Association of cytochrome P450 genetic polymorphisms with neoadjuvant chemotherapy efficacy in breast cancer patients

Background The enzymes of the cytochrome P450 family (CYPs) play an important role in the metabolism of a great variety of anticancer agents; therefore, polymorphisms in genes encoding for metabolizing enzymes and drugs transporters can affect drug efficacy and toxicity. Methods The genetic polymorphisms of cytochrome P450 were studied in 395 patients with breast cancer by RLFP analysis. Results Here, we studied the association of functionally significant variant alleles of CYP3A4, CYP3A5, CYP2B6, CYP2C8, CYP2C9 and CYP2C19 with the clinical response to neoadjuvant chemotherapy in breast cancer patients. A significant correlation was observed between the CYP2C9*2 polymorphism and chemotherapy resistance (OR = 4.64; CI 95% = 1.01 – 20.91), as well as between CYP2C9*2 heterozygotes and chemotherapy resistance in women with nodal forms of breast cancer and a cancer hereditary load (OR = 15.50; CI 95% = 1.08 – 826.12) when the potential combined effects were examined. No significant association between chemotherapy resistance and the other examined genotypes and the potential combined clinical and tumour-related parameters were discovered. Conclusion In conclusion, CYP2C9*2 was associated with neoadjuvant chemotherapy resistance (OR = 4.64; CI 95% = 1.01 – 20.91) in the population of interest.

toxicity. CYP2C8 and CYP2C9 are considered highly variable genes and have more than 14 and 34 polymorphic alleles, respectively (www.imm.ki.se/CYPalleles). Most of the CYP2C9 polymorphisms are known to decrease the enzymatic activity of the enzyme. The CYP2C8*3, CYP2C9*2 and CYP2C9*3 polymorphic alleles frequently occur among Caucasians and lead to non-synonymous mutations, which result in decreased activity of CYP2C8 and CYP2C9 [7]. The CYP2C8 enzyme is involved in the metabolism of cyclophosphamide, ifosphamide and paclitaxel, while CYP2C9 metabolizes cyclofosphamide, ifosphamide and tamoxifen and activates tegafur [5,8].
At least 28 CYP2C19 variant alleles have been previously described (www.imm.ki.se/CYPalleles), 9 of which encode for inactive enzymes. Mutations in exon 5 (CYP2C19*2) and exon 4 (CYP2C19*3) are the most common polymorphisms. Both CYP2C19*2, which causes a 40nucleotide deletion and a frameshift, and CYP2C19*3, which leads to a premature stop codon, result in the production of a truncated protein without enzymatic activity [9]. CYP2C19 plays a role in the metabolism of the anticancer drugs cyclofosphamide, ifosphamide, tamoxifen and thalidomide [8].
CYP2B6 may have as many as 29 polymorphisms, but only CYP2B6*5 and CYP2B6*7 are likely to be functionally significant and result in decreased enzymatic activity [10,11]. CYP2B6 is involved in the activation of anticancer drugs such as cyclophosphamide and ifosphamide [5].
CYP3A is the P450 cytochrome with the widest range of drug substrates. For the majority of people, it is also the most abundant cytochrome in the liver. CYP3A4 activity has a wide range of individual activity, up to a 40-fold difference, which may be related to the individual's health status, environment, hormones or genetics. Over 30 CYP3A4 alleles have been described, including 18 associated, nonsynonymous mutations (www.imm.ki.se/CYPalleles). The most common allele is CYP3A4*2, which is found in Caucasians and encodes for a protein with decreased activity. A number of upstream polymorphisms have also been detected. The most extensively studied of these is CYP3A4*1B. This polymorphism alters the putative transcriptional regulatory element -nifidipine oxidase specific element (NFSE), which is known to be required for the expression of CYP3A4 [12]. Although the CYP3A4*1B allele was initially shown to result in a 1.5-fold increase in transcription in vitro, subsequent reports have indicated no change in its enzymatic activity. Currently, over 11 different CYP3A5 alleles have been identified. Individuals homozygous for the CYP3A5*3 allele appear to not express a functional version of the CYP3A5 enzyme due to a cryptic splice site that results in the incorporation of intronic sequence in the mature mRNA and the production of a truncated protein due to a premature appearance of termination codon [13]. This is important in anticancer therapies as CYP3A is involved in the metabolism of many drugs, including cyclophpsphamide, ifosfamide, docetaxel, paclitaxel, etoposide, irinotecan, tamoxifen, imatenib, gefenitib and vinca-alkaloids [4].
We studied the association of the functionally significant variant alleles of CYP3A4, CYP3A5, CYP2B6, CYP2C8, CYP2C9 and CYP2C19 with the clinical response to neoadjuvant chemotherapy in breast cancer patients.

Patients
From 1991 to 2007, 395 women (mean age of 52.00 ± 9.89) with a morphologically confirmed diagnosis of breast cancer at stages T 1-4 N 0-3 M 0 were observed at the Tomsk Cancer Research Institute in the Tomsk Scientific Center of the RAMS. Of these, 261 were treated with one of the following neoadjuvant chemotherapies: CMF or CMXeloda (Cyclophosphamide, Methotrexate and either Fluorouracil or Xeloda), FAC (Fluorouracil, Adreamicin and Cyclophosphamide) and CAF or CAXeloda (Cyclophosphamide, Adreamicin and either Fluorouracil or Xeloda). The chemotherapeutic effect was estimated after 2-4 chemotherapy courses through ultrasonic scanning and mammography according to the World Health Organization criteria. Complete remission (CR) was observed when no breast tumour; partial remission (PR) when the reduction in the tumour area was 50-100%; and stable disease (SD) when the tumour area was reduced 0-50, progressive disease (PD) was recorded if the tumour area increased or if a new lesion was detected. Complete remission was achieved in 3% of the patients; the partial remission rate was 47%; the stable disease rate was 46% and 4% of the patients showed progression of the disease. The patients were divided into two groups according to the results of the neoadjuvant therapy; patients classified as having CR, PR or SD formed the positive response group, while patients classified as showing PD made up the negative chemotherapy response group. The local ethical review boards approved the study protocol (Protocol N3 from 08.06.2005), and all patients provided written, informed consent before acceptance into the study in accordance with the Declaration of Helsinki.
The PCR reactions were carried out in a 20 μl volume and contained 1x PCR buffer, 1-2 mM MgCl 2 , 250 μM dNTPs, 0.5 μM primers, 2 U Taq DNA polymerase (Medigen, Russia) and 30 ng of genomic DNA. Thermal cycling was performed using an initial denaturation time of 3 min at 94°C followed by 33 cycles of 15 s at 94°C, 30 s at 55°C and 30 s at 72°C. A terminal extension time of 5 min at 72°C was used. The reaction products were digested with the appropriate restriction endonucleases (SibEnzyme Ltd., Russia) ( Table 1) for 6 hour at 37°C. The fragments were resolved by PAGE on 10% TBE gels and were subsequently stained with ethidium bromide (10 mg/ml) to visualize the bands using the "VersaDoc System" (Bio-Rad, USA).

Statistical analysis
The allelic and genotype frequencies for all examined genes were calculated for descriptive purposes, and correlation analyses were performed to investigate the relationship between the genotypes and the neoadjuvant chemotherapy outcome. The data were analyzed using EpiInfo 6.0. Differences in the distributions of the variables were analyzed by the χ 2 or Fisher's exact tests, where the cell numbers were less than 5. We also calculated the odds ratios (OR), 95% confidence intervals (CI) and the levels of significance. A p value less than or equal to 0.05 was considered statistically significant. All of the genotype distributions were analyzed for Hardy-Weinberg equilibrium.

Results and discussion
The distribution of the cytochrome P450 genotypes and allele frequencies in the breast cancer patients The cytochrome P450 genetic polymorphisms CYP2C8*2, CYP2C8*3, CYP2C9*2, CYP2C9*3, CYP2C19*2, CYP2C19*3, CYP3A4*1B, CYP3A4*2, CYP3A5*3 and CYP2B6*5 were investigated in 395 female breast cancer patients. Table 2 shows the distribution of the cytochrome P450 genotypes and allele frequencies in the patients. The distribution of the genotype frequencies in breast cancer patients is in agreement with the expected frequencies for the majority of the investigated polymorphisms. Significant deviation from Hardy-Weinberg equilibrium was observed for CYP2C19*2 (p = 0.003) and CYP2B6*5 (p = 0.0008). This may be indicative of the functional significance of this locus or the variants being in non-equilibrium, linking this locus to breast cancer.
The frequencies of the CYP2C8*2 and CYP2C8*3 mutant alleles were 0.39% and 8.33% in the investigated group of breast cancer patients. According to the literature, the CYP2C8*2 allele is known to be found only in African Americans, which carry the allele at a frequency of 18%, while the CYP2C8*3 mutant allele occurs primarily in Caucasians at a frequency of 13%. Neither allele has so far been identified in Asians [18,19].
In this group of breast cancer patients the frequency of the CYP2C9*2 and CYP2C9*3 mutant alleles was 10.36% * -fragments of the wild type alleles, fragments of the mutant type alleles parentheses. and 9.90%, which was not significantly different from the number found in the literature for Caucasians, i.e., 8-19% and 0-8.5% [1,7,[20][21][22][23] for CYP2C9*2 and CYP2C9*3, respectively, but it is significantly higher than that found in African Americans, who carry the alleles at a frequency of 1% and 0.5%, respectively. The CYP2C9*2 allele has not been detected in a Chinese population, and the frequency of the CYP2C9*3 allele was 2-2.6% [22,24]. The CYP2C19*2 mutant allele was found at a frequency of 21.9% in this study, which is not significantly different from previous reports in the literature for Caucasians (15 20%) [1,7] and Japanese (26.7%) [25]. The CYP2C19*3 mutant allele frequency was 1.03% in our group of breast cancer patients, which was not different from the value found in  the literature data for Caucasians (0.0%) or Asians (3.75%) [26]. This allele occurs only in heterozygocity. None of the investigated patients were homozygotes for the CYP2C19*3 allele. The CYP2B6*5 mutant allele frequency was 8.1%, which was not significantly lower than the published value for Caucasians (9.1%) and higher than that published for Asians (4.2%). The frequency of the CYP3A4*1B mutant allele was 2.57% in this group of breast cancer patients; the literature reports significant interethnic variations in this allele (2-9% for Caucasians, 35-66% for African Americans and 0% for Taiwanese and Chinese [27]). The CYP3A4*2 mutant allele was not present in the investigated group of breast cancer patients, while in its frequency in Caucasians ranges from 1.1% in a German population to 4.5% in a Portuguese population [28].
The frequency of the CYP3A5*3 polymorphic allele in the total population sampled was 93.75%. This agrees with the previously published data, which reports a total frequency of the CYP3A5*3 allele of 91.7-94.2% in Caucasians [14,27,29] and 66.7-75% in Asians [30].
The cytochrome P450 polymorphisms and the chemotherapy efficacy in the breast cancer patients Table 3 shows the distribution of the polymorphic variants of the analyzed genes in the breast cancer patients along  with the known efficacy of the neoadjuvant chemotherapy for the CYP2 family. The study of the association of the CYP2B6*5, CYP2C8*2, CYP2C8*3, CYP2C9*2, CYP2C9*3, CYP2C19*2 and CYP2C19*3 variants with a negative response to neoadjuvant chemotherapy in breast cancer patients included an odds ratio evaluation indicating the probability of poor chemotherapeutic efficacy in individuals with certain genotypes. The observed distribution of the most common genotypes in breast cancer patients are in Hardy-Weinberg equilibrium with the exception of CYP2B6*5 (p = 0.0198) and CYP19*2 (p = 0.0003). The enrichment of the CYP2B6*5 and CYP19*2 alleles in this group may be due to the population possessing its own pool of alleles, resulting in different frequencies of unfavorable alleles. The risk of an insufficient response to breast cancer neoadjuvant chemotherapy in CYP2C9*2 heterozygotes was 4.64-fold higher (OR = 4.64, p = 0.02) than in patients with the wild type allele. The impaired efficacy of neoadjuvant chemotherapy in patients containing the CYP2C9*2 mutant allele might be a result of the enzymatic activity of CYP2C9, which is involved into the chain reaction responsible for the conversion of the cyclophpsphamide prodrug into an active metabolite [29,30]. The risk of poor neoadjuvant chemotherapy efficacy is much higher in CYP2C8*2 heterozygotes (OR = 24.30) and slightly higher in CYP2C8*3 heterozygotes (OR = 3.08) at a level of statistical significance close to the proposed level (p = 0.08 and p = 0.09, respectively, versus p < 0.05). Furthermore, the CYP2C9*3 (OR = 9.15) and CYP2C19*2 (OR = 6.13) mutant type genotypes and heterozygotic CYP2C19*3 (OR = 4.82) genotype are associated with a low efficacy of neoadjuvant chemotherapy; however, these associations were not found to be statistically significant (p > 0.05). It appears that the CYP2B6*5, CYP2C8*2, CYP2C8*3, CYP2C9*3, CYP2C19*2 and CYP2C19*3 mutant alleles are not a factor in resistance to neoadjuvant chemotherapies in breast cancer patients. Table 4 shows the distribution of the polymorphic variants of the genes studied here in breast cancer patients along with the known efficacy of the neoadjuvant chemotherapy for the CYP3 family. Gene polymorphisms association with neoadjuvant chemotherapy efficacy were analyzed for the CYP3A4*1B and CYP3A5*3 polymorphisms. CYP3A4*2 was not included in the study because no heterozygotes or homozygotes containing the CYP3A4*2 mutant allele were found in the population. For CYP3A4*1B heterozygotes, the risk of a negative response to neoadjuvant chemotherapy is 24.9-fold higher than in carriers of the wild type allele, which has a level of significance (p = 0.08) that approaches the significance threshold (p = 0.05). No significant association between CYP3A5*3 genotypes and chemotherapy resistance were discovered.
The analyses performed here cover the distribution of cytochrome genetic polymorphisms and their association with the known efficacy of neoadjuvant chemotherapy in breast cancer patients known to have a cancer hereditary load (cancer cases in patient's relatives), a clinical form of cancer, degree of malignancy and histological type of tumour. In CYP2C9*2 heterozygotes with a high hereditary load, the risk of tumour resistance to neoadjuvant chemotherapy was 7.6-fold higher than wild type homozygotes (OR = 7.6, p = 0.04) ( Table 5). A similar association was observed in patients with the nodal form of breast cancer. The risk of tumour resistance to chemotherapy for CYP2C9*2 heterozygotes was 6.83-fold higher than in wild type homozygotes (Table 5; OR = 6.83, p = 0.019). In CYP2C9*2 heterozygotes with nodal form of cancer and a cancer hereditary load, the risk of resistance to chemotherapy was 15.5-fold higher than those with the wild type genotype (OR = 15.5, p = 0.02). Patients with other clinical features and genetic variants of the investigated cytochromes showed no statistically important association with neoadjuvant chemotherapy efficacy.