Numerous studies have suggested that beige adipocyte abundance is correlated with improved metabolic performance, but direct evidence showing that beige adipocyte expansion protects animals from the development of obesity is missing. Previously, we have described that the liver kinase b1 (LKB1) regulates beige adipocyte renaissance in subcutaneous inguinal white adipose tissue (iWAT) through a class IIa histone deacetylase 4 (HDAC4)-dependent mechanism. This study investigates the physiological impact of persistent beige adipocyte renaissance in energy homeostasis in mice. Here we present that the transgenic mice H4-TG, overexpressing constitutively active HDAC4 in adipocytes, showed beige adipocyte expansion in iWAT at room temperature. H4-TG mice exhibited increased energy expenditure due to beige adipocyte expansion. They also exhibited reduced adiposity under both normal chow and high-fat diet (HFD) feeding conditions. Specific ablation of beige adipocytes reversed the protection against HFD-induced obesity in H4-TG mice. Taken together, our results directly demonstrate that beige adipocyte expansion regulates adiposity in mice and targeting beige adipocyte renaissance may present a novel strategy to tackle obesity in humans.
Two types of UCP1+ adipocytes have been characterized by their differences in anatomic localization, developmental origin and molecular signature in rodents (Harms & Seale 2013, Kajimura et al. 2015). The classical brown adipocytes are present in interscapular and perirenal adipose tissues, and they originate from Myf5+/Pax7+ skeletal muscle progenitors (Seale et al. 2008). The beige adipocytes are formed and clustered within the subcutaneous inguinal white adipose tissue (iWAT) in response to β-adrenergic receptor (βAR) stimulation (Seale et al. 2011, Walden et al. 2012, Wu et al. 2012, Cohen et al. 2014). Beige adipocytes are either derived from mural progenitors that reside in iWAT (Lee et al. 2012, Wu et al. 2012, Vishvanath et al. 2015, Berry et al. 2016) or they may arise from transdifferentiation of white adipocytes (Cinti 2009, Barbatelli et al. 2010, Rosenwald et al. 2013).
Several studies using anatomical and gene profiling approaches have indicated that interscapular brown fat from human infants and neck brown fat from human adults shares features with classical murine brown adipocytes, while supraclavicular fat of adult humans mainly consists of beige adipocytes (Sharp et al. 2012, Cypess et al. 2013, 2015, Jespersen et al. 2013, Lidell et al. 2013, Lee et al. 2014). However, specific functions of beige adipocytes have not been addressed for two major reasons. Beige adipocytes tend to be activated together with brown adipocytes in response to the same stimuli, and the tools currently used for genetic manipulations in mice cannot distinguish between these two cell types.
The beige adipocytes formed postnatally in iWAT can lose thermogenic gene expression and multilocular morphology at the adult stage, but βAR stimuli can restore their ‘beigeing’ properties, a phenomenon termed as ‘beige adipocyte renaissance’ (Lee et al. 2015, Wang et al. 2017). We have demonstrated that beige adipocyte renaissance is governed by cyclic adenosine 3’,5’-monophosphate (cAMP) and histone deacetylase 4 (HDAC4) signaling in white adipocytes (Wang et al. 2017). We have shown that HDAC4 is activated in adipocytes of both iWAT and interscapular brown adipose tissue (iBAT) from pan adipocyte-specific liver kinase b1 (Lkb1)-knockout mice (the Lkb1AKO mice), but its activation is necessary for beige adipocyte renaissance in iWAT. The mice expressing constitutively active HDAC4 in adipocytes (H4-TG) exhibit persistent beige adipocyte renaissance in iWAT without affecting thermogenic capacity in iBAT and improved metabolic performance under both normal chow and HFD conditions. Ablating beige adipocytes in H4-TG mice abolished the protective effects on HFD-induced obesity, suggesting that beige adipocytes can regulate adiposity.
Materials and methods
All animal experiments were approved by the UCSF IACUC Committee in adherence to NIH guidelines and policies. Rosa-iDTR (JAX#007900) mice were obtained from JAX. Ucp1-Cre (JAX#024670) mice were provided by Dr Evan Rosen. A 5.6 kb fragment of the ap2 promoter (provided by Dr Ronald Evans) and a 4 kb fragment encoding GFP-HDAC4.S245/467/632A (mutation of the three Salt-inducible kinase phosphorylation sites to alanine) were used to generate the ap2-GFP-HDAC4.3A transgenic mouse. The transgenic mice (line C and D) had been backcrossed into C57bl/6J background for more than five generations before metabolic analysis. Both male and female mice were used for thermogenic studies. Male mice were used in metabolic studies because they were more prone to HFD-induced obesity and other metabolic abnormalities (Yang et al. 2014). Mice were housed in a temperature-controlled environment at 22°C under a 12-h light:darkness cycle with free access to water and food (PicoLab Rodent Diet 20, #5053). For thermoneutral housing, ~4-week-old mice were placed in a 30°C rodent chamber (Power Scientific RIS52SD Rodent Incubator) for additional 4 weeks.
Browning of white adipose tissue
For cold stimulation experiments, mice were placed in an 8°C rodent chamber (Power Scientific RIS52SD Rodent Incubator) for 7 days. For CL 316,243 (CL) treatment, mice were injected intraperitoneally with 1 mg/g daily for 7 days. To identify beige adipocytes, subcutaneous inguinal fat (iWAT) was isolated and Ucp1 RNA and protein levels were analyzed. Paraffin sections and hematoxylin and eosin (H&E) staining were performed at AML Labs using standard protocols.
Beige adipocyte ablation
Three-week-old male mice were injected with a single diphtheria toxin (DT) intraperitoneally at the dose of 10–25 ng/mouse (Biological Laboratories Inc.).
Total energy expenditure (TEE) and respiratory exchange ratio (RER) were measured using CLAMS (Columbus Instruments), installed under a constant temperature-controlled environment, under a 12-h light:darkness cycle with free access to food and water. To determine βAR-induced TEE, CL (1 mg/g) was injected intraperitoneally at ~10:00 h of the last day of CLAMS run. The raw data were normalized per mouse (Butler & Kozak 2010, Tschop et al. 2012). The energy expenditure of physical activity (PAEE) and the thermogenic effect of food (TEF) were calculated as described before (Abreu-Vieira et al. 2015). MRI scans for fat and lean mass were performed using an Echo MRI-100 instrument according to manufacturer’s instructions. Investigators were blinded to mice genotypes for CLAMS, which were performed by the UCSF Diabetes and Endocrinology Research Center Metabolic Research Unit. For HFD studies, six-week-old male mice were transferred to a 60% fat diet (Research Diets, D12492) at room temperature. Body weight was monitored weekly. For insulin tolerance test (ITT), mice were fasted 4–6 h before intraperitoneal administration of insulin (Humulin; 0.75 U/kg). Serum glucose was measured from tail vein at indicated time points with a glucometer (Contour, Bayer). Serum TAG levels were measured using Infinity Triglycerides Reagent kit (Thermo Scientific), and serum insulin levels were determined using mouse insulin ultrasensitive ELISA kit (Alpco).
Tissues were fixed in 10% formalin and processed and stained at AML Laboratories. Cell size was measured using ImageJ. Adipocyte size distribution was calculated using total adipocyte numbers counted in different areas from multiple sections.
Cold tolerance test (CTT)
Six- to 8-week-old male and female mice were single-housed with free access to food and water during CTT. Rectal temperature was monitored hourly during 4°C cold challenge with a BAT-12 Microprobe Thermometer (Physitem Instruments). Cold challenge was started at 11:00 h (4 h after light on).
Tissue catecholamine measurement
Catecholamine levels were measured by ELISA (Rocky Mountain Diagnostics, # BA E-6600) according to the manufacturer’s protocol. Tissues were homogenized in homogenization buffer (0.01 N HCl, 1 mM EDTA, 4 mM Na2S2O5) and centrifuged at 15,000 g for 15 min at 4°C. The supernatants were collected and stored at −80°C prior to quantification. All measurements were normalized to total tissue protein concentration.
Respiration ex vivo
For measurements of tissue respiration activity ex vivo, the minced inguinal fat tissues were added in the chamber that contained prewarmed respiration buffer at 37°C. Then the tissue oxygen consumption rates were measured using a Clarke electrode (Strathkelvin Instruments) using the established protocol (Li & Graham 2012). Oxygen consumption was normalized to tissue weight.
RNA from cells and tissues was extracted using RNeasy Mini Kit (Qiagen). Total RNA (1 μg) was reverse-transcribed by iScript cDNA synthesis kit (Bio-Rad) and cDNA was used for real time PT-PCR (CFX384, Bio-Rad), using 2 ng of cDNA template and a primer concentration of 400 nM. Values were normalized to 36b4 housekeeping gene. Primer sequences were listed in Supplementary Table 1 (see section on supplementary data given at the end of this article). Pooled cDNA from iWAT of ~8-week-old male WT and H4-TG mice were used for Affymetrix Mouse Gene Array 1.0 using the protocol provided by manufacturer. The data were deposited in NCBI GEO (# GSE106310). Gene ontology analysis was performed using Metascape at
Tissues were homogenized on ice in lysis-buffer (50 mM Tris–HCl, 150 mM NaCl, 1 mM EDTA, 6 mM EGTA, 20 mM NaF, 1% Triton X-100 and protease inhibitors) for 15–20 min. After centrifugation at 13,000 rpm for 15 min, supernatants were reserved for protein quantification followed by SDS-PAGE analysis. The following antibodies were used: anti-Hdac4 (Cell Signaling Technology, #7628), anti-Hsp90 (Santa Cruz Biotechnology, #SC-7947) and anti-Ucp1 (Sigma, #U6382).
We used GraphPad Prism 6.0 to assess data for normal distribution and similar variance between groups. Data were presented as the mean ± s.d. Statistical significance was determined using an unpaired two-tailed Student’s t test with unequal variance between groups: ns: not significant, *P < 0.05 and **P < 0.01. We selected sample size for animal experiments based on numbers typically used in similar published studies. We did not perform randomization of animals or predetermine sample size by a statistical method.
Mice expressing constitutively active HDAC4 exhibit beige adipocyte expansion in iWAT
We have employed pan adipocyte-specific Adiponectin-Cre and brown/beige adipocyte-specific Ucp1-Cre to investigate roles of liver kinase b1 (LKB1) in adipocytes to demonstrate that LKB1 regulates two distinct pathways in brown and beige adipocytes. In brown adipocytes, LKB1 appears to govern thermogenic capacity independently of HDAC4 (Masand et al. 2018); but in white adipocytes, it controls HDAC4 activity (through salt-inducible kinases/SIKs) to regulate beige adipocyte renaissance non-cell autonomously (Wang et al. 2017). We reasoned that HDAC4 activation in white adipocytes could promote beige adipocyte renaissance in iWAT without affecting adaptive thermogenesis in BAT, and this unique requirement of HDAC4 for beige adipocyte renaissance could enable us to address beige adipocyte-specific function in energy homeostasis. In this report, we have generated a transgenic mouse line, expressing constitutively active HDAC4 (GFP-HDAC4.3A with all SIKs phosphorylation sites mutated to alanine) under the control of an AP2 promoter (Fig. 1A, top). The transgene was expressed in both white and brown adipose tissues, but not in muscle and liver in two separate lines (line C and line D) (Fig. 1A, bottom, Supplementary Fig. 3A). The line D (referred as H4-TG#D throughout this study) was selected for further detailed characterizations. Immunoblots further confirmed that the transgene GFP-HDAC4.3A was over-expressed in mature adipocytes, not in stromal vascular fraction cells (Supplementary Fig. 1A). The male H4-TG#D mice had similar Ucp1 expression and multilocular brown adipocyte morphology, and they were similarly cold resistant as WT mice at ~8-week of age (Supplementary Fig. 1B, C, D and E). In contrast, we have observed increased Ucp1 mRNA levels in iWAT from these H4-TG#D male mice at 8 weeks of age, although there were no detectable differences in Ucp1 transcript levels at 3 weeks of age (Fig. 1B). Histological analysis (H&E staining) also confirmed the presence of multilocular beige adipocyte expansion in iWAT of adult H4-TG#D mice (Fig. 1C), which resembled the beige adipocyte expansion phenotype observed in Lkb1AKO mice (Wang et al. 2017). Gender did not affect beige adipocyte expansion caused by constitutively HDAC4 action, because elevated thermogenic gene expression and multilocular morphology were also observed in the iWAT of female H4-TG#D mice at ~8 weeks of age (Supplementary Fig. 2A and B). Beige adipocyte presence in H4-TG#D mice also required sympathetic inputs, as thermoneutral housing reduced Ucp1 mRNA and UCP1 protein levels in iWAT of H4-TG mice (Fig. 1B and Supplementary Fig. 6A). On the other hand, CL316243 (a β3 specific agonist, abbreviated as CL) administration did not further induce UCP1 protein levels in adult H4-TG#D mice (Fig. 1D). Notably, the H4-TG#C mice also exhibited elevated thermogenic gene expression and UCP1 protein levels in iWAT (Supplementary Fig. 3B and C).
We have described that adipocyte LKB1 deficiency led to HDAC4 activation in adipocytes, which promoted beige adipocyte renaissance in iWAT (Fig. 2A) (Wang et al. 2017). The gene profiling study revealed that there were over 500 genes were upregulated in iWAT by LKB1 deficiency in white adipocytes. We then performed microarray experiment in the H4-TG#D mice and compared the transcriptomic changes in the iWAT caused by adipocyte LKB1 deficiency (in Lkb1AKO mice) and constitutively active HDAC4 overexpression (in H4-TG mice). In fact, about half of genes upregulated in the iWAT of Lkb1AKO mice were also upregulated in H4-TG#D mice (Fig. 2B). Detailed gene ontology (GO) analysis on these shared regulated genes suggested that the top four enriched GO terms were related to muscle and metabolic functions (Fig. 2C), consistent with the histological observation of the expansion of thermogenic active beige adipocytes in iWAT. The q-PCR analysis further confirmed that many thermogenic signature genes and metabolic genes were upregulated in iWAT of H4-TG#D mice (Fig. 2D). Collectively, adipocyte HDAC4 activation is necessary and sufficient to promote beige adipocyte expansion in iWAT of adult mice. The AP2 promoter was not exclusively active in adipocytes (Lee et al. 2013, Mullican et al. 2013). The sympathetic activity mainly regulates the catecholamine turnover of in the iWAT, and the issue catecholamine levels were not affected (WT 18.8 vs H4-TG 16.2 pg/µg protein). The contribution of macrophage to catecholamine secretion is under debate (Qiu et al. 2014, Fischer et al. 2017, Reitman 2017). Inflammation and macrophage marker genes were not altered in iWAT of H4-TG#D mice (Supplementary Fig. 4A), suggesting that catecholamines produced in immune cells is unlikely causing the beige adipocyte expansion in H4-TG#D mice. However, we cannot fully rule out the possibility that HDAC4 in other cell types may lead to the beige adipocyte expansion in iWAT.
Previously, we have characterized an Ucp1-Cre inducible diphtheria toxin receptor (iDTR)-mediated cell ablation system (the Ucp1-Cre;ROSA-iDTR mice), where beige adipocytes can be permanently ablated in iWAT upon DT injection (Fig. 3A) (Wang et al. 2017). We then generated H4-TG#D;Ucp1-Cre;Rosa-iDTR (abbreviated as H4-TG#D;Cre+) mice. DT injection at 3 weeks of age ablated beige adipocytes in H4-TG#D;Cre+ mice (Fig. 3B and C). In these mice, CL was unable to induce Ucp1 expression anymore at adult stage (Fig. 3D). Thus, overexpression of constitutively active GFP-HDAC4.3A was not sufficient to convert white adipocytes to beige adipocytes cell autonomously in beige adipocyte-ablated H4-TG#D;Cre+ mice. Instead, HDAC4 activation in white adipocytes maintained beigeing properties of preexisting beige adipocytes in iWAT, suggesting that beige adipocyte expansion in H4-TG#D mice was formed through the same persistent ‘renaissance’ mechanism observed in the Lkb1AKO mice.
H4-TG mice show increased energy expenditure
Beige adipocytes are capable of UCP1-dependent and -independent thermogenesis (Shabalina et al. 2013, Kazak et al. 2015, Ikeda et al. 2017). We then measured whether beige adipocyte expansion in iWAT contributed to TEE at room temperature. Indeed, there was about 50% increase of thermogenic capacity of iWAT from H4-TG#D mice ex vivo (Supplementary Fig. 4B). Indirect calorimetry experiments showed that 8-week-old H4-TG#D male mice exhibited increased basal TEE particularly at night (Fig. 4A), although CL-induced EE was not affected (Supplementary Fig. 5). Scatter plots confirmed that elevated TEE in H4-TG#D mice was independent of body weight (Fig. 4B). RER was not altered (Fig. 4C). Food intake and physical activities were slightly increased (Fig. 4D and E). Regarding the partitioning of TEE, the percentage of energy expenditure of physical activity (%PAEE) was increased from 14 to 20% at night, but that of thermogenic effect of food (%TEF) was not altered in the H4-TG#D mice (Fig. 4F).
The contribution of adaptive thermogenesis to TEE is dependent on ambient temperature (Cannon & Nedergaard 2011, Maloney et al. 2014, Abreu-Vieira et al. 2015, Ganeshan & Chawla 2017). We next determined whether the absence of thermogenic active beige adipocytes in H4-TG#D mice at thermoneutrality would affect energy expenditure. At thermoneutrality, Ucp1 mRNA and protein levels were decreased, and brown adipocytes exhibited unilocular morphology in both WT and H4-TG#D mice (Supplementary Fig. 6A, B and C). Elevated Ucp1 expression in the iWAT was also abolished in H4-TG#D mice (Fig. 1B and Supplementary Fig. 6A). Basal TEE at night in H4-TG#D mice was not increased (Supplementary Fig. 6D and E). RER and food intake were not different (Supplementary Fig. 6F and G). Physical activity in H4-TG#D mice remained elevated at thermoneutrality, although the %PAEE was no longer elevated (Supplementary Fig. 6H and I). These data suggest that expanded beige adipocytes in iWAT and increased physical activity contributed to increased basal TEE in H4-TG#D mice at room temperature.
H4-TG mice show improved metabolic performance
Thus, we employed H4-TG#D mice as a model system to study metabolic consequences of beige adipocyte expansion. Under normal chow feeding, H4-TG#D mice had reduced body weight in an age-dependent manner (Fig. 5A). At ~6 months of age, the reduced body weight in H4-TG#D mice was due to reduced fat mass, but not lean mass (Fig. 5B and C). Food intake was also increased at night (Fig. 5D), possibly secondary to reduced obesity. H4-TG#D mice had smaller adipocytes (measured by histological analysis on H&E sections) and less macrophage infiltration (determined by q-PCR analysis of Cd68 and F4/80) in epididymal WAT (eWAT) (Fig. 5E, F and G). Serum triglyceride, insulin and glucose levels were all reduced in these mice (Fig. 5H, I and J), suggesting that H4-TG#D mice showed reduced adiposity and improved glucose and lipid metabolism.
We have demonstrated that the persistent beige adipocyte population in Lkb1AKO mice is sustained during HFD-induced obesity (Wang et al. 2017). Since H4-TG mice and Lkb1AKO mice exhibited beige adipocyte expansion in iWAT through the same ‘renaissance’ mechanism, we also analyzed whether H4-TG#D male mice could also maintain beige adipocyte expansion even after HFD. Histological analysis confirmed the presence of multilocular beige adipocytes in the iWAT after 4–5 weeks HFD (Supplementary Fig. 7B). Immunblot analysis also showed that Ucp1 protein levels remained elevated (Supplementary Fig. 7A). Additional beige adipocyte markers that were upregulated in H4-TG#D mice at normal chow also remained elevated after 4-week HFD (Fig. 2D and Supplementary Fig. 9A).
Then we investigated whether H4-TG#D mice were also protected against HFD-induced metabolic abnormalities. Indeed, H4-TG#D mice showed reduced body weight on HFD (as early as in 2 weeks), and the difference in weight gain between H4-TG and control mice became greater with longer HFD feeding (Fig. 6A). The reduced body weight under HFD was also observed in the H4-TG#C mice (Supplementary Fig. 8). The reduced body weight in H4-TG#D mice under HFD was only contributed by reduced fat mass, while the lean mass was not affected (Fig. 6B and C). Food intake was slightly upregulated at day in H4-TG#D mice (Fig. 6D). Histological analysis showed the reduced adipocyte size in the eWAT of H4-TG#D mice (Fig. 6E and F). The reduced adiposity under HFD (and NC) was not due to defects in adipogenesis, because mRNA levels of adipogenesis markers were not reduced in the eWAT of H4-TG#D mice (Supplementary Fig. 10A and B). Notably, there were less crown-like structures in the eWAT of H4-TG#D mice after HFD (Fig. 6E). Additional q-PCR analysis of macrophage markers confirmed that adipose inflammation was reduced in H4-TG#D mice (Fig. 6G). Furthermore, H4-TG#D mice were insulin sensitive; they showed enhanced glucose clearance in ITT (Fig. 6H and I). Collectively, H4-TG#D mice displayed metabolic improvements, such as reduced adiposity and enhanced insulin sensitivity, under normal chow and HFD.
Beige adipocyte ablation in H4-TG mice reverses protection against HFD-induced obesity
In order to determine whether thermogenic activity of beige adipocytes in H4-TG contributed to the metabolic phenotypes, we generated H4-TG mice in Ucp1-knockout background (referred as H4-TG;Ucp1 KO mice). There was an expansion of multilocular beige-like adipocytes in iWAT of Ucp1 KO mice, as previously reported (Enerback et al. 1997, Liu et al. 2003). These multilocular but UCP1− adipocytes remained in iWAT of H4-TG;Ucp1 KO mice (Supplementary Fig. 11B). The q-PCR analysis showed thermogenic gene program in iWAT from H4-TG;Ucp1 KO mice was further induced compared to that of H4-TG or Ucp1 KO mice alone (Supplementary Fig. 11A). Under HFD, H4-TG;Ucp1 KO mice exhibited further reduced body weight, although there was no significant body weight difference between WT and Ucp1 KO mice during 4-week HFD (Supplementary Fig. 11C). These observations suggest that enhanced UCP1-dependent thermogenic activity from expanded beige adipocytes in H4-TG mice does not contribute to their metabolic phenotypes under HFD. But the caveat of this approach is that UCP1 deficiency alone can induce compensatory thermogenic mechanisms that ultimately affect systemic insulin sensitivity (Kazak et al. 2015, Ikeda et al. 2017).
Then, we used the beige adipocyte ablation approach described earlier to investigate whether beige adipocytes can regulate metabolic behavior (Fig. 3A). We generated four groups of mice: Cre− (genotype: Rosa-iDTR), Cre+ (genotype: Ucp1-Cre;Rosa-iDTR), H4-TG#D;Cre− (genotype: H4-TG#D;Rosa-iDTR) and H4-TG#D;Cre+ (genotype: H4-TG#D;Ucp1-Cre;Rosa-iDTR) (Fig. 7A). The postnatal beige adipocytes were ablated in the Cre+ and H4-TG#D;Cre+ after DT injection at 3 weeks of age, and they remained absent for 2 months (Fig. 7B, C and D). While DT injection did not affect beige adipocyte abundance in the H4-TG#D;Cre− mice (Fig. 7D). Importantly, beige adipocyte ablation in H4-TG#D;Cre+ mice did not affect Ucp1 mRNA levels, multilocular morphology in iBAT and cold resistance in 4°C cold test (Supplementary Fig. 12A, B and C). We then determined metabolic behaviors of beige adipocyte-ablated H4-TG#D mice under HFD (Fig. 7B). Beige adipocyte-ablated H4-TG#D (H4-TG#D;Cre+) mice gained more weight compared with H4-TG#D mice without beige ablation (H4-TG#D;Cre−) during initial weeks of HFD (Fig. 7E and F). This effect was not due to alteration of adipocyte numbers in adipose tissues by DT-mediated cell ablation, as beige ablation in WT mice had no effect. However, H4-TG#D;Cre+ mice remained improved insulin sensitivity compared with H4-TG#D;Cre− mice (Fig. 7G and H), suggesting that HDAC4 activation in white adipocytes in H4-TG mice might regulate insulin sensitivity independently of beige adipocyte expansion. Thus, changes of ITT are not simply a consequence of reduced adiposity in the H4-TG#D mice. These observations also suggest that beige adipocytes in H4-TG#D mice specifically regulate adiposity under an obesogenic diet presumably through an UCP1-dependent mechanism.
Numerous studies have observed that beige adipocyte expansion correlated with improved metabolic performance in rodents and humans; however, the casualty between them has not been established. This is partially due to a conundrum facing the adipocyte research community that no perfect Cre lines are available to target white, beige or brown adipocyte specifically. Previously, we have used both Adiponectin-Cre and Ucp1-Cre to generate pan adipocyte-specific and brown/beige adipocyte-specific Lkb1-knockout mice (Lkb1AKO and Lkb1BKO mice respectively) and identified that LKB1-SIKs-HDAC4 pathway in white adipocytes regulated beige adipocyte maintenance in iWAT in a non-cell autonomous fashion, a process called beige adipocyte renaissance (Wang et al. 2017). In fact, LKB1 deficiency in beige adipocytes themselves did not affect beige adipocyte dynamics in iWAT. Hence, the beige adipocyte expansion phenotype in many Adiponectin promoter-driven overexpression or knockout mouse models should be interpreted with caution. Combinational studies using Adiponectin-Cre and Ucp1-Cre are needed to determine whether the pathways of interest regulate beige adipocyte formation cell autonomously or non-cell autonomously.
It has been demonstrated that beige adipocytes may have additional roles in energy metabolism rather than UCP1-dependent and -independent adaptive thermogenesis (Kajimura et al. 2015, Kazak et al. 2015, Ikeda et al. 2017). However, experimental evidence definitely showing specific functions of beige adipocytes is still lacking. One possible route to address beige adipocyte-specific function is to identify genetic programs that only affect beige adipocytes. We showed that HDAC4 activation by LKB1 deficiency was specifically required for beige adipocyte renaissance in iWAT, and it did not regulate adaptive thermogenesis in brown adipocytes (Wang et al. 2017, Paulo et al. 2018). On the other hand, HDAC4 loss of function did not affect adaptive thermogenesis in brown adipocytes in either WT mice or in adipocyte LKB1 deficiency background (Wang et al. 2017). Consistent with our previous report, constitutively active HDAC4 overexpression in adipocytes drives the persistent beige adipocyte renaissance in iWAT, similar to HDAC4 overactivation caused by LKB1 deficiency in adipocytes. This specific requirement of HDAC4 in beige adipocyte expansion enabled us to explore beige adipocyte-specific function in energy homeostasis using the H4-TG mice. In combination with a novel beige adipocyte ablation system (Wang et al. 2017), we showed that specific ablation of UCP1+ beige adipocytes did reverse protection against HFD-induced obesity in H4-TG mice, establishing a causal link between beige adipocyte abundance and adiposity. Furthermore, this novel beige adipocyte ablation approach will be also applicable to address numerous aspects of beige adipocyte biology.
There are some caveats in the current study. The promiscuous AP2 promoter used in H4-TG mice could overexpress constitutively active HDAC4 in non-adipocyte cells including adipogenic progenitors, neurons, macrophages and endothelial cells (Lee et al. 2013, Mullican et al. 2013, Shan et al. 2013, Jeffery et al. 2014). These potential off-target effects of constitutively active HDAC4 might indirectly contribute to beige adipocyte expansion in iWAT and improved metabolic performance, such as improved insulin sensitivity and elevated physical activity. Additional studies using Adiponectin promoter-driven conditional (spatial and/or temporal) overexpression of constitutively active HDAC4 will address this concern.
Although expanded beige adipocytes in both Lkb1AKO and H4-TG mice were present during short-term HFD (~4 weeks), they did lose their beigeing properties after long-term HFD (~10 weeks) (Supplementary Fig. 9A and B) (Wang et al. 2017). Therefore, we only performed 4–5 week HFD experiments in H4-TG mice after beige adipocyte ablation, to establish the causality relationship between beige adipocyte abundance and reduced adiposity. Whether the presence of beige adipocytes during early-phase HFD contributes to metabolic improvements in H4-TG mice after long-term HFD remains unknown. Regardless, our study provides direct evidence that beige adipocytes could regulate adiposity. Future studies are warranted to use these animal models to investigate whether thermogenic or non-thermogenic function of beige adipocytes is required for the anti-obesity effect in H4-TG mice.
This is linked to the online version of the paper at
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This work is supported by National Institutes of Health (NIH) grants DK105175, UCSF Diabetes Research Center P30DK063720, UCSF Nutrition Obesity Research Center P30DK098722, and research grants from Hillblom Foundation and Juvenile Diabetes Research Foundation (B W). E P is supported by a fellowship grant from Hillblom Foundation.
Author contribution statement
B W, E P and D W planned the experiments and wrote the manuscript. D W characterized H4-TG mice phenotype and E P performed beige adipocyte ablation studies. P H analyzed microarray data. Y Z assisted with mouse genotyping.
The authors thank X-J Yang (McGill University) for providing the GFP-HDAC4.3A plasmid. They thank C Paillart for assistance with CLAMS experiments and L Ouyang (The Salk Institute for Biological Studies) for microarray experiment.
BarbatelliGMuranoIMadsenLHaoQJimenezMKristiansenKGiacobinoJPDe MatteisRCintiS 2010 The emergence of cold-induced brown adipocytes in mouse white fat depots is determined predominantly by white to brown adipocyte transdifferentiation. American Journal of Physiology: Endocrinology and Metabolism 298 E1244–E1253. (https://doi.org/10.1152/ajpendo.00600.2009)
CypessAMWhiteAPVernochetCSchulzTJXueRSassCAHuangTLRoberts-TolerCWeinerLSSzeC, 2013 Anatomical localization, gene expression profiling and functional characterization of adult human neck brown fat. Nature Medicine 19 635–639. (https://doi.org/10.1038/nm.3112)
FischerKRuizHHJhunKFinanBOberlinDJvan der HeideVKalinovichAVPetrovicNWolfYClemmensenC, 2017 Alternatively activated macrophages do not synthesize catecholamines or contribute to adipose tissue adaptive thermogenesis. Nature Medicine 23 623–630. (https://doi.org/10.1038/nm.4316)
IkedaKKangQYoneshiroTCamporezJPMakiHHommaMShinodaKChenYLuXMaretichP, 2017 UCP1-independent signaling involving SERCA2b-mediated calcium cycling regulates beige fat thermogenesis and systemic glucose homeostasis. Nature Medicine 23 1454–1465. (https://doi.org/10.1038/nm.4429)
JespersenNZLarsenTJPeijsLDaugaardSHomoePLoftAde JongJMathurNCannonBNedergaardJ, 2013 A classical brown adipose tissue mRNA signature partly overlaps with brite in the supraclavicular region of adult humans. Cell Metabolism 17 798–805. (https://doi.org/10.1016/j.cmet.2013.04.011)
KazakLChouchaniETJedrychowskiMPEricksonBKShinodaKCohenPVetrivelanRLuGZLaznik-BogoslavskiDHasenfussSC, 2015 A creatine-driven substrate cycle enhances energy expenditure and thermogenesis in beige fat. Cell 163 643–655. (https://doi.org/10.1016/j.cell.2015.09.035)
PauloEWuDWangYZhangYWuYSwaneyDLSoucherayMJimenez-MoralesDChawlaAKroganNJ, 2018 Sympathetic inputs regulate adaptive thermogenesis in brown adipose tissue through cAMP-Salt inducible kinase axis. Scientific Reports 8 11001. (https://doi.org/10.1038/s41598-018-29333-6)
VishvanathLMacPhersonKAHeplerCWangQAShaoMSpurginSBWangMYKusminskiCMMorleyTSGuptaRK 2015 Pdgfrbeta mural preadipocytes contribute to adipocyte hyperplasia induced by high-fat-diet feeding and prolonged cold exposure in adult mice. Cell Metabolism 23 350–359. (https://doi.org/10.1016/j.cmet.2015.10.018)
WaldenTBHansenIRTimmonsJACannonBNedergaardJ 2012 Recruited vs. nonrecruited molecular signatures of brown, ‘brite,’ and white adipose tissues. American Journal of Physiology: Endocrinology and Metabolism 302 E19–E31. (https://doi.org/10.1152/ajpendo.00249.2011)