Genomic Resistance Patterns to Second-Generation Androgen Blockade in Paired Tumor Biopsies of Metastatic Castration-Resistant Prostate Cancer

Purpose Patients with castration-resistant prostate cancer (CRPC) receive second-generation androgen-deprivation therapy, but frequently experience relapse or do not respond. Understanding the genetic mechanisms of resistance will help to identify strategies and biomarkers that are essential for the next line of therapy. Patients and Methods We analyzed whole exomes of patient-matched pre- and post-treatment tumors from patients with CRPC. These patients had received the secondary androgen-deprivation therapy agent, abiraterone, which suppresses androgens to below castration levels, or enzalutamide, which competitively inhibits the key androgen signaling effector, androgen receptor. Results We observed that abiraterone-resistant tumors harbored alterations in AR and MYC, whereas enzalutamide-resistant tumors gained alterations in cell-cycle pathway genes, such as mutation in cyclin-dependent kinase N2A (CDKN2A) or amplification of CDK6. Experimentally, overexpressing cell-cycle kinases promoted enzalutamide resistance in androgen-sensitive LnCAP cells that was mitigated via CDK4/6 blockade—palbociclib and ribociclib. Conclusion CDK4/6-mediated resistance observed in preclinical experiments suggests that CDK4/6 amplifications may sufficiently promote enzalutamide resistance in CRPC, and that these patients may respond to palbociclib or ribociclib. The overall observations suggest that, in genomically selected advanced CRPC, clinical strategies against abiraterone- or enzalutamide-resistant tumors may require treatment strategies that are tailored to the resistance mechanisms that are specific to those patient subpopulations.


INTRODUCTION
Prostate cancer is among the most prevalent adult malignancies in men. 1 Patients with metastatic prostate cancer receive primary androgen-deprivation therapy (ADT), and whereas many patients achieve a response, almost all develop castration-resistant prostate cancer (CRPC; Fig 1A). For patients with CRPC, the standard of care includes secondgeneration inhibitors of androgen receptor (AR) signaling, including abiraterone 2 and enzalutamide. 3 These agents effectively prolong survival, but all patients eventually develop resistance. Moreover, considering the potential wider usage of abiraterone from recent findings on the benefit of adding abiraterone and prednisone to primary ADT in hormone-sensitive advanced prostate cancer, 4 understanding the resistance mechanisms that are specific to these agents is even more critical.
relationship to treatment resistance has been incompletely characterized.
Whereas genomic studies of metastatic prostate cancer have demonstrated genomic alterations in AR and its pathway, the genomic characterization of paired biopsy samples from living patients with CRPC before secondary ADT initiation and at the time of resistance have been limited as a result of the logistical challenges of obtaining repeated tumor biopsies and tumor heterogeneity in metastatic prostate cancer. Although difficulties in obtaining repeated biopsies persist and may not translate to standard of care, we hypothesize that molecular interrogation of such paired pre-and post-treatment CRPC biopsies provides an opportunity to define how individuals resist therapy with higher precision. Results may complement previous findings, identify genetic events that are specific to abiraterone or enzalutamide resistance, and provide a rationale for combined and sequential therapy to improve patient outcomes.

PATIENTS AND METHODS
Methods and any associated references are available in the Appendix.

RESULTS
We obtained biopsies from patients with CRPC before either abiraterone or enzalutamide, and at the time of radiographic progression, we obtained a second biopsy at a radiographically matched site, when possible ( Fig 1A and Data Supplement). When insufficient tumor material was obtained from the same site or undergoing sampling was unsafe for the patient, we proceeded to examine postresistant tumors of the patient at a distinct site. We next performed whole-exome sequencing for each biopsy along with germline DNA. After assessment of pathology and whole-exome sequencing quality metrics (Appendix) for the 15 patients who were included in this clinical series, results from seven patients were available for analysis (Data Supplement). We also examined clinical information for each patient (Data Supplement), including prostate-specific antigen (PSA) levels ( Fig  1B), treatment history, and radiographic images (Data Supplement). We primarily used therapy duration and changes in PSA level to define clinical response. 10 We confirmed soft tissue progression using Response Evaluation Criteria in Solid Tumors (RECIST v1.1) 11 criteria and bone disease progression using protocol by Prostate Cancer Working Group 2 12 criteria (Appendix). Overall, we classified acquired resistance in patients as an initial demonstration of a PSA response-a 50% decrease in PSA level 12 -and being on therapy for . 6 months, with the remaining patients being intrinsically resistant ( Fig 1B). By this measure, three patients (Abi-1, Enza-1, and Enza-3) exhibited acquired resistance, and four patients (Abi-2, Abi-3, Abi-4, and Enza-2) were intrinsically resistant.
We then performed mutation, copy number, and phylogenetic analyses of these tumors to nominate putative genetic correlates of resistance by therapeutic class (Appendix). In each pre-and posttreatment tumor, we identified focally amplified and mutated genes (Data Supplement). In abiraterone patients, one patient (Abi-2), who was clinically classified as intrinsically resistant, harbored a wellcharacterized AR resistance mutation (L702H) 13,14 in the post-treatment sample that was not detected in the pretreatment sample (0 of 62 reads and 17 of 46 reads in pre-and post-treatment tumors, respectively; Fig 2). In two additional intrinsic patients (Abi-3 and Abi-4), both pre-and  post-treatment samples harbored focal amplification of AR (Fig 3A). Although our observations associate AR with abiraterone resistance, one patient with pre-existing AR focal amplification (Abi-1) demonstrated an initial 50% decrease in PSA level response before ultimately developing resistance ( Fig 1B). Of interest, we detected a focal amplification in chromosome 8q that involved MYC only in the post-treatment sample (Fig 3B). In preclinical studies, MYC overexpression sufficiently promoted   resistance to AR suppression and bicalutamide. 15,16 Our result associates MYC with abiraterone resistance independent of AR status, which suggests that genetic changes beyond AR may also contribute to clinical abiraterone resistance.
We then examined genetic evolution in the context of clinical resistance to enzalutamide. In one patient (Enza-1) with paired biopsy samples that were obtained from the same site (Fig 4A), a P81L mutation in CDKN2A was only detected in the resistant tumor (Figs 4B and 4C and Data Supplement). This is a clinically observed cancer mutation 17 that is also adjacent to a hotspot location (R80). 18 In addition, relative to wild-type CDKN2A, P81L is functionally defective when overexpressed in melanoma cells. 19 The posttreatment tumor from a second enzalutamideresistant patient (Enza-2) had chr7q (spanning CDK6), whereas AR amplification was detected at both time points (Fig 5). CDK6 regulates cell-cycle progression by phosphorylating and inhibiting the tumor suppressor protein, RB. Because the genetic loss of all RB family members promotes the constitutive activation of CDK signaling, we also investigated alterations of RB family proteins (RB1, RBL1, and RBL2; Data Supplement). Neither deletion, nor hotspot mutations were found. In the last acquired-resistance patient (Enza-3), we did not detect alterations in cell-cycle genes or oncogenic pathways that had been previously associated with ADT resistance.
The observation of cell-cycle up-regulation specifically from these enzalutamide-resistant patients suggests the activity of cell-cycle kinases in enzalutamide resistance. We sought to confirm this clinical observation by determining whether overexpression of CDK4/6 kinases promoted resistance in preclinical models. We followed the schematics in Figure 6A and used open reading frames that contained CDK4 or CDK6 to overexpress these genes in enzalutamide-sensitive LnCAP cells. 20,21 LnCAP cells were used to examine the efficacy of enzalutamide in preclinical applications, 20 and have recently been used to study acquired resistance to enzalutamide. 21 After confirming overexpression by immunoblotting (Data Supplement), we mimicked ADT by first culturing each resulting cell line in media that was supplemented with androgenfree media (charcoal-stripped serum [CSS]) for 3 days and, subsequently, in both CSS and enzalutamide. We observed significant differences in ADT proliferation, as CDK4/6-expressing cells continued to proliferate, whereas luciferaseexpressing negative control cells failed to do so (P , .005; two-tailed t test; Fig 6B).
In combination with the estrogen inhibitor letrozole, the CDK4/6 inhibitor palbociclib has recently been approved by the US Food and Drug Administration for treatment of estrogen receptor-positive breast cancers. 22 Another CDK4/6 inhibitor, ribociclib, has demonstrated efficacy in RB wildtype 23 and AR mutant prostate cancer cells. 24 In two clinical trials, CDK4/6 inhibition was thought to benefit prostate cancers that express wild-type RB (ClinicalTrials.gov identifiers: NCT02059213 and NCT02555189). Specifically, palbociclib has been proposed for use in metastatic prostate cancers in combination with several agents that target androgen biosynthesis (ClinicalTrials.gov identifier: NCT02059213), whereas ribociclib has been proposed for use in combination with enzalutamide in metastatic CRPCs that express wild-type RB (ClinicalTrials.gov identifier: NCT02555189). Thus, we hypothesized that patients who experience relapse after enzalutamide or who achieve minimal response to enzalutamide, in whom post-treatment tumors specifically harbor cell-cycle mutations, including CDK4/6, are strong candidates for combination therapies of enzalutamide and CDK4/6 inhibitors. To test the clinical potential of combining ADT with CDK4/6 inhibitors specifically in enzalutamide-resistant CRPCs with CDK4/6 amplifications, we again used our preclinical model in which CDK4/6 sufficiently promoted enzalutamide resistance. Specifically, we examined whether ribociclib or palbociclib could ablate resistance to ADT (CSS and enzalutamide) in CDK4/6-expressing LnCAP cells. Indeed, the originally resistant CDK4/6-expressing cell lines failed to proliferate when cultured in ribociclib or palbociclib in combination with androgen deprivation (CSS and enzalutamide; Fig 6B).
Our experimental results support the rationale for using palbociclib or ribociclib specifically in enzalutamide-resistant metastatic CRPCs that have CDK4/6 amplifications in patient cases of clinical resistance that are either intrinsic or acquired.

DISCUSSION
In summary, we used a paired biopsy approach and have associated the clinical resistance of abiraterone with AR alterations and, in one patient, with MYC. Two enzalutamide-resistant patients harbored aberrations in cell-cycle pathway genes.
Our preclinical data demonstrate that CDK4/6 overexpression sufficiently drove enzalutamide resistance, but this phenotype was abrogated by CDK4/6 inhibitors. Clinically, our results suggest that some abiraterone-resistant patients may benefit from improved AR inhibition, specifically those with AR amplifications or mutations. For enzalutamide-resistant patients, we identify the specific cell-cycle mutations, CDKN2A and CDK4/6, as biomarkers that may predict whether an enzalutamide-resistant patient could benefit from combination therapy that involves CDK4/6 inhibition and enzalutamide. In this study, we do not disambiguate enzalutamide resistance from general ADT resistance; however, other forms of ADT resistance, including AR-splice variants, that promote enzalutamide resistance 22 between cell-free DNA and tumor requires extensive evaluation in patients with prostate cancer.
In summary, although the sample number is small as a result of the difficulty of obtaining matched tissue biopsies in patients with CRPC, to our knowledge this is the first report of genomic changes in pre-and postabirateroneor enzalutamide-treated patients, and upon validation in larger cohorts, our results provide a rationale for the development of new therapeutic approaches.