Abstract 4355: Elucidation of the different roles of CDK8 and CDK19 in colorectal cancer (CRC) using CRISPR gene editing technology

Abstract

Abstract Background CDK8 and CDK19 are two highly homologous cyclin-dependent kinases involved in gene transcription via the Mediator complex. Individual kinase module paralogs (CDK8 or CDK19) in complex with cyclin C, Med13 and Med12 regulate transcription through phosphorylation of the C-terminal domain of RNAPol II. Despite their structural similarities, emerging data suggest possible distinct transcriptional regulatory functions. There are few studies of CDK19 function, but CDK8 has been described as an oncogene in CRC, required for β-catenin mediated transcription and tumor cell proliferation. In addition, CDK8 knock down reportedly inhibited the growth of HT29 and Colo205 CRC xenografts. Results Recently, we identified a series of 3,4,5-trisubstituted pyridines (exemplified by CCT251545) as small molecule chemical probes of CDK8 and CDK19.1 Although our compounds bind selectively to these closely related biochemical targets, it is not known whether both are equally inhibited, whether inactivation of one (or both) is required for tumor growth control or if there is compensation in terms of kinase function and/or other roles in the Mediator complex. Using single and double CRISPR/Cas9 knockout clones, we aim to elucidate the roles of CDK8 and CDK19 in human CRC cells. We used the SW620 CRC cell line due to its sensitivity in vivo to our CDK8/19 inhibitor CCT251921. Single and double homozygous knockout (KO) clones were generated after CRISPR/Cas9 transfection of plasmids targeting CDK8 and CDK19. KO clones were identified by western blotting and validated by qPCR and exome sequencing analyses. Functional assays were performed in vitro and PD/PK and efficacy studies with CCT251921 in vivo using selected single and double KO clones. Substrates including pRNAIISer5, E2F1Ser375 and pSTAT1Ser727were also quantified by several independent assay platforms in in vitro and ex vivo tumor samples. mRNA profiling of ex vivo samples from PD/PK and efficacy studies was also undertaken. CRISPR deletion of neither CDK8 (in contrast to literature reports using RNAi in different CRC xenografts) nor CDK19 inhibited the growth of SW620 CRC cells in vitro or in vivo. Interestingly, the CDK19 knockout cells remained responsive to two chemically distinct probes in vitro and/or in vivo, with clear inhibition of the pSTAT1Ser727 engagement biomarker, whereas the CDK8 and CDK8/19 knockouts were far less responsive and showed significantly lower levels of pSTAT1Ser727. Our data suggest that there are certain independent functions of CDK8 and CDK19 that might not be simply interchangeable and that the presence or absence of one may also impact on the behaviour of the other.