Cytochrome 450 1B1 (CYP1B1) polymorphisms associated with response to docetaxel in Castration-Resistant Prostate Cancer (CRPC) patients

Prostate cancer is the most common malignancy in men and the second leading cause of cancer death among males in the Western World [1]. Approximately 70 to 80% of patients with advanced prostate cancer respond to medical or surgical castration [2]. When tumours become refractory to androgen withdrawal therapy, most systemic treatments offer modest benefit in terms of overall-survival (OS) [3].

Two multicenter phase III randomized clinical trials, TAX 327 and SWOG 9916, showed a survival advantage in Castration-Resistant Prostate Cancer (CRPC) patients treated with taxane-based chemotherapy [4, 5]. In both trials the rates of prostate-specific-antigen (PSA) response and OS were significantly higher in the docetaxel groups compared to mitoxantrone and prednisone. In particular, PSA responses were 45% and 50%, while OS was 18.9 and 17.5 months in docetaxel-treated patients, in the TAX 327 and in the SWOG 9916 trial, respectively. On the basis of these findings, taxane-based chemotherapy is now considered the standard first-line therapy for CRPC patients.

Despite the relative success in the treatment of these patients, high variability in the clinical response to docetaxel has been observed [6]. This variability can be partially attributed to a poor understanding of inter-individual differences in the pharmacokinetics and pharmacodynamics of docetaxel. Due to the clinical relevance of docetaxel, genetic markers with predictive power to assess inter-patient differences in clinical response are urgently needed.

There is evidence from preclinical and clinical studies that the cellular response to docetaxel is linked to expression and/or activity of cytochrome-P450 1B1 (CYP1B1). Although docetaxel is not directly metabolized by CYP1B1 [7, 8], CYP1B1 overexpression resulted in a significant decrease in docetaxel sensitivity in transfected hamster ovary cells [9]. CYP1B1 was also induced by docetaxel treatment as a mechanism of resistance in several breast cancer cell lines; the same effect was not observed in breast tumours following treatment with other drugs, nor was it observed for docetaxel applied to cell lines derived from tumours without significant hormone-related etiology [10]. Recent evidence suggests that this is probably due to the action of 4-hydroxyestradiol, the major CYP1B1 metabolite. The 4-OHE2-derived quinone has been shown to impair docetaxel antitumour activity through both an interference with the microtubule stabilization induced by docetaxel, and by direct structural alteration of docetaxel [11].

CYP1B1 activity is regulated by several functional non-synonymous single-nucleotide-polymorphisms (SNPs). The 4326C>G SNP (rs1056836), leading to the 432LeuVal (432LV) amino-acid transition, is associated with increased catalytic activity of CYP1B1 [12, 13], while the 142C>G SNP (rs10012), leading to 48ArgGly (48RG) transition, resulted in increased CYP1B1 gene expression [12, 14], without alterations in catalytic properties unless in combination with other functional alleles. Finally, the 4390A>G SNP (rs 1800440), leading to 453AsnSer (453NS) transition, has been associated with decrease in protein expression due to an increase in CYP1B1 degradation [15].

Interestingly, the homozygous variant of the 4326C>G SNP (CYP1B1*3-allele) has been associated with enhanced synthesis of 4-OHE2, increasing the intracellular ratio with the minor CYP1B1 metabolite 2-OHE2 [8], and CRPC patients carrying two copies of CYP1B1*3 had a significantly shorter OS after docetaxel-based therapies [11].

However, no data on response were reported, and further studies are needed to clarify the role of CYP1B1 genotype and docetaxel activity in CRPC patients. For this purpose we retrospectively evaluated the correlation between the 142C>G, 4326C>G and 4390G>A SNPs and clinical outcome in 60 docetaxel-treated CRPC patients.

This study investigated selected CYP1B1 polymorphisms as potential biomarkers of activity/resistance to docetaxel in CRPC, and provides evidence that the CYP1B1 4326C>G polymorphism may be associated with significantly lower response rate and shorter PFS/OS after docetaxel treatment.

Therapeutic options against prostate cancer have been expanded considerably in the last decade by the introduction of docetaxel for the treatment of CRPC patients [4, 5]. Docetaxel is a second-generation drug of taxane family, with marked activity against in vitro and in vivo models of prostate cancer, where it has been shown to induce apoptosis, inhibit angiogenesis and modulate expression of several signaling pathways [19]. The preclinical promise translated in clinical practice and docetaxel is effective as monotherapy and combination therapy across a variety of tumour types, including prostate cancer. The activity of docetaxel against CRPC has been confirmed in the present work by the percentage of PSA response and median OS, which were 48.3% and 20.4 months, respectively.

The reasons for therapeutic success or failure are, however, elusive and few studies have evaluated possible genetic markers to select patients most likely to respond to docetaxel.

A major mechanism of resistance to taxanes may be represented by the mutation/overexpression of tubulin. In particular, β-tubulin-class-I mutations were identified in 33% of 49 tissues from non-small-cell lung cancer (NSCLC) patients, none of whom had an objective response to paclitaxel [20], but we still lack the validation of mutational analysis or gene expression quantification of tubulin-isotypes as predictive tests to tailor treatment with taxanes.

Other studies have evaluated polymorphisms of the multidrug resistance (MDR1) gene, whose product P-glycoprotein functions as an ATP-dependent exporter of several drugs, including paclitaxel and docetaxel, from cells. Subjects homozygous for C in position 3435 (3435C>T) had higher MDR1 mRNA expression in leukocytes than subjects with the TT genotype [21]. This SNP together with the 2677G>A polymorphism, has been associated with better response in NSCLC patients treated with docetaxel-cisplatin chemotherapy, but no significant correlations were found with response in ovarian cancer patients treated with paclitaxel [22, 23].

In CRPC patients receiving docetaxel alone, individuals carrying a diplotype consisting of the 1236C-2677G-3435C MDR1 linked alleles had significantly improved overall survival, but no data were available on response [24].

Assessing germline genetic polymorphisms as either predictive or prognostic markers is very appealing, especially in the CRPC patients. In these patients primary tumours, as well as bone metastasis, are not resected so that the avaibility of tumour material can be problematic. SNPs are inherited genetic variants harboured by all the cells and their analysis can be easily performed in blood and is easier to adopt in the routine clinical setting than tumour expression arrays, which need biopsies of patient's tumours, laser-microdissection and sophisticated infrastructure.

Therefore, in the present study we evaluated functional polymorphisms of CYP1B1, which play an important role in cancer risk and progression as well as in the metabolism of cancers modulated by sex hormones. In particular, estrogen exposure has been implicated in the disease aetiology of prostate cancer, and CYP1B1 is up-regulated in prostate cancer [25]. Previous studies showed that CYP1B1 polymorphisms may be involved in cancer risk, alone or in combination with other factors [8]. In particular, the 4326GG allele was associated with increased cancer incidence in a Caucasian population [17], where the prevalence of this genotype in CRPC patients (34%) was comparable to that observed in the present study (36%).

However, this study focused on the analysis of CYP1B1 polymorphisms in order to establish their possible relationships with clinical outcome in CRPC patients treated with docetaxel.

The primary end-point of this study was the correlation with response and significant response differences were observed according to CYP1B1 4326C>G polymorphism. Patients carrying the CYP1B1 4326CC or the 4326CG genotype experienced positive clinical response in a significantly higher rate than CYP1B1 4326GG patients.

Since previous studies reported a linkage-disequilibrium between the CYP1B1 4326GG and the 412GG polymorphisms [26], we also performed a statistical analysis in patients harbouring the CYP1B1 4326GG and 412GG genotype. These patients experienced a significantly lower rate of response, with respect to patients harbouring the other genotypes.

Of note, the patients carrying both the CYP1B1 4326GG and the 4390AG or 4390GG genotype also experienced a significantly lower response rate with respect to all the patients harbouring the other genotypes. However, the patients carrying the CYP1B1 412GG and the 4390AG or the 4390GG genotype had similar response rate with respect to patients harbouring the other genotypes, suggesting that the 4326GG genotype might be the best pharmacogenetic marker of lower prevalence of response to docetaxel in CRPC patients.

Furthermore, the CYP1B1 4326GG genotype was associated with significantly shorter PFS and OS, whereas no correlations were observed between both the grouped CYP1B1 4326GG and 412GG and the grouped CYP1B1 4326GG and 4390AG or 4390GG genotypes with PFS.

Given the small number (N = 60) of patients enrolled in the study, in order to evaluate whether other prognostic factors could potentially explain their short survival, we performed corrections for multiple comparison and checked carefully several known baseline demographic, pathological and biochemical characteristics that predicted PSA decline and survival in previous studies in CRPC [27–29]. The occurrence of visceral metastasis and anaemia were associated with shorter OS, while age, ECOG-PS, and the docetaxel schedule were not associated with outcome. However, in the multivariate analysis, including all the significant variables from the univariate analysis, the CYP1B1 C4326C genotype remained an independent predictive parameter of death risk.

The underlying molecular mechanisms for the association between CYP1B1 4326GG genotype and the observed clinical outcome to docetaxel are not entirely clear. In view of previous studies demonstrating no correlation between docetaxel pharmacokinetics and CYP1B1 genotype [7], as well as in vitro studies indicating that docetaxel is not directly metabolized by CYP1B1 [30], this association could most likely be caused by indirect interactions. The CYP1B1-mediated 17β-estradiol metabolites, such as the catechol estrogens, quinonal and semiquinonal catechol estrogens, and methoxyestrogens can bind to tubulin at colchicine binding site [31, 32], and can interfere with the microtubule stabilizing effect of docetaxel. This hypothesis is supported by a previous work by Sissung et al. [11], showing that the 4-OHE2-derived estrogen quinone completely abrogates the polymerization of tubuline. CYPY1B1-mediated estrogen metabolites that bind tubulin are more prevalent in the 4326GG genotype, and CRPC patients with this genotype had a significantly shorter OS. These results were further supported by a case report of a man with residual disease after radical prostatectomy, treated successfully with docetaxel. After only two cycles of therapy, a complete remission was obtained and then consolidated with additional cycles of docetaxel and radiotherapy. Prospective genetic analysis showed that this patient was heterozygous for the 4326CG (432LeuVal) polymorphism [33].

However in the study by Sissung et al. [11] no association was observed between 4326GG genotype and time-to-progression. In contrast, breast cancer patients carrying at least one C allele experienced a significantly shorter PFS following high-dose paclitaxel-based combination chemotherapy [34]. Similarly, in our study, significant correlations were observed between PFS and OS in patients grouped according to the 4326C>G polymorphism. However no reproducible significant associations between genotype and outcome were found for CYP1B1 polymorphism in a recent study in ovarian cancer patients treated with carboplatin and taxane regimens in the Scottish Randomised Trial in Ovarian Cancer phase III trial [23]. These controversial results suggest that pharmacogenetic associations may not always be reproducible when explored in small size series, without standardized unbiased methods, as well as in different settings for tumour type, stage and treatment. Further studies in larger homogeneous populations are warranted in order to understand the utility of candidate polymorphisms in the prediction of outcome in specific clinical settings.

Given the high variant allele frequencies of these polymorphisms, the statistical power to detect associations with these SNPs and response is greatly improved. Furthermore, we used very conservative statistical methods and correction for multiple comparisons to evaluate these relationships and reduce the chances of spurious findings. However, as most previous studies in this cancer setting, the present research is limited by the small sample and residual confounding could have resulted in our findings. Larger prospective studies are needed to validate these preliminary results, after adjustment for known prognostic factors. Furthermore, our data can only suggest a prognostic role for the CYP1B1-4326GG polymorphism, while its predictive role in docetaxel resistance should be evaluated in a prospective randomized study including a non-docetaxel containing arm.

The data from the present investigational study provide the first evidence that the CYP1B1-4236GG polymorphism is linked to the clinical response to docetaxel, and may represent a potential new tool for treatment optimization. These results should prompt to perform multicenter prospective clinical trials, to validate CYP1B1 polymorphisms as possible biomarkers of docetaxel activity in CRPC patients.