnicity of their four grandparents verified patient ethnicity. Standard state criteria were used. Three of four grandparents from the same ethnic group defined a patient's ethnicity, except for Hawaiians, for whom any grandparent being Hawaiian superseded other ethnic classifications and resulted in the patient as being defined as Hawaiian. Interviews were conducted by one of four female research associates, all of whom had completed graduate work in social sciences as well as extensive training in interviewing cancer patients. Interviews took an average of one hour.
Measures
During the participant interviews the questions from the EORTC QLQ-C30 version 1.0 were self-administered [3]. The questionnaire consists of 30 items each written to assess aspects of QoL (see Appendix [additional file 1]). The items are grouped into five functional scales (physical, role, cognitive, emotional, and social), three symptom scales (pain, fatigue, and nausea / vomiting), one global health status scale, five symptom items (dyspnea, insomnia, appetite loss, constipation, and diarrhea) and one financial difficulties item. Responses are either dichotomous (yes or no) for the physical and role scales, or Likert-type for the others.
Additionally, questions assessing Karnofsky Performance Status (KPS) [25] and receipt of chemotherapy were asked. This information was obtained to assess the correlations with the QLQ-C30 as a validity measure. The KPS has been frequently employed in clinical trials and QoL research [26], and assesses a person's ability to perform normal activities. The scale ranges from zero (deceased) to 100 (no impairment). The KPS is generally rated by an observer, and in this study, the interviewers were trained to provide a KPS score for each patient. Chemotherapy is a treatment known to have a significant impact on QoL and was recorded as either receipt of chemotherapy (coded one) or no chemotherapy received (coded zero).
Data Analysis
Several methods have been developed for measuring DIF [27], of which two major categories exist: classical and modern. Modern psychometrics is mostly defined by item response theory (IRT), and this technique provides important advantages over the classical methods (commonly referred to as classical test theory, or simply CTT) [28-32]. When the assumption of unidimensionality (i.e., a single latent trait is influencing the items) is met, invariance and information are the two primary advantages of IRT over CTT modeling. Under invariance, item difficulties (see below) are independent of the sample (i.e., independent of overall ability) and person abilities are independent of the items (i.e., independent of the particular items responded to). IRT models have the invariance property because both item and person characteristics are estimated simultaneously within the model. Information refers to the item information function (IIF) that is computed for each item through IRT modeling. The IIF indicates how well an item discriminates between persons at varying ability levels.
Because of these advantages, an IRT approach has been adopted for the present analyses. Specifically, a one-parameter (Rasch) model using the marginal maximum likelihood procedure has been employed. The one-parameter model was chosen over the two-parameter model because it provides more invariant estimates of the item difficulties and requires fewer cases [33]. In Rasch modeling the slope (discrimination) parameter is set to 1.0 for all of the items. Detailed explanations of the one- and two-parameter IRT models, and maximum likelihood estimation can be found in the literature [28,30-34].
Even though Rasch modeling requires fewer cases than more complex IRT models, the number of individuals available in our sample was still less than desired. In IRT modeling, the general standard of reliability is to have 99% confidence that the parameter estimates are within one half logit of the stable value. To meet this standard using Rasch models, a minimum sample of 108 is recommended [35]. The Japanese and Caucasian groups are above this minimum, but the Hawaiian and Filipino groups are not. However, even with the smallest group, Hawaiian, the confidence levels obtained are still within an interpretable range. The Hawaiian and Filipino groups are both large enough to have 99% confidence that the parameter estimates are within one logit of the stable value [35]. Therefore, while extra caution is needed when interpreting the analyses in the Hawaiian and Filipino groups, potentially meaningful results can still be obtained.
Because the purpose of this study was to assess DIF resulting from ethnic differences, the data were grouped into four ethnic categories: Caucasian, Japanese, Filipino, and Hawaiian. IRT models were calculated for each ethnic group using the PARSCALE software application. The entire QLQ-C30 was used for the IRT models and not subcategories based on functional and symptom scales. Although Cella et al. [1] recommended using subcategories, the present research was interested in DIF with respect to the higher-order QoL construct and not the specific subcategories. The presence of this single higher-order construct is supported by research conducted by Gotay et al. [36] that employed confirmatory factor analysis.
Using the above IRT model within each ethnic group, an ability score was obtained for each participant (recall that ability refers to the level of the construct assessed, in this case QoL) and a difficulty score was obtained for each item. The term difficulty is borrowed from IRT models used to assess item difficulties for educational testing. More difficult items are ones that are less likely to be answered correctly. When assessing QoL there is clearly no right or wrong answer. However, even though difficulty may seem an inappropriate term when assessing QoL, it still in fact conveys the same type of information.
For an educational testing measure, more difficult items are those that require a higher ability in order to be answered in the direction of higher ability (i.e., correctly). For a quality of life measure, more difficult items are those that require a higher QoL in order to be answered in the direction of higher QoL. For example, question 2 from the QLQ-C30 asks, "Do you have any trouble taking a long walk?" and question 3 asks, "Do you have any trouble taking a short walk outside of the house?" It is reasonable to suggest that a person would need a higher QoL to take a long walk than to take a short one. Therefore, it is more difficult to provide a "no" response to question 2 than it is for question 3.
After obtaining the item difficulties, group differences that were consistent across all items (and were assumed to be reflective of true differences in the population) were statistically controlled by standardizing the item difficulties within each ethnic group [32]. This ensured that the means and standard deviations were equivalent across groups, and that the model parameters were on a common scale. Hence, items found to exhibit DIF reflected the presence of bias, and not actual group differences. Using the item difficulties, intraclass correlations (note that because the item difficulties were standardized, the intraclass correlations are equivalent to Pearson correlations) were calculated between each ethnic group. Two groups are considered to be of the same type (i.e., having equivalent rank order of the item difficulties) when the correlation between them is greater than 0.98 [37].
Next, the analysis of differential item functioning was made using the PARSCALE software. Because of the number of items that were analyzed (n = 30), the α-level was set to .0017 (using the Bonferroni adjustment, 0.05 / 30). Each item difficulty was compared across all four groups and the items that showed statistically significant differences exhibited DIF, which implied item bias.
Finally, an examination of the reliability and validity (validity was assessed using concurrent measures) of the QLQ-C30 was made with both the original unadjusted QLQ-C30 scores and with DIF-adjusted scores. Two methods were used for calculating the adjusted scores. The first involved simply removing the items shown to exhibit DIF. For the second, the effect of ethnicity was partialed out of the DIF items using standard regression techniques for computing a partial correlation. Reliability was assessed with coefficient α. Calculating the correlation between overall QLQ-C30 score and two measures known to be related to QoL, receipt of chemotherapy (yes or no) and the Karnofsky Performance Status Scale [38], assessed validity. When comparing the results from the unadjusted and adjusted measures, a small change in reliability and validity would be consistent with a hypothesis that the QLQ-C30 maintains the same psychometric properties even in the presence of DIF.
Results
Prior to conducting analyses, seven of the 366 patients were removed from the sample because their ethnicity was unknown. Missing values for items on the QLQ-C30 were replaced by the mean values for the respective ethnic groups. For example, if an individual who was Japanese left a question blank, the value was replaced by the mean of the given question for the Japanese group (rounded to the nearest integer value). Out of 10,770 possible values (359 patients multiplied by 30 items on the QLQ-C30), there were only 11 (0.1%) missing values that needed to be imputed in this way. Patient characteristics are shown in Table 1.
Table 1 Means (M) and standard deviations (s) of demographic and cancer measures
Ethnicity
Measure Caucasian
n = 121 Japanese
n = 133 Filipino
n = 61 Hawaiian
n = 51 Total
n = 366
M s M s M s M s M s
Age* 63.4 11.8 63.7 11.9 58.6 13.4 57.3 14.6 61.9 12.8
Female (%) 51.2 50.2 58.6 49.4 50.8 50.4 64.7 48.3 55.7 49.7
Education (years)* 14.2 3.0 12.9 2.9 11.9 4.0 12.1 2.6 13.0 3.2
Married (%) 63.0 48.5 73.5 44.3 76.3 42.9 59.2 49.7 68.5 46.5
Cancer Site
Breast (%)* 29.8 45.9 44.4 49.9 19.7 40.1 37.3 48.8 34.4 47.6
Prostate (%) 32.2 46.9 27.1 44.6 31.1 46.7 11.8 32.5 27.3 44.6
Advanced Stage (%) 13.2 34.0 9.0 28.8 25.0 43.7 7.8 27.2 12.9 33.5
Performance Status 87.4 11.0 87.7 10.0 84.3 10.4 85.1 11.4 86.7 10.6
Comorbidity (%) 58.7 49.4 62.4 48.6 55.7 50.1 62.7 48.8 60.1 49.0
Chemo (%) 7.5 26.4 10.5 30.8 15.3 36.3 12.5 33.4 10.6 30.8
Hormone (%) 19.1 39.5 30.6 46.3 16.1 37.1 25.0 43.8 23.7 42.6
Radiation (%) 37.8 48.7 43.7 49.8 27.6 45.1 27.1 44.9 36.8 48.3
Surgery (%) 83.5 37.3 80.3 39.9 77.0 42.4 94.1 23.8 82.7 37.8
Notes. Advanced Stage was defined as regional or distant disease. Performance Status is the Karnofsky Performance Status, which ranges from 0 to 100 (where 0 = deceased and 100 = no sign of disease). Comorbidity indicates the presence of a comorbid disease. Chemo, Hormone, Radiation, and Surgery indicate receipt of the respective treatment. Statistically significant (p < .01) differences across ethnic groups are indicated with an asterisk (*).
The intraclass correlations between group difficulties of the items are shown in Table 2. An examination of the values indicates that none of the ethnic groups is exactly the same type (using a cutoff intraclass correlation of 0.98 [37]), although the Caucasian and Japanese groups are the closest (0.945). The two groups most different from one another are the Hawaiians and Filipinos (0.792). These intraclass correlations give an indication of how many individual items will show significant DIF. Because the Hawaiians and Filipinos have the lowest intraclass correlation, these groups can be expected to have the largest number of items with significant DIF. The groups with the highest intraclass correlation, Caucasians and Japanese, can be expected to have the least.
Table 2 Item Difficulty Intraclass Correlations by Ethnic Group
Caucasian Japanese Filipino Hawaiian
Caucasian
Japanese 0.945
Filipino 0.896 0.863
Hawaiian 0.871 0.939 0.792
Note. Because the item difficulties were standardized (mean = 0, standard deviation = 1), the intraclass correlations are equivalent to the Pearson correlations.
Of the 30 items examined, 12 (40%) showed significant DIF (see Table 3 for the IRT difficulty parameters and significance tests across all of the ethnic groups). When compared to Caucasians, five items existed that were significantly biased against Filipinos, and three that were significantly biased in favor. When compared to Japanese, three items existed that were significantly biased against Filipinos, and five were significantly biased in favor. This supports the stated hypothesis that biased items would exist for Filipinos when compared to Caucasian and Japanese groups. Further, when Filipinos were compared to Hawaiians, there were five significantly biased items. The items that appear consistently biased against Filipinos, independent of the comparison group, are ones representing nausea and vomiting (items14 and 15) and financial difficulties (item 28). Items that appear consistently biased for Filipinos are one representing physical functioning (items 1, 2, and 29) and cognitive functioning (item 25, remembering).
Table 3 IRT difficulty parameters for the QLQ-C30 items
Standardized Difficulty
Item Content Caucasian Japanese Filipino Hawaiian χ2(3) p
01 Physical 1.09 1.64 1.11 0.74 27.7 *
02 Physical 0.72 0.83 0.19 1.11 15.2 *
03 Physical -1.51 -1.29 -2.35 -0.61 11.1 .01
04 Physical -1.12 -0.78 -1.24 -0.88 2.3 .51
05 Physical -2.15 -2.39 -2.97 -2.29 2.5 .47
06 Role 0.81 0.51 0.51 0.61 7.3 .06
07 Role -1.79 -0.80 -0.73 -0.38 22.2 *
08 Dyspnea -0.15 -0.01 -0.06 0.17 4.2 .24
09 Pain 0.90 0.89 0.67 0.68 5.3 .15
10 Fatigue 1.50 0.94 1.25 1.02 28.8 *
11 Insomnia 1.05 0.68 0.93 1.02 11.0 .01
12 Fatigue 0.23 0.35 0.51 0.43 5.2 .15
13 Appetite Loss -0.11 -0.39 0.00 -0.34 5.2 .16
14 Nausea -0.96 -1.25 -0.30 -0.61 19.4 *
15 Nausea -2.23 -2.89 -1.35 -3.43 19.5 *
16 Constipation -0.34 0.01 0.16 0.16 27.2 *
17 Diarrhea -0.44 -0.55 -0.73 -1.04 12.1 .01
18 Fatigue 1.26 1.20 1.08 1.41 1.3 .72
19 Pain -0.35 -0.14 -0.13 0.11 9.0 .03
20 Cognitive -0.42 -0.35 -0.42 -0.24 1.0 .80
21 Emotional 0.44 0.18 0.42 0.01 11.1 .01
22 Emotional 1.18 0.91 1.27 0.56 21.7 *
23 Emotional 0.49 0.51 0.58 0.72 2.5 .47
24 Emotional 0.47 0.30 0.54 -0.03 11.0 .01
25 Cognitive 0.54 0.89 0.13 0.92 30.2 *
26 Social 0.07 0.23 0.21 -0.24 4.7 .19
27 Social 0.71 0.57 0.28 0.23 15.2 *
28 Financial 0.26 0.30 1.12 0.49 47.0 *
29 Physical 0.08 0.14 -0.33 -0.01 15.1 *
30 Overall -0.22 -0.22 -0.37 -0.31 2.0 .56
Notes. Standardized difficulty parameters are given for each ethnic group. For a given item, values shown in bold are significantly greater than values shown in italics. Values shown in plain type do not differ significantly from other values. The χ2(3) column provides the chi-square value (with 3 degrees of freedom) for the omnibus test for differences in difficulty across all four ethnic groups. The p column is the p-value for the chi-square test. An asterisk (*) indicates that the p-value is less the alpha-value, which was set to 0.0017 (0.05 / 30). Because Rasch modeling was used, all items have equal slope parameters, set to 1.0.
Some other findings related to item bias were also noteworthy. For one emotional functioning item (item 22, worry), Hawaiians show significantly less difficulty. For Caucasians, a role functioning question (item7, work at job) and the constipation question (item16) were significantly less difficult for them. For a social functioning item (item 27, social activities), Caucasians and Japanese had greater difficultly than Hawaiians and Filipinos.
Finally, an assessment of the reliability and validity of the QLQ-C30, both with the unadjusted overall QoL scores and the DIF-adjusted scores, were made. Coefficient α was calculated for the unadjusted scale, the scale adjusted by deleting DIF items, and the scale adjusted by partialling ethnicity from DIF items. The values were .94, .92, and .93, respectively.
The correlations between unadjusted overall QoL scores with receipt of chemotherapy and Karnofsky Performance Status were r = -.14 (p = .01) and r = .59 (p < .0001), respectively. After the removal of the QLQ-C30 items shown to exhibit DIF, the absolute values of the correlations dropped slightly to r = -.12 (p = .02) and r = .56 (p < .0001). Using the item partialling adjustment, correlations of r = -.13 (p = .02) and r = .58 (p < .0001) were obtained. The low correlation between chemotherapy and QoL may be related to persons with high cancer stage being less likely to receive chemotherapy.
Note that adjustment for DIF (both the deletion and partialling methods) resulted in a reduction in the level of ethnic differences for overall QoL score. This in turn caused the variance in QoL score to also be reduced, which led to the lower α coefficients and lower correlations. The reduction was too small, however, for it to be considered a result of anything other than a mathematical artifact, suggesting that DIF may not have an impact on the psychometric properties of the QLQ-C30.
Discussion
Overview
Differential item functioning was found in several of the items contained in the EORTC QLQ-C30, implying that not all of the items assess quality of life equally across ethnicity. This supports theories and research suggesting that ethnicity influences perceptions of health and sickness [16-19], and indicates that caution is necessary when comparing scores between ethnic groups. Because responses to the questionnaire are dependent on a factor unrelated to QoL, namely ethnicity, persons who are equal on the underlying construct of QoL will not necessarily respond equally on the questionnaire. Should it be necessary to compare QoL scores between ethnic groups, a special consideration must be made for those items that exhibit DIF.
Several items existed that showed statistically significant bias with respect to Filipinos, supporting the main hypothesis of the study. The factors of physical functioning, nausea and vomiting, cognitive functioning, social functioning, and financial difficulties are potentially biased. This implies that Filipinos will indicate either higher or lower scores on these items when compared to the other ethnic groups, even when their global QoL is the same. Hence, the global QoL as measured by the QLQ-C30 will be biased with respect to Filipinos.
However, the presence of DIF did not appear to alter reliability and validity estimates. Neither removal nor adjustment of items shown to exhibit DIF had an appreciable impact on any of the estimated values. Although replication is clearly needed to support this and all of the findings presented here, this suggests that the deletion or modification of the items exhibiting DIF is not necessary, and may even be inappropriate. So, assuming the existence of DIF is reflective of real ethnic differences in how certain items are interpreted, what are its implications and how should it be addressed?
Implications
Although detecting DIF is relatively straightforward, determining its cause and the best method to alleviate its presence is more difficult. If the reasons for DIF had been suspected to arise because of poorly written questions that are not interpreted consistently, then rewording or even removal of items is probably necessary. Even though the QLQ-C30 has been tested to ensure that the questions are worded properly [39], there still remains the possibility that ethnic groups not previously examined will have different interpretations. However, in the present sample, all individuals were English speaking, lived in Hawai'i, and had the opportunity to ask for clarification if they were confused. Therefore, the likelihood of poor wording being the primary cause of the DIF found here is reduced, and rewording or removal of items is probably unwarranted.
If poor wording or misinterpretation of items cannot explain the presence of DIF, then this suggests that the definition of QoL is not equivalent across ethnic groups. The questions exhibiting DIF provide relevant information for assessing how these definitions differ. Specifically, the DIF questions are not consistent measures of the QoL construct across all ethnic groups, providing little QoL-related information for one ethnic group but valuable information for another. Hence, not only is it important that these items be kept, but it is necessary that special attention be paid to them when comparing QoL across ethnic groups.
An examination of the items showing significant DIF between the Caucasians and the other ethnic groups shows that Caucasians are better able to work at a job or to do household jobs, and experience less constipation. Because overall QoL had been controlled, this finding may suggest that Caucasians value these factors more strongly than others in terms of global QoL. In order for Caucasians to achieve the same level of QoL held by others they may need better performance with their work and they may need greater freedom from constipation. If the Caucasians were equal with others on these items, their underlying QoL would likely be less.
The Hawaiians, when compared with others, showed significantly lower scores on the emotional functioning item related to worry. This may suggest that for Hawaiians, when compared to the other groups, freedom from worry is more important in relation to their global QoL. Similarly, the items that were biased against the Filipinos may reflect aspects of QoL that are not particularly relevant to the Filipino population. For example, there was significant bias in the items reflecting financial difficulties. Perhaps in the Filipino worldview financial success is less of a necessity, and when hardships of costly treatments are endured, the impact on QoL is less severe.
An important question to consider for future research is what cultural differences could explain the different interpretation of certain items. For example, why did the item related to working appear to have greater importance for Caucasians? Perhaps in Caucasian society work attributes are more highly prized than in other societies. This is consistent with the current lore regarding Caucasian and Asian cultures. When compared to Asians, Caucasians place greater value on independence and individual achievement [40]. Therefore, for Caucasians, not being able to work may be particularly detrimental to quality of life. If so, this may have significance beyond simply the assessment of quality of life in cancer patients. This may provide important evidence for culture-based theories comparing individualist versus collectivist societies.
Recommendations
It must be emphasized that before any recommendations can be made for adjusting scores or modifying the QLQ-C30 based on DIF, the results found here must be replicated. This is an exploratory study, which we hope will lead to increased research in this area. While this sample from Hawai'i certainly contributes to the current lore, it is by no means representative enough to indicate, by itself, how corrections can be made for a questionnaire that is used in countries all over the world. Future research needs to expand the study of DIF to additional countries, cultures, and ethnic groups. Through continued research in this area, a clearer picture will emerge as to how the definition of QoL varies across ethnic groups.
Assuming that the findings presented here are replicated and that future research demonstrates which items are consistently biased and to what degree the bias exists, one approach to correct DIF might be to weight items based on ethnicity. Items that are biased against a particular group should be given less weight in the calculation of overall QoL for individuals in that group. By placing greater emphasis on the items that are most relevant for a particular group, most biases could be eliminated. The weighting scheme should be based on the level of bias that exists. For example, items severely biased against a certain group should be given little weight when calculating that group's overall QoL score. But items biased in favor should be given greater weight. From the present study, it is suggested that items related to work and constipation should probably be given greater emphasis when assessing Caucasians. Similarly, items related to financial difficulties and nausea should probably be given less weight for Filipinos, whereas items related to physical functioning should be given more.
Another approach that could be used is similar to the one of the methods employed here for removing bias from the DIF items. If an item is biased, the effect of ethnicity could be partialed out of it using standard regression procedures (controlling for the effect of ethnicity). This would make all ethnic groups equivalent on the particular item. However, this approach could prove cumbersome in practice because additional statistical procedures would be required, as compared to simply weighting items.
Finally, it is recommended that DIF not be seen solely as a problem that is to be eliminated. Items exhibiting DIF may reflect key differences in how QoL is defined across ethnic groups, and important information can be obtained from these items for understanding cultural differences. Therefore, it is not recommended that any of the biased items be removed, or even reworded. Rather, expanded cultural studies should be undertaken to further explain the multifaceted QoL construct.
Limitations
The primary strength of the methods employed here is the ability to detect isolated items exhibiting DIF (i.e., item bias). However, a limitation with this approach is that it cannot detect item bias if the bias exists within most or all of the items. For example, if all of the items on the QLQ-C30 were biased against Filipinos, then this approach would have failed to detect the biases. Group differences would have appeared to reflect genuine differences in QoL, and not bias. However, because of the extensive research that has been conducted assessing the validity of the QLQ-C30, the risk of such widespread bias is reduced.
Another limitation arises when statistically controlling for group differences by standardizing the item difficulties within each group. In order for this controlling procedure to be effective, it is necessary for relatively few of the items to be biased. If too many items are biased, then it is not necessarily only quality of life that is being controlled, but also factors related to the bias. In this study, the issue is probably not serious because there were many items (60%) that did not show statistically significant bias. With the possible exception of the Hawaiian and Filipino comparisons, the QLQ-C30 was suited for assessing item bias with the methods employed here.
Finally, a two-parameter IRT methodology may provide a greater level of item information (recall that information refers to how well an item discriminates between persons at varying ability levels). Some of the items on the QLQ-C30 may provide more information than others in assessing QoL, and a two-parameter model would indicate the amount of information provided by each item. In the present study, the sample size was not large enough to obtain an optimal IRT solution when incorporating two-parameters, but future QoL research is planned in which a two-parameter IRT model will be used. This future research will also examine each of the QoL subscales in addition to overall QoL, allowing for a comparison between the two approaches.
Conclusion
In conclusion, this research suggests that quality of life, as assessed by the EORTC QLQ-C30, is at least partially dependent on one's ethnic origin. The culture from which one belongs is an important determinant of how one defines QoL. This may explain why previous research has shown Filipinos to have a lower QoL than other ethnic groups [21]. Filipinos may be equal to other groups with respect to underlying QoL, but different in terms of the aspects that characterize their QoL. This study suggests that persons from different ethnic backgrounds do not define QoL in exactly the same manner, and research involving QoL that employs different ethnic groups cannot ignore this important issue.
Authors' contributions
IP designed the study, carried out all of the analyses, and drafted the manuscript.
CG handled the data collection and editing of the manuscript.
Supplementary Material
Additional File 1
Pagano and Gotay Appendix.doc. Appendix: The QLQ?C30 version 1.0 with Functional / Symptom Scales Indicated
Click here for file
Acknowledgements
This study was supported by a grant from the National Cancer Institute (CA61711). We are grateful for assistance in data collection by Jeffrey Stern, Mary Lynn Fiore, Akiko Lau, Malia Wilson, Daniella Dumitriu, Florence Yee, and Cris Yamabe. The participation of Kaiser-Permanente Hawaii, Kuakini Medical Center, St. Francis Medical Center, Straub Clinic and Hospital, and Queens Medical Center is also greatly appreciated.
==== Refs
Cella DF Lloyd SR Wright BD Cross-cultural instrument equating: Current research and future directions 1992 707 715
Quality of life and clinical trials [editorial] [see comments] Lancet 1995 346 1 2
Aaronson NK Ahmedzai S Bergman B Bullinger M Cull A Duez NJ Filiberti A Flechtner H Fleishman SB de Haes JC The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology J Natl Cancer Inst 1993 85 365 376 8433390
Aaronson NK Bakker W Stewart AL van Dam FSAM van Zandwijk Y Kirkpatrick A Aaronson NK and Beckmann JH Multidimensional approach to the measurement of quality of life in lung cancer trials The quality of life of cancer patients 1987 New York, Raven Press 63 82
Schipper H Clinch JJ Olweny CLM B.Spilker Quality of life studies: Definitions and conceptual issues Quality of Life and Pharmacoeconomics in Clinical Trials 1996 2nd Philadelphia, Lippincott-Raven Publishers 11 23
Irvine DM Vincent L Graydon JE Bubela N Fatigue in women with breast cancer receiving radiation therapy Cancer Nurs 1998 21 127 135 9556939 10.1097/00002820-199804000-00006
Longman AJ Braden CJ Mishel MH Side effects burden in women with breast cancer Cancer Pract 1996 4 274 280 9004574
Broeckel JA Jacobsen PB Balducci L 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
Longman AJ Braden CJ Mishel MH Side-effects burden, psychological adjustment, and life quality in women with breast cancer: pattern of association over time Oncol Nurs Forum 1999 26 909 915 10382190
Ganz PA Hirji K Sim MS Schag CA Fred C Polinsky ML Predicting psychosocial risk in patients with breast cancer Med Care 1993 31 419 431 8501990
Nayfield SG Hailey BJ Quality of life assessment in cancer clinical trials 1991 Bethesda, MD, Division of Cancer Preention and Control, National Cancer Institute, U. S. Department of Health and Human Services 1920492
Moinpour CM Feigl P Metch B Hayden KA Meyskens FLJ Crowley J Quality of life end points in cancer clinical trials: review and recommendations J Natl Cancer Inst 1989 81 485 495 2646455
Johnson JR Temple R Food and Drug Administration requirements for approval of new anticancer drugs Cancer Treat Rep 1985 69 1155 1159 4042094
Henderson-James D Spilker B Spilker B An industry perspective Quality of life assessment in clinical trials 1990 New York, Raven Press 183 192
Cella DF Quality of life: the concept J Palliat Care 1992 8 8 13 1279152
Brisbane FL Womble M Working with African Americans: the professional's handbook 1992 Chicago, IL, HRDI International Press
Flaherty JA Gaviria FM Pathak D Mitchell T Wintrob R Richman JA Birz S Developing instruments for cross-cultural psychiatric research J Nerv Ment Dis 1988 176 257 263 3367140
Kleinman A Eisenberg L Good B Culture, illness, and care: clinical lessons from anthropologic and cross-cultural research Ann Intern Med 1978 88 251 258 626456
Marin G Marin BVO Research with Hispanic Populations Applied Social Research Methods Series 1991 23 Newberry Park, CA, Sage
Marshall PA Cultural influences on perceived quality of life Semin Oncol Nurs 1990 6 278 284 2274725 10.1016/0749-2081(90)90030-9
Gotay CC Holup J Pagano IS Ethnic differences in quality of life among early breast and prostate cancer survivors Psycho-Oncology 2001 11 103 113 11921326 10.1002/pon.568
Johnson CD Wicks MN Milstead J Hartwig M Hathaway DK Racial and gender differences in quality of life following kidney transplantation Image J Nurs Sch 1998 30 125 130 9775552
Bjorner JB Kreiner S Ware JE Damsgaard MT Bech P Differential item functioning in the Danish translation of the SF-36 J Clin Epidemiol 1998 51 1189 1202 9817137 10.1016/S0895-4356(98)00111-5
Teresi JA Kleinman M Ocepek-Welikson K Ramirez M Gurland B Lantigua R Holmes D Applications of item response theory to the examination of the psychometric properties and differential item functioning of the Comprehensive Assessment and Referral Evaluation dementia diagnostic scale among samples of Latino, African-American, and White non-Latino elderly Research on Aging 2000 22 738 773
Karnofsky DA Burchenal JH Macleod CM The clinical evaluation of chemotherapeutic agents in cancer Evaluation of Chemotherapeutic Agents in Cancer 1949 New York, Columbia University Press 191 205
Batel-Copel LM Kornblith AB Batel PC Holland JC Do oncologists have an increasing interest in the quality of life of their patients? A literature review of the last 15 years Eur J Cancer 1997 33 29 32 9071895 10.1016/S0959-8049(96)00414-5
Osterlind SJ Test Item Bias 1983 Beverly Hills, CA, SAGE Publications
Nunnally JC Bernstein IH Psychometric Theory 1994 Third New York, McGraw-Hill, Inc. 752
Holland PW Wainer H Differential Item Functioning 1993 Hillsdale, NJ, Lawrence Erlbaum Associates
Hays RD Morales LS Reise SP Item response theory and health outcomes measurement in the 21st century Med Care 2000 38 II28 42 10982088
Hambleton RK Emergence of item response modeling in instrument development and data analysis Med Care 2000 38 II60 5 10982090
Embretson SE Reise SP Item Response Theory For Psychologists 2000 Mahwah, NJ, Lawrence Erlbaum Associates, Inc.
Cella D Chang CH A discussion of item response theory and its applications in health status assessment Med Care 2000 38 II66 72 10982091
Baker FB Item Response Theory: Parameter Estimation Techniques 1992 New York, Marcel Dekker, Inc.
Linacre JM Sample size and item calibration stability Rasch Measurement Transactions 1994 7 328
Gotay CC Blaine D Haynes SN Holup J Pagano IS Assessment of quality of life in a multicultural cancer patient population Psychological Assessment 2002 14 439 450 12501569 10.1037/1040-3590.14.4.439
Angoff WH Holland PW and Wainer H Perspectives on differential item functioning methodology Differential Item Functioning 1993 Hillsdale, NJ, Lawrence Erlbaum Associates 3 23
Schaafsma J Osoba D The Karnofsky Performance Status Scale re-examined: a cross-validation with the EORTC-C30 Qual Life Res 1994 3 413 424 7866359 10.1007/BF00435393
Aaronson NK Cull A Kaasa S Sprangers MAG Spilker B The European Organization for Research and Treatment of Cancer (EORTC) modular approach to quality of life assessment in oncology: An update Quality of Life and Pharmacoeconomics in Clinical Trials 1995 2nd New York, Raven Press
Gotay CC Baider L, Cooper GL and Kaplan De-Nour A Cultural variation in family adjustment to cancer Cancer and the family 1996 New York, Wiley 31 52
|
16209720
|
PMC1277838
|
CC BY
|
2021-01-04 16:38:13
|
no
|
Health Qual Life Outcomes. 2005 Oct 7; 3:60
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-60
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-631623616710.1186/1477-7525-3-63ResearchDifferential effect sizes of growth hormone replacement on quality of life, well-being and health status in growth hormone deficient patients: a meta-analysis Deijen Jan Berend [email protected] Lucia I [email protected] Joost [email protected] Madeleine L [email protected] Department of Clinical Neuropsychology, Free University, van der Boechorststraat 1, 1081 BT Amsterdam, the Netherlands2 Department of Endocrinology, VU University Medical Center, PO Box 7057, 1007 MB, Amsterdam, the Netherlands2005 19 10 2005 3 63 63 3 8 2005 19 10 2005 Copyright © 2005 Deijen et al; licensee BioMed Central Ltd.2005Deijen 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
Patients with growth hormone deficiency (GHD) frequently report to suffer from an impaired Quality of Life (QoL) and growth hormone (GH) substitution is found to improve this. However, the same test may be used for measuring QoL, well-being or health status in different studies. QoL has been defined as the subjective appraisal of one's current life based primarily on psychological function. The most important in the appraisal of well-being is mental function and concerning health status patients evaluate physical function as most important. To differentiate the effects of GH replacement on psychological variables in patients with GHD we carried out a number of meta-analyses, classifying questionnaires into instruments measuring QoL, psychological well-being and health status.
Methods
We searched the electronic databases PUBMED and PiCarta from 1985 to 2004. Studies were included that evaluated the effect of GH on patient-reported outcomes in adults with GHD (aged 18 years and above). According to generally accepted definitions we classified the questionnaires as instruments measuring QoL, well-being and health status. By means of meta-analyses the average effect size (d) for QoL, well-being and health status was calculated.
Results and Discussion
Based on open studies GH replacement is found to improve QoL with a small effect size (d = 0.18), well-being with a medium effect size (d = 0.47) and health status with a small effect size (d = 0.26). As the effect size of well-being is most pronounced the generally reported effects of GH replacement on QoL may be overestimated and actually reflect the effect on well-being.
Conclusion
To get more insight in the specific psychological effects of GH treatment it is recommended that instruments selected for these studies should be more consistently classified as instruments measuring QoL, well-being or health status.
==== Body
Background
The concept of Quality of life (QoL) is frequently used in reports of studies in patients with growth hormone deficiency (GHD). It is already 4 decades ago that the first report was published on QoL in relation to growth hormone deficiency. In this particular study the effect of GH substitution on QoL in a GHD patient is described [1]. From then on, it has been frequently reported that patients with GHD suffer from an impaired Quality of Life (QoL), and that growth hormone (GH) substitution improves their condition [2-6]. With regard to GHD, a subnormal QoL in adults with GHD is inferred from the observations that these patients feel less energetic, are emotionally more labile, and experience disturbances in sex life and feelings of social isolation at a significantly higher frequency than controls [2,3,6,7]. With respect to GH substitution in GHD adults, the effects of 10 years of GH replacement on psychological well-being have been evaluated. Overall scores for the NHP, energy levels and emotional reaction improved in the GH-treated group compared to an untreated group [8]. In addition, one year of discontinuation of GH treatment in a study in GHD patients led to a decrease in QoL (psychological complaints and depression). This effect was counteracted after restart of GH therapy resulting in reduced anxiety and depression and improved QoL [9]. In another study withdrawal of GH treatment from adults with GHD had detrimental psychological effects (decreased energy, increased tiredness, pain, irritability and depression) [10]. The studies above exemplify that the psychological effects of GH replacement are being measured with a variety of instruments. Moreover, the same tests may be used as an instrument measuring QoL in one and well-being or health status in another study.
The concept of QoL with particular relevance to patients with GH deficiency has been defined as "the social and psychological well-being assessed from the patient's perspective". Elements which contribute to a person's QoL are their levels of emotional, cognitive and social functioning [11]. Indeed, it is generally acknowledged that the QoL of a patient is not only defined by quantitative factors of a disease, for example the severity of GH deficiency, but also by psychosocial factors. Functional impairments, not being able to perform personal goals, unemployment and relational problems should also be taken into account when measuring the influence of a disease for a patient [12].
A decade ago editorial attention in the Lancet was paid to the conceptual and methodological difficulties pertaining to the concept of QoL [13]. This editorial included the study of Gill and Feinstein [14] who sampled 75 articles with "quality of life' in their titles. Only 15% of the sample included definitions of QoL and in only 13% of the cases patient-rated QoL measures – as opposed to 'objective' questionnaires – were used. From the perspective of patients, QoL and health status have been found to be distinct constructs. QoL has been defined as 'the subjective appraisal of one's current life based primarily on psychological function and to a lesser degree on physical functioning'. Indeed, when rating QoL, patients give greater emphasis to mental health than to physical functioning. This pattern is reversed for appraisal of health status, for which physical function is more important than mental health [15]. In a second commentary in the Lancet the importance of differentiating health status from QoL is also addressed. It is stated that using a health status questionnaire, which is measuring how people feel about their health, to assess QoL may lead to misleading conclusions. In case of health status patients report how they feel mainly about their physical health, whereas in case of reporting well-being patients exhibit feelings of depression, anxiety and energy [16]. Thus, patient reported outcomes of physical status may be specifically measured by a health status questionnaire such as the Short-form Health Survey (SF-36) [17], whereas mental status may be measured by a well-being questionnaire such as the Psychological General Well-being Index (PGWB) [18]. It may be clear that such generic measures of health status and well-being are not measures of QoL, although often categorized as such. Therefore, with respect to the measurement of QoL in patients with GHD, a few disease-specific QoL questionnaires have been developed, the QoL-AGHDA being one of these. This is a condition-specific QoL measure. All items of this instrument are expressed as unsatisfied needs [19]. More recently, the psychometric properties of a new individualized questionnaire, the A-RHDQoL, measuring perceived impact of age-related hormonal decline on QoL in older men, have been reported. The questionnaire is individualized because respondents only rate those domains that are relevant to them. Both the impact of age-related hormonal decline on life domains and the importance of each domain to the individual are taken into account [10].
A complete picture of the effects of GH replacement on psychological status in GHD patients can be inferred from changes determined with a QoL scale, in conjunction with an established measure of health status and another of well-being. However, up until now these scales have not been used consequently to measure specifically one of these concepts. As a consequence, in spite of a number of studies on the effects of GH treatment on psychological status, the relative contribution of GH treatment on changes in QoL, well-being and health status is not known yet.
In order to differentiate the effects of GH replacement on psychological variables in patients with GHD we carried out a number of meta-analyses, distinguishing between QoL, psychological well-being and health status.
Methods
Search strategy
We searched the electronic databases PUBMED and PiCarta from 1985 to 2004. PiCarta is an integrated multimaterial database with request-facilities and offering access to online resources and electronic documents.
Studies were included that evaluated the effect of GH on patient-reported outcomes in adults with GHD (aged 18 years and above). The following search terms were used: growth hormone, mood, health status, well-being and quality of life.
Study selection
Two investigators independently examined manuscripts for inclusion. Eligible studies were reports providing quantitative data about the effect of GH therapy on patient-reported outcomes in GH deficient adults. Studies had to be placebo-controlled or designed as a cross-over/parallel or open clinical trial. Questionnaires had to be used to measure patient-reported outcomes. Case reports, review articles and studies in which the psychometric quality of the used questionnaire was unknown were excluded. Furthermore, studies on GH therapy for other diseases (for instance Turner syndrome, Prader Willi Syndrome, fibro-myalgia, etc.) were not included in this meta-analysis.
Statistical analysis
We carried out a series of meta-analyses using a random effects model. The meta-analyses were performed by means of the statistical package Comprehensive Meta-analysis (Biostat, Inc, USA) [20]. This program is used to determine d-values (effect sizes). The most commonly used measures of effect size are the standardized mean difference (d) and the correlation coefficient (r). The effect size is a simple quantitative measure that provides one useful index of the importance of an effect. The effect size index d standardizes the raw effect size as expressed in the measurement unit of the dependent variable by dividing it by the common SD of the measures in their respective populations [21]. We calculated the difference prior to and after GH therapy, divided by the pooled standard deviation of the two measurements. Effect sizes (d's) were calculated, averaged for each study and pooled. Effect size d = 0.2–0.5 is conceived as a small effect, d = 0.5–0.8 as a medium effect and d ≥ 0.8 as a large effect. A medium effect size is conceived as one large enough to be visible to the naked eye [21].
Questionnaires
The most frequently encountered questionnaires measuring QoL, well-being or health status encountered in the studies included in the meta-analysis are described below.
The Nottingham Health Profile (NHP) is a frequently used health status questionnaire in GH deficient patients that measures physical, emotional and social distress. It consists of the subscales emotional reactions, energy, pain, physical mobility, sleep and social isolation [22]. The Psychological General Well Being Schedule (PGWB) measures self-perceived affective and emotional states [18]. Subscales include anxiety, depressed mood, positive well-being, self-control, general health and vitality. The Hopkins Symptom Checklist (HSCL) is a questionnaire for the assessment of psychological and somatic complaints [23]. The Profile of Mood States (POMS) is a 32-item questionnaire with subscales depression, anger, fatigue, vigor and tension [24] and the State-Trait Anxiety Inventory (STAI) is a questionnaire to assess state and trait anxiety [25]. The Quality of Life Assessment of Growth Hormone Deficiency in Adults (QoL-AGHDA) [19] is especially designed to assess relevant aspects of GHD.
Distinguishing between QoL, well-being and health status
As is pointed out above, QoL is conceived as the total of psychosocial determinants and physical functioning assessed from the patient's perspective. We made the QoL concept operational by "the subjective judgment of the quality of daily functioning related to psychological or physical capabilities" and classified the questionnaires accordingly as QoL. In addition, as well-being is perceived as feelings of depression, anxiety and energy, and health status as feelings about physical health we defined well-being as perceived mental health and health status as perceived physical health. According to the above criteria we classified the questionnaires into the three categories. If instruments have multiple domains measuring QoL, well-being or health status, we classified these domains separately. The result of our classification is summarized in Table 1.
Table 1 Classification of questionnaires into instruments measuring quality of life, psychological well-being or health status
Quality of Life • Qol Assessment of Growth Hormone Deficiency in Adults (QoL-AGHDA)
• Satisfaction with physical activity (VAS-score)
• Sick leave, Hospital days, Doctor visits
• Nottingham Health Profile (NHP, part 1) scale: Physical mobility
• Quality of Life Scale (QLS)
• Life Fulfilment Scale
Psychological well-being • Psychological General Well-being Scale (PGWB), except General Health scale
• NHP scales: Emotional reactions, Social isolation, Energy.
• Minnesota Multiphasic Personality Inventory-2 (MMPI-2) scale: Depression.
• Hamilton Depression Scale (HDS).
• Beck Depression Inventory (BDI).
• Profile Of Mood States (POMS).
• Sjöberg mood questionnaire.
• State-Trait Anxiety Inventory (STAI) subscale: State anxiety.
• Kellner Symptom Questionnaire (KSQ).
• Symptom Checklist (SCL-90): Anxiety, Depression.
• Mental Fatigue Questionnaire (MFQ).
• Hospital Anxiety and Depression scale.
Health status • General health questionnaire (GHQ).
• NHP scales: Overall score, Pain, Sleep.
• PGWB scale: General health.
• Hopkins Symptom Checklist (HSCL).
• Leisure time physical activity (VAS-score)
• SCL-90 scale: Somatic complaints, Sleep.
Results
Fifteen studies met our inclusion criteria for analysis of the effect of GH on patient-reported outcomes and were included in this meta-analysis. Study characteristics are shown in Table 2. Total number of patients is 830 and follow-up with a maximum of 24 months was analyzed.
Table 2 Included studies on GH therapy and psychological variables
First author, year [ref] N Mean age (range) (years) Duration therapy (months) Trial design Tests
Ahmad, 2001 [28] 46 Unknown 3 Open QoL-AGHDA
Baum, 1998 [29] 40 Median 51 (24–64) 18 Controlled NHP, PGWB, GHQ, MMPI-2
Burman, 1995 [30] 36 46 (28–57) 9 Controlled NHP, PGWB, HSCL
Carroll, 1997 [31] 38 42.9 6 Controlled NHP, PGWB
Cuneo, 1998 [32] 83 41.2 12 Controlled-open NHP
Degerblad, 1990 [33] 6 20–38 3 Controlled Sjoberg, POMS
Deijen, 1998 [34] 48 27 (19–37) 24 Controlled-open POMS vigor, STAI state, HSCL
Giusti, 1998 [35] 26 51 (21–74) 6 Controlled KSQ, HDS
Hernberg, 2001 [36] 304 M: 50.8
F: 48.6 12 Open QoL-AGHDA, VAS
Murray, 1999 [37] 65 38.7 (17–72) 8 Open QoL-AGHDA, PGWB
Sartorio, 1995 [38] 8 29.6 (25–34) 6 Open Dysphoria, Anxiety
Soares, 1999 [39] 9 39.4 (28–52) 12 Controlled-open HDS, BDI
Stouthart, 2003 [9] 20 21 (17–27) 12 Open QLS, STAI, POMS, SCL-90, HSCL
Wallymahmed, 1997 [40] 30 35 24 Controlled-open NHP, MFQ
Wiren, 1998 [41] 71 45 (19–76) 24 Open NHP, PGWB
BDI = Beck Depression Inventory; GHQ = General Health Questionnaire; HSCL = Hopkins Symptom Check list; HDS = Hamilton Depression Scale; MFQ = Mental Fatigue Questionnaire; MMPI-2 = Minnesota Multiphasic Personality Inventory-2; NHP = Nottingham Health Profile; KSQ = Kellner Symptom Questionnaire; PGWB = Psychological General Well Being Schedule; POMS = Profile of Mood States; SCL-90 = Symptoms Check List-90; Sjöberg = Sjöberg mood questionnaire; STAI = State-Trait Anxiety Inventory; QLS = Quality of Life Scale; QoL-AGHDA = Quality of Life Assessment of Growth Hormone Deficiency in Adults; VAS = Visual Analogue Scale.
A series of meta-analyses on open-label studies was carried out on our classification of instruments differentiating QoL, psychological well-being and health status. As the data were too limited to distinguish between different treatment lengths we analyzed the effects of pooled treatment durations. GH replacement with an average duration of 8.6 (±4.0) months (based on 26 d's from 9 studies) improves QoL with a small effect size (p = 0.001; d = 0.18, 0.07–0.29 [CI]). With regard to psychological well-being after GH replacement with an average duration of 9.2 (±5.1) months (86 d's from 13 studies) an increase with a medium effect size is found (p < 0.001; d = 0.47, 0.36–0.57 [CI]). Finally, GH replacement with an average duration of 9.4 (±4.0) months (31 d's from 10 studies) increases health status with a small effect size (p < 0.001; d = 0.26, 0.14–0.37 [CI]). Thus, the largest effect is found for well-being, followed by health status and then Qol (Figure 1).
Figure 1 Average effect sizes for quality of life, health status and well-being.
Discussion
This present series of meta-analyses evaluated the effects of GH substitution on psychological parameters in GH deficient subjects, separately analyzed for QoL, psychological well-being and health status. Our aim was to examine whether GH would improve QoL, well-being and health status differently. Therefore we classified the psychological tests we encountered in the studies as being an instrument measuring QoL, well-being or health status.
After differentiating QoL from well-being and health status we determined the effect sizes obtained from the meta-analyses on GH replacement. GH replacement appeared to improve well-being the most, followed by health status and then QoL. The effect sizes indicate that the effect on well-being is more than twice as large as that concerning QoL and nearly twice as large as that concerning health status. It may be argued that this result may be associated with differences in the psychometric properties of the scales for QoL, well-being and health status. However, an effect size is a standardized, dimensionless number, which allows the comparison of the results of different studies or the results of different tests within one study [26]. In short, effect size is a simple quantitative measure that provides one useful index of the importance of an effect. In addition, a distinction can be made between 'small', 'medium' and 'large' effect sizes. A medium effect size is conceived as one large enough to be visible to the naked eye [21]. Thus, particularly the medium effect size of well-being is substantial while the effect sizes of health status and QoL seem to be of less importance. This suggests that the generally reported effects of GH replacement on QoL may be overestimated and actually reflect the effect on psychological well-being. However, as the effect size index only pertains to statistical effects our data are not indicative of the clinical relevance of the present results. In order to determine clinical relevance we should have been looking at the minimum clinically important difference for each scale. Unfortunately, such a clinical determination was not possible because of the variety of instruments. We can only conclude that a larger effect size may be associated with a more important clinical effect, but this needs not necessarily be the case. In contrast, it is even conceivable that the larger effect size observed for well-being reflects a smaller clinical relevance than the small effect size observed for QoL.
Finally, it is important to note that a positive effect of GH on psychological outcomes has been found mainly in open studies lacking a control group. The effects of GH treatment on patient reported outcomes have been compared to placebo in a meta-analysis we performed earlier on the same database [27]. The purpose of that analysis was to determine whether GH replacement has any beneficial effect on psychological variables. Therefore, the data of all questionnaires were pooled disregarding measuring QoL, well-being or health status. GH treatment effects on these averaged psychological functions were then compared with placebo. The overall effect of GH treatment with a median duration of 6 months was found not to be better than placebo. After reporting these data we decided to perform the present meta-analysis with a quite different objective, that is identifying which variable may be most sensitive to the effects of GH treatment.
It cannot be excluded that the enhanced well-being we observed in open studies in addition to the absence of a difference between GH treatment and placebo point to placebo-effects. The observed improvement in psychological outcomes in the open studies can therefore be attributed to other factors than GH. The attention and care given to the patients with GHD in the open studies may improve patient-reported outcomes more substantially than the contribution of GH itself.
At this point it is important to note that the present meta-analysis is based only on a selected subgroup of studies, because a lot of studies did not meet our inclusion criteria. For instance, in spite of pooling, our sample size appeared still to be quite small (i.e. 830 patients). In addition, the variability of patient characteristics, the diversity in study designs including the variety of dependent variables and publication bias may distort the results. A number of moderator or confounding variables may have attributed to the variance in effect sizes. These variables can be assumed to be sex, age, medial history (radiotherapy), dose of GH and severity or type of GHD. The limited amount of data pertaining to these confounders including the differential treatment effects in patients with either childhood-onset/adulthood-onset GHD or isolated GHD/multiple pituitary hormone deficiencies did not allow to control for these confounders.
The present meta-analysis may therefore have lead to unjustified conclusions. However, with respect to publication bias, it is known that especially reports lacking positive treatment effects are not published. Thus, if such studies had been published and be part of the meta-analysis, they would have resulted in smaller effect sizes than those reported here.
Conclusion
From the present meta-analysis we may conclude that the psychological effects of GH treatment are not quite clear yet. A variety of instruments have been used to determine specific effects on QoL, well-being and health status. However, up until now the instruments have not been reliably classified into those measuring QoL, well-being or health status. The inconsistent classification of psychological questionnaires may have lead to unjustified conclusions concerning the psychological effects of GH therapy. The results of the present meta-analysis based on generally accepted definitions of these concepts indicate that the effects of GH treatment are most obvious with respect to well-being, followed by health status and QoL. It may thus well be true that the frequently reported effects of GH on QoL are overvalued. Therefore, to get more insight into the precise nature of the psychological effects of GH therapy we recommend that in future studies more uniform classifications of psychological outcomes should be used.
Authors' contributions
JBD and LIA made substantial contributions to conception and design, acquisition, analysis, and interpretation of data and writing the manuscript. JW was involved in the design of the study and data acquisition, performed the statistical analysis and participated in drafting the manuscript. MLD was involved in interpretation of data and in revising the manuscript. All authors read and approved the final manuscript.
==== Refs
Raben MS Growth hormone. 2. Clinical use of human growth hormone N Engl J Med 1962 266 82 86 14489585
Bjork S Jonsson B Westphal O Levin JE Quality of life of adults with growth hormone deficiency: a controlled study Acta Paediatr Scand Suppl 1989 356 55 59 2816358
Deijen JB de Boer H Blok GJ van der Veen EA Cognitive impairments and mood disturbances in growth hormone deficient men Psychoneuroendocrinology 1996 21 313 322 8817729 10.1016/0306-4530(95)00050-X
McGauley GA Cuneo RC Salomon F Sonksen PH Psychological well-being before and after growth hormone treatment in adults with growth hormone deficiency Horm Res 1990 33 52 4 2245968
Mitchell CM Joyce S Johanson AJ Libber S Plotnick L Migeon CJ Blizzard RM A retrospective evaluation of psychosocial impact of long-term growth hormone therapy Clin Pediatr (Phila) 1986 25 17 23 3943250
Rosen T Wiren L Wilhelmsen L Wiklund I Bengtsson BA Decreased psychological well-being in adult patients with growth hormone deficiency Clin Endocrinol (Oxf) 1994 40 111 116 8306469
Rikken B van Busschbach J le Cessie S Manten W Spermon T Grobbee R Wit JM Impaired social status of growth hormone deficient adults as compared to controls with short or normal stature. Dutch Growth Hormone Working Group Clin Endocrinol (Oxf) 1995 43 205 211 7554316
Gibney J Wallace JD Spinks T Schnorr L Ranicar A Cuneo RC Lockhart S Burnand KG Salomon F Sonksen PH Russell-Jones D The effects of 10 years of recombinant human growth hormone (GH) in adult GH-deficient patients J Clin Endocrinol Metab 1999 84 2596 2602 10443645 10.1210/jc.84.8.2596
Stouthart PJ Deijen JB Roffel M Delemarre-van de Waal HA Quality of life of growth hormone (GH) deficient young adults during discontinuation and restart of GH therapy Psychoneuroendocrinology 2003 28 612 626 12727130 10.1016/S0306-4530(02)00045-8
McMillan CV Bradley C Gibney J Healy ML Russell-Jones DL Sonksen PH Psychological effects of withdrawal of growth hormone therapy from adults with growth hormone deficiency Clin Endocrinol (Oxf) 2003 59 467 475 14510909 10.1046/j.1365-2265.2003.01870.x
McGauley G Cuneo R Salomon F Sonksen PH Growth hormone deficiency and quality of life Horm Res 1996 45 34 37 8742116
Muldoon MF Barger SD Flory JD Manuck SB What are quality of life measurements measuring? BMJ 1998 316 542 545 9501721
Quality of life and clinical trials Lancet 1995 346 1 2
Gill TM Feinstein AR A critical appraisal of the quality of quality-of-life measurements JAMA 1994 272 619 626 7726894 10.1001/jama.272.8.619
Smith KW Avis NE Assmann SF Distinguishing between quality of life and health status in quality of life research: a meta-analysis Qual Life Res 1999 8 447 459 10474286 10.1023/A:1008928518577
Bradley C Importance of differentiating health status from quality of life Lancet 2001 357 7 8 11197385 10.1016/S0140-6736(00)03562-5
Ware JE JrSherbourne CD The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection Med Care 1992 30 473 483 1593914
Dupuy HJ Matsson ME, Furberg CD, Elinson I The psychological general well-being (PGWB) index Assessment of Quality of Life in Clinical Trials of Cardiovascular Therapy 1984 New York: Le Jacq Publications
McKenna SP Doward LC Alonso J Kohlmann T Niero M Prieto L Wiren L The QoL-AGHDA: an instrument for the assessment of quality of life in adults with growth hormone deficiency Qual Life Res 1999 8 373 383 10472170 10.1023/A:1008987922774
Borenstein M Rothstein H Comprehensive Meta Analysis 1999 New Jersey, USA;
Cohen J A power primer Psychol Bull 1992 112 155 159 10.1037/0033-2909.112.1.155
Hunt SM McKenna SP McEwen J Williams J Papp E The Nottingham Health Profile: subjective health status and medical consultations Soc Sci Med [A] 1981 15 221 229 6973203
Derogatis LR Lipman RS Rickels K Uhlenhuth EH Covi I The Hopkins Symptom Checklist Pharmapsychiatry 1974 7 79 110
Shacham S A shortened version of the Profile of Mood States J Pers Assess 1983 47 305 306 6886962
Spielberger CD Test manual for the State-Trait Anxiety Inventory-STAI form Y 1980 Palo Alto, California: Consulting Psychologist Press 7391775
Prentice D Miller D When small effects are impressive Psychological Bulletin 1992 112 160 164 10.1037/0033-2909.112.1.160
Arwert LI Deijen JB Witlox J Drent ML The influence of growth hormone (GH) substitution on patient-reported outcomes and cognitive functions in GH-deficient patients: a meta-analysis Growth Horm IGF Res 2005 15 47 54 15701572 10.1016/j.ghir.2004.11.004
Ahmad AM Hopkins MT Thomas J Ibrahim H Fraser WD Vora JP Body composition and quality of life in adults with growth hormone deficiency, effects of low-dose growth hormone replacement Clin Endocrinol (Oxf) 2001 54 709 717 11422104 10.1046/j.1365-2265.2001.01275.x
Baum HB Katznelson L Sherman JC Biller BM Hayden DL Schoenfeld DA Cannistraro KE Klibanski A Effects of physiological growth hormone (GH) therapy on cognition and quality of life in patients with adult-onset GH deficiency J Clin Endocrinol Metab 1998 83 3184 3189 9745423 10.1210/jc.83.9.3184
Burman P Broman JE Hetta J Wiklund I Erfurth EM Hagg E Karlsson FA Quality of life in adults with growth hormone (GH) deficiency: response to treatment with recombinant human GH in a placebo-controlled 21-month trial J Clin Endocrinol Metab 1995 80 3585 3590 8530603 10.1210/jc.80.12.3585
Carroll PV Littlewood R Weissberger AJ Bogalho P McGauley G Sonksen PH Russell-Jones DL The effects of two doses of replacement growth hormone on the biochemical, body composition and psychological profiles of growth hormone-deficient adults Eur J Endocrinol 1997 137 146 153 9272102 10.1530/eje.0.1370146
Cuneo RC Judd S Wallace JD Perry-Keene D Burger H Lim-Tio S Strauss B Stockigt J Topliss D Alford F Hew L Bode H Conway A Handelsman D Dunn S Boyages S Cheung NW Hurley D The Australian Multicenter Trial of Growth Hormone (GH) Treatment in GH-Deficient Adults J Clin Endocrinol Metab 1998 83 107 116 9435425 10.1210/jc.83.1.107
Degerblad M Almkvist O Grunditz R Hall K Kaijser L Knutsson E Ringertz H Thoren M Physical and psychological capabilities during substitution therapy with recombinant growth hormone in adults with growth hormone deficiency Acta Endocrinol (Copenh) 1990 123 185 193 2220259
Deijen JB de Boer H van der Veen EA Cognitive changes during growth hormone replacement in adult men Psychoneuroendocrinology 1998 23 45 55 9618751 10.1016/S0306-4530(97)00092-9
Giusti M Meineri I Malagamba D Cuttica CM Fattacciu G Menichini U Rasore E Giordano G Impact of recombinant human growth hormone treatment on psychological profiles in hypopituitary patients with adult-onset growth hormone deficiency Eur J Clin Invest 1998 28 13 19 9502182 10.1046/j.1365-2362.1998.00239.x
Hernberg-Stahl E Luger A Abs R Bengtsson BA Feldt-Rasmussen U Wilton P Westberg B Monson JP Healthcare consumption decreases in parallel with improvements in quality of life during GH replacement in hypopituitary adults with GH deficiency J Clin Endocrinol Metab 2001 86 5277 5281 11701692 10.1210/jc.86.11.5277
Murray RD Skillicorn CJ Howell SJ Lissett CA Rahim A Smethurst LE Shalet SM Influences on quality of life in GH deficient adults and their effect on response to treatment Clin Endocrinol (Oxf) 1999 51 565 573 10594517 10.1046/j.1365-2265.1999.00838.x
Sartorio A Molinari E Riva G Conti A Morabito F Faglia G Growth hormone treatment in adults with childhood onset growth hormone deficiency: effects on psychological capabilities Horm Res 1995 44 6 11 7649527
Soares CN Musolino NR Cunha NM Caires MA Rosenthal MC Camargo CP Bronstein MD Impact of recombinant human growth hormone (RH-GH) treatment on psychiatric, neuropsychological and clinical profiles of GH deficient adults. A placebo-controlled trial Arq Neuropsiquiatr 1999 57 182 189 10412515
Wallymahmed ME Foy P Shaw D Hutcheon R Edwards RH MacFarlane IA Quality of life, body composition and muscle strength in adult growth hormone deficiency: the influence of growth hormone replacement therapy for up to 3 years Clin Endocrinol (Oxf) 1997 47 439 446 9404442 10.1046/j.1365-2265.1997.2801076.x
Wiren L Bengtsson BA Johannsson G Beneficial effects of long-term GH replacement therapy on quality of life in adults with GH deficiency Clin Endocrinol (Oxf) 1998 48 613 620 9666873 10.1046/j.1365-2265.1998.00462.x
|
16236167
|
PMC1277839
|
CC BY
|
2021-01-04 16:38:15
|
no
|
Health Qual Life Outcomes. 2005 Oct 19; 3:63
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-63
|
oa_comm
|
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-161624889810.1186/1479-5868-2-16ResearchOut-of-home food outlets and area deprivation: case study in Glasgow, UK Macintyre Sally [email protected] Laura [email protected] Steven [email protected] Cate [email protected] Social & Public Health Sciences Unit, Medical Research Council, Glasgow, UK2 Department of Geography, Queen Mary, University of London, London, UK3 School of Exercise & Nutrition Science, Deakin University, Australia2005 25 10 2005 2 16 16 1 8 2005 25 10 2005 Copyright © 2005 Macintyre et al; licensee BioMed Central Ltd.2005Macintyre 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 a popular belief that out-of-home eating outlets, which typically serve energy dense food, may be more commonly found in more deprived areas and that this may contribute to higher rates of obesity and related diseases in such areas.
Methods
We obtained a list of all 1301 out-of-home eating outlets in Glasgow, UK, in 2003 and mapped these at unit postcode level. We categorised them into quintiles of area deprivation using the 2004 Scottish Index of Multiple Deprivation and computed mean density of types of outlet (restaurants, fast food restaurants, cafes and takeaways), and all types combined, per 1000 population. We also estimated odds ratios for the presence of any outlets in small areas within the quintiles.
Results
The density of outlets, and the likelihood of having any outlets, was highest in the second most affluent quintile (Q2) and lowest in the second most deprived quintile (Q4). Mean outlets per 1,000 were 4.02 in Q2, 1.20 in Q4 and 2.03 in Q5. With Q2 as the reference, Odds Ratios for having any outlets were 0.52 (CI 0.32–0.84) in Q1, 0.50 (CI 0.31 – 0.80) in Q4 and 0.61 (CI 0.38 – 0.98) in Q5. Outlets were located in the City Centre, West End, and along arterial roads.
Conclusion
In Glasgow those living in poorer areas are not more likely to be exposed to out-of-home eating outlets in their neighbourhoods. Health improvement policies need to be based on empirical evidence about the location of fast food outlets in specific national and local contexts, rather than on popular 'factoids'.
==== Body
Background
Obesity is associated with a range of disorders including coronary heart disease, diabetes, kidney failure, osteoarthritis, cancer, back pain, and psychological damage [1]. Rates of overweight and obesity are high, and rising, in developed countries, and considerable concern has been expressed in a number of countries about this increase [2-4], described by the Chief Medical Officer for England as 'a ticking time bomb' [5].
The increasing prevalence of overweight and obesity has been linked to increasing physical inactivity and changes in eating patterns [6]. There has been an increase in the consumption of foods outside the home, and increases in portion size in out-of-home outlets (particularly 'fast food' outlets) [7-9]. In the UK in 2002–3, 27% of expenditure on food and drinks (excluding alcohol) was spent on consumption outside the home [10].
Foods eaten outside the home are often higher in energy and fat than foodstuffs prepared at home ; and it has been suggested that:
'the high energy densities of many fast foods challenge human appetite control systems with conditions for which they were not designed. Among regular consumers this is likely to result in the accidental consumption of excess energy and hence to promote weight gain and obesity' [7] p187.
Studies of fast food restaurant use have shown positive associations with intake of total energy and percent fat, and negative associations with intakes of fibre [11,12]. A longitudinal study found that people who ate meals from fast-food restaurants more than twice a week, at both baseline and 15 year follow up, gained 4.5 kilos more weight and had 104% greater increase in insulin resistance than people who ate less than one meal at a fast-food restaurant each week [13]. One study in the USA found, at the state level, a correlation between density measures of fast-food restaurants (per resident and per square mile) and obesity rates [14], although another study found that childhood overweight was not associated with proximity to fast-food restaurants in Cincinnati [15].
Dietary intake among poorer socio-economic groups is less likely to meet current nutritional guidelines for fruit and vegetable consumption [16], and more likely to be high in fat, salt and sugar (typically features of fast food) [17]. In the UK this applies to socio-economic status (SES) whether measured by household occupational social class or area-based measures of deprivation [18,19]. In developed countries obesity rates in women rise linearly with decreasing SES; the pattern for men is less straightforward and there are less steep, and in some places no or non linear, SES gradients in men [18-21]. However, area deprivation has consistently been shown to be related to overweight and obesity, even in models which take individual socio-demographic characteristics into account [22-24].
This observation has led to researchers to hypothesise that deprived areas may have fewer outlets selling healthy foods at affordable prices [25], but be better supplied by fast food outlets selling foods that are high fat and energy dense [26]. Evidence on the first point is equivocal, some studies finding poorer access in more deprived areas to supermarkets selling foods recommended in current dietary guidelines [27-30], but other studies finding either no socio-economic differences in shop locations or more supermarkets in poorer areas [31-37].
There is less available evidence on the second point. A study in Melbourne, Australia, found that there was a dose-response relation between SES and the density of fast food outlets, with people living in the poorest SES areas having 2.5 times greater exposure to fast food outlets than people in the wealthiest areas [26]. A study in New Orleans found that people living in low income and predominantly black areas had significantly more exposure to fast food restaurants [38]. Cummins et al found that in England and Scotland there was a higher density of McDonald's restaurants per thousand population in more deprived areas [39]. Although these are the only existing empirical studies on this issue, it is commonly suggested that residents of poorer neighbourhoods are more exposed to fast food outlets, e.g. in a frequently cited comprehensive review of obesity it is stated that:
'Poorer neighbourhoods tend to have fewer recreation amenities, be less safe, and have a higher concentration of fast food outlets' [20]p133.
Scotland has high rates of premature mortality which have often been attributed to poor diets, in particular high consumption of sugar, fried foods, and carbonated soft drinks [18,40]. Glasgow has a reputation for poor health and poor diet [41] and contains a higher proportion of deprived neighbourhoods than any other area in Scotland [42]. However, Glasgow has a considerable range of health and deprivation indices. We therefore considered it a good location for a case study which aimed to test the hypothesis that fast food outlets are more likely to be found in poorer neighbourhoods in the UK.
Design and Methods
Identification and classification of food outlets
We obtained lists of out-of-home food outlets in Glasgow in 2003 from the Food Safety Unit of Glasgow City Council. All food premises must be registered with the Council in accordance with the Food Safety (General Food Hygiene) Act 1995. A database was created containing all food outlets listed as restaurant, café or takeaway by the Council, and the full unit postcode of each outlet. This categorisation was mutually exclusive and based on the main activity of the outlet. However in practice these categories were not mutually exclusive; many restaurants or cafes (including Indian restaurants, fish and chip restaurants, burger restaurants, coffee shops and ice cream parlours) also provide a take away service. Restaurants include fine dining independent restaurants, vegetarian restaurants, ethnic restaurants, and fast food restaurants so their sales of high energy high fat foods may vary considerably. We separated restaurants into restaurants (independent and chain) or fast food chain. Chain fast food restaurants are national or international multi-outlet companies or franchises, whose outlets provide tables and chairs, but no crockery or cutlery, and counter service only; in Glasgow these include McDonalds, Burger King, KFC, Pizza Hut, and Wimpy. Many independent restaurants in Glasgow specialise in 'ethnic' food and are categorised in the local yellow pages as such (for example, Italian, Indian, Thai, Chinese etc). Many takeaways in Glasgow sell a range of foodstuffs (e.g. chips, pies, pizzas, curry, kebabs, burgers) all of which are typically energy dense.
Classification of neighbourhood deprivation
Our primary measure of neighborhood deprivation was the Scottish Index of Multiple Deprivation (SIMD), created by the government for monitoring and planning purposes. The SIMD is calculated using data such as current income, employment, health, education, skills and training, telecommunications, and housing at the level of data zones [43]. Data zones are statistical areas which nest within local authority boundaries and are smaller than postcode sectors or wards. Data zones are intended to be effective in identifying small areas with particular social characteristics, and are therefore more internally homogeneous than postcode sectors. Look-up tables are available which relate individual post codes to data zones [44]. As the SIMD score increases the level of deprivation increases. We divided data zones into quintiles by SIMD scores (Quintile 1 covers 138 data zones, and quintiles 2–5 each cover 139). The mean population per data zone was highest (at 859.2) in quintile 1 (the least deprived) and lowest (at 815.1) in quintile 5 (the most deprived); the overall mean was 832.7 persons, standard deviation 152.6.
Initially we also used another area-based measure of deprivation, Carstairs scores, in addition to SIMD. This was because we were unsure about the appropriate spatial scale at which to measure access to out of home food outlets, and wished to check whether the spatial scales used made any difference to the analysis. Carstairs scores are based on the proportions of: overcrowded households, heads of household in social classes IV and V, unemployed male heads of household, and non-owner occupied properties, at the post code sector level using data in from the 2001 Census [42]. Post code sectors in Glasgow have a mean population size of 5556 and standard deviation of 3109. Four postcode sectors (G1 3, G2 2, G2 5, G2 6), containing 102 outlets (57 restaurants, 5 fast food chain restaurants, 21 cafes and 29 takeaways), do not have Carstairs scores, because of small populations. We found the results using Carstairs (results not shown, available from authors) followed a similar pattern to that using SIMD but we report the latter analysis only, as all food outlets and areas of the city were included, and data zones were smaller and less variable in size than post code sectors.
We are not assuming that people are restricted to food outlets within their own data zone. However, small areas in the UK are very clustered by deprivation (i.e. deprived areas tend to be adjacent to other deprived areas, and affluent areas to affluent areas) so if out of home outlets are concentrated in deprived areas these are also likely to effect the exposure of residents in adjacent, also deprived, areas.
Analysis strategy
The mean number of eating outlets per 1000 persons was calculated using population data (2001 Census) for each data zone within each SIMD quintile. Statistical comparison between quintiles in outlet density was determined by ANOVA. Accepted level of significance was p < 0.05. More than half of the data zones had no outlets of any kind, so we also used Logistic Regression to determine the probability, in terms of Odds Ratios (OR) and Confidence Intervals (CI), of data zones within quintiles having any of the various outlets. Data zones grouped within Quintile 2 were used as the reference category since they contained most food outlets of all kinds. All analysis used SPSS Version 12.0 for Windows.
We also used MapInfo Professional Client 6.0 to map the different outlets on a base map of Glasgow to describe their distribution geographically to complement the statistical analysis by deprivation (map 1).
Results
We identified 1301 out-of-home eating outlets in Glasgow; 339 restaurants, 30 fast food chain restaurants, 303 cafes and 629 takeaways. Thirty five per cent of these outlets, and nearly 50% of the fast food chain restaurants, were located in Q2, the second least deprived quintile (see table 1). The highest density per thousand population of each type of outlet, and for all combined, was in Q2 and the lowest density was in Q4. Differences between quintiles were statistically significant for restaurants and for takeaways.
Table 1 Mean number of food premises per 1000 people per SIMD Quintile
Restaurants Fast food chains Cafés Takeaways All outlets
Mean N Mean N Mean N Mean N Mean N
*SIMD Quintile (population)
1 Most Affluent (118,568) 0.58 66 0.02 3 0.37 43 0.61 71 1.58 183
2 (114,744) 1.51 164 0.13 14 0.77 88 1.61 192 4.02 458
3 Middling (116,074) 0.48 58 0.05 6 0.61 70 1.36 149 2.51 283
4 (115,178) 0.12 15 0.00 0 0.26 29 0.82 92 1.20 136
5 Most deprived (113,305) 0.29 36 0.07 7 0.61 73 1.06 125 2.03 241
Total (577,869) 0.60 339 0.05 30 0.52 303 1.09 629 2.27 1301
Sig (ANOVA) 0.045 0.300 0.260 0.041 0.092
*SIMD Quintile 1 includes 138 Data zones while Quintiles 2–5 include 139 Data zones each.
Table 2 outlines the results from the logistic regression analysis. Using Q2 as the reference category there was a statistically significant reduced probability of having any restaurants in data zones within Q3 (OR 0.45, 95% CI 0.24 to 0.82), Q4 (OR 0.21, 95% CI 0.10–0.43), and Q5 (OR 0.18, 95% CI 0.09–0.40). The probability of restaurants being present in data zones in Q1 was also lower but confidence intervals included 1 (OR 0.77, 95% CI 0.45–1.33). Confidence intervals were wide for fast food chain restaurants given their small number and estimates are thus unreliable. For cafes, the probability of data zones having any outlets was lower in each quintile than in the reference category, but only in Q4 did the confidence interval for the estimated odds ratio not include 1 (OR 0.40, 95% CI 0.22–0.73). Takeaways showed no clear pattern with the only statistically significant result being for Q1 (OR 0.49, 95% CI 0.29–0.81). Finally, combining all types of outlet, the probability of data zones having any outlet was lower in quintiles 1, 3, 4 and 5 than 2, though for Q3 the difference was non-significant (95% CI 0.56–1.42). There was no evidence of a linear relationship between any of the categories of outlet and deprivation. Overall there was no evidence that poorer areas in Glasgow were more likely to contain any out-of-home eating outlets.
Table 2 Probability of data zones within SIMD quintiles containing any outlets (in comparison to quintile 2); Odds Ratios (OR) and 95% Confidence intervals (CI). Significant Odds ratios (OR) in Italics.
Restaurants Fast food chains Cafés Takeaways All outlets
OR 95%CI OR 95%CI OR 95%CI OR 95%CI OR 95%CI
Quintile
1 Affluent 0.77 0.45–1.33 0.49 0.12–2.10 0.61 0.35–1.06 0.49 0.29–0.81 0.52 0.32–0.84
2 (reference) 1.00 1.00 1.00 1.00 1.00
3 Middling 0.45 0.24–0.82 0.83 0.25–2.78 0.90 0.53–1.52 1.23 0.76–1.98 0.89 0.56–1.42
4 0.21 0.10–0.43 0.00 0.00-0.00 0.40 0.22–0.73 0.71 0.43–1.16 0.50 0.31–0.80
5 Deprived 0.18 0.09–0.40 0.66 0.18–2.38 0.77 0.46–1.32 0.78 0.48–1.27 0.61 0.38–0.98
Overall sig. 0.00 0.89 0.03 0.01 0.01
*SIMD Quintile 1 includes 138 Data zones while Quintiles 2–5 include 139 Data zones each.
The map displays the location of all outlets and shows that out-of-home food outlets tend to be located in the city centre (which is a retail centre, theatre/concert hall/cinema/club/pub area, Central Business District, and transport hub containing the main bus station and two train stations); along arterial highways (Great Western Road, Maryhill Road, London Road); and in cosmopolitan areas which are busy with workers during the day and have active night-time economies (e.g. Byres Road, Pollokshaws Road). From maps overlain with quintiles of area deprivation it appears that all types of outlet tend to be more common in City Centre and West End type areas, and to be less common in the peripheral deprived social housing areas such as Easterhouse, Castlemilk, Drumchapel, and Pollok, and in the most affluent housing areas in the City (maps not shown, available from authors).
Figure 1 Map of Restaurants, Fast food Chain restaurants, Cafes and Takeaways in Glasgow City, 2003.
Discussion
Whereas studies in Melbourne and New Orleans have found that fast food outlets were more concentrated in low income and black areas, we have found that neither out-of-home outlets in general, nor takeaways or fast food chain restaurants in particular, were more likely to be found in more deprived areas of Glasgow. To the contrary, we found that there were more outlets in the second most affluent category, and that outlets tended to be located in inner city or West End areas which attract customers during the day (e.g. retail and business centres) and evening (e.g. leisure centres). There are few outlets in areas which are primarily residential.
These findings may be unique to Glasgow and its particular history of urban development and planning, but we suspect the pattern we observed may be common in other UK cities. This is because restaurants and takeaways are likely to be located where there is most potential custom both during the day and evenings, and such demand is higher in retail, transport and commercial centres, areas with high density of entertainment facilities such as cinemas, theatres and pubs, and along arterial highways with much passing traffic. 'Gentrification' has involved the movement of socio-economically advantaged individuals into Glasgow City centre (e.g. 'the Merchant City') and the West End. Deprived areas in Glasgow are primarily residential and have often been noted to be lacking a whole range of local amenities (such as public transport, schools etc), especially the peripheral public housing estates (Castlemilk, Eaterhouse, Pollok, Drumchapel; see map) which were built immediately after the second world war to alleviate appalling housing conditions in the inner city. Thus the location of out-of-home eating outlets may reflect rational responses on the part of owners or franchisees to principles of supply and demand. This pattern may be different from those observed in the USA, where deprived populations may be concentrated more in the inner city and wealthier people dispersed to suburbs [34,45], and from Australia, where it has been argued by some that there is less spatial segregation along social and economic lines [46]. The different measures of deprivation used in the New Orleans, Melbourne, and Glasgow studies may also contribute to different findings.
Despite the fact that the location of out-of-home eating outlets may be a rational response to likely demand, there seems to be a prevailing assumption that such outlets, particularly fast food outlets, are targeted at deprived communities and are part of the reason for poorer diets and higher obesity levels in poor places. In an article in The Observer, a national Sunday newspaper, David Smith wrote about Shettleston, a relatively deprived area in Glasgow, which has one the worst life expectancy records in the UK. He gave considerable emphasis to the density of fast food outlets:
'At the front of Celtic FC's Parkhead Stadium, children queue for a burger bar at 10 in the morning....(and) Bridgeton, within the Shettleston constituency, is possibly the alcohol and fast food capital of Britain. Within a radius of just 200 yards around the metro station there are nine pubs, an off licence and seven takeaways'
and he quotes several local people similarly emphasising fast food availability:
"The children go to Burger Kings or McDonalds, and there's nothing you can do"...."There is a fast food shop at every corner. Going to those places becomes a habit" [47].....
While such stereotypes are common, they are not supported by the findings of this study. Though there are a number of out-of-home outlets along Shettleston Road and London Road (see map) their presence there is no more dense than along other main roads including ones in more middle class areas such as the West End. Just as perceptions on the part of the mass media, policymakers and food activists to the effect that deprived communities have poorer access to healthy foods at affordable prices may not be borne out by the empirical evidence, [34] perceptions held about a greater exposure among deprived communities to unhealthy diets may not necessarily be empirically substantiated.
There are clearly some limitations to our study. It is restricted to one city (as are those undertaken in Melbourne and New Orleans). By definition, we have not been able to map the location of mobile fast food outlets (e.g. vans selling pies, burgers etc) and it may be that these target poorer residential areas. It should also be noted that although takeaways and the chain restaurants primarily serve what is conventionally called fast food (i.e. high in fat, energy, and salt), they, and other restaurants and cafes, may serve healthier options instead of, or as well as, fast foods. Our interpretation of the maps is descriptive rather than using a more sophisticated GIS approaches.
Our findings may not appear consistent with those from another study in which we assessed the location of McDonald's restaurants in England and Scotland in relation to deprivation in 2005, and found that these restaurants were significantly more likely to be found in more deprived neighbourhoods[39]. Large global chains such as McDonald's have sophisticated marketing systems that allow them to pin-point with some accuracy the 'optimum' location based on geo-demographics, distance from headquarters, distribution points and main transport links and intersections, and sales figures in existing outlets [9,48] whereas the locational strategies of independent outlets may be more locally based. It may also be that individual chains avoid areas where others are located, so that if one chain tends to be located in more deprived neighbourhoods, others will locate in less deprived ones and so balance each other out.
The association between area level social deprivation and the availability of foods prepared outside the home, in particular fast foods, must be investigated in the context of the local availability of land, the price of real estate and the ease of obtaining planning permission. Differences in these factors may explain the differences between the findings reported here and those reported nationally for a single large global chain, and in other countries.
With increasing evidence of the possible risks to health of some types of fast food (because of portion sizes, energy density, and fat and salt content), it is important to establish whether the popular assumption that proximity to fast food outlets tends to lead to greater consumption of such foods and subsequently higher rates of obesity and poor health is substantiated. Further critical evaluation of the role of access to foods eaten outside the home in the aetiology of obesity is warranted. This relationship may well be more complex than simple proximity to an outlet, and may vary with macro and more local cultural and socioeconomic factors. As we have previously argued [49], it is important that health promotion policies in relation to the predicted obesity epidemic are based on robust empirical evidence and sensitivity to cultural and socioeconomic context, rather than on untested assumptions or 'factoids'.
Acknowledgements
Sally Macintyre and Laura McKay are employed by the UK Medical Research Council, from whom Steven Cummins holds a fellowship. We are grateful to the Food Safety Unit at Glasgow City Council for supplying the data, and Anne Ellaway and Mary-Kate Hannah for comments and help with the analysis.
==== Refs
House of Commons Health Committee Obesity. Third Report of Session 2003-04, Volume 1. Report, together with formal minutes 2004 London, The Stationery Office
Royal College of Physicians Royal College of Paediatrics and Child Health Faculty of Public Health Medicine Storing up problems: The medical case for a slimmer nation 2004 London, RCP Publications
World Health Organization Obesity: preventing and managing the global epidemic. Report of a WHO consultation. WHO Technical Series 2000 Geneva, World Health Organization
Flegal KM Caroll MD Ogden CL Johnson CL Prevalence and trends in obesity among US adults, 1999-2000 Journal of American Medical Association 2002 288 1723 1727 10.1001/jama.288.14.1723
Department of Health Health Check. Annual report of the Chief Medical Officer 2002 2003 London, Department of Health
Prentice AM Jebb SA Obesity in Britain: gluttony or sloth? British Medical Journal 1995 311 437 439 7640595
Prentice AM Jebb SA Fast foods, energy density and obesity: a possible mechanistic link Obesity Reviews 2003 4 187 194 14649369 10.1046/j.1467-789X.2003.00117.x
French SA Story M Jeffrey RW Environmental Influences on Eating and Physical Activity Annual Review of Public Health 2001 22 309 335 11274524 10.1146/annurev.publhealth.22.1.309
Schlosser E Fast Food Nation 2001 Boston, Houghton Mifflin
Department for Environment Food and Rural Affairs Family Food 2002-2003 2004 London, TSO
McCrory A Fuss PJ Hays NP Vinken AG Greenberg AS Roberts SB Overeating in America: association between restaurant food consumption and body fatness in healthy adult men and women ages 19 to 80 Obesity Research 1999 7:564-71
French SA Harnack L Jeffrey RW Fast food restaurant use among women in the Pound of Prevention study: dietary, behavioral and demographic correlates. International Journal of Obesity Related Metabolic Disorders 2000 24 1353 1359 10.1038/sj.ijo.0801429
Pereira M Kartashov A Ebbeling C Van Horn L Slattery M Jacobs D Ludwig D Fast-food habits, weight gain, and insulin resistance (the CARDIA study): 15-year prospective analysis Lancet 2005 365 36 42 15639678 10.1016/S0140-6736(04)17663-0
Maddock J The relationship between obesity and the prevalence of fast food restaurants:state-level analysis American Journal of Health Promotion 2004 19 137 143 15559714 10.1093/heapro/dah201
Burdette H Whitaker R Neighbourhood playgrounds, fast food restaurants, and crime: relationships to overweight in low income pre-school children Preventive Medicine 2004 38 57 63 14672642 10.1016/j.ypmed.2003.09.029
Billson H Pryer JA Nichols R Variation in fruit and vegetable consumption among adults in Britain. An analysis from the dietary and nutritional survey of British adults. European Journal of Clinical Nutrition 1999 53 946 952 10602352 10.1038/sj.ejcn.1600877
Leather S The making of modern malnutrition; an overview of food poverty in the UK 1996 London, Caroline Walker Trust
Shaw A McMunn A Field J The Scottish Health Survey 1998: vol 1 2000 1 Edinburgh, The Stationery Office
Sprosten K Primatesta P Health Survey for England 2002 2004 London, TSO
Swinburn B Caterson I Seidall J James W Diet, nutrition and prevention of excess weight gain and obesity Public Health Nutrition 2004 7 123 146 14972057 10.1079/PHN2003585
Macintyre S Hunt K Socioeconomic position, gender and health: how do they interact? Journal of Health Psychology 1997 2 315 334
Ellaway A Anderson A Macintyre S Does area of residence affect body size and shape? International Journal of Obesity 1997 21 304 308 9130028 10.1038/sj.ijo.0800405
van Lenthe FJ Mackenbach JP Neighbourhood deprivation and overweight: the GLOBE study International Journal of Obesity Related Metabolic Disorders 2002 26 234 240 10.1038/sj.ijo.0801841
Sundquist J Malstrom M Johansson SE Cardiovascular risk factors and the neighbourhood environment: a multilevel analysis International Journal of Epidemiology 1999 28 841 845 10597980 10.1093/ije/28.5.841
Sooman A Macintyre S Anderson A Scotland's health - a more difficult challenge for some? The price and availability of healthy foods in socially contrasting localities in the West of Scotland Health Bull 1993 51 276 284
Reidpath DD Burns C Garrard J Mahoney M Townsend M An ecological study of the relationship between social and environmental determinants of obesity Health & Place 2002 8 141 145 11943585 10.1016/S1353-8292(01)00028-4
Morland K Wing S Deiz Roux A Poole C Neighbourhood characteristics associated with the location of food stores and food service places American Journal of Preventive Medicine 2002 22 23 29 11777675 10.1016/S0749-3797(01)00403-2
Guy C David G Measuring physical access to 'healthy foods' in areas of social deprivation: a case study in Cardiff International Journal of Consumer Studies 2004 28 222 234 10.1111/j.1470-6431.2003.00340.x
Chung C Myers S Do the poor pay more for food? An analysis of grocery store availability and food price disparities Journal of Consumer Affairs 1999 33 276 296
Zenk S Schulz A Israel B Kames S Bao S Wilson M Neighborhood racial composition, neighbourhood poverty, and the spatial accessibility of supermarkets in metropolitan Detroit American Journal of Public Health 2005 95 660 667 15798127 10.2105/AJPH.2004.042150
Cummins S Macintyre S The location of food stores in urban areas: a case study in Glasgow British Food Journal 1999 101 545 553 10.1108/00070709910279027
Cummins S Macintyre S A systematic study of an urban foodscape: the price and availability of food in Greater Glasgow Urban Studies 2002 39 2115 2130 10.1080/0042098022000011399
White M Williams E Raybould S Adamson A Bunting J Mathers J Do food deserts exist? A multi-level geographical analysis of the relationship between retail food access, socio-economic position and dietary intake.(Final report to Food Standards Agency) 2004 Newcastle, School of Population and Health Sciences, University of Newcastle upon Tyne
Winkler E Turrell G Patterson C Does living in a disadvantaged area mean fewer opportunities to purchase fresh fruit and vegetables in the area? Findings from the Brisbane food study Health & Place
van Lenthe FJ Brug J Mackenbach JP Neighbourhood inequalities in physical inactivity: the role of neighbourhood attractiveness, proximity to local facilities and safety in the Netherlands Social Science & Medicine 2005 60 763 775 15571894 10.1016/j.socscimed.2004.06.013
Yen IH Kaplan GA Neighborhood social environment and risk of death: multilevel evidence from the Alameda County Study American Journal of Epidemiology 1999 149 898 907 10342798
Competition Commission Supermarkets: A report on the supply of groceries from multiple stores in the United Kingdom (3 vols). CM 4842 2000 London, The Stationery Office
Block JP Scribner RA DeSalvo KB Fast Food, Race/Ethnicity, and Income American Journal of Preventive Medicine 2004 27 211 217 15450633 10.1016/j.amepre.2004.06.007
Cummins S McKay L Macintyre S McDonald's restaurants and neighbourhood deprivation in Scotland and England American Journal of Preventive Medicine
Scottish Office Department of Health Eating for Health, a diet action plan for Scotland 1996 Edinburgh, Scottish Office
Wrieden WL Connaghan J Morrison C Tunstall-Pedoe H Secular and socio-economic trends in compliance with dietary targets in the north Glasgow MONICA population surveys 1986-1995;did social gradients widen? Public Health Nutrition 2004 7 835 842 15482607 10.1079/PHN2004636
McLoone P Carstairs scores for Scottish postcode sectos from the 2001 Census 2004 Glasgow, MRC Social and Public Health Sciences Unit
Scottish Executive Scottish Index of Mutliple Deprivation 2004
Scottish Executive Scottish Neighbourhood Statistics Datazones 2004 (CD ROM)
Acevedo-Garcia D Lochner KA Kawachi I and Berkman L Residential Segregation and Health Neighbourhoods and Health 2003 Oxford, Oxford University Press 265 281
Turrell G Blakely T Patterson C Oldenburg B A multilevel analysis of socioeconomic (small area) differences in household food purchasing behaviour Journal of Epidemiology & Community Health 2004 58 208 215 14966233 10.1136/jech.2003.011031
Smith D You'll be lucky to reach 60 here. But its not the Third World, it's East Glasgow. Shettleston's diet of chips, fags and booze means that life expectancy is actually falling in one of the most deprived parts of the UK The Observer, 2004 London, England,
Sault J Toivanen O Waterson M Fast food- the early years: Geography and the growth of a chain store in the UK Warwick Economic Research Papers no 655 2002 Coventry, University of Warwick
Cummins A Macintyre S 'Food deserts' - evidence and assumption in health policy making British Medical Journal 2002 325 436 438 12193363
|
16248898
|
PMC1277840
|
CC BY
|
2021-01-04 16:38:04
|
no
|
Int J Behav Nutr Phys Act. 2005 Oct 25; 2:16
|
utf-8
|
Int J Behav Nutr Phys Act
| 2,005 |
10.1186/1479-5868-2-16
|
oa_comm
|
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-261626289310.1186/1476-072X-4-26MethodologyOpportunities for using spatial property assessment data in air pollution exposure assessments Setton Eleanor M [email protected] Perry W [email protected] C Peter [email protected] Spatial Sciences Research Laboratory, Geography Department, University of Victoria, PO BOX 3050 STN CSC, Victoria, B.C., V8W 3P5, Canada2005 31 10 2005 4 26 26 14 9 2005 31 10 2005 Copyright © 2005 Setton et al; licensee BioMed Central Ltd.2005Setton 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 epidemiological studies examining the relationships between adverse health outcomes and exposure to air pollutants use ambient air pollution measurements as a proxy for personal exposure levels. When pollution levels vary at neighbourhood levels, using ambient pollution data from sparsely located fixed monitors may inadequately capture the spatial variation in ambient pollution. A major constraint to moving toward exposure assessments and epidemiological studies of air pollution at a neighbourhood level is the lack of readily available data at appropriate spatial resolutions. Spatial property assessment data are widely available in North America and may provide an opportunity for developing neighbourhood level air pollution exposure assessments.
Results
This paper provides a detailed description of spatial property assessment data available in the Pacific Northwest of Canada and the United States, and provides examples of potential applications of spatial property assessment data for improving air pollution exposure assessment at the neighbourhood scale, including: (1) creating variables for use in land use regression modelling of neighbourhood levels of ambient air pollution; (2) enhancing wood smoke exposure estimates by mapping fireplace locations; and (3) using data available on individual building characteristics to produce a regional air pollution infiltration model.
Conclusion
Spatial property assessment data are an extremely detailed data source at a fine spatial resolution, and therefore a source of information that could improve the quality and spatial resolution of current air pollution exposure assessments.
==== Body
Background
Many epidemiological studies examining the relationships between adverse health outcomes and exposure to air pollutants use ambient air pollution measurements as a proxy for personal exposure levels [1]. Because the number of fixed outdoor monitoring sites within a city usually is limited, ambient pollution measurements often are extrapolated to areas between monitors, thus disregarding any neighbourhood-scale spatial variation in pollution levels. Recent research suggests that some neighbourhoods within a city can be disproportionately exposed to air pollution and that these differences may influence health outcomes [2]. A major constraint to moving toward exposure assessments and epidemiological studies of air pollution at a neighbourhood level is the lack of readily available data at appropriate spatial resolutions.
Spatial property assessment data (SPAD) were identified as a potential data source for exposure research through an ongoing study, funded by Health Canada via the British Columbia Centre for Disease Control, examining the effects of air pollution on birth outcomes and subsequent development of health outcomes associated with exposure to air pollution for a birth cohort of 90,000. The study area encompasses the Georgia Basin Puget Sound airshed, located in the Pacific Northwest of the United States and Canada and encompassing approximately 10 million hectares of land and marine environments. SPAD (considered here to be made up of both tabular assessment data on building characteristics and spatial data that show the location of each property) for every assessed parcel of land are generally available for the airshed as spatially referenced digital databases, suitable for use with common geographic information systems (GIS). These data may increase the resolution and accuracy of variables used in exposure assessment models and epidemiological analyses, but the authors have found few published studies using SPAD for developing exposure assessments or in epidemiological analyses of air pollution impacts on health.
The purpose of this paper is to describe SPAD and to illustrate their potential utility for neighbourhood level exposure estimates and epidemiological research. Three possible uses of SPAD are examined, including: (1) creating variables for use in land use regression modelling of neighbourhood levels of ambient air pollution; (2) enhancing wood smoke exposure estimates by mapping fireplace locations; and (3) using data available on individual building characteristics to produce a regional air pollution infiltration model.
Results and discussion
Typical SPAD characteristics
Tabular assessment data generally incorporate two kinds of information: 1) property addresses or property identifiers that can be used for spatial referencing in conjunction with digital street networks or digital cadastral maps; and 2) descriptive variables including building characteristics, building and land values, and land use information on which annual property tax assessments are based. Table 1 provides examples of common variables in tabular assessment data that may be important to air pollution exposure assessments and epidemiological analyses.
Table 1 Common variables in tabular assessment data
Location School District #, Area #, Township Range, Jurisdiction #, Neighbourhood #, Street Address, Street Direction, Street Type, ZIP Code, City, Property Identifier
Land Appraisal Date, Property Size, Property Use Code, Land Use Code, Electricity, Water, Sewer, Street Surface Type, # of Dwelling Units, # of Outbuildings, # of Improvements, Building Permit.
Sales Sale Date, Sale Price, Sales Excise Number, Deed Type, Qualification Code, Multiple Sales, Land Value, Improvements Value.
Improvements Improvement Type, Structure Use, Building Type, # of Stories, Year Built, Total Square Footage, # of Bedrooms, Predominant Heating Type, Fireplace, Structural Quality.
SPAD are created when tabular property assessment data are spatially referenced, either by linking property addresses to a digital street network, or by linking property identifiers to a digital cadastral map. Spatial referencing allows for complex queries of the tabular assessment, the results of which can be mapped. In effect, the spatial resolution of SPAD is the individual parcel size, generally much finer than other spatially referenced data commonly used in exposure assessment and epidemiological analyses, such as Census data.
Data format and availability
SPAD are developed and maintained by numerous jurisdictions throughout North America, therefore data format and content may differ significantly among jurisdictions. The following discussion highlights some of the differences and associated issues in British Columbia and Washington State, as shown in Figure 1. In British Columbia, tabular assessment data are collected by the BC Assessment Authority and maintained as a tabular database, while each taxing municipality or regional district develops its own cadastral data that can be linked with assessment data to create SPAD. By agreement, BC Assessment uses a unique property identifier for each assessment record, and the same property identifiers are used by taxing jurisdictions when developing cadastral data, thereby enabling a linkage between the tabular assessment data from BC Assessment and the jurisdiction's cadastral data using GIS. In Washington State, each county is responsible for developing both the tabular assessment data and the cadastral data. In this regard, there is some advantage in the British Columbia system, in that one single authority collects and maintains the tabular assessment data and so these are standard throughout the entire province. Unfortunately, when multiple jurisdictions are responsible for developing tabular assessment data and/or cadastral data, there can be significant differences in format among jurisdictions. For example, in British Columbia, cadastral data are often available in ESRI© GIS formats, but in some cases are only available in AutoCad© format. The latter is primarily an engineering and drafting application and the format may not always translate easily into GIS formats. In Washington State, neither the tabular assessment data nor the cadastral data may be standardized among counties, as each develop and maintain their own information systems. In this case, it is possible that some tabular assessment data (i.e., presence of air conditioner) are collected for one county, but not for the adjacent county, or that different GIS applications are used by different counties.
Figure 1 The development of SPAD differs between British Columbia and Washington State.
Access to SPAD (or its constituent tabular and cadastral data) is markedly different in British Columbia in comparison to Washington State. In British Columbia, researchers must negotiate data sharing or purchasing agreements with each jurisdiction in order to access SPAD, and may also have to purchase additional tabular assessment data directly from the BC Assessment Authority in order to develop SPAD specific to the research question. In Washington State, SPAD are available for download through each county's internet site, or may be ordered directly from each county at no cost or for a small fee (i.e., for CD writing and postage). In many cases, due to large file sizes, the tabular assessment data and the spatial cadastral data are provided separately, and must be linked by the researcher using GIS to create the final SPAD.
Linking tabular assessment data using property addresses or identifiers to produce SPAD is not always trouble free. In cases where tabular assessment data and the spatial cadastral data are provided by the same jurisdiction, linking the two datasets often is easily accomplished. In Washington State, for example, where each county develops and maintains its own SPAD, we were able to download the tabular assessment data and the spatial cadastral data, and link each record with a 98 percent success rate. For the British Columbia portion of the airshed, we initially purchased tabular assessment data from the BC Assessment Authority and spatially referenced them using the included property addresses and a commercially available digital street network with ESRI© ArcGIS 8.1. Approximately 1.1 million records were received from BC Assessment for the entire Georgia Basin airshed, which is comprised of 26 separate taxing jurisdictions. Linking between the tabular assessment data and the street network was successfully completed for approximately 83 percent of the records, with the number of links in urban areas better than in rural regions (89 percent versus 67 percent respectively). The lower success rate in rural regions is generally due to incomplete or non-standard street addresses (i.e., post office boxes or rural post offices rather than street addresses) in the tabular assessment data. Also, the road network (circa 2003) did not contain information on the most recent subdivisions and new construction, so those properties were excluded by default. We subsequently acquired cadastral data from each of the 26 taxing authorities, and achieved an average success rate of 96 percent when linking the tabular data provided by BC Assessment Authority. Obviously, linking tabular assessment data to cadastral data is preferred; however, in jurisdictions without digital cadastral data, using a digital street network may be the only option, and link success rates may vary widely.
Developing variables from SPAD for use in land use regression models of neighbourhood pollution levels
When adequate measured data are not available, neighbourhood level exposure assessments may use outdoor pollution levels derived by models that require land use data as inputs. For example, land use regression (LUR) models have been used to predict traffic-related air pollution levels for neighbourhood areas depending on nearby roads, traffic volume, population density, and land uses [3-5]; these predicted levels were then used as indicators of exposure for epidemiological analyses. In their 1997 study, Briggs, Collins et al. used land cover data interpreted from aerial photographs, as well as building density (six classes) derived from local planning maps in a LUR model to predict spatial surfaces of nitrogen dioxide (NO2) levels in three European cities [4]. In 2003, Brauer et al. used 100 m raster grids of population density in a LUR model to predict fine particulate (PM2.5) levels at over 10,000 residential addresses in Sweden and the Netherlands [5]. The 100 m raster grids of population density were developed by national agencies from population registries that record the current residential address for most of the population. In research currently underway in the Pacific Northwest, Brauer, Henderson, et al. have included the area of commercial land, provided by local government as a digital map, as a predictor in a LUR model of traffic-related air pollution in Vancouver, British Columbia [6].
SPAD provide a unique opportunity to develop neighbourhood-level variables for use in LUR models. Whereas developing land use data from air photo interpretation or local planning maps may not be feasible for large study areas, there are no standard population registries in North America, and local digital land use maps may not be readily available, SPAD can be used to develop variables measuring building density, population density, residential unit density, and commercial land use (among others). In fact, SPAD may present an opportunity to significantly improve the spatial resolution of these kinds of density measures since SPAD are essentially individual-level data (i.e., available for every parcel), in contrast with widely used census data which are only available pre-aggregated for fixed census areas. Because SPAD are individual-level data, density measures can be based on any area(s) defined by the researcher, rather than restricted to existing census areas which may not adequately define the true areas of interest. Perhaps more importantly, current GIS can easily create spatial surfaces of density given several distance parameters (i.e., calculate density for every 10 m × 10 m cell in the study area, based on the number of residential units within 100 m of the cell centre). Figures 22 and 3 provide an illustration of the improved spatial resolution in measuring density using SPAD. Figure 2 shows residential density per hectare using SPAD, but reported for census boundaries. Note that the large census area near the top of the figure is shown with a residential density of >0 – 5. In Figure 3, residential density per hectare was calculated from SPAD using a GIS kernel function, and shows that the same large census area near the top in fact has a range of residential densities. When making neighbourhood level exposure assessments, variations on the scale of several hundred metres may be important. Note that the area shown in Figures 2 and 3 is approximately 4.8 hectares in size.
Figure 2 Residential unit density reported for census areas.
Figure 3 Residential units density surface calculated using a GIS kernel function.
SPAD also provide very detailed information about land use. In British Columbia, properties have a designated actual use code, organized in an hierarchical fashion. For example, a property's Level 1 (Property Code) designation may be 'major industry'; the Level 2 (Actual Use Code) designation may be 'primary metal industry', and its Level 3 (Manual Class Code) designation may be 'primary smelting and refining'. Similarly, a residential property may be designated as residential, single family residence, and 1 1/2 storey good condition. In British Columbia, there are 950 unique Level 3 designations. In Washington State, the property use codes used by counties generally correspond to the standard Land Use Coding Manual created by the Urban Renewal Administration, Housing and Home Finance Agency and Bureau of Public Roads (1965), which contains 4 levels of classification (see for more information).
Other sources of land use data are provided in highly generalized formats with pre-defined land use classes which may not be optimal for researchers. For example, DMTI © Spatial produces a commercially available land use dataset for GIS use, with the following land use classes: commercial, government and institutional, open area, parks and recreational, residential, resource and industry, and waterbody (see for more information). Figures 4 and 5 illustrate the different spatial distributions of commercial and industrial classes based on DMTI © Spatial data and SPAD (including properties coded as industrial or business, assuming that these classifications in SPAD are comparable to commercial and resource/industrial classifications in the DMTI © Spatial land use dataset), suggesting that significantly different results could be obtained for the same LUR model, depending on which data set is employed. This is not to suggest that the DMTI © Spatial data set is of poor quality, instead, it should be noted that this data set and others like it have been prepared for specific purposes and for use at general spatial scales that may not be adequate for neighbour-level exposure assessment. Also shown, in Figure 6, is a density map (square footage of business and industrial buildings per hectare) based on SPAD. If commercial and industrial activity is meant to act as a surrogate for air pollution, it is argued here that the density map produced with SPAD could provide a much more accurate measure of the level of commercial/industrial activity than do simple land use maps.
Figure 4 Commercial land use from DMTI Spatial Inc.
Figure 5 Commercial land use from SPAD.
Figure 6 Density of commercial square footage derived from SPAD using a GIS kernel function.
What is not readily apparent in Figures 5 and 6 is the high level of additional detail on land use inherent in the SPAD, which can be used to further refine land use classifications. In the above example, parcels from SPAD were selected based on the first level of description, the Property Code. Table 2 provides a summary of the Actual Use Code for all parcels selected and illustrates the wide variety of land uses included in the more general Property Code classification. This additional detail provides significant flexibility to researchers in terms of developing surrogate indicators of ambient air pollution based on land use, as they can include or exclude properties from the indicator based on more refined conceptual links to ambient air pollution. For example, researchers may choose to include parking lots since vehicle use may be concentrated there, but exclude vacant properties as no current activity occurs.
Table 2 Detailed information from SPAD for commercial land use
'Actual Use' Classification Parcels (n) 'Actual Use' Classification Parcels (n)
Storage and Warehousing – closed 100 Department Store 4
Stores and Services – Commercial 97 Fast Food Restaurant 4
Office Building (primary use) 70 Automobile sales – lot 3
Vacant 53 Industrial – Vacant 3
Parking Lot 27 Self-Serve Service Station 3
Commercial – strata lot 25 Shopping Center – neighbourhood 3
Automobile Paint Shop/Garage 23 Food Market 2
Stores and Offices 15 Metal Fabricating Industry 2
Automobile dealership 13 Shopping Center – regional 2
Shopping Center 10 Bakery and Biscuit Manufacturing 1
Shopping Center – community 10 Bowling Alley 1
Convenience Store/Service Station 9 Car Wash 1
Motel and Auto Court 9 Clothing Industry 1
Restaurant 9 Confectionary Manufacturing 1
Lumber Yard or Building Supplies 8 Furniture and Fixtures Industry 1
Service Station 8 Marine and Navigational Facilities 1
Hotel 5 Sash and Door Industry 1
Neighbourhood Pub 5 Soft Drink Bottling 1
Neighbourhood Store 5 Storage and Warehousing – cold 1
Bank 4 Stores and Living Quarters 1
Bus Company 4 Transportation Equipment 1
Using SPAD to estimate exposure to wood smoke
Exposure to wood smoke has been associated with negative health impacts, particularly for children and the elderly [7-9] and there is increasing interest in developing models to predict spatial estimates of wood smoke levels in order to provide spatially refined estimates that do not rely on individual surveys or monitoring campaigns. Spatial estimates of residential wood burning have been included in regional emissions inventories prepared for air quality management purposes and so a very brief overview of the methods used for emissions inventory purposes is provided here. In general, the contribution of residential wood burning to regional air quality is estimated by applying an emission factor to the proportion of households thought to have a wood burning appliance. Both the emission factor and the proportion of households are often derived from telephone surveys conducted in the region of interest. An example of this approach, employed for eight regions in British Columbia, is described in a recent report produced by the British Columbia Ministry of Water, Land and Air Protection [10]. Recent research by Tian et al. describes an approach in which a number of spatial variables are used to predict the proportion of wood-burning households, similar to the LUR models described above [11]. In their study, Tian et al. found that elevation, age (retired or ages 34-54), presence of farm income, and owner occupied residences predicted the number of households using wood as a primary heating source (as per the 1990 US Census) for census block groups. While it is not clear how this improves on the data already available from the US Census (at least for 1990 and 2000), this method could be used where US Census data do not exist, i.e., Canada.
Using SPAD, it is possible to locate wood burning appliances, and to map predominant heating source (i.e., electric baseboards, electric radiant, forced hot air, electric forced hot air, gas forced hot air, oil forced hot air, heat pump, hot water, etc.), as shown in Figures 7 and 8. This spatial information provides an opportunity to greatly increase the spatial resolution of wood smoke estimates over those derived from census variables and regional telephone surveys.
Figure 7 Map of fireplace locations.
Figure 8 Map of primary heat sources based on SPAD.
In the context of epidemiological studies, Larson et al. have used SPAD in conjunction with other spatial variables in order to predict fine particulate (PM2.5) levels associated with wood smoke for a large epidemiological study currently underway in the Georgia Basin Puget Sound Airshed [12]. Preliminary results suggest that building age, population density, and number of fireplaces are relatively strongly correlated with measured PM2.5 in the study area. A range of socio-economic variables are more weakly correlated. Of particular interest, this approach negates the need for additional information on wood-burning practices and emissions factors by relating spatial variables derived from SPAD and other sources (i.e., Census data) directly to actual measures of PM2.5 on cold clear evenings.
Infiltration modelling using SPAD
Population level epidemiological studies of air pollution commonly use an indirect approach to exposure assessment by assigning exposure levels based on outdoor ambient air pollution levels at the residential location, even though an increasing number of personal monitoring studies have shown that exposure measurements based on ambient monitoring are usually lower than those derived from personal monitoring [13]. Strong associations have been found between indoor and outdoor PM2.5 concentrations which indicate that a significant proportion of indoor fine particles are of outdoor origin [14], and other studies have identified specific building characteristics that influence infiltration rates, for example, type of basement, and year of construction [15].
SPAD contain a variety of information on building characteristics (Table 3) that could be incorporated into a regional infiltration model when used in conjunction with data on external conditions, such as climate factors, wind shielding, wind speed and direction. Such an infiltration model could provide a more complete picture of indoor pollutant levels, the spatial distributions of infiltration rates, and the impacts of indoor exposure to total exposure levels and health outcomes. The authors currently are developing an infiltration model for the Georgia Basin Puget Sound airshed, based on SPAD and an indoor/outdoor PM2.5 monitoring program.
Table 3 Variables common in SPAD that may be used in a regional infiltration model
Land Variables Property Size, Property Use, Topography, Building Permit Class.
Building Variables Improvement Type, Structure Use, Building Type, # of Stories, Year Built, Total Square Footage, Predominant Construction Type, # of Bedrooms, Predominant Heating Type, Air Conditioning, Fireplace, Structural Quality.
Conclusion
Considering that many exposure assessments and epidemiological analyses of the impacts of air pollution on health have been undertaken at regional scales, and that only recently have researchers begun to investigate neighbourhood-level variation in pollutant levels, it is not surprising that the authors could not find any published exposure assessments or epidemiological studies of air pollution that made use of SPAD. This paper illustrates that SPAD are a readily available data source that may provide an opportunity for conducting air pollution exposure assessment at neighbourhood level scales. SPAD also provide highly detailed information on building characteristics that may prove useful for modelling indoor levels of ambient-origin air pollution based on building infiltration characteristics, and there may be some utility in using SPAD to develop or refine indicators of socio-economic status. Some limitations to using SPAD are also apparent: SPAD are very large datasets which require GIS software and expertise to clean and extract the required subset of data in order to avoid slow processing times; and issues of comparability between GIS formats and data content may arise when a study area encompasses more than one jurisdiction. Limitations notwithstanding, the authors expect to see increasing uses of SPAD for exposure assessment and epidemiological analyses in the future, as researchers continue to investigate spatial variations in pollutant levels and other factors affecting exposure at increasingly finer scales.
Methods
SPAD were developed for the Canadian (southwest British Columbia) portion of the airshed by spatially referencing tabular property assessment data provided by the province to cadastral (parcel) data provided by municipal governments. For the American portion of the airshed (a portion of Washington State) the data were acquired in a readily useable format from each county. These data are used to illustrate the typical characteristics of SPAD, and to identify issues for using SPAD in terms of format, attributes and availability. Conceptual applications of SPAD to exposure assessment are demonstrated using SPAD from British Columbia and Washington State.
Authors' contributions
ES and PH acquired and processed the spatial property assessment data. ES prepared the manuscript with support from PH. PK reviewed and edited the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This research has been funded by the BC Centre for Disease Control, via a grant provided by Health Canada as part of the ongoing Border Air Quality Strategy agreement between Canada and the United States.
==== Refs
Williams FLR Ogston SA Identifying populations at risk from environmental contamination from point sources Occup Environ Med 2002 59 2 8 11836461 10.1136/oem.59.1.2
O'Neill MS Jerrett M Kawachi L Levy JL Cohen AJ Gouveia N Wilkinson P Fletcher T Cifuentes L Schwartz J Health, wealth, and air pollution: Advancing theory and methods Environmental Health Perspectives 2003 111 1861 1870 14644658
Clench-Aas J Bartonova A Bohler T Gronskei KE Sivertsen B Larssen S Air pollution exposure monitoring and estimating Part I. Integrated air quality monitoring system Journal of Environmental Monitoring 1999 1 313 319 11529128 10.1039/a902775k
Briggs DJ Collins S Elliott P Fischer P Kingham S Lebret E Pryl K Van Reeuwijk H Smallbone K Van der Veen A Mapping urban air pollution using GIS: a regression-based approach International Journal of Geographical Information Science 1997 11 699 718 10.1080/136588197242158
Brauer M Hoek G van Vliet P Meliefste K Fischer P Gehring U Heinrich J Cyrys J Bellander T Lewne M Brunekreef B Estimating long-term average particulate air pollution concentrations: Application of traffic indicators and geographic information systems Epidemiology 2003 14 228 239 12606891 10.1097/00001648-200303000-00019
Brauer M Henderson S Jerrett M Beckerman B Land Use Regression Modeling of Nitrogen Oxides and Fine Particulate Matter in the Greater Vancouver Regional District: November 8 - 11; Blaine, Washington. 2005
Salam MT Li YF Langholz B Gilliland FD Early-life environmental risk factors for asthma: Findings from the children's health study Environmental Health Perspectives 2004 112 760 765 15121522
Boman BC Forsberg AB Jarvholm BG Adverse health effects from ambient air pollution in relation to residential wood combustion in modern society Scandinavian Journal of Work Environment & Health 2003 29 251 260
Larson TV Koenig JQ Wood Smoke - Emissions and Noncancer Respiratory Effects Annual Review of Public Health 1994 15 133 156 8054078 10.1146/annurev.pu.15.050194.001025
British Columbia Ministry of Water LAP Residential Wood Burning Emissions in British Columbia 2004 Victoria, BC,
Tian YQ Radke JD Gong P Yu Q Model development for spatial variation of PM2.5 emissions from residential wood burning Atmospheric Environment 2004 38 833 843 10.1016/j.atmosenv.2003.10.040
Larson T Su J Baribeau A Buzzelli M Setton E Brauer M A Spatial Model of Urban Winter Woodsmoke Concentrations: ; Blaine, Washington. 2005
Toivola M Alm S Reponen T Kolari S Nevalainen A Personal exposures and microenvironmental concentrations of particles and bioaerosols Journal of Environmental Monitoring 2002 4 166 174 11871701 10.1039/b108682k
Rojas-Bracho L Suh HH Oyola P Koutrakis P Measurements of children's exposures to particles and nitrogen dioxide in Santiago, Chile Science of the Total Environment 2002 287 249 264 11993967 10.1016/S0048-9697(01)00987-1
Chang TJ Huang MY Wu YT Liao CM Quantitative prediction of traffic pollutant transmission into buildings Journal of Environmental Science and Health Part a-Toxic/Hazardous Substances & Environmental Engineering 2003 38 1025 1040 10.1081/ESE-120019861
|
16262893
|
PMC1277841
|
CC BY
|
2021-01-04 16:39:06
|
no
|
Int J Health Geogr. 2005 Oct 31; 4:26
|
utf-8
|
Int J Health Geogr
| 2,005 |
10.1186/1476-072X-4-26
|
oa_comm
|
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-261626289310.1186/1476-072X-4-26MethodologyOpportunities for using spatial property assessment data in air pollution exposure assessments Setton Eleanor M [email protected] Perry W [email protected] C Peter [email protected] Spatial Sciences Research Laboratory, Geography Department, University of Victoria, PO BOX 3050 STN CSC, Victoria, B.C., V8W 3P5, Canada2005 31 10 2005 4 26 26 14 9 2005 31 10 2005 Copyright © 2005 Setton et al; licensee BioMed Central Ltd.2005Setton 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 epidemiological studies examining the relationships between adverse health outcomes and exposure to air pollutants use ambient air pollution measurements as a proxy for personal exposure levels. When pollution levels vary at neighbourhood levels, using ambient pollution data from sparsely located fixed monitors may inadequately capture the spatial variation in ambient pollution. A major constraint to moving toward exposure assessments and epidemiological studies of air pollution at a neighbourhood level is the lack of readily available data at appropriate spatial resolutions. Spatial property assessment data are widely available in North America and may provide an opportunity for developing neighbourhood level air pollution exposure assessments.
Results
This paper provides a detailed description of spatial property assessment data available in the Pacific Northwest of Canada and the United States, and provides examples of potential applications of spatial property assessment data for improving air pollution exposure assessment at the neighbourhood scale, including: (1) creating variables for use in land use regression modelling of neighbourhood levels of ambient air pollution; (2) enhancing wood smoke exposure estimates by mapping fireplace locations; and (3) using data available on individual building characteristics to produce a regional air pollution infiltration model.
Conclusion
Spatial property assessment data are an extremely detailed data source at a fine spatial resolution, and therefore a source of information that could improve the quality and spatial resolution of current air pollution exposure assessments.
==== Body
Background
Many epidemiological studies examining the relationships between adverse health outcomes and exposure to air pollutants use ambient air pollution measurements as a proxy for personal exposure levels [1]. Because the number of fixed outdoor monitoring sites within a city usually is limited, ambient pollution measurements often are extrapolated to areas between monitors, thus disregarding any neighbourhood-scale spatial variation in pollution levels. Recent research suggests that some neighbourhoods within a city can be disproportionately exposed to air pollution and that these differences may influence health outcomes [2]. A major constraint to moving toward exposure assessments and epidemiological studies of air pollution at a neighbourhood level is the lack of readily available data at appropriate spatial resolutions.
Spatial property assessment data (SPAD) were identified as a potential data source for exposure research through an ongoing study, funded by Health Canada via the British Columbia Centre for Disease Control, examining the effects of air pollution on birth outcomes and subsequent development of health outcomes associated with exposure to air pollution for a birth cohort of 90,000. The study area encompasses the Georgia Basin Puget Sound airshed, located in the Pacific Northwest of the United States and Canada and encompassing approximately 10 million hectares of land and marine environments. SPAD (considered here to be made up of both tabular assessment data on building characteristics and spatial data that show the location of each property) for every assessed parcel of land are generally available for the airshed as spatially referenced digital databases, suitable for use with common geographic information systems (GIS). These data may increase the resolution and accuracy of variables used in exposure assessment models and epidemiological analyses, but the authors have found few published studies using SPAD for developing exposure assessments or in epidemiological analyses of air pollution impacts on health.
The purpose of this paper is to describe SPAD and to illustrate their potential utility for neighbourhood level exposure estimates and epidemiological research. Three possible uses of SPAD are examined, including: (1) creating variables for use in land use regression modelling of neighbourhood levels of ambient air pollution; (2) enhancing wood smoke exposure estimates by mapping fireplace locations; and (3) using data available on individual building characteristics to produce a regional air pollution infiltration model.
Results and discussion
Typical SPAD characteristics
Tabular assessment data generally incorporate two kinds of information: 1) property addresses or property identifiers that can be used for spatial referencing in conjunction with digital street networks or digital cadastral maps; and 2) descriptive variables including building characteristics, building and land values, and land use information on which annual property tax assessments are based. Table 1 provides examples of common variables in tabular assessment data that may be important to air pollution exposure assessments and epidemiological analyses.
Table 1 Common variables in tabular assessment data
Location School District #, Area #, Township Range, Jurisdiction #, Neighbourhood #, Street Address, Street Direction, Street Type, ZIP Code, City, Property Identifier
Land Appraisal Date, Property Size, Property Use Code, Land Use Code, Electricity, Water, Sewer, Street Surface Type, # of Dwelling Units, # of Outbuildings, # of Improvements, Building Permit.
Sales Sale Date, Sale Price, Sales Excise Number, Deed Type, Qualification Code, Multiple Sales, Land Value, Improvements Value.
Improvements Improvement Type, Structure Use, Building Type, # of Stories, Year Built, Total Square Footage, # of Bedrooms, Predominant Heating Type, Fireplace, Structural Quality.
SPAD are created when tabular property assessment data are spatially referenced, either by linking property addresses to a digital street network, or by linking property identifiers to a digital cadastral map. Spatial referencing allows for complex queries of the tabular assessment, the results of which can be mapped. In effect, the spatial resolution of SPAD is the individual parcel size, generally much finer than other spatially referenced data commonly used in exposure assessment and epidemiological analyses, such as Census data.
Data format and availability
SPAD are developed and maintained by numerous jurisdictions throughout North America, therefore data format and content may differ significantly among jurisdictions. The following discussion highlights some of the differences and associated issues in British Columbia and Washington State, as shown in Figure 1. In British Columbia, tabular assessment data are collected by the BC Assessment Authority and maintained as a tabular database, while each taxing municipality or regional district develops its own cadastral data that can be linked with assessment data to create SPAD. By agreement, BC Assessment uses a unique property identifier for each assessment record, and the same property identifiers are used by taxing jurisdictions when developing cadastral data, thereby enabling a linkage between the tabular assessment data from BC Assessment and the jurisdiction's cadastral data using GIS. In Washington State, each county is responsible for developing both the tabular assessment data and the cadastral data. In this regard, there is some advantage in the British Columbia system, in that one single authority collects and maintains the tabular assessment data and so these are standard throughout the entire province. Unfortunately, when multiple jurisdictions are responsible for developing tabular assessment data and/or cadastral data, there can be significant differences in format among jurisdictions. For example, in British Columbia, cadastral data are often available in ESRI© GIS formats, but in some cases are only available in AutoCad© format. The latter is primarily an engineering and drafting application and the format may not always translate easily into GIS formats. In Washington State, neither the tabular assessment data nor the cadastral data may be standardized among counties, as each develop and maintain their own information systems. In this case, it is possible that some tabular assessment data (i.e., presence of air conditioner) are collected for one county, but not for the adjacent county, or that different GIS applications are used by different counties.
Figure 1 The development of SPAD differs between British Columbia and Washington State.
Access to SPAD (or its constituent tabular and cadastral data) is markedly different in British Columbia in comparison to Washington State. In British Columbia, researchers must negotiate data sharing or purchasing agreements with each jurisdiction in order to access SPAD, and may also have to purchase additional tabular assessment data directly from the BC Assessment Authority in order to develop SPAD specific to the research question. In Washington State, SPAD are available for download through each county's internet site, or may be ordered directly from each county at no cost or for a small fee (i.e., for CD writing and postage). In many cases, due to large file sizes, the tabular assessment data and the spatial cadastral data are provided separately, and must be linked by the researcher using GIS to create the final SPAD.
Linking tabular assessment data using property addresses or identifiers to produce SPAD is not always trouble free. In cases where tabular assessment data and the spatial cadastral data are provided by the same jurisdiction, linking the two datasets often is easily accomplished. In Washington State, for example, where each county develops and maintains its own SPAD, we were able to download the tabular assessment data and the spatial cadastral data, and link each record with a 98 percent success rate. For the British Columbia portion of the airshed, we initially purchased tabular assessment data from the BC Assessment Authority and spatially referenced them using the included property addresses and a commercially available digital street network with ESRI© ArcGIS 8.1. Approximately 1.1 million records were received from BC Assessment for the entire Georgia Basin airshed, which is comprised of 26 separate taxing jurisdictions. Linking between the tabular assessment data and the street network was successfully completed for approximately 83 percent of the records, with the number of links in urban areas better than in rural regions (89 percent versus 67 percent respectively). The lower success rate in rural regions is generally due to incomplete or non-standard street addresses (i.e., post office boxes or rural post offices rather than street addresses) in the tabular assessment data. Also, the road network (circa 2003) did not contain information on the most recent subdivisions and new construction, so those properties were excluded by default. We subsequently acquired cadastral data from each of the 26 taxing authorities, and achieved an average success rate of 96 percent when linking the tabular data provided by BC Assessment Authority. Obviously, linking tabular assessment data to cadastral data is preferred; however, in jurisdictions without digital cadastral data, using a digital street network may be the only option, and link success rates may vary widely.
Developing variables from SPAD for use in land use regression models of neighbourhood pollution levels
When adequate measured data are not available, neighbourhood level exposure assessments may use outdoor pollution levels derived by models that require land use data as inputs. For example, land use regression (LUR) models have been used to predict traffic-related air pollution levels for neighbourhood areas depending on nearby roads, traffic volume, population density, and land uses [3-5]; these predicted levels were then used as indicators of exposure for epidemiological analyses. In their 1997 study, Briggs, Collins et al. used land cover data interpreted from aerial photographs, as well as building density (six classes) derived from local planning maps in a LUR model to predict spatial surfaces of nitrogen dioxide (NO2) levels in three European cities [4]. In 2003, Brauer et al. used 100 m raster grids of population density in a LUR model to predict fine particulate (PM2.5) levels at over 10,000 residential addresses in Sweden and the Netherlands [5]. The 100 m raster grids of population density were developed by national agencies from population registries that record the current residential address for most of the population. In research currently underway in the Pacific Northwest, Brauer, Henderson, et al. have included the area of commercial land, provided by local government as a digital map, as a predictor in a LUR model of traffic-related air pollution in Vancouver, British Columbia [6].
SPAD provide a unique opportunity to develop neighbourhood-level variables for use in LUR models. Whereas developing land use data from air photo interpretation or local planning maps may not be feasible for large study areas, there are no standard population registries in North America, and local digital land use maps may not be readily available, SPAD can be used to develop variables measuring building density, population density, residential unit density, and commercial land use (among others). In fact, SPAD may present an opportunity to significantly improve the spatial resolution of these kinds of density measures since SPAD are essentially individual-level data (i.e., available for every parcel), in contrast with widely used census data which are only available pre-aggregated for fixed census areas. Because SPAD are individual-level data, density measures can be based on any area(s) defined by the researcher, rather than restricted to existing census areas which may not adequately define the true areas of interest. Perhaps more importantly, current GIS can easily create spatial surfaces of density given several distance parameters (i.e., calculate density for every 10 m × 10 m cell in the study area, based on the number of residential units within 100 m of the cell centre). Figures 22 and 3 provide an illustration of the improved spatial resolution in measuring density using SPAD. Figure 2 shows residential density per hectare using SPAD, but reported for census boundaries. Note that the large census area near the top of the figure is shown with a residential density of >0 – 5. In Figure 3, residential density per hectare was calculated from SPAD using a GIS kernel function, and shows that the same large census area near the top in fact has a range of residential densities. When making neighbourhood level exposure assessments, variations on the scale of several hundred metres may be important. Note that the area shown in Figures 2 and 3 is approximately 4.8 hectares in size.
Figure 2 Residential unit density reported for census areas.
Figure 3 Residential units density surface calculated using a GIS kernel function.
SPAD also provide very detailed information about land use. In British Columbia, properties have a designated actual use code, organized in an hierarchical fashion. For example, a property's Level 1 (Property Code) designation may be 'major industry'; the Level 2 (Actual Use Code) designation may be 'primary metal industry', and its Level 3 (Manual Class Code) designation may be 'primary smelting and refining'. Similarly, a residential property may be designated as residential, single family residence, and 1 1/2 storey good condition. In British Columbia, there are 950 unique Level 3 designations. In Washington State, the property use codes used by counties generally correspond to the standard Land Use Coding Manual created by the Urban Renewal Administration, Housing and Home Finance Agency and Bureau of Public Roads (1965), which contains 4 levels of classification (see for more information).
Other sources of land use data are provided in highly generalized formats with pre-defined land use classes which may not be optimal for researchers. For example, DMTI © Spatial produces a commercially available land use dataset for GIS use, with the following land use classes: commercial, government and institutional, open area, parks and recreational, residential, resource and industry, and waterbody (see for more information). Figures 4 and 5 illustrate the different spatial distributions of commercial and industrial classes based on DMTI © Spatial data and SPAD (including properties coded as industrial or business, assuming that these classifications in SPAD are comparable to commercial and resource/industrial classifications in the DMTI © Spatial land use dataset), suggesting that significantly different results could be obtained for the same LUR model, depending on which data set is employed. This is not to suggest that the DMTI © Spatial data set is of poor quality, instead, it should be noted that this data set and others like it have been prepared for specific purposes and for use at general spatial scales that may not be adequate for neighbour-level exposure assessment. Also shown, in Figure 6, is a density map (square footage of business and industrial buildings per hectare) based on SPAD. If commercial and industrial activity is meant to act as a surrogate for air pollution, it is argued here that the density map produced with SPAD could provide a much more accurate measure of the level of commercial/industrial activity than do simple land use maps.
Figure 4 Commercial land use from DMTI Spatial Inc.
Figure 5 Commercial land use from SPAD.
Figure 6 Density of commercial square footage derived from SPAD using a GIS kernel function.
What is not readily apparent in Figures 5 and 6 is the high level of additional detail on land use inherent in the SPAD, which can be used to further refine land use classifications. In the above example, parcels from SPAD were selected based on the first level of description, the Property Code. Table 2 provides a summary of the Actual Use Code for all parcels selected and illustrates the wide variety of land uses included in the more general Property Code classification. This additional detail provides significant flexibility to researchers in terms of developing surrogate indicators of ambient air pollution based on land use, as they can include or exclude properties from the indicator based on more refined conceptual links to ambient air pollution. For example, researchers may choose to include parking lots since vehicle use may be concentrated there, but exclude vacant properties as no current activity occurs.
Table 2 Detailed information from SPAD for commercial land use
'Actual Use' Classification Parcels (n) 'Actual Use' Classification Parcels (n)
Storage and Warehousing – closed 100 Department Store 4
Stores and Services – Commercial 97 Fast Food Restaurant 4
Office Building (primary use) 70 Automobile sales – lot 3
Vacant 53 Industrial – Vacant 3
Parking Lot 27 Self-Serve Service Station 3
Commercial – strata lot 25 Shopping Center – neighbourhood 3
Automobile Paint Shop/Garage 23 Food Market 2
Stores and Offices 15 Metal Fabricating Industry 2
Automobile dealership 13 Shopping Center – regional 2
Shopping Center 10 Bakery and Biscuit Manufacturing 1
Shopping Center – community 10 Bowling Alley 1
Convenience Store/Service Station 9 Car Wash 1
Motel and Auto Court 9 Clothing Industry 1
Restaurant 9 Confectionary Manufacturing 1
Lumber Yard or Building Supplies 8 Furniture and Fixtures Industry 1
Service Station 8 Marine and Navigational Facilities 1
Hotel 5 Sash and Door Industry 1
Neighbourhood Pub 5 Soft Drink Bottling 1
Neighbourhood Store 5 Storage and Warehousing – cold 1
Bank 4 Stores and Living Quarters 1
Bus Company 4 Transportation Equipment 1
Using SPAD to estimate exposure to wood smoke
Exposure to wood smoke has been associated with negative health impacts, particularly for children and the elderly [7-9] and there is increasing interest in developing models to predict spatial estimates of wood smoke levels in order to provide spatially refined estimates that do not rely on individual surveys or monitoring campaigns. Spatial estimates of residential wood burning have been included in regional emissions inventories prepared for air quality management purposes and so a very brief overview of the methods used for emissions inventory purposes is provided here. In general, the contribution of residential wood burning to regional air quality is estimated by applying an emission factor to the proportion of households thought to have a wood burning appliance. Both the emission factor and the proportion of households are often derived from telephone surveys conducted in the region of interest. An example of this approach, employed for eight regions in British Columbia, is described in a recent report produced by the British Columbia Ministry of Water, Land and Air Protection [10]. Recent research by Tian et al. describes an approach in which a number of spatial variables are used to predict the proportion of wood-burning households, similar to the LUR models described above [11]. In their study, Tian et al. found that elevation, age (retired or ages 34-54), presence of farm income, and owner occupied residences predicted the number of households using wood as a primary heating source (as per the 1990 US Census) for census block groups. While it is not clear how this improves on the data already available from the US Census (at least for 1990 and 2000), this method could be used where US Census data do not exist, i.e., Canada.
Using SPAD, it is possible to locate wood burning appliances, and to map predominant heating source (i.e., electric baseboards, electric radiant, forced hot air, electric forced hot air, gas forced hot air, oil forced hot air, heat pump, hot water, etc.), as shown in Figures 7 and 8. This spatial information provides an opportunity to greatly increase the spatial resolution of wood smoke estimates over those derived from census variables and regional telephone surveys.
Figure 7 Map of fireplace locations.
Figure 8 Map of primary heat sources based on SPAD.
In the context of epidemiological studies, Larson et al. have used SPAD in conjunction with other spatial variables in order to predict fine particulate (PM2.5) levels associated with wood smoke for a large epidemiological study currently underway in the Georgia Basin Puget Sound Airshed [12]. Preliminary results suggest that building age, population density, and number of fireplaces are relatively strongly correlated with measured PM2.5 in the study area. A range of socio-economic variables are more weakly correlated. Of particular interest, this approach negates the need for additional information on wood-burning practices and emissions factors by relating spatial variables derived from SPAD and other sources (i.e., Census data) directly to actual measures of PM2.5 on cold clear evenings.
Infiltration modelling using SPAD
Population level epidemiological studies of air pollution commonly use an indirect approach to exposure assessment by assigning exposure levels based on outdoor ambient air pollution levels at the residential location, even though an increasing number of personal monitoring studies have shown that exposure measurements based on ambient monitoring are usually lower than those derived from personal monitoring [13]. Strong associations have been found between indoor and outdoor PM2.5 concentrations which indicate that a significant proportion of indoor fine particles are of outdoor origin [14], and other studies have identified specific building characteristics that influence infiltration rates, for example, type of basement, and year of construction [15].
SPAD contain a variety of information on building characteristics (Table 3) that could be incorporated into a regional infiltration model when used in conjunction with data on external conditions, such as climate factors, wind shielding, wind speed and direction. Such an infiltration model could provide a more complete picture of indoor pollutant levels, the spatial distributions of infiltration rates, and the impacts of indoor exposure to total exposure levels and health outcomes. The authors currently are developing an infiltration model for the Georgia Basin Puget Sound airshed, based on SPAD and an indoor/outdoor PM2.5 monitoring program.
Table 3 Variables common in SPAD that may be used in a regional infiltration model
Land Variables Property Size, Property Use, Topography, Building Permit Class.
Building Variables Improvement Type, Structure Use, Building Type, # of Stories, Year Built, Total Square Footage, Predominant Construction Type, # of Bedrooms, Predominant Heating Type, Air Conditioning, Fireplace, Structural Quality.
Conclusion
Considering that many exposure assessments and epidemiological analyses of the impacts of air pollution on health have been undertaken at regional scales, and that only recently have researchers begun to investigate neighbourhood-level variation in pollutant levels, it is not surprising that the authors could not find any published exposure assessments or epidemiological studies of air pollution that made use of SPAD. This paper illustrates that SPAD are a readily available data source that may provide an opportunity for conducting air pollution exposure assessment at neighbourhood level scales. SPAD also provide highly detailed information on building characteristics that may prove useful for modelling indoor levels of ambient-origin air pollution based on building infiltration characteristics, and there may be some utility in using SPAD to develop or refine indicators of socio-economic status. Some limitations to using SPAD are also apparent: SPAD are very large datasets which require GIS software and expertise to clean and extract the required subset of data in order to avoid slow processing times; and issues of comparability between GIS formats and data content may arise when a study area encompasses more than one jurisdiction. Limitations notwithstanding, the authors expect to see increasing uses of SPAD for exposure assessment and epidemiological analyses in the future, as researchers continue to investigate spatial variations in pollutant levels and other factors affecting exposure at increasingly finer scales.
Methods
SPAD were developed for the Canadian (southwest British Columbia) portion of the airshed by spatially referencing tabular property assessment data provided by the province to cadastral (parcel) data provided by municipal governments. For the American portion of the airshed (a portion of Washington State) the data were acquired in a readily useable format from each county. These data are used to illustrate the typical characteristics of SPAD, and to identify issues for using SPAD in terms of format, attributes and availability. Conceptual applications of SPAD to exposure assessment are demonstrated using SPAD from British Columbia and Washington State.
Authors' contributions
ES and PH acquired and processed the spatial property assessment data. ES prepared the manuscript with support from PH. PK reviewed and edited the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This research has been funded by the BC Centre for Disease Control, via a grant provided by Health Canada as part of the ongoing Border Air Quality Strategy agreement between Canada and the United States.
==== Refs
Williams FLR Ogston SA Identifying populations at risk from environmental contamination from point sources Occup Environ Med 2002 59 2 8 11836461 10.1136/oem.59.1.2
O'Neill MS Jerrett M Kawachi L Levy JL Cohen AJ Gouveia N Wilkinson P Fletcher T Cifuentes L Schwartz J Health, wealth, and air pollution: Advancing theory and methods Environmental Health Perspectives 2003 111 1861 1870 14644658
Clench-Aas J Bartonova A Bohler T Gronskei KE Sivertsen B Larssen S Air pollution exposure monitoring and estimating Part I. Integrated air quality monitoring system Journal of Environmental Monitoring 1999 1 313 319 11529128 10.1039/a902775k
Briggs DJ Collins S Elliott P Fischer P Kingham S Lebret E Pryl K Van Reeuwijk H Smallbone K Van der Veen A Mapping urban air pollution using GIS: a regression-based approach International Journal of Geographical Information Science 1997 11 699 718 10.1080/136588197242158
Brauer M Hoek G van Vliet P Meliefste K Fischer P Gehring U Heinrich J Cyrys J Bellander T Lewne M Brunekreef B Estimating long-term average particulate air pollution concentrations: Application of traffic indicators and geographic information systems Epidemiology 2003 14 228 239 12606891 10.1097/00001648-200303000-00019
Brauer M Henderson S Jerrett M Beckerman B Land Use Regression Modeling of Nitrogen Oxides and Fine Particulate Matter in the Greater Vancouver Regional District: November 8 - 11; Blaine, Washington. 2005
Salam MT Li YF Langholz B Gilliland FD Early-life environmental risk factors for asthma: Findings from the children's health study Environmental Health Perspectives 2004 112 760 765 15121522
Boman BC Forsberg AB Jarvholm BG Adverse health effects from ambient air pollution in relation to residential wood combustion in modern society Scandinavian Journal of Work Environment & Health 2003 29 251 260
Larson TV Koenig JQ Wood Smoke - Emissions and Noncancer Respiratory Effects Annual Review of Public Health 1994 15 133 156 8054078 10.1146/annurev.pu.15.050194.001025
British Columbia Ministry of Water LAP Residential Wood Burning Emissions in British Columbia 2004 Victoria, BC,
Tian YQ Radke JD Gong P Yu Q Model development for spatial variation of PM2.5 emissions from residential wood burning Atmospheric Environment 2004 38 833 843 10.1016/j.atmosenv.2003.10.040
Larson T Su J Baribeau A Buzzelli M Setton E Brauer M A Spatial Model of Urban Winter Woodsmoke Concentrations: ; Blaine, Washington. 2005
Toivola M Alm S Reponen T Kolari S Nevalainen A Personal exposures and microenvironmental concentrations of particles and bioaerosols Journal of Environmental Monitoring 2002 4 166 174 11871701 10.1039/b108682k
Rojas-Bracho L Suh HH Oyola P Koutrakis P Measurements of children's exposures to particles and nitrogen dioxide in Santiago, Chile Science of the Total Environment 2002 287 249 264 11993967 10.1016/S0048-9697(01)00987-1
Chang TJ Huang MY Wu YT Liao CM Quantitative prediction of traffic pollutant transmission into buildings Journal of Environmental Science and Health Part a-Toxic/Hazardous Substances & Environmental Engineering 2003 38 1025 1040 10.1081/ESE-120019861
|
16242026
|
PMC1277842
|
CC BY
|
2021-01-04 16:38:37
|
no
|
Int Semin Surg Oncol. 2005 Oct 20; 2:22
|
latin-1
|
Int Semin Surg Oncol
| 2,005 |
10.1186/1477-7800-2-22
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-311620737310.1186/1743-0003-2-31ResearchA preliminary study of clinical assessment of left unilateral spatial neglect using a head mounted display system (HMD) in rehabilitation engineering technology Tanaka Toshiaki [email protected] Shunichi [email protected] Hiroyuki [email protected] Shuichi [email protected] Tohru [email protected] Department of Physical Therapy, School of Health Sciences Sapporo Medical University, Sapporo, Hokkaido, Japan2 Sapporo Shuyukai Hospital, Sapporo, Hokkaido, Japan3 AdIn Research, Inc., Sapporo, Hokkaido, Japan4 Research Center for Advanced Science and Technology, The University of Tokyo. Tokyo, Japan2005 5 10 2005 2 31 31 5 4 2005 5 10 2005 Copyright © 2005 Tanaka et al; licensee BioMed Central Ltd.2005Tanaka 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.
Purpose
Unilateral spatial neglect (USN) is a common syndrome in which a patient fails to report or respond to stimulation from the side of space opposite a brain lesion, where these symptoms are not due to primary sensory or motor deficits. The purpose of this study was to analyze an evaluation process system of USN in various visual fields using HMD in order to understand more accurately any faults of USN operating in the object-centred co-ordinates.
Method
Eight stroke patients participated in this study and they had Left USN in clinical test, and right hemisphere damage was checked by CT scan. Assessments of USN were performed the BIT common clinical test (the line and the stars cancellation tests) and special tests the zoom-in condition (ZI) condition and the zoom-out condition (ZO) condition. The subjects were first evaluated by the common clinical test without HMD and then two spatial tests with HMD. Moreover, we used a video-recording for all tests to analyze each subject's movements.
Results
For the line cancellation test under the common condition, the mean percentage of the correct answers at the left side in the test paper was 94.4%. In the ZI condition, the left side was 61.8.% and the right side was 92.4.%. In the ZO condition, the left side was 79.9% and the right side was 91.7.%. There were significant differences among the three conditions. The results of the stars cancellation test also showed the same tendency as the line bisection test.
Conclusion
The results showed that the assessment of USN using a technique of HMD system may indicate the disability of USN more than the common clinical tests. Moreover, it might be hypothesized that the three dimensional for USN test may be more related to various damage and occurrence of USN than only the two dimensional test.
Unilateral spatial neglecthead mounted display systemvirtual realityclinical assessment
==== Body
Introduction
Unilateral spatial neglect (USN) is a common syndrome in which a patient fails to report or respond to stimulation from the side of space opposite a brain lesion, where these symptoms are not due to primary sensory or motor deficits [1]. Patients with severe neglect often collide with objects, ignore food on one side of the plate, and in general tend to rely on just one side of the body [2]. Patients with USN of the left hemispace require longer hospital stays and have more difficulty resuming activities of daily living [3]. Katz et al. [4] reported that impairment and disability levels of RBD patient with and without USN were clearly different. Neglect is associated with lower performance on measures of impairment, as well as on measures of disability in ADL. Recently, several studies have singled out USN as one of the major disruptive factors impeding functional recovery and rehabilitation success [5].
Progress in the treatment of USN has been hampered by an inadequate understanding and examination of the underlying involved mechanisms [6]. One problem has been the underrepresentation of left hemisphere-damaged patients in many studies, despite several reports which indicated no significant differences in the frequency of neglect [7]. The situation is further complicated by the existence of competing theoretical models [8,9], different lesion locations, and considerable variation in the reported incidence among right-brain-damaged patients [10]. Little attention has been paid to systematic behavioral assessment of patients with USN. As a result, there has been a largely unquestioned assumption that the diverse assessment procedures all provide an accurate measurement of the same underlying deficit.
From a rehabilitation perspective, the traditional assessment of USN centers on a variety of simple perceptual motor tasks. Investigations have used line crossing [11], cancellation task [12] and more recently, an indented reading test [13]. However, there is no single standardized battery of tests currently available for the assessment of USN. Also, performance rating of these tasks cannot be related to the specific difficulties encountered in everyday life. Rehabilitation prospects of brain-damaged patients are rendered more specific and realistic by a consideration of their behavioral strengths and deficits within a functional framework [14]. The development of an objective behavioral test of everyday skills relevant to neglect would provide therapists and clinicians with a more precise description of a patient's capabilities, which would encourage a more robust grounding for rehabilitation.
An analysis of USN can be explained with a space coordinate system theory. The boundaries of the neglected space are not constant in as much as the neglect patients'performance is influenced by the relevant system of spatial coordinates; egocentric or allocentric co-ordinates. Egocentric co-ordinates specify locations relative to the viewer [15], whereas allocentric co-ordinates code their position independent of viewpoint [16]. Clinical evidence from visuospatial neglect suggests that some patients neglect one side of each individual object in a scene, rather than just one side of the scene as a whole. For example, in copying a lateral array of objects, right-hemisphere patients may reproduce only the right side of the objects, but produce these for each of the objects in the scene including those on the extreme left [17]. This is suggestive evidence for neglect operating in the object-centred allocentric co-ordinates. Driver and Halligan suggested that USN can be object-centered in the sense of operating relative to the principal axes [18]. However, copying evidence is not conclusive.
Several sensory manipulations may be temporarily effective for improving unilateral spatial neglect. Karnath indicated the effectiveness of neck vibration [19]. Pizzamiglio et al. also adopted an effective means of optokinetic stimulation [20]. Rossetti et al. investigated the effect of prism adaptation on neglect symptoms, including the pathological shift of the subjective midline to the right [21]. They reported that all patients exposed to the optical shift of the visual field to the right were improved in their manual body-midline demonstration and on their classical neuropsychological tests. However, these manipulations have not yet succeeded in bringing about a consistent improvement of neglect.
Virtual reality (VR) refers to computer-generated, usually visual, representation of real-world objects in which a user can navigate or manipulate the environment [22]. The most well-known approach is " immersive, " where the real world is opaque to the user and he or she is provided the sensation of interacting directly with the computer-generated objects. In other approaches, VR shares certain attributes similar to a three-dimensional computer-aided design (CAD). In immersive VR, a head-mounted display (HMD) is worn and its position in space is tracked. As the user moves his or her head, aspects of the computer-generated object appropriate to the HMD position are displayed. Virtual reality (VR) has many advantages over other ADL rehabilitation techniques and offers the potential to develop a human performance testing and training environment [23] and also a VR system for training individuals with unilateral spatial neglect to cross streets in a safe and vigilant manner. [24]. VR can give human versatile sensory information artificially and easily for the visual, vestibular, and the somatic sensations. Recently, VR has been investigated in a few studies using devices for compensation of visual sensory. For example, there is one approach where HMD gives a patient with Parkinson' disease an emphasized visual input in order to improve a frozen gait of the patient [25]. HMD has a function which can focus on a certain object or to limit the surrounding environmental conditions, and to offer versatile visual information. Therefore, HMD can produce the object-centred co-ordinates for a USN patient.
The purpose of this study was to analyze an evaluation process system of USN in various visual fields using HMD in order to understand more accurately any faults of USN operating in the object-centred co-ordinates. Moreover, we constructed a new device that uses rehabilitation engineering technology for assessing and training of USN.
The following hypothesis was verified that a special evaluation process system with HMD for USN can be more accurate and detailed than the common clinical test for USN. It may be assumed that the significant difference between the common evaluation of USN and the special test in the object-centred co-ordinates was produced by the result of using HMD.
However, there were a few limitations of this study. There was the possibility of low validity of the results because of the small number of subjects. There was also a limitation about discussion of concerning the mechanism of USN because of the damaged part of the brain and the versatility of coping mechanisms.
Methods
1. Subjects
Eight patients who had suffered a stroke (mean age 67.1 years old) participated in this study after gaining their informed consent. The patients were tested for the presence of any neglect for activities of daily living (ADL) by two therapists. Two medical doctors checked the right hemisphere damage of all subjects by CT (computed tomography) or MRI (magnetic resonance imaging). Individuals with weak visual acuity, dementia, hemianopsia, apraxia or those being left-handed were excluded. The subjects could sit on an ordinary chair by themselves. The period from the appearance of disease to study assessment was 4–27 weeks (Table 1).
Table 1 Characteristics of patients
Patient No. Age (years) Dignosis Lesion* Time of rehabilitation onset (weeks) FIM-M
1 75 I FTP 6 30
2 65 I BgFPT 1 38
3 64 H Th 1 61
4 63 H Bg 1 35
5 56 I PT 1 85
6 70 I Bg 1 33
7 79 I FPT 1 86
8 68 I BgFPT 2 72
Abbreviations: I: infarction, H; hemorrhage, F: frontal lobe, P: parietal lobe; T; temporal lobe, Bg; basal ganglia, Th; thalamus. FIM; Functional Independence measure Motor.
*all lesions were right sided.
2. Functional assessment
The Functional Independence Measure (FIM) was executed as an ADL evaluation [26,27]. The FIM motor sub scores (FIM-M) was used for measure of disability as the best predictors of rehabilitation length of stay for stroke. Moreover, two therapists evaluated the patients who exhibited specific neglect behaviors in ADL using a special checklist (Table 2). The checklist used a modified version of Halligan's checklist [28]. The therapists were requested to score the checklist in terms of those behaviors they considered to be related to as visual neglect, as opposed to poor performance that might be expected to follow concomitant disorders such as problems of motor coordination or initiation.
Table 2 Checklist of Everyday Neglect Behaviors
1. Dose the patient show difficulties when: talking or communicating with others
2. Dose the patient neglect the left/right side of personal space?
3. Dose the patient show difficulties in eating?
4. Dose the patient show difficulties in grooming (self-care skills, washing, bathing, etc)
5. Does the patient show difficulties in dressing?
6. Does the patient show difficulties in body movement transferring (from a bed to W/C,etc)?
7. Does the patient show difficulties in locomotion 1 (the patient collides against objects and wall on the affected side. The patient can not negotiate a W/C between doors, kerbs, etc.)?
8. Does the patient show difficulties in locomotion 2 (the patient turns toward the direction of the affected side.)
9. Does the patient show difficulties during PT exercise?
10. Does the patient show difficulties during OT excercise?
3-1. Evaluation for USN
3-1-1. Common clinical test (Figure 1)
Figure 1 Analysis method for line and star cancellation test.
To asses neglect, the widely used line and star cancellation tests as included in the Behavioral Inattention Test (BIT) were given to the subjects [29]. We used the BIT Japanese version which was modified by Ishiai et al [30].
For the line cancellation test (score range from 0 to 36 points), the subjects were presented with a single sheet of paper on which 6 lines in varying orientations were drawn, 18 on each side. They were instructed to make a mark through all of the lines. Left- sided neglect was indicated by a failure to mark more lines on the left side than on the right. Degree of neglect was assessed by the proportion of lines omitted relative to the total number of lines. The line cancellation test sheet was divided into right and left portions and a right and then a left correct answer rates were analyzed. 34 points were set as a cutoff value.
For the star cancellation test (score range from 0 to 54 points), the A4 stimulus sheet contained 56 targets (small stars) pseudo-randomly interspersed with distracter items. The targets actually fell into six columns, with two additional targets which were located centrally. The experimenter clearly indicated the full extent of the sheet and crossed out the two central targets as an example to the subject. The subject was then asked to cancel the remaining small stars. The number of targets omitted in each lateral half of the sheet was counted. The star cancellation test sheet was divided into six areas (left-left, middle-left, right-left areas and right-right, middle-right, left-right areas) and was analyzed using the correct rate for six areas. 51 points were set as a cutoff value.
3-2. Special test with HMD (Figure 2)
Figure 2 Experimental setup for the HMD (head mounted display) system.
(a) Experimental apparatus
The main experimental apparatus includes a digital camera, HMD (GT270, Canon Inc.), and a digital video camera. HMD is a glass type display method (270,000 pixel, effective pixel number is 99.99%, weight is 150 g) that consists of two TFT liquid crystal panels. The digital camera takes a picture of a test sheet on the desk, and HMD presents the subject from the digital camera. Moreover, the subject's head movement was recorded by a digital video camera as a qualitative motion analysis.
(b) Assessments of USN with HMD (Figure 3)
Figure 3 Two special tests of USN with HMD.
We attempted to find the degree that USN alters when the co-ordinate of the subject's visual field was carried out as object-centered by HMD. Therefore, we used two different lens of the digital camera in order to change visual field and then HMD displayed the test paper to the subject as the two special tests as follows;
1) Special test 1: the zoom-in (ZI) condition which can display only the test paper using combined HMD and a DV camera.
2) Special test 2: the zoom-out (ZO) condition which can display 0.7 times special condition1 by changing the lens.
3-3. Procedure
The subjects sat on a wheelchair if needed or a straight back chair sitting in an up-right position as a starting point. The test paper on a desk was placed at a midline of each subject's body. All tasks were done without any restriction as to time.
The subjects were first evaluated by a normal test without HMD as the common clinical test and then two spatial tests with HMD. The line cancellation test was scored using the correct rate and then the score divided into two areas; right and left. The star cancellation test was scored using the correct rate for six areas (left-left, middle-left, right-left areas and right-right, middle-right, left-right areas) in which the test paper was divided (Figure 1). All subjects performed in random order the common clinical test and two special tests (ZI, ZO). The examiner confirmed the HMD monitor as the display from the image of the digital camera. Moreover, the movements of head, trunk, and upper/lower extremities were were qualitatively analyzed during these tests for finding an abnormal movement.
4. Data analysis
All statistics were performed using SPSS statistical software (7.5.2 J). An ANOVA or Student's t test was used as a comparison between the common clinical test and the two special tests with HMD. Moreover, a Student's t test or an ANOVA was used for a comparison within the line cancellation test and the star cancellation test, respectively. Multivariate ANOVA tests were performed in each group and Shėffe post hoc tests were performed if significant differences were found at the 5 % significance level.
The qualitative analysis of head, trunk, and upper/lower extremity movement during all tests was performed by the digital video camera in a sagittal or a frontal plane.
Results
In this study, the average of FIM-M of all subjects was 53.0 ± 21.6 points (Table 1). The subject needs maximal or moderate assistance for some performance of ADL.
As the common clinical test for USN, in the first evaluation of the frequency of presence of neglect for ADL (Table 3), 75 percent of all subjects admitted a USN symptom in activities of dressing. For example, a patient with USN cannot easily put on their clothes on the left side. Moreover, 62.5 percent of the subjects admitted a USN symptom in activities of transferring, and locomotion (Table 3). According to the motion analysis of head motion in the common clinical test, the subjects began searching from the right side in both the line and the star cancellation tests. In a normal performance, the head naturally rotated from the right to the left to follow a movement during the line cancellation test. However, the head movement to their left was insufficient for searching from the right side in the both tests. For the line cancellation test under the common condition, the mean percentage of the correct answers at the left side in the test paper was 94.4%. The right side was 100 %. Nobody fell below the cutoff value (Table 4) [30]. For the star cancellation test under the common clinical test (Table 5), the mean percentage of the correct answers at the left- left area was 91.1 %. The middle-left area was 89.2 % and the right-left side was 84.4 %. The mean percentage of the correct answers at the right-right was 92.9 %, middle-right was 96.4 %, and left-right area was 81.8 %. Three subjects fell below the cutoff value as an abnormal [30].
Table 3 Ratio of USN symptoms in ADL
n = 8 Ratio of USN (%)
talking or communicating with others 4 50.0
neglecting the left side of bed space 2 25.0
eating 1 12.5
grooming (self-care skills, washing, bathing,etc) 2 25.0
dressing 6 75.0
transferring (from a bed to W/C.etc) 5 62.5
lecomotion 1 negotiatin a W/C between doors, kerbs, etc. 5 62.5
lecomotion 2 the patient turns toward the direction of the affected side. 5 62.5
during PT exercise 6 75.0
during OT excercise 7 87.5
Table 4 Mean percentage of correct answers of the line cancellation test in the common method.
Mean percentage of correct answers (%)
left side of test sheet 95.1 ± 13.8
right side of test sheet 100 ± 0
Table 5 Mean percentage of correct answers of the star cancellation test in the common method.
Percentage of correct answers (%)
Correct answers of left-left 91.1 ± 13.7
Correct answers of right-left 81.8 ± 31.1
Correct answers of mid-left 89.3 ± 8.6
Correct answers of mid-right 96.4 ± 5.9
Correct answers of left-right 84.4 ± 30.1
Correct answers of right-right 92.9 ± 14.0
For the special test with HMD, in the motion analysis of head motion, the subjects began searching from the right side in both the line and the star cancellation tests. However, seven subjects kept rotating only on the right side. They did not rotate to the left side. For the line cancellation test under the ZI condition in the special test with HMD (Table 6), the mean percentage of the correct answers at the left side in the test paper was 61.8 %. The right side was 92.4 %. For the ZO condition, the mean percentage of the correct answers at the left side in the test paper was 79.9 %. The right side was 91.7 %. In both ZI and ZO conditions, the left score was significantly greater than the right score (p < 0.05). There was a significant difference between the common clinical test and ZI conditions of the special test with HMD for the left side score (p < 0.05). For the star cancellation test under the ZI condition in the special test with HMD (Table 7.), the mean percentage of the correct answers at the left- left area was 60.7 %. The middle-left area was 69.6 % and the right-left side was 77.9 %. The mean percentage of the correct answers at the right-right was 87.5 %, middle-right was 92.9 %, and left-right area was 87.0 %. For the ZO condition, the mean percentage of the correct answers at the left- left area was 69.7 %. The middle-left area was 70.8 % and the right-left side was 77.9 %. The mean percentage of the correct answers at the right-right was 97.9 %, middle-right was 87.5 %, and a left-right area was 92.4 %.
Table 6 Mean percentage of correct answers of the cancellation test in three conditions
correct answers for left side (%) correct answers for right side (%)
Common 95.1 ± 13.8ab 100 ± 0
ZI 61.8 ± 34.3a 92.3 ± 11.1
ZO 79.8 ± 37.6a 91.7 ± 14.5
a significant difference between right and left (p < 0.05)
b significant difference between common and ZI (p < 0.05)
Table 7 Mean percentage of correct answers of the star cancellation test in three conditions
Mean percentage of correct answers(%)
Common ZI ZO
Correct answers of left-left 91.1 ± 13.7 60.7 ± 47.0 66.7 ± 51.6
Correct answers of right-left 81.8 ± 31.1 87.0 ± 10.2 69.7 ± 38.4
Correct answers of mid-left 89.3 ± 8.6 69.6 ± 37.4 70.8 ± 42.3
Correct answers of mid-right 96.4 ± 5.9 92.9 ± 6.4 87.5 ± 13.7
Correct answers of left-right 84.4 ± 30.1 77.9 ± 37.0 69.9 ± 38.4
Correct answers of right-right 92.9 ± 14.0 87.5 ± 14.3 97.9 ± 4.9
Discussion
All subjects reported that the HMD presented a brighter, clearer image almost at real time and there was no discomfort in wearing the HMD. In this study, HMD can be shown as if the subject was looking at a 52 inch display screen 2 m away. Moreover, a change in the range of indirect vision field became possible by operating the input method using the HMD with a computer.
A digital camera was used for projecting the test sheet on the liquid crystal screen of the HMD. This camera was fixed, so that the test sheet reflected on the liquid crystal screen of HMD did not move, even if the head did during a test. This implies that the special test with HMD produced a better suited condition of the object-centred allocentric co-ordinates than the common condition test did. In this study, ZI condition was the same as that of the object-centred allocentric co-ordinates.
For motion analysis during the special test with HMD, the results showed that the subjects had the tendency to mainly focus on the right side of the test sheet under the conditions of ZI and ZO as compared to the common clinical test for USN. In a viewing the video recording as a qualitative motion analysis, when subject performed special test with HMD, there was a tendency that the subject tried to concentrate more on the right side of the test sheet. It may be that the subject's neglect was enhanced by HMD. Since the special test with HMD produced the object-centered allocentric coordinate, the subject focused more on the test sheet itself than the common clinical test. This means that if the subject pays too much attention to an object, it may be risk factor that he/she ignores the left side. Moreover, Ishiai et al. examined USN patient's eye movement using an eye camera [31]. The eye movement of a healthy person and the patients with homonymous hemianopia who have no USN symptoms could maintain a central focus. However, the patients with homonymous hemianopia who also have USN symptoms veered to the right side and their eyes did not move to the left side. HMD might be able to better clarify the left neglected area because the patients can concentrate on the object (test sheet) by limiting the viewing area as compared with the common clinical test.
The correct answer rate of the left space under ZI and ZO conditions was significantly lower than those in the common clinical test. Moreover, the correct answer rate which rose under the ZO condition was slightly greater than that of the ZI condition. It might be considered that the ZI condition placed a greater focus on an object more than the ZO condition. These results indicated that when the patients with USN concentrated on an object, their USN symptoms were more aggravated. The subjects'dressing, transferring, and locomotion of checklist by Halligan et al. indicated high percentage of presence of USN symptom [28]. Although the common BIT did not sufficiently show USN where the correct answer rate score of left space was more than 80%, the special test with HMD indicated USN where the correct answer rate score of the left space was about 60%. The HMD test may be able to better find a USN symptom which can not be easily discovered by the common clinical test.
In our former study, the use of the HMD improved the neglect symptoms in all subjects who had right cerebral hemisphere damage [32]. Rossetti et al. investigated the effect of prism adaptation on neglect symptoms, including the pathological shift of the subjective midline to the right [33]. They reported that all patients exposed to the optical shift of the visual field to the right were improved in their manual body-midline demonstration and on their classical neuropsychological tests. Lee et al. [34], Woo and Mandelmant [35] also suggested the effectiveness of the Fresnel prism when placed on a spectacle lens for improving various visual-field losses. The improvement induced by the HMD indicates that a signal is given to the brain that stimulates the natural recovery process in the same manner as the prism adaptation method. Moreover, the HMD system may lead to the further correction of left neglect than a Fresnel prism placed on a spectacle lens. Since a high power Fresnel prism membrane for obtaining a wide field of view is not clear, the prism produces a distortion of a real image and has lowered capabilities of visual acuity. By contrast, the HMD has the possibility of obtaining various fields of view without deterioration of visual acuity.
The HMD system has the advantage of being non-invasive, safe, and one can easily change the size of the visual field. Although the standard clinical examinations [36,37] were mainly used in a horizontal two-dimensional plane, the HMD system can easily produce a standard clinical examination related more closely to ADL in other planes, frontal or sagittal plane. On the other hand, the HMD system has to develop greater portability, a lighter weight and a decreased delay of response between the computer and the HMD regarding a transformation of data. The system's delay time is 50 m seconds. Therefore, the HMD system needs a higher level of technology of processing, recording and displaying a changed visual field of view in near or real time.
Technique of the HMD system may play an important role in the neuropsychological rehabilitation of unilateral spatial neglect as an evaluation device. Bowen et al. performed a systematic review of publish reports. They reported that 17 of which directly compared right brain damage (RBD) and left brain damage (LBD) and USN occurs more frequently after RBD than LBD was apparently supported by a systematic review of published data. However, an accurate estimate of the rates of occurrence and recovery after stroke could not be derived. They suggested that different USN disorders may exist, which may require type-specific rehabilitation approaches. Our system may have clinical implication for new assessment because HMD can change versatile visual input to fit each patient's degree of USN. Because, a clinical assessment for USN may be able to use various images in HMD by a computer such as change of colors and partial enlarge or reduce of real image, and to produce suitable visual information in HMD for each patient who has USN.
In this research, HMD evaluation could produce the condition of an object-centred allocentric co-ordinate. This means that our system can focus on the evaluation of the allocentric system to a greater degree than the egocentric system. A future study should be able to produce the condition of an egocentric system. In this case, a HMD display should be synchronized with a small CCD camera to be placed on the head or trunk. Moreover, eye and head movements should be measured in order for an analysis of eye – head or eye – hand coordination. It may be that eye and head movements are related to USN symptoms. In addition, we should identify the mechanisms behind the effectiveness of the HMD system and gather more from the patients.
In conclusion, the results showed that the assessment of USN using an HMD system may clarify the left neglect area which can not be easily observed in the clinical evaluation for USN. Moreover, it might be hypothesized that the USN test using HMD may display greater accuracy and be able to assess the occurrence and degree of USN more than the common clinical test. HMD can produce an artificially versatile environment ass compared to the common clinical evaluation.
==== Refs
Heilman KM Heilman KM, Valenstein E Neglect and related disorders Clinical Neuropsychology 1979 New York: Oxford University Press
Friedland RP Weinstein EA Hemi-inattention and hemisphere specialization: introduction and historical review Adv Neurol 1977 18 1 13 411354
Wade DT Skilbeck CE Hewer RL Predicting Barthel ADL score 6 months after an acute stroke Arch Phys Med Rehabil 1983 64 24 28 6849630
Katz N Hartman-Maeir A Ring H Sorken N Functional disability and rehabilitation outcome in right hemisphere damaged patints with and without unilateral spatial neglect Arch Phys Med Rehabil 1999 80 379 384 10206598 10.1016/S0003-9993(99)90273-3
Denes G Semenza C Stoppa E Lis A Unilateral spatial neglect and recovery from hemiplegia: follow-up study Brain 1982 105 543 552 7104665
Campell DC Oxbury JM Recovery from unilateral spatial neglect Cortex 1976 12 303 312 1009767
Ogden JA Anterior-posterior interhemispheric differences in loci of lesions producing visual hemineglect Brain Cognition 1985 4 59 75 10.1016/0278-2626(85)90054-5
Bisiach e luzzatti c unilateral neglect of representational space Cortex 1978 14 129 133 16295118
Kinsbourne M A model for the mechanism of unilateral neglect of space Trans Am Neurol Assoc 1970 95 143 146 5514359
Hier DB Mondlock J Caplan LR Recovery of behavioral abnormalities after right hemisphere stroke Neurology 1983 33 345 350 6681880
Albert ML A Simple test of visual neglect Neurology 1973 23 658 664 4736313
Diller L Weinberg J Hemi-inattention in rehabilitation: the evolution of rational remediation program Adv Neurol 1977 18 63 82 920526
Calpan B Assessment of unilateral neglect: a new reading test J Clinical and Experimental Neuropsychology 1987 9 359 364
Goldstein G Some recent developments in clinical neuropsychology Clin Psychology Rev 1981 1 245 268 10.1016/0272-7358(81)90006-4
Karnath HO Disturbed coordinate transformation in the neural representation of space as the crucial mechanism leading to neglect Neuropsychological Rehabilitation 1994 4 147 150
Calvanio R Petrone PN Levine DN Left visual spatial neglect in both environment-centered and body-centered Neurology 1987 37 1179 1183 3601081
Ogden JA Jeannerod M The "neglected" left hemisphere and its contribution to visuospatial neglect neuropsysiological and neuropsychologocal aspects of spatial neglect 1987 Amsterdam: North Holland
Diver J Halligan PW Can visual neglect operate in object-centered co-ordinatres? An affirmative single-case study Coginitive Neuropsychology 1991 8 475 496
Karnath HO Subjective body orientation in neglect and the interactive contribution of neck muscle proprioception and vestibular stimulation Brain 1994 117 1001 1012 7953584
Pizzamiglio L Frasca R Guariglia C Incoccia C Antonucci G Effect of optokinetic stimulation in patients with visual neglect Cortex 1990 26 535 540 2081391
Rossetti Y Rode G Pisella L Farne A Li L Boisson D Perenin M-T Prism adaptation to a rightward optical deviation rehabilitates left hemispatial neglect Nature 1998 395 166 169 9744273 10.1038/25988
Majoros AE Boyle E Salvendy G Maintainability Handbook of human factors and ergonomics 1997 New York. John Wiley & Sons, Inc 1587 1590
Lee JH Ku J Cho W A virtual reality system for the assessment and rehabilitation of the activities of daily living Cyberpsychol Behav 2003 6 383 388 14511450 10.1089/109493103322278763
Weiss PL Naveh Y Katz N Design and testing of a virtual environment to train stroke patients with unilateral spatial neglect to cross a street safely Occup Ther Int 2003 10 39 55 12830318
Prothero JD The treatment of akinesia using virtual images Master's the-sis 1993 Industrial Engineering, University of Washington
Granger CV Hamilton BB Keith RA Zielezny M Sherwin FS Advances in functional assessment for medical rehabilitation Top Geriartr Rehabil 1986 1 59 74
Keith RA Granger CV Hamilton BB Sherwin FS Eisenberg MG, Grzesiak RC The functional independence measure: a new tool for rehabilitation Advances in Clinical Rehabilitation 1987 New York: Springer-Verlag 3503663
Halligan PW Cockburn J Wilson BA The behavioral assessment of visual neglect Neuropsychological Rehabilitation 1991 1 5 32
Wilson BA Cockburn J Halligan PW Behavioural inattention test 1987 Thames Valley Test Company, England
Isiai S Behavioural inattention test, Japanese edition 1999 Shinkoh Igaku Shuppan, Co,Ltd Tokyo, Japan
Ishiai S Eye movement patterns in unilateral spatial neglect Advances in Neurological Science 1996 40 506 513
Tanaka T Shirogane S Ohyanagi T Izumi T Yumoto H Ino S Ifukube T Application of head mounted display system for left unilateral special neglect 14th International Congress of the World Confederation for Physical Therapy 2003 Barcelona, Spain Proceedings, RR-PO-0982
Lee AG Perez AM Improving awareness of peripheral visual field using sectorial prism J Am Optom Assoc 1999 70 624 628 10561920
Woo GC Mandelman T Fresnel prism therapy for right hemianopia Am J Optom & Physiol Optics 1983 60 739 743
Shenkenberg T Bradford DC Ajax ET Line bisection and unilateral visual neglect in patients with neurologic impairment Neurology 1980 30 509 517 7189256
Gianotti G Messerli P Tissot R Qualitative analysis of unilateral spatial neglect in relation to laterality of cerebral lesions J Nueol Neurosurg Psychiat 1972 35 545 550
Bowen A McKenma K Tallis C Reasons for variability in the reported rate of occurrence of unilateral spatial neglect after stroke Stroke 1999 30 1196 1202 10356099
|
16207373
|
PMC1277843
|
CC BY
|
2021-01-04 16:37:39
|
no
|
J Neuroengineering Rehabil. 2005 Oct 5; 2:31
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-31
|
oa_comm
|
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-231620217110.1186/1476-511X-4-23ResearchActivity of peroxisomal enzymes, and levels of polyamines in LPA-transgenic mice on two different diets Eliassen Knut A [email protected] Bjørn P [email protected] Aud [email protected] Harald [email protected]ønning Helle [email protected] Srdjan [email protected] Kåre [email protected] Department Basic Sciences and Aquatic Medicine, Norwegian School of Veterinary Science, P.O. Box 8146 Dep. No-0033 Oslo, Norway2 Department of Oral Biology, University of Oslo, P.O. Box 1052 Blindern, No-0316 Oslo, Norway3 Department of Pathology, Aker University Hospital, No-0514 Oslo, Norway4 Institute of Medical Genetics, University of Oslo, P.O. Box 1036 Blindern, Oslo, Norway5 Department of Medical Genetics, Ullevål University Hospital, No-0407 Oslo, Norway2005 4 10 2005 4 23 23 10 8 2005 4 10 2005 Copyright © 2005 Eliassen et al; licensee BioMed Central Ltd.2005Eliassen 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 man, elevated levels of plasma lipoprotein (a)(Lp(a)) is a cardiovascular risk factor, and oxidized phospholipids are believed to play a role as modulators of inflammatory processes such as atherosclerosis. Polyamines are potent antioxidants and anti-inflammatory agents. It was therefore of interest to examine polyamines and their metabolism in LPA transgenic mice.
Concentration of the polyamines putrescine, spermidine and spermine as well as the activity of peroxisomal polyamine oxidase and two other peroxisomal enzymes, acyl-CoA oxidase and catalase were measured. The mice were fed either a standard diet or a diet high in fat and cholesterol (HFHC). Some of the mice in each feeding group were in addition given aminoguanidine (AG), a specific inhibitor of diamine oxidase, which catalyses degradation of putrescine, and also inhibits non-enzymatic glycosylation of protein which is implicated in the aetiology of atherosclerosis in diabetic patients. Non-transgenic mice were used as controls.
Results
Intestinal peroxisomal polyamine oxidase activity was significantly higher in LPA transgenic mice than in the non-transgenic mice, while intestinal peroxisomal catalase activity was significantly lower. Hepatic β-oxidation increased in Lp(a) transgenic mice fed the HFHC diet, but not in those on standard diet.
Hepatic spermidine concentration was increased in all mice fed the HFHC diet compared to those fed a standard diet, while spermine concentration was decreased. With exception of the group fed only standard diet, transgenic mice showed a lower degree of hepatic steatosis than non-transgenic mice. AG had no significant effect on hepatic steatosis.
Conclusion
The present results indicate a connection between peroxisomal enzyme activity and the presence of the human LPA gene in the murine genome. The effect may be a result of changes in oxidative processes in lipid metabolism rather than resulting from a direct effect of the LPA construct on the peroximal gene expression.
==== Body
Background
Elevated levels of plasma lipoprotein (a) Lp(a) are a significant cardiovascular risk factor in man [1]. We have earlier reported development of arteriosclerosis in aorta of mice transgenic for cDNA representing the human gene for Lp(a), hLPA, on a standard diet [2,3], while the non-transgenic mice only sporadically developed arteriosclerosis. In the LPA cDNA-transgenic animals, apolipoprotein(a), (apo(a)), occurs free in plasma, and we found a significant correlation between the plasma apo(a) concentration and the size of aortic lesions.
The present investigation, based on tissue samples from the animals in the above-mentioned study, was undertaken to uncover if polyamines could influence on the atherosclerotic development. There were no leftovers of blood and the aortic wall for polyamine measurements. The liver and kidney were therefore chosen.
The rationale for examining the polyamines, spermidine and spermine, is that their positive charges strongly interact with phospholipids and inhibit lipid peroxidation [4], and to a certain extent protect liposome from oxidation [5]. In addition, spermine exhibits an anti-inflammatory effect [5] and exerts an antagonistic action on platelet aggregation [6]. Since cell proliferation in the vascular walls is important in the development of atherosclerotic lesions, knowledge of levels and metabolism of polyamines are per se of interest because of their well-known importance for cell growth and differentiation. In addition it was of interest to measure the activity of the peroxisomal polyamine oxidase, an enzyme important in converting polyamines [7]. Examination of the activity of peroxisomal enzymes was of interest also because the peroxisomal proliferators activated receptors (PPARs) are known to be involved in the development of arteriosclerosis (for review, see [8]. We have earlier shown that feeding rats a diet enriched in polyamines resulted in a decrease in hepatic polyamine oxidase and catalase activity, which could be restored by simultaneously supplementing the diet with clofibrate, a peroxisomal proliferator and a hypolipidemic drug [9].
The fact that the hypolipidemic drug, clofibrate, changed the activity of polyamine oxidase indicates that the fat content of tissues may influence polyamine metabolism. It was therefore of special interest to examine if fat loading had any effect on liver polyamine content, and on polyamine oxidase activity.
In the present study we found that polyamine oxidase activity was higher in transgenic mice than in non-transgenic animals. In order to examine whether this reflects a general increase in peroxisomal enzymes, we measured the activity of two other peroxisomal enzymes, catalase and β-oxidase. Catalase decomposes H2O2, a product of oxidase activity, and thereby protects the cell against the toxic effect of H2O2, while peroxisomal β-oxidation plays the important physiological role of oxidation very long fatty acids and the side chain of cholesterol.
Some groups of mice were treated with AG because AG inhibits the formation of non-enzymatic glycosylation of proteins [10], which are implicated in the aetiology of diabetic complications, including arteriosclerosis [10,11], and because AG is a well-known specific inhibitor of diamine oxidase, which catalyses degradation of putrescine, the precursor of the polyamines; spermidine and spermine.
We here report that the introduction of the human LPA gene into FVB mouse resulted in no significant changes in the polyamine concentration in liver and kidneys, but it changed the activity of three peroxisomale enzymes namely: intestinal polyaminoxidase and catalase and hepatic acyl-CoA oxidase (responsible for β-oxidation). There seemed to be less hepatic steatosis in the transgenic mice compared to the control FVB mice.
In mice fed a HFHC diet compared to those fed a standard low fat diet, the hepatic spermidine and spermine concentrations were different, resulting in a lower ratio between spermidine and spermine concentrations in the HFHC fed mice.
Results
The results did not suggest a sex difference with respect to any of the measured parameters.
Body weight
Body-weight decreased by 2–5% during the experimental period. Animals treated with aminoguanidine showed the highest weight-loss. The reduction in the body weight may be related to the relatively high age of the mice. Mice on the high fat diet, showed an intermittent increase in body weight of 2–9%, with a peak about 3 weeks after start. However, mean body weight at the end of the treatment period, was not significantly different from those of the groups fed a standard diet (data not shown).
Polyamine oxidase activity
The small intestinal polyaminoxidase activity was significantly (p < 0.05) higher in the transgenic mouse groups than in the non-transgenic. The polyamine oxidase activity in the small intestine was about twice as high as in the liver. Data showing the effects of the treatments on intestinal polyamine oxidase activity are presented in Fig 1. Hepatic polyamine oxidase activity was 0.41 ± 0.06 nmol/min·mg and was not affected by the diet fat content, AG treatment, or transgenity.
Figure 1 Intestinal polyamine oxidase activity in LPA transgenic and non-transgenic mice. The mice were fed two different diets, one high in fat and cholesterol (HFHC diet), the other a standard diet low in fat. Some of the animals from each group (see Table 1) were treated with 0.5% AG (aminiguanidine) in the drinking water. After 100 days of treatment the activity of the enzyme was assayed. The number of mice in each group is given with numbers within each column. Experimental details are given in Materials and Methods. The histograms represent means with SD indicated. Population means that were significantly different from the non-transgenic mice are denoted with asterisk, * (p < 0.05).
Catalase activity
Figure 2 shows the catalase activity measured in homogenates from the small intestine. In all treatment groups the catalase activity in the transgenic animals was significant lower than in the non-transgenic mice. In transgenic mice, given aminoguanidine in the drinking water, the catalase activity was significantly higher in those on the HFHC diet than in those on the standard diet (Fig. 2). AG treatment of transgenic mice on the standard diet reduced the intestinal catalase activity significantly.
Figure 2 Intestinal catalase activity in LPA transgenic and non-transgenic mice. The histograms represent means with SD indicated. Population means that were significantly different from the non-transgenic mice (p < 0.05) are denoted with *. Population means denoted with • indicate difference from the corresponding mice fed the HFHC diet and ⊗ indicate significant differences from the corresponding aminoguanidine (AG) treated mice. Experimental details are given in the legend to Fig. 1 and in Materials and Methods section.
There were only minor and insignificant differences in hepatic catalase activity between transgenic and non-transgenic mice and between the two diets. The activity was roughly the same as in the small intestine of non-transgenic mice.
AG in the drinking water to mice on the HFHC diet resulted in an increased hepatic catalase activity from about 1.5 to 3.0 μmol H2O2/s·mg protein.
Peroxisomal β-oxidation (hepatic acyl-CoA oxidase activity)
Peroxisomal β-oxidation in livers of transgenic mice fed the HFHC diet and aminoguanidine in the drinking water was significantly (p < 0.05) higher than in the non-transgenic mice on the same treatment (Fig. 3). In corresponding experiment with mice on the same diet, but without aminoguanidine (Fig. 3), the difference observed did not reach statistical significance.
Figure 3 Hepatic peroxisomal β-oxidation activity in LPA transgenic mice and non-transgenic mice. Experimental details are given in the legend to Fig. 1 and in Materials and Methods section. The histograms represent means with SD indicated. Population means that were significantly different (p < 0.05) from non-transgenic mice are denoted with *. Population means denoted with ♦ indicate significant differentce from the corresponding mice fed a HFHC diet.
Intestinal diamine oxidase activity
The diamine oxidase activity in the proximal part of the small intestine was powerfully inhibited in animals given AG, irrespective of their genetic status. The mean ± SD values of diamine oxidase activity expressed as amount putrescine turnover pr. min and g tissue was 220 ± 160 and 17 200 ± 6200 pmol for the aminoguanidine treated and the non-treated animals, respectively. This is in agreement with the established inhibitory effect of aminoguanidine on diamine oxidase activity [12].
Tissue polyamines
The hepatic concentrations of putrescine wearied around the level of detection of the assay. Evaluation of these data could therefore not be carried out.
With exception of mice on standard diet and aminoguanidine treatment, transgenic and non-transgenic mice exhibited no significant differences in hepatic spermidine and spermine concentrations (Figs. 4 and 5)
Figure 4 Hepatic spermidine concentration in LPA transgenic and non-transgenic mice. The columns represent means with SD indicated. Population mean that is significantly different from that of non-transgenic mice is denoted with asterisk, *. Δ Indicate significant difference from the corresponding groups fed a standard diet. Experimental details are given in the legend to Fig 1 and in the Materials and Methods section.
Figure 5 Hepatic spermine concentration in LPA transgenic and non-transgenic mice. The columns represent means with SD indicated. Population mean that is significantly different (p < 0.05) from that of the non-transgenic mice is denoted with asterisk, *. Δ Indicate significant difference (p < 0.05) from the corresponding HFHC fed groups. Experimental details are given in the legend to Fig. 1 and in the Materials and Methods section.
The hepatic concentration of spermidine in mice fed the atherogenic diet, HFHC; was significantly (p < 0.05) lower than in mice given standard diet (Fig. 4). On the other hand, a slight, but significant increase (p < 0.05) in spermine concentration was observed in mice fed the HFHC diet without AG in the drinking water, compared to those fed the standard diet (Fig. 5). The spermidine/spermine ratio was 0.76 for mice on the HFHC diet, and 1.32 for those on standard diet.
In kidney there were no differences between transgenic and non-transgenic mice. Thus for the non-transgenic mice the putrescine, spermidine and spermine concentrations were (73 ± 36) nmol/g, (363 ± 83) nmol/g and (606 ± 101) nmol/g, respectively, and the corresponding values for the transgenic animals were 66 ± 22, 398 ± 170 and 642 ± 154 nmol/g. Furthermore, there was no difference between mice on the two diets, or between AG treated and untreated mice.
Hepatic steatosis
In the fatty infiltrations, there was no sign of inflammatory cells. Hepatic steatosis was graded from 1 to 3. Typical sections of liver with steatosis of grade 1, 2 and 3, together with sections of a normal liver are shown in Fig. 6. Steatosis occurred in the liver of all, but four mice. The grade of steatosis was lower in the transgenic mice than in non-transgenic mice, but the difference was statistical significant (p < 0.05) only for those on a standard diet with AG in the drinking water and for those fed only a high fat diet (Fig. 7).
Figure 6 Scoring of hepatic steatosis. A) Normal liver with no lipid deposits, grade 0, B) Liver tissue with steatosis grade 1 (slight steatosis), C) grade 2, that between slight and serve stearosis, D) Liver tissue with steatosis grade 3 (serve steatisis), almost all cells are affected and many with large fat vacuoles. Magnification 1000x.
Figure 7 Hepatic steatosis (grade 0 to 3) in LPA transgenic and non-transgenic mice. The columns represent means with SD indicated. Population means that are significantly (p < 0.05) different from the non-transgenic mice are denoted with asterisks, *. Experimental details are given in the legend to Fig. 1 and in the Materials and Methods section.
Discussion
Enzyme activities
Introduction of cDNA representing the LPA gene together with the transferrin promotor, into the mouse genome resulted in changes in activity of the peroxisomale enzymes polyamine oxidase and catalase in the small intestine, but not in the liver (Figs. 2, 3, 4). This is in contrast to the fact that the LPA gene is expressed in liver, but not in the intestines [13], indicating that the effect on the intestinal enzymes must be secondary. Since the genes coding for polyamine oxidase (Locus ID no Mm 212503) and for acyl-CoA oxidase (Locus ID no. Mm 11430) are located on different mouse chromosomes, 7 and 11, respectively, it is unlikely that the observed changes in these enzyme activities were caused by the introduction of the LPA gene with a transferrin promotor, in or near the regulator region of the gene for these enzymes.
The increased intestinal polyamine oxidase activity in LPA transgenic mice compared to non-transgenic mice (Fig. 1), indicates an altered peroxisomal polyamine metabolism. When the activity of the H2O2 producing polyamine oxidase enzyme is increased in the transgenic mice, an increase in the catalase activity should be anticipated. However, the opposite was the case.(Fig. 2) Increased polyamine oxidase activity can therefore not be explained by a general increase in activity of peroxisomal enzymes. The increase in H2O2 production from enhanced polyamine oxidase activity is probably marginal, compared with H2O2 production from other sources. The observed changes in hepatic peroxisomal β-oxidation (Fig. 3) cannot explain the decrease in the catalase activity because peroxisomal β-oxidation did not change in the control mice and even increased in the LPA-transgenic mice on HFHC diet. The increased peroxisomal β-oxidation in LPA transgenic mice on the HFHC diet is in agreement with reports showing that diets high in fat, particular those containing hydrogenated fat, cause increased of peroxisomal β-oxidation [14,15]. In non-transgenic mice the same diet caused no significant increase of peroxisomal β-oxidation. It follows that mice, transgenic with respect to LPA may be more prone than non-transgenic mice to increase of peroxisomal β-oxidation probably because of the low fat content (Fig. 2).
The observed increase in polyamine oxidase and β-oxidase activity together with a decrease in catalase activity indicate that also other oxidative processes are changed as a result of the introduction of the LPA gene. This seems interesting in the view of the possible role of oxidized phospholipids as modulators of inflammatory processes. Thus, modified phospholipids accumulate at sites of inflammation such as atherosclerotic lesions [16]. Furthermore, oxidized phospholipids are shown to induce expression of atherosclerosis-related genes [17]. Additional evidence has been provided by the use of murine natural monoclonal IgM antibody, EO6, which binds to oxidation-specific epitopes on oxidized low-density lipoprotein. A high correlation between plasma Lp(a) and EO6 was then found [18].
The conclusion so far must be that changes in the three peroxisomal enzyme activities in LPA-transgenic mice is more likely to be caused by altered lipid metabolism in these animals, rather than by a direct effect of the LPA construction on the expression of the peroxisomal genes.
Polyamines
In the kidneys, neither the LPA gene nor the HFHC diet and AG treatment seemed to influence the polyamine levels significantly (data not shown).
The hepatic spermidine and spermine concentration were almost unaffected by aminoguanidine treatment (Fig. 4 and 5). The same was true for the kidneys. Although the polyamine concentration in the standard diet was much higher than in the HFHC diet, the HFHC diet resulted in higher hepatic spermidine concentration. It seems not likely that differences in the intestinal micro flora caused the observed differences in hepatic polyamine content between the two feeding groups.
Clofibrate treatment, that decreases plasma lipid levels has earlier been found to increase the hepatic content of spermidine and reduce that of spermine [9], which is contrary to the effect of the HFHC diet.
The HFHC fed group had higher levels of hepatic fat, as judged by the amount of fat filled vacuoles (steatosis) than those on the standard diet. However, we found no significant correlation between hepatic steatosis and hepatic levels of spermidine or spermine. In spite of that, we suggest that hepatic lipid level may be of importance for the hepatic polyamine level. The fact that spermine binds to negatively charged lipids, e.g. phospholipids [19] may in part explain the increased spermine concentration in liver.
Spermidine in plasma can bind to HDL [20], which is known to inhibit the formation of oxidized low-density lipoprotein (LDL) [21]. Polyamines, especially spermine, are also potent antioxidants and anti-inflammatory agents [5]. The polyamine concentrations have further been reported to inhibit platelet aggregation in hypercholesterolemic rabbits [6]. Studies of the role of polyamines in atherogenesis/thrombogenesis are therefore called for.
The spermidine/spermine ratio is known to be high in fast growing tissue. Thus, in rat liver at birth the ratio is 4.5, whereas this ratio at 9 months is 0.8 [22]. One can only speculate if the cell turnover rate for liver cells in mice fed the HFHC diet, that have a hepatic spermidine/spermine ratio on 0.76, is reduced compared to that of mice on a standard diet where the corresponding ratio was 1.32.
Liver steatosis
For two of the feeding groups there was significantly lower degree of hepatic steatosis in the transgenic mice than in the non-transgenic animals (Fig 7) suggesting differences in lipid metabolism. We have earlier reported a significantly higher rate of aortic lesions in LPA transgenic mice than in non-transgenic animals and a positive correlation between apo(a) level and size of aortic lesions [2,3]. This, together with the fact that inflammatory cells only could be seen in the aortic fat infiltrations, not in that of the liver, indicate that there are different mechanisms involved in the deposition of lipids in aortic lesions and in liver cells.
The AG treated mice on the HFHC diet had less steatosis than the untreated animals but the differences were far from significant. This indicates that glycosylation is not participating in the formation of hepatic steatosis in mice.
Conclusion
An increase in polyamine oxidase activity in hLPA transgenic mice, and an increase in β-oxidation in transgenic mice fed a HFHC diet, together with a decrease in catalase activity may indicate that also other oxidative processes are changed as a result of the introduction of the LPA gene. Such changes would be of great interest because oxidation products are known to be important in the development of arteriosclerosis.
Changes in the polyamine pattern upon increased fat intake are also noteworthy. However, in this study we have not been able to show any connection between polyamines/polyamine-metabolism and the arteriosclerotic lesion seen in the same mice [2,3]. Further studies are warranted.
Materials and methods
Reagents
Horseradish peroxidase (EC 1. 11. 1. 7) (Type VI. RZ between 250 and 330 U/mg), N1-acetyl-spermine, 4-aminoantipyrine, palmitoyl-CoA, NAD+, FAD, HEPES, hexanediamine, putrescine 2HCl, spermidine 3HCl, spermine 4HCl, and mannitol were purchased from Sigma Chemical Co. (St. Louis, MO., USA). Perhydrol (H2O2) was obtained from E. Merck (Darmstadt, Germany). All other reagents were of analytical grade. [1,4-14C]-Putrescine dihydrochloride (spes.act. 2.11 GBq/mmol) was obtained from Amersham Pharmacia Biotech, Ltd. (Rainham, Essex, UK).
Animals and feeding
This stock of mice have also been used in a parallel investigation of effects of the LPA gene on the development of aortic lesions [2,3]. Forty-one transgenic mice with cDNA representing the human LPA gene linked to the mouse transferrin promotor were studied. The mice were of a hybrid genetic background of strain C57BL/6 crossed with strain SJL [2,3]. The original transgenic breeder mice were a generous gift from Dr. Richard M. Lawn, Falk Cardiovascular Research Center, Stanford University, CA., USA. Twenty-nine non-transgenic littermates from breeding between transgenic and non-transgenic mice were used as controls. DNA analyses of white blood cells were performed to ascertain the LPA transgenic status of the mice. The mice were of both sexes and between 49 and 67 weeks of age when the study started (see Table 1).
Table 1 Diet, treatments, age and sex distribution of LPA transgenic and non-transgenic mice
Non-transgenic mice Transgenic mice
Diet and treatment No. Sex Age (days), mean ± SD No. Sex Age (days), mean ± SD
Standard 12 F 6, M 6 481 ± 46 11 F 4, M 7 469 ± 36
Standard + aminoguanidine 9 F 5, M 4 501 ± 55 7 F 3, M 4 498 ± 53
HFHC 11 F 8, M 3 467 ± 16 17 F 8, M 9 486 ± 42
HFHC + aminoguanidine 7 F 5, M 2 498 ± 51 5 F 2, M 3 476 ± 20
F = females
M = males
Half of the animals were fed a standard mouse diet (RMI (E) SQS, Special Diets Services, Witham, Essex, England) containing 2.6% fat (crude oil), where saturated fat accounted for 20%. The other half of the animals, were fed an atherogenic semi-purified diet high in fat and cholesterol (HFHC diet), containing 1.25% cholesterol, 18.4% regular butter and 0.5% sodium cholate (ICN Pharmaceuticals, Inc. 1731 Asse-Relegem, Belgium). Several of the animals on each diet were given 0.1% (w/v) aminoguanidine in the drinking water (Table 1). All the treatments lasted for 100 days.
The amounts of polyamines in the standard diet were: 110 nmol/g putrescine, 240 nmol/g spermidine and 45 nmol/g spermine. The corresponding figures for putrescine and spermidine in the HFHC diet were 1.0 and 3.5 nmol/g, respectively. Spermine could not be detected.
Females and males were housed separately, with 1–5 mice per cage. They had a 12 hrs. light/dark cycle, a constant temperature of 21°C and a relative humidity of 65%. Sentinel mice were used to run a FELASA-style health-monitoring scheme. The mice were clinically healthy throughout the experimental period. The mice were raised in The Laboratory Animal Unit at Norwegian School of Veterinary Science (accredited by the Association for Assessment and Accreditation of Laboratory Animal Care and Use International, Brussels, Belgium), and they were kept according to the regulations of the Norwegian Gene Technology Act of 1994. The study was approved by the Norwegian Animal Research Authority.
Blood sampling and preparation of tissue samples for enzyme measurements
The mice were anaesthetized by a subcutaneous injection of 0.05 ml/10 g body-weight of a 1:1 mixture of Fenatyl/Midazolum (Hypnorm "Janssen" diluted 10x with water and Dormicum "Roche" 5 mg/ml) prior to blood sampling, which was done by an incision in vena saphena. The mice were afterwards killed by neck dislocation, and tissues for analysis were removed as quickly as possible. After removal, the liver was weighed, and transferred into ice-cold mannitol-medium (300 mmol/L mannitol, 25 mmol/L HEPES, 1 mmol/L EGTA, pH 7.2). A 10% (w/v) homogenate was prepared by 2 strokes in a Potter-Elvehjem homogenizer equipped with a Teflon® piston. The homogenate was centrifuged for 1 min at 1075 × gav. The resulting supernatants were transferred into several small vials, and stored at -20°C until analyzed.
A 20 cm long proximal segment of the small intestine, starting about 0.5 cm from the pyloric sphincter, was prepared. It was cut open longitudinally, rinsed with ice-cold 0.9% (w/v) NaCl, blotted against filter paper, and homogenized for 15 s in ice-cold mannitol medium with an Ultra-Turrax® at 25,000 rpm. The resulting homogenate was subsequently centrifuged at 1075 × gav for 15 min and the supernatant was stored at -20°C until it was analyzed. Prior to assays, one part of the homogenate was mixed with one part of ice-cold mannitol medium, and then Triton X-100 was added to a resulting concentration of 0.05% (v/v).
Enzyme assays
Catalase (EC 1.11.1.6) activity was assayed by monitoring the decomposition of H2O2 at 240 nm and 25°C essentially as described by Bergmeyer et al [23]. Catalase is labile in dilute solution. Therefore, homogenates were diluted 10 times with 20 mmol/L potassium phosphate buffer, pH 7.4 immediately prior to analysis.
Peroxisomal β-oxidation was assayed as palmitoyl-CoA-dependent NAD+-reduction, according to Hovik and Osmundsen [24]. Polyamine oxidase (polyamine: oxygen oxidoreductase, EC 1.5.3.3) was assayed spectrophotometrically as described by Hayashi et al. [25]. Diamine oxidase, EC 1.4.3.6 was determined radiometrically, as described by Okuyama and Kobayashi [26].
Catalase activity, peroxisomal β-oxidation and polyamine oxidase activity were measured in 6 randomly selected mice from each of 7 groups of animals, whereas all 5 mice in the transgenic group on HFHC diet and AG were examined. As seen in Fig. 3 the numbers of mice in some other groups were also less than 6, due to accidental loss of some samples.
Polyamine assays
Tissue specimens were weighed and thereafter homogenized in 4 volumes of 5% trichloroacetic acid. Hexane diamine was added as an internal standard. The homogenates were kept on ice for 1 hr, followed by centrifugation at 2°C for 10 min at 5000 gav. The supernatant was stored at -20°C until analyzed. The polyamines were dansylated [27,28] and separated by HPLC [29] on a Radial PAK-A column (Waters, Milford, Ma., USA).
Assay of proteins
Protein assay were carried out using the biuret method [30] with Boehringer Precimat (Boehringer Mannheim, Mannheim, Germany) as protein standard.
Liver steatosis
Liver tissue for morphometric studies were removed and immediately placed in a buffered formaldehyde solution. The examination was made routinely on sections of liver, in paraffin blocks, stained with Oil red O and counterstained with hematoxylin. Hepatic steatosis was identified by light microscopy as empty vacuoles in the cytoplasm of liver cells (Fig. 7). For definitive identification of lipid, frozen sections from separate specimens were stained with Oil Red O. The same pathologist examined the sections blindly, and all measurements were performed four times. The evaluation of visible fat in the liver cells was assessed using a semi quantitative scale from 0 to 3. Zero indicates that no vacuolated liver cells could be detected. In the slightest degree of steatosis (grade 1) tiny vacuoles are found in small parts of the lobule, whereas in the severe case (grade 3) almost every liver cell is affected and most cells have large vacuoles. "Moderate" (grade 2) denotes fatty changes between that of slight and severe steatosis (Fig. 7).
Statistical analysis
To calculate the significance of differences between non-transgenic and transgenic mice on the same treatment we used an unpaired, two tailed t-test, employing the GraphPad Prism v.2.0 program (GraphPad software Inc., San Diego, USA).
The values given in the text and legends to Figures are means, with SD as indicated.
List of abbreviations
AGE, advanced glycosoationtion end products; AG, aminoguanidine; HFHC, diet high in fat and cholesterol; LPA, gene encoding apo(a).
Authors' contributions
SD was responsible for the testing of transgenicity of animals used in this study. BPB and HO were responsible for measurements of β-oxidation, polyamine oxidase and catalase activity. AS did the pathological examinations. HR measured the polyamines. KB and KAE conceived the study and performed its design and coordination. KAE drafted the manuscript. All authors read and approved the final version of the manuscript.
Acknowledgements
We gratefully thank Dr. Richard M. Lawn for the generous gift of LPA transgenic mice. The skillful technical assistance of Patrica A. Engen, Bente B. Gehrken, Tove Norén, Inger E. Nossen and Toril Woldene is highly appreciated.
==== Refs
Djurovic S Berg K Epidemiology of Lp(a) lipoprotein: its role in atherosclerotic/thrombotic disease Clinical Genetics 1997 52 281 292 9520118
Svindland A Berg K Eliassen K Lawn RM Djurovic S Alestrom P Noren T Smith A Histopathology of arterial lesions in LPA transgenic mice on cholesterol-enriched chow Atherosclerosis 2000 153 349 354 11164423 10.1016/S0021-9150(00)00430-5
Berg K Svindland A Smith AJ Lawn RM Djurovic S Alestrom A Alestrom P Eliassen K Spontaneous atherosclerosis in the proximal aorta of LPA transgenic mice on a normal diet Atherosclerosis 2002 163 99 104 12048126 10.1016/S0021-9150(01)00772-9
Tadolini B Cabrini L Landi L Varani E Pasquali P Polyamine binding to phospholipid-vesicles and inhibition of lipid-peroxidation Biochemical and Biophysical Research Communications 1984 122 550 555 6466327 10.1016/S0006-291X(84)80068-6
Lovaas E Carlin G Spermine – an antioxidant and anti inflammatory agent Free Radical Biology and Medicine 1991 11 455 461 1663062 10.1016/0891-5849(91)90061-7
de la Pena HC Sosa-Melgarejo JA Ramos RR Mendez JD Inhibition of platelet aggregation by putrescine, spermidine, and spermine in hypercholesterolemic rabbits Archives of Medical Research 2000 31 546 550 11257319 10.1016/S0188-4409(00)00238-1
Seiler N Bolkenius FN Rennert OM Interconversion, catabolism and elimination of the polyamines Medical Biology 1981 59 334 346 7040832
Duval C Chinetti G Trottein F Fruchart JC Staels B The role of PPARs in atherosclerosis Trends in Molecular Medicine 2002 8 422 430 12223313 10.1016/S1471-4914(02)02385-7
Brodal BP Eliassen KA Ronning H Osmundsen H Effects of dietary polyamines and clofibrate on metabolism of polyamines in the rat Journal of Nutritional Biochemistry 1999 10 700 708 15539269 10.1016/S0955-2863(99)00058-3
Brownlee M Vlassara H Kooney A Ulrich P Cerami A Aminoguanidine prevents diabetes-induced arterial-wall protein cross-linking Science 1986 232 1629 1632 3487117
Li YM Steffes M Donnelly T Liu C Fuh H Basgen J Bucala R Vlassara H Prevention of cardiovascular and renal pathology of aging by the advanced glycation inhibitor aminoguanidine Proceedings of the National Academy of Sciences of the United States of America 1996 93 3902 3907 8632987 10.1073/pnas.93.9.3902
Schayer RW Kennedy J Smiley RL Studies on histamine-metabolizing enzymes in intact animals II Journal of Biological Chemistry 1953 205 739 748 13129253
Teivainen PA Eliassen KA Rubin EM Djurovic S Berg K Human apoB contributes to increased serum total apo(a) levels in LPA transgenic mice Lipids in Health and Disease 2004 3 8 15134578 10.1186/1476-511X-3-8
Neat CE Thomassen MS Osmundsen H Effects of high-fat diets on hepatic fatty acid oxidation in the rat. Isolation of rat liver peroxisomes by vertical-rotor centrifugation by using a self-generated, iso-osmotic, percoll gradient Biochemical Journal 1981 196 149 159 6272750
Thomassen MS Christiansen EN Norum KR Characterization of the stimulatory effect of high-fat diets on peroxisomal beta-oxidation in rat liver Biochemical Journal 1982 206 195 202 7150240
Furnkranz A Leitinger N Regulation of inflammatory responses by oxidized phospholipids, structure-function relationships Current Pharmaceutical Design 2004 10 915 921 15032694 10.2174/1381612043452929
Furnkranz A Schober A Bochkov VN Bashtrykov P Kronke G Kadl A Binder BR Weber C Leitinger N Oxidized Phospholipids Trigger Atherogenic Inflammation in Murine Arteries Arterioscler Thromb Vasc Biol 2005 25 633 638 15591214 10.1161/01.ATV.0000153106.03644.a0
Bergmark C Oxidized Phospholipids and Lp(a) Lipoprotein, a role in Innate Immunity? "Genomics and Genetic Applications in Medicine" Symposium organized by SK Thoresens foundation and The Norwegian Academy of Sciences, Oslo, May 19 – May 21, 2005 44 46 Abstract
Tadolini B Cabrini I Varani E Sechi AM Spermine binding and aggregation of vesicles of different phospholipid composition Biogenic Amines 1985 3 87 96
CatrosQuemener V Leray G Moulinoux JP Havouis R deCertaines JD Chapman J Tumour growth modifies intravascular polyamine transport by plasma lipoproteins in the mouse Biochimica et Biophysica Acta-Lipids and Lipid Metabolism 1997 1346 30 37 10.1016/S0005-2760(97)00016-7
Navab M Hama SY Cooke CJ Anantharamaiah GM Chaddha M Jin L Subbanagounder G Faull KF Reddy ST Miller NE Fogelman AM Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: step 1 Journal of Lipid Research 2000 41 1481 1494 10974056
Russell DH Polyamines as biochemical markers of normal and malignant growth Progress in cancer Research and Therapy 1973 9 1 13
Bergmeyer HU Grassl M Walter HE Catalase Verlag Chemie, Weinhein, Germany 1983 II 165 167
Hovik R Osmundsen H Peroxisomal beta-oxidation of long-chain fatty acids possessing different extents of unsaturation Biochemical Journal 1987 247 531 535 3426550
Hayashi H Yoshida H Hashimoto F Okazeri S Changes in polyamine-oxidizing capacity of peroxisomes under various physiological conditions in rats Biochimica et Biophysica Acta 1989 991 310 316 2719974
Okuyama T Kobayashi Y Determination of diamine oxidase activity by liquid scintillation counting Archives of Biochemistry and Biophysics 1961 95 242 250 14481667 10.1016/0003-9861(61)90141-2
Smith TA Best GR Polyamines in barley seedlings Phytochemistry 1977 16 841 843 10.1016/S0031-9422(00)86676-5
Seiler N Wiechmann M Microdetermination of spermine and spermidine as 1-dimethylaminonaphthalene-5-sulfonic acid derivatives Hoppe Seylers Z Physiol Chem 1967 348 285 1290 5589206
Seiler N Knodgen B Eisenbeiss F Determination of di- and polyamines dimethylaminonaphthalene-1-sulfonyl derivatives Journal of Chromatography 1978 145 29 39 621245
Gornall AG Bardawill CJ David MM Determination of serum proteins by means of the biuret reaction Journal of Biological Chemistry 1949 177 751 766
|
16202171
|
PMC1277844
|
CC BY
|
2021-01-04 16:39:19
|
no
|
Lipids Health Dis. 2005 Oct 4; 4:23
|
utf-8
|
Lipids Health Dis
| 2,005 |
10.1186/1476-511X-4-23
|
oa_comm
|
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-241620970710.1186/1476-511X-4-24ResearchThe influence of walking performed immediately before meals with moderate fat content on postprandial lipemia Pfeiffer Martina [email protected] Tanja [email protected] Caspar [email protected] Paolo C [email protected] INW Nutrition Biology, Department of Agriculture and Food Science, Swiss Federal Institute of Technology Zurich, Universitätsstrasse 2, 8092 Zurich, Switzerland2005 6 10 2005 4 24 24 16 9 2005 6 10 2005 Copyright © 2005 Pfeiffer et al; licensee BioMed Central Ltd.2005Pfeiffer 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
Postprandial lipemia is an independent risk factor for coronary heart disease. Single bouts of moderate exercise may lower this risk, but the minimum duration of moderate intensity exercise that still lowers postprandial lipemia is not known. We, therefore, performed a dose-response study with a normal, daily life setting, to identify the minimum duration of moderate intensity walking that lowers postprandial lipemia in sedentary, healthy young men.
Methods
Sixteen men performed three activity trials (30, 60, or 90 min of treadmill walking at 50% of their individual VO2max) and a control trial with no physical activity in a repeated measures crossover design. The subjects walked immediately before ingestion of the first of two mixed meals, which were served 3 h apart. The meals had a moderate fat content (0.5 g per kg body mass and 33% of total energy per meal) and a macronutrient composition corresponding to current recommendations. Each meal provided one third of the subject's estimated daily energy requirement. Venous blood samples were taken in the fasted state, and then hourly for 6 h after the first meal to assess the postprandial phase. Postprandial lipemia (the incremental area under the curve (dAUC) of triacylglycerol) was compared with a mixed model analysis and Tukey's adjustment.
Results
Postprandial lipemia (dAUC of triacylglycerol) was, compared to the control trial, +2% (P = 1.00), -14% (P = 0.24), and -15% (P = 0.23) in the 30, 60, and 90 min walking trials, respectively.
Conclusion
Moderate intensity walking of 60 and 90 min duration slightly, but insignificantly, reduced postprandial lipemia after two mixed meals with moderate fat content in sedentary, healthy young men, compared to inactivity. Therefore, it should be reconsidered if the acute exercise-induced reduction in postprandial lipemia usually observed in studies using high fat meals is of importance in a real, daily life setting.
==== Body
Background
Postprandial lipemia describes the blood triacylglycerol (TAG) content after meal intake, and is an independent risk factor for atherosclerosis and coronary heart disease [1-4]. The latter is the leading cause of death in industrialized countries and is rapidly becoming a primary cause of death worldwide [5]. Interventions that have the potential to attenuate the postprandial lipemic response are, therefore, valuable tools for lowering the risk of cardiovascular diseases.
Endurance exercise is known to positively influence postprandial lipemia [6], and endurance athletes show a lower postprandial lipemic response than sedentary people [7]. However, it seems to be the acute response to single exercise bouts rather than the improved endurance capacity per se that has the favourable effect [8], because postprandial lipemia increases rapidly with detraining [9,10]. Furthermore, the intensity at which endurance exercise is usually performed is rather high, which makes endurance exercise not a very attractive option for reducing the cardiovascular risk for sedentary people.
A single exercise bout with moderate intensity only was also shown to improve postprandial lipemia when performed immediately before meal ingestion, but usually the exercise duration used was one hour or longer [11-13]. Since about 60% of the world's population already does not meet the minimum recommendation of daily physical exercise [14], which is 30 min of moderate intensity activity on most, preferably all, days of the week [15], it is unlikely that sedentary people would engage in exercise of long duration. Therefore, it seems important to determine the influence of moderate exercise with shorter duration on postprandial lipemia, but few studies have investigated this exercise mode. We are aware of two studies in which the exercise bouts were performed just before ingestion of a meal and corresponded to an intensity and duration required for health-maintaining physical activity, e.g. moderate intensity exercise with a duration of less than one hour [16,17]. Whereas Murphy et al. [16] found a significant influence of 30 min of brisk walking (60% of maximal oxygen uptake, VO2max) on postprandial lipemia in overweight or obese and older subjects, Petridou et al. [17] found a non-significantly lowered TAG response of 17% after 45 min of cycling at 62% of predicted maximal heart rate in sedentary but otherwise healthy young men.
In postprandial lipemia studies, a high fat meal (e.g. 1 g fat per kg body mass, BM) is usually provided to the subjects [18] to induce a strong increase of postprandial lipemia [19]. However, such high fat meals do not represent normal meals either in absolute or relative fat content, and do not correspond to the current recommendations about macronutrient composition [20]. As even 15–30 g of fat already elevate TAG significantly [18], high fat meals are not a precondition to study postprandial lipemia. In order to obtain information on postprandial lipemia that is relevant to real life, it seems more reasonable to choose normal mixed meals with a moderate fat content, as has been done recently in two studies investigating the influence of exercise on postprandial lipemia [17,21]. Additionally, free-living individuals consume sequential meals during the course of the day, but we know of only one study [16] where more than one meal was provided to the subjects. Since second meal effects of postprandial lipemic responses to mixed meals have been reported [22], the results of the studies with just one test meal do not accurately reflect real life [23].
The aim of the present study was, therefore, to determine the minimum duration of walking with a moderate intensity that would significantly lower postprandial lipemia in sedentary, healthy young men in a normal, daily life setting. The subjects performed three activity trials (30, 60, or 90 min of treadmill walking at 50% of their individual VO2max), and a control trial with no physical activity. Immediately after exercising they ingested the first of two mixed meals, which were served 3 h apart. The meals had a moderate fat content (0.5 g per kg BM and 33% of total energy per meal), a macronutrient composition corresponding to current recommendations [20], and each one provided one third of the subject's estimated daily energy requirement.
Results
Exercise sessions
Mean relative oxygen uptake with the 30, 60, and 90 min walking sessions was 49.8 (0.2)%, 50.1 (0.1)%, and 50.0 (0.2)% of VO2max, respectively. Data comparing mean metabolic responses between the three exercise sessions are presented in Table 1.
Table 1 Metabolic responses during 30, 60, and 90 min treadmill walking in the activity trials
30 min 60 min 90 min
Mean SEM Mean SEM Mean SEM
Energy expenditure [kJ per exercise session] 879a 46 1799b 83 2630c 126
Energy expenditure [kJ·min-1] 29 2 30 1 29 1
Respiratory exchange ratio 0.95a 0.01 0.92b 0.01 0.91b 0.01
Fat oxidation [g per exercise session] 4a 1 12b 2 21c 2
Fat oxidation [g·min-1] 0.12a 0.03 0.20ab 0.03 0.23b 0.02
Carbohydrate oxidation [g per exercise session] 44a 3 80b 6 110c 6
Carbohydrate oxidation [g·min-1] 1.46a 0.11 1.33ab 0.10 1.22b 0.06
Heart rate [min-1] 122 4 125 4 127 3
Heart rate [% HRmax] 64 2 65 2 66 1
Rating of perceived exertion* 11.0a 0.4 11.5ab 0.4 11.8b 0.4
SEM, standard error of the mean; HRmax, maximum heart rate.
*Borg rating of relative perceived exertion, 6–20 scale.
Significant difference between means not sharing a common letter (P < 0.05; n = 15 for respiratory data in 30 and 60 min walking trials; n = 16 for other data).
Blood chemistry
The postprandial TAG profile did not differ statistically between the trials (P = 0.06; Fig. 1). Highest values were observed 5 h after breakfast and were 2.63 (0.28) mmol· L-1, 2.65 (0.26) mmol· L-1, 2.47 (0.28) mmol· L-1, and 2.39 (0.26) mmol· L-1 with the control, 30 min, 60 min, and 90 min walking trials, respectively, whereas mean postprandial TAG values (area under the curve (AUC) divided by six) were 1.78 (0.17) mmol· L-1, 1.82 (0.17) mmol· L-1, 1.71 (0.18) mmol· L-1, and 1.71 (0.16) mmol· L-1, respectively. Compared to the control trial the increase in AUC (dAUC) of TAG was +2% (P = 1.00), -14% (P = 0.24), and -15% (P = 0.23) for the 30, 60, and 90 min walking trials, respectively (Fig. 2). To detect a statistically significant difference in dAUC of TAG between the inactivity and the 60 or 90 min walking trials a sample size of 279 and 182 subjects, respectively, would have been required (Power = 0.80).
Figure 1 Triacylglycerol concentration for the inactivity and the activity (30, 60, or 90 min treadmill walking) trials. For better readability the symbols of the different trials at the same time points are slightly staggered. Values are means (standard error of the mean). Overall mixed model analysis with repeated measurements, effect of trial: P = 0.06.
Figure 2 Incremental area under the curve (dAUC) for postprandial triacylglycerol (mean, standard error of the mean) for the inactivity and the activity (30, 60, or 90 min treadmill walking) trials. Overall mixed model analysis, effect of trial: P = 0.07.
Insulin values were highest one hour after breakfast in all trials. A second peak was observed one hour after lunch (Fig. 3). Compared to the control trial postprandial insulinemia (dAUC) was lowered by 9% (P = 0.79), 16% (P = 0.54), and 19% (P = 0.20) for the 30, 60, and 90 min walking trials, respectively (Fig. 4). There was a significantly lower insulin to glucagon ratio one hour after breakfast in the 90 min exercise trial compared to the control trial (P = 0.01). No difference between the postprandial glucose values of all trials either in the profile over time (P = 0.50) or dAUC (P = 0.95) was observed. Profiles of glucose, glucagon, fatty acids (FA), and glycerol are presented in Figure 5A–D.
Figure 3 Insulin concentration for the inactivity and the activity (30, 60, or 90 min treadmill walking) trials. For better readability the symbols of the different trials at the same time points are slightly staggered. Values are means (standard error of the mean). Overall mixed analysis with repeated measurements, effect of trial: P = 0.16.
Figure 4 Incremental area under the curve (dAUC) for postprandial insulin (mean, standard error of the mean) for the inactivity and the activity (30, 60, or 90 min treadmill walking) trials. Overall mixed model analysis, effect of trial: P = 0.25.
Figure 5 Concentrations of glucose (A), glucagon (B), fatty acids (C), and glycerol (D) for the inactivity and the activity (30, 60, or 90 min treadmill walking) trials. For better readability the symbols of the different trials at the same time points are slightly staggered. Values are means (standard error of the mean). Significant differences between means not sharing a common letter (P < 0.05).
Discussion
In a dose-response study we examined the influence of 30, 60, and 90 min of moderate intensity walking just before meal ingestion to determine the minimum duration necessary to lower postprandial lipemia in sedentary, healthy young men in a normal, daily life setting. Our findings indicate that 60 or 90, but not 30 min of brisk walking induced a modest but statistically insignificant reduction of postprandial lipemia compared to an inactive control.
This is in contrast to most of the available literature, because in most studies a reduction of postprandial lipemia after a moderate intensity exercise bout was found [11-13,16]. The main difference between those studies and the current study was the type or number of test meals. Our test meals had only a moderate fat content (0.5 g per kg BM and 33% of total energy per meal) and were composed of ordinary, commercially available food. Additionally, to simulate the course of a normal day we served two meals (breakfast and lunch), each meal providing one third of the subject's estimated daily energy requirement. In three studies with healthy subjects who exercised immediately before ingestion of a meal, a significant reduction in postprandial lipemia was observed after 60 min walking at 60% of VO2max [13], 90 min at 50% of VO2max [11], and 90 min at 65% of VO2max [12] (postprandial lipemia (dAUC) was lowered by 38%, 49%, and 39%, compared to inactivity, respectively). However, in these studies high fat meals were served (95–101 g, 81–92% of total energy). Since in our study the 60 and 90 min walking sessions did not significantly lower postprandial lipemia, we assume that our test meals with the moderate fat content only and the macronutrient composition according to the current recommendations [20] may have largely contributed to the non-significant response observed.
Besides the amount of fat, the other macronutrients can also influence postprandial lipemia. Carbohydrates in a mixed meal provoke an insulin response, which plays an important role in the regulation of plasma TAG concentration [8], e.g. through down-regulating lipoprotein lipase (LPL) activity in skeletal muscle [24]. This insulin effect is in opposition to the acute effect of exercise (90 min of cycling at 85% of the lactate threshold) that increased muscle LPL activity in the immediate 3 to 4 h after exercise, relative to rest in men [25]. Consequently, it could also be assumed that in our study the suppressing effect of insulin on muscle LPL activity after the meal intake exceeded the stimulating effect of exercise. However, since according to Gill [8] a reduction in postprandial lipemia after moderate intensity exercise would be caused rather through decreased very-low-density lipoprotein (VLDL) production than increased TAG clearance, this is not very likely to have been the case. The moderate exercise sessions elevated FA and glycerol levels significantly, but only for a short time. Nevertheless, it seems relevant and worth mentioning, since it has not always been done in previous studies, that in lipemia studies including exercise, the TAG measurements are corrected for glycerol in order not to overestimate TAG values immediately and up to one hour after exercise.
In one study where three mixed meals with a fat content (47% of total energy) above the recommendation [20] were served, 30 min of brisk walking (60% of VO2max) performed before breakfast lowered postprandial lipemia significantly [16]. However, the subjects were overweight or obese and older (34–66 years of age) than in our study (20–30 years). Postprandial lipemia seems to behave differently in healthy, untrained adults of different age groups [26]. Therefore, the influence of an exercise bout on postprandial lipemia might also be influenced by the age of the subjects, and explain the differing findings of that study and ours.
Most similar to our study in terms of design as well as investigated population group was the study conducted by Petridou et al. [17]. In that study, subjects cycled for 45 min at moderate intensity (62% of predicted maximal heart rate) just before ingestion of a meal with a fat content of 0.65 g per kg BM and 35% of total energy. They reported similar results to those in our study, finding also only a modest (-17%) and statistically insignificant effect of the exercise bout on TAG dAUC.
Although postprandial lipemia was not lowered significantly with the exercise session used in this study setting, health benefits due to the exercise bout are nevertheless expected via improved insulin sensitivity, increased energy expenditure, and fat oxidation. Gill et al. [27] stated that in non-diabetic subjects, the insulin response to dynamic metabolic stress is likely to provide the best indirect measure of insulin sensitivity. In this study, postprandial insulinemia tended to decrease the longer the exercise session lasted, with no difference in glycemia, indicating that tissue sensitivity to insulin might have increased [28]. Furthermore, physical activity accumulating an energy expenditure of 4200 kJ per week is associated with a 30% reduction of all-cause mortality rates [29]. Our subjects could reach this weekly amount of energy expenditure by performing, for example, five walking sessions of 30 min, one session of 60 min plus three of 30 min, or one session of 60 min plus one of 90 min. Fat oxidation in these three situations would be 18 g, 23 g, and 32 g per week, respectively, indicating that greater fat oxidation can be obtained when performing fewer but longer activity sessions. Interestingly, five sessions of 30 min equal the current minimum recommendation of physical activity [15]. None of our subjects had difficulties in accomplishing the exercise bouts and, according to the rating of the Borg scale, they perceived the exercise intensity as light. Thus, this kind of physical activity is absolutely suitable to include in daily living.
We conclude that 60 and 90, but not 30 min of moderate intensity walking slightly reduced postprandial lipemia after two mixed meals with moderate fat content in sedentary, healthy young men compared to an inactive control. However, the reduction was not statistically significant. It seems, therefore, that a single exercise session of moderate intensity and with a duration of 90 min or less is insufficient to significantly reduce postprandial lipemia in these subjects in a normal daily life setting, i.e. when normal mixed meals with a moderate fat content are served. However, health benefits can nevertheless be achieved at these exercise intensities through other mechanisms.
Methods
The study was approved by the Ethical Committee of the Swiss Federal Institute of Technology Zurich and was carried out according to a repeated measures cross over design. All participants of the study gave written informed consent.
Subjects
Sixteen healthy, normal-weighed, sedentary men participated in the study (Table 2). They were non-smokers and all normotriacylglycerolemic and euglycemic according to the classification of the NCEP [30] and the WHO [31], respectively. One subject had borderline high values of total cholesterol and low-density lipoprotein (LDL) cholesterol while the others were in the normal range (Table 3).
Table 2 Characteristics of the male subjects (n = 16)
Characteristic Mean SEM
Age [y] 24.8 0.8
Height [m] 1.81 0.02
Body mass [kg] 68.9 1.6
Body mass index [kg· m-2] 21.1 0.5
VO2max [mL· min-1· kg-1 body mass] 41.2 0.8
Maximum heart rate [min-1] 192 2
SEM, standard error of the mean; VO2max, maximal oxygen uptake.
Table 3 Fasting biochemical variables of the subjects (n = 16)
Plasma parameter Mean SEM
Triacylglycerol [mmol· L-1] 1.03 0.09
Total cholesterol [mmol· L-1] 3.99 0.19
HDL cholesterol [mmol· L-1] 1.27 0.06
LDL cholesterol [mmol· L-1] 2.26 0.15
Glucose [mmol· L-1] 5.01 0.08
Insulin [pmol· L-1] 91.3 5.0
Glucagon [pmol· L-1] 20.2 1.2
Glycerol [μmol· L-1] 55.4 3.1
Fatty acids [mmol· L-1] 0.43 0.05
SEM, standard error of the mean; HDL, high-density lipoprotein; LDL, low-density lipoprotein.
Preliminary tests
Each subject performed two preliminary walking tests on a treadmill to determine his fitness capacity. In the first test, which also served to familiarise subjects with the methods of the study, subjects walked twice for approximately 20 min on the treadmill (PULSAR, H-P-COSMOS Sports Medical GmbH, Nussdorf-Traunstein, Germany) at 5 and 6 km· h-1, respectively, and with an increasing treadmill inclination of 3% every 3 min. The relationship between work rate and oxygen uptake was established in this test and the results were used to determine the work rate in the activity trials resulting in 50% of VO2max.
VO2max was determined in the second test. Subjects walked at 6 or 7 km· h-1, and the treadmill inclination was increased by 5% every 3 min until exhaustion. The test was designed to be finished after 10 to 12 min. VO2max was considered to be valid when at least two of the following three criteria were met: 1) respiratory exchange ratio (RER) >1.1; 2) heart rate within 10 beats per min of the predicted maximum (220 beats per min minus age); 3) rating of perceived exertion ≥19 on the Borg 6–20 scale.
Main trials
Each subject undertook four randomized trials (one control trial with no activity and three activity trials) at intervals of at least four days. To ensure similar baseline conditions, subjects were only allowed to follow activities of daily life on the two days prior to the test days. On the day preceding the test days, the intake of alcohol and caffeine-containing drinks was limited to at most 1–2 glasses of beer or wine and 2 cups of coffee, or 4 cups of tea, or 1 litre of a caffeine-containing soft drink. Additionally, subjects received a standardized evening meal (spaghetti, tomato sauce, cheese, and apple sauce) to prepare at home, providing 0.2 g fat per kg BM, 2.3 g carbohydrates per kg BM, 0.4 g protein per kg BM, and 53 kJ per kg BM. The subjects arrived at the laboratory by public transport after an overnight fast of at least 12 h. The compliance with the instructions was checked with questionnaires that were filled out by the subjects just after arrival at the laboratory.
A catheter was placed into an antecubital or forearm vein 20 min after arrival and a fasting blood sample (8.7 mL) was taken. Subjects then walked on the treadmill for 30, 60, or 90 min at 50% of their VO2max, depending on the respective trial, or rested in the control trial. During walking, respiration and heart rate (Polar Vantage NV, Polar Electro Oy, Kempele, Finland) were measured continuously. The subjects rated their perceived exertion after every 30 min of exercise on the Borg 6–20 scale. The walking session was briefly interrupted after 30 min (in the 60 and 90 min sessions) and after 60 min (in the 90 min session) to flush the catheter with heparin free saline (NaCl 0.9%, B. Braun Medical AG, Emmenbrücke, Switzerland). In the meantime, subjects were allowed to drink water ad libitum. A second fasting blood sample was taken just after the exercise bout. Breakfast was served shortly after, consisting of commercially available cereals, yoghurt, cream, and chocolate drink providing 0.5 g fat per kg BM, 1.6 g carbohydrates per kg BM, 0.4 g protein per kg BM, and 53 kJ per kg BM. Subjects received lunch three hours after breakfast, consisting of commercially available bread, cheese, chocolate cream, cream, and orange juice providing 0.5 g fat per kg BM, 1.6 g carbohydrates per kg BM, 0.5 g protein per kg BM, and 53 kJ per kg BM. Blood samples were taken every hour for six hours after breakfast to assess the postprandial period. The catheter was flushed with heparin free saline every 30 min to keep it patent. Water was available ad libitum during the postprandial period in the first trial, and the ingested amount was replicated in the following trials. The subjects stayed at the laboratory and pursued only seated activities until the end of the test day.
Indirect calorimetry and energy expenditure calculations
Oxygen uptake and carbon dioxide production during physical activity were determined using a pulmonary gas exchange system (Quark b2, Cosmed, Rome, Italy). Energy expenditure, fat, and carbohydrate oxidation during physical activity were calculated according to the Weir formula [32] and a table of the nonprotein respiratory quotient [33], assuming that the urinary nitrogen excretion was negligible.
Blood sampling and analyses
Fasting blood samples were analysed for TAG, insulin, glucagon, total cholesterol, LDL cholesterol, high-density lipoprotein (HDL) cholesterol, FA, glycerol, and glucose. Postprandial blood samples were analysed for TAG, insulin, glucagon, FA, glycerol, and glucose. Venous blood was collected into a 7.5 mL EDTA tube for analyses of TAG, insulin, total cholesterol, LDL cholesterol, HDL cholesterol, glycerol, and FA. A protease-inhibitor (Trasylol, Bayer AG, Leverkusen, Germany) was immediately added to an aliquot of the EDTA blood for analysis of glucagon. An additional 1.2 mL EDTA/Fluoride tube was collected for analysis of glucose. The tubes were stored on ice until centrifugation (3700 rpm, 8 °C, 12 min; g = 1800; Omnifuge 2.0 RS, Heraeus Sepatech, Osterode, Germany). After separation of venous samples, aliquots of serum were stored at -80 °C. Analyses were conducted after conclusion of the entire experimental period.
A baseline fasting blood sample from each subject's first experimental day was analyzed by enzymatic colorimetric methods for TAG, total cholesterol (both Hitachi Modular P system, Roche Diagnostics, Basel, Switzerland), HDL (Cobas Integra800 analyzer, Roche Diagnostics, Basel, Switzerland), and LDL (calculated by the Friedewald formula [34]).
All fasting and postprandial blood samples were analysed for TAG, FA, glycerol, and glucose by enzymatic colorimetric methods, using a centrifugal analyzer (Cobas-Mira, Roche, Basel, Switzerland). Quality control sera (Roche Diagnostics, Basel, Switzerland) were used to ensure accuracy and precision. The TAG values were corrected by subtracting the glycerol values. Insulin and glucagon were analysed in duplicates by radioimmunoassay, using half of the prescribed amount of kit reagents (LINCO Research, St. Charles, Missouri, USA). The centrifugation steps (glucagon: 4200 rpm, 4 °C, 30 min and insulin: 4700 rpm, 4 °C, 30 min) were carried out in a conventional laboratory centrifuge (Varifuge RF, Heraeus Sepatech, Hanau, Germany) and radioactivity was measured using a gamma counter (Cobra II, Packard Gamma Counter, Minnesota, USA).
Data analyses and statistics
The AUC and the dAUC TAG were calculated using the trapezoidal method [35]. The measurement just before meal intake was taken as the base value. Statistical analyses were performed using SAS statistical software (version 8.2 for Windows, SAS institute Inc., Cary, NC, USA). The results of the responses during exercise, AUC, dAUC, and blood measurements at specific time points were compared with a mixed model analysis with Tukey's adjustment. The postprandial responses over time were compared with a mixed model analysis with repeated measurements and Tukey's adjustment. The factors for the analyses were subject, period, and trial. The fasting value before activity was used as an additional factor for the postprandial responses over time. Data are presented as mean and standard error of the mean (SEM). A p-value of less than 0.05 was considered significant.
Authors' contributions
MP was responsible for designing, planning, and accomplishing the study, the analyses of the results, as well as drafting the manuscript. TL participated in the accomplishment of the study and analyses of the blood samples. CW participated in the design and coordination of the study. PCC participated in designing the study, the analyses of the results, and drafting the manuscript.
Acknowledgements
This project was supported by the Swiss Federal Council of Sports (FCS) and the Swiss Foundation for Nutrition Research (SFEFS). The test meals were provided by Coop, Switzerland and bio-familia AG, Switzerland.
==== Refs
Hyson D Rutledge JC Berglund L Postprandial lipemia and cardiovascular disease Curr Atheroscler Rep 2003 5 437 444 14525676
Patsch W Esterbauer H Foger B Patsch JR Postprandial lipemia and coronary risk Curr Atheroscler Rep 2000 2 232 242 11122749
Roche HM Gibney MJ The impact of postprandial lipemia in accelerating atherothrombosis J Cardiovasc Risk 2000 7 317 324 11143761
Burdge GC Calder PC Plasma cytokine response during the postprandial period: a potential causal process in vascular disease? Br J Nutr 2005 93 3 9 15705218 10.1079/BJN20041282
Hu FB Willett WC Optimal diets for prevention of coronary heart disease JAMA 2002 288 2569 2578 12444864 10.1001/jama.288.20.2569
Ziogas GG Thomas TR Harris WS Exercise training, postprandial hypertriglyceridemia, and LDL subfraction distribution Med Sci Sports Exerc 1997 29 986 991 9268954
Merrill JR Holly RG Anderson RL Rifai N King ME DeMeersman R Hyperlipemic response of young trained and untrained men after a high fat meal Arteriosclerosis 1989 9 217 223 2493784
Gill JM Exercise and postprandial lipid metabolism – an analysis of the current evidence Eur J Lipid Sci Technol 2004 106 110 121 10.1002/ejlt.200300906
Gill JM Caslake MJ McAllister C Tsofliou F Ferrell WR Packard CJ Malkova D Effects of short-term detraining on postprandial metabolism, endothelial function, and inflammation in endurance-trained men: dissociation between changes in triglyceride metabolism and endothelial function J Clin Endocrinol Metab 2003 88 4328 4335 12970306 10.1210/jc.2003-030226
Hardman AE Lawrence JE Herd SL Postprandial lipemia in endurance-trained people during a short interruption to training J Appl Physiol 1998 84 1895 1901 9609782
Katsanos CS Moffatt RJ Acute effects of premeal versus postmeal exercise on postprandial hypertriglyceridemia Clin J Sport Med 2004 14 33 39 14712164 10.1097/00042752-200401000-00006
Katsanos CS Grandjean PW Moffatt RJ Effects of low and moderate exercise intensity on postprandial lipemia and postheparin plasma lipoprotein lipase activity in physically active men J Appl Physiol 2004 96 181 188 12949025 10.1152/japplphysiol.00243.2003
Zhang JQ Thomas TR Ball SD Effect of exercise timing on postprandial lipemia and HDL cholesterol subfractions J Appl Physiol 1998 85 1516 1522 9760349
World Health Organisation The world health report 2002 2002 World Health Organisation. Geneva, Switzerland
Pate RR Pratt M Blair SN Haskell WL Macera CA Bouchard C Buchner D Ettinger W Heath GW King AC Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine JAMA 1995 273 402 407 7823386 10.1001/jama.273.5.402
Murphy MH Nevill AM Hardman AE Different patterns of brisk walking are equally effective in decreasing postprandial lipaemia Int J Obes Relat Metab Disord 2000 24 1303 1309 11093292 10.1038/sj.ijo.0801399
Petridou A Gerkos N Kolifa M Nikolaidis MG Simos D Mougios V Effect of exercise performed immediately before a meal of moderate fat content on postprandial lipaemia Br J Nutr 2004 91 683 687 15137919 10.1079/BJN20041097
Dubois C Beaumier G Juhel C Armand M Portugal H Pauli AM Borel P Latge C Lairon D Effects of graded amounts (0–50 g) of dietary fat on postprandial lipemia and lipoproteins in normolipidemic adults Am J Clin Nutr 1998 67 31 38 9440372
Parks EJ Recent findings in the study of postprandial lipemia Curr Atheroscler Rep 2001 3 462 470 11602066
Institute of Medicine Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (macronutrients). A report of the panel on macronutrients, subcommittees on upper reference levels of nutrients and interpretation and uses of dietary reference intakes, and the standing committee on the scientific evaluation of dietary reference intakes 2002 Institute of Medicine of the National Academies. Washington, D.C.
Kolifa M Petridou A Mougios V Effect of prior exercise on lipemia after a meal of moderate fat content Eur J Clin Nutr 2004 58 1327 1335 15054409 10.1038/sj.ejcn.1601968
Fielding BA Callow J Owen RM Samra JS Matthews DR Frayn KN Postprandial lipemia: the origin of an early peak studied by specific dietary fatty acid intake during sequential meals Am J Clin Nutr 1996 63 36 41 8604667
Burdge GC Jones AE Frye SM Goodson L Wootton SA Effect of meal sequence on postprandial lipid, glucose and insulin responses in young men Eur J Clin Nutr 2003 57 1536 1544 14647218 10.1038/sj.ejcn.1601722
Preiss-Landl K Zimmermann R Hammerle G Zechner R Lipoprotein lipase: the regulation of tissue specific expression and its role in lipid and energy metabolism Curr Opin Lipidol 2002 13 471 481 12352010 10.1097/00041433-200210000-00002
Perreault L Lavely JM Kittelson JM Horton TJ Gender differences in lipoprotein lipase activity after acute exercise Obes Res 2004 12 241 249 14981216
Issa JS Diament J Forti N Postprandial lipemia: influence of aging Arq Bras Cardiol 2005 85 15 19 16041449
Gill JM Herd SL Tsetsonis NV Hardman AE Are the reductions in triacylglycerol and insulin levels after exercise related? Clin Sci (Lond) 2002 102 223 231 11834142
Mikines KJ Sonne B Farrell PA Tronier B Galbo H Effect of physical exercise on sensitivity and responsiveness to insulin in humans Am J Physiol 1988 254 E248 E259 3126668
Kesaniemi YK Danforth E JrJensen MD Kopelman PG Lefebvre P Reeder BA Dose-response issues concerning physical activity and health: an evidence-based symposium Med Sci Sports Exerc 2001 33 S351 S358 11427759
National Cholesterol Education Program (NCEP) Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Third report of the national cholesterol education program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III) final report Circulation 2002 106 3143 3421 12485966
Alberti KG Zimmet P Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus 1999 World Health Organisation. Geneva, Switzerland
Weir JB New methods for calculating metabolic rate with special reference to protein metabolism J Physiol 1949 1 9 15394301
Peronnet F Massicotte D Table of nonprotein respiratory quotient: an update Can J Sport Sci 1991 16 23 29 1645211
Friedewald WT Levy RI Fredrickson DS Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge Clin Chem 1972 18 499 502 4337382
Food and Agriculture Organisation, World Health Organisation Carbohydrates in human nutrition: report of a joint FAO/WHO expert consultation Food and Nutrition Paper No:66 Rome:FAO 1998
|
16209707
|
PMC1277845
|
CC BY
|
2021-01-04 16:39:19
|
no
|
Lipids Health Dis. 2005 Oct 6; 4:24
|
utf-8
|
Lipids Health Dis
| 2,005 |
10.1186/1476-511X-4-24
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-511624201710.1186/1475-2875-4-51Case StudyResearch influence on antimalarial drug policy change in Tanzania: case study of replacing chloroquine with sulfadoxine-pyrimethamine as the first-line drug Mubyazi Godfrey M [email protected] Miguel A [email protected] National Institute For Medical Research (NIMR), Department of Health Systems and Policy Research, Centre for Enhancement of Effective Malaria Interventions (CEEMI), P.O Box 9653, Dar es Salaam, United Republic of Tanzania2 Alliance For Health Policy and Systems Research (AHPSR), World Health Organization, Geneva, Switzerland2005 20 10 2005 4 51 51 3 9 2005 20 10 2005 Copyright © 2005 Mubyazi and Gonzalez-Block; licensee BioMed Central Ltd.2005Mubyazi and Gonzalez-Block; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Research is an essential tool in facing the challenges of scaling up interventions and improving access to services. As in many other countries, the translation of research evidence into drug policy action in Tanzania is often constrained by poor communication between researchers and policy decision-makers, individual perceptions or attitudes towards the drug and hesitation by some policy decision-makers to approve change when they anticipate possible undesirable repercussions should the policy change as proposed. Internationally, literature on the role of researchers on national antimalarial drug policy change is limited.
Objectives
To describe the (a) role of researchers in producing evidence that influenced the Tanzanian government replace chloroquine (CQ) with sulfadoxine-pyrimethamine (SP) as the first-line drug and the challenges faced in convincing policy-makers, general practitioners, pharmaceutical industry and the general public on the need for change (b) challenges ahead before a new drug combination treatment policy is introduced in Tanzania.
Methods
In-depth interviews were held with national-level policy-makers, malaria control programme managers, pharmaceutical officers, general medical practitioners, medical research library and publications officers, university academicians, heads of medical research institutions and district and regional medical officers. Additional data were obtained through a review of malaria drug policy documents and participant observations were also done.
Results
In year 2001, the Tanzanian Government officially changed its malaria treatment policy guidelines whereby CQ – the first-line drug for a long time was replaced with SP. This policy decision was supported by research evidence indicating parasite resistance to CQ and clinical CQ treatment failure rates to have reached intolerable levels as compared to SP and amodiaquine (AQ). Research also indicated that since SP was also facing rising resistance trend, the need for a more effective drug was indispensable but for an interim 5–10 year period it was justifiable to recommend SP that was relatively more cost-effective than CQ and AQ. The government launched the policy change considering that studies (ethically approved by the Ministry of Health) on therapeutic efficacy and cost-effectiveness of artemisinin drug combination therapies were underway. Nevertheless, the process of communicating research results and recommendations to policy-making authorities involved critical debates between policy makers and researchers, among the researchers themselves and between the researchers and general practitioners, the speculative media reports on SP side-effects and reservations by the general public concerning the rationale for policy change, when to change, and to which drug of choice.
Conclusion
Changing national drug policy will remain a sensitive issue that cannot be done overnight. However, to ensure that research findings are recognised and the recommendations emanating from such findings are effectively utilized, a systematic involvement of all the key stakeholders (including policy-makers, drug manufacturers, media, practitioners and the general public) at all stages of research is crucial. It also matters how and when research information is communicated to the stakeholders. Professional organizations such as the East African Network on Malaria Treatment have potential to bring together malaria researchers, policy-makers and other stakeholders in the research-to-drug policy change interface.
==== Body
Introduction
Malaria is still the leading cause of morbidity and mortality in sub-Saharan Africa especially in young children and pregnant women [1]. Considering the limited health budgets and the rising cost of medical services, the increasing trends of drug resistance raise critical public health concerns, as this constrains the provision of adequate treatment in countries where the disease is endemic. The increasing evidence on Plasmodium falciparum parasite resistance to chloroquine (CQ) has prompted some countries to revise their treatment guidelines [2]. In the last decade, the immediately considered alternative first-line drug in some southern African countries, such as South Africa, Botswana, Malawi and Kenya, was SP, but now, countries such as Burundi and Rwanda have already opted for artemisinin drug combination therapy while several others are considering to do so [3]. In many countries, the hesitation by ministries of health to make a policy change decision has been over-influenced by economic budget considerations [4,5]. In Tanzania, the critical nature of the decision to switch to a new first-line drug was closely linked with the estimated budget for the policy change of US $100.8 million [6], which represented 3.4% of the GNP.
Located between latitudes 1°S and 12°S and longitudes 30°E and 40°E, the United Republic of Tanzania covers an area of 945,050 km2, including 59,050 km2 of inland waters [7]. Malaria has been the most life-threatening public disease in terms of morbidity and mortality in Tanzania since the colonial era, and today has a critical influence on the poverty cycle. One recent study showed that malaria reduces the national economic growth by 1.3% [8]. The same study revealed that contributing to about 40,000 deaths annually malaria puts nearly 35 million Tanzanian population at risk. Furthermore, records show malaria accounting to a third of all outpatient visits, a third of inpatient admissions and a third of deaths among children under the age of five years admitted to hospitals [9,10]. Also the national annual health-facility-based statistics for the last ten years (1995–2004) indicate malaria as the highest cause of outpatient attendance in people aged 5 years and above and deaths in hospitalized patients of all age groups.
In Tanzania, the capacity to respond to this malaria problem and other health crises is highly constrained by the meagre Government spending of US $4 per capita on health, with a range of 8–10% of total government budget [11]. The large external debt exerts pressure on the Government's limited resources and in its struggle to eradicate poverty. Tanzania is one of the poorest countries with a per capita income of US $170 per year, about 27% of the population spending less than $0.50 per day on overall needs, 48% spending less than $0.65 per day on basic needs while the real annual economic growth-rate was about 4.9% by year 1999 [12].
For the last 45 years or so, CQ has been the first line drug in nearly all malaria endemic SSA countries. This is due to its ready availability in kiosks, shops and drug stores, as well as from formal health facilities, its relatively low cost per dose and its safety [13]. Until the last day of July 2001, CQ was officially the first-line drug for the treatment of uncomplicated malaria in Tanzania, SP being the second-line drug while quinine was the third-line drug for severe or complicated malaria [14]. As in many other tropical countries, the treatment of malaria has ranged from self-medication using traditional medicines to the use of modern pharmaceuticals [15]. In the context of high levels of poverty, exacerbated by diseases and external pressures, it is initially difficult to implement large-scale and sustainable policy changes even if there is available evidence to support them. This has been the case with recent changes in the malaria treatment policy in Tanzania.
Conceptual framework
The paper is part of a series of case-studies undertaken by the Alliance for Health Policy and Systems Research (AHPSR) to identify the production and utilization of research evidence for policy making in developing countries. An attempt has been made to describe an enabling environment for the demand for research evidence and its utilization, including research funding, research priority setting, institutional support and commissioning. Case-studies were prepared to encourage discussion about the processes and mechanisms which affect support for research and its impact and identification of the challenges in setting research priorities, decision-makers' support for research and the benefits gained from the research process and its results.
In the Tanzanian case-study, particular attention was paid to describing the role of diverse mechanisms and actors in the research-to-policy change process. The impact of HPSR was analysed by observing research inputs and decision outputs in specific policy development situations [16]. Research inputs would be studied from the supply side by analysing problems of HPSR dissemination, and from the demand side through an examination of the participation of researchers as part of the policy-making process. The influence of different types of knowledge – from empirical findings in data-driven design situations to broad conceptual policy frameworks was explored.
Research is greatly under-utilized as a tool to guide health policy formulation, improvement and practice, particularly in the challenges of scaling up interventions and improving access to services in developing health systems. There is a general lack of formal interfaces and linkage strategies to ensure that research supports policy development. To improve utilization, it is important to reveal the policy framework through which researchers can or have influenced decisions. This involves looking at how policy-makers and service managers can use or have effectively used the available research information and how researchers can be or have been promoted to interact among themselves and other key stakeholders in the policy arena [16]. In Tanzania and East Africa, a number of policy initiatives and research projects have supported research undertakings, evidence production and utilization. The Tanzania National Health Research Forum (TANHERF) has been the prime priority-setting mechanism since 1999 and has led to the establishment of an institutional framework for mediating the communication of research evidence to policy-makers [17]. The NIMR and TEHIP have functioned as secretariat for the Forum [18,19] (see Figures 1 &2).
Figure 1 Actors in the research-to-policy interaction process in Tanzania.
Figure 2 Time-line of key events in the Research-to-Policy process (summary).
In this paper, a description is given on the process of changing the malaria treatment policy by replacing CQ with SP as the first-line antimalarial drug in Tanzania. Focus is on the process and role of researchers in providing evidence to and their interaction with policy-makers and other stakeholders and the challenges ahead before a government decision to switch from SP to artemisinin drug combination therapy.
Study Methods
Scope of the study and study design
The case study was exploratory in design, undertaken between June, 2001 and November, 2002 to assess the value of existing institutional mechanisms and interfaces in Tanzania with reference to a recent national antimalarial drug policy change. Focus has been on the role of, and methods used by researchers and research institutions in producing and communicating evidence on antimalarial drug resistance situation and cost-effectiveness of alternative drugs to policy-makers and other stakeholders (medical practitioners, local pharmaceutical manufacturers, traders and community representatives). The study was built on the conception that research is greatly under-utilized as a tool to guide the health policy-making process, policy improvement and practice, considering the emerging challenges to scaling up interventions and improving access to services in developing health systems of countries such as Tanzania. Also considering that policy-making is greatly influenced by health professionals, industry and to some extent the public [16] it was imperative to analyse the relationship between research undertakings to produce evidence for explicit policy decision and the actual utilization of evidence by policy decision-makers.
Sampling
The target was to approach people who played one or several of the following roles or by virtue of their designations: research experience on malaria drugs for at least 5 years, members of various committees involved in discussions on drug policy issues at national level, general practitioners, and lecturers in health policy aspects at high learning institutions. In total, 21 officers were interviewed, including 10 senior malaria researchers, two policy-makers, one officer working with the national pharmacy board, four national-level malaria programme managers, five general medical practitioners (of who one was district- and one a regional medical officer), two university teaching hospital lecturers, two medical research publications and documentations officers, two heads of medical research institutions. Of all the respondents, eight were medical doctors with long experience in malaria case management.
Major themes covered in the indepth interviews and document review
Generally interviewees were asked for their opinions about (i) initiatives that were available to produce research evidence to inform national policy-makers about the efficacy and effectiveness of different antimalarial drugs, the actors involved in such research initiatives and the financial, institutional and political environment to support such initiatives in Tanzania. This component looked at what prompted researchers develop interest in antimalarial drug studies (including how the research topic was identified and priorities set), and how were the studies funded and how were the reports from such studies communicated to policy-makers and other potential stakeholders (ii) research reports showing that due to increasing trend in parasite resistance to CQ and its treatment failures compared to other antimalarial drugs the government urgently needed to revise its malaria treatment guidelines (iii) as potential research-to-policy actors, experiences they have had seeing research evidence utilised by national antimalarial drug policy-makers. This component observed the mechanisms involved in bringing together researchers and policy-makers to policy discussion tables on aspects related to malaria treatment using various antimalarial drugs and the challenges faced so far and those ahead. The document review complemented the interviews by observing the local and international institutions involved in the antimalarial research-to-policy interplay since independence, the funding of the research, the output from several studies commissioned by the government, the WHO and bilateral and multilateral organisations supporting health research and development.
Data collection methods
In 2001, WHO Geneva through technical experts of the Alliance for Health Policy and Systems Research (AHPSR), an initiative of the WHO and the GFHR provided a standard study framework highlighting key themes to be covered by all the case studied supported by the AHPSR. The framework was validated for its applicability in the context of the country and topic of the case study in question. It also provided guidance for the analysis of the study results. Based on such a framework, research instruments were designed to allow the collection of information through (i) a review of official reports and other documents (ii) in-depth interviews with malaria researchers, policy-makers, programme managers and health workers and at institutional and national levels and the national pharmacy board officers (iii) participant observation (the listed first author in this paper participated in the study in the costs, effects and cost-effectiveness of changing the first-line antimalarial drug policy in Tanzania, a study which was commissioned by the NMCP and undertaken by researchers from four research institutions as identified later. Additional information was obtained from district and regional medical officers' meetings, participation in health research and policy conferences held in the country as organized by NIMR, TMA, TPHA and the Multilateral Initiative on Malaria.
Data handing and analysis
Most of the respondents suggested to be confronted as many times as it would be found necessary for additional information or clarity needed, so did not see the need for being tape-recorded. Therefore, interviews were recorded through hand written notes and the interviewer was keen in noting down all the important explanations expressed by the individual respondents. With reference to the study themes/questions and as is recommended in case studies [20], the analysis exercise looked at the content, context, actors, process and pattern of the views expressed by different individual respondents. Attempt was made to note the similarities and contrasts and the possible explanations for the contrasting views.
Results
A summary of the content, patterns, context, actors and processes involved in the research-to-policy process has been presented in Table 2. The subsequent sections give a more detailed account of the interface and interplay between and among the actors involved in the process.
Table 2 Factors facilitating and those hindering the production, dissemination and utilisation of research evidence to guide policy formulation process
Factor Barrier Factors Facilitating Factors
Contextual -Poverty: (i) Possibility that the drug proposed to replace the existing one may not be cost-effective given the poverty situation facing the majority of the residents (ii) Resource poor government – meagre health budget, high national debt crisis -UN agencies e.g. WHO and bilateral agencies e.g. DFID, SDC, USAID, DANIDA etc. readiness to assist technically and financially
-Other countries in the Region also changing their national treatment policy
Actors/Institutional -Fear by drug manufacturers and traders mainly when they still have huge stocks of the drug proposed to be replaced
-Perceptions by doctors/clinicians based on their experiences with prescription/use of alternative antimalarial drugs
-Perceptions of some biomedical researchers and national level policy decision makers
Sometimes contrasting/overlapping research evidence about drug resistance and cure rates of various drugs (lack of/delayed consensus)
-Anticipated repercussions about (i) drug's side effects (ii) poor compliance by drug users and sometimes by drug administrators
-Sustainability in government health budget should donors pull out/terminate assistance or when external assistance is not guaranteed -Involvement of key stakeholders in research
-Formation and operation of credible Regional organizations such as EANMAT
-Presence formal interface between researchers and policy-makers i.e. institutional and policy frameworks such as TANHERF and professional associations such as national Drug Task Policy Force, MAT, TPHA, and NMAC
-Strong local and biomedical research capacity supported by Northern Institutions, bilateral organisations such as DANIDA, DFID, SDC, USAID and multilateral agencies such as TDR
Content -Cost of alternative drugs
-Cost of implementing national policy change
-Some studies carried out on too small scale in terms of population sample size and area coverage to justify representation of the national picture
-Delay in reporting/disseminating research evidence
-Delay in policy-makers to make informed decisions based on research evidence and recommendations
-Poor communication of research evidence: some reports being too long, some being too technically/professionally written -Availability of local research evidence on drug resistance
-Detailed research reports (i) e.g. Abdulla et al. [1] on cost-effectiveness analysis of alternative treatment policy options and Research synthesis (ii) e.g. brief reports to feedback policy-makers
Formulation of the need for policy change
The period from mid 1980s marked an increasing interest in antimalarial drug resistance among biomedical researchers in Tanzania. Thus, several small-scale clinical trials were conducted by researchers from the NIMR in collaboration with local and foreign universities, funded by bilateral agencies such as SDC, DANIDA, USAID, DFID and multilateral agencies including WHO, UNDP and UNICEF. Findings from each of such studies gave evidence that CQ was increasingly facing resistance and treatment failures.
Spread of drug resistance and recommended level to prompt policy change
The need for changing the policy came from accumulated evidence and increasing debates on increasing trend of P. falciparum resistance to CQ, that has been observed in different parts of the country [21-26] since the mid-1950s (also see Table 1 & Figure 1), through to 1970s [27] and the last two decades [4,28]. The Minister of Health, Hon. Anna Abdallah, informed the Parliament of the United Republic of Tanzania during the 2002–2003 budgetary session that her Ministry's decision to suspend CQ as the first-line drug was based on sound evidence pointing to the high cure-rate failure of about 60% while the SP cure rate was 85–90% and was more cost-effective than other antimalarials [27]. While WHO recommends policy change to an alternative drug when the treatment failure reached 25%, evidence from different sentinel sites in the country indicated that up to the time of policy change, CQ treatment failure rate [6] had already reached 52% (ranging 28–72%), 9.5% for SP (ranging 6–32%), and other drugs such as amodiaquine (AQ) and quinine was less than 4.6% (ranging 3.5% – 6%). The increasing numbers of malaria morbidity and mortality in the country, from year to year, was associated by local and foreign researchers with the increasing trend of parasite resistance to CQ. Concern was, therefore, raised about the need to review and improve the national malaria treatment policy guidelines. Another important reason for the increased enthusiasm for policy change towards the end of 1990s is that some countries such as Kenya, Botswana, Malawi and South Africa had already revised their national drug policy guidelines whereby SP had replaced CQ as the first-line drug [6].
Table 1 Trends in antimalarial drug resistance in Tanzania, 1950s–1990s
Period Resistance pattern to chloroquine up to 1999
Early 1950s A dose of 2.5 mg/kg is still efficacious (grace period)
Mid-1970s Owing to slow increase in resistance, the therapeutic dose was gradually increased to the maximum safe level of 25 mg/kg (alert period)
Late 1980s Resistance to the maximum dose began to reach levels of significant public health importance (alert period)
Mid-1990s WHO recommended drug policy action if resistance (total treatment failure) reaches 25% (change period)
Late 1990s Tanzania established sentinel sites in nine regions to monitor resistance by standard methods
Source: Ministry of Health, 1999.
Actors in research to policy
WHO Geneva [through TDR]
Between April–May 1996, WHO through TDR organized an inter-country workshop in Tanga Region, Tanzania aimed at training biomedical researchers and discussing the improved protocol for testing the therapeutic efficacy of CQ, SP and other antimalarials. This workshop was organized to support enhancement of research capacity in collecting, documenting and reporting evidence on antimalarial drug resistance as a milestone for guiding national drug policy review process.
EANMAT
Formed in 1997 and supported by DFID, DANIDA and several other agencies, the East African Network on Monitoring Antimalarial Treatment (EANMAT) was established to bring malaria researchers and policy-makers from Ministries of Health in the three East African countries, Kenya, Uganda and Tanzania although Rwanda joined the Network later in 1999. EANMAT's mission is to have a network that would enable the regular monitoring of treatment outcome of the commonly used first and second-line anti-malarial drugs, based on which rational anti-malarial treatment policies would be developed. EANMAT has a secretariat coordinating the activities of member countries according to its constitution and has generated important malaria treatment database that has contributed to the review and modifications of malaria treatment policies in member countries, whereby in Tanzania the data started to be effectively utilized at policy dialogue level in May 1999. EANMAT also provided funds for the second study phase on the efficacy of AQ, Lapdap and other alternative anti-malarial drugs and produced a standard malaria treatment protocol for East African countries for consideration by countries [4].
Sentinel sites for antimalarial drug resistance monitoring
In 1997, the MoH of Tanzania through its NMCP, in collaboration with EANMAT, decided to select nine areas in different parts of the country to act as sentinel sites for monitoring antimalarial drug resistance. These sites were selected on a number of criteria, including areas with different socio-economic characteristics and accessibility to antimalarial drug sources and varying malaria endemicity and drug resistance patterns. This was followed by the NMCP with support from WHO and other donors to commission a number of studies from research institutions like NIMR and IHRDC, university teaching hospitals, and other institutions to monitor drug resistance and report findings to the Government for consideration. From their inception, the sentinel sites have provided an environment for producing evidence based on which the National Task Force on Antimalarial Drug Policy and EANMAT developed a policy-brief summary to feedback to the national policy makers, who were finally convinced of the widespread CQ resistance and clinical treatment failures throughout the country, which had never happened previously.
The National Task Force on Antimalarial Drug Policy
The National Task Force on Antimalarial Drug Policy was formulated in May 1999 as a sub-committee of the National Malaria Advisory Committee (NMAC) in consultation with the EANMAT and WHO country office in Dar Es Salaam. This body is comprised of interdisciplinary professionals, some of them from the MoH acting as key policy decision-makers (including the Director of Preventive Services, NIMR's Director General, Muhimbili National Hospital, the National Pharmacy Board, Integrated Management of Childhood Illnesses (IMCI) and the Medical Stores Department). WHO Country office also has been participating in several activities of the Task Force and the NMAC and in providing the necessary and feasible technical and material assistance.
On 23rd July 1999, the Task Force developed a three-page summary, drawing on evidence from clinical trials in the sentinel sites. This information was supplemented with a review of national health management information (HMIS) records, as a research policy-brief to highlight to national policy-makers the trend in antimalarial drug resistance, the status of malaria-related morbidity and mortality, and provide immediate suggestions for short and long term intervention towards effective and sustainable malaria control in the country [10]. The evidence presented in such a brief document was exactly the same as it appeared in the technical research report by Abdulla et al [6] warning of the increase of resistance to chloroquine. On that ground, it was recommended that SP should be adopted as the first-line drug in the interim period, while efforts to find the most suitable alternative were underway. It was recommended that the decision to change the policy should be interim because of the increasing evidence on high SP resistance in various parts such as Muheza and Kilombero districts [7]. By that time, neighbouring countries like Malawi and Kenya, as well as South Africa and Botswana had already switched to SP as their first-line drug in 1992 and 1996 respectively [4,24], while records on SP resistance in Malawi indicating to have had remained below 10% over the past six years [6].
Multi-centre collaborative study on cost-effectiveness of CQ, SP and AQ
In the same year 1999, shortly before the Task Force presented a summary policy brief paper, a consultancy contract was commissioned by the NMCP and WHO-AFRO to NIMR, IHRDC and LSHTM to undertake a systematic cost-effectiveness analysis of alternative antimalarial drugs (SP, CQ and AQ) and to cost the policy change to an alternative regimen, and finally to inform the NMCP, the research community and national policy-makers. The study projected that within the next 10-year-period, the cost of transforming the new treatment policy from CQ to SP as first-line, would be half that of maintaining CQ as the first-line because the decreasing effectiveness of CQ led to recurring episodes and repeated treatment of patients. It was furthermore revealed that within the said period the cost of using SP instead of CQ would cost around US $0.46 per operational failure averted, or US $33 per death averted, considering changes in outpatient drug cost only. The study further concluded that the cost of adopting AQ would be marginally lower than that of using SP, but due to increasing incidences of side-effects, AQ should be reserved as the second-line, therefore, recommending that SP be introduced as an interim first-line drug anticipating that within the next 10-year period a better drug would be identified [6,25].
TANHERF
Launched in 1999 through consultative strategy facilitated by the Commission For Health Research and Development (COHRED), the Tanzania National Health Research Forum (TANHERF) is composed of partner institutions in health research and their representatives (see Figure 1) and is an inclusive body whose mission is to ensures that each partner has clear defined role, is considered as an asset and shares in the ownership of the mechanism. Its main function is to ensure that evidence based information is correctly utilized by policy-makers and health managers in order to facilitate the provision of better health services to populations. Working more closely with the NIMR where its Secretariat is based, TANHERF is a consultative body to policy and decision-makers defining health research priorities and research undertakings, coordination, collaboration, dissemination of findings and utilization of research results into policy-oriented decision-making. TANHERF receives and approves reports from the Essential National Health Research Coordinating Committee and the National Health Research Ethics Committee and is a custodian for the dissemination of all the national health research results and through its collaborating institutions, it supports health research publications including research on drug resistance and antimalarial treatment options [28]. As some of its members are also members of the previously mentioned drug task force and the NMAC, TANHERF has influenced indirectly the policy change process.
Consensus building
Researchers and members of the Task Force recognized that numerous and sometimes conflicting evidence had been presented by researchers through technical research reports and/or presentations at scientific conferences and publications in peer reviewed journals. The issue of how to reconcile the evidence and deliver a common, simple and clear message to policy-makers was raised. The solution proposed was to organize stakeholder workshops to identify, discuss and synthesize the most pertinent research and health facility-based information concerning malaria treatment and the way forward. The NMCP in liaison with the NMAC, national Drug Policy Task Force and WHO Country office organized malaria workshops and meetings with departmental heads and directors within the MoH and other partner institutions such as the pharmaceutical industry, bednet manufacturing industries, research institutions such as NIMR, IHRDC and academic institutions such as the MUCHS, St. Augustine University Teaching Hospital – Bugando and KCMC, and intervention projects such as Tanzania Essential Health Interventions Project (TEHIP) and the Adult Morbidity and Mortality Project (AMMP). One of the latest workshops held was the one of May 1999 in Bagamoyo in Tanzania supported by the WHO Roll Back Malaria (RBM) Programme. Based on the presentations made by several participants during the workshops, the discussion focused on developing a common way of understanding the trend of drug resistance and the pros and cons of whether to maintain the status quo with CQ as the first-line drug or to switch to alternative drugs like SP, AQ, chlorproguanil-dapsone (Lapdap) or others.
Despite the intense discussion about alternative drug regimens, there was a common appreciation that CQ treatment failure rates at different levels and in different drug monitoring sentinel sites were alarming enough to justify the need for change. The remaining question that was left for Parliament to answer in liaison with the MoH and other partners (e.g. NMCP, NMAC, Drug Policy Task Force) was whether to replace CQ either with SP as a single first-line therapy or in combination with other regimens. RBM, through its representative in East Africa, expressed readiness to support technically and materially where necessary and feasible to facilitate any government policy change decision. As one of the respondents argued, this was a very important message to the country policy-makers because of the budget implications of such a change.
The actual policy change process and current perceptions of the change
After the national Task Force on antimalarial drug policy had presented its policy brief, a series of newspapers and some private radio stations began to inform the public that CQ was no longer a recommended drug for malaria treatment and that the Government was considering replacing it with a new drug. This information caused public concern and debates erupted in different parts of the country about the rationality for the change. Those involved were the general public, the research community, traders, the pharmaceutical industry, and health-care providers in the public and private health facilities. To maintain public confidence, the Minister of Health gave out a press release that indicated the Government stand concerning the treatment guidelines to be followed while strategies were underway to make an appropriate decision. Nobody was assured of the time of the actual policy change and which type of new treatment guideline would be recommended, as a senior health manager remarked while interviewed in this study:
"It was not clear when the policy would have changed, as sometimes information that was coming out in the press releases last year was controversial. There is a time when the Minister for Health seemed to criticize the information that came out in one of the local newspapers that CQ was no longer effective, and advised the public to remain patient until his ministry gets sufficient evidence. On the other hand, researchers continued to disseminate information indicating high levels of resistance and suggesting for finding out a more suitable drug".
Also, according to several respondents, the policy change was not in the interests of the pharmaceutical manufacturers and traders, who had built up large stocks of CQ and had profited much from its familiarity among most of their client populations. Drug supply companies had already invested in small vans to deliver CQ with a banner 'CHLOROQUINE' on their sides.
Medical practitioners and biomedical researchers considered it was too early to change the policy. One of the reasons stated was that many people still believed that CQ was effective despite variations between different areas and that SP resistance was reported to be on the increase. Some high-ranking government officers at parliamentary and ministerial levels identified themselves as being among those who were still using CQ effectively. Detractors also relied on a few incidences of patients who had had side effects with other antimalarial drugs and also pointed paucity of information available regarding SP resistance. This fact is similar to one found in the report by Abdulla et al [6] warning that:
"Anecdotal evidence indicates that many health professionals are unaware of the extent of resistance to CQ and do not perceive an urgent need for change".
It was also expressed that the decision to change the policy was a very sensitive issue considering the financial implications of the change, both to the government and to the users of the drugs on one hand, and the lack of expertise to manage the change and the uncertainty of treatment outcomes in the use of the new drug.
The Government's official announcement of the policy change came out of the media in 2000, although the actual implementation officially started on 1st August 2001. Before this Government policy-decision was passed, approval by the national parliament was sought through the speech by the Minister of Health by then to the Parliament while presenting his ministry's annual budget [6]. Scientific facts played a prominent role, as the Minister pointed out that based on the routine health facility-based morbidity and mortality statistics on malaria and biomedical research evidence, the MoH was convinced that it was high time for the government to replace CQ with a more cost-effective first-line drug that obviously could alter the existing treatment regimen as a whole. The agenda for change was presented based on a summary report by the mentioned Drug Policy Task Force, supplemented by collective information from HMIS records submitted to the MoH, drawn from all regions and policy-research related workshop/conference proceedings, describing the malaria situation.
A similar speech was the one presented by the new Minister of Health, this time Hon. Anna Abdallah during the 2002–2003 Budget Session for her ministry [27] in 2002. With these two speeches, the members of the Parliament (MPs) were convinced of the need for change, albeit the decision to switch would be interim, given that more clinical and cost-effectiveness studies were ongoing in some sentinel sites to establish more evidence regarding the most appropriate treatment option. This was justified by the presence of the 5-year Interdisciplinary Monitoring Programme for Antimalarial Drugs in Tanzania (IMPACT-Tz) whose initiation involved the MoH as a key stakeholder. The IMPACT-Tz project is a collaborative venture between the U.S Centres for Disease Control (CDC), IHRDC, NIMR, LSHTM, AMMP, TEHIP, MUCHS, WHO and district health authorities [27,30].
All the respondents in this study expressed the need for continued research on alternative drug regimens that are cost-effective if used in real situations, and to publish systematically organized policy-briefs to keep policy-makers and national programme managers informed of the malaria treatment situation and its socio-economic consequences. It was clarified that each research project should consider the presentation of concise policy feedback report by arranging to meet directly, where possible, with policy-makers and programme managers to present their findings and recommendations and share discussions with them whereby various questions can be answered on the spot and reaching agreement on the way forward. It was emphasised that the policy reports should be in the simplest possible language rather than using too technical scientific jargons which would discourage most of the non-technical scientific readers among whom are some policy-makers, planners and programme managers.
Discussion
Factors for the delaying or speeding up of policy change process
As this case study shows (Table 2), the process of research-to-policy change is not simple, it is lengthy and tortuous, facing numerous challenges from the initiation of the research to producing evidence for policy-makers and other key stakeholders to be convinced. Thus, research intended to build consensus with policy-makers based on the evidence it has produced would be successful depending on the following:
(i) research evidence communicated clearly and timely to policy-decision makers. It needs to be communicated in a language which can convey the message to policy-making bodies, as many cannot understand biomedical technical jargon: research reports that are too technical and too long face the risk of being discarded by the audience, who are non-professional in the field or who have limited time to read/listen the whole report due to their other official commitments or personal responsibilities;
(ii) consistence of the evidence produced from different studies on the same research topic/theme: as research evidence from different studies may be more or less overlapping or contradicting each other, policy-makers find it difficult to make explicit policy change decisions;
(iii) the scale of the study on which recommendations for policy change are based: policy-makers would be interested in research evidence drawn from a wide range of study populations and evidence drawn from different country sentinel sites are more likely to be acceptable the purpose of a nationwide policy change. Sometimes, evidence from multi-country studies within the region are preferred to small scale studies undertaken in one or several areas in the country. As is evident from the present case-study, EANMAT reports have contributed substantially to convince the Tanzanian policymakers as has been the case with the rest of East African NMCPs;
(iv) the anticipated repercussions of making a policy change decision: such a decision cannot be made in isolation of the standpoint of view of the key stakeholders, especially drug manufacturers and drug traders, as these may need to be given an assurance on issues such as stocks of drug materials or circulation, which have an implication on the profit and loss accounts of their firms;
(v) policy-makers sometimes lack confidence about the reported efficacy and safety of the recommended drugs and may use a few speculative media reports and personally observed treatment outcomes to justify their reasons for delaying a policy change. As Ralston [31] has remarked, research and its synthesis and discussion helped allay the fear of uncertainty and supported action in the face of marked opposition;
(vi) more often, the government making decisions after learning from what other countries especially the neighbouring ones have done or are considering doing: If other countries have not yet changed or are considering a change to their national drug policies in a way which is different from the researchers recommendations within the country, policy decision may take longer time to change than expected by the local evidence producers/researchers;
(vi) the production of research evidence is costly: besides the budget required for the research itself and its dissemination, the cost of implementing a nationwide policy change (e.g. production of new guidelines, training of health workers, supply of newly recommended drug to all health service delivery levels, replacement of the stocks of the previously recommended drug, advertisement/publicity of the new treatment guidelines through the media, etc.) is expensive to the provider's e.g. government's side, besides potential cost associated with health care seeking and drug utilisation by the target members of the public. Unless policy-makers are convinced to reason without doubt about how these potential constraints would be minimised or avoided, no wonder that it will take time for the recommended new treatment policy to be approved.
Potential challenges to the next drug combination therapy change
In response to recommendation by WHO to countries facing classical antimalarials to implement artemisin-derivatives-based combination therapy [32], the Tanzanian Government through the MoH has already explicitly announced its strategy for replacing SP monotherapy with artemisinin-based combination therapy (ACT) with effect from 2006 [33]. Such a policy move faces similar challenges such as reservations/inadequate faith by the research and drug prescribing community who anticipate immediate parasite resistance to ACT mainly due to the majority poor residents potential failure to comply with the treatment schedules because of the drug cost, individual tastes and preferences driven by their perceptions of the drug with the sulpha component and availability of alternative drugs in the liberal retail market (as some people may prefer monotherapy such as AQ or other drugs). The concern about cost related barriers both to the provision and utilisation of artemisinin drug combination therapy related services have been documented by other authors [2,34-37]. There is also failure of the drug users to take the full course/dose of the recommended drug as some of them may prefer the drug taken once and for all rather than the one whose dose has to be completed by taking it several times. Another failure may be on the side of some health practitioners to administer the drug as recommended in the policy guidelines due to lack of training on the administration of the new drug or other motives driving their practices e.g. private-for-profit health care drug administrators may be driven by the preferences of alternative types of drugs by their clients or may prescribe alternative drugs which they access easily and cheaply in the market and which apparently are more profitable.
Conclusions and recommendations
Despite taking longer than expected, the eventual decision by the Tanzanian government to revise its malaria treatment policy in 2000 and to enforce its effective implementation in 2001 marks the explicit policy decision output backed up with research evidence. Also, the commitment by the MoH to support clinical trials e.g. the IMPACT-Tz study and the one by Abdulla et al., can be seen as intermediate explicit policy decision points. The present case study shows that the road towards changing a nationwide drug policy is long and multifaceted. However, the ultimate decision to change the national policy depends also on evidence backed by demand-driven systematic research. As research plays a role in modifying political, technical and social values, the following is suggested: (a) need to involve key stakeholders at all stages of research to ensure it is always demand-driven (b) need to consider how and when research information should be communicated to or between different stakeholders (e.g. inviting policy-makers and the media in policy-brief workshops/meetings), as sometimes done by EANMAT, TEHIP, NIMR, COSTECH and other institutions/organizations involved in the research-to-policy partnership. As stated by Eisenberg [38], EANMAT contributed to globalize the evidence and localize the decisions, thus helping to obtain support for change (c) the imperativeness of appreciating that is unusual for all policy-makers are able to interpret research findings presented in too technical language, so a simplified version of the biomedical language is crucial for the optimal utilisation of the evidence presented (d) need for continued support to build capacity both in research, evidence presentation and in policy analysis in developing health systems (e) The importance of multi-country evidence on drug resistance to influence national policy change decision is mostly inescapable especially in countries located in the same region (e.g. EANMAT or WENMAT countries] as part of globalization, albeit it is up to the individual countries to find out the feasibility of whatever policy change decision move in the local context and localise the decisions, as some observers recommend [36].
Abbreviations
AFRO African Region
DANIDA Danish International Development Cooperation
DFID Department for International Development (UK)
GFHR Global Forum for Health Research
IHRDC Ifakara Health Research and Development Centre
IMPACT-Tz Interdisciplinary Monitoring Programme for Antimalarial Drugs in Tanzania
KCMC Kilimanjaro Christian Medical College
LSHTM London School of Hygiene and Tropical Medicine
MUCHS Muhimbili University College of Health Sciences
NIMR National Institute for Medical Research (Tanzania)
NMCP National Malaria Control Programme
SDC Swiss Development Cooperation
STI Swiss Tropical Institute
UNDP United Nations Development Programme
UNICEF United Nations Children's Fund
USAID United States Agency for International Development
TEHIP Tanzania Essential Health Interventions Project
TDR UNICEF/World Bank/UNDP/WHO Special Programme for Research and Training in Tropical Diseases
TMA Tanzania Medical Association
TPHA Tanzania Public Health Association
WENMAT: West African Network on Monitoring Malaria Treatment
WHO World Health Organization
Authors' contributions
GMM (a GMP PhD candidate) was the principal investigator, participated in all the case study stages from its design through research report writing and developed the first and final drafts of the manuscript. As the Manager of the AHPSR, M.A.G.B substantially provided technical advice on the design of the study protocol, data collection tools and analtytical framework, critically reviewed the case study report and furnishing it and in the writing of the manuscript.
Acknowledgements
Financial support from the AHPSR, although neither the latter is responsible for the personal opinions and other information presented by the authors; useful comments from Prof. Anne Mills of the LSHTM who is the Chairperson of the Board of Directors for the AHPSR; all the interviewees for their valuable information and time; Dr. T.K. Mutabingwa, Chairman of EANMAT and Dr. Martha Lemnge (by then Director of Amani Medical Research Centre, Tanzania an currently the Director of NIMR Bombo Centre) for responding to this study and their extra time and document support; Dr. Indira Pathmanathan and Dr. C.Y.K Yesudian (Tata Institute, Mumbai, India) for some guidance in case study design and report writing; Dr. Don de Savigny (TEHIP-Tanzania and STI, Basel) for interview and support with literature; Mr. Michael Munga for assisting in several interviews; Joseph R. Mwanga for advice on qualitative interview approaches; Dr. Teresa Lander on behalf of the AHPSR for assisting in editing the case study technical report; Dr. Andrew Y. Kitua, NIMR Director General, for commenting on the study report and approving this case study for publication.
==== Refs
UNICEF Malaria: a major cause of child death and poverty in Africa 2004 United Nations Children's Fund (UNICEF), New York
Breman JG Alilio MS Mills A Conquering the intolerable burden of malaria: what is needed: a summary Am J Trop Med Hyg 2004 71 1 15 15331814
WHO Facts on ACTs (Artemisinin-based Combination Therapies): an update on recent progress in policy and access to treatment 2004 World Health Organization
EANMAT Monitoring antimalarial drug resistance within national malaria control programmes: the EANMAT experience Trop Med Int Health 2001 6 1 8 11251892 10.1046/j.1365-3156.2001.00679.x
Mutabingwa T Nzila A Mberu E Nduati E Winstanley P Hills E Watkins W Chlorproguanil-dapsone for treatment of drug resistant falciparum malaria in Tanzania Lancet 2001 358 1218 1223 11675058 10.1016/S0140-6736(01)06344-9
Abdulla S Goodman C Coleman P Mubyazi G Kikumbih N Okorosobo T The Costs, Effects and Cost-Effectiveness of Changing the First-Line Drug for the Treatment of Malaria in Tanzania Health Economics and Financing Programme (HEFP), LSHTM Working Paper No 01/00 produced in February 2000
Mubyazi GM The role of Research in changing the first-line antimalarial drug policy in Tanzania. Working Paper 2003 Alliance For Health Policy and Systems Research (AHPSR), Geneva
Schellenberg JA Mponda H Abdulla S Nathan R Improving child survival in rural Tanzania: Insecticide treated nets for malaria control in the Kilombero Valley. The KINET Project Overview A report submitted by the Swiss Tropical Institute in collaboration with the Ifakara Health Research and Development Centre to the Ministry of Health of Tanzania and the Swiss Development Co-operation 2000
Kitua AY Antimalarial drug policy in Tanzania: making a systematic change Lancent 1999 354 2
Ministry of Health – Tanzania Summary report of the Task Force on Antimalarial Drug Policy, Tanzania Ministry of Health, Dar Es Salaam 1999, July 23
Ministry of Health – Tanzania & WHO-Country Office, Tanzania Plan of action for accelerated implementation of malaria control in Tanzania Unpublished Report, January 1999 & Revised in February 1999
Ministry of Health – Tanzania National guidelines for malaria diagnosis and treatment: Malaria Control Series 2000 Ministry of Health, Dar Es Salaam
Kitua AY Antimalarial drug policy in Tanzania: making a systematic change Lancent 1999 354 32
Mubyazi GM The role of Research in changing the first-line antimalarial drug policy in Tanzania. Working Paper 2003 Alliance For Health Policy and Systems Research (AHPSR), Geneva
Mubyazi GM Public perceptions of malaria and their utilization of modern and traditional medicines in Korogwe district, Tanzania Tanzania Health Research Bulletin 2005 7 88 93
Hanney SR Gonzalez-Block M Kogan M The utilisation of health research in policy-making: concepts, examples and methods of assessment A report of the Research Policy and Cooperation Department 2002 World Health Organization, Geneva
TANHERF-Tanzania Tanzania National Health Research Forum, Priority Setting Workshop 15–21 February 1999, Arusha International Conference Centre Final Report 1999 ENHR Secretariat, NIMR Dar Es Salaam
TEHIP Completing the circle: bringing research into development: antimalarial combination therapy A Newsletter of the IDRC/MoH/Tanzania Essential Health Interventions Project 2000 15
TEHIP On evidence-based planning TEHIP Newsletter 1999 1
Yin RK Case study research: Designs and Methods 2004 2 Sage Publications: Thousand Oaks, CA. London
Clyde D Observation on monthly pyrimethamine (Daraprim) prophylaxis in an East African village East African Medical Journal 1954 3 41 46 13151067
Clyde D Shute G Resistance of Plasmodium falciparum in Tanganyika to pyrimethamine administered at weekly intervals Trans R Soc Trop Med Hyg 1957 51 505 513 13496184 10.1016/0035-9203(57)90039-1
Antimalarial Drug Task Force-Tanzania Drug policy: implementation of new treatment in Tanzania: the rationale for change to guide policy implementation Tanzania Health Research Bulletin 2000 2 10 13
Mutabingwa TK Maxwell CA Sia IG Msuya FH Mkongewa S Vannithone S Curtis J Curtis F A trial of proguanil-dapsone in comparison with sulfadoxine pyrimethamine for the clearance of Plasmodium falciparum infections in Tanzania Trans R Soc Trop Med Hyg 2001 95 433 438 11579891 10.1016/S0035-9203(01)90207-X
Goodman C Abdulla S Coleman P Mubyazi G Would the replacement of chloroquine with SP be cost-effective in Tanzania? Malaria Consortium 2000
Goodman C Abdulla S Coleman P Mubyazi G Kikumbih N Okorosobo T Mills A Roberts J Choosing the first line drug for malaria treatment-how can cost-effectiveness analysis inform policy? Economics of Infection, Chapter 2 Open University Press, London, UK
TEHIP Ministry of Health clarifies changes on malaria treatment TEHIP Newsletter 2003 13
TANHERF-Tanzania Tanzania National Health Research Forum, Priority Setting Workshop 15–21 February 1999, Arusha International Conference Centre Final Report 1999 ENHR Secretariat, NIMR Dar Es Salaam
Rønn A Msangeni H Mhina J High level of resistance of Plasmodium falciparum to salfadoxine-pyrimethamine in children in Tanzania Trans R Soc Trop Med Hyg 1996 54 197 181
TEHIP Completing the circle: bringing research into development: Antimalarial combination therapy TEHIP Newsletter 15
Ralston S Voltair's Bastards The dictatorship of reason in the West 1992 New York, the Free Press 584 585
Medecins sans Frontieres Act Now – Malaria Report Médecins sans Frontières (MSF) (April 2003: ) and updated with detailed country cases and protocols for ACT treatment at
MoH-Tanzania Diagnosis and treatment of malaria: Orientation Guide for District Trainers. Malaria Control Series 10 National Malaria Control Programme 2005
Wyss K Malaria Update A commissioned discussion paper established in the context of the mandate t751-14 of the Swiss Development Cooperation 2002
Bloland PB Combination therapy for malaria in Africa: hype or hope? Bull World Health Organ 2000 78 1378 1388 11196485
Hanson K Goodman C Lines J Meek S Bradley D Mills A The economics of malaria control interventions Global Forum for Health Research 2004 WHO, Geneva
Greenwood B Malaria first, roll-back expectations Bull World Health Organ 2000 78 1451 1453
Eisenberg JM Globalize the evidence, localize the decision. Evidence-based evidence and international diversity Health affairs 2002 2 167
|
16242017
|
PMC1277846
|
CC BY
|
2021-01-04 16:37:30
|
no
|
Malar J. 2005 Oct 20; 4:51
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-51
|
oa_comm
|
==== Front
Nutr JNutrition Journal1475-2891BioMed Central London 1475-2891-4-261620217510.1186/1475-2891-4-26ResearchValidity of a self-administered food frequency questionnaire (FFQ) and its generalizability to the estimation of dietary folate intake in Japan Ishihara Junko [email protected] Seiichiro [email protected] Hiroyasu [email protected] Manami [email protected] Shoichiro [email protected] JPHC FFQ Validation Study Group 1 Epidemiology and Prevention Division, Research Center for Cancer Prevention and Screening, National Cancer Center, 5-1-1 Tsukiji, Chuo-ku, Tokyo, 104-0051 Japan2 Department of Public Health Medicine, Majors of Medical Sciences, Graduated School of Comprehensive Medical Sciences, University of Tsukuba, Ibaraki, Japan3 Statistics and Cancer Control Division, Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo, Japan4 Public Health, Department of Social and Environmental Medicine, Graduate School of Medicine, Osaka University, Osaka, Japan2005 5 10 2005 4 26 26 12 7 2005 5 10 2005 Copyright © 2005 Ishihara et al; licensee BioMed Central Ltd.2005Ishihara 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 an epidemiological study, it is essential to test the validity of the food frequency questionnaire (FFQ) for its ability to estimate dietary intake. The objectives of our study were to 1) validate a FFQ for estimating folate intake, and to identify the foods that contribute to inter-individual variation of folate intake in the Japanese population.
Methods
Validity of the FFQ was evaluated using 28-day weighed dietary records (DRs) as gold standard in the two groups independently. In the group for which the FFQ was developed, validity was evaluated by Spearman's correlation coefficients (CCs), and linear regression analysis was used to identify foods with large inter-individual variation. The cumulative mean intake of these foods was compared with total intake estimated by the DR. The external validity of the FFQ and intake from foods on the same list were evaluated in the other group to verify generalizability. Subjects were a subsample from the Japan Public Health Center-based prospective Study who volunteered to participate in the FFQ validation study.
Results
CCs for the internal validity of the FFQ were 0.49 for men and 0.29 and women, while CCs for external validity were 0.33 for men and 0.42 for women. CCs for cumulative folate intake from 33 foods selected by regression analysis were also applicable to an external population.
Conclusion
Our FFQ was valid for and generalizable to the estimation of folate intake. Foods identified as predictors of inter-individual variation in folate intake were also generalizable in Japanese populations. The FFQ with 138 foods was valid for the estimation of folate intake, while that with 33 foods might be useful for estimating inter-individual variation and ranking of individual folate intake.
folateFFQinternal validityexternal validityinter-individual variation
==== Body
Introduction
Owing to their ease of administration and low burden on the subject, the assessment of dietary intake in epidemiological studies is often assessed by means of food frequency questionnaires (FFQs) [1]. The chief limitation of FFQs, however, is that serious errors can occur if foods governing inter-individual differences in intake of certain nutrients are omitted from the list. Folate is particularly prone to such error, because it is derived from a variety of foods of both animal and plant origin, not all of which can be included in an FFQ. Foods that contribute to inter-individual differences may fail to be included in food lists, potentially confounding estimation of folate intake by the FFQ.
Another implication of inter-individual variation in the intake of folate or any other nutrient is that such variation may be the underlying determinant of associations between food intake and disease. The preventive effect of certain foods on specific diseases is usually the effect of a particular nutrient contained in the food. Such associations are more likely to be detected when inter-individual differences in intake of the nutrient are larger. Conversely, even if a food contains high levels of a particular nutrient, association will be weak if consumption among individuals is similar. The identification of foods that contribute to inter-individual variation in nutrient intake among the population is therefore an important component of any investigation of nutrients responsible for associations between food intake and disease.
According to the National Nutrition Examination Survey of the United States Population [2] and the National Nutrition Survey in Japan [3], folate intake between the countries is similar among both middle-aged and older age groups. Intake among younger groups, however, is much lower in Japan. This trend is presumably due to differences in the foods that contribute to folate intake. Specifically, folate-fortified food and supplements are large contributors to folate intake for individuals in some Western countries, and questionnaires especially designed to assess folate intake include these items [4,5]. Because folate-fortified foods are not available in Japan and supplement consumption is low, however, folate intake is almost exclusively from natural sources. Nevertheless, little is known about those foods that are the main sources of folate intake among Japanese.
A FFQ was developed and validated for the estimation of dietary intake for the JPHC study. Spearman's correlation coefficients (CCs) between serum folate and folate intake estimated by this FFQ using a biomarker as reference was 0.26 in men [6]. In the present report, we evaluated the validity of this FFQ in the population subgroup for which the FFQ was originally developed using dietary records (DRs) as standard. We then attempted to identify foods that most contributed to folate intake, and those responsible for the differences in intake between individuals. We subsequently repeated these analyses in a second subgroup that was independent of the population for which the FFQ was originally developed, i.e., an external population, to assess its generalizability in Japan. The objectives of this study were: 1) to validate this self-administered FFQ as a means of estimating folate intake, and identify foods that contribute to individual intake and inter-individual variation in folate intake in the population for which the FFQ was developed; and 2) to determine the validity of this FFQ, designed for a study cohort, in estimating the intake of folate and identifying foods that predict inter-individual variation in folate in the general Japanese population.
Materials and methods
Study Subjects
Two validation studies of the FFQ were conducted in a subsample of participants in the JPHC Study, a large population-based prospective study that involved medical examination of the study participants. The target population of the JPHC Study consisted of two cohorts, the first started in 1990 (Cohort I) and the second in 1993 (Cohort II). The aim of the cohort study was to investigate the association between various lifestyle factors such as diet and chronic diseases. Despite Japan's small geographical area, considerable regional variation in diet is seen, and study sites were selected to be representative of the whole country. Cohort I was drawn mainly from the northeast part of the country and Cohort II from the southwest (Figure 1). The study design and participants in the entire cohort have been described previously [7].
Figure 1 Study sites of the JPHC Study.
The present FFQ validation study was conducted in subsamples of Cohort I and Cohort II. The Cohort I study was initiated in February 1995 to validate the FFQ for use in a 5-year follow-up survey [8], given that it was originally developed based on data from 3-day weighed DRs in a random sample from this Cohort [9], while the Cohort II study was to evaluate the generalizability of the FFQ independent of the population for which it was originally developed [10]. Respective numbers and recruitment areas were 247 volunteers from the Ninohe, Yokote, Saku and Chubu (previously named Ishikawa) public health center areas, and 392 volunteers from Mito, Kashiwazaki, Chuo-higashi, Kamigoto, Miyako and Suita. For the present report, we analyzed the data of the 215 and 350 subjects in Cohorts I and II, respectively, for whom the 28-day DR and FFQ data were complete.
Data Collection
Data collection has been described in detail elsewhere [8,10]. In brief, each subject completed two FFQs and 28-day DRs (Figure 2). The first FFQ was administered to provide data to compare with the second FFQ as a means of evaluating reproducibility, and the second FFQ was administered to obtain data to compare with the DRs to evaluate its validity. Only data from the second FFQ for validity has been used in this paper.
Figure 2 Sequence of data collection for the JPHC FFQ Validation Study.
DRs were collected over 7 consecutive days in each of the 4 seasons, except in Chubu (2 seasons). Local dietitians instructed the subjects to weigh all foods and beverages with the scales and measuring utensils provided, and to record results in a specially designed booklet. The subjects in Cohort I, however, were instructed to use standardized portion sizes for some foods that were difficult to weigh (semi-weighed DRs). The subjects described each food, method of preparation, and the names of the dishes in detail. They also reported all dietary supplements used, if any. At the end of each season, the DRs were reviewed in a standardized manner, and each food was coded by local dietitians.
The self-administered semi-quantitative FFQ consisted of 138 food items and 14 supplementary questions on dietary habits and use of supplements. The validity of the questionnaire in regard to the intake of energy, other nutrients and foods, as well as the use of dietary supplements is described elsewhere [10-13].
Dietary intakes of folate according to the DRs and the FFQ were calculated using the Standardized Tables of Food Composition, 5th ed. [14]. Mean daily intake of folate for the 28 days (14 days in Chubu) was calculated based on the DRs of each subject. Because none of the subjects used folic acid supplementation, dietary supplements were not included in the calculation.
Statistical analysis
Means and standard deviation of folate intakes from the DRs and FFQ were calculated by sex for Cohort I and Cohort II. Spearman's rank CCs were calculated for crude intake and energy-adjusted intake by the residual method, and were corrected for the attenuating effect of random intra-individual error (deattenuation) in Cohort I subjects to evaluate internal validity in the population for which the FFQ was developed. Deattenuation was done using the following formula: Deattenuated , where r is the observed correlation, λx is the ratio of intra- to inter-subject variation, and nx is number of dietary records for each subject [15]. The same analysis was performed for Cohort II subjects to evaluate external validity
The percentage contribution of each food to total folate intake was computed based on the DRs of Cohort I subjects, and the 20 foods contributing most were listed based on their percentage contribution. Percentage contributions of the same 20 foods in the DRs of Cohort II were calculated and their actual rank of percentage contribution in Cohort II was determined.
Linear regression analysis with stepwise selection was used to identify foods that contributed to inter-individual variation, with folate intake from each food item according to the FFQ of Cohort I used as the explanatory variable, and total folate intake according to the DRs as the response variable. A model (partial) R-square value for the selected food items was computed. Cumulative mean intake from each food item on the list was calculated, and compared to total intake according to the DRs by Spearman's CCs to evaluate validity. Cumulative intake from the same food items in the Cohort II subjects and their CCs from the DR data were calculated to evaluate external validity.
Results
Daily folate intake as assessed by DRs and FFQ as well as the Spearman's rank CCs between the two measurements by cohort group and sex are shown in Table 1. Mean daily intake of folate based on the FFQ in Cohort I was significantly overestimated compared to the DR data. Both crude and adjusted CCs were higher in men than women in Cohort I but were similar in Cohort II (Table 1).
Table 1 Folate intake (μg/day) assessed by the dietary records and food frequency questionnaire, and their correlation coefficients.
DR1 FFQ2 Spearman correlation
Mean ± SD Median Range Mean ± SD Median Range % difference of mean Crude Energy-adjusted3 Deattenuated
Cohort I
Men (n = 102) 425 ± 103 427 210–735 473 ± 231 444 119–1807 11 0.49 0.40 0.57
Women (n = 113) 389 ± 106 380 153–667 476 ± 287 419 146–1978 22 0.29 0.35 0.47
Cohort II
Men (n = 174) 467 ± 156 443 197–1280 421 ± 190 370 85–1178 -10 0.33 0.50 0.63
Women (n = 176) 426 ± 112 417 198–980 454 ± 237 397 4–1498 7 0.42 0.48 0.63
1 DR, dietary records
2 FFQ, food frequency questionnaire.
3 Folate was adjusted for total energy intake using the residual method.
The 20 foods that made the greatest contribution to total folate according to the DR in Cohort I are listed in Table 2. The list consists of various foods, mainly from plant sources such as vegetables, with spinach making the highest contribution followed by rice, green tea, cabbage, eggs and beer. These 20 foods contributed 55.2% of total intake in men and 52.9% in women. The contribution of the same 20 foods to folate intake in Cohort II subjects was 44.9% for men and 43.2% for women. Gyokuro, a type of green tea, made the second highest contribution in Cohort II, but was not among the 20 in Cohort I. Other food items with the highest contribution in Cohort II but not Cohort I were kamairi-cha (pan-fried green tea), bread, tomatoes, and pumpkins in both sexes; purple laver in men; and sweet potatoes and komatsuna (a green leafy vegetable) in women. When both kinds of green tea were excluded, however, these foods accounted for only 4.6% of total intake in men and 6.2% in women.
Table 2 Foods that contribute to folate intake and their cumulative percentage contribution to total intake as assessed by dietary records.
Men Cohort I Cohort II Women Cohort I Cohort II
Food item Rank1 % Rank2 % Food items Rank1 % Rank2 %
Spinach, leaves 1 6.5 3 5.4 Spinach, leaves 1 7.1 3 5.6
Well-milled rice 2 6.4 4 5.0 Green tea, sencha, infusion 2 6.1 1 10.1
Green tea, sencha, infusion 3 5.6 1 9.1 Cabbage, head 3 4.7 5 3.4
Cabbage, head 4 4.9 5 3.6 Well-milled rice 4 4.5 4 3.6
Chicken eggs, whole 5 3.8 6 3.6 Chicken eggs, whole 5 3.5 6 3.1
Beer 6 3.3 7 3.0 Aspargus, shoots 6 2.8 36 0.7
Radish, root with skin 7 2.6 8 2.5 Chiken offal, liver 7 2.5 15 1.2
Pork offal, liver 8 2.4 35 0.7 Radish, root with skin 8 2.5 7 2.4
Aspargus, shoots 9 2.3 34 0.7 Bracken, young shoots 9 1.9 72 0.3
Natto, itohiki-natto (whole fermented using Bacillus natto) 10 2.0 13 1.4 Natto, Itohiki-natto (Whole fermented using Bacillus natto) 10 1.9 10 1.7
Miso, rice- koji miso, dark yellow type 11 1.9 46 0.5 Broccoli, florets 11 1.8 8 2.0
Chiken offal, liver 12 1.8 12 1.4 Miso, rice- koji miso, dark yellow type 12 1.8 55 0.4
Chinese cabbage, head 13 1.8 11 1.7 Chinese cabbage, head 13 1.7 11 1.6
Bracken, young shoots 14 1.6 67 0.3 Ordinary liquid milk 14 1.5 20 1.1
Broccoli, florets 15 1.6 9 1.8 Shoyu: soy sauce, Koikuchi-shoyu (Common type) 15 1.5 23 1.0
Shoyu: soy sauce, Koikuchi-shoyu (Common type) 16 1.6 17 1.1 Head letucce, crisp type, head 16 1.5 22 1.1
Carrots, European type, root with skin 17 1.4 16 1.1 Purple laver, toasted 17 1.5 18 1.1
Potatoes, tuber 18 1.3 23 0.9 Pork offal, liver 18 1.4 34 0.7
Head lettuce, crisp type, head 19 1.3 22 1.0 Carrots, European type, root with skin 19 1.4 17 1.1
Leaf mustard, leaves 20 1.2 75 0.2 Potatoes, tuber 20 1.3 26 0.9
Total 55.2 44.9 52.9 43.1
1 Food items are listed in order of % contribution to total folate intake based on dietary records of Cohort I (by rank in Cohort I).
2 Rank of % contribution to total folate intake determined on the basis of the dietary records of Cohort II.
The foods that best predicted inter-individual variation in dietary folate and the validity (correlation coefficients) of folate intake based on those foods are shown in Table 3. A total of 33 foods are listed with the cumulative R-square value of 0.59. The food that best predicted variation of intake was green tea, which contributed 12–15% of total intake. No other food predictive of variation contributed more than 1% of total folate intake. The cumulative folate intake from the 33 foods contributed approximately 30% of total intake in both men and women. The CC of intake from the 33 foods in the internal population (Cohort I) was 0.46 in men and 0.28 in women (indicated as "internal" in Table 3), and had approximately the same level of validity as the data from the full 138-food FFQ.
Table 3 Foods most predictive of inter-individual variation in dietary folate, and their correlation coefficients with intake from DR.
Foods selected by regression analysis1 Male and female Male Female
Partial Cumulative Cohort I (internal) Cohort II (external) Cohort I (internal) Cohort II (external)
R-Square1 R-Square1 Cumulative Spearman3 Cumulative Spearman3 Cumulative Spearman3 Cumulative Spearman3
mean intake correlation mean intake correlation mean intake correlation mean intake correlation
μg/day (%)2 μg/day (%)2 μg/day (%)2 μg/day (%)2
Green tea (sencha) 0.096 0.096 72 (15.3) 0.31 55 (13.0) 0.28 55 (11.6) 0.17 66 (14.6) 0.25
Dried small fish 0.070 0.166 73 (15.4) 0.33 55 (13.1) 0.27 56 (11.7) 0.19 67 (14.7) 0.26
Horse mackerel, sardine 0.039 0.204 74 (15.7) 0.37 57 (13.5) 0.27 57 (12.0) 0.20 68 (15.0) 0.24
Cake 0.031 0.236 74 (15.7) 0.37 57 (13.6) 0.27 58 (12.1) 0.19 69 (15.1) 0.24
Miso soup 0.030 0.266 84 (17.8) 0.37 62 (14.8) 0.28 66 (13.8) 0.24 73 (16.1) 0.27
Luncheon Meat 0.032 0.298 84 (17.8) 0.38 62 (14.8) 0.28 66 (13.8) 0.24 73 (16.1) 0.27
Ham, loin 0.022 0.320 84 (17.8) 0.38 62 (14.8) 0.29 66 (13.8) 0.24 73 (16.1) 0.27
Cream for coffee 0.014 0.335 85 (17.9) 0.37 63 (14.9) 0.28 66 (13.9) 0.24 74 (16.2) 0.27
Stewed pork, Western style 0.012 0.346 85 (17.9) 0.37 63 (15.0) 0.28 66 (14.0) 0.24 74 (16.2) 0.27
Mayonnaise 0.011 0.357 85 (17.9) 0.37 63 (15.0) 0.28 66 (14.0) 0.24 74 (16.2) 0.27
Worcester sauce 0.013 0.370 85 (17.9) 0.37 63 (15.0) 0.28 66 (14.0) 0.24 74 (16.2) 0.27
Kamaboko (fish paste product) 0.014 0.384 85 (18.0) 0.37 63 (15.0) 0.28 67 (14.0) 0.24 74 (16.3) 0.27
Lettuce 0.011 0.395 86 (18.3) 0.37 64 (15.3) 0.30 68 (14.3) 0.25 75 (16.5) 0.27
Bean sprouts 0.015 0.410 88 (18.7) 0.37 66 (15.7) 0.29 71 (14.9) 0.23 77 (16.9) 0.26
Peaches 0.017 0.427 89 (18.7) 0.38 66 (15.8) 0.29 71 (14.9) 0.23 77 (17.0) 0.26
Sausage, Wieners 0.011 0.437 89 (18.7) 0.38 66 (15.8) 0.29 71 (14.9) 0.23 77 (17.0) 0.26
Chocolate 0.009 0.446 89 (18.8) 0.38 67 (15.9) 0.29 71 (15.0) 0.23 78 (17.1) 0.26
Octopus 0.008 0.455 89 (18.8) 0.38 67 (15.9) 0.29 71 (15.0) 0.23 78 (17.1) 0.26
Salted fish 0.008 0.463 90 (19.1) 0.38 68 (16.0) 0.30 73 (15.2) 0.23 78 (17.2) 0.26
Apples 0.010 0.473 92 (19.5) 0.38 69 (16.3) 0.31 75 (15.8) 0.24 80 (17.6) 0.26
Sweet pepper 0.012 0.485 94 (19.8) 0.39 70 (16.6) 0.31 77 (16.1) 0.26 81 (17.9) 0.27
Udon 0.011 0.496 95 (20.0) 0.39 71 (17.0) 0.30 78 (16.3) 0.25 82 (18.1) 0.27
Grilled chicken 0.014 0.510 95 (20.1) 0.39 72 (17.1) 0.30 78 (16.4) 0.25 83 (18.2) 0.27
Pickled plums 0.009 0.519 95 (20.1) 0.39 72 (17.1) 0.30 78 (16.4) 0.25 83 (18.2) 0.27
Papaya 0.009 0.528 96 (20.2) 0.39 73 (17.3) 0.29 79 (16.5) 0.25 83 (18.4) 0.27
Black tea 0.007 0.536 96 (20.4) 0.39 73 (17.4) 0.30 79 (16.7) 0.24 84 (18.5) 0.28
Green tea (bancha, genmaicha) 0.007 0.542 105 (22.1) 0.42 83 (19.8) 0.33 88 (18.6) 0.25 95 (21.0) 0.35
Chicken liver 0.007 0.550 122 (25.8) 0.45 98 (23.2) 0.32 101 (21.2) 0.28 107 (23.6) 0.34
Bananas 0.008 0.557 126 (26.6) 0.46 101 (24.1) 0.32 104 (21.9) 0.27 111 (24.3) 0.34
Rice mixed with other grains 0.005 0.562 128 (27.1) 0.45 103 (24.5) 0.33 107 (22.5) 0.27 112 (24.8) 0.34
Well-milled rice 0.011 0.573 152 (32.2) 0.49 124 (29.5) 0.30 125 (26.2) 0.29 128 (28.3) 0.34
Yushi-dofu 0.006 0.579 153 (32.3) 0.49 125 (29.6) 0.30 125 (26.3) 0.29 129 (28.4) 0.34
Bitter gourds 0.011 0.590 158 (33.3) 0.46 128 (30.5) 0.30 130 (27.2) 0.28 132 (29.1) 0.35
1 Foods were selected by stepwise regression analysis using data from the food frequency questionnaire of Cohort I men and women. Partial and cumulative R-Square values were calculated in the process of performing the regression analysis.
2 Percent of total folate according to the food frequency questionnaire.
3 Spearman's correlation coefficients between cumulative intake and total intake based on dietary records
When the same 33-food list was used to compute intake in the external population (Cohort II), the cumulative folate intake contribution was again 30% (indicated as "external" in Table 3), and CC was 0.30 in men and 0.35 in women, showing approximately the same level of validity as in the internal population.
Discussion
In this study, we evaluated the validity of a FFQ as a means of estimating dietary folate intake in the population for which the FFQ was originally developed. We also attempted to identify foods that differentiated the level of folate intake in individuals by stepwise regression, and tested the validity of assessing folate intake based on the intake of these foods. The results were also cross-validated in an independent population to evaluate generalizability.
Validity of the FFQ in estimating folate intake was moderate in both the internal and external populations. In previous studies, the validity of questionnaires in estimating energy-adjusted dietary folate intake varied from 0.2 to 0.6, depending on the study population [15-21]. In studies that reported intake from the diet and from dietary supplements separately, CC was 6–21% higher for folate intake that included supplements than for intake from diet alone. Although none of the subjects used supplements that contained folic acid, which had only just become available at the time of the study, the validity of our FFQ in estimating folate from the diet was relatively high, probably because the dietary folate intake of our subjects was as high as that of supplement users in some of the previously cited studies.
Although the largest proportion of folate intake was from vegetables, these did not necessarily explain the differences in intake between individuals. For example, spinach, which is very rich in folate and was one of the highest contributors to mean folate intake, could not explain inter-individual variation because it was consumed by almost every subject. In contrast, green tea contributed greatly to both individual intake and to inter-individual variation, probably because consumption was strongly dependent on individual preference. Although the analysis provided us with information about foods that predicted inter-individual variation in folate intake, some foods which had a moderate partial R-square value contributed less than 0.1% to total folate intake, such as luncheon meat, ham, and so on. These may have been selected by chance alone. In this kind of analysis, even unimportant contributors to the cumulative R-square value may be statistically significant, but can nevertheless be ignored [1]. In any case, it is noteworthy that individuals could be ranked by folate intake based on only 33 foods, with the same level of validity as with the long FFQ.
Any analysis of the possible effects of folate intake on disease also requires analysis of the effect of food items which contribute to total intake and inter-individual variation of folate. It is of great research interest to determine whether the association between food intake and disease is the result of folate intake. For example, we might hypothesize that folate intake may help explain the protective effect of green tea on gastric cancer described in the recent report of the JPHC study [22]. The mechanism of carcinogenesis through DNA instability and methylation abnormalities as a result of folate deficiency has been studied in animal and in vitro studies [23-25], and an association between folate and gastric cancer has been reported in a number of case-control studies [26-29]. Other prospective studies, however, have failed to show a consistent association between green tea and gastric cancer [30-32]. We speculate that the association was strong in the JPHC population owing to the large contribution of green tea to the variation in folate intake, which was not seen in the other populations.
One of the strengths of our study is the precision of the reference data. The ratio of intra- to inter-individual variation in our data was somewhat higher (1.9–4.8) than in several studies in Western countries [33,34]. When intra-individual variation is larger, an increased number of dietary assessment days is required to obtain a valid standard. Although intra-individual variation among our subjects was high, we had more than a sufficient numbers of days (28 days) of data to represent the true intake of the individuals, because the number of DR days needed to estimate true intake within 20% of the true mean with our data was only about 17 to 19 days according to our analysis. By comparison, the greatest number of days of dietary assessment used as standard in previous studies was 14 days [35].
In addition, our analysis is unique because we used regression analysis to identify the foods most predictive for inter-individual variation in dietary folate, and then evaluated both the internal and external validity of the intake of those foods. It is important to test external validity, because there is no assurance that explanatory variables selected by regression analysis are valid in an external population [1]. To our knowledge, this is the first study to attempt to identify foods that contribute to inter-individual variation of folate intake in Japan, where folate intake is almost exclusively from natural sources. We developed a list of foods that contribute to inter-individual variation in folate intake in the population for which FFQ was originally developed, and tested the generalizability of the results to an external population. The two populations covered various geographic areas throughout Japan. Our results imply the possibility that a shorter questionnaire which specifically targets folate intake in Japanese populations can be developed.
One limitation of our study is that because the subjects needed to be highly motivated to complete the 28-day DRs, they were not a randomly selected sample. Mean folate intake based on DRs was slightly higher than in the entire cohort, probably because the validation study subjects were likely more health conscious and consumed more vegetables. If all subjects had consumed a similar amount of certain foods, inter-individual variation in the food might have been falsely low. The generalizability of questionnaire results needs to be established with care.
Conclusion
Our FFQ is valid for estimating folate intake and is generalizable to the Japanese population. Although some foods such as spinach, rice and cabbage contributed more to total intake of folate, they did not necessarily contribute to inter-individual variation. Validity of the estimation of folate intake based on intake of the 33 foods that most contributed to inter-individual variation was about the same as that based on the original 138-food FFQ. Although folate is contained in a wide variety of foods, 33 foods in the FFQ were sufficient to account for the inter-individual variation of intake and ranking of individuals' intake in Japan. We concluded that the FFQ with 138 foods was valid for the estimation of folate intake, and that the FFQ with 33 foods might be useful for estimating inter-individual variation and the ranking of individual folate intake.
Competing interests
The author(s) declare that they have no competing interests.
List of Abbreviations
FFQ: food frequency questionnaire
DR: dietary record
CC: correlation coefficients
JPHC Study: Japan Public Health Center-based prospective Study
Authors' contributions
JI performed the data analysis and drafted the manuscript
SY participated in the design of the study, coordinated the study and helped with analysis and with the preparation of the manuscript
HI helped with analysis and with the preparation of the manuscript
MI helped with analysis and with the preparation of the manuscript
ST participated in the design of the study and helped to draft the manuscript. He was principal investigator of the JPHC Study
All authors have read and approved the final manuscript.
Acknowledgements
This study was supported by grants-in-aid for Cancer Research and for the Third-Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labor and Welfare of Japan and for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Junko Ishihara is an Awardee of Research Resident Fellowship from the foundation of Promotion of Cancer Research (Japan) for the 3rd Term Comprehensive 10-Year-Strategy for Cancer Control.
The authors wish to express their appreciation to the local staff in each study area, especially to the local dietitians for their efforts in conducting the dietary survey. The investigators in the validation study of the self-administered food frequency questionnaire in the JPHC Study (the JPHC FFQ Validation Study Group) and their affiliations at the time of the study were: S. Tsugane, S. Sasaki, and M. Kobayashi, Epidemiology and Biostatistics Division, National Cancer Center Research Institute East, Kashiwa; T. Sobue, S. Yamamoto, and J. Ishihara, Cancer Information and Epidemiology Division, National Cancer Center Research Institute, Tokyo; M. Akabane, Y. Iitoi, Y. Iwase, and T. Takahashi, Tokyo University of Agriculture, Tokyo; K. Hasegawa, and T. Kawabata, Kagawa Nutrition University, Sakado; Y. Tsubono, Tohoku University, Sendai; H. Iso, Tsukuba University, Tsukuba; S. Karita, Teikyo University, Tokyo; the late M. Yamaguchi, and Y. Matsumura, National Institute of Health and Nutrition, Tokyo.
==== Refs
Willett WC Nutrition Epidemiology 1998 second Oxford, UK, Oxford University Press
Ervin RB Wright JD Wang CY Kennedy-Stephenson J Dietary intake of selected vitamins for the United States Population: 1999-2000 Advance data from vital and health statistics 2004 Hyattsville, Maryland, National Center for Health Statistics 2004 1250
The Study Circle for Health and Nutrition Information Kokumin Eiyo no Genjou (Results of National Nutrition Survey in Japan, 2002, Ministry of Health, Labour and Welfare, Japan) 2004 Tokyo, Japan, Daiichi Shuppan Publishers. (in Japanese)
Pufulete M Emery PW Nelson M Sanders TA Validation of a short food frequency questionnaire to assess folate intake British Journal of Nutrition 2002 87 383 390 12064348 10.1079/BJNBJN2002518
Yen J Zoumas-Morse C Pakiz B Rock CL Folate intake assessment: validation of a new approach J Am Diet Assoc 2003 103 991 1000 12891147
Iso H Moriyama Y Yoshino K Sasaki S Ishihara J Validity of a self-administered food frequency questionnaire used in the 5-year follow-up survey for the JPHC Study to assess folate, vitamin B6, B12 intake: comparison with dietary records and blood level J Epidemiol 2003 13(1 Suppl) S98 101 12701636
Watanabe S Tsugane S Sobue T Konishi M Baba S for the JPHC Study group Study design and organization of JPHC Study J Epidemiol 2001 11 (6 Suppl) S3 S7 11763137
Tsugane S Kobayashi M Sasaki S Tsubono Y M A Validity and reproducibility of a self-administered food frequency questionnaire in the JPHC Study Cohort I: study design, conduct and participant profiles J Epidemiol 2003 13(1 Suppl) S2 S12 12701628
Tsubono Y Takamori S Kobayashi M Takahashi T Iwase Y Iitoi Y Akabane M Yamaguchi M Tsugane S A data-based approach for designing a semiquantitative food frequency questionnaire for a population-based prospective study in Japan J Epidemiol 1996 6 45 53 8795957
Ishihara J Sobue T Yamamoto S Yoshimi I Sasaki S Kobayashi M Takahashi T Iitoi Y Akabane M Tsugane S Validity and reproducibility of a self-administered food frequency questionnaire used in the 5-year follow-up survey of the JPHC Study Cohort II: study design, participant profile and results in comparison with Cohort I J Epidemiol 2003 13(1 Suppl) S134 S147 12701641
Tsugane S Kobayashi M Sasaki S Validity of a self-administered food frequency questionnaire used in the 5-year follow-up survey of the JPHC Study Cohort I: comparison with dietary records for main nutrients J Epidemiol 2003 13(1 Suppl) S51 S56 12701631
Ishihara J Sobue T Yamamoto S Sasaki S Akabane M Tsugane S Validity and reproducibility of a self-administered questionnaire to determine dietary supplement users among Japanese European Journal of Clinical Nutrition 2001 55 360 365 11378809 10.1038/sj.ejcn.1601164
Sasaki S Kobayashi M Tsugane S Validity of a self-administered food frequency questionnaire used in the 5-year follow-up survey of the JPHC Study Cohort I: comparison with dietary records for food groups J Epidemiol 2003 13(1 Suppl) S57 S63 12701632
Science and Technology Agency Standard Tables of Food Composition in Japan, the fifth revised edition. 2002 Tokyo, Printing Bureau, Ministry of Finance
Kabagambe EK Baylin A Allan DA Siles X Spiegelman D Campos H Application of the method of triads to evaluate the performance of food frequency questionnaires and biomarkers as indicators of long-term dietary intake American Journal of Epidemiology 2001 154 1126 1135 11744518 10.1093/aje/154.12.1126
Flagg EW Coates RJ Calle EE Potischman N Thun MJ Validation of the American Cancer Society Cancer Prevention Study II Nutrition Survey Cohort Food Frequency Questionnaire Epidemiology 2000 11 462 468 10874556 10.1097/00001648-200007000-00017
Rimm EB Giovannucci EL Stampfer MJ Colditz GA Litin LB Willett WC Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals American Journal of Epidemiology 1992 135 1114 1126 1632423
Munger R Folsom A Kushi L Kaye S Sellers T Dietary assessment of older Iowa women with a food frequency questionnaire: nutrient intake, reproducibility, and comparison with 24-hour dietary recall interviews Am J Epidemiol 1992 136 192 200 1415141
Feskanich D Marshall J Rimm EB Litin LB Willett WC Simulated validation of a brief food frequency questionnaire Annals of Epidemiology 1994 4 181 187 8055118
Rockett HR Breitenbach M Frazier AL Witschi J Wolf AM Field AE Colditz GA Validation of a youth/adolescent food frequency questionnaire Preventive Medicine 1997 26 808 816 9388792 10.1006/pmed.1997.0200
Messerer M Johansson SE Wolk A The validity of questionnaire-based micronutrient intake estimates is increased by including dietary supplement use in Swedish men Journal of Nutrition 2004 134 1800 1805 15226472
Sasazuki S Inoue M Hanaoka T Yamamoto S Sobue T Tsugane S Green Tea Consumption and Subsequent Risk of Gastric Cancer by Subsite: The JPHC Study Cancer Causes Control 2004 15 483 491 15286468 10.1023/B:CACO.0000036449.68454.42
Steinmetz KA Potter JD Vegetables, fruits and cancer. II. Mechanisms Cancer Causes and Control 1991 2 427 442 1764568 10.1007/BF00054304
Kim Y Pogribny I Basnakian A Miller J Selhub J James S Mason J Folate deficiency in rats induces DNA strand breaks and hypomethylation within the p53 tumor suppressor gene Am J Clin Nutr 1997 65 46 52 8988912
Jacob RA Gretz DM Taylor PC James SJ Pogribny IP Miller BJ Henning SM Swendseid ME Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women 1998 128 1204 1212 9649607
La Vecchia C Ferraroni M D'Avanzo B Decarli A S F Selected micronutrient intake and the risk of gastric cancer Cancer Epidemiol Biomarkers Prev 1994 3 393 398 7920206
Gonzalez CA Riboli E Badosa J Batiste E Cardona T Pita S Sanz JM Torrent M Agudo A Nutritional factors and gastric cancer in Spain American Journal of Epidemiology 1994 139 466 473 8154470
Nomura AM Hankin JH Kolonel LN Wilkens LR Goodman MT Stemmermann GN Case-control study of diet and other risk factors for gastric cancer in Hawaii (United States) Cancer Causes and Control 2003 14 547 558 12948286 10.1023/A:1024887411846
Mayne ST Risch HA Dubrow R Chow WH Gammon MD Vaughan TL Farrow DC Schoenberg JB Stanford JL Ahsan H West AB Rotterdam H Blot WJ Fraumeni JFJ Nutrient intake and risk of subtypes of esophageal and gastric cancer Cancer Epidemiol Biomarkers Prev 2001 10 1055 1062 11588131
Koizumi Y Tsubono Y Nakaya N Nishino Y Shibuya D Matsuoka H Tsuji I No association between green tea and the risk of gastric cancer: pooled analysis of two prospective studies in Japan Cancer Epidemiol Biomarkers Prev 2003 12 472 473 12750246
Hoshiyama Y Kawaguchi T Miura Y Mizoue T Tokui N Yatsuya H Sakata K Kondo T Kikuchi S Toyoshima H Hayakawa N Tamakoshi A Ohno Y Yoshimura T Japan Collaborative Cohort Study Group A prospective study of stomach cancer death in relation to green tea consumption in Japan British Journal of Cancer 2002 87 309 313 12177800 10.1038/sj.bjc.6600487
Tsubono Y Nishino Y Komatsu S Hsieh CC Kanemura S Tsuji I Nakatsuka H Fukao A Satoh H Hisamichi S Green tea and the risk of gastric cancer in Japan New England Journal of Medicine 2001 344 632 636 11228277 10.1056/NEJM200103013440903
Sempos CT Johnson NE Smith EL Gilligan C Effects of intraindividual and interindividual variation in repeated dietary records American Journal of Epidemiology 1985 121 120 130 3964987
Hunt WC Leonard AG Garry PJ Goodwin JS Components of variation in dietary data for an elderly population Nutrition Research 1983 3 433 444
Messerer M Johansson SE Wolk A Sociodemographic and health behaviour factors among dietary supplement and natural remedy users European Journal of Clinical Nutrition 2001 55 1104 1110 11781678 10.1038/sj.ejcn.1601272
|
16202175
|
PMC1277847
|
CC BY
|
2021-01-04 16:39:31
|
no
|
Nutr J. 2005 Oct 5; 4:26
|
utf-8
|
Nutr J
| 2,005 |
10.1186/1475-2891-4-26
|
oa_comm
|
==== Front
Reprod HealthReproductive Health1742-4755BioMed Central London 1742-4755-2-81624203010.1186/1742-4755-2-8ResearchWomen's autonomy, education and contraception use in Pakistan: a national study Saleem Shabana [email protected] Martin [email protected] Reproductive Health Centre, Federal Government Services Hospital, Sector G\6, Islamabad Pakistan2 International Centre for Health and Society, Department of Epidemiology and Public Health, University College London, UK2005 21 10 2005 2 8 8 24 8 2005 21 10 2005 Copyright © 2005 Saleem and Bobak; licensee BioMed Central Ltd.2005Saleem and Bobak; 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 proposed that the autonomy of women is one of the mechanisms of how education influences contraceptive use in developing countries. We tested this hypothesis in a national sample of women in Pakistan.
Methods
We used the 2000 Pakistan Reproductive Health and Family Planning Survey, which interviewed a national sample of ever married women aged 15–49 years (n = 6579). Women's decision autonomy was estimated from 9 questions on who makes decisions at home; movement autonomy was based on 6 questions on whether women need permission to visit places outside home. A number of socio-demographic variables were used in multivariate analysis to investigate the independent association between autonomy and lifetime and current contraception use and to assess the extent to which autonomy mediates the association between education and contraception use.
Results
Decision autonomy was significantly associated with both lifetime and current contraception use; after controlling for covariates, the odds ratios for the highest vs. the lowest quintile were 1.8 (1.4–2.4) and 2.0 (1.4–2.8), respectively. Movement autonomy was not consistently associated with contraceptive use. Contraceptive use was strongly associated with women's education but this relation was not mediated by women's autonomy.
Conclusion
Women's decision autonomy is significantly associated with contraceptive use but it does not appear to mediate the link between woman's education and contraception.
==== Body
Introduction
Family planning is an important issue for many developing countries worldwide, including South Asia. In Pakistan, despite a governmental programme supporting family planning and despite the improvements over the last few decades, total fertility rate remains high (4.8 in 2000) and current contraception use remains relatively low (20% in 2000) [1]. In 2004, Pakistan had lower contraception use than most other Muslim countries [2].
Fertility and contraceptive use in developing countries are associated with various markers of socioeconomic status, most prominent of which is women's education [3,4]; the well documented link between female education and use of contraception plays an important role in development of family planning policies in lower income countries.
In parts of South Asia, and elsewhere, women have a considerably lower social status and autonomy than men [4-7], and their low status and autonomy seems to be associated with lower fertility control [4,6,8]. Several reports showed a positive association between women's autonomy and contraception use [4,9,10]. Improving women's education has been seen one way to increase their status and autonomy [4,5,7,11], and it has been proposed that autonomy acts as a mediator of the link between education and contraception use [4,8,12].
This paper, using population data from Pakistan in 2000, has two objectives. First, to investigate the relation between women's autonomy and contraception use, and second, to assess the extent to which women's autonomy mediates the association between education and contraception use. The report is based on a secondary analysis of an existing dataset, and the choice of dimensions of women's autonomy was therefore restricted to what was available in the dataset. However, since both sets of questions (on movement and decision autonomy) consisted of very similar or identical questions that have previously been used, the results should be well comparable with previous studies.
Methods
The survey
We used data from the 2000 Pakistan Reproductive Health and Family Planning Survey (PRHFPS) [1]. The survey used a multi-stage sampling method to randomly select 7332 households (details have been described elsewhere [1]). In each selected household, ever-married women 15–49 years old were asked to participate in an interview. Interviews were conducted by specially selected and trained female interviewers between October 2000 and January 2001. The interview collected extensive information on household composition and socioeconomic circumstances and on women's socioeconomic, demographic, reproductive and family characteristics.
Measurements
The basic demographic characteristics used in this analysis include province, urban/rural area of residence and women's age group and the number of living children. The standard of living of the household was characterised by the type of water supply, toilet facility and house construction. Women's education was taken as the basic measure of their socioeconomic status, and husband's education and employment in agriculture were taken as measures of husband's socioeconomic status (men employed in agriculture tend to have low income). The categorisation of these variables is shown in table 1.
Table 1 Numbers of women, contraceptive use (ever and current) and high (top quintile) decision and movement autonomy by socio-demographic characteristics.
Number (%) Contraception ever (%) Contraception currently* (%) Decision autonomy % Movement autonomy %
Total sample 6579 (100) 40.2 28.0 20.0 21.4
Province
Punjab 2895 (44.2) 44.8 31.5 24.7 26.6
Sindh 1791 (27.3) 32.7 24.0 15.8 12.7
NWFP 1167 (18.1) 43.3 26.4 19.5 15.6
Balochistan 606 (8.9) 29.9 21.6 6.8 29.2
Islamabad 120 (1.7) 64.0 48.3 43.3 41.7
p-value <0.001 <0.001 <0.001 <0.001
Area
Major urban 1524 (23.1) 62.0 47.3 30.4 34.1
Other urban 1302 (19.8) 46.7 30.4 19.3 22.6
Rural 3753 (57.1) 29.1 19.4 16.1 15.8
p-value <0.001 <0.001 <0.001 <0.001
Married
Yes 6361 (96.7) 40.8 28.1 18.5 20.4
No 218 (3.3) 23.4 23.4 63.8 50.0
p-value <0.001 0.127 <0.001 <0.001
Age group
<19 404 (6.1) 7.7 4.7 6.7 7.4
20–24 1081 (16.4) 23.7 13.8 11.2 10.9
25–29 1410 (21.4) 38.9 24.7 16.0 16.4
30–34 1233 (18.7) 48.7 33.4 21.9 20.9
35–39 1036 (15.8) 50.5 38.8 26.3 28.1
40+ 1415 (21.5) 48.5 36.0 28.3 33.8
p-value <0.001 <0.001 <0.001 <0.001
Water supply
Piped 2310 (35.1) 53.7 38.7 23.7 25.1
Well in residence 3168 (48.2) 36.0 24.5 19.0 21.0
Other 110 (16.7) 24.3 15.9 15.2 14.6
p-value <0.001 <0.001 <0.001 <0.001
House construction
Katcha 2305 (35.0) 23.9 14.9 14.6 15.8
Semi-pacca 1673 (25.4) 40.7 26.7 20.8 20.1
Pacca 1895 (28.8) 55.5 41.1 23.5 26.9
Flat/house 490 (7.4) 64.4 49.0 32.0 34.3
Other 216 (3.3) 21.3 15.1 12.5 13.0
p-value <0.001 <0.001 <0.001 <0.001
Toilet facility
Flush 3258 (49.5) 54.4 39.6 24.7 26.2
Other in house 1047 (15.9) 33.1 20.5 14.1 15.1
No facility 2274 (34.6) 23.3 15.0 16.1 17.3
p-value <0.001 <0.001 <0.001 <0.001
Education
None 4604 (70.0) 32.7 22.1 17.8 18.3
1–5 yrs 804 (12.2) 48.8 34.2 21.4 23.3
6–10 yrs 801 (12.2) 63.7 47.0 25.6 31.6
11+ yrs 370 (5.6) 64.1 46.9 32.4 33.8
p-value <0.001 <0.001 <0.001 <0.001
Living children
None 869 (13.2) 2.1 0.6 12.2 14.5
1–2 1747 (26.6) 31.2 20.0 18.0 17.7
3–4 1884 (28.6) 50.2 34.6 24.2 24.4
5+ 2079 (31.6) 54.7 40.3 21.2 24.6
p-value <0.001 <0.001 <0.001 <0.001
Husband's education
None 2561 (38.9) 28.7 19.2 20.7 21.5
1–5 yrs 1013 (15.4) 37.5 24.3 17.8 16.8
6–10 yrs 1929 (29.3) 48.4 34.2 19.7 21.1
11+ yrs 1076 (16.4) 55.5 41.2 21.0 25.9
p-value <0.001 <0.001 <0.001 <0.001
Husband works in agriculture**
No 5029 (79.0) 44.9 30.9 21.0 22.6
Yes 1338 (21.0) 25.3 17.9 9.3 12.4
p-value <0.001 <0.001 <0.001 <0.001
* among non-pregnant women (n = 6372)
** among married women = 6361
Of the several dimensions of women's autonomy described in the literature [4,7], two were assessed in this study: decision autonomy and movement autonomy. Decision autonomy was estimated from 9 questions on decision making (e.g. children's health care, education, buying/selling property, what to cook etc) [1]. The responses were scored as follows: 2 points for decisions made by the woman; 1 point by decisions made jointly by both the woman and her husband; and 0 for all of decisions taken by others. We used the Cronbach's alpha coefficient to assess whether individual questions in the scale measured the same one underlying factor (the higher coefficient, the more internally consistent is the scale; values larger than 0.6 are considered acceptable). The Cronbach's alpha was 0.78, indicating a good internal consistency. The sum of valid (non-missing) responses was divided by the number of valid responses, resulting in the final score with values between 0 (no autonomy) and 1 (full autonomy).
The movement autonomy scale was based on 6 questions on whether permission by husband or a senior family member was required to go to several places (market, health centre, relatives' home etc)[1] The responses were scored as 2 (no permission required), 1 (depends) and 0 (permission always required). The Cronbach's alpha of these 6 sub-questions was 0.87, suggesting high internal consistency. The score was calculated identically as for decision autonomy, with the final score ranging from 0 (no autonomy) to 1 (full autonomy). For both scales, the questions were very similar to those used in previous studies; responses were not weighted.
An overall autonomy score, combining both dimensions, was also calculated. However, the correlation between the two individual scores (decision and movement) was relatively low (r = 0.35), and in exploratory analyses the overall score predicted contraception use less well than the individual scores, possibly because of combining two different dimensions of autonomy dilutes the effects of each scale. For these reasons, the overall score was not used in the final analyses.
The outcome of interest was contraceptive use. Women reported whether they ever or currently used any contraception, and they indicated the method they used. There were no differences in results between all and "modern" contraception methods (the latter excluded withdrawal and abstinence), and we therefore report the results on all types of contraception.
Statistical analysis
The autonomy scores were distributed asymmetrically, with considerably more women with low autonomy than with high autonomy (figures 1 and 2). Women were therefore classified into quintiles of these three scores; since more than half of women reported no movement autonomy, there were only 4 categories.
Figure 1 Distribution of the decision autonomy score.
Figure 2 Distribution of the movement autonomy score.
Contraceptive use and autonomy were first tabulated by socio-demographic characteristics; statistical significance of the associations was assessed by chi square test. The odds ratios of contraceptive use (separately for ever and current use) by quintiles of decision and movement autonomy scores were estimated in logistic regression. (Since the data were collected in 367 primary sampling units, the sub-command "cluster" in STATA [13] was used to allow for potential autocorrelation within the sampling units). Second, the odds ratios were adjusted for the four basic demographic variables (province, urban/rural area of residence, age group and the number of living children). Third, the odds ratios were adjusted for all socio-demographic variables in table 1.
In the final step, we estimated the possible contribution of women's autonomy to the educational differences in contraceptive use. This was done by comparing the odds ratios by women's education before and after inclusion of women's autonomy into the following models: (i) crude (only education); (ii) adjusted for demographic factors (i.e. education plus province, urban/rural area of residence, age group and the number of living children); and (iii) fully adjusted (i.e. education plus all other socio-demographic variables listed in table 1). A reduction in odds ratio, after controlling for autonomy, would indicate a potential mediating role of autonomy. The reduction in odds ratios was quantified as [(odds ratio(adjusted for autonomy) – odds ratio(not adjusted for autonomy)) / (odds ratio(not adjusted for autonomy) -1)]. Since it is the difference in point estimates which matters, confidence intervals were not reported for this part of the analysis.
Results
There were 6579 women with valid data (table 1). The life time prevalence of contraceptive use was 40%, and 28% of women were current users. There were marked and statistically significant differences in contraceptive use and autonomy scores by all socio-demographic characteristics. In exploratory multivariate analyses, most socio-demographic variables were associated with both indicators contraception use; the most prominent predictors of contraception use were the number of living children and women's education (not shown in table). For example, the fully adjusted odds ratio for the highest vs. the lowest educational category was 3.6 (2.5–5.0) for ever use and 2.4 (1.8–3.3) for current use. Distribution of both autonomy scores was highly asymmetrical, indicating low levels of autonomy of most women (figures 1 and 2).
Table 2 shows the association between the two types of autonomy with contraceptive use (ever and current, respectively). In crude analyses (column 2), decision autonomy was strongly associated with both lifetime and current contraceptive use; the odds ratios for the highest vs. the lowest quintile were 4.8 (3.8–6.0) and 5.0 (3.7–6.9), respectively. The higher odds ratios for current use, compared with ever use, is probably due to more recent and slightly more precise information on current use. Controlling for demographic variables reduced the odds ratios considerably (but further adjustment for all variables used in table 1 did not change the results). Nevertheless, in the full model, decision autonomy remained significantly associated with contraception use. The relation between movement autonomy and contraception use was considerably weaker than that of decision autonomy, and it was not linear. In the full model, the highest odds ratio was seen for the 4th quintile.
Table 2 Odds ratios (95% confidence intervals) of contraceptive use ever by quintile of decision, movement and combined autonomy score.
Type of autonomy Quintile Crude Adjusted for demographic factors* Fully adjusted**
Contraception use ever
Decision autonomy
1 1.0 1.0 1.0
2 2.16 (1.72–2.70) 1.49 (1.15–1.94) 1.39 (1.06–1.82)
3 2.98 (2.37–3.74) 1.59 (1.22–2.07) 1.42 (1.08–1.86)
4 4.88 (3.91–6.09) 2.12 (1.64–2.75) 1.81 (1.38–2.37)
5 4.82 (3.84–6.04) 1.88 (1.45–2.46) 1.82 (1.37–2.41)
p-value for trend < 0.001 < 0.001 < 0.001
Movement autonomy
1+2 1.0 1.0 1.0
3 0.97 (0.80–1.18) 1.23 (1.01–1.51) 1.25 (1.02–1.48)
4 1.43 (1.16–1.77) 1.51 (1.22–1.88) 1.54 (1.26–1.86)
5 1.72 (1.45–2.05) 1.14 (0.96–1.36) 1.13 (0.95–1.33)
p-value for trend < 0.001 0.010 0.006
Current contraceptive use
Decision autonomy
1 1.0 1.0 1.0
2 2.36 (1.74–3.20) 1.82 (1.34–2.49) 1.73 (1.26–2.39)
3 3.16 (2.31–4.33) 1.88 (1.36–2.60) 1.73 (1.25–2.40)
4 5.34 (3.92–7.26) 2.54 (1.85–3.50) 2.23 (1.61–3.08)
5 5.04 (3.67–6.92) 2.15 (1.55–2.99) 2.01 (1.44–2.81)
p-value for trend < 0.001 < 0.001 < 0.001
Movement autonomy
1+2 1.0 1.0 1.0
3 0.92 (0.75–1.13) 1.13 (0.91–1.39) 1.13 (0.92–1.38)
4 1.36 (1.09–1.69) 1.29 (1.04–1.60) 1.29 (1.03–1.58)
5 1.66 (1.39–1.98) 1.06 (0.89–1.27) 1.03 (0.86–1.23)
p-value for trend < 0.001 0.174 0.288
* adjusted for province, urban/rural area of residence, age group and the number of living children
** adjusted for province, urban/rural area of residence, age group, the number of living children, type of toilet, type of house, type of water supply, education, husband education and husband's occupation in agriculture.
Finally, we examined the potentially mediating effect of autonomy on the association between women's education and contraceptive use (table 3). We did so by adjusting the effect of education for autonomy and assessing the change in odds ratios. The changes in the odds ratios (between models without and with autonomy) in models adjusted for demographic and other covariates were modest, up to 11%. The small impact of adjustment for autonomy suggests that autonomy is not a major mediator (or confounder) of the link between education and contraception.
Table 3 Changes in the odds ratios of current contraceptive use by education after controlling for autonomy, at different levels of adjustment. Percentages in parentheses indicate reduction in odds ratios after adding autonomy into the model.
Ever use Current use
Without autonomy Additionally adjusted for... Without autonomy Additionally adjusted for...
Decision autonomy Movement autonomy Both autonomy scores Decision autonomy Movement autonomy Both autonomy scores
Crude model
Education
None 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
1–5 yrs 1.96 1.85 (-11%) 1.92 (-4%) 1.83 (-10%) 1.83 1.72 (-13%) 1.79 (-5%) 1.71 (-14%)
6–10 yrs 3.60 3.32 (-11%) 3.46 (-5%) 3.24 (-14%) 3.12 2.85 (-13%) 2.99 (-6%) 2.80 (-15%)
11+ yrs 3.66 3.09 (-23%) 3.52 (-5%) 3.04 (-23%) 3.10 2.59 (-24%) 2.98 (-6%) 2.57 (-25%)
Adjusted for demographic factors*
Education
None 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
1–5 yrs 2.25 2.21 (-3%) 2.27 (+2%) 2.20 (-4%) 1.93 1.89 (-4%) 1.94 (+1%) 1.90 (-3%)
6–10 yrs 4.21 4.14 (-2%) 4.22 (0%) 4.11 (-3%) 3.17 3.09 (-4%) 3.18 (0%) 3.11 (-3%)
11+ yrs 5.35 5.07 (-6%) 5.33 (0%) 5.03 (-7%) 3.69 3.47 (-8%) 3.69 (0%) 3.51 (-7%)
Fully adjusted**
Education
None 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
1–5 yrs 1.65 1.62 (-5%) 1.65 (0%) 1.60 (-8%) 1.43 1.40 (-8%) 1.43 (0%) 1.40 (-8%)
6–10 yrs 2.81 2.75 (-3%) 2.78 (-2%) 2.69 (-7%) 2.12 2.06 (-5%) 2.11 (-1%) 2.05 (-6%)
11+ yrs 3.56 3.35 (-8%) 3.51 (-2%) 3.27 (-11%) 2.41 2.27 (-10%) 2.40 (-1%) 2.26 (-11%)
* adjusted for province, urban/rural area of residence, age group and the number of living children
** adjusted for province, urban/rural area of residence, age group, the number of living children, type of toilet, type of house, type of water supply, education, husband education and husband occupation in agriculture.
Discussion
In this secondary analysis of a large nationwide survey of women in Pakistan, we found that, in addition to a range of socio-demographic variables, women decision autonomy was significantly associated with contraceptive use. The data suggested, however, that autonomy did not mediate the association between women's education and contraceptive use.
The major strengths of the study are its representativeness, large sample size and comprehensive information on participating women and their household. The main weakness, on the other hand, is the cross-sectional design which may, in some aspects, obscure the temporality. For example, women's autonomy is partly derived from the number of living children, and the number of living children also influences contraception use; it is therefore difficult to clearly establish the temporality and causality of the effects. This problem, fortunately, does not affect the main focus of this paper. Education is a long-term and relatively stable characteristic, unlikely to be affected by the number of living children or autonomy. Similarly, it is unlikely that contraceptive use influences autonomy.
The second potential limitation is the opportunistic nature of these analyses, i.e. the fact that we relied on secondary data in defining the autonomy scores. Measurement of women's status is complex [7], with no general consensus on definition and most important autonomy dimensions. Using only two dimensions, from a number of those that have been suggested [4], is an oversimplification. However, similar or identical sets of questions have been used before and they appear to be useful to indicate women's autonomy in Pakistan and similar countries [9,10].
Our results confirmed the well known effects of most aspects of socioeconomic environment on contraceptive use (not reported in detail in this paper). From the various variables available, women's education had the most prominent role. This is consistent with most of the literature from South Asia and elsewhere [3,4,9,10].
The main focus of this analysis was on women's autonomy. Several findings deserve a note. First, the distribution of both autonomy scores was skewed towards low autonomy levels (figures 1 and 2). It has been pointed out that the western view, seeing low autonomy as negative, is not necessarily correct [7,14]. However, if women's decision authority is indeed associated with fertility and other health related characteristics, as our and others' results suggest, then the low levels of decision autonomy are of concern.
Second, decision autonomy remained significantly associated with contraceptive use, even after controlling for a battery of socio-demographic variables. The fully adjusted effects were not huge but they still indicate an approximately two-fold difference in contraceptive use between women with least and most autonomy. This is not negligible. By contrast, movement autonomy was not associated with contraceptive use after adjustment for other variables. This is consistent with a recent study of Pakistan which found only a limited role of women's mobility and uptake of reproductive health services [14].
Finally, only a few studies investigated the mediating effect of autonomy (on the influence of education) in individual level data, with conflicting results. While in Bangladesh women's autonomy played a major role [12], analysis of Indian data found that autonomy did not mediate the link between education and contraception [15]. In our data, both autonomy scores were associated with women's education (and most other socio-demographic characteristics) but they did not appear to mediate the effect of women's education on contraceptive use. High women's autonomy generally is generally seen as desirable (although its significance may be different in different settings [7,14]); however, our findings suggest that the impact of women's education on contraceptive use is independent from either decision or movement autonomy.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SS and MB jointly developed the principal idea for the analyses. SS obtained and analysed the data, reviewed the literature and commented on the draft. MB supervised the analyses and drafted the paper.
Acknowledgements
SS was funded by a scholarship from the Pakistani government.
==== Refs
Hakim A Sultan M ud din Ahmad F Pakistan Reproductive Health and Family Planning Survey (2000-01) Preliminary report 2001 Islamabad, National Institute of Population Studies
Population growth and its implications 2004 Islamabad, National Institute of Population Studies
Castro MT Women's education and fertility: results from 26 Demographic and Health Surveys Studies in Family Planning 1995 26 187 202 7482677
Jejeebhoy SJ Women's education, autonomy and reproductive behaviour: experience from developing countries 1995 Oxford, Clarendon Press
Jejeebhoy SJ Sathar ZA Women's Autonomy in India and Pakistan: The Influence of Religion and Region Population & Development Review 2001 27 687 712 10.1111/j.1728-4457.2001.00687.x
Dyson T Moore M On Kinship Structure, Female Autonomy, and Demographic Behavior in India Population & Development Review 1983 9 35 60
Mason KO The Status of Women: Conceptual and Methodological Issues in Demographic Studies Sociological Forum 1986 1 284 300 10.1007/BF01115740
Mason KO The Impact of Women's Social Position on Fertility in Developing Countries Sociological Forum 1987 2 718 745 10.1007/BF01124382
Fikree FF Khan A Kadir MM Sajan F Rahbar MH What Influences Contraceptive Use among Young Women in Urban Squatter Settlements of Karachi, Pakistan? International Family Planning Perspectives 2001 27 130 136
Al Riyami A Afifi M Mabry RM Women's autonomy, education and employment in Oman and their influence on contraceptive use Reprod Health Matters 2004 12 144 154 15242223 10.1016/S0968-8080(04)23113-5
Mason KO How family position influence married women's autonomy and power in five Asian countries REvision of a paper prepared for the CICRED Seminar on Women and the Family, Paris, 1997 1997 Paris
Cleland J Kamal N Sloggett A Jeffrey R and Basu A Links between fertility regulation and the schooling and autonomy of women in Bangladesh Girls schoooling, autonomy and fertlity change in South Asia 1996 New Dehli, Sage
StataCorp. Stata statistical software: Release 8 1995 College Station, TX, Stata Corporation
Mumtaz Z Salway S 'I never go anywhere': extricating the links between women's mobility and uptake of reproductive health services in Pakistan Soc Sci Med 2005 60 1751 1765 15686807 10.1016/j.socscimed.2004.08.019
Moursund A Kravdal O Individual and community effects of women's education and autonomy on contraceptive use in India Population Studies 2003 57 285 301 10.1080/0032472032000137817
|
16242030
|
PMC1277848
|
CC BY
|
2021-01-04 16:38:16
|
no
|
Reprod Health. 2005 Oct 21; 2:8
|
utf-8
|
Reprod Health
| 2,005 |
10.1186/1742-4755-2-8
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-621623231910.1186/1742-4690-2-62ResearchContribution of the C-terminal tri-lysine regions of human immunodeficiency virus type 1 integrase for efficient reverse transcription and viral DNA nuclear import Ao Zhujun [email protected] Keith R [email protected] Éric A [email protected] Xiaojian [email protected] Laboratory of Molecular Human Retrovirology, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba R3E 0W3, Canada2 Department of Medical Microbiology, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba R3E 0W3, Canada3 Laboratory of Human Retrovirology, Institut de Recherches Cliniques de Montréal, Département de microbiologie et immunologie, Faculté de Médecine, Université de Montréal, Montréal, Quebec H2W 1R7, Canada2005 18 10 2005 2 62 62 5 8 2005 18 10 2005 Copyright © 2005 Ao et al; licensee BioMed Central Ltd.2005Ao 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 addition to mediating the integration process, HIV-1 integrase (IN) has also been implicated in different steps during viral life cycle including reverse transcription and viral DNA nuclear import. Although the karyophilic property of HIV-1 IN has been well demonstrated using a variety of experimental approaches, the definition of domain(s) and/or motif(s) within the protein that mediate viral DNA nuclear import and its mechanism are still disputed and controversial. In this study, we performed mutagenic analyses to investigate the contribution of different regions in the C-terminal domain of HIV-1 IN to protein nuclear localization as well as their effects on virus infection.
Results
Our analysis showed that replacing lysine residues in two highly conserved tri-lysine regions, which are located within previously described Region C (235WKGPAKLLWKGEGAVV) and sequence Q (211KELQKQITK) in the C-terminal domain of HIV-1 IN, impaired protein nuclear accumulation, while mutations for RK263,4 had no significant effect. Analysis of their effects on viral infection in a VSV-G pseudotyped RT/IN trans-complemented HIV-1 single cycle replication system revealed that all three C-terminal mutant viruses (KK215,9AA, KK240,4AE and RK263,4AA) exhibited more severe defect of induction of β-Gal positive cells and luciferase activity than an IN class 1 mutant D64E in HeLa-CD4-CCR5-β-Gal cells, and in dividing as well as non-dividing C8166 T cells, suggesting that some viral defects are occurring prior to viral integration. Furthermore, by analyzing viral DNA synthesis and the nucleus-associated viral DNA level, the results clearly showed that, although all three C-terminal mutants inhibited viral reverse transcription to different extents, the KK240,4AE mutant exhibited most profound effect on this step, whereas KK215,9AA significantly impaired viral DNA nuclear import. In addition, our analysis could not detect viral DNA integration in each C-terminal mutant infection, even though they displayed various low levels of nucleus-associated viral DNA, suggesting that these C-terminal mutants also impaired viral DNA integration ability.
Conclusion
All of these results indicate that, in addition to being involved in HIV-1 reverse transcription and integration, the C-terminal tri-lysine regions of IN also contribute to efficient viral DNA nuclear import during the early stage of HIV-1 replication.
==== Body
Background
The integrase (IN) of human immunodeficiency virus type 1 (HIV-1) is encoded by the pol gene and catalyzes integration of viral cDNA into host chromosome, an essential step in HIV-1 replication. In addition to mediating the integration process, HIV-1 IN also participates in different steps during viral life cycle, including reverse transcription and viral DNA nuclear import [1-6]. During early phase of the HIV-1 replication cycle, after virus entry into target cells, another pol gene product, reverse transcriptase (RT), copies viral genomic RNA into double-stranded cDNA which exists within a nucleoprotein preintegration complex (PIC). The PIC also contains viral proteins including RT, IN, nucleocapsid (NC, p9), Vpr and matrix (MA, p17) and this large nucleoprotein complex is capable of actively translocating into the cell nucleus, including that of non-dividing cells (reviewed in reference [7]). This feature is particularly important for the establishment of HIV-1 replication and pathogenesis in exposed hosts, since the infection of postmitotic cells including tissue macrophages, mucosal dendritic cells as well as non-dividing T cells may be essential not only for viral transmission and dissemination, but also for the establishment of persistent viral reservoirs.
HIV-1 IN is composed of three functional domains, an N-terminal domain, a central catalytic core domain and a C-terminal domain, all of which are required for a complete integration reaction. The N-terminal domain harbors an HHCC-type zinc binding domain and is implicated in the multimerization of the protein and contributes to the specific recognition of DNA ends [8-10]. The core domain of IN contains the highly conserved DDE motif which is important for catalytic activity of the protein [11,12]. The C-terminal domain was shown to possess nonspecific DNA binding properties [13,14]. Some mutations within this region cause a drastic loss of virus infectivity without affecting the enzymatic activity of IN in vitro [2,13-16]. There are three conserved sequences in the C-terminus of IN that are essential for HIV-1 replication. Regions C (235WKGPAKLLWKGEGAVV) and N (259VVPRRKAK) are conserved in all known retroviruses and the 211KELQKQITK motif falls within the so-called glutamine-rich based region (sequence Q) of lentiviruses [17]. Alteration of each of the three sequences such as Q214L/Q216L, K215A/K219A, W235E, K236A/K240A, K244A/E246A, RRE263-5AAH resulted in loss of viral replication [15-18]. However, the mechanism(s) underlying the loss of viral infectivity remains controversial.
A number of studies have demonstrated the karyophilic properties of IN implicating that this protein may play an important role for PIC nuclear import [3,19-23]. However, the definition of nuclear localization signals (NLSs) in IN as well as their contribution to HIV-1 PIC nuclear import still remain to be determined. Previous report has suggested an atypical bipartite NLS (186KRK and 211KELQKQITK) by showing that IN mutants K186Q and Q214/216L in these regions lost the protein nuclear localization and their inability to bind to karyopherin α in vitro [3]. However, in attempt to analyze the effect of these mutants during HIV-1 replication, other studies did not reveal the importance of these IN mutants (K186Q and Q214/216L) for viral nuclear import; rather they appear to be required for reverse transcription, integration or undefined post-nuclear entry steps [16,18,23]. Also, another IN amino acid sequence IIGQVRDQAEHLK (aa161–173), was initially identified as an atypical NLS, which is required for viral DNA nuclear import [19]. However, reassessments of this putative NLS function failed to confirm this conclusion [24,25]. Some reports have also acknowledged that IN localization could result from passive diffusion of the protein and its DNA binding property [26,27], but DNA binding alone does not fully explain a rapid, ATP- and temperature-dependent nuclear import of IN [20]. It has recently been reported that the nuclear translocation of HIV-1 IN can be attributed to its interaction with a cellular component, human lens epithelium-derived growth factor/transcription coactivator p75 (LEDGF/p75) and LEDGF/p75 was also shown to be a component of HIV PIC [28,29]. However, whether this IN/LEDGF/p75 interaction plays an important role for HIV-1 nuclear import still remains to be elucidated, since HIV-1 infection and replication in LEDGF/p75-deficient cells was equivalent to that in control cells, regardless whether cells were dividing or growth arrested [29]. Thus, even though extensive studies have been dedicated in this specific research field, the contribution of HIV-1 IN to viral PIC nuclear import remains to be defined.
In this study, we have performed substitution mutational analysis to investigate the contribution of different C-terminal regions of IN to protein nuclear localization and their effects on HIV-1 replication. Our results showed that mutations of lysine residues in two tri-lysine regions, which are located within previously described Region C and sequence Q [17] in the C-terminal domain of HIV-1 IN, impaired protein nuclear localization, while mutations of arginines at amino acid position of 263 and 264 in the distal part of the C-terminal domain of IN had no significant effect. Moreover, we assessed the effect of these IN mutants during HIV-1 single cycle infection mediated by VSV-G pseudotyped RT/IN trans-complemented viruses. Results showed that, while all three C-terminal mutant viruses differentially affected HIV-1 reverse transcription, the KK240,4AE mutant exhibited most profound inhibition on this step, whereas KK215,9AA significantly impaired viral DNA nuclear import.
Results
The C-terminal domain of HIV-1 integrase (IN) is required for the nuclear localization of IN-YFP fusion protein
In this study, we first investigated the intracellular localization of HIV-1 IN and delineated the region(s) of IN contributing to its karyophilic property. A HIV-1 IN-YFP fusion protein expressor (CMV-IN-YFP) was generated by fusing a full-length HIV-1 IN cDNA (amplified from HIV-1 HxBru molecular clone [30]) to the 5' end of YFP cDNA in a CMV-IN-YFP expressor, as described in Materials and Methods. Transfection of CMV-IN-YFP expressor in 293T cells resulted in the expression of a 57 kDa IN-YFP fusion protein (Fig. 1B, lane 2; Fig. 2B, lane 1), whereas expression of YFP alone resulted in a 27 kDa protein (Fig. 2B, lane 5). Given that HeLa cells have well-defined morphology and are suitable for observation of intracellular protein distribution, we tested the intracellular localization of YFP and IN-YFP by transfecting CMV-IN-YFP or CMV-YFP expressor in HeLa cells. After 48 hours of transfection, cells were fixed and subjected to indirect immunofluorescence assay using primary rabbit anti-GFP antibody followed by secondary FITC-conjugated anti-rabbit antibodies. Results showed that, in contrast to a diffused intracellular localization pattern of YFP (data not shown), the IN-YFP fusion protein was predominantly localized in the nucleus (Fig 1C, a1), confirming the karyophilic feature of HIV-1 IN.
Figure 1 Subcellular localization of the wild-type and truncated HIV integrase fused with YFP. A) Schematic structure of HIV-1 integrase-YFP fusion proteins. Full-length (1–288aa) HIV-1 integrase, the N-terminus-truncated mutant (51–228aa) or the C-terminus-truncated mutant (1–212aa) was fused in frame at the N-terminus of YFP protein. The cDNA encoding for each IN-YFP fusion protein was inserted in a SVCMV expression plasmid. B) Expression of different IN-YFP fusion proteins in 293T cells. 293T cells were transfected with each IN-YFP expressor and at 48 hours of transfection, cells were lysed, immunoprecipitated with anti-HIV serum and resolved by electrophoresis through a 12.5% SDS-PAGE followed by Western blot with rabbit anti-GFP antibody. The molecular weight markers are indicated at the left side of the gel. C) Intracellular localization of different IN-YFP fusion proteins. HeLa cells were transfected with each HIV-1 IN-YFP fusion protein expressor and at 48 hours of transfection, cells were fixed and subjected to indirect immunofluorescence using rabbit anti-GFP and then incubated with FITC-conjugated anti-rabbit antibodies. The localization of each fusion protein was viewed by Fluorescence microscopy with a 50× oil immersion objective. Upper panel is fluorescence images and bottom panel is DAPI nucleus staining.
Figure 2 Effect of different IN C-terminal substitution mutants on IN-YFP intracellular localization. A) Diagram of HIV-1 IN domain structure and introduced mutations at the C-terminal domain of the protein. The position of lysines in two tri-lysine regions and introduced mutations are shown at the bottom of sequence. B) The expression of the wild-type and mutant IN-YFP fusion proteins were detected in transfected 293T cells by using immunoprecipitation with anti-HIV serum and Western blot with rabbit anti-GFP antibody, as described in figure 1. The molecular weight markers are indicated at the left side of the gel. C) Intracellular localization of different HIV-1 IN mutant-YFP fusion proteins in HeLa cells were analyzed by fluorescence microscopy with a 50× oil immersion objective. The nucleus of HeLa cells was simultaneously visualized by DAPI staining (lower panel).
To delineate the karyophilic determinant in HIV-1 IN, two truncated IN-YFP expressors CMV-IN50–288-YFP and CMV-IN1–212-YFP were generated. In CMV-IN50–288-YFP, the N-terminal HH-CC domain of IN (aa 1–49) was deleted and in CMV-IN1–212-YFP, the C-terminal domain (aa 213–288) was removed (Fig. 1A). Transfection of each truncated IN-YFP fusion protein expressor in 293T cells resulted in the expression of IN50–288-YFP and IN1–212-YFP at approximately 52 kDa and 48 kDa molecular mass respectively (Fig. 1B, lanes 3 and 4). We next investigated the intracellular localization of truncated IN-YFP fusion proteins in HeLa cells by using indirect immunofluorescence assay, as described above. Results showed that the IN50–288-YFP was predominantly localized in the nucleus with a similar pattern as the wild-type IN-YFP fusion protein (Fig. 1C, compare b1 to a1). However, IN1–212-YFP fusion protein was excluded from the nucleus, with an accumulation of the mutant protein in the cytoplasm (Fig 1C, c1). These results were also further confirmed by using rabbit anti-IN antibody immunofluorescence assay (data not shown). Taken together, our data show that the C-terminal domain of HIV-1 IN is required for its nuclear accumulation.
Two tri-lysine regions in the C-terminal domain of IN are involved in the protein nuclear localization
The C-terminal domain of HIV-1 IN contains several regions that are highly conserved in different HIV-1 strains, including Q, C and N regions [17]. Interestingly, in regions Q and C, sequences of 211KELQKQITK and 236KGPAKLLWK possess high similarity in terms of numbers and position of lysine residues and therefore, we term them proximal tri-lysine region and distal tri-lysine region, respectively (Fig. 2A). All of these lysine residues are highly conserved in most HIV-1 strains [31]. To test whether these basic lysine residues could constitute for a possible nuclear localization signal for IN nuclear localization, we specifically introduced substitution mutations for two lysines in each tri-lysine region and generated INKK215,9AA-YFP and INKK240,4AE-YFP expressors (Fig. 2A). In the conserved N region, there is a stretch of four basic residues among five amino acids (aa) 262RRKAK. To characterize whether this basic aa region may contributes to IN nuclear localization, we replaced an arginine and a lysine at positions of 263 and 264 by alanines in this region and generated a mutant (INRK263,4AA-YFP). The protein expression of different IN-YFP mutants in 293T cells showed that, like the wild type IN-YFP, each IN-YFP mutant fusion protein was detected at similar molecular mass (57 kDa) in SDS-PAGE (Fig 2B, lanes 1 to 4), while YFP alone was detected at position of 27 kDa (lane 5). Then, the intracellular localization of each IN mutant was investigated in HeLa cells by using similar methods, as described above. Results showed that, while the wild type IN-YFP and INRK263,4AA-YFP still predominantly localized to the nucleus (Fig. 2C, a1 and d1), both INKK215,9AA-YFP and INKK240,4AE-YFP fusion proteins were shown to distribute throughout the cytoplasm and nucleus, but with much less intensity in the nucleus (Fig 2C, a1 and b1). These data suggest that these lysine residues in each tri-lysine regions are required for efficient HIV-1 IN nuclear localization.
Production of VSV-G pseudotyped HIV-1 IN mutant viruses and their effects on HIV-1 infection
Given that two di-lysine mutants located in the C-terminal domain of IN are involved in HIV-1 IN nuclear localization, we next evaluated whether these IN mutants would affect the efficiency of HIV-1 infection. To specifically analyze the effect of IN mutants in early steps of viral infection, we modified a previously described HIV-1 single-cycle replication system [32] and constructed a RT/IN/Env gene-deleted HIV-1 provirus NLlucΔBglΔRI, in which the nef gene was replaced by a firefly luciferase gene [33]. Co-expression of NLlucΔBglΔRI provirus with Vpr-RT-IN expressor and a vesicular stomatitis virus G (VSV-G) glycoprotein expressor will produce viral particles that can undergo a single-round of replication, since RT, IN and Env defects of provirus will be complemented in trans by VSV-G glycoprotein and Vpr-mediated RT and IN trans-incorporation [32]. This single cycle replication system allows us to introduce different mutations into IN gene sequence without differentially affecting viral morphogenesis and the activity of the central DNA Flap. After different IN mutations KK215,9AA, KK240,4AE and RR263,4AA were introduced into Vpr-RT-IN expressor, we produced VSV-G pseudotyped HIV-1 IN mutant virus stocks in 293T cells. In order to specifically investigate the effect of IN mutants on early steps during HIV-1 infection prior to integration, an IN class I mutant D64E was also included as control. After each viral stock was produced (as indicated in Fig. 3A), similar amounts of each virus stock (quantified by virion-associated RT activity) were lysed and virus composition and trans-incorporation of RT and IN of each virus stock were analyzed by Western blot analysis with anti-IN and anti-HIV antibodies, as described in Materials and Methods. Results showed that all VSV-G pseudotyped IN mutant viruses had similar levels of Gagp24, IN and RT, as compared to the wild-type virus (Fig. 3A), indicating that trans-incorporation of RT and IN as well as HIV-1 Gag processing were not differentially affected by the introduced IN mutations.
Figure 3 Production of different single-cycle replicating viruses and their infection in HeLa-CD4-CCR5-β-Gal cells. A). To evaluate the trans-incorporation of RT and IN in VSV-G pseudotyped viral particles, viruses released from 293T cells transfected with NLlucΔBglΔRI provirus alone (lane 6) or cotransfected with different Vpr-RT-IN expressors and a VSV-G expressor (lane 1 to 5) were lysed, immunoprecipitated with anti-HIV serum. Then, immunoprecipitates were run in 12% SDS-PAGE and analyzed by Western blot with rabbit anti-IN antibody (middle panel) or anti-RT and anti-p24 monoclonal antibody (upper and lower panel). B) The infectivity of trans-complemented viruses produced in 293 T cells was evaluated by MAGI assay. HeLa-CD4-CCR5-LTR-β-Gal cells were infected with equal amounts (at 10 cpm/cell) of different IN mutant viruses and after 48 hours of infection, numbers of β-Gal positive cells (infected cell) were monitored by X-gal staining. Error bars represent variation between duplicate samples and the data is representative of results obtained in three independent experiments.
To test the infectivity of different IN mutant viruses in HeLa-CD4-CCR5-LTR-β-Gal cells, we first compared the infectivity of VSV-G pseudotyped wild type virus and the D64E mutant virus. At 48 hours post-infection with equivalent amount of each virus stock (at 1 cpm RT activity/cell), the number of β-Gal positive cells was evaluated by MAGI assay, as described previously [34]. Results showed that the number of infected cells (β-Gal positive cells) for D64E mutant reached approximately 14% of the wild type level (data not shown). This result is consistent with a previous report showing that, in HeLa MAGI assay, the infectivity level of class I IN integration-defect mutant was approximately 20 to 22% of wild type level [15]. It indicates that, even though the IN mutant D64E virus is defective for integrating viral DNA into host genome, tat expression from nucleus-associated and unintegrated viral DNAs can activate HIV-1 LTR-driven β-Gal expression in HeLa-CD4-CCR5-LTR-β-Gal cells. Indeed, several studies have already shown that HIV infection leads to selective transcription of tat and nef genes before integration [2,35,36]. Therefore, this HeLa-CD4-CCR5-LTR-β-Gal cell infection system provides an ideal method for us to evaluate the effect of different IN mutants on early steps of viral infection prior to integration. We next infected HeLa-CD4-CCR5-LTR-β-Gal cells with different VSV-G pseudotyped IN mutant viruses at higher infection dose of 10 cpm RT activity/cell and numbers of β-Gal positive cells were evaluated by MAGI assay after 48 hours of infection. Interestingly, results showed that the IN mutant D64E virus infection induced the highest level of β-Gal positive cells, whereas infection with viruses containing IN mutants KK215,9AA, KK240,4AE or RK263,4AA yielded much lower levels of β-Gal positive cells, which only reached approximately 11%, 5% or 26% of the level of D64E virus infection (Fig. 3B). Based on these results, we reasoned that these IN C-terminal mutants blocked infection mostly by affecting earlier steps of HIV-1 life cycle, such as reverse transcription and/or viral DNA nuclear import steps, which are different from the action of D64E mutant on viral DNA integration.
Effect of IN mutants on viral infection in dividing and non-dividing C8166 T cells
To further test whether these C-terminal mutants could induce similar phenotypes in CD4+ T cells, we infected dividing and non-dividing (aphidicolin-treated) C8166 CD4+ T cells with equal amounts of VSV-G pseudotyped IN mutant viruses (at 5 cpm of RT activity/cell). Since all IN mutant viruses contain a luciferase (luc) gene in place of the nef gene, viral infection can be monitored by using a sensitive luc assay which could efficiently detect viral gene expression from integrated and unintegrated viral DNA [33]. After 48 hours of infection, equal amounts of cells were lysed in 50 μl of luc lysis buffer and then, 10 μl of cell lysates was used for measurement of luc activity, as described in Materials and Methods. Results showed that the D64E mutant infection in dividing C8166 T cells induced 14.3 × 104 RLU of luc activity (Fig. 4A), which was approximately 1000-fold lower than that in the wild type virus infection (data not shown). This level of luc activity detected in D64E mutant infection is mostly due to nef gene expression from the unintegrated DNA [33]. In agreement with the finding by MAGI assay described in figure 3, the Luc activity detected in KK215,9AA, KK240,4AE and RK263,4AA mutant samples were approximately 13%, 5% and 36% of level of D64E mutant infection (Fig. 4A). In parallel, infection of different IN mutants in non-dividing C8166 T cells was also evaluated and similar results were observed (Fig. 4B).
Figure 4 Effect of IN mutants on viral infection in dividing and nondividing C8166 T cells. To test the effect of different IN mutants on HIV-1 infection in CD4+ T cells, dividing (panel A) and non-dividing (aphidicolin-treated, panel B) C8166 T cells were infected with equal amount of VSV-G pseudotyped IN mutant viruses (at 5 cpm/cell). For evaluation of the effect of different IN mutants on HIV-1 envelope-mediated infection in CD4+ T cells, dividing C8166 T cells were infected with equal amount of HIV-1 envelope competent IN mutant viruses (at 10 cpm/cell) (panel C). After 48 hours of infection, HIV-1 DNA-mediated luciferase induction was monitored by luciferase assay. Briefly, the same amount (106 cells) of cells was lysed in 50 ul of luciferase lysis buffer and then, 10 μl of cell lysate was subjected to the luciferase assay. Error bars represent variation between duplicate samples and the data is representative of results obtained in three independent experiments.
To test whether these IN mutants had similar effects during HIV-1 envelope-mediated single cycle infection, we produced virus stocks by co-transfecting 293T cells with a HIV-1 envelope-competent NLlucΔRI provirus with each Vpr-RT-IN mutant expressor, as described in Materials and Methods. Then, dividing CD4+ C8166 cells were infected with each virus stock (at 10 cpm RT activity/cells). At 48 hours post-infection, cells were collected and measured for luc activity. Results from figure 4C showed that, similar to results obtained from VSV-G pseudotyped virus infection (Fig. 4A), the Luc activity detected in cells infected by HIV-1 envelope competent KK215,9AA, KK240,4AE and RK263,4AA mutant viruses were approximately 13.5%, 6% and 29% of level of D64E mutant infection (Fig. 4C). All of these results confirm the data from HeLa-CD4-CCR5-LTR-β-Gal infection (Fig. 3) by using either VSV-G- and HIV-1 envelope-mediated infections and suggest again that the significantly attenuated infection of KK215,9AA, KK240,4AE and RK263,4AA mutant viruses may be due to their defect(s) at reverse transcription and/or viral DNA nuclear import steps.
Effects of IN mutants on reverse transcription, viral DNA nuclear import and integration
All results so far suggest that these C-terminal mutants might significantly affect early steps during HIV-1 replication. To directly assess the effect of these IN C-terminal mutants on each early step during viral infection, we analyzed the viral DNA synthesis, their nuclear translocation and integration following each IN mutant infection in dividing C8166 cells. Levels of HIV-1 late reverse transcription products were analyzed by semi-quantitative PCR after 12 hours of infection with HIV-1 specific 5'-LTR-U3/3'-Gag primers and Southern blot, as previously described [32,37]. Also, intensity of amplified HIV-1 specific DNA in each sample was evaluated by laser densitometric scanning of bands in Southern blot autoradiograms (Fig. 5A). Results showed that total viral DNA synthesis in both KK215,9AA and RK263,4AA infection reached approximately 61% and 46% of that of the wild type (wt) virus infection (Fig. 5A and 5B). Strikingly, in KK240,4AA sample, detection of viral DNA synthesis was drastically reduced, which only reached 21% of viral DNA level in WT sample (Fig. 5A and 5B). These results indicate that all three C-terminal mutants negatively affected viral reverse transcription during viral infection and KK240,4AA mutant exhibited most profound effect.
Figure 5 Effects of different IN mutants on HIV-1 reverse transcription and DNA nuclear import. Dividing C8166 T cells were infected with equal amounts of different HIV-1 IN mutant viruses. A) At 12 hours post-infection, 1 × 106 cells were lysed and the total viral DNA was detected by PCR using HIV-1 LTR-Gag primers and Southern blot. B) Levels of HIV-1 late reverse transcription products detected in panel A were quantified by laser densitometry and viral DNA level of the wt virus was arbitrarily set as 100%. Means and standard deviations from two independent experiments are presented. C) At 24 hours post-infection, 2 × 106 cells were fractionated into cytoplasmic and nuclear fractions as described in Materials and Methods. The amount of viral DNA in cytoplasmic and nuclear fractions were analyzed by PCR using HIV-1 LTR-Gag primers and Southern blot (upper panel, N. nuclear fraction; C. cytoplasmic fraction). Purity and DNA content of each subcellular fraction were monitored by PCR detection of human globin DNA and visualized by specific Southern blot (lower panel). D). The percentage of nucleus-associated viral DNA relative to the total amount of viral DNA for each mutant was also quantified by laser densitometry. Means and standard deviations from two independent experiments are shown.
Meanwhile, the nucleus- and cytoplasm-associated viral DNA levels were analyzed at 24 hours post-infection in C8166 T cells. The infected cells were first gently lysed and separated into nuclear and cytoplasmic fractions by using a previously described fractionation technique [37]. Then, levels of HIV-1 late reverse transcription products in each fraction were analyzed by semi-quantitative PCR, as described above. Results revealed differential effects of C-terminal mutants on HIV-1 DNA nuclear import. In the wt, D64E and RK263,4AA virus-infected samples, there were respectively 70%, 72% and 68% of viral DNA associated with nuclear fractions (Fig. 5C (upper panel, lanes 1 and 2; 3 and 4; 9 and 10) and 5D). For KK240,4AE mutant, approximately 51% of viral DNA was nucleus-associated (Fig. 5C (upper panel, lane 7 and 8) and 5D). Remarkably, in KK215,9AA infected sample, viral cDNA was found predominantly in the cytoplasm and only approximately 21% of viral DNA was associated with the nuclear fraction (Fig. 5C (upper panel, lane 5 and 6) and 5D). Meanwhile, the integrity of fractionation procedure was validated by detection of β-globin DNA, which was found solely in the nucleus and levels of this nucleus-associated cellular DNA were similar in each nuclear sample (Fig. 5C, lower panel).
Even though the C-terminal mutants were shown to significantly affect HIV-1 reverse transcription and/or nuclear import, the various low levels of nucleus-associated viral DNA during the early stage of replication (Fig. 5C) may still be accessible for viral DNA integration. To address this question, 1 × 106 dividing C8166 T cells were infected with equivalent amounts of each single cycle replicating virus stock (5 cpm/cell), as indicated in figure 6 and after 24 hours of infection, the virus integration level was checked by using a previously described sensitive Alu-PCR technique [32], Results revealed that, while the wt virus resulted in an efficient viral DNA integration (Fig. 6, upper panel; lanes 1 and 2), there was no viral DNA integration detected in D64E mutant (lanes 3 to 4) and in all three C-terminal mutant infection samples (lanes 5 to 10), although similar levels of cellular β-globin gene were detected in each sample (Fig. 6, middle panel). These results suggest that, in addition to affecting HIV-1 reverse transcription and nuclear import, all three C-terminal IN mutants tested in this study also negatively affected viral DNA integration. Overall, all of these results indicate that all three IN C-terminal mutants are belonged to class II mutants, which affected different early steps during HIV-1 replication. Among these mutants, the KK240,4AE showed the most profound inhibition on reverse transcription and the KK215,9AA, and to a lesser extent, KK240,4AE, impaired viral DNA nuclear translocation during early HIV-1 infection in C8166 T cells.
Figure 6 Effect of IN mutants on HIV-1 proviral DNA integration. Dividing C8166 T cells were infected with equal amounts of different HIV-1 IN mutant viruses. At 24 hours post-infection, 1 × 106 cells were lysed and serial-diluted cell lysates were analyzed by two-step Alu-PCR and Southern blot for specific detection of integrated proviral DNA from infected cells (Upper panel). The DNA content of each lysis sample was also monitored by PCR detection of human β-globin DNA and visualized by specific Southern blot (middle panel). The serial-diluted ACH-2 cell lysates were analyzed for integrated viral DNA and as quantitative control (lower panel). The results are representative for two independent experiments.
Discussion
In this study, we performed mutagenic studies to analyze different regions in the C-terminal domain of HIV-1 IN that contribute to protein nuclear localization as well as their effects on virus infection. First, our analyses showed that specific lysine mutations introduced in two highly conserved tri-lysine regions in the C-terminal domain of HIV-1 IN impaired protein nuclear accumulation. Second, infection experiments revealed that all three C-terminal mutant viruses (KK215,9AA, KK240,4AE and RK263,4AA) exhibited more severe defect of induction of β-Gal positive cells and luc activity, as compared to an IN class 1 mutant D64E virus, in CD4+ HeLa-β-Gal cells, dividing and non-dividing C8166 T cells. It suggests that all three C-terminal mutant virus infections may have defects at steps prior to integration. Further analysis of total viral DNA synthesis, viral DNA nuclear import and integration indicates that all three C-terminal mutants displayed a class II mutant profile. Even though all of them reduced viral reverse transcription levels, the mutant KK240,4AE showed the most profound inhibitory effect. In addition, the mutant KK215,9AA, and to a lesser extent, KK240,4AE, impaired viral DNA nuclear translocation. These IN mutant-induced defects do not appear to result from various effects of mutants on Gag-Pol processing and maturation given that RT and IN were complemented in trans in this HIV-1 single-cycle infection system. Rather, the effect of different IN mutants on reverse transcription and viral DNA nuclear import is likely originated from a role of mutants within the maturing PIC complexes.
Previous work by Gallay et al., have proposed an atypical bipartite NLS (186KRK and 211KELQKQITK) in HIV-1 IN by finding that IN mutants K186Q and Q214/216L lost their karyophilic feature and their ability to bind to karyopherin α in vitro [3]. Even though these results were confirmed by Petit and colleagues by studying the intracellular localization of HIV-1 Flag-IN [18], other studies, using GFP-IN fusion protein, did not reveal the importance of K186Q and Q214/216L mutations for HIV-1 IN nuclear localization [16,23,27]. Therefore, the definition of region(s) in HIV-1 IN contributing to the protein nuclear localization is still controversial. In this study, we investigated the intracellular localization of several IN-YFP fusion proteins including the C-terminal-deletion mutant IN1–212-YFP, substitution mutants INKK215,9AA-YFP and INKK240,4AE-YFP and found that all of these IN fusion mutants impaired protein nuclear accumulation. It suggests that two C-terminal tri-lysine regions 211KELQKQITK and 236KGPAKLLWK contribute to IN nuclear localization. Interestingly, the study by Maertens et al also showed that the fusion of HIV-1 IN C-terminal fragment alone with GFP rendered fusion protein to be exclusively in the nucleus, speculating that the C-terminal domain may have a role in HIV-1 nuclear import [28]. However, at this moment, we still could not exclude the possibility that the IN nuclear accumulation could be facilitated by the DNA binding ability of IN protein, as suggested by Devroe et al [27]. It has to be noted that two studies have previously observed the nuclear localization of GFP-IN fusion proteins although the C-terminal domain of IN was deleted from the fusion protein [23,28]. It has also been shown that both N-terminal zinc binding domain and the central core domain of HIV-1 IN are involved in its interaction with a cellular protein, human lens epithelium-derived growth factor/transcription coactivator p75 (LEDGF/p75) and this IN/LEDGF/p75 interaction is required for GFP-IN nuclear localization [28]. However, our deletion analysis by using IN-YFP fusion protein failed to reveal the importance of both N-terminal and core domains for IN nuclear localization (Fig. 1). One explanation for this discrepancy could be different orientations of fusion proteins used in our study (IN-YFP) and other studies (GFP-IN). It is possible that different forms of fusion proteins may differentially affect the ability of IN to interact with LEDGF/p75 and consequently affect their ability for nuclear targeting. Therefore, it would be interesting to test whether INKK215,9AA-YFP and INKK240,4AE-YFP could loss their ability to interact with LEDGF/p75. These studies are underway.
An important question that needs to be addressed is the impact of nuclear localization-defective IN mutants on HIV-1 replication. Given that most IN mutants characterized so far are classified as class II mutants that cause pleiotropic damage including defects in viral morphogenesis, reverse transcription and integration [16,38], we used a previously described VSV-G pseudotyped HIV-1 RT/IN trans-complement single-cycle replication system [32,39] to minimize differential effects of IN mutants on virus maturation. Also, in our infection experiments, a specific integration-defective class I mutant D64E virus was introduced in order to monitor the viral gene expression from unintegrated HIV-1 DNA species that are already translocated into nucleus during virus infection. It is known that certain levels of selected viral gene expression (tat and nef) from unintegrated viral DNA species are detected during this Class I mutant infection [2,35,36]. Interestingly, our infection analysis revealed that more profound infection defects were found for all three IN C-terminal mutant viruses KK215,9AA, KK240,4AE and RK263,4AA than D64E mutant virus in Hela-CD4-CCR5-β-Gal cells, dividing and non-dividing C8166 T cells (Fig. 3 and 4). These results suggest that these C-terminal IN mutants may affect early steps such as reverse transcription and/or nuclear import and consequently result in a reduced level of viral DNA in the nucleus, which is accessible for tat and nef expression, To understand the mechanism(s) underlying replication defects of each C-terminal mutant, levels of total reverse transcription were analyzed during early viral infection. Consistent with a previous study [6], infection with D64E mutant virus did not affect reverse transcription as compared to wt virus infection. However, all three C-terminal mutants display various levels of impaired HIV-1 reverse transcription (Fig. 5A and 5B). The mutant KK240,4AE showed strongest inhibition of reverse transcription (21% compared to the wt level (100%)), while mutants KK215,9AA and RK263,4AA reached to 61% and 46% (Fig. 5A and 5B). These data indicate that all of these IN mutants, especially KK240,4AA, negatively affect reverse transcription at early viral infection. Consistently, recent studies have shown that the C-terminal domain of IN contributes to efficient reverse transcription and this domain of IN was able to bind to heterodimeric RT [6,40,41]. It is possible that these C-terminal mutants, especially for KK240,4AE, may disrupt the interaction between IN and RT and result in decreased viral cDNA synthesis.
Subsequently, we examined levels of nucleus- and cytoplasm-associated viral DNA during early virus infection. Results clearly show that the nuclear localization defective mutant KK215,9AA leads to significantly reduced levels of viral DNA in the nucleus, as compared to the wt and D64E viruses (Fig. 5C and 5D). It suggests that the Q region is in fact important for HIV-1 nuclear import. Consistently, a recent study by Lu et al also observed that infection of K215A/K219A mutant induced more than 3-fold lower luc activity compared to class I IN mutant D64N/D116N [16]. Moreover, similar to our experimental system, their study revealed that, in the context of VSV-G pseudotyped virus infection in Jurkat cells, 2-LTR circle DNA levels of K215A/K219A and Q214L/Q216L were significantly lower than other mutants V165A and C130G, even though the inhibition of viral reverse transcription mediated by these mutants were comparable [16]. In addition, KK240,4AE mutant also showed a modest impairment of viral DNA nuclear import (Fig. 5C and 5D). In fact, this mutant exhibited the most profound infection defect, compared to other two mutants (KK215,9AA and RK263,4AA) (Fig. 3 and 4). This may be due to combined effects of this mutant on both reverse transcription and viral DNA nuclear import, as shown in Fig. 5. One interesting question is whether such profound infection defect of KK240,4AE mutant virus could be due to a structural alteration by replacing glutamic acid (E) for lysine at position of 244. It seems to be unlikely since 1) the effect of this mutant on nuclear import was not as dramatic as KK215,9AA mutant (as shown in Fig. 5); 2) Wiskerchen et al have reported that infection of MAGI cells with two other IN mutants K236A/K240A and K244A/E246A mutants, that are located in the same region as our KK240,4AE mutant, resulted in 0 and 4 β-Gal positive cells, while infection of class I IN mutants produced 700 to 1400 β-Gal positive cells [15]. All of these observations suggest that this region indeed plays an important role for IN activities during early stage of virus infection prior to integration. Also, it has to be noted that although similar inhibition of reverse transcription was seen for KK215,9AA and RK263,4AA mutants, RK263,4AA mutant induced two to three fold higher level of β-Gal positive cells and luc activity than KK215,9AA mutant (Fig. 3 and 4). This is expected since KK215,9AA affected both reverse transcription and nuclear import, while RK263,4AA mutant only impaired reverse transcription (Fig. 5). In addition, our analysis could not detect viral DNA integration in each C-terminal mutant infection (Fig. 6), even though they displayed various low levels of nucleus-associated viral DNA (Fig. 5C). It suggests that these IN mutants may also negatively affect viral integration during their infection. Alternatively, it could be possible that these mutants may have additional defect(s) at an undefined postnuclear entry step that is required for viral DNA integration, as suggested by Lu et al [16]. Consistently, their recent reports have shown that several IN mutants in same regions, including K215A/K219A, E244A and R262A/K264A, completely lost virus replication ability in CD4+ Jurkat T cells [16,42].
Up to now, the mechanism(s) underlying the action of HIV-1 IN in viral PIC nuclear import is still unclear. Since IN is a component of viral PIC, at least two factors may affect the contribution of IN to viral PIC nuclear import: first, IN needs to directly or indirectly associate with viral DNA and/or other PIC-associated proteins in order to participate in driving viral DNA into the nucleus; second, IN needs to have a NLS and/or bind to other karyophilic proteins for nuclear translocation. Any mutation disrupting one of these two abilities would affect IN's action for viral DNA nuclear import. A recent study evaluated the effect of several IN core domain mutants targeting key residues for DNA recognition on HIV-1 replication and indicated that, while all of these IN mutants maintained their karyophilic properties, viruses harboring these mutants still severely impaired viral DNA nuclear import [4]. In our study, both KK215,9AA and KK240,4AE mutants clearly lost their karyophilic properties and negatively affected viral DNA nuclear import. However, it is still premature to define these regions acting as IN NLS, even though a previously described IN mutant Q214/216L, which is also located in proximal tri-lysine domain, has been shown to reduce IN-karyopherin α interaction in vitro [3]. More studies are required for further characterization of molecular mechanisms underlying the action of these IN mutants during HIV-1 DNA nuclear import.
Conclusion
Taken together, the results presented here highlight that all three C-terminal mutants tested in this study resulted in drastic loss of viral infectivity that were due to defects in different early steps of viral replication. Specific lysine mutations introduced in the tri-lysine regions of the C-terminal domain of HIV-1 IN, especially for KK215,9AA, impaired protein nuclear accumulation and HIV-1 PIC nuclear import. Although all of C-terminal mutants inhibited viral reverse transcription to different extents, KK240,4AE mutant exhibited most profound effect on this step. These results suggest that the tri-lysine regions (211KELQKQITK and 236KGPAKLLWK) in the C-terminal of IN are important for HIV-1 reverse transcription and/or nuclear import. More studies are underway to further characterize the mechanisms involved in the action of these regions during early steps of HIV-1 replication.
Materials and methods
Construction of different IN expressors and HIV-1 RT/IN defective provirus
The full-length wild-type HIV-1 IN cDNA was amplified by polymerase chain reaction (PCR) using HIV-1 HxBru strain [30] as template and an engineered initiation codon (ATG) was placed prior to the first amino acid (aa) of IN. The primers are 5'-IN-HindIII-ATG (5'-GCGCAAGCTTGGATAGATGTTTTTAGATGGAA-3') and 3'-IN-Asp718 (5'-CCATGTGTGGTACCTCATCCTGCT-3'). The PCR product was digested with HindIII and Asp718 restriction enzymes and cloned in frame to 5' end of EYFP cDNA in a pEYFP-N1 vector (BD Biosciences Clontech) and generated a IN-YFP fusion expressor. Also, cDNA encoding for truncated IN (aa 50 to 288 or aa 1 to 212) was amplified by PCR and also cloned into pEYFP-N1 vector. The primers for generation of IN50-288 cDNA are IN50-HindIII-ATG-5'(5'– GCGCAAGCTTGGATAGATGCATGGACAAGTAG-3) and 3'-IN-Asp718 and primers for amplifying IN1-212 cDNA are IN-HindIII-ATG-5' and IN-212-XmaI-3'(5'-CAATTCCCGGGTTTGTATGTCTGTTTGC-3). IN substitution mutants INKK215,9AA-YFP, INKK240,4AE-YFP and INRK263,4AA-YFP, were generated by a two-step PCR-based method [43] by using a 5'-primer (5'-IN-HindIII-ATG), a 3'-primer (3'-IN-Asp718) and complementary primers containing desired mutations. Amplified IN cDNAs harboring specific mutations were then cloned into pEYFP-N1 vector. To improve the expression of each IN-YFP fusion protein, all IN-YFP fusing cDNAs were finally subcloned into a SVCMV vector, which contains a cytomegalovirus (CMV) immediate early gene promoter [43].
To construct HIV-1 RT/IN defective provirus NLlucΔBglΔRI, we used a previously described HIV-1 envelope-deleted NLlucΔBglD64E provirus as the backbone (kindly provided by Dr. Irvin S.Y. Chen). In this provirus, the nef gene was replaced by a firefly luciferase gene [33]. The ApaI/SalI cDNA fragment in NLlucBglD64E was replaced by the corresponding fragment derived from a HIV-1 RT/IN deleted provirus R-/ΔRI [32] and generated a RT/IN deleted provirus NLlucΔBglΔRI, in which RT and IN gene sequences were deleted while a 194-bp sequence harboring cPPT/CTS cis-acting elements was maintained. To restore HIV-1 envelope gene sequence in NLlucΔBglΔRI provirus, the SalI/BamHI cDNA fragment in this provirus was replaced by a corresponding cDNA fragment from a HIV-1 envelope competent provirus R-/ΔRI [32] and the resulting provirus is named as NLlucΔRI. To functionally complement RT/IN defects of NLlucΔBglΔRI, a CMV-Vpr-RT-IN fusion protein expressor [32] was used in this study. Co-transfection of NLlucΔBglΔRI, CMV-Vpr-RT-IN and a vesicular stomatitis virus G (VSV-G) glycoprotein expressor results in the production of VSV-G pseudotyped HIV-1 that can undergo for single cycle replication in different cell types [32]. To investigate the effect of IN mutants on viral replication, different mutants KK215,9AA, KK240.4AE, RK263,4AA or D64E were introduced into CMV-Vpr-RT-IN expressor by PCR-based method as described above and using a 5'-primer corresponding to a sequence in RT gene and including a natural NheI site (5'-GCAGCTAGCAGGGAGACTAA-3'), a 3'-primer (3'-IN-stop-PstI, 5'– CTGTTCCTGCAGCTAATCCTCATCCTG-3') and the complementary oligonucleotide primers containing desired mutations. All IN mutants were subsequently analyzed by DNA sequencing to confirm the presence of mutations or deletions.
Cell lines and reagents
Human embryonic kidney 293T, HeLa and HeLa-CD4-CCR5-β-Gal cells were maintained in Dulbecco's Modified Eagles Medium (DMEM) supplemented with 10% fetal calf serum (FCS). Human C8166 T-lymphoid cells were maintained in RPMI-1640 medium. Antibodies used in the immunofluorescent assay, immunoprecipitation or western blot are as follows: The HIV-1 positive human serum 162 and anti-HIVp24 monoclonal antibody used in this study were previously described [44]. The rabbit anti-GFP and anti-IN antibodies were respectively obtained from Molecular Probes Inc and through AIDS Research Reference Reagent Program, Division of AIDS, NIAID, NIH. Aphidicolin was obtained from Sigma Inc.
Cell transfection and immunofluorescence assay
DNA transfection in 293T and HeLa cells were performed with standard calcium phosphate DNA precipitation method. For immunofluorescence analysis, HeLa cells were grown on glass coverslip (12 mm2) in 24-well plate. After 48 h of transfection, cells on the coverslip were fixed with PBS-4% paraformaldehyde for 5 minutes, permeabilized in PBS-0.2% Triton X-100 for 5 minutes and incubated with primary antibodies specific for GFP or HIV-1 IN followed by corresponding secondary FITC-conjugated antibodies. Then, cells on the coverslip were viewed using a computerized Axiovert 200 inverted fluorescence microscopy (Becton Deckson Inc).
Virus production and infection
Production of different single-cycle replicating virus stocks and measurement of virus titer were previously described [32]. Briefly, 293T cells were co-transfected with RT/IN defective NLlucΔBglΔRI provius, a VSV-G expressor and each of CMV-Vpr-RT-IN (wt/mutant) expressor. To produce HIV-1 envelope competent single cycle replicating virus, 293T cells were co-transfected with NLlucΔRI and different CMV-Vpr-RT-IN (wt/mutant) expressors. After 48 hours of transfection, supernatants were collected and virus titers were quantified by RT activity assay [43].
To test the effect of IN mutants on virus infection, equal amounts of virus were used to infect HeLa-CCR5-CD4-β-Gal cells, dividing and non-dividing C8166 T cells. To compare the infection of each viral stock in HeLa-CCR5-CD4-β-Gal cells, numbers of infected cells (β-Gal positive cells) were evaluated by the MAGI assay 48 hours post-infection (p.i) as described previously [34]. To infect CD4+ T cells, dividing or aphidicolin-treated non-dividing C8166 T cells (with 1.3 μg/ml of aphidicolin) were infected with equivalent amounts of single cycle replicating viruses (5 cpm/cell) for 2 hours. Then, infected cells were washed and cultured in the absence or presence of the same concentration of aphidicolin. At 48 hours post-infection, 1 × 106 cells from each sample were collected, washed twice with PBS, lysed with 50 μl of luciferase lysis buffer (Fisher Scientific Inc) and then, 10 μl of cell lysate was subjected to the luciferase assay by using a TopCount®NXT™ Microplate Scintillation & Luminescence Counter (Packard, Meriden) and the luciferase activity was valued as relative luciferase units (RLU). Each sample was analyzed in duplicate and the average deviation was calculated.
Immunoprecipitation and Western blot analyses
For detection of IN-YFP fusion proteins, 293T cells transfected with each IN-YFP expressor were lysed with RIPA lysis buffer and immunoprecipitated using human anti-HIV serum. Then, immunoprecipitates were run in 12% SDS-PAGE and analyzed by Western blot using rabbit anti-GFP antibody. To analyze virion-incorporation of IN and virus composition, 293T cells were co-transfected with NLlucΔBglΔRI provirus and each of CMV-Vpr-RT-IN (wt/mutant) expressors. After 48 hours, viruses were collected, lysed with RIPA lysis buffer and immunoprecipitated with human anti-HIV serum. Then, immunoprecipitates were run in 12% SDS-PAGE and analyzed by Western blot with rabbit anti-IN antibody and anti-p24 monoclonal antibody or anti-HIV serum.
HIV-1 reverse-transcribed and integrated DNA detection by PCR and Southern blotting
C8166 T cells were infected with equal amount of the wt or IN mutant viruses for 2 hours, washed for three times and cultured in RPMI medium. To detect total viral DNA synthesis, at 12 hours post-infection, equal number (1 × 106 cells) of cells were collected, washed twice with PCR washing buffer (20 mM Tris-HCl, pH8.0, 100 mM KCl), and lysed in lysis buffer (PCR washing buffer containing 0.05% NP-40, 0.05% Tween-20). Lysates were then incubated at 56°C for 30 min with proteinase K (100 μg/ml) and at 90°C for 10 min prior to phenol-chloroform DNA purification. To detect viral cDNA from each sample, all lysates were serially diluted 5-fold and subjected to PCR analysis. The primers used to detect late reverse transcription products were as following: 5'-LTR-U3, 5'-GGATGGTGCTTCAAGCTAGTACC-3' (nt position 8807, +1 = start of BRU of transcription initiation); 3'-Gag 5'-ACTGACGCTCTCGCACCCATCTCTCTC-3' (nt position 329). The probe for southern blot detection was generated by PCR with a 5'-LTR-U5 oligonucleotide, 5'-CTCTAGCAGTGGCGCCCGAACAGGGAC-3' (nt position 173) and the 3'-Gag oligo. PCR was carried out using 1× HotStar Taq Master Mix kit (QIAGEN, Mississauga, Ontario), as described previously [32].
To analyze nucleus- and cytoplasm-associated viral DNA, a subcellular fractionation of infected C8166 T cells (2 × 106) was performed after 24 hours of infection, as described previously [37]. Briefly, infected cells were pelleted and resuspended in ice-cold PCR lysis buffer (washing buffer containing 0,1% NP-40). After a 5-min incubation on ice, the nucleus was pelleted by centrifugation, washed twice with PCR wash buffer, and lysed in lysis buffer (0,05% NP-40, 0,05% Tween-20). Then, both cytoplasmic sample (supernatant from the first centrifugation) and the nuclear sample were treated with proteinase K and used for PCR analysis, as described above.
Integrated proviral DNA was detected in cell lysates by a modified nested Alu-PCR [32], in which following the first PCR, a second PCR was carried-out to amplify a portion of the HIV-1 LTR sequence from the first Alu-LTR PCR-amplified products. The first PCR was carried out by using primers including 5'-Alu oligo (5'-TCCCAGCTACTCGGGAGGCTGAGG-3') and 3'-LTR oligo (5'-AGGCAAGCTTTATTGAGGGCTTAAGC-3') (nt position 9194) located respectively in the conserved region of human Alu sequence and in HIV-1 LTR. The primer used for both of the second nested PCR and for generating a probe are 5'-NI: 5'-CACACACAAGGCTACTTCCCT-3' and 3'-NI: 5'-GCCACTCCCCAGTCCCGCCC-3'. As a control, the first and second PCR primer pairs were also used in parallel to detect integrated viral DNA from serially diluted ACH-2 cells, which contain one viral copy/cell, in a background of uninfected C8166 cellular DNA.
To evaluate the DNA content of extracted chromosomal DNA preparations, detection of human β-globin gene was carried-out by PCR, as described previously [37]. All final PCR products were electrophoresed through 1.2% agarose gel and transferred to hybridization transfer membrane (GeneScreen Plus, PerkinElmer Life Sciences), subjected to Southern hybridization by using specific PCR DIG-Labeling probes (Roche Diagnostics, Laval, Que) and visualized by a chemiluminescent method. Densitometric analysis was performed using a Personal Molecular Imager (Bio-Rad) and Quantity One software version 4.1.
Authors' contributions
Z-J Ao designed and performed experiments, constructed most IN mutants and wrote the manuscript. KR Fowke provided technique support and critically evaluated the manuscript. EA Cohen participated in the design of the study and critically evaluated the manuscript. X-J Yao designed the study and coordinated it. All authors read and approved the final manuscript.
Acknowledgements
We would like to thank Nicole Rougeau, John Rutherford and Andres Finzi for their technical support. We also thank Dr. Irvin S.Y. Chen for kindly providing NLlucBglD64E provirus and Dr. Kevin Coombs for critical reading of the manuscript. We are also grateful to Drs. M. Emerman and D. Grandgenett for the HeLa-CD4-CCR5-β-Gal cells and anti-IN antiserum that were obtained through the AIDS Research Reference Reagent Program, Division of AIDS, NIAID, NIH. Eric A. Cohen is the recipient of the Canada Research Chair in Human Retrovirology. This work was supported by a Canadian Institutes of Health Research (CIHR) grant (HOP-63013) to X.J.Y.
==== Refs
Bukovsky A Gottlinger H Lack of integrase can markedly affect human immunodeficiency virus type 1 particle production in the presence of an active viral protease J Virol 1996 70 6820 6825 8794322
Engelman A Englund G Orenstein JM Martin MA Craigie R Multiple effects of mutations in human immunodeficiency virus type 1 integrase on viral replication J Virol 1995 69 2729 2736 7535863
Gallay P Hope T Chin D Trono D HIV-1 infection of nondividing cells through the recognition of integrase by the importin/karyopherin pathway Proc Natl Acad Sci U S A 1997 94 9825 9830 9275210 10.1073/pnas.94.18.9825
Ikeda T Nishitsuji H Zhou X Nara N Ohashi T Kannagi M Masuda T Evaluation of the functional involvement of human immunodeficiency virus type 1 integrase in nuclear import of viral cDNA during acute infection J Virol 2004 78 11563 11573 15479797 10.1128/JVI.78.21.11563-11573.2004
Nakamura T Masuda T Goto T Sano K Nakai M Harada S Lack of infectivity of HIV-1 integrase zinc finger-like domain mutant with morphologically normal maturation Biochem Biophys Res Commun 1997 239 715 722 9367835 10.1006/bbrc.1997.7541
Wu X Liu H Xiao H Conway JA Hehl E Kalpana GV Prasad V Kappes JC Human immunodeficiency virus type 1 integrase protein promotes reverse transcription through specific interactions with the nucleoprotein reverse transcription complex J Virol 1999 73 2126 2135 9971795
Piller SC Caly L Jans DA Nuclear import of the pre-integration complex (PIC): the Achilles heel of HIV? Curr Drug Targets 2003 4 409 429 12816349 10.2174/1389450033490984
Lee SP Xiao J Knutson JR Lewis MS Han MK Zn2+ promotes the self-association of human immunodeficiency virus type-1 integrase in vitro Biochemistry 1997 36 173 180 8993331 10.1021/bi961849o
Ellison V Gerton J Vincent KA Brown PO An essential interaction between distinct domains of HIV-1 integrase mediates assembly of the active multimer J Biol Chem 1995 270 3320 3326 7852418 10.1074/jbc.270.7.3320
Heuer TS Brown PO Photo-cross-linking studies suggest a model for the architecture of an active human immunodeficiency virus type 1 integrase-DNA complex Biochemistry 1998 37 6667 6678 9578550 10.1021/bi972949c
Engelman A Craigie R Identification of conserved amino acid residues critical for human immunodeficiency virus type 1 integrase function in vitro J Virol 1992 66 6361 6369 1404595
van Gent DC Groeneger AA Plasterk RH Mutational analysis of the integrase protein of human immunodeficiency virus type 2 Proc Natl Acad Sci U S A 1992 89 9598 9602 1409671
Eijkelenboom AP Lutzke RA Boelens R Plasterk RH Kaptein R Hard K The DNA-binding domain of HIV-1 integrase has an SH3-like fold Nat Struct Biol 1995 2 807 810 7552753 10.1038/nsb0995-807
Lutzke RA Plasterk RH Structure-based mutational analysis of the C-terminal DNA-binding domain of human immunodeficiency virus type 1 integrase: critical residues for protein oligomerization and DNA binding J Virol 1998 72 4841 4848 9573250
Wiskerchen M Muesing MA Human immunodeficiency virus type 1 integrase: effects of mutations on viral ability to integrate, direct viral gene expression from unintegrated viral DNA templates, and sustain viral propagation in primary cells J Virol 1995 69 376 386 7983732
Lu R Limon A Devroe E Silver PA Cherepanov P Engelman A Class II integrase mutants with changes in putative nuclear localization signals are primarily blocked at a postnuclear entry step of human immunodeficiency virus type 1 replication J Virol 2004 78 12735 12746 15542626 10.1128/JVI.78.23.12735-12746.2004
Cannon PM Byles ED Kingsman SM Kingsman AJ Conserved sequences in the carboxyl terminus of integrase that are essential for human immunodeficiency virus type 1 replication J Virol 1996 70 651 657 8523588
Petit C Schwartz O Mammano F The karyophilic properties of human immunodeficiency virus type 1 integrase are not required for nuclear import of proviral DNA J Virol 2000 74 7119 7126 10888652 10.1128/JVI.74.15.7119-7126.2000
Bouyac-Bertoia M Dvorin JD Fouchier RA Jenkins Y Meyer BE Wu LI Emerman M Malim MH HIV-1 infection requires a functional integrase NLS Molecular Cell 2001 7 1025 1035 11389849 10.1016/S1097-2765(01)00240-4
Depienne C Mousnier A Leh H Rouzic EL Dormont D Benichou S Dargemont C Characterization of the nuclear import pathway for HIV-1 integrase J Biol Chem 2001 276 18192 18107 10.1074/jbc.M009029200
Petit C Schwartz O Mammano F Oligomerization within virions and subcellular localization of human immunodeficiency virus type 1 integrase J Virol 1999 73 5079 5088 10233971
Pluymers W Cherepanov P Schols D Clercq ED Debyser Z Nuclear localization of human immunodeficiency virus type 1 integrase expressed as a fusion protein with green fluorescent protein Virology 1999 258 327 332 10366569 10.1006/viro.1999.9727
Tsurutani N Kubo M Maeda Y Ohashi T Yamamoto N Kannagi M Masuda T Identification of critical amino acid residues in human immunodeficiency virus type 1 IN required for efficient proviral DNA formation at steps prior to integration in dividing and nondividing cells J Virol 2000 74 4795 4806 10775618 10.1128/JVI.74.10.4795-4806.2000
Dvorin JD Bell P Maul GG Yamashita M Emerman M Malim MH Reassessment of the roles of integrase and the central DNA flap in human immunodeficiency virus type 1 nuclear import J Virol 2002 76 12087 12096 12414950 10.1128/JVI.76.23.12087-12096.2002
Limon A Devroe E Lu R Ghory HZ Silver PA Engelman A Nuclear localization of human immunodeficiency virus type 1 preintegration complexes (PICs): V165A and R166A are pleiotropic integrase mutants primarily defective for integration, not PIC nuclear import J Virol 2002 76 10598 10607 12368302 10.1128/JVI.76.21.10598-10607.2002
Kukolj G Jones MS Skalka M Subcellular localization of avian sarcoma virus and human immunodefiiency virus type 1 integrase. J Virol 1997 71 843 847 8985428
Devroe E Engelman A Silver PA Intracellular transport of human immunodeficiency virus type 1 integrase J Cell Sci 2003 116 4401 4408 13130095 10.1242/jcs.00747
Maertens G Cherepanov P Pluymers W Busschots K De Clercq E Debyser Z Engelborghs Y LEDGF/p75 is essential for nuclear and chromosomal targeting of HIV-1 integrase in human cells J Biol Chem 2003 278 33528 33539 12796494 10.1074/jbc.M303594200
Llano M Vanegas M Fregoso O Saenz D Chung S Peretz M Poeschla EM LEDGF/p75 determines cellular trafficking of diverse lentiviral but not murine oncoretroviral integrase proteins and is a component of functional lentiviral preintegration complexes J Virol 2004 78 9524 9537 15308744 10.1128/JVI.78.17.9524-9537.2004
Yao XJ Subbramanian RA Rougeau N Boisvert F Bergeron D Cohen EA Mutagenic analysis of human immunodeficiency virus type 1 Vpr: role of a predicted N-terminal alpha-helical structure in Vpr nuclear localization and virion incorporation J Virol 1995 69 7032 7044 7474123
Kuiken C Foly B Hahn B Marx P McCutchan F Mellors J Wolinsky SM Korber B HIV Sequence Compendium 2001 Los Alamos National Laboratory 2001
Ao Z Yao X Cohen EA Assessment of the role of the central DNA flap in human immunodeficiency virus type 1 replication by using a single-cycle replication system J Virol 2004 78 3170 3177 14990738 10.1128/JVI.78.6.3170-3177.2004
Poon B Chen IS Human immunodeficiency virus type 1 (HIV-1) Vpr enhances expression from unintegrated HIV-1 DNA J Virol 2003 77 3962 3972 12634356 10.1128/JVI.77.7.3962-3972.2003
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 b-galactosidase gene. J Virol 1992 66 2232 2239 1548759
Brussel A Sonigo P Evidence for gene expression by unintegrated human immunodeficiency virus type 1 DNA species J Virol 2004 78 11263 11271 15452245 10.1128/JVI.78.20.11263-11271.2004
Wu Y Marsh JW Selective transcription and modulation of resting T cell activity by preintegrated HIV DNA Science 2001 293 1503 1506 11520990 10.1126/science.1061548
Simon JHM Malim MH The human immunodeficiency virus type 1 Vif protein modulates the postpenetration stability of viral nucleoprotein complexes J Virol 1996 70 5297 5305 8764040
Nakajima N Lu R Engelman A Human immunodeficiency virus type 1 replication in the absence of integrase-mediated dna recombination: definition of permissive and nonpermissive T-cell lines J Virol 2001 75 7944 7955 11483739 10.1128/JVI.75.17.7944-7955.2001
Wu X Liu H Xiao H Conway JA Hunter E Kappes JC Functional RT and IN incorporated into HIV-1 particles independently of the Gag/Pol precursor protein Embo J 1997 16 5113 5122 9305652 10.1093/emboj/16.16.5113
Hehl EA Joshi P Kalpana GV Prasad VR Interaction between human immunodeficiency virus type 1 reverse transcriptase and integrase proteins J Virol 2004 78 5056 5067 15113887 10.1128/JVI.78.10.5056-5067.2004
Zhu K Dobard C Chow SA Requirement for integrase during reverse transcription of human immunodeficiency virus type 1 and the effect of cysteine mutations of integrase on its interactions with reverse transcriptase J Virol 2004 78 5045 5055 15113886 10.1128/JVI.78.10.5045-5055.2004
Lu R Ghory HZ Engelman A Genetic analyses of conserved residues in the carboxyl-terminal domain of human immunodeficiency virus type 1 integrase J Virol 2005 79 10356 10368 16051828 10.1128/JVI.79.16.10356-10368.2005
Yao XJ Kobinger G Dandache S Rougeau N Cohen E HIV-1 Vpr-chloramphenicol acetyltransferase fusion proteins: sequence requirement for virion incorporation and analysis of antiviral effect Gene Ther 1999 6 1590 1599 10490769 10.1038/sj.gt.3300988
Yao XJ Mouland AJ Subbramanian RA Forget J Rougeau N Bergeron D Cohen EA Vpr stimulates viral expression and induces cell killing in human immunodeficiency virus type 1-infected dividing Jurkat T cells J Virol 1998 72 4686 4693 9573232
|
16232319
|
PMC1277849
|
CC BY
|
2021-01-04 16:36:39
|
no
|
Retrovirology. 2005 Oct 18; 2:62
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-62
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-1131621267510.1186/1465-9921-6-113ResearchAsthma and COPD in cystic fibrosis intron-8 5T carriers. A population-based study Dahl Morten [email protected]ærg-Hansen Anne [email protected] Peter [email protected] Børge G [email protected] Department of Clinical Biochemistry, Herlev University Hospital, DK-2730 Herlev, Denmark2 Department of Clinical Biochemistry, Rigshospitalet, Copenhagen University Hospital, DK-2100 Copenhagen Ø, Denmark3 Department of Respiratory Medicine, Hvidovre University Hospital, DK-2650, Hvidovre, Denmark4 The Copenhagen City Heart Study, Bispebjerg University Hospital, DK-2200 Copenhagen N, Denmark2005 9 10 2005 6 1 113 113 17 4 2005 9 10 2005 Copyright © 2005 Dahl et al; licensee BioMed Central Ltd.2005Dahl 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
Carriers of cystic fibrosis intron-8 5T alleles with high exon-9 skipping could have increased annual lung function decline and increased risk for asthma or chronic obstructive pulmonary disease (COPD).
Methods
We genotyped 9131 individuals from the adult Danish population for cystic fibrosis 5T, 7T, 9T, and F508del alleles, and examined associations between 11 different genotype combinations, and annual FEV1 decline and risk of asthma or COPD.
Results
5T heterozygotes vs. 7T homozygous controls had no increase in annual FEV1 decline, self-reported asthma, spirometry-defined COPD, or incidence of hospitalization from asthma or COPD. In 5T/7T heterozygotes vs. 7T homozygous controls we had 90% power to detect an increase in FEV1 decline of 8 ml, an odds ratio for self-reported asthma and spirometry-defined COPD of 1.9 and 1.7, and a hazard ratio for asthma and COPD hospitalization of 1.8 and 1.6, respectively. Both 5T homozygotes identified in the study showed evidence of asthma, while none of four 5T/F508del compound heterozygotes had severe pulmonary disease. 7T/9T individuals had annual decline in FEV1 of 19 ml compared with 21 ml in 7T homozygous controls (t-test:P = 0.03). 6.7% of 7T homozygotes without an F508del allele in the cystic fibrosis transmembrane conductance regulator gene reported asthma vs. 11% of 7T/9T individuals with an F508del allele (χ2:P = 0.01) and 40% of 7T homozygotes with an F508del allele (P = 0.04). 7T homozygotes with vs. without an F508del allele also had higher incidence of asthma hospitalization (log-rank:P = 0.003); unadjusted and adjusted equivalent hazard ratios for asthma hospitalization were 11 (95%CI:1.5–78) and 6.3 (0.84–47) in 7T homozygotes with vs. without an F508del allele.
Conclusion
Polythymidine 5T heterozygosity is not associated with pulmonary dysfunction or disease in the adult Caucasian population. Furthermore, our results support that F508del heterozygosity is associated with increased asthma risk independently of the 5T allele.
==== Body
Background
Asthma and chronic obstructive pulmonary disease (COPD) are caused by complex interactions between environmental and genetic factors. A putative genetic risk factor for asthma and COPD is the cystic fibrosis transmembrane conductance regulator (CFTR) gene [1-3]. This gene encodes a cAMP-regulated channel with chloride activity in pulmonary epithelia. When channel activities are absent, cystic fibrosis with life-threatening airways obstruction due to thickened secretions and secondary pulmonary infection develop [4]. The most common cause of cystic fibrosis is homozygosity for the phenylalanine-508 deletion (F508del), explaining about 70% of cystic fibrosis worldwide [4,5].
We previously showed that persons heterozygous for a F508del deletion are overrepresented among people with asthma [1,6]. Another more common variant, the 5T allele, could likewise be involved in asthma [7] or COPD. This variation is in the polythymidine tract of the CFTR gene and has mainly been associated with congenital bilateral absence of the vas deferens, a monosymptomatic form of cystic fibrosis [8-10]. However, it may also be associated with increased risk of obstructive lung disease, particularly bronchiectasis [9-14]. Because most previous studies on lung disease in 5T carriers were based on case patients [2,9-24], currently we know little about the risk for obstructive lung disease in 5T carriers in the general population.
Three common alleles are known in the polythymidine tract, 5T, 7T, and 9T. The polythymidine tract is situated in intron-8 near the acceptor splice site for exon-9 [25,26]. The shorter this polythymidine tract is, the more often exon-9 is skipped from CFTR mRNA. Transcripts missing exon-9 increases from 1%–13% in 9T homozygotes [27-29] to 12%–25% in 7T homozygotes [13,27-30] to 66%–90% in 5T homozygotes [13,27,31,32]. CFTR mRNA without exon-9 leads to a protein with no chloride channel activity [33,34]. Thus, carriers of 5T with high exon-9 skipping have reduced channel activities and could have increased susceptibility for obstructive lung disease. This could be particularly relevant for 5T carriers exposed to additional risk factors for lung disease such as tobacco smoke or familial predisposition to lung disease. Variations in the genes for mannose-binding lectin and α1-antitrypsin have been studied as modifiers of cystic fibrosis lung disease [35-37] and could also potentially influence risk of lung disease in 5T heterozygotes. Allele frequencies in whites are approximately 5% for the 5T allele, 84% for 7T, and 11% for 9T [25,26].
We hypothesised that carriers of the 5T allele have increased annual lung function decline and increased risk for asthma or COPD. To test this hypothesis, we genotyped 9131 individuals from the adult Danish population for the 5T, 7T, and 9T alleles in the CFTR gene. We combined polythymidine and F508del genotypes [1], and examined associations between 11 different genotype combinations, and annual FEV1 decline and risk of asthma or COPD. We also examined whether other common risk factors for lung disease or variations in the genes for mannose-binding lectin and α1-antitrypsin significantly add to risk of lung disease in 5T carriers.
Methods
Subjects participated in the 1976–78, 1981–83, and/or 1991–94 examination of the Copenhagen City Heart Study, a prospective epidemiological study initiated in 1976–78 [38]. Participants aged 20 years and above were selected randomly after age stratification into 5-year age groups from among residents of Copenhagen. Of the 17180 individuals invited, 10135 participated, 9259 gave blood, and 9131 were genotyped for the polythymidine tract variants of the cystic fibrosis conductance membrane regulator (CFTR) gene. Details of study procedures and some characteristics of non-responders are described elsewhere [38,39]. More than 99% were Whites of Danish descent. All participants gave written informed consent, and Herlev University Hospital and the ethics committee for Copenhagen and Frederiksberg approved the study (# 100.2039/91).
Participants filled out a self-administered questionnaire, which was validated by the participant and an investigator on the day of attendance. Participants reported on long-term occupational exposure to dust or welding fumes, pulmonary symptoms (dyspnea, wheezing, bringing up phlegm), familial predisposition to asthma (having at least one sibling with asthma), smoking habits (current smoker, ex-smoker, never-smoker), type of smoking and daily tobacco consumption. An estimate of life-time tobacco exposure (in packyears) was calculated as: daily tobacco consumption (g) times duration of smoking (years) divided by 20 (g/pack). If at least once during the study period participants aswered "Yes" to the question "Do you suffer from asthma?", we recorded they had self-reported asthma. Medication for asthma / bronchitis was "Yes" to the question "Do you daily take medication for asthma / bronchitis?" Additional information on hospitalizations due to asthma (ICD8: 493; ICD10: J45–46) and COPD (ICD8: 491–492; ICD10: J41–44) was drawn from the Danish National Hospital Discharge Register from May 1st 1976 through December 31st 2000. We confirmed in the Danish National Hospital Discharge Register covering all hospital discharges in Denmark, that no participants in the sample were ever hospitalized for cystic fibrosis.
Forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) were measured with an electronic spirometer (model N403, Monaghan, Littleton, Colo.) at the 1976–78 and 1981–83 examinations and with a dry wedge spirometer (Vitalograph, Maidenhead, UK) at the 1991–94 examination. At each examination, three sets of values were obtained, and as a criterion for correct performance of the procedure, at least two measurements of FEV1 and FVC differing by less than 5% had to be produced. The highest set of FEV1 and FVC were used in the analyses as percentage of predicted value using internally derived reference values based on a subsample of healthy never smokers [40]. Annual decline in FEV1 (ml/year) was calculated as FEV1 (ml) obtained at the latest measurement minus the FEV1 value obtained at the first measurement, times 365.25 divided by the number of days between the two measurements (in years-1). Spirometry defined COPD was FEV1<80% predicted and FEV1/FVC<0.7, excluding self-reported asthma [41].
We amplified the polythymidine tract variants in intron-8 by nested polymerase chain reaction using the primerpairs: 5'-TAATGGATCATGGGCCATGT-3'and 5'-ACAGTGTTGAATGTGGTGCA-3' (first step reaction), and 5'-CCGCCGCTGTGTGTGTGTGTGTGTTTTT-3' and 5'GCTTTCTCAAATAATTCCCC-HEX-3' (second step reaction) (mismatch underlined) [8]. Products of 52 bp (5T allele), 53 bp (6T allele), 54 bp (7T allele), and 56 bp (9T allele) were seperated by capillary electrophoresis on an ABI 310 sequenator. Tamra 350 marker was added to samples before analysis, and each analysis ran dummy standard, water control, and positive controls. The F508del allele in the CFTR gene [1], S and Z alleles in the Serine Protease Inhibitor-A1 gene [42], and B, C, and D alleles in the Mannose-Binding Lectin-2 gene [43] were identified using polymerase chain reaction followed by restriction enzyme digestion as described. Diagnoses of polythymidine alleles in 5T/F508del genotypes, 5T/5T, 6T/7T, and 69 randomly selected 5T/9T, 7T/9T, 7T/7T, 5T/7T, 9T/9T genotypes were confirmed by sequencing. All 7T/7T F508del genotypes were re-analyzed to confirm their diagnosis, using sequencing (7T/7T) and RFLP-PCR (F508del). The number of TG repeats adjacent to the 5T allele in 5T/F508del and 5T/5T genotypes were determined by sequencing. For each polythymidine allele, expected exon-9 skipping was half the middle value of the ranges of skipping observed in homozygotes [32]; expected exon-9 skipping was not estimated in individuals with F508del heterozygosity.
Linkage disequilibrium between the 9T and F508del alleles was tested by the linkage utility program "EH" , which estimates allele and haplotype frequencies with and without allelic association. The linkage disequilibrium coefficient D was calculated as D = P22 - p2q2, where P22 is the observed frequency of the 9T/F508del haplotype, p2 is the frequency of the F508del allele in the general population and q2 is the population frequency of the 9T allele. The degree of linkage disequilibrium was expressed as D' = D/Dmax × 100%.
Statistical analysis was performed with SPSS; for power calculations, NCSS-PASS and StatMate were used. P < 0.05 on a two-sided test was considered significant. Pearson's χ2-test or analysis of variance (ANOVA) was used for overall comparisons between several genotypes; Pearson's or Fisher's Exact χ2-test were used for post-hoc two-genotype comparisons. The most common genotype combination in the population, 7T homozygosity without F508del, was used as reference group for statistical comparisons. We evaluated asthma and COPD prevalences between genotypes using unadjusted and adjusted logistic regression with Wald's test as a measure of significance; the adjusted model included gender, age at study entry (deciles), and packyears at study entry (never smokers and deciles). We evaluated asthma and COPD incidences between genotypes using the log-rank test [42-44]. Unadjusted and adjusted Cox regression with forced entry examined time to disease by using hazard ratios (relative risks) and 95% confidence intervals; the adjusted model included gender, age at study entry (deciles), tobacco use during follow-up (never smokers and deciles), and FEV1 % predicted at study entry (deciles). We tested possible interactions between the 5T/7T genotype and smoking habits, long-term occupational exposure to dust or welding fumes, familial predisposition to asthma, α1-antitrypsin MS genotype, α1-antitrypsin MZ genotype, or mannose-binding lectin deficiency in predicting FEV1 at study entry in ANCOVA models.
Results
Characteristics of participants are given in Table 1; genotypes are ordered according to predicted increased skipping of exon-9 of the cystic fibrosis transmembrane conductance regulator gene, stratified for presence or absence of F508del heterozygosity. Among the 9,131 participants selected randomly from the Danish general population, 352 (3.9%) were 5T heterozygotes and 249 (2.7%) were F508del heterozygotes. Expected numbers of 5T and F508del heterozygotes according to the Hardy Weinberg equilibrium were 349 and 246, respectively. Allele frequencies did not differ from those predicted by the Hardy Weinberg equilibrium (χ2-test for 7T allele: P = 0.84; 9T allele: P = 0.60; 6T allele: P = 0.98; 5T allele: P = 0.42; F508del allele: P = 0.19). The novel intron-8 polythymidine tract variant, the 6T allele [45], was identified in four individuals. The 9T and F508del alleles were in linkage disequilibrium with a degree of linkage of 98% (χ2-test: P < 0.001).
Table 1 Characteristics of subjects by intron-8 polythymidine tract and F508del genotype
Polythymidine 9T/9T 7T/9T 7T/7T 6T/7T 5T/9T 5T/7T 5T/5T 9T/9T 7T/9T 7T/7T 5T/9T
Expected exon-9 skipping, % 7 13 18 ≥18 43 48 78 - - - -
F508del heterozygosity yes yes yes yes P-value
Women / Men 44 / 39 841 / 699 3,818 / 3,087 2 / 2 22 / 18 171 / 137 1 / 1 13 / 10 127 / 90 4 / 1 2 / 2 0.99
Genotype frequency, % 0.9 16.9 75.6 0.0 0.4 3.4 0.0 0.3 2.4 0.1 0.0
Smoking before study entry, packyears* 16 ± 2.1 16 ± 0.5 15 ± 0.2 13 ± 10 18 ± 3.0 14 ± 1.1 8.4 ± 12 13 ± 4.0 14 ± 1.3 18 ± 10 14 ± 10 0.81
Age at study entry, years 46 ± 1.4 47 ± 0.3 47 ± 0.2 46 ± 6.3 47 ± 2.0 46 ± 0.7 39 ± 8.9 48 ± 2.6 48 ± 0.9 41 ± 5.6 46 ± 6.3 0.63
FEV1 at study entry, %pred. 87 ± 1.9 90 ± 0.4 90 ± 0.2 83 ± 8.8 96 ± 2.8 90 ± 1.0 84 ± 12 94 ± 3.7 89 ± 1.2 84 ± 7.9 101 ± 8.8 0.24
Smoking during follow-up, g/day† 9.0 ± 1.1 8.8 ± 0.3 8.9 ± 0.1 11 ± 5.0 8.1 ± 1.6 7.5 ± 0.6 6.3 ± 7.1 7.9 ± 2.1 7.1 ± 0.7 8.0 ± 4.5 8.0 ± 5.0 0.24
Follow-up, years 23 ± 0.14 23 ± 0.03 23 ± 0.02 23 ± 0.66 23 ± 0.21 23 ± 0.08 24 ± 0.93 23 ± 0.27 23 ± 0.09 24 ± 0.59 24 ± 0.66 0.97
Values are number of individuals, percentages, or mean ± SD. P-values by Pearson's χ2 test or analysis of variance. *Calculated as daily tobacco use (g/day) × duration of smoking (years) / 20 (g/pack). †The average amount of tobacco used (in g/day) at the different examinations attended.
Annual decline in FEV1
Annual decline in FEV1 did not differ between 5T heterozygotes or homozygotes vs. 7T homozygous controls (Fig. 1). 7T/9T individuals had annual decline in FEV1 of 19 ml compared with 21 ml in 7T homozygous controls (t-test: P = 0.03; Fig. 1). None of the other genotype combinations differed from 7T homozygous controls. The analysis had 90% power to detect differences in annual FEV1 decline of 14 ml in 9T/9T, 3.8 ml in 7T/9T, 61 ml in 6T/7T, 23 ml in 5T/9T, 8 ml in 5T/7T, 31 ml in 9T/9T F508del, 9 ml in 7T/9T F508del, 72 ml in 7T/7T F508del, and 72 ml in 5T/9T F508del individuals vs. 7T homozygous controls.
Figure 1 Annual FEV1 decline by intron-8 polythymidine tract and F508del genotype. Values are mean and SEM. *P = 0.03 compared with 7T homozygotes without F508del.
Asthma
Prevalence of self-reported asthma did not differ between 5T heterozygotes or homozygotes vs. 7T homozygous controls (Ps ≥ 0.10; data not depicted). However, self-reported asthma differed between genotypes overall (χ2: P = 0.02); eleven percent of 7T/9T individuals with F508del (χ2: P = 0.01) and 40% of 7T homozygotes with F508del (χ2: P = 0.04) had asthma vs. 6.7% of 7T homozygous controls (data not depicted). None of the other genotype combinations differed from 7T homozygous controls.
Unadjusted odds ratios for self-reported asthma were 1.7 (95%CI:1.1–2.7) in 7T/9T individuals with F508del and 9.2 (1.5–55) in 7T homozygotes with F508del vs. 7T homozygous controls (Fig. 2, upper panel). After adjusting for gender, age at study entry, and packyears at study entry, equivalent odds ratios for self-reported asthma were 1.7 (1.0–27) in 7T/9T individuals with F508del and 27 (2.2–327) in 7T homozygotes with F508del (Fig. 2, lower panel). The analysis had 90% power to detect an odds ratio for asthma of 3.0 for 9T/9T, 1.4 for 7T/9T, 23 for 6T/7T, 4.2 for 5T/9T, 1.9 for 5T/7T, 5.8 for 9T/9T F508del, 2.1 for 7T/9T F508del, 18 for 7T/7T F508del, and 23 for 5T/9T F508del individuals vs. 7T homozygous controls.
Figure 2 Odds ratios for self-reported asthma by intron-8 polythymidine tract and F508del genotype. 7T homozygotes without F508del was used as reference group. The adjusted model included gender, age at study entry, and packyears at study entry. Error bars are 95% confidence intervals. Self-reported asthma = "Yes" at least once during the study period to the question "Do you suffer from asthma?".
Incidence of hospitalization from asthma during 24 years follow-up did not differ between 5T heterozygotes or homozygotes versus 7T homozygous controls (Table 2). However, incidence of asthma hospitalization was increased in 7T homozygotes with F508del compared with 7T homozygous controls (Table 2). Unadjusted and after adjusting for gender, age at study entry, tobacco consumption, and FEV1 % predicted at study entry, the hazard ratio for asthma hospitalization was 11 (1.5–78) and 6.3 (0.84–47) in 7T homozygotes with F508del vs. 7T homozygous controls. None of the other genotype combinations differed from 7T homozygous controls (Table 2). The analysis had 90% power to detect a hazard ratio for asthma hospitalization of 2.7 for 9T/9T, 1.4 for 7T/9T, 15 for 6T/7T, 3.7 for 5T/9T, 1.8 for 5T/7T, 4.9 for 9T/9T F508del, 2.0 for 7T/9T F508del, 13 for 7T/7T F508del, and 15 for 5T/9T F508del individuals vs. 7T homozygous controls.
Table 2 Incidences and hazard ratios for asthma hospitalisation by intron-8 polythymidine tract and F508del genotype during 24 years follow-up
Poly-T Expected exon-9 skipping, % F508del heterozygosity n Incidence n/10000 person-years P-value* Unadjusted HR (95%CI) Adjusted† HR (95%CI) 90% power‡ HR
9T/9T 7 83 9.8 0.83 1.2 (0.28–4.7) 1.1 (0.27–4.4) 2.7
7T/9T 13 1540 9.3 0.60 1.1 (0.76–1.6) 1.1 (0.77–1.6) 1.4
7T/7T 18 6905 8.4 - 1.0 1.0 -
6T/7T ≥18 4 0 0.77 - - 15
5T/9T 43 40 10 0.85 1.2 (0.17–8.6) 1.2 (0.17–8.9) 3.7
5T/7T 48 308 5.3 0.35 0.63 (0.23–1.7) 0.53 (0.17–1.7) 1.8
5T/5T 78 2 0 0.84 - - 25
9T/9T - yes 23 0 0.49 - - 4.9
7T/9T - yes 217 11 0.47 1.3 (0.59–3.1) 1.3 (0.55–2.9) 2.0
7T/7T - yes 5 87 0.003 11 (1.5–78) 6.3 (0.84–47) 13
5T/9T - yes 4 0 0.77 - - 15
*P-values are for the comparison with 7T/7T individuals without the F508del deletion by log-rank test. †Cox regression adjusted for gender, age at study entry, tobacco use during follow-up, and FEV1 % predicted at study entry. ‡90% power to detect a hazard ratio (HR) of asthma at 2-sided P < 0.05. 95%CI = 95% confidence interval. Hospitalizations from asthma (ICD8: 493; ICD10: J45–46) were drawn from the Danish National Discharge Register from 1976 through 2000.
Chronic obstructive pulmonary disease (COPD)
Prevalence of spirometry defined COPD did not differ between 5T heterozygotes or homozygotes vs. 7T homozygous controls (Ps ≥ 0.22) and did not differ between genotypes overall (χ2: P = 0.51) (data not depicted). Unadjusted and adjusted odds ratios for spirometry defined COPD did not differ between genotypes (Fig. 3). The analysis had 90% power to detect an odds ratio for COPD of 2.5 for 9T/9T, 1.3 for 7T/9T, 19 for 6T/7T, 3.4 for 5T/9T, 1.7 for 5T/7T, 4.6 for 9T/9T F508del, 1.8 for 7T/9T F508del, 15 for 7T/7T F508del, and 19 for 5T/9T F508del individuals vs. 7T homozygous controls.
Figure 3 Odds ratios for spirometry defined COPD by intron-8 polythymidine tract and F508del genotype. 7T homozygotes without F508del was used as reference group. The adjusted model included gender, age at study entry, and packyears at study entry. Error bars are 95% confidence intervals. COPD = FEV1<80% predicted and FEV1/FVC<0.7, excluding self-reported asthma.
Incidence of hospitalization from COPD during 24 years follow-up was reduced in 5T/7T individuals vs. 7T homozygous controls (Table 3). Unadjusted and after adjusting for gender, age at study entry, tobacco consumption and FEV1 % predicted at study entry, the hazard ratio for COPD was 0.47 (0.23–0.95) and 0.49 (0.23–1.0) in 5T/7T individuals vs. 7T homozygous controls (Table 3). There was a trend toward increased incidence of COPD hospitalization in 6T/7T individuals; unadjusted and adjusted hazard ratio for COPD hospitalization was 4.9 (0.69–35) and 7.6 (1.0–55) in 6T/7T individuals vs. 7T homozygous controls (Table 3). Other genotypes did not differ in COPD risk from 7T homozygous controls. The analysis had 90% power to detect a hazard ratio for COPD of 2.3 for 9T/9T, 1.3 for 7T/9T, 11 for 6T/7T, 3.0 for 5T/9T, 1.6 for 5T/7T, 3.8 for 9T/9T F508del, 1.7 for 7T/9T F508del, 9.7 for 7T/7T F508del, and 11 for 5T/9T F508del individuals vs. 7T homozygous controls.
Table 3 Incidences and hazard ratios for COPD hospitalisation by intron-8 polythymidine tract and F508del genotype during 24 years follow-up
Poly-T Expected exon-9 skipping, % F508del heterozygosity n Incidence n/10000 person-years P-value* Unadjusted HR (95%CI) Adjusted† HR (95%CI) 90% power‡ HR
9T/9T 7 83 40 0.10 1.8 (0.89–3.6) 1.7 (0.85–3.5) 2.3
7T/9T 13 1540 21 0.70 0.95 (0.75–1.2) 0.99 (0.78–1.3) 1.3
7T/7T 18 6905 22 - 1.0 1.0 -
6T/7T ≥18 4 105 0.08 4.9 (0.69–35) 7.6 (1.0–55) 11
5T/9T 43 40 21 0.90 0.92 (0.23–3.7) 0.75 (0.19–3.0) 3.0
5T/7T 48 308 11 0.03 0.47 (0.23–0.95) 0.49 (0.23–1.0) 1.6
5T/5T 78 2 0 0.73 - - 19
9T/9T - yes 23 0 0.25 - - 3.8
7T/9T - yes 217 25 0.73 1.1 (0.63–1.9) 1.1 (0.62–1.9) 1.7
7T/7T - yes 5 0 0.59 - 9.7
5T/9T - yes 4 0 0.63 - - 11
*P-values are for the comparison with 7T/7T individuals without the F508del deletion by log-rank test. †Cox regression adjusted for gender, age at study entry, tobacco use during follow-up, and FEV1 % predicted at study entry. ‡90% power to detect a hazard ratio (HR) of COPD at 2-sided P < 0.05. 95%CI = 95% confidence interval. Hospitalizations from COPD (ICD8: 491–492; ICD10: J41–44) were drawn from the Danish National Discharge Register from 1976 through 2000.
5T homozygotes and 5T/F508del compound heterozygotes
One of two 5T homozygous smokers reported having asthma and took daily medication for respiratory disease (Table 4). The other homozygous individual showed evidence of airway obstruction with reversibility and was referred for further examination and treatment of asthma. None of four 5T/F508del compound heterozygotes had clinical signs of severe pulmonary disease (Table 4).
Table 4 Pulmonary status of 5T homozygotes and 5T/F508del compound heterozygotes sampled from the general population
Poly-T* F508del heterozygosity Age Gender Smoking status FEV1 Self-reported asthma‡ Medication for asthma / bronchitis¶ Hospitalization Often bothered by
years %predicted reversibility† asthma** COPD** dyspnoea wheezing phlegm
TG12-5T/TG12-5T 32 M current smoker 92 - yes yes no no yes yes no
TG11-5T/TG11-5T 62 F current smoker 67 30% no no no no no no no
TG11-5T yes 33 F current smoker 115 - no no no no no no no
TG11-5T yes 62 M never smoker 121 - no no no no no no no
TG12-5T yes 65 F ex-smoker 79 - no no no no no no no
TG11-5T yes 70 M current smoker 128 - no no no no no no no
*Number of TG repeats adjacent to the polythymidine tract included. †FEV1 30 minutes after inhalation of 0.5 mg terbutaline minus FEV1 at 0 minutes divided by FEV1 at 0 minutes times 100%; only individuals with FEV1/FVC<0.7 were tested for FEV1 reversibility. ‡"Yes" to "Do you suffer from asthma?" ¶"Yes" to "Do you daily take medication for asthma / bronchitis?" **Hospitalizations from asthma (ICD8: 493; ICD10: J45–46) and COPD (ICD8: 491–492; ICD10: J41–J44) were drawn from the Danish National Discharge Register from 1976 through 2000.
Context-dependent associations for 5T/7T genotype
There was no interaction between 5T/7T genotype and smoking status (P = 0.78), occupational exposure to dust or welding fumes (P = 0.10), familial asthma (P = 0.37), α1-antitrypsin MS genotype (P = 0.64), α1-antitrypsin MZ genotype (P = 0.47), or mannose-binding lectin deficiency (P = 0.73) in predicting FEV1 % predicted at study entry.
Discussion
This study shows that polythymidine 5T heterozygosity is not associated with increased annual decline in FEV1 or risk of asthma or COPD in the adult Caucasian population; these results are independent of age, gender, tobacco smoking, and other potential confounders. Interestingly, however, both 5T homozygotes showed evidence of asthma. Furthermore, our results support that F508del heterozygosity is associated with increased asthma risk independently of the 5T allele.
Because 1 in 26 carries a 5T allele in this population, it is indeed important that 5T heterozygosity does not increase risk of obstructive lung disease in the population at-large. It appears that the 5T allele causes lung disease only in very rare circumstances [9-14], leaving the average heterozygous individual unaffected by obstructive lung disease. Previous results suggest that penetrance of pulmonary manifestations in 5T carriers might depend on the length of an adjacent TG repeat [46,47]. This could be particularly relevant for 5T homozygotes and compound heterozygotes. In 5T heterozygotes, however, longer TG repeats seem less likely to affect risk of pulmonary disease. This is because 5T heterozygosity was not associated with risk of lung disease in this study although predicted TG12 and TG13 allele frequency in 5T carriers in our population was 31% [47]. Other additional genetic variations have also been shown to influence exon-9 skipping in 5T carriers, but to a lesser degree than the TG repeat.
Because all 5T/F508del compound heterozygotes were free from severe pulmonary disease, the 5T allele did not appear to explain our previous results [1,6] suggesting that F508del heterozygosity may be overrepresented among asthmatics. A few recent studies also support this observation [2,19,48], while others have found no [20,21,49] or negative associations [50]. In the present analyses, 7T/9T and 7T/7T individuals with F508del heterozygosity had higher prevalences of self-reported asthma, and 7T/7T individuals with F508del heterozygosity also had higher incidence of hospitalization from asthma. F508del heterozygosity was only associated with increased asthma risk in individuals without the 5T allele, indicating that our previous observations are independent of influence from this allele. In addition, both 5T homozygotes showed evidence of asthma supporting the hypothesis that CFTR variations may be associated with asthma [2,19].
To identify factors in the population that significantly add to risk of lung disease in 5T heterozygotes, we tested for interactions between 5T/7T genotype and potential risk factors for lung disease, but found no significant interactions. Garred [35] and coworkers found a worse prognosis in cystic fibrosis patients with MBL deficiency. We were not able to extend this finding, since lung function in 5T or F508del heterozygotes was not reduced by MBL deficiency. Previous studies by Mahadeva [36] and Frangolias [37] showed that pulmonary disease severity in cystic fibrosis patients were unaffected by α1-antitrypsin S and Z alleles. In line with this, we also observed no increased risk for pulmonary dysfunction in 5T carriers with α1-antitrypsin MS or MZ genotypes.
In the present study, bias caused by investigators' knowledge of disease or risk-factor status seems unlikely, because we selected from a general population and genotyped our sample without knowledge of disease status or lung function test results. Selection bias is possible if severe lung disease in some individuals with 5T genotypes prevented them from participating in our study; however, expected and observed numbers of these genotypes according to the Hardy-Weinberg equilibrium were similar. The 2.7% frequency of F508del heterozygosity found in this study is in accordance with the 2.9% frequency of F508del heterozygosity observed in another previous study of the Danish population [51]. Annual decline in FEV1 was reduced in 7T/9T individuals and incidence of COPD hospitalization was reduced in 5T/7T individuals. If correction for multiple comparisons was performed, these significant findings become nonsignificant. Therefore, and because reduced COPD risk in 5T/7T individuals is less biologically plausible, the findings are likely due to chance alone rather than representing real phenomena. Misclassification of genotypes is unlikely, because diagnoses were confirmed by sequencing a subsample of different poly-T variants.
Conclusion
Polythymidine 5T heterozygosity was not associated with increased annual decline in FEV1 or risk of asthma or COPD in adults in this population-based study; however, both 5T homozygotes showed evidence of asthma. Furthermore, our results also support that F508del heterozygosity may be associated with increased asthma risk independently of the 5T allele.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Morten Dahl, Anne Tybjærg-Hansen, and Børge G. Nordestgaard carried out the genotyping and statistical analysis. Peter Lange helped collect the data and was involved in the statistical analysis. All investigators participated in designing the study and in writing the paper, and all authors read and approved the final version of the manuscript.
Acknowledgements
We thank Birgit Hertz, Hanne Damm and Nina D. Kjersgaard for expert technical assistance. The Danish Heart Foundation and the Danish Lung Association supported this study.
==== Refs
Dahl M Tybjærg-Hansen A Lange P Nordestgaard BG ΔF508 heterozygosity in cystic fibrosis and susceptibility to asthma Lancet 1998 351 1911 1913 9654257 10.1016/S0140-6736(97)11419-2
Tzetis M Efthymiadou A Strofalis S Psychou P Dimakou A Pouliou E CFTR gene mutations – including three novel nucleotide substitutions – and haplotype background in patients with asthma, disseminated brochiectasis and chronic obstructive pulmonary disease Hum Genet 2001 108 216 221 11354633 10.1007/s004390100467
Hoffjan S Nicolae D Ober C Association studies for asthma and atopic diseases: a comprehensive review of the literature Resp Res 2003 4 14 10.1186/1465-9921-4-14
Boucher RC New concepts of the pathogenesis of cystic fibrosis lung disease Eur Respir J 2004 23 146 158 14738247 10.1183/09031936.03.00057003
The Cystic Fibrosis Genetic Analysis Consortium Worldwide survey of the ΔF508 mutation – report from the cystic fibrosis genetic analysis consortium Am J Hum Genet 1990 47 354 359 2378364
Dahl M Nordestgaard BG Lange P Tybjærg-Hansen A Fifteen-year follow-up of pulmonary function in individuals heterozygous for the cystic fibrosis phenylalanine-508 deletion J Allergy Clin Immunol 2001 107 818 823 11344348 10.1067/mai.2001.114117
Griesenbach U Geddes DM Alton EWFW The pathogenic consequences of a single mutated CFTR gene Thorax 1999 54 suppl 2 S19 S23 10451687
Chillón M Casals T Mercier B Bassas L Lissens W Silber S Mutations in the cystic fibrosis gene in patients with congenital absence of the vas deferens N Engl J Med 1995 332 1475 1480 7739684 10.1056/NEJM199506013322204
Pignatti PF Bombieri C Benetazzo M Casartelli A Trabetti E Gilè LS CFTR gene variant IVS8-5T in disseminated bronchiectasis Am J Hum Genet 1996 58 889 892 8644755
Kerem E Rave-Harel N Augarten A Madgar I Nissim-Rafinia M Yahav Y A cystic fibrosis transmembrane conductance regulator splice variant with partial penetrance associated with variable cystic fibrosis presentations Am J Respir Crit Care Med 1997 155 1914 1920 9196095
Bombieri C Benetazzo M Saccomani A Belpinati F Gilè LS Luisetti M Complete mutational screening of the CFTR gene in 120 patients with pulmonary disease Hum Genet 1998 103 718 722 9921909 10.1007/s004390050897
Castellani C Bonizzato A Pradal U Filicori M Foresta C La Sala GB Evidence of mild respiratory disease in men with congenital absence of the vas deferens Respir Med 1999 93 869 875 10653048 10.1016/S0954-6111(99)90052-5
Noone PG Pue CA Zhou Z Friedman KJ Wakeling EL Ganeshananthan M Lung disease associated with the IVS8 5T allele of the CFTR gene Am J Respir Crit Care Med 2000 162 1919 1924 11069835
Noone PG Knowles MR 'CFTR-opathies': disease phenotypes associated with cystic fibrosis transmembrane regulator gene mutations Respir Res 2001 2 328 332 11737931 10.1186/rr82
Andrieux J Audrézet MP Frachon I Leroyer C Roge C Scotet V Quantification of CFTR splice variants in adults with disseminated bronchiectasis, using the TaqMan flourogenic detection system Clin Genet 2002 62 60 67 12123489 10.1034/j.1399-0004.2002.620108.x
Lee JH Choi JH Namkung W Hanrahan JW Chang J Song SY A haplotype-based molecular analysis of CFTR mutations associated with respiratory and pancreatic diseases Hum Mol Genet 2003 12 2321 2332 12952861 10.1093/hmg/ddg243
Casals T De-Gracia J Gallego M Dorca J Rodríguez-Sanchón B Ramos MD Bronchiectasis in adult patients: an expression of heterozygosity for CFTR gene mutations? Clin Genet 2004 65 490 495 15151509 10.1111/j.0009-9163.2004.00265.x
King PT Freezer NJ Holmes PW Holdsworth SR Forshaw K Sart DD Role of CFTR mutations in adult bronchiectasis Thorax 2004 357 358 15047968 10.1136/thx.2003.020263
Lázaro C de Cid R Sunyer J Soriano J Giménez J Álvarez M Missense mutations in the cystic fibrosis gene in adult patients with asthma Hum Mutat 1999 14 510 519 10571949 10.1002/(SICI)1098-1004(199912)14:6<510::AID-HUMU10>3.0.CO;2-O
de Cid R Chomel JC Lazaro C Sunyer J Baudis M Casals T CFTR and asthma in the French EGEA study Eur J Hum Genet 2001 9 67 69 11175304 10.1038/sj.ejhg.5200578
Castellani C Quinzii C Altieri S Mastella G Assael BM A pilot survey of cystic fibrosis clinical manifestations in CFTR mutation heterozygotes Genet Test 2001 5 249 254 11788092 10.1089/10906570152742317
Marchand E Verellen-Dumoulin C Mairesse M Delaunois L Brancaleone P Rahier JF Frequency of cystic fibrosis transmembrane conductance regulator gene mutations and 5T allele in patients with allergic bronchopulmonary aspergillosis Chest 2001 119 762 767 11243954 10.1378/chest.119.3.762
Eaton TE Miller PW Garrett JE Cutting GR Cystic fibrosis transmembrane conductance regulator gene mutations: do they play a role in the aetiology of allergic bronchopulmonary aspergillosis? Clin Exp Allergy 2002 32 756 761 11994102 10.1046/j.1365-2222.2002.01361.x
Friedman KJ Heim RA Knowles MR Silverman LM Rapid characterization of the variable length polythymidine tract in the cystic fibrosis (CFTR) gene: association of the 5T allele with selected CFTR mutations and its incidence in atypical sinopulmonary disease Hum Mutat 1997 10 108 115 9259194 10.1002/(SICI)1098-1004(1997)10:2<108::AID-HUMU3>3.0.CO;2-G
Kiesewetter S Macek M Davis C Curristin SM Chu CS Graham C A mutation in CFTR produces different phenotypes depending on chromosomal background Nat Genet 1993 5 274 278 7506096 10.1038/ng1193-274
Cuppens H Teng H Raeymaekers P De Boeck C Cassiman JJ CFTR haplotype backgrounds on normal and mutant CFTR genes Hum Mol Genet 1994 3 607 614 7520797
Chu CS Trapnell BC Curristin S Cutting GR Crystal RG Genetic basis of variable exon 9 skipping in cystic fibrosis transmembrane conductance regulator mRNA Nat Genet 1993 3 151 156 7684646 10.1038/ng0293-151
Teng H Jorissen M Van Poppel H Legius E Cassiman JJ Cuppens H Increased proportion of exon 9 alternatively spliced CFTR transcripts in vas deferens compared with nasal epithelial cells Hum Mol Genet 1997 6 85 90 9002674 10.1093/hmg/6.1.85
Larriba S Bassas L Giménez J Ramos MD Segura A Nunes V Testicular CFTR splice variants in patients with congenital absence of the vas deferens Hum Mol Genet 1998 7 1739 1744 9736775 10.1093/hmg/7.11.1739
Mak V Jarvi KA Zielenski J Durie P Tsui LC Higher proportion of intact exon 9 CFTR mRNA in nasal epithelium compared with vas deferens Hum Mol Genet 1997 6 2099 2107 9328474 10.1093/hmg/6.12.2099
Rave-Havel N Kerem E Nissim-Rafinia M Madjar I Goshen R Augarten A The molecular basis of partial penetrance of splicing mutations in cystic fibrosis Am J Hum Genet 1997 60 87 94 8981951
Manson A Huxley C Skipping of exon 9 of human CFTR in YAC-transgenic mice Genomics 2001 77 127 134 11597137 10.1006/geno.2001.6630
Delaney SJ Rich DP Thomson SA Hargrave MR Lovelock PK Welsh MJ Cystic fibrosis transmembrane conductance regulator splice variants are not conserved and fail to produce chloride channels Nat Genet 1993 4 426 431 7691356 10.1038/ng0893-426
Strong TV Wilkinson DJ Mansoura MK Devor DC Henze K Yang Y Expression of an abundant alternatively spliced form of the cystic fibrosis transmembrane conductance regulator (CFTR) gene is not associated with a cAMP-activated chloride conductance Hum Mol Genet 1993 2 225 230 7684641
Garred P Pressler T Madsen HO Frederiksen B Svejgaard A Høiby N Association of mannose-binding lectin gene heterogeneity with severity of lung disease and survival in cystic fibrosis J Clin Invest 1999 104 431 437 10449435
Mahadeva R Westerbeek RC Perry DJ Lovegrove JU Whitehouse DB Carroll NR Alpha1-antitrypsin deficiency alleles and the TaqI GA allele in cystic fibrosis lung disease Eur Respir J 1998 11 873 879 9623690 10.1183/09031936.98.11040873
Frangolias DD Ruan J Wilcox PJ Davidson GF Wong LTK Berthiaume Y Alpha1-antitrypsin deficiency alleles in cystic fibrosis lung disease Am J Respir Cell Mol Biol 2003 29 390 396 12689922 10.1165/rcmb.2002-0271OC
Schnohr P Jensen G Lange P Scharling H Appleyard M The Copenhagen City Heart Study – Østerbroundersøgelsen. Tables with data from the third examination 1991–1994 Eur Heart J Suppl 2001 3 H H1 H83
Jensen G Epidemiology of chest pain and angina pectoris, with special reference to treatment needs Acta Med Scand Suppl 1984 682 1 120 6587759
Lange P Nyboe J Jensen G Schnohr P Appleyard M Ventilatory function impairment and risk of cardiovascular death and of fatal or non-fatal myocardial infarction Eur Respir J 1991 4 1080 1087 1756841
British Thoracic Society guidelines for the management of chronic obstructive pulmonary disease Thorax 1997 52 suppl 5 S1 S28 9474238
Dahl M Tybjærg-Hansen A Lange P Vestbo J Nordestgaard BG Change in lung function and morbidity from chronic obstructive pulmonary disease in α1-antitrypsin MZ heterozygotes: a longitudinal study of the general population Ann Intern Med 2002 136 270 279 11848724
Dahl M Tybjærg-Hansen A Schnohr P Nordestgaard BG A population-based study of morbidity and mortality in mannose-binding lectin deficiency J Exp Med 2004 199 1391 1399 15148337 10.1084/jem.20040111
Bojesen SE Tybjærg-Hansen A Nordestgaard BG Integrin β3 Leu33Pro homozygosity and risk of cancer J Natl Cancer Inst 2003 95 1150 1157 12902444
Viel M Leroy C Georges MD Claustres M Bienvenu T Novel length variant of the polypyrimidine tract within the splice acceptor site in intron 8 of the CFTR gene: consequences for genetic testing using standard assays Eur J Hum Genet 2004 15340362
Cuppens H Lin W Jaspers M Costes B Teng H Vankeerberghen A Polyvariant mutant cystic fibrosis transmembrane conductance regulator genes. The polymorphic (Tg)m locus explains the partial penetrance of the T5 polymorphism as a disease mutation J Clin Invest 1998 101 487 496 9435322
Groman JD Hefferon TW Casals T Bassas L Estivill X Georges MD Variation in a repeat sequence determines whether a common variant of the cystic fibrosis transmembrane conductance regulator gene is pathogenic or benign Am J Hum Genet 2004 74 176 179 14685937 10.1086/381001
Aznarez I Zielenski J Siminovitch K Tsui LC Increased frequency of CFTR mutations and variants among asthma patients Pediatr Pulmonol 1999 suppl 19 208 10213261
Mennie M Gilfillan A Brock DJH Liston WA Heterozygotes for the delta F508 cystic fibrosis allelele are not protected against bronchial asthma Nat Med 1995 1 978 979 7489375 10.1038/nm1095-978b
Schroeder SA Gaughan DM Swift M Protection against brochial asthma by CFTR delta F508 mutation: a heterozygote advantage in cystic fibrosis Nat Med 1995 1 703 705 7585155 10.1038/nm0795-703
Schwartz M Brandt NJ Koch C Lanng S Schiøtz PO Genetic analysis of cystic fibrosis in Denmark. Implications for genetic counseling, carrier diagnosis and prenatal diagnosis Acta Paediatr 1992 81 522 526 1392366
|
16212675
|
PMC1277850
|
CC BY
|
2021-01-04 16:36:26
|
no
|
Respir Res. 2005 Oct 9; 6(1):113
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-113
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-1151621613010.1186/1465-9921-6-115ResearchA confocal microscopic study of solitary pulmonary neuroendocrine cells in human airway epithelium Weichselbaum Markus [email protected] Malcolm P [email protected] Elisha J [email protected] Philip J [email protected] Darryl A [email protected] Asthma and Allergy Research Institute, Sir Charles Gairdner Hospital, Nedlands, 6009, Western Australia2 Centre for Asthma, Allergy and Respiratory Research, University of Western Australia, 60093 Department of Physiology, University of Western Australia, Nedlands, 6009, Western Australia4 Heart Research Institute, Royal North Shore Hospital, The University of Sydney NSW 2006 Australia5 James Hogg iCAPTURE center for Cardiovascular and Respiratory Research, St. Pauls Hospital, University of British Columbia, Vancouver, BC V6Z 1Y6, Canada2005 10 10 2005 6 1 115 115 9 5 2005 10 10 2005 Copyright © 2005 Weichselbaum et al; licensee BioMed Central Ltd.2005Weichselbaum et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Pulmonary neuroendocrine cells (PNEC) are specialized epithelial cells that are thought to play important roles in lung development and airway function. PNEC occur either singly or in clusters called neuroepithelial bodies. Our aim was to characterize the three dimensional morphology of PNEC, their distribution, and their relationship to the epithelial nerves in whole mounts of adult human bronchi using confocal microscopy.
Methods
Bronchi were resected from non-diseased portions of a lobe of human lung obtained from 8 thoracotomy patients (Table 1) undergoing surgery for the removal of lung tumors. Whole mounts were stained with antibodies to reveal all nerves (PGP 9.5), sensory nerves (calcitonin gene related peptide, CGRP), and PNEC (PGP 9.5, CGRP and gastrin releasing peptide, GRP). The analysis and rendition of the resulting three-dimensional data sets, including side-projections, was performed using NIH-Image software. Images were colorized and super-imposed using Adobe Photoshop.
Results
PNEC were abundant but not homogenously distributed within the epithelium, with densities ranging from 65/mm2 to denser patches of 250/mm2, depending on the individual wholemount. Rotation of 3-D images revealed a complex morphology; flask-like with the cell body near the basement membrane and a thick stem extending to the lumen. Long processes issued laterally from its base, some lumenal and others with feet-like processes. Calcitonin gene-related peptide (CGRP) was present in about 20% of PNEC, mainly in the processes. CGRP-positive nerves were sparse, with some associated with the apical part of the PNEC.
Conclusion
Our 3D-data demonstrates that PNEC are numerous and exhibit a heterogeneous peptide content suggesting an active and diverse PNEC population.
==== Body
Background
Pulmonary neuroendocrine cells (PNEC) are specialized airway epithelial cells that occur as solitary cells or as clusters called neuroepithelial bodies (NEB) [1]. They are located in the nasal respiratory epithelium, laryngeal mucosa [2] and throughout the entire respiratory tract from the trachea to the terminal airways [3]. In the fetal lung they are frequently located at the branching points of airway tubules, and in humans are present by 10 weeks gestation [4]. Neuroendocrine cells are bottle- or flask-like in shape, and reach from the basement membrane to the lumen. They can be distinguished by their profile of bioactive amines and peptides namely serotonin, calcitonin, calcitonin gene-related peptide (CGRP), chromogranin A, gastrin-releasing peptide (GRP) and cholecystokinin [4,5]. NEB may play a role as hypoxic-sensitive airway chemoreceptors [6], and an oxygen-sensitive potassium channel coupled to an oxygen sensory protein has been demonstrated in their lumenal membrane in the rabbit [7]. They are also considered to be involved in regulating localized epithelial cell growth and regeneration through a paracrine mechanism whereby their bioactive peptides are released into the environment [8]. Peptides and amines released by PNEC are involved in normal fetal lung development including branching morphogenesis [9]. The best-characterized peptides are GRP, the mammalian form of bombesin, and CGRP, which exert direct mitogenic effects on epithelial cells and exhibit many growth factor-like properties [10].
The majority of data available on the morphology, distribution, peptide expression and function of PNEC and NEB have been obtained from animal studies [11,12]. In human airways, the morphology of NEB have been studied ultrastructurally during the fetal and perinatal stage of lung development, and their peptides identified using immunogold-labeled antibodies where they are colocalized in the dense core vesicles in the cytoplasm [4,13-15]. However, there is little data describing the three dimensional morphology and peptide distribution in adult human airways where both PNEC and NEB are reported to be sparse [16,17]. It has been suggested that PNEC may play a role in mediating airway remodelling in normal lungs and in naturally occurring pulmonary disease where they increase in number [8,18].
The innervation of fetal and postnatal NEB has been also studied ultrastructurally in humans where both adrenergic and cholinergic nerve endings have been observed [4], in rabbits [19] and rats and dogs [20,21]. In rats, vagal nodose afferents traced using the carbocyanine dye DiI, terminate within NEB, but they are not positive for the sensory nerve marker CGRP [22] whereas the epithelium is richly innervated with CGRP- and Substance P (SP)- containing nerve terminals in guinea pigs [23], rats [22,24] and pigs [25]. In guinea pigs most of these afferents arise in the jugular ganglia [26,27]. However, little is known about the relationship between nerves and PNEC.
The aims of this study were to characterize the three dimensional morphology of PNEC, their distribution, and their relationship to the epithelial nerves in whole mounts of adult human bronchi using confocal microscopy. The peptides CGRP and GRP were examined for their consistency as markers of PNEC. Protein gene product 9.5 (PGP 9.5) was used as a marker of PNEC and for epithelial nerves, and CGRP for sensory nerves. The investigation was restricted to the solitary PNEC because NEB appear to be extremely rare in adult human lung [17].
Methods
Human airway tissue
A small section of bronchus was resected from the non-diseased portion of a lobe from a human lung obtained at thoracotomy from 8 patients undergoing surgery for the removal of lung tumors (Table 1). Three subjects were life-long non-smokers. The sample was removed from the freshly excised lobe on ice and fixed in Streck Tissue Fixative (Streck Laboratories, US). The airway segment(s) ranged from 3 to 6 mm inner diameter and were up to 1 cm in length. They were cut open lengthwise and the airway wall carefully dissected to create thin sheets that comprised epithelium and mucosa while discarding smooth muscle and cartilage. As a standard antigen retrieval method, tissues were microwaved for 20 min in citrate buffer (pH 6.0) and afterwards blocked for one hour in PBS pH 7.4 containing 1% bovine serum albumin. The tissue was cut into pieces of approximately 5 mm2 area -usually about 10 pieces – which are referred to as wholemounts, and all further treatment was carried out in 96-well culture plates equipped with anti-evaporation lids.
Table 1 Patient Demographic Data
Patient Gender Age Smoking status Disease
1 Male 65 Smoker SSC
2 Male 67 Smoker LSC
3 Female 40 Non-smoker Adeno
4 Female 37 Non-smoker LSC
5 Female 61 Ex-Smoker Adeno
6 Female 74 Smoker SSC
7 Male 70 Smoker SSC
8 Female 72 Smoker SSC
Abbreviations: SSC, small cell carcinoma; LSC, Large cell carcinoma; Adeno, Adenocarcinoma.
Preparation and staining of whole mounts
Whole mounts were stained with antibodies to reveal all nerves (PGP 9.5), sensory nerves (CGRP), and PNEC (PGP 9.5, CGRP and GRP). The PGP 9.5 antibodies (monoclonal and polyclonal) were obtained from UltraClone, UK and used at a dilution of 1/100 and 1/500, respectively. Antibodies to GRP (polyclonal) and CGRP (monoclonal and polyclonal) were purchased from Dako, NSW, Australia. The dilutions were as follows: GRP, 1/200; polyclonal CGRP, 1/400, monoclonal CGRP, 1/100. The secondary antibodies (anti-mouse and anti-rabbit) conjugated to Alexa-488 and Alexa-543, respectively were obtained from Molecular Probes, MA and used in a dilution of 1/200. Typically, 10 μl of antibody solution was used for each well. Control experiments to test for auto-fluorescence and non-specific staining were carried out using non-immune rabbit and mouse sera as described previously [28]. The tissues were incubated with primary antibodies overnight (4°C) in the presence of 0.3% Triton X-100 to enhance permeabilization. After washing 3 × 20 min in PBS, fluorophor-conjugated secondary antibodies were applied overnight (4°C). After further washing with PBS, the preparations were mounted in 90% glycerol containing p-phenylethylenediamine (1 mg/ml) to reduce bleaching of the fluorochromes. Custom-made slides were used that enabled imaging the specimen from both sides. The coverslips were raised with spacers (Imaging spacers, Sigma) in order to minimize compression of the specimens. The edges of the coverslips were sealed with nail polish to prevent evaporation of the mounting medium.
Confocal microscopy
Wholemount pieces were double-stained in combinations of poly- and monoclonal antibodies and imaged using confocal microscopy (Biorad MRC-1000, Comos Software 7.0) as previously described [29]. For highest magnification, the focus depth was increased at 1 micron steps during scanning. The analysis and rendition of the resulting three-dimensional data sets, including side-projections, was performed using NIH-Image software (Version 1.61b12). Images were colorized and super-imposed with Adobe Photoshop 5 to reveal the complex structure of these cells. The Image Processing Tool Kit plug-in (Raindeer Software) was used to measure PNEC density. From each patient, a minimum of four fields of adequate staining quality and optimal signal to noise ratio, were imaged at a magnification of × 10. The PNEC were manually marked out and the software automatically calculated the PNEC density per area. Because the PNEC numbers were not homogeneously distributed in most of the bronchi sampled, no attempt was made to calculate the mean density of these cells for each bronchus.
Results
PNEC staining with PGP9.5
Solitary PNEC were abundant within the epithelium of the bronchi of all eight human lungs examined when stained with PGP 9.5 or GRP. The numbers were not evenly distributed, ranging from 65 to 100/mm2 over the area of any one wholemount, but with denser patches comprising 150 to 260/mm2 observed less frequently in several of the lungs (Figure 1). Staining with the neural marker PGP 9.5 typically revealed PNEC with a flask-like shape when viewed from the side (Figure 2a, 90° rotation). The cell body was located near the basement membrane with the apical part of the cell comprising a characteristic thick stem that extended to the lumen surface (Figure 2b and 2c). The overall height of the cells averaged 50.1 ± 6.7 μm (SD, n = 21) in four lungs. Processes issued from the cell body along the basal region of the epithelium and also toward the lumen (Figure 2a, 90° rotation, Fig 2b and 2c). These processes have a dendritic-like appearance when viewed as a projection from the lumen because their three-dimensional morphology cannot be readily appreciated from this aspect (Figure 2a, 0° rotation; Figure 2d). Varicose nerves ascended in close association with the PNEC stem to reach an apical nerve plexus (Figure 2 and 3a, 3b) that lies just below the luminal surface of the epithelium. Figure 3a shows an example of an isolated patch of varicose epithelial nerves taken at low power (lumen view, 0 degrees). The apical disposition of these nerves is seen in the 90 degree rotation. No correlation was found between the distribution of PNEC and nerves in the epithelium.
Figure 1 Whole mount of mucosa from a human bronchus stained with gastrin releasing peptide (GRP) and imaged from the luminal surface with a confocal microscope. The low power projection reveals an abundance of pulmonary neuroendocrine cells (PNEC) in the epithelium. Bar = 500 μm. Inset: higher power view revealing the morphology of PNEC. Where the epithelial surface is flat (ie. parallel with the cover slip), the view is from the top looking down on the cell body and processes but where they lie on the edge of a mucosal fold their flask-like shape is revealed. Bar = 50 μm.
Figure 2 (a). High power projection of two PNEC and associated nerves imaged from the lumen and stained with the neural marker, protein gene product 9.5 (PGP 9.5). To reveal the shape of the cell body and the diverse structure of its processes the data set has been rotated to enable viewing the PNEC from the side. The upper panel is the conventional view from the lumen surface. The middle panel is rotated at 45 deg, and the lower panel shows the side view at 90 degrees. Nerves are present in close apposition to the PNEC. Bar = 20 μm. (b & c) Projections of typical flask-like PNEC stained with PGP 9.5 and imaged from the lumen over the edge of a mucosal fold. The cell bodies are seen from their sides (thus a cut off line for the surface of the epithelium with the lumen is not clearly seen). Processes of varying morphology arise from the cell bodies and varicose nerves are present near the base and apex of the PNEC with individual nerve fibers rising through the epithelium. The apex of the cell is brightly stained in (b). Bars = 10 μm. (d) Four PNEC viewed from a flat area of the airway lumen showing dendritic-like processes. This low power projection extends through a depth of 50 μm and includes nerves that lie below the basement membrane.
Figure 3 (a) Representative views of airway mucosa from four lungs double-stained for protein gene product 9.5 (PGP9.5, green) and gastrin releasing peptide (GRP, red). The upper panel is the lumen view, the middle one is rotated through 30 deg and the lower one through 90 deg, ie view from the side. Strings of varicose nerves are present in the epithelium. The dark holes indicate the location of goblet cells. PNEC are inconspicuous in the upper view but become more apparent when the field is projected at an angle. The lower panels demonstrate that GRP is present predominantly in the PNEC processes, whereas PGP 9.5 is restricted to the cell body with its apical stem. Nerves feature strongly in the apical epithelium. Bar = 50μm. Boxed area: This PNEC has been turned through 90 degrees so that the prominent processes now point upwards, and enlarged (right, upper panel). A 90 deg rotation of the projection reveals that the PNEC processes stain strongly for GRP whereas the cell body and apical stem stain mainly for PGP 9.5. Some of the processes are lumen-directed, other processes with feet-like appearance are directed toward the lamina propria. Bar = 10μm. (b) Three PNEC in a field from another lung shown from the lumen (upper) and rotated through 90 deg (lower). Two of the PNEC are predominantly GRP positive whereas the third PNEC stains strongly for PGP 9.5. The upper stem of the middle cell body stains yellow indicating that the PGP 9.5 and the GRP staining are about equal. Bar = 10 μm. (c) A single PNEC shown as individual fields: PGP9.5 only (left), GRP only (middle), composite PGP9.5 + GRP (right). GRP reveals fine processes that issue from the cell body. In contrast to PGP 9.5, GRP does not stain the cell nucleus. Bar = 10 μm. (d) A single PNEC in close association with a nerve terminal. A nerve rises from the base of the PNEC, climbs through the epithelium along the PNEC stem and spreads laterally in the apical epithelium where it exhibits enlarged terminal varicosities. Upper panel: lumen view, middle panel: 45 deg rotation, lower panel, 90 degree rotation. Bar 10 μm. (e) Lumen view of mucosa where the epithelium is tilted showing three PNEC from an angle. Patches of fine varicose nerves are present. Some of the nerves lie close to the stems of two of the PNEC (arrow heads). Bar = 25 μm. Boxed area(right): High power view after rotating shows two PNEC within a patch of nerves. The left PNEC is strongly PGP9.5 positive. The right PNEC has several processes in close apposition to nerves that rise through the epithelium to form an apical nerve plexus. Nerves in the apical epithelium lie close to the central stems of both PNEC. Bar = 10 μm.
Co-localization of GRP and PGP9.5 in PNEC
All PGP 9.5-positive PNEC were also positive for GRP, however, the PGP 9.5 staining intensity of individual PNEC varied considerably. GRP exhibited significantly higher detail of the processes whereas PGP 9.5 stained predominantly the cell body with its prominent stem (Figure 3a, 30° and 90° rotations). This is strikingly shown in the PNEC enclosed in the boxed area of Figure 3a which has been rotated and enlarged. Most PNEC exhibited a dominance of one marker over the other (Figure 3a, 90 degree rotation, and Figure 3b). Individual staining of a single PNEC for both markers is shown in Figure 3c where detail of the thick and fine processes are revealed with GRP.
Double staining for GRP and PGP9.5 in PNEC associated with nerves
The association of ascending and apical nerves with PNEC in the epithelium was more readily appreciated with double staining. Figure 3d shows a nerve rising from the base of a PNEC that extends upwards along the stem and terminates near the luminal surface. Figure 3e shows three PNEC, two of which are in a patch of nerves (arrows, boxed region). The right PNEC and its processes are in close proximity to a network of varicose nerves. When a rotation was performed on the field (boxed area, right) nerves arising from the base of this PNEC ascended along its stem to join the apical nerve plexus. Other nerves travelling from the lamina propria into the epithelium accompanied the GRP-stained processes towards the lumen.
Double staining for CGRP and PGP 9.5
When PNEC were double stained for CGRP and PGP 9.5, CGRP was present in 22 ± 9% (SD, 10 fields, 4 lungs) of all PNEC stained (Figure 4a). CGRP typically stained the processes but was faint or absent in the cell body. Figure 4a (boxed area, right) shows a PNEC side-view with lumen-directed processes where one is particularly strongly stained for CGRP. It was also present in the shorter processes directed towards the lamina propria (Figure 4b, left) where three pairs of PNEC display quite diverse morphology. Unlike the staining pattern observed with GRP, CGRP is chiefly present in thicker, more proximal part of the processes.
Figure 4 (a) Lumen view of a small area of epithelium representative of a whole mount of airway mucosa double-stained with calcitonin gene-related peptide (CGRP, red) and PGP9.5 (green). CGRP is present in terminal processes of some PNEC. Bar = 100 μm. Boxed area (right): Side projection of the PNEC where two lumen-directed processes and the long stem of the cell body (green) ascend to the apical epithelium. From this view the right process is strongly CGRP positive. Bar = 20 μm. (b) Wholemount of airway mucosa double-stained for CGRP (red) and PGP 9.5 (green) showing diverse morphology of PNEC. Upper panel is the lumen view, middle is a rotated through 45 degrees and the lower through 90 degrees. From left to right: A pair of PNEC with CGRP in the processes. The next two have fine processes that arise from the cell body and the apex of the stem. The right hand pair of PNEC exhibits branching of the main stem close to the lumen. Bar = 50 μm.
Staining for CGRP in nerves and PNEC
Faint staining of CGRP could be detected in PNEC cell bodies but it was much less than that in the processes. Figure 5a demonstrates that one of the lumen-directed processes stains stronger for CGRP than the other processes and the cell body. When PNEC were present in a field containing CGRP-positive nerves, they appeared to make contact with the PNEC. These contacts were characterized by brightly stained, enlarged terminal varicosities indicative of nerve endings, suggestive of innervation of PNEC by CGRP-positive nerves (Figure 5b).
Figure 5 (a) Two PNEC stained with CGRP, shown as lumen view (upper panels) and rotated through 90 deg (lower panels). Each of the latter reveal a brightly staining process that issues from their cell body toward lumen whereas the cell bodies (left side of each panel) and other processes show weaker staining. Bar = 20 μm. (b) Wholemount of airway mucosa stained for CGRP. A network of weakly staining CGRP varicose nerves and a single PNEC is present. The top panel is the lumen view (0°), followed by rotations of 30°, 60° and 90°. One fiber exhibits brightly stained varicosities at the apparent point of contact with the PNEC. Rotating the field demonstrates that nerves run from below the base of the PNEC prior to ascending to it. Other ascending fibers are present in the left side of the field. The horizontal lines are a consequence of the limited number of Z steps in the confocal data set. Bar = 20 μm.
Discussion
This is the first report characterizing PNEC morphology in three dimensions in human airways. The use of whole mounts of mucosal tissue enabled the direct demonstration of PNEC abundance over large areas of airway epithelium. Rotation of 3-D images revealed the complexity of the PNEC body and its processes that issue laterally from its base, some lumen directed, others feet-like that were directed toward the lamina propria. PGP 9.5 and GRP were reliable markers, both staining all PNEC, whereas CGRP was present in about 20% of the PNEC population. However, the distribution of staining varied widely among cells, with GRP and CGRP mainly present in the processes, but with GRP occurring also in the cell body to varying degrees. The variation exhibited in the morphology of the PNEC and its differing peptide profiles suggests that these cells may be in a dynamic state in the epithelium.
PNEC were abundant when stained with either PGP 9.5 or GRP. Previous studies in human airways that investigated PNEC density used conventional cross sections and revealed numbers ranging from 1.05 PNEC/cm basement membrane (corresponding to 4 PNEC per 10,000 epithelial cells) using neuron specific enolase [17] to 12.5 PNEC/cm, or 41 PNEC per 10,000 epithelial cells, using chromogranin A as a marker [30]. In the current study, our measurements reveal that the density of PNECs ranges from 65/mm2 to 260/mm2 within an individual wholemount. There did not appear to be a homogenous distribution of PNEC either within or between wholemounts. Areas of high PNEC density appeared to be juxtaposed to areas of sparse cell numbers. Direct comparisons between the numbers provided in our study and those reported previously are confounded by the considerable discrepancy between the proportions of PNEC revealed by markers used to identify PNEC [17,30]. We used PGP 9.5 to label the whole PNEC population, whereas Boers reported that only 14% of all PNEC showed PGP 9.5 immunoreactivity [30]. Furthermore, our study indicates that all PNEC contain both PGP 9.5 and GRP, whereas Gosney et al., [17] found GRP present in 65% of all PNEC stained with neuron specific enolase and Boers demonstrated GRP in 59% of all PNEC stained with chromogranin A [30]. The reasons for these discrepancies are unknown, although the greater sensitivity and ability of confocal microscopy to resolve cell types and contents may account for the observed differences. Similarly, the size of the airway may also influence the PNEC distribution. In rat lungs at least, the density of NEB/PNEC appears to be dependent on airway size, with a greater density of cells observed in proximal airways compared to more distal generations [31]. Only large cartilaginous airways (3–6 mm ID) were available for this study. Although multiple wholemount pieces were analysed from each airway, our study was limited to 1 piece of bronchial tissue per subject and as such we are unable to comment on the overall distribution of PNEC throughout the lungs.
The majority of PNEC were not in close association with epithelial nerves, partly because nerves were generally sparse and tended to occur in patches as we reported recently [25]. When present, PGP 9.5 positive nerve fibres were generally observed to be in close apposition with the PNEC cell body, its stem, apex, and processes. Unfortunately, the immunofluorescent staining used in this study does not permit the distance between the nerve fibres and the PNEC to be measured accurately, and side projections computed from data obtained by scanning from the luminal surface are subject to loss of resolution. Nonetheless, when images were viewed from the lumen through a series of rotations over 360 degrees, it was clear that the nerves lie very close to these structures, within one micrometre. These are likely to be sensory nerves because they are varicose, with enlarged varicosities in the terminal region [25], and in most lungs stained positively though weakly for CGRP. CRRP-positive nerves endings appeared to terminate near or on the processes issuing from the base of the PNEC where bright terminal enlarged varicosities were seen. Afferent and efferent nerves have been characterized ultrastructurally on cells within NEB in fetal and neonate humans [4] and adult animals [21,22] but as far as we are aware, not on single PNEC. This approach is hampered because of the apparent paucity of epithelial sensory nerves in humans. The infrequent patches of PGP 9.5-positive epithelial nerves stain very weakly in humans [25,32], or not at all [33,34] for SP or CGRP [35] in contrast to rats [24] and pigs [25] where they are abundant.
In humans, NEB decrease in frequency with age and are rare in adult lung. Gosney et al.,[16] observed only three NEB after searching through preparations from 5 post-mortem specimens. Our data support this finding, as only two NEB were observed and were confined to a single lung after scanning several hundred whole mount PIECES from eight adult lungs. Brouns and colleagues recently demonstrated a complex innervation pattern of pulmonary NEBs in rat airways, comprising both sensory vagal nerves as well as non-vagal CGRP/SP positive nerves [36]. The physiological role of the innervation of NEB is not well understood. It has been proposed that the nerve endings at the base of the NEB subserve an axon reflex, presumably arising in the NEB itself and possibly penetrating to deeper tissues such as the airway smooth muscle [4]. There may also be local reflex connections through peripheral ganglia. Hypoxia detected by the O2 sensor in the NEB is presumed to release mediators that stimulate vagal afferents, but no central nervous reflexes have been identified [37]. Recent advances in microscopic techniques with increased sensitivity may shed more light on the morphological basis for many of the suggested functions of NEB innervation.
In our study, we used immunofluorescently-labeled antibodies to PGP 9.5, GRP and CGRP to detect PNEC in adult airway epithelium. PGP 9.5 stained all the cell bodies fairly uniformly but was weaker in the processes so that their fine ends frequently were not revealed. All PGP 9.5 positive cells were also positive for GRP, suggesting that it may also be a reliable marker for identifying PNEC. It was predominantly observed in the processes, but in many cells it stained the cell body and stem region either partially or completely. However, less than a quarter of PNEC exhibited CGRP immunoreactivity, with the greatest intensity displayed in the thick processes. Although CGRP is often used in animal studies as a marker for quantitative studies of NEB and PNEC [38], we have shown that CGRP is not a reliable marker of the PNEC populations in human adult epithelium, as only a subset exhibited CGRP immunoreactivity. These markers, used in conjunction with three dimensional imaging and image rotation, have revealed the overall morphology of the PNEC that hitherto has not been appreciated using conventional light and electron microscopy. PGP 9.5 revealed the variety of shapes that the cell body can attain, most often flask or bottle-shaped with the base at the basement membrane and its long stem extending to the lumen where its tip was often more brightly stained. Some PNEC exhibited branching of the main stem close to the lumen.
GRP staining revealed a striking PNEC morphology with thick processes issuing laterally from near the base of the cell body upwards toward the apical epithelium and along the basement membrane. In addition GRP also stained fine processes originating from the side of the PNEC body that were not readily detected with PGP 9.5. The thick and thin processes of the PNEC, revealed by our 3-D confocal microscopy, may be the conduits that effect delivery of the bioamines and peptides proposed to be secreted by PNEC.
Bioamines and peptides contained within the PNEC have been proposed to be secreted into the adjacent epithelium and lamina propria in response to such stimuli as hypoxia [6,11]. GRP and CGRP have been shown to have mitogenic and growth factor like influences [39] and may have a direct influence on epithelial regeneration and an indirect one via local vasodilation of the adjacent bronchial vasculature. Our confocal microscopic study demonstrates PNEC with heterogeneous peptide content, suggesting an active and diverse PNEC population is present in adult human airway epithelium.
In this study, lung tissue samples were derived from a diverse group of patients ranging from 39 to 74 years of age that were undergoing thoracotomy for removal of lung tumors. The small number of patients precluded the correlation of our results to gender or smoking history. Thus it is difficult to determine to what extent data presented in this study represent the steady-state versus disease-specific remodeling of the airway epithelium.
Conclusion
Our 3D-data demonstrates that PNEC are numerous and exhibit a heterogeneous peptide content suggesting an active and diverse PNEC population. Valuable insights into the biology of cells identified in this study may come from a better understanding of their abundance, morphology and innervation comparing normal lung tissue versus injured or diseased lungs.
Authors' contributions
MW, EJH carried out the sample preparation and confocal microscopy and reviewed the manuscript. MPS, PJT and DAK conceived of the program, participated in the design and coordination of this study, and drafted the manuscript.
Acknowledgements
The authors would like to thank the cardiothoracic surgeons and theatre staff at Sir Charles Gairdner Hospital and Pathology staff at the PathCentre for provision of lung specimens. This work was supported by the University of Western Australia Research Grants Scheme and the Sir Charles Gairdner Hospital Research Foundation.
==== Refs
Lauweryns JM Van Ranst L Protein gene product 9.5 expression in the lungs of humans and other mammals. Immunocytochemical detection in neuroepithelial bodies, neuroendocrine cells and nerves Neurosci Lett 1988 85 311 6 2966308 10.1016/0304-3940(88)90584-8
Johnson EW Eller PM Jafek PW Protein gene product 9.5-like and calbindin-like immunoreactivity in the nasal respiratory mucosa of perinatal humans Anat Rec 1997 247 38 45 8986301 10.1002/(SICI)1097-0185(199701)247:1<38::AID-AR6>3.0.CO;2-B
Adriaensen D Scheuermann DW Neuroendocrine cells and nerves of the lung Anat Rec 1993 236 70 85 7685156 10.1002/ar.1092360111
Stahlman MT Gray ME Ontogeny of neuroendocrine cells in human fetal lung. I. An electron microscopic study Lab Invest 1984 51 449 463 6482387
Polak JM Becker KL Cutz E Gail DB Goniakowska-Witalinska L Gosney JR Lauweryns JM Linnoila I McDowell EM Miller YM Lung endocrine cell markers, peptides and amines Anat Rec 1993 236 169 171 8507003 10.1002/ar.1092360120
Lauweryns JM Cokelaere M Deleersynder M Liebens M Intrapulmonary neuro-epithelial bodies in newborn rabbits. Influence of hypoxia, hyperoxia, hypercapnia, nicotine, reserpine, L-DOPA and 5-HTP Cell Tissue Res 1977 182 425 440 922815 10.1007/BF00219827
Youngson C Nurse C Yeger H Cutz E Oxygen sensing in airway chemoreceptors Nature 1993 365 153 155 8371757 10.1038/365153a0
Reynolds SD Giangreco A Power JH Stripp BH Neuroepithelial bodies of pulmonary airways serve as a reservoir of progenitor cells capable of epithelial regeneration Am J Pathol 2000 156 269 78 10623675
Sunday ME Hua J Dai HB Nusrat A Torday AJ Bombesin increases fetal lung growth and maturation in utero and in organ culture Am J Respir Cell Mol Biol 1990 3 199 205 2390263
Li K Nagalla SR Spindel ER A rhesus monkey model to characterize the role of gastrin-releasing peptide (GRP) in lung development. Evidence for stimulation of airway growth J Clin Invest 1994 94 1605 1615 7929836
Cutz E Introduction to pulmonary neuroendocrine cell system, structure-function correlations Microsc Res Tech 1997 37 1 3 9144617 10.1002/(SICI)1097-0029(19970401)37:1<1::AID-JEMT1>3.0.CO;2-U
Sorokin SP Ebina M Hoyt RF Development of PGP 9.5- and calcitonin gene-related peptide-like immunoreactivity in organ cultured fetal rat lungs Anat Rec 1993 236 213 225 8507009 10.1002/ar.1092360126
Stahlman MT Gray ME Colocalization of peptide hormones in neuroendocrine cells of human fetal and newborn lungs: an electron microscopic study Anat Rec 1993 236 206 212 8507008 10.1002/ar.1092360125
Stahlman MT Gray ME Immunogold EM localization of neurochemicals in human pulmonary neuroendocrine cells Microsc Res Tech 1997 37 77 91 9144624 10.1002/(SICI)1097-0029(19970401)37:1<77::AID-JEMT8>3.0.CO;2-T
Wang D Yeger H Cutz E Expression of gastrin-releasing peptide receptor gene in developing lung Am J Respir Cell Mol Biol 1996 4 409 416 8624245
Gosney JR Sissons MC O'Malley JA Quantitative study of endocrine cells immunoreactive for calcitonin in the normal adult human lung Thorax 1985 40 866 869 2866609
Gosney JR Sissons MC Allibone RO Neuroendocrine cell populations in normal human lungs: a quantitative study Thorax 1988 43 878 882 3065973
Gosney JR Pulmonary neuroendocrine cell system in pediatric and adult lung disease Microsc Res Tech 1997 37 107 113 9144627 10.1002/(SICI)1097-0029(19970401)37:1<107::AID-JEMT11>3.0.CO;2-V
Lauweryns JM Van Lommel A Ultrastructure of nerve endings and synaptic junctions in rabbit intrapulmonary neuroepithelial bodies: a single and serial section analysis J Anat 1987 151 65 83 3654362
Van Lommel A Lauweryns JM Neuroepithelial bodies in the Fawn Hooded rat lung: morphological and neuroanatomical evidence for a sensory innervation J Anat 1993 183 553 566 8300432
Van Lommel A Lauweryns JM De Leyn P Wouters P Schreinemakers H Lerut T Pulmonary neuroepithelial bodies in neonatal and adult dogs: histochemistry, ultrastructure, and effects of unilateral hilar lung denervation Lung 1995 173 13 23 7776703
Adriaensen D Timmermans JP Brouns I Berthoud HR Neuhuber WL Scheuermann DW Pulmonary intraepithelial vagal nodose afferent nerve terminals are confined to neuroepithelial bodies, an anterograde tracing and confocal microscopy study in adult rats Cell Tissue Res 1998 293 395 4 9716729 10.1007/s004410051131
Lundberg JM Hokfelt T Martling CR Saria A Cuello C Substance P-immunoreactive sensory nerves in the lower respiratory tract of various mammals including man Cell Tissue Res 1984 235 251 261 6200231 10.1007/BF00217848
Baluk P Nadel JA McDonald DM Substance P-immunoreactive sensory axons in the rat respiratory tract, a quantitative study of their distribution and role in neurogenic inflammation J Comp Neurol 1992 319 586 598 1377714 10.1002/cne.903190408
Lamb JP Sparrow MP Three-dimensional mapping of sensory innervation with substance p in porcine bronchial mucosa, comparison with human airways Am J Respir Crit Care Med 2002 166 1269 1281 12403698 10.1164/rccm.2112018
Hunter DD Undem BJ Identification and substance P content of vagal afferent neurons innervating the epithelium of the guinea pig trachea Am J Respir Crit Care Med 1999 159 1943 1948 10351943
Undem BJ Myers AC Barnes PJ Autonomic Ganglia Autonomic Control of the Respiratory System 1997 United Kingdom: Harwood Academic Publications 87 118
Sparrow MP Weichselbaum M McCray PB Development of the innervation and airway smooth muscle in human fetal lung Am J Respir Cell Mol Biol 1999 20 550 560 10100986
Weichselbaum M Sparrow MP A confocal microscopic study of the formation of ganglia in the airways of fetal pig lung Am J Respir Cell Mol Biol 1999 21 607 620 10536120
Boers JE den Brok JL Koudstaal J Arends JW Thunnissen TFB Number and proliferation of neuroendocrine cells in normal human airway epithelium Am J Respir Crit Care Med 1996 154 758 763 8810616
Larson SD Schelegle ES Hyde DM Plopper CG The Three-Dimensional Distribution of Nerves Along the Entire Intrapulmonary Airway Tree of the Adult Rat and the Anatomical Relationship Between Nerves and Neuroepithelial Bodies Am J Respir Cell Mol Biol 2003 28 592 599 12707015 10.1165/rcmb.4889
Hislop AA Wharton J Allen KM Polak JM Haworth SG Immunohistochemical localization of peptide-containing nerves in human airways, age-related changes Am J Respir Cell Mol Biol 1990 3 191 198 1975191
Laitinen LA Laitinen A Panula PA Partanen M Tervo K Tervo T Immunohistochemical demonstration of substance P in the lower respiratory tract of the rabbit and not of man Thorax 1983 38 531 536 6193599
Chanez P Springall D Vignola AM Moradoghi-Hattvani A Polak JM Godard PM Bousquet J Bronchial mucosal immunoreactivity of sensory neuropeptides in severe airway diseases Am J Respir Crit Care Med 1998 158 985 990 9731035
Joos GF Germonpre PR Pauwels RA Role of tachykinins in asthma Allergy 2000 55 321 337 10782516 10.1034/j.1398-9995.2000.00112.x
Brouns I Van Genechten J Hayashi H Gomi T Burnstock G Timmermans J-P Adriaensen D Dual Sensory Innervation of Pulmonary neuroepithelial Bodies Am J Respir Cell Mol Biol 2003 28 275 285 12594053 10.1165/rcmb.2002-0117OC
Adriaensen D Brouns I Van Genechten J Timmermans JP Functional morphology of pulmonary neuroepithelial bodies: Extremely complex airway receptors Anat Rec 2003 270 25 40 10.1002/ar.a.10007
Avadhanam KP Plopper CG Pinkerton KE Mapping the distribution of neuroepithelial bodies of the rat lung. A whole-mount immunohistochemical approach Am J Pathol 1997 150 851 859 9060823
Coleridge HM Coleridge JC Cherniack NS, Widdicombe JG Reflexes evoked from the tracheobronchial tree and lungs Handbook of Physiology, Section 3: The Respiratory System, Control of Breathing Part I 1986 II Washington DC: American Physiological Society 395 429
|
16216130
|
PMC1277851
|
CC BY
|
2021-01-04 16:36:25
|
no
|
Respir Res. 2005 Oct 10; 6(1):115
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-115
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-671623618010.1186/1477-7819-3-67ResearchClinico-morphological patterns of breast cancer including family history in a New Delhi hospital, India-A cross-sectional study Saxena Sunita [email protected] Bharat [email protected] Anju [email protected] Ashok [email protected] [email protected] Nandagudi S [email protected] Institute Of Pathology-ICMR, Safdarjung Hospital Campus, New Delhi – 110029. India2 Department of Surgery, Safdarjung Hospital, New Delhi – 110029. India3 Emeritus Scientist (Statistics), Indian Council of Medical Research, New Delhi – 110029. India2005 13 10 2005 3 67 67 16 8 2005 13 10 2005 Copyright © 2005 Saxena et al; licensee BioMed Central Ltd.2005Saxena 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
Breast cancer is the second most common malignancy among women, next to cervix cancer. Understanding its pathogenesis, morphological features and various risk-factors, including family history holds a great promise for the treatment, early detection and prevention of this cancer.
Patients and methods
In an attempt to evaluate the clinico-morphological patterns of breast cancer patients, including their family history of breast and/or other cancers, a detailed analysis of 569 breast cancer cases diagnosed during the years 1989–2003 was carried out. Mean and standard deviation and Odds ratios along with 95% confidence intervals were estimated. χ2/Fisher's exact test were employed to test for proportions.
Results
Mean age of the patient at presentation was 47.8 years, ranging from 13–82 years. Among the various histo-morphological types, Infiltrating duct carcinoma (IDC) was found to be commonest type i.e. in 502 cases (88.2%), followed by infiltrating lobular carcinoma (ILC) in 21 cases (3.7%) and other types forming 9(1%). Out of 369 cases where TNM staging was available, stage IIIB (35.2%) was the commonest. Lymph node positivity was observed in 296 cases (80.2%). Out of 226 cases evaluated for presence of family history, 47 cases (20.7%) revealed positive family history of cancer, among which breast or ovarian cancer were the commonest type (72.0%). Patients below 45 years of age had more frequent occurrence of family history as compared to above 45 years. Amongst familial cases, Infiltrating duct carcinoma was the commonest form accounting for 68.8% cases while ILC was found to be in a higher proportion (12.5%) as compared to non- familial cases (5.4%).
Conclusion
Among the various determining factors for development of breast cancer and for its early detection, family history of cancer forms one of the major risk factor. It is important to take an appropriate history for eliciting information pertaining to occurrence of cancers amongst the patients' relatives there by identifying the high risk group. Educating the population about the risk factors would be helpful in early detection of breast cancer.
==== Body
Background
Breast Cancer is the most frequent cancer in women worldwide with 1.05 million new cases every year and represents over 20% of all malignancies among females [1]. Over 50% of breast cancer incidence occurs in the developed world. High-risk areas include Europe and North America. The lowest rates are reported from Africa and Asia. However, it still ranks as the commonest cancer among women in these regions. Incidence of breast cancer is increasing in most of the countries, including the areas, which have had previously low rates [2-4]. It is estimated that in 2001 there were approximately 80,000 new breast cancer cases in India [5]. The population based cancer registry data from the various parts of the country, has revealed breast cancer as the commonest cancer among women in Delhi, Mumbai, Ahmedabad, Calcutta and Trivandrum. In the rest of the other Indian registries, breast cancer is listed as the second leading site among women [6,7]. The age standardized incidence rates vary between 9–32 per 100,000 women. An increasing trend in the incidence rates of the breast cancer has been reported from the various registries of National Cancer Registry Project [7]. This malignancy accounts for 19–34% of all cancer cases among women nationally. While the epidemiological studies for breast cancer carried out in India have largely focused on risk factors such as, religion, age at menarche, menopause and reproductive history, not much attention has been paid on role of family history even though genetic predisposition is responsible for 5–10% of all breast cancers [8-17]. It is well known that the hereditary factors play a far greater role in women for the development of breast cancer [18-20]. Data so far available is from the western studies. Several clinical and morphological parameters such as histological type of tumor, tumor grade, axillary lymph node involvement, bilaterality etc. have been established as the predictors of tumor behavior in breast cancer patients. These prognostic factors are indicators of the inherent aggressiveness of the tumor as well as of the extent of the disease and based on these factors, treatment decisions are being taken up by the clinicians. The association of these prognostic factors with familial cancers has not been documented.
The present study attempts to describe some of the clinico-pathological features of the breast cancer cases seen at a tertiary level hospital in the city of Delhi. The data has been analyzed by various parameters such as age, religious groups, morphological patterns of tumor, lymph node status, TNM staging and a significant parameter lately included i.e. familial history.
Patients and methods
A total of 788 specimens in the form of cytology and biopsy, which were clinically diagnosed as breast cancer cases were received at Institute of Pathology (IOP) during the period 1989–2003. The following eligibility criteria were employed for inclusion of cases into the study, (a) patients with complete clinical details, containing the personal proforma (b) patients who were residents of India and (c) histologically/cytologically confirmed cases of breast cancer.
Of the 788 cases, 569 cases (72.2%) satisfied the eligibility criteria and were included for the present analysis. The detailed breakup regarding 218 ineligible cases was as follows: 104 were benign lesions on histopathological examination and 40 were inflammatory lesions. Eight cases were of cystosarcoma phylloides. In 42 cases, the tissue submitted was inadequate and in the rest 25 cases clinical history was not available. For all those cases, which satisfied the eligibility criteria, clinical history and brief information on socio demographic parameters viz. age, religion, marital status and parity were extracted from the clinical records. Information relating to bilaterality of the cancer was recorded after the pathological diagnosis. Data on family history of cancers were recorded up to second degree after an in depth interview with the patient or with the person accompanying the patient. A thorough rapport was established before collecting the information regarding the family history. Wherever possible, patients were requested to bring the diagnostic reports relating to their family members to confirm the authenticity of family history of cancer. Collection of information relating to the family history was started only after 1998. The clinical extent of the disease at presentation was assessed according to the UICC TNM classification [21]. The religion wise distribution of the population for the city of Delhi was obtained from the reports of the latest census [22].
Clinical breast and axillary lymph node examination
In clinically selected cases, triple approach including clinical examination coupled with ultrasonographic examination (USG) and mammography was also carried out. USG was carried out to assess the extent of the axillary lymph nodes more precisely. Axillary lymph node examination was done clinically as well as by ultrasound. Most of the patients of breast cancer that we receive are locally advanced cancers and have palpable lymph nodes. As a matter of policy, axillary clearance was thus done in most patients. There was also a small percentage of patients who did not have palpable lymph nodes,- for them axillary lymph node sampling was done and if the level 1 lymph nodes were found to be positive, further management was done.
Statistical analysis
The data were tabulated and analyzed. Mean and standard deviation (SD) were computed for quantitative data. The statistical significance of associations between the various qualitative parameters was evaluated through Chi-square (χ2) test/Fisher's exact test (two tail). In view of skewness in the age data the statistical significance of means was tested through the non-parametric Mann Whitney test (in case of two groups) or Kruscal Wallis one- way analysis of variance for more than two groups. P value ≤ 0.05 was considered statistically significant (S). Odds ratio (OR) along with 95% confidence intervals (95% CI) were also estimated, wherever applicable.
Results
Age
In the present series the average age of the patient at presentation was 47.8 years with a standard deviation of 12.2 years. The youngest patient was 13 years and oldest patient was 86 years old. The commonest age group of incidence was 45–54 years (31.8%). Nearly 22% of cases were below 40 years while 16% of cases were above the age of 65 years.
Religion
Analysis of data by major religious groups indicated that 87%, 7.7% and 5.3% belonged to Hindus, Muslims and Christian communities respectively. It may be noted that while Christian population constituted only 0.88% of the total population in the city of Delhi as per the reports of latest census of 2001 the cancer cases in this community accounted for 5.3%. The mean age at presentation according to various religions was found to be 48.1 (SD = 12.28), 42.4 (SD = 9.15) and 50.8 (SD = 12.59) years for Hindu, Muslim and Christians respectively. Christians had higher age at presentation. The differences in the mean age at presentation among different groups were found to be statistically significant (P < 0.05).
Histo-morphological types
In the present study, various morphological variants and patterns were observed. The histomorphological types seen among 569 female breast cancers indicated that there were 502 cases (88.2%) with histology of IDC not otherwise specified (NOS), which was found to be the most common type. This was followed in decreasing order by infiltrating lobular carcinoma in 21 cases (3.7%); colloid carcinoma in 6 cases (1.1%), ductal carcinoma-in-situ in 6 cases (1.1%), metaplastic type in 5 cases (0.9%), schirrous carcinoma in 5 cases (0.9%), apocrine type in 4 cases (0.7%) and the rest 20 cases (3.5%) with other types of carcinoma (Table 1).
Table 1 Distribution of cases according to histomorphological types
Histomorphological type No. %
Infiltrating duct carcinoma. [IDC] 502 88.2
Infiltrating lobular carcinoma. [ILC] 21 3.7
Colloid carcinoma. 6 1.1
Ductal carcinoma-in-situ 6 1.1
Metaplastic carcinoma. 5 0.9
Schirrous carcinoma. 5 0.9
Apocrine carcinoma. 4 0.7
Medullary carcinoma. 6 1.1
Tubular carcinoma. 3 0.5
Cribriform carcinoma. 3 0.5
Mixed ductal & lobular carcinoma. 5 0.9
Other carcinomas 3 0.5
Total 569 100
Others consisted of Epidermoid, Papillary and juvenile Secretory carcinomas
Reproductive history
Out of 569 cases of breast cancer, it was observed that 6 cases (1.1%) were unmarried. Out of these, 3 were in the age group 20–29 years and remaining 3 cases were in the age group 30–39 years. Amongst the 563 married women, four cases (0.7%) revealed a history of nulliparity. Out of these 4 cases, 3 cases were in 40–49 years and remaining one case was in the age group 50–59 years old.
Bilaterality
It was noted that out of 569 cases, 564 (99.1%) cases presented with unilateral breast lump that was proven to be cancerous. Five cases presented with bilateral breast lumps, subsequently proven to be cancerous. Out of these, 2 were synchronous and 3 were metachronous. Among the 5 bilateral cases, 3 cases revealed a positive family history of cancer.
TNM staging
Of the 569 cases, TNM staging was available for 369 patients. The most commonly observed stage of presentation was IIIB with 130 cases (35.2%) followed in decreasing order of frequency by stages IIIA with 100 cases (27.1%), IIB with 60 (16.3%) cases, IV with 29 cases (7.9%) and stage I with only 5 cases (1.4%) (Table 2). Of the 29 metastatic cases, 6 cases had supraclavicular lymph nodes, 18 had hepatic and 5 cases had both hepatic and skin deposits. The mean age of cases at presentation was found to be 45.0 years (SD = 14.15) for cases in stage IIA as compared to 51.1(SD = 13.67) years for cases in stage IV. The mean age at presentation showed an increasing trend for the various stages, viz – stage IIB (48.4 years, SD = 10.69), IIIA (49.8 years, SD = 10.76), IIIB (50.6 years, SD = 11.6) and was found to be statistically significant (p < 0.05).
Table 2 Distribution of cases by stage, lymph node status and mean age.
Characteristics TNM grading No. Percentage Mean Age (SD)
TNM stage [N = 369]
Stage I T1N0M 5 1.4 N.C
Stage IIA 45 12.2 45.0(14.15)
T0N1M0 2
T1N1M0 5
T2N0M0 38
Stage IIB 60 16.3 48.4(10.69)
T2N1M0 37
T3N0M0 23
Stage IIIA 100 27.1 49.8(10.76)
T2N2M0 1
T3N1M0 67
T3N2M0 32
Stage IIIB 130 35.2 50.6(11.8)
T4 any NM0 126
Any T N3M0 4
Stage IV Any T any NM1 29 7.9 51.1(13.67)
Lymphnode status (N = 369)
Present 296 80.2 48.5(11.71)
Absent 73 19.8 45.8(13.43)
Percentages are against the total number of cases.
NC-Mean age (years) was not calculated due to small number of cases.
Lymph node status
Lymph node positivity was analysed histologically in all the 369 cases. Histopathological confirmation of lymph node involvement pN was observed in 296 cases (80.2%) (Table 2). The mean age of cases with lymph node positivity was found to be 48.5 (11.7) years as compared to 45.8 (SD = 13.43) years in negative cases. However, the differences were found to be statistically not significant
Association of religion with histo-pathological type, TNM staging and lymph node status
An analysis was carried out to understand, if any relationship existed between the religious affiliation of patient with histo-pathological type, TNM staging and lymph node status as the cultural patterns are different in the three religions (Table 3). The three histo-pathological types of disease viz. IDC, ILC and "all other tumors combined together" were considered for the above analysis. Percentage patients with these histopathological types were found to be similar in all the three religious groups (p = 0.66). Percentage patients presenting with localized tumors (TNM stage I & II) were found to be slightly higher amongst Christians (39.1%) as compared to other two religions (29.6% and 24.1%). However, the differences were not found to be statistically significant amongst the three religious groups (p = 0.84). Similarly, even the lymph node involvement was also observed to be almost similar in all the religious groups (p = 0.38).
Table 3 Association of religion with histological type, TNM stage and lymph node status [Percentage distribution]
Characteristics/Religion Hindu Muslim Christian
Histological type [N = 569] % % %
IDC 88.7 84.1 86.7
ILC 3.2 6.8 6.7
Others 8.0 9.1 6.7
TNM stage [N = 369]
I 1.3 0.0 4.3
IIa 12.3 13.8 8.7
IIb 16.0 10.3 26.1
IIIa 27.4 24.1 26.1
IIIb 34.9 44.8 30.4
IV 8.2 6.8 4.3
Lymphnode status [N = 369]
Present 79.2 79.3 91.3
Absent 20.8 20.7 8.7
Percentages are against the total number of cases seen in each religion.
Family history
Another important epidemiological parameter included from the year 1998 was family history. The information on family history was available for 226 cases. Out of these 226 cases, 47 cases (20.7%) revealed family history of cancer. Further, stratified analysis of data, by two broad groups of less than or equal to 44 years and above 44 years revealed that, of the 93 cases below the age group of 45 years, 27 (29.0%) women gave family history of cancer in their first or/and second degree relatives. Similarly, in the other age group of above 45 years, the frequency of cancer was reported to be in 20 (15.0%) women of the 133 women belonging to above 45 years of age. The differences in the occurrence of cancer amongst the first and second-degree relatives between the above two age groups were found to be statistically significant (p < 0.05). In the families of patients below 45 years a higher incidence of cancer was noted as compared to patients above 45 years of age. All women who had family history of breast cancer in their families were considered as cases, while the rest of women were taken as controls, for estimation of odd's ratio in the age groups below and above 45 years. This was found to be 2.3 (95% CI: 1.15–4.68, p < 0.02), indicating more frequent occurrence of cancer in relatives of patients below 45 years of age. Seventy two percent of cases (n = 34) with a positive family history revealed occurrence of breast or ovarian cancers amongst their family members. Remaining 28% of patients (n = 13) with a positive family history reported other cancers in their relatives such as lymphomas, thyroid cancer, prostate and colorectal carcinomas etc. Four patients (8.5%) gave the history of multiple cancers in their family members.
Further, an attempt was made to analyze the similarity in the histo-pathological type of tumor between the familial and non-familial groups of cases. It was observed that IDC formed as the most common type of tumor in both groups of cancer (68.8% and 75.5%). However, ILC was present in a higher proportion in familial group (12.5%) as compared to non-familial cases (5.4%). Further, "all other tumors combined together" were found to be 18.7% and 19.1% respectively. The differences, was not found to be statistically significant possibly due to small number of cases observed under the familial group (p > .10)
As regards the number of family members being affected with the cancers, two patients (4.3%) revealed history of cancer in more than 4 family members amongst their first and/or second- degree relatives. Fifteen patients (31.9%) had a history of cancer in two -three family members. Further, thirty patients (63.8%) mentioned that there was only one case of cancer amongst their relatives of first and second degree. It was observed that on an average, 1.51 first and/or second degree family members had suffered from cancer in these 47 families.
Discussion
Breast cancer incidence rates are increasing worldwide. In India, it is the most common cancer among women in many regions and has overtaken cervix cancer, which was the commonest cancer a decade ago. The continuing rise in breast cancer incidence has created an urgent need to develop strategies for prevention. Breast cancer appears to have a complex etiology, possibly with interplay of many causal factors including hormonal, genetic and environmental factors operating over a long period. Although several risk factors have been well defined, the interactions of the various etiological factors are yet to be completely understood. Moreover, analysis of newer parameters like family history could be helpful in screening high risk women for developing breast cancer, followed by planning of future, preventive and treatment modalities.
Age of the cancer patient is an important factor both for occurrence and management of the case. Average age of the patients seen in the six hospital based cancer registries in the National Cancer Registry Project (NCRP) network for the period 1994–98 was found to range from 44.2 years in Dibrugarh to 49.6 years in Bangalore and Chennai registries. In the present study the average age of the breast cancer case at presentation was found to be 47.9 years. Our findings are in agreement with the findings of the NCRP network [23]. Similarly, the average age of breast cancer patients has been reported to be 50 to 53 years in various population-based registries located in different parts of the country [7]. Similar hospital based studies carried-out at Delhi and Jaipur have also reported that the average age of breast cancer cases to be as 46.8 and 47 years [24,25]. The average age of occurrence of breast cancer amongst US white females has been reported to be 61.0 years [26]. The average age of occurrence of the breast cancer in India reveals that the disease occurs a decade earlier, as compared to western countries. The reason for early age of occurrence amongst Indian females needs to be further studied. A similar viewpoint has been put forward by a study conducted by Borovanova [27] in the Czech population. In their study, they found a shift of cancer more towards younger women.
Epidemiological studies carried out in the country have shown variation in the incidence of breast cancer among different religious groups such as Hindus, Muslims, Christians, Parsi and Buddhists. Breast cancer incidence by religion for greater Bombay population indicated highest incidence rates among Parsis and Christians and lowest rates among Jains and Buddhists. In another study from Chennai, the rates have been shown to be highest among Christians followed by Hindus and Muslims. Similarly in Trivandrum study, the incidence rates have been reported to be highest among Christians followed by Hindus and Muslims. Paymaster et al [14], in an analysis of hospital cancer cases by major religious groups have shown the variation in relative frequency between religious groups and in particular, high percentage among Parsi women compared to other religious communities. The reason for high incidence of breast cancer in Parsi community is as a result of their more westernized life-style, conserved genetic pool, high frequency of consanguineous marriages and higher age at the time of marriage and child birth [28]. In our study, on comparison of relative frequencies of cancer cases by various religious groups with that of population distribution of Delhi by these religious groups revealed that Christians had a higher percentage of cancer. Although Christians accounted for only 0.88% in the general population but relative frequency of cancer cases in the present study were 5.8%. The other two religions had more or less similar proportions. Although, the age at presentation amongst Christian patients was highest (51.5 years), as compared to other two religions, no significant difference was noticed in stage at presentation and lymph node status. The possible explanation for high incidence and higher age at presentation amongst Christians may be the same as in Western population.
Few studies of international variation in breast cancer have considered tumor histology. The histological type of carcinoma whether it is invasive or an in situ type are features which carry an inbuilt understanding of their general behavior pattern. In a study conducted by Wynder et al [4], comedo carcinoma and medullary carcinoma were found to be more frequent in Tokyo. Cases from a broader range of hospitals in Boston and Tokyo revealed occurrence of tumors of better prognostic types viz. intraductal, medullary and colloid types in Tokyo than in Boston. It is important to understand the relationship of histological type to etiology, and to allow separation of entities with distinct etiologies.
Histology as a prognostic factor has been well documented. Patients with histology of Infiltrating duct carcinoma (IDC)(NOS) have a poor survival compared to other types [17]. In the present study among the different histomorphological types, Infiltrating duct Carcinoma (NOS) was found to be the most common type i.e. in 502 cases (86.9%). The histological distribution of female breast cancers seen during the years 1984–93 in the hospital based cancer registries under the network of NCRP revealed that, in Mumbai, Bangalore and Thiruvananthapuram 83.3%, 85.0%, and 83.3% cases had a histology of infiltrating duct carcinoma while, 2.7%, 1.7%, 1.6%, belonged to lobular carcinoma, and the remaining 14.0%, 13.3%, 15.2% cases respectively were of other histopathological subtypes [23]. Infiltrating duct carcinoma was the commonest type seen in our series. The findings of the present study are almost in agreement with the findings reported by the hospital based cancer registries under the NCRP network. In the US population also, infiltrating ductal type of breast carcinoma was found to be the commonest histological type [29]. Among the age-correlation with various histomorphological types, prognostically better tumours like juvenile secretory were more commonly observed amongst younger females while infiltrating ductal carcinoma and lobular carcinoma remained the two most common cancers amongst elderly patients. It was also noted that ILC was present in a higher proportion in familial group (12.5%) as compared to non-familial cases (5.4%) in the present study, however it was not found statistically significant.
It is well known that breast cancer cases diagnosed at an earlier stage have a more favorable prognosis compared to those detected at late stage. However, because lack of awareness, fear of disease and psychological reasons, most of the patients in our country try to ignore or hide the disease and by the time they come to the hospital, the disease is already in the late stages. In the present study, over 90% of cases are in stages II, III and IV. The findings of the present study reveal that breast cancer frequently presents at higher stage i.e. 36.1% cases presenting with Stage IIIB in Indian population. This reflects a need for awareness and to initiate programs for early diagnosis of the cancer.
Numerous studies carried out in India and western populations have identified various reproductive factors generally associated with breast cancer [9-11,13,29-31]. A case-control study to identify risk factors for breast cancer carried out in Mumbai, India indicated that single women compared to married women had 4–5 fold higher risk for development of breast cancer in the age group of 40–54 years and 55 and above [14]. In another study it has been shown that nulliparous women had 2.2-fold higher risk than parous women [16]. High incidence of breast cancer among Parsi women was partly due to more unmarried women, late age at marriage and first child, less children and consanguinity of marriage [28]. Nulliparity and late age at first birth are the consistently observed reproductive risk factors [32]. In our study only 6 cases (1.0%) were found to be unmarried and four married woman (0.6%) were nulliparous.
Family history is another risk-factor for breast carcinoma. It has been noted that women who have first degree relative with breast cancer have a risk two to three times that of general population, the risk further increased if the relative was affected at an early age and/or had bilateral disease [20]. There is a greater risk if more than one close relative is affected, if breast cancer has occurred at a young age in a family member or if a patient has bilateral disease [30]. One of the explanations for familial aspects of breast cancer is germline mutation in BRCA1, BRCA2, p53 and other genes [33]. These cellular genes, which comprise dominantly acting oncogenes and recessively acting tumour suppressor genes, have been shown to contribute to genetic predisposition to variety of human cancers. In the present study, 20.2% cases revealed a positive family history of cancer. Out of these, higher percentages of cases 29.0% were observed in females under 44 years of age as compared to 15% in women of above 45 years of age. The presence of family history doubled the risk of subsequent breast cancer among younger women. In a similar study by Marcus et al, the relative risk of breast cancer was found to be doubled by the presence of a family history of breast cancer and amplified by younger age. It has been stated that there might be considerable underestimation of hereditary breast cancers. The studies carried-out at Jaipur and Delhi have reported the family history of cancer as 10% and 8% in their series. With documentation of pedigree, familial breast cancer (FBCs) may constitute as high as one third of the total incidence of breast cancers and approximately one forth of them would fall into the subset of hereditary breast cancer (HBC). It has been observed that familial breast cancer patients have an improved rate of survival, thereby indicating importance of noting familial cancer cases [33]. The present study clearly implies the importance of taking an appropriate history for eliciting family histories from the relatives. It is crucial to elicit detailed personal and family history extending back to atleast three generations, checking the medical records including pathology reports, wherever possible it is better to complete an accurate pedigree and taking family history from both maternal and paternal side of family. History of cancers among family members helps in identifying high risk groups, who can be counselled and subjected to careful follow ups with early diagnostic modalities and can even choose certain therapies. Improved ways of follow up with study of various interacting genes would also be useful to identify high-risk groups. Therefore, it is imperative to include parameters like family history for better understanding of breast cancer causation and predisposition.
Breast cancer usually presents with a single hard lump as is evident in the present study with occurrence of bilaterality in only 5 out of total 569 cases. In an earlier study, bilaterality was reported in 3% of breast cancer patients [34]. FBC have been observed to be having improved survival, as a result of stringent monitoring which may help in diagnosis at early stage [30,35]. Clinical stage is another factor implicated. In a study conducted by Langlands et al [35], FBC cases had a lower stage of presentation. However, in our study, no such association was observed.
Conclusion
Continuing increased incidence of breast cancer has added urgency to investigations of control and preventive measures. It is important to incorporate primary and secondary prevention measures of breast cancer as number of affected individuals is rising and the age of onset is shifting towards younger age groups. Efforts should be made to detect breast cancer at the very early stage through periodic screening of high-risk groups either by physical self-examination or by self-breast examination. Mammography will be difficult to implement in Indian situation for the control of breast cancer. Efforts should be made to elicit information related to factors like family history of breast cancer, thereby identifying high-risk groups. There is a need to educate both high risk and all women about the importance of Breast self Education (BSE) through press and electronic media, thus more cases can be diagnosed at an early stage and effective treatment can be given to those women and their lives can be saved.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SS, The study was carried out under the overall supervision and guidance, have given final approval of the version.
BR Involved in the diagnostic and analytical part of the study, alongwith preparation of the manuscript
AB, Involved in the diagnostic and analytical part of the study
As B, Collection of data
C, Collection of clinical details & preparation of manuscript
NSM, Design, analysis and interpretation of data and preparation of the manuscript
Funding
None declared
==== Refs
Parkin DM Bray F Ferlay J Pisani P Estimating the world cancer burden: Globocon 2000 Int J Cancer 2001 94 153 156 11668491 10.1002/ijc.1440
Parkin DM Pisani P Ferlay J Estimates of worldwide incidence of 25 major cancers in 1990 Int J Cancer 1999 80 827 841 10074914 10.1002/(SICI)1097-0215(19990315)80:6<827::AID-IJC6>3.0.CO;2-P
Parkin DM Global cancer statistics in the year 2000 Lancet Oncol 2001 2 532 542 10.1016/S1470-2045(01)00486-7
Wynder EL Kajitani T Kuno J Lucas JC JrDepalo A Farrow J A comparison of survival rates between American and Japanese patients with breast cancer Surg Gynecol Obstet 1963 117 196 200 14048012
Murthy NS Juneja A Sehgal A Parbhakar AK Luthra UK Cancer projection by the turn of the century – Indian scene Ind J Cancer 1990 27 74 82
Siddiqi M Sen U Mondal SS Patel DD Yeole BB Jussawala DJ Pratinidhi A Sunny L Yeole BB Jussawala DJ Rao BN Sunny L Yeole BB Jussawala DJ Varsha S Sunny L Nair MK Varghese C Nair MK Gangadharan P Jayalekshmi P Cancer statistics from non-ICMR registries: Population based registries, CRAB (Cancer registry Abstract). Newsletter of the National Cancer Registry Project of India 2001 8 47 59
National Cancer registry Programme Consolidated report of the population based cancer registries 1990–1996 Indian Council of Medical Research, New Delhi 2001
Badwe RA Gangawal S Mittra I Desai PB Clinico-pathological features and prognosis of breast cancer in different religious communities in India Ind J Cancer 1990 27 220 228
De Ward F Breast cancer incidence in nutritional status with particular reference to body weight and height Cancer Res 1975 35 3351 3356 1104147
Gajalakshmi CK Shanta V Risk factors for female breast cancer. A hospital based case-control study in Madras, India Acta Oncol 1991 30 569 574 1892673
Jussaawala DJ Yeole BB Natekar MV Histological and epidemiological features of breast cancer in different religious groups in greater Bombay J Surg Oncol 1981 18 269 279 6273657
Jussawaala DJ Jain DK Breast cancer and religion in Greater Bombay women: An epidemiological study of 2130 women over a 9-year period Br J Cancer 1970 36 634 638
Jussawalla DJ Yeole BB Natekar MV Narayanan R Epidemiology of breast cancer in India Indian J Cancer 1975 12 231 242 1193654
Paymaster JC Gangadharan P Some observations on the epidemiology of cancer of breast in women in western India Int J Cancer 1972 10 443 450 4668492
Rao DN Ganesh B Estimate of cancer incidence in India in 1991 Indian J Cancer 1998 35 10 18 9847465
Rao DN Ganesh B Desai PB Role of reproductive factors in breast cancer in a low-risk area: a case-control study Br J Cancer 1994 70 129 32 8018523
Rao DN Dinshaw KA Epidemiological review: In Risk factors and survival rates Hospital Cancer Registry Division of Epidemiology and Biostatistics Tata Memorial Hospital Mumbai 1999 1 7
Anderson DE Some characteristics of familial breast cancer Cancer 1971 28 1500 1504 5127797
Ellisen Leif W Daniel Harber A Hereditary Breast Cancer Ann Rev Med 1998 49 425 436 9509273 10.1146/annurev.med.49.1.425
Skolnick MH Cannon Albright LA Genetic predisposition to breast cancer Cancer 1992 70 1747 1754 1516030
UICC Harmer MH TNM classification of malignant Tumors 1978 Third International Union Against Cancer, Geneva
Census of India Census Data Online 2001 [accessed April, 2005]
National Cancer registry Program Ten year consolidated report of the Hospital Based Cancer Registries, 1984–1993, An assessment of the burden and care of cancer patients Indian Council of Medical Research, New Delhi 2001
Gupta P Sharma RG Verma M Review of breast cancer cases in Jaipur region J Ind Med Assoc 2002 100 282 284
Goel AK Seenu V Shukla NK Raina V Breast cancer presentation at a regional cancer center Natl Med J Ind 1995 8 6 9
Parkin DM Whelan SL Ferlay J Storm H Cancer Incidence in Five Continents Vol.VIII. International Agency for research on Cancer [IARC], Lyon, France, IARC Scientific Publication No. 155 2002
Borovanova T Soucek P Breast cancer: An overview of factors affecting the onset and development of the disease Cab Lek Cesk 2002 141 80 89
Paymaster TC Gangadharan P Cancer in Parsi community of Bombay Int J Cancer 1970 5 426 431 5452072
Correa P The epidemiology of cancer of the breast Am J Clin Pathol 1975 64 720 727 173175
Claus EB Rish N Thomson WD Age at onset as an indicator of familial risk of breast cancer Am J Epidemiol 1990 131 961 972 2188501
Anderson DE Genetic considerations in breast cancer Breast cancer: Early and Late 1970 Chicago, yearbook Medical Publishers, Inc 27 36
Kesley JL Gammon MD John EM Reproductive factors and breast cancer Epidemiol Rev 1993 15 36 47 8405211
Saxena S Szabo CI Chopin S Barjhoux L Sinilnikova O Lenoir G Goldgar DE Bhatanagar D BRCA1 and BRCA2 in Indian Breast Cancer patients Hum Mut 2002 20 473 474 12442273 10.1002/humu.9082
Monsour EG Kavdin PM Dressler L Prognostic factors in early breast carcinoma Cancer 1994 74 381 400 8004612
Langlands AO Kerr GR Bloomer SM Familial breast cancer Oncol 1976 2 41 45
|
16236180
|
PMC1277852
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Oct 13; 3:67
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-67
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7343ehp0113-00011915687047ResearchCommentaryWill Investments in Large-Scale Prospective Cohorts and Biobanks Limit Our Ability to Discover Weaker, Less Common Genetic and Environmental Contributors to Complex Diseases? Foster Morris W. 1Sharp Richard R. 21Department of Anthropology, University of Oklahoma, Norman, Oklahoma, USA2Center for Medical Ethics and Health Policy, Baylor College of Medicine, Houston, Texas, USAAddress correspondence to M.W. Foster, Department of Anthropology, 455 W. Lindsey, Room 505C, University of Oklahoma, Norman, OK 73019 USA. Telephone: (405) 325-2491. Fax: (405) 325-7386. E-mail:
[email protected] publication was made possible by grant ES11174 from the National Institute of Environmental Health Sciences (NIEHS) and grant HG02691 from the National Human Genome Research Institute. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS, National Human Genome Research Institute, or National Institutes of Health.
The authors declare they have no competing financial interests.
2 2005 4 11 2004 113 2 119 122 22 6 2004 3 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Increasing the size of prospective cohorts and biobanks is one approach to discovering previously unknown contributors to complex diseases, but it may come at the price of concealing contributors that are less common across all the participants in those larger studies and of limiting hypothesis generation. Prospective cohorts and biobanks constitute significant, long-term investments in research infrastructure that will have ongoing consequences for opportunities in biomedical research for the foreseeable future. Thus, it is important to think about how these major additions to research infrastructure can be designed to be more productive in generating hypotheses for novel environmental contributors to complex diseases and to help identify genetic and environmental contributors that may not be common across the larger samples but are more frequent within local or ancestral subsets. Incorporating open-ended inquiries and qualitative information about local communal and ecologic contexts and the political, economic, and other social structures that affect health status and outcome will enable qualitative hypothesis generation in those localized contexts, as well as the collection of more detailed genealogic and family health history information that may be useful in designing future studies. Using communities as building blocks for larger cohorts and biobanks presents some practical and ethical challenges but also enhances opportunities for interdisciplinary, multilevel investigations of the multifactorial contributors to complex diseases.
biobankscommunitiescomplex diseasegene–environment interactionprospective cohortsqualitative methodsresearch design
==== Body
Of the approximately 30,000 genes in the entire human genome, > 1,500 genetic variants have been discovered in which a single allele (either as a homozygote or heterozygote) is sufficient for a single gene or Mendelian disorder such as Huntington’s disease to develop (National Center for Biotechnology Information 2004). However, relatively few variants have been confirmed for complex diseases such as cancer, heart disease, and diabetes in which both susceptibility genes and environmental contributors are required for the disease to develop (Botstein and Risch 2003; Hirschhorn et al. 2002). The slow pace in identifying and confirming genetic contributors for complex diseases is due primarily to the difficulties of detecting relatively weak, incremental genetic effects as well as to the possibility that even moderate or strong effects involving a genetic contributor may require the co-occurrence of one or more environmental contributors (Hodgson and Popat 2003).
Similarly, although the identity and function of some environmental contributors to complex diseases such as cancer are well known (toxicants such as asbestos, behaviors such as smoking, viruses such as human papilloma virus), almost all of these known contributors have been identified as such because they have relatively strong effects on disease susceptibility. At the same time, however, a significant proportion of environmental contributors remain unknown for many complex diseases. For example, only one-third of the breast cancer cases in the United States can be accounted for by known risk factors (Stevens 2002). The overwhelming remainder involves either candidate risk factors that are known but have not yet been confirmed as such (which raises the cases accountable to ~50%) or risk factors that are not recognized as such at all. Moreover, even already-identified risk factors for disease such as diet, tobacco, and hormones each are composed of complicated combinations of behaviors and toxicants whose roles in carcinogenesis are not well understood (Brennan 2002). Smoking, for instance, is a contextually shaped behavior that can take a variety of often culturally specific forms as it exposes those who perform it (and others around them) to > 300 different toxicants (Chassin et al. 2000; Frohlich et al. 2002).
In response to these current limitations, a number of researchers have suggested scaling up research sample sizes to provide greater statistical power for identifying and confirming genetic and environmental contributors to complex diseases (Caporaso 2002; Collins 2004; Little et al. 2003; Millikan 2002). Efforts in scaling up sample sizes involve significant national and private investments in research infrastructure. Governmental and nonprofit funding agencies as well as for-profit ventures in various countries are in the process of planning or assembling larger scientific resources to meet that perceived need.
Some of these larger sample collections are in the form of prospective cohorts that recruit healthy participants with the intention of following their health status over a number of years. For example, the National Institute of Child Health and Human Development along with the National Institute of Environmental Health Sciences, the Centers for Disease Control and Prevention, and the U.S. Environmental Protection Agency has been planning a National Children’s Study designed to follow 100,000 children and their parents over multiple decades (National Children’s Study 2004), and the National Cancer Institute (NCI) has recently issued a new call for proposals for funding large prospective cohorts (NCI 2003). The NCI already funds the Black Women’s Cohort (64,500 participants) and the California Teachers Study (133,479 participants) among other large prospective studies (NCI 2004). The NCI announcements of funding for prospective cohorts explicitly contrast them with previous investments in cross-sectional or case–control studies, characterizing cohorts as more flexible, longer-lasting investments in research infrastructure. Most recently, the National Human Genome Research Institute, in collaboration with the National Heart, Lung, and Blood Institute, has requested information from researchers in planning a national cohort of 500,000 participants (National Institutes of Health 2004). In Europe, there is a long tradition of birth cohort studies that extend decades into adulthood, with recent investments in new birth cohorts by the United Kingdom and planning for a “mega” cohort by the European Union (Kogevinas 2002).
A related kind of resource, often called biobanks, incorporates members of national or regional populations for which extensive retrospective medical records, DNA samples, and other health-related information are available to researchers (Austin et al. 2003). Some biobanks also function as prospective cohorts. The deCODE project, for instance, already has isolated genes that appear to contribute to osteoporosis, stroke, diabetes, and several other complex diseases using historical and contemporary health information and DNA samples from more than 100,000 residents of Iceland (deCODE Genetics 2004), although some of those findings may turn out to be limited to rarer familial factors. Similar biobanks are being assembled in Estonia (an open-ended number of participants), the United Kingdom (500,000 participants), Quebec (60,000 participants), and Japan (300,000 participants). In the United States, Howard University has announced the formation of a biobank with samples from participants who identify themselves as African Americans (Kaiser 2003).
Investments made today in prospective cohorts and biobanks that are projected to be used (and funded) for decades to come will have significant consequences for determining both the opportunities and the limits of future research into genetic and environmental contributors to complex diseases. Although it will be possible to establish new cohorts and biobanks in the future, it will be several years before prospectively recruited participants develop diseases of interest in sufficient numbers for analysis. Moreover, funding for additional cohorts in future years will compete with the costs of maintaining ongoing cohorts, which likely will limit future growth in this research infrastructure. Consequently, as cohorts and biobanks are being planned, it is important to consider the methodologic implications that their increased scales may have for identifying genetic and environmental contributors that may be more locally variable in effect. Locally variable or less common contributors nonetheless can have significant effects on health disparities, raising questions about the equitable distribution of research benefits in the case of large, expensive cohorts that may not be designed to attend to smaller-scale contexts.
Is Bigger Always Better?
Although larger cohorts or biobanks likely will help identify many genetic and environmental contributors that are more common among their members, they will be less likely to help identify those less common contributors that are rare among most participants. Indeed, the additional power that a larger cohort provides to detect weaker common effects simultaneously can mask those contributors that are localized primarily within subsets of the larger sample, depending on how cohort information is collected and analyzed. For instance, a genetic variant that is more frequent among individuals of a particular ancestry but rare among others may not be detected in a sample of 100,000 participants recruited using such inclusion criteria as regional or national residence or occupation. Similarly, an environmental contributor that is specific to exposures resulting from a local ecologic feature or a locally specific behavior also could be lost in a large, multisite cohort, even though it may be a significant determinant of disease. This means that the ways in which participants are categorized and recruited for a particular cohort or biobank and in which their information is collected and analyzed will affect what studies using that resource may find as well as what they may miss.
A criticism of the UK Biobank, for instance, has been that it has no specific plans to incorporate a familial component into its recruitment strategy (Wright et al. 2002). Family members (particularly sibling pairs and parents) provide greater power for separating genetic effects from the background noise of nongenetic effects. In addition, there also tend to be correlations in common environmental and gene–environment interactions among close relatives compared with random, unrelated individuals. Thus, the larger size of a cohort may not necessarily increase its power to detect genetic or environmental contributors to complex diseases.
That situation is complicated further by the possibility that the same complex disease may have multiple genetic and environmental contributors that are neither necessary nor sufficient for a similar phenotype to be expressed (Smith and Lusis 2002). In the cases of type 2 diabetes and systemic lupus erythematosus, for example, different candidate genes have been proposed from studies of geographically and ancestrally differing patient populations, although some of those will not be confirmed (Kelly et al. 2002; Stern 2002). In the case of breast cancer (as for the vast majority of other cancers), not all confirmed environmental contributors need be present for the disease to develop. With the additional variable of gene–environment interactions, it may well be that some significant (although still minority) proportions of the incidence of most complex diseases are attributable to intersections of locally varying combinations of genetic and environmental contributors some or even many of which may not be detectable in large multisite samples. To the extent that those polygenic and polyenvironmental contributors are nonrandomly distributed among and across populations, a large cohort or biobank may fail to detect some or even most of these unless it is structured to support more intensive study of subsets of participants.
The greater cost of larger cohorts, however, tends to mean that fewer and often less precise measures are obtained for each participant, a situation that actually can reduce the power of a larger sample (Wong et al. 2003). Sampling costs also can reduce the ability to collect information that is most productive for hypothesis generation. Because it is expensive to investigate family histories and environmental exposure histories for large numbers of participants (Barbour 2003), large cohort studies tend to collect participant information through closed-ended questions—that is, by giving participants a range of predetermined answers to predetermined questions and forcing them to choose among them (UK Biobank 2002). For environmental exposures, closed-ended questions are useful in testing hypotheses about established or suspected contributors but are of limited value in identifying previously unsuspected contributors whether those are localized or more common (Foster and Aston 2003). With respect to ancestry, some studies allow participants to indicate more than one ethnic or racial background but without eliciting additional information that may be more informative about how genetic variants are distributed in the extensive middle ground between immediate family members and large population categories such as European American or African American.
These limited, closed-ended responses frequently are used as proxies for a shared population history (in the case of ancestry) or for shared environmental exposures (or both) for purposes of sample stratification. The difficulty, however, is that such broad, decontextualed proxies often are treated as units of analysis rather than as heuristic means to disambiguate or discover specific ancestral and environmental contributors to disease or to provide a degree of diversity within the sample frame.
Identity alone, however, is not causal and may not even necessarily be predictive. First, not all factors linked to a given identity necessarily contribute to disease expression or to the expression of the same diseases. Second, only some environmental and ancestral factors are shared among those with a common identity. Third, only some of those with a common identity necessarily share those linked factors. Social identity does become a more powerful predictor, however, when it intersects with ecology in a specific locality. Sharing both a social identity and a locality increases the likelihood that and the extent to which a social community will regulate the actions of its members according to some standard of appropriateness (and, hence, manifest many of the same behavioral environmental factors), the likelihood that community members are exposed to many of the same ambient factors in the physical environment, and the degree of access to prevention, surveillance, and treatment available to community members. Locality also may limit significantly the number of ancestries shared among co-residents while increasing the likelihood that some are related by more immediate genealogic connections.
Communities as Building Blocks
These critiques suggest that a significant challenge in constructing large-scale cohorts and biobanks is to design a study with a large number of participants that nonetheless gathers rich data on individuals and the contexts that affect their health, providing flexibility for discovering unanticipated data fields and new categories within existing fields. One solution may be to use the local communities in which individuals are everyday members—a naturally occurring social middle ground between single participants and very large ethnic and other categories—as building blocks for constructing large prospective samples. Local communities also would be appropriate contexts for recruiting parents and siblings to enrich the familial component of cohorts and biobanks.
Local communities, of course, may be quite variable in form, ranging from relatively well-defined residential clusters or towns in rural areas to neighborhoods or social networks within large metropolitan areas. What defines a localized community, however, is that its members share similar interactional conventions, a consequence of their everyday encounters with one another, as well as similar ambient or background exposures due to the local physical environment.
The idea that locality or place may affect health is not new (Durkheim 1951). However, the last decade has seen a revival of interest in theorizing and conceptualizing that relationship (Curtis and Rees-Jones 1998; Kearns and Joseph 1993; Macintyre et al. 1993; Tunstall et al. 2004). In contrast with the prevailing epidemiologic focus on individual risk factors, this revival has emphasized collective or contextual effects that may mediate the effects of individual-level variables such that the health status of individuals depends to some extent on the social and physical environments in which individuals grow up and live (Schwarz 1994; Susser 1994). The proponents of this approach argue that collective or contextual “area effects” are complex, multilevel interactions involving phenomena or forces ranging from global, national, or regional social structures that determine opportunities and limitations for well-being (including economic systems and conditions, health care systems and access, political structures and equity, and widespread cultural beliefs and social practices) to more localized communal beliefs, practices, and conditions to diverse intracommunity patterns of individual agency (Macintyre et al. 2002; Popay et al. 1998). Thus, rather than adopt the traditional epidemiologic practice of isolating and testing one environmental factor at a time while attempting to control for the effects of others, a more appropriate method of analysis may be to embrace the complexity of multilevel collective or contextual contributors.
Fine-grained information about contextual effects in local communities offers two primary advantages in studies of environmental contributors to disease susceptibility. First, those data provide additional background information that can be used to better interpret responses to standardized questions, but in ways that still allow comparison across the larger sample. For instance, the same ethnic identity or household income level can indicate differing health risks and outcomes depending on such locally variable contributors as beliefs about health and illness, familial and communal social dynamics and networks, and political and economic structures (Krieger 2001; Williams 2003). Each of these parameters (along with others) helps shape everyday life in ways that can have differing consequences for behaviors that may expose individuals to environmental toxins and may be further differentiated by local variations in physical environments and the ambient exposures that those offer. Detailed investigations of these local differences can augment an understanding of the pathways by which social and ecologic factors contribute to disease susceptibility or can explain why a risk factor does not appear to be as predictive for a specific sub-population (Frohlich et al. 2001).
Second, detailed local investigations allow many more opportunities for hypothesis generation, which then can be tested across the larger sample. Epidemiologic tests for the statistical significance of associations between established proxies such as ethnicity or socioeconomic status and disease incidence or mortality offer few opportunities for generating novel hypotheses about environmental contributors, mainly because those proxies summarize rather than disaggregate specific environmental factors. In contrast to proxies that summarize information, a community-specific approach that produces large amounts of in-depth information about a broad range of aspects of everyday life provides many specific possibilities for generating hypotheses (Brown 2003; Thompson and Gifford 2000). Indeed, generating hypotheses in small-scale contexts is preferable to doing so across large multisite samples because the former is more amenable to qualitative studies of the different ways in which a large number of factors interact with one another, whereas the latter is more suited to testing hypotheses about a limited number of well-defined, measurable data points.
One of the primary difficulties in using large samples to detect gene–environment interactions is that most nongenetic influences are difficult to measure such that they often are dismissed as being beyond investigation in large samples (Wright et al. 2002). Rather than simply ignore those influences in a larger sample because they cannot be measured accurately or efficiently using existing metrics, qualitative, community-specific approaches offer the possibility of developing a functional understanding of how their effects are achieved, which may help develop accurate, efficient measures that then can be applied in quantitative analyses of larger samples. For example, qualitative data gathered using a “life course” approach can be analyzed to identify biologic and social factors that affect health throughout life in a cumulative manner (both independently and interactively), develop measures of their effects, and describe chains or pathways of risk by which linked exposures raise the likelihood of disease expression (Ben-Shlomo and Kuh 2002; Hallqvist et al. 2004; Kuh and Ben-Shlomo 1997; Kuh et al. 2003). Thus, qualitative methods such as ethnography may become an interdisciplinary companion to epidemiology (Kaufman and Cooper 2001; O’Campo 2003).
Practical and Ethical Challenges
Taking a community-specific approach does raise several logistical and ethical issues that may be problematic. An immediate reaction to our proposal is likely to be concern about the additional cost of recruiting participants to comprise both local community units and the overall cohort, as well as the additional cost of in-person elicitation of open-ended ethnographic and genealogic information. However, given the significant investments already required by very large prospective cohorts or retrospective biobanks, incurring additional costs to enrich the information collected, particularly with respect to hypothesis generation, should be seen as enhancing the value of what will become long-term investments in biomedical infrastructure. A less expensive alternative could be to recruit some but not all participants as members of community units, with the idea that hypothesis generation need not involve all cohort or biobank participants. Indeed, community units may be selected within the larger scale of the study as a whole in two ways: as models that are representative of most study participants (and so have a likelihood of generating hypotheses that may be tested quantitatively across most participants to identify more common contributors) or as efforts to make the cohort more diverse by including participants whose identities contain elements (e.g., ancestry, residence, occupation, household income) that may evidence some contributors to disease differing from those that are more common within the larger cohort. Both strategies add value to the cohort as a whole, albeit in different ways.
With respect to the latter strategy, a frequent problem in making a participant pool more diverse is that including subjects who may represent minority experiences of disease does not necessarily ensure sufficient power to stratify the sample to quantitatively analyze the less common contributors that may affect those more diverse participants. However, by recruiting some of those minority participants as members of community units, they can be oversampled by the greater detail of information collected rather than by attempting to recruit larger numbers of participants who fit those less common inclusion criteria.
A community-specific strategy also presents several ethical challenges. For example, investigators will need to consider when the additional subject interactions become overly burdensome on particular populations—for example, minority communities that may have been studied extensively in the past. Collecting large amounts of in-depth data about participants, family members, and local communities also presents somewhat greater ethical challenges than do responses to closed-ended questions. Although maintaining confidentiality is a requirement in both cases, it is more difficult to anticipate the risks that might accrue from open-ended inquiries. Moreover, gathering additional information about communities as wholes and about third-party relatives may entail the potential for risks to others than just study participants. For example, published indications of greater genetic susceptibility to a disease among individuals of a specific ancestry or of greater environmental risks to those who reside in a particular place or pursue a particular lifestyle may put those with that ancestry, residence, or lifestyle at a greater risk for discrimination or stigmatization.
At the same time, community-specific investigation often creates a stronger relationship between researchers and participants that should tend to produce greater trust and, hence, more extensive and accurate responses as well as reduced attrition in multiyear and multidecade studies. Emphasizing communities makes it possible to engage pre-existing social organizations and networks in evaluating (and possibly modifying) ethical protections and recruitment strategies, in assisting in participant recruitment and liaison, in actually collecting some study information, and in helping construct local interpretations of the information collected (Sharp and Foster 2000). This greater attention to local contexts should result in greater participant influence in shaping how research is done and greater investigator awareness of local community needs.
Conclusion
The future of biomedical research should reside both in “small science” and in “big science.” The two approaches are not necessarily mutually exclusive, although the larger scale of the latter may limit the scale of information that is collected from participants. We believe that larger cohorts and biobanks need not preclude smaller, finer-grained investigations of community-specific influences on disease. In fact, qualitative, community-specific investigations are not only possible within the context of those increasingly large-scale investigations but can provide opportunities for additional hypothesis generation as well as facilitate the multilevel analysis of individual, contextual, and structural factors that contribute to complex diseases.
==== Refs
References
Austin MA Harding S McElroy C 2003 Genebanks: a comparison of eight proposed international genetic databases Community Genet 6 1 37 45 12748437
Barbour V 2003 UK Biobank: a project in search of a protocol? Lancet 361 1734 1738 12767753
Ben-Shlomo Y Kuh D 2002 A life course approach to chronic disease epidemiology: conceptual models, empirical challenges, and interdisciplinary perspectives Int J Epidemiol 31 285 293 11980781
Botstein D Risch N 2003 Discovering genotypes underlying human phenotypes: past successes for Mendelian disease, future approaches for complex disease Nat Genet 33 suppl 228 237 12610532
Brennan P 2002 Gene-environment interaction and aetiology of cancer: what does it mean and how can we measure it? Carcinogenesis 23 3 381 387 11895852
Brown P 2003 Qualitative methods in environmental health research Environ Health Perspect 11 1789 1798 14594634
Caporaso NE 2002 Why have we failed to find the low penetrance genetic constituents of common cancers? Cancer Epidemiol Biomarkers Prev 11 12 1544 1549 12496041
Chassin L Presson CC Pitts SC Sherman SJ 2000 The natural history of cigarette smoking from adolescence to adulthood in a midwestern community sample: multiple trajectories and their psychosocial correlates Health Psychology 19 3 223 231 10868766
Collins FS 2004 The case for a US prospective cohort study of genes and environment Nature 429 6990 475 477 15164074
Curtis S Rees-Jones I 1998 Is there a place for geography in the analysis of health inequality? Soc Health Illness 20 645 672
deCODE Genetics 2004. From genes to drugs. Reykjavik, Iceland:deCODE Genetics. Available: http://www.decode.com/ [accessed 14 October 2004].
Durkheim E 1951 [1897]. Suicide: A Study in Sociology (Spaulding JA, Simpson G, trans). New York:Free Press.
Foster MW Aston CE 2003 A practice approach for identifying previously unsuspected environmental contributors to systemic lupus erythematosus and other complex diseases Environ Health Perspect 111 593 597 12676621
Frohlich KL Corin E Potvin L 2001 A theoretical proposal for the relationship between context and disease Sociol Health Illness 23 6 776 797
Frohlich KL Potvin L Gauvin L Chabot P 2002 Youth smoking initiation: disentangling context from composition Health Place 8 3 155 166 12135639
Hallqvist J Lynch J Bartley M Lang T Blane D 2004 Can we disentangle life course processes of accumulation, critical period, and social mobility? An analysis of disadvantaged socio-economic positions and myocardial infarction in the Stockholm Heart Epidemiology Program Soc Sci Med 58 1555 1562 14759698
Hirschhorn JN Lohmueller K Byrne E Hirschhorn K 2002 A comprehensive review of genetic association studies Genet Med 4 2 45 61 11882781
Hodgson SV Popat S 2003 Polymorphic sequence variants in medicine: a challenge and an opportunity Clin Med 3 3 260 264 12848262
Kaiser J 2003 African-American population biobank proposed Science 300 1485 12791950
Kaufman JS Cooper RS 2001 Commentary: considerations for use of racial/ethnic classification in etiologic research Am J Epidemiol 154 291 298 11495850
Kearns RA Joseph AE 1993 Space in its place: developing the link in medical geography Soc Sci Med 37 711 717 8211286
Kelly JA Moser KL Harley JB 2002 The genetics of systemic lupus erythematosus: putting the pieces together Genes Immunity 3 suppl 1 S71 S85 12215907
Kogevinas M 2002. Expression of interest for an integrated project: European Union birth-cohort on child health and human development. Luxembourg:Community Research and Development Information Service. Available: http://eoi.cordis.lu/docs/int_28461.doc [accessed 14 October 2004].
Krieger N 2001 Theories of social epidemiology in the 21st century: an ecosocial perspective Int J Epidemiol 30 668 677 11511581
Kuh D Ben-Shlomo Y eds. 1997. A Life Course Approach to Chronic Disease Epidemiology. Oxford, UK:Oxford University Press.
Kuh D Ben-Shlomo Y Lynch J Hallqvist J Power C 2003 Life course epidemiology J Epidemiol Community Health 57 778 783 14573579
Little J Khoury MJ Bradley L Clyne M Gwinn M Lin B 2003 The Human Genome Project is complete. How do we develop a handle for the pump? Am J Epidemiol 157 8 667 673 12697570
Macintyre S Ellaway A Cummins S 2002 Place effects on health: how can we conceptualise, operationalise, and measure them? Soc Sci Med 55 125 139 12137182
Macintyre S Mciver S Soomans A 1993 Area, class, and health: should we be focusing on places or people J Soc Policy 22 213 234
Millikan R 2002 The changing face of epidemiology in the genomics era Epidemiology 13 4 472 480 12094104
National Center for Biotechnology Information 2004. Online Mendelian Inheritance of Man. National Center for Biotechnology Information. Available: http://www.ncbi.nlm.nih.gov/Omim/mimstats.html [accessed 14 October 2004].
National Children’s Study 2004. What Is the National Children’s Study? Available: http://nationalchildrensstudy.gov/about/mission/overview.cfm [accessed 14 October 2004].
National Institutes of Health 2004. Request for Information: Design and Implementation of a Large-Scale Prospective Cohort Study of Genetic and Environmental Influences on Common Diseases. NOT-OD-04-041. Bethesda, MD:National Institutes of Health. Available: http://grants2.nih.gov/grants/guide/notice-files/NOT-OD-04–041.html [accessed 14 October 2004].
NCI 2003. Cohort Studies in Cancer Epidemiology (PAR-04-011). National Cancer Institute. Available: http://grants1.nih.gov/grants/guide/pa-files/PAR-04-011.html [accessed 14 October 2004].
NCI 2004. Consortium of Cohorts: Comparison of Selected Characteristics among Large Cohorts. Available: http://ospahome.nci.nih.gov/cohort/table.htm [accessed 14 October 2004].
O’Campo P 2003 Invited commentary: advancing theory and methods for multilevel models of residential neighborhoods and health Am J Epidemiol 157 9 13 12505885
Popay J Williams G Thomas C Gatrell A 1998 Theorising inequalities in health: the place of lay knowledge Sociol Health Illness 20 619 644
Schwarz S 1994 The fallacy of the ecological fallacy—the potential misuse of a concept and the consequences Am J Pub Health 85 819 824
Sharp RR Foster MW 2000 Involving study populations in the review of genetic research J Law Med Ethics 28 41 51 11067631
Smith DJ Lusis AJ 2002 The allelic structure of common disease Hum Mol Genet 11 2455 2461 12351581
Stern MP 2002 The search for type 2 diabetes susceptibitity genes using whole-genome scans: an epidemiologist’s perspective Diabetes Metab Res Rev 18 2 106 113 11994901
Stevens RG 2002 Lighting during the day and night: possible impact on risk of breast cancer Neuroendocrinol Lett 23 suppl 2 57 60 12163850
Susser M 1994 The logic in ecological. 1. The logic of analysis Am J Public Health 84 825 829 8179056
Thompson SJ Gifford SM 2000 Trying to keep a balance: the meaning of health and diabetes in an urban Aboriginal community Soc Sci Med 51 10 1457 1472 11077950
Tunstall HVZ Shaw M Dorling D 2004 Places and health J Epidemiol Community Health 58 6 10 14684719
UK Biobank 2002. Protocol for the UK Biobank: a study of genes, environment and health. Available: http://www.ukbiobank.ac.uk/documents/draft_protocol.pdf [accessed 14 October 2004].
Williams GH 2003 The determinants of health: structure, context and agency Sociol Health Illness 25 131 154
Wong MY Day NE Luan JA Chan KP Wareham NJ 2003 The detection of gene-environment interaction for continuous traits: should we deal with measurement error by bigger studies or better measurement? Int J Epidemiol 32 1 51 57 12690008
Wright AF Carothers AD Campbell H 2002 Gene-environment interactions—the BioBanks UK study Pharmacogenomics J 2 75 82 12049178
|
15687047
|
PMC1277853
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 4; 113(2):119-122
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7343
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7254ehp0113-00012315687048ResearchReviewDevelopmental Neurotoxicity of Pyrethroid Insecticides: Critical Review and Future Research Needs Shafer Timothy J. 1Meyer Douglas A. 2Crofton Kevin M. 121Neurotoxicology Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, USA2Curriculum in Toxicology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USAAddress correspondence to T.J. Shafer, Neurotoxicology Division, MD B105-05, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 USA. Telephone: (919) 541-0647. Fax: (919) 541-4849. E-mail:
[email protected] thank D. Ray (Medical Research Council, UK) and Bayer CropScience for graciously making available unpublished data for this review; L. Sheets (Bayer CropScience) and S. Padilla (U.S. EPA) for comments on a previous version of the manuscript; and J. Harrill (University of North Carolina at Chapel Hill) and J. Havel (CSC Corporation) for their assistance with figures and graphics.
Preparation of this document has been funded wholly by the U.S. Environmental Protection Agency. This document has been subjected to review by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.
The authors declare they have no competing financial interests.
2 2005 14 10 2004 113 2 123 136 14 5 2004 14 10 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Pyrethroid insecticides have been used for more than 40 years and account for 25% of the worldwide insecticide market. Although their acute neurotoxicity to adults has been well characterized, information regarding the potential developmental neurotoxicity of this class of compounds is limited. There is a large age dependence to the acute toxicity of pyrethroids in which neonatal rats are at least an order of magnitude more sensitive than adults to two pyrethroids. There is no information on age-dependent toxicity for most pyrethroids. In the present review we examine the scientific data related to potential for age-dependent and developmental neurotoxicity of pyrethroids. As a basis for understanding this neurotoxicity, we discuss the heterogeneity and ontogeny of voltage-sensitive sodium channels, a primary neuronal target of pyrethroids. We also summarize 22 studies of the developmental neurotoxicity of pyrethroids and review the strengths and limitations of these studies. These studies examined numerous end points, with changes in motor activity and muscarinic acetylcholine receptor density the most common. Many of the developmental neurotoxicity studies suffer from inadequate study design, problematic statistical analyses, use of formulated products, and/or inadequate controls. These factors confound interpretation of results. To better understand the potential for developmental exposure to pyrethroids to cause neurotoxicity, additional, well-designed and well-executed developmental neurotoxicity studies are needed. These studies should employ state-of-the-science methods to promote a greater understanding of the mode of action of pyrethroids in the developing nervous system.
age-dependent toxicitybiologically based dose–response modeldevelopmental neurotoxicitymode of actionphysiologically based pharmacokinetic modelpyrethroidrisk assessmentvoltage-sensitive sodium channel
==== Body
Pyrethroid insecticides have been used in agricultural and home formulations for more than 30 years and account for approximately one-fourth of the worldwide insecticide market (Casida and Quistad 1998). Currently, 16 pyrethroids are registered for use in the United States in a large variety of agricultural or consumer products (Bryant and Bite 2003). Often, pyrethroids are sold and/or used as mixtures containing a combination of two or more compounds (Farm Chemicals Handbook 1997). Exposure to pyrethroids has been widely documented in humans, including exposure of pregnant women, infants, and children (Berkowitz et al. 2003; Huedorf et al. 2004; Schettgen et al. 2002; Whyatt et al. 2002). Although the acute toxicity of these compounds to adults has been well characterized, the potential for developmental toxicity of pyrethroids is not well understood.
In the present review we focus on the potential for neurotoxicity after developmental exposure to pyrethroid insecticides. We also consider the current state and quality of scientific data that could be used to support risk decisions related to pyrethroid developmental and age-dependent neurotoxicity. Specifically, in this review we a) provide a brief overview of the toxicity of this class of compounds; b) review pyrethroid effects on voltage-sensitive sodium channels (VSSCs), a primary mode of action of pyrethroids; c) discuss the developmental profiles of VSSCs; d) provide examples of the results of perturbation of VSSCs during development by other insults; e) discuss the evidence for age-related sensitivity to this class of compounds; f ) summarize and critique studies of pyrethroid neurotoxicity after developmental exposure; and g) make recommendations regarding future research needs related to the developmental neurotoxicity of pyrethroids.
In addition to being important to scientists interested in characterizing the neurotoxicity of these compounds, this information will be useful when considering the scientific data needed to inform risk decisions related to pyrethroid insecticides. Under the Food Quality Protection Act (FQPA; 1996), the U.S. Environmental Protection Agency (EPA) is required to include a default 10× safety factor (uncertainty factor) in risk decisions to protect against potentially greater sensitivity of developing individuals to toxic insult. This factor can be adjusted only if compelling scientific data exist regarding age-related differences in sensitivity. Furthermore, developing individuals must be considered under FQPA requirements for cumulative risk assessments (classes of compounds with the same mode of action). The quality of the scientific data used to support these and other risk decisions is an important component of scientifically based risk assessment. In addition, information regarding mode of action improves the scientific basis for risk decisions (Brock et al. 2003; Mileson et al. 1998; Sonich-Mullin et al. 2001), including those related to developmental neurotoxicity (Costa 1998; Tilson 2000a, 2000b).
The U.S. EPA has recently released the revised cumulative risk assessment of the organophosphate class of insecticides (U.S. EPA 2002) and has requested that registrants of these insecticides submit developmental neurotoxicity studies to the agency. In the near future, the U.S. EPA must consider developmental and cumulative risk for other classes of insecticides, including pyrethroids. Thus, in this review we focus on issues of mode of action and age-dependent and developmental neurotoxicity as related to risk decisions under the FQPA.
Overview of Pyrethroid Toxicity
The pyrethroid class of insecticides was derived from natural compounds (the pyrethrins) isolated from the Chrysanthemum genus of plants (Casida 1980). Although natural pyrethrins do have insecticidal activity, they also are inherently unstable when exposed to light. Therefore, the pyrethrin structure was modified to produce more stable compounds that retained the desirable insecticidal and toxicologic properties (Valentine 1990). All pyrethroids contain several common features: an acid moiety, a central ester bond, and an alcohol moiety (Figure 1). The acid moiety contains two chiral carbons; thus, pyrethroids typically exist as stereoisomeric compounds. Furthermore, some compounds also contain a chiral carbon on the alcohol moiety, which allows for three chiral carbons and a total of eight different stereoenantiomers. Pyrethroid insecticidal activity (Elliot et al. 1974), acute mammalian neurotoxicity (Gray and Soderlund 1985), and effects on VSSCs (Lund and Narahashi 1982) are stereospecific, indicating the presence of specific binding sites. For some compounds, several commercial products are available that differ in stereoisomer content. For example, allethrin is a mixture of all possible allethrin stereoisomers, d-allethrin contains only the 1R isomers, bioallethrin contains only the 1R-trans isomers, and S-bioallethrin is enriched in the S stereoisomer of the 1R-trans isomers (Figure 2).
The acute mammalian neurotoxicity of pyrethroids has been well characterized, and several comprehensive reviews of pyrethroid toxicity, metabolism, and actions are available (Kaneko and Miyamoto 2001; Narahashi 2001; Ray 2001; Soderlund et al. 2002). Verschoyle and colleagues (Verschoyle and Aldridge 1980; Verschoyle and Barnes 1972) conducted structure–activity relationship studies with a series of pyrethroids and described two generalized syndromes after acute exposure. Based on toxic signs in the rat, pyrethroids have been divided into two types: a) compounds that produce a syndrome consisting of aggressive sparring, increased sensitivity to external stimuli, and fine tremor progressing to whole-body tremor and prostration (type I or T syndrome); and b) compounds that produce a syndrome consisting of pawing and burrowing, profuse salivation, and coarse tremor progressing to choreoathetosis and clonic seizures (type II or CS syndrome) (Verschoyle and Aldridge 1980). Analogous toxic signs have been observed in mice (Lawrence and Casida 1982; Staatz et al. 1982) and cockroaches (Gammon et al. 1981; Scott and Matsumura 1983). Structurally, a key difference between type I and type II pyrethroids is the absence or presence, respectively, of a cyano group at the α carbon of the 3-phenoxybenzyl alcohol moiety of the compound. Thus, the type I/II or T/CS nomenclatures are useful as general classification schemes and are widely used in the published literature. However, a few pyrethroids do not fit neatly into these schemes because they produce signs related to both syndromes (Verschoyle and Aldridge 1980; for review see Soderlund et al. 2002). Further, these schemes are based on doses of pyrethroids that cause overt neurotoxicity and thus may not apply to either low-dose or developmental exposures. Because it conveys useful structural information, in this review we use the type I/II classification system.
Effects of Pyrethroids on VSSCs
The primary mode of pyrethroid action in both insects and mammals is disruption of VSSC function. Perturbation of sodium channel function by pyrethroids is stereospecific (Lund and Narahashi 1982); those stereoisomers that are the most potent disruptors of VSSC function also have the most potent insecticidal or toxicologic activity (Ray 2001). Pyrethroids slow the activation, or opening, of VSSCs. In addition, they slow the rate of VSSC inactivation (or closing) and shift to more hyperpolarized potentials the membrane potential at which VSSCs activate (or open) (for review, see Narahashi 1996). The result is that sodium channels open at more hyperpolarized potentials (i.e., after smaller depolarizing changes in membrane potential) and are held open longer, allowing more sodium ions to cross and depolarize the neuronal membrane. In general, type II compounds delay the inactivation of VSSCs substantially longer than do type I compounds. Type I compounds prolong channel opening only long enough to cause repetitive firing of action potentials (repetitive discharge), whereas type II compounds hold open the channels for such long periods of time that the membrane potential ultimately becomes depolarized to the point at which generation of action potentials is not possible [depolarization-dependent block (Figure 3)]. These differences in prolongation of channel open times are hypothesized to contribute to the differences in the CS and T syndromes after exposure to type II and I pyrethroids, respectively (for review, see Ray 2001).
Mammalian VSSCs are composed of one α and two β subunits. Ten separate α subunits (Table 1; Ogata and Ohishi 2002) and four different β subunits (Isom 2002) have been identified and are expressed in a tissue-, region-, and time-specific manner. With one exception (the NaX subunit), α subunits all comprise VSSCs when expressed individually or with β subunits. The α subunit forms the pore of the channel and determines its major functional characteristics, whereas the β subunits are auxiliary proteins that influence gating properties, localization in the membrane, and interactions with cytoskeletal proteins (Isom 2001, 2002). The diverse functional roles of VSSC, such as generating action potential spikes, amplifying sub-threshold depolarizations, regulating repetitive firing and generating after-depolarizations, depend on the numerous potential combinations of α and β subunits (Ogata and Ohishi 2002). The types of VSSCs expressed in different regions, their relative sensitivity, and their functional role may all contribute to the manifestation of pyrethroid effects.
VSSC Heterogeneity and Pyrethroid Effects
All available evidence indicates that pyrethroids bind to the α subunit of the VSSC. Trainer et al. (1997) expressed only the Nav1.2 α subunit in Chinese hamster ovary cells and found that the presence of the α subunit is sufficient for pyrethroids to produce their characteristic effects on sodium channel function in mammalian cells. This conclusion is supported by additional research demonstrating that pyrethroids alter currents produced by expression of Nav1.2 (Smith and Soderlund 1998) or Nav1.8 (Smith and Soderlund 2001) in oocytes in the absence of coexpression with β subunits. Interestingly, coexpression of the β1 subunit with Nav1.2 increased the sensitivity of this channel compared with expression of Nav1.2 alone (Smith and Soderlund 1998), indicating that the β subunit modulates the affinity of pyrethroid interaction with the channel. Mutations in the α subunit of both insects (Lee and Soderlund 2001; Smith et al. 1997) and mammals (Vais et al. 2000, 2001; Wang et al. 2001) alter the sensitivity of VSSCs to pyrethroid effects, supporting the conclusion that pyrethroids interact with the α subunit.
The relative susceptibility of the 10 different VSSC α subunits to pyrethroids is not well understood. Differential sensitivity of VSSCs to pyrethroids was first reported by Tatebayashi and Narahashi (1994). In a comparison of tetramethrin effects on tetrodotoxin-sensitive (TTX-S) versus -resistant (TTX-R) sodium channels in dorsal root ganglion neurons, TTX-R channels were demonstrated to be more sensitive to perturbation by tetramethrin (Tatebayashi and Narahashi 1994). However, TTX-R or TTX-S channels may arise from several different VSSC α subunits (Table 1). Although not all α subunits have been examined, differences in sensitivity to pyrethroids were reported after expression of different subunits in vitro (details provided in Table 1). For example, Nav1.2 (Smith and Soderlund 1998) is sensitive to type II but not type I compounds, whereas Nav1.8 (Smith and Soderlund 2001) is sensitive to both. Interactions of pyrethroids with other sodium channel α subunits have not been investigated to date. Importantly, the pyrethroid sensitivity of VSSC subunits and splice variants expressed during development has yet to be examined.
Developmental Expression of VSSC
VSSCs show complex regional and temporal ontogeny, which is briefly summarized in Table 1. In general, embryonically expressed forms of VSSCs are replaced by expression of adult forms as neurodevelopment proceeds. For example, high expression of Nav1.3 during embryonic periods (Albrieux et al. 2004) diminishes as expression of Nav1.2 increases in early postnatal periods in rodents (Felts et al. 1997), and expression of Nav1.2 at immature nodes of Ranvier is replaced by Nav1.6 as myelination proceeds (Boiko et al. 2001; Jenkins and Bennett 2002). Similar changes are observed with the β subunits, because β3 expression is replaced by β1 and β2 (Shah et al. 2001). Alternatively spliced forms of the VSSC subunits also contribute to developmental differences in expression because the Nav1.2, Nav1.3, and Nav1.6 subunits all have splice variants that are expressed in rodents from embryonic through early postnatal ages (Gustafson et al. 1993; Plummer et al. 1997; Sarao et al. 1991). Given the previously reported differences in α subunit sensitivity to pyrethroids, the complex ontogeny of VSSC expression could result in altered sensitivity (either increases or decreases) of the developing nervous system to perturbation by various pyrethroids. In addition, understanding the timing and localization of expression of the most pyrethroid-sensitive VSSCs during neurodevelopment could help in understanding and explaining effects reported after developmental exposure. With respect to age-dependent toxicity of pyrethroids, research to date indicates that toxicokinetic and not toxicodynamic factors account for differences in susceptibility between young and adult animals (Cantalamessa 1993; Sheets et al. 1994); however, toxicodynamic factors have not been systematically examined.
Disruption of VSSC Function and Expression during Development
Evidence from mutation and knockout models demonstrates that perturbation of VSSC function during development impairs nervous system structure and function. Several examples are discussed below for illustrative purposes. These examples demonstrate the plausibility that perturbations in VSSC function by pyrethroids during development could result in adverse consequences in the developing nervous system.
Knockout and mutant mouse models of sodium channel α subunits demonstrate varying degrees of adverse outcomes associated with loss or alteration of specific channel subunits. When mRNA for the Nav1.2 subunit was reduced by approximately 85%, mice exhibited reduced levels of electrical excitability, had high levels of apoptotic neurons in the brainstem and cortex, and died from severe hypoxia within 1–2 days of birth (Planells-Cases et al. 2000). In contrast, mutation of the gene encoding the Nav1.6 subunit resulted in development of hindlimb paralysis, skeletal muscle atrophy by postnatal day (PND)10, and death by PND20 (Porter et al. 1996). Atrophy was specific to muscle innervated by spinal and not oculomotor neurons (Porter et al. 1996). Finally, Nav1.8 knockout mice survived to adulthood and exhibited normal behavior, although sensation of some types of noxious stimuli was lost or diminished (Akopian et al. 1999; Laird et al. 2002).
In humans, perturbation of nervous system development has been associated with altered VSSC structure or function. Recent advances in molecular genetics have identified in genes coding for VSSC subunits a number of mutations that result in neuronal hyperexcitability due to subtle changes in channel gating and inactivation (see Meisler et al. 2001, their Table 3). These mutations have been linked to various forms of epilepsy in humans, providing evidence that changes in VSSC function can give rise to clinically definable disease (Claes et al. 2001; Escayg et al. 2001; Meisler et al. 2002; Noebels 2002; Wallace et al. 2001). Mouse models expressing these mutant ion channels have been constructed, facilitating the study of these diseases (Kearney et al. 2001; Meisler et al. 2001). It is noteworthy that pyrethroids, like these mutations, alter VSSC activation, inactivation, and neuronal excitability. The mechanisms and magnitude of mutational versus pyrethroid effects are different, as would be the duration of effect (dependent on exposure for pyrethroids vs. permanent for mutations). Because of these differences, results from mutation and knockout models may not be predictive of developmental exposure to pyrethroids. Notably, potential interactions between pyrethroids and these mutations to VSSCs have not yet been examined.
Phenytoin, an anticonvulsant that blocks VSSCs as well as other ion channels (Catterall 1999), has been demonstrated to disrupt nervous system structure and function after developmental exposure (Adams et al. 1990). In humans, the use of anticonvulsants during pregnancy has been associated with a number of defects and malformations, which collectively are referred to as fetal hydantoin syndrome, and include microcephaly and intellectual impairment. Studies in animal models support the human findings (Hatta et al. 1999; Ohmori et al. 1997, 1999; Schilling et al. 1999; Vorhees et al. 1995). Thus, developmental exposure to this drug, which acts on VSSCs, can produce significant alterations in nervous system structure and function. It should also be noted that, although phenytoin is used as an example, there are currently no data to suggest that developmental exposure to pyrethroids results in similar effects.
Age-Related Differences in Sensitivity to Pyrethroids
The magnitude of the age-related toxicity of pyrethroids appears to be much larger than for many other pesticide classes, but the number of studies is small. Whether this age-related neurotoxicity includes both type I and type II compounds is currently unclear. In neonatal versus adult rats, the acute lethality of the type II pyrethroid deltamethrin was 16-fold greater in young animals (Sheets et al. 1994). Concentration data indicate that the age dependency was due to lower metabolic capabilities in the young rats (Sheets et al. 1994). Similarly, the type II pyrethroid cypermethrin was 17-fold, and the type I pyrethroid permethrin was 6-fold more lethal in PND8 rats compared with adults; metabolic inhibitors were used to demonstrate that toxicokinetic factors were responsible for this age-dependent susceptibility (Cantalamessa 1993). In contrast, evidence has been presented that two type I pyrethroids, cismethin and permethrin, did not have any age-dependent toxicity (Sheets 2000).
Age-related sensitivity to pyrethroids may be influenced by dose. In a symposium report, Sheets (2000) argued that the age-dependent sensitivity of pyrethroids is apparent only at high acute doses. This report contained data suggesting a lack of age-dependent differences in the behavioral toxicity of type I and type II pyrethroids at doses below those causing overt toxicity. However, age-dependent differences in pyrethroid neurotoxicity have not been thoroughly studied at the lower end of the dose–response relationship (sublethal doses). The scientific basis for decisions related to the FQPA could be strengthened by additional studies comparing the relative susceptibility of differential sensitivity between young and adult animals, particularly at sublethal doses. For example, replication of Sheets’s (2000) report and expansion to include additional compounds would provide useful information regarding sensitivity differences between developing and adult animals.
Pyrethroid Developmental Neurotoxicity Studies
A total of 22 studies were evaluated for this review (Tables 2–4), including 19 peer-reviewed publications (Table 2), unpublished studies (Muhammad and Ray, unpublished data; see Table 3), and regulatory studies provided by Bayer AG (Table 4; Ivens et al., unpublished data; Jekat et al., unpublished data). The studies conducted by Muhammad and Ray (unpublished data) consisted of several similarly treated “cohorts” for both S-bioallethrin and deltamethrin. Rather than present the overall findings for each of these two compounds, the results of individual “cohorts” are summarized in Table 3 to provide more detailed information. Tables 2–4 contain a summary of important information from each study, including test compound/formulation, animal species, dosing period, and major findings. Because the vehicle used and route of exposure can have profound influence on the expression of pyrethroid neurotoxicity in adult rats (Crofton et al. 1995), this information is included as well.
Allethrin (in the form of allethrin, d-allethrin, bioallethrin, and S-bioallethrin) and permethrin are the only type I pyrethroids for which peer-reviewed studies of potential developmental neurotoxicity have been conducted. Of the type II compounds, results of developmental studies have been published for deltamethrin, cypermethrin, fenvalerate, and cyhalothrin, and data regarding the developmental neurotoxicity of cyfluthrin (Jekat et al., unpublished data) have been submitted to the U.S. EPA. Thus, no developmental neurotoxicity studies exist for many pyrethroids.
Rodents were the sole animal models used in these studies: 13 studies used rats and 9 studies used mice. No studies were conducted specifically to examine species differences, nor could any clear species-dependent effects be discerned. The choice of rats or mice seemed to be based on a) previous use of that species in the laboratory or b) whether or not the study was designed to replicate (in whole or part) results published previously by other investigators. A systematic comparison of factors that underlie potential species differences in neurotoxic responses could provide useful information regarding the extrapolation of data from animals to humans. For example, Nav1.3 expression in rodents appears to be primarily embryonic, yet in humans considerable expression in adults has been reported (Whitaker et al. 2000, 2001). How this and other species differences influence neurotoxic responses has not been investigated.
Several studies reported persistent changes in behavior and/or neurochemistry in animals examined long after exposure had stopped. Eriksson’s group (Ahlbom et al. 1994; Eriksson and Fredriksson 1991; Eriksson et al. 1993; Eriksson and Nordberg 1990) has reported that mice exposed to pyrethroids during PND10–16 exhibit increased motor activity and lack of habituation, as well as changes in density of muscarinic acetylcholine receptor (mAChR) binding for as long as 5 months (Talts et al. 1998) after cessation of exposure. Given the short half-lives for pyrethroids (Anadón et al. 1991, 1996; for review, see Kaneko and Miyamoto 2001), these effects are likely due to exposure during development and not residual tissue concentrations of pyrethroids. Studies conducted by Eriksson and co-workers used bioallethrin and deltamethrin, which contain only two and predominantly one stereoisomer, respectively. Thus, effects can be ascribed to the compound that has insecticidal activity (vs. studies conducted with formulated products). In addition, dose–response relationships have been demonstrated for bioallethrin (Ahlbom et al. 1994), and the replication of effects, both behavioral and biochemical, within this laboratory has been consistent over several studies. Others have also reported persistent changes in behavior and/or biochemistry, including learning (Moniz et al. 1990), motor activity (deltamethrin only; Husain et al. 1992), sexual behavior (Lazarini et al. 2001), mAChR binding (Aziz et al. 2001; Malaviya et al. 1993), and blood–brain barrier permeability (Gupta et al. 1999a).
There were several studies that examined both motor activity and mAChR expression after developmental exposure to pyrethroids. A summary of effects on these end points, independent of dose, exposure period, and other parameters, is provided in Table 5. In all of these studies, quinuclidinyl benzilate (QNB) binding was used to measure mAChR expression. QNB is a nonspecific antagonist for this receptor (Watling et al. 1995) and does not discriminate between mAChR subtypes (M1–M5). Measurement of QNB binding may in fact be one of the more comparable end points across these numerous studies. In addition, many but not all of these studies examined mAChR expression at PND17 and/or 4 months of age.
Comparison of pyrethroid effects on QNB binding across studies does not reveal clear trends in reported effects between laboratories. In preweanling animals, across all compounds and treatment protocols, QNB binding was reported to increase in six studies, decrease in two studies, and not change in four studies (Table 5). In cortical tissue, the data for PND17 are more consistent in that five of eight studies reported increases in mAChR expression. If only the various forms of allethrin are considered, four studies reported increases and two reported no change in QNB binding when measured on PND17. Persistent alterations in mAChR in adulthood after developmental exposure are less clear, with three studies reporting increases, three reporting decreases, and five reporting no change in QNB binding. Considering only allethrin forms again, QNB binding increased or decreased in two studies each and was unchanged in three studies.
Differences in a number of important variables may underlie some of the inconsistencies in QNB binding data. One difference is exposure route. Two studies used inhalation exposure (Ivens et al., unpublished data; Jekat et al., unpublished data), whereas exposure in the remainder of the studies was via oral gavage (Table 5). A comparison of effects in Tables 2–5 suggests that this is not a tenable explanation for these inconsistencies because results do not correlate to route. Another variable that differed between laboratories was the formulation of allethrin used. Allethrin, like all pyrethroids, exists as several different stereoisomers (Figure 2), and the insecticidal and toxic effects of pyrethroids are highly stereospecific. These studies employed allethrin formulations with differing contents of allethrin stereoisomers; two groups used d-allethrin (Ivens et al., unpublished data; Tsuji et al. 2002), one used bioallethrin (Eriksson group: Ahlbom et al. 1994; Ericksson and Fredriksson 1991; Eriksson and Nordberg 1990; Talts et al. 1998), and two used S-bioallethrin (Muhammad and Ray, unpublished data; Pauluhn and Schmuck 2003). Again, data in Table 5 suggest that this is not a tenable explanation because d-allethrin and bioallethrin result in either increases or no effects on mAChR binding. An additional variable in these data sets is the specific methods used in the competitive binding experiments. Competition experiments with carbachol were used in several studies to distinguish between high- and low-affinity QNB binding sites (Ahlbom et al. 1994; Eriksson and Fredriksson 1991; Eriksson and Nordberg 1990; Ivens et al., unpublished data; Jekat et al., unpublished data; Talts et al. 1998). Two studies (Ahlbom et al. 1994; Eriksson and Nordberg 1990) reported that bioallethrin increased the percentage of low-affinity binding sites in PND17 mice, an effect not reported in adult mice, despite changes in the density of muscarinic binding (Eriksson and Fredriksson 1991; Talts et al. 1998). Ivens et al. (unpublished data) did not find changes in the percentages of high- and low-affinity sites, even though they did report changes in the density of QNB binding sites in PND17 animals. In some cases, the relative proportion of high- and low-affinity sites was not investigated even though changes in density were reported (Muhammad and Ray, unpublished data). The ability to distinguish high- and low-affinity sites, and effects thereon, is dependent on the number of points included on the agonist competition curve. Studies conducted by the group at Bayer (Ivens et al., unpublished data; Jekat et al., unpublished data) used seven different concentrations of carbachol, whereas studies conducted by Eriksson’s group (Ahlbom et al. 1994; Eriksson and Fredriksson 1991; Eriksson and Nordberg 1990) used 18 concentrations of carbachol (Eriksson P, personal communication). This information was typically not available to evaluate and may account for some reported differences, because use of too few points may preclude detection of changes in the low-affinity site. Overall, the data across laboratories indicate that changes in QNB binding may not be a robust response to developmental exposure to pyrethroids and that conditions may need to be more carefully controlled in order to observe changes.
A smaller number of studies examined potential alterations in catecholaminergic systems. Both deltamethrin (Lazarini et al. 2001) and bioallethrin (Muhammad and Ray, unpublished data) were reported to increase 3,4-dihydroxyphenylacetic acid (DOPAC) levels in the adult striatum after developmental exposure. However, developmental exposure to a commercial product containing fenvalerate had no effect on monoamine levels in the striatum (Moniz et al. 1999). Malaviya et al. (1993) reported that binding of 3H-spiroperidol to striatal membranes from PND21 rats was decreased and increased, respectively, after gestational and lactational exposure to a commercial product containing fenvalerate, whereas binding was increased after only lactational exposure to a commercial product containing cypermethrin. Thus, similar to the muscarinic cholinergic system, the dopaminergic system may be affected by developmental exposure to pyrethroids, but studies examining this system have reported inconsistent results to date.
Eriksson and co-workers have consistently reported increased motor activity and a lack of habituation after exposure to pyrethroids (Ahlbom et al. 1994; Eriksson et al. 1993; Talts et al. 1998). A comparison of effects of pyrethroids on motor function between laboratories is not as consistent. Muhammad and Ray (unpublished data) observed effects on motor activity in some cohorts but not in others. After inhalation exposure to bioallethrin (Tsuji et al. 2002) or d-allethrin (Ivens et al., unpublished data), no effects on activity or habituation were reported. By contrast, inhalation exposure to cyfluthrin resulted in hyperactivity and decreased habituation in female mice (Jekat et al., unpublished data). Several additional studies also examined other measures of open field or motor activity (Gomes et al. 1991a; Husain et al. 1992, 1994; Lazarini et al. 2001). Reports of effects in these studies were also variable (Table 2). The reasons for the discrepant nature of these findings are unknown.
A small number of studies tested cognitive functions (Table 2). Two studies reported that bioallethrin exposure during PND10–16 (via different routes) had no significant effect on performance in the Morris water maze at 5 (Talts et al. 1998) and 11 (Tsuji et al. 2002) months of age. Other studies reported decreases in avoidance and Y-maze learning (Aziz et al. 2001; Husain et al. 1994; Moniz et al. 1990) or no change in avoidance behavior (Gomes et al. 1991a). A major confounder in the Y-maze and avoidance studies is the use of commercial formulations rather than technical compound.
There are several common weaknesses in the developmental studies that temper the scientific strength of some individual reports, as well as the data set when taken as a whole. A key weakness is problematic statistical analyses. Most behavioral studies [with the exception of Ivens et al. (unpublished data), Jekat et al. (unpublished data), and Tsuji et al. (2002)] used multiple pups from the same litter without correction in the statistical analysis. The sampling of multiple pups from the same litter inflates the sample size and increases the probability of a type I statistical error (Abbey and Howard 1973; Holson and Pearce 1992; Muller et al. 1985; Reily and Meyer 1984). When biochemical end points were examined, statistical analyses often lacked robustness or, in some cases, were absent. In several studies examining receptor binding, results were compared (and significant differences found) using multiple Student’s t-tests. Use of multiple t-tests can easily increase the probability of a type I error (Muller et al. 1985). These study designs should use statistical models that control for multiple comparisons (e.g., analysis of variance with appropriate post hoc test for comparisons of different group means). Meta-analyses or other statistical approaches to examine related data sets from the same and different laboratories could help strengthen conclusions when effect magnitude is small but have not been conducted to date.
An additional limitation common to these reports was a lack of tissue concentration data. None of the studies reported pyrethroid blood or brain concentrations from dams or pups. Such information would have greatly facilitated comparisons between studies and would also be useful to compare target tissue concentrations in the test species with exposure estimates in pregnant women (see Whyatt et al. 2002).
Lack of information about the stereoisomer composition and/or purity of the test compound was a serious confound in some reports. Such information is important to be able to compare studies generated in different laboratories, as discussed above for the different allethrin products. In addition, several studies used formulated products rather than purified compound (Aziz et al. 2001; Gupta et al. 1999a, 1999b; Husain et al. 1992, 1994; Malaviya et al. 1993). Formulated pesticide products typically contain solvents, emulsifying agents, petroleum distillates, and other “inerts” (Farm Chemicals Handbook 1997), many of which are known or suspected to have neurotoxic properties. Although use of formulated products may provide a more real-life exposure situation, lack of information on the content of proprietary formulations hampers comparisons between studies and often precludes attributing effects directly to the pyrethroid.
Several other limitations should also be noted. The number of time points examined in these studies typically was three or fewer, one of which was often a measurement in adult animals. Considerable ontogeny of both behavioral responses as well as biochemical end points is well established. Thus, the tendency of most studies to examine a “snapshot in time” may miss important ontogenic shifts induced by these compounds. Dosing duration and age at exposure are two other important factors. Although a number of studies examined the period of PND10–15, the choice of dosing periods in the present studies was variable, and, to date, there has not been a systematic evaluation of potentially sensitive developmental periods. An additional consideration regarding dosing periods is the differential rates of neurodevelopment in rodents versus humans. Thus, studies such as those conducted by Whyatt et al. (2002) could potentially provide important information about exposure to the developing fetus. In addition, the effects of sex were not always considered in the present studies, with a few exceptions (e.g., Gomes et al. 1991b; Moniz et al. 1999). Also related to this topic is the relative distribution of males and females in a litter. In some cases, culling information was readily available; however, many studies provided no or insufficient information to evaluate this variable.
Although not necessarily a limitation, there is a significant conceptual gap between the variety of behavioral, biochemical, and physiologic end points studied to date (Tables 1–4). The relationships, if any, between these biochemical and behavioral changes have yet to be established. In addition, the relationship between the end points examined in the present studies and the major action of pyrethroids, disruption of VSSC function, is also unknown. Only one study to date has examined changes in VSSC expression (Muhammad and Ray, unpublished data). The relationship between biochemical alterations and pyrethroid-induced developmental neurotoxicity could be strengthened by better characterization of neurochemical mode(s) of action of pyrethroid neurotoxicity. Establishing mode-of-action pathways increases confidence that reported effects are the result of pyrethroid action, particularly when the magnitude of those effects is small.
Conclusions and Recommendations for Future Research
Several research needs in the area of developmental neurotoxicity are apparent from this review. These include additional information regarding potential differences underlying age-dependent sensitivity to pyrethroids, clarification of changes in behavioral and biochemical end points, and linking these end points to VSSCs or other cellular targets. In considering these potential areas for future research, determining the priority of addressing different research questions often depends on individual perspectives. In this context, a different conceptual approach to conducting future research may improve the resulting data’s usefulness for the purpose of risk decisions.
Biologically based dose–response (BBDR) models (Andersen and Dennison 2001) describe the relationships between different components of the continuum between exposure to and the adverse effects of a chemical (Figure 4). For example, such a model has recently been constructed for the developmental neurotoxicity of perchlorate (Jarabek et al. 2002). Mode-of-action models strengthen science in two important ways. First, the uncertainty regarding animal-to-human extrapolations can be reduced if a toxicant’s mode of action in an animal model is demonstrated to be relevant to humans (Cohen et al. 2004; Meek et al. 2003; Sonich-Mullin et al. 2001). Second, these models often provide insight into research needs by identifying data gaps and research needs. For pyrethroids, much of the future research needs can be described in the context of the type of data that would be useful in constructing a BBDR for this class of compounds, or for individual compounds within this class. A cornerstone of a BBDR model is a physiologically based pharmacokinetic (PBPK) model that describes the relationship between exposure and target tissue dose (Andersen and Dennison 2001). Additional pharmacokinetic information in animal models as well as additional pharmacokinetic and exposure information in humans is needed. For pyrethroids, this will involve defining the relationship between maternal and fetal compartments, and the involvement of oral (including lactation), inhalation, and dermal exposures to the newborn. Current data indicate that some exposure does occur to pregnant mothers, infants, and children, resulting in low internal doses (Berkowitz et al. 2003; Heudorf et al. 2004; Schettgen et al. 2002). However, insufficient information is available to adequately evaluate the range of internal doses of pyrethroids in humans. These data will be valuable in quantitative extrapolations of exposure from animals to humans (Andersen and Dennison 2001). Pharmacokinetic information is available comparing acute high-dose exposures in neonatal versus adult animals (Cantalamessa 1993; Sheets et al. 1994). However, only a limited number of compounds have been examined to date, and no information is available for ages before PND11.
Another component of a BBDR model is a physiologically based pharmacodynamic (PBPD) model (Andersen et al. 1992; Conolly 2002). PBPD models are quantitative models that describe the mode of action of a chemical. A benefit of PBPD models is identification of research gaps that are critical to link key events in the mode of action to adverse outcomes. Currently available studies of pyrethroid developmental neurotoxicity have examined a wide variety of end points but have not sought to link target tissue events (e.g., receptor activation, changes in ion channel function) to consequent biochemical, physiologic, or behavioral outcomes. Future studies need to target the large data gap between the target site (e.g., VSSCs) and adverse outcomes. For example, can the sequence of biochemical processes be described that, when perturbed by pyrethroids, result in changes in end points such as motor activity or mAChR binding? If changes in sodium currents alter neuronal firing rate, how does this then lead to alterations in neurodevelopment? Considerable information supports involvement of VSSCs in the mode of action of acute pyrethroid neurotoxicity, yet the potential role of VSSCs in developmental neurotoxicity of pyrethroids has not been examined. Future research on the developmental neurotoxicity of pyrethroids should endeavor to fill these research gaps. These studies must be designed and conducted so as to avoid the limitations mentioned in the preceding section. Such studies of the developmental neurotoxicity of these compounds can strengthen the scientific basis for risk decisions. The most efficient use of scientific resources will be to design those additional studies to fit into a BBDR scheme.
Figure 1 Structures of pyrethroids for which developmental neurotoxicity has been examined. Developmental neurotoxicity studies have been conducted using either technical compound or formulations of the seven pyrethroids illustrated; the numbers in parentheses after each compound name indicate the number of studies that have been conducted using that compound or a formulation containing that compound. Only one stereoisomer is illustrated for each compound.
Figure 2 Eight possible stereoisomers of allethrin (A–H). The inset lists allethrin-containing products and the stereoisomer content of each.
Figure 3 Pyrethroid effects on neuronal excitability. This schematic depicts pyrethroid effects on individual channels, whole-cell sodium currents, and action potentials. Depolarization opens VSSCs (top left) allowing sodium to enter the cell. To limit sodium entry and depolarization length, VSSCs inactivate and must return to a “resting” state before reopening. Pyrethroids inhibit the function of two different “gates” that control sodium flux through VSSCs (top right), delaying inactivation (indicated by double arrows between states) of the channel and allowing continued sodium flux (Open*). If sodium current through an entire cell is measured, depolarization leads to a rapidly inactivating current under normal circumstances (bottom left, Sodium current). Pyrethroid-modified VSSCs remain open when depolarization ends (bottom right, Sodium current), resulting in a “tail” current (the notch at the end of example currents). If membrane voltage is examined, depolarization under normal circumstances generates a single action potential (bottom left). VSSCs modified by type I compounds (bottom right, Action potential) depolarize the cell membrane above the threshold for action potential generation, resulting in a series of action potentials (repetitive firing). Type II compounds cause greater membrane depolarization, diminishing the sodium electrochemical gradient and subsequent action potential amplitude. Eventually, membrane potential becomes depolarized above the threshold for action potential generation (depolarization-dependent block).
Figure 4 Major elements in a proposed BBDR model for pyrethroid neurotoxicity and research needs for the PBPK and PBPD components. Boxes with question marks indicate that the sequence of events between changes in the target and adverse effects has not been completely elucidated.
Table 1 Sodium channel α subunit nomenclature and effects of pyrethroids.a
α subunit Older names TTX sensitivity Tissue expression Developmental expression Effect of pyrethroids
Nav1.1 Rat I, HBSCI, GPBI, SCN1A TTX-S CNS, PNS, Purkinje, HP pyramidal cells, spinal motor neurons, somatic localization Not detected in HP during development, detectable in CB Purkinje cells at PND15, detected at PND2 in SC; strong expression in motor neuronsb Not tested to date
Nav1.2 Rat II, HBSCII, HBA TTX-S CNS, forebrain, substantia nigra, HP mossy fibers, CB molecular layer, axonal localization In HP, increase between GD17 and PND30; in CB granule cells on PND15 and Purkinje cells on PND2; detected at all ages in SCb Splice variant expressed during developmentc Cypermethrin-induced tail currents detectable at > 30 nM in rat 1.2 (adult splice variant) co-expressed with β1 subunits; reported insensitive to permethrin or cismethrind
Nav1.3 Rat III TTX-S CNS and DRG HP expression at GD17, increasing at PND2, then decreasing to barely detectable at PND30. Detected at GD17 in CB neuroepithelium, decreasing thereafter, similar in SCb; developmentally regulated splice variante Not tested to date
Nav1.4 SkM1, μ1 TTX-S Skeletal muscle Increases with agef Only slightly modified by 10 μM deltamethrin when expressed in HEK 293t cellsg
Nav1.5 SkM2, H1 TTX-R Uninnervated skeletal muscle, heart, brain mRNA expressed in rat PND0 limbic structures and medulla; expressed in fetal and adult human brainh Not tested to date
Nav1.6 NaCh6, PN4, Scn8a, CerIII TTX-S CNS, DRG (all diameter neurons), node of Ranvier–peripheral nerve Truncated form expressed from GD12 to PND7, full-length mRNA expression is slight at GD14 and increases with agei Not tested to date
Nav1.7 NaS, hNE-NA, PN1 TTX-S DRG (all diameter neurons) CNS, Schwann cells All DRG neurons at PND2, increased during developmentb Not tested to date
Nav1.8 SNS, PN3, NaNG TTX-R DRG (small diameter neurons) Expression beginning at GD15 with adult levels by PND7; largely in unmyelinated C-fibersj Sensitive to both cismethrin and cypermethrin at thresholds of 500 nM and 30 nM, respectively k
Nav1.9 NaN, SNS2, PN5, NaT, SCN12A TTX-R DRG (small diameter neurons) Expression beginning at GD17 with adult levels by PND7; largely in unmyelinated C-fibers j Not tested to date
Nax Nav2.1, Nav2.3 Na-G, SCL11 ? Heart, uterus, skeletal muscle, astrocytes, DRG Transient between PND2 and 15 in HP; peak expression at PND2 in CB, SC; large DRG neurons, GD17 to PND30b Not tested to date
Abbreviations: CB, cerebellum; CNS, central nervous system; DRG, dorsal root ganglion; GD, gestation day; HP, hippocampus; PND, postnatal day; PNS, peripheral nervous system; SC, spinal cord; TTX, tetrodotoxin; TTX-R, TTX resistant; TTX-S, sensitive to TTX.
a Data in the first four columns are based on information presented by Goldin et al. (2000) and Novakovic et al. (2001).
b Felts et al. (1997).
c Sarao et al. (1991).
d Smith and Soderlund (1998).
e Gustafson et al. (1993).
f Kallen et al. (1990).
g Wang et al. (2001).
h Donahue et al. (2000).
i Plummer et al. (1997).
j Benn et al. (2001).
k Smith and Soderlund (2001).
Table 3 Summary of developmental neurotoxicity studies with pyrethroid compounds in NMRI mice dosed once daily on PND10–16 (Muhammad and Ray, unpublished data).
Compound Dose/route/vehicle Effects Comments
d-Allethrin, 93% purity (cis/trans) Experiment 13 0.7 mg/kg egg lecithin/peanut oil (1:10) 40% fat emulsion 4 months: 0 effect on motor activity; 0 effect on mAChR (QNB) binding in CTX Strengths: each chemical was examined in several cohorts in this study; closely replicates methodology of Eriksson and co-workers (see Table 2) for motor activity measurements; examined vehicle differences; technical compounds of known purity (100% for deltamethrin and 95.2% for S-bioallethrin)
S-Bioallethrin (trans) Experiment 17a 0.7, 3.5 mg/kg egg lecithin/peanut oil (1:10) 40% fat emulsion 4 months: ↑motor activity, habituation (slow mobile counts), 0 effect on mAChR in CTX Limitations: not published, peer-reviewed or submitted to any regulatory agency; litter was not used as statistical unit; statistical models not well described; t-tests used for biochemical measures; date of study unknown, circa mid-1990s
S-Bioallethrin (trans) Experiment 19a Attempt to replicate experiment 17a 4 months: ↑mAChR in CTX, CB (3.5 mg/kg); ↑mAChR brainstem (0.7 and 3.5 mg/kg); ↓habituation (slow mobile counts) by 0.7 mg/kg dose; ↑DOPAC, HVA in striatum; ↑saxitoxin binding in CB and MB, ↓in CTX
S-Bioallethrin (trans) Experiment 25a 0.7 mg/kg, corn oil PND17: 0 effect on mAChR in CTX 4 months: no data provided, despite mention that motor activity and mAChR were assessed
S-Bioallethrin (trans) Experiment 26a 0.7 mg/kg, corn oil 4 months: significant delay in habituation of slow rearing, fast rearing, total rearing, and rearing time; 0 effect on mobile activity and time, 0 effect on mAChR
Deltamethrin Experiment 12 0.7 mg/kg, egg lecithin/peanut oil (1:10) 40% fat emulsion 4 months: ↑ rearing time fast and total mobile counts slow, fast, and total rearing; delayed habituation of counts, slow mobile counts, and mobile time mAChR not examined
Deltamethrin Experiment 23 0.7 mg/kg, corn oil 4 months: ↑mAChR in CTX; no effect on any measure of motor activity
Deltamethrin Experiment 25 0.7 mg/kg, corn oil PND17: ↑mAChR; motor activity not examined
Deltamethrin Experiment 26 0.7 mg/kg, corn oil 4 months: significant delay in habituation of slow mobile counts, mobile and rearing time; 0 change in mAChR (increased but not significant)
Abbreviations: CB, cerebellum; CTX, cortex; HVA, homovanillic acid; MB, midbrain.
Table 2 Summary of peer-reviewed developmental neurotoxicity studies with pyrethroids.a
Species/compound Dose/route/vehicle Dosing period Effects Reference Comments
Rat (Wistar)
Cyhalothrin (type II) 0.02% in drinking water; 0.4% sucrose + “cyhalothrin vehicle” PND0–21 ↓learning avoidance latencies at PND90, 0 effect on motor activity in pup Moniz et al. 1990 Strengths: maternal behavior examined in Moniz et al., 1990 (no effect); culling described but not even across studies (culled to 5, 6, and 8 pups/dam) Limitations: commercial product, unknown vehicle (“cyhalothrin vehicle”) composition; dosing time frame not clear, but thought to be GD0–PND0 (Gomes et al. 1991a, 1991b); inappropriate statistical models; minimal description of results; not clear that litter is statistical unit (numbers of replicates in figure legends do not always agree with number of treatment groups)
0.018%; 1 mL dermal, daily; “cyhalothrin vehicle” “Entire pregnancy” Delayed development of fur, ear/eye opening, and testes descent. PND90: ↓hole-board head dips; 0 effect avoidance; and locomotion in open field Gomes et al. 1991a
0 change in sexual behaviors in males or females Gomes et al. 1991b
Fenvalerate (type II) 10 mg/kg, i.p.; saline GD18 and PND2–5 0 effect: testis descent, weight, monoamine levels, stereotyped behavior, locomotion, rearing ↓pup weight on PND21, ↓ductus deferens and seminal vesicle weight; female sexual behavior at PND120 Moniz et al. 1999 Strengths: litter as statistical unit; more complete and appropriate statistical analysis, but still some incorrect uses of t-test (Moniz et al. 1999); maternal weight examined/reported; Lazarini et al. (2001) considered sex differences; only papers examining reproductive behavior; culling, male/female ratios described and even. Housing as adults described
Deltamethrin (type II) 0.08 mg/kg, p.o. “deltamethrin vehicle” GD6–15, once daily PND21: ↑rearing in males; 0 effect on locomotion frequency in males or females Lazarini et al. 2001
PND60 males: ↓immobility time in forced swim test ↑DOPAC, DOPAC/DA, NA 0 effect on 5HT, 5HIAA, HVA/DA; 0 effect in PND60 females Limitations: deltamethrin commercial product; unknown (“deltamethrin vehicle”) vehicle composition; purity of fenvalerate not known; discrepancies between text and figures in Moniz et al. (1999 their Figure 3); differences in control testes descent day in Moniz et al. (1999) vs. Gomes et al. (1991, 1991b) (19 vs. 23 days)
Mouse (NMRI)
Bioallethrin (type I) 0.72 and 72 mg/kg 20% fat emulsion (egg lecithin/peanut oil) PND10–16, once daily PND17: ↑mAChR density and altered ratio of high- and low- affinity QNB binding sites in CTX but not HP with deltamethrin and bioallethrin at low (0.7 mg/kg) but not high doses Eriksson and Nordberg 1990 Strengths: consistent demonstration of increased motor activity and lack of habituation with bioallethrin and deltamethrin; dosing occurs over a critical period of brain development; dose response demonstrated for bioallethrin for behavior and biochemistry effects present 3.5–4 months postdosing; behavior, biochemistry measured in same animals; changes in mAChR binding in CTX ~10% at 4 months, but changes not observed after 5 months (bioallethrin); consistent effects over several different studies; history of publications with motor activity and QNB binding
Deltamethrin (type II) 0.71 and 1.2 mg/kg 20% fat emulsion (egg lecithin/peanut oil) 0 change in nAChR density
Bioallethrin (type I) 0.7 mg/kg, p.o.; 20% fat emulsion (egg lecithin/peanut oil) 4 months: ↑motor activity with lack of habituation; ↓mAChR density in CTX; 0 change in mAChR in HP, STR Eriksson and Fredriksson 1991
Deltamethrin (type II) 0.7 mg/kg, p.o.; 20% fat emulsion (egg lecithin/peanut oil) 4 months: ↑motor activity with lack of habituation; 0 change in mAChR in CTX, HP, STR Limitations: statistical analysis of biochemical data increases the possibility of type I error; unclear that litter is unit of treatment; in some cases, changes as small as 1–3% reported as significant (biochemistry); sex differences not considered/included; toxicity observed at high dose of deltamethin and bioallethrin by Eriksson and Nordberg (1990), with tolerance developing by the fourth day of dosing
Bioallethrin 0.42, 0.70, 42 mg/kg, p.o.; 20% fat emulsion (egg lecithin/peanut oil) PND17: ↑mAChR density in CTX; ↑low-affinity QNB (mAChR) binding 4 months: ↑motor activity with lack of habituation; ↓mAChR density in CTX Ahlbom et al. 1994
Bioallethrin 0.7 mg/kg, p.o.; 20% fat emulsion (egg lecithin/peanut oil) 4 treatment groups: vehicle as pup and 5 months; VB, vehicle as pup, bioallethrin at 5 months; BV, bioallethrin as pup, vehicle at 5 months; BB, bioallethrin as pup and 5 months PND10–16, once daily; again at 5 months for 7 days, once daily 5 months: ↓motor activity with lack of habituation in BB and BV groups Performance in H2O maze: reversal in BB groups; 0 effect in BV, VB groups mAChR density in CTX: ↑in BB treatment group; 0 effect in BV, VB groups Talts et al. 1998
Rat (Wistar)
Deltamethrin (type II) 0.7 mg/kg, i.p.; propylene glycol PND9–13 Examined on PNDs 12, 15, 21, and 30: delayed cerebellar cytogenesis and morphogenesis of interneurons, vascular damage with focal degeneration; ↓brain and body weight Patro et al. 1997 Strengths: only study examining morphology; culled litters to equal numbers; time course examined; within- litter dosing design Limitations: effects may be due to decreased growth, not direct neurotoxicity; inappropriate statistical models; toxicity; inappropriate statistical models; no control for “maternal” neglect effects in control vs. treated pups
Rat (Druckrey)
Cypermethrin Experiment 2: 5 mg/kg, p.o. (corn oil vehicle) PND10–13, 17, or 30 ↑BBB permeability at PNDs 13, 17, and 30 by 71, 61, and 80%; effect recovered by PND60 following withdrawal on PND18 Gupta et al. 1999a Strengths: control data demonstrate maturation of BBB; within-paper replication of effect; technical grade (94.5% purity) cypermethrinb Limitation: litter was not the statistical unit
Experiment 3: 2.5 mg/kg, p.o. (corn oil vehicle), (1/100 LD50) PND10–17 ↑BBB permeability by 28%
Allethrin 18 hr/day inhalation of vapors; unknown commercial product containing 3.6% Allethrin, 96% kerosine, 0.3% stabilizer PND2–19 ↓body (23%) and brain (17%) weights; ↑BBB permeability, LH levels on PND10 but not PND18; ↑(small) in conjugated dienes (measure of lipid peroxidation) on PND10; ↓GSH 17% on PND10; ↑GSH by 28% on PND18 Gupta et al. 1999b Strengths: replication of fluorescence levels on PND10 compared with Gupta et al. (1999a); litters culled to 8 pups/dam (size of litter is known) Limitations: unknown formulation; exposure to kerosine > > allethrin; no kerosine control
Deltamethrin 1.0 mg/kg, p.o., deltamethrin formulation in corn oil GD14–20 Delayed surface righting reflex 6 and 12 weeks postnatal: ↑AChE activity; ↑GAP-43 immunohistochemistry (both % area and total number of positive cells); ↓QNB Bmax; ↓relearning in Y-maze task Aziz et al. 2001 Strengths: examined two time points; behavioral and biochemical changes Limitations: unknown formulation, corn oil used as “control”; unclear that litter is statistical unit; maze learning procedure is poorly described, and “relearning” is poorly defined
Rat (Wistar)
Deltamethrin 7 mg/kg, p.o. 2.8% EC formulation, peanut oil GD5–21 ↓weight of unspecified brain regions at PND22(?);↑resorptions and neonatal death; delayed surface righting, eye opening, fur development, incisor eruption, and pinna detachment; ↓grip strength; ↓motor activity at PNDs 21 and 42; altered regional polyamine levels Husain et al. 1992 Strengths: work uniquely covers effects of pyrethroids on different periods of perinatal development from shortly after conception to postweaning, and suggests that effects may depend on the exposure period (includes Malaviya et al. 1993). However, different compounds were utilized; effects on maternal parameters, general toxicity recorded; litter size adjusted to an average of 8 pups/litter
Fenvalerate 10 mg/kg, p.o.; 20% EC formulation, peanut oil Delayed surface righting, eye opening, fur development, incisor eruption, and pinna detachment; ↓grip strength; 0 effect on motor activity; altered regional polyamine levels Limitations: formulated products used; lack of relevant vehicle controls; general or less specific toxicity may be indicated by changes in fur development, pinna detachment; statistical models are often inappropriate; descriptions of comparisons (data sets) used for statistical tests are sometimes unclear or confusing; not clear that litter is the statistical unit
Cypermethrin 15 mg/kg, p.o.; 25% EC formulation; peanut oil Delayed surface righting, eye opening, fur development, incisor, eruption and pinna detachment; 0 effect on motor activity; altered regional polyamine levels
Deltamethrin 7 mg/kg; 2.8% EC formulation, corn oil PND22–37 ↓hippocampal weight without effect on other brain regions; ↑mitochondrial monamine oxidase and microsomal AChE without effect on Na/K ATPase; ↑spontaneous locomotor activity; ↓conditioned avoidance response; altered regional polyamine levels Husain et al. 1994 Strengths: work uniquely covers effects of pyrethroids on different periods of perinatal development from shortly after conception to postweaning, and suggests that effects may depend on the exposure period (includes Malaviya et al. 1993). However, different compounds were utilized; effects on maternal parameters, general toxicity recorded; litter size adjusted to an average of 8 pups/litter
Rat (Charles Wistar) Limitations: formulated products used; lack of relevant vehicle controls; general or less specific toxicity may be indicated by changes in fur development, pinna detachment; statistical models are often inappropriate; descriptions of comparisons (data sets) used for statistical tests are sometimes unclear or confusing; not clear that litter is the statistical unit
Fenvalerate 10 mg/kg. p.o.; corn oil GD5–21 (gestational exposure) or PND1–15 (lactational exposure) Biochemical outcomes measured at 3 weeks of age 0 effect on dam weight, food/water intake, gestation length, no. of offspring, sex ratio Gestational exposure: ↓MAO, Na/K-ATPase activity; spiroperidol binding; ↑AChE activity Lactational exposure: ↓MAO, AChE activity; ↑spiroperidol, QNB binding Malaviya et al. 1993
Cypermethrin 15 mg/kg, p.o.; corn oil GD5–21 (gestational exposure) or PND1–15 (lactational exposure) Biochemical outcomes measured at 3 weeks of age 0 effect on dam weight, food/water intake, gestation length, no. of offspring, sex ratio Gestational exposure: 0 effect on MAO, Na/K-ATPase, AChE activity; spiroperidol binding ↓ QNB binding Lactational exposure: ↓Na/K-ATPase, AChE activity; ↑ spiroperidol, QNB binding Malaviya et al. 1993
Rat (Wistar)
d-Allethrin 0.43–74.2 mg/m3 Inhalation; unknown vehicle PND10–16, 6hr/day 0 Effects on weight gain, motor activity, mAChR density when assessed on PND17 and 4 months Tsuji et al. 2002 Strengths: measured air levels of allethrin during exposure; provides additional exposure information; multiple dose levels; litter controlledc
0 effect in Morris water maze at 11 months Limitations: absence of positive controls; this would demonstrate that lack of effect is true negative
Mouse (ICR)
Permethrin (cis or trans) Experiment 1: 0.33 to 33 μg/ml cis-permethrin or 33 μg/mL trans-permethrin in drinking water; 0.33 μg/mL DMSO vehicle PND0–21 0 effect on weight in dam, pups; concentration-dependent decrease in c-fos mRNA in cerebellum at PND21; trend toward decrease in BDNF mRNA at PND21; 0 effect on β -actin mRNA Imamura et al. 2002 Strengths: water consumption (ingested dose) measured; replication of c-fos decrease by different routes of exposure; similar findings following in vitro exposure to cerebellar granule cells (Imamura et al. 2000)
Experiment 2: 1 mg/day cis-permethrin, p.o. corn oil PND0–35 ↓ c-fos mRNA at PND21 only; 0 effect on β-actin mRNA at any time Limitations: did not use litter as statistical unit. 3–4 samples/litter; BDNF data variable
Mouse (NMRI)
Deltamethrin 0.7 mg/ml p.o.; 20% fat emulsion (egg lecithin/peanut oil) Hypothermic, normothermic, and hyperthermic groups PND10–16 Pup mortality in hypothermic groups (control and S-bioallethrin), including cannibalism; hypothermic pups displayed reduced motility; body weight gain PND10–17 was affected by conditions of hypothermia, hyperthermia; rectal temperature was affected by environmental temperature, differences in temperature between control and deltamethrin-treated animals were present in hypothermic but not hyperthermic animals; environmental temperature altered brain weight, with effects of S-bioallethrin and deltamethrin observed only in hypothermic animals; both deltamethrin and S-bioallethrin decreased brain/body weight ratios in hypothermic animals; QNB binding: on PND17, mAChR density was increased in both sexes by S-bioallethrin in hypothermic and normothermic groups; no differences were observed in the hyperthermic group or in the deltamethrin-treated groups Pauluhn and Schmuck 2003 Strengths: technical compound of known purity used (99.8% for deltamethrin and 95.7% for S-bioallethrin); statistical analysis using ANOVAs; randomized selection of pups and dams for treatment groups from a pool. Limitations: pup mortality observed in control, S-bioallethrin groups with no information provided regarding number of pups lost/cannibalized; replacement pups came from a pool of pups that had been housed under “normal conditions,” which likely differed in temperature from group that lost pups (hypothermic pups); sample size for various end points is difficult to determine; examined only PND17 animals; not known if temperature differences could contribute to long-term changes in mAChR expression; randomized assignment of pups to dams does not control for maternal effects; did not demonstrate that typical p.o. dosing causes hypothermia; because of design of study (incomplete block design), comparisons between vehicle and pyrethroid treatments cannot be made; study design was to compare effects of different temperature conditions within these treatments
S-bioallethrin 0.7 mg/mL p.o.; 20% fat emulsion (egg lecithin/peanut oil) Hypothermic, normothermic and hyperthermic groups PND10–16
Abbreviations: 5HIAA, 5-hydroxyindoleacetic acid; 5HT, serotonin; AChE, acetylcholinesterase; ANOVA, analysis of variance; BBB, blood–brain barrier; BDNF, brain-derived neurotropic factor; Bmax, maximum number of binding sites; CTX, cortex; DA, dopamine; DMSO, dimethyl sulfoxide; EC, emulsifiable concentrate; GAP-43, growth-associated protein 43; GSH, glutathione; HP, hippocampus; HVA, homovanillic acid; i.p., intraperitoneal; LD50, dose lethal to 50%; LH, luteinizing hormone; MAO, monoamine oxidase; NA, noradrenaline; nAChR, nicotinic acetylcholine receptor; p.o., per os; STR, stratum.
a Publications by the same group of authors are indicated by shading; in some cases, comments are made on groups of papers published by the same group of authors rather than on individual papers.
b Not reported in original publication (Gupta et al. 1999a); data from A.K. Agarwal (personal communication).
c Not reported in original publication (Tsuji et al. 2002); data from R. Tsuji (personal communication).
Table 4 Summary of data from studies in NMRI mice (dosed once daily on PND10–16) submitted to the U.S. EPA.
Compound Dose/route/vehicle Effects References Comments
d-Allethrin 0.15, 4, or 100 mg/m3 6 hr/day, inhalation; polyethylene glycol PND17: motor activity: increased habituation in 0.15 mg/m3 females when compared to control; effects not dose-related; mAChR: 25% ↑in QNB in cortex, smaller changes in hippocampus and striatum; nAChR: 40–60% ↓in cortex, hippocampus, and striatum in both sexes; AChE: ↑ by 70–80% in striatum but not significant due to large variability; ChAT: 0 effect Ivens et al., unpublished data Strengths: technical compound, 95% purity; group sizes of 10; litter was statistical unit; good statistical analysis, males and females considered separately; second control group was included; closely replicates methodology of Eriksson and co-workers (see Table 2) for motor activity measurements
Limitations: not peer-reviewed or published; some biochemical measurements were variable and not dose-related
4 months: motor activity: no significant effects; mAChR: 0 effect; nAChR: large sporadic changes but no clear sex- or dose-related trends; AChE: 0 effect; ChAT: 0 effect
Cyfluthrin 6, 15 or 50 mg/m3, 6 hr/day, inhalation; polyethylene glycol All pups died in 50 mg/m3 dose group; 15 mg/m3 pups had clinical signs including “clonic seizures” (probably tremors and/or choreoathetosis); ↓pup weight in 15 mg/m3 and in 5 mg/m3 females Jekat et al., unpublished data Strengths: technical compound, 96.8% purity; group sizes of 10; litter was statistical unit; good within-lab replicability for motor activity [comparison of data with Ivens et al. (unpublished data)]; closely replicates methodology of Eriksson and co-workers (see Table 2) for motor activity measurements
PND17: no measurements Limitations: not peer-reviewed or published; only examined adults; general toxicity observed; QNB data variable, no dose-related effects, difficult to compare with other studies because presented either as dpm or percent of control
4 months: motor activity: 15 mg/m3 females were hyperactive and had decreased habituation in horizontal and vertical v activity;mAChR: ↓QNB binding (not statistically significant) of ~22% in 15 mg/m3 males
Abbreviations: AChE, acetylcholinesterase; ChAT, choline acetyltransferase; nAChR, nicotinic acetylcholine receptor.
Table 5 Summary of effects on mAChR and motor activity after developmental exposure to pyrethroids.
MAChR expressiona Motor activity
Compoundb Preweaning Adult Preweaning Adult Reference
d-Allethrin ↑CTX 0 CTX ↑HB 0 Ivens et al., unpublished data
d-Allethrin 0 0 ND 0 Tsuji et al. 2002
Bioallethrin 0 CTX ↑CTXc ND ↑MA, ↓HB Muhammad and Ray, unpublished data
Bioallethrin ND 0 CTX ND ↑MA, ↓HB Talts et al. 1998
Bioallethrin/bioallethrin ND ↑ CTX ND ↑MA, ↓HB
Bioallethrin ND ↓CTX; 0 HP, STR 0 MA, 0 HB ↑MA, ↓HB Eriksson and Fredriksson 1991
Bioallethrin ↑CTX ↓CTX ND ↑MA, ↓HB Ahlbom et al. 1994
Bioallethrin ↑CTX ND ND ND Eriksson and Nordberg 1990
S-Bioallethrin ↑CTX ND ND ND Pauluhn and Schmuck 2003
Cyfluthrin ND 0 CTX ND In females,↑MA, ↓HB Jekat et al., unpublished data
Cypermethin ↓STR (gestation experiment), ↑STR (lactation experiment) ND ND ND Malaviya et al. 1993
Deltamethrin ND ↓HP ND ND Aziz et al. 2001
Deltamethrin ↑CTX ↑CTX ND ↓HB Muhammad and Ray, unpublished data
Deltamethrin ND 0 CTX, HP, STR 0 MA, 0 HB ↑MA, ↓HB Eriksson and Fredriksson 1991
Deltamethrin ↓HP ND ND ND Eriksson and Nordberg 1990
Deltamethrin 0 CTX ND ND ND Pauluhn and Schmuck 2003
Fenvalerate 0 STR (gestation experiment), ↑STR (lactation experiment) ND ND ND Malaviya et al. 1993
Abbreviations: 0, end point was examined and was not affected by treatment; CTX, cortex; HB, habituation; HP, hippocampus; MA, motor activity; ND, not determined; STR, striatum.
a As measured by QNB binding.
b Compounds are arranged in alphabetical order.
c An increase in QNB binding was observed in one “cohort” but was not consistently observed in all “cohorts” in studies by this group.
See Table 3 for complete details.
==== Refs
References
Abbey H Howard E 1973 Statistical procedures in developmental studies on a species with multiple offspring Dev Psychobiol 6 329 335 4793362
Adams J Vorhees CV Middaugh LD 1990 Developmental neurotoxicity of anticonvulsants: human and animal evidence on phenytoin Neurotoxicol Teratol 12 203 214 2196420
Ahlbom J Fredriksson A Eriksson P 1994 Neonatal exposure to a type-I pyrethroid (bioallethrin) induces dose-response changes in brain muscarinic receptors and behaviour in neonatal and adult mice Brain Res 645 318 324 8062093
Akopian AN Souslova V England S Okuse K Ogata N Ure J 1999 The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways Nat Neurosci 2 541 548 10448219
Albrieux M Platel J Dupuis A Villaz M Moody WJ 2004 Early expression of sodium channel transcripts and sodium current by Cajal-Retzius cells in the preplate of the embryonic mouse neocortex J Neurosci 24 1719 1725 14973256
Anadón A Martinez-Larrañaga MR Diaz MJ Bringas P 1991 Toxicokinetics of permethrin in the rat Toxicol Appl Pharmacol 110 1 8 1871768
Anadón A Martinez-Larrañaga MR Fernandez-Cruz ML Diaz MJ Fernandez MC Martinez MA 1996 Toxicokinetics of deltamethrin and its 4’-HO-metabolite in the rat Toxicol Appl Pharmacol 141 8 16 8917670
Andersen ME Dennison JE 2001 Mode of action and tissue dosimetry in current and future risk assessments Sci Total Environ 274 3 14 11453304
Andersen ME Krishnan K Conolly RB McClellan RO 1992 Mechanistic toxicology research and biologically-based modeling: partners for improving quantitative risk assessments CIIT Act 12 1 7
Aziz MH Agrawal AK Adhami VM Shukla Y Seth PK 2001 Neurodevelopmental consequences of gestational exposure (GD14-GD20) to low dose deltamethrin in rats Neurosci Lett 300 161 165 11226636
Benn SC Costigan M Tate S Fitzgerald M Woolf CJ 2001 Developmental expression of the TTX-resistant voltage-gated sodium channels Nav1.8 (SNS) and Nav1.9 (SNS2) in primary sensory neurons J Neurosci 21 6077 6085 11487631
Berkowitz GS Obel J Deych E Lapinski R Godbold J Liu Z 2003 Exposure to indoor pesticides during pregnancy in a multiethnic, urban cohort Environ Health Perspect 111 79 84 12515682
Boiko T Rasband MN Levinson SR Caldwell JH Mandel G Trimmer JS 2001 Compact myelin dictates the differential targeting of two sodium channel isoforms in the same axon Neuron 30 91 104 11343647
Brock WJ Rodricks JV Rulis A Dellarco VL Gray GM Lane RW 2003 Food safety: risk assessment methodology and decision-making criteria Int J Toxicol 22 435 451 14680991
Bryant R Bite MG eds. 2003 Global Insecticide Directory. 3rd ed. Orpington, Kent, UK:Agranova.
Cantalamessa F 1993 Acute toxicity of two pyrethroids, permethrin, and cypermethrin in neonatal and adult rats Arch Toxicol 67 510 513 8240001
Casida JE 1980 Pyrethrum flowers and pyrethroid insecticides Environ Health Perspect 34 189 202 6993201
Casida JE Quistad GB 1998 Golden age of insecticide research: past, present, or future? Annu Rev Entomol 43 1 16 9444749
Catterall WA 1999 Molecular properties of brain sodium channels: an important target for anticonvulsant drugs Adv Neurol 79 441 456 10514834
Claes L Del Favero J Ceulemans B Lagae L Van Broeckhoven C De Jonghe P 2001 De novo mutations in the sodium-channel gene SCN1A cause severe myoclonic epilepsy of infancy Am J Hum Genet 68 1327 1332 11359211
Cohen SM Klaunig J Meek ME Hill RN Pastoor T Lehman-McKeeman L 2004 Evaluating the human relevance of chemically-induced animal tumors Toxicol Sci 78 181 186 14737005
Conolly RB 2002 The use of biologically based modeling in risk assessment Toxicology 181–182 275 279
Costa L 1998 Biochemical and molecular neurotoxicology: relevance to biomarker development, neurotoxicity testing and risk assessment Toxicol Lett 102–103 417 421
Crofton KM Kehn LS Gilbert ME 1995 Vehicle and route dependent effects of a pyrethroid insecticide, deltamethrin, on motor function in the rat Neurotoxicol Teretol 17 489 495
Donahue LM Coates PW Lee VH Ippensen DC Arze SE Poduslo SE 2000 The cardiac sodium channel mRNA is expressed in the developing and adult rat and human brain Brain Res 887 335 343 11134623
Elliot M Farnham AW Janes NF Needham PH Pulman DA 1974. Insecticidally active conformations of pyrethroids. In: Mechanism of Pesticide Action (Kohn GK, ed). Washington, DC:American Chemical Society, 80–91.
Eriksson P Fredriksson A 1991 Neurotoxic effects of two different pyrethroids, bioallethrin and deltamethrin, on immature and adult mice: changes in behavioral and muscarinic receptor variables Toxicol Appl Pharmacol 108 78 85 2006507
Eriksson P Johansson U Ahlbom J Fredriksson A 1993 Neonatal exposure to DDT induces increased susceptibility to pyrethroid (bioallethrin) exposure at adult age—changes in cholinergic muscarinic receptor and behavioural variables Toxicology 77 21 30 8442015
Eriksson P Nordberg A 1990 Effects of two pyrethroids, bioallethrin and deltamethrin, on subpopulations of muscarinic and nicotinic receptors in the neonatal mouse brain Toxicol Appl Pharmacol 102 456 463 2315914
Escayg A Heils A MacDonald BT Haug K Sander T Meisler MH 2001 A novel SCN1A mutation associated with generalized epilepsy with febrile seizures plus—and prevalence of variants in patients with epilepsy Am J Hum Genet 68 866 873 11254445
Farm Chemicals Handbook 1997. Farm Chemicals Handbook ‘97. Willoughby, OH:Meister Publishing.
Felts PA Yokoyama S Dib-Hajj S Black JA Waxman SG 1997 Sodium channel alpha-subunit mRNAs I, II, III, NaG, Na6 and hNE (PN1): different expression patterns in developing rat nervous system Brain Res Mol Brain Res 45 71 82 9105672
Food Quality Protection Act of 1996 1996. Public Law 104-170.
Gammon DW Brown MA Casida JE 1981 Two classes of pyrethroid action in the cockroach Pest Biochem Physiol 15 181 191
Goldin AL Barchi RL Caldwell JH Hofmann F Howe JR Hunter JC 2000 Nomenclature of voltage-gated sodium channels Neuron 28 365 368 11144347
Gomes MS Bernardi MM Spinosa HS 1991a Pyrethroid insecticides and pregnancy: effect on physical and behavioral development of rats Vet Hum Toxicol 33 315 317 1897124
Gomes MS Bernardi MM Spinosa HS 1991b Effects of prenatal insecticide exposure on the sexual development of rats Vet Hum Toxicol 33 427 428 1746129
Gray AJ Soderlund DM 1985. Mammalian toxicology of pyrethroids. In: Insecticides (Hutson DH, Roberts TR, eds). New York:Wiley, 193–248.
Gupta A Agarwal R Shukla GS 1999a Functional impairment of blood-brain barrier following pesticide exposure during early development in rats Hum Exp Toxicol 18 174 179 10215108
Gupta A Nigam D Gupta A Shukla GS Agarwal AK 1999b Effect of pyrethroid-based liquid mosquito repellent inhalation on the blood-brain barrier function and oxidative damage in selected organs of developing rats J Appl Toxicol 19 67 72 9989480
Gustafson TA Clevinger EC O’Neill TJ Yarowsky PJ Krueger BK 1993 Mutually exclusive exon splicing of type III brain sodium channel alpha subunit RNA generates developmentally regulated isoforms in rat brain J Biol Chem 268 18648 18653 8395514
Hatta T Ohmori H Murakami T Takano M Yamashita K Yasuda M 1999 Neurotoxic effects of phenytoin on post-natal mouse brain development following neonatal administration Neurotoxicol Teratol 21 21 28 10023798
Heudorf U Angerer J Drexler H 2004 Current internal exposure to pesticides in children and adolescents in Germany: urinary levels of metabolites of pyrethroid and organophosphorous insecticides Int Arch Occup Environ Health 77 67 72 14551781
Holson RR Pearce B 1992 Principles and pitfalls in the analysis of prenatal treatment effects in multiparous species Neurotoxicol Teratol 14 221 228 1635542
Husain R Malaviya M Seth PK Husain R 1992 Differential responses of regional brain polyamines following in utero exposure to synthetic pyrethroid insecticides: a preliminary report Bull Environ Contam Toxicol 49 402 409 1498477
Husain R Malaviya M Seth PK Husain R 1994 Effect of deltamethrin on regional brain polyamines and behaviour in young rats Pharmacol Toxicol 74 211 215 8090688
Imamura L Hasegawa H Kurashina K Hamanishi A Tabuchi A Tsuda M 2000 Repression of activity-dependent c-fos and brain-derived neurotrophic factor mRNA expression by pyrethroid insecticides accompanying a decrease in Ca2+ influx into neurons J Pharmacol Exp Ther 295 1175 1182 11082455
Imamura L Hasegawa H Kurashina K Matsuno T Tsuda M 2002 Neonatal exposure of newborn mice to pyrethroid (permethrin) represses activity-dependent c-fos mRNA expression in cerebellum Arch Toxicol 76 392 397 12111003
Isom LL 2001 Sodium channel beta subunits: anything but auxiliary Neuroscientist 7 42 54 11486343
Isom LL 2002 Beta subunits: players in neuronal hyperexcitability? Novartis Found Symp 241 124 138 11771642
Jarabek A Bruins R Choudhury H Collette T Crofton K Dellarco V 2002. Perchlorate Environmental Contamination: Toxicological Review and Risk Characterization. External Review Draft. NCEA-1-0503. Washington, DC:Office of Research and Development, U.S. Environmental Protection Agency.
Jenkins SM Bennett V 2002 Developing nodes of Ranvier are defined by ankyrin-G clustering and are independent of paranodal axoglial adhesion Proc Natl Acad Sci USA 99 2303 2308 11842202
Kallen RG Sheng ZH Yang J Chen LQ Rogart RB Barchi RL 1990 Primary structure and expression of a sodium channel characteristic of denervated and immature rat skeletal muscle Neuron 4 233 242 2155010
Kaneko H Miyamoto J 2001. Pyrethroid chemistry and metabolism. In: Handbook of Pesticide Toxicology, Vol 2: Agents (Krieger R, Doull J, Ecobichon D, eds). San Diego:Academic Press, 1263–1288.
Kearney JA Plummer NW Smith MR Kapur J Cummins TR Waxman SG 2001 A gain-of-function mutation in the sodium channel gene SCN2A results in seizures and behavioral abnormalities Neuroscience 102 307 317 11166117
Laird JM Souslova V Wood JN Cervero F 2002 Deficits in visceral pain and referred hyperalgesia in Nav1.8 (SNS/PN3)-null mice J Neurosci 22 8352 8356 12351708
Lawrence LJ Casida JE 1982 Pyrethroid toxicology: mouse intracerebral structure-toxicity relationships Biochem Physiol 18 9 14
Lazarini CA Florio JC Lemonica IP Bernardi MM 2001 Effects of prenatal exposure to deltamethrin on forced swimming behavior, motor activity, and striatal dopamine levels in male and female rats Neurotoxicol Teratol 23 665 673 11792535
Lee SH Soderlund DM 2001 The V410M mutation associated with pyrethroid resistance in Heliothis virescens reduces the pyrethroid sensitivity of house fly sodium channels expressed in Xenopus oocytes Insect Biochem Mol Biol 31 19 29 11102831
Lund AE Narahashi T 1982 Dose-dependent interaction of the pyrethroid isomers with sodium channels of squid membranes Neurotoxicology 3 11 24 6290951
Malaviya M Husain R Seth PK Husain R 1993 Perinatal effects of two pyrethroid insecticides on brain neurotransmitter function in the neonatal rat Vet Hum Toxicol 35 119 122 8385837
Meek ME Bucher JR Cohen SM Dellarco V Hill RN Lehman-McKeeman LD 2003 A framework for human relevance analysis of information on carcinogenic modes of action Crit Rev Toxicol 33 591 653 14727733
Meisler MH Kearney J Ottman R Escayg A 2001 Identification of epilepsy genes in human and mouse Annu Rev Genet 35 567 588 11700294
Meisler MH Kearney JA Sprunger LK MacDonald BT Buchner DA Escayg A 2002 Mutations of voltage-gated sodium channels in movement disorders and epilepsy Novartis Found Symp 241 72 81 11771652
Mileson BE Chambers JE Chen WL Dettbarn W Ehrich M Eldefrawi AT 1998 Common mechanism of toxicity: a case study of organophosphorus pesticides Toxicol Sci 41 8 20 9520337
Moniz AC Bernardi MM Souza-Spinosa HS Palermo-Neto J 1990 Effects of exposure to a pyrethroid insecticide during lactation on the behavior of infant and adult rats Braz J Med Biol Res 23 45 48 2386848
Moniz AC Cruz-Casallas PE Oliveira CA Lucisano A Florio JC Nicolau AA 1999 Perinatal fenvalerate exposure: behavioral and endocrinology changes in male rats Neurotoxicol Neuroteratol 21 611 618
Muller KE Barton CN Benignus VA 1985 Recommendation for appropriate statistical practice in toxicological experiments Neurotoxicology 5 113 126 6542184
Narahashi T 1996 Neuronal ion channels as the target sites of insecticides Pharmacol Toxicol 78 1 14 8821391
Narahashi T 2001. Neurophysiological effects of insecticides. In: Handbook of Pesticide Toxicology. Vol 1: Principles (Krieger R, Doull J, Ecobichon D, eds). San Diego:Academic Press, 335–350.
Noebels JL 2002 Sodium channel gene expression and epilepsy Novartis Found Symp 241 109 120 11771641
Novakovic SD Eglen RM Hunter JC 2001 Regulation of Na+ channel distribution in the nervous system Trends Neurosci 24 473 478 11476887
Ogata N Ohishi Y 2002 Molecular diversity of structure and function of the voltage-gated Na+ channels Jpn J Pharmacol 88 365 377 12046980
Ohmori H Ogura H Yasuda M Nakamura S Hatta T Kawano K 1999 Developmental neurotoxicity of phenytoin on granule cells and Purkinje cells in mouse cerebellum J Neurochem 72 1497 1506 10098854
Ohmori H Yamashita K Hatta T Yamasaki S Kawamura M Higashi Y 1997 Effects of low-dose phenytoin administered to newborn mice on developing cerebellum Neurotoxicol Teratol 19 205 211 9200141
Patro N Mishra SK Chattopadhyay M Patro IK 1997 Neurotoxicological effects of deltamethrin on the postnatal development of cerebellum in rat J Biosci 22 117 130
Pauluhn J Schmuck G 2003 Critical analysis of potential body temperature confounders on neurochemical endpoints caused by direct dosing and maternal separation in neonatal mice: a study of bioallethrin and deltamethrin interactions with temperature on brain muscarinic receptors J Appl Toxicol 23 9 18 12518331
Planells-Cases R Caprini M Zhang J Rockenstein EM Rivera RR Murre C 2000 Neuronal death and perinatal lethality in voltage-gated sodium channel alpha(II)-deficient mice Biophys J 78 2878 2891 10827969
Plummer NW McBurney MW Meisler MH 1997 Alternative splicing of the sodium channel SCN8A predicts a truncated two-domain protein in fetal brain and non-neuronal cells J Biol Chem 272 24008 24015 9295353
Porter JD Goldstein LA Kasarskis EJ Brueckner JK Spear BT 1996 The neuronal voltage-gated sodium channel, SCN8A, is essential for postnatal maturation of spinal, but not oculomotor, motor units Exp Neurol 139 328 334 8654536
Ray DE 2001. Pyrethroid insecticides: mechanisms of toxicity, systemic poisoning syndromes, paresthesia, and therapy. In: Handbook of Pesticide Toxicology: Vol 2: Agents (Krieger R, Doull J, Ecobichon D, eds). San Diego:Academic Press, 1289–1303.
Reily EP Meyer LJ 1984 Consideration for the design, implementation and interpretation of animal models of fetal alcohol effects Neurotoxicol Teratol 6 97 101
Sarao R Gupta SK Auld VJ Dunn RJ 1991 Developmentally regulated alternative RNA splicing of rat brain sodium channel mRNAs Nucleic Acids Res 19 5673 5679 1658739
Schettgen T Heudorf U Drexler H Angerer J 2002 Pyrethroid exposure of the general population-is this due to diet Toxicol Lett 134 141 145 12191872
Schilling MA Inman SL Morford LL Moran MS Vorhees CV 1999 Prenatal phenytoin exposure and spatial navigation in offspring: effects on reference and working memory and on discrimination learning Neurotoxicol Teratol 21 567 578 10492391
Scott JG Matsumura F 1983 Evidence for two types of toxic action of pyrethroids on susceptible and DDT-resistant German cockroaches Pestic Biochem Physiol 19 141 150
Shah BS Stevens EB Pinnock RD Dixon AK Lee K 2001 Developmental expression of the novel voltage-gated sodium channel auxiliary subunit β3, in rat CNS J Physiol 534 763 776 11483707
Sheets LP 2000 A consideration of age-dependent differences in susceptibility to organophosphorus and pyrethroid insecticides Neurotoxicology 21 57 64 10794385
Sheets LP Doherty JD Law MW Reiter LW Crofton KM 1994 Age-dependent differences in the susceptibility of rats to deltamethrin Toxicol Appl Pharmacol 126 186 190 8184428
Smith TJ Lee SH Ingles PJ Knipple DC Soderlund DM 1997 The L1014F point mutation in the house fly Vssc1 sodium channel confers knockdown resistance to pyrethroids Insect Biochem Mol Biol 27 807 812 9474777
Smith TJ Soderlund DM 1998 Action of the pyrethroid insecticide cypermethrin on rat brain IIa sodium channels expressed in Xenopus oocytes Neurotoxicology 19 823 832 9863771
Smith TJ Soderlund DM 2001 Potent actions of the pyrethroid insecticides cismethrin and cypermethrin on rat tetrodotoxin-resistant peripheral nerve (SNS/PN3) sodium channels expressed in Xenopus oocytes Pestic Biochem Physiol 70 52 61
Soderlund DM Clark JM Sheets LP Mullin LS Piccirillo VJ Sargent D 2002 Mechanisms of pyrethroid neurotoxicity: implications for cumulative risk assessment Toxicology 171 3 59 11812616
Sonich-Mullin C Fielder R Wiltse J Baetcke K Dempsey J Fenner-Crisp P 2001 IPCS conceptual framework for evaluating a MOA for chemical carcinogenesis Regul Toxicol Pharmacol 34 146 152 11603957
Staatz CG Bloom AS Lech JJ 1982 Effect of pyrethroids on [3 H]kainic acid binding to mouse forebrain membranes Toxicol Appl Pharmacol 64 566 569 7135402
Talts U Fredriksson A Eriksson P 1998 Changes in behavior and muscarinic receptor density after neonatal and adult exposure to bioallethrin Neurobiol Aging 19 545 552 10192213
Tatebayashi H Narahashi T 1994 Differential mechanism of action of the pyrethroid tetramethrin on tetrodotoxin-sensitive and tetrodotoxin-resistant sodium channels J Pharmacol Exp Ther 270 595 603 8071852
Tilson HA 2000a New horizons: future directions in neurotoxicology Environ Health Perspect 108 suppl 3 429 441 10852840
Tilson HA 2000b Neurotoxicology risk assessment guidelines: developmental neurotoxicology Neurotoxicology 21 189 194 10794399
Trainer VL McPhee JC Boutelet-Bochan H Baker C Scheuer T Babin D 1997 High affinity binding of pyrethroids to the alpha subunit of brain sodium channels Mol Pharmacol 51 651 657 9106631
Tsuji R Kobayashi K Ikeda M Yoshioka T Yamada T Seki T 2002 Lack of changes in brain muscarinic receptor and motor activity of mice after neonatal inhalation exposure to d-allethrin J Appl Toxicol 22 423 429 12424746
U.S. EPA 2002. Organophosphate Pesticides: Revised Cumulative Risk Assessment. Washington, DC:Office of Pesticide Programs, Office of Prevention, Pesticides and Toxic Substances. Available: http://www.epa.gov/pesticides/cumulative/rra-op/ [accessed 22 December 2004].
Vais H Atkinson S Eldursi N Devonshire AL Williamson MS Usherwood PN 2000 A single amino acid change makes a rat neuronal sodium channel highly sensitive to pyrethroid insecticides FEBS Lett 470 135 138 10734222
Vais H Williamson MS Devonshire AL Usherwood PN 2001 The molecular interactions of pyrethroid insecticides with insect and mammalian sodium channels Pest Manag Sci 57 877 888 11695180
Valentine WM 1990 Toxicology of selected pesticides, drugs, and chemicals. Pyrethrin and pyrethroid insecticides Vet Clin North Am Small Anim Pract 20 375 382 2180183
Verschoyle RD Aldridge WN 1980 Structure-activity relationships of some pyrethroids in rats Arch Toxicol 45 325 329 7447703
Verschoyle RD Barnes JM 1972 Toxicity of natural and synthetic pyrethrins to rats Pestic Biochem Physiol 2 308 311
Vorhees CV Acuff-Smith KD Schilling MA Moran MS 1995 Prenatal exposure to sodium phenytoin in rats induces complex maze learning deficits comparable to those induced by exposure to phenytoin acid at half the dose Neurotoxicol Teratol 17 627 632 8747744
Wallace RH Scheffer IE Barnett S Richards M Dibbens L Desai RR 2001 Neuronal sodium-channel alpha1-subunit mutations in generalized epilepsy with febrile seizures plus Am J Hum Genet 68 859 865 11254444
Wang SY Barile M Wang GK 2001 A phenylalanine residue at segment D3-S6 in Nav1.4 voltage-gated Na+ channels is critical for pyrethroid action Mol Pharmacol 60 620 628 11502895
Watling KJ Kebabian JW Neumeyer JL eds. 1995. The RBI Handbook of Receptor Classification and Signal Transduction. Natick, MA:Research Biochemicals International.
Whitaker WR Clare JJ Powell AJ Chen YH Faull RL Emson PC 2000 Distribution of voltage-gated sodium channel alpha-subunit and beta-subunit mRNAs in human hippocampal formation, cortex, and cerebellum J Comp Neurol 422 123 139 10842222
Whitaker WR Faull RL Waldvogel HJ Plumpton CJ Emson PC Clare JJ 2001 Comparative distribution of voltage-gated sodium channel proteins in human brain Brain Res Mol Brain Res 88 37 53 11295230
Whyatt RM Camann DE Kinney PL Reyes A Ramirez J Dietrich J 2002 Residential pesticide use during pregnancy among a cohort of urban minority women Environ Health Perspect 110 507 514 12003754
|
15687048
|
PMC1277854
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 14; 113(2):123-136
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7254
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7473ehp0113-00013715687049ResearchArticlesAirborne Multidrug-Resistant Bacteria Isolated from a Concentrated Swine Feeding Operation Chapin Amy Rule Ana Gibson Kristen Buckley Timothy Schwab Kellogg Johns Hopkins Bloomberg School of Public Health, Department of Environmental Health Sciences, Baltimore, Maryland, USAAddress correspondence to K. Schwab, Johns Hopkins Bloomberg School of Public Health, Department of Environmental Health Sciences, 615 N. Wolfe St., Room E6620, Baltimore, MD 21205-2103 USA. Telephone: (410) 614-5753. Fax: (410) 955-9334. E-mail:
[email protected] thank the swine grower for providing access to the swine operation. We also thank K. Carroll and D. Flayhart for speciating the bacterial isolates and reviewing the manuscript; W. Merz for reviewing the manuscript; and E. Silbergeld for reviewing the manuscript and providing many helpful insights.
This research was supported by the Center for a Livable Future at the Johns Hopkins Bloomberg School of Public Health and the National Institute for Occupational Safety and Health, Education Research Center, Pilot Project Research Training Award (T42/CCT31049-09). A.C. is a Howard Hughes Medical Institute Predoctoral Fellow.
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 137 142 2 8 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. The use of nontherapeutic levels of antibiotics in swine production can select for antibiotic resistance in commensal and pathogenic bacteria in swine. As a result, retail pork products, as well as surface and groundwaters contaminated with swine waste, have been shown to be sources of human exposure to antibiotic-resistant bacteria. However, it is unclear whether the air within swine operations also serves as a source of exposure to antibiotic-resistant bacterial pathogens. To investigate this issue, we sampled the air within a concentrated swine feeding operation with an all-glass impinger. Samples were analyzed using a method for the isolation of Enterococcus. A total of 137 presumptive Enterococcus isolates were identified to species level using standard biochemical tests and analyzed for resistance to erythromycin, clindamycin, virginiamycin, tetracycline, and vancomycin using the agar dilution method. Thirty-four percent of the isolates were confirmed as Enterococcus, 32% were identified as coagulase-negative staphylococci, and 33% were identified as viridans group streptococci. Regardless of bacterial species, 98% of the isolates expressed high-level resistance to at least two antibiotics commonly used in swine production. None of the isolates were resistant to vancomycin, an antibiotic that has never been approved for use in livestock in the United States. In conclusion, high-level multidrug-resistant Enterococcus, coagulase-negative staphylococci, and viridans group streptococci were detected in the air of a concentrated swine feeding operation. These findings suggest that the inhalation of air from these facilities may serve as an exposure pathway for the transfer of multidrug-resistant bacterial pathogens from swine to humans.
air samplingairborne bacteriaantibiotic resistanceCAFOconcentrated swine feeding operationmultidrug-resistant bacteria
==== Body
The development and persistence of multidrug-resistant bacteria pose increasing challenges to public health (Institute of Medicine 1998). Although the use of antibiotics in human medicine has influenced the emergence of antibiotic-resistant bacteria, the use of antibiotics in animal agriculture has markedly contributed to this critical problem as well (Cohen and Tauxe 1986; Gorbach 2001; Institute of Medicine 1998; National Research Council 1999; van den Boogard and Stobberingh 1999). In animal agriculture, antibiotics are administered for therapeutic purposes to treat infections, prophylactic purposes in advance of observed symptoms, and nontherapeutic purposes to promote growth and improve feed efficiency (Wegener 2003). In general, antibiotics are administered at higher concentrations for therapeutic and prophylactic use and lower concentrations for nontherapeutic use (Wegener 2003). It has been estimated that the nontherapeutic use of antimicrobials in livestock production comprises 60–80% of total antimicrobial production in the United States (Mellon et al. 2001). The swine industry alone uses an estimated 10.3 million pounds of antibiotics annually for nontherapeutic purposes. Among the antibiotics used are ampicillin, bacitracin, erythromycin, lincomycin, virginiamycin, and tetracycline (Food and Drug Administration 2004), some of which are important in human clinical medicine. The use of antibiotics for nontherapeutic purposes such as growth promotion has been shown to select for resistance to high concentrations of antibiotics in both pathogenic and commensal bacteria in swine (Aarestrup et al. 2000a, 2000b; Bager et al. 1997; Jensen et al. 2002; Wegener et al. 1999). For this reason, attention has been given to retail pork products as a source of human exposure to antibiotic-resistant bacteria (Donabedian et al. 2003; Gambarotto et al. 2001; Hayes et al. 2003; Sorensen et al. 2001; White et al. 2001). Yet the ingestion of pork products is not the only pathway of exposure for the transfer of resistant organisms from swine to humans. Environmental pathways of exposure may be equally important.
Along with the pork products, more than 110 million tons of swine waste—containing antibiotic-resistant bacteria—is produced at swine concentrated animal feeding operations (CAFOs) in the United States each year (Environmental Defense 1997). The practice of storing this waste in pits and open-air lagoons and subsequently applying the waste to land can lead to the contamination of soils and nearby surface and groundwaters. Several studies have reported the appearance of antibiotic residues and antibiotic-resistant bacteria in surface and groundwaters proximal to swine CAFOs (Campagnolo et al. 2002; Chee-Sanford et al. 2001). Campagnolo et al. (2002) suggested that swine waste may be a source of antimicrobial drugs in surface and groundwaters near swine facilities, and Chee-Sanford et al. (2001) found that groundwater can be affected by swine waste and serve as a potential source of exposure to antibiotic-resistance genes.
However, few studies have examined the air within swine CAFOs as an additional source of environmental exposure to antibiotic-resistant bacterial pathogens. It has been well documented that the air within swine CAFOs is highly contaminated with bacteria, yeasts, and molds. Mean total bacterial concentrations can range from 104 to 107 colony forming units (CFU)/m3 (Clark et al. 1983; Cormier et al. 1990; Crook et al. 1991; Predicala et al. 2002). Specific bacteria detected in the air of swine CAFOs have included the following potential human pathogens: Enterococcus, Staphylococcus, Pseudomonas, Bacillus, Listeria, and Escherichia coli (Cormier et al. 1990; Crook et al. 1991; Predicala et al. 2002). Yet, to date, these airborne pathogens have not been assessed for resistance to antibiotics that are commonly used in both swine production and clinical medicine. Hamscher et al. (2003) assessed the presence of antibiotics in dust samples collected at a swine production facility over two decades. Several different antibiotics, including tetracycline, tylosin (an analog to erythromycin), and chloramphenicol, could be detected in 90% of the dust samples tested (Hamscher et al. 2003). In abstract form within conference proceedings, Zahn et al. (2001) reported on the presence of tylosin and tylosin-resistant bacteria in the air released from three mechanically ventilated swine CAFOs. Their study indicated that tylosin-resistant bacteria, primarily Corynebacterium, accounted for 80% of total culturable bacteria detected. These results provided the first evidence of airborne antibiotic-resistant bacteria in swine CAFOs.
The goal of this study was to test air samples collected within a swine CAFO for the presence of antibiotic-resistant enterococci, gram-positive, catalase-negative cocci that are not only members of the normal intestinal flora of humans and animals but also capable of causing a variety of human and animal infections [National Nosocomial Infections Surveillance (NNIS) 2001]. Resistance to erythromycin, clindamycin, tetracycline, and virginiamycin [an analog to quinupristin/dalfopristin, which is used to treat vancomycin-resistant Enterococcus faecium infections in humans (Johnson and Livermore 1999)] was investigated. These drugs (or their analogs) have been approved for use in swine production for growth promotion, feed efficiency, and therapeutic purposes. Resistance to vancomycin also was tested. Vancomycin, an analog to avoparcin, which has been used extensively in animal agriculture in Europe, has never been approved for use in livestock in the United States.
Materials and Methods
Study site.
The study site is a swine finishing CAFO located in the Mid-Atlantic United States. The CAFO consists of two tunnel-ventilated swine houses built atop 12-ft deep concrete pits where swine waste is stored before periodic siphoning into a transport truck for off-farm disposal by land application. Each house has the capacity to hold 2,500 hogs; however, during the sampling period approximately 1,500 hogs were being housed in each building. Air sampling at the swine facility was conducted on 9 December 2003 and 5 January 2004.
Collection of air samples.
Air samples were collected at a calibrated flow rate of 12.5 L/min using all-glass impingers (AGI-30; Ace Glass, Vineland, NJ) designed to collect respirable particles, including bioaerosols, with an aerodynamic diameter < 5 μm. Impingers were autoclaved and filled with 20 mL phosphate-buffered saline (PBS) before sampling. On the first sampling day, sampling was conducted over a 30-min period. On the second sampling day, sampling was conducted for 60 min in order to increase yield. For the longer sampling period, the impinger solution was replenished with distilled deionized H2O to maintain the sampler collection efficiency (Lin et al. 1997) and avoid increasing liquid salinity. All sampling equipment was placed on top of a table (1.5 m from the ground) within an empty swine stall located within the facility approximately 30 m from the south wall of the swine facility where air exits through ventilation fans. At the time of sampling, four of eight 32-inch ventilation fans were in operation to maintain a farm-operator–designated target temperature of 21°C within the facility. Temperature and relative humidity were monitored throughout the sampling periods and were 22°C ± 1°C and 76 ± 4% respectively. Impingers were stored and transported back to the laboratory at 4°C.
Bacterial isolation and speciation.
Approximately 3 hr after the last air sample was collected, impinger liquid samples were analyzed in the laboratory. Because no standard method exists regarding the isolation of Enterococcus from air, the standard methods used for the isolation of Enterococcus from recreational water were modified to accommodate the air samples [U.S. Environmental Protection Agency (EPA) 2000]. All broths and agars were obtained from Becton Dickinson (Sparks, MD). Three 10-fold dilutions (using PBS as the diluent) of the impinger samples were plated (100 μL/plate) in duplicate on mE agar. Negative control plates were made by plating 100 μL of both the replenishing fluid that was transported to the site and the dilution liquid. All plates were incubated for 48 hr at 41.5°C under aerobic conditions. All resulting colonies were counted, and counts from dilution plates containing 30–300 CFU were used in back-calculations to determine the concentration of isolated bacteria per cubic meter of air within the swine CAFO. Colonies from sample dilution plates that ranged from pink to red in color (indicative of Enterococcus colonies) were streaked onto Enterococcosel agar and incubated for 24 hr at 41.5°C under aerobic conditions. CFUs characteristic of Enterococcus that formed a black precipitate on the Enterococcosel agar plates were considered presumptive Enterococcus (U.S. EPA 2000). Presumptive Enterococcus isolates were archived in a 20% glycerol, tryptic soy broth solution at –80°C for subsequent speciation and antimicrobial susceptibility testing.
All presumptive Enterococcus isolates, as well as the quality control strains E. faecium 19434 and Enterococcus faecalis 29212 (American Type Culture Collection, Manassas, VA), were streaked from –80°C archived stocks onto both tryptic soy agar and tryptic soy agar No. 2 with 5% defibrinated sheep blood (Quad Five, Ryegate, MT) and incubated for 24 hr at 37°C. All of the media formulations and test interpretations used have been described previously (Murray et al. 2003). Gram stains were prepared on all isolates to verify the presence of gram-positive cocci. Each isolate was tested for the production of catalase in the presence of 3% hydrogen peroxide. Catalase-positive isolates were identified as Staphylococcus species (except for one isolate, which was further tested for oxidase activity and identified as Micrococcus luteus). Each Staphylococcus isolate was inoculated onto 0.5 mL rabbit plasma (Becton Dickinson) to test for the production of coagulase. Catalase-negative isolates were differentiated further by pyrrolidonyl-arylamidase activity using Remel’s PYR kit (Remel, Lenexa, KS). The following biochemical tests were performed on the isolates displaying pyrrolidonyl-arylamidase activity: mannitol, arabinose, sorbitol, raffinose, lactose, and sucrose carbohydrate fermentation tests; arginine deamination; acidification of methyl-α-d-glucopyranoside; pyruvate utilization; and isolate pigmentation.
Antimicrobial susceptibility testing.
Antimicrobial susceptibility testing was conducted using the minimal inhibitory concentration (MIC) agar dilution method [National Committee for Clinical Laboratory Standards (NCCLS) 2002]. E. faecalis 29212 was used as the quality control reference strain. Susceptibility to erythromycin, clindamycin, virginiamycin (streptogramin A and B combination), tetracycline, and vancomycin was tested. Erythromycin, clindamycin, tetracycline, and vancomycin were obtained from Sigma (St. Louis, MO). Virginiamycin was obtained from Research Products International Corp. (Mt. Prospect, IL). Concentrations of antibiotics tested ranged from 0.5 μg/mL to 256 μg/mL for erythromycin and tetracycline, 0.03 μg/mL to 128 μg/mL for clindamycin, 0.03 μg/mL to 32 μg/mL for virginiamycin, and 0.03 μg/mL to 64 μg/mL for vancomycin.
In preparation for the agar dilution tests, the air sample isolates, as well as the MIC reference strain E. faecalis 29212, were streaked from –80°C archived stocks onto tryptic soy agar No. 2 with 5% defibrinated sheep blood (QuadFive, Ryegate, MT) and incubated for 24 hr at 37°C. After 24 hr, each isolate was suspended in 3 mL Mueller-Hinton broth with a sterile cotton swab and adjusted to a 0.5 McFarland standard using a Vitek colorimeter (Hach, Loveland, CO). Two hundred microliters of each suspension was transferred to a well within a Cathra replicator plate (Oxoid Inc., Ogdensburg, NY) and replicated with 1-mm pins in accordance with NCCLS guidelines onto Mueller-Hinton agar plates that were previously prepared with the appropriate concentrations of antibiotics (NCCLS 2002). Plates were incubated for 24 hr at 37°C under aerobic conditions. After 24 hr, the plates were read manually and MICs were determined. Specifically, the MIC was recorded as the minimum antibiotic concentration that completely inhibited bacterial growth. According to the MIC, isolates were categorized as susceptible, intermediate, or resistant to each antibiotic using the following MIC breakpoints established by the NCCLS for Enterococcus: erythromycin, susceptible ≤ 0.5 μg/mL, intermediate 1–4 μg/mL, and resistant ≥ 8 μg/mL; clindamycin, susceptible ≤ 0.5 μg/mL, intermediate 1–2 μg/mL, and resistant ≥ 4 μg/mL; virginiamycin, susceptible ≤ 1 μg/mL, intermediate 2 μg/mL, and resistant ≥ 4 μg/mL; tetracycline, susceptible ≤ 4 μg/mL, intermediate 8 μg/mL, and resistant ≥ 16 μg/mL; and vancomycin, susceptible ≤ 4 μg/mL, intermediate 8–16 μg/mL, and resistant ≥ 32 μg/mL (NCCLS 2002).
Results
Bacterial concentrations in air and bacterial identification.
The mean concentration of presumptive Enterococcus present in the air of the swine CAFO on both 9 December 2003 and 5 January 2004 was 4 × 104 CFU/m3. After bacterial speciation was completed on 137 presumptive Enterococcus isolates, only 47 out of 137 isolates (34%) were confirmed to be Enterococcus (Table 1). Forty-four isolates (32%) were identified as staphylococci, 45 isolates (33%) were viridans group streptococci, and 1 isolate was identified as Micrococcus luteus (Table 1).
Antibiotic resistance.
Ninety-eight percent (121 of 124) of the bacterial isolates that grew successfully during the antimicrobial susceptibility tests were resistant to high levels of at least two antibiotics commonly used in swine production (erythromycin, clindamycin, virginiamycin, or tetracycline), and 93% of the isolates (115 of 124) were resistant to three antibiotics commonly used in swine production. Individually, 98% of the isolates were resistant to erythromycin, 94% were resistant to clindamycin, 90% were resistant to tetracycline, and 37% were resistant to virginiamycin. None of the isolates displayed resistance to vancomycin. Because none of the E. avium, E. pseudoavium, or E. raffinosus isolates (all belonging to the Enterococcus physiologic group I) grew successfully on the control or antibiotic-amended MIC plates after being suspended as 0.5 McFarland standard solutions, MIC data for these isolates were not determined. MIC distributions among all other isolates were similar for erythromycin, clindamycin, tetracycline, and vancomycin, regardless of bacterial genus or species (Tables 2 and 3). For instance, across all organisms, most isolates (96%) had MICs > 256 μg/mL for erythromycin (Tables 2 and 3). In contrast, resistance to virginiamycin was more prevalent among coagulase-negative staphylococci versus Enterococcus or Streptococcus isolates (Tables 2 and 3). Phenotypes of antibiotic resistance among the bacterial isolates appear in Table 4.
Discussion
In this study, multidrug-resistant Enterococcus, coagulase-negative staphylcocci, and viridans group streptococci were isolated from the air of a swine CAFO. Ninety-eight percent of the isolates were resistant to at least two of the following antibiotics: erythromycin, clindamycin, virginiamycin, and tetracycline, all of which are approved for use in swine production for growth promotion. In contrast, none of the isolates were resistant to vancomycin, which has never been approved for use in swine production in the United States. These results support the findings of previous reports that nontherapeutic use of antibiotics results in the presence of antibiotic-resistant bacteria in swine (Aarestrup et al. 2000a, 2000b; Bager et al. 1997; Jensen et al. 2002; Wegener et al. 1999). In addition, these results provide evidence that in the absence of nontherapeutic antibiotic use—vancomycin in this case—no resistance is detected among bacteria present in the swine environment.
Furthermore, these findings suggest that, in addition to the ingestion of retail pork products (Gambarotto et al. 2001; Hayes et al. 2003; Sorensen et al. 2001; White et al. 2001) and surface and groundwaters in the vicinity of swine CAFOs (Campagnolo et al. 2002; Chee-Sanford et al. 2001), the inhalation of air within swine operations may serve as another exposure pathway for the transfer of multidrug-resistant bacteria from swine to humans. These data are especially relevant to the health of swine CAFO workers, their direct contacts in the community, and possibly nearby neighbors of swine CAFOs.
The types of bacteria detected within the air of the swine facility investigated in this study are associated with a variety of human infections. Enterococcus, particularly some of the species isolated in this study including E. faecalis and E. faecium, has emerged as one of the leading causes of nosocomial bacteremias, urinary tract infections, and wound infections in the United States (NNIS 2001). Similarly, coagulase-negative staphylococci are the third most common causes of nosocomial infections and the most common causes of nosocomial bacteremias. The presence of multidrug-resistant Enterococcus and coagulase-negative staphylococci in patients significantly limits the treatment options available for these life-threatening infections. Although viridans group streptococci are part of the normal flora of the human respiratory tract, they also have been implicated as the cause of infective endocarditis and life-threatening septicemias in neutropenic patients. In addition, viridans group streptococci have been implicated as reservoirs of erythromycin-resistance genes, possibly capable of transferring resistance determinants to more pathogenic species including Streptococcus pneumoniae and Streptococcus pyogenes (Bryskier 2002).
Of particular concern to the health of individuals with direct or indirect contact with swine environments is the finding of virginiamycin-resistant gram-positive bacteria in the air of the swine CAFO. Virginiamycin, a streptogramin A and B combination, which has been used extensively as a growth promoter in swine, is an analog to quinupristin-dalfopristin, an injectable streptogramin A and B combination that is often the drug of last resort for multidrug-resistant gram-positive infections characterized by methicillin-resistant Staphylococcus aureus and glycopeptide-resistant E. faecium and coagulase-negative staphylococci (Johnson and Livermore 1999). Bacteria expressing resistance to virginiamycin are cross-resistant to quinupristin-dalfopristin, and a previous study has suggested that the transfer of streptogramin-resistant Enterococcus can occur between animals and humans in the livestock environment (Jensen et al. 1998). Thus, the inhalation of virginiamycin-resistant gram-positive bacteria in the swine environment could contribute to the appearance of quinupristin-dalfopristin–resistant gram-positive infections in humans, leaving few or no treatment options for the affected individual.
The finding of airborne clindamycin-resistant gram-positive bacteria in this study also is a potential concern to public health. Clindamycin is indicated for the treatment of human staphylococcal and streptococcal pneumonia (among other aerobic and anaerobic infections). Specifically, clindamycin has been used for the treatment of community-acquired methicillin-resistant S. aureus (Marcinak and Frank 2003). Clindamycin also has been shown to be significantly more potent than penicillin in inhibiting both invasive and noninvasive group A streptococci such as S. pyogenes (Mascini et al. 2001). The findings of airborne clindamycin-resistant coagulase-negative staphylococci and viridans group streptococci in the swine environment raise the question as to whether these organisms could serve as reservoirs of clindamycin-resistant genes [as well as reservoirs of erythromycin-resistant genes (Bryskier 2002)], passing on clindamycin resistance determinants to more pathogenic species as described above.
Furthermore, exposure to virginiamycin-, erythromycin-, clindamycin-, and tetracycline-resistant Enterococcus, coagulase-negative staphylococci, and viridans group streptococci through the inhalation of contaminated air could lead to the colonization of these multidrug-resistant organisms in both the nasal passages (Aubry-Damon 2004) and the lungs of swine CAFO workers, potentially making the workers themselves reservoirs of antibiotic-resistant organisms. Coexposures to other aerosols and gases in the swine environment such as organic dusts, molds, and ammonia have been shown to induce symptoms associated with chronic bronchitis, including a persistent cough characterized by expectoration (Mackiewicz 1998). The presence of this type of cough can increase the potential for secondary spread of antibiotic-resistant organisms into the community, where additional individuals could serve as reservoirs of multidrug-resistant bacteria.
Moreover, the tunnel-ventilated design of swine CAFOs, which moves air outside of the facilities at a high flow rate, could create a situation where neighbors living downwind of the ventilation fans also could be directly exposed to airborne multidrug-resistant bacteria. An epidemiologic study by Wing and Wolf (2000) indicated that people who live in the vicinity of swine CAFOs experience elevated rates of headaches, runny noses, sore throats, excessive coughing, and diarrhea compared with people living in communities that are not situated near livestock operations. The findings of airborne multidrug-resistant bacteria in a swine CAFO in our study raise the question as to whether airborne bacteria also could travel beyond the confines of the swine CAFO on ventilation fan air currents, directly contacting nearby neighbors and potentially contributing to health effects such as those observed in the Wing and Wolf study. Because populations living in areas where swine CAFOs are built already may experience higher rates of certain diseases because of lack of access to appropriate health care (Weber et al. 1989), investigating airborne exposures to multidrug-resistant bacteria among these at-risk populations is an important area for future research.
In addition to potential airborne exposures occurring among individuals living near swine CAFOs, the results of this study could have broader public health implications. Specifically, one may question whether airborne exposures to multidrug-resistant bacteria could be occurring and contributing to health problems around other environmental sources of animal or human waste, including land application areas for animal waste and human sludge, and human wastewater treatment facilities. Endotoxins, exotoxins, and other chemical components in dusts associated with animal waste and human sludge have been linked to hypersensitivity reactions among individuals living near land application areas (Lewis and Gattie 2002). These reactions have been shown to result in increased susceptibility to serious respiratory infections, including those caused by S. aureus (Lewis and Gattie 2002). Thus, the presence of high concentrations of multidrug-resistant staphylococci and other bacterial pathogens amidst endotoxin-containing dust from animal and human waste could pose unique health concerns to people living near land application areas.
Conclusions
In summation, the findings of this study suggest that the inhalation of air from swine CAFOs may serve as an additional environmental exposure pathway for the transfer of multidrug-resistant bacterial pathogens from swine to humans. Given the growing interest in reservoirs of antibiotic resistance genes associated with large-scale livestock operations (Nandi et al. 2004), our findings in this investigation emphasize the importance of studying multiple genera of bacteria in different environmental media as sources of human exposure to antibiotic resistance genes.
Table 1 Airborne bacteria isolated from a swine CAFO using methods for the isolation of Enterococcus species.
Bacteria No. of isolates (%)
Enterococcus 47 (34)
E. avium 5 (4)
E. dispar 4 (3)
E. durans 2 (1)
E. faecalis 6 (4)
E. faecium 1 (< 1)
E. hirae 14 (10)
E. mundtii 1 (< 1)
E. pseudoavium 2 (1)
E. raffinosus 1 (< 1)
Other 11 (8)
Staphylococcus 44 (32)
S. aureus 1 (< 1)
Coagulase-negative staphylococci 43 (31)
Streptococcus
Viridans group streptococci 45 (33)
Micrococcus luteus 1 (< 1)
Total 137 (100)
Table 2 MIC distributions for five antibiotics observed in airborne Enterococcus collected from a swine CAFO.
No. of bacterial isolates with the following MICs (μg/mL)
Bacteria, antibiotic ≤ 0.5 1 2 4 8 16 32 64 128 256 > 256 %S %I %R
Enterococcus (n = 38)
Erythromycin 1 37 0 3 97
Clindamycin 1 1 1 1 3 8 23a 3 0 97
Virginiamycin 19 5 5 9 63 13 24
Tetracycline 1 2 7 6 17 5 3 5 92
Vancomycin 38 100 0 0
E. dispar (n = 4)
Erythromycin 4 0 0 100
Clindamycin 1 1 2a 0 0 100
Virginiamycin 4 100 0 0
Tetracycline 3 1 0 0 100
Vancomycin 4 100 0 0
E. durans (n = 2)
Erythromycin 2 0 0 100
Clindamycin 1 1a 0 0 100
Virginiamycin 1 1 0 50 50
Tetracycline 1 1 50 50 0
Vancomycin 2 100 0 0
E. faecalis (n = 6)
Erythromycin 1 5 17 0 83
Clindamycin 1 1 2 2a 17 0 83
Virginiamycin 2 3 1 83 0 17
Tetracycline 2 1 2 1 0 0 100
Vancomycin 6 100 0 0
E. faecium (n = 1)
Erythromycin 1 0 0 100
Clindamycin 1 0 0 100
Virginiamycin 1 0 0 100
Tetracycline 1 0 0 100
Vancomycin 1 100 0 0
E. hirae (n = 14)
Erythromycin 14 0 0 100
Clindamycin 2 12a 0 0 100
Virginiamycin 8 2 4 57 14 29
Tetracycline 1 2 1 8 2 0 7 93
Vancomycin 14 100 0 0
Other (n = 11)
Erythromycin 11 0 0 100
Clindamycin 1 1 2 1 6a 0 0 100
Virginiamycin 5 2 2 2 64 18 18
Tetracycline 2 4 4 1 0 0 100
Vancomycin 11 100 0 0
Abbreviations: %I, percent intermediate; %R, percent resistant; %S, percent susceptible.
a MIC is > 128 μg/mL.
Table 3 MIC distributions for five antibiotics observed in airborne Staphylococcus and Streptococcus collected from a swine CAFO.
Number of bacterial isolates with the following MICs (μg/mL)
Bacteria, antibiotic ≤ 0.5 1 2 4 8 16 32 64 128 256 > 256 %S %I %R
Staphylococcus (n = 43)a
S. aureus (n = 1)
Erythromycin 1 0 0 100
Clindamycin 1b 0 0 100
Virginiamycin 1 0 100 0
Tetracycline 1 0 0 100
Vancomycin 1 100 0 0
Coagulase-negative
staphylococci (n = 42)
Erythromycin 42 0 0 100
Clindamycin 1 2 39b 0 2 98
Virginiamycin 4 2 2 21 13 14 5 81
Tetracycline 2 1 1 5 12 13 7 1 7 0 93
Vancomycin 7 3 30 2 100 0 0
Streptococcus (n = 43)c
Viridans group streptococci
Erythromycin 1 1 1 40 0 0 100
Clindamycin 2 2 2 9 28 5 0 95
Virginiamycin 29 7 4 3 84 9 7
Tetracycline 1 8 17 10 7 2 0 98
Vancomycin 43 100 0 0
Abbreviations: %I, percent intermediate; %R, percent resistant; %S, percent susceptible.
a Analyzed using the breakpoints for Enterococcus.
b MIC is > 128 μg/mL.
c Analyzed using the following breakpoints: erythromycin, susceptible ≤ 0.25 μg/mL, intermediate 0.5 μg/mL, and resistant ≥ 1.0 μg/mL; clindamycin, susceptible ≤ 0.5 μg/mL, intermediate 1–2 μg/mL, and resistant ≥ 4 μg/mL; virginiamycin, susceptible ≤ 1 μg/mL, intermediate 2 μg/mL, and resistant ≥ 4 μg/mL; tetracycline, susceptible ≤ 2 μg/mL, intermediate 4 μg/mL, and resistant ≥ 8 μg/mL; vancomycin, susceptible ≤ 1 μg/mL, intermediate and resistant not available (NCCLS 2002).
Table 4 Phenotypes of antibiotic resistance among airborne bacteria collected from a swine CAFO.
Bacteria Antibiotic resistance pattern No. of isolates (%)
Enterococcus
E. dispar (n = 4) Ery, Clin, Tet 4 (100)
E. durans (n = 2) Ery, Clin 1 (50)
Ery, Clin, Virg 1 (50)
E. faecalis (n = 6) Tet 1 (17)
Ery, Clin, Tet 4 (66)
Ery, Clin, Tet, Virg 1 (17)
E. faecium (n = 1) Ery, Clin, Tet, Virg 1 (100)
E. hirae (n = 14) Ery, Clin 1 (7)
Ery, Clin, Tet 9 (64)
Ery, Clin, Tet, Virg 4 (29)
Other Enterococcus (n = 11) Ery, Clin, Tet 9 (82)
Ery, Clin, Tet, Virg 2 (18)
Staphylococcus aureus (n = 1) Ery, Clin, Tet 1 (100)
Coagulase-negative staphylococci (n = 42) Ery, Tet 1 (2)
Ery, Clin, Tet 8 (19)
Ery, Clin, Virg 6 (14)
Ery, Virg, Tet 1 (2)
Ery, Clin, Tet, Virg 26 (62)
Viridans group streptococci (n = 43) Tet 2 (5)
Ery, Clin 1 (2)
Ery, Tet 2 (5)
Ery, Clin, Tet 35 (81)
Ery, Clin, Tet, Virg 3 (7)
Abbreviations: Clin, clindamycin; Ery, erythromycin; Tet, tetracycline; Virg, virginiamycin.
==== Refs
References
Aarestrup FM Agerso Y Gerner-Smidt P Madsen M Jensen LB 2000a Comparison of antimicrobial resistance phenotypes and resistance genes in Enterococcus faecalis and Enterococcus faecium from humans in the community, broilers, and pigs in Denmark Diagn Microbiol Infect Dis 37 127 137 10863107
Aarestrup FM Kruse H Tast E Hammerum AM Jensen LB 2000b Associations between the use of antimicrobial agents for growth promotion and the occurrence of resistance among Enterococcus faecium from broilers and pigs in Denmark, Finland, and Norway Microb Drug Resist 6 63 70 10868809
Aubry-Damon H Grenet K Sall-Ndiaye P Che D Cordeiro E Bougnoux ME 2004 Antimicrobial resistance in commensal flora of pig farmers Emerg Infect Dis 10 873 879 15200822
Bager F Madsen M Christensen J Aarestrup FM 1997 Avoparcin used as a growth promoter is associated with the occurrence of vancomycin-resistant Enterococcus faecium on Danish poultry and pig farms Prev Vet Med 31 95 112 9234429
Bryskier A 2002 Viridans group streptococci: a reservoir of resistant bacteria in oral cavities Clin Microbiol Infect 8 65 69 11952717
Campagnolo ER Johnson KR Karpati A Rubin CS Kolpin DW Meyer MT 2002 Antimicrobial residues in animal waste and water resources proximal to large-scale swine and poultry feeding operations Sci Total Environ 299 89 95 12462576
Chee-Sanford JC Aminov RI Krapac IJ Garrigues-Jeanjean N Mackie RI 2001 Occurrence and diversity of tetracycline resistance genes in lagoons and groundwater underlying two swine production facilities Appl Environ Microbiol 67 1494 1502 11282596
Clark S Rylander R Larsson L 1983 Airborne bacteria, endotoxin and fungi in dust in poultry and swine confinement buildings Am Ind Hyg Assoc J 44 537 541 6613856
Cohen ML Tauxe RV 1986 Drug-resistant Salmonella in the United States: an epidemiologic perspective Science 234 964 969 3535069
Cormier Y Tremblay G Meriaux A Brochu G Lavoie J 1990 Airborne microbial contents in two types of swine confinement buildings in Quebec Am Ind Hyg Assoc J 51 304 309 2353639
Crook B Robertson JF Glass SA Botheroyd EM Lacey J Topping MD 1991 Airborne dust, ammonia, microorganisms, and antigens in pig confinement houses and the respiratory health of exposed farm workers Am Ind Hyg Assoc J 52 271 279 1951065
Donabedian SM Thal LA Hershberger E Perri MB Chow JW Bartlett P 2003 Molecular characterization of gentamicin-resistant Enterococci in the United States: evidence of spread from animals to humans through food J Clin Microbiol 41 1109 1113 12624037
Environmental Defense 1997. Animal Waste Summary. New York:Environmental Defense. Available: http://www.hogwatch.org/maps/index [accessed 15 June 2004].
Food and Drug Administration 2004. FDA Approved Animal Drug Products. Blacksburg, VA:Drug Information Laboratory, Virginia/Maryland Regional College of Veterinary Medicine.
Gambarotto K Ploy MC Dupron F Giangiobbe M Denis F 2001 Occurrence of vancomycin-resistant enterococci in pork and poultry products from a cattle-rearing area of France J Clin Microbiol 39 2354 2355 11376091
Gorbach SL 2001 Antimicrobial use in animal feed—time to stop N Engl J Med 345 1202 1203 11642237
Hamscher G Pawelzick HT Sczesny S Nau H Hartung J 2003 Antibiotics in dust originating from a pig-fattening farm: a new source of health hazard for farmers? Environ Health Perspect 111 1590 1594 14527837
Hayes JR English LL Carter PJ Proescholdt T Lee KY Wagner DD 2003 Prevalence and antimicrobial resistance of enterococcus species isolated from retail meats Appl Environ Microbiol 69 7153 7160 14660361
Institute of Medicine 1998. Antimicrobial Resistance: Issues and Options, Workshop Report, Forum on Emerging Infections. Washington, DC:National Academy Press.
Jensen LB Hammerum AM Aarestrup FM van den Bogaard AE Stobberingh EE 1998 Occurrence of satA and vgb genes in streptogramin-resistant Enterococcus faecium isolates of animal and human origins in the Netherlands Antimicrob Agents Chemother 42 3330 3331 10049241
Jensen LB Hammerum AM Bager F Aarestrup FM 2002 Streptogramin resistance among Enterococcus faecium isolated from production animals in Denmark in 1997 Microb Drug Resist 8 369 374 12523635
Johnson AP Livermore DM 1999 Quinupristin/dalfopristin, a new addition to the antimicrobial arsenal Lancet 354 2012 2013 10636359
Lewis DL Gattie DK 2002 Pathogen risks from applying sewage sludge to land Environ Sci Technol 36 286A 293A
Lin X Willeke K Ulevicius V Grinshpun S 1997 Effect of sampling time on the collection efficiency of all-glass impingers Am Ind Hyg Assoc J 58 480 488
Mackiewicz B 1998 Study on exposure of pig farm workers to bioaerosols, immunologic reactivity and health effects Ann Agric Environ Med 5 169 175 9860819
Marcinak JF Frank AL 2003 Treatment of community-acquired methicillin-resistant Staphylococcus aureus in children Curr Opin Infect Dis 16 265 269 12821819
Mascini EM Janszea M Schoulsb LM Verhoefa J Van Dijk H 2001 Penicillin and clindamycin differentially inhibit the production of pyrogenic exotoxins A and B by group A streptococci Int J Antimicrob Agents 18 395 398 11691576
Mellon M Benbrook C Benbrook KL 2001. Hogging It: Estimates of Antimicrobial Abuse in Livestock. Cambridge, MA:Union of Concerned Scientists Publications.
Murray PR Baron EJ Jorgensen JH Pfaller MA Yolken RH 2003. Manual of Clinical Microbiology. 8th ed. Washington, DC:American Society for Microbiology Press.
Nandi S Maurer JJ Hofacre C Summers AO 2004 Gram-positive bacteria are a major reservoir of class 1 antibiotic resistance integrons in poultry litter Proc Natl Acad Sci USA 101 7118 7122 15107498
National Research Council 1999. The Use of Drugs in Food Animals: Benefits and Risks. Washington, DC:National Academy Press.
NCCLS 2002. Performance Standards for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolates from Animals; Approved Standard. 2nd ed. M31-A2. Wayne, PA:National Committee for Clinical Laboratory Standards.
NNIS 2001 National Nosocomial Infections Surveillance (NNIS) system report, data summary from January 1992–June 2001 Am J Infect Control 29 404 421 11743489
Predicala BZ Urban JE Maghirang RG Jerez SB Goodband RD 2002 Assessment of bioaerosols in swine barns by filtration and impaction Curr Microbiol 44 136 140 11815859
Sorensen TL Blom M Monnet DL Frimodt-Moller N Poulsen RL Espersen F 2001 Transient intestinal carriage after ingestion of antibiotic-resistant Enterococcus faecium from chicken and pork N Engl J Med 345 1161 1166 11642232
U.S. EPA 2000. Improved Enumeration Methods for the Recreational Water Quality Indicators: Enterococci, and Escherichia coli EPA/821/R-97/004. Washington, DC:U.S. Environmental Protection Agency.
van den Boogard AE Stobberingh EE 1999 Antibiotic usage in animals: impact on bacterial resistance and public health Drugs 58 589 607 10551432
Weber D Rutala W Samsa G Sarubbi F King L 1989 Epidemiology of tuberculosis in North Carolina, 1966 to 1986: analysis of demographic features, geographic variation, AIDS, migrant workers, and site of infection South Med J 92 1204 1214 2799438
Wegener HC 2003 Antibiotics in animal feed and their role in resistance development Curr Opin Microbiol 6 439 445 14572534
Wegener HC Aarestrup FM Jensen LB Hammerum AM Bager F 1999 Use of antimicrobial growth promoters in food animals and Enterococcus faecium resistance to therapeutic antimicrobial drugs in Europe Emerg Infect Dis 5 329 335 10341169
White DG Zhao S Sudler R Ayers S Friedman S Chen S 2001 The isolation of antibiotic-resistant salmonella from retail ground meats N Engl J Med 345 1147 1154 11642230
Wing S Wolf S 2000 Intensive livestock operations, health, and quality of life among eastern North Carolina residents Environ Health Perspect 108 233 238 10706529
Zahn JA Anhalt J Boyd E 2001 Evidence for transfer of tylosin and tylosin-resistant bacteria in air from swine production facilities using sub-therapeutic concentrations of tylan in feed [Abstract] J Anim Sci 79 189
|
15687049
|
PMC1277855
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):137-142
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7473
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7467ehp0113-00014315687050ResearchArticlesTopophilia and the Quality of Life Ogunseitan Oladele A. Department of Environmental Health, Science, and Policy, University of California, Irvine, California, USAAddress correspondence to O.A. Ogunseitan, Department of Environmental Health, Science, and Policy, University of California, Irvine, CA 92697-7070 USA. Telephone: (949) 824-6350. Fax: (949) 824-2056. E-mail:
[email protected] am grateful for the assistance of A. Castro, C. Yi, S. Ly, and M. Poulin. The research benefited from the expertise and insight of E.A. Holman and from the inspirational work of M. Lin in the nexus of architecture and the human experience.
This project was funded by grants from the Claire Trevor School of the Arts and the Program in Industrial Ecology at the University of California at Irvine.
The author declares he has no competing financial interests.
2 2005 22 11 2004 113 2 143 148 30 7 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. With this research I tested the hypothesis that individual preferences for specific ecosystem components and restorative environments are significantly associated with quality of life (QOL). A total of 379 human subjects responded to a structured 18-item questionnaire on topophilia and to the 26-item World Health Organization’s Quality of Life (WHOQOL-Bref) instrument. Confirmatory factor analyses revealed four domains of topophilia (ecodiversity, synesthetic tendency, cognitive challenge, and familiarity) and four domains of QOL (physical, psychological, social, and environmental). Synesthetic tendency was the strongest domain of topophilia, whereas the psychological aspect of QOL was the strongest. Structural equation modeling was used to explore the adequacy of a theoretical model linking topophilia and QOL. The model fit the data extremely well: χ2 = 5.02, p = 0.414; correlation = 0.12 (p = 0.047). All four domains of topophilia were significantly correlated with the level of restoration experienced by respondents at their current domicile [for cognitive challenge: r = 0.19; p < 0.01; familiarity: r = 0.12; p < 0.05; synesthetic tendency: r = 0.18; p < 0.01; ecodiversity (the highest value): r = 0.28; p < 0.01]. Within ecodiversity, preferences for water and flowers were associated with high overall QOL (r = 0.162 and 0.105, respectively; p < 0.01 and 0.05, respectively). Within the familiarity domain, identifiability was associated with the environmental domain of QOL (r = 0.115; p < 0.05), but not with overall QOL. These results provide a new methodologic framework for linking environmental quality and human health and for implementing evidence-based provision of restorative environments through targeted design of built environments to enhance human QOL.
ecosystemsmental healthnaturequality of liferestorative environmentsstresstopophilia
==== Body
Mental stress is an underestimated and growing component of disease burden in many parts of the world (Saxena et al. 2003). The roles of both natural and constructed environments in relieving mental stress have long been suspected but are poorly understood. Tuan (1999) originally defined topophilia as the affective bond between people and place or environmental setting. Topophilia is presumed to be a vivid and personal experience, but research is scarce on the determinants of individual preferences and on the potential health benefits derived from such experiences. The few existing studies have not adequately deconstructed the confounding of affective and cognitive processes in aesthetic response versus tangible health outcomes (e.g., Parsons 1991; Ulrich 1979). Furthermore, quantitative assessments of the values associated with the considerable financial investment by societies in naturalistic environmental design, landscape architecture, and ecosystem conservation through wildland natural preserves are rare. When available, the results of such studies are often inconclusive or contradictory. The proximate causes of particular topophilia are embedded in measurable characteristics: environmental perception, defined as the response of senses to external stimuli and purposeful activity; attitude, or ingrained cultural stances; and values, the rank-ordered conception of preferences that emerge following a personalized exercise in trade-offs among alternative scenarios. Environmental designers have long exploited the basic ideas of topophilia to create presumably attractive surroundings that restore mental health based on the use of materials, sensory stimuli, and arrangements that remind people of the place and environmental settings that are comforting and/or associated with healing potential (Carlson 2000; Porteous 1996).
In the tradition of environmental psychology, “restorative environments” are defined as specific geographical contexts that renew diminished functional capabilities and enhance coping strategies and resources for managing stress (Hartig and Staats 2003). There is also general consensus that measuring restoration according to this definition is complicated. In urban cultures where restorative environments are conventionally linked to few and remote vestiges of forest wilderness or pristine water views, it is increasingly important to understand the role of landscape design and public art in providing sanctuaries where a sense of balance can be restored to hectic lifestyles. However, parameters such as age, sex, ethnic background, and socioeconomic status have powerful influences on individual and group perception of restorative environments as defined by artificial public spaces in confined urban centers (Hartig et al. 2003; Laumann et al. 2003). In this research project I sought to identify common features of preferred restorative environments in a sample population according to the categories typically associated with topophilia: synesthetic tendency (commingling of sensory stimuli and the memory of place), environmental familiarity, cognitive challenge, and ecodiversity (Janzen 1998; Tuan 1993, 1999). The specification of topophilic preferences is potentially more informative if the preferences are linked to tangible benefits for human health and welfare. In this regard, the literature on restorative environments has lacked a quantitative measure of restoration, although there have been some preliminary empirical excursions into the putative linkages between individual environment preferences and restoration (Staats et al. 1997, 2003; van den Berg et al. 2003). Those studies typically request that stressed subjects declare their preference (e.g., which is more “beautiful”?) for either a forested landscape or the concrete world of an urban downtown. Hence, the fine-level characteristics of built versus natural ecosystems have not been adequately captured (Ulrich 1993).
In the present study, I assessed the mental health profile of respondents to the topophilia survey by means of the World Health Organization’s quality of life survey instrument (WHOQOL-Bref). The WHO defines quality of life (QOL) as “an individual’s perception of their position in life in the context of the culture and value systems in which they live, and in relation to their goals, expectations, standards and concerns” (WHOQOL Group 1994, 1995, 1998a, 1998b). Both the 100-question (WHOQOL-100) and 26-question (WHOQOL-Bref) versions have been validated across many cultures, in several countries, and for different contexts of health, well-being, and occupational stress (Kuyken et al. 1994; Nasermoaddeli et al. 2003; Skevington et al. 2004; Weatherall et al. 2004). WHOQOL-Bref is recommended when it is advisable to minimize the time burden on respondents. Furthermore, it has been shown to have excellent psychometric properties of reliability, and it performed well in tests of validity across the four domains of health, namely, physical health, psychological well-being, social relationships, and environmental support (Saxena et al. 2001).
As a seminal exploration of the linkages between QOL and preferred environmental and ecosystem features, the present study explicitly posed the hypothesis that those exhibiting high QOL are more likely to describe their domicile as providing access to restorative environments defined by specific components of the landscape. I further hypothesized that preference for the specific topophilia domain of ecodiversity is associated with high QOL. The hypotheses were tested by statistical analyses of responses to structured questionnaires. The population sampled for this study identified preferences for water bodies, flowers, and spatial familiarity restoration.
The results provide insight into specific aspects of ecosystems and artificial landscapes that are more likely to support restoration and the enhancement of QOL. Importantly, the set of methods developed here provides a strategy for future investigations addressing the response of diverse populations in different urban environments to various aspects of natural and artificial topography.
Materials and Methods
Human subject pool.
Human subjects for the study were recruited at the University of California at Irvine between August 2001 and August 2002. For most respondents, the campus represented both residential and work environments during periods of concentrated academic activity. It is partly for this reason that construction and landscape developmental plans for many campuses recognize the need to provide oases for recreation, reflection, and mental restoration. However, there has never been a systematic study of preferences for landscape design relative to the level of restoration experienced after study-induced stress or fatigue. Respondents were recruited from well-visited locations across campus, including library, bookstore, restaurant, and athletic fields. The average amount of time required for completion of the questionnaire was 15 min. The recruitment material was approved by the institutional review board for research on human subjects at the University of California, Irvine.
A cover letter introduced the research project and informed potential respondents that participation is voluntary and confidentiality is assured throughout the entire process. Each survey was denoted by a numerical identifier. Self-reported information was collected on baseline characteristics such as sex, age, level of education attained, marital status, and ethnic background. Information was also collected on the location of permanent domicile and on the length of time that respondents have spent living and working or schooling at the specific campus.
Topophilia rating.
Restorative environment is used in the context of this study to mean a place associated with relief from mental stress or fatigue. There are few standardized quantitative measures of the specific components of restorative environments (Laumann et al. 2001). In this study, composite measures for environmental perception and preferences for specific ecosystem components and landscape design were integrated in an 18-item questionnaire (questionnaire items 8–25 in Table 1). This measure of topophilia was developed using the theoretical foundations provided by the work of Tuan and existing theories of restorative environments (Betrabet 1996; Hartig and Evans 1993; Hertzog et al. 2003; Tuan 1999). Respondents rated their preferences for specific categories of ecosystem components and environmental and landscape design characteristics on a scale of 1 to 10, with 10 being most effective toward respondent’s expectation of restoration experience. This set of questions focused on the level of importance that respondents accorded to ecosystem components regardless of whether or not they have current access or they expect to actually experience the benefits of exposure to the items being rated. For example, the questions in this category were framed as follows: “Rate the following characteristics (or sensory qualities/ecological components) of an environment according to your expectation of how effective they will be in making you feel refreshed or experience restoration, on a scale of 1–10 with 10 being most effective.”
Confirmatory factor analyses identified four specific domains underlying topophilia: cognitive challenge (e.g., complexity and coherence), synesthetic tendency (e.g., colors and sounds), ecodiversity (e.g., water bodies and trees), and familiarity (e.g., identifiability and privacy). For these domains, statistical factor loadings all exceeded 0.60, and Cronbach α-values ranged from 0.68 to 0.87 (Table 2). The last question in the section on topophilia ratings asked respondents to actually rate the campus according to the number and kinds of accessible restorative environments, on a scale of 1 to 10 with 10 representing saturation (i.e., all subcategories within topophilia are accessible). This question addressed the extent to which various environmental elements were not only present, but also provided satisfying restorative effects in respondents’ current environment. The question was phrased as follows: “On a scale of 1 to 10, rate your current home environment according to the abundance and variety of restorative environments that are accessible to you.”
Assessing QOL.
I used the brief version of the WHO’s QOL survey instrument (WHOQOL-Bref) in this study to assess the QOL of respondents according to the four minor domains of physical health (seven categorical items), psychological welfare (six items), social relationships (three items), and environmental support (eight items). The four minor domains were statistically modeled to produce an overall score for the QOL for each respondent. The reliability of the associations between the observed variables and the latent domain of QOL was excellent, according to the consistently high Cronbach α-values computed for the models (Table 2). WHOQOL-Bref instrument was used with permission from the WHO (Üstün TB, personal correspondence). A syntax file for checking the data and computing domain scores was obtained from M. Power (University of Edinburgh, Scotland). The WHOQOL-Bref scores were created and interpreted exactly as specified by the WHOQOL Group ( 1994, 1995, 1998a, 1998b). Factor loadings for all four domains exceeded 0.6, and Cronbach α-values ranged from 0.71 to 0.77 (Table 2).
Statistical analyses.
Descriptive statistics, correlation coefficients, and regression analyses were conducted using SPSS statistical software (version 12.0; SPSS Inc., Chicago, IL). Structural equation modeling to identify relationships among the domains of topophilia and QOL was conducted using Amos software (version 5.0; SPSS, Inc.).
Results
Human subjects.
Table 1 shows the descriptive statistics and general properties of the sample population. A total of 379 respondents completed the questionnaire. The average age of respondents was 23 years, ranging from 17 to 60 years. Females represented 58% of the sample population. The sample was ethnically diverse, but of those who registered their ethnicity, more respondents (24%) claimed Asian ethnicity than others (17% Caucasian, 4% Hispanic, 3% African American, and 4% mixed ethnicity). The majority (88%) of the sample population reported being single (9% married; 3% divorced or separated). Most respondents (79%) were pursuing undergraduate degree programs, and a large majority (88%) reported themselves to be healthy at the time of the survey.
Statistical model.
It was important to first determine whether the responses to questions posed to assess topophilia clustered together in easily recognizable groups. Indeed, confirmatory factor analyses demonstrated four domains underlying topophilia: ecodiversity (questionnaire items 20–25 in Table 1), synesthetic tendency (items 15–19), cognitive challenge (items 8–11), and familiarity (items 12–14). Structural equation modeling showed that all four domains loaded onto the latent construct of topophilia. The strongest domain was synesthetic tendency (0.84), and the weakest domain was cognitive challenge (0.37) (Figure 1).
Four major domains of human experience are also generally recognized to contribute to human self-reporting of QOL. Figure 1 shows the results of confirmatory factor analyses demonstrating that the four recognized domains of WHOQOL-Bref (i.e., physical health, psychological well-being, social relationships, and environmental support) also loaded highly on the underlying latent construct of QOL. These factor loadings are comparable with those identified in an international population sample by the WHOQOL Group ( 1994, 1995, 1998a, 1998b). The strongest domain was psychological well-being (0.81), and the weakest domain was social relationships (0.66).
I also used structural equation modeling to test the relationship between the latent variable of topophilia and the overall QOL scores based on WHOQOL-Bref. The statistical model showed extremely good fit with the data, linking observed overall QOL score and the latent variable of topophilia that was derived from all the four major domains: χ2 (df = 5, n = 379) = 5.02 (p = 0.414). The correlation between topophilia and QOL score is 0.12 (p = 0.047) (Figure 1). The smallest loading factor among the four underlying determinants of topophilia was 0.37 for the domain of cognitive challenge. Therefore, I tested a new model without the cognitive challenge domain, and the fit between the data and model improved slightly: χ2 (df = 2, n = 379) = 1.84 (p = 0.398). For this new model, the correlation between topophilia and QOL remained at 0.12 (p = 0.040). Therefore, I judged the model with all four domains of topophilia to be the best model, although further research is warranted to improve the factor loading for the cognitive challenge domain, which currently includes questions on complexity, mystery, coherence, and texture.
Variance and correlations among the domains of topophilia and QOL. Topophilia.
On a scale of 1 to 10, with 10 being the most effective in supporting a restorative experience, the mean rating of topophilia subcategories ranged from the lowest observed value of 4.75 (SD = 2.67) for complexity to the highest observed value of 7.90 (SD = 2.32) for the presence of trees (Table 1). The mean (±SD) rating of restoration opportunities attributed to respondents’ location was 7.1 ± 1.9, also on a scale of 1–10, with 10 being the most saturated with opportunities for experiencing restoration.
Quality of life.
Most respondents ranked their QOL very highly (mean ~ 3.98 ± 0.81; on a scale of 1–5, with 5 being the highest QOL). Similarly, most respondents were satisfied with their health status (mean = 3.69 ± 0.89). Respondents mostly felt that their lives are meaningful (mean = 3.76 ± 0.96), and most enjoyed a healthy physical environment (mean = 3.54 ± 0.83). The computed scores for the four domains of QOL were reasonably high, consistent with scores observed by WHOQOL Group ( 1994, 1995, 1998a, 1998b) for healthy international populations. The computed score for the physical health domain was the highest (mean = 15.22 ± 2.24; on a scale of 1–20, with 20 being the highest). The lowest domain score was for the environment domain (mean = 14.38 ± 2.33). The overall QOL computed from the domain scores was also high (mean = 14.69 ± 2.11) (Table 1).
Correlations.
Table 3 shows the correlation matrix between the domains of topophilia and the domains of QOL. The data show that only the “ecodiversity” category of topophilia was significantly correlated to the overall QOL (r = 0.123; p < 0.05), and within this category, the presence of flowers (r = 0.162; p < 0.01) and proximity to lakes/ocean (r = 0.129; p < 0.05) were significantly correlated with the overall QOL. All the major categories of topophilia were significantly correlated with the rating of opportunities for restoration at the current domicile of the respondents, but the domain of familiarity was significant only at the p = 0.05 level, whereas cognitive challenge, synesthetic tendency, and ecodiversity were significant at the p = 0.01 level (Table 3).
Discussion
What are the tangible health benefits to its citizens of society’s investment in ecologic conservation, environmental design, and expensive landscape architecture? There is near universal agreement that these investments are justifiable, but until now there have been no straightforward methodologies for providing quantitative answers to this question because of the widely acknowledged variations in individual preferences and valuation of environmental quality across regional, national, political, and cultural boundaries. This study linked, for the first time, a standardized globally validated measure of human QOL with the indicators of human preferences for ecosystem attributes that have been associated with restorative environments. In addition to providing this linkage, the results of this study also suggest a quantitative strategy for proactive assessment of user preferences for specific landscape features before the implementation of environmental design initiatives aimed at enhancing public health and welfare.
This study was conducted primarily among an educated youthful population sample inhabiting a societal microcosm. This is considered an important strength of the study in the sense that both the population and site are supported by considerable societal economic expenditure as an investment in future generations. However, appropriate caution is warranted before the data can be extrapolated to major urban centers—for example, in the construction of large parks for populations having lower levels of education, different ethnic composition, or different kinds of stressors. That said, it is important to note that the WHOQOL-Bref model scores observed in this study are not significantly different from those measured for healthy populations in most parts of the world (Saxena et al. 2001; Skevington et al. 2004) (Figure 1).
This study yielded two major findings: a) The overall QOL score is significantly associated with high rating of topophilia, and b) environmental and landscape design strategies associated with cognitive challenge—complexity, coherence, and the use of textural stimulation—are less effective in creating impressions of environmental restoration, whereas ecologic designs using ecodiversity themes—particularly the presence of flowers, lakes, or oceans—are generally perceived as providing restorative environments. The implications of these two major findings are discussed in the following sections.
Linkage of topophilia, restoration, and QOL.
The major finding of this study is that a statistically valid model explicitly connects a standardized measure of the overall QOL scores with the latent construct of topophilia (correlation = 0.12; p = 0.047). Furthermore, all the factor loadings from the four precisely defined domains (ecodiversity, synesthetic tendency, environmental familiarity, and cognitive challenge) were significant, and the reliability according to Cronbach α-values was very good for the latent construct of topophilia (Table 2). These findings provide a strong tool for studies attempting to bridge the current epistemologic gap between personal preferences for environmental or ecologic resources and mental health. There is a long history of research on the theoretical underpinnings of the specific identities of person–environment interactions that enhance the restorative experience (Betrabet 1996; Kaplan 1995, 2001; King et al. 2002; Korpela et al. 2001). However, empirical validations of these theoretical constructs are rare. Among the dominant theories of restorative environments is attention restoration theory (ART), which posits that intensive or prolonged use of directed attention leads to fatigue of the mechanisms that serve it, and that the recovery of effective functioning (restoration) is enabled by experience of certain components of a restorative environment (Hertzog et al. 2003; Kaplan 1995). ART is particularly relevant to populations encamped in densely populated geographical locations with the fatigue-prone occupations. According to ART, restorative environments are characterized by four features: “being away,” “extent,” “fascination,” and “compatibility” (Hertzog et al. 2003). The topophilia domains used in the present study differ substantially from ART features, although there are overlaps. For example, certain aspects of “being away” and “compatibility” are captured by the “environmental familiarity” index used in this study. Similarly, the “extent” feature of ART is most similar to the “cognitive challenge” category, whereas the “fascination” feature of ART is most similar to the “synesthetic tendency” construct used here. Perhaps the most salient advantage of the strategy used here is the explicit presentation of “ecodiversity” as a category. In ART, the main focus is to explain why people prefer natural environments to artificial (built) environments. This limitation has prevented empirical analysis of just what part of nature people find extensive, fascinating, or compatible. The finding of the present research eliminates this limitation and provides a solid context for further empirical testing of the determinants of restorative environments.
Ecodiversity themes are paramount in the environmental restoration experience.
The results of this research further buttress previous findings that when presented with opportunities for restoration, people rank proximity to natural/wildlife environments higher than landscape or urban constructions that overemphasize complex designs or artificial sensory stimulation, although these latter criteria can also contribute to the overall restoration experience. Specifically, the presence of flowers and water bodies are identified in this study as major factors that are associated with QOL and the experience of restorative environments. This level of pinpointing has been previously difficult to establish because most research on environmental preferences have relied on composite measures of “nature,” such as photographs of forests or nature hikes (Hartig et al. 1994). Specifically, van den Berg et al. (2003) noted that the absence of mediational analyses in past research has led to inadequate evidence for the intricacies of the theoretically sound and empirically supported line of reasoning that people typically demonstrate a fondness for nature more than the built environment. The functional accounting of environmental preferences suggests that individuals are attracted to environments that provide tangible benefits to health and that the level of attraction depends on the baseline of measurable health status (Hertzog et al. 2003). To use a pertinent metaphor, drivers whose automobiles rarely run out of gas are also more likely to pay attention to their fuel gauges and to know the locations of the best refueling stations, being picky about the cost of fuel and brand name of each station. That is, they are more likely to indulge in preferential rating of refueling stations than drivers who are stressed and less attentive. To bring this metaphor home to the present study, those who maintain a high QOL are also more likely to rank high on topophilia and to more clearly identify those components of the environment that afford high levels of regular restoration.
In addition to pointing out the positive associations between specific components of ecodiversity and mental health, it is also noteworthy to emphasize the surprising finding that none of the components of the synesthetic sensory stimuli category showed strong statistical association with QOL. So, for example, the anecdotal linkages that have been made in the academic literature, and even in commercial enterprises regarding the health benefits of listening to sounds associated with wildlife and natural settings (e.g., ocean waves, wind-rustled leaves, cricket sounds), are not strongly supported here. However, it is equally important to note the subjective nature of such preferences, and a much larger subject sample may be required to reach firm conclusions in this direction.
Conclusion
This study demonstrated a statistically significant association between QOL and topophilia using a standardized, internationally validated measure of QOL developed by the mental health group of the WHO, and a new construct of environmental preferences defined by the latent variable of topophilia. Synesthetic tendency is the strongest domain of topophilia, and psychological well-being is the strongest domain of QOL. Furthermore, the study demonstrated in the sample population that the appreciation of ecologic diversity is the strongest component of topophilia that is associated with QOL. Within the ecodiversity subdomain, the appreciation of flowers and water bodies are correlated with high QOL, but not the presence of animals, trees, or hilly terrains. These findings are consistent with other findings regarding the ubiquitous preference of natural environments instead of built environments, in the sense that no strong associations were observed between environmental features of complexity and coherence, which are typically assumed to be artificial features. In addition, there were no strong associations between the experience of sensory stimuli, such as sound, smell, or color, and QOL, possibly because of a high level of variance in the latent variable entitled synesthetic tendency. This study provides a new empirical way of assessing restoration and other health benefits that have been theoretically associated with human experience of specific ecosystem components. The approach presented here should be valuable for proactive environmental and landscape design with the aim of providing mental restoration after stress and fatigue.
Figure 1 The statistical model of the association between topophilia, QOL, and their proximate determinants. Structural equation modeling was used to generate confirmatory loading factors for the relationships between each of the questionnaire items for topophilia and the standardized WHOQOL-Bref model. Boxes at level T-1 represent the four major domains of topophilia that were revealed by principal components analysis of responses to rating preferences for questionnaire items included in the boxes at level T-2. Level T-0 in the oval shape represents the latent variable of topophilia. Similarly, boxes at level Q-1 represent the four major domains of QOL identified through principal components analysis of responses to questionnaire items at level Q-2. Level Q-0 in the rectangular shape represents measured values for QOL. The factor values not in parentheses are from this study; comparative values for an international field trial of WHOQOL-Bref are included (in parentheses) from the general instrument validation study reported by Skevington et al. (2004).
Table 1 Descriptive statistics of the sample population and the summary statistics of responses to the composite questionnaire (n = 379).
Questionnaire item Minimum Maximum Mean ± SD
1 Participant’s sex 0 (Male) 1 (Female) 0.58 ± 0.495
2 Date of birth (year only) 1943 1986 1979 ± 5.477
3 Level of education completed 0 5 2.09 ± 1.488
4 Marital status 0 4 0.19 ± 0.605
5 Ethnicity 1 9 5.46 ± 3.468
6 Level of restoration experienced off campus 1 5 2.56 ± 5.152
7 Level of restoration experienced on campus 1 5 3.11 ± 7.411
8 Environmental complexity rating 1 10 4.75 ± 2.673
9 Environmental mystery rating 1 10 5.04 ± 2.661
10 Environmental coherence rating 1 10 5.97 ± 2.354
11 Environmental texture rating 1 10 5.83 ± 2.442
12 Environmental identifiability rating 1 10 7.29 ± 2.633
13 Spaciousness rating 1 10 7.69 ± 2.555
14 Privacy rating 1 10 7.08 ± 2.496
15 Colors rating 1 10 7.38 ± 2.385
16 Smells rating 1 10 7.24 ± 2.566
17 Sounds rating 1 10 7.35 ± 2.521
18 Light rating 1 10 7.59 ± 2.410
19 Tactile (touch stimulation) rating 1 10 6.06 ± 2.467
20 Flowers rating 1 10 7.27 ± 2.515
21 Trees rating 1 10 7.90 ± 2.324
22 Animals rating 1 10 6.00 ± 2.814
23 Flowing water rating 1 10 7.68 ± 2.605
24 Lake or ocean rating 1 10 7.79 ± 2.538
25 Hills or mountains rating 1 10 7.11 ± 2.642
26 Campus rating on topophilia criteria 1 10 7.08 ± 1.942
27 Currently ill? 1 5 1.02 ± 0.259
28 How do you rate your quality of life? 1 5 3.98 ± 0.814
29 How well are you satisfied with your health? 1 5 3.69 ± 0.893
30 What extent does physical pain hamper you? 1 5 4.13 ± 0.999
31 Need medical treatment to function? 1 5 4.38 ± 0.893
32 Enjoy life? 1 5 3.88 ± 0.826
33 Feel life to be meaningful? 1 5 3.76 ± 0.956
34 Able to concentrate? 1 5 3.31 ± 0.883
35 Safe in daily life? 1 5 3.86 ± 0.787
36 Healthy physical environment? 1 5 3.54 ± 0.830
37 Enough energy for daily life? 1 5 3.73 ± 0.808
38 Accept your bodily appearance? 1 5 3.57 ± 0.976
39 Enough money to meet your needs? 1 5 3.32 ± 1.205
40 Information that you need available? 1 5 3.71 ± 0.775
41 Opportunity for leisure activities? 1 5 3.27 ± 0.964
42 Able to get around? 1 5 3.91 ± 0.935
43 Satisfied with your sleep? 1 5 3.27 ± 1.048
44 Satisfied with ability for daily activities? 1 5 3.68 ± 0.826
45 Satisfied with capacity for work? 1 5 3.55 ± 0.922
46 Satisfied with yourself? 1 5 3.78 ± 0.916
47 Satisfied with your personal relationships? 1 5 3.68 ± 1.038
48 Satisfied with your sex life? 1 5 3.31 ± 1.186
49 Satisfied with support from friends? 1 5 3.91 ± 0.915
50 Satisfied with conditions of living space? 1 5 3.71 ± 0.961
51 Satisfied with access to health care? 1 5 3.68 ± 0.982
52 Satisfied with transport? 1 5 3.65 ± 1.131
53 How often do you have negative feelings? 1 5 3.59 ± 0.864
54 Physical domain 6.86 20.00 15.2153 ± 2.24354
55 Psychological domain 6.67 20.00 14.5933 ± 2.49269
56 Social domain 4.00 20.00 14.6029 ± 3.38493
57 Environment domain 5.00 20.00 14.3765 ± 2.33073
58 Overall QOL score 7.18 19.58 14.6985 ± 2.11191
59 Ecodiversity ratings factor 1.00 10.00 7.3687 ± 1.94399
60 Synesthetic tendency ratings factor 1.00 10.00 7.3917 ± 2.09584
61 Cognitive ratings factor 1.00 10.00 5.5342 ± 1.79975
62 Familiarity ratings factor 1.00 10.00 7.3611 ± 2.01276
Table 2 Cronbach α-value estimates of statistical reliability for the associations between observed variables (minor domains) and the two latent variables of topophilia and QOL (major domains).
Domains α-Value
Topophilia
Ecodiversity 0.833
Synesthetic tendency 0.870
Cognitive challenge 0.746
Familiarity 0.684
QOL
Physical health 0.717
Psychological well-being 0.777
Social relationships 0.715
Environmental support 0.751
Table 3 Matrix of correlation coefficients among QOL, topophilia, and respondent experience of restoration.
QOL domains
Topophilia domains Physical Psychological Social Environmental Overall QOL Level of restoration at current location
Cognitive (0.009)
Complexity −0.004 0.029 0.043 −0.030 0.017 0.189**
Mystery −0.019 −0.002 0.095 −0.056 0.016
Coherence −0.016 −0.062 0.003 −0.005 −0.024
Texture −0.061 −0.078 0.047 −0.029 −0.031
Familiarity (0.082)
Identifiability 0.041 0.006 0.072 0.115* 0.069 0.118*
Spaciousness 0.049 0.014 0.095 0.009 0.054
Privacy 0.034 0.092 0.067 −0.006 0.062
Synesthetic tendency (0.077)
Colors 0.036 0.049 0.059 0.041 0.059 0.183**
Smells −0.013 0.010 0.084 −0.005 0.030
Sounds 0.062 0.081 0.058 0.030 0.071
Lighting 0.083 0.092 0.062 0.072 0.095
Tactile −0.002 −0.005 0.084 −0.010 0.033
Ecodiversity (0.123*)
Flowers 0.128* 0.063 0.188* 0.106* 0.162** 0.282**
Trees 0.087 0.012 0.082 0.073 0.084
Animals −0.023 −0.033 0.063 0.024 0.021
Flowing water 0.053 −0.005 0.055 0.059 0.055
Lake/ocean 0.082 0.064 0.136* 0.105* 0.129*
Hills/mountain 0.040 0.039 0.075 0.067 0.075
Values in parentheses are correlation coefficients between the overall QOL and each of the major domains of topophilia tested as a group.
*Pearson correlation coefficients are significant at the 0.05 level (two-tailed).
**Pearson correlation coefficients are significant at the 0.01 level (two-tailed).
==== Refs
References
Betrabet G 1996 The garden as a restorative environment: a theoretical perspective J Ther Horticult 8 15 20
Carlson A 2000. Aesthetics and the Environment: The Appreciation of Nature, Art, and Architecture. London:Routledge.
Hartig T Bowler PA Wolf A 1994 The psychological ecology of ecological restoration Restor Manag Notes 12 133 137
Hartig T Evans GW 1993 Psychological foundations of nature experience Adv Psychol 96 427 457
Hartig T Evans GW Jamner LD Davis DS Gärling T 2003 Tracking restoration in natural and urban field settings J Environ Psychol 23 109 123
Hartig T Staats H 2003 Guest editors’ introduction: restorative environments J Environ Psychol 23 103 107
Hertzog TR Maguire CP Nebel MB 2003 Assessing the restorative components of environments J Environ Psychol 23 159 170
Janzen D 1998 Gardenification of wildland nature and the human footprint Science 279 1312 1313
Kaplan S 1995 The restorative benefits of nature: toward an integrated framework J Environ Psychol 15 169 182
Kaplan S 2001 Meditation, restoration, and the management of mental fatigue Environ Behav 33 480 506
King A Stokols D Talen E Brassington ES Killingsworth R 2002 Theoretical approaches to the promotion of physical activity: forging a transdisciplinary paradigm Am J Prev Med 23 15 25 12133734
Korpela KM Hartig T Kaiser FG Fuhrer U 2001 Restorative experience and self-regulation in favorite places Environ Behav 33 572 589
Kuyken W Orley J Hudelson P Sartorius N 1994 Quality of life assessment across cultures Int J Ment Health 23 5 27
Laumann K Garling T Stormark KM 2001 Rating scale measures of restorative components of environments J Environ Psychol 21 31 34
Laumann K Garling T Stormark KM 2003 Selective attention and heart rate responses to natural and urban environments J Environ Psychol 23 25 134
Nasermoaddeli A Sekine M Hamanishi S Kagamimori S 2003 Associations between sense of coherence and psychological work characteristics with changes in quality of life in Japanese civil servants: a 1-year follow-up study Ind Health 41 236 241 12916754
Parsons R 1991 The potential influences of environmental perception on health J Environ Psychol 11 1 23
Porteous JD 1996. Environmental Aesthetics: Idea, Politics and Planning. London:Routledge.
Saxena S Carlson D Billington R WHOQOL Group. 2001 The WHO quality of life assessment instrument (WHOQOL-Bref): the importance of its items for cross-cultural research Qual Life Res 10 711 721 11871592
Saxena S Sharan P Saraceno B 2003 Budget and financing of mental health services: baseline information on 89 countries from WHO’s Project Atlas J Ment Health Policy Econ 6 135 143 14646006
Skevington SM Lotfy M O’Connell KA WHOQOL Group. 2004 The World Health Organization’s WHOQOL-Bref quality of life assessment: psychometric properties and results of the international field trial. A report from the WHOQOL group Qual Life Res 13 299 310 15085902
Staats H Gattersleben B Hartig T 1997 Change in mood as a function of environmental design: arousal and pleasure on a simulated forest hike J Environ Psychol 17 283 300
Staats H Kieviet A Hartig T 2003 Where to recover from attentional fatigue: an expectancy-value analysis of environmental preference J Environ Psychol 23 147 157
Tuan Y-F 1993. Passing Strange and Wonderful: Aesthetics, Nature, and Culture. Washington, DC:Island Press.
Tuan Y-F 1999. Topophilia: A Study of Environmental Perception, Attitudes, and Values. New York:Columbia University Press.
Ulrich RS 1979 Visual landscapes and psychological well-being Landsc Res 4 17 23
Ulrich RS 1993. Biophilia, biophobia, and natural landscapes. In: The Biophilia Hypothesis (Kellert SR, Wilson EO, eds). Washington, DC:Island Press, 73–137.
Van den Berg AE Koole SL van der Wulp NY 2003 Environmental preference and restoration: (how) are they related? J Environ Psychol 23 135 146
Weatherall M McPherson K Taylor W Simpson R 2004 Avoiding pitfalls of correlation coefficients in the assessment of measurement instruments in rehabilitation research Clin Rehab 18 186 194
WHOQOL Group. 1994 Development of the WHOQOL: rationale and current status Int J Mental Health 23 24 56
WHOQOL Group. 1995 The World Health Organization Quality of Life Assessment (WHOQOL): position paper from the World Health Organization Soc Sci Med 41 1403 1409 8560308
WHOQOL Group. 1998a Development of the World Health Organization WHOQOL-Bref quality of life assessment Psych Med 28 551 558
WHOQOL Group. 1998b The World Health Organization Quality of Life Assessment (WHOQOL): development and general psychometric properties Soc Sci Med 46 1569 1585 9672396
|
15687050
|
PMC1277856
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):143-148
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7467
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7421ehp0113-00014915687051ResearchArticlesDevelopmental Exposure to Low-Dose PBDE-99: Effects on Male Fertility and Neurobehavior in Rat Offspring Kuriyama Sergio N. Talsness Chris E. Grote Konstanze Chahoud Ibrahim Institute of Clinical Pharmacology and Toxicology, Department of Toxicology, Charité University Medical School Berlin, Campus Benjamin Franklin, Berlin, GermanyAddress correspondence to I. Chahoud, Charité University Medical School Berlin, Campus Benjamin Franklin, Department of Toxicology, Garystrasse 5, 14195 Berlin, Germany. Telephone: 49-30-8445-1750. Fax: 49-30-8445-1761. E-mail:
[email protected] thank H. Marburger and B. Woelffel for exemplary technical assistance and C. Gericke for valuable support on statistical analysis and graphic layout.
This work was supported by UBA–Forschungsund Entwicklungsvorhaben grant 29965221/04.
The authors declare they have no competing financial interests.
2 2005 4 11 2004 113 2 149 154 16 7 2004 4 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. In utero exposure to a single low dose of 2,2′,4,4′,5-pentabromodiphenyl ether (PBDE-99) disrupts neurobehavioral development and causes permanent effects on the rat male reproductive system apparent in adulthood. PBDEs, a class of flame retardants, are widely used in every sector of modern life to prevent fire. They are persistent in the environment, and increasing levels of PBDEs have been found in biota and human breast milk. In the present study we assessed the effects of developmental exposure to one of the most persistent PBDE congeners (PBDE-99) on juvenile basal motor activity levels and adult male reproductive health. Wistar rat dams were treated by gavage on gestation day 6 with a single low dose of 60 or 300 μg PBDE-99/kg body weight (bw). In offspring, basal locomotor activity was evaluated on postnatal days 36 and 71, and reproductive performance was assessed in males at adulthood. The exposure to low-dose PBDE-99 during development caused hyperactivity in the offspring at both time points and permanently impaired spermatogenesis by the means of reduced sperm and spermatid counts. The doses used in this study (60 and 300 μg/kg bw) are relevant to human exposure levels, being approximately 6 and 29 times, respectively, higher than the highest level reported in human breast adipose tissue. This is the lowest dose of PBDE reported to date to have an in vivo toxic effect in rodents and supports the premise that low-dose studies should be encouraged for hazard identification of persistent environmental pollutants.
developmentendocrine active compoundsin utero exposurelow-dose effectsmale fertilityneurobehaviorPBDE-99
==== Body
Flame retardants have been shown to possess a wide spectrum of toxicity in laboratory animals, and they are present in every level of the food chain. However, this class of substances is important in modern society because of their ability to save lives by limiting the consequences of fires. One class of these chemicals is the polybrominated diphenyl ethers (PBDEs), used in plastics, textiles, foams, and electronic circuitry to avoid fire propagation. They are additives mixed into polymers and, as such, are not chemically bound and can leach into the surrounding environment (de Wit 2002). Because of their high lipophilicity and persistence, PBDEs have been found in sewage sludge, sediment, biota, and humans (Darnerud et al. 2001). Between 1972 and 1997, analysis of human milk has shown a 60-fold increase of PBDE levels in Swedish women (Meironyte et al. 1999), and a recent article has reported much higher levels in human breast adipose tissue in the San Francisco Bay area (She et al. 2002).
PBDEs display a molecular structure similar to that of polychlorinated biphenyls (PCBs), and therefore, one would expect that they also exhibit a different spectrum of toxicity among the 209 possible congeners. PBDE-99 is one of the most persistent congeners detected in almost all environmental samples (Darnerud et al. 2001; de Wit 2002; McDonald 2002). Neurobehavioral toxicity seems to be the most sensitive target of PBDEs in rodents that has been reported to date, although pronounced effects on thyroid homeostasis have been also reported (Hallgren and Darnerud 2002; Hallgren et al. 2001; Zhou et al. 2001, 2002). In rats and mice, pre- and/or postnatal exposure to PBDE-99 has been shown to cause permanent neurobehavioral disturbances in offspring at doses below that able to cause maternal toxicity (Branchi et al. 2002, 2003; Eriksson et al. 2001, 2002; Viberg et al. 2002). One study suggests that cholinergic nicotinic receptors may also be a target for PBDEs, because they found a decrease in α-bungarotoxin binding in hippocampus in mice neonatally exposed to PBDE-153 (Viberg et al. 2003a). However, the mechanism(s) underlying PBDE-induced toxicity is not clear.
There is growing evidence that male reproductive health has been deteriorating over the last few decades. Studies in France, Belgium, Denmark, and Great Britain have reported a significant temporal decline in human semen quality in the last half-century (Auger et al. 1995; Irvine et al. 1996; Van Waeleghem et al. 1996). During the same time, the numbers of hypospadias and cryptorchidisms appear to be increasing (Cour-Palais 1966; Czeizel 1985; Czeizel et al. 1981; Kallen and Winberg 1982; Matlai and Beral 1985; Yucesan et al. 1993), and a similar trend has been observed in the incidence of testicular cancer, which is now the most common malignancy of young men (Adami et al. 1994; Boyle et al. 1987; Forman and Moller 1994; Hakulinen et al. 1986; Nethersell et al. 1984; Pike et al. 1987). Despite the fact that an extensive review of the published data suggests a temporal decline in human sperm production, methodologic bias hinders a final conclusion. The etiology of the malignancies is unknown, but biological plausibility and experimental evidence support the hypothesis that environmental pollutants are acting as endocrine-active compounds. For example, in rodents, reduced sperm counts have been observed at adulthood when animals were pre- and/or postnatally exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (Faqi et al. 1998b; Theobald and Peterson 1997), PCBs (Faqi et al. 1998a; Hsu et al. 2003; Kuriyama and Chahoud 2004), and the pesticides deltamethrin (Andrade et al. 2002) and lindane (Dalsenter et al. 1997). Nevertheless, there is scant information regarding possible effects of PBDEs on male reproduction.
Because of the increasing levels of PBDEs found in human and biota samples, we conducted the present study to examine the effects of a single low dose of 60 or 300 μg PBDE-99/kg body weight (bw) on gestation day (GD)6 on neurobehavior and male reproductive health in rat offspring. Assuming that fat content in rats is approximately 14% of total body weight, the doses used in this study are approximately 6 and 29 times, respectively, higher than the highest level reported by She et al. (2002; 72.2 μg/kg fat) in human breast adipose tissue It has been demonstrated that PBDE-99 possesses a long half-life (~ 41.6 days in female rat) (Geyer et al. 2004), and we therefore treated the dams on GD6 in order to assess the effects of PBDE-99 during embriofetal and lactational (~ 37 days) periods. Because PBDEs have the ability to interfere with thyroid hormone (TH) homeostasis, we included a reference group for TH-mediated effects by adding the goitrogen 6-n-propyl thiouracil (PTU) in the drinking water of pregnant females.
Material and Methods
Animals and treatment.
Wistar rats (HsdCpb:WU; Fa. Harlan-Winkelmann, Borchen, Germany) weighing 200 ± 15 g were allowed to acclimatize for 2 weeks. The rats were exposed to constant light/dark periods of 12 hr each, a temperature of 21 ± 1°C, and 50 ± 5% relative humidity. Rodent chow (Altromin 1324; Altromin GmbH, Lage, Germany) and tap water were available ad libitum. Two nongravid females were placed with one male for 3 hr, and the day of sperm detection in the vaginal smear was considered day 0 of gestation. The gravid females were randomly assigned among the four groups and housed individually in type III Macrolon cages with stainless steel covers and wood shavings (Altromin GmbH). 2,2′,4,4′,5-Pentabromodiphenyl ether (PBDE-99; 98% pure), lot number VL02, was purchased from LGC Promochem GmbH (Wesel, Germany). Pregnant rats were treated orally by gavage with a single dose of 60 μg PBDE/kg (n = 20) or 300 μg PBDE-99/kg (n = 19) on GD6. The control pregnant rats (n = 16) received the vehicle, peanut oil, in a volume of 10 mL/kg bw on the same day. An additional group was administered the goitrogen PTU (6-n-propyl-2-thiouracil; Sigma-Aldrich Chemicals GmbH, Steinheim, Germany), which served as a reference control for TH effects. PTU was given to the gravid dams by adding 5 mg/L PTU in the drinking water on GD7–21. Dams were allowed to deliver, and the litter size was not artificially altered. The experimental protocol has the approval of the National Animal Protection Law (Tierschutzgesetz BGBI. IS. 3082, 2002).
Postnatal reflex and developmental landmarks.
Developmental landmarks (eruption of incisors, fur development, eye opening, and testes descent) and postnatal reflexes were evaluated in all pups (control, n = 163; PTU, n = 200; 60 μg PBDE, n = 218; 300 μg PBDE, n = 200). Starting on postnatal day (PND)3, we monitored the offspring for the development of spontaneous cliff-drop aversion reflex, and beginning on PND18, we examined their ability to stay on a rotating rod for 3 min at 7 rpm.
Locomotor activity.
Circadian motility was measured over 24-hr periods on PND36 and PND71 in individual offspring using a Mobilitron (FU-Berlin/Eisenberger GmbH, Dillenburg, Germany), a device that monitors the locomotion of the animal at 5-min intervals using three infrared photocells per cage. Habituation in the Mobilitron took place for a 24-hr period before testing began in order to allow the animals to adjust to their new environment and the solitary accommodation before measurements were taken. The locomotor activity of one male and one female per litter per group (one animal per cage) were evaluated before puberty (PND36) and after puberty (PND71). The animals were randomly assigned in the Mobilitron (which allows simultaneous measurement of 48 cages) to avoid confounding factors. The method has been described in detail by Thiel et al. (1989).
Reproductive assessment of adult male offspring.
At adulthood (~ PND140), 12 males/group (from different litters) were killed by decapitation. Trunk blood was collected for hormone analysis, and organ weights (thymus, spleen, liver, testis, epididymis, seminal vesicle, and ventral prostate) were recorded. The right testis and caudal epididymis were kept in saline buffer for spermatid and sperm counts, respectively.
Spermatid number.
The testis was minced and homogenized for 1 min in 10 mL 0.9% NaCl containing 0.5% Triton X-100 at medium speed in an IKA-RW 15 Tissuemizer (Janke and Kunkel, Staufen in Breisgau, Germany). The number of homogenization-resistant spermatids was counted in a Buerker hemocytometer (Brand GmbH, Wertheim, Germany). Daily sperm production was calculated, dividing the number of homogenization-resistant spermatids by 6.1 (Robb et al. 1978).
Sperm count and morphology.
Cauda epididymis was minced and homogenized for 1 min in 10 mL 0.9% NaCl containing 0.5% Triton X-100 at medium speed in an IKA-RW 15 Tissuemizer. The number of homogenization-resistant sperm was counted in a hemocytometer (Buerker). For sperm morphology, the ductus deferens was rinsed with 1 mL 0.9% NaCl to obtain a sperm suspension. Aliquots were stained with 2% eosin to assess the percentage of morphologically abnormal sperm by evaluating 200 sperm/animal.
Testosterone and luteinizing hormone levels.
After decapitation, trunk blood was collected and allowed to clot on an ice bath (4°C) for 2 hr. Serum was collected via centrifugation of clotted samples (2,500 rpm for 15 min) and stored at –20°C for later analyses. Serum testosterone and luteinizing hormone (LH) were measured using the ELISA kit purchased from DRG Diagnostics GmbH (Marburg, Germany). Testosterone was measured in crude rat serum, which is reliable for comparison among groups, but matrix effects cause uncertainties with respect to absolute values.
Male reproductive performance.
Adult male offspring (± 150 days of age; n = 15–19 animals/group), representing all litters, were mated with untreated females (1:1) daily for 14 days to determine whether the males were fertile and could sire normal offspring. Vaginal smears were collected daily and examined for the presence of sperm. The day of sperm detection in vaginal smears was considered day 0. The dams were sacrificed on GD21 and the uterus was excised. The uterine and fetal weights and the numbers of implantations, resorptions, and fetuses were determined. The fetuses were examined for external anomalies and sexed.
Male sexual behavior.
Approximately on PND160, 20 males/group (representing all litters) were mated with untreated females in estrus (1:1), and the sexual behavior of each mating was recorded for 20 min under blue light illumination (black light, 75W; Osram, Berlin, Germany) using a video camcorder (Hi8 Handycam CCD-V800E, Sony, Tokyo, Japan). The recorded videos, which provide a permanent record and the opportunity for replay, were evaluated by a trained observer in a blind fashion. The phase of the estrous cycle of the untreated females was predetermined by examining vaginal smears. The method was previously described in detail by Chahoud and Faqi (1998).
Statistical analyses.
The statistical analyses were performed with SPSS software, version 11.5 for Windows (SPSS Inc., Chicago, IL, USA). Data for males and females within each group were tested by the Student t-test. Data with normal distribution were analyzed by analysis of variance (ANOVA) followed by the Dunnett t-test. The equality of survival distributions was examined using the Kruskal-Wallis test followed by the Mann-Whitney U test. Proportions were analyzed by Fisher’s exact test, and statistical differences were considered significant when p < 0.05.
Results
Spontaneous behavior and developmental landmarks.
Spontaneous behavior and developmental landmarks are shown in Figure 1. Except for cliff-drop aversion reflex, no differences between sexes were detected in the statistical analysis, and therefore, the male and female data were pooled for the analysis of all the developmental landmarks and the rotating rod reflex. The age at fur development, testes descent, and the ability to master the rotating rod test were not different between treated and control animals (data not shown). However, cumulative survival function analysis for the age at eye opening, eruption of incisors, and the cliff-drop aversion reflex revealed a statistically significant difference among the groups. The onset of eye opening was earlier in the PTU-treated litters than among controls (Figure 1A), and the eruption of incisors was delayed in the groups treated with PTU or 300 μg/kg PBDE compared with controls (Figure 1B). The development of the cliff-drop aversion reflex was significantly delayed in both PTU-exposed male and female offspring as well as in males exposed to the 300 μg dose of PBDE-99 (Figure 1C,D).
Locomotor activity.
Using the Mobilitron apparatus, individual locomotion of rats was measured over 24 hr in young and pubertal offspring. Statistical analysis revealed no difference between the sexes for all groups tested, and therefore, the data from the males and females are presented together. On PND36, the total light beam interruption (LBI) count per day was significantly greater in the PTU and 300 μg/kg PBDE groups (Figure 2A). The number of active hours per day was longer in the PBDE 300 μg/kg group, an effect not seen in the PTU group (Figure 2B). The qualitative analysis (i.e., LBI count per phase and duration of activity per phase) on the same day (PND36) confirms what was observed in the quantitative analysis. Both the PTU and 300 μg/kg PBDE groups were more active during the active phases compared with control, and the duration of the active phases were also longer even though there were no statistically significant differences in the number of active phases (Figure 2C,D). An active phase is defined when the animal begins to move (associated with LBIs) until a pause (a period of no LBI) is observed. No differences compared with control were seen in the 60 μg PBDE group on PND36. At puberty (PND71), the quantitative analysis indicated that the two PBDE groups were hyperactive compared with controls. In other words, both the LBI counts and duration of activity per day were significantly increased in 60 μg PBDE and 300 μg PBDE groups. In that age (PND71), no statistically significant qualitative differences were observed among the groups (Figure 3).
Body and organ weights of adult male offspring (PND140).
Body, liver, thymus, and spleen weights of adult male offspring are given in Table 1. We observed no changes in body, liver, and thymus weights related to the treatment. Pre- and postnatal exposure to PBDE-99 and gestational exposure to PTU produced a significant increase in absolute spleen weight (Table 1). However, when spleen weight was expressed as a ratio of body weight (relative weight), only animals in the 60 μg PBDE-99 group exhibited the same trend (Table 1).
Male fertility and reproductive performance.
Table 2 shows the reproductive organ weights, as well as sperm and spermatid counts, sperm morphology, and steroid hormone levels. We found no significant differences in the absolute testis and epididymis weights; however, when expressed as a percentage of body weight (relative weight), the PTU and PBDE 300 μg groups had smaller testes, whereas the epididymis relative weights were decreased in all three treatment groups compared with controls. No differences were observed in prostate and seminal vesicle (absolute and relative) weights (Table 2). The lower testis and epididymis weights were accompanied by reductions in sperm and spermatid counts as well as daily sperm production. Reductions in testicular spermatid count and sperm count from caudal epididymis were observed in all treatment groups (Table 2). It is noteworthy that daily sperm production was reduced by approximately 30% from controls. The decrease in sperm production was not associated with poor sperm quality because the percentage of abnormal sperm was within normal limits in all groups. Testosterone and LH levels were also not affected, suggesting a minor role for steroid hormones in the impairment of sperm production. When the litter mates of the animals analyzed for sperm counts were mated with untreated females for fertility studies, exposed males could sire offspring similar to the control males (Table 3). Uterine weight, litter size, and numbers of implantations, resorptions, and viable fetuses were within the normal range of control (Table 3).
Sexual behavior.
Pre- and postnatal exposure to PTU or either dose of PBDE-99 did not impair sexual behavior of the adult male offspring. Ejaculatory and mounting latencies, intromission frequency and latency, and number of penetrations were normal when all groups were compared with controls (Table 4). However, the number of animals that had two or more ejaculations during 20 min of mating was significantly lower in the PBDE-exposed animals. Approximately 50% of controls had a second ejaculation, whereas only 39% and 21% of the males from the 60 μg PBDE and 300 μg PBDE groups, respectively, achieved a second ejaculation (Table 4).
Discussion
In the present study we found consistent evidence that exposure to low doses of PBDE-99 during critical periods of development affects motor activity and permanently impairs spermatogenesis in adult rat offspring. This is the lowest dose of PBDE reported to date to have an in vivo toxic effect in rodents. We observed neurobehavioral changes on PND36 and PND71 in offspring exposed to the highest dose (300 μg/kg bw), whereas behavioral changes were present only at puberty in the lowest dose tested (60 μg/kg bw; Figures 2 and 3). Other groups have reported neurodevelopmental disturbances in rodents exposed to PBDEs (Branchi et al. 2002, 2003; Eriksson et al. 2001; Viberg et al. 2003a, 2003b), and this system seems to be the most sensitive to PBDE-induced toxicity. Supporting our data, Eriksson et al. (2001) found that neonatal exposure (single dose on PND10) to PBDE-99 or PBDE-47 disrupts spontaneous behavior and causes hyperactivity in mice that appears to be permanent and to worsen with age (Eriksson et al. 2001). The crucial role of THs during brain development is well known, and disturbances in TH homeostasis (e.g., PTU exposure) can cause serious impairment in neurologic development (Porterfield 1994). In support of our findings, previous studies have reported hyperactivity in rodent offspring after pre- and postnatal hypothyroidism induced by goitrogens such as PTU (Akaike et al. 1991; Davenport and Hennies 1976; Goldey et al. 1995; Tamasy et al. 1986). However, the behavioral changes in PBDE-99 groups (persistent at least until PND71) observed in this study are not similar to our reference group for TH-mediated effects (PTU; transient hyperactivity only on PND36), suggesting that the neurotoxicity induced by PBDE-99 may stem from mechanism(s) other than those caused by PTU. Even though more mechanistic studies are lacking, the cholinergic system seems to be affected after neonatal exposure to PBDE, because Viberg et al. (2003a) found reduced amounts of nicotinic receptors in the hippocampus of exposed animals using an α-bungarotoxin assay, and the response of the cholinergic agent nicotine was altered in mice neonatally exposed to PBDE-99 (Viberg et al. 2002, 2003a). In that way, hyperactivity induced by PBDE-99 might be explained by changes in the cholinergic system during pre-and postnatal exposure. Nevertheless, one should not rule out other mechanisms because some hydroxylated PBDE metabolites have been shown to possess high binding affinities to TH receptors (Marsh et al. 1998). It is plausible to hypothesize that also the binding of the PBDE-99 molecule or its metabolite to the TH receptor in the developing brain could cause neurobehavioral disturbance in offspring.
Increasing evidence suggests that continuous exposure to environmental pollutants is related to the postulated deterioration of male reproductive health in the last 50 years. This hypothesis highlights the need for more experimental studies that employ doses relevant to environmental/human exposure scenarios in order to elucidate possible mechanisms involved in such a decline. In this study, developmental exposure to low-dose PBDE-99 not only caused persistent neurobehavioral effects but also permanently affected adult male reproductive health (Table 2). This is the first report on effects of PBDE-99 on male reproductive performance because our survey of the literature failed to find data on this topic. Questions regarding persistent chemical contamination typically focus on bioaccumulation, neurotoxicity, and carcinogenicity. However, the male reproductive system has been shown to be a very sensitive end point when the insult occurs during critical periods of development (Andrade et al. 2002; Cooke et al. 1992; Dalsenter et al. 1997; Faqi et al. 1998b; Kuriyama and Chahoud 2004; Sharpe et al. 1995). In the mating study, no effect on fertility was seen when males were mated with untreated females, which is not inconsistent with the observed decrease in daily sperm count of the littermates (Tables 2 and 3). In rats, sperm production can be reduced up to 90% without compromising fertility (Aafjes et al. 1980; Kirby et al. 1992). On the other hand, relatively small changes in sperm production in men may have severe consequences for human reproduction (Zenick and Clegg 1989). Because the normal human sperm count is near the threshold for the number of sperm needed to ensure reproductive competence, sperm count is a sensitive and validated end point for reproductive toxicology assessment. The growth and maturation of the developing testis as well as the maintenance of spermatogenesis are regulated by several endocrine and paracrine factors. Among them, thyroid function during early life has a major impact on regulating testicular growth and function. When rats are made hypothyroid during a critical window of neonatal development, permanent increases in adult testis size and sperm production have been observed (Cooke et al. 1992). However, this effect occurs only when rats are hypothyroid during the first week of postnatal development (Cooke et al. 1992). Using a dose 200-fold lower than that reported by Cooke et al. (1992), we observed that prenatal hypothyroidism induced by PTU caused an opposite effect, namely, decreased sperm production and testis size. Impaired spermatogenesis and reduced testicular weight seen in males exposed to both doses of PBDE-99 might also be correlated to alterations in TH concentrations. However, the mechanisms underlying these effects observed both in PTU-exposed and PBDE-99–exposed animals remain to be elucidated. This study demonstrates for the first time that exposure to a low dose of PBDE-99, which resembles the human exposure levels, causes permanent impairment of spermatogenesis in rats. These findings encourage further investigation on mechanistic studies in order to assess the hazard of flame retardants on human reproductive health. Moreover, the issue of synergistic or additive effects when PBDEs are combined to other persistent pollutants (e.g., PCBs and DDT) remains to be elucidated.
Figure 1 Cumulative survival function of the spontaneous reflex, cliff-drop aversion, and the developmental landmarks eye opening and eruption of incisors from rat offspring after pre- and postnatal (via milk) low-dose PBDE-99 (60 or 300 μg/kg bw) exposure. (A) Time of eye opening. (B) Time of eruption of incisors. (C) Time to develop the cliff-drop aversion reflex in male offspring. (D) Time to develop the cliff-drop aversion reflex in female offspring. Groups were compared using the Kruskal-Wallis test followed by the Mann-Whitney U test, and differences were considered statistically significant when p < 0.05: PTU in (A) and (D); PTU and PBDE 300 in (B) and (C).
Figure 2 Locomotor activity of rat offspring after pre- and postnatal (via milk) low-dose PBDE-99 (60 or 300 μg/kg bw) exposure: quantitative and qualitative analysis on PND36 showing total activity (LBI) and duration of activity per day and per active phase. (A) LBI counts per day. (B) Duration (hours) of activity per day. (C) LBI counts per active phase. (D) Duration of activity (minutes) per active phase. Bars represent mean ± SEM.
*p < 0.05; significances were detected by ANOVA, followed by the Dunnett t-test when p < 0.05.
Figure 3 Locomotor activity of rat offspring after pre- and postnatal (via milk) low-dose PBDE-99 (60 or 300 μg/kg bw) exposure: quantitative and qualitative analysis on PND71 showing total activity (LBI) and duration of activity per day and per active phase. (A) LBI counts per day. (B) Duration (hours) of activity per day. (C) LBI counts per active phase. (D) Duration of activity (minutes) per active phase. Bars represent mean ± SEM.
*p < 0.05; significances were detected by ANOVA, followed by the Dunnett t-test when p < 0.05.
Table 1 Absolute and relative (percent of body weight) organ weights from adult offspring (PND140) exposed pre- and postnatally (via milk) to PBDE-99 (n = 12/group).
PBDE
Parameters Control PTU 60 μg/kg bw 300 μg/kg bw
Body weight (g) 311.7 ± 8.3 335.9 ± 9.9 320.5 ± 5.9 334.9 ± 8.6
Liver weight (g) 10.43 ± 0.47 11.17 ± 0.53 10.82 ± 0.21 11.26 ± 0.39
Percent bw 3.35 ± 0.08 3.31 ± 0.07 3.38 ± 0.04 3.36 ± 0.07
Thymus weight (g) 0.34 ± 0.02 0.34 ± 0.03 0.36 ± 0.02 0.32 ± 0.03
Percent bw 0.11 ± 0.006 0.10 ± 0.008 0.11 ± 0.005 0.10 ± 0.007
Spleen weight (g) 0.55 ± 0.01 0.63 ± 0.03* 0.60 ± 0.02* 0.60 ± 0.02*
Percent bw 0.17 ± 0.004 0.19 ± 0.006 0.19 ± 0.005* 0.18 ± 0.004
Absolute and relative organ weights were analyzed using ANOVA followed by the Dunnett t-test. Values are presented as mean ± SEM.
* p < 0.05.
Table 2 Reproductive organ weights, hormone levels, sperm number, and daily sperm production in adult offspring (PND140) exposed pre- and postnatally (via milk) to PBDE-99 (n = 12/group).
PBDE
Parameters Control PTU 60 μg/kg bw 300 μg/kg bw
Testis weight (g) 1.57 ± 0.06 1.47 ± 0.10 1.58 ± 0.03 1.53 ± 0.04
Percent bw 0.51 ± 0.02 0.44 ± 0.03* 0.49 ± 0.01 0.46 ± 0.01*
Epididymis weight 0.58 ± 0.02 0.55 ± 0.02 0.56 ± 0.01 0.58 ± 0.02
Percent bw 0.19 ± 0.01 0.17 ± 0.01* 0.18 ± 0.002* 0.17 ± 0.007*
Seminal vesicle weight empty (g) 0.99 ± 0.04 1.11 ± 0.04 1.00 ± 0.03 1.04 ± 0.05
Percent bw 0.32 ± 0.01 0.33 ± 0.01 0.31 ± 0.01 0.31 ± 0.01
Prostate (g) 0.38 ± 0.03 0.40 ± 0.02 0.40 ± 0.01 0.43 ± 0.03
Percent bw 0.12 ± 0.01 0.12 ± 0.01 0.12 ± 0.003 0.13 ± 0.008
Spermatid (106) 266.2 ± 7.5 198.6 ± 10.5* 182.8 ± 7.6* 175.0 ± 5.7*
Daily sperm production (106) 43.6 ± 1.2 32.6 ± 1.7* 30.0 ± 1.2* 28.7 ± 0.9*
Sperm number (106) 189.6 ± 11.7 143.2 ± 8.0* 134.7 ± 6.4* 156.3 ± 8.1*
Abnormal sperm (%) 6.3 ± 0.8 7.7 ± 1.0 5.6 ± 0.5 7.9 ± 1.0
LH (ng/mL) 10.8 ± 1.2 12.4 ± 1.3 14.4 ± 2.1 10.3 ± 1.1
Testosterone (ng/mL) 8.7 ± 1.2 10.0 ± 1.4 7.5 ± 1.0 8.4 ± 1.4
Absolute and relative organ weights were analyzed using ANOVA followed by the Dunnett t-test. Values are presented as mean ± SEM.
* p < 0.05.
Table 3 Reproductive performance of adult male offspring exposed pre- and postnatally (via milk) to PBDE-99.
PBDE
Parameters Control PTU 60 μg/kg bw 300 μg/kg bw
No. of dams 19 18 15 19
Body weight gain (%) 49.3 46.3 47.5 50.9
Uterine weight (g) 73.6 ± 4.3 77.8 ± 1.7 71.2 ± 2.7 71.1 ± 3.5
Implantations (n) 214 203 161 217
Implantations/litter 11.3 ± 0.18 11.3 ± 0.11 10.7 ± 0.15 11.4 ± 0.12
Viable fetuses/litter 10.8 ± 0.19 10.9 ± 0.11 10.1 ± 0.15 10.3 ± 0.17
Total resorptions (%) 9 (4) 6 (3) 10 (6) 21 (10)
Fetal weight/litter (g) 4.70 ± 0.10 4.66 ± 0.06 4.96 ± 0.13 4.85 ± 0.07
Sex ratio (male/female) 47.3/52.7 47.4/52.6 46.4/53.6 42.9/57.1
Male rats were pre- and postnatally exposed to a low dose (60 μg or 300 μg/kg bw) of PBDE-99 and mated with non-exposed females. Values are presented as mean ± SEM.
Table 4 Male sexual behavior of adult offspring exposed pre- and postnatally (via milk) to PBDE-99.
Parameters Control PTU PBDE 60 PBDE 300
No. of animals with ejaculation/total no. (%) 17/20 (85) 18/20 (90) 19/20 (95) 17/20 (85)
Mounting latency (sec) 24.7 ± 3.9 23.4 ± 4.5 41.0 ± 9.6 27.2 ± 4.9
Intromission latency (sec) 50.7 ± 11.6 37.2 ± 7.3 54.7 ± 10.1 35.1 ± 5.5
Ejaculatory latency (min) 11.1 ± 0.7 8.4 ± 1.0 12.9 ± 1.1 12.5 ± 0.8
Intromission frequency (n/min) 0.99 ± 0.08 0.97 ± 0.12 1.06 ± 0.08 1.23 ± 0.10
No. of penetrations before the first ejaculation 19.8 ± 1.5 19.5 ± 2.5 21.3 ± 1.5 24.7 ± 2.0
Percent of animals with two or more ejaculations 53 71 39 21*
Animals were pre- and postnatally exposed to a low dose of PBDE-99 (60 μg or 300 μg/kg bw) or PTU (5 mg/L). Adult offspring were mated with nonexposed females. Values are presented as mean ± SEM.
* p < 0.05.
==== Refs
References
Aafjes JH Vels JM Schenck E 1980 Fertility of rats with artificial oligozoospermia J Reprod Fertil 58 345 351 7431265
Adami HO Bergstrom R Mohner M Zatonski W Storm H Ekbom A 1994 Testicular cancer in nine northern European countries Int J Cancer 59 33 38 7927900
Akaike M Kato N Ohno H Kobayashi T 1991 Hyperactivity and spatial maze learning impairment of adult rats with temporary neonatal hypothyroidism Neurotoxicol Teratol 13 317 322 1886541
Andrade AJ Araujo S Santana GM Ohi M Dalsenter PR 2002 Reproductive effects of deltamethrin on male offspring of rats exposed during pregnancy and lactation Regul Toxicol Pharmacol 36 310 317 12473415
Auger J Kunstmann JM Czyglik F Jouannet P 1995 Decline in semen quality among fertile men in Paris during the past 20 years N Engl J Med 332 281 285 7816062
Boyle P Kaye SB Robertson AG 1987 Changes in testicular cancer in Scotland Eur J Cancer Clin Oncol 23 827 830 3653199
Branchi I Alleva E Costa LG 2002 Effects of perinatal exposure to a polybrominated diphenyl ether (PBDE 99) on mouse neurobehavioural development Neurotoxicology 23 375 384 12387364
Branchi I Capone F Alleva E Costa LG 2003 Polybrominated diphenyl ethers: neurobehavioral effects following developmental exposure Neurotoxicology 24 449 462 12782110
Chahoud I Faqi AS 1998 An optimized approach for the assessment of sexual behavior in male rats Reprod Toxicol 12 667 671 9875700
Cooke PS Porcelli J Hess RA 1992 Induction of increased testis growth and sperm production in adult rats by neonatal administration of the goitrogen propylthiouracil (PTU): the critical period Biol Reprod 46 46 154
Cour-Palais IJ 1966 Spontaneous descent of the testicle Lancet 1 7452 1403 1405 4160959
Czeizel A 1985 Increasing trends in congenital malformations of male external genitalia Lancet 1 8426 462 463 2857840
Czeizel A Erodi E Toth J 1981 An epidemiological study on undescended testis J Urol 126 524 527 6116808
Dalsenter PR Faqi AS Webb J Merker HJ Chahoud I 1997 Reproductive toxicity and toxicokinetics of lindane in the male offspring of rats exposed during lactation Hum Exp Toxicol 16 146 153 9088967
Darnerud PO Eriksen GS Johannesson T Larsen PB Viluksela M 2001 Polybrominated diphenyl ethers: occurrence, dietary exposure, and toxicology Environ Health Perspect 109 suppl 1 49 68 11250805
Davenport JW Hennies RS 1976 Perinatal hypothyroidism in rats: persistent motivational and metabolic effects Dev Psychobiol 9 1 67 82 943352
de Wit CA 2002 An overview of brominated flame retardants in the environment Chemosphere 46 583 624 11999784
Eriksson P Jakobsson E Fredriksson A 2001 Brominated flame retardants: a novel class of developmental neurotoxicants in our environment? Environ Health Perspect 109 903 908 11673118
Eriksson P Viberg H Jakobsson E Orn U Fredriksson A 2002 A brominated flame retardant, 2,2′,4,4′,5-pentabromodiphenyl ether: uptake, retention, and induction of neurobehavioral alterations in mice during a critical phase of neonatal brain development Toxicol Sci 67 98 103 11961221
Faqi AS Dalsenter PR Mathar W Heinrich-Hirsch B Chahoud I 1998a Reproductive toxicity and tissue concentrations of 3,3’,4,4′- tetrachlorobiphenyl (PCB 77) in male adult rats Hum Exp Toxicol 17 151 156 9587783
Faqi AS Dalsenter PR Merker HJ Chahoud I 1998b Reproductive toxicity and tissue concentrations of low doses of 2,3,7,8- tetrachlorodibenzo-p -dioxin in male offspring rats exposed throughout pregnancy and lactation Toxicol Appl Pharmacol 150 383 392 9653070
Forman D Moller H 1994 Testicular cancer Cancer Surv 19–20 323 341
Geyer HJ Schramm KW Darnerud PO Aune M Feicht EA Fried KW 2004 Terminal elimination half-lives of the brominated flame retardants TBBPA, HBCD and lower brominated PBDEs in humans Organohalogen Compounds 66 3867 3872
Goldey ES Kehn LS Rehnberg GL Crofton KM 1995 Effects of developmental hypothyroidism on auditory and motor function in the rat Toxicol Appl Pharmacol 135 67 76 7482541
Hakulinen T Andersen A Malker B Pukkala E Schou G Tulinius H 1986 Trends in cancer incidence in the Nordic countries. A collaborative study of the five Nordic cancer registries Acta Pathol Microbiol Immunol Scand 288 suppl 1 151
Hallgren S Darnerud PO 2002 Polybrominated diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs) and chlorinated paraffins (CPs) in rats—testing interactions and mechanisms for thyroid hormone effects Toxicology 177 227 243 12135626
Hallgren S Sinjari T Hakansson H Darnerud PO 2001 Effects of polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) on thyroid hormone and vitamin A levels in rats and mice Arch Toxicol 75 200 208 11482517
Hsu PC Li MH Guo YL 2003 Postnatal exposure of 2,2′,3,3’,4,6′-hexachlorobiphenyl and 2,2′,3,4′,5′,6-hexachlorobiphenyl on sperm function and hormone levels in adult rats Toxicology 187 117 126 12699901
Irvine S Cawood E Richardson D MacDonald E Aitken J 1996 Evidence of deteriorating semen quality in the United Kingdom: birth cohort study in 577 men in Scotland over 11 years Br Med J 312 467 471 8597676
Kallen B Winberg J 1982 An epidemiological study of hypospadias in Sweden Acta Paediatr Scand 293 suppl 1 21
Kirby JD Jetton AE Cooke PS Hess RA Bunick D Ackland JF 1992 Developmental hormonal profiles accompanying the neonatal hypothyroidism-induced increase in adult testicular size and sperm production in the rat Endocrinology 131 559 565 1639007
Kuriyama SN Chahoud I 2004 In utero exposure to low-dose 2,3’,4,4′,5-pentachlorobiphenyl (PCB 118) impairs male fertility and alters neurobehavior in rat offspring Toxicology 202 3 185 197 15337582
Marsh G Bergman A Bladh LG Gillner M Jakobsson E 1998 Synthesis of p-hydroxybromodiphenyl ethers and binding to the thyroid hormone receptor Organohalogen Compounds 37 305 307
Matlai P Beral V 1985 Trends in congenital malformations of external genitalia [Letter] Lancet 1 8420 108 2857006
McDonald TA 2002 A perspective on the potential health risks of PBDEs Chemosphere 46 745 755 11999798
Meironyte D Noren K Bergman A 1999 Analysis of polybrominated diphenyl ethers in Swedish human milk. A time-related trend study, 1972–1997 J Toxicol Environ Health A 58 329 341 10580757
Nethersell AB Drake LK Sikora K 1984 The increasing incidence of testicular cancer in East Anglia Br J Cancer 50 377 380 6466547
Pike MC Chilvers CE Bobrow LG 1987 Classification of testicular cancer in incidence and mortality statistics Br J Cancer 56 83 85 3040061
Porterfield SP 1994 Vulnerability of the developing brain to thyroid abnormalities: environmental insults to the thyroid system Environ Health Perspect 102 suppl 2 125 130 7925183
Robb GW Amann RP Killian GJ 1978 Daily sperm production and epididymal sperm reserves of pubertal and adult rats J Reprod Fertil 54 103 107 712697
Sharpe RM Fisher JS Millar MM Jobling S Sumpter JP 1995 Gestational and lactational exposure of rats to xeno-estrogens results in reduced testicular size and sperm production Environ Health Perspect 103 1136 1143 8747020
She J Petreas M Winkler J Visita P McKinney M Kopec D 2002 PBDEs in the San Francisco Bay Area: measurements in harbor seal blubber and human breast adipose tissue Chemosphere 46 697 707 11999793
Tamasy V Meisami E Vallerga A Timiras PS 1986 Rehabilitation from neonatal hypothyroidism: spontaneous motor activity, exploratory behavior, avoidance learning and responses of pituitary—thyroid axis to stress in male rats Psychoneuroendocrinology 11 91 103 3704068
Theobald HM Peterson RE 1997 In utero and lactational exposure to 2,3,7,8-tetrachlorodibenzo-p -dioxin: effects on development of the male and female reproductive system of the mouse Toxicol Appl Pharmacol 145 124 135 9221831
Thiel R Chahoud I Schwabe R Neubert D 1989 Device for monitoring locomotor activity of 120 animals: motility of offspring of dams exposed to haloperidol Neurotoxicology 10 621 628 2626220
Van Waeleghem K De Clercq N Vermeulen L Schoonjans F Comhaire F 1996 Deterioration of sperm quality in young healthy Belgian men Hum Reprod 11 325 329 8671218
Viberg H Fredriksson A Eriksson P 2002 Neonatal exposure to the brominated flame retardant 2,2′,4,4′,5- pentabromodiphenyl ether causes altered susceptibility in the cholinergic transmitter system in the adult mouse Toxicol Sci 67 104 107 11961222
Viberg H Fredriksson A Eriksson P 2003a Neonatal exposure to polybrominated diphenyl ether (PBDE 153) disrupts spontaneous behaviour, impairs learning and memory, and decreases hippocampal cholinergic receptors in adult mice Toxicol Appl Pharmacol 192 95 106 14550744
Viberg H Fredriksson A Jakobsson E Orn U Eriksson P 2003b Neurobehavioural derangements in adult mice receiving decabrominated diphenyl ether (PBDE 209) during a defined period of neonatal brain development Toxicol Sci 76 112 120 12915714
Yucesan S Dindar H Olcay I Okur H Kilicaslan S Ergoren Y 1993 Prevalence of congenital abnormalities in Turkish school children Eur J Epidemiol 9 373 380 8243591
Zenick H Clegg ED 1989. Assessment of male reproductive toxicity: a risk assessment approach. In: Principles and Methods of Toxicology (Hayes W, ed). New York:Raven Press, 275–309.
Zhou T Ross DG DeVito MJ Crofton KM 2001 Effects of short-term in vivo exposure to polybrominated diphenyl ethers on thyroid hormones and hepatic enzyme activities in weanling rats Toxicol Sci 61 76 82 11294977
Zhou T Taylor MM DeVito MJ Crofton KM 2002 Developmental exposure to brominated diphenyl ethers results in thyroid hormone disruption Toxicol Sci 66 105 116 11861977
|
15687051
|
PMC1277857
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 4; 113(2):149-154
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7421
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7417ehp0113-00015515687052ResearchArticlesA Revised Probabilistic Estimate of the Maternal Methyl Mercury Intake Dose Corresponding to a Measured Cord Blood Mercury Concentration Stern Alan H. Division of Science Research and Technology, New Jersey Department of Environmental Protection, Trenton, New Jersey, USA; and Division of Environmental and Occupational Health, University of Medicine and Dentistry of New Jersey–School of Public Health, Piscataway, New Jersey, USAAddress correspondence to A.H. Stern, Division of Science Research and Technology, New Jersey Department of Environmental Protection, 401 E. State St., Trenton, NJ 08625 USA. Telephone: (609) 633-2374. Fax: (609) 777-2852. E-mail:
[email protected] work was supported, in part, under a contract with the U.S. Environmental Protection Agency (3W-1182-NAGX).
The author declares he has no competing financial interests.
2 2005 4 11 2004 113 2 155 163 14 7 2004 3 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. In 2001, the U.S. Environmental Protection Agency (EPA) adopted a revised reference dose (RfD) for methyl mercury (MeHg) of 0.1 μg/kg/day. The RfD is based on neurologic developmental effects measured in children associated with exposure in utero to MeHg from the maternal diet. The RfD derivation proceeded from a point of departure based on measured concentration of mercury in fetal cord blood (micrograms per liter). The RfD, however, is a maternal dose (micrograms per kilogram per day). Reconstruction of the maternal dose corresponding to this cord blood concentration, including the variability around this estimate, is a critical step in the RfD derivation. The dose reconstruction employed by the U.S. EPA using the one-compartment pharmacokinetic model contains two areas of significant uncertainty: It does not directly account for the influence of the ratio of cord blood:maternal blood Hg concentration, and it does not resolve uncertainty regarding the most appropriate central tendency estimates for pregnancy and third-trimester–specific model parameters. A probabilistic reassessment of this dose reconstruction was undertaken to address these areas of uncertainty and generally to reconsider the specification of model input parameters. On the basis of a thorough review of the literature and recalculation of the one-compartment model including sensitivity analyses, I estimated that the 95th and 99th percentiles (i.e., the lower 5th and 1st percentiles) of the maternal intake dose corresponding to a fetal cord blood Hg concentration of 58 μg/L are 0.3 and 0.2 μg/kg/day, respectively. For the 99th percentile, this is half the value previously estimated by the U.S. EPA.
cord bloodmaternalmercurymethyl mercuryMonte Carloone-compartmentpharmacokineticprobabilisticreference doseRfD
==== Body
In 2001, the U.S. Environmental Protection Agency (EPA) adopted a revised reference dose (RfD) for methyl mercury (MeHg) of 0.1 μg/kg/day (U.S. EPA 2001a, 2001b; Rice et al. 2003), relying heavily on the assessment conducted by the National Research Council (NRC 2000). The RfD is based on neurologic developmental effects measured in children associated with exposure in utero to MeHg from the maternal diet. The NRC and U.S. EPA assessments employed a benchmark dose approach to derive the lower 95% confidence interval on the fetal cord blood mercury concentration, doubling the proportion of children in the lowest 5% of performance on tests of neurologic performance. The NRC identified a cord blood concentration of 58 μg/L (ppb) total Hg based on analysis of the individual test judged to give the most sensitive and robust response, whereas the U.S. EPA identified a range of cord blood concentrations of 46–79 μg/L based on consideration of several tests. These values are in fact concentrations, whereas the RfD is a dose—in this case, the maternal intake dose. The reconstruction of the maternal MeHg intake dose that resulted in the observed cord blood Hg concentration is a critical step in the RfD derivation. The dose reconstruction requires a pharmacokinetic model linking dose and blood concentration. Two different types of pharmacokinetic models have been used for this purpose—a physiologically based pharmacokinetic (PBPK) model (Clewell et al. 1999) and the one-compartment model (Stern 1997; Swartout and Rice 2000). In both types of models, the relationship between cord blood concentration and dose depends on several physiologic and metabolic parameters. The values of these parameters vary among individuals. The population variability in the value of these parameters results in variability in the output of these models. Thus, there is no unique relationship between a given cord blood Hg concentration and a single maternal intake dose. Rather, this relationship is described by a probability distribution. For the RfD to be appropriately protective and inclusive of the variability in the population, the estimate of intake dose must itself be a distribution that describes this variability. This requires that the inputs to these pharmacokinetic models be in the form of probability distributions. The calculation of the outputs of these models using probability distributions has been accomplished through the use of Monte Carlo simulation. Both the PBPK and pharmacokinetic models for MeHg have been analyzed in this manner and have yielded estimates of variability in maternal dose reconstruction that are quite comparable. These include two separate analyses of the one-compartment model (Stern 1997; Swartout and Rice 2000) and one analysis of the PBPK model (Clewell et al. 1999; Stern et al. 2002).
Despite the close agreement regarding variability among these analyses, significant uncertainty remains regarding the appropriate central tendency estimates (e.g., means, medians) for the model parameters and, consequently, for the output of the model. This uncertainty results from the differing assumptions and different data employed among these three analyses (NRC 2000). In part, these differences result from uncertainty as to whether these parameters need to reflect conditions during pregnancy and from lack of specificity as to the period of pregnancy. Recognizing this source of uncertainty, the NRC (2000) assessment recommended separating the central tendency and variability aspects of the dose reconstruction. This approach was adopted by the U.S. EPA (2001b). In this approach, nondistributed (i.e., single value) central tendency estimates were selected for each parameter of the one-compartment model. The resulting output of the model represented a central tendency estimate of the maternal intake dose. This value was divided by an uncertainty factor derived from the analysis of variability generated from the probabilistic analysis of the distributions of the parameters of the one-compartment model. The value of the uncertainty factor was normalized to the central tendency estimate based on the ratio of the 50th percentile to the 1st percentile (i.e., lower 99th percentile) of the distribution of the dose derived from variability analysis. The resulting value is an estimate of the dose that accounts for 99% of the interindividual variability in the dose reconstruction (NRC 2000; Stern et al. 2002). Although this approach makes the uncertainty in the selection of the central tendency estimates explicit, it does not resolve that uncertainty.
Another significant source of uncertainty in the dose reconstruction was the failure to adequately incorporate the cord blood:maternal blood Hg ratio. The one-compartment model specifically estimates the relationship between maternal MeHg intake and maternal blood Hg concentration. To extend this relationship to fetal cord blood Hg concentration, it is necessary to estimate the maternal blood concentration corresponding to a measured cord blood concentration. This is done by applying an empirically derived ratio. In the NRC (2000) assessment, evaluation of a limited set of data led to the suggestion that this ratio was close to 1.0. On the basis of that conclusion, the benchmark cord blood Hg concentration was assumed to adequately describe the corresponding maternal blood Hg concentration. A preliminary analysis of the literature by the U.S. EPA (2001) suggested that the correct value for this ratio was likely to be > 1.0. The U.S. EPA RfD derivation thus included the uncertainty in the cord blood:maternal blood Hg ratio as a factor in its overall uncertainty factor adjustment. Nonetheless, the U.S. EPA did not identify a central tendency estimate for the cord:maternal ratio and did not account for the variability around that estimate. Thus, the appropriate contribution of this ratio to the overall RfD remains an unresolved source of uncertainty.
In an attempt to resolve both of these sources of uncertainty and to examine the extent to which a more complete analysis might alter the value currently employed in the RfD derivation, I have undertaken a reanalysis of the dose reconstruction. I have revisited the scientific literature and have reevaluated both central tendency estimates and overall distributions for each of the input parameters in the one-compartment model with a specific emphasis on pregnancy and, more specifically, on third-trimester–specific values. I have also incorporated the results of a recent analysis of the distribution of the cord blood:maternal blood Hg ratio (Stern and Smith 2003) into the overall analysis of the maternal dose reconstruction.
The One-Compartment Model
For consistency with the existing U.S. EPA RfD approach for MeHg, in this analysis I focus on the one-compartment model rather than the PBPK model for dose reconstruction. The one-compartment model is used in contrast to the PBPK model because it employs a relatively small set of parameters whose distributions within the population are generally well characterized. This facilitates a more accurate and transparent assessment of variability in the dose estimate. The one-compartment model predicts the relationship between MeHg intake dose and MeHg blood concentration under steady-state conditions in a closed system—in this case, the maternal system. The model is expanded to address the relationship between maternal intake dose and fetal cord blood by applying an empirically derived ratio to estimate the maternal blood concentration corresponding to the measured cord blood concentration.
The one-compartment model [NRC 2000; World Health Organization (WHO) 1990] as modified to include the cord blood:maternal blood Hg ratio, can be expressed as
where D is the maternal intake dose of MeHg (micrograms per kilogram per day), C is the measured Hg concentration in cord blood (for this analysis, the cord blood concentration is assumed to be the BMD (benchmark dose) value of 58 μg/L identified by the NRC (2000), R is the ratio of cord blood Hg concentration/maternal blood Hg concentration (unitless), b is the rate constant for elimination of MeHg from the blood (per day), V is the maternal blood volume (liters), W is the maternal body weight (kilograms), A is the fraction of the ingested dose that is absorbed (unitless), and F is the fraction of the absorbed dose that is present in the blood at steady state (unitless).
Materials and Methods
I searched the scientific and medical literature for data relating to the parameters of the one-compartment model. Priority was given to data for pregnant women, particularly data specific to the third trimester of pregnancy. To generate both central tendency estimates and descriptions of the distributions of the parameters, data were selected only if they were available as subject-specific values, summary percentiles, or graphic representations from which distributional data could be deduced (e.g., histograms). When two different appropriate data sets were available for a model parameter, the determination of whether to combine the data sets was based on a test (t-test and/or the nonparametric Mann-Whitney test as appropriate) of the hypothesis that the two data sets arose from the same underlying distribution. Data were tested for fit to parametric distributions (generally normal or log-normal) using curve-fitting software (BestFit for Windows, version 2.0d; Palisade Corp., Newfield, NY). In the absence of reasonable fits to parametric distributions, distributions were described by empirical distributions (e.g., cumulative probability distributions). Probabilistic (Monte Carlo) analysis of the one-compartment model was carried out using @Risk for Windows (version 3.5.2; Palisade Corp.). Latin hypercube sampling was employed to provide adequate representation of the tails of distributions. Sampling was accomplished with 5,000 iterations because this gave good interiteration stability in the moments of the output distribution. Results of the model simulations are based on the average of five separate simulations of 5,000 iterations each. Sensitivity analyses of central tendency estimates and variability results were carried out as indicated in the text. Unless otherwise indicated, all statistical analyses were carried out using Statistica (version 6.1; StatSoft Inc., Tulsa, OK).
Derivation of Probability Distributions for Model Parameters
Table 1 summarizes the probability distributions derived for each of the parameters in the one-compartment model. The detailed rationale for each parameter follows. In each case, an estimate of the true uncertainty (i.e., lack of knowledge) in the derivation of each parameter is provided. Consistent with the lack of knowledge, this estimate is semiquantitative using a relative scale of high, medium, and low uncertainty, based on professional judgment. This estimate is then employed in the subsequent sensitivity analysis.
C—the measured concentration of Hg in fetal cord blood.
This is an empirically derived constant. It is a biomarker of fetal MeHg exposure derived from the benchmark dose analysis relating neurodevelopmental performance to the measured concentration of total Hg in cord blood (NRC 2000). Although the benchmark dose analysis is subject to uncertainty, this value, once derived, is carried forward as the point of departure in the pharmacokinetic analysis. The uncertainty in this value is dealt with elsewhere in the overall risk assessment. The NRC (2000) identified a value of 58 μg/L based on its selection of the most appropriate test of developmental performance from the Faroe Islands study. The U.S. EPA (2001a, 2001b) identified a range of values bracketing the NRC selection from several tests in that study. To simplify this analysis, the single value of 58 μg/L is used. Because this value is a constant, the variability in the dose estimate is unaffected by the specific value selected, and the central tendency estimate of the dose scales linearly with the selected value. This value reflects measurement of the concentration of total Hg. As discussed below, and in detail by Stern and Smith (2003), the correct value for the purposes of dose reconstruction is the concentration of MeHg + MeHg-derived inorganic Hg. This value is closely estimated by the concentration of total Hg in cord blood.
R—the cord blood:maternal blood Hg ratio.
The derivation of this value is discussed in detail by Stern and Smith (2003). Briefly, the ratio was calculated from 10 separate studies meeting the selection criteria. The mean and standard deviation of the ratio from each study were calculated from the raw data, or estimated by probabilistic simulation, and showed a generally close agreement across studies for both parameters. Each study was described as the corresponding log-normal distribution, and the distributions from the individual studies were sampled in a meta-analysis to generate a summary distribution. The recommended summary distribution, mean ± SD = 1.7 ± 0.93, is used in this analysis. This ratio has historically been expressed as cord blood Hg/maternal blood Hg. However, the one-compartment model requires the reciprocal form. To avoid confusion, the historical form of the ratio is retained, and its reciprocal (1/R) is calculated in the model.
Based on the analysis by Stern and Smith (2003), an estimate of low true uncertainty is applied to this parameter.
b—the elimination rate constant for MeHg in maternal blood.
As noted above for C, total Hg rather than MeHg was measured in cord blood. Smith et al. (1994) and Smith and Farris (1996) point out that if the half-life in the blood of a dose of MeHg is measured as MeHg per se, it will be shorter than if it is measured as total Hg. This occurs because, even though the inorganic metabolite is not rapidly cleared from the blood, it is no longer measured as MeHg. The authors therefore suggest expressing the elimination rate of MeHg from the blood based on MeHg-specific measurements. This is appropriate if the rate of MeHg elimination is employed to directly estimate the period of time during which MeHg is present in the blood to exert a toxic effect (i.e., area-under-the-curve considerations). In this analysis, however, the elimination rate of MeHg is employed to calculate the steady-state balance between the ingestion of MeHg and its elimination. For this purpose, the inherent toxicity of MeHg does not enter into consideration. Thus, the appropriate metric is the rate of elimination of all the Hg that entered the blood as ingested MeHg. In maternal blood, this is closely estimated by the concentration of total Hg (NRC 2000; Stern and Smith 2003).
Three studies report half-lives/elimination rates measured directly in blood from nonpregnant females as well as males, Sherlock et al. (1984), Miettinen et al. (1971), and Kershaw et al. (1980). In the study by Miettinen et al. (1971), five men and one woman ate a single fish meal containing added 203Hg. The total mass of Hg in the fish meal, including endogenous Hg, was 22 μg. Neither the dose nor the body weight of the subjects is presented. However, if a body weight of 70 kg is assumed, the dose can be estimated at 0.3 μg/kg. The mean half-life was 49.87 days. Kershaw et al. (1980) administered an intravenous dose of MeHg ranging from 18.1 to 21.8 μg/kg/day to four men. The mean half-life was 52.98 days. Al-Shahristani and Shihab (1974) measured the decrease in Hg concentration along hair strands of 48 subjects of both sexes and ages from 6 months to 66 years in Iraq who were defined as “patients” as a result of the MeHg poisoning epidemic 8–12 months previously. The mean half-life was estimated to be 72 days with a wide range (36–189 days). The maximum hair concentrations are not provided, and the corresponding intake doses cannot be estimated from the available data. However, given that the subjects were still recognized as patients up to a year after the exposure, it is likely that their exposures were quite high. Because of the likelihood that the exposures resulted in frank toxicity, and given the inability to address the extent to which these elevated exposures may have affected the half-life (see below), the data from Al-Shahristani and Shihab (1974) were not further considered.
Cox et al. (1989) presented data from the Iraqi poisoning episode for 55 women (of the initial cohort of 83) who were exposed to MeHg-treated grain while pregnant and whose exposure was sufficiently large to allow x-ray fluorescence analysis of Hg in single hair strands. The decrease in Hg concentration was followed in their hair after they eliminated the MeHg-treated grain from their diets. Data on individual elimination rate as a function of exposure are not provided, but it is reported that hair concentration ranged from 10 to 670 ppm. This range would correspond to a steady-state daily intake of approximately 1.2–79.6 μg/kg/day assuming a 62-kg woman (NRC 2000). The population mean dose is not given. Data were presented for 122 hair samples from the 55 subjects. For generating summary and distributional statistics, it was assumed that each subject contributed equally to the total number of analyzed hair strands. Data were presented by Cox et al. (1989) as a histogram with the half-life binned in 5-day intervals. The frequency for the midpoint value for each half-life bin was read off the graph. The mean half-life for the 122 samples was 47.17 days. Although these data do not specifically represent third-trimester conditions, they are the only elimination rate data reported for a pregnant population and therefore appear to be the most applicable to the present analysis. However, the elevated level of exposure in this population raises concerns that the elimination rate may vary as a function of dose. Sherlock et al. (1984) present data on half-life and dose for 20 subjects given an experimental dose of MeHg. The subjects ingested MeHg in prepared fish meals two to four times per week over a 96-day period. The mean half-life was 50.20 days. The doses varied within a relatively small range from 0.5 to 3.6 μg/kg/day (mean = 1.6 μg/kg/day). The mean dose in the Sherlock et al. study is encompassed by the estimated dose range among the subjects in the Cox et al. (1989) study. It is clear, however, from the Cox et al. report that a significant fraction of the study population had exposures in the upper end of the range. In the Sherlock et al. (1984) data set, the half-life is significantly correlated with the dose (rSpearman = 0.52). Based on linear regression, the half-life increased by 3.21 days for each microgram per kilogram per day increase in dose (p = 0.04). The nominal half-life at zero dose (i.e., the y-intercept) is 45 days.
As shown in Table 2, with the exception of the Cox et al. (1989) data, the mean half-lives of the studies for which the dose is known or can be derived increase with increasing MeHg dose. This is consistent with the relationship observed in the Sherlock et al. (1984) data. Thus, although the upper range of exposures in Cox et al. (1989) exceeds the doses in the other studies by a factor of at least four, that study yields the smallest half-life, with a mean value close to that predicted at zero dose from the Sherlock et al. (1984) data. Note, however, that the Cox et al. (1989) estimate is specific to the period of pregnancy. During pregnancy, MeHg crosses the placenta and is retained in the fetal compartment. Given a mean cord blood:maternal blood ratio of 1.7 (Stern and Smith 2003), the transfer of MeHg to the fetus is, in essence, an additional route of maternal elimination and would be expected to result in a shorter half-life in pregnant women than in nonpregnant adults. Nonetheless, maternal elimination of MeHg across the placenta does not preclude a concentration dependence for other routes of maternal elimination. There does not appear to be any reliable way to account for the relative influence of each of these competing processes and thus to quantitatively adjust the half-life estimate from Cox et al. (1989) to account for concentration dependence. The Cox et al. data are therefore selected as the most appropriate estimate of half-life (and the elimination rate constant) during pregnancy with the caveat that the true value of the central tendency for lower doses may be smaller than the reported value. It should, however, be noted that, given the relatively wide range of exposures represented in this data set, the estimate of variability may be less uncertain than the central tendency estimate.
As shown in Figure 1, these data are not closely fit by parametric distributions, and there is a suggestion of a bimodal distribution. They are therefore described by an empirical distribution whose minimum and maximum are calculated as mean ± 3 SD.
The pregnancy-specific nature of the Cox et al. (1989) half-life data, and its lower mean value, consistent with fetal transport, implies that the data are relatively specific. The methodology of tracking elimination through hair segments is also relatively accurate. On the other hand, the possibility that the central tendency estimate is biased high due to the high dose to this population produces some uncertainty in this estimate. An overall estimate of medium true uncertainty is therefore assigned to this parameter.
V—the maternal blood volume.
Given the half-life of MeHg in maternal blood, the concentration in cord blood reflects maternal exposures during the last half of the third trimester (NRC 2000). Therefore, the most appropriate measure of maternal blood volume (V) is the blood volume during the third trimester of pregnancy. Blood volume increases during pregnancy starting at the sixth to eighth week and reaching a maximum at 32–34 weeks with a volume about 40% larger than the nonpregnant state (Fredriksen 2001).
Limited data on third-trimester blood volumes in the United States from which estimates of population variability can be derived are available from early work. Only data generated < 30 days before delivery were considered for the purposes of this analysis. Thomson et al. (1938) present data on third-trimester blood volume for 11 women (mean ± SD = 5.42 ± 0.92 L). Caton et al. (1951) present third-trimester data for 10 women (mean ± SD = 5.74 ± 0.97 L). Analysis of these data sets indicates that they are consistent with the hypothesis that both arise from the same underlying distribution (pt-test = 0.44). They are therefore combined into a single data set (n = 21, mean ± SD = 5.57 ± 0.93 L). Although these data are relatively old, Hytten (1985), in summarizing the more recent state of knowledge of blood volume at 40 weeks of pregnancy, reported a value of 5.50 L. The earlier studies thus appear to compare well with more recent data. Thomson et al. (1938) also report the measured body weight corresponding to the blood volumes. Third-trimester blood volume and body weight are moderately correlated (rSpearman = 0.49), but the correlation is marginally significant (p = 0.13). Body weight and blood volume are known to be correlated in general (Fredriksen 2001; Hytten 1985). Therefore, the lack of significance for this correlation in the Thomson et al. (1938) data likely reflects the small sample size rather than the absence of a meaningful correlation.
As seen in Figure 2, the combined data set is not well described by parametric distributions, and there is a suggestion of polymodality. These data are therefore described as an empirical distribution, with minimum and maximum values chosen as mean – 2 SD and mean + 2.5 SD, respectively. The nonsymmetrical bounds are chosen to accommodate the right-skewness of the data. In the probabilistic analysis, samples of V and maternal body weight at delivery (W) are mutually constrained by the correlation coefficient derived from the Thomson et al. (1938) data.
These data are highly specific to the target population. Although the combined data set for this parameter is relatively small, the overall variance is also small (coefficient of variation = 0.17), as would be expected given the physiologic constraints on blood volume. Thus, it is not likely that a larger sample would significantly change the parameters of the resulting distribution. Despite the age of the data, they appear consistent with more contemporary estimates. This parameter is therefore judged to have a low true uncertainty.
A—the fraction of the dose that is absorbed.
This parameter is a measure of the fraction of the mass of ingested MeHg that is absorbed into the body as a whole. As such, it cannot be measured directly by sampling of blood but requires either whole-body counting of labeled MeHg or long-term monitoring of Hg2+ and MeHg excretion in the absence of prior or subsequent confounding exposures. Consequently, there are few human data for this parameter. Miettinen et al. (1971) fed measured portions of fish containing a known activity of radio-labeled MeHg to 14 adult subjects, six females and eight males. Whole-body counting was used to measure the amount of activity remaining after 3–4 days, and urine and feces counting was used to measure the excreted activity during that period. Miettinen et al. (1971) used these measurements to estimate the fraction of the intake dose that was retained/absorbed. Measurements were taken after the 3–4 day period presumably to allow elimination of MeHg that had become associated with the lining of the gastrointestinal tract but not absorbed into the body proper. The mean retention was estimated at 94%. However, during this 3–4 day period, some fraction of the absorbed dose was metabolized to Hg2+ and excreted. The original Miettinen et al. (1971) estimate does not account for this metabolism and thus slightly underestimates the correct value of this parameter. Previous assessments of the one-compartment model (Stern 1997; Swartout and Rice 2000) used this original Miettinen et al. estimate. However, Miettinen et al. (1971) also provide individual estimates of the whole-body half-life of MeHg. This allows an estimate of the fraction of the absorbed activity that was eliminated during the 3–4 days due to metabolism. The corrected value of A is estimated as A = D/(M + U), where D is the administered MeHg activity, M is the administered MeHg lost to metabolism, and U is the administered activity that is retained at 3–4 days.
The value of A estimated using this approach (mean ± SD = 0.97 ± 0.016) is 3.4% larger than the original estimate. The mean values for males and females are almost identical, and the entire data set is therefore used to estimate the distribution of A. Aberg et al. (1969) also reported that “almost 100%” of administered MeHg label was absorbed. Although the available data do not relate to pregnancy, given the absorption of close to 100% in the nonpregnant state, it does not seem likely that absorption during pregnancy would be substantially altered.
The data set is not well fitted by any parametric function, and there is some suggestion of a bimodality, with a secondary peak at > 99% absorption (Figure 3). The data are therefore fit to an empirical cumulative probability distribution. Given the narrow range of the data, the minimum and maximum values are selected empirically (0.940 and 0.999, respectively).
Although the available data are not third-trimester– or pregnancy-specific, they are precise and describe a small range of variability. Given the likelihood that these values, apparently resulting from simple uptake phenomena, would not change as a result of the pregnant state, this parameter is judged to have low true uncertainty.
F—the fraction of the absorbed dose that is present in the blood at steady state.
This parameter refers to the fraction of the absorbed dose of MeHg found in the maternal blood after distribution among the various tissues. Functionally, it is derived from the same data as the elimination rate constant by back extrapolating the linear fit of the relationship between time (T) and the log of the Hg concentration (or radio-tracer activity) in the blood to T = 0. This gives the concentration (or activity) in the blood at the theoretical point at which MeHg in blood had equilibrated with other tissues, but before it started to be eliminated from the blood. This value, however, must be converted to the mass of MeHg in the total blood volume and then expressed as a fraction of the absorbed mass of MeHg. Thus, F is calculated from
where CT0 is the concentration of MeHg in blood at T = 0 (micrograms per liter), V is the blood volume (liters), and D is the absorbed dose (micrograms).
Sherlock et al. (1984), Kershaw et al. (1980), and Smith et al. (1994) do not report blood volume. Although Kershaw et al. (1980) and Smith et al. (1994) report values for F, it appears that blood volumes used in their calculations were based on an assumed deterministic relationship between body weight on blood volume (Allen PV, personal communication). In the case of Kershaw et al. (1980), it appears that blood volume was calculated as 7% of body mass in kilograms. Individual values of F for the nonpregnant state can be estimated from Sherlock et al. (1984) based on their data for the percentage of intake dose distributed in 1 kg of blood by assuming a deterministic relationship between body weight and blood volume. Thus, for consistency with the approach in Kershaw et al. (1980), blood volume was estimated for the Sherlock et al. (1984) data by assuming that blood volume (L) is equal to 7% of body weight (kilograms). However, estimation of F using such an approach underestimates the overall variability in F. For the Sherlock et al. (1984) data, a further deterministic assumption must be made to estimate the volume corresponding to 1 kg of blood (i.e., blood density). This introduces an additional small loss in variability in the distribution of F. Standard values for blood density are generally given as 1.06 kg/L (Reinking 2002; U.S. EPA 2002). Miettinen et al. (1971), although providing activity versus time data, provided no data on blood volume or body weight. These data cannot therefore be used to estimate F without an arbitrary assumption of blood volume.
The Sherlock et al. (1984) data show a moderately negative correlation between the fraction of the dose in 1 kg of blood and body weight (r = –0.47, p = 0.04). The authors suggest that this is a function of the larger volume of distribution to be expected with larger body masses. Swartout and Rice (2000) included this correlation in their simulation of the one-compartment model. This explanation is consistent with the positive correlation between body weight (W) and blood volume (V) of approximately the same magnitude derived from the data of Thomson et al. (1938). However, note that the relationship with body weight derived from Sherlock et al. (1984) refers to the percentage of the intake found in 1 kg of blood. As body weight increases, the total blood volume also increases, and the fraction of the total blood Hg that is found in a fixed volume (or mass) of blood (e.g., 1 kg) decreases proportionally. This will be the case even if the fraction of the MeHg intake found in the total blood volume remains constant. Therefore, the correlation observed by Sherlock et al. (1984) does not appear to reflect a relationship that is independent of the correlation between body weigh and blood volume. Thus, having already specified that correlation in the model inputs for W and V, it does not appear necessary to additionally specify a correlation between W and F.
In both Kershaw et al. (1980) and Smith et al. (1994), all subjects were men. In the Sherlock et al. (1984) study, 6 of the 20 subjects were women. The calculated values of F for the women (mean = 4.59%) were significantly less than those for the men (mean = 5.58%; pt-test = 0.014; pMann-Whitney = 0.011). However, multiple regression analysis of F shows that when both sex and body weight are included as independent variables, weight is significant (p = 0.03) but sex is not. Thus, sex does not appear to have an influence on F independent of body weight. The values of F calculated from Kershaw et al. (1980; mean = 5.63%) and Sherlock et al. (1984) (mean = 5.28%) are clearly different from those from Smith et al. (1994) (mean = 7.7%; pKruskal-Wallis = 0.005). The Sherlock et al. (1984) and Kershaw et al. (1980) data, on the other hand, are not statistically different (pMann-Whitney = 0.35), the small size of the Kershaw et al. (1980) data set notwithstanding. Therefore, although the Sherlock et al. (1984) and Kershaw et al. (1980) data sets for F appear compatible, the Smith et al. (1994) data are incompatible. Differences in body weight between the Smith et al. subjects and those of Sherlock et al. or Kershaw et al. do not appear to account for this difference, and the reason for this difference is not clear. However, it is possible that the unknown relationship employed by Smith et al. (1994) to estimate blood volume from body weight may contribute to this difference. In light of the good agreement between the Sherlock et al. (1984) and Kershaw et al. (1980) estimates of F, these data sets are selected as the basis for a distribution for this parameter.
The combined Sherlock et al. (1984) and Kershaw et al. (1980) data are consistent with both normal and log-normal distributions by standard goodness-of-fit tests. There is little difference in how well these data are fitted by these distributions. This may be a function of the relatively small size of the combined data set (n = 24). In the absence of further indications, the normal distribution is chosen to avoid extreme values in the more elongated upper tail of the log-normal distribution. Figure 4 shows the fit of the maximum likelihood normal distribution to these data.
Given the lack of pregnancy-specific data for this parameter, the multiple uncertainties in deriving the parameter from the available data, the incompatibility of one of the three available data sets, and the uncertainty in the specification of the distributional form, this parameter is assigned a high true uncertainty estimate.
W—Maternal body weight.
Although cord blood Hg concentration largely reflects maternal intake during the third trimester, the changing nature of maternal body weight during this period and the lack of a unique reference point during this period, other than delivery, make maternal weight at delivery the most appropriate measure of maternal weight for this analysis. Directly measured population-based data on maternal weight at delivery for the U.S. population are not available. The Centers for Disease Control and Prevention (CDC) does, however, collect individual-specific data from 19 participating states on weight gain during pregnancy and prepregnancy weight as part of their Pregnancy Risk Assessment Monitoring System (PRAMS) database (CDC 2004b). Data for both parameters are partly self-reported. Maternal weight at delivery is calculated as the sum of these two quantities. Although the databases for weight gain and prepregnancy weight are intrinsically linked, corresponding data for individuals were available only through internal CDC codes. Linked data were therefore provided directly in tabulated form (Whitehead N, personal communication). These data are given in Table 3.
Although these data are specific to delivery, the fact that they result partly from self-reported information produces some uncertainty. In addition, they represent data from only 19 of the 50 U.S. states. There is also a suggestion of some uncertainty in the tails of the distribution. Overall, this parameter is associated with a medium degree of true uncertainty.
Results
Model Outputs
Table 4 summarizes the distribution of the values estimated for D, the maternal intake dose corresponding to 58 μg Hg/L cord blood. Values are the mean of five separate simulations. The last two rows in Table 4 express the variability in the estimate of the maternal dose relative to its central tendency estimate. The ratio of the 50th percentile estimate to the 5th percentile estimate reflects the “distance” between the median maternal dose and the maternal dose at which 95% of the fetal population is predicted to achieve 58 μg/L (i.e., the lower 5% estimate of dose). Likewise, the ratio of the 50th percentile to the 1st percentile reflects the distance between the median maternal dose and the maternal dose at which 99% of the fetal population is predicted to achieve 58 μg/L (i.e., the lower 1% estimate of dose). Based on this analysis, the lower 5% estimate of dose is about a factor of 3 lower than the median dose, and the lower 1% estimate of dose is a factor of 4 lower than median dose.
Sensitivity Analyses
The derivation of the distributions for each of the input parameters is associated with varying degrees of true uncertainty (i.e., lack of knowledge). In the context of this analysis, this includes both the uncertainty in estimating the true values of the central tendency and percentiles of the input parameters, and to the uncertainty in specifying their correct distributional form. In theory, true uncertainty is reducible through the use of more or better data. On the other hand, variability is an inherent property of the data and is not reducible through more or better data.
Sensitivity analysis of variability.
Table 5 presents a sensitivity analysis of the contributions of each of the model input parameters to the variability in the model output. This is accomplished in the Monte Carlo simulation by setting each model parameter, in turn, to its fixed (point estimate) mean value—that is, by removing each parameter’s contribution to the output variability. The resulting model output is then compared with the model output reflecting the contribution of all of the parameters to the output variability. The magnitude of the percent difference between these outputs is a measure of the contribution of each parameter to the total variability. A negative percent difference indicates that the parameter functions in the full model to increase variability, and a positive difference indicates that the parameter functions to decrease variability. In this analysis, differences in the outputs are assessed using the ratio of the 50th percentile value to the 5th or 1st percentile value (i.e., the normalized lower 95th, and 99th percentiles of the maternal intake dose resulting in 58 μg/L Hg in cord blood). This reflects the variability in the portion of the output containing the sensitive population. Clearly, R makes the largest contribution to output variability, with much smaller contributions (in decreasing order) by b, F, and W. V and A make no significant contribution to the variability.
Table 5 also presents a summary of the estimates of true uncertainty for each of the input parameters as discussed in the preceding section. Considering the contributions of each parameter to output variability and true uncertainty together, R, the cord blood:maternal blood Hg ratio, is the parameter that clearly has the greatest influence on variability. However, it is associated with a low level of true uncertainty and thus makes little contribution to uncertainty in output variability. The elimination rate constant, b, makes the next largest contribution to variability, but its influence is only about 20% of R. It is associated with a moderate level of true uncertainty. Its overall contribution to uncertainty in the output variability is therefore relatively small. F, the fraction of the dose residing in the blood, is highly uncertain but makes only a small contribution to variability. Its contribution to uncertainty in output variability is therefore also relatively small. All other parameters add little or no uncertainty to the output variability.
Sensitivity analysis of central tendency.
If the central tendency estimates of the input parameters (i.e., means, medians) in this analysis are inaccurate, they could bias the estimates of the central tendency and percentiles of the model output even if the variability in the output distribution is accurately estimated. Because three of the input parameters are judged to have a low degree of true uncertainty (Table 5), it is unlikely that they will contribute significantly to uncertainty in the central tendency estimate of the output. Therefore, I focused on the variables with medium or high uncertainty. Because b and W are judged to have medium uncertainty, I assumed, for the purposes of this analysis, that the true mean value could vary from their original estimate by ± 10%. For F, with high uncertainty, it was assumed that the true mean value could vary by ± 20% from its original estimate. These parameters were therefore allowed to assume one of three fixed values: their original estimate of mean value, +10% (or 20% for F) of their original estimate, and –10% (or 20% for F) of their original estimate. The remaining parameters were fixed at the original estimates of their mean values. The model was calculated using Monte Carlo sampling (5,000 iterations) to yield the various possible combinations of mean maternal dose. Table 6 presents the results of this analysis.
The minimum and maximum changes reflect the outcomes where the central tendencies of each of the input parameters are simultaneously altered in the direction that produces the greatest change in the output. The values between the 25th and 75th percentiles reflect a more likely combination of altered input values resulting in a more likely estimate of the uncertainties in the central tendency estimates. Under the assumptions of this analysis, the true uncertainty in the central tendency estimates of the most uncertain input parameters is most likely to influence the estimate of maternal dose by ≤ 20%. In addition, because of the unexplained difference in the value of F estimated from the Smith et al. (1994) data compared with the combined Kershaw et al. (1980) and Sherlock et al. (1984) data, this analysis was rerun assuming that the maximum value of F could be 48% larger than the original central tendency estimate [i.e., equal to the mean value from the analysis in Smith et al. (1994)]. The results of this analysis (not shown) indicate that, as expected, the use of a larger value of F results in a smaller value at the lower end of the range of likely outcomes. The difference, however (a 28% decrease at the 25th percentile compared with the 20% decrease seen for the analysis presented in Table 6), is not dramatic and suggests that, in the context of the overall uncertainty in the estimate of central tendency, incorporating the Smith et al. (1994) value for F would have a relatively small effect on the central tendency estimate.
Discussion
This reanalysis of the maternal MeHg dose resulting in 58 μg Hg/L in fetal cord blood was undertaken to address two specific needs that have arisen since the original NRC and U.S. EPA analyses. The first was to incorporate the cord blood:maternal blood ratio in the analysis. This was accomplished by integrating the distribution of R derived in the recent analysis of Stern and Smith (2003). The second need was to reduce the uncertainty in the estimate of the central tendency of the maternal dose. This was accomplished by reassessing each of the input parameters in the one-compartment model. This also allowed a reassessment of the variability in the estimate of maternal dose. As expected based on both its central tendency estimate of 1.7 (as opposed to the original implicit estimate of 1.0) and the significant variability around this value (Stern and Smith 2003), the cord blood:maternal blood Hg ratio (R) had a significant influence on the estimate of maternal dose. In addition to proportionally decreasing the central tendency estimate, this parameter significantly increased the variability in the estimate, resulting in a 37% increase in “distance” between the median and the lower 5th percentile and a 49% increase in the distance between the median and the lower 1st percentile. The sensitivity analysis of variability shows that this parameter is the largest source of variability in the estimate of maternal dose.
Much of the previous uncertainty in the central tendency estimate of maternal dose resulted from the lack of temporal specificity in the model parameters regarding the period of pregnancy. In the present analysis, I made a specific effort to identify distributional data that are specific to pregnancy and, whenever possible, specific to the third trimester of pregnancy—the period of gestation corresponding most closely to the period of accumulation of measured Hg in cord blood. Appropriate pregnancy-specific data were identified for four of the six parameters in the model (excluding C, the empirical value selected as the point of departure). Third-trimester–specific data sets were identified for three of these four. As shown in Table 7, the two largest differences in central tendency values between the present analysis and the U.S. EPA analysis are for R and W. The value of W used in the present analysis is 21% larger than the value chosen by the U.S. EPA. The value used in the present analysis is based on recent data that are specific to delivery and are reasonably consistent with independent estimates of pregnancy weight gain and prepregnancy weight. R is entered into the numerator of the model as its reciprocal, and W appears in the denominator of the model. Thus, the effect of increasing the value of both of these parameters relative to the U.S. EPA estimates is to decrease the central estimate of the maternal dose. Based on the values in Table 7, the central tendency estimate of maternal dose corresponding to 58 μg Hg/L cord blood is 0.7 μg/kg/day. Note that this value differs from the mean value derived from the full distributional analysis (Table 4), because the equation for the model is nonlinear, and therefore, the mean of the means (i.e., the central tendency estimate) is not equivalent to the overall mean. The corresponding U.S. EPA central tendency estimate is 1.1 μg/kg/day (U.S. EPA 2001). When both R and W are set to their U.S. EPA central tendency values, the effect of the remaining central tendency values from the present analysis is to increase the dose estimate to 1.4 μg/kg/day. Thus, although individually the changes in b, V, F, and A are relatively small compared with their values in the U.S. EPA analysis, the combined effect of the changes is influential. Apart from the extent of change in the central tendency estimates, the increased specificity and in-depth evaluation presented here can reduce the level of uncertainty present in the existing U.S. EPA analysis.
In addition to the reassessment of the central tendency estimates, this analysis yields a revised analysis of the variability in the estimate of maternal dose. The analysis presented as part of the NRC report (2000) and discussed in Stern et al. (2002) concluded that dividing the central tendency estimate of dose by a factor of 2–3 would account for 95–99% of the variability in the dose estimate. Based largely on that analysis, the U.S. EPA chose a value for the 50th percentile:1st percentile ratio of maternal dose of 3.0 (U.S. EPA 2001). The central tendency estimate was divided by this value to obtain an estimate of the 1st percentile (i.e., lowest 99th percentile) of maternal intake dose. In the present analysis, a value of 4.0 is estimated for the 50th percentile:1st percentile ratio, and a value of 2.7 for the 50th percentile:5th percentile ratio (Table 4).
To address the uncertainties in the estimates of the central tendency of the individual model parameters, the U.S. EPA followed the recommendation of the NRC to decouple the central tendency and variability estimates of maternal dose. By doing so, the U.S. EPA could explicitly discuss central tendency values and separately fold the variability into the overall consideration of uncertainty factors. For a cord blood Hg concentration of 58 μg/L, this approach yields an estimate of the 1st percentile of maternal dose of 0.4 μg/kg/day (i.e., 1.1 μg/kg/day/3.0) based on the U.S. EPA’s central tendency and variability estimates. To the extent that the present analysis has substantially reduced the uncertainty in the estimates of central tendency, it can be argued that this decoupling is no longer necessary. Rather, I believe that it is now appropriate to directly estimate the lowest 1st (or any other) percentile of maternal dose corresponding to 58 μg Hg/L cord blood directly from the output of the one-compartment model. With reference to Table 4, a maternal intake dose of 0.2 μg/kg/day would correspond to the 1st percentile, and a dose of 0.3 μg/kg/day would correspond to the 5th percentile. Thus, the present analysis supports an estimate that is half of the value implicit in the U.S. EPA analysis. Following this approach, no uncertainty factor would be required to account for pharmacokinetic variability in the estimate of maternal dose in the calculation of the RfD. Other uncertainty factors could still be applied to the derivation of the RfD as appropriate to address database factors as well as the remaining fetal variability in the system including toxicodynamic factors.
Figure 1 Comparison of Cox et al. (1989) half-life data to maximum likelihood normal and log-normal distributions.
Figure 2 Combined third-trimester blood volume data from Thomson et al. (1938) and Caton et al. (1951) compared with maximum likelihood normal and log-normal distributions.
Figure 3 Comparison of Miettinen et al. (1971) data on the fraction of MeHg absorbed to maximum likelihood normal and log-normal distributions.
Figure 4 Comparison of combined Sherlock et al. (1984) and Kershaw et al. (1980) data on fraction of the dose in maternal blood to the maximum likelihood normal distribution.
Figure 5 Comparison of maternal delivery weight data (Whitehead N, personal communication) to the maximum likelihood log-normal distribution.
Table 1 Summary of distributions selected for parameters of the one-compartment model.
Distributions
Parameter Source Data Probability
C (concentration of Hg in cord blood) NRC 2000 58 μg/L
R (ratio of Hg concentration in cord blood to maternal blood Stern and Smith 2003 Log-normal
Mean ± SD = 1.7 ± 0.9
b [rate constant for elimination of MeHg from blood; note that this parameter is reported as half-life of MeHg in blood (T1/2), related to b as b = ln 0.5/T1/2] Cox et al. 1989 Empirical T1/2 (days) Relative probability
20 2.46
25 1.64
30 5.74
35 8.20
40 12.30
45 17.21
50 14.75
55 25.41
60 7.38
65 4.10
70 0.82
Minimum = 15
Maximum = 75
V (maternal blood volume) Thomson et al. 1938, Caton et al. 1951 Empirical (correlated with W) V (L) Cumulative probability
4.480 0.05
4.530 0.1
4.970 0.25
5.280 0.5
6.310 0.75
6.408 0.85
6.694 0.9
7.380 0.95
Minimum = 3.707
Maximum = 7.902
Correlated with W – r = 049
A (fraction of dose absorbed) Miettinen et al. 1971 (modified to account for metabolism) Empirical A Cumulative probability
0.947 0.071
0.960 0.286
0.971 0.5
0.983 0.786
0.996 0.929
Minimum = 0.940
Maximum = 0.999
F (fraction of absorbed dose that is present in the blood at steady state) Sherlock et al. 1984, Kershaw et al. 1980 Normal
Mean ± SD = 0.052 ± 0.0095
W (maternal body weight) CDC 2004b Log-normal
Mean ± SD = 80.9 ± 16.3 kg
Correlation with V – r = 0.49
Table 2 MeHg half-life and elimination rate as a function of intake dose.
Study No. Mean dose μg/kg/day (range) Mean T1/2 (days) Mean elimination rate constant (b)
Cox et al. 1989 55 (1.2–79.6)a 47.2 0.0147
Miettinen et al. 1971 6 0.3b 49.9 0.0142
Sherlock et al. 1984 20 1.6 (0.5–3.6) 50.2 0.0140
Kershaw et al. 1980 4 20.0 (18.1–20.9) 53.0 0.0133
a Estimated from reported hair Hg concentrations based on the one-compartment pharmacokinetic model, and assuming a 62-kg body weight.
b Estimated assuming uniform intake of the tracer dose and 70-kg body weight.
Table 3 Calculated percentiles of maternal weight (kg) at delivery (CDC 2004b; Whitehead N, personal communication).
Cumulative percentile Maternal weight at delivery (based on PRAMS data)
1st 52.66
5th 59.02
10th 63.00
20th 67.36
30th 71.11
40th 74.59
50th 77.91
60th 81.62
70th 86.23
80th 92.07
90th 102.15
95th 112.64
99th 135.93
Table 4 Model output of selected percentiles of the maternal intake dose of MeHg corresponding to 58 μg Hg/L cord blood.
Distribution of maternal MeHg intake dose (percentile) Maternal intake dose (μg/kg/day)
Mean ± SD 0.993 ± 0.702
1st 0.202
5th 0.301
10th 0.373
50th 0.812
50th/5th 2.700
50th/1st 4.020
Table 5 Estimated contributions of model parameters to variabilitya and true uncertaintyb in the model output.
Contribution R b V A F W
50th percentile/5th percentile −37.0 −8.3 < 0.1 +0.4 −5.9 −2.0
50th percentile/1st percentile −49.1 −9.8 +0.3 −1.4 −7.2 −2.1
True uncertainty L M L L H M
See text (“The One-Compartment Model”) for an explanation of the model paremeters.
a Percent change in normalized 5th and 1st percentile values for models with fixed parameters compared to the full variability model.
b Estimates of the degree of true uncertainty in the specification of the model parameters: H, high; M, medium; L, low.
Table 6 Sensitivity analysis of the estimates of the central tendency for selected model parameters.
Outcomes from the combination of alternative central tendency estimates for selected model parameters Estimated mean dose Percent change from original mean dose
Original mean maternal dose 0.71a —
Minimum alternative value 0.47 −34
Maximum alternative value 1.04 +46
25th percentile alternative value 0.57 −20
75th percentile alternative value 0.84 +18
Values represent the change in the mean maternal dose resulting from various combinations of high and low values for input parameters. See text (“Sensitivity analysis of central tendency”) for explanation of parameter selection.
a This value differs from the mean estimated from the full distributional analysis.
Table 7 Comparison of central tendency estimates in the present analysis to central tendency estimates in the U.S. EPA RfD derivation (U.S. EPA 2001).
Parameter Current centrala tendency estimate Pregnancy specific? Third-trimester specific? EPA central tendency estimate Pregnancy specific? Third-trimester specific?
R 1.7 Yes Yes 1.0 (implicit) No No
b 0.0147 day−1 (47 days) Yes No 0.014 day−1 (50 days) No No
V 5.6 Lb Yes Yes 5 Lc Yes Yes
W 80.9 kg Yes Yes 67 kg Yes No
A 0.97 No No 0.95 No No
F 0.052 No No 0.059 No No
a Means of fitted distributions (see Table 1).
b U.S. women.
c Nigerian women.
==== Refs
References
Aberg B Ekman L Falk R Greitz U Persson G Snihs J-O 1969 Metabolism of methyl mercury (203 Hg) compounds in man—excretion and distribution Arch Environ Health 19 478 484 5822994
Al-Shahristani H Shihab KM 1974 Variation of biological half-life of methylmercury in man Arch Environ Health 28 342 344 4829091
Caton WL Roby CC Reid DE Caswell R Maletskos CJ Fluharty RG 1951 The circulating red cell volume and body hematocrit in normal pregnancy and the puerperium Am J Obstet Gynecol 61 1207 1217 14838019
CDC 2002. Births: final data for 2001. Natl Vital Stat Rep 51(2). Available: http://www.cdc.gov/nchs/data/nvsr/nvsr51/nvsr51_02.pdf [accessed 24 June 2004].
CDC 2004a. NHANES 2001–2002 Data Files. Atlanta, GA:Centers for Disease Control and Prevention, National Center for Health Statistics. Available: http://www.cdc.gov/nchs/about/major/nhanes/nhanes01–02.htm [accessed 21 April 2004].
CDC 2004b. Pregnancy Risk Assessment Monitoring System. Atlanta, GA:Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion. Available: http://www.cdc.gov/reproductivehealth/srv_prams.htm#1 [accessed 12 October 2004].
Clewell HJ Gearhart JM Gentry PR Covington TR VanLandingham CB Crump KS 1999 Evaluation of the uncertainty in an oral reference dose for methylmercury due to interindividual variability in pharmacokinetics Risk Anal 19 547 558 10765421
Cox C Clarkson TW Marsh DO Amin-Zaki L Tikriti S Myers GG 1989 Dose response analysis of infants prenatally exposed to methyl mercury: an application of a single-compartment model to single-strand hair analysis Environ Res 49 318 332 2473897
Fredriksen MC 2001 Physiologic changes in pregnancy and their effect on drug disposition Semin Perinatol 25 120 123 11453606
Hytten F 1985 Blood volume changes in normal pregnancy Clinics Hematol 14 601 612
Kershaw TG Dhahir PH Clarkson TW 1980 The relationship between blood levels and dose of methylmercury in man Arch Environ Health 35 28 36 7189107
Miettinen JK Rahola T Hatula K Rissanen K Tillander M 1971 Elimination of 203 Hg-methylmercury in man Ann Clin Res 3 116 122 4997252
NRC (National Research Council) 2000. Toxicological Effects of Methylmercury. Washington, DC:National Academy Press.
Reinking L 2002. Composition, properties and functions of blood, In: Cardiopulmonary Physiology. Millersville, PA:Millersville University. Available: http://muweb.millersville.edu/~biology/biology.455/chapter2.doc [accessed 12 October 2004].
Rice DC Schoeny R Mahaffey K 2003 Methods and rationale for derivation of a reference dose for methylmercury by the U.S. EPA Risk Anal 23 107 115 12635727
Sherlock J Hislop J Newton D Topping G Whittle K 1984 Elevation of mercury in human blood from controlled chronic ingestion of methylmercury in fish Human Toxicol 3 117 131
Smith JC Allen PV Turner MD Most B Fisher HL Hall LL 1994 The kinetics of intravenously administered methyl mercury in man Toxicol Appl Pharmacol 128 251 256 7940540
Smith JC Farris FF 1996 Methyl mercury pharmacokinetics in man: a reevaluation Toxicol Appl Pharmacol 136 245 252 8661350
Stern AH 1997 Estimation of the interindividual variability in the one-compartment pharmacokinetic model for methylmercury: implications for the derivation of a reference dose Reg Toxciol Pharmacol 25 277 288
Stern AH Clewell HJ Swartout J 2002 An objective uncertainty factor adjustment for methylmercury pharmacokinetic variability Human Ecol Risk Assess 8 885 894
Stern AH Smith AE 2003 An assessment of the cord blood:maternal blood methylmercury ratio: implications for risk assessment Environ Health Perspect 111 1465 1470 12948885
Swartout J Rice G 2000 Uncertainty analysis of the estimated ingestion rates used to derive the methylmercury reference dose Drug Chem Toxicol 23 293 306 10711403
Thomson KJ Hershheimer A Gibson JG Evans WA 1938 Studies on the circulation in pregnancy Am J Obstet Gynecol 36 48 59
U.S. EPA 2001a. Methylmercury Reference Dose for Chronic Oral Exposure. U.S. Environmental Protection Agency, Integrated Risk Information System (IRIS). Washington, DC: U.S. Environmental Protection Agency. Available: http://www.epa.gov/iris/subst/0073.htm [accessed 22 April 2004].
U.S. EPA 2001b. Water Quality Criterion for the Protection of Human Health—Methylmercury. Final. EPA-823-R-01-001. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2002. System Requirements and Design for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK) Windows® Version—32-Bit Version, Appendix C, IEUBKwin Parameter Dictionary, EPA OSWER No. 9285.7-43. Washington, DC:U.S. Environmental Protection Agency.
WHO 1990. Methylmercury. Environmental Health Criteria 101. Geneva:World Health Organization, International Programme on Chemical Safety.
|
15687052
|
PMC1277858
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 4; 113(2):155-163
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7417
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7329ehp0113-00016415687053ResearchArticlesMetals in Urine and Peripheral Arterial Disease Navas-Acien Ana 123Silbergeld Ellen K. 4Sharrett A. Richey 1Calderon-Aranda Emma 45Selvin Elizabeth 12Guallar Eliseo 1231Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland, USA2Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins Medical Institutions, Baltimore, Maryland, USA3Johns Hopkins Center for Excellence in Environmental Public Health Tracking, Baltimore, Maryland, USA4Department of Environmental Health Sciences, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland, USA5Sección de Toxicología, Cinvestav, MéxicoAddress correspondence to E. Guallar, Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University Bloomberg School of Public Health, 2024 East Monument St., Room 2-639, Baltimore, MD 21205-2223 USA. Telephone: (410) 614-0574. Fax: (410) 955-0476. E-mail:
[email protected]. was supported in part by an American Heart Association Scientist Development Award (0230232N). A.N.-A. and E.G. were supported by the Johns Hopkins Center of Excellence in Environmental Public Health Tracking. E.S. was supported by National Heart, Lung, and Blood Institute grant T32 HL07024. E.K.S. and E.C.-A. were supported in part by National Institute of Environmental Health Sciences grant 1 R21 ES11717.
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 164 169 14 6 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Exposure to metals may promote atherosclerosis. Blood cadmium and lead were associated with peripheral arterial disease (PAD) in the 1999–2000 National Health and Nutrition Examination Survey (NHANES). In the present study we evaluated the association between urinary levels of cadmium, lead, barium, cobalt, cesium, molybdenum, antimony, thallium, and tungsten with PAD in a cross-sectional analysis of 790 participants ≥40 years of age in NHANES 1999–2000. PAD was defined as a blood pressure ankle brachial index < 0.9 in at least one leg. Metals were measured in casual (spot) urine specimens by inductively coupled plasma–mass spectrometry. After multivariable adjustment, subjects with PAD had 36% higher levels of cadmium in urine and 49% higher levels of tungsten compared with noncases. The adjusted odds ratio for PAD comparing the 75th to the 25th percentile of the cadmium distribution was 3.05 [95% confidence interval (CI), 0.97 to 9.58]; that for tungsten was 2.25 (95% CI, 0.97 to 5.24). PAD risk increased sharply at low levels of antimony and remained elevated beyond 0.1 μg/L. PAD was not associated with other metals. In conclusion, urinary cadmium, tungsten, and possibly antimony were associated with PAD in a representative sample of the U.S. population. For cadmium, these results strengthen previous findings using blood cadmium as a biomarker, and they support its role in atherosclerosis. For tungsten and antimony, these results need to be interpreted cautiously in the context of an exploratory analysis but deserve further study. Other metals in urine were not associated with PAD at the levels found in the general population.
antimonyatherosclerosiscadmiumleadmetalsperipheral arterial diseasetungsten
==== Body
Peripheral arterial disease (PAD) is a consequence of atherosclerotic occlusion of blood flow in the muscular arteries of the lower extremities. Because traditional risk factors for atherosclerosis, such as age, smoking, hypertension, and diabetes, do not completely explain the distribution of PAD in the population, there is considerable interest in identifying novel, nontraditional risk factors. Certain metals may promote atherosclerosis by increasing oxidative stress [e.g., by catalyzing the production of reactive oxygen species or inhibiting their degradation (Stohs and Bagchi 1995)] or by affecting other cardiovascular risk factors [e.g., by increasing blood pressure levels (Nawrot et al. 2002; Revis et al. 1981)]. Arsenic, for instance, has been associated with PAD and with other forms of atherosclerosis in populations exposed to arsenic in drinking water (Carter et al. 2003; National Research Council 1999). However, evidence for an association between most metals and PAD is limited.
Blood cadmium and lead, at levels well below current safety standards, were associated with an increased prevalence of PAD in the 1999–2000 National Health and Nutrition Examination Survey (NHANES) (Navas-Acien et al. 2004). Cadmium and lead have been associated with other cardiovascular end points, such as myocardial infarction and stroke, in some studies (Lustberg and Silbergeld 2002; Ponteva et al. 1979) but not all (Pocock et al. 1988; Staessen et al. 1996). For several other metals, including barium, cobalt, molybdenum, antimony, and thallium, data on their role in human atherosclerosis are sparse (Brenniman et al. 1979; Guo et al. 1992; Heim et al. 2002; Schnorr et al. 1995). There are no epidemiologic data on the potential cardiovascular effects of cesium or tungsten.
The objective of this study was to evaluate the association of metal levels in urine with the prevalence of PAD in NHANES 1999–2000, a representative sample of the civilian, noninstitutionalized U.S. population. In epidemiologic studies, PAD is usually defined by the presence of a low (< 0.9) blood pressure ankle-brachial index (ABI)—that is, when the systolic blood pressure measured at the ankle is < 90% of the systolic blood pressure measured at the arm (Selvin and Erlinger 2004). We hypothesized that cadmium and lead would be positively associated with PAD, as determined by an ABI < 0.9. In addition, we explored the association of PAD with urinary barium, cobalt, cesium, molybdenum, antimony, thallium, and tungsten, without prior hypotheses. These metals were selected for study in NHANES 1999–2000 by the National Center for Health Statistics (NCHS). The widespread exposure to these metals in the general population [National Center for Environmental Health (NCEH) 2003] and the lack of relevant population data for most of them underscore the public health relevance of this study.
Materials and Methods
Study population.
The NCHS has conducted a series of cross-sectional health and nutrition surveys beginning in the 1960s. These surveys have used a stratified, multistage probability cluster design to provide data representing the noninstitutionalized U.S. population (NCHS 2004a). In 1999, NHANES became a continuous survey and also began collecting information on ABI in men and women ≥40 years of age (NCHS 2004b).
Among the 9,965 participants in NHANES 1999–2000, one-third of the study population ≥6 years of age were randomly selected to obtain urinary measurements of metals (n = 2,465). NHANES 1999–2000 included a detailed lower-extremity examination that involved ABI measurements in subjects ≥40 years of age with no bilateral amputations and weighing < 400 lb (n = 2,875, of whom 796 had a determination of urinary metals). From these 796 participants, we excluded one participant with an ABI value > 1.5 [usually related to noncompressible vessels in the legs (Newman et al. 1993)] and five participants with no information on smoking status or urine levels of creatinine, for a final sample size of 790 participants. A few remaining participants had missing data for some of the metals (sample size available for each metal shown in Table 1). The response rate for the household interview and the physical exam components of NHANES 1999–2000 were 82 and 76%, respectively (NCHS 2004b).
PAD.
A specific protocol was used to measure ABI in NHANES 1999–2000 (NCHS 2004b). The measurements of blood pressure used for ABI were different from the measurements of blood pressure used to define hypertension. Supine systolic blood pressure was measured on the right arm (brachial artery) and both ankles (posterior tibial arteries) using a Doppler device, the Parks Mini-Lab IV, model 3100 (Parks Medical Electronics, Aloha, OR, USA). If the participant had a condition that would interfere with blood pressure reading in the right arm, the left arm was used. Systolic blood pressure was measured twice at each site for participants 40–59 years of age and once at each site for participants ≥60 years of age. The measurements for left and right ABI were obtained by dividing the mean ankle systolic blood pressure in each side by the mean brachial systolic blood pressure. PAD was defined as an ABI value < 0.90 in at least one leg (Newman et al. 1993).
Urinary metals.
A casual (or spot) urine specimen was collected from the participants after confirmation of no background contamination in collection materials (NCHS 2004b). Urinary levels of cadmium, lead, barium, beryllium, cobalt, cesium, molybdenum, platinum, antimony, thallium, and tungsten were measured at the Environmental Health Sciences Laboratory of the Centers for Disease Control and Prevention (CDC) and the National Center for Environmental Health (NCEH) by inductively coupled plasma–mass spectrometry (PerkinElmer/SCIEX model 500; PerkinElmer, Shelton, CT, USA) using a multielement analytical technique following published protocols (NCHS 2004c; Paschal et al. 1998). Cadmium levels in urine were corrected for interference from molybdenum oxide (NCHS 2004c). Urine Standard Reference Material 2670 from the National Institute of Standards and Technology (Gaithersburg, MD, USA) was used for external calibration and spiked pools prepared at the laboratory were used for internal quality control. Quality control samples included both bench and blind samples. The ranges for the interassay coefficients of variation for each metal are shown in Table 1. The limits of detection varied by metal, from 0.01 μg/L for thallium to 0.85 μg/L for molybdenum (Table 1). The levels of beryllium and platinum were below the limit of detection in > 98% of participants and were not considered further. For the other metals, the percentage of subjects with levels below the limit of detection ranged from 0.5% for cesium to 30% for tungsten (Table 1). A level equal to the limit of detection divided by the square root of 2 was imputed for those subjects with levels below the limits of detection.
Other variables.
Information on age, sex, race/ethnicity, education, and smoking was based on self-report. Urinary creatinine was determined using an enzymatic reaction measured with a CX3 analyzer (NCHS 2004b).
Statistical analysis.
All statistical analyses were performed using SUDAAN software (Research Triangle Institute, Research Triangle Park, NC, USA) to account for the complex sampling design and weights in the NHANES 1999–2000 metal subsample. The jackknife “leave-one-out” method was used to obtain appropriate standard errors of all estimates.
Urinary metal levels were right skewed and were log-transformed to improve normality. For each metal, the ratio of the geometric mean urinary level and its 95% confidence interval (CI) comparing subjects with PAD versus subjects without PAD were estimated using linear regression models on log-transformed metal levels. For risk analyses, logistic regression was used to obtain the adjusted odds ratios and 95% CIs of PAD comparing the 75th with the 25th percentile of the distribution of each metal assuming a log-linear relationship between urinary metals and PAD. We also used restricted cubic spline transformations to assess nonlinear relationships. All models were adjusted for age, sex, race, and education level. In addition, we evaluated the impact of further adjusting for smoking status, a source of cadmium, lead, and cobalt (International Agency for Research on Cancer 1986). Metal levels were based in one casual urine sample and thus depend on urine dilution. Urinary metal levels were adjusted for urinary creatinine to correct for differences in urine dilution in spot urine samples, but because of the limitations of creatinine concentration to adjust for urine dilution (Ikeda et al. 2003), we present results both with and without creatinine adjustment. We also repeated the analyses excluding subjects fasting < 8 hr to evaluate the impact of postprandial or fasting state and rapid renal clearance of some metals. No substantial changes were observed (data not shown). We also assessed confounding by other cardiovascular risk factors, including hypertension, hypercholesterolemia, diabetes, estimated glomerular filtration rate (Coresh et al. 2002), and C-reactive protein, by adding them one at a time to the multivariable models. No noticeable changes were observed (data not shown), and because of the limited sample size, these additional variables were not included in the final models.
Results
Metal levels in urine.
Table 1 describes urinary metal levels in the study sample. The geometric means were lowest for tungsten, 0.07 μg/L, and highest for molybdenum, 37.7 μg/L. The median level and the interquartile range for each metal by participant characteristics are shown in Figure 1. Men had higher levels than women for most metals, and cadmium and lead tended to increase with age. Compared with whites, blacks and Mexican Americans had higher levels of cadmium, lead, and tungsten, and blacks had higher levels of cesium, molybdenum, antimony, and thallium. Current smokers had higher levels of cadmium, lead, antimony, and tungsten compared with never smokers.
Metals and PAD.
The overall prevalence of PAD in the study sample was 5.4% (54 cases). After adjusting for demographic variables, smoking status, and creatinine in urine, subjects with PAD had 36% (95% CI, 1 to 83) higher mean levels of cadmium in urine compared with subjects without PAD (Table 2). Levels of lead, barium, cobalt, cesium, molybdenum, antimony, and thallium were similar in subjects with and without PAD. Subjects with PAD had 49% (95% CI, −10 to 249) higher mean levels of tungsten. The association of cadmium and tungsten with PAD was also evident in logistic models (Table 3, Figure 2). The odds ratio for PAD comparing the 75th with the 25th percentile of the cadmium distribution was 3.05 (95% CI, 0.97 to 9.58). The corresponding odds ratio for tungsten was 2.25 (95% CI, 0.97 to 5.24). For antimony, no association was apparent in linear models (Tables 2 and 3), but nonlinear models showed an increase in the prevalence of PAD at very low levels and a plateau of risk > 0.1 μg/L (Figure 2).
Discussion
Cadmium.
Cadmium levels in urine were 36% higher in subjects with PAD than in those without PAD. This association was similar but actually stronger than the previously reported association between blood cadmium and PAD in NHANES 1999–2000, where blood cadmium was 16% higher in PAD cases than in noncases (Navas-Acien et al. 2004). The stronger association for urinary compared with blood cadmium probably reflects the fact that urinary cadmium is a more reliable biomarker of chronic cadmium exposure than blood cadmium (ATSDR 1999a; Trzcinka-Ochocka et al. 2004). The association between cadmium levels and PAD was not explained by smoking status, and adjusting for smoking decreased only slightly the magnitude of the association. In our study, an increased prevalence of PAD was associated with low levels of urinary cadmium, below the levels reported in workers exposed to cadmium occupationally (Olsson et al. 2002). For example, only two subjects had cadmium levels > 3 μg/g of creatinine, the Occupational Safety and Health Administration safety standard for cadmium in urine (Occupational Safety and Health Administration 2003). In the NHANES population, because occupational exposure is expected to be relatively rare, it is likely that exposure to cadmium occurred mainly through cigarette smoking, inhalation of airborne cadmium in ambient air (usually higher near coal-fired power plants and municipal waste incinerators), or consumption of some foods (highest levels in shellfish, liver, and kidney meats) (ATSDR 1999b).
Several mechanisms may explain an increased risk of atherosclerosis with cadmium, including the catalysis of reactive oxygen species (Stohs and Bagchi 1995; Vaziri et al. 2001), the promotion of lipid peroxidation (Ding et al. 2000; Yiin et al. 1999), the depletion of glutathione and protein-bound sulf-hydryl groups (Stohs and Bagchi 1995), the production of inflammatory cytokines (Heo et al. 1996), and the down-regulation of nitric oxide production (Demontis et al. 1998; Vaziri et al. 2001). Cadmium has also induced atherosclerosis and hypertension in some animal models in vivo (Revis et al. 1981). However, these mechanistic studies were typically conducted at considerably higher exposures than those corresponding to the urinary concentrations observed in the present study, and hence their relevance to human atherogenesis and to PAD is uncertain.
Epidemiologic studies of cadmium and cardiovascular disease are also limited. Ecologic studies have found associations of cardiovascular mortality rates with cadmium levels in air (Carroll 1966) and in soil and water (Houtman 1993). Two small case–control studies found higher blood cadmium in subjects with myocardial infarction than in controls (Adamska-Dyniewska et al. 1982; Ponteva et al. 1979), but a cross-sectional study in Belgium found no association between blood cadmium and the prevalence of cardiovascular disease (Staessen et al. 1996). Finally, several autopsy studies have found associations between tissue cadmium levels and atherosclerotic lesions (Aalbers and Houtman 1985; Voors et al. 1982). Additional studies, particularly with a prospective design, are needed to confirm the role of cadmium in the pathogenesis of PAD and to determine its role in other cardiovascular end points.
Lead.
Lead in urine was not associated with PAD in this study. This result is in contrast to the association observed between blood lead and PAD in NHANES 1999–2000, where blood lead levels were 14% higher in cases of PAD than in noncases (Navas-Acien et al. 2004). This discrepancy could be related to the fact that urinary lead levels in this population were low and may have considerable fluctuations when evaluated in spot urine samples. Under these conditions, urinary lead is considered a less reliable biomarker of exposure than blood lead (ATSDR 1999b). Indeed, previous cohort studies of lead and cardiovascular disease have used blood lead as the biomarker of exposure. Blood lead was positively associated with cardiovascular mortality in NHANES II (Lustberg and Silbergeld 2002) and with coronary heart disease incidence in Denmark (Moller and Kristensen 1992), although an earlier study in British men did not show an association between blood lead and cardiovascular disease incidence (Pocock et al. 1988). Several mechanisms support a role for lead in atherosclerosis. Lead increases blood pressure (Nawrot et al. 2002; Revis et al. 1981), and experimental studies show that lead promotes oxidative stress (Stohs and Bagchi 1995), stimulates inflammation (Heo et al. 1996), and induces endothelial damage (Vaziri et al. 2001). The role of lead in the development of atherosclerosis, however, needs to be further investigated in mechanistic studies at low levels of exposure and in prospective studies in humans using appropriate biomarkers of chronic exposure.
Tungsten.
Tungsten levels in urine were 49% higher in subjects with PAD than in those without PAD. This analysis is among the first to examine the role of tungsten in any health status indicator, and it needs to be interpreted cautiously in the context of multiple comparisons. Little is known regarding tungsten toxicity and carcinogenicity (ATSDR 2003), and there are insufficient data, either from animal or human studies, on its cardiovascular effects (Lagarde and Leroy 2002). It is known, however, that tungsten is thrombogenic and proinflammatory (Byrne et al. 1997). In fact, these properties have motivated the clinical use of tungsten coils for the occlusion of intracranial aneurysms, varicocele veins, and other abnormal vascular connections (Butler et al. 2000; Peuster et al. 2002). In addition, tungsten may interfere with the biologic effect of molybdenum, an essential trace element that acts as cofactor in several proteins (Johnson et al. 1974; Nell et al. 1980). For instance, tungsten inhibits xanthine oxidase (Johnson et al. 1974), an antioxidant molybdoenzyme with a role in endothelial dysfunction (Berry and Hare 2004) and in the maintenance of the vessel wall integrity.
Sources of tungsten exposure include occupational use of tungsten inert gas, welding, and sometimes drinking water. The amounts of tungsten in foods and in ambient air are generally unknown (ATSDR 2003). Urban settings tend to have higher levels of tungsten in the air because tungsten can be released from industrial sources and waste incinerators (tungsten filaments are used in incandescent light bulbs) (ATSDR 2003). In response to the lack of knowledge on the health effects of tungsten exposure coupled with evidence of widespread human exposure, the CDC/NCEH nominated tungsten in 2002 to the National Toxicology Program as a priority candidate for toxicologic assessment (U.S. Department of Health and Human Services 2003). Our findings suggest that these studies should also include the evaluation of the relationship of tungsten to cardiovascular end points and its potential mechanisms of action.
Antimony.
For antimony, there was an increase in the prevalence of PAD in subjects with low levels compared with those at the limit of detection, and the risk remained increased > 0.1 μg/L. The general population is exposed to antimony in food, in drinking water, and in ambient air (ATSDR 1992a). At levels found in antimony smelting plants, antimony has been related to pneumoconiosis and dermatitis (McCallum 1989). Although antimony is a well-known toxic metal at high doses, there is only one available study of chronic antimony exposure and cardiovascular end points in humans (Schnorr et al. 1995). This study compared coronary mortality in Mexican-American antimony smelter workers with other groups of workers, but the findings were inconclusive. Interestingly, antimony shares similar chemical and toxicologic properties with arsenic (Gebel 1997), and they are frequent coexposures (Gebel et al. 1998). Arsenic exposure has been associated with PAD, and “blackfoot disease” is a classic sign of high arsenic exposure (Carter et al. 2003; National Research Council 1999). Unfortunately, arsenic levels were not measured in NHANES 1999–2000, and we cannot discard the possibility that the observed association between PAD and antimony was due to arsenic coexposure. Further studies are required on antimony and cardiovascular end points.
Other metals.
Barium, cobalt, cesium, molybdenum, and thallium in urine were not associated with PAD in this representative sample of the U.S. population. Data on the possible role of these metals in atherosclerosis are scarce. Ecologic studies have found positive associations with cardiovascular disease for barium (Brenniman et al. 1979) and thallium (Heim et al. 2002) and a negative association for molybdenum (Guo et al. 1992). Thallium is a poison at high dose (ATSDR 1992b). At low dose, thallium is used in cardiac imaging and thought to be relatively safe (Ranhosky and Kempthorne-Rawson 1990). Cobalt exposure in the hard metal industry results in cardiomiopathy (Seghizzi et al. 1994), but no information is available on its atherogenicity.
Limitations and strengths.
Our analysis was limited by a relatively small number of cases, because NHANES 1999–2000 measured metals in urine in just one-third of survey participants. This sample size limited the investigation of interactions (e.g., lead and cadmium, or tungsten and molybdenum) and makes it possible that we missed weak associations between some metals and PAD. At the same time, our prior hypotheses regarding the associations of cadmium and lead with PAD were specific, but for the other metals our analyses were exploratory and need to be confirmed in future studies.
Other limitations include the cross-sectional design of the study, the possibility of residual confounding by socioeconomic status or urbanization, the use of a single measurement of urinary metals, and the lack of 24-hr urine collection to better account for short-term variability in metal excretion and urine dilution. To correct for urine dilution, we adjusted for urinary creatinine. However, creatinine in urine is a marker of both urine dilution and creatinine production, and it is associated with factors that affect production such as sex, age, race, and muscle mass. The adequacy of correcting for creatinine has been questioned (Ikeda et al. 2003), and in fact the Second National Report on Human Exposure to Environmental Chemicals (NCEH 2003), based on NHANES data, presented metal concentrations in urine both ways, with and without correction for urinary creatinine. In our data, the findings using models with and without adjustment for creatinine were similar and do not affect the conclusions.
Despite these limitations, this study is the first systematic investigation of a panel of metals in urine with PAD. The use of a representative sample of the U.S. population, rigorous laboratory methods with extensive quality control, and the availability of a standardized procedure to measure ABI add to the strengths of this study.
Conclusions
Urinary cadmium was strongly associated with PAD in a representative sample of the general U.S. population. This finding strengthens similar results from a previous study in the same population using blood cadmium as a biomarker of exposure (Navas-Acien et al. 2004) and further supports a possible role for cadmium in atherosclerosis. Tungsten was associated with PAD throughout the range of exposure, whereas antimony showed a positive association at low levels that reached a plateau > 0.1 μg/L. These associations need to be interpreted cautiously and require confirmation in future epidemiologic studies and supporting evidence from mechanistic research. Finally, no association between PAD and other metals in urine (lead, barium, cobalt, cesium, molybdenum, and thallium) was evident at the levels found in the general population. The observation of widespread exposure, particularly to poorly studied metals such as tungsten and antimony, supports further research on the role of metals and PAD.
Figure 1 Metal levels in urine (μg/L) by participant characteristics. Horizontal lines, interquartile ranges; squares, medians; dotted vertical line, the geometric mean for the overall study sample.
Figure 2 Odds ratios of PAD by metal levels in urine. The curves are odds ratios adjusted for age, sex, race, education, smoking, and urinary creatinine based on restricted cubic spline transformations. The reference value (odds ratio = 1) was set at the 10th percentile of the distribution for each metal. The bar histograms represent the frequency distribution of each metal in the study sample. The tick marks at the bottom of the histogram represent the metal level of the cases of PAD.
Table 1 Metal levels (μg/L) in urine.
Cadmium Lead Barium Cobalt Cesium Molybdenum Antimony Thallium Tungsten
Sample size 728 790 704 790 790 728 725 776 751
Geometric mean 0.36 0.79 1.28 0.31 4.09 37.7 0.11 0.16 0.07
Percentile
10th 0.10 0.20 0.30 0.10 1.50 11.0 < LOD 0.06 < LOD
25th 0.19 0.50 0.70 0.18 2.80 21.2 0.07 0.10 < LOD
50th 0.36 0.90 1.40 0.33 4.50 41.2 0.11 0.18 0.06
75th 0.67 1.50 2.60 0.53 6.80 71.3 0.17 0.26 0.13
90th 1.16 2.30 4.70 0.81 9.40 126.1 0.29 0.38 0.26
Maximum 12.8 31.5 42.2 556.6 67.4 683.5 5.70 0.86 4.74
LOD 0.06 0.10 0.08 0.07 0.10 0.85 0.04 0.01 0.04
Percent < LOD 1.5 1.8 3.3 3.4 0.5 0.6 9.0 1.3 30.0
Range of CV (%) 1.2–4.7 1.0–5.3 1.4–6.2 1.8–6.0 1.5–9.2 0.6–5.6 1.3–4.8 1.2–8.5 1.1–2.8
Abbreviations: CV, coefficient of variation; LOD, limit of detection.
Table 2 Ratios (95% CIs) of the geometric means of metal levels in urine (μg/L) in PAD cases versus noncases.
Cases Noncases Model 1a Model 2b Model 3c
Cadmium 49 679 1.81 (1.24–2.62) 1.62 (1.19–2.21) 1.36 (1.01–1.83)
Lead 54 736 1.09 (0.86–1.37) 1.08 (0.85–1.38) 0.92 (0.74–1.15)
Barium 45 659 0.99 (0.67–1.47) 0.99 (0.68–1.45) 0.82 (0.60–1.11)
Cobalt 54 736 1.13 (0.80–1.59) 1.13 (0.82–1.57) 0.98 (0.69–1.40)
Cesium 54 736 1.05 (0.83–1.32) 1.12 (0.89–1.42) 0.96 (0.79–1.16)
Molybdenum 49 679 0.97 (0.66–1.42) 1.08 (0.74–1.57) 0.91 (0.72–1.15)
Antimony 49 676 1.18 (0.92–1.51) 1.17 (0.92–1.50) 1.03 (0.87–1.22)
Thallium 54 722 0.97 (0.71–1.34) 1.08 (0.78–1.48) 0.94 (0.74–1.19)
Tungsten 51 700 1.75 (0.98–3.10) 1.67 (0.96–2.89) 1.49 (0.90–2.49)
a Adjusted by age, sex, race, and education.
b Further adjusted by smoking status (never/former/current).
c Further adjusted by urinary creatinine.
Table 3 Odds ratio (95% CIs) of PAD comparing the 75th with the 25th percentile of the metal distribution.
Cases Noncases Model 1a Model 2b Model 3c
Cadmium 49 676 2.67 (1.40–5.07) 2.14 (1.11–4.13) 3.05 (0.97–9.58)
Lead 54 736 1.17 (0.81–1.69) 1.17 (0.78–1.76) 0.89 (0.45–1.78)
Barium 45 659 1.02 (0.67–1.56) 1.07 (0.72–1.58) 0.88 (0.57–1.36)
Cobalt 54 736 1.21 (0.65–2.23) 1.22 (0.67–2.20) 1.01 (0.33–3.14)
Cesium 54 736 1.08 (0.73–1.60) 1.19 (0.78–1.80) 0.91 (0.33–2.48)
Molybdenum 49 679 0.98 (0.60–1.60) 1.10 (0.69–1.77) 0.83 (0.49–1.41)
Antimony 49 676 1.25 (0.93–1.68) 1.30 (0.95–1.77) 1.15 (0.81–1.63)
Thallium 54 722 0.96 (0.53–1.73) 1.18 (0.60–2.32) 0.87 (0.30–2.52)
Tungsten 51 700 2.45 (1.12–5.37) 2.23 (1.03–4.82) 2.25 (0.97–5.24)
a Adjusted by age, sex, race, and education.
b Further adjusted by smoking status (never/former/current).
c Further adjusted by urinary creatinine.
==== Refs
References
Aalbers TG Houtman JP 1985 Relationships between trace elements and atherosclerosis Sci Total Environ 43 255 283 4012298
Adamska-Dyniewska H Bala T Florczak H Trojanowska B 1982 Blood cadmium in healthy subjects and in patients with cardiovascular diseases Cor Vasa 24 441 447 7184698
ATSDR 1992a. Toxicological Profile for Antimony. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
ATSDR 1992b. Toxicological Profile for Thallium. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
ATSDR 1999a. Toxicological Profile for Cadmium. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
ATSDR 1999b. Toxicological Profile for Lead. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
ATSDR 2003. Toxicological Profile for Tungsten. Draft. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
Berry CE Hare JM 2004 Xanthine oxidoreductase and cardiovascular disease: molecular mechanisms and pathophysiological implications J Physiol 555 589 606 14694147
Brenniman GR Namekata T Kojola WH Carnow BW Levy PS 1979 Cardiovascular disease death rates in communities with elevated levels of barium in drinking water Environ Res 20 318 324 546641
Butler TJ Jackson RW Robson JY Owen RJ Delves HT Sieniawska CE 2000 In vivo degradation of tungsten embolisation coils Br J Radiol 73 601 603 10911782
Byrne JV Hope JK Hubbard N Morris JH 1997 The nature of thrombosis induced by platinum and tungsten coils in saccular aneurysms Am J Neuroradiol 18 29 33 9010517
Carroll RE 1966 The relationship of cadmium in the air to cardiovascular disease death rates JAMA 198 267 269 5952924
Carter DE Aposhian HV Gandolfi AJ 2003 The metabolism of inorganic arsenic oxides, gallium arsenide, and arsine: a toxicochemical review Toxicol Appl Pharmacol 193 309 334 14678742
Coresh J Astor BC McQuillan G Kusek J Greene T Van Lente F 2002 Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate Am J Kidney Dis 39 920 929 11979335
Demontis MP Varoni MV Volpe AR Emanueli C Madeddu P 1998 Role of nitric oxide synthase inhibition in the acute hypertensive response to intracerebroventricular cadmium Br J Pharmacol 123 129 135 9484863
Ding Y Gonick HC Vaziri ND 2000 Lead promotes hydroxyl radical generation and lipid peroxidation in cultured aortic endothelial cells Am J Hypertens 13 552 555 10826409
Gebel T 1997 Arsenic and antimony: comparative approach on mechanistic toxicology Chem Biol Interact 107 131 144 9448748
Gebel TW Suchenwirth RH Bolten C Dunkelberg HH 1998 Human biomonitoring of arsenic and antimony in case of an elevated geogenic exposure Environ Health Perspect 106 33 39 9417766
Guo W Li JY King H Locke FB 1992 Diet and blood nutrient correlations with ischemic heart, hypertensive heart, and stroke mortality in China Asia Pac J Public Health 6 200 209 1345447
Heim M Wappelhorst O Markert B 2002 Thallium in terrestrial environments—occurrence and effects Ecotoxicology 11 369 377 12463684
Heo Y Parsons PJ Lawrence DA 1996 Lead differentially modifies cytokine production in vitro and in vivo Toxicol Appl Pharmacol 138 149 157 8658504
Houtman JP 1993 Prolonged low-level cadmium intake and atherosclerosis Sci Total Environ 138 31 36 8259493
Ikeda M Ezaki T Tsukahara T Moriguchi J Furuki K Fukui Y 2003 Bias induced by the use of creatinine-corrected values in evaluation of beta2-microgloblin levels Toxicol Lett 145 197 207 14581173
International Agency for Research on Cancer. 1986 Tobacco smoking IARC Monogr Eval Carcinog Risks Hum 38 1 403
Johnson JL Waud WR Cohen HJ Rajagopalan KV 1974 Molecular basis of the biological function of molybdenum. Molybdenum-free xanthine oxidase from livers of tungsten-treated rats J Biol Chem 249 5056 5061 4368927
Lagarde F Leroy M 2002 Metabolism and toxicity of tungsten in humans and animals Met Ions Biol Syst 39 741 759 11913143
Lustberg M Silbergeld E 2002 Blood lead levels and mortality Arch Intern Med 162 2443 2449 12437403
McCallum RI 1989 The industrial toxicology of antimony. The Ernestine Henry lecture 1987 J R Coll Physicians Lond 23 28 32 2926715
Moller L Kristensen TS 1992 Blood lead as a cardiovascular risk factor Am J Epidemiol 136 1091 1100 1462969
NCEH 2003. Second National Report on Human Exposure to Environmental Chemicals. Atlanta, GA:National Center for Environmental Health, U.S. Department of Health and Human Services.
NCHS 2004a. National Health and Nutrition Examination Survey. Hyattsville, MD:National Center for Health Statistics. Available: http://www.cdc.gov/nchs/nhanes.htm [accessed 15 October 2004].
NCHS 2004b. 1999–2000 National Health and Nutrition Examination Survey. Hyattsville, MD:National Center for Health Statistics. Available: http://www.cdc.gov/nchs/about/major/nhanes/NHANES99_00.htm [accessed 15 October 2004].
NCHS 2004c. NHANES 1999–2000 Public Release Data Set. Lab 06HM—Urinary Barium, Beryllium, Cadmium, Cobalt, Cesium, Molybdenum, Lead, Platinum, Antimony, Thallium, and Tungsten. Hyattsville, MD:National Center for Health Statistics. Available: http://www.cdc.gov/nchs/data/nhanes/frequency/lab06hm_doc.pdf [accessed 15 October 2004].
National Research Council 1999. Arsenic in Drinking Water. Washington DC: National Academy Press.
Navas-Acien A Selvin E Sharrett AR Calderon-Aranda E Silbergeld E Guallar E 2004 Lead, cadmium, smoking, and increased risk of peripheral arterial disease Circulation 109 3196 3201 15184277
Nawrot TS Thijs L Den Hond EM Roels HA Staessen JA 2002 An epidemiological re-appraisal of the association between blood pressure and blood lead: a meta-analysis J Hum Hypertens 16 123 131 11850770
Nell JA Annison EF Balnave D 1980 The influence of tungsten on the molybdenum status of poultry Br Poult Sci 21 193 202 6930313
Newman AB Siscovick DS Manolio TA Polak J Fried LP Borhani NO 1993 Ankle-arm index as a marker of atherosclerosis in the Cardiovascular Health Study. Cardiovascular Heart Study (CHS) Collaborative Research Group Circulation 88 837 845 8353913
Occupational Safety and Health Administration 2003 Occupational Safety and Health Standards. Toxic and Hazardous Substances. Cadmium CFR 1910.1027 135 229
Olsson IM Bensryd I Lundh T Ottosson H Skerfving S Oskarsson A 2002 Cadmium in blood and urine—impact of sex, age, dietary intake, iron status, and former smoking—association of renal effects Environ Health Perspect 110 1185 1190 12460796
Paschal DC Ting BG Morrow JC Pirkle JL Jackson RJ Sampson EJ 1998 Trace metals in urine of United States residents: reference range concentrations Environ Res 76 53 59 9466897
Peuster M Fink C von Schnakenburg C Hausdorf G 2002 Dissolution of tungsten coils does not produce systemic toxicity, but leads to elevated levels of tungsten in the serum and recanalization of the previously occluded vessel Cardiol Young 12 229 235 12365168
Pocock SJ Shaper AG Ashby D Delves HT Clayton BE 1988 The relationship between blood lead, blood pressure, stroke, and heart attacks in middle-aged British men Environ Health Perspect 78 23 30 3203640
Ponteva M Elomaa I Backman H Hansson L Kilpio J 1979 Blood cadmium and plasma zinc measurements in acute myocardial infarction Eur J Cardiol 9 379 391 456396
Ranhosky A Kempthorne-Rawson J 1990 The safety of intravenous dipyridamole thallium myocardial perfusion imaging. Intravenous Dipyridamole Thallium Imaging Study Group Circulation 81 1205 1209 2107985
Revis NW Zinsmeister AR Bull R 1981 Atherosclerosis and hypertension induction by lead and cadmium ions: an effect prevented by calcium ion Proc Natl Acad Sci USA 78 6494 6498 6947240
Schnorr TM Steenland K Thun MJ Rinsky RA 1995 Mortality in a cohort of antimony smelter workers Am J Ind Med 27 759 770 7611310
Seghizzi P D’Adda F Borleri D Barbic F Mosconi G 1994 Cobalt myocardiopathy. A critical review of literature Sci Total Environ 150 105 109 7939581
Selvin E Erlinger TP 2004 Prevalence and risk factors for peripheral arterial disease in the U.S.: results from the National Health and Nutrition Examination Survey (NHANES), 1999–2000 Circulation 110 738 743 15262830
Staessen JA Buchet JP Ginucchio G Lauwerys RR Lijnen P Roels H 1996 Public health implications of environmental exposure to cadmium and lead: an overview of epidemiological studies in Belgium. Working Groups J Cardiovasc Risk 3 26 41 8783028
Stohs SJ Bagchi D 1995 Oxidative mechanisms in the toxicity of metal ions Free Radic Biol Med 18 321 336 7744317
Trzcinka-Ochocka M Jakubowski M Razniewska G Halatek T Gazewski A 2004 The effects of environmental cadmium exposure on kidney function: the possible influence of age Environ Res 95 143 150 15147919
U.S. Department of Health and Human Services. 2003 National Toxicology Program: Announcement of and Request for Public Comments on Substances Nominated to the National Toxicology Program (NTP) for Toxicological Studies and Study Recommendations Made by the NTP Interagency Committee for Chemical Evaluation and Coordination (ICCEC) Fed Reg 68 42068 42071
Vaziri ND Ding Y Ni Z 2001 Compensatory up-regulation of nitric-oxide synthase isoforms in lead-induced hypertension; reversal by a superoxide dismutase-mimetic drug J Pharmacol Exp Ther 298 679 685 11454931
Voors AW Shuman MS Johnson WD 1982 Additive statistical effects of cadmium and lead on heart-related disease in a North Carolina autopsy series Arch Environ Health 37 98 102 7073331
Yiin SJ Chern CL Sheu JY Tseng WC Lin TH 1999 Cadmium-induced renal lipid peroxidation in rats and protection by selenium J Toxicol Environ Health A 57 403 413 10478822
|
15687053
|
PMC1277859
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):164-169
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7329
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7389ehp0113-00017015687054ResearchArticlesComparison of Ultrastructural Cytotoxic Effects of Carbon and Carbon/Iron Particulates on Human Monocyte-Derived Macrophages Long John F. 1Waldman W. James 2Kristovich Robert 3Williams Marshall 4Knight Deborah 2Dutta Prabir K. 31Department of Veterinary Biosciences,2Department of Pathology,3Department of Chemistry, and4Department of Molecular Virology, Immunology, and Medical Genetics, Ohio State University, Columbus, Ohio, USAAddress correspondence to J.F. Long, Department of Veterinary Biosciences, Ohio State University, College of Veterinary Medicine, 1925 Coffey Rd., Columbus, OH 43210 USA. Telephone: (614) 688-5940. Fax: (614) 292-6473. E-mail:
[email protected] thank E. Handley for her work in the preparation of specimens for electron microscopy.
This study was supported in part by an Interdisciplinary Seed Grant from the Office of Research, Ohio State University, and National Science Foundation–funded Environmental Molecular Sciences Institute at Ohio State University (CHE 0089147).
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 170 174 17 6 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. In this study, we tested the hypothesis that the presence of iron in carbon particulates enhances ultrastructural perturbation in human monocyte-derived macrophages (MDMs) after phagocytosis. We used 1-μm synthetic carbon-based particulates, designed to simulate environmental particulates of mass median aerodynamic diameter ≤ 2.5 μm (PM2.5). Cultures of human MDMs or T-lymphocytes (as a nonphagocytic control) were exposed to carbon or carbon/iron particulates for various time periods and examined by transmission electron microscopy for ultrastructural changes. T-cells failed to internalize either of the particulates and showed no organelle or nuclear changes. Conversely, MDMs avidly phagocytized the particulates. MDMs treated with C particulates exhibited morphologic evidence of macrophage activation but no evidence of lysis of organelles. In contrast, MDMs treated with C/Fe particulates exhibited coalescence of particulate-containing lysosomes. This phenomenon was not observed in the case of C particulates. By 24 hr there was a tendency of the C/Fe particulates to agglomerate into loose or compact clusters. Surrounding the compact C/Fe agglomerates was a uniform zone of nearly total organelle lysis. The lytic changes diminished in proportion to the distance from the agglomerate. In such cells, the nucleus showed loss of chromatin. Although C particles induced no detectable oxidative burst on treated MDMs, C/Fe particles induced a nearly 5-fold increase in the extracellular oxidative burst by treated MDMs compared with untreated controls. Iron bound to C particles catalyzed the decomposition of hydrogen peroxide to generate hydroxyl radicals. Results of these studies suggest that, among particulates of similar size, biologic activity can vary profoundly as a function of particulate physicochemical properties.
carboncarbon/ironcytotoxicitymacrophagesultrastructure
==== Body
Respirable particulates containing transition metals such as iron, vanadium, nickel, and copper are known to catalyze the generation of reactive oxygen species (ROSs) such as the highly damaging hydroxyl radical (Smith et al. 2000). Sequelae from ROS production can include lipid peroxidation of cell membranes, oxidative modification of proteins, oxidation of DNA, and alterations in calcium homeostasis of the cell.
The process of phagocytosis allows uncontrolled entry of iron-containing particulates into cells because these particulates have bypassed control by the protein transferrin (Hardy and Aust 1995) and its cognate receptor. Transferrin is a transport factor normally found in serum (Darling and Morgan 1994). Although iron is essential for life, its uncontrolled entry into cells has the potential for damaging the cell by oxidative stress (Smith et al. 2000). Iron release from phagocytized iron is greatest at a low pH (e.g., present in lysosomes containing iron particulates), indicating that iron can be mobilized inside macrophages after phagocytosis (Gilmour et al. 1996).
Valavanidis et al. (2000), studying the generation of hydroxyl radicals by urban suspended particulate matter (PM), concluded that the presence of trace metals, especially iron in the form of soluble ferrous and ferric salts, play an important role for oxidant-generating activity. There is also variation in bioactivity of iron in different compounds. Fubini et al. (1999) points out that not every iron-containing solid is active; iron oxide, for example, was mostly inactive.
Our group has examined two naturally occurring zeolites, including forms whose biologic activity is reported to range from highly pathogenic (erionite) to essentially benign (mordenite). We found that on exposure to the same mass of a specific type of particulate, the oxidative burst increases with decreasing particle size but remains relatively independent of zeolite composition. On the other hand, the Fenton reaction depends on the type of zeolite, suggesting that the surface structure of the iron on the zeolite plays an important role (Fach et al. 2002). In another study, we found that the mutagenic potential of mordenite was not enhanced by the addition of ferrous ion. Conversely, mutagenicity of erionite was significantly enhanced by the addition of ferrous ions. These results suggested that although the cytotoxicity of mordenite and erionite may be related to the ability of these fibers to transport iron into a cell, the different coordination state of iron associated with the two fiber surfaces is critical for inducing genotoxic damage (Fach et al. 2003).
In the present study, we used synthetic 1-μm carbon and carbon/iron particulates, designed to simulate environmental particulates of mass median aerodynamic diameter ≤ 2.5 μm (PM2.5) that occur as airborne pollution. We made ultrastructural observations and compared human blood monocyte-derived macrophages (MDMs; known to avidly phagocytize PM) with human T-cells (as nonphagocytic controls) on exposure to C or C/Fe particulates. To assess the bioactivity of C and C/Fe particulates, luminescent assays using luminol for oxidative bursts were conducted after exposure of macrophages to the particulates. In addition, the ability of the particulates to produce hydroxyl radicals from hydrogen peroxide (H2O2) was studied (Fenton reaction).
Materials and Methods
Synthesis of C and C/Fe particulates.
The procedures followed were adapted from a previously published procedure for the synthesis of a carbonaceous negative image of a zeolite (Johnson et al. 1997). The synthesis of C particle replicas proceeds through an acid-catalyzed condensation reaction of phenol and paraformaldehyde monomers using zeolite Y as a template for the reaction. The zeolite was acidified by ion exchange with NH4Cl, followed by decomposition of the ion-exchanged ammonium ions under vacuum at 500°C. Iron was added to the appropriate particles by ion exchange with 0.01 M ferrous sulfate (3 cycles, 30 min each) before decomposition of the ammonium ions. The zeolite was cooled to room temperature, and solid phenol (0.21 g phenol/g zeolite) was added. A weak vacuum was pulled on the system, and the temperature was raised slowly to 60°C. Solid paraformaldehyde (0.25 g paraformaldehyde/g zeolite) was added to the zeolite/phenol material, and the temperature was raised very slowly to 120°C in a nitrogen environment. At this point the solid material turns red as the phenol/paraformaldehyde cross-linked polymer forms. The solid material was held at 120°C for 5 hr to allow for complete polymerization. The zeolite/polymer mix was pyrolyzed at 800°C for 19 hr under flowing argon. The zeolite template was removed by etching in 48% hydrofluoric acid for 6 hr. The size of the particles was confirmed by scanning electron microscopy (SEM). Bulk analysis of C/Fe particulates found aluminum content of 1.38%, silicon 0.33%, iron 0.46%, and the rest carbon.
Preparation of target cells.
Peripheral blood mononuclear cells (PBMCs) were separated by Ficoll-Hypaque density gradient centrifugation from fresh heparinized blood collected by venipuncture from healthy, nontransfused consenting volunteers, as previously described (Boyum 1968; Waldman et al. 1992). This study was approved by the Ohio State University institutional review board, human subjects protocol 94HO344. For isolation and differentiation of monocytes, PBMCs were suspended in mononuclear leukocyte culture medium [complete Dulbecco’s modified Eagle’s medium (DMEM)] (Waldman et al. 1992) supplemented with 10 ng/mL monocyte-colony–stimulating factor (M-CSF; R&D Systems, Minneapolis, MN), transferred to six-well plastic tissue culture plates (Costar, Corning Inc., Corning, NY) at a concentration of approximately 2 × 107 cells/well (2 mL/well), and incubated at 37°C in a humidified atmosphere of 10% CO2/90% air for 48 hr to allow adherence of monocytes. Nonadherent cells were then removed, and adherent monocytes were washed twice with Seligman’s buffered saline solution (SBSS) and supplied with fresh complete M-CSF–free leukocyte culture medium. Cells were incubated for an additional 10–12 days with medium changes at 48-hr intervals to allow differentiation into the macrophage phenotype and were labeled MDMs.
T-lymphocytes were isolated from PBMCs by negative selection with a commercially available cocktail of monoclonal antibodies and complement (T Lympho-kwik, One Lambda, Inc., Los Angeles, CA), using methods based on those of Clouse et al. (1987) as detailed elsewhere (Waldman et al. 1992). To remove any residual monocytes, cells were incubated in leukocyte culture medium for 1 hr in plastic tissue culture flasks (Costar) at 37°C in a humidified atmosphere of 10% CO2/90% air, after which nonadherent cells were recovered for use in experiments.
To assess homogeneity of populations prepared in this manner, samples were suspended in SBSS, stained with fluorescein isothio-cyanate–conjugated monoclonal antibodies specific for CD3, CD4, CD8, and CD14 (Gen Trak, Liberty, NC) as previously described (Waldman et al. 1992), and analyzed (5,000 cells) using a Coulter Epics XL flow cytometer (Beckman Coulter, Inc., Fullerton, CA). As controls for nonspecific staining, cells were reacted with appropriate isotypically matched irrelevant murine antibodies. MDMs routinely marked 90–95% positive for CD14 with undetectable levels of T-cell contamination (as indicated by the absence of CD3+ cells). T-cells routinely marked 90–95% positive for CD3 and 25–35% positive for CD8, with the remainder positive for CD4, and had undetectable levels of monocyte contamination (as indicated by the absence of CD14+ cells).
Treatment of cells with particulates.
C/Fe particulates 1 μm in diameter as well as similarly sized C particulates (without iron) were sterilized by steam autoclave and suspended in serum-free DMEM. MDMs differentiated as described above were washed with phosphate-buffered saline (PBS) and supplied with fresh complete DMEM (2.0 mL/well). T-cells were likewise washed and suspended in fresh medium in six-well culture plates (~ 5 × 106 cells in 2 mL/well). Particulates were sonicated and immediately added to cultures at a non-toxic concentration of 5 μg/cm2 surface area. Trypan blue dye exclusion was used to demonstrate that the exposure of 5 μg/cm2 particles was not toxic to the cells. Plates were centrifuged for 10 min at 300 × g immediately after addition of particulates and then incubated at 37°C in a humidified atmosphere of 10% CO2/90% air. After various periods of incubation (2–24 hr), MDMs were washed twice with PBS and harvested for fixation by gentle scraping after 15 min of incubation in 0.01% EDTA/PBS at 4°C. Concurrently, free particulates were separated from T-cells by density gradient centrifugation through Ficoll-Hypaque (Histopaque, Sigma, St. Louis, MO), which allows sedimentation of free particulates but not T-cells. T-cells recovered from the gradient were washed twice in PBS before fixation.
Transmission electron microscopy.
Suspended cells were washed twice in PBS and then fixed for 18–24 hr in 3% cacodylate buffered glutaraldehyde and again pelleted by centrifugation. Cells were washed twice in 0.1 M cacodylate, postfixed in 1% S-collidine–buffered osmium tetroxide (1 hr), and then dehydrated in graded ethanol washes. Specimens were embedded in Medcast (Ted Pella, Inc., Redding, CA). Blocks were cured for a minimum of 12 hr at 60°C. Thin sections (~100 nm) were cut from cured blocks using an ultramicrotome (LKB Nova; LKB, Stockholm, Sweden) and mounted on 2-mm 300-mesh copper grids. Grids were heavy metal stained using a standard two-step uranyl acetate/lead citrate technique and then examined and photographed at 60 kV with a Philips 300 transmission electron microscope (Philips Electronic Instrument Co., Mahwah, NJ).
Assay of particulate-induced oxidative burst.
Human PBMCs were plated in 96-well Optilux culture plates (BD Falcon, Palo Alto, CA) and allowed to differentiate into MDMs as described above. Immediately before assay, culture medium was removed, and cells were washed twice with PBS before addition of ice-cold serum-free RPMI 1640 (Cellgro-Mediatech, Herndon, VA), 100 μL/well. A stock solution of 100 mM luminol was prepared with 20 mg luminol (5-amino-2,3-dihydro-1,4-phthalazinedione, sodium salt; Sigma)/mL dimethyl sulfoxide and stored at −20°C until use. Immediately before assays, luminol stock solution was thawed, diluted 1:100 (1 mM) in ice cold RPMI, and added to culture wells (100 μL/well). Particulates (C or C/Fe) were sonicated and added to culture wells at a concentration of 5 μg/cm2, four replicate wells per treatment. As negative controls, each experiment included 4 replicate wells containing MDMs, medium, and luminol but no particulates. Plates were sealed with UV-sterilized Top Seals (Packard, Meriden, CT) and centrifuged at 4°C (400 × g, 3 min). Luminescence was immediately measured (time 0) with a Top-Count Microplate Scintillation and Luminescence Counter (Packard), after which plates were placed in a 37°C incubator. Subsequently, luminescence was measured at 10 min intervals, with plates being incubated at 37°C between counts.
Luminescence indices were calculated by dividing the mean luminescence counts per minute (cpm) of four replicate particulate-treated wells by the mean cpm of four negative control wells at each time point.
Measurement of hydroxyl radical production by C and C/Fe particles.
Five milligrams of synthetic C or C/Fe particles were weighed and suspended in 500 μL of PBS (pH 7.4) solution. To this suspension, 100 μL of 5,5-dimetheyl-1-pyrolline-N-oxide (DMPO; 97% purity; Aldrich Chemical Company, St. Louis, MO) was added, along with 50 μL of a 30% hydrogen peroxide solution (Mallinkrodt Baker, Inc., Phillipsburg, NJ). The solution was shaken in the dark for 15 min. The particles were removed by centrifugation, and the solution was analyzed in a capillary column by electron spin resonance (ESR) spectroscopy (Bruker ESP300 ESR spectrometer; Bruker, Ettlinger, Germany). ESR was performed with a modulating frequency of 100 kHz with a modulating amplitude of 2.090 G. The microwave power was 0.632 mW. The samples were scanned 10 times with the average taken as the representative spectra.
Results
For the untreated cells (controls), the nucleus showed the expected rim of heterochromatin. The cytoplasmic organelles appeared normal. The microvilli showed expected formations. Nucleoli were ultrastructurally normal.
An SEM image of the C/Fe synthetic particulates is shown in Figure 1; the uniformity in size is apparent. The C particles are of similar morphology, reflecting the inorganic template used to form both C particles. Because of the extreme heat and acidity required during particle synthesis, there was the expected lack of digestion of the particles in the lysosome. Hence, there was difficulty in morphologically differentiating phagosomes and phagolysosomes in this regard.
Comparison of C and C/Fe particulates on generating an oxidative burst in the MDMs.
To assess bioactivity related to these synthesized particulates, we performed luminol assays to measure the release of ROSs on phagocytosis (Nadeau et al. 1996). These experiments showed that the oxidative burst reached its peak in the C/Fe-exposed cells after 20 min and then gradually diminished (Figure 2). For the C-exposed cells, no oxidative burst was detected.
Data presented in Figure 2 provide evidence of the differential abilities of the particulates in inducing stress as a function of their physicochemical characteristics. Although C particles induced no detectable oxidative burst in treated MDMs, C/Fe particles induced a nearly 5-fold increase in extracellular oxidative burst by treated MDMs compared with untreated controls.
Impact of the C particulates on the MDMs.
To examine the kinetics of cell/particulate interaction, we fixed the cells 2–4 and 24 hr after exposure to particulates.
At 2 hr postexposure, ultrastructural signs of macrophage activation were evident with generalized dilatation of endoplasmic reticulum (ER). Numerous C particulates were seen within intact lysosomes (Figure 3). At 24 hr, the C particulates were often in small clumps and without evidence of surrounding lysosomal membrane. There was no evidence of adjacent organelle lysis or of agglomeration of particulates (Figure 4).
Impact of the C/Fe particulates on the MDMs.
By the first sampling (4 hr), particulates of similar size and contours as the stock preparation of C/Fe particulates (as seen by SEM) were present within cells. Many particulates were in lysosomes with a detectable membrane surrounding them (Figure 5). In others the membrane could not be followed. Varying degrees of dilation of the ER could be seen in regions adjacent to particulates. The nucleus at this point appeared normal.
By 24 hr, the particulates were sometimes partially or markedly aggregated (Figures 6 and 7). Although there was extensive dilatation of the ER, other organelles such as mitochondria appeared morphologically normal. The nucleus at this point continued to appear morphologically normal. In occasional cells (< 10%) still in the 24 hr fixation group, the process seemed more advanced. The agglomerated mass of particulates appeared spherical in overall shape and compact (Figure 7). There was a nearly uniform zone of total organelle lysis surrounding the agglomerate, with less than total lysis surrounding the lytic zone. This lysis extended to include the outer cell membrane. The nucleus by this time had undergone loss of chromatin. The nuclear membrane appeared to still be intact and surrounded the nucleus, which was lacking in chromatin.
Impact of the C and C/Fe particulates on T-cells.
The control T-cells and T-cells exposed to the C or C/Fe particulates showed no detectable ultrastructural changes. Random fields revealed healthy appearing cells. No evidence of internalized C or C/Fe particulates could be seen in the experimentally exposed group or controls. In both categories, the nuclei were generally oval and had the expected abundant heterochromatin. No ultrastructural changes were detected between the two categories.
Agglomeration of particulates.
To determine the role of serum proteins in the agglomeration of C/Fe particulates, free particulates were incubated for 24 hr in complete or serum-free leukocyte culture medium, washed in distilled water, and examined by SEM. The C/Fe particulates, which had been suspended in the serum-free medium, showed no tendency to agglomerate (Figure 8). In contrast, the C/Fe particulates incubated in complete medium (10% pooled human serum) showed a profound degree of agglomeration (Figure 9) with tightly packed clusters of particulates in a spherical conformation. The clusters were remarkably similar in size (~ 10–12 μm) to those observed intracellularly.
Particle-bound iron and hydroxyl radical formation.
We observed the propensity of the particles to catalyze the formation of hydroxyl radicals from hydrogen peroxide by ESR spectroscopy. The procedure involved trapping of the hydroxyl radical with DMPO (Shi et al. 2003). The four-line ESR signal (1:2:2:1 quartet) characteristic of 2,2-dimethyl-5-hydroxy-1-pyrrolidinyloxyl (DMPO-OH) with a hyperfine splitting constant of 14.3 G was clearly observed after exposure of C/Fe particles to hydrogen peroxide (Figure 10). The signal after exposure of C particles to hydrogen peroxide and DMPO was smaller by a factor of 10 (Figure 10). Thus, C/Fe particles form hydroxyl radicals with much greater propensity than do C particles. The radicals formed in the presence of carbon must arise from trace levels of iron present in the sample. It is known that the zeolite template itself contains trace iron (McNicol and Pott 1970).
Discussion
The physical properties of both the C and C/Fe particulates were such that sections for ultrastructural evaluation could be prepared without sectioning artifacts. The occasional absence of lysosomal membranes around particulates (e.g., Figure 4) might be attributed to cell-harvesting procedures. The fact that there was no distortion or displacement of organelles would indicate that the scraping and other aspects of the harvesting procedure did not hinder the interpretation.
Epidemiologic studies have demonstrated that there is a direct correlation between exposure to small airborne particulates and human disease. However, the mechanism(s) by which these particulates exacerbate pulmonary or cardiovascular disease is not known. To begin to address the role of iron in these processes, we synthesized carbon-based particles with or without iron. It was interpreted that cell features associated with activation seen in the MDMs exposed to the C particulates only were a consequence of the activation brought about by the extensive phagocytosis of particulates. There was no lysis of organelles in cells treated with C particles.
Consider the functional significance of some of the C/Fe-exposed cells having intact lysosomal membranes and ER, whereas others did not appear to be related to the length of time of exposure: At the 4-hour sampling, both structures seemed to be morphologically intact (but with dilated ER), whereas at 24 hr, the lysosomal membranes around the agglomerates were often ruptured and there was extensive ER dilation and vacuole formation (approximately in proportion to distance away from the agglomerate). Thus, both time and formation of the large agglomerate were positively correlated with the intracellular lesions. From a functional standpoint, it seems likely that sequelae from generation of ROSs could require time as well as sufficient concentration in an agglomerate to reach the threshold to alter the ultrastructural morphology.
We also observed the following regarding the phenomenon of intracytoplasmic agglomeration of C/Fe particulates: At 4 hr postexposure, numerous particulates had become internalized. These were interpreted as being within phagosomes because the membrane could generally be followed around these variably sized particulates. At 24 hr, the particulates were often seen to be forming into much larger clusters but still generally within the lysosomal membranes. At an apparently more advanced phase, the particulates had sometimes formed a dense spherical cluster, surrounded by a zone of nearly total organelle lysis. The cause of such an intracellular lesion is not certain, but it would appear to be compatible with that resulting from reactive oxygen metabolite generation from the C/Fe particles.
To explain the propensity of the C/Fe particulates to form spherical clusters within cells, we tested cell-free medium with and without serum. The agglomerates formed after the addition of C/Fe particulates (but not with C-only particulates) providing that serum was present. The spherical clusters were often approximately 10 μm in overall diameter whether in the cellular or noncellular system. The reason why the individual particulates in the intracellular clusters were smaller than particulates initially ingested by the macrophage is not apparent. One explanation may be that the large C/Fe conglomerate formation underwent a fracturing process associated with the cutting effect by the microtome blade. Given that during synthesis of the C/Fe particulates, the material is subjected to extreme acidification procedures, the acidification reached in a lysosome would not be sufficient to decompose the particle. Hence, it seems unlikely that degradation on a purely chemical basis could have occurred in the biologic environment of the lysosome and that the apparent smaller particle size is an artifact of the preparation associated with the slicing of a dense agglomeration of graphitic particles.
Regarding the occurrence of agglomerates of particulates observed within cells, two possibilities seem to exist as to their formation. One possibility is that the agglomerates developed outside of the cell in the culture medium (Figure 9) and were subsequently phagocytized as a preformed agglomerate. The demonstrated ability of serum proteins to bring about particulate agglomeration makes this a possibility. It has been recently noted that lung-lining liquid (which would contain surfactant) modifies PM2.5 to bring about particle aggregation (Kendall et al. 2002). The other possibility is that individual or small groups of particulates were initially phagocytized and became subsequently agglomerated within the cell. If there were areas of sufficient lysis and loss of intracellular structure in these areas, it would seem feasible that forces (hydrophobicity and others) could enable intracellular clusters of particulates to develop into agglomerates (Figure 7). In our consideration, this last proposal seems more likely in that only individual or small clusters were seen to have been phagocytized in the earliest examinations (4 hr) after exposure of the cell to particulates had occurred.
The T-lymphocytes, similarly exposed, failed to internalize either the C particulates or the C/Fe particulates and subsequently showed no ultrastructural lesions. Thus, we suggest that the failure to internalize the C/Fe particulates avoids the uncontrolled entry of iron into the cell and thus enables the T-lymphocytes to avoid the particulate-induced damage.
Although the oxidative burst of MDMs exposed to the C/Fe particulates occurred early (~20 min) postexposure, the morphologic evidence of ultrastructural change was not detected until much later. The lesions were not seen at the 4 hr sampling but were seen at the 24 hr sampling. The reason the morphologic changes were not evident until later is not clear. One possibility is that cell injury may have occurred early but did not manifest morphologic changes until later. Another possibility is that the lesions were a consequence of the concentration of the possibly toxic products resulting from the coalescence of multiple small phagolysosomes with their contents. Still another possibility is that the physical compression on adjacent organelles as a consequence of the mass of the agglomerate might be considered. Of these, it seems most feasible that the lesions were related to concentration of possibly toxic secondary products produced by a chemical reaction from the agglomerate. The C/Fe particulates contain iron on the surface, and the cell has been demonstrated to undergo an oxidative burst on phagocytosis of the particulates. A likely oxidant produced by the cell during the oxidative burst is hydrogen peroxide. The C/Fe particle can promote the formation of hydroxyl radicals via the Fenton reaction:
The ultrastructural changes noted in this study are consistent with macrophage activation and damage, which could be explained by the uncontrolled formation of ROSs.
Summary and Conclusions
Synthetic C and C/Fe particulates (1 μm) were given to cultures of human T-cells and MDMs. The T-cells failed to ingest either particles and showed no ultrastructural changes. The MDMs avidly ingested both type of particles. In contrast, those receiving C particulates showed only ultrastructural changes associated with cell activation. Those receiving C/Fe particulates by 24 hr showed evidence of clustering and coalescence of particulates. A highly discrete, concentrated mass of particulates was sometimes surrounded by a zone of total organelle lysis. Evidence that the C/Fe particulates were bioactive was demonstrated by a nearly 5-fold increase in oxidative burst by treated MDMs. Similar cells exposed to C particulates showed no increase in this regard. The synthetic C/Fe particulates also produced hydroxyl radicals on exposure to hydrogen peroxide. We hypothesize that the formation of intracellular ROSs is responsible for the ultrastructural changes observed. Results of these studies demonstrate that particle-induced ultrastructural changes depend on phagocytosis and suggest that, among respirable particulates of similar size, biologic activity can vary profoundly as a function of particulate physicochemical properties.
Figure 1 C/Fe particulates seen with SEM (magnification 10,000×). Note size and degree of uniformity. Bar = 1 μm.
Figure 2 Luminol assay to assess levels of oxidative burst by human MDMs after exposure to C or C/Fe particulates. Error bars indicate mean ± SD of four replicates.
Figure 3 Human MDMs (magnification, 11,000×) 2 hr after exposure to C particulates. Note particulates within intact lysosomes. Bar = 2 μm.
Figure 4 Human MDMs (magnification, 9,000×) 24 hr after exposure to C particulates. Note clumps of individual particulates free in cytoplasm. Bar = 2 μm.
Figure 5 Human MDMs (magnification, 11,000×) at 4 hr after exposure to C/Fe particulates. Note particulates within lysosomes (arrow points to membranes). Note dilatation of ER. Bar = 2 μm.
Figure 6 Human MDMs (magnification, 11,000×) 24 hr after exposure to C/Fe particulates. Note tendency for fine particulates to form a loose agglomerate. Some particulates are still within membrane-bound phagolysosomes. The nucleus appears to be nearly structurally normal. Bar = 2 μm.
Figure 7 Human MDMs (magnification, 4,900×) 24 hr after exposure to C/Fe particulates. Note chromatolysis of nucleus and the dense cluster of fine particulates with a surrounding zone of lysis. The lysis extends entirely to cell membrane, which is still apparently intact morphologically. Bar = 2 μm.
Figure 8 SEM preparation (magnification, 1,000×) after exposure in serum-free medium to suspension of C/Fe particulates. Note no tendency to agglomerate. Bar = 10 μm.
Figure 9 SEM preparation (magnification, 1,000×) after exposure in serum-containing medium to C/Fe particulates. Note tendency to agglomerate in about 10–12 μm clusters. Bar = 10 μm.
Figure 10 Electronic paramagnetic resonance spectroscopy spectra of the DMPO-OH adduct for C particles (A) and C/Fe particles (B). Both spectra shown on same scale (10 Gauss).
==== Refs
References
Boyum A 1968 Separation of leukocytes from blood and bone marrow Scand J Clin Lab Invest 21 77 89 5637478
Clouse K Adams P Sheridan J Orosz C 1987 Rapid method for purification of human T lymphocytes for further functional studies J Immunol Methods 105 253 262 3500984
Darling DC Morgan SJ 1994. Animal Cells, Culture and Media. New York:John Wiley & Sons.
Fach E Kristovich R Long JF Waldman WJ Dutta Prabir K Williams MV 2003 The effect of iron on the biological activities of erionite and mordenite Environ Int 29 451 458 12705942
Fach E Waldman WJ Williams M Long J Meister RK Dutta Prabir K 2002 Analysis of the biological and chemical reactivity of zeolite-based aluminosilicate fibers and particulates Environ Health Perspect 110 1087 1096 12417479
Fubini B Giamello E Mollo L Zanetti G Eborn SK Aust AE 1999 Zeolites as model solids for investigations on the role of iron at the solid-liquid interface in particulate toxicity Res Chem Intermed 25 95 109
Gilmour PS Brown DM Lindsay TG Beswick PH MacNee W Donaldson K 1996 Adverse effects of PM10 particles, involvement of iron in generation of hydroxyl radical Occup Environ Med 53 817 822 8994401
Hardy JA Aust AE 1995 Iron in asbestos chemistry and carcinogenicity Chem Rev 95 97 118
Johnson SA Brigham ES Ollivier PJ Mallouk TE 1997 The effect of micropore topology on the structure and properties of zeolite polymer replicas Chem Mater 9 2448 2458
Kendall M Tetley TD Wigzell E Hutton B Nieuwenhuijsen M Luckham P 2002 Lung lining liquid modifies PM(2.5) in favor of particle aggregation: a protective mechanism Am J Physiol Lung Cell Mol Physiol 282 L109 L114 11741822
McNicol BD Pott GT 1970 Fe3+ ions in crystalline aluminosilicate frameworks: electron spin resonance, phosphorescence, and thermal studies J Chem Soc D Chem Commun 7 438
Nadeau D Vincent R Kumarathasan P Brook J Dufresne A 1996 Cytotoxicity of ambient air particles to rat lung macrophages: comparison of cellular and functional assays Toxicol In Vitro 10 2 161 172 20650194
Shi T Schins RPF Knaapen AM Kuhlbusch T Pitz M Heinrich J 2003 Hydroxyl radical generation by electron paramagnetic resonance as a new method to monitor ambient particulate matter composition J Environ Monit 5 550 556 12948226
Smith KR Veranth JM Hu AA Lighty JS Aust AE 2000 Interleukin-8 levels in human lung epithelial cells are increased in response to coal fly ash and vary with the bioavailability of iron, as a function of particle size and source of coal Chem Res Toxicol 13 118 125 10688536
Valavanidis A Salika A Theodoropoulou A 2000 Generation of hydroxyl radicals by urban suspended particulate air matter. The role of iron ions Atmos Environ 34 2379 2386
Waldman WJ Adams PW Orosz CG Sedmak DD 1992 T lymphocyte activation by cytomegalovirus-infected, allogeneic cultured human endothelial cells Transplantation 54 887 896 1332226
|
15687054
|
PMC1277860
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):170-174
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7389
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7252ehp0113-00017515687046ResearchArticlesExposure to PCBs and p,p′-DDE and Human Sperm Chromatin Integrity Rignell-Hydbom Anna 1Rylander Lars 1Giwercman Aleksander 2Jönsson B.A.G. 1Lindh Christian 1Eleuteri Patrizia 3Rescia Michele 3Leter Giorgio 3Cordelli Eugenia 3Spano Marcello 3Hagmar Lars 11Department of Occupational and Environmental Medicine, Lund University Hospital, Lund, Sweden2Fertility Centre, Malmö University Hospital, Malmö, Sweden3Section of Toxicology and Biomedical Sciences, ENEA Casaccia Research Centre, Rome, ItalyAddress correspondence to A. Rignell-Hydbom, Department of Occupational and Environmental Medicine, Lund University, SE-221 85 Lund, Sweden. Telephone: 46-46-177280. Fax: 46-46-173669. E-mail:
[email protected] work was supported by grants from the European Commission, Quality of Life and Management of Living Resources, Key Action Four on Environment and Health (contract QLK4-CT-2001-00202), the Swedish Research Council, the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning, and the Medical Faculty, Lund University.
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 175 179 13 5 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Persistent organochlorine pollutants (POPs) such as polychlorinated biphenyls (PCBs) and dichlorodiphenyldichloroethylene (p,p′-DDE), the major metabolite of dichlorodiphenyltrichloroethane (DDT), are stable lipophilic compounds widely found in the environment and in the general population. They can enter the food chain, and their negative impact on male reproduction is currently under active scrutiny. To explore the hypothesis that environmental exposure to these compounds is associated with altered sperm chromatin structure integrity in human sperm, we conducted a study of 176 Swedish fishermen (with low and high consumption of fatty fish, a very important exposure source of POPs). We determined serum levels of 2,2′,4,4′,5,5′-hexachlorobiphenyl (CB-153) and p,p′-DDE, and we used the sperm chromatin structure assay (SCSA) to assess sperm DNA/chromatin integrity. When CB-153 serum levels (individual dose range, 39–1,460 ng/g lipid) were categorized into equally sized quintiles, we found an association with the DNA fragmentation index (%DFI). A significantly lower %DFI was found in the lowest CB-153 quintile (< 113 ng/g lipid) compared with the other quintiles; there was a similar tendency, although not statistically significant, between %DFI and p,p′-DDE. These results suggest that POP exposure may have a slight negative impact on human sperm chromatin integrity.
DDEpolychlorinated biphenylssperm chromatin integritysperm chromatin structure assay (SCSA)
==== Body
Polychlorinated biphenyls (PCBs), widely used in the past in cutting oils and lubricants and as an electric insulator, were restricted or totally banned in the 1970s in most developed countries, together with the insecticide dichlorodiphenyltrichloroethane (DDT). However, because of their high persistence to both biotic and abiotic degradation and their ability to bioaccumulate, these persistent organochlorine pollutants (POPs) continue to be a potential health hazard for the general population as they enter the food chain.
In Sweden, the consumption of fatty fish, such as salmon and herring, from the Baltic Sea off the Swedish east coast represents a major exposure source of PCBs, DDT, and its major metabolite, dichlorodiphenyldichloroethylene (p,p′-DDE). Fatty fish species from the Baltic Sea are much more contaminated with PCBs and p,p′-DDE than are corresponding fish from the Swedish west coast (Bergqvist et al. 1989). This is also the case with other POPs such as polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs). This was reflected in higher average plasma levels of dioxin-like POPs among east coast fishermen (290 pg/g lipid) than among west coast fishermen (139 pg/g lipid) and men from the general Swedish population (123 pg/g lipid) (Svensson et al. 1995).
PCBs are not a uniform group of compounds with similar biologic effects. Theoretically there are 209 PCB congeners, varying in the degree of chlorination and the position of chlorine atoms, which affect their stability and toxicity. In reality, fewer can be detected in the environment. The PCB congener 2,2′,4,4′,5,5′-hexachlorobiphenyl (CB-153) is a useful biomarker of dietary exposure to POPs because it correlates very well with both total PCB concentration (Gladen et al. 1999; Glynn et al. 2000; Grimvall et al. 1997), the 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) equivalent (TEQ), and the total POP derived TEQ (Brouwer et al. 1995; Gladen et al. 1999). Another relevant biomarker is the antiandrogenic compound p,p′-DDE, which is present in relatively high serum concentrations in men consuming fatty fish from the Baltic Sea (Sjodin et al. 2000).
Several studies on wildlife and laboratory animals have shown that exposure to PCBs and p,p′-DDE is capable of interfering with reproductive and endocrine functions (Cooke et al. 1996; Faqi et al. 1998; Fry 1995; Guillette et al. 1994, 1996; Hsu et al. 2003b; Kim 2001).
Human studies have shown that accidentally high exposures to PCBs and PCDFs have a negative effect on male reproductive function (Guo et al. 2000; Hsu et al. 2003a). Even more interesting is that lower exposure levels to PCBs, relevant for the general population in Western countries, have been associated with effects on sperm motility, sperm concentration, and total sperm count (Bush et al. 1986; Dallinga et al. 2002; Hauser et al. 2003a; Richthoff et al. 2003). Sperm motility seems to be especially vulnerable.
Sperm DNA integrity is essential for the accurate transmission of genetic information, and sperm chromatin abnormalities or DNA damage may result in male infertility (Agarwal and Said 2003). There are only a few human studies relating PCBs and p,p′-DDE levels in biologic fluids to sperm genetic integrity, which seems pivotal for the full expression of individual fertility potential (Hauser et al. 2003b; Rozati et al. 2002).
Among the variety of new methods to study sperm genetic integrity, sperm chromatin structure assay (SCSA) is considered as one of the most stable, robust, and objective (Evenson et al. 2002; Perreault et al. 2003). SCSA seems particularly fit for epidemiologic surveys because only a small (0.1 mL) amount of semen is needed for the analysis, and it can be frozen, stored, and assayed at the end of the study, minimizing interassay variation (Perreault et al. 2000). SCSA has been used in a number of epidemiologic studies among men exposed to pesticides, lead, styrene, solvents, and air pollution (Bonde et al. 2002; Kolstad et al. 1999; Larsen et al. 1998b; Perreault et al. 2000; Sanchez-Pena et al. 2004; Selevan et al. 2000) but not previously among POP-exposed subjects.
The aim of this study was to investigate whether serum levels of CB-153 and p,p′-DDE were associated with sperm chromatin damage assessed by SCSA.
Materials and Methods
Study population.
Cohorts of fishermen from the Swedish east and west coasts were established in 1988 (Svensson et al. 1995). In 2000, a postal questionnaire, focused mainly on fracture incidence, was sent to 3,505 west coast fishermen and 1,678 east coast fishermen, born 1935 or later (Figure 1). The questionnaire included a question about whether the subjects were interested in more information on a study of semen function. Among the 2,614 subjects (east, n = 848; west, n = 1,766) who responded to this specific question, 479 (east, n = 171; west, n = 308) wanted more information about the semen study. We contacted these subjects and another 169 east coast fishermen who had become members of the east coast fishermen’s union after the closure of the cohorts. From the east coast, 130 of 340 men wanted to participate and gave their written informed consent. The corresponding figures from the west coast were 136 of 308. Thirty-four subjects from the east coast and 37 from the west coast were excluded for logistical reasons, changes of mind, sickness, or recent vasectomy during the field study period. In the end, 195 men participated in the semen study, and the results of standard semen analyses have been published previously (Rignell-Hydbom et al. 2004). Because of limited amounts of semen, samples from only 176 men could be used for SCSA.
Nonparticipants.
The nonparticipants from the fishermen’s cohort had similar age distribution (median, 52 years; range, 29–67 years) as the participants in the present study [median, 48 years; range, 29–67 years]. The participants had on average 2.0 children. We do not have any directly comparable data for the nonparticipants, but a previous study showed that during 1973–1991 fishermen’s wives on average gave birth to 2.0 infants (Rylander et al. 1995). In addition, the body mass index (BMI) distributions and fraction of smokers were very similar among the participants and the nonparticipants.
Questionnaire.
Approximately 2 weeks before telephone contact, a questionnaire regarding lifestyle and medical and reproductive history was sent out to the fishermen. In this way, the participants had time to get acquainted with the questions that they were interviewed on later. During the telephone contact, an agreement was reached on time and date for collection of semen and blood samples at the subject’s home. The participants received information on the procedures for collecting the semen samples both in verbal and written form. The study was approved by the Ethical Committee at Lund University.
Mobile laboratory unit and semen and blood sampling.
A mobile laboratory unit was established for this study. The subjects were asked to keep 3 days’ abstinence time before sample collection (median, 3 days; range, 1–21 days), which took place in the participant’s homes. Immediate semen analyses were performed within 1 hr after ejaculation (Rignell-Hydbom et al. 2004). Two tubes with 200-μL aliquots of undiluted raw semen, collected 30 min after liquefaction, were directly put into a box with dry ice and shortly thereafter transferred into a freezer at −80°C. Venous blood samples were collected and centrifuged in the mobile laboratory, and sera were frozen at −80°C for later analysis.
Sperm chromatin structure assay.
The frozen samples were transported for flow cytometry (FCM) SCSA analysis to the Section of Toxicology and Biomedical Sciences, ENEA Casaccia, Rome, Italy. The samples were quickly thawed in a 37°C water bath and analyzed immediately. The SCSA was applied following the procedure described elsewhere (Evenson et al. 2002; Spano et al. 2000). A total of 1–2 × 106 cells were treated with a detergent solution (pH 1.2) containing 0.1% Triton X-100, 0.15 M NaCl, and 0.08 N HCl for 30 sec and then stained with 6 mg/L of purified acridine orange (AO; Molecular Probes, Eugene, OR, USA) in a phosphate-citrate buffer, pH 6.0. All measurements began 3 min after AO staining. Cells were analyzed by a FACScan (Becton Dickinson, San Jose, CA, USA) equipped with an air-cooled argon ion laser and standard optical filters to collect green and red fluorescence. A total of 10,000 events were accumulated for each measurement. Under these experimental conditions, when excited with a 488 nm light source, AO, when intercalated with double-stranded DNA emits green fluorescence, whereas AO associated with single-stranded DNA emits red fluorescence. Thus, sperm chromatin damage can be quantified by the FCM measurements of the metachromatic shift from green (native, double-stranded DNA) to red (denatured, single-stranded DNA) fluorescence and displayed as red (fragmented DNA) versus green (DNA stainability) fluorescence intensity cytogram patterns. Off-line analysis of the flow cytometric data was carried out by using dedicated software (SCSASoft, SCSA Diagnostics, Brookings, SD, USA). Computer gates are used to determine the proportion of spermatozoa with increased levels of red and green fluorescence, respectively. We have expressed the extent of DNA denaturation in terms of DNA fragmentation index (DFI), which is the ratio of red to total (red plus green) fluorescence intensity (Evenson et al. 2002). The DNA fragmentation index (DFI) value was calculated for each sperm cell in a sample, and the resulting DFI frequency profile for the entire sperm population was obtained (Figure 2A). The normal population of sperm with no detectable DNA damage forms a unimodal distribution. The fraction of sperm with higher red fluorescence intensity represents the population of abnormal sperm with detectable DNA damage. It is expressed as the percentage of sperm showing DNA fragmentation (%DFI). Additionally, we have also considered the fraction of high-DNA-stainable (HDS) cells, which represent immature spermatozoa with incomplete chromatin condensation. The percentage of HDS cells was calculated by setting an appropriate gate on the bivariate cytogram (Figure 2B) and considering those events that exhibit green fluorescence intensity higher than the upper border of the main cluster of the sperm population with a nondetectable %DFI as immature spermatozoa.
For the flow cytometer setup and calibrations, a reference semen sample retrieved from the laboratory repository was used. Samples were measured twice during independent FCM sessions, and the average value was used. Results from the two measurements were highly correlated (DFI, r = 0.96; HDS, r = 0.96).
Determination of CB-153 and p,p′-DDE.
The levels of CB-153 and p,p′-DDE were determined as previously described (Rignell-Hydbom et al. 2004). Briefly, CB-153 and p,p′-DDE were extracted from serum by solid-phase extraction (Isolute ENV+; IST, Hengoed, UK) using on-column degradation of the lipids and analysis by gas chromatography mass spectrometry. 13C12-Labeled CB-153 and 13C12-labeled p,p′-DDE were used as internal standards. The selected ion monitoring of p,p′-DDE was performed at m/z 318, whereas m/z 330 was used for the internal standard. The relative standard deviations, calculated from samples analyzed in duplicate at different days, for CB-153 was 7% at 0.6 ng/mL (n = 76) and 5% at 1.5 ng/mL (n = 37) and for p,p′-DDE was 12% at 0.6 ng/mL (n = 56) and 7% at 2.4 ng/mL (n = 50). The detection limits were 0.05 ng/mL for CB-153 and 0.1 ng/mL for p,p′-DDE. The analyses of CB-153 and p,p′-DDE were part of the Round Robin inter-comparison program (H. Drexler, Institute and Out-Patient Clinic for Occupational, Social and Environmental Medicine, University of Erlangen-Nuremberg, Erlangen, Germany) with analysis results within the tolerance limits.
Determination of lipids by enzymatic methods.
Serum concentrations of triglycerides, cholesterol, and phospholipids were determined by enzymatic methods using triglycerides and cholesterol from Boehringer-Mannheim (Mannheim, Germany) and phospholipids from Waco Chemicals (Neuss, Germany). The total lipid concentration in serum was calculated by summation of the amounts of triglycerides, cholesterol, and phospholipids. In these calculations, the average molecular weights of triglycerides and phospholipids were assumed to be 807 and 714. For cholesterol we used an average molecular weight of 571, assuming that the proportion of free and esterified cholesterol in serum was 1:2.
Hormone analyses.
Serum concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), and estradiol were analyzed with immunofluorometric techniques. The total assay variation coefficients were 2.9, 2.6, and 8.1%, respectively. Serum testosterone and sexual-hormone–binding globulin (SHBG) were measured by commercially available immunoassays. The total assay variation coefficients were 5.5 and 4.6%, respectively. Inhibin B levels were assessed using a specific immunometric method, as previously described, with a detection limit of 15 ng/L and intra-assay and total assay variation coefficients < 7% (Groome et al. 1996).
Statistical analysis.
In linear regression models, we evaluated the effect of CB-153 and p,p′-DDE as exposure variables on the outcome variables %DFI and HDS (Table 1). The CB-153 and p,p′-DDE variables were treated as continuous variables (untransformed and log transformed) as well as categorized variables (into equally sized quintiles). CB-153 and p,p′-DDE levels correlated strongly (r = 0.78), and these variables were not included in the model at the same time. Accordingly, no interaction analyses were performed. Because of the skewed distributions of the %DFI and HDS variables, we also tested whether log transformation of these variables better fulfilled the model assumptions, which was checked by means of residual analysis. As potential confounders, we initially considered age (as a continuous variable), current smoking status (yes/no), abstinence time (as continuous or categorized into 0–2, > 2–4, > 4–6, > 6 days), BMI (as continuous), and levels of testosterone, SHBG, FSH, LH, estradiol, and inhibin B in serum and the testosterone:SHBG ratio. However, there were no associations between the exposure variables and abstinence time (rs < 0.03), smoking (mean difference in CB-153 and p,p′-DDE < 7%), inhibin B (rs < 0.06), LH (rs < 0.02), testosterone (rs < 0.08), or estradiol (rs < 0.04). Thus, the above-mentioned variables were excluded from further analyses. The remaining variables were included in the models, one at a time, together with the exposure variable if they showed any association (p < 0.20) with %DFI or HDS. If the adjusted effect estimates (i.e., the effect of exposure on %DFI and HDS, respectively) differed < 15% from the crude estimates, we only present the crude estimates.
Results
There was no significant correlation between the SCSA parameters ln %DFI and ln HDS (r = −0.097, p = 0.20). Serum levels of CB-153 and p,p′-DDE ranged from 39 to 1,460 and from 40 to 2,251 ng/g lipid, respectively (medians, 189 and 240 ng/g lipid).
In univariate analysis ln CB-153 was associated with ln %DFI (r = 0.27, p < 0.001; Figure 3). However, when age, which was strongly associated with %DFI, simultaneously was included in the model, this association was no longer significant (p = 0.28). On the other hand, when CB-153 was categorized into five equally sized quintiles, there seemed to be an effect (Figure 4). The quintile with the lowest exposure had significantly lower levels of %DFI compared with the other quintiles (p < 0.001). This effect remained when age was included in the model (p = 0.006). The four highest exposed quintiles (> 113 ng/g lipid) had 41% (95% CI, 11–78) higher %DFI compared with the lowest exposed quintile. None of the other potential confounders changed this estimate more than marginally. In addition, the DFI levels in the four highest exposed quintiles did not differ from each other (all p-values > 0.20). Regarding p,p′-DDE exposure, the pattern was less clear. When age was included in the model (p = 0.10), the lowest exposed group (< 136 ng/g lipid) did not significantly differ from the four highest exposed quintiles (22%; 95% CI, −4 to 53; Figure 5). However, the exposure–response pattern for p,p′-DDE with respect to %DFI was similar as for CB-153 but not statistically significant. Neither CB-153 nor p,p′-DDE was associated with HDS (all p-values > 0.25).
Discussion
The main result of the present study was a positive association between serum levels of CB-153 and %DFI, indicating that POP exposure might affect sperm DNA integrity. A certain fraction of sperm with DNA breaks is always present in normal ejaculates. Clinical studies using SCSA have, however, demonstrated that the fecundability of a couple is negatively correlated with %DFI when it exceeds 20% (Spano et al. 2000) and becomes negligible for DFI > 30% (Evenson et al. 1999).
Low participation rates and potential selection bias are of great concern in all human semen studies. Both age and demonstrated fertility have an impact on participation rate in semen studies (Larsen et al. 1998a). In the present study, the age distributions as well as the average number of children were very similar among the participants and the nonparticipants. Therefore, we do not suspect that selection bias concerning these factors is of major concern for the interpretation of the results.
A positive correlation (r = 0.30, p < 0.001) between %DFI and the age of the men enrolled in this study was found. Such an association has been shown previously (Singh et al. 2003; Spano et al. 1998), suggesting less efficient apoptotic mechanisms operating during or after spermatogenesis in aging men. Our study showed, however, an association between serum levels of CB-153 and %DFI, also after adjusting for age. The association was nonlinear, indicating a threshold effect. On the other hand, no associations were found between serum levels of CB-153 and HDS, the SCSA parameter that mirrors the fraction of sperm with defects of the proteic component of the chromatin that characterizes immature spermatozoa (Evenson et al. 2002). The correlation between CB-153 and p,p′-DDE was high in our study, and the exposure–response pattern for p,p′-DDE with respect to the outcome variable %DFI showed results pointing into the same direction as for CB-153, but the results were weaker and not statistically significant.
The association between the CB-153 levels and %DFI is in accordance with our recent reports on a correlation between CB-153 and decreased sperm motility (Richthoff et al. 2003; Rignell-Hydbom et al. 2004). A circumstantial evidence supporting the biologic relevance of our observation was that sperm motility was shown to decrease with increasing levels of %DFI (Giwercman et al. 2003). Most of the published SCSA studies, based on infertility patients, report weak, negative significant associations between the %DFI and the parameters from semen quality assessment. In the present study, we found a weak to moderate association between %DFI and sperm motility (r = −0.37), which is in close agreement with a study by Giwercman et al. (2003).
There are only two previous studies regarding the association between POP exposure and sperm chromatin damage in humans. In a study carried out in India, 21 infertile men were compared with 32 men with normal semen analyses and evidence of conception (Rozati et al. 2002). Sperm nuclear chromatin integrity was assessed by the chromatin condensation assay using AO staining, and the DNA integrity was monitored under a fluorescent microscope. There was a significant positive correlation between seminal total PCB level and the percentage of single-stranded DNA in sperm. In the other study, carried out in the United States, the neutral single-cell microgel electrophoresis assay (Comet assay) was used to assess DNA integrity in 212 male partners of subfertile couples (Hauser et al. 2003b). The authors did not find any statistically significant associations between sperm DNA damage and serum levels of any individual PCB congeners, sum of PCBs, or p,p′-DDE, leading to the conclusion that there were no strong relationships. In the American study, CB-153 serum levels ranged from 9 to 421 ng/g lipid (median, 44 ng/g lipid), which is much lower compared with the present study (median, 189 ng/g lipid). The serum levels of p,p′-DDE were, however, rather similar (median, 254 ng/g lipid) with our present study.
The present study has two advantages compared with the previous ones assessing the association between POP exposure and DNA integrity. First, the study population includes men from the general population and not patients from infertility clinics. Second, SCSA is a computerized technique that ensures a theoretical detection sensitivity advantage, the analysis being based on a large enough number of cells.
A potential mechanism whereby PCBs may produce DNA damage is through hydroxylated PCB metabolites (OH-PCBs), which are found in human serum, in relatively high concentrations (Sjodin et al. 2000). These metabolites can be further oxidized to form semiquinons and quinons (McLean et al. 2000; Schlezinger et al. 1999), which are reactive electrophiles that may induce free-radical–mediated oxidative DNA damage and strand breaks (Li and Trush 1993). PCBs added to human hepatic cell lines increased DNA adduct formation (Borlak et al. 2003). Moreover, PCB quinones inhibited topoisomerase II activity (Srinivasan et al. 2002), which is of key importance for sperm nucleus remodeling.
In addition to more direct mechanisms of toxicity, it has to be considered that experimental data show that several POPs interact with steroid hormone homeostasis and thereby act as “endocrine disruptors” (Bonefeld-Jorgensen et al. 2001; Kester et al. 2002; Portigal et al. 2002; Shevtsov et al. 2003), but the relevance of these effects for sperm chromatin damage is unclear.
In conclusion, we found a statistically significant association between serum levels of CB-153 and %DFI, and a similar tendency although not significant for p,p′-DDE, in adult men. Further studies are needed to clarify the mechanism responsible for the association between POP exposure and the sperm chromatin integrity.
Figure 1 Flow chart for recruitment process of participants in the study.
Figure 2 (A) Frequency distribution histogram of the DFI. The area located to the right of the main peak (which includes normal sperm with non-detectable DFI) represents the region where the sperm with detectable levels of fragmented DNA accumulate (%DFI). (B) SCSA scattergram of red (fragmented DNA, x-axis) versus green (double-stranded DNA, y-axis) fluorescence intensity of the same semen sample. Cytogram dots represent single spermatozoa with dual-parameter green and red fluorescence values acquired at 10-bit resolution (1,024 channels) on the flow cytometer. Debris (B, left) was excluded from the analysis. The region for calculating the fraction of immature sperm with HDS is indicated by the line. The line indicates the threshold for HDS (channel 550 on the y-axis).
Figure 3 The association between the logarithm of the serum concentration of CB-153 and the logarithm of the DFI (r = 0.27, p < 0.001).
Figure 4 The association between serum concentration of CB-153 (divided into five groups) and the logarithm of DFI (p < 0.001).
Figure 5 The association between serum concentration of p,p′-DDE (divided into five groups) and the logarithm of DFI (p = 0.10).
Table 1 Distribution of SCSA results, exposure variables, and potential confounders in 176 Swedish fishermen.
Mean ± SD Median 5–95%
SCSA outcome variable
%DFI 19 ± 12 15 6–39
HDS (%) 10 ± 7 8 4–25
Exposure variables
CB-153 (ng/g lipid) 233 ± 178 189 63–552
p,p′-DDE (ng/g lipid) 334 ± 307 240 80–887
Potential confounder that did not fulfill the inclusion criteria for multivariate models
Current smoker (%) 23
Abstinence time (days) 3.8 ± 2.7 3.0 1–9
BMI (kg/m2) 27 ± 3.3 27 22–34
Serum LH (IU/L) 2.9 ± 1.2 2.7 1.3–5.2
Serum testosterone (nmol/L) 12.9 ± 5.8 11.7 6.7–22.6
Serum estradiol (pmol/L) 91 ± 43 83 50–158
Serum inhibin (ng/L) 192 ± 69 181 98–308
Serum FSH (IE/L) 4.0 ± 2.2 3.4 1.7–8.5
Serum SHBG (nmol/L) 31.4 ± 12.2 30.9 14.6–54.2
Serum HBG:testosterone 0.44 ± 0.21 0.41 0.23–0.76
Confounder included in the multivariate models
Age (years) 47 ± 9 48 32–63
==== Refs
References
Agarwal A Said TM 2003 Role of sperm chromatin abnormalities and DNA damage in male infertility Hum Reprod Update 9 331 345 12926527
Bergqvist PA Adami HO Mohner M Zatonski W Storm H Ekbom A 1989 Dioxins in cod and herring from the seas around Sweden J Natl Cancer Inst 88 727 733
Bonde JP Joffe M Apostoli P Dale A Kiss P Spano M 2002 Sperm count and chromatin structure in men exposed to inorganic lead: lowest adverse effect levels Occup Environ Med 59 234 242 11934950
Bonefeld-Jorgensen EC Andersen HR Rasmussen TH Vinggaard AM 2001 Effect of highly bioaccumulated polychlorinated biphenyl congeners on estrogen and androgen receptor activity Toxicology 158 141 153 11275356
Borlak J Hock A Hansen T Richter E 2003 DNA adducts in cultures of polychlorinated biphenyl-treated human hepatocytes Toxicol Appl Pharmacol 188 81 91 12691726
Brouwer A Ahlborg UG Van den Berg M Birnbaum LS Boersma ER Bosveld B 1995 Functional aspects of developmental toxicity of polyhalogenated aromatic hydrocarbons in experimental animals and human infants Eur J Pharmacol 293 1 40 7545581
Bush B Bennett AH Snow JT 1986 Polychlorobiphenyl congeners, p,p ′-DDE, and sperm function in humans Arch Environ Contam Toxicol 15 333 341 3090950
Cooke PS Zhao YD Hansen LG 1996 Neonatal polychlorinated biphenyl treatment increases adult testis size and sperm production in the rat Toxicol Appl Pharmacol 136 112 117 8560463
Dallinga JW Moonen EJ Dumoulin JC Evers JL Geraedts JP Kleinjans JC 2002 Decreased human semen quality and organochlorine compounds in blood Hum Reprod 17 1973 1979 12151423
Evenson DP Jost LK Marshall D Zinaman MJ Clegg E Purvis K 1999 Utility of the sperm chromatin structure assay as a diagnostic and prognostic tool in the human fertility clinic Hum Reprod 14 1039 1049 10221239
Evenson DP Larson KL Jost LK 2002 Sperm chromatin structure assay: its clinical use for detecting sperm DNA fragmentation in male infertility and comparisons with other techniques J Androl 23 25 43 11780920
Faqi AS Dalsenter PR Merker HJ Chahoud I 1998 Reproductive toxicity and tissue concentrations of low doses of 2,3,7,8-tetrachlorodibenzo-p -dioxin in male offspring rats exposed throughout pregnancy and lactation Toxicol Appl Pharmacol 150 383 392 9653070
Fry DM 1995 Reproductive effects in birds exposed to pesticides and industrial chemicals Environ Health Perspect 103 suppl 7 165 171 8593865
Giwercman A Richthoff J Hjollund H Bonde JP Jepson K Frohm B 2003 Correlation between sperm motility and sperm chromatin structure assay parameters Fertil Steril 80 1404 1412 14667876
Gladen BC Longnecker MP Schecter AJ 1999 Correlations among polychlorinated biphenyls, dioxins, and furans in humans Am J Ind Med 35 15 20 9884741
Glynn AW Wolk A Aune M Atuma S Zettermark S Maehle-Schmid M 2000 Serum concentrations of organochlorines in men: a search for markers of exposure Sci Total Environ 263 197 208 11194153
Grimvall E Rylander L Nilsson-Ehle P Nilsson U Stromberg U Hagmar L 1997 Monitoring of polychlorinated biphenyls in human blood plasma: methodological developments and influence of age, lactation, and fish consumption Arch Environ Contam Toxicol 32 329 336 9096084
Groome NP Illingworth PJ O’Brien M Pai R Rodger FE Mather JP 1996 Measurement of dimeric inhibin B throughout the human menstrual cycle J Clin Endocrinol Metab 81 1401 1405 8636341
Guillette LJ Jr Gross TS Masson GR Matter JM Percival HF Woodward AR 1994 Developmental abnormalities of the gonad and abnormal sex hormone concentrations in juvenile alligators from contaminated and control lakes in Florida Environ Health Perspect 102 680 688 7895709
Guillette LJ Jr Pickford DB Crain DA Rooney AA Percival HF 1996 Reduction in penis size and plasma testosterone concentrations in juvenile alligators living in a contaminated environment Gen Comp Endocrinol 101 32 42 8713642
Guo YL Hsu PC Hsu CC Lambert GH 2000 Semen quality after prenatal exposure to polychlorinated biphenyls and dibenzofurans Lancet 356 1240 1241 11072947
Hauser R Chen Z Pothier L Ryan L Altshul L 2003a The relationship between human semen parameters and environmental exposure to polychlorinated biphenyls and p,p ′-DDE Environ Health Perspect 111 1505 1511 12948891
Hauser R Singh NP Chen Z Pothier L Altshul L 2003b Lack of an association between environmental exposure to polychlorinated biphenyls and p,p ′-DDE and DNA damage in human sperm measured using the neutral comet assay Hum Reprod 18 2525 2533 14645167
Hsu PC Huang W Yao WJ Wu MH Guo YL Lambert GH 2003a Sperm changes in men exposed to polychlorinated biphenyls and dibenzofurans JAMA 289 2943 2944 12799401
Hsu PC Li MH Guo YL 2003b Postnatal exposure of 2,2′,3,3’,4,6’-hexachlorobiphenyl and 2,2′,3,4′,5′,6-hexachlorobiphenyl on sperm function and hormone levels in adult rats Toxicology 187 117 126 12699901
Kester MH Bulduk S van Toor H Tibboel D Meinl W Glatt H 2002 Potent inhibition of estrogen sulfotransferase by hydroxylated metabolites of polyhalogenated aromatic hydrocarbons reveals alternative mechanism for estrogenic activity of endocrine disrupters J Clin Endocrinol Metab 87 1142 1150 11889178
Kim IS 2001 Effects of exposure of lactating female rats to polychlorinated biphenyls (PCBs) on testis weight, sperm production and sertoli cell numbers in the adult male offspring J Vet Med Sci 63 5 9 11217063
Kolstad HA Bonde JP Spano M Giwercman A Zschiesche W Kaae D 1999 Change in semen quality and sperm chromatin structure following occupational styrene exposure. ASCLEPIOS Int Arch Occup Environ Health 72 135 141 10392560
Larsen SB Abell A Bonde JP 1998a Selection bias in occupational sperm studies Am J Epidemiol 147 681 685 9554607
Larsen SB Giwercman A Spano M Bonde JP 1998b A longitudinal study of semen quality in pesticide spraying Danish farmers. The ASCLEPIOS Study Group Reprod Toxicol 12 581 589 9875693
Li Y Trush MA 1993 DNA damage resulting from the oxidation of hydroquinone by copper: role for a Cu(II)/Cu(I) redox cycle and reactive oxygen generation Carcinogenesis 14 1303 1311 8392444
McLean MR Twaroski TP Robertson LW 2000 Redox cycling of 2-(x ’-mono, -di, -trichlorophenyl)-1,4-benzoquinones, oxidation products of polychlorinated biphenyls Arch Biochem Biophys 376 449 455 10775433
Perreault SD Aitken RJ Baker HW Evenson DP Huszar G Irvine DS 2003 Integrating new tests of sperm genetic integrity into semen analysis: breakout group discussion Adv Exp Med Biol 518 253 268 12817694
Perreault SD Rubes J Robbins WA Evenson DP Selevan SG 2000 Evaluation of aneuploidy and DNA damage in human spermatozoa: applications in field studies Andrologia 32 247 254 11021516
Portigal CL Cowell SP Fedoruk MN Butler CM Rennie PS Nelson CC 2002 Polychlorinated biphenyls interfere with androgen-induced transcriptional activation and hormone binding Toxicol Appl Pharmacol 179 185 194 11906248
Richthoff J Rylander L Jonsson BA Akesson H Hagmar L Nilsson-Ehle P 2003 Serum levels of 2,2,4,4,5,5-hexachlorobiphenyl (CB-153) in relation to markers of reproductive function in young males from the general Swedish population Environ Health Perspect 111 409 413 12676591
Rignell-Hydbom A Rylander R Giwercman A Jonsson BAG Nilsson-Ehle P Hagmar L 2004 Exposure to CB-153 and p,p ′-DDE and male reproductive function Hum Reprod 19 2066 2075 15284211
Rozati R Reddy PP Reddanna P Mujtaba R 2002 Role of environmental estrogens in the deterioration of male factor fertility Fertil Steril 78 1187 1194 12477510
Rylander L Stromberg U Hagmar L 1995 Decreased birth-weight among infants born to women with a high dietary intake of fish contaminated with persistent organochlorine compounds Scand J Work Environ Health 21 368 375 8571093
Sanchez-Pena LC Reyes BE Lopez-Carrillo L Recio R Moran-Martinez J Cebrian ME 2004 Organophosphorous pesticide exposure alters sperm chromatin structure in Mexican agricultural workers Toxicol Appl Pharmacol 196 108 113 15050412
Schlezinger JJ White RD Stegeman JJ 1999 Oxidative inactivation of cytochrome P-450 1A (CYP1A) stimulated by 3,3’,4,4′-tetrachlorobiphenyl: production of reactive oxygen by vertebrate CYP1As Mol Pharmacol 56 588 597 10462547
Selevan SG Borkovec L Slott VL Zudova Z Rubes J Evenson DP 2000 Semen quality and reproductive health of young Czech men exposed to seasonal air pollution Environ Health Perspect 108 887 894 11017895
Shevtsov S Petrotchenko EV Pedersen LC Negishi M 2003 Crystallographic analysis of a hydroxylated polychlorinated biphenyl (OH-PCB) bound to the catalytic estrogen binding site of human estrogen sulfotransferase Environ Health Perspect 111 884 888 12782487
Singh NP Muller CH Berger RE 2003 Effects of age on DNA double-strand breaks and apoptosis in human sperm Fertil Steril 80 1420 1430 14667878
Sjodin A Hagmar L Klasson-Wehler E Bjork J Bergman A 2000 Influence of the consumption of fatty Baltic Sea fish on plasma levels of halogenated environmental contaminants in Latvian and Swedish men Environ Health Perspect 108 1035 1041 11102293
Spano M Bonde JP Hjollund HI Kolstad HA Cordelli E Leter G 2000 Sperm chromatin damage impairs human fertility. The Danish First Pregnancy Planner Study Team Fertil Steril 73 43 50 10632410
Spano M Kolstad AH Larsen SB Cordelli E Leter G Giwercman A 1998 The applicability of the flow cytometric sperm chromatin structure assay in epidemiological studies. ASCLEPIOS Hum Reprod 13 2495 2505 9806274
Srinivasan A Robertson LW Ludewig G 2002 Sulfhydryl binding and topoisomerase inhibition by PCB metabolites Chem Res Toxicol 15 497 505 11952335
Svensson BG Nilsson A Jonsson E Schutz A Akesson B Hagmar L 1995 Fish consumption and exposure to persistent organochlorine compounds, mercury, selenium and methylamines among Swedish fishermen Scand J Work Environ Health 21 96 105 7618064
|
15687046
|
PMC1277861
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):175-179
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7252
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.6935ehp0113-00018015687055ResearchArticlesAn Approach to Evaluation of the Effect of Bioremediation on Biological Activity of Environmental Contaminants: Dechlorination of Polychlorinated Biphenyls Ganey Patricia E. 1Boyd Steven A. 2Departments of 1Pharmacology and Toxicology, and2Crop and Soil Science, Institute of Environmental Toxicology, Michigan State University, East Lansing, Michigan, USAAddress correspondence to P.E. Ganey, Department of Pharmacology and Toxicology, 214 Food Safety and Toxicology Building, Michigan State University, East Lansing, MI 48824 USA. Telephone: (517) 432-1761. Fax: (517) 432-2310. E-mail:
[email protected] article is based on a presentation at the conference “Bioremediation and Biodegradation: Current Advances in Reducing Toxicity, Exposure and Environmental Consequences” (http://www-apps.niehs.nih.gov/sbrp/bioremediation.html) held 9–12 June 2002 in Pacific Grove, California, and sponsored by the National Institute of Environmental Health Sciences (NIEHS) Superfund Basic Research Program. The overall focus of this conference was on exploring the research interfaces of toxicity reduction, exposure assessment, and evaluation of environmental consequences in the context of using state-of-the art approaches to bioremediation and biodegradation. The Superfund Basic Research Program has a legacy of supporting research conferences designed to integrate the broad spectrum of disciplines related to hazardous substances.
This work was supported by grant ES04911 from the National Institutes of Health.
The authors declare they have no competing financial interests.
2 2005 9 12 2004 113 2 180 185 23 12 2003 19 5 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. The effectiveness of bioremediation efforts is assessed traditionally from the loss of the chemical of interest. In some cases, analytical techniques are coupled with evaluation of toxicity to organisms representative of those found in the affected environment or surrogate organisms. Little is known, however, about the effect of remediation of environmental chemicals on potential toxicity to mammalian organisms. We discuss both an approach that employs mammalian cell system bioassays and the criteria for selection of the assays. This approach has been used to evaluate the biological response to mixtures of polychlorinated biphenyls (PCBs) before and after remediation by reductive dechlorination. The dechlorination process used results in accumulation of congeners substituted in only the ortho and para positions and containing fewer chlorines than the starting mixtures. Evaluation of the dechlorinated mixture reveals a loss of biological activity that could be ascribed to coplanar PCBs not containing chlorine in the ortho positions. Conversely, biological activity associated with ortho-substituted PCB congeners is unaffected or increased by remediation. Thus, the results of the bioassays are consistent with the remediation-induced change in the profile of PCB congeners and the known mechanisms of action of PCBs. The results emphasize a need for evaluation of the products of remediation for biological activity in mammalian systems. Furthermore, the approach outlined demonstrates the potential to assess the impact of remediation on a range of biological activities in mammalian cells and thus to estimate positive and negative effects of remediation strategies on toxicity. Future needs in this area of research include assays to evaluate biological effects under conditions of exposure that mimic those found in the environment and models to extrapolate effects to assess risk to people and wildlife.
bioassaycytochrome P450dechlorinationinsulinin vitro fertilizationneutrophilPCBtranscriptionuterine contraction
==== Body
Biological remediation technologies offer the advantage of partial or complete destruction of contaminants within a site. The ultimate goal of remediation is conversion of toxic organic contaminants to simple, less-toxic constituents, although for some chemicals, incomplete conversion occurs and stable intermediates are formed. The effectiveness of remediation strategies is traditionally evaluated from the disappearance of the chemical of interest. This approach does not consider that end products or intermediates produced during remediation may be toxic. Furthermore, the potential exists that remediation may result in products for which the toxic response is greater than for the parent compound or for which the target of toxicity is different, and these possibilities would not be detected. Accordingly, from the standpoint of assessing risk, it is important to understand the biological activity or toxicity of the end products and stable intermediates. Thus, the question becomes, Are the products or intermediates of bioremediation less toxic than the starting materials?
The anticipated answer to this question is yes; however, there is a dearth of evidence to support this assumption, particularly with respect to effects on mammalian systems. There are some reports of decreased toxic effects after remediation of contaminants, using mammalian systems to evaluate toxicity (Mousa et al. 1996, 1998; Quensen et al. 1998). On the other hand, some evidence suggests that products formed during remediation or breakdown of environmental chemicals have greater biological activity than the starting materials. For example, DDE [1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene], a major environmental transformation product of DDT [1,1,1-trichloro-2,2-bis(p-chlorophenyl) ethane], is a more potent androgen receptor antagonist than its parent compound (Kelce et al. 1995). In addition, products of microbial reductive dechlorination of polychlorinated biphenyls (PCBs) are more effective than parent PCB mixtures at stimulating uterine contractions in vitro (Bae et al. 2001). Similarly, chemical remediation may result in products with increased biological activity. For example, pyrene, a four-ringed polycyclic aromatic hydrocarbon, can be degraded with ozone. This ozonation results in the formation of at least 10 major products, some of which are more mutagenic than pyrene itself (Sasaki et al. 1995). The initial products formed from ozonation of a variety of polycyclic aromatic hydrocarbons in aqueous solution cause greater inhibition of the ability of mammalian cells to communicate through gap junctions compared with the parent compounds (Upham et al. 1997; Weis et al. 1998). These reports emphasize the need for investigators to consider the biological activity not only of the parent contaminants, but also of their stable transformation products produced during remediation.
Bioassays Commonly Used to Assess Effectiveness of Remediation
Investigators have not ignored the question of whether loss of biological activity accompanies remediation. The approaches used include bioassays using organisms representative of those we expect to find in the affected environment or surrogate organisms or plants. For example, the survival, growth, and reproduction of a variety of marine organisms exposed to sediments or soil collected from contaminated sites before and after remediation have been used to assess effectiveness of some remediation strategies [Deanovic et al. 1999; Kemble et al. 2000; McGann et al. 2003; Tabak et al. 2003; U.S. Environmental Protection Agency (EPA) 1989]. Toxicity to earthworms has been used to evaluate the effects of methods of removal of contaminants from soil (Chang et al. 1997; Maenpaa et al. 2002; Saterbak et al. 1999; U.S. EPA 1988). Luminescent bacterial assays such as the commercially available Microtox assay have also been used widely (Ahtiainen et al. 2002; Dorn and Salanitro 2000; Frische and Hoper 2003; Kemble et al. 2000; Layton et al. 1999). This technique is based on the observation that some bacteria (e.g., Vibrio fischeri) luminesce in proportion to cellular metabolism; accordingly, toxicity to the microorganisms is detected as a decrease in the intensity of luminescence. A solid-phase application of this method offers an advantage in that it permits exposure of bacteria to sediment-bound contaminants (Kemble et al. 2000). An integrated approach to ecotoxicologic evaluation involves combinations of these methods (Frische 2003).
These approaches yield valuable information regarding effectiveness of remediation and help focus additional remediation strategies. As with all bioassays, each has advantages and disadvantages, some of which relate to sensitivity, cost, versatility of application, reliability, rapidity, reproducibility, and relationship to health risk. A comprehensive discussion of these is not within the scope of this work. However, none of these bioassays addresses the potential biological activity of products of remediation in mammalian systems that may represent more specific and/or integrated functions relevant to human health. In the remainder of this article, we review an approach to the evaluation of toxicity of products of remediation in mammalian systems.
Bioassays Employing Mammalian Cell Systems
The concept that products of remediation may have biological activity in mammalian systems has not been studied extensively. Investigators associated with the Michigan State University Superfund Program Project began an effort a number of years ago as part of a Bioremediation Product Evaluation Core to address the issue. The working hypothesis was that products of remediation have different biological activities compared with those of the starting compounds or mixtures. We developed a list of assays of biological activity that relied on the strengths and expertise of the toxicologists within the group (Table 1). Generally, criteria for useful bioassays include sensitivity over a range of concentrations of test chemical, low rate of false-positive and false-negative responses, ease and rapidity of the assay, reproducibility of results, and reasonable cost. How well the end point being measured reflects a biological response of interest in humans or animals may also be important. For purposes of using results from an assay for risk assessment, it is helpful to have a reference value for toxicity, namely, a response known to be associated with toxicity in whole organisms. Assays selected for use in the Bioremediation Product Evaluation Core met many of these criteria. Additional criteria for inclusion in the Core were that assays were performed routinely within a laboratory and that the expected results were relatively uncomplicated in interpretation. These latter two criteria precluded the use of whole-animal studies, so the assays selected involved in vitro methodology. With this approach, the list developed covers a variety of cellular functions including intracellular signaling, intercellular communication, proliferation and cell death, gene expression, measures of integrated cellular function and integrated tissue function, and aryl hydrocarbon (Ah) receptor function (important for dioxin-like contaminants) (Table 1). Accordingly, although the list is not exhaustive, many possible responses to chemical insult are represented. Additional measures not represented on this list that would be useful include whole-animal assessments and assays that measure endocrine disruption, neurotoxicity, genotoxicity, or mutagenicity.
In evaluating remediation products, we selected specific bioassays for initial examination on the basis of current knowledge of the mechanism of action of the parent compound of interest. For example, for dioxin-like chemicals (e.g., PCBs) one of the first avenues of investigation was the effects on cytochrome P450 induction based on the known activity of these compounds to increase cytochrome P4501A. Similarly, for chemicals known to disrupt intracellular signaling, such as some of the polycyclic aromatic hydrocarbons (Burdick et al. 2003; Patten Hitt et al. 2002), first priority for analysis was given to examination of activation of mitogen-activated protein kinases or alterations in neutrophil function. Initial studies using this approach were aimed at evaluation of products of bioremediation of PCBs. One promising remediation technique for PCBs is the removal of chlorines by microorganisms. We review results of these studies below.
Evaluation of Products of Reductive Dechlorination of PCBs
Polychlorinated biphenyls are among the most widely distributed environmental contaminants. Commercial PCB mixtures were manufactured in the United States between 1929 and 1978 and used for a variety of industrial purposes. An estimated 1.4 billion pounds of PCBs have been produced worldwide and approximately several hundred million pounds have been released into the environment. Commercial PCBs (e.g., Aroclors) typically consist of 60–90 of the 209 possible congeners, each of which differs in the positions and/or numbers of chlorines on the biphenyl ring. Several characteristic PCB mixtures differ in the extent of chlorination and specific congener composition. Common examples are Aroclors 1242, 1248, and 1254, which contain 42, 48, and 54% chlorine by weight, respectively. Because of their lipophilic properties, PCBs tend to accumulate in biological tissue and in environments rich in organic matter, such as sediments.
PCB mixtures found in the environment often do not match any of the known commercial formulations because they have been subjected to congener-selective environmental processes, for example, reductive dechlorination by anaerobic bacteria (Bedard and Quensen 1995; Quensen et al. 1988, 1990). Reductive dechlorination is a microbially mediated process that removes chlorine from biphenyl with replacement by hydrogen, resulting in a product mixture in which the average number of chlorines is substantially diminished. Chlorines substituted in the meta and para positions are preferentially removed by this process; ortho chlorines are rarely removed. In situ reductive dechlorination has been documented in anaerobic sediments at numerous locations, and six distinct dechlorination patterns have been observed, giving rise to six recognizable profiles of congeners in the dechlorination products (Bedard and Quensen 1995).
As mentioned above, PCBs comprise 209 individual congeners, and a variety of toxic effects mediated by multiple mechanisms accompany this structural diversity. Effects include neurotoxicity, induction of enzymes involved in xenobiotic metabolism, alterations in reproductive function, hepatotoxicity, carcinogenicity, and effects on cells that mediate innate and specific immunity (Safe 1994). In applying Occam’s Razor, one can think of PCBs as falling into two groups in terms of structure and mechanisms of action (Figure 1). Coplanar PCBs lack ortho substitution, bind with high affinity to the Ah receptor, and mediate many of their effects through changes in gene expression initiated by binding to this receptor. Noncoplanar PCBs, which contain chlorine in one or more of the four ortho positions, are poor ligands for the Ah receptor. The mechanisms of their biological effects are in many cases unknown but often involve initial changes in cell signaling (Fischer et al. 1998).
Studies were undertaken to compare the biological activity of Aroclor mixtures of PCBs with the activity of products of their reductive dechlorination. The dechlorination process employed resulted in accumulation of congeners substituted in only the ortho and para positions and containing fewer chlorines than the starting mixtures (Mousa et al. 1996; Quensen et al. 1998). For example, 2,2′,4-trichlorobiphenyl represented 4% (on a molar basis) of the total mixture before dechlorination and 16% of the dechlorinated product. For more detailed description of the congener profile of the remediation products, the reader is referred to Bae et al. (2001), Ganey et al. (2000), and Mousa et al. (1998).
Table 2 is a summary of the results of examination of biological activity. Coplanar, dioxin-like PCBs induce cytochrome P4501A through an Ah receptor–mediated mechanism (Sanderson et al. 1996), and the potency for this effect can be compared with the potency of dioxin (2,3,7,8-tetrachlorodibenzo-p-dioxin) to generate a toxic equivalency factor (TEF) for individual congeners (Safe 1993). TEF values can then be used to determine the toxic equivalents (TEQs) for mixtures of chemicals. This approach has been used for risk assessment of dioxin-like compounds, although it is not without limitation (Li and Hansen 1996; Safe 1998). The ability of products of dechlorination of Aroclor mixtures to induce cytochrome P4501A activity, monitored as ethoxyresorufin-O-deethylase activity, was examined in the rat liver hepatoma cell line H4IIE. Parent Aroclors 1242 and 1254 were compared with products of their dechlorination by microorganisms collected from two different sites, Silver Lake, Massachusetts, and River Raisin, Michigan. Aroclors were evaluated at concentrations ranging from 0.04 to 2.5 μg/well (250 μL/well), and the dechlorination products were used at molar equivalent concentrations based on biphenyl concentration (biphenyl concentration is unaffected by dechlorination). Both potency and efficacy of induction of the Aroclor mixtures were diminished by dechlorination (Mousa et al. 1998; Quensen et al. 1998). The decrease in potency was dependent on the extent of removal of the coplanar and mono-ortho-substituted PCBs, consistent with the known mechanism of this effect. For example, the TEQ for nondechlorinated Aroclor 1242 derived from the assay was 3.1, whereas the TEQ for the dechlorinated mixture was below the limit of detection (0.06). These values were in agreement with TEQs calculated from the known composition of the nondechlorinated and dechlorinated mixtures, 5.7 and < 0.08, respectively.
In vitro fertilization is reflective of reproductive capacity. Epidemiologic studies assessing the effects of human exposure to PCBs on fertility and reproduction have yielded various results: some indicate a negative effect of PCBs on fertility, whereas others report no association (Axmon et al. 2001, 2002; Dallinga et al. 2002; Rozati et al. 2002; Yu et al. 2000). In experimental animals dioxin-like chemicals, including some PCBs, cause reproductive toxicity (Birnbaum and Tuomisto 2000; Peterson et al. 1993; Petroff et al. 2001). For example, administration of heavily chlorinated, noncoplanar PCB congeners to male rats decreases several markers of sperm function (Hsu et al. 2003). Exposure of female mice to the coplanar congener 3,3′,4,4′-tetrachlorobiphenyl decreases reproductive capacity (Huang et al. 1998a), and exposure of pregnant mice to Aroclor 1242 or to 3,3′,4,4′-tetrachlorobiphenyl alters fertility in male offspring (Fielden et al. 2001; Huang et al. 1998b). In addition, coplanar PCBs inhibit in vitro fertilization of murine eggs (Huang et al. 1998a). Products of dechlorination of Aroclors 1242 and 1254 were compared with the parent Aroclors for the ability to inhibit in vitro fertilization of mouse gametes (Mousa et al. 1996, 1998). Aroclor 1254 decreased the percentage of fertilized eggs and increased the percentage of degenerated eggs at 10 ppm and 20 ppm. The products of reductive dechlorination used at equivalent molar concentrations produced less of an adverse effect on fertilization and did not cause gamete degeneration. Similarly, the negative effects of Aroclor 1242 on fertilization were not observed with its product of dechlorination. Based on the observations that coplanar PCBs and heavily chlorinated, noncoplanar PCBs alter reproductive capacity, this result was consistent with the loss of these congeners due to dechlorination.
Environmental exposure to PCBs has been associated with increased risk of cancer in some but not all studies (Demers et al. 2002; Gammon et al. 2002; Kimbrough et al. 2003; Laden et al. 2002; Lucena et al. 2001; Stellman et al. 2000; Woolcott et al. 2001). The transcription factor activator protein-1 (AP-1) is a protein that regulates gene expression and has been implicated in tumorigenesis. Using the rat liver epithelial cell line WB-344, transfected with AP-1–binding DNA and a luciferase reporter gene, the ability of remediation products of Aroclors to induce AP-1 activity was determined. Native Aroclors (2 μg/mL) caused a 2- to 3-fold increase in induction of AP-1 transcription, whereas dechlorinated products (equivalent molar concentration) had no effect on AP-1–mediated transcription (Mousa et al. 1998). Stimulation of AP-1–mediated transcription is attributed to more heavily chlorinated, noncoplanar PCBs; thus, these results are consistent with the loss of heavily chlorinated congeners upon dechlorination.
Exposure to PCBs has been associated with decreased gestation length in several epidemiologic studies (Bercovici et al. 1983; Taylor et al. 1989; Wassermann et al. 1982). Because uterine contractions actuate parturition, the effects of PCBs on contractility of pregnant rat uteri were examined. Aroclor 1242 stimulated contraction of uteri isolated from pregnant rats in a concentration- and time-dependent manner (Bae et al. 1999, 2001). A concentration of 100 μM nondechlorinated Aroclor 1242 increased contraction frequency, whereas smaller concentrations were without effect (Bae et al. 2001). The potency of various Aroclor mixtures to increase uterine contraction frequency was inversely related to chlorine content, suggesting that this effect was mediated by less heavily chlorinated congeners. Results with native Aroclors were compared with the effects of Aroclors that had been dechlorinated by microorganisms collected from the Hudson River basin. Compared with the response to unaltered Aroclor 1242, the dechlorinated mixture shifted the concentration–response curve to the left, such that 10 μM of the dechlorinated mixture caused an increase in uterine contraction frequency. Similarly, the cumulative concentration–response curve of the dechlorinated Aroclor 1254 was shifted to the left relative to that of the unaltered Aroclor 1254. In fact, parent Aroclor 1254 did not stimulate contractions with exposure up to 300 μM, yet the dechlorinated mixture exerted a powerful stimulatory response in terms of both effective concentration range (30 μM increased contraction frequency) and efficacy. Thus, dechlorination produced a mixture with uterine-stimulating activity from a relatively nonactive Aroclor mixture.
PCB exposure has been associated with alterations in immune status in humans (Belles-Isles et al. 2002; Van Den Heuvel et al. 2002) and experimental animals (Arena et al. 2003; De Krey and Kerkvliet 1995; De Krey et al. 1994). In addition, cells of both specific (e.g., lymphocytes) and innate (e.g., neutrophils) immunity are affected by PCBs (Fernlof et al. 1997; Ganey et al. 1993; Suh et al. 2003). For example, noncoplanar PCBs stimulate neutrophils to produce reactive oxygen species, specifically superoxide anion (Ganey et al. 1993). In addition, PCBs increase superoxide anion production in response to subsequent stimulation with phorbol myristate acetate (PMA). The ability of Aroclor 1242 to cause generation of reactive oxygen species in neutrophils was compared with the ability of its products of dechlorination by microorganisms from Silver Lake or River Raisin (Ganey et al. 2000). Exposure of rat neutrophils in vitro to Aroclor 1242 at 10 μg/mL increased PMA-stimulated superoxide anion generation. Exposure of neutrophils to products of dechlorination of Aroclor 1242 at equivalent molar concentrations caused similar increases in superoxide anion production (Ganey et al. 2000). Accordingly, dechlorination did not diminish the ability of the mixtures to activate neutrophils. On the other hand, parent Aroclor 1254 did not increase superoxide anion production in PMA-stimulated neutrophils, but its dechlorination products did. Thus, like the effects observed for stimulation of uterine contractility, dechlorination induced biological activity in a nonactive Aroclor mixture. These results are consistent with the accumulation of noncoplanar PCBs in the dechlorination products.
Increased incidence of diabetes has been associated with high concentrations of PCBs or other organochlorine chemicals in serum (Glynn et al. 2003; Longnecker et al. 2001). In addition, Aroclor mixtures of PCBs stimulate the release of insulin from the rat clonal cell line RINm5F (Fischer et al. 1996). This effect is mediated by noncoplanar PCBs (Fischer et al. 1998). RINm5F cells were exposed to Aroclor 1242 or 1254 (10 μg/mL) or their products of dechlorination by River Raisin or Silver Lake microorganisms (equivalent molar concentrations), and insulin release was examined. Both parent Aroclor mixtures caused release of insulin within 30 min of exposure. The magnitude of response to the mixtures of dechlorinated Aroclors was similar or greater when compared with the non-dechlorinated parent mixtures (Ganey et al. 2000). These results are consistent with the observed accumulation of ortho-substituted, noncoplanar PCBs in the mixtures produced by reductive dechlorination.
Taken together, these results demonstrate that a variety of responses can be observed after exposure of mammalian cell systems to products of remediation. In the case of the studies described above for remediation of PCBs, the responses followed what would be expected based on structure and known biological activity of the chemicals. That is, Ah receptor–mediated activities diminished because of the removal of coplanar congeners via meta and para dechlorination processes, and biological activities mediated by non-coplanar PCBs were enhanced or unchanged. These studies were guided by knowledge of some of the mechanisms of action of PCBs. For remediation processes aimed at chemicals for which less is known about effects in mammalian systems, studies similar to those described above may reveal unexpected results.
Summary and Future Needs
Several important aspects of evaluation were not addressed in this series of experiments. For this specific case of remediation of PCBs, no measure of neurotoxicity was performed. This is an important deficit because the neurotoxic effects of PCBs have been demonstrated experimentally and suggested by results of epidemiologic studies (Schantz et al. 2001, 2003; Seegal 1996). Because many neurotoxic effects are associated with non-coplanar PCBs (Kodavanti and Tilson 1997; Wong et al. 2001), one would expect effects of the products of remediation to be similar or greater than those of the parent Aroclors.
All the assays used were in vitro assays that represent selected functions that occur within a whole organism. This approach does not address issues of exposure, including relevant routes of exposure to environmental contaminants and their remediation products. In addition, the duration of exposure during in vitro assays is short and does not mimic longer-term, often-repeated exposures that occur naturally. Issues of bioavailability are not considered when performing in vitro assays. This includes bioavailability from an environmental engineering point of view (e.g., how much of the contaminant is not bound to soil constituents) and from the perspective of toxicology (e.g., how much of the exposure dose interacts with target tissue). These issues can best be addressed using whole-organism studies, which, as mentioned above, are costly and inconvenient. In addition, biologically based toxicokinetic and toxicodynamic modeling could be used to address issues of extrapolation to human risk. In the future, approaches to include these considerations must be developed.
Thus, it should be emphasized that the approach described above to evaluate effects of products of remediation in a variety of in vitro assays employing mammalian cells represents a beginning. Using this approach, the biological activity of remediation products is compared with activity of the parent compound, such that relative activity is assessed. Although this is a useful component in determination of the effectiveness of remediation, it stops short of estimating potential health risk of the remediation products. Comprehensive evaluation of the biological activity of remediation products will necessitate far more extensive in vitro and in vivo testing, the use of validated extrapolation models to assess risk to people and wildlife, and epidemiologic correlates. It seems unlikely that this type of effort will arise from any single institution. It is more likely to be achieved through a consortium of institutions or a government-based testing facility that can amass the expertise and resources required.
Despite these limitations, several points can be drawn from these remediation assessment evaluations. First, the overarching message is that it is important to evaluate the biological activity of products of remediation and also of stable intermediates produced during remediation. As seen in the series of experiments presented above, the products of remediation are not necessarily devoid of biological activity. When compared with the parent compound, activity of remediation products may be decreased, unchanged, or increased. It is also possible that biological activity of remediation products may be qualitatively different from the activity of the starting compound. Furthermore, although not observed in the studies described above, when bacteria are used in remediation processes, it is possible that bacterial by-products unrelated to the chemical contaminant itself are produced that have biological activity in some cellular systems. Another important point to be made is that a better understanding of the mechanisms of biological effects of contaminants will permit a more directed approach to evaluation of the activity of the remediation products. The selection of bioassays to be used as well as the specific details of experimental design can be based on known mechanisms of action of the parent compounds. Finally, knowledge of the spectrum of biological activities associated with remediated chemicals and their stable intermediates will provide the basis for more accurate risk assessment and guide remediation needs and approaches.
Figure 1 Structure of coplanar and noncoplanar PCBs. 3,3′,4,4′,5-Pentachlorobiphenyl is the representative coplanar PCB depicted. 2,2′,4,4′-Tetrachlorobiphenyl is the representative noncoplanar PCB depicted.
Table 1 Examples of assays used to assess the biological activity of remediation products.
Assay Biological functions represented
Induction of cytochrome P450 enzymes Receptor-mediated activity (Ah receptor)
Activation of mitogen-activated protein kinases Intracellular signaling
Disruption of gap junctional intercellular communication Intercellular signaling, cell death
Activation of AP-1 transcription factor Gene expression
Alteration in neutrophil function Cellular function, cell death
Stimulation of insulin release Cellular function
Contraction of uterine muscle in vitro Integrated tissue function
Alteration in fertilization in vitro Integrated tissue/organ system function
Stimulation of lymphocyte proliferation Proliferation, cell death
Table 2 Summary of effects of biological activity of dechlorinated PCBs.
Biological activity Effect of parent Aroclor Type of PCBs mediating effect Effect of dechlorinated products Reference
Cytochrome P450 activity Induction Coplanar None Mousa et al. 1998; Quensen et al. 1998
In vitro fertilization Reduction Coplanar None Mousa et al. 1996, 1998
AP-1-mediated transcription Induction More heavily chlorinated, noncoplanar None Mousa et al. 1998
Uterine contraction Stimulation Less heavily chlorinated, noncoplanar Greater stimulation Bae et al. 2001
Neutrophil function Activation Noncoplanar Same or greater activation Ganey et al. 2000
Insulin secretion Stimulation Noncoplanar Stimulation Ganey et al. 2000
==== Refs
References
Ahtiainen J Valo R Jarvinen M Joutti A 2002 Microbial toxicity tests and chemical analysis as monitoring parameters at composting of creosote-contaminated soil Ecotoxicol Environ Saf 53 323 329 12568471
Arena SM Greeley EH Halbrook RS Hansen LG Segre M 2003 Biological effects of gestational and lactational PCB exposure in neonatal and juvenile C57BL/6 mice Arch Environ Contam Toxicol 44 272 280 12520400
Axmon A Rylander L Stromberg U Dyremark E Hagmar L 2001 Polychlorinated biphenyls in blood plasma among Swedish female fish consumers in relation to time to pregnancy J Toxicol Environ Health A 64 485 498 11732699
Axmon A Rylander L Stromberg U Hagmar L 2002 Female fertility in relation to the consumption of fish contaminated with persistent organochlorine compounds Scand J Work Environ Health 28 124 1232 12019589
Bae J Mousa MA Quensen JR Boyd SA Loch-Caruso R 2001 Stimulation of contraction of pregnant rat uterus in vitro by non-dechlorinated and microbially dechlorinated mixtures of polychlorinated biphenyls Environ Health Perspect 109 275 282 11333189
Bae J Peters-Golden M Loch-Caruso R 1999 Stimulation of pregnant rat uterine contraction by the polychlorinated biphenyl (PCB) mixture Aroclor 1242 may be mediated by arachidonic acid release through activation of phospholipase A2 enzymes J Pharmacol Exp Ther 289 1112 1120 10215694
Bedard DL Quensen JF III 1995. Microbial reductive dechlorination of polychlorinated biphenyls. In: Ecological and Applied Microbiology: Microbial Transformation and Degradation of Toxic Organic Chemicals (Young LY, Cerniglia CE, eds). New York:Wiley-Liss, 127–216.
Belles-Isles M Ayotte P Dewailly E Weber JP Roy R 2002 Cord blood lymphocyte functions in newborns from a remote maritime population exposed to organochlorines and methylmercury J Toxicol Environ Health A 65 165 182 11820504
Bercovici B Wassermann M Cucos S Ron M Wassermann D Pines A 1983 Serum levels of polychlorinated biphenyls and some organochlorine insecticides in women with recent and former missed abortions Environ Res 30 169 174 6403348
Birnbaum LS Tuomisto J 2000 Non-carcinogenic effects of TCDD in animals Food Addit Contam 17 275 288 10912242
Burdick AD Davis JW Liu KF Hudson LG Shi H Monske ML 2003 Benzo(a )pyrene quinones increase cell proliferation, generate reactive oxygen species, and transactivate the epidermal growth factor receptor in breast epithelial cells Cancer Res 63 7825 7833 14633709
Chang LW Meier JR Smith MK 1997 Application of plant and earthworm bioassays to evaluate remediation of a lead-contaminated soil Arch Environ Contam Toxicol 32 166 171 9069192
Dallinga JW Moonen EJ Dumoulin JC Evers JL Geraedts JP Kleinjans JC 2002 Decreased human semen quality and organochlorine compounds in blood Hum Reprod 17 1973 1979 12151423
Deanovic L Connor VM Knight AW Maier KJ 1999 The use of bioassays and toxicity identification evaluation (TIE) procedures to assess recovery and effectiveness of remedial activities in a mine drainage-impacted stream system Arch Environ Contam Toxicol 36 21 27 9828258
De Krey GK Baecher-Steppan L Fowles JR Kerkvliet NI 1994 Polychlorinated biphenyl-induced suppression of cytotoxic T lymphocyte activity: role of prostaglandin-E2 Toxicol Lett 74 211 220 7871545
De Krey GK Kerkvliet NI 1995 Suppression of cytotoxic T lymphocyte activity by 2,3,7,8-tetrachlorodibenzo-p -dioxin occurs in vivo , but not in vitro , and is independent of corticosterone elevation Toxicology 97 105 112 7716775
Demers A Ayotte P Brisson J Dodin S Robert J Dewailly E 2002 Plasma concentrations of polychlorinated biphenyls and the risk of breast cancer: a congener-specific analysis Am J Epidemiol 155 629 635 11914190
Dorn PB Salanitro JP 2000 Temporal ecological assessment of oil contaminated soils before and after bioremediation Chemosphere 40 419 426 10665408
Fernlof G Gadhasson I Podra K Darnerud PO Thuvander A 1997 Lack of effects of some individual polybrominated diphenyl ether (PBDE) and polychlorinated biphenyl (PCB) congeners on human lymphocyte functions in vitro Toxicol Lett 90 189 197 9067487
Fielden MR Halgren RG Tashiro CH Yeo BR Chittim B Chou K 2001 Effects of gestational and lactational exposure to Aroclor 1242 on sperm quality and in vitro fertility in early adult and middle-aged mice Reprod Toxicol 15 281 292 11390173
Fischer LJ Seegal RF Ganey PE Pessah IN Kodavanti PR 1998 Symposium overview: toxicity of non-coplanar PCBs Toxicol Sci 41 49 61 9520341
Fischer LJ Zhou H Wagner MA 1996 Polychlorinated biphenyls release insulin from RINm5F cells Life Sci 59 2041 2049 8950306
Frische T 2003 Ecotoxicological evaluation of in situ bioremediation of soils contaminated by the explosive 2,4,6-trinitrotoluene (TNT) Environ Pollut 121 103 113 12475067
Frische T Hoper H 2003 Soil microbial parameters and luminescent bacteria assays as indicators for in situ bioremediation of TNT-contaminated soils Chemosphere 50 415 427 12656263
Gammon MD Wolff MS Neugut AI Eng SM Teitelbaum SL Britton JA 2002 Environmental toxins and breast cancer on Long Island. II. Organochlorine compound levels in blood Cancer Epidemiol Biomarkers Prev 11 686 697 12163320
Ganey PE Quensen JF Mousa MA Boyd SA Wagner MA Fischer LJ 2000 Biological activity associated with non-coplanar polychlorinated biphenyls after microbial dechlorination of Aroclor 1242 and Aroclor 1254 Environ Toxicol Chem 19 1311 1316
Ganey PE Sirois JE Denison M Robinson JP Roth RA 1993 Neutrophil function after exposure to polychlorinated biphenyls in vitro Environ Health Perspect 101 430 434 8119254
Glynn AW Granath F Aune M Atuma S Darnerud PO Bjerselius R 2003 Organochlorines in Swedish women: determinants of serum concentrations Environ Health Perspect 111 349 355 12611665
Hsu PC Li MH Guo YL 2003 Postnatal exposure of 2,2′,3,3′,4,6′-hexachlorobiphenyl and 2,2′,3,4′,5′,6-hexachlorobiphenyl on sperm function and hormone levels in adult rats Toxicology 187 117 126 12699901
Huang A Lin S Inglis R Powell D Chou K 1998a Pre- and postnatal exposure to 3,3′,4,4′-tetrachlorobiphenyl. II. Effects on the reproductive capacity and fertilizing ability of eggs in female mice Arch Environ Contam Toxicol 34 209 214 9469864
Huang A Powell D Chou K 1998b Pre- and postnatal exposure to 3,3′,4,4′-tetrachlorobiphenyl. I. Effects on breeding ability and sperm fertilizing ability in male mice Arch Environ Contam Toxicol 34 204 208 9469863
Kelce WR Stone CR Laws SC Gray LE Kemppainen JA Wilson EM 1995 Persistent DDT metabolite p,p ′-DDE is a potent androgen receptor antagonist Nature 375 581 585 7791873
Kemble NE Hardesty DG Ingersoll CG Johnson BT Dwyer FJ MacDonald DD 2000 An evaluation of the toxicity of contaminated sediments from Waukegan Harbor, Illinois, following remediation Arch Environ Contam Toxicol 39 452 461 11031305
Kimbrough RD Doemland ML Mandel JS 2003 A mortality update of male and female capacitor workers exposed to polychlorinated biphenyls J Occup Environ Med 45 271 282 12661184
Kodavanti PR Tilson HA 1997 Structure-activity relationships of potentially neurotoxic PCB congeners in the rat Neurotoxicology 18 425 441 9291492
Laden F Ishibe N Hankinson SE Wolff MS Gertig DM Hunter DJ 2002 Polychlorinated biphenyls, cytochrome P450 1A1, and breast cancer risk in the Nurses’ Health Study Cancer Epidemiol Biomarkers Prev 11 1560 1565 12496044
Layton AC Gregory B Schultz TW Sayler GS 1999 Validation of genetically engineered bioluminescent surfactant resistant bacteria as toxicity assessment tools Ecotoxicol Environ Saf 43 222 228 10375425
Li MH Hansen LG 1996 Enzyme induction and acute endocrine effects in prepubertal female rats receiving environmental PCB/PCDF/PCDD mixtures Environ Health Perspect 104 712 722 8841756
Longnecker MP Klebanoff MA Brock JW Zhou H 2001 Polychlorinated biphenyl serum levels in pregnant subjects with diabetes Diabetes Care 24 1099 1101 11375377
Lucena F Allam MF Costabeber IH Villarejo ML Navajas RF 2001 Breast cancer risk factors: PCB congeners Eur J Cancer Prev 10 117 119 11263587
Maenpaa KA Kukkonen JV Lydy MJ 2002 Remediation of heavy metal-contaminated soils using phosphorus: evaluation of bioavailability using an earthworm bioassay Arch Environ Contam Toxicol 43 389 398 12399909
McGann M Alexander CR Bay SM 2003 Response of benthic foraminifers to sewage discharge and remediation in Santa Monica Bay, California Mar Environ Res 56 299 342 12648961
Mousa MA Ganey PE Quensen JF Madhukar BV Chou K Giesy JP 1998 Altered biological activities of commercial polychlorinated biphenyl mixtures after microbial reductive dechlorination Environ Health Perspect 106 suppl 6 1409 1418 9860899
Mousa MA Quensen JF Chou K Boyd SA 1996 Microbial dechlorination alleviates inhibitory effects of PCBs on mouse gamete fertilization in vitro Environ Sci Technol 30 2087 2092
Patten Hitt E DeLong MJ Merrill AH Jr 2002 Benzo(a )pyrene activates extracellular signal-related and p38 mitogen-activated protein kinases in HT29 colon and adenocarcinoma cells: involvement in NAD(P)H:quinone reductase activity and cell proliferation Toxicol Appl Pharmacol 183 160 167 12383707
Peterson RE Theobald HM Kimmel GL 1993 Developmental and reproductive toxicity of dioxins and related compounds: cross-species comparisons Crit Rev Toxicol 23 283 335 8260069
Petroff BK Roby KF Gao X Son D Williams S Johnson D 2001 A review of mechanisms controlling ovulation with implications for the anovulatory effects of polychlorinated dibenzo-p -dioxins in rodents Toxicology 158 91 107 11275352
Quensen JF III Boyd SA Tiedje JM 1990 Dechlorination of four commercial polychlorinated biphenyl mixtures (Aroclors) by anaerobic microorganisms from sediments Appl Environ Microbiol 56 2360 2369 16348249
Quensen JF III Mousa MA Boyd SA Sanderson TJ Froese KL Giesy JP 1998 Reduction of aryl hydrocarbon receptor-mediated activity of polychlorinated biphenyl mixtures due to anaerobic microbial dechlorination Environ Toxicol Chem 17 806 813
Quensen JF III Tiedje JM Boyd SA 1988 Reductive dechlorination of polychlorinated biphenyls by anaerobic microorganisms from sediments Science 242 752 754 17751997
Rozati R Reddy PP Reddanna P Mujtaba R 2002 Role of environmental estrogens in the deterioration of male factor fertility Fertil Steril 78 1187 1194 12477510
Safe S 1993 Development of bioassays and approaches for the risk assessment of 2,3,7,8-tetrachlorodibenzo-p -dioxin and related compounds Environ Health Perspect 101 suppl 3 317 325 8143638
Safe SH 1994 Polychlorinated biphenyls (PCBs): environmental impact, biochemical and toxic responses, and implications for risk assessment Crit Rev Toxicol 24 87 149 8037844
Safe SH 1998 Development validation and problems with the toxic equivalency factor approach for risk assessment of dioxins and related compounds J Anim Sci 76 134 141 9464894
Sanderson JT Aarts JM Brouwer A Froese KL Denison MS Giesy JP 1996 Comparison of Ah receptor-mediated luciferase and ethoxyresorufin-O -deethylase induction in H4IIE cells: implications for their use as bioanalytical tools for the detection of polyhalogenated aromatic hydrocarbons Toxicol Appl Pharmacol 137 316 325 8661358
Sasaki J Arey J Harger W 1995 Formation of mutagens from the photooxidations of 2-4 ring-PAH Environ Sci Technol 29 1324 1335 22192029
Saterbak A Toy RJ Wong DCL McMain BJ Williams MP Dorn PB 1999 Ecotoxicological and analytical assessment of hydrocarbon-contaminated soils and application to ecological risk assessment Environ Toxicol Chem 18 1591 1607
Schantz SL Gasior DM Polverejan E McCaffrey RJ Sweeney AM Humphrey HE 2001 Impairments of memory and learning in older adults exposed to polychlorinated biphenyls via consumption of Great Lakes fish Environ Health Perspect 109 605 611 11445515
Schantz SL Widholm JJ Rice DC 2003 Effects of PCB exposure on neuropsychological function in children Environ Health Perspect 111 357 376 12611666
Seegal RF 1996 Epidemiological and laboratory evidence of PCB-induced neurotoxicity Crit Rev Toxicol 26 709 737 8958469
Stellman SD Djordjevic MV Britton JA Muscat JE Citron ML Kemeny M 2000 Breast cancer risk in relation to adipose concentrations of organochlorine pesticides and biphenyls in Long Island, New York Cancer Epidemiol Biomarkers Prev 9 1241 1249 11097233
Suh J Kang JS Yang KH Kaminski NE 2003 Antagonism of aryl hydrocarbon receptor-dependent induction of CYP1A1 and inhibition of IgM expression by di-ortho -substituted polychlorinated biphenyls Toxicol Appl Pharmacol 187 11 21 12628580
Tabak HH Lazorchak JM Lei L Khodadoust AP Antia JE Bagchi R 2003 Studies on bioremediation of polycyclic aromatic hydrocarbon-contaminated sediments: bioavailability, biodegradability, and toxicity issues Environ Toxicol Chem 22 473 482 12627632
Taylor PR Stelma JM Lawrence CE 1989 The relation of polychlorinated biphenyls to birth weight and gestational age in the offspring of occupationally exposed mothers Am J Epidemiol 129 395 406 2492144
Upham BL Boddy B Xing XS Trosko JE Masten SJ 1997 Non-genotoxic effects of selected pesticides and their disinfection by-products on gap junctional intercellular communication Ozone Sci Eng 19 351 369
U.S. EPA 1988. Protocols for Short-term Toxicity Screening of Hazardous Waste Sites. EPA 600-3-88-029. Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research Laboratory, Office of Research and Development.
U.S. EPA 1989. Short-term Methods for Estimating Chronic Toxicity of Effluents and Receiving Waters to Freshwater Organisms, 2nd ed. EPA/600/4-89/001. Cincinnati, OH:U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory.
Van Den Heuvel RL Koppen G Staessen JA Hond ED Verheyen G Nawrot TS 2002 Immunologic biomarkers in relation to exposure markers of PCBs and dioxins in Flemish adolescents (Belgium) Environ Health Perspect 110 595 600 12055051
Wassermann M Ron M Bercovici B Wassermann D Cucos S Pines A 1982 Premature delivery and organochlorine compounds: polychlorinated biphenyls and some organochlorine insecticides Environ Res 28 106 112 6179776
Weis LM Rummel AM Masten SJ Trosko JE Upham BL 1998 Bay or bay-like regions of polycyclic aromatic hydrocarbons were potent inhibitors of gap junction intercellular communication Environ Health Perspect 106 17 22 9417772
Wong PW Garcia EF Pessah IN 2001 ortho -Substituted PCB95 alters intracellular calcium signaling and causes cellular acidification in PC12 cells by an immunophilin-dependent mechanism J Neurochem 76 450 463 11208908
Woolcott CG Aronson KJ Hanna WM SenGupta SK McCready DR Sterns EE 2001 Organochlorines and breast cancer risk by receptor status, tumor size, and grade (Canada) Cancer Causes Control 12 395 404 11545454
Yu ML Guo YL Hsu CC Rogan WJ 2000 Menstruation and reproduction in women with polychlorinated biphenyl (PCB) poisoning: long-term follow-up interviews of the women from the Taiwan Yucheng cohort Int J Epidemiol 29 672 677 10922344
|
15687055
|
PMC1277862
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 9; 113(2):180-185
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.6935
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.6934ehp0113-00018615687056ResearchArticlesUse of Ecotoxicological Tools to Evaluate the Health of New Bedford Harbor Sediments: A Microbial Biomarker Approach Ford Timothy 1Jay Jenny 2Patel Anand 2Kile Molly 3Prommasith Phanida 3Galloway Tamara 4Sanger Ross 4Smith Karen 4Depledge Mike 41Department of Microbiology, Montana State University, Bozeman, Montana, USA2Department of Civil and Environmental Engineering, University of California Los Angeles, Los Angeles, California, USA3School of Public Health, Harvard University, Boston, Massachusetts, USA4Plymouth Environmental Research Center, University of Plymouth, Plymouth, United KingdomAddress correspondence to T. Ford, 109 Lewis Hall, Department of Microbiology, Montana State University, Bozeman, MT 59717 USA. Telephone: (406) 994-2901. Fax: (406) 994-4926. E-mail:
[email protected] article is based on a presentation at the conference “Bioremediation and Biodegradation: Current Advances in Reducing Toxicity, Exposure and Environmental Consequences” (http://www-apps.niehs.nih.gov/sbrp/bioremediation.html) held 9–12 June 2002 in Pacific Grove, California, and sponsored by the National Institute of Environmental Health Sciences (NIEHS) Superfund Basic Research Program. The overall focus of this conference was on exploring the research interfaces of toxicity reduction, exposure assessment, and evaluation of environmental consequences in the context of using state-of-the art approaches to bioremediation and biodegradation. The Superfund Basic Research Program has a legacy of supporting research conferences designed to integrate the broad spectrum of disciplines related to hazardous substances.
This publication was made possible by grant 5 P42 ES05947 from the NIEHS, National Institutes of Health (NIH). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS, NIH.
The authors declare they have no competing financial interests.
2 2005 8 12 2004 113 2 186 191 23 12 2003 26 5 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. We have been investigating microbial communities in sediments from New Bedford Harbor (NBH), Massachusetts, USA, for a number of years. NBH is a U.S. Environmental Protection Agency–designated Superfund site heavily contaminated with polychlorinated biphenyls, polycyclic aromatic hydrocarbons, and heavy metals. Microorganisms are thought to contribute to the fate and distribution of contaminants in NBH through a variety of mechanisms, including direct transformations and formation of soluble and insoluble species. Our more recent research has focused on changes in microbial community structure and function in response to exposure to toxic contaminants, with the ultimate goal of using microbes as ecotoxicological tools. Microbial diversity, as measured by restriction fragment-length polymorphism analysis, changes along pollution gradients, with an apparent increase in diversity at the most contaminated sites, concomitant with an increase in genetic relatedness. Current work on microbial communities examines the presence of arsenic-resistance genes in NBH isolates. In collaboration with the Plymouth Environmental Research Center, Plymouth University, United Kingdom, we have also used more conventional ecotoxicological approaches to examine the health of the NBH biota.
metal-resistance genesmicrobial diversityRAMPRFLP
==== Body
Participants of the “Roundtable Discussion on Biological Activity of Remediation Products” held at the Asilomar Conference on Bioremediation and Biodegradation, 9–12 June 2002, presented a framework for discussion that highlights the limitations of monitoring techniques for site remediation (Figure 1) (Ford et al. 2002). Research at the New Bedford Harbor (NBH), Massachusetts, USA, a U.S. Environmental Protection Agency (U.S. EPA)–designated Superfund site, has focused partly on developing ecological biomarkers of contaminant exposure for use in monitoring remediation of contaminated sites.
The New Bedford Harbor Superfund Site
New Bedford Harbor is located approximately 40 miles south of Boston (Figure 2). The harbor is a poorly flushed estuary with a long history of metal contamination. Our early research focused on characterizing metal distribution throughout the harbor sediments (Ford et al. 1998; Shine et al. 1995). Sediment from the upper harbor has much higher concentrations of copper, zinc, chromium, lead, and cadmium compared with the lower and outer harbor (Figure 3) (Ford et al. 1998).
The harbor has also been heavily contaminated with polychlorinated biphenyls (PCBs) from capacitor manufacturers in the 1930s and 1940s. PCB usage peaked at about 2 million pounds/year between 1973 and 1975. Sediment concentrations of PCBs were measured as high as 100,000 ppm in the upper harbor. In 1964 a hurricane barrier was constructed that restricted the flushing rates of the estuary and altered flow patterns. Eighteen thousand acres of the harbor and outer bay are currently closed to commercial and recreational fishing (Nelson et al. 1996).
In 1987 a pilot study was conducted to examine dredging and disposal options to remove the most contaminated sediments from the harbor. In 1990 a Record of Decision was signed to remove approximately 10,000 cubic yards of sediment with PCB concentrations > 4,000 ppm. Dredging of this hot spot was completed in the fall of 1995. However, the dredged sediments currently remain in a confined disposal facility at the side of the harbor, prior to eventual removal for land-filling out of state.
Over the years, a number of ecological studies have examined the effects of contaminant exposure on NBH biota. For example, total PCBs in mummichogs have been measured by the U.S. EPA Narragansett Laboratories at 300 μg/g dry weight in the upper harbor, with decreasing concentrations toward Buzzards Bay (Nelson et al. 1996). Black et al. (1998) found concentrations of PCBs in mummichog livers of 35 μg total PCBs/g dry weight associated with increased mortality, reduced survival of progeny, and greater spinal abnormalities.
Monitoring Remediation of Contaminated Sediment
This raises the question: How do we evaluate remediation of contaminated sediment? The U.S. EPA has mandated a long-term monitoring program for NBH and the surrounding waters that includes the collection and analysis of sediment chemistry, bioaccumulation tests, and sediment characterization (acute toxicity tests, grain size and texture, and benthic community structure) (Nelson et al. 1996). For example, using the 75% benthic community abundance measure, the upper harbor is dominated by the polycheate Streblospio benedicti, whereas the lower harbor is dominated by the clam Mulinia lateralis; the outer harbor is dominated by Ostracoda (Nelson et al. 1996). A long-term remediation strategy would therefore be evaluated based on a reduction in concentration of sediment contaminants and a return to community abundance that reflects a less-contaminated state. A number of problems occur, however. Measures of sediment chemistry do not reflect the bioavailable portion of contaminants present, and in fact, a decrease in contaminant concentration could be accompanied by an increase in bioavailability through an inappropriate remediation strategy (e.g., addition of nutrients to stimulate microbial activity; oxygenation of sediments through dredging). Community abundance and diversity is also problematic. For example, Nacci et al. (1999) have shown that Fundulus heteroclitis indigenous to this site actually have an inherited tolerance that allows them to survive in large numbers.
Our program focuses on using rapid ecotoxicological approaches to monitoring the health of NBH. This includes development of microbial biomarkers of contaminant exposure and the application of the rapid assessment of marine pollutants (RAMP) technique developed at the Plymouth Environmental Research Center in the United Kingdom.
The specific aims of our research program are to a) develop a suite of microbial molecular biomarkers that will indicate the bioavailability of contaminants and the potential for adverse ecological effects; b) evaluate the physiological and biochemical responses in the biota to validate the microbial biomarker approach and at the same time provide additional tools to characterize the stress of the aquatic ecosystem; and c) in the long term, provide multiple probes for the analysis of ecosystem health, which can be used as monitoring or screening tools for environmental decision makers who are evaluating remediation alternatives for contaminated aquatic sediments.
It should be noted that the information reported in this article reflects our progress to date and does not seek to provide answers for all the specific aims of our research program. Analysis of gene presence alone in highly contaminated marine sediments is extremely complex. The eventual goal of monitoring gene expression for a suite of microbial biomarkers requires a long-term and research-intensive program.
Past and Current Research
Microbial biomarkers.
Our initial approach to developing microbial biomarkers of exposure to bioavailable contaminants focused on examination of microbial molecular diversity along pollution gradients. If contaminants are not bioavailable, then diversity should not be affected by their presence. We evaluated species diversity by extracting DNA from sediment and subjecting it to restriction fragment length polymorphism (RFLP) analysis of the 16S rRNA genes (Figure 4). We then applied a cluster analysis to the resulting fragment patterns (known as operational taxonomic units) to compare genetic distances. Although our results suggested greater diversity at contaminated sites relative to less-contaminated sites, they also suggested an increased genetic relatedness. This may be consistent with a more constrained (stressed) environment with a wide diversity of organic carbon sources (Sorci et al. 1999).
As with higher organisms, a contaminated environment is likely to select for contaminant-resistant organisms, and measurements of diversity may be misleading. This certainly appears to be the case in NBH, where diversity increases with higher contaminant (and organic carbon) loads. An alternative approach that reflects ongoing research in our laboratory is to look for the presence of specific genes that convey resistance to toxic metals or the ability to degrade toxic organic compounds. For NBH, both approaches are possible. If these genes are present, this provides at least preliminary evidence that the organisms may be exposed to bioavailable contaminants. The eventual goal of the research program is to quantitatively assess both the presence of selected genes, on the assumption that copy number will increase with increasing pollution, and their expression (see “Research Directions”).
Evaluation of metal resistance.
As a starting point for this research, we exploited the ability of bacteria to develop metal resistance as a microbial biomarker. The genes that convey arsenic (As) resistance were used as a model, as they have been well characterized in the literature. The arsenic resistance operon, known as ars, encodes a detoxification system that includes reduction of As(V) to As(III) by the soluble reductase ArsC, followed by extrusion from the cell by the membrane pump ArsB alone or in conjunction with ATPase ArsA (Rosen 1999; Silver 1998; Xu et al. 1998). The ars operon has been found on gram-negative and gram-positive bacteria (Diorio et al. 1995; Silver 1998) on plasmids (Kaur and Rosen 1992), transposons (Summers 1992), and the chromosome of Escherichia coli (Carlin et al. 1995) and Thiobacillus ferrooxidans (Butcher et al. 2000). Ars genes have been observed in Desulfovibrio Ben-RA, isolated from an Australian reed bed (Macy et al. 2000); Pseudomonas fluorescens MSP3, isolated from seawater (Prithivirajsingh et al. 2001); and aerobic bacteria from sewage and As-enriched creek waters (Saltikov and Olson 2002). However, the prevalence of these genes in aerobes and anaerobes from a range of environmental systems is unknown.
Shallow water sediment samples (18 μg/g As, dry weight) were collected from a contaminated NBH field site using a sterilized gravity coring device (Wildco, Inc., Buffalo, NY) and placed on ice for transfer to the laboratory. For aerobic microbiology, bacteria were extracted from sediment (Engelen et al. 1998), serially diluted, and plated onto marine agar plates (Difco Diagnostic Systems, Sparks, MD) amended with concentrations of As (as sodium arsenate) ranging from 0.1 to 10 mM. For anaerobic microbiology, a portion of the core was transferred to the anaerobic chamber (5% hydrogen, 15% carbon dioxide, 80% nitrogen). Triplicate aliquots were diluted in anaerobic solutions of 0.5 M sodium chloride and 0.05 M Tris buffer (pH 7.8) and plated onto basal salt sulfate-reducing media (with lactate, acetate, pyruvate, and butyrate as carbon sources) amended with As (as sodium arsenate) ranging from 0.1 to 10 mM. Individual colonies of As-resistant bacteria were picked (from plates containing arsenate concentrations that inhibited growth and diversity of bacteria) and transferred at least 3 times per isolate before analysis on plates containing 10 and 1 mM arsenate for aerobes and anaerobes, respectively.
Marine agar (for aerobic isolates) and sulfate-reducing agar (for anaerobic isolates) plates were poured with sodium arsenate concentrations of 0, 10, 20, 50, 100, and 150 mM. Aerobic and anaerobic isolates were allowed to grow on the plates at room temperature for 7 and 14 days, respectively. Tolerance was determined as the highest concentration in which growth occurred on both duplicate plates.
Isolates were genetically grouped using RFLP. Cells were lysed by heat, sonication, and alternate rapid freeze/thaw. The 16S rRNA gene was amplified by polymerase chain reaction (PCR) from the crude DNA, RFLP was performed using the restriction endonuclease known as HhaI, and RFLP profiles were grouped manually, allowing approximately 5% variation in fragment length within groups. Of 200 bacterial species isolated, RFLP grouping revealed 14 groups of aerobes and 9 groups of anaerobes.
Plasmid DNA was extracted using a modified phenol/chloroform extraction (25:42:1 phenol/chloroform/isoamyl alcohol) (Ausubel et al. 1995), which denatures and eliminates chromosomal DNA. Genomic DNA was also extracted by a phenol/chloroform method designed to extract total cellular DNA that would yield primarily chromosomal DNA. Although the plasmid extraction method affords virtual certainty that PCR-amplified genes from such samples are located on plasmids, the chromosomal DNA extraction method is less definitive. Although we expect plasmid DNA to be greatly reduced in these extracts, it is possible that large plasmids could be extracted along with the chromosome.
Nested primer sets were developed for the amplification of arsA, arsB, and arsC, the three genes on the ars operon that encode for ArsA ArsB, and ArsC, respectively (Table 1). E. coli with the pUM3 plasmid (kindly provided by B. Rosen) known to contain the ars operon was used as a positive control for plasmid extraction and PCR, and E. coli K12 was used as the chromosomal DNA positive control. Each PCR reaction contained 25 μL HotStarTaq Master Mix (Qiagen, Valencia, CA), 0.1 μM of each primer (Great American Gene Company, Ramona, CA), 0.2 μg of template DNA, and distilled water to give a final volume of 50 μL. Amplification of arsA, arsB, and arsC were multiplexed in the same reaction and run in a Geneamp 2400 thermocycler (PerkinElmer, Norwalk, CT) (15 min at 95°C, 30 cycles of 1 min at 94°C, 1 min at 50°C, 1 min at 72°C, 10 min extension at 72°C, and hold at 4°C).
For 16 aerobic isolates representing 10 RFLP groups, and 7 anaerobes representing 4 RFLP groups, plasmid and chromosomal DNA were screened for the presence of arsA, arsB, and arsC (Figure 5; Table 2). ars Genes were observed in chromosomal extracts of all but three strains tested and 10 plasmid DNA extracts. As previously described, the chromosomal DNA preparation is likely to contain plasmid DNA. However, in six cases, ars genes were detected in plasmid DNA preparations and not in chromosomal DNA preparations, strongly suggesting that they are present on plasmid DNA. Conversely, in 15 cases, ars genes were detected in chromosomal DNA extracts but not in plasmid DNA extracts, suggesting that the genes are either present on chromosomal DNA or on very large plasmids not extracted in the plasmid DNA methodology. This distinction is important both from an ecological viewpoint and for the use of genetic markers as indicators of contaminant stress. Mobile genetic elements may be more rapidly disseminated within the sediment microbial communities in response to bioavailable contaminants and hence may provide a better indicator of contaminant exposure than chromosomal DNA (Ford 2000). Further hybridization studies are clearly warranted on these isolates to better distinguish between chromosomal and plasmid genes.
In 11 cases, all three ars genes were observed together; however, in a number of cases only one or two of the genes were observed. Because the genes are part of the same operon and are regulated together, the absence of observed arsB or arsC along with other ars genes is probably indicative of variations in the gene that decreased the homology with our primer set. Because the arsenite extrusion pump (ArsB) can function alone, absence of observable arsA could indicate nonhomology, or an operon without arsA. In 20 of the 22 cases when any of the ars genes were observed, arsC was present.
Arsenic tolerance levels were determined for each isolate (Table 2). Tolerance range is from 20 to 150 mM, There is no clear relationship between tolerance level and the prevalence of the ars genes. All the anaerobes were able to tolerate at least 50 mM As added to the medium. Tolerance is likely dependent on the speciation in the medium, which we did not measure.
This work shows that the ars genes are prevalent in NBH sediments among both aerobic and anaerobic bacteria with a diverse group of 16S rRNA RFLP patterns. This has important implications for As cycling, as the form of As extruded by this detoxification mechanism [As(III)] is the more mobile and toxic form. The nested primer method described here is capable of amplifying these genes from a contaminated site. These findings may be applied to the use of ars genes as a biomarker for bioavailable As in the field.
As mentioned previously, NBH is contaminated with a wide range of metals and organics, and there are many potential gene targets to use as microbial biomarkers. The As-resistance system is only one example, and our laboratory is currently investigating the presence of a number of other genes (see final section). However, presence alone is likely to be a poor indicator of exposure to bioavailable contaminants, and our current focus is on optimizing RNA extraction from NBH sediments for evaluation of gene expression.
Rapid assessment of marine pollution.
The second component of our research program is to examine more traditional ecotoxicological indicators in higher organisms using the RAMP approach (Wells et al. 2001). The eventual aim is to correlate responses in higher organisms with the microbial approach. For example, if a genotoxic response to a specific pollutant (or mixture) in an invertebrate species increases along a pollution gradient, we might expect increased expression of a specific gene in the microbial population. In these studies we have focused on the biochemical and physiological activity of the Atlantic ribbed mussel, Geukensia demissa, which is extremely common in NBH and the surrounding coastal areas of Buzzards Bay. Geukensia lives partially within the surficial sediments, making it an excellent candidate for this study, as it is directly exposed to high levels of sediment contamination (Figure 6).
The RAMP approach was developed at the Plymouth Environmental Research Center in the United Kingdom. This approach combines chemical residue analysis with measurement of a range of biological responses to determine the ecological health of a marine ecosystem. Approaches include a) evaluating the physiological status of the organism by monitoring its heart rate or condition index; b) evaluating genotoxicity by observing micronucleus formation; c) evaluating cellular status by measures of cell viability and lysosomal integrity; and d) evaluating immunotoxicity through measures of spontaneous cytotoxicity.
A summary of our findings from the RAMP survey found that PCBs and polycyclic aromatic hydrocarbons (PAHs) in the mussel tissue were greatest at the inner harbor site and decreased along a pollution gradient out toward the control site in Buzzards Bay (Figure 7). Chromosomal damage was greatest at the most highly polluted sites, and immune function, heart rate, and cell viability all decreased with increasing pollution (Galloway et al. 2002). Figure 8, adapted from Galloway et al. (2002), illustrates this relationship for heart rate and chromosomal damage. Significant differences in PCB and PAH tissue residues were detected among sites using immunoassay techniques (RaPID assay; Ohmicron Environmental Diagnostics, Inc., Newtown, PA). However, no significant differences were observed in metal concentrations in mussel tissues (copper, cadmium, lead, As, mercury, and nickel) throughout the area. Multivariate canonical correlation analysis indicated that PCB and PAH concentrations were strongly associated with biomarkers of genotoxicity (micronucleus formation), immunotoxicity (spontaneous cytotoxicity), and physiological impairment (heart rate) (Galloway et al. 2002).
Research Directions
Our current goal is to target other resistance or catabolic genes that may be more prevalent than ars genes to use as microbial biomarkers. We have begun to evaluate sediments for the presence of biphenyl-degrading genes; the biphenyl degrading (bph) gene cluster implicated in the degradation of PCBs to chlorobenzoates through the 2,3-deoxygenation pathway (Furukawa and Kimura 1995). Increased copy number/expression of the bph genes is expected at the sediment–water interface as a response to both an overall increase in PCBs and an increase in the more readily biodegradable fraction (Erb and Wagner-Döbler, 1993). Once we have optimized our methodologies for detecting bph genes, we propose to expand the research in the following directions:
Real-time PCR to detect changes in copy number of genes across a pollution gradient
Extraction of mRNA from sediments to assess gene expression
Develop fluorescent in situ hybridization (FISH) probes for mRNA to detect genes in situ
As a long-term goal, be able to examine a number of metal-resistance systems and catabolic genes for PCB degradation concurrently, using DNA array technologies
Validate all methodologies with more traditional biomarkers of exposure (RAMP).
Microbial biomarkers as ecotoxicological tools.
Our long-term goal is to develop multiple probes to evaluate ecological health in marine ecosystems. Our approach will be to develop multiple FISH probes (or micro-arrays) to rapidly hybridize genes that are actively expressed in response to contaminant stress. These biomarkers should correlate with stress (biomarker) responses in higher organisms. We expect microbial biomarkers to be a rapid and sensitive measure of exposure to bioavailable contaminants, as microbes are ubiquitous in the environment, have no migratory behavior, and integrate responses to multiple stressors.
Figure 1 General model of the toxic risks of the remediation products of contaminated sites.
Figure 2 Map of New Bedford Harbor sites.
Figure 3 Metals in New Bedford Harbor. Figure adapted from Ford et al. (1998).
Figure 4 Examples of unique RFLP patterns from New Bedford Harbor area sediments. M designates the marker standard. The 684 base-pair fragment present in all lanes was generated as a result of RSA1 endonuclease digestion of the pCRII vector and used as an internal reference. Figure modified from Sorci et al. (1999).
Figure 5 Electrophoresis gels showing PCR-amplified products of arsA, arsB, and arsC in (A) genomic and (B) plasmid DNA extracts of As-resistant New Bedford Harbor isolates. Numbers refer to specific isolates listed in Table 2.
Figure 6 Atlantic ribbed mussels were used as the bioindicator organism for New Bedford Harbor. Their heart rate was monitored using an infrared heart rate monitor attached to the surface of the mussel’s shell. Photograph courtesy of Ross Sanger, University of Plymouth.
Figure 7 Mussel tissue burdens of PCBs, PAHs, and selected metals (Hg, Cd, Ni, Pb, Cu, As). Figure adapted from Galloway et al. (2002).
Figure 8 Examples of RAMP assays applied to Atlantic ribbed mussels from New Bedford Harbor. Figure adapted from Galloway et al. (2002).
Table 1 Nested primer sets for ars genes.a
ars Gene Outer primer sequence Amplicon size Inner primer sequence Amplicon size
arsA Fb TAT TTC CTG CGC CAC GGC GAT 389 F CTG CTG GTC AGT ACC GAT 300
Rc GAA GGC GAA TGG TGT GAC R GAT ATG GTC AAA CGT CAG
arsB F CCG GTG GTG TGG AAT ATT GT 409 F GTT GCT GGA TGA GTC AGG CT 259
R ACT CCG TGA ATC CCA GTT R GTA TCG GAA ATA CCG GC
arsC F CTG ATA TGA GCA ACA TCA CTA TTT 446 F ATC ATA ACC CAG CCT GC 341
R ATT TCA GCC GTT TTC CTG CTT CA R CTG CGC ATC CTG TAG GAT ARC
a Primer set design was based on ars operon nucleotide sequence of resistance factor R773 (Chen et al. 1986). Primer3 software was used for primer design (version 1.0; Whitehead Institute for Biomedical Research, MIT, Boston, MA; Rozen and Skaletsky 2000).
b Forward.
c Reverse.
Table 2 Presence of arsA, arsB, and arsC on chromosomes and plasmids from 23 randomly selected aerobic and anaerobic New Bedford Harbor isolates and their respective As tolerance.
Isolate Genomic
Plasmid
Tolerance (mM As)
arsA arsB arsC arsA arsB arsC
A1 X X 20
A2 X X X 20
A3 X X X 20
A5 X 20
A6 X X 20
A7 X X 20
A9 X 20
A18 X X X X X 100
A21 X X X X X 50
A24 X X X X X 150
A27 X X X X X 50
A36 –
A44 X X 100
A62 50
A73 X X X 50
A86 X X X 50
N10 50
N11 X X X X 50
N16 X X X X 50
N1 X X X X X X 50
N6 X X ND 150
N4a X X X X X ND
N3a X X X X 150
Abbreviations: A, aerobic isolate; N, anaerobic isolate; ND, not determined.
a Originally isolated on cadmium but also show arsenate resistance.
==== Refs
References
Ausubel FM Brent R Kingston RE Moore DD Seidman JG Smith JA eds. 1995. Current Protocols in Molecular Biology, Vol 1. New York:Wiley.
Black DE Gutjahr-Gobell R Pruell RJ Bergen B Mills L McElroy AE 1998 Reproduction and polychlorinated biphenyls in Fundulus heteroclitus (Linnaeus) from New Bedford Harbor, Massachusetts, USA Environ Toxicol Chem 17 1405 1414
Butcher BG Deane SM Rawlings DE 2000 The chromosomal arsenic resistance genes of Thiobacillus ferrooxidans have an unusual arrangement and confer increased arsenic and antimony resistance to Escherichia coli Appl Environ Microbiol 66 1826 1833 10788346
Carlin A Shi W Dey S Rosen BP 1995 The ars operon of Escherichia coli confers arsenical and antimonial resistance J Bacteriol 177 981 986 7860609
Chen C-M Misra TK Silver S Rosen BP 1986 Nucleotide sequence of the structural genes for an anion pump. The plasmid-encoded arsenical resistance operon J Biol Chem 261 15030 15038 3021763
Diorio C Cai J Marmor J Shinder R DuBow MS 1995 An Escherichia coli chromosomal ars operon homolog is functional in arsenic detoxification and is conserved in gram-negative bacteria J Bacteriol 177 2050 2056 7721697
Engelen B Meinken K von Wintzingerode F Heuer H Malkomes H-P Backhaus H 1998 Monitoring impact of a pesticide treatment on bacterial soil communities by metabolic and genetic fingerprinting in addition to conventional testing procedures Appl Environ Microbiol 64 2814 2821 9687435
Erb RW Wagner-Döbler I 1993 Detection of polychlorinated biphenyl degradation genes in polluted sediments by direct DNA extraction and polymerase chain reaction Appl Environ Microbiol 59 4065 4073 8285706
Ford T Sorci J Ika R Shine JP 1998 Interactions between metals and microbial communities in New Bedford Harbor, Massachusetts Environ Health Perspect 106 1033 1039 9703489
Ford TE 2000 Response of marine microbial communities to anthropogenic stress J Aquat Ecosys Stress Recov 7 75 89
Ford TE Geesy G Barkay T Cerniglia C Ganey P Giesy J 2002. Roundtable discussion on biological activity of remediation products. In: Proceedings of the Asilomar Conference on Bioremediation and Biodegradation: Current Advances in Reducing Toxicity, Exposure and Environmental Consequences, 9–12 June 2002, Asilomar, California. Published on CD by J. Kukor. New Brunswick, NJ:Rutgers University.
Furukawa K Kimura N 1995 Biochemistry and genetics of PCB metabolism Environ Health Perspect 103 suppl 5 21 23 8565902
Galloway TS Sanger RC Smith KL Fillmann G Readman JW Ford TE 2002 Rapid assessment of marine pollution using multiple biomarkers and chemical immunoassays Environ Sci Technol 36 2219 2226 12038833
Kaur P Rosen BP 1992 Plasmid-encoded resistance to arsenic and antimony Plasmid 27 29 40 1531541
Macy J Santini J Pauling B O’Neill A Sly L 2000 Two new arsenate/sulfate-reducing bacteria: mechanisms of arsenate reduction Arch Microbiol 173 49 57 10648104
Nacci D Coiro L Champlin D Jayaraman S McKinney R Gleason TR 1999 Adaptations of wild populations of the estuarine fish Fundulus heteroclitus to persistent environmental contaminants Mar Biol 134 9 18
Nelson WG Bergen BJ Benyi SJ Morrison G Voyer RA Strobel CJ 1996. New Bedford Harbor Long-term Monitoring Assessment Report: Baseline Sampling. EPA/600/R-96/097. Narragansett, RI:U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division.
Prithivirajsingh S Mishra SK Mahadevan A 2001 Detection and analysis of chromosomal arsenic resistance in Pseudomonas fluorescens strain MSP3 Biochem Biophys Res Commun 280 1393 1401 11162686
Rosen BP 1999 Families of arsenic transporters Trends Microbiol 7 207 212 10354596
Rozen S Skaletsky H 2000. Primer3 on the WWW for general users and biologist programmers. In: Bioinformatics Methods and Protocols (Krawetz, Misener SS, eds). Series on Methods in Molecular Biology. Totawa, NJ:Humana Press, 365–386. Code available: http://fokker.wi.mit.edu/primer3
Saltikov CW Olson BH 2002 Homology of Escherichia coli R773 arsA, arsB , and arsC genes in arsenic-resistant bacteria isolated from raw sewage and arsenic-enriched creek waters Appl Environ Microbiol 68 280 288 11772637
Shine J Raveendra I Ford T 1995 Multivariant statistical examination of spatial and temporal patterns of heavy metal contamination in New Bedford Harbor marine sediments Environ Sci Technol 29 1781 1788 22176450
Silver S 1998 Genes for all metals–a bacterial view of the Periodic Table. 1996 Thom Award Lecture J Ind Microbiol Biotechnol 20 1 12 9523453
Sorci J Paulauskis JD Ford T 1999 16S rRNA restriction fragment length polymorphism analysis of bacterial diversity as a biomarker of ecological health in polluted sediments from New Bedford harbor, Massachusetts Mar Poll Bull 38 663 675
Summers AO 1992 Untwist and shout: a heavy metal responsive regulator J Bacteriol 174 3097 3101 1577681
Wells PG Depledge MH Butler JN Manock JJ Knap AH 2001 Rapid toxicity assessment and biomonitoring of marine contaminants—exploiting the potential of rapid biomarker assays and microscale toxicity tests Mar Poll Bull 42 799 804
Xu C Zhou T Kuroda M Rosen BP 1998 Metalloid resistance mechanisms in prokaryotes J Biochem 123 16 23 9504403
|
15687056
|
PMC1277863
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 8; 113(2):186-191
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.6934
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7337ehp0113-00019215687057ResearchArticlesUrinary Creatinine Concentrations in the U.S. Population: Implications for Urinary Biologic Monitoring Measurements Barr Dana B. 1Wilder Lynn C. 2Caudill Samuel P. 1Gonzalez Amanda J. 2Needham Lance L. 2Pirkle James L. 11National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia, USA2Agency for Toxic Substances and Disease Registry, Atlanta, Georgia, USAAddress correspondence to D.B. Barr, Centers for Disease Control and Prevention, 4770 Buford Hwy, Mailstop F17, Atlanta, GA 30341 USA. Telephone: (770) 488-7886. Fax: (770) 488-0142. E-mail:
[email protected] thank the National Center for Health Statistics for their thorough review and thoughtful input into this article.
The authors declare they have no competing financial interests.
2 2005 23 9 2004 113 2 192 200 18 6 2004 23 9 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Biologic monitoring (i.e., biomonitoring) is used to assess human exposures to environmental and workplace chemicals. Urinary biomonitoring data typically are adjusted to a constant creatinine concentration to correct for variable dilutions among spot samples. Traditionally, this approach has been used in population groups without much diversity. The inclusion of multiple demographic groups in studies using biomonitoring for exposure assessment has increased the variability in the urinary creatinine levels in these study populations. Our objectives were to document the normal range of urinary creatinine concentrations among various demographic groups, evaluate the impact that variations in creatinine concentrations can have on classifying exposure status of individuals in epidemiologic studies, and recommend an approach using multiple regression to adjust for variations in creatinine in multivariate analyses. We performed a weighted multivariate analysis of urinary creatinine concentrations in 22,245 participants of the Third National Health and Nutrition Examination Survey (1988–1994) and established reference ranges (10th–90th percentiles) for each demographic and age category. Significant predictors of urinary creatinine concentration included age group, sex, race/ethnicity, body mass index, and fat-free mass. Time of day that urine samples were collected made a small but statistically significant difference in creatinine concentrations. For an individual, the creatinine-adjusted concentration of an analyte should be compared with a “reference” range derived from persons in a similar demographic group (e.g., children with children, adults with adults). For multiple regression analysis of population groups, we recommend that the analyte concentration (unadjusted for creatinine) should be included in the analysis with urinary creatinine added as a separate independent variable. This approach allows the urinary analyte concentration to be appropriately adjusted for urinary creatinine and the statistical significance of other variables in the model to be independent of effects of creatinine concentration.
biomonitoringcreatininecreatinine adjustmenturine
==== Body
Biologic monitoring (i.e., biomonitoring) is used to assess human exposures to environmental and workplace chemicals. The most commonly used matrices for biomonitoring are blood (and its components, e.g., serum and plasma) and urine. The average blood volume of an individual changes an average of 80 mL/kg body weight (Guyton and Hall 2000) and remains relatively constant for a healthy individual who maintains a given body weight; thus, changes in blood concentrations of selected environmental and workplace chemicals in individuals or populations can be readily evaluated. For example, in the Third National Health and Nutrition Examination Survey (NHANES III) (Brody et al. 1994), blood lead concentrations demonstrated the decline in the concentrations of lead in the U.S. population between the Second NHANES (NHANES II), 1976–1980, and the first phase of NHANES III, 1988–1991 (Pirkle et al. 1994). Blood has also been used to evaluate exposures to lipophilic compounds, such as polychlorinated dibenzo-p-dioxins, polychlorinated biphenyls, and organochlorine insecticides. These chemicals are reported in blood and serum based on their lipid content, which varies among individuals and within an individual, especially after eating. Adjusting based upon lipid content allows direct comparisons of their concentrations within and among individuals, regardless of the amount of lipid in the blood, and also comparisons among various biologic matrices, such as blood and adipose tissue (Phillips et al. 1989).
Urine also is a widely used matrix for biomonitoring, especially for nonpersistent chemicals (i.e., chemicals that have short biologic half-lives), such as some current-use pesticides, metals, and drugs. One of the major advantages of using urine in biomonitoring is its ease of collection for spot or grab (untimed) urine samples but not for 24-hr urine voids, because 24-hr collection can be cumbersome, often resulting in improper or incomplete collection. Therefore, spot urine samples, whether first-morning voids or “convenience” samples, are generally used for biomonitoring. The major disadvantages of spot urine samples include the variability in the volume of urine and the concentrations of endogenous and exogenous chemicals from void to void. How to best adjust the urinary concentrations of environmental chemicals in a manner analogous to the adjustment of the concentrations of lipophilic chemicals in blood samples remains a subject of research.
Variations in urinary analyte concentrations from changing water content in urine have been eliminated using urinary excretion rate (UER) calculations (Rigas et al. 2001). To calculate the UER, the metabolite concentration in urine is multiplied by the volume of the void and then divided by the duration of time the void was accumulating in the bladder. This model assumes that the entire bladder is emptied with each void and that the entire sampling void volume is known. Because this is based on the mass in the sample, variability in urine concentrations from urine dilution is removed, particularly for analytes where the rate of excretion varies with the urine flow (Boeniger et al. 1993). However, because the void volume and times of previous and current voids are required, this approach is often not practical for epidemiologic studies, especially those studies involving young children or large population groups.
Urinary creatinine concentrations, specific gravity, and osmolality are common methods for adjusting dilution and for determining whether a spot urine sample is valid for assessing chemical exposures. The most widely used method is creatinine adjustment that involves dividing the analyte concentration (micrograms analyte per liter urine) by the creatinine concentration (grams creatinine per liter urine). Analyte results are then reported as weight of analyte per gram of creatinine (micrograms analyte per gram creatinine).
Many studies have documented that creatinine-adjusted urinary metabolite concentrations correlate better with blood, serum, or plasma concentrations of the parent chemical than the unadjusted concentrations, suggesting that creatinine-adjusted analyte concentrations may serve as good surrogates for size-related dose (Cline et al. 1989; Hill et al. 1995a; Shealy et al. 1997; To-Figueras et al. 1997). However, these studies typically self-correct for size variation because each data pair is from a single individual. Thus, children, who have blood volumes that are proportionately smaller (80 mL less per kilogram body weight), would have higher blood concentrations of a chemical after the absorption of the same amount of chemical after an exposure compared with adults with an identical exposure. Similarly, their lower urinary creatinine concentrations would increase the creatinine-adjusted urine concentration of the metabolite compared with that of an adult with an identical exposure. Therefore, the paired urine and blood values from children and adults can be easily used to determine the relationship between matrices within an individual, but this does not necessarily mean that the creatinine-adjusted metabolite concentrations can be used to accurately compare exposures among the study participants.
Creatinine concentrations also are used to determine whether the spot urinary sample is valid. The guidelines of the World Health Organization (WHO) for valid urine samples for occupational monitoring often are used. The WHO recommends that if a sample is too dilute (creatinine concentration < 30 mg/dL) or too concentrated (creatinine concentration > 300 mg/dL), another urine void should be collected (WHO 1996) and analyzed for creatinine and the target chemical. These guidelines have been adopted for biomonitoring in the U.S. workplace (Lauwerys and Hoet 1993). The U.S. Department of Transportation defines an acceptable urine specimen for the screening of selected drugs of abuse as one that has a creatinine concentration of ≥5 mg/dL and a specific gravity of 1.001–1.020 (Barbanel et al. 2002). Urine of “normal” persons would be unlikely to be excluded using these criteria (Barbanel et al. 2002).
Urine creatinine concentrations were used to adjust the urinary concentrations of pesticides and metabolites of pesticides and phthalates in subsets of adults participating in NHANES III. These “creatinine-corrected” concentrations (micrograms analyte per gram creatinine) were reported in addition to the unadjusted concentrations in micrograms analyte per liter urine (Blount et al. 2000; Hill et al. 1995b). These reports also used the WHO’s recommendation for exclusion of samples, regardless of age (these were all adults), sex, or race/ethnicity.
Because urinary creatinine concentrations are so widely used to adjust or correct urinary concentrations of environmental and workplace chemicals or their metabolites, the formation of urinary creatinine and the ways in which various factors may affect its concentration are important to review. Creatinine is a waste product formed by the spontaneous, essentially irreversible dehydration of body creatine and creatine phosphate from muscle metabolism. A total of 94–98% of total creatine is accumulated within skeletal muscle. The rate of creatinine formation is fairly constant, with approximately 2% of body creatine converted to creatinine every 24 hr; this rate decreases with age in adults.
Creatinine is cleared from the body through the kidney primarily by glomerular filtration. However, 15–20% of the creatinine in urine can occur by active secretion from the blood through the renal tubules (Boeniger et al. 1993). The rate of secretion can vary substantially among persons because of various genetic and biologic factors. Researchers have found a high correlation between urinary creatinine concentrations and muscle mass (Edwards and Whyte 1959; Fuller and Rich 1982); higher urinary creatinine concentrations in men than in women (Bjornsson 1979; Turner and Cohn 1975); decreased urinary creatinine concentrations in adults with increasing age, probably because of a general decline in muscularity and glomerular filtration rate (GFR) (Alessio et al. 1985; Drive and McAlevy 1980); and seasonal variation in creatinine concentrations in children (Freeman et al. 1995; O’Rourke et al. 2000). In addition, persons with a high red meat intake have a higher urinary creatinine concentration than do those on a low-red-meat diet (Lykken et al. 1980). The effects of these factors and others on urinary creatinine concentrations have been reviewed (Boeniger et al. 1993).
Because of the relatively constant excretion rate of creatinine into the urine (which makes urinary creatinine concentration inversely proportional to urine flow rate), creatinine adjustment has been used to normalize analyte concentrations in spot samples for occupational and environmental exposure monitoring. This approach reportedly works well for individual occupational exposure analysis (e.g., preshift and postshift samples from the same person) if the analyte measured behaves similarly to creatinine in the kidney (Boeniger et al. 1993). However, if the analyte is excreted predominantly through passive secretion in the kidney, the analyte secretion will vary with urine flow rate and creatinine adjustment would not correct for urine concentration/dilution.
Urinary creatinine concentration data have been used to adjust urinary concentrations of environmental and workplace chemicals, primarily in adults. Thus, most of the published urinary creatinine concentration data are for adults. However, as more emphasis is placed on children’s health issues and assessment of their exposures to environmental contaminants, biomonitoring of younger populations is also increasing (Needham and Sexton 2000; O’Fallon et al. 2000).
Our study objectives were to document the normal range of urinary creatinine concentrations among various demographic groups, evaluate the influence demographic variations in creatinine concentrations can have on biologic monitoring measurements, and explore methods to appropriately adjust urinary analytes using creatinine that take into account demographic differences in urinary creatinine levels. In this article, we present urinary creatinine concentrations in samples collected during 1988–1994 throughout the United States from NHANES III participants. We describe the distribution of urinary creatinine concentrations within this population by age, sex, and race/ethnicity for persons ≥6 years of age. We also examine other factors that can affect urinary creatinine concentrations, such as body mass index (BMI), fat-free mass (FFM), and health status: kidney function, hyperthyroidism, hypertension, and diabetes (Boeniger et al. 1993). In addition, we compare urinary creatinine concentrations in urine samples collected at three different times throughout the day (morning, afternoon, and early evening). Finally, we propose a multiple regression approach to adjusting urinary analytes for differences in creatinine concentration. This information will greatly assist researchers, occupational health physicians, risk assessors, public health officials, and other users of urinary biomonitoring data to better analyze and interpret urinary biomonitoring measurements.
Materials and Methods
NHANES III, which was conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC), was a 6-year survey during 1988–1994 designed to measure the health and nutrition status of the civilian, noninstitutionalized U.S. population ≥2 months of age. National population estimates and estimates for the three largest racial/ethnic subgroups in the U.S. population (non-Hispanic white, non-Hispanic black, and Mexican American) can be derived from each of the two individual 3-year phases (1988–1991 and 1992–1994) and from the full 6-year survey.
Sampling selection for NHANES III was based on a complex multistage area probability design. Children younger than 5 years, adults ≥60 years of age, non-Hispanic blacks, and Mexican Americans were oversampled to allow an adequate number of sample persons in these demographic groups from which population-based estimates could be derived. However, urine samples were not collected for children < 6 years of age. Data were collected through a household interview, and a standardized physical examination was conducted in a mobile examination center. Urine specimens for analyses, including those for measuring creatinine concentrations, were collected during this examination throughout the day. Pre-examination procedures depended on the age and health status of the individual. For example, persons > 12 years of age were asked to fast for 2–12 hr, depending on appointment times, and persons with known diabetes or < 12 years of age were asked to eat a normal diet before the examination. Sociodemographic information and medical histories of the survey participants and their families were collected during the household interviews. Details of the sample design have been published (Ezzati et al. 1992). The data set used in our analysis is a part of the public release data set for NHANES III (NCHS 2004a).
Laboratory methods
During the physical examinations, urine specimens were collected, stored cold (2–4°C) or frozen, and sent to the Fairview University Medical Center (Minneapolis, MN), where they were analyzed for creatinine using an automated colorimetric determination based on a modified Jaffe reaction using a Beckman Synchron AS/ASTRA clinical analyzer (Beckman Instruments, Brea, CA) (Jaffe 1886). The laboratory and method were certified according to guidelines set forth in the Clinical Laboratory Improvement Act and Amendment (1988).
Demographic covariates
Age was reported at the time of the household interview as the age in years at the last birthday. Age categories used in our statistical analyses were 6–11, 12–19, 20–29, 30–39, 40–49, 50–59, 60–69, and ≥70 years. A composite racial/ethnic variable based on reported race and ethnicity was created to define three major racial/ethnic groups: non-Hispanic black, non-Hispanic white, and Mexican American. Persons who self-reported race as none of the three major racial/ethnic groups were included in the overall estimates but excluded from analyses in which race/ethnicity was the stratification variable.
Health status definitions
The health status of participants was considered in the data analysis. All participants were tested for a variety of physical conditions that have been reported to potentially affect urinary creatinine concentrations. Participants were not screened for a given condition if they reported having been previously informed by a physician as having one of the conditions. Clinical parameters for determining the health status of individuals are summarized in Table 1.
The GFR used for kidney function analysis was calculated using the equation derived from the Modification of Diet in Renal Disease (MDRD) study (Coresh et al 2002, 2003; Levey et al. 2003), in which serum creatinine, age, sex, and race were used. Serum creatinine measurements for the MDRD study and the NHANES III study were performed in different laboratories, and a laboratory bias was observed (Coresh et al. 2002). Thus, serum creatinine values in the NHANES III data set were calibrated to be more comparable with the laboratory data obtained in the MDRD study by subtracting 0.23 mg/dL from each value (Coresh et al. 2002, 2003). For our analysis, we considered persons to have kidney dysfunction if their GFR was < 60 mL/min/1.73 m2, indicative of moderately or severely decreased kidney function (Levey et al. 2003).
Statistical analysis
We analyzed data using the NHANES III analytic guidelines (NCHS 2004b) for sample size and coefficient of variation to ensure reliability of estimates. Survey-specific sample weights were used in statistical analyses. Arithmetic means, selected percentiles of urinary creatinine concentrations, and their respective confidence intervals were calculated using SAS release 8 (SAS Institute, Cary, NC) and the SUDAAN (release 7.5.6; Research Triangle Institute International, Research Triangle Park, NC) Proc Descript procedure. SUDAAN incorporates the NHANES III sampling weights and adjusts for the complex sample design of NHANES III. Sample weights account for the unequal probabilities of selection resulting from the cluster design and the planned oversampling of certain subgroups. Oversampling of children, the elderly, non-Hispanic blacks, and Mexican Americans necessitated the use of sampling weights in all analyses to produce national estimates of prevalence and associated variances. Because Proc Descript does not provide design effect estimates for distribution percentiles, we multiplied the design effect associated with a mean by 30 or 80 [i.e., the NCHS-recommended sample size for estimating a proportion of 0.50 (n = 30) or a proportion of 0.10 or 0.90 (n = 80) when the design effect is 1.0 (NCHS 2004b)]. If this product was larger than the actual sample size, we determined that the percentile estimate should not be reported. All distribution percentiles reported met this criterion.
The collective data set of urinary creatinine values was slightly skewed toward higher values; however, logarithmic transformation did not improve the shape of the distribution. Because the results were only slightly skewed and variance estimates obtained using SUDAAN software were robust, we chose not to transform the urinary creatinine results for the analysis.
An analysis of covariance was used to correct for demographic covariates before comparing concentrations among demographic groups and daily collection times. Statistical significance was set at p < 0.05.
Similar to the approach used by Wilder et al. (unpublished data), we used multiple linear regression models to study the influence of standard demographic variables on urinary creatinine concentration and additional factors previously reported to affect urinary creatinine concentrations. Nine variables were evaluated, although all variables were not used in the final model: race/ethnicity, sex, age, BMI, FFM, diabetes status, hypertension status, hyperthyroid disease status, and kidney disease status. FFM was calculated using a sex-and age-specific bioelectrical impedance analysis equation reported by Deurenberg et al. (1991). Height, weight, age, sex, and reactance measurements (ohms at 50 kHz) were used in the equation to derive individual estimates of FFM. Reactance measurements were available only for persons ≥12 years of age.
Our analysis comprised 22,245 valid creatinine values in urine samples collected during 1988–1994. Although we did not perform a thorough analysis of the rate of nonresponse and its possible effects on our analyses, we did evaluate the potential effects of differential nonresponse using the method of Flegal et al. (1991). We analyzed major demographic variables obtained from the interview data for persons with urinary creatinine values and persons without urinary creatinine values. For each variable, we compared the observed mean urinary creatinine level with the expected mean value for persons in the interviewed sample after we adjusted for that variable. The comparison assumed no statistical significance from differential nonresponse if the estimates were within 10% of the expected means (Flegal et al. 1991). We did not detect bias resulting from differential nonresponse for any of the previously listed variables.
Results
The weighted urinary creatinine arithmetic means, medians, 10th and 90th percentiles, and their respective upper and lower 95% confidence intervals (CIs) are shown in Table 2. The data are shown both collectively and divided into age, race/ethnicity, and sex categories. No data were excluded from the distribution analysis. Non-Hispanic blacks had significantly greater concentrations of urinary creatinine than did all other racial/ethnic groups, across all age groups (p < 0.0006; Figure 1). On average, blacks had 33.43 and 34.25 mg more creatinine per deciliter of urine than did Mexican Americans and non-Hispanic whites, respectively. Adult (i.e., ≥20 years of age) males had significantly greater urinary creatinine than did adult females (p < 0.0001).
The percentage of individuals in each demographic group that had urinary creatinine concentrations outside the WHO exclusionary guidelines is shown in Table 3. Recently, Wilder et al. (unpublished data) reported that these exclusionary criteria should be re-evaluated for urine samples taken from children. In that study (410 children 1–8 years of age), 12% of all children fell below the guideline value, and 0% were too concentrated. Up to 8% of the NHANES samples examined had urinary creatinine concentrations < 30 mg/dL, whereas < 3% had concentrations > 300 mg/dL. Although these percentages differed for each demographic category, more samples were considered “too dilute” than “too concentrated.”
We did not have the information to classify the diabetic status or kidney function of persons 6–19 years of age; thus, we first limited our multiple linear regression analysis to subjects ≥20 years of age to determine the effects of diabetes and kidney function on urinary creatinine. For subjects ≥20 years of age, statistically significant categorical independent variables in the model included race/ethnicity, sex, diabetic status, kidney function status, and age group. The continuous independent variable BMI was also a statistically significant factor. There were statistically significant interactions between race and diabetic status (p = 0.0022), between race and kidney function status (p = 0.0073), between race and age group (p = 0.0028), between sex and age group (p = 0.0260), and between diabetic status and age group (p = 0.0133). Hyperthyroidism, hypertension, and FFM were not significant factors in the model and thus were not included in the final model.
Participants with diabetes tended to have lower urinary creatinine levels than did those without diabetes, and the magnitude of the decrease varied significantly among the three racial/ethnic groups studied and among the age categories. For example, non-Hispanic black participants with diabetes had urinary creatinine levels 34.2 mg/dL lower (p < 0.0001) than those without diabetes in the same ethnic group, whereas no significant differences were observed in the other racial/ethnic groups. Similar variation was observed for persons with diabetes in different age group categories. For example, urinary creatinine levels for persons with diabetes 30–39 years of age were 40.6 mg/dL lower (p = 0.011) than those without diabetes in the same age group.
The effect of kidney dysfunction on urinary creatinine concentration was not the same across racial/ethnic groups. Non-Hispanic whites with kidney dysfunction had urinary creatinine levels 10.7 mg/dL (p = 0.0047) higher than those without kidney disease, whereas the levels for Mexican Americans with kidney disease were 15.5 mg/dL (p = 0.0329) lower than those without.
So that we could include children and adolescents in our analyses, we next performed multiple linear regression analyses that included all ages. Subjects ≥20 years of age were only included if they could be classified as not having diabetes and as not having moderately or severely decreased kidney function.
Coefficients from the multiple linear regression model are presented in Table 5. The R2 of the model was 0.175. Statistically significant categorical independent variables in the model included race/ethnicity, sex, and age group. Neither hyperthyroidism nor hypertension was a significant factor in the model. The continuous independent variable BMI was also a statistically significant factor. Statistically significant interactions were observed between race and age group (p = 0.0002) and between sex and age group (p < 0.0001).
According to the model results, the effect of age category on urinary creatinine concentrations differed among each racial/ethnic group. Among Mexican Americans, urinary creatinine levels for 20- to 29-year-olds were 44.3 mg/dL higher (p < 0.0001) than those for 50- to 59-year-olds. Among non-Hispanic whites, this difference was 55.8 mg/dL (p < 0.0001), and among non-Hispanic blacks, 57.5 mg/dL (p < 0.0001).
BMI also was significantly related (p < 0.0001) to urinary creatinine concentrations. According to the model results, every unit increase in BMI was associated with a 1.30 mg/dL increase in urinary creatinine. Thus, persons with a BMI at the 90th percentile (31.37 kg/m2) would be expected to have urinary creatinine levels about 8.6 mg/dL higher than persons of the same demographic group but with a BMI at the median (24.75 kg/m2). However, when FFM is included in the model, it interacts strongly with BMI. For example, at the median FFM (2574.97 units), a one-unit increase in BMI results in a 0.92-mg/dL increase in urinary creatinine. At the 75th percentile FFM (2692.15 units), a one-unit increase in BMI is associated with a 0.5-mg/dL increase in urinary creatinine. At the 25th percentile FFM (2462.47 units), a one-unit increase in BMI is associated with a 1.33-mg/dL increase in urinary creatinine. Thus, at higher FFM, BMI has a smaller effect on urinary creatinine.
Discussion
Biomonitoring of exposure is used in the workplace to evaluate a person’s chemical exposure during the workday and to provide some standard measure for allowable individual workplace exposures. When timed urine excretion (to determine UER) or 24-hr samples are not collected, the chemical measurement is routinely adjusted using creatinine to correct for urine concentration/dilution in spot samples.
For occupational monitoring, the WHO has recommended exclusionary guidelines for urinary creatinine concentrations to identify individual samples that are invalid for chemical analysis. The rationale behind these guidelines is that urine samples with extremely low creatinine concentrations are too dilute and may impair detection of low levels of toxicants, whereas samples with extremely high creatinine concentrations indicate dehydration, which could have changed the kidney’s secretion, excretion, and/or reabsorption of the target chemical. Therefore, analysis of either dilute or concentrated spot samples would not result in an analyte concentration representative of actual exposures. Typical statistical rules of exclusion of outliers would exclude the upper and lower 1 or 5% of the population. However, our data indicate that in some demographic categories, almost no one would be excluded using these criteria. In other demographic categories, as many as 20% of the participants would be excluded. These data support the findings recently reported by Wilder et al. (unpublished data). For example, essentially no Mexican-American female adults ≥70 years of age had urinary creatinine > 300 mg/dL. However, in the same demographic group, about 19% of the samples would be excluded because their urinary creatinine concentrations were < 30 mg/dL.
The WHO guidelines may have been established for occupational monitoring using a workforce with less diversity than the U.S. workforce. If only non-Hispanic white males 20–60 years of age are considered, approximately 10% of the samples would have been excluded, 5% for each exclusionary criterion. Among both sexes in this age range or women alone, approximately 15% of samples would have been excluded, with the majority (9–13%) excluded for being too dilute. In the U.S. population as a whole, samples from nearly 10 million women could be excluded using criteria that were likely not established using data from women. Clearly, with the change in the composition of the modern U.S. workforce to include women, multiple racial/ethnic groups, and older workers because of the increasing retirement age, the guidelines for sample exclusion should be re-evaluated to reflect the results shown in Table 2. In addition, a special reconsideration, or perhaps elimination, of the lower limit of acceptable creatinine concentration should be given. As analytical technology for measuring environmental toxicants or their metabolites in urine samples has dramatically improved over the last several decades, driving the limits of detection very low, detection of chemicals in urine samples considered “dilute” is much less likely to be an issue of concern. Rather, intermittent or low-level exposures will likely have a greater effect on the ability for a given marker of exposure to be measured with current analytical technology.
We observed a small but statistically significant increase in creatinine concentrations in the morning compared with the afternoon and evening. Although we have no information suggesting the morning urine collections in NHANES III were first morning voids, our analyses appear consistent with the general thought that urine from a first morning void is more concentrated.
In the early 1980s, biomonitoring for nonoccupational, environmental exposures became an important exposure assessment tool in epidemiologic studies evaluating environmental exposure risks. In these studies, 24-hr samples were costly and logistically impractical to collect. Therefore, in keeping with the most common approach in workplace monitoring, spot urine samples were collected and chemical measurements were adjusted using creatinine. This approach was generally considered the only valid way to adjust spot urine samples for comparison across groups, even though limited data were available to evaluate the validity of this adjustment. With the increase in the number of child health studies in the 1990s, including assessing in utero exposures by analyzing the urine of pregnant women, the variation in creatinine concentrations among different age groups has become increasingly apparent. Several researchers have noted significant differences in chemical exposures among children and adults (Aprea et al. 2000; Heudorf and Angerer 2001; Mills and Zahm 2001; Wilder et al., unpublished data), and most have recognized and reported that creatinine adjustment elevates the urinary chemical concentrations in children compared with adults.
The differences between children and adults are due partly to differences in lean muscle mass. Children and the elderly tend to have less muscle than active adults. Accordingly, children have lower FFM than adults. Because lean muscle produces the vast majority of creatinine in the body, we evaluated the relation between FFM and urinary creatinine. Indeed, FFM and urinary creatinine were significantly associated (r = 0.222; p < 0.0001); however, the magnitude of their correlation was much lower than expected. When FFM is considered in the linear regression model, it accounts for much, but not all, of the significant associations with age, sex, and race. Because bioimpedance analysis is not performed in most studies collecting biomonitoring data for exposure assessments, age, sex, and race can be used in concert as a surrogate for FFM. Further, because the FFM accounts for a significant proportion of the variation of creatinine, creatinine-adjusted measurements may serve as a useful surrogate for estimating the size-related dose of an individual (Barr et al. 2004).
Urinary biomonitoring measurements are used to assess exposures of individuals and population groups. For an individual, if the urinary chemical level is divided by the creatinine concentration to adjust for dilution, one must recognize that the urinary creatinine concentration varies by age, sex, and race/ethnicity (Mage et al. 2004). Therefore, it would be best for “normal” or “reference” ranges for creatinine-adjusted urinary levels to be available for separate demographic groups, (e.g., children, adolescents, and adults), rather than just for the total population. The Second National Report on Human Exposure to Environmental Chemicals (National Center for Environmental Health 2003) provides separate reference ranges for 116 chemicals by age, sex, and race/ethnicity. In addition, the report provides reference ranges for non-creatinine-adjusted levels.
For population groups, public health scientists use the creatinine-adjusted urinary chemical level in two types of models. In model 1, the creatinine-adjusted urinary chemical level is a dependent variable, and other variables are regressed against it to determine significant predictors of exposure to that chemical. In model 2, the creatinine-adjusted urinary chemical level is an independent variable used to determine if that chemical exposure is a significant predictor of a disease outcome. In both models, the urinary chemical concentration is typically divided by the urinary creatinine level, and the resulting concentration, expressed per weight of creatinine, is the variable used.
In model 1, where the creatinine-corrected urinary level is the dependent variable, independent variables may be unrelated to the chemical concentration itself but related to the urinary creatinine concentration. In such a case, the independent variable could potentially achieve statistical significance only because it is related to urinary creatinine. Because age, sex, and race/ethnicity all relate to urinary creatinine, this possibility would have to be considered if they were significant predictors of creatinine-corrected urinary chemical levels.
In model 2, a similar problem could exist in which the creatinine-corrected urinary level may be a significant predictor of a health outcome only because the health outcome is related to urinary creatinine levels, not to the levels of the chemical. This would be a less likely scenario than model 1 but is possible because the urinary level is a ratio of a chemical concentration divided by urinary creatinine concentration.
A straightforward solution to both of these potential problems in interpreting multiple regression results is to separate the urinary chemical concentration from the urinary creatinine concentration in the regression models. For model 1, the dependent variable would be the urinary chemical concentration, unadjusted for creatinine. Urinary creatinine concentration would be included in the multiple regression as an independent variable. In this manner, the urinary chemical concentration is adjusted for urinary creatinine, because urinary creatinine is an independent variable, and other covariates in the model are also adjusted for urinary creatinine. Statistical significance of independent variables would therefore not be due to association with urinary creatinine concentration.
Similarly, in model 2, urinary chemical concentration (unadjusted for creatinine) would be included with urinary creatinine as independent variables to predict the health outcome. The health outcome and the urinary chemical concentration variables are adjusted for creatinine by the urinary creatinine independent variable, so any association of the health outcome with chemical concentration would not be influenced by a relationship with urinary creatinine levels.
The present study has several limitations. First, some of the variables used in our evaluation of the data such as the bioimpedance measurements and serum creatinine measurements were available only for persons > 12 years of age. Second, fasting times may have differed among participants and no dietary variables were considered in the analysis. Third, children < 6 years of age were not evaluated. Fourth, first morning void samples were not targeted for collection, so few were likely present in our study; therefore, these findings may not be directly applicable to first morning void samples. Last, upper-bound confidence intervals could not be established for seven of the 90th-percentile estimates given for creatinine levels in different age, sex, and racial/ethnic demographic groups.
Conclusions
Generally, in epidemiologic studies it is not practical to collect 24-hr urine samples or, when young children are involved, even first morning voids. Therefore, spot samples are generally the urine samples that are analyzed for assessing human exposures to many chemicals. The urinary concentrations of these chemicals are often reported on a weight/volume basis and a creatinine-adjusted basis. However, urinary creatinine concentrations differ dramatically among different demographic groups; thus, biomonitoring studies using creatinine concentrations to adjust the concentrations of environmental and occupational chemical concentrations should seriously consider the impact these findings will have on the data. For an individual, the creatinine-adjusted concentration of an analyte should be compared with a “reference” range derived from persons in a similar demographic group (e.g., children with children, adults with adults). For multiple regression analysis of population groups, we recommend that the analyte concentration (unadjusted for creatinine) be included in the multiple regression analysis with urinary creatinine added as a separate independent variable. This approach allows the urinary analyte concentration to be appropriately adjusted for urinary creatinine and the statistical significance of other variables in the model (e.g., age, sex, race/ethnicity) to be independent of effects of urinary creatinine concentration.
Figure 1 Mean urinary creatinine concentrations (mg/dL) for each sex and racial/ethnic group by age group.
Table 1 Clinical parameters for designation of health status of individuals in NHANES III (1988–1994) survey.
Health status Clinical parameter
Diabetesa Blood glucose > 126 mg/dL after 8-hr fast
Hypertensionb Systolic value > 140 mm Hg or diastolic > 90 mm Hg
Hyperthyroidism Serum thyroid-stimulating hormone > 5 μU/mL
Kidney dysfunction Glomerular filtration rate < 60 mL/min/1.73 m2
a Also included individuals who were told by a physician that they had diabetes.
b Also included individuals who were told by one physician two or more time or by two or more physicians that they were hypertensive. Systolic and diastolic measurements were the average of three measurements.
Table 2 Weighted quantiles (95% CIs) of urinary creatinine concentrations (mg/dL) in the NHANES III (1988–1994) study population in persons 6–90 years of age.
Race/ethnicity, age (years) All
Male
Female
No. 10th 50th 90th Mean No. 10th 50th 90th Mean No. 10th 50th 90th Mean
Alla
All 22,245 33.54 (32.07–35.22) 118.6 (115.6–121.4) 237.2 (234.7–241.1) 130.4 (128.2–132.7) 10,610 49.56 (46.08–53.30) 137.2 (134.2–141.0) 254.4 (249.9–262.1) 148.3 (145.3–151.3) 11,635 27.36 (26.04–28.90) 99.49 (97.15–102.9) 217.7 (212.6–224.0) 113.5 (110.7–116.3)
6–11 3,078 42.84 (38.77–46.55) 98.09 (93.81–102.2) 163.1 (157.9–173.4) 102.1 (98.91–105.2) 1,590 49.91 (43.18–55.95) 97.84 (92.79–103.7) 164.7 (158.1–179.6) 104.4 (100.3–108.5) 1,488 33.22 (29.54–40.47) 98.34 (91.40–104.0) 160.6 (153.4–171.7) 99.48 (95.27–103.7)
12–19 3,095 62.14 (56.47–67.04) 150.2 (145.1–158.7) 271.2 (263.0–283.4) 161.5 (156.7–166.2) 1,461 65.27 (60.08–75.83) 151.9 (145.3–163.6) 271.2 (258.8–285.3) 163.6 (157.3–169.9) 1,634 56.04 (46.63–64.40) 149.5 (140.5–158.7) 271.6 (261.3–290.1) 159.3 (153.4–165.1)
20–29 3,438 47.45 (42.53–53.42) 153.8 (147.1–160.7) 275.4 (266.4–294.4) 161.8 (156.6–166.9) 1,608 71.64 (61.62–79.74) 172.8 (161.6–185.3) 297.2 (283.5–324.2) 183.0 (175.4–190.6) 1,830 37.24 (31.64–44.04) 132.8 (126.1–141.6) 246.6 (236.4–264.6) 141.0 (135.0–146.9)
30–39 3,259 31.15 (28.80–36.01) 128.8 (121.5–135.8) 245.7 (239.0–259.1) 138.0 (132.4–143.5) 1,438 44.77 (39.90–55.96) 150.5 (140.2–162.1) 263.3 (251.8–285.2) 157.9 (150.1–165.7) 1,821 27.36 (24.80–29.49) 106.9 (100.6–113.9) 227.7 (215.3–240.0) 118.8 (112.6–125.0)
40–49 2,542 26.32 (23.20–30.79) 119.0 (112.3–124.6) 226.2 (216.4–238.8) 124.6 (120.1–129.1) 1,203 43.24 (33.38–54.56) 146.9 (140.5–154.0) 252.3 (235.8–265.8) 149.7 (143.0–156.4) 1,339 20.49 (17.75–24.31) 89.62 (80.26–96.92) 195.1 (185.0–207.8) 100.6 (95.91–105.3)
50–59 1,823 26.80 (25.02–29.92) 98.43 (92.63–102.9) 206.0 (195.2–217.1) 108.1 (103.8–112.5) 838 39.06 (33.30–47.73) 123.5 (114.4–136.4) 227.7 (216.9–243.5) 131.8 (123.6–139.9) 985 22.54 (20.73–25.22) 73.09 (65.66–81.02) 165.5 (155.0–178.2) 86.06 (80.66–91.46)
60–69 2,243 30.01 (27.67–32.95) 94.22 (89.12–98.97) 193.6 (187.0–200.7) 105.5 (101.8–109.2) 1,134 43.54 (39.06–52.11) 121.4 (114.0–127.0) 213.4 (206.3–231.8) 126.4 (121.2–131.6) 1,109 23.64 (21.53–28.25) 75.37 (69.29–82.09) 167.4 (159.3–179.3) 87.91 (82.50–93.32)
≥70 2,767 29.37 (27.41–31.37) 86.23 (82.36–90.57) 179.6 (175.3–189.1) 97.99 (95.14–100.8) 1,338 43.77 (39.62–50.19) 107.4 (103.0–115.4) 199.2 (188.6–210.9) 117.5 (112.8–122.2) 1,429 23.90 (21.86–26.68) 69.14 (65.23–74.63) 166.5 (157.2–180.0) 84.51 (80.87–88.15)
Non-Hispanic white
All 8,150 30.94 (29.31–33.10) 112.7 (109.7–115.9) 229.5 (224.7–236.1) 124.6 (122.0–127.2) 3,820 45.91 (41.88–50.56) 133.0 (129.6–137.2) 249.2 (242.3–259.3) 144.0 (140.3–147.8) 4,330 25.27 (24.03–26.63) 92.09 (87.71–96.48) 205.9 (200.5–212.9) 106.1 (103.2–109.0)
6–11 800 42.75 (38.08–47.69) 98.11 (92.57–103.8) 155.1 (149.0–166.7) 99.92 (95.85–104.0) 413 48.74 (42.61–59.12) 97.19 (90.56–103.9) 158.6 (149.1–169.9) 102.1 (96.98–107.3) 387 32.95 (28.89–40.18) 99.00 (90.86–107.6) 152.0 (145.5–169.4) 97.48 (92.8–102.16)
12–19 790 55.90 (50.31–63.98) 147.1 (139.2–156.0) 261.4 (248.8–278.3) 156.0 (149.8–162.1) 348 61.84 (55.72–74.85) 147.4 (138.3–159.2) 252.0 (237.8–275.1) 155.7 (147.8–163.6) 442 47.86 (39.55–61.86) 145.3 (135.6–156.6) 269.6 (252.1–301.7) 156.2 (147.5–164.9)
20–29 879 42.72 (36.40–50.15) 143.3 (137.1–155.0) 271.7 (258.8–296.4) 154.8 (148.2–161.4) 388 63.81 (53.31–79.41) 169.7 (159.6–185.2) 299.4 (281.2–333.8) 181.7 (171.4–192.1) 491 31.55 (26.10–40.42) 120.3 (110.2–132.1) 233.7 (214.7–246.5) 128.9 (121.9–136.0)
30–39 1,025 30.03 (27.68–34.07) 123.3 (115.2–131.9) 237.0 (231.2–254.5) 133.3 (126.5–140.1) 437 42.21 (36.31–53.47) 146.3 (132.9–162.3) 252.5 (241.1–286.5) 153.7 (143.4–163.9) 588 25.83 (23.08–29.71) 103.3 (94.32–108.3) 221.3 (208.0–232.5) 113.2 (106.4–120.0)
40–49 893 23.40 (20.58–28.72) 113.9 (106.6–120.2) 219.2 (208.7–234.9) 119.8 (115.0–124.6) 422 37.14 (27.78–49.52) 142.3 (136.0–152.0) 244.5 (225.5–260.1) 145.0 (137.2–152.9) 471 18.61 (16.70–22.96) 78.56 (71.33–95.15) 182.9 (172.8–220.2) 94.46 (89.08–99.83)
50–59 884 26.06 (23.96–28.89) 93.95 (87.35–100.4) 203.3 (186.8–218.4) 104.3 (98.83–109.8) 409 37.74 (32.01–46.84) 117.6 (107.5–133.7) 226.0 (213.8–244.7) 129.0 (118.3–139.6) 475 22.64 (20.46–25.30) 70.44 (60.83–77.02) 154.8 (144.1–170.3) 81.36 (75.52–87.20)
60–69 963 30.04 (27.35–33.20) 90.41 (84.88–97.18) 189.0 (184.0–197.1) 102.7 (98.60–106.8) 495 43.84 (39.83–54.25) 121.0 (110.5–125.2) 210.3 (201.7–231.6) 124.8 (119.4–130.2) 468 22.76 (20.24–27.22) 72.55 (64.84–80.14) 162.9 (153.6–174.0) 83.35 (77.53–89.17)
≥70 1,916 28.69 (26.60–30.86) 84.89 (79.77–88.55) 176.9 (171.2–187.6) 96.03 (92.84–99.22) 908 42.90 (38.70–49.49) 107.0 (101.1–114.9) 196.4 (184.3–209.6) 116.2 (111.0–121.3) 1,008 23.44 (21.27–27.17) 66.73 (63.62–72.34) 160.9 (152.0–173.8) 82.30 (78.16–86.44)
Non-Hispanic black
All 6,664 57.24 (54.37–61.00) 153.3 (149.6–158.1) 282.6 (277.7–289.5) 165.4 (162.3–168.5) 3,117 72.84 (68.31–76.59) 170.3 (164.5–177.6) 298.5 (292.7–310.3) 181.9 (177.3–186.4) 3,547 49.64 (45.94–53.27) 140.1 (136.5–144.4) 265.1 (257.6–272.6) 151.3 (147.8–154.8)
6–11 1,060 53.72 (47.81–58.87) 113.8 (110.1–120.9) 201.2 (192.5–211.3) 120.9 (116.3–125.6) 553 54.58 (47.09–60.66) 113.2 (107.4–120.6) 199.5 (188.2–209.9) 120.3 (114.8–125.8) 507 52.64 (44.80–59.58) 115.6 (108.8–122.1) 203.9 (192.5–215.0) 121.6 (115.7–127.5)
12–19 1,113 83.12 (74.04–92.62) 179.4 (172.1–187.3) 310.9 (302.2–325.6) 193.1 (185.4–200.8) 530 88.38 (76.57–102.1) 188.5 (179.9–200.5) 322.3 (313.1–343.1) 203.9 (193.8–214.0) 583 75.86 (65.30–87.54) 172.3 (163.0–182.3) 295.0 (279.4–317.6) 182.4 (173.6–191.2)
20–29 1,098 82.04 (69.40–93.87) 193.4 (188.1–202.9) 315.0 (301.5–332.9) 200.1 (192.8–207.5) 484 90.23 (76.67–115.3) 207.0 (193.1–224.3) 339.9 (316.1–377.4) 214.7 (202.5–227.0) 614 77.77 (61.32–89.35) 185.2 (175.2–194.3) 292.9 (285.9–315.4) 188.0 (179.6–196.2)
30–39 1,120 64.14 (59.23–68.77) 164.4 (155.0–173.6) 284.9 (272.8–299.6) 172.0 (165.7–178.3) 480 82.70 (72.56–97.43) 186.1 (176.9–197.6) 312.0 (290.2–326.5) 193.1 (184.2–202.1) 640 56.48 (48.35–63.58) 148.7 (140.9–157.0) 267.4 (252.3–283.1) 155.3 (146.8–163.9)
40–49 798 53.76 (47.98–65.24) 152.8 (140.8–169.3) 275.2 (260.6–288.1) 164.2 (155.8–172.6) 359 78.50 (68.15–92.70) 180.6 (171.0–192.2) 293.5 (279.4–321.4) 189.5 (181.2–197.7) 439 44.54 (36.66–53.85) 130.0 (119.9–146.6) 238.3 (226.1–267.4) 142.9 (132.2–153.7)
50–59 475 35.78 (28.68–48.23) 134.6 (118.0–150.0) 245.3 (228.4–264.5) 164.2 (155.8–172.6) 210 67.50 (57.98–81.14) 165.3 (151.2–174.7) 269.7 (242.4–NE) 169.0 (157.8–180.1) 265 26.01 (22.35–36.53) 111.0 (95.86–125.3) 217.8 (195.0–232.7) 117.2 (105.9–116.7)
60–69 557 47.22 (41.17–54.59) 115.9 (107.1–126.2) 224.5 (210.3–242.4) 140.0 (130.7–149.3) 279 62.87 (49.25–75.93) 150.1 (139.9–162.2) 270.2 (245.7–288.8) 158.6 (149.3–167.9) 278 41.38 (36.32–50.60) 96.15 (90.57–103.2) 186.4 (171.7–207.9) 108.8 (100.8–116.7)
≥70 443 38.79 (34.55–46.27) 110.9 (105.6–120.3) 209.8 (203.8–224.2) 129.3 (122.4–136.3) 222 49.05 (40.90–56.87) 130.0 (11.2–145.9) 220.8 (204.7–NE) 136.0 (126.7–145.4) 221 34.58 (30.28–42.77) 104.1 (93.44–109.7) 203.4 (185.0–NE) 112.2 (103.5–120.9)
Mexican American
All 6,496 38.35 (35.69–42.13) 123.3 (120.2–126.4) 236.5 (231.6–243.7) 132.9 (129.7–136.1) 3,253 50.52 (45.74–55.94) 138.2 (133.2–144.5) 252.5 (245.1–264.9) 147.2 (142.4–151.9) 3,243 30.80 (28.08–34.80) 106.0 (102.7–110.1) 218.3 (211.2–224.9) 117.6 (114.3–120.9)
6–11 1,083 31.92 (26.14–37.65) 87.99 (82.28–92.45) 154.4 (142.6–166.6) 92.24 (87.67–96.82) 548 32.22 (25.61–41.30) 89.53 (84.26–97.65) 160.3 (144.3–173.5) 94.76 (89.59–99.93) 535 30.10 (24.10–38.54) 85.55 (77.88–95.20) 152.2 (135.4–165.8) 89.57 (82.58–96.55)
12–19 1,039 57.50 (50.37–35.25) 140.0 (134.9–145.8) 249.0 (236.3–265.6) 148.2 (142.1–154.3) 518 57.72 (47.04–70.12) 142.3 (133.6–152.3) 255.7 (237.5–275.7) 151.5 (141.3–161.6) 521 56.75 (46.27–65.77) 133.6 (127.5–145.7) 240.4 (226.5–262.8) 144.8 (138.4–151.3)
20–29 1,311 50.96 (42.23–61.01) 148.9 (142.8–156.9) 261.9 (247.8–282.7) 155.4 (150.2–160.7) 664 65.14 (52.75–77.76) 166.9 (157.8–174.4) 276.3 (258.8–297.8) 168.9 (162.4–175.4) 647 38.47 (33.15–48.69) 126.9 (117.1–137.6) 246.5 (230.1–269.9) 138.5 (132.0–145.0)
30–39 979 36.71 (32.00–44.12) 132.4 (126.9–138.5) 251.8 (236.7–265.9) 139.9 (133.5–146.3) 464 52.12 (45.63–63.38) 152.2 (143.9–159.7) 270.9 (259.3–285.2) 160.9 (153.1–168.6) 515 29.57 (25.79–33.52) 108.8 (101.0–115.5) 216.1 (189.2–236.6) 116.7 (108.9–124.6)
40–49 738 36.33 (31.16–45.31) 126.5 (118.7–136.6) 227.7 (217.3–240.6) 133.0 (126.6–139.3) 376 53.58 (45.62–66.69) 146.3 (138.0–157.9) 244.7 (231.8–263.5) 153.6 (146.1–161.1) 362 30.05 (23.07–40.38) 105.3 (90.92–120.9) 202.8 (190.3–216.5) 111.2 (101.6–120.7)
50–59 367 27.21 (22.37–33.48) 99.12 (85.91–113.7) 196.2 (188.0–210.9) 109.1 (100.2–118.1) 177 46.38 (37.21–60.52) 125.2 (106.7–151.5) 218.5 (202.3–NE) 134.5 (122.6–146.4) 190 18.60 (15.41–26.66) 71.09 (58.38–91.33) 170.7 (153.9–183.6) 85.27 (76.36–94.18)
60–69 641 21.39 (19.67–27.51) 88.47 (79.08–97.59) 196.8 (184.4–211.1) 99.68 (94.73–104.6) 326 28.99 (19.82–47.54) 111.2 (102.1–130.0) 201.6 (174.7–217.9) 116.9 (108.4–125.3) 315 20.89 (14.85–25.71) 66.85 (57.74–80.61) 192.2 (157.1–NE) 86.36 (77.82–94.90)
≥70 338 27.77 (21.93–35.03) 94.66 (87.62–104.8) 174.8 (160.4–201.0) 99.46 (91.72–107.2) 180 59.55 (50.02–70.02) 116.8 (101.7–135.7) 190.8 (175.7–NE) 123.8 (114.3–133.3) 158 20.84 (16.97–29.49) 67.79 (55.72–81.33) 145.7 (116.5–NE) 76.47 (66.0–86.94)
NE, could not be reliably estimated.1
a All population data, including those individuals not grouped into one of the three race/ethnicity categories, are presented.
Table 3 Percentage of each demographic group in NHANES III (1988–1994) whose urinary creatinine concentrations (mg/dL) fell outside the WHO guideline range (i.e., < 30 mg/dL or > 300 mg/dL).
All
Male
Female
Race/ethnicity, age(years) No. < 30 mg/dL > 300 mg/dL No. < 30 mg/dL > 300 mg/dL No. < 30 mg/dL > 300 mg/dL
All
All 22,245 7.7 3.3 10,610 4.0 4.6 11,635 11 2.2
6–11 3,078 4.7 0.1 1,590 2.9 0.1 1,488 6.7 0.1
12–19 3,095 2.3 6.5 1,461 1.6 6.0 1,634 3.1 7.0
20–29 3,438 5.2 6.9 1,608 3.4 10 1,830 7.0 4.2
30–39 3,259 8.4 4.2 1,438 4.3 6.4 1,821 12 2.0
40–49 2,542 11 2.5 1,203 5.9 3.8 1,339 16 1.3
50–59 1,823 12 0.9 838 6.0 1.5 985 17 0.3
60–69 2,243 9.3 0.6 1,134 3.9 1.2 1,109 14 0.1
≥70 2,767 10 0.7 1,338 3.5 1.1 1,429 15 0.5
Non-Hispanic white
All 8,150 8.8 3.0 3,820 4.5 4.2 4,330 13 1.8
6–11 800 4.3 0.0 413 2.6 0.0 387 6.1 0.0
12–19 790 3.0 6.1 348 2.0 4.6 442 3.9 7.6
20–29 879 6.2 6.4 388 3.9 10 491 8.4 3.0
30–39 1,025 9.2 4.0 437 4.9 6.2 588 14 1.8
40–49 893 13 2.3 422 7.0 3.5 471 19 1.1
50–59 884 13 0.6 409 7.4 1.1 475 18 0.2
60–69 963 9.3 0.4 495 3.0 0.8 468 15 0.0
≥70 1,916 11 0.8 908 3.6 1.2 1,008 15 0.5
Non-Hispanic black
All 6,664 2.8 7.1 3,117 1.5 9.8 3,547 3.8 4.8
6–11 1,060 3.4 0.6 553 2.7 0.4 507 4.2 0.8
12–19 1,113 0.6 12 530 0.2 15 583 1.1 8.5
20–29 1,098 1.7 13 484 1.6 17 614 1.7 9.5
30–39 1,120 2.8 7.6 480 1.8 12 640 3.5 4.5
40–49 798 3.3 5.8 359 1.4 8.2 439 4.9 3.7
50–59 475 6.9 3.5 210 1.1 5.8 265 12 1.6
60–69 557 2.4 2.6 279 1.0 54 278 3.4 0.7
≥70 443 4.9 1.1 222 3.6 1.2 221 5.9 0.6
Mexican American
All 6,496 6.5 3.1 3,253 4.4 4.3 3,243 8.8 1.8
6–11 1,083 8.9 0 548 8.0 0.0 535 9.8 0.0
12–19 1,039 2.8 4.2 518 2.0 5.0 521 3.5 3.4
20–29 1,311 4.8 5.4 664 3.8 6.5 647 6.1 3.9
30–39 979 6.7 3.5 464 3.9 5.4 515 9.7 1.4
40–49 738 6.5 2.1 376 4.0 4.0 362 9.2 0.2
50–59 367 10 1.5 177 3.3 3.3 190 16 0.0
60–69 641 15 0.3 326 10 0.8 315 19 0.0
≥70 338 11 0.0 180 2.8 0.0 158 19 0.0
Table 4 Weighted mean urinary creatinine concentration (mg/dL) for each collection time frame during the day.
Collection time frame No. Mean creatinine (mg/dL) Contrasted to morning Contrasted to afternoon Contrasted to evening
Morning 10,621 133.5 NA p = 0.058 p = 0.001
Afternoon 7,190 128.6 p = 0.058 NA p = 0.27
Evening 4,434 126.1 p = 0.001 p = 0.27 NA
NA, not applicable. The concentrations were corrected for age, race/ethnicity, sex, and BMI. Each mean was contrasted to the means of other collection time frames using an analysis of covariance test to determine whether they were statistically different.
Table 5 Coefficients of the independent variables from the multiple linear regression model of urinary creatinine concentrations (dependent variable).
Independent variable
Variable Coefficient ± SE p-Value
Intercept 53.51 ± 6.83 < 0.0001
Race/ethnicity
Non-Hispanic white (1) −7.33 ± 5.00 0.1486
Non-Hispanic black (2) 20.82 ± 5.68 0.0006
Mexican American (3) 0.00 ± 0.00 NA
Sex
Male (1) 34.59 ± 4.14 < 0.0001
Female (2) 0.00 ± 0.00 NA
Age group (years)
6–11 (1) 12.55 ± 5.24 0.0026
12–19 (2) 62.90 ± 5.64 < 0.0001
20–29 (3) 43.56 ± 5.70 < 0.0001
30–39 (4) 29.78 ± 5.78 < 0.0001
40–49 (5) 16.65 ± 6.42 0.0125
50–59 (6) −1.17 ± 6.24 0.8524
60–69 (7) −8.47 ± 4.81 0.0847
≥70 (8) 0.00 ± 0.00 NA
BMI (continuous) 1.30 ± 0.19 < 0.0001
Race/ethnicity × age group
(1) × (1) 16.19 ± 6.09 0.0106
(1) × (2) 16.14 ± 6.67 0.0192
(1) × (3) 10.74 ± 6.68 0.1141
(1) × (4) 4.34 ± 5.66 0.4469
(1) × (5) −2.40 ± 6.94 0.7308
(1) × (6) −0.82 ± 5.73 0.8864
(1) × (7) 6.99 ± 4.86 0.1569
(1) × (8) 0.00 ± 0.00 NA
(2) × (1) 8.64 ± 6.48 0.1886
(2) × (2) 24.28 ± 6.48 0.0005
(2) × (3) 28.19 ± 6.50 0.0001
(2) × (4) 15.01 ± 7.12 0.0403
(2) × (5) 14.69 ± 7.77 0.0648
(2) × (6) 14.98 ± 8.27 0.0762
(2) × (7) 8.58 ± 6.35 0.1826
(2) × (8) 0.00 ± 0.00 NA
(3) × (1) 0.00 ± 0.00 NA
(3) × (2) 0.00 ± 0.00 NA
(3) × (3) 0.00 ± 0.00 NA
(3) × (4) 0.00 ± 0.00 NA
(3) × (5) 0.00 ± 0.00 NA
(3) × (6) 0.00 ± 0.00 NA
(3) × (7) 0.00 ± 0.00 NA
(3) × (8) 0.00 ± 0.00 NA
Sex × age group
(1) × (1) −30.64 ± 4.26 < 0.0001
(1) × (2) −30.44 ± 5.86 < 0.0001
(1) × (3) 11.57 ± 5.30 0.0339
(1) × (4) 6.01 ± 7.16 0.4051
(1) × (5) 15.86 ± 5.53 0.0061
(1) × (6) 12.53 ± 7.57 0.1045
(1) × (7) 9.39 ± 5.51 0.0944
(1) × (8) 0.00 ± 0.00 NA
(2) × (1) 0.00 ± 0.00 NA
(2) × (2) 0.00 ± 0.00 NA
(2) × (3) 0.00 ± 0.00 NA
(2) × (4) 0.00 ± 0.00 NA
(2) × (5) 0.00 ± 0.00 NA
(2) × (6) 0.00 ± 0.00 NA
(2) × (7) 0.00 ± 0.00 NA
(2) × (8) 0.00 ± 0.00 NA
NA, not applicable. Numbers in parentheses correspond to the specific racial/ethnic group, sex, or age group for which the interaction term was derived.
==== Refs
References
Alessio L Berlin A Dell’Orto A Toffoletto F Ghezzi I 1985 Reliability of urinary creatinine as a parameter used to adjust values of urinary biological indicators Int Arch Occup Environ Health 55 99 106 3988361
Aprea C Strabi M Novelli MT Lunghini L Bozzi N 2000 Biologic monitoring of exposure to organophosphorus pesticides in 195 Italian children Environ Health Perspect 108 521 525 10856025
Barbanel CS Winkelman JW Fischer GA King AJ 2002 Confirmation of the Department of Transportation criteria for a substituted urine specimen J Environ Med 44 407 416
Barr DB Bravo R Weerasekera G Caltabiano LM Whitehead RD Jr Olsson AO 2004 Concentrations of dialkyl phosphate metabolites of organophosphorus pesticides in the U.S. population Environ Health Perspect 112 186 200 14754573
Bjornsson TD 1979 Use of serum creatinine concentrations to determine renal function Clin Pharmacokinet 4 200 222 383355
Blount BC Silva MJ Caudill SP Needham LL Pirkle JL Sampson EJ 2000 Levels of seven urinary phthalate metabolites in a human reference population Environ Health Perspect 108 979 982 11049818
Boeniger MF Lowry LK Rosenberg J 1993 Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: a review Am Ind Hyg Assoc J 54 615 627 8237794
Brody DJ Pirkle JL Kramer RA Flegal KM Matte TD Gunter EW 1994 Blood lead levels in the US population JAMA 272 277 283 8028140
Cline RE Hill RH Jr Phillips DL Needham LL 1989 Pentachlorophenol measurements in body fluids of people in log homes and workplaces Arch Environ Contam Toxicol 18 475 481 2774665
Clinical Laboratory Improvement Act and Amendment 1988. Baltimore, MD:Centers for Medicare and Medicaid Services. Available: http://www.cms.hhs.gov/clia/ [accessed 16 December 2004].
Coresh J Astor BC Greene T Eknoyan G Levey AS 2003 Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey Am J Kidney Dis 41 1 1 12 12500213
Coresh J Astor BC McQuillan G Kusek J Green T Van Lente F 2002 Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate Am J Kidney Dis 39 5 920 929 11979335
Deurenberg P Van der Kooy K Leenen R West-Strate JA Seidell JC 1991 Sex and age specific prediction formulas for estimating body composition from bioelectrical impedance: a cross-validation study Int J Obesity 15 17 25
Driver AG McAlevy MT 1980 Creatinine height index as a function of age Am J Clin Nutr 33 2057 2062 7416074
Edwards KDG Whyte HM 1959 Creatinine excretion and body composition Clin Sci 18 361 366 13819422
Ezzati TM Massey JT Waksburg J Chu A Maurer KR 1992 Sample design: Third National Health and Nutrition Examination Survey Vital Health Stat 2 113 1 39
Flegal KM Keyl RM Nieto FJ 1991 Differential misclassification arising from nondifferential errors in exposure measurement Am J Epidemiol 134 10 1233 1244 1746532
Freeman NC Wainman T Lioy PJ Stern AH Shupack SI 1995 The effect of remediation of chromium waste sites on chromium levels in urine of children living in the surrounding neighborhood J Air Waste Manag Assoc 45 604 614 7664166
Fuller L Rich AJ 1982 An index of lean body mass from 24th urinary creatinine excretion [Abstract] Proc Nutr Soc 41 A104
Guyton AC Hall JE 2000. Textbook of Medical Physiology. 10th ed. Philadelphia:W.B. Saunders Company.
Heudorf U Angerer J 2001 Metabolites of organophosphorous insecticides in urine specimens from inhabitants of a residential area Environ Res 86 80 87 11386745
Hill RH Jr Ashley DL Head SL Needham LL Pirkle JL 1995a p -Dichlorobenzene exposure among 1,000 adults in the United States Arch Environ Health 50 277 280 7677426
Hill RH Jr Head SL Baker S Gregg M Shealy DB Bailey SL 1995b Pesticide residues in urine of adults living in the United States: reference range concentrations Environ Res 71 99 108 8977618
Jaffe M 1886 Ueber den niederschlag welchen pikrinsaure in normalen harn erzeugt und uber eine neue reaction des kreatinins [in German] Z Physiol Chem 10 391 400
Lauwerys RR Hoet P 1993. Industrial Chemical Exposure: Guidelines for Biological Monitoring. Boca Raton, FL:Lewis Publishers.
Levey AS Coresh J Balk E Kausz AT Levin A Steffes MW 2003 National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification Ann Intern Med 139 137 147 12859163
Lykken GI Jacob RA Muno JM Sandstead HH 1980 A mathematical model of creatine metabolism in normal males—comparison between theory and experiment Am J Clin Nutr 33 2674 2685 7435436
Mage DT Allen R Gondy G Smith W Barr DB Needham LL 2004. Estimating pesticide dose from urinary pesticide concentration data by creatinine correction in the Third National Health and Nutrition Examination Survey. J Expo Anal Environ Epidemiol 10.1038/sj.jea.7500343.
Mills PK Zahm SH 2001 Organophosphate pesticide residues in urine of farmworkers and their children in Fresno County, California Am J Ind Med 40 571 577 11675626
National Center for Environmental Health 2003. Second National Report on Human Exposure to Environmental Chemicals. Atlanta, GA:Centers for Disease Control and Prevention. Available: www.cdc.gov/exposurereport [accessed 5 June 2003].
NCHS 2004a. NHANES Public Release Data Sets. Hyattsville, MD:National Center for Health Statistics. Available: http://www.cdc.gov/nchs [accessed 12 February 2004].
NCHS 2004b. NHANES Analytic Guidelines. Hyattsville, MD:National Center for Health Statistics. Available: http://www.cdc.gov/nchs/data/nhanes/nhanes_general_guidelines_june_04.pdf [accessed 12 February 2004].
Needham LL Sexton K 2000 Assessing children’s exposure to hazardous environmental chemicals: an overview of selected research challenges and complexities J Expo Anal Environ Epidemiol 10 611 629 11138654
O’Fallon LR Collman GW Dearry A 2000 The National Institute of Environmental Health Sciences’ research program on children’s environmental health J Expo Anal Environ Epidemiol 10 630 637 11138655
O’Rourke MK Lizardi PS Rogan SP Freeman NC Aguirre A Saint CG 2000 Pesticide exposure and creatinine variation among young children J Expo Anal Environ Epidemiol 10 672 681 11138659
Phillips DL Pirkle JL Burse VW Bernert JT Henderson LO Needham LL 1989 Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding Arch Environ Contam Toxicol 18 495 500 2505694
Pirkle JL Brody DJ Gunter EW Kramer RA Paschal DC Flegal KM 1994 The decline in blood lead levels in the United States JAMA 272 284 291 8028141
Rigas ML Okino MS Quackenboss JJ 2001 Use of a pharmacokinetic model to assess chlorpyrifos exposure and dose in children, based on urinary biomarker measurements Toxicol Sci 61 374 381 11353146
Shealy DB Barr JR Ashley DL Patterson DG Jr Camann DE Bond AE 1997 Correlation of environmental carbaryl measurements with serum and urinary 1-naphthol measurements in a farmer applicator and his family Environ Health Perspect 105 510 513 9222136
To-Figueras J Sala M Otero R Barrot C Santiago-Silva M Rodamilans M 1997 Metabolism of hexachlorobenzene in humans: association between serum levels and urinary metabolites in a highly exposed population Environ Health Perspect 105 78 83 9074885
Turner WJ Cohn S 1975 Total body potassium and 24-hour creatinine excretion in healthy males Clin Pharm Ther 18 405 412
WHO 1996. Biological Monitoring of Chemical Exposure in the Workplace. Vol 1. Geneva:World Health Organization.
|
15687057
|
PMC1277864
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 23; 113(2):192-200
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7337
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7523ehp0113-00020115687058ResearchArticlesAmbient Air Pollution and Atherosclerosis in Los Angeles Künzli Nino Jerrett Michael Mack Wendy J. Beckerman Bernardo LaBree Laurie Gilliland Frank Thomas Duncan Peters John Hodis Howard N. Divisions of Environmental Health and Biostatistics, Department of Preventive Medicine, and Atherosclerosis Research Unit, Division of Cardiovascular Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California, USAAddress correspondence to N. Künzli, Keck School of Medicine University of Southern California, Division of Environmental Health, 1540 Alcazar St. CHP 236, Los Angeles, CA 90033-9013 USA. Telephone: (323) 442-2870. Fax: (323) 442-3272. E-mail:
[email protected] work was supported in part by the National Institute on Aging [grants R01AG-13860 (Vitamin E Atherosclerosis Prevention Study) and R01AG-17160 (B-Vitamin Atherosclerosis Intervention Trial)], the National Institute of Environmental Health Sciences (grants P30 ES07048, 5P01ES11627), the Wright Foundation, the Hastings Foundation, and the Health Effects Institute.
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 201 206 26 8 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Associations have been found between long-term exposure to ambient air pollution and cardiovascular morbidity and mortality. The contribution of air pollution to atherosclerosis that underlies many cardiovascular diseases has not been investigated. Animal data suggest that ambient particulate matter (PM) may contribute to atherogenesis. We used data on 798 participants from two clinical trials to investigate the association between atherosclerosis and long-term exposure to ambient PM up to 2.5 μm in aerodynamic diameter (PM2.5). Baseline data included assessment of the carotid intima-media thickness (CIMT), a measure of subclinical atherosclerosis. We geocoded subjects’ residential areas to assign annual mean concentrations of ambient PM2.5. Exposure values were assigned from a PM2.5 surface derived from a geostatistical model. Individually assigned annual mean PM2.5 concentrations ranged from 5.2 to 26.9 μg/m3 (mean, 20.3). For a cross-sectional exposure contrast of 10 μg/m3 PM2.5, CIMT increased by 5.9% (95% confidence interval, 1–11%). Adjustment for age reduced the coefficients, but further adjustment for covariates indicated robust estimates in the range of 3.9–4.3% (p-values, 0.05–0.1). Among older subjects (≥60 years of age), women, never smokers, and those reporting lipid-lowering treatment at baseline, the associations of PM2.5 and CIMT were larger with the strongest associations in women ≥60 years of age (15.7%, 5.7–26.6%). These results represent the first epidemiologic evidence of an association between atherosclerosis and ambient air pollution. Given the leading role of cardiovascular disease as a cause of death and the large populations exposed to ambient PM2.5, these findings may be important and need further confirmation.
air pollutionatherosclerosisparticulate matter
==== Body
A large body of epidemiologic evidence suggests associations between ambient air pollution and cardiovascular mortality and morbidity (Peters and Pope 2002; Pope et al. 2004). All of these studies focus on events occurring at a late stage of vascular disease processes. The impact of air pollution on the underlying preclinical conditions remains poorly understood. We hypothesize that current levels of ambient particulate matter (PM) up to 2.5 μm in aerodynamic diameter (PM2.5) may contribute to atherosclerosis, leading to subclinical anatomical changes that play a major role in cardiovascular morbidity and mortality later in life. Animal studies support our hypothesis by showing that inhalation of ambient PM promotes oxidative lung damage, including alveolar and systemic inflammatory responses (Becker et al. 1996; Dye et al. 2001; Fujii et al. 2002; Goto et al. 2004; Suwa et al. 2002; van Eeden et al. 2001).
We investigated the association between residential ambient PM2.5 and carotid artery intima-media thickness (CIMT) using pre-randomization baseline data from two recent clinical trials conducted in Los Angeles, California (Hodis et al. 2002). CIMT is a well-established quantitative measure of generalized atherosclerosis that correlates well with all of the major cardiovascular risk factors, with coronary artery atherosclerosis, and with clinical cardiovascular events (Mack et al. 2000). It is an established tool for investigating the contribution of long-term exposures such as smoking or passive smoking to subclinical stages of atherosclerosis at any given age (Diez-Roux et al. 1995; Howard et al. 1994, 1998). This is the first study to assess the association of atherosclerosis with air pollution.
Materials and Methods
Population and health assessment
We used baseline health data from two randomized, double-blind, placebo-controlled clinical trials conducted at the University of Southern California Atherosclerosis Research Unit (Hodis et al. 2002). The Vitamin E Atherosclerosis Progression Study (VEAPS) investigated the effects of vitamin E on the progression of atherosclerosis measured by CIMT. The B-Vitamin Atherosclerosis Intervention Trial (BVAIT) focused on the effect of vitamin B supplements on the progression of atherosclerosis (trial in progress). Baseline assessment in both trials included CIMT measured between 1998 and 2003 using the same standardized methods (Hodis et al. 2002; Selzer et al. 1994, 2001). Recruitment of volunteers occurred over the entire Los Angeles Basin, covering a geographic area of approximately 64,000 km2.
Eligible subjects for the VEAPS trial (n = 353) were men and women ≥40 years of age with slightly increased LDL cholesterol (≥3.37 mmol/L) but with no clinical signs or symptoms of cardiovascular disease (CVD) (Hodis et al. 2002). Subjects with diabetes, diastolic blood pressure > 100 mm Hg, thyroid disease, serum creatinine > 0.065 mmol/L, life-threatening diseases, or high alcohol intake were excluded.
BVAIT (n = 506) had a similar design to that of VEAPS. Men and women > 40 years of age were prescreened to meet study criteria (fasting plasma homocysteine ≥8.5 μmol/L; postmenopausal for women; no evidence of diabetes, heart disease, stroke, or cancer). Subjects were excluded on the basis of any clinical signs or symptoms of CVD, diabetes or fasting serum glucose ≥140 mg/dL, triglyceride levels ≥150 mg/dL, serum creatinine > 1.6 mg/dL, high blood pressure, untreated thyroid disease, life-threatening disease with prognosis < 5 years, or high alcohol intake.
Thus, our study included “healthy” subjects with biomarkers (elevated LDL cholesterol or homocysteine) that suggested an increased risk of future CVDs (n = 859). Fifty-eight subjects were excluded in the exposure assignment process because they lived outside the area with PM2.5 data. Three subjects had missing data in at least one of the covariates used in the models. Our total sample consisted of 798 participants.
Health measures, including CIMT
Our main outcome of interest is CIMT. In both trials, high-resolution B-mode ultrasound images of the right common carotid artery were obtained before the intervention (baseline) with a 7.5-MHz linear array transducer attached to an ATL Ultramark-4 Plus Ultrasound System (Ultramark, Bothell, WA). We used this baseline CIMT measurement as the outcome. Details of this highly reproducible method are published (Hodis et al. 2002; Selzer et al. 1994, 2001). Blood pressure, height, and weight were measured with standard procedures.
The baseline questionnaires included an assessment of all major CVD risk factors and covariates, including clinical events, diet, use of prescription medications, physical activity, current and past smoking and passive smoking, and vitamin supplements. Age, education, and other sociodemographic factors were available for each subject. Fasting blood samples were also drawn for lipid measurements. Data used in our analyses were collected with the same tools in both trials.
Exposure assignment
To assess exposure we chose a novel approach derived from a geographic information system (GIS) and geostatistics. This method allows for assignment of long-term mean ambient concentrations of PM2.5 to the ZIP code area of each subject’s residential address (Künzli and Tager 2000). The resulting surface of PM2.5 covered the entire Los Angeles metropolitan area. The surface is derived from a geostatistical model and data from 23 state and local district monitoring stations (during 2000). These monitors are located across the Los Angeles region to characterize urban levels of pollution. To assign exposure, PM2.5 data were interpolated using a combination of a universal kriging model with a quadratic drift and a multi-quadric radial basis function model (Bailey and Gatrell 1995; Burrough and McDonnell 1998). We averaged the two surfaces based on 25-m grid cells. Examination of errors from the universal model showed that > 50% of the study area had assigned values within 15% of monitored concentrations, whereas 67% were within 20%. The larger errors were on the periphery of our study area, where the density of study participants was the lowest. We linked the ZIP code centroids of each subject with the exposure surface through a geocoding database [Environmental Systems Research Institute (ESRI) 2004]. Figure 1 illustrates the PM2.5 surface with the geolocated ZIP codes. Individually assigned PM2.5 data had a range from 5.2 to 26.9 μg/m3 (mean, 20.3), thus exceeding the range observed across 156 metropolitan areas used in the largest cohort study of air pollution and mortality (Pope et al. 2002). All models were implemented with ArcScript from ESRI (Redlands, CA).
Statistical analyses
We tested the univariate and multivariate associations between CIMT and ambient PM2.5 using linear regression analyses. Extensive residual diagnostics indicated some heteroskedasticity, which was rectified with the natural log-transformed CIMT. We adjusted for factors that were statistically associated with both CIMT and ambient PM2.5 (age, male sex, low education, and low income). Next, we expanded the models using covariates that were associated with either PM2.5 or CIMT, including indicator variables for current second-hand smoke exposure and current and former personal smoking. We then added covariates that play a role in atherosclerosis such as blood pressure, LDL cholesterol, or proxy measures such as reporting treatment with antihypertensives or lipid-lowering medications at study entry. These factors may affect the pathophysiologic pathways linking air pollution exposure and atherosclerosis (Ross 1999); thus, such models may overadjust the coefficients. We chose this conservative approach to test the sensitivity of the effect estimates under a broad range of model assumptions.
There is increasing evidence that host factors such as age, sex, or underlying disease and risk profiles may modify the effects of air pollution (Pope et al. 2002; Zanobetti and Schwartz 2002). Furthermore, the finding of atherosclerosis in PM-exposed rabbits was based on a hyperlipidemic trait (Suwa et al. 2002). Therefore, we also stratified by sex, age (< 60 years, ≥60 years), smoking status, and lipid-lowering drug therapy.
Results
Table 1 summarizes the main characteristics of the study population and among main subgroups. Table 2 presents the percent change in CIMT in association with a 10 μg/m3 contrast in ambient PM2.5 concentrations for three cross-sectional regression models. The unadjusted model indicates a 5.9% [95% confidence interval (CI), 1–11%] increase in CIMT per 10 μg/m3 PM2.5. For the observed contrast between lowest and highest exposure (20 μg/m3 PM2.5), this corresponds to a 12.1% (2.0–23.1%) increase in CIMT. The only covariate with a substantial effect on the point estimate was age, which reduced the effect from 5.9 to 4.3% (0.4–9%) per 10 μg/m3 PM2.5. This change agrees with the age-related effect modification. Otherwise, effect estimates across the models remained robust, in the range of 3.9–4.3% with p-values from 0.05 to 0.1. To corroborate the exposure–response relationship, we also categorized PM2.5 levels into quartiles. Figure 2 shows the adjusted mean CIMT across these four groups of equal sample size at the mean levels of the covariates (age, sex, education, and income). The trend across the exposure groups was statistically significant (p = 0.041). The unadjusted means of CIMT among these quartiles of exposure were 734, 753, 758, and 774 μm, respectively.
The associations between CIMT and PM2.5 were substantially stronger among 109 subjects reporting lipid-lowering medication at study entry, both in men and in women (Table 2, Figure 3). The crude effect reached 15.8% (2–31%) per 10 μg/m3 PM2.5, with adjusted values ranging between 12 and 16%. Despite the small sample size, p-values of all models were mostly < 0.1 and often < 0.05.
Results also suggest significant age and sex interactions, with much larger effects in women and in the older age group (Figure 3). Effect estimates in women were statistically significant and typically in the range of 6–9% per 10 μg/m3 PM2.5. Associations were strongest among women ≥60 years of age (n = 186), leading to crude estimates of 19.2% (9–31%). Adjusted coefficients ranged from 14 to 19%, being statistically significant in all models and sensitivity analyses.
Among never smokers (n = 502), the effect estimate reached 6.6% (1.0–12.3%). The estimate was small and not significant in current (n = 30) and former smokers (n = 265).
Discussion
Our study presents the first evidence for an association between CIMT and long-term exposure to ambient air pollution. As recently reviewed in a statement of the American Heart Association (Brook et al. 2004) substantial epidemiologic and experimental evidence suggests a contribution of ambient air pollutants on cardiovascular mortality and morbidity. However, these studies focus on acute and subacute effects on cardiac autonomic function, inflammatory or thrombogenic markers, arrhythmia, myocardial infarction, cardiovascular hospital admission, and death. The only outcome considered in long-term air pollution studies has been mortality. The relative risks for acute effects on mortality have been substantially smaller than those observed for long-term associations (Pope et al. 2002; Samet et al. 2000). As shown previously, cohort studies are capable of capturing acute and chronic effects of air pollution on the course of diseases that ultimately lead to premature death (Künzli et al. 2001). In contrast, time-series and panel studies investigate only the associations of event occurrence with the most recent exposure (Künzli et al. 2001). Thus, if air pollution has both acute and cumulative long-term effects, one expects larger mortality coefficients in cohort studies. CIMT reflects long-term past exposure; thus, we provide the first evidence for chronic effects of air pollution on atherogenesis that may in part explain the above mentioned discrepancy between acute and long-term risk estimates (Pope et al. 2002; Samet et al. 2000).
There are several major aspects to be considered in the interpretation of this new finding, mainly the strength in the exposure assignment, the limited evidence for bias, the differences in effects within subgroups, and plausibility.
Exposure assignment
The individual residence-based assignment of exposure represents a substantial improvement over most studies that have relied on central monitors or on binary road buffers combined with basic interpolation (Hoek et al. 2002; Pope et al. 2004). As a sensitivity analysis, we used weighted least-squares models with the weights specified as the inverse of the standard errors from the universal kriging model to down-weight estimates with larger error. In addition, we implemented models based solely on the universal kriging estimate. In both instances results were robust and similar to what we found with our main model.
Time–activity studies show that people spend most of their time in or around home, and our restriction of exposure assessment on residential address captures the most relevant part of exposure (Leech et al. 2002). PM2.5 generally displays spatially homogeneous distributions across small areas such as neighborhoods and blocks, and as a result, the ambient conditions at the ZIP code centroid likely reflect the levels expected at home outdoors (Roosli et al. 2000). PM2.5 of outdoor origin will also penetrate indoors, and correlations between long-term outdoor PM concentrations and indoor levels of PM from outdoor origin is high (Sarnat et al. 2000). Exposure to ambient air pollution while working and during commute are not included in our exposure term but are considered to be a relevant source of exposure (Riediker et al. 2003). Although most likely a random misclassification with biases toward the null, the errors may affect subgroups differently, thus explaining part of the observed interactions.
In Los Angeles, no clear trends have been observed in PM2.5 concentrations over the past 5–10 years. The year 2000 surface characterizes the prevailing mean PM2.5 concentrations across several years and can be considered a measure of long-term past exposure. This year also sits in the middle of the baseline recruitment period. Overall, the various limitations in our exposure assignment may add some random error, biasing results toward weaker associations (Thomas et al. 1993).
We also assigned ambient ozone to ZIP code centroids. Inclusion of ozone in the models had no impact on the PM2.5 coefficients or the SEs. Ozone and PM2.5 were not correlated (r = –0.17), and the PM2.5 estimates were not substantially different in low-and high-ozone regions. The estimates of association for ozone were positive but not statistically significant and much smaller than for PM2.5. This finding must be put in context of the specific challenges in determining long-term exposure to ozone, which are substantially different than in the case of PM exposure. In contrast to PM2.5 from outdoor origin, ambient ozone levels have lower correlations with personal exposure (Avol et al. 1998; Sarnat et al. 2000, 2002); therefore, the ability to detect effects of ozone will likely be reduced due to greater misclassification.
Biases
Our subjects were a nonrandom sample of “healthy” volunteers with above-average education, meeting strict inclusion criteria for the two clinical trials. Although we cannot exclude some systematic selection biases affecting the cross-sectional data, it is unlikely that subjects with preclinical signs of atherosclerosis would have been more likely to volunteer if they lived in more polluted areas. Although the selection of subjects limits the generalization to other populations, we do not expect this to lead to over- or underestimating the cross-sectional associations. The two trials recruited subjects independently; thus, the effects may be compared across trials to evaluate the potential influence of selecting volunteers. The populations differed with regard to age, smoking habits, baseline LDL and treatment, blood pressure, active and passive smoking, and other relevant factors; thus, the PM2.5 coefficients were smaller and were not statistically significant in the VEAPS trial with its younger population. However, after taking these factors into account, the associations with ambient PM2.5 were similar. For example, among elderly women of VEAPS (n = 70) and BVAIT (n = 116), the effect estimate was 18.1% (–0.1 to 36.3.%) and 13.6% (2.8–24.4.%), respectively. There is some evidence for larger effects in subjects with cardiovascular risk factors, indicated by prescriptions of lipid-lowering treatment. Our trials excluded subjects with clinically manifest CVDs. Moreover, if air pollution amplifies systemic inflammation among those prone to atherosclerosis, exclusion of subjects with high LDL may be a source of bias. One may expect effect estimates in a less selected, less healthy population to be larger than those reported.
The wealth of baseline data from these clinical trials offered the opportunity to control for a broad array of covariates. Apart from the effect of age adjustment, estimates were robust to numerous combinations of covariates, including income, education, active and passive tobacco smoke, cardiovascular prescriptions, vitamin intake, and physical activity. Uncontrolled or residual confounding appears to be an unlikely explanation for these results. Among women, adjustment for hormone replacement therapies did not affect the PM2.5 estimates.
In previous studies, we found that spatial autocorrelation in the residuals could affect the size and significance of pollution coefficients (Jerrett et al. 2003a). We investigated spatial autocorrelation of the unstandardized residuals. We assessed autocorrelation with first-order, adjusted first-order, and second-order spatial weight matrices based on nearest neighbor contiguity, but we found no evidence of spatial autocorrelation. This supports the conclusion that the models supply efficient unbiased estimates (Jerrett et al. 2003b). As part of our sensitivity analyses, we also derived PM2.5 surfaces using different interpolations and weighted least squares with weights equal to the inverse of the standard error of the exposure estimate. All approaches produced very similar results.
Evidence for effect modification
The data suggest substantial interactions with age, sex, smoking, and underlying cardiovascular risk factors. Given the reduced sample size among subgroups, the recruitment of volunteers, and the cross-sectional nature of the data, it is difficult to fully explore the causes of the observed modifications of associations and to establish susceptibility profiles. If the exposure misclassifications differed across subgroups, part of the interactions may be explained by differential exposure error. The sex and age difference could also be an artifact due to measurement error in the assigned exposure because time spent in commuting and location of work places may be different in men and women and in the young and elderly. Empirical studies on mobility suggest women have smaller activity spaces than men and younger groups, meaning they tend to spend more time in and around the home (Kwan and Lee 2004), and the same is probably true of the elderly compared with younger groups. Exposure measurement error may be reduced in those spending more time at home, leading to stronger effects (Thomas et al. 1993). Moreover, differences in statistical power may play a role as well; as shown at least for the 25–40-year age range, power to detect effects on CIMT is larger in women than in men (Stein et al. 2004).
The finding that those reporting prescriptions of lipid-lowering medications at baseline showed stronger associations of CIMT with PM2.5 merits further investigation. This result agrees with the observed effects of PM on atherosclerosis in experiments conducted in hyperlipidemic rabbits (Goto et al. 2004; Suwa et al. 2002). The systemic inflammatory and atherogenic reaction in these rabbits was related to the amount of PM contained in the alveolar macrophages. In our study, being under lipid-lowering therapy is an indicator for risk profiles prone to atherogenesis. Those subjects were mostly men (64%) and, on average, older, more often active or passive smokers, and almost twice as likely to report antihypertensive treatment. The systemic response to ambient PM may amplify and expand the oxidation of LDL cholesterol among these susceptible subjects, consequently contributing to injury in the artery wall (Goto et al. 2004; Ross 1999). Investigations of short-term effects of ambient air pollution on mortality also suggest that underlying risk profiles such as diabetes may amplify susceptibility to ambient PM (Zanobetti and Schwartz 2002), and similar findings have been shown with smoking and diabetes mellitus in association with CIMT (Karim et al. 2005). To clarify the relevance of lipid status, it would be interesting to investigate our hypothesis among cohorts with familial hypercholesteremia (Wiegman et al. 2004; Wittekoek et al. 1999).
As shown in Figure 3, the size of the point estimate was larger among the older subjects. Future research needs to clarify whether air pollution contributes to atherosclerosis only after a certain age or early on. Effects of air pollution on lung development have been observed during adolescence and may be a result of both pulmonary and chronic systemic inflammatory effects (Gauderman et al. 2002); thus, it is conceivable that atherogenic responses may occur early in life. The age dependence of the effects may also be codetermined by genetic factors (Humphries and Morgan 2004; Ross 1999).
We also observed larger effects in women. If other cardiovascular risk factors such as occupational exposures dominate atherosclerosis in men, we would expect a smaller effect signal and less precision in the estimates among men. We also hypothesize that interactions may reflect biologic causes. If premenopausal women are protected against atherosclerosis by endogenous hormones, loss of hormonal protection would lead to increased vulnerability after menopause (Kannel et al. 1976). This could explain part of the interaction by both age and sex.
Active and passive smoking did not confound results in either the total sample or among subgroups. Adjustment for active tobacco smoke led to a slight increase in the effect estimate; thus, residual confounding is unlikely to overestimate the effects. However, PM2.5 associations were clearly stronger in never smokers compared with smokers (data not shown). This gradient was also observed in all subgroups with significant PM2.5 associations (Figure 3). Oxidative and inflammatory effects of smoking may dominate to such an extent that the additional exposure to ambient air pollutants may not further enhance effects along the same pathways. The difference in the effects of PM2.5 in smokers and nonsmokers needs further investigation. The American Cancer Society cohort study does not reveal a clear pattern of a smoking interaction for the association of ambient air pollution and cardiovascular death (Krewski et al. 2004; Pope et al. 2004). In the Study on Air Pollution and Lung Diseases in Adults (SAPALDIA), associations between air pollution and level of pulmonary function did not differ by smoking status (Ackermann-Liebrich et al. 1997).
Some U.S. studies indicate effect modification of air pollution by socioeconomic status (SES) with much stronger effects among the less educated (Pope et al. 2002). The cause of this interaction pattern is not well understood. SES status was rather homogeneous in these mostly well-educated volunteers, providing little power to investigate interactions of pollution with SES. If lower SES also positively modifies effects of air pollution on atherosclerosis, our population would provide an underestimate of the health effects in the general population (O’Neill et al. 2003). Further research on samples representative of the population will be needed to assess whether the high SES in the clinical trials biases the effects toward the null.
Future research should focus on identifying factors that determine susceptibility to PM2.5. We are initiating studies on subjects with inflammatory metabolic syndromes prone to accelerated atherosclerosis such as postmenopausal women, diabetics, or obese or physically inactive people. To corroborate the cross-sectional findings, follow-up studies are ultimately needed to investigate the association of concurrent levels of air pollution exposure with the progression of CIMT.
Plausibility
From a biologic perspective, our results support the hypothesis that long-term exposure to ambient PM contributes to systemic inflammatory pathways, which are a relevant aspect of atherogenesis (Ross 1999). The findings indicate a biologically plausible link between the observed acute effects of ambient air pollution on systemic inflammation (Glantz 2002) and the long-term consequences of sustained vascular inflammation leading to increased atherosclerosis and, ultimately, cardiovascular death (Hoek et al. 2002; Pope et al. 2004). Among susceptible people, this may lead to artery wall lesions similar to those observed in the rabbit model (Fujii et al. 2002; Suwa et al. 2002). In these hyperlipidemic rabbits, 4-week PM exposure was associated with the progression of atherosclerotic lesions, coupled with an enhanced release of bone marrow monocytes. These precursors of macrophages play an important role in the atherogenic inflammatory responses (Goto et al. 2004; Ross 1999; Suwa et al. 2002). Given the central role of oxidized LDL in the initiation and progression of atherogenesis, suggestions that the plasma of automotive workers with high exposure to traffic exhaust is more susceptible to oxidation is also of interest (Sharman et al. 2002).
As a quantitative plausibility check, we compared the size of the PM2.5 effects with effects of other risk factors on CIMT. Using smoking and environmental tobacco smoke (ETS) as a model for air pollution exposure, the size of our estimates appear plausible (Diez-Roux et al. 1995; Howard et al. 1994). Associations of ETS and current levels of air pollution with various respiratory outcomes are similar and support the notion of common underlying pathways (Künzli 2002). Smoking and ETS associate with stiffer and thicker artery walls, reflecting the systemic effect of these exposures (Howard et al. 1994; Mack et al. 2003). Exposure to ETS was associated with 2–3% thicker intima-media, which approximate the effects observed for a 10 μg/m3 change in PM2.5 (Diez-Roux et al. 1995; Howard et al. 1994). Using never smokers without ETS exposure as the referent group in our data, never smokers with ETS at home had 0.9% (–2.7 to 4.5%) thicker artery walls; former smokers’ CIMT was increased on average by 3.4% (0.7–6.3%), and the 30 current smokers had 5% (–1.5 to 11.6%) thicker CIMT. The trend across these four categories of tobacco exposure was statistically significant. As shown in Table 1, smokers were underrepresented in these volunteers of well-educated participants.
The observed percent change in CIMT corresponds to an increase in the thickness of approximately 20–40 μm per 10 μg/m3 contrast in PM2.5. This difference in CIMT translates into some 3–6% increase in the long-term risk for myocardial infarction (O’Leary et al. 1999). Pope et al. (2004) reported that long-term exposure to PM2.5 was associated with an 18% (14–23%) increase in ischemic heart disease. Effect sizes reported here concur with these findings, indicating that a fraction of the total effect of ambient PM on cardiovascular mortality may be mediated through sustained long-term effects of air pollution on atherosclerosis (Künzli et al. 2001). This is in line with the proposed model (Künzli et al. 2001) in which some of the effects observed in cohort studies must reflect long-term contributions of air pollution to the underlying disease progression, whereas in other cases, air pollution contributes only to triggering of cardiovascular events or death (Bell et al. 2004; Künzli et al. 2001; Peters and Pope 2002).
From a biologic and policy perspective, we emphasize that PM2.5 probably serves as a surrogate for the mixture of urban air pollution and constituents of PM. It is premature to conclude that PM2.5 and its constituents are the atherogenic culprit per se. Atherosclerosis results from complex processes that may include a combination of various urban pollutants, host factors, and pathways that ultimately lead to the findings of a CIMT–PM2.5 association.
In conclusion, we have presented the first epidemiologic evidence supporting the idea of a chronic vascular response to respiratory and systemic effects of PM exposure. Given the leading role of heart disease as a cause of death in most westernized countries and the growing contribution in developing countries, these findings may be of high public health relevance. Further investigations need to focus on susceptible groups and follow-up of cohorts to investigate the effect of air pollution on the progression of CIMT.
Figure 1 ZIP code locations of the study population geocoded on the PM2.5 surface, modeled with 2000 PM2.5 data, and distribution of individually assigned concentrations.
Figure 2 Mean CIMT ± 1 SE among quartiles of the PM2.5 distribution. The y-axis shows mean CIMT levels at the population average of the adjustment covariates (age, sex, education, and income). The first quartile is the reference group.
Figure 3 Percent difference and 95% CI in CIMT associated with a 10 μg/m3 contrast in ambient PM2.5 in all subjects and in subgroups. Lipid-LT, lipid-lowering therapy. All estimates are based on the cross-sectional linear model with log intima-media thickness as the dependent variable and home outdoor PM2.5 as the independent variable, adjusted for sex, age, education, and income. Numbers in parentheses are numbers of subjects per group. Data are ordered by size of point estimate; the null effect line is indicated by a dash.
Table 1 Description of assigned exposure (outdoor concentration in 2000) and CIMT, and main characteristics of the study population at the time of baseline measurements in the total sample, men, women, women ≥60 years of age, and subjects under lipid-lowering therapy.
Characteristic Total sample (n = 798) Males (n = 443) Females (n = 355) Females ≥60 years (n = 186) Lipid-lowering therapy (n = 109)
PM2.5 (μg/m3) 20.3 ± 2.6 20.1 ± 2.7 20.5 ± 2.4 20.7 ± 2.3 20.0 ± 2.5
Ozone (ppb, annual mean of daily maximum) 89.2 ± 17.9 89.6 ± 18.5 88.8 ± 17.3 87.1 ± 17.2 88.5 ± 18.6
CIMT (μm) 755 ± 148 767 ± 166 740 ± 118 775 ± 120 788 ± 140
Age (years) 59.2 ± 9.8 58.3 ± 10.3 60.4 ± 8.9 67.3 ± 5.3 63.3 ± 10.0
Diastolic blood pressure (mm Hg) 77.8 ± 9.2 79.2 ± 8.8 75.9 ± 9.3 74.8 ± 9.5 78.1 ± 8.9
Systolic blood pressure (mm Hg) 127.2 ± 16.3 126.7 ± 16.0 127.8 ± 16.6 130.5 ± 16.7 130.9 ± 16.2
LDL cholesterol (mg/dL) 137.9 ± 29.5 137.0 ± 30.9 139.0 ± 27.6 136.4 ± 26.9 125.7 ± 33.7
White (%) 67.3 67.7 66.8 65.0 69.7
Smoking status (%)
Never smokers 62.9 62.8 63.1 62.9 53.2
Former smokers 33.2 33.4 33.0 33.3 44.0
Current smokers 3.8 3.6 3.9 3.8 2.8
ETS at home (%) 33.5 21.9 47.9 55.4 37.5
Lipid-lowering therapy (%) 13.7 15.3 11.5 15.1 100
Antihypertensive prescriptions (%) 26.2 26.6 25.6 33.3 42.2
ETS, environmental tobacco smoke. Data are mean ± SD except where indicated.
Table 2 Percent change (and 95% CI) in CIMT (μm) associated with a 10 μg/m3 change in ambient outdoor PM2.5 concentration at the residential ZIP code in the total population (n = 798).a
Total sample (798)
Females ≥60 years (186)
Lipid-lowering therapy (109)
Modela (with adjustment factors in the model) Percent change p-Value Percent change p-Value Percent change p-Value
None (unadjusted estimate) 5.9 (1.0–10.9) 0.018 19.2 (8.8–30.5) 0.001 15.8 (2.1–31.2) 0.024
Age, sex, education, incomeb 4.4 (0.0–9.0) 0.056 15.7 (5.7–26.6) 0.002 13.3 (0–28.5) 0.051
All above plus active and passive smoking, multivitamins, alcohol 4.2 (–0.2–8.9) 0.064 13.8 (4.0–24.5) 0.002 13.3 (–0.3–28.8) 0.060
a Unadjusted association (crude model) and estimates from two multivariate models; 95% CIs of the estimates are shown in parentheses. The relative effects are based on a linear model with log intima-media thickness as dependent variable.
b Factors with univariate associations with both, CIMT and PM2.5.
==== Refs
References
Ackermann-Liebrich U Leuenberger P Schwartz J Schindler C Monn C Bolognini G 1997 Lung function and long term exposure to air pollutants in Switzerland. Study on Air Pollution and Lung Diseases in Adults (SAPALDIA) Team Am J Respir Crit Care Med 155 1 122 129 9001300
Avol EL Navidi WC Colome SD 1998 Modeling ozone levels in and around Southern California homes Environ Sci Technol 32 463 468
Bailey T Gatrell A 1995. Interactive Spatial Data Analysis. New York:Longman.
Becker S Soukup JM Gilmour MI Devlin RB 1996 Stimulation of human and rat alveolar macrophages by urban air particulates: effects on oxidant radical generation and cytokine production Toxicol Appl Pharmacol 141 2 637 648 8975789
Bell ML Samet JM Dominici F 2004 Time-series studies of particulate matter Annu Rev Public Health 25 247 280 15015920
Brook RD Franklin B Cascio W Hong Y Howard G Lipsett M 2004 Air pollution and cardiovascular disease: a statement for healthcare professionals from the Expert Panel on Population and Prevention Science of the American Heart Association Circulation 109 21 2655 2671 15173049
Burrough P McDonnell R 1998. Principles of Geographical Information Systems. New York:Oxford University Press.
Diez-Roux AV Nieto FJ Comstock GW Howard G Szklo M 1995 The relationship of active and passive smoking to carotid atherosclerosis 12–14 years later Prev Med 24 1 48 55 7740015
Dye JA Lehmann JR McGee JK Winsett DW Ledbetter AD Everitt JI 2001 Acute pulmonary toxicity of particulate matter filter extracts in rats: coherence with epidemiologic studies in Utah Valley residents Environ Health Perspect 109 suppl 3 395 403 11427389
ESRI 2004. ArcGIS. Redlands, CA:Environmental Systems Research Institute. Available: http://www.esri.com/software/arcgis/index.html [accessed 29 December 2004].
Fujii T Hayashi S Hogg JC Mukae H Suwa T Goto Y 2002 Interaction of alveolar macrophages and airway epithelial cells following exposure to particulate matter produces mediators that stimulate the bone marrow Am J Respir Cell Mol Biol 27 1 34 41 12091243
Gauderman WJ Gilliland GF Vora H Avol E Stram D McConnell R 2002 Association between air pollution and lung function growth in Southern California children: results from a second cohort Am J Respir Crit Care Med 166 1 76 84 12091175
Glantz SA 2002 Air pollution as a cause of heart disease. Time for action J Am Coll Cardiol 39 6 943 945 11897433
Goto Y Hogg JC Shih CH Ishii H Vincent R van Eeden SF 2004 Exposure to ambient particles accelerates monocyte release from the bone marrow in atherosclerotic rabbits Am J Physiol Lung Cell Mol Physiol 287 L79 85 15003928
Hodis HN Mack WJ LaBree L Mahrer PR Sevanian A Liu CR 2002 Alpha-tocopherol supplementation in healthy individuals reduces low-density lipoprotein oxidation but not atherosclerosis: the Vitamin E Atherosclerosis Prevention Study (VEAPS) Circulation 106 12 1453 1459 12234947
Hoek G Brunekreef B Goldbohm S Fischer P van den Brandt PA 2002 Association between mortality and indicators of traffic-related air pollution in the Netherlands: a cohort study Lancet 360 9341 1203 1209 12401246
Howard G Burke GL Szklo M Tell GS Eckfeldt J Evans G 1994 Active and passive smoking are associated with increased carotid wall thickness. The Atherosclerosis Risk in Communities Study Arch Intern Med 154 11 1277 1282 8203995
Howard G Wagenknecht LE Burke GL Diez-Roux A Evans GW McGovern P 1998 Cigarette smoking and progression of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study JAMA 279 2 119 124 9440661
Humphries SE Morgan L 2004 Genetic risk factors for stroke and carotid atherosclerosis: insights into pathophysiology from candidate gene approaches Lancet Neurol 3 4 227 235 15039035
Jerrett M Burnett R Goldberg M Sears M Krewski D Catalan R 2003a Spatial analysis for environmental health research: concepts, methods, and examples J Toxicol Environ Health A 66 16–19 1783 1810 12959844
Jerrett M Burnett RT Willis A Krewski D Goldberg MS DeLuca P 2003b Spatial analysis of the air pollution-mortality relationship in the context of ecologic confounders J Toxicol Environ Health A 66 16–19 1735 1777 12959842
Kannel WB Hjortland MC McNamara PM Gordon T 1976 Menopause and risk of cardiovascular disease: the Framingham study Ann Intern Med 85 4 447 452 970770
Karim R Buchanan TA Hodis HN Li Y Mack WJ 2005 The association of smoking and subclinical atherosclerosis in type 2 diabetes: modification by duration of diabetes Diabet Med 22 1 81 87 15606696
Krewski D Burnett RT Goldberg MS Hoover K Siemiatycki J Abrahamowicz M 2004 Validation of the Harvard Six Cities Study of particulate air pollution and mortality N Engl J Med 350 2 198 199 14711928
Künzli N 2002 The public health relevance of air pollution abatement Eur Respir J 20 1 198 209 12166570
Künzli N Medina S Kaiser R Quenel P Horak FJ Studnicka M 2001 Assessment of deaths attributable to air pollution: should we use risk estimates based on time series or on cohort studies? Am J Epidemiol 153 11 1050 1055 11390322
Künzli N Tager I 2000 Long-term health effects of particulate and other ambient air pollution: research can progress faster if we want it to Environ Health Perspect 108 915 918 11049809
Kwan MP Lee J 2004. Geovisualization of human activity patterns using 3D GIS: a time-geographic approach. In: Spatially Integrated Social Science: Examples in Best Practice (Goodchild MF, Janelle DG, eds). New York:Oxford University Press, 48–66.
Leech JA Nelson WC Burnett RT Aaron S Raizenne ME 2002 It’s about time: a comparison of Canadian and American time-activity patterns J Expo Anal Environ Epidemiol 12 6 427 432 12415491
Mack WJ Islam T Lee Z Selzer RH Hodis HN 2003 Environmental tobacco smoke and carotid arterial stiffness Prev Med 37 2 148 154 12855214
Mack WJ LaBree L Liu C Selzer RH Hodis HN 2000 Correlations between measures of atherosclerosis change using carotid ultrasonography and coronary angiography Atherosclerosis 150 2 371 379 10856529
O’Leary DH Polak JF Kronmal RA Manolio TA Burke GL Wolfson SK Jr 1999 Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study Collaborative Research Group N Engl J Med 340 1 14 22 9878640
O’Neill MS Jerrett M Kawachi I Levy JI Cohen AJ Gouveia N 2003 Health, wealth, and air pollution: advancing theory and methods Environ Health Perspect 111 1861 1870 14644658
Peters A Pope CA 2002 Cardiopulmonary mortality and air pollution Lancet 360 9341 1184 1185 12401239
Pope CA Burnett RT Thun MJ Calle EE Krewski D Ito K 2002 Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution JAMA 287 9 1132 1141 11879110
Pope CA Burnett RT Thurston GD Thun MJ Calle EE Krewski D 2004 Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease Circulation 109 1 71 77 14676145
Riediker M Williams R Devlin R Griggs T Bromberg P 2003 Exposure to particulate matter, volatile organic compounds, and other air pollutants inside patrol cars Environ Sci Technol 37 10 2084 2093 12785511
Roosli M Braun-Fahrlander C Künzli N Oglesby L Theis G Camenzind M 2000 Spatial variability of different fractions of particulate matter within an urban environment and between urban and rural sites J Air Waste Manag Assoc 50 7 1115 1124 10939205
Ross R 1999 Atherosclerosis—an inflammatory disease N Engl J Med 340 2 115 126 9887164
Samet JM Dominici F Curriero FC Coursac I Zeger SL 2000 Fine particulate air pollution and mortality in 20 U.S. cities, 1987–1994 N Engl J Med 343 24 1742 1749 11114312
Sarnat J Koutrakis P Suh H 2000 Assessing the relationship between personal particulate and gaseous exposures of senior citizens living in Baltimore, MD J Air Waste Manag Assoc 50 7 1184 1198 10939211
Sarnat JA Long CM Koutrakis P Coull BA Schwartz J Suh HH 2002 Using sulfur as a tracer of outdoor fine particulate matter Environ Sci Technol 36 24 5305 5314 12521154
Selzer RH Hodis HN Kwong-Fu H Mack WJ Lee PL Liu CR 1994 Evaluation of computerized edge tracking for quantifying intima-media thickness of the common carotid artery from B-mode ultrasound images Atherosclerosis 111 1 1 11 7840805
Selzer RH Mack WJ Lee PL Kwong-Fu H Hodis HN 2001 Improved common carotid elasticity and intima-media thickness measurements from computer analysis of sequential ultrasound frames Atherosclerosis 154 1 185 193 11137099
Sharman JE Coombes JS Geraghty DP Fraser DI 2002 Exposure to automotive pollution increases plasma susceptibility to oxidation Arch Environ Health 57 6 536 540 12696650
Stein JH Douglas PS Srinivasan SR Bond MG Tang R Li S 2004 Distribution and cross-sectional age-related increases of carotid artery intima-media thickness in young adults. The Bogalusa Heart Study Stroke 35 2782 2787 15514185
Suwa T Hogg JC Quinlan KB Ohgami A Vincent R van Eeden SF 2002 Particulate air pollution induces progression of atherosclerosis J Am Coll Cardiol 39 6 935 942 11897432
Thomas D Stram D Dwyer J 1993 Exposure measurement error: influence on exposure-disease relationships and methods of correction Annu Rev Publ Health 14 69 93
van Eeden SF Tan WC Suwa T Mukae H Terashima T Fujii T 2001 Cytokines involved in the systemic inflammatory response induced by exposure to particulate matter air pollutants (PM10 ) Am J Respir Crit Care Med 164 5 826 830 11549540
Wiegman A Hutten BA de Groot E Rodenburg J Bakker HD Buller HR 2004 Efficacy and safety of statin therapy in children with familial hypercholesterolemia: a randomized controlled trial JAMA 292 3 331 337 15265847
Wittekoek ME de Groot E Prins MH Trip MD Buller HR Kastelei JJ 1999 Differences in intima-media thickness in the carotid and femoral arteries in familial hypercholesterolemic heterozygotes with and without clinical manifestations of cardiovascular disease Atherosclerosis 146 2 271 279 10532683
Zanobetti A Schwartz J 2002 Cardiovascular damage by airborne particles: are diabetics more susceptible? Epidemiology 13 5 588 592 12192230
|
15687058
|
PMC1277865
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):201-206
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7523
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7200ehp0113-00020715687059Environmental MedicineGrand RoundsA Case of Bowen’s Disease and Small-Cell Lung Carcinoma: Long-Term Consequences of Chronic Arsenic Exposure in Chinese Traditional Medicine Lee Linda 1Bebb Gwyn 21University of Alberta, Faculty of Medicine, Edmonton, Alberta, Canada2Tom Baker Cancer Centre, Calgary, Alberta, CanadaAddress correspondence to G. Bebb, Tom Baker Cancer Centre, 1331, 29th St. NW, Calgary, Ab, T2N 4N2, Canada. E-mail:
[email protected] thank K. Smith for photographing the skin lesions and B. Sheehan and N. Murray for reviewing the manuscript.
The authors declare they have no competing financial interests.
2 2005 20 12 2004 113 2 207 210 22 4 2004 20 12 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Chronic arsenic toxicity occurs primarily through inadvertent ingestion of contaminated water and food or occupational exposure, but it can also occur through medicinal ingestion. This case features a 53-year-old lifetime nonsmoker with chronic asthma treated for 10 years in childhood with Chinese traditional medicine containing arsenic. The patient was diagnosed with Bowen’s disease and developed extensive-stage small-cell carcinoma of the lung 10 years and 47 years, respectively, after the onset of arsenic exposure. Although it has a long history as a medicinal agent, arsenic is a carcinogen associated with many malignancies including those of skin and lung. It is more commonly associated with non–small-cell lung cancer, but the temporal association with Bowen’s disease in the absence of other chemical or occupational exposure strongly points to a causal role for arsenic in this case of small-cell lung cancer. Individuals with documented arsenic-induced Bowen’s disease should be considered for more aggressive screening for long-term complications, especially the development of subsequent malignancies.
arsenicBowen’s diseasecase reportChinese traditional medicinechronic toxicitysmall-cell lung carcinoma
==== Body
A 53 year-old nonsmoking Chinese man with a long-standing diagnosis of asthma presented with a 1-week history of worsening shortness of breath, productive cough, and pleuritic chest pain. Chest X ray confirmed right-middle lobe consolidation suggestive of pneumonia. His symptoms resolved with a 5-day course of antibiotics, but a follow-up chest X ray demonstrated worsening of the consolidation. Bronchoscopy revealed a right middle-lobe tumor, and the biopsy was positive for small-cell lung carcinoma.
The patient was born in Hong Kong and immigrated to Canada around 1990 at 40 years of age. According to his medical history, at 6 years of age he was diagnosed with asthma, which was treated with Chinese traditional medicine (CTM) pills known to contain arsenic until he was 16 years of age, when he was diagnosed with Bowen’s disease. In 1993, a skin biopsy of a lesion on his right elbow was positive for squamous cell carcinoma in situ. He had never smoked cigarettes, and he drank alcohol only once a month. He had no other known exposures to arsenic through drinking water, diet, or occupation.
When seen by the medical oncologist, he complained of a cough that produced clear sputum and dull pain over the right costal margin, but he denied constitutional symptoms. A physical exam revealed an afebrile Asian male with no palpable lymphadenopathy. Chest examination was consistent with right middle lobe consolidation. The liver edge was palpable 3 cm below the costal margin but was neither nodular nor tender. Multiple 3–8 cm scaly, purple-brown, well-demarcated tear-drop–shaped patches were distributed over the skin of his trunk and extremities; these patches were consistent with his history of Bowen’s disease. His neurologic examination was normal.
Computed tomography of the patient’s chest and abdomen confirmed a right hilar mass completely obstructing the right middle-lobe bronchus with distal pneumonitis and large mediastinal nodes. Multiple hypodense liver lesions suggestive of metastatic disease were also found.
A diagnosis of extensive stage small-cell lung cancer was made. The patient entered a phase III clinical trial comparing cisplatin and etoposide to cisplatin and irinotecan and was randomized to the latter arm. Because more than 35 years had passed since the cessation of his known arsenic exposure, direct testing to confirm a diagnosis of arsenic exposure would probably have served little purpose. He achieved a complete remission but died of complications of progressive disease 10 months later.
Discussion
Arsenic, a naturally occurring metal, is best known as a poison and can generate both acute and chronic toxicity. Exposure can occur through air, water, soil, and food. Less well known is that arsenic has been used medicinally to treat a variety of illnesses since the times of Hippocrates and Galen. In the 18th and 19th centuries, Fowler’s solution (containing 1% arsenic trioxide) was prescribed to treat asthma, psoriasis, syphilis, and chronic myelogenous leukemia (Waxman and Anderson 2001). Although the association of inorganic arsenic intake and malignancy was first documented in 1887 (Hutchison 1887), Fowler’s solution continued to be popularized as a health tonic and was listed in the British Pharmaceutical and Therapeutic Products handbook as recently as 1958 (Ratnaike 2003).
With the advent of other pharmaceuticals and the discovery of the complications of chronic arsenic exposure, therapeutic arsenic use has declined. By the mid-1990s, the only indication for arsenic was in the treatment of trypanosomiasis (Waxman and Anderson 2001). However, intravenous arsenic trioxide has recently been shown to induce remission in patients with acute promyelocytic leukemia (Soignet et al. 1998), prompting a resurgence in its therapeutic use. In 2000, the U.S. Food and Drug Administration (FDA) approved the drug Trisenox (arsenic trioxide; Cell Therapeutics, Inc., Seattle, WA, USA) only 3 years after studies were initiated (FDA 2001).
In our patient, chronic arsenic exposure occurred by consuming Chinese herbal anti-asthmatic preparations in Hong Kong, most likely between 1956 and 1966. Although these medicines were widely available throughout Asia, their use was first described in Singapore by Tay and Seah (1975). They identified preparations containing high concentrations of inorganic arsenic ranging from 25 to 107,000 mg/L, with six of these preparations having been imported from Hong Kong.
Arsenic continues to be either a main constituent or a contaminant in many traditional and herbal medicines. An analysis of Asian traditional medicines by Garvey et al. (2001) revealed that 4 of the 54 (7.4%) sampled pills would result in a daily arsenic dosage of > 0.1 mg/day, whereas other pills contained significant quantities of mercury or lead. Containing daily dosages of arsenic that ranged from 0.140 to 16.1 mg, these pills were indicated for the treatment of asthma, headache, fever, and children’s ailments and to clear the kidneys and lungs (Garvey et al. 2001). Although manufactured in Southeast Asia, two of the pills were purchased in the United States (Garvey et al. 2001). In recent decades, there has been growing interest and availability of traditional Asian medicines. Currently, the FDA has not imposed any standard limits for arsenic in food or medicine except in animals treated with veterinary drugs [Agency for Toxic Substances and Disease Registry (ATSDR) 2004].
The toxicity of arsenic depends on its chemical state. Inorganic arsenic in its trivalent form is more toxic than pentavalent arsenic (Hughes 2002). By binding to thiol or sulfhydryl groups on proteins, As(III) can inactivate over 200 enzymes (Abernathy et al. 1999; Hughes 2002). This is the likely mechanism responsible for arsenic’s widespread effects on the liver, lungs, kidneys, spleen, gastrointestinal tract, and keratin-rich tissues. As(V) can replace phosphate, which is involved in many biochemical pathways resulting in the depletion of compounds such as adenosine-5′-triphosphate (ATP) (Hughes 2002). Increased levels of reactive oxidants in plasma (Wu et al. 2001) and markers of oxidative damage in arsenic-related skin conditions (Matsui et al. 1999) suggest that long-term damage from chronic arsenic exposure is mediated through the generation of reactive oxygen species.
Investigations of the effects of arsenic in animals have been problematic (Wang JP et al. 2002), and it is uncertain whether inorganic arsenic itself or resultant methylated metabolites that form in vivo are responsible for the carcinogenic effect. The few studies that have been successful, however, confirm arsenic’s carcinogenicity. Arsenic is likely a cocarcinogen that inhibits DNA repair and enhances the activity of other directly genotoxic agents (Andrew et al 2003; Beyermann 2002; Rossman et al. 2002). The cellular response to arsenic exposure seems to be concentration dependent. At high concentrations (> 50 μM), arsenic is able to induce an apoptotic response in vitro, a phenomenon probably exploited in its use to treat leukemia (Jimi et al 2004; Lunghi et al 2004). At lower concentrations (< 25 μM), evidence of genomic stress can be observed in the form of nuclear accumulation of p53, but apoptosis is not generally seen (Dong 2002). Environmental exposure to arsenic in a chronic low-dose manner likely leads to the gradual accumulation of genomic damage without apoptosis. Other observations suggest that the role of arsenic as a cocarcinogen may be mediated by inhibition of DNA repair and increased expression of cyclin D1 (Vogt and Rossman 2001). Differences in the expression pattern of p53 have also been attributed to the down-regulation of gene expression by alteration of promoter methylation status (Mass and Wang 1997). Arsenic has been shown to modulate cell signaling by inducing mitogen-activated protein kinases to change gene expression (Beyersmann 2002; Yang and Frenkel 2002).
Diagnosis of arsenic intoxication is often difficult because clinical presentation varies depending on route of exposure, chemical form, dose, and time elapsed since exposure. Furthermore, because arsenic affects multiple systems, poisoning can present with a wide variety of signs and symptoms. In acute arsenic poisoning, initial symptoms are gastrointestinal in nature due to the direct toxic effect of arsenic on intestinal epithelial cells. Clinical features include colicky abdominal pain, nausea, vomiting, bloody or rice-water diarrhea, and excessive salivation. Other manifestations include acute psychosis, cardiomyopathy, pulmonary edema, renal failure, skin rash, anemia, and encephalopathy (Ratnaike 2003). Quantitative studies can be performed on blood and urine in acute arsenic poisoning to confirm a suspected diagnosis. Because arsenic is cleared from blood within 10 hr (Hindmarch 2002), a urine arsenic level is usually more useful in cases of recent ingestion within 1–3 days (Buchet et al. 1981). Residual traces of arsenic in hair and nail samples may confirm arsenic exposure but can be subject to external contamination and cannot reliably date time of exposure (Hindmarch 2002). Presence of anemia, leukopenia, thrombocytopenia, or eosinophilia on complete blood count, basophilic stippling on the peripheral smear, or elevated liver transaminases is consistent with arsenic exposure but is not specific.
Like acute arsenic poisoning, the clinical features of chronic arsenic exposure are multi-systemic. Symptoms include malaise, weakness, decreased appetite, weight loss, and a sensory peripheral neuropathy that progresses to glove and stocking anesthesia (Ratnaike 2003). However, the hallmark of long-term arsenic exposure involves cutaneous changes such as hyperkeratosis, hyperpigmentation, Mee’s lines on nails, and malignant skin changes including Bowen’s disease, squamous-call carcinoma, and basal-cell carcinoma (Centano et al. 2002; Wong SS 1998). Although cutaneous changes develop slowly over time (up to 3–7 years for pigmentation changes and keratoses and up to 40 years for skin cancer), they may occur after lower doses than those causing neuropathy or anemia (ATSDR 2004). Studies on populations with chronic exposure to arsenic through drinking water show an association with increased cardiovascular disease (Tseng et al. 2003; Wang CH et al. 2002), peripheral vascular disease (Wang CH et al. 2002; Wang et al. 2003; Yu et al. 2002), cerebrovascular disease (Wang CH et al. 2002), respiratory disease (Milton and Rahman 2002), and diabetes (Wang et al. 2003). The most serious long-term consequence of arsenic exposure is increased risk for malignancy. Arsenic is now a recognized carcinogen associated with increased incidence of skin, lung, liver, bladder, and kidney malignancies (Chen et al. 1992).
A positive dose–response relationship between arsenic exposure and its chronic health effects has been observed. In a large population study in West Bengal, the prevalence of both skin lesions and hyperpigmention increased with the concentration of arsenic in drinking water (Mazumder et al. 1998). For levels of arsenic > 0.80 mg/L, the prevalence of skin lesions and hyperpigmentation in males was 10.7 and 22.7 per 100, respectively (Mazumder et al. 1998). A similar trend was noted in a Taiwanese study; for arsenic levels > 0.60 mg/L, the prevalence of skin cancer by 60 years of age was 92.0 in 1,000 (Tseng et al. 1968). In a recent review of the dose–response relationship between arsenic consumption through drinking water and its adverse health effects, Yoshida et al. (2004) noted that skin lesions are the most sensitive feature and often the earliest nonmalignant effect of arsenic exposure. Although a dose–response relationship between arsenic and certain malignancies including lung cancer was first identified in 1989 on the basis of a maximum 22-year latency period (Wu et al. 1989), more recent data have quantified that only arsenic exposure levels > 0.64 mg/L are associated with a significant increase in lung cancer mortality (Guo 2004).
Medicinal arsenic ingestion typically results in prolonged toxic exposure at doses higher than those present in contaminated water (Garvey et al. 2001; Tay and Seah 1975). Because both types of exposure involve trivalent arsenic and occur through the same mechanism of oral consumption followed by gastrointestinal absorption, it is possible that epidemiologic data from studies of drinking-water exposure may be applied to medicinal arsenic exposure. In 1998, case reports of three patients with chronic arsenic poisoning from CTMs in Singapore document that all three had cutaneous changes including basal-cell carcinoma and squamous-cell carcinoma, one had lung cancer, and one had liver cancer (Wong ST et al. 1998).
Another case review of 17 patients from Singapore selected for arsenic-induced cutaneous changes found that 15 were exposed through CTMs (Wong SS et al. 1998). All the patients had Bowen’s disease that developed after a long average latency period of 39 years. Since 1995, CTMs containing > 5 mg/L inorganic arsenic have been banned in Singapore (Wong SS et al. 1998).
Skin lesions generally precede the onset of internal malignancies. In a study of patients who took Fowler’s solution (containing 1% arsenic trioxide) during 1945–1969, approximately 50% had arsenic-related skin changes (Cuziek et al. 1982). A follow-up report 10 years later demonstrated excess bladder cancer mortality in the subgroup of patients with skin changes (Cuziek et al. 1992). In a study in Japan, Miki et al. (1982) reported that of 31 patients with Bowen’s disease and increased arsenic levels through drinking water, 10 had invasive skin cancers and 10 had internal malignancies, including 7 patients with pulmonary cancers. The authors hypothesized a timeline in which exposure to arsenic was followed by Bowen’s disease within 10 years, invasive skin cancers after 20 years, and pulmonary cancers after 30 years. Given the long period for the development of arsenic-induced malignancy, it may be too early to see cases arising from areas currently affected by contaminated drinking water such as West Bengal, Bangladesh, and China. In our patient, exposure likely started at approximately 6 of age, with his Bowen’s disease and pulmonary cancer diagnosed 10 and 47 years later, respectively.
Multiple studies have demonstrated that arsenic exposure is a documented risk factor for the development of lung carcinoma. This was best shown in a study in Nakajo, Japan (Nakadaira et al. 2002), in which some residents were exposed to well water with inorganic arsenic levels as high as 400 mg/L during 1954–1959. Of 454 inhabitants who underwent medical examinations in 1959,93 (20.5%) were diagnosed as having chronic arsenic poisoning on the basis of physical signs including cutaneous changes. Twenty-nine years after the exposure was terminated, exposed male patients demonstrated an excess mortality rate from lung cancer: the ratio of observed deaths to expected deaths from lung cancer was 7.0:0.64. Although arsenic exposure is more commonly associated with non–small-cell lung cancer, small-cell carcinoma incidence was also increased when compared with control groups, thereby supporting a causal relationship between arsenic and small-cell lung cancer. However, smoking was a confounding factor that was not addressed in the study design.
Bowen’s disease (squamous-cell carcinoma in situ), which can arise as a consequence of both arsenic and exposure to ultraviolet (UV) radation, seems a natural platform from which to study carcinogenic changes. A large population-based Danish cohort study confirmed that patients with Bowen’s disease have an excess risk of nonmelanomatous skin cancer and found a 2-fold increase in the risk of lung cancer in male patients with Bowen’s disease on sun-protected areas (Jaeger et al. 1999). The different mutagenic mechanisms associated with arsenic compared with those of other genotoxic agents have been reflected in mutational spectra of specific genes. Subtle differences have been noted in the mutation spectrum of UV-induced Bowen’s disease, in which point mutations are common, in contrast to arsenic-induced skin lesion, for which few p53 mutations were observed (Castren et al. 1998; Hsieh et al. 1994). Whether such observations can be translated from cutaneous lesions to bronchial neoplasms is unclear, but it seems likely that a different mutational spectrum may be seen in arsenic-induced versus smoking-induced small-cell lung cancers.
Small-cell lung cancer is an aggressive tumor that metastasizes early. Patients often present with extensive disease, which has a poor prognosis. It is extremely rare in young nonsmokers, so such a diagnosis should provoke a search for other risk factors. In the present case, the only risk factor was a remote 10-year history of arsenic ingestion through CTMs. Physical examination revealed the presence of cutaneous lesions that had been previously biopsied to show the presence of Bowen’s disease. Although peripheral neuropathy was absent, we felt that his cutaneous and lung neoplasms served to indirectly confirm his history of remote chronic arsenic exposure.
Conclusion
Chronic arsenic toxicity is a clinical diagnosis. It can be difficult to elicit a clear history of exposure either to contaminated food or well water, or through occupational exposure. However, it is a diagnosis that should be considered if there is a clear history of traditional or herbal medication use, particularly for the treatment of asthma, psoriasis, or syphilis. Moreover, chronic arsenic toxicity should be suspected in anyone presenting with cutaneous changes such as hyperkeratosis, hyperpigmentation, Mee’s lines on nails, or malignant skin changes such as Bowen’s disease with or without concomitant peripheral neuropathy.
Arsenic is a recognized etiologic factor in Bowen’s disease and a known risk factor for lung cancer. In our patient, it is likely that each condition developed independently following arsenic exposure, with skin pathology preceding lung cancer. Although a unifying pathophysiologic mechanism remains to be elucidated, patients with a history of arsenic exposure or ingestion of antiasthmatic CTMs require additional vigilance for signs of skin changes that may herald other malignancies. As chronic arsenic exposure through contaminated drinking water continues in many areas of the world, a large population may be at risk for latent malignancy, particularly if skin changes have already been noted. Because the role of chemopreventative approaches in these patients remains to be proven, such individuals should be considered candidates for chemoprevention trials.
==== Refs
References
Abernathy CO Liu YP Longfellow D Aposhian HV Beck B Fowler B 1999 Arsenic: health effects, mechanisms of actions, and research issues Environ Health Perspect 107 593 597 10379007
Andrew AS Karagas MR Hamilton JW 2003 Decreased DNA repair gene expression among individuals exposed to arsenic in United States drinking water Int J Cancer 104 263 268 12569548
ATSDR 2004. Clincal Evaluation, Arsenic Toxicity Case Study. Atlanta, GA:Agency for Toxic Substances and Disease Registry.
Beyersmann D 2002 Effects of carcinogenic metals on gene expression Toxicol Lett 127 63 68 10.1016/S0378-4274(01)00484-2.12052642
Buchet JP Lanwery R Roels H 1981 Comparison of the urinary excretion of arsenic metabolites after a single oral dose of sodium arsenite, monomethyl arsonate, or dimethyl arsenite in man Int Arch Occup Environ Health 48 71 79 6894292
Castren K Ranki A Welsh JA Vahakangas KH 1998 Infrequent p53 mutations in arsenic-related skin lesions Oncol Res 10 475 482 10223623
Centeno JA Mullick FG Martinez L Page NP Gibb H Longfellow D 2002 Pathology related to chronic arsenic exposure Environ Health Perspect 110 suppl 5 883 886 12426152
Chen CJ Chen CW Wu MM Kuo TL 1992 Cancer potential in liver, lung, bladder and kidney due to ingested inorganic arsenic in drinking water Br J Cancer 66 888 892 1419632
Cuzick J Evans S Gillman M Price Evans DA 1982 Medicinal arsenic and internal malignancies Br J Cancer 45 904 911 6212076
Cuzick J Sasieni P Evans S 1992 Ingested arsenic, keratoses, and bladder cancer Am J Epidemiol 136 417 421 1415161
Dong Z 2002 The molecular mechanisms of arsenic-induced cell transformation and apoptosis Environ Health Perspect 110 suppl 5 757 759 12426127
FDA (Food and Drug Administration) 2001. Making medical progress: a look at FDA approvals in 2000. FDA Consum 35(2):7–8. Available: http://www.fda.gov/fdac/features/2001/201_med.html [accessed 5 January 2005].
Garvey GJ Hahn G Lee RV Harbison RG 2001 Heavy metal hazards of Asian traditional remedies Int J Environ Health Res 11 63 71 11260788
Guo HR 2004 Arsenic level in drinking water and mortality of lung cancer Cancer Causes Control 15 171 177 10.1023/BCACO.0000019503.02851.b0.15017129
Hindmarch JT 2002 Caveats in hair analysis in chronic arsenic poisoning Clin Biochem 35 1 11 11937073
Hsieh LL Chen HJ Hsieh JT Jee SH Chen GS Chen CJ 1994 Arsenic-related Bowen’s disease and paraquat-related skin cancerous lesions show no detectable ras and p53 gene alterations Cancer Lett 86 59 65 7954356
Hughes MF 2002 Arsenic toxicity and potential mechanisms of action Toxicol Lett 133 1 16 12076506
Hutchison J 1887 Arsenic cancers Br Med J 2 1280 1281
Jaeger AB Gramkow A Hjalgrim H Melbye M Frisch M 1999 Bowen disease and risk of subsequent malignant neoplasms: a population-based cohort study of 1147 patients Arch Dermatol 135 790 793 10411153
Jimi S Uchiyama M Takaki A Suzumiya J Hara S 2004 Mechanisms of cell death induced by cadmium and arsenic Ann NY Acad Sci 1011 325 331 15126309
Lunghi P Tabilio A Lo-Coco F Pelicci P Bonati A 2004. Arsenic trioxide (ATO) and MEK1 inhibition synergize to induce apoptosis in acute promyelocytic leukemia cells. Leukemia; 10.1038/sj.leu.2403585 [11 November 2004].
Mass MJ Wang L 1997 Arsenic alters cytosine methylation patterns of the promoter of the tumor suppressor gene p53 in human lung cells: a model for a mechanism of carcinogenesis Mutat Res 386 263 277 9219564
Matsui M Nishigori C Toyokuni S Takada J Akaboshi M Ishikawa M 1999 The role of oxidative DNA damage in human arsenic carcinogenesis: detection of 8-hydroxy-2′-deoxyguanosine in arsenic-related Bowen’s disease J Invest Dermatol 113 26 31 10417614
Mazumder DN Das Gupta J Santra A Pal A Ghose A Sarkar S 1998 Chronic arsenic toxicity in West Bengal—the worst calamity in the world J Indian Med Assoc 96 4 7, 18 9601181
Miki Y Kawatsu T Matsuda K Machino H Kubo K 1982 Cutaneous and pulmonary cancers associated with Bowen’s disease J Am Acad Dermatol 6 26 31 6211469
Milton AH 2002 Respiratory effects and arsenic contaminated well water in Bangladesh Int J Environ Health Res 12 175 179 12400554
Nakadaira H Endoh K Katagiri M Yamamoto M 2002 Elevated mortality from lung cancer associated with arsenic exposure for a limited duration J Occup Environ Med 44 291 299 11911031
Ratnaike RN 2003 Acute and chronic arsenic toxicity Postgrad Med J 79 391 396 12897217
Rossman TG Uddin AN Burns FJ Bosland MC 2002 Arsenite cocarcinogenesis: an animal model derived from genetic toxicology studies Environ Health Perspect 110 suppl 5 749 752 12426125
Soignet SL Maslak P Wang ZG Jhanwar S Calleja E Dardashti LJ 1998 Complete remission after treatment of acute promyelocytic leukemia with arsenic trioxide N Engl J Med 339 1341 1348 9801394
Tay CH Seah CS 1975 Arsenic poisoning from anti-asthmatic herbal preparations Med J Aust 2 424 428 1186550
Tseng CH Chong CK Tseng CP Hsueh YM Chiou HY Tseng CC 2003 Long-term arsenic exposure and ischemic heart disease in arseniasis-hyperendemic villages in Taiwan Toxicol Lett 137 15 21 12505429
Tseng WP Chu HM How SW Fong JM Lin CS Yeh S 1968 Prevalence of skin cancer in an endemic area of chronic arsenicism in Taiwan J Natl Cancer Inst 40 453 463 5644201
Vogt BL Rossman TG 2001 Effects of arsenite on p53, p21 and cyclin D expression in normal human fibroblasts—a possible mechanism for arsenite’s comutagenicity Mutat Res 478 159 168 11406180
Wang CH Jeng JS Yip PK Chen CL Hsu LI Hsueh YM 2002 Biological gradient between long-term arsenic exposure and carotid atherosclerosis Circulation 105 1804 1809 11956123
Wang JP Qi L Moore MR Ng JC 2002 A review of animal models for the study of arsenic carcinogenesis Toxicol Lett 133 17 31 10.1016/S0378-4274(02)00086-3 [Online 14 May 2002].12076507
Wang SL Chiou JM Chen CJ Tseng CH Chou WL Wang CC 2003 Prevalence of non-insulin-dependent diabetes mellitus and related vascular diseases in southwestern arseniasis-endemic and nonendemic areas in Taiwan Environ Health Perspect 111 155 159 12573898
Waxman S Anderson KC 2001 History of the development of arsenic derivatives in cancer therapy Oncologist 6 suppl 2 3 10 11331434
Wong SS Tan KC Goh CL 1998 Cutaneous manifestations of chronic arsenicism: review of seventeen cases J Am Acad Dermatol 38 179 185 9486671
Wong ST Chan HL Teo SK 1998 The spectrum of cutaneous and internal malignancies in chronic arsenic toxicity Singapore Med J 39 171 173 9676149
Wu MM Chiou HY Wang TW Hsueh YM Wang IH Chen CJ 2001 Association of blood arsenic levels with increased reactive oxidants and decreased antioxidant capacity in a human population of northeastern Taiwan Environ Health Perspect 109 1011 1017 11675266
Wu MM Kuo TL Hwang YH Chen CJ 1989 Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases Am J Epidemiol 130 1123 1132 2589305
Yang C Frenkel K 2002 Arsenic-mediated cellular signal transduction, transcription factor activation, and aberrant gene expression: implications in carcinogenesis J Environ Pathol Toxicol Oncol 21 331 342 12510962
Yoshida T Yamauchi H Fan Sun G 2004 Chronic health effects in people exposed to arsenic via the drinking water: dose-response relationships in review Toxicol Appl Pharmacol 198 243 252 15276403
Yu HS Lee CH Chen GS 2002 Peripheral vascular diseases resulting from chronic arsenical poisoning J Dermatol 29 123 130 11990246
|
15687059
|
PMC1277866
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 20; 113(2):207-210
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7200
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.6984ehp0113-00021115687060Children's HealthArticlesChlorpyrifos Accumulation Patterns for Child-Accessible Surfaces and Objects and Urinary Metabolite Excretion by Children for 2 Weeks after Crack-and-Crevice Application Hore Paromita 1Robson Mark 1Freeman Natalie 1Zhang Jim 1Wartenberg Daniel 1Özkaynak Halûk 2Tulve Nicolle 2Sheldon Linda 2Needham Larry 3Barr Dana 3Lioy Paul J. 11Environmental and Occupational Health Sciences Institute, Exposure Measurement and Assessment Division, Rutgers University and the University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, Piscataway, New Jersey, USA2National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, USA3Contemporary Pesticide Laboratory, Centers for Disease Control and Prevention, Atlanta, Georgia, USAAddress correspondence to P.J. Lioy, 170 Freling-huysen Rd., EOHSI Floor 3, Piscataway, NJ 08854 USA. Telephone: (732) 445-0150. Fax: (732) 445-0116. E-mail:
[email protected] thank the following organizations for support with portions of the project: U.S. Environmental Protection Agency (EPA)/National Exposure Research Laboratory technical services contract; U.S. EPA University Partnership Agreement CR827033; National Institute of Environmental Health Sciences grants P30-ES05022 and ESO7148-17; Dow Agro Sciences Grant-in-Aid, Indiana to the Environmental and Occupational Health Sciences Institute.
The article has been subjected to U.S. EPA agency review and approved for publication.
The authors declare they have no competing financial interests.
2 2005 23 9 2004 113 2 211 219 27 1 2004 23 9 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. The Children’s Post-Pesticide Application Exposure Study (CPPAES) was conducted to look at the distribution of chlorpyrifos within a home environment for 2 weeks after a routine professional crack-and-crevice application and to determine the amount of the chlorpyrifos that is absorbed by a child living within the home. Ten residential homes with a 2- to 5-year-old child in each were selected for study, and the homes were treated with chlorpyrifos. Pesticide measurements were made from the indoor air, indoor surfaces, and plush toys. In addition, periodic morning urine samples were collected from each of the children throughout the 2-week period. We analyzed the urine samples for 3,5,6-trichloropyridinol, the primary urinary metabolite of chlorpyrifos, and used the results to estimate the children’s absorbed dose. Average chlorpyrifos levels in the indoor air and surfaces were 26 (pretreatment)/120 (posttreatment) ng/m3 and 0.48 (pretreatment)/2.8 (posttreatment) ng/cm2, respectively, reaching peak levels between days 0 and 2; subsequently, concentrations decreased throughout the 2-week period. Chlorpyrifos in/on the plush toys ranged from 7.3 to 1,949 ng/toy postapplication, with concentrations increasing throughout the 2-week period, demonstrating a cumulative adsorption/absorption process indoors. The daily amount of chlorpyrifos estimated to be absorbed by the CPPAES children postapplication ranged from 0.04 to 4.8 μg/kg/day. During the 2 weeks after the crack-and-crevice application, there was no significant increase in the amount of chlorpyrifos absorbed by the CPPAES children.
biomarkerchildchildrenchlorpyrifoscrack-and-creviceindoor chemical usepesticide
==== Body
Eighty percent of U.S. households use pesticides more than once a year in and around their homes (Davis et al. 1992; Whitmore et al. 1994). Many of the pesticides applied indoors are semivolatile, with vapor pressures ranging from 10−2 to 10−8 mm Hg (Dalaker and Naifeh 1997). Once applied indoors, semivolatile pesticides can vaporize from treated surfaces and can distribute in and on targeted and nontargeted surfaces and objects (Byrne et al. 1998; Gurunathan et al. 1998; Lewis et al. 2001; Wright et al. 1984). This raises concern about exposures because U.S. householders, including children, can spend up to 90% of their time indoors within or around treated areas (Savage et al. 1981). Children in pesticide-treated homes may be exposed to pesticides via multiple routes and from multiple media. Given their inherent biologic vulnerabilities and characteristic behaviors that are different from those of adults, children can be particularly susceptible to the effects of pesticides (Aprea et al. 2000; Bearer 1995; Freeman et al. 1997; Guzelian et al. 1992; Reed et al. 1999).
In 1996, the Food Quality Protection Act (FQPA) mandated that contributions from all routes of exposure and from all possible sources be considered when setting food tolerance levels for pesticides, paying particular attention to address the potential risks to infants and small children (FQPA 1996). Several studies have used direct and indirect measures to try to estimate the total pesticide uptake by children via the inhalation, dermal, and nondietary ingestion routes after an indoor pesticide application (Byrne et al. 1998; Gurunathan et al. 1998; Lewis et al. 2001). Pesticide body burden levels estimated from environmental concentrations have been reported after either broadcast (Gurunathan et al. 1998) or homeowner/professional crack-and-crevice applications (Byrne et al. 1998; Lewis et al. 2001). No studies thus far have serially collected biomarker samples from children residing within treated homes to allow a comparison between body burden estimated from environmental data and body burden estimated from biomarker levels. Given that information regarding pesticide uptake by children in treated homes is needed to assess the health risks for exposed children, the lack of information on the time course of body burden levels after professional indoor application is a gap in the currently available research.
A detailed multimedia/multipathway 10-home residential study, referred to as the Children’s Post-Pesticide Application Exposure Study (CPPAES), was conducted to provide information on the release and movement of chlorpyrifos, a semivolatile pesticide (vapor pressure, 1.87 × 10−5 mm Hg at 20°C), within a residential environment and within children living in this environment over time after an application. The scientific approach involved collecting environmental samples from a treated home coupled with biomarker samples from a child living in the treated home, for 2 weeks after a routine crack-and-crevice application of chlorpyrifos. CPPAES was designed to evaluate the extent of aggregate chlorpyrifos exposure for children living within treated homes. The general concept for this study was outlined during a workshop held by the International Life Sciences Institute (ILSI 2000). The study was carried out between 1999 and 2001, before the U.S. Environmental Protection Agency (EPA) phased out indoor residential use of organophosphate pesticide.
Materials and Methods
Study design.
Ten residential homes (identified as H1–H10) were selected for CPPAES based on the criteria that they applied pesticides on a routine basis and had a child between 2 and 5 years of age who spent most of his or her time indoors at home. Each of the CPPAES homes was located in urban areas within New Jersey. The homes varied in size (34–96 m2) and style. For the protection of human subjects, the study design was thoroughly evaluated and approved by the institutional review board committee of the University of Medicine and Dentistry of New Jersey and the U.S. EPA.
Pesticide application.
The commercial product Dursban 2.E. or Dursban L.O. containing the insecticide chlorpyrifos [O,O-diethyl-O-(2-isopropyl-6-methyl-4-pyrimidinyl) phosphorothioate, CAS No. 2921-88-2] was applied to each of the CPPAES homes reportedly as a 0.25–0.5% water emulsion. Until recently and throughout the study period, chlorpyrifos was one of the most commonly used household insecticides within the United States used by homeowners, renters, and professional applicators to control cockroaches, fleas, and termites (U.S. EPA 2000). A licensed pesticide applicator applied the pesticide solution to each of the homes via a crack-and-crevice mode of application. The applications were made using a hand-pump compressed air sprayer (tank capacity, 1 gallon) with a pin stream nozzle, spraying with a downward-directed nozzle tip 12–16 inches from the floor. Applications were made to the cracks and crevices of the homes and in some cases along the perimeters of the walls behind appliances or furniture. The applications lasted approximately 15 min per home as the applicator examined each home for cracks and crevices and evidence of roach trails. Approximately 60–700 mL of the chlorpyrifos solution was reportedly sprayed in each CPPAES home. A sample of the pesticide solution applied within each home (except H1) was collected from the pesticide applicator, and the samples were analyzed in the laboratory. The amount of pesticide applied in each home was then based on the estimated volume of the pesticide solution applied. Although the study was designed to make uniform applications in each home, the analytical results indicated that the amount of chlorpyrifos applied within homes H8–H10 (4.1 × 10−7 to 4.3 × 10−6 g) was considerably lower than what was applied in homes H2–H7 (0.07–0.6 g). A sample of the pesticide application solution was not available for H1; however, based on the chlorpyrifos levels measured in the indoor air postapplication, the applied amount in H1 was probably similar to amounts applied in homes H2–H7.
In homes H3, H5, and H8, the pesticide was applied in all rooms. For homes H1, H2, and H4, the pesticide was applied in all the rooms except the bathrooms. For homes H6, H7, and H9, it was not applied in the parents’ bedrooms; for H10, pesticide was not applied in two of the bedrooms. During the crack-and-crevice application, the study participants left the treated homes and no sampling was conducted. After the application, re-entry did not occur for 3 hr. An exception was H10, where during this time the participants restricted their movements to the untreated portions of the house rather than vacating the home. The windows in all of the homes were “cracked” open during this 3-hr period.
Sampling scheme.
A 2-week multimedia sampling effort was carried out before and after an indoor crack-and-crevice application. Environmental samples were collected over time for measuring chlorpyrifos in the indoor environment. Simultaneously, biomarker samples were obtained from the participating children living within the treated homes. Preapplication measurements were made from the CPPAES homes on the day before the day of pesticide application. A crack-and-crevice pesticide application was then made to each of the CPPAES homes on what is designated day 0. Postapplication measurements were made on days 1, 2, 3, 5, 7, 9, and 11 after the day of application. The sampling scheme is presented in Table 1.
Samples collected.
In each CPPAES home, measurements were taken in two rooms that had been treated with the pesticide: either in the child’s main play area, designated “A,” and/or in or near the child’s bedroom area, “B.”
Time-weighted average measurements for chlorpyrifos vapor and aerosol were obtained in chlorpyrifos-treated rooms (H1–H9 samples collected in area A; H10 samples collected in area B). The indoor air samples were collected using a low-flow pump with a PM10 (particulate matter ≤10 μg) inlet and a carbon-impregnated filter. Collection and extraction methods for the air samples were developed by Gurunathan et al. (1998). The sampling time per sample spanned the time interval between each visit (i.e., days –1 to 0, 0–1, 1–2, 2–3, 3–5, 5–7, 7–9, and 9–11). Postsampling, the filters were extracted in 10 mL toluene via sonication and concentrated down to a sample volume of 5 mL. Mean recoveries for chlorpyrifos from laboratory controls were 101% [coefficient of variation (CV), 7.8%].
Surface wipe samples were collected within treated rooms on days –1, 1, 2, 3, 5, 7, 9, and/or 11 from nontargeted surfaces (i.e., areas not directly sprayed with the pesticide). The Lioy-Weisel-Wainman (LWW) sampling method as described by Gurunathan et al. (1998), and the method of Lioy et al. (2000) was used to collect the pesticide wipe samples by the movement of a C18-impregnated Teflon filter (moistened with isopropanol) within a 100 cm2 template.
The LWW sampler was used to collect wipe samples from smooth surfaces in both areas A and B. The wipe samples obtained from area A were collected from floor surfaces. All of the wipe samples obtained from area B, except for H1, H2, and H8 (day –1), were also collected from floor surfaces. Two-week area B samples collected from homes H1 and H2 were obtained from a dresser, 0.1–0.8 m above the floor. The H8 day –1 area B sample was collected from a windowsill, 0.6 m above the floor. Because of limited resources, except for in homes H2, H9, and H10, samples from area B were not collected on days 5 and 9. No LWW samples were collected from area B in H8 as the floor was carpeted. Each LWW wipe sample was collected from a different location within each home to prevent the surface activation previously noted by Gurunathan et al. (1998). However, whenever possible, the sampled areas were adjacent to the previous samples. Postsampling, the LWW filters were extracted in 5 mL of isopropanol via sonication. Mean recoveries for chlorpyrifos from laboratory controls were 106% (CV, 4.9%). The data were used to estimate the amount of chlorpyrifos distributed on open surfaces in the treated home environment.
Chlorpyrifos measurements were also made on samples collected from indicator toys placed within CPPAES treated rooms (H1–H3, H5–H9 samples collected in area A; H4 and H10 samples collected in area B). Toys were placed indoors immediately after the 3-hr re-entry period, and each was sequentially removed for chlorpyrifos analysis on days 1, 2, 3, 5, 7, 9, and 11 postapplication. A duplicate toy was collected on days 2, 5, and 9 postapplication from homes H3–H10. For H2, a duplicate sample was available only for day 2. For H5, a duplicate toy was collected on days 2, 5, 9, and 11 postapplication. “Sweet Stuffs” toys from the “The First Years Collection” purchased at Toys R Us (Watchung, NJ) (surface areas ~ 125–150 cm2) were used as the indicator toys. They were placed in a birdcage to limit the children’s interactions with the toys but at the same time not sheltering the toys from the movement of pesticide in the air. Plush toys were used because they are a potential sink for pesticides accessible to children within residential homes. Moreover, plush toys can serve as a surrogate for any sorbant medium present indoors with polyfoam filler, such as furniture upholstery and bedding.
A combination of toy surface wipes and toy extractions were analyzed from the duplicate toys to evaluate both dislodgeable and total components of the pesticide in/on the duplicate toys. A surface wipe of the duplicate toys was collected before the full extraction of the toys. The surface of each plush toy was wiped using isopropanol-impregnated swabs. The swab wipes were then extracted in 10 mL of isopropanol via sonication and concentrated down to a sample volume of 5 mL. Mean recoveries for chlorpyrifos from laboratory controls were 102% (CV, 3.2%). The plush toys were then extracted in 200 mL of hexane via sonication and concentrated down to a sample volume of 5 mL (Gurunathan et al. 1998). Mean recoveries for chlorpyrifos from laboratory controls were 96% (CV, 6.2%).
To estimate chlorpyrifos bioaccumulation in the CPPAES children, urine samples were collected and analyzed for 3,5,6-trichloropyridinol (TCPy), the primary urinary metabolite for chlorpyrifos (Nolan et al. 1984). First-morning-void urine samples were collected from the CPPAES children on each of the sampling days −1 (preapplication), 1, 2, 3, 5, 7, 9, and 11. These urine samples were designed to represent the contact of the children with chlorpyrifos on days −1, 0, 1, 2, 4, 6, 8, and 10, respectively, and estimate body burden, although there would be higher uncertainty in these values because it was a first void and not a 24-hr average (Wessels et al. 2003). The preapplication urine sample was collected as a baseline urine measure for TCPy concentration. Only 10% of the urine samples collected were not the first morning voids, and only two children missed more than one morning void (H5 and H9). The urine samples were analyzed by the Centers for Disease Control and Prevention. The samples were analyzed for TCPy using a slightly modified version of the method described by Hill et al. (1995) using a 3-hr derivatization process. The analytical limit of detection (LOD) for TCPy concentration using this method was 1.0 μg/L for a 4-mL sample. Results for both creatinine (CR) adjusted (micrograms TCPy/grams CR) and non-CR-adjusted (micrograms TCPy/liters urine) TCPy concentrations are reported in this study. The TCPy levels for 5 of the 80 urine samples were reported as less than the analytical LOD. We assumed a value of 0.5 × LOD for these samples, which is a generally accepted method of reporting data below the LOD (U.S. EPA 1999).
Translating the non-CR-adjusted morning void TCPy concentrations (micrograms TCPy/liters urine) to estimated daily TCPy excretion (micrograms per kilogram per day) required an assumption of 0.5 L/day daily urinary excretion rate for children between 0 and 4 years of age (Lentner 1981). The CR-adjusted first-morning-void TCPy concentrations (micrograms TCPy/grams CR) required an assumption of 25 mg CR/kg/day excretion rate (Hay et al. 1997) to estimate daily excretion of TCPy (micrograms per kilogram per day). However, there are uncertainties associated with both estimates. Based on the daily TCPy excreted amounts, the daily estimated amounts of chlorpyrifos absorbed by each of the CPPAES children via all routes were calculated using the approach of Byrne et al. (1998):
Chemical analysis.
We used a capillary gas chromatograph (Hewlett-Packard 5860 Series II; all equipment from Hewlett-Packard, Wilmington, DE) equipped with an HP Nickel 63 electron capture detector and an Autosampler Injector 7673 for chlorpyrifos analysis of the air, surface wipe, and toy samples. We used HP Chem Station chromatography software to quantify the concentration of chlorpyrifos in all of the samples. A split/splitless injector was maintained at 250°C. The detector temperature was held at 325°C. A 60-m (0.25 mm inner diameter DB-1) fused silica capillary column with 0.25 μm film thickness (J&W Scientific, Folsom, CA) was used. Under splitless conditions, the column was temperature programmed from 50°C to 190°C at 30°C/min (held for 28 min), and from 190°C to 270 °C at 70°C/min, and held at 270°C for 16 min, altogether resulting in a run time of approximately 50 min per sample run. Helium was used as the carrier gas (flow rate, 1.0 mL/min). Nitrogen was used as the makeup gas (flow rate, 65 mL/min). An injection volume of 1 μL was maintained for all of the samples.
Instrument quality assurance and quality control.
Standard solutions for chlorpyrifos ranged from 0.0012 to 2.4 μg/mL. These were analyzed with every gas chromatograph run, and calibration curves were generated for the concentration range of interest. The results were used to generate a linear regression equation (r
2 = 0.99). Replicates of independent standard solutions (prepared by Chem Service, Linden, NJ) were included with each sample run to evaluate the performance of the gas chromatograph. Pesticide recoveries from the independent standards (n = 10) were within 2% of the reported values, with CVs < 2.1%. All solvent blanks remained free of chlorpyrifos. Where no peaks were detected, the sample results were reported as nondetects (ND). The instrument LOD for chlorpyrifos was 0.0011 μg/mL.
Statistical analyses.
CPPAES was designed specifically to study the mechanisms of release and exposure to semivolatile pesticides over a 2-week period postapplication. Thus, the emphasis of the study was on the time course of accumulation and elimination of a pesticide in a variety of media in the same home. Thus, it was not a population-based study. Because three of the homes received approximately five orders of magnitude lower amounts of chlorpyrifos, the CPPAES homes were divided into two groups based on application rate: “high” (H1–H7) and “low” (H8–H10). We used the Wilcoxon signed ranks test to compare the pretreatment and the 2-week posttreatment chlorpyrifos levels as measured from the indoor environment (air, dust, plush toys) and from the children (chlorpyrifos absorbed dose) within these groups. Given the mechanistic design of the study, there was a small sample size, and a nonparametric analysis method was employed to examine between group data. Using the nonparametric Mann-Whitney U-test, we compared the extent of daily average postapplication chlorpyrifos levels between the “high” and the “low” homes. This type of study was previously recommended as part of a modeling workshop (ILSI 2000).
Results
Indoor air samples.
Based on estimated chlorpyrifos application rates for homes H1–H7 (> 4.3 × 10−6 g) and homes H8–H10 (4.1 × 10−7 to 4.3 × 10−6 g) (Table 2), air concentrations of chlorpyrifos were categorized into two groups designated “high” and the “low” homes, respectively. Box plots of the indoor air chlorpyrifos concentrations measured throughout the 2-week period are presented in Figure 1.
Surface wipes.
The LWW wipe sample results obtained from the samples collected in the main play areas (A) and bedroom areas (B) are found in Tables 3 and 4, respectively. Box plots of the chlorpyrifos surface loadings for the 2-week period are presented in Figures 2 and 3, respectively.
Chlorpyrifos levels in the main play areas of the “high” homes (H3–H6) were considerably greater than the levels measured in homes H8–H10. For days 0–10, the average ranged from 3.1 to 6.9 ng/cm2 (H3–H6) and from 0.17 to 1.7 ng/cm2 (H8–H10). Despite the lower chlorpyrifos application rates in H8–H10, chlorpyrifos levels were detected; in fact, H10 chlorpyrifos levels (days 0–10 mean = 1.7 ng/cm2) were higher than levels measured in homes H1, H2, and H7 (days 0–10, mean range = 0.4–1.0 ng/cm2). The preapplication level in H10 (1.0 ng/cm2) suggested another source contributed to H10 chlorpyrifos levels. A potential source could be previous pesticide applications made within the home. LWW area A chlorpyrifos surface loadings in H3–H6 peaked between days 1 and 2 postapplication (mean = 13 ng/cm2), and the levels were significantly greater than the preapplication levels (< 0.9 ng/cm2; p = 0.006). After the peak day, the loading gradually declined approaching pretreatment levels by day 11 (mean = 0.9 ng/cm2). Surface loadings in homes H8 and H10 did not follow the same decay pattern. The highest loading postapplication for H8 (1.6 ng/cm2) was observed on day 7, and for H10 (2.1 ng/cm2) on day 3. Postapplication surface loadings in homes H8–H10 (mean = 0.93 ng/cm2) were only slightly greater than preapplication levels (mean = 0.44 ng/cm2). Levels reached or approached pretreatment levels on day 11 (mean = 0.55 ng/cm2).
Chlorpyrifos levels measured in the bedroom areas were generally lower than levels measured in the main play areas for H1–H7, excluding H5. In fact, except for H3 and H5, the highest postapplication surface loadings measured in the bedroom areas were only slightly greater than the preapplication levels (range = 0.18–0.82 ng/cm2; pretreatment = 0.28–0.70 ng/cm2). The highest LWW area A and area B surface loadings were measured in H5, with loadings peaking on day 1 postapplication (range = 21.2–23.8 ng/cm2; pretreatment levels were ND). After the peak day, loadings in H5 gradually declined approaching pretreatment levels on day 11.
Plush toys.
Chlorpyrifos levels found in/on the plush toys are presented in Table 5 and illustrated in Figure 4. Chlorpyrifos concentrations in/on the plush toys increased throughout the 2-week sampling period for all homes. A similar trend was observed by Gurunathan et al. (1998), after a broadcast application of chlorpyrifos. On day 1, the plush toy chlorpyrifos concentrations for CPPAES homes H1–H10 averaged 197 ng/toy, reaching 634 ng/toy on day 11. Overall, levels measured within homes H1–H7 were significantly higher than levels in homes H8–H10 (p = 0.000). Measured chlorpyrifos levels were the highest in H5 throughout the 2-week period. Less than 5% (mean ± SD = 1.6 ± 2.0%; n = 26) of the chlorpyrifos was wiped off the plush toys (mean ± SD = 3.4 ± 2.6 ng). These amounts were significantly less than the amounts of chlorpyrifos obtained from the toys after full extraction (mean ± SD = 519 ± 606 ng; p = 0.000).
Biomonitoring.
We estimated chlorpyrifos levels absorbed by the CPPAES children by quantifying the amount of chlorpyrifos metabolite TCPy that was excreted by the children on the sampled days. The amount of TCPy excreted by the CPPAES children and the corresponding absorbed doses derived from both the non-CR-adjusted and the CR-adjusted TCPy results are presented in Table 6 and illustrated in Figures 5 and 6, respectively. However, CR is at lower levels in children, and there is probably a higher level of variability due to the lack of a 24-hr sample. The CPPAES children excreted on average approximately 0.25 μg TCPy/kg/day (non-CR adjusted; n = 10) or 0.34 μg TCPy/kg/day (CR adjusted; n = 10) preapplication. The estimated average chlorpyrifos absorbed doses were 0.55 μg chlorpyrifos/kg/day (non-CR adjusted) and 0.85 μg chlorpyrifos/kg/day (CR adjusted). The amount of TCPy excreted by the children postapplication on average per day ranged from 0.21 to 0.28 μg TCPy/kg/day (non-CR adjusted) and from 0.31 to 0.51 μg TCPy/kg/day (CR adjusted). The corresponding daily average postapplication chlorpyrifos absorbed doses ranged from 0.53 to 0.7 μg chlorpyrifos/kg/day (non-CR adjusted) and from 0.77 to 1.3 μg chlorpyrifos/kg/day (CR adjusted). A significant increase was not observed in the amount of chlorpyrifos absorbed by the CPPAES children during the 2-week period after the crack-and-crevice application.
Discussion
CPPAES combined extensive multimedia monitoring efforts within residential homes for a 2-week period after a crack-and-crevice application of chlorpyrifos with simultaneous biomonitoring of the children residing within the treated homes. Biomonitoring of the chlorpyrifos metabolite enabled us to quantify the extent of aggregate exposure to the pesticide for a child living within a treated residence and estimate the body burden levels. Although previous studies have examined the time-series distribution of chlorpyrifos within an indoor environment, no studies thus far have concurrently measured the time-series urine levels from children that lived within the pesticide-treated homes and spent most of their time indoors. Moreover, because three of the homes (H8–H10) received approximately five orders of magnitude lower amounts of the chlorpyrifos, the reduced level of application gave us an opportunity to investigate the distribution of the pesticide within the home and the children after different application rates.
Some of the findings from CPPAES were in agreement with other studies that have demonstrated that semivolatile pesticides applied indoors within a home can contaminate the indoor air (Byrne et al. 1998; Gurunathan et al. 1998; Lewis et al. 2001; Wright and Leidy 1978, 1980; Wright et al. 1981) and nontargeted indoor surfaces (Gurunathan et al. 1998; Wright 1976; Wright and Jackson 1975).
Chlorpyrifos applied inside the 10 CPPAES homes was detected within the treated room indoor air throughout the 2-week postapplication period. Mostly, higher pesticide levels were detected from the CPPAES homes that received a greater application rate (except H7). For homes H1–H6, 2-week postapplication indoor air levels ranged from 22 to 816 ng/m3; H8–H10 levels ranged from 2.2 to 31 ng/m3. Comparatively, overall CPPAES concentrations in the indoor air were either similar or considerably lower than some of the reported studies. For instance, a study conducted by Wright and Leidy (1978) measured chlorpyrifos concentrations in the air within vacant rooms after a crack-and-crevice application of 0.5 or 1% chlorpyrifos solution. Pesticide measurements were made from the indoor air throughout a 3-day period after a crack-and-crevice application. Chlorpyrifos levels in the indoor air as measured immediately after the indoor application ranged from 600 to 2,700 ng/m3. A more recent study was conducted by Byrne et al. (1998) to estimate chlorpyrifos levels within pesticide-treated homes for a 10-day period after a crack-and-crevice application made with a 0.5% pesticide solution. The study was conducted in three residential homes. An estimated 3.3–3.9 g of chlorpyrifos was applied to each of the homes. Preapplication indoor air levels from the CPPAES homes were more or less comparable with measurements collected by Byrne et al. (1998) from two of the three treated homes (< 20 ng/m3). The highest indoor air chlorpyrifos level measured postapplication in the CPPAES study (816 ng/m3), however, was lower than the maximum concentration (2,300 ng/m3) observed by Byrne et al. (1998).
As a measure of the extent of nontarget deposition of the chlorpyrifos within the CPPAES homes after the crack-and-crevice application, postapplication surface loading measurements were made from nontreated surfaces within the treated homes. The highest postapplication chlorpyrifos loadings, as measured via wipe sampling from nontargeted surfaces within the CPPAES children’s main play areas and main living areas, were observed within homes H1–H7 (range = 0.03–24.6 ng/cm2). However, not all of the measured postapplication loadings from homes H1–H7 were higher than the corresponding levels from homes H8–H10 (range = 0.08–3 ng/cm2). Higher measured loadings in the children’s main play areas were not always accompanied with higher loadings in the children’s main living areas (except H5). Factors such as cleaning of the homes and tracking in or out of home soil/dust most likely contributed to the 2-week distribution of the indoor measured surface loadings. The levels observed on the indoor surfaces in the CPPAES were similar but somewhat higher than levels observed in the Minnesota Children’s Pesticide/National Human Exposure Assessment Survey study (median = 0.34 and 0.42 ng/cm2 for two different rooms in each home; maximum = 3.64 and 14.4 ng/cm2 for the same rooms) (Lioy et al. 2000). The latter were obtained in homes that used pesticides such as chlorpyrifos but were not necessarily measured immediately postapplication.
Other studies have reported similar or lower indoor levels of chlorpyrifos after crack-and-crevice treatments. In a study conducted by Wright and Jackson (1975), chlorpyrifos measurements were made from nontargeted surfaces (aluminum pie plates) placed within vacant dormitory rooms for an 8-day period after indoor crack-and-crevice pesticide applications with either 0.5 or 1% chlorpyrifos solutions. Chlorpyrifos deposition levels measured from the 0.5 or 1% treated areas ranged from 0.4 to 3.5 ng/cm2 and 0.4 to 11.3 ng/cm2, respectively, with overall pesticide levels decreasing throughout the 8-day period. The higher measured nontargeted surface loadings found in the present study, compared with levels measured in the reported studies with a greater application rate, may have resulted from a number of reasons. For instance, although the intention of this study was to sample from nontargeted surfaces, some of the nontargeted surfaces may have accidentally been applied with chlorpyrifos. Some of the variability observed in the surface concentrations may have resulted from the different sampling techniques that were used between the studies. Moreover, less activity within the treated rooms, such as walking or children playing, particularly in the dormitory study conducted by Wright and Jackson (1975), may have contributed to lower pesticide loadings on the nontargeted surfaces because of less redistribution and resuspension of the indoor dust.
In this study, we also examined pesticide levels on nontreated surfaces such as plush toys because children living within pesticide-treated homes may come into contact with contaminated objects, such as toys, within a home environment (Gurunathan et al. 1998). Moreover, similar sorbant surfaces such as furniture upholstery can also contain pesticides that children residing within treated homes can be exposed to. Chlorpyrifos concentrations measured from the plush toys that were placed within homes H1–H7 were significantly greater than levels measured from toys placed within homes H8–H10. H1–H7 chlorpyrifos levels ranged from 87 to 1,949 ng/toy; H8–H10 levels ranged from 7 to 221 ng/toy. Chlorpyrifos concentrations in/on the CPPAES plush toys increased throughout the 2-week sampling period, demonstrating a cumulative trend.
An increase in chlorpyrifos levels within the CPPAES homes provided an opportunity for increased exposure postapplication. However, although an increase was observed in the amount of chlorpyrifos measured from the CPPAES homes after the crack-and-crevice application, a significant increase was not observed in the amount of chlorpyrifos absorbed by the CPPAES children during the 2-week period after the crack-and-crevice application (Figures 5 and 6). Moreover, even though chlorpyrifos levels as measured from the various media within the indoor environment were considerably greater in the “high” homes compared with the “low” homes (indoor air ~ 10-fold; indoor surfaces ~ 4-fold; plush toys ~ 8-fold), postapplication daily absorbed chlorpyrifos doses measured from the “high” home children were only slightly greater (~ 2-fold) than levels measured from the “low” home children, essentially indicating that the children in fact were not coming into contact with all of the chlorpyrifos within the indoor environment, and the body burden levels could have been due to multiple sources, a point previously described by Krieger et al. (2003). The children’s activities may in fact have played an important role in determining how much pesticide each child actually absorbed. Total absorbed doses of chlorpyrifos as estimated for the children residing within the CPPAES treated homes (< 4.8 μg/kg/day) were within a factor of 2.5 of the chlorpyrifos doses estimated by Byrne et al. (1998) (< 2.1 μg/kg/day). The potential absorbed doses for children residing within three chlorpyrifos-treated homes were calculated by Byrne et al. (1998) using environmental data gathered after a crack-and-crevice application. The estimated body burden levels, however, could not be compared with the environmental results because body burden levels were not measured for children by Byrne et al. (1998).
Most (~ 97%) of the postapplication CPPAES children’s estimated absorbed doses (range = 0.02–4.8 μg/kg/day) were lower than the U.S. EPA oral reference dose (RfD) value of 3 μg/kg/day [based upon a no observed effect level of 30 μg/kg/day; calculated without the additional 10× safety factor added by FQPA (1996) to protect young children]. However, most (88%) of the 10 CPPAES children’s estimated absorbed doses exceeded the revised RfD value of 0.3 μg/kg/day (calculated including the additional 10× safety factor) by up to 1,600%. The EPA in their final risk assessment for chlorpyrifos had considered a safety factor of 3 as opposed to a more conservative FQPA safety factor of 10, which reduced the number of estimated exceedances. Only 29% of the CPPAES children’s estimated absorbed doses exceeded the RfD of 1 μg/kg/day (calculated using the safety factor of 3).
Comparison of results from CPPAES and Gurunathan et al. (1998) suggests that selection of the application method will greatly influence the children’s exposures and dose received from pesticides applied indoors. In particular, comparison of the results of these two studies has indicated that estimated pesticide body burden levels for children living within homes after a broadcast application of a semivolatile pesticide were considerably greater than the measured body burden levels for the children living within crack-and-crevice–treated homes. For instance, the total absorbed doses of chlorpyrifos for children residing within the crack-and-crevice–treated homes were considerably lower than even the nondietary estimated doses from Gurunathan et al. (1998) (208–356 μg/kg/day). One possible reason may have been that the pesticide levels in the indoor environment on child-accessible objects after the broadcast application by Gurunathan et al. (1998) were considerably greater than the levels measured in the crack-and-crevice studies. This was not unexpected because, compared with the Gurunathan et al. (1998) broadcast application, the CPPAES crack-and-crevice application method required a smaller volume of the pesticide applied. For instance, whereas approximately 296–473 mL of chlorpyrifos formulation containing ~ 0.07–1.8 g of chlorpyrifos was applied to the CPPAES homes, approximately 2,000 mL of a chlorpyrifos formulation yielding 12 g of chlorpyrifos was applied to surfaces in each Gurunathan et al. (1998) apartment. The highest indoor air chlorpyrifos level measured postapplication in the study by Gurunathan et al. (1998) was 7,000 ng/m3. Whereas cumulative pesticide concentrations measured from the CPPAES toys were < 1,949 ng/toy, plush toy cumulative pesticide concentrations measured by Gurunathan et al. (1998) reached levels > 30,000 ng/toy. Consequently, residents of a crack-and-crevice–treated home would be potentially exposed to lower amounts of the pesticide.
Some data gaps introduce uncertainties during interpretation of CPPAES postapplication urinary TCPy results. For instance, there is limited information available on the natural variability in background urinary TCPy levels, a point that needs to be kept in mind because the values are relatively low. Moreover, both CR-adjusted and non-CR-adjusted TCPy data have inherent limitations. For example, no studies have systematically evaluated the validity of using CR adjustment for children. Moreover, the accuracy of TCPy values derived from samples with CR levels ≤30 mg/dL urine is questioned by Lauwerys and Hoet (1993) as being too dilute to provide valid results.
Conclusions
CPPAES results indicate that when chlorpyrifos is applied properly via a crack-and-crevice mode of application, the application does not lead to a significant increase in the children’s chlorpyrifos body burden levels. Although an increase was observed in the amount of chlorpyrifos measured from the CPPAES homes after the pesticide application, CPPAES findings indicated that the children living within the crack-and-crevice–treated homes were actually not coming into contact with most of the chlorpyrifos that was present in the indoor environment. Thus, pesticide body burden levels estimated for children living within crack-and-crevice–treated homes, which are considerably lower than levels estimated for children living within homes treated via broadcast application (Gurunathan et al. 1998), had other sources. Essentially, adjusting the mode of application so as to spray the pesticide around pest-infested targeted areas rather than an entire surface area of a house greatly reduced the amount of pesticide that children living within treated homes would potentially be exposed to and uptake after an application.
Figure 1 Box plots for chlorpyrifos concentrations in indoor air (ng/m3) for (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10). Note that the y-axis on each plot is not the same.
Figure 2 Box plots for chlorpyrifos surface loadings (main play areas, LWWA) (ng/cm2) for (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10). Note that the y-axis on each plot is not the same.
Figure 3 Box plots for chlorpyrifos surface loadings (bedroom areas, LWWB) (ng/cm2) for (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10). Note that the y-axis on each plot is not the same.
Figure 4 Box plots for chlorpyrifos concentrations within reference plush toys (ng/toy) for (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10). Note that the y-axis on each plot is not the same.
Figure 5 Box plots for daily TCPy excreted amounts measured from the CPPAES children postapplication (H1–H7 vs. H8–H10) (μg TCPy/kg/day). Non-CR-adjusted (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10); CR-adjusted (C) “high” homes (H1–H7) and (D) “low” homes (H8–H10).
Figure 6 Box plots for daily chlorpyrifos absorbed doses calculated for the CPPAES children postapplication (H1–H7 vs. H8–H10) (μg chlorpyrifos/kg/day). Non-CR-adjusted (A) “high” homes (H1–H7) and (B) “low” homes (H8–H10); CR-adjusted (C) “high” homes (H1–H7) and (D) “low” homes (H8–H10).
Table 1 Sample collection scheme for the CPPAES homes.
Preapplication Day of application nth day postapplication
Day −1 0 1 2 3 4 5 7 9 11
Table 2 Indoor air measurements for chlorpyrifos within treated rooms (ng/m3).
Days
Home identification −1–0 0–1 1–2 2–3 3–5 5–7 7–9 9–11 Average (days 0–10)
“High” homes
H1 3.4 179 195 178 132 123 87 73 138
H2 10 121 130 71 29 31 39 22 63
H3 7.2 338 207 153 155 107 73 69 157
H4 58 312 203 164 145 158 102 122 172
H5 14 816 648 709 587 386 294 299 534
H6 115 196 44 55 41 45 46 50 68
H7 18 32 14 45 5.5 4.0 4.4 6.3 16
Average 32 285 206 196 156 122 92 92 —
Median 14 196 195 153 132 107 73 69 —
SD 41 257 210 233 199 129 95 99 —
“Low” homes
H8 3.8 4.5 2.8 4.0 4.8 3.6 3.7 2.2 3.7
H9 12 18 21 18 20 21 19 19 19
H10 24 24 25 23 29 28 28 31 27
Average 13 15 16 15 18 17 17 17 —
Median 12 18 21 18 20 21 19 19 —
SD 10 9.9 12 9.9 12 13 12 14 —
Table 3 Surface loading measurements for chlorpyrifos from nontargeted surfaces within treated rooms (main play areas, LWWA) (ng/cm2).
Day
Home identification −1 1 2 3 5 7 9 11 Average (days 0–10)
“High” homes
H1 ND 1.89 1.03 1.02 NA 0.71 NA 0.49 1.03
H2 0.10 0.49 0.60 0.59 0.24 0.18 0.29 0.33 0.39
H3 ND 2.55 6.04 4.39 1.96 2.81 2.69 1.36 3.11
H4 ND 24.6 10.9 4.48 3.40 1.46 0.75 0.61 6.60
H5 ND 21.2 10.1 7.93 5.26 1.71 1.33 0.83 6.90
H6 0.85 16.5 9.6 7.7 6.6 2.2 0.82 0.83 6.32
H7 0.57 0.46 0.42 0.45 0.46 0.25 0.67 0.67 0.48
Average 0.51 9.7 5.5 3.8 3.0 1.3 1.1 0.73 —
Median 0.57 2.6 6.0 4.4 2.7 1.5 0.79 0.67 —
SD 0.38 10.7 4.8 3.2 2.6 1.0 0.85 0.33 —
“Low” homes
H8 0.21 0.24 0.41 0.91 1.3 1.6 1.2 0.81 0.91
H9 0.12 0.25 0.23 0.19 0.24 0.10 0.09 0.08 0.17
H10 1.0 1.8 2.1 2.1 1.6 1.5 2.0 0.8 1.70
Average 0.44 0.75 0.91 1.08 1.05 1.04 1.09 0.55 —
Median 0.21 0.25 0.41 0.91 1.26 1.46 1.18 0.78 —
SD 0.49 0.88 1.00 0.99 0.72 0.82 0.96 0.41 —
NA, not available.
Table 4 Surface loading measurements for chlorpyrifos from nontargeted surfaces within treated rooms (bedroom areas, LWWB) (ng/cm2).
Day
Home identification −1 1 2 3 5 7 9 11 Average (days 0–10)
“High” homes
H1 ND 0.18 0.10 0.12 NA 0.18 NA 0.16 0.15
H2 0.63 NA 0.27 0.18 0.10 0.05 0.07 0.16 0.14
H3 ND 2.7 4.7 3.1 NA 1.1 NA 0.03 2.3
H4 0.28 0.29 0.47 0.26 NA 0.20 NA 0.30 0.30
H5 ND 23.8 21.8 23.0 NA 6.6 NA 3.1 15.7
H6 0.49 0.82 0.41 0.40 NA 0.34 NA 0.23 0.44
H7 0.70 0.49 0.41 0.38 NA 0.15 NA 0.27 0.34
Average 0.53 4.7 4.0 3.9 0.10 1.20 0.07 0.61 —
Median 0.56 0.66 0.41 0.38 0.10 0.20 0.07 0.23 —
SD 0.19 9.4 8.0 8.5 — 2.4 — 1.1 —
“Low” homes
H8 0.21 NA NA NA NA NA NA NA —
H9 0.23 0.27 0.17 0.26 0.11 0.10 0.26 0.28 0.21
H10 1.6 1.3 2.3 1.5 3.0 1.9 2.8 1.8 2.1
Average 0.67 0.76 1.24 0.86 1.54 1.02 1.50 1.03 —
Median 0.23 0.76 1.24 0.86 1.54 1.02 1.50 1.03 —
SD 0.78 0.70 1.5 0.84 2.0 1.3 1.76 1.1 —
NA, not available.
Table 5 Chlorpyrifos levels in/on reference plush toys placed within treated rooms (ng/toy).
Day
Home identification 1 2 3 5 7 9 11
“High” homes
H1 329 761 911 957 578 665 721
H2 189 278 342 343 362 427 442
H3 344 445 672 625 824 746 753
H4 150 247 300 420 374 374 962
H5 481 926 1,615 1,495 1,275 1,480 1,949
H6 302 328 457 384 434 566 588
H7 87 145 221 284 293 437 552
Average 269 447 646 644 592 671 852
Median 302 328 457 420 434 566 721
SD 135 289 490 440 350 382 512
“Low” homes
H8 7.3 10 11 13 13 18 22
H9 45 62 81 96 130 139 134
H10 35 76 87 144 156 157 221
Average 29 50 60 84 100 105 126
Median 35 62 81 96 130 139 134
SD 19 35 42 66 76 75 100
Table 6 Amount of TCPy excreted in urine calculated for the CPPAES children (μg/kg/day).
Home Identification
Day H1 H2 H3 H4 H5 H6 H7 H8 H9 H10
Non-CR-adjusted
−1 0.22 NAa 0.11 0.02a 0.34b 0.64 0.28 0.14 0.11 0.14a
(0.53) (NA)a (0.28) (0.04)a (0.83)b (1.6) (0.68) (0.35) (0.27) (0.35)a
1 0.14 0.29 0.05 0.07 0.24a,b 0.46 0.66 0.21 0.20a 0.33
(0.35) (0.71) (0.14) (0.18) (0.60)a,b (1.1) (1.6) (0.53) (0.49)a (0.81)
2 0.22 0.44 0.08 0.02a,c 0.26 0.39 0.25 0.15 0.30 0.15
(0.53) (1.1) (0.21) (0.05)a,c (0.64) (0.97) (0.63) (0.37) (0.73) (0.37)
3 0.22 0.37 0.09 0.19 0.29b 0.71 0.32a 0.14 0.28 0.24
(0.53) (0.91) (0.22) (0.48) (0.71)b (1.8) (0.78)a (0.35) (0.69) (0.58)
5 0.24 0.34 0.05 0.27 0.27 0.64 0.086a 0.12 0.35 0.02c
(0.58) (0.84) (0.12) (0.67) (0.68) (1.6) (0.21)a (0.30) (0.86) (0.04)c
7 0.39 0.28 0.04a 0.06a 0.21 0.39 0.22 0.17 0.29 0.08
(0.97) (0.70) (0.10)a (0.15)a (0.52) (0.97) (0.56) (0.42) (0.71) (0.21)
9 0.20 0.25 0.01c 0.24 0.43 0.50 0.28 0.33 0.13b 0.10
(0.49) (0.62) (0.02)c (0.58) (1.1) (1.2) (0.69) (0.82) (0.33)b (0.25)
11 0.15a 0.48 0.14 0.24 0.34 0.46 0.086a 0.18a 0.08b 0.02c
(0.38)a (1.2) (0.34) (0.58) (0.83) (1.1) (0.21)a (0.44)a (0.19)b (0.04)c
CR-adjusted
−1 0.30 NAa 0.09 NAa 0.68b 0.70 0.18 0.21 0.16 0.43a
(0.74) (NA)a (0.22) (NA)a (1.7)b (1.7) (0.44) (0.51) (0.40) (1.1)a
1 0.38 0.35 0.08 0.10 1.9a,b 0.73 0.26 0.23 0.70a 0.39
(0.93) (0.87) (0.19) (0.25) (4.8)a,b (1.8) (0.64) (0.56) (1.7)a (0.96)
2 0.30 0.20 0.09 NAa 0.50 0.83 0.21 0.18 0.53 0.12
(0.74) (0.50) (0.22) (NA)a (1.2) (2.0) (0.52) (0.45) (1.3) (0.30)
3 0.23 0.38 0.10 0.30 0.40b 0.90 1.9a 0.16 0.27 0.42
(0.56) (0.93) (0.25) (0.74) (1.0)b (2.2) (4.6)a (0.40) (0.67) (1.0)
5 0.45 0.40 0.12 0.38 0.43 0.68 0.35a 0.11 0.48 NA
(1.1) (0.99) (0.29) (0.93) (1.1) (1.7) (0.86)a (0.27) (1.2) (NA)
7 0.38 0.21 NA 0.21a 0.50 0.53 0.33 0.32 0.22 0.12
(0.93) (0.51) (NA) (0.51)a (1.2) (1.3) (0.82) (0.79) (0.54) (0.29)
9 0.63 0.35 NA 0.22 0.30 0.50 0.33 0.37 0.13b 0.17
(1.5) (0.87) (NA) (0.53) (0.74) (1.2) (0.82) (0.91) (0.31)b (0.42)
11 0.60a 0.35 0.17 0.38 0.28 0.75 0.32a 0.82a 0.17b NA
(1.5)a (0.87) (0.43) (0.93) (0.68) (1.9) (0.79)a (2.0)a (0.42)b (NA)
NA, not available. Chlorpyrifos-absorbed doses within parentheses.
a Sample dilute: urine samples with CR levels < 30 mg/dL urine (Lauwerys and Hoet 1993).
b Not morning void urine sample.
c Analyte (TCPy) concentrations were < 1 μg/L (LOD for a 4-mL sample). For these a value of 0.5 × LOD (i.e., 0.5 μg/L) was assumed. Daily total urine volume excretion was assumed to be 0.5 L (Lentner 1981); CPPAES children’s body weights H1–H10 = 25, 14, 25, 14, 16, 14, 14, 18, 15, and 14 kg, respectively. Daily CR excretion rate was assumed to be 25 mg of CR/kg/day (average of the 20–30 mg of CR/day excretion rate for children suggested by Hay et al. (1997). Chlorpyrifos absorbed doses were calculated using the equation presented in “Materials and Methods.”
==== Refs
References
Aprea C Strambi M Novelli MT Lunghini L Bozzi N 2000 Biologic monitoring of exposure to organophosphorus pesticides in 195 Italian children Environ Health Perspect 108 521 525 10856025
Bearer CF 1995 Environmental health hazards: how children are different from adults Future Child Crit Issues Child Youths 5 2 11 26
Byrne SL Shurdut BA Saunders DG 1998 Potential chlorpyrifos exposure to residents following standard crack-and-crevice treatment Environ Health Perspect 106 725 731 9799188
Dalaker J Naifeh M 1997. Poverty in the United States. U.S. Bureau of the Census, Current Population Reports. Series P60-201. Washington, DC:U.S. Government Printing Office.
Davis JR Brownson RC Garcia R 1992 Family pesticide use in the home, garden, orchard, and yard Arch Environ Contam Toxicol 22 260 266 1616309
Food Quality Protection Act of 1996 (FQPA) 1996. Public Law 104–170.
Freeman NCG Ettinger A Barry M Rhoads G 1997 Hygiene and food related behaviors associated with blood lead levels of young children from lead contaminated homes J Expo Assess Environ Epidemiol 7 1 15
Gurunathan S Robson M Freeman N Buckley B Roy A Meyer R 1998 Accumulation of chlorpyrifos on residential surfaces and toys accessible to children Environ Health Perspect 106 9 16 9417768
Guzelian PS Henry CJ Olin SS 1992. Similarities and Differences between Children and Adults: Implications for Risk Assessment. Washington, DC:ILSI Press.
Hay WW JrGroothuis JR Hayward AR Levin MJ eds. 1997. Current Pediatric Diagnosis and Treatment. 13th ed. Stamford, CT:Appleton and Lange.
Hill RH Jr Shealy DB Head SL Williams CC Bailey SL Gregg M 1995 Determination of pesticide metabolites in human urine using an isotope dilution technique and tandem mass spectrometry J Anal Toxicol 19 5 323 329 7500620
ILSI 2000. Aggregate Exposure Assessment: Model Evaluation and Refinement Workshop Report. An ILSI Health and Environment Sciences Institute Workshop Report. Washington, DC:International Life Sciences Institute.
Krieger RI Dinoff TM Williams RL Zhang X Ross JH Aston LS 2003 Preformed biomarkers in produce inflate human organophosphate exposure assessments [Letter] Environ Health Perspect 111 A688 A691 14527854
Lauwerys RR Hoet P 1993. Industrial Chemical Exposure: Guidelines for Biological Monitoring. Boca Raton, FL:Lewis Publishers.
Lentner C 1981. Geigy Scientific Tables. 8th ed., vol. 1. West Caldwell, NJ:Medical Education Division, Ciba Geigy Corporation.
Lewis RG Fortune CR Camann DE 2001 Movement and deposition of two organophosphorus pesticides within a residence after interior and exterior applications Air Waste Manag Assoc 51 339 351
Lioy PJ Edwards RD Freeman NCG Gurunathan S Pellizzari E Adgate JL 2000 Housedust levels of selected insecticides and a herbicide measured by the EL and LWW samplers and comparisons to hand rinses and urine metabolites J Expo Anal Environ Epidemiol 10 327 340 10981727
National Academy of Sciences 1982. An Assessment of the Health Risks for Seven Pesticides Used for Termite Control. Washington, DC:National Academy Press.
Nolan RJ Rick DL Freshour NL Saunders JH 1984 Chlorpyrifos: pharmacokinetics in human volunteers Toxicol Appl Pharmacol 73 8 15 6200956
Reed K Jimenez M Freeman NCG Lioy PJ 1999 Quantification of children’s hand and mouthing activities J Expo Anal Environ Epidemiol 9 513 520 10554153
Savage EP Keefe TJ Wheeler HW Mounce L Helwig L Applehans F 1981 Household pesticide usage in the United States Arch Environ Health 36 304 309 7316568
U.S. EPA 1999. Classification of Food Forms with Respect to Level of Blending. HED Standard Operating Procedure 99.6. Washington, DC:Office of Pesticide Programs Health Effects Division, U.S. Environmental Protection Agency.
U.S. EPA 2000. Chlorpyrifos Revised Risk Assessment and Risk Mitigation Measures. Washington, DC:Office of Pesticide Programs, U.S. Environmental Protection Agency. Available: http://www.epa.gov/oppsrrd1/op/chlorpyrifos/comerqs.htm#1 [accessed 20 May 2004].
Wessels D Barr DB Mendola P 2003 Use of biomarkers to indicate exposure of children to organophosphate pesticides: implications for a longitudinal study of children’s environmental health Environ Health Perspect 111 1939 1946 14644670
Whitmore RW Immerman FW Camarin DE Bond AE Lewis RG Schaum JL 1994 Non-occupational exposures to pesticides for residents of two U.S. cities Arch Environ Contam Toxicol 26 47 59 8110023
Wright CG 1976 Insecticide movement following application to crevices in rooms Arch Environ Contam Toxicol 4 4 492 500 63269
Wright CG Jackson MD 1975 Insecticide residues in non-target areas of rooms after two methods of crack-and-crevice application Bull Environ Contam Toxicol 13 123 128 48390
Wright CG Leidy RB 1978 Chlorpyrifos residues in air after application to crevices in rooms Bull Environ Contam Toxicol 19 314 344
Wright CG Leidy RB 1980 Insecticide residues in the air of buildings and pest control vehicles Bull Environ Contam Toxicol 24 582 589 6155163
Wright CG Leidy RB Dupree HE 1981 Insecticides in the ambient air of rooms following their application for control of pests Bull Environ Contam Toxicol 26 548 553 7236915
Wright CG Leidy RB Dupree HE 1984 Chlorpyrifos and diazinon detection on surfaces in dormitory rooms Bull Environ Contam Toxicol 32 259 264 6201215
|
15687060
|
PMC1277867
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 23; 113(2):211-219
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.6984
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7382ehp0113-00022015687061Children's HealthArticlesOrganochlorine Pesticides and Male Genital Anomalies in the Child Health and Development Studies Bhatia Rajiv 1Shiau Rita 1Petreas Myrto 2Weintraub June M. 1Farhang Lili 13Eskenazi Brenda 41San Francisco Department of Public Health, San Francisco, California, USA2Hazardous Materials Laboratory, Department of Toxic Substances Control, California Environmental Protection Agency, Berkeley, California, USA3Public Health Institute, Berkeley, California, USA4Center for Children’s Environmental Health Research, School of Public Health, University of California at Berkeley, Berkeley, California, USAAddress correspondence to J.M. Weintraub, San Francisco Department of Public Health, 1390 Market St., Suite 910, San Francisco, CA 94102 USA. Telephone: (415) 252-3800. Fax: (415) 252-3964. E-mail:
[email protected] thank B. van den Berg and B. Cohn for making the Child Health and Development Studies (CHDS) specimens available for this study and R. Christianson for sharing her wealth of knowledge about the CHDS database.
This study was funded by the National Institute of Environmental Health Sciences (R29 ES09042).
The authors declare they have no competing financial interests.
2 2005 4 11 2004 113 2 220 224 2 7 2004 3 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Increasing rates of cryptorchidism and hypospadias in human populations may be caused by exogenous environmental agents. We conducted a case–control study of serum levels of p,p′-dichlorodiphenyltrichloroethane (DDT) and its major metabolite, p,p′-dichlorodiphenyldichloroethylene (DDE), and cryptorchidism and hypospadias in the Child Health and Development Study, a longitudinal cohort of pregnancies that occurred between 1959 and 1967, a period when DDT was produced and used in the United States. Serum was available from the mothers of 75 male children born with cryptorchidism, 66 with hypospadias, and 4 with both conditions. We randomly selected 283 controls from the cohort of women whose male babies were born without either of these conditions. Overall, we observed no statistically significant relationships or trends between outcomes and serum measures. After adjusting for maternal race, triglyceride level, and cholesterol level, compared with boys whose mothers had serum DDE levels < 27.0 ng/mL, boys whose mothers had serum DDE levels ≥61.0 ng/mL had odds ratios of 1.34 [95% confidence interval (CI), 0.51–3.48] for cryptorchidism and 1.18 (95% CI, 0.46–3.02) for hypospadias. For DDT, compared with boys whose mothers had serum DDT levels < 10.0 ng/mL, boys whose mothers had serum DDT levels ≥20.0 ng/mL had adjusted odds ratios of 1.01 (95% CI, 0.44–2.28) for cryptorchidism and 0.79 (95% CI, 0.33–1.89) for hypospadias. This study does not support an association of DDT or DDE and hypospadias or cryptorchidism.
cryptorchidismDDEDDThypospadiasinsecticidesmale genital anomaliesorganochlorinepregnancy
==== Body
Hypospadias (an abnormal opening of the urethra) and cryptorchidism (a failure of one or both testicles to descend) are two relatively common male genital congenital anomalies, occurring in 35 of 10,000 and 40 of 10,000 births, respectively (Paulozzi 1999). Recent evidence has suggested that hypospadias and cryptorchidism have increased in frequency in some populations (Paulozzi 1999). This is of public health significance, particularly because of the well-known association between cryptorchidism and testicular cancer (Wilson and Foster 1985; Toppari et al. 1996), which also appears to be on the increase (Toppari et al. 1996).
In mammals, cryptorchidism depends on mullerian-inhibiting hormone, androgens, and intra-abdominal pressure (Wilson and Foster 1985). The formation of the male external genitalia in utero also occurs under the influence of androgens (Baskin et al. 2001). There is suggestion that increasing rates of hypospadias and cryptorchidism may be caused partly by early exposure to endocrine-disrupting chemicals in the environment (Beard et al. 1984; Berkowitz et al. 1995; Depue 1984; Hjertkvist et al. 1989; Jackson 1988; Landrigan et al. 2003; McBride et al. 1991; Sharpe and Skakkebaek 1993; Sweet et al. 1974; Swerdlow et al. 1983; Toppari et al. 1996). This hypothesis is supported by the higher rates of urogenital anomalies, including cryptorchidism, in studies of sons exposed in utero to diethylstilbestrol, a potent estrogen (Cosgrove et al. 1977; Gill et al. 1979).
Although the use of one potent environmental endocrine disruptor, dichlorodiphenyl-trichloroethane (DDT), has been banned in the United States since 1972, concerns about its potential health effects continue because of its worldwide use in the eradication of malaria, its accumulation in the human food chain, and its long half-life in human tissues due to its lipophilicity (Clarkson 1995). DDT functions as an exogenous estrogen (Bulger and Kupfer 1983) and, in animal studies, has been shown to suppress Leydig cell development, alter secretion of mullerian-inhibiting substance by the Sertoli cells, and cause negative feedback inhibition at the fetal pituitary gland (Sharpe and Skakkebaek 1993). p,p′-Dichlorodiphenyl-dichloroethylene (DDE), the persistent metabolite of p,p′-dichlorodiphenyltrichloroethane (DDT), acts as an androgen receptor antagonist at target tissues and has been shown to inhibit the action of testosterone (Danzo 1997; Kelce et al. 1995). DDT or DDE may alter sex hormone metabolism, reducing available testosterone to tissues (Guillette et al. 1995). Adverse reproductive system effects associated with in utero DDT or DDE exposure in male animals include abnormal development of ovarian tissue (Fry and Toone 1981), reduced penis size (Guillette and Guillette 1996), reduced testosterone levels (Subramanian et al. 1987), reduced male rat anogenital distance (Gray et al. 2001), hypospadias (Gray et al. 2001), cryptorchidism (Facemire et al. 1995; Gray et al. 2001), impaired reproductive capacity (Bowerman et al. 1995), low sperm density (Facemire et al. 1995), and abnormal sperm (Facemire et al. 1995).
Few epidemiologic studies provide information on the relationship between male reproductive disorders and organochlorine insecticides in humans. García-Rodríguez et al. (1996) found higher rates of orchidopexy (a surgical procedure to correct cryptorchidism) in districts in Spain with more intensive farming. However, Longnecker et al. (2002) found no clear evidence of an effect of DDE on hypospadias or cryptorchidism among subjects in the Collaborative Perinatal Project. We report on an analysis of serum levels of DDT and DDE in a nested case–control study of cryptorchidism and hypospadias in a longitudinal cohort of pregnancies that occurred between 1959 and 1967, a period when DDT was produced and used in the United States but before the beginning of the rise in prevalence of hypospadias in the United States (Baskin et al. 2001).
Materials and Methods
The Child Health and Development Studies (CHDS) is a longitudinal cohort study of 20,754 pregnancies among women enrolled in the San Francisco Bay Area Kaiser Foundation Health Plan between 1959 and 1967 (van den Berg et al. 1988). Nearly 92% of all eligible pregnancies invited to participate were included in the cohort. Subscribers to this prepaid health plan represented an economically and ethnically diverse urban population.
Pregnant women were interviewed shortly after recruitment and later during their pregnancies. Maternal and pediatric medical records were abstracted throughout the follow-up period. Abstracted records contained information from every child visit until at least 5 years of age, including physician’s diagnosis and treatment, test results, and anthropometric measures (Christianson et al. 1981). Children averaged at least three visits per year from birth until 5 years of age. The study’s rate of attrition was extremely low; the CHDS observed 89.4% of live-born children until 5 years of age.
The ascertainment of congenital anomalies occurred through routine medical procedures and practices and appropriate specialized tests and referrals. The CHDS itself did not implement any additional routine or special tests to ascertain congenital anomalies. At least two physicians and one biostatistician reviewed abstracted information to ensure uniform recording. Congenital anomalies were classified based on the International Classification of Diseases, 7th Revision [World Health Organization (WHO) 1957]. An anomaly was coded as definite only if a physician considered the diagnosis certain or if confirmed by surgery or laboratory tests. Abstracted data were subsequently computer coded (Christianson et al. 1981).
Blood samples were obtained from the pregnant women at the time of enrollment, during each subsequent trimester, and immediately after delivery. At least one sample was obtained for 89% of pregnancies. Serum samples were subsequently divided into a package of four 2-mL vials and stored at −20°C at the National Institutes of Health (Bethesda, MD).
For this nested case–control study, we used a subset of males who were followed by the CHDS for at least 2 years and for whom computer records showed that at least one serum sample had been collected and stored. We restricted the subset to this minimum length of follow-up because cryptorchidism was coded by CHDS as an anomaly only if it persisted until 2 years of age.
Among the 9,345 males followed until 2 years of age, 101 had cryptorchidism, 73 had hypospadias, and 6 had both. We were able to obtain serum from the mothers of 75 subjects with cryptorchidism, 66 subjects with hypospadias, and 4 subjects with both conditions. We randomly selected 283 controls from the remaining male singleton births who were followed to 2 years of age, did not have hypospadias or cryptorchidism, and had one recorded serum sample. There were no matching criteria for controls.
Laboratory assays.
If available, the serum conservators provided the last pregnancy serum of the mother (n = 86); otherwise, samples from the postpartum period were provided (n = 334). Given the long half-life of DDT and DDE and the high correlation among DDE levels measured at different times during gestation (Longnecker et al. 1999), these serum samples should accurately reflect body burdens over the entire pregnancy. National Cancer Institute staff in Frederick, Maryland, retrieved the requested archived serum, placed a 1.5-mL aliquot of the sample into a separate vial, assigned a study identification number, and shipped the samples overnight on dry ice to the Hazardous Materials Laboratory of the State of California in Berkeley, California, where they were stored at below −20°C until laboratory analysis. Standard quality assurance procedures included rigorous calibration procedures, traceability of all standards, and internal review and audit. Method (reagent) blanks and laboratory controls were performed on either 10% of the samples or at least one with every batch of samples, whichever was greater. Internal standard recovery was performed on every chemical group on every sample. The laboratory staff were blind to the identity of the samples and to case/control status.
Laboratory personnel performed analyses in batches of 12 samples. Each batch consisted of nine study subject samples, one method blank, one laboratory control (fortified bovine serum), and a standard reference material [SRM 1589a, a human serum from the National Institute of Standards and Technology (Gaithersburg, MD)]. Batches included a consistent proportion of cases and controls. Unbeknown to laboratory staff, some batches included samples of pooled CHDS serum in place of a study subject sample to assess performance and to also facilitate future interlaboratory standardization.
Analytical methods are described in detail elsewhere (Petreas et al. 2003). Briefly, serum was thawed, and 1 mL was pipetted into a 15-mL test tube. Internal standards [polychlorinated biphenyl (PCB) congeners 14, 65, and 166 and tetrachloromethyl xylene (TCMX; AccuStandard, Inc., New Haven, CT)] were added before denaturing the proteins with 1 mL of acetic acid (Fisher Scientific, Pittsburgh, PA). Solvents employed were nanograde isooctane (Mallinckrodt, Paris, KY), trace environmental analysis grade hexane (99.9%), methanol (99.9%), dichloromethane (99.9%), pesticide residue grade acetone (99.9%), and toluene (99.9%) (Burdick & Jackson, Muskegon, MI). The analytes in the serum were then extracted with hexane:dichloromethane (90:10, vol:vol), and the extract was passed through a glass column filled with Florisil. The analytes were eluted with hexane followed by hexane: dichloromethane (1:1, vol:vol). The eluates were combined and concentrated, and recovery standards (pentachloronitrobenzene, PCB-30, PCB-204, and PCB-209) were added. We used six-level calibration curves with concentrations encompassing expected ranges for each analyte. Analysis was performed by gas chromatography/electron capture detection (Hewlett Packard 6890; Agilent Technologies, Palo Alto, CA) equipped with 60-m DB-XLB (Agilent Technologies) and Rtx-5ms capillary gas chromatography columns (Resick Corporation, Bellefonte, PA). Total lipids were calculated from total cholesterol and triglycerides (Phillips et al. 1989). We determined total cholesterol and triglycerides enzymatically in a small aliquot of serum at the Clinical and Epidemiological Research Laboratory, Boston Children’s Hospital (Boston, MA), and results were reported both as nanograms per milliliter of serum and as nanograms per gram lipid.
We used recoveries of internal standards (PCB-14, PCB-65, PCB-166 and TCMX) to gauge overall data quality for all analytes across all serum batches. Recoveries were between 81 and 99%, and no corrections were made to the measurements. Control charts on the performance of the laboratory controls (reagent blanks, fortified bovine serum, and SRM 1589a) were maintained for all analytes across all batches to ensure that results were within quality control (QC) criteria. Of the samples analyzed, 420 were from participants and 20 were blind laboratory controls interspersed among the actual samples serving as external QC controls. The identity of the 20 external QC controls was revealed only at the end of the analyses, and results were used to assess precision among batches. Based on these external QC samples, within-batch precision [expressed as the intrabatch coefficient of variation (CV%)] was 2.71% for DDT and 3.01% for DDE. The interbatch CV% was 9.97% for DDT and 9.11% for DDE.
Statistical methods.
Preliminary analyses involved univariate examination of serum measures and covariates of interest using summary statistics. Among the available CHDS information, we selected variables known from the literature to be related to exposures or outcomes as potential covariates: maternal age, prepregnancy body mass index (BMI), parity, maternal ethnicity, maternal place of birth, maternal occupation before pregnancy, birth weight, gestational age, date of blood draw, and season of birth. We plotted the cumulative distribution as well as density and quintile plots. We performed bivariate analysis of covariates and case/control status using logistic regression. We performed bivariate analysis between covariates and the distribution of serum measures using linear regression for continuous variables and analysis of variance for categorical variables. Because 25% of all observations were missing either height or prepregnancy weight, we imputed the prepregnancy BMI for those women who were missing only prepregnancy weight by calculating median weight gained during pregnancy for women in each pregnancy BMI category and applied this weight change to the women whose weight was measured at the same point during their pregnancy.
We examined the relationship between exposure and case status for both cryptorchidism and hypospadias using logistic regression. For models that included serum measures as continuous variables, we used log-transformed values. For categorical analysis, we identified quartile cut points based on the distribution of each measure among the whole study sample. All regression analyses included cholesterol and triglycerides (milligrams per deciliter serum) as separate continuous variables.
We included all covariates in a backward stepwise elimination model (exclusion criteria p < 0.20) to evaluate the influence of these potential confounders on the effect of exposures to DDT or DDE on cryptorchidism and hypospadias. Separate models were evaluated for cryptorchidism and for hypospadias; models for all cases combined were also examined. Finally, we assessed the effect of the ratio of DDE to DDT on the outcomes as a way to evaluate the effect of recentness of exposure to DDT. All analyses were performed using the open source statistical program RGui, version 1.3.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Most maternal, child, and delivery characteristics examined were similar among cases and controls (Table 1); a higher proportion of hypospadias cases were born to white mothers, and mothers with offspring with cryptorchidism had higher prepregnancy BMI than did control mothers. Mean DDE and DDT levels, whether expressed as a serum concentration (nanograms per milliliter serum) or lipid adjusted (micrograms per gram lipid), did not differ among cryptorchidism cases, hypospadias cases, and controls (Table 2).
Discussion
Our analysis did not find a statistically significant adverse association between maternal serum measures of DDT or DDE and cryptorchidism or hypospadias among pregnancies enrolled in the CHDS in California in the 1960s, when levels of exposure were considerably higher than they are today. The results of our study are consistent with those reported on hypospadias and cryptorchidism and DDE in archived maternal serum from the Collaborative Perinatal Project, a study conducted concurrently with the CHDS. The Collaborative Perinatal Project found that after adjusting for maternal race, triglyceride level, and cholesterol level, compared with boys whose mothers had serum DDE levels in the lowest quintile (< 21.4 ng/mL), boys whose mothers had serum DDE levels in the highest quintile (≥85.6 ng/mL) had ORs of 1.3 (95% CI, 0.7–2.4) for cryptorchidism and 1.2 (95% CI, 0.6–2.4) for hypospadias (Longnecker et al. 2002). Although our cut points differed from those in the Collaborative Perinatal Project, our results were comparable, with adjusted OR comparing boys whose mothers had serum DDE levels in the highest quartile (≥61.0 ng/mL) to boys whose mothers had serum DDE levels in the lowest quartile (< 27.0 ng/mL) at 1.34 (95% CI, 0.51–3.48) for cryptorchidism and 1.18 (95% CI, 0.46–3.02) for hypospadias. The similarity in the results between these two studies was found despite differences in the geographic location of the subjects (northern California vs. 12 centers across the United States), participants’ health care services (prepaid health plan vs. university-based practice), and case inclusion criteria (cryptorchidism present after 2 years of age vs. cryptorchidism diagnosed in first year of life).
Our study had several strengths. It sampled a population from a large prospective cohort study undertaken at a single site with excellent subject retention and reliable information on a number of relevant covariates. Only two subjects in our sample had used hormones in the interval from 6 months before the last menstrual period up to the pregnancy, so the sample was not subject to bias because of the effects of hormone use.
The study had consistent procedures for identifying and confirming congenital anomalies. Testicular descent is a dynamic process in that the prevalence of undescended testes may decrease with age (John Radcliffe Hospital Cryptorchidism Study Group 1992). Unlike the Collaborative Perinatal Project (Longnecker et al. 2002), the CHDS coded cases of cryptorchidism only if observed for 2 years, which allowed for spontaneous testicular descent and decreased the possibility of case misclassification. However, we could not exclude cases of cryptorchidism first observed after 1 year of age, which may have lead to a misdiagnosis of “retractile testes.” Nevertheless, the prevalence of cryptorchidism in the CHDS and Collaborative Perinatal Project studies was identical (1.08%), although the prevalence of hypospadias in the Collaborative Perinatal Project was about 25% higher (0.96 vs. 0.78%). Any misclassification of cases would be nondifferential with respect to exposure and would therefore attenuate any findings of an association.
The CHDS enrolled subjects at a time of high U.S. use of organochlorine insecticides (Kutz et al. 1991). Serum measures of DDE found in this study (43 ng/mL or 5.2 μg/g lipid) are comparable with those estimated from samples of body fat during the late 1960s and are among the highest recorded in U.S. populations (Kutz et al. 1991). Our results are comparable with measures of DDE (54 ng/mL or 6.9 μg/g lipid) from another analysis of CHDS archived maternal serum (James et al. 2002) as well as with those (43 ng/mL) found in an analysis of archived serum collected between 1964 and 1971 in northern California (Krieger et al. 1994). Notably, the CHDS mothers had DDE levels that were higher than the recovery-adjusted DDE levels in the Collaborative Perinatal Project mothers (34.3 ng/mL and 4.24 μg/g lipid) (Longnecker et al. 2002).
Serum measures in this population for DDE are also comparable with those associated with reproductive effects seen in eagles in their natural environments. In a study of nestling eagles, Bowerman et al. (1995) showed that productivity, measured as the number of young eagles observed in occupied nests, varied inversely with DDE levels measured in plasma of the same populations of eagles, which ranged from 5 to 40 ng/g. The serum DDE levels in our study are sufficient to inhibit androgen activity, based on in vitro effects. Kelce et al. (1995) found that 63.6 ng/mL DDE was sufficient to inhibit androgen receptor transcriptional activity in vitro.
Our study was limited in that serum was not available for all cases in the cohort, and this could have resulted in a bias. If the missing cases all had high DDT or DDE levels, then selection bias could have caused us to miss a positive association. An unmeasured confounder could also explain these results if levels of DDT or DDE varied with a protective factor for hypospadias or cryptorchidism. For example, fish consumption may promote fetal growth, a possible protective factor against the disorders studied, and may be associated with higher human cumulative exposure to organochlorines (Olsen and Secher 2002; Olsen et al. 1990, 1993). Notably, one U.S. study found no relationship between dietary consumption of fish and serum DDE levels (Laden et al. 1999).
Few models exist to adequately examine the combined effects of multiple chemicals with potential endocrine activity. Competing effects due to environmental agents with similar exposure pathways may obscure relationships between cause and effect in epidemiologic studies. For example, an androgen antagonist may oppose the effects of an estrogen agonist on the human pituitary because both androgens and estrogens inhibit luteinizing hormone secretion. Similarly, if DDE and DDT have differing effects, they could oppose each other and obscure any direct associations with adverse outcomes.
In summary, our study does not provide epidemiologic support for a causal adverse relationship between DDT or DDE and cryptorchidism or hypospadias. Our sample size was adequate to identify an approximate doubling of risk for the outcomes under study relative to the range of serum measures; however, the study may lack sufficient power to find a more modest effect with the observed exposure levels. Overall, DDT and DDE remain appropriate candidates for the study of environmental endocrine effects, especially given their wide use, their environmental persistence, their documented adverse reproductive effects in animals, and effects on other reproductive outcomes. Although this study does not support an association of DDT or DDE and hypospadias and cryptorchidism, the continued use of DDT in vector control in developing countries and its global distribution warrant further inquiry.
Table 1 Characteristics of mothers and male offspring in a nested case–control study of U.S. participants in the CHDS, 1959–1967.
Characteristics Cryptorchidism cases (n = 75) Hypospadias cases (n = 66) Controls (n = 283)
Maternal characteristics
Mean ± SD [age (years)] 27.9 ± 6.5 26.4 ± 6.0 26.6 ± 6.2
Race [no. (%)]
White 46 (61.3) 49 (74.2) 173 (61.1)
Latino 1 (1.3) 1 (1.5) 9 (3.2)
Black 21 (28.0) 9 (13.6) 82 (29.0)
Asian 5 (6.7) 4 (6.1) 10 (3.5)
Other 1 (1.3) 1 (1.5) 7 (2.5)
Unknown 1 (1.3) 2 (3.0) 2 (0.7)
Highest level of education completed [no. (%)]
< 12th grade 14 (18.7) 9 (13.6) 50 (17.7)
High school graduate/trade school 44 (58.7) 37 (56.0) 160 (56.5)
College graduate 13 (17.3) 10 (15.2) 38 (13.4)
Unknown 4 (5.3) 10 (15.2) 35 (12.4)
Household income [no. (%)]
< $5,000 15 (20.0) 10 (15.2) 45 (15.9)
$5,000–9,999 32 (42.7) 31 (47.0) 119 (42.0)
$10,000–14,999 10 (13.3) 9 (13.6) 38 (13.4)
≥$15,000 2 (2.7) 1 (1.5) 3 (1.1)
Unknown 16 (21.4) 15 (22.8) 78 (27.5)
Ever smoked [no. (%)]
Yes 31 (41.3) 28 (42.4) 117 (41.3)
No 36 (48.0) 27 (40.9) 111 (39.2)
Unknown 8 (10.7) 11 (16.7) 55 (19.4)
Years lived on a farm before age 15 [no. (%)]
None 37 (49.3) 39 (59.1) 142 (50.2)
1–4 5 (6.7) 6 (9.1) 19 (6.7)
≥5 14 (18.7) 4 (6.1) 34 (12.0)
Unknown 19 (25.3) 17 (25.8) 88 (31.1)
Maternal place of birth [no. (%)]
California 30 (40.0) 24 (36.4) 85 (30.0)
Southeastern United States 15 (20.0) 7 (10.6) 66 (23.3)
Other U.S. states 20 (26.7) 18 (27.3) 73 (25.8)
Non-United States 6 (8.0) 7 (10.6) 24 (8.5)
Unknown 4 (5.3) 10 (15.2) 35 (12.4)
Parity [no. (%)]
0 24 (32.0) 22 (33.3) 84 (29.7)
1–2 29 (38.7) 30 (45.5) 113 (39.9)
≥3 21 (28.0) 12 (18.2) 78 (27.6)
Unknown 1 (1.3) 2 (3.0) 8 (2.8)
Median [IQR] prepregnancy BMI (kg/m2) 22 [20–25] 21 [20–24] 21 [20–24]
Median [IQR] age at menarche (years) 13 [12–13] 12 [11–13] 12 [11–13]
Child’s characteristics
Season of birth [no. (%)]
January–March 17 (22.7) 23 (34.8) 65 (23.0)
April–June 20 (26.7) 12 (18.2) 68 (24.0)
July–September 22 (29.3) 18 (27.3) 70 (24.7)
October–December 16 (21.3) 13 (19.7) 80 (28.3)
Method of delivery [no. (%)]
Vaginal 74 (98.7) 62 (93.9) 274 (96.8)
Cesarian 1 (1.3) 4 (6.1) 9 (3.2)
Small for gestational age [no. (%)] 6 (8.0) 9 (13.6) 36 (12.7)
Preterm birth [no. (%)] 2 (2.7) 3 (4.5) 21 (7.4)
Median [IQR] gestational age (weeks) 40 [39–41] 40 [38–41] 40 [29–41]
Median [IQR] birth weight (g) 3,374 [3,048–3,671] 3,260 [2,948–3,622] 3,345 [3,005–3,657]
IQR, interquartile range.
Table 2 Serum concentration distributions for organochlorine compounds in a nested case–control study of U.S. participants in the CHDS, 1959–1967.
Serum measures Cryptorchidism cases [n = 75; median (IQR)] Hypospadias cases [n = 66; median (IQR)] Controls [n = 283; median (IQR)]
DDE serum concentration (ng/mL) 43.0 (32.0–60.0) 41.0 (30.2–57.8) 43.0 (32.0–56.5)
DDE, lipid adjusted (μg/g lipid) 5.3 (3.9–7.3) 4.6 (3.5–6.6) 5.2 (3.8–6.9)
DDT serum concentration (ng/mL) 12.1 (8.7–18.1) 9.5 (7.5–14.2) 11.1 (8.4–16.1)
DDT, lipid adjusted (μg/g lipid) 1.4 (1.0–2.0) 1.2 (0.9–1.6) 1.4 (1.0–1.9)
Serum cholesterol concentration (g/L) 2.5 (2.1–3.0) 2.6 (2.2–3.0) 2.5 (2.1–3.0)
Serum triglycerides concentration (g/L) 1.7 (1.4–2.4) 2.1 (1.5–2.6) 1.8 (1.4–2.3)
Total serum lipid concentration (g/L) 8.0 (6.7–9.5) 8.6 (7.3–9.9) 8.1 (7.0–9.5)
IQR, interquartile range.
Table 3 Adjusted ORs (95% CI) for birth defects among male offspring by DDE level in mother’s serum, CHDS, 1959–1967.
DDE (ng/mL serum) No. of Cases No. of Controls Adjusteda OR (95% CI) Adjustedb OR (95% CI)
All cases
< 27.0 21 42 Reference Reference
27.0–43.9 53 107 0.95 (0.50–1.77) 0.99 (0.52–1.89)
44.0–60.9 35 83 0.79 (0.41–1.55) 0.86 (0.43–1.70)
≥61.0 28 51 1.02 (0.50–2.09) 1.24 (0.58–2.63)
p-Value for trend 0.89 0.72
Cryptorchidism
< 27.0 10 42 Reference Reference
27.0–43.9 30 107 1.16 (0.52–2.60) 1.17 (0.51–2.66)
44.0–60.9 19 83 0.94 (0.40–2.24) 0.95 (0.39–2.30)
≥61.0 16 51 1.29 (0.52–3.22) 1.34 (0.51–3.48)
p-Value for trend 0.77 0.75
Hypospadias
< 27.0 12 42 Reference Reference
27.0–43.9 24 107 0.73 (0.33–1.62) 0.81 (0.36–1.84)
44.0–60.9 16 83 0.61 (0.26–1.43) 0.68 (0.28–1.64)
≥61.0 14 51 0.86 (0.35–2.10) 1.18 (0.46–3.02)
p-Value for trend 0.7 0.82
a Adjusted for cholesterol and triglyceride levels.
b Adjusted for cholesterol level, triglyceride level, and maternal race.
Table 4 Adjusted ORs (95% CI) for birth defects among male offspring by DDT level in mother’s serum, CHDS, 1959–1967.
DDE (ng/mL serum) No. of Cases No. of Controls Adjusteda OR (95% CI) Adjustedb OR (95% CI)
All cases
< 10.0 65 117 Reference Reference
10.0–14.9 29 87 0.59 (0.35–0.99) 0.63 (0.37–1.07)
15.0–19.9 24 37 1.13 (0.62–2.06) 1.25 (0.66–2.36)
≥20.0 19 42 0.77 (0.41–1.44) 0.89 (0.46–1.72)
p-Value for trend 0.62 0.98
Cryptorchidism
< 10.0 32 117 Reference Reference
10.0–14.9 12 87 0.50 (0.24–1.02) 0.49 (0.23–1.01)
15.0–19.9 20 37 1.95 (0.99–3.83) 2.04 (1.00–4.18)
≥20.0 11 42 0.95 (0.43–2.07) 1.01 (0.44–2.28)
p-Value for trend 0.42 0.38
Hypospadias
< 10.0 34 117 Reference Reference
10.0–14.9 18 87 0.70 (0.37–1.32) 0.81 (0.42–1.56)
15.0–19.9 5 37 0.45 (0.16–1.24) 0.45 (0.16–1.28)
≥20.0 9 42 0.66 (0.28–1.52) 0.79 (0.33–1.89)
p-Value for trend 0.15 0.30
a Adjusted for cholesterol and triglyceride levels.
b Adjusted for cholesterol level, triglyceride level, and maternal race.
==== Refs
References
Baskin LS Himes K Colborn T 2001 Hypospadias and endocrine disruption: is there a connection? Environ Health Perspect 109 1175 1183 11713004
Beard CM Melton LJ III O’Fallon WM Noller KL Benson RC 1984 Cryptorchism and maternal estrogen exposure Am J Epidemiol 120 707 716 6149686
Berkowitz GS Lapinski RH Godbold JH Dolgin SE Holzman IR 1995 Maternal and neonatal risk factors for cryptorchidism Epidemiology 6 127 131 7742397
Bowerman WW Giesy JP Best DA Kramer VJ 1995 A review of factors affecting productivity of bald eagles in the Great Lakes region: implications for recovery Environ Health Perspect 103 suppl 4 51 59 7556024
Bulger WH Kupfer D 1983 Estrogenic action of DDT analogs Am J Ind Med 4 163 173 6188376
Christianson RE van den Berg BJ Milkovich L Oechsli FW 1981 Incidence of congenital anomalies among white and black live births with long-term follow-up Am J Public Health 71 1333 1341 7315998
Clarkson TW 1995 Environmental contaminants in the food chain Am J Clin Nutr 61 3 suppl 682S 686S 7879738
Cosgrove MD Benton B Henderson BE 1977 Male genitourinary abnormalities and maternal diethylstilbestrol J Urol 117 220 221 833973
Danzo BJ 1997 Environmental xenobiotics may disrupt normal endocrine function by interfering with the binding of physiological ligands to steroid receptors and binding proteins Environ Health Perspect 105 294 301 9171990
Depue RH 1984 Maternal and gestational factors affecting the risk of cryptorchidism and inguinal hernia Int J Epidemiol 13 311 318 6149198
Facemire CF Gross TS Guillette LJ Jr 1995 Reproductive impairment in the Florida panther: nature or nurture? Environ Health Perspect 103 suppl 4 79 86 7556029
Fry DM Toone CK 1981 DDT-induced feminization of gull embryos Science 213 922 924 7256288
García-Rodríguez J García-Martín M Nogueras-Ocaña M Lunadel-Castillo J García ME Olea N 1996 Exposure to pesticides and cryptorchidism: geographical evidence of a possible association Environ Health Perspect 104 1090 1095 8930551
Gill WB Schumacher FB Bibbo M Straus FH Schoenberg HW 1979 Association of diethylstilbestrol exposure in utero with cryptorchidism, testicular hypoplasia, and semen abnormalities J Urol 122 36 39 37351
Gray LE Jr Ostby J Furr J Wolf CJ Lambright C Parks L 2001 Effects of environmental antiandrogens on reproductive development in experimental animals Hum Reprod Update 7 248 264 11392371
Guillette LJ Jr Gross TS Gross DA Rooney AA Percival HF 1995 Gonadal steroidogenesis in vitro from juvenile alligators obtained from contaminated or control lakes Environ Health Perspect 103 suppl 4 31 36 7556021
Guillette LJ Jr Guillette EA 1996 Environmental contaminants and reproductive abnormalities in wildlife: implications for public health? Toxicol Ind Health 12 537 550 8843570
Hjertkvist M Damber JE Bergh A 1989 Cryptorchidism: a registry based study in Sweden on some factors of possible aetiological importance J Epidemiol Community Health 42 324 329 2575645
Jackson MB 1988 The epidemiology of cryptorchidism. John Radcliffe Hospital Cryptorchidism Research Group Horm Res 30 153 156 2907891
James RA Hertz-Picciotto I Willman E Keller JA Charles MJ 2002 Determinants of serum polychlorinated biphenyls and organochlorine pesticides measured in women from the Child Health and Development Study Cohort, 1963–1967 Environ Health Perspect 110 617 624 12117636
John Radcliffe Hospital Cryptorchidism Study Group 1992 Cryptorchidism: a prospective study of 7500 consecutive male births, 1984–8 Arch Dis Child 67 892 899 1355643
Kelce WR Stone CR Laws SC Gray LE Kemppainen JA Wilson EM 1995 Persistent DDT metabolite p,p ′-DDE is a potent androgen receptor antagonist Nature 375 581 585 7791873
Krieger N Wolff MS Hiatt RA Rivera M Vogelman J Orentreich N 1994 Breast cancer and serum organochlorines: a prospective study among white, black, and Asian women J Natl Cancer Inst. 86 589 599 8145274
Kutz FW Wood PH Bottimore DP 1991 Organochlorine pesticides and polychlorinated biphenyls in human adipose tissue Rev Environ Contam Toxicol 120 1 82 1899728
Laden F Neas LM Spiegelman D Hankinson SE Willett WC Ireland K 1999 Predictors of plasma concentrations of DDE and PCBs in a group of U.S. women Environ Health Perspect 107 75 81 9872720
Landrigan P Garg A Droller DBJ 2003 Assessing the effects of endocrine disruptors in the National Children’s Study Environ Health Perspect 111 1678 1682 14527850
Longnecker MP Klebanoff MA Brock JW Zhou H Gray KA Needham LL 2002 Maternal serum level of 1,1-dichloro-2,2-bis(p -chlorophenyl)ethylene and risk of cryptorchidism, hypospadias, and polythelia among male offspring Am J Epidemiol 155 313 322 11836195
Longnecker MP Klebanoff MA Gladen BC Berendes HW 1999 Serial levels of serum organochlorines during pregnancy and postpartum Arch Environ Health 54 110 114 10094288
McBride ML van den Steen N Lamb CW Gallagher RP 1991 Maternal and gestational factors in cryptorchidism Int J Epidemiol 20 964 970 1686874
Olsen SF Grandjean P Weihe P Videro T 1993 Frequency of seafood intake in pregnancy as a determinant of birth weight: evidence for a dose dependent relationship J Epidemiol Community Health 47 436 440 8120495
Olsen SF Olsen J Frische G 1990 Does fish consumption during pregnancy increase fetal growth? A study of the size of the newborn, placental weight and gestational age in relation to fish consumption during pregnancy Int J Epidemiol 19 971 977 2084030
Olsen SF Secher NJ 2002 Low consumption of seafood in early pregnancy as a risk factor for preterm delivery: prospective cohort study [Abstract] Br Med J 324 447 11859044
Paulozzi LJ 1999 International trends in rates of hypospadias and cryptorchidism Environ Health Perspect 107 297 302 10090709
Petreas M She J Brown FR Winkler J Windham G Rogers E 2003 High body burdens of 2,2′,4,4′-tetrabromodiphenyl ether (BDE-47) in California women Environ Health Perspect 111 1175 1180 12842770
Phillips DL Pirkle JL Burse VW Bernert JT Jr Henderson LO Needham LL 1989 Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding Arch Environ Contam Toxicol 18 495 500 2505694
Sharpe RM Skakkebaek NE 1993 Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract? Lancet 341 1392 1395 8098802
Subramanian AN Tanabe S Tatsukawa R Saito S Miyazak N 1987 Reduction in the testosterone levels by PCBs and DDE in Dall’s porpoises of northwestern North Pacific Mar Pollut Bull 18 643 646
Sweet RA Schrott HG Kurland R Culp OS 1974 Study of the incidence of hypospadias in Rochester, Minnesota, 1940–1970, and a case-control comparison of possible etiologic factors Mayo Clin Proc 49 52 57 4855617
Swerdlow AJ Wood KH Smith PG 1983 A case-control study of the aetiology of cryptorchidism J Epidemiol Community Health 37 238 244 6137506
Toppari J Larsen JC Christiansen P Giwercman A Grandjean P Guillette LJ Jr 1996 Male reproductive health and environmental xenoestrogens Environ Health Perspect 104 suppl 4 741 803 8880001
van den Berg BJ Christianson RE Oechsli FW 1988 The California Child Health and Development Studies of the School of Public Health, University of California at Berkeley Paediatr Perinat Epidemiol 2 265 282 3070486
Wilson JD Foster DW 1985. Williams Textbook of Endocrinology. 7th ed. Philadelphia:W.B. Saunders.
WHO 1957. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death, 7th Revision. Geneva:World Health Organization.
|
15687061
|
PMC1277868
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb 4; 113(2):220-224
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7382
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/ehp.7111ehp0113-00022515687062Children's HealthArticlesRelation of Trihalomethane Concentrations in Public Water Supplies to Stillbirth and Birth Weight in Three Water Regions in England Toledano Mireille B. 1Nieuwenhuijsen Mark J. 12Best Nicky 1Whitaker Heather 13Hambly Peter 1de Hoogh Cornelis 1Fawell John 4Jarup Lars 1Elliott Paul 11Small Area Health Statistics Unit, Department of Epidemiology and Public Health, Faculty of Medicine, and2Department of Environmental Science and Technology, Imperial College London, London, United Kingdom3Department of Statistics, Open University, Milton Keynes, United Kingdom4Independent Consultant, Buckinghamshire, United KingdomAddress correspondence to P. Elliott, Department of Epidemiology and Public Health, Faculty of Medicine, Imperial College London, St. Mary’s Campus, Norfolk Pl., London W2 1PG, UK. Telephone: 44-0-20-7594-3328. Fax: 44-0-20-7402-2150. E-mail:
[email protected] thank the following for their helpful contribution to the study: A. Gowers, J. Bennett, I. Maitland, N. Cobley, K. Konstantinou, D. Fecht, S. Cockings, D. Briggs, V. Barnard, and S. Fawell. We are also grateful to Northumbrian Water, United Utilities Water (formerly North West), and Severn Trent Water for providing the trihalomethane data and to the Office for National Statistics for providing the health data used in the study.
The Small Area Health Statistics Unit is funded by a grant from the Department of Health, Department of the Environment, Food and Rural Affairs, Environment Agency, Health and Safety Executive, Scottish Executive, Welsh Assembly Government and Northern Ireland Department of Health, Social Services and Public Safety. The views expressed in this publication are those of the authors and not necessarily those of the funding departments.
The authors declare they have no competing financial interests. J. Fawell consults on issues relating to water supply and safety for both government and industry.
2 2005 21 10 2004 113 2 225 232 23 3 2004 21 10 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. We investigated the association between total trihalomethanes (TTHMs) and risk of stillbirth and low and very low birth weight in three water regions in England, 1992–1998; associations with individual trihalomethanes (THMs) were also examined. Modeled estimates of quarterly TTHM concentrations in water zones, categorized as low (< 30 μg/L), medium (30–59 μg/L), or high (≥60 μg/L), were linked to approximately 1 million routine birth and stillbirth records using maternal residence at time of birth. In one region, where there was a positive socioeconomic deprivation gradient across exposure categories, there was also a positive, significant association of TTHM with risk of stillbirth and low and very low birth weight. Overall summary estimates across the three regions using a random-effects model to allow for between-region heterogeneity in exposure effects showed small excess risks in areas with high TTHM concentrations for stillbirths [odds ratio (OR) = 1.11; 95% confidence interval (CI), 1.00–1.23), low birth weight (OR = 1.09; 95% CI, 0.93–1.27), and very low birth weight (OR = 1.05; 95% CI, 0.82–1.34). Among the individual THMs, chloroform showed a similar pattern of risk as TTHM, but no association was found with concentrations of bromodichloromethane or total brominated THMs. Our findings overall suggest a significant association of stillbirths with maternal residence in areas with high TTHM exposure. Further work is needed looking at cause-specific stillbirths and effects of other disinfection by-products and to help differentiate between alternative (noncausal) explanations and those that may derive from the water supply.
chemicaldisinfectioninfant low birth weightpregnancy outcomestillbirthtrihalomethaneswater pollutionwater purification
==== Body
Chlorination has been the main means for disinfecting municipal drinking water in many countries, including the United Kingdom, for many decades. The added chlorine reacts with naturally occurring organic matter to form a wide range of unwanted halogenated organic compounds, often referred to as disinfection by-products (DBPs). Among the most widely occurring by-products are trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles, and haloketones (Krasner et al. 1989; Nieuwenhuijsen et al. 2000b). Besides organic matter and chlorine dose, factors affecting the composition and concentration of DBPs include residence time in the distribution system, temperature, pH, and bromide levels (Chen and Weisel 1998; Krasner et al. 1989, 1994; Singer 1999; Stevens et al. 1989).
Of the DBPs, a group of four THMs [chloroform, bromodichloromethane (BDCM), dibromochloromethane (DBCM), and bromoform] generally occur at the highest concentrations in drinking water and are the DBPs for which standards are most commonly set. In consequence, they are routinely measured throughout water supplies and have been used as the exposure index in various epidemiologic studies that have examined the relationship between DBPs and adverse birth outcomes. These studies have, however, used a variety of study designs and methods of exposure assessment, and findings to date have been inconsistent (Nieuwenhuijsen et al. 2000a). Although some have reported significant excess risks of low birth weight or small size for gestational age (Bove et al. 1995; Gallagher et al. 1998; Kallen and Robert 2000; Kramer et al. 1992; Wright et al. 2003), others have not (Dodds et al. 1999; Jaakkola et al. 2001; Kanitz et al. 1996; Savitz et al. 1995; Yang et al. 2000). A number of studies conducted in Canada have found significant excess risks of stillbirths with higher total THM concentrations (Dodds et al. 1999, 2004) and chloroform and BDCM (King et al. 2000), whereas studies in other countries have found no excess risk (Aschengrau et al. 1993; Bove et al. 1995; Jaakkola et al. 2003). Evidence for an association with spontaneous abortion is sparse; two studies have reported significant excess risks (80–100%) (Aschengrau et al. 1989; Waller et al. 1998, 2001), whereas Savitz et al. (1995) found a smaller (20%), nonsignificant excess risk. Of the 10 studies to date that have examined various congenital anomalies, eight have shown excess risks for some of the congenital defects (Aschengrau et al. 1993; Bove et al. 1995; Cedergren et al. 2002; Dodds and King 2001; Dodds et al. 1999; Hwang et al. 2002; Klotz and Pyrch 1999; Magnus et al. 1999), particularly for all defects (Aschengrau et al. 1993; Bove et al. 1995; Hwang et al. 2002; Magnus et al. 1999), neural tube defects (Bove et al. 1995; Dodds and King 2001; Dodds et al. 1999; Klotz and Pyrch 1999; Magnus et al. 1999), and urinary defects (Aschengrau et al. 1993; Hwang et al. 2002; Magnus et al. 1999).
Despite these inconsistencies, the large number of people exposed to chlorinated water supplies means that potentially the population-attributable risk is high, even though the available evidence suggests that the risk, if any, for stillbirth and low birth weight in relation to THMs is small. Here we present the results of the first U.K. study to examine this question.
Materials and Methods
Public water supplies in the United Kingdom are statutorily divided into water zones, each zone covering a population of < 50,000. Less than 1% of households in the United Kingdom have private water supplies. The study was carried out in regions covered by three water companies in the north and midlands of England: Northumbrian, United Utilities (formerly North West), and Severn Trent (Figure 1), for which published data on THMs showed a wide range of exposure variation across water zones, and data on water zone boundaries were available. Northumbrian Water supplies approximately 2.6 million people across 120 water zones; United Utilities Water, 6.8 million people across 315 water zones; and Severn Trent Water, 7.4 million people across 300 water zones.
Where available, digital boundaries of water zones and their identification codes were obtained from each water company for each year under study. Alternatively, paper maps providing such details were obtained and digitized in-house, using the ArcInfo (version 7.02; Environmental Systems Research Institute, Redlands, CA, USA) geographical information system (GIS). Boundary data were available for the following years: Northumbrian, 1997; United Utilities, 1992–1997, and Severn Trent, 1993–1998.
Individual postal-coded records were extracted from the national birth and stillbirth registers held at the U.K. Small Area Health Statistics Unit (SAHSU; London, UK). Low birth weight was defined as < 2,500 g, and very low birth weight as < 1,500 g. Registration of all stillbirths is a legal requirement in the United Kingdom, providing a national register with high levels of ascertainment. Since the end of 1992, stillbirths are legally defined as fetal deaths after 24 completed weeks of gestation. Birth and stillbirth records were subsequently linked to data on water supply for the three regions. A link between postal code and water zone was created using point-in-polygon methods within the GIS software to allocate each postal code to its water supply zone. Postal code locations were derived from the historical postal code file for Great Britain, developed by SAHSU. This file traces postal codes back in time and assigns a grid coordinate for each postal code in each year. To take account of changes in the location of both postal codes and water zone boundaries over time, a separate link was created for each year of the study period.
Exposure data.
We used THM concentrations as the marker for chlorination byproducts in this study. Water samples are routinely collected and analyzed from each water zone using random samples at the tap. Under the regulations operating during the study period, the standard sampling frequency for THMs was a minimum of four samples per annum. However, if there was a breach of the standard of 100 μg/L for total THMs (TTHMs) as a rolling 3-month average (or, where samples were too few, a maximum concentration > 100 μg/L in any one sample), the sampling frequency increased to a minimum of 12 or 24 per annum depending on the zone size. Conversely, if the TTHM concentration was consistently < 50% of the standard, a reduced frequency of a minimum of one per annum could be used. The number of THM samples that had been collected and recorded in each zone was highly variable, ranging from 1 to 80 measurements in a year (Whitaker et al. 2004). The mean number of samples per region, per year, was 11.2 for United Utilities Water, 6.3 for Severn Trent Water, and 4.5 for Northumbrian Water. In addition, a number of THM measurements were below the limit of detection, with percentages ranging across regions as follows: chloroform, 3.5% (United Utilities) to 23% (Severn Trent); BDCM, 4.9% (United Utilities) to 22% (Severn Trent); DBCM, 16% (Severn Trent) to 82% (Northumbrian); and bromoform, 21% (Severn Trent) to 85% (United Utilities) (Whitaker et al. 2003b).
Because of the small number of THM measurements in some water zones, the need for quarterly (3 months) estimates (to allow for trimester weighted exposure estimates), and the problem of measurements below the limit of detection, it was necessary to model the raw THM data to obtain more robust estimates of the mean TTHM concentration in each zone. This was done using a hierarchical mixture model in the software WinBUGS (Bayesian inference using Gibbs sampling) (Spiegelhalter et al. 1996), as described elsewhere (Whitaker et al. 2004). Briefly, modeling was carried out separately for each water company and year. In each case, the data were transformed to approximate normality using an appropriate Box–Cox transformation (Box and Cox 1964). The model calculated the mean annual individual THM concentrations for each water zone and subsequently assigned an estimated water source type to each water zone depending on the four THM levels within each zone. We fitted a three-component mixture model in which zones were assumed to belong to one or some mixture of three components that we labeled “ground,” “lowland surface,” and “upland surface” waters (the components may not strictly correspond to these three water source types; we simply aimed to group waters with similar THM profiles, which are more likely to be shared among water of the same source type). Constraints were imposed on the model such that, for example, on average, chloroform concentrations were highest in the two “surface water” components and bromoform concentrations were highest in the “groundwater” component. These constraints were based on a priori knowledge about the relative concentrations of different THMs in different water sources, and were necessary to identify the three components in the mixture model. The hierarchical model was assigned over the zone-specific mean individual THM concentrations, enabling zones to “borrow” information from other zones with the same water source type. This resulted in more stable estimates for zones where few samples were taken. For measurements under the detection limit, we modeled to obtain an estimate between zero and the detection limit, rather than arbitrarily assigning half or two-thirds the detection limit, which is common practice. Seasonal variation was taken into account by estimating a quarterly effect common to all zones supplied by the same source type. These quarterly zone mean THM estimates were then back-transformed onto the original scale and summed to give TTHM levels.
The postal code of the maternal residence at the year of birth was used to identify the water zone of interest and hence the appropriate exposure status for each birth record. Because the final trimester may be the most relevant trimester of pregnancy for both low birth weight and stillbirth (Kline et al. 1989; Pless 1994), we obtained exposure status by calculating a weighted average of the modeled quarterly TTHM estimates for the appropriate zone for the last 93 days before the date of birth. The weighting was based on the proportion of the trimester falling into each quarterly period. Because data on gestation weeks at birth were unavailable, we were unable to allow for pregnancies that had not gone to term. For full-term pregnancies, the last 93 days would equate to the third trimester. For a premature fetus, this period would be the last 93 days of the pregnancy, which will include part of the second trimester. Births occurring in the first 93 days of the first year of the study for each company were excluded. Finally, the weighted average TTHM estimate associated with each birth record was categorized into one of three predefined exposure categories: low (< 30 μg/L), medium (30–59 μg/L), or high (≥60 μg/L). These were chosen with reference to the published literature on the possible associations of birth outcomes with TTHMs (Nieuwenhuijsen et al. 2000a) and with regard to the distribution of TTHM levels across the three water regions.
Study population size.
The study population comprised all births in the water regions for a varying number of years between 1992 and 1998. A total of six (0.02%), 3,471 (0.68%), and 24,157 (4.5%) births were excluded in the Northumbrian, United Utilities, and Severn Trent areas, respectively. Reasons for exclusion included births occurring in water zones that could not be assigned an exposure estimate because of zone code-boundary mismatches; postal codes of births falling into gaps or overlaps between different water zone boundaries; lack of water zone identification provided with the zone boundary data; or water supplied to that postal code not from one of the three study water companies but from another water company or a private water supply. There were no material differences in the birth weight and stillbirth profiles of these excluded births and those that were retained in the study. A further 592 (2.9%), 12,247 (2.5%), and 13,888 (2.8%) multiple births were excluded in the three regions, respectively. This left 20,624 total (live and still) births in Northumbrian in 1997, 412,973 in United Utilities in 1993–1997 (data for 1992 were omitted because of a national change in the definition of a stillbirth), and 486,974 in Severn Trent for 1993–1998. Analysis of birth weight was restricted to live-birth records with a birth weight > 200 g (99% of birth weights < 200 g were recorded as zero), giving, for these analyses, 20,452 live births in Northumbrian in 1997, 467,597 live births in United Utilities in 1992–1997, and 481,255 in Severn Trent in 1993–1998.
Statistical methods.
Using the statistical package S-Plus (Insightful, Seattle, WA, USA), we performed descriptive analysis for all three outcomes—stillbirth, low birth weight, and very low birth weight—in each region separately, as well as univariate and multiple logistic regression modeling with adjustment for measured potential confounders. Sex and maternal age (for which individual-level information was available) were considered as potential confounders, as was socioeconomic deprivation measured at the small-area level, according to location of the postal code of maternal residence at the time of birth. Maternal age was represented in five categories: ≤ 20, 21–25, 26–30, 31–35, and ≥ 36 years. Deprivation was measured by quintiles of the Carstairs index (Carstairs and Morris 1991), a combination of four indicators from the 1991 census at the level of enumeration district (the smallest geographic area for which British census data are available, with, on average, 400 people): the percentage of people with no car, percentage living in overcrowded housing, percentage with the head of household in social class IV (partly skilled occupations) or V (unskilled occupations), and the percentage of men unemployed. In the regression models, only potential confounders that led to a significant (p < 0.05) change in the model deviance or led to > 5% change in the log odds ratio (OR), in at least one of the three study regions, were included in the final models (Greenland and Rothman 1998a). Interaction parameters between THM category and all other covariates were tested in the final models.
Generalized additive models were fitted using smoothing splines (Hastie and Tibshirani 1990), to examine the shape of association, and to check linearity assumptions, for both continuous THM estimates and Carstairs deprivation score in each water region separately. In addition, a multilevel model with random water zone effects was fitted using the glmmPQL function in S-Plus (version 6.2, function available from Modern Applied Statistics with S-Plus library; Insightful) (Leyland and Goldstein 2001; Venables and Ripley 2002) to check for residual clustering of the outcomes within water zones. Finally, tests for heterogeneity of the ORs associated with THM exposure across the three water regions were performed and a random-effects model was used to obtain an overall summary estimate of the effect of THM allowing for heterogeneity in the region-specific estimates (Dersimonian and Laird 1986). All analyses were carried out for TTHMs, chloroform, BDCM, and total brominated THMs (the sum of BDCM, DBCM, and bromoform). (Levels of DBCM and bromoform were often below the detection limit and too low for categorization and meaningful analysis in the three regions under study.)
Results
Descriptive analysis.
Descriptive data are shown in Table 1. Mean prevalence across the three regions ranged from 5.2 to 5.4 per 1,000 live and stillbirths for stillbirth, and from 61.5 to 64.8 and 9.1 to 10.7 per 1,000 live births for low and very low birth weight, respectively, whereas mean birth weight ranged from 3,337 to 3,351 g. Northumbrian was the most deprived region, and Severn Trent was the most affluent (mean Carstairs scores of 1.54 and 0.65, respectively). Maps for each region showing TTHM exposure classification by water zone and quarter are shown in Figures 2–4.
Average TTHM concentrations were similar for Northumbrian and United Utilities (56.6 and 52.0 μg/L, respectively), whereas the average concentration in Severn Trent was somewhat lower (35.8 μg/L); the average concentration in each of the three exposure categories was, however, similar in all three regions (Table 1). A tendency for increasing deprivation across the exposure categories (low to high) was seen in United Utilities but not in the other two regions. A pattern of higher rates of stillbirth and low and very low birth weight and lower mean birth weight was also seen across increasing exposure categories in the United Utilities region. In Severn Trent, there was a tendency for the reverse pattern for low and very low birth weight but not for stillbirths.
Regression models.
Univariate logistic regression analysis for stillbirths and low and very low birth weight confirmed a trend of increasing prevalence with higher TTHM concentrations in United Utilities but not in the other regions. In United Utilities, the unadjusted ORs [95% confidence intervals (CIs)] for stillbirth in medium versus low and high versus low-exposure categories were 1.21 (1.04–1.41) and 1.34 (1.15–1.57), respectively; for low birth weight they were 1.20 (1.15–1.25) and 1.37 (1.31–1.43), respectively; for very low birth weight, they were 1.15 (1.03–1.28) and 1.32 (1.18–1.48), respectively.
Table 2 shows the results of the multiple logistic regression analysis for stillbirths and low and very low birth weight for each water region after adjusting for potential confounders. Again, in the United Utilities region, the risk was always highest and significant in the high-exposure category, with intermediate risk in the medium-exposure category. In Severn Trent, no statistically significant association was found between risk of stillbirths and low birth weight and TTHM concentrations, although for very low birth weight the risk in the high-TTHM areas was lower than in the low-TTHM areas (OR = 0.90; 95% CI, 0.82–0.99). Nonsignificant excess risks in medium- and high-TTHM areas relative to low-TTHM areas were found in Northumbrian for each birth outcome; CIs were wide, reflecting the much smaller numbers of births included in this region. No significant interactions between TTHM exposure and any of the potential confounders were found in the multivariate analysis.
The ORs associated with TTHM exposure showed statistically significant heterogeneity between water regions, for both low and very low birth weight but not for stillbirths (Table 2, notes). Allowing for this heterogeneity using a random-effects model to obtain overall summary estimates of the TTHM effects, small excess risks were found in the high- compared with low-exposure areas for stillbirths and low and very low birth weight of 11% (95% CI, 0–23%), 9% (95% CI, –7–27%), and 5% (95% CI, –18–34%), respectively, with intermediate risks in the medium-exposure areas (Table 2). Only results for stillbirths were statistically significant.
Among the individual THMs, chloroform showed a similar pattern of risk for stillbirths and low and very low birth weight to that of TTHM, both for the overall summary estimates across the three regions and in each individual region. Concentrations of BDCM and total brominated THMs did not show any association with risk of stillbirths or low or very low birth weight (data not shown).
Analysis using smoothing splines showed that at concentrations up to approximately 80 μgL, the relationship between TTHM and each of the birth outcomes was consistent with linearity (at concentrations > 80 μg/L CIs were very wide, because this represented ≤5% of the births in each region) (plots not shown). Sensitivity analysis excluding births from wards where the proportion of ethnic minority groups was ≥20%, and use of empirical annual mean TTHM estimates did not materially alter the results (Toledano 2004). Similarly, multilevel modeling including random water zone effects had negligible impact on the regression coefficients and their standard errors (data not shown).
Discussion
This is the largest study yet conducted of the association between DBPs in the public water supply, as measured by TTHMs, and stillbirth and birth weight. In the United Utilities region, we found a trend of increasing prevalence of low and very low birth weight and stillbirth from low- to medium- to high-exposure areas, but this was not apparent in the other regions. There was also a socioeconomic deprivation gradient across exposure categories in this region. There was strong evidence of heterogeneity between water regions in the effect of exposure to TTHMs for low and very low birth weight but not for stillbirths. A random-effects model was therefore used to obtain an overall summary estimate of the exposure effect because it allows for different biases and unmeasured factors in the different study regions and incorporates the heterogeneity of effects in the analysis of overall risk associated with TTHM. In the random-effects analysis, we found small but statistically significant excess risk in the high-TTHM exposure areas for stillbirths.
This study is approximately twice as large as all the previous studies combined on low birth weight (Bove et al. 1995; Dodds et al. 1999; Gallagher et al. 1998; Jaakkola et al. 2001; Kallen and Robert 2000; Kanitz et al. 1996; Kramer et al. 1992; Savitz et al. 1995; Wright et al. 2003) and four times the size of all other studies combined on stillbirths (Aschengrau et al. 1993; Bove et al. 1995; Dodds et al. 1999, 2004; Kallen and Robert 2000). Although one of the main strengths of this study is its size, this and its retrospective nature simultaneously limit the options available for exposure assessment. Clearly, it is not possible to obtain individual tap water samples at each maternal residence, or direct measures of individual exposure, in such a large-scale study, and there is an inevitable trade-off between specificity of the exposure assessment and study power. For these reasons, most studies have used an ecologic measure for exposure assessment (Aschengrau et al. 1993; Bove et al. 1995; Hwang et al. 2002; Jaakkola et al. 2001; Kallen and Robert 2000; Kanitz et al. 1996; Klotz and Pyrch 1999; Kramer et al. 1992; Wright et al. 2003; Yang et al. 2000); some, like us, have incorporated modeled ecologic exposure estimates to improve the exposure classification (Dodds and King 2001; Dodds et al. 1999; Gallagher et al. 1998; King et al. 2000), and only a few have obtained individual-level exposure information (Dodds et al. 2004; Savitz et al. 1995; Shaw et al. 2003; Waller et al. 1998, 2001). To the extent that all these approaches are bound to lead to exposure misclassification, varying degrees of error (both Berkson and classical error) (Nieuwenhuijsen et al. 2000b) will result, leading to loss of power and/or bias in estimates of exposure–disease associations, most likely (but not necessarily) toward the null (no effect).
In this study we used modeled ecologic quarterly estimates of TTHM concentrations for all birth locations in the study, taking into account THM profiles commonly associated with particular water sources and seasonal variation (Whitaker et al. 2004) to provide an improved and more robust exposure assessment. One particular advantage is that the exposures are estimated with comparable precision across all the zones and quarters because of the hierarchical links built into the model, which is important given the variable number of raw measurements available in different zones. Nonetheless, inevitably there will be a degree of exposure misclassification because all mothers in one water zone were assigned the same (ecologic) exposure estimate. No account was taken of the potential mobility of mothers during pregnancy and consumption of water outside the home, other activities affecting THM exposure such as swimming, and possible variability in THM concentrations within a water zone.
The possibility of exposure measurement error from residential mobility during pregnancy cannot be ruled out, because an American study found that > 20% of pregnant women moved residence between the time of conception and delivery (Shaw and Halinka 1991). Of course, if mothers move but remain within the same water supply zone, this should not introduce substantial measurement error unless within-zone variability is greater than between-zone variability (which is not the case for our data). Mobility from zone to zone could also result from the home and workplace being in different water zones. However, recent research on tap-water–related activities among pregnant women in the United Kingdom suggests that possible consumption of water outside the home is unlikely to be a major source of exposure misclassification (Kaur et al. 2004). For example, on average, only 18% of total fluid ingestion by study participants was cold tap water, and only 30% of this tap water was consumed outside the home (Kaur et al. 2004). Moreover, women drank almost equal amounts of cold tap water and bottled water at home, but at work and elsewhere they drank almost three times more bottled water than cold tap water (Kaur et al. 2004). The effects of variations in individual behaviors (e.g., ingestion, showering and bathing habits) on actual THM uptake, and their implications for this epidemiologic study, have been explored in a simulation study. This showed that a moderate to strong correlation (~ 0.6–0.8) could be expected between concentrations of chloroform in tap water and actual uptake by pregnant women, even when there is no information on individual behavior (Whitaker et al. 2003a). Furthermore, analysis of THM data inone of our study water regions (United Utilities) showed that between-zone variation was consistently larger than within-zone variation for both chloroform and BDCM, the main THMs. This suggests that water zone means are a valid way of differentiating exposure to THMs between individuals (Keegan et al. 2001). Taken together, the above suggests that our methods provided a valid approach to estimating TTHM exposure of individuals for use in our epidemiologic study.
To date, total THMs have been the main focus of epidemiologic investigation. However, total THMs may not be a good marker of the individual THMs (e.g., brominated compounds) and other by-products (e.g., haloacetates) that have recently been implicated with respect to adverse birth outcomes (King et al. 2000; Klotz and Pyrch 1999; Swan and Waller 1998; Wright et al. 2003). For example, we found only a moderate correlation between total THMs and the various individual THMs (Keegan et al. 2001; Whitaker et al. 2003b). In our study of the individual THMs, we found an association with chloroform but not with the brominated compounds. Findings of our overall summary analyses reflected in particular trends in United Utilities region; although differences in results between our water regions might partly be accounted for by differing sociodemography, they might also have been caused by differing composition of the DBPs or the presence of other substances or factors that are strongly correlated with THMs in one region but not in the others.
An important issue is the extent to which our results might be explained by unmeasured or uncontrolled confounding. We had only limited data on potential confounders, and information on potentially important risk factors, such as maternal smoking habits and gestational age, was not available. Some previous studies have shown a much stronger association between TTHM exposure and low birth weight for term births only (Gallagher et al. 1998), whereas others have detected no consistent associations of low birth weight among all births or term births (Jaakkola et al. 2001; Wright et al. 2003). Others have observed an increased risk, in particular, of small size for gestational age with high TTHM exposure (Bove et al. 1995; Kramer et al. 1992; Wright et al. 2003). It is not yet clear, therefore, whether the underlying association between low birth weight and TTHM concentrations reflects a risk for babies born prematurely but of appropriate size for their gestational age or fetal growth retardation among babies born at term. Although some of the discrepancies between studies may have been due to differences in design, Wright et al. (2003) recently reported that confounding by gestational age had a substantial impact on the association between birth weight and TTHM concentrations. Unfortunately, we were unable to examine this in our study because data on gestational age are not included on the routine birth records.
The diverse etiologic routes to low birth weight might be a possible explanation for the observed heterogeneity in effect of TTHM on low and very low birth weight but not on stillbirths. For example, there could be differing proportions of small-for-gestational-age and low-birth-weight preterm births among the three study regions (e.g., reflecting differences in ethnic minority mix), with a stronger association of THM exposure with one of these pathways to becoming a low-birth-weight baby, but not the other.
Another possible explanation for the observed heterogeneity in effect of TTHM exposure on low and very low birth weight could relate to differences in baseline rates between the regions. Severn Trent and Northumbrian were found to have a higher prevalence of low and very low birth weight in the low-exposure areas than did United Utilities. If the effects of TTHM exposure are additive rather than multiplicative, this could lead to heterogeneity of relative effect measures such as ORs (Greenland and Rothman 1998b). However, this would not explain an apparent inverse association for very low birth weight in the high-exposure category seen in Severn Trent. The reasons for the different baseline rates across regions are unclear and merit further investigation.
We did have information on socioeconomic deprivation at small-area scale. In contrast to the other two regions, the high-exposure areas in the United Utilities region tended to be more deprived than the low-exposure areas. This was an unexpected finding. Both stillbirth and low birth weight are related to deprivation (higher rates among lower social classes) (Dummer et al. 2000; Nordstrom and Cnattingius 1996; Parker et al. 1994; Rodriguez et al. 1995). Comparison of ORs without adjustment for deprivation (Carstairs index) with those after adjustment in the United Utilities region showed that ORs were reduced by up to about one-half, suggesting the possibility of residual confounding. This was explored in more detail using data from the Health Survey for England (Erens and Primatesta 1997) on smoking habits and ethnicity for women of reproductive age living in the United Utilities region; these data showed that higher proportions of women of nonwhite origin (7.5 vs. 1.3%) and women who smoke (39.7 vs. 31.6%) resided in high- than in low-exposure areas. Analysis of data for London from the St. Mary’s Maternity Information System (Chapple 1997) showed increased relative risks (ranging from 1.4 to 3.2) among offspring of women who smoke and women who are of nonwhite origin, for each of the birth outcomes under study. Using these data, the higher proportion of nonwhite women living in areas of high compared with low TTHM concentrations in the United Utilities region would explain only around 13, 4, and 5% of the excess risk for stillbirth, low birth weight, and very low birth weight, respectively, whereas the higher proportion of women smokers living in areas of high compared with low TTHM concentrations would explain only around 3, 5, and 3%, respectively, of the excess risk (Toledano 2004). These excesses are generally less than or similar to the difference between the unadjusted and adjusted (for deprivation) risk estimates for each of the birth outcomes, suggesting that inclusion of the Carstairs index may have adequately adjusted for deprivation-related effects in the United Utilities region. Nonetheless, residual confounding by socioeconomic deprivation cannot be excluded. Excess risks in areas of high deprivation relative to areas of low deprivation (after adjustment for all other potential confounders and TTHM category) across the three water regions were, on average, 15–20 times the magnitude of those found in association with areas of high relative to low TTHM exposure, after adjustment for socioeconomic deprivation and other potential confounders.
If, however, our results reflect some causal association rather than confounding or other source of bias, what could be the potential mechanisms? The THMs have been studied in laboratory animals and appear to show little reproductive or developmental toxicity Nieuwenhuijsen et al. 2000a). In addition, recent studies found no association between swimming and excess risk of various birth outcomes (Klotz and Pyrch 1999; Nieuwenhuijsen et al. 2002; Waller et al. 1998), even though the potential for THM exposure and uptake during swimming may be high (Chu and Nieuwenhuijsen 2002; Whitaker et al. 2003a). Nevertheless, THMs may be acting as a surrogate measure for other chlorination by-products (e.g., the HAAs). These show some capacity for developmental effects but only at very high doses (Nieuwenhuijsen et al. 2000a). To date, they have not been a focus for epidemiologic investigation because of the lack of routinely collected data on these compounds. Klotz and Pyrch (1999) found only small associations of HAAs and haloacetonitriles with neural tube defects, but study power was low and CIs were wide. Other chlorination by-products, including the highly mutagenic chlorinated furanone MX, show little or no reproductive or developmental toxicity except at very high doses (International Programme on Chemical Safety 2000). However, not all potential chlorination byproducts have been identified yet. In addition, not all those that are known have been comprehensively studied for reproductive and developmental toxicity, and in most cases the substances have been studied separately rather than as a mixture, to which humans are generally exposed.
In summary, our findings overall suggest a significant association of stillbirths with maternal residence in high-TTHM exposure areas. Further work is needed to examine cause-specific stillbirths and effects of other DBPs and to explore the possibility of residual confounding at the individual level to help differentiate between alternative (non-causal) explanations and those that may be due to the water supply. The finding of significant heterogeneity between regions in the effect of TTHMs on risk of low and very low birth weight also deserves further study to understand better the reasons for heterogeneity, including possible differences in composition of other DBPs between water regions. Although the limited data from laboratory and epidemiologic studies do not so far indicate a causal association between exposure to THMs and stillbirth in humans, it would seem appropriate that water suppliers continue to follow the current policy of reducing THMs and other DBPs in public water supplies, as far as is consistent with maintaining effective control against waterborne microbiologic disease.
Figure 1 Locations of the study water company supply regions in Great Britain.
Figure 2 Maps showing water-supply-zone-level TTHM exposure categories for each quarter: Northumbrian Water, 1997: (A) January–March; (B) April–June; (C) July–September; (D) October–December.
Figure 3 Maps showing water supply-zone-level TTHM exposure categories for each quarter, United Utilities Water, 1997: (A) January–March; (B) April–June; (C) July–September; (D) October–December.
Figure 4 Maps showing water supply-zone-level TTHM exposure categories for each quarter, Severn Trent Water, 1997: (A) January–March; (B) April–June; (C) July–September; (D) October–December.
Table 1 Descriptive data for the study population, by water region and TTHM category, 1992–1998.
TTHM (μg/L)
Stillbirths
Low birth weight
Very low birth weight
Water region/TTHM category Carstairs score[mean (5th, 95th percentile)] Mean (5th, 95th percentile) No.a Prevalence(95% CI) No.a Prevalence(95% CI) No.a Prevalence(95% CI) No.a Birth weight[g; mean (5th, 95th percentile)]
Northumbrian
Low 1.68 (–2.83, 5.55) 18.0 (8.3, 29.0) 6 4.8 (1.0–8.6) 80 64.1 (50.5–77.7) 12 9.6 (4.2–15.0) 1,248 3,350 (2,380, 4,200)
Medium 1.53 (–3.30, 6.88) 48.1 (34.2, 58.9) 58 5.7 (4.2–7.1) 638 62.9 (58.2–67.6) 114 11.2 (9.2–13.3) 10,142 3,346 (2,410, 4,220)
High 1.54 (–3.62, 7.82) 71.5 (61.0, 88.2) 47 5.1 (3.7–6.6) 607 67.0 (61.8–72.1) 93 10.3 (8.2–12.3) 9,062 3,337 (2,380, 4,200)
Overall 1.54 (–3.45, 7.25) 56.6 (27.0, 81.1) 111 5.4 (4.4–6.4) 1,325 64.8 (61.4–68.2) 219 10.7 (9.3–12.1) 20,452 3,342 (2,390, 4,210)
United Utilities
Low −0.13 (–3.88, 5.82) 19.2 (6.4, 29.4) 192 4.3 (3.7–5.0) 2,665 50.6 (48.7–52.5) 405 7.7 (6.9–8.4) 52,662 3,396 (2,490, 4,260)
Medium 0.88 (–3.58, 7.58) 46.0 (32.6, 58.5) 1,194 5.3 (5.0–5.6) 15,882 59.9 (59.0–60.8) 2,336 8.8 (8.5–9.2) 265,030 3,356 (2,410, 4,220)
High 1.90 (–3.36, 8.21) 71.9 (60.8, 88.9) 824 5.8 (5.4–6.2) 10,197 68.0 (66.7–69.3) 1,521 10.1 (9.6–10.7) 149,905 3,326 (2,360, 4,200)
Overall 1.12 (–3.57, 7.71) 52.0 (19.0, 81.1) 2,210 5.4 (5.1–5.6) 28,744 61.5 (60.8–62.2) 4,262 9.1 (8.8–9.4) 467,597 3,351 (2,409, 4,220)
Severn Trent
Low 0.54 (–3.54, 6.84) 11.2 (2.2, 28.9) 920 5.1 (4.7–5.4) 11,401 63.5 (62.4–64.6) 1,786 9.9 (9.5–10.4) 179,605 3,343 (2,390, 4,220)
Medium 0.86 (–3.57, 7.99) 44.0 (31.3, 57.8) 1,233 5.3 (5.0–5.6) 14,845 64.4 (63.4–65.4) 2,290 9.9 (9.5–10.3) 230,653 3,331 (2,381, 4,203)
High 0.26 (–3.61, 6.13) 70.7 (60.7, 88.6) 378 5.2 (4.7–5.7) 4,326 60.9 (59.2–62.7) 610 8.6 (7.9–9.3) 70,997 3,344 (2,410, 4,220)
Overall 0.65 (–3.57, 7.41) 35.8 (2.8, 72.5) 2,531 5.2 (5.0–5.4) 30,572 63.5 (62.8–64.2) 4,686 9.7 (9.5–10.0) 481,255 3,337 (2,399, 4,220)
Data are mean Carstairs scores, TTHM (μg/L) concentrations, prevalence and 95% CIs of stillbirths per 1,000 total births, low and very low birth weight per 1,000 live births, and mean birth weight (g). Prevalence of stillbirths, mean Carstairs score (the lower the score, the more affluent the area), and mean TTHM were based on total births for Northumbrian, 1997; United Utilities, 1993–1997; and Severn Trent, 1993–1998. Birth weight variables were based on live births for Northumbrian, 1997; United Utilities, 1992–1997; Severn Trent, 1993–1998. TTHM was categorized as follows: low, < 30 μg/L; medium, 30–59 μg/L; and high, ≥60 μg/L.
a Number of stillbirths, low-birth-weight births, and very-low-birth-weight births, and for birth weight, number of live births.
Table 2 Adjusted ORsa (95% CIs) for stillbirths and low and very low birth weight by TTHM category and by water region and overall, 1992–1998.
Water region/TTHM category Stillbirthsb Low birth weight Very low birth weight
Northumbrian
Low 1.00 1.00 1.00
Medium 1.19 (0.51–2.75) 1.02 (0.80–1.30) 1.20 (0.66–2.18)
High 1.09 (0.46–2.55) 1.11 (0.87–1.41) 1.11 (0.61–2.03)
United Utilities
Low 1.00 1.00 1.00
Medium 1.16 (1.00–1.35) 1.11 (1.07–1.16) 1.09 (0.98–1.21)
High 1.21 (1.03–1.42) 1.19 (1.14–1.24) 1.20 (1.07–1.34)
Severn Trent
Low 1.00 1.00 1.00
Medium 1.03 (0.95–1.13) 1.00 (0.98–1.03) 1.00 (0.94–1.06)
High 1.04 (0.93–1.18) 0.98 (0.95–1.02) 0.90 (0.82–0.99)
Overall summaryc
Low 1.00 1.00 1.00
Medium 1.06 (0.99–1.15) 1.05 (0.96–1.15) 1.03 (0.96–1.10)
High 1.11 (1.00–1.23) 1.09 (0.93–1.27) 1.05 (0.82–1.34)
a ORs for stillbirths are adjusted for maternal age and Carstairs quintile and based on total births for Northumbrian, 1997; United Utilities, 1993–1997; and Severn Trent, 1993–1998. Regression analysis for birth weight variables is based on live births for Northumbrian, 1997; United Utilities, 1992–1997; and Severn Trent, 1993–1998. ORs for low birth weight are adjusted for maternal age, Carstairs quintile, sex of baby, and year of study (year was omitted in the case of Northumbrian). ORs for very low birth weight are adjusted for maternal age, Carstairs quintile, and year of study (year was omitted in the case of Northumbrian).
b Overall summary estimates for stillbirths are shown from the random-effects model for consistency with the birth weight estimates even though statistically significant heterogeneity between water regions was not found. However, results from a fixed-effects model were virtually identical.
c Overall summary estimates were obtained from random-effects model combining the region-specific exposure ORs allowing for heterogeneity between regions. p-Values for tests for heterogeneity (medium:low, high:low) from random-effects model were as follows: stillbirths (0.449, 0.339), low birth weight (0.000, 0.000), and very low birth weight (0.322, 0.001).
==== Refs
References
Aschengrau A Zierler S Cohen A 1989 Quality of community drinking water and the occurrence of spontaneous abortion Arch Environ Health 44 283 290 2554824
Aschengrau A Zierler S Cohen A 1993 Quantity of community drinking water and the occurrence of late adverse pregnancy outcomes Arch Environ Health 48 105 113 8476301
Bove FJ Fulcomer MC Klotz JB Esmart J Dufficy EM Savrin JE 1995 Public drinking water contamination and birth outcomes Am J Epidemiol 141 850 862 7717362
Box GEP Cox DR 1964 An analysis of transformations J R Stat Soc Series B 26 211 252
Carstairs V Morris R 1991. Socio-economic Status and Health in Scotland. Aberdeen:Aberdeen University Press.
Cedergren MI Selbing AJ Lofman O Kallen B 2002 Chlorination byproducts and nitrate in drinking water and risk of congenital cardiac defects Environ Res A 89 124 130
Chapple JC 1997. The north west Thames experience—the St Mary’s Maternity Information System (SMMIS). In: The Effective Use of Maternity Data (Steer PJ, Beard RW, eds). London:Clinical Accountability, Service Planning and Evaluation, 21–27.
Chen W Weisel C 1998 Halogenated DBP concentrations in a distribution system J Am Water Works 90 151 163
Chu H Nieuwenhuijsen MJ 2002 Distribution and determinants of trihalomethanes levels in UK indoor swimming pools Occup Environ Med 59 243 247 11934951
Dersimonian R Laird N 1986 Meta-analysis in clinical trials Control Clin Trials 7 177 188 3802833
Dodds L King W 2001 Relation between trihalomethane compounds and birth defects Occup Environ Med 58 443 446 11404448
Dodds L King W Allen AC Armson BA Fell DB Nimrod C 2004 Trihalomethanes in public water supplies and risk of stillbirth Epidemiology 15 2 179 186 15127910
Dodds L King W Woolcott C Pole J 1999 Trihalomethanes in public water supplies and adverse birth outcomes Epidemiology 3 233 237 10230830
Dummer TJ Dickinson HO Pearce MS Charlton ME Parker L 2000 Still birth risk with social class and deprivation: no evidence for increasing inequality J Clin Epidemiol 53 2 147 155 10729686
Erens B Primatesta P eds. 1997. Health Survey for England: 1996. Vols. 1 and 2. London:The Stationery Office.
Gallagher MD Nuckols JR Stallones L Savitz DA 1998 Exposure to trihalomethanes and adverse pregnancy outcomes Epidemiology 9 484 489 9730025
Greenland S Rothman KJ 1998a. Introduction to stratified analysis. In: Modern Epidemiology (Rothman KJ, Greenland S, eds). 2nd ed. Philadelphia:Lippincott Williams & Wilkins, 253–279.
Greenland S Rothman KJ 1998b. Measures of effect and measures of association. In: Modern Epidemiology (Rothman KJ, Greenland S, eds). 2nd ed. Philadelphia:Lippincott Williams & Wilkins, 47–64.
Hastie T Tibshirani R 1990. Generalized Additive Models. London:Chapman & Hall.
Hwang B-F Magnus P Jaakkola JJK 2002 Risk of specific birth defects in relation to chlorination and the amount of natural organic matter in the water supply Am J Epidemiol 156 374 382 12181108
International Programme on Chemical Safety 2000. Environmental Health Criteria 216: Disinfectants and Disinfection By-products. Geneva:World Health Organization.
Jaakkola JJK Hwang B-F Magnus P 2003 Risk of stillbirth in relation to water chlorination and the amount of natural organic matter [Abstract] Epidemiology 14 5 S55
Jaakkola JJK Magnus P Skrondal A Hwang B-F Becher G Dybing E 2001 Foetal growth and duration of gestation relative to water chlorination Occup Environ Med 58 437 442 11404447
Kallen BAJ Robert E 2000 Drinking water chlorination and delivery outcome—a registry based study in Sweden Reprod Toxicol 14 303 309 10908833
Kanitz S Franco Y Patrone V Caltabellotta M Raffo E Riggi C 1996 Association between drinking water disinfection and somatic parameters at birth Environ Health Perspect 104 516 520 8743439
Kaur S Nieuwenhuijsen MJ Ferrier H Steer P 2004 Exposure of pregnant women to tap water related activities in London and validation of the questionnaire Occup Environ Med 61 454 460 15090668
Keegan T Whitaker H Nieuwenhuijsen MJ Toledano MB Elliott P Fawell J 2001 The use of routinely collected trihalomethane drinking water data for epidemiological purposes Occup Environ Med 58 447 452 11404449
King WD Dodds L Allen AC 2000 Relation between stillbirth and specific chlorination by-products in public water supplies Environ Health Perspect 108 883 886 11017894
Kline J Stein Z Susser M 1989. Conception to Birth: Epidemiology of Prenatal Development. New York:Oxford University Press.
Klotz JB Pyrch LA 1999 Neural tube defects and drinking water disinfection by-products Epidemiology 10 383 390 10401872
Kramer MD Lynch CF Isacson P Hanson JW 1992 The association of waterborne chloroform with intrauterine growth retardation Epidemiology 3 407 413 1391132
Krasner SW McGuire MJ Jacaugelo JG Patania NL Reagen KM Aieta EM 1989 The occurrence of disinfection byproducts in U.S. drinking water J Am Water Works Assoc 81 41 53
Krasner SW Sclimenti MJ Means EG 1994 Quality degradation: implications for DBP formation J Am Water Works Assoc 86 34 47
Leyland AH Goldstein H 2001. Multilevel Modelling of Health Statistics. New York:John Wiley & Sons, 29.
Magnus P Jaakkola JJK Skrondal A Alexander J Becher G Krogh T 1999 Water chlorination and birth defects Epidemiology 10 513 517 10468423
Nieuwenhuijsen MJ Northstone K Golding J the ALSPAC team. 2002 Swimming and birthweight Epidemiology 13 725 728 12410017
Nieuwenhuijsen MJ Toledano MB Eaton NE Fawell J Elliott P 2000a Chlorination disinfection by-products in water and their association with adverse reproductive outcomes: a review Occup Environ Med 57 73 85 10711274
Nieuwenhuijsen MJ Toledano MB Elliott P 2000b Uptake of chlorination disinfection by-products; a review and a discussion of its implications for exposure assessment in epidemiological studies J Expos Anal Environ Epidemiol 10 586 599
Nordstrom ML Cnattingius S 1996 Effects on birth weights of maternal education, socio-economic status, and work-related characteristics Scand J Soc Med 24 1 55 61 8740877
Parker JD Schoendorf KC Kiely JL 1994 Associations between measures of socioeconomic status and low birth weight, small for gestational age, and premature delivery in the United States Ann Epidemiol 4 4 271 278 7921316
Pless IB 1994. Epidemiology of Childhood Disorders. New York:Oxford University Press.
Rodriguez C Regidor E Gutierrez-Fisac JL 1995 Low birth weight in Spain associated with sociodemographic factors J Epidemiol Community Health 49 1 38 42 7707003
Savitz DA Andrews KW Pastore LM 1995 Drinking water and pregnancy outcome in central North Carolina: source, amount and trihalomethane levels Environ Health Perspect 103 592 596 7556013
Shaw GM Halinka LH 1991 Residential mobility during pregnancy for mothers of infants with or without congenital cardiac anomalies Arch Environ Health 46 310 312
Shaw GM Ranatunga D Quach T Neri E Correa A Neutra R 2003 Trihalomethane exposures from municipal water supplies and selected congenital malformations Epidemiology 14 191 199 12606885
Singer PC 1999 Humic substances as precursors for potentially harmful disinfection by-products Water Sci Technol 40 25 30
Spiegelhalter D Thomas A Best N Gilks W 1996. BUGS 0.5, Bayesian Inference Using Gibbs Sampling Manual (Version ii). Cambridge, UK:Medical Research Council Biostatistics Unit. Available: http://www.mrc-bsu.cam.ac.uk/bugs [accessed 6 January 2005].
Stevens AA Moore LA Miltner RJ 1989 Formation and control of non-trihalomethane disinfection by-products J Am Water Works Assoc 81 54 60
Toledano MB 2004. Chlorination By-products and Adverse Birth Outcomes: A Retrospective Semi-ecologic Study in Three Regions in England [PhD Thesis]. London:University of London.
Venables WN Ripley BD 2002. Modern Applied Statistics with S. 4th ed. New York:Springer.
Waller K Swan SH DeLorenze G Hopkins B 1998 Trihalomethanes in drinking water and spontaneous abortion Epidemiology 9 134 140 9504280
Waller K Swan S Windham GC Fenster L 2001 Influence of exposure assessment methods on risk estimates in an epidemiologic study of trihalomethane exposure and spontaneous abortion J Expos Anal Environ Epidemiol 11 522 531
Whitaker H Best N Nieuwenhuijsen MJ Wakefield J Fawell J Elliott P 2004. Modelling exposure to disinfection byproducts in drinking water for an epidemiological study of adverse birth outcomes. J Expos Anal Environ Epidemiol 10.1038/sj.jea.7500380.
Whitaker H Nieuwenhuijsen MJ Best N 2003a The relationship between water chloroform levels and uptake of chloroform: a simulation study Environ Health Perspect 111 688 694 12727595
Whitaker H Nieuwenhuijsen MJ Best N Fawell J Gowers A Elliott P 2003b Description of trihalomethane levels in three UK water companies J Expos Anal Environ Epidemiol 13 17 23
Wright JM Schwartz J Dockery DW 2003 Effect of trihalomethane exposure on fetal development Occup Environ Med 60 173 180 12598663
Yang C-Y Cheng B-H Tsai S-S Wu T-N Lin M-C Lin K-C 2000 Association between chlorination of drinking water and adverse pregnancy outcome in Taiwan Environ Health Perspect 108 765 768 10964797
|
15687062
|
PMC1277869
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 21; 113(2):225-232
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/ehp.7111
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences 10.1289/txg.7273ehp0113-00023315687063ToxicogenomicsArticleGlobal Gene Expression Profiling in Whole-Blood Samples from Individuals Exposed to Metal Fumes Wang Zhaoxi 1Neuburg Donna 2Li Cheng 2Su Li 1Kim Jee Young 1Chen Jiu Chiuan 1Christiani David C. 131Department of Environmental Health, Occupational Health Program, Harvard School of Public Health, Boston, Massachusetts, USA2Department of Statistical Sciences, Dana Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts, USA3Pulmonary and Critical Care Unit, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USAAddress correspondence to D. Christiani, 665 Huntington Ave., I-1402, Boston, MA 02115 USA. Telephone: (617) 432-3323. Fax: (617) 432-3441. E-mail:
[email protected] work was supported by grant T32 ES07069 from the National Institute of Environmental Health Sciences and by research grants ES009860, CA074386, CA090578, and CA092824 from the National Institutes of Health.
The authors declare they have no competing financial interests.
2 2005 22 11 2004 113 2 233 241 21 5 2004 22 11 2004 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright. Accumulating evidence demonstrates that particulate air pollutants can cause both pulmonary and airway inflammation. However, few data show that particulates can induce systemic inflammatory responses. We conducted an exploratory study using microarray techniques to analyze whole-blood total RNA in boilermakers before and after occupational exposure to metal fumes. A self-controlled study design was used to overcome the problems of larger between-individual variation interferences with observations of relatively smaller changes caused by environmental exposure. Moreover, we incorporated the dichotomous data of absolute gene expression status in the microarray analyses. Compared with nonexposed controls, we observed that genes with altered expression in response to particulate exposure were clustered in biologic processes related to inflammatory response, oxidative stress, intracellular signal transduction, cell cycle, and programmed cell death. In particular, the preinflammatory cytokine interleukin 8 and one of its receptors, chemokine receptor 4, seemed to play important roles in early-stage response to heavy metal exposure and were down-regulated. Furthermore, most observed expression variations were from nonsmoking exposed individuals, suggesting that smoking profoundly affects whole-blood expression profiles. Our study is the first to demonstrate that with a paired sampling study design of pre- and postexposed individuals, small changes in gene expression profiling can be measured in whole-blood total RNA from a population-based study. This technique can be applied to evaluate the host response to other forms of environmental exposures.
functional pathwaygene expression profilinginflammationoccupational particulate exposurewhole-blood total RNA
==== Body
Exposure to ambient particulate air pollution is associated with increases in morbidity and mortality from respiratory and cardiovascular diseases (Godleski et al. 2000). The welding process generates high levels of metal fume containing respirable particles. Epidemiologic studies have shown that acute exposure to welding fume is associated with metal-fume fever (Mueller and Seger 1985) and increased reversible respiratory symptoms (El-Zein et al. 2003a; Wolf et al. 1997). There was an increased prevalence of inflammatory lung diseases, such as asthma and chronic bronchitis, among welders (El-Zein et al. 2003b). Additionally, accumulating epidemiologic evidence in the last decade has pointed to the associations of particulate exposure with adverse cardiovascular effects (Dockery et al. 1993; Mann et al. 2002; Peters et al. 2000, 2001a; Pope et al. 2002). Limited evidence indicates that welding-fume exposure also may be associated with increased cardiovascular events (Sjogren et al. 2002).
It has been proposed that inhaled particulates from air pollution may cause systemic alterations by the release of inflammatory cytokines subsequent to pulmonary inflammation, which plays an important role in the pathogenesis of atherosclerosis and coronary diseases. Indeed, elevated ambient particulate levels have been shown to be associated with increased levels of inflammatory markers, such as white blood cell (WBC) counts (Schwartz 2001), C-reactive protein (CRP; Peters et al. 2001b; Seaton et al. 1999), and fibrinogen (Pekkanen et al. 2000; Schwartz 2001) in both cross-sectional and longitudinal epidemiologic observations. In the experimental setting, animal studies have revealed that concentrated ambient particulate exposures increase the total WBC counts and the differential count of circulating neutrophils (Clarke et al. 2000; Gordon et al. 1998) in both healthy animals and those with pulmonary hypertension. Intratracheal instillation of residual oil fly ash (ROFA) can induce a significant elevation of plasma fibrinogen in cardiopulmonary-compromised rats (Gardner et al. 2000). Suwa et al. (2002) in their important work showing progressive atherosclerosis related to particulate exposure in hyperlipidemic rabbits also noted an increase in circulating polymorphonuclear leukocyte counts caused by exposures to particulate matter (PM) with a mass median aerodynamic diameter ≤10 μm (PM10).
However, most previous studies evaluated only downstream markers for systemic inflammatory responses. Direct human evidence is still lacking that shows particulates can induce systemic inflammation, although previous human studies and animal experiments did generate data, suggesting the involvement of inflammatory responses in particulate-mediated acute cardiac events. If particulate-mediated systemic inflammation were responsible for the observed adverse effects on the cardiovascular system, we would expect to see corresponding changes in mRNA expression for particulate-mediated systemic inflammation. The study described in this article addresses this mechanistic gap by investigating the systemic inflammatory response to welding-fume exposure using cDNA microarray technology on whole-blood total RNA. Blood samples were collected from welders and nonwelding controls before and after the work shift. We hypothesized that welding-fume exposure would be associated with systemic inflammation, as indicated by the findings that genes involved in systemic inflammation have significantly altered expressions. Furthermore, previous epidemiologic studies have shown that cigarette smoking significantly affects CRP, fibrinogen, and WBC levels (Frohlich et al. 2003; Smith et al. 2003). Therefore, we also hypothesized that smoking status would significantly affect the association between welding fume and the various systemic inflammatory gene expressions.
Microarray technology provides a format for the simultaneous measurement of the expression of thousands of genes in a single experimental assay and quickly becomes one of most the powerful and versatile tools for genomics and biomedical research (Murphy 2002). Peripheral blood is an essential tissue type for biomedical and clinical research because of its critical roles in immune response and metabolism. Furthermore, considering the simplicity and ease of collection, peripheral blood is also essential for discovery of biomarkers of hematologic diseases and surrogate markers of a wide range of nonhematologic disorders. Thus, applying microarray technology on peripheral blood may provide new insights of variations in global gene expression specifically associated with states of normal and disease and has the potential of applying the technology in disease detection and diagnosis. However, with the challenges unique to the blood sample, including complex composition of heterogeneous cell types and ex vivo changes of expression profiles induced by different handling and processing methods, it is difficult to apply microarray technology on whole-blood total RNA, and there are few previous publications of such research. To this end, this study is also an exploratory research with the purpose of developing proper methods for applying microarray technology on whole-blood total RNA.
Materials and Methods
Study population.
The study was approved by the institutional review board of the Harvard School of Public Health, and written informed consent was obtained from each subject. The study population consisted of 28 welding apprentices, instructors, and union officers, recruited and monitored at an apprentice welding school (Union Local 29, Quincy, MA). All 18 exposed subjects actively welded in the workshop, whereas 10 nonexposed controls stayed in the office or classroom of the same building during the work period. Blood samples were collected from each subject before and after the welding workshop. A self-administered questionnaire was used to obtain relevant information, including respiratory symptoms and diseases, smoking history, and occupational history. Exposure to fine particulate matter (particulate matter with a mass median aerodynamic diameter ≤2.5 μm, PM2.5) was assessed using KTL cyclones (GK2.05SH; BGI Inc., Waltham, MA). The air sample was collected on a 37-mm polytetrafluoroethylene membrane filter (Gelman Laboratories, Ann Arbor, MI), and the mass concentration was determined as previously described (Kim et al. 2003).
Blood measurements.
Complete blood counts of all blood samples were carried out at Path Lab Inc. (Portsmouth, NH). The blood parameters included total WBC count with differential, red blood cell count, platelet count, hemoglobin, hematocrit, and erythrocyte indices (mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, and red cell distribution width).
RNA preparation.
Immediately after the blood was drawn, we added TRI Reagent BD (Molecular Research Center, Inc., Cincinnati, OH) and mixed to stabilize the whole-blood total RNA. The stabilized samples were transported to our laboratory on dry ice and stored at −80°C until RNA extraction. Total RNA was isolated later from 10 mL of whole blood according to manufacturer protocols and purified using the RNeasy mini kit (Qiagen, Chatsworth, CA). The yield and quality of RNA were assessed by spectrophotometry and the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).
Microarray hybridization.
For genomewide expression profiling, we used Affymetrix Human Genome U133A GeneChips (Affymetrix, Santa Clara, CA), which allow detection of approximately 22,215 gene expression probe sets. All RNA samples were sent to and analyzed at the Microarray Core Facility of the Dana-Farber Cancer Institute (Boston, MA), according to the manufacturer’s manual. The baseline and postexposure RNA samples from each subject were processed together in one batch of microarray analysis to minimize inherent variations. The quality of microarrays analysis was initially assessed by examination of the 3′ to 5′ ratios of five housekeeping controls on U133A GeneChips.
Normalization and data extraction.
We used DNA-Chip Analyzer 1.3 (dChip; http://www.dchip.org/) software to normalize the raw microarray signals and then calculate the model-based expression values using a default perfect-match–only model with outlier detection. dChip software applied an invariant set normalization method, which chose a subset of perfect-matched probes with small within-subset rank difference in the two microarrays to serve as the basis for fitting a normalization curve (Li and Wong 2001a, 2001b). The outlier detection algorithm of the dChip software allowed further quality assessment of microarray data by cross-referencing one array with other arrays through a modeling approach to identify problematic arrays (Li and Wong 2001b). To have a better fit in the model for more precise estimations of expression values, we included 10 additional microarrays in data normalization and extraction. The detection of whether a gene was expressed (present) or not expressed (absent) in a RNA sample was carried out by Affymetrix Microarray Suite (MAS) 5.0 software (Affymetrix) using one-sided Wilcoxon’s signed-ranked algorithm (Detection Calls; Liu et al. 2002).
Microarray data analysis.
Initially we evaluated gene expression changes by comparing the large fold-changes of expression values between baseline and postexposed microarrays in both exposed (welders) and nonexposed (controls) subjects. Then, we focused on using the paired t-test in the dChip software package to flush out genes with small expression changes in response to metal particulate exposure. The results of the paired t-test were adjusted by standard errors associated with each gene expression value. Because dChip software only gave an expression value to each gene on an array without discriminating whether the gene was expressed, we attempted to incorporate the Detection Calls information generated from the Affymetrix MAS 5.0 software package in the data analyses.
Hierarchical clustering.
The analyses were carried out by dChip software using a hierarchical clustering algorithm (Eisen et al. 1998) with average-linkage method. After linear transformation to standardize the expression values across all selected samples, the distance between two genes was calculated as 1-absolute standard correlation coefficient and was used in the subsequent repeated process to build phylogenetic tree of genes and samples.
Clustering analyses using gene ontology.
Testing for significant change in a single gene is difficult to accomplish given the stringent criteria for significance in the multiple-test background of > 20,000 probe sets. Therefore, we focused instead on assessing the biologic functions enriched with genes identified from the paired t-test, using the annotations defined by the Gene Ontology Consortium (GO; http://www.geneontology.org/; Hakak et al. 2001). The GO annotations are structured, controlled vocabulary for describing the roles of genes in any organism. The probability of observing a particular number of genes in one GO biologic process (bioprocess) was tested using hypergeometric distribution as previously described (Tavazoie et al. 1999). Briefly, we addressed the problem as what is the probability of observation of at least (x) genes of a certain GO bioprocess annotation in a list of (k) genes from paired t-test results, given the background that there are (n) genes with the same GO annotation from total (m ) genes (total annotated genes or a subclass of GO annotated genes on the entire Affymetrix array). The p-values were calculated using the following formula:
In this study, we focused only on the gene annotations of GO biologic process and used Affymetrix nonredundant build of human GO annotations downloaded 18 May 2004. The lists of genes were uploaded to Affymetrix NetAffx Analysis Center (http://www.affymetrix.com/analysis/index.affx) to obtain the numbers of genes within each GO bioprocess.
A potential problem of significance testing using GO annotations is that the hypergeometric distribution p-values are biased and sensitive to the total genes (m) used in the tests, which are not truly representing the entire genome because of the selection biases in array design and incomplete process of GO annotation. Furthermore, the problem of multiple testing is difficult to adjust because the GO bioprocesses are highly interrelated and genes are often assigned into multiple GO bioprocesses. Because GO has a multiple-level structure of directed acyclic graphs with each level of bioprocess linked through multiple parent–child relationships, there are one or more pathways that could be identified by tracing back from any GO bioprocess to the top using true-path–rule logic relationships. Thus, we adopted a conservative approach of testing the hypergeometric distributions. First, we used three numbers of the total genes (m) corresponding to the top three levels of GO bioprocess. A GO bioprocess was regarded as significant when it had a p < 0.005 at the lowest testing level and a p < 0.05 at the immediate upper level. A functional pathway was regarded as significant when it had three consecutive GO bioprocesses tested significantly or had two consecutive significant bioprocesses but also tested significantly in other pathway(s). The results were visualized using GoSurfer Soft Mining Tool (https://www.affymetrix.com/analysis/query/go_analysis.affx).
Statistical analysis.
Statistical analyses were performed using SAS version 6.12 (SAS Institute Inc., Cary, NC). Exposure status was dichotomized as nonexposed controls and welders. Study population characteristics between controls and welders were compared using two-sample t-tests, Wilcoxon rank-sum tests with exact p-values, and Fisher’s exact test. The mean (SD) values of the PM2.5 concentrations were determined for controls and welders. Wilcoxon rank-sum tests with exact p-values were performed to compare the PM2.5 concentrations in controls and welders and also in smokers and nonsmokers. To account for the repeated measurements, linear mixed models with an interaction term for exposure status and smoking status were used for analysis. A generalized autoregressive covariance structure was used to account for the exponential decay of the correlation function as the interval between the measurements increases (Verbeke and Molenberghs 1997). Restricted maximum likelihood was used to estimate the covariance parameters. Baseline mean (SEM) levels of systematic inflammatory markers in peripheral blood were calculated in controls and welders according to smoking status. Linear mixed models were used to compare the baseline levels of the systemic inflammatory markers in controls and welders and in smokers and nonsmokers. The effect of age on the baseline levels of the systemic inflammatory markers also was investigated. The level of significance for all analyses was set at 0.05.
Results
Study population characteristics.
A combined approach of using intraindividual (self-pairing samples) and interindividual controls was implemented in the study to a) minimize the biologic variability among individuals and b) compare more precisely the gene expression profiles of different exposure states. Eighteen welders and 10 nonexposed controls at an apprentice welding school were recruited, and blood samples were collected at two time points: baseline and after 5.8 ± 0.6 hr of exposure to welding fume. After microarray hybridization, we had complete data sets on 44 arrays from 15 welders and 7 controls available for subsequent analysis. Among the 6 excluded study subjects, 1 withdrew during the study, 4 had low yield or poor quality of RNA extraction, and 1 had a poor quality of microarray hybridization. Population demographic data are summarized in Table 1. All study subjects were male, including 18 Caucasians, 3 Hispanics, and 1 African American. The age and smoking status were comparable between exposed and control groups.
During the welding workshop, the welders were exposed to metal fume and airborne PM from shielded metal arc welding, gas tungsten arc welding, plasma arc cutting, and grinding, with carbon steel being the most commonly used base metal. The controls were exposed primarily to ambient levels of PM while performing bookwork and office tasks at the welding school. In this study, particulate samples were collected from all controls and welders. With comparable mean sampling times between controls and welders, the median PM2.5 concentrations of welders were significantly higher than those of nonexposed controls (p < 0.01). However, there were no significant differences in PM2.5 concentrations according to smoking status in welders (p = 0.9). Previous occupational exposures, as measured by years of boilermaking, were not significantly different between controls and welders (Table 1). Moreover, before the day of sample collection, all controls and 12 of 15 welders had at least a 5-day period of non-welding or nonboilermaking to wash out the effects of previous metal-fume exposure. Among three welders with shorter than 5-day washout periods, two performed welding 1 day before the welding workshop and one performed welding 2 days before the workshop.
Systemic inflammatory marker levels in peripheral blood.
All study subjects, including both controls and welders, had their blood cell counts within the normal ranges and had similar baseline profiles of major systematic inflammatory markers. Controls and welders were not found to have significantly different mean baseline CRP (p = 0.4), fibrinogen (p = 0.8), absolute neutrophil count (p = 0.1), and absolute WBC counts (p = 0.1). However, smokers were found to have significantly higher mean baseline WBC (p < 0.01) and neutrophil (p < 0.001) levels than nonsmokers, among welders as well as in the entire studied population.
The changes of the systemic inflammatory markers across two time points were not significant in controls except for a significant increase of fibrinogen [25 mg/dL; 95% confidence interval (95% CI), 4–45] in the post-exposure measurements. In contrast, there was a significant increase in total WBC counts (mean change, +1.2 ×103/μL; 95% CI, 0.6–1.8) in nonsmoking welders but not in smoking welders (mean change, +0.3 × 103/μL; 95% CI, −1.3 to 1.8). Relative and absolute neutrophil counts were also increased significantly in nonsmoking welders (p < 0.02) but not in smoking welders (p > 0.7). The change profiles of CRP levels were opposite those of WBC and neutrophil counts, with a significant increase in smokers (p = 0.02) and a nonsignificant change in nonsmoking welders (p = 0.4). Fibrinogen levels did not change significantly between postexposure and baseline in both smoking and nonsmoking welders (p ≥0.6). Overall, our observations of acute metal-fume exposure were consistent with previous epidemiologic findings that increased levels of inflammatory markers were associated with elevated ambient particulate levels (Pekkanen et al. 2000; Peters et al. 2001b; Schwartz 2001; Seaton et al. 1999)
Finding genes with large expression variations by fold-change analysis.
Initially, we tried to find genes with large alterations of expressions between baseline and postexposed microarrays by comparing the large n-fold-changes of expression values in both exposed (welders) and nonexposed (controls) subjects. In both the welder and control groups, there was no gene with a 2-fold greater difference of the mean expression levels between baseline and postexposure arrays and an absolute difference > 50. Moreover, for each pair of baseline and postexposed arrays from the same subject, we found that few genes had large fold-changes (median number of genes, 20; range, 1–123) regardless of exposure status. In addition, the correlation coefficients of the raw expression values across entire probe sets were high between baseline and postexposed arrays from the same subject (median, 0.971; range, 0.949–0.988). These observations suggested that the real signals of changes in gene expression profiling in response to occupational metal exposure were very small, which could be the compound results of mixed cell types and large amounts of hemoglobin RNA in the whole-blood samples.
Identifying genes with altered expressions by paired t-test.
When all 22,215 probes on the U133A GeneChip were included in the paired t-test, we found more genes (p < 0.05) in welders (533 genes from 546 probes) than in controls (86 genes from 88 probes) (Table 2). Considering the absolute gene expression status, we further found that probes identified by the paired t-test in controls had a larger proportion of noninformative probes (60.5%) that had absent calls assigned by the Detection Calls algorithm in every tested array compared with those in welders (47.3%). Regarding the entire set of probes on GeneChip, our data set had an overall 49.0% of noninformative probes among all baseline arrays. The initial observations suggested there were only random variations and no statistically significant changes in whole-blood expression between postexposed and baseline samples in individuals without metal particulate exposure. We then conducted a series of paired t-tests in several subsets of genes, which had Present calls in at least one, 10%, 25%, and 50% arrays. With the increase of Present calls, the number of genes identified by paired t-tests dropped, but the difference in the numbers of identified genes between welders and controls increased (Table 2). Taken together, consistent findings of more genes identified by paired t-tests in welders than in controls suggested there were alterations of global gene expression profiling in the whole-blood total RNA in response to acute metal-fume exposure. In addition, only one gene, RIO kinase 3 (RIOK3), was identified and down-regulated in both welders and controls.
Sample clustering using genes identified in paired t-tests.
Genes identified by paired t-tests were used to classify RNA samples in hierarchical clustering analyses to further evaluate the expression patterns in samples categorized by different collection time points, smoking status, and metal-fume exposure status. We tested various lists of genes obtained from paired t-tests in controls, welders, nonsmoking welders, and smoking welders on the original expression data of baseline and postexposure arrays, as well as the data of log2-transformed expression ratios of postexposure over baseline. The clustering results neither revealed any distinct pattern of gene expressions with any kind of combination of selected genes and RNA samples nor showed any subgroup of samples or genes with similar expression patterns. However, in general, we found that > 70% of samples had the baseline and postexposure arrays of the same individual always clustered next to each other, regardless their exposure status (Figure 1). When RNA samples of non-smoking controls and welders were clustered with genes identified from paired t-tests of nonsmoking welders, all study participants had their baseline and postexposure arrays grouped together in the phylogenetic tree of sample clustering. Furthermore, samples of controls and welders seemed to be randomly mixed in any sample clustering analyses, including those using the data of log2-transformed expression ratios (data not shown). These observations further demonstrate that the real signals of gene expression changes caused by occupational metal exposure were smaller than the interindividual variations.
Functional clustering using gene ontology.
Next, the genes identified by paired t-tests were evaluated by hypergeometric distribution testing based on GO annotations to define any bioprocesses enriched with the identified genes. To minimize the noise of the false-positive genes on the paired t-test, we applied a set of highly stringent criteria to define the significant GO bioprocesses and functional pathways and further observed the trends and distribution of the significant bioprocesses in four subsets of genes, with increasing percentage of Present calls among all arrays (at least one, 10%, 25%, and 50% arrays). The results are shown in Figure 2. With a decrease in the available numbers of genes and an increase in the percentages of Present calls, the main structures of GO bioprocesses were preserved in both welders and controls except for a few low-level bioprocesses that disappeared. In the nonexposed group, we did not find that genes were significantly enriched in any functional pathways except for two statistically significant bioprocesses: response to DNA damage stimulus (GO ID 6974) and nucleotide-excision repair (GO ID 6289). These two bioprocesses also existed in the welders but were not statistically significant. However, in contrast to the controls, in the welder group many GO bioprocesses were found to be significantly enriched with genes having significant alterations of expression after exposure to metal fume. Some of the GO bioprocesses tested significantly across all subsets with different Present calls. In subsets including genes with lower Present calls, the significant bioprocesses were distributed more discretely, with fewer functional pathways identified. With the increase of Present calls, more significant functional pathways showed up in welders by connecting discrete bioprocesses with newly appeared ones. In the subset of at least 50% Present calls, most significant bioprocesses were in the interconnected functional pathways. In the metal-exposed welders, functional pathways related to nucleic acid metabolism (including RNA metabolism and DNA metabolism), and cellular morphogenesis disappeared with the increase of Present calls.
We identified eight functional pathways with significant enrichment of genes having altered expressions in response to metal-fume exposure in the subset of genes having Present calls in > 50% arrays (Table 3). These functional pathways contained many GO bioprocesses related to proinflammatory and immune responses, oxidative stress, phosphate metabolism, cell proliferation, and programmed cell death. Moreover, we identified 35 genes from these significant pathways that had altered expression levels in welding fume–exposed individuals in comparison with their own baseline samples (Table 4). Among the identified genes, we found several genes involved in every aspect of the inflammatory response, including proinflammatory mediators, cytokine receptors, downstream signal transduction genes, and cytotoxic granulysin.
Smoking effects on gene expression profiling.
We assessed further the effects of smoking on acute particulate exposure expression profiles. Of 15 welders and 7 nonexposed controls, there were 6 smoking welders and 1 smoking control. It appeared that most observed expression alterations were from nonsmokers exposed to welding fume because the number of genes identified from the paired t-test and the cluster of genes in GO bioprocesses were comparable between this subgroup of welders and the entire welding group (Table 5). In contrast to nonsmoking welders, fewer genes were identified from the paired t-test in welding smokers, and they had different patterns of gene clustering. A similar finding was observed in the analysis of the peripheral WBC count as described in the preceding section, and our results suggest that smoking may alter expression profiles in whole-blood total RNA and is a confounding factor in the study of particulate exposure-induced gene expression profiling changes.
Discussion
In the present study, small expression alerations in several genes, caused by short-term occupational exposure to metal particulates, could be detected in whole-blood total RNA by paired t-tests. Based on GO annotations, the significant genes were clustered in functional pathways related to proinflammatory and immune responses, oxidative stress, phosphate metabolism, cell proliferation, and programmed cell death, suggesting systemic reactions in peripheral blood in response to environmental particulate exposure. Moreover, the observations were confounded by smoking because most variations were observed in non-smoking welders exposed to welding fume.
Accumulating evidence proved that microarray technology for the investigation of global gene expression profiling is a powerful tool for basic biologic research and laboratory investigations of patient materials, especially in the field of cancer research and toxicology. Although this technology had been successfully applied on fractionated blood samples (Klein et al. 2001; Locati et al. 2002) such as peripheral blood mononuclear cells (PBMCs), successful studies of gene expression profiles in whole-blood total RNA have been limited because of the difficult challenges of heterogeneous cell types and potential ex vivo changes from blood handling and processing. Compared with fractionated blood samples, whole-blood total RNA had lower detection sensitivities mainly caused by a large amount of hemoglobin RNA from reticulocytes, which contributes up to 70% of the total RNA isolated from whole blood (Affymetrix 2003a, 2003b). PBMCs have a more uniform cell population, containing lymphocytes and monocytes but excluding red blood cells and granulocytes (eosinophils, basophils, neutrophils), and are the most transcriptionally active cells in blood (DePrimo et al. 2003). However, the extra fractionation procedure for PBMCs requires a prolonged period before RNA stabilization, which results in significant ex vivo changes in gene expression profiling (Affymetrix 2003a; Pahl and Brune 2002). In this study, because all blood samples were collected within 1 day, it was beyond the capacity of our laboratory to fractionate all blood samples in a timely fashion. Thus, the whole-blood total RNA was extracted and applied in all subsequent microarray assays.
Compared with person-to-person variations of gene expressions, the exposure-induced gene expression changes were smaller. Regardless of exposure status, a pair of baseline and postexposed microarrays of the same subject often had a higher correlation coefficient of raw signals across entire probe sets than a pair of baseline microarrays randomly selected, and most pairs were clustered next to each other in sample clustering analyses. In addition, excess hemoglobin RNA and mixed cell types in the whole blood made it more difficult to observe the real changes in gene expression profiles. Under such circumstances, we were able to control better the biologic variability among individuals and obtain more sensitive and precise measurements on gene expression profiles by using self-paired controls. In our experiments, this test identified more genes in the exposed group (139 genes) than in nonexposed controls (17 genes), with Present calls in at least 50% arrays.
Affymetrix U133A GeneChip contains > 20,000 probes for measuring gene expressions in a single hybridization experiment. One major issue in data analysis is to determine whether changes in gene expression are experimentally significant, with the background of thousands of individual genes tested simultaneously. On a GeneChip, many genes are functionally interrelated or have unknown functions, and there are multiple probe sets detecting the same gene. In addition, the weak signals of exposure-induced changes made it very difficult, or even impossible, to conduct a valid multiple testing adjustment. With these considerations, we did not perform any adjustments on the results of the paired t-test in the present study.
An alternative approach in the statistics of multiple testing is to estimate the false discovery rate (FDR) by random permutations within the same data set (Tusher et al. 2001). We estimated the FDR of paired t-test results by permutating each pair of baseline and postexposed arrays 500 times using dChip 1.3 software. There was a lower FDR (median, 30.2%) in the exposed welders than in non-exposed controls (median, 112%), suggesting that the paired t-test results of the exposed group contained genes with real changes in expressions in response to occupational exposure. However, the permutation tests through dChip software did not adjust for the problem that multiple probe sets detect the same gene on a GeneChip, so the estimated FDRs could be inflated. Nevertheless, knowing that an approximately 30% FDR was associated with a set of genes from the paired t-test limits our ability to identify individual genes with statistically significant changes in expression in response to particulate exposure. Instead of further testing the significant change in a single gene, we focused on identifying significant pattern changes of biologic process in the genes identified from the paired t-test, using the annotations defined by GO. The underlying hypothesis is that several genes of one functional bioprocess change their expressions in response to environmental challenge because genes are highly networked and coordinated and do not act alone. Although one gene change may be small and difficult to be detected accurately in a significance test, the significant enrichment of genes with small changes in a biologic process and a functional pathway may be assessable.
In this study we identified 35 genes from eight significant functional pathways that had altered expression levels after metal-fume exposure. The most interesting finding was the identification of several genes involved in every aspect of the inflammatory response, including proinflammatory mediators, cytokine receptors, downstream signal transduction genes, and cytotoxic granulysin. Five genes (IL8, IL1A, CXCR4, RALBP1, and SCYE1) have been implicated in chemotaxis of the early inflammatory response, especially IL8, which is a critical mediator for neutrophil-dependent acute inflammation (Mukaida 2000, 2003). IL8 has a wide range of actions on different cell types, including neutrophils, lymphocytes, monocytes, endothelial cells, and fibroblasts. IL8 is produced from various cell types in response to a wide variety of stimuli, including proinflammatory cytokines, microbes and their products, and environmental changes such as hypoxia, reperfusion, and hyperoxia. Previous studies on ROFA-exposed workers found an increase in proinflammatory cytokines and polymorphonuclear cells in the nasal lavage fluid, indicating that the particulate exposure resulted in acute upper airway inflammation (Hauser et al. 1995; Woodin et al. 1998). In our study, IL8 and other cytokines and receptor genes were transcriptionally down-regulated in whole-blood total RNA in response to metal particulate exposure.
Our findings that genes with altered expressions in whole-blood total RNA in response to metal particulate exposure were clustered in the functional pathways related to inflammatory and immune responses support the hypothesis that particulates induce systemic inflammation. It has been well documented that particulate air pollutants can cause both pulmonary parenchymal (Nel et al. 2001; Pope 2000) and airway inflammation (Peden 2001). These particulate-mediated local inflammatory responses conform to those epidemiologic observations that exposure to particulate air pollutants can lead to asthma exacerbation, increased pulmonary infections, decreased pulmonary functions, increased hospitalizations due to pulmonary and/or airway diseases, and increased mortality. Recent studies using high-throughput technology for gene expression profiling have added to our understanding of particulate-mediated local inflammation underlying those adverse effects on lungs and airway in response to air pollution. Increased RNA expression for stress response, inflammatory, and repair-related genes were observed in Sprague-Dawley rats after intratracheal instillation of ROFA (Nadadur and Kodavanti 2002). In co-cultures of alveolar macrophages and primary human bronchial epithelial cells, mRNA levels of tumor necrosis factor (TNF)-α, granulocyte macrophage colony stimulating factor (GM-CSF), interleukin IL1β, IL6, and IL8 were increased within 2 hr (p < 0.05) after exposure to 100 μg/mL of PM10 (Fujii et al. 2002), and mRNA levels of leukemia inhibitory factor (LIF), GM-CSF, IL1α, and IL8 in primary human bronchial epithelial cells were increased after exposure to PM10 (Fujii et al. 2001).
In this study, we also demonstrated that it was critical to apply a dichotomous definition of absolute gene expression status, that is, expressed versus nonexpressed, in the data mining of the microarray data. Many algorithms currently used for microarray analysis retrieve the expression data from raw signals as continuous data and do not distinguish the distinct dichotomous biologic status of a gene. If one gene was not expressed in a RNA sample, there was always a meaningless expression value being generated that could not be distinguished accurately from other samples that expressed the same gene. In reality there should be no mRNA in a sample when a gene is not expressed. If the expression values generated by an algorithm truly represented reality, the data for expressed and nonexpressed genes should have different distributions. Therefore, without distinguishing expression status, a large number of meaningless data from nonexpressed genes would have deteriorating effects on a statistical analysis that assumed a normal distribution of data. Our observations that more functional pathways were associated with high content of Present calls in welders support this hypothesis. Furthermore, based on the absolute expression status, microarray data may be divided into three categories: consistently not expressed, turned on or off, and continuously expressed in different experimental conditions. The first category of genes was noninformative, and the analyses of the second category of genes were very complicated and difficult. Only the last category of genes, those with a high percentage of Present calls across all arrays, was suitable for parametric statistical analysis. At present, the Detection call algorithm of Affymetrix MAS 5.0 is the only one available for determining the absolute gene expression status, with limitations on both sensitivity and specificity to distinguish low-level expressed genes from nonexpressed genes (Liu et al. 2002).
In conclusion, using a repeated measure design, peripheral blood gene expression profiles revealed that environmental exposures to metal fume in healthy individuals produced observable changes in gene expression clustered in biologic processes related to inflammatory, oxidative stress, phosphate metabolism, cell proliferation, and programmed cell death. Smoking modified the observed responses. Finally, our study demonstrates the utility of paired sampling pre- and postexposure in an at-risk population.
Figure 1 Cluster analysis 44 RNA samples using 139 genes identified by paired t-test in welders. The clustering display was generated by dChip software with two-way data clustering. Each row represents an individual gene, and each column corresponds to an individual array. Gene expression values were standardized and color coded relative to the mean: blue, values less than the mean; red, values greater than the mean. RNA samples from the same individual were labeled with the same sample ID with different suffixes, representing different collection time points. Smoking status: N, nonsmoking; S, smoking. Exposure status: N, controls; Y, welders. Time point: B, baseline; P, postexposure. Experiment: the number indicates the hybridization batch in which a sample was analyzed.
Figure 2 GoSurfer graphic view of hypergeometric distribution testing of gene clustering. Each node represents a GO biologic process, and a line connecting nodes represents parent–child relationship in the top-down direction. Because GO allows multiple parent–child relationships toward one biologic process but GoSurfer only plots one upstream and one downstream relationship for each node, one biologic process may appear several times in the GoSurfer plot. Red nodes represent significant GO bioprocesses tested by hypergeometric distribution as described in “Materials and Methods.” Numbered GO bioprocesses were used in calculation in hypergeometric distribution testing: 1, biologic process; 2.1, cellular process; 2.2, development; 2.3, physiologic processes; 3.1, cell communication; 3.2, cell growth and/or maintenance; 3.3, metabolism; 3.4, response to external stimulus; 3.5; response to stress; and 3.6, death.
Table 1 Demographics of study population.
Welders Nonexposure controls p-Value
No. of subjects 15 7
No. of smokers (%) 6 (40) 1 (14) 0.35*
Age, years 32 (22–46) 40 (19–57) 0.69**
Years of boilermaking 3 (2–20) 3 (1.5–33) 0.61**
Number with hypertension (%) 1 (7) 2 (29) 0.23*
Welding fume exposure (PM2.5 concentration, mg/m3) 2.44 (1.30–3.42) 0.04 (0.02–0.17) < 0.001**
Unless specified, values are expressed as median (range) and were tested by the median test.
*Fisher’s exact test.
**Wilcoxon rank-sum test with exact p-value.
Table 2 Genes identified by paired t-test: postexposure versus baseline microarrays.
Gene with Present calls in Welders (28 arrays/14 pairs) Nonexposed controls (16 arrays/8 pairs) Gene ratio (welders:controls)
All arrays 533 86 6.20
At least one array 281 34 8.26
At least 10% arrays 236 28 8.43
At least 25% arrays 186 23 8.09
At least 50% arrays 139 17 8.18
Table 3 Results of hypergeometric testing using annotations from the Gene Ontology Consortium (GO).a
Genes from paired t-testb
Functional pathway GO ID Biologic processes Genes annotated on array Welders Controls
1 9605 Response to external stimulus 959 18 0
42330 Taxis 88 5 0
6935 Chemotaxis 88 5 0
30595 Immune cell chemotaxis 2 1 0
30593 Neutrophil chemotaxis 1 1 0
2 9605 Response to external stimulus 959 18 0
9607 Response to biotic stimulus 653 15 0
6952 Defense response 595 12 0
6955 Immune response 548 12 0
45087 Innate immune response 147 5 0
6954 Inflammatory response 145 5 0
42119 Neutrophil activation 1 1 0
30593 Neutrophil chemotaxis 1 1 0
3 6950 Response to stress 616 16 2
9605 Response to external stimulus 959 18 0
9611 Response to wounding 213 7 0
6954 Inflammatory response 145 5 0
42119 Neutrophil activation 1 1 0
30593 Neutrophil chemotaxis 1 1 0
4 9607 Response to biotic stimulus 653 15 0
6950 Response to stress 616 16 2
9613 Response to pest/pathogen/parasite 364 11 0
6954 Inflammatory response 145 5 0
42119 Neutrophil activation 1 1 0
30593 Neutrophil chemotaxis 1 1 0
9615 Response to viruses 28 2 0
5 9605 Response to external stimulus 959 18 0
9607 Response to biotic stimulus 653 15 0
6979 Response to oxidative stress 34 3 0
6950 Response to stress 616 16 2
6 8219 Cell death 315 7 1
12501 Programmed cell death 292 7 1
6915 Apoptosis 291 7 1
6916 Anti-apoptosis 60 3 1
7 8283 Cell proliferation 762 14 1
7049 Cell cycle 491 14 1
67 DNA replication and chromosome cycle 127 6 1
84 S phase of mitotic cell cycle 100 4 0
6260 DNA replication 99 4 0
6270 DNA replication initiation 14 2 0
8 6793 Phosphorus metabolism 468 9 1
6796 Phosphate metabolism 468 9 1
16311 Dephosphorylation 78 4 0
a Genes were identified from paired t-test with Present calls in at least 50% arrays. Annotations are from Gene Ontology Consortium (http://www.geneontology.org/).
b All listed biologic processes tested significantly in hypergeometric distribution testing in welders but nonsignificantly in nonexposed controls.
Table 4 Genes with altered expressions in response welding-fume exposure.
Welders
Controls
Accession numbera Gene namea Gene symbola Fold change Paired p-value Fold change Paired p-value
NM_00584 Interleukin 8 IL8 −1.22 0.004 1.02 0.923
NM_000575 Interleukin 1, alpha IL1A −1.12 0.035 −1.04 0.673
NM_003467 Chemokine (C-X-C motif) receptor 4 CXCR4 −1.26 0.038 1.02 0.900
NM_004757 Small inducible cytokine subfamily E, member 1 (endothelial monocyte-activating) SCYE1 −1.12 0.017 1.03 0.769
NM_006788 ralA binding protein 1 RALBP1 −1.16 0.036 −1.09 0.629
NM_004111 Flap structure-specific endonuclease 1 FEN1 −1.13 0.038 1.02 0.864
NM_000416 Interferon gamma receptor 1 IFNGR1 −1.39 0.014 −1.10 0.483
NM_078481 CD97 antigen CD97 1.21 0.049 −1.00 0.930
NM_002339 Lymphocyte-specific protein 1 LSP1 1.21 0.040 1.14 0.225
NM_012483 Granulysin GNLY −1.18 0.034 1.05 0.577
NM_001766 CD1D antigen, d polypeptide CD1D −1.22 0.002 −1.09 0.394
NM_001828 Charot-Leyden crystal protein CLC −1.21 0.033 −1.18 0.540
NM_000633 B-cell CLL/lymphoma 2 BCL2 −1.12 0.010 1.08 0.198
NM_006144 Granzyme A (granzyme 1, cytotoxic T-lymphocyte-associated serine esterase 3) GZMA −1.27 0.045 1.00 0.985
NM_080549 Protein tyrosine phosphatase, non-receptor type 6 PTPN6 1.10 0.032 −1.05 0.612
NM_000345 Synuclein, alpha (non A4 component of amyloid precursor) SNCA −1.28 0.030 1.02 0.907
NM_002656 Pleiomorphic adenoma gene-like 1 PLAGL1 −1.20 0.039 −1.11 0.318
NM_201397 Glutathione peroxidase 1 GPX1 1.14 0.035 −1.01 0.897
NM_004417 Dual specificity phosphatase 1 DUSP1 −1.21 0.035 −1.09 0.331
NM_001752 Catalase CAT −1.22 0.044 −1.07 0.521
NM_004383 c-src tyrosine kinase CSK 1.18 0.013 −1.02 0.736
NM_006999 Polymerase (DNA directed) sigma POLS −1.08 0.046 1.03 0.689
NM_002835 Protein tyrosine phosphatase, non-receptor type 12 PTPN12 −1.23 0.042 −1.02 0.810
NM_145906 RIO kinase 3 (yeast) RIOK3 −1.25 0.008 −1.05 0.737
NM_181742 Origin recognition complex, subunit 4-like (yeast) ORC4L −1.13 0.032 −1.06 0.596
NM_052811 ret finger protein 2 RFP2 −1.13 0.037 1.03 0.750
NM_002577 p21 (CDKN1A)-activated kinase 2 PAK2 −1.20 0.020 1.03 0.751
NM_002848 Protein tyrosine phosphatase, receptor type, O PTPRO −1.14 0.041 −1.12 0.209
NM_015374 unc-84 homolog B (C. elegans) UNC84B 1.12 0.044 −1.03 0.462
NM_004359 Cell division cycle 34 CDC34 1.17 0.013 −1.06 0.426
NM-002958 RYK receptor-like tyrosine kinase RYK −1.19 0.033 1.01 0.976
NM_014826 CDC42 binding protein kinase alpha (DMPK-like) CDC42BPA −1.21 0.020 1.04 0.843
NM_016839 RNA binding motif, single stranded interacting protein 1 RBMS1 −1.18 0.023 −1.09 0.535
NM_016113 Transient receptor potential cation channel, subfamily V, member 2 TRPV2 1.08 0.041 −1.07 0.208
NM_032454 Serine/threonine kinase 19 STK19 −1.11 0.044 1.07 0.456
a From Affymetrix NetAffx Analysis Center (http://www.affymetrix.com/analysis/index.affx).
Table 5 Effects of smoking on acute metal exposure expression profiles.
Welders
Nonexposed controls
All welders Nonsmokers Smokers All controls Nonsmokers
Number of arrays 30 18 12 14 12
Paired t-test
No. of genes with Present calls in all arrays 533 419 251 86 104
No. of genes with Present calls in at least 50% arrays 139 154 85 17 16
Hypergeometric distribution test in gene with Present call in at least 50% arraysa
1b 9605c Response to external stimulusd (959)e 18f 18f 8f 0f 0f
42330 Taxis (88) 5 3 1 0 0
6935 Chemotaxis (88) 5 3 1 0 0
30595 Immune cell chemotaxis (2) 1 1 0 0 0
30593 Neutrophil chemotaxis (1) 1 1 0 0 0
2 9605 Response to external stimulus (959) 18 18 8 0 0
9607 Response to biotic stimulus (653) 15 16 6 0 0
6952 Defense response (595) 12 13 6 0 0
6955 Immune response (548) 12 13 5 0 0
45087 Innate immune response (147) 5 3 0 0 0
6954 Inflammatory response (145) 5 3 0 0 0
42119 Neutrophil activation (1) 1 1 0 0 0
30593 Neutrophil chemotaxis (1) 1 1 0 0 0
3 6950 Response to stress (616) 16 17 4 2 2
9605 Response to external stimulus (959) 18 18 8 0 0
9611 Response to wounding (213) 7 4 1 0 0
6954 Inflammatory response (145) 5 3 0 0 0
42119 Neutrophil activation (1) 1 1 0 0 0
30593 Neutrophil chemotaxis (1) 1 1 0 0 0
4 9607 Response to biotic stimulus (653) 15 16 6 0 0
6950 Response to stress (616) 16 17 4 2 2
9613 Response to pest/pathogen/parasite (364) 11 10 3 0 0
6954op Inflammatory response (145) 5 3 0 0 0
42119 Neutrophil activation (1) 1 1 0 0 0
30593 Neutrophil chemotaxis (1) 1 1 0 0 0
9615 Response to viruses (28) 2 1 0 0 0
5 9605 Response to external stimulus (959) 18 18 8 0 0
9607 Response to biotic stimulus (653) 15 16 6 0 0
6979 Response to oxidative stress (34) 3 2 0 0 0
6950 Response to stress (616) 16 17 4 2 2
6 8219 Cell death (315) 7 8 5 1 1
12501 Programmed cell death (292) 7 8 5 1 1
6915 Aptosis (291) 7 8 5 1 1
6916 Anti-apoptosis (60) 3 5 2 0 0
7 8283 Cell proliferation (762) 14 17 5 1 1
7049 Cell cycle (491) 14 14 4 1 1
67 DNA replication and chromosome cycle (127) 6 5 1 1 1
84 S phase of mitotic cell cycle (100) 4 5 0 0 0
6260 DNA replication (99) 4 5 0 0 0
6270 DNA replication initiation (14) 2 1 0 0 0
8 6793 Phosphorus metabolism (468) 9 5 3 1 0
6796 Phosphate metabolism (468) 9 5 3 1 0
16311 Dephosphorylation (78) 4 2 0 0 0
a Italic numbers indicate that the functional pathway was tested significantly in hypergeometric distribution testing.
b Functional pathway.
c GO identification number (from GO Consortium: http://www.geneontology.org/)
d Biologic process.
e The number of genes on U133A array belonging to the functional pathway.
f The number of genes identified by paired t-test belonging to the functional pathway.
==== Refs
References
Affymetrix 2003a. An Analysis of Blood Processing Methods to Prepare Samples for GeneChip Expression Profiling. Affymetrix Technical Notes. Santa Clara, CA:Affymetrix. Available: http://www.affymetrix.com/support/technical/technotes/blood_technote.pdf [accessed 21 October 2004].
Affymetrix 2003b. Globin Reduction Protocol: A Method for Processing Whole Blood RNA Samples for Improved Array Results. Affymetrix Technical Notes. Santa Clara, CA:Affymetrix. Available: http://www.affymetrix.com/support/technical/technotes/blood2_technote.pdf [accessed 21 October 2004].
Clarke R Coull B Reinisch U Catalano P Killingsworth CR Koutrakis P 2000 Inhaled concentrated ambient particles are associated with hematologic and bronchoalveolar lavage changes in canines Environ Health Perspect 108 1179 1187 11133399
DePrimo S Wong L Khatry D Nicholas S Manning W Smolich B 2003 Expression profiling of blood samples from an SU5416 Phase III metastatic colorectal cancer clinical trial: a novel strategy for biomarker identification BMC Cancer 3 3 12657164
Dockery D Pope C Xu X Spengler J Ware J Fay ME 1993 An association between air pollution and mortality in six U.S. cities N Engl J Med 329 1753 1759 8179653
Eisen MB Spellman PT Brown PO Botstein D 1998 Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci USA 95 14863 14868 9843981
El-Zein M Malo J Infante-Rivard C Gautrin D 2003a Prevalence and association of welding related systemic and respiratory symptoms in welders Occup Environ Med 60 655 661 12937186
El-Zein M Malo J Infante-Rivard C Gautrin D 2003b Incidence of probable occupational asthma and changes in airway calibre and responsiveness in apprentice welders Eur Respir J 22 513 518 14516144
Frohlich M Sund M Lowel H Imhof A Hoffmeister A Koenig W 2003 Independent association of various smoking characteristics with markers of systemic inflammation in men. Results from a representative sample of the general population (MONICA Augsburg Survey 1994/95) Eur Heart J 24 1365 1372 12871694
Fujii T Hayashi S Hogg JC Mukae H Suwa T Goto Y 2002 Interaction of alveolar macrophages and airway epithelial cells following exposure to particulate matter produces mediators that stimulate the bone marrow Am J Respir Cell Mol Biol 27 34 41 12091243
Fujii T Hayashi S Hogg JC Vincent R Van Eeden SF 2001 Particulate matter induces cytokine expression in human bronchial epithelial cells Am J Respir Cell Mol Biol 25 265 271 11588002
Gardner S Lehmann J Costa D 2000 Oil fly ash-induced elevation of plasma fibrinogen levels in rats Toxicol Sci 56 175 180 10869466
Godleski J Godleski JJ Verrier RL Koutrakis P Catalano P Coull B 2000 Mechanisms of morbidity and mortality from exposure to ambient air particles Res Rep Health Eff Inst 91 5 88 discussion 89–103.10817681
Gordon T Nadziejko C Schlesinger R Chen LC 1998 Pulmonary and cardiovascular effects of acute exposure to concentrated ambient particulate matter in rats Toxicol Lett 96–97 285 288
Hakak Y Walker J Li C Wong W Davis K Buxbaum J 2001 Genome-wide expression analysis reveals dysregulation of myelination-related genes in chronic schizophrenia Proc Natl Acad Sci USA 98 4746 4751 11296301
Hauser R Elreedy S Hoppin J Christiani D 1995 Upper airway response in workers exposed to fuel oil ash: nasal lavage analysis Occup Environ Med 52 353 358 7795759
Kim Y Wand P Hauser R Mukherjee S Herrick F Christiani C 2003 Association of expired nitric oxide with occupational particulate exposure Environ Health Perspect 111 676 680 12727593
Klein U Tu Y Stolovitzky G Mattioli M Cattoretti G Husson H Freedman A 2001 Gene expression profiling of B cell chronic lymphocytic leukemia reveals a homogeneous phenotype related to memory B cells J Exp Med 194 1625 1638 11733577
Li C Wong W 2001a Model-based analysis of oligonucleotide arrays: model validation, design issues and standard error application Genome Biol 2 0032.1 0032.11
Li C Wong W 2001b Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection Proc Natl Acad Sci USA 98 31 36 11134512
Liu W Mei R Di X Ryder T Hubbell E Dee S 2002 Analysis of high density expression microarrays with signed-rank call algorithms Bioinformatics 18 1593 1599 12490443
Locati M Deuschle U Massardi M Martinez F Sironi M Sozzani S 2002 Analysis of the gene expression profile activated by the CC chemokine ligand 5/RANTES and by lipopolysaccharide in human monocytes J Immunol 168 3557 3562 11907119
Mann J Tager I Lurmann F Segal M Quesenberry C Lugg M 2002 Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia Environ Health Perspect 110 1247 1252 12460805
Mueller E Seger D 1985 Metal fume fever—a review J Emerg Med 2 271 274 3910713
Mukaida N 2000 Interleukin-8: an expanding universe beyond neutrophil chemotaxis and activation Int J Hematol 72 391 398 11197203
Mukaida N 2003 Pathophysiological roles of interleukin-8/CXCL8 in pulmonary diseases Am J Physiol Lung Cell Mol Physiol 284 L566 L577 12618418
Murphy D 2002 Gene expression studies using microarrays: principles, problems, and prospects Adv Physiol Educ 26 256 270 12443997
Nadadur SS Kodavanti UP 2002 Altered gene expression profiles of rat lung in response to an emission particulate and its metal constituents J Toxicol Environ Health A 65 1333 1350 12227955
Nel AE Diaz-Sanchez D Li N 2001 The role of particulate pollutants in pulmonary inflammation and asthma: evidence for the involvement of organic chemicals and oxidative stress Curr Opin Pulm Med 7 20 26 11140402
Pahl A Brune K 2002 Gene expression changes in blood after phlebotomy: implications for gene expression profiling Blood 100 1094 1095 12130480
Peden DB 2001 Air pollution in asthma: effect of pollutants on airway inflammation Ann Allergy Asthma Immunol 87 12 17 11770676
Pekkanen J Brunner EJ Anderson HR Tiittanen P Atkinson RW 2000 Daily concentrations of air pollution and plasma fibrinogen in London Occup Environ Med 57 818 822 11077010
Peters A Dockery D Muller J Mittleman M 2001a Increased particulate air pollution and the triggering of myocardial infarction Circulation 103 2810 2815 11401937
Peters A Frohlich M Doring A Immervoll T Wichmann HE Hutchinson WL 2001b Particulate air pollution is associated with an acute phase response in men; results from the MONICA-Augsburg Study Eur Heart J 22 1198 1204 11440492
Peters A Liu E Verrier R Schwartz J Gold D Mittleman M 2000 Air pollution and incidence of cardiac arrhythmia Epidemiology 11 11 17 10615837
Pope CA III 2000 What do epidemiologic findings tell us about health effects of environmental aerosols? J Aerosol Med 13 335 354 11262440
Pope C Burnett R Thun M Calle E Krewski D Ito K 2002 Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution JAMA 287 1132 1141 11879110
Schwartz J 2001 Air pollution and blood markers of cardiovascular risk Environ Health Perspect 109 405 409 11427390
Seaton A Soutar A Crawford V Elton R McNerlan S Cherrie J 1999 Particulate air pollution and the blood Thorax 54 1027 1032 10525563
Sjogren B Fossum T Lindh T Weiner J 2002 Welding and ischemic heart disease Int J Occup Environ Health 8 309 311 12412847
Smith M Kinmonth A Luben R Bingham S Day N Wareham N 2003 Smoking status and differential white cell count in men and women in the EPIC-Norfolk population Atherosclerosis 169 331 337 12921986
Suwa T Hogg J Quinlan K Ohgami A Vincent R van Eeden S 2002 Particulate air pollution induces progression of atherosclerosis J Am Coll Cardiol 39 935 942 11897432
Tavazoie S Hughes J Campbell M Cho R Church G 1999 Systematic determination of genetic network architecture Nat Genet 22 281 285 10391217
Tusher V Tibshirani R Chu G 2001 Significance analysis of microarrays applied to the ionizing radiation response Proc Natl Acad Sci USA 98 5116 5121 11309499
Verbeke G Molenberghs G 1997. Linear Mixed Models in Practice: A SAS-Oriented Approach. New York:Springer.
Wolf C Pirich C Valic E Waldhoer T 1997 Pulmonary function and symptoms of welders Int Arch Occup Environ Health 69 350 353 9192220
Woodin M Hauser R Liu Y Smith T Siegel P Lewis D 1998 Molecular markers of acute upper airway inflammation in workers exposed to fuel-oil ash Am J Respir Crit Care Med 158 182 187 9655727
|
15687063
|
PMC1277870
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb 22; 113(2):233-241
|
utf-8
|
Environ Health Perspect
| 2,004 |
10.1289/txg.7273
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a0008415687039PerspectivesEditorialGuest Editorial: Satellite Remote Sensing Can Improve Chances of Achieving Sustainable Health Patz Jonathan University of Wisconsin–Madison, Madison, Wisconsin, E-mail:
[email protected] Patz, MD, MPH, is an associate professor in the Center for Sustainability and the Global Environment at the University of Wisconsin–Madison, an adjunct associate professor at the Johns Hopkins Bloomberg School of Public Health, and an affiliate scientist of the National Center for Atmospheric Research.
2 2005 113 2 A84 A85 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
The Global Earth Observation System of Systems (GEOSS) is a welcomed cooperative in an era when we are becoming “data rich but knowledge poor.” With the proliferation of satellite platforms, each monitoring different characteristics of the earth’s surface and atmosphere at varying resolutions, the task of using a combination of satellite databases has been intimidating and often not possible without large analytical effort. Also, one of the most pressing challenges across the field of environmental health is obtaining accurate exposure assessments. A system that can help integrate, for instance, meteorological, air, and water pollution and soil and food contamination will improve risk assessment. Remotely sensed data are especially useful in monitoring changes in broad area or earth system disturbances; two that are especially pertinent to disease emergence include global climate change and land use change.
Many diseases or health outcomes are sensitive to climatic conditions, from mortality and morbidity due to extreme heat, cold, drought, or storms, to vector- or waterborne infectious diseases. One clear application for remotely observed data in climate-health studies is that of thermal mapping with high resolution thermal infrared aircraft (Lillesand et al. 2004). Looking at urban sprawl in aggregate, Kalnay and Cai (2003) estimated a mean surface warming due to urban sprawl and land-use change to be 0.27°C (0.49°F) for the continental United States. Thermal imagery has been combined with Landsat Thematic Mapper data in many cities; one example, Dallas, Texas, shows an urban heat island effect of 5–11°C compared to surrounding rural areas (Aniello et al. 1995). A shocking 22,000–35,000 heat-related deaths occurred across Europe during two weeks in August 2003 (IFRC 2004); if we are building cities that can raise temperatures by several degrees, we are certainly not helping future situations under climate change scenarios.
Infectious disease epidemics occur at a local or sometimes regional scale, and one key challenge to accurate vulnerability analysis is incorporating land use change projections with future projections of global climate change. For example, Hurricane Mitch, a devastating storm that hit Central America in 1998, demonstrates the combined effects of land use and extreme weather: 9,600 people perished, widespread illness from water-and vector-borne diseases ensued, and 1 million people were left homeless. Areas with extensive deforestation, with settlements on degraded hillsides or floodplains, suffered the greatest morbidity and mortality (Cockburn et al. 1999). The importance of land-cover features as a buffer to severe floods emerged as essential to long-term prevention of injuries and fatalities from floods (Glantz and Jamieson 2000).
Increasingly in recent years, meteorological satellite data has been used to model the spatial and seasonal dynamics of infectious disease transmission and develop affordable early warning systems for malaria (Thomson et al. 1997). Other climate-sensitive disease studies have combined climate and land use data to develop predictive models. Using Landsat Thematic Mapper satellite imagery, Glass et al. (2000) found that El Niño/Southern Oscillation (ENSO)-related heavy rainfall, with subsequent increase in the rodent population, preceded human cases of hantavirus pulmonary syndrome in the American southwest. And in the Bay of Bengal, Colwell and colleagues (Colwell 1996; Lobitz et al. 2000) were able to predict cholera epidemics by using AVHRR (Advanced Very High Resolution Radiometer), TOPEX/Poseidon (TOPography EXperiment for Ocean Circulation), and SeaWiFS (Sea-viewing Wide Field-of-view Sensor) remotely sensed data to determine sea-surface temperature and turbidity, sea-surface altitude, and marine algal blooms, respectively.
Using Pacific and Indian Ocean sea-surface temperature anomalies, coupled with satellite normalized difference vegetation index data, Linthicum et al. (1999) found that Rift Valley fever outbreaks could be predicted up to 5 months in advance of outbreaks in East Africa. One limitation to the use of remote sensing for the study of vector-borne disease epidemics has been cloud cover during the most critical period key to transmission for some diseases—the rainy season. Now, with the arrival of the synthetic aperture radar (SAR) that can penetrate through clouds, this problem is being resolved.
Land-use practices have had many positive impacts on human health, largely by increasing food supply, shelter, and sanitation. Nevertheless, land-use practices have also led to unintended health consequences. Road and dam construction, irrigation, habitat fragmentation, and urban sprawl all modify the transmission of infectious disease (Patz et al. 2004). Irrigation in the tropics increases the habitat and breeding sites for schistosomiasis and malaria. Dam construction has led to proliferation of the mosquito Culex pipiens and subsequent filariasis, or elephantiasis, near the Aswan High Dam in the southern Nile Delta (Thompson et al. 1996).
The biodiversity monitoring of GEOSS is also relevant to human health: an estimated 75% of human diseases are zoonotic, having links to either wildlife or domestic animals (Taylor et al. 2001). Lyme disease is one example of a disease linked to forest fragmentation in the eastern United States, with subsequent proliferation of deer and white-footed mice key in the pathogen’s life cycle. Combining remotely sensed land use data with statistical software to analyze habitat fragmentation patterns could, therefore, potentially enhance Lyme disease risk predictions.
Finally, lessons for building resilience against unpredictable catastrophes are emerging from the recent tragic tsunami in the Indian Ocean that, at last report, has killed upwards of 150,000 people, with many more injured or at risk of infectious diseases. Improved satellite early warning systems are already under discussion, but additional evidence is emerging about high survival rates of people in areas with intact coral reefs and mangroves. These types of land use change are best studied with satellite remote sensing in combination with local ground-truth data.
In summary, the goals of GEOSS’ 10-year international collaboration to greatly improve data compatibility and communication across earth-observing systems has particular relevance to human health. The goals of disaster reduction, water resource management, ocean and marine resource management, air-quality monitoring, biodiversity monitoring, and sustainable land use management could not be more central to understanding human population vulnerability across the generations.
==== Refs
References
Aniello C Morgan K Busbey A Newland L 1995 Mapping micro-urban heat islands using Landsat TM and a GIS Comput Geosciences 21 8 965 969
Cockburn A St Clair J Silverstein K 1999 The politics of “natural” disaster: who made Mitch so bad? Int J Health Serv 29 459 462 10379461
Colwell RR 1996 Global climate and infectious disease: the cholera paradigm Science 274 2025 2031 8953025
Glantz M Jamieson D 2000 Societal response to Hurricane Mitch and intra- versus intergenerational equity issues: whose norms should apply? Risk Anal 20 869 882 11314736
Glass GE Cheek JE Patz JA Shields TM Doyle TJ Thoroughman DA 2000 Using remotely sensed data to identify areas of risk for hantavirus pulmonary syndrome Emerg Infect Dis 63 3 238 247 10827113
IFRC (International Federation of Red Cross and Red Crescent Societies) 2004. World Disaster Report 2004. New York:Oxford University Press.
Kalnay E Cai M 2003 Impact of urbanization and land-use change on climate Nature 423 6939 528 531 12774119
Lillesand TM Kiefer RW Chipman JW 2004. Remote Sensing and Image Interpretation. 5th ed. New York:John Wiley & Sons, Inc.
Linthicum KJ Anyamba A Tucker CJ Kelley PW Myers MF Peters CJ 1999 Climate and satellite indicators to forecast Rift Valley fever epidemics in Kenya Science 285 5426 397 400 10411500
Lobitz B Beck L Huq A Wood B Fuchs G Faruque AS 2000 Climate and infectious disease: use of remote sensing for detection of Vibrio cholerae by indirect measurement Proc Natl Acad Sci USA 97 1438 1443 10677480
Patz JA Daszak P Tabor GM Aguirre AA Pearl M Epstein J 2004 Unhealthy landscapes: policy recommendations on land use change and infectious disease emergence Environ Health Perspect 112 1092 1098 15238283
Taylor LH Latham SM Woolhouse ME 2001 Risk factors for human disease emergence Philos Trans R Soc Lond B Biol Sci 356 983 989 11516376
Thompson DF Malone JB Harb M Faris R Huh OK Buck AA 1996 Bancroftian filariasis distribution and diurnal temperature differences in the southern Nile delta Emerg Infect Dis 2 234 235 8903237
Thomson MC Connor SJ Milligan PJM Flasse SP 1997 Mapping malaria risk in Africa: what can satellite data contribute? Parasitol Today 13 313 318 15275058
|
15687039
|
PMC1277871
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A84-A85
|
utf-8
|
Environ Health Perspect
| 2,005 |
10.1289/ehp.113-a84
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a00085PerspectivesEditorialNote from the Editors: Toxicogenomics Update Goehl Thomas J. Editor-in-Chief, EHP, NIEHS, Research Triangle Park, North Carolina, E-mail:
[email protected] Kenneth S. Toxicogenomics Editor, EHP ,University of Louisville Health Sciences Center, Louisville, Kentucky2 2005 113 2 A85 A85 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
EHP is continually evolving to meet the needs of our readership. Our intention in publishing a separate section in toxicogenomics for the past 2 years has been to feature research in this emerging field and to use our news articles as an educational tool to explain basic toxicogenomics principles to our general readership.
Feedback from our readership and the editorial board has indicated that our current coverage of the field of toxicogenomics might not be achieving the intended goals. After careful deliberations with stakeholders, the editors have decided that a better way to meet the needs of our readership and the field of toxicogenomics is to incorporate the Toxicogenomics section into the monthly issues. This change will be implemented beginning this month.
The major advantage of this approach is that we can print toxicogenomics articles shortly after acceptance rather than holding them for publication in a quarterly section. In addition, published articles will gain wider exposure among the general readership, helping to better fold toxicogenomics into the environmental health research portfolio. Another advantage is that we will able to more efficiently achieve our educational goals by making the field more accessible to our general readership.
What does the change mean for coverage?
A toxicogenomics research section will be included within each monthly issue when research papers are available.
EHP will continue news coverage through a variety of formats including updates, investigative articles, and policy discussions as warranted.
To meet our educational goals, we have compiled all toxicogenomics feature articles (Focus articles) in a Toxicogenomics Primer that will be placed on a CD to be included with the March issue and made available at various toxicogenomics meetings.
We will incorporate the toxicogenomics calendar, fellowships and grants, new books, and book reviews within the monthly sections.
We are working to enhance the Toxicogenomics section of the EHP website by adding a compilation of news and research articles, “Notes from the Field,” and Resources.
Editorials on toxicogenomics will continue to be solicited.
We look forward to your continued support of EHP and the Toxicogenomics section.
|
0
|
PMC1277872
|
CC0
|
2021-01-04 23:40:55
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A85
|
utf-8
|
Environ Health Perspect
| 2,005 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a0008615687041PerspectivesCorrespondenceOzone: Unrealistic Scenarios Schwartz Joel American Enterprise Institute for Public Policy Research Washington, DC, E-mail:
[email protected] Patrick Davis Robert E. Department of Environmental Sciences University of Virginia Charlottesville, VirginiaThe authors declare they have no competing financial interests.
2 2005 113 2 A86 A87 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Knowlton et al. (2004) argued that increasing temperatures associated with climate change will increase urban ozone and related health risks. They have disregarded important factors in reaching this conclusion.
During the last 20 years, nationwide exceedances of the federal 1-hr ozone standard declined 90%, and the June–August average of daily 1-hr peak ozone levels declined 10% (Schwartz et al., in press), presumably with ensuing declines in ozone-related mortality. Ozone declined despite a roughly 1°C increase in urban temperatures during the last few decades (Karl et al. 1988). Knowlton et al. (2004) did not explain why we should expect the future to be the opposite of the past.
Knowlton et al. (2004) used ozone-precursor [nitrogen oxides (NOx) and volatile organic compound (VOC)] emissions estimates for 1996 to predict ozone levels in the 2050s. However, even current emissions are substantially lower than 1996 levels, while, as shown below, already-adopted U.S. Environmental Protection Agency (EPA) requirements will eliminate most remaining ozone-precursor emissions, even after accounting for growth.
The U.S. EPA (2003) estimated that between 1996 and 2001, total emissions of NOx and VOC declined, 10 and 14%, respectively. [The U.S. EPA updated its trend estimates in November 2004 (U.S. EPA 2004a) and now believes the decline was even steeper, although these new estimates were obviously not available to Knowlton et al.] During 2003 and 2004, the U.S. EPA capped total NOx from coal-fired power plants and industrial boilers at 60% below 2000 levels (U.S. EPA 1998a, 2004b). A range of emissions data show the average automobile’s NOx emissions rate declined 4–9% per year between 1995 and 2001, with greater improvements for vehicles up to 4 years old (Pokharel et al. 2003; Schwartz 2003). Total driving is increasing < 2% per year, resulting in large net NOx declines (Texas Transportation Institute 2004).
Data on heavy-duty diesel vehicles are sparse, but there is every reason to believe that diesel NOx has also declined. The U.S. EPA has tightened NOx standards for new on- and off-road diesels several times over the last 15 years, and also recently required additional NOx reductions from in-use 1993–1998 model year trucks (U.S. EPA 2002a, 2004c, 2004d).
VOCs have declined far more than NOx and far more than U.S. EPA estimates. The U.S. EPA’s official VOC inventory understates significantly the gasoline-vehicle contribution to total VOCs (Watson et al. 2001). Real-world data show the average automobile’s VOC emission rate is declining 11–15% per year, again much more rapidly than driving is increasing, and with a more rapid decline for recent models (Pokharel et al. 2003; Schwartz 2003). The U.S. EPA also recently implemented VOC reductions for other sources (U.S. EPA 1998b, 2002b, 2004e).
Overall, between 1996 and 2004, anthropogenic NOx and VOC emissions likely declined, respectively, at least 25 and 50%—declines overlooked by Knowlton et al. (2004).
Emission declines will continue. For example, a vehicle fleet meeting the U.S. EPA’s “Tier 2” automobile standards, implemented in 2004, on-road diesel standards set for 2007, and off-road diesel standards set for 2010, will emit 90% less NOx and VOCs per mile over their lifetime than the current average vehicle, resulting in huge emissions declines, even with predicted increases in driving (Schwartz 2003; U.S. EPA 2000a, 2000b, 2004c).
Knowlton et al. (2004) assume ozone-precursor emissions several times greater than any plausible future scenario. Their projections of future ozone and related health impacts are therefore unrealistically high.
Heat-related mortality has also declined, by 70% nationwide since the 1960s, despite warming urban climates, with the hottest and most humid cities achieving the greatest risk reductions (Davis et al. 2003). These health improvements resulted from a range of adaptive technologies and processes, including increased air conditioning, changes in building design, physiologic adaptations, and improved emergency medicine.
Nevertheless, because of a single major blackout on a warm day in 2003, Knowlton et al. (2004) maintain that “air conditioning may not really be an appropriate ‘fix’ for adapting to climate change.” Air conditioning is clearly a vital adaptive technology that has saved countless lives. One study reported a relative risk of death on hot days of 1.7 for people with no air conditioning compared to those with central air (Rogot et al. 1992).
The nondiscriminating reader might be impressed by the downscaling of a general circulation model using a regional mesoscale model to predict localized differences in future air-pollution related mortality, but the complexity of the models is irrelevant in the face of Knowlton et al.’s failure to temper their theoretical exercise with real-world data. Had Knowlton et al. (2004) accounted for observed historical health and pollution trends and future emission-reduction requirements, they would have arrived at a markedly different story.
==== Refs
References
Davis RE Knappenberger PC Michaels PJ Novicoff WM 2003 Changing heat-related mortality in the United States Environ Health Perspect 111 1712 1718 14594620
Karl TR Diaz HF Kukla G 1988 Urbanization: its detection and effect in the United States climate record J Climate 1 1099 1123
Knowlton K Rosenthal JE Hogrefe C Lynn B Gaffin S Goldberg R 2004 Assessing ozone-related health impacts under a changing climate Environ Health Perspect 112 1557 1563 15531442
Pokharel SS Bishop GA Stedman DH Slott R 2003 Emissions reductions as a result of automobile improvement Environ Sci Technol 37 5097 5101 14655694
Rogot E Sorlie PD Backlund E 1992 Air-conditioning and mortality in hot weather Am J Epidemiol 136 106 116 1415127
Schwartz J 2003. No Way Back: Why Air Pollution Will Continue to Decline. Washington, DC:American Enterprise Institute.
Schwartz J Hayward SF Kahlbaum D In press. Air Quality in America. Washington, DC:American Enterprise Institute.
Texas Transportation Institute 2004. Congestion Data for Your City. Available: http://mobility.tamu.edu/ums/congestion_data/ [accessed 25 October 2004].
U.S. EPA 1998a. Addendum to the Regulatory Impact Analysis for the NOx SIP Call, FIP, and Section 126 Petitions. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 1998b National Volatile Organic Compound Emission Standards for Consumer Products; Final Rule Fed Reg 63 48819 48847
U.S. EPA 2000a Control of Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements; Final Rule Fed Reg 65 6698 6870
U.S. EPA 2000b. Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements. EPA420-R-00-026. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2002a. Health Assessment Document for Diesel Engine Exhaust. EPA/600/8-90/057F. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2002b. Regulatory Announcement: Emission Standards for New Nonroad Engines. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2003. 1970–2001 Average Annual Emissions, All Criteria Pollutants. Washington, DC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/ttn/chief/trends/trends01/trends2001_aug2003.zip [accessed 25 October 2004].
U.S. EPA 2004a. 1970–2002 Average Annual Emissions, All Criteria Pollutants. Washington, DC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/ttn/chief/trends/trends02/trendsreportallpollutants111504.xls [accessed 22 November 2004].
U.S. EPA 2004b. NOx Budget Trading Program: 2003 Progress and Compliance Report. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2004c. Final Regulatory Analysis: Control of Emissions from Nonroad Diesel Engines. Washington, DC:U.S. Environmental Protection Agency.
U.S. EPA 2004d. Heavy Duty Diesel Engine Settlement Information. Washington, DC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/compliance/civil/programs/caa/diesel/index.html [accessed 25 October 2004].
U.S. EPA 2004e. National Emission Standards for Hazardous Air Pollutants. Washington, DC:U.S. Environmental Protection Agency. Available: http://www.epa.gov/ttn/atw/mactfnlalph.html [accessed 18 October 2004].
Watson JG Chow JC Fujita EM 2001 Review of volatile organic compound source apportionment by chemical mass balance Atmos Environ 32 1567 1584
|
15687041
|
PMC1277873
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A86-A87
|
utf-8
|
Environ Health Perspect
| 2,005 |
10.1289/ehp.113-1277873
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a00089PerspectivesErrataErrata 2 2005 113 2 A89 A89 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
Siemiatycki et al. have detected some errors in their review of occupational carcinogens published in the November 2004 issue of EHP [
Environ Health Perspect 112:1447–1459 (2004)]. Specifically, they inadvertently included in their list of Group 2B (possible) carcinogens some substances that had been downgraded to Group 3 (not classifiable) in a subsequent IARC (International Agency for Research on Cancer) Monograph. Their corrections are as follows:
On page 1449, 3rd column, “113 possible human occupational carcinogens (IARC Group 2B; Table 5)” should be replaced by “110 possible human occupational carcinogens (IARC Group 2B; Table 5).”
On page 1454, Table 5 (under “Respirable dusts and fibers”), glass wool, rock wool, and slag wool fireproofing should not have appeared in the listing of Group 2B human carcinogens because they were downgraded to Group 3 in the latest monograph to address these substances (IARC 2002a); special purpose glass fibers such as E-glass and “475” glass fibers are not used in the “Reinforced plastic industry” but rather in “High-efficiency air filtration media and battery separator media” (IARC 2002a). A corrected version of this section of Table 5 is presented below.
On page 1459, Table 8, as a result of the previous corrections, the last section of Table 8 (Current rating 2B) should be modified as follows: the total number of substances with this rating should read 110 instead of 113; the number of substances unrated by IARC in 1987 should read 36 instead of 39; and the number of substances unrated by the World Health Organization (WHO) in 1964 should read 104 instead of 107. A corrected version of Table 8 is presented below.
Siemiatycki et al.’s review of the IARC Monographs was intended to cover volumes 1–83. With these corrections, the tables and text are complete.
Table 5 Substances and mixtures that have been evaluated by IARC as possible (Group 2B) human carcinogens and that are occupational exposures [corrected section only].
Substance or mixture Occupation or industry in which the substance is founda IARC Monograph volume (year)b Human evidencec Animal evidencec
Respirable dusts and fibers
Palygorskite (long fibers > 5 μm) Miners and millers; production of waste absorbents, fertilizers, and pesticides Vol. 68 (1997b) Inadequate Sufficient
Refractory ceramic fibers Production; furnace insulators; ship builders; heat-resistant fabric manufacture Vol. 81 (2002a) Inadequate Sufficient
Special-purpose glass fibers such as E-glass and “475” glass fibers High-efficiency air filtration media; battery separator media Vol. 81 (2002a) Not available Sufficient
Table 8 Evolution in knowledge regarding current (2003) IARC occupational carcinogens (corrected version).
Earlier evaluation
Current rating Past rating IARC 1987 WHO 1964
1 (n = 28) 1 19 13
2A 4 ]— 4
2B 1
3 0 NA
Unrated 4 11
Total 28 28
2A (n = 27) 1 0 0
2A 16 ]— 0
2B 6
3 2 NA
Unrated 3 27
Total 27 27
2B (n = 110) 1 0 1
2A 2 ]— 5
2B 63
3 9 NA
Unrated 36 104
Total 110 110
NA, not applicable.
==== Refs
References
IARC 1987. Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. IARC Monogr Eval Carcinog Risk Chem Hum (suppl 7).
IARC 1997b. Silica, Some Silicates, Coal Dust and Para-aramid Fibrils. IARC Monogr Eval Carcinog Risks Hum 68.
IARC 2002a. Man-Made Vitreous Fibres. IARC Monogr Eval Carcinog Risks Hum 81.
WHO 1964. Prevention of Cancer. Report of a WHO Committee. Technical Report Series 276. Geneva:World Health Organization.
|
0
|
PMC1277874
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A89
|
utf-8
|
Environ Health Perspect
| 2,005 |
nan
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a0009015687043EnvironewsForumNatural Disasters: Building a Tsunami Warning System Schmidt Charles W. 2 2005 113 2 A90 A90 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
The 26 December 2004 tsunami that devastated coastal areas in Indonesia, Malaysia, Thailand, Myanmar, India, the Maldives, Sri Lanka, and Somalia was produced by a magnitude 9.0 earthquake 155 kilometers southwest of northern Sumatra. The tsunami inundated coasts with waves more than 30 feet high every 30–40 minutes for several hours. It killed more than 150,000 people and injured millions, making it one of the worst natural disasters the modern world has ever seen.
Could populations have been warned in advance? Not with the tsunami monitoring networks that exist in the affected area today, say officials with the National Oceanic and Atmospheric Administration (NOAA). Unlike the Pacific Ocean, which is wired for tsunami alerts by the United Nations Intergovernmental Oceanographic Commission, the Indian Ocean is largely devoid of comparable sensor technologies that detect earthquakes and issue tsunami warnings to affected countries. According to NOAA officials, future warnings could be enhanced by the creation of an Indian Ocean tsunami warning center that would deploy coastal tide gauges to measure the amplitude of waves near tsunami source areas such as fault zones and volcanoes, and would establish a communications infrastructure to send and receive alerts.
Tsunami buoys also would be useful in an early warning system, NOAA officials say. At present, a network of six buoys is deployed by NOAA off the Aleutian Islands in the Pacific Ocean. This network, called Deep Assessment and Reporting of Tsunamis, is composed of two parts: a sea floor sensor and a buoy that relays tsunami information to warning centers on the ground by satellite communication. NOAA estimates that about 50 such buoys are needed to adequately cover the world’s oceans. On 6 January 2005 Senator Joe Lieberman (D–CT) proposed a $30 million package to develop these additional buoys, which the U.S. Congress is currently considering. Australian scientists are also designing an Indian Ocean tsunami alert system that is expected to cost US$20 million.
Still more help could come from the sky. On 10 January 2005, NOAA announced that its scientists had measured the height of the December 26 tsunami using data from four Earth-orbiting satellites passing over the affected area at the time. NOAA geophysicist Walter H.F. Smith said in an agency press release that the best application of satellite data to improve tsunami hazard forecasts may be in helping to map the ocean floor from space.
To be truly effective, any tsunami warning system will need to be part of an overall disaster reduction strategy, says Helen Wood, senior advisor for systems and services in NOAA’s National Environmental Satellite, Data, and Information Service. “You don’t want a separate system that segregates tsunamis from earthquakes or cyclones, because populations on the coast are at risk from any and all of those things,” she says.
Wood heads the secretariat for the Global Earth Observation System of Systems, an emerging international coalition that has set disaster reduction as one of nine priority areas [see “Terra Cognita: Using Earth Observing Systems to Understand Our World,” p. A98 this issue]. “We had already identified tsunami and related instrumentation as a necessity,” Wood says. “Now there’s a sense of urgency. [The December 26 tsunami] was a catalyst for action. It’s sad, but disaster often mobilizes people at all levels to take action.”
The tragedy of hindsight. Similar disasters may be mitigated in the future by better warning sytems.
Waves of destruction. Composite satellite images created by NOAA measure the tsunami’s height as it spread from the quake epicenter. The ability to make depth surveys from space may lead to improvements in models that forecast the hazardous effects of tsunamis.
|
15687043
|
PMC1277875
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A90
|
utf-8
|
Environ Health Perspect
| 2,005 |
10.1289/ehp.113-a90
|
oa_comm
|
==== Front
Environ Health PerspectEnviron. Health PerspectEnvironmental Health Perspectives0091-67651552-9924National Institute of Environmental Health Sciences ehp0113-a00093EnvironewsForumEHPnet: The Earth Observing System Dooley Erin E. 2 2005 113 2 A93 A93 Publication of EHP lies in the public domain and is therefore without copyright. All text from EHP may be reprinted freely. Use of materials published in EHP should be acknowledged (for example, ?Reproduced with permission from Environmental Health Perspectives?); pertinent reference information should be provided for the article from which the material was reproduced. Articles from EHP, especially the News section, may contain photographs or illustrations copyrighted by other commercial organizations or individuals that may not be used without obtaining prior approval from the holder of the copyright.
==== Body
In 1991, the National Aeronautics and Space Administration (NASA) embarked upon the extensive and comprehensive Earth Science Enterprise, a program to help us better understand what is happening on the Earth’s surface. The centerpiece of the Earth Science Enterprise is known as The Earth Observing System (EOS). The EOS website, located at http://eospso.gsfc.nasa.gov/, is designed to keep scientists, educators, and the news media informed about the system and its activities.
EOS is composed of a series of satellites conducting long-term global observations of the land surface, biosphere, solid Earth, atmosphere, and oceans. The first EOS satellite, Terra, was launched into orbit in December 1999. NASA hopes the data gathered by the system, the first to provide integrated measurements of processes occurring on Earth, can be used to help better predict climate change, oversee conservation efforts, and improve forecasting to help farmers, fishers, and other weather-dependent tradespeople.
The For Scientists section of the website provides links to overviews of the field experiments that EOS conducts, and contains information on what sorts of measurements EOS takes and what types of instruments are used. Also within this section are links to brochures, reference books, data product catalogs, a bibliography of EOS-related publications, and a calendar of workshops and conference seminars conducted by EOS personnel.
Teachers looking for EOS-related resources can visit the For Educators section. The EOS Education Project, with a link located under the Educational Links subhead, has been developed for teachers and students at all levels to teach about topics including geographic information systems and remote sensing. The project offers outreach programs for teachers and Internet-based classes. This section also features a directory of other EOS-related teaching materials, including brochures, fact sheets, posters, lithographs, and lesson plans that can be downloaded or ordered for free. Also available are online tools for calculating when a satellite will pass overhead and tracking various satellites’ current location in the sky. Links to other websites include the main NASA Earth Science Enterprise educational website and the homepages of related agencies such as the National Science Foundation.
News writers have their own section on the website. Available under For News Media is the Earth Observatory Newsroom, which contains press releases on the latest EOS research, NASA news announcements, and a list of newly published research. This section also features the EOS Global Change Media Directory 2001. This resource lists 343 scientists working in fields such as ozone chemistry, global warming, and ecosystems who have expressed interest in working with the media. Science writers’ guides to individual spacecraft provide extensive background material and contact information.
The site’s Data Services section provides links to the different types of data that the EOS and its partners provide. This section links to NASA’s Global Change Master Directory website, which provides extensive data sets on 13 topics, including agriculture, human dimensions, oceans, and sun–Earth interactions. Subscribers can add to the data sets and participate in an online discussion list.
|
0
|
PMC1277876
|
CC0
|
2021-01-04 23:40:56
|
no
|
Environ Health Perspect. 2005 Feb; 113(2):A93
|
utf-8
|
Environ Health Perspect
| 2,005 |
nan
|
oa_comm
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.