Abstract
PRMT5 catalyzes the arginine symmetric demethylation of histone and non-histone proteins. PRMT5 is overexpressed in lymphoma and plays a key role in lymphomagenesis through the repression of several tumor suppressors, including P53, while promoting the expression of lymphoma-drivers, such as c-MYC.
Currently, three PRMT5 inhibitors are being clinically evaluated in B-cell lymphoma patients, however, the mechanism of action of these molecules remain poorly understood. Thus, to identify the mechanism of action of the potent selective PRMT5 inhibitor, GSK3203591 (GSK-591), we performed a genome-wide CRISPR/Cas9 KO screen in the human mantle cell line, Z-138 and a validation CRISPR /Cas9 screen in the DLBCL cell line, OCI-LY19. Our genome-wide CRISPR screen identified 316 sensitizing genes and 89 genes involved in resistance. The top sensitizing gene was TP53, which validated our screens, as it is a well characterized target of PRMT5. Furthermore, we found that TP53 deletion and the hot spot TP53 R248Wmutation are biomarkers of resistance to GSK-591.
RNA-binding protein MUSASHI 2 (MSI2) was identified as the top-ranked driver of resistance to GSK-591. MSI2 plays a key role in hematopoietic stem cells activation, myeloid leukemia and CLL, however, its function in lymphoma remains unknown. MSI2 is overexpressed in DLBCL patients (n=96). Relapsed MCL and DLBCL primary samples express high levels of both PRMT5 and MSI2. Furthermore, MSI2 depletion decreased cell viability and sensitized lymphoma cells to PRMT5 inhibitor in vitro and vivo. In contrast, overexpression of MSI2 reverted sensitivity to GSK-591 in lymphoma cells. Consistent with our genetic studies, inhibition of MSI2 using Ro 08-2750 (Ro) (Minuesa et al. Nat Commun. 2019), conferred sensitivity to GSK-591, and the combination therapy was synergistic inducing cytotoxicity.
In order to identify specific downstream targets of MSI2 that may contribute to resistance to PRMT5 inhibition in lymphoma, we performed MSI2-HyperTRIBE, a recently discovered technology that allows mapping the MSI2 targeting network (Nguyen et al. Nat Commun. 2020). Using MSI2-HyperTRIBE, we were able to identify MSI2 targets in lymphoma cells. Moreover, we found that Ro treatment significantly blocked MSI2 binding activity, while GSK-591 had no effect on MSI2 targets.
To uncover the mechanism involved in the synergy of PRMT5 and MSI2 inhibitors, we performed RNA-sequencing of Z-138 cells treated with GSK-591, Ro or the combination. Gene set enrichment analysis (GSEA) demonstrated a loss in the c-MYC pathway upon drug combination.MSI2 RNA-IP assays showed that MSI2 binds c-MYC and this interaction is disrupted by the combination of GSK-591 and Ro. We observed that the combination of MSI2 and PRMT5 inhibitors does not affect c-MYC mRNA levels but, rather, controls translation. Interestingly, in the human B cell line P493-6 that express a conditional, tetracycline-regulated c-MYC, the depletion of c-MYC significantly increased the anti-proliferative activity of GSK-591 and Ro alone or in combination.
To further investigate whether there are other functionally relevant downstream targets of MSI2 promoting resistance to PRMT5 inhibition, we overlapped the differentially expressed genes upon treatment with GSK-591 and Ro (RNA-Seq), the Ro-direct targets in lymphoma (MSI2-HyperTRIBE) and the resistance genes to PRMT5 inhibition (whole-genome CRISPR screen) and we identified the anti-apoptotic protein BCL-2 as a common target of the PRMT5-MSI2 axis. We established first that the combination of GSK-591 and Ro blocked the binding of MSI2 to BCL-2 mRNA. We also found that the drug combination significantly decreased BCL-2 mRNA and protein abundance. Furthermore, Z-138 and OCI-LY19 BCL-2 KO cells are more sensitive to GSK-591 and Ro alone or in combination. Using venetoclax, we found that BCL-2 inhibition enhanced GSK-591 activity in vitro and in vivo. Thus, these data confirmed that BCL-2 is a key driver of resistance to PRMT5 inhibition.
Overall, our study uncovered a novel oncogenic axis, PRMT5/MSI2/c-MYC/BCL-2 that drives resistance to PRMT5-targeted therapy in lymphoma. We demonstrated that TP53 LOF and MSI2 expression could be used as biomarkers for patient stratification. Moreover, we proposed two novel drug combination strategies, with venetoclax or a MSI2 inhibitor, to be considered in further clinical studies with PRMT5 inhibitors.
Disclosures
Barbash: GlaxoSmithKline: Current Employment, Current equity holder in publicly-traded company. Batlevi: Viatris: Current holder of individual stocks in a privately-held company; Juno/Celgene: Consultancy; Pfizer: Current holder of individual stocks in a privately-held company; Bayer: Research Funding; ADC Therapeutics: Consultancy; TG Therapeutics: Consultancy; Medscape: Honoraria; BMS: Current holder of individual stocks in a privately-held company; Memorial Sloan Kettering Cancer Center: Current Employment; Kite Pharma: Consultancy; Life Sciences: Consultancy; Seattle Genetics: Consultancy; Karyopharm: Consultancy; Regeneron: Current holder of individual stocks in a privately-held company; Moderna: Current holder of individual stocks in a privately-held company; TouchIME: Honoraria; GLG Pharma: Consultancy; Dava Oncology: Honoraria; Xynomic: Research Funding; Roche/Genentech: Research Funding; Novartis: Research Funding; Epizyme: Research Funding; Janssen: Research Funding; Autolus: Research Funding. Younes: AZ: Current Employment, Other: Senior Vice President, Global Head of Haematology (Early and Late Stage) Oncology R&D at AstraZeneca.
Major functional bias for mitochondrial complexes in genome-wide CRISPR screens
