Introduction: Given that PD-L1 is a crucial immune checkpoint in regulating T-cell responses, the aim of this study was to explore the impact of PD-L1 gene polymorphisms and the interaction with cooking with solid fuel on susceptibility to tuberculosis (TB) in Chinese Han populations. Methods: A total of 503 TB patients and 494 healthy controls were enrolled in this case-control study. Mass spectrometry technology was applied to genotype rs2297136 and rs4143815 of PD-L1 genes. The associations between single nucleotide polymorphism (SNPs) and TB were assessed using unconditional logistic regression analysis. Marginal structural linear odds models were used to estimate the gene-environment interactions. Results: Compared with genotype CC, genotypes GG and CG+GG at rs4143815 locus were significantly associated with susceptibility to TB (OR: 3.074 and 1.506, respectively, p < 0.05). However, no statistical association was found between rs2297136 SNP and TB risk. Moreover, the relative excess risk of interaction between rs4143815 of the PD-L1 gene and cooking with solid fuel was 2.365 (95% CI: 1.922–2.809), suggesting positive interactions with TB susceptibility. Conclusion: The rs4143815 polymorphism of the PD-L1 gene was associated with susceptibility to TB in Chinese Han populations. There were significantly positive interactions between rs4143815 and cooking with solid fuel.

Tuberculosis (TB) is a major disease accounting for a large number of cases and deaths worldwide. According to the Global Tuberculosis Report 2020 by the World Health Organization (WHO) [1], approximately 10 million new TB cases occurred, and approximately 1.2 million people died from TB globally in 2019. China ranked third among countries with the highest TB burden, next to India and Indonesia, involving approximately 8.4% of the world’s TB cases.

Generally, humans infected with Mycobacterium tuberculosis (Mtb) have a wide spectrum of clinical severity, suggesting a complex host-pathogen interaction. The polarized effector T-cell response is believed to be critical for combating Mtb in the human immune system, and this process is regulated by various costimulatory signals to maintain homeostatic balance [2]. Programed cell death-1 (PD-1) is a cell-surface co-inhibitory receptor upregulated on CD4+ and CD8+ T cells in patients with TB, and its ligand PD-L1 (also called CD274 and B7-H1) is an immune checkpoint glycoprotein expressed broadly on immune cells and tumor cells [2]. Previous studies have demonstrated the induction of PD-L1 on T cells and antigen-presenting cells, such as macrophages and dendritic cells, during the process of Mtb stimulation. Furthermore, upon ligation with PD-L1, the PD-1 cytoplasmic domain is tyrosine phosphorylated and binds with src homology 2-domain-containing tyrosine phosphatase (SHP-2), leading to dephosphorylation and deactivation of TCR signaling molecules and finally accelerated apoptosis of activated T cells [3]. Blockade of the PD-1/PD-L1 axis is supposed to restore T-cell functions and increase resistance to tumor growth and various infectious diseases [4]. However, regarding Mtb infection, both clinical and mouse studies reported enhanced susceptibility to TB after PD-1/PD-L1 blockade, which indicates a more complicated relationship between them [5].

Single nucleotide polymorphisms (SNPs) play a vital role in altering gene expression and the function of the encoded protein. To date, several SNPs of the PD-1 gene have been identified in relation to cancers, autoimmune diseases, and infectious diseases such as HBV and HCV infection [6]. A few studies have investigated the associations with Mtb infection and have suggested significant results in selected PD-1 variants. Similarly, polymorphisms of the PD-L1 gene, such asrs2297136 and rs4143815, have been proved to be associated with diseases such as cancer and type 1 diabetes [7, 8], and few research has reported the relevance of polymorphisms to TB risk. Considering the important role of PD-L1 in immune response and its tight relationship with PD-1, we speculated that genomic variations in PD-L1 may play a role in TB development.

Furthermore, indoor air pollution (IAP) like kitchen smoke has been suspected and proven to contribute to TB [9, 10], particularly in children [11]. Fuel types used in households are mainly categorized into solid fuel (coal and biomass) and non-solid fuel worldwide. People cooking or heating with coal and biomass are exposed to harmful kitchen smoke containing various chemical constituents, such as nitrogen oxides and particulate matter (PM) [11], which could compromise the immune function of the respiratory system and make individuals more vulnerable to infections [12]. It is estimated 3 billion people, about 40% of the world’s population, use stoves that run solid fuels [13]. China’s rural populations still mostly rely on coal and biomass fuels for cooking and heating (61% and 15% of total energy use, respectively) [14]. Population-based studies have suggested a possible association between cooking with solid fuel and predisposition to TB. Moreover, biological evidence has shown that IAP including smoke from solid fuel may initiate an inflammatory response in the lung and lead to T-cell proliferation and cytokine production, whereas lowering the antimycobacterial activity of macrophages and alteration of gene expression has been reported [11]. Air pollution particles can activate certain proteins, such as ERK1/2 and STAT, along with the production proinflammatory cytokines, which are able to modulate PD-L1 expression in lung tissues and cells [15]. Therefore, we speculated that there might be associations between the harmful smoke and PD-L1 on TB. This study aimed to explore the impact of PD-L1 variations on TB in Chinese Han populations and examine whether there is any interaction between the PD-L1 gene and cooking with solid fuel.

Source of Subject

We estimated the sample size of 445 subjects for patients and controls, respectively, based on the unmatched case-control study (estimated minor allele frequency in the rs2297136 locus was 0.21, OR = 1.65, α = 0.05, β = 0.1). Stratified sampling was used to select the study samples. According to the two-stage sampling principle, four centers for disease control (CDCs) (i.e., Hongjiang City CDC, Qidong County CDC, Yueyanglou District CDC, and Yueyang County CDC) at the county level out of 122 CDCs were randomly chosen across the province, and TB cases newly diagnosed were randomly selected as case groups. All cases were confirmed according to the unified TB diagnostic criteria formulated by the Ministry of Health of China. The control group was selected using the same strategy, in which the eligible controls were randomly selected from permanent residents of Xin’ansi Community. Our previous study outlined the specific inclusion criteria [10].

For both groups, all subjects were Chinese Han population aged ≥18 years and were not related within three generations. In addition, we excluded people with primary immune deficiency, HIV infection, severe malnutrition, long-term hormone use, tumors, diabetes, and organ transplants.

Information and Sample Collection

A self-designed questionnaire was used to collect demographic and environmental information, including gender, educational background, body mass index (BMI), BCG vaccination history, alcohol consumption, tea-drinking, kitchen ventilation facilities, and exposure to solid fuel fumes from cooking. In this study, household cooking fuels were classified as solid fuel and non-solid fuel. The former included wood, coal, crop wastes, etc.; the latter included liquefied, natural gas, etc. Information on fuel exposure mainly covered the type of primary cooking fuel, exposure frequency, and duration. Individuals who cooked for their households using solid fuels were classified as cooking with solid fuel.

Furthermore, we collected 5 mL of venous blood from each participant and stored it at 4°C until use, using ethylenediaminetetraacetic acid (EDTA) anticoagulant tubes. Thereafter, we extracted the genome from peripheral white blood cells using the Wizard Genomic DNA kit and stored at −80°C.

Selection of SNPs and Genotyping

The selected SNPs must fulfill at least one of the following criteria: being associated with TB or other infectious diseases according to existing evidence. Only polymorphisms with minor allele frequency >5% were included to ensure statistical efficacy. After searching the dbSNP database for the corresponding SNPs and their frequencies (http://www.ncbi.nlm.nih.gov), we finally determined two SNPs (rs2297136 and rs4143815) of the PD-L1 gene. Using Assay Design 3.1 (Sequenom) and the site sequences of the two SNPs identified in the gene bank, we designed suitable primers for genotyping. Quality of the synthesized primers was checked by MALDI-TOF (matrix-assisted laser desorption ionization time-of-flight mass spectrometry).

A 5-μl PCR reaction system was used, including 1 μL PCR primer mix, 0.1 μL dNTP (25 mm), 1.8 μL ddH2O, 0.2 μL Hotstar, 0.5 μL 10*PCR buffer, 0.4 μL MgCl2 (25 mm), and 1 μL DNA (20–50 ng). Predegeneration was conducted for 2 min at 95°C, amplification was carried out for 45 cycles (56°C for 30 s, 95°C for 30 s, 72°C for 60 s), followed by extension at 72°C for 5 min. The digestive enzyme reaction was carried out with 0.3 μL SAP enzyme, 0.17 μL SAP buffer, and 1.53 μL ddH2O. The reaction time was 40 min at 37°C, followed by 5 min at 85°C. One-base extension reactions required 0.041 iplex enzyme, 0.2 μL terminator mix, 0.2 μL iplex buffer, 0.94 μL extend primer mix, and 0.619 μL ddH2O. The conditions for the reaction were 94°C predegeneration for 30 s, 40 cycles of amplification containing 3 min extension at 72°C and 5 cycles at 80°C, 52°C, and 94°C, respectively, for 5 s. Subsequently, we used a 6-mg resin plate and 384-well SpectroCHIP (Sequenom) chip for resin purifying and spotting (MassARRAYNanodispenser RS1000). Molecular weight differences can be interpreted as differences in base composition due to SNP polymorphisms and finally used for determining SNP typing. More details were described in our previous publication [10].

Statistical Analysis

We used Epidata 3.0 and SAS 9.2 for data input and analysis and the χ2 test for grouped data and Hardy-Weinberg equilibrium. We used unconditional logistic regression models with adjustments for possible confounders like marital status, etc. to estimate the odds ratios and 95% CIs and the relative excess risk of interaction (RERI) to analyze the additive gene-environment interactions if the main effect of SNPs in the PD-L1 gene on TB was meaningful. Marginal structural linear odds models [16] were used for the point estimation and interval estimation of RERI. An RERI >0 indicates a positive interaction.

Ethics Declarations

The studies involving human participants were reviewed and approved by the Medical Ethical Committee of Xiangya School of Public Health, Central South University (Ref No. XYGW-2018-11). All experiments obeyed the Declaration of Helsinki. We obtained the informed consent from participants of the study.

A case-control study was used to explore the associations of the selected factors with TB susceptibility, in which 503 TB patients and 494 healthy controls were included. Statistically significant differences (p < 0.05) were found in terms of BMI, history of BCG vaccination, smoking, alcohol consumption, tea drinking, cooking with solid fuel, and kitchen ventilation facilities. However, no statistical significance (p > 0.05) was observed in sex, age, marital status, and educational background between the two groups (Table 1).

Table 1.

Demographic characteristics and associated risk factors in tuberculosis patients versus healthy controls

TB patients (n = 503)Healthy controls (n = 494)χ2p value
n%n%
Sex 
 Male 369 73.36 348 70.45 1.048 0.306 
 Female 134 26.64 146 29.55   
Age, years 
 18–30 68 13.52 78 15.79 1.908 0.592 
 31–50 190 37.77 178 36.03   
 51–70 167 33.20 153 30.97   
 >70 78 15.51 85 17.21   
Marital status 
 Married 345 68.59 330 66.80 0.364 0.546 
 Other 158 31.41 164 33.20   
Educational background 
 Primary school or below 198 39.36 220 44.53 2.824 0.244 
 Junior high school 161 32.01 148 29.96   
 Senior high school or above 144 28.63 126 25.51   
BMI, kg/m2 
 <18.5 180 35.79 169 34.21 9.310 0.010a 
 18.5–24.9 297 59.05 274 55.47   
 ≥25.0 26 5.17 51 10.32   
History of BCG vaccination 
 Yes 105 20.87 141 28.54 7.884 0.005a 
 No 398 79.13 353 71.46   
Smoking 
 Yes 298 59.24 240 48.58 11.403 <0.001a 
 No 205 40.76 254 51.42   
Alcohol drinking 
 Yes 91 18.09 77 15.59 1.116 0.291 
 No 412 81.91 417 84.41   
Tea drinking 
 Yes 247 49.11 293 59.31 10.457 0.001a 
 No 256 50.89 201 40.69   
Cooking with solid fuel 
 Yes 299 59.44 250 50.61 7.864 0.005a 
 No 204 40.56 244 49.39   
Kitchen ventilation facilities 
 Unavailable 295 58.65 217 43.93 21.620 <0.001a 
 Available 208 41.35 277 56.07   
TB patients (n = 503)Healthy controls (n = 494)χ2p value
n%n%
Sex 
 Male 369 73.36 348 70.45 1.048 0.306 
 Female 134 26.64 146 29.55   
Age, years 
 18–30 68 13.52 78 15.79 1.908 0.592 
 31–50 190 37.77 178 36.03   
 51–70 167 33.20 153 30.97   
 >70 78 15.51 85 17.21   
Marital status 
 Married 345 68.59 330 66.80 0.364 0.546 
 Other 158 31.41 164 33.20   
Educational background 
 Primary school or below 198 39.36 220 44.53 2.824 0.244 
 Junior high school 161 32.01 148 29.96   
 Senior high school or above 144 28.63 126 25.51   
BMI, kg/m2 
 <18.5 180 35.79 169 34.21 9.310 0.010a 
 18.5–24.9 297 59.05 274 55.47   
 ≥25.0 26 5.17 51 10.32   
History of BCG vaccination 
 Yes 105 20.87 141 28.54 7.884 0.005a 
 No 398 79.13 353 71.46   
Smoking 
 Yes 298 59.24 240 48.58 11.403 <0.001a 
 No 205 40.76 254 51.42   
Alcohol drinking 
 Yes 91 18.09 77 15.59 1.116 0.291 
 No 412 81.91 417 84.41   
Tea drinking 
 Yes 247 49.11 293 59.31 10.457 0.001a 
 No 256 50.89 201 40.69   
Cooking with solid fuel 
 Yes 299 59.44 250 50.61 7.864 0.005a 
 No 204 40.56 244 49.39   
Kitchen ventilation facilities 
 Unavailable 295 58.65 217 43.93 21.620 <0.001a 
 Available 208 41.35 277 56.07   

ap < 0.05.

The genotypic frequencies of PD-L1 polymorphisms are shown in Table 2. After adjusting for the covariates of sex, age, marital status, educational background, body mass index, smoking, alcohol consumption, tea-drinking, BCG vaccination, cooking with solid fuel, and kitchen ventilation facilities, multivariate unconditional logistic regression analysis showed that rs4143815 of the PD-L1 gene was associated with susceptibility to TB. In detail, genotypes GG and CG+GG at rs4143815 (compared with genotype CC) were significantly associated with increased risk of TB, with adjusted ORs reaching 3.074 and 1.506, respectively. However, no statistical association was observed between the rs2297136 variant and TB risk (Table 2).

Table 2.

PD-L1 gene polymorphisms in tuberculosis patients versus healthy controls

TB patientsHealthy controlsORc (95% CI)ORab (95% CI)
n%n%
rs4143815 
 CC 320 63.62 370 74.90 
 CG 143 28.43 110 22.27 1.503 (1.125–2.009)a 1.296 (0.941–1.787) 
 GG 40 7.95 14 2.83 3.304 (1.765–6.183)a 3.074 (1.609–5.871)a 
 CG+GG 183 36.38 124 25.10 1.706 (1.299–2.241)a 1.506 (1.115–2.035)a 
rs2297136 
 TT 277 55.07 290 58.70 
 TC 203 40.36 181 36.64 1.174 (0.906–1.522) 1.131 (0.864–1.482) 
 CC 23 4.57 23 4.66 1.047 (0.574–1.909) 0.777 (0.412–1.466) 
 TC+CC 226 44.93 204 41.30 1.160 (0.902–1.491) 1.088 (0.838–1.413) 
TB patientsHealthy controlsORc (95% CI)ORab (95% CI)
n%n%
rs4143815 
 CC 320 63.62 370 74.90 
 CG 143 28.43 110 22.27 1.503 (1.125–2.009)a 1.296 (0.941–1.787) 
 GG 40 7.95 14 2.83 3.304 (1.765–6.183)a 3.074 (1.609–5.871)a 
 CG+GG 183 36.38 124 25.10 1.706 (1.299–2.241)a 1.506 (1.115–2.035)a 
rs2297136 
 TT 277 55.07 290 58.70 
 TC 203 40.36 181 36.64 1.174 (0.906–1.522) 1.131 (0.864–1.482) 
 CC 23 4.57 23 4.66 1.047 (0.574–1.909) 0.777 (0.412–1.466) 
 TC+CC 226 44.93 204 41.30 1.160 (0.902–1.491) 1.088 (0.838–1.413) 

ap< 0.05.

bAdjusted for the covariates of sex, age, marital status, educational background, body mass index, smoking, alcohol drinking, tea drinking, BCG vaccination, cooking with solid fuel. and kitchen ventilation facilities.

Furthermore, marginal structural linear odds models were used to examine the interactions between the selected sites of the PD-L1 gene and the selected environmental factors. After adjusting for the covariates of sex, age, marital status, educational background, body mass index, smoking, alcohol consumption, tea-drinking, BCG vaccination, and kitchen ventilation facilities, the RERI between rs4143815 of the PD-L1 gene and cooking with solid fuel was 2.365 (95% CI: 1.922–2.809), suggesting positive interactions with TB susceptibility (Table 3).

Table 3.

Impact of interactions between rs4143815 and cooking with solid fuel on incidence of tuberculosis

Cooking with solid fuelrs4143815RERIcRERIaa (95% CI)
CCCG+GG
No 1.121 1.247 2.365 (1.922–2.809)b 
Yes 1.104 2.472 
Cooking with solid fuelrs4143815RERIcRERIaa (95% CI)
CCCG+GG
No 1.121 1.247 2.365 (1.922–2.809)b 
Yes 1.104 2.472 

RERI, relative excess risk of interaction. RERI >0 suggests positive interactions.

aAdjusted for the covariates of sex, age, marital status, educational background, body mass index, smoking, alcohol drinking, tea drinking, BCG vaccination, and kitchen ventilation facilities.

bp< 0.05.

Tuberculosis involves complex Mtb-human interactions, in which many different cell types, including T cells, dendritic cells, macrophages, and natural killer cells, participate in the progression from asymptomatic form to clinical disease [17]. As a cell-surface receptor expressed on most of these cells, PD-L1 exerts a modulatory role in human innate and particularly adaptive immune responses by binding to PD-1, expanding Tregs, and suppressing efficient clearance against pathogens and tumor cells [2]. The human gene PD-L1 is located on chromosome 9p24.2 [18], where numerous variants are associated with several diseases, except TB [8]. Therefore, this study aimed to evaluate the possible relationship between PD-L1 polymorphisms and TB and gene-environment interactions with cooking with solid fuel.

The results showed PD-L1rs4143815 SNP were associated with susceptibility to TB. Individuals with GG/CG+GG genotypes at the locus displayed a significantly increased vulnerability to TB than those carrying the CC genotype. To our knowledge, this finding is novel since no relevant study has been reported previously. The variant rs4143815 is located at the 3′-untranslated region (3′-UTR) of the PD-L1 gene, and previous investigations have shown that structural variations such as duplications and translocations would cause aberrant PD-L1 expression in several cancers [19]. Furthermore, studies have shown that miRNAs can bind to the 3′-UTR of messenger RNA (mRNA) and regulate gene translational repression [18]. As a genetic variation within 3′-UTR of the PD-L1 gene, rs4143815 SNP is supposed to disturb the suppressive function by affecting the binding capacity of PD-L1 mRNA to its corresponding miRNA miR-570, thus leading to aberrant PD-L1 expression.

To date, numerous cancer types have been associated with the rs4143815 variant. Xie et al. [20] found that the rs4143815 GG genotype had higher levels of PD-L1 and was associated with an increased risk of hepatocellular carcinoma (HCC) compared with CC genotypes, whereas individuals with the CC genotype displayed better prognosis compared with the CG and GG genotypes. Similarly, several studies have demonstrated that G allele carriers confer an increased susceptibility to cancers including gastric adenocarcinoma, ovarian cancer compared with CC genotypes [21]. Moreover, an elevated PD-L1 expression among them was reported, suggesting that the G allele in rs4143815 of PD-L1 may increase PD-L1 expression by affecting the binding force between miR-570 and mRNA [21]. In addition, pathological features, including tumor size, differentiation degree, and tumor node metastasis (TNM) stage, were associated with rs4143815 SNP [20]. However, controversy still exists regarding the function of PD-L1 rs4143815 in diseases. In a previous study, the GG genotype of rs4143815 decreased the risk of breast cancer [22]. In addition, an insignificant association of rs4143815 SNP with the risk of colorectal cancer, non-small cell lung cancer (NSCLC), and esophageal squamous cell carcinoma has been reported in previous studies [23‒25]. Further, research on type 1 diabetes suggested a larger stem-loop structure for G allele carriers, which might facilitate the interaction between the 3′-UTR of PD-L1 and miR-570, resulting in lower PD-L1 expression [8]. Overall, discrepancies among studies may be attributed to the essential differences between these diseases, and the substantial heterogeneity of study designs, ethnicities, and confounders control might have some role. In this study, the mechanism mentioned above may explain our findings; however, further investigations on genetic mutation and PD-L1 expression among TB patients should be conducted to confirm this.

In the rs2297136SNP, our results suggested no significant association with TB risk. However, previous studies showed that it is a binding site for miR-324-5p located in the 3′-UTR of the PD-L1 gene. The mechanism by which the rs2297136 polymorphism acts on the expression of PD-L1 protein is supposed to be similar to that of the rs4143815 SNP, with the vital step being the interaction between miR-324-5p and PD-L1 mRNA [7, 20]. Several studies have been performed to confirm the process and examine the associations of the rs2297136 SNP with multiple diseases. Previous studies proposed that rs2297136 TT and TC genotypes contributed to the poor prognosis of multiple myeloma and colorectal cancer, respectively [26, 27], and a higher level of PD-L1 mRNA was demonstrated [28]. In contrast, Xie et al. [20] revealed a significant association between the TT/TC genotypes and better prognosis of HCC. In NSCLC, Du et al. [24] reported a higher risk of NSCLC among G allele carriers, with the transcriptional activity of the PD-L1 gene being suppressed in a luciferase assay. Lee et al. [29] found a better prognosis for patients carrying the C allele after chemotherapy. In addition, another study showed that individuals with GG genotypes had lower PD-L1 expression levels and poor prognosis of gastric cancer [7], indicating the complex role of mutation genotypes on PD-L1 expression and disease progression. Moreover, the current knowledge regarding the correlations between rs2297136 and infectious diseases, including TB, is limited; more studies should be conducted despite a preliminary view of the field provided by this study.

Furthermore, marginal structural linear odds model analysis suggested that rs4143815 of the PD-L1 gene had positive interactions with cooking with solid fuel on TB susceptibility. It is well known that cooking with solid fuel is one of major sources of IAP, which has been implicated in the occurrence and sequela of various diseases. However, evidence remains inconsistent in TB research currently. Many epidemiological studies have attempted to clarify the dose-response relationship, but the results were equivocal. Some studies reporting a significant association [9, 10], whereas others did not [30]. The discrepancy among studies may be due to the substantial heterogeneity in exposure definitions, accuracy of exposure monitoring, and quantitative effect estimates. In addition, kitchen ventilation facilities and room structures may have indirect effects by altering the exposure concentration. Although conflict exists in population studies, biological evidence supports these associations. Basic studies have shown that biomass smoke from different types of solid fuels contains a complex mixture of approximately 200 chemicals, such as carbon monoxide, varying sizes of PM, sulfur and nitrogen oxides, polycyclic aromatic hydrocarbons (PAH), aldehydes, free radicals and non-radical oxidizing species, and volatile organic compounds. Chronic exposure to these hazardous substances can impair mucociliary clearance of the lung and facilitate persistent inflammation through the secretion of pro-inflammatory cytokines [31]. PAH can inhibit the immune function, thereby weakening the immunity of the human body [10]. PM such as PM2.5 can act as a carrier for several toxic components, promote inflammatory responses, and influence gene transcription [31]. In addition, gene-environment interactions could be mediated by epigenetic pathways [32, 33]. Strong evidence suggests that epigenetic mechanisms mediate the effects of gene-environmental factors on certain disease susceptibility and development [34]. Studies have shown that Mtb induces epigenetic changes in the host and plays an important role in host immune evasion strategies [35]. Epigenetically regulatory T cells have been reported to have closely relationship in gene-environment interactions [32, 36]. Among those cells, the differentiation of Tregs is very crucial due to its role in the suppression of inflammation [32, 36]. Studies have indicated the expression of PD-L1 is associated with Human iTreg differentiation by a complex system of genetic and environmental and their interactions [37]. These mechanisms provided the biological plausibility for associations between the risk factor of interest and Mtb infection. They revealed that there might be some cross-talk in the mechanism between the environmental factor and the PD-L1 pathway. In this regard, our findings indicate rs4143815 SNP of the PD-L1 gene are at a higher risk of contracting TB if exposed to cooking with solid fuel. Behavioral interventions including reducing exposure to cooking with solid fuel will not only lower TB incidence from the perspective of the main effect but also decrease their risk of TB by reducing interactions. The novel linkages between modifiable factors and genetic polymorphisms may provide new directions for TB prevention and are helpful in understanding precise prevention strategies among different populations.

This study had some limitations. First, we detected limited SNPs of the PD-L1 gene, and many other genetic and environmental backgrounds among the population may cause some improbability in the study. Hence, more studies on gene-gene, gene-environment, and environment-environment interactions should be conducted to better elucidate the etiology of TB. Second, this study did not examine levels of PD-L1 protein, leaving deep mechanisms behind the observed associations on TB susceptibility not been determined. Thus, further in-depth studies are still needed to validate our findings on molecular levels. Third, this study was a case-control study, so our findings need to be validated in a larger cohort study in the future in the Chinese Han population. In addition, the participants in this study were all limited to Chinese Han population, considering the probable existence of genetic deviations among different ethnic populations, external validation studies in other populations are still required to determine the generalizability of our findings. Despite these limitations, this study precisely defined exposure to cooking with solid fuel, and adjusting for covariates such as age, smoking, alcohol consumption, and kitchen ventilation facilities, the results should be reliable.

This study suggests that rs4143815 of the PD-L1 gene is associated with a significantly increased risk of TB in a Chinese population, and there were positive interactions between rs4143815 SNP and cooking with solid fuel. These findings are helpful for studying the pathogenesis of TB infection and identifying the high-risk populations of acquiring TB, which may contribute to implementing the more targeted measures against developing TB. Further studies in larger populations are needed to verify these findings.

We thank our partners Dr. Bai Liqiong and Dr. Xu Zuhui (Hunan Institute of Tuberculosis Prevention and Treatment) for their input into this work.

The study protocol was reviewed and approved by the Xiangya School of Public Health Central South University Ethics Review Committee Ref No. XYGW/11/2018. All subjects enrolled in the study were over 19 years old and so parental consent was not required. The written informed consent was obtained from each participant before conducting the study. Investigations were carried out in accordance with the principles of the Helsinki Declaration of 1975, as revised in 2008.

The authors have no conflicts of interest to declare.

This research was funded by Mega Project of Research on the Prevention and Control of HIV/AIDS, Viral Hepatitis Infectious Diseases, Grant No. 2018ZX10302302001008; National Natural Science Foundation of China, Grant No. 81803298 and 81800275; and Hunan Provincial Natural Science Foundation of China, Grant No. 2020JJ4762. The funding organization played no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript.

M.W. and M.C. conceived and planned the study. J.W., K.T., M.C., H.Z., and Q.W. were involved in planning and supervised the data collection. Z.L., C.W., R.A., Y.L., and H.T. contributed to the data collection. K.T., J.W., Q.W., Z.L., and H.Z. performed the statistical analyses. K.T. wrote the draft and the final version of the manuscript. K.T., M.W., M.C., H.Z., Q.W., H.T., and L.C. contributed to the literature search, language editing, and manuscript revision. M.W. and J.W. contributed to the overall project supervision and manuscript revision. All authors read and approved the final version of the manuscript.

Additional Information

Kun Tang and Jing Wang have contributed equally to this work.

The data that support the findings of this study are not publicly available as they include some data from Hunan Institute of Tuberculosis Prevention and Treatment and consent forms did not include broad data sharing. The data generated and/or analyzed during this study are available from the corresponding author upon reasonable request.

1.
World Health Organization
.
Global tuberculosis report
;
2020
. https://apps.who.int/iris/handle/10665/336069?locale-attribute=en&
2.
Suarez
GV
,
Melucci Ganzarain
CDC
,
Vecchione
MB
,
Trifone
CA
,
Marin Franco
JL
,
Genoula
M
, et al
.
PD-1/PD-L1 pathway modulates macrophage susceptibility to Mycobacterium tuberculosis specific CD8(+) T cell induced death
.
Sci Rep
.
2019
;
9
(
1
):
187
.
3.
Rota
G
,
Niogret
C
,
Dang
AT
,
Barros
CR
,
Fonta
NP
,
Alfei
F
, et al
.
Shp-2 is dispensable for establishing T cell exhaustion and for PD-1 signaling in vivo
.
Cell Rep
.
2018
;
23
(
1
):
39
49
.
4.
Huang
C
,
Ge
T
,
Xia
C
,
Zhu
W
,
Xu
L
,
Wang
Y
, et al
.
Association of rs10204525 genotype GG and rs2227982 CC combination in programmed cell death 1 with hepatitis B virus infection risk
.
Medicine
.
2019
;
98
(
35
):
e16972
.
5.
Anand
K
,
Sahu
G
,
Burns
E
,
Ensor
A
,
Ensor
J
,
Pingali
SR
, et al
.
Mycobacterial infections due to PD-1 and PD-L1 checkpoint inhibitors
.
ESMO Open
.
2020
;
5
(
4
):
e000866
.
6.
Salmaninejad
A
,
Khoramshahi
V
,
Azani
A
,
Soltaninejad
E
,
Aslani
S
,
Zamani
MR
, et al
.
PD-1 and cancer: molecular mechanisms and polymorphisms
.
Immunogenetics
.
2018
;
70
(
2
):
73
86
.
7.
Wu
Y
,
Zhao
T
,
Jia
Z
,
Cao
D
,
Cao
X
,
Pan
Y
, et al
.
Polymorphism of the programmed death-ligand 1 gene is associated with its protein expression and prognosis in gastric cancer
.
J Gastroenterol Hepatol
.
2019
;
34
(
7
):
1201
7
.
8.
Qian
C
,
Guo
H
,
Chen
X
,
Shi
A
,
Li
S
,
Wang
X
, et al
.
Association of PD-1 and PD-L1 genetic polymorphyisms with type 1 diabetes susceptibility
.
J Diabetes Res
.
2018
;
2018
:
1614683
.
9.
Patel
V
,
Foster
A
,
Salem
A
,
Kumar
A
,
Kumar
V
,
Biswas
B
, et al
.
Long-term exposure to indoor air pollution and risk of tuberculosis
.
Indoor Air
.
2021
;
31
(
3
):
628
38
.
10.
Chen
M
,
Deng
J
,
Su
C
,
Li
J
,
Wang
M
,
Abuaku
BK
, et al
.
Impact of passive smoking, cooking with solid fuel exposure, and MBL/MASP-2 gene polymorphism upon susceptibility to tuberculosis
.
Int J Infect Dis
.
2014
;
29
:
1
6
.
11.
Obore
N
,
Kawuki
J
,
Guan
J
,
Papabathini
SS
,
Wang
L
.
Association between indoor air pollution, tobacco smoke and tuberculosis: an updated systematic review and meta-analysis
.
Publ Health
.
2020
;
187
:
24
35
.
12.
Ibironke
O
,
Carranza
C
,
Sarkar
S
,
Torres
M
,
Choi
HT
,
Nwoko
J
, et al
.
Urban air pollution particulates suppress human T-cell responses to Mycobacterium tuberculosis
.
Int J Environ Res Public Health
.
2019
;
16
(
21
):
4112
.
13.
World Health Organization
.
Global Health observatory (GHO)data
;
2019
. Available from: https://www.who.int/gho/phe/indoor_air_pollution/exposure/en
14.
Zheng
X
.
Chinese household energy consumption Report2015
.
Beijing
:
Science Press
;
2016
.
15.
Rui
C
,
Defu
L
,
Lingling
W
,
Jiahui
D
,
Richeng
X
,
Yuanyuan
Y
, et al
.
Cigarette smoke or motor vehicle exhaust exposure induces PD-L1 upregulation in lung epithelial cells in COPD model rats
.
COPD
.
2022
;
19
(
1
):
206
15
.
16.
VanderWeele
TJ
,
Vansteelandt
S
.
A weighting approach to causal effects and additive interaction in case-control studies: marginal structural linear odds models
.
Am J Epidemiol
.
2011
;
174
(
10
):
1197
203
.
17.
Xu-Monette
ZY
,
Zhang
M
,
Li
J
,
Young
KH
.
PD-1/PD-L1 blockade: have we found the key to unleash the antitumor immune response
.
Front Immunol
.
2017
;
8
:
1597
.
18.
Zou
J
,
Wu
D
,
Li
T
,
Wang
X
,
Liu
Y
,
Tan
S
.
Association of PD-L1 gene rs4143815 C>G polymorphism and human cancer susceptibility: a systematic review and meta-analysis
.
Pathol Res Pract
.
2019
;
215
(
2
):
229
34
.
19.
Kataoka
K
,
Shiraishi
Y
,
Takeda
Y
,
Sakata
S
,
Matsumoto
M
,
Nagano
S
, et al
.
Aberrant PD-L1 expression through 3'-UTR disruption in multiple cancers
.
Nature
.
2016
;
534
(
7607
):
402
6
.
20.
Xie
Q
,
Chen
Z
,
Xia
L
,
Zhao
Q
,
Yu
H
,
Yang
Z
.
Correlations of PD-L1 gene polymorphisms with susceptibility and prognosis in hepatocellular carcinoma in a Chinese Han population
.
Gene
.
2018
;
674
:
188
94
.
21.
Tan
D
,
Sheng
L
,
Yi
QH
.
Correlation of PD-1/PD-L1 polymorphisms and expressions with clinicopathologic features and prognosis of ovarian cancer
.
Cancer Biomark
.
2018
;
21
(
2
):
287
97
.
22.
Karami
S
,
Sattarifard
H
,
Kiumarsi
M
,
Sarabandi
S
,
Taheri
M
,
Hashemi
M
, et al
.
Evaluating the possible association between PD-1 (Rs11568821, Rs2227981, Rs2227982) and PD-L1 (Rs4143815, Rs2890658) polymorphisms and susceptibility to breast cancer in a sample of southeast Iranian women
.
Asian Pac J Cancer Prev
.
2020
;
21
(
10
):
3115
23
.
23.
Catalano
C
,
da Silva Filho
MI
,
Frank
C
,
Jiraskova
K
,
Vymetalkova
V
,
Levy
M
, et al
.
Investigation of single and synergic effects of NLRC5 and PD-L1 variants on the risk of colorectal cancer
.
PLoS One
.
2018
;
13
(
2
):
e0192385
.
24.
Du
W
,
Zhu
J
,
Chen
Y
,
Zeng
Y
,
Shen
D
,
Zhang
N
, et al
.
Variant SNPs at the microRNA complementary site in the B7H1 3′untranslated region increase the risk of nonsmall cell lung cancer
.
Mol Med Rep
.
2017
;
16
(
3
):
2682
90
.
25.
Zhou
RM
,
Li
Y
,
Liu
JH
,
Wang
N
,
Huang
X
,
Cao
SR
, et al
.
Programmed death-1 ligand-1 gene rs2890658 polymorphism associated with the risk of esophageal squamous cell carcinoma in smokers
.
Cancer Biomark
.
2017
;
21
(
1
):
65
71
.
26.
Zhang
HLD
,
Zhou
X
,
Liu
Y
,
Li
M
,
Liu
S
.
The influence of PD-L1 rs2297136 polymorphism on the clinical outcomes of postoperative colorectal cancer patients receiving capecitabine-based adjuvant chemotherapy
.
China Oncol
.
2019
;
29
(
10
):
795
802
.
27.
Kasamatsu
T
,
Awata
M
,
Ishihara
R
,
Murakami
Y
,
Gotoh
N
,
Matsumoto
M
, et al
.
PDCD1 and PDCD1LG1 polymorphisms affect the susceptibility to multiple myeloma
.
Clin Exp Med
.
2020
;
20
(
1
):
51
62
.
28.
Wang
L
,
Wang
H
,
Chen
H
,
Wang
WD
,
Chen
XQ
,
Geng
QR
, et al
.
Serum levels of soluble programmed death ligand 1 predict treatment response and progression free survival in multiple myeloma
.
Oncotarget
.
2015
;
6
(
38
):
41228
36
.
29.
Lee
SY
,
Jung
DK
,
Choi
JE
,
Jin
CC
,
Hong
MJ
,
Do
SK
, et al
.
PD-L1 polymorphism can predict clinical outcomes of non-small cell lung cancer patients treated with first-line paclitaxel-cisplatin chemotherapy
.
Sci Rep
.
2016
;
6
:
25952
.
30.
Emuron
D
,
Siddharthan
T
,
Morgan
B
,
Pollard
SL
,
Grigsby
MR
,
Goodman
D
, et al
.
Lack of an association between household air pollution exposure and previous pulmonary tuberculosis
.
Lung
.
2019
;
197
(
6
):
793
801
.
31.
Capistrano
SJ
,
van Reyk
D
,
Chen
H
,
Oliver
BG
.
Evidence of biomass smoke exposure as a causative factor for the development of COPD
.
Toxics
.
2017
;
5
(
4
):
36
.
32.
Potaczek
DP
,
Harb
H
,
Michel
S
,
Alhamwe
BA
,
Renz
H
,
Tost
J
.
Epigenetics and allergy: from basic mechanisms to clinical applications
.
Epigenomics
.
2017
;
9
(
4
):
539
71
.
33.
AlashkarAlhamwe
B
,
Alhamdan
F
,
Ruhl
A
,
Potaczek
DP
,
Renz
H
.
The role of epigenetics in allergy and asthma development
.
Curr Opin Allergy Clin Immunol
.
2020
;
20
(
1
):
48
55
.
34.
Potaczek
DP
,
AlashkarAlhamwe
B
,
Miethe
S
,
Garn
H
.
Epigenetic mechanisms in allergy development and prevention
.
Handb Exp Pharmacol
.
2022
;
268
:
331
57
.
35.
Sengupta
S
,
Pattanaik
KP
,
Mishra
S
,
Sonawane
A
.
Epigenetic orchestration of host immune defences by Mycobacterium tuberculosis
.
Microbiol Res
.
2023
;
273
:
127400
.
36.
Acevedo
N
,
AlashkarAlhamwe
B
,
Caraballo
L
,
Ding
M
,
Ferrante
A
,
Garn
H
, et al
.
Perinatal and early-life nutrition, epigenetics, and allergy
.
Nutrients
.
2021
;
13
(
3
):
724
.
37.
Hu
M
,
AlashkarAlhamwe
B
,
Santner-Nanan
B
,
Miethe
S
,
Harb
H
,
Renz
H
, et al
.
Short-chain fatty acids augment differentiation and function of human induced regulatory T cells
.
Int J Mol Sci
.
2022
;
23
(
10
):
5740
.