Peroxisome proliferator-activated receptor γ (PPARγ) coactivator 1α (PGC-1α) is the founding member of a family of transcriptional coactivators that also includes PGC-1β (also called PERC) and PGC-1-related coactivator (reviewed in Puigserver and Spiegelman, 2003). PGC-1α is expressed in heart, kidney, brown adipose tissue (BAT), and brain and was originally described as a cold-inducible coactivator controlling adaptive thermogenesis in BAT and skeletal muscle by stimulating mitochondrial biogenesis and oxidative metabolism. Fasting induces hepatic PGC-1α expression thereby increasing gluconeogenesis, whereas in skeletal and cardiac muscle exercise increases PGC-1α, mitochondrial biogenesis, and respiration. Thus, PGC-1α expression seems finely tuned to reflect the cellular energy needs, with conditions of increased energy demands, such as cold, physical activity, or fasting inducing its expression. PGC-1α accomplishes these diverse functions through activating nuclear receptors like the PPARs, thyroid hormone receptors (TR), estrogen-related receptors (ERR), glucocorticoid receptor (GR) and hepatocyte nuclear factor (HNF) 4α, and other transcription factors, including nuclear respiratory factors (NRF) and FOXO1 (Figure 1). PGC-1α is dependent on other coregulators like the steroid receptor coactivator-1 (SRC-1) for its activity [Puigserver et al. 1999] and [Picard et al. 2002] .

Figure 1.

Schematic Representation of the Different Functions of PGC-1α in Its Target Tissues

See text for explanation.

Lin and coworkers now report in Cell the phenotypic analysis of PGC-1α−/− mice, which illustrates the diversity of PGC-1α physiology (Lin et al., 2004). The high postnatal mortality in PGC-1α−/− mice already underscores its importance for survival. Future studies are required to determine the cause of excessive neonatal death. Interestingly, gluconeogenesis is constitutively activated and uncoupled of nutritional status in surviving PGC-1α−/− mice. This constitutive activation of gluconeogenesis was subsequent to the induction of gluconeogenic genes by the elevated levels of C/EBPβ. These in vivo data are in apparent conflict with the predictions made from their in vitro studies and the defective gluconeogenesis in isolated PGC-1α−/− hepatocytes (Lin et al., 2004). Furthermore, acute adenoviral-mediated downmodulation of hepatic PGC-1α with small interfering RNA (RNAi) was reported to decrease gluconeogenesis leading to fasting hypoglycemia (Koo et al., 2004). The studies in PGC-1α−/−hepatocytes and with adenovirus-mediated PGC-1α RNAi address, however, acute hepatic depletion of PGC-1α, which is in contrast to the studies in animals with a genetic PGC-1α deficiency in all tissues. In such animals, both chronic compensation, as well as signals of nonhepatic PGC-1α-deficient tissues, are confounders. Temporally and spatially controlled PGC-1α mutants will be useful to clarify this issue.

PGC-1α−/− mice also have reduced mitochondrial function and defects in oxidative phosphorylation (OXPHOS). This is not only reflected by abnormal gluconeogenesis, which is dependent on ATP generated by OXPHOS, but also by impaired cold resistance. Adaptive thermogenesis in rodents is mainly a function of BAT in which mitochondrial oxidation is partially uncoupled of energy production through the action of uncoupling protein 1 (UCP-1). UCP-1 expression is reduced in PGC-1α−/− BAT. Furthermore, PGC-1α−/− mice have severe defects in expression of genes involved in OXPHOS and mitochondrial function in muscle. This muscle energy (ATP) deficit is further underscored by the activation of AMP-activated protein kinase subsequent to the reduced cellular ATP/AMP ratio. Paradoxically, whereas in man a decrease in PGC-1α, OXPHOS genes, and impaired mitochondrial activity in muscle has been linked with obesity, insulin resistance and predisposition to diabetes (Petersen et al., 2004 and references therein), PGC-1α−/− mice are lean and resistant to diet-induced obesity reflective of a substantial increase in energy expenditure. How can we explain this?

Energy expenditure is accounted for by adaptive thermogenesis and physical activity. In the wake of the decreased mitochondrial function and adaptive thermogenesis in PGC-1α−/− mice, the most likely culprit to explain the increased energy expenditure is an increased physical activity. In addition to being hyperactive, PGC-1α−/− mice also have behavioral abnormalities, stimulus-induced myoclonus, exaggerated startle responses, dystonic posturing, frequent limb clasping, and showed spongiform lesions in striatum and brain stem. Both these clinical and pathological findings are reminiscent of neurodegenerative disorders of movement, most notably of Huntington’s disease (HD), an autosomal dominant degenerative brain disorder, characterized by choreiform movements and dementia, that begins in adulthood and neuropathologically strikes the striatum. HD is the consequence of an expansion of CAG trinucleotide repeats occurring in the open reading frame of huntingtin, now encoding a protein with an expanded polyglutamine (polyQ) tract. Interestingly, homozygous and heterozygous HD patients have similar phenotypes, suggesting that the gain of function of huntingtin is toxic to neurons. The leading hypothesis to explain the HD phenotype is that the expanded polyQ tract in huntingtin interferes with other cellular processes, such as mitochondrial function. Mitochondrial dysfunction has since a long-time been linked with neurodegenerative diseases, a finding again highlighted by the decreased expression of mitochondrial genes in PGC-1α−/− mice. Other coregulators such as the CREB binding protein (CBP), the mammalian Sir2 homolog SIRT1, and TAFII130 have been linked with neurobehavioral abnormalities and neurodegeneration (Araki et al., 2004). In view of this, the role of PGC-1α and other coregulators with impact on energy homeostasis should be carefully explored in neurodegenerative diseases.

From a therapeutic perspective, PGC-1α−/− mice further validate the notion that stimulation of mitochondrial activity could be useful in the context of diseases such as obesity and associated metabolic disorders. It is interesting to note that an increase in physical activity and dietary restriction, the cornerstone of the therapy of obesity and metabolic syndrome, are both inducing PGC-1α activity. The molecular mechanism through which exercise and dietary restriction induce PGC-1α activity are not known in detail, however, several signaling pathways are being proposed. These include protein kinase A subsequent to activation of G protein-coupled receptors such as the glucagon and β3-adrenergic receptor, calcium/calmodulin-dependent protein kinase IV and calcineurin A, p38 mitogen-activated protein kinase (reviewed in [Puigserver and Spiegelman 2003] and [Kelly and Scarpulla 2004] ), S6 kinase 1, an effector of mammalian target of rapamycin, acting in an ancient nutrient-sensing signaling pathway (Um et al., 2004), and cyclin T1/cyclin-dependent kinase 9 (Sano et al., 2004). Whereas the above pathways directly target PGC-1α, another potential avenue consist of developing so-called selective nuclear receptor modulators that favor the recruitment of particular cofactors (e.g., PGC-1α, SRC-1) to nuclear receptors, as to induce a distinct and favorable biological response such as mitochondrial activation (Picard et al., 2002). Examples of nuclear receptors involved in energy homeostasis for which such modulators may be useful include the PPARs and ERRα. Finally other signaling pathways, perhaps working in parallel with PGC-1α to enhance mitochondrial activity, such as stimulating local generation of the TR ligand 3,5,3′-triiodothyronine through the induction of type 2 iodothyronine deiodinase, must be explored. It is clear that research on PGC-1α has energized the entire mitochondria field and has increased the chances that strategies to turbocharge mitochondrial activity may one day translate into therapies not only for metabolic disease, but perhaps also for certain cardiac problems and neurodegenerative diseases.