The transition from lean to obese states involves systemic metabolic remodeling that impacts insulin sensitivity, lipid partitioning, inflammation, and glycemic control. Here, we have taken a pharmacological approach to test the role of a nutrient-regulated chromatin modifier, lysine-specific demethylase (LSD1), in obesity-associated metabolic reprogramming. We show that systemic administration of an LSD1 inhibitor (GSK-LSD1) reduces food intake and body weight, ameliorates nonalcoholic fatty liver disease (NAFLD), and improves insulin sensitivity and glycemic control in mouse models of obesity. GSK-LSD1 has little effect on systemic metabolism of lean mice, suggesting that LSD1 has a context-dependent role in promoting maladaptive changes in obesity. In analysis of insulin target tissues we identified white adipose tissue as the major site of insulin sensitization by GSK-LSD1, where it reduces adipocyte inflammation and lipolysis. We demonstrate that GSK-LSD1 reverses NAFLD in a non-hepatocyte-autonomous manner, suggesting an indirect mechanism potentially via inhibition of adipocyte lipolysis and subsequent effects on lipid partitioning. Pair-feeding experiments further revealed that effects of GSK-LSD1 on hyperglycemia and NAFLD are not a consequence of reduced food intake and weight loss. These findings suggest that targeting LSD1 could be a strategy for treatment of obesity and its associated complications including type 2 diabetes and NAFLD.

The root cause of obesity is an imbalance between energy intake and expenditure. The resulting body weight gain is associated with remodeling of systemic metabolism. The metabolic remodeling gives rise to interrelated defects including insulin resistance, ectopic lipid deposition, and chronic, low-grade inflammation (1,2). These defects predispose people with obesity to more serious chronic diseases including cardiovascular disease, type 2 diabetes (T2D), and nonalcoholic fatty liver disease (NAFLD), which together comprise much of the public health burden of obesity.

Obesity causes insulin resistance in white adipose tissue (WAT) (3,4), thereby increasing lipolysis and the trafficking of free fatty acids (FFA) from WAT to the liver (5,6). This oversupply of FFA to the liver contributes to accumulation of triglycerides that promote hepatic steatosis, the defining characteristic of NAFLD (68). Systemic insulin resistance caused by obesity increases pancreatic β-cell workload and thereby also predisposes to T2D, which manifests when the β-cell adaptive response fails to produce adequate insulin to meet demand (6,9). The hyperinsulinemia associated with successful β-cell adaptation promotes lipogenesis in the liver and exacerbates hepatic steatosis, underscoring the systemic nature of metabolic defects in obesity (10,11). Despite the urgent global need to ameliorate complications associated with obesity, there are few therapeutic options to treat obesity itself (2). As obesity is a complex multiorgan disease, therapies that address its root cause or simultaneously target the secondary metabolic defects may be necessary to mitigate the associated complications. Unfortunately, current therapeutic efforts for obesity are limited by an incomplete understanding of how the nutrient environment is interpreted by various metabolically relevant tissues to evoke adaptive or maladaptive changes in obesity (12). Overall, there is an unmet need to understand context-dependent regulation of metabolism in the obese compared with the lean state.

Adaptation of systemic metabolism to changing nutrient states involves several layers of regulation that affect the intake, storage, and expenditure of energy. These processes are controlled at the whole-organism level through the endocrine and central nervous systems (13,14). Whether these systemic controls are augmented by unified tissue-intrinsic mechanisms is currently unclear. One such potential mechanism is the transcriptional regulation of metabolism, which is mediated by transcription factors and coregulators whose functions are coupled to nutrient state (1519). For example, intermediary metabolites have been shown to regulate the enzymatic activities of several coregulators that modify the epigenome (20,21). Changes in cellular metabolism lead to altered abundance of substrates for these coregulators, thereby altering their enzymatic activities to change epigenomic states and gene expression (22). Posttranslational modifications of coregulators in response to hormonal or nutrient signals provide an additional layer of regulation that links coregulator function to nutrient state (15,17). Genetic inactivation of transcriptional coregulators has revealed pervasive roles for these enzymes in governing nutrient responses of metabolically relevant tissues (1518,2327). For instance, SIN3A is an insulin-regulated corepressor of the glucokinase gene that restrains hepatic lipogenesis in the fasted state. In the absence of SIN3A, the lipogenic effect of insulin is dampened (18). Overall, nutrient regulation confers context-specific functions to transcriptional coregulators, and therefore these enzymes could underlie context-dependent regulatory mechanisms specific to lean or obese states. Given recent drug development efforts to target transcriptional coregulators, understanding how the compendium of these enzymes impacts metabolism could lead to new therapies for metabolic disease.

We and others have identified key roles for lysine-specific histone demethylase 1A (LSD1/KDM1A) in context-dependent regulation of metabolism. The demethylase activity of LSD1 requires flavin adenine dinucleotide as a cofactor, linking cellular metabolism to the nutrient environment (26). Given its nutrient responsiveness and broad expression pattern (28), we reasoned that LSD1 would be well positioned to mediate systemic changes to metabolism. LSD1 is part of transcriptional complexes in the nucleus that regulate nutrient responses in various cell types including hepatocytes, adipocytes, and pancreatic β-cells (21,2732). LSD1 restrains the insulin secretory response by pancreatic β-cells in the fed state (27), whereas in the liver, LSD1 promotes insulin-stimulated lipogenesis in response to feeding (2931). Furthermore, beiging of WAT during cold exposure requires LSD1 to activate thermogenesis (32,33). While these studies indicate that LSD1 regulates context-specific functions of several metabolically relevant tissues, it is unknown whether LSD1 is involved in the remodeling of systemic metabolism in response to changing nutrient states.

In the current study, we used a pharmacological approach to test the role of LSD1 in metabolic remodeling associated with obesity. We show that systemic LSD1 inhibition reduces hyperphagia and weight gain, improves insulin sensitivity, and prevents hyperglycemia in obesity models, while having no discernable metabolic effects in lean mice. Mechanistically, we found that LSD1 inhibition reduces adipose tissue inflammation and ameliorates NAFLD. In intervention studies, we demonstrate that LSD1 inhibition can reverse Western diet–induced weight gain and hyperinsulinemia as well as enhance insulin signaling. Together, our data suggest that LSD1 could be a therapeutic target in metabolic disease that simultaneously addresses the root cause of obesity and its associated complications.

Mice

Mice homozygous for the leptin receptor (BKS.Cg-Dock7m+/+Leprdb/J, no. 000642) (hereafter referred to as db/db) and the respective lean control animals (db/+) as well as C57BL/6J mice (no. 000664) were purchased from The Jackson Laboratory (34,35). Lsd1fl/fl mice (provided by the laboratory of Michael Rosenfeld, UCSD) (36) were bred with db/db mice to generate Lsd1fl/fldb/db mice. Liver-specific Lsd1 deletion was induced with intravenous injection of AAV8-TBG-iCre virus (2.5 × 1011 genome copies/mouse). As controls, Lsd1fl/+db/db and Lsd1+/+db/db mice were injected with the virus in parallel. All mice were housed and bred in vivaria approved by the Association for Assessment and Accreditation of Laboratory Animal Care located in the School of Medicine, University of California, San Diego (UCSD), following standards and procedures approved by the UCSD Institutional Animal Care and Use Committee. Mice were weaned at 4 weeks, maintained on a 12-h light cycle, and fed ad libitum with water and standard rodent chow (PicoLab Rodent Diet 20 5053), a Western diet (TD.88137 Envigo Teklad) containing 42% kcal from fat, or a high-fat diet (HFD) (D12492; Research Diets) containing 60% kcal from fat, unless otherwise indicated. Both male and female mice were used for all studies on db/db mice, and male mice only were used for Western diet and HFD studies. Mice received GSK-LSD1 (500 μg/kg/mouse, SML1072; Sigma-Aldrich) or vehicle (veh) (0.9% NaCl) daily via intraperitoneal injections. SP2509 (25 mg/kg/mouse, HY-12636; MedChemExpress), dissolved in 10% DMSO, 10% Tween80 in PBS, or veh (10% DMSO, 10% Tween80 in PBS) were injected intraperitoneally three times per week. Body weight and blood glucose levels were monitored weekly.

Metabolic Studies

For glucose tolerance tests, mice were fasted for 5 h and orally gavaged with 2 mg/g body wt glucose. Plasma glucose levels were measured in blood samples from the tail vein at baseline, 15 min, 30 min, 60 min, 90 min, and 120 min after gavage with a Bayer Contour glucometer. For insulin tolerance tests, an insulin solution (0.8–2.0 units/kg body wt i.p.) was administered to fasted (5 h) mice and glucose levels were monitored as described above. For pyruvate tolerance tests, mice starved overnight (16 h) were injected with a pyruvate solution (1 mg/g body wt i.p.) and plasma glucose levels were analyzed as described above.

ELISA

Plasma insulin levels were measured after 5 h of fasting or before and 10 min after a glucose gavage (2 mg/g body wt) of fasted (5 h) mice via the Mouse Ultrasensitive Insulin ELISA or Mouse Insulin ELISA kit (ALPCO). Norepinephrine levels were measure in homogenized WAT samples with ELISA according to the manufacturer’s instructions (Abcam).

Acute Insulin Response

For determination of tissue insulin responsiveness, mice were starved overnight (16 h) and injected with 2.0 units/kg body wt i.p. of an insulin solution. Tissues were harvested 15 min after insulin stimulation and flash frozen in liquid N2. Lysates were then prepared of gWAT, liver, and skeletal muscle, and samples were subsequently analyzed via Western blotting.

To measure insulin sensitivity in vitro, differentiated adipocytes were preincubated with GSK-LSD1 or veh overnight. The next day, cells were serum starved in Krebs buffer for 2 h and stimulated with 10 mmol/L insulin for 10 min. Treatment with GSK-LSD1 or veh was continued throughout the experiment. Adipocytes were then washed with ice-cold PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific) containing protease inhibitor and phosphatase inhibitor, and protein content was determined with commercial kits. Insulin responsiveness in adipocytes was then quantified through visualizing Akt and phosphorylated (p)-Akt via Western blotting.

Western Blotting

Tissues were homogenized with RIPA buffer (Thermo Fisher Scientific) and freshly added protease inhibitor cocktail (Roche) and phosphatase inhibitor (Sigma-Aldrich). Lysates of gWAT (20 μg), liver (40 μg), and skeletal muscle (40 μg) of db/db mice and of liver (30 μg) from Lsd1ΔLdb/db mice were analyzed with SDS-PAGE on 12% Bis-Tris gels with equal amount of protein loading. Proteins were visualized after transfer onto a polyvinylidene fluoride blotting membrane (GE Healthcare) and incubation with specific primary antibodies with horseradish peroxidase–conjugated secondary antibodies. Western blot primary antibodies included: rabbit anti-mouse Akt (no. 4691S; Cell Signaling Technology), rabbit anti-mouse p-AktSer473 (no. 4060S; Cell Signaling Technology), mouse anti-mouse GAPDH (no. AM4300; Thermo Fisher Scientific), rabbit anti-mouse Hsl (no. 4107; Cell Signaling Technology), rabbit anti-mouse p-HslSer660 (no. 4126; Cell Signaling Technology), rabbit anti-mouse LSD1 (no. 17721; Abcam), and mouse anti-mouse vinculin (no. ab18058; Abcam). Western blot secondary antibodies included goat anti-rabbit horseradish peroxidase (no. 4010-05; Southern Biotech) and ECL sheep anti-mouse (no. NA931V; GE Healthcare).

Histology

Immunohistochemical analysis was performed on sections of paraformaldehyde-fixed and paraffin-embedded tissues. Histology was performed by the UCSD Histology and Immunohistochemistry Core. Routine hematoxylin-eosin (H-E) stain was performed on adipose tissues, liver, and skeletal muscle. The size of adipocytes was determined through quantifying at least 100 adipocytes with ImageJ analysis software (National Institutes of Health).

For analysis of crown-like structures (CLS), adipose tissues were stained with rat anti-F4/80 primary antibody (no. MCA497B; Bio-Rad Laboratories) and goat anti-rat horseradish peroxidase polymer (no. AH-100; Cell IDX) followed by chromogenic 3,3′-diaminobenzidine (no. 95041-478; VWR) and counterstained with Val Hematoxylin (no. VLT8014G20; Biocare Medical). Slides were imaged with a slide scanner machine (NanoZoomer; Hamamatsu Photonics), and blinded samples were analyzed with ImageJ analysis software. The number of CLS was quantified on whole tissue sections.

RNA Extraction, Quantitative RT-PCR, and mRNA-Sequencing Library Preparation

Total RNA was isolated in Trizol from homogenized tissue and cells and purified using RNeasy columns and RNase-free DNase Digestion according to the manufacturer’s instructions (QIAGEN). The quality and quantity of the total RNA were monitored and measured with a NanoDrop (NanoDrop Technologies, Wilmington, DE) according to the manufacturer’s instructions. For quantitative PCR analysis, cDNA was synthesized with the iScript cDNA Synthesis Kit (Bio-Rad Laboratories) and 500 ng isolated RNA per reaction. Template cDNA (20 ng per reaction) and iQ SYBR Green Supermix (Bio-Rad Laboratories) were used for real-time PCR with gene-specific primers (Supplementary Table 1) and Tbp as a housekeeping gene on a CFX96 Real-Time PCR Detection System (Bio-Rad Laboratories).

For mRNA-sequencing (mRNA-seq) library preparation, total RNA was assessed for quality with an Agilent 4200 TapeStation, and samples with an RNA integrity number >8.0 were used to generate RNA-sequencing libraries with the Illumina Stranded mRNA Prep (Illumina, San Diego, CA). For each sample, 500 ng RNA was processed according to the manufacturer’s instructions. Resulting libraries were multiplexed and sequenced with 100 base pair paired end reads to a depth of ∼50 million reads per sample on an Illumina NovaSeq 6000. Samples were demultiplexed with bcl2fastq Conversion Software (Illumina).

mRNA-seq Data Analysis

mRNA-seq reads were mapped to the NCBI37/mm9 (mouse) genome with STAR 2.4.0f1 (--outSAMstrandField intronMotif, --outFilterMultimapNmax 1, and --runThreadN 5 options), excluding reads mapping to multiple loci. Genes with mean reads per kilobase of transcript per million reads mapped) ≥1 in at least one experimental group were considered to be expressed (total of 15,224 genes), and all nonexpressed genes were excluded from downstream analyses. Cuffdiff was used to assess expression differences for all pairwise comparisons, with P < 0.01 considered significant. For volcano plots, P values = 0 were graphed as 0.00001. K-means clustering was performed with R. All differentially expressed genes identified in this study are listed in Supplementary Table 2AC.

Gene Ontology

Functional categories related to differentially expressed genes and links between each pair of categories were identified with Metascape as described at https://metascape.org/. All gene ontology (GO) terms enriched among differentially expressed genes are listed in Supplementary Table 2D and E. Statistically enriched pathways from Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, and GO (biological process) were hierarchically clustered into a tree based on κ-statistical similarities among their gene memberships. A 0.3 κ-score was applied as the threshold to define clusters, with each representing a group of similar functional categories. A subset of representative terms from each cluster was automatically selected by Metascape and converted into a network, where terms with similarity >0.3 are connected by edges. Specifically, terms with the most significant P values from each of the clusters were depicted as network nodes, with the constraint of having a maximum of 15 terms per cluster and 250 terms in total, with node size representative of the degree of enrichment. For clarity, a representative term was selected to represent each cluster.

Lipolysis and Hsl Phosphorylation in Primary Adipocytes

Lipolysis was measured in differentiated adipocytes derived from the stromal vascular fraction of subcutaneous WAT (sWAT). In detail, sWAT harvested from 5-week-old db/db mice was minced and digested with collagenase D (2.5 mg/mL; Sigma) for 30 min shaking at 37°C. The digestion was stopped with addition of 10 mL culture medium (DMEM/F12 containing 10% FBS, 100 units/mL penicillin, and 0.1 mg/mL streptomycin), and the homogenate was filtered through a 100-μm cell strainer. After a 5-min centrifugation at 500g, the pellet containing the stromal vascular fraction was collected and incubated with erythrocyte lysis buffer for 5 min at room temperature. Next, the stromal vascular fraction was precipitated, and the cell pellet was then resuspended in culture medium and seeded into a T75 flask.

Once the preadipocytes reached 80% confluency, the cells were seeded into 12 wells and grown to confluency. Adipocyte differentiation was induced with addition of 0.1 μmol/L dexamethasone, 450 μmol/L isobutylmethylxanthine, 2 μg/mL insulin, and 1 μmol/L rosiglitazone (day 0). After 3 days, the medium was changed to culture medium containing 2 μg/mL insulin and 1 μmol/L rosiglitazone. On day 5, the medium was changed to culture medium and the adipocytes were grown for another 2 days.

Differentiated adipocytes were preincubated with 1 μmol/L GSK-LSD1 or veh overnight. The next day (day 7), adipocytes were serum starved in DMEM containing 2% fatty acid free BSA supplemented with 10 μmol/L isoproterenol or control. After 1 h, the incubation media was replaced, and the release of fatty acids (FUJIFILM Wako Laboratory Chemicals) and glycerol (Sigma-Aldrich) was determined after 4 h. Hsl phosphorylation was assessed with Western blotting in adipocytes harvested after 1 h of isoproterenol treatment. Treatment with GSK-LSD1 or veh was continued throughout the experiment. Data were normalized to cellular protein content as described above.

Lipid and Metabolic Analysis

Lipid levels were analyzed in plasma and liver samples as previously described (37,38). Blood was drawn via the tail vein from mice fasted for 5 h. Total cholesterol and triglyceride levels were determined in plasma and liver extracts with commercially available kits (Sekisui Diagnostics). Further, plasma β-hydroxybutyrate (Cayman Chemical) and ALT and AST (Sigma-Aldrich) levels were determined with enzymatic kits. Nonesterified fatty acids (NEFA) (FUJIFILM Wako Laboratory Chemicals) were measured in plasma obtained from overnight fasted mice.

Metabolic Cages

For determination of energy expenditure and food intake, db/db mice administered veh or GSK-LSD1 were placed into metabolic cages. Over a 48-h time frame, carbon dioxide production (VCO2), oxygen consumption (VO2), food intake, consumed water, and spontaneous motor activity were analyzed with the Comprehensive Lab Animal Monitoring System (CLAMS) (Columbus Instruments) at the UCSD Animal Phenotyping Core. The respiratory quotient was calculated as the ratio of VCO2 to O2.

Pair-Feeding Study

A pair-feeding experiment was performed on 4-week-old, individually housed db/db mice (both males and females), which were randomly assigned to the respective treatment groups. To establish pair-feeding conditions, a group of mice was administered with GSK-LSD1 as described above and fed a normal chow diet ad libitum. Food intake was measured daily between 4:00 and 6:00 p.m., and the average daily food consumption was calculated across all mice in this treatment group with separate calculations for male and female mice. Next, the veh-treated pair-fed mice were provided with the calculated average amount of food consumed by the GSK-LSD1–treated mice over the last 24 h. This procedure was then continued daily for 6 weeks. Leftover food of the pair-fed group, if any, was weighted and removed from the cage before new food was added. As a control group, veh-administered db/db mice were fed normal chow diet ad libitum for 6 weeks.

Glucose Production in Primary Hepatocytes

Primary hepatocytes were isolated from 5-week-old db/db mice by perfusion of the liver with EDTA to dissociate the cells, followed by Percoll centrifugation as previously described (39). Hepatocytes were seeded into collagen-coated 12-well plates at 300.000 cells/well; cultured in DMEM containing 10% FBS, 100 units/mL penicillin, and 0.1 mg/mL streptomycin; and preincubated with 1 μmol/L GSK-LSD1 or veh overnight. The next day, hepatocytes were serum starved in Krebs buffer (125 mmol/L NaCl, 4 mmol/L KCl, 1 mmol/L CaCl2, 1 mmol/L MgCl2, 0.85 mmol/L KH2PO4, 1.25 mmol/L Na2HPO4, 15 mmol/L NaHCO3, 10 mmol/L HEPES, and 0.2% fatty acid–free BSA) for 1 h and gluconeogenesis was stimulated with addition of 1 mmol/L pyruvate and 10 mmol/L lactate. After 2 h, glucose production was measured with the HK assay kit (Sigma-Aldrich). Cells were lysed in RIPA buffer (Thermo Fisher Scientific) containing protease inhibitor and phosphatase inhibitor, and protein content was determined with commercial kits.

Statistics

Statistical analyses was performed with GraphPad Prism 8 (GraphPad Software). Normality was tested via Shapiro-Wilk test, and F tests were performed to analyze equal variances. Data that passed both tests were analyzed with two-tailed Student t test for two-group comparisons and one-way ANOVA for comparison of multiple groups (more than two) followed by Tukey post hoc testing. For data with multiple variables, e.g., glucose measurements over time, two-way ANOVA for repeated measurements followed by Tukey post hoc test or Fisher least significant difference post hoc testing was performed. All data are presented as mean ± SEM. P values <0.05 were considered significant.

Study Approval

All mouse experiments were approved by the UCSD Institutional Animal Care and Use Committee.

Data and Resource Availability

mRNA-seq data have been submitted to Gene Expression Omnibus (GEO). All other data generated or analyzed in this study are included in the published article (and Supplementary Material). No applicable resources were generated or analyzed during the current study.

LSD1 Inhibition Prevents Hyperglycemia in db/db Mice

To investigate whether LSD1 contributes to metabolic dysfunction in obesity, we studied the effect of systemic LSD1 inhibition in db/db mice. Four-week-old db/db mice and lean db/+ control mice were injected daily with GSK-LSD1, an LSD1 inhibitor, or vehicle (veh), and metabolic parameters were monitored longitudinally for 6 weeks (Fig. 1A). As expected (40), veh-treated db/db mice rapidly gained weight and became hyperglycemic at ∼6 weeks of age (Fig. 1B). GSK-LSD1 reduced body weight gain of db/db mice (Fig. 1B). Remarkably, blood glucose levels of GSK-LSD1–treated db/db mice were comparable with levels in lean control mice (Fig. 1C). GSK-LSD1–treated db/db mice also exhibited improved glucose tolerance (Fig. 1D). Notably, GSK-LSD1 had no effect on body weight or blood glucose levels in lean db/+ mice (Supplementary Fig. 1A and B), suggesting that LSD1 regulates maladaptive changes to metabolism in obesity. Together, these findings show that systemic LSD1 inhibition with GSK-LSD1 prevents development of diabetes in db/db mice, a model of obesity and T2D. Supporting these findings, similar effects were observed using the structurally distinct LSD1 inhibitor SP2509 (Supplementary Fig. 1CG).

Figure 1

Systemic LSD1 inhibition prevents the development of hyperglycemia and improves insulin sensitivity in db/db mice. A: Four-week-old male and female db/db mice received daily intraperitoneal injections of GSK-LSD1 or veh for 6 weeks. As a control, lean db/+ mice were injected with veh. Metabolic measurements were conducted at the indicated time points. Body weight (B) and blood glucose levels (C) were measured weekly (n = 10 mice/group). D: Blood glucose levels at indicated time points after a glucose bolus via oral gavage (n = 10 mice/group). E: Fasting plasma insulin levels at baseline and after 1, 3, and 6 weeks of GSK-LSD1 or veh treatment (n = 6–10 mice/group). F: Plasma insulin levels before and 10 min after glucose gavage (n = 8–10 mice/group). G: Blood glucose levels at indicated time points after insulin injection (2.0 units/kg body wt i.p.) (n = 10 mice/group). H: Blood glucose levels at indicated time points after intraperitoneal pyruvate injection (n = 10 mice/group). I: Immunoblot analysis of p-AktSer473, Akt, and vinculin in gWAT of GSK-LSD1– or veh-treated db/db mice injected with insulin or saline. J: Quantification of p-AktSer473–to–Akt ratio as fold change compared with veh-treated mice without insulin injection (n = 4 mice/group). Data are presented as mean ± SEM. Statistical differences were calculated with two-way ANOVA with Tukey post hoc analysis (in BD, FH, and J). One-way ANOVA with Tukey post hoc analysis was performed to analyze statistical differences between three or more groups (E). Unless otherwise indicated, significance is shown between GSK-LSD1– and veh-treated mice. *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05; ##P < 0.01; ###P < 0.001. AU, arbitrary units; GTT, glucose tolerance test; ITT, insulin tolerance test; ns, not significant; PTT, pyruvate tolerance test; wks, weeks.

Figure 1

Systemic LSD1 inhibition prevents the development of hyperglycemia and improves insulin sensitivity in db/db mice. A: Four-week-old male and female db/db mice received daily intraperitoneal injections of GSK-LSD1 or veh for 6 weeks. As a control, lean db/+ mice were injected with veh. Metabolic measurements were conducted at the indicated time points. Body weight (B) and blood glucose levels (C) were measured weekly (n = 10 mice/group). D: Blood glucose levels at indicated time points after a glucose bolus via oral gavage (n = 10 mice/group). E: Fasting plasma insulin levels at baseline and after 1, 3, and 6 weeks of GSK-LSD1 or veh treatment (n = 6–10 mice/group). F: Plasma insulin levels before and 10 min after glucose gavage (n = 8–10 mice/group). G: Blood glucose levels at indicated time points after insulin injection (2.0 units/kg body wt i.p.) (n = 10 mice/group). H: Blood glucose levels at indicated time points after intraperitoneal pyruvate injection (n = 10 mice/group). I: Immunoblot analysis of p-AktSer473, Akt, and vinculin in gWAT of GSK-LSD1– or veh-treated db/db mice injected with insulin or saline. J: Quantification of p-AktSer473–to–Akt ratio as fold change compared with veh-treated mice without insulin injection (n = 4 mice/group). Data are presented as mean ± SEM. Statistical differences were calculated with two-way ANOVA with Tukey post hoc analysis (in BD, FH, and J). One-way ANOVA with Tukey post hoc analysis was performed to analyze statistical differences between three or more groups (E). Unless otherwise indicated, significance is shown between GSK-LSD1– and veh-treated mice. *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05; ##P < 0.01; ###P < 0.001. AU, arbitrary units; GTT, glucose tolerance test; ITT, insulin tolerance test; ns, not significant; PTT, pyruvate tolerance test; wks, weeks.

Close modal

To identify mechanisms leading to improved glycemic control after GSK-LSD1 treatment, we measured plasma insulin levels and performed insulin and pyruvate tolerance tests in veh-administered and GSK-LSD1–treated db/db mice. Fasting plasma insulin levels were significantly reduced after 3 and 6 weeks of GSK-LSD1 treatment (Fig. 1E). Moreover, GSK-LSD1–treated mice showed a twofold increase in plasma insulin levels after glucose stimulation compared with baseline, whereas glucose stimulation did not significantly increase insulin levels in the veh group (Fig. 1F), indicating improved coupling of insulin secretion with blood glucose following LSD1 inhibition. Insulin tolerance tests revealed increased insulin sensitivity in the GSK-LSD1 treatment group compared with veh-administered db/db mice (Fig. 1G). Finally, db/db mice that received GSK-LSD1 exhibited lower glucose excursions following pyruvate challenge compared with veh-administered mice (Fig. 1H), suggesting that systemic LSD1 inhibition also reduces hepatic glucose production.

Insulin resistance in peripheral insulin target tissues is a major characteristic of obesity (6,41). We therefore assessed signal transduction downstream of the insulin receptor by measuring p-AktSer473 in WAT, liver, and skeletal muscle following insulin injection of db/db mice. Insulin-stimulated p-AktSer473 was higher in gonadial WAT (gWAT) of the GSK-LSD1 group compared with the veh group (Fig. 1I and J). By contrast, LSD1 inhibition had little effect on p-AktSer473 in skeletal muscle and liver (Supplementary Fig. 2). These results suggest that systemic LSD1 inhibition improves glucose homeostasis of db/db mice in part through enhanced insulin sensitivity of adipose tissue.

LSD1 Inhibition Reduces Adipose Inflammation and Lipolysis

To examine effects of GSK-LSD1 on adipose tissue, we first measured adipose tissue weight and adipocyte size in gWAT, sWAT, and brown adipose tissue. The tissue weights of gWAT and sWAT were significantly reduced in the GSK-LSD1–treated group, indicating a loss of fat mass following LSD1 inhibition (Supplementary Fig. 3AC). The fat mass loss was not a result of reduced adipocyte size (Fig. 2A and Supplementary Fig. 3DH).

Figure 2

LSD1 inhibition reduces adipose tissue inflammation and lipolysis in db/db mice. Representative images of gWAT sections stained with H-E (A) and detection of CLS using an F4/80 antibody in gWAT (B) after 6 weeks of daily GSK-LSD1 or veh administration to male and female db/db mice. As a control, lean db/+ mice were injected with veh. Red arrows highlight CLS. Scale bars = 100 μm. C: Quantification of F4/80+ CLS in gWAT relative to tissue size (n = 4 mice/group). D: Quantitative PCR analysis of inflammatory genes in gWAT. Transcript levels in GSK-LSD1–treated relative to veh-treated db/db mice (Rel. to veh.) (n = 6–11 mice/group). E and F: Lipolysis in differentiated adipocytes isolated from db/db mice after preincubation with GSK-LSD1 or veh and stimulation with isoproterenol. FFA release (E) and glycerol release (F) (n = 3 mice). G: Plasma NEFA levels (n = 6–10 mice/group). H: Networks of genes upregulated (network nodes in red) or downregulated (network nodes in blue) by GSK-LSD1 compared with vehicle (P < 0.01 by Cuffdiff) in gWAT from db/db mice following 6 weeks of treatment, shown as clustered functional categories (n = 4 mice/group). Data are presented as mean ± SEM. One-way ANOVA with Tukey post hoc analysis was performed to analyze statistical differences between three or more groups (C and G) or multiple unpaired t tests were performed to determine differences between two groups (DF). *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05.

Figure 2

LSD1 inhibition reduces adipose tissue inflammation and lipolysis in db/db mice. Representative images of gWAT sections stained with H-E (A) and detection of CLS using an F4/80 antibody in gWAT (B) after 6 weeks of daily GSK-LSD1 or veh administration to male and female db/db mice. As a control, lean db/+ mice were injected with veh. Red arrows highlight CLS. Scale bars = 100 μm. C: Quantification of F4/80+ CLS in gWAT relative to tissue size (n = 4 mice/group). D: Quantitative PCR analysis of inflammatory genes in gWAT. Transcript levels in GSK-LSD1–treated relative to veh-treated db/db mice (Rel. to veh.) (n = 6–11 mice/group). E and F: Lipolysis in differentiated adipocytes isolated from db/db mice after preincubation with GSK-LSD1 or veh and stimulation with isoproterenol. FFA release (E) and glycerol release (F) (n = 3 mice). G: Plasma NEFA levels (n = 6–10 mice/group). H: Networks of genes upregulated (network nodes in red) or downregulated (network nodes in blue) by GSK-LSD1 compared with vehicle (P < 0.01 by Cuffdiff) in gWAT from db/db mice following 6 weeks of treatment, shown as clustered functional categories (n = 4 mice/group). Data are presented as mean ± SEM. One-way ANOVA with Tukey post hoc analysis was performed to analyze statistical differences between three or more groups (C and G) or multiple unpaired t tests were performed to determine differences between two groups (DF). *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05.

Close modal

A key mechanism leading to impaired insulin sensitivity in WAT is local inflammation (1,3). WAT inflammation is characterized by infiltration of macrophages, which form CLS around dead adipocytes (4,42) and produce chemokines and cytokines that promote adipose tissue insulin resistance (3,4). Histological evaluation revealed abundant CLS in gWAT from veh-administered db/db mice more so than in sWAT and brown adipose tissue (Fig. 2A and Supplementary Fig. 3D and E). CLS were rare in gWAT from GSK-LSD1–treated db/db mice (Fig. 2A), suggesting that GSK-LSD1 reduces macrophage accumulation in gWAT. Quantification of macrophages based on staining for the macrophage-specific epitope F4/80 confirmed reduced macrophage numbers in gWAT after GSK-LSD1 treatment (Fig. 2B and C). In line with this result, gene expression studies revealed downregulation of inflammatory genes, such as interleukin 1β (Il1β) and interleukin 6 (Il6) in gWAT (Fig. 2D). These observations indicate that LSD1 inhibition decreases adipose inflammation in obesity. Considering the association between inflammation and adipose tissue insulin resistance (3), the reduction of adipose inflammation by GSK-LSD1 could contribute to the improvement in adipose insulin sensitivity (Fig. 1H and I). To determine whether the LSD1 inhibitor has direct effects on insulin sensitivity at the level of adipocytes, we measured insulin-stimulated p-AktSer473 following GSK-LSD1 treatment in differentiated adipocytes of db/db mice (Supplementary Fig. 3IK). In contrast to our observations in vivo, GSK-LSD1 did not alter insulin signaling in vitro in adipocytes (Supplementary Fig. 3J and K). These data suggest that LSD1 inhibition improves adipose tissue insulin sensitivity indirectly.

Lipolysis is tightly regulated by circulating insulin. WAT insulin resistance is associated with augmented hydrolysis of triglycerides, resulting in increased levels of circulating NEFA (41,43). To test whether GSK-LSD1 could exert adipocyte-autonomous effects on lipolysis, we preincubated differentiated adipocytes from db/db mice with GSK-LSD1 or veh and then treated with isoproterenol, which induces lipolysis by activating β-adrenergic signaling. GSK-LSD1 reduced isoproterenol-stimulated FFA release concomitant with a trend toward reduced glycerol release (Fig. 2E and F). Moreover, GSK-LSD1 reduced plasma NEFA levels in db/db mice, suggesting that systemic LSD1 inhibition reduces WAT lipolysis in vivo (Fig. 2G). Importantly, GSK-LSD1 did not change WAT norepinephrine levels, suggesting that the beneficial effects of LSD1 inhibition on adipose tissue lipolysis occur independent of changes in sympathetic tone (Supplementary Fig. 3L). Together, these observations indicate that LSD1 inhibition corrects several WAT defects associated with obesity including insulin resistance, inflammation, and augmented lipolysis.

To identify potential mechanisms whereby GSK-LSD1 inhibits lipolysis and adipose inflammation, we performed mRNA-seq of gWAT from GSK-LSD1–treated or veh-administered db/db mice as well as veh-treated db/+ mice as a lean control. Pairwise comparisons between groups yielded a total of 2,269 genes that were differentially expressed for at least one comparison (P < 0.01 by Cuffdiff) (Supplementary Table 2AC). Obesity exerted a more prominent effect than LSD1 inhibition, as reflected by 1,777 genes being differentially expressed between db/db and db/+ gWAT compared with 731 genes being altered by GSK-LSD1 treatment in db/db mice. K-means clustering of all differentially expressed genes identified several clusters of genes strongly regulated by obesity whose obesity-associated changes were partially reversed by systemic LSD1 inhibition (Supplementary Fig. 4A). These findings led us to predict that GSK-LSD1 improves adipose tissue function in db/db mice by reversing maladaptive transcriptional changes in obesity. LSD1 inhibition resulted in 359 upregulated and 372 downregulated genes (Supplementary Fig. 4B), of which nearly all were changed in the opposite direction by obesity (Supplementary Fig. 4A). GO analysis identified enrichment of inflammation-associated functional categories among genes downregulated by GSK-LSD1 (Fig. 2H and Supplementary Table 2D and E), which included cell surface receptors (e.g., Ccr5, Cd14, and Tlr4) involved in recruitment and/or activation of immune cells as well as transcription factors downstream of inflammatory signaling pathways (e.g., Fos). GSK-LSD1–downregulated genes also included genes with indirect roles in inflammation whose inactivation has been shown to improve adipose tissue function in obesity. These include enzymes involved in extracellular matrix remodeling (Mmp14) (44), activation of lipogenic gene expression (Cyp1b1) (45), and repression of insulin signaling (Ptprj) (46). As inflammation has been shown to promote lipolysis (47,48), we predicted that reduced inflammation could underlie effects of GSK-LSD1 on lipolysis. These transcriptional changes could reflect differences in immune cell recruitment (Fig. 2H) or, alternatively, cell-autonomous roles of LSD1 in adipocytes or the immune cells themselves. To test the possibility that LSD1 directly regulates these genes in adipocytes, we treated adipocytes differentiated from db/db mice with GSK-LSD1. GSK-LSD1 did not repress these inflammation genes ex vivo (Supplementary Fig. 4C), suggesting the in vivo effect of GSK-LSD1 on these genes is indirect. This result also indicates that downregulation of these inflammation genes can be uncoupled from the effect of GSK-LSD1 on lipolysis, which is intact in this ex vivo experiment (Fig. 2E). As we also observed no ex vivo effect of GSK-LSD1 on insulin signal transduction (Supplementary Fig. 3J and K), these findings collectively suggest that LSD1 regulates lipolysis independent of changes in insulin signaling and repression of inflammation genes.

To narrow down the step of the lipolysis pathway that is affected by LSD1 inhibition, we tested whether GSK-LSD1 interferes with the canonical phosphorylation cascade that activates lipolysis. To this end, we assessed phosphorylation of hormone-sensitive lipase (Hsl), which is one of the distal lipolytic enzymes activated downstream of the β-adrenergic receptor (49). Robust Hsl phosphorylation in response to isoproterenol in GSK-LSD1–treated adipocytes (Supplementary Fig. 4D and E) indicates that LSD1 inhibition does not interfere with—and in fact hyperactivates—the phosphorylation cascade downstream of β-adrenergic signaling. Overall, these results suggest that the effect of LSD1 inhibition on lipolysis likely involves distal effects at the level of lipolytic enzymes or the lipid droplet. In support of this model, examination of lipolysis and lipid droplet genes in mRNA-seq data revealed upregulation of two genes encoding lipolysis inhibitors, G0s2 (50) and Plin5 (51,52), in gWAT of GSK-LSD1–treated db/db mice (Supplementary Table 2A). These genes are strong candidates for future mechanistic studies of GSK-LSD1’s effect on lipolysis.

LSD1 Inhibition Ameliorates Liver Steatosis in db/db Mice

Obesity and insulin resistance are closely associated with NAFLD (68). Having observed that GSK-LSD1 prevents weight gain, increases insulin sensitivity, and reduces FFA release in db/db mice, we asked whether GSK-LSD1 ameliorates obesity-associated changes in the liver. Livers from veh-treated db/db mice showed clear morphological and histological signs of steatosis (Fig. 3A and B). In contrast, livers from db/db mice treated with GSK-LSD1 were grossly and histologically normal, resembling livers from lean db/+ mice. In addition, H-E staining of liver sections from GSK-LSD1–treated mice showed a reduction in microvesicular steatosis compared with veh-treated db/db mice (Supplementary Fig. 5A). In line with these findings, GSK-LSD1 significantly reduced liver weight in db/db mice (Fig. 3C). Steatosis results from a combination of increased lipid delivery from the circulation and increased lipogenesis by hepatocytes, leading to excess intrahepatic accumulation of lipids (7,8). LSD1 inhibition in db/db mice decreased both circulating and hepatic triglycerides to levels comparable with those of lean db/+ mice (Fig. 3D and E). Of note, plasma and hepatic cholesterol levels as well as circulating levels of the ketone body β-hydroxybutyrate were unaffected by GSK-LSD1 administration (Supplementary Fig. 5BD). Effects of GSK-LSD1 on liver morphology and circulating lipid levels in db/db mice were recapitulated by SP2509 (Supplementary Fig. 5EG). To determine whether GSK-LSD1 improves clinical biomarkers of NAFLD, we measured plasma levels of AST and ALT. Treatment of db/db mice with GSK-LSD1 led to a significant reduction in AST activity and a slight, albeit not significant, decrease in ALT activity (Fig. 3F and G). These results indicate that systemic LSD1 inhibition prevents liver pathology caused by obesity. In line with the GSK-LSD1–mediated reduction of steatosis, lipogenic genes such as elongation of very long chain fatty acids protein 6 (Elovl6), fatty acid synthase (Fasn), and stearyl-CoA desasturase (Scd1) were downregulated by GSK-LSD1 treatment (Fig. 3H). LSD1 inhibition further reduced expression of pyruvate kinase isoenzyme R/L (Pklr), which catalyzes the last step in the glycolytic pathway and is required for de novo lipogenesis from glucose. Moreover, lower expression of fatty acid binding protein 1 (Fabp1) and cluster of differentiation 36 (Cd36) indicates a possible effect of GSK-LSD1 on liver fatty acid uptake in db/db mice. Liver inflammatory markers Il1β, Il6, Il10, and tumor necrosis factor α (Tnfα) did not significantly differ between treatment groups (Supplementary Fig. 5H). Taken together, these observations suggest that systemic LSD1 inhibition impacts both lipid partitioning and hepatocyte metabolic pathways to ameliorate liver steatosis in obesity.

Figure 3

LSD1 inhibition protects against liver steatosis in db/db mice. A: Representative images of livers in db/db mice treated daily with GSK-LSD1 or veh for 6 weeks. As a control, lean db/+ mice were injected with veh (n = 13–14 mice/group). B: H-E stain of liver sections. Scale bars = 100 μm. C: Liver weight (n = 13–14 mice/group). D: Fasting plasma triglyceride levels at indicated time points (n = 9–10 mice/group). E: Hepatic triglyceride levels (n = 4 mice/group). AST (F) and ALT (G) activity in plasma of GSK-LSD1– or veh-treated mice for 6 weeks (n = 8–10 mice/group). H: Quantitative PCR analysis of genes associated with hepatic lipid metabolism in liver. Transcript levels relative to (Rel.) db/+ mice. (n = 3–4 mice/group). Data are presented as mean ± SEM. Statistical differences were calculated with one-way ANOVA (C and EH) or two-way ANOVA (D) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant; wks, weeks.

Figure 3

LSD1 inhibition protects against liver steatosis in db/db mice. A: Representative images of livers in db/db mice treated daily with GSK-LSD1 or veh for 6 weeks. As a control, lean db/+ mice were injected with veh (n = 13–14 mice/group). B: H-E stain of liver sections. Scale bars = 100 μm. C: Liver weight (n = 13–14 mice/group). D: Fasting plasma triglyceride levels at indicated time points (n = 9–10 mice/group). E: Hepatic triglyceride levels (n = 4 mice/group). AST (F) and ALT (G) activity in plasma of GSK-LSD1– or veh-treated mice for 6 weeks (n = 8–10 mice/group). H: Quantitative PCR analysis of genes associated with hepatic lipid metabolism in liver. Transcript levels relative to (Rel.) db/+ mice. (n = 3–4 mice/group). Data are presented as mean ± SEM. Statistical differences were calculated with one-way ANOVA (C and EH) or two-way ANOVA (D) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant; wks, weeks.

Close modal

Hepatocyte-Specific Lsd1 Deletion Does Not Improve Glycemia or Liver Steatosis in db/db Mice

We reasoned that a reduction in steatosis following GSK-LSD1 treatment could result from altered lipid partitioning from storage tissues or from cell-autonomous effects of LSD1 in hepatocytes. While reduced circulating NEFA and triglycerides following LSD1 inhibition supports a potential effect on lipid partitioning (Figs. 2G and 3D), LSD1 is expressed in hepatocytes (Supplementary Fig. 6A) and could have hepatocyte-autonomous effects on lipogenesis or lipid uptake. To analyze hepatocyte-specific functions of LSD1, we deleted Lsd1 in hepatocytes of 6-week-old Lsd1fl/fldb/db mice using AAV8-TBG-iCre (referred to as Lsd1ΔLdb/db mice hereafter) (Supplementary Fig. 6B and C). In contrast to systemic LSD1 inhibition, liver-specific Lsd1 deletion did not alter body weight or blood glucose levels (Supplementary Fig. 6D and E). Furthermore, glucose tolerance, plasma insulin levels, and insulin sensitivity were similar in Lsd1ΔLdb/db and control db/db mice (Supplementary Fig. 6FH). Accordingly, hepatocyte-specific Lsd1 deletion did not improve pyruvate tolerance, reflecting unaltered hepatic gluconeogenesis (Supplementary Fig. 6I). Further supporting this conclusion, LSD1 inhibition in primary hepatocytes isolated from db/db mice did not alter glucose production from lactate and pyruvate (Supplementary Fig. 6J). Together, these findings suggest that improvements in glucose homeostasis following systemic LSD1 inhibition are not mediated by a direct effect on hepatocytes.

We further examined the effects of hepatocyte-specific Lsd1 deletion on liver steatosis in db/db mice. Lsd1ΔLdb/db mice exhibited liver histology, liver weight, plasma triglycerides, ALT, and AST levels similar to those of db/db control mice (Supplementary Fig. 6KO). However, we observed a reduction in hepatic triglyceride levels in Lsd1ΔLdb/db mice (Supplementary Fig. 6P), suggesting a hepatocyte-autonomous role of LSD1 in the regulation of hepatic lipid metabolism. Despite the observed effect of hepatic Lsd1 deletion on hepatic triglycerides, our findings support the overall conclusion that the beneficial effects of systemic LSD1 inhibition on glucose homeostasis and liver steatosis are not a direct result of LSD1 inhibition in hepatocytes.

Food Intake Is Reduced by Systemic LSD1 Inhibition

The improvements in glucose homeostasis and liver health following systemic LSD1 inhibition coincided with reductions in weight gain (Fig. 1B). Changes in body weight typically result from imbalanced caloric intake and energy expenditure. To investigate whether GSK-LSD1–mediated effects on weight gain are caused by decreased food consumption or increased energy expenditure, GSK-LSD1–treated or veh-administered db/db mice were placed in metabolic cages and monitored over the course of 48 h. This experiment revealed that food intake was reduced by GSK-LSD1 treatment (Fig. 4A and Supplementary Fig. 7A). In contrast, GSK-LSD1 had no significant impact on energy expenditure as measured by oxygen consumption and carbon dioxide production (Fig. 4B and C), resulting in a similar respiratory quotient between GSK-LSD1-treated and veh-administered mice (Supplementary Fig. 7B). We also found no evidence for changes in animal activity levels (Supplementary Fig. 7C). Fluid intake was reduced by GSK-LSD1 (Supplementary Fig. 7D and E), which is likely a secondary effect of normalized glucose levels. Importantly, long-term measurements of food intake revealed no difference between GSK-LSD1–treated or veh-administered lean db/+ mice (Supplementary Fig. 7F). Overall, these findings suggest that systemic LSD1 inhibition reduces weight gain in db/db mice in part by attenuating hyperphagia.

Figure 4

Food intake is reduced in GSK-LSD1–treated db/db mice. A: Male and female mice were injected daily with GSK-LSD1 or vehicle (veh) for 5 weeks and then placed into metabolic cages. With use of CLAMS, food intake was monitored over 48 h and is shown as food consumed per day (n = 4–5 mice/group/day). Oxygen consumption (VO2) (B) and carbon dioxide production (VCO2) (C) (n = 4–5 mice/group). Data presented as mean ± SEM. Statistical differences between two groups were calculated with an unpaired two-tailed Student t test (A) or two-way ANOVA (B and C) with Tukey post hoc analysis. *P < 0.05.

Figure 4

Food intake is reduced in GSK-LSD1–treated db/db mice. A: Male and female mice were injected daily with GSK-LSD1 or vehicle (veh) for 5 weeks and then placed into metabolic cages. With use of CLAMS, food intake was monitored over 48 h and is shown as food consumed per day (n = 4–5 mice/group/day). Oxygen consumption (VO2) (B) and carbon dioxide production (VCO2) (C) (n = 4–5 mice/group). Data presented as mean ± SEM. Statistical differences between two groups were calculated with an unpaired two-tailed Student t test (A) or two-way ANOVA (B and C) with Tukey post hoc analysis. *P < 0.05.

Close modal

Metabolic Effects of Systemic LSD1 Inhibition Are Independent of Food Intake

Pair feeding of db/db mice with lean controls has been shown to prevent obesity and hyperglycemia (53). Having observed reduced food intake and weight gain in GSK-LSD1–treated mice, we reasoned that some of the metabolic benefits of systemic LSD1 inhibition could result from its effect on feeding. Therefore, we conducted a pair-feeding experiment in which food consumption in GSK-LSD1–treated db/db mice was measured daily and the same amount was then provided to a second group of db/db mice receiving veh, with a third group of veh-administered ad libitum–fed db/db mice serving as a control (Fig. 5A). GSK-LSD1–treated mice consumed significantly less food over the course of the study compared with veh-administered ad libitum–fed mice, whereas the veh-administered pair-fed group consumed the same amount of food as the GSK-LSD1–treated group as defined per experimental design (Fig. 5B and Supplementary Fig. 8A). Body weight gain did not differ between the three groups over the course of the intervention (Supplementary Fig. 8B). Blood glucose levels in GSK-LSD1–treated db/db mice were significantly lower than in veh-administered pair-fed db/db mice (Fig. 5C). Likewise, GSK-LSD1–treated db/db mice showed improved glucose tolerance relative to veh-administered pair-fed db/db mice (Fig. 5D), indicating that GSK-LSD1 prevents hyperglycemia in db/db mice independent of food intake. Accordingly, fasting plasma insulin levels remained high in pair-fed compared with ad libitum–fed mice but were significantly lower in the GSK-LSD1–treated group (Fig. 5E). Together, these observations suggest that GSK-LSD1 improves insulin sensitivity in db/db mice independent of its effect on food intake.

Figure 5

GSK-LSD1–mediated improvement of metabolic dysfunction is independent of reduced food intake. A: Study design of pair-feeding experiment. Four-week-old male and female db/db mice were injected daily with veh or GSK-LSD1 for 6 weeks and fed a normal chow diet ad libitum. A third group of mice received veh and was pair-fed to GSK-LSD1–treated mice. B: Food consumption was monitored daily and is shown as cumulative food intake over the 6-week study (n = 8–11 mice/group). C: Blood glucose levels measured weekly. Asterisks indicate statistical differences between veh (black) and GSK-LSD1 (red) groups. D: Blood glucose levels at indicated time points after a glucose bolus via oral gavage (n = 8–11 mice/group). E: Fasting plasma insulin levels at the indicated time points (n = 8–11 mice/group). F: Fasting plasma triglycerides levels at indicated time points (n = 8–11 mice/group). G: Representative images of liver sections stained with H-E. Scale bars = 100 μm (n = 4 mice/group). H: Liver weight (n = 8–11 mice/group). I: Representative images of gWAT sections stained against F4/80 for detection of CLS. Red arrows indicate CLS. Scale bars = 100 μm. J: Quantification of F4/80+ CLS in gWAT relative to tissue size (n = 8–9 mice/group). Data presented as mean ± SEM. Statistical differences were calculated with one-way ANOVA (D, F, H, and J) or two-way ANOVA (B, C, and E) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001; ###P < 0.001. GTT, glucose tolerance test; wks, weeks.

Figure 5

GSK-LSD1–mediated improvement of metabolic dysfunction is independent of reduced food intake. A: Study design of pair-feeding experiment. Four-week-old male and female db/db mice were injected daily with veh or GSK-LSD1 for 6 weeks and fed a normal chow diet ad libitum. A third group of mice received veh and was pair-fed to GSK-LSD1–treated mice. B: Food consumption was monitored daily and is shown as cumulative food intake over the 6-week study (n = 8–11 mice/group). C: Blood glucose levels measured weekly. Asterisks indicate statistical differences between veh (black) and GSK-LSD1 (red) groups. D: Blood glucose levels at indicated time points after a glucose bolus via oral gavage (n = 8–11 mice/group). E: Fasting plasma insulin levels at the indicated time points (n = 8–11 mice/group). F: Fasting plasma triglycerides levels at indicated time points (n = 8–11 mice/group). G: Representative images of liver sections stained with H-E. Scale bars = 100 μm (n = 4 mice/group). H: Liver weight (n = 8–11 mice/group). I: Representative images of gWAT sections stained against F4/80 for detection of CLS. Red arrows indicate CLS. Scale bars = 100 μm. J: Quantification of F4/80+ CLS in gWAT relative to tissue size (n = 8–9 mice/group). Data presented as mean ± SEM. Statistical differences were calculated with one-way ANOVA (D, F, H, and J) or two-way ANOVA (B, C, and E) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001; ###P < 0.001. GTT, glucose tolerance test; wks, weeks.

Close modal

The feeding-independent effects of GSK-LSD1 on glucose homeostasis suggested that systemic LSD1 inhibition could ameliorate complications of obesity independent of altered food intake. Indeed, LSD1 inhibition reduced plasma triglycerides, improved liver steatosis, and reduced liver weight and plasma ALT activity in comparisons with veh-administered pair-fed db/db mice (Fig. 5F–H and Supplementary Fig. 8C), indicating that these effects are independent of reduced feeding. Similarly, the effect of LSD1 inhibition on adipose tissue inflammation was independent of reduced food intake, as evidenced by reduced abundance of F4/80+ CLS in comparisons with veh-treated pair-fed db/db mice (Fig. 5I and J). Overall, these experiments show that beneficial effects of GSK-LSD1 on glucose homeostasis, liver health, and adipose inflammation occur independent of its hyperphagia-reducing effect.

GSK-LSD1 Prevents Obesity and Improves Insulin Signaling in Western Diet–Fed Mice

db/db mice are deficient for signaling by the satiety hormone leptin and therefore do not fully recapitulate human obesity, in which leptin function is intact. Therefore, we determined whether effects of GSK-LSD1 can be observed in mice that develop obesity following consumption of a high-energy Western diet (54). C57BL/6J mice were fed a Western diet (42% kcal from fat, 34% sucrose by weight) for 11 weeks concomitant with daily injections of GSK-LSD1 or veh, with veh-treated C57BL/6J mice fed a normal chow diet serving as a control (Fig. 6A). Western diet feeding resulted in substantial weight gain of veh-treated mice, while GSK-LSD1-treated mice fed the Western diet remained lean, exhibiting body weights similar to those of normal chow-fed mice (Fig. 6B). Notably, GSK-LSD1 also prevented obesity in response to consumption of an HFD (60% kcal from fat) (Supplementary Fig. 9A). While blood glucose levels did not change in response to either diet (Fig. 6C and Supplementary Fig. 9B), GSK-LSD1 improved glucose tolerance in mice fed an HFD but not a chow diet (Supplementary Fig. 9C). Plasma insulin levels increased in response to Western diet feeding of veh-administered mice but remained low in the GSK-LSD1 group (Fig. 6D). The reduction in insulin levels without corresponding increases of blood glucose in GSK-LSD1–treated mice suggests that LSD1 inhibition improves insulin sensitivity during Western diet feeding, as we observed in db/db mice.

Figure 6

GSK-LSD1 reverses metabolic dysfunction due to diet-induced obesity. A: Ten-week-old male C57BL/6J wild-type mice fed a Western diet (WD) were injected daily with GSK-LSD1 or veh for 11 weeks. As a control, a group of WT mice received veh and was kept on a normal chow diet (NCD). Body weight (B) and blood glucose levels (C) measured weekly (n = 7–8 mice/group). Asterisks indicate statistical differences between GSK-LSD1– and veh-treated mice on WD unless otherwise stated. D: Fasting plasma insulin levels at indicated time points (n = 6–8 mice/group). E: Overview of intervention treatment protocol. After 11 weeks of veh administration, the veh group was split into one group continuing veh administration (black), whereas the other half began to receive GSK-LSD1 daily for another 7 weeks (blue). Likewise, the GSK-LSD1 group was split into one group continuing GSK-LSD1 administration after 11 weeks (red), whereas the other half began to receive veh daily for another 7 weeks (brown). F: Body weight across all treatment groups (Western diet, n = 4 mice/group for week 11–16, n = 2–4 mice/group, for weeks 17 and 18, and n = 7 mice for normal chow diet group). Asterisks indicate statistical differences between WD-fed mice administered veh (black) and the intervention group (blue). G: Fasting plasma insulin levels 2 weeks after drug intervention (normal chow diet veh, n = 7 mice, and Western diet, n = 4 mice). H: Blood glucose levels at indicated time points after insulin injection (0.8 units/kg body wt i.p.) at week 17. Data shown relative (rel.) to time point 0 min (n = 2–4 mice/group). I: Body weight (Western diet, n = 4 mice/group for week 11–16, n = 2–4 mice/group for weeks 17 and 18). Data are shown as mean ± SEM. Statistical differences were calculated with one-way ANOVA (G and J) or two-way ANOVA (BD, F, H, and I) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05; ##P < 0.01; ###P < 0.001. ITT, insulin tolerance test; ns, not significant; wks, weeks.

Figure 6

GSK-LSD1 reverses metabolic dysfunction due to diet-induced obesity. A: Ten-week-old male C57BL/6J wild-type mice fed a Western diet (WD) were injected daily with GSK-LSD1 or veh for 11 weeks. As a control, a group of WT mice received veh and was kept on a normal chow diet (NCD). Body weight (B) and blood glucose levels (C) measured weekly (n = 7–8 mice/group). Asterisks indicate statistical differences between GSK-LSD1– and veh-treated mice on WD unless otherwise stated. D: Fasting plasma insulin levels at indicated time points (n = 6–8 mice/group). E: Overview of intervention treatment protocol. After 11 weeks of veh administration, the veh group was split into one group continuing veh administration (black), whereas the other half began to receive GSK-LSD1 daily for another 7 weeks (blue). Likewise, the GSK-LSD1 group was split into one group continuing GSK-LSD1 administration after 11 weeks (red), whereas the other half began to receive veh daily for another 7 weeks (brown). F: Body weight across all treatment groups (Western diet, n = 4 mice/group for week 11–16, n = 2–4 mice/group, for weeks 17 and 18, and n = 7 mice for normal chow diet group). Asterisks indicate statistical differences between WD-fed mice administered veh (black) and the intervention group (blue). G: Fasting plasma insulin levels 2 weeks after drug intervention (normal chow diet veh, n = 7 mice, and Western diet, n = 4 mice). H: Blood glucose levels at indicated time points after insulin injection (0.8 units/kg body wt i.p.) at week 17. Data shown relative (rel.) to time point 0 min (n = 2–4 mice/group). I: Body weight (Western diet, n = 4 mice/group for week 11–16, n = 2–4 mice/group for weeks 17 and 18). Data are shown as mean ± SEM. Statistical differences were calculated with one-way ANOVA (G and J) or two-way ANOVA (BD, F, H, and I) with Tukey post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001; #P < 0.05; ##P < 0.01; ###P < 0.001. ITT, insulin tolerance test; ns, not significant; wks, weeks.

Close modal

The above findings collectively demonstrate that GSK-LSD1 prevents metabolic defects in several models of obesity. However, an ideal therapy for metabolic disease would reverse preexisting defects. Therefore, we performed a drug intervention study to ask whether GSK-LSD1 improves metabolic health in obese mice (Fig. 6E). After 11 weeks of Western diet feeding and veh treatment, mice were split into two groups receiving either GSK-LSD1 for an additional 7 weeks or continuation of veh treatment. Remarkably, switching treatment groups from veh to GSK-LSD1 lowered body weight and plasma insulin levels significantly in Western diet–fed mice (Fig. 6F and G). As blood glucose levels did not change in any of the treatment groups (Supplementary Fig. 10A), this result suggests that GSK-LSD1 ameliorates insulin resistance in mice with preexisting obesity. Indeed, insulin tolerance tests revealed that insulin sensitivity was improved to a similar extent whether GSK-LSD1 was administered concomitant with Western diet feeding or initiated after the establishment of obesity (Fig. 6H). Altogether, these observations support the potential for systemic LSD1 inhibition as a therapy for obesity and T2D.

GSK-LSD1 is an irreversible inhibitor, which could limit the ability to reverse the effects of this drug following treatment withdrawal (e.g., when target weight is achieved in previously obese individuals). To test whether the effects of GSK-LSD1 on systemic metabolism are reversible, we performed a washout study in Western diet–fed mice treated with GSK-LSD1 for 11 weeks, at which point mice were switched to veh treatment (Fig. 6E). Western diet–fed mice that received GSK-LSD1 or veh for the duration of the study were used for comparison. The GSK-LSD1–mediated prevention of Western diet–induced body weight gain was completely reversible once the mice were taken off treatment, with body weights of the group taken off GSK-LSD1 converging with those of mice treated with veh for the duration of the study (Fig. 6I). Similarly, beneficial effects of GSK-LSD1 on liver, gWAT, and sWAT weight were reversed when Western diet–fed mice were taken off the drug (Supplementary Fig. 10BD). Overall, our findings indicate that systemic LSD1 inhibition prevents and corrects hallmark metabolic defects associated with obesity in a manner that is readily reversible on drug withdrawal.

Despite the growing obesity pandemic, there is an unmet need for efficient and safe therapies for obesity and its associated complications. Here, we report that systemic LSD1 inhibition in rodent models of obesity and T2D causes weight loss and ameliorates obesity-induced complications including hyperglycemia and NAFLD. LSD1 inhibition decreases food intake, body weight, and fat mass in diet-induced and genetic models of obesity. Moreover, LSD1 inhibitors prevent hyperglycemia and fatty liver disease in db/db mice. These effects are accompanied by improved adipose tissue function as shown by reduced tissue inflammation and increased insulin sensitivity. LSD1 inhibition does not affect metabolism in lean mice, indicating that LSD1 plays a context-specific role during overfeeding. This is of importance for clinical applications, as LSD1 inhibition does not cause anorexia or hypoglycemia, suggesting physiological set points are maintained during LSD1 inhibition. Overall, these findings identify LSD1 as a potential therapeutic target to promote weight loss and prevent T2D and NAFLD in people with obesity.

The beneficial effects of LSD1 inhibition on metabolism occurred independent of reduced food intake, indicating direct effects on tissues involved in nutrient metabolism. Our findings suggest that LSD1 inhibition in WAT contributes to the beneficial effect of GSK-LSD1 on obesity-associated metabolic complications. In support of this conclusion, improved systemic insulin sensitivity after GSK-LSD1 treatment was associated with improved insulin signaling in WAT but not in skeletal muscle or liver. In addition, GSK-LSD1 reversed several obesity-associated defects in adipose tissue that are known to contribute to metabolic syndrome, including inflammation and excessive lipolysis. LSD1 has been reported to have a cell-autonomous role in hepatocytes, promoting steatosis through transcriptional activation of lipogenic enzyme genes (29). Consistent with this finding, we observed lower hepatic triglyceride levels after hepatocyte-specific Lsd1 deletion. However, we found that hepatocyte-specific Lsd1 deletion did not recapitulate the beneficial effect of systemic GSK-LSD1 treatment on NAFLD or hepatic glucose production, suggesting that systemic LSD1 inhibition indirectly ameliorates hepatic steatosis and excessive gluconeogenesis.

It is well established that obesity-associated changes in adipose tissue can have pleiotropic effects on systemic metabolism (1,4,43,55). Excess caloric intake and the subsequent expansion of fat mass result in increased infiltration of macrophages and local secretion of proinflammatory cytokines, which precipitate metabolic defects in adipocytes that lead to increased circulating FFA (4,43). This sets in motion a vicious cycle of additional macrophage recruitment, further inflammation, lipolysis, and insulin resistance (1,5,6,43,5561). Our finding that GSK-LSD1 inhibits lipolysis in isolated adipocytes indicates effects of LSD1 on lipolysis independent of inflammation and insulin signaling, suggesting that a direct effect of LSD1 on lipolysis could initiate a chain of events leading to adipose tissue inflammation and insulin resistance in vivo. In support of this model, inhibition of lipolysis has been shown to improve insulin sensitivity, glucose tolerance, adipose tissue inflammation, and liver health during obesity (43,62). In addition to the GSK-LSD1–mediated reduction in lipolysis in isolated adipocytes, we also found decreased circulating FFA levels in vivo. It is known that fatty acids released from adipose promote hepatic glucose production and are the main substrate for hepatocyte triglyceride synthesis (5,59,6367). Consequently, lipolysis-derived FFA promote hepatic gluconeogenesis and ectopic lipid deposition in hepatocytes (43,62), providing a plausible link between lipolysis inhibition by GSK-LSD1 and indirect metabolic improvements in the liver. Our findings suggest that GSK-LSD1 inhibits lipolysis at a site distal to the phosphorylation cascade linking the β-adrenergic receptor to lipolytic enzymes. Examination of transcriptomes of gWAT from GSK-LSD1–treated mice revealed upregulation of two genes encoding lipid droplet proteins that inhibit the enzymatic activity of the key lipase Atgl. Further studies will be necessary to determine exactly how LSD1 regulates lipolysis as well as clarify the interrelationship of lipolysis inhibition, reduced adipose inflammation, and improved insulin signaling following systemic LSD1 inhibition. Inducible white adipocyte–specific Lsd1 deletion would facilitate such mechanistic studies and determine the relative contribution of adipocyte LSD1 to aberrant metabolic remodeling in obesity. Nevertheless, it remains possible that LSD1’s cell-autonomous functions in several cell types including but not limited to hepatocytes, adipocytes, and immune cells have additive effects in promoting obesity-associated defects.

Genetic studies in lean mice have revealed that a major function of LSD1 is to promote mitochondrial metabolism in adipocytes (33,68). Lsd1 deletion in white and brown fat via adiponectin-Cre leads to increased adipose tissue mass and weight gain (68). Moreover, this model develops exaggerated glucose intolerance during HFD feeding compared with Lsd1-intact controls (68). Our finding that pharmacological LSD1 inactivation reduces adipose tissue mass and improves glucose tolerance during obesity seemingly contrasts with the findings of these genetic studies. It is possible that pharmacological inhibition of LSD1’s enzymatic activity has an effect distinct from genetic deletion of Lsd1, which can impact other functions such as scaffolding of proteins within transcriptional complexes. Alternatively, the use of constitutive Cre recombinase to delete Lsd1 may have confounding effects on adipocyte development (33).

The selective effect of LSD1 inhibition during overfeeding suggests that LSD1 has context-dependent functions in the regulation of systemic metabolism. We found that GSK-LSD1 reduces food intake and body weight only in models of genetic or diet-induced obesity, with no effect on these parameters in lean, chow-fed mice. Similarly, LSD1 inhibition prevents hyperglycemia in db/db mice but is of little effect on glycemia or glucose tolerance in lean, nondiabetic mice. Obesity and T2D are known to disrupt homeostatic feedback mechanisms regulating appetite, weight gain, and blood glucose (13). Our observations open the possibility that LSD1 is involved in the maladaptive changes to systemic metabolism that result in higher defended body weight and blood glucose. Effects of LSD1 inhibition on food intake, adipose lipolysis, hepatic triglycerides, and gluconeogenesis suggest that LSD1 could evoke shared tissue-autonomous mechanisms for metabolic rewiring during obesity. While our pharmacological approach does not definitively implicate a single target tissue responsible for these effects, simultaneously targeting several relevant tissues with GSK-LSD1 treatment provides the advantage of testing the coordinated effects of LSD1. Future work should address whether and how obesity-associated changes to LSD1 and its associated epigenomic program occur in endocrine and neuronal systems involved in feeding behavior as well as in peripheral tissues directly involved in fuel metabolism.

One limitation of the study is that only male mice were analyzed in the Western diet study. Importantly, Western or HFD feeding evokes sex-dependent effects on metabolism, with female mice being less susceptible to Western or HFD-induced body weight gain, glucose intolerance, and insulin resistance (6972). We understand that sex differences are a major concern given that sex is a variable affecting a multitude of physiological and pathological processes (73); thus, future studies should include both sexes. However, no sex differences were observed in experiments performed with db/db mice in this study, suggesting that the effect of LSD1 inhibition in this model is similar in both males and females.

The development of weight loss therapies to treat obesity and its complications has proven to be a considerable challenge. None of the obesity drugs with long-standing U.S. Food and Drug Administration approval are widely used due to the risk of adverse events (74,75). The recently approved glucagon-like peptide 1 receptor agonist semaglutide holds promise for safe and effective obesity treatment in combination with lifestyle interventions (76). With the hope that broad adoption of semaglutide mirrors results of clinical trials indicating durable weight loss and reduced complications of obesity (76,77), the next generation of obesity drugs may well be tailored to patients at risk for specific complications such as NAFLD, for which there is currently no approved therapy beyond implementing lifestyle changes. Epigenomic regulation has recently emerged as a novel regulatory layer in energy homeostasis, opening a new avenue for potential treatment strategies for metabolic disease (78,79). Evidence that histone deacetylases (HDACs) play a role in glucose homeostasis by regulating the function of hepatocytes and insulin-producing pancreatic β-cells positioned these enzymes as potential drug targets in metabolic disease (78). Multiple HDAC inhibitors are currently being investigated for the treatment of obesity and T2D (78), but currently none have been approved for this clinical application. Here, we show that inhibition of the histone demethylase LSD1 has beneficial effects on glucose homeostasis and NAFLD independent of changes in food intake, suggesting that LSD1 could be a novel therapeutic target for correcting epigenomic defects in metabolic disease. The feasibility of LSD1 targeting in people with obesity will depend on a favorable safety profile. Our results describing metabolic benefits of GSK-LSD1 provide a strong rationale for clinical trials investigating the safety and efficacy of LSD1 inhibitors in obesity and NAFLD.

This article contains supplementary material online at https://doi.org/10.2337/figshare.21035914.

Acknowledgments. The authors thank members of the Sander laboratory and Drs. Martin Myers (University of Michigan), Alan Saltiel (UCSD), and Jerrold Olefsky (UCSD) for helpful discussions. The authors thank Fenfen Liu (UCSD), Vivian Lin (UCSD), and Johanna Fleischman (UCSD) for technical assistance.

Funding. This work was supported by Larry L. Hillblom Foundation fellowship 2021-D-008-FEL (B.R.), JDRF postdoctoral fellowship 3-PDF-2014-193-A-N and Diabetes Research Center Pilot and Feasibility grant P30 DK063491 (M.W.), John G. Davies Endowed Fellowship in Pancreatic Research S1079-1002614 and S1105-1002847-AWD (B.R. and M.W.), Foundation Leducq grant 16CVD01 (P.L.S.M.G.), National Institutes of Health grants R01 DK068471 and R01 DK078803 (M.S.), and UCSD School of Medicine Microscopy Core grant NINDS P30 NS047101.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. B.R. designed research studies, conducted experiments, acquired data, analyzed data, and wrote the manuscript. D.P.P., H.Z., C.N., A.R.H., and I.O. conducted experiments and acquired data. P.L.S.M. designed research studies. M.W. designed research studies, conducted experiments, acquired data, analyzed data, and wrote the manuscript. M.S. designed research studies, analyzed data, and wrote the manuscript. M.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Saltiel
AR
,
Olefsky
JM
.
Inflammatory mechanisms linking obesity and metabolic disease
.
J Clin Invest
2017
;
127
:
1
4
2.
Sarma
S
,
Sockalingam
S
,
Dash
S
.
Obesity as a multisystem disease: trends in obesity rates and obesity-related complications
.
Diabetes Obes Metab
2021
;
23
(
Suppl. 1
):
3
16
3.
Lee
YS
,
Wollam
J
,
Olefsky
JM
.
An integrated view of immunometabolism
.
Cell
2018
;
172
:
22
40
4.
Weisberg
SP
,
McCann
D
,
Desai
M
,
Rosenbaum
M
,
Leibel
RL
,
Ferrante
AW
Jr
.
Obesity is associated with macrophage accumulation in adipose tissue
.
J Clin Invest
2003
;
112
:
1796
1808
5.
Czech
MP
.
Insulin action and resistance in obesity and type 2 diabetes
.
Nat Med
2017
;
23
:
804
814
6.
Roden
M
,
Shulman
GI
.
The integrative biology of type 2 diabetes
.
Nature
2019
;
576
:
51
60
7.
Chen
Z
,
Yu
Y
,
Cai
J
,
Li
H
.
Emerging molecular targets for treatment of nonalcoholic fatty liver disease
.
Trends Endocrinol Metab
2019
;
30
:
903
914
8.
Friedman
SL
,
Neuschwander-Tetri
BA
,
Rinella
M
,
Sanyal
AJ
.
Mechanisms of NAFLD development and therapeutic strategies
.
Nat Med
2018
;
24
:
908
922
9.
Kahn
SE
.
The relative contributions of insulin resistance and beta-cell dysfunction to the pathophysiology of Type 2 diabetes
.
Diabetologia
2003
;
46
:
3
19
10.
Li
S
,
Brown
MS
,
Goldstein
JL
.
Bifurcation of insulin signaling pathway in rat liver: mTORC1 required for stimulation of lipogenesis, but not inhibition of gluconeogenesis
.
Proc Natl Acad Sci U S A
2010
;
107
:
3441
3446
11.
Mehran
AE
,
Templeman
NM
,
Brigidi
GS
, et al
.
Hyperinsulinemia drives diet-induced obesity independently of brain insulin production
.
Cell Metab
2012
;
16
:
723
737
12.
Priest
C
,
Tontonoz
P
.
Inter-organ cross-talk in metabolic syndrome
.
Nat Metab
2019
;
1
:
1177
1188
13.
Myers
MG
Jr
,
Affinati
AH
,
Richardson
N
,
Schwartz
MW
.
Central nervous system regulation of organismal energy and glucose homeostasis
.
Nat Metab
2021
;
3
:
737
750
14.
Cahill
GF
Jr
.
Fuel metabolism in starvation
.
Annu Rev Nutr
2006
;
26
:
1
22
15.
Mihaylova
MM
,
Vasquez
DS
,
Ravnskjaer
K
, et al
.
Class IIa histone deacetylases are hormone-activated regulators of FOXO and mammalian glucose homeostasis
.
Cell
2011
;
145
:
607
621
16.
Li
P
,
Fan
W
,
Xu
J
, et al
.
Adipocyte NCoR knockout decreases PPARγ phosphorylation and enhances PPARγ activity and insulin sensitivity
.
Cell
2011
;
147
:
815
826
17.
Liu
Y
,
Dentin
R
,
Chen
D
, et al
.
A fasting inducible switch modulates gluconeogenesis via activator/coactivator exchange
.
Nature
2008
;
456
:
269
273
18.
Langlet
F
,
Haeusler
RA
,
Lindén
D
, et al
.
Selective inhibition of FOXO1 activator/repressor balance modulates hepatic glucose handling
.
Cell
2017
;
171
:
824
835.e18
19.
Hakvoort
TB
,
Moerland
PD
,
Frijters
R
, et al
.
Interorgan coordination of the murine adaptive response to fasting
.
J Biol Chem
2011
;
286
:
16332
16343
20.
Reid
MA
,
Dai
Z
,
Locasale
JW
.
The impact of cellular metabolism on chromatin dynamics and epigenetics
.
Nat Cell Biol
2017
;
19
:
1298
1306
21.
Haws
SA
,
Leech
CM
,
Denu
JM
.
Metabolism and the epigenome: a dynamic relationship
.
Trends Biochem Sci
2020
;
45
:
731
747
22.
Carrer
A
,
Parris
JLD
,
Trefely
S
, et al
.
Impact of a high-fat diet on tissue Acyl-CoA and histone acetylation levels
.
J Biol Chem
2017
;
292
:
3312
3322
23.
Lempradl
A
,
Pospisilik
JA
,
Penninger
JM
.
Exploring the emerging complexity in transcriptional regulation of energy homeostasis
.
Nat Rev Genet
2015
;
16
:
665
681
24.
Mihaylova
MM
,
Shaw
RJ
.
Metabolic reprogramming by class I and II histone deacetylases
.
Trends Endocrinol Metab
2013
;
24
:
48
57
25.
Rosen
ED
.
Epigenomic and transcriptional control of insulin resistance
.
J Intern Med
2016
;
280
:
443
456
26.
Hino
S
,
Sakamoto
A
,
Nagaoka
K
, et al
.
FAD-dependent lysine-specific demethylase-1 regulates cellular energy expenditure
.
Nat Commun
2012
;
3
:
758
27.
Wortham
M
,
Liu
F
,
Fleischman
JY
, et al
.
Nutrient regulation of the islet epigenome controls adaptive insulin secretion
.
bioRxiv
.
2019
:
742403
28.
Tabula Muris Consortium
;
Overall coordination
;
Logistical coordination
;
Organ collection and processing
;
Library preparation and sequencing
;
Computational data analysis
;
Cell type annotation
;
Writing group
;
Supplemental text writing group
;
Principal investigators
.
Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris
.
Nature
2018
;
562
:
367
372
29.
Liu
Y
,
Lin
H
,
Jiang
L
, et al
.
Hepatic Slug epigenetically promotes liver lipogenesis, fatty liver disease, and type 2 diabetes
.
J Clin Invest
2020
;
130
:
2992
3004
30.
Fleet
T
,
Zhang
B
,
Lin
F
, et al
.
SRC-2 orchestrates polygenic inputs for fine-tuning glucose homeostasis
.
Proc Natl Acad Sci U S A
2015
;
112
:
E6068
E6077
31.
Pan
D
,
Mao
C
,
Wang
YX
.
Suppression of gluconeogenic gene expression by LSD1-mediated histone demethylation
.
PLoS One
2013
;
8
:
e66294
32.
Duteil
D
,
Tosic
M
,
Schüle
R
.
Lsd1, a metabolic sensor of environment requirements that prevents adipose tissue from aging
.
Adipocyte
2017
;
6
:
298
303
33.
Duteil
D
,
Metzger
E
,
Willmann
D
, et al
.
LSD1 promotes oxidative metabolism of white adipose tissue
.
Nat Commun
2014
;
5
:
4093
34.
Hummel
KP
,
Dickie
MM
,
Coleman
DL
.
Diabetes, a new mutation in the mouse
.
Science
1966
;
153
:
1127
1128
35.
Chua
SC
Jr
,
Chung
WK
,
Wu-Peng
XS
, et al
.
Phenotypes of mouse diabetes and rat fatty due to mutations in the OB (leptin) receptor
.
Science
1996
;
271
:
994
996
36.
Wang
J
,
Scully
K
,
Zhu
X
, et al
.
Opposing LSD1 complexes function in developmental gene activation and repression programmes
.
Nature
2007
;
446
:
882
887
37.
Gordts
PL
,
Nock
R
,
Son
NH
, et al
.
ApoC-III inhibits clearance of triglyceride-rich lipoproteins through LDL family receptors
.
J Clin Invest
2016
;
126
:
2855
2866
38.
Ramms
B
,
Patel
S
,
Sun
X
, et al
.
Interventional hepatic apoC-III knockdown improves atherosclerotic plaque stability and remodeling by triglyceride lowering
.
JCI Insight
2022
;
7
:
e158414
39.
Gordts
PL
,
Bartelt
A
,
Nilsson
SK
, et al
.
Impaired LDL receptor-related protein 1 translocation correlates with improved dyslipidemia and atherosclerosis in apoE-deficient mice
.
PLoS One
2012
;
7
:
e38330
40.
Chick
WL
,
Like
AA
.
Studies in the diabetic mutant mouse. 3. Physiological factors associated with alterations in beta cell proliferation
.
Diabetologia
1970
;
6
:
243
251
41.
Petersen
MC
,
Shulman
GI
.
Mechanisms of insulin action and insulin resistance
.
Physiol Rev
2018
;
98
:
2133
2223
42.
Murano
I
,
Barbatelli
G
,
Parisani
V
, et al
.
Dead adipocytes, detected as crown-like structures, are prevalent in visceral fat depots of genetically obese mice
.
J Lipid Res
2008
;
49
:
1562
1568
43.
Perry
RJ
,
Camporez
JG
,
Kursawe
R
, et al
.
Hepatic acetyl CoA links adipose tissue inflammation to hepatic insulin resistance and type 2 diabetes
.
Cell
2015
;
160
:
745
758
44.
Li
X
,
Zhao
Y
,
Chen
C
, et al
.
Critical role of matrix metalloproteinase 14 in adipose tissue remodeling during obesity
.
Mol Cell Biol
2020
;
40
:
e00564-19
45.
Liu
X
,
Huang
T
,
Li
L
, et al
.
CYP1B1 deficiency ameliorates obesity and glucose intolerance induced by high fat diet in adult C57BL/6J mice
.
Am J Transl Res
2015
;
7
:
761
771
46.
Krüger
J
,
Brachs
S
,
Trappiel
M
, et al
.
Enhanced insulin signaling in density-enhanced phosphatase-1 (DEP-1) knockout mice
.
Mol Metab
2015
;
4
:
325
336
47.
Foley
KP
,
Chen
Y
,
Barra
NG
, et al
.
Inflammation promotes adipocyte lipolysis via IRE1 kinase
.
J Biol Chem
2021
;
296
:
100440
48.
Yang
X
,
Zhang
X
,
Heckmann
BL
,
Lu
X
,
Liu
J
.
Relative contribution of adipose triglyceride lipase and hormone-sensitive lipase to tumor necrosis factor-α (TNF-α)-induced lipolysis in adipocytes
.
J Biol Chem
2011
;
286
:
40477
40485
49.
Anthonsen
MW
,
Rönnstrand
L
,
Wernstedt
C
,
Degerman
E
,
Holm
C
.
Identification of novel phosphorylation sites in hormone-sensitive lipase that are phosphorylated in response to isoproterenol and govern activation properties in vitro
.
J Biol Chem
1998
;
273
:
215
221
50.
Yang
X
,
Lu
X
,
Lombès
M
, et al
.
The G(0)/G(1) switch gene 2 regulates adipose lipolysis through association with adipose triglyceride lipase
.
Cell Metab
2010
;
11
:
194
205
51.
Wang
H
,
Bell
M
,
Sreenivasan
U
, et al
.
Unique regulation of adipose triglyceride lipase (ATGL) by perilipin 5, a lipid droplet-associated protein
.
J Biol Chem
2011
;
286
:
15707
15715
52.
Kuramoto
K
,
Okamura
T
,
Yamaguchi
T
, et al
.
Perilipin 5, a lipid droplet-binding protein, protects heart from oxidative burden by sequestering fatty acid from excessive oxidation
.
J Biol Chem
2012
;
287
:
23852
23863
53.
Ishida
E
,
Kim-Muller
JY
,
Accili
D
.
Pair feeding, but not insulin, phloridzin, or rosiglitazone treatment, curtails markers of β-cell dedifferentiation in db/db mice
.
Diabetes
2017
;
66
:
2092
2101
54.
Kleinert
M
,
Clemmensen
C
,
Hofmann
SM
, et al
.
Animal models of obesity and diabetes mellitus
.
Nat Rev Endocrinol
2018
;
14
:
140
162
55.
Glass
CK
,
Olefsky
JM
.
Inflammation and lipid signaling in the etiology of insulin resistance
.
Cell Metab
2012
;
15
:
635
645
56.
Czech
MP
,
Tencerova
M
,
Pedersen
DJ
,
Aouadi
M
.
Insulin signalling mechanisms for triacylglycerol storage
.
Diabetologia
2013
;
56
:
949
964
57.
Wu
H
,
Ballantyne
CM
.
Metabolic inflammation and insulin resistance in obesity
.
Circ Res
2020
;
126
:
1549
1564
58.
Yang
A
,
Mottillo
EP
.
Adipocyte lipolysis: from molecular mechanisms of regulation to disease and therapeutics
.
Biochem J
2020
;
477
:
985
1008
59.
Guilherme
A
,
Henriques
F
,
Bedard
AH
,
Czech
MP
.
Molecular pathways linking adipose innervation to insulin action in obesity and diabetes mellitus
.
Nat Rev Endocrinol
2019
;
15
:
207
225
60.
Han
MS
,
White
A
,
Perry
RJ
, et al
.
Regulation of adipose tissue inflammation by interleukin 6
.
Proc Natl Acad Sci U S A
2020
;
117
:
2751
2760
61.
Ying
W
,
Riopel
M
,
Bandyopadhyay
G
, et al
.
Adipose tissue macrophage-derived exosomal miRNAs can modulate in vivo and in vitro insulin sensitivity
.
Cell
2017
;
171
:
372
384.e12
62.
Wang
L
,
Pydi
SP
,
Zhu
L
, et al
.
Adipocyte Gi signaling is essential for maintaining whole-body glucose homeostasis and insulin sensitivity
.
Nat Commun
2020
;
11
:
2995
63.
Petersen
MC
,
Vatner
DF
,
Shulman
GI
.
Regulation of hepatic glucose metabolism in health and disease
.
Nat Rev Endocrinol
2017
;
13
:
572
587
64.
Lewis
GF
,
Carpentier
AC
,
Pereira
S
,
Hahn
M
,
Giacca
A
.
Direct and indirect control of hepatic glucose production by insulin
.
Cell Metab
2021
;
33
:
709
720
65.
Donnelly
KL
,
Smith
CI
,
Schwarzenberg
SJ
,
Jessurun
J
,
Boldt
MD
,
Parks
EJ
.
Sources of fatty acids stored in liver and secreted via lipoproteins in patients with nonalcoholic fatty liver disease
.
J Clin Invest
2005
;
115
:
1343
1351
66.
Perry
RJ
,
Samuel
VT
,
Petersen
KF
,
Shulman
GI
.
The role of hepatic lipids in hepatic insulin resistance and type 2 diabetes
.
Nature
2014
;
510
:
84
91
67.
Alves-Bezerra
M
,
Cohen
DE
.
Triglyceride metabolism in the liver
.
Compr Physiol
2017
;
8
:
1
8
68.
Zeng
X
,
Jedrychowski
MP
,
Chen
Y
, et al
.
Lysine-specific demethylase 1 promotes brown adipose tissue thermogenesis via repressing glucocorticoid activation
.
Genes Dev
2016
;
30
:
1822
1836
69.
Hwang
LL
,
Wang
CH
,
Li
TL
, et al
.
Sex differences in high-fat diet-induced obesity, metabolic alterations and learning, and synaptic plasticity deficits in mice
.
Obesity (Silver Spring)
2010
;
18
:
463
469
70.
Pettersson
US
,
Waldén
TB
,
Carlsson
PO
,
Jansson
L
,
Phillipson
M
.
Female mice are protected against high-fat diet induced metabolic syndrome and increase the regulatory T cell population in adipose tissue
.
PLoS One
2012
;
7
:
e46057
71.
Huang
KP
,
Ronveaux
CC
,
Knotts
TA
,
Rutkowsky
JR
,
Ramsey
JJ
,
Raybould
HE
.
Sex differences in response to short-term high fat diet in mice
.
Physiol Behav
2020
;
221
:
112894
72.
Casimiro
I
,
Stull
ND
,
Tersey
SA
,
Mirmira
RG
.
Phenotypic sexual dimorphism in response to dietary fat manipulation in C57BL/6J mice
.
J Diabetes Complications
2021
;
35
:
107795
73.
Mauvais-Jarvis
F
,
Arnold
AP
,
Reue
K
.
A guide for the design of pre-clinical studies on sex differences in metabolism
.
Cell Metab
2017
;
25
:
1216
1230
74.
Lazzaroni
E
,
Ben Nasr
M
,
Loretelli
C
, et al
.
Anti-diabetic drugs and weight loss in patients with type 2 diabetes
.
Pharmacol Res
2021
;
171
:
105782
75.
Aaseth
J
,
Ellefsen
S
,
Alehagen
U
,
Sundfør
TM
,
Alexander
J
.
Diets and drugs for weight loss and health in obesity - An update
.
Biomed Pharmacother
2021
;
140
:
111789
76.
Wilding
JPH
,
Batterham
RL
,
Calanna
S
, et al.;
STEP 1 Study Group
.
Once-weekly semaglutide in adults with overweight or obesity
.
N Engl J Med
2021
;
384
:
989
1002
77.
Newsome
PN
,
Buchholtz
K
,
Cusi
K
, et al.;
NN9931-4296 Investigators
.
A placebo-controlled trial of subcutaneous semaglutide in nonalcoholic steatohepatitis
.
N Engl J Med
2021
;
384
:
1113
1124
78.
Dewanjee
S
,
Vallamkondu
J
,
Kalra
RS
, et al
.
The emerging role of HDACs: pathology and therapeutic targets in diabetes mellitus
.
Cells
2021
;
10
:
1340
79.
Ling
C
,
Rönn
T
.
Epigenetics in human obesity and type 2 diabetes
.
Cell Metab
2019
;
29
:
1028
1044
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.