Sarco(endo)plasmic reticulum Ca 2+ -ATPase (SERCA) uncoupling in skeletal muscle, and mitochondrial uncoupling via uncoupling protein 1 (UCP1) in brown/beige adipose tissue are two primary mechanisms implicated in energy expenditure. Here, we investigated the effects of glycogen synthase kinase 3 (GSK3) inhibition via lithium chloride (LiCl) treatment on SERCA uncoupling in skeletal muscle and UCP1 expression in adipose. C2C12 and 3T3-L1 cells treated with LiCl had increased SERCA uncoupling and UCP1 protein levels, respectively, ultimately raising cellular respiration; however, this was only observed when LiCl treatment occurred throughout differentiation. In vivo, LiCl treatment (10 mg/kg/day) increased food intake in chow-fed and high-fat diet (HFD, 60% kcal) fed male mice without increasing body mass – a result attributed to elevated daily energy expenditure. In soleus muscle, we determined LiCl treatment promoted SERCA uncoupling via increased expression of SERCA uncouplers, sarcolipin and/or neuronatin, under chow and HFD-fed conditions. We attribute these effects to the GSK3 inhibition observed with LiCl treatment as partial muscle-specific GSK3 knockdown produced similar effects. In adipose, LiCl treatment inhibited GSK3 in inguinal WAT (iWAT) but not in brown adipose tissue under chow-fed conditions, which in turn led to an increase in UCP1 in iWAT and a beiging-like effect with a multilocular phenotype. We did not observe this beiging-like effect and increase in UCP1 when mice were fed a HFD, as LiCl could not overcome the ensuing overactivation of GSK3. Nonetheless, our study establishes novel regulatory links between GSK3 and SERCA uncoupling in muscle and GSK3 and UCP1 and beiging in iWAT.

Lithium is a well-known GSK3 inhibitor that has been commonly used in the treatment of bipolar disorder (). Although higher doses (i.e.,1.0mM serum concentration) taken over a prolonged period of time have been associated with weight gain and obesity (), low dose lithium supplementation produces the opposite effect (). We have recently shown that trace levels of lithium found naturally in water negatively correlates with the prevalence of obesity (). In mice, others have shown that low dose lithium supplementation (provided as lithium chloride [LiCl], 10 mg/kg/day) for 14 weeks attenuated high-fat diet induced obesity and atherosclerosis (). In skeletal muscle specifically, we have shown that male mice fed this same dose for 6 weeks have reduced GSK3 activation, leading to increased calcineurin activation, peroxisome proliferator-activated receptor-gamma coactivator 1-alpha (PGC-1α) content and fatigue resistance (). Calcineurin is a Ca-dependent phosphatase that promotes PGC-1α expression and fatigue resistance (). Furthermore, recent studies have shown that calcineurin stimulates both muscle- and adipose-based thermogenesis by increasing SLN and UCP1 in muscle and adipose tissue, respectively (). In muscle, calcineurin is well-known to be counteracted by GSK3 (). Thus, in the present study, we used a cell and rodent model approach to determine whether GSK3 inhibition via LiCl supplementation would enhance muscle-based and adipose-based thermogenic mechanisms: SERCA uncoupling and UCP1 expression.

Both adipose-based (via UCP1) and muscle-based (via SERCA uncoupling) thermogenesis have shown promise in the fight against obesity – a metabolic disorder caused by a chronic imbalance between excessive caloric intake and a deficit in energy expenditure (). Thus, discovering novel cellular targets that could promote both adipose-based and muscle-based thermogenesis would likely aid in offsetting the onset or progression of obesity. First recognized for its role in the regulation of glycogen synthase (), glycogen synthase kinase 3 (GSK3) has been recently identified as a significant contributor to numerous disease states including obesity and diabetes (). GSK3 is a constitutively active serine/threonine kinase which exists in two isoforms: GSK3α and GSK3β, though the latter is most dominant in adipose and muscle tissues (). It has been suggested that the overactivation of GSK3 may contribute to the onset of obesity. For example, GSK3β overexpression in mice resulted in increased body mass and adiposity along with impaired glucose tolerance and insulin sensitivity (). Mechanistically, the association between GSK3 activity and obesity may in part be mediated through adaptive thermogenesis. Recently, it has been demonstrated that GSK3 negatively regulates brown adipose based-thermogenesis, where the expression of thermogenic genes including, UCP1 are suppressed by GSK3 activity (). However, analysis was restricted to the effects of GSK3 on brown adipocytes, and the role of GSK3 on regulating browning of WAT remains unknown. Furthermore, the role of GSK3 in regulating muscle-based thermogenesis and SERCA uncoupling remains unknown. This is important as it was muscle-specific overexpression of GSK3β in mice that resulted in increased adiposity, even under chow-fed conditions ().

Adaptive thermogenesis is the cellular process where, in response to prolonged cold exposure or caloric excess, energy expenditure and heat production are increased resulting in a greater combustion of metabolic substrates (). In mammals, brown (BAT)/beige adipose tissue and skeletal muscle are the primary sites for adaptive thermogenesis. Beige adipose and BAT are characterized by an abundance of mitochondria and high uncoupling protein 1 (UCP1) content, which acts to uncouple the proton gradient from ATP synthesis, ultimately dissipating the stored energy in the form of heat. In skeletal muscle, the sarco(endo)plasmic reticulum Ca-ATPase (SERCA) pump is a major energy (ATP) consumer and is a known mediator of muscle-based thermogenesis (). Specifically, the SERCA pump catalyzes the active transport of Cafrom the cytosol to the sarcoplasmic reticulum (SR), a process that is important for muscle relaxation. Based on its structure, SERCA has 2 Caand 1 ATP binding sites (), which suggests that under optimal conditions SERCA can transport 2 Caions for every 1 ATP hydrolyzed (). Although in vivo, the presence of SERCA uncouplers such as sarcolipin (SLN) and the newly identified neuronatin (NNAT), as well as changes in membrane lipid composition or ryanodine receptor (RyR) Caleak, makes the SERCA pump much less efficient thereby lowering the apparent coupling ratio and increasing energy expenditure and heat release ().

Lithium is known to have insulin sensitizing effects and Choi et al., (2010) found that LiCl treatment, with the same dose used in this study, lowered fasting glucose in high fat fed mice (). Here, we did not find any effect of LiCl supplementation on glucose or insulin tolerance in either chow-fed ( Figure 10 ) or high-fat fed mice ( Figure 11 ).

Glucose tolerance tests (GTT) and insulin tolerance tests (ITT) in control and LiCl treated control (CON), high fat diet (HFD), and HFD-lithium (HFD-Li) fed mice. A and B) GTT curves and corresponding area under the curve (AUC) analysis. C and D) ITT curves and corresponding AUC analysis. For A and C: *p < 005, **p < 0.01, ***p < 0.001 and ****p < 0.0001 for HFD vs chow; and # p < 005, ## p < 0.01, ### p < 0.001 and #### p < 0.0001 for HFD-Li vs chow with a two-way ANOVA and a Tukey’s post hoc test. For B and D: ***p < 0.001 and ****p < 0.0001 with a one-way ANOVA and a Tukey’s post-hoc test (n = 22-24 per group).

Figure 11 Glucose tolerance tests (GTT) and insulin tolerance tests (ITT) in control and LiCl treated control (CON), high fat diet (HFD), and HFD-lithium (HFD-Li) fed mice. A and B) GTT curves and corresponding area under the curve (AUC) analysis. C and D) ITT curves and corresponding AUC analysis. For A and C: *p < 005, **p < 0.01, ***p < 0.001 and ****p < 0.0001 for HFD vs chow; and # p < 005, ## p < 0.01, ### p < 0.001 and #### p < 0.0001 for HFD-Li vs chow with a two-way ANOVA and a Tukey’s post hoc test. For B and D: ***p < 0.001 and ****p < 0.0001 with a one-way ANOVA and a Tukey’s post-hoc test (n = 22-24 per group).

Glucose tolerance tests (GTT) and insulin tolerance tests (ITT) in control and LiCl treated chow fed mice. A and B) GTT curves and corresponding area under the curve (AUC) analysis. C and D) ITT curves and corresponding AUC analysis (n = 12 per group).

Figure 10 Glucose tolerance tests (GTT) and insulin tolerance tests (ITT) in control and LiCl treated chow fed mice. A and B) GTT curves and corresponding area under the curve (AUC) analysis. C and D) ITT curves and corresponding AUC analysis (n = 12 per group).

In contrast with chow-fed mice, LiCl supplementation did not lead to any alterations in iWAT or BAT in HFD fed mice ( Figure 9 ). There were no differences in inhibitory serine9 phosphorylation of GSK3β between HFD and HFD-Li groups when expressed as a ratio of total GSK3β ( Figure 9 A). When assessed separately, our results show that the HFD led to a significant ∼44% reduction in serine9 phosphorylated GSK3β in both the HFD (p = 0.01) and HFD-Li (p = 0.01) groups compared with control (F = 5.8, p = 0.009). This occurred without any changes in total GSK3β (F = 0.14, p = 0.86). This suggests that LiCl supplementation could not overcome the overactivation of GSK3 in iWAT observed with high fat feeding. When measuring protein content in iWAT, there were significant reductions found in several mitochondrial-related proteins with no change in UCP1 in both HFD and HFD-Li groups compared with chow-fed mice ( Figure 9 B and C). In BAT, there were no differences observed for UCP1 or most mitochondrial proteins, except for citrate synthase, which was upregulated under both HFD conditions ( Figure 9 B and D). Finally, histological analyses did not reveal a beige-like phenotype in iWAT from the HFD-Li group ( Figure 9 E).

LiCl supplementation did not inhibit GSK3 in iWAT or BAT from mice fed a high fat diet (HFD). A) Inhibitory serine9 phosphorylation of GSK3 in iWAT and BAT from control (CON; standard chow) HFD and HFD-Li mice (n = 8 per group). B-D) Representative Western blot images and analyses of mitochondrial proteins PGC-1α, citrate synthase (CS), cytochrome c oxidase subunit IV (COXIV), pyruvate dehydrogenase E1-alpha subunit (PDH) and UCP1 in iWAT (C) and BAT (D). E) H&E stain of iWAT and BAT sections from chow, HFD and HFD-Li mice (scale bar = 200 μm).For B and C: *p < 005, **p < 0.01 with a one-way ANOVA and Tukey’s post-hoc test. Western blot data are presented as relative to chow. All values are means ± SEM.

Figure 9 LiCl supplementation did not inhibit GSK3 in iWAT or BAT from mice fed a high fat diet (HFD). A) Inhibitory serine9 phosphorylation of GSK3 in iWAT and BAT from control (CON; standard chow) HFD and HFD-Li mice (n = 8 per group). B-D) Representative Western blot images and analyses of mitochondrial proteins PGC-1α, citrate synthase (CS), cytochrome c oxidase subunit IV (COXIV), pyruvate dehydrogenase E1-alpha subunit (PDH) and UCP1 in iWAT (C) and BAT (D). E) H&E stain of iWAT and BAT sections from chow, HFD and HFD-Li mice (scale bar = 200 μm).For B and C: *p < 005, **p < 0.01 with a one-way ANOVA and Tukey’s post-hoc test. Western blot data are presented as relative to chow. All values are means ± SEM.

We next examined the effects of LiCl treatment on both iWAT and BAT in chow fed and HFD fed mice. In chow fed mice and as early as 6 weeks of treatment, LiCl significant increased GSK3 serine9 phosphorylation in iWAT ( Figure 8 A); however, this was not observed in BAT ( Figure 8 A). In turn, LiCl treatment resulted in significant elevations of UCP1 and several other mitochondria-related proteins including PGC-1α in iWAT but not BAT ( Figure 8 B and C). Histological analysis demonstrated for the first time that GSK3 inhibition associated with LiCl supplementation leads to a significant ‘beiging’ effect on iWAT, taking on a multilocular-like phenotype ( Figure 8 D and E). As SERCA-mediated Casignalling has been shown to contribute to energy homeostasis in beige adipocytes (), we next examined whether LiCl supplementation would alter protein levels of SERCA2 and RYR in iWAT. However, we did not find any differences in either SERCA2 or RYR content ( Figure 8 F).

LiCl supplementation inhibits GSK3 and promotes a beige-like phenotype in iWAT from mice fed a chow diet. A) Inhibitory serine9 GSK3 phosphorylation is elevated in iWAT but not BAT after LiCl treatment (n = 11-12 per group). B and C) Western blot analyses of mitochondrial proteins UCP1, cytochrome c (Cyto C), cytochrome c oxidase subunit IV (COXIV), citrate synthase (CS), pyruvate dehydrogenase E1-alpha subunit (PDH), and PGC-1α in iWAT (B) and BAT (C) from control and LiCl-fed mice (n = 6-12 per group, apart from UCP1 where 18 per group were used). D) H&E stain of iWAT and BAT sections from control and LiCl-fed mice (scale bar = 100 μm). E) Percent of multilocular adipocytes quantified using imageJ. F) Western blot quantification of SERCA2 and RYR in iWAT (n = 4 per group). *p < 0.05 using a Student’s t-test. All western blot data are presented as relative to control. All values are means ± SEM.

Figure 8 LiCl supplementation inhibits GSK3 and promotes a beige-like phenotype in iWAT from mice fed a chow diet. A) Inhibitory serine9 GSK3 phosphorylation is elevated in iWAT but not BAT after LiCl treatment (n = 11-12 per group). B and C) Western blot analyses of mitochondrial proteins UCP1, cytochrome c (Cyto C), cytochrome c oxidase subunit IV (COXIV), citrate synthase (CS), pyruvate dehydrogenase E1-alpha subunit (PDH), and PGC-1α in iWAT (B) and BAT (C) from control and LiCl-fed mice (n = 6-12 per group, apart from UCP1 where 18 per group were used). D) H&E stain of iWAT and BAT sections from control and LiCl-fed mice (scale bar = 100 μm). E) Percent of multilocular adipocytes quantified using imageJ. F) Western blot quantification of SERCA2 and RYR in iWAT (n = 4 per group). *p < 0.05 using a Student’s t-test. All western blot data are presented as relative to control. All values are means ± SEM.

Acknowledging the fact that Li also has GSK3-independent effects, we generated a partial muscle-specific GSK3 knockdown (GSK3) mouse colony. These mice heterozygously express the flox sequence flanking both GSK3 α and β, while also heterozygously expressing HSA-Cre to direct skeletal muscle-specific knockdown of GSK3. In turn, the GSK3mice display a ∼55% reduction in GSK3β in the soleus compared with GSK3control mice ( Figure 7 A). Similar to LiCl treatment and under a chow diet, we did not find any differences in rates of Cauptake or SERCA activity when examined separately, however, a significant reduction in SERCA’s apparent coupling ratio, particularly at 500 nM of [Cawas detected ( Figure 7 B- D). Moreover, while we did not observe any differences in RYR content, both SLN and NNAT were elevated in muscles from GSK3mice compared with GSK3mice ( Figure 7 E).

Muscle specific partial GSK3 knockdown (GSK3 mKD ) promotes SERCA uncoupling in soleus muscles from chow fed male C57BL/6J mice. A) GSK3β protein content assessed via western blotting in soleus muscles from GSK3 floxed (GSK3 floxed , control) and GSK3 mKD mice (n = 4 per group). B) Rates of Ca 2+ uptake in soleus muscles from GSK3 floxed and GSK3 mKD mice (n = 4 per group). C) SERCA activity in soleus muscles from GSK3 floxed and GSK3 mKD mice (n = 4 per group). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ) (n = 4 per group). E) Western blot analysis of SERCA1, SERCA2, RYR, SLN and NNAT (n = 3-4 per group). *p < 0.05, **p < 0.01, using an independent Student’s t-test. All values are means ± SEM.

Figure 7 Muscle specific partial GSK3 knockdown (GSK3 mKD ) promotes SERCA uncoupling in soleus muscles from chow fed male C57BL/6J mice. A) GSK3β protein content assessed via western blotting in soleus muscles from GSK3 floxed (GSK3 floxed , control) and GSK3 mKD mice (n = 4 per group). B) Rates of Ca 2+ uptake in soleus muscles from GSK3 floxed and GSK3 mKD mice (n = 4 per group). C) SERCA activity in soleus muscles from GSK3 floxed and GSK3 mKD mice (n = 4 per group). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ) (n = 4 per group). E) Western blot analysis of SERCA1, SERCA2, RYR, SLN and NNAT (n = 3-4 per group). *p < 0.05, **p < 0.01, using an independent Student’s t-test. All values are means ± SEM.

Western blot analysis revealed that the consumption of a HFD significantly reduced GSK3β serine9 phosphorylation, although this effect was blunted with LiCl treatment ( Figure 6 A). Further, PGC-1α content appeared to be elevated in HFD-Li supplemented mice in comparison to chow-fed controls, however this was not significantly different ( Figure 6 B). With respect to SERCA, we identified a significant reduction in Cauptake in the HFD-Li group when compared to untreated counterparts, whereas there were no differences between groups in SERCA activity ( Figure 6 B and C). We also found a significant reduction in SERCA’s apparent coupling ratio in the soleus of HFD-Li group compared with HFD ( Figure 6 D). Notably, the rates of Cauptake, activity, and the calculated apparent coupling ratio were obtained at a 1500 nM [Casince Cauptake in soleus muscles from HFD fed mice (with or without LiCl) did not reach levels below 1000 nM [Ca. To identify the cellular mechanisms leading to the reduction in apparent coupling ratio observed with a HFD, we examined the protein levels of SERCA uncoupling proteins SLN and NNAT as well as RYR. Our results show a significant increase in SLN and NNAT content in HFD-Li vs HFD soleus; however, there were no differences in RYR ( Figure 6 E).

LiCl supplementation inhibits GSK3 and promotes SERCA uncoupling in soleus muscles from male C57BL/6J high fat diet (HFD)-fed mice. A) Inhibitory serine9 phosphorylation of GSK3 and PGC-1α protein in soleus muscles from control (CON; standard chow) HFD and HFD-Li mice (n = 8-11 per group). B) Rates of Ca 2+ uptake in soleus muscles obtained from HFD and HFD-Li mice (n = 3-4 per group with each n representing pooled soleus from 3 mice). C) SERCA activity in soleus muscles obtained from HFD and HFD-Li mice (n = 3-4 per group with each n representing pooled soleus from 3 mice). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ). E) Western blot analyses of SERCA uncouplers SLN and NNAT as well as RYR (n = 6-8 per group). For A, **p < 0.01 with a one-way ANOVA and Tukey’s post-hoc test. For B-E. *p < 0.05, ***p < 0.001 with a Student’s t-test. All values are means ± SEM.

Figure 6 LiCl supplementation inhibits GSK3 and promotes SERCA uncoupling in soleus muscles from male C57BL/6J high fat diet (HFD)-fed mice. A) Inhibitory serine9 phosphorylation of GSK3 and PGC-1α protein in soleus muscles from control (CON; standard chow) HFD and HFD-Li mice (n = 8-11 per group). B) Rates of Ca 2+ uptake in soleus muscles obtained from HFD and HFD-Li mice (n = 3-4 per group with each n representing pooled soleus from 3 mice). C) SERCA activity in soleus muscles obtained from HFD and HFD-Li mice (n = 3-4 per group with each n representing pooled soleus from 3 mice). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ). E) Western blot analyses of SERCA uncouplers SLN and NNAT as well as RYR (n = 6-8 per group). For A, **p < 0.01 with a one-way ANOVA and Tukey’s post-hoc test. For B-E. *p < 0.05, ***p < 0.001 with a Student’s t-test. All values are means ± SEM.

SERCA coupling ratio was tested in the soleus muscle from LiCl-supplemented and control mice. This muscle was chosen based on its known expression of SERCA uncouplers, SLN and NNAT (). Under a chow diet, western blot analysis showed a significant increase in inhibitory serine9 phosphorylation on GSK3β in soleus muscles from LiCl treated mice vs control ( Figure 5 A). When measuring rates of Cauptake and SERCA activity separately, we did not find any significant effect of LiCl treatment, however, LiCl treatment resulted in a significant reduction in the apparent coupling ratio, particularly at 1000 nM [Ca Figure 5 B-D). The promotion of SERCA uncoupling was associated with a significant increase in NNAT but not SLN ( Figure 5 E). We also did not observe any changes in SERCA or RYR content ( Figure 5 E).

LiCl supplementation inhibits GSK3 and promotes SERCA uncoupling in soleus muscles from male C57BL/6J chow-fed mice. A) Inhibitory serine9 phosphorylation of GSK3β in soleus muscles from control and LiCl-treated mice (n = 5 per group). B) Rates of Ca 2+ uptake in soleus muscles from control and LiCl-treated mice (n = 5 per group). C) SERCA activity in soleus muscles from control and LiCl-treated mice (n = 5 per group). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ) (n = 5 per group). E) Western blot analysis of SERCA1, SERCA2, RYR, SLN and NNAT. For SERCA2 and RYR, these targets were probed for on the same membrane and have therefore have the same ponceau. *p < 0.05, **p < 0.01, using an independent Student’s t-test. All values are means ± SEM.

Figure 5 LiCl supplementation inhibits GSK3 and promotes SERCA uncoupling in soleus muscles from male C57BL/6J chow-fed mice. A) Inhibitory serine9 phosphorylation of GSK3β in soleus muscles from control and LiCl-treated mice (n = 5 per group). B) Rates of Ca 2+ uptake in soleus muscles from control and LiCl-treated mice (n = 5 per group). C) SERCA activity in soleus muscles from control and LiCl-treated mice (n = 5 per group). D) Apparent coupling ratio (Ca 2+ uptake divided by SERCA activity at matching [Ca 2+ ] free ) (n = 5 per group). E) Western blot analysis of SERCA1, SERCA2, RYR, SLN and NNAT. For SERCA2 and RYR, these targets were probed for on the same membrane and have therefore have the same ponceau. *p < 0.05, **p < 0.01, using an independent Student’s t-test. All values are means ± SEM.

To determine if LiCl could enhance muscle-based and adipose-based thermogenesis in vivo, we treated chow-fed and HFD fed mice with a dose we had previously shown to cause GSK3 inhibition in muscle (). Under a chow diet, 12 weeks of LiCl supplementation (10 mg/kg/day via drinking water) did not alter body mass, however, LiCl treatment appeared to increase food consumption with a significant difference in cumulative food intake by week 12 compared with control ( Figure 3 A and B). Daily food intake in the LiCl group also tended to be higher compared with control but this was not statistically significant ( Figure 3 C). Furthermore, body composition analysis showed no significant differences between LiCl and control ( Figure 3 D and E). This apparent increase in food consumption without an increase in body mass or percent of body fat points towards an increase in energy expenditure with LiCl supplementation, which was observed across light, dark and daily periods ( Figure 3 F and G). In both absolute and relative to lean mass, Oconsumption was significantly elevated in the LiCl group at both 6 weeks and 11 weeks of treatment. Importantly, these changes in energy expenditure could not be explained by any differences in cage ambulation ( Figure 3 H). Similar findings were also observed in mice fed a HFD (60% kcal from fat), a condition that would benefit from adaptive thermogenesis. Like the chow-fed mice, no differences in body mass between HFD and HFD-Li groups were observed after 12 weeks of feeding ( Figure 4 A) despite the HFD-Li group having the highest levels of cumulative food consumption and daily food intake ( Figure 4 B and C). We presume this to be due to elevated daily energy expenditure, similar to what was found in the chow study ( Figure 3 ). However, at the time of this experiment we did not have the Promethion Metabolic Cages nor Small Animal DXA scanner in place to conduct measures of Oconsumption or body composition.

LiCl supplementation does not alter body mass despite increases in food intake in male C57BL/6J high fat diet (HFD)-fed mice. A) Weekly analysis of body mass throughout the 12-week LiCl treatment period (n = 24 per group). B) Cumulative food intake throughout the 12-week LiCl treatment period (n = 24 per group). C) Daily food intake in control and LiCl fed mice (n = 24 per group). For A and B, two-way ANOVA was used to test the main effects of time, diet, and their potential interaction. **p < 0.01, ***p < 0.001, ****p < 0.0001 for HFD vs control (CON, chow); # p < 0.05, ## p < 0.01, ### p < 0.001, #### p < 0.0001 for HFD-Li vs control; † p < 0.05 for HFD vs HFD-Li. For C, a one-way ANOVA with a Tukey’s post hoc test was used, ***p < 0.001. All values are means ± SEM.

Figure 4 LiCl supplementation does not alter body mass despite increases in food intake in male C57BL/6J high fat diet (HFD)-fed mice. A) Weekly analysis of body mass throughout the 12-week LiCl treatment period (n = 24 per group). B) Cumulative food intake throughout the 12-week LiCl treatment period (n = 24 per group). C) Daily food intake in control and LiCl fed mice (n = 24 per group). For A and B, two-way ANOVA was used to test the main effects of time, diet, and their potential interaction. **p < 0.01, ***p < 0.001, ****p < 0.0001 for HFD vs control (CON, chow); # p < 0.05, ## p < 0.01, ### p < 0.001, #### p < 0.0001 for HFD-Li vs control; † p < 0.05 for HFD vs HFD-Li. For C, a one-way ANOVA with a Tukey’s post hoc test was used, ***p < 0.001. All values are means ± SEM.

LiCl supplementation does not alter body mass despite increases in food intake due to increases in daily energy expenditure in male C57BL/6J chow-fed mice. A) Weekly analysis of body mass throughout the 12-week LiCl treatment period (n = 10-12 per group). B) Cumulative food intake throughout the 12-week LiCl treatment period (n = 10-12 per group). C) Daily food intake in control and LiCl fed mice (n = 10-12 per group). D and E) Body fat and fat-free mass measured through small animal DXA (n = 10-12 per group). F) Absolute energy expenditure (VO 2 ) in control and LiCl treated mice at 6 and 11 weeks of treatment (n = 10-12 per group). G) Relative energy expenditure (per lean mass) in control and LiCl treated mice at 6 and 11 weeks of treatment (n = 10-12 per group). H) Cage activity in control and LiCl treated mice (n = 10-12 per group). For A, B, D-F, a two-way repeated ANOVA was used to test the main effects of time, lithium, and their potential interaction. For C and G, comparisons between control and LiCl were made using independent Student t-tests, *p < 0.05. All values are means ± SEM.

Figure 3 LiCl supplementation does not alter body mass despite increases in food intake due to increases in daily energy expenditure in male C57BL/6J chow-fed mice. A) Weekly analysis of body mass throughout the 12-week LiCl treatment period (n = 10-12 per group). B) Cumulative food intake throughout the 12-week LiCl treatment period (n = 10-12 per group). C) Daily food intake in control and LiCl fed mice (n = 10-12 per group). D and E) Body fat and fat-free mass measured through small animal DXA (n = 10-12 per group). F) Absolute energy expenditure (VO 2 ) in control and LiCl treated mice at 6 and 11 weeks of treatment (n = 10-12 per group). G) Relative energy expenditure (per lean mass) in control and LiCl treated mice at 6 and 11 weeks of treatment (n = 10-12 per group). H) Cage activity in control and LiCl treated mice (n = 10-12 per group). For A, B, D-F, a two-way repeated ANOVA was used to test the main effects of time, lithium, and their potential interaction. For C and G, comparisons between control and LiCl were made using independent Student t-tests, *p < 0.05. All values are means ± SEM.

The acute effects of LiCl treatment were also investigated in differentiated 3T3-L1 adipocytes. As observed above with C2C12 myocytes, 3 days of LiCl treatment did not alter the protein levels of UCP1 nor PGC-1α ( Figure 2 F). This again indicates that the ability of LiCl to affect UCP1 and PGC-1α levels requires that the treatment occur during differentiation.

We next examined the effects of LiCl treatment on UCP1 content in 3T3-L1 adipocytes. As with C2C12 cells, 3T3-L1 cells were treated with 0.5 mM LiCl for 10 days during differentiation. As expected, there was a significant increase in inhibitory serine9 phosphorylation of GSK3β with LiCl treatment compared to non-treated cells ( Figure 2 A). This inhibition was associated with an increase in cellular respiration ( Figure 2 B) and increased UCP1 and PGC-1α content ( Figure 2 C), further supporting the role of GSK3 in negatively regulating UCP1 expression in adipocytes. However, LiCl treatment had no impact on the mitochondrial footprint, indicating no increase in mitochondrial abundance ( Figure 2 D-E).

LiCl treatment inhibits GSK3 and increases UCP1 and cellular respiration in 3T3-L1 adipocytes. A) Increased inhibitory serine9 phosphorylation of GSK3β with 0.5 mM LiCl treatment (n = 5-6 per group). B) Resting cellular respiration rates in LiCl-treated and control 3T3-L1 adipocytes (n = 3-4 per group). C) UCP1 and PGC-1α protein content measured with western blotting (n = 6-8 per group). D-E) Mitochondrial footprint and branch length in 3T3-L1 adipocytes treated with LiCl. F) Acute (3 days) of 0.5 mM LiCl treatment in 10-day differentiated 3T3-L1 adipocytes did not alter UCP1 or PGC-1α protein content (n = 3 per group). *p < 0.05 using a Student’s t-test, with each n representing a technical replicate. All western blot data are presented as relative to control. All values are means ± SEM.

Figure 2 LiCl treatment inhibits GSK3 and increases UCP1 and cellular respiration in 3T3-L1 adipocytes. A) Increased inhibitory serine9 phosphorylation of GSK3β with 0.5 mM LiCl treatment (n = 5-6 per group). B) Resting cellular respiration rates in LiCl-treated and control 3T3-L1 adipocytes (n = 3-4 per group). C) UCP1 and PGC-1α protein content measured with western blotting (n = 6-8 per group). D-E) Mitochondrial footprint and branch length in 3T3-L1 adipocytes treated with LiCl. F) Acute (3 days) of 0.5 mM LiCl treatment in 10-day differentiated 3T3-L1 adipocytes did not alter UCP1 or PGC-1α protein content (n = 3 per group). *p < 0.05 using a Student’s t-test, with each n representing a technical replicate. All western blot data are presented as relative to control. All values are means ± SEM.

We have previously shown that LiCl treatment enhances myoblast differentiation (), which could influence our findings. Therefore, we next examined the acute effects of LiCl in 10-day differentiated myoblasts. Our results show that 3 days of LiCl treatment after C2C12 cells were already differentiated produced no effect on coupling ratio or RYR content ( Figure 1 J and K). This suggests that the effect of LiCl on lowering SERCA efficiency, presumably through increased RYR Caleak, only manifests when the treatment occurs throughout differentiation.

First, we questioned whether 0.5 mM LiCl would alter SERCA coupling ratio in C2C12 myoblasts, which is a dose that we have previously shown inhibits GSK3 in this cell line (). Cells were treated with or without LiCl for 7 days during differentiation. This resulted in a significant increase in inhibitory serine9 GSK3 phosphorylation ( Figure 1 A). PGC-1α is also negatively regulated by GSK3 (), and LiCl increased PGC-1α protein content nearly 4-fold ( Figure 1 A); however, this occurred without any significant changes in other mitochondrial markers ( Figure 1 B). Nonetheless, LiCl-treated C2C12 cells had elevated cellular respiration rates ( Figure 1 C) and 10 mM MgCllowered respiration rates in control and LiCl treated cells ( Figure 1 D). This reduction is presumed to have been caused by an inhibition of ryanodine receptor (RYR) Caleak that minimizes SERCA activity (). Calculating SERCA’s energetic contribution as the difference in respiration with and without 10 mM MgClrevealed a near 2-fold increase with LiCl treatment ( Figure 1 E). Corresponding well with this data, we observed a significant reduction in Cauptake with no change in SERCA activity, leading to a significant reduction in SERCA’s apparent coupling ratio with LiCl treatment ( Figure 1 F-H). This suggests that the increase in SERCA’s energy consumption is due to a reduction in Catransport efficiency. To investigate the cellular mechanisms leading to this reduction, we examined the protein levels of SERCA1a/2a, SLN, NNAT, and RYR1. No changes were detected in either SERCA isoform, SLN, or NNAT; however, we did observe a significant upregulation in RYR1 ( Figure 1 I). Thus, the increase in SERCA energy expenditure resulting from a reduction in SERCA’s apparent coupling ratio may be due to enhanced RYR Caleak in LiCl treated C2C12 cells.

LiCl treatment inhibits GSK3, increases cellular respiration and promotes SERCA uncoupling in C2C12 myocytes. A) Increased inhibitory serine9 phosphorylation of GSK3β and PGC-1α protein with 0.5 mM LiCl treatment (n = 5-6 per group). B) No changes in mitochondrial proteins: pyruvate dehydrogenase (PDH), cytochrome c, cytochrome c oxidase subunit IV (COXIV), or citrate synthase were found with 0.5 mM LiCl treatment (n = 3 per group).C) Resting cellular respiration rates in LiCl-treated and control C2C12 cells (n = 6 per group). Cellular respiration in the presence and absence of 10 mM MgCl 2 (D) to calculate SERCA’s energetic contribution (E) (n = 3 per group). F) Ca 2+ uptake is significantly reduced without altering SERCA activity (G), leading to a significant reduction in SERCA’s apparent coupling ratio (H) (n = 6 per group). I) Western blot analyses of SERCA isoform, SLN, NNAT and RYR1 in LiCl-treated and control C2C12 cells (n = 5-6 per group). J-K) Acute (3 days) of 0.5 mM LiCl treatment in 10-day differentiated C2C12 cells did not alter SERCA coupling ratio or RYR content (n = 3 per group). *p < 0.05, **p < 0.01 using a Student’s t-test, with each n representing a technical replicate. All western blot data are presented as relative to control. All values are means ± SEM.

Figure 1 LiCl treatment inhibits GSK3, increases cellular respiration and promotes SERCA uncoupling in C2C12 myocytes. A) Increased inhibitory serine9 phosphorylation of GSK3β and PGC-1α protein with 0.5 mM LiCl treatment (n = 5-6 per group). B) No changes in mitochondrial proteins: pyruvate dehydrogenase (PDH), cytochrome c, cytochrome c oxidase subunit IV (COXIV), or citrate synthase were found with 0.5 mM LiCl treatment (n = 3 per group).C) Resting cellular respiration rates in LiCl-treated and control C2C12 cells (n = 6 per group). Cellular respiration in the presence and absence of 10 mM MgCl 2 (D) to calculate SERCA’s energetic contribution (E) (n = 3 per group). F) Ca 2+ uptake is significantly reduced without altering SERCA activity (G), leading to a significant reduction in SERCA’s apparent coupling ratio (H) (n = 6 per group). I) Western blot analyses of SERCA isoform, SLN, NNAT and RYR1 in LiCl-treated and control C2C12 cells (n = 5-6 per group). J-K) Acute (3 days) of 0.5 mM LiCl treatment in 10-day differentiated C2C12 cells did not alter SERCA coupling ratio or RYR content (n = 3 per group). *p < 0.05, **p < 0.01 using a Student’s t-test, with each n representing a technical replicate. All western blot data are presented as relative to control. All values are means ± SEM.

Discussion

In this study, we examined whether GSK3 inhibition, via LiCl supplementation, would promote SERCA uncoupling in C2C12 cells and in skeletal muscles obtained from chow-fed and high fat-fed mice. We also examined whether LiCl supplementation would enhance UCP1 expression in 3T3-L1 adipocytes, and in iWAT and BAT from chow-fed and high fat-fed mice.

2 experiments that indirectly inhibit SERCA by lowering SR Ca2+ leak ( 3 Smith I.C.

Bombardier E.

Vigna C.

Tupling A.R. ATP consumption by sarcoplasmic reticulum Ca(2)(+) pumps accounts for 40-50% of resting metabolic rate in mouse fast and slow twitch skeletal muscle. 2+ leak as enhanced RYR Ca2+ leak is known to cause futile Ca2+ cycling with SERCA, thereby increasing energy expenditure and heat release ( 42 Ikeda K.

Kang Q.

Yoneshiro T.

Camporez J.P.

Maki H.

Homma M.

Shinoda K.

Chen Y.

Lu X.

Maretich P.

Tajima K.

Ajuwon K.M.

Soga T.

Kajimura S. UCP1-independent signaling involving SERCA2b-mediated calcium cycling regulates beige fat thermogenesis and systemic glucose homeostasis. 43 da Costa D.C.

Landeira-Fernandez A.M. Thermogenic activity of the Ca2+-ATPase from blue marlin heater organ: regulation by KCl and temperature. 44 Berchtold M.W.

Brinkmeier H.

Muntener M. Calcium ion in skeletal muscle: its crucial role for muscle function, plasticity, and disease. 45 Maclennan D.H.

Zvaritch E. Mechanistic models for muscle diseases and disorders originating in the sarcoplasmic reticulum. 18 Markussen L.K.

Winther S.

Wicksteed B.

Hansen J.B. GSK3 is a negative regulator of the thermogenic program in brown adipocytes. 2+ transport efficiency. Using cellular models first, our results demonstrate that LiCl treatment enhanced respiration in C2C12 cells and 3T3-L1 adipocytes. In C2C12 cells, the increase in respiration was accounted for, at least in part, by a ∼2-fold increase in SERCA’s energetic contribution to respiration as estimated with MgClexperiments that indirectly inhibit SERCA by lowering SR Caleak (). This increase in SERCA’s energy expenditure was likely due to enhanced SERCA uncoupling which was mediated through an increase in RYR rather than any changes in known regulatory proteins that can uncouple the SERCA pump. An increase in RYR can lower SERCA’s transport efficiency via Caleak as enhanced RYR Caleak is known to cause futile Cacycling with SERCA, thereby increasing energy expenditure and heat release (). In 3T3-L1 adipocytes, we attribute the increase in energy expenditure to an increase in PGC-1α and UCP1, which was expected based on previous work showing that GSK3 negatively regulated UCP1 expression in BAT (). However, in both adipocytes and myocytes, the effect of LiCl treatment was abolished when treatment occurred acutely, for 3 days, after cellular differentiation. Together, these results show that in cultured adipocytes and myocytes, LiCl supplementation and GSK3 inhibition throughout cellular differentiation can increase energy expenditure via alterations in UCP1 and SERCA Catransport efficiency.

34 Kurgan N.

Bott K.N.

Helmeczi W.E.

Roy B.D.

Brindle I.D.

Klentrou P.

Fajardo V.A. Low dose lithium supplementation activates Wnt/beta-catenin signalling and increases bone OPG/RANKL ratio in mice. 35 Hamstra S.I.

Kurgan N.

Baranowski R.W.

Qiu L.

Watson C.J.F.

Messner H.N.

MacPherson R.E.K.

MacNeil A.J.

Roy B.D.

Fajardo V.A. Low-dose lithium feeding increases the SERCA2a-to-phospholamban ratio, improving SERCA function in murine left ventricles. 46 Hamstra S.I.

Whitley K.C.

Baranowski R.W.

Kurgan N.

Braun J.L.

Messner H.N.

Fajardo V.A. The role of phospholamban and GSK3 in regulating rodent cardiac SERCA function. 25 Whitley K.C.

Hamstra S.I.

Baranowski R.W.

Watson C.J.F.

MacPherson R.E.K.

MacNeil A.J.

Roy B.D.

Vandenboom R.

Fajardo V.A. GSK3 inhibition with low dose lithium supplementation augments murine muscle fatigue resistance and specific force production. To determine if these effects would translate to the in vivo setting, we treated mice with 10 mg/kg/day of LiCl for 12 weeks under chow-fed and high fat-fed conditions. Though we have previously shown that this dose results in a serum concentration that is much lower than that used in our cell culture experiments (0.02 mM vs 0.5 mM, respectively) (), this dose and duration were effective in inhibiting GSK3 and increasing PGC-1α in murine soleus muscle (). Here, our results show that LiCl supplementation did not alter body mass despite increasing total food intake in both chow and HFD conditions. We attribute this effect to an increase in daily energy expenditure in LiCl treated mice. However, since this was only measured in chow fed conditions, we can only assume that this was also the case in HFD fed mice.

3 Smith I.C.

Bombardier E.

Vigna C.

Tupling A.R. ATP consumption by sarcoplasmic reticulum Ca(2)(+) pumps accounts for 40-50% of resting metabolic rate in mouse fast and slow twitch skeletal muscle. 2+ transport efficiency may influence total daily energy expenditure, given the energetic nature of skeletal muscle. In chow fed mice, the reduction in SERCA Ca2+ transport efficiency was observed at 1000 nM [Ca2+] free , however, statistical significance was lost at 500 nM [Ca2+] free . Similarly, when fed a HFD, LiCl treatment significantly reduced SERCA coupling ratio, though we could only measure this at 1500 nM [Ca2+] free as Ca2+ uptake in these samples did not reach levels below 1000 nM. While the exact reason for this remains unknown, we believe that this finding is indicative of an impairment in Ca2+ uptake, or excessive leak of Ca2+ from the SR with a HFD ( 47 Jain S.S.

Paglialunga S.

Vigna C.

Ludzki A.

Herbst E.A.

Lally J.S.

Schrauwen P.

Hoeks J.

Tupling A.R.

Bonen A.

Holloway G.P. High-fat diet-induced mitochondrial biogenesis is regulated by mitochondrial-derived reactive oxygen species activation of CaMKII. In skeletal muscle, the SERCA pumps account for ∼50% of resting metabolic rate and was estimated to account for 22-25% of total daily energy expenditure (). Thus, the SERCA pumps may be viewed as a metabolic hub in muscle, and alterations in SERCA Catransport efficiency may influence total daily energy expenditure, given the energetic nature of skeletal muscle. In chow fed mice, the reduction in SERCA Catransport efficiency was observed at 1000 nM [Ca, however, statistical significance was lost at 500 nM [Ca. Similarly, when fed a HFD, LiCl treatment significantly reduced SERCA coupling ratio, though we could only measure this at 1500 nM [Caas Cauptake in these samples did not reach levels below 1000 nM. While the exact reason for this remains unknown, we believe that this finding is indicative of an impairment in Cauptake, or excessive leak of Cafrom the SR with a HFD (). Nonetheless, under both chow fed and high fat fed conditions, LiCl treatment significantly reduced SERCA’s apparent coupling ratio which could contribute at least in part to the increase in daily energy expenditure.

2+ transport efficiency by increasing the expression of SERCA uncoupling proteins. In chow fed mice, LiCl treatment significantly increased NNAT without altering SLN content. This sole increase in NNAT content in chow fed conditions was sufficient in lowering SERCA Ca2+ transport efficiency. Conversely, in high fat fed mice, both SLN and NNAT were significantly upregulated with LiCl treatment. This is interesting since both SLN ( 48 Gamu D.

Juracic E.S.

Fajardo V.A.

Rietze B.A.

Tran K.

Bombardier E.

Tupling A.R. Phospholamban deficiency does not alter skeletal muscle SERCA pumping efficiency or predispose mice to diet-induced obesity. 49 Braun J.L.

Teng A.C.T.

Geromella M.S.

Ryan C.R.

Fenech R.K.

MacPherson R.E.K.

Gramolini A.O.

Fajardo V.A. Neuronatin promotes SERCA uncoupling and its expression is altered in skeletal muscles of high-fat diet-fed mice. mKD mice, the results presented in this study establish a novel regulatory link between GSK3 and the expression of both SLN and NNAT and SERCA Ca2+ transport efficiency in skeletal muscle. Unlike C2C12 cells, the promotion of SERCA uncoupling in vivo did not occur with any changes in RYR content, which highlights an important difference between cell culture and in vivo models. Instead, in chow fed and high fat fed mice, LiCl treatment lowered SERCA Catransport efficiency by increasing the expression of SERCA uncoupling proteins. In chow fed mice, LiCl treatment significantly increased NNAT without altering SLN content. This sole increase in NNAT content in chow fed conditions was sufficient in lowering SERCA Catransport efficiency. Conversely, in high fat fed mice, both SLN and NNAT were significantly upregulated with LiCl treatment. This is interesting since both SLN () and NNAT () content were recently shown to be lowered in soleus muscles obtained from mice fed a HFD. It is thus tempting to speculate that the enhancement in GSK3 activation observed here with high fat feeding (and blunted with LiCl treatment) could contribute to the lowered expression of SLN and NNAT content in mice fed a HFD. In further support of this, partial skeletal muscle specific GSK3 knockdown increased both SLN and NNAT in soleus muscle ultimately reducing SERCA’s apparent coupling ratio. While future studies in our lab will investigate the effects of high fat feeding in the GSK3mice, the results presented in this study establish a novel regulatory link between GSK3 and the expression of both SLN and NNAT and SERCA Catransport efficiency in skeletal muscle.

In adipose tissue, we questioned whether LiCl treatment could increase UCP1 in iWAT and BAT by inhibiting GSK3, ultimately contributing to changes in daily energy expenditure. Under chow fed conditions, our results show that LiCl treatment significantly inhibited GSK3 in iWAT, which in turn increased UCP1 and PGC-1α content and promoted a multilocular histological phenotype. To our knowledge, ours is the first study establishing a beiging-like effect of LiCl and GSK3 inhibition in iWAT obtained from mice. However, there was no effect of LiCl on UCP1 or PGC-1α in BAT – a result we attributed to an inability to inhibit GSK3 in this specific adipose depot. Interestingly, when mice were fed a HFD, LiCl treatment did not alter UCP1 or PGC-1α content nor the histological appearance of iWAT, which we also attribute to an inability to inhibit GSK3. Similar to our findings in skeletal muscle, GSK3 was more active in iWAT from high fat fed mice with significantly lowered serine9 phosphorylation. However, unlike skeletal muscle, LiCl treatment could not attenuate this effect. Thus in combination with our in vitro results, these findings demonstrate that when GSK3 is inhibited with LiCl, UCP1 and PGC-1α is increased and may elevate energy expenditure. Future studies from our lab will investigate the effects of more potent GSK3 inhibitors and adipose-tissue specific GSK3 knockdown on iWAT beiging.

23 Choi S.E.

Jang H.J.

Kang Y.

Jung J.G.

Han S.J.

Kim H.J.

Kim D.J.

Lee K.W. Atherosclerosis induced by a high-fat diet is alleviated by lithium chloride via reduction of VCAM expression in ApoE-deficient mice. The enhancement of GSK3 activation observed in both adipose and muscle obtained from mice fed a HFD may suggest a role for GSK3 overactivation in diet-induced obesity. In this context, previous work conducted by Choi et al., (2010) found that treating mice with LiCl for 10 mg/kg/day (same dose used in this study) reduced the weight gained after a HFD (). Moreover, lithium is known to have insulin sensitizing effects and Choi et al., (2010) found that LiCl treatment in high fat fed mice lowered fasting glucose levels. However, in our hands we did not find any alterations in body mass or glucose handling in chow fed or high fat fed conditions. It should be noted that Choi et al., (2021) used a HFD that only comprised 20% fat, whereas we provided a 60% fat diet in our study, which strongly induced obesity in mice and may have overpowered any effect of LiCl. Moreover, the increase in food intake found here in chow fed and HFD fed mice treated with LiCl could have masked the effect of LiCl on lowering body mass and potentially glucose handling. Thus, our study is limited in that we did not conduct pair-feeding experiments. Our study is also limited in that we only utilized male mice, primarily to avoid potential confounding effects of hormonal shifts in female mice. However, future studies should examine the effects of LiCl treatment and GSK3 inhibition in female mice.

50 Davis J.

Desmond M.

Berk M. Lithium and nephrotoxicity: Unravelling the complex pathophysiological threads of the lightest metal. 22 Mangge H.

Bengesser S.

Dalkner N.

Birner A.

Fellendorf F.

Platzer M.

Queissner R.

Pilz R.

Maget A.

Reininghaus B.

Hamm C.

Bauer K.

Rieger A.

Zelzer S.

Fuchs D.

Reininghaus E. Weight Gain During Treatment of Bipolar Disorder (BD)—Facts and Therapeutic Options. 51 Peselow E.D.

Dunner D.L.

Fieve R.R.

Lautin A. Lithium carbonate and weight gain. Nonetheless, our study does provide novel insight into the effect of lithium on appetite and food intake. Lithium is conventionally used in the treatment of bipolar disorder although clinical use of lithium is prescribed at higher doses (serum concentration of 0.5-1.0 mM) in order to overcome the blood brain barrier and exert beneficial effects on mental health (). At this high dose, it has been suggested that some patients undergoing lithium therapy are at an increased risk of developing obesity (), perhaps due to an increase in appetite (). Here, we observed that low dose LiCl supplemented mice also had increased appetite, consuming more food than controls, which is consistent with high dose lithium therapy. However, at low doses of LiCl, our results show that this increase in appetite and food intake was met with an increase in SERCA uncoupling and/or UCP1 expression, which we believe attenuated any additional weight gain expected with an increase in food intake. Whether the promotion of SERCA uncoupling and UCP1 in skeletal muscle and adipose tissue, respectively, is lost with higher doses of lithium should be investigated with future studies.

In conclusion, our study examined the effects of GSK3 inhibition via low dose LiCl supplementation on SERCA uncoupling in skeletal muscle and UCP1 expression in adipose tissue. Our results provide novel regulatory connections between GSK3 and the expression of NNAT and SLN and SERCA uncoupling in skeletal muscle; and for the first time show that GSK3 inhibition can promote a beiging-like effect in iWAT.