Abstract
Objectives
Germline genetic variation may affect clinical outcomes of cancer patients. We applied a candidate-gene approach to evaluate the effect of putative markers on survival of patients with pancreatic cancer. We also examined gene-radiotherapy and gene-chemotherapy interactions, aiming to explain inter-individual differences in treatment outcomes.
Methods
In total, 211 patients with pancreatic cancer were recruited in a population-based study. Sixty-four candidate genes associated with cancer survival or treatment response were selected from existing publications. Genotype information was obtained from a previous GWAS dataset. The main effect of genetic variation and gene-specific treatment interactions on overall survival were examined by proportional hazards regression models.
Results
Fourteen genes showed evidence of association with pancreatic cancer survival. Among these, rs1760217, located at the DPYD gene, rs17091162 at SERPINA3 and rs2231164 at ABCG2 had the lowest P-values of 10−4.60, 0.0013 and 0.0023, respectively. We also observed that two genes, RRM1 and IQGAP2, had significant interactions with radiotherapy in association with survival, and two others, TYMS and MET, showed evidence of interaction with 5-FU and erlotinib, respectively.
Conclusions
Our study suggested significant associations between germline genetic polymorphisms and overall survival in pancreatic cancer, as well as survival interactions between various genes and radiotherapy and chemotherapy.
Keywords: Pancreatic neoplasms, Survival, Genetic heterogeneity, Polymorphism, single nucleotide, Prognosis
Introduction
Pancreatic cancer is the fourth-most frequent cause of cancer death in the United States. Because of the aggressiveness and generally late diagnosis of the disease, five-year survival is less than 5% and median survival is less than 6 months for all stages of pancreatic cancer combined.1
The poor prognosis of pancreatic cancer varies appreciably across patient characteristics, tumor stage, treatment and compliance. Systemic treatment has been shown to improve outcome.2 In particular, surgery is the only potentially curative treatment, but only about 20-25% of cases are diagnosed at a resectable state. Therefore, radiotherapy and chemotherapy have been implemented as adjuvant or palliative strategies, but their effects on the prognosis of pancreatic cancer vary among patients. Previous studies have suggested that genetic polymorphisms may influence pancreatic cancer prognosis and treatment efficacy.3-7 However, most of these studies were based on only a few genes or polymorphisms and gene-treatment interactions have not been systematically explored.
Screening for disease loci on a genome-wide scale has become available with the development of high-throughput genotyping. Genome-wide association studies (GWAS) of disease risk have been carried out in pancreatic cancer.8,9 Susceptibility regions including the ABO gene and loci in chromosomes 13q22.1, 1q32.1 and 5p15.33 have been found to be associated with risk of pancreatic cancer.8,9 GWAS studies which produce massive amounts of genetic information can be adapted to examine the role of genetic variation in prognosis and response to specific treatments in patients with pancreatic cancer.10
In the present study, we selected from existing literature publications a total of 64 candidate genes, all having evidence of involvement in carcinogenesis-related molecular processes or drug metabolic pathways such as DNA repair, oxidative stress, cell cycle and signal transduction. Using genotype information of 211 subjects from our population-based case-control study that were in the PanScan GWAS,8,9 we performed a systematic analysis of the prognostic impact of the genetic variation. In addition, with available detailed treatment information on our study subjects, we explored potential gene-radiotherapy and gene-chemotherapy interactions on survival.
Materials and Methods
Study Population
The present study population consists of a subset of participants from a population-based case-control study of pancreatic cancer in Connecticut conducted between 2005 and 2009, described previously.11 Briefly, 216 cases in the study were selected as a representative sample for inclusion in PanScan.8,9 Of the 216, 211 had available follow-up information and were included in the present analysis. These cases were diagnosed between February 2005 and June 2008, and followed through June 2010, with an average follow-up of 3.4 years (minimum 1.9 years, maximum 5.3 years).
Data Collection
Clinical information was obtained from patients' medical records, from in-person interviews, and from clinical records maintained by the Connecticut Tumor Registry (CTR). The collected information included date of case diagnosis, age at diagnosis, sex, race, SEER stage of tumor, fact of treatment by surgery, radiotherapy or chemotherapy, and all specific chemotherapeutic agents used. Fact and date of death were obtained from the CTR records as well as from the Social Security Death Index, using Social Security Number for identification. Overall survival time was calculated from the date of pathological diagnosis to the date of death, last follow-up or last treatment visit listed in the CTR, or June 30, 2010. The study was approved by the Human Investigation Committee of Yale University and the Connecticut Department of Public Health. Written informed consent was obtained from each patient following detailed explanation of the study, after which in-person interview and phlebotomy were conducted.
Candidate Gene Selection and Genotyping
Selection of candidate genes was based on possible involvement in pancreatic cancer survival or in associated biological processes. By searching the PUBMED database, we identified reported candidate genes for pancreatic cancer survival, or for response to radiotherapy or to the five chemotherapy agents 5-FU, gemcitabine, erlotinib, capecitabine, and cisplatin/oxaliplatin commonly used in pancreatic cancer. We also searched for relevant articles cited in the references of the reviews.10,12,13 Altogether, 64 genes were manually selected. For our analysis, all candidate-gene single nucleotide polymorphisms (SNPs) that had originally been genotyped in the PanScan study were eligible if the minor allele frequency was at least 0.01. Detailed methods of the genotyping are available elsewhere.9
Statistical Analysis
We used multivariate proportional hazards regression to test the effect of genetic polymorphism on overall survival with adjustment for age, sex, race, disease stage and treatment (surgery, radiotherapy and chemotherapy). Further, we examined whether variation in these genes modified the effects of specific cancer treatment (radiotherapy, 5-FU, gemcitabine, erlotinib, capecitabine, or cisplatin/oxaliplatin). Tests for gene-treatment interactions were limited to the top hit of each gene in the main-effect analysis on cancer survival, in order to decrease the number of tests performed. Both main effects and interaction terms for genetic polymorphism and specific treatment were included in the regression models adjusted for age, sex, race, disease stage, and fact of treatment other than by the one under interaction. Nominal statistical significance was set at a level of 0.05 and all analyses were done using an extended version of the Generalized Linear Interactive Modeling (GLIM) computer program that includes conditional and proportional hazards regression.14 Our analysis was driven by predefined hypotheses and is exploratory, thus we did not perform explicit numerical Bonferroni corrections to the P-values. For the analysis of main effects, even though we used a candidate gene approach, each gene had a variable number of SNPs genotyped in the original PanScan assay. Because it is unclear how many independent tests are involved in the sets of SNPs within each gene, we have chosen to present only the top variant in each gene as a main effect. The given P-values can be regarded as “gene-wide” significant by consideration of how many likely independent haplotype groups describe the gene for 95% of individuals.
Results
Clinical characteristics of the patients are shown in Table 1. Among the 211 patients, 181 (85.8%) had died during the follow-up. The average follow-up of the study subjects was 3.4 years. In total, 66 patients underwent surgery, 145 received radiotherapy and 137 chemotherapy. For the specific chemotherapeutic agents, 26 patients received 5-FU, 98 received gemcitabine. 23 were treated with erlotinib, 20 were given capecitabine and 15 received cisplatin/oxaliplatin.
Table 1. Basic Characteristics of Study Subjects.
| Variable | Case Patients | (%) |
|---|---|---|
| Mean (SD) age at recruitment, y | 67.3 (9.80) | |
| Median (range) age at recruitment, y | 67.9 (37.2-83.9) | |
| ≤60 y | 52 | 24.6% |
| >60 y | 159 | 75.4% |
| Gender | ||
| male | 121 | 57.4% |
| female | 90 | 42.6% |
| Race | ||
| white | 196 | 92.9% |
| non-white | 15 | 7.11% |
| Vital Status at end of Follow-up | ||
| alive | 30 | 14.2% |
| dead | 181 | 85.8% |
| SEER Stage | ||
| in situ | 2 | 0.95% |
| localized | 25 | 11.8% |
| regional | 101 | 47.9% |
| distant | 71 | 33.6% |
| unstaged | 11 | 5.21% |
| Radiation Therapy | ||
| yes | 145 | 68.7% |
| no | 59 | 28.0% |
| not known | 7 | 3.32% |
| Surgery | ||
| yes | 66 | 31.3% |
| no | 145 | 68.7% |
| not known | 0 | 0.00% |
| Chemotherapy | ||
| yes | 137 | 64.9% |
| no | 72 | 34.1% |
| not known | 2 | 0.95% |
| 5-FU | ||
| yes | 26 | 12.3% |
| no | 174 | 82.5% |
| not known | 11 | 5.21% |
| Gemcitabine | ||
| yes | 98 | 46.4% |
| no | 101 | 47.9% |
| not known | 12 | 5.69% |
| Erlotinib | ||
| yes | 23 | 10.9% |
| no | 177 | 83.9% |
| not known | 11 | 5.21% |
| Capecitabine | ||
| yes | 20 | 9.48% |
| no | 180 | 85.3% |
| not known | 11 | 5.21% |
| Cisplatin/Oxaliplatin | ||
| yes | 15 | 7.11% |
| no | 185 | 87.7% |
| not known | 11 | 5.21% |
In the study, 1,186 polymorphisms associated with 64 candidate genes were examined with respect to overall survival. In the multivariate proportional hazards models, 14 of these genes were nominally associated with survival after adjustment for age, sex, race, disease stage and treatments. The allele frequencies and P-value of the top hit in each gene are summarized in Table 2. In particular, the strongest and most significant association was found in the gene DPYD at rs1760217 (P =10−4.60). The DPYD risk CT/TT genotypes were significantly associated with reduced survival (HR = 1.89; 95%CI = 1.39-2.57), compared with the CC genotype. The SNPs rs17091162 at SERPINA3 and rs2231164 at ABCG2 also showed nearly significant associations with survival, with P-values of 0.0013 and 0.0023, respectively. Subjects carrying the AC/CC genotypes at rs17091162 showed reduced survival compared to individuals with the AA genotype: adjusted HR = 1.57 (95%CI = 1.18-2.08), while individuals carrying the rs2231164 AG/GG genotypes exhibited better survival than patients carrying the AA genotype, with adjusted HR of 0.62 (95%CI = 0.45-0.85). Eleven additional genes had main effects of nominal statistical significance: XPA, CHEK1, MMP3, MAPK10, KRAS, GNAS, IQGAP2, SLC29A1, CCND1, ABO and TGM3. The top hit found in ABO, rs2073828, differed from the most significant variant identified in the PanScan study, rs505922.9 Little evidence of association with survival was found for rs505922: HR = 1.16 (95%CI = 0.93-1.44), P=0.20.
Table 2. Survival Hazard Ratios for Polymorphisms in Candidate Genes in Pancreatic Cancer.
| Gene | Chr | No. of SNPs genotyped | Top variant | A1∣A21 | Frequency of A2 in HapMap-CEU | Frequency of A2 in patients | HR2 | 95%CI LL | 95%CI UL | P-value |
|---|---|---|---|---|---|---|---|---|---|---|
| DPYD | 1 | 230 | rs1760217 | C∣T | 0.750 | 0.806 | 1.89 | 1.39 | 2.57 | 10−4.60 |
| SERPINA3 | 14 | 8 | rs17091162 | A∣C | 0.800 | 0.764 | 1.57 | 1.18 | 2.08 | 0.0013 |
| ABCG2 | 4 | 31 | rs2231164 | A∣G | 0.083 | 0.171 | 0.62 | 0.45 | 0.85 | 0.0023 |
| XPA | 9 | 5 | rs3176757 | C∣T | 0.252 | 0.209 | 1.46 | 1.12 | 1.89 | 0.0063 |
| CHEK1 | 11 | 7 | rs2298483 | A∣G | 0.917 | 0.879 | 0.64 | 0.46 | 0.89 | 0.011 |
| MMP3 | 11 | 4 | rs645419 | A∣G | 0.429 | 0.521 | 1.31 | 1.06 | 1.61 | 0.011 |
| MAPK10 | 4 | 68 | rs6815306 | A∣C | 0.607 | 0.581 | 0.74 | 0.59 | 0.93 | 0.011 |
| KRAS | 12 | 18 | rs10842514 | C∣T | 0.400 | 0.491 | 0.77 | 0.63 | 0.95 | 0.012 |
| GNAS | 20 | 12 | rs6123832 | C∣T | 0.450 | 0.374 | 1.33 | 1.05 | 1.67 | 0.017 |
| IQGAP2 | 5 | 112 | rs153317 | C∣T | 0.542 | 0.569 | 1.31 | 1.04 | 1.66 | 0.021 |
| SLC29A1 | 6 | 5 | rs324148 | C∣T | 0.308 | 0.256 | 0.77 | 0.61 | 0.97 | 0.022 |
| CCND1 | 11 | 2 | rs649392 | A∣G | 0.383 | 0.441 | 1.26 | 1.02 | 1.55 | 0.035 |
| ABO | 9 | 19 | rs2073828 | A∣G | 0.568 | 0.600 | 0.77 | 0.61 | 0.99 | 0.039 |
| TGM3 | 20 | 19 | rs214814 | A∣G | 0.908 | 0.863 | 1.41 | 1.01 | 1.97 | 0.040 |
| UGT1A7 | 2 | 46 | rs1604144 | A∣G | 0.683 | 0.756 | 1.28 | 0.99 | 1.65 | 0.055 |
| PMS2 | 7 | 6 | rs2228006 | A∣G | 0.867 | 0.867 | 0.74 | 0.55 | 1.00 | 0.056 |
| MTHFR | 1 | 11 | rs1801133 | C∣T | 0.242 | 0.391 | 1.24 | 0.99 | 1.54 | 0.061 |
| LOH12CR1 | 12 | 29 | rs918115 | C∣T | 0.667 | 0.754 | 1.28 | 0.98 | 1.66 | 0.061 |
| ERCC1 | 19 | 6 | rs11615 | C∣T | 0.650 | 0.621 | 1.22 | 0.97 | 1.52 | 0.082 |
| RAD51 | 15 | 3 | rs2412547 | A∣C | 0.533 | 0.528 | 1.20 | 0.98 | 1.48 | 0.082 |
| XRCC3 | 14 | 4 | rs861539 | C∣T | 0.417 | 0.372 | 1.20 | 0.97 | 1.50 | 0.10 |
| MDM2 | 12 | 2 | rs769412 | A∣G | 0.058 | 0.062 | 1.47 | 0.95 | 2.28 | 0.10 |
| PAPD7 | 5 | 9 | rs274700 | C∣T | 0.858 | 0.876 | 0.76 | 0.55 | 1.05 | 0.11 |
| RRM1 | 11 | 5 | rs1662172 | A∣G | 0.575 | 0.562 | 1.22 | 0.96 | 1.55 | 0.11 |
| EGFR | 7 | 71 | rs2241054 | A∣G | 0.683 | 0.775 | 1.23 | 0.95 | 1.59 | 0.11 |
| CES2 | 16 | 5 | rs11568310 | C∣T | 0.027 | 0.021 | 0.58 | 0.28 | 1.21 | 0.12 |
| EGF | 4 | 19 | rs2298999 | C∣T | 0.600 | 0.562 | 1.18 | 0.95 | 1.46 | 0.13 |
| ABCB1 | 7 | 67 | rs4148749 | C∣G | 0.008 | 0.026 | 1.78 | 0.88 | 3.61 | 0.13 |
| UMPS | 3 | 1 | rs13146 | C∣T | 0.150 | 0.128 | 1.28 | 0.93 | 1.76 | 0.13 |
| TP73 | 1 | 17 | rs3765736 | C∣T | 0.392 | 0.405 | 0.85 | 0.68 | 1.07 | 0.16 |
| SERPINA1 | 14 | 12 | rs6575424 | C∣T | 0.383 | 0.372 | 0.86 | 0.69 | 1.07 | 0.17 |
| NOS3 | 7 | 2 | rs3918227 | A∣C | 0.867 | 0.927 | 0.76 | 0.53 | 1.10 | 0.17 |
| ERCC5 | 13 | 8 | rs4150351 | A∣C | 0.183 | 0.147 | 0.82 | 0.61 | 1.11 | 0.18 |
| IL1RN | 2 | 4 | rs315952 | C∣T | 0.742 | 0.704 | 1.17 | 0.92 | 1.49 | 0.20 |
| TP53 | 17 | 5 | rs2078486 | A∣G | 0.900 | 0.905 | 0.79 | 0.56 | 1.12 | 0.20 |
| GABRA4 | 4 | 17 | rs1512130 | A∣G | 0.642 | 0.671 | 1.16 | 0.92 | 1.46 | 0.22 |
| DNM3 | 1 | 87 | rs10489286 | C∣T | 0.283 | 0.180 | 1.19 | 0.90 | 1.58 | 0.22 |
| ATR | 3 | 13 | rs9867946 | G∣T | 1.000 | 0.981 | 1.67 | 0.69 | 4.04 | 0.22 |
| DCTD | 4 | 19 | rs17074237 | A∣G | 0.117 | 0.062 | 1.35 | 0.85 | 2.14 | 0.22 |
| ERBB2 | 17 | 2 | rs1810132 | C∣T | 0.724 | 0.699 | 0.86 | 0.67 | 1.10 | 0.24 |
| GSTM1 | 1 | 1 | rs3754446 | G∣T | 0.676 | 0.694 | 1.15 | 0.90 | 1.48 | 0.25 |
| MET | 7 | 26 | rs2237717 | C∣T | 0.408 | 0.434 | 1.13 | 0.91 | 1.39 | 0.26 |
| GSTP1 | 11 | 4 | rs6591256 | A∣G | 0.433 | 0.340 | 0.88 | 0.70 | 1.12 | 0.29 |
| CDA | 1 | 12 | rs12404655 | A∣G | 0.208 | 0.220 | 0.86 | 0.65 | 1.15 | 0.30 |
| EPC2 | 2 | 11 | rs6759083 | A∣G | 0.768 | 0.825 | 1.16 | 0.86 | 1.57 | 0.33 |
| CYP3A5 | 7 | 11 | rs1419745 | A∣G | 0.017 | 0.038 | 1.34 | 0.75 | 2.42 | 0.34 |
| CDKN2A | 9 | 5 | rs3731217 | G∣T | 0.850 | 0.879 | 0.86 | 0.63 | 1.18 | 0.37 |
| GGH | 8 | 4 | rs4617146 | C∣T | 0.125 | 0.237 | 0.89 | 0.69 | 1.15 | 0.37 |
| CHRNA3 | 15 | 10 | rs6495309 | C∣T | 0.200 | 0.216 | 0.89 | 0.68 | 1.16 | 0.38 |
| MLH1 | 3 | 8 | rs9857293 | C∣T | 0.983 | 0.983 | 1.51 | 0.57 | 3.95 | 0.39 |
| MSH2 | 2 | 14 | rs10495944 | A∣G | 0.819 | 0.886 | 0.86 | 0.61 | 1.21 | 0.40 |
| XRCC1 | 19 | 7 | rs939461 | A∣C | 0.108 | 0.121 | 0.88 | 0.65 | 1.20 | 0.40 |
| PARP1 | 1 | 9 | rs1805411 | A∣C | 0.825 | 0.846 | 0.89 | 0.67 | 1.19 | 0.44 |
| FGFR4 | 5 | 1 | rs376618 | A∣G | 0.233 | 0.258 | 0.91 | 0.71 | 1.17 | 0.45 |
| ATM | 11 | 11 | rs11212570 | A∣G | 0.850 | 0.934 | 1.18 | 0.75 | 1.87 | 0.46 |
| PTGS2 | 1 | 4 | rs3918304 | A∣G | 0.000 | 0.007 | 0.61 | 0.12 | 3.01 | 0.52 |
| UGT1A1 | 2 | 13 | rs4148328 | C∣T | 0.450 | 0.348 | 1.08 | 0.85 | 1.36 | 0.54 |
| XRCC2 | 7 | 5 | rs3218527 | A∣C | 0.000 | 0.012 | 0.70 | 0.18 | 2.70 | 0.59 |
| TCEB1 | 8 | 7 | rs2433208 | A∣G | 0.083 | 0.116 | 1.09 | 0.76 | 1.57 | 0.65 |
| GPT | 8 | 1 | rs1063739 | A∣C | 0.617 | 0.571 | 0.96 | 0.76 | 1.21 | 0.72 |
| CDC5L | 6 | 7 | rs10484624 | G∣T | 0.000 | 0.028 | 0.91 | 0.43 | 1.90 | 0.79 |
| TYMP | 22 | 1 | rs470119 | A∣G | 0.650 | 0.573 | 0.98 | 0.80 | 1.21 | 0.87 |
| ERCC2 | 19 | 4 | rs1799787 | C∣T | 0.268 | 0.299 | 1.01 | 0.81 | 1.26 | 0.94 |
| TYMS | 18 | 2 | rs2847153 | A∣G | 0.800 | 0.777 | 0.99 | 0.77 | 1.28 | 0.96 |
Allele 1∣Allele 2
Hazard ratio for allele 2 with respect to allele 1 as discussed in text, adjusted for age, sex, race, stage and treatments
Abbreviations: Chr, chromosome; HR, hazard ratio; 95%CI, 95% confidence interval; LL, lower limit; UL, upper limit
We next investigated genetic modification of radiotherapy on cancer survival. Table 3 shows four genes (RRM1, IQGAP2, TP73, and XRCC3) that may interact with radiotherapy, and at least one interaction (RRM1, P=0.00063) is likely to be statistically significant even if adjusted for multiple comparisons inherent in considering 64 gene interactions. Among patients who did not receive radiotherapy, the HR associated with RRM1 rs1662172 AG/GG genotypes was 0.90 (95%CI = 0.67-1.21), compared with the AA genotype. In patients who did receive radiotherapy, the HR associated with rs1662172 AG/GG genotypes was 2.18 (95%CI = 1.44-3.31) indicating poorer survival. Similarly, polymorphisms in IQGAP2, TP73 and XRCC3 were nominally associated with pancreatic cancer survival among patients who underwent radiotherapy, and no significant associations were found in subjects who did not have radiotherapy. For IQGAP2, in patients who received radiotherapy, the HR associated with rs153317 CT/TT genotypes was 2.33 (95%CI = 1.43-3.80), P=0.00074, and this risk of reduced survival is likely statistically significant considering 64 gene interactions.
Table 3. Survival Hazard Ratios for Gene-Radiation Interactions among Radiation-Treated and Untreated Patients in Pancreatic Cancer.
| Gene | Variant | genotypes1 | HR (95%CI)2 | P interaction | |
|---|---|---|---|---|---|
| Treated | Untreated | ||||
| RRM1 | rs1662172 | AA/(AG+GG) | 2.18(1.44-3.31) | 0.90(0.67-1.21) | 0.00063 |
| IQGAP2 | rs153317 | CC/(CT+TT) | 2.33(1.43-3.80) | 1.13(0.80-1.46) | 0.0088 |
| TP73 | rs3765736 | CC/(CT+TT) | 0.60(0.40-0.90) | 1.00(0.76-1.33) | 0.040 |
| XRCC3 | rs861539 | CC/(CT+TT) | 1.74(1.13-2.67) | 1.06(0.82-1.37) | 0.050 |
Reference genotype/risk genotypes
Adjusted for age, sex, race, stage, and treatments except radiation
Finally, we explored gene-chemotherapy interactions on survival. As shown in Table 4, 22 genes showed nominally significant interactions with specific chemotherapeutic agents, but none was unambiguously significant after consideration of the 64 interactions examined. In one gene--MET--survival was reduced among patients with SNP rs2237717 who had taken erlotinib: HR=3.23 (95%CI =1.67-6.24), P=0.00050 for CT/TT genotypes compared to CC. This result is of borderline significance after considering the number of interaction SNPs examined. Additionally, the interaction of TYMS and 5-FU was associated with a low nominal P-value for interaction (P = 0.0012). Among individuals carrying the rs2847153 AG/GG genotypes and not receiving 5-FU, mortality risk did not increase compared with those carrying the AA genotype. However, in subjects carrying AG/GG genotypes and receiving 5-FU treatment, there was a 2.56-fold increased mortality risk (95%CI = 1.33-4.95).
Table 4. Hazard Ratios for Gene-Chemotherapy Interactions among Specific Chemotherapy-Treated and Untreated Patients.
| Drug | Gene | Variant | Genotypes1 | HR (95%CI)2 | P interaction | |
|---|---|---|---|---|---|---|
| Treated | Untreated | |||||
| 5-FU | ||||||
| TYMS | rs2847153 | AA/(AG+GG) | 2.56(1.33-4.95) | 0.83(0.63-1.10) | 0.0012 | |
| ERCC1 | rs11615 | CC/(CT+TT) | 0.51(0.29-0.92) | 1.36(1.07-1.72) | 0.0040 | |
| UGT1A7 | rs1604144 | CC/(CT+TT) | 0.33(0.13-0.85) | 1.43(1.09-1.89) | 0.0055 | |
| XRCC1 | rs939461 | AA/(AC+CC) | 3.95(1.45-10.75) | 0.79(0.57-1.10) | 0.0068 | |
| CDKN2A | rs3731217 | GG/(GT+TT) | 0.29(0.12-0.69) | 0.99(0.69-1.41) | 0.015 | |
| CDKN1A | rs876581 | AA/(AG+GG) | 2.95(1.07-8.13) | 0.81(0.54-1.23) | 0.018 | |
| CHRNA3 | rs6495309 | CC/(CT+TT) | 0.36(0.16-0.84) | 0.97(0.74-1.28) | 0.024 | |
| ERCC5 | rs4150351 | AA/(AC+CC) | 0.43(0.21-0.88) | 0.96(0.68-1.35) | 0.033 | |
| MMP3 | rs645419 | AA/(AG+GG) | 2.46(1.31-4.60) | 1.22(0.98-1.52) | 0.036 | |
| UGT1A1 | rs4148328 | CC/(CT+TT) | 0.57(0.30-1.11) | 1.19(0.92-1.52) | 0.038 | |
| Gemcitabine | ||||||
| GPT | rs1063739 | AA/(AC+CC) | 1.34(0.96-1.87) | 0.71(0.51-0.98) | 0.0076 | |
| GSTP1 | rs6591256 | AA/(AG+GG) | 1.17(0.86-1.59) | 0.64(0.45-0.90) | 0.011 | |
| CDA | rs12404655 | AA/(AG+GG) | 1.22(0.82-1.82) | .61(0.40-0.93) | 0.018 | |
| IQGAP2 | rs153317 | CC/(CT+TT) | 1.05(0.79-1.39) | 1.80(1.26-2.58) | 0.019 | |
| GABRA4 | rs1512130 | AA/(AG+GG) | 0.89(0.63-1.25) | 1.49(1.06-2.08) | 0.036 | |
| XRCC2 | rs3218527 | AA/(AC+CC) | 14.8(1.68-130.2) | 0.51(0.10-2.49) | 0.038 | |
| Erlotinib | ||||||
| DCTD | rs17074237 | AA/(AG+GG) | 0.28(0.08-0.97) | 1.96(1.18-3.25) | 0.0013 | |
| MET | rs2237717 | CC/(CT+TT) | 3.23(1.67-6.24) | 1.05(0.84-1.31) | 0.0017 | |
| SLC29A1 | rs324148 | CC/(CT+TT) | 1.38(0.86-2.19) | 0.66(0.51-0.86) | 0.010 | |
| CDKN1A | rs876581 | AA/(AG+GG) | 3.23(0.92-11.35) | 0.86(0.57-1.30) | 0.029 | |
| MSH2 | rs10495944 | AA/(AG+GG) | 2.12(0.78-5.72) | 0.76(0.53-1.10) | 0.049 | |
| Capecitabine | ||||||
| MMP3 | rs645419 | AA/(AG+GG) | 0.44(0.16-1.19) | 1.38(1.12-1.71) | 0.023 | |
| SERPINA3 | rs17091162 | AA/(AC+CC) | 4.35(1.64-11.54) | 1.46(1.09-1.94) | 0.032 | |
| Cisplatin/Oxaliplatin | ||||||
| GGH | rs4617146 | CC/(CT+TT) | 3.70(1.21-11.34) | 0.81(0.62-1.06) | 0.014 | |
| MMP3 | rs645419 | AA/(AG+GG) | 0.56(0.24-1.28) | 1.37(1.11-1.70) | 0.036 | |
Reference genotype/risk genotypes
Adjusted for age, sex, race, stage, and other treatments
Discussion
We compiled from existing literature publications a list of 64 candidate genes, all having evidence of involvement in carcinogenesis-related molecular processes or drug metabolic pathways. Among pancreatic cancer patients in a population-based follow-up study, we then evaluated the effects of these candidate genes on survival and explored effect modification with radiotherapy and chemotherapy, using existing GWAS genotype information. We found three genes that showed statistically significant associations or nearly so with pancreatic cancer survival. We also explored gene-radiotherapy and gene-chemotherapy effects and observed that several genetic polymorphisms appear to interact with specific cancer treatment, thus contributing to the prognosis of pancreatic cancer. To our knowledge, this is the first systematic study of genetic effects and modifiers of radiotherapy and chemotherapy on pancreatic cancer survival.
In this study, we found that a number of genetic markers differed with respect to survival in pancreatic cancer. In particular, DPYD rs1760217 CT/TT genotype was associated with an increased risk of poorer survival. DPYD is located on chromosome 1p22 and encodes the enzyme dihydropyrimidine dehydrogenase (DPD). DPD catalyzes the first degradation step of pyrimidine and fluoropyrimidines.15 Seventy-80% of administered 5-FU is normally degraded in vivo by DPD. DPD deficiency is recognized as an important pharmacogenetic factor in the etiology of severe 5-FU-associated toxicity.16,17 Previous studies have shown that DPD expression and activity in pancreatic tumors were higher than in normal tissues.18 Our study implicates a role of DPYD in the outcome of pancreatic cancer. It is possible that rs1760217 or other genetic variants in DPYD may functionally affect DPD expression or activity in survival. Alpha-1 antitrypsin, encoded by SERPINA3, is a serine protease inhibitor and functions in the inflammatory response. Previous studies show that alpha-1 antitrypsin is overexpressed in pancreatic cancer specimens.19,20 It has also been reported that alpha-1 antitrypsin could be a predictive marker of response to gemcitabine and pancreatic cancer survival.10 Increased serum levels of alpha-1 antitrypsin have been associated with shorter survival in pancreatic cancer.10,21 Our data suggest that genetic polymorphism in SERPINA3 likely influences survival in patients with pancreatic cancer, adding to the prognostic importance of this gene.
Recent studies have shown that ABCG2 genetic variants are associated with survival in breast, lung, and prostate cancer.22-25 Our results suggest a role of ABCG2 in survival in pancreatic cancer as well. ABCG2 encodes an ATP-binding cassette cellular xenobiotic exporter that has been postulated to play a role in multidrug resistance.22,23 Our finding of improved survival HR = 0.62 associated with the rs2231164 variant may involve reduced function of this gene.
Due to potential interrelatedness of genetic predisposition, treatment response and prognosis,26 we hypothesized that genetic polymorphisms may interact with radiotherapy or chemotherapy and modify survival in pancreatic cancer. Radiotherapy is widely used as a palliative or adjuvant treatment. However, a beneficial effect of radiation on pancreatic cancer remains inconsistent and varies among patients.27-29 In our study, we found two genes (RRM1 and IQGAP2) that showed interactions with radiotherapy. Neither gene demonstrated a main effect on prognosis. However, subjects carrying the RRM1 rs1662172 AG/GG genotypes and receiving radiation had a significant 2.18-fold increased risk of shorter survival than those subjects carrying the AA genotype. RRM1 encodes one subunit of ribonucleoside-diphosphate reductase, an enzyme required for the synthesis of deoxyribonucleotides involved in DNA synthesis in dividing cells. It is possible that this variant or others in RRM1, under the effects of cellular radiation damage, enhances the ability of tumor cells to survive radiation treatment.
Systemic chemotherapy remains the standard of care in metastatic pancreatic cancer. In addition, adjuvant chemotherapy after resection of localized and locally advanced cancer has been found to improve the outcome of pancreatic cancer. Currently, 5-FU, gemcitabine, erlotinib, capecitabine, and cisplatin/oxaliplatin are commonly used chemotherapeutic agents.30-33 It is now generally understood that choice of drug treatment is best based on individual patient metabolism of and response to chemotherapeutic agents. In our study, we examined whether any of the five chemotherapy agents were associated with survival differences according to variants in any of the 64 genes. We found 22 genes that may interact with specific chemotherapy agents with nominal interaction P-values less than 0.05, but no unambiguous interactions after consideration of the number of comparisons. The strongest evidence for a gene-chemotherapy interaction was observed in TYMS rs2847153 and 5-FU (P for interaction= 0.0012). 5-FU has been used in the treatment of gastrointestinal cancers for more than 50 years, acting principally as a thymidine synthase (TS) inhibitor. The TYMS gene, which is located on 18p11, encodes TS. TS functions as a folate-dependent enzyme that catalyzes the conversion of deoxyuridine-5′-monophosphate to deoxythymidine-5′-monophosphate, representing the only de novo source of thymidine for DNA synthesis.34 Genetic variants in this gene have been found to influence patient response to 5-FU. A previous study using an ex-vivo lymphoblastoid cell-line model system showed that TYMS rs2847153 was significantly associated with 5-FU cytotoxicity.35 Our results further support the involvement of TYMS with 5-FU and their interactive effect on clinical outcome in pancreatic cancer.
Erlotinib is an EGFR small molecule tyrosine kinase inhibitor, which has been FDA approved to be used in combination with gemcitabine as a first line treatment for advanced pancreatic cancer.36 The pharmacologic effects of erlotinib have not been well characterized in patients with pancreatic cancer. In our study, five genes have potential interactions with erlotinib. Among these, the strongest evidence points to DCTD, which has been previously reported as a pharmacogenetic marker for gemcitabine37,38 and catalyzes the nucleotide substrate for TS; and MET, which is a proto-oncogene receptor tyrosine kinase and is associated with tumor growth and metastasis.39
Overall, our study has a number of strengths. First, the case patients in the current analysis were included in the PanScan study, which completed a GWAS involving 550,000 SNPs across the human genome. Taking advantage of the rich genotyping information provided by the GWAS, we used a candidate-gene strategy to highlight a number of genes with respect to survival in pancreatic cancer. GWAS provides a new tool for discovery and exploration of survival differences in pancreatic cancer. Second, with detailed treatment information, our study has systematically explored genetic modulation of radiotherapy and chemotherapy on prognosis in pancreatic cancer. If validated, our findings may help to direct individualized decisions for pancreatic cancer treatment in the future. Third, the cases in our study were obtained from a population-based study, generally representative of pancreatic cancer cases in Connecticut. Fourth, all of the genes selected for investigation here have evidence in the literature supporting their involvement in cancer treatment or survival.
This study however has some limitations. Six of the seven SNPs in the genes that we have discussed are intronic, and the other, rs153317 in IQGAP2, is 289 nucleotides downstream (3′) of the gene. Functional effects of these specific SNPs are uncertain and our results could reflect associations with other functional variants in these genes. In addition, cases with very short survival of only a few weeks after diagnosis (approximately 25% of eligible cases) were not included in our study, thus our results may be slightly shifted toward those of individuals with better survival or less aggressive disease. However, our overall mortality of 86% at an average of 3.4 years post-diagnosis suggests that the survival experience of our patients is typical for pancreatic cancer. Also, since our study was hypothesis-generating and exploratory, we did not explicitly perform Bonferroni corrections, though multiple comparisons should be considered. Nevertheless, the results are likely to be of importance because of the biological and etiological context of the genes involved, and a number of the findings even with Bonferroni correction would be statistically significant or nearly so. Even with more than 200 patients, more than half of whom had succumbed to their disease, the limited sample size of our study did not provide enough power to look at three-way interactions. The exploratory results of interactions between genes and specific cancer treatments should therefore be considered cautiously. Detection of significant associations needs replication analysis in independent sample sets with larger sample sizes, as well as functional confirmation of the genetic variation.
In conclusion, the present study identified a potential role of germline genetic polymorphisms on clinical outcome of pancreatic cancer. Genetic factors may not only influence the prognosis of pancreatic cancer, but also contribute to inter-individual differences in patients undergoing radiotherapy and chemotherapy. Our study adds to the growing evidence for future individualized therapy in pancreatic cancer.
Acknowledgments
The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
The cooperation of 30 Connecticut hospitals, including Stamford Hospital, in allowing patient access, is gratefully acknowledged. This study was approved by the State of Connecticut Department of Public Health Human Investigation Committee. Certain data used in this study were obtained from the Connecticut Tumor Registry in the Connecticut Department of Public Health. The authors assume full responsibility for analyses and interpretation of these data. No funders of the study had any involvement in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.
Financial Acknowledgments: Funded by US National Cancer Institute, National Institutes of Health Grant 5R01-CA098870 (HAR) and Contract HHSN261200800001E (SJC), and by the Overseas Study Program of the China Scholarship Council, Beijing, China (HZ).
Financial support: Harvey A. Risch, Stephen J. Chanock
Administrative support: Harvey A. Risch, Stephen J. Chanock, Patricia Hartge
Abbreviations
- CTR
Connecticut Tumor Registry
- GLIM
Generalized Linear Interactive Modeling
- GWAS
Genome-wide association study
- PUBMED
US National Library of Medicine publications citation database
- SEER
Surveillance, Epidemiology and End Results (US Cancer Registry Network)
- SNP
Single nucleotide polymorphism
Footnotes
Authors' Disclosures of Potential Conflicts of Interest: The authors have no potential conflicts of interest.
Author Contributions
Conception and design: Harvey A. Risch, Stephen J. Chanock and Patricia Hartge for the PanScan Consortium, Dhanpat Jain, Mark S. Kidd, M. Wasif Saif
Provision of study materials or patients: Harvey A. Risch, Herbert Yu, Lingeng Lu, Stephen J. Chanock, Patricia Hartge
Collection and assembly of data: Hongmei Zeng, Harvey A. Risch, Stephen J. Chanock and Patricia Hartge for the PanScan Consortium, Herbert Yu
Data analysis and interpretation: Hongmei Zeng, Harvey A. Risch, Lingeng Lu
Manuscript writing: Hongmei Zeng, Herbert Yu, Lingeng Lu, Dhanpat Jain, Mark S. Kidd, M. Wasif Saif, Stephen J. Chanock and Patricia Hartge for the PanScan Consortium, Harvey A. Risch
Final approval of manuscript: Hongmei Zeng, Herbert Yu, Lingeng Lu, Dhanpat Jain, Mark S. Kidd, M. Wasif Saif, Stephen J. Chanock and Patricia Hartge for the PanScan Consortium, Harvey A. Risch
References
- 1.Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2009. CA Cancer J Clin. 2009;59:225–249. doi: 10.3322/caac.20006. [DOI] [PubMed] [Google Scholar]
- 2.Van Cutsem E, Aerts R, Haustermans K, et al. Systemic treatment of pancreatic cancer. Eur J Gastroenterol Hepatol. 2004;16:265–274. doi: 10.1097/00042737-200403000-00004. [DOI] [PubMed] [Google Scholar]
- 3.Dong X, Jiao L, Li Y, et al. Significant associations of mismatch repair gene polymorphisms with clinical outcome of pancreatic cancer. J Clin Oncol. 2009;27:1592–1599. doi: 10.1200/JCO.2008.20.1111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Okazaki T, Jiao L, Chang P, et al. Single-nucleotide polymorphisms of DNA damage response genes are associated with overall survival in patients with pancreatic cancer. Clin Cancer Res. 2008;14:2042–2048. doi: 10.1158/1078-0432.CCR-07-1520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kamikozuru H, Kuramochi H, Hayashi K, et al. ERCC1 codon 118 polymorphism is a useful prognostic marker in patients with pancreatic cancer treated with platinum-based chemotherapy. Int J Oncol. 2008;32:1091–1096. [PubMed] [Google Scholar]
- 6.Li D, Li Y, Jiao L, et al. Effects of base excision repair gene polymorphisms on pancreatic cancer survival. Int J Cancer. 2007;120:1748–1754. doi: 10.1002/ijc.22301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Li D, Frazier M, Evans DB, et al. Single nucleotide polymorphisms of RecQ1, RAD54L, and ATM genes are associated with reduced survival of pancreatic cancer. J Clin Oncol. 2006;24:1720–1728. doi: 10.1200/JCO.2005.04.4206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Petersen GM, Amundadottir L, Fuchs CS, et al. A genome-wide association study identifies pancreatic cancer susceptibility loci on chromosomes 13q22.1, 1q32.1 and 5p15.33. Nat Genet. 2010;42:224–228. doi: 10.1038/ng.522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Amundadottir L, Kraft P, Stolzenberg-Solomon RZ, et al. Genome-wide association study identifies variants in the ABO locus associated with susceptibility to pancreatic cancer. Nat Genet. 2009;41:986–990. doi: 10.1038/ng.429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Strimpakos AS, Syrigos KN, Saif MW. Pharmacogenetics in pancreatic cancer. JOP; Highlights from the 45th ASCO annual meeting; Orlando, FL, USA. May 29-June 2, 2009; 2009. pp. 357–360. [PubMed] [Google Scholar]
- 11.Risch HA, Yu H, Lu L, et al. ABO blood group, Helicobacter pylori seropositivity, and risk of pancreatic cancer: a case-control study. J Natl Cancer Inst. 2010;102:502–505. doi: 10.1093/jnci/djq007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Saif MW. Translational research in pancreatic cancer. JOP; Highlights from the “44th ASCO Annual Meeting”; Chicago, IL, USA. May 30 - June 3, 2008; 2008. pp. 398–402. [PubMed] [Google Scholar]
- 13.Raimondi S, Maisonneuve P, Lowenfels AB. Epidemiology of pancreatic cancer: an overview. Nat Rev Gastroenterol Hepatol. 2009;6:699–708. doi: 10.1038/nrgastro.2009.177. [DOI] [PubMed] [Google Scholar]
- 14.Francis BGM, Payne C. Release 4 manual. Oxford: Clarendon Press; 1993. The GLIM system: generalised linear interactive modelling. [Google Scholar]
- 15.Sobrero AF, Aschele C, Bertino JR. Fluorouracil in colorectal cancer--a tale of two drugs: implications for biochemical modulation. J Clin Oncol. 1997;15:368–381. doi: 10.1200/JCO.1997.15.1.368. [DOI] [PubMed] [Google Scholar]
- 16.Saif MW, Ezzeldin H, Vance K, et al. DPYD*2A mutation: the most common mutation associated with DPD deficiency. Cancer Chemother Pharmacol. 2007;60:503–507. doi: 10.1007/s00280-006-0392-5. [DOI] [PubMed] [Google Scholar]
- 17.van Kuilenburg AB. Dihydropyrimidine dehydrogenase and the efficacy and toxicity of 5-fluorouracil. Eur J Cancer. 2004;40:939–950. doi: 10.1016/j.ejca.2003.12.004. [DOI] [PubMed] [Google Scholar]
- 18.Nakahara O, Takamori H, Tanaka H, et al. Clinical significance of dihydropyrimidine dehydrogenase and thymidylate synthase expression in patients with pancreatic cancer. Int J Clin Oncol. 2010;15:39–45. doi: 10.1007/s10147-009-0008-2. [DOI] [PubMed] [Google Scholar]
- 19.Trachte AL, Suthers SE, Lerner MR, et al. Increased expression of alpha-1-antitrypsin, glutathione S-transferase pi and vascular endothelial growth factor in human pancreatic adenocarcinoma. Am J Surg. 2002;184:642–647. doi: 10.1016/s0002-9610(02)01105-4. discussion 647-648. [DOI] [PubMed] [Google Scholar]
- 20.Orchekowski R, Hamelinck D, Li L, et al. Antibody microarray profiling reveals individual and combined serum proteins associated with pancreatic cancer. Cancer Res. 2005;65:11193–11202. doi: 10.1158/0008-5472.CAN-05-1436. [DOI] [PubMed] [Google Scholar]
- 21.Matsubara J, Ono M, Honda K, et al. Survival prediction for pancreatic cancer patients receiving gemcitabine treatment. Mol Cell Proteomics. 2010;9:695–704. doi: 10.1074/mcp.M900234-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ishikawa T, Nakagawa H. Human ABC transporter ABCG2 in cancer chemotherapy and pharmacogenomics. J Exp Ther Oncol. 2009;8:5–24. [PubMed] [Google Scholar]
- 23.Ross DD, Nakanishi T. Impact of breast cancer resistance protein on cancer treatment outcomes. Methods Mol Biol. 2010;596:251–290. doi: 10.1007/978-1-60761-416-6_12. [DOI] [PubMed] [Google Scholar]
- 24.Muller PJ, Dally H, Klappenecker CN, et al. Polymorphisms in ABCG2, ABCC3 and CNT1 genes and their possible impact on chemotherapy outcome of lung cancer patients. Int J Cancer. 2009;124:1669–1674. doi: 10.1002/ijc.23956. [DOI] [PubMed] [Google Scholar]
- 25.Hahn NM, Marsh S, Fisher W, et al. Hoosier Oncology Group randomized phase II study of docetaxel, vinorelbine, and estramustine in combination in hormone-refractory prostate cancer with pharmacogenetic survival analysis. Clin Cancer Res. 2006;12:6094–6099. doi: 10.1158/1078-0432.CCR-06-1188. [DOI] [PubMed] [Google Scholar]
- 26.Savas S, Liu G. Genetic variations as cancer prognostic markers: review and update. Hum Mutat. 2009;30:1369–1377. doi: 10.1002/humu.21078. [DOI] [PubMed] [Google Scholar]
- 27.Roy R, Maraveyas A. Chemoradiation in pancreatic adenocarcinoma: a literature review. Oncologist. 2010;15:259–269. doi: 10.1634/theoncologist.2009-0272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gutt R, Liauw SL, Weichselbaum RR. The role of radiotherapy in locally advanced pancreatic carcinoma. Nat Rev Gastroenterol Hepatol. 2010;7:437–447. doi: 10.1038/nrgastro.2010.98. [DOI] [PubMed] [Google Scholar]
- 29.Sultana A, Tudur Smith C, Cunningham D, et al. Systematic review, including meta-analyses, on the management of locally advanced pancreatic cancer using radiation/combined modality therapy. Br J Cancer. 2007;96:1183–1190. doi: 10.1038/sj.bjc.6603719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Burris HA, 3rd, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol. 1997;15:2403–2413. doi: 10.1200/JCO.1997.15.6.2403. [DOI] [PubMed] [Google Scholar]
- 31.Cascinu S, Falconi M, Valentini V, et al. Pancreatic cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2010;21 5:v55–v58. doi: 10.1093/annonc/mdq165. [DOI] [PubMed] [Google Scholar]
- 32.Hess V, Salzberg M, Borner M, et al. Combining capecitabine and gemcitabine in patients with advanced pancreatic carcinoma: a phase I/II trial. J Clin Oncol. 2003;21:66–68. doi: 10.1200/JCO.2003.04.029. [DOI] [PubMed] [Google Scholar]
- 33.Heinemann V, Labianca R, Hinke A, et al. Increased survival using platinum analog combined with gemcitabine as compared to single-agent gemcitabine in advanced pancreatic cancer: pooled analysis of two randomized trials, the GERCOR/GISCAD intergroup study and a German multicenter study. Ann Oncol. 2007;18:1652–1659. doi: 10.1093/annonc/mdm283. [DOI] [PubMed] [Google Scholar]
- 34.Johnston PG, Drake JC, Trepel J, et al. Immunological quantitation of thymidylate synthase using the monoclonal antibody TS 106 in 5-fluorouracil-sensitive and -resistant human cancer cell lines. Cancer Res. 1992;52:4306–4312. [PubMed] [Google Scholar]
- 35.Peters EJ, Kraja AT, Lin SJ, et al. Association of thymidylate synthase variants with 5-fluorouracil cytotoxicity. Pharmacogenet Genomics. 2009;19:399–401. doi: 10.1097/FPC.0b013e328329fdec. [DOI] [PubMed] [Google Scholar]
- 36.Moore MJ, Goldstein D, Hamm J, et al. Erlotinib plus gemcitabine compared with gemcitabine alone in patients with advanced pancreatic cancer: a phase III trial of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol. 2007;25:1960–1966. doi: 10.1200/JCO.2006.07.9525. [DOI] [PubMed] [Google Scholar]
- 37.Gilbert JA, Salavaggione OE, Ji Y, et al. Gemcitabine pharmacogenomics: cytidine deaminase and deoxycytidylate deaminase gene resequencing and functional genomics. Clin Cancer Res. 2006;12:1794–1803. doi: 10.1158/1078-0432.CCR-05-1969. [DOI] [PubMed] [Google Scholar]
- 38.Okazaki T, Javle M, Tanaka M, et al. Single nucleotide polymorphisms of gemcitabine metabolic genes and pancreatic cancer survival and drug toxicity. Clin Cancer Res. 2010;16:320–329. doi: 10.1158/1078-0432.CCR-09-1555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Wong HH, Lemoine NR. Pancreatic cancer: molecular pathogenesis and new therapeutic targets. Nat Rev Gastroenterol Hepatol. 2009;6:412–422. doi: 10.1038/nrgastro.2009.89. [DOI] [PMC free article] [PubMed] [Google Scholar]