Gemcitabine and Genetic Polymorphisms in Pancreatic Cancer
In the current study, we evaluated the relationships between genetic polymorphisms involved in the gemcitabine metabolic pathway and clinical outcomes in advanced pancreatic cancer patients treated with first-line gemcitabine-based chemotherapy. The efficacy of gemcitabine-based chemotherapy for advanced pancreatic cancer is still unsatisfactory, so defining predictive markers of clinical outcomes is important. Although there have been efforts to reveal the relationship between variable chemotherapeutic agents and gene polymorphisms, the effects of gene polymorphisms in pancreatic cancer patients treated with gemcitabine have only recently been studied. We found five SNPs in the CMPK1, DCTD, POLS and CDC5L genes that were statistically associated with OS, and six SNPs in the CMPK1, SLC29A1 and SH2D5 genes that were associated with TTP in analyses using a dominant genetic model. We also detected associations between tumor progression and two SNPs in the CMPK1 and SLC28A1 genes in a dominant model analysis. Additionally, in a recessive genetic model analysis, one SNP in the POLS and ESR2 genes was found to be associated with OS and TTP.
We observed that four SNPs in the CMPK1 gene were associated with therapeutic efficacy. In particular, CMPK1 360C>T was associated with OS, TTP and therapeutic response. Patients with the CC genotype showed a longer OS and TTP than those with CT and TT genotypes. In haplotype analyses, the frequencies of the GGCCTT haplotype of the CMPK1 gene, and the GGCCGTGT haplotype of the CMPK1 and DCK genes were higher in the progressive group than in the nonprogressive group. CMPK1, the enzyme that phosphorylates dFdCMP to dFdCDP, plays an essential role in gemcitabine activation; however, there have been few studies on genetic polymorphism of the CMPK1 gene. Among these four SNPs that have shown differences in the frequency according to tumor response and clinical outcomes, three have never been evaluated with respect to gemcitabine therapy in pancreatic cancer patients and one SNP, CMPK1 240G>T, was reported to not significantly affect enzyme activity and protein level in a previous in vitro study. We suggest that the genetic polymorphisms in the CMPK1 gene, especially CMPK1 360C>T, which showed consistent correlations with OS, TTP and tumor progression, would be potential indicators for clinical outcome in patients treated with gemcitabine-based chemotherapy. CMPK1 360C>T is located in the 3′-UTR, which may affect gene transcription and expression of the CMPK1 gene. Studies addressing the effects of CMPK1 360C>T on gene transcription and gene expression will be helpful to understand how this SNP affects therapeutic efficacy.
Gemcitabine enters the cell through members of the nucleoside transporter family. In the present study, we evaluated 12 SNPs in three genes, SLC28A1, SLC28A3 and SLC29A1. Among these SNPs, the frequency differences of two SNPs in SLC28A1 and SLC29A1 genes according to either TTP or tumor progression were observed in univariate analyses. In haplotype analysis, combinations associated with differences in disease progression were also observed. There have been several reports that the expression level of SLC28A1 and SLC29A1 is related to therapeutic response and survival in lung cancer and pancreatic cancer. On the other hand, other previous in vitro studies showed no functional alterations according to five SNPs in SLC28A3 and SLC29A1. In the studies about the relationship between SNPs and therapeutic efficacy in cancer patients, two studies involved pancreatic cancer and the other one involved lung cancer. The therapeutic efficacy was unrelated to rs2290272, rs8187758, rs2242046 and rs2242047 in SLC28A1, and rs10868138 in SLC28A3, but gemcitabine toxicity was related to rs7867504 in SLC28A3. A recent pharmacokinetic study of gemcitabine in patients with pancreatic cancer included several SNPs in the SLC29A1 gene, including SLC29A1 −55+1681G>A, but they did not find a significant difference. The results reported in the literature, including the present study, have been inconsistent. Therefore, genetic polymorphisms of transporter genes in patients treated with gemcitabine should be confirmed through further clinical studies.
We also observed associations between OS and POLS −35+8748T>G (rs274713) and POLS −35+6255A>G (rs274717). In haplotype analysis, CTCGA in TLE4 was more prevalent in the progressive group than in the nonprogressive group. These genes are involved in the regulation of transcription or DNA replication. TLE4 protein is a member of the Gro/TLE family and indirectly modulates the process of transcription through multiple transcription factors. There is only one study exploring how the TLE4 gene affects gemcitabine toxicity, but no studies on the response to gemcitabine therapy. The POLS gene is involved in DNA replication, mitosis and cell division. POLS −35+8748T>G and POLS −35+6255A>G have not been evaluated in terms of tumor response in in vivo or in vitro studies. Only one SNP in POLS, POLS 836+107G>A (rs2279653), was previously shown to be associated with the response to gemcitabine in vitro, however this SNP was not associated with the therapeutic efficacy in our study. These genes are involved in DNA replication, DNA repair and transcription, which are all in a common pathway targeted by a number of chemotherapeutic agents. Hence, changes in these genes might result in decreased tumor response to gemcitabine-based chemotherapy such as the results of a previous study regarding polymorphisms of DNA repair genes.
It was also found that DCTD 491+759C>T is related with decreased OS. The CGATAACA haplotype in the DCTD gene and CTCAACCGATAACA haplotype in the CDA and DCTD genes were more frequent in disease progressive group than nonprogressive group. DCTD is responsible for the inactivation of gemcitabine-active metabolites. Earlier studies have demonstrated a decline in enzyme activity in the presence of DCTD 172A>G (rs35932500) and no associations between DCTD 315T>C (rs4742) and tumor response, toxicity, TTP or OS. These SNPs were initially included in our study, but were excluded from the statistical analysis owing to low call rate or MAF. An association between therapeutic efficacy and three SNPs in SH2D5, CDC5L and ESR2 genes, which were reported to be correlated with gemcitabine in each one study, was also observed. The role of these genes in gemcitabine metabolism has not been clearly defined, therefore further validation studies are needed.
Previous studies on genetic polymorphisms and gemcitabine have mainly focused on the CDA gene, which is involved in inactivation of gemcitabine metabolites. Among six SNPs in the CDA gene finally evaluated in our study, rs2072671, rs602950 and rs532545 have been shown to have clinical significance in previous studies. However, it was not possible to reproduce the results observed in the previous studies. Differences of the haplotype frequencies of only the CDA and DCTD genes were observed according to the tumor progression. There are some limitations in this study. There were differences in clinical characteristics of patients and chemotherapy schedules between studies, which may contribute to the discordant results. This study is a retrospectively designed study that carries some basic limitations including not strictly controlled evaluation schedules and missing clinical variables, compared with the prospective study. Therefore, we have processed data by removing samples without important clinical data such as prechemotherapeutic CA19-9, and therapeutic response data, resulting in modest sample size for the number of SNPs. Because of the limited sample size, some associations may have been missed or noticed only by chance, especially in haplotype analysis, and we could not perform a validation study. In addition, statistical significance was not maintained after the multiple testing corrections, which may be attributed by the large number of SNPs included in this study. Therefore, the possible bias of the current study may make definitive conclusions difficult and have influenced the relationship between genetic polymorphisms and clinical outcomes in patients treated with gemcitabine-based chemotherapy, which needs to be interpreted with caution and to be prospectively validated. Although the relatively small sample size is the major limitation in this retrospective explorative study, this is not a small scale study when considering the low incidence of pancreatic cancer in the Korean population. Large-scale surveys targeting a number of SNPs and prospective cohort studies for validation of our results are needed. Additionally, now there are a lot of chemotherapeutic agents used in combination with gemcitabine and therefore studies focused on the patient group of each combination would be useful.
Discussion
In the current study, we evaluated the relationships between genetic polymorphisms involved in the gemcitabine metabolic pathway and clinical outcomes in advanced pancreatic cancer patients treated with first-line gemcitabine-based chemotherapy. The efficacy of gemcitabine-based chemotherapy for advanced pancreatic cancer is still unsatisfactory, so defining predictive markers of clinical outcomes is important. Although there have been efforts to reveal the relationship between variable chemotherapeutic agents and gene polymorphisms, the effects of gene polymorphisms in pancreatic cancer patients treated with gemcitabine have only recently been studied. We found five SNPs in the CMPK1, DCTD, POLS and CDC5L genes that were statistically associated with OS, and six SNPs in the CMPK1, SLC29A1 and SH2D5 genes that were associated with TTP in analyses using a dominant genetic model. We also detected associations between tumor progression and two SNPs in the CMPK1 and SLC28A1 genes in a dominant model analysis. Additionally, in a recessive genetic model analysis, one SNP in the POLS and ESR2 genes was found to be associated with OS and TTP.
We observed that four SNPs in the CMPK1 gene were associated with therapeutic efficacy. In particular, CMPK1 360C>T was associated with OS, TTP and therapeutic response. Patients with the CC genotype showed a longer OS and TTP than those with CT and TT genotypes. In haplotype analyses, the frequencies of the GGCCTT haplotype of the CMPK1 gene, and the GGCCGTGT haplotype of the CMPK1 and DCK genes were higher in the progressive group than in the nonprogressive group. CMPK1, the enzyme that phosphorylates dFdCMP to dFdCDP, plays an essential role in gemcitabine activation; however, there have been few studies on genetic polymorphism of the CMPK1 gene. Among these four SNPs that have shown differences in the frequency according to tumor response and clinical outcomes, three have never been evaluated with respect to gemcitabine therapy in pancreatic cancer patients and one SNP, CMPK1 240G>T, was reported to not significantly affect enzyme activity and protein level in a previous in vitro study. We suggest that the genetic polymorphisms in the CMPK1 gene, especially CMPK1 360C>T, which showed consistent correlations with OS, TTP and tumor progression, would be potential indicators for clinical outcome in patients treated with gemcitabine-based chemotherapy. CMPK1 360C>T is located in the 3′-UTR, which may affect gene transcription and expression of the CMPK1 gene. Studies addressing the effects of CMPK1 360C>T on gene transcription and gene expression will be helpful to understand how this SNP affects therapeutic efficacy.
Gemcitabine enters the cell through members of the nucleoside transporter family. In the present study, we evaluated 12 SNPs in three genes, SLC28A1, SLC28A3 and SLC29A1. Among these SNPs, the frequency differences of two SNPs in SLC28A1 and SLC29A1 genes according to either TTP or tumor progression were observed in univariate analyses. In haplotype analysis, combinations associated with differences in disease progression were also observed. There have been several reports that the expression level of SLC28A1 and SLC29A1 is related to therapeutic response and survival in lung cancer and pancreatic cancer. On the other hand, other previous in vitro studies showed no functional alterations according to five SNPs in SLC28A3 and SLC29A1. In the studies about the relationship between SNPs and therapeutic efficacy in cancer patients, two studies involved pancreatic cancer and the other one involved lung cancer. The therapeutic efficacy was unrelated to rs2290272, rs8187758, rs2242046 and rs2242047 in SLC28A1, and rs10868138 in SLC28A3, but gemcitabine toxicity was related to rs7867504 in SLC28A3. A recent pharmacokinetic study of gemcitabine in patients with pancreatic cancer included several SNPs in the SLC29A1 gene, including SLC29A1 −55+1681G>A, but they did not find a significant difference. The results reported in the literature, including the present study, have been inconsistent. Therefore, genetic polymorphisms of transporter genes in patients treated with gemcitabine should be confirmed through further clinical studies.
We also observed associations between OS and POLS −35+8748T>G (rs274713) and POLS −35+6255A>G (rs274717). In haplotype analysis, CTCGA in TLE4 was more prevalent in the progressive group than in the nonprogressive group. These genes are involved in the regulation of transcription or DNA replication. TLE4 protein is a member of the Gro/TLE family and indirectly modulates the process of transcription through multiple transcription factors. There is only one study exploring how the TLE4 gene affects gemcitabine toxicity, but no studies on the response to gemcitabine therapy. The POLS gene is involved in DNA replication, mitosis and cell division. POLS −35+8748T>G and POLS −35+6255A>G have not been evaluated in terms of tumor response in in vivo or in vitro studies. Only one SNP in POLS, POLS 836+107G>A (rs2279653), was previously shown to be associated with the response to gemcitabine in vitro, however this SNP was not associated with the therapeutic efficacy in our study. These genes are involved in DNA replication, DNA repair and transcription, which are all in a common pathway targeted by a number of chemotherapeutic agents. Hence, changes in these genes might result in decreased tumor response to gemcitabine-based chemotherapy such as the results of a previous study regarding polymorphisms of DNA repair genes.
It was also found that DCTD 491+759C>T is related with decreased OS. The CGATAACA haplotype in the DCTD gene and CTCAACCGATAACA haplotype in the CDA and DCTD genes were more frequent in disease progressive group than nonprogressive group. DCTD is responsible for the inactivation of gemcitabine-active metabolites. Earlier studies have demonstrated a decline in enzyme activity in the presence of DCTD 172A>G (rs35932500) and no associations between DCTD 315T>C (rs4742) and tumor response, toxicity, TTP or OS. These SNPs were initially included in our study, but were excluded from the statistical analysis owing to low call rate or MAF. An association between therapeutic efficacy and three SNPs in SH2D5, CDC5L and ESR2 genes, which were reported to be correlated with gemcitabine in each one study, was also observed. The role of these genes in gemcitabine metabolism has not been clearly defined, therefore further validation studies are needed.
Previous studies on genetic polymorphisms and gemcitabine have mainly focused on the CDA gene, which is involved in inactivation of gemcitabine metabolites. Among six SNPs in the CDA gene finally evaluated in our study, rs2072671, rs602950 and rs532545 have been shown to have clinical significance in previous studies. However, it was not possible to reproduce the results observed in the previous studies. Differences of the haplotype frequencies of only the CDA and DCTD genes were observed according to the tumor progression. There are some limitations in this study. There were differences in clinical characteristics of patients and chemotherapy schedules between studies, which may contribute to the discordant results. This study is a retrospectively designed study that carries some basic limitations including not strictly controlled evaluation schedules and missing clinical variables, compared with the prospective study. Therefore, we have processed data by removing samples without important clinical data such as prechemotherapeutic CA19-9, and therapeutic response data, resulting in modest sample size for the number of SNPs. Because of the limited sample size, some associations may have been missed or noticed only by chance, especially in haplotype analysis, and we could not perform a validation study. In addition, statistical significance was not maintained after the multiple testing corrections, which may be attributed by the large number of SNPs included in this study. Therefore, the possible bias of the current study may make definitive conclusions difficult and have influenced the relationship between genetic polymorphisms and clinical outcomes in patients treated with gemcitabine-based chemotherapy, which needs to be interpreted with caution and to be prospectively validated. Although the relatively small sample size is the major limitation in this retrospective explorative study, this is not a small scale study when considering the low incidence of pancreatic cancer in the Korean population. Large-scale surveys targeting a number of SNPs and prospective cohort studies for validation of our results are needed. Additionally, now there are a lot of chemotherapeutic agents used in combination with gemcitabine and therefore studies focused on the patient group of each combination would be useful.
Source...