Liver-specific ATP-citrate lyase inhibition by bempedoic acid decreases LDL-C and attenuates atherosclerosis

Abstract

Despite widespread use of statins to reduce low-density lipoprotein cholesterol (LDL-C) and associated atherosclerotic cardiovascular risk, many patients do not achieve sufficient LDL-C lowering due to muscle-related side effects, indicating novel treatment strategies are required. Bempedoic acid (ETC-1002) is a small molecule intended to lower LDL-C in hypercholesterolemic patients, and has been previously shown to modulate both ATP-citrate lyase (ACL) and AMP-activated protein kinase (AMPK) activity in rodents. However, its mechanism for LDL-C lowering, efficacy in models of atherosclerosis and relevance in humans are unknown. Here we show that ETC-1002 is a prodrug that requires activation by very long-chain acyl-CoA synthetase-1 (ACSVL1) to modulate both targets, and that inhibition of ACL leads to LDL receptor upregulation, decreased LDL-C and attenuation of atherosclerosis, independently of AMPK. Furthermore, we demonstrate that the absence of ACSVL1 in skeletal muscle provides a mechanistic basis for ETC-1002 to potentially avoid the myotoxicity associated with statin therapy.

Figure 1. ETC-1002-CoA inhibits ACL and mediates β1-selective AMPK activation.

Recombinant human ACL was incubated in the presence of the indicated concentrations of ETC-1002-CoA (a–c) or ETC-1002 (d), and (a) coenzyme A (CoA), (b) citrate or (c) ATP. Conversion of [¹⁴C]-citrate to [¹⁴C]-acetyl-CoA was measured in counts per minute (CPM). Recombinant human (e) AMPKα1β1γ1 and (f) AMPKα1β2γ1 complexes were incubated in the presence of the indicated concentrations of ETC-1002, ETC-1002-CoA, AMP or A-769622. (g) Structure of ETC-1002 and the biochemical reaction that generates ETC-1002-CoA acyl-CoA synthetase (ACS). ETC-1002-CoA/CoASH enzyme ACL kinetic analyses are single measures, and representative results from n=3 independent experiments shown; AMPK activity was determined by time-resolved fluorescence resonance energy transfer expressed as mean±s.e.m. of triplicate measures. (a) Ki calculated by Michaelis–Menten kinetic analysis.

Figure 2. Differentiation of ETC-1002 from the myotoxic effects of statins.

[¹⁴C]-ETC-1002 (10 μM) conversion to [¹⁴C]-ETC-1002-CoA in human liver microsomes was determined in the presence of 30 μM unlabelled (a) monocarboxylic and (b) dicarboxylic saturated competitive fatty acid substrates. RH7777 cells were treated with negative control (Control) or Slc27a2 (ACSVL1) siRNA; n=6, for 48 h, and (c) ACSVL1 expression determined by western blot, and [¹⁴C]-ETC-1002-CoA synthesis determined; n=3. (d) Both control and Slc27a2 siRNA-treated RH7777 cells were treated with vehicle, ETC-1002 (30 μM) or atorvastatin (0.5 μM), and de novo cholesterol synthesis measured over 4 h; n=6 transfections. (e) Relative ACSVL1 expression in microsomes prepared from human liver, kidney and skeletal muscle (SkM). (f) ETC-1002-CoA synthetase activity in microsomes prepared from human liver and skeletal muscle; palmitate (10 μM) used as a positive control for skeletal muscle microsome viability. (g) ACSVL1 expression was measured in primary human hepatocytes and primary human skeletal muscle (SkM) cells by western blotting and expressed relative to β-actin. (h) Primary human myotubes were treated with vehicle (Veh.), simvastatin 10 μM (Simva), atorvastatin 10 μM (Atorva) or ETC-1002 100 μM in the presence of [¹⁴C]-glucose for 12 h, and incorporation into non-saponifiable lipids determined; or treated with 0.1–100 μM simvastatin, atorvastatin or ETC-1002 for 48 h and (i) cytotoxicity was measure by GF-AFC/bis-AAF-R110 cleavage, and (j) Caspase 3,7 activity (DEVD cleavage) determined. Data for myotube cytotoxicity assays are expressed as mean±s.e.m., n=2 performed in triplicate. Data for liver (n=50 donors pooled) and kidney (n=50 donors pooled) microsome preparations are representative of two independent experiments performed in duplicate, and expressed as mean±s.d. Data for human skeletal muscle microsomes (n=4 pooled donors) are the mean±s.e.m. of two independent experiments performed in triplicate. Multiple comparisons were made using an one-way ANOVA, Bonferroni's multiple comparisons test; *P<0.05.

Figure 3. ETC-1002 mediates effects on lipid metabolism independently of AMPK β1.

Apoe^−/− and Apoe^−/−/Ampkβ1^−/− (DKO) mice were fed a HFHC diet for 12 weeks with or without ETC-1002 targeted to achieve a 30 mg kg⁻¹ per day dose and (a) total and phosphorylated (T172) AMPKα and total and phosphorylated (S79) ACC measured in livers from Apoe^−/− (n=4) and DKO (n=3). (b) Slc27a2 (ACSVL1) expression in liver and tibialis anterior from untreated Apoe^−/− mice (n=6). (c) Representative fast performance liquid chromatography tracings from chow- and HFHC diet-fed±ETC-1002 Apoe^−/− and DKO mice. Plasma (d) very-low-density lipoprotein (VLDL), (e) LDL, (f) HDL (n=4 for chow-fed mice and 6 for remaining treatment groups) and (g) total cholesterol (n=10), liver (h) cholesterol and (i) triglycerides (n=4 for chow-fed and 6 for remaining treatment groups) determined at the end of study. (j) RER was measured over 24 h following 10 weeks on diet and mean RER during (k) dark (19:00–07:00) and (l) light (07:00–19:00) cycles calculated. (m) Total lipid synthesis in primary hepatocytes isolated from WT and AMPK β1 KO mice treated for 4 h with the indicated concentration of ETC-1002, A-769622 or vehicle (CON) in the presence of ³H-acetate (n=3–8 independent experiments performed in duplicate). (n) Srebf2, Ldlr, Pcsk9, Hmgr, Acly, Srebf1c and Slc27a2 gene expression (n=4 chow-fed and 5 for remaining treatment groups) determined at the end of study. (o) Representative images for immunohistochemistry staining of LDLR in frozen liver sections from ETC-1002-treated Apoe^−/− and DKO mice (20 × ); scale bar, 100 μM. Data are expressed as mean±s.e.m. Multiple comparisons were made using an *one-way (within Apoe^−/− treatment groups) or #two-way ANOVA (between Apoe^−/− and DKO treatment groups); Bonferroni's multiple comparisons test; *and ^#P<0.05.

Figure 4. ETC-1002 treatment reduces the progression of atherosclerosis.

Apoe^−/− and Apoe^−/−/Ampkβ1^−/− (DKO) mice were fed a HFHC diet for 12 weeks with or without ETC-1002 targeted to achieve a 30 mg kg⁻¹ per day dose. (a) Sections of the aortic sinus from control and ETC-1002-treated mice were stained with haematoxylin and eosin and atherosclerotic lesion size determined; n=8 for all treatment groups except ETC-1002-treated Apoe^−/− (n=7). (b) Total cholesterol (TChol) was measured in whole aorta (n=4 for Apoe^−/− and 6 for DKO HFHC-fed mice) and (c) plasma SAA measured at the end of study (n=10). Data are expressed as mean±s.e.m. Multiple comparisons were made using an *one-way (within Apoe^−/− treatment groups) or #two-way ANOVA (between Apoe^−/− and DKO treatment groups); Bonferroni's multiple comparisons test; *and ^#P<0.05. NS, not significant.

Figure 5. ACL suppression increases LDLR.

(a) Primary human hepatocytes were treated with vehicle or ETC-1002 (100 μM) for 18 h in the presence of [³H]-H₂O or pretreated with the indicated concentrations of ETC-1002 for 1 h, following a 3 h [¹⁴C]-acetate pulse. Counts incorporated into the non-saponifiable (sterols), and saponifiable (fatty acids) lipid fractions were determined and expressed as % vehicle. (b) Primary human hepatocytes were treated with vehicle, atorvastatin (0.5 μM) or ETC-1002 (100 μM) for 36 h and total intracellular cholesterol concentrations were determined in cell lysates. (c) Media was assayed for apoB by enzyme-linked immunosorbent assay. (d) PHHs were treated with ETC-1002 for the indicated time points and the apoB concentrations were measured in the media. (e) PHH were pre-labelled overnight with [¹⁴C]-oleate and the effects of ETC-1002 treatment on oleate-derived counts in the media determined at the indicated time points. PHH hepatocytes were treated with ETC-1002 and atorvastatin, and (f) mRNA levels for HMGR, SREBF2, PCSK9 and LDLR were determined by reverse transcription–quantitative PCR or (g) LDL receptor activity determined; A-769662 (10 μM). (h) RH7777 cells were treated with the indicated concentrations of ETC-1002 or atorvastatin (0.5 μM), and [¹⁴C]-lactate incorporation into cholesterol (Chol), cholesteryl esters (CE) and triglycerides determined. (i) LDLR-mediated DiI–LDL uptake in response to the indicated concentrations of ETC-1002±atorvastatin (0.5 μM). RH7777 cells were subjected to negative control (NC) or Acly siRNA-mediated gene silencing (n=3) and analysed for (j) ACL LDLR protein levels, or (k) LDLR-mediated DiI–LDL uptake. Data for PHH are expressed as mean±s.e.m. of n=2 donors performed in triplicate (a–c,f,g; except A-769662 treatment was 1 donor) or representative of two independent experiments showing similar results (d,e). (h) IC₅₀ and (i) EC₅₀ determinations were made using nonlinear curve fit model. Multiple comparisons were made using (a,j,k) unpaired Student's t-test or (b,c,f,g) an one-way ANOVA, Bonferroni's multiple comparisons test; *P<0.05.

Figure 6. The mechanism of action of ETC-1002.

In liver, ETC-1002 (bempedoic acid) is activated to ETC-1002-CoA by ACSVL1, and subsequently inhibits ACL. Similar to inhibition of HMG-CoA reductase (HMGR) by statins, inhibition of liver ACL by ETC-1002-CoA results in the suppression of cholesterol synthesis and compensatory LDLR upregulation and LDL particle clearance from the blood. Skeletal muscle does not express ACSVL1 and is unable to convert ETC-1002 to its active form. Therefore, ETC-1002 does not suppress the synthesis of cholesterol or the associated biological intermediates that are required to maintain normal muscle cell function, or promote the associated toxicity. Green and red arrows indicate the effects of ETC-1002 and statins, respectively.