PAI-1–regulated extracellular proteolysis governs senescence and survival in Klotho mice

Significance

Plasminogen activator inhibitor-1 (PAI-1) is an essential mediator of cellular senescence in vitro and is one of the biochemical fingerprints of senescence in vivo. Klotho-deficient (kl/kl) mice display a complex phenotype reminiscent of human aging and exhibit age-dependent increases in PAI-1 in tissues and in plasma. Thus, we hypothesized that PAI-1 contributes to the aging-like phenotype of kl/kl mice. We observed that either genetic deficiency or pharmacological inhibition of PAI-1 in kl/kl mice was associated with reduced evidence of senescence, preserved organ structure and function, and a fourfold increase in median lifespan. These findings indicate that PAI-1 is a critical mediator of senescence in vivo and defines a novel target for the prevention and treatment of age-related disorders in man.

Abstract

Cellular senescence restricts the proliferative capacity of cells and is accompanied by the production of several proteins, collectively termed the “senescence-messaging secretome” (SMS). As senescent cells accumulate in tissue, local effects of the SMS have been hypothesized to disrupt tissue regenerative capacity. Klotho functions as an aging-suppressor gene, and Klotho-deficient (kl/kl) mice exhibit an accelerated aging-like phenotype that includes a truncated lifespan, arteriosclerosis, and emphysema. Because plasminogen activator inhibitor-1 (PAI-1), a serine protease inhibitor (SERPIN), is elevated in kl/kl mice and is a critical determinant of replicative senescence in vitro, we hypothesized that a reduction in extracellular proteolytic activity contributes to the accelerated aging-like phenotype of kl/kl mice. Here we show that PAI-1 deficiency retards the development of senescence and protects organ structure and function while prolonging the lifespan of kl/kl mice. These findings indicate that a SERPIN-regulated cell-nonautonomous proteolytic cascade is a critical determinant of senescence in vivo.

Advanced age contributes to the development of frailty and disease in humans, but the fundamental mechanisms that drive physiological aging are incompletely understood (1, 2). Cellular senescence, which halts the proliferative capacity of cells, is associated with the manifestation of the senescence-associated secretory phenotype (3) and the production and secretion of a distinct set of proteins (2, 4), including insulin-like growth factor-binding proteins (IGFBPs), interleukins (ILs), transforming growth factor type β (TGF-β), and plasminogen activator inhibitor-1 (PAI-1) (5), collectively termed the “senescence-messaging secretome” (SMS) (6). In addition to this pattern of protein production and secretion, senescent cells display a distinctive morphology, and can be identified by increased expression of senescence-associated β-galactosidase (7). The tumor suppressor and proapoptotic protein p53 plays a central role in inducing replicative senescence by regulating the transcription of genes involved in cell cycle arrest and apoptosis, including the cyclin-dependent kinase inhibitors p16^Ink4a and p21 (8). Senescence can be triggered by a number of factors, including DNA damage (9), oncogene induction (10), and oxidative stress (11). Although the relationship between cellular senescence and physiological aging remains an area of intense investigation, it is becoming increasingly evident that the two processes are fundamentally linked. Senescent cells accumulate in aging tissues and have been hypothesized to disrupt tissue regeneration, which may reflect cell-nonautonomous effects of the SMS (6).

In the last decade, numerous examples of genetically modified mice with phenotypes reminiscent of human aging have been described and investigated. These include the BubR1^H/H progeroid (12) and Klotho-deficient (kl/kl) mice (13). BubR1^H/H progeroid mice exhibit an age-dependent increase in the expression levels of PAI-1 in numerous locations, including white adipose tissue, skeletal muscle, and the eye (12). BubR1^H/H mice have a shortened average lifespan (24 wk) and develop various aging-like phenotypic abnormalities, including sarcopenia, cataracts, arterial stiffening, and impaired wound healing (14). Klotho functions as an aging-suppressor protein by impeding the development of senescence in vitro and in vivo through inhibition of the Wnt (15), TGF-β (16), and IGF1 signaling pathways (17). Thus, kl/kl mice exhibit a rapidly progressive phenotype after weaning that includes a truncated lifespan (8–12 wk), renal sclerosis, arteriosclerosis, emphysema, and osteoporosis (13). Membrane-bound Klotho forms a heterodimer with fibroblast growth factor (FGF) receptors (FGFRs) generating a high-affinity receptor for FGF23. Signals transduced by FGF23 via the Klotho–FGFR complex inhibit 1,25-(OH)₂ vitamin D₃ and parathyroid hormone synthesis and promote renal phosphate excretion. kl/kl mice exhibit a remarkable increase in plasma levels of FGF23, as well as significant increases in serum levels of calcium, phosphate, vitamin D₃, and creatinine (13). Interestingly, kl/kl mice also have an age-dependent increase in plasma PAI-1 levels as well as increased PAI-1 expression in a number of tissues including kidney, aorta, and heart (18). Because PAI-1 is necessary and sufficient to induce replicative senescence in vitro downstream of p53 (19) and is markedly increased in kl/kl mice, we hypothesized that PAI-1 is a critical determinant of the phenotypic abnormalities developed by kl/kl mice. Here we examined the impact of PAI-1 on senescence and physiological aging in vivo by breeding kl/kl and PAI-1–deficient (pai-1^−/−) mice to generate kl/kl mice with partial (kl/klpai-1^+/−) or complete (kl/klpai-1^−/−) PAI-1 deficiency.

Results

PAI-1 Deficiency Prolongs the Survival of kl/kl Mice.

We systematically monitored the effect of PAI-1 deficiency on the growth, vigor, and survival of littermate kl/kl (n = 26), kl/klpai-1^+/− (n = 39), kl/klpai-1^−/− (n = 25), and WT (n = 16) mice (Fig. 1) all in the same genetic background (75% C57BL/6J, 25% C3J). We observed that either partial or complete absence of PAI-1 prolonged the survival of kl/kl mice. Log-rank analysis indicated that the survival curves for the WT, kl/kl, kl/klpai-1^+/−, and kl/klpai-1^−/− mice differed significantly (P < 0.0001). The median survival of kl/kl mice was 58 d, and this value increased with PAI-1 deficiency: 2.8-fold (163 d) in kl/klpai-1^+/− mice and 4.2-fold (246 d) in kl/klpai-1^−/− mice. Whereas all of the kl/kl mice died within 120 d, 65% of kl/klpai-1^+/− and 82% of kl/klpai-1^−/− mice were alive beyond 120 d. Although median survival indicates a dose–response relationship between genotype and mortality (P = 0.0002 by log-rank test for trend), the mean lifespan increased similarly in kl/klpai-1^+/− (250 ± 169 d) and kl/klpai-1^−/− (254 ± 123 d) mice (mean ± SD), corresponding to 4.2- and 4.5-fold increases, respectively. Furthermore, we achieved a similar prolongation of lifespan in kl/kl mice (n = 11) (Fig. 1B) by the administration of an orally active small-molecule PAI-1 antagonist, TM5441, whose pharmacokinetic properties, toxicity, and specificity have been described recently (20). In contrast with the inconsistent effects based on sex of a low phosphate diet on survival in kl/kl mice (21), both males and females appear to benefit from complete PAI-1 deficiency. However, survival of kl/klpai-1^+/− females (n = 19) was not as long as that of males (n = 20) of the same genotype (median survival 121 d vs. 315 d, mean lifespan 208 ± 151 d vs. 285 ± 182, respectively; P = 0.16). Nevertheless, kl/klpai-1^+/− females do live longer than kl/kl females (n = 14) (median survival 121 d vs. 58 d, mean lifespan 208 ± 151 d vs. 57 ± 18, respectively; P = 0.0004) (Fig. S1). This improvement in survival was also associated with evidence of increased overall vigor and health, as kl/klpai-1^−/− mice exhibited near-normal weight gain over time (Fig. 1 A and C) and spontaneous physical activity (Fig. 1D).

Fig. 1.

Effects of PAI-1 deficiency in Klotho mice. (A) Size and appearance of 8-wk-old littermate mice. (B) Survival curve. Log-rank analysis showed that the survival curves for the WT, kl/kl, kl/klpai-1^+/−, and kl/klpai-1^−/− mice differed significantly (P < 0.0001). (C) Bodyweight measurements starting from 3 wk of age. *P = 0.002, **P = 0.0001, and ***P = 0.0003. (D) Open field physical activity measurements recorded as distance traveled in 20 min in age-matched animals. *P = 0.027, **P = 0.018, and ***P = 0.036. Data are plotted as mean ± SD.

PAI-1 Deficiency Normalizes Senescence and Telomere Length in kl/kl Mice.

To characterize the extent of senescence in kl/kl mice and how PAI-1 deficiency affects it, plasma levels of the SMS factors IGFBP-3 and IL-6, and telomere length were measured in liver, aorta, tail, and kidney tissue. We observed that kl/kl mice had increased levels of IGFBP-3 (Fig. 2A) compared with WT mice (P = 0.02). With partial (P = 0.03 vs. kl/kl mice) or complete PAI-1 deficiency (P = 0.02 vs. kl/kl mice), IGFBP-3 levels did not significantly differ from those seen in WT mice. Similarly, we found that compared with levels in WT mice, kl/kl mice had a 13-fold increase (P = 0.02) in plasma levels of proinflammatory cytokine IL-6, which functions in the acquisition of the senescent phenotype in vitro (Fig. 2B). Compared with kl/kl mice, IL-6 levels were reduced by 79% in kl/klpai-1^+/− (P = 0.03) and 83% in kl/klpai-1^−/− (P = 0.0001) mice. These observations suggest that the elevated PAI-1 levels in kl/kl mice are a dominant factor in contributing to increases in plasma IGFBP-3 and IL-6 and further augment the senescent phenotype in these mice.

Fig. 2.

Effect of PAI-1 deficiency on plasma levels of SMS factors and ATLR in various tissues from age-matched littermate Klotho mice. (A) Determination of IGFBP-3 levels in plasma samples (n = 6 per group). *P = 0.02, **P = 0.03, and ***P = 0.02. (B) Quantitation of circulating IL-6 levels (n = 6 to 14). *P = 0.02, **P = 0.03, and ***P = 0.0001. (C) Quantitation of ATLR by qRT-PCR in liver, aorta, tail, and kidney tissue (n = 6 to 14). *P < 0.05 compared with WT, ^#P < 0.05 compared with kl/kl, and ^$P < 0.05 compared with kl/klpai-1^+/−. Data are plotted as mean ± SD.

Although elevated plasma levels of SMS components may reflect systemic senescence, they are nonspecific in nature and do not provide precise identification of which tissues are actually senescent. To address this limitation, telomere length was determined in several different tissues. Liver, aorta, and tail tissue samples from kl/kl mice displayed moderate but significant decreases in the average telomere length ratio (ATLR), whereas renal tissue had 16% longer ATLRs compared with those of WT animals (Fig. 2C). In contrast, ATLRs of liver and tail tissues from kl/klpai-1^+/− and liver, aorta, and tail, tissues from kl/klpai-1^−/− mice were significantly longer than those of kl/kl mice. These findings indicate that PAI-1 deficiency provides partial protection of telomere integrity in numerous tissues. The anomalous preservation of telomere length in renal tissue from kl/kl mice may reflect the lack of turnover in renal cells and the early induction of replicative senescence in the kidneys of kl/kl mice (22).

Because the kidneys are one of the most severely compromised organs in kl/kl mice, we also examined kidneys for biomarkers of senescence, including p16^Ink4a and p21. We detected strong immunostaining for p16^Ink4a localized in the nuclei of tubules in kidney tissues from kl/kl mice (Fig. 3), but not for p21. In contrast, kidney sections from kl/klpai-1^−/− mice had only minimal evidence of p16^Ink4a accumulation (Fig. 3). Quantitative real-time PCR (qRT-PCR) analysis showed that the relative expression of p16^Ink4a in kidneys from kl/kl mice is 3.2-fold higher than that in WT mice (P = 0.001) (Fig. 3E). In kl/klpai-1^+/− and kl/klpai-1^−/− mice, p16^Ink4a expression was reduced by 80% (P = 0.04) and 92% (P = 0.0001) compared with the kl/kl mice, respectively, and in kl/klpai-1^−/− mice by 78% compared with the levels seen in WT animals (P = 0.0001).

Fig. 3.

Effect of PAI-1 deficiency on renal p16^Ink4a expression in Klotho-deficient mice. (A and C) Control immunostaining of kidney sections from kl/kl and kl/klpai-1^−/− mice, respectively, in the absence of the anti-p16^Ink4a antibody. (B and D) Immunodetection of p16^Ink4a-positive cells in kidney sections from kl/kl and kl/klpai-1^−/− mice, respectively. (E) Quantitation of relative p16^Ink4a expression in the kidney. qRT-PCR analysis was performed on total RNA samples purified from WT (n = 8), kl/kl (n = 6), kl/klpai-1^+/− (n = 4), and kl/klpai-1^−/− (n = 8) kidneys. *P = 0.001, **P = 0.0001, and ***P = 0.04. Data are plotted as mean ± SD. (Magnification: A–D, 60×.)

Effect of PAI-1 Deficiency on the Biochemical Hallmarks of kl/kl Mice.

In an effort to explain the effects of PAI-1 deficiency on the kl/kl phenotype, we measured plasma levels of factors that are biochemical hallmarks of kl/kl mice, including FGF23, vitamin D₃, calcium, phosphate, creatinine, and PAI-1 (Table 1). As expected, kl/kl mice displayed a more than 1,200-fold increase in FGF23 levels [2.8 × 10⁵ ± 1.4 × 10⁵ pg/mL vs. 225 ± 65 pg/mL in WT mice (P = 0.004)], reflecting the loss of functioning receptors for FGF23. Both kl/klpai-1^+/− and kl/klpai-1^−/− mice exhibited a nearly 98% reduction in plasma FGF23 levels compared with kl/kl mice [3.4 × 10³ ± 2.1 × 10³ pg/mL (P < 0.0001) and 3.9 × 10³ ± 0.9 × 10³ pg/mL (P = 0.0001), respectively]. Similarly, vitamin D₃ levels were reduced in kl/klpai-1^+/− (P = 0.032) and kl/klpai-1^−/− (P = 0.0003) mice compared with the levels in kl/kl mice. Interestingly, partial or complete PAI-1 deficiency had only a marginal impact on serum levels of calcium, phosphate, and creatinine in kl/kl mice. As expected, PAI-1 antigen was not detectable in plasma from kl/klpai-1^−/− mice, and levels in kl/klpai-1^+/− animals were reduced by nearly 50% compared with those from kl/kl mice (P < 0.05). In addition, PAI-1 expression levels were reduced in tissues from kl/klpai-1^+/− mice compared with those of kl/kl mice (Fig. S2).

Table 1.

Effects of PAI-1 deficiency on blood levels of the biochemical hallmarks of klotho mice

Circulating factors assayed	WT	kl/kl	kl/klpai-1+/−	kl/klpai-1−/−
	n = 4 to 5	n = 6 to 9	n = 5 to 6	n = 6 to 10
Phosphate, mg/dL	7.7 ± 1.6	14.0 ± 2.8*	10.8 ± 2.6*	12.6 ± 2.7*
Calcium, mg/dL	7.9 ± 0.6	10.4 ± 0.9*	10.1 ± 1.2*	10.6 ± 0.8*
Creatinine, mg/dL	0.22 ± 0.11	0.31 ± 0.33	0.21 ± 0.11	0.18 ± 0.06
PAI-1, ng/mL	1.8 ± 0.4	45.2 ± 4.8*	24.3 ± 2.7*	0.0
FGF23, pg/mL	225 ± 65	295,657 ± 139,709*	3,924 ± 1,316*,#	3,795 ± 870*,#
Vitamin D3, pg/mL	ND	1359 ± 145	698 ± 419#	314 ± 244#
Aldosterone, pg/mL	102 ± 20	160 ± 55*	ND	223 ± 96*
ALP, U/mL	20 ± 7	21 ± 6	17 ± 7	19 ± 5

ND, not determined.

^*P < 0.05 compared with WT.

^#P < 0.05 compared with kl/kl.

PAI-1 Deficiency Preserves Organ Structure in kl/kl Mice.

As reported previously, kl/kl mice develop emphysema that is characterized by a progressive, age-dependent enlargement of air spaces and associated alveolar destruction (Fig. 4) (23). Histological analysis of lung tissues from kl/klpai-1^+/− and kl/klpai-1^−/− (Fig. 4) mice showed that PAI-1 deficiency primarily prevents alveolar enlargement. Consistent with the preservation of pulmonary structural integrity, pulmonary function was also maintained with PAI-1 deficiency. We found that kl/kl mice had a 40% decrease in PaO₂ levels (P = 0.018) in arterial blood samples. Arterial oxygenation normalized with partial (P = 0.05) and complete (P = 0.02) PAI-1 deficiency in kl/kl mice (Fig. 4E). These results indicate that PAI-1 is an important contributor to the emphysematous changes observed in kl/kl mice.

Fig. 4.

Effects of PAI-1 deficiency on lung morphology and function. Masson’s trichrome staining of lung sections in (A) WT, (B) kl/kl, (C) kl/klpai-1^+/−, and (D) kl/klpai-1^−/− mice. (Scale bars: 150 μm.) (E) Partial pressure of oxygen (PaO₂) measurements in arterial blood (n = 4 for each group). *P = 0.018, **P = 0.05, and ***P = 0.02. Data are plotted as mean ± SD.

Finally, we analyzed mice for evidence of ectopic calcification, which has been reported to increase with age in kl/kl mice. Whereas the age-matched WT littermate mice had no detectable calcification, we observed prominent calcium deposits in the kidneys of kl/kl mice (Fig. 5 A and B) (P = 0.002). However, analysis of kidneys from kl/klpai-1^+/− (Fig. 5C) and kl/klpai-1^−/− (Fig. 5D) mice showed that ectopic calcification areas were significantly reduced by 41% (P = 0.03) and 96% (P < 0.0001), respectively. To test the effect of PAI-1 deficiency on the impaired osteogenic signaling observed in kl/kl mice (24, 25), we measured the serum levels of aldosterone and alkaline phosphatase (ALP) activity (Table 1). Although aldosterone levels were significantly higher in kl/kl mice than that of WT animals, it was not altered significantly in kl/klpai-1^−/− mice. Furthermore, we did not observe any difference in ALP activity among the mice groups studied here. This observation indicates that partial or complete loss of PAI-1 expression protects against age-induced ectopic calcification in kl/kl mice without altering serum levels of phosphate, calcium, ALP and aldosterone.

Fig. 5.

Effects of PAI-1 deficiency on ectopic calcification characterized by von Kossa staining. Ectopic calcification analysis was performed on kidney (n = 5 for each group) sections. Although samples from (A) WT mice showed no detectable calcification, kidney sections from (B) kl/kl mice had remarkable calcified deposits. (C) Partial (kl/klpai-1^+/−) or (D) complete PAI-1 deficiency (kl/klpai-1^−/−), on the other hand, significantly decreased calcification levels. (Scale bars: 40 μm.) (E) Quantitative analyses using Image-Pro-6.3 software showed that the percentages of calcified areas in kidneys were reduced by 41% in kl/klpai-1^+/− and 96% in kl/klpai-1^−/− mice compared with kl/kl mice. *P = 0.002, **P = 0.03, and ***P = 0.0001. Data are plotted as mean ± SD.

Discussion

PAI-1 is expressed in senescent cells and tissues, and is recognized as a primary component of the SMS. Most mammalian models of aging investigated thus far exhibit evidence of increased PAI-1 expression (12, 14). Furthermore, genetic or therapeutic interventions that prolong survival or reduce senescence in tissues in vivo are coincidentally associated with reductions in PAI-1. To our knowledge, this in vivo study is the first to investigate systematically the role of PAI-1 not only in the development of senescence, but also in the aging-like pathology of a mammal. The results from this study suggest that the onset of physiological aging can be delayed by modulating PAI-1, which subsequently prevents the nuclear accumulation of senescence marker p16^Ink4a and maintains the structural and functional integrity of vital organs. Furthermore, the ability of a small-molecule PAI-1 antagonist to augment survival to a similar extent in kl/kl mice indicates that the observed effects are likely cell-nonautonomous. The protective effects of partial or complete PAI-1 deficiency are in agreement with previous work from our laboratory indicating that transgenic overexpression of PAI-1 is sufficient to induce several aging-like phenotypic abnormalities, including age-dependent spontaneous coronary thrombosis, systemic amyloid deposition, and hair loss (26).

The normalization of FGF23 and vitamin D₃ levels in partial or complete PAI-1 knockout models strongly indicates that PAI-1 directly influences FGF23 signaling in kl/kl mice. Recent observations demonstrate that FGF23 is highly sensitive to cleavage by the serine protease furin, which is rapidly inhibited by PAI-1 (27, 28). In addition to FGF23, kl/kl mice have augmented expression of other furin substrates, including IGF1, TGF-β, MMP2, and MMP9 (29, 30). PAI-1 is known to regulate the proteolytic activation and/or clearance of many of these proteins. The precise identity and function of other proteases that are inhibited by PAI-1 and that contribute to the Klotho phenotype merits further investigation.

Because FGF23 signaling is impaired in kl/kl mice, the negative feedback inhibition on vitamin D₃ synthesis is dysfunctional. High vitamin D₃ and phosphate levels in kl/kl mice likely stimulate FGF23 synthesis in bone continuously, whereas the elevated PAI-1 levels reduce the proteolytic clearance of FGF23. Together, these combined effects on production, signaling, and metabolism likely explain the >1,200-fold increase in plasma FGF23 levels observed in kl/kl mice. Recent reports indicate that dietary deficiency of phosphate, zinc, and calcium significantly improves the lifespan of klotho mice (21, 25, 31). Although we did not detect any significant changes in the levels of phosphate and calcium in kl/klpai-1^−/− mice, we observed that PAI-1 deficiency significantly prolongs the survival of kl/kl mice indicating that, in addition to the mineral homeostasis, PAI-1–regulated extracellular proteolysis strongly influences the aging phenotype.

Our findings also reveal a previously unrecognized role of PAI-1 in modulating the effects of FGF23. The increased plasma levels of PAI-1 in kl/kl mice are not surprising, but deserve some mechanistic explanation. Numerous factors likely contribute to the Klotho phenotype, and prominent on that list are the effects of TGF-β (16), Wnt (15), and IGF1 (17, 32). Importantly, all three of these factors can directly induce PAI-1 expression. As the phenotype matures, PAI-1 production is likely further augmented by the effects of progressive hypoxemia, the induction of the p53 pathway, aldosterone excess (25), and elevated levels of other components of the SMS, including IL-6, interferons, TGF-β, and IGFBPs.

Discovery of the Klotho gene has shed light into the molecular mechanisms of tissue calcification. The elevated phosphate and calcium levels in kl/kl mice certainly contribute to the pattern of ectopic calcification (33, 34). Recently, hyperaldosteronism was reported to be the major inducer of the osteog enic signaling, which was partially reversed by spironolactone without normalizing plasma levels of vitamin D₃, FGF23, calcium, and phosphate (24, 25). Furthermore, Lim et al. reported that FGF23 treatment reduced the aldosterone-induced expression of osteogenic factors in vitro (35). In addition, PAI-1 deficiency drastically reduced FGF23 levels and calcification without altering aldosterone and ALP levels in kl/kl mice (Table 1). These findings suggest that PAI-1, which is regulated by aldosterone (36), plays an unexpected but pivotal role in tissue calcification. The reduced plasma levels of vitamin D₃ and prevention of tissue calcification in kl/klpai-1^−/− mice require the restoration of Klotho-independent FGF23 signaling. The present study strongly indicates that PAI-1 plays a direct role in the regulation of FGF23 signaling in kl/kl mice.

The reduced IGFBP-3 levels in kl/kl mice provide an important mechanistic insight into the protective effects of PAI-1 deficiency on the Klotho phenotype. It was recently demonstrated that the PAI-1–IGFBP-3 cascade promotes stress-induced senescence in human breast fibroblasts (37). Expression levels of IGFBP-3 are increased in response to senescence-inducing stimuli. However, the proteolytic metabolism of IGFBP-3 by tissue-type plasminogen activator (t-PA) prevents the induction of cellular senescence in vitro. Because PAI-1 is the main physiological inhibitor of t-PA, IGFBP-3 appears to be a critical downstream target of PAI-1–induced senescence. These findings provide a possible mechanistic explanation for the prosenescent effects of increased PAI-1 in kl/kl mice and suggest a role for an extracellular cascade of secreted proteins in the regulation of cellular senescence and physiological aging.

It was recently shown that membrane-bound Klotho prevents the retinoic-acid-inducible gene-I–induced expression of IL-6 and -8 both in vitro and in vivo, suggesting that the antiaging function of Klotho also includes the suppression of inflammation (38). It is interesting to note that with PAI-1 deficiency, levels of both IGFBP-3 and IL-6 were normalized in kl/kl mice. This confirms and extends the recent observations by López-Andrés et al. that modulating the activity of a key member of the SMS can also normalize the levels of other SMS factors (39). This indicates that IGFBP-3 and IL-6 are downstream from PAI-1 in the senescence pathway.

In human fibroblasts, telomere shortening initiates senescence through a pathway that involves ataxia telangiectasia mutated (ATM), p53, and p21^CIP1, but not p16^Ink4a (9). Thus, the presence of p16^Ink4a-positive but not p21-positive cells in kidneys of kl/kl mice suggests that the pathway to senescence and accelerated aging in the renal tissue of kl/kl mice is likely mediated by p16^Ink4a and occurs independently of the p53 pathway. These findings in the kidney of preserved telomere length, together with augmented p16^Ink4a expression, suggest that renal cells undergo senescence without the requisite cellular division to shorten telomeres (22).

The results presented here indicate that PAI-1 is a critical contributor to, and not merely a marker of, senescence in vivo (12), and that novel therapies targeting PAI-1 or other components of the SMS (40) may prevent senescence and age-related pathologic changes in humans, including arteriosclerosis and emphysema. The development of small-molecule, orally active, selective PAI-1 antagonists, such as TM5441 and others (41), will allow these hypotheses to be tested prospectively.

Materials and Methods

Animals and Animal Care.

The original genetic background of kl/kl mice, a kind gift from Makoto Kuro-o (Jichi Medial University, Shimotsuke, Tochigi, Japan), was composed of C57BL/6J and C3H/J (13). PAI-1–deficient mice were obtained from The Jackson Laboratory (background strain C57BL/6J; strain name B6.129S2-Serpine1tm1Mlg/J; stock no. 002507). Klotho mice and PAI-1–deficient mice were crossed to generate the heterozygous dihybrid mice (KL/kl-pai-1^+/−). All mice used in this study were littermates generated by breeding these KL/kl-pai-1^+/− mice and thus of the same mixed background. They were housed in a temperature-controlled environment with a daily 14:10 h light–dark cycle and had unlimited access to food (standard rodent chow diet; Harlan Teklad) and water. All experimental protocols were approved by the Institutional Animal Care and Use Committee of Northwestern University.

PAI-1 Inhibitor TM5441.

TM5441 was derived from the original hit compound TM5007 (42) and the lead compound TM5275 (43) through an extensive structure-activity relationship (44). We have recently described the pharmacokinetic properties and toxicity, and specificity of TM5441 (20). TM5441 was administered at 100 mg/kg⋅d⁻¹ mixed in the chow.

Histological Methods.

Tissues harvested from mice were fixed in formalin for 24–48 h and then processed overnight and embedded in paraffin. Tissues sectioned at 6-μm thickness were stained with Masson’s trichrome to visualize tissue morphology. Ectopic calcification was analyzed by von Kossa staining of kidneys (n = 5 for each study group). The extent of calcification was quantified in multiple kidney sections from each mouse by using Image-Pro-6.3 image-processing software and the results presented as the percent calcified areas in kidneys (Fig. 3). To detect the senescent cells, tissue sections were immunostained overnight at 4 °C with a primary antibody to p16^Ink4a antigen (Cell Applications), and then the antigen was visualized with an HRP-conjugated secondary antibody (goat anti-mouse IgG at 1:500 dilution; Santa Cruz Biotechnology, Inc.). One-Step AEC Solution (BioGenex) was used as the substrate for antigen detection.

Quantitation of Factors in Plasma and Serum.

Plasma levels of IGFBP-3, IL-6, and FGF23 were measured using the ELISA kits Quantikine Immunoassay kit from R&D Systems (catalog no. MGB300), Becton-Dickinson (catalog no. 555220), and Immutopics (catalog no. 60-6300), respectively, by following the manufacturers’ suggested protocols. Vitamin D₃ levels in plasma were measured using the mouse 1,25-(OH)₂VitD3 HVD3 ELISA kit from NovaTein Biosciences (catalog no. NB-E20523). Calcium (catalog no. 0150-250), creatinine (catalog no. 0430-120), and phosphate (catalog no. 0830-125) levels were determined by using kits from Stanbio Laboratories. Plasma PAI-1 levels were measured with the Murine PAI-1 Total Antigen Assay from Molecular Innovations. A colorimetric assay kit and an ELISA kit from Abcam were used to measure serum levels of ALP and aldosterone (catalog nos. ab83369 and ab136933, respectively).

Measuring Partial Oxygen Pressure in Arterial Blood.

PaO₂ levels in the arterial blood were measured in mice anesthetized with pentobarbital (75 mg/kg body weight, administered i.p.). After adequate anesthesia was achieved, we performed a tracheostomy and sutured a 20-gauge angiocatheter into the trachea. Mice were then placed on a small rodent ventilator (MiniVent; Harvard Apparatus) with the following settings: a respiratory rate of 150 breaths per minute, a tidal volume of 8 mL/kg body weight, and an FiO₂ of 0.21 (room air) as described previously (45). The animals were ventilated for 15 min before a thoracotomy was performed and then 200 μL arterial blood were collected into a heparinized syringe via direct puncture of the left ventricle. The arterial blood sample was processed for gas analysis using the Stat Profile pHOx blood gas analyzer (Nova Biomedical).

Quantitation of ATLR.

Genomic DNA isolated from various tissues was used to measure telomere length by qRT-PCR as previously described with minor modification (46, 47). Briefly, telomere repeats were amplified using specially designed primers, which were then compared with the amplification of a single-copy gene, the 36B4 gene (acidic ribosomal phosphoprotein PO), to determine the ATLR. One hundred nanograms of genomic DNA template were added to each 20 μL reaction containing forward and reverse primers (250 nM each for telomere primers, and 500 nM each for the 36B4 primers), SsoAdvanced SYBR Green Supermix (Bio-Rad USA), and nuclease-free water. A serially diluted standard curve of 100 ng to 3.125 ng per well of template DNA from a WT mouse sample was included on each plate for both the telomere and the 36B4 reactions to facilitate ATLR calculation. Critical values were converted to nanogram values according to the standard curves, and nanogram values of the telomere (T) reaction were divided by the nanogram values of the 36B4 (S) reaction to yield the ATLR. The primer sequences for the telomere portion were as follows: 5′-CGG TTT GTT TGG GTT TGG GTT TGG GTT TGG GTT TGG GTT-3′ and 5′-GGC TTG CCT TAC CCT TAC CCT TAC CCT TAC CCT TAC CCT-3′. The primer sequences for the 36B4 single copy gene portion were as follows: 5′-ACT GGT CTA GGA CCC GAG AAG-3′ and 5′-TCA ATG GTG CCT CTG GAG ATT-3′. Cycling conditions for both primer sets (run in the same plate) were 95 °C for 10 min, 30 cycles of 95 °C for 15 s, and 55 °C for 1 min for annealing and extension.

qRT-PCR.

Tissues harvested from subject mice were snap-frozen in liquid nitrogen. Excess tissue was removed under a dissecting microscope. RNA was isolated using the Qiagen RNAeasy Mini Kit (Qiagen) using the manufacturer’s protocol. cDNA was generated from the RNA using the qScript cDNA Supermix (Quanta Biosciences). qRT-PCR was performed using the SsoAdvanced SYBR Green Supermix (Bio-Rad USA). Forward and reverse primers used for p16^Ink4a expression were, respectively: 5′-AGGGCCGTGTGCATGACGTG-3′ and 5′-GCACCGGGCGGGAGAAGGTA-3′; for PAI-1 expression, respectively: 5′-ACGCCTGGTGCTGGTGAATGC-3′ and 5′-ACGGTGCTGCCATCAGACTTGTG-3′; and for GAPDH expression, respectively: 5′-ATGTTCCAGTATGACTCCACTCACG-3′ and 5′-GAAGACACCAGTAGACTCCACGACA-3′ (IDT, Inc.).

Behavioral Characterization.

To test the effect of PAI-1 deficiency on the level of physical activity of kl/kl mice, the open field test for spontaneous horizontal activity was performed on age-matched animals at the Northwestern University Mouse Behavioral Phenotyping Core Laboratory as described previously (13).

Statistical Analyses.

We present the averaged values obtained for each study as mean ± SD. Statistical significance was assigned to a comparison by using an unpaired, two-tailed Student t test when comparing two groups. Statistical significance for the survival of groups was established by the log-rank analysis of Kaplan–Meier plots using GraphPad Prism 4 software. P < 0.05 was considered statistically significant.