Investigation of BTLA tagging variants with risk of esophagogastric junction adenocarcinoma

The morbidity of esophagogastric junction adenocarcinoma (EGJA) is promoting rapidly, both in developing and developed contries [1–3]. EGJA comprises a vital portion esophageal and gastric cancer, with an increasing ratio. It is reported that EGJA is a common fatal tumor in China. EGJA is regarded as an entity with a specific clinical feature and molecular profile. The potential protective factor or a real cause of EGJA is unclear. Thus, an understanding of the potential risk factors influencing the development EGJA biology may be helpful to diagnosis and prognostic assessment for the supervision of EGJA patients.

During the activation of T lymphocytes, they can express some receptors for receiving various signals. B- and T-lymphocyte attenuator (BTLA), also named CD272, is a most recently identified and studied member of the immune globulin (Ig) superfamily [4–7]. BTLA is a glycoprotein and it contains two tyrosine-based inhibitory motifs [8]. During activation, BTLA is not expressed on T helper type 2 (Th2) cells, but Th1 cells. The expression of BTLA on T cells participates in negative regulation of T cell and then leads to an decreased T-lymphocytes proliferation [9]. Recently, many investigations have focused on the relationship of BTLA with inflammation, autoimmune disease and cancer. Shi et al. reported that BTLA-herpes virus entry mediator (HVEM) checkpoint axis might be implicated in the regulation of inflammation in liver [10]. A previous study indicated that the up-regulation of BTLA gene expression and soluble BTLA (sBTLA) was validated in thymoma-associated myasthenia gravis [11]. A prognostic investigation showed that the levels of immune checkpoints sBTLA could be considered as a biomarker for unresectable pancreatic adenocarcinoma cases with a poor survival [12]. A functional study identified that IFN-γ level in circulating T-lymphocytes could be promoted by inhibiting BTLA/HVEM pathway [13]. Additionally, Feng et al. [14] and Lan et al. [15] reported that the level of BTLA expression in gastric carcinoma (GC) might be a useful biomarker for the evalution of GC prognosis.

Single-nucleotide polymorphisms (SNPs) in BTLA gene are likely to affect the role of BTLA protein. Some studies have kept a watchful eye on the correlation of BTLA variants with the development of cancer [16–18]. Fu et al. reported that the frequencies of BTLA rs1844089 and rs2705535 SNPs may alter the risk of breast cancer [17]. In Polish population, it was found that BTLA rs1982809 G>A, a 3′-UTR SNP, might be a low-penetrating risk factor for the development of renal cell carcinoma [18]. In addition, another study indicated that BTLA rs1982809 G and rs2705511 C alleles were more frequent in patients with chronic lymphocytic leukemia compared with healthy controls [16]. In view of the vital role in cancer development and progress, we supposed that BTLA SNPs might be correlated with EGJA susceptibility. Here, BTLA tagging SNPs (rs16859629 T>C, rs1982809 G>A, rs2171513 G>A and rs3112270 A>G) were selected. The aim of the present study was to identify the association of BTLA tagging SNPs with risk of EGJA.

The present study involved 1236 new incident EGJA patients and 1540 cancer-free controls. Among these patients, 393 cases patients diagnosed with EGJA and treated at two affiliated hospitals of Fujian Medical University [Union Hospital (Fuzhou, China) and Fujian Cancer Hospital (Fuzhou, China)] from January 2014 to June 2018. In addition, 843 patients with EGJA were from Jiangsu University People’s Hospital (Zhenjiang, China) from January 2008 to June 2018. Siewert type was used in our study [19]. Here, all EGJA cases included were Siewert type II (their centre within 1 cm proximal and 2-cm distal of the anatomical cardia). All included EGJA cases were diagnosed at the first time with histopathological test. For EGJA cases, the major included criteria were: (a) individuals who did not have a history of other cancers, (b) without any immunological diseases and (c) EGJA patients were not treated wtih any chemotherapy and/or radiotherapy before the enrolment. We recruited 1540 cancer-free subjects as controls matching to the EGJA patients by sex, year of birth (±5 year) and ethnicity (Eastern Chinese Han nationality). They were from the hospitals mentioned above for regular health examination. The major included criteria for controls were: (a) cancer-free individuals, (b) without any immunological diseases, (c) sex and age matching to EGJA cases and (d) Han nationality who living in Eastern China. Each patitcipant signed a consent form. The experimental protocol was authorized by the ethics committees of the Jiangsu University.

The tagging SNPs of BTLA [from 112458030 to 112504757 in chromosome 3 (extending 5 Kb, upstream and downstream, respectively)] were structured and collected from Chinese populations via Genome Variation Server data. The criteria of tagging SNPs selection were described in our previous studies [20,21].

Genomic DNA was extracted from the colletced blood samples with the Promega DNA Kit (Promega, Madison, U.S.A.), according to the explanatory memorandum. A 2-μl DNA was droped in NanoDrop ND-1000 spectrophotometer (Wilmington, U.S.A.) to evaluate concentration and purity of DNA sample.

The genotypes of BTLA rs16859629, rs1982809, rs2171513 and rs3112270 SNPs were analyzed using a SNPscan Kit (Genesky Biotechnologies Inc., Shanghai, China) as described previously [22–24]. PCR process was conducted in a 20-μl mixture volume in 96-well plates. ABI 3730xl DNA Analyzer was used to identify the genotype. The data of the sequencing were read by GeneMapper 4.1 (AppliedBiosystems, U.S.A.). One hundred and eleven DNA specimens were randomly chosen for repeat genotyping by another person in a blind fashion, and the obtained variants were concordant.

For each locus in BTLA gene, an online χ² test was used to assess the Hardy–Weinberg equilibrium (HWE) [25]. The Student t test was performed to deal with continuous variables of demographic characteristics between two groups. And χ² test was harnessed to handle the categorical variables (e.g., age, sex, cigarette using and alcohol consumption) and variant distributions of BTLA SNPs between two groups. The haplotypes of BTLA gene were evaluated by SHESIS software [26]. Odds ratios (ORs) and 95% confidence intervals (CIs) were used to evaluate the strength of the correlation of BTLA SNPs with the risk of EGJA. Multiple logistic regression analysis was harnessed to check the distribution of BTLA rs16859629, rs1982809, rs2171513 and rs3112270 genotypes between two groups. Subgroup analyses between the BTLA variants and characteristic variables were also conducted. The adjusted P values, ORs and 95% CIs were calculated by adjustment for age, sex, cigarette using and drinking. A P < 0.05 (two-way tests) was defined as significance in all statistical tests. All statistical analyses described previously were performed in SAS 9.4 software (SAS Institute Inc., Cary, NC, U.S.A.). Using PS software (http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize), the power value (α = 0.05) was calculated [27,28]. We also used the false-positive report probability (FPRP) to determine the significant findings [29].

Table 1 summarizes age, sex, cigarette using and alcohol consumption in two groups. EGJA patients had a mean age of 64.28 ± 8.64 years. The age and sex ratio was not significant between two groups (P = 0.408 and P = 0.485, respectively). The success rate of BTLA rs16859629, rs1982809, rs2171513 and rs3112270 genotyping was of high quality (>97%) (Table 2). We pressented the data of minor allele frequency (MAF) in Table 2. In control group, the frequencies of genotype distribution met HWE (Table 2).

The genotype distributions and frequencies of BTLA rs16859629, rs1982809, rs2171513 and rs3112270 genotypes are presented in Table 3. In a single SNP analysis, the BTLA rs2171513 G>A genotype frequencies were 62.83% (GG), 32.67% (GA) and 4.50% (AA) in EGJA patients and 63.70% (GG), 32.27% (GA) and 4.03% (AA) in the cancer-free controls. When the BTLA rs2171513 GG genotype was defined as the reference, the BTLA rs2171513 GA genotype was not correlated with the susceptibility for EGJA (GA versus GG: adjusted OR = 1.04, 95% CI: 0.88–1.22, P = 0.668); the BTLA rs2171513 AA genotype was not correlated with the susceptibility for EGJA (AA versus GG: adjusted OR = 1.23, 95% CI: 0.83–1.81, P = 0.302). In addition, the BTLA rs2171513 GA/AA genotypes did not conferred the risk to EGJA in the dominant model (GA/AA versus GG: adjusted OR = 1.06, 95% CI: 0.90–1.24, P = 0.497). In the recessive genetic compared model, when the BTLA rs2171513 GG/GA genotypes were defined as a reference, the BTLA rs2171513 AA genotype was not correlated with susceptibility for EGJA (AA versus GG/GA: adjusted OR = 1.21, 95% CI: 0.83–1.78, P = 0.327) (Table 3). No association was also found between BTLA rs3112270 A>G, rs1982809 G>A and rs16859629 T>C SNPs and the susceptibility of EGJA (Table 3).

Table 4 presents the variant frequencies of BTLA rs1982809 SNP in stratification analysis. When we conducted an adjustment for gender, age and alcohol consumption, we identified that the BTLA rs1982809 G>A was associated with an increased susceptibility of EGJA for ever smokers (AA versus GG: adjusted OR = 2.09, 95% CI 1.08–4.07, P = 0.030; and AA versus GA/GG: adjusted OR = 1.99, 95% CI 1.04–3.82, P = 0.039). We found that there was no significant association between BTLA rs1982809 G>A SNP and the risk of EGJA in other subgroups.

No association was found between the BTLA rs2171513 G>A, rs3112270 A>G and rs16859629 T>C SNPs and the susceptibility of EGJA in subgroup analyses (data was not shown).

Using haplotype constructing software mentioned above [26], we observed 12 BTLA gene haplotypes. We identified that TAAG haplotype with the order of BTLA rs16859629, rs1982809, rs2171513 and rs3112270 SNPs might increase the susceptibility of EGJA (OR = 3.07, 95% CI = 1.41–6.71; P = 0.003). However, other observed BTLA gene haplotypes did not alter the susceptibility of EGJA (Table 5).

Using PS software (http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize), the power value (α = 0.05) was calculated [27,28]. For BTLA rs1982809 G>A SNP, the power value was 0.631 in AA versus GG genetic model and 0.589 in AA versus GG/GA genetic model among ever smokers. In haplotype comparison, T_rs16859629A_rs1982809A_rs2171513G_rs3112270 haplotype could increase the susceptibility of EGJA (power value, 0.830).

The incidence of EGJA is increasing in both the East and Western countries. It is reported that altered lifestyle and lower chronic Helicobacter pylori infection may result in an increasing incidence of EGJA [30,31]. The etiology of EGJA may be attribute to gene and environment factors. Recent evidence suggested that the variants of immune and inflammatory response related genes could alter the risk of cancer [21,32–35]. Considering an important role of BTLA gene in immune, we chose BTLA tagging SNPs (rs16859629, rs1982809, rs2171513 and rs3112270) and explored their effects on the development of EGJA. Here, we identified that BTLA TAAG haplotype with the order of rs16859629, rs1982809, rs2171513 and rs3112270 SNPs might be associated with the development of EGJA.

BTLA rs1982809 G>A SNP locates in 3′-UTR, which could participate in post-transcriptional control. Recently, studies have been conducted to identify a potential effect of BTLA rs1982809 locus on the development of malignancy. BTLA rs1982809 polymorphism, a 3′-UTR SNP, was found to be associated with the development of renal cell carcinoma in Polish populations [18]. Another case–control study also found that BTLA rs1982809 polymorphism were associated with chronic lymphocytic leukemia [16]. Subsequently, in the same study, the funcional investigation demonstrated that the presence of BTLA rs1982809 G allele was correlated with lower expression of BTLA mRNA in lymphocyte as compared with rs1982809 A allele [16]. In the present study, we first studied the relationship between BTLA rs1982809 locus and cancer risk in Asians. We found this SNP might not alter the overall EGJA risk. However, BTLA rs1982809 locus was identified as a risk factor to EGJA in smoking subgroup, which was similar to the previous reports [16,18]. The results suggested that the role of BTLA rs1982809 G>A polymorphism may be influenced by environmental factors. However, the subjects included in smoking subgroup were related small, these findings may be underpowered. In the future, more case–control studies should be conducted to evaluate whether BTLA rs1982809 G>A polymorphism might inhibit the function of B and T cells and influence the susceptibility of cancer.

In the present case–control study, the BTLA haplotypes were also constructed. We found BTLA T_rs16859629A_rs1982809A_rs2171513G_rs3112270 haplotype might influence the risk of EGJA. However, this rare BTLA haplotypes only altered the susceptibility of a minor fraction of the EGJA patients. We first expolre the association of BTLA haplotypes with cancer risk in Asians. Our findings should be verified in the future studies.

It is necessary to acknowledge the limitations in the present case–control study. First, the present study was designed as hospital-based. Although the frequencies of genotype distribution in BTLA rs16859629, rs1982809, rs2171513 and rs3112270 SNPs met HWE and the MAFs of these selected SNPs in control group were close to the database for Chinese, the bias might have happened. Second, we only included four risk factors (gender, age, smoking and alcohol consumption). And other potential environment factors (e.g. body mass index, intake of vegetable and fruit, education level and economic income) were not considered. Thus, the potential interactions between gene and these environment factors could not addressed. Third, the participants included were related small in some subgroups, the observations may be insufficient evidence to identify a relationship with a definitive power. Fourth, in the present study, the biological functions of BTLA SNPs were not studied. Finally, only four BTLA tagging SNPs (rs16859629, rs1982809, rs2171513 and rs3112270) were selected, which could not fully assess the total hereditary susceptibility in BTLA gene.

To conclude, this investigation suggests that BTLA T_rs16859629A_rs1982809A_rs2171513G_rs3112270 haplotype may increase the susceptibility of EGJA. More studies with multiple environment factors should be carried out to evaluate whether BTLA variants may influence the susceptibility of cancer in the future.

We appreciate all subjects who participated in this study.

All authors contributed significantly to this study. Conceived and designed the experiments: W.T., S.C., Performed the experiments: C.L., J.L., W.T., Analyzed the data: M.K., Contributed reagents/materials/analysis tools: S.C., Wrote the manuscript: M.K., W.T.

The authors declare that there are no competing interests associated with the manuscript.

This study was supported in part by 333 Talent Training Project of Organization Department in Jiangsu Province [grant number BRA2017147] and Young and Middle-aged Talent Training Project of Health Development Planning Commission in Fujian Province [grant number 2016-ZQN-25].

FPRP

false-positive report probability

MAF

minor allele frequency

SNP

single-nucleotide polymorphism