Negative regulators of survival and proliferation (NRSPs) identified by CRISPR activation studies in multiple myeloma (MM) vs. other neoplasias: Therapeutic implications

Abstract Background - Methods: Hundreds of cell lines have been examined by genome-sale CRISPR knockout (KO) studies to identify functional drivers of tumors, but only few lines have been systematically examined with CRISPR activation (CRISPRa). We reasoned that CRISPRa studies can identify genes whose activation may suppress tumor cell fitness and serve as negative regulators of tumor cell survival/proliferation (NRSPs). To this end, we performed genome-scale CRISPRa studies (dCas9-P65-HSF transcriptional activation system; Calabrese sgRNA library, typically 6 sgRNAs/gene) in 10 genotypically diverse multiple myeloma (MM) cell lines at multiple timepoints and compared results with similar publicly available or in-house CRISPRa studies in other cancers (n=17; including lymphoma, leukemia, NSCLC, melanoma, ovarian, prostate cancer; with up to 3 lines per tumor type). Results: Based on multiple quantitative metrics, hundreds of genes were recurrently identified as NRSPs (e.g. 500 NRSPs in 5 or more lines). Integrated analyses of our CRISPRa data with molecular annotation for the respective MM lines did not identify any obvious bias in the CRISPRa-based detection of NRSP that relates to e.g., DNA copy number status; chromatin accessibility; or baseline levels of transcript expression. Top recurrent and pronounced NRSPs included several known proapoptotic genes (e.g., death receptors; caspases; BH3 only Bcl-2 family members) or other tumor suppressive molecules (e.g. CDK inhibitors), providing reassuring positive controls. Moreover, many previously unrecognized NRSPs were identified including transcription factors (TFs)/cofactors; chromatin remodeling genes and epigenetic regulators; signaling molecules regulating survival/proliferation; RNA binding proteins; nucleic acid-sensing and antiviral restriction factors; metabolic regulators; E3 ligases and other proteostatic regulators; DNA repair genes; solute carrier transporters, among others. No correlation is currently observed between the pattern of NRSPs for individual MM cell lines and their respective molecular subtype (e.g., t(4;14), t(14;16), t(11;14), etc). Some clusters of NRSPs are shared between MM and non-MM lines; and others are more prominent in MM, e.g. certain plasma cell-related TFs or endoplasmic reticulum regulators. We validated the functional relevance of several NRSPs with individual sgRNAs for CRISPRa; using focused sgRNA libraries for CRISPRa in MM cells in vitro and in NSG mice in vivo. We also performed single cell RNA-seq of MM cells transduced with pools of sgRNAs (CROP-seq) or bulk RNA-seq of MM cells transduced with individual sgRNAs for select NRSPs. Many NRSPs have undetectable/low transcript levels, without recurrent deletions/mutations, in MM cell lines or patient-derived MM cells but are expressed in healthy and malignant cells from different non-hematopoietic tissues. Thus, the NRSP effect of these genes in MM cells may involve lineage-inappropriate activation of molecular programs that are critical for, or at least tolerated in, other lineages but suppress MM cell fitness. Among the top 200 NRSPs with recurrent expression in MM cells, analyses of MMRF CoMMpass data revealed downregulation of ≥1 NRSPs in ~50% of paired MM samples after relapse vs. before treatment, suggesting that suppression of such NRSP genes can contribute, alone or in concert, to enhanced MM cell fitness in clinical relapses. Some NRSPs defined by CRISPRa are potent dependencies by CRISPR KO, including the MM-preferential, lineage-defining, TF PRDM1. CROP-Seq and bulk RNA-seq of MM cells with CRISPRa of PRDM1 (vs. other NRSPs or control sgRNAs) revealed that PRDM1 overexpression is toxic to MM cells by suppressing other dependencies, including IRF4, another lineage-defining TF and master MM dependency. Conclusions: Collectively,our studies have identified a large collection of NRSP genes whose overexpression suppresses tumor cell fitness in MM, with several of them also being relevant for other hematologic neoplasias. Several canonical tumor suppressor genes are also identified by CRISPRa, but most NRSPs have not been previously examined for their potential to suppress tumor cell fitness in MM or other neoplasias. Our study highlights the concept that engineered overexpression of NRSPs has intriguing potential for therapeutic applications in MM and other hematologic neoplasias.