(357az) Engineering Improved CRISPRi Repressors for Targeted Gene Regulation in Cancer Cells

The ability to silence gene expression is vital for performing robust whole-genome screens and discovering non-coding transcriptional regulatory motifs in cancer cells. CRISPR interference (CRISPRi) technology has emerged as a powerful tool enabling site-specific transcriptional repression for these applications. CRISPRi works by deploying an RNA-guided, catalytically inactive endonuclease, dCas9, to deliver conjoined transcriptional repressor domains near the transcription start site of target genes. Almost all CRISPRi platforms implement direct fusions of dCas9 with the Krüppel-associated box (KRAB) domain from the human protein KOX1. This domain is traditionally known as KRAB, but here is called KOX1-KRAB to prevent ambiguity. Previous studies demonstrated improved CRISPRi transcriptional repression by combining KOX1-KRAB with additional repressor domains or by using KRAB domains from other human proteins. Despite this progress, CRISPRi still suffers from incomplete repression, high performance variability across cell lines, and general overreliance on the KOX1-KRAB effector.

To address these limitations, we sought to engineer enhanced CRISPRi systems by discovering novel repressor fusions and characterizing their mechanisms of action. We hypothesize that incorporating several repressors with complimentary transcriptional modalities can lead to more efficient silencing of target genes. To explore this concept, we first identified putative transcriptional repressor domains and assessed their activity using a reporter assay in HEK293T cells. Eleven novel, non-KRAB protein domains were compatible with the CRISPRi system, and one achieved greater silencing than KOX1-KRAB. Next, we computationally analyzed gene expression patterns of co-factors that interact with the novel CRISPRi effectors we discovered (as well as KOX1-KRAB). Through this, we discovered these domains recruit co-factors from diverse transcriptional regulation pathways with highly variable expression across human cell lines. To evaluate whether combining domains with diverse modalities can synergistically improve CRISPRi activity, we generated a library of bipartite repressors combining top-performing KRAB domains with the novel domains from our initial experiments. Screening in HEK293T cells revealed four unique variants each achieving 20-30% improved gene silencing compared to current CRISPRi platforms. To build upon these findings, we are assembling and screening tripartite repressor domains to evaluate potential synergistic activity with three domains directly fused together in conjunction with dCas9. Furthermore, we are assessing the generalizability of our improved CRISPRi effectors by targeting endogenous genes across several cancer cell lines. We envision these efforts will mark a significant step forward in applying the CRISPRi platform to discover therapeutic targets in cancer cells from many lineages.