Lipolysis in Brown Adipocytes Is Not Essential for Cold-Induced Thermogenesis in Mice

The Stimulated Lipolysis Is Abolished in CGI-58-Deficient iBAT

As mentioned under the Introduction, CGI-58 involves at least two steps of cytosolic LD lipolysis: i) interacting with LD coat proteins (perilipins), and ii) activating ATGL as a coactivator. Consequently, CGI-58 and ATGL mutations cause overlapping, yet distinct phenotypes (Fischer et al., 2007; Lord et al., 2016). In test tubes, CGI-58-deficient iBAT homogenates had significantly reduced lipase activity, which was further suppressed, though not significantly, by an ATGL inhibitor Atglistatin (Figure S1C). This finding was not surprising given the presence of ATGL protein in the homogenate of CGI-58-deficient iBAT (Figure S1D) and relatively low levels of CGI-58 expression in other cell types that constitutes ∼50% of BAT (Rosenwald et al., 2013). Consistent with a previous study showing a major contribution of hormone sensitive lipase (HSL) to the in vitro lipase activity of adipose tissue homogenates (Schweiger et al., 2006), the predominant lipase activity in the homogenates of CGI-58-deficient and control iBATs was from HSL (Figure S1E), which is capable of hydrolyzing TG in vitro (Sztalryd et al., 1995). Despite abundant HSL and ATGL proteins in the homogenate of CGI-58-deficient iBAT (Figure S1D), CGI-58 deficiency, like ATGL or HSL inhibitor treatment, completely abolished forskolin-stimulated lipolysis in live iBAT explants (Figure 1J and S1F) demonstrating an essential role of CGI-58, ATGL, or HSL in mediating the stimulated lipolysis.

BAT-KO Mice Have Normal Energy Expenditure and Body Temperature at Room Temperature or Thermoneutrality

A major function of BAT is to dissipate energy as heat. However, BAT-KO mice only showed a minor reduction in total energy expenditure at room temperature (22°C) (Figure S1G). There were no significant differences in oxygen consumption, respiratory exchange ratio (RER, indicative of metabolic substrate preference), total physical activity, food intake, and core body temperature between BAT-KO and control mice housed at either room temperature or thermoneutrality (30°C) (Figure S1G). As expected, thermoneutrality reduced total energy expenditure and oxygen consumption and increased RER in normal mice. Similar responses were observed in BAT-KO mice (Figure S1G). These findings demonstrate that LD lipolysis in BAT has limited impact on whole-body energy balance and thermogenesis at room temperature or thermoneutrality.

BAT-KO Mice Are Not Cold Sensitive

At room temperature, BAT-KO mice did not develop hypothermia even during fasting (Figure 2A). To determine whether defective LD lipolysis in BAT influences cold-induced thermogenesis under different nutritional status, we subjected chow-fed BAT-KO mice to acute cold exposure (4°C) and monitored changes in intra-rectal temperatures. BAT-KO mice were not cold sensitive when food was absent (Figure 2B). Relative to littermate controls, these animals had higher body temperatures when they were provided food during acute cold exposure (Figure 2C). Similar responses were observed in HFD-fed mice when the core body temperature was continuously monitored by Telemetry (Figure 2D and 2E). During chronic cold exposure when mice had free access to food and water, the core body temperature was significantly higher in the first 2 days, and stayed at a slightly higher level thereafter in BAT-KO mice than controls (Figure 2F). In addition, relative to littermate controls, BAT-KO mice had higher total energy expenditure during chronic, though not acute, cold exposure (Figure S2A), which was not a result of increased physical activity (Figure S2B).

Open in a separate window

Figure 2

BAT-KO Mice Are Not Cold Sensitive

(A) Intrarectal temperatures of 20-week-old HFD-fed BAT-KO (KO) and control (Con) mice at 22°C. n = 4-5/group.

(B, C) Intrarectal temperatures of 10-11-week-old chow-fed mice exposed 4°C in the absence or presence of diet. n = 5/group.

(D) Core body temperatures of 28-week-old HFD-fed mice exposed to 4°C in the absence of food. n = 5-6/group.

(E) Core body temperatures of 24-week-old HFD-fed mice exposed to 4°C in the presence of food. n = 5-6/group.

(F) Core body temperatures of 18-week-old HFD-fed mice during 7 days of cold exposure. n = 6/group.

CGI-58 Deletion in BAT Does Not Reduce Whole-Body Thermogenic Capacity

It was reported that iBAT UCP1 expression was reduced in mice lacking ATGL in all adipose tissues (Ahmadian et al., 2011). However, we did not observe a significant reduction of UCP1 protein in the iBAT of BAT-KO mice on chow (Figure 3A). Due to a ∼50% increase of the total protein per iBAT in BAT-KO mice relative to the controls (Figure 1G), the total amount of UCP1 protein per iBAT was actually increased in these mice. To gain additional information as to why mice deficient in BAT lipolysis are not cold sensitive, we measured iBAT expression levels of UCP1 and other genes relevant to thermogenesis in mice treated with the β3 agonist CL316,243 for 4 days or cold for 7 days. Under the basal condition, i.e., at room temperature without β3 agonist or cold treatment, iBAT UPC1 mRNA and protein levels did not differ between the two genotypes (Figure 3B-3E). The protein expression of tyrosine hydroxylase (TH), a marker of sympathetic innervation (Foster and Bartness, 2006), increased substantially in the CGI-58-deficient iBAT (Figure 3D) suggesting augmented sympathetic drive, which may partly explain its hyperplasic phenotype (Figure 1F) and reserved UCP1 expression. CGI-58 ablation in BAT also had no effects on iBAT levels of mRNAs for peroxisome proliferator-activated receptor (PPAR)-α, PR domain-containing 16 (PRDM16), and PPAR-γ, but increased iBAT levels of mRNAs for PPAR-γ coactivator (PGC)-1α and type II iodothyronine deiodinase (Dio2) (Figure 3B). Despite this, UCP1 protein expression in the CGI-58-deficient iBAT did show a blunted response to β3 agonist (Figure 3C) or cold (Figure 3D), which was not correlated with levels of mRNA for β3 adrenergic receptor (Figure S3A). Multiple transcriptional factors are implicated in transcriptional regulation of UCP1, including fatty acid-activated PPARα and PGC-1α. The response of PPARα and PGC-1α to β3 agonist was also attenuated (Figure 3B). CGI-58 deficiency may suppress PPAR signaling in BAT by limiting PPAR ligands from cytosolic LD lipolysis as seen in the ATGL-deficient heart (Haemmerle et al., 2011). Nonetheless, attenuated UCP1 expression under cold did not reduce total energy expenditure (Figure S2A).

Open in a separate window

Figure 3

CGI-58 Deletion in BAT Does Not Reduce Whole-body Thermogenic Capacity

(A) Western blots and densitometry of iBAT UCP1 protein in 14-week-old chow-fed mice.

(B) Relative iBAT expression levels of thermogenesis-related genes from the mice injected with CL316,243 for 4 days (CL) or vehicle (Saline). n = 5/group.

(C) Western blots and densitometry of iBAT UCP1 in the same mice described under 3B.

(D) Western blots and densitometry analyses of iBAT UCP1 and TH proteins in 18-week-old HFD-fed mice at 22°C (Basal) or 4°C for 7 days (Cold).

(E) UCP1 mRNA levels in the iBAT of the mice described under 3D. n = 4-5/group.

(F and G) Core body temperatures (F) and oxygen consumption rates (G) in 23-week-old mice injected with a bolus of CL316,243 at thermoneutrality. n = 6/group.

(H) Plasma levels of FFAs in thermoneutrally housed 11-week-old mice at 15 min after a bolus of CL316,243 injection. n = 6-7/group.

(I) Oxygen consumption rates (OCRs) measured by Seahorse assays in isolated primary brown adipocytes from 28-day-old mice. Oli., oligomycin; Iso., isoproterenol; R/A, rotenone/antimycin.

Cold-acclimated UCP1-deficient mice were able to defend body temperature via shivering-induced thermogenesis (Golozoubova et al., 2001). To eliminate muscle shivering and assess nonshivering thermogenic capacity in BAT-KO mice, we injected a bolus of the β3 agonist CL316,243 into the mice housed at thermoneutrality. BAT-KO and control mice had similar increases in core body temperatures (Figure 3F) within 3 h of injection. Although the average increase in oxygen consumption per mouse was significantly lower (∼20% less) in BAT-KO mice than controls (Figure 3G) during this period, it reached the 2-fold level normally seen in CL316,243-injected mice (Grujic et al., 1997). When fold changes in oxygen consumption rates were plotted as a function of body weight or fat body mass that showed a trend towards decrease in BAT-KO mice, the difference was not statistically significant (Figure S3B). The modest reduction of CL316,243-stimulated oxygen consumption and FFA release (Figure 3H) in BAT-KO mice likely resulted from their reduced reserve of white fat (Figure 1H), a fat type whose β3 adrenergic receptor was reported to play a predominant role in determining total energy expenditure levels in mice following β3 agonist stimulation (Grujic et al., 1997). Importantly, isolated CGI-58-deficient primary brown adipocytes had no changes in basal and maximum oxygen consumption and responded to isoproterenol similarly as controls under normal Seahorse assay conditions (Figure 3I). When FFAs were completely depleted in the assay medium by adding a very high concentration (2%) of fatty acid-free albumin, the isoproterenol-stimulated and maximum oxygen consumption started to show a significant suppression in CGI-58-deficient brown adipocytes (Figure S3C). This observation is consistent with a previous study using ATGL and HSL inhibitors (Li et al., 2014) and suggests that CGI-58-mediated cytosolic LD lipolysis becomes critical in fueling brown adipocyte mitochondrial respiration only when exogenous FFAs are not available, a condition that does not normally occur in vivo. Overall, our results demonstrate that inhibiting the stimulated LD lipolysis in BAT does not reduce thermogenic capacity in vivo in mice.

BAT-KO Mice Increase Combustion of Circulating Fuels During Cold Exposure

Brown adipocytes may directly use FFAs, glucose and other chemicals derived from the blood as thermogenic substrates without firstly storing them in cytosolic LDs, though it was reported that utilization of TGs stored in cytosolic LDs of brown adipocytes plays a predominant role in acute cold-induced thermogenesis (Ma and Foster, 1986; Ouellet et al., 2012). In the fed state, diet is a major source of energetic substrates present in the blood. BAT-KO mice had increased calorie intake during chronic cold exposure (Figure 4A). The respiratory exchange ratio (RER) was significantly higher in these animals during the first 5 days of cold exposure (Figure 4B) indicating increased combustion of carbohydrates during this period. To provide direct evidence for this, we examined glucose-initiated thermogenesis during acute cold exposure. After a bolus of glucose, BAT-KO mice relative to controls showed a markedly attenuated response in blood glucose increases and a significant rise in intra-rectal temperatures (Figure 4C and 4D). These results were consistent with increased glucose uptake by iWAT and by iBAT as an organ during cold exposure (Figure 4E). It was also consistent with increased mRNA for glucose transporter 1 (Glut1) in these fat depots (Figure 4F and S4A). The mRNA level of glucose transporter 4 (Glut4), which is responsible for insulin-stimulated glucose uptake, decreased in iBAT and remained unchanged in iWAT in BAT-KO mice (Figure 4F and S4A). Insulin-stimulated glucose uptake per mg wet tissue was not significantly altered in adipose and muscle tissues (Figure S4B), except CGI-58-deficient iBAT whose total glucose uptake should be considered elevated because of its increased wet weight.

Open in a separate window

Figure 4

BAT-KO Mice Increase Combustion of Circulating Fuels During Cold Exposure

(A and B) Average daily calorie intake (A) and respiratory exchange ratio (RER) (B) of 18-week-old HFD-fed BAT-KO (KO) and control (Con) mice during 7 days of cold exposure. n = 5-6/group.

(C and D) Changes in blood glucose levels (C) and intrarectal temperatures (D) in 12-week-old HFD-fed mice after a bolus of glucose injection (i.p., at 1.5 g/kg body weight). The mice were fasted for 7h (4h at room temperature and 3h at cold) prior to glucose injection. n = 4-5/group.

(E) Tissue glucose uptake in 20-week-old HFD-fed mice during acute cold exposure. The total glucose uptake of iBAT was calculated based on the total iBAT weight. n = 4-5/group.

(F) Levels of mRNAs for glucose transporter 1 (Glut1) and glucose transporter 4 (Glut4) in iBAT in mice housed at 22°C (Basal) or 4°C for 7 days (Cold). n = 5/group.

(G) Levels of mRNAs for CD36, lipoprotein lipase (LPL), and fatty acid transporter protein 1 (FATP1) in iBAT in 14-week-old HFD-fed mice housed at 22°C. n = 5/group.

(H) Immunoblots and densitometry of CD36 protein in iBAT in mice exposed to cold for 7 days or injected daily with CL316,243 for 4 days.

Compared to glucose, FFAs are energy-rich molecules and normally the major fuel for thermogenesis. CD36, lipoprotein lipase (LPL) and fatty acid transporter 1 (FATP1) play important roles in iBAT uptake and transport of FFAs (Bartelt et al., 2011; Wu et al., 2006). We observed that iBAT levels of mRNAs for CD36 and LPL, but not FATP1, increased significantly in BAT-KO mice (Figure 4G). Expression of CD36 protein, a fatty acid transporter critically implicated in cold-induced thermogenesis (Bartelt et al., 2011), also elevated in the iBAT of BAT-KO mice after chronic cold exposure or CL316,243 treatment (Figure 4H). These findings imply that CGI-58-deficient BAT may have increased fatty acid uptake from the blood, especially during chronic cold adaptation. Collectively our results suggest that CGI-58-deficient BAT may compensate its defective LD lipolysis by increasing combustion of blood fuels including glucose and perhaps FFAs for thermogenesis.

WAT Lipolysis Is Essential for Thermogenesis During Fasting

When BAT-KO mice can use dietary nutrients to fuel thermogenesis in the fed state, they must rely on endogenous fuel during fasting. Fasting mobilizes energy stores mainly by inducing hepatic glycogenolysis and WAT lipolysis. It has been shown that mice lacking ATGL globally or in all adipose tissues are cold sensitive (Ahmadian et al., 2011; Haemmerle et al., 2006) suggesting a critical role of adipose lipolysis in cold-induced thermogenesis. During fasting, mice lacking CGI-58 in BAT only were not cold sensitive (Figure 2). Given this, we hypothesized that WAT lipolysis is essential for fueling thermogenesis during fasting. To test this hypothesis, we created a mouse line with CGI-58 inactivation in both BAT and WAT (FAT-KO mice) (Figure S5A). FAT-KO mice showed no alterations in weight gain and food intake despite increased fat body mass and weights of all adipose depots (Figure 5A-D). Like BAT-KO mice, FAT-KO mice also showed an enlarged white fat-looking iBAT (Figure 5E) and large cytosolic LD deposition in iBAT adipocytes (Figure 5F) resulting in increased brown adipocyte size (Figure 5G). Relative to controls, they also had increased total amounts of DNA and protein per iBAT (Figure S5B). As expected, FAT-KO mice were completely resistant to lipolysis stimulation (Figure 5H). Failure of an adrenergic agonist to stimulate lipolysis in FAT-KO mice demonstrates an essential role of CGI-58 in mediating the stimulated lipolysis. Like mice lacking ATGL globally (Haemmerle et al., 2006), FAT-KO mice housed at ambient temperature also developed hypothermia during fasting (Figure 5I). They were cold sensitive when food was removed during cold exposure (Figure 5J). Since mice lacking CGI-58 in BAT only were not cold sensitive during fasting, our results suggest that it is the WAT lipolysis that fuels thermogenesis in these animals in the fasted state.

Open in a separate window

Figure 5

WAT Lipolysis Is Essential for Thermogenesis During Fasting

(A) Weight gain of FAT-KO and control mice on chow (n = 9-12) or HFD (n = 10-13. Mice were 6 weeks old at Week 0.

(B) Calorie intake on the fifth week of chow (n = 5-6/group) or HFD (n = 5-6/group) treatment.

(C) Body composition of 12-week-old HFD-fed mice. n = 6/group.

(D) Adipose depot/body weight ratios in 24-week-old HFD-fed mice. n = 7/group.

(E) Gross appearance of iBATs from 24-week-old HFD-fed mice.

(F) H&E staining of iBATs from 14-week-old HFD-fed mice.

(G) Adipocyte size of iBATs from 24-week-old HFD-fed mice. For each sample, 5,000 cells were measured. n = 5/group.

(H) In vivo lipolysis capacity of 10-week-old HFD-fed mice.

(I) Intrarectal temperatures of 26-week-old HFD-fed mice during cold exposure in the absence of food. n = 5-6/group.

(J) Core body temperatures of 23-week-old HFD-fed mice during cold exposure in the absence or presence of food. n = 5-6/group.

(K) Western blots and densitometry of UCP1 protein in iBAT in 22-week-old HFD-fed mice.

(L) Levels of UCP1 mRNA in iBAT in 22-week-old HFD-fed mice. n = 4/group.

Despite cold intolerance during fasting, FAT-KO mice were able to tolerate cold when they were provided food during cold exposure (Figure 5J) suggesting that inhibiting adipose lipolysis causes cold intolerance by limiting thermogenic fuels, not by impairing thermogenic machinery. In FAT-KO mice fed a HFD, a chow diet, or a chow diet coupled with β3 agonist treatment, iBAT levels of UCP1 protein and mRNA did not decrease when normalized to GAPDH (Figure 5K, 5L, and S5C), and the levels actually elevated when considering the relative total amounts of DNA and protein per iBAT depot (Figure S5B). In HFD-fed mice treated with CL316,243, we did see a lower level of UCP1 protein in the CGI-58-deficient iBAT after normalization to GAPDH (Figure S5D). Again, when considering the relative total amounts of protein per iBAT, the total amount of UCP1 protein was largely unchanged in the CGI-58-deficient iBAT depot. However, unlike BAT-KO mice, FAT-KO mice had a drastic reduction in thermogenic capacity as evidenced by blunted increases in core body temperatures (Figure S5E) and oxygen consumption after treatment with CL316,243 at thermoneutrality (Figure S5F and S5G). Thus, WAT lipolysis plays a key role in determining whole-body thermogenic capacity.

CGI-58 Deletion in BAT Stimulates WAT Browning

Cold-exposed BAT-KO mice relative to controls exhibited higher body temperatures when they were provided food during cold exposure (Figure 2). It appeared that BAT-KO mice had a higher set point of whole-body thermogenesis in the brain, which could activate multiple compensatory thermogenic mechanisms through SNS. We did observe increased protein expression of TH, a marker for sympathetic innervation, in iBAT (Figure 3D) and iWAT (Figure 6A). WAT browning is an important cold adaptative mechanism often induced by sympathetic activation. Indeed, BAT-KO versus control mice housed at room temperature (a moderate cold condition compared to thermoneutrality), exposed to cold, or treated with the β3 agonist CL316,243 for 4 days all showed augmented WAT browning as evidenced by increased emergence of adipocytes containing multilocular LDs and enhanced UCP1 immunohistochemical staining in iWAT (Figure 6B and 6C), a fat pad sensitive to browning factors. Increased WAT browning of BAT-KO mice was further supported by elevated UCP1 mRNA and protein levels in iWAT after chronic cold exposure or β3 agonist treatment (Figure 6D and 6E). Under the basal condition (22°C and saline treatment), UCP1 mRNA levels in iWAT also increased significantly in BAT-KO mice when Student-t-test, not Two-way ANOVA, was used for statistical analysis (Figure 6D).

Open in a separate window

Figure 6

CGI-58 Deletion in BAT Stimulates WAT Browning

(A) Western blots and densitometry of tyrosine hydroxylase (TH) in iWAT in 18-week-old HFD-fed mice exposed to cold for 7 days.

(B) H&E staining of iWATs from 14-week-old chow-fed BAT-KO (KO) and control (Con) mice (top), 18-week-old HFD-fed mice (middle), and 18-week-old HFD-fed mice exposed to cold for 7 days (bottom).

(C) Immunostaining of UCP1 protein in iWAT in 18-week-old HFD-fed mice or 18-week-old HFD-fed mice exposed to cold for 7 days.

(D and E) Levels of UCP1 mRNA (D) and protein (E) in iWAT in 18-week-old HFD-fed mice at 22°C (Basal, or B) or 4°C for 7 days (Cold, or C) (n = 4-5/group), or 14-week-old HFD-fed mice injected with CL316,243 for 4 days (CL) or saline (S) (n = 5/group).

UCP1 transcription in beige adipocytes may activate UCP1-cre recombinase, resulting in excision of floxed CGI-58 alleles. CGI-58 protein expression levels in iWAT decreased modestly in BAT-KO mice after cold or β3 agonist stimulation (Figure 6A and 6E), but this does not answer whether CGI-58-floxed alleles were excised in beige adipocytes because of un-browned white adipocytes present in this depot. To circumvent this issue, we crossed BAT-KO mice with Rosa26-eYFP reporter mice (Srinivas et al., 2001). In these mice, UCP1 promoter activation resulted in excision of the loxP-floxed STOP sequence before eYFP, allowing eYFP expression and tracking of UCP1-positive beige adipocytes. Immunostaining of CGI-58 in iWAT sections revealed that CGI-58 expression disappeared from only a small proportion of eYFP-positive beige adipocytes in BAT-KO mice (Figure S6A). This observation suggests that UCP1-cre is less efficient in beige than brown adipocytes in excising floxed CGI-58 alleles, which is consistent with previous studies using the same UCP1-cre mouse line (Altshuler-Keylin et al., 2016; Kong et al., 2014). Given the normal UCP1 expression in CGI-58-deficient iBAT (Figure 3), our results indicate that UCP1 expression or beige adipocyte recruitment is not coupled with CGI-58 expression, which may imply that the stimulated adipose LD lipolysis is dispensable for UCP1 expression or WAT browning.

Mitochondrial respiration capacity is expected to be higher in thermogenic beige adipocytes than classical white adipocytes. To determine if this is the case in BAT-KO mice, we measured oxygen consumption of iWAT cell suspensions. As expected, the iWAT live cell suspension from BAT-KO mice relative to controls had significantly elevated basal and maximum oxygen consumption rates (Figure S6B), indicating that beige adipocytes in BAT-KO mice are functionally active.

Adipose browning can be induced by increased sympathetic innervation and/or direct actions of humoral and hormonal factors on adipocytes. TH protein levels in iWAT increased in BAT-KO mice housed at room temperature or exposed to the cold for 7 days (Figure 6A). The iWAT mRNA for β3 adrenergic receptor also elevated in these mice (Figure S6C). Thus, defective LD lipolysis in BAT may enhance WAT browning by increasing sympathetic innervation. To examine this possibility, we chemically and selectively denervated the sympathetic nerves by injecting the catecholaminergic neurotoxin 6-hydroxydopamine (6-OHDA) into one of two iWAT pads as described (Vaughan et al., 2014). Surgical denervation was not chosen because it cuts both sensory and sympathetic nerves. Ten days after 6-OHDA injection and cold exposure, TH protein levels in the denervated fat pad relative to the shamed operated pad of the same mouse decreased 56% and 60% in control and BAT-KO mice, respectively, which was associated with 62% and 69% decreases of UCP1 protein levels in the respective groups (Figure S6D). Consistently beige adipocytes in the denervated iWAT fat pad decreased substantially as revealed by H&E staining and UCP1 immunostaining (Figure S6E and S6F). These results suggest that WAT browning depends largely on sympathetic innervation, and that the increased sympathetic innervation into iWAT explains, in a large part, the increased WAT browning in BAT-KO mice.

Lipase Activity Assays

Tissue lipase activity was assayed as described previously (Wei et al., 2007). Snap-frozen iBATs were quickly weighed, homogenized in ice-cold buffer (50 mM Tris-HCI, pH 7.4, 250 mM sucrose, 1 mM EDTA), and sonicated briefly. Protein contents in homogenates were analyzed using a BCA kit. Lipase activity was initiated by mixing 20 μl of homogenates containing 1 μg of total proteins, 10 μl of 5 μM 4-mtehylumbelliferyl heptanoate (4-MUH) for assays using HSL inhibitor 0079-1A (NovoNordisk), or 2 μg of total proteins and 20 μl of 10 μM 4-MUH for assays including ATGL inhibitor Atglistatin (Cat.#: 1469924-27-3, Cayman), in Reaction Buffer (20 mM Tris-HCI, pH 8.0, 1 mM EDTA, 300 μM taurodeoxycholate), and 60 μl of Reaction Buffer in 96-well-plates and incubating at room temperature with shaking for 5 min. The fluorescence was then continuously recorded over a 10 min-period at 355 nm as excitation and 460 nm as emission wavelengths at 22°C in a plate reader.

In vivo Lipolysis Assays

The mice were injected intraperitoneally with isoproterenol at 10 mg/kg body weight around 10AM. Plasma samples were collected before and 15 min after isoproterenol injection. Plasma concentrations of FFA and glycerol were analyzed using Free Fatty Acids, Half Micro Test kit (Cat.#: 11383175001, Roche), or NEFA-HR (2) (Cat.#: 434-91795 and 434-91995, Wako Diagnostics) and Free Glycerol Reagent (Sigma).

Chronic CL-316,243 Treatment

CL-316,243 at 1 mg/kg body weight or saline were injected into mice daily around 10AM for 4 days. On day 5, the mice were sacrificed around 1PM without additional injection.

Ex vivo Lipolysis Assays

The iBATs of BAT-KO and control mice were surgically removed and washed with PBS. Tissue pieces (∼12 mg) were preincubated in DMEM containing 40 μM Atglistatin or vehicle DMSO for 8 h at 37°C, or 10 μM HSL inhibitor 0079-1A (NovoNordisk) or vehicle DMSO for 1 h at 37°C. Thereafter, the medium was replaced by DMEM containing 2% fatty acid-free BSA in the presence or absence of 20 μM forskolin and 40 μM Atglistatin or 10 μM HSL inhibitor, and incubated for an additional 60 min at 37°C. The medium was then collected for FFA assays using NEFA-HR (2) (Wako Diagnostics).

Immunohistochemistry

The iWAT was fixed in 10% formalin solution, paraffin-embedded, and then sectioned at 5 μm thickness. The sections were deparaffinized and rehydrated, followed by antigen retrieval using 10 mM sodium citrate (pH6.0) for 20min. After antigen retrieval, non-specific sites were blocked by incubating the sections in TBS containing 10% normal goat serum and 0.2% Tween 20 for 40 min. The sections were then incubated with a primary antibody against UCP1 (1:200) for overnight at 4°C, followed by three washes, 5 min each, with TBS-Tween 20 solution and incubation for 1h with a biotinylated secondary antibody (1:200) included in VECTASSTAIN Elite ABC HRP Kit (Peroxidase, Rabbit IgG (Cat.#: PK-6101, Vector Laboratories). After three washes, the sections were incubated with the reagents in VECTASSTAIN Elite ABC HRP Kit for 30 min and then with 3,3′-diaminobenzidine tetrahydrochloride (DAB) (Thermo Scientific Pierce) for 5-10 min. After three washes, cell nuclei were stained with hematoxylin (Ricca Chemical Company). The sections were dehydrated, sealed, and examined under a microscope.

Glucose Disposal and Insulin Sensitivity Tests

For glucose tolerance test (GTT), the mice were fasted overnight (16 h) and then injected i.p. with glucose solution at 1.5 g/kg body weight. Blood glucose concentrations were measured using Bayer Contour Glucometer (Bayer Healthcare, LLC.) at 0, 15, 30, 60, and 120 min after glucose injection. For glucose-stimulated insulin secretion, plasma samples were collected during GTT and plasma insulin concentrations were measured (Millipore). For insulin tolerance test (ITT), the mice were fasted for 6 h during the daytime cycle, followed by an intraperitoneal injection of recombinant human insulin in saline at a dose of 0.75 U/kg body weight. Blood glucose levels were monitored at 0, 15, 30, 60 and 120 min after insulin injection.

Analysis of Adipocyte Size

The adipose tissue cell sizing was done in osmium-fixed urea-isolated adipocytes as described previously (Etherton et al., 1977) with some modifications. Briefly, the adipose tissues from different fat depots of the mice on HFD for 17 weeks were excised aseptically, sliced to ∼50 mg in size and rinsed twice in 0.154 M NaCl at 37°C to wash off free lipids on the tissues. The tissue slices were transferred to scintillation vials containing 1.2 mL of 50mM Colloidine HCl buffer and 2 mL of 3% osmium tetroxide in 0.05 M Colloidine HCl buffer (pH 7.4 at 37°C). The sample vials were kept in a ventilated fume hood at room temperature for 72 h. The osmium-colloidine buffer was removed and replaced by 10 mL of 0.154 M NaCl for 24 h. Subsequently, 10 mL 8M Urea in 0.154 M NaCl was added to the sample vials and the vials were intermittently swirled by hand and kept in room temperature for 48 h. This solubilized the connective tissue in the adipose tissue slices resulting in a suspension of adipocyte cells. The adipocyte cell suspension was passed through 250 μM nylon membrane to remove any of the remaining tissue debris. The cells were rinsed with 0.01% Triton-X in distilled water and then used to profile cell size distribution using Multisizer^® Coulter Counter (Beckman coulter Ireland Inc.).

Tissue Lipid Analysis

Total lipids were extracted as described previously (Temel et al., 2007). To determine tissue lipid concentrations, a total of ∼40-50 mg of tissue was sliced and extracted with 1:1 CHCl₃/MeOH. The organic phase was analyzed for lipid species. Enzymatic assays were used to determine total cholesterol (TC), free cholesterol (FC), and triglyceride (TG) concentrations of the samples. TC and FC were measured with enzymatic assay kit from Wako (Cat.#: 439-17501 for TC and Cat.#: 435-35801 for FC, Wako). TG was measured enzymatically using the Free Glycerol Reagent (Cat.#: F6428, Sigma-Aldrich) and Triglyceride Reagent (Cat.#: T2449, Sigma-Aldrich). Protein concentrations in the delipidated samples were determined using a Lowry assay (Lowry et al., 1951).

Glucose Uptake Assays

To measure insulin-stimulated glucose uptake, the mice were fasted overnight (16h), and then injected with human insulin at 0.75 mU/Kg and 10 μCi of C¹⁴-2-Deoxyglucose (2-DG). Forty-five minutes after 2-DG injection, the mice were sacrificed and tissues collected. For cold-induced glucose uptake, overnight fasted mice were exposed to cold (4°C) for 30 min before injection of 10 μCi of [³H]-2-DG as a tracer and then sacrificed for tissue collection after additional 45 min cold exposure. DG is transported into tissues and phosphorylated to 2-deoxyglucose 6-phosphate (2-DGP). Tissue contents of 2-DGP were used to reflect their glucose uptake efficiency. To determine tissue 2-DGP content, samples were homogenized and 2-DGP was separated from 2-DG using ion exchange column (Chromatography Columns #731-6211, Bio Rad), and then counted as described (Kim et al., 2000).

Protein Extraction and Immunoblotting

The tissues were homogenized and lysed for 30 min in RIPA buffer containing 0.1% SDS with a protease and phosphatase inhibitor cocktail (Sigma), followed by sonication and centrifugation. To avoid fat contamination, a 26G_1/2 syringe was used to collect the supernatant. The protein concentration of the supernatant was determined using a bicinchoninic acid (BCA) kit (BCA, Pierce, Rockford, IL). For immunoblotting, proteins in the supernatant were denatured by heating, separated by SDS-PAGE, and then transferred onto a nitrocellulose membrane (Bio-Rad). The membrane was incubated in a 5% non-fat milk blocking buffer (TBS-Tween 20) for 1 h, followed by incubation with a primary antibody in TBS-Tween 20 containing 5%BSA for overnight at 4°C. After 3 washes with TBS-Tween 20, the membrane was incubated with a secondary antibody, washed, and developed with Enhanced Chemiluminescence Detection Reagents (ECL, Thermo Fisher). The protein signals were imaged using a Bio-Rad ChemiDoc System.

Quantitative Real-time Polymerase Chain Reaction

Total RNAs were isolated using TRIzol^® Reagents. RNAs were reverse transcribed into cDNA using TaqMan^® Reverse Transcription Reagents (Applied Biosystems) and tissue mRNA levels were determined by qPCR using Stratagene Mx3005p PCR machine (Agilent Technologies). Reactions were done in duplicate for each biological sample using SYBR^® Green Real-Time PCR Master Mix (Invitrogen). The relative mRNA expression level for each gene was calculated by the 2(^-DDCt) method and normalized to GAPDH, which was arbitrarily set to 1.

Arginine Tolerance Test (ATT)

Mice were fasted for 6 h and then injected i.p. with arginine in saline at 1 g/kg body weight. Plasma insulin concentrations were measured using Ultra-Sensitive Mouse Insulin ELISA Kit (Crystal Chem) at 0, 5, 15, 30 min after arginine injection.

Chemical Denervation of the Sympathetic Nerves in iWAT Pads

Chemical denervation of sympathetic nerves in the iWAT was performed using the catecholaminergic neurotoxin 6-hydroxy-dopamine hydrochloride (6-OHDA) as described (Vaughan et al., 2014). Briefly, 6-OHDA was prepared at 9 mg/ml in 0.15 M NaCI containing 1% ascorbic acid. A total of 24 μl (2 μl per injection at 12 loci) of 6-OHDA solution was injected into the right side iWAT pad per mouse and a total of 24 μl of vehicle was injected into the left side iWAT pad as a shamed operated control. For the determination of the effect of sympathetic innervation on iWAT browning, prior to denervation, the mice were housed at thermoneutrality (30°C) for 3 weeks to minimize the basal level differences in WAT browning between BAT-KO and control mice. After denervation, the mice were recovered at thermoneutrality for 7 days and then at room temperature for 3 days, followed by 6 days of the intermittent cold exposure (12 h at 4°C in the light cycle and 12 h at room temperature during the dark cycle). Twenty-four hours before sacrifice, the mice were housed at 4°C continuously to maximally induce browning. All of the mice survived this cold exposure regimen after iWAT sympathetic denervation.

Seahorse Analysis of Primary Brown Adipocytes

Primary brown adipocytes were prepared from the stromal vasculature fraction (SVF) isolated from iBAT. In brief, iBAT was isolated and digested with PBS containing 10 mM CaCl₂, Collagenase D (1.5 U/ml) and Dispase II (2.4 U/ml) for ∼45 min. After digestion, the SVF was collected by centrifugation at 700×g for 10 min at 37°C, and then filtered through a 40 μm nylon mesh cell strainer. The cells in the SVF were cultured in the Maintaining Medium (DMEM/F12 supplemented with 10% FBS). For Seahorse studies, 20,000 cells were plated onto each well of XF96 Cell Culture Microplates. When the cells reached at ∼95% confluency, the induction medium [1 nM 3,3′5-Triiodo-L-thyronine (T3), 5 μg/ml insulin, 125 μM indomethacin, 2 μg/ml dexamethasone, 0.5 mM 3-isobutyl-1-methyxanthine (IBMX), 0.5 μM rosiglitazone] was added. After 48 h induction, the cells were maintained in the Maintaining Medium supplemented with insulin (5 μg/ml), T3 (1 nM), and rosiglitazone (0.5 μM) for additional 48 h induction. The cells were fully differentiated and used for Seahorse analysis after additional 3-4 days of incubation in the fresh Maintaining Medium containing Insulin (5 μg/ml), T3 (1 nM), and rosiglitazone (1 μM).

XF^e96 Extracellular Flux Analyzer was used for Seahorse studies. Briefly, the basal respiration was recorded in the untreated cells, followed by inhibition of the coupled respiration by oligomycin (1 μM). The sympathetic activation-associated uncoupled respiration was then measured after addition of isoproterenol (1 μM). The maximum respiration capacity of the cells was assessed after addition of 2 μM carbonyl cyanide-4-(thfluomrthoxy)phenylhydrazone (FCCP). Lastly, non-mitochondrial respiration was measured after addition of rotenone/antimycin A (0.5 μM).

Oxygen Consumption Rates of Live Cell Suspensions of iWATs

The iWATs from 10-week-old BAT-KO and control mice were collected and digested in PBS containing 10 mM CaCI₂, Collagenase D (1.5 U/ml) and Dispase II (2.4 U/ml) at 37°C for 40 min. The cell suspensions were then used directly for Seahorse analysis. The basal and maximal oxygen consumption rates were normalized by the DNA amount in each cell suspension preparation.

Calculation of Quantitative Insulin Sensitive Check Index (QUICKI)

Plasma insulin and glucose concentrations were determined in overnight fasted BAT-KO and control mice. QUICKI was calculated as 1 / [log(fasting insulin at μU/ml) + log(fasting glucose at mg/dL)].

We thank Dr. Rudolf Zechner at University of Graz for sharing HSL inhibitor 0079-1A. This work was supported in part by Award Numbers R01DK085176 (L.Y.), R01DK111052-01 (L.Y.), R01DK107544 (B.X.), and R01HL107500 (B.X.) from the National Institute of Diabetes and Digestive and Kidney Diseases, 17GRNT33670590 (L.Y.) and 15GRNT25710256 (H.S.) from AHA, and ADA 7-13-BS-159 (H.S.).