ADH1B, the adipocyte-enriched alcohol dehydrogenase, plays an essential, cell-autonomous role in human adipogenesis

Abstract

Alcohol dehydrogenase 1B (ADH1B) is a primate-specific enzyme which, uniquely among the ADH class 1 family, is highly expressed both in adipose tissue and liver. Its expression in adipose tissue is reduced in obesity and increased by insulin stimulation. Interference with ADH1B expression has also been reported to impair adipocyte function. To better understand the role of ADH1B in adipocytes, we used CRISPR/Cas9 to delete ADH1B in human adipose stem cells (ASC). Cells lacking ADH1B failed to differentiate into mature adipocytes manifested by minimal triglyceride accumulation and a marked reduction in expression of established adipocyte markers. As ADH1B is capable of converting retinol to retinoic acid (RA), we conducted rescue experiments. Incubation of ADH1B-deficient preadipocytes with 9-cis-RA, but not with all-transretinol, significantly rescued their ability to accumulate lipids and express markers of adipocyte differentiation. A homozygous missense variant in ADH1B (p.Arg313Cys) was found in a patient with congenital lipodystrophy of unknown cause. This variant significantly impaired the protein's dimerization, enzymatic activity, and its ability to rescue differentiation in ADH1B-deficient ASC. The allele frequency of this variant in the Middle Eastern population suggests that it is unlikely to be a fully penetrant cause of severe lipodystrophy. In conclusion, ADH1B appears to play an unexpected, crucial and cell-autonomous role in human adipocyte differentiation by serving as a necessary source of endogenous retinoic acid.

Keywords: 9-cis retinoic acid; ADH1B; adipocyte differentiation; alcohol dehydrogenase 1B; human adipose stem cells.

Fig. 1.

ADH1B deficiency in ASC alters lipid droplet formation and triglyceride content. Data were obtained in ASC, ASC with a CRISPR-Cas9-mediated ADH1B-knockout (KO), and ASC transduced with a Cas9/scramble gRNA plasmid corresponding to control (CTL) cells. (A) Timeline representation of the ASC differentiation process using a hormonal cocktail. IBMX: 3-isobutyl-1-methylxanthine; D0: day 0 (undifferentiated state); D10: day 10; D20: day 20. (B) ADH1B expression in ASC during adipocyte differentiation and validation of ADH1B KO in ASC at D0 and at D20. Numbers on the left correspond to molecular weight markers (kDa). Western blot images are representative of three independent experiments. (C) Adipocyte differentiation assessed by Oil Red-O lipid staining. ASC preadipocytes were studied during adipocyte differentiation for 20 d. First and second lines: representative pictures of cell dishes by optical microscopy. Images are representative of three independent experiments. Third and fourth lines: representative images of fluorescence microscopy after staining of intracellular lipids (Oil Red-O, red) and nuclei (DAPI, blue). Images are representative of three independent experiments. (D) Quantification of Oil Red-O fluorescence normalized to DNA content (DAPI). Results are expressed as means ± SEM of three independent experiments. (E) Intracellular triglyceride contents were measured at D0 and D20 in ASC, CTL, and ADH1B KO cells. The measurements are representative of three independent experiments. (F) Protein expression of adipocyte markers obtained by western blotting during in vitro adipocyte differentiation of ASC cells at D0 and D20. Numbers on the left correspond to molecular weight markers (kDa). Western blot images are representative of three independent experiments. PPARγ: peroxisome proliferator-activated receptor-gamma; C/EBPα: CCAAT/enhancer-binding protein-alpha; SREBP-1c: sterol regulatory element-binding protein-1c; FAS: fatty acid synthase. Numbers on the left correspond to molecular weight markers (kDa). Western blot images are representative of three independent experiments. **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s. designates nonspecific bands.

Fig. 2.

ADH1B deficiency can be bypassed by 9-cis RA treatment, but not by rosiglitazone. Data were obtained in ASC, ASC with a CRISPR-Cas9-mediated ADH1B-knockout (KO), and ASC transduced with a Cas9/scramble gRNA plasmid corresponding to control (CTL) cells. Adipocyte differentiation was induced in the presence or not of 9-cis RA (9-cis RA). Cells are studied at D20 postinduction. (A) Schematic representation of the link between ADH1B, 9-Cis RA, and adipocyte differentiation. The yellow square corresponds to 9-Cis RA. (B) Adipocyte differentiation assessed by Oil Red-O lipid staining. First line: representative pictures of cell dishes by optical microscopy. Second line: representative images of fluorescence microscopy after staining of intracellular lipids (Oil Red-O, red) and nuclei (DAPI, blue). Images are representative of three independent experiments. (C and G) Quantification of Oil Red-O fluorescence normalized to DNA content (DAPI) at D20. Results are expressed as means ± SEM of three independent experiments. (D) Intracellular triglyceride contents were measured at D20 in ASC, CTL, and ADH1B KO cells. The measurements are representative of three independent experiments. (E) Protein expression of adipocyte markers obtained by western blotting during in vitro adipocyte differentiation of ASC at D20. Numbers on the left correspond to molecular weight markers (kDa). Western blot images are representative of three independent experiments. PPARγ: peroxisome proliferator-activated receptor-gamma; C/EBPα: CCAAT/enhancer-binding protein-alpha; SREBP-1c: sterol regulatory element-binding protein-1c; FAS: fatty acid synthase. (F) Representative images of fluorescence microscopy after staining of intracellular lipids (Oil Red-O, red) and nuclei (DAPI, blue) in ADH1B-KO ASC. The first line depicts KO cells without rosiglitazone, while the second line shows those treated with it. (G) Quantification of Oil Red-O fluorescence normalized to DNA content (DAPI) at D20. Results are expressed as means ± SEM of three independent experiments. **P < 0.01, ****P < 0.0001, n.s.: nonsignificant.

Fig. 3.

The block in adipogenesis is not rescued by retinol or small doses of ATRA. Data were obtained in ASC, ASC with a CRISPR-Cas9-mediated ADH1B-knockout (KO), and ASC transduced with a Cas9/scramble gRNA plasmid corresponding to control (CTL) cells. Adipocyte differentiation was induced in the presence or not of 9-cis RA, all-trans retinol (ATRetinol) and all-trans RA (ATRA). Cells are studied at D20 postinduction. (A and E) Adipocyte differentiation assessed by Oil Red-O lipid staining. Representative images of fluorescence microscopy after staining of intracellular lipids (Oil Red-O, red) and nuclei (DAPI, blue) are depicted. Images are representative of three independent experiments. (B and F) Quantification of Oil Red-O fluorescence normalized to DNA content (DAPI) at D20. Results are expressed as means ± SEM of three independent experiments. (C and G) Intracellular triglyceride contents were measured at D20 in ASC, CTL, and ADH1B KO cells. The measurements are representative of three independent experiments. (D) Protein expression of adipocyte markers obtained by western blotting during in vitro adipocyte differentiation of ASC at D20. Numbers on the left correspond to molecular weight markers (kDa). Western blot images are representative of three independent experiments. PPARγ: peroxisome proliferator-activated receptor-gamma; C/EBPα: CCAAT/enhancer-binding protein-alpha; SREBP-1c: sterol regulatory element-binding protein-1c; FAS: fatty acid synthase. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 4.

Identification of an ADH1B homozygous variant affecting ADH1B homodimerization and enzyme activity in a patient with generalized lipoatrophy. (A) Genealogical tree and segregation analysis for the ADH1B variant in the family investigated herein. The arrow indicates the proband. +, normal allele; M, mutant allele. (B) Top panel: schematic representation of ADH1B transcript sequence (NM_000668.6) displaying the location of the variant identified. Bottom panel: schematic representation of ADH1B protein sequence comprising 375 amino acids. The prediction of protein domain organization was based on the UniProt database (protein reference: P00325). (C) Classification of amino acids in order of increasing hydrophobicity using a Fauchère and Pliska plot and showing the major change induced by the p.Arg313Cys variant. (D) Prediction of the effect of p.Arg313Cys variant on ADH1B 3D structure. This was performed using the MIZTLI software (https://miztli.biokerden.eu/) and the 1u3u ADH1B structure from Protein Data Bank (https://www.wwpdb.org/). Arg313 was replaced by a cysteine residue using FASPR (23). Dashed lines represent noncovalent hydrogen bonds. The p.Arg313Cys variant induces the loss of a number noncovalent interactions with other ADH1B residues, which is predicted to affect the protein stability. (E) Model of the 3D structure of ADH1B extracted from SWISS MODEL repository and corresponding to the crystal structure of a human homodimer of ADH1B (https://swissmodel.expasy.org—protein reference: P00325). The location of the p.Arg313Cys variant identified in patient 1 is indicated by a star. (F) Transfection of HEK 293 cells with plasmids encoding wild-type (WT) ADH1B and the p.Arg313Cys mutated form of the protein with a C-terminal Flag tag. Cellular extracts were subjected to native-PAGE (Upper panel) vs. SDS-PAGE (Middle panel) to evaluate the impact of the variant on ADH1B homodimerization. *: The lowest band on the gel corresponds to the one obtained with the anti-Flag antibody, since the same membrane was blotted to reveal tubulin. (G) Assessment of ADH activity in HEK 293 cells stably expressing either WT or mutant forms of ADH1B. The measurements are representative of three independent experiments. (H) WB analysis performed on ADH1B KO cells nucleofected with plasmids encoding ADH1B WT, mutant, or green fluorescent protein (GFP). (I) Optical microscopy images captured from ADH1B KO cells complemented with ADH1B WT, mutant, or GFP at D15 after adipocyte differentiation induction. (J) Measurement of intracellular triglyceride levels at D15 in ADH1B KO cells complemented with ADH1B WT, mutant, or GFP as a negative control. The measurements are representative of two independent experiments in duplicates. *P < 0.05, **P < 0.01.