Design and synthesis of novel (S)-methadone analogs as non-competitive N-methyl-D-aspartate receptor antagonists

Dextromethadone (DMD, REL-1017), the (S)-enantiomer of racemic methadone, is a non-competitive antagonist of the NMDAR, and has the potential to become a new orally administered single-agent NMDAR antagonist approved for the treatment of MDD. Currently, Phase 3 clinical trials for DMD are in progress, aiming to enroll approximately 1000 patients diagnosed with MDD. Should the Phase 3 trials reproduce the positive outcomes observed in Phase 2, DMD could provide a safe and well-tolerated rapid treatment option for millions of patients. In order to elucidate the structure-activity relationship (SAR) of DMD for NMDAR, and to develop new efficient NMDAR non-competitive antagonists, this Ph.D. project aims at synthesizing (S)-methadone analogs by modifying its three main molecular features. At the beginning of the project, a virtual library of about 300 synthesizable DMD analogs has been designed by modifying its three main structural domains: aryl moieties, ethyl ketone group, and N,N-dimethylpropyl-2-amino group, respectively Ars, COR1, and CH2CH(NMe2)R2 (in Figure 35). For initial in silico screening of the virtual library and throughout the project we have conducted computational studies, performed in collaboration with Prof. Andrea Cavalli (IRB, Bellinzona, CH), to optimize and guide the synthetic efforts. During the PhD project we have synthesized 40 new DMD analogs (achieved through three different synthesis generations), that have been tested for NMDAR antagonistic activity (by Evotec Italia S.r.l company), by means of in vitro FLIPR™ assays, using cell lines stably overexpressing diheteromeric recombinant human NMDAR, containing GluN1 plus one amongst GluN2A, GluN2B, GluN2C, or GluN2D subunit. A selected pool of new NMDAR antagonists among those synthesized over the three years has been further investigated by ligand displacement experiments on other relevant CNS receptors to assess potential off-target effects. In conclusion, we have set up the way for the development of new lead candidates with a DMD-like structure, and with off-patent structures. The novel analogs have been structurally modified through a rational lead optimization approach to reach the optimal response, both in terms of modulating efficacy and selectivity at NMDAR subtypes, and finally tested to assess possible off-target interactions at CNS relevant receptors. Two compounds, namely DMD-oCl-simm and DMD-R5 have been selected for further preclinical in vivo behavioral studies in mice models of depression, because they possess the best compromise between ease of production, in vitro NMDARs activity profile, and other CNS off-target interactions, for possible future pharmaceutical product development. After three generations of synthesized compounds, we can confirm that lipophilic substitutions on the two Ars of DMD, (especially on the aryl ortho position) showed the best results in terms of NMDAR non-competitive antagonism. During a short research period at the IRB (Bellinzona, CH) in Prof. Andrea Cavalli’s group new findings had led us to hypothesize that the hydrophobic interactions of the modified Ars with the lipophilic residues in the transmembrane domains may play a pivotal role in determining whether or not the potential non-competitive antagonists can reach the phencyclidine (PCP) binding site and how efficient its “travel” is. It should be mentioned that the NMDAR cavities are very narrow spaces in which the steric hindrance of the modified compounds need also to be carefully considered. The good balance between lipophilicity and small dimensions may suggest a possible SAR explanation on why chlorine substituted analogs (compared to others) seem to be the best compromise to efficiently reach the PCP binding site. Further computational analysis will be crucial to obtain accurate simulations, and find additional insights that may justify the undoubtedly favored substitution on the ortho position of the two Ars in DMD analogs.