In this issue of “Small Molecule Highlights” we bring you 5 new small molecules out of recent drug discovery journals. These molecules target a diverse selection of targets, including BRAF, ATR, SARS-CoV-2 3C-like protease, ADAMTS7, and LPAR1. Enjoy!
KIN-2787
KIN-2787: Paradoxical activation of RAF dimers (homo- and hetero-) continues to be a confounding issue in the development of specific RAF inhibitors as a monotherapy. While advances have been made in pan-RAF inhibitors that can block dimer-dependant signaling, the current state of the art (i.e. naporafenib and belvarefenib) are marred by poor physicochemical properties, namely, solubility. The team at Kinnate Biopharma looked to remedy this by optimizing the naporafenib scaffold to increase its solubility and circumvent compound metabolism. Towards this end, the medicinal chemists investigated different alcohol containing appendages to increase solubility. Incorporation of the chiral alaninol substituent was hypothesized to disrupt crystal packing thereby aiding solubility, as well as protect the solvent exposed alcohol group from oxidative metabolism. The pyrrolidine urea group was deployed to stabilize the DFG out conformation, with the pyrrolidine group and its pendant trifluoromethyl substituent making key lipophilic interactions within the DFG vacated back pocket. The morpholino-substituent was maintained from naporafenib as it bestowed the requisite kinase selectivity. In cellular efficacy studies, KIN-2787 was shown to potently inhibit phosphorylation of ERK in monomeric class I altered cell lines (A375 EC50 = 62 nM and Colo800 EC50 = 103 nM), dimer-driven class II altered cell lines (NCI-H2405 EC50 = 10 nM, BxPC-3 EC50 = 51 nM, and OV-90 EC50 = 26 nM) and heterodimer class III altered cell lines (WM3629 EC50 = 9 nM, and CAL-12T EC50 = 18 nM). Crucially, relatively potent pERK inhibition was observed in a standard cell model of drug-induced paradoxical activation (IPC-298 cells; EC50 = 265 nM). Kinase selectivity was assessed across a panel of wild-type and mutant kinases (Reaction Biology; 688 kinases in total), with only DDR (both wild-type and mutant) showing up as significant off-target kinases. The compound demonstrated good hepatocyte stability across species (% remaining @ 60 min = 85/72/69/62 in human/rat/dog/mouse) and respectable oral bioavailability in rat, mouse and dog (F(%) = 44/82/96 in mouse/rat/dog). Pharmacodynamic evaluation of KIN-2787 took place in xenograft models exemplifying class I (A375), class II (BxPC-3), and class III (WM3629) NRAS driven cancers where BRAF in mutated. In all cases, dose-dependant tumor growth inhibition, and regression at the highest doses tested, was observed with no overt signs of toxicity. Efficacy was also demonstrated in NRAS driven models harboring BRAF WT protein (SK-MEL-2 in NOD SCID mice). Given the excellent PD data, a phase I dose escalation trial (NCT04913285) was conducted to examine plasma exposure and appropriate dosing (300 mg/kg BID gave an average plasma concentration of 250 nM). A follow-up safety and efficacy study in patients with BRAF and/or NRAS mutant positive solid tumors is currently recruiting (NCT04913285).
Reference: https://doi.org/10.1021/acs.jmedchem.3c01830
RP-3500 (Camonsertib)
RP-3500 (Camonsertib): This latest report from the team at Repare Therapeutics provides a very nice example of pharmacophore driven drug optimization. The target, ataxia telangiectasia and Rad3-related (ATR) kinase, is a key protein in multiple DNA-damage response pathways, offers opportunities for synthetic lethality with several proteins, including ATM and BRCA1/2. Taking AZ20 (AstraZeneca) as a starting point, the medicinal chemistry team quickly evaluated requirements for efficient binding and selectivity towards ATR. Their assessment revealed a key hydrogen bond donor interaction requiring the pyrazole of RP-3500, and a need for structural rigidification of the original AZ20 scaffold to orient this substituent correctly (via azaindazole ring). A further breakthrough in SAR was realized with the identification of the tertiary alcohol group which provided desirable physicochemical properties and increased metabolic stability. These modifications bolstered potency (HeLa IC50 = 0.9 nM) and bestowed exquisite selectivity for ATR (i.e. 30-fold selectivity over mTOR demonstrated in LoVo cells). Pharmacokinetic characterization of RP-3500 revealed acceptable to good oral bioavailability across species (F(%) = 72/45/113/56 in mouse/rat/dog/monkey) with rat exhibiting the highest clearance of all the species tested. Unfortunately, erythroid cell toxicity is a commonly encountered issue observed with ATR inhibitors in vivo. To minimize the impact of anemia in live subjects, RP-3500 was subject to a variety of dosing regimes to deconvolute the kinetics of erythroblast depletion/recovery in response to the compound. Intermittent dosing schedules (i.e. 3 days on/4 days off) were found to be the most effective. Furthermore, the authors showed that coadministration of RP-3500 with PARP inhibitors (Niraparib) concurrently on an intermittent dosing schedule could effectively control tumor growth in Granta-519 tumor-bearing mice without severely impacting the erythroid cellular compartment. With a number of active and recruiting clinical trials in play, Camosertinib looks to ameliorate the dose-limiting toxicities that beleaguered previous ATR inhibitors.
Reference: https://doi.org/10.1021/acs.jmedchem.3c01917
CMX990
CMX990: Leveraging its drug repurposing library ReFRAME (Repurposing, Focused Rescue, and Accelerated Medchem), medicinal chemists at Calibr found numerous starting points that could be quickly pushed to active antivirals at the height of the COVID pandemic. A number of initial hits were found to target SARS-CoV-2 3C-like protease (CLpro); a protein responsible for the processing/cleavage of virally encoded polyproteins. As a starting point, a peptidomimetic compound developed by Pfizer (PF-00835231) was selected and its structure optimized to increase efficacy and decrease clearance. The first major advance in the SAR was incorporation of the trifluoromethoxymethyl ketone as the cysteine engaging warhead substituent. This group provided a 3-fold increase in potency and improved in vitro (human liver microsome) by > 30-fold. Tethering of the middle portion of the compound via proline further enhanced the pharmacokinetic properties of the compound, leading to enhanced oral exposure. Characterization of CMX990 in cells revealed potent activity in HeLa cell models of SARS-CoV-2 alpha, delta, and omicron infection (63 nM, 35 nM, and 37 nM, respectively). ADMET analysis confirmed lower clearance in human liver hepatocytes and microsomes compared to Nirmatrelvir (> 3-fold improvement) and no observed hERG liabilities (IC50 > 10 μM) or CYP inhibition (IC50 > 45 μM across 5 major cytochromes). Measures of genotoxicity (i.e. Ames and MNT tests) were also negative. In pharmacokinetic studies, CMX990 exhibited tractable to good oral bioavailability in all species tested other than monkey (F(%) = 14.5/12.2/52.8/1.1 in mouse/rat/dog/monkey). Follow up MTD studies in CD-1 mice and beagle revealed no sustained adverse effects, even at doses of 1000 mg/kg/day for 5-days. Remarkably, CMX990 was progressed to a Phase I clinical trial in only 10 months after its synthesis, however, no further clinical development has been disclosed. Taken as a whole, CMX990 is a sobering reminder of the need to develop pharmaceuticals against targets without a human homolog.
Reference: https://doi.org/10.1021/acs.jmedchem.3c01938
BAY-9835
BAY-9835: Matrix metalloprotease ADAMTS7 has been implicated it the development and progression of coronary artery disease (CAD). Multiple genome-wide association studies, as well as knockout and catalytic domain loss of function studies, have made ADAMTS7 an intriguing target for therapeutic intervention. However, selectivity across the 19-member zinc metalloprotease family has historically been difficult to attain. The medicinal chemists at Bayer took an in silico design approach, looking for tractable differences between the catalytic domains of ADAMTS7 and other related metalloproteases to tease apart tractable pockets that could be leveraged for selectivity. The scientist used a known ADAMTS4/5/7 inhibitor (Eli Lilly; https://doi.org/10.1021/acs.jmedchem.7b00650) as a jump off point and gradually tuned its selectivity through iterative rounds of SAR. One of the most impactful design decisions was the incorporation of ortho-biphenyl carboxamide to bolster selectivity for ADAMTS7 over the closely related MMP12. Judicious placement of fluorine on this substituent helped fix metabolic soft spots and contributed to the overall bioavailability of the compound. Additionally, the methyl pyrazole moiety provided a good trade-off between potency, selectivity, and pharmacokinetic properties. In vitro, BAY-9835 exhibited activity against ADAMTS7 (IC50 = 6 nM) and exquisite selectivity in a panel of related metalloproteases (1121x/1654x/375x/5467x/962x/17316x selectivity against ADAMTS4/ ADAMTS5/ADAM8/ADAM10/ADAM17/MMP2). Lower selectivity was observed for ADAMTS12 (5x). The authors hinted at further work aimed at improving selectivity by modifying the methyl pyrazole substituent, leveraging a potential steric clash with Phe354 of ADAMTS12. Pharmacokinetic characterization of the compound revealed low blood clearance (CLblood (L/kg/h) = 1.1/0.55/0.02 in mouse/rat/dog) and great bioavailability (F(%) = 100/96/77 in mouse/rat/dog) across species. In toxicity studies, no inhibition of cardiac ion channels (hERG, hNav1.5, hCav1.2) was observed below 10 μM, and tests for genotoxicity (Ames and MNT) were negative. In vivo, BAY-9835 was well tolerated in a two-week chronic dosing study in rats (50 mg/kg/day, p.o.) with no observable signs of toxicity. In summation, BAY-9835 possesses desirable activity and selectivity, along with tractable pharmacokinetic properties, which set it apart as one of the first orally bioavailable and selective molecules targeting ADAMTS7.
Reference: https://doi.org/10.1021/acs.jmedchem.3c02036
ACT-1016-0707
ACT-1016-0707: Idorsia Pharmaceuticals recently disclosed their lead optimization of an LPAR1 receptor antagonist aimed at overcoming the poor PK of their antecedent molecule(s) (compound 3). More specifically, low unbound fraction and high clearance in rat made PK and dose predictions in humans untenable. Towards this end, the medicinal chemistry team looked at reducing the lipophilicity of the original scaffold to find the optimal balance of efficacy and in vivo PK. Contraction of the central piperidine ring to an azetidine offered tractable gains in solubility and metabolic stability. Also, dispersing polarity across the molecule (i.e introduction of the chloropyridine) had a similar effect. Finally, replacing the succinic acid group with a sulfamide circumvented transport via organo anion trasporters (OATP) and ameliorated enterohepatic recirculation issues observed in close congeners. In Tango EDG2-bla U2OS, ACT-1016-0707 exhibited potent and selective inhibition of the LPAR1 (IC50 = 2.9 nM; corrected for unbound fraction, IC50 >10 μM in LPAR2 and LPAR3 reporter cell lines). ADME and PK characterization revealed low clearance in hepatocytes (CLHep (μL/min/106 cells) = 7.4/n.d./<2 in rat/mouse/dog) and oral absorption (F(%) = 49/79/52 in rat/mouse/dog) warranting further evaluation in vivo. An LPA-induced skin vascular leakage mouse model was used to assess PD. In the model, ACT-1016-0707 was shown to significantly reduce vascular leakage at a dose of 30 mg/kg. Follow up toxicity tests (Ames and Eurofins Panlab safety screen) didn’t flag any substantial genotoxicity of promiscuity issues. In scope, the structural evolution of preclinical candidate ACT-1016-0707 nicely highlights the design considerations needed to arrive at a compound with a balanced ADMET, PK and efficacy profile.
Reference: https://doi.org/10.1021/acs.jmedchem.3c01827