Given that there are a variety of resistance mechanisms related to androgen receptor (AR) that can confer castration resistance (e.g. AR amplification or constitutive activation via splice variants or ligand-binding domain mutations), they evaluated AR signaling and, despite observing uniform nuclear localization, saw an absence of persistent AR signaling nor any increase in AR expression, the presence of splice variants, or prevalence of point mutations. This suggested to the authors that the mechanisms of resistance triggered after “short term” AR blockade may differ from those observed in CRPC.
Comparison of residual tumors to matched untreated tumors, the authors identified by RNA sequencing revealed significant upregulation of transcriptional programs associated with epithelial-to-mesenchymal transition (EMT) and chromatin modification. EMT markers were similarly upregulated in treated benign tissues, but not in CRPC – although this comparison was in a small number of samples. They clearly showed loss of E-cadherin, a prototypical event although known not to be sufficient to declare EMT, and authors suggest an intermediate state on the E-to-M spectrum. Stemness-related changes were seen in all treated samples irrespective of transformed status and in CRPCs, suggesting EMT is a distinct survival program for near-term adaptive response to AR pathway inhibition.
Insight into a potential mechanism was revealed with whole-genome sequencing on a subset of samples, with the pre-treatment specimens being less commonly sequenced due to less available material for analysis. No consistent trend of nominated driver mutations was observed, limited in part by small numbers. PyClone did not identify any resistant clones in the 4 samples that could be resolved. Taken together, this is supportive of an intrinsic resistance mechanism.
When pursuing any subclonal loss of heterozygosity after treatment among the three subjects with best response, RNA sequencing nominated two candidates, among them SNAI2, known EMT regulator. Furthermore, the copy number of SNAI2 in the pretreatment samples was observed to correlate with response. Cells deficient in SNAI2 were more likely to be lost from a sample after neoadjuvant treatment with LHRH agonist, AR inhibition, and AR synthesis inhibition.
This study is a clever deployment of a “window of opportunity” study and provides further evidence that AR indifference and resistance is not monolithic across tumor, individual, or clinical setting, such as in this study of tripartite AR targeting for a modest interval. Numbers were perhaps too small to adequately use WGS-based analysis to categorize drivers, and, as the authors acknowledge, the known heterogeneity in prostate cancer mean these results should be interpreted with caution. It is intriguing to imagine how such observations could be capitalized upon, such as how to abrogate the transitory intermediate EMT state to increase tumor susceptibility to short-term ADT in the absence of a preexisting susceptibility (e.g. SNAI2 loss), including in the salvage setting or intermittent ADT
Written by: Jones Nauseef, MD, PhD, Assistant Professor of Medicine, Division of Hematology and Medical Oncology at Weill Cornell Medicine, and Assistant Attending physician at NewYork-Presbyterian Hospital
Read the Abstract