Prostate cancer is the 2nd leading cause of mortality in men in developed countries. For metastatic prostate cancer patients, the 1st-line treatment relies on hormone therapy. However, the efficacy of androgen deprivation therapy remains limited in time and most patients eventually develop castration-resistant prostate cancer (CRPC), while remaining androgen-dependent.

Docetaxel is currently the standard of care for metastatic CPRC. It has been shown that testosterone levels within metastatic tumoral tissue from men receiving hormone therapy were significantly higher than those from primitive tumors of untreated prostate cancers. Among the mechanistic explanations for this observation, it has been shown that CYP17A1, a key enzyme in de novo steroid synthesis localized in testis and adrenal gland, is up-regulated in CRPC metastases. The existence of de novo CYP17A1-dependent androgen biosynthesis at the tumor level has supported the development of novel antiandrogens, including abiraterone acetate (AA), an irreversible CYP17A1 inhibitor. Based on a placebo-controlled phase III trial, demonstrating that abiraterone prolonged overall survival (14.8 vs 10.9 months) and increased PSA response rate (29% vs 6%) in patients with metastatic CRPC who previously received docetaxel, AA was recently approved by the FDA and French Health Authorities. AA is well-tolerated and main toxicities are urinary tract infections (2%) and a syndrome of secondary mineralocorticoid excess characterized by fluid overload, hypertension and hypokaliema (1% to 4% of grade 3-4).

Almost concomitantly, a novel taxane-class cytotoxic agent, cabazitaxel, has proven efficacy in CRPC treatment after failure to docetaxel, and has recently been approved by the FDA and French Health Authority. Although cabazitaxel exhibits a less favorable toxicity profile, this precise context creates a need to dispose of objective individual criteria so as to orientate patients to treatment towards AA or towards cabazitaxel. To this purpose, several approaches are of potential interest for identifying good candidates for a treatment by AA: tumor-specific TMPRSS2-ERG gene fusion measurement, circulating tumor cell analysis, tumoral CYP17A1 expression, analysis of splicing forms of the androgen receptor. However, the clinical relevance of these potential predictive factors remains to be established in this setting.

Pharmacogenetics examines germinal gene polymorphisms likely to influence the pharmacodynamics of anticancer agents. Encouraging results have recently been reported by our group for irinotecan pharmacogenetics with concrete possibilities of individual dose adaptations, and very recently by other investigators for sunitinib pharmacogenetics. Concerning AA, one can hypothesize that tumors with elevated CYP17A1 expression will be more likely to respond better to AA. This hypothesis is indirectly supported by the observation that in CPRC patients receiving AA, PSA-based response is higher in patients with elevated pre-treatment blood concentration of DHEA and androstenedione.

The CYP17A1 gene presents numerous single nucleotide polymorphisms (SNPs), whose frequencies of rare alleles are at least 12%. Their functional impact has been suggested for nine of them, which were linked either to the risk of developing prostate cancer or to survival of prostate cancer patients. So far, no study has examined the links between these polymorphisms and the efficacy of a CYP17A1 inhibitor. Also, relationships with the efficacy of androgen deprivation therapy have recently been reported for SNPs of genes involved in the membrane-transport testosterone and dehydroepiandrosterone, namely SLCO2B1 and SLCO1B3. One can make the hypothesis that gene polymorphisms of these transporters may play a role for the intratumoral concentration of testosterone locally-produced through the mediation of CYP17A1 activity.

To resume, two second-line treatments of metastatic CRPC cancers are currently available, thus is raising the question in practice of which treatment is more appropriate for a given patient. Herein, the present study proposes an original pharmacogenetic approach in order to highlight a relationship between AA activity and patient's genetic profile. Ultimately, this could reveal evidences of genetic predispositions for potentially good responders to AA treatment.

Inclusion Criteria:

Age > 18 years. Histologically confirmed prostate adenocarcinoma. ECOG ≤ 2. Evidence of metastatic disease by the presence of documented locoregional or distant metastases on CT scan of the abdomen and/or pelvis, or bone scintigraphy.

Patients who have had disease progression during or after prior docetaxel chemotherapy regimen, defined as:

Progressive measurable disease : At least a 20% increase in the sum of the longest diameters of measurable lesions over the smallest sum observed -or- the appearance of one or more new measurable lesions as assessed by CT scan. Soft tissue disease progression defined by modified RECIST 1.1 criteria (baseline lymph node size must be ≥ 2.0 cm to be considered target or evaluable lesion).

OR Bone Scan Progression: appearance of 2 or more new lesions on bone scan. OR Increasing serum PSA level: Two consecutive increases in PSA levels documented over a previous reference value obtained at least one week apart are required. If the third PSA value is less than the second, an additional fourth test to confirm a rising PSA is acceptable. A minimum starting value of 2.0 ng/mL is required for study entry.

NOTE: Androgen ablative therapy may have included either medical or surgical castration.

At least one prior chemotherapy regimen of docetaxel. At least 28 days had to have elapsed between the withdrawal of antiandrogens and enrolment, except LH-RH agonist therapy that must be continued throughout this study for patients who were already treated by it.

Hormonal castration confirmed biologically (testosterone < 0.5 ng/ml). Patient with adequate organ function10. Patient able to swallow abiraterone acetate whole as a tablet.

Information delivered to patient and informed consent signed by the patient or legal representative.

Patient affiliated with a health insurance system.

Exclusion Criteria:

Patients already treated with abiraterone acetate. Known hypersensitivity or allergy to abiraterone or any of the excipients Patients suffering from severe or moderate hepatic impairment (Child-Pugh Class B or C), active or symptomatic viral hepatitis or renal impairment.

Any radiation within 28 days prior to study entry. Patient with central nervous system (CNS) metastasis or with history of CNS metastasis.

Patient treated for a cancer other than prostate cancer, with the exception of basal cell carcinoma, within the past 5 years.

Treatment on another therapeutic clinical trial within 28 days before enrolment Prior treatment with novel hormonal agents including enzalutamide, orteronel, ARN509, EPI100 and novel non hormonal treatments including cabozantinib, alpharadin.

Patients with uncontrolled hypertension, heart disease clinically significant (such as myocardial infarction or recent arterial thrombotic events, severe or unstable angina, heart failure class III-IV NYHA (appendix 5) or with a measurement of the cardiac ejection fraction <50%), within 6 months of randomization.

Any significant concurrent medical illness that in the opinion of the Investigator would preclude protocol therapy.

Permanent contraindication to corticosteroids. Patients who have partners of childbearing potential who are not willing to use refusing 2 methods of birth control with adequate barrier protection during the study and for 13 weeks after last study drung administration.

Patient with history of poor compliance or current or past psychiatric conditions or severe acute or chronic medical conditions that would interfere with the ability to comply with the study protocol.

Patient enables to give informed consent.