br Results and discussion The synthesis
Results and discussion The synthesis of core aldehyde building block 2 commenced with known aldehyde 5. The Horner–Wadsworth-Emmons olefination of the latter (using commercially available 6) gave cinnamate ester 7. Facile hydrogenation of the double bond in 7 was achieved over Raney nickel without the disruption of the benzyl ether. Selective hydrolysis of the methyl ester in resulting tert-butyl 3-phenylpropionate 8 and subsequent mixed anhydride reduction of carboxylic Levonorgestrel 9 furnished target building block 1 (Scheme 1). The diverse set of periphery spirocyclic building blocks 3a–r was synthesized from progenitor 1 as summarized in Scheme 2. Sodium alkoxide, generated via treatment of 1 with sodium hydride in DMF, was alkylated with a set of alkyl halides to furnish, after Boc group removal, building blocks 3a–g. Multigram-scale oxidation of 1 with PDC furnished ketone 10 which served as a starting material for preparation of the rest of building blocks 3h–r. The Wittig reaction of 10 with phosphonium ylides (generated by treatment of the respective phosphonium salts with n-BuLi) furnished respective olefins, all of which, except for unsubstituted methylene compound, were obtained as difficult-to-separate mixtures of E- and Z-isomers. After brief fractionation on silica gel, the olefin intermediates were hydrogenated over 10% Pd on carbon using ammonium formate as a source of hydrogen. Upon removal of the Boc group, a set of building blocks 3h–o was obtained. Considering the previous success in grafting polar heterocycles onto the periphery of FFA1 agonists, we designed 1,2,4-oxadiazole containing spirocyclic piperidine building blocks 3p–r. The synthesis of the latter was achieved via a multistep reaction sequence which included Horner–Wadsworth–Emmons olefination, reduction of the α,β-unsaturate ester, alkaline hydrolysis of its saturated counterpart to provide common carboxylic acid starting material 11. Activation of the carboxylic functionality in the latter with isobutyl chloroformate followed by the reaction with the respective amidoximes, cyclodehydration of the resulting initial adducts on treatment with TBAF in refluxing toluene, and Boc group removal furnished target 1,2,4-oxadiazoles 3p–r (Scheme 2). Spirocyclic building blocks 3a–r thus synthesized (as well as commercially available unsubstituted 1-oxa-9-azaspiro[5.5]undecane 12) were then employed in reductive amination of aldehyde 2 and, after brief fractionation of the tertiary amine products, the Boc protecting group was removed using TFA in CH2Cl2, followed by counter-ion exchange with HCl in diethyl ether, to furnish the target compounds 4a–s (Scheme 3). It should be noted that, owing to difficulties we faced in purification of the hydrochloride products, which required repeated manipulation of semi-purified material, the isolated yields of compounds 4a–s achieved in this study were relatively low (see ESI). Nonetheless, the synthetic strategy described above allowed the preparation of the required quantities of compounds 4a–s for the initial in vitro biological profiling of the lead compounds thus identified—for in vitro ADME and in vivo PK characterization. Further optimization of the synthesis of the frontrunner compounds will be needed before further development of the series disclosed herein. Potential FFA1 agonists 4a–s thus synthesized were tested for FFA1 activation using calcium flux assay employing Chinese hamster ovary (CHO) cells engineered to stably express human FFA1. All compounds were tested in dose–response (% FFA1 activation) mode in order to determined EC50 values. The latter, as well as % of maximum efficacy achieved for active compounds relative to reference agonist GW9508 (used at 5μM concentration throughout the study) are given in Table 1. On examination of the FFA1 activation data presented in Table 1 it becomes clear that in the newly designed series, increased lipophilicity (in combination with the basic nitrogen center counter-balancing the lipophilic periphery) is the principal driver of potency, as it was described for FFA1 agonists in a number of instances.10, 17 This should be expected for receptors modulated by endogenous free fatty acids and for the ligands intended to mimic the lipophilic character of the latter. Indeed, the hydrophilicity imparted by unsubstituted 1-oxa-9-azaspiro[5.5]undecane periphery (compound 4s) was detrimental to the agonistic potency. However, introduction of alkoxy substituents of increasing size (compound 4a–g) gradually brought the potency back into low-micromolar range. Of particular interest is the influence of alkyl substituents in the pyran portion of the new FFA1 agonist scaffold of the potency (exemplified by compounds 4h–o). Here, introduction of lipophilic benzyl periphery results in a significant potency increase, with the most compound (4l) having EC50 comparable to that of Takeda’s phase III compound TAK-875 (EC50=0.014μM). Taking into account the size of the pilot set of compounds studied in this work, this attests to the success of the initial FFA1 design idea implemented in the new molecular scaffold.