ERCC2 2251A>C genetic polymorphism was highly correlated with early relapse in high-risk stage II and stage III colorectal cancer patients: A preliminary study

Background Early relapse in colorectal cancer (CRC) patients is attributed mainly to the higher malignant entity (such as an unfavorable genotype, deeper tumor invasion, lymph node metastasis and advance cancer stage) and poor response to chemotherapy. Several investigations have demonstrated that genetic polymorphisms in drug-targeted genes, metabolizing enzymes, and DNA-repairing enzymes are all strongly correlated with inter-individual differences in the efficacy and toxicity of many treatment regimens. This preliminary study attempts to identify the correlation between genetic polymorphisms and clinicopathological features of CRC, and evaluates the relationship between genetic polymorphisms and chemotherapeutic susceptibility of Taiwanese CRC patients. To our knowledge, this study discusses, for the first time, early cancer relapse and its indication by multiple genes. Methods Six gene polymorphisms functional in drug-metabolism – GSTP1 Ile105Val, ABCB1 Ile1145Ile, MTHFR Ala222Val, TYMS double (2R) or triple (3R) tandem repeat – and DNA-repair genes – ERCC2 Lys751Gln and XRCC1 Arg399Gln – were assessed in 201 CRC patients using a polymerase chain reaction-restriction fragment-length polymorphism (PCR-RFLP) technique and DNA sequencing. Patients were diagnosed as either high-risk stage II (T2 and 3 N0 M0) or III (any T N1 and 2 M0) and were administered adjuvant chemotherapy regimens that included 5-fluorouracil (5FU) and leucovorin (LV). The correlations between genetic polymorphisms and patient clinicopathological features and relapses were investigated. Results In this study, the distributions of GSTP1 (P = 0.003), ABCB1 (P = 0.001), TYMS (P < 0.0001), ERCC2 (P < 0.0001) and XRCC1 (P = 0.006) genotypes in the Asian population, with the exception of MTHFR (P = 0.081), differed significantly from their distributions in a Caucasian population. However, the unfavorable genotype ERCC2 2251A>C (P = 0.006), tumor invasion depth (P = 0.025), lymph node metastasis (P = 0.011) and cancer stage (P = 0.008) were significantly correlated with early relapse. Patients carrying the ERCC2 2251AC or2251CC genotypes had a significantly increased risk of early relapse (OR = 3.294, 95% CI, 1.272–8.532). Conclusion We suggest that ERCC2 2251A>C alleles may be genetic predictors of early CRC relapse.


Background
The primary treatment for colorectal cancer (CRC) is resection of the primary tumor. After surgery, patients are frequently administered adjuvant chemotherapy to eliminate cancer cells that may have metastasized [1]. Despite chemotherapy, CRC remains the third major cause of cancer-related death in Taiwan, accounting for >3,000 deaths per year [2]. The overall five-year survival is 50-60% in European countries [3], a result similar to that in Taiwan [4]. The primary cause of death is distant and locoregional relapses. Notably, CRC relapse is strongly correlated with chemotherapeutic drug response [1,5,6].
Genetic polymorphisms in drug-targeted genes [7,8], metabolizing enzymes [9], and DNA-repairing enzymes [10] have been linked to inter-individual differences in the efficacy and toxicity of numerous drugs. Several studies have investigated various gene expressions and chemotherapeutic drug responses of cancers. For instance, a polymorphic 28-bp tandem double repeat polymorphism in the regulatory region of the TYMS gene is correlated with a better response to 5-fluorouracil (5 FU) chemotherapy than the triple repeat in the polymorphism [11]. A common polymorphism in the MTHFR gene (677C>T; Ala222Val) increases the efficacy of fluoropyrimidinebased chemotherapy [12]. In Caucasian patients with advanced CRC who are treated with oxaliplatin, 5 FU and leucovorin (LV), ERCC1 118 T/T, XRCC1 (Arg ->Gln substitution in exon 10), ERCC2 751AC, and ERCC2 751CC genotypes are independently associated with poor progression-free survival and short-term survival [13,14]. Glutathione S-transferases (GSTs) participate in the detoxification of platinum compounds and are important mediators of intrinsic and acquired resistance to oxaliplatin [8].
Advanced CRC is one of the most chemotherapy-resistant human malignancies. The cytotoxic agent with the most consistent antitumor activity is 5 FU. However, significant variability in drug response can occur among cancer patients treated with the same medications [15]. Elevated levels of thymidylate synthase (TS) are correlated with resistance to 5 FU and a poor clinical outcome [16][17][18][19].
Conventional regimens for treating cancer patients with chemotherapy do not account for interpatient variability in the expression of particular target genes. Such variability results in unpredictable tumor responses and host toxicity. This hospital-based study investigates the role of multiple genetic polymorphisms of six metabolizing and DNA-repair genes (GSTP1, Glutathione S-transferase P1; ABCB1, multidrug resistance 1; MTHFR, methylenetetrahydrofolate reductase; TYMS, thymidylate synthase; ERCC2, excision repair cross-complementing rodent repair deficiency, complementation group 2; XRCC1, X-ray cross-complementing 1) in literature, genomic databases, and the Medline database [10][11][12][20][21][22][23]. The correlations between single nucleotide polymorphisms of the six candidate genes and clinicopathological features of 201 Taiwanese CRC patients, in addition to the relationship between genetic variants and post-therapy early relapse, were also analyzed to elucidate the roles of genotypes of these six genes as predictors of response of CRC patients following 5 FU/LV chemotherapy. Furthermore, this study reviewed current literature regarding the distribution of these six candidate genes and their genotypes in CRC patients of different ethnic groups, and compared differences between Taiwanese CRC patients and those of other races. This is the first investigation of clinical outcome using multiple chemotherapeutic drug-related genetic polymorphisms for Taiwanese patients with advanced CRC. The ability to predict with a high degree of accuracy which patients are likely to respond to treatment and those who are unlikely to respond will significantly influence the design of new treatment regimens.

Patients and specimens
Enrolled in this prospective study were 201 Taiwanese high-risk International Union Against Cancer (UICC) stage II and stage III CRC patients (median age, 62.09 ± 12.67 years) who were admitted to the Department of Surgery at Kaohsiung Medical University Hospital for elective surgery. Patients with other malignant diseases in their medical history were excluded. All 201 patients underwent radical resection for a primary lesion. Radical resection was defined as any gross residual tumor that did not remain in the surgical bed, and the surgical resection margin is pathologically negative for tumor invasion. All patients were diagnosed as either high-risk stage II (T2 and 3 N0 M0) or III (any T N1 and 2 M0). Patients with risk factors for relapse (tumor poorly differentiated, tumor perforation, number of lymph nodes examined <12 or lymphatic/vascular invasion) were considered high-risk stage II cases. Written informed consent was obtained from all subjects and/or guardians for use of their blood samples. The follow-up endpoint was June 2007. Sample acquisition and subsequent use were approved by the institutional review board at the Kaohsiung Medical University Hospital. Patients were administered six 8-week cycles of adjuvant chemotherapy. Each cycle consisted of leucovorin 500 mg/m 2 administered as a 2-h infusion and given weekly for six doses, and 5 FU 500 mg/m 2 administered as an intravenous bolus 1 h after the start of leucovorin infusion and repeated weekly for 6 doses. This cycle was then repeated after a 2-week rest period. Postoperative surveillance consisted of comprised medical history, physical examination, and laboratory studies, including assessing serum carcinoembryonic antigen (CEA) levels at 3-month intervals. Abdominal ultrasonography or computed tomography was performed at 6-month intervals, and chest radiography, bone scans, and total colonoscopy were performed annually. Patients were followed up at 3month intervals for 2 years and at 6-month intervals thereafter. Median follow-up time was 34.2 months (range, 24-40 months). Clinical stage and pathological features of primary tumors were defined according to criteria of the American Joint Commission on Cancer/International Union Against Cancer (AJCC/UICC) [24].
Development of new post-operative recurrent or metastatic lesions was defined as postoperative relapse. Early relapse was defined as local recurrence (tumor growth restricted to the anastomosis or the region of the primary operation) or distant metastasis (distant metastasis or diffuse peritoneal seeding) within 1 year following radical resection.

DNA extraction
Constitutional gene polymorphisms were analyzed via DNA extraction from 4 ml peripheral blood using a PURE-GENE ® DNA Isolation Kit (Gentra Systems, Inc., Minneapolis, MN, USA).

PCR-RFLP
All genomic DNA from patients were examined using the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) approach to determine the genotypes of GSTP1, ABCB1, MTHFR, TYMS, ERCC2, and XRCC1. Following digestion with suitable restriction enzymes, PCR fragments were separated on a 2.5-3.0% agarose gel and visualized after ethidium bromide staining. This study also utilized the automated sequencing approach to confirm PCR-RFLP results (Figs. 1, 2, 3, 4). All primers utilized in this study were designed by using Primer3 freeware [25]. Table 1 presents the primer sequences and restriction enzymes. The PCR reaction volume was 40 μL. The PCR conditions, for the GSTP1 poly-morphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, annealing at 65°C for 10 s, and 72°C for 25 s. A 433-bp product was amplified following digestion with BsmAI. Digestion generated fragments of 329 and 104 bp for the A allele, and fragments of 222, 107, and 104 bp for the Gallele. The PCR conditions for the ABCB1 polymorphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, annealing at 58°C for 15 s, and 72°C for 20 s. A 244-bp product was amplified following digestion with DpnII. Digestion generated fragments of 172 and 72 bp for the C allele, and a 244-bp fragment for the T allele. The PCR conditions for the MTHFR polymorphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, annealing at 70°C for 20 s, and 72°C for 25 s. A 400-bp product was amplified following digestion with HinfI. Digestion produced a 400-bp fragment for the C allele, and 318 and 82 bp fragments for the T allele. The PCR conditions for the TYMS polymorphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, and 72°C for 35 s. A 240-and 212-bp product was amplified for the 3R allele and 2R allele, respectively. The PCR conditions for the ERCC2 polymorphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, annealing at 64°C for 10 s, and 72°C for 20 s. A 149-bp product was amplified following digestion with PstI. Digestion generated 143-and 6-bp fragments for the A allele, and fragments of 80, 63, and 6 bp for the C allele. The PCR conditions for the XRCC1 polymorphism were as follows: 95°C for 5 min, 35 cycles at 95°C for 30 s, annealing at 62°C for 20 s, and 72°C for 20 s. A 208-bp product was amplified following digestion with MspI. Digestion produced 156-and 52-bp fragments for the G allele, and a 208-bp fragment for the A allele.

Statistical analysis
All data were analyzed using the Statistical Package for the Social Sciences Version 12.0 (SPSS, Inc., Chicago, IL, USA). The correlation between a polymorphism and relapse status was assessed using the relative risk ratio and a 95% confidence interval (CI). This is an exploratory and hypothesis-generating study. To clarify the clinical significance of these combined genotypes as predictors of postoperative relapse, a multivariate adjustment was performed by logistic regression analysis. The two-sided Pearson χ 2 test was applied to assess the differences in distributions of genotypes between different races. A value of P < 0.05 was considered statistically significant. PCR-RFLP analysis and automated sequencing of GSTP1 A313G

Correlation between early relapse and clinicopathological data
No statistical correlations existed between early relapse status and sex, age, tumor size, tumor location and histology (P > 0.05; Table 2). However, depth of tumor invasion, lymph node metastasis and cancer stage were significantly related to early relapse (P < 0.05). Risk of early relapse was higher in T3+T4 that had a deeper invasion than did T2 (OR, 1.205; 95% CI, 1.127-1.289; P = 0.025). Stage III patients had a higher early relapse rate than stage II patients (OR, 2.897; 95% CI, 1.243-6.381; P = 0.011), and early relapse was higher in cases with lymph node metastasis than those with no lymph node metastasis (OR, 2.987; 95% CI, 1.293-6.899; P = 0.008).

Comparison of genotype distributions between Taiwanese CRC patients and those of other races
To compare the genotypes of Taiwanese CRC patients with those of other races (Table 3)

Correlation between genetic polymorphisms with clinicopathological data
This study assessed correlations between genetic polymorphisms and clinicopathological features of 201 Taiwanese CRC patients (Table 4). No statistically significant correlations existed between genotype distributions and sex, age, tumor site or location, depth of tumor invasion, lymph node metastasis, cancer stage, or histology (all P > 0.05).

Correlation between early relapse and genetic polymorphisms
The correlations between gene polymorphisms (GSTP1 313A>G, ABCB1 3435C>T, MTHFR 667C>T, TYMS double or triple tandem repeats, ERCC2 2251A>C and XRCC1 1196G>A) and patients with or without early recurrence were examined. Statistical analyses indicate that genotype polymorphisms of ERCC2 were strongly correlated between patients with recurrent and non-recurrent tumors (P = 0.006; Table 5); however, GSTP1, ABCB1, MTHFR, TYMS, and XRCC1 genotypes polymorphisms were not correlated (P > 0.05; Table 5). The Taiwanese CRC patients with ERCC2 2251AC and ERCC2 2251CC genotypes have a risk of recurrence 3.294 times greater than that of those with other genotypes (P = 0.01; 95% CI, 1.272-8.532).

Discussion
This is an exploratory and hypothesis-generating study. This study attempted to move beyond single genetic polymorphisms to a more comprehensive investigation that identifies genomic variants and patterns and performed Figure 4 PCR-RFLP analysis of TYMS repeat polymorphisms. Figure 4 presents PCR-RFLP analysis results for TYMS repeat polymorphisms. A 240-bp product was amplified for the 3R allele and a 212-bp product was amplified for the 2R allele.

PCR-RFLP analysis of TYMS repeat polymorphisms
early relapse analysis. This study investigated six functional genomic polymorphisms in genes, which have different enzyme functions, or expressions, and DNA repair by PCR-RFLP assay and cycling sequencing. This is the first comprehensive study to investigate genotype frequencies of six gene polymorphisms in Taiwanese CRC patients. Additionally, this study analyzed the genotypes of 201 Taiwanese CRC patients and demonstrated that genotype distributions have ethnic variations.
In 5FU-related genes, the incidence of the TYMS gene promoter 3R/3R genotype in the 201 cases in this study is similar to that in Japanese [34] and Chinese [35] populations, and higher than that in Caucasian populations [36]. Neither the TYMS 2R allele nor the TYMS 3R allele is predominant in Caucasians; however, the TYMS 3R allele is predominant in Asian populations. For the MTHFR C677T polymorphism, the incidence of the 677CC genotype in this Taiwanese CRC population was higher than that for Caucasians [31], and Japanese [32], and Korean [33] populations (Table 3). Notably, the C allele and T allele frequencies in Caucasian and Asian populations are similar.
This study assessed the ABCB1 3435C>T of 201 Taiwanese CRC patients. The frequency of the TT genotype in the 201 Taiwanese patients is similar to that in Japanese [29] and Korean [30] patients, but lower than that in Caucasian populations [28]. The allele frequency of C and T did not differ between Caucasians and Asians.
The genotype data for cell detoxification-related gene polymorphism, GSTP1 313A>G, are similar to that for 408 Taiwanese CRC patients in a study by Yen et al. [27]. Caucasian populations have an equal incidence of AA homozygotes and AG heterozygotes [26]; however, the Taiwanese population has more AA homozygotes than AG heterozygotes. The allele frequency of Ile is also more predominant in Taiwanese patients than in Caucasian patients.
Since the genetic polymorphisms of DNA-repair enzymes may influence DNA adduct levels [40][41][42], the particular degree of DNA repair capacity can be utilized to identify genetically high-risk individuals for human cancers [43]. Resistance to platinum agents has been attributed to increased tolerance to platinum DNA adducts, enhanced DNA repair, or decreased drug accumulation [44]. Proteins of the nucleotide excision repair (NER) pathway, in particular, are believed play a central role in repair of DNA damage caused by platinum compounds. The enzymes related to the DNA repair system, XRCC1 and ERCC2, were similar to those identified in other studies of Taiwanese CRC patients [38]. Although very few studies have  (14) a International Union Against Cancer investigated these two polymorphisms in Asian CRC populations, the incidence of XRCC1 G1196A in a Korean population [39] is similar to that in the 201 Taiwanese patients in this study. However, the frequency of the XRCC1 GG homozygous and GA heterozygous patients in Caucasian [37] and Asian populations differ significantly. Moreover, the frequency of the G allele and A alleles do not differ between Caucasians and Asians. For ERCC2 2251A>C, the genotype and allele frequencies are dissimilar between Taiwanese and Caucasians [37]. Neither the A allele nor C allele has been identified as predominant in Caucasians, whereas the A allele is predominant in the Taiwanese (Table 3).
In 2004, Shirao et al. compared the efficacy, toxicities, and pharmacokinetics of an oral regimen consisting of uracil/ tegafur (UFT) and LV between Japanese patients and Caucasian patients in the United States [45]. Although the response rate did not differ (36.4% for Japanese patients and 34.1% for patients in the United States), a difference existed in toxicity profile, specifically the incidence of diarrhea -9% in the Japanese population and 22% in the Caucasian population in the United States. Since genotype frequencies were similar between Japanese and Taiwanese populations, we suggest that cancer chemotherapy efficacy, toxicities, and pharmacokinetics would be similar to those for Japanese patients. Consequently, we hypothesize that these polymorphisms screened have the ability to predict toxicity, clinical outcome and survival in Taiwanese CRC patients. This hypothesis and suggestion warrant further investigation. Studies focusing on acute relapse risk of CRC and these metabolizing and DNA-repair genes are limited and controversial. In addition to the statistically significant role of early relapse, including depth of tumor invasion (P = 0.025), lymph node metastasis (P = 0.011), and cancer stage (P = 0.008), analytical results in this study indicate ERCC2 2251A>C AC and CC gene polymorphismswere statistically significant in predicting early relapse.
The DNA repair systems play an important role in maintaining genomic integrity and preventing carcinogenesis [46]. At least four pathways for DNA repair operate on specific DNA damage types [47]. Nucleotide excision repair is the primary pathway in humans. The ERCC2 gene plays a dominant role in nucleotide excision repair and basal transcription, both of which are crucial to the elimination of bulky DNA adducts. Furthermore, the ERCC2 protein is essential for nucleotide excision repair activity [48,49]. The ERCC2 gene consists of 23 exons at 19q13.3. Several single nucleotide polymorphisms have been identified in the coding part of ERCC2, of which Ile199Met (C/ G), His201Tyr (C/T), Asp312Asn (G/A), and Lys751Gln (A/C) result in amino acid changes; however, codon Arg156Arg (C/A) and Asp711Asp (C/T) are silent polymorphisms. The functional effect of ERCC2 polymorphisms remains unclear. The polymorphisms at codon 312 and 751 have been analyzed extensively for their potential ability to increase lung cancer risk [49,50].
Although an earlier study showed that ERCC2 homozygous variant individuals were at increased risk of adenomatous polyps [51], associations between DNA repair gene polymorphisms and CRC have not been explored extensively. This study suggests that CRC patients who have the ERCC2 2251AC (751Lys/Gln) or 2251CC (751Glu/Gln) genotypes have a significantly increased early relapse risk (OR = 3.294; 95% CI, 1.272-8.532), and no statistically significant correlation exists between genotype distributions and clinicopathological features (all P > 0.05).
This study examined potential genetic predictors of CRC relapse. A recent study [52] determined that the C(Glu)  [42,[54][55][56][57][58], we hypothesize that the malignant entity of the C (Glu) allele of ERCC2 Lys751Gln is responsible for poor prognosis of early relapse among CRC patients receiving similar chemotherapeutic regimens in this study. In 2001, Park reported the finding of a significant relationship between clinical response to chemotherapy (combined oxaliplatin and 5FU) and ERCC2 Lys751Gln polymorphisms. Patients with the Gln/Gln genotype had significantly shorter survival or increased relative risk of dying when compared with the Lys/Lys group [10]. These experimental results suggest that ERCC2 2251A>C may indeed be of functional importance in Taiwanese CRC. Aggressive chemotherapy or combined radiotherapy will be considered for patients with this poor prognostic genotype.

Conclusion
In conclusion, analytical data in this study suggest that the polymorphism ERCC2 A2251C is associated with risk of CRC early relapse in a Taiwanese population. Additionally, this study analyzed six gene polymorphisms associated with the clinical outcome of Taiwanese CRC patients. These genotypic frequencies vary among ethnic groups. Studies with relatively larger sample sizes are needed to elucidate the effects of these polymorphisms on early relapse risk in this population. However, experimental results in this study serve as a basis for large-scale correlational studies on the relevance of these variants in predicting early relapse of CRC in a Taiwanese population.

Competing interests
The author(s) declare that they have no competing interests.