High-throughput screening identifies small molecule inhibitors of thioesterase superfamily member 1: Implications for the management of non-alcoholic fatty liver disease

Abstract

Objective: Thioesterase superfamily member 1 (Them1) is a long chain acyl-CoA thioesterase comprising two N-terminal HotDog fold enzymatic domains linked to a C-terminal lipid-sensing steroidogenic acute regulatory transfer-related (START) domain, which allosterically modulates enzymatic activity. Them1 is highly expressed in thermogenic adipose tissue, where it functions to suppress energy expenditure by limiting rates of fatty acid oxidation, and is induced markedly in liver in response to high fat feeding, where it suppresses fatty acid oxidation and promotes glucose production. Them1^-/- mice are protected against non-alcoholic fatty liver disease (NAFLD), suggesting Them1 as a therapeutic target.

Methods: A high-throughput small molecule screen was performed to identify promising inhibitors targeting the fatty acyl-CoA thioesterase activity of purified recombinant Them1.Counter screening was used to determine specificity for Them1 relative to other acyl-CoA thioesterase isoforms. Inhibitor binding and enzyme inhibition were quantified by biophysical and biochemical approaches, respectively. Following structure-based optimization, lead compounds were tested in cell culture.

Results: Two lead allosteric inhibitors were identified that selectively inhibited Them1 by binding the START domain. In mouse brown adipocytes, these inhibitors promoted fatty acid oxidation, as evidenced by increased oxygen consumption rates. In mouse hepatocytes, they promoted fatty acid oxidation, but also reduced glucose production.

Conclusion: Them1 inhibitors could prove attractive for the pharmacologic management of NAFLD.

Keywords: Acyl-CoA thioesterase; Brown adipose tissue; Energy homeostasis; Insulin resistance; Lipid metabolism; Obesity associated metabolic disorders.

Figure 1

Small molecule inhibitors targeting Them1 activity identified by high-throughput screening. Reactions were performed in 384-well microplates with recombinant His-tagged human Them1 (Them1; 125 nM), myristoyl (C₁₄)-CoA (C₁₄-CoA; 25 μM) and compounds (12.5 μM), and incubated for 60 min at 22 °C. (A) Z′ factors were calculated from reactions containing no enzyme [(+) control] or no C₁₄-CoA [(−) control; 25 μM]. Plate ID indicates the unique number assigned for each plate in the compound library. (B) Scatter plot for the normalized percent inhibition (NPI) of 360,705 library compounds (yellow dots [(+) control)]; green dots [(−) control] and gray dots [compounds (12.5 μM)]). (C) Representative 384-well microplate from the high-throughput screen. Compounds were considered as potential inhibitors based on a NPI ≥ 30%. Negative values are attributable to the absorbance of a minority of compounds. The Z′ factor is shown. (D) Representative compounds that exhibited inhibition of Them1 activity (open circles) or exceeded the maximum IC₅₀ screening criteria of 20 μM (closed circles). Data represent mean (s.e.m.) of triplicate determinations. Where not visible, standard error bars are contained within the symbol sizes.

Figure 2

Small molecule inhibitors selectively targeting Them1 activity. (A) Reactions were performed with recombinant His-tagged human Them1 (Them1; 125 nM), myristoyl-CoA (C₁₄-CoA; 25 μM) and compounds, and incubated for 60 min at 22 °C. IC₅₀ values of compounds targeting Them1 are listed. Data represent mean (s.e.m.) of triplicate determinations. Compounds selected for scaffold optimization are highlighted in pink. (B) Flow chart describing identification of Them1 START domain-dependent inhibitors from the 360,705 library compounds. Counter screens targeted ∗Acot isoforms followed by ^# truncated Them1 containing only the 2 thioesterase (Thio) domains but lacking the START domain (Them1-ΔSTART).

Figure 3

Structure activity relationship (SAR) expansion of commercially available small molecule inhibitors with START-domain dependent Them1 activity. Structural features of U and W series small molecules that selectively inhibit Them1 activity through its START domain. Acot activities were determined with recombinant His-tagged human Them1 (Them1; 125 nM), myristoyl-CoA (C₁₄-CoA; 25 μM) and compounds, and incubated for 60 min at 22 °C. IC₅₀ values are listed. Binding interactions were quantified as K_d values by microscale thermophoresis using recombinant Them1 labeled with Monolith RED-Tris-NTA (100 nM). Data represent mean (s.e.m.) of triplicate determinations. Where not visible, standard error bars are contained within the symbol sizes.

Figure 4

Them1 inhibitors increase fatty acid oxidation rates in cultured primary brown adipocytes. Oxygen consumption rate (OCR) values in primary brown adipocytes cultured from wild type and A-Them1Tg mice following stimulation with norepinephrine (NE; 1 μM), and treatment for 30 min at 10 μM with compounds (A) U1, (B) U2, (C) W1 and (D) W2. The bar graphs in each panel represent relative values of area under the curve (AUC) following NE stimulation. Data in the graphs represent the mean ± s.e.m. of 2–3 independent experiments. Where not visible, standard error bars are contained within the symbol sizes. ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Figure 5

Them1 inhibitors increase fatty acid oxidation rates and suppress glucose production in cultured primary hepatocytes. (A) Oxygen consumption rate (OCR) values in primary hepatocytes following stimulation with palmitic acid conjugated with fatty acid-free BSA (300 μM) and treatment with compound U1 (33 μM). The bar graphs in each panel represent relative values of area under the curve (AUC) following stimulation with palmitatic acid. Data in the graphs represent the mean ± s.e.m. of 2 independent experiments. Where not visible, standard error bars are contained within the symbol sizes. ∗P < 0.05; wild type (DMSO) vs. L-Them1Tg (DMSO). (B) Relative glucose production compared to wild type controls in serum-starved primary hepatocytes were determined by appearance of glucose in media following addition of DMSO or compounds U1 (32 μM, left and right panels) or W1 (20 μM, left and right panels), and pyruvate (2 mM) and lactate (20 mM) as gluconeogenic substrates. Data in the graphs represent the mean ± s.e.m. of 2 independent experiments. ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Figure 6

START domain-dependent inhibition of Them1 activity by freshly synthesized compound U1 and novel structural derivatives. Scaffold modifications of freshly synthesized compound U1 and novel structural derivatives (U6 – U17). Reactions were performed in 384-well microplates with recombinant His-tagged human Them1 (Them1; 125 nM), myristoyl-CoA (C₁₄-CoA; 25 μM) and compounds, and incubated for 60 min at 22 °C. Data represent the mean (s.e.m.) of triplicate determinations. Where not visible, error bars are contained within the symbol sizes.

Figure 7

Compound U1 and structural derivatives selectively bind the Them1 START domain. Binding interactions were quantified as K_d values using recombinant His-tagged proteins for (A) full-length Them1 [Them1; (100 nM)] and a truncated Them1 either containing only the START domain but lacking the 2 thioesterase domains [START; (100 nM)], or (B) a truncated Them1 either containing only the 2 thioesterase domains but lacking the START domain [Them1-ΔSTART; (100 nM)], full-length human Acot12 [Acot12; (100 nM)] or human PC-TP [PC-TP; 25 nM)]. Data represent the mean (s.e.m.) of triplicate determinations. Where not visible, error bars are contained within the symbol sizes. Compound K_d values targeting Them1 or START are listed.

Figure 8

Novel Them1 inhibitors increase oxygen consumption rates (OCR) in cultured primary brown adipocytes. (A–E) OCR values in primary brown adipocytes cultured from wild type and transgenic adipose tissue-specific Them1 overexpression (A-Them1Tg) mice following stimulation with norepinephrine (NE; 1 μM), and treatment for 30 min with the following synthesized compounds: (A) U6 (35 μM) and U9 (50 μM), (B) U11 (10 μM), (C) U12 (10 μM), (D) U14 (10 μM) and U15 (30 μM), and (E) U8 (30 μM) and U16 (60 μM). The bar graphs in each panel represent relative values of area under the curve (AUC) following NE stimulation. Data in the graphs represent the mean ± s.e.m. of 2 independent experiments. Where not visible, standard error bars are contained within the symbol sizes. ∗, P < 0.05; ∗∗∗, P < 0.001.

Figure 9

Evidence supporting a pathogenic role of Them1 in human NAFLD.Them1 mRNA expression levels measured in brown adipose tissue (BAT), liver and white adipose tissue from human subjects were gleaned from the NCBI GEO or ArrayExpress databases: (A) BAT samples from a human subject who volunteered to underdo biopsy at thermoneutrality (26 ± 1.2 °C) and under cold exposure (18.2 ± 2.1 °C) (Accession #: E-MTAB-4031). (B) Liver samples from lean subjects and patients with non-alcoholic fatty liver disease (NAFLD). Lean, n = 10; NAFLD, n = 10. (Accession #: GSE135776). (C) Liver samples from patients with NAFLD or non-alcoholic steatohepatitis (NASH). NAFLD, n = 51; NASH, n = 47. (Accession #: GSE167523). (D) Subcutaneous white adipose tissue (scWAT) from lean and obese human subjects. Lean, n = 10; Obese, n = 10. (Accession #: GSE162653). (E) scWAT and visceral WAT (vWAT) in obese human subjects. scWAT, n = 10; vWAT, n = 83. (Accession #: GSE135251). Data in the graphs represent the mean ± s.e.m. ∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001.