GPR180 is a component of TGFβ signalling that promotes thermogenic adipocyte function and mediates the metabolic effects of the adipocyte-secreted factor CTHRC1

Abstract

Activation of thermogenic brown and beige adipocytes is considered as a strategy to improve metabolic control. Here, we identify GPR180 as a receptor regulating brown and beige adipocyte function and whole-body glucose homeostasis, whose expression in humans is associated with improved metabolic control. We demonstrate that GPR180 is not a GPCR but a component of the TGFβ signalling pathway and regulates the activity of the TGFβ receptor complex through SMAD3 phosphorylation. In addition, using genetic and pharmacological tools, we provide evidence that GPR180 is required to manifest Collagen triple helix repeat containing 1 (CTHRC1) action to regulate brown and beige adipocyte activity and glucose homeostasis. In this work, we show that CTHRC1/GPR180 signalling integrates into the TGFβ signalling as an alternative axis to fine-tune and achieve low-grade activation of the pathway to prevent pathophysiological response while contributing to control of glucose and energy metabolism.

Fig. 1. GPR180 is upregulated in BAT and required for the brown phenotype.

a Workflow of target identification by transcriptomic analysis of human supraclavicular BAT, subcutaneous WAT and hMADS cells differentiated into beige and white adipocytes. DESeq2 detects DE genes based on a generalized linear model using the negative binomial distribution. Effect of GPR180 silencing in human beige adipocytes on b UCP1 protein (n = 9; p < 0.0001) and c mitochondrial respiration (n = 5; p = 0.0016 for cAMP-stimulated uncoupled respiration). d Quantification of lentiviral GPR180 overexpression on mRNA level in white adipocytes (n = 3; p < 0.0001) and its effect on e UCP1 protein (n = 9; p = 0.0021) and f mitochondrial oxygen consumption (n = 5; p = 0.0437 for cAMP-stimulated uncoupled respiration and p = 0.0161 for maximal respiration). Data are presented as mean ± SEM. Statistical analysis was performed by two-sided Student´s t-test (b), one-way ANOVA with Dunnett’s post-hoc test (d, e) or two-way ANOVA with Tukey’s post-hoc test (c, f). Significance is indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. cAMP cyclic adenosine monophosphate, GPR180 G protein-coupled receptor 180, hMADS human multipotent adipose-derived stem cells, OCR oxygen consumption rate, RFP red fluorescent protein, scBAT supraclavicular brown adipose tissue, UCP1 Uncoupling protein 1.

Fig. 2. Metabolic derangements in GPR180 knockout mice are caused by dysfunctional BAT.

UCP1 protein in a iBAT (p = 0.0425) and b iWAT (p = 0.0027) of GPR180 global knockout mice and their wild-type littermates (n = 7) fed chow diet and housed at room temperature (RT). c Energy expenditure in male mice with deleted GPR180 (n = 6) on chow diet and housed at RT (AUC p < 0.0001). d Representative images including rainbow scale bar indicating temperature range with min 28 °C and max 36 °C and e quantification of surface temperature in male GPR180 global knockout mice and wild-type littermates (n = 6; p = 0.0065). f Intraperitoneal glucose tolerance test in 12-weeks-old male mice housed at RT and fed chow diet (WT n = 6, GPR180^−/− n = 7; p = 0.047 at 15 min, p = 0.0001 at 30 min and p < 0.0001 at 60 min). g Body weight gain (p = 0.0212 at 6 weeks, p = 0.0036 at 7 weeks, p = 0.0006 at 8 weeks, p = 0.0003 at 9 weeks, p = 0.0004 at 10 weeks, p < 0.0001 at 11 and 12 weeks), and h hepatic lipid accumulation (p = 0.008856) in male mice housed RT and fed HFD for 12 weeks (WT n = 11, GPR180^−/− n = 12). i Glucose tolerance test in 12-weeks-old male mice housed at thermoneutrality (TN) for 8 weeks prior the test and fed chow diet (n = 5). j Body weight gain and k hepatic lipid accumulation in animals housed at TN and fed HFD (n = 5). Representative blots and quantification of the UCP1 protein levels in l iBAT and m iWAT (p = 0.0473) of male adipocyte-specific GPR180 (aGPR180) knockout mice and fl/fl controls (fl/fl control n = 11, aGPR180 knockout n = 10). n Energy expenditure in male aGPR180 knockout mice and fl/fl controls (fl/fl control n = 6, aGPR180 knockout n = 5; p = 0.0473; 0.0051; 0.0450; 0.0098; 0.0139; 0.0412; 0.0277; 0.0212; 0.0278). o Representative images including rainbow scale bar indicating temperature range with min 28 °C and max 36 °C and p quantification of surface temperature (p = 0.0255 for basal and p = 0.0086 for post CL-316,423) in male aGPR180 knockout mice and fl/fl controls (fl/fl control n = 9, aGPR180 n = 8). q Glucose tolerance test in male aGPR180 knockouts and fl/fl controls (n = 4; p = 0.0002 at 15 min and p = 0.0243 at 30 min) 2 weeks after tamoxifen (TAM) gavage (2 mg/animal) in two consecutive days while mice were housed at RT and fed chow diet. r Body weight gain in aGPR180 knockout mice fed HFD housed at RT (fl/fl control n = 6, aGPR180 knockout n = 7; p = 0.0419 at 11 weeks and p = 0.0011 at 12 weeks). Data are presented as mean ± SEM. Statistical significance was calculated using two-sided (a–c, h, k) and one-sided Student´s t-test (l, m) or two-way ANOVA with repeated measurements followed by Sidak post-hoc test (e–g, i, j, p–r) and Fisher’ LSD multiple comparison test (n). Area under the curve was calculated to compare energy expenditure in the global knockout mice (c). Statistical differences are indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. AUC area under the curve, CL CL-316,243, GPR180, G protein-coupled receptor 180, HFD high-fat diet, HSP90 Heat shock protein 90, iBAT interscapular brown adipose tissue, iWAT inguinal white adipose tissue, RT room temperature, TAM tamoxifen, UCP1 Uncoupling protein 1, WT wild-type.

Fig. 3. GPR180 is not a GPCR, but a component of TGFβ signalling pathway enhancing mature beige adipocytes function.

a KEGG pathway analysis of differentially expressed genes (DEGs) of human beige adipocytes with ablated GPR180. Phosphorylation of SMAD3 at serine 423 in b non-starved human beige adipocytes after GPR180 silencing (n = 6; p = 0.0002), c beige adipocytes treated with different concentrations of TGFβ1 in combination with knockdown of GPR180 or TGFβR2 (ctrl siRNA vs siGPR180 p = 0.0416 for 1 pg/ml TGFβ1, p = 0.0129 for 100 pg/ml TGFβ1, p = 0.0034 for 1 ng/ml TGFβ1 and p = 0.0146 for 10 ng/ml TGFβ1) and d in white adipocytes overexpressing GPR180 (n = 6; p = 0.0050). Representative images of epitope tag immunostaining (green), nuclei stained by Hoechst (blue), e N-terminal HA tag and f C-terminal V5 tag in hMADS cells overexpressing modified GPR180; scale bar 100 µm. Experiment was performed 3 times with similar results. Long-term TGFβ1 treatment (72 h) dose-dependently promotes g UCP1 protein (n = 6; p = 0.0084 for 0.1 ng/ml and p < 0.0001 for 1 ng/ml) and h mitochondrial respiration (n = 5; cAMP uncoupled respiration p = 0.0020 for 1 pg/ml TGFβ1, p = 0.0334 for 10 pg/ml, p = 0.0008 for 100 pg/ml and p = 0.0012 for 1 ng/ml; maximal respiration p = 0.0205 for 1 pg/ml TGFβ1, p = 0.0178 for 10 pg/ml, p < 0.0001 for 100 pg/ml and p = 0.0127 for 1 ng/ml TGFβ1) in mature human beige adipocytes. Effect of i pharmacological (p = 0.0062 for 1 µM and p = 0.0017 for 10 µM) and j genetic (p = 0.0002) inhibition of TGFβR1 on UCP1 protein level in beige hMADS cells (n = 6). Effect of TGFβR2 silencing on k UCP1 expression (n = 6; p < 0.0001) and l HSL phosphorylation at serine 660 (n = 6; p = 0.0126). m Mitochondrial oxygen consumption rate following knockdown of individual TGFβ receptors in mature beige adipocytes (n = 5; p = 0.0082 for cAMP-stimulated uncoupled respiration and p < 0.0001 for maximal respiration). Effect of SMAD2 (p = 0.0001) and SMAD3 (p < 0.0001) knockdown on n UCP1 protein level (n = 6) and o mitochondrial respiration (n = 5; p = 0.0039 for SMAD3 and p = 0.0044 for SMAD2 cAMP-stimulated uncoupled respiration; p = 0.0001 for SMAD3 and p = 0.0011 for SMAD2 maximal respiration) in beige hMADS cells. Data are shown as average ±SEM. Statistical analysis was performed using two-sided Student´s t-test (b, j–l), one-way ANOVA with Dunnett’s post-hoc test (c, d, g, i, n) or two-way ANOVA with Tukey post-hoc test (h, m, o) and significance is indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. cAMP cyclic adenosine monophosphate, GPR180 G protein-coupled receptor 180, HSL Hormone sensitive lipase, HSP90 Heat shock protein 90, OCR oxygen consumption rate, RFP red fluorescent protein, SMAD3 Mothers against decapentaplegic homolog 3, TGFβ1 Transforming growth factor β1, TGFβR1 Transforming growth factor β receptor type 1, TGFβR2 Transforming growth factor β receptor type 2, UCP1 Uncoupling protein 1.

Fig. 4. CTHRC1 induces SMAD3 phosphorylation via GPR180.

a Immunostaining of TGFβ proteins in conditioned media of white and beige hMADS adipocytes with recombinant TGFβ as a positive control and identification of secreted proteins by hMADS adipocytes. b TGFβ1 levels in conditioned media of preadipocytes and mature white and beige hMADS adipocytes (n = 3). c Representative blots and quantification of SMAD3 phosphorylation in white adipocytes overexpressing GPR180 while treated with TGFβ neutralizing antibody (n = 6; p < 0.0001 for RFP IgG vs GPR180 IgG; p = 0.0064 for GPR180 IgG vs GPR180 anti-TGFβ and p = 0.0408 for RFP anti-TGFβ vs GPR180 anti-TGFβ). d SMAD3 phosphorylation in beige adipocytes following CTHRC1 treatment in combination with GPR180 ablation (n = 6; p = 0.0012 for PBS ctrl siRNA vs CTHRC1 ctrl siRNA and p = 0.0030 for CTHRC1 ctrl siRNA vs CTHRC1 siGPR180). e Time-dependent effect of acute CTHRC1 stimulus on SMAD3 phosphorylation after the adipocytes were starved for 2 h (n = 6; p = 0.0103 for 30 min and p = 0.0073 for 60 min). f Effect of CTHRC1 and TGFβ1 treatment (1 h) on SMAD3 phosphorylation in beige hMADS cells in presence of IgG and neutralizing anti-TGFβ antibody (n = 6; p = 0.0133 ctrl vs CTHRC1 within IgG, p < 0.0001 ctrl vs TGFβ1 within IgG, p < 0.0001 CTHRC1 vs TGFβ1 within IgG, p = 0.0442 ctrl vs CTHRC1 within anti-TGFβ, p < 0.0001 ctrl vs TGFβ1 within anti-TGFβ and p = 0.0095 CTHRC1 vs TGFβ1 within anti-TGFβ). g Representative images with scale bar 100 µm and quantification of h phosphorylated SMAD3 (green) positive nuclei (blue) (n = 15; p = 0.0002 for CTHRC1 and p < 0.0001 for TGFβ1) and i total SMAD3 (green) shuttling in response to CTHRC1 and TGFβ1 (n = 1400; p < 0.0001 for both CTHRC1 and TGFβ1). Experiment was repeated independently twice with similar results. j Representative western blots and quantification of the binding of CTHRC1 tagged with HiBiT in control and stable GPR180 knockout HEK-293T cells (n = 6; p < 0.0001). k Effect of CTHRC1 stimulus (1 h) on SMAD3 phosphorylation in TGFβR1 kinase inhibitor pre-treated cells (n = 4; p = 0.0356 for CTHRC1, p < 0.0001 for TGFβ1 and p < 0.0001 for CTHRC1 vs TGFβ1 within DMSO). SMAD3 phosphorylation in response to CTHRC1 in combination with silencing of l TGFβR1 (n = 6; p = 0.0053 for PBS ctrl siRNA vs CTHRC1 ctrl siRNA, p = 0.0137 for CTHRC1 ctrl siRNA vs CTHRC1 siTGFβR1) and m TGFβR2 (n = 8; p < 0.0001 for PBS ctrl siRNA vs CTHRC1 ctrl siRNA, p < 0.0001 for CTHRC1 ctrl siRNA vs CTHRC1 siTGFβR2 and p = 0.0275 for PBS siTGFβR2 vs CTHRC1 siTGFβR2) in human beige adipocytes. n Luciferase activity in naïve and GPR180 knockout HEK-293T cells transfected with plasmid expressing 4× SMAD binding elements (SBE) upstream of luciferase in response to CTHRC1 and TGFβ1 stimulus for 18 h (n = 4; p < 0.0001 for both treatments in naïve HEK-293T cells and p = 0.0261 for TGFβ1 in GPR180 ablated HEK293Ts cells). o SMAD3 phosphorylation in beige hMADS cells treated with different concentrations of TGFβ1 in combination with CTHRC1 (n = 3; p < 0.0001 for 300 ng/ml, p = 0.0362 for 100 ng/ml and p = 0.0177 for 3000 ng/ml). p SMAD3 phosphorylation in response to TGFβ1 (300 pg/ml) in combination with increasing dose of CTHRC1 (n = 3; p = 0.0003 for 100 ng/ml, p = 0.0099 for 300 ng/ml, p < 0.0001 for 1000 ng/ml and p = 0.0004 for 3000 ng/ml). q Phosphorylation of SMAD3 in response to CTHRC1 and TGFβ1 (n = 3; CTHRC1 p = 0.0021, TGFβ 1 pg/ml p = 0.0281 and 10 pg/ml p < 0.0001). r Regulation of CTHRC1 in beige adipocytes following TGFβ1 treatment (n = 6; p < 0.0001 for both 0.1 and 1 ng/ml). Data are shown as average ±SEM. Statistical analysis was performed by two-sided Student’s t-test (j, o, p), one-way (e, h, i, q, r) or two-way ANOVA (c, d, f, k–m) with Dunnett’s or Tukey post-hoc test, respectively. Significant difference is indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. CTHRC1 Collagen triple helix repeat containing 1, DMSO dimethyl sulfoxide, GPR180 G protein-coupled receptor 180, hMADS human multipotent adipose-derived stem cells, IgG immunoglobulin G, n.d. not detected, PBS phosphate buffered saline, RFP red fluorescent protein, SMAD3 Mothers against decapentaplegic homolog 3, TGFβ1 Transforming growth factor β1, TGFβR1 Transforming growth factor β receptor type 1, TGFβR2 Transforming growth factor β receptor type 2.

Fig. 5. CTHRC1, an adipokine that requires GPR180 to enhance the beige adipocyte phenotype.

a Representative blots and b quantification of CTHRC1 protein in human adipocytes (n = 4; p = 0.0011) and c cell-conditioned media (n = 7; p = 0.0408). d Expression of CTHRC1 in supraclavicular brown and subcutaneous white adipose tissue of 6 healthy volunteers with detectable BAT (p = 0.3008). Effect of long-term (72 h) CTHRC1 treatment on e UCP1 protein (n = 4; p = 0.0001 for 100 ng/ml and p = 0.0064 for 500 ng/ml) and f mitochondrial respiration (n = 5; basal p = 0.0102, cAMP-stimulated uncoupled p < 0.0001 for 100 ng/ml and p = 0.0021 for 500 ng/ml, maximal p = 0.0207) in mature human beige adipocytes. g Effect of long-term (72 h) CTHRC1 (500 ng/ml) and TGFβ1 (1 ng/ml) on mitochondrial respiration in beige hMADS cells in presence of IgG (control) and neutralizing anti-TGFβ antibody (1 μg/ml) (n = 6; basal CTHRC1 IgG p = 0.0376, CTHRC1 anti-TGFβ p = 0.0489, TGFβ IgG p = 0.0474; uncoupled CTHRC1 IgG p = 0.0372, CTHRC1 anti-TGFβ p = 0.0079; coupled TGFβ IgG p = 0.0449; cAMP-stimulated uncoupled CTHRC1 IgG p = 0.0407, CTHRC1 anti-TGFβ p = 0.0440, TGFβ IgG p = 0.0317; maximal CTHRC1 IgG p = 0.0478, CTHRC1 anti-TGFβ p = 0.0310). h UCP1 levels (n = 6; siGPR180 in PBS p = 0.0016, CTHRC1 in ctrl siRNA p < 0.0001 and siGPR180 in CTHRC1 treatment p < 0.0001) and i mitochondrial respiration (n = 10; p < 0.0001 basal, p < 0.0001 coupled, p = 0.0094 for CTHRC1 treatment in ctrl siRNA and p = 0.0013 for siGPR180 in PBS treated in cAMP-stimulated uncoupled respiration, p < 0.0001 maximal respiration) in adipocytes following long-term (72 h) CTHRC1 treatment (500 ng/ml) in combination with GPR180 knockdown. Effect of CTHRC1 silencing on (j) UCP1 protein (ctrl siRNA n = 8, siCTHRC1 n = 9; p = 0.0002) and k mitochondrial oxygen consumption (n = 5; p = 0.0311 for uncoupled respiration, p = 0.0326 for cAMP-stimulated uncoupled respiration) in hMADS cells. Data are shown as mean ± SEM. Statistical analysis was performed by unpaired Student´s t-test (b, c, j), paired Student´s t-test (d), one-way ANOVA with Dunett’s post-hoc-test (e) or two-way ANOVA with Sidak and Tukey post-hoc tests (f–i, k). Significance is indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. cAMP cyclic adenosine monophosphate, CTHRC1 Collagen triple helix repeat containg 1, GPR180 G protein-coupled receptor 180, hMADS human multipotent adipose-derived stem cells, IgG imunoglobulin G, OCR oxygen consumption rate, PBS phosphate buffered saline, scBAT supraclavicular brown adipose tissue, TGFβ Transforming growth factor β, UCP1 Uncoupling protein 1, WAT white adipose tissue.

Fig. 6. CTHRC1 requires GPR180 to ameliorate metabolic disturbances in obesity.

Effect of AAV-mediated CTHRC1 overexpression on a body weight (p = 0.0174 week 10, p = 0.0191 week 11 and p = 0.0133 week 12), b fasting blood glucose (p = 0.03070 and c circulating FFA in 20-weeks-old male C57Bl/6 N mice challenged with HFD (n = 5). HSL phosphorylation (at Serine 660) in d iBAT and e iWAT of mice with increased circulating CTHRC1 (n = 5). Gene expression of selected brown adipocyte markers in f iWAT (p = 0.0280 for Ucp1, p = 0.0377 for Cidea and p = 0.0320 for Pgc1α) and g iBAT (p = 0.0111 for Cidea and p = 0.0191 for Cpt1β) of animals overexpressing CTHRC1 (n = 5). h UCP1 protein in iBAT following CTHRC1 treatment in male mice exposed to HFD. Effect of CTHRC1 overexpression on i energy expenditure (AUC p = 0.0014) and j respiratory exchange ratio (effect of CTHRC1 overexpression p = 0.0005) in C57Bl/6 N mice challenged with HFD AAV was injected prior to acclimatization in metabolic cages and measurement (n = 7). k Intraperitoneal glucose tolerance test in wild-type and GPR180 knockout mice fed with HFD and overexpressing stuffer or CTHRC1 (n = 8–9; p = 0.0160 for Gpr180^−/− vs WT at 30 min, p = 0.0011 for Gpr180^−/− vs WT at 60 min, p = 0.0020 for Gpr180^−/− vs WT at 90 min, p = 0.0366 for WT stuffer vs WT CTHRC1 at 90 min, p = 0.0041 for Gpr180^−/− vs WT at 120 min, p = 0.0160 for WT stuffer vs WT CTHRC1 at 120 min). Data are presented as mean ± SEM. Statistical significance was calculated using two-sided Student´s t-test (b–j) or two-way ANOVA with repeated measurements followed by Sidak post-hoc test (a, j, k). Area under the curve was calculated to compare energy expenditure in CTHRC1 overexpressing mice (i). Statistical differences are indicated as *p < 0.05; (k) * WT CTHRC1 vs WT stuffer, # WT stuffer vs GPR180^−/− stuffer, ^#p < 0.05, ^##p < 0.01 and ^###p < 0.001. AAV adeno-associated virus, AUC area under the curve, CIDEA Cell death inducing DFFA like effector a, CPT1β Carnitine palmitoyltransferase 1β, CTHRC1 Collagen triple helix repeat containing 1, DIO2 Iodothyronine deiodinase 2, ELOVL3 Fatty acid elongase 3, GPR180 G protein-coupled receptor 180, HFD high-fat diet, HSL Hormone sensitive lipase, HSP90 Heat shock protein 90, iBAT interscapular brown adipose tissue, iWAT inguinal white adipose tissue, NEFA non-esterified fatty acids, PGC1A PPARγ coactivator 1α, RER respiratory exchange ratio, UCP1 Uncoupling protein 1, WT wild-type.

Fig. 7. Regulation of GPR180 and CTHRC1 in human.

a Expression of GPR180 in WAT of normal weight men and individuals with obesity and normal glucose tolerance (NGT), obesity with prediabetes and obesity with diabetes (n = 57; p = 0.0021 for individuals with obesity and NGT, p = 0.0003 for individuals with obesity and prediabetes and p = 0.0013 for patients with obesity and diabetes). b Expression of GPR180 in paired samples of SAT and isolated adipocytes (n = 21; p = 0.0077). Correlation of adipocyte GPR180 level with c adipocyte size (n = 54), d suppression of fatty acid release during EHC (n = 55) and e resting energy expenditure (n = 31). f Incidence of circulating CTHRC1 in normal weight individuals, and patients affected by obesitybut NGT, patients having obesity and prediabetes and/or type 2 diabetes (n = 85; p = 0.02533) and its association with g basal energy expenditure and h energy expenditure measured during the steady state of EHC (n = 20). i Schematic illustration of the signalling mechanism identified in this study. Data are shown as mean ± SEM. Statistical analysis was performed by one-way ANOVA with Dunett post-hoc test (a), paired Student´s t-test (b) or Fisher´s exact test (f). For association of GPR180 expression in WAT or circulating CTHRC1 with metabolic parameters Pearson’s correlation coefficient was calculated (c–e, g, h). Significant differences are indicated as *p < 0.05, **p < 0.01 and ***p < 0.001. CTHRC1 Collagen triple helix repeat containing 1, EHC euglycaemic hyperinsulinaemic clamp, FFA free fatty acids, GPR180 G protein-coupled receptor 180, REE resting energy expenditure, SAT subcutaneous adipose tissue, SMAD3 Mothers against decapentaplegic homolog 3, TGFβ1 Transforming growth factor β1, TGFβR1 Transforming growth factor β receptor type 1, TGFβR2 Transforming growth factor β receptor type 2, UCP1 Uncoupling protein 1.