The von Hippel-Lindau Tumor Suppressor Gene and Kidney Cancer: Insights into Oxygen Sensing and Treating Cancers Caused by Undruggable Mutations

VHL tumor suppressor protein (pVHL) inactivation is the usual initiating (“truncal”) event in the most common form of kidney cancer, clear cell renal cell carcinoma (ccRCC). pVHL forms a ubiquitin ligase that targets the HIF transcription factor for degradation. HIF2, but not HIF1, promotes ccRCC. Binding to pVHL requires that the HIFalpha subunit be prolyl hydroxylated by one of the 3 oxygen-sensitive EglN prolyl hydroxylases, which are 2-oxoglutarate (2-OG)-dependent dioxygenases (as are the TET DNA demethylases and KDM histone demethylases). IDH mutant cancers accumulate the 2-OG competitor 2-hydroxyglutarate (2-HG). Our work contributed to the development of VEGF inhibitors/HIF2 inhibitors for ccRCC and 2-HG inhibitors for IDH mutant leukemia. Our discovery that thalidomide reprograms cereblon to destroy the IKZF1/3 myeloma oncoproteins has also galvanized interest in small molecule degraders. Not all ccRCCs respond to VEGF inhibitors/HIF2 inhibitors and 2-HG inhibitors are fairly inactive against IDH mutant gliomas. Synthetic lethality (SL) should be a source of alternative drug targets for such tumors, and more generally, for cancers linked to “undruggable” mutations. We have identified new potential SL interactors for VHL (CDK4/6 and ITGAV) and mutant IDH (DHODH and GSK3b) and now propose further validation and mechanistic studies. Intriguingly, Cyclin D1, the partner for CDK4/6, is a HIF2 target in ccRCC, but the SL between VHL and CDK4/6 is not HIF2-dependent. We also embarked on SL screens in Drosophila cells because paralog compensation likely causes many false-negatives in genetic screens with human cells. We created an “up” screen for chemicals and gene knockouts that can degrade a protein of interest. We used it to discover that Spautin-1 is a cereblon-independent IKZF1 degrader and are pursuing the underlying mechanism. We are also applying it to various undruggable oncoproteins (e.g., c-Myc, K-Ras, b-Catenin). Failure to downregulate Cyclin D1 causes resistance of VHL-/- ccRCC to HIF2 inhibitors in a pRB- independent manner. We are using biochemical approaches to identify the relevant Cyclin D1 substrate(s). We unexpectedly found that HIF1 and HIF2 bind to approximately the same genomic sites, yet recruit different proteins. We are validating these associated proteins in biochemical and functional assays. We are also using substrate-trapping conditions to recover non-HIF EglN substrates and non-histone KDM substrates. We are continuing our efforts to use somatic gene editing with CRISPR to make a murine model of HIF2- dependent VHL-/- ccRCC. We have also identified new pVHL-dependent secreted proteins including PTH-LH, which might cause the cachexia and hypercalcemia seen in some ccRCC patients. ccRCC is highly immunogenic, but the reason is unknown. We found that HIF2 drives the expression of many endogenous retroviruses, some of which can be translated and presented as HLA-bound peptides. We are examining additional ccRCC cell lines and tumors for such peptides and whether they are immunogenic.