Risk of cervical squamous cell carcinoma associated with a single nucleotide polymorphism in the RAD18 gene in the Chinese population and its significance as a predictive biomarker.

Cellular DNA is continuously damaged by a range of endogenous and exogenous sources. If not sensed and repaired efficiently, DNA damage leads to genome instability and eventually cancer. The maintenance of genomic integrity is fundamental for cell survival and controlled cell growth. Indeed, most cancer cells exhibit genome instability, often arising from defects in DNA replication defects and faulty repair events. [ 1 , 2 ] Trans-lesion synthesis (TLS) is a DDT mechanism that permits DNA synthesis using templates containing bulky DNA lesions such as ultraviolet (UV)-induced cyclobutane pyrimidine dimers, benzo[a]pyrene dihydrodiol epoxide (BPDE) adducts, 8-oxodG, and many others. [ 3 ] DNA damage tolerance (DDT) pathways are largely coordinated by mono- or polyubiquitination of the replicative clamp proliferating cell nuclear antigen (PCNA). [ 4 , 5 ] Mono-ubiquitination of the homotrimeric DNA polymerase processivity factor PCNA contributes to TLS polymerase recruitment at sites of DNA damage. [ 6 ] Although several E3 ubiquitin ligases control this modification, Rad18 is a central regulator, required for both types of PCNA ubiquitination. [ 6 – 9 ] The RAD18 gene, located on human chromosomes 3p24-p25, plays a crucial role in post-replication repair in various organisms from yeast to humans. Loss of Rad18 increases mutation rates in cells and sensitizes them to DNA damage, illustrating the importance of the DDT pathways in genome stability and cell survival. [ 10 , 11 ] However, overexpression of Rad18 is also deleterious, as it disrupts the proper assembly of some DNA repair foci [ 12 ] and leads to inappropriate PCNA ubiquitination and TLS polymerase recruitment in the absence of DNA damage. [ 13 ] These events could perturb DNA repair or processive DNA replication and increase mutagenesis, consistent with the fact that Rad18 is upregulated in certain cancers. [ 14 – 16 ] Thus, tight control of Rad18 levels and activity promotes genome maintenance. The genetic variation of the Rad18 gene is closely related to tumorigenesis. The RAD18 Arg302Gln polymorphism is associated with the risk of colorectal cancer (CRC). [ 17 ] The frequency of RAD18 Gln302Gln polymorphisms in non-small-cell lung cancer (NSCLC) is much higher than in normal controls. [ 18 ] Another study suggested that, although no base mutation had been found in both NSCLC cancer cell lines and NSCLC cancer tissues, the frequency of SNPs at codon 302 tended to be higher in NSCLC patients compared to healthy volunteers. [ 19 ] It has also been reported that genetic polymorphism of some RAD18 loci is correlated with the prognosis of cancer patients [ 20 ] and the side effects of platinum-chemotherapy. [ 21 ] Cervical cancer remains a major cause of female mortality worldwide, particularly in developing countries that have limited screening programs. [ 22 ] Only a small fraction (~1%) of women with cervical human papillomavirus (HPV) infection develop cervical neoplasia, [ 23 ] but the factors determining the risk of progression are incompletely understood. The host genetic variation is a major determinant of the likelihood of cervical neoplasia in HPV-affected women. Further research is needed such as the potential for genetic risk score analysis in combination with other measures to identify subsets of women at particularly high risk of cervical neoplasia. [ 24 , 25 ] To date, there is no literature involving genetic variations of the RAD18 gene and susceptibility to cervical cancer. We designed a case-control study based on a large sample population, selecting 6 SNP loci from RAD18 and detecting their distribution in the peripheral blood cell genomes of 650 cases of CIN III, 580 cases of cervical squamous cell carcinoma (CSCC), and 1320 normal healthy controls. We also investigated the relationship between different SNP genotypes and susceptibility to CSCC and CIN III, and conducted correlation analyses on corresponding clinical parameters related to prognosis, with the aim of better understanding the role of specific SNP genotypes in the carcinogenic process of CSCC.

Conceived and designed the experiments: Feng Ye, Rui Zhang. Performed the experiments: Rui Zhang, Jianping Kong, Yun Li, Hanzhi Wang, Qi Cheng, Caiyun Zhou, Minghua Yu. Analyzed the data: Jianping Kong, Yitong Wang. Contributed reagents/materials/analysis tools: Feng Ye, Hanzhi Wang, Qi Cheng. Wrote the paper: Feng Ye, Rui Zhang. Conceptualization: Feng Ye, Jianping Kong. Funding acquisition: Feng Ye. Investigation: Rui Zhang, Yun Li, Hanzhi Wang. Methodology: Hanzhi Wang, Qi Cheng. Validation: Hanzhi Wang. Caiyun Zhou, Minghua Yu. Writing – original draft: Rui Zhang. Writing – review & editing: Feng Ye.

We conducted a study looking at the role of selected SNPs in CSCC and its precursor lesion CIN III. In total, 580 CSCC patients and 650 CIN III patients from the Women’s Hospital, School of Medicine, Zhejiang University China were investigated and diagnoses were confirmed by 2 pathologists. In total, 1320 healthy volunteers were enrolled as controls in this study. The inclusion criteria for the normal healthy controls were as follows: no history of tumors, no cervical cytological finding, no endometriosis, and no immune-related disease. Among CSCC, the expression of the RAD18 protein was detected by immunohistochemistry on paraffin sections from 187 CSCC patient samples. Our research was approved by the Medical Ethical Committee of the Women’s Hospital, School of Medicine, Zhejiang University (No. 2004002). All patients signed informed consent for this molecular research. The clinical and pathohistological characteristics examined were: age, tumor family history, FIGO stage, tumor size, differentiation grade, lymph node metastasis, vascular involvement, stromal invasion, vaginal wall extension, parametrial extension and endometrial extension. The above data were obtained from the records of the archives of the Women’s Hospital, School of Medicine, Zhejiang University. Before the patients began their treatment, 2 mL of peripheral blood was collected in EDTA-anticoagulant tubes. Genomic DNA was extracted from the blood, using a whole blood gDNA extraction kit (Sangon Bio Co., Shanghai, China), following the manufacturer’s instructions. DNA quantification and quality were assessed using a NanoDrop 2000 (Thermo Fisher, Wilmington). Genomic DNA was dissolved in deionized water and frozen until use. A total of 6 tag-SNPs were picked up from a SNP library ( https://www.ncbi.nlm.nih.gov/snp/ ). We utilized the filter option (filters activated: SNP missense, 5’-UTR, 3’-UTR, minor allele frequency [MAF] from 0.05 to 0.5) to obtain 6 effective SNPs in the RAD18 gene (rs373572[A/G], rs615967[A/G], rs193920[A/C/T], rs250403[A/G], rs250404[C/T] and rs34927291[A/C/T]). One of these 6 SNPs was missensed in the coding region, 2 in the 5’-UTR region and 3 in the 3’-UTR region. Based on the work of Kisaki et al, [ 26 ] we designed allele-specific primers to detect specific nucleotides at each SNP site, then amplified them by PCR, and detected the products by agarose gel electrophoresis, and the different positive electrophoretic bands were used to determine the identity of nucleotide present. This method of detecting genetic variations in nucleotides is a modified allele-specific primer extension reaction. To prevent mismatched primer extensions, an artificial mismatched base was introduced at the second position from the 3’ terminal of the forward primer. The specific allele-specific primers and PCR product length are shown in Table S1 (Supplemental Digital Content, https://links.lww.com/MD/P735 ). PCR was conducted in a total volume of 20 µL and contained 20 ng DNA, 5.0 pmol forward and reverse primers, 0.25 mM dNTP and 1.0 U of Taq DNA polymerase (TAKARA Co., Dalian, China). Thermal cycling was carried out using a PCR Thermal Cycler S1000 (BIO-RAD) and programmed as follows: 10 minutes at 94°C, followed by 35 cycles at 94°C for 30 seconds, 56−60°C for 30 seconds, 72°C for 30 seconds, and a final elongation at 72°C was performed for 5 minutes. Then, electrophoresis was conducted using a 2% agarose gel, and the products were stained with ethidium bromide. All experimental results were validated by 2 technicians under double-blind conditions. The sections were firstly incubated with Rabbit anti-RAD18 Polyclonal antibody (1:200, 18333-1-AP, Proteintech), and then incubated with Dako Envision TM Peroxidase (Dako Diagnostica, Hamburg, Germany), and visualized with 3,3’-diaminobenzidine tetrahydrochloride (Dako). All slides were counterstained with hematoxylin. Immunohistochemical results were scored as follows: 0: <5% positive cells; 1: 5% to 25% positive cells; 2: 26% to 75% positive cells; 3: more than 76% positive cells. Stain intensity was scored as follows: 0, no staining; 1: faint-yellow; 2: brown-yellow; 3: dark-brown. The expression level (sum of the 2 scores) was finally defined as follows: (0), + (1–2), ++ (3–4), +++ (5–6). All the evaluations were made by 2 independent pathologists, unaware of the clinical data. We compared the allele frequencies of the SNPs in the RAD18 gene between the healthy control group and the patient groups with CIN III or CSCC. The distribution of the RAD18 SNPs genotype in all of the patients and the healthy controls was tested for adherence to the Hardy-Weinberg equilibrium. The binary logistic regression analysis was used to obtain odds ratios (ORs), 95% confidence intervals (CIs), and P values. The normal control group acted as the reference. The OR and 95% CI were both adjusted for age, sex, and smoking status using an unconditional logistic regression model. FDR adjusted p values were corrected using the Benjamin–Hochberg method for multiple testing corrections. The relationship between the genotype distribution frequency and the clinicopathological parameters was examined using the Kruskal–Wallis H test. Multinomial regression analysis was performed among the different groups for different genotypes. The immunohistochemistry looking at protein expression was assessed using the Kruskal–Wallis H test and the Mann–Whitney U test A P -value of ≤.05 was considered to be statistically significant. The statistical analyses were performed using SPSS (IBM Corp., Armonk) software (Version 18.0 for Windows).

Table 1 shows the genotypes and allele frequencies for the 6 genetic polymorphism loci of RAD18 (rs373572, rs615967, rs193920, rs250403, rs250404, rs34927291). All genotype frequency distributions met the requirements of the Hardy–Weinberg equilibrium (Table S2, Supplemental Digital Content, https://links.lww.com/MD/P735 ). Association between RAD18 genetic variants and the risk of CIN III and CSCCs. All P -values are adjusted for age, number of sexual partners, age at first intercourse, parities (including full-term pregnancy and abortion at or after 28 wk) and age at first full-term pregnancy. The frequency of the genotype distribution found that 4 genetic polymorphisms (rs373572, rs193920, rs250404, rs34927291) were not associated with the risk of CIN III or CSCC. The AA, AG, and GG genotype frequencies of RAD18 rs615967 were 36.1%, 47.3%, and 16.6% in normal controls; 31.1%, 49.4%, and 19.5% in the CIN III group and 26.6%, 49.7%, and 23.8% in the CSCC group, respectively. Patients with the rs615967 GG genotype had a significantly higher risk of CIN III [OR = 1.369; 95% CI: 1.042−1.800] and CSCC [OR = 1.952; 95% CI: 1.475−2.582). We also found that the frequency of G alleles at the rs615967 in the CIN III group (575/1300, 44.2%) and CSCC group (564/1160, 48.6%) were significantly higher than those in normal controls (1062/2640, 40.2%). The OR of the G allele in the CIN III group was 1.178 (95% CI: 1.030−1.348) and 1.406 (95% CI: 1.224−1.616) in the CSCC group. Carriers of the G allele (AG + GG) at rs615967 were associated with a higher risk of CIN III (OR = 1.255; 95% CI: 1.027−1.534) and CSCC (OR = 1.565; 95% CI: 1.261−1.942). The AA, AG, and GG genotype frequencies of RAD18 rs250403 were 60.9%, 33.1%, and 6.0% in the controls; 54.6%, 32.9%, and 12.5% in the CIN III group and 45.2%, 32.2%, and 22.6% in the CSCC group, respectively. Patients carrying the heterozygote AG genotype rs250403 also had a significantly elevated risk of CSCC (OR = 1.313; 95% CI: 1.053−1.638). The G allele frequencies of rs250403 in the CIN III (376/1300, 28.9%) and CSCC groups (449/1160, 38.7%) were higher than those in normal controls (595/2640, 22.5%). The G allele was associated with a higher risk of both CIN III (OR = 1.399; 95% CI: 1.203−1.626) and CSCC (OR = 2.170; 95% CI: 1.869−2.520), respectively. Carriers of the G allele (AG + GG) at rs250403 were associated with a higher risk of CIN III (OR = 1.295; 95% CI: 1.071−1.566) and CSCC (OR = 1.891; 95% CI: 1.552−2.304). Stratified analysis was conducted to analyze the association between the RAD18 rs250403 and rs615967 genotypes and age, number of sexual partners, age at first intercourse, number of parities, age at first parity, and HR-HPV infection status. There was no enrichment between CIN III and CSCC and RAD18 rs250403, rs615967 genetic polymorphism (Tables 2 and 3 ). Association between RAD18 rs250403 polymorphisms and the risk for CIN III and CSCCs stratified by the sexual, reproductive history. Stratified analysis were applied by the Kruskale–Wallis H test. A P value <.05 was considered significant. Association between RAD18 rs615967 polymorphisms and the risk for CIN III and CSCCs stratified by the sexual, reproductive history. Stratified analysis were applied by the Kruskale–Wallis H test. A P value <.05 was considered significant. Correlations with RAD18 rs250403 and rs615967 polymorphisms and the clinicopathological characteristics of CSCC are shown in Tables 4 and 5 . Association between RAD18 rs250403 polymorphism and the risk for cervical carcinoma stratified by clinical pathological characteristics. Stratified analysis were applied by the Kruskale–Wallis H test. A P value <.05 was considered significant. Association between RAD18 rs615967 polymorphism and the risk for cervical carcinoma stratified by clinical pathological characteristics. Stratified analysis were applied by the Kruskale–Wallis H test. A P value <.05 was considered significant. Stratified analysis was performed based on age, tumor family history, FIGO stage, tumor size, differentiation grade, lymph node metastasis, vascular involvement, stromal invasion, vaginal wall extension, parametrial extension, and endometrial extension. We found that the polymorphism of RAD18 rs250403 was significantly correlated with differentiation grade (χ 2  = 8.750, P  = .003), lymph node metastasis (χ 2  = 4.758, P  = .029), and vascular involvement (χ 2  = 4.082, P  = .043), while RAD18 rs615967 was significantly correlated with tumor family history (χ 2  = 6.012, P  = .014), differentiation grade (χ 2  = 11.435, P  = .001), and lymph node metastasis (χ 2  = 6.719, P  = .010). We analyzed the linkage disequilibrium between the genotype frequencies of rs250403 (A/G) and rs615967 (A/G), as these 2 genetic polymorphisms were significantly associated with the risk of CIN III and CSCC. As shown in Table 6 , when compared with the reference haplotype (AA-AA), the haplotypes of AG-GG (OR = 1.827; 95% CI: 1.176−2.840), GG-AA (OR = 2.033; 95% CI: 1.100−3.760), GG-AG (OR = 2.436; 95% CI: 1.410−4.210), and GG-GG (OR = 3.433; 95% CI: 1.900−6.202) were significantly associated with an increased risk of CIN III. for CSCCs, a higher risk was detected with AG-AG (OR = 1.674; 95% CI: 1.182−2.370), AG-GG (OR = 2.789; 95% CI: 1.771−4.393), GG-AA (OR = 4.529; 95% CI: 2.549−8.047), GG-AG (OR = 6.647; 95% CI: 4.011−11.015), and GG-GG (OR = 7.192; 95% CI: 4.061−12.736). These data indicated that the linkage mode of rs250403 (A/G) and RAD18 rs615967 (A/G) was associated with an elevated risk for CIN III and CSCC and the riskiest genetic linkage mode was GG-GG. Therefore, these specific linkage patterns were associated with a higher risk of CIN III or CSCC. The haplotypes of AG-GG, GG-AA, GG-AG, and GG-GG at rs250403 and rs615967 in the RAD18 gene may act as a genetic predictive biomarker for susceptibility of CIN III and/or CSCC. RAD18 haplotype of rs250403(A/G) and rs615967(A/G) and the risk of CIN III and CSCCs. Genotypes are composed of 2 polymorphic sites: rs250403 (A/G), rs615967 (A/G). All P -values are adjusted for age, number of sexual partners, age at first intercourse, parities (including full-term pregnancy and abortion at or after 28 wk) and age at first full-term pregnancy. As shown in Figure 1 A to C, among the 187 cases of CSCC, the frequencies of AA, AG, and GG genotypes of rs250403 were 115 (61.5%), 59 (31.6%), and 13 (6.9%), respectively. When the rs250403 (AA) was used as the control group, there was no significant difference in the expression of RAD18 protein when compared with the different genotype groups (χ 2  = 1.729, P  = .421). RAD18 expression in CSCC with different genetic polymorphisms as determined by immunohistochemistry (40× objective, magnified). (A) rs250403-AA; (B) rs250403-AG; (C) rs250403-GG; (D) rs615967-AA; (E) rs615967-AG; (F) rs615967-GG. The distinct brown coloration is mainly located in the nucleus of the positive cells. The frequencies of AA, AG and GG genotypes of rs615967 in the 187 CSCC patients were 65 (34.8%), 89 (47.6%), and 33 (17.6%). When rs615967 (AA) was used as the control, the expression of the RAD18 protein with the rs615967 (AG) genotype decreased by approximately 15%, and the rs615967 (GG) genotype decreased by approximately 33.9%. The expression of the RAD18 protein in patients with rs615967 (AG) and rs615967 (GG) were significantly lower than in patients with rs250403 (AA) (χ 2  = 11.598, P  = .003) (Fig. 1 D–F).

The development of cervical cancer is strongly associated with genital infection from oncogenic types of HPV. However, the majority of women infected with HPV never develop cancer. Pedigree studies show that cervical cancer has a significant heritability factor and genetic predisposing factors may influence the likelihood of sensitivity to, or persistence of HPV infection, as well as the rate of tumor development. [ 27 ] This suggests that genomic stability and genetic susceptibility play a critical role in the etiology of the genetic susceptibility of cervical cancer. Many studies, including 2 genome-wide association studies, have identified susceptibility loci and genetic variants in cervical cancer. [ 28 – 30 ] Our previous studies have also found that 2 SNP loci in the SMUG1 gene are significantly correlated with susceptibility to cervical cancer and HR-HPV infection, further supporting the important role of genomic genetic stability in cervical cancer. [ 31 ] Here, we determined whether the polymorphism of 6 SNPs within the RAD18 gene with MAF values of more than 5% was associated with the occurrence, progression, and prognostic risk of CIN III or CSCC. We found that the polymorphism of 4 SNPs (rs373572, rs193920, rs250404, rs34927291) did not differ in distribution among CIN III, CSCC or and healthy control groups, while there were significant differences in genotype distribution between the rs615967 (A/G) and rs250403 (A/G) loci. Furthermore, the GG homozygosity at rs615967 or carrying of the G allele (AG + GG) increased the risk of developing CIN III or CSCC. The GG homozygotes at rs250403 and those carrying the G allele (AG + GG) also have the same distribution and higher risk, especially the GG homozygotes at rs250403 and have an OR value of 5.089 in CSCC. We compared the RAD18 rs250403 (A/G) and rs615967 (A/G) haplotypes with the reference genotype AA-AA and found that haplotypes AG-GG, GG-AA, GG-AG, and GG-GG were significantly associated with an increased risk of CIN III. In addition, in CSCC, the risk of haplotypes possessing AG-AG, AG-GG, GG-AA, GG-AG, and GG-GG was much higher. Especially when both loci exhibited a G allele, the impact on disease susceptibility was much greater than when these 2 loci were analyzed separately. When both loci exhibited haplotypes of the GG homozygous type (GG-GG), the OR values for CIN III and CSCC were 3.433 and 7.192, respectively. A higher OR value, combined with statistical significance, indicated a synergistic effect between the rs250403 and rs615967 genetic polymorphisms in the RAD18 gene. This synergistic effect may promote the development of CIN III, ultimately leading to cervical cancer. As is well-known, persistent infection with high-risk HPV is a prerequisite for the occurrence of cervical cancer. Approximately 99.7% of cervical cancer cases are caused by persistent genital high-risk human papillomavirus infection. [ 32 ] It is interesting that although the rs250403 and rs615967 polymorphisms are significantly associated with the occurrence of CIN III and CSCC, when the stratified analysis was conducted on high-risk HPV infection, patient age, age of first sexual intercourse, frequency of childbirth, and age of first childbirth, (which are considered to be associated with an increased risk of cervical cancer), [ 33 , 34 ] we found that these characteristics were not associated with the polymorphisms of the 2 SNPs suggesting that their pathogenic role was not through increased susceptibility to high-risk HPV, but through the pathogenic process after HPV infection. In 2007, Kanzaki et al detected the RAD18 SNP (Arg302Gln) gene polymorphism in 100 colorectal cancer patients and 200 healthy controls in the Japanese population. They found a significant difference in genotype frequency between the control and the patient groups. In the control group, the frequencies of the Arg/Arg, Arg/Gln, and Gln/Gln genotypes were 43.0%, 45.5%, and 11.5%, respectively, while in colorectal cancer patients, they were 32.0%, 50.0%, and 18.0%, respectively. Compared with the control group with the Arg/Arg genotype, colorectal cancer patients with homozygous Gln/Gln (A/A) genotype showed the most significant increase in risk (OR = 2.10), and the Gln allele enhances susceptibility to colorectal cancer development. [ 17 ] Another study on 159 non-small cell lung cancer patients also showed a correlation between RAD18 -Arg302Gln polymorphism and the risk of non-small cell lung cancer in humans. The frequency of Gln/Gln genotype in non-small cell lung cancer patients (20.7%) was significantly higher than that in the healthy control group (11.5%), and the Gln/Gln genotype was detected to increase risk in non-small cell lung cancer patients (OR = 2.63). [ 18 ] The RAD18 SNP (Arg302Gln) in these 2 studies was located in the coding region of the RAD18 gene, but we are also very interested in the genetic variations in the noncoding region. Considering that over 90% of the associated genetic variations in the human genome are in the noncoding region of the genome, the noncoding region has the highest genome-wide association studies heritability (5-fold), and the genetic variations in these regions are crucial for understanding human phenotypic variations. [ 35 ] Therefore, we selected 6 SNPs with MAF values of more than 5% in the noncoding region of the RAD18 gene, all of which are located in the 5’-UTR promoter region or the 3’-UTR. Among them, rs615967, which is significantly associated with cervical cancer susceptibility, is located in the 5’-UTR promoter region and rs250403 is located in the 3’-UTR. These genetic variations in noncoding regions may affect gene function through the regulation of transcription, post-transcriptional modifications, and translational processes. Therefore, further functional studies are needed to elucidate their regulatory mechanisms. Considering that RAD18 rs615967 (A/G) is located in the 5‘-UTR promoter region, we believe that genetic variations may affect gene expression. Therefore, we measured the expression of RAD18 protein in the pathological tissues of 187 CSCC patients. We found that RAD18 protein was significantly reduced in the patients with the rs615967-AG and GG genotypes, indicating that the effect of RAD18 SNP (rs615967) on cervical cancer susceptibility may be due to changes in RAD18 expression, leading to a decrease in the ability to repair damaged genomes, resulting in genomic instability and tumorigenesis.We also found that although rs250403 (A/G) located in the 3’-UTR has a higher susceptibility risk for cervical cancer (rs250403 GG with an OR = 5.089), the different genotypes of rs250403 (A/G) did not lead to differences in protein expression. Therefore, we speculate that the rs250403 (A/G) SNP may change the spatial structure of protein functional domains by altering non synonymous changes in amino acid sequences, thereby affecting the level of DNA repair activity in cells, inducing genomic instability, and ultimately leading to cervical cancer. Furthermore, we analyzed the correlation between the polymorphisms of rs250403 and rs615967 and some clinical pathological features related to the prognosis of cervical cancer. We found that the polymorphisms of rs615967 and rs250403 were significantly correlated with lymph node metastasis and tumor differentiation. In addition, rs615967 was also associated with tumor family history and rs250403 was correlated with the degree of vascular involvement. Our findings are consistent with the results of other studies. There is a significant correlation between RAD18 SNP (Arg302Gln) gene polymorphisms and clinicopathological parameters in colorectal cancer, especially in terms of the degree of differentiation (OR = 7.00) and lymph node metastasis (OR = 3.71). In patients with elevated differentiation and lymph node metastasis (N1), the detection frequency of the Gln allele was higher. [ 17 ] Another study also found that in patients with colorectal cancer, the disease-free survival (DFS) in GG genotype patients with the RAD18 SNP of rs373572 was low. Compared with AG or AA genotype patients, the 1-year, 3-year, and 5-year DFS in patients with the GG genotype and the rs373572 RAD18 SNP were 86.7%, 53.3%, and 45.7%, respectively, while the 1-year, 3-year, and 5-year DFS in AG/AA genotype patients were 94.9%, 78.9%, and 74.2%, respectively. Especially in stage I colorectal cancer patients, the GG genotype is seen more commonly in patients with recurrent disease, making it a potential negative prognostic factor for early colorectal cancer diagnosis. [ 20 ] To summarize, these results indicated that the polymorphisms of RAD18 rs250403 and rs615967 were associated with disease susceptibility, disease progression, and prognosis in CIN III and CSCC. Some specific high-risk haplotypes (AG-GG, GG-AA, GG-AG, and GG-GG) linked by rs250403 and rs615967 serve as genetic biomarkers for predicting susceptibility to CIN III and CSCC. This is the first report providing evidence for the association between the RAD18 gene polymorphism and human cervical cancer risk.

We thank International Science Editing ( http://www.internationalscienceediting.com ) for editing this manuscript.