Fragments improve solubility: GlaxoSmithKline’s BD2 inhibitors

The epigenetic readers known as bromodomains have been popular anticancer targets for fragment-based approaches. But rapid success in generating potent molecules has not led to equally rapid success in the clinic, in part due to toxicity. Many early molecules inhibited both of the bromodomains (BD1 and BD2) present in the four BET family proteins, and some evidence suggests that BD2-selective inhibitors would be better tolerated. Indeed, last year we wrote about AbbVie’s selective clinical compound. Now GlaxoSmithKline has just reported a new selective inhibitor in J. Med. Chem.

 

GlaxoSmithKline had previously discovered the BD2-selective molecule GSK620, which unfortunately suffered from low solubility in FaSSIF (fasted state simulated intestinal fluid). To try to improve this molecule, they turned to fragments. Although the screening details are not described here, the company’s first fragment screens against bromodomains from a decade ago provided dozens of crystallographically-characterized starting points. Compound 6 binds in a similar manner to GSK620 and has good ligand efficiency.

 

 

Compound 6 shows equal potency against BD1 and BD2 of BRD4, but merging the five-membered core with GSK620 led to BD2-selective compound (S)-11. (Although compound 6 contains a pyrrole, the NH was inconveniently positioned and thus the researchers explored other five-membered heterocycles during scaffold hopping; the paper describes furan and pyrazole series.) Further optimization of the furan, in part based on earlier SAR, ultimately led to GSK743.

 

This molecule showed greater than 1000-fold selectivity for the BD2 bromodomain of BRD4 over the BD1 domain, and >300-fold selectivity for the BD2 domains of BRD2, BRD3, and BRDT as well as selectivity against a large panel of other bromodomains. More extensive profiling revealed it to be clean against CYP3A4, hERG, and other potential off-targets, and it was also negative in an Ames test for mutagenicity. Pharmacokinetics and oral bioavailability were also reasonable in both rat and dog. The compound had potent antiproliferative activity against acute myeloid leukemia cell lines. Finally, FaSSIF solubility was at least 20-fold better than for GSK620.

 

This is a nice example of fragment-based scaffold hopping, akin to another example from GlaxoSmithKline we highlighted last year. Whether GSK743 ultimately advances will probably depend on how other molecules in the class perform. Biology will have the final say, but fragments – combined with elegant medicinal chemistry – provided the tools to answer the questions.