Administration of negative allosteric modulators of GluN2B subunit-containing NMDA receptors such as Ro 25-6981 (1) and ifenprodil (2) results in neuroprotective effects. In this study, the phenol of 1 and 2 was replaced bioisosterically by an indazole to inhibit glucuronidation. The γ- and β-aminoalcohols 10 and 11 were prepared without installing a protective group at the indazole ring using the ketone 6 as a common intermediate. All four stereoisomeric γ- and β-aminoalcohols 10 and 11 were obtained by diastereoselective reduction of ketones 7 and 9 followed by separation of enantiomers. The analogously structured γ-aminoalcohol (1S,2S)-10c (Ro 25-6981 bioisostere) and β-aminoalcohol (1R,2R)-11c (ifenprodil bioisostere) exhibited high GluN2B affinity (Ki = 50 and 66 nM, respectively) and high to moderate inhibitory activity in two-electrode voltage clamp experiments. The indazole bioisosteres 10 and 11 showed higher metabolic stability than 1. In the presence of uridinyldiphosphate activated glucuronic acid, glucuronidation of 10 and 11 was not observed.

In order to obtain novel antagonists of GluN2B subunit containing NMDA receptors, aryloxiranes were opened with benzylpiperidines. Phenyloxiranes 6 and (indazolyl)oxirane 15 were opened regioselectively at the position bearing the aryl moiety. Reaction of the resulting β-aminoalcohols 7 and 16 with carboxylic acids under Mitsunobu conditions (DIAD, PPh3) led to rearrangement and after ester hydrolysis to the regioisomeric β-aminoalcohols 9 and 18. This strategy allows the synthesis of amino-ifenprodil 12 as well using phthalimide in the Mitsunobu reaction. Unexpectedly, the isomeric (indazolyl)oxirane 21 reacted with benzylpiperidines to afford both regioisomeric β-aminoalcohols 22 and 23. In radioligand receptor binding studies, the indazolyl derivative 18a, which can be regarded as indazole bioisostere of ifenprodil, showed high GluN2B affinity (Ki = 31 nM). Replacement of the benzylic OH moiety of ifenprodil by the NH2 moiety in amino-ifenprodil 12 also resulted in low nanomolar GluN2B affinity (Ki = 72 nM). In TEVC experiments, 18a inhibited the ion flux to the same extent as ifenprodil proving that the phenol of ifenprodil can be replaced bioisosterically by an indazole ring maintaining affinity and inhibitory activity. Whereas 10-fold selectivity was found for the ifenprodil binding site over σ1 receptors, only low preference for the GluN2B receptor over σ2 receptors was detected. The log D7.4 value of 18a (log D7.4 = 2.08) indicates promising bioavailability.

Negative allosteric modulation of GluN2B subunit-containing NMDA receptors prevents overstimulation, resulting in neuroprotective effects. Since the phenol of prominent negative allosteric modulators is prone to rapid glucuronidation, its bioisosteric replacement by an indazole was envisaged. The key step in the synthesis was a Sonogashira reaction of non-protected iodoindazoles with propargylpiperidine derivatives. Modification of the alkynyl moiety allowed the introduction of several functional groups. The synthesized indazoles showed very high GluN2B affinity but limited selectivity over σ receptors. Molecular dynamics simulations revealed the same molecular interactions with the ifenprodil binding site as the analogous phenols. In two-electrode voltage-clamp experiments, enantiomeric 3-(4-benzylpiperidin-1-yl)-1-(1H-indazol-5-yl)propan-1-ols (S)-10a and (R)-10a displayed higher inhibitory activity than ifenprodil. In contrast to phenolic GluN2B antagonists, the indazoles were not conjugated with glucuronic acid. It can be concluded that the phenol of potent GluN2B antagonists can be replaced bioisosterically by an indazole, retaining the high GluN2B affinity and activity but inhibiting glucuronidation.