Inhibition of class IIa histone deacetylase (HDAC) enzymes have been suggested as a therapeutic strategy for a number of diseases, including Huntington's disease. Catalytic-site small molecule inhibitors of the class IIa HDAC4, -5, -7, and -9 were developed. These trisubstituted diarylcyclopropanehydroxamic acids were designed to exploit a lower pocket that is characteristic for the class IIa HDACs, not present in other HDAC classes. Selected inhibitors were cocrystallized with the catalytic domain of human HDAC4. We describe the first HDAC4 catalytic domain crystal structure in a "closed-loop" form, which in our view represents the biologically relevant conformation. We have demonstrated that these molecules can differentiate class IIa HDACs from class I and class IIb subtypes. They exhibited pharmacokinetic properties that should enable the assessment of their therapeutic benefit in both peripheral and CNS disorders. These selective inhibitors provide a means for evaluating potential efficacy in preclinical models in vivo.

No abstract

No abstract

Histone deacetylase (HDAC) inhibitors have received considerable attention as potential therapeutics for a variety of cancers and neurological disorders. Recent publications on a class of pimelic diphenylamide HDAC inhibitors have highlighted their promise in the treatment of the neurodegenerative diseases Friedreich’s ataxia and Huntington’s disease, based on efficacy in cell and mouse models. These studies’ authors have proposed that the unique action of these compounds compared to hydroxamic acid-based HDAC inhibitors results from their unusual slow-on/slow-off kinetics of binding, preferentially to HDAC3, resulting in a distinctive pharmacological profile and reduced toxicity. Here, we evaluate the HDAC subtype selectivity, cellular activity, absorption, distribution, metabolism and excretion (ADME) properties, as well as the central pharmacodynamic profile of one such compound, HDACi 4b, previously described to show efficacy in vivo in the R6/2 mouse model of Huntington’s disease. Based on our data reported here, we conclude that while the in vitro selectivity and binding mode are largely in agreement with previous reports, the physicochemical properties, metabolic and p-glycoprotein (Pgp) substrate liability of HDACi 4b render this compound suboptimal to investigate central Class I HDAC inhibition in vivo in mouse per oral administration. A drug administration regimen using HDACi 4b dissolved in drinking water was used in the previous proof of concept study, casting doubt on the validation of CNS HDAC3 inhibition as a target for the treatment of Huntington’s disease. We highlight physicochemical stability and metabolic issues with 4b that are likely intrinsic liabilities of the benzamide chemotype in general.

Using an iterative structure−activity relationship driven approach, we identified a CNS-penetrant 5-(trifluoromethyl)-1,2,4-oxadiazole (TFMO, 12) with a pharmacokinetic profile suitable for probing class IIa histone deacetylase (HDAC) inhibition in vivo. Given the lack of understanding of endogenous class IIa HDAC substrates, we developed a surrogate readout to measure compound effects in vivo, by exploiting the >100-fold selectivity compound 12 exhibits over class I/IIb HDACs. We achieved adequate brain exposure with compound 12 in mice to estimate a class I/IIb deacetylation EC 50 , using class I substrate H4K12 acetylation and global acetylation levels as a pharmacodynamic readout. We observed excellent correlation between the compound 12 in vivo pharmacodynamic response and in vitro class I/IIb cellular activity. Applying the same relationship to class IIa HDAC inhibition, we estimated the compound 12 dose required to inhibit class IIa HDAC activity, for use in preclinical models of Huntington's disease.