In order to address the Cyp inhibition
In order to address the Cyp inhibition issue, we tested the possibility of further changes in the heterocycle combined with reducing the electron density in the ring system by swapping the central aniline nitrogen atom for a ketone. details the synthesis of these analogs, which started with 7-bromobenzosuberone. Oxime installation followed by cyclization with 4-methoxyphenylacetic BKM120 produced the tricyclic 2-(4-methoxy-benzyl)-7-methyl-9,10-dihydro-1-oxa-3-aza-benzoazulen-4-one core . Suzuki coupling produced tricyclic (IC=0.30μM), while addition of methylamine into intermediate produced the corresponding tricyclic imidazole core which allowed production of analogs and (IC’s=1.1μM and 0.38μM, respectively). All three analogs were modest GSMs with IC’s >300nM. Unfortunately, all three analogs remained potent CYP3A4 inhibitors (IC’s=0.3–0.5μM). While the 4-methoxybenzyl substituent did not allow direct comparison of activity to the thiazole series, we were encouraged that the first analogs containing the central ketone were modestly active and wanted to test expanding the heterocyclic ring to a 6-membered pyridyl ring. A first pyridyl analog was synthesized with the suberone core using a slight variant of established chemistry (). Negishi coupling of and was followed by -oxidation and rearrangement to the pyridyl 2-nitrile . Formation of the -butyl amide followed by deprotonation and alkylation of the pyridyl methyl group gave . Dehydration back to the nitrile followed by acid-catalyzed cyclization and Ullman coupling gave the desired compound . This initial 6-membered ring analog had an Aβ1–42 IC=1μM and was still a CYP3A4 inhibitor (IC=0.3μM). Hoping to both improve the potency of the analogs and further improve the Cyp inhibitor profile we next moved to the well-established 6,11-dihydro-5-benzopyrido[2,3-][1,4]diazepin-5-one core, where existing chemistry would allow for additional modifications of both the steric and electronic environment of the inhibitor. Synthetically we accessed the desired analogs by selectively installing a 6-benzyl group on 2,6-dichloronicotinc acid using a Negishi coupling protocol. In order to test steric blocking of the pyridyl group, was bismethylated and then saponified to produce . Separately, the left-hand heterocycle was installed on 4-fluoronitro intermediates . Reduction to the diamine followed by thermal condensation with the preformed chloronicotinic acids produced the desired analogs –, albeit in modest chemical yield. The desired regioisomer was typically favored, but purification was necessary to produce pure material (). The Aβ1–42 activity and inhibition of CYP3A4 for compounds – is summarized in . The parent 4-methylimidazole analog was the most potent inhibitor in the series (IC=50nM), with activity similar to that of the more potent compounds in the thiazine series – (IC=35nM). The inhibition of CYP3A4, however, was not improved by changing to this core (IC=0.15μM). In this particular series, the chloroimidazole was 2–3 fold more active than the corresponding methyltriazole . Somewhat surprisingly, fluorine was poorly tolerated at R, with over a 10-fold loss in Aβ1–42 potency and an increase in CYP3A4 inhibition. Interestingly, the chloroimidazoles and had a dramatically improved CYP3A4 profile (IC >13μM). This was not solely the result of a solubility change, as the assay is monitored for drug precipitation. The data suggest that in this series the CYP3A4 SAR closely tracks with heterocycle basicity, with the methyl imidazole being most potent, followed by the triazole, while the weakly basic chloroimidazole is a poor binder. More analogs would be needed to confirm these observations, however the series overall suffered from modest potency at Aβ1–42, so additional analogs were not prepared. As a benchmark, compound was also examined at 30mpk in the 3xTg-AD mouse model, but no activity was observed. In summary, starting from a class of aniline triazines GSMs, we designed multiple series of conformationally restrained, tricyclic GSMs. The 2-benzyl-9,10-dihydro-4-1-thia-3,4-diaza-benzoazulenes including compound were effective in an animal model of Aβ1–42 production at high exposure. In an effort to explore the SAR and separate efficacy from inhibition of CYP3A4 activity, a variety of additional tricyclic cores was explored. While we were able to construct additional scaffolds with moderate to good in vitro potency, the combination of high potency without inhibition of CYP3A4 remains an issue to be addressed.