Background: Treatment with kinase inhibitors, although initially effective in the clinic, frequently leads to resistance and relapse. Multiple mechanisms of resistance have been identified including mutation of the original targeted kinase or upregulation of downstream or parallel signaling pathways. Many oncogenic kinases are clients of HSP90, including KIT, B-RAFV600E and EML4/ALK and remain dependent on HSP90 even when secondary mutations arise. In addition, proteins in many signaling pathways also rely on HSP90 activity and are depleted when HSP90 is inhibited. Treatment with an HSP90 inhibitor is therefore a promising approach for addressing multiple mechanisms of kinase inhibitor resistance.

AT13387, a potent, fragment-derived HSP90 inhibitor, is currently under evaluation in a Phase II gastrointestinal stromal tumor (GIST) trial in combination with imatinib. AT13387 is effective in both imatinib-sensitive and imatinib-resistant GIST models, demonstrating HSP90 inhibition can overcome kinase inhibitor-resistance caused by secondary mutations in the drug target client, in this case KIT. Here we describe the activity of AT13387 in vemurafenib-resistant melanoma models. Multiple mechanisms of vemurafenib resistance have been identified. These frequently involve upregulation of the AKT (e.g. PTEN loss) or MEK/ERK (e.g. Ras mutation, BRAF splice variants) signaling pathways, rather than mutation at the vemurafenib-binding site in BRAFV600E.

Results: The effect of HSP90 inhibition on acquired vemurafenib resistance was investigated by generating resistant clones from the A375, vemurafenib-sensitive, cell line by continuous exposure to 2 μM vemurafenib. Proliferation and signalling in the vemurafenib-resistant pool of cells were insensitive to vemurafenib but potently inhibited by AT13387, indicating that HSP90 inhibition can overcome an acquired mechanism of vemurafenib resistance.

In established cell lines, two different mechanisms of vemurafenib resistance were investigated. The A2058 cell line is PTEN-null and overcomes BRAFV600E inhibition through upregulation of the AKT pathway, whilst RPMI7951 overexpresses COT, upregulating the MEK/ERK pathway and bypassing inhibition of BRAFV600E. The proliferation of both these cell lines was inhibited by AT13387 with GI50s of 31 and 37 nM respectively. AT13387 treatment depleted the levels of the client proteins, BRAF and CRAF, and in the RPMI7951 cell line, levels of COT were also seen to decrease. Signaling through both the AKT and MEK/ERK pathways was inhibited as indicated by a decrease in the levels of phosphoAKT, phosphoS6 and phosphoERK. This suggests that HSP90 inhibition can overcome upregulation of either the AKT or MEK/ERK pathways and so counteract multiple mechanisms of vemurafenib resistance.

AT13387 and vemurafenib could be combined in both the A2058 and RPMI7951 cell lines without any antagonistic effects, whilst culturing A375 cells in the presence of both AT13387 and vemurafenib prevented the development of vemurafenib-resistant clones. These data suggest that not only can AT13387 treatment overcome established resistance but that in combination it may also be a means of delaying the emergence of this resistance.

Conclusions: We have demonstrated that AT13387 can overcome kinase inhibitor resistance in both imatinib-resistant GIST and vemurafenib-resistant melanoma models. This suggests HSP90 inhibition may be a promising approach for combating multiple mechanisms of resistance to kinase inhibitors in the clinic and that AT13387 combinations may have the potential to delay the emergence of resistance.

This content is only available via PDF.