Pharmacogenetics of irinotecan and oxaliplatin in advanced colorectal cancer
Kweekel, D.M.
Citation
Kweekel, D. M. (2009, May 26). Pharmacogenetics of irinotecan and oxaliplatin in advanced colorectal cancer. Retrieved from https://hdl.handle.net/1887/13820
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden Downloaded from: https://hdl.handle.net/1887/13820
Note: To cite this publication please use the final published version (if applicable).
SUMMARY
&
NEDERLANDSE SAMENVATTING
196
197 SUMMARY
SUMMARY
Colorectal cancer is one of the leading causes of cancer-related death in many industrialised countries. If the tumor is detected early, curative resection is often possible. However, most patients either present with, or develop, distant metastases for which resection is not an option.
These patients will then receive palliative therapy with traditional chemotherapeutic agents such as fluorouracil (and –derivatives), irinotecan and oxaliplatin; newer agents include the VEGF antibody bevacizumab and the EGFR inhibitor cetuximab. Since the introduction of irinotecan and oxaliplatin, median overall survival for colorectal cancer patients has increased to about two years after diagnosis of metastases. However, toxicity is a main concern for both agents. Oxaliplatin and irinotecan have distinct toxicity patterns, of which cumulative neurotoxicity (oxaliplatin) and febrile neutropenia and diarrhea (irinotecan) are typical examples. This thesis focuses on the study of various genetic polymorphisms with regard to the efficacy and toxicity of oxaliplatin and irinotecan in metastatic colorectal cancer (CRC) patients.
In chapter 1, we provide an overview of various candidate polymorphisms that were described in pharmacogenetic association studies in colorectal cancer patients. We developed a fast and reliable method for the simultaneous genotyping of XRCC1 polymorphisms Arg399Gln and Tyr576Ser using the Pyrosequencing technology. With this method, described in chapter 2, it is possible to obtain both genotypes within 3 hours of genomic DNA isolation from blood. A possible limitation of using blood (germline) genotypes in association analyses of drug efficacy is, that the tumor genotype may be different from germline as a result of tumorigenesis. However, this is unlikely if genes are studied that do not contribute to the process of tumor development. Therefore, in chapter 3, we compared the germline genotypes of 11 pharmacological candidate SNPs (single nucleotide polymorphisms) with the genotypes found in tumor tissue. For these SNPs, we found that the use of formalin-fixed, paraffin-embedded tumor tissue is a valid alternative to EDTA-blood, and that it can be used in pharmacogenetic studies of CRC. Also, our findings validate the use of EDTA-blood in CRC treatment efficacy studies, at least for the genetic variations that were described. For the GSTP1 rs1695 SNP, there was a small discrepancy between tumor and blood genotypes (3.0%) that may be attributed to the effects of multiple testing.
In the second part of this thesis, we focus specifically on irinotecan pharmacogenetics.
Irinotecan is a potent drug used in the palliative chemotherapy of metastatic colorectal cancer (MCRC), but it is also highly toxic and potentially lethal. Irinotecan is metabolised in vivo to SN38, which is the active substance. SN38, on its turn, is metabolised to SN38G, which is the inactive glucuronide metabolite. The irinotecan product leaflet includes suggestions for dose adaptation in various patient categories. Chapter 4 gives an overview of both patient characteristics and pharmacogenetic markers associated with irinotecan toxicity. It also provides dose recommendations for specific conditions and the UGT1A1 genotype. In chapter 5, we discuss the influence of a particular genotype, UGT1A1*28, on irinotecan efficacy and susceptibility to toxicity. We hypothezised that patients, homozygous for the low-activity allele of this enzyme, would not only have a higher risk of toxicity (and therefore receive more frequent dose reductions) but also have a higher response rate. However, response rates were found similar for all UGT1A1 genotypes. As expected, the risk of febrile neutropenia
198
was significantly higher in UGT1A1*28 homozygotes, but this difference in toxicity did not result in a lower dose, more dose adaptations or a lower number of chemotherapy courses administered. The main reason for dose reduction or therapy discontinuation in our patient population was found to be severe diarrhea, and not febrile neutropenia. We found no association between severe diarrhea and the UGT1A1 *28 genotype within the therapeutic regimen used in our study.
Although UGT1A1*28 and irinotecan efficacy were not associated, we found that the GSTP1 codon 105 polymorphism Ile>Val was related to time to progression on first-line treatment of irinotecan plus capecitabine therapy. GSTP1 is a polymorphic enzyme involved in the conjugation of glutathione to several drugs, such as oxaliplatin. The GSTP1 enzyme containing a valine residue at position 105 is less active compared to the wild-type protein.
Irinotecan or SN38 are not known substrates of this enzyme, but in vitro experiments suggested that higher GSTP1 expression in the cell nucleus reduces irinotecan cytotoxicity.
In chapter 6, we describe that Ile/Ile patients do not benefit, in terms of time to progression, from addition of irinotecan to first-line therapy with capecitabine. This is in contrast with the efficacy of irinotecan in other GSTP1 genotypes, who have significantly longer time to progression when receiving the irinotecan combination regimen. These findings may be explained either by enzymatic differences of GSTP1 (although irinotecan and SN38 are not known substrates of this enzyme) or by differential activation of the cell cycle regulating JNK-pathway, in which GSTP1 plays an important role.
In the third part of this thesis, the main focus is on oxaliplatin. This platinum compound, in contrast to many other platinum derivatives, is effective in colorectal cancer. In chapter 7 we describe the pharmacokinetics, -dynamics and -genetics of oxaliplatin. Understanding of the mechanism of action and cellular detoxification routes is essential when selecting candidate genes in a clinical pharmacogenetic association study. Polymorphisms in the aforementioned GSTP1 codon 105, and in DNA-repair enzymes such as ERCC1 have been studied in several MCRC patient populations. ERCC1 is a member of the nucleotide excision repair family, and plays an important role in the excision of damaged or inappropriate nucleotides from a DNA-strand. ERCC1 is a highly conserved gene, and (complete) loss of this enzyme seems incompatible with life. Few genetic variations are known in this gene, but one of these is the codon 118 substitution of AAC into AAT. Both codons result in incorporation of the same amino acid, but AAT is not frequently used. The ERCC1 enzyme resulting from codon 118 AAT may have different characteristics, as a result from differences in amino acid incorporation kinetics and hence folding of the protein. In chapter 8 we describe the in vitro experiments we carried out with an ERCC1-negative chinese hamster ovary (CHO) cell line (UV20). UV20 cell lines transfected with ERCC1 plasmids containing either the 118 AAC or AAT sequence were equally sensitive towards oxaliplatin, and also equally sensitive as compared to an ERCC1 positive CHO cell line. Similarly, there were no differences in DNA-repair activity as determined by the COMET-assay. Besides in vitro sensitivity, we also studied the clinical effects of oxaliplatin and the intratumoral ERCC1 expression in MCRC patients. We found no association between the ERCC1 codon 118 (and C8092A) genetic variations and protein expression in the tumors, as determined by immunohistochemistry.
Also, there was no association of these polymorphisms with time to progression or overall survival of MCRC patients receiving oxaliplatin-based chemotherapy.
199 SUMMARY
In chapter 9 we discuss associations of the GSTP1 codon 105 polymorphism with oxaliplatin efficacy and toxicity. Although some studies with gastrointestinal cancers found that homozygote Val-allele carriers have longer overall survival when receiving platinum- based chemotherapy, we found no association of genotype with progression-free or overall survival after start of oxaliplatin therapy. Of all toxicities with a reported incidence of at least 5%, severe vomiting was more frequent in the Val/Val genotype compared to the other patients. Neurotoxicity was studied in patients who had received a cumulative oxaliplatin dose of at least 500mg/m2. A previous study in patients receiving this minimal cumulative dose suggested that Val/Val patients are more at risk of developing grades 3-4 neurotoxicity.
In that particular study, neurotoxicity was determined using an oxaliplatin-specific neurotoxicity scale. However, we found no association between the GSTP1 genotype and the incidence of (grades 3-4) neurotoxicity using the CTC scale. The difference in these results may at least partly be attributed to the use of different neurotoxicity scales. In this chapter, we furthermore argue that patient selection is very important in pharmacogenetic studies, especially when considering second- or third-line therapies. Imbalances amongst genotype groups in, e.g. performance status or number of prior therapies, may influence (progression-free) survival and constitute an important source of bias. Therefore, in our opinion, pharmacogenetic researchers should take care when combining groups of patients with different regimens and/or different lines of therapy, and correct for this type of imbalances when presenting their data.
Chapter 10 describes the results of our study on oxaliplatin efficacy and toxicity using a SNP array. The commercially available array that was used contains 100 SNPs located in 55 different genes, all of which are involved in DNA repair pathways. Using a pre-defined 2-step selection process, we identified two potential new markers for oxaliplatin efficacy in the ERCC5 (excision repair cross-complementing group 5) and the ATM (ataxia telangiectasia mutated) genes, but none for oxaliplatin toxicity. A main concern with this type of study involving relatively few samples is the occurrence of false-negative findings. The risk of false-negative findings can only be minimized by increasing sample size, which, in turn, is not always feasible. The results of this study, although explorative in nature, may nonetheless serve as a basis for new candidate SNP studies of genes located in the various DNA repair pathways, especially ATM and ERCC5.
In the final chapter of this thesis we discuss the findings described in the previous chapters.
We conclude that treatment of MCRC patients with cytotoxic chemotherapy has significantly improved overall survival, but that treatment effects remain largely unpredictable. Efforts are being undertaken aiming to find single markers that are associated with toxicity or anti-tumor effects of both oxaliplatin and irinotecan. Promising developments have been made in the field of the pharmacogenetics of toxicity, of which the test for UGT1A1*28 genotype (and dose adaptation recommendations) are one clear example. Regarding the pharmacogenetics of drug efficacy, GSTP1 genotype may be a promising marker to assess the added value of irinotecan. However, we should not forget the important influence of other patient characteristics such as age, performance score and lever enzyme function.
These characteristics give equally valuable information as compared to pharmacogenetics.
In fact, these parameters need to be considered complementary to each other, and the combination of markers and clinical parameters may enable us to improve our prediction of chemotherapy effects in the individual in the (near) future.
