The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 60:1494-1509 (2005)
© 2005 The Gerontological Society of America
Peroxisome Proliferator-Activated Receptor 
Coactivator 1 in Caloric Restriction and Other Models of Longevity
J. Christopher Corton1, and
Holly M. Brown-Borg2
1 United States Environmental Protection Agency, Research Triangle Park, North Carolina.
2 Department of Pharmacology, Physiology and Therapeutics, University of North Dakota School of Medicine and Health Sciences, Grand Forks.
Address correspondence to J. Christopher Corton, PhD, Division of Environmental Carcinogenesis, NHEERL, US-EPA, MD B105-01, Research Triangle Park, NC 27711. E-mail: corton.chris{at}epa.gov
 |
Abstract
|
|---|
Dietary restriction of calories (caloric restriction [CR]) increases longevity in phylogenetically diverse species. CR retards or prevents age-dependent deterioration of tissues and an array of spontaneous and chemically induced diseases associated with obesity including cardiovascular disease, diabetes, and cancer. An understanding of the molecular mechanisms that underlie the beneficial effects of CR will help identify novel dietary, pharmacological, and lifestyle strategies for slowing the rate of aging and preventing these diseases as well as identify factors which modulate chemical toxicity. Here, we review the involvement of transcriptional coactivator proteins, peroxisome proliferator-activated receptor (PPAR) 
coactivator 1 (PGC-1) 
and ß, and regulated nuclear receptors (NR) in mediating the phenotypic changes found in models of longevity which include rodent CR models and mouse mutants in which insulin and/or insulin-like growth factor-I signaling is attenuated. PGC-1 is transcriptionally or posttranslationally regulated in mammals by: 1) forkhead box "other" (FoxO) transcription factors through an insulin/insulin-like growth factor-I -dependent pathway, 2) glucagon-stimulated cellular AMP (cAMP) response element binding protein, 3) stress-activated kinase signaling through p38 mitogen-activated protein kinase, and 4) the deacetylase and longevity factor sirtuin 1 (SIRT1). PGC-1 and PGC-1ß regulate the ligand-dependent and -independent activation of a large number of NR including PPAR and constitutive activated receptor (CAR). These NR regulate genes involved in nutrient and xenobiotic transport and metabolism as well as resistance to stress. CR reverses age-dependent decreases in PGC-1, PPAR, and regulated genes. Strategies that target one or multiple PGC-1-regulated NR could be used to mimic the beneficial health effects found in models of longevity.
CALORIC restriction (CR) has broad effects on increasing life span and attenuating chronic diseases of aging that is unequaled by any pharmacological or other dietary intervention. CR has prolongevity effects in disparate invertebrate and vertebrate species including nematodes, fruit flies, spiders, fish, and reptiles (1). In many rodent studies, CR regimens increase mean and maximum life span, increase resistance to a wide variety of physical and chemical stressors, and reduce the incidence and onset of age-related diseases (2–4).
In animals the development of an efficient energy storage and utilization machinery has allowed survival during times of food shortages through appropriate management of available energy. However, these abilities are ill-suited to the lifestyles of modern society, as evidenced by the fact that obesity has reached epidemic proportions in most industrialized nations. More than 60% of U.S. adults are now overweight or obese, predisposing millions of Americans to chronic diseases including cardiovascular disease (CVD), diabetes mellitus, and certain forms of cancer (5,6). Although high consumption of calorie-rich foods and sedentary lifestyles are considered the major risk factors for the high incidence of obesity, genetic and socioeconomic factors also play a significant role (5).
Laboratory animals fed an ad libitum diet can be considered models for sedentary humans who are at risk for obesity and associated diseases. CR counteracts the trend for laboratory rodents and primates to progressively add body fat during aging. In rodent models, CR decreases the incidence of many diseases associated with obesity (7). Recent studies of CR in rhesus monkeys (8,9) have produced physiological effects that parallel those observed in rodents including reduction of established risk factors for diabetes and CVD. However, the effects of CR on monkey mortality and morbidity require further study. Although there are indications that biomarkers of CR associated with increased longevity are more frequently observed in long-lived humans (10), and there are abundant epidemiological data linking excessive caloric intake to many chronic diseases, evidence is lacking whether CR can retard aging processes in humans (11).
CR has been used to describe a wide variety of protocols but is generally defined as a reduction in the number of calories consumed to a level less than that eaten by ad libitum fed animals without any observable malnutrition. Malnutrition is important to monitor, because it can cause secondary effects predisposing animals to a number of diseases including cancer. Two paradigms have proven effective in increasing life span and disease resistance in rodents. In the more common paradigm, animals are provided a daily food allotment that is typically 30%–40% less than the ad libitum consumption of a control population; this feeding regimen leads to initial decreases and then usually stabilization of body weight. In the second paradigm, used in only a limited number of studies, animals are subjected to fasting every other day (EOD) (e.g., 12–15) or once a week (16,17). EOD fasting results in lower body weights although there is one mouse strain that loses little or no weight (13). Animals on the EOD regimen live longer than do animals on an ad libitum diet (12,13). Animals on either of the fasting protocols share some of the beneficial effects of CR (14–17). It should be noted that many of the phenotypic effects of CR cannot be adequately attributed to the CR itself or to the fasting before killing that some protocols follow.
Recent studies have shed new light on the molecular basis for CR effects. Here, we review the involvement of transcriptional coactivator proteins collectively called peroxisome proliferator-activated receptor (PPAR) coactivator-1 (PGC-1) in the phenotypic effects found in CR as well as other models of longevity. PGC-1 is negatively regulated by a well conserved insulin/insulin-like growth factor (IGF-I)-dependent pathway involved in longevity and stress resistance. We examined the: 1) role of insulin/IGF-I signaling in longevity, 2) regulation of PGC-1 expression and activity by fasting or CR-induced changes in insulin/IGF-I signaling, 3) regulation of nuclear receptor (NR) activity by PGC-1, 4) roles of PGC-1-regulated NR in CR effects, 5) role of PGC-1 in other models of longevity, and 6) reversal of age-dependent decreases in PGC-1 and regulated genes by CR.
|
PGC-1 FAMILY MEMBERS INVOLVED IN RESPONSES TO CR
|
|---|
Animals have evolved highly regulated systems to maintain nutrient levels within tight limits, even during times of food deprivation. Blood glucose levels are determined by glucose uptake in peripheral tissues and the hormonal modulation of glucose production. Glucose production is controlled by both transcriptional and posttranslational mechanisms. During short-term fasting, exercise, or times of stress, glucose is released from glycogen stores by posttranslational activation of glycogenolysis enzymes. When glycogen stores are depleted, metabolic gene expression programs including those involved in gluconeogenesis are activated to maintain appropriate glucose levels. Furthermore, triglycerides are released from fat stores in adipocytes by the action of lipases and are carried to the liver and other organs to be used for fatty acid oxidation.
PGC-1 family members have emerged as central regulators of the adaptive responses to caloric deprivation. The first member of the family to be characterized was PGC-1, a protein that interacts with the NR PPAR, and acts as a master regulator of brown fat differentiation during cold adaptation (18). PGC-1 also carries out essential roles in white fat differentiation, mitochondrial biogenesis, and muscle fiber-type switching (19). PGC-1 and related family member PGC-1ß (also called PGC-1 estrogen receptor coactivator [PERC]) coordinately regulate genes involved in gluconeogenesis and fatty acid ß-oxidation in a number of organs during fasting (19–24). An additional structurally related family member has been preliminarily characterized called PGC-1-related coactivator [PRC; (25)] that may also be involved in mitochondrial biogenesis (26). PGC-1 family members, like other coactivators, modulate transcription by providing bridging interactions between DNA binding transcription factors and the transcriptional machinery in the putative absence of any direct interaction with DNA themselves. Ligand-dependent and -independent activation of NR by PGC-1 family members results in increases in transcription initiation of target genes.
The PGC-1 family members share homology within specific regions of their primary protein sequence relevant to their functions as coactivators (19). Their N termini contain an acidic transcriptional activation domain important for interaction with NR. Within this region are embedded at least three LXXLL motifs (where L is leucine and X is any amino acid) which provide a protein interface in an -helical region essential for interactions between NR and coactivators. The sequences flanking the LXXLL motif provide a role in determining NR selectivity. The C termini of PGC-1 family members contain an RNA-binding domain and an arginine–serine-rich domain, both of which are likely important for interaction with messenger RNA (mRNA) processing complexes (27). These structural domains in the N and C termini have been conserved in rodents and humans (19) suggesting that underlying mechanisms of transcriptional regulation by PGC-1 family members during times of nutritional deprivation may be conserved.
Two mouse lines in which the PGC-1 gene was disrupted have been recently reported (23,24). Both lines exhibit cold intolerance, decreases in hepatocyte respiration rates, and neurological lesions. The lines also possess a number of interesting differences including the level of basal and fasting-induced expression of gluconeogenesis genes, and a fasting-induced hepatic steatotic phenotype which may be due to reductions in fatty acid oxidation rate and activation of triglyceride synthesis. These mice will be useful models for understanding the role of PGC-1 in fasting and CR.
|
ROLE OF INSULIN/IGF-I SIGNALING IN LONGEVITY
|
|---|
Insulin and IGF-I play contributory roles in diseases of aging, whereas CR counteracts their effects. An insulin-regulated pathway is important in determining longevity in diverse species (Figure 1). In mammals, insulin and IGF-I bind to their respective receptors triggering a cascade of events including activation of phosphatidylinositol 3-kinase (PI3K) and serine–threonine protein kinases (Akt-1/Akt-2/protein kinase B [PKB]). In the worm Caenorhabditis elegans, this pathway determines responses to longevity and environmental stress (28). Mutations in C. elegans which inactivate the insulin/IGF-I pathway, including Daf-2, the receptor for insulin/IGF-I or the PI3K ortholog Age-1, result in increased longevity as well as increased thermotolerance and antioxidant defenses. These effects require reversal of negative regulation of the stress resistance factor, Daf-16 (29–33). Daf-16 encodes a transcription factor containing a "winged helix" or "forkhead" DNA binding domain. (These factors are now called forkhead box [Fox] factors.) Overexpression of Daf-16 in worms (33) or an ortholog in flies (34) extends life span. Daf-16 regulates the expression of a large number of genes involved in xenobiotic metabolism and stress resistance, many of which when individually inactivated or overexpressed in C. elegans have incremental effects on longevity (30–32).
View larger version (53K):
[in this window]
[in a new window]
|
Figure 1. Regulation of peroxisome proliferator-activated receptor (PPAR) coactivator 1 (PGC-1) by insulin. Insulin and insulin-like growth factor-I (IGF-I) bind to their respective receptors triggering a cascade of events including activation of phosphatidylinositol 3-kinase (PI3K), p66Shc, and serine-threonine protein kinases (Akt-1/Akt-2/protein kinase B [PKB]). Mutations in Caenorhabditis elegans which inactivate the insulin/IGF-I pathway including Daf-2, the receptor for insulin/IGF-I, the PI3K ortholog (Age-1), or Akt result in increased longevity. These effects require reversal of negative regulation of the stress resistance factor, Daf-16. In times of nutrient availability, binding of insulin or insulin-like molecules to their respective receptors leads to activation of PI3K and Akt, inactivation of Daf-16 in C. elegans and forkhead box "other" (FoxOs) in mammals by phosphorylation and subsequent transport of these factors out of the nucleus. FoxOs regulate the activity and expression of the transcriptional coactivator, PGC-1. PGC-1 family members control the activity of a large number of nuclear receptors. FoxO and PGC-1 activity is also negatively regulated by cellular AMP (cAMP) response element binding protein binding protein (CBP), through acetylation during conditions of oxidative stress. The deacetylase sirtuin 1 (SIRT1), positively regulated by FoxO1 and p53, negatively regulates the effects of CBP by deacetylating FoxO family members and PGC-1. FoxO is also phosphorylated by the Jun N-terminal kinase (JNK)-1 under conditions of environmental stress resulting in nuclear localization and activation of PGC-1
|
|
Mammalian homologs of Daf-16 fall into the FoxO (O for "other") family of forkhead box factors including FoxO1 (also known as FKHR), FoxO3a (FKHR-L1), FoxO4 (AFX) (35), and the less characterized FoxO6 (36). In times of nutrient availability, binding of insulin or insulin-like molecules to their respective receptors leads to activation of PI3K and Akt and phosphorylation of Daf-16 in C. elegans and FoxOs in mammals. Phosphorylated FoxO factors are recognized by 14-3-3 proteins which facilitate transport out of the nucleus. When insulin signaling is decreased (such as during fasting or CR) there are not only increases in nuclear/cytoplasmic FoxO ratios but FoxO factor expression as well (37–40). Although the C. elegans Daf-16 appears to be uniquely regulated by insulin-like molecules, the mammalian FoxOs are negatively regulated by insulin (41) as well as growth factors, cytokines (42), an insulin/IGF-I effector, p66Shc probably through activation of Akt/PKB (43), and serum/glucocorticoid-regulated kinase-1 (SGK-1), a p53-activated protein kinase (44).
IGF-I, like insulin, stimulates PI3K and Akt/PKB and inhibits the function of FoxO1 (45–52), FoxO3, and FoxO4 (53) through phosphorylation and nuclear export (Figure 1). Phosphorylation and inactivation of FoxO factors by IGF-I and insulin probably occur through different mechanisms, as the sites of phosphorylation are not the same (48,54). Although there is presently no evidence that IGF-I regulates PGC-1, IGF-I does inhibit the expression of gluconeogenesis genes (55–57) by a mechanism similar to that of insulin albeit through different membrane receptors (57).
Mutations in mammalian genes which downregulate insulin/IGF-I signaling increase longevity. Genes which determine growth hormone (GH) secretion (Ghrhr, Pit1, Prop1) or signaling (Ghr) are partially or completely inactivated in dwarf mice and result in decreased insulin and IGF-I levels and increased longevity (58,59) (discussed in greater detail below). Mice nullizygous for p66Shc exhibit increased ability to respond to oxidative stress (60) and, along with IGF-I receptor heterozygote mice that lack wild-type activation of p66Shc (61), live longer than do wild-type mice. Adipocyte-specific deletion of the insulin receptor gene also increases longevity (62). In another model of longevity, transgenic mice containing the murine uroplasminogen activator under the control of the A-crystallin gene promoter (MUPA mice) have lower serum IGF-I levels possibly because the mice eat less (63).
CR in rodents also has dramatic effects on insulin signaling. CR uncouples insulin/IGF-I signaling to FoxO factors by markedly reducing plasma IGF-I and insulin levels over the life span of rats (64,65). These decreases in circulating insulin and IGF-I levels are associated with decreased Akt phosphorylation in liver (39), decreased PI3K expression in muscle (66), and increases in the expression of FoxO family members by fasting (37,38) or CR (39,40). In addition, a diet low in the amino acid methionine led to decreases in insulin and IGF-I and increases in life span without altering food intake (67). However, given that the mice lost weight on this diet, further studies are necessary to determine if the diet alters absorption, excretion, and physical activity.
There is evidence that long-lived humans such as centenarians have decreased insulin signaling. In the Baltimore Longitudinal Study of Aging, a decreased insulin level is one predictor of longer life in males (10). Long-lived humans (i.e., >85 years) have lower IGF-I plasma levels, and these levels are affected by polymorphisms in IGF-I receptor and phosphatidyl inositol 3-kinase (PI3K) genes (68). Women with one copy of a variant of the GH gene were 20% less likely to die at any given age than were women with other variants of GH (69). Obesity can lead to insulin resistance, higher circulating levels of insulin/IGF-I, and increased incidence of an array of diseases that may have as their basis uncontrolled insulin signaling (5,70). Overall, the studies reviewed in this section demonstrate that downregulation of insulin/IGF-I signaling results in increases in the activity of Daf-16 and FoxO factors consistently found in diverse models of longevity.
|
SIRTUIN 1 AND JUN N-TERMINAL KINASE SIGNALING TO FOXO
|
|---|
FoxO factors may be a key link to other mechanisms that affect longevity. Like Daf-16 in C. elegans, mammalian FoxO family members carry out functions that determine cell survival during times of stress including regulation of apoptosis, cell-cycle checkpoint control, and oxidative stress resistance (35,71). Activation of FoxO3 or FoxO4 leads to increases in cell-cycle G1 arrest (72,73) and increases in apoptosis (73,74) presumably as a way to eliminate cells damaged by oxidative stress. In addition to negative regulation by insulin signaling, FoxO factors are regulated by the cAMP response element binding protein (CREB) binding protein (CBP) or a related protein called p300. The activity of CBP and p300 can enhance trans-activation of DNA binding transcription factors through histone acetylation and dissolution of chromatin structure which facilitates assembly of transcription complexes on regulated promoters. Overexpression of CBP (75) or p300 (76) enhances the ability of FoxO factors to trans-activate gene expression through this mechanism. However, CBP also exhibits negative effects on FoxO factors through direct acetylation of FoxO proteins (72–76) triggered on exposure of cells to hydrogen peroxide (72,74) or other treatments that induce oxidative stress including ultraviolet light (74) and heat (75).
The sirtuin 1 (SIRT1) protein can reverse the negative regulation of FoxO family members by CBP. SIRT1 is a mammalian SIR2 family member of conserved NAD+-dependent deacetylases that have roles in gene silencing, DNA repair, and aging (77,78), and is activated by polyphenolic compounds including resveratrol (found in red wine) which increase longevity in C. elegans and D. melanogaster in a SIR2-dependent manner (79). In D. melanogaster, SIRT2 is also required for CR to increase longevity (80). Like PGC-1, SIRT1 levels are increased during CR in rat liver and are negatively regulated by insulin and IGF-I (81). Additionally, the related family member SIRT3, a mitochondrial protein, exhibited increased expression in white and brown fat upon CR (82). SIRT1 gene expression is under positive control by FoxO3 and p53 (83). Interestingly, SIRT1 at least, partially reverses the acetylation of FoxO1, FoxO3a, and FoxO4 caused by CBP or p300 (72–76). Deacetylation is coincident with direct interactions between SIRT1 and FoxO family members (72–76), enhanced in response to acetylation during oxidative stress (72,73). SIRT1 has divergent effects on FoxO-regulated gene expression including positive regulation of cell-cycle arrest or stress genes p27, GADD45, or MnSOD (72,73,75), and negative regulation of PEPCK involved in gluconeogenesis and the proapoptotic regulator, Bim (74). These important findings provide a logical framework by which CR, through increased expression and activity of SIRT1, can protect cells during oxidative stress by enhancing G1 arrest and decreasing apoptosis, allowing time for FoxO-regulated genes to repair the damage.
FoxOs are regulated by c-Jun N-terminal kinase 1 (JNK-1), which serves as a molecular sensor for various stressors. In C. elegans, JNK-1 directly interacts with and phosphorylates Daf-16 and in response to heat stress, JNK-1 promotes the translocation of Daf-16 into the nucleus. Overexpression of JNK-1 in C. elegans leads to increases in life span and increased survival after heat stress (84). In D. melanogaster, mild activation of JNK leads to increased stress tolerance and longevity (85) dependent on an intact dFoxO (86). JNK-1 activity antagonizes insulin-like signaling to dFoxO leading to nuclear translocation. Expression of activated JNK-1 in insulin-producing cells in the D. melanogaster brain is sufficient for life-span extension (86). Oxidative stress can also induce JNK-1-dependent phosphorylation and nuclear translocation of mammalian FoxO4, even in the presence of serum growth factors which allow for a basal level of PKB activity (87). Thus, when PKB activity is relatively low (as opposed to insulin-stimulated PKB activity), the JNK-1 signaling pathway acts opposite to and dominates PKB. At higher concentrations of hydrogen peroxide following treatment with insulin, PKB activation increases to dominate JNK-1 activity (87). A further understanding of this interplay of signaling will help to determine the relative contribution of PKB and JNK-1 in modulating the ability of FoxO to regulate different sets of cell survival genes under different environmental conditions.
|
REGULATION OF PGC-1 GENE EXPRESSION
|
|---|
PGC-1 is tightly regulated by environmental stimuli and hormones that respond to altered glucose levels. Insulin negatively regulates gluconeogenesis and fatty acid ß-oxidation in mammals through effects on the PGC-1 promoter as well as FoxO interactions with PGC-1 itself (see below). When insulin levels are low, PGC-1 and PGC-1ß genes are induced by fasting in rodents (19,88–90). PGC-1 was also induced in the livers of mice (91) and rats (92) after longer term CR. A 5-day severe CR induced PGC-1 mRNA expression in skeletal muscle of obese but not lean humans (93). A 10-week hypocaloric diet in obese persons resulted in 
7 kg loss of weight and increases in PGC-1 in subcutaneous fat (94). In reporter gene assays, insulin as well as ectopic expression of an activated form of Akt suppressed PGC-1 promoter activity. FoxO1 induced PGC-1 expression through insulin-responsive sequences in the PGC-1 promoter (95) (Figure 2). Inhibition of FoxO1 activity by Akt can account for the inhibitory effect of insulin on PGC-1 promoter activity. Consistent with this, inactivation of the insulin receptor in mouse liver increased the expression of PGC-1 (96), and chronic activation of Akt in mouse heart led to decreased expression of PGC-1 (97).
PGC-1 is transcriptionally regulated by glucagon, a hormone produced and secreted by pancreatic alpha cells in response to hypoglycemia. Glucagon, acting via cAMP, is a major activator of gluconeogenesis genes in the liver. Treatment of hepatocytes with 8-bromo-cAMP, a cell-permeable form of cAMP which activates protein kinase A (PKA), induced mRNA expression of gluconeogenic genes. CREB, which is activated by glucagon through a cAMP-activated PKA-dependent pathway, increased the expression of PGC-1 through binding to the PGC-1 promoter (Figure 2). Overexpression of PGC-1 in CREB-null mice rescued the expression of gluconeogenic genes and restored glucose homeostasis (88). Additionally, ectopic expression of a mitochondrial sirtuin (SIRT3) in brown adipocytes increases PGC-1 expression, most likely through increased phosphorylation and activation of CREB (82).
Conditions that induce mitochondrial biogenesis also increase PGC-1 promoter activity. Nitric oxide produced by endothelial nitric oxide synthase controls mitochondrial biogenesis possibly through increased expression of PGC-1 as well as other transcription factors. This process is mediated by cyclic guanosine 3',5'-phosphate, resulting from activation of "soluble" guanylate cyclase (98). The mechanisms by which PGC-1 expression is increased by cyclic guanosine 3',5'-phosphate have not been identified. Exercise leads to elevated intracellular calcium levels resulting in the activation of calcium/calmodulin-dependent protein kinase IV (CaMKIV) and calcineurin in skeletal muscle. Activated CaMKIV phosphorylates CREB, which increases transcription at CREB-responsive elements in the PGC-1 promoter. CaMKIV and calcineurin increase the transcriptional activity of myocyte enhancer factor 2 (MEF2), which in combination with another transcription factor, Nuclear Factor of Activated T-cells (NFAT), binds to at least one MEF2 binding site in the PGC-1 promoter (99). Exercise also increases PGC-1 gene expression through p38 mitogen-activated protein kinase (MAPK) pathway in skeletal muscle (100). It is worthwhile to note that long-term CR abrogates the age-dependent decline in skeletal muscle aerobic function by preventing the decline of mitochondrial oxidative capacity (101) which may have at its basis increases in PGC-1 expression and activity.
Caloric deprivation switches off adenosine triphosphate (ATP)-consuming processes such as biosynthetic pathways, while switching on catabolic processes that generate ATP including fatty acid uptake and oxidation (102). These processes are controlled in part by 5'-AMP-activated protein kinase (AMPK), through increased cellular AMP concentrations that accompany a fall in the cellular ATP/AMP ratio. Overexpression of the AMPK subunit AMPK2 was shown to extend life span in C. elegans (103). Basal level expression of PGC-1 in skeletal muscle was decreased in mice deficient in the AMPK2 subunit (104). Exposure of mice to ß-guanidinopropionic acid, a creatine analog which activates AMPK, led to increases in expression of PGC-1 possibly due to increases in CaMKIV (105). However, evidence exists for adaptive responses to fasting and CR being AMPK-independent (106).
|
POSTTRANSLATIONAL REGULATION OF PGC-1 ACTIVITY
|
|---|
PGC-1 is regulated by three types of posttranslational modifications. Activated under conditions of stress, p38 MAPK phosphorylates PGC-1 at three sites within a region involved in repression of transcriptional activation. Phosphorylation results in increased activation of genes involved in ß-oxidation of fatty acids (107,108). Overexpression of an N-terminal region of PGC-1 containing the NR interaction domain overcomes repression of target gene expression (109). This N-terminal region contains a binding site for p160 myb binding protein, a repressor of PGC-1 activation of mitochondrial genes (110). Thus, the negative regulation of PGC-1 by p160 myb-binding protein is relieved by p38 MAPK phosphorylation of the PGC-1 N terminus. This mechanism provides a logical link between conditions of stress requiring higher metabolic demands and the increased expression of genes involved in energy generation.
PGC-1 activity is potentiated by arginine methylation by a protein arginine methyltransferase (PMRT1), another NR coactivator. Methylation by PRMT1 of two or three of the arginine residues found within the C-terminal glutamic acid-rich region of PGC-1 plays an important role in the coactivator function of PGC-1 and ability to activate genes involved in mitochondrial biogenesis (111). Further work is needed to determine when PMRT1 is activated, how arginine methylation changes the transcriptional activation properties of PGC-1, and the relative role of this modification in PGC-1 function.
PGC-1 activity is dependent on the acetylation state of the protein (Figure 1). PGC-1 is acetylated in vivo, and acetylation is augmented by p300 (112). SIRT1 can form a complex with PGC-1 facilitating deacetylation. The interactions between SIRT1 and PGC-1 are distinct from those of SIRT1 with p53 and FoxO3A (112). Interaction with SIRT1 and deacetylation of PGC-1 result in increases in the expression of gluconeogenesis genes, suppression of glycolytic genes, and increases in hepatic glucose output. Interestingly, SIRT1 does not regulate the effects of PGC-1 on mitochondrial gene expression (113).
|
REGULATION OF NR ACTIVITY BY PGC-1 FAMILY MEMBERS
|
|---|
NR are critical modulators of developmental and physiological processes, and are targets of drugs as well as chemicals of environmental significance. NR are divided into classes based on their properties as regulatable DNA binding transcription factors. Class I NR including estrogen and glucocorticoid receptors bind as homodimers to response elements and shuttle between the cytoplasm and nucleus depending on availability of ligand. Most of the NR regulated by PGC-1 and PGC-1ß that play known or putative roles in response to nutrient deprivation are class II NR, i.e., they heterodimerize with family members of another NR called retinoid X receptors (RXR). These heterodimers bind to NR DNA response elements found in the promoters of regulated genes consisting of two direct repeats (DR) of TGACCT separated by 1 to 5 nucleotides (DR-1 to DR-5). Class III NR include estrogen receptor-related (ERR) family members that bind as monomers to an extended direct repeat with a consensus sequence TGACCTTGA. Hepatocyte nuclear factor 4 (HNF-4) is a class IV NR which binds as a homodimer to a DR-1 element. For most NR, ligand binding brings about a conformational change within the ligand binding domain resulting in exposure of a protein interface recognized by coactivator proteins including PGC-1. Assembly of transcriptional complexes on regulated genes subsequently follows.
PGC-1 family members mediate effects on gene expression through both ligand-dependent and -independent activation of NR. Differences between PGC-1 family members exist in the spectrum of NR with which they interact and activate. PGC-1 interacts with a large number of NR including constitutive activated receptor (CAR), ERR, farnesoid X receptor (FXR), glucocorticoid receptor, HNF-4, liver X receptor (LXR), PPAR, PPAR, pregnane X receptor (PXR), vitamin D receptor, and RXR (19,21,114,115) (Table 1). PGC-1ß interacts with or activates ERR family members (21,115), PPAR, thyroid hormone receptor ß (20,116), PPAR, and ER (22). NR require one of the N-terminal LXXLL motifs of PGC-1 for ligand-dependent interaction including ER (117), LXR (118), PPAR (20,109,119), FXR (120), RXR (121), and glucocorticoid receptor (122). The LXXLL motif in PGC-1 is also important for ligand-independent interactions with NR including HNF-4 (89) and CAR (123). FXR (124), PPAR (19), LXR (118), ERR, and ERR (125,126) interact with a region of PGC-1 distinct from other NR. In contrast, PRC activates a number of non-NR transcription factors (Nrf-1, Nrf-2) involved in mitochondrial biogenesis (26). PGC-1 and ß exhibit differences in side-by-side NR activation. HNF-4 is preferentially activated by PGC-1 whereas PPAR is preferentially activated by PGC-1ß. These differences may explain the targeted activation of gluconeogenesis genes by PGC-1 (but not PGC-1ß) through HNF-4 and the activation of fatty acid ß-oxidation and mitochondrial energy production genes by both PGC-1 and PGC-1ß through PPAR (116).
|
RECIPROCAL REGULATION OF FOXO1 AND NR
|
|---|
FoxO1 regulates genes involved in gluconeogenesis in both hepatocytes and in the intact mouse liver through direct interactions with the PGC-1 protein in a manner that is inhibited by Akt phosphorylation (127). There is also evidence that NR are regulated directly as FoxO1 is a cofactor for NR-regulated transcription (Figure 3). HNF-4 (128), PPAR (129), CAR, and PXR (130) interact with FoxO1 with different functional consequences. Unphosphorylated (i.e., activated) FoxO1 inhibited HNF-4 from binding to its DNA binding site; in the presence of insulin, FoxO1 was phosphorylated (through Akt/PKB) and the inhibition of HNF-4 DNA binding activity was reversed (128). PPAR-dependent gene expression is inhibited by FoxO1 through disruption of PPAR/RXR heterodimer binding to peroxisome proliferator response elements (PPREs), and PPAR/RXR inhibited FoxO1 trans-activation activity through an unknown mechanism (129). FoxO1 was found to be a coactivator of CAR- and PXR-mediated transcription, whereas drugs that activated either CAR or PXR blocked FoxO1 activation (130). Insulin or a constitutively active mutant of Akt effectively blocked the ability of FoxO1 to coactivate CAR and PXR in vitro. Thus, insulin could repress these NR by FoxO1 phosphorylation and inactivation via Akt, whereas FoxO1-mediated transcription is repressed by drugs that activate CAR and PXR (130). Further experiments are needed to determine if these findings, based mostly on in vitro effects, have relevance in the intact animal and whether these effects are PGC-1-independent.
|
ROLE OF PGC-1-REGULATED NR IN CR
|
|---|
Many of the NR regulated by PGC-1 are altered upon short- or long-term CR and regulate genes involved in energy utilization and stress responses. In this review, we focus on PPAR and CAR that have known or plausible roles in mediating effects of CR.
PPAR
PPAR has well characterized roles in fasting and emerging roles in CR. PPAR, like other PPAR subtypes, is activated by a large, structurally diverse group of amphipathic chemicals. Most notably, PPAR is activated by peroxisome proliferators (PP) including hypolipidemic agents that increase the number and size of peroxisome organelles in mouse and rat liver (131). PPAR is also activated by a large number of compounds that are either found endogenously within the cell (e.g., fatty acids) or that are dietary components (e.g., pristanic acid, phytanic acid, and resveratrol) (132–134). In addition, the naturally occurring lipid oleylethanolamide (OEA) is a PPAR agonist and requires PPAR for regulation of satiety and body weight (135). PPAR agonists elicit a predictable course of adaptive responses in the liver, including peroxisome proliferation, induction of lipid-metabolizing genes, hepatomegaly, and eventually liver cancer (131). Studies in wild-type and PPAR-null mice demonstrated that the phenotypic effects of PP exposure in the liver require a functional PPAR (131,136).
PPAR plays a key role in carbon source utilization during fasting. PPAR was upregulated in liver, duodenum, jejunum, colon, thymus (137), and pancreas (138) by fasting and downregulated in pancreatic ß-cells by glucose (139). In contrast, PPAR was downregulated in the rat gastrocnemius muscle after a 48-hour fast, even though a large number of genes involved in fatty acid ß-oxidation, known to be regulated by PPAR, were simultaneously increased in expression (140). During fasting, PPAR-null mice exhibit defects in the ability to regulate genes involved in fatty acid ß- and 
-oxidation and ketogenesis in liver, kidney, and heart in parallel with an inability to maintain proper levels of blood glucose or ketone bodies (141–147). Although fasted PPAR-null mice can still mobilize and transport triglycerides from fat stores to the liver, fasting results in severe micro- and macrovesicular steatosis because of impairment of fatty acid ß- and -oxidation (142,143). The induction by fasting of pyruvate dehydrogenase kinase 4 (PDK4), a key gene controlling carbon source utilization was found to be PPAR-dependent in a number of tissues (148) but PPAR-independent in mouse muscles (149). In humans, the Leu162Val polymorphism of PPAR was associated with a decreased level of fasting serum triglyceride in glucose-tolerant white persons (150).
PPAR also plays roles in long-term CR. The expression of the PPAR gene but not protein was increased in the livers of CR mice (151). Examination of the transcript profiles in the livers of wild-type and PPAR-null mice revealed that 20% of all gene expression changes were PPAR-dependent after a 4-week 65% CR (91). Despite the fact that PPAR regulates fatty acid ß-oxidation genes in the liver after fasting and PP exposure, genes involved in fatty acid ß-oxidation were induced by CR through a PPAR-independent mechanism. However, the induction of fatty acid -hydroxylases Cyp4a10 and Cyp4a14 by CR was PPAR-dependent (91).
In addition to roles in lipid metabolism, PPAR likely carries out other unexpected functions during CR. PPAR was partially or completely required for CR to downregulate acute phase genes (C4bp, C9, Mbl1, Orm1, Saa4) responsive to inflammatory cytokines (91), possibly through the ability of PPAR to negatively regulate transcription factors important in inflammatory responses (131). The downregulation by CR of the epidermal growth factor receptor (Egfr), a key gene involved in carcinogenesis, required PPAR, whereas downregulation of GHR was PPAR-independent. Thus, PPAR is required for regulation of a subset of CR-responsive genes in the liver involved in fatty acid metabolism, inflammation, and cell growth. Further experiments are required to determine if these changes in gene expression are functionally relevant to suppression of inflammation and cancer found in CR animals.
CR protects the liver from a wide range of environmental stressors, many of which induce damage through inflammatory mediators (152,153). Like CR, PPAR regulates responses to diverse forms of stress. Wild-type mice preexposed to PPAR agonists exhibit decreased cellular damage, increased tissue repair, and decreased mortality after exposure to a number of physical and chemical stressors in liver (154–158). Additionally, hepatocytes from PPAR-null mice exhibit decreased resistance to oxidative stress induced by chemical exposure (159). PPAR was required for CR to protect the liver from damage induced by the hepatotoxicant thioacetamide. PPAR-null mice fed a CR diet were as sensitive to hepatotoxicity as were wild-type or PPAR-null mice on an ad libitum diet. The mechanism of CR protection likely includes enhanced repair of damage as compensatory cell proliferation was increased earlier and to a greater extent in wild-type CR mice compared to PPAR-null CR mice in the absence of major differences in the amount of tissue damage (91). The molecular basis for the PPAR-dependent increased cell proliferation in CR mice is not known, but it is possible that PPAR is required for energy production needed for tissue repair or optimal expression of repair genes (154,157). Given that PGC-1 expression is increased with CR (91,92), PGC-1 through regulation of PPAR may be required for CR to protect the liver from chemical-induced stress. Support for this hypothesis comes from studies showing that the livers of diabetic mice in which PGC-1 is induced (19) are protected from toxicity induced by acetaminophen, carbon tetrachloride and bromobenzene due to robust tissue repair (160,161). Moreover, like CR, the protection observed in diabetic mice treated with acetaminophen required PPAR (157). Taken together, these data indicate the importance of PGC-1-mediated regulation of PPAR in CR-induced hepatoprotection. Future experiments should test directly the role of PGC-1 in CR stress resistance.
CAR
A number of enzyme classes work in concert to dispose of toxic endogenous or xenobiotic compounds including carcinogens, drugs, and toxins. The NR CAR regulates phase I, phase II, and phase III pathways of oxidative metabolism, conjugation, and transport of xenobiotics (162). CAR exhibits constitutive activity as an NR but in the presence of exogenous compounds, CAR can be further activated by two mechanisms. Phenobarbital induces activation of CAR through increases in CAR phosphorylation. Other activators bind to CAR directly including the planar hydrocarbon 1,4-bis[2-(3,5-dichloropyridyloxy)] benzene. Both processes lead to increased interaction of CAR with binding sites in the promoters of regulated genes. CAR expression was induced in the livers of mice fasted for 24 or 48 hours (163).
Decreases in thyroid hormone levels may be associated with changes in metabolic rate upon CR (164). CAR controls the expression of genes involved in thyroid hormone metabolism induced upon fasting. Expression of Cyp2b10 and members of the uridine 5'-diphosphate-glucosyltransferase and sulfotransferase families (phase I and II genes) are induced in parallel with decreases in T3 and T4 levels by fasting in wild-type but not CAR-null mice (165). CAR-null mice on a 40% CR diet had a higher basal metabolic rate and lost more body weight than did wild-type mice (165). These results demonstrate that CAR controls the expression of xenobiotic metabolism genes that regulate thyroid hormone levels during CR. It is worth noting the parallel between the increased expression of xenobiotic metabolism genes by CR through CAR and in the long-lived C. elegans daf-2 mutant (compared to wild-type animals). In the daf-2 mutant, constitutively active Daf-16 increases the expression of drug detoxification genes including cytochrome P450 family members, short-chain dehydrogenases/reductases, uridine 5'-diphosphate-glucosyltransferases, and glutathione-S-transferases (166) suggesting a general role for these genes in stress resistance and possibly longevity in multiple species.
|
ROLE OF PGC-1 AND REGULATED NR IN DWARF MICE WITH INCREASED LONGEVITY
|
|---|
Dwarf mice have defects in GH action and share a number of beneficial phenotypic characteristics with rodents on CR diets. Mutations that lead to the dwarf phenotype are found in transcription factors that determine development of the anterior pituitary (Pit1, Prop1), as well as the GH releasing hormone receptor (Ghrhr) and the GH receptor (Ghr). GH elicits many of its effects through regulation of IGF-I in the liver, and dwarf mice have very low circulating levels of IGF-I. In addition, dwarf mice exhibit decreased levels of insulin and increased insulin sensitivity (167). These shared decreases in insulin and IGF-I may be the basis for beneficial phenotypic effects in common with CR and dwarf mutations. Like CR animals, dwarf mice are protected from spontaneous and chemically induced cancer (168–171), age-dependent declines in cognitive function (172), and age-related declines in immune function and collagen cross-linking (173). Fibroblasts from Snell (Pit1), Ames, and GHR-null dwarf mice exhibited increased resistance to diverse physical and chemical stressors in vitro compared to wild-type fibroblasts (174,175).
Evidence that insulin/IGF-I-regulated pathways are disrupted in dwarf mice comes from a number of studies. Snell mouse livers exhibited decreased insulin-stimulated phosphorylation of insulin receptor substrate-1 (IRS-1), IRS-2, PI3K, and Akt (176,177). Ames (Prop1) mice also had decreases in insulin-stimulated signaling including decreases in PI3K activity and Akt phosphorylation (167). Similar changes in the direction and sometimes magnitude of changes were observed in insulin/IGF-I signaling after CR in wild-type mice and in ad libitum dwarf GHR-null mice including decreases in insulin and IGF-I levels and Akt phosphorylation, with parallel increases in FoxO1 mRNA, CREB phosphorylation, and PGC-1 protein levels (39). Importantly, CR brought about additive changes in GHR-null mice in most of these facets of insulin signaling indicating that similar mechanisms are behind the changes in insulin signaling. However, there were differences in the two models including increases in Akt2 mRNA, FoxO1 protein, AMPK phosphorylation, and PGC-1 mRNA expression in dwarf mice only. In contrast, CR uniquely increased expression of SIRT1. These studies provide strong evidence that insulin/IGF-I-regulated signaling is decreased resulting in activation of FoxO1 and PGC-1, in both CR wild-type mice and ad libitum dwarf mice.
Dwarf mutations and CR result in overlaps in expression of liver genes, some of which are regulated by PGC-1. The effect of the GHR-null or Snell dwarf mutations and CR led to a partial overlap in the transcript profiles (178). A more recent study in which a greater number of genes were queried also showed that the Ames mutation and CR affect overlapping sets of genes and suggest that the additive effects of the dwarf mutation and CR on life span arise from their additive effects on the level of expression of some genes and from their independent effects on other genes (40). The Ames mutation and, to a lesser extent, CR had effects on a number of gene categories that are under control of PGC-1 including increases in fatty acid ß-oxidation and xenobiotic metabolism. Many of these same genes were also regulated to greater levels in young and old Snell dwarf mice (179) or by a 4-week CR in wild-type mice (91). In Ames, Snell, and Little mice, CAR (NR1I3, I4) and target genes Cyp2b9, Cyp2b10, Sult-N, and Sult1a1, as well as PXR (NR1I2) and target gene Cyp3a11 (179,180) exhibited increased expression indicating that these PGC-1-regulated NR may also be activated under these conditions.
Many genes under control of PPAR were constitutively regulated in dwarf mice. In Snell dwarf mice these included those involved in ß- and -oxidation of fatty acids (Acox1, Cyp4a10, Cyp4a14) and those involved in stress responses (the chaperonin, Tcp1
) and CVD (fibrinogen, Fib). The levels of some of these gene products were also altered in other dwarf mouse models including Ames, Little (Ghrhr mutant mice), and GHR-null mice (181). The constitutive increases in PPAR-regulated genes may be partly due to increased expression of PPAR mRNA and protein observed in the livers of control Snell dwarf mice (181). However, PPAR gene and protein levels were not altered in Ames dwarf mice compared to phenotypically normal controls (151).
The molecular basis for some of these gene expression changes in dwarf mice may be the antagonistic relationship between GH signaling and PPAR-regulated gene expression. STAT5b, a GH-inducible transcription factor, inhibited the ability of PPAR to activate PPAR-dependent reporter gene transcription by endogenous or xenobiotic PP in vitro (182,183), and a transcript profiling study in the livers of GH-treated rats revealed a subset of PPAR-regulated genes that were downregulated including CYP4A3 as well as PPAR itself (184). Mice with defects in GH-responsive STAT5b transcription factor exhibited constitutive increases in PPAR-regulated gene expression (185). Given the role of PPAR in mediating the transcriptional responses of CR (91), PGC-1-regulated PPAR may be one common pathway operational in dwarf mouse mutants and CR mice.
|
PGC-1 AND REGULATED NR IN EXERCISE AND HIBERNATION
|
|---|
A number of other models of longevity and/or health have features in common with CR and dwarf mice including altered expression of PGC-1 and PPAR. Exercise improves overall health and offsets the risk of age-related disease (186). The effect of exercise in rodents is primarily on median but not maximum life span, suggesting that although beneficial in terms of slowing age-dependent diseases, exercise may not retard aging at the molecular level (187). Exercise offsets many age-related gene expression changes observed in the hearts of sedentary mice, indicating that adaptive physiological mechanisms induced by exercise can retard many effects of aging on the transcriptional profile in heart muscle (188). PGC-1 (19,189,190) and PRC (190) but not PGC-1ß (90) are induced during exercise. PGC-1 promotes muscle fiber-type switching from fast-twitch fibers which rely only on glucose as an energy source (glycolytic; type IIa and IIb) to slow-twitch which rely on glucose and fatty acids (oxidative; type I) (191). PGC-1 expression was positively regulated in muscle by CaMKIV operating through CREB (99) (Figure 2). PPAR levels are also increased in the muscles of men undergoing endurance training (192). Many genes involved in fatty acid oxidation, mitochondrial biogenesis, and mitochondrial oxidative phosphorylation are regulated by PGC-1 in muscle (see below).
Hibernation can extend life span in at least one mammal (193). In hibernating 13-lined ground squirrels, Akt phosphorylation and kinase activity were significantly reduced (194) and PGC-1, PPAR, and regulated genes were coordinately increased during hibernation (195–197). Thus, an increase in PGC-1 activation of PPAR is a common feature in a number of models of mammalian longevity.
|
REVERSAL OF AGE-DEPENDENT CHANGES BY CR
|
|---|
Top
Abstract
PGC-1 Family Members Involved...
