Piperazine Oxadiazole Inhibitors of Acetyl-CoA ... - ACS Publications

Dec 2, 2013 - Aaron Siegmund,. †. John G. Allen,. †. Hong Shu,. †. Christopher Fotsch, ... James Busby,. ‡. Minghan Wang,. ‡. James Meyer,. ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/jmc

Piperazine Oxadiazole Inhibitors of Acetyl-CoA Carboxylase Matthew P. Bourbeau,*,† Aaron Siegmund,† John G. Allen,† Hong Shu,† Christopher Fotsch,† Michael D. Bartberger,† Ki-Won Kim,‡ Renee Komorowski,‡ Melissa Graham,‡ James Busby,‡ Minghan Wang,‡ James Meyer,§ Yang Xu,§ Kevin Salyers,§ Mark Fielden,∥ Murielle M. Véniant,‡ and Wei Gu‡ †

Therapeutic Discovery, ‡Metabolic Diseases Research, §Pharmacokinetics/Drug Metabolism, and ∥Comparative Biology & Safety Sciences, Amgen Inc., 1 Amgen Center Dr, Thousand Oaks, California 91320, United States S Supporting Information *

ABSTRACT: Acetyl-CoA carboxylase (ACC) is a target of interest for the treatment of metabolic syndrome. Starting from a biphenyloxadiazole screening hit, a series of piperazine oxadiazole ACC inhibitors was developed. Initial pharmacokinetic liabilities of the piperazine oxadiazoles were overcome by blocking predicted sites of metabolism, resulting in compounds with suitable properties for further in vivo studies. Compound 26 was shown to inhibit malonyl-CoA production in an in vivo pharmacodynamic assay and was advanced to a long-term efficacy study. Prolonged dosing with compound 26 resulted in impaired glucose tolerance in diet-induced obese (DIO) C57BL6 mice, an unexpected finding.



INTRODUCTION Metabolic syndrome is a combination of related conditions that has been shown to affect >20% of the adult population in the United States. It is associated with an increased risk of type 2 diabetes and cardiovascular disease.1 Specific complications that result from metabolic syndrome include insulin resistance, high blood pressure, central obesity, decreased HDL cholesterol, and elevated triglycerides. The growing prevalence of metabolic syndrome in the population has not gone unnoticed, and a number of different approaches for the treatment of this condition are being investigated.2,3 Although there is not a clear understanding of all of the contributing factors to the pathogenesis of metabolic syndrome, abnormal fatty acid metabolism is considered to be a key element.4,5 Studies have demonstrated positive associations between insulin resistance and intracellular accumulations of triglycerides and fatty acid metabolites in insulin-responsive tissues such as muscle and liver. Fatty acid metabolites such as acyl-CoA, diacylglycerol, and ceramides have been shown to interfere with insulin signaling.6,7 Accumulation of intracellular fatty acid metabolites depends on fatty acid uptake and the rates of fatty acid synthesis and oxidation. As a result, drugs that block fatty acid uptake, inhibit fatty acid synthesis, and/or increase fatty acid oxidation might prove useful for the treatment of metabolic syndrome. The acetyl-CoA carboxylases (ACCs) are a class of enzymes that catalyze the conversion of acetyl-CoA to malonyl-CoA and, given their role in fatty acid synthesis and metabolism, have emerged as attractive targets for the treatment of metabolic syndrome.8 There are two characterized isoforms of ACC (known as ACC1 and ACC2) that have divergent functions.9 ACC1 is located in the cytosol and is primarily expressed in liver and fat tissue. ACC1 catalyzes the initiation of fatty acid © 2013 American Chemical Society

synthesis in a rate-limiting manner. ACC2 is primarily expressed in muscle tissue, where it is associated with the mitochondrial membrane. Malonyl-CoA produced by ACC2 serves as a negative regulator of carnityl palmitate transfer protein 1 (CPT1). CPT1 is the protein primarily responsible for transport of fatty acyls across the mitochondrial membrane for subsequent β-oxidation. Thus, malonyl-CoA production by ACC2 activity serves to directly inhibit mitochondrial fatty acid oxidation.10 Given the roles of the ACCs in both the synthesis and metabolism of fatty acids, inhibition of the ACCs would be expected to decrease fatty acid synthesis while simultaneously increasing fatty acid oxidation. ACC activity has been extensively studied in mouse knockout models. Attempts to generate whole animal ACC1−/− mice resulted in embryonic lethality.11 However, liver-specific ACC1 knockout mice (LACC1−/−) have been successfully generated by the groups of Wakil and Kusunoki.12,13 In Wakil’s case, a 75% reduction in liver malonyl-CoA levels was observed along with a 50% decrease in fatty acid synthesis in the liver. Conversely, liver samples from the LACC1−/−mice from Kusunoki’s lab showed no change in malonyl-CoA levels or fatty acid synthesis but did show an increase in ACC2 protein expression level and activity, suggesting that ACC2 was being upregulated to compensate for the ACC1 deletion. Several groups have developed ACC2 knockouts that are viable and have normal life spans. The most extensively characterized ACC2−/− mice were generated by Wakil and coworkers.14−16 These mice exhibit reduced malonyl-CoA levels and increased fatty acid oxidation in muscle and heart tissue relative to wild-type mice as well as increased total energy Received: October 15, 2013 Published: December 2, 2013 10132

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Scheme 1a

a

Reagents and conditions: (a) potassium (4-cyanophenyl)trifluoroborate, KOAc, 1,1-bis[(di-tert-butyl-p-methylaminophenyl]palladium(II) chloride, 3:1 MeCN/water, 90 °C, 29%; (b) NH2OH(aq), cat. AcOH, EtOH, 80 °C, propionic anhydride, pyridine, 80 °C, 32%.

Scheme 2a

Reagents and conditions: (a) nPrI, Cs2CO3, 20 °C, DMF, 99%; (b) chloro(2-dicyclohexylphosphino-2′,4′,6′-tri-isopropyl-1,1′-biphenyl)[2-(2aminoethyl)phenyl]palladium(II) methyl-tert-butylether adduct, NaOtBu, dioxane, 80 °C, 30−84%; (c) TFA; (d) CNBr, DIEA, CH2Cl2, 0 °C, 25− 88% for two steps; (e) NH2OH-HCl, DIEA, EtOH, 90 °C; (f) 2,2,5-trimethyl-1,3-dioxane-4,6-dione, pyridine, 90 °C, 41% for two steps or (EtCO)2O, 120 °C, 25−51% for two steps. a

(hACC2 IC50 = 841 nM, hACC1 IC50 = 4280 nM). Given the lethality of ACC1 embryonic knockout and the conflicting phenotypes observed between the two conditional ACC1−/− mouse strains, we were uncertain regarding the extent to which pharmacological ACC1 inhibition would be necessary and/or tolerated in vivo. For an initial proof-of-concept tool, we felt that a compound with dual ACC1/2 inhibitory activity had the greatest chance of producing the largest metabolic effect. However, there could potentially be an advantage to having an ACC2-selective compound if ACC1 inhibition proved to be unnecessary or a safety concern; therefore, we chose to evaluate ACC1/ACC2 selectivity to guide efforts toward compounds with greater ACC2 selectivity for subsequent studies.

expenditure. When placed on a high-fat/high-calorie diet, the ACC2−/− mice showed improved insulin sensitivity and lipid profiles as well as reduced weight gain and adiposity relative to wild-type mice. More recently, the groups of Cooney and Lowell have published phenotypic analysis of ACC2−/− mice generated by their respective groups.17,18 The ACC2−/− mice from Cooney’s lab show reduced malonyl-CoA levels in heart and muscle tissue as well as increased whole-body fatty acid oxidation but do not show improvements in any metabolic parameters relative to wild-type mice when fed a high-fat/highcalorie diet. In the case of the ACC2−/− mice described by Lowell’s group, only a modest reduction in malonyl-CoA levels was observed in heart tissue, and no additional metabolic changes were apparent. The cause of the observed differences in the metabolic phenotypes of the three published ACC2−/− mouse models is not obvious. Driven largely by the initial ACC2 knockout data published by Wakil, there have been significant efforts targeting ACC inhibition as a treatment for metabolic syndrome.19−25 Our own work in this area began with a high-throughput screening (HTS) campaign initially targeting human ACC2 (hACC2), with follow-up counter screening against human ACC1 (hACC1). Our screening campaign identified biphenyl oxadiazole 1 as a starting point for further investigation



CHEMISTRY Synthesis of Biphenyl Oxadiazole (2, Scheme 1). 1Bromo-2-methyl-4-proproxylbenzene underwent palladium-catalyzed cross coupling with potassium (4-cyanophenyl)trifluoroborate to afford nitrile 27. Nitrile 27 was sequentially treated with hydroxylamine and propionic anhydride to afford oxadiazole 2. Synthesis of Piperazine Oxadiazoles from Mono-BocProtected Piperazines (3−9, Scheme 2). Phenyl piperazines 31, 33, 35, 37, 39, 41, and 43 were prepared by

10133

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Scheme 3a

a

Reagents and conditions: (a) 29, chloro(2-dicyclohexylphosphino-2′,4′,6′-tri-isopropyl-1,1′-biphenyl)[2-(2-aminoethyl)phenyl]palladium(II) methyl-tert-butylether adduct, NaOtBu, dioxane, 120 °C, 55%; (b) CNBr, KOAc, MeOH, 72%; (c) NH2OH-HCl, DIEA, EtOH, 90 °C; (d) (EtCO)2O, 120 °C, 26% for two steps.

Scheme 4a

Reagents and conditions: (a) CNBr, DIEA, CH2Cl2, 0 °C, 99%; (b) NH2OH(aq), EtOH, 20 °C, 99%; (c) (EtCO)2O, pyridine, 120 °C, 47%; (d) HCl, dioxane, 20 °C, 95%; (e) PhI, N,N,-dimethylglycine, CuI, Cs2CO3, dioxane, 95 °C, 44% or RBr, Cs2CO3, DMF, 20 °C, 65−98%; (f) chloro(2dicyclohexylphosphino-2′,4′,6′-tri-isopropyl-1,1′-biphenyl)[2-(2-aminoethyl)phenyl]palladium(II) methyl-tert-butylether adduct, NaOtBu, dioxane, 100 °C, 8−29%. a

Scheme 5a

Reagents and conditions: (a) nPrI, Cs2CO3, 20 °C, DMF, 88−93%; (b) S-2-methylpiperazine, Pd2(dba)3, rac-BINAP, NaOtBu, toluene, 110 °C, 65−99%; (c) CNBr, DIEA, CH2Cl2, 0 °C, 99%; (d) NH2OH-HCl, DIEA, EtOH, 90 °C; (e) (EtCO)2O, 120 °C, 47−56%.

a

followed by condensation with cyanogen bromide and diisopropylethylamine afforded amino nitriles 32, 34, 36, 38, 40, 42, and 44. Sequential reaction of the amino nitrile with

palladium-catalyzed cross-coupling of the appropriate aryl halide with mono-Boc-protected piperazines using Buchwald’s X-Phos palladacycle.26 Subsequent cleavage of the Boc group 10134

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Scheme 6a

a Reagents and conditions: (a) EtI, Cs2CO3, DMF, 20 °C, 32−95%; (b) for 63 and 66: S-2-methylpiperazine, Pd2(dba)3, BINAP, NaOtBu, toluene, 120 °C, 41−51%; for 68: Pd2(dba)3, 2-dicyclohexylphosphino-2,6-diisopropoxy-1,1-biphenyl, NaOtBu, toluene, 100 °C, 55%; (c) CNBr, DIEA, CH2Cl2, 0 °C, 100%; (d) NH2OH, EtOH, 20 °C, quant; (e) (EtOCO)2O, pyridine, 80 °C, 30−31%; (f) (CH3CF2CO)2O, 70 °C, 21−23%.

Scheme 7a

a

Reagents and conditions: (a) NH2OH-HCl, DIEA, EtOH, 20 °C, 100%; (b) AcNHCR1R2COOH, HATU, DIEA, 80 °C, 53−67%.

11−14. Analogue 15 was prepared from commercially available 55. Synthesis of Chlorophenyl and Trifluoromethylphenyl Analogues (16−17, Scheme 5). 4-Bromo-3-chlorophenol and 4-bromo-3-(trifluoromethyl)phenol were alkylated with 1iodopropane to afford propyl ethers 56 and 59, respectively. Compounds 56 and 59 were then cross coupled with (S)-2methylpiperazine, affording piperazines 57 and 60. Reaction with cyanogen bromide and diisopropylethylamine afforded amino nitriles 58 and 61. Sequential reaction of the amino nitrile with hydroxylamine and propionic anhydride afforded oxadiazoles 16 and 17. Synthesis of Pyridine and Difluoro Analogues (18−21, Scheme 6). Bromopyridines 62 and 65 were cross-coupled with (S)-2-methylpiperazine mediated by Pd2(dba)3 and racBINAP to afford piperazines 63 and 66. Aryl bromide 52 was cross-coupled with (S)-2-methylpiperazine mediated by Pd2(dba)3 and RuPhos to afford piperazine 68. Treatment of 63, 66, and 68 with cyanogen bromide and diisopropylethyl-

hydroxylamine and 2,2,5-trimethyl-1,3-dioxane-4,6-dione afforded desired piperazine oxadiazoles 3−9. Synthesis of Trifluoromethylpiperazine Oxadiazole (10, Scheme 3). 2-(Trifluoromethyl)piperazine and 1-bromo2-methyl-4-propoxybenzene were coupled using Buchwald’s XPhos palladacycle to form trifluoromethylpiperazine 45. The secondary amine was then treated with cyanogen bromide and potassium acetate, and resulting amino nitrile 46 was reacted sequentially with hydroxylamine and propionic anhydride, affording trifluoromethylpiperazine oxadiazole 10. Synthesis of Phenoxyether Analogues (11−15, Scheme 4). (S)-tert-Butyl 3-methylpiperazine-1-carboxylate was treated with cyanogen bromide and diisopropylethylamine to afford amino nitrile 47. The amino nitrile was reacted sequentially with hydroxylamine and propionic anhydride followed by Boc deprotection with TFA to afford trifluoromethylpiperazine oxadiazole 50. Aryl ethers 51−54 were prepared from 4-bromo-3-methylphenol and were crosscoupled with trifluoromethylpiperazine oxadiazole 50 to afford 10135

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Scheme 8a

Reagents and conditions: (a) BocNHCR1R2COOH, HATU, DIEA, DMF, 80 °C; (b) TFA, CH2Cl2, 20 °C; (c) Ac2O, DIEA, CH2Cl2, 20 °C, 42− 44% over three steps. a

Figure 1. Initial development from HTS hit. Data are reported as the mean ± SD, n = 3.

recombinant cell line in which the ACC2 coding sequence was expressed under the control of an inducible tetracycline promoter utilizing the CHOK1-TREx platform.27 Compound 1 was found to inhibit the production of malonyl-CoA in the ACC2 cellular assay and was, in fact, more potent in this cellbased assay than in the biochemical assay (hACC2 IC50 = 841 nM, cell IC50 = 359 nM). The ACC2 protein expressed in the CHOK1 cells had a 24 amino acid deletion at the C-terminus of the protein. These residues are thought to be responsible for the mitochondrial association of ACC2. The cellular ACC2 protein also possessed a constitutively activating S221A mutation. It is possible that these changes resulted in the unexpected cell shift. Nevertheless, we felt that this cellular assay would serve as a useful tool for assessing the ability of advanced compounds to inhibit directly malonyl-CoA production in cells and thus we used the cellular assay to profile key compounds as the SAR of the oxadiazole series developed. Initial efforts to improve potency by modifying the propyl ether of 1 were unsuccessful. Therefore, a more conservative approach was undertaken, and analogues with one F or Me at each unsubstituted position of the two phenyl rings were prepared. Of these eight analogues, the majority lost activity in the hACC2 and hACC1 enzymatic assays (data not shown). Compound 2, however, was more promising, showing improved activity in the ACC2 and ACC1 enzymatic assays (IC50 = 126 and 1460 nM, respectively) as well as the ACC2 cellular assay (IC50 = 103 nM) (Figure 1). With this result in hand, we attempted to replace the central phenyl ring of 2 with a saturated ring in an effort to break the aromaticity of the lead molecules and potentially to improve overall physical properties.28 These efforts resulted in the synthesis of piperazine oxadiazole 3, which exhibited ACC inhibition comparable to 2. Encouraged by the results obtained with 3, we next examined more substituted piperazine analogues (Table 1). Methylation of the R3 position resulted in a modest boost in ACC2 potency, with S enantiomer 5 in particular showing inhibition of ACC2 in the enzymatic assay of 71 nM. Removing the R1 methyl group, as shown in 6, resulted in a loss of potency, indicating

amine afforded amino nitriles 64, 67, and 69. Sequential treatment of 64 and 67 with hydroxyl amine and propionic anhydride afforded pyridylpiperazine oxadiazoles 19 and 20, respectively. Sequential treatment of 69 and 64 with hydroxyl amine and ethyl-2,2-difluoropropionate afforded difluorooxadiazoles 18 and 21, respectively. Synthesis of Oxadiazole Acetamide Analogues (22− 24, Scheme 7). Amino nitrile 69 was treated with hydroxylamine to form carboximidamide 70. HATU-mediated coupling of 70 with N-acetylglycine afforded oxadiazole acetamide 22. HATU-mediated coupling of 70 with (S)- and (R)-2acetamidopropanoic acid afforded oxadiazole acetamides 23 and 24, respectively. Synthesis of Dimethyl and Cyclobutyl Oxadiazole Acetamide Analogues (25−26, Scheme 8). Carboximidamide 70 was subjected to HATU-mediated coupling with Bocalpha-methylalanine. Boc cleavage with TFA and treatment of the resulting free amine with acetic anhydride afforded dimethyl oxadiazole acetamide 25. Separately, carboximidamide 70 was subjected to HATU-mediated coupling with 1-(tertbutoxycarbonylamino)cyclobutanecarboxylic acid. Boc cleavage with TFA and treatment of the resulting free amine with acetic anhydride afforded cyclobutyl oxadiazole acetamide 26.



DISCUSSION HTS of the Amgen sample collection identified biphenyl oxadiazole 1 as a starting point for further investigation (Figure 1). In our primary enzymatic assay, which measured the conversion of acetyl-CoA to malonyl-CoA by LC/MS/MS, 1 was found to inhibit hACC2 and hACC1 activity with IC50’s of 841 and 4280 nM, respectively, and was found to be competitive with acetyl-CoA. After determining that we could conveniently quantify malonyl-CoA levels by LC/MS/MS, we sought to develop a cellular assay that also measured malonylCoA reduction in a similar manner. Attempts to measure ACC inhibition in native cell lines was hampered by the relatively low baseline levels of malonyl-CoA (FAO, L6, and H9C2 cells were examined). To circumvent this issue, we constructed a 10136

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Table 1. SAR Investigations on Piperazinea

entry

R1

R2

R3

hACC2 IC50 (nM)

hACC1 IC50 (nM)

3 4 5 6 7 8 9 10

Me Me Me H H Me Me Me

H H H H Me H H H

H Me (R) Me (S) Me (S) H Et (S) iPr (S) CF3

391 ± 135 178 ± 62 71 ± 9 2580 ± 226 >10 000 231 ± 70 335 ± 30 166 ± 28

4010 ± 1350 1028 ± 102 1910 ± 322 >10 000 >10 000 3930 ± 240 >10 000 492 ± 36

a

ACC2 enzymatic ratio. In addition to alkyl ethers, phenyl ether 14 was also examined, but it showed a significant reduction in potency relative to the propyl and ethyl ethers. Given the potency of 12 in our ACC2 enzymatic assay, we profiled 12 in our ACC2 cellular assay, where it exhibited an IC50 of 39 nM. (Figure 2) Compound 12 was then evaluated for in vitro microsomal stability where it had rapid human and rat microsomal turnover (HLM and RLM). In an in vivo pharmacokinetic (PK) study in rats, 12 also exhibited a high clearance of approximately 78% of liver blood flow, indicating good in vitro prediction of in vivo clearance. Additionally, all attempts to utilize 12 in in vivo studies measuring changes in tissue levels of malonyl-CoA were met with no success. Therefore, our next area of focus was to reduce PK liabilities of 12. To this end, 12 was analyzed using the Metasite metabolism prediction program.29 Three potential sites of metabolism were identified: the ethyl ether, the aryl methyl group, and the oxadiazole ethyl group. A number of different analogues were made in an attempt to block the predicted sites of metabolism (Figure 3). Efforts to block the metabolism of the aryl ether by replacement of the alkyl group with a CF3 led to an analogue with significantly reduced potency (15). The aryl methyl group could be replaced with either a chloro (16) or CF3 (17) group with a modest loss of ACC2 enzymatic potency. In the case of CF3 analogue 17, there was some improvement seen in the HLM and RLM values. A similar improvement in the HLM and RLM values was observed by difluorination of the benzylic position of the oxadiazole ethyl group (18) as well as the replacement of the phenyl group with pyridine isomers (19 and 20). In all three cases, however, there was a similar loss of ACC2 enzymatic potency. Combination of the pyridine A ring with the gem difluoroethyl oxadiazole did not result in additive improvement in the HLM and RLM values (21). Although the previously discussed examples gave us some encouragement that we could improve upon the poor microsomal stability of 12, we were much more excited by the improvements that we saw by substituting the ethyl oxadiazole group with acetamides (Figure 4). Compound 22 retained reasonable potency in the ACC2 enzyme assay but exhibited almost no oxidative metabolism in either human or rat liver microsomes. Methylation at the position alpha to the acetamide (23 and 24) led to compounds with a slight increase in microsomal turnover. However, these compounds exhibited significantly better ACC enzyme activity and, in the case of 23, potency in the ACC2 cellular assay was in the single-digit nanomolar range. Dimethyl and cyclobutyl analogues 25 and 26

Data are reported as the mean ± SD, n ≥ 3.

that substitution on both the phenyl and piperazine rings improved potency. Shifting the methyl group to the R2 position of the piperazine ring (7) resulted in a loss of enzyme activity (IC50 > 10 000 nM). Attempts to place groups larger than methyl at the R3 position (8, 9, and 10) resulted in reduced potency relative to 5, indicating that methyl was the optimal substituent at the R3 position for this set of compounds. Holding the new methyl piperazine central ring constant, we examined variations to the aryl ether portion of the molecule. We found that we were able to vary the length of the alkyl chain and retain some potency (Table 2). In particular, methyl, ethyl, Table 2. SAR Investigations on Alkyl Ethera

entry

R1

hACC2 IC50 (nM)

hACC1 IC50 (nM)

hACC1/hACC2

5 11 12 13 14

Pr Me Et iPr Ph

71 ± 9.5 486 ± 37 37 ± 3.2 273 ± 31 789 ± 72

1910 ± 322 >10 000 4570 ± 258 >10 000 >10 000

27× >20× 123× >36× >13×

a

Data are reported as the mean ± SD, n ≥ 3.

and isopropyl ethers were examined (11, 12, and 13). Ethyl ether 12 not only retained potency on the ACC2 enzymatic assay but also showed a significant improvement in the ACC1/

Figure 2. Analysis of 12. In vivo PK analysis was performed in male Sprague−Dawley rats, n = 3. 10137

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Figure 3. Efforts to address metabolic stability.

Figure 4. Oxadiazole amide analogues.

stability, these compounds were further evaluated for their potential as in vivo proof-of-concept tools. In particular, 26 was extensively profiled (Figure 5). As had been observed with previous compounds, 26 was more potent in the cell-based assay than in the in vitro assay (cell IC50 = 8.2 nM). Mouse ACC1 and ACC2 (mACC1 and mACC2) enzyme activity was similar to that observed in the corresponding human ACC assays. Compound 26 was readily formulated at concentrations needed for in vivo dosing and possessed low clearance and high exposure in both mouse and rat PK studies. There was no inhibition of human cytochrome P450 3A4 or 2D6 observed with 26, and the plasma protein binding of 26 was determined to be 97.5−98.6% across species. Compound 26 was profiled in an in vitro pharmacology panel consisting of 116 receptors and

showed improved microsomal stability relative to the monomethyl analogues with retention of the excellent cellular potency. It was surprising that these amide analogues showed such a profound improvement in metabolic stability over 12, especially considering that difluoro analogue 18, which was also designed to address metabolism of the oxadiazole ethyl group, did not show such a significant improvement. Our working hypothesis for this observation centers on the amide’s effect on lowering cLogP values. The amide analogues have cLogP values that are ∼1.5−2.5 log units lower than 12, and it may be that the increased polarity reduces recognition by P450 enzymes and thus reduces metabolism.30 Given the favorable ACC1/2 enzyme potency of the oxadiazole amides coupled with the improved microsomal 10138

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Figure 5. PK properties of 26. In vivo PK analysis was performed in male Sprague−Dawley rats, n = 3, and male C57BL6 mice, n = 3, as indicated.

was conducted in diet-induced obese (DIO) C57BL6 mice with 10 mice per dose group. The mice were administered daily (po) for 31 days with vehicle or 15, 35, and 75 mg/kg/day of 26. Body-weight changes were monitored daily, and food intake was monitored weekly. After 21 days of dosing, blood samples were taken to measure glucose, insulin, triglycerides, cholesterol, free fatty acids, alanine transaminase (ALT), and aspartate transaminase (AST) levels. At 28 days of dosing, an oral glucose tolerance test (GTT) was performed. Three days later, the mice were euthanized, and tissues were collected to measure terminal PD and other parameters. The 75 mg/kg dose group exhibited severe weight loss on the tenth day of administration and therefore had to be dropped from the study because of animal welfare concerns (Figure 7). The 10 and the 35 mg/kg doses were tolerated for the duration of the study, and the 35 mg/kg dose showed a small but significant decrease in weight gain relative to the control group, with no concomitant change in food intake. Surprisingly, however, there was a significant increase in plasma glucose levels for the animals administered 26 at both 10 and 35 mg/kg relative to control. Additionally, when the animals were subjected to an oral GTT, the mice in the 35 mg/kg dose group exhibited increased glucose AUC relative to the control group. It is not clear whether these observations resulted directly from ACC inhibition. Terminal PD conducted 25 h after the last dose of 26 showed significant inhibition of malonyl-CoA in both liver and heart tissue. Of particular note, the 35 mg/kg dose showed malonyl-CoA reductions of 42% in liver and 67% in heart, with a corresponding unbound drug concentration of 618 nM. This data is very similar to the 8 h time point from the time-course PD experiment (44% malonyl-CoA reduction in liver and 68% malonyl-CoA reduction in heart, with an unbound drug concentration of 882 nM), suggesting that the PK/PD relationship of drug concentration to malonyl-CoA lowering remained relatively constant through the course of the study. In addition to the previously described findings, there was a slight but significant (∼10%) elevation of cholesterol levels in the 35 mg/kg group of the study relative to control. There were no significant differences in insulin, triglyceride, free fatty acid, ALT, or AST levels in either dose group relative to the control group. Our initial foray into ACC as a target for the treatment of metabolic syndrome/diabetes was largely inspired by the pioneering ACC2 knockout work of Wakil’s group.14−16 Subsequent studies by Cooney and Lowell called into question the metabolic benefits of ACC2 knockout.17,18 We developed a

enzymes (Cerep, Inc., Redmond, WA) and an in vitro panel of 100 kinases (Ambit Biosciences, San Diego, CA). The only additional target on which 26 demonstrated activity was the benzodiazepine receptor (IC50 = 710 nM), which was thought to be of little impact on the in vivo end points that we were going to measure. Compound 26 was also nonmutagenic in a bacterial mutation (Ames) assay. Thus, we believed 26 possessed the required potency, selectivity, and PK properties to be a useful in vivo proof-of-concept tool. We next advanced 26 to a mouse pharmacodynamic (PD) study to examine the effect on malonyl-CoA levels in liver (ACC1-rich tissue) and heart (ACC2-rich tissue) (Figure 6). At

Figure 6. (a) Dose-response PD study. C57BL6 mice, n = 3 per dose group. Malonyl-CoA measured by LC/MS/MS 1 h post dose. * = p value < 0.05. Free drug concentration (plasma concentration × unbound fraction in nanomolar): 25 mg/kg: 708 ± 76, 50 mg/kg: 988 ± 89, and 100 mg/kg: 1482 ± 141. (b) Time-course PD study, C57BL6 mice, n = 3 per dose group. Malonyl-CoA measured by LC/ MS/MS 1 h post dose. * = p value < 0.05. Free drug concentration (plasma concentration × unbound fraction in nanomolar): 1 h: 1165 ± 68, 4 h: 1142 ± 133, 8 h: 882 ± 43, and 12 h: 498 ± 59.

25, 50, and 100 mg/kg po doses, 26 showed a significant reduction in malonyl-CoA levels in both the liver and heart 1 h postdose. There appeared to be an effective maximal effect at ∼85% inhibition, consistent with observations made by scientists at Pfizer conducting similar experiments with panACC inhibitors.25 In a 12 h time course study, a 25 mg/kg dose of 26 showed >50% reduction in malonyl-CoA levels between 4 and 8 h in liver and 8 to 12 h in heart tissue. At the 8 h time point, the unbound plasma drug concentration was 882 nM, approximately 10-fold higher than the mACC2 enzyme IC50 and 2-fold higher than the mACC1 enzyme IC50. Having confirmed that dosing 26 resulted in significant reduction of malonyl-CoA levels in vivo, we next chose to study 26 in a long-term in vivo study to evaluate the effect of ACC inhibition on a broad set of metabolic parameters. The study 10139

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Figure 7. Effect of long-term ACC inhibition with 26 on DIO C57BL6 mice. * = p value < 0.05, *** = p value < 0.001. (a) Change in body weight over 27 days of dosing. (b) Change in fed plasma glucose levels after 21 days of dosing. (c) Glucose tolerance test after 28 days of dosing. (d) Percent reduction of malonyl-CoA relative to control after 31 days of dosing. Free drug concentration levels (plasma concentration × unbound fraction in nanomolar): 15 mg/kg: 95 ± 27 and 35 mg/kg: 618 ± 106.

Notes

series of oxadiazole ACC inhibitors with varying degrees of ACC1/ACC2 selectivity, the most potent members of which were from an oxadiazole amide subseries. Oxadiazole amide 26 possessed good in vivo PK properties and was shown to lower malonyl-CoA levels in C57BL6 mice in both liver and heart tissue, suggesting in vivo inhibition of ACC1 and ACC2. A 31 day efficacy experiment was conducted by dosing 26 in DIO mice. Terminal PD studies showed a significant reduction in malonyl-CoA levels in liver and heart tissue, and a decrease in body-weight gain relative to control animals was observed. However, the glucose tolerance of the mice dosed with 26 actually worsened relative to control mice, and serum glucose levels also increased. There were no beneficial changes to any fatty acid related measurements observed in the animals, which, given the reportedly direct effect of ACCs on such metabolic parameters, was surprising. In a publication that appeared subsequent to the completion of our experiments, scientists at Sanofi-Aventis observed a similar lack of desired efficacy with their ACC inhibitors in long-term studies conducted in both DIO mice as well as in Zucker diabetic fatty (ZDF) rats.31 We feel that these results, coupled with the ACC2 knockout studies of Cooney and Lowell, suggest that ACC inhibition may not, in fact, be a useful strategy for the treatment of metabolic disease/ diabetes, at least as represented in rodent models of disease.



The authors declare the following competing financial interest(s): The corresponding author is an employee and stock holder of Amgen, Inc.

■ ■

DEDICATION This paper is dedicated to the memory of Kevin Salyers, our colleague and friend. ABBREVIATIONS USED Ac, acetyl; ACC, acetyl-CoA carboxylase; AcOH, acetic acid; ALT, alanine transaminase; AST, aspartate transaminase; AUC, area under the curve; BINAP, 2,2′-bis(diphenylphosphino)1,1′-binaphthyl; Boc, tert-butyloxycarbonyl; CHOK1, Chinese hamster ovary K1 cells; CL, clearance; CL int, intrinsic clearance; CMax, maximum plasma concentration; CNBr, cyanogen bromide; CPT1, carnityl palmitate transfer protein 1; Cs2CO3, cesium carbonate; CuI, copper iodide; dba, 1,5diphenylpenta-1,4-dien-3-one; DIEA, diisopropylethylamine; DIO, diet-induced obese; DMF, dimethylformamide; (EtCO)2O, propionic anhydride; EtI, ethyl iodide; EtOH, ethyl alcohol; %F, bioavailability; GTT, glucose tolerance test; hACC, human acetyl-CoA carboxylase; HATU, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate; HCl, hydrochloric acid; HDL, high-density lipoprotein; HLM, human liver microsomes; HTS, high-throughput screen; iv, intravenous dosing; po, oral dosing; KOAc, potassium acetate; LC/MS/MS, liquid chromatography/mass spec/mass spec; mACC, mouse acetyl-CoA carboxylase; MeCN, acetonitrile; MeOH, methyl alcohol; NaOtBu, sodium tert-butoxide; NH2OH, hydroxylamine; nPrI, 1-iodopropane; PD, pharmacodynamic; PK, pharmacokinetic; RLM, rat liver microsomes; RuPhos, 2dicyclohexyl(2′,6′-diisopropoxybiphenyl-2-yl)phosphine; SAR, structure−activity relationship; THF, tetrahydrofuran; TFA, trifluoroacetic acid; t1/2, half life; Vdss, volume of distribution;

ASSOCIATED CONTENT

S Supporting Information *

Experimental procedures and synthetic procedures for all final compounds and noncommercially available intermediates. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: 805-447-9430; E-mail: [email protected]. 10140

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141

Journal of Medicinal Chemistry

Article

Weight, Fat Mass, or Food Intake. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 7598−7603. (19) Corbett, J. W. Review of Recent Acetyl-CoA Carboxylase Inhibitor Patents: Mid-2007−2008. Expert Opin. Ther. Pat. 2009, 19, 943−956. (20) Bourbeau, M. P.; Allen, J. G.; Gu, W. Recent Advances in AcetylCoA Carboxylase Inhibitors. Annu. Rep. Med. Chem. 2010, 45, 95−108. (21) Bourbeau, M. P. ACC Inhibitors in Development. In New Therapeutic Strategies for Type 2 Diabetes: Small Molecule Approaches; Jones, R. M., Ed.; Royal Society of Chemistry, Cambridge, 2012; pp 464−500. (22) Yamashita, T.; Kamata, M.; Endo, S.; Yamamoto, M.; Kakegawa, K.; Watanabe, H.; Miwa, K.; Yamano, T.; Funata, M.; Sakamoto, J-i.; Tani, A.; Mol, C. D.; Zou, H.; Dougan, D. R.; Sang, B.; Snell, G.; Fukatsu, K. Design, Synthesis, and Structure-Activity Relationships of Spirolactones Bearing 2-Ureidobenzothiophene as Acetyl-CoA Carboxylases Inhibitors. Bioorg. Med. Chem. Lett. 2011, 21, 6314−6318. (23) Kamata, M.; Yamashita, T.; Kina, A.; Funata, M.; Mizukami, A.; Sasaki, M.; Tani, A.; Funami, M.; Amano, N.; Fukatsu, K. Design, Synthesis, and Structure-Activity Relationships of Novel SpiroPyridines as Acetyl-CoA Carboxylase Inhibitors. Bioorg. Med. Chem. Lett. 2012, 22, 3643−3647. (24) Kamata, M.; Yamashita, T.; Kina, A.; Tawada, M.; Mizukami, A.; Sasaki, M.; Tani, A.; Nakano, Y.; Watanabe, Y.; Furuyama, N.; Funami, M.; Amano, N.; Fukatsu, K. Symmetrical Approach of SpiroPyrazolidinediones as Acetyl-CoA Carboxylase Inhibitors. Bioorg. Med. Chem. Lett. 2012, 22, 4769−4772. (25) Freeman-Cook, K. D.; Amor, P.; Bader, S.; Buzon, L. M.; Coffey, S. B.; Corbett, J. W.; Dirico, K. J.; Doran, S. D.; Elliott, R. L.; Esler, W.; Guzman-Perez, A.; Henegar, K. E.; Houser, J. A.; Jones, C. S.; Limberakis, C.; Loomis, K.; McPherson, K.; Murdande, S.; Nelson, K. L.; Phillion, D.; Pierce, B. S.; Song, W.; Sugarman, E.; Tapley, S.; Tu, M.; Zhao, Z. Maximizing Lipophilic Efficiency: The Use of FreeWilson Analysis in the Design of Inhibitors of Acetyl-CoA Carboxylase. J. Med. Chem. 2012, 55, 935−942. (26) Biscoe, M. R.; Fors, B. P.; Buchwald, S. L. A New Class of Easily Activated Palladium Precatalysts for Facile C−N Cross-Coupling Reactions and the Low Temperature Oxidative Addition of Aryl Chlorides. J. Am. Chem. Soc. 2008, 130, 6686−6687. (27) Yao, F.; Svensjö, T.; Winkler, T.; Lu, M.; Eriksson, C.; Eriksson, E. Tetracycline Repressor, tetR, Rather than the tetR-Mammalian Cell Transcription Factor Fusion Derivatives, Regulates Inducible Gene Expression in Mammalian Cells. Hum. Gene Ther. 1998, 9, 1939− 1950. (28) Lovering, F.; Bikker, J.; Humblet, C. Escape from Flatland: Increasing Saturation as an Approach to Improving Clinical Success. J. Med. Chem. 2009, 52, 6752−6756. (29) Cruciani, G.; Carosati, E.; De Boeck, B.; Ethirajulu, K.; Mackie, C.; Howe, T.; Vianello, R. MetaSite: Understanding Metabolism in Human Cytochromes from the Perspective of the Chemist. J. Med. Chem. 2005, 48, 6970−6979. (30) Lewis, D. F. V.; Jacobs, M. N.; Dickins, M. Compound Lipophilicity for Substrate Binding to Human P450s in Drug Metabolism. Drug Discovery Today 2004, 9, 530−537. (31) Glien, M.; Haschke, G.; Pfenninger, A.; Zoller, G.; Keil, S.; Müller, M.; Herling, A. W.; Schmoll, D. Stimulation of Fat Oxidation, but No Sustained Reduction of Hepatic Lipids by Prolonged Pharmacological Inhibition of Acetyl CoA Carboxylase. Horm. Metab. Res. 2011, 43, 601−606.

X-Phos, 2-dicyclohexylphosphino-2′,4′,6′-triisopropyl-1,1′-biphenyl; ZDF rats, Zucker diabetic fatty rats



REFERENCES

(1) Ford, E. S.; Giles, W. H.; Dietz, W. H. Prevalence of the Metabolic Syndrome Among US Adults. JAMA 2002, 287, 356−359. (2) Moller, D. E.; Kaufman, K. D. Metabolic Syndrome: A Clinical and Molecular Perspective. Annu. Rev. Med. 2005, 56, 45−62. (3) Grundy, S. M. Drug Therapy of the Metabolic Syndrome: Minimizing the Emerging Crisis in Polypharmacy. Nat. Rev. Drug Discovery 2006, 5, 295−309. (4) Bjorntorp, P. Fatty Acids, Hyperinsulinema, and Insulin Resistance: Which Comes First? Curr. Opin. Lipidol. 1994, 5, 166−174. (5) McGarry, J. D. What if Minkowski Had Been Ageusic? An Alternative Angle on Diabetes. Science 1992, 258, 766−770. (6) Savage, D. B.; Petersen, K. F.; Shulman, G. I. Disordered Lipid Metabolism and the Pathogenesis of Insulin Resistance. Physiol. Rev. 2007, 87, 507−520. (7) Shulman, G. I. Cellular Mechanisms of Insulin Resistance. J. Clin. Invest. 2000, 106, 171−176. (8) Ploakis, S. E.; Guchhait, R. B.; Zwergel, E. E.; Lane, M. D.; Cooper, T. G. Acetyl Coenzyme A Carboxylase System of Escherichia coli: Studies on the Mechanisms of the Biotin Carboxylase- and Carboxyltransferase-Catalyzed Reactions. J. Biol. Chem. 1974, 249, 6657−6667. (9) McGarry, J. D.; Foster, D. W. Regulation of Hepatic Fatty Acid Oxidation and Ketone Body Production. Annu. Rev. Biochem. 1980, 49, 395−420. (10) Rasmussen, B. B.; Holmback, U. C.; Volpi, E.; Morio-Liondore, B.; Paddon-Jones, D.; Wolfe, R. R. Malonyl Coenzyme A and the Regulation of Functional Carnitine Palmitoyltransferase-1 Activity and Fat Oxidation in Human Skeletal Muscle. J. Clin. Invest. 2002, 110, 1687−1693. (11) Abu-Elheiga, L.; Matzuk, M. M.; Kordari, P.; Oh, W.; Shaikenov, T.; Gu, Z.; Wakil, S. J. Mutant Mice Lacking Acetyl-CoA Carboxylase 1 Are Embryonically Lethal. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 12011−12016. (12) Mao, J.; DeMayo, F. J.; Li, H.; Abu-Elheiga, L.; Gu, Z.; Shaikenov, T.; Kordari, P.; Chirala, S. S.; Heird, W. C.; Wakil, S. J. Liver-Specific Deletion of Acetyl-CoA Carboxylase 1 Reduces Hepatic Triglyceride Accumulation without Affecting Glucose Homeostasis. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8552−8557. (13) Harada, N.; Oda, Z.; Hara, Y.; Fujinama, K.; Okawa, M.; Ohbuchi, K.; Yonemoto, M.; Ikeda, Y.; Ohwaki, K.; Aragane, K.; Tamai, Y.; Kununoki, J. Hepatic De Novo Lipogenesis Is Present in Liver-Specific ACC1-Deficient Mice. Mol. Cell. Biol. 2007, 27, 1881− 1888. (14) Abu-Elheiga, L.; Matzuk, M. M.; Abo-Hashema, K. A.; Wakil, S. J. Continuous Fatty Acid Oxidation and Reduced Fat Storage in Mice Lacking Acetyl-CoA Carboxylase 2. Science 2001, 291, 2613−2616. (15) Abu-Elheiga, L.; Oh, W.; Kordari, P.; Wakil, S. J. Acetyl-CoA Carboxylase 2 Mutant Mice are Protected Against Obesity and Diabetes Induced by High-Fat/High-Carbohydrate Diets. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 10207−10212. (16) Choi, C. S.; Savage, D. B.; Abu-Elheiga, L.; Zhu, Z. X.; Kim, S.; Kulkarni, A.; Distefano, A.; Hwang, Y. J.; Reznick, R. M.; Codella, R.; Zhang, D.; Cline, G. W.; Wakil, S. J.; Shulman, G. I. Continuous Fat Oxidation in Acetyl-CoA Carboxylase 2 Knockout Mice Increases Total Energy Expenditure, Reduces Fat Mass, and Improves Insulin Sensitivity. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 16480−16485. (17) Hoehn, K. L.; Turner, N.; Swarbrick, M. M.; Wilks, D.; Preston, E.; Phua, Y.; Joshi, H.; Furler, S. M.; Larance, M.; Hegarty, B. D.; Leslie, S. J.; Pickford, R.; Hoy, A. J.; Kraegen, E. W.; James, D. E.; Cooney, G. J. Acute or Chronic Upregulation of Mitochondrial Fatty Acid Oxidation Has No Net Effect on Whole-Body Energy Expenditure or Adiposity. Cell Metab. 2010, 11, 70−76. (18) Olson, D. P.; Pulinilkunnil, T.; Cline, G. W.; Shulman, G. I.; Lowell, B. B. Gene Knockout of Acc2 Has Little Effect on Body



NOTE ADDED AFTER ASAP PUBLICATION After this paper was published ASAP on December 11, 2013, corrections were made to Scheme 7 and the footnotes of Schemes 4, 5, and 6. The corrected version was reposted December 17, 2013.

10141

dx.doi.org/10.1021/jm401601s | J. Med. Chem. 2013, 56, 10132−10141