Genotype frequencies in Russian and Yakut patients.
There is substantial interindividual variability in the efficacy and tolerability of anticancer drugs. Such differences can be greater between individuals of different ethnicities. The clinical studies demonstrate that individuals from Asia (East Asia) are more susceptible to the effects of platinum-containing chemotherapies than their Western counterparts. To determine whether population-related genomics (i.e., frequencies of DNA polymorphisms) contribute to differences in patient outcomes, polymorphisms in 109 genes involved mainly in xenobiotic metabolism, DNA repair, the cell cycle, and apoptosis were tested in Russian (Caucasians) and Yakut (North Asians) ovarian cancer patients receiving cisplatin-based chemotherapy. Totally, 232 polymorphisms were genotyped in individual DNA samples using conventional PCR and arrayed primer extension technology. Single nucleotide polymorphisms (SNPs) in more than 30 genes were found to be associated with one or more of clinical end points (i.e., tumor response, progression-free survival, overall survival, and side effects). However, all associations between SNPs and clinical outcomes were specific for each of ethnic group studied. These findings let us to propose the existence of distinctive ethnic-related characteristics in molecular mechanisms determining the sensitivity of patients to platinum drug effects.
- DNA polymorphisms
- ethnic diversity
- ovarian cancer
There is substantial interindividual variability in the efficacy and tolerability of pharmaceuticals, including anticancer drugs. Such differences can be greater between individuals of different ethnicities . Currently, pharmacoethnicity, or ethnic diversity in drug effectiveness and/or toxicity, is an increasingly recognized factor for accounting interindividual variations in drug response . Although the reasons underlying ethnic diversity in drug response are likely multifactorial , the results of numerous population studies suggest that they may be attributed, at least in part, to the interpopulation differences in frequencies of DNA polymorphisms – inherited variations at the DNA sequence level [4, 5, 6]. In terms of FST, the most commonly used measure of population differentiation; the proportion of such differences is 5–13% of total genetic diversity depending on the type of polymorphic markers chosen . The opponents of ethnic-/race-based explorations in pharmacogenomics often consider these portions of variation as non-essential in the context of considerably larger proportions of within population variation which represents the average difference between members of the same population and accounts for 87–95% of total variance [7, 8]. Nevertheless, significant differences in the population prevalence of functionally impaired allelic variants of genes may create a potential for ethnic differences in responses to drugs that are detoxified (transported or targeted) by the proteins that are encoded by those genes [9, 10, 11]. A prominent example how population-based genetic differences can affect the drug response is the significantly greater risk for Stevens-Johnson syndrome and toxic epidermal necrolysis among East/Southeast Asian carbamazepine users, particularly Han-Chinese, Thais, and Malaysians, that has been associated with
Keeping all that in mind, we carried out a comparative study aimed to explore the genetic bases of differences between Asian and Caucasian cancer patients in their sensitivity to the effects of platinum-containing chemotherapy. Platinum-based drugs are among the most widely used cytotoxic agents for the treatment of many types of cancer . The first information about lesser tolerance of Asian patients to standard, approved for Europeans, doses of platinum-containing regimens came from Japanese physicians . In both individual small studies and some common arm trials conducted in Japan and by Southwest Oncology Group, the higher frequency of toxicity, particularly hematologic toxicity, was registered in Asian patients than non-Asians (mostly Caucasians) [1, 17]. Moreover, it was also found that the incidence of toxicity was still higher among Asians even after appropriate dose reduction . Although some comparative pharmacogenetic studies have been conducted, the reasons underlying the higher sensitivity/toxicity of Asians to the systemic platinum-containing therapy are not yet well understood [18, 19, 20]. To assess the effect of population genomics on difference in patient response, we comparatively explored the results of cisplatin-based chemotherapy in Russian (Caucasians) and Yakut (North Asians) ovarian cancer patients. Principal component analysis, performed by us using genotype data of a common set of 125,000 genome-wide SNPs, demonstrated significant differences between gene pools of Asian and non-Asian populations (Figure 1).
The estimates, obtained using the same set of polymorphic markers, showed that a portion of variation accounted for population-related differences, FST, in allele frequency between Russians and Yakuts was as high as 0.08, creating the potential for searching a causative polymorphism(s) with corresponding prevalence in population frequency. In the current study such candidates were searched among 232 polymorphisms from 109 genes involved mainly in xenobiotic metabolism, DNA repair, the cell cycle, and apoptosis.
2. Materials and methods
Ovarian cancer patients were identified and treated between 2003 and 2007 years at the N. N. Blokhin Cancer Research Centre and the Yakutsk Republic Cancer Clinic. Once identified, patients were invited to participate and were enrolled after they signed an informed consent. Detailed procedures of patient enrolment and data collection have been described previously [21, 22]. Briefly, unrelated Russian and Yakut women with morphologically confirmed epithelial ovarian carcinoma, who had received no previous chemotherapy or radiation therapy, were recruited. The upper age limit was 65 years. Exclusion criteria were serious concomitant diseases (diabetes, uncontrolled hypertension, myocardial infarction within the last 6 months, etc.), and clinically significant hearing impairment (grade 2 or higher). To ascertain ethnicity, women completed a questionnaire about their ancestry; only self-described Russian and Yakut patients with no history of interethnic marriages in the past two generations were recruited. Before the initiation of chemotherapy, venous blood samples were obtained for genetic testing. The chemotherapy regimen was intravenous cisplatin (100 mg/m2) plus cyclophosphamide (600 mg/m2) on day 1, every 3 weeks, for a maximum of 6 cycles. Intraperitoneal chemotherapy and radiotherapy were not allowed. Toxicity of the treatment was described according to standard National Cancer Institute Common Toxicity Criteria version 2.0 . All patients were assessed for the maximal grades of nephrotoxicity, ototoxicity, neurotoxicity, emesis, neutropenia, anemia, and thrombocytopenia.
The tumor response was assessed every 2 cycles. After the completion of chemotherapy, the patients were followed-up for disease relapse and survival. Patients with progressive disease were treated with second-line chemotherapy, mostly taxane based. The study protocol and informed consent form were approved by the Ethics Committee of the N. N. Blokhin Cancer Research Centre.
DNA was isolated from the venous blood samples (leukocytes) using a conventional approach including proteinase K treatment with subsequent phenol-chloroform extraction . Some polymorphisms (Table 1) were genotyped using a polymerase chain reaction restriction fragment length polymorphism (RFLP)-based technique or determined directly through evaluation of their PCR product lengths.
|Polymorphism||#rs ID||Genotypes (No. patients)*||P|
|rs3957357||CC (43/60)||CT (49/22)||TT (12/5)||0.0007|
|0/0 (47/28)||+/0** (57/59)||NA||0.0754|
|rs1799735||AGG/AGG (83/75)||AGG/− (16/12)||−/− (5/0)||0.1054|
|GSTM3 Val224Ile||rs7483||Val/Val (39/30)||Val/Ile (57/37)||Ile/Ile (8/20)||<0.0001|
|GSTP1 Ile105Val||rs1695||Ile/Ile (41/67)||Ile/Val (53/17)||Val/Val (10/3)||<0.0001|
|GSTP1 Ala114Val||rs1138272||Ala/Ala (80/84)||Ala/Val (24/3)||―***||0.0001|
|0/0 (18/22)||+/0** (86/65)||NA||0.2123|
|rs11615||TT (43/53)||TC (46/29)||CC (15/5)||0.0146|
|rs3212986||CC (61/52)||CA (37/31)||AA (6/4)||0.9351|
|ERCC2 Asp312Asn||rs1799793||Asp/Asp (34/66)||Asp/Asn (50/19)||Asn/Asn (20/2)||<0.0001|
|ERCC2 Lys751Gln||rs13181||Lys/Lys (28/67)||Lys/Gln (54/18)||Gln/Gln (22/2)||<0.0001|
|XRCC1 Arg194Trp||rs1799782||Arg/Arg (94/69)||Arg/Trp (10/18)||―||0.0398|
|XRCC1 Arg280His||rs25489||Arg/Arg (95/80)||Arg/His (9/6)||His/His (0/1)||0.5007|
|XRCC1 Arg399Gln||rs25487||Arg/Arg (49/40)||Arg/Gln (45/39)||Gln/Gln (10/8)||0.9752|
|TP53 Arg72Pro||rs1042522||Arg/Arg (52/47)||Arg/Pro (40/35)||Pro/Pro (12/5)||0.3733|
|−/− (100/74)||−/ins (4/13)||―||0.0097|
|rs2031920||CC (102/68)||CT (2/18)||―||<0.0001|
|rs6413432||TT (91/67)||TA (12/20)||AA (1/0)||0.0753|
|rs2070676||CC (82/78)||CG (20/9)||GG (2/0)||0.0908|
Other polymorphisms were genotyped using a microarray “DNA repair single nucleotide polymorphism detection test” (version 2, Asper Biotech, Tartu, Estonia). The microarray genotypes 228 SNPs in 106 genes involved in1 DNA repair, cell cycle control, apoptosis, and xenobiotic metabolism. Most of manually genotyped loci were also in the list of the microarray’s polymorphisms and served as controls of genotyping efficiency. To check into account the potential mistakes in genotyping with the microarray , all polymorphisms, which were associated with any clinical endpoints, were additionally tested using the RFLP method.
2.3. Statistical analysis
A permutation exact test, a two-sided Fisher exact test, and a χ2 test were used to determine the relationship between the variables and alleles/genotypes tested. Correlations between survival and genotype or genetic polymorphism were assessed using the Kaplan-Meier product limit method and the log-rank test. The significance of associations was set at P < 0.05 . The statistical analyses were performed using the Statistica software (version 6.0, StatSoft, Inc., Tulsa, OK, USA) or the IBM SPSS Statistics software package (version 19, SPSS, Inc., IBM Company, IBM Corporation, Armonk, NY, USA), GraphPadInStat (version 3.00, GraphPad Software, San Diego, CA, USA), and the PowerMarker software (version 3.0) .
3. Results and discussion
During 2003–2007 years, 104 Russian patients and 87 Yakut patients were enrolled in the study. The median age of patients was 52 and 51 years, respectively. The majority of patients in both groups had stage III disease (72 and 53 women, respectively). Stages I, II, and IV were detected in 14, 6, and 10 Russian patients and 1, 8, and 25 Yakut patients. A total of 21 Russian patients and 11 Yakut patients were receiving adjuvant chemotherapy and were not eligible for evaluation of tumor response (i.e., they had no residual disease after surgery). Overall response rates, comprising complete and partial responders, were 85% in the Russian group and 58% in the Yakut group.
The median progression-free survival (PFS) in the Russian group was 12 months, and the median overall survival (OS) was 55 months. In the Yakut group, both intervals were shorter —8 and 29 months, respectively. However, being adjusted for disease stage, the values became similar to those for the Russian group.
More contrast results were obtained in analysis of occurrence of adverse events. To assess the association between genotype and the toxicity of the treatment used, patients were classified as having good or poor tolerance to treatment (grades 3–4 of neutropenia, grades 2–4 of anemia, grades 2–4 of neuropathy, grades 3–4 of emesis, all grade of thrombocytopenia, nephrotoxicity, and ototoxicity were considered as clinically significant toxicities). Comparison of the frequencies of side effects registered in Russian and Yakut patients confirmed higher toxicity of platinum-based regimens for patients of Asian origin than for Europeans. Particularly, Yakut patients suffered more frequently than Russians from nephrotoxicity and severe emesis (P = 0.027 and P = 0.061, respectively), which both were known to be the most common adverse events observed in regimens using cisplatin .
Genetic testing of the patients from our groups was performed in two stages. At first, we explored the associations between outcomes of a cisplatin-cyclophosphamide regimen in Russian and Yakut ovarian cancer patients and some most common polymorphisms in several genes [21, 22], among the tested genes, were glutathione S-transferase (GST) genes (
One of the most significant associations found in the first part of the study was the correlation between the survival time intervals and
Analysis of genotype distribution in the population groups for toxicities revealed that occurrence of nephrotoxicity and severe emesis in Yakut patients correlated with
|Gene name||#rs ID||Clinical outcomes**|
When genotype distributions in both groups were compared significant differences in population genotype frequencies were noted for 10 of 19 polymorphisms studied but only 5 of them were associated with clinical outcomes (Tables 1 and 2).
Evaluation of the direction/strength of the associations showed that they were not correlated with the differences in population frequencies of corresponding genotypes. For example, the absence of correlation between the
Taken in the context of other data, the obtained results generally supported the role of ethnicity as an additional reason for differences in the outcomes of clinical trials in which the same treatment is used. At the same time, the results of genetic testing suggested that a single genotypic difference was unlikely to account for the observed ethnic variation in toxicity and survival. Moreover, they also suggested that assessing traditionally tested common polymorphisms in GST and DNA repair genes is not enough for relevant description of lesser tolerability of Asians (North/East Asians) to the effects of platinum-containing chemotherapies and further studies involving more polymorphic markers are required.
In the second part of our study, we systematically investigated the associations between patients’ outcomes and SNPs in more than 100 genes using the microarray “DNA repair single nucleotide polymorphism detection test” (version 2) [36, 37]. Like similar genotyping panels, the list of genes tested comprised candidate genes involved in key pathways of cellular response to different drugs [38, 39], including many genes that are related to the cisplatin pathway (platinum pathway) . A total of 213 SNPs from 228 genotyped SNPs were new (i.e., they did not include SNPs from the first stage) and 27 SNPs were associated with one or more of the assessed clinical end points (Table 3).
|Gene name||#rs ID||Clinical outcomes|
Increasing number of polymorphisms yielded an association with tumor response. To assess the association, patients who achieved a complete remission were compared with those without it (i.e., patients with partial response, stable and progressive disease). In the Russian group, a significant difference in complete response was observed according to polymorphism in the
Genotypes of seven SNPs were associated with differences in PFS in the Russian group (Table 3). The most significant SNPs were rs1142345 (A/G) in
Thiopurine S-methyltransferase (TPMT) is a cytosolic methylating enzyme with unknown physiological role . However, this enzyme is known to be able to catalyze the S-methylation of some aromatic and heterocyclic compounds, particularly thio-compounds (e.g., 6-mercaptopurine and 6-thioguanine). Discussing the associations revealed between SNPs in
The tested SNPs in
In addition to
Totally, 16 SNPs were associated with the side effects of chemotherapy. Thirteen such SNPs were revealed in the Russian group and three SNPs in Yakuts (Table 3). A total of 6 of 13 SNPs were associated with an incidence of severe neutropenia in Russian patients. A strong association was estimated for SNP rs1052536 in
Another side effect for which multiple associations were found in the Russian group was anemia. SNPs in
Three SNPs in
The protein products of
Only one gene from the list above was also present among the genes whose polymorphisms were associated with the adverse reactions in Yakut patients, namely,
The second gene whose allelic variants were associated with a side effect of chemotherapy in Yakut patients (i.e., severe emesis) was
Intergroup comparison of genotypes generated with the microarrays revealed substantial differences in population frequencies of alleles and genotypes for many polymorphic markers. More than half of all markers differed significantly in the occurrence of their allelic variants in Russian and Yakut patients.
The proportion of significant genotype frequency differences resembled the results obtained in the first part of the study where a smaller number of polymorphic markers was involved (Table 2). Furthermore, the results of the comparisons of population-related associative spectra were also the same: there were no identical correlations for any of significant polymorphisms. All associations between the polymorphic markers and clinical outcomes were specific for each of the ethnic group studied.
These findings are generally compatible with the results of the HapMap project studying of the toxicity of platinum compounds (i.e., cisplatin and carboplatin) to lymphoblastoid cell lines from three groups of racially different individuals . One can propose that the failure to detect common associations/commonly associated polymorphisms in our two groups was due to distinctive ethnic-related characteristics in the molecular mechanisms determining the sensitivity of patients to platinum drugs. Hence the difference in platinum drug sensitivity might not exclusively depend on the difference in variant frequencies of given polymorphisms. Another, but not exclusive, explanation of the findings could be a limitation of the number of polymorphisms tested and a possible omission of other potentially important markers. The latest may be mainly due to the misunderstanding of molecular phenotype(s) of the particular drug(s) . The more relevant is the molecular phenotype, the higher is the potential to optimize the use of a particular drug. For some drugs, such as fluorouracil, irinotecan, and mercaptopurine, some relevant variants (i.e.,
The importance of DNA repair, particularly nucleotide excision repair, for platinum cytotoxicity is widely accepted . However, the overall contribution of even the most common genetic variants to predictions of response to platinum-based therapy is not yet well established [70, 71]. In principle, the situation with other “canonical” pathways affecting mainly cisplatin pharmacokinetics could be described the same way . Therefore, the role of additional mechanisms that are not directly related to cisplatin cellular processing has also been proposed . The results of our study overrepresented with the associations with polymorphisms in genes for different metabolic enzymes (
In summary, comprehensive exploration of genotypes of polymorphisms in more than 100 genes in ovarian cancer patients from Russian and Yakut ethnic groups, receiving cisplatin-based chemotherapy, revealed pronounced differences in associative spectra between them. Taken in the context of absence of correlations between the associations and polymorphic genotype frequencies, the differences suggest a potential for distinct ethnic-related molecular mechanisms determining the sensitivity of patients to platinum drug effects. The mechanisms are thought to be associated with activity of different metabolic enzymes, including those involved in processing the reactive oxygen species. These genetic findings and differential responses to platinum-based chemotherapy between ethnic groups suggest that future genetic testing may be invaluable not only in predicting chemotherapy response but also in deciding the most appropriate chemotherapy regimen. It may be possible to identify in detail the susceptibility differences to chemotherapy sensitivity at the molecular level and harness this for therapeutic gains.
This study was supported by grants from the Programs “Fundamental Researches for Development of Biomedical Technologies” and “Molecular and Cell Biology” of the Russian Academy of Sciences and Russian Foundation for Basic Research.
- Alternative variant – “genes which are involved in”.