Review

Genetic Epidemiology of Tuberculosis Susceptibility:

Impact of Study Design
Catherine M. Stein*
Department of Epidemiology and Biostatistics, and Tuberculosis Research Unit, Case Western Reserve University, Cleveland, Ohio, United States of America

non-diseased individuals (‘‘controls’’). In TB, this is complicated,

Abstract: Several candidate gene studies have provided because the pathogenesis of TB can be thought of as a two-stage
evidence for a role of host genetics in susceptibility to process [18]. The first stage consists of latent Mtb infection (LTBI),
tuberculosis (TB). However, the results of these studies in which Mtb establishes a productive infection but does not
have been very inconsistent, even within a study popula- produce symptoms. LTBI is diagnosed by a positive tuberculin skin
tion. Here, we review the design of these studies from a test (TST) and/or positive interferon-c response assay (IGRA) in the
genetic epidemiological perspective, illustrating important
absence of clinical signs and symptoms of full-blown disease [19,20].
differences in phenotype definition in both cases and
controls, consideration of latent M. tuberculosis infection Definitive diagnosis of pulmonary TB requires the recovery of Mtb
versus active TB disease, population genetic factors such as from sputum and cultivation in culture or detection of acid-fast
population substructure and linkage disequilibrium, poly- bacilli (AFB) on smear [19,21]. Studies have shown that AFB smear
morphism selection, and potential global differences in M. is less sensitive than culture, and that AFB smear grade could reflect
tuberculosis strain. These considerable differences between differences in disease severity [21]. Smear-negative, culture-positive
studies should be accounted for when examining the TB is also a problem in developing countries [21]. Thus, the method
current literature. Recommendations are made for future used to diagnose TB could affect the comparability of studies, and
studies to further clarify the host genetics of TB. these differences could reflect variation in disease severity or even
potential misclassification of disease status, generating a significant
impact on the type I and type II error of studies. Here, we will first
Introduction review the various diagnostic criteria used for TB disease, then the
Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is a clinical characterization of study controls, and how these differences
growing public health problem in the era of the HIV/AIDS in study design may affect the interpretation of results across studies.
pandemic. Among the one-third of the world infected by Mtb [1], As stated by Möller and colleagues [22], studies of TB are
almost 8 million new cases of TB occur annually, with 2 million ‘‘exquisitely sensitive to phenotype definition’’. Different criteria
deaths attributed to the disease each year. Only 10% of those have been used to diagnose TB in different study sites. Here, we
individuals infected by Mtb go on to develop clinical disease, and focus on studies of the NRAMP1 (SLC11A1) gene, which has been
disease presentation itself is heterogeneous, suggesting that studied most extensively (Table 1). To summarize, some studies
host factors play a large role in disease susceptibility and natural have used the gold standard definition for TB diagnosis based on
history. An increased understanding of the host response to growth of Mtb in culture [19], though other studies only diagnosed
Mtb will facilitate the development of new vaccines and TB patients based on positive AFB smear. Some studies had
therapeutics [2]. heterogeneous diagnostic criteria, classifying together cases
Several studies have suggested a role for host genetics in TB diagnosed by smear or culture or symptoms. Other studies have
susceptibility. Support for genetic susceptibility to TB in humans combined pulmonary and extrapulmonary TB cases in the analysis
was first provided by twin studies [3,4], animal models [5–8], then [23–27]. Notice that of the 12 studies demonstrating an association
later segregation analyses [9,10]. Countless candidate gene studies between NRAMP1 and TB, only four used culture positivity as
have been conducted, as well as seven genome-wide linkage scans their diagnosis method. Could these differences in diagnostic
[11–17]. However, there is a great deal of inconsistency across criteria disguise differences in disease severity across populations?
these studies. Among studies of any candidate gene, there are
always several reports that provide both positive and negative Citation: Stein CM (2011) Genetic Epidemiology of Tuberculosis Susceptibility:
evidence for an association with TB. Within genome scans, there Impact of Study Design. PLoS Pathog 7(1): e1001189. doi:10.1371/journal.
has been replication of some results across two of the studies ppat.1001189
[14,15], but there is very little replication across the remaining Editor: Marianne Manchester, University of California San Diego, United States of
papers. America
There are a number of key components of the design of these Published January 20, 2011
studies that may explain the inconsistency in the literature. The Copyright: ß 2011 Catherine M. Stein. This is an open-access article distributed
objective of this review is to discuss these issues, illustrated with under the terms of the Creative Commons Attribution License, which permits
examples from the TB genetics literature, and propose some unrestricted use, distribution, and reproduction in any medium, provided the
original author and source are credited.
approaches for taking a more thorough approach to the study of
Funding: Our research is supported in part by the National Institutes of Health
TB genetics.
grants R01HL096811 from NHLBI, Tuberculosis Research Unit (grant N01-AI95383
and HHSN266200700022C/N01-AI70022 from NIAID), and Resource Grant
Impact of Study Design RR03655 from the NCRR. The funders had no role in study design, data collection
and analysis, decision to publish, or preparation of the manuscript.
Phenotype Definition Competing Interests: The author has declared that no competing interests
The first step in any epidemiological study is to define the criteria exist.
used to diagnose disease. Then, one must define what is meant by * E-mail: catherine.stein@case.edu

PLoS Pathogens | www.plospathogens.org 1 January 2011 | Volume 7 | Issue 1 | e1001189

 Table 1. Summary of TB association genetic studies of NRAMP1/SLC11A1, including TB diagnostic criteria, characterization of
controls, and whether there was an association with any SNP in the gene.

Population (Reference) TB Diagnostic Criteria Characterization of Controls Association?

Gambia [80] Smear + Healthy blood donors Yes

Gambia [81] Smear + Healthy blood donors
Malawi [82] Smear + OR culture Unrelated with no history Yes
+ OR histology of infectious disease
Morocco [37] Culture + Healthy family members
Tanzania [83] Culture + Blood donors Yes
Guinea [36] Microscopy (smear +? Culture +?) Unaffected relatives
South Africa [32] Smear + OR culture + Unrelated healthy Yes
Caucasian and African Culture + OR past diagnosis Household members Yes
American [26] in close contact
Caucasian [84] Culture + OR response Clinic patients without Yes
to TB treatment infectious disease
Caucasian, African Culture + Tuberculin skin
American, and Asian [27] test positive
Cambodia [85] Smear + Hospital/clinic patients Yes
China [86] Smear + OR culture + OR symptoms Unrelated healthy males Yes
and radiological evidence; males only
Japan [87] Smear + OR culture + No history of TB disease Yes
Japan [88] Smear + Random clinic patients Yes
Taiwanese [89] Culture + Clinic patients without
pulmonary disease
Japan [90] Smear + Healthy blood donors
without history of pulmonary
or inflammatory disease
Thai [91] Culture + Healthy blood bank donors
China [92] Culture + Hospital patients and Yes
healthy blood donors
Korea [93] Culture + (unclear) No history of TB disease Yes
Japan [94] Smear + OR culture + Unrelated healthy
Poland [38] Culture + TST negative

Smear + refers to AFB smear positive. ‘‘Culture +’’ could include more stringent definitions such as culture positive, smear positive, and radiological evidence consistent
with TB.
This table is limited to studies published in English so that case and control definitions could be determined. It is also limited to studies of pulmonary TB in all age
groups.
doi:10.1371/journal.ppat.1001189.t001

Related to this is the definition of controls. It is unknown in but never develop LTBI [15,19]. Characterization of controls in
many of these studies whether or not the ‘‘controls’’ were latently TB genetics studies has differed widely (examples in Table 1).
infected with Mtb, as evidenced by either a TST or IGRA. Recent Many studies have utilized population controls, similar to the
studies have suggested some genes may actually be related to LTBI approach taken in recent large genome-wide association studies
and not progression to TB [15,28,29], while other studies have (GWAS) [33], i.e., by using blood bank donors. The disadvantage
suggested some genes may differentiate between LTBI and active of this design is possible misclassification bias [34]—the chance
TB disease [30,31]. This is important in truly understanding the that some of these ‘‘controls’’ may never become affected for TB,
role of these genes in disease pathogenesis and progression. If which is problematic when the disease is common [35]. By
controls are latently infected, and there is an association seen contrast, other studies have utilized unaffected household
between a gene and TB, that suggests the gene influences members [26,31,36,37] or have conducted thorough clinical
progression from LTBI to TB. However, if controls are uninfected, evaluation with TST in those without disease [27,30,38]; in these
it is unclear whether an association implies susceptibility for situations, exposure in unaffected individuals is known, so these
developing active disease or just acquisition of LTBI. are true controls in the epidemiological sense. Note that only one
Finally, the selection of controls is not trivial. In a case-control of the NRAMP1 associations was observed in studies where
study, controls should be similar to cases in every way possible exposure has been quantified (Table 1).
except for the presence of disease. In studies of TB, this means
controls should be exposed to infectious TB cases, so that they Epidemiological Study Design
have the opportunity to acquire infection and then progress to The vast majority of genetic epidemiological studies, not just for TB
active TB disease. Some studies conducted in TB-endemic settings but for other complex traits as well, tend to be case-control studies.
assume all individuals are exposed to TB [25,32]. However, Such studies are easier to conduct because they do not require
studies have shown individuals may be persistently exposed to Mtb cooperation of the entire family, and a greater number of cases can be

PLoS Pathogens | www.plospathogens.org 2 January 2011 | Volume 7 | Issue 1 | e1001189

recruited. One major advantage of family-based designs for the study Population Differences—More Than Just Geography
of infectious diseases is the characterization of exposure in the A typical explanation for differing results by population is
‘‘controls’’, as discussed above. Individuals living in the same household population differentiation [22,50], including genetic heterogeneity
have a high likelihood of exposure to an infectious TB case, thereby or inestimable polygenic effects. Another important genetic
influencing the probability that they too will develop TB [39–41]. As difference between populations is in linkage disequilibrium (LD).
described above, epidemiological characterization of exposure is Early studies of TB genetics were restricted to well-character-
important in order to construct a valid case-control study. ized markers within genes (studies of SLC11A1/NRAMP1 in
Another advantage of family-based studies is the ability to Table 1 are examples). Often these markers were exonic or
account for population substructure. Hidden population stratifi- restriction fragment length polymorphisms. The underlying
cation may result in bias (false positive results) [42] or false assumption of the power and design of such studies is that the
negative results [43]. Studies of TB genetics have been conducted polymorphism being analyzed is the causal polymorphism.
in many admixed populations, including African Americans There are millions of single nucleotide polymorphisms (SNPs)
[26,27,44–46], Mexicans [30], and South African ‘‘Coloureds’’ throughout the genome [51,52]. Because of the LD structure in the
[11,14,32,47,48]. Some of these studies [11,26,45,46] have genome, certain SNPs can be used to ‘‘tag’’ haplotypes, such that one
employed family-based designs. Other studies have examined or a few SNPs capture information about LD structure [53]. Many
potential population substructure by analyzing genomic control trait-associated SNPs (.40%) are intergenic or intronic, suggesting an
markers: one study in South Africa utilized ,25 markers [47,48], important role for non-coding SNPs in complex disease [54]. This
and another study utilized .200 markers [49]. Marchini et al. [43] serves as a reminder that disease risk alleles may actually be in LD with
point out genomic control markers will not adequately correct for genotyped markers, which serve as ‘‘tags’’ for haplotypes on which the
population substructure if too few markers are used, but it is causal allele may reside. This is illustrated by Figure 1, where we
difficult to enumerate a sufficient number of markers in consider an underlying disease allele that is not directly genotyped but
populations of African descent. It is unclear if other studies were surrounded by flanking markers. The ability to detect association with
able to account for population substructure. It may be impossible the region where the disease allele resides depends entirely on the
for existing study cohorts to incorporate family-based designs or strength of LD between the unobserved risk allele and flanking
retrospectively evaluate population stratification, but this clearly markers. As patterns of LD differ between study populations, the
may explain some of the heterogeneity among studies. specific trait-associated SNPs will consequently differ.

Figure 1. Impact of variation in linkage disequilibrium (LD) in detection of disease risk alleles. For all three scenarios, D is the underlying
disease risk allele. (A) There is strong LD between D and marker #1 (M1), and weak LD between D and M2. In this situation, association will be
detected with M1, depending on study power based on sample size, strength of genetic effect, and minor allele frequencies. (B) There is no LD
between M1 and D but strong LD between M2 and D. Here, association will be detected only with M2 (again, depending on power). (C) There is weak
LD throughout the region. Association will likely not be detected.
doi:10.1371/journal.ppat.1001189.g001

PLoS Pathogens | www.plospathogens.org 3 January 2011 | Volume 7 | Issue 1 | e1001189

 The impact of LD differences between study populations is in LD in the reference population, and existence of still-unknown
further illustrated in Figure 2. Here, LD patterns in NRAMP1 were risk alleles all complicate replication across studies.
plotted using HapMap reference populations representative of Another controversial issue is the study of common versus rare
those populations where NRAMP1 has been studied: Caucasians genetic variants. The common disease–common variant (CDCV)
in Utah, United States (CEU), Yoruba in Nigeria (YRI), Maasai in hypothesis posits that genetic risk for common diseases will often
Kenya (MKK), Han Chinese (CHN), and African Americans in be due to common risk alleles [59]. This is in contrast to the
the US Southwest (ASW). These LD plots were generated using common disease rare variant (CDRV) hypothesis, which states
default parameters in the Genome Variation Server (http://gvs.gs. that a significant proportion of common chronic diseases are
washington.edu/GVS/), with no minor allele frequency cutoff. influenced by the summation of effects of multiple low frequency
African populations (YRI and MKK) have very little LD because variants in the same gene, where tagging SNPs will not be useful in
they are older populations, and their LD patterns differ. Newer identifying a single haplotype because no single haplotype exists
populations (CEU and CHN) have much greater LD, and recently [60]. Most candidate gene studies assume the CDCV hypothesis.
admixed populations (ASW) also exhibit LD between SNPs, but Recent sequencing studies [61,62] have detected rare SNPs in the
there are differences. Also note that the SNPs themselves (rs TLR family of genes; these could be important, but massive studies
numbers) differ between populations, illustrating how different will be needed in order to detect disease associations at a
polymorphisms exist within the same genes across world statistically significant threshold. In addition, copy number
populations. A perfect illustration of this phenomenon is provided variants (CNVs) have recently attracted attention in their
by Velez et al. [26], who analyzed a number of SNPs within association with complex traits, such as HIV acquisition and
NRAMP1. Though they did not observe statistical association with progression and autoimmune diseases [63]. These are also
the markers that were examined in early studies, they did observe considered rare variants, so we are again faced with all of the
association with intronic and exonic SNPs. If they had not challenges of testing the CDRV hypothesis.
conducted such extensive genotyping, they may have missed these The above discussion focuses on population genetics of humans.
associations. Another related issue is variation in Mtb strains. Researchers have
Only a few other studies have accounted for global variations in categorized Mtb into six main bacterial strain lineages that are
LD by analyzing several SNPs within candidate genes of interest. associated with particular geographical regions [64], as well as
Some studies have selected tag SNPs based on relevant HapMap differences in clinical presentation [65] and rate of progression to
reference populations [26,28,31,46]. Other studies have sequenced active TB disease [66]. So, not only do different diagnostic criteria,
genes of interest first to identify novel SNPs within the gene(s), then as discussed above, potentially reflect differences in disease
analyzed association with those SNPs [55–58]. Though other severity, but specific Mtb strains may also influence disease
studies did not utilize LD in their selection of SNPs, they later severity. A recent study suggests a host genotype x Mtb genotype
estimated LD between markers in their dataset, and used this interaction, whereby the TLR2 genotype is associated with TB
analysis to guide haplotype analysis [48,49]. Since LD patterns caused by the Beijing strain [67]. Very few studies have the
differ by population, it should not be surprising that genetic capacity to examine this potential host by Mtb interaction, but it
association results differ, especially given the limited number of could easily be a potential explanation for differences between
markers analyzed per gene. There are many implications of this studies.
variation. Differences in the strength of LD between the actual
disease locus and genotyped markers will affect the power to detect Complex Genetic Effects
association to markers (Figure 1). In populations with weaker LD Complex traits such as TB are likely influenced by several
such as African populations, denser SNP observed maps are factors, including gene–gene interaction and gene–environment
necessary to detect association effects with untyped disease loci. interaction. Few studies have investigated gene-gene interactions
Thus, variation in number of polymorphisms analyzed, differences in the context of human TB. Many gene products (e.g., Toll-like

Figure 2. Linkage disequilibrium (LD) of the NRAMP1 gene for HapMap reference populations. Yoruba (YRI), Maasai (MKK), Han Chinese
(CHN), Utah Caucasians (CEU), and African Americans (ASW) are shown. The strength of LD is illustrated using the color scale shown in the figure key.
doi:10.1371/journal.ppat.1001189.g002

PLoS Pathogens | www.plospathogens.org 4 January 2011 | Volume 7 | Issue 1 | e1001189

receptors [TLRs]) are known to interact biologically [68], and memory response via positive TST and/or IGRA does not
interaction effects have been demonstrated in mouse models of TB necessarily imply the presence of viable Mtb bacilli. In the US as
[69]. A recent study identified interactions between the NOS2A well as other public health systems, individuals with positive TST
gene and IFNGR1 and TLR4 [45]. Interestingly, both IFNGR1 and are treated as though there are viable organisms present, adding
TLR4 showed no evidence of significant main effects in this further confusion to this phenotype. According to Parrish et al.,
analysis. Another study by the same research group found there is a 2%–23% lifetime probability of developing TB after
interaction between NRAMP1 and TLR2, but TLR2 did not itself acquisition of Mtb infection (LTBI) [73]. This illustrates the
have a significant main effect [26]. This suggests many important heterogeneity in this clinical group, since the risk of progression to
genes may influence TB in combination with other genes, but this active TB may depend on a variety of known and unknown risk
could be overlooked because their individual effects did not meet factors. Furthermore, prophylaxis of LTBI with isoniazid (INH) is
criteria for statistical significance. Motsinger-Reif et al. used the standard of care in many research settings, so that many
multifactor dimensionality reduction to identify a potential gene– individuals with ‘‘LTBI’’ based on positive TST/IGRA, geneti-
gene interaction between TLR4 and the TNF-a gene (TNF) [70]. cally predisposed to develop TB, may not. One way to investigate
In addition, it is well known that HIV influences the pathogenesis the role of host genetics in LTBI would be to compare TST (or
of TB, but most genetic epidemiological studies have been IGRA) positive individuals that develop incident TB to those that
restricted to HIV seronegative individuals. Our work [31] showed do not. Ideally, such a study would not include individuals on INH
an interaction between HIV and the TNF receptor 1 gene. prophylaxis, though that is unethical in many settings. For these
Because many studies have excluded HIV-positive individuals, this reasons, some may argue that it is more relevant to study TB
hypothesis remains relatively unexplored. Similar to the TNF-a genetics, and not LTBI, from a public health standpoint.
pathway, the type I and II interferon pathways have been Thus, it is essential to take a multidisciplinary approach [74] to
associated with both TB and HIV pathogenesis [71], and so develop an all-encompassing picture of the natural history of Mtb
should also be considered for future studies of gene–HIV infection and disease. Few studies have examined the genetics of
interactions. The challenge of examining interaction effects is TB immunology [15,31,75–77]. Gene expression studies using
the requirement of even larger sample sizes, as discussed by Velez microarrays may also shed light on host responses to Mtb [78].
et al. [45]. Proteomic studies will further elucidate host factors involved in
pathogenesis. These various approaches should be analyzed
Conclusions, Recommendations, and Future together to hopefully identify more meaningful clinical groups.
Directions For example, genomic, proteomic, and immunologic data,
collectively, may better capture the heterogeneity in latently
As reviewed recently by Möller et al. [22], the body of work infected individuals.
showing statistical associations between candidate genes and TB Additional complicating factors in comparing geographically
continues to grow. This does not include potential unpublished diverse studies are potential population substructure and LD
studies that failed to find significant associations and are not differences among populations. We recommend that future studies
readily available due to publication bias [22]. Even in the analyze enough SNPs to capture LD in their study population.
published body of literature, however, there is a great deal of Analyses of a few markers within a gene no longer advance the
inconsistency between marker-trait associations, so we are far from field, particularly in light of LD differences between populations.
reaching a consensus regarding genes involved in TB risk. Even with advances in genotyping, many studies of ‘‘old’’ markers
This review focused on methodological reasons for inconsisten- continue to be published. The choice of a reference population for
cy across studies. One important factor is the diagnostic criteria for tag SNP selection is not trivial [62]; thus, dense SNP mapping may
TB disease, which have differed dramatically across studies. be necessary, particularly in studies of African populations. If it is
Resources available for TB diagnosis differ by country, which is impossible to rigorously examine genes in this way, publishing the
confounded when there has been conflict [72]. Differences in LD patterns in the study data [28,45,48,49] is a good start.
diagnostic criteria across studies can reflect differences in TB Furthermore, studies in admixed populations should attempt to
severity and may lead to misclassification of cases as controls; this examine population substructure to minimize this source of bias.
would have a significant impact on the type I and type II error of Populations also differ in the Mtb strain lineage that caused TB;
studies. It is impossible to standardize the diagnostic definitions future studies examining host gene by Mtb gene interaction are
used across all study sites, but researchers should be mindful of warranted. Finally, as in all genetic epidemiological studies of
such differences when interpreting their findings. We strongly complex traits, genes may act in complex ways. Genes may
recommend that researchers characterize the level of exposure to interact with other genes and/or epidemiological factors; these
Mtb in individuals without disease, which should include TST/ potential relationships should not be overlooked. Furthermore, too
IGRA and careful epidemiological characterization. New studies many researchers (authors and journal reviewers alike) focus too
could utilize the household contact design, which facilitates the much on p-values. All p-values must be reported, even if greater
characterization of all stages of Mtb exposure, infection, and than 0.05. Markers with p-values greater than 0.05 may still be
disease [41]. When the household contact study design is not important in their interaction with other markers or environmen-
feasible, spousal controls are also ideal because of persistent and tal factors. Researchers should collect sufficient data to explore
prolonged exposure. these meaningful biological effects.
Recall that TB follows two stages of pathogenesis, and LTBI There are GWAS of TB forthcoming. Given the issues discussed
precedes TB disease. Recent studies suggest that LTBI may have in this review, we must interpret the findings of those GWAS
unique genetic influences [15,28,29]. Persons with LTBI constitute cautiously. Will these studies be underpowered due to the
a major impediment to TB control efforts [73]. Since many heterogeneity among TB cases and controls? A recent summary
ongoing vaccine development efforts will focus either preventing analysis of published GWAS found the reported SNP–trait
LTBI or progression to TB, it is important to understand host associations attaining significance (p,1025) had a median odds
factors that influence containment of Mtb infection. However, the ratio of 1.33, with an interquartile range of 1.20–1.61 [54]; thus,
study of the genetics of LTBI is also not trivial. Indication of T cell the effect sizes of SNPs identified through GWAS are relatively

PLoS Pathogens | www.plospathogens.org 5 January 2011 | Volume 7 | Issue 1 | e1001189

small. Furthermore, the proportion of heritability explained by data should be collected in individuals without TB to better
these variants ranges between 1% and 50% [79]. TB GWAS may understand LTBI and risk of progression to TB, and population
provide new clues into the host biology of TB pathogenesis, but genetic factors should be carefully characterized and considered in
the overall clinical relevance of these SNPs will be limited. In the analysis.
addition, GWAS of other complex traits have often merged data
across ongoing research studies. Because of the dramatic Accession Numbers for Genes Mentioned in This Paper
heterogeneity among studies described in this review, meta- (GeneIDs from EntrezGene)
analyses of TB genetic association studies should be conducted
TLR2 (7097); SLC11A1, aka NRAMP1 (6556); IFNGR1 (3459);
with care.
TLR4 (7099); TNF (7124); TNFSF1A, aka TNF receptor 1 (7132);
In sum, we have barely scratched the surface in understanding
NOS2A (4843).
the genetic determinants of TB pathogenesis. Because of the
significant public health impact of TB, additional studies are
necessary, and should be multidisciplinary in nature. Future Acknowledgments
studies should carefully consider phenotype definition and genetic Thank you to Drs. Thomas Hawn and W. Henry Boom for helpful
epidemiological principles when designing, analyzing, and inter- discussion. Thank you also to two anonymous reviewers, whose suggestions
preting findings. Ideally, culture confirmation for pulmonary TB and challenges greatly strengthened this manuscript.
should be conducted where feasible, thorough epidemiological

References
1. Raviglione M, Snider D, Kochi A (1995) Global epidemiology of tuberculosis: Manosuthi W, Kantipong P, Kanitwittaya S, Sura T, Khusmith S, Tokunaga K,
morbidity and mortality of a worldwide epidemic. JAMA 273: 220–226. Sawanpanyalert P, Keicho N (2009) Genome-wide SNP-based linkage analysis
2. Kaufmann SH, Baumann S, Nasser EA (2006) Exploiting immunology and of tuberculosis in Thais. Genes Immun 10: 77–83.
molecular genetics for rational vaccine design against tuberculosis. Int J Tuberc 18. Comstock G (1982) Epidemiology of tuberculosis. Am Rev Resp Dis 125: 8–15.
Lung Dis 10: 1068–1079. 19. [No authors listed] (2000) Diagnostic Standards and Classification of
3. Comstock G (1978) Tuberculosis in twins: a re-analysis of the Prophit Study. Am Tuberculosis in Adults and Children. This official statement of the American
Rev Resp Dis 117: 621–624. Thoracic Society and the Centers for Disease Control and Prevention was
4. Kallmann F, Reisner D (1943) Twin studies on the significance of genetic factors adopted by the ATS Board of Directors, July 1999. This statement was endorsed
in tuberculosis. Am Rev Tuberculosis 47: 549–574. by the Council of the Infectious Disease Society of America, September 1999.
5. Blackwell J, Barton C, White J, Searle S, Baker A, Williams H, Shaw M (2004) Am J Respir Crit Care Med 161: 1376–1395.
Genomic organizaton and sequence of the human NRAMP gene: identification 20. Nyendak M, Lewinsohn DA, Lewinsohn D (2008) The use of interferon-gamma
and mapping of a promotor region polymorphism. Mol Med 1: 194–205. release assays in clinical practice. In: Davies P, Barnes PF, Gordon SB, eds.
6. Flynn J, Goldstein M, Chan J, Triebold K, Pfeffer K, Lowenstein C, Schreiber R, Clinical tuberculosis. London: Hodder Arnold.
Mak T, Bloom B (1995) Tumor necrosis factor-alpha is required in the 21. Garay S (2004) Pulmonary tuberculosis. In: Rom WN, Garay SM, eds.
protective immune response against Mycobacterium tuberculosis in mice. Tuberculosis. Philadelphia: Lippincott Williams & Wilkins. pp 345–394.
Immunity 2: 561–572. 22. Moller M, de Wit E, Hoal EG (2010) Past, present and future directions in
7. Kramnik I, Dietrick W, Demant P, Bloom B (2000) Genetic control of resistance human genetic susceptibility to tuberculosis. FEMS Immunol Med Microbiol 58:
to experimental infection with virulent Mycobacterium tuberculosis. Proc Natl 3–26.
Acad Sci 97: 8560–8565. 23. Ma X, Reich R, Wright J, Tooker H, Teeter L, Musser J, Graviss E (2003)
8. Lurie M (1941) Heredity, constitution and tuberculosis: An experimental study. Association between Interleukin-8 gene alleles and human susceptbility to
Am Rev Tuberc 44: 1–125. tuberculosis disease. J Infect Dis 188: 349–355.
9. Shaw MA, Collins A, Peacock CS, Miller EN, Black GF, Sibthorpe D, Lins- 24. Rossouw M, Nel HJ, Cooke GS, van Helden PD, Hoal EG (2003) Association
Lainson Z, Shaw JJ, Ramos F, Silveira F, Blackwell JM (1997) Evidence that between tuberculosis and a polymorphic NFkB binding site in the interferon c
genetic susceptibility to Mycobacterium tuberculosis in a Brazilian population is gene. Lancet 361: 1871–1872.
under oligogenic control: linkage study of the candidate genes NRAMP1 and 25. Taype CA, Castro JC, Accinelli RA, Herrera-Velit P, Shaw MA, Espinoza JR
TNFA. Tuberc Lung Dis 78: 35–45. (2006) Association between SLC11A1 polymorphisms and susceptibility to
10. Stein CM, Nshuti L, Chiunda AB, Boom WH, Elston RC, Mugerwa RD, different clinical forms of tuberculosis in the Peruvian population. Infect Genet
Iyengar SK, Whalen CC (2005) Evidence for a major gene influence on tumor Evol 6: 361–367.
necrosis factor-alpha expression in tuberculosis: path and segregation analysis. 26. Velez DR, Hulme WF, Myers JL, Stryjewski ME, Abbate E, Estevan R,
Hum Hered 60: 109–118. Patillo SG, Gilbert JR, Hamilton CD, Scott WK (2009) Association of SLC11A1
11. Bellamy R, Beyers N, McAdam K, Ruwende C, Gie R, Samaai P, Bester D, with tuberculosis and interactions with NOS2A and TLR2 in African-Americans
Meyer M, Corrah T, Collin M, Camidge D, Wilkinson D, Hoal-van Helden E, and Caucasians. Int J Tuberc Lung Dis 13: 1068–1076.
Whittle H, Amos W, van Helden P, Hill A (2000) Genetic susceptibility to 27. Motsinger-Reif AA, Antas PR, Oki NO, Levy S, Holland SM, Sterling TR
tuberculosis in Africans: A genome-wide scan. Proc Natl Acad Sci 97: (2010) Polymorphisms in IL-1beta, vitamin D receptor Fok1, and Toll-like
8005–8009. receptor 2 are associated with extrapulmonary tuberculosis. BMC Med Genet
12. Jamieson S, Miller E, Black G, Peacock C, Cordell H, Howson J, Shaw M, 11: 37.
Burgner D, Xu W, Lins-Lainson Z, Ramos F, Silveira F, Blackwell J (2004) 28. Thye T, Browne EN, Chinbuah MA, Gyapong J, Osei I, Owusu-Dabo E,
Evidence for a cluster of genes on chromosome 17q11-q21 controlling Brattig NW, Niemann S, Rusch-Gerdes S, Horstmann RD, Meyer CG (2009)
susceptibility to tuberculosis and leprosy in Brazilians. Genes and Immunity 5: IL10 haplotype associated with tuberculin skin test response but not with
46–57. pulmonary TB. PLoS ONE 4: e5420. doi:10.1371/journal.pone.0005420.
13. Miller E, Jamieson S, Joberty C, Fakiola M, Hudson D, Peacock C, Cordell H, 29. Cobat A, Gallant CJ, Simkin L, Black GF, Stanley K, Hughes J, Doherty TM,
Shaw M, Lins-Lainson Z, Ramos F, Silveira F, Blackwell J (2004) Genome-wide Hanekom WA, Eley B, Jais JP, Boland-Auge A, van HP, Casanova JL, Abel L,
scans for leprosy and tuberculosis susceptibility genes in Brazilians. Genes Hoal EG, Schurr E, Alcais A (2009) Two loci control tuberculin skin test
Immun 5: 63–67. reactivity in an area hyperendemic for tuberculosis. J Exp Med 206: 2583–2591.
14. Cooke GS, Campbell SJ, Bennett S, Lienhardt C, McAdam KP, Sow O, 30. Flores-Villanueva PO, Ruiz-Morales JA, Song CH, Flores LM, Jo EK,
Gustafson P, Mwangulu F, van HP, Fine P, Hoal EG, Hill AV (2008) Mapping Montano M, Barnes PF, Selman M, Granados J (2005) A functional promoter
of a Novel Susceptibility Locus Suggests a Role for MC3R and CTSZ in Human polymorphism in monocyte chemoattractant protein-1 is associated with
Tuberculosis. Am J Respir Crit Care Med 178: 203–207. increased susceptibility to pulmonary tuberculosis. J Exp Med 202: 1649–1658.
15. Stein CM, Zalwango S, Malone LL, Won S, Mayanja-Kizza H, Mugerwa RD, 31. Stein CM, Zalwango S, Chiunda AB, Millard C, Leontiev DV, Horvath AL,
Leontiev DV, Thompson CL, Cartier KC, Elston RC, Iyengar SK, Boom WH, Cartier KC, Chervenak K, Boom WH, Elston RC, Mugerwa RD, Whalen CC,
Whalen CC (2008) Genome scan of M. tuberculosis infection and disease in Iyengar SK (2007) Linkage and association analysis of candidate genes for TB
Ugandans. PLoS ONE 3: e4094. doi:10.1371/journal.pone.0004094. and TNFalpha cytokine expression: evidence for association with IFNGR1, IL-
16. Baghdadi JE, Orlova M, Alter A, Ranque B, Chentoufi M, Lazrak F, 10, and TNF receptor 1 genes. Hum Genet 121: 663–673.
Archane MI, Casanova JL, Benslimane A, Schurr E, Abel L (2006) An 32. Hoal EG, Lewis LA, Jamieson SE, Tanzer F, Rossouw M, Victor T,
autosomal dominant major gene confers predisposition to pulmonary tubercu- Hillerman R, Beyers N, Blackwell JM, Van Helden PD (2004) SLC11A1
losis in adults. J Exp Med 203: 1679–1684. (NRAMP1) but not SLC11A2 (NRAMP2) polymorphisms are associated with
17. Mahasirimongkol S, Yanai H, Nishida N, Ridruechai C, Matsushita I, Ohashi J, susceptibility to tuberculosis in a high-incidence community in South Africa.
Summanapan S, Yamada N, Moolphate S, Chuchotaworn C, Chaiprasert A, Int J Tuberc Lung Dis 8: 1464–1471.

PLoS Pathogens | www.plospathogens.org 6 January 2011 | Volume 7 | Issue 1 | e1001189

33. (2007) Genome-wide association study of 14,000 cases of seven common diseases 58. Thuong NT, Hawn TR, Thwaites GE, Chau TT, Lan NT, Quy HT, Hieu NT,
and 3,000 shared controls. Nature 447: 661–678. Aderem A, Hien TT, Farrar JJ, Dunstan SJ (2007) A polymorphism in human
34. Edwards BJ, Haynes C, Levenstien MA, Finch SJ, Gordon D (2005) Power and TLR2 is associated with increased susceptibility to tuberculous meningitis. Genes
sample size calculations in the presence of phenotype errors for case/control Immun 8: 422–428.
genetic association studies. BMC Genet 6: 18. 59. Reich DE, Lander ES (2001) On the allelic spectrum of human disease. Trends
35. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JPA, Genet 17: 502–510.
Hirschhorn JN (2008) Genome-wide association studies for complex traits: 60. Bodmer W, Bonilla C (2008) Common and rare variants in multifactorial
consensus, uncertainty and challenges. Nat Rev Genet 9: 356–369. susceptibility to common diseases. Nat Genet 40: 695–701.
36. Cervino A, Lakiss S, Sow O, Hill A (2000) Allelic association between the 61. Ma X, Liu Y, Gowen BB, Graviss EA, Clark AG, Musser JM (2007) Full-exon
NRAMP1 gene and susceptibility to tuberculosis in Guinea-Conakry. Ann Hum resequencing reveals toll-like receptor variants contribute to human susceptibility
Genet 64: 507–512. to tuberculosis disease. PLoS ONE 2: e1318. doi:10.1371/journal.
37. El Baghdadi J, Remus N, Benslinane A, El Annaz H, Chentoufi M, Abel L, pone.0001318.
Schurr E (2004) Variants of the human NRAMP1 gene and susceptibility to 62. Baker A, Randhawa A, Shey M, de Kock M, Kaplan G, Adams M,
tuberculosis in Morocco. Int J Tuberc Lung Dis 7: 599–602. Hanekom WA, Boom WH, Hawn TR, Stein CM (2009) Comparison of
38. Dubaniewicz A, Jamieson SE, Dubaniewicz-Wybieralska M, Fakiola M, genotype frequencies in Toll-like receptor genes in Ugandans, South Africans,
Nancy ME, Blackwell JM (2005) Association between SLC11A1 (formerly and African HapMap populations. Presented at the American Society of Human
NRAMP1) and the risk of sarcoidosis in Poland. Eur J Hum Genet 13: 829–834. Genetics 59th annual meeting; 20–24 October 2009; Honolulu, Hawaii, United
39. Chapman JS, Dyerly MD (1964) Social and other factors in intrafamilial States of America. 1808/T/Poster Board #357.
transmission of tuberculosis. Am Rev Respir Dis 90: 48–60. 63. McCarroll SA, Altshuler DM (2007) Copy-number variation and association
40. Enarson D, Chiang C-Y, Murray J (2004) Global epidemiology of tuberculosis. studies of human disease. Nat Genet 39: S37–S42.
In: Rom WN, Garay SM, eds. Tuberculosis. Philadelphia: Lippincott Williams 64. Gagneux S, Small PM (2007) Global phylogeography of Mycobacterium
& Wilkins. pp 13–30. tuberculosis and implications for tuberculosis product development. Lancet
41. Guwattude D, Nakakeeto M, Jones-Lopez E, Maganda A, Chiunda A, Infect Dis 7: 328–337.
Mugerwa R, Ellner J, Bukenya G, Whalen C (2003) Tuberculosis in household 65. Thwaites G, Caws M, Chau TT, D’Sa A, Lan NT, et al. (2008) The relationship
contacts of infectious cases in Kampala, Uganda. Am J Epidemiol 158: 887–898. between Mycobacterium tuberculosis genotype and the clinical phenotype of
42. Satten GA, Flanders WD, Yang Q (2001) Accounting for unmeasured pulmonary and meningeal tuberculosis. J Clin Microbiol 46: 1363–1368.
population substructure in case-control studies of genetic association using a 66. de Jong BC, Hill PC, Aiken A, Awine T, Antonio M, Adetifa IM, Jackson-
novel latent-class model. Am J Hum Genet 68: 466–477. Sillah DJ, Fox A, Deriemer K, Gagneux S, Borgdorff MW, McAdam KP,
43. Marchini J, Cardon LR, Phillips MS, Donnelly P (2004) The effects of human Corrah T, Small PM, Adegbola RA (2008) Progression to active tuberculosis,
population structure on large genetic association studies. Nat Genet 36: but not transmission, varies by Mycobacterium tuberculosis lineage in The
512–517. Gambia. J Infect Dis 198: 1037–1043.
44. Ma X, Reich R, González A, Pan X, Fothergill A, Starke J, Teeter L, Musser J, 67. Caws M, Thwaites G, Dunstan S, Hawn TR, Lan NT, Thuong NT,
Graviss E (2003) No evidence for association between the polymorphism in the Stepniewska K, Huyen MN, Bang ND, Loc TH, Gagneux S, Van SD,
39 untranslated region of interleukin-12B and human susceptibility to Kremer K, van der SM, Small P, Anh PT, Chinh NT, Quy HT, Duyen NT,
tuberculosis. J Infect Dis 188: 1116–1118. Tho DQ, Hieu NT, Torok E, Hien TT, Dung NH, Nhu NT, Duy PM, van
45. Velez DR, Hulme WF, Myers JL, Weinberg JB, Levesque MC, Stryjewski ME, Vinh CN, Farrar J (2008) The influence of host and bacterial genotype on the
Abbate E, Estevan R, Patillo SG, Gilbert JR, Hamilton CD, Scott WK (2009) development of disseminated disease with Mycobacterium tuberculosis. PLoS
NOS2A, TLR4, and IFNGR1 interactions influence pulmonary tuberculosis Pathog 4: e1000034. doi:10.1371/journal.ppat.1000034.
susceptibility in African-Americans. Hum Genet 126: 643–653. 68. Berrington WR, Hawn TR (2007) Mycobacterium tuberculosis, macrophages,
46. Velez DR, Wejse C, Stryjewski ME, Abbate E, Hulme WF, Myers JL, Estevan R, and the innate immune response: does common variation matter? Immunol Rev
Patillo SG, Olesen R, Tacconelli A, Sirugo G, Gilbert JR, Hamilton CD, 219: 167–186.
Scott WK (2010) Variants in toll-like receptors 2 and 9 influence susceptibility to
69. Bafica A, Scanga CA, Feng CG, Leifer C, Cheever A, Sher A (2005) TLR9
pulmonary tuberculosis in Caucasians, African-Americans, and West Africans.
regulates Th1 responses and cooperates with TLR2 in mediating optimal
Hum Genet 127: 65–73.
resistance to Mycobacterium tuberculosis. J Exp Med 202: 1715–1724.
47. Barreiro LB, Neyrolles O, Babb CL, Tailleux L, Quach H, McElreavey K,
70. Gomez LM, Anaya JM, Vilchez JR, Cadena J, Hinojosa R, Velez L, Lopez-
Helden PD, Hoal EG, Gicquel B, Quintana-Murci L (2006) Promoter variation
Nevot MA, Martin J (2007) A polymorphism in the inducible nitric oxide
in the DC-SIGN-encoding gene CD209 is associated with tuberculosis. PLoS
synthase gene is associated with tuberculosis. Tuberculosis (Edinb) 87: 288–294.
Med 3: e20. doi:10.1371/journal.pmed.0030020.
71. Giacomini E, Remoli ME, Gafa V, Pardini M, Fattorini L, Coccia EM (2009)
48. Moller M, Nebel A, Valentonyte R, Van Helden PD, Schreiber S, Hoal EG
IFN-beta improves BCG immunogenicity by acting on DC maturation. J Leukoc
(2009) Investigation of chromosome 17 candidate genes in susceptibility to TB in
Biol 85: 462–468.
a South African population. Tuberculosis (Edinb) 89: 189–194.
72. Gustafson P, Gomes VF, Vieira CS, Jensen H, Seng R, Norberg R, Samb B,
49. Davila S, Hibberd ML, Hari DR, Wong HE, Sahiratmadja E, Bonnard C,
Alisjahbana B, Szeszko JS, Balabanova Y, Drobniewski F, van Crevel R, van de Naucler A, Aaby P (2001) Tuberculosis mortality during a civil war in Guinea-
Vosse E, Nejentsev S, Ottenhoff THM, Seielstad M (2008) Genetic association Bissau. JAMA 286: 599–603.
and expression studies indicate a role of toll-like receptor 8 in pulmonary 73. Parrish NM, Dick JD, Bishai WR (1998) Mechanisms of latency in
tuberculosis. PLoS Genet 4: e1000218. doi:10.1371/journal.pgen.1000218. Mycobacterium tuberculosis. Trends Microbiol 6: 107–112.
50. Deng HW (2001) Population admixture may appear to mask, change or reverse 74. Comas I, Gagneux S (2009) The past and future of tuberculosis research. PLoS
genetic effects of genes underlying complex traits. Genetics 159: 1319–1323. Pathog 5: e1000600. doi:10.1371/journal.ppat.1000600.
51. International HapMap Consortium (2003) The International HapMap Project. 75. Wheeler E, Miller EN, Peacock CS, Donaldson IJ, Shaw MA, Jamieson SE,
Nature 426: 789–796. Blackwell JM, Cordell HJ (2006) Genome-wide scan for loci influencing
52. International HapMap Consortium (2005) A haplotype map of the human quantitative immune response traits in the Belem family study: comparison of
genome. Nature 437: 1299–1320. methods and summary of results. Ann Hum Genet 70: 78–97.
53. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, 76. Hawn TR, Misch EA, Dunstan SJ, Thwaites GE, Lan NT, et al. (2007) A
Higgins J, Defelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, common human TLR1 polymorphism regulates the innate immune response to
Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D (2002) The lipopeptides. Eur J Immunol 37: 2280–2289.
structure of haplotype blocks in the human genome. Science 296: 2225–2229. 77. Shey MS, Randhawa AK, Bowmaker M, Smith E, Scriba TJ, de Kock M,
54. Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS, Mahomed H, Hussey G, Hawn TR, Hanekom WA (2010) Single nucleotide
Manolio TA (2009) Potential etiologic and functional implications of genome- polymorphisms in toll-like receptor 6 are associated with altered lipopeptide- and
wide association loci for human diseases and traits. Proc Natl Acad Sci U S A mycobacteria-induced interleukin-6 secretion. Genes Immun 11: 561–572.
106: 9362–9367. 78. Thuong NT, Dunstan SJ, Chau TT, Thorsson V, Simmons CP, Quyen NT,
55. Tang NL, Fan HP, Chang KC, Ching JK, Kong KP, Yew WW, Kam KM, Thwaites GE, Thi Ngoc Lan N, Hibberd M, Teo YY, Seielstad M, Aderem A,
Leung CC, Tam CM, Blackwell J, Chan CY (2009) Genetic association between Farrar JJ, Hawn TR (2008) Identification of tuberculosis susceptibility genes with
a chemokine gene CXCL-10 (IP-10, interferon gamma inducible protein 10) and human macrophage gene expression profiles. PLoS Pathog 4: e1000229.
susceptibility to tuberculosis. Clin Chim Acta 406: 98–102. doi:10.1371/journal.ppat.1000229.
56. Dissanayeke SR, Levin S, Pienaar S, Wood K, Eley B, Beatty D, Henderson H, 79. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ,
Anderson S, Levin M (2009) Polymorphic variation in TIRAP is not associated McCarthy MI, Ramos EM, Cardon LR, Chakravarti A, Cho JH,
with susceptibility to childhood TB but may determine susceptibility to TBM in Guttmacher AE, Kong A, Kruglyak L, Mardis E, Rotimi CN, Slatkin M,
some ethnic groups. PLoS ONE 4: e6698. doi:10.1371/journal.pone.0006698. Valle D, Whittemore AS, Boehnke M, Clark AG, Eichler EE, Gibson G,
57. Hawn TR, Dunstan SJ, Thwaites GE, Simmons CP, Thuong NT, Lan NT, Haines JL, Mackay TF, McCarroll SA, Visscher PM (2009) Finding the missing
Quy HT, Chau TT, Hieu NT, Rodrigues S, Janer M, Zhao LP, Hien TT, heritability of complex diseases. Nature 461: 747–753.
Farrar JJ, Aderem A (2006) A polymorphism in Toll-interleukin 1 receptor 80. Bellamy R, Ruwende C, Corrah T, McAdam K, Whittle H, Hill A (1998)
domain containing adaptor protein is associated with susceptibility to meningeal Variations in the NRAMP1 gene and susceptibility to tuberculosis in West
tuberculosis. J Infect Dis 194: 1127–1134. Africans. N Engl J Med 338: 640–644.

PLoS Pathogens | www.plospathogens.org 7 January 2011 | Volume 7 | Issue 1 | e1001189

81. Awomoyi A, Marchant A, Howson J, McAdam K, Blackwell J, Newport M Shirakawa T, Kishi F (2000) Genetic variants of NRAMP1 and active
(2002) Interleukin-10, Polymorphism in SLC11A1 (formerly NRAMP1), and tuberculosis in Japanese populations. Clin Genet 58: 74–76.
susceptibility to tuberculosis. J Infect Dis 186: 1808–1814. 89. Liaw Y-S, Wu J-JT, Wu C-H, Hung C-C, Lee C-N, Yang P-C, Luh K-T, Kuo S-
82. Fitness J, Floyd S, Warndorff D, Sichali L, Malema S, Crampin A, Fine P, Hill A H (2002) Variations in the NRAMP1 gene and susceptibility of tuberculosis in
(2004) Large-scale candidate gene study of tuberculosis susceptibility in the Taiwanese. Int J Tuberc Lung Dis 6: 454–460.
Karonga district of Nothern Malawi. Am J Trop Med Hyg 71: 341–349. 90. Akahoshi M, Ishihara M, Remus N, Uno K, Miyake K, Hirota T, Nakashima K,
83. Soborg C, Andersen AB, Range N, Malenganisho W, Friis H, Magnussen P, Matsuda A, Kanda M, Enomoto T, Ohno S, Nakashima H, Casanova JL,
Temu MM, Changalucha J, Madsen HO, Garred P (2007) Influence of Hopkin JM, Tamari M, Mao XQ, Shirakawa T (2004) Association between
candidate susceptibility genes on tuberculosis in a high endemic region. Mol IFNA genotype and the risk of sarcoidosis. Hum Genet 114: 503–509.
Immunol 44: 2213–2220. 91. Vejbaesya S, Chierakul N, Luangtrakool P, Sermduangprateep C (2007)
84. Ma X, Wright J, Reich R, Teeter L, El Sahly H, Awe R, Musser J, Graviss E NRAMP1 and TNF-alpha polymorphisms and susceptibility to tuberculosis in
(2002) 59 dinucleotide repeat polymorphism of NRAMP1 and susceptibility to Thais. Respirology 12: 202–206.
tuberculosis among Caucasian patients in Houston, Texas. Int J Tuberc Lung 92. Leung KH, Yip SP, Wong WS, Yiu LS, Chan KK, Lai WM, Chow EY, Lin CK,
Dis 6: 818–823.
Yam WC, Chan KS (2007) Sex- and age-dependent association of SLC11A1
85. Delgado J, Baena A, Thim S, Goldfeld A (2002) Ethnic-specific genetic
polymorphisms with tuberculosis in Chinese: a case control study. BMC Infect
associations with pulmonary tuberculosis. J Infect Dis 186: 1463–1468.
Dis 7: 19.
86. Liu W, Cao W, Zhang C, Tian L, Wu X, Habbema J, Zhao Q, Zhang P, Xin
93. Ryu S, Park Y-K, Bai G-H, Kim S-J, Park S-N, Kang S (2000) 39UTR
Z-T, Li C, Yang H (2004) VDR and NRAMP1 gene polymorphisms in
susceptibility to pulmonary tuberculosis among the Chinese Han population: a polymorphisms in the NRAMP1 gene are associated with susceptibility to
case-control study. Int J Tuberc Lung Dis 8: 428–434. tuberculosis in Koreans. Int J Tuberc Lung Dis 4: 577–580.
87. Abe T, Iinuma Y, Ando M, Yokoyama T, Yamamoto T, Nakashima K, 94. Kusuhara K, Yamamoto K, Okada K, Mizuno Y, Hara T (2007) Association of
Takagi N, Baba H, Hasegawa Y, Shimokata K (2003) NRAMP1 polymor- IL12RB1 polymorphisms with susceptibility to and severity of tuberculosis in
phisms, susceptibility and clinical features of tuberculosis. J Infect 46: 215–220. Japanese: a gene-based association analysis of 21 candidate genes.
88. Gao P-S, Fujishima S, Mao X-Q, Remus N, Kanda M, Enomoto T, Dake Y, Int J Immunogenet 34: 35–44.
Bottini N, Tabuchi M, Hasegawa N, Yamaguchi K, Tiemessen C, Hopkin J,

PLoS Pathogens | www.plospathogens.org 8 January 2011 | Volume 7 | Issue 1 | e1001189