MCP1 SNPs and Pulmonary Tuberculosis in Cohorts from West Africa, the USA and Argentina: Lack of Association or Epistasis with IL12B Polymorphisms

The monocyte chemotactic protein-1 (MCP-1) is a chemokine that plays an important role in the recruitment of monocytes to M. tuberculosis infection sites, and previous studies have reported that genetic variants in MCP1 are associated with differential susceptibility to pulmonary tuberculosis (PTB). We examined eight MCP1 single nucleotide polymorphisms (SNPs) in a multi-ethnic, case-control design that included: 321 cases and 346 controls from Guinea-Bissau, 258 cases and 271 controls from The Gambia, 295 cases and 179 controls from the U.S. (African-Americans), and an additional set of 237 cases and 144 controls of European ancestry from the U.S. and Argentina. Two locus interactions were also examined for polymorphisms in MCP1 and interleukin 12B (IL12B), another gene implicated in PTB risk. Examination of previously associated MCP1 SNPs rs1024611 (−2581A/G), rs2857656 (−362G/C) and rs4586 (+900C/T) did not show evidence for association. One interaction between rs2857656 and IL12B SNP rs2288831 was observed among Africans but the effect was in the opposite direction in Guineans (OR = 1.90, p = 0.001) and Gambians (OR = 0.64, p = 0.024). Our data indicate that the effect of genetic variation within MCP1 is not clear cut and additional studies will be needed to elucidate its role in TB susceptibility.


Introduction
Approximately one third of the world's population is infected with Mycobacterium tuberculosis (Mtb) with a global burden of TB disease in 2009 of 9.4 million incident cases, 14 million prevalent cases and more than 1.6 million deaths. According to the 2010 WHO global report on TB, the great majority of cases were in the South-East Asia, Africa and the Western Pacific (35%, 30% and 20%, respectively). An estimated 11-13% of incident cases were HIV-positive and approximately 80% of these cases were in Africa [1]. However, the majorities of those infected with Mtb maintain a latent state and do not convert to clinical disease but do remain at risk of progressing to active TB later. Factors that can modulate progression to active TB include gender, anemia, smoking and alcohol consumption as well as bacterial and host genetic factors [2][3][4]. In addition, increasing rates of TB and HIV have been highly correlated and a large percentage of TB cases are HIVpositive. Nonetheless, a substantial proportion of risk remains unexplained [5].
Evidence from human and animal studies indicates that Mtb clearance is genetically regulated [6]. Twin studies, genome-wide linkage and association analyses as well as candidate gene studies support the notion that human genetic factors play a role in the development of TB [7][8][9] and the majority of genes that have been implicated so far are in immunological pathways [8]. However, studies implicating genetic loci or specific genes have sometimes been inconsistent, possibly due to heterogeneity in phenotype definitions and study populations. Such heterogeneities are common challenges in global studies of etiologically complex traits, such as risk for Mtb infection and progression to TB [10].
The chemokine (C-C motif) ligand 2/monocyte chemotactic protein 1 gene (CCL2/MCP1) encodes the CCL2/MCP-1 protein, a member of the CC chemokine subfamily that is characterized by a two cysteine residue motif proximal to the amino-terminus of the protein. The MCP-1 chemokine plays a key role in the granulomatous reaction in lung tissue and Mtb containment in mouse models occurs through an interaction with the cognate receptor, chemokine (C-C motif) receptor 2 (CCR2), expressed on monocytes, macrophages, CD4+ T cells and immature dendritic cells [11]. The chemokine MCP-1 was associated with severe tuberculosis and was proposed as a marker of disease severity [12].
The 17q11-q21 chromosomal region encompassing MCP1 was initially identified as a candidate for TB susceptibility in linkage analyses of multi-case tuberculosis and leprosy families from Brazil and the critical interval was subsequently refined to 17q11.12 [13]. An MCP1 promoter variant has been associated with increased susceptibility to pulmonary TB (PTB), which is mediated through the inhibition of cytokine IL-12p40 production and is required for IFNc-induced protection from PTB. Functional studies by Flores-Villanueva et al. showed that the GG genotype at MCP1 ''22518'' (alias 22581 used in this paper, rs1024611) had the highest MCP-1 plasma levels and lowest IL-12p40 plasma concentrations in TB patients [14]; IL-12p40 (encoded by IL12B) is required for IFNcinduced protection from PTB and the GG homozygotes had 56 higher odds of developing TB than AA homozygotes [14]. The same SNP has also been associated with modulation of risk for spina bifida, coronary artery disease, and HIV-1, suggesting that rs1024611 could be a pleiotropic mutation with effects on different but key biological pathways [15][16][17].
Based on the potential biological role of MCP1 in PTB susceptibility and previous evidence implicating this gene, we undertook this study to investigate the association of gene variants of MCP1 with susceptibility to PTB in two West African populations (Guinea-Bissau and The Gambia) and to replicate the results in African-Americans and samples of European ancestry from North and South America. We assayed eight SNPs in MCP1 in DNA samples from 321 PTB cases and 346 controls from Guinea-Bissau, 258 PTB cases and 271 controls from The Gambia, 295 cases and 179 controls that are African-Americans, and 237 cases and 144 controls of European ancestry from North and South America. Genetic data were evaluated for association with PTB risk. In addition, motivated by our previous findings of an association between IL12B and PTB and the known biological interaction between the gene products, we assessed whether polymorphisms in IL12B modify PTB susceptibility due to MCP1 variation [18].

Study Populations
Detailed clinical and demographic information for subjects from Guinea-Bissau, The Gambia, the United States (African-Americans) and North and South Americans of European ancestry has been previously published [19][20][21]. A summary of basic demographic characteristics and study samples sizes is provided in Table 1.
Guinea-Bissau. This case-control study was conducted at the Bandim Health Project (BHP), a demographic surveillance site in Bissau, the capital of Guinea-Bissau. The incidence of PTB in this area is among the highest in the world, 470/100,000. Our Guinean cohort consisted of Papel (25%), Balanta (17%), Manjaco (14%), Fulani (13%), Mancanha (10%), Mandinka (7%), and other ethnicities (12%). Cases were residents or long-term guests of Bissau, aged greater than 15 years and newly diagnosed with PTB using three sputum examinations for acid fast bacteria or clinical criteria by the World Health Organization's definition of active pulmonary TB [22]. No culture confirmation of TB was available in Bissau during the study period, as facilities were destroyed during a civil war; 218/321 (68%) cases were smear positive. Patients with newly diagnosed TB were enrolled when they started antitubercular treatment at local health centers. During the inclusion period from November 2003 to November 2005, 438 TB patients were screened at local health centers: 344 subjects met inclusion criteria and provided written informed consent, and from these we could obtain 321 DNA samples.
Healthy controls were recruited from the study area from May 2005 to November 2005. A random sample of 200 houses was selected from the database of all subjects living in the study area; houses with a recorded case of TB within the past 2 years were excluded from the sampling. Exclusion criteria for controls included the presence of cough for more than 2 weeks, history of TB and TB in the household within the last 2 years to avoid households with a high-risk of active TB. The composition of the case and control samples was different in terms of sex and ethnicity. These differences are due to the sampling strategy as controls were derived from healthy nuclear families; hence more healthy married couples were collected, whereas TB patients are more often males. The ethnic differences are due to willingness of healthy subjects to give blood, which was not the same across the ethnic groups, whereas most TB patients agreed to participate in the study regardless of their ethnic background. All controls were unrelated to cases. Analyses were also performed excluding cases who only had clinical diagnoses to test for the effect of these samples on our results (Table S1).
All subjects were interviewed by field assistants, using a standardized questionnaire on ethnicity, environmental factors and prior exposure to TB. Permission to perform HIV tests was obtained for cases but not for controls, as requiring HIV testing would have negatively impacted participation in the study. Venous blood samples were collected from all subjects. Ethical approval was granted by the 'Unidade de Coordenacao de Estudos e Pesquisas em material de Saude' (Ministry of Health) in Guinea-Bissau. All adults and children's guardians signed a written informed consent to the study.
The Gambia. Between June 2002 and October 2004, PTB cases and their household contacts were enrolled in a prospective cohort study in the Greater Banjul region of The Gambia, where about 750,000 people live, representing more than 50% of the total Gambian population (http://www.columbia.edu/,msj42/ index.htm). Our Gambian cohort consisted of Mandinka (36%), Jola (28%), Wolof (12%), Fulani (9%), and other ethnicities (15%). According to WHO 2007 burden estimates, the incidence of TB in The Gambia is 258/100,000 and 11% of new TB cases are HIV positive (http://www.who.int/globalatlas/predefinedReports/TB/ PDF_Files/gmb.pdf). Recruitment took place at the major government TB clinic and the Medical Research Council (MRC) outpatient clinic, and consisted of sputum smear positive pulmonary TB cases at least 15 years old, who had at least one household contact living with them [19]. HIV positive patients were excluded from the study, and all included patients had two positive sputum smear samples for acid-fast bacilli and Mtb isolated upon culture. An index case was defined as the first TB case identified in a household. Household contact controls were defined as individuals living the majority of the time on the same compound as the index TB case, sharing meals and identifying a common household head. Written informed consent was obtained from all subjects, including parents/guardians for minors. The study was approved by the combined Gambia Government/MRC National Ethics Committee of The Gambia.
Subjects were interviewed by field assistants, using a standardized questionnaire on ethnicity, environmental factors and prior exposure to TB. Permission to perform HIV tests was asked for cases but not for controls. Venous blood samples were collected from all subjects, and from these, 258 case DNA samples and 271 control DNA samples (unrelated to cases) were genotyped for analyses. All samples were archived in the National Gambian DNA Bank and used in compliance with the bank guidelines [23].

African-Americans, European ancestry Americans and
Argentineans. Participants were ascertained through the North Carolina or South Carolina TB Control Programs, U.S.A., or as patients at the outpatient clinic at F.J. Muñ iz Hospital in Buenos Aires, Argentina, between 2002 and 2006. Criteria for inclusion as TB cases were: a) age 14 years of age or older and cultureconfirmed PTB, or b) younger than 14 years old and either culture-confirmed or clinically diagnosed PTB that included a positive tuberculin skin test plus an infiltrate or hilar adenopathy on chest x-ray. Some of the cases only had clinical diagnoses (7% for the African-American and European-American cases and 4% for the Argentinean cases). To test for the effect of these samples on the results they were excluded in a set of sensitivity analyses (Table S1).
Individuals were eligible to participate if TB had been diagnosed in the past, or if they were currently receiving TB treatment. All TB cases remained eligible if they also had a diagnosis of extrapulmonary TB. Family members of eligible TB cases, who themselves had a history of TB, were enrolled as part of a multi-case family if review of their records established diagnosis of either pulmonary or extrapulmonary TB. Thus, a small portion of our study subjects enrolled as part of a multi-case family had extrapulmonary TB only.
Severity of TB disease was assessed by presence of acid-fast bacilli (AFB) in sputum smears (7% African-Americans, 4% European-Americans/Argentineans) or x-ray evidence of cavitary lesions. We attempted to document HIV status through medical record review for all subjects. However, participation in this study did not require that the individual authorize review of HIV test results.
Unaffected individuals who were in close contact with cases, such as household contacts such as spouses and partners, and relatives such as parents and siblings, were enrolled as controls. Written informed consent was obtained from all subjects or their legal representatives before participation in the study. Human experimentation guidelines of the U.S. Department of Health and Human Services and those of the participating research institutions were followed. The protocol was IRB-approved at Duke University Medical Center, the North and South Carolina

DNA extraction and genotyping
SNPs were selected based on either being associated with PTB in previous studies (rs1024611, rs2857656 and rs4586) or being a haplotype tagging SNP in the MCP1 gene [14,24,25]. Tags were selected from HapMap phase III samples: African-Americans (from the SW USA), Africans (Yoruba, Maasai, and Luhya), Mexicans (from Los Angeles, USA) and Northern and Western Europeans (Centre d'Etude du Polymorphisme Humain (CEPH) family samples from Utah, USA) (http://www.hapmap.org). To focus analysis on common variants for which these samples have the greatest statistical power to detect effects, SNPs with minor allele frequency greater or equal to 0.1 and located in a region extending 3 kb on either side of the gene were identified from the HapMap and Genome Variation Server databases and were grouped into bins of highly correlated SNPs (r 2 greater than or equal to 0.80). A single ''tagSNP'' was selected from each bin for genotyping. A summary of the SNPs examined is provided ( Figure 1). All DNA samples were extracted using a standard salting-out procedure (Guinea-Bissau and The Gambia) or the Puregene method from Gentra systems (African-Americans and European-Americans/Argentineans). DNA purities were estimated spectrophotometrically, and final concentrations were determined by PicoGreen. One of the SNPs (rs1024611) genotyped in Guinea-Bissau and The Gambia samples was genotyped by TaqMan assay (ABI, Applera International Inc, Foster City, CA, USA) in 10 ml reaction volume, using the Rotor-Gene 3000 (Corbett Robotics Pty Ltd, Brisbane, Queensland, Australia) and the ABI 7500 real-time PCR system. Fluorescence curves were analyzed with the Rotor-Gene Software version 6 and the 7500 Sequence Detection Software version 1.2.1 for allelic discrimination. The remaining SNPs in all populations (Guinea-Bissau, The Gambia, African-Americans, and European-Americans/Argentineans) were genotyped using TaqMan assays on an ABI 7900 HT with genotype calling performed using ABI SDS software. All SNPs used in this study had genotyping call rates of 95% or better (mean call rates of 98%) and quality control duplicate sample match rates of 100%.

Bioinformatics Tools
SNP base pair (bp) position and function was identified using the SNPper (http://snpper.chip.org) database NCBI Build 35.1 ( Figure 1). The HapMap database (http://www.hapmap.org) was used to obtain linkage disequilibrium (LD) and genotype information from the Yoruba and CEPH populations.

Statistical Methods
All analyses were performed separately for the four cohorts (Guinea-Bissau, The Gambia, African-Americans, and European-Americans/Argentineans). For Guinea-Bissau and The Gambia tests for deviations from Hardy-Weinberg Equilibrium (HWE) were performed using PLINK statistical software [26]. Tests for deviations from HWE in African-Americans and European-Americans/Argentineans were calculated using genetic data analysis (GDA) software using one case and one control from each pedigree [27]. Statistical significance for these analyses was determined using Fisher's exact test.
Pairwise LD was characterized, standard summary statistics D9 and r 2 , and haplotype frequencies were calculated using Haplo-View statistical software [28]. Haplotype blocks were assigned, using the D9 confidence interval algorithm created by Gabriel et al. 2002 [29]. Haplotype analyses were performed Guinea Bissau and The Gambia with 3 and 8-marker sliding windows using PLINK software. Analyses were run using haplotype-bases association test with generalized linear models (GLMs) adjusting for age, sex, and ethnic group. Haplotype analyses only included common haplotypes (haplotype frequency $0.05) and statistical significance was assessed with 1,000 permutations to generate empirical p values.
Single locus tests of association in Guinea-Bissau and The Gambia were performed using logistic regression models with PLINK software [26] assuming an additive genetic model. Odds ratios (ORs) and confidence intervals (CI) were reported for all statistical models. Confounding by age, ethnicity, and sex was evaluated in logistic regression models; for inclusion in the final model we required that a change in effect size for the SNP be greater than or equal to 0.05. As a result, all regression models were performed with an adjustment for age, ethnicity, and sex. Unadjusted models are presented in Table S2. A 2-degree of freedom genotypic test of association was also performed with PLINK software for Guinea-Bissau and The Gambia for modelfree tests of association.
For African-Americans and European-Americans/Argentineans single locus additive genotypic tests of association were performed with generalized estimating equations (GEE) using the indepen-  dence correlation matrix implemented in STATA 11.0 statistical software (College Station, TX). GEE performs a valid test of gene6gene and gene6environment interactions in mixed family and case-control data [30]. We also performed GEE analyses adjusting for the potential confounders, age and sex for all analyses. For European-Americans/Argentineans we did the analyses two ways. First, we performed the analyses in these two cohorts separately; second, we incorporated ascertainment site in the models, using the combined data (Table S3). Because there was no evidence of a recruitment site effect between Argentinian and European-American cohorts, we pooled these samples in subsequent analyses.
Two locus interaction analyses were performed to test for interactions between MCP1 and IL12B based on the observation in previous studies that a functional relationship exists between MCP1 promoter polymorphisms and IL12B concentrations [14]. The main effects for the polymorphisms examined in IL12B in our cohort have been previously published in Morris et al. 2011 [18] (a list of those variants is provided in Table S4). Two locus interactions were examined between MCP1 and IL12B polymorphisms with a MAF greater than 0.05 within a population and were performed with logistic regression for Guinea Bissau and The Gambia and with GEE for African-Americans and European-Americans/Argentineans using STATA 11.0 statistical software   (College Station, TX). These analyses were performed adjusting for the same covariates used in single locus tests of associations. A Bonferroni correction for multiple testing was used to adjust for multiple testing for single locus and gene-gene interactions. Examination of allele and genotype frequency differences for cases with and without HIV indicated no evidence for significant differences between the two groups in any of the populations examined; as a result analyses were performed pooling cases with and without HIV. Sensitivity analyses were also performed by excluding clinically diagnosed TB cases for Guinea Bissau,  African-Americans, and European-Americans/Argentineans. Overall, these analyses showed no difference in the association results.
Finally, in order to assess the concordance of results across study populations meta-analyses of single SNP associations were performed using PLINK software [26]. Meta-analyses were run both including and excluding the European/Argentinean study population.

Guinea-Bissau and The Gambia
Single locus tests of association did not identify a statistically significant association at any of the SNPs examined with either the logistic regression model for allelic association (Table 2) Table 3). SNP rs2530797 was in LD with both previously associated SNPs rs1024611 (D9 = 1; r 2 = 0.03) and rs2857656 (D9 = 1; r 2 = 0.08) in the Guinea-Bissau population (Figure 2A and 2B). No statistically significant haplotype associations were found in the Guinean and Gambian cohorts (Table S5).
Interaction analyses revealed evidence for a gene-gene interaction between MCP1 SNP rs2857656 and IL12B SNP rs2288831 (Table  S6, OR INT = 1.90, 95% CI [1.31-2.77], p = 0.001) in Guinea-Bissau; however, this result was not in the same direction in The Gambia, despite being statistically significant (p = 0.024). Neither of these results was significant after correction for multiple testing.

African-Americans and European-Americans/ Argentineans
Single locus tests of association did not identify a statistically significant association at any of the SNPs in the African-American and European-American/Argentinean populations (Table 4). LD plots for African-American and for European-American/Argentinean controls are shown in Figure 3, respectively 3A-B and 3C-D. There were no statistically significant interactions between MCP1 and IL12B in European-Americans/Argentineans; however, in African-Americans there was one statistically significant interaction between MCP1 rs3917891 and IL12B rs11574790 (OR INT = 0.28, 95% CI [0.13-0.65], p = 0.003) ( Table S6). None of the associations were statistically significant after correction for multiple testing.
The strongest evidence for concordant single SNP association across study populations was for MCP1 rs2530797, (OR meta = 1.30, random effect meta-analysis p = 0.051). Cochrane's Q statistic (p$0.36) and the I 2 index of heterogeneity (I#3.36, scale 0-100) indicated little evidence for heterogeneity for this association across study populations for these SNPs. The results did not change when removing the European American/ Argentinian sample (data not shown).

Discussion
In the present study we examined eight SNPs in MCP1 for association with PTB in two African populations, one African-American population, and populations of European ancestry from North and South America. We focused on variants that had been shown to be associated with either increased or decreased TB risk, although previous studies were inconsistent, possibly reflecting differences both in genetic structure and phenotype definitions.
We did not observe any statistically significant association at the SNPs studied in Guineans, Gambians, African-Americans and European-Americans/Argentineans. Examination of all previously associated SNPs did not provide evidence for association in any of our populations. We observed a statistically significant interaction between MCP1 and IL12B in the West African cohorts; however, the association was in the opposite direction in the two populations, indicating that this is likely to be spurious.
Our data is in contrast to the majority of that published to date (Table S7). Specifically for the SNPs we genotyped that have been previously examined we found: 1) rs1024611 (22581A/G) associations with PTB have been reported in Mexican, Korean, Ghanaian, Zambian, Tunisian, Moroccan and Peruvian cohorts [14,24,[31][32][33][34]. This SNP, originally reported by Flores-Villanueva et al., promoted subsequent genetic studies of MCP1 and TB [14]. However, while in Ghanaians the ''G'' allele and the ''AG+GG'' genotypes were found to confer protection from PTB (OR = 0.81) [24], in a Moroccan sample only the ''GG'' genotype showed the same effect (OR = 0.35) [33]; in all other populations typed the ''G'' allele associated with increased risk (e.g. OR = 2.63 in Koreans, OR = 1.29 in Peruvians) [14,34]. We also observed a trend in the direction of increased risk. The heterogeneous effect of the ''G'' allele was also reported in a meta-analysis by Feng et al. (Table S7) [35]. 2) rs2857656 (2362G/C): in Ghanaians, the ''C'' allele and ''CG+CC'' genotypes were associated with protection from Statistical models for African-Americans included an adjustment for age and sex and European-Americans/Argentineans also included an adjustment for ascertainment site. 1 A dominant model was used to calculate the association p value because the number of individuals in the rare homozygous class was below 5 in cases, controls, or both. 2 OR is for additive model except for those instances where a dominant model was used. doi:10.1371/journal.pone.0032275.t004 Table 4. Cont.
disease (OR = 0.83) [24]. In the same population, when SNPs rs1024611 and rs2857656 were combined in an extended haplotype including a 14 bp insertion/deletion, rs3917887 (haplotype: ''rs1024611 G-rs2857656 C-rs3917887 del''), the protective effect size increased (OR = 0.78) [36], though the combination of genetic data with functional assays indicated that the role of rs1024611 was small at best. According to Thye et al. rs2857656 was the variant driving the protective effect  originally observed in Ghanaians (see Table S7). Our data provide no evidence for association of rs2857656. 3) rs4586 (+900C/T): the ''C'' allele and the ''CC+TC'' genotypes were associated with increased risk of TB (OR = 1.34 and 1.94, respectively), in a pediatric cohort from Northern China that was heterogeneous with respect to phenotype definitions, ranging from pulmonary TB (35%) to extra-pulmonary TB (26%) and TB meningitis (39%) [25]; moreover, the association was only found in males. This SNP was not significantly associated with TB in the Ghanaian study [24] nor in ours.
Analyses of extended haplotypes in our Guineans and Gambians, encompassing polymorphisms associated with protection from TB in Ghanaians [36], did not reveal any association signal.
Taken together our data do not support any of these previously associated variants.
We also tested for interactions between MCP1 and IL12B based on prior experimental evidence supporting an interaction between these two genes [14] and our own recent study showing an IL12B association with PTB susceptibility in the same African ancestry samples examined in this paper [18]. Although we did observe some weak evidence for risk-modulating interactions in Guineans, Gambians and African-Americans, none of the MCP1-IL12B effects remained statistically significant after correction for multiple testing, and, most importantly, they were not always in the same direction. In conclusion, our data indicate that there was no evidence for a genetic interaction between MCP1 and IL12B with respect to susceptibility to PTB.
In our study there were aspects that are worth discussing. We used two different ascertainment centers for the collection of European-ancestry samples, one from the Southeastern U.S. and another from Argentina, and they included a small number of HIV-positive cases as well as cases of extrapulmonary TB. In order to account for this we included ascertainment site as a variable in our models and performed sensitivity analyses excluding HIVpositive and extrapulmonary TB cases. Site, HIV status and extrapulmonary TB did not influence the significance of our results. Both the Guinea-Bissau and The Gambia population samples showed evidence for confounding by ethnic groups, but we dealt with this by adjustment for ethnicity. Finally, there were some limitations regarding power to detect effect sizes previously found. Within our Gambian cohort we had approximately 80% power to detect OR ranging between ,0.49 or .1.8 with a MAF of 0.20, while, in our Guinea Bissau cohort with the same MAF we had 80% power to detect OR ranging between ,0.53 .1.70. Although we were underpowered to detect associations of the effect size reported in the Ghanaian population (OR = 0.81 for rs1024611 and OR = 0.83 for rs2857656) [24], we failed to detect effects in the same direction as those previously published for both rs1024611 and rs2857656 in our West African samples. These discordant results are cause for caution in interpreting the role of MCP1 in TB susceptibility. However, it is possible that the lack of replication is due to unmeasured variables interacting with the MCP1 SNPs. Although we would have been able to have increased power by pooling the Guinea Bissau and Gambian cohorts, several studies have shown significant genetic heterogeneity across African populations, even within limited geographical areas [20,37], and for this reason we chose not to pool our populations but instead to meta-analyze the results. Meta-analysis did not detect statistically significant, robust results across the studies. Our findings of interactions in opposite directions in these two cohorts support the decision not to pool.
In conclusion, this study did not replicate associations with TB previously observed in MCP1. Although this is a highly relevant candidate gene, our data indicate that the effect of genetic variation within MCP1 is not clear cut and additional studies will be needed to elucidate its role in TB susceptibility.

Supporting Information
Table S1 Sensitivity analysis including only culture confirmed or smear positive TB cases adjusted for covariates.
(DOC)   Table   S6 Top (p,0.05) gene6gene interaction (MCP16IL12B) results across cohorts. Two locus interactions are presented between MCP1 and IL12B polymorphisms with a MAF greater than 0.05 within a population. These analyses were performed with logistic regression for Guinea Bissau and The Gambia and with GEE for African-Americans and European-Americans/Argentineans using STATA 11.0 statistical software (College Station, TX) and were performed adjusting for the same covariates used in single locus tests of associations. A Bonferroni correction for multiple testing was used to adjust for multiple testing for single locus and gene6gene interactions. The gene6gene interactions results are presented according to increasing p values. (DOC)