Discovery of a potent HMG-CoA reductase degrader that eliminates statin-induced reductase accumulation and lowers cholesterol

Abstract

Statins are inhibitors of HMG-CoA reductase, the rate-limiting enzyme of cholesterol biosynthesis, and have been clinically used to treat cardiovascular disease. However, a paradoxical increase of reductase protein following statin treatment may attenuate the effect and increase the side effects. Here we present a previously unexplored strategy to alleviate statin-induced reductase accumulation by inducing its degradation. Inspired by the observations that cholesterol intermediates trigger reductase degradation, we identify a potent degrader, namely Cmpd 81, through structure-activity relationship analysis of sterol analogs. Cmpd 81 stimulates ubiquitination and degradation of reductase in an Insig-dependent manner, thus dramatically reducing protein accumulation induced by various statins. Cmpd 81 can act alone or synergistically with statin to lower cholesterol and reduce atherosclerotic plaques in mice. Collectively, our work suggests that inducing reductase degradation by Cmpd 81 or similar chemicals alone or in combination with statin therapy can be a promising strategy for treating cardiovascular disease.

Fig. 1

Lovastatin causes a substantial accumulation of HMGCR protein. a–c Male C57BL/6J mice (n = 3 per group) were gavaged with vehicle or lovastatin (60 mg/kg/day) once daily for 7 days. Livers were harvested for immunoblotting and RT-qPCR. a Immunoblotting analysis of HMGCR protein from membrane fractions. Clathrin heavy chain (CHC) was a loading control. b Quantifications of HMGCR protein shown in a. The HMGCR protein level of mice gavaged with vehicle was defined as 1. c Quantifications of Hmgcr mRNA level by RT-qPCR. The Hmgcr mRNA level of vehicle-treated mice was defined as 1. Cyclophilin was used as the reference gene. d, e CHO-7 cells were treated with indicated concentrations of lovastatin for 16 h, then harvested for immunoblotting and RT-qPCR. d Immunoblotting analysis of HMGCR protein. e Dose–response curves of HMGCR protein and Hmgcr mRNA levels. HMGCR protein and Hmgcr mRNA levels of DMSO-treated cells were defined as 1. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was a reference gene. f, g SRD-13A cells were treated with lovastatin for 16 h. f HMGCR protein was analyzed by immunoblotting. g Hmgcr mRNA levels were measured by RT-qPCR. Data are from three independent experiments and presented as mean ± SD. ^***P < 0.001, unpaired two-tailed Student’s t-test. Source data are provided as a Source Data File. Uncropped immunoblots are shown in Supplementary Fig. 9

Fig. 2

A reporting system that measures HMGCR degradation. a Schematic of the HMGCR (TM1-8)-GFP fusion protein. TM, transmembrane. b, c CHO-7 cells stably expressing HMGCR (TM1-8)-GFP (CHG) were incubated with cholesterol or 24,25-DHL at indicated concentrations for 16 h and harvested for immunoblotting. b Endogenous HMGCR (IgG-A9) and overexpressed HMGCR (TM1-8)-GFP (polyclonal rabbit anti-GFP) protein were analyzed by immunoblotting. Asterisk represents a non-specific band. c Quantification of endogenous and overexpressed HMGCR protein levels in response to cholesterol or 24,25-DHL shown in b. The mean intensity of HMGCR protein bands in DMSO-treated cells was defined as 100. d, e CHG cells were incubated with varying concentrations of 24,25-DHL for 16 h. Cells were then fixed for immunofluorescence analysis. d Representative images showing a dose-dependent decrease of GFP signals following 24,25-DHL treatment. Scale bar, 20 μm. e Dose–response curves of HMGCR (TM1-8)-GFP fluorescent intensity to varying concentrations of cholesterol or 24,25-DHL. The GFP intensity of DMSO-treated cells was defined as 100. The mean EC₅₀ values of cholesterol and 24,25-DH were 30.4 μM and 1.5 μM, respectively. Data are from three independent experiments and presented as mean ± SD. Source data are provided as a Source Data File. Uncropped immunoblots are shown in Supplementary Fig. 9

Fig. 3

Identification of essential structural features of HMGCR degrader. a, b CHG cells were incubated with indicated compounds for 16 h. The GFP intensity of DMSO-treated cells was defined as 100. a Dose–response curves of HMGCR (TM1-8)-GFP fluorescent intensity to cholesterol analogs with different modifications at C-4 position. 4,4-Dimethyl cholesterol (7) was able to degrade HMGCR protein. b Dose–response curves of HMGCR (TM1-8)-GFP fluorescent intensity to 4,4-dimethyl cholesterol derived analogs with different moieties at C-7 position. 4,4-Dimethyl 7β-hydroxyl cholesterol (35) had improved activity of inducing HMGCR degradation. Data are from three independent experiments and presented as mean ± SD. Source data are provided as a Source Data File

Fig. 4

Summary of the EC₅₀ values of various compounds. Compounds bearing the 3β-hydroxyl group, 4,4-dimethyl groups and 7β-hydroxyl group on the steroid ring but differing in the side chain length or terminal moieties (R) were synthesized. CHG cells were incubated with indicated compounds for 16 h and the mean intensity of GFP from three independent experiments were measured. The EC₅₀ values were determined with dose–response curves by Prism software. Those with EC₅₀ values less than 0.75 μM are highlighted in shadow. Source data are provided as a Source Data File. Dose–response curves of indicated compounds are shown in Supplementary Fig. 3

Fig. 5

Cmpd 81 induces HMGCR ubiquitination and degradation. a, b CHO-7 cells were treated with Cmpd 81 at indicated concentrations for 16 h. 24,25-DHL was used as a positive control. a Immunoblotting analysis of HMGCR protein. b Quantification of Cmpd 81-induced HMGCR degradation shown in a. HMGCR protein of DMSO-treated cells was defined as 100. The EC₅₀ value of Cmpd 81 was 0.39 μM. c CHO-7 cells were treated with Cmpd 81 in the presence (+) or absence (−) of 10 μM MG-132 for 5 h. MG-132 prevented Cmpd 81-induced HMGCR degradation. d CHO-7 cells were treated with Cmpd 81 and 10 μM MG-132 for 3 h. Lysates were immunoprecipitated and pellets were probed for anti-ubiquitin (P4D1) and anti-HMGCR (A9). e CHO-7 cells were transfected with indicated plasmids and treated as in d. Lysates were immunoprecipitated with anti-Flag M2 beads, and pellets were probed for anti-ubiquitin (HA) and anti-HMGCR (Flag). f CHO-7 cells were transfected with indicated plasmids, then treated with Cmpd 81 for 5 h. Cmpd 81 promoted the degradation of WT but not ubiquitin sites mutated (K89R/K248R) HMGCR protein. g CHO-7 and SRD15 cells were treated with Cmpd 81 and the HMGCR protein was subjected for immunoblotting analysis. Data are from three independent experiments and presented as mean ± SD. Source data are provided as a Source Data File. Uncropped immunoblots are shown in Supplementary Fig. 10

Fig. 6

Cmpd 81 prevents lovastatin-induced HMGCR accumulation and ameliorates diet-induced hyperlipidemia and atherosclerosis. a Immunoblotting and quantification of HMGCR protein from CHO-7 cells. Data are from three independent experiments. b Male C57BL/6J mice (n = 3 per group) fed on chow diet were gavaged once daily with Cmpd 81 (60 mg/kg/day) with or without lovastatin (60 mg/kg/day) for 10 days. HMGCR proteins from liver membrane fractions were analyzed and quantified by immunoblotting. c–g Male C57BL/6J mice (n = 5 per group) fed with chow diet or medium fat medium cholesterol (MFMC) diet, were gavaged once daily with Cmpd 81 (60 mg/kg/day) with or without lovastatin (60 mg/kg/day) for 6 weeks. Blood and livers were harvested and examined for total cholesterol (TC) levels in the serum (c), triglyceride (TG) levels in the serum (d), TC levels in liver (e), TG levels in liver (f), and lipoprotein profiles of cholesterol levels in pooled serum using FPLC analysis (g). h–j Male Ldlr^−/− mice (n = 5 per group) fed on western diet (WD) were gavaged once daily with indicated Cmpd 81 (60 mg/kg/day) and lovastatin (60 mg/kg/day) for 20 weeks. h FPLC analysis of cholesterol contents in different lipoproteins. i Representative en face Sudan IV staining of aortas. Scale bar, 200 μm. j Quantifications of atherosclerotic lesions shown in i. Data are presented as mean ± SD. P values were determined by one-way ANOVA followed by Dunnett’s multiple comparisons test. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^#P < 0.05. Source data are provided as a Source Data File. Uncropped immunoblots are shown in Supplementary Fig. 10

Fig. 7

A working model of Cmpd 81 in lowering cholesterol synthesis. All clinically used statins dramatically induce compensatory increase of HMGCR, and the increased HMGCR proteins would blunt statins’ efficacy. However, Cmpd 81 remarkably degrades statins-induced HMGCR protein through the ubiquitin-proteasome pathway to lower the cholesterol synthesis, and to improve the efficacy of statins