Allelic-Specific Regulation of xCT Expression Increases Susceptibility to Tuberculosis by Modulating microRNA-mRNA Interactions

Tuberculosis (TB) is the leading cause of death from a single infectious agent globally, and the development of multidrug resistance represents a serious health concern, particularly in the developing world. Novel effective treatments are urgently required. xCT expression is known to increase susceptibility to TB, and certain polymorphisms in the gene encoding this protein interrupt the binding of microRNA and prevent its suppression. Taking advantage of the FDA approval for the use of sulfasalazine (SASP), which inhibits xCT-mediated cystine transport in humans, we demonstrate how host genotype-specific therapies tailored to the xCT genotype can improve TB outcomes.

and host immunity and the inflammatory response are also closely linked with disease outcomes (5). Accordingly, certain genetic variants that affect the regulation of cellmediated immune and inflammatory responses are also associated with susceptibility and host responses to TB (6)(7)(8)(9)(10)(11). The differences in susceptibility to TB in particular ethnicities lends further credence to the role of genetic polymorphisms. Strong evidence suggests that polymorphisms in genes encoding Toll-like receptor (TLR), gamma interferon (IFN-␥), tumor necrosis factor alpha (TNF-␣), interleukin 12 (IL-12), IL-10, monocyte chemoattractant protein 1 (MCP-1), matrix metalloproteinase 1 (MMP-1), and the vitamin D receptor (VDR) are involved (12)(13)(14)(15)(16). Further functional studies are required to fully characterize the impact of human genetic variation on TB and to identify additional candidate polymorphisms in genes that might be of interest. One such gene is SLC7A11, which encodes the protein xCT.
xCT is a subunit of the x c Ϫ cysteine-glutamate antiporter: it is responsible for the Na ϩ -independent transport of cystine into a cell in exchange for glutamate. This process is essential to cellular redox homeostasis, as it effectively increases the synthesis of glutathione, which serves as an antioxidant (17,18). xCT is chaperoned by the CD98/Slc3a2 subunit, but unlike the constitutively expressed CD98, xCT expression is upregulated by oxidative stress. This is an important regulatory mechanism of the growth of both tumor cells and intracellular pathogens, such as M. tuberculosis. Accordingly, the genetic disruption of xCT and the chemical modulation of xCT with sulfasalazine (SASP), an inhibitor of xCT-mediated cystine transporters, have both been found to significantly decrease bactericidal burden and inflammation-related tissue damage in mice with TB (19).
SASP exerts inhibitory effects against tumor growth, invasion, and metastasis in many types of cancer (20,21). As SASP has already received FDA approval for the treatment of ulcerative colitis, this gives it an advantage over other potential novel treatments for TB. However, the outcome of SASP treatment is still likely to be influenced by the genetics of the host. For example, a previous report demonstrated that a single-nucleotide polymorphism (SNP) in the human LTA4H promoter, rs17525495 TT, is associated with 2.3-fold higher LTA4H protein expression levels than the CC genotype. This SNP has been demonstrated to critically influence the response to anti-inflammatory dexamethasone treatment in TB meningitis (22,23). This indicates that host genotype-specific therapies can optimize treatments for M. tuberculosis, as each patient can be matched to the treatment regimen that they are most likely to benefit from. However, for this strategy to have any hope of working, the impact of genetic polymorphisms on treatment response must be fully characterized.
It is clear that the identification of genetic variants that regulate xCT gene expression during M. tuberculosis infection are necessary if treatment of TB with SASP is to be successfully translated to the clinic. However, the associations between genetic polymorphisms in xCT and TB susceptibility remain largely unknown, and it is unclear how these variants modulate xCT expression or the effects of SASP treatment. To this end, we conducted a case-control study in a Chinese cohort to identify functional SNPs in the xCT gene and their links to TB susceptibility. We also investigated the effects and mechanisms underlying the effects of these functional SNPs on M. tuberculosis-induced xCT expression.

RESULTS
xCT gene polymorphism is associated with susceptibility to tuberculosis. To analyze the relationship between xCT gene polymorphism and susceptibility to tuberculosis, four SNPs were screened with the MassARRAY platform, in a cohort of 914 active TB cases and 936 controls. Of these SNPs, rs4131888 and rs11764488 are located within the known intron region of the xCT gene, while rs7674870 and rs13120371 are located in the 3= untranslated region (UTR). The prevalence of these four SNPs significantly differed between patients with active TB and healthy controls (Table 1). However, there was no linkage disequilibrium between rs13120371 and the other three SNPs (see Fig. S1 in the supplemental material). For this reason, we focused on rs13120371 at the genotype level. Individuals with rs13120371 AA genotypes exhibited increased TB susceptibility compared with those with TT genotypes, as determined using an additive model (odds ratio [OR], 1.38; 95% confidence interval [CI], 1.01 to 1.89; P ϭ 0.044) ( Table 2), indicating that the A allele is associated with an increased risk of M. tuberculosis infection. Notably, pulmonary tuberculosis is more common in males than in females. While the frequency of allele A among males in the active TB group was significantly increased compared with that in the control group, the same was not true for female subjects (Table 3).
SNP rs13120371 A is associated with increased disease severity. Our previous report suggested that the rs13120371 polymorphism might affect the inflammatory response to M. tuberculosis. Thus, we investigated IFN-␥ production by peripheral blood mononuclear cells (PBMCs) in patients with different rs13120371 genotypes from our cohort. Patients carrying the rs13120371 AA genotype had significantly higher numbers of M. tuberculosis antigen-specific IFN-␥ spot-forming cells (SFCs) than those carrying the GG genotype ( Fig. 1A and B). There was also a significant difference in erythrocyte sedimentation rate (ESR) levels among pulmonary TB patients with different rs13120371 genotypes, with those with the AA genotype having increased levels compared to those with the GG genotype (Fig. 1C). Then, using high-resolution computed tomography (HRCT), lung damage in these patients was quantified using a score based on radiographic manifestations, including the presence  HRCT scores were significantly higher in patients carrying the rs13120371 AA genotype than in those with the rs13120371 GG genotype (Fig. 1D). This difference remained apparent 2 years after the completion of anti-TB treatment, with patients carrying the rs13120371 AA genotype still displaying significantly higher HRCT scores than those with the rs13120371 GG genotype (Fig. 1E). There were no significant differences in the prevalences of the other three SNPs (Fig. S2 to S4), however, suggesting that the rs13120371 polymorphism specifically is associated with long-term lung damage and disease outcome in patients with TB. SNP rs13120371 influences xCT gene expression. SNP rs13120371 is located in the 3= UTR of the xCT gene, which suggests that it may influence gene expression in response to M. tuberculosis infection. To test this hypothesis, we isolated PBMCs from healthy controls with different rs13120371 genotypes and assessed xCT mRNA expression levels in these cells following stimulation with heat-killed M. tuberculosis. Cells  isolated from individuals carrying the rs13120371 AA genotype expressed significantly higher levels of xCT mRNA than those with the GG genotype ( Fig. 2A). Consistent with these results, xCT protein expression levels were significantly higher in PBMCs isolated from TB patients with the rs13120371 AA genotype than in those carrying the GG genotypes (Fig. 2B). However, there was no significant difference in CXCL1, CXCL2, or IL1B mRNA expression levels between genotypes (Fig. 2C to E). Expression of these genes was previously found to be significantly increased by xCT gene deletion (19). xCT gene expression is suppressed by miR-142-3p. TargetScan predicted that the xCT gene is the target for miR-142-3p. To confirm whether miR-142-3p regulates the expression of xCT, we transfected miR-143-3p mimics, miR-143-3p inhibitors, or control oligonucleotides into 293T cells and investigated the impact on xCT expression. miR-143-3p overexpression significantly reduced the expression of xCT compared to that in the control, while inhibition of miR-143-3p significantly increased xCT mRNA expression ( Fig. 3A and B). Thus, the xCT gene was confirmed as the target gene of miR-143-3p. To determine whether miR-143-3p was also involved in immune regulation in TB, we infected differentiated Thp-1 macrophages with the H37Ra strain of M. tuberculosis in vitro. Infection with M. tuberculosis reduced miR-143-3p expression levels and increase xCT mRNA expression levels in a time-dependent manner ( Fig. 3C; see also Fig. S5). To validate these results, we measured miR-142-3p expression in PBMCs isolated from healthy controls, patients with active TB, and TB patients who had been cured. miR-142-3p expression was significantly downregulated in patients with active TB compared with that in healthy controls, but this reduction was reversed after successful anti-TB treatment. xCT gene expression showed the opposite pattern, being significantly increased in PBMCs taken from patients with active TB compared to that in the healthy control but significantly reduced in the cured group compared to that in the active TB group (Fig. 3D and E).
SNP rs13120371 regulates xCT expression through modulating the binding of miR-142-3p to the 3= UTR of the xCT gene. miRNASNP 2.0 software analysis predicted that the SNP rs13120371 lies within a putative binding site for miR-142-3p in the 3= UTR of the xCT gene. In addition, the G allele matches the predicted seed region of miR-142-3p with a gain effect (Fig. 4A), indicating that the G allele resulted in increased affinity for miR-142-3p. To determine whether rs13120371 influences miR-143-3p binding activity, we performed a luciferase reporter assay, constructing two different reporter vectors containing the xCT gene 3= UTR with the AA or GG genotype and transfecting them into 293T cells along with miR-143-3p mimics or a control plasmid. We found that luciferase activity significantly decreased following transfection with pMIR-3=-UTR-G (Fig. 4B). These data indicated that SNP rs13120371 may affect the binding of miR-142-3p to the 3= UTR of the xCT gene.

Treatment with SASP alleviates the increased bacterial burden experienced by patients with the AA rs13120371 polymorphism.
To investigate whether rs13120371 polymorphism impacts M. tuberculosis clearance, monocyte-derived macrophages (MDMs) isolated from individuals with different genotypes were infected with M. tuberculosis H37Ra at multiplicity of infection (MOI) of 10. As expected, bacterial burden was significantly higher in cells from patients carrying the AA genotype than in those carrying the GG genotype (Fig. 5A). However, when the MDMs were pretreated with SASP (200 M), the bacterial burden was significantly decreased in cells from the AA genotype group compared with that in untreated AA genotype cells. However, treatment with SASP had no significant impact on cells from the AG or GG groups (Fig. 5B).

DISCUSSION
TB remains a major threat to human health, especially in developing countries (1). An increasing number of candidate gene and genome-wide association studies have focused on the contributions of the human genetic background to TB susceptibility and resistance (24,25). For example, our previous study demonstrated that increased xCT expression was associated with the development of active TB (19). This finding led to our further more thorough investigation of the impact of xCT polymorphisms on TB in the present study, where we identified a functional genetic variation (rs13120371) in the 3= UTR of the xCT gene. This polymorphism was strongly associated with susceptibility to TB in our cohort. ESR levels and HRCT scores also significantly differed in TB patients with different rs13120371 genotypes. Recently, the results of a genome-wide scan for microRNA (miRNA)-associated SNPs suggested that SNP rs13120371 was associated with cytotoxicity (26). These results all suggest that the rs13120371 polymorphism is associated with inflammation status and disease phenotype through its impact on host cytotoxic response. miRNAs are endogenous noncoding RNAs which negatively regulate gene expression by binding to the 3= UTRs of their target genes. miRNA-mediated translational repression is dependent on sequence complementarity (27), and this may be influenced by a range of factors, including polymorphisms in the miRNAs themselves or in their targets (28,29). Thus, SNPs may alter the function of their host gene and influence a number of cellular processes, including the clinical outcome of and susceptibility to complex diseases such as TB (30)(31)(32). In the present study, we determined that the rs13120371 SNP of the xCT gene lies within the putative binding site of miR-142-3p, and a luciferase reporter assay confirmed that rs13120371 polymorphisms influence miR-143-3p binding activity. Decreased expression of miR-142-3p in CD4 ϩ T cells and peripheral blood in TB patients was previously reported (33), and a miR-142-3p gain- xCT Polymorphism Increases TB Susceptibility of-function assay resulted in the downregulation of N-Wasp expression and decreased mycobacterial intake by macrophages (34). These results are in line with the findings of the present study and represent an initial step toward the comprehensive characterization of the role of miRNAs in TB.
Host immunity against M. tuberculosis is complex, and both innate and adaptive immune components are crucial for protection against disease (35). A key factor influencing susceptibility to TB is IFN-␥ and the underlying Th1 immune response. In the present study, we observed that patients who carried the rs13120371AA genotype had significantly increased numbers of M. tuberculosis antigen-specific IFN-␥ SFCs than those carrying the GG genotype ( Fig. 1A and B). Furthermore, xCT expression was significantly increased in TB patients with the AA genotype compared to that in those with the GG genotype. Bacterial burden was also significantly higher in cells with the AA genotype than in those with GG genotype. Taken together, these results confirm that the rs13120371 AA SNP affects the inflammatory response and outcome of TB.
Tobin et al. previously demonstrated that the response of patients with TB to treatment with glucocorticoids is influenced by host LTA4H genotype (36). Glucocorticoid therapy is only of benefit in patients with the genotype rs17525495 TT, who are characterized by a strong inflammatory response, but there is no benefit to patients with the low-inflammatory rs17525495 CC genotype. In fact, the disease worsened, and the risk of death increased in such patients (36). Similarly, in the present study, we found that macrophages with the rs13120371 AA genotype, which experienced high levels of bacterial burden, were more sensitive to SASP treatment. This suggests that a simple genotyping assay for the presence of high-activity alleles could effectively screen for patients likely to benefit from particular treatments. This has the potential to improve TB outcomes and alleviate the high mortality rate of this disease.
In conclusion, genetic polymorphisms in the xCT gene, which is involved in the host immune response, could have a significant impact on the control and progression of TB. The functional rs13120371 SNP in the 3= UTR region of the xCT gene was confirmed to alter the expression of this gene by affecting the ability of miRNA to bind to it. Furthermore, with the advantage of FDA approval for use in humans, SASP represents a promising novel adjunctive treatment for TB in patients with the rs13120371 AA genotype. A shift toward such individually tailored treatment regimens would have the potential to improve TB outcomes across the world.

MATERIALS AND METHODS
Ethics statement. This study was approved by the Institutional Review Board of the Shenzhen University School of Medicine, China, and informed written consent was obtained from each participant. All experiments and samplings were carried out in accordance with ethical and biosafety protocols approved via the institutional guidelines.
Human subjects and samples. We established a case-control cohort, including 914 patients with pulmonary TB and 936 healthy controls (HC) from Shenzhen Third People's Hospital (Shenzhen, China). Diagnosis of active TB was based on WHO guidelines and according to the clinical symptoms, chest radiography, and microscopy for acid-fast bacilli (AFB), sputum and/or bronchoalveolar lavage fluid M. tuberculosis culture, and response to anti-TB chemotherapy. Patients with allergic diseases, diabetes, cancer, or HIV or who were immunocompromised were excluded. Healthy controls with normal chest radiographic findings and no clinical history of TB were recruited. M. tuberculosis-specific IFN-␥ release assays (IGRA) were used to differentiate individuals with latent TB infection (LTBI) from healthy controls. All SNPs were in Hardy-Weinberg equilibrium (HWE) in the diseased and healthy groups (P Ͼ 0.05).
ELISPOT assay. PBMCs from participants were obtained from whole blood by Ficoll-Hypaque density gradient centrifugation (Ficoll-Paque Plus; Amersham Biosciences) and then resuspended in Lympho-Spot medium (U-CyTech Bioscience, The Netherlands). Then, 2 ϫ 10 5 cells were seeded in duplicates in 96-well plates (MultiScreen-IP; Millipore) precoated with anti-IFN-␥ capture monoclonal antibody (eBioscience). Cells were stimulated with the peptide pool (ESAT-6 amino acids [aa] 21 to 40, aa 51 to 70, and aa 71 to 90 and CFP-10 aa 21 to 40, aa 51 to 70, and aa 66 to 85) for 24 h at 37°C with 5% CO 2 as described previously (37). PBMCs in medium alone or stimulated with phytohemagglutinin (Sigma) at 2.5 g/ml were used as negative or positive controls, respectively. Biotinylated anti-IFN-␥ detection monoclonal antibody (eBioscience) was added for 4 h, followed by the addition of streptavidin-alkaline phosphatase conjugate (Pierce Biotechnology) for 1 h. After a washing step, the nitroblue tetrazolium-BCIP (5-bromo-4-chloro-3-indolylphosphate; Sigma) chromogenic substrate was added. The individual spots were counted by use of an automated image analysis system ELISPOT reader (BioReader 4000 Pro-X; Biosys, Germany). SNP selection and genotyping. Genomic DNA was prepared from whole peripheral blood with the