Peroxisome proliferator-activated receptor γ (PPARγ) plays an important role in the inhibition of cell growth by promoting cell-cycle arrest, and PPARγ activation induces the expression of p16INK4α (CDKN2A), an important cell-cycle inhibitor that can induce senescence. However, the role of PPARγ in cellular senescence is unknown. Here, we show that PPARγ promotes cellular senescence by inducing p16INK4α expression. We found several indications that PPARγ accelerates cellular senescence, including enhanced senescence-associated (SA)-β-galactosidase staining, increased G1 arrest and delayed cell growth in human fibroblasts. Western blotting studies demonstrated that PPARγ activation can upregulate the expression of p16INK4α. PPARγ can bind to the p16 promoter and induce its transcription, and, after treatment with a selective PPARγ agonist, we observed more-robust expression of p16INK4α in senescent cells than in young cells. In addition, our data indicate that phosphorylation of PPARγ decreased with increased cell passage. Our results provide a possible molecular mechanism underlying the regulation of cellular senescence.
Cellular senescence, also known as replicative senescence, is a process during which cells lose their proliferative potential after a limited number of population doublings (PDs). Senescence is accompanied by a specific set of changes in morphology and in gene expression. For example, senescence-associated β-galactosidase (SA-β-gal) activity appears, the cell cycle is irreversibly arrested at the G1 phase and expression of cyclin-dependent kinase inhibitor (CDKI) increases (Hayflick, 1965; Wong and Riabowol, 1996). In addition, CDK4 and CDK6 can be inhibited by p16INK4α (CDKN2A), an important cell-cycle inhibitor that can induce senescence (Collins and Sedivy, 2003). The p16INK4α protein is thought to be an important biomarker of aging in vivo (Krishnamurthy et al., 2004) because its accumulation can trigger the onset of cellular senescence (Duan et al., 2001).
Caloric restriction (CR) increases the lifespan of many organisms. Some studies indicate that SIRT1 might mediate the effects of CR by repressing peroxisome proliferator-activated receptor γ (PPARγ) activity (Picard et al., 2004). PPARγ is a member of the nuclear receptor superfamily of ligand-activated transcription factors. PPARγ plays an important role the induction of cellular differentiation and the inhibition of cell growth by promoting cell-cycle arrest (Chang and Szabo, 2000; Elstner et al., 1998; Kubota et al., 1998; Mueller et al., 1998; Sarraf et al., 1998; Tontonoz et al., 1997). Guan et al. reported that PPARγ activation induces the expression of p16INK4α and is accompanied by G1 arrest (Guan et al., 1999). Gizard et al. also demonstrated that PPARα, another member of the receptor superfamily to which PPARγ belongs, can bind to the p16 promoter and increase p16INK4α expression (Gizard et al., 2005). These data suggest that PPARγ can accelerate cellular senescence by regulating p16INK4α expression. However, a role for PPARγ in senescence and the mechanisms by which PPARγ regulates p16INK4α remain poorly understood. In the present study, we demonstrate a role for PPARγ in cellular senescence. Moreover, we identify the molecular mechanisms by which PPARγ regulates the expression of p16. Our results indicate that PPARγ dephosphorylation might play an important role in cellular senescence.
Sustained PPARγ overexpression induced premature senescence and irreversible cell-growth arrest
The senescent state of normal human diploid fibroblast cells is characterized by enlarged, flattened cells with prominent lipofuscin granules and irreversible growth arrest (Hayflick and Moorhead, 1961). To determine the effects of PPARγ expression on cellular senescence, young 2BS (PD25) and WI-38 (PD20) cells were transfected with the expression plasmids pcDNA3.1 and pcDNA-PPARγ (termed vector and PPARγ, respectively). After sustained selection with G418, the transformants were obtained. Western blot results (Fig. 1A) confirmed that PPARγ overexpression significantly increased the expression of the wild-type PPARγ in 2BS and WI-38 cells. The transformants were analyzed for the relative senescence markers. Untransfected young, middle-aged and senescent 2BS cells were also analyzed for comparison (see supplementary material Fig. S1). To further analyze the effect of PPARγ activation on senescence, the influence of the selective PPARγ agonists troglitazone (20 μM) and pioglitazone (10 μM) (see supplementary material Fig. S1), which have reduced activity towards PPARα or PPARβ (Willson et al., 1996), on senescence markers was also tested.
PPARγ activation causes a senescence-like cell morphology and increased SA-β-gal activity
The specific senescence-associated marker pH 6.0 optimum β-galactosidase (SA-β-gal) was assayed by X-gal staining. Virtually all treated PPARγ cells (PD42) were strongly stained blue, with gross enlargement and flattened morphology resembling senescent cells (see supplementary material Fig. S1A). However, no significant morphological changes were observed in untreated vectors (PD42), which retained a refractive cytoplasm with long thin projections, and only a few dispersed cells were SA-β-gal-stained (Fig. 1B). For untreated PPARγ and treated vector cells (all at PD42), the positive ratio was comparable with that of middle-aged cells (PD42) (see Fig. 1B and supplementary material Fig. S1A). To determine whether the effect of PPARγ is a general accompaniment to senescence, we extended our study to investigate other normal human diploid fibroblast WI-38 cells. We obtained similar results (Fig. 1B).
The activation of PPARγ leads to growth inhibition
To observe the impact of PPARγ on cell proliferation, the growth curves for vector and PPARγ cells (all at PD42) were compared. The curve of treated PPARγ cells approached that of senescent cells (PD62), showing near complete inhibition of growth (see Fig. 1C and supplementary material Fig. S1B). Untreated PPARγ and treated vector cells had almost the same growth rate or growth potential as middle-aged cells (PD42); by contrast, untreated vector cells had stronger growth potential than untreated PPARγ and treated vector cells (see Fig. 1C and supplementary material Fig. S1B).
PPARγ activation accelerated G1 cell-cycle arrest
To clarify the mechanisms underlying growth-rate inhibition, the cell-cycle profile was analyzed by flow cytometry. Each experiment was performed at least three times and representative data are shown in Fig. 1D. Flow-cytometry assay revealed that PPARγ overexpression or agonist treatment increased the proportion of 2BS cells in the G0-G1 phases (P<0.05 vs vehicle-treated vector cells), and the greatest increase was observed when cells were both transfected with PPARγ and treated with agonist (P<0.001 vs vehicle-treated vector cells) (see Fig. 1D and supplementary material Fig. S1C). Thus, PPARγ activation might influence cell proliferation by influencing cell-cycle progress.
PPARγ activation results in a reduction of the 2BS replicative lifespan
The replicative senescence of normal human diploid fibroblasts is directly correlated to the number of PDs rather than to the growth and metabolic time (Dai and Enders, 2000; Hayflick, 1965). After completing a finite number of divisions, cells enter permanent growth arrest. To determine the effects of PPARγ overexpression on the lifespan of 2BS cells, the number of PDs for PPARγ-transfected and vector-transfected cells from the same batch of young cells was counted. The results revealed that PPARγ cells treated with troglitazone ceased cell division approximately 12 PDs earlier than did vector cells treated with troglitazone. Notably, treated PPARγ cells ceased dividing 17 PDs earlier than did untreated vector cells and 19 PDs earlier than did normal cells. Untreated PPARγ and treated vector cells had a slightly shorter lifespan than did untreated vector cells (Table 1).
|Cells .||Cumulative PDs (means ± s.e.) .|
|Cells .||Cumulative PDs (means ± s.e.) .|
P⩽0.05 vs vehicle-treated vector-transfected cells
Silencing PPARγ delays a senescence-like state and induces lifespan extension
To determine the effects of PPARγ silencing on cellular senescence, young 2BS and WI-38 cells were transfected with the expression plasmids pSilencer 2.1-U6 neo and pSilencer-PPARγ (termed RNAi vector and siPPARγ, respectively). After sustained drug selection, G418-resistant cell clones were obtained. Using western blotting, we detected the expression of PPARγ in vector-transfected and siPPARγ-transfected cells. We found that, in siPPARγ-transfected cells, PPARγ levels decreased significantly relative to RNAi-vector-transfected cells (Fig. 2A). Transfected cells were then analyzed for relative senescence markers.
PPARγ silencing inhibits SA-β-gal activity
No SA-β-gal activity was observed in treated and untreated siPPARγ cells (all at PD51), whereas treated RNAi-vector cells (PD51) were strongly stained blue. Only sporadic SA-β-gal-positive cells were observed in untreated RNAi vectors (PD51) (Fig. 2B). We obtained similar results with WI-38 cells (Fig. 2B).
siPPARγ promotes cell growth
The impact of PPARγ gene-specific silencing on cell growth was evaluated. The curve of treated and untreated siPPARγ cells (all at PD51) advanced quickly, indicative of a strong proliferation potential. Control-treated and untreated RNAi-vector cells (all at PD51) grew slower than did the siPPARγ cells. Moreover, untreated RNAi-vector cells grew slightly faster than did the treated RNAi-vector cells (Fig. 2C).
siPPARγ postponed G1 cell-cycle arrest
Treated and untreated siPPARγ cells (all at PD51) exhibited significantly postponed irreversible growth arrest, whereas treated and untreated RNAi-vector cells (all at PD51) exhibited an increased proportion of 2BS cells in the G0-G1 phases. Furthermore, treated RNAi-vector cells had a slightly higher percentage of cells in the G0-G1 phases than did the corresponding untreated RNAi-vector cells (Fig. 2D).
siPPARγ results in a finite extension of 2BS replicative lifespan
To determine the effects of PPARγ gene-specific silencing on the lifespan of 2BS cells, the number of PDs for siPPARγ-transfected and RNAi-vector-transfected cells from the same batch of young cells was counted. The lifespan of treated siPPARγ cells was about 17 PDs longer than treated RNAi-vector cells and about 13 PDs longer than untreated RNAi-vector cells. Treated RNAi-vector cells had a slightly shorter lifespan than did the untreated RNAi-vector cells. No significant lifespan extension was observed in untreated siPPARγ cells or treated siPPARγ cells (Table 2). Taken together, these findings indicated that PPARγ activation could weaken the replicative capacity, reduce the replicative lifespan, and ultimately promote the onset of senescence of 2BS and WI-38 cells, whereas RNAi-mediated silencing of PPARγ gene exerted the opposite effects.
|Cells .||Cumulative PDs (means ± s.e.) .|
|RNAi vector (vehicle)||59±1|
|RNAi vector (troglitazone)||55±1|
|Cells .||Cumulative PDs (means ± s.e.) .|
|RNAi vector (vehicle)||59±1|
|RNAi vector (troglitazone)||55±1|
P⩽0.05 vs vehicle-treated RNAi-vector-transfected cells
PPARγ-mediated regulation of p16INK4α expression
Previous work reported that PPARγ activation induces the expression of p16INK4α (Guan et al., 1999) and that accumulation of p16INK4α triggers the onset of cellular senescence (Duan et al., 2001). Therefore, we hypothesized that p16 is a PPARγ target gene. First, we determined whether PPARγ overexpression could upregulate p16 protein expression. Western blot analysis revealed that PPARγ overexpression or troglitazone treatment enhanced p16 protein levels in 2BS and WI-38 cells, and the greatest increase was observed when cells were both overexpressed with PPARγ and treated with troglitazone (Fig. 3A). To further verify this finding, we silenced the PPARγ gene and examined the p16INK4α protein levels in 2BS and WI-38 cells. Western blot assays revealed markedly reduced p16INK4α expression in the siPPARγ-transfected cells compared with RNAi-vector-transfected cells. Furthermore, PPARγ-agonist treatment did not increase p16INK4α levels (Fig. 3B). To determine the specificity and efficiency of siPPARγ, we detected the expression of related proteins in RNAi-vector-transfected, negative-transfected [the expression plasmids pSilencer-negative control small interfering RNA (siRNA) with sequences that have no homology to any known mammalian gene, termed as negative] and siPPARγ-transfected cells by western blotting. Our findings show that, in siPPARγ-transfected cells, the PPARγ level reduced markedly compared with RNAi-vector- and negative-transfected cells. However, the levels of PPARα and PPARβ in siPPARγ-transfected cells were similar to RNAi-vector- and negative-transfected cells (see supplementary material Fig. S2A). These data identified the specificity and efficiency of siPPARγ. Western blot assays revealed a decreased p16INK4α expression in siPPARγ-transfected cells compared with negative-transfected cells (see supplementary material Fig. S2B). The results were consistent with Fig. 3B, which confirmed further that the knockdown of PPARγ caused a reduction in p16INK4α expression. These results suggest that PPARγ activation upregulates the expression of p16INK4α.
PPARγ binds to the PPRE-containing region of the endogenous p16 promoter
We examined in vivo DNA-binding activity of PPARγ to the p16 promoter. A peroxisome proliferator response element (PPRE) has been identified in the p16 promoter sequence (Fig. 4A) and is located at position –1023 relative to the translation initiation site (Gizard et al., 2005). We found that PPARγ bound the PPRE both in young cells and in senescent cells, and both young and senescent cells had similar amplification without troglitazone stimulation (Fig. 4B). However, the binding of PPARγ was more robust in senescent cells than in young cells after stimulation with troglitazone or pioglitazone (Fig. 4B,D). Real-time PCR can increase the accuracy and precision of chromatin immunoprecipitation (ChIP) measurements, allowing for the detection of changes of less than twofold (Johnson and Bresnick, 2002). Quantitative real-time PCR analyses revealed that, after treatment with PPARγ agonists, there was a significant increase in the binding activity of PPARγ in both young and senescent 2BS cells, but the magnitude of change was much greater in senescent 2BS cells (Fig. 4C,E). Following the addition of GW9662, a PPARγ antagonist (Bendixen et al., 2001; Wright et al., 2000), the binding of PPARγ was weaker in senescent cells than in young cells (Fig. 4F). These results indicate that PPARγ binds to the p16 promoter in vivo and imply that the role of PPARγ in p16 transcriptional regulation is ligand dependent. As negative controls, PCR amplification using primers covering a region located immediately downstream of the –1023 site (PPRE) failed to yield a significant signal (Fig. 4G); neither the irrelevant antibody control (β-actin) nor the negative control (no antibody sample) had amplification products. Taken together, these results demonstrated the specificity of PPARγ immunoprecipitation and PCR amplification.
Ligand-activated PPARγ induces transcription of p16
Subsequently, we analyzed the effects of PPARγ on the expression of the p16 reporter gene, in which the luciferase gene is driven by a p16 promoter. 2BS cells were transfected with various combinations of plasmids that expressed pcDNA3.1, pcDNA-PPARγ, pSilencer 2.1-U6 neo or pSilencer-PPARγ together with wild-type and mutant p16-Luc (luciferase). Cells were then treated with troglitazone (20 μM), pioglitazone (10 μM) or GW9662 (10 μM). Reporter activity was enhanced by treatment with PPARγ agonists, and this activity was higher in cell lysates that contained PPARγ from senescent cells (about 13-fold) than in those that contained PPARγ from young cells (about fourfold) (Fig. 5A). In WI-38 cells, reporter activity was also higher in cell lysates that contained PPARγ from senescent cells (about tenfold) than in those that contained PPARγ from young cells (about sixfold) (Fig. 5A). Mutation of p16-Luc or siPPARγ resulted in a significant, albeit incomplete, reduction of p16 promoter activation (Fig. 5A,B). Identical results were obtained in cells treated with GW9662 (10 μM) (Fig. 5A). Taken together, these data suggest that ligand-activated PPARγ induced transcriptional activity of the p16 promoter, and this induction was stronger in senescent cells than in young cells.
Silencing the p16 gene by RNA interference
In order to further verify that PPARγ could promote senescence via p16INK4α, we evaluated the role of p16 in mediating the senescence effects of PPARγ activation in 2BS cells. Young 2BS cells (PD25) were transfected with the expression plasmids pSilencer 2.1-U6 neo and pSilencer-p16 (termed RNAi vector and sip16, respectively). After sustained selection with G418, the transformants (all at PD51) were obtained. Western blot results revealed that, in sip16-transfected cells, p16 levels decreased significantly relative to RNAi-vector-transfected cells (Fig. 6A). It is noteworthy that, similar to what occurred in siPPARγ-transfected cells, the SA-β-gal-positive staining ratio was drastically decreased in treated and untreated sip16-transfected cells (Fig. 6B, Fig. 2B). Interestingly, sip16-transfected cells proliferated at a much higher rate than did RNAi-vector-transfected cells. Moreover, sip16-transfected cells did not respond to troglitazone treatment (Fig. 6C). Accordingly, the ratios of cells in G0-G1 phases were reduced in sip16-transfected cells and were not affected by troglitazone treatment (Fig. 6D). Altogether, these data indicate that the silencing of p16INK4α caused resistance to PPARγ-agonist-induced senescence.
PPARγ expression in young and senescent cells
Because PPARγ DNA-binding activity and transcriptional activity increased as cells approached the end of their replicative lifespan in culture, we further evaluated the expression patterns of PPARγ during successive passages. We therefore measured endogenous mRNA and protein expression levels of PPARγ and p16INK4α. Using reverse transcription-polymerase chain reaction (RT-PCR), PPARγ mRNA, p16 mRNA and GAPDH mRNA levels were analyzed in young (PD25), middle-aged (PD42) and senescent (PD62) 2BS cells. Our findings demonstrate that the levels of PPARγ mRNA were similar in young, middle-aged and senescent cells, whereas the p16 mRNA levels were increased in senescent cells relative to other groups (Fig. 7A). In addition, western blot analysis was performed on the 2BS and WI-38 cells. Protein levels of PPARγ did not change appreciably as cells aged, whereas p16INK4α protein levels increased as cells aged in culture (Fig. 7B).
Phosphorylation of PPARγ represses its transactivating function
The above findings indicate that PPARγ levels remain constant in young and senescent cells, but they do not explain why the transcriptional activity of PPARγ increased in senescent cells. Phosphorylation modifications reportedly affect PPARγ transcriptional activity (Hu et al., 1996; Lazennec et al., 2000; Han et al., 2000); therefore, we analyzed the phosphorylation state of PPARγ in young and senescent 2BS and WI-38 cells. We found that the phosphorylation level of PPARγ decreased as cells aged (Fig. 8A). To evaluate the effect of PPARγ phosphorylation on its transcriptional function, we created a mutant of pcDNA-PPARγ, in which the serine at position 84, the potential phosphorylation site, was changed to alanine [pcDNA-PPARγ (S84A)]. We then analyzed the effects of PPARγ phosphorylation on the expression of the p16 reporter gene in which the luciferase gene is driven by the p16 promoter. 2BS cells were transfected with various combinations of plasmids that expressed pcDNA3.1(–), pcDNA-PPARγ (WT) or mutant pcDNA-PPARγ(S84A) together with wild-type p16-Luc. Cells were then treated with troglitazone (20 μM) or pioglitazone (10 μM). The results revealed that the untreated mutant PPARγ (S84A) exhibited greater transactivation (about 4.5-fold) than untreated wild-type PPARγ (about twofold). Furthermore, the treated S84A mutant activated luciferase expression to a significantly greater extent (about 6.5-fold) than treated wild-type PPARγ (about threefold). Our finding indicated that the S84A mutant showed a high level of transactivation, comparable with that of wild-type PPARγ (Fig. 8B). Taken together, these data suggest that phosphorylation of PPARγ represses its transcriptional activity.
Senescence is the state or process of aging at the cellular level, and is thought to relate to age-related diseases and tumorigenesis (Campisi, 2001; Campisi, 2005; Weinstein and Ciszek, 2002). Senescence may be described as accumulated DNA damage, a limited number of cell divisions and a decreased ability to remove free radicals (Weinstein and Ciszek, 2002). CR has been proposed to extend lifespan and slow aging. Some studies suggest that SIRT1, the primary molecule mediating the effects of CR, functions by repressing PPARγ activity (Picard et al., 2004). PPARγ might play an important role in cellular senescence; however, the function of PPARγ in the progression of cellular senescence has never been described. In the present study, we demonstrated that PPARγ activation promoted cellular senescence (Figs 1, 2), an effect attributed to the induction of p16. p16INK4α is an important cell-cycle inhibitor, and its accumulation triggers the onset of cellular senescence (Duan et al., 2001). An upregulation of p16INK4α expression by PPARγ agonists provides a mechanism for senescence being repressed by CR and also a model (Fig. 9) that regulates cellular senescence.
The nuclear receptor superfamily of PPARs regulates the transcription of numerous target genes after dimerizing with the retinoid X receptor (RXR) and binding to PPRE (a specific DNA-binding site) (Mangelsdorf et al., 1995). PPREs usually consist of a direct repeat of the hexanucleotide AGGTCA sequence, separated by one or two nucleotides (DR1 or DR2) (Michalik et al., 2004). Recently, one report indicated that PPARα can bind to PPRE in the p16 promoter and increase p16INK4α expression (Gizard et al., 2005). Because the nuclear receptor superfamily of PPARs can bind to the same DNA-binding site, PPRE, we hypothesize that PPARγ can also bind to PPRE in the p16 promoter and thereby regulate p16INK4α expression. Western blot analysis demonstrated that PPARγ overexpression and PPARγ-agonists treatment enhanced endogenous expression of p16INK4α, whereas RNA interference of PPARγ inhibited the expression of p16INK4α (Fig. 3). These results suggested that the activation of PPARγ induced the expression of p16INK4α in 2BS and WI-38 cells.
In the search for the molecular mechanisms involved in the regulation by PPARγ of p16, PPARγ activation was found to upregulate p16 promoter activity, thus identifying for the first time a molecular mechanism by which PPARγ directly interferes with senescence progression. We observed that both DNA-binding activity and transcriptional activity of PPARγ were increased several fold in senescent cells, and the increased activity is ligand-dependent (Figs 4, 5).
2BS cells were previously isolated from human fetal lung fibroblast tissue and have been fully characterized (Tang et al., 1994). The maximum population doubling of human diploid fibroblasts is limited in culture, so they are widely used as a model of cellular senescence. Previous reports have suggested that fibroblast cellular senescence occurs as a consequence of a `genetic program' (Goldstein, 1990). This program has been partially characterized by gene expression patterns during the progression of successive passages. Sasaki et al. reported that PPARγ mRNA levels do not significantly differ by sex or age in lung tissue (Sasaki et al., 2002). However, Ye et al. found that the expression of PPARγ at both the mRNA and protein level in adipose tissue of older rats was dramatically decreased and, likewise, that the expression of PPARγ mRNA in omental adipose tissue of elderly men was significantly decreased (Ye et al., 2006). These conflicting data regarding PPARγ expression patterns might be due to the various roles of PPARγ in different tissues during the progression of successive passages. Because PPARγ activation increases during aging, fat-deposit size reportedly declines and lipids are redistributed to muscle, bone marrow and other tissues (Moerman et al., 2004; Kirkland et al., 2002). Our results indicate that the mRNA and protein level of PPARγ in young cells is similar to that in senescent cells (Fig. 7). We found that PPARγ activity increased in senescent cells, although the expression of PPARγ did not appear to be upregulated.
Research has demonstrated that PPARγ activity is modulated by phosphorylation, which inhibits PPARγ transcriptional activity (Hu et al., 1996; Lazennec et al., 2000; Han et al., 2000). To determine why the transcriptional activity of PPARγ increases as cells age, we analyzed the phosphorylation state of PPARγ in both young and senescent 2BS and WI-38 cells. As expected, the levels of PPARγ phosphorylation decreased as cells aged (Fig. 8A). We also found that mutant PPARγ (mutation of serine 84 to alanine, which abolished phosphorylation of PPARγ) has higher transcriptional activity than wild-type PPARγ (Fig. 8B). These results indicate that phosphorylation of PPARγ represses the transcriptional activity of PPARγ itself. Three possible mechanisms may explain the modulation of PPARγ transactivation activity: (a) phosphorylation of PPARγ might decrease its affinity for its cognate ligand; (b) the phosphorylation of PPARγ could influence its interactions with co-repressors and/or coactivators of transcription; and (c) phosphorylation of PPARγ might promote its degradation by the ubiquitin-proteasome system in response to ligand activation. According to the first mechanism, hyperphosphorylated PPARγ should exhibit decreased ligand binding. Our results indicated that the DNA-binding activity and transcriptional activity of PPARγ increased upon agonist treatment and these changes were much greater in senescent cells than that in young cells (Figs 4, 5). This indicated that, in young cells, hyperphosphorylation of PPARγ decreases its transcriptional activity. However, hyperphosphorylation of PPARγ does not alter its DNA-binding activity, because the DNA-binding activity of PPARγ is similar in untreated young cells and untreated senescent cells (Fig. 4). The second pattern is currently under investigation in our laboratory. Finally, because PPARγ levels remained constant during successive passages (Fig. 7), we hypothesize that the third mechanism is not involved in PPAR-regulated senescence in 2BS and WI-38 cells.
p16INK4α is an important cell-cycle inhibitor that can induce senescence and repress tumor cell growth (Collins and Sedivy, 2003). Its loss or inactivation is correlated with cell immortality (Ruas and Peters, 1998). Although p16INK4α plays an important role in cellular senescence and tumorigenesis, its transcriptional control is poorly understood. Here, we report, for the first time to our knowledge, the molecular mechanism by which PPARγ upregulates the expression of p16INK4α in 2BS and WI-38 cells. Our results, which show that p16INK4α is required for PPARγ-agonist-induced senescence (Fig. 6), verify this assertion. However, recent studies have proposed p16 regulatory pathways that are distinct from those described here. Ohtani et al. proposed a model in which the upregulation of p16INK4α depended on the accumulation of Ets1 and the absence of interference by Id1 during senescence (Ohtani et al., 2001). Zheng et al. found that Id1 regulated p16INK4α levels through interactions with E47 (Zheng et al., 2004). Wang et al. reported that a 24-kDa protein might inhibit the expression of p16INK4α by interacting with the INK4α transcription silence element (ITSE) (Wang et al., 2001). These findings do not contradict the present conclusions. It is clear that the p16 promoter is subject to multiple levels of control (Hara et al., 1996; Jacobs et al., 1999); therefore, p16 regulation cannot be explained by a single isolated pathway.
In summary, in senescent cells, dephosphorylation of PPARγ increases its transcriptional activity. The increased PPARγ activity might be one reason for the elevated p16INK4α expression in senescent 2BS and WI-38 cells, which in turn contributes to the onset of cellular senescence (Fig. 9). Various polyunsaturated fatty acids, the major dietary constituents, are specific ligands for PPARγ (Forman et al., 1995; Kahn et al., 2000). Much research shows that fatty acids can influence the transcriptional activity of PPARγ. In our study, PPARγ accelerated senescence by inducing p16INK4α expression in a ligand-dependent manner; therefore, PPARγ might underlie a key switch between exterior factors (such as diet) and interior factors (such as the p16 gene). Finally, we demonstrated a gene-environment effect induced by PPARγ-agonist administration, which results in senescence-like growth arrest with p16INK4α expression. The effect was more marked in senescent cells than in young cells. Our study might offer an opportunity to investigate gene-environment interactions associated with cellular senescence in health and disease.
Materials and Methods
Antibodies, reagents and plasmids
Antibodies against PPARγ (SC-7273), PPARα (SC-9000), PPARβ (SC-1983) and β-actin (SC-1616) were purchased from Santa Cruz Biotechnology. Anti-phospho-PPARγ (05-816) and anti-p16 (MS-887-P1) antibodies were purchased from Upstate. Pioglitazone and GW9662 were purchased from Cayman. Troglitazone was purchased from Calbiochem or Cayman. The PPARγ expression plasmid cloned in pcDNA3.1 was constructed as described previously (Fu et al., 2001). p16 cDNA in pBluescript was a kind gift from David Beach (Howard Hughes Medical Institute, Cold Spring Harbor, NY). The full-length p16 cDNA (800 bp) was placed into the expression vector pcDNA3.1 in both orientations. A p16-promoter fragment that contained 1040 bp was obtained via PCR from the 3070-bp pGL2-Basic vector, which was generously provided by Gorden Peters (Imperial Cancer Research Fund Laboratories, London, UK).
The siRNA was designed as reported previously. The sequence of the sense strand of PPARγ siRNA was 5′-GCCCTTCACTACTGTTGAC-3′ (Kelly et al., 2004); p16 siRNA was 5′-AGAACCAGAGAGGCTCTGA-3′ (Zhou et al., 2004); and negative control siRNA was 5′-TTCTCCGAACGTGTCACGT-3′ (Genechem). The hairpin-siRNA template oligonucleotides were chemically synthesized with 5′-phosphate, 3′-hydroxyl, and two base overhangs on each strand. Then the template oligonucleotides were inserted into the BamHI and HindIII sites of the pSilencer 2.1-U6 neo vector.
Cell culture and transfection
Human embryonic lung diploid fibroblast 2BS cells (obtained from the National Institute of Biological Products, Beijing, China) were previously isolated from female fetal lung fibroblast tissue and have been fully characterized (Tang et al., 1994). The current expected lifespan is approximately PD70. 2BS cells are considered to be young at PD30 or below and to be fully senescent at PD55 or above. Human embryonic lung diploid fibroblast WI-38 cells (ATCC number: CCL75) were obtained from the Chinese Academy of Sciences (Shanghai, China). The current expected lifespan is approximately PD55. WI-38 cells are considered to be young at PD25 or below and to be fully senescent at PD50 or above. Cells were maintained in Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal bovine serum (FBS, Hyclone), 100 units/ml penicillin, and 100 μg/ml streptomycin at 37°C in 5% CO2.
Young cells were grown to 80-90% confluence. Expression plasmids were transfected with Lipofectin reagent (Life Technologies) according to the manufacturer's instructions. Pools of stable transformants were obtained by sustained selection of 300 μg/ml G418 (Life Technologies). PDs were calculated with the formula PD=log(n2/n1)/log2, where n1 is the number of cells seeded and n2 is the number of cells recovered (Shay and Wright, 1989).
Cells were detached and seeded into 96-well plates, with 2000 cells per well. After overnight incubation, cells were treated with 20 μM troglitazone or 10 μM pioglitazone diluted in DMSO (Sigma). All cells received identical volumes of DMSO and were exposed to each drug for 6 days; medium and drug were changed every 48 hours. At the indicated times, cells were stained with 20 μl 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT, 10 mg/ml in PBS; Sigma) for 3 hours and then dissolved with DMSO. The optical density at 570 nm was determined.
Cell-cycle analysis and synchronization
When cells reached 70-80% confluence, they were washed with PBS, detached with 0.25% trypsin and fixed with 75% ethanol overnight. After treatment with 1 mg/ml RNase A (Sigma) at 37°C for 30 minutes, cells were resuspended in 0.5 ml of PBS and stained with propidium iodide in the dark for 30 minutes; DNA contents were measured by fluorescence-activated cell sorting on a FACScan flow cytometry system (BD Biosciences). The data were analyzed using CellFiT software.
For synchronization, 2BS cells were rendered quiescent by serum deprivation for 48 hours and then stimulated to re-enter the cell cycle by the addition of serum to a final concentration of 10%. G1-phase cells were harvested at 8 hours after serum stimulation.
Cells were washed twice in PBS, fixed for 3-5 minutes at room temperature in 3% formaldehyde and washed with PBS again. Then cells were incubated overnight at 37°C without CO2 in a freshly prepared staining solution [1 mg/ml 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal), 40 mM citric acid/sodium phosphate, pH 6.0, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, 2 mM MgCl2] (Dimri et al., 1995). At least 200 cells were counted in randomly chosen fields from each culture well.
Cells were lysed in modified radioimmune precipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.25% Na-deoxycholate, 1% NP-40, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 0.2 mM sodium orthovanadate, 1 mM NaF). Protein concentration of each sample was determined by BCA protein assay reagent (Pierce); 100-150 μg of protein was electrophoresed on 15% SDS-polyacrylamide gel and transferred to nitrocellulose membrane (Millipore). The membrane was blocked and then incubated with the primary antibody in 5% non-fat dry milk in TBST (10 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.05% Tween 20) overnight at 4°C. After washing, the blots were incubated with secondary antibody conjugated to horseradish peroxidase (Amersham Biosciences) at 1:40,000 in TBST for 1 hour at room temperature. Proteins were visualized with Chemiluminescent Substrate (Pierce) according to manufacturer's instructions.
Substitution mutations were generated using the QuikChange mutagenesis kit (Stratagene) according to the manufacturer's instructions. Synthetic oligonucleotides that contained the desired bases were used in the mutagenesis. The sequence of mutations for nucleotides at positions –1023 and –1022 (GG-CC) of the 5′-flanking region of the p16 promoter was 5′-GTGTGAACCAGACAGGACAGTATTT-3′ (GG-CC mutation underlined) (Gizard et al., 2005); mutant PPARγ with the substitution of serine-84 to alanine was 5′-GTGGAGCCTGCAGCTCCACCTTATTATTC-3′ (TCT-GCT mutation) (Adams et al., 1997). DNA that incorporated the desired mutations was transformed into XL1-Blue supercompetent cells. Plasmid DNA was prepared and the presence of the mutations was confirmed by sequencing.
ChIPs were performed using the Chromatin Immunoprecipitation Assay kit (Upstate) according to manufacturer's instructions. For each experimental condition, 1×106 cells were used. At 48 hours before harvesting, cells were pre-treated with 20 μM troglitazone, 10 μM pioglitazone, 10 μM GW9662, or 1% DMSO (vehicle) as control. Cells were sonicated and lysates immunoprecipitated using the indicated antibodies. To amplify PPRE regions of the p16 promoter, the following sequences of the primers were used: p16 PPREs, 5′-GCACTCATATTCCCTTCCCCCCT-3′; p16 PPREa, 5′-GGAAGGACGGACTCCATTCTCAAAG-3′. Control p16 element was located immediately downstream of PPRE. The sequences of the primers used were as follows: p16 controls, 5′-GAAGCTGGTCTTTGGATCACTGTGC-3′; p16 controla, 5′-GACGGGGGAGAATTCTGCCTGT-3′ (Gizard et al., 2005). All primers were synthesized at Sunbio Biotechnology (Beijing, China).
Real-time PCR and data analysis
Two-step real-time PCR was performed with 5 μl of DNA and 400 nM primers diluted to a final volume of 50 μl in SYBR Green Master Mix (Applied Biosystems). Accumulation of fluorescent products was monitored by real-time PCR using an Applied Biosystem 7300 real-time PCR system (Applied Biosystems). A melting curve was generated to ensure that a single peak of the predicted Tm was produced and no primer-dimer complexes were present. Single amplicon generation was verified by agarose gel electrophoresis. No PCR products were observed in the absence of template. Sequence Detector software (version 1.3.1) was used for data analysis and relative fold induction was determined by the comparative threshold cycle (CT). Fold-enrichments were determined by the method described in the Applied Biosystems User Bulletin and data analysis followed the methodology described in a recent report (Frank et al., 2001). Fold differences were calculated by correcting for each signal concentration with the concentration of input signal for each sample [(signal concentration)/(input concentration)]. Real-time RT-PCR data figures were generated using Microsoft Excel (Microsoft Corporation, USA).
2BS cells were plated in six-well culture plates in triplicate for each condition at an initial concentration of 2×105 cells/well. pGL2 luciferase reporter constructs driven by the indicated 1 μg p16 wild-type (p16 WT) or mutant (p16 mutant) promoter fragments were co-transfected with 3 μg pcDNA3.1 (vector), wild-type pcDNA-PPARγ (PPARγ or WT), mutant pcDNA-PPARγ (S84A), pSilencer 2.1-U6 neo (RNAi vector) or pSilencer-PPARγ (siPPARγ) expression plasmid. The amount of plasmid in the transfection mixture was equalized to 4 μg by adding pcDNA3.1 vector. Renilla luciferase reporter plasmid pRL-CMV (10 ng) was also co-transfected into each well as an internal control. After 24 hours, cells were treated with 20 μM troglitazone, 10 μM pioglitazone or 10 μM GW9662. Luciferase activity was assessed with a dual-luciferase reporter assay system (Promega) according to manufacturer's instructions after 48 hours of drug treatment. The enzyme activity was normalized for efficiency of transfection on the basis of Renilla luciferase activity levels and reported as relative light units (RLU). All reporter assays were performed in triplicate on at least two individual experiments and standard errors are denoted by bars in the figures.
Total RNA was isolated from 2BS cells by RNeasy kit (QIAGEN). After denaturing the total RNA at 70°C for 10 minutes, cDNA was synthesized with oligo-dT primer and reverse transcriptase (Invitrogen). PCR amplification was performed using specific primers for PPARγ as follows: PPARγs, 5′-GAGCCCAAGTTTGAGTTTGC-3′; PPARγa, 5′-TGGAAGAAGGGAAATGTTGG-3′. The sequences of the primers used for p16 were: p16s, 5′-CCCAACGCACCGAATAGT-3′; p16a, 5′-ATCTAAGTTTCCCGAGGTT-3′. The sequences of the primers used for GAPDH were: GAPDHs, 5′-CGAGTCAACGGATTTGGTGGTAT-3′; GAPDHa, 5′-AGCCTTCTCCATGGTGAAGAC-3′. PCR products were loaded onto an agarose gel and stained with ethidium bromide.
The data are reported as mean ± s.d. of the indicated number of experiments. Values were assessed by pairwise (one-way analysis of variance, ANOVA). In all cases, P⩽0.05 and P⩽0.01 was considered significant.
We thank Wengong Wang for helpful discussions. This work was supported by grants from the National Basic Research Programs of China (No. 2007CB507400) and the National Natural Science Foundation of China (No. 30671064).