Distinct roles for REV-ERBα and REV-ERBβ in oxidative capacity and mitochondrial biogenesis in skeletal muscle

The following mouse strains used were either purchased from the Jackson Laboratory and/or were bred at the Scripps Research Institute–Florida (Scripps Florida). Male B6.Cg-Nr1d1^tm1Ven/LazJ, stock # 018447 (REV-ERBα^-/-)[21]; Male and female REV-ERBβ KO mice were generated as previously described[22]. Wild-type (WT) littermates were used as controls for both knock out strains. This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Animal care and experimental protocols used in this study were approved by the Scripps Florida Institutional Animal Care and Use Committee (Assurance number: D16-00726). Mice were housed in groups of 3–5 in 12 h light/12 h dark cycles at 23°C and fed standard chow (Harlan 2920X), unless otherwise stated. Mice had access to chow and water ad libitum. Mice were sacrificed by CO₂ asphyxiation for tissue processing.

Food intake was monitored in normal chow-fed mice using the BioDAQ system (Research Diets) following vendor recommended procedures. The feeding patterns of pre-acclimatized, individually housed mice were continuously recorded for 3 consecutive days. The daily number of bouts, percent of time spent in bouts and total chow consumed were obtained. A bout constitutes a visit or set of visits (no more than 5 seconds apart) to the food hoppers. Each bout has a duration and amount of food obtained during it. Data represent the average day or night intake. Average values are compared by 2-tailed Student’s t test to determine significance.

Mice were individually placed and acclimatized in a Comprehensive Laboratory Monitoring System (CLAMS; Columbus Instruments) at 23°C for 48 h. Afterwards, VO2, VCO2, food intake, and spontaneous locomotor activity were measured for the indicated time periods. Respiratory exchange ratio (RER) and energy expenditure (EE) were calculated using the following equations: RER = VCO2/VO2; EE (kcal/h) = (3.815 + 1.232 X RER) x VO₂. EE was normalized by lean body mass. Both raw and average values for RER are presented.

Total, fat, lean, and fluid mass of mice was measured by Nuclear Magnetic Resonance using the Minispec LF-NMR (Brucker Optics) analyzer.

To generate murine REV-ERBα or REV-ERBβ retroviral vectors, mouse REV-ERB sequences were inserted into the MIGR1 vector (Addgene) using the XhoI site and further screened for orientation. MIGR1 empty vector was used as a control. MIGR1 was a gift from Warren Pear (Addgene, plasmid # 27490)[23]. To knock down REV-ERBβ expression, 3 different shRNAmirs (TransOmic Technologies, Inc.) were inserted, individually into the pLMPd-Ametrine vector[24]. The shRNAmirs used were top-scoring designs from shERWOOD analysis (TransOmic Technologies, Inc.). Each shRNAmir insert was PCR amplified using primers specific for common regions in the flanking miR-30 sequences, which included XhoI and EcoRI sites. pLMPd-Ametrine containing a CD8 shRNA was used as a control.

Plat-E cells (Cell Biolabs, Inc.) were cultured in DMEM containing 10% fetal bovine serum, 2mM L-glutamine, and 1% penicillin/streptomycin at 37°C under standard culture conditions. Plat-E cells were seeded at 350,000 cells per mL in a 10cm dish the day before transfection. 5μg total plasmid DNA was transfected via Fugene6 reagent (Promega) according to manufacturer’s protocol. Viral supernatant was harvested 48 and 72 hours post transfection and stored at -80.

C2C12 myoblasts were obtained from ATCC (Manassas, VA) and grown in Dulbecco modified Eagle’s medium (DMEM; 4.5 g/l D-Glucose; Gibco) supplemented with 10% fetal bovine serum, 2mM L-glutamine, and 1% penicillin (50units/mL)/streptomycin (50ug/mL). Prior to myogenic differentiation, C2C12s were plated at 30,000 cells/mL in 24 well plates and left to adhere overnight. The following day, cells were washed once with PBS and incubated overnight in 1mL total of 50% retrovirus conditioned media and 50% normal culture media containing Polybrene (8ug/mL, Santa Cruz Biotech). 24 hours later, retrovirus was removed and replaced with normal growth media for another 24 hours before myogenic differentiation was induced. One transduction was sufficient to obtain >90% GFP/Ametrine-positive cells. Myogenic differentiation into myotubes was induced after cells reached confluency by adding DMEM supplemented with 2% horse serum (HS), 2mM L-glutamine, and 1% penicillin/streptomycin at 37°C in a humidified incubator under 5% CO₂ for 6 days. Overexpression and knock down was verified by qPCR analysis.

Retrovirally transduced C2C12 cells were washed with PBS, trypsinized and incubated at 37°C for 20 min with 100nM MitoTracker Red^FM dyes (Molecular Probes). MitoTracker Red probe is a red-fluorescent dye that stains mitochondria in living cells and its accumulation is dependent upon membrane potential (Molecular Probes). To measure viability in the samples, a Fixable Viability Dye (Invitrogen/Thermo Fisher) was added to the cultures during the Mitotracker staining step. This dye can be used to stain both live and fixed cells to irreversibly label dead cells prior to analysis. Samples were washed three times in PBS and flow cytometric analysis was performed on a BD LSRII (BD Biosciences) instrument and analyzed using FlowJo software (TreeStar).

Mitochondrial content was measured using methods previously described[25]. Briefly, genomic DNA was extracted using the Blood & Cell culture DNA Mini Kit (Qiagen) per manufacturer’s instructions from C2C12 samples. DNA was quantified via a standard curve method in order to dilute all samples to 3ng/ul in TE buffer. Once normalized for concentration, samples were subject to RT-qPCR using a cocktail of SYBR green PCR master mix, template DNA, and mtDNA target specific primer pairs, plus H2O. Each sample was run in triplicate and data were averaged. This step was repeated in separate wells using nuclear DNA specific primers and the ratio between mtDNA and nuclear DNA was quantified. Dissociation curves were calculated for all samples to ensure the presence of a single PCR product.

Muscle was harvested from WT and REV-ERBβ KO mice after a 5 hour fast, at ZT6 (Zeitgeber 6–6 hours into the murine nocturnal period). Total RNA was isolated from tissues by guanidinium thiocyanate/phenol/chloroform extraction[15]. RNA concentrations were adjusted in order to load 500ng of RNA per cDNA reaction. Total RNA was isolated from C2C12 cells using a Quick-RNA MicroPrep kit (Zymo Research) using the manufacturers protocol. RNA concentrations were adjusted in order to load 1μg of RNA per cDNA reaction. cDNA was synthesized using qScript^TM cDNA SuperMix synthesis kit (Quanta Biosciences). Quantification of each transcript by quantitative real time-polymerase chain reaction (qRT-PCR) was performed using SYBR Green dye (Quanta Biosciences) to detect dsDNA synthesis, and analyzed using cycling threshold (Ct) values. Relative expression was determined using the ΔΔCT method and normalized to the housekeeping gene 18s. Quantitative RT-PCR was performed with a 7900HT Fast Real Time PCR System (Applied Biosystems) using SYBR Green (Roche) as previously described. Primers were designed using Primer3 (primer3.sourceforge.net↗). Specificity and validation of the primers were determined using an In silico PCR software program (genome.ucsd.edu↗) and melting curve analysis to eliminate the possibility of primer-dimer artifacts and check reaction specificity. Primer sequences can be found in Table 1.

All data are expressed as mean±S.E.M. Statistical analysis was performed using GraphPad Prism6 software. For multiple comparisons, two-way ANOVA was performed. Post-test analysis (ANOVA) was performed correcting for multiple comparisons using the Sidak-Bonferroni method. For all other independent data sets, since the data analyzed were not random, statistical significance was determined using two-tailed unpaired Student’s t-tests with no correction for multiple comparisons[26]. ANCOVA analysis was performed using SPSS Statistics (IBM). (N per group per study is indicated in each figure legend). Significance was assessed as follows: *p<0.05, ** p<0.01, *** p<0.005, **** p<0.001.

We previously demonstrated that REV-ERBα was a critical regulator of muscle mitochondrial biogenesis[15]. To determine the cell autonomous role of REV-ERBβ versus REV-ERBα, we utilized the C2C12 cell line, a well-studied cell culture model commonly used to investigate myogenesis[27–30]. We retrovirally overexpressed REV-ERBα, REV-ERBβ, or empty vector retrovirus in proliferating C2C12 myoblasts two days prior to the induction of differentiation (Fig 1A). MitoTracker Red staining, indicative of functional mitochondria, revealed that overexpression of both REV-ERBs enhanced mitochondrial biogenesis, demonstrated by increased fluorescence in the REV-ERB overexpressing cells relative to empty vector control. Compared to undifferentiated cells (Day 0; D0), the cells had differentiated and acquired a muscle specific phenotype by Day 6 (D6). Very little difference was observed in expression of Myogenin (Myog), Troponin I slow (Tnni1) and Troponin 1 fast (Tnni2), at Day 6 (D6) between groups[17] (Fig 1B). No overt viability issues were observed by day 6. (S1A and S1B Fig) Overexpression of both REV-ERBs also resulted in increased mitochondrial content, indicated by the increased expression of mitochondrial NADH dehydrogenase I (mt-Nd1), Cytochrome c oxidase I (mt-Co1,) and Cytochrome c oxidase II (mt-Co2), compared to empty vector control cells (Fig 1C). These data suggest that the REV-ERBs can drive mitochondrial biogenesis in vitro.

The REV-ERBs have been demonstrated to be core members of the circadian clock and participate in the regulation of a diverse set of metabolic processes, thus linking control of our daily rhythms with maintenance of metabolic homeostasis[8]. Given the changes in mitochondrial content observed in the REV-ERB overexpressing cells, we next wanted to determine how expression of the clock genes correlated with expression of genes involved in mitochondrial function. Overexpression of REV-ERBα (gene name Nr1d1) significantly repressed almost all molecular clock genes, including REV-ERBβ (Nr1d2), Bmal1 (Arntl), Clock, the Cryptochromes (Cry1, Cry2), and the Period genes (Per1, Per2, Per3), which is consistent with its role as a transcriptional repressor (Fig 2A). Overexpression of REV-ERBα also repressed expression of Cd36, a gene which is involved in lipid uptake[31]. In contrast, RORα and PGC1α (Ppargc1a), the master regulator of mitochondrial biogenesis were upregulated with REV-ERBα overexpression. Additionally, genes encoding enzymes of fatty acid β-oxidation, notably carnitine palmitoyltransferase 1b (Cpt1b), long chain acyl-CoA dehydrogenase (Acadl), short chain acyl-CoA dehydrogenase (Acads), Ucp2, and Ucp3, genes involved in skeletal muscle metabolism, were also upregulated with REV-ERBα overexpression, consistent with previously published work (Fig 2A)[15, 32]. REV-ERBα also repressed expression of Sterol regulatory element binding-protein 1 (Srebf1), Stearoyl-CoA desaturase-1 (Scd1), Fatty acid synthase (Fasn), genes involved in lipogenesis, and Il6, a pleiotropic muscle myokine that has been shown to be involved in muscle growth, myogenesis, and regulation of energy metabolism (S1C Fig) [33]. Overexpression of REV-ERBβ had similar effects on circadian clock and mitochondrial metabolism genes, with the exception that it potently repressed REV-ERBα expression (Fig 2B). The negative regulation of REV-ERBα on REV-ERBβ expression and vice-versa is consistent their ability to negatively regulate their own expression through conserved ROREs/RevREs in their promoter regions[34, 35]. Interestingly, and in contrast to REV-ERBα overexpression, REV-ERBβ overexpression led to increased expression of Srebf1 and Fasn, but had no effects on expression of Scd1 or Il6 (S1C Fig). While it appears that REV-ERBβ was not as effective as REV-ERBα in the overexpression studies, this could be due to the differences in overexpression levels between the two receptors, with REV-ERBα being more highly expressed than REV-ERBβ. Regardless, the increased expression of genes involved in mitochondrial metabolism was consistent with the MitoTracker staining, mitochondrial content, and skeletal muscle metabolism genes. These results demonstrate an inverse relationship between the expression of the core circadian clock genes and the REV-ERBs and indicate that overexpression of REV-ERBα and REV-ERBβ can modulate the expression of genes driving fatty acid β-oxidation and lipid metabolism in C2C12 cells. However, under some circumstances, REV-ERBβ can not completely fulfill REV-ERBα’s role.

Work previously performed assessing REV-ERBβ’s function in C2C12s suggested that it plays a significant role in the regulation of skeletal muscle lipid homeostasis. However, this analysis was performed in C2C12 cells using a dominant negative form of REV-ERBβ that lacked the ligand binding domain[17]. Using this approach, Ramakrishnan and colleagues reported that dominant negative REV-ERBβ largely repressed genes involved in skeletal muscle energy expenditure and lipid catabolism[17]. These results were in line with data demonstrating that loss of REV-ERBα in skeletal muscle led to decreased mitochondrial content and function[15]. Given the overlap in function observed in our overexpression studies, we hypothesized that loss of REV-ERBβ would yield similar results as loss of REV-ERBα. Therefore, to fully assess how loss of REV-ERBβ affects myogenesis, we retrovirally overexpressed a REV-ERBβ shRNA or a CD8 control shRNA in proliferating C2C12 myoblasts 2 days prior to the induction of differentiation (Fig 3A). Strikingly, knock down of REV-ERBβ enhanced mitochondrial biogenesis, indicated by the increased staining with MitoTracker Red. Analysis of Myog, Tnni1, and Tnni2 indicated that compared to undifferentiated cells (Day 0, D0), the cells had differentiated and acquired a muscle specific phenotype by Day 6 (D6). No overt viability issues were observed by day 6. (S2A and S2B Fig) Furthermore, loss of REV-ERBβ appeared to significantly increase expression of Tnni2. (Fig 3B) Knock down of REV-ERBβ also increased mitochondrial content, indicated by the increased expression of mt-Nd1, mt-Co1, and mt-Co2, compared to CD8 shRNA control cells. Interestingly, while knock down of REV-ERBβ led to increased expression of REV-ERBα and RORα, it also de-repressed all of the core molecular clock genes. Like overexpression of REV-ERBα, overexpression of RORα has also been shown to regulate the expression of genes involved in skeletal muscle metabolism[36]. Knock down of REV-ERBβ also resulted in increased expression of almost all genes analyzed that were involved in lipid and skeletal muscle metabolism, including Cd36, Cpt1b, Acadl, Ucp2, and Ucp3, genes involved in skeletal muscle, which was consistent with the MitoTracker staining. No change was observed in Acads. However, knock down of REV-ERBβ resulted in decreased expression of PGC1α (Ppargc1a), Srebf1, Fasn, and Il6, but no change was observed in Scd1 (Fig 3D and S2C Fig). While these results were largely in contrast to previously published work using a dominant negative REV-ERBβ construct, these results were at least consistent with the notion that in C2C12 cells, REV-ERBβ plays a role in the regulation of muscle metabolism and energy expenditure. Interestingly, these data are in contrast to the overexpression data and show a correlation between the circadian clock and skeletal muscle metabolic gene expression.

Given the surprising results we observed in C2C12 cells, we next wanted to determine whether loss of REV-ERBβ affected skeletal muscle gene expression in vivo in a similar manner. Using REV-ERBβ-deficient mice[22], we isolated (whole quadriceps) skeletal muscle from wild-type (WT) and REV-ERBβ knock out (KO) mice at Zeitgeber 6 (ZT6), 6 hours post lights-on and the time point in which peak REV-ERB mRNA expression occurs[37]. qRT-PCR analysis indicated that, similar to the C2C12 data, loss of REV-ERBβ de-repressed most of the core clock genes, including REV-ERBα, RORα, Bmal1, and Clock. We observed no effects on Cry2 or Per3 (Fig 4A). Similar results were observed in REV-ERBα KO muscle. (Fig 4B) Loss of REV-ERBβ also led to increased expression of genes involved in lipid and skeletal muscle metabolism (Cd36, Cpt1b, Acadl, Acads) and energy expenditure (Ucp2, Ucp3) (Fig 4A). However, unlike the C2C12 cells, REV-ERBβ KO mice also had increased expression of PGC1α (Ppargc1a). Similar to the C2C12s, REV-ERBβ KO mice also exhibited decreased expression of Srebf1, Fasn, and Il6, with little effect observed on Scd1. (S3A Fig) Moreover, loss of REV-ERBβ appeared to shift the muscle towards a more oxidative phenotype as we observed increased expression of tropomyosin 3 (Tpm3), a marker of oxidative type 1 fiber and myosin heavy chain IIa (Myh2), a myosin heavy chain gene more associated with oxidative fibers than others. The expression of Myh4, a marker of type 2b glycolytic fibers, was unaltered (Fig 4A). The shift towards a more oxidative phenotype in REV-ERBβ KO mice was in stark contrast to what was observed in REV-ERBα KO mice, which was largely a shift towards a less oxidative phenotype demonstrated by decreased expression of PGC1α, Cpt1b, Acadl, Acads, Ucp2, Ucp3, Myh2, and Tpm3 and is consistent with previously published results[15]. (Fig 4B) However, not all changes in gene expression between REV-ERBβ KO and REV-ERBα KO mice were conflicting. REV-ERBα KO mice also demonstrated decreased expression of the lipogenic genes Srebf1 and Scd1 in muscle while loss of REV-ERBα de-repressed Fasn and Il6 (S3B Fig) However, unlike the muscle tissue, gene expression analysis of livers from REV-ERBβ KO and REV-ERBα KO mice demonstrated similar trends in gene expression (G6Pase, Cd36, Clock) or no effects (Pepck) on gene expression in vivo (S3C and S3D Fig). These data are consistent with previously published work indicating that REV-ERBα and REV-ERBβ have largely overlapping roles in liver metabolism, with REV-ERBα acting as the dominant factor[19, 20]. These data suggest that in liver, REV-ERBβ appears to have overlapping functions to REV-ERBα. However, in skeletal muscle, the effects mediated by the loss of REV-ERBβ are in contrast to those of REV-ERBα.

Due to the increased expression of genes involved in mitochondrial biogenesis in the REV-ERBβ KO mice compared to WT controls, we wanted to determine whether this translated to effects on metabolism and any other circadian driven behavior. Real-time oxygen consumption (VO₂), carbon dioxide production (VCO₂), ambulatory activity and food consumption were monitored in twelve week-old male mice fed a normal chow diet using the CLAMS system. Energy expenditure (EE) and RER (Respiratory exchange ratio) were calculated based on the gas exchange as described in the Methods section (Fig 5A–5C). Body composition was obtained just prior to the CLAMS analysis, with no differences observed in lean, fat, and fluid masses or total body weight between groups (S4A Fig). The RER is a surrogate indicator of the caloric source being utilized for energy production by an animal. It varies between 0.7, where lipids are the predominant fuel source, to 1.0, where carbohydrates are the main contributors. The REV-ERBβ KO mice showed a delay in the typical circadian drop in RER that occurs from the dark to light phase transition when mice shift from high activity and feeding to resting and sleep (Fig 5A and 5B). This prolonged high RER cycle indicates an increased relative contribution of carbohydrates as an energy source in the KO animals. This could be due to their increased chow consumption, observed during the light phase while in the CLAMS (S4B Fig). This light-phase specific increase in eating was further confirmed in REV-ERBβ KO mice tested in the BioDAQ system, which uses highly sensitive feeding monitoring hoppers adapted to regular housing cages (Fig 5E, S4D Fig). Additionally, an increase in energy expenditure was detected in association with the observed increased daytime chow consumption in the REV-ERBβ KO mice (Fig 5C). Nighttime energy expenditure in REV-ERBβ KO mice was comparable to WT mice, despite the KO animals being less active during this time (Fig 5D). Whether the increase in energy expenditure in the KO mice during the daytime can be attributed to elevated feeding and digestive processes per se remains to be tested. Ambulatory activity during the light-phase was similar between the groups and thus, cannot explain that. However, upon fasting, the KO’s demonstrated decreased RER relative to WT controls, which may be a result from a faster transition to fat burning (Fig 5F). Regardless, the increased caloric output can, at least in part, account for the absence of weight gain by the KO’s in spite of their higher caloric intake (S4E Fig). These data indicate that REV-ERBβ regulates circadian behaviors and metabolism in vivo.