Cooperative control of oxidative metabolism by PGC‐1α and PPARβ: implications for exercise‐induced mitochondrial remodelling in skeletal muscle

Mitochondria undergo robust morphological and functional changes in response to a number of environmental conditions. For instance, physical exercise, myogenesis and cold exposure are known to induce mitochondrial biogenesis in skeletal muscle. The transcriptional co-activator peroxisome proliferator-activated receptor (PPAR) gamma coactivator-1 alpha (PGC-1α) is an established regulator of skeletal muscle mitochondrial biogenesis, though recent evidence (reviewed in Islam et al. 2019) has also identified additional protein with important roles in mitochondrial remodelling, including the nuclear hormone receptor PPAR-beta (PPARβ). PPARβ heterodimerizes with the retinoic acid receptor to bind specific DNA sequences known as PPAR response elements (PPREs), which are present on the promoters of various genes involved in oxidative metabolism. In the absence of a ligand (e.g. free-fatty acids), PPARβ’s association with corepressors (e.g. nuclear receptor corepressor-1; NCoR1) inhibits its transcriptional activity, whereas ligand-binding and/or post-translational modification leads to the recruitment of coactivators (e.g. PGC-1α) that increase PPARβ’s transcriptional activity at PPREs. Importantly, PPARβ promotes mitochondrial biogenesis through a dual mechanism involving enhanced expression of the nuclear respiratory factor 1 (NRF-1) and stabilization of existing PGC-1α protein (Koh et al. 2017). Given that both PGC-1α and PPARβ are involved in the transcriptional control of mitochondrial biogenesis (Islam et al. 2019), a functional interaction between these two regulatory protein (Koh et al. 2017) implicates the PGC-1α/PPARβ axis as an important mediator of exercise-induced mitochondrial remodelling in skeletal muscle. In the current issue of The Journal of Physiology, Lima and colleagues (Lima et al. 2019) employed in silico and in vitro models to investigate the potential involvement of the PGC-1α/PPARβ axis in the transcriptional control of uncoupling protein 3 (UCP3) expression in skeletal muscle. First, the authors used public ChIPseq datasets to demonstrate that the UCP3 promoter contains putative PPREs that overlap with binding sites for PGC-1α. Second, Lima et al. (2019) examined UCP3 mRNA levels and PGC-1α interactions at the UCP3 promoter during myogenesis. Coincident with an increase in UCP3 mRNA levels, myogenesis increased the presence of PGC-1α at a binding site on the UCP3 promoter that overlapped with a known PPRE. Supporting the co-involvement of PGC-1α and PPARβ in the regulation of UCP3 gene expression in skeletal muscle cells, silencing PGC-1α abolished the increase in UCP3 mRNA following treatment with a PPARβ ligand. Third, chronic treatment of myotubes with palmitate enriched the presence of PGC-1α at the PPRE on the UCP3 promoter, activated PPRE-driven gene transcription, and increased UCP3 mRNA levels. Moreover, UCP3 knockdown in myoblasts overexpressing PGC-1α abolished the palmitate-induced increase in mitochondrial respiratory function, highlighting UCP3 as an essential mediator of PGC-1α’s ability to enhance mitochondrial function in response to metabolic overload. Collectively, these findings exemplify the importance of the PGC-1α/PPARβ axis in the induction of UCP3 gene expression. An additional noteworthy observation was the increase in PGC-1α activity despite reductions in PGC-1α mRNA/protein levels during myogenesis and/or in response to chronic palmitate exposure. Importantly, this observation demonstrates that PGC-1α’s activity can dissociate from changes in its mRNA/protein content under conditions of increased metabolic stress. In their article, Lima et al. (2019) provide an interesting discussion on the interactions between the PGC-1α/PPARβ axis and UCP3 expression in response to myogenesis and fatty acid exposure. In an attempt to the extend the authors’ discussion, we consider the implications of their findings within the context of exercise-induced mitochondrial biogenesis in skeletal muscle. Specifically, we discuss (1) the importance of the PGC-1α/PPARβ axis in the transcriptional control of exercise-induced mitochondrial biogenesis and (2) the implications of the dissociation between PGC-1α activity and its mRNA/protein content. The findings of Lima and colleagues (2019) extend several recent reports highlighting the dynamic interplay and functional redundancies between PGC-1α and PPARβ in the control of skeletal muscle mitochondrial biogenesis (reviewed in Islam et al. 2019). For instance, both PGC-1α and PPARβ are downstream targets of the cellular energy sensor AMP kinase (AMPK), a bona fide molecular transducer of mitochondrial biogenesis. Further, both proteins are upstream activators of NRF-1, a key transcription factor that coordinates the expression of nuclear- and mitochondrial-encoded genes required for mitochondrial biogenesis. Moreover, PPARβ overexpression in murine skeletal muscle enhances AMPK phosphorylation via increased NRF-1-mediated transcription of the calcium/calmodulin-dependent protein kinase kinase beta (CaMKKβ) (Koh et al. 2017). In addition to sharing upstream activators and downstream targets involved in mitochondrial biogenesis, PGC-1α is a known coactivator of PPARβ-mediated transcription at PPREs and is itself one of PPARβ’s transcriptional targets. PPARβ has also recently been implicated in the post-translational control of PGC-1α protein stability by preventing its ubiquitination and proteasomal degradation (Koh et al. 2017). Finally, both PGC-1α and PPARβ can autoregulate their own expression in a myocyte enhancer factor 2 (MEF2)-dependent manner (Koh et al. 2019). Coupling the aforementioned observations with the work of Lima and colleagues (2019), we speculate that the PGC-1α/PPARβ axis forms the basis of a molecular network involved in the regulation of exercise-induced mitochondrial biogenesis. Specifically, we propose a model (Fig. 1) where exercise-induced activation of AMPK stimulates the PGC-1α/PPARβ axis leading to the coordinated upregulation of nuclear- and mitochondrial-encoded gene transcription via enhanced NRF-1 expression and/or coactivation of existing NRF-1 protein. Because NRF-1 is also a transcription factor for CaMKKβ, the NRF-1-mediated increase in CaMKKβ expression may further stimulate the PGC-1α/PPARβ axis via enhanced AMPK phosphorylation by CaMKKβ (Koh et al. 2017). Concurrently, given that PPARβ stabilizes existing PGC-1α protein (Koh et al. 2017) and that metabolic stress enriches the presence of PGC-1α at PPREs (Lima et al. 2019), is it plausible that exercise enhances PGC-1α’s coactivation of PPARβ-mediated transcription on the promoters of key mitochondrial biogenic genes (e.g. NRF-1 and PGC-1α). This latter process may be further bolstered by the ongoing stimulation of PGC-1α/PPARβ promoter activities by their respective MEF2-dependent positive feedback loops, which would presumably increase cellular levels of PGC-1α and PPARβ (Koh et al. 2019). While distinct components of this hypothetical model have been confirmed in separate experiments, this speculative pathway remains to be examined as a whole in a single comprehensive study. Although our model highlights the transcriptional activation of gene expression following exercise, it is also possible that the control of mitochondrial biogenesis by PGC-1α/PPARβ is facilitated by a reduction in the activity of NCoR1, a transcriptional corepressor that is recruited to PPREs on the UCP3 promoter (Yamamoto et al. 2011). Importantly, as NCoR1 mRNA/protein expression is reduced in response to metabolic stress (e.g. fatty acid exposure, exercise) thereby facilitating the induction of oxidative gene expression (Yamamoto et al. 2011), the contribution of transcriptional de-repression in the control of mitochondrial biogenesis by PGC-1α/PPARβ provides an important avenue for future research in exercised human skeletal muscle. Lima et al.’s (2019) observation of a dissociation between PGC-1α activity and protein/mRNA content in response to cellular stressors (i.e. myogenesis and chronic palmitate exposure) has widespread implications for the field of exercise physiology. Changes in PGC-1α mRNA expression are a frequently utilized outcome measure when examining the potency of an exercise stimulus, with the common interpretation of larger post-exercise increases in PGC-1α mRNA expression as greater activation of existing PGC-1α protein (based on the MEF2 positive feedback loop). However, the apparent dispensability of altered PGC-1α mRNA expression and protein levels for enhanced PGC-1α activity reported by Lima and colleagues (2019) questions the utility of measuring changes in PGC-1α mRNA to infer changes in PGC-1α protein activity following acute exercise. As such, the methods employed by Lima et al. (2019) to assess PGC-1α activity (e.g. qChIP and luciferase activity experiments) or other indices of protein activity (e.g. post-translational modifications and/or changes in subcellular localization) may be more appropriate for approximating changes in PGC-1α activity following exercise. The dissociation between PGC-1α activity and content may also help explain previous seemingly paradoxical reports of exercise training resulting in reduced levels of proteins that are involved in metabolic homeostasis (e.g. Nrf2, SIRT1) (Islam et al. 2019). For example, a training-induced decrease in protein content may be paralleled by an increase in the activity of that same protein, thereby alleviating the need for an increase in protein content. Further, this observation may also explain why PGC-1α protein levels dissociate from mitochondrial content in distinct human fibre types (Gouspillou et al. 2014). Specifically, although highly oxidative type I fibres contain lower levels of PGC-1α protein content compared to less oxidative type IIA fibres (Gouspillou et al. 2014), it is possible that the pool of PGC-1α protein in type I fibres is more active than PGC-1α protein in type IIA fibres. Although future work is needed to measure changes in protein activity following training and PGC-1α activity in different human fibre types, these two scenarios highlight physiological phenomena that may also be explained by dissociations between protein content and activity. Overall, Lima and colleagues (2019) conducted a series of elegant experiments to demonstrate the importance of the PGC-1α/PPARβ axis on UCP3 gene expression in skeletal muscle. While these authors did not examine the effects of exercise per se, we believe their findings have major implications for the molecular regulation of exercise-induced mitochondrial biogenesis in skeletal muscle. None. Both authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed. None. We thank our supervisor Dr Brendon J. Gurd for his careful evaluation of our article and the constructive feedback he provided.