Abstract 3778: Genome-Scale CRISPRa and CRISPRi screening for lncRNA drivers of prostate cancer progression

Background: Overtreatment of indolent prostate cancer (PC) and delayed treatment of aggressive PC is common due to suboptimal risk stratification tools, thus warranting identification of novel prognostic biomarkers. Although a few long non-coding RNAs (lncRNAs) with biomarker potential in PC are known, the majority of lncRNAs remain uncharacterized. Here, we aimed to identify novel lncRNA biomarker candidates. We hypothesized that strong candidates would have a functional role in driving PC progression in addition to their expression being linked to PC prognosis, and we therefore combined functional CRISPR screening with lncRNA expression profiling of PC patients. Methods: Total RNA sequencing (RNAseq) data was generated from 31 adjacent normal (AN) and 125 tumor samples from 141 clinically localized PC patients, along with 17 primary tumor samples from metastatic PC patients. Raw reads were mapped to the hg38 reference genome and kallisto was used for quantification. CRISPR activating (CRISPRa) and CRISPR interference (CRISPRi) screens were performed in the LNCaP PC cell line stably expressing either dCas9-VP64 or dCas9-KRAB, respectively. Cells were transduced in duplicate with custom single guide RNA (sgRNA) libraries targeting 20,306 and 20,474 lncRNA transcripts of interest using 72,281 and 72,360 sgRNAs (CRISPRa and CRISPRi, respectively). Cells were harvested and DNA extracted from an early (day 4 post-transduction) and a late (day 17-21) timepoint and next-generation sequenced. MAGeCK was used for data analysis. Results: To identify lncRNAs with biomarker potential in PC, we analyzed lncRNA expression in total RNAseq data from 158 PC patients. Using differential expression analysis and cox regression analysis with biochemical recurrence as endpoint, we identified 6,928 lncRNAs with biomarker potential. To investigate if any of these had a functional role in driving PC progression, we performed CRISPRa and CRISPRi screens to assess how lncRNA activation/inhibition affected PC cell proliferation. Based on the screens, lncRNA candidates with the most prominent phenotypes (normalized read count difference >200 and log-fold change >33% between the early and late timepoint for ≥3 sgRNAs in both replicates) were selected for individual validation. This identified 7 (CRISPRa) and 8 (CRISPRi) negative hits (decreased cell proliferation) along with 5 (CRISPRa) and 2 (CRISPRi) positive hits (increased cell proliferation). Individually activated/inhibited LNCa