Discovery of a Lead Triphenylethanamine Cholesterol Ester Transfer Protein (CETP) Inhibitor

Abstract

Lead optimization of the diphenylpyridylethanamine (DPPE) and triphenylethanamine (TPE) series of CETP inhibitors to improve their pharmaceutical profile is described. Polar groups at the N-terminus position in the DPPE series resulted in further improvement in potency and pharmaceutical properties concomitant with retaining the safety, efficacy, and pharmacokinetic (PK) profile. A structure–activity relationship observed in the DPPE series was extended to the corresponding analogs in the more potent TPE series, and further optimization resulted in the identification of 2-amino-N-((R)-1-(3-cyclopropoxy-4-fluorophenyl)-1-(3-fluoro-5-(1,1,2,2-tetrafluoroethoxy)phenyl)-2-phenylethyl)-4,4,4-trifluoro-3-hydroxy-3-(trifluoromethyl)butanamide (13). Compound 13 demonstrated no significant changes in either mean arterial blood pressure or heart rate in telemetry rats, had an excellent PK profile, and demonstrated robust efficacy in human CETP/apo-B-100 dual transgenic mice and in hamsters.

Atherosclerosisis a hallmark of coronary heart disease (CHD) characterized by the formation of cholesterol laden plaques that lead to the thickening of the arterial wall, decreased blood flow, and increased propensity for rupture resulting in acute clinical events such as myocardial infarction.^1,2 Uptake of modified low-density lipoprotein cholesterol (LDL-C) particles by macrophages in the arterial wall contributes to the formation of plaque, and there is a well-established correlation between circulating levels of LDL-C and the risk of CHD.³ Current lipid modifying therapies for the treatment of CHD target lowering circulating levels of LDL-C via inhibition of the 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase⁴ and Niemann-Pick C1-Like 1 (NPC1L1) inhibition.⁵ Reverse cholesterol transport (RCT) is the process of transferring cholesterol from peripheral cells (especially those in the arterial wall) to high density lipoprotein (HDL) and then LDL in the circulation, followed by uptake in the liver and biliary excretion. Increasing high-density lipoprotein (HDL) is known to facilitate RCT,⁶ and there is an inverse correlation between the levels of circulating HDL-C and CHD.³ These observations suggest that agents, which increase HDL-C, have the potential to be beneficial for CHD due, in part, to their effects on RCT, but potentially also due to other antiatherosclerotic effects that have been ascribed to HDL.⁷ Currently approved therapies (e.g., Niacin) demonstrate relatively modest effects on HDL-C levels and are of questionable benefit to patients with CHD.⁸

The plasma glycoprotein cholesterol ester transfer protein (CETP) is secreted by the liver and mediates the net transfer of cholesterol ester from HDL to LDL and VLDL in exchange for triglycerides.⁹ Inhibition of CETP by small molecule inhibitors such as anacetrapib,¹⁰ evacetrapib,¹¹ and antisense oligonucleotides¹² has resulted in reduced LDL-C and increased HDL-C levels in the clinic. However, it has not yet been demonstrated that increasing HDL-C levels through this mechanism, which should enhance RCT from the periphery, is antiatherosclerotic. The phase III study with evacetrapib was terminated due to lack of efficacy. However, the Phase III studies with anacetrapib were concluded successfully in 2017 whereupon clinical proof of concept was achieved and inhibition of CETP led to an increase in HDL-C and prevention of CHD.¹³

We have disclosed diphenylpyridylethanamine (DPPE) 2(14) and triphenylethanamines (TPE) 3(15) as potent CETP inhibitors (Figure 1). In an effort to further improve the pharmaceutical properties of the compounds in the DPPE and TPE series, while maintaining potency, excellent safety, and PK profile, we continued to explore polar substituents at the N-terminus position. In the course of optimizing the N-terminus hydroxy alkyl amine analogs in the DPPE series, we identified diol, 4(16) (Figure 1), with significantly improved solubility albeit with a very low level of CETP inhibitory activity. However, hydroxyamine 3 (Figure 1) was potent, so we sought to combine the diol moiety with a terminal trifluoromethyl group to increase the potency of the diol analogs. The first analog synthesized to test this hypothesis was the bis trifluoromethyl diol N-terminus amide, 7.

Figure 1.

DPPE and TPE series CETP inhibitors.

Synthesis of diol 7 was accomplished from the intermediate chiral amine 5a,¹⁷ which was converted to the vinylamide 6a and oxidized to give diol 7 directly. The corresponding monoalcohol 8 was prepared from amine 5a and the commercially available (4,4,4)-trifluoromethyl-3-hydroxy-3-trifluoromethyl-1-butanoic acid (Scheme 1).

Scheme 1. Synthesis of Compounds 6a–9b.

Reagents and conditions: (a) (4,4,4)-trifluoro-3-(trifluoromethyl)but-2-enoic acid, isobutylchloroformate, chloroform, 39% yield; (b) KMnO₄, acetone, K₂CO₃, 23% yield; (c) (4,4,4)-trifluoromethyl-3-hydroxy-3-trifluoromethyl-1-butanoic acid, IIDQ resin, MeCN, 14% yield; (d) (4,4,4)-trifluoro-3-(trifluoromethyl)but-2-enoic acid, IIDQ resin MeCN, 62% yield; (e) NaOCl, pyridine N-oxide, MeCN, chiral separation ChiralPak AD (IPA/heptane, 90/10) isomer 1, 6c elution at 46 min and isomer 2, 6d elution at 65 min; (f) NH₃/MeOH, 38% yield for 9a and 52% yield for 9b.

Simultaneous optimization of the B-ring (Scheme 1) resulted in identification of the 3-fluoro-5-tetrafluoroethoxy substituted aryl group in 5b. The corresponding intermediate vinylamide 6b was oxidized to epoxides 6c and 6d. The diastereomers were separated and the single diastereomers treated with ammonia to generate corresponding amino alcohols 9a and 9b (Scheme 1).

We have previously disclosed further optimization of the A-ring and the discovery of the more potent TPE series.¹⁵ Incorporation of the amino-alcohol N-terminus groups into the 4-fluorophenyl substituted TPE series yielded 10a and secondary and tertiary amine analogs 10b–10d (Scheme 2).

Scheme 2. Synthesis of Compounds 10a–13.

Reagents and conditions: See Supporting Information for reagents, conditions, and individual compound yields.

As the A-ring was further optimized for CETP potency,¹⁵ the 4-fluoro-3-isopropoxy group was identified, and this was subsequently used to prepare analogs 11a to 11e. Combining the 4-fluoro-3-isopropoxy group in the A-ring and the tetrafluoroethoxy group in the B-ring, compound 11a was prepared by the route shown in Scheme 2. Additional SAR at the N-terminus was explored and the diol amide (11b), diol (11c), and amino alcohol (11d) were prepared as described in Scheme 2. Exploring additional alkoxy groups on the A-ring resulted in the tertiary butyl ether (12) and the cyclopropyl ether (13).

All compounds were assayed for CETP inhibitory activity in the scintillation proximity and whole plasma assays (SPA and WPA, respectively, described in the Supporting Information), and data are included in Table 1.

Table 1. CETP Inhibitory Activity, Microsomal Stability, and hERG Flux Data for DPPE and TPE Analogs.

compound	CETP scintillation proximity IC50 (μM)b	CETP human whole plasma IC50 (μM)b	liver microsomal stabilityc (% remaining at 10 min) human, mouse	hERG flux IC50 (μM)c
7	0.035	1.4	ND	ND
8	0.12	3.8	92, 91	22
9aa (isomer 1)	0.053	5.9	100, 100	>80
9ba (isomer 2)	0.080	0.94	100, 100	>80
10aa (isomer 2)	0.011	0.43	100, 72	>80
10ba (isomer 2)	0.14	6.2	62, 87	13
10ca (isomer 2)	0.23	7.1	71, 54	63
10da (isomer 2)	3.3	30	58, 82	16
11a (isomer 1)	0.003	0.056	95, 90	44
11b (isomer 1)	0.001	0.084	89, 100	11
11c (isomer 1)	0.005	0.058	100, 100	12
11d (isomer 1)	0.026	0.20	81, 68	>80
11e	0.006	0.27	2, 16	ND
12 (isomer 1)	0.005	0.14	100, 100	ND
13 (isomer 1)	0.002	0.062	100, 100	>80

^aDiastereomers were derived from the corresponding chiral epoxide intermediates.

^bScintillation proximity assay and human whole plasma assay IC₅₀ values were measured in duplicate at six concentrations and the mean values were used to calculate IC₅₀ values.

^cProtocols for assays included in Table 1 are included in the Supporting Information.

We were gratified to see that diol amide 7 (SPA IC₅₀ 0.035 μM) was significantly more potent than 4 (SPA IC₅₀ 8.5 μM); however, further optimization was required to achieve the level of WPA potency observed with the amino alcohol 3.¹⁶ The corresponding monoalcohol 8 was found to be 3-fold less potent than diol 7; however, preparation of the amino alcohol 9b led to the first submicromolar compounds in the WPA in this series. Consistent with our earlier SAR, the TPE series were more potent than DPPE (e.g., 9b versus 11a). Substitution on the amine resulted in loss in CETP WPA potency (10b, 10c, and 10d). Overall, compound 11a had excellent CETP potency and was advanced to PK/PD studies and safety evaluation. Further modification to the N-terminus resulted in either decreased WPA CETP potency (11d and 11e), decreased liver microsomal stability (11e), or in vitro hERG liabilities (11b and 11c)¹⁸ compared to 11a. Thus, with the N-terminus amino alcohol amide group fixed, further A-ring ether optimization to increase the microsomal stability led to 12 and 13.

For all amino-alcohols, both diastereomers were synthesized and tested; without exception the most active antipodes were synthesized from the faster eluting chiral epoxide (e.g., 6c, Scheme 1). As described, we were able to determine the absolute stereochemistry at the quaternary center for intermediates 5a–f; however, we have been unable to obtain a single crystal structure or determine by NMR the absolute stereochemistry of the carbon bearing the amino group in compound 13.

Compounds 11a and 13 were 2-fold more potent than torcetrapib, 1 in the WPA, had high liver microsomal stability in both mouse and human, and, importantly, did not inhibit hERG below 44 μM. Based on the overall profile of these compounds, both 11a and 13 were advanced to PD evaluation in the hCETP/apoB-100 dual transgenic mouse¹⁹ (Figure 2). Compound 11a dosed at 1 mg/kg and compound 13 dosed at 0.5 mg/kg demonstrated comparable effects on suppression of CE transfer activity to torcetrapib dosed at 10 mg/kg in hCETP/apo-B-100 dual transgenic mice (Figure 2). Compound 13 demonstrated a greater effect in this model than compound 11a.

Figure 2.

Effects on 3H-CE transfer activity in hCETP/apoB-100 dual transgenic mice for 11a and 13 compared to torcetrapib, 1.

In addition, the exposure profile in rats was higher for compound 13 with bioavailability of 39% (Table 3). Compound 13 was selected as the lead compound in this series and advanced further to cynomologous monkey PK. We were gratified to observe the improved solubility of compound 13 in vehicles such as 10%Cremophor:10%EtOH:80%H₂O which were compatible with our longer term toxicology models and the excellent PK exposure in both rodents and monkeys consistent with an improved pharmaceutical profile for compounds in this series (Table 2). Compound 13 was also evaluated in the hamster model for CETP efficacy and demonstrated a 32% increase in HDL cholesterol when dosed at 10 mg/kg with a plasma exposure of 900 nM (Figure 3.)

Table 2. Pharmacokinetic Summary for Compounds 11a and 13 in Rat, Mouse, and Monkey^a.

	rat (compound 11a)	rat (compound 13)	mouse (compound 13)	monkey (compound 13)
vehicle	IVb: Vehicle A	IVb: Vehicle A	IVb: Vehicle A	IVb: Vehicle C
	POc: Vehicle A	POc: Vehicle B	POc: Vehicle B	POc: Vehicle B
dose(mg/kg)	1 IVb; 10 POc	1 IVb; 10 POc	5 IVb; 10 POc	1 IVb; 5 POc
Vssd(L/kg)	0.5	1.5	0.4	3.7
Cle(mL/min/kg)	5.4	2.9	1.9	0.7
t1/2f(h)	1.9	25	5	69
%Fg	12	39	18	27

^aVehicle A is 10% Cremophor/10% EtOH/80%H₂O; Vehicle B is 15% Miglyol 812/30% Capmul MCM/15% Triacetin/40% Cremorphor RH40; Vehicle C is 70%PEG-400/10%EtOH/20%H₂O.

^bIntravenous administration.

^cOral administration.

^dVolume of distribution.

^eClearance.

^fHalf-life.

^gBioavailability. Additional details available in Supporting Information.

Figure 3.

FPLC trace demonstrating effects of compound 13 (10 mg/kg) on plasma cholesterol profiles in hamsters (additional details in Supporting Information).

In light of the liability profile of torcetrapib with respect to blood pressure elevation and the induction of glucocorticoid and aldosterone synthase²⁰ (CYP11B1 and CYP11B2), we tested compound 13 in H295R cells for its ability to increase CYP11B1 and CYP11B2 mRNA levels. Torcetrapib, after 24 h of incubation, increased expression of the two genes potently (0.002 to 0.006 μM) and effectively (∼9-fold of control for CYP11B1 and 29 to 68-fold of control for CYP11B2). Angiotensin-II (0.1 μM) increased both CYP11B1 and CYP11B2 expression to ∼4-fold of control at 24 h. In contrast, compound 13 did not significantly affect the expression of either CYP11B1 (1.3-fold of control) or CYP11B2 (0.9 to 2.1-fold of control) demonstrating that compound 13, at a high concentration (10 μM) relative to concentrations that inhibit CETP in plasma (0.062 μM), lacks the aldosterone synthase liability of torcetrapib in H295R cells. To further evaluate the cardiac safety profile of compound 13, a telemetry study was performed in rats dosed at 10 mg/kg, IV (2.5 mg/kg infused over 5 min), followed by 7.5 mg/kg IV (infused over 55 min) (n = 7). There were no changes in mean, systolic or diastolic BP, heart rate, or locomotor activity at an observed C_max of 23.5 μM in the study.

In conclusion, we explored polar groups at the N-terminus in the DPPE series and subsequently further optimized the polar groups in the more potent TPE series to identify lead compound 13 with excellent WPA potency and an acceptable in vitro liability profile. PK evaluation of compound 13 in rodents and monkeys in vehicles compatible with longer term toxicology protocols demonstrated an increased vehicle solubility and exposure in both species consistent with an improved pharmaceutics profile. Compound 13 demonstrated robust HDL elevation in both dual transgenic mice and hamster models of efficacy, and significantly, compound 13 did not demonstrate effects on aldosterone synthase in H295R cells or on blood pressure or heart rate in telemetry rats. Given the clinical proof of concept achieved with anacetrapib for elevation of HDL-C through inhibition of CETP, this continues to be an exciting approach to the treatment of CHD.

Experimental Section

Experimental details for synthesis and analysis of compounds 5a, 5b, 7, 8, 9a, 9b, 10a–d, 11a–e, 12, and 13 can be found in the Supporting Information.

Glossary

ABBREVIATIONS

CETP

cholesterol ester transfer protein

CHD

coronary heart disease

HDL-C

high-density lipoprotein cholesterol

LDL-C

low-density lipoprotein cholesterol.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.9b00086.

• Experimental procedures and analytical data for compounds 5a to 13, and in vitro and in vivo assay conditions and protocols (PDF)

The authors declare no competing financial interest.