RRM1 *151A>T, RRM1 ‐756T>C, and RRM1 ‐585T>Gis associated with increased susceptibility of lung cancer in Chinese patients

Abstract Ribonucleotide reductase M1 (RRM1) is a crucial gene in DNA repair. Recent studies have shown that RRM1 expression can be a powerful predictor of survival or chemotherapy sensitivity in patients presenting with carcinomas who are treated with adjuvant gemcitabine‐based chemotherapy including lung cancer. However, the relationship between the single nucleotide polymorphisms (SNP) of RRM1 and the susceptibility of lung cancer to chemotherapy has not been well addressed. We detected six tag SNPs of RRM1 genotypes in a cohort of 1007 patients with primary lung cancer and 1007 age‐ and sex‐matched population controls using SNaPshot detection technology. Logistic regression, odds ratios (OR), and 95% confidence intervals were calculated to estimate lung cancer risk associated with SNP genotypes and haplotypes, after adjusting for case–control matching factors. Compared with the T/T and A/T genotype of RRM1 *151A>T, the A/A genotype had an increased risk for overall lung cancer (adjusted OR, 1.33). Additionally, the T/T+T/C genotypes of RRM1 ‐756T>C were risk factors that increased the susceptibility to lung cancer (adjusted OR 1.54, as compared with the C/C genotype). While the T/T+G/T genotypes of RRM1 ‐585T>G behaved as protective factors to increase the susceptibility to lung cancer (adjusted OR 0.44, as compared with the C/C genotype). In summary, this is the first study to systematically identify the relationship between the polymorphisms of RRM1 and individual susceptibility to lung cancer. It is anticipated that the RRM1 *151A>T, RRM1 ‐756T>C, and RRM1 ‐585T>G polymorphisms will improve the predictive prognosis of lung cancer sensitivity.

Introduction and 31%, respectively) than that at advanced stage (III/ IV) [3]. Thus, to identify certain population who may have susceptibility to lung cancer and diagnose lung cancer in early stage are crucial to improve treatment outcome. Since the genetic characteristics has been proved to contribute to lung cancer development, many molecular epidemiological studies have been conducted to evaluate the relationship between lung cancers and the genetic variety, such as single nucleotide polymorphisms (SNP) in genes which may be involved in lung cancer development. The DNA repair system plays an important role in protecting against mutagenesis and carcinogenesis. These genes are mainly involved in DNA maintenance and repair, carcinogen metabolism, cell cycle regulation, apoptosis, and so on. Some of the SNPs have been consistently shown to correlate with lung cancer susceptibility, including XPD [4], APEX, XRCC [5], POLG2, RECQL4 [6], and so on.
The ribonucleotide reductase M1 (RRM1) gene, a crucial gene of the nucleotide excision repair (NER) system, is important in the DNA damage removal pathway, which encodes the regulatory subunit of ribonucleotide reductase. This is the rate-limiting step in deoxyribonucleotide formation, and the only known enzyme that converts ribonucleotides to deoxyribonucletides, which is required for DNA polymerization and repair. Furthermore, RRM1 is the molecular target of gemcitabine, and its mRNA expression levels are related to the efficacy of gemcitabine therapy [7,8]. The RRM1 polymorphism -37 C<A and haplotype -was shown to be associated with susceptibility to gemcitabine in cancer patients including lung cancer [9,10]. However, the relationship between the ribonucleotide reductase M1 (RRM1) gene and susceptibility to lung cancer has not been well addressed.

Study subjects
Patients, diagnosed with lung cancer based on the pathological features, were recruited between June 2009 and August 2012 at Zhejiang Cancer Hospital. After omitting samples that failed to be genotyped, the final number of study subjects consisted of 1007 patients with primary lung cancer and 1007 age-and sex-matched population controls. In cancer cases, the peripheral blood samples were collected from patients with primary lung cancer who had not previously received surgical treatment or radiotherapy or chemotherapy.

SNP target determinations
For SNP site selection, we searched potential functional polymorphisms in SNP in the 5′ promoter region, nonsynonymous SNP, SNP in the 5′ untranslated region, synonymous SNP in the 3′ untranslated region, and synonymous SNP. In addition, tagSNPs were identified to examine common patterns of variation in the RRM1 gene regions. All SNPs were selected to account for all having a minor allele frequency (MAF) of 5% from the HapMap database. Haplotype frequencies and the linkage disequilibrium coefficient were estimated using Haploview software 4.2 [11]. HapMaptag SNPs in a pairwise linkage with a disequilibrium threshold were defined by a correlation coefficient (R 2 ) of 0.80. Hardy-Weinberg proportions in the controls and haplotypes were assessed by Fisher's exact test.

SNaPshot multiplex single base extension reaction
Extension reaction (10 μL) included 5 μL buffer from the SNaPshot Multiplex Kit (Applied Biosystems Co., Ltd., Foster City, CA, USA), 2 μL multiple PCR products after purification, 1 μL extension primer mix, and 2 μL of ultrapure water. The conditions of the PCR amplification were as follows: 96°C for 1 min followed by 96°C for 10 sec, 55°C for 5 sec, 60°C for 30 sec, which was repeated for 28 cycles, and then held at 4°C. Next, 0.5 μL of a Liz120 size standard and 9 μL of a Hi-Di mix were placed into 0.5 μL of the purified extension product, which was then denatured at 95°C for 5 min, then sequenced using an ABI 3730XL DNA analyzer. The original data were analyzed by GeneMapper 4.1 software (Applied Biosystems Co., Ltd., Foster City, California, USA).

Statistical analysis
Differences in gender, age, and genotypic frequencies that were found between cases and control subjects were X. Xu et al. RRM1 Polymorphisms and Lung Cancer Susceptibility evaluated by Pearson's chi-square (χ 2 ) test. The dominant and recessive models were used to assess the risk of the SNP genotype in lung cancer. The reference group was the minor homozygous genotype among controls. The associations between RRM1 genotypes and risk of lung cancer were estimated by computing the odds ratios (ORs) and their 95% confidence intervals (CIs) from both univariate and multivariate logistic regression analyses. The ORs were also adjusted for age and gender by the use of frequency matching. These statistical analyses were performed with the SPSS statistical software package (SPSS v.18.0; Chicago, IL, USA). All probability values were two sided and P values with an alpha value <0.05 were considered statistically significant.

Patient characteristics
For SNP genotyping, more than 99% of the samples were successfully determined and a representative plot for genotyping is shown in Figure 1A and B. There were no statistically significant differences between cases and controls in frequency distribution of sex, age, and lifestyle in the 1007 patients with lung cancer and 1007 controls (all Ps > 0.05). Allelic distributions in controls were in good agreement with those in the CHBHapMap populations. All genotypic distributions were compatible with the Hardy-Weinberg equilibrium (P > 0.05). Of the 1007
Genotyping data frequencies for six tagSNPs in RRM1 are shown in Table 1. The distributions of these RRM1 genotypes were then compared, respectively, among cases and controls. It was found that the frequencies of RRM1 *151A>T A/A, A/T, and T/T genotypes among cancer cases were significantly different as those among controls (6.8%, 43.3%, and 49.9% vs. 7.0%, 41.9%, and 51.1%, P = 0.001). There were also significant differences observed in RRM1 -702G>A and RRM1 -585T>G (P = 0.03 and P = 0.01). However, the frequencies of RRM1 -756T>C and RRM1 -269C>A genotypes were almost the same as those among in cancer cases and controls (P = 0.11 and P = 0.19).
We used logistic regression analysis to examine potential associations between polymorphisms in the RRM1 genes and the risk of lung cancer after adjusting for age, gender, and lifestyles. Adjusted ORs are represented in Table 1. Both dominant and recessive models were considered for each SNP.
Compared with the T/T and A/T genotypes of RRM1 *151A>T, the A/A genotype had an increased risk for overall lung cancer (adjusted OR, 1.33; 95% CI, 1.11-1.60) by recessive models. Additionally, the T/T+T/C genotypes of RRM1 -756T>C behaved as risk factor of enhanced susceptibility of lung cancer (adjusted OR 1.54; 95% CI, 1.10-2.15) as compared with the C/C genotype. While the T/T+G/T genotypes of RRM1 -585T>G behaved as protective factors in increasing the susceptibility of lung cancer (adjusted OR, 0.44; 95% CI, 0.24-0.84) as compared with the C/C genotype. However, there was no evidence that showed any association between the heterozygous genotypes of both RRM1 -269C>A and RRM1 -702G>A sites and relative risk of lung cancer development.

Association between SNPs and risk of lung cancer pathology subtypes
Since lung cancers are derived from different cell types and carcinogenesis of different subtypes of lung cancer There were 857 patients with NSCLC and 150 patients with SCLC among the total of 1007 lung cancer cases. In dominant models, it was found that RRM1 *151A>T had more than eightfold increased risk associated with the polymorphisms in patients with SCLC, with an adjusted OR of 8.54 (95% CI, 1.02-71.3; P = 0.48) for the A/ A+A/T genotype as compared with the T/T genotype. However, there was no evidence that showed any association between the polymorphisms of RRM1 and SCLC. In addition, we found that there was a significant association between the polymorphisms of RRM1 -585T>G and NSCLC with an adjusted OR of 0.62 (95% CI, 0.36-1.05; P = 0.07) by dominant models in patients with non-small-cell lung cancer. Thus, the RRM1 -756T>C had an adjusted OR of 1.72 (95% CI, 1.19-2.48; P = 0.00). In the recessive model, RRM1 *151A>T had a remarkable association with the susceptibility of lung cancer with 1.43 (95% CI, 1.17-1.74; P = 0.00).
The potential interactions between these six genetic polymorphisms and age on risk of lung cancer were analyzed in a logistic regression model. However, no effect was observed (data not shown), apparently due to the small numbers involved.

Haplotype analysis
The haplotype frequencies of RRM1among cases and controls were estimated and the results are presented in Table 2. No significant difference was observed between cases and controls in terms of haplotype frequencies of the RRM1 *151A>T and RRM1 *316C>A polymorphisms. We also performed linkage disequilibrium analysis to examine the linkage between the polymorphisms at the two loci. The two locus disequilibrium was statistically significant (R 2 = 1), which indicated that the two polymorphisms were completely linked in our study population. There was no significant linkage disequilibrium found among the other four polymorphisms of RRM1 for R 2 < 0.8.

Discussion
Carcinogenesis is a multistep process which involved genetic and environmental risk factors. Common polymorphisms in many genes, including those involved in DNA repair, have been shown to have individual susceptibility to cancers. It has been proved that polymorphisms of DNA repair genes, such as the polymorphisms of XPA -4G>A (rs1800975), ERCC2 862G>A (rs1799793) [12], MSH3 3133G>A (rs26279), and PMS1 639G>A (rs5742938) [13], are associated with the risk of lung cancer.
RRM1 is involved in several important DNA repair processes (e.g., nucleotide excision repair, interstrand crosslink repair, ROS-induced single-strand break repair, and double-strand break repair), and it resides on chromosome segment 11p15.5 [14] in a region of frequent loss of heterozygosity (LOH). Clinical and cell biological studies suggest that LOH is an important tumor-suppressor gene region in lung cancer [15,16]. Moreover, RRM1 is a metastasis suppressor gene through PTEN-regulated pathways in lung cancer [17]. Furthermore, it encodes the regulatory subunit of ribonucleotide reductase which was inhibited by gemcitabine metabolites (5′ diphosphate) [18]. In a recent study, ribonucleotide reductase might play a key role in suppressing the invasive capacity and anchorageindependent growth of human lung cancers through the ability of the ribonucleotide reductase small subunit p53R2 to bind ERK kinase 2, thereby suppressing MEK-ERK activity [19].
Many recent works have mainly focused on whether RRM1 can be used as a molecular predictor in lung cancer by different methods of biomarker analysis, including quantitative real-time PCR analysis, immunohistochemistry, western blot analysis, and automated quantitative analysis [7,[20][21][22]. A meta-analysis of 18 studies evaluating a gemcitabine-based regimen in patients with advanced NSCLC (n = 1243) was recently published [21]. Patients with low tumor RRM1 expression assessed using either qRT-PCR or IHC survived 3.94 months longer (95% CI, 2.15-5.73; P = 0.001) than those patients with high RRM1 expression. Meanwhile, RRM1 expression has been shown to be a powerful predictor of survival or chemotherapy susceptibility in patients with carcinomas, such as pancreatic cancer [23], advanced nasopharyngeal carcinoma [24], and gastric cancer [25], treated with adjuvant gemcitabine-based chemotherapy. Since the specimen of core needle biopsies derived from patient tumors are always inconvenient to obtain, and the technologies of immunohistochemical methods, mRNA expression and quantitative in situ protein analyses, are complicated, it is necessary to identify a more generally applicable methodology. Genotyping of SNPs is more convenient and stable; additionally, genotyping is assisted by the relative ease of obtaining peripheral blood. However, whether the polymorphism of RRM1 was related to individual risk of developing lung cancer has not been well investigated.
Genetic polymorphisms may affect protein structure, function, stability, or folding. The most common form of polymorphism in the human genome is an SNP, and some SNPs have been shown to correlate with cancer sensitivity including lung cancer. In this study, we selected six SNP sites in RRM1, four sites located at 3′ untranslated region (UTR), one located at 5′ UTR, and the other two sites were located at the promoter. In addition, micro-RNAs, 21-24 nucleotide-long small noncoding RNA gene products, can regulate gene expression by base pairing with the 3′ UTR region of target mRNAs.
Finally, we identified three common polymorphisms of RRM1 that were associated with lung cancer in the Chinese population. On the basis of the analysis of 1007 patients and 1007 frequency-matched controls, the A/A genotype of polymorphism RRM1 *151A>T and the T/T+T/C genotype of RRM1 -756T>C was positively associated with an increased risk to the disease. While the T/T+G/T genotype of RRM1 -585T>G was associated with a decreased risk for tumorigenesis, suggesting that these SNPs may have potential function to influence the amplification and expression of RRM1. Subgroup analysis was also conducted, and polymorphisms of RRM1were both associated with NSCLC and SCLC.
However, our study also has its limitations. The primary limitation of our study was the use of a single center study and consequently low power of analysis in statistical measurements, which can lead to both false negative as well as false positive findings. Moreover, such findings should thus be considered preliminary [26].
In summary, to the best of our knowledge, this is the first study to systematically identify the relationship between the tagSNPs of RRM1 and individual susceptibility to lung cancer. Hopefully, the RRM1 *151A>T, RRM1 -756T>C, and RRM1 -585T>G will improve the prediction of lung cancer sensitivity. However, these results suggest a few potential candidates for further investigation in larger studies. Furthermore, the regulatory pathway associated with potentially surrogate SNPs remains to be explored.