Investigation of ICOS, CD28 and CD80 polymorphisms with the risk of hepatocellular carcinoma: a case–control study in eastern Chinese population

Single nucleotide polymorphisms (SNPs) in immune related gene may influence the susceptibility of cancer. We selected inducible T cell costimulator (ICOS) rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A SNPs and assessed the potential relationship of these SNPs with hepatocellular carcinoma (HCC) risk. A total of 584 HCC cases and 923 healthy controls were recruited. And SNPscan™ genotyping assay was used to obtain the genotypes of ICOS, CD28 and CD80 polymorphisms. We found that ICOS rs10932029 T>C polymorphism significantly increased the risk of HCC (additive model: adjusted odds ratio (OR), 1.59; 95% confidence interval (CI), 1.13–2.22; P=0.007; homozygote model: adjusted OR, 1.12; 95% CI, 0.31–4.03; P=0.867; dominant model: adjusted OR, 1.58; 95% CI, 1.14–2.19; P=0.007 and recessive model: adjusted OR, 1.02; 95% CI, 0.28–3.68; P=0.974). However, ICOS rs4404254 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A SNPs were not associated with the risk of HCC. To evaluate the effects of ICOS rs10932029 T>C on HCC risk according to different age, gender, chronic hepatitis B virus (HBV) infection, tobacco consumption and drinking status, we carried out a stratification analysis. We found that ICOS rs10932029 T>C polymorphism might increase the risk of HCC in male, ≥53 years, never smoking, never drinking and non-chronic HBV infection subgroups. Our study highlights that ICOS rs10932029 T>C polymorphism may confer the susceptibility to HCC. It may be beneficial to explore the relationship between variants in immune related genes and the development of HCC.


Introduction
Hepatocellular carcinoma (HCC) remains a major public health problem worldwide, especially in China. [1] The etiology of HCC is very complicated. It is reported that many environmental factors and unhealthy lifestyles may influence the development and progress of HCC. The potential risk factors contributing to HCC are chronic hepatitis B virus (HBV) infection, aflatoxin, foods preserved by salting, smoking and drinking et al. [2][3][4] Although a growing number of investigations have focused on the etiology of HCC, it is not fully understood. It is suggested that an individual's hereditary factor is also implicated in pathogenesis of HCC. Recently, a number of studies reported that some immune related gene variants might play important roles in the development of HCC [5][6][7].
The process of T-cell activity is very complex. Several transmembrane receptor/ligand pairs cooperate with the T-cell receptor to inhibit or enhance the activity of T cells [8]. The CD28 immunoglobulin superfamily involves the co-inhibitory molecules CTLA-4 and PD-1 as well as the costimulatory molecules inducible T-cell costimulator (ICOS) and CD28. ICOS gene shares homology with human CD28 gene [9]. Recently, it has been identified that ICOS may be up-regulated along with T-lymphocyte activation and then interacts with its ligand (ICOSL). Finally, these processes promote T-lymphocyte proliferation and T helper 2 (Th2) differentiation [10]. Nagase et al. [11] reported that ICOS+Foxp3+ tumor infiltrating lymphocytes were associated with prognosis of gastric cancer and effector regulatory T cell (Treg) correlated with Helicobacter pylori. In addition, a previous study suggested that Treg, especially ICOS + Foxp3 + Treg, might be increased in the HCC microenvironment and predict reduced survival [12]. Based on the vital roles of participation in both T-lymphocyte proliferation and Th2 differentiation, any variant of ICOS gene may influence the development and carcinogenesis of HCC. The ICOS gene is polymorphic, which is located on chromosome 2 in humans. Several ICOS polymorphisms [e.g. rs10932029 T>C, rs4404254 T>C, rs4675379 G>C, rs10932037 C>T (ISV1+173T>C) and rs10183087 A>C] have been established. Among these single nucleotide polymorphisms (SNPs), ICOS rs10932029 T>C and rs4404254 T>C were most widely studied for their susceptibility to various cancers [13][14][15][16]. However, the observed results remain inconsistent rather than conclusive.
CD28 is expressed by most T cells, which competes with CTLA-4 for B7 binding and promotes T-cell proliferation. Recently, some epidemiological studies indicated the potential relationship between CD28 rs3116496 T>C (IVS3 +17T>C) variants and cancer susceptibility. Several publications reported that CD28 rs3116496 TT genotype conferred a low penetrance risk to breast cancer and cervical cancer [17,18]. However, the association between CD28 rs3116496 T>C (IVS3 +17T>C) variants and HCC risk remains unknown.
CD80 (also B7-1) is a protein expressed on activated B cells, dendritic cells and monocytes, which provides a costimulatory signal for T-lymphocyte activation and survival. It is the ligand for CD28 (for auto-regulation and intercellular association). Wu et al. [13] reported that CD80 rs7628626 C>A variants were not associated with the risk of CRC; however, CD80 rs7628626 C>A variants were closely related to regional lymph node metastasis and aggressive tumor progression. Thus, CD80 rs7628626 C>A may be implicated in the development of cancer.
Here, we selected ICOS rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A polymorphisms and carried out a hospital-based case-control study to explore the potential association of ICOS, CD28 and CD80 SNPs with the risk of HCC.

Subjects
A total of 584 cases with HCC and non-cancer controls (n=923) were recruited. HCC cases were enrolled in Fuzong Clinical Medical College and Union Clinical Medical College of Fujian Medical University, Fuzhou, China. Controls were included voluntarily, who participated in a routine medical check-up. All participants were eastern Chinese Han population and unrelated. HCC patients underwent operation, and the pathological findings were confirmed by two experienced pathologists. Controls were fully matched with HCC cases in terms of sex and age. Each participant signed a written informed consent. Risk factors (smoking and drinking) and demographic variables were collected by an interview. Hepatitis B surface antigen (HBsAg) was measured. The criteria of 'smoker' and 'drinker' were described in the previous study [19]. The corresponding data are presented in Table 1. The whole blood was donated by each participant and stored immediately at −80 • C until use. The study protocol was approved by Institutional Review Board at Fujian Medical University.

Selection of SNPs
The polymorphisms of ICOS, CD28 and CD80 gene were selected based on publications, [13][14][15][16][17][18] in which polymorphisms were studied the association with the risk of cancer. Finally, ICOS rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A were selected and studied. The primary information of ICOS rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A SNPs is summarized in Table 2.

DNA extraction and genotyping
Using the DNA Purification Kit (Promega, Madison, U.S.A.), we extracted the genomic DNA from lymphocytes. The obtained DNA was stored at −80 • C until use. The concentration and purity were measured by micro-spectrophotometer. SNPscan™ genotyping assay (Genesky Biotechologies Inc., Shanghai, China), a double ligation and multiplex fluorescence PCR, [20] was used to analyze the variants of ICOS rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A polymorphisms. The success rates of ICOS rs4404254 T>C,  rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A genotyping are shown in Table 2. For quality control, four percent of overall DNA samples were randomly selected and analyzed. And the reproducibility was 100%.

Statistical analysis
Age of participants was described as the mean + − standard deviation (SD   [28].

Baseline characteristics
The information of demographics (age and sex) and selected susceptibility factors (status of chronic HBV infection, smoking and drinking) are summarized in Table 1. As demonstrated in   Table 4). However, in our study, no significant association of ICOS rs4404254 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A variants with the risk of HCC was found.
We used a software to calculate the power value (α = 0.05). For ICOS rs10932029 T>C, the power value was 0.940 in additive model and 0.942 in dominant model.

Association of ICOS rs10932029 T>C polymorphism with HCC in different stratification groups
To evaluate the effects of ICOS rs10932029 T>C on HCC risk according to different age, gender, chronic HBV infection, smoking and drinking status, we carried out a subgroup analysis. Table 5 lists frequencies of ICOS rs10932029 T>C variants in the stratified analysis. After adjustment by logistic regression analysis with these risk factors, we found that ICOS rs10932029 T>C polymorphism might be associated with an increased risk of HCC in some subgroups [

Discussion
HBV is considered as an important risk factor in the development of HCC. However, the incidence of HCC alters materially between similarly chronic HBV infection subjects, suggesting that hereditary factor may contribute to its development. Of late, a number of studies reported that immune related gene variants might be associated with the development of HCC [29][30][31][32][33]. In consideration of the role of ICOS, CD28 and CD80 genes in tumor immunity, we chose ICOS rs4404254 T>C, rs10932029 T>C, CD28 rs3116496 T>C and CD80 rs7628626 C>A SNPs to explore their potential roles in the etiology of HCC. In this case-control study, we found that ICOS rs10932029 T>C polymorphism was associated with the risk of HCC. In the stratified analysis, we found that ICOS rs10932029 T>C polymorphism might be associated with the risk of HCC in male, ≥53 years, never smoking, never drinking and non-chronic HBV infection subgroups. Rs10932029 T>C polymorphism is located on first intron region of ICOS gene, [16] where a number of splicing and regulatory components may interact with it [34]. Recently, several case-control studies have assessed the relationship of ICOS rs10932029 T>C polymorphism with cancer risk [15,16,35,36]; however, the results are controversial. Several epidemiological studies reported that ICOS rs10932029 T>C polymorphism was not associated with the risk of cancer [16,35,36]. However, Xu et al. [15] found that compared with ICOS rs10932029 TT genotype and T allele, the ICOS rs10932029 CT genotype and C allele conferred a significantly increased susceptibility to breast cancer, and this correlation was also identified in a validation cohort. In addition, a previous study indicated that compared with ICOS rs10932029 TT genotype, ICOS rs10932029 CT genotype was associated with a higher rate of disease progression in B-cell chronic lymphocytic leukemia patients [35]. In this case-control study, we found ICOS rs10932029 T>C locus might be associated with an increased risk of HCC, which was similar to the results of the previous study [15]. In the future, the potential role of ICOS rs10932029 T>C on influencing the expression of ICOS in HCC patient blood cells should be assessed to support our findings.
There are some limitations that should be acknowledged. First, all participants were recruited in two local hospitals in Fuzhou City, China. These subjects might not fully represent the eastern Chinese Han population. Second, only four important SNPs in ICOS, CD28 and CD80 genes were selected, which lack sufficient power to assess the total inherited risk in these genes. In the future, a tagging or a fine-mapping study is needed to further explore the potential association between SNPs in ICOS, CD28 and CD80 gene and the development of HCC. Third, in the present study, there is lack of the data about the expression or function of ICOS associated with rs10932029 T>C polymorphism. Finally, for lack of information for co-variates (e.g. body mass index, diet, lifestyle and so on), a more precise assessment was not carried out.
In summary, our study highlights that ICOS rs10932029 T>C polymorphism was associated with the susceptibility of HCC, especially in male, ≥53 years, never smoking, never drinking and non-chronic HBV infection subgroups. Our primary study shows that immune related gene variants may be advantageous for exploring susceptible to HCC.