Demographic characteristics the of study population
The connection between TNFAIP8 SNPs and OC risk was explored in Heilongjiang Province of China. The basic information of all individuals was summarized in Table 1. The average ages of cases and controls were 53.24±10.54 and 54.32±9.58 years, respectively. Furthermore, no significant difference was observed between these two groups (P = 0.261). Also, there was no significant difference of body mass index (BMI) between two gourps (P = 0.814). In addition, there were no significant differences between the cases and controls in Parity, complication and smoking history (P > 0.05). However, positive significance (P = 0.023) between the case and control groups was presented in family cancer history (ovarian cancer).
The relationship of TNFAIP8 polymorphism with OC risk
In this case-control study, three SNPs (rs11064, rs1045241, and rs1045242) which are located in the 3′ UTR, which is a binding site for the regulation of gene expression by microRNAs (miRNAs) in TNFAIP8 gene were selected and analyzed [21]. The genotype frequencies of each SNP conformed to the Hardy-Weinberg equilibrium among controls (P > 0.05 for all). Displayed in Table 2, TNFAIP8 rs11064 A-allele (COR: 0.690, 95%CI: 0.491-0.971, P = 0.033 and AOR: 0.709, 95%CI: 0.504-0.997, P = 0.048) and rs1045242 G-allele (COR: 1.619, 95%CI: 1.129-2.323, P = 0.009 and AOR: 1.628, 95%CI: 1.132-2.342, P = 0.009) are risk factors for OC. However, the allele of TNFAIP8 rs1045241 had no effect on the risk of OC (P > 0.05).
For further examination, we conducted the correlation between the genotypes of SNPs and OC risk by logistic regression analysis under the codominant, dominant, and recessive models (Table 3). Our results showed that rs11064 was significantly associated with increased OC susceptibility in codominant model (GG/AA, COR: 0.200, 95%CI: 0.057-0.706, P = 0.012 and AOR: 0.205, 95%CI: 0.058-0.726, P = 0.014) and recessive model (GG/AA+AG, COR: 0.209, 95%CI: 0.060-0.731, P = .014 and AOR: 0.212, 95%CI: 0.060-0.744, P = 0.016). Also, we demonstrated that rs1045242 was related to a higher risk of OC under codominant model (AG/AA, COR: 1.670, 95%CI: 1.091-2.558, P = 0.018 and AOR: 1.703, 95%CI: 1.108-2.618, P = 0.015) and dominant model (AG+GG/AA, COR: 1.736, 95%CI: 1.149-2.623, P = 0.009 and AOR: 1.761, 95%CI: 1.162-2.670, P = 0.008). However, there was no significant association between TNFAIP8 rs1045241 and OC risk.
Stratification analysis between TNFAIP8 SNPs and OC risk based on age, smoking history, complication, and family history
Aiming to deeply analyze the relationships of TNFAIP8 genotypes with OC susceptibility, we divided age into ≤ 54 years old and > 54 years old, whether smoking, whether complication (patients with diabetes and cardio-cerebrovascular disease), and whether there is family history of OC. It revealed that rs1045242 mutation (AG+GG/AA) would significantly increase risk of OC (OR: 2.048, 95%CI: 1.116-3.757, P = 0.021) at age ≤ 54 years old (Supplementary Table 1). In subjects with no smoking history, the rs11064 mutation (GG) was a protective factor for OC (OR: 0.164, 95%CI: 0.036-0.742, P = 0.019). On the contrary, the rs1045242 mutation (AG+GG) was a risk factor for OC (OR: 2.670, 95%CI: 1.141-6.247, P = 0.024) in subjects with smoking history (Supplementary Table 2). As showed in Supplementary Table 3 and Table 4, the rs1045242 mutation (AG+GG) was a risk factor for OC in subjects with no complication (OR: 1.829, 95%CI: 1.109-3.018, P = 0.018) and no family history of OC (OR: 1.746, 95%CI: 1.150-2.650, P = 0.009). The rs11064 GG genotype was a protective factor for OC in subjects with no family history of OC (OR: 0.205, 95%CI: 0.058-0.724, P = 0.014).
Correlation between TNFAIP8 SNPs and clinicopathological characteristics of OC
The correlation between three TNFAIP8 genotypes and the clinicopathologic data of OC is illustrated in Table 4. It was found that rs1045242 was related to an increased risk in OC patients with III/IV FIGO stage (P = 0.040 and P = 0.013, respectively) and presence of recurrence (P = 0.043 and P = 0.034, respectively) both under codominant and dominant models. For rs1045242, it was confirmed that AG+GG genotype was significantly associated with an increased OC risk in residual tumor more than 1 cm (P = 0.019). rs1045241 SNP was strongly significant associated with FIGO stage (P = 0.025) and residual tumor (P = 0.033) under dominant model. Furthermore, rs11064 SNP was observed to be positively related to FIGO stage both under codominant (P = 0.024) and dominant (P = 0.006) models.