Identifying Long Noncoding RNA Dependencies Using CRISPR Interference (CRISPRi)-Based Platform in Multiple Myeloma

To identify therapeutically actionable genetic dependencies we have pursued various approaches to derive a deeper understanding of the oncogenic hallmarks of myelomagenesis. We have studied the long noncoding RNA (lncRNA) landscape in multiple myeloma (MM) and identified a large number of differentially expressed lncRNAs in MM versus normal plasma cells. These lncRNAs presumably drive the tumorigenesis and MM cell growth, and in turn be susceptible to therapeutic intervention. To this end, we have developed and utilized a CRISPR interference (CRISPRi)-based platform for decoding and targeting the lncRNA dependencies (LongDEPs) in MM.

In this study, we have used RNA-seq of patient-derived CD138+ MM cells (n=360) and MM cell lines (n=70) to generate a priority list of 913 expressed intergenic lncRNAs. Then, to systematically interrogate the role of these lncRNAs in MM cell growth, we have performed a CRISPRi viability screen transducing 3 MM cell lines engineered to express a dCAS9-KRAB fusion protein, with a pooled library consisting of 7 sgRNAs against each of the 913 transcription start sites (TSS) and 576 negative control sgRNAs. Relative representation of sgRNAs was assessed by deep sequencing after 3 weeks and analyzed using the MAGeCK robust rank aggregation (RRA) algorithm. The most enriched or depleted sgRNAs were further tested in a secondary CRISPRi viability screen. Focusing on depleted sgRNAs, we have identified >30 unique LongDEPs; which were further validated via an antisense oligonucleotide (ASO)-based loss-of-function study in a panel of MM cell lines (n=11).

A comparative transcritpomic analysis comparing data from 360 newly-diagnosed and clinically-annotated MM patients and 16 healthy donors showed significant upregulation of these LongDEPs in MM patient cells. Of note, specific longDEPs were found selectively upregulated in genetically-defined patient subsets, including high-risk MM carrying t(4;14), 1q gain or del17p. Moreover, at least 18 LongDEPs were identified as independent risk-predictors of clinical outcome in newly-diagnosed MM patients.

The lncRNA RROL was identified as a leading LongDEP, with a dependency score on a par with positive controls such as IRF4 or MYC. This lncRNA is specifically overexpressed in MM patients after disease relapse, and its higher expression in newly diagnosed MM patients could predict a worse clinical outcome. We have validated the essential role of RROL in support of the proliferation and survival of MM cells both in vitro and in vivo in NOD SCID mice, using ASO-based loss-of-function studies. To explain this effect, we have characterized its role in the control of the pro-survival de novo lipogenesis (DNL) pathway via an unbiased lipid profiling and by measuring the incorporation of C ¹⁴-radiolabeled glucose into the lipid pool. Mechanistically, we have shown that RROL promotes the DNL pathway via transcriptional regulation of rate-limiting enzymes including ACC1. Using in vitro (RNA protein pull down) and in cellulo (RNA yeast-3-hibrid) assays, we have identified the transcription factor c-MYC as a relevant protein interactor of RROL. This interaction occurs at the chromatin level and is required for i) MYC occupancy at DNL gene loci (e.g. ACC1), as shown by both ChIP-qPCR and single molecule dual RNA FISH coupled with immunofluorescence; ii) MYC interaction with a number of transcriptional co-activators, including WDR82, as assessed in vitro in 3 MM cell lines using co-immunoprecipitation followed by Mass spectrometry (Co-IP/MS) and in cellulo using the proximity-dependent biotin identification assay (BioID) in Flp-In T-REx cells expressing a FLAG-BirA*-MYC fusion protein. Overall, our data indicate that RROL provides the chromatin scaffold to assemble a transcriptionally activated ribonucleoprotein complex - minimally composed by RROL, MYC and WDR82 - at gene regulatory loci of DNL rate-limiting enzymes.

To develop therapeutic inhibitors of LongDEPs, starting with RROL, we have tested >70 ASOs following a multi-step screening approach. The anti-MM activity of 2 leading compounds was demonstrated in vitro and in vivo in 2 clinically relevant animal models, including a BLI-based orthotopic model.

In conclusion, our work establish LongDEPs as an additional source of genetic dependencies in MM paving the way for their biologic, clinical and therapeutic characterization in this disease context.