The Sticky Wicket of Undruggable Mutations: Can Targeted Molecular Degradation Help?

Undruggable targets, that is, proteins that are not amenable to small molecule inhibition, are a major challenge in therapeutic target drug development. One approach to target proteins that are intractable to conventional pharmacological approaches is to induce selective degradation through ubiquitination. So-called “molecular glue” degraders (MGDs) are an exciting class of small molecule drugs that induce or stabilize interaction between proteins, in some cases leading to selective protein degradation. Thalidomide analogues (IMiDs) act as MGDs, and many these are already in widespread clinical use in the treatment of multiple myeloma, B-cell malignancies, and myelodysplastic syndromes with a deletion in chromosome 5q. Binding of IMiDs to their primary target, the substrate receptor Cereblon (CRBN), confers a neomorphic activity to the E3 ubiquitin ligase complex and reprograms its selectivity, thus diverting the ligase to drive multiple rounds of target ubiquitination (IKZF1 and IKZF3 in the case of multiple myeloma) and leading to proteasomal degradation of the target protein substrates. Thus far, MGDs have largely been discovered serendipitously, and identification of novel agents is a major challenge in the field.

Dr. Steven Corsello and colleagues previously devised an ambitious high-throughput screening project through the repurposing of clinical and preclinical drugs.¹  “PRISM” is a multiplexed screening method comprising libraries of 4,518 oncology and nononcology small-molecule compounds assayed against a panel of 578 human cancer cell lines spanning 24 tumor types. This screening strategy was harnessed by Dr. Mikołaj Słabicki to discover novel MGDs by correlating drug sensitivity with the expression level of 499 E3 ligase components. The primary validation screens identified the aryl sulphonamides indisulam and tasisulam that are known MGDs, alongside the pleiotropic CDK inhibitor CR8 as the most potent inducers of cytotoxicity across the pooled cell lines tested. This was associated with increased expression of the E3 ubiquitin ligase adaptor proteins DCAF15 and DDB1, respectively. Dependency on these E3 ubiquitin ligase-adaptor proteins was confirmed by CRISPR-based knockout.

Proteomic analysis, followed by in vitro polyubiquitination and proteasome inhibition assays, revealed that DDB1-associated sensitivity to CR8 is orchestrated by a ubiquitin ligase complex that rapidly destabilizes the substrate, cyclin K (CycK). The composition of the ligase core was further elucidated using a genome-wide CRISPR screen that revealed the DDB1-CUL4B-RBX1 system as the molecular determinant of resistance to CR8. Surprisingly, additional reporter assays aiming to unravel the regulators of substrate degradation failed to identify any canonical substrate receptors. Intriguingly, the cyclin-dependent kinase CDK12 was identified as a crucial component for CR8-induced degradation of CycK. CDK12 does not interact physiologically with the adaptor protein DDB1, but serves as drug-induced substrate receptor, linking DDB1 to the target for ubiquitination.

The extensive study of the crystal structure of the DDB1-CR8-CDK12-CycK complex revealed for the first time the ability of a molecular glue to provide the structural framework for the enzymatic activity of an E3 ubiquitin ligase complex. Indeed, CR8 catalyzes a rapid ubiquitination process by engaging complementarily, but not directly interfering with components of the ligase core. The presence of a solvent-exposed moiety of CR8 induces de novo contacts with CDK12, a nonconstitutive component of the E3 ligase, and is sufficient for overcoming the affinity threshold of the ligase, thus conferring gain-of-function glue properties to the compound.