Novel Function of Rev-erbα in Promoting Brown Adipogenesis

To study the function ofin brown fat, we first examined its abundance in brown and white adipose tissue. Rev-erbα protein is highly enriched in BAT as compared to its low expression in WAT, at a level that is comparable to the liver (). In contrast, thetarget gene in the molecular clock circuit, Bmal1, display an opposite distribution pattern in adipose tissues with higher level in WAT than BAT. To investigate thefunction ofin brown fat development, we generated-null () mice using gene-targeted ES cell clone obtained from the KOMP consortium, with the null allele identified as a 380 bp PCR amplification (). The homozygote null mice are viable, but adult mice show reduced fertility, as reported previously. The brown fat ofmice display complete ablation of(), while mRNA level of its direct target gene,, is markedly induced as compared to wild-type controls (WT), as expected. Next, we examined the effects of loss offunction on BAT development in embryonic and neonatal stages in WT, heterozygote and homozygote mutants. As BAT forms late during development with fully mature features detected after E16.5, we first analyzed BAT formation at E18.5 embryos by haematoxylin and eosin staining. BAT in WT mice at this stage of development display fully organized, densely stained structures with three separate lobes in between the dermis and underlying muscle layer, as shown in. In contrast, the size of interscapular BAT inmice is markedly reduced as compared to that of the WT littermates (), with large part of the structure replaced by white adipose tissue. In addition, the structural organization of these brown fat pads with the underlying muscular layer is severely disrupted, with skeletal muscle interspersed within the brown fat. BAT ofembryos are also smaller, but their juxtaposition with adjacent muscle is largely maintained. Therefore, the loss ofmarkedly impairs the formation, structural integrity and characteristics of BAT. Rev-erbα Rev-erbα in vivo Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα Bmal1 Rev-erbα Rev-erbα Rev-erbα Rev-erbα in vivo Fig. 1A Fig. 1B ²³ Fig. 1C ²⁴ Fig. 1D Fig. 1D −/− −/− −/− +/−

We further examined BAT formation in early postnatal development of day 5 (P5), which allows us access to tissues and quantitative analysis. Similar substantially reduced size and disrupted structure of BAT is evident in themice through examination of one side of the interscapular fat pad (), withaffected to a lesser extent. Reduction of theBAT size is accompanied by diminished mitochondrial-rich cytoplasmic staining characteristic of brown fat and its replacement by increased lipid droplets as compared to controls (). Analysis of total protein in BAT reveals nearly absence of UCP-1 expression and marked reduction of C/EBPβ in these mice, corroborating the histological observation of loss of brown fat identity (). Likely due to impaired brown fat development, the BAT mass as determined by its tissue weight (), or tissue to body weight ratio () is ~16% and ~31% lower in neonatalandmice than that of the littermate controls, respectively. Their body weight at this developmental stage is not altered (). Similar to what we found in embryonic development and neonatal mice, brown fat weight in 10 weeks old adultandmice remains to be significantly lower (). Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα −/− +/− −/− +/− −/− +/− −/− Fig. 2A Fig. 2B Fig. 2C Fig. 2D Fig. 2E Fig. 2F Fig. S1A

Interestingly, in addition to impaired BAT formation, adultdisplay ~40% reduction of WAT mass relative to the WT, as indicated by measured epididymal fat pad weight () or NMR analysis of total body fat (&). In addition, NMR analysis reveals a significant reduction of total lean mass in Revmice, but not in Revmice (). The body weight of adultmice are also significantly lower than that of the WT controls (), whereas their heart weights are not affected (). As lean mass accounts for over 80% of total body weight, the reduced lean weight likely contributed to the lower body weight observed in Revmice together with lower fat mass. The reduced amount of white adipose tissue observed in adult Revmice is consistent with previously reported Rev-erba positive regulation of adipogenesis, and shRNA silencing ofin 3T3-L1 cells further demonstrates this function (). Interestingly, despite the reduction of brown adipose tissue in adultmice, the H/E histology is largely comparable to that of the WT (), and cold tolerance test reveals similar responses (). Furthermore, gene expression analysis indicate induction of UCP-1 mRNA inmice, although brown adipogenic markers (), and mitochondrial genes () are down-regulated (). These intriguing findings suggest that despite the marked impairment of brown adipose tissue development during embryonic and neonatal stages, adultmice do generate functional brown fat. Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα Myf5, Pparg, Fabp4 Pgc1a and Cox7a1 Rev-erbα −/− −/− +/− −/− −/− −/− −/− −/− −/− Fig. S1B Fig. S1C S1D Fig. S1E Fig. S1F Fig. S1G ^{20 25} Fig. S2 Fig. S3A Fig. S3B Fig. S3C

Based on the finding thatpromotes BAT development, we next explored whether it directly participates in brown adipogenesis in a cell-autonomous manner using genetic loss- and gain-of-function studies in brown adipocyte progenitors. Brown adipocytes arise from mesenchymal precursors that first commit to a brown progenitor fate and subsequently undergo terminal differentiation. Initial analysis of Rev-erbα protein level in mesenchymal precursors and brown preadipocytes, as compared to white preadipocytes and myoblasts, indicate that it is more abundant in C3H10T1/2 (10T1/2) mesenchymal stem cells and HIB1B brown preadipocytes (), two cell types that are capable of differentiating into mature brown adipocytes under specific induction conditions. Using stable transfection of a retroviral short hairpin RNA (shRNA) construct targeting(), we generated stable knockdown clones of shRev-expressing 10T1/2 mesenchymal precursors, which results in near absence of Rev-erbα protein as compared to cells expressing scrambled control shRNA () (). When subjected to a brown adipocyte induction cocktail, scrambled control cells display robust progression to mature adipocyte, with ~48% exhibiting typical round morphology and muti-locular lipid accumulation as indicated by phase-contrast images (). In comparison,silencing markedly reduces the percentage of cells with mature adipocyte morphology to ~22%. BODIPY staining of neutral lipids accumulation as an index of mature differentiation confirms the reduction of lipid droplets incells. Moreover, MitoTracker Red staining of functional mitochondria in live cells is substantially weaker in-deficient cells than that of the. The observed lower mitochondrial abundance suggests that loss ofimpairs mitochondrial biogenesis associated with mature brown adipocyte differentiation. Furthermore, gene expression analysis at beginning, day 4 and 9 of differentiation demonstrates a markedly suppressed brown adipogenic gene program (). The induction of brown-specific differentiation markers,and, are significantly attenuated incells as compared to their robust expression (~7-8 fold induction) in the controls. The lack of induction of adipogenic genes,andduring the differentiation time course indicates suppressed adipogenesis (). We also found a moderate ~2-fold up-regulation ofmRNA level at day 9 of brown adipogenic induction, with significantly attenuated expression in thecells. Collectively, these effects ofdeficiency in 10T1/2 cells indicate that it modulates the lineage commitment and differentiation of mesenchymal precursor cells into brown adipocytes. Rev-erbα Rev-erbα shRev shSC Rev-erbα shRev Rev-erbα shSC Rev-erbα Ucp-1 Prdm16 shRev Cebpβ, Cebpα, Pparγ FABP4 Rev-erbα shRev Rev-erbα ²⁶ Fig. 3A Fig. 3B Fig. 3C Fig. 3D Fig. 3D

Next, we determined the role of Rev-erbα in terminal differentiation stage of brown adipogenesis through gain- and loss-of-function studies in HIB1B brown preadipocytes. Similar to the phenotype observed with stableshRNA depletion in mesenchymal precursors, inhibition of Rev-erbα in committed brown preadipocytes markedly suppresses lipid accumulation and mitochondrial abundance at day 6 of differentiation, as indicated by reduced BODIPY and MitoTracker staining relative to scrambled control cells (). This significantly attenuated differentiation is accompanied by marked suppression of brown specific marker gene expression (,and, and adipogenic factors (,and(). Stable Rev-erbα depletion in these cells achieved ~70% reduction oftranscript. Rev-erbα Ucp-1 Dio2 Prdm16) Cebpβ Cebpα Pparγ) Rev-erbα Fig. 4A Fig. 4B

Using stable expression of acDNA plasmid (Rev-cDNA) in HIB1B cells, we tested the effect ofoverexpression on brown adipocyte terminal differentiation. Forced expression of-cDNA markedly elevatedprotein level over vector control, pcDNA3 (). Consistent with its protein expression,transcript is increased ~6-fold in-cDNA-expressing cells as indicated by RT-qPCR analysis (). BODIPY and MitoTracker staining to assess mature differentiation demonstrate increased lipid accumulation and mitochondrial abundance in-cDNA-expressing cells at day 6 of differentiation than controls, indicating that ectopic expression ofenhances terminal differentiation (). Of note, at the same differentiation stage, the pcDNA3 vector control display less lipid accumulation and mitochondrial staining than scrambled shRNA control shown in, suggesting a slight inhibitory effect of pcDNA3 vector on brown differentiation. Gene expression analysis of-overexpressing HIB1B cells further demonstrates marked induction of brown adipocyte-specific markers over vector controls, together with moderate up-regulation of genes involved in adipogenesis (). Rev-erbα Rev-erbα Rev Rev-erbα Rev-erbα Rev Rev Rev-erbα Rev-erbα Fig. 5A Fig. 5C Fig. 5B Fig. 5B Fig. 5C

To assess the impact of complete loss offunction on brown preadipocyte differentiation, we isolated primary brown preadipocytes from WT andmice and generated immortalized cell clones using SV40 large T antigen retroviral transduction. As shown inimmortalized WT primary brown preadipocytes exhibit rounded mature brown adipocyte morphology with robust lipid accumulation and mitochondrial formation upon brown adipogenic induction. In sharp contrast, differentiation ofbrown preadipocytes is severely blocked, as they fail to progress to mature adipocytes and display barely detectable amount of lipid and mitochondria staining. Nearly abolished brown adipogenic marker gene induction incells relative to WT controls further confirms the severe defect in differentiation (). Taken together, these findings indicate that Rev-erbα positively regulates terminal differentiation in lineage-committed brown preadipocytes. Rev-erbα Rev-erbα Rev-erbα Rev-erbα −/− −/− −/− Fig. 6A,B Fig. 6C

As a ligand-dependent nuclear receptor, the transcriptional activity of Rev-erbα is amenable to regulation by small molecule ligands, and both its agonist, SR9011, and antagonist, SR8278, display metabolic effects. Therefore, we tested the effect of pharmacological activation and blockade of Rev-erbα activity on brown adipocyte differentiation using SR9011 and SR8278, respectively. In line with results from genetic interventions,activation by SR9011 enhances, whereas its inhibition by SR8278 impairs brown preadipocyte terminal differentiation in comparison to DMSO vehicle control, as indicated by lipid and mitochondrial staining shown in. Repression and induction ofmRNA in response to SR9011 and SR8278, respectively, confirms the modulation ofactivity by these ligands in brown adipocytes (). SR8278 suppression of brown adipocyte marker genes expression further validates the findings on differentiation as assessed by morphological progression (). Interestingly, although SR9011 significantly up-regulates brown-specific markers, its effect on adipogenic factors are not evident. Together, these results demonstrate that modulation of Rev-erbα activity by pharmacological means is sufficient to impact brown adipogenesis. ^{27 28 29} Fig. 7A,B Fig. 7B Fig. 7B Rev-erbα Bmal1 Rev-erbα

TGF-β cascade is a key developmental signal that suppresses brown adipogenesis. Interestingly, molecular clock components in epidermal stem cells regulate various signaling cascades, including the TGF-β pathway, and a recently published ChIP-Seq dataset revealedbinding peaks on certain TGF-β pathway gene regulatory regions. We thus tested whethercould modulate TGF-β signaling to impact brown adipogenesis. Using the TGF-β-responsive luciferase reporter, SBE4-Luc, to assess TGF-β pathway activity, we found significantly increased reporter activity in-deficient 10T1/2 cells as compared to SC control upon TGF-β stimulation, indicating that loss offunction promotes TGF-β signaling transduction (). Smad3 phosphorylation in response to TGF-β activation ultimately transduces TGF-β signaling to the nucleus to modulate gene transcription. Therefore, we examined Smad3 phosphorylation in WT andbrown preadipocytes. Robust Smad3 phosphorylation is detected in WT cells only in response to TGF-β1, as expected. In contrast,cells already display a moderate level of Smad3 phosphorylation at basal condition that is enhanced by TGF-β1 treatment (). Non-responsiveness of Smad3 to BMP4 stimulation validates the specificity of TGF-β1 signaling. Interestingly, total Smad3 protein level is increased inpreadipocytes, indicating thatimpacts Smad3 expression. To address potential direct transcriptional regulation ofon genes of the TGF-β signaling pathway, we screened for RORE elements, in −2 kb and first intron of their regulatory regions to identify putativebinding sites. Among the genes screened, including the ligandsand, the receptorsand, and signal transducer,response elements were identified in promoter regions ofand, as listed in. Using specific primers flanking these identified sites (), chromatin immunoprecipitation qPCR (ChIP-qPCR) analysis detectsenrichment onandpromoters similar to, or higher than that of the knowntarget,(). In line with this finding, absence ofin brown preadipocytes leads to a ~15-fold induction ofmRNA, and a tendency for higherexpression (). These results demonstrate thatis a bona-fide transcription repressor of TGF-β pathway genes and loss of its repression leads to enhanced TGF-β signaling in brown adipocyte progenitors. ^{30 31 32 33 34 28} Fig. 8A ³⁵ Fig. 8B Fig S3 Fig. S4 Fig. 8C Fig. 8D Rev-erbα Rev-erbα Rev-erbα Rev-erbα Rev-erbα-null Rev-erbα-null Rev-erbα-null Rev-erbα Rev-erbα Rev-erbα Tgfb1 Tgfb2 Tgfbr1 Tgfbr2 Smad3 Rev-erbα Tgfbr2 Smad3 Rev-erbα Tgfbr2 Smad3 Rev-erbα Bmal1 Rev-erbα Smad3 Tgfbr2 Rev-erbα

Collectively, findings from our study suggest a model in whichtranscriptional control of the TGF-β pathway suppresses TGF-β activity, and attenuation of this inhibitory signal of brown fat development ultimately promotes brown adipocyte lineage determination and terminal differentiation (). Rev-erbα Fig. 8E

Mice were maintained in the Houston Methodist Hospital Research Institute mice facility at ambient temperature under a constant 12:12 light dark cycle, with light on at 7:00 AM (ZT0). All experimental protocols were approved by the IACUC animal care research committee of the Houston Methodist Research Institute, and carried out in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. Mice with global ablation ofwere generated using gene targeted embryonic stem cells, Velocigene ES clone 11705A-E7, C57BL/6NTac origin, obtained from KOMP repository. Briefly, ES cell clones were injected into albino C57BL/6J-mouse blastocysts. The chimeric mice obtained were mated with C57BL/6J-mice for germ line transmission and the targetedallele was confirmed by polymerase chain reaction (PCR) analysis using tail DNA. Offspring were backcrossed to C57BL/6J, and heterozygotes were bred to obtain homozygote null and littermate wild-type controls used in the study. Thehomozygotes were born at a normal Mendelian ratio without obvious developmental defects. Tissue samples were obtained at ambient temperature at the same time of the day to ensure comparable circadian gene expression. Rev-erbα Tyr Tyr Rev-erbα Rev-erbα

Mesenchymal stem cell line, C3H10T1/2, and brown preadipocyte cell line, HIB1B, were obtained from ATCC and maintained at subconfluence between passages in DMEM with 10% FBS and antibiotics. C3H10T1/2 differentiation to brown adipocytes was conducted as describedwithout BMP7 pretreatment. Briefly, cells were grown to confluency and subjected to brown induction media of DMEM containing 10% FBS, insulin (20 nM), T3 (1 nM), isobutylmethylxanthine (0.5 mM), dexamethasone (5 mM), and Rosiglitazone (1 μM) for 3 days. Cells were then switched to maintenance media of DMEM containing 10% FBS supplemented with only insulin and T3 for 9 additional days. Fully differentiated brown phenotype with significant lipid accumulation andexpression occur at 9 days (D9) after induction. Differentiation of HIB-1B and primary brown adipocytes were induced using similar cocktails with induction media for two days and maintenance media for four days. ⁴⁷ Ucp-1

shRNA construct in pGIPZ vector (V2LMM_25660) and scrambled control vector (RHS4346) were purchased from Open Biosystems, and mouse full-lengthcDNA clone (in pcDNA3 vector) was obtained from OriGene. C3H10T1/2 or HIB1B cell lines expressingshRNA or cDNA were transiently transfected using FuGENE 6 reagent, and selection of stably transfected clones by Puromycin (for shRNA) or Neomycin (for pcDNA3) were started 48 hours following transfection and maintained for 7–10 days. The entire transfected plate was subcultured in order to minimize the effect of confluency on efficiency of differentiation, as described previously. Rev-erbα Rev-erbα Rev-erbα MC203687

Primary brown adipocytes and preadipocytes were isolated from interscapular brown adipose tissue pad of 4-week-old mice, as described. Briefly, tissues were digested by Type I collagenase digestion in the presence of 1% BSA at 37 °C for 30 minutes. The suspension was filtered, centrifuged, and the top fat layer was collected as adipocytes and pellet containing the stromal vascular fraction was resuspended and plated. Preadipocytes in the stromal vascular fraction were passaged once prior to adipogenic differentiation. Immortalization of isolated primary preadipocytes was performed using by retroviral SV-40 Large T antigen transformation and puromycin selection, as described. ^{48 47}

BODIPY 493/503 (Life Technologies) staining of neutral lipids was carried out at a concentration of 1 mg/ml after formaldehyde fixation and incubation for 30 minutes. MitoTracker Red (MitoTracker Deep Red FM, 100 nM, Life Technologies) staining was applied to live cells at 37°C for 30 minutes before fixation to stain functional mitochondria, according to manufacturer’s protocol.

RNeasy miniprep kits (Qiagen) were used to isolate total RNA from snap-frozen tissues or cells. Tissues samples are collected at times as indicated and cell samples were obtained at non-synchronized normal culture conditions. cDNA was generated using q-Script cDNA Supermix kit (Quanta Biosciences) and quantitative PCR was performed using a Roche 480 Light Cycler with Perfecta SYBR Green Supermix (Quanta Biosciences), as described. Relative expression levels were determined using the comparative Ct method to normalize target genes to 36B4 internal control or compared to controls as indicated. ⁴⁹

40–50 μg of total protein from tissues or cell homogenates were used for each sample on SDS-PAGE gel. After electrophoresis, protein were transferred to PVDF membrane, blotted using specific primary and secondary antibody and detected by chemiluminescence (Supersignal; Pierce Biotechnology), as previously described. Smad3 phosphorylation in immortalized primary brown preadipocytes was assessed at 1 hour after indicated ligand treatment. The primary antibodies used were: Rev-erbα, PA5-29865 (Thermo Scientific), Bmal1, AB93806 (Abcam), UCP-1, AB3038 (Millipore); CEBPβ, sc-150, TBP, sc-204 (Santa Cruz); Smad3, 04-1035, phosphor-Smad3 (Ser 423/425), 07-1389 (Millipore). ⁴⁹

Immunoprecipitation was performed using Bmal1 Rev-erbα antibody (PA5-29865) or control rabbit IgG plus protein A/G beads, as described. Briefly, cells were fixed by formaldehyde, lysed and sonicated to shear the chromatin. The immunoprecipitated chromatin fragments were eluted, treated with proteinase K and purified using Qiaquick PCR purification kit (Qiagen). Real-time PCR using Perfecta SYBR Green Supermix (Quanta Biosciences) was carried out with an equal volume (4 ul) of each reaction with specific primers. Sequences of the specific primers flanking the identified RORE sites are listed in. TBP were included as negative controls, and the known Rev-erbα target Bmal1 as a positive control. Data was expressed as fold enrichment over IgG after normalization to 1% input. ⁵⁰ supplementary Fig. S3

C3H10T1/2 cells were seeded to ~80% confluency overnight in 24 well plates. Transient transfection using FuGENE 6 (Roche) was carried out in four replicates, as described. Per well, the transfection mixture contains 150 ng of TGF-β-responsive SBE4-Luc luciferase reporter (Addgene Plasmid 16495) together with 20 ng of Renilla luciferase (pRL-TK, Promega) as an internal control. TGF-β1 (2 ng/ml) were added 16 hours after transfection and luciferase activity was measured using Dual-Glo luciferase assay system (Promega) 24 hours following ligand treatment. Reporter luciferase values were normalized to Renilla readings and expressed as fold of induction over controls. ^{39 51}

Data is expressed as Mean ± SE. Statistical differences between mean values of two groups were assessed by two-tailed, unpaired Student’s t test Student’stest. P ≤ 0.05 is considered as statistically significant. t