MCP1 SNPs and pulmonary tuberculosis in cohorts from West Africa, the USA and Argentina: lack of association or epistasis with IL12B polymorphisms.

Abstract

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.

Full text

Page 1

MCP1 SNPs and Pulmonary Tuberculosis in Cohorts from
West Africa, the USA and Argentina: Lack of Association
or Epistasis with IL12B Polymorphisms
Digna R. Velez Edwards1,2,3, Alessandra Tacconelli4, Christian Wejse5,6,7, Philip C. Hill8,9, Gerard A. J.
Morris8, Todd L. Edwards1,10, John R. Gilbert1, Jamie L. Myers1, Yo Son Park1, Martin E. Stryjewski11,
Eduardo Abbate12, Rosa Estevan12, Paulo Rabna5, Giuseppe Novelli4,13, Carol D. Hamilton14,15, Richard
Adegbola8,16, Lars Østergaard6, Scott M. Williams2, William K. Scott1*, Giorgio Sirugo4,8*
1Dr. John T. Macdonald Foundation Department of Human Genetics and Miami Institute of Human Genomics, University of Miami, Miami, Florida, United States of
America, 2Center for Human Genetics Research, Vanderbilt University, Nashville, Tennessee, United States of America, 3Vanderbilt Epidemiology Center, Institute of
Medicine and Public Health, and Department of Obstetrics and Gynecology, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America, 4Centro
di Genetica, Centro di Ricerca Scientifica, Ospedale San Pietro FBF, Rome, Italy, 5Bandim Health Project, Danish Epidemiology Science Centre and Statens Serum Institute,
Bissau, Guinea-Bissau, 6Department of Infectious Diseases, Aarhus University Hospital, Skejby, Denmark, 7Center for Global Health, School of Public Health, Aarhus
University, Skejby, Denmark, 8MRC Laboratories, Fajara, The Gambia (West Africa), 9Centre for International Health, University of Otago School of Medicine, Dunedin,
New Zealand, 10Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt University School of Medicine, Nashville, Tennessee,
United States of America, 11Division of Infectious Diseases, Department of Medicine, Centro de Educacio ´n Me ´dica e Investigaciones Clı ´nicas ‘‘Norberto Quirno’’ (CEMIC),
Buenos Aires, Argentina, 12Department of Medicine, Hospital F. J. Mun ˜iz, Buenos Aires, Argentina, 13Dipartimento di Biopatologia e Diagnostica per Immagini,
Universita ` di Tor Vergata, Rome, Italy, 14Family Health International 360, Research Triangle Park, North Carolina, United States of America, 15Duke University Medical
Center, Durham, North Carolina, United States of America, 16Bill & Melinda Gates Foundation, Seattle, Washington, United States of America
Abstract
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 (22581A/G), rs2857656 (2362G/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.
Citation: Velez Edwards DR, Tacconelli A, Wejse C, Hill PC, Morris GAJ, et al. (2012) MCP1 SNPs and Pulmonary Tuberculosis in Cohorts from West Africa, the USA
and Argentina: Lack of Association or Epistasis with IL12B Polymorphisms. PLoS ONE 7(2): e32275. doi:10.1371/journal.pone.0032275
Editor: Pere-Joan Zealand, Fundacio ´ Institut d9Investigacio ´ en Cie `ncies de la Salut Germans Trias i Pujol - Universitat Auto `noma de Barcelona - CIBERES, Spain
Received September 22, 2011; Accepted January 25, 2012; Published February 27, 2012
Copyright: ? 2012 Velez Edwards et al. 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 author and source are credited.
Funding: This study was supported by the Medical Research Council UK (MRC) award G0000690 to GS; MRC (UK) The Gambia Unit core-funding to PCH and RA;
grants from the Danish Medical Research Council, the Danish Society of Respiratory Medicine, the Danish Council of Development Research to CW and LO, and by
the United States National Institutes of Health grant R01 HL68534 to WKS. The funders had no role in study design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: sirugo.giorgio@fbfrm.it (GS); BScott@med.miami.edu (WKS)
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
[2–4]. In addition, increasing rates of TB and HIV have been
highly correlated and a large percentage of TB cases are HIV-
positive. 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
PLoS ONE | www.plosone.org1 February 2012 | Volume 7 | Issue 2 | e32275

Page 2

support the notion that human genetic factors play a role in the
development of TB [7–9] 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 IFNc-
induced 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 [15–17].
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].
Materials and Methods
Study Populations
Detailed clinical and demographic information for subjects from
Guinea-Bissau, The Gambia, the United States (African-Ameri-
cans) and North and South Americans of European ancestry has
been previously published [19–21]. A summary of basic
demographic characteristics and study samples sizes is provided
in Table 1.
Guinea-Bissau.
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
This case-control study was conducted at the
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org2February 2012 | Volume 7 | Issue 2 | e32275

Page 3

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 standard-
ized 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. Mun ˜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 culture-
confirmed 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 institu-
tions were followed. The protocol was IRB-approved at Duke
University Medical Center, the North and South Carolina
Departments of Public Health (USA), Centro de Educacio ´n
Me ´dica e Investigaciones Clı ´nicas ‘‘Norberto Quirno’’ (CEMIC),
the F.J. Mun ˜iz Hospital, Buenos Aires, Argentina, and the
University of Miami Miller School of Medicine.
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
Table 1. Demographic data summary.
Guineans (Guinea-Bissau)GambiansAfrican-Americans
European Americans/
Argentinians
Cases
(N=321)
Controls
(N=346)
Cases
(N=258)
Controls
(N=271)
Cases
(N=295)
Controls
(N=179)
Cases
(N=237)
Controls
(N=144)
Age, mean 6 SD, (years) 37.08613.7335.58612.4033.34613.6329.09613.1445.48617.86 51.91621.24 39.20617.6140.67620.28
Sex (%)
Male60.4449.7169.3869.7465.9783.5352.2062.12
Female35.5650.2930.6230.2634.0216.4747.737.0
Ethnicity (%)
Balanta 15.2619.36------
Fulani14.9511.5612.18 6.27----
Jola--21.4334.12
Mancanha8.1012.43------
Mandinka7.487.5139.0832.55----
Manjaco19.009.54------
Papel20.2529.77--
Wolof--11.3413.33----
Other14.959.83 15.9713.73--
doi:10.1371/journal.pone.0032275.t001
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org3 February 2012 | Volume 7 | Issue 2 | e32275

Page 4

HapMap and Genome Variation Server databases and were
grouped into bins of highly correlated SNPs (r2greater 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 spectro-
photometrically, 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,andEuropean-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 r2, 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 model-
free 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-
Figure 1. MCP1 Gene Structure and SNP Positions. MCP1 is oriented 59 to 39 with SNPs indicated above the gene with rs numbers and position
relative to coding sequence start site.
doi:10.1371/journal.pone.0032275.g001
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org4February 2012 | Volume 7 | Issue 2 | e32275

Page 5

Table 2. Guineans and Gambians single locus association results adjusted for age, ethnicity and sex.
PopulationMarkerGenotype
Genotype Counts
OR2
95% CI
Additive
p-Value
CasesControlsLowerUpper
Guineansrs1024611*GG17211.230.921.630.163
AG 123103
AA174217
rs10246101
TT200.64 0.361.15 0.137
AT 2238
AA288305
rs3760396CC001.00 0.462.24 0.999
CG 1616
GG293 323
rs2857656 CC 5663 0.970.761.230.786
CG 150158
GG107119
rs4586TT35 29 1.050.811.360.702
CT129143
CC149169
rs39178911
TT1100.890.60 1.350.604
CT6970
CC244 261
rs41416652CC00----
CT10
TT 309342
rs2530797CC57 1.39 0.952.020.088
CT7159
TT240 277
Gambiansrs1024611*GG18 151.000.751.340.991
AG80 93
AA 138144
rs10246101
TT100.930.511.700.818
AT 2429
AA219 230
rs37603961
CC001.59 0.614.180.346
CG 118
GG 231253
rs2857656CC44500.940.72 1.220.623
CG116129
GG80 81
rs4586TT 2220 1.040.781.38 0.798
CT96 109
CC126131
rs3917891 TT660.86 0.591.25 0.427
CT 5663
CC181 187
rs41416652CC00----
CT 00
TT247 266
rs25307971
CC53 1.020.631.630.951
CT 38 44
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org5 February 2012 | Volume 7 | Issue 2 | e32275

Page 6

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 polymor-
phisms 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
PopulationMarker Genotype
Genotype Counts
OR2
95% CI
Additive
p-Value
Cases ControlsLowerUpper
TT 204216
*Indicates a statistically significant deviation from HWE in either cases or controls.
1A 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.
2OR is for additive model except for those instances where a dominant model was used.
doi:10.1371/journal.pone.0032275.t002
Table 2. Cont.
Table 3. Additive, dominant, and recessive regression models for SNPs previously associated in published studies adjusted for age,
ethnicity and sex.
Population MarkerModelOR
95% CI
p-Value
LowUpper
Guineansrs1024611AA(referent) vs AG vs GG1.230.921.630.163
AA(referent) vs AG&GG 1.360.961.930.079
AA&AG(referent) vs GG0.960.46 2.010.915
rs2857656GG(referent) vs GC vs CC0.970.761.23 0.786
GG(referent) vs GC&CC0.970.681.380.851
GG&GC(referent) vs CC 0.940.611.460.792
Gambiansrs1024611AA(referent) vs AG vs GG1.000.75 1.340.991
AA(referent) vs AG&GG0.94 0.651.350.744
AA&AG(referent) vs GG1.280.622.650.499
rs2857656 GG(referent) vs GC vs CC 0.94 0.721.22 0.623
GG(referent) vs GC&CC0.900.62 1.330.611
GG&GC(referent) vs CC 0.940.581.500.783
African-Americansrs1024611AA(referent) vs AG vs GG1.270.831.930.272
AA(referent) vs AG&GG1.310.812.100.270
AA&AG(referent) vs GG1.33 0.335.410.687
rs2857656GG(referent) vs GC vs CC1.120.60 2.090.717
GG(referent) vs GC&CC1.210.761.920.418
GG&GC(referent) vs CC0.660.351.25 0.204
European-Americans/Argentineansrs1024611AA(referent) vs AG vs GG 1.180.632.230.599
AA(referent) vs AG&GG1.780.59 5.350.303
AA&AG(referent) vs GG0.96 0.362.53 0.933
rs2857656 GG(referent) vs GC vs CC1.12 0.602.09 0.717
GG(referent) vs GC&CC1.82 0.615.470.285
GG&GC(referent) vs CC0.840.332.170.723
doi:10.1371/journal.pone.0032275.t003
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org6 February 2012 | Volume 7 | Issue 2 | e32275

Page 7

(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,
Figure 2. Guinea-Bissau and The Gambia HaploView plots for controls (MCP1). LD plots are presented for Guinea Bissau controls (A and B)
and for The Gambia controls (C and D) including both D9 and r2. All figures are oriented 59 to 39, right to left, relative to the gene orientation on the
minus strand. D9 (shades of red) and r2(shades of black) are indicated in percentages within squares in the LD plots, with solid blocks without
numbers indicating D9 and r2=1. Strong LD is indicated by red or dark gray, while pink and light gray and white indicate uninformative and low
confidence values, respectively. LD Blocks were created with the default algorithm in HaploView that creates 95% confidence bounds on D9
considered being in strong LD where 95% of the comparisons made are informative. The haplotype blocks were created using HaploView program,
version 4.1.
doi:10.1371/journal.pone.0032275.g002
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org7 February 2012 | Volume 7 | Issue 2 | e32275

Page 8

Table 4. African-Americans and European-Americans/Argentineans single locus association results using an additive GEE model
adjusted for age and sex.
PopulationMarkerGenotype
Genotype Counts
OR2
95% CI
Additive
p-Value
CasesControlsLowUpper
African-Americansrs10246111
GG841.310.832.050.243
AG9149
AA 189 123
rs10246101
TT201.070.54 2.130.848
AT 3622
AA250151
rs37603961
CC10 1.290.692.39 0.425
CG2614
GG 259158
rs2857656 CC38 270.99 0.71 1.380.947
CG 14081
GG11165
rs4586TT42240.800.581.110.179
CT12460
CC12188
rs39178911
TT740.720.431.200.207
CT59 50
CC222117
rs41416652 CC10----
CT41
TT 284174
rs25307971
CC 1111.56 0.91 2.690.100
CT73 39
TT 202133
European-Americans/Argentiniansrs1024611 GG68 321.18 0.632.23 0.599
AG95 53
AA70 55
rs1024610 TT560.70 0.19 2.550.585
AT 56 35
AA 171 102
rs37603961
CC350.550.231.290.168
CG 4432
GG183 103
rs2857656 CC 69331.120.602.09 0.717
CG 96 54
GG71 54
rs4586 TT44 391.370.682.79 0.379
CT102 57
CC8345
rs3917891TT00----
CT22
CC228 140
rs414166521
CC12 2 1.54 0.613.88 0.356
CT 47 17
TT176121
rs2530797CC14 171.020.452.31 0.971
CT 9753
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org8February 2012 | Volume 7 | Issue 2 | e32275

Page 9

African-Americans,
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.
andEuropean-Americans/Argentineans.
Results
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) or using the
2 degree of freedom genotypic test (results not shown). A borderline
association was observed in the Guinea-Bissau populationbutnot in
The Gambia at MCP1 marker rs2530797 with an OR=1.39, 95%
CI [0.95–2.02], p=0.088 (Table 2). The association became less
statistically significant for the dominant model (CC (referent) versus
CT & TT, OR=1.25, 95% CI [0.33–4.78], p=0.742) and the
recessive model (results not shown). Detailed analyses of the
previously associated SNPsrs1024611 andrs2857656 demonstrated
a borderline significant association in the Guinea-Bissau population
for rs1024611 under the dominant model (AA (referent) versus AG
& GG, OR=1.36, 95% CI [0.96–1.93], p=0.079) (Table 3). SNP
rs2530797 was in LD with both previously associated SNPs
rs1024611 (D9=1; r2=0.03) and rs2857656 (D9=1; r2=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, ORINT=1.90, 95% CI[1.31–2.77], p=0.001) inGuinea-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/Argenti-
nean 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 interac-
tion between MCP1 rs3917891 and IL12B rs11574790 (ORINT=
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, (ORmeta=
1.30, random effect meta-analysis p=0.051). Cochrane’s Q
statistic (p$0.36) and the I2index 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, Tuni-
sian, Moroccan and Peruvian cohorts [14,24,31–34]. This
SNP, originally reported by Flores-Villanueva et al., promot-
ed 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].
rs2857656 (2362G/C): in Ghanaians, the ‘‘C’’ allele and
‘‘CG+CC’’ genotypes were associated with protection from
2)
PopulationMarker Genotype
Genotype Counts
OR2
95% CI
Additive
p-Value
CasesControls LowUpper
TT11769
Statistical models for African-Americans included an adjustment for age and sex and European-Americans/Argentineans also included an adjustment for ascertainment
site.
1A 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.
2OR is for additive model except for those instances where a dominant model was used.
doi:10.1371/journal.pone.0032275.t004
Table 4. Cont.
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org9 February 2012 | Volume 7 | Issue 2 | e32275

Page 10

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 wassmall at best. Accordingto Thye
et al. rs2857656 was the variant driving the protective effect
Figure 3. African-Americans and European-Americans/Argentineans HaploView plots for controls (MCP1). LD plots are presented for
African-American controls (A and B) and for European-American/Argentinean controls (C and D) including both D9 and r2. All figures are oriented 59 to
39, right to left, relative to the gene orientation on the minus strand. D9 (shades of red) and r2(shades of black) are indicated in percentages within
squares in the LD plots, with solid blocks without numbers indicating D9 and r2=1. Strong LD is indicated by red or dark gray, while pink and light
gray and white indicate uninformative and low confidence values, respectively. LD Blocks were created with the default algorithm in HaploView that
creates 95% confidence bounds on D9 considered being in strong LD where 95% of the comparisons made are informative. The haplotype blocks
were created using HaploView program, version 4.1.
doi:10.1371/journal.pone.0032275.g003
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org10February 2012 | Volume 7 | Issue 2 | e32275

Page 11

originally observed in Ghanaians (see Table S7). Our data
provide no evidence for association of rs2857656.
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.
3)
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 HIV-
positive 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
confirmed or smear positive TB cases adjusted for
covariates.
(DOC)
Sensitivity analysis including only culture
Table S2
and Gambians unadjusted for age, sex and ethnicity.
(DOC)
Single locus tests of association in Guineans
Table S3
Argentineans single locus association results using an
additive GEE model unadjusted for age and sex.
(DOC)
African-Americans and European-Americans/
Table
MCP16IL12B interaction analyses.
(DOC)
S4
IL12B polymorphismsexaminedin
Table S5
sliding window haplotype analysis.
(DOC)
Guineans and Gambians 8 and 3 marker
Table
(MCP16IL12B) results across cohorts. Two locus interac-
tions 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)
S6
Top (p, ,0.05) gene6geneinteraction
Table S7
pulmonary tuberculosis.
(DOC)
Genetic association of MCP1 variants with
Acknowledgments
We thank Dr. Luca Lavra, Centro di Ricerca, Ospedale San Pietro FBF,
for help with editing.
Author Contributions
Conceived and designed the experiments: DRVE PCH CW CDH WKS
GS. Performed the experiments: GAJM JRG JLM MES EA PR. Analyzed
the data: DRVE TLE SMW WKS GS. Contributed reagents/materials/
analysis tools: YSP RE GN RA LO. Wrote the paper: DRVE AT SMW
WKS GS.
References
1. World Health Organization (2010) Global tuberculosis control. WHO report
2010 World Health Organization.
2. Flynn JL, Chan J (2001) Immunology of tuberculosis. Annu Rev Immunol 19:
93–129.
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org 11February 2012 | Volume 7 | Issue 2 | e32275

Page 12

3. Moller M, de Wit E, Hoal EG (2010) Past, present and future directions in
human genetic susceptibility to tuberculosis. FEMS Immunol Med Microbiol 58:
3–26.
4. Lienhardt C, Bennett S, Del PG, Bah-Sow O, Newport M, et al. (2002)
Investigation of environmental and host-related risk factors for tuberculosis in
Africa. I. Methodological aspects of a combined design. Am J Epidemiol 155:
1066–1073.
5. Frieden TR, Sterling TR, Munsiff SS, Watt CJ, Dye C (2003) Tuberculosis.
Lancet 362: 887–899.
6. Berrington WR, Hawn TR (2007) Mycobacterium tuberculosis, macrophages,
and the innate immune response: does common variation matter? Immunol Rev
219: 167–186.
7. Comstock GW (1978) Tuberculosis in twins: a re-analysis of the Prophit survey.
Am Rev Respir Dis 117: 621–624.
8. Vannberg FO, Chapman SJ, Hill AV (2011) Human genetic susceptibility to
intracellular pathogens. Immunol Rev 240: 105–116.
9. Sirugo G, Hennig BJ, Adeyemo AA, Matimba A, Newport MJ, et al. (2008)
Genetic studies of African populations: an overview on disease susceptibility and
response to vaccines and therapeutics. Hum Genet 123: 557–598.
10. Stein CM, Baker AR (2010) Tuberculosis as a complex trait: impact of genetic
epidemiological study design. Mamm Genome 22: 91–99.
11. Saunders BM, Britton WJ (2007) Life and death in the granuloma:
immunopathology of tuberculosis. Immunol Cell Biol 85: 103–111.
12. Hasan Z, Cliff JM, Dockrell HM, Jamil B, Irfan M, et al. (2009) CCL2 responses
to Mycobacterium tuberculosis are associated with disease severity in
tuberculosis. PLoS ONE 4: e8459.
13. Blackwell JM, Black GF, Peacock CS, Miller EN, Sibthorpe D, et al. (1997)
Immunogenetics of leishmanial and mycobacterial infections: the Belem Family
Study. Philos Trans R Soc Lond B Biol Sci 352: 1331–1345.
14. Flores-Villanueva PO, Ruiz-Morales JA, Song CH, Flores LM, Jo EK, et al.
(2005) A functional promoter polymorphism in monocyte chemoattractant
protein-1 is associated with increased susceptibility to pulmonary tuberculosis.
J Exp Med 202: 1649–1658.
15. Jensen LE, Etheredge AJ, Brown KS, Mitchell LE, Whitehead AS (2006)
Maternal genotype for the monocyte chemoattractant protein 1 A(-2518)G
promoter polymorphism is associated with the risk of spina bifida in offspring.
Am J Med Genet A 140: 1114–1118.
16. Modi WS, Goedert JJ, Strathdee S, Buchbinder S, Detels R, et al. (2003) MCP-
1-MCP-3-Eotaxin gene cluster influences HIV-1 transmission. AIDS 17:
2357–2365.
17. Szalai C, Duba J, Prohaszka Z, Kalina A, Szabo T, et al. (2001) Involvement of
polymorphisms in the chemokine system in the susceptibility for coronary artery
disease (CAD). Coincidence of elevated Lp(a) and MCP-1 -2518 G/G genotype
in CAD patients. Atherosclerosis 158: 233–239.
18. Morris GA, Edwards DR, Hill PC, Wejse C, Bisseye C, et al. (2011) Interleukin
12B (IL12B) genetic variation and pulmonary tuberculosis: a study of cohorts
from The Gambia, Guinea-Bissau, United States and Argentina. PLoS ONE 6:
e16656.
19. Hill PC, Brookes RH, Fox A, Fielding K, Jeffries DJ, et al. (2004) Large-scale
evaluation of enzyme-linked immunospot assay and skin test for diagnosis of
Mycobacterium tuberculosis infection against a gradient of exposure in The
Gambia. Clin Infect Dis 38: 966–973.
20. Olesen R, Wejse C, Velez DR, Bisseye C, Sodemann M, et al. (2007) DC-SIGN
(CD209), pentraxin 3 and vitamin D receptor gene variants associate with
pulmonary tuberculosis risk in West Africans. Genes Immun 8: 456–467.
21. Velez DR, Hulme WF, Myers JL, Stryjewski ME, Abbate E, et al. (2009)
Association of SLC11A1 with tuberculosis and interactions with NOS2A and
TLR2 in African-Americans and Caucasians. Int J Tuberc Lung Dis 13:
1068–1076.
22. Gustafson P, Gomes VF, Vieira CS, Rabna P, Seng R, et al. (2004) Tuberculosis
in Bissau: incidence and risk factors in an urban community in sub-Saharan
Africa. Int J Epidemiol 33: 163–172.
23. Sirugo G, Schim van der Loeff M, Sam O, Nyan O, Pinder M, et al. (2004) A
national DNA bank in The Gambia, West Africa, and genomic research in
developing countries. Nat Genet 36: 785–786.
24. Thye T, Nejentsev S, Intemann CD, Browne EN, Chinbuah MA, et al. (2009)
MCP-1 promoter variant 2362C associated with protection from pulmonary
tuberculosis in Ghana, West Africa. Hum Mol Genet 18: 381–388.
25. Feng WX, Mokrousov I, Wang BB, Nelson H, Jiao WW, et al. (2011) Tag SNP
Polymorphism of CCL2 and Its Role in Clinical Tuberculosis in Han Chinese
Pediatric Population. PLoS ONE 6: e14652.
26. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, et al. (2007)
PLINK: a tool set for whole-genome association and population-based linkage
analyses. Am J Hum Genet 81: 559–575.
27. Abecasis GR, Cookson WO (2000) GOLD–graphical overview of linkage
disequilibrium. Bioinformatics 16: 182–183.
28. Barrett JC, Fry B, Maller J, Daly MJ (2005) Haploview: analysis and
visualization of LD and haplotype maps. Bioinformatics 21: 263–265.
29. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, et al. (2002) The
structure of haplotype blocks in the human genome. Science 296: 2225–2229.
30. Hancock DB, Martin ER, Li YJ, Scott WK (2007) Methods for interaction
analyses using family-based case-control data: conditional logistic regression
versus generalized estimating equations. Genet Epidemiol 31: 883–893.
31. Buijtels PC, van de Sande WW, Parkinson S, Petit PL, van der Sande MA, et al.
(2008) Polymorphism in CC-chemokine ligand 2 associated with tuberculosis in
Zambia. Int J Tuberc Lung Dis 12: 1485–1488.
32. Ben-Selma W, Harizi H, Boukadida J (2011) MCP-1 -2518 A/G functional
polymorphism is associated with increased susceptibility to active pulmonary
tuberculosis in Tunisian patients. Mol Biol Rep 38: 5413–5419.
33. Arji N, Busson M, Iraqi G, Bourkadi JE, Benjouad A, et al. (2012) The MCP-1
(CCL2) -2518 GG genotype is associated with protection against pulmonary
tuberculosis in Moroccan patients. J Infect Dev Ctries 6: 73–78.
34. Ganachari M, Ruiz-Morales JA, Gomez de la Torre Pretell JC, Dinh J,
Granados J, et al. (2010) Joint effect of MCP-1 genotype GG and MMP-1
genotype 2G/2G increases the likelihood of developing pulmonary tuberculosis
in BCG-vaccinated individuals. PLoS One 5: e8881.
35. Feng WX, Flores-Villanueva PO, Mokrousov I, Wu XR, Xiao J, et al. (2011)
CCL2-2518 (A/G) polymorphisms and tuberculosis susceptibility: a meta-
analysis. Int J Tuberc Lung Dis Dec 2.
36. Intemann CD, Thye T, Forster B, Owusu-Dabo E, Gyapong J, et al. (2011)
MCP1 haplotypes associated with protection from pulmonary tuberculosis.
BMC Genet 12: 34.
37. Tishkoff SA, Reed FA, Friedlaender FR, Ehret C, Ranciaro A, et al. (2009) The
genetic structure and history of Africans and African Americans. Science 324:
1035–1044.
MCP1 Polymorphisms and Pulmonary Tuberculosis
PLoS ONE | www.plosone.org 12 February 2012 | Volume 7 | Issue 2 | e32275