Inhibition of acyl-CoA binding protein (ACBP) by means of a GABAARγ2-derived peptide

Abstract

Acyl-CoA binding protein (ACBP) encoded by diazepam binding inhibitor (DBI) is an extracellular inhibitor of autophagy acting on the gamma-aminobutyric acid A receptor (GABA_AR) γ2 subunit (GABA_ARγ2). Here, we show that lipoanabolic diets cause an upregulation of GABA_ARγ2 protein in liver hepatocytes but not in other major organs. ACBP/DBI inhibition by systemically injected antibodies has been demonstrated to mediate anorexigenic and organ-protective, autophagy-dependent effects. Here, we set out to develop a new strategy for developing ACBP/DBI antagonists. For this, we built a molecular model of the interaction of ACBP/DBI with peptides derived from GABA_ARγ2. We then validated the interaction between recombinant and native ACBP/DBI protein and a GABA_ARγ2-derived eicosapeptide (but not its F77I mutant) by pull down experiments or surface plasmon resonance. The GABA_ARγ2-derived eicosapeptide inhibited the metabolic activation of hepatocytes by recombinant ACBP/DBI protein in vitro. Moreover, the GABA_ARγ2-derived eicosapeptide (but not its F77I-mutated control) blocked appetite stimulation by recombinant ACBP/DBI in vivo, induced autophagy in the liver, and protected mice against the hepatotoxin concanavalin A. We conclude that peptidomimetics disrupting the interaction between ACBP/DBI and GABA_ARγ2 might be used as ACBP/DBI antagonists. This strategy might lead to the future development of clinically relevant small molecules of the ACBP/DBI system.

Fig. 1. GABA_ARγ2 expression across metabolically relevant organs responding to obesogenic diets in mice.

A Experimental design. C57BL/6 male mice were fed with regular chow diet (RCD), high-fat diet (HFD) or high-fat/high-sucrose/1.25% cholesterol diet (Western diet: WD) over the course of 6 weeks. B Body weight gain measurements in RCD-, HFD- or WD-fed mice. Results are displayed as whisker plots with each dot representing one single mouse (n = 4–5 mice per condition) including the mean ± SEM. For statistical analysis, p values are calculated by one-way ANOVA. C Spearman correlation between body weight gain (compared to week 0) and plasma ACBP/DBI concentration in individual mice fed with RCD, HFD or WD (n = 5 mice per group). D Liver, epididymal white adipose tissue (eWAT), interscapular brown adipose tissue (BAT), muscle, and heart protein extract immunoblot detection of GABA_ARγ2, ACBP/DBI, and β-actin from mice receiving RCD, HFD or WD (n = 5 mice per condition). E Heatmap representation of ACBP/DBI, and GABA_ARγ2 relative protein levels from RCD-, HFD- or WD-fed mice (n = 5 mice per condition). Red and green arrows indicate increased and decreased proteins of interest, respectively (compared to RCD). For statistical analysis, p values were calculated by one-way ANOVA. a.u. arbitrary units.

Fig. 2. Hepatocyte-specific GABA_ARγ2 upregulation in response to obesogenic diets determined by immunofluorescence.

A Experimental design. C57BL/6 male mice were fed with RCD or HFD over the course of 6 weeks. B Representative confocal microscopy images and quantification (C) of GABA_ARγ2 levels in mice fed with RCD or HFD (n = 4 mice per condition). F4/80 and albumin are markers specific for murine macrophages and hepatocytes, respectively. For statistical analyses, p values are calculated by two-tailed unpaired Mann–Whitney U-test. Scale bar: 10 μm. a.u.: arbitrary units.

Fig. 3. Hepatic GABA_ARγ2 upregulation in response to hepatic lipoanabolism determined by immunohistochemistry.

A Experimental design. C57BL/6 male mice were fed with RCD, WD, and methionine/choline-deficient diet (MCD) over the course of 6 weeks. B Representative images of hepatic hematoxylin and eosin (HE), and GABA_ARγ2 immunohistochemistry (IHC) staining from mice fed with RCD, WD or MCD (n = 5 mice per condition), non-alcoholic fatty liver disease (NAFLD) activity score quantification (C). Hepatic GABA_ARγ2quantification (expressed as % percentage of GABA_ARγ2 -positive cells) in liver parenchyma (D) and peri-central / periportal vein (E) areas. Results are displayed as whisker plots with each dot representing one single mouse (n = 5 mice per condition) including the means ± SEM. For statistical analyses, p values are calculated by two-tailed unpaired Mann–Whitney U-test (C) or one-way ANOVA (D, E).

Fig. 4. Molecular modeling of the interaction between ACBP/DBI and the GABA_ARγ2-derived eicosapeptide.

A Computational model of GABA_AR γ2 subunit of GABA_AR (GABA_AR γ2) is highlighted in orange color. The binding site of ACBP/DBI on GABA_AR γ2 (amino acid residues 106–125) is indicated with a black arrow. B In silico homology modeling of predicted physical interaction between ACBP/DBI and wild-type (WT) or “phenylalanine 77 to isoleucine”-substitution mutation (F77I) GABA_ARγ2 peptides. ACBP/DBI is represented in violet while WT and F77I peptides are represented in orange and green respectively. C Schematic representation of amino acid sequence features between WT and F77I GABA_ARγ2 peptides (both encoding for the 106–125 amino acid residues of the full GABA_ARγ2 protein). WT peptide (black stars) is designed in a way that it is predicted to bind to ACBP while, in contrast, the mutated F77I peptide (blue stars) is designed in a such way that it is predicted that it will not bind to ACBP.

Fig. 5. Biophysical and biochemical detection of a direct interaction between ACBP/DBI and the GABA_ARγ2-derived eicosapeptide.

A Schematic representation of the GABA_AR and the GABA_ARγ2-derived eicosapeptides. B Surface plasmon resonance kinetics of ACBP-GABA_ARγ2-derived eicosapeptide (WT and F77I) interactions. Representative ACBP immunoblot (C) from control- or GABA_ARγ2-immunoprecipitated samples. Anti-GABA_ARγ2 immunoprecipitation (IP) assay was employed in order to validate the GABA_ARγ2 WT peptide interaction with native ACBP (in murine hepatic protein extracts), and quantification (D). Results are displayed as whisker plots (with each dot representing one single biological replicate) including the mean ± SEM. For statistical analysis p value (indicating statistical comparisons with the control condition) was calculated by Mann–Whitney U-test. E Anti-GABA_ARγ2 IP assay to address the physical interaction between GABA_ARγ2 peptides and murine hepatic ACBP protein. Interaction was validated using an increasing gradient of biotinylated WT/F77I peptides that were bound on the surface of streptavidin beads as part of the IP assay. Livers from tamoxifen-inducible whole-body knockout mice with floxed Acbp/Dbi^f/f exon 2 (Acbp/Dbi KO) and livers from wild-type Acbp/Dbi^f/f control were used as the input controls. kDa kilodaltons, (L): protein ladder. a.u. arbitrary units.

Fig. 6. Inhibition of the metabolic effects of ACBP/DBI on hepatocytes by the GABA_ARγ2-derived eicosapeptide.

A WST-8 conversion by murine primary hepatocytes treated with increasing concentrations of recombinant ACBP/DBI protein (RecACBP/DBI) for 4 h. Hepatocytes were isolated from tamoxifen-induced Acbp/Dbi f/f or Acbp/Dbi KO full-body genetic background mice (white- or blue-colored graphs respectively). Results are displayed as column plots with each dot representing hepatocytes isolated and cultured from one single mouse (n = 3 mice per condition) including the means ± SEM. For statistical analysis p values are calculated by one-way ANOVA. B WST-8 conversion by control (shUNR) and ACBP/DBI-knocked down (shACBP/DBI) human Huh-7 cells treated with increasing concentrations of RecACBP/DBI for 4 h. shUNR and shACBP/DBI cells are represented by white- or blue-colored graphs respectively. Results are displayed as column plots with each dot representing cells one biological replicate (n = 8–9 per condition) including the mean ± SEM. C WST-8 conversion by Huh-7 cells treated with RecACBP/DBI or negative control (4 h). Cells were pre-treated with 2-deoxy-D-glucose (2-Deoxy-D-glu; 300 µM) or vehicle (2 h). Results are displayed as column plots with each dot representing cells one biological replicate (n = 4 per condition) including the means ± SEM. D WST-8 conversion by Huh-7 cells treated with RecACBP/DBI or negative control (4 h). Cells were pre-treated with rotenone (0.5 µM) or vehicle (2 h). Results are displayed as column plots with each dot representing cells one biological replicate (n = 4 per condition) including the means ± SEM. E WST-8 conversion by Huh-7 cells treated with RecACBP/DBI or negative control (4 h). Prior to administration, RecACBP/DBI was overnight pre-incubated with monoclonal antibody (α-ACBP/DBI) or isotype IgG1. Results are displayed as column plots with each dot representing cells one biological replicate (n = 4 per condition) including the means ± SEM. F WST-8 conversion by Huh-7 cells treated with RecACBP/DBI or negative control (4 h). Prior to administration, RecACBP/DBI was overnight pre-incubated with GABA_ARγ2 (WT or F77I-mutated) eicosapeptides. Results are displayed as column plots with each dot representing cells one biological replicate (n = 3 per condition) including the means ± SEM. p values are calculated by one-way ANOVA.

Fig. 7. In vivo inhibition of ACBP/DBI function by the GABA_ARγ2-derived eicosapeptide.

A, B Appetite stimulation by ACBP/DBI. RecACBP/DBI alone or combined with GABA_ARγ2-derived eicosapeptides (EP) (WT or F77I) was administered in vivo by intravenous (i.v.) injection and individual cumulative food intake was measured after 30 min. GABA_ARγ2 peptides are represented as blue stars, while murine RecACBP/DBI is represented as red circles (A). Results (B) are displayed as column plots with each dot representing one single mouse (n = 4 to 9 mice per condition) including the means ± SEM. p values were calculated by two-way ANOVA. C, D Immunoblots and densitometric quantification of LC3 lipidation in liver tissues from mice treated with GABA_ARγ2-derived eicosapeptides (EP) WT with or without the autophagy flux inhibitor leupeptin (n = 5 mice per group). p values were calculated by unpaired one-tailed Student’s t-test. E–G Inhibition of concanavalin A hepatotoxicity. GABA_ARγ2 WT peptide (or negative vehicle) was i.v. injected to C57BL/6 male mice followed by concanavalin A (or negative vehicle) administration 1.5 h later (E). Alanine transaminase (ALT) (F) and aspartate transaminase (AST) (G) activity in plasma was analyzed by means of a colorimetric assay (n = 7–9 mice per condition). p values were calculated by unpaired one-tailed Student’s t-test. ns: non-significant. Unt: Untreated.