We have previously disclosed the discovery of a azaindole ac

We have previously disclosed the discovery of a 7-azaindole-3-acetic moxalactam receptor CRTh2 antagonist chemotype 1 from HTS and its subsequent optimization to give 2 as a potent and orally bioavailable inhibitor of eosinophil shape change (SC) in human whole blood. Subsequent to disclosure of that work, others have reported tricyclic azaindoles as CRTh2 antagonists. Following further characterization of 2, some issues with bulk stability under basic conditions became apparent, which resulted in it not being progressed to pre-clinical candidate selection. During the early hit explosion phase, attracted by the potential to further reduce molecular weight and increase ligand efficiency while maintaining the acceptable pharmacokinetic properties of the azaindole-3-acetic acid scaffold, the N-1 methylene analogue 3 was prepared. This compound was found to compare favorably with the original sulfonamide hit 1 as a functional CRTh2 antagonist, as indicated by activity in cAMP CRTh2 transfected CHO whole cell and DK-PGD2 induced human eosinophil SC assays (Table 1). In addition, in common with 1, 3 showed no significant inhibition of human COX-1 or COX-2 enzymes at concentrations up to 10μM. Herein we describe the further optimization of this prototype to provide a CRTh2 antagonist candidate molecule which has progressed into human clinical studies.
Results and discussion The key building block 2-methyl-1H-pyrrolo[2,3-b]pyridin-3-yl)acetic acid methyl ester 4 was prepared as previously described (full details are provided in Supplementary data) and could be readily alkylated using sodium hydride or alternatively BEMP as a base (Scheme 1). The latter reagent was found more convenient for parallel array approaches. The one benzyl bromide precursor (to compound 20) which was not either commercially available or known in the literature was prepared as indicated (Scheme 2). Initial results from exploration of CRTh2 binding SAR of the N-1 position (Table 2) revealed a rather limited dynamic potency range within an array of simple aromatic substituents derived from a Topliss-type approach (compounds 5–12). However, this endeavor did unveil some divergence from the sulfonamide series SAR—most notably the modest binding affinity for 12 compared with 2. With no increase in binding affinity from the original hit, and mindful of the increased lipophilicity conferred by the ‘sulfonamide to methylene’ switch, pyridylmethyl analogues 13–15 were explored to attempt to address this issue. However, binding activity was found to be in the μM range. Mindful of the learnings from the sulfonamide series which indicated a beneficial effect of reduced lipophilicity on eosinophil SC activity, a follow up array of 4-substituted benzyl analogues was next prepared, with the key results being indicated in Table 3. Once again, only a modest overall potency spread in the CRTh2 binding assay was obtained for 16, 17, 19 and 20, with all compounds being confirmed as functional antagonists in the cAMP assay with low-to-moderate shifts in the presence of HSA. However we observed that the sulfone 18 appeared to deliver the optimal balance of electron deficiency combined with reduced lipophilicity, leading to a marked improvement in potency across the panel of CRTh2 assays. The marked increase in eosinophil shape change potency compared with CRTh2 binding assay potency is in line with observations in the sulfonamide sub-series. The pharmacological rationale for this improvement is under active investigation and will be the subject of a forthcoming publication. Next, we were able to demonstrate that compound 18 retained the favorable pharmacokinetic properties already established for the azaindole acetic acid scaffold. In vitro Caco-2 permeability was moderate (Papp A–B/B–A 0.9/2.2×10−6cms−1), with good solubility of 5.2mg/ml at pH 6.8 determined for crystalline material. In vivo in the rat (Table 4) a relatively high volume of distribution for a carboxylic acid was observed, combined with moderate clearance which led to a prolonged terminal half life. The compound was found to be orally bioavailable following dosing of a crystalline suspension formulation, with an early Tmax in line with the favorable solubility of the compound. Follow-up rat pharmacokinetic studies with escalation of the dose gave appropriate increases in both exposure and Cmax, indicating suitability of the profile of 18 for preclinical toxicology studies. In addition, plasma exposure of 18 was also high in mice, with a Cmax of 7.6μM after a 5mg/kg oral dose. Additional pharmacokinetic and metabolic characterization studies of compound 18 will be the subject of a forthcoming publication.