HSP90 inhibition without heat shock response

In this issue of Blood, Bhatia and colleagues describe a heat shock protein 90 (HSP90) C-terminal dimerization inhibitor with mechanistic differences that distinguish it from other clinically unsuccessful N-terminal ATPase binding compounds.¹  Can graveyard raiding of an old therapeutic target with a new strategy bring long awaited success?

Developing targeted therapeutics for cancer is quite complicated because multiple redundant mechanisms bypass that particular agent’s target of action. In addition, many therapeutic targets relevant to cancer also have normal functions that prevent targeting with small molecules. No such target more represents this dilemma than HSP90. HSP90 represented a promising target because multiple mutated or aberrantly expressed oncogenes depend on HSP90 for protein stabilization. Based on structural variations induced by their activating mutation, there was a rush to develop agents that inhibited HSP90, which should cause destabilization of oncogene-induced proteins. Such drugs would be hypothesized to have a dramatic clinical benefit, even in the presence of mutated oncogene proteins not responsive to standard kinase inhibitors. In addition, as many mutated proteins depend more on HSP90 for stabilization and tumors have been reported to have more active HSP90 ATPase activity,²  multiple animal studies demonstrated that a significant therapeutic window existed between the mutated proteins and the normal homolog proteins, thereby yielding an acceptable therapeutic index.³  Unfortunately, clinical trials of multiple different HSP90 inhibitors were unsuccessful due to both lack of clinical efficacy and toxicity. Some of these toxicities, including night blindness^4,5  and cardiac toxicity,^6,7  are unacceptable findings, which ultimately have prevented therapeutics from moving forward. How can such a promising target so utterly fail?

Bhatia and colleagues provide insight on why previous HSP90 inhibitors have failed, and based on this insight, developed a peptide inhibitor that prevents dimerization of the C-terminal portion of HSP90. Virtually all HSP90 inhibitors developed to date bind in the ATP binding site in the N-terminal domain of HSP90 protein. Although inhibiting HSP90, this yielded an HS1 transcriptional response, which activated alterative heat shock proteins (HSP70, HSP40, or HSP27) with either redundant or compensatory antiapoptotic function, thereby maintaining the stability of mutated proteins and/or preventing cell death. Utilizing a different strategy, Bhatia and colleagues used structural combinatorial monitoring to develop small peptides, which prevent dimerization of HSP90 protein in the C-terminal region of the protein. This novel approach with their peptide compound aminoxyrone did not promote an HS1 transcriptional response and resistance to cancer cell death, as seen with earlier compounds explored by others. The authors show also that aminoxyrone is effective in preclinical models of chronic myeloid leukemia (CML) similar to N-terminal–directed HSP90 inhibitors.^8,9  Unlike these older agents, aminoxyrone does not promote an HS1 response. Aminoxyrone also effectively depleted BCR-ABL–containing CML stem cells, the cellular origin of this disease that requires long-term treatment with BCR-ABL–targeted therapeutics such as imatinib. For a well-characterized oncogene-driven disease, this aminoxyrone was effective in vitro. In addition, aminoxyrone was shown to have in vivo activity in a K562 cell line model with evidence of tumor pharmacodynamic modulation of STAT5a and Crkl phosphorylation without modulation of other HS1 targets. Supporting data are also provided from primary tumor cells derived from tumors with other mutated oncogenes, including FLT3 ITD⁺ acute myeloid leukemia and Philadelphia chromosome–like acute lymphoblastic leukemia. In addition, preclinical activity was also shown with chronic lymphocytic leukemia, a disease consistently demonstrated to be sensitive in vitro to HSP90 inhibition.¹⁰ 

Where does aminoxyrone as a therapeutic go from here? Although in vivo activity was demonstrated against the K562 cell line in vivo, it is unlikely that the supramolar concentrations required will allow effective translation to clinical trials in patients. Peptide therapeutics are challenging to develop due to production issues, cellular penetration, and delivery. Even if aminoxyrone cannot be translated to the clinic, this study shows that derivative molecules that bind to the C-terminal region of HSP90 and prevent dimerization may have a favorable clinical impact on cancers dependent on HSP90 client proteins, which often represents an active oncogene. Provided preclinical toxicity, pharmacology, and pharmacodynamic studies of these compounds look favorable, it will lie upon the pharmaceutical developers to move forward with the initial phase 1 trials despite multiple negative HSP90 inhibitor trials published to date. The concept presented in this paper differentiating aminoxyrone as a HSP90 C-terminal dimerization inhibitor justifies this and gives hope that a long described therapeutic target for cancer might actually come to fruition.