Lipolysis drives expression of the constitutively active receptor GPR3 to induce adipose thermogenesis

Abstract

Thermogenic adipocytes possess a therapeutically appealing, energy-expending capacity, which is canonically cold-induced by ligand-dependent activation of β-adrenergic G protein-coupled receptors (GPCRs). Here, we uncover an alternate paradigm of GPCR-mediated adipose thermogenesis through the constitutively active receptor, GPR3. We show that the N terminus of GPR3 confers intrinsic signaling activity, resulting in continuous Gs-coupling and cAMP production without an exogenous ligand. Thus, transcriptional induction of Gpr3 represents the regulatory parallel to ligand-binding of conventional GPCRs. Consequently, increasing Gpr3 expression in thermogenic adipocytes is alone sufficient to drive energy expenditure and counteract metabolic disease in mice. Gpr3 transcription is cold-stimulated by a lipolytic signal, and dietary fat potentiates GPR3-dependent thermogenesis to amplify the response to caloric excess. Moreover, we find GPR3 to be an essential, adrenergic-independent regulator of human brown adipocytes. Taken together, our findings reveal a noncanonical mechanism of GPCR control and thermogenic activation through the lipolysis-induced expression of constitutively active GPR3.

Keywords: G protein-coupled receptor; GPCR; GPR3; adrenergic receptor; brown adipose tissue; constitutively active; energy expenditure; lipolysis; thermogenesis; transcription.

Graphical abstract

Figure 1

The constitutively active receptor GPR3 is the most cold-induced Gs-coupled GPCR in thermogenic adipose tissue (A) Schematic depicting canonical ligand-dependent (solid line) versus hypothesized transcriptional control (dotted line) of Gs-coupled receptors in thermogenic adipocytes. (B) Induction of Gs-coupled receptors in brown (left) and subcutaneous (right) white adipose depots during adaptation to cold. Statistical significance for each receptor at individual time points is indicated in Table S1 (BAT) and Table S2 (scWAT). (C) cAMP accumulation in COS-7 cells transfected with increasing concentrations of GPR3 plasmid; gene expression data presented in log scale. (D) Schematic depicting the bioluminescence resonance energy transfer (BRET) assay used to assess. (E) G protein recruitment to wild-type (WT) and DRY-mutant GPR3. (F) Scheme depicting the BRET assay used to assess. (G–I) (G) cAMP levels produced by WT and N-terminal truncations of GPR3 and cAMP production induced by N-terminal GPR3 fragment aa18-27 on (H) WT GPR3 and (I) cannabinoid 1 receptor (CB1). (J) Tissue panel of cold-induced fold changes in Gpr3 expression. (K) Differential levels of cold-induced Gpr3 expression in BAT adipocytes (Ad) and stromal vascular fraction (SVF). (L) In situ hybridization (ISH) of Gpr3 mRNA (red) in BAT of thermoneutral-housed or cold-challenged mice. Nuclei in BAT are stained with DAPI (blue). For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (K and J) or Bonferroni's multiple comparisons test (G). See also Figure S1.

Figure S1

Cold-induced GPCR expression in mouse tissues and Gpr3 transcription in β-less mice housed at thermoneutrality, related to Figures 1 and 2 (A) transcriptional regulation of established BAT activating Gs-coupled receptors in BAT during adaptation to cold. (B) induction of Gs-coupled receptors in brown (left) and subcutaneous (right) white adipose depots during adaptation to cold (non-normalized values from Figure 1B). (C) tissue panel of cold-induced Gpr3 expression. (D) in situ hybridization (ISH) of Gpr3 mRNA (red) in scWAT, E, scWAT (high magnification. Dotted arrow: Unilocular adipocyte. Solid arrow: Multilocular adipocyte), and, F, eWAT of thermoneutral-housed or cold-challenged mice. BAT Gpr3 expression in, G, thermoneutral-acclimated β-less mice and wildtype controls. For all panels, error bars represent ±SEM, p ≤ 0.05=∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (C) or Bonferroni's multiple comparisons test (A).

Figure 2

A lipolytic signal controls cold-induced expression of Gpr3 (A) Cold-induced Gpr3 expression in BAT. (B) Lipolytic activity in ex vivo eWAT from wild-type control and β-less mice stimulated with either isoproterenol (ISO) or forskolin (Fsk). (C–E) (C) Protein levels of adipose triglyceride lipase (ATGL) and hormone-sensitive lipase (HSL) in BAT, (D) gene expression of Atgl and Hsl, and (E) lipolytic activity in primary subcutaneous white adipocytes following ISO treatment with or without lipase inhibitors (ATGL inhibitor, Atglistatin; HSL inhibitor, 76-0079). (F and G) BAT Gpr3 expression in (F) acute cold-induced (RT, room temperature) and (G) 3-week cold-adapted DAKO mice and control littermates. (H) Gpr3 expression in primary brown adipocytes following ISO treatment with or without lipase inhibitors (ATGL inhibitor, Atglistatin; HSL inhibitor, CAY10499). (I) Regulation of brown adipocyte Gpr3 expression by the lipolytic activator, SR-3420, with or without lipase inhibitors (ATGL inhibitor, Atglistatin; HSL inhibitor, CAY10499). (J) Gpr3 expression in brown adipocytes following SR-3420 treatment with or without lipid oxidation inhibitors. (K) BAT gene expression from mice given PPARα (fenofibrate) or PPARγ (rosiglitazone) agonists for 2 weeks. (L) Small interfering RNA (siRNA)-mediated knockdown of the lipid-activated nuclear receptors. (M) Impact on norepinephrine (NE)-induced Gpr3 expression in brown adipocytes. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (A, B, D, F–J, L, and M) or Bonferroni's multiple comparisons test (E and K). See also Figure S1.

Figure S2

Characterization of the in vitro and in vivo LV delivery models and Gpr3 OE primary adipocyte model, related to Figure 3 (A) gene expression from brown adipocytes following lentiviral (LV)-mediated overexpression of wildtype (WT) and DRY-mutant GPR3. The shaded region indicates the physiological range of maximal cold-induced Gpr3 expression in BAT. (B–F) (B) Fluorescent visualization of BAT (BF=bright field), (C) physical activity, (D) food intake, (E) bodyweight, and (F) UCP1 staining in BAT from mice injected with LV particles carrying either Gfp or Gpr3. (G) schematic of Gpr3 OE mice, in which the Gpr3 coding region is preceded by a synthetic CAG promoter and lox-STOP-lox cassette (TAM=tamoxifen). (H) representative light microscopy images of primary brown and subcutaneous white adipocytes with and without TAM-induced Gpr3 expression. For all panels, error bars represent ±SEM.

Figure 3

Gpr3 overexpression activates the adipose thermogenic program independently of sympathetic signaling (A–C) (A) Mitochondrial respiration, (B) fatty acid (FA) uptake, and (C) gene expression from brown adipocytes expressing wild-type (WT) or DRY-mutant GPR3. (D–H) (D) Schematic depicting the site-directed, high-titer lentiviral (LV) injections used to (E) overexpress Gfp or Gpr3 in BAT and assess (F) energy expenditure, (G) tissue weights, and (H) Ucp1 gene expression. (I–K) (I) Gene expression, (J) FA uptake, and (K) mitochondrial respiration of primary brown adipocytes with or without tamoxifen (TAM)-induced Gpr3 expression. (L and M) (L) BAT cAMP levels and (M) tissue-specific triglyceride (TG)-derived FA uptake in C-3BO mice and control littermates. (N and O) (N) Schematic of BAT denervation protocol used to assess 18-fluorodeoxyglucose (¹⁸F-FDG) uptake and (O) Ucp1 gene expression of C-3BO mice and control littermates. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (E–M and O) or Bonferroni's multiple comparisons test (A–C). Box plots are presented as box: 25^th to 75^th percentile, and whiskers: min to max. See also Figures S2 and S3.

Figure S3

Characterization of the Gpr3 OE primary adipocyte model (continued) and C-3BO mouse model, related to Figure 3 (A–C) (A) thermogenic gene expression, (B) fatty acid (FA) uptake, and (C) mitochondrial respiration of primary subcutaneous white adipocytes with and without TAM-induced Gpr3 expression. (D and E) Adrenergic receptor gene expression from primary (D) brown and (E) subcutaneous white adipocytes with and without TAM-induced Gpr3 expression. (F–I) (F) Gpr3 expression levels across fat depots, (G) representative interscapular BAT (iBAT) images, (H) thermogenic gene expression in BAT, and (I) thermogenic gene expression in scWAT of chow-fed C-3BO mice and control littermates. (J–L) (J) tyrosine hydroxylase immunohistochemistry (IHC), (K) quantified 18-Fluorodeoxyglucose (¹⁸F-FDG) uptake in BAT one week after denervation surgery (PET=positron emission tomography), and (L) hematoxylin and eosin (H&E)-staining of sham and denervated iBAT of C-3BO mice and control littermates. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test. Box plots are presented as box: 25^th to 75^th percentile and whiskers: min to max.

Figure S4

Phenotyping of the C-3BO mouse model, related to Figure 4 (A–F) (A) bodyweights, (B) lean and fat mass, (C) tissue weights, (D) food intake (average per day), (E) physical activity (average per 15 min), and (F) energy expenditure of chow-fed C-3BO mice and control littermates. (G) food intake (average per day) of C-3BO mice and control littermates after transition to high fat diet (HFD). (H) change in HFD-induced energy expenditure between CL-316,243 (CL) and saline-injected mice and C-3BO and control littermates. (I and J) (I) bodyweight gain and (J) tissue weights of CL/saline and C-3BO/control cohorts (after 1-week HFD-challenge). (K) pathway analysis of gene networks specifically induced in C-3BO mice. (L) gene expression of CL/saline and C-3BO/control cohorts (after 1-week HFD-challenge). For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, t test (A, C, J, and L), two-way ANOVA (B and I), or Fisher’s exact test (K).

Figure 4

Dietary fat potentiates GPR3-mediated thermogenic activation (A–D) Indirect calorimetry and respiratory exchange ratio (RER) during the transition from chow to high fat diet (HFD) for (A and B) C-3BO mice and control littermates and (C and D) mice injected daily with 1 mg/kg CL-316,243 (CL). (E–G) (E) Heat map, (F), pathway enrichment, and (G) quantification of genes induced by CL-treatment and in C-3BO mice under chow and HFD-fed conditions. Error bars represent ±SEM, p ≤ 0.001 = ∗∗∗, Fisher’s exact test (F) or Wilcoxon signed-rank tests (G). See also Figure S4.

Figure 5

GPR3 activation of thermogenic adipocytes counteracts metabolic disease (A and B) (A) Bodyweight (BW) gain and (B) body composition of C-3BO mice and control littermates over the course of an 8-week high fat diet (HFD) challenge. (C–G) (C) Tissue weights, (D) liver triglycerides (TG), (E) energy expenditure (EE), (F) BAT thermogenic gene expression, and (G) glucose tolerance of C-3BO mice and control littermates during HFD challenge. (H and I) (H) cAMP levels in BAT 1 week after tamoxifen (TAM) administration, and (I) indirect calorimetry of obese I-3BO mice and control littermates following 3 consecutive days of TAM-treatment by oral gavage. (J and K) (J) Weight loss and (K) tissue weights 1 week after TAM-administration. (L–N) (L) TG-derived fatty acid (FA) uptake, (M) glucose uptake, and (N) glucose tolerance in HFD-fed I-3BO mice and control littermates. (O–R) (O) Schematic depicting the site-directed adeno-associated virus (AAV) injections used to (P) overexpress Gfp, Gpr3, or Adrb3 in BAT and assess, (Q) energy expenditure during chow to HFD transition, (R) HFD-induced EE versus BW, and (S) BAT thermogenic gene expression. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (C, D, F, H, K, L, and M), Bonferroni's multiple comparisons test (S), or two-way ANOVA (A, B, G, J, and N). See also Figure S5.

Figure S5

Characterization of the C-3BO, I-3BO, AAV-modified, and B-3KO mouse models, related to Figures 5 and 6 (A) bodyweight gain of C-3BO mice and control littermates challenged with high fat diet (HFD) (independent experiment from the study in Figure 3). (B–E) (B) food intake (average per day), (C) lean mass, (D) representative interscapular BAT images, and (E) thermogenic gene expression in scWAT of HFD-fed C-3BO mice and control littermates. (F) thermogenic gene expression in BAT of HFD-fed I-3BO mice and control littermates. (G and H) (G) HFD-induced energy expenditure and (H) food intake (average per day) of HFD-fed I-3BO mice and control littermates. (I) HFD-induced day and night energy expenditure in mice infected with adeno-associated virus (AAV) particles carrying either Gfp, Adrb3, or Gpr3. (J) schematic for conditional deletion of Gpr3 in the B-3KO mouse model. (K) Gpr3 in situ hybridization in BAT from B-3KO and control littermates. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (E, F, and I) or two-way ANOVA (A and C).

Figure 6

BAT Gpr3 is required for thermogenic activity in vitro but is compensated in vivo (A) BAT thermogenic gene expression from high fat diet (HFD)-fed B-3KO mice and control littermates. (B–E) (B) HFD-induced weight gain, (C) cold tolerance, (D) glucose tolerance, and (E) cold-induced BAT gene expression from B-3KO mice and control littermates (RT, room temperature). (F) Norepinephrine (NE)-induced energy expenditure in anesthetized B-3KO mice and control littermates at thermoneutrality following acute cold-challenge. (G) Indirect calorimetry of B-3KO mice and control littermates during the transition from chow to HFD. (H–J) (H) Mitochondrial respiration, (I) fatty acid (FA) uptake, and (J) gene expression in brown adipocytes following siRNA-mediated Gpr3 knockdown. For all panels, error bars represent ±SEM, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test. Box plots are presented as box: 25^th to 75^th percentile, and whiskers: min to max. See also Figure S5.

Figure 7

GPR3 is an essential activator of human thermogenic adipocytes (A) Structural location (snake plot) and disease association (table) of human GPR3 variant, A27G. (B) Functional consequence of A27G mutation on GPR3 cAMP-inducing activity. (C) Correlation between BAT GPR3 expression and body mass index (BMI) in glucose tolerant and glucose intolerant individuals. (D) Schematic of GPR3 loss-of-function and gain-of-function studies in patient-derived, non-immortalized brown adipocytes. (E) Gene expression of siRNA-mediated GPR3 knockdown and 4-h vehicle or norepinephrine (NE) treatment of patient-derived, non-immortalized brown adipocytes. (F and G) Pathway analysis of gene networks (F) reduced by GPR3 depletion and (G) induced by GPR3 activation in patient-derived, non-immortalized brown adipocytes. (H) Gene expression following siRNA-mediated GPR3 knockdown and 4-h vehicle or NE treatment of patient-derived, non-immortalized brown adipocytes. (I) Change in the gene expression of Gs-coupled GPCRs in human brown adipocytes following siRNA-mediated GPR3 knockdown. (J) Correlation between GPR3 and ADRB1 expression in human BAT. (K and L) Correlation between (K) GPR3 and ADRB2 and (L) GPR3 and UCP1 expression in human scWAT. (M) GPR3 expression in scWAT before and after bariatric surgery (NonOB, non-obese; OB, obese; PostOB, post-obese). (N–P) (N) Gene expression, (O) leak respiration, and (P) fatty acid (FA) uptake in human subcutaneous white adipocytes in which GPR3 expression has been induced by CRISPR/Cas9-engineering. (Q) Pathway analysis of gene networks induced by GPR3 in CRISPR/Cas9-engineered human subcutaneous white adipocytes. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test (E, H, I, and M–P), two-way ANOVA (B), simple linear regression (C and J–L), or Fisher’s exact test (F, G, and Q). Box plots are presented as box: 25^th to 75^th percentile, and whiskers: min to max. See also Figures S6 and S7.

Figure S6

Characterization of GPR3 in human thermogenic adipocytes, related to Figure 7 (A) GPR3 expression of transfected COS-7 cells for BRET-analysis. (B–D) (B) representative light microscopy images and, (C and D), gene expression of patient-derived, non-immortalized brown adipocytes following siRNA mediated GPR3 knockdown and 4 hours vehicle or norepinephrine (NE) treatment. (E and F)) (E) heat map of gene regulation and, specifically, (F) genes in the de novo cholesterol synthesis pathway changed by GPR3 depletion in patient-derived, non-immortalized brown adipocytes. (G) heat map of genes induced by GPR3 activation in patient-derived, non-immortalized brown adipocytes. (H) gene expression of patient-derived, non-immortalized brown adipocytes following siRNA mediated GPR3 knockdown and 4 hours vehicle or NE treatment. (I) gene expression in human brown adipocytes following siRNA mediated GPR3 knockdown. For all panels, error bars represent ±SEM, p ≤ 0.05 = ∗, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, p ≤ 0.0001 = ∗∗∗∗, t test.

Figure S7

Characterization of GPR3 in human thermogenic adipocytes (continued), related to Figure 7 (A and B) Correlations between (A) GPR3 and ADRB2 as well as GPR3 and ADRB3 expression in human BAT and (B) GPR3 and ADRB1 as well as GPR3 and ADRB3 expression in human scWAT. (C) schematic depicting the analysis in GTEx of human GPCR co-regulation with GPR3 (G protein-coupling data based on Inoue et al. 2019). (D) top ten GPCRs from the GTEx analysis that are negatively correlated with GPR3. (E and F) (E) gene expression and (F) heat map of global gene profiling of human subcutaneous white adipocytes with CRISPR-engineered GPR3 overexpression. (G) model comparing the canonical ligand-based activation of GPCRs versus transcriptional induction of constitutively active receptors in the control of adipose thermogenesis. (H) change in the expression of 336 GPCRs, which do not primarily signal through Gs-coupling in human brown adipocytes following siRNA mediated GPR3 knockdown. For all panels, error bars represent ±SEM, p ≤ 0.01 = ∗∗, p ≤ 0.001 = ∗∗∗, t test (E and H) or simple linear regression (A and B).