Hsp90 inhibition differentially destabilises MAP kinase and TGF-beta signalling components in cancer cells revealed by kinase-targeted chemoproteomics
- Equal contributors
1 Department of Vertebrate Genomics, Max-Planck Institute for Molecular Genetics, Ihnestrasse 73, 14195 Berlin, Germany
2 Cellzome AG, Meyerhofstrasse 1, 69117 Heidelberg, Germany
3 International Center for Information Technology, University of Bonn, 53113 Bonn, Germany
4 Cell Division Group, IRB-Barcelona, PCB, c/Baldiri Reixac 10-12, 08028 Barcelona, Spain
5 Institucio Catalana de Recerca i Estudis Avançats, Passeig Lluís Companys 23 08010 Barcelona, Spain
BMC Cancer 2012, 12:38 doi:10.1186/1471-2407-12-38Published: 25 January 2012
The heat shock protein 90 (Hsp90) is required for the stability of many signalling kinases. As a target for cancer therapy it allows the simultaneous inhibition of several signalling pathways. However, its inhibition in healthy cells could also lead to severe side effects. This is the first comprehensive analysis of the response to Hsp90 inhibition at the kinome level.
We quantitatively profiled the effects of Hsp90 inhibition by geldanamycin on the kinome of one primary (Hs68) and three tumour cell lines (SW480, U2OS, A549) by affinity proteomics based on immobilized broad spectrum kinase inhibitors ("kinobeads"). To identify affected pathways we used the KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway classification. We combined Hsp90 and proteasome inhibition to identify Hsp90 substrates in Hs68 and SW480 cells. The mutational status of kinases from the used cell lines was determined using next-generation sequencing. A mutation of Hsp90 candidate client RIPK2 was mapped onto its structure.
We measured relative abundances of > 140 protein kinases from the four cell lines in response to geldanamycin treatment and identified many new potential Hsp90 substrates. These kinases represent diverse families and cellular functions, with a strong representation of pathways involved in tumour progression like the BMP, MAPK and TGF-beta signalling cascades. Co-treatment with the proteasome inhibitor MG132 enabled us to classify 64 kinases as true Hsp90 clients. Finally, mutations in 7 kinases correlate with an altered response to Hsp90 inhibition. Structural modelling of the candidate client RIPK2 suggests an impact of the mutation on a proposed Hsp90 binding domain.
We propose a high confidence list of Hsp90 kinase clients, which provides new opportunities for targeted and combinatorial cancer treatment and diagnostic applications.