Identification of Genes Promoting Resistance to Enzalutamide

In 2012, the US Food and Drug Administration approved enzalutamide (Xtandi) for the treatment of patients with metastatic castration-resistant prostate cancer who have previously received docetaxel. In most patients, enzalutamide halts progression of the disease for a short period of time, but patients usually relapse, which is associated with poor prognosis. Identifying the genes and pathways that the cancer cells use to overcome enzalutamide are of interest, because new interventions aimed at preventing relapse after enzalutamide therapy would have therapeutic value to late-stage prostate cancer patients. In this study, we took advantage of RNA interference techniques and an in vitro model of castration-resistant prostate cancer cells to identify genes that drive resistance to enzalutamide. Overall design: We transduced one-third of a commercially-available barcoded lentiviral shRNA library (DECIPHER Module 1 from Cellecta) into C4-2B cells. The shRNA library contained shRNA clones targeting approximately 5045 human genes. The shRNA integrated into the chromosomal DNA of the transduced C4-2B cells and stably knocked down messenger RNA transcripts. Our methodology called for the establishment of three populations of transduced C4-2B cells, which were maintained for six days before the cells were harvested and genomic DNA extracted. The first population (hereafter, "Initial") is a control that was maintained in the absence of enzalutamide and DMSO, which is the solvent for enzalutamide. The second population ("EnzNeg") was maintained in the presence of DMSO, but not enzalutamide, to control for the effects of the solvent. The third population ("EnzPos") was a test population maintained in the presence of both DMSO and enzalutamide. Genomic DNA from two replicates from Initial and three replicates from each of EnzNeg and EnzPos were subjected to targeted deep sequencing on an Illumina HiSeq 2500, for a total of eight samples. The targeting protocol specifically promoted the sequencing of the bar-coded shRNA in the genome. In this analysis, we count the number of times each shRNA bar-code was sequenced in the eight samples, and calculate the ratio of the abundance of each shRNA in EnzPos versus both control groups. Our hypothesis is that if a shRNA-targeted gene helps cancer cells to achieve resistance when selected by enzalutamide, we will observe fewer bar-codes for those genes relative to both controls (which translates into more mRNA for the gene in EnzPos than both controls). A given shRNA was considered a “hit” if it showed at least two-fold abundance decrease in enzalutamide-positive relative to both enzalutamide-negative control and initial samples.