Leveraging selective degradation of CBP and EP300 for potent anti-cancer activity

Numerous cancers have shown dependency on one of two paralog histone acetyltransferases, CREB binding protein (CBP) or E1A binding protein (EP300). Genomic screens have identified a bidirectional synthetic lethal relationship for these two proteins, such that, for example, EP300-mutant gastric and colorectal malignancies strongly depend on CBP function for growth and survival. Additionally, the divergent biology of these targets can underly selective lineage dependencies. For example, hematological cancers are exquisitely sensitive to loss of EP300, even in the presence of functionally intact CBP. Previously developed inhibitors that disrupt activity of both enzymes have resulted in hematopoietic toxicity, and selectively drugging these targets to improve the therapeutic window has been a significant challenge. Here we describe our platform to develop fast, potent, and selective CBP and EP300 degraders that remove only one of the targets, while sparing the activity of its paralog. We elucidate the mechanisms underlying selective degradation through cellular, structural, and biophysical assays of ternary complex formation and ubiquitination between the target proteins and the E3 ligase VHL. We find that our selective degraders likely drive differential target ubiquitination and degradation through differences in ternary complex stability and the induction of divergent orientations between the bromodomains of the respective target relative to VHL. Additionally, we demonstrate that CBP and EP300 selective degraders show strong anti-proliferative effects across numerous oncology indications that are dependent on either target. Importantly, treatment with our selective degraders at efficacious doses is not associated with hematopoietic toxicity, providing evidence that specific targeting of one of the paralogs provides a wider therapeutic window than dual inhibition.