Adipose ABHD6 regulates tolerance to cold and thermogenic programs

Characterization of AT-ABHD6–KO mice at RT.

Chow diet–fed AT-ABHD6–KO mice did not differ in their body weight, food intake, as well as fat and lean mass from their controls (Supplemental Figure 1, G–I). Oral glucose tolerance test (OGTT) showed no differences between AT-ABHD6–KO, Cre, and Fl/Fl control mice in their glycemia or insulinemia (Supplemental Figure 1, J and K). However, AT-ABHD6–KO mice displayed better insulin sensitivity (P < 0.05) in insulin tolerance test (ITT) (Supplemental Figure 1L). Taken together, under standard diet and RT conditions, deletion of adipose ABHD6 in adult mice increases whole-body insulin sensitivity but has no effect on overall fat mass as well as glucose and energy homeostasis.

AT-ABHD6–KO mice show elevated energy expenditure under cold and are resistant to cold-induced hypothermia.

We further examined whether adipose ABHD6 displays a regulatory role in energy balance and adaptive thermogenesis that is autonomous to mature adipocytes, under cold-stress conditions. Control (Fl/Fl and Cre) and AT-ABHD6–KO mice were exposed to 4°C for 3 hours or 24 hours. While there was no difference in body weight and glycemia between KO and controls in all conditions (Supplemental Figure 2, A and B), rectal temperature was maintained more stably at a higher level in AT-ABHD6–KO mice than in controls, during the 3 hours cold exposure and after 24 hours cold exposure (Supplemental Figure 2, C and D). In order to continuously collect various metabolic parameters in a temperature-controlled environment, Fl/Fl, Cre, and KO mice were implanted with intraperitoneal temperature probes and were individually housed in metabolic cages (Comprehensive Lab Animal Monitoring System, CLAMS). Following 24 hours acclimatization at RT, respiration, food intake, core body temperature, and locomotor activity were recorded at RT for another 24 hours. This was followed by gradual decrease of cage temperature to 4°C over a 12-hour period, and metabolic parameters were then recorded at 4°C for 24 hours. When housed at RT, KO and control mice exhibited a similar and stable core body temperature. But under cold conditions, core body temperature dropped to a much lower level in control mice than in the KO mice (Figure 2, A and B). AT-ABHD6–KO mice displayed significantly higher oxygen consumption in light cycle and carbon dioxide production and energy expenditure (EE) in both light and dark cycles at 4°C (Figure 2, C–E), compared with the controls. There were no differences between the groups, in respiratory exchange ratio (Figure 2F), physical activity, and food consumption (Supplemental Figure 2, E and F). To eliminate RT-related cold stress, in another set of experiments, the CLAMS temperature was adjusted to 30°C for 5 consecutive days, and measurements were performed during the last 24 hours of the protocol. All the metabolic parameters were comparable between Fl/Fl and KO mice under thermoneutrality (Supplemental Figure 2, G and H), similar to the RT condition. Although it was previously reported that RT increases the basal metabolic rate in rodents (26), the intensity of the cold stress under 22°C appears to be insufficient to stimulate significant metabolic changes in the absence of ABHD6, especially given that the mice used in this study were born and housed under RT and adapted to their environmental condition.

Open in a separate window

Figure 2

AT-ABHD6–KO mice show elevated energy expenditure under cold and are resistant to cold-induced hypothermia.

(A) Core body temperature. (B) Average core body temperature over 24 hours (light + dark cycles). (C) Volume O₂ (V_O2). (D) Volume CO₂ (V_CO2). (E) Energy expenditure (EE; kcal/kg metabolic mass/h). (F) Respiratory exchange ratio (RER). n = 5–9 mice/group; 1-way ANOVA and Tukey’s post hoc test; Fl/Fl vs. KO: *P < 0.05, **P < 0.01; Cre vs. KO: ^#P < 0.05, ^##P < 0.01.

The in vivo studies using AT-ABHD6–KO mice generated by adipoq-Cre–driven ABHD6 deletion helped in understanding the overall significance of adipose ABHD6 in the control of cold-induced thermogenesis. However, as this model cannot differentiate the role of individual fat depots for in vivo measurements, we performed several ex vivo experiments to assess and better understand the effect of depot-specific ABHD6 deletion.

Glucose and oxidative metabolism are enhanced in iBAT from AT-ABHD6–KO mice under cold exposure.

Because adipose ABHD6 deletion potentiates adaptive thermogenesis under cold stress, we investigated the contribution of the iBAT fuel partitioning and metabolic responses. Glucose uptake was measured in vivo, using [¹⁸F]-fluorodeoxyglucose ([¹⁸F]-FDG), by μPET/CT, after exposing the mice (KO and Fl/Fl) at 30°C and 10°C for 48 hours. Cold exposure (10°C) markedly induced the fractional (Ki) and net uptake (Km) of [¹⁸F]-FDG in both KO and control mice, compared with thermoneutrality (30°C) (Figure 3, A and B). Although large variation observed in [¹⁸F]-FDG data among animals in both groups under cold and thermoneutral conditions, glucose uptake showed a trend toward increased levels in KO versus Fl/Fl mice under cold (P = 0.1) (Figure 3, A and B). Notably, the changes were more evident based on the PET/CT images, which demonstrated an intense focal uptake of [¹⁸F]-FDG in iBAT of AT-ABHD6–KO mice in response to cold, compared with the control (Figure 3C). Glucose oxidation (after 3 hours cold induction) and glucose utilization (after 24 hours cold induction), measured ex vivo, were enhanced in iBAT from KO versus Fl/Fl mice (P < 0.05; Figure 3D). The tissue content and extracellular release of lactate, an indicator of glycolysis, were also significantly elevated by 2-fold in iBAT from KO mice (P < 0.05; Figure 3E). Also, the extracellular acidification rate (ECAR), an index of glycolysis, measured using the Seahorse system, was elevated in KO mice after cold exposure (P < 0.05; Figure 3F).

Open in a separate window

Figure 3

Glucose and oxidative metabolism are enhanced in iBAT from AT-ABHD6–KO mice under cold exposure.

(A) Fractional (Ki) [¹⁸F]-FDG uptake, determined by PET/CT. (B) Net (Km) [¹⁸F]-FDG uptake, determined by PET/CT. (C) Representative sagittal view of PET/CT images following 10 MBq [¹⁸F]-FDG injection. The red arrows indicate iBAT localization. (D) Ex vivo glucose oxidation (¹⁴C-glucose) and utilization (³H-glucose). (E) Tissue lactate content and ex vivo lactate release from tissue explants. (F) Extracellular acidification rate (ECAR) on tissue explants. (G) Oxidative metabolism index, analyzed by [¹¹C]-acetate PET tracer (P = 0.03). (H) Mean standard uptake value (SUV) of [¹¹C]-acetate over the first 500 seconds of acquisition after tracer injection (P = 0.01). (A and B) n = 15 mice/group; (D–H) 5–16 mice/group. (A, B, and G) Student’s t test; (D–F) 1-way ANOVA and Tukey’s post hoc test; (H) 2-way ANOVA and Tukey’s post hoc test. The effects of genotype: *P < 0.05; the effects of cold induction: ^##P < 0.01, ^####P < 0.0001.

PET/CT studies revealed no changes in the Ki and Km of [¹⁸F]-FDG in visceral perigonadal WAT (gWAT) (Supplemental Figure 3, A and B) and subcutaneous inguinal WAT (iWAT) (data not shown) in the KO mice, compared with controls, regardless of temperature conditions. This is in agreement with gWAT ex vivo glucose metabolism, showing no difference between KO and Fl/Fl mice (Supplemental Figure 3C). However, insulin-stimulated ex vivo glucose uptake in gWAT was significantly elevated in cold-exposed KO mice, compared with controls (Supplemental Figure 3D), and this was in line with the increased glucose transporter type 4 (Glut4) expression in gWAT from cold-induced KO mice (Supplemental Figure 3E).

The PET analysis using [¹¹C]-acetate, to measure iBAT mitochondrial oxidative capacity in vivo (16), indicated that even though iBAT oxidative metabolism was induced by cold in both KO and Fl/Fl mice, the increase was significantly higher (P < 0.05) in AT-ABHD6–KO mice (Figure 3G). Accordingly, the [¹¹C]-acetate uptake kinetics curve (Figure 3H) showed that iBAT from KO mice exhibited higher uptake of [¹¹C]-acetate in response to cold (P < 0.05). Next, we investigated if there are any changes in the mitochondrial biogenesis and function in iBAT. Measurement of mitochondrial respiratory complexes in iBAT showed that the expression of different subunits from complexes I–V was similar in AT-ABHD6–KO and Fl/Fl mice (Supplemental Figure 3F). In addition, isolated iBAT mitochondria were examined for substrate-driven respiration. The oxygen consumption rate (OCR) in the iBAT mitochondria in response to pyruvate/malate (state 2), ADP (state 3), and trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP) (maximal) were comparable between the groups at RT; nevertheless, all these parameters were elevated in both KO and Fl/Fl mice in response to cold stress (Supplemental Figure 3G). Finally, no differences were observed between KO and Fl/Fl mice in the in vivo oxidative activity of gWAT and iWAT, measured using [¹¹C]-acetate (data not shown). Thus, the results show that adipose ABHD6 deletion results in enhanced glucose uptake, glycolysis, and oxidative metabolism in iBAT without apparent difference in the white fat depots.

AT-ABHD6–KO mice show induction of cold-adaptive thermogenesis mechanisms in gWAT distinct from enhanced adipose UCP1 expression.

Adaptive thermogenesis is generally considered to involve the classical UCP1-dependent mechanisms of BAT activation and WAT browning (3, 4). Besides, our earlier work in whole-body ABHD6-KO mice indicated a potential thermogenic role for UCP1, which was upregulated in all the adipose tissues of KO mice (22). In order to examine the role of UCP1 in the enhanced adaptive thermogenesis seen in the AT-ABHD6–KO mice, we measured UCP1 mRNA and protein levels in fat tissues from animals kept at 30°C or RT and in cold-challenged mice (4°C). Because 3 hours of cold exposure is reported to be sufficient to promote BAT UCP1-dependent metabolism (27), in addition to 24 hours of cold induction, a group of mice were kept 3 hours at 4°C for assessing the effects of acute cold induction. As expected, UCP1 expression was upregulated in fat depots following cold stress in both control and KO mice, in agreement with the known function of UCP1 in adaptive thermogenesis (Figure 4, A and B, and Figure 5, A–C). Our data indicated that cold-mediated adipose UCP1 induction was more prominent after 24 hours versus 3 hours of exposure in both genotypes (Figure 5, A–C). In support of these results, previous reports have also shown that longer duration of cold is essential for restoration of total adipose thermogenic capacity (28). Immunohistochemistry results indicated slightly higher UCP1 protein expression in ABHD6-deficient iBAT at RT (Figure 4A); however, no significant differences were found after 3 hours or 24 hours of cold induction in either mRNA or protein levels in all 3 fat tissues (Figure 4, A–C, and Figure 5, A–C). To eliminate the possible thermogenic activity of adipose depots under RT, we also assessed the thermogenic markers in mice housed at thermoneutrality (30°C) for 5 days. Under thermoneutral conditions UCP1 expression was much lower in iBAT and gWAT, and there were no differences between Fl/Fl and KO groups (Supplemental Figure 4, A and B).

Open in a separate window

Figure 4

Adipose-specific deletion of ABHD6 does not alter UCP1 expression.

(A) UCP1 immunostaining. Scale bar: 100 μm. (B) UCP1 protein level. (C) UCP1 and PGC1α protein levels.

Open in a separate window

Figure 5

AT-ABHD6–KO mice show induction of cold-adaptive thermogenesis mechanisms in gWAT distinct from enhanced adipose UCP1 expression.

(A–C) Ucp1, Pgc1a, Ppara, and Pparg mRNA expression. n = 6–18 mice/group; 1-way ANOVA and Tukey’s post hoc test. The effects of genotype: *P < 0.05, **P < 0.01; the effects of cold induction: ^##P < 0.01, ^####P < 0.0001.

We further analyzed the expression of other genes (Pgc1a, Ppara, and Pparg) known to be involved in WAT browning, BAT activation, and thermogenesis (29). Similar to Ucp1, Pgc1a was also significantly downregulated at 30°C, and there were no differences between KO and Fl/Fl mice (Supplemental Figure 4, A and B). However, in response to cold exposure, PGC1α protein expression was upregulated markedly in the gWAT from KO mice (Figure 4C). Pgc1a mRNA level was elevated under cold in all 3 fat depots of both genotypes, but AT-ABHD6–KO mice exhibited significantly higher levels (3 hours and 24 hours of cold induction), compared with the Fl/Fl mice (Figure 5, A–C). PPARs were similarly expressed in the gWAT of KO and Fl/Fl mice under both 30°C and 22°C conditions (Figure 5A and Supplemental Figure 4, A and B). Expression of both Ppara and Pparg was decreased in gWAT of Fl/Fl mice upon 24 hours cold stress, whereas no such decrease was seen in the KO mice (Figure 5A). Thus, in concert with PGC1α, after 24 hours cold exposure, there was a significantly elevated level of Ppara and Pparg mRNA in the gWAT of AT-ABHD6–KO versus Fl/Fl mice (Figure 5A). No change was observed in iWAT and iBAT’s expression of PPARs between AT-ABHD6–KO and Fl/Fl under any temperature condition (Figure 5, B and C, and Supplemental Figure 4, A and B). Thus, the above results show that the increased whole-body thermogenic response of the AT-ABHD6–KO mice to cold is associated with elevated levels of PGC1α, PPARα, and PPARγ in gWAT, but without any concomitant changes in UCP1 level, compared with Fl/Fl mice.

Smaller gWAT adipocytes and iBAT lipid droplets in cold-exposed AT-ABHD6–KO mice.

In response to hypothermia, adipose tissue is rapidly remodeled with reduction of adipocyte size and lipid droplets. Histological analysis of adipose tissues from Fl/Fl and KO mice showed that cold exposure per se increased the proportion of smaller sized adipocytes in gWAT and of smaller sized lipid droplets in iBAT, with no apparent differences in iWAT (Supplemental Figure 4, C–E). Quantitation of images of H&E-stained tissue sections indicated a greater frequency of smaller adipocytes in gWAT (but not in iWAT) and lipid droplets in iBAT of cold-exposed KO mice versus control (Supplemental Figure 4, F–H), which corresponded to significantly smaller average adipocyte area in gWAT, as well as diminished average lipid droplet size in iBAT (Supplemental Figure 4I). There was a reduction in the size of adipocytes in the iWAT of KO mice only at RT (Supplemental Figure 4I). Following 24 hours cold exposure, there was a decrease in the weights of gWAT, iWAT, and iBAT depots in both genotypes (Supplemental Figure 4J).

AT-ABHD6–KO mice show enhanced antiinflammatory markers in gWAT.

White adipocyte size is a strong determinant of lipolysis responsiveness, and gWAT adipocyte size, in particular, is known to positively associate with insulin resistance and inflammation (30). The reduced gonadal adipocyte size in KO mice is in line with the changes in inflammatory markers and insulin sensitivity in these mice. Antiinflammatory marker Arg1 expression was induced in the gWAT from KO mice at RT (Supplemental Figure 4K). After 24 hours cold induction, Arg1, Il10, and Adipoq mRNA levels were significantly higher in the ABHD6-deficient gWAT versus Fl/Fl mice (Supplemental Figure 4, K and L). Besides, the cold-induced increase in adiponectin receptor-2 expression in gWAT was much higher in AT-ABHD6–KO mice than in Fl/Fl mice (Supplemental Figure 4L). In the iBAT, the mRNA expression of Il10, Il4, and Arg1 was almost undetectable (data not shown).

Increased cold-stimulated activity of the lipolytic arm of the GL/FFA cycle in gWAT of AT-ABHD6–KO mice.

GL/FFA cycle has been proposed as a heat-producing and metabolic signal-generating pathway in many cell types (7, 10, 31). Generally, WAT and BAT respond differently under stimulated conditions (32). In WAT there is enhanced lipolytic release of FFAs, which are supplied to BAT for oxidation and heat production and to other energy-demanding organs (33). Because deletion of adipose ABHD6, which besides MAGL catalyzes the last step of the lipolysis arm of the GL/FFA cycle, led to higher energy expenditure and better cold tolerance, we examined the thermogenesis-regulatory role of ABHD6 and how it is related to the GL/FFA cycle in WAT and/or BAT. We first assessed whether the lipolytic arm of this cycle is enhanced in various adipose depots in association with the expression and activity of the lipolytic enzymes.

Twenty-four–hour cold stress induced ex vivo lipolysis in gWAT from both Fl/Fl and KO mice, measured as glycerol and FFA release. Cold-stimulated glycerol and FFA release was higher in ABHD6-ablated gWAT (P < 0.05), even though the basal levels at RT remained unchanged (Figure 6A). Thus, while acute (2 hours) ISO stimulation led to lowered lipolysis in ABHD6-deleted gonadal adipocytes versus control (Figure 1, C and F), 24 hours cold exposure enhanced the lipolysis in the KO gWAT (Figure 6A). Cold exposure of KO mice for 24 hours led to altered expression of GL/FFA cycle genes, as seen below, which could not have happened during acute ISO treatment ex vivo. Consistently, expression of lipolytic enzymes ATGL, HSL, and MAGL, at mRNA and protein levels, was elevated in gWAT from cold-induced AT-ABHD6–KO mice (Figure 6, C and D). Also, ABHD5 protein, an activator of ATGL, and activating phosphorylation of HSL (Ser563) were increased in the gWAT of KO mice (Figure 6D). Expression of aquaglyceroporin 7, the major glycerol transporter in adipocytes, was also increased at mRNA level in gWAT from cold-treated AT-ABHD6–KO mice (Figure 6G), probably to facilitate glycerol efflux during accelerated lipolysis. Lipolysis measured as glycerol and FFA release from explants of iBAT of the KO versus Fl/Fl mice was slightly reduced at RT and under cold (Figure 6B). Expression of most of the lipolytic enzymes in iBAT was unchanged between KO and Fl/Fl mice at RT and after cold (Figure 6, E and F). Notably, cold exposure led to elevated mRNA expression of MAGL (Mgll) in iBAT from both groups (Figure 6E). Also, it is interesting to note that ABHD6 mRNA and protein expression in control mice was markedly induced by cold in iBAT but not in gWAT (Supplemental Figure 5, A and B). Plasma analysis showed modest reductions in TG, glycerol, and NEFAs in KO mice at RT, with no apparent changes between the control and KO mice under cold (Figure 6H). Finally, as shown in Supplemental Figure 5C, gWAT and iBAT 18s levels (used as a reference gene in this study) revealed stable mRNA levels across different experimental conditions (RT vs. cold).

Open in a separate window

Figure 6

Increased cold-stimulated activity of the lipolytic arm of the GL/FFA cycle in gWAT of AT-ABHD6–KO mice.

(A and B) Ex vivo glycerol and NEFA released from AT explants (7 mice/group). (C–F) Expressions of lipolytic enzymes at mRNA level (6–12 mice/group) and at protein level (4 mice/group). (G) Aquaporin 7 (Agp7) mRNA expression. (H) Plasma TG, glycerol, and NEFA levels (11–12 mice/group). Statistics were calculated by 1-way ANOVA and Tukey’s post hoc test, and for C–F to compare Fl/Fl vs. KO, Student’s t test was used. The effects of genotype: *P < 0.05, **P < 0.01, ***P < 0.001; the effects of cold induction: ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001.

Adipose-specific deletion of ABHD6 does not alter β-adrenergic receptor 3–mediated lipolysis.

In order to further unravel the signaling pathways involved in enhanced expression and activation of lipases in the gWAT of cold-exposed AT-ABHD6–KO mice, we assessed the role of β-adrenergic receptors (β-ARs). β-AR3, encoded by Adrb3, is expressed predominantly in the fat tissues and is involved in WAT browning and BAT activation mostly through UCP1-dependent mechanisms (34). To investigate if stimulation of β-AR3–mediated lipolytic responses at RT can recapitulate the cold-induced changes in the lipolysis pathway, AT-ABHD6–KO and Fl/Fl mice were injected with a selective β-AR3 agonist ({"type":"entrez-nucleotide","attrs":{"text":"CL316243","term_id":"44896132","term_text":"CL316243"}}CL316243; 1 mg/kg/d). Plasma samples were collected at 0, 15, and 30 minutes of the first injection (acute response) and at the end of the protocol (2 days). Plasma glucose level decreased similarly in both groups by acute CL treatment (Supplemental Figure 5D). As expected, circulating glycerol, NEFA, and TG concentrations were all increased within 30 minutes postinjection in both groups, though a significantly lower level of TG was observed in the KO mice (Supplemental Figure 5, E–G). No significant differences were found in the ex vivo lipolysis either in gWAT or in iBAT (Supplemental Figure 5, H and I) between CL-treated AT-ABHD6–KO versus Fl/Fl mice. Consistently, plasma parameters (TG, glycerol, and NEFA) measured at the end of the protocol (2-day treatment) were comparable between groups (Supplemental Figure 5J). Under cold-stress condition, among the 3 β-ARs, only Adrb2 mRNA level was significantly increased in the gWAT of AT-ABHD6–KO mice (Supplemental Figure 5K). Nevertheless, similar Adrb3 mRNA levels between KO and Fl/Fl (Supplemental Figure 5, K and L) and comparable responses to the CL treatment in both genotypes suggest that β-AR3 signaling is likely not altered in the KO mice. However, whether the increase in gWAT Adrb2 expression in cold-exposed ABHD6-KO mice is under the influence of the SNS, endocrine factors, or adipocyte autonomous signaling (MAG/PPAR-dependent) is not clear. Thus, we further investigated whether adipocyte β-AR expression is dependent on the activity of PPARs, because we found elevated Ppara and Pgc1a expression in gWAT and our earlier study indicated increased function of PPARs in whole-body ABHD6-KO mice (22). Differentiated 3T3-L1 adipocytes were treated with WY14643 (WY) and pioglitazone (Pio), agonists of PPARα and PPARγ, respectively, and the expression of β-ARs was assessed. Interestingly, only Adrb2 expression was significantly induced in adipocytes treated with PPARα agonist (P < 0.001; Supplemental Figure 5M), which is in line with the cold exposure data. Adrb3 mRNA expression was similar between DMSO and PPARα agonist–treated cells (Supplemental Figure 5M), whereas Adrb1 mRNA was not detectable in 3T3-L1 adipocytes (data not shown). Taken together, the data suggest that enhanced cold tolerance with adipose ABHD6 deletion occurs independently of changes in β-AR3 agonism and that it might be conducted via a PPARα-β-AR2-GL/FFA cycle axis.

Increased cold-stimulated activity of the lipogenesis arm of the GL/FFA cycle in gWAT of AT-ABHD6–KO mice.

As thermogenesis by GL/FFA cycling depends on both lipolysis and lipogenesis arms of this pathway, we examined if lipogenesis per se and expression of lipogenic genes is also altered in the KO mice under cold stress. Glycerol-3-phosphate acyltransferase 3 (Gpat3), diacylglycerol acyltransferase 1 (Dgat1), Lipin 1 (Lpin1), and lysophosphatidic acid acyltransferase (Agpat2) mRNAs were all upregulated in the gWAT from cold-exposed AT-ABHD6–KO mice (Figure 7A). Additional genes that contribute to GL/FFA subcycles, including lipid phosphate phosphatases (Lpp3), diacylglycerol kinase δ (Dgkd), and acylglycerol kinase (Agk), were similarly expressed in both Fl/Fl and KO mice at RT and after cold stress (Figure 7B). In accordance with previous reports (35), glycerol kinase (Gk) expression in gWAT was induced in response to cold, without difference between the 2 groups of mice (Figure 7B). Figure 7C summarizes all the observed changes in the expression of GL/FFA cycle and subcycle enzyme genes in gWAT of cold-challenged mice. The expression of the lipogenic enzymes glycerol-3-phosphate acyltransferase, mitochondrial (Gpam); Gpat3; and Dgat1 and subcycle enzyme Gk in iBAT were increased in both KO and Fl/Fl mice after cold exposure (Figure 7, D and E). Expression of Gk and Agpat2 was slightly but significantly increased in the KO versus Fl/Fl iBAT after cold exposure (Figure 7, D and E).

Open in a separate window

Figure 7

Increased cold-stimulated activity of the lipogenesis arm of the GL/FFA cycle in gWAT of AT-ABHD6–KO mice.

(A and B) Gene expressions of lipogenic enzymes (6–18 mice/group). (C) GL/FFA cycle and subcycle enzyme gene expression levels in gWAT from cold-challenged mice. Red arrows correspond to the upregulated enzymes in the KO mice (vs. Fl/Fl), black arrows refer to enzymes with similar mRNA expressions between genotypes, and the blue arrow indicates the upregulation of the enzymes in response to cold, independent of genotype. (D and E) Gene expressions of GL/FFA subcycle enzymes (6–18 mice/group). (F) Ex vivo de novo lipogenesis (¹⁴C-glucose). Statistics by 1-way ANOVA and Tukey’s post hoc test, except (A–D) Student’s t test to compare Fl/Fl vs. KO. The effects of genotype: *P < 0.05, **P < 0.01; the effects of cold induction: ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001.

Notably, in iBAT from both genotypes, cold exposure led to elevated mRNA expression of Mgll, some lipogenic enzymes, and Gk, which suggests increased intracellular reusage of lipolytic glycerol in iBAT. In agreement with the gene expression changes, ex vivo de novo lipogenesis from glucose was significantly higher in the gWAT from the cold-exposed KO mice versus Fl/Fl (P < 0.01; Figure 7F). Although lipogenesis in BAT explants was induced by cold, no difference was found between genotypes, regardless of temperature (Figure 7F). Moreover, there were no changes between groups following 2-day treatment at RT with β-AR3 agonist CL in body weight, rectal temperature, blood glucose, or gWAT and iBAT de novo lipogenesis (Supplemental Figure 6, A–D).

Cold exposure

KO, Cre, and Fl/Fl male mice at the age of 14–16 weeks were randomly divided into 2 groups, one kept at RT and the other in the cold. The cold group mice were kept at 4°C for either 3 hours or 24 hours. For acute cold exposure (3 hours), mice were singly housed in prechilled cages without food and bedding but with free access to water. During 24 hours cold induction, mice were individually housed with minimum bedding on a 12-hour light/12-hour dark cycle with ad libitum access to chow diet and water. Blood samples were collected and rectal temperature was recorded.

β-AR3 stimulation by {"type":"entrez-nucleotide","attrs":{"text":"CL316243","term_id":"44896132","term_text":"CL316243"}}CL316243 treatment

KO and Fl/Fl male mice (14–16 weeks old) were treated with {"type":"entrez-nucleotide","attrs":{"text":"CL316243","term_id":"44896132","term_text":"CL316243"}}CL316243 (MilliporeSigma; 1 mg/kg/d) by intraperitoneal injection for 2 days. Blood samples were collected and rectal temperature was recorded before and after treatment. For acute treatment, mice were fasted overnight, and blood was collected from the tail vein before and at 15 and 30 minutes after {"type":"entrez-nucleotide","attrs":{"text":"CL316243","term_id":"44896132","term_text":"CL316243"}}CL316243 injection.

In vivo metabolic studies

Mouse adipose depots

For further analysis, gWAT, iWAT, and iBAT fat depots were dissected (Supplemental Figure 1D).

Cell culture and treatment

Ex vivo metabolic studies

All the ex vivo metabolic studies were done on 14- to 16-week-old male Fl/Fl and KO mice.

Mitochondrial preparations and oxygen consumption

iBAT mitochondria were prepared from freshly dissected tissue (63), and the protein content was measured. Mitochondria were suspended in an assay medium (MAS: 70 mM sucrose, 220 mM mannitol, 5 mM KH₂PO₄, 5 mM MgCl₂, 2 mM HEPES, 1 mM EGTA, 0.2% fatty acid–free BSA, pH 7.2), and 10 μg of mitochondrial protein (in 50 μL) was loaded per well in Seahorse XF24 assay plates. To promote adherence, the plate was centrifuged at 4°C, 20 minutes, at 2000g, and 475 μL of MAS was carefully delivered to each well without disturbing the settled mitochondria. The plate was placed at 37°C for 4 minutes to equilibrate the temperature and then transferred to the XF24 Analyzer (Seahorse Bioscience). To measure the OCR, the following substances were injected into the wells, sequentially (final concentrations): 5 mM pyruvate and 5 mM malate, 3.5 mM ADP, 4 μM FCCP, and 5 μM each antimycin A and rotenone.

Analysis of MAG species

Analysis of 1- and 2-MAG species was done as described previously (20). In brief, total lipids from adipose samples were extracted by Folch method. 1-MAG and 2-MAG were separated on thin-layer chromatography plates, and the corresponding bands were scraped and processed for FFA species analysis by HPLC. Individual quantities of MAG species were assessed from the measured FFA.

Nuclear isolation and MAG analysis

Nuclei were purified from freshly dissected gWAT (64), with slight modifications. Minced adipose tissue was added to 4 volumes (w/v) of 0.3 M sucrose buffer (0.3 M sucrose, 20 mM KH₂PO₄, 20 mM Na₂HPO₄ at pH 7.2, 137 mM NaCl, 3 mM KCl, 0.02% Triton X-100), plus protease and phosphatase inhibitors (1 mM phenylmethylsulfonyl fluoride, 2 μg/mL aprotinin, 0.7 μg/mL pepstatin, 0.5 μg/mL leupeptin, and 10 mM sodium orthovanadate), and homogenized in an ice-cold Dounce homogenizer. The homogenate was filtered through a 70 μm cell strainer (Corning Life Sciences) and then centrifuged through a 1.4 M sucrose cushion. The nuclear pellet underneath the sucrose layer was gently resuspended in about the same volume of 0.3 M sucrose buffer. The purity of isolated nuclear fraction was determined using immunoblotting for specific organelle markers, such as lamin A/C (nuclear), GAPDH (cytosolic), and CII/CIII (mitochondrial). Isolated nuclei were used for lipid extraction by Folch method, followed by separation of MAG species and HPLC quantification, as detailed above.

Tissue histology and analysis

Freshly dissected adipose tissues were fixed (10% paraformaldehyde), dehydrated, embedded in paraffin, and sectioned. Sections were stained with H&E and examined for UCP1 expression (anti-UCP1 antibody from Abcam). Images were processed for cell size and number using ImageJ Version 1.48v (National Institutes of Health, Bethesda, Maryland, USA). Utilizing a semiautomated method (65), the average area, frequency distribution, and number of adipocytes in the gWAT were estimated. iBAT lipid droplet size was also assessed. Briefly, images were converted to 8-bit gray scale with the threshold adjusted to mid gray scale range, converted to binary, and segmented with watershed plug-in; and finally, the lipid droplet particles were analyzed automatically.

Immunofluorescence microscopy

3T3-L1 adipocytes were serum-starved (3 hours), fixed (4% paraformaldehyde), permeabilized, (100 μM digitonin), and incubated with anti-ABHD6 antibody (Cell Signaling Technology) and Alexa Fluor 488–conjugated secondary antibody (Thermo Fisher Scientific) (antibodies are all listed in Supplemental Table 1). Pictures were taken using fluorescence microscope and analyzed by ImageJ Version 1.48v (National Institutes of Health, Bethesda, Maryland, USA).

Immunoblotting

Tissues were digested with lysis buffer (20 mM Tris-HCl, pH 7.2, containing 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% v/v Triton X-100, 0.1% SDS, and protease inhibitors), and after protein quantification, 25–40 μg protein was used for Western blot analysis. Antibodies are listed in Supplemental Table 1.

RNA extraction and quantitative PCR

Total RNA was isolated from tissues/cells using the RNeasy Mini Kit (Qiagen). Following reverse transcription of 2 μg RNA to cDNA, gene expression was determined by quantitative PCR using SYBR Green. All gene expression analyses were run in duplicate and normalized to 18s. Primer sequences are listed in Supplemental Table 2.

Plasma analysis

Glycerol, NEFAs, and triglycerides were measured using commercially available kits (Supplemental Table 1).

Statistics

Results are expressed as means ± SEM. All data were analyzed and figures were prepared using GraphPad Prism 6 (GraphPad Software). Statistical differences between 2 groups were assessed by unpaired, 2-tailed Student’s t test and between multiple groups using 1-way or 2-way ANOVAs, as indicated. A P value of less than 0.05 was considered statistically significant.

This study was supported by funds from Canadian Institutes of Health Research (to MP and SRMM) and the Intercentre Collaborative Projects from the Research Network on Cardiometabolic Health, Diabetes and Obesity of the Fonds de Recherche du Québec — Santé (to MP and ACC). MP holds the Canada Research Chair in Diabetes and Metabolism. PP is a recipient of Montreal Diabetes Research Center, CRCHUM Grand Labo Postdoctoral, and Mitacs Elevate Fellowships. ACC is the recipient of the Canada Research Chair in Molecular Imaging of Diabetes. We thank the Cellular Physiology & Metabolomics and Rodent Phenotyping core facilities and the animal house facility of CRCHUM. We acknowledge the Sherbrooke Molecular Imaging Center for μPET/CT acquisitions and analyses and thank Jean-François Beaudoin, Maxime Paillé, and Mélanie Archambault for their technical assistance and Ouhida Benrezzak for her support with the ethical committee. We thank Stefan Offermanns (Max-Planck-Institute, Bad Nauheim, Germany), J. Mark Brown (Cleveland Clinic, Cleveland, Ohio, USA), and Jennifer Estall (University of Montreal) for the Adipoq-Cre mice, ABHD6 antibody, and PGC1α antibody, respectively. We also gratefully acknowledge Gareth Lim (University of Montreal), J. Mark Brown (Cleveland Clinic), Richard G. Kibbey (Yale University, New Haven, Connecticut, USA), and Lawrence Kazak (McGill University) for helpful discussions. The graphical abstract was created with BioRender.com.