|Home | About | Journals | Submit | Contact Us | Français|
A novel series of biaryl amides was identified as RORγt inhibitors through core replacement of a starting hit 1. Structure–activity relationship exploration on the biaryl moiety led to discovery of potent RORγt inhibitors with good oral bioavailability and CNS penetration. Compounds 9a and 9g demonstrated excellent in vivo efficacy in EAE mice dose dependently with once daily oral administration.
T helper (Th) 17 cells, a lineage of CD4+ effector T cells characterized by the production of IL-17A and IL-17F, are pathogenic in human autoimmune inflammatory diseases including multiple sclerosis (MS).1−4 The presence of IL-17 was detected in MS lesions, and Th17 cells were observed in the infiltrations of mouse experimental autoimmune encephalomyelitis (EAE) central nervous system (CNS).5,6 Differentiation and function of Th17 cells are controlled by the transcription factor retinoic acid receptor-related orphan receptor-gamma-t (RORγt).7−9,11 It has been shown that the genetic deficiency of RORγt in mice severely impaired Th17 cell differentiation and conferred resistance to EAE.10 RORγt inhibitors has potential utility in reducing the activity of Th17 cells and therefore can be developed as therapeutic agents for the treatment of Th17 cell mediated autoimmune diseases.12−18
A few small molecule RORγt inhibitors have been reported in literature.19 Digoxin,20 SR1001,21 and ursolic acid22 were first reported to inhibit RORγt and ameliorate EAE in mice via intraperitoneal administration. Other small molecular RORγt inhibitors23−31 were later disclosed. Recently, we reported discovery of thiazole ketone amides (e.g., 2) and thiophene ketone amides (e.g., 3) as novel RORγt inhibitors based on a high throughput screening (HTS) hit 1 (Figure (Figure11).32 These ketones, especially the thiophene ketones, showed good RORγt activities but were poorly orally bioavailable and lack of CNS penetration that is believed to be important for developing an effective oral MS drug. In this Letter, we report the discovery of novel biaryl amides as first potent, orally bioavailable, and CNS penetrant RORγt inhibitors, which demonstrated EAE in vivo efficacy dose dependently via oral administration.
The lack of CNS penetration of thiazole/thiophene ketones was attributed to their ketone moiety as the nonketone thiazole amide 1 is CNS penetrant with a brain-to-blood ratio (Br/Bl) of 1.5 in a mouse CNS study (i.p., 2 mg/kg).33 Encouraged by the CNS data of 1, we conducted thiazole core replacement with a number of different aromatic rings (4a–4i), aiming to identify a suitable scaffold for multiproperty optimization (Table 1). Among the five-membered ring analogues, 2,4-substituted thiophene 4c showed the best RORγt potency in the FRET assay.32,34 The heteroatom such as oxygen and nitrogen on the ring decreased (4a) and even abolished (4d, 4f, 4h) RORγt activities. For the six-membered ring analogues, para-substituted aryl amide (4i) showed better RORγt potency than the meta-substituted one (4g). Because of its reasonable RORγt potency, good CNS penetration (Br/Bl = 2.0), improved ligand efficiency (LE) and lipophilic ligand efficiency (LLE) (0.33 and 2.3 for 4i compared to 0.29 and 1.9 for 1, respectively),35 and easy modification/diversification, the aryl amide 4i was used as the new chemistry starting point for optimization.
In order to explore the structure–activity relationship (SAR) of the biaryl moiety of the amide, a versatile synthesis of the general structures of biaryl amides was developed (Scheme 1).36 Biaryl amines 7 were prepared from either bromoanilines 5 through Suzuki coupling with aryl boronic acids or from reaction of aryl bromides with aniline boronic esters 6, obtained from 5. Coupling 7 with acids A, or acid chlorides B, or perfluorophenyl esters C afforded the desired biaryl amides 8 or 9. The biaryl amides could also be prepared by first coupling of 5 with A to form amides 10, which were converted to the target compounds directly via Suzuki coupling, or via its boronic ester intermediate 11.
We investigated the binding mode of compound 4i and its derivatives in RORγt LBD based on the cocrystal structure of a similar aryl amide with RORγt LBD (pdb code: 4NIE).37 The perpendicular confirmation of the two aryl rings in the left-hand side (LHS) of the amides provided preferred intermolecular interactions with the surrounding hydrophobic residues in the RORγt LBD and was believed to be important for the RORγt binding affinity (Figure (Figure2).2). Subsequently, the substitutions on the ortho-positions of the two aryl rings, which force the two aryls to take perpendicular conformation, were studied extensively, and the key SAR of the biaryls was summarized in Table 2. Nonsubstituted biphenyl amide 8a showed a RORγ FRET pIC50 of 6.3. Adding a Cl group on ortho-position of the central phenyl ring (8b) enhanced RORγt activity. Keeping the ortho-Cl on the central phenyl ring, adding a hydrophobic group on ortho-position of the terminal phenyl ring provided potent RORγt inhibitors (8c–8f) with pIC50s ≥ 8.0 in the FRET assay. Biaryl amides 8c–8f also showed good cellular activities in the Th17 cell differentiation assay (pIC50 > 6.0).32−34 Obviously, the cLogPs of 8c–8f are relatively high (4.4–5.1, from ChemBioDraw Ultra 12.0). Replacing −OiPr moiety (8f) with −CH2NMe2 (8g, cLogP 3.7) or changing the phenyl ring to a pyridine ring (8h, cLogP 4.0) or other heteroaromatic rings such as pyrole (8i, cLogP 3.5) lowered RORγt potency, which also resulted in essentially no activity in Th17 cell differentiation assay. These findings indicate that the binding pocket where the LHS aryl occupies is hydrophobic and unable to tolerate some polar moieties.
We then fixed the OCF3 group at ortho-position of the LHS terminal aryl and studied SAR of substitution on the central phenyl ring of the amides (Table 2). Adding a hydrophobic group such as methyl (8l) or CF3 (8m) on the central phenyl ring increased Th17 potency while replacing the central phenyl ring with pyridine (8k) significantly decreased both RORγt and Th17 potency. Although polar groups like carboxylic acid (8o) decreased the RORγt potency dramatically, certain polar groups such as acetyl (8n) were found to be tolerated on the central phenyl. Encouraged by this, a number of heteroaromatic rings were introduced on the central phenyl and the resulting compounds (e.g., 8p, 8q) showed good RORγt activity in both FRET and Th17 assays. It is good to see that the cLogP of 8p and 8q are relatively lower (3.6 and 3.0, respectively), resulting in higher LLE values (4.7 and 5.4, respectively) although their molecular weights are higher.
We next added a second substituent to the central phenyl ring to constrain the preferred perpendicular conformation. As expected, additional substituent CN (8r), Me (8s), or Cl (8t) boosted RORγt activities in both FRET and Th17 assays.
Compound 8t was used as a tool compound for RORγt biological studies because of its excellent in vitro activities as well as good oral exposure and CNS penetration.38 Encouraged by the profile of 8t, we incorporated the previous SAR learnings and further optimized the LHS biaryl part as well as right-hand side (RHS) sulfone part of the amides, trying to obtain a molecule with more balanced profile (Table 3). Changing the ethyl sulfone in 8t with a methyl sulfone (9a) resulted in a similar RORγt potency and CNS penetration. However, replacing the methyl sulfone with a primary sulfonamide (9b) basically eliminates the CNS penetration although the RORγt and Th17 potency remained, possibly due to introduction of two more H-bond donors as well as increase of topological polar surface area (tPSA) in 9b. Switching OCF3 (9a) to OCF2 (9c) lowered its CNS penetration. The CNS penetration was further decreased when OCF2 (9c) was replaced by a CN group (9d). With a Cl group in the para-position of LHS phenyl and only one substituent (F, Me, or Cl) in the ortho-position of central phenyl, compounds 9f–9h showed good RORγt potency and CNS penetration. Compared to methyl sulfone 9h, the ethyl sulfone 9i demonstrated the best CNS penetration (Br/Bl = 2.0). Clearly, the data of CNS penetration were well correlated to values of tPSA and/or cLogP. As a result, LLE value is relatively low for those biaryl amides with better CNS penetration (Table 3).
Several representative compounds were evaluated for their mouse PK profile (Table 4). Biaryl amides 8d, 8e, and 9a demonstrated good PK profile with oral bioavailabilities of 75%, 102%, and 100%, respectively. Compound 9g was only evaluated via po administration and showed excellent oral exposure.
With good Th17 activity and mouse oral exposure, we then evaluated 9a and 9g in EAE mice where Th17 cells play a critical role (Figure (Figure33).33 Compounds 9a and 9g were orally administered once daily at 3 doses (1, 3, and 10 mg/kg) to EAE mice from the day of immunization. Compared to the control, the treatment with 9a or 9g resulted in a delay and significant reduction in clinical severity of EAE in a dose-dependent manner. Compared to thiazole ketone amide 2, which only showed EAE efficacy up to day 20 at 100 mg/kg twice daily dosing,32 the biaryl amides 9a and 9g are much more efficacious. This could be attributed to their good in vitro activities as well as much improved oral exposure and CNS penetration. However, it should be noted that although 9g had more brain exposure than 9a, it exhibited less efficacy than 9a in EAE experiments, indicating that there might be additional factors such as “free” brain concentration affecting in vivo efficacy.
In summary, we have discovered a novel series of biaryl amides as RORγt inhibitors. Detailed SAR study on the LHS biaryl moiety of the amides led to discovery of potent RORγt inhibitors with excellent oral bioavailability and CNS penetration. The key compounds 9a and 9g demonstrated a dose-dependent EAE efficacy in mice when administrated orally once a day. Further optimization on sulfone moiety of the biaryl amides to balance potency and some developability properties such as solubility is ongoing.
We thank Wei Zhang, Jing Zhang, Gang An, Jinqiang Zhang, Melanie Huxdorf, Yingli Ma, Shuai Wang, Hui Lei, Hongtao Lu, and Zhong Zhong for their help and discussions.
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.
The authors declare no competing financial interest.