Association of CTLA‐4, TNF alpha and IL 10 polymorphisms with susceptibility to hepatocellular carcinoma

Our aim was to evaluate the association of genetic polymorphisms of immunoregulatory molecules with susceptibility to hepatocellular carcinoma (HCC). The polymorphisms in CTLA‐4 (−318 T/C, CT60 G/A), TNF (−238 G/A, −308 G/A) and IL10 (−592 C/A, −819 C/T) were genotyped by PCR and DNA sequencing. The functional relevance of the polymorphisms was examined by ELISAs, in vitro lymphocyte proliferation assay and cytotoxic assay. The CTLA‐4 −318 TC/TT, CTLA‐4 CT60 GG, IL10 −592 CA and −819 CT/TT variants, CTLA‐4 −318 T and IL 10 −819 T alleles were positively associated with HCC risk (P < .05). While TNF −238 AA variant, TNF −238 A allele were associated with decreased risk of HCC (P < .05). Furthermore, combinations of CTLA‐4 −318 TC/TT and TNF −238 GG/GA; CTLA‐4 −318 TC/TT and IL 10 −819 CC; CTLA‐4 −318 CC and IL 10 −819 CT/TT in patients with HCC were statistically significant (P < .05). Peripheral blood mononuclear cells (PBMCs) carrying −318 TC/TT genotypes exhibited significantly lower proliferation rates, decreased IL‐2, IL‐4 levels, fewer cytolytic activities and elevated TGF‐β levels. For IL 10 −819 C/T, the CC genotype was significantly associated with higher proliferation rate, decreased TGF‐β, IL‐10 levels and higher cytolytic activities (P < .05). For TNF −238 G/A, the AA genotype only had association with serum IL‐2, IL‐4 (P < .05). In addition, we also found that CTLA‐4 −318 T/C, IL‐10 −819 T/C variants, combinations of CTLA‐4 −318 CC with IL 10 −819 CT or TT, CTLA‐4 −318 TC or TT with IL 10 −819 CT or TT were associated with the severity of HCC. These findings suggest that CTLA‐4 −318 TC/TT and IL 10 −819 CT/TT could promote the pathogenesis of HCC, which might be related with down‐regulation of Th1/Th2‐type cytokines and/or up‐regulation of Th3‐type cytokines.

for autoimmune diseases and certain types of cancer 3,6 ; however, the mechanism by which polymorphism of CTLA-4 acts to inhibit cancer remains unclear. Recently, several studies focused on the association of CTLA-4 single nucleotide polymorphisms (SNPs) with susceptibility to various types of cancer, such as gastric cancer, cervical cancer, colorectal cancer and lung cancer. [7][8][9][10] Also, a few case-control studies focused on the association of CTLA-4 SNPs with susceptibility to HCC. 11,12 However, due to the small sample size and the limited amount of data, the relationship between CTLA-4 SNPs and risk of HCC was not fully understood.
There are various factors associated with the development of HCC, and the host immune response has been highlighted as a genetic biomarker for the disease with the production of several cytokines. 13 Cytokines released from a variety of activated T cells play key roles in the regulation of immune responses, and SNPs in the regulatory sequences of genes are presumed to be associated with the production of cytokines. 14 Therefore, polymorphisms in genes coding cytokines can also influence the production and function of these proteins, susceptibility or disease progression. Among them, TNF-α and IL-10 are of particular interest, as the genes for these cytokines are located in the central major histocompatibility complex (MHC).
Tumour necrosis factor alpha (TNF-α) is one of the major pro-inflammatory cytokines involved in tumour proliferation, invasion and metastasis. 15 Furthermore, it was reported that the expression of TNF-α is regulated at the transcriptional level and various SNPs in TNF have been associated with TNF-α level. In particular, the polymorphisms at position −238 and −308 in the promoter region have been commonly studied. 16 As an anti-inflammatory cytokine, IL-10 is produced by Th2 lymphocytes, monocytes and B lymphocytes. It inhibits macrophage-dependent cytokine synthesis by Th1 cells and thus regulates balance between cellular and humoral immune responses. 17 The IL 10 presents its transcript governed by SNPs in the promoter region of the gene. Numerous polymorphic sites have been identified in the IL 10 gene, including −592 C/G and −819 C/T, which appear to be mostly related to SNPs in the promoter and might influence the circulating IL-10 levels. 18 Host immune responses are known determinants of cancer susceptibility. Cancer-bearing hosts have different antitumour responses, whereas the underlying mechanisms are not fully understood. In the current study, we hypothesized that SNPs in CTLA 4, TNF and IL 10 can interfere with T cell function, which might be a genetic susceptibility factor for HCC. To test this hypothesis, we investigate the possible associations of CTLA-4, TNF and IL 10 genotypes with the susceptibility to HCC. We also functionally characterized the effects of these SNPs by analysing T cell proliferation, cytotoxicity and cytokine production using PBMCs with diverse genotypes. Additionally, the synergistic effects of these genotypes were also determined.

| DNA extraction and genotyping
Genomic DNA of all subjects was extracted from peripheral leucocytes using QIAamp DNA mini Kit (Qiagen GmbH, Hilden, Germany). DNA purity and concentration were determined by an ultraviolet spectrophotometer and then kept at −20°C refrigerator.
Genotyping for polymorphisms in CTLA-4 (−318 T/C, CT60 G/A), TNF (−238 G/A, −308 G/A) and IL 10 (−592 C/A, −819 C/T) was carried out using PCR, with a 25 μL reaction solution containing 100 ng genomic DNA, 1 × PCR buffer, 2 mmol/L MgCl 2 , 200 μmol/L dNTPs, 2 unit DNA polymerase (Takara, Japan) and 1 μmol/L of specific primer mix (Shenggong, China). The reaction conditions were as follows: initial denaturation for 2 minutes at 95°C, followed by 40 cycles of denaturing at 95°C for 30 seconds, annealing at 59°C for 30 seconds, extension at 72°C for 30 seconds and a final extension at 72°C for 5 minutes. The primer pairs employed were listed in Table 1. SNP genotyping was performed using a 3730 DNA sequencer (Applied Biosystems).

| Peripheral blood mononuclear cells isolation and culture
Peripheral blood mononuclear cells (PBMCs) were separated from heparinized venous blood using Ficoll-Hypaque

| Cytotoxic assay
The cytotoxic assays were done by using a non-radioactive cytotoxicity assay kit (Promega) according to the manufacturer's instructions and as described previously. 19 PBMCs treated with PHA were used as effector cells, and HepG2 liver cancer cells were used as target cells.

| Statistical analysis
The genotype frequencies in patients and controls were tested for conformity with Hardy-Weinberg equilibrium using chisquared test. The homozygous form of the most common allele was used as the reference variant. For the sake of easy calculations, the frequencies of genotypes and alleles in Tables 3, 4, 5 and 6 were calculated in double, and the frequencies between the two groups were compared using chisquared test. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to evaluate the relative disease risk. Student's t test or one-way analysis of variance (ANOVA) was used to compare differences in lymphocyte proliferation, cytokine concentrations and cytolytic activity between the different genotypes. All statistical analyses and graphics were conducted using SPSS 22.0 software (SPSS Inc) and GraphPad Prism 5.0 (GraphPad Inc). P value < .05 was considered statistically significant. Table 2 summarizes the main characteristics of the two groups. There was statistical difference of distribution on alcohol consuming between patients and controls (P < .05). No significant differences were observed between the two groups in regard to age, sex distributions and smoking status (P = .518, P = .468 and P = .059). .145

Note:
All values are given in pg/mL.
For −318 T/C, the difference between TC + TT and CC was calculated. For −238 G/A, the difference between GG and GA + AA was calculated. For −819 T/C, the difference between CC and CT + TT was calculated.

| Prevalence of CTLA-4 polymorphisms in patients with HCC
CTLA-4 −318 T/C, CT60 G/A allele and genotype frequencies in patients with HCC and controls are shown in Table 3

| Prevalence of TNF and IL 10 polymorphisms in patients with HCC
Distributions of −238 G/A, −308 G/A polymorphisms in TNF and −592 C/A, −819 T/C polymorphisms in IL10 were in Hardy-Weinberg equilibrium in case and control groups (P > .05). Tables 4 and 5 show the association analysis of the above genotypes and HCC risk as ORs. For −238 G/A, we found that in comparison with healthy controls, patients with HCC had significantly lower AA genotype with an OR of 0.483 (95% CI: 0.279-0.836), P < .05 for the codominant model. This result was confirmed in the dominant model GA + AA with an OR of 0.724 (95% CI: 0.557-0.941), P = .016, and recessive model AA with an OR of 0.508 (95% CI: 0.295-0.877), P = .013. For alleles, a strong negative association with HCC was observed for the A allele, with an OR of 0.699 (95% CI: 0.509-0.961), P = .027. Nevertheless, no significant associations were found between genotypes or allele frequencies of −308 G/A and HCC risk (P > .05). Table 5 shows a marginal positive association between the genotype CA in −592 C/A and HCC, with an OR of 1.

| Prevalence of CTLA-4 and TNF or IL 10 polymorphism combinations in patients with HCC
In order to investigate whether CTLA-4 and TNF or IL10 polymorphisms could act together to have an additive effect, the associations of combinations of these genotypes with HCC were analysed (Table 6). Statistically significant models with the strongest positive odds ratios from Tables 3, 4

| Association of CTLA-4, TNF and IL 10 polymorphisms with serum cytokine levels
The serum IL-2, IL-4, TGF-β, IL-10 and TNF-α levels in patients with HCC and controls were analysed using the same model mentioned above. As shown in Table 7, the CTLA-4 −318 T/C, TNF −238 G/A and IL 10 −819 T/C genotypes had no associations with IL-2, IL-4, TGF-β, IL-10 and TNF-α levels in the healthy controls (P > .05). However, for patients with HCC, compared with CC genotype, the TC and TT genotypes in −318 were significantly associated with a decrease in serum IL-2 (P < .01) and IL-4 (P = .027), and an increase in serum TGF-β (P = .048) levels. For TNF −238 G/A, the GG genotype was significantly associated with a decreased IL-2 level (P = .026) and an increased IL-4 level (P = .023) compared with GA and AA genotypes. Similarly, the CT and TT genotypes in IL 10 −819 were significantly associated with increase in TGF-β (P < .01), IL-10 (P < .01) levels compared with CC genotype. Together, we can conclude from these results that CTLA-4 −318 TC/TT, TNF −238 GG and IL 10 −819 CT/TT genotypes are associated with HCC, possibly through modulating the shift from Th1/ Th2 to Th3-type cytokines.

| Effects of CTLA-4, TNF and IL 10 polymorphisms on T cell proliferation and cytokine production
MTT assay was performed to evaluate the proliferation of T lymphocytes (Figure 1), while cytokine levels (IL-2, IL-4, TGF-β, IL-10 and TNF-α) in the cell culture supernatants were detected by ELISAs (Figure 2). MTT assays showed that PBMCs carrying CTLA-4 −318 TC/TT genotypes had significantly lower cell proliferation rates after stimulation with PHA (P < .05), while PBMCs with IL 10 −819 CC had significantly higher rate of proliferation rate. Analysis of IL-2, IL-4, TGF-β, IL-10 and TNF-α levels in cell culture supernatant showed that PBMCs with CTLA-4 −318 TC/TT genotypes secreted significantly lower IL-2, IL-4 but higher TGF-β (P < .05). Similarly, for IL 10 −819 T/C, the CC genotype had association with lower levels of TGF-β, IL-10 (P < .05). These results mirror what is seen in the serum and clearly suggest that CTLA-4 −318 TC/TT and IL 10 −819 CC genotypes have significant effects on cell proliferation and/or cytokine production. F I G U R E 2 Differential levels of cytokine production in PBMCs from healthy individuals carrying different CTLA-4 −318 T/C, TNF −238 G/A and IL 10 −819 C/T genotypes. PBMCs extracted from healthy individuals were stimulated with PHA. Levels of IL-2, IL-4, TGF-β, IL-10 and TNF-α were detected in the cell culture supernatant. * TT + TC vs CC, P < .05. # CC vs CT + TT, P < .05

CTLA-4, TNF and IL 10 genotypes on liver cancer cells
For cytotoxic assay, the effector and target cells were PBMCs carrying different CTLA-4, TNF, IL 10 genotypes and HepG2 cells, respectively. As shown in Figure 3, a significant decrease in tumour lysate of HepG2 cancer cells was observed in PBMCs with CTLA-4 −318 TC/TT compared with −318 CC genotype (P < .05). Similarly, for IL 10 −819 T/C, PBMCs with −819 CC genotype exhibited a significant increase in tumour lysate of HepG2 cancer cells compared with PBMCs with −819 CT/TT (P < .05). However, no significant change was found between the genotypes in TNF −238.

| DISCUSSION
The multifactorial natures of HCC are fully recognized, but genetic factors are considered strong determinants of these diseases, which has encouraged scientists to search for the genes responsible. 20 Many studies have confirmed a critical role of T cell response in the antitumour effect, and therefore, genes encoding molecules involved in T cell response may potentially affect the carcinogenesis of cancers. The present study aims to assess the associations of CTLA-4, TNF and IL 10 polymorphisms with susceptibility to HCC.
CTLA-4 is a structural homolog for CD28 and expressed on activated T cells. 21 CTLA-4 can function as a negative costimulatory regulator of T cell activation and modulate immune responses of the body. 22 CTLA-4 is polymorphic, and several important SNPs in CTLA-4 have been reported to be associated with a range of different malignancies. [23][24][25] Although a large number of studies suggest that +49 G/A polymorphism in CTLA-4 is significantly associated with HCC, only few studies have been conducted to investigate the associations between −318 T/C, CT60 G/A variants and HCC risk. Thus, we performed a comprehensive case-control study to explore the association of CTLA-4 −318 T/C and CT60 G/A polymorphisms with HCC risk in a Chinese Han population. The results showed that patients with HCC had obviously higher frequencies of −318 TC/TT genotypes and −318 T alleles, which implied that −318 TC/TT in CTLA-4 were positively associated with HCC risk. To our knowledge, this is the first study about the association of CTLA-4 −318 T/C and CT60 G/A SNPs with susceptibility to HCC, which may contribute to our understanding of the role of CTLA-4 SNPs in the pathogenesis of HCC.
TNF-α is a powerful pro-inflammatory factor involved in several biological processes such as macrophage activation, recruitment of inflammatory cells and amplification of proinflammatory cytokines. 26 Polymorphisms in the promoter of TNF may increase its transcription level and thus cytokine production. 27 Many studies have reported the associations of −238 G/A and/or −308 G/A polymorphisms in TNF with susceptibility to HCC, but the results are conflicting. [28][29][30] Our findings indicated that TNF −238 GG genotype was associated with increased risk of HCC. These findings are consistent with some previous reports but not with others. [28][29][30] Polymorphisms in IL10, another immunomodulatory gene, have also been implicated in HCC. Two polymorphisms (−592 C/A and −819 C/T) in IL 10 were examined to find possible associations with HCC. Our results showed that −592 CA genotype was marginally associated with HCC. Also, we found that −819 CT/TT genotypes were significantly associated with HCC. This is consistent with previous studies, but contradicts that reported in other works. 31,32 These inconsistent results may be partly due to small sample sizes, racial and ethnic differences, and publication bias.
Next, in order to explore whether the interaction of the above SNPs had an additive effect, association of combinations of CTLA-4 and TNF or IL10 genotypes with HCC risk was analysed. We found that combinations of CTLA-4 −318 TC or TT and TNF −238 GG or GA; CTLA-4 −318 TC or TT and IL10 −819 CC; CTLA-4 −318 CC and IL10 −819 CT or TT were strongly associated with susceptibility to HCC. The increase in the odds ratios compared with the single SNPs alone suggests that there may be an additive effect of having high-risk genotypes at multiple loci.
Effective antitumour responses require CD4 + Th cells and CD8 + cytotoxic T lymphocytes (CTLs). 33 To further characterize the function of these polymorphisms, the associations of polymorphisms in CTLA-4 −318 T/C, TNF −238 G/A and IL 10 −819 C/T with T cell activation and cytokine secretion were investigated. Our analyses showed that PBMCs carrying CTLA-4 −318 TC/TT and IL 10 −819 CT/TT genotypes showed lower proliferation rates and cytolytic activities. The cytokine profile in vitro was consistent with serum cytokine profiles in patients with HCC. Furthermore, CTLA-4 −318 TC/TT had significant associations with lower IL-2 (Th1-type cytokine), IL-4 (Th2-type cytokine) and higher TGF-β levels (Th3-type cytokine), and IL 10 −819 CT/TT had significant associations with higher TGF-β and IL-10 (Th2-type cytokine) levels. This shift from Th1/Th2 to Th3 might be responsible for facilitating tumour progression by subverting various cellular immune surveillance mechanisms. Moreover, the present study provides evidence that CTLA-4 −318 T/C and IL 10 −819 T/C variants were associated with the severity of HCC. Therefore, we may reasonably expect that individuals who carry the CTLA-4 −318 TC/TT and IL 10 −819 CT/TT genotypes might negatively associated with T A B L E 8 Genotype frequencies of CTLA-4, TNF and IL 10 in HCC patients with different clinicopathologic stages of disease T cell proliferation and function, which might be likely the underlying mechanisms conferring HCC susceptibility. Cumulatively, these data demonstrate that CTLA-4 −318 T/C and IL 10 −819 T/C can affect susceptibility to HCC by altering the immune status of an individual. Additional studies using a larger sample size and haplotype analysis with other SNPs may lead to better understanding of these variants in the pathogenesis of HCC.