Association of PTPN22-C1858T Polymorphism With Susceptibility to Mycobacterium tuberculosis and Mycobacterium leprae Infection: A Meta-Analysis

It was previously published that single-nucleotide polymorphism rs2476601 (PTPN22 [protein tyrosine phosphatase non-receptor type 22]-C1858T) might be related to increased sensibility to Mycobacterium tuberculosis and M. leprae infection. However, the results were inconclusive despite a high degree of similarity between both parameters. Herein, we carried out this meta-analysis to systematically summarize and articulate the correlation between PTPN22-C1858T polymorphism and mycobacterial infection. The susceptibility of PTPN22-C1858T carriers with autoimmune conditions receiving immunosuppressive therapy to M. tuberculosis and M. leprae infection was determined. A systematic retrieval of studies on relevance of PTPN22-C1858T polymorphism to susceptibility of M. tuberculosis or M. leprae infection was performed in Chinese National Knowledge Infrastructure, PubMed and Embase databases. We regarded Odds ratios (ORs) and 95% confidence intervals (CIs) as the determined effect size. Finally, four and two case-control studies on tuberculosis and leprosy, respectively, were included. In all genetic models, without indicated association between PTPN22-C1858T polymorphism and tuberculosis’s susceptibility. [C versus T: OR = 0.22 (95% CI: 0.09–0.50, PH = 0.887); CT versus CC: OR = 0.21 (95% CI: 0.09–0.49, PH = 0.889); TT+CT versus CC: OR = 0.21 (95% CI: 0.09–0.49, PH = 0.889)]. A significantly increased risk of leprosy was perceived in patients with the PTPN22-C1858T polymorphism [C versus T: OR = 2.82 (95% CI: 1.02–7.81, PH = 0.108)]. While the PTPN22-C1858T polymorphism is irrelevant to higher susceptibility to the infection of M. tuberculosis in Caucasians and Asians, it is relevant to increased susceptibility to the infection of M. leprae. However, the results of M. leprae are supposed to interpreted with prudence owing to the limited quantity of studies and heterogeneity. Further well-designed studies with sufficient populations are required to verify our conclusions.


INTRODUCTION
Worldwide, tuberculosis is one of the principal lethal causes, ranking above human immunodeficiency virus infection/acquired immunodeficiency syndrome. Despite being a preventable and curable disease, the annual number of people died from tuberculosis is 1.5 million approximately. (1). Tuberculosis accounts for a great share of global burden of disease, particularly among vulnerable populations, and exerts significant impacts on 10 million people annually (2). By recent estimates, over 200,000 new leprosy cases are diagnosed annually. At the end of 2018, the prevalence rate of leprosy is 0.2/10,000 and a total of 184,212 leprosy cases has been registered (3). Leprosy can be cured using effective and affordable multidrug therapy; however, it is a major problem in resource-poor tropical and warm temperate countries and keeps endemic in several low-and middle-income countries globally (4,5). Available evidence indicates that inferior living conditions are possible to raise the risk of leprosy (6). Moreover, the correlation between tuberculosis or leprosy and sociodemographic risk markers such as crowded living conditions, poor sanitation, and poverty has been validated across diverse geographical settings, both in ecological-and individual-level studies (4,7).
Tuberculosis and leprosy are caused by mycobacteria (8). In humans, tuberculosis is predominantly caused by Mycobacterium tuberculosis and occasionally by other components of M. tuberculosis complex, and Mycobacterium leprae result in leprosy. M. tuberculosis and M. leprae are obligate pathogens with a similar morphology and high cell wall lipid content, which is responsible for their increased chromaticity, resistance, and pathogenicity. Moreover, these mycobacteria can cause chronic granuloma in most infected individuals.
Locating on chromosome 1p13, the protein tyrosine phosphatase non-receptor type 22 (PTPN22) gene encodes the protein lymphoid tyrosine phosphatase (LYP), which regulates the activation of protein kinases to modulate intracellular tyrosine phosphorylation incidence and induce various biological effects. PTPN22-C1858T is the most talked about single-nucleotide polymorphism (SNP) in the field of autoimmunity (9); this SNP causes R620W substitution in PTPN22 gene's C-terminal region. PTPN22-C1858T was identified as a missense SNP in PTPN22 gene's exon 14 by using the candidate gene approach in 2004. Moreover, genome-wide association studies showed a correlation between PTPN22-C1858T polymorphism and increased risk of autoimmune diseases (10-15) as well as bacterial infections (16). In addition to disease-causing microorganisms in the common disease spectrum, the susceptibility of individuals carrying PTPN22-C1858T SNP to tuberculosis and leprosy has been proposed in preceding studies (17)(18)(19).
Herein, we performed this meta-analysis to systematically summarize and articulate the correlation between PTPN22-C1858T polymorphism and the risk of mycobacterial infection.

Retrieval of Studies
Two investigators searched literatures from Chinese National Knowledge Infrastructure (CNKI), PubMed and Embase databases independently and systematically, which were updated on April 15, 2020. Retrieved items including "tuberculosis or (pulmonary tuberculosis) or (mycobacterium tuberculosis) or (tubercle bacillus) or (tubercle bacillus) or mtb)", "leprosy or (leprosy bacillus) or (mycobacterium leprae) or (mycogerms leprae) or (wholemycobacterium leprae)", "PTPN22 OR LYP", "polymorphism OR poly-morphism" and susceptibility" were searched in the National Institutes of Health (PubMed) and European (Embase) databases without limitation. The key words "PTPN22基因多态性" and "结核" were searched in the CNKI database. The references in retrieved studies and related reviews were also under reviewing by two investigators without communications.

Inclusion Criteria and Excluding Criteria
All studies pertaining to PTPN22 polymorphisms and infection susceptibility were electronically retrieved. Thereafter, casecontrol studies were identified from their titles and abstracts. The full article of manuscripts fulfilled the inclusion criteria was perused. Inclusion criteria were as follows: (a) studies associated between PTPN22 polymorphisms and infection susceptibility; (b) the category of the study is case-control study among human; (c) original research with detailed explanation of the sample size; (d) with odds ratios (ORs) and its corresponding 95% confidence intervals (CIs); (e) with clear determination of genotype frequency; (f) genotype distribution in controls complying with Hardy-Weinberg equilibrium (HWE) law. Excluding criteria were as follows: (a) non-conformance to the inclusion criteria; (b) article type: abstract, review, comment, and letter; (c) studies with insufficient or no explanation of the sample size; (d) studies without genotype data; (e) repeated publications; (f) family-based studies. If several studies were with same population, the larger one will be included. According to the inclusion criteria and excluding criteria, two investigators screened out eligible literatures independently through titles, abstracts, and full-texts. We settled all dissents through discussion. The third researcher was consulted in the absence of an agreement. Finally, two studies were discussed and excluded after our discussions.

Data Withdrawal
Two investigators withdrawn the data from all included studies independently. The extracted data included the name of first author, year of publication, subject resource, ethnicity, infectious type, diagnostic criteria for infection, genotyping method, number of participants in the control and case groups, cases' and controls' characteristics, all subjects' PTPN22-C1858T genotype frequency and HWE. Therein, populations from all included studies were classified into two ethnic: Caucasians and Asians. Two investigators verified the withdrawn data and achieved agreements on the conclusive data. In the case of a disagreement, the original data was re-verified and re-discussed to reach consensus. If disagreements are still subsistent, we Abbreviations: CI, confidence interval; CNKI, Chinese National Knowledge Infrastructure; HWE, Hardy-Weinberg equilibrium; IFN, interferon; NOS, Newcastle-Ottawa Scale; OR, odds ratio; PTPN22, protein tyrosine phosphatase non-receptor type 22; SNP, single-nucleotide polymorphism; TCR, Tcell receptor. consulted for final judgement from a third investigator. In the end, a study was discussed through third investigator.

Literature Assessment
The Newcastle-Ottawa Scale (NOS) was applied to assess the quality of six eligible studies in meta-analysis by two independent investigators (20) (Methods were shown in Supplement 2). All study were assessed by using a "star system" on the basis of three categories: selection of case and control, subjects' comparability, and exposure's ascertainment. The maximum score is 9 on the NOS scale. If a study with a score ≥ 7, it is considered as a highquality study. The details of assessment has been shown in Table S3.

Statistical Methods
The PRISMA checklist was followed throughout the course of this meta-analysis strictly. (21). The HWE was used to assess genotyping errors in control groups by Chi-squared test in all studies (P < 0.05 was significant). ORs and its corresponding 95% CI is was applied to determine the extent of correlation between PTPN22-C1858T polymorphism and susceptibility to mycobacterial infection. We calculated the pooled ORs in allelic comparison (PTPN22: C versus T), the heterozygote model (PTPN22: CT versus CC), and the dominant model (PTPN22: CT+TT versus CC). P < 0.05 was significant according to Z-test. Heterogeneity was assessed by Q test (P < 0.1 was significant) and I-squared statistic (I 2 : 0%-25%, without heterogeneity; I 2 : 25%-50%, moderate heterogeneity; I 2 : 50%-75%, large heterogeneity; I 2 : 75%-100% extreme heterogeneity; I 2 > 50% represented significant inconsistency in our metaanalysis) (22). We applied a fixed-effect or a random-effects model to pool effect sizes depending on data heterogeneity (23).
Heterogeneity was analyzed by using meta-regression model (P < 0.1 was considered significant) with pathogen (M. tuberculosis and M. leprae) and ethnicity and subgroup analysis stratified by pathogen and ethnicity. No deviation from HWE was shown among controls in all included studies, which are used for further meta-analysis. Additionally, we assessed publication bias by Begg's funnel plots (24). An asymmetric plot with P < 0.05 was identified a significant publication bias. Besides, we also performed sensitivity analysis to estimate the pooled ORs of PTPN22-C1858T polymorphism in each study. The results of the meta-analysis were recalculated after excluding each studies. All statistical analyses were performed by Stata 14.0 software (StataCorp, College Station, TX, USA). Except for certain conditions defined a specific P value, a two-tailed P value < 0.05 was regarded as significant.

Study Retrieval
The study selection process during meta-analysis of association of rs2476601(PTPN22-C1858T) polymorphism with M. tuberculosis and M. leprae infection is shown in Figure 1. Following the initial retrieval of 25 publications through a database search (eleven from PubMed, eight from Embase, and six from CNKI), 18 records were selected after the removal of seven duplicates. Moreover, after careful review of the title and abstract, eight publications were rejected because of their irrelevance to this meta-analysis. The remaining ten publications were full-article reviewed; of these, four were excluded. One of them was related to other SNPs in PTPN22, one was not related to tuberculosis or lepriasis, and two were not case-control studies. Finally, six case-control studies (17)(18)(19)(25)(26)(27) consisting of 2,160 participants (cases = 901; controls = 1,259) were included in metaanalysis. General characteristics of the six studies are in Table 1 while the genotype distribution of subjects is shown in Table 2.
The Relevance Between PTPN22-C1858T Polymorphism and the Susceptibility to Mycobacterial Infection We explored that PTPN22-C1858T polymorphism associates susceptibility to mycobacterial infection. After Q-test and Isquared statistics in various genetic models, it indicated significant heterogeneity. Therefore, we used the randomeffects model in meta-analysis. No genetic model showed correlation between PTPN22-C1858T polymorphism and increased susceptibility to mycobacterial infection [C versus G: OR = 0.59 (95% CI: 0.15-2.29, P H = 0.000) ( Figure 2); CT versus CC: OR = 0.58 (95% CI: 0.14-2.36, P H = 0.000); CT+TT versus CC: OR = 0.58, (95% CI: 0.14-2.36, P H = 0.000)] (     Table 2). Then, a further meta-analysis was performed in M. tuberculosis infection on the four included studies according to ethnicity (18,19,26,27). All three genetic models showed significant homogeneity among Caucasians; therefore, a fixedeffects model was applied. However, all of involved genetic models didn't illustrate an association with significance between PTPN22-C1858T polymorphism and increased susceptibility to mycobacterial infection in Caucasians  Table 2). What's more, all three genetic models showed heterogeneity among Asians and we applied a random-effects model in Asian subgroup. No association with significance between PTPN22-C1858T polymorphism and probability to M.  Table 2).

PTPN22-C1858T Polymorphism Correlated With Increased Susceptibility to M. leprae Infection
The relevance between PTPN22-C1858T polymorphism and increased possibility to M. leprae infection in Caucasian and Asian populations was investigated in two studies (17,25). Owing to limited quantity of studies and heterogeneity (P = 0.108 and I-squared = 61.2%) ( Figure 3B Table 2).

Meta-Regression Was Applied to Analyzing Heterogeneity
Ethnicity and pathogen were included in meta-regression model for heterogeneity analysis. The coefficient of ethnicity is -0.060 (P = 0.955) and the coefficient of pathogen is 2.560 (P = 0.060) in all patients with Mycobacterial infection ( Table 3), which indicated that pathogen (M. tuberculosis/M. leprae) was able to account for the heterogeneity of the associations between mycobacterial infections and PTPN22-C1858T polymorphism.

Publication Bias
No publication bias on relavance between PTPN22-C1858T polymorphism and possibility to tuberculosis (P = 0.734) or leprosy (P = 1.000) was reflected in Begg's funnel plot. Meanwhile, we obtained a symmetrical funnel plots (Figure 4).

Sensitivity Analysis
In order to determine impact from each study on pooled ORs for PTPN22-C1858T polymorphism, we performed sensitivity analysis by the means of deleting one of single study each time in every genetic model. The pooled ORs of both four studies related to M. tuberculosis infection ( Figure S1A) and two studies related to M. leprae infection ( Figure S1B) showed no significant association with PTPN22-C1858T polymorphism in above mentioned genetic models.

DISCUSSION
In this meta-analysis, we analyzed four eligible case-control studies including 595 subjects with tuberculosis and 697 controls and two studies including 306 subjects with leprosy and 562 controls. Our results indicate that PTPN22-C1858T polymorphism is uncorrelated with increased susceptibility to M. tuberculosis infection both among Caucasian and Asian populations in all genetic models but is related to increased susceptibility to M. leprae infection. However, further studies are required to validate the relevance between PTPN22-C1858T polymorphism and leprosy because of the limited quantity of studies and indicated heterogeneity among studies.
In addition to mycobacteria, a number of studies have focused on the relevance of PTPN22-C1858T polymorphism with raised susceptibility to infections by other bacteria (16,28,29). First, it was confirmed that individuals with PTPN22-C1858T polymorphism are predisposed to invasive infections with Streptococcus pneumoniae (16). Second, allogeneic hematopoietic stem cell transplant recipients have been found to have a significantly lower risk of posttransplantation bacterial infections after accepting corresponding allografts from PTPN22-C1858T polymorphic donors (28). Third, PTPN22-C1858T polymorphic patients with chronic mucocutaneous candidiasis showed a higher incidence of bacterial pulmonary infections than normal controls (29). Moreover, other PTPN22 SNPs such as rs33996649 (PTPN22-G788A) might be related to raised susceptibility to tuberculosis. One study showed   (19). Detrimental effect of PTPN22-C1858T polymorphisminduced immune dysregulation on delayed-type hypersensitivity of tuberculosis and leprosy has been extensively studied. Herein, we discussed the relevance of PTPN22-C1858T polymorphism and increased risk of tuberculosis and leprosy in humans (17)(18)(19)(25)(26)(27). The PTPN22 gene encodes for LYP, which render an impact on negative regulation to immunity (30). Moreover, PTPN22-C1858T carriers have variable immune patterns (31). Under physiological conditions, LYP selectively inhibits positive selection during a self-tolerant and immunological T-cell repertoire's formation (32,33), dephosphorylates the suppressive tyrosine located at Src family kinases' C-terminus, and downregulate T-cell receptor (TCR) signaling by combining with growth factor receptor-bound protein 2 adapter (34)(35)(36)(37). Patients with PTPN22-C1858T polymorphism show decreased calcium mobilization that is TCR-induced (38), reduced TCR-CD3z, Zap-70 and Lck's phosphorylation (39), and decreased interleukin-2 production (39). However, some of these studies have yielded inconsistent conclusions. The absolute expression of PTPN22 and proportion of T-cell subsets should be considered during assessment of the overall effect of PTPN22-C1858T polymorphism in T-cellmediated immunity (40). Furthermore, increased susceptibility to mycobacterial infections is closely correlated with the immune status and genetic milieu of hosts. Most healthy individuals with a normal functioning immune system are resistant to M. tuberculosis or M. leprae and do not develop clinical tuberculosis or leprosy (41,42). Host genetics accounts for a large proportion of inherited mycobacterial diseases; epidemiological investigations and functional genetic studies have identified several genes involved in mycobacterial infection. These studies provide information about the inherited predisposition to these infections (43)(44)(45). Hence, we conducted this meta-analysis to determine and articulate the relevance between PTPN22-C1858T polymorphism and risk of tuberculosis and leprosy.
Increased intracellular expression of LYP in patients with mycobacterial infection (46,47), the involvement of PTPN22 in downregulation of T-cell function (30), and the participation of PTPN22 in invasive bacterial infection (16) have been considered as evidence to conclude that PTPN22-C1858T polymorphism increases the risk of mycobacterial disease onset (17). Moreover, PTPN22 status has been shown to cause lymphoproliferative diseases in animal models. Splenomegaly and lymphadenectasis accompanied by initiative germinal center formation and higherlevel antibodies are subtle immune changes observed in PTPN22knockout mice (32). In addition, PTPN22 promotes K63-linked polyubiquitination of the Toll-like receptor signal central promoter TRAF3 to upregulate type I interferons (IFNs) and promote type I IFN-dependent biological effects, causing immune cells to trigger a host defense response. In contrast, PTPN22-C1858T carriers show decreased TRAF3 K63-linked polyubiquitination and type I IFN production (48).
Roles of PTPN22-C1858T in infection and prevention mainly are beneficial to the patients with autoimmune diseases. PTPN22-C1858T is more often discussed in autoimmune disease. With the indispensable treatment with glucocorticoids, patient with autoimmune disease are suffering the from the risk of various infection. Additionally, we tend to claim that population with PTPN22-C1858T is more likely to gain autoimmune disease and the diseases itself carry a high risk of various infection. To this analysis, we are expecting our conslusions are able to provide useful informations for human to evade the potential infection and suggestion for the usage of immunosuppressors, especially for patients with autoimmune disease. To be more concrete, patients with PTPN22-C1858T are recommand a lower dose of immunosuppressors after diagnosis of autoimmne disease and they also are advised to notice all possiablity of infection.
We would like to emphasize a couple of strengths in our metaanalysis. First, we applied subgroup analysis to demonstrate the differences in the correlation between PTPN22-C1858T polymorphism and increased risk of mycobacterial infection. What's more, the conclusion was verified in meta-regression model. Additionally, two independent investigators searched literatures without limitations so that we well controlled the selection bias. Besides, the result from Begg's funnel plot revealed no evident publication bias, which render reliability to our conclusions. Finally, most of each studies in our meta-analysis attained a score representing high quality with regards to NOS quality assessment.
Nevertheless, we also acknowledge several limitations in this meta-analysis. In the first place, other factors contributing to mycobacterial infection were not considered due to the availability of limited information. Second, the results of M. leprae were obtained from two studies. Third, with a low allele frequency less than 1%, PTPN22-C1858T is a rare SNP in Asian population. Therefore, PTPN22-C1858T polymorphism would be rarely detected in most populations except Caucasians of Northern European descent (49). Fourth, the absence of cases and controls with TT in the six included studies may be attributed to their low survival rate for gene mutations. Fifth, tuberculosis and leprosy are considered curable diseases worldwide, especially in developed regions. Moreover, the annual incidence rate of these disease is declining. Furthermore, the diagnosis of leprosy is often not considered outside leprosy-endemic areas (50). Finally, the occurrence of a latency period following infection (42) and lengthy incubation periods of M. leprae (from a month to over 40 years) (50) might have excluded individuals with recessive infection or no infected (in the incubation period) to be registered as positive cases or were registered as controls.
In conclusion, we manifested that this PTPN22-C1858T polymorphism was uncorrelated with raised susceptibility to M. tuberculosis in Caucasians and Asians. In contrast, PTPN22-C1858T polymorphism related to increased susceptibility to M. leprae infection. However, the results of M. leprae are supposed to interpreted with. Further well-designed studies with sufficient populations are required to verify our conclusions.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
YC and SL conceived the topic and analyzed data. SL wrote this manuscript and YC modified it. JY provided statistical suggestions. SL and YZ evaluated the quality of the six included studies. SL, YZ, XW, and TC searched references.Y, ZJZ, and ZZ provided the overall direction of series studies and funding support. All authors contributed to the article and approved the submitted version.