Search

Amplification and/or overexpression of the PAK1 gene is common in several malignancies, and inhibition of PAK1 by small molecules has been shown to impede the growth and survival of such cells. Potent inhibitors of PAK1 and its close relatives, PAK2, and PAK3, have been described, but clinical development has been hindered by recent findings that PAK2 function is required for normal cardiovascular function in adult mice. A unique allosteric PAK1-selective inhibitor, NVS-PAK1-1, provides a potential path forward, but has relatively modest potency in cells. Here, we report the development of BJG-05-039, a PAK1-seletive degrader consisting of the allosteric PAK1 inhibitor NVS-PAK1-1 conjugated to lenalidomide, a recruiter of the E3 ubiquitin ligase substrate adaptor Cereblon (CRBN). BJG-05-039 induced degradation of PAK1, but not PAK2, and displayed enhanced anti-proliferative effects relative to its parent compound in PAK1-dependent, but not PAK2-dependent, cell lines. Notably, BJG-05-039 promoted sustained PAK1 degradation and inhibition of downstream signaling effects at ten-fold lower dosage than NVS-PAK1-1. Our findings suggest that selective PAK1 degradation may confer more potent pharmacological effects compared with catalytic inhibition and highlight the potential advantages of PAK1-targeted degradation.

e22512 Background: Activating PDGFRA mutations are seen in about 10% of gastrointestinal stromal tumors (GIST). Of these, the majority are one specific PDGFRA mutation, D842V, which confers resistance to all clinically approved tyrosine kinase inhibitors (TKI). Much has been learned about the mechanism by which TKIs work in GIST using cell lines derived from human GIST that bear activating KIT mutations. However, despite enormous effort to develop cell lines derived from human tumors over the past 15 years, there are currently no human cell models of PDGFRA D842V GIST. With the development of potent PDGFRA D842V TKI, such as crenolanib, avapritinib, and DCC2618 it is of interest to understand how these drugs affect PDGFRA-mutant GIST biology. Methods: We sought to convert the KIT-driven GIST-T1 cell line to be PDGFRA-driven by transduction with PDGFRA D842V. Endogenous KIT expression was knocked-out using CRISPR Cas9. Knockout was verified at protein and DNA levels. Clonal KIT knockout cell lines were tested for sensitivity to KIT and PDGFRA TKI both in vitro and in vivo models. Kinase dependence was determined by measuring the effect on viability after KIT and/or PDGFRA knockdown by siRNA. Results: GIST-T1 cells transduced with PDGFRA D842V display highly phosphorylated PDGFRA, imatinib resistance, and PDGFRA dependence. Cell lines that express both endogenous mutant KIT and exogenous PDGFRA D842V maintain some dependence on KIT, but less so than the parental KIT-driven GIST-T1 cell line. Complete knockout of KIT protein expression was achieved in GIST-T1 D842V cells using CRISPR-Cas9. GIST-T1 D842V KITko cell lines demonstrate sensitivity to TKI avapritinib, crenolanib, and DCC2618. Xenografts were established using GIST-T1 D842V KITko cell lines and are sensitive to avapritinib in vivo. Conclusions: The GIST-T1 D842V KITko cell line demonstrates ability to switch kinase allegiance resulting in PDGFRA-dependence and KIT-independence. This GIST-T1 D842V KITko cell line represents a novel model of PDGFRA-mutant GIST and can be used to understand the unique biological aspects of these tumors, particularly the response to new potent PDGFRA kinase inhibitors such as avapritinib which will have immense clinical relevance.