• Open Access

Aging alters PPARγ in rodent and human adipose tissue by modulating the balance in steroid receptor coactivator-1

Authors


Frédéric Picard, Laval Hospital Research Center, Y3106 Pavillon Marguerite-d’Youville, Hôpital Laval, 2725 Chemin Ste-Foy, Québec, QC G1V 4G5, Canada. Tel.: (418) 656 8711 ext 3737; fax: (418) 656 4942; e-mail: frederic.picard@crhl.ulaval.ca

Summary

Age is an important risk factor for the development of metabolic diseases (e.g. obesity, diabetes and atherosclerosis). Yet, little is known about the molecular mechanisms occurring upon aging that affect energy metabolism. Although visceral white adipose tissue (vWAT) is known for its key impact on metabolism, recent studies have indicated it could also be a key regulator of lifespan, suggesting that it can serve as a node for age-associated fat accretion. Here we show that aging triggers changes in the transcriptional milieu of the nuclear receptor peroxisome proliferator-activated receptor gamma (PPARγ) in vWAT, which leads to a modified potential for transactivation of target genes upon ligand treatment. We found that in vWAT of mice, rats and men, aging induced a specific decrease in the expression of steroid receptor coactivator-1 (SRC-1), whose recruitment to PPARγ is associated with improved insulin sensitivity and low adipogenic activity. In contrast, aging and oxidative stress did not impact on PPARγ expression and PPARγ ligand production. Age-induced loss of PPARγ/SRC-1 interactions increased the binding of PPARγ to the promoter of the adipogenic gene aP2. These findings suggest that strategies aimed at increasing SRC-1 expression and recruitment to PPARγ upon aging might help improve age-associated metabolic disorders.

Introduction

Most diseases associated with the metabolic syndrome (e.g. obesity, type 2 diabetes and atherosclerosis) appear through middle age (Toth & Tchernof, 2000), and indeed age has been shown to be an important risk factor for several of them (Lemieux et al., 1999; Pascot et al., 1999). Aging leads to profound changes in energy metabolism. Muscle mass, fatty acid oxidation and mitochondrial function in skeletal muscle decrease upon the lifespan (Toth & Tchernof, 2000), while there is a concomitant and rather proportional increase in body fat content until very old age, not only targeting visceral adipose tissue but nonadipose tissues as well, including bone marrow (Kirkland et al., 2002). Pioneer studies have established that fat cell number increases with age in adipose depots (Bertrand et al., 1978, 1980; Kirkland & Dax, 1984) whereas lipolytic control is partially lost (Dax et al., 1981; Kirkland & Dax, 1984; Yki-Jarvinen et al., 1986; Gregerman, 1994). Visceral fat accretion is central for age-associated insulin resistance since removal of visceral adipose depots in aged rats restores glucose tolerance to the levels found in young rats (Gabriely et al., 2002). Pathways contributing to altered adipocyte functions during old age have been associated with a reduction in adipogenic factors (such as C/EBPα and C/EBPδ) and increases in the expression of CUG triplet repeat-binding protein (CUGBP1) (Karagiannides et al., 2006) and that of the anti-adipogenic C/EBPβ-LIP isoform (Karagiannides et al., 2001). However, the molecular mechanisms responsible for age-induced adipogenesis, fat accumulation and insulin resistance are not fully determined, especially in middle-age individuals (Yki-Jarvinen et al., 1986; Catalano et al., 2005; Escriva et al., 2007).

Among the most studied hallmarks of aging is progressively increasing oxidative damage to proteins, lipids and DNA (Ashok & Ali, 1999; Droge, 2002, 2003). Increased levels of oxidative stress in white adipose tissue (WAT) are also observed upon fat accretion and obesity (Nakao et al., 2000; Furukawa et al., 2004; Park et al., 2007). Several reports have suggested that reactive oxygen species (ROS) production is tightly linked to the development of an inflammatory response and insulin resistance in adipocytes, which could contribute to macrophage infiltration in WAT (Cinti et al., 2005). Based on these observations and the fact that peroxisome proliferator-activated receptor gamma (PPARγ) is a key node for adipogenesis, insulin sensitivity, lipid metabolism and adipokine expression in adipocytes (Heikkinen et al., 2007), this study was aimed at assessing how PPARγ transcriptional activity is modulated during the development of middle-age-associated adipose tissue accretion, and whether the concomitant increase in ROS production can contribute to this process. We show that age does not influence PPARγ expression, but rather impacts on PPARγ activity by reducing the expression of steroid receptor coactivator (SRC) in rodent and humans, which limits its recruitment to PPARγ and triggers lipid accumulation through a modified transcriptional program.

Results

Aging modifies PPARγ transcriptional potency

To first assess whether aging modulates the ability of PPARγ to improve lipid metabolism and insulin sensitivity, young (4 months), middle-aged (12 months) and old (24 months) mice were injected with the PPARγ agonist pioglitazone for 1 week (Rocchi et al., 2001). As expected, 12 and 24 old mice weighed more than their 4-month-old counterparts (Fig. 1A) (P < 0.05). In contrast, consistent with the middle-age fat accretion process in the visceral area, the weight of the epididymal adipose depot was higher in 12, but not 24-month-old mice (Fig. 1B). Pioglitazone treatment did not affect total body weight (Fig. 1A) but slightly increased (= 0.06) the relative weight of the epididymal fat depot (Fig. 1B). In all age groups, pioglitazone strongly decreased triglyceridemia, albeit slightly more in the 12-month group (−45%) than in the other groups (−32% and −36%) (Fig. 1C).

Figure 1.

 Effects of a 1-week pioglitazone treatment (10 mg kg−1 day−1) on body weight (A), epididymal fat mass (B) and plasma triglyceride levels (C) in young (4 months), middle-aged (12 months) and old (24 months) male mice. n = 5, * and † indicate significant effects of age and drug respectively (P < 0.05).

Consistent with the above findings, the transcriptional response to pioglitazone in visceral WAT (vWAT) was robustly modified by age. In 12-month-old mice, pioglitazone treatment resulted in a significant, stronger upregulation of aP2, LPL, ATGL, adiponectin (acrp30) and Glut4 mRNA levels than it did in 4 and 24-month-old mice (Fig. 2) (P < 0.05). In contrast, pioglitazone had no effect on IRS1, IRS2, p85 and eroL1 gene expression in either group (Fig. 2). These changes cannot be attributed to a modification in PPARγ itself, as both mRNA (Fig. 3A) and protein (Fig. 3B) levels were not altered in 12-month-old mice compared to those in 4 months old. However, PPARγ protein level in vWAT was higher in 24-month-old mice (Fig. 3B). Macrophage infiltration [evidenced by a significant increase in MCP-1 mRNA levels in 24-month mice only (4 months: 0.67 ± 0.09; 12 months: 0.97 ± 0.19; 24 months: 2.25 ± 0.64, P = 0.0295)], could have contributed to modulate PPARγ expression in this group as PPARγ is highly expressed in these cells (Li et al., 2000).

Figure 2.

 Impact of aging on PPARγ transactivation of target genes by pioglitazone. mRNA expression levels of target genes were quantified in epididymal white adipose tissue from C57BL/6 male mice aged of 4, 12 and 24 months old and treated with pioglitazone for 1 week (10 mg kg−1 day−1) (n = 5). All genes were corrected for levels of GAPDH as housekeeping gene. Data are presented as fold change over vehicle treatment (dotted line). * indicates significant effects of age (P < 0.05).

Figure 3.

 Effect of age on PPARγ expression. mRNA (A–C) and protein (B–D) expression levels were quantified in epididymal vWAT from C57BL/6 male mice aged of 4, 12 and 24 months old (n = 15) (A,B) and in omental vWAT from obese men (n = 10) (C,D). * indicates significant effects of age (P < 0.05).

To determine whether these observations were conserved in men, vWAT extracts from obese patients undergoing biliopancreatic duodenal switch surgery were analyzed. Consistent with published observations in young and middle-aged nonobese men (Imbeault et al., 2001), but contrasting with a previous study showing decreased PPARγ mRNA upon aging in rhesus monkeys (Hotta et al., 1999), PPARγ expression (mRNA, Fig. 3C; protein Fig. 3D) did not differ significantly upon aging in obese men, which indicates that the obesity status does not impact on the relationship between age and PPARγ levels. These data further suggest that the potency of PPARγ to stimulate its transcriptional program is stronger during middle age, in a manner that is independent from PPARγ level itself.

Age-related oxidative stress increases PPARγ activity

To unravel the molecular mechanisms involved in age-induced increases in PPARγ activity, an in vitro model using 3T3-L1 adipocytes was set up in which aging was mimicked by chronic treatment of differentiated adipocytes with a low dose of hydrogen peroxide (H2O2). Such treatment resulted in an increase in lipid accumulation (Fig. 4A) that was concomitant with the induction of p21 and p53 gene expression (Fig. 4B), suggesting commitment toward cellular senescence. As in middle-aged mice and men (Fig. 3), PPARγ protein levels remained unchanged in H2O2-treated adipocytes (data not shown). Insulin-stimulated glucose uptake in control adipocytes was lower than in H2O2-treated cells (Fig. 4C) (P < 0.001). Moreover, consistent with previous studies (Furukawa et al., 2004) insulin-mediated inhibition of lipolysis was blunted in H2O2-treated adipocytes (Fig. 4D) (P < 0.005). In addition, treatment with H2O2 significantly stimulated the expression of aP2 and LPL but not that of IRS1, IRS2, p85 or adiponectin (Fig. 4E,F). These findings indicate that H2O2-treated adipocytes feature characteristics of fat accumulation and insulin resistance that are typical of those found upon aging.

Figure 4.

 Hydrogen peroxide treatment mimics the effects on aging in 3T3-L1 adipocytes. Differentiated adipocytes (D4) were treated with vehicle or hydrogen peroxide (H2O2) (20 μm) for 4 days (D8) and then used for (A) Oil red O staining; (B) mRNA expression levels of p21 and p53; (C) insulin-stimulated glucose uptake assays; (D) lipolysis assays and (E,F) mRNA expression levels of genes involved in lipid and glucose metabolism. Results presented are representative of at least three independent experiments done in triplicate. * and † indicate significant effects of H2O2 and norepinephrine respectively (P < 0.05). (G,H) Hydrogen peroxide increases PPARγ activity without inducing the production of PPARγ ligands. (G) 293T cells pre-treated or not with H2O2 for 4 days (20 μm) were co-transfected with a β-gal expression vector (Picard et al., 2004) and plasmids either empty or containing PPARγ cDNA. Twelve hours later, cell media was replaced with media containing pioglitazone (1 μm) or not. Twenty-four hours later, luciferase activity was quantified in cell extracts, and corrected with β-gal activity to control for transfection efficiency. * and † indicate significant effects of H2O2 and pioglitazone, respectively (P < 0.05). (H) 293T cells were co-transfected with a β-gal expression vector (Picard et al., 2004) and a plasmid containing PPARγ cDNA. Twenty-four hours later, cell media was replaced with spent media from 3T3-L1 adipocytes treated with the indicated compounds (3T3-L1-conditioned medium) or fresh media containing rosiglitazone (1 μm) or not (nonconditioned medium). Twenty-four hours later, luciferase activity was quantified in cell extracts, and corrected with β-gal activity to control for transfection efficiency. * indicates significant effects of indicated compound (P < 0.05). Results presented are the mean ± SEM of three independent experiments done in triplicate.

To test whether oxidative stress has a direct impact on PPARγ transcriptional activity, a reporter gene construct was used in which response elements for PPARγ (PPRE) are cloned upstream of the luciferase gene. H2O2-treated 293T cells co-transfected with this construct and PPARγ cDNA showed higher transactivation than vehicle-treated cells (Fig. 4G) (P < 0.05), suggesting that oxidative stress directly increases PPARγ activity. The same system was then used to test the hypothesis that age-associated oxidative stress increases PPARγ activity through the production of endogenous ligands within adipocytes (Kim et al., 1998). Spent media from 3T3-L1 adipocytes treated for 24 h with increasing concentrations of H2O2 were harvested, added to 293T cells co-transfected as above, and luciferase activity quantified 24 h later. No significant change in PPARγ-driven luciferase activity was observed in these conditions compared to fresh media (Fig. 4H). In contrast, rosiglitazone kept in the spent media for a same amount of time retained its ability to activate PPARγ (Fig. 4H). Finally, except for the known PPARγ ligand linoleic acid, no change in PPARγ activity was detected when anti-oxidants were added in the 3T3-L1 media to control for any residual H2O2-induced oxidative stress. These findings suggest that oxidative stress, a well-established feature of aging, does not increase PPARγ activity through a change in the endogenous production of PPARγ ligands.

Aging modifies PPARγ transcriptional milieu and causes an imbalance in nuclear cofactors

Faced with the findings that PPARγ appears more active during middle age and upon oxidative stress despite the absence of change in PPARγ protein expression levels and PPARγ ligand production, we tested the hypothesis that aging is associated with altered levels of nuclear cofactors, which bind to and modify the activity of transcription factors (Feige & Auwerx, 2007). The expression levels of 28 known nuclear cofactors were quantified by qPCR in epididymal WAT from 4, 12 and 24-month-old male mice. Significant modifications in mRNA levels were observed for AEBP1, SRC-1, SRC-3, and SMRT (Table 1 and Fig. 5A). Importantly, pioglitazone treatment for 1 week had no effect by itself on SRC-1 levels in young and old mice (data not shown). Further analysis revealed that SRC-1 protein levels were reduced in both 12 and 24-month-old mice compared to those in 4-month-old counterparts (Fig. 5B). This decrease did not appear to be gender specific, as a similar pattern was observed in both male and female mice (Fig. 5C). 3T3-L1 mouse adipocytes treated with H2O2 also had reduced SRC-1 mRNA and protein levels compared to vehicle-treated cells (Fig. 5D,E) (P < 0.05). To determine whether these changes were species specific, a similar screening was performed in epididymal vWAT of male rats aged 4 and 18 months old. A conserved modification in SRC-1 protein level was found upon aging in this species (Fig. 5F). Finally, a similar, conserved reduction in SRC-1 protein levels was found in middle-aged obese men compared with their younger counterparts (Fig. 5G) (P = 0.005). The mRNA levels of the other cofactors (AEBP1, SRC-3 and SMRT) were not modified upon aging in men (data not shown). Taken together, these findings indicate that the diminution of SRC-1 is a specific, conserved characteristic of aging that causes an imbalance in nuclear cofactors, which could contribute to modify the PPARγ transcriptional milieu.

Table 1.   Gene expression of nuclear cofactors in visceral adipose tissue (epididymal depot) of male mice aged 4, 12 or 24 months old
Cofactor4 months n = 1412 months n = 1524 months n = 15anova (P)
  1. Gene expression was quantified by real-time qPCR and corrected for expression of GAPDH as a housekeeping gene (no significant modulation of GAPDH upon aging). Results are mean ± SEM, n = 15.

  2. NS, not significant.

AEBP10.95 ± 0.120.90 ± 0.081.47 ± 0.150.0019
CBP1.25 ± 0.130.99 ± 0.080.95 ± 0.11NS
Bridge11.22 ± 0.221.08 ± 0.151.03 ± 0.11NS
ARA701.08 ± 0.150.92 ± 0.061.05 ± 0.12NS
PRIP1.41 ± 0.550.83 ± 0.160.78 ± 0.32NS
SRC-11.29 ± 0.091.11 ± 0.040.89 ± 0.070.0009
TIF2 (SRC-2)1.20 ± 0.230.96 ± 0.170.76 ± 0.19NS
pCIP (SRC-3)1.39 ± 0.151.05 ± 0.090.87 ± 0.160.0428
TORC21.17 ± 0.070.95 ± 0.041.03 ± 0.09NS
VinexinB1.32 ± 0.630.80 ± 0.220.99 ± 0.44NS
SIRT11.13 ± 0.250.91 ± 0.071.21 ± 0.19NS
NCoR1.08 ± 0.170.99 ± 0.051.04 ± 0.10NS
SMRT1.39 ± 0.131.04 ± 0.170.53 ± 0.100.0005
RIP1401.22 ± 0.111.11 ± 0.090.93 ± 0.10NS
TRAP2201.45 ± 0.620.85 ± 0.130.78 ± 0.28NS
HIC51.41 ± 0.151.11 ± 0.081.03 ± 0.12NS
ANT11.20 ± 0.111.14 ± 0.051.08 ± 0.08NS
PGC-1α3.25 ± 0.623.25 ± 0.613.29 ± 0.77NS
Pin10.98 ± 0.171.09 ± 0.171.07 ± 0.11NS
RLP111.24 ± 0.091.11 ± 0.071.16 ± 0.09NS
CCPG1.36 ± 0.340.90 ± 0.110.92 ± 0.09NS
CoCoA0.68 ± 0.060.68 ± 0.090.68 ± 0.07NS
Med141.36 ± 0.331.10 ± 0.170.97 ± 0.75NS
TBL12.01 ± 0.292.13 ± 0.261.89 ± 0.20NS
TAZ2.04 ± 0.251.93 ± 0.211.75 ± 0.20NS
REA1.29 ± 0.121.39 ± 0.101.63 ± 0.19NS
Chd91.22 ± 0.151.09 ± 0.101.09 ± 0.10NS
ASXL11.20 ± 0.340.91 ± 0.181.22 ± 0.33NS
Figure 5.

 Aging modulates the expression of the nuclear cofactor SRC-1 in rodent and human adipocytes. (A–C) SRC-1 mRNA (A) (n = 15) and protein (B, C) expression levels were quantified in epididymal vWAT from male (A–C) (n = 4) and female (C) (n−4) C57BL/6 mice aged 4, 12 and 24 months old. (D,E) SRC-1 mRNA (D) and protein (E) expression levels in 3T3-L1 adipocytes treated or not with H2O2 as in Figure 4. (F) SRC-1 protein levels in male rats aged 4 or 18 months old (n = 4). (G) SRC-1 protein levels in omental tissue of young (23 ± 1 years old) and middle-aged (40 ± 1 years old) obese men matched for BMI and co-morbidities (n = 4). n.s.b. is non-specific band. Bars bearing different letters are significantly different from each other (P < 0.05).

Aging is associated with altered SRC-1/PPARγ interactions

The decrease in SRC-1 levels observed upon aging and oxidative stress could likely impact on its recruitment to PPARγ. Consistent with this concept, coimmunoprecipitation of protein extracts from 3T3-L1 adipocytes treated with H2O2 showed that oxidative stress causes a loss of PPARγ-SRC-1 physical interactions in adipocytes (Fig. 6A). Moreover, reduced SRC-1 docking to PPARγ was also found in vWAT of 19-month-old rats compared with their 2-month-old counterparts (Fig. 6B). To determine whether loss of SRC-1 recruitment was associated with increased PPARγ transactivation, binding to DNA PPRE in the promoter of the adipogenic marker aP2 was assessed by electro mobility shift assay. Increased PPARγ binding was observed both in 3T3-L1 adipocytes treated with H2O2 (Fig. 6C) and in nuclear extracts from vWAT of middle-aged men compared with their respective controls (Fig. 6D). As these results indicated that PPARγ could be more active on adipogenic genes because of impaired SRC-1 docking upon aging, the hypothesis that pharmacological ligands specifically recruiting SRC-1 to PPARγ would be less active in old cells was thus tested using H2O2-treated adipocytes. As expected, rosiglitazone increased LPL expression in both control and H2O2-treated adipocytes (Fig 6E) (P < 0.05). In contrast, treatment with the PPARγ modulator and partial agonist FMOC-l-leucine (Rocchi et al., 2001), which induces the specific recruitment and docking of SRC-1 to PPARγ (Rocchi et al., 2001), failed to stimulate LPL expression in H2O2-treated adipocytes as it did in untreated cells (Fig. 6E). In addition, FMOC-l-leucine did not significantly increase insulin-stimulated glucose uptake in H2O2-treated adipocytes as it did in control cells (Fig. 6F). These findings further demonstrate that the decreased expression of SRC-1 upon aging and oxidative stress triggers an impaired cofactor recruitment to PPARγ, which leads to increased transactivation of adipogenic genes, but not of those involved in insulin sensitization.

Figure 6.

 Aging modulates SRC-1/PPARγ interactions and PPARγ DNA binding ability. (A) Co-immunoprecipitation between PPARγ and SRC-1 in nuclear extracts of 3T3-L1 adipocytes treated or not with hydrogen peroxide as in Fig. 4. (B) Co-immunoprecipitation between PPARγ and SRC-1 in pooled extracts from vWAT of 2 and 19-month-old rats (n = 6). (C) Electromobility shift assay (EMSA) using nuclear extracts from (A) and radiolabeled aP2-PPRE probes. Nonradioactive probes were used for competition. (D) EMSA using pooled nuclear extracts (n = 4) from vWAT of obese men aged 23 ± 1 or 40 ± 1 years and radiolabeled aP2-PPRE probes. (E,F) 3T3-L1 adipocytes treated as in Fig. 4 were co-treated with the PPARγ agonists rosiglitazone (10−7 m) or FMOC-l-leucine (10−5 m) (Rocchi et al., 2001). Twenty four hours later, cells were used for LPL mRNA quantification and insulin-stimulated glucose uptake assays. * indicates a significant difference with DMSO-treated cells (P < 0.05). Results in cells are representative of three independent experiments done in triplicate.

Discussion

Several studies have shown that mammalian aging is associated with an increase in fat accretion and insulin resistance, which in humans mostly occurs after the fourth decade of life (Toth & Tchernof, 2000). Here, our findings in aged rodents and humans demonstrate that an imbalance in specific nuclear cofactors, namely that of SRC-1, impacts on PPARγ transactivation of target genes and contributes to the induction of lipid accumulation and insulin resistance.

It is well established that aging is associated with important changes in gene expression in multiple tissues, including heart (Lee et al., 2002), skeletal muscle (Lee et al., 1999), adipose tissue (Higami et al., 2004; Linford et al., 2007) and the brain (Lee et al., 2000). We therefore proposed that aging could induce an imbalance in the nuclear transcriptional milieu that would contribute to such changes. Screening of 28 nuclear cofactors in vWAT of mice, rats and humans, which represents only 10% of currently known cofactors (approximately 300), indicated that aging triggers a conserved and very specific reduction in SRC-1, an important component of the transcriptional complex of PPARγ. This decline occurred in men despite massive obesity and its consequences, suggesting direct effect of aging. Whether this findings apply in the general population remains to be determined. The reduction in SRC-1 is consistent with previous reports in the Japanese quail (Charlier et al., 2006) and in rat brain (Matsumoto, 2002) and spinal cord (Ranson et al., 2003) that showed a similar decrease in SRC-1 mRNA and protein levels upon aging. In addition, as SRC-1 has an important role in modulating steroid hormone signaling through its interactions with nuclear receptors (O’Malley, 2007), loss of SRC-1 in target tissues could likely contribute to the diminution of hormone sensitivity and action observed upon aging, including in macrophages infiltrated in vWAT, as SRC-1 is expressed in these cells (Li et al., 2000).

The increased potency of PPARγ in 12-month-old mice in this study is consistent with the observation that lipid accumulation induced by the PPARγ agonist ciglitazone was more pronounced in middle-aged humans (Schipper et al., 2008). This appears to be independent of changes in PPARγ level itself in mice and men (Fig. 3), as PPARγ was not increased during this period. This contrasts with results in aged rhesus monkeys (Hotta et al., 1999) showing continuous reduction in PPARγ, suggesting species specificity of this response.

Lipid accumulation represents a major determinant of the age-associated augmentation of adipose tissue mass, as a recent study using incorporation of radioactivity into genomic DNA in human populations showed that the number of adipocytes remains constant through adulthood (Spalding et al., 2008), in line with previous studies in rodents [reviewed in (Cartwright et al., 2007)]. Together with our findings, these observations suggest that targeting lipogenesis and lipid uptake in aging adipocytes to reduce fat mass could promote a stronger impact on age-associated fat accretion than blocking adipogenesis. Moreover, this work provides further support to the concept that in overweight, insulin-resistant middle-aged humans and rodents, PPARγ retains its ability to promote lipid accumulation, but not its insulin-sensitizing properties (Picard & Auwerx, 2002), which is in line with a reduced probability of SRC-1 recruitment upon ligand binding (Rocchi et al., 2001). Indeed, because recruitment of SRC-1 has been demonstrated to be a key factor contributing to the PPARγ-induced insulin sensitization, strategic stimulation of SRC-1 expression and docking to PPARγ in old subjects to restore the imbalance in nuclear cofactors could represent a promising therapeutic avenue. It should be noted, however that our findings in visceral fat cannot at present be generalized to other depots (especially subcutaneous adipose tissue), as they are shown to respond differently to aging.

Oxidative stress and ROS production are an established characteristic of aging and are associated with insulin resistance in adipocytes (Furukawa et al., 2004; Lin et al., 2005; Shimoyama et al., 2006; Park et al., 2007; Sautin et al., 2007) as well as in other tissues (Laurent et al., 2008). This is consistent with a report of increased mitochondrial activity and PCG-1 levels in vWAT of FIRKO mice (Katic et al., 2007), which have an increased longevity. In addition, oxidative stress modulates several other pathways and transcription factors (other than PPARγ), such as NFkB (Park et al., 2006), C/EBPα (Pessler-Cohen et al., 2006), and Foxo1 (Subauste & Burant, 2007), which likely contributed to the altered cell biology. Although our model has obvious limitations and should be interpreted with caution, in this study, a low dose of oxidative stress was found to mimic the features of old adipocytes, including reduced SRC-1 docking to PPARγ as seen in vivo (Fig. 6B).

The nuclear cofactor Sirt1, suggested as a node in the regulation of aging (Bordone et al., 2007; Boily et al., 2008; Guarente, 2008), has been shown to modulate adipocyte, liver and muscle energy metabolism (Picard et al., 2004; Rodgers et al., 2005; Baur et al., 2006; Lagouge et al., 2006). In this study, no significant change in vWAT Sirt1 mRNA or protein upon aging was observed (either in rodent or in men). However, because of the broad impact of SRC-1 on transcriptional activation, loss of SRC-1 in H2O2-treated cells and in aged adipocytes could have impaired the activity of many interacting nuclear cofactors such as PGC-1α (Puigserver et al., 1999; Picard et al., 2002), which is itself an important partner of Sirt1 (Rodgers et al., 2008). Current work is ongoing to test whether loss of SRC-1 upon aging alters the interaction and activity of Sirt1/PGC-1 complexes.

Finally, this study adds to the observations that visceral adipose tissue might play a major role in aging and age-associated metabolic diseases. Because the amount of vWAT has been shown to impact on longevity both in mice (Blüher et al., 2003; Chiu et al., 2004; Heikkinen et al., 2009) and rats (Muzumdar et al., 2008), further studies aiming at addressing whether adipose tissue SRC-1 could possibly have a role in the control of lifespan would be of particular interest.

Experimental procedures

Materials and oligonucleotides

Pioglitazone was purchased from Axxora (San Diego, CA, USA) and rosiglitazone from Cayman Chemical Company (Ann Arbor, MI, USA). All chemicals, except when specified, were purchased from Sigma (Oakville, ON, USA). Anti-PPARγ (N-20 and H-100), anti-SRC1 (M-341), anti-mouse or rabbit IgG antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The oligonucleotide sequences used for various experiments described in the manuscript are available upon request.

Animals and treatments

Male C57BL/6 mice (kindly provided by NIA, USA) were cared for and handled in conformance with the Canadian Guide for the Care and Use of Laboratory Animals, and protocols were approved by our institutional animal care committee. Mice used in the aging cohort (n = 25 per age group) were killed by ketamine–xylazine injection 1 week after their arrival. A subgroup of mice were treated by IP injection of vehicle (n = 5) or pioglitazone (n = 5) (10 mg kg−1 day−1) for 7 days and then killed. In all experiments, adipose tissue samples were immediately harvested and snap frozen in liquid nitrogen. Samples of adipose tissue from 4 and 18-month-old male F344/NHsd rats were purchased from NIA Aged Rodent tissue Bank (Bethesda, MD, USA) (n = 4). WAT samples from 2 and 19-month-old Sprague–Dawley rats (n = 6) were kindly donated by the Aging Rat Colony Infrastructure of the Quebec Network for Research on Aging. Triglyceride levels were evaluated according to the Roche Diagnostics Triglyceride assay kit (Montreal, QC, USA).

Obese patients

Omental WAT from obese men (n = 10 per age group) was obtained during ongoing bariatric surgery and also quickly snap frozen in liquid nitrogen and stored at −80 °C. Men of this study were recruited through the bariatric surgery schedule of Laval Hospital. The study included men aged 18.8–65.8 years (body mass index 55.1 ± 11.4 kg m−2, range 40.0–83.0 kg m−2). None of the subjects were diabetic but 10 had hypertension (and were on anti-hypertensive therapy) and four had dyslipidemia and were treated with a statin. Five subjects were smokers. Because subjects with these disorders were evenly distributed across age groups, excluding these subjects from the analyses did not alter the present results. Approbations by the medical ethics committees of Laval Hospital were obtained. All subjects provided written informed consent before their inclusion in the study.

Cell culture and transient transfections

3T3L1 and 293T cells (ATCC) were grown in Dulbecco’s modified Eagle’s medium high glucose with 10% fetal bovine serum supplemented with 4 and 2 mm, respectively, of glutamine in a 5% CO2 environment. 3T3L1 cells were differentiated, 2 days after confluence (D0), in the same medium complemented with 10 μg mL−1 insulin, 0.25 mm 3-isobutyl-1-methyl-xanthine and 1 μm dexamethasone. After 2 days (D2), medium was supplemented with only 10 μg mL−1 insulin and replaced every 2 days until terminal differentiation (D10). For oxidative stress studies, 20 μm H2O2 was added at D4 in the differentiation medium and replaced every other day. For transient transfections, 293T cells pretreated or not with H2O2 were seeded and transfected 6 h later for 12 h with 250 ng of reporter vector and 50 ng of PPARγ expression vector as described (Picard et al., 2004). The next day, the transfection medium was removed and cells were induced with DMSO or 10−6m pioglitazone for 24 h. Luciferase and β-galactosidase activities were then measured (Picard et al., 2002). Each transfection experiment was repeated three times.

Glucose uptake

Glucose transport was measured using the glucose analogue 2-[3H]-2-deoxy-d-glucose (PerkinElmer, Woodbridge, ON, Canada). Cells were glucose deprived for 5 h, and then medium was removed and cells were incubated with 10−5, 10−6, 10−7 m of insulin for 45 min. Then, cells were washed with saline buffer (NaCl 0.9%), and medium containing 3H-2DG was added to the cells for 8 min. After that period, cells were washed three times with saline buffer, lysed in 50 mm NaOH for 30 min, and radioactivity was measured. Data were corrected for total protein quantity assessed in each condition by the Bradford assay.

Lipolysis

Insulin (indicated concentrations) and norepinephrine (1 μmol L−1) were added to the 3T3-L1 culture media. At the end of the incubation, incubation media were frozen until the measurement of glycerol using reagent kits (Sigma) as described (Festuccia et al., 2006).

RNA extraction and RTQ-PCR

RNA was isolated by the acid guanidinium thiocyanate/phenol/chloroform method. 1.5 μg of total RNA was reverse transcribed at 42 °C for 1 h with the SuperScriptTM Reverse Transcriptase (Invitrogen, Burlington, CA, USA), followed by 15 min inactivation at 70 °C. Quantitative PCR was carried out using a Rotor Gene 3000 system (Montreal Biotech, Montreal, QC, Canada). Chemical detection of the PCR products was achieved with SYBR Green Jumpstart Taq ReadyMix without MgCl2 (Sigma). Relative level of gene expression was determined via a standard curve composed of a mix from all the cDNA. Results were normalized to the expression level of the housekeeping gene indicated in the figures. Primer sequences are available on request.

Protein extracts

Tissues were homogenized and incubated in lysis buffer (20 mm Tris pH 7.4, 140 mm NaCl, 1 mm CaCl2, 1 mm MgCl2, 10% glycerol, 1% NP40) for 1 h at 4 °C on a wheel. Lysates were centrifuged 10 min at 16 110 g and protein concentrations were determined with Bradford assay.

Nuclear protein extraction

For nuclear extracts, tissues were homogenized in ice-cold buffer A [10 mm HEPES pH 7.9, 10 mm KCl, 2 mm MgCl2, 0.1 mm ethylenediaminetetraacetic acid (EDTA), 1 mm DL-dithiothreitol (DTT) and diluted 1:1000 Protease Inhibitor Cocktail (PIC)]. The homogenates were centrifuged 1 min at 3660 g at 4 °C to eliminate unbroken tissues. After 20 min on ice, 0.1 volume of 10% NP40 was added and the supernatants were vortexed for 30 s. The supernatants were then centrifuged for 1 min at 11 000 g. The nuclear pellet was suspended in 50 μL in ice-cold buffer B (20 mm HEPES pH 7.9, 420 mm NaCl, 1.5 mm MgCl2, 0.1 mm EDTA, 1 mm DTT, 1:1000 PIC and 25% glycerol), incubated for 30 min at 4 °C with high shaking, and centrifuged for 15 min at 13 200 rpm at 4 °C. The supernatants were collected and protein concentrations were determined with the Bradford assay.

Western blot analysis

Proteins were subjected to SDS–PAGE on polyacrylamide gels and electrophoretically transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were incubated for 1 h at room temperature in blocking buffer [50 mm Tris–HCl pH 7.5, 150 mm NaCl, 0.02% Tween 20, 0.04% NP40 (wash buffer) and 5% milk], and overnight at 4 °C with primary antibodies in wash buffer containing 1% bovine serum albumin (BSA). PVDF membranes were then washed three times in wash buffer for 8 min and incubated for 1 h with either anti-mouse or anti-rabbit immunoglobulin G conjugated to horseradish peroxidase in wash buffer containing 1% BSA at room temperature. After three washes for 8 min, the immunoreactive bands were detected by the chemiluminescent method.

Coimmunoprecipitation assay

Cells were lysed in IP buffer (150 mm NaCl, 1% NP40, 50 mm Tris pH 8, 1:1000 PIC), and an aliquot was taken as input. Cell lysates were precleared with protein A-sepharose beads (GE Healthcare, Baie d’Urfé, QC, Canada) for 1 h at 4 °C and then centrifuged 5 min at 6100 g. Supernatants were immunoprecipitated with adequate antibody overnight at 4 °C, and mouse IgG were used as negative control. Immunoprecipitates were washed once with IP buffer, twice with WB (0.25 m KCl in PBS) and then subjected to SDS–PAGE electrophoresis.

Electro mobility shift assay

Tissue nuclear extracts were incubated in binding buffer [ZHENG; 10 mm Tris pH 7.9, 40 mm KCl, 10% glycerol, 0.05% NP40, 1 mm DTT, 1 μg μL−1 poly (dI:dC)] for 15 min at room temperature. Then, a radiolabeled double-stranded oligonucleotide probe (0.5 μm) was added for 10 min, and complexes were subjected to electrophoresis on a polyacrylamide gel.

Statistical analysis

Data are presented as mean ± SEM. Statistical differences were analyzed by anova and Fisher’s t-test (ad hoc) when appropriate. A P-value < 0.05 was considered significant.

Acknowledgments

We would like to thank Pierrette Gaudreau and Guylaine Ferland for providing us with 19-month-old male Sprague–Dawley rats from the Aging Rat Colony Infrastructure of the Quebec Network for Research on Aging (http://www.rqrv.com). We also thank the bariatric surgeons at Laval Hospital for their help on human adipose tissue sampling, and Yves Deshaies for critical reading of the manuscript. S. Miard was a recipient of FER and CRHL studentships. F Picard holds a CIHR New Investigator Award (MSH-123508). This work was supported by an operating grant from the CIHR (MOP-66967).

Ancillary