Towards ‘best-in-class’ Hsp90 inhibitors: Design, biosynthesis and pre-clinical profiles

5724

Heat-shock protein 90 (Hsp90) is a promising target for anticancer drug development due to its maintenance of stability and function of a large number of client proteins involved in all of the hallmark traits of cancer (cell growth, apoptosis avoidance, uncontrolled proliferation, angiogenesis and metastasis). Inhibition of Hsp90 therefore affects multiple abnormalities in tumour growth regulation and potentially has a wide spectrum of anticancer effects.

Hsp90 inhibitors in advanced development for cancer can broadly be classified into two groups: semi-synthetic derivatives of naturally occurring ansamycins and small molecules, such as mimetics of adenosine and non-purine based resorcinol compounds. Semi-synthetic ansamycins such as 17-AAG, 17-DMAG and IPI-504 contain or can generate a common ‘toxicophore’ - the quinone moiety that undergoes redox cycling in vivo, generating superoxide radicals and forming covalent adducts with glutathione, both of which contribute to off-target non-selective toxicities such as dose-limiting liver toxicity. Small molecule inhibitors are still at the early stage of their development and few have progressed to clinical trials. Complex natural products are often found to have better target selectivity than their simpler small molecules.

We have previously characterised macbecin I as an alternative lead to geldanamycin as an Hsp90 inhibitor. It has a higher binding affinity than geldanamycin (Kd 0.24 μM vs 1.2 μM by ITC) to Hsp90 and is more water-soluble than geldanamycin (solubility 81 μM vs 1.7 μM). Through genetic engineering of the macbecin biosynthetic pathway we have now prepared a series of derivatives that lack the undesired quinone moiety, and do not form covalent adducts with glutathione. Additional structural modifications in the ansa ring have also made the compounds more ready to adopt the active conformation for binding to Hsp90. The most potent compound of this series has a binding affinity (Kd) of 3 nM to Hsp90, and is much better tolerated in vivo than 17-AAG (MTTD ≥ 250 mg/kg vs ~50 mg/kg i.p. in mice). It showed equivalent in vivo tumor growth inhibition in xenograft models at ≤ 1/3 MTTD to that of 17-AAG at MTTD. The improved therapeutic index coupled with its potent inhibition of Hsp90 makes the compound potentially the ‘best-in-class’ Hsp90 inhibitor described to-date.