Hydantoin in region F bridging the phenoxybenzene and
Hydantoin in region F bridging the phenoxybenzene and benzimidazole moieties was replaced by cyclic building blocks, diketopiperazine (25) and imidazolidinone (26), and linear bonds, amide (27) and urea (28). Analogs 25–28 possessing the linkers other than hydantoin were not tolerated in LBD of FXR, suggesting that hydantoin in region F would be necessary in order to retain binding affinity and antagonistic activity. We finally turned our attention to the external aromatic ring of 12 (region G) to verify the impact on antagonism and prepared 29–32 since removal of the external ring indicated a tendency to greatly lower antagonism against FXR in the hit-to-lead approach. 4-Phenoxy derivatives in analogs (29–31) retained potency for FXR in both assays. Likewise IWP-L6 mg 12, in particular 30 revealed nanomolar potency against FXR in both assays (0.007 ± 0.002 μM and <0.001 nM). Molecular modeling studies implied that the hydroxyl group of the 4-phenoxyl ring in 30 (blue) seemed to be protruding into solvent environment without large adverse effects on the binding mode of 12 (yellow) complexed with LBD of FXR, and to have no hydrogen-bonding interaction with the binding site of FXR. (Fig. S2) This result supports the observation that 30 retains potency for FXR at the same level as 12. Replacement of the external phenyl ring with an aliphatic ring (32) failed to significantly affect both activities. It is indicated that when phenoxyl derivatives are in region G, no significant decrease in inhibition against FXR was observed. The results on the building blocks of 12 are briefly summarized in Fig. 3. The regions and substitution pattern, which could preserve antagonism against FXR of 12, are as follows; (a) in region D, the small substituents at positions 5 or 6 on benzimidazole are tolerated for LBD in FXR (12, 22); (b) hydantoin in region F is the moiety that cannot be altered in terms of retention of inhibition; and (c) para-substitution on the phenyl ring in region G can be employed to maintain inhibition against FXR at or below nanomolar levels (29–31). More important to the chemotype as represented by 12, 22 and 30, an isobutyryl piperidine moiety in regions B and C including the S-configuration in region A was essential to sustain the antagonism against FXR. To confirm the validity of the in vitro activity, the cytotoxic activities of 13–32 were evaluated by tetrazolium (MTT) colorimetric assays according to the same protocol as described previously. (Table S2) Analogs 22 and 30, which exhibited robust antagonism, had no cytotoxicity. Thus, our structure-activity analyses of 12 culminated in the identification of 30 that preserved equivalent activity of 12. In regards to other biological characterizations, 12 prevented the activation of promoters related to FXR activated by the endogenous and synthetic FXR agonists, and was without agonism and antagonism against other nuclear receptors except FXR. Like 12, compound 30 was also further investigated. (Figs 3, S3 and S4A–S4J) The downstream genes of FXR by real-time RT-PCR as shown previously were analyzed. Cells cultured in six well plates were treated with different concentrations of 30 in the presence or absence of endogenous agonist, CDCA (Fig. 4) and synthetic agonist, GW4064 (Fig. S3). After 24 h, cells were collected. CDCA and GW4064 induced the expression of BSEP, SHP, and OSTα.5, 6, 24 Analog 30 inhibited the activation of promoters related to FXR activated by CDCA and GW4064. Next we evaluated the interaction with nine nuclear receptors including FXR, retinoid X receptor α (RXRα), peroxisome proliferator-activated receptor α, γ, and δ (PPAR α, γ, and δ), liver X receptor α and β (LXR α and β), vitamin D receptor (VDR), retinoic acid receptor α (RARα) and TGR5 according to published methods. Analog 30 prevented CDCA-induced FXR activity (Fig. S4A) and had no agonist (alone) or antagonist (plus the corresponding agonists) effects toward any of these receptors except FXR (Figs. S4B–S4J).