Design, synthesis and biological evaluation of novel FFA1/GPR40 agonists: New breakthrough in an old scaffold
Zheng Li, Chunxia Liu, Jianyong Yang, Jiaqi Zhou, Zhiwen Ye, Dazhi Feng, Na Yue, Jiayi Tong, Wenlong Huang, Hai Qian
ABSTRACT: Based on an old phenoxyacetic acid scaffold, CPU014 (compound 14) has been identified as a superior agonist by comprehensive exploration of structure-activity relationship. In vitro toxicity study suggested that CPU014 has lower risk of hepatotoxicity than TAK-875. During acute toxicity study (5 to 500 mg/kg), a favorable therapeutic window of CPU014 was observed by evaluation of plasma profiles and liver slices. Moreover, CPU014 promotes insulin secretion in a glucose-dependent manner, while no GLP-1 secretion has been enhanced. Other than good pharmacokinetic properties, CPU014 significantly improved glucose tolerance both in normal and diabetic models without the risk of hypoglycemia. These subversive findings provided a safer candidate CPU014, which is currently in preclinical study to assess its potential for the treatment of diabetes.
1.Introduction
Type 2 diabetes mellitus (T2DM), a prevalent metabolic disorder, is mainly displayed as impaired glucose tolerance followed by insulin insufficient and/or resistance.[1, 2] Many typical insulin secretagogues including sulfonylureas and meglitinides are widely used in clinic.[3] But nevertheless, these drugs usually have the high risk of hypoglycemia because their effects on promoting insulin secretion are independent of the concentration of plasma glucose.[4] The free fatty acid receptor 1 (FFA1), also known as G-protein coupled receptor 40 (GPR40), has emerged as a potential anti-diabetic target in the last decade.[5]
In pancreatic β-cells, FFA1 amplifies the glucose-stimulated insulin secretion, which affords the potential advantage of reducing the possibility of hypoglycemia in comparison to sulfonylureas.[6-8] Furthermore, specific tissue expression of FFA1 decreases the risk of FFA1-induced adverse effects.[9] As summarized in recent review[10], many structurally diverse FFA1 agonists have been reported (Figure 1).[11-21] More recently, a series of novel FFA1 agonists were also explored based on hybrid of quercetin and oleic acid.[22, 23] Several advanced compounds AMG-837, LY2881835, and TAK-875 have been evaluated in clinical trials for the management of T2DM.[5, 10] However, TAK-875 was terminated due to sign of hepatotoxicity[24] even though clinical evidences indicate the incidence of increased 3-fold alanine transaminase only slightly higher than placebo group.[25, 26] The exact molecular basis causing hepatotoxicity is unclear, but it is widely accepted that the hepatotoxicity is not target-related because FFA1 is low expressed in liver.[7, 27, 28] Recently, several studies suggested that TAK-875-induced hepatotoxicity might relate to the hepatobiliary transporters inhibition,[29, 30] the suppression of mitochondrial respiration, and covalent binding of acyl glucuronide with hepatocytes.[31]
In our previous researches, several scaffolds were introduced to explore more structure types of FFA1 agonists.[32-38] To avoid β-oxidation of phenylpropanoic acid moiety, several medicinal chemistry teams have replaced the phenylpropanoic acid with a phenoxyacetic acid scaffold, while the resulting phenoxyacetic acid derivatives revealed a significantly lower agonistic activity than the corresponding phenylpropanoic acid derivatives (Figure 2A and B).[18, 39] Therefore, the phenoxyacetic acid scaffold was abandoned until satisfactory results (e.g., compound 2, 2-(4-((4′-ethoxy-2′,6′-dimethyl-[1,1′-biphenyl]-3-yl)methoxy)-2-fluorophenoxy)acetic acid) were obtained in our laboratory (Figure 2C).[40, 41] Although the agonistic activity of compound 2 was still inferior to the corresponding phenylpropanoic acid derivative 1, the subversive findings have provided us with valuable guidance for further structure optimization. Herein, we describe further structural optimization of present series and our efforts toward the discovery of superior agonist (Figure 3).
Current structural optimizations often tend to provide compounds with higher lipophilicity, a critical property of drug-likeness because high lipophilicity is related to metabolic instability, high promiscuity, poor absorption, and thereby a higher risk of attrition.[42-46] To counteract this, the ligand efficiency (LE) and ligand lipophilicity efficiency (LLE) are increasingly being implemented to guide the structural optimization away from highly lipophilic and oversized candidates.[47, 48] Therefore, the LE and LLE, as well as lipophilicity (LogD7.4) were monitored in our SAR studies. All of these efforts ultimately subverted our previous bias for phenoxyacetic acid moiety and identified a new and safer agonist 14 (CPU014, EC50 = 15.7 nM), which revealed better in vitro and in vivo potency as well as a lower risk of hepatotoxicity than TAK-875.
2.Results and Discussion
2.1.Chemistry
The synthetic route of compounds 3 and 5-24 is illustrated in Scheme 1. Intermediates 3a or 3b were produced from the nitro reduction of intermediates 2a or 2b, which were prepared by Williamson ether synthesis of methyl chloroacetate with phenols 1a or 1b. The intermediates 6a-j were synthesized by standard alkylation of phenols 5a or 5b, which were derived from a Suzuki coupling reaction of (3-formylphenyl) boronic acid with bromobenzene 4a or 4b. Compounds 3a or 3b were connected with 6a-j by reductive amination reaction in the presence of NaBH3CN, followed by ester hydrolytic reaction with LiOH·H2O to provide the target compounds 3, 5-8 and 19-24. The target compounds 9-18 were obtained from the same process of Suzuki coupling and reductive amination reaction.
The synthetic route of molecules 4, 25 and 26 is depicted in Scheme 2. Aldehyde 11a was reduced with NaBH4 and treated with thionyl chloride, followed by condensation with 3-fluoro-4-nitrophenol, affording the intermediate 12a. After stirring for 24 h, the intermediate 12a was reduced by catalytic amount of Pd-C under hydrogen atmosphere to provide 13a. Condensation of the intermediate 13a with methyl chloroacetate in the presence of K2CO3, followed by basic hydrolysis, gave derivative 4. The isoxazole intermediate 15a was synthesized
by standard Suzuki coupling by treating (3-formylphenyl) boronic acid with commercially available isoxazole borate 14a. The reaction of methoxyamine hydrochloride and 16a in DMSO afforded the intermediate 17a, which was treated with 4-hydroxybenzaldehyde to give compound 18a. Aldehydes 15a or 18a were connected with 3b by reductive amination reaction using NaBH3CN, followed by basic hydrolysis to provide the target compounds 25 and 26.
2.2.SAR study
Besides in vitro agonistic activity, the metrics LE and LLE, as well as LogD7.4 were taken full advantage of in the direction of structural optimization. As shown in Table 1, the central oxygen atom was firstly replaced by an amine due to the significant lipophilicity lowering effect. Surprisingly, the obtained compound 3 turned out to be approximately 3-fold as potent as the parent compound 2 and resulted in markedly decreased lipophilicity (LogD7.4 value: 3.04 vs 3.61) and improved LLE value. Interestingly, switching position between oxygen atom and amine of compound 3 provided compound 4, which led to a significant drop of potency. One reasonable explanation is that the hydrogen-bonding interaction of carboxylic acid was interrupted by high dissociation energy generated from carboxylic acid and adjacent amino. Subsequently, we explored the importance of ortho-fluoro at C ring, a substituent introduced by Takeda to improve the metabolic stability and slightly increase potency.[11]
Indeed, removing the ortho-fluoro on compound 3 to provide compound 5 gave an 8-fold decrease in potency despite a slight advantage in terms of lipophilicity. Previous studies on biphenyl series revealed that various substituents were tolerated in the para-position of A ring.[13, 14] With this information in mind, the propoxy-substituent was incorporated at A ring to give compound 6, which resulted in a slight loss of agonistic activity despite its higher lipophilicity than the parent compound 3. Extension of the para-ethyoxyl with hydrophilic substituents were explored (7 and 8) and led to significantly decreased LogD7.4 values, but unfortunately also reduced potency and LE values due to the oversized groups.
Having identified the para-aminophenoxy acetic acid moiety (C ring), we next directed our efforts to the comprehensive SAR studies in present series (Table 2). The unsubstituted biphenyl (9) was tolerated in phenoxyacetic acid series and revealed a significant improvement on lipophilicity and LLE value. Introducing a chloro substituent in the B ring (10) did not led to further improvement in potency, which prompted an exploration in the ortho-position of A ring. For ortho-substituted analogs, the relationship between the potency and size of ortho substituent is roughly one reversed U-shape curve: the potency of compound 9 (2-H, 1.20 Å) < 11 (2-F, 1.47 Å) < 12 (2-Cl, 1.75 Å) < 13 (2-Me, 1.80 Å) > 14 (2-CF3, 2.20 Å) > 15 (i-Pr), indicating that the steric effect of ortho-substituent might be crucial for potency.[49]
Among them, compound 13 not only revealed the strongest agonistic activity, but also possessed a more appropriate lipophilicity (LogD7.4 = 2.41) and LLE value (5.9) compared with the corresponding chloro-substituted analog 12 and trifluoromethyl analog 14. However, the hydrophilic substitution of an ortho-nitrile (16) or ortho-methoxy (17) turned out a significant erosion of potency, implying that the lipophilicity of ortho-position substitution was also crucial for agonistic activity. The para-methyl analog 18 was found to be more potent than TAK-875 but only equipotent with parent compound 12, despite the higher lipophilicity of analog 18. Next, various substituents were explored in para-position of A ring based on the optimal compound 13. Introduction of ethyoxyl substituent (19) maintained the potency but also more lipophilic, which renders compound 19 has lower LE and LLE values than 13.
Similar to SAR of dimethyl series (6), incorporating the propoxy substituent (20) resulted in reduced potency back to the level of analog 6. Replacing hydrophobic substituents (19 and 20) with phenolic hydroxyl group was also explored (21) and led to a markedly reduced lipophilicity (LogD7.4 = 1.85) and increased LLE value while significantly decreasing activity. Introducing larger hydrophilic groups on the para-position of A ring, obtained compounds 22, 23 and 24 were found to be inferior to analogs with smaller hydrophobic substituents (e.g., 19), despite significant advantages in terms of lipophilicity. The 3,5-dimethylisoxazole scaffold has been reported to increase potency,[50] however, introduction of 3,5-dimethylisoxazole in compound 25 resulted in a lower potency than that of biphenyl scaffold (e.g., 3). Moreover, the oxime ether 26, the optimal scaffold in our previous report,[35] only showed moderate potency on FFA1.
2.3.Molecular modeling study
To better understand the high potency and SAR, we performed a docking study of compound 13 based on X-ray structure of FFA1 reported as co-complex with TAK-875.[51] As shown in Figure 4, derivative 13 take same binding mode of TAK-875 with slightly deflection of biphenyl scaffold, and the docking model can also explain SAR reasonably. Fluobenzene of compound 13 formed an edge-on interaction with Trp174, and the carboxylic acid of compound 13 was coordinated by crucial residues Arg183, Arg2258, Tyr91 and Tyr2240 forming an anchor point. This anchor point renders the carboxylic acid very sensitive to even small modifications (compounds 1 vs 2, and 3 vs 4). Moreover, the amine of compound 13 formed an additional ionic bond with carboxylic acid of Leu138, consistent with better potency of amino analogs compared to corresponding oxygen analogs such as compound 2 and TAK-875. In conclusion, compound 13 fitted well to the binding pocket by using it effectively.
2.4.OGTT in ICR mice
Compounds with high potency (3, 12, 13, 14, 18 and 19) were selected to assess the glucose-lowering effects in mice (Figure 5). As shown in Figure 5D, compounds 3, 12, 14 and 19 markedly decreased the levels of blood glucose compared to TAK-875 at the same dose. Gratifyingly, the glucose tolerance was significantly improved in compound 13-treated group, and the blood glucose AUC0−120min was significantly lower than that of TAK-875. Moreover, the plasma glucose curve of 13 tends to flatten after 1 h rather than decreased continually, indirectly suggesting its low risk of hypoglycemia.
2.5.The risk of hepatotoxicity
Though the exact mechanism of hepatotoxicity caused by TAK-875 is not clear at present, some experts speculated that TAK-875-induced hepatobiliary transporter inhibition might be one of the factors resulting in liver injury.[29, 30] So, the effects of selected compounds 13, 14, 19, and 23 on the hepatobiliary transporter were evaluated by measuring the accumulation of d8-Taurocholic acid (d8-TCA, Figure 6). Similar to TAK-875, incubation with these compounds (25 µM) significantly reduced the accumulation of d8-TCA (Figure 6A). Incubation with 25 µM TAK-875 decreased biliary excretion index (BEI) of d8-TCA to 17.7% (Figure 6B), suggesting that only 17.7% of d8-TCA taken up by hepatocytes was effluxed into bile. Compounds 13, 14, 19, and 23 also displayed moderate inhibitory effect on the efflux of bile acids, with BEI values of 12.3%, 29.6%, 19.3% and 14.3%, respectively. Fortunately, compound 14 (BEI: 29.6%) revealed significantly reduced risk of hepatotoxicity compared to TAK-875 (BEI: 17.7%). Moreover, the robustly glucose-lowering effects of compound 14 might provide a large enough therapeutic window to reduce the risk of hepatotoxicity. Therefore, compound 14 could be a safer candidate for the treatment of diabetes though compound 13 exhibited a stronger activity in vitro and in vivo.
2.6.Pharmacokinetic evaluation
Table 3 shows the oral PK parameters of compounds 13 and 14, two of the most typical analogs in this series. Compounds 13 and 14 presented good PK parameters, particularly in the low clearance, high maximum concentration, and long plasma half-life. Compound 14 even show that a 3‒fold higher plasma exposure (AUC0−24h = 3909.33 µg/mL·h) than TAK-875 (AUC0−24h = 1064.23 µg/mL·h). These results suggested that our strategy has successfully improved the PK profiles by replacing the β-carbon of phenylpropanoic acid moiety with oxygen atom. Moreover, compound 14 exhibited a higher maximum concentration (Cmax = 321.67 µg/mL) than that of compound 13 (Cmax = 186.53 µg/mL), which could be attributed to the metabolic stability of trifluoromethyl group at compound 14.
2.7.Insulinotropic effect and the risk of hypoglycemia
Based on these positive results, the insulinotropic effect of compound 14 was subsequently evaluated in Sprague−Dawley rats. As shown in Figure 7, oral administration of derivative 14 significantly suppressed blood glucose excursion in rats. Moreover, the plasma glucose AUC0−120min (Figure 7C) indicated that the glucose-lowering effect of derivative 14 (10 mg/kg) was slightly stronger than that of TAK-875. Besides, the level of insulin was almost unchanged from -60 min to 0 min at fasting glucose levels (Figure 7B), while plasma insulin level was sharply increased after glucose load at 0 min. Meanwhile, no significant increase in the level of insulin has been observed in compound 14-treated group without glucose load (Figure 7B). These results suggested that compound 14 enhanced the glucose-dependent insulin secretion, which may reduce the risk of hypoglycemia. As we expected, the fasting plasma glucose levels of compound 14-treated mice only slightly changed, and no statistical difference was observed between compound 14 and vehicle-treated groups (Figure 8). In contrast, the blood glucose levels of glibenclamide-treated mice turned out to be far below the fasting state. Therefore, this result suggested that compound 14 might be a safer insulin secretagogue compared to glibenclamide.
2.8.GLP-1 secretion
To explore whether the robustly glucose-lowering effect of compound 14 is dependent on full agonist mediated GLP-1 secretion, a further in vivo study was performed to detect the GLP-1 concentration after oral administration of compound 14 in mice. As shown in Figure 9A, there is no significant difference for GLP-1 concentration between vehicle, TAK-875, and compound 14 treated groups. Combined treatment with compound 14 and linagliptin, a well-known DPP-4 inhibitor[52], only slightly increased the secretion of GLP-1 compared to linagliptin alone, and no significant difference compared to the partial agonist TAK-875 combined with linagliptin (Figure 9B). All of these results indicated that compound 14 could not enhance GLP-1 secretion as a partial agonist similar to TAK-875.
2.9.Glucose-lowering effects in HF/STZ mice
To further assess the glucose-lowering effects in diabetic condition, compound 14 was investigated in male HF/STZ mice, a high-fat fed and streptozotocin-treated diabetic model with decreased β-cell capacity and insulin resistance, two typical features of type 2 diabetes.[53] As shown in Figure 10, the hyperglycemia was dramatically suppressed in 14-treated group, and the AUC0−120min was equivalent to that of TAK-875. This positive result indicated that the superior agonist 14 has sufficient effectiveness for the control of hyperglycemia state in type 2 diabetic mice.
2.10.The risk of acute toxicity
In order to further evaluate the safety of compound 14 in vivo, a one-week acute toxicity study at oral doses of 5 to 500 mg/kg was subsequently assessed in mice. At the end of toxicity study, several blood parameters related to hepatotoxicity were analyzed by automatic biochemical analyzer. In addition, the isolated liver from each treated group were evaluated by paraffin section to observe the histopathologic change. As shown in Figure 11, no significant increase in the levels of ALT and AST even at 500 mg/kg, preliminary indicating a low risk of compound 14-induced liver toxicity despite it has a moderate inhibitory effect on hepatobiliary transporter in vitro. Just as the hypotoxicity of compound 14 observed in blood parameters, the liver slices also confirmed this conclusion (Figure 11). There was no obvious histological change in the liver slices from compound 14-treated group (5 to 500 mg/kg) compared with that of control group. In this acute toxicity study, no distinct side-effects were observed. These results suggested compound 14, with relative low effective dose and high toxic dose, revealed excellent therapeutic window renders it a suitable molecule for further development.
3.Conclusion
In conclusion, starting from previously abandoned phenoxyacetic acid moiety, we identified a superior agonist 14 (CPU014, EC50 = 15.7 nM) based on SAR study and toxicity screening. Compound 14 revealed significantly improved efflux of bile acids compared to TAK-875, and exhibited low hepatotoxicity in an acute toxicity study even at 500 mg/kg. In addition, compound 14 presented significant glucose-lowering effects and insulinotropic effects in vivo, though no significant increased GLP-1 secretion in compound 14 treated group. Compound 14 has good PK profiles, particularly in the low clearance (CL = 7.47 mL/h/kg), high maximum concentration (Cmax = 321.67 µg/mL), and long plasma half-life (T1/2 = 4.84 h). All of these results subverted the previous bias for phenoxyacetic acid moiety and provided a new and safer agonist CPU014 for the treatment of T2DM.
4.Experimental section
4.1.General chemistry
All starting materials, reagents and solvents were obtained from commercial sources. Purifications of chromatography were performed by silica gel and detected by thin layer chromatography using UV light at 254 and 365 nm. Melting points were measured on RY-1 melting-point apparatus. NMR spectra were recorded on a Bruker ACF-300Q instrument (300 MHz for 1H NMR and 75 MHz for 13C NMR spectra), chemical shifts are expressed as values (ppm) relative to tetramethylsilane as internal standard, and coupling constants (J values) were given in hertz (Hz). LC/MS spectra were recorded on a Waters LC-MS system (ESI). Analytical HPLC was performed on a Shimadzu LC-20AT instrument equipped with Shimadzu C18 reversed-phase column (5 µm, 150 mm × 4.6 mm). All target compounds have >95% purity. Elemental analyses were performed by the Heraeus CHN-O-Rapid analyzer and were within 0.4% of the theoretical values. TAK-875 was synthesized by published procedures.[12]
Methyl 2-(4-aminophenoxy)acetate (3a). To a solution of 4-nitrophenol (4.0 g, 22.1 mmol) in 30 mL acetonitrile was added methyl chloroacetate (3.2 mL, 30.6 mmol) and K2CO3 (8.8 g, 63.7 mmol) at room temperature. The solution was heated to reflux for 6 h. The reaction mixture was filtered and evaporated to afford a colorless oil product.
To a solution of the obtained product
4.7 g in ethanol was added Pd/C (5%, 0.5 g) and the mixture was stirred for 18 h at room temperature in a hydrogen atmosphere. Insoluble matters were removed using Celite, and the filtrate was concentrated in vacuo to give the desired product (3.9 g, 74%) as a white solid. 1H NMR (300 MHz, DMSO-d6) δ: 6.76–6.72 (m, 4H), 5.26 (s, 2H), 4.73 (s, 2H), 3.75 (s, 3H). Methyl 2-(4-amino-2-fluorophenoxy)acetate (3b). The title compound was obtained from 2-fluoro-4-nitrophenol according to the method described for the synthesis of compound 3a in 67% yield as a white solid. 1H NMR (300 MHz, DMSO-d6) δ: 6.94-6.90 (m, 1H), 6.79-6.77 (m, 1H), 6.54, 6.50 (dd, J = 2.2, 12.5 Hz, 1H), 5.28 (s, 2H), 4.76(s, 2H), 3.75 (s, 3H).General procedure for 5a-b. The bromobenzene (1 equiv) and (3-formylphenyl)boronic acid (1 equiv) were dissolved in a mixture of 1 M sodium carbonate solution (15 mL), EtOH (5mL) and toluene (15 mL). After nitrogen substitution, Pd(PPh3)4 (0.05 equiv) was added. The reaction mixture was stirred at 80 °C under nitrogen atmosphere for 12 h.
The reaction mixture was cooled, and water (15 mL) was added. The mixture was diluted with ethyl acetate (15 mL), and the insoluble material was filtered off through Celite. The organic layer of the filtrate was washedwith brine, dried over anhydrous sodium sulfate, and concentrated in vacuo. The residue was purified by column chromatography using a mixture of petroleum ether/ethyl acetate (10:1, v/v) as eluent to afford the desired product as a solid.4’-Hydroxy-2’-methyl-3-biphenylcarbaldehyde (5a). Yield: 92%; white solid; 1H NMR (300 MHz, CDCl3) δ: 10.04 (s, 1H), 7.87 (d, J = 7.6 Hz, 1H), 7.66 (t, J = 1.7 Hz, 1H), 7.59-7.53(m, 2H), 7.42 (d, J = 7.4 Hz, 1H), 6.89 (d, J = 1.5 Hz, 1H), 6.76-6.72 (m, 1H), 4.67 (s, 1H), 1.98 (s, 3H).4’-Hydroxy-2’,6’-dimethyl-3-biphenylcarbaldehyde (5b). Yield: 83%; white solid; 1H NMR (300 MHz, CDCl3) δ: 10.05 (s, 1H), 7.96 (d, J = 7.8 Hz, 1H), 7.68 (t, J = 1.5 Hz, 1H),7.61-7.55 (m, 2H), 6.62 (s, 2H), 4.69 (s, 1H), 1.97 (s, 6H).General procedure for 6a-j. To a solution of 5a-b (1 equiv) and alkyl halide (1.2 equiv) in acetone was added K2CO3 (2 equiv) and a catalytic amount of KI at room temperature. The reaction mixture was heated to reflux with stirring overnight.
Then reaction mixture was cooled followed by filtration and the filtrate was concentrated under vacuum. The residue was purified by column chromatography using a mixture of petroleum ether/ethyl acetate (10:1, v/v) as eluent to afford a white solid.4′-ethoxy-2′,6′-dimethyl-[1,1′-biphenyl]-3-carbaldehyde (6a). Yield: 78%; 1H NMR (300 MHz, CDCl3) δ: 10.06 (s, 1H), 7.98 (d, J = 7.6 Hz, 1H), 7.69 (t, J = 1.7 Hz, 1H), 7.63-7.57 (m,2H), 6.63 (s, 2H), 4.02 (q, J = 6.7 Hz, 2H), 1.96 (s, 6H), 1.31 (t, J = 6.7 Hz, 3H).2′,6′-dimethyl-4′-propoxy-[1,1′-biphenyl]-3-carbaldehyde (6b). Yield: 73%; 1H NMR (300 MHz, CDCl3) δ: 10.06 (s, 1H), 7.98 (d, J = 7.6 Hz, 1H), 7.69 (t, J = 1.7 Hz, 1H), 7.63-7.57 (m,2H), 6.63 (s, 2H), 3.93 (t, J = 6.5 Hz, 2H), 1.92 (s, 6H), 1.73–1.69 (m, 2H), 0.99 (t, J = 7.3 Hz,3H).General procedure for target compounds 3 and 5-24. To a solution of 6a-j (1 equiv) and 3a-b (1 equiv) in MeOH (10 mL) and THF (20 mL) was added portionwise NaBH3CN (3 equiv) and the mixture was stirred for 12 h.
The reaction mixture was pouring into ice water (10 mL), and extracted with ethyl acetate (3 × 15 mL), washed with saturated brine (2 × 15 mL) prior to drying over anhydrous sodium sulfate. After filtration and concentrate, the residue was purified by column chromatography using a mixture of petroleum ether/ethyl acetate (5:1, v/v) as eluent to afford a white solid. To a solution of the obtained solid (1 equiv) in 2:3:1 THF/MeOH/H2O (18 mL) was added LiOH·H2O (1.5 equiv). After stirring for 4 h, the volatiles were removed under reduced pressure. The residue was acidified (PH: 5-6) with 1N hydrochloric acid solution, and extracted with ethyl acetate (4 × 25 mL), washed with saturated brine (2 × 10 mL) prior to drying.
After filtration and concentrate, the residue was purified by column chromatography using a mixture of petroleum ether/ethyl acetate (1:1, v/v) as eluent to afford a white solid.2-(4-(((4′-ethoxy-2′,6′-dimethyl-[1,1′-biphenyl]-3-yl)methyl)amino)-2-fluorophenoxy)ace tic acid (3). Yield: 67%; HPLC purity: 96.5%; m.p. 102-104 °C; 1H NMR (300 MHz, DMSO-d6)δ: 7.43 – 7.26 (m, 2H), 7.04 (s, 1H), 6.95 (d, J = 7.3 Hz, 1H), 6.81 (t, J = 9.3 Hz, 1H), 6.65 (s, 2H),6.42, 6.37 (dd, J = 14.1, 2.5 Hz, 1H), 6.29 (d, J = 8.9 Hz, 1H), 4.51 (s, 2H), 4.26 (s, 2H), 4.01 (q, J= 7.0 Hz, 2H), 1.87 (s, 6H), 1.31 (t, J = 7.0 Hz, 3H). 13C NMR (75 MHz, DMSO-d6) δ: 170.78, 157.55, 144.89, 140.61, 140.49, 136.93, 134.27, 128.87, 128.77, 128.07, 125.92, 117.63, 113.55,108.23, 101.25, 66.90, 63.16, 47.14, 21.17, 15.23. ESI-MS m/z: 422.2 [M-H]-. Anal. calcd. ForC25H26FNO4: C, 70.91; H, 6.19; N, 3.31; Found: C, 70.84; H, 6.08; N, 3.16.
4.2.FLIPR Assay
CHO cells stably expressing FFA1 were seeded into 96-well plates and incubated 16 h at 37 °C. After, the culture medium was removed and washed with 100 µL of Hank’s Balanced Salt Solution. Then, cells were incubated in loading buffer (containing 2.5 µg/mL Fluo 4-AM, 2.5 mmol/L probenecid and 0.1% fatty acid-free BSA) for 1 h at 37 °C. Various concentrations of compounds or γ-linolenic acid (Sigma) were added and the signals of calcium flux were monitored by FLIPR Tetra system (Molecular Devices). The agonistic activities were calculated as [(A−B)/(C−B)]×100 (increased concentration of calcium (A) in the compound group and (B) in vehicle group, and (C) in 10 µM γ-linolenic acid group). The EC50 value was calculated using GraphPad InStat version 5.00 (GraphPad software, San Diego, CA, USA).
4.3.Determination of LogD7.4
40 µL of 10 mM stock solution was added 1980 µL phosphate buffer solution (0.01 M, pH = 7.4) and 1980 µL 1-octanol (Sigma), obtaining 100 µM final concentration of the test compounds. The solution was shaken at 700 rpm for 24 h and left for 1 h to separate the phases. The 1-octanol phase was diluted x10 with 0.1% formic acid in methanol and MilliQ H2O (4 : 1) prior to analysis on HPLC, and the buffer phase was analyzed directly. The logD7.4 values were calculated by dividing the peak area (mAU*min) at 254 nm, and peak areas were corrected by two calibration points per compound per solvent.
4.4.Molecular Docking
Docking simulations were performed using MOE (version 2008.10, The Chemical Computing Group, Montreal, Canada) based on the crystal structure of FFA1 (PDB ID: 4PHU). Other ligands and water were removed, and then prepared with Protonate 3D. Subsequently, a Gaussian Contact surface was drawn and the active site was isolated. The ligand poses were filtered using Pharmacophore Query Editor. Compound 13 were docked into binding site with Pharmacophore method and then ranked with the London dG scoring function. For energy minimization, MOE Forcefield Refinement was used and ranked with the London dG scoring function.
4.5.Animals and Statistical Analysis
8 weeks old male SD rats, 10 weeks old ICR mice (male and female) and 10 weeks old male C57BL/6 mice were purchased from Comparative Medicine Centre of Yangzhou University (Jiangsu, China) and acclimatized for one week. Animal room was kept constant 12 h light/black cycle at 23 ± 2 °C with relative humidity 50 ± 10%. Animals were allowed ad libitum. All animal experimental protocols were approved by the ethical committee at China Pharmaceutical University and conducted according to the Laboratory Animal Management Regulations in China and adhered to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication NO. 85-23, revised 2011). Statistical analyses were carried out by GraphPad InStat version 5.00 (San Diego, CA, USA), and analyzed by Student’s t test or one-way ANOVA with Tukey’s multiple-comparison post hoc test.
4.6.Pharmacokinetic Study
SD rats were fasted for 12 h, weighted, and randomized into 2 groups (n = 4 per group). At 5, 15, 30, 45 min and 1, 2, 4, 6, 8, 12, 18, 24 h after oral dosing of 13 and 14 (10 mg/kg), blood samples were collected and centrifuged at 6000 rpm for 10 min to separate plasma. Plasma proteins were precipitated with methanol containing an internal standard, mixed and centrifuged at 14000 rpm for 14 min. The supernatant was diluted and 5 µL of supernatant was analyzed by Waters LC-MS/MS. Pharmacokinetic profiles were performed by DAS 2.1.1.
4.7.OGTT in mice
ICR mice were fasted for 12 h, weighted, and randomized into 8 groups (n = 6). Mice were administrated orally with vehicle, TAK-875 (20 mg/kg), or selected compounds (20 mg/kg) and subsequently dosed orally with 3 g/kg glucose solution after 30 min. Blood samples were collected before drug administration (-30 min), before glucose load (0 min), and at 15, 30, 60 and 120 min post-dose. The plasma glucose was measured by blood glucose test strips (SanNuo ChangSha, China).
4.8.OGTT in Rats.
SD rats were fasted for 12 h, weighted, and randomized into 4 groups (n = 6 per group). Rats were administrated orally with vehicle, TAK-875 (10 mg/kg), or compound 14 (10 mg/kg) and subsequently dosed orally with 3 g/kg glucose solution after 1 h. Another group (compound 14 + no glucose group) administrated orally with compound 14 (10 mg/kg) and subsequently dosed orally with vehicle after 60 min. Blood samples were collected immediately before drug administration (-60 min), before glucose load (0 min), and at 15, 30, 60 and 120 min post-dose. The blood glucose was measured by blood glucose test strips (SanNuo ChangSha, China). The level of insulin was measured by a rat insulin ELISA kit (Mercodia, Sweden).
4.9.GLP-1 secretion study
1) Glucose-stimulated GLP-1 secretion: compound 14 (20 mg/kg) and TAK-875 (20 mg/kg) were orally administered to overnight fasting male ICR mice (n = 8 per group). Compounds were given 0.5 h prior to 4 g/kg glucose load. After 5 min, blood samples were collected. 2) Basal GLP-1 secretion: male ICR mice were fasted for 6 h. Linagliptin (3 mg/kg) was orally given to all the mice except vehicle control group. One hour later, compound 14 (20 mg/kg) and TAK-875 (20 mg/kg) were orally given to the mice. After 4 h, blood samples were collected and placed in Eppendorf tubes containing DPP-4 inhibitor (DPP-4−010, Millipore) with a final concentration of 1% blood samples and 25 mg/mL EDTA to measure serum active GLP-1[7−36 amide] levels by using a mouse GLP-1 ELISA kit (RAYbiotech, America).
4.10.OGTT in HF/STZ Model
C57BL/6 mice were fed with high-fat diet (45% calories from fat, Mediscience Ltd., Yangzhou, China) ad libitum for 4 weeks and then injected intraperitoneally (i.p.) with STZ (80 mg/kg). The mice were fed with high-fat-diet for another 4 weeks, and diabetes was verified by glucose tolerance. The obtained type 2 diabetic HF/STZ mice were fasted for 12 h, weighted, and randomized into 3 groups (n = 6 per group). Mice were administrated orally with vehicle, TAK-875 (20 mg/kg), or compound 14 (20 mg/kg) and subsequently dosed orally with 2 g/kg glucose solution after 30 min. Blood samples were collected immediately before drug administration (-30 min), before glucose load (0 min), and at 15, 30, 60 and 120 min post-dose. The blood glucose was measured by blood glucose test strips (SanNuo ChangSha, China).
4.11.The Risk of Hypoglycemia
Normal SD rats were fasted for 12 h and randomized into 3 groups (n = 6 per group). Compound 14 (30 mg/kg), glibenclamide (10 mg/kg), or vehicle was orally administered, and blood was collected immediately before administration (0 min) and at 30, 60, 90, 120 and 180 min after administration and measure blood glucose as described above.
4.12.d8-TCA Uptake and Biliary Excretion Inhibition Assay[21]
Rat hepatocytes were isolated from SD rats and cultured according to reported method.[29] A sandwich configuration model was established as previously reported.[54] After incubation with selected compounds, TAK-875 or troglitazone in standard HBSS (Sigma) for 15 min, sandwich-cultured rat hepatocytes (SCRHs) were rinsed two times with warm standard HBSS or Calcium-free HBSS and pre-incubated at 37 °C for 15 min. After removing the buffer, the hepatocytes were incubated with standard HBSS containing test compounds in presence of 1 µM d8-TCA (TRC Canada) for another 15 min, respectively. After incubation, the solution was aspirated from cells, uptake was terminated by washing three times with ice-cold PBS. The d8-TCA concentration in hepatocytes was analyzed by LC-MS/MS in negative ion mode (m/z 522 to m/z 128). The biliary excretion index (BEI) was calculated as follows: BEI = [AHBSS–AHBSS (Ca2+-free)]/AHBSS × 100%, where AHBSS and AHBSS (Ca2+-free) represent the accumulated d8-TCA in the standard buffer treatment wells (hepatocytes+bile) and Ca2+-free buffer treatment wells (hepatocytes), respectively.
4.13.Acute Toxicity Study
ICR mice (n = 6 per group, males and females) were dosed daily with the vehicle and compound 14 (5, 50 and 500 mg/kg) by oral administration for 7 days. Any abnormal state was recorded. At the end of treatment, mice were fasted for 12 h, blood samples were collected and serum was separated. ALT and AST were measured using automatic biochemical analyzer (Beckman Coulter, AU5811, Tokyo, Japan). The liver tissue were isolated immediately after sacrifice and washed with ice-cold saline before fixed in 10% (v/v) formalin. The sections were embedded in paraffin after dehydrate. Four-micron sections were cut and stained with H&E for histopathological assessment.
Supporting information
Synthetic procedures and characterization of final compounds and intermediates not described in the manuscript.
Declaration of interest
The authors declare no competing financial interest.
Acknowledgements
This study was supported by grants from the National Natural Science Foundation of China (Grants 81673299 and 81273376).
Abbreviations
FFA1, free fatty acid receptor 1; T2DM, type 2 diabetes mellitus; SAR, structure-activity relationship; LE, ligand efficiency; LLE, ligand lipophilicity efficiency; PK profiles, pharmacokinetic profiles; DMSO, dimethylsulfoxide; FLIPR, fluorometric imaging plate reader; CHO, Chinese hamster ovary; PDB, Protein Data Bank; OGTT, oral glucose tolerance test; Cmax, maximum plasma concentration; CL: plasma clearance; T1/2, plasma half-life time; AUC, area under curve; HF/STZ mice, high-fat fed and low-dose streptozotocin-treated mice; SD rats, Sprague−Dawley rats; BEI, biliary excretion index; SCRHs, sandwich-cultured rat hepatocytes; Fasiglifam ALT, alanine aminotransferase; AST, aspartate transaminase; UREA, urea nitrogen; CREA, serum creatinine; CKMB, creatine kinase isoenzyme; GLB, globulin; ALB, albumin; TP, total protein; TG, triglyceride; TC, total cholesterol.