SLC11A1 (NRAMP1) Polymorphisms and Tuberculosis Susceptibility: Updated Systematic Review and Meta-Analysis

Background Natural resistance associated macrophage protein 1 (NRAMP1), encoded by the SLC11A1 gene, has been described to regulate macrophage activation and be associated with infectious and autoimmune diseases. The relation between SLC11A1 polymorphisms and tuberculosis susceptibility has been studied in different populations. Methods We systematically reviewed published studies on SLC11A1 polymorphisms and tuberculosis susceptibility until September 15, 2010 and quantitatively summarized associations of the most widely studied polymorphisms using meta-analysis. Results In total, 36 eligible articles were included in this review. In Meta-analysis, significant associations were observed between tuberculosis risk and widely studied SLC11A1 polymorphisms with summarized odds ratio of 1.35 (95%CI, 1.17–1.54), 1.25 (95% CI, 1.04–1.50), 1.23 (95% CI, 1.04–1.44), 1.31 (95%CI, 1.08–1.59) for 3′ UTR, D543N, INT4, and 5′ (GT)n, respectively. Heterogeneity between studies was not pronounced, and the associations did not remarkably vary in the stratified analysis with respect to study population and study base. Conclusions The association between SLC11A1 polymorphisms and tuberculosis susceptibility observed in our analyses supports the hypothesis that NRAMP1 might play an important role in the host defense to the development of tuberculosis.


Introduction
Host genetic susceptibility to infectious disease has been widely studied in recent years, which is helpful for high-risk population identification and therefore promotes diseases prevention and early diagnosis [1,2]. Moreover, such study also contributes to clarify potential mechanisms underlying host defense to the disease development. Natural resistance associated macrophage protein 1 (NRAMP1), encoded by the SLC11A1 gene, has multiple effects on macrophage activation and has been reported to play an important role in host innate immune response against infections [3].
Tuberculosis (TB), caused by infection of Mycobacterium tuberculosis, remains a major challenge to global public health. As estimated, that one-third of the world's population is infected, but that only a minority of those infected ever develop TB [4]. Host genetic susceptibility, together with some environmental and lifestyle factors, has been suggested to contribute to such clinical diversity [5,6]. In 1998, for the first time, relation between SLC11A1 polymorphisms and TB susceptibility was reported in a population from West Africa [7]. Later, the association of several SLC11A1 loci have been extensively investigated, including 39 UTR (1729+55del4), D543N (Asp543Asn), INT4 (469+14G/C), and 59 promoter (GT)n. However, the results were not consistent between the studies. A meta-analysis, which based on literature review until December 2004, suggested an ethnicity specific effect of SLC11A1 polymorphisms on TB risk [8]. In the past five years, the number of original studies addressing this topic has doubled. Therefore, it is necessary to update the meta-analysis which might provide more solid evidence and minimize potential bias caused by limited publications in the past.
In this article, we performed a systematic review and metaanalysis, based on literature identification until 15 September 2010, to summarize associations between the most widely studied SLC11A1 polymorphisms and TB susceptibility.

Results
A total of 336 articles were achieved by literature search, from the PubMed, EMBASE and CBM databases, using different combination of key terms. As shown in Figure 1, after excluding those overlapped between the databases, 217 abstracts were retrieved for detailed evaluation. Forty seven studies addressing the association of SLC11A1 polymorphisms and TB were identified, and full-text article retrieve excluded 11 of them (please refer to Table S1 for more detailed information). Finally, 38 studies from 36 articles, 29 in English [7,and 7 in Chinese [37][38][39][40][41][42][43], were included in this review and meta-analysis.
As shown in Table S2 and Table S3, among the included articles, 25 were conduced in Asians and 11 in non-Asians (3 from Africa, 4 from Europeans and 2 from Americans). Thirteen studies and 3 studies specifically addressed pulmonary TB and extrapulmonary TB, respectively. Seven studies were population based. HIV status of the studied population was considered in 24 studies. There were 13 studies matched controls with cases for major covariates (age, sex, ethnicity, and area of residence). For 39 UTR, D543N, INT4 and 59 (GT)n, meta-analyses were conduced within 30, 29, 20 and 12 studies, respectively. Figure 2 shows the associations between SLC11A1 39 UTR polymorphism and TB. Meta-analysis suggested that TGTGcarriage (TGTG2/2 and TGTG2/+) might be a risk factor for TB with a summarized OR of 1.35 (95%CI, 1.17-1.54) as compared to TGTG+/+ genotype. Medium heterogeneity between studies (p,0.01; I 2 = 48.41%) was observed. No evident publication bias was found (p = 0.99 for Begg rank correlation analysis; p = 0.59 for Egger weighted regression analysis). In the stratified analysis, the strength of the association was most evident for Asians, and marginal results were observed for Africans and Westerns. No pronounced difference was found according to different study base.
As shown in Figure 3, good homogeneity was observed between studies addressing D543N polymorphism (p,0.01; I 2 = 61.26%) and A allele carriage was significantly associated with TB (OR, 1.24; 95% CI, 1.04-1.49). No substantial publication bias was observed (p = 0.39 for Begg rank correlation analysis; p = 0.49 for Egger weighted regression analysis). In the subgroup analyses, significant relations were observed for Africans and Westerns.
As shown in Figure 4, a significant association was found for INT4 C allele carriage (CC+CG) with increased risk of TB as compared to GG genotype (OR, 1.23; 95% CI, 1.04-1.44). Medium heterogeneity between studies (p = 0.08; I 2 = 32.47%) was observed. No significant publication bias was observed (p = 0.12 for Begg rank correlation analysis; p = 0.35 for Egger weighted regression analysis). The association was not significant any more in Westerns in the subgroup analysis.
Meta-analysis of the association between TB and SLC11A1 59 promoter (GT)n polymorphism was shown in Figure 5. As compared to the most commonly distributed allele 3, carriage of other alleles was significantly related to an increased risk of TB with a summarized OR of 1.31 (95%CI, 1.08-1.59). No substantial heterogeneity was observed between studies (p,0.01; I 2 = 61.78%). No publication bias was found (p = 0.49 for Begg rank correlation analysis; p = 0.51 for Egger weighted regression analysis). In the stratified analyses, similar results were observed between subgroups according to study base, and significant association was found only for Africans.
After excluding studies specifically addressing extra-pulmonary TB, the associations between SLC11A1 polymorphisms and TB were not substantially changed (see Table S4). Similar associations were observed as well when stratified the analysis on pulmonary TB (see Table S4), which suggests the effect of SLC11A1 polymorphisms might not be influenced by disease types.

Discussion
This review addressed the associations between SLC11A1 polymorphisms and TB susceptibility reported until September 2010. Thirty six articles addressing the most widely studied SLC11A1 polymorphisms (39 UTR, D543N, INT4 and 59 (GT)n) were identified, and their effects were summarized by means of metaanalysis. Significant associations with TB susceptibility were observed for all these four loci. Strength of the associations in the subgroup analyses with respect to study population was not consistent. However, significant association has been observed for all populations (Africans, Asians and Westerns) with at least one of the four loci. Limited number of studies in the stratified analysis and heterogeneity between studies might explain, at least in part, the existing inconsistency.
TB is a serious public health problem in worldwide. Early diagnosis, drug resistance, vaccine and HIV co-infection are major factors influencing efficiency of TB prevention and control [44]. In the past decades, the association between host genetic polymorphisms and TB susceptibility has been widely studied as well. Most of the target loci were localized in genes participating immune response [45]. The NRAMP1 protein is an integral membrane protein expressed exclusively in the lysosomal compartment of monocytes and macrophages [3]. It has been reported that NRAMP1 regulates macrophage activation and is associated with TB susceptibility. Underlying mechanisms have been proposed based on its function as metal transporter [46]. After phagocytosis, NRAMP1 is targeted to the membrane of the microbe-containing phagosome, where it mediates transport of iron and other cations. Iron is essential for biological functions, both for host immune defense and mycobacterial growth. Therefore, NRAMP1 has also been suggested to play a role in determining host susceptibility to other intracellular pathogens and autoimmune diseases [47]. Positive association between SLC11A1 polymorphisms and these infections and diseases has provided strong evidence for this hypothesis.
In 2006, Li HT and colleagues reported a meta-analysis on SLC11A1 polymorphisms and TB which based on 14 eligible studies published until December 2004 [8]. In this article, less common alleles at the four most widely studied loci were generally shown to confer an increased risk of TB on their carriers from African and Asian, whereas they were not statistically associated with TB in those of European origin. Authors suggested this ethnic-specific observation may explain in part by Africans and Asians have greater susceptibility to TB than those of Europeans. Meanwhile, authors also mentioned that this hypothesis needed to be proved in future research due to the limited number of included publications and subjects. Our present updated meta-analysis, based on 36 eligible studies until September 2010, did not present an ethic-specific effect of SLC11A1 polymorphisms. Three ethnic groups were defined with respect to the distribution of study population in the included studies, e.g. Asians, Africans and Westerns (Europeans and Americans). Statistically significant or marginal associations with TB susceptibility were found for the four studied loci in all of the three ethnic groups. Large number of included articles in the present study makes the evidence stronger to propose a consistent effect of SLC11A1 polymorphisms in different populations. In addition, population-based study has shown to be more powerful for such genetic association analysis. However, only minority of included studies (7 of 36) were conducted as populationbased and no substantial different or more pronounced effect was observed as compared to those hospital-based studies.
The stratified analyses suggested different study population and study base might partly explain the moderate heterogeneity between studies observed in our analyses. Further sensitivity analyses were performed as well to explore potential origin of the heterogeneity (see Table S4). When restrict to analyses on pulmonary TB or excluding those studies specifically on extrapulmonary TB, better homogeneity between studies was observed. Nevertheless, the effects of the polymorphisms were not substantially influenced by such stratified analyses, which suggested that the role of NRAMP1 in the development of TB might not be disease type specific. Further studies addressing the effect of NRAMP1 on pulmonary TB and extra-pulmonary TB, respectively, are warranted to verify this hypothesis.
There are some limitations to this systematic review that should be kept in mind. First, the potential confounding effect of age, sex and ethnicity was not controlled for in more than a half of included studies, which makes the interpretation of the results and stratified analyses difficult. Second, because not all necessary information could be obtained from all included studies, more detailed sub-grouping analysis (e.g. by HIV status or type of TB) could not be performed. Third, the crude division of ethnics groups into 'Asian', 'African', and 'Western' makes the analyses be prone to bias. Further studies from different populations are warranted to verify current findings. Fourth, included studies were restricted to those published in English or Chinese in our study which might introduce potential bias into data analysis as well. Fifth, only 17 included studies mentioned whether their study population was in HWE for the investigated variants [11,14,16,17,[23][24][25][26][27][28][29][30][33][34][35]39,43]. Based on the data provided by the articles and own calculations, significant deviations from HWE (p,0.05) in controls were observed for three studies for any of their studied polymorphisms [12,38,42]. Their results should be interpreted with more caution. We therefore repeated the meta-analyses after exclusion of these studies. However, this exclusion did not materially affect the results (see Table S4).
In conclusion, this systematic review summarized the associations between SLC11A1 polymorphisms and TB susceptibility. Our results suggested a consistent association between SLC11A1 polymorphisms and TB in different populations, which supports the hypothesis that NRAMP1 might play an important role in the host defense to the development of TB. However, due to the moderate strength of the associations, their values to be used for risk prediction should be considered cautiously.

Literature search
Studies addressing the association between SLC11A1 polymorphisms and TB were identified by searching for articles in the PubMed, EMBASE and Chinese BioMedical Literature (CBM) Database until 15 September 2010 [5]. Various combinations of the terms ''tuberculosis'', ''NRAMP1'', ''SLC11A1'', ''polymorphism'' and ''susceptibility'' were used to screen for potentially relevant studies. Additional studies were also identified by means of cross-referencing.

Inclusion and exclusion criteria
Case-control or cohort studies presenting original data on the associations between SLC11A1 polymorphisms and TB were included. Only the most widely studied polymorphisms were considered: 39 UTR (1729+55del4) (rs17235416), D543N (As-p543Asn) (rs17235409), INT4 (469+14G/C) (rs3731865), and 59 promoter (GT)n (rs34448891). Exclusion criteria were: 1) review articles; 2) studies in languages other than English or Chinese; 2) studies on other polymorphisms other than the target four polymorphisms; 3) previous articles addressing meta-analysis of the associations between SLC11A1 polymorphisms and TB. If the eligible study was reported in duplication, the article published in English or published earlier was included in this review.

Data extraction
For all studies, we extracted the following data from original publications: first author and year of publication; distribution of genotypes for each polymorphism among cases and controls; characteristics of the study design and the study population (study base, numbers and mean age of cases and controls, TB diagnosis, HIV status, source of controls, matching criteria and host ethnicity).

Statistical analysis
Hardy-Weinberg Equilibrium (HWE) was examined in controls by asymptotic Pearson's chi-square test for each polymorphism in each study. The association between polymorphism and TB was estimated by means of odds ratios (OR) and corresponding 95% confidence intervals (CI) comparing cases to controls. Codominant model was used for 39 UTR, D543N, and INT4. The effect of allele 3 carriage was assessed for 59 (GT)n. Meta-analyses were carried out using Comprehensive Meta-Analysis (V2.0, Biostat, Englewood, NJ, USA). Random effects models were used for meta-analysis, taking into account the possibility of heterogeneity between studies which was tested by the Q test and I 2 test. Stratified analyses were conducted with respect to study base (hospital or population based) and host ethnicity. The latter was categorized into Africans, Asians and Westerns (Europeans and Americans). Because of the limited number of publications in Americans, they were sub-grouped to Westerns combined with studies from Europeans. Begg rank correlation method and Egger weighted regression method were used to statistically assess publication bias (p,0.05 was considered indicative of statistically significant publication bias). Sensitivity analyses were performed after excluding studies specifically on extra-pulmonary TB, or restricting the analysis specifically on pulmonary TB.