Targeting both ends of Hsp90: Chimeric N-terminal inhibitors and novobiocin-derivatized C-terminal inhibitors

Abstract

The 90 kDa heat shock proteins (Hsp90) are ubiquitously expressed in cells, which are integrally involved in the cell signaling, proliferation, and survival. Many proteins in tumor cells are dependent upon Hsp90 for their stability, refolding, and maturation. Additionally, inhibition of Hsp90 influences all six hallmarks of cancer: a documented feature unique to Hsp90. Thus, Hsp90 has emerged as one of the most promising targets for the treatment of cancer.

Hsp90 is a homodimer, containing three domains. The N-terminal domain has an ATP-binding site that binds natural products geldanamcyin and radicicol. The middle domain is highly charged and has an affinity for co-chaperones and client proteins. The C-terminal domain is suspected to have an additional ATP-binding site that binds cisplatin and the natural product, novobiocin.

Geldanamycin and radicicol are potent Hsp90 inhibitors. Co-crystal structures of these molecules bound to Hsp90, as well as other biological studies, provide evidence that the quinone moiety of geldanamycin and the resorcinol moiety of radicicol are key functional groups that contribute to their inhibitory activities. Therefore, it was reasonable to propose a seco-chimeric compound that contained both functional groups. Upon synthesis and biological evaluation of this lead compound, a series of related analogs were designed and synthesized. Preliminary data gave rise to structure-activity relationships that formed a great foundation for further investigation of this class of inhibitors. At the same time, biological investigations with macrocyclic chimeric compounds resulted in more potent inhibitors. Continued optimization of this class of inhibitors provides a guide for future studies.

There are debates about why geldanamycin exhibits increased potency in cellular over in vitro assays. The co-crystal structure of geldanamycin bound to Hsp90 revealed that this compound adopts a cup-shaped conformation when the amide bond is isomerized from trans to cis. To improve the binding affinity for the chimeric inhibitors, a cis-amide functionality was incorporated. However, biological evaluations did not produce the expected results. It is believed that in hindsight, the cis-amide group does not allow the molecule to adopt the cup-shaped conformation.

The aminocoumarin antibiotic, novobiocin, was reported to bind to the Hsp90 C-terminal ATP binding site. Structural modification of this compound led to an increase of 1000-fold in anti-proliferactive assays. This class of inhibitors can also induce degradation of Hsp90 client proteins. However, unlike the N-terminal ATP binding site, there is no reported co-crystal structure of Hsp90 C-terminus. Four photolabile novobiocin analogs were designed, synthesized, and incubated with Hsp90. Proteolytic studies indicated these compounds selectively bind to the same location as the natural product and are capable of labeling the protein. Additional studies are still ongoing.