IL1A polymorphisms is a risk factor for colorectal cancer in Chinese Han population: a case control study

The current research involved a total of 248 CRC patients (143 males and 105 females) and 463 controls (265 males and 198 females) with unrelated Chinese Han ancestries. All CRC cases were diagnosed and confirmed by two independent pathological examinations. As for the eligible CRC cases selection, the individuals without other types of cancer, familial adenomatous polyposis, hereditary nonpolyposis colorectal cancer or undergone previous radiotherapy and chemotherapy were included. With regard to healthy controls, the subjects who had suffered from chronic or severe endocrine and metabolic diseases, digestive system disorders and malignant diseases were excluded from this study. The controls were polyp free individuals at recruitment. Moreover, the individuals with family colorectal disease and cancer history were excluded from control group as well.

By reading the previous publications of IL1A polymorphisms, we selected SNPs which could impact cancer risk, and searched the genetic data of them in dbSNP database (https://​www.​ncbi.​nlm.​nih.​gov/​snp/​) and 1000 Genomes database (http://​www.​internationalgen​ome.​org/​). Only the SNPs whose minor allele frequency (MAF) beyond 5% in Asian populations were included in this study in order to achieve adequate statistical power. Finally, five candidate variations rs3783550, rs3783546, rs2856838, rs1609682, and rs3783521 in IL1A gene were selected for genotyping. Five millilitres venous blood was collected from the subjects when they were recruited in this study. Genomic DNA was extracted from the blood with the Whole Blood Genomic DNA Purification Kit (GoldMag, Xi’an, China). The PCR primers used in multiplexed PCR assay were designed by Agena Bioscience Assay Design Suite V2.0 software and are showed in Additional file 1: Table S1[10]. In order to improve the PCR efficiency and ensure that the amplification primers, extension primers and extension products could be differed by their molecular weight, a tag 10-base sequence (5′-ACGTTGGATG-3′) was added to the 5′ end of each amplification primer. SNP genotyping was carried out with the usage of the Agena Nanodispenser (Agena Bioscience, San Diego, USA) and MassARRAY iPLEX platform (Agena Bioscience, San Diego, USA) [10]. The procedure for SNP genotyping was described as follows. First, a first round PCR was performed to increase and concentrate the target genomic fragments containing the polymorphisms. Second, using the products obtained from the last step as the templates, only one ‘mass-modified’ nucleotide was added to the polymorphic locus at the end of the extension primer fragment in extension reaction. Third, the analyte mixture was dropped to a SpectroCHIP Array by Agena Bioscience Nanodispenser and further processed by the matrix-assisted laser desorption/ionization—time of flight (MALDI-TOF) mass spectrometry. With this method, the analytes could be differed according to their flight time in the system and the nucleotide at the each SNP locus was identified. Finally, the genotyping results were managed and presented in the Agena Bioscience TYPER version 4.0 software [10, 24].

SPSS 18.0 software (SPSS, Chicago, IL, USA) was employed for primary statistical analysis. The heterozygosity for the five SNPs in control cohort was assessed by the Fisher’s exact test to detect their compliance with Hardy-Weinberg equilibrium (HWE). Allele and genotype frequencies of each SNP in CRC cases and healthy controls were calculated and evaluated for the difference with Pearson’s χ² test. Logistic regression analysis with adjustment for age and gender was applied to the odds ratio (OR) and 95% confidence interval (CI) assessment [4]. Adjustment process is usually implemented in the statistical analysis to eliminate the influence of confounding factors, such as age and gender. In all statistical analysis, p ≤ 0.05 was regarded as statistical difference. The associations of the five candidate variants with the CRC risk were examined in codominant, dominant and recessive genetic models using SNPstats online software (https://​www.​snpstats.​net/​start.​htm). Additionally, we used the Haploview software package, version 4.2 (http://​analysis.​bio-x.​cn/​myAnalysis.​php) for linkage disequilibrium (LD) construction and haplotype evaluation [30]. Using the genotyping results, we first constructed the LD pattern among genetic variants. The status of LD was analyzed using two parameters: r² and D’. And D’ = 1 is defined as complete linkage disequilibrium. The allele combination of the variants in one LD block was considered as haplotype. We selected the haplotypes with frequency greater than 5% to evaluate whether the haplotypes were associated with CRC risk. These associations were detected using SHEsis software with available OR (95% CI) and p value.

HaploReg v4.1 database (https://​pubs.​broadinstitute.​org/​mammals/​haploreg/​haploreg.​php) was applied for SNP functional annotation using bioinformatics methods. Expression quantitative trait loci (eQTL) analysis were carried out with the data available on the Genotype-Tissue Expression (GTEx) project (https://​gtexportal.​org/​home/​). The impact of each candidate SNP on IL1A expression was determined in several human tissues and p-values were provided for statistical significance assessment.

The demographic information of the enrolled CRC patients and healthy controls are listed in Additional file 1: Table S2. After applicable statistical evaluation, there was no significant difference in gender observed between patients and controls. The mean age [± standard deviation (SD)] of the patient group was 58.69 ± 12.77 years at diagnosis and of the controls was 50.65 ± 11.79 years at recruitment.

Basic information and allele frequencies of the IL1A polymorphisms are presented in Table 1. None of the five SNPs was excluded because of the deviation from HWE at the 5% p level in the control population. The minor allele of each SNP was assumed a risk factor compared to the wild-type allele. Results of the genetic model analysis are showed in Table 2. The frequency of the homozygous “A/A” genotype of the IL1A SNP rs2856838 differed significantly between patients and controls (8.9% vs. 5.2%, p < 0.05). An association between allele “A/A” of rs2856838 and an increased risk of CRC was observed under the recessive model (adjusted OR = 1.98, 95% CI: 1.05–3.72, p = 0.036) after adjustments for age and gender.

In accordance with the stratified analysis by gender, we found that the IL1A variants rs3783550, rs2856838, rs1609682 and rs3783521 did not show any significant association with CRC risk in males, whereas the statistical significance was detected in females (Table 3). The frequency of homozygous “T/T” genotype of IL1A SNPs rs3783550 and rs1609682 differed significantly between patients and controls in females (rs3783550: 15.2% vs. 8.1%; rs1609682: 15.4% vs. 8.1%, p < 0.05), respectively. Moreover, rs3783550 and rs1609682 were associated with an increased risk of CRC based on the results of the recessive model (rs3783550: OR = 2.17, 95% CI: 1.03–4.60, p = 0.043; rs1609682: OR = 2.20, 95% CI: 1.04–4.65, p = 0.040). The frequency of homozygous “A/A” genotype of rs2856838 differed significantly between patients and controls in females (13.3% vs. 6.1%, p < 0.05) as well. The recessive genetic model also provided the evidence for the role of rs2856838 in CRC susceptibility among Chinese Han women (OR = 2.58, 95% CI: 1.13–5.87, p = 0.024). Additionally, rs3783521 homozygous genotype “G/G” distributed differently between the case and control group (15.2% vs. 8.1%, p < 0.05). Further genetic model analysis revealed that SNP rs3783521 was linked to an increased risk of CRC in females according to the recessive model (OR = 2.13, 95% CI: 1.01–4.49, p = 0.048). However, no statistically significant association was detected between SNP rs3783546 and CRC risk in Chinese Han women.

Finally, five IL1A polymorphisms (rs3783550, rs3783546, rs2856838, rs1609682, rs3783521) were mapped to a 10-kb LD block and displayed three haplotypes with frequencies of more than 5% in our subjects (Table 4). In Fig. 1, the red squares of the IL1A LD block represented significant linkage between the five SNPs. The LD degree is displayed by standard color schemes with bright red for very strong LD (D’ = 1). According to the haplotype analysis, there were no statistically significant differences appearing between patients and controls among any of the IL1A haplotype frequencies.

In order to evaluate the possible function harbored by the five selected variants, we performed a bioinformatics analysis using HaploReg v4.1 database. All the variations were successfully predicted as regulatory SNPs with different biological functions (Additional file 1: Table S3). Furthermore, eQTL analysis uncovered remarkable relationships between these SNPs and IL1A expression in skin, testis or pituitary, which indicated the effects of different genotypes on gene expression (Additional file 1: Table S4).