Discovery of Chromane Propionic Acid Analogues as Selective Agonists of GPR120 with in Vivo Activity in Rodents

Abstract

GPR120 (FFAR4) is a fatty acid sensing G protein coupled receptor (GPCR) that has been identified as a target for possible treatment of type 2 diabetes. A selective activator of GPR120 containing a chromane scaffold has been designed, synthesized, and evaluated in vivo. Results of these efforts suggest that chromane propionic acid 18 is a suitable tool molecule for further animal studies. Compound 18 is selective over the closely related target GPR40 (FFAR1), has a clean off-target profile, demonstrates suitable pharmacokinetic properties, and has been evaluated in wild-type/knockout GPR120 mouse oGTT studies.

Type 2 diabetes mellitus (T2DM), a worldwide public health problem, is a metabolic disorder characterized by hyperglycemia, pancreatic β-cell impairment, impaired insulin secretion, and insulin resistance in various tissues, including skeletal muscle, adipose tissue, and liver.¹⁻³ Inflammation has been suggested to be a major contributor to this condition,⁴ and chronic complications such as cardiovascular disease, retinopathy, neuropathy, and nephropathy can be encountered by those afflicted with this disease.⁵

GPR120 (FFAR4) is a long-chain fatty acid sensing G protein-coupled receptor (GPCR) that is expressed in intestine, lung, macrophages, and adipose tissue.^6,7 Stimulation of this receptor by free fatty acids has been reported to promote secretion of glucagon-like peptide-1 (GLP-1).⁸ Activation has also been demonstrated to display insulin sensitizing effects in vivo,⁷ and as such GPR120 could be an attractive drug target.^9,10

Due to the above considerations, a suitable agonist of GPR120 has been proposed to be desirable as a possible treatment for type 2 diabetes.¹¹⁻¹⁵ Due to the presence of multiple free fatty acid receptors that are able to be stimulated by structurally similar ligands and the manifestation of the complex pharmacology that can result,¹⁶ we counterscreened against GPR40 (FFAR1) to identify compounds that are selective for activation of GPR120.^17,18 Herein is described our recent efforts to identify selective and potent agonists of GPR120 that have suitable pharmacokinetic profiles to support in vivo studies.

During the optimization of GPR120 agonists spiropiperidine 1 and benzofuran 2 (Figure 1), which were previously disclosed by our colleagues,¹⁹⁻²² it was discovered that replacement of the scaffold core with a chromane system, as in propionic acid 3, resulted in maintained agonist activity on GPR120.²³

Figure 1.

Comparison of core structures.

We decided to optimize this chromane series and identify a selective agonist via an in vitro evaluation using IP1 and β-Arrestin2 (β-Arr2) assays to measure on target activation of human (h) GPR120 in conjunction with an IP1 assay to determine activation of hGPR40.

Toward this end, we explored the effects of modifying the propionic acid side chain via substitution on the carbon chain and replacement of the acid with various isosteres. Multiple examples were prepared and tested, but most exhibited a loss of activity. The most promising derivatives are summarized in Table 1.

Table 1. SAR Overview of a Promising Series of Chromanes as GPR120 Agonists^a.

^aValues are the average of at least two experiments, each in 10-point titrations. Unless noted, all analogues tested were generally deemed to be full GPR120 agonists under assay conditions.

Both the R and S chromane enantiomers, 3 and 4, respectively, were found to be agonists of GPR120, but we focused on the R chromane scaffold for the majority of structure–activity relationship (SAR) development due to its slightly better in vitro activity in multiple examples.

Most substitutions on the side chain were found to be detrimental to activity, but it was discovered that a cyclopropionic acid was tolerated and could improve selectivity over GPR40 as observed in chromane 5. The stereochemical configuration of the cyclopropane was critical, as the (R,R,R)-isomer 5 displayed more potency than the (R,S,S)-cyclopropane diastereomer 6 for GPR120 activation.

The tetrazole functionality was identified as a possible carboxylic acid bioisostere, exemplified here by compound 7. This modification was accompanied by an attenuation of potency, and the tetrazole 7 seemed to display partial agonist activity when compared to a standard GPR120 agonist under the conditions used in the hGPR120 IP1 assay (>50% activation is considered full).

Concurrent to the SAR exploration on the acid side chain, the aryl substitution at the chromane 6-position was also modified (Table 2). One observed trend was that appropriate substitution ortho to the fluorine resulted in an improvement of in vitro potency on GPR120. Furthermore, analogues containing a trifluoromethoxy group para to the fluorine were not as selective against GPR40 as analogues where the substituent was replaced with a cyclobutoxy group. Replacement of the phenyl ring with a series of pyridine analogues was evaluated, and the 6-cyclobutoxy-3-fluoropyridin-2-yl derivative 16 displayed better potency on GPR120 than fluoro pyridines 14 or 15. Similar to the SAR observed when employing a phenyl ring, substitution ortho to the fluorine in the methoxy pyridine analogue 17 improved potency on GPR120.

Table 2. Chromane Biaryl SAR Optimization^a.

^aValues are the average of at least two experiments, each in 10-point titrations. Unless noted, all analogues tested were generally deemed to be full GPR120 agonists under assay conditions.

Further exploration of substitution on the cyclobutyl ring revealed that trans-methoxy substitution at the 3-position was tolerated in analogues 18 and 19. Aryl groups could also be used to replace the cyclobutyl ring with retention of potency, exemplified here by the methyl thiazoles 20 and 21.

Upon identification of these potent and selective compounds, promising derivatives were studied to determine their pharmacokinetic (PK) profiles in rats via IV dosing (Table 3). Oral bioavailability was >70% for compounds throughout this series (vide infra, Table 7), so optimization of rat half-life through lowering intrinsic clearance was a main factor that guided SAR development and tool compound selection. All compounds were >99% bound to plasma proteins. Analysis of intrinsic clearance (Cl_int) for this series suggested that chromanes 18 and 19 may be good choices as potential tool compound candidates due to their comparatively low Cl_int and longer half-lives in rats. Consequently, these compounds were selected for further in vivo evaluation.

Table 3. Rat IV Pharmacokinetic Data of Selected Chromane Derived GPR120 Agonists^a.

Compound	Clp	t1/2	Vdss	PPB	Clint
3	9.5	1.3	0.64	99.84	10500
11	5.4	1.3	0.45	99.93	8700
16	9.1	0.74	0.23	99.77	4500
18	1.3	2.1	0.24	99.76	560
19	3.4	2.4	0.49	99.68	760
20	1.6	0.93	0.14	99.97	6200
21	1.6	1.2	0.15	99.93	3600

^aCl_p, plasma clearance (mL/min/kg); t_1/2, terminal half-life (h); Vd_ss, volume of distribution at steady state (L/kg); PPB, plasma protein binding (% bound extrapolated from 10% plasma); Cl_int, intrinsic clearance (mL/min/kg). IV cassette dosing (DMSO/PEG400/H₂O = 20/60/20; 0.1 mg/kg).

Table 7. Pharmacokinetic Profile of Compound 18 Across Multiple Species^a.

property	rat	mouse	Rhesus
AUCIV (μM·h)	35.9	8.1	57.8
Clp (mL/min/kg)	1.1	9.9	0.33
Vdss (L/kg)	0.3	1.4	0.27
t1/2 (h)	4.6	3.0	9.8
MRT (h)	4.6	2.4	13.5
Foral (%)	93	76	82
PPB (% bound)	99.76	99.67	99.78
Clint (mL/min/kg)	470	3400	150

^aAUC_iv, area under the plasma concentration vs time curve following IV dosing; Cl_p, plasma clearance; Vd_ss, volume of distribution at steady state; t_1/2, terminal half-life; MRT, mean residence time; F_oral, oral bioavailability; PPB, plasma protein binding; Cl_int, intrinsic clearance. IV dosing in DMSO/PEG400/H₂O = 20/60/20 (rat 1 mg/kg, mouse 2 mg/kg, Rhesus 0.5 mpk). PO dosing via suspension in methylcellulose 0.5% (rat 2 mg/kg, mouse 30 mg/kg, Rhesus 1 mpk).

Promising propionic acid replacements were next evaluated in combination with the optimized aryl groups from Table 2. SAR trends were maintained in the case of the cyclopropane containing analogues (Table 4); substitution ortho to the fluorine improved activity, and trans-methoxy substitution was tolerated.

Table 4. Cyclopropionic Acid Analogues^a.

^aValues are the average of at least two experiments, each in 10-point titrations. Unless noted, all analogues tested were generally deemed to be full GPR120 agonists under assay conditions.

Congeners combining the best aryl groups and a replacement of the carboxylic acid with the tetrazole moiety were also synthesized and evaluated (Table 5). In the case of the tetrazole series, methoxy substituion ortho to the fluorine provided a more marked improvement in potency on GPR120 than did the difluoro substitution. The analogues with weaker potency seemed to display partial activity on hGPR120 under the conditions of the IP1 assay. This partial activity was not observed in the β-Arr2 assay, however, and upon identifying analogues with improved potency this partial behavior returned to full agonism. Importantly, compound 26, which displayed similar potency and full agonist activity in a mouse GPR120 IP1 assay (results not shown), demonstrated in vivo efficacy in a mouse oral glucose tolerance test (oGTT) (Figure 2). These results suggest that exploration of the tetrazole series or other carboxylic acid bioisosteres could potentially be an area for future investigation.

Table 5. Tetrazole Analogues^a.

^aValues are the average of at least two experiments, each in 10-point titrations. Unless noted, all analogues tested were generally deemed to be full GPR120 agonists under assay conditions.

Figure 2.

Glucose lowering effects as measured by mouse oGTT. The following doses were used, resulting in the corresponding plasma concentrations (C_p) at 3 h: 18, 3 mg/kg [n = 3, 0.75 μM (s = 0.04 μM)]; 10 mg/kg [n = 3, 1.17 μM (s = 0.13 μM)]; 30 mg/kg [n = 3, 3.12 μM (s = 1.02 μM); 24, 30 mg/kg [n = 3, 37.4 μM (s = 3.43 μM)]; 26, 100 mg/kg [n = 3, 40.6 μM (s = 16.4 μM)]. (A) Blood glucose (mg/dL) over time (minutes). (B) Net blood glucose area under the curve (AUC). Vehicle (veh); dextrose (dex) challenge; Tx = vehicle and compound administration; mpk = mg/kg. **P < 0.01 compared to vehicle, ***P < 0.0001 compared to vehicle; two-way ANOVA.

We evaluated compounds 18, 24, and 26in vivo using the mouse oral glucose tolerance test (oGTT).²⁴Figure 2 summarizes the results of compounds 18 (3, 10, 30 mg/kg) in a dose titration and single dose evaluation of 24 (30 mg/kg) and 26 (100 mg/kg). All three compounds showed reduction of blood glucose, demonstrating that propionic acid replacements such as the cyclopropionic acid and the tetrazole are capable of exhibiting in vivo glucose lowering activity. Furthermore, compound 18 showed a dose-dependent glucose reduction with a plateau of efficacy observed at 10 mg/kg thus defining the minimum dose that afforded maximal efficacy (MED_max) in this model.

In a further set of wild-type/GPR120 knockout mouse oGTT experiments, compounds 18 and 19 both displayed glucose lowering at the superpharmacologic dose of 150 mg/kg in wild-type mice (Figure 3). Meanwhile, no efficacy was observed in the GPR120 knockout mice, suggesting that glucose lowering observed at very high doses is still linked to GPR120 activation. The in vitro profile of compounds 18 and 19 on mouse (m) GPR120 and mGPR40 was similar to that observed for the human receptor (Table 6).

Figure 3.

(A,B) Mouse oGTT of compound 18 in wild-type [150 mg/kg; n = 4, 3 h C_p= 43.9 μM (s = 15.8 μM)] and GPR120 knockout [150 mg/kg; n = 4, 3 h C_p = 53.4 μM (s = 14.0 μM)] mice. (C,D) Mouse oGTT of compound 19 in wild-type [150 mg/kg; n = 4, 3 h C_p = 33.5 μM (s = 21.0 μM)] and GPR120 knockout [150 mg/kg; n = 4, 3 h C_p = 27.6 μM (s = 6.03 μM)] mice. (A,C) Blood glucose (mg/dl) over time (minutes). (B,D) Net blood glucose area under the curve (AUC). Vehicle (veh); dextrose (dex) challenge; Tx = vehicle and compound administration; mpk = mg/kg. **P < 0.01 compared to vehicle, two-way ANOVA.

Table 6. Mouse in Vitro Data for Compounds 18 and 19^a.

	mGPR120 EC50 (nM)
Compound	IP1	β-Arr2	mGPR40 IP1 EC50 (nM)
18	6.4	6.8	>10k
19	5.6	4.6	>10k

^aValues are the average of at least two experiments, each in 10-point titrations. Unless noted, all analogues tested were generally deemed to be full GPR120 agonists under assay conditions.

Compound 18 was evaluated for potency in common off-target in vitro assays. Ion channel activity was determined to be minimal (IC₅₀s at Ca_v1.2 = 28 μM,²⁵ IKr > 60 μM,²⁶ Na_v1.5 = 27 μM²⁷), and the cytochrome P-450 inhibition was also negligible (CYP3A4, CYP2D6, CYP2C9 IC_50s > 50 μM).²⁸ Additionally, compound 18 was characterized in a general broad panel screen of enzyme, ion channel, and GPCR assays and displayed minimal activity, including at PPAR alpha, delta, and gamma (>10 μM).

The pharmacokinetic profile of compound 18 was evaluated in multiple species, including rat in a higher dose study, and determined to be suitable for further in vivo evaluation in various PD models (Table 7). High oral bioavailabilities were observed as well as relatively long half-lives in these species. The pharmacokinetic profile in Rhesus was particularly encouraging, as it suggests that this series could be optimized toward efficient oral dosing in primates.

In summary, the optimization of a potent and selective series of chromane derived GPR120 agonists was described. Selected compounds from the series exhibit pharmacokinetic properties in the rat that supported their use as in vivo tools. Three compounds demonstrated glucose lowering in mouse oGTT studies, including cyclopropionic acid 24 and tetrazole 26, which may offer new future directions for teams pursuing GPR120 agonists. Finally, two compounds (18 and 19) demonstrated mechanism-based glucose lowering even at superpharmacologic exposures in mice, suggesting that these compounds will be good tools for further in vivo studies, which will be reported in due course.

Glossary

ABBREVIATIONS

FFAR1

free fatty acid receptor 1

FFAR4

free fatty acid receptor 4

T2DM

type 2 diabetes mellitus

GLP-1

glucagon like peptide 1

GPR120

G-protein coupled receptor 120

GPCR

G-protein coupled receptor

GPR40

G-protein coupled receptor 40

IP1

d-myo-inositol-1-phosphate

oGTT

oral glucose tolerance test

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.6b00394.

• Experimental procedures and analytical data for the synthesis of all compounds; GPR120 and GPR40 in vitro assay procedures; PK protocol; oGTT protocols (PDF)

Author Present Address

^# (U.S.) Jones Day, 250 Vesey Street, New York, NY 10281, USA.

Author Present Address

^⊥ (M.K.) Charles River Laboratories, 334 South St., Shrewsbury, MA 01545, USA.

The authors declare no competing financial interest.