Published ahead of print on November 9, 2006, doi:10.1164/rccm.200607-970OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200607-970OC
A Polymorphism in the P2X7 Gene Increases Susceptibility to Extrapulmonary Tuberculosis1 Mycobacterial Research Programme, Centenary Institute, Newtown, Australia; 2 Discipline of Medicine, Central Clinical School, and 3 Discipline of Medicine, Western Clinical School, University of Sydney, Sydney, Australia; 4 Chest Clinic, Royal Prince Alfred Hospital, Camperdown, Australia; 5 Department of Medicine, Liverpool Clinical School, University of New South Wales, Sydney, Australia; and 6 Woolcock Institute of Medical Research, Camperdown, Australia Correspondence and requests for reprints should be addressed to Prof. W. J. Britton, Ph.D., M.B., Centenary Institute of Cancer Medicine and Cell Biology, Locked Bag No. 6, Newtown, 2042, NSW, Australia. E-mail: wbritton{at}med.usyd.edu.au
Rationale: Genetic variation influences susceptibility to clinical tuberculosis (TB). Activation of the P2X7 receptor on human macrophages induces killing of mycobacteria. We have identified polymorphisms in the P2X7 gene that markedly reduce this killing. Objective: To determine if polymorphisms in P2X7 are associated with increased risk of TB, the prevalence of four polymorphisms was assessed in individuals from Southeast Asia, where the majority of patients with TB in our study originate. The association of these polymorphisms with clinical TB was subsequently investigated in two separate case-control cohorts and the function of P2X7 was assessed in subjects from one cohort. Methods: Genotyping of P2X7 polymorphisms was performed from subjects in a nested case-control study of a longitudinal refugee cohort and a separate case-control study. The functional capacity of P2X7 was investigated by measuring ATP-mediated mycobacterial killing and apoptosis. Results: Only the 1513A-C polymorphism was present in Southeast Asians and the allele was associated with reduced killing of Mycobacterium tuberculosis. The 1513C allele was strongly associated with extrapulmonary, but not pulmonary, TB in the first (odds ratio, 3.8; 95% confidence interval, 1.6–9.0) and second cohorts (odds ratio, 3.7; 95% confidence interval, 1.7–8.0). ATP-mediated killing of mycobacteria was ablated in macrophages from subjects homozygous for the 1513C allele and significantly impaired in macrophages from heterozygous subjects. There was strong correlation between the degree of mycobacterial killing and ATP-induced apoptosis. Conclusions: The 1513C allele increases susceptibility to extrapulmonary TB, and this defect is associated with the reduction in the capacity of macrophages to kill M. tuberculosis.
Key Words: macrophage • tuberculosis • P2X7 • genes • susceptibility
Tuberculosis (TB) remains an enormous global health problem, resulting in 2 million deaths per year (1, 2). The majority of active TB results from reactivation of dormant infection (3), and the lifetime risk of reactivation for a 25-year-old with latent TB infection (LTBI) is 7.3% (4). HIV coinfection and other environmental causes of reduced T-cell immunity predispose to the progression to active TB in a proportion of subjects (5). Genetic variability also contributes to the risk of developing active TB. Polymorphisms in the HLA haplotype and in the genes for the vitamin D3 receptor, SCL11A1, IFN-  promoter, and mannose binding lectin (MBL) have all been associated with increased susceptibility to Mycobacterium tuberculosis, although, individually, the attributable risk of each polymorphism is modest (6). Therefore, the genetic influence is likely to be polygenic in nature. Many of the identified susceptibility alleles act at the level of the macrophage. Macrophages are the principal host cells for intracellular mycobacterial replication, and they are ultimately responsible for killing internalized bacilli through the generation of reactive nitrogen and oxygen intermediates (RNI, ROI) and the destructive action of lysosomal enzymes (7). Activation of the purinergic receptor P2X7 by extracellular ATP also induces killing of intracellular mycobacteria, independently of RNI or ROI (8–10).
P2X7 receptors are highly expressed on macrophages (11, 12). Activation of P2X7 causes an immediate opening of a cation-selective channel, and the influx of Ca2+, resulting in the induction of the caspase cascade, with resultant apoptosis (13–15), as well as the activation of phospholipase D (PLD) (16). PLD promotes phagosome–lysosome fusion, leading to the death of mycobacteria (8, 10, 17). We have identified several single nucleotide polymorphisms (SNPs) in P2X7 in whites that impair ATP-mediated mycobacterial killing (18). The most common of these, the 1513A To determine whether polymorphisms of P2X7 confer susceptibility to TB, we conducted a nested case-control study involving predominantly Southeast Asian refugees with LTBI, comparing the frequency of these polymorphisms in those refugees who developed TB over 15 years versus those who remained free of active disease. The reproducibility of these findings, as well as the effect of ATP treatment on mycobacterial killing, was then examined in a separate case-control study in patients with recently diagnosed TB.
Study Cohorts The prevalence of the 1513A  C, 1729TA, 946GA, and 151+1 gt SNPs in P2X7 was determined in 77 randomly chosen Southeast Asian individuals (see online supplement for further details). The Liverpool cohort comprised refugees who arrived in Sydney between 1984 and 1998, and were found to have LTBI at initial postarrival screening. LTBI was defined as a tuberculin skin test (TST) of 10 mm or greater and a normal chest X-ray (CXR). Refugees with active TB, diagnosed at premigration screening, were excluded from the cohort. Cases of TB that arose within the study population after arrival in Australia were identified by linkage with a database of TB cases who were notified to the NSW Department of Health between 1984 and 1998, as described previously (4, 28). For the analysis reported here, the linkage was extended to include cases registered between January 1999 and November 2003. As before, registered cases were accepted if there was microbiological or histologic confirmation of the diagnosis (28). In 2003, there were 119 such cases. We randomly selected 554 control subjects from the study population who did not develop TB during the follow-up period. Cases and control subjects were selected so that there were similar proportions with baseline TSTs of 10 to 14 mm and of 15 mm or greater in both groups. The Sydney cohort consisted of HIV-negative patients with active TB, confirmed microbiologically or histologically, who were recruited from TB clinics in Sydney between 2001 and 2004. Control subjects were recruited through the TB clinics and had evidence of LTBI, with a TST of 15 mm or greater and a normal CXR. Blood was collected for molecular and macrophage studies. Differences in ATP-mediated killing of M. tuberculosis were determined in 10 healthy subjects with loss-of-function polymorphisms in P2X7 and in six wild-type (WT) subjects. Ethics approval was obtained from the Central Sydney, Southwestern and Sydney West Area Health Services, the NSW Department of Health, and the University of Sydney. All participants donated blood after written consent.
Genotyping
Human Monocyte–derived Macrophage Cultures, Mycobacterial Killing, and Apoptosis Assays
Statistics
We have previously established that loss-of-function polymorphisms in P2X7 are associated with impaired ability to kill mycobacteria via ATP in vitro (18, 19). We examined the prevalence of four of these loss-of-function polymorphisms in healthy whites and in subjects from Southeast Asia, because the majority of patients with TB and their control subjects come from that region. Of these four, only the 1513A C SNP was present in Southeast Asians (Table 1). Therefore, we examined the association of the 1513C allele with TB in two independent case-control studies.
Increased Prevalence of the 1513C Allele in Patients with TB in the Liverpool Cohort The features of the two study groups are shown in Table 2. The database of refugees screened between 1984 and 1998 comprised 24,652 subjects. Of these, 8,609 had a TST reaction of greater than 10 mm and a normal CXR at the initial postarrival screening. Linkage of the two databases revealed 148 TB cases in the study population, of whom 119 had a TST of 10 mm or greater on enrollment. From these 119 cases, 90 (76%) were contactable and 86 agreed to participate (72% of all cases and 96% of those contacted). Five hundred and fifty-four control subjects were selected at random from the database. Of these, 167 agreed to participate (30% of selected control subjects and 89% of those contacted).
Univariate analysis of genotyping data revealed a significant association between the 1513C allele and TB in the Liverpool cohort with an OR of 1.7 (95% CI, 1.0–2.9; p < 0.05; Table 3). This association was only evident for extrapulmonary disease with an OR of 4.0 (95% CI, 1.7–9.3; p  0.001). In contrast, there was no significant relationship between the 1513C allele and pulmonary disease. Logistic regression analysis, adjusting for age, sex, BCG status, and TST, confirmed this strong association between the 1513C allele and extrapulmonary disease (adjusted OR, 3.8; 95% CI, 1.6–9.0; p < 0.01).
Increased Prevalence of 1513C Allele in a Second Cohort of Patients with TB The association between the 1513C allele and active TB was further tested by determining the prevalence of this allele in a second, separate cohort of recently diagnosed patients with TB and control subjects. In this Sydney cohort, univariate analysis revealed a significant association between the 1513C allele and active TB (OR, 1.9; 95% CI, 1.1–3.4; p < 0.05; Table 4). As in the Liverpool cohort, the association was only evident for extrapulmonary disease with an OR of 3 (95% CI, 1.5–6.1; p < 0.01). Analysis of the data for lymph node TB, separately from other forms of TB, revealed an association between the 1513C allele and the development of the lymph node TB, which was similar to the susceptibility of individuals with extrapulmonary TB (OR, 3.0; 95% CI, 1.2–7.6) There was also no significant relationship between the 1513C allele and pulmonary disease in this cohort. Among control subjects in this cohort, the prevalences of the 1513C allele were as follows: 24% in those born in Vietnam, Cambodia, or Laos; 31% in those born in other parts of Asia; and 46% in those born in other parts of the world (p = 0.17). Logistic regression analysis confirmed that the strong association between the 1513C allele and extrapulmonary disease was independent of age, sex, and region of origin (adjusted OR, 3.7; CI, 1.7–8.1; p = 0.001).
Reduced ATP-mediated Killing of Mycobacteria in Macrophages from Subjects with SNPs in P2X7 The effect of the SNPs in P2X7 on mycobacterial killing was examined in patients with TB and control subjects from the Sydney cohort along with 10 healthy subjects with a number of loss-of-function polymorphisms in P2X7. Macrophages were activated with IFN- , infected with Mycobacterium bovis BCG or M. tuberculosis, and then stimulated with ATP. Macrophages from subjects heterozygous for the 1513AC or one other nonfunctional allele also showed a significant impairment in ATP-mediated killing of M. tuberculosis, (p < 0.01; Figure 1). Moreover, ATP-mediated killing was abrogated in macrophages from 1513C homozygous and compound heterozygous individuals (p < 0.05, 1513C homozygotes/compound heterozygotes vs. WT subjects; p < 0.05, 1513C homozygotes/compound heterozygotes vs. single heterozygotes). Macrophages from patients with TB in the Sydney cohort, who were heterozygous for the 1513AC polymorphism in P2X7, displayed a significant impairment in their ability to kill M. bovis BCG after ATP stimulation, as compared with P2X7 WT macrophages (p < 0.0001; Figure 2A), and this ATP-mediated mycobacterial killing was ablated in macrophages from 1513C homozygous patients with TB. This effect was also apparent in macrophages from 1513C heterozygous and homozygous control subjects (Figure 2B). Hence, polymorphisms in P2X7 impair ATP-induced killing of both M. tuberculosis and M. bovis BCG (Figures 1 and 2) (18) in a cumulative fashion, regardless of disease status.
Reduced ATP-induced Apoptosis in Macrophages from Subjects with the 1513C SNP in P2X7 The capacity of ATP to induce apoptosis in BCG-infected macrophages was examined in individuals from the Sydney cohort. There was a significant reduction in ATP-induced apoptosis in macrophages from heterozygous and homozygous subjects in both patients with TB and control subjects (Figure 3). Spearman's rank test demonstrated a strong correlation between the levels of ATP-induced apoptosis and BCG killing in macrophages from patients with TB and control subjects (rs2 = 0.65, p < 0.0001; Figure 4).
These studies of two clinical cohorts provide strong evidence that the 1513A C loss-of-function polymorphism in P2X7 increases susceptibility to extrapulmonary TB. There is considerable polymorphic variation in P2X7 with four well-described SNPs leading to loss of receptor function (18). Three of these (151+1 gt, 946GA, 1729TA) were not present in subjects of Southeast Asian ethnicity. Therefore, our analyses focused on the polymorphism at nucleotide position 1513 of the gene. Although polymorphisms in over 10 other genes have been associated with active TB, only polymorphisms in the genes for HLA-DR molecules, SCL11A1, the vitamin D3 receptor, and MBL have been found to confer susceptibility to TB in multiple studies. The strength of our data lies in the reproducibility of the association between the 1513AC SNP in P2X7 and development of extrapulmonary TB in two independent populations. Although previous studies have revealed genetic associations with pulmonary TB, only a small number of genes have been associated with an increase in susceptibility to extrapulmonary TB, including MBL with pediatric meningitis (29), SCL11A1 with pleurisy (30), and IL-1Ra with extrapulmonary TB (31). Extrapulmonary TB is a major cause of morbidity and mortality (32). The mechanism of predisposition to extrapulmonary disease in individuals with the 1513C allele probably lies at the level of the macrophage. Mycobacteria reside in phagosomes within macrophages and killing of these mycobacteria requires fusion of the phagosome with a lysosome. ATP stimulation of P2X7 induces a range of cellular changes, including activation of PLD, phospholipase A2, and mitogen-activated protein kinase, which results in phagosome-lysosome fusion (9, 33). This ATP-mediated killing of mycobacteria by macrophages is independent of a range of other antimycobacterial mechanisms, such as ROI, RNI, SCL11A1, Fas ligation, and complement-mediated lysis (8). The source of ATP in vivo is unclear, but infection of macrophages with M. tuberculosis does result in increased ATP levels within the extracellular medium (34). Furthermore, respiratory epithelial cells infected with parainfluenza virus release ATP, which binds to purinergic receptors, triggering changes in ion transport and fluid accumulation in the respiratory tract (35). Although the concentration of ATP released into the immunologic synapse between infected macrophages and T cells within a granuloma has not been determined, it may be sufficient to stimulate mycobacterial killing by human macrophages in vivo.
The 1513A
Our study reflects the polygenic nature of the susceptibility to TB. Not all subjects with LTBI who harbor the 1513C SNP developed clinical TB, and some patients with TB were WT at this polymorphic site. Of 12 subjects who were WT or heterozygous for the 1513A
The authors thank K. Ng and C. V. Nhan for their recruitment of subjects for the Liverpool cohort study; B. De Souza, S. Simpson, R. Chidiac, W. Krause, M. Pavlovic, K. Mills, B. Hales, P. Wong, L. Tsang, D. Barnes, D. Frakes, D. Cruz, J. Branley, M. Peters, and E. Lagaac for their assistance in the recruitment of subjects for the Sydney-wide study; and S. Fuller, D. Williams, M. Karunakaran, S. Tran, N. Field, and K. Hall for their technical assistance. They are grateful to all the subjects who participated in this study.
Supported by the National Health and Medical Research Council of Australia, the Community Health Anti-Tuberculosis Association, the Cecilia Kilkeary Foundation, the South Western Area Health Service, and the New South Wales Department of Health, through its research infrastructure grant to the Centenary Institute. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200607-970OC on November 9, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form July 17, 2006; accepted in final form November 8, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
, IL-18, and matrix metalloproteinase (MMP)-9 from macrophages, and shedding of CD23 and CD62L from lymphocytes (
15 mm) using the Mantel-Haenszel method. Multivariate logistic regression was performed to control for the potential confounding effects of age, sex, bacille Calmette-Guérin (BCG) status, TST category, and region of origin. Differences in mycobacterial killing and apoptosis between the two groups were analyzed by the Mann-Whitney U test, and were considered significant when p values were less than 0.05. The correlation between mycobacterial killing and apoptosis was determined by Spearman's rank test. 
