Endothelial progeria induces adipose tissue senescence and impairs insulin sensitivity through senescence associated secretory phenotype

Abstract

Vascular senescence is thought to play a crucial role in an ageing-associated decline of organ functions; however, whether vascular senescence is causally implicated in age-related disease remains unclear. Here we show that endothelial cell (EC) senescence induces metabolic disorders through the senescence-associated secretory phenotype. Senescence-messaging secretomes from senescent ECs induced a senescence-like state and reduced insulin receptor substrate-1 in adipocytes, which thereby impaired insulin signaling. We generated EC-specific progeroid mice that overexpressed the dominant negative form of telomeric repeat-binding factor 2 under the control of the Tie2 promoter. EC-specific progeria impaired systemic metabolic health in mice in association with adipose tissue dysfunction even while consuming normal chow. Notably, shared circulation with EC-specific progeroid mice by parabiosis sufficiently transmitted the metabolic disorders into wild-type recipient mice. Our data provides direct evidence that EC senescence impairs systemic metabolic health, and thus establishes EC senescence as a bona fide risk for age-related metabolic disease.

Fig. 1. Replicative senescent EC impairs adipocyte function through the SASP.

a SA-β-Gal staining of 3T3-L1 adipocytes treated with the conditioned medium (CM) derived from proliferating young or replicative senescent EC. b CDK inhibitor expression in 3T3-L1 adipocytes treated with the control medium, or CM derived from proliferating young or replicative senescent EC (n = 8 biologically independent samples for control medium group; n = 6 biologically independent samples for EC-CM groups). c SASP factor expression in 3T3-L1 adipocytes treated with the control medium, or CM derived from proliferating young or replicative senescent EC (n = 4 biologically independent samples each). d Immunoblotting for the insulin signal pathway in 3T3-L1 adipocytes treated with the control medium (control), CM derived from proliferating young EC (young), or CM derived from replicative senescent EC (aged) in the presence or absence of insulin treatment. e Quantitative analysis for IRβ and Akt activation in response to insulin (n = 3 biologically independent samples each). f Immunoblotting for IRS-1 and IRS-2 in 3T3-L1 adipocytes treated with the control medium, or indicated EC-CM. Quantitative analysis was also shown (n = 3 biologically independent samples each). Non-repeated ANOVA with post hoc analysis of Fisher’s PLSD was used for difference evaluation between the groups (b, c, e, f). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Bars: 100 μm. Source data are provided as a Source Data file.

Fig. 2. Premature senescent EC impairs adipocyte function through the SASP.

a SA-β-Gal staining in 3T3-L1 adipocytes treated with CM derived from control ECs (EC/GFP-CM) or premature senescent ECs (EC/TERF2DN-CM). b SPiDER-β-Gal staining in 3T3-L1 adipocytes treated with CM derived from control ECs (EC/GFP-CM) or premature senescent ECs (EC/TERF2DN-CM). c CDK inhibitor expression in 3T3-L1 adipocytes treated with the control medium, EC/GFP-CM or EC/TERF2DN-CM (n = 6 biologically independent samples each). d Immunoblotting for the insulin signal pathway and IRS-1 in insulin-stimulated 3T3-L1 adipocytes treated with the control medium, EC/GFP-CM or EC/TERF2DN-CM. Non-repeated ANOVA with post hoc analysis of Fisher’s PLSD was used for statistical analysis. Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, and ****P < 0.0001. Bars: 100 μm. Source data are provided as a Source Data file.

Fig. 3. Senescent EC impairs adipocyte functions by inducing oxidative stress.

a Superoxide was detected (as red fluorescence) in 3T3-L1 adipocytes treated with the control medium, or CM derived from proliferating young or replicative senescent EC. b Superoxide was detected in 3T3-L1 adipocytes treated with the control medium, or CM derived from control ECs (EC/GFP-CM) or premature senescent ECs (EC/TERF2DN-CM). c CDK inhibitor expression in 3T3-L1 adipocytes treated with the indicated EC-CM in the presence or absence of βNMN or NAC (n = 4 biologically independent samples each). A two-tailed Student’s t test was used for difference evaluation between the two groups. Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, and ^#not significant. d SA-β-Gal staining in 3T3-L1 adipocytes treated with the indicated EC-CM in the presence or absence of βNMN or NAC. e Immunoblotting for the insulin signal pathway and IRS-1 in insulin-stimulated 3T3-L1 adipocytes treated with the indicated EC-CM in the presence or absence of βNMN or NAC. Bars: 100 μm. Source data are provided as a Source Data file.

Fig. 4. Generation of EC-specific progeroid mice.

a CDK inhibitor expression in EC and non-EC isolated from the lung and WAT of WT or Tie2-TERF2DN-Tg mice (line #16) (n = 6 biologically independent samples for WT; n = 5 biologically independent samples for Tg). b SASP factor expression in EC and non-EC isolated from the WAT of WT or Tie2-TERF2DN-Tg mice (line #16) (n = 6 biologically independent samples each). c SA-β-Gal staining in EC and non-EC isolated from the lung of WT or Tie2-TERF2DN-Tg mice (line #16). SA-β-Gal-positive cells were counted (n = 5 biologically independent samples each). d EC and non-EC were isolated from the lung of young (7W) or aged (70W) mice. Cells were stained with SA-β-Gal, and staining-positive cells were counted (n = 6 biologically independent samples each). e Principalcomponent analysis for gene expression profiles assessed by DNA microarray in ECs isolated from the lung of young WT (7W), aged WT (70W), or Tie2-TERF2DN-Tg mice (20-week old). A two-tailed Student’s t test was used for difference evaluation between the two groups (a–d). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, and ****P < 0.0001. Bars: 100 μm. Source data are provided as a Source Data file.

Fig. 5. EC-specific progeria impairs systemic metabolic health.

a Body weight of WT or Tie2-TERF2DN-Tg (line #16) mice fed normal chow (NC) (n = 7 biologically independent animals for WT; n = 6 biologically independent animals for Tg). b, c Body weight (b) and body fat ratio (c) in WT or Tie2-TERF2DN-Tg mice fed NC at the age of 20 weeks old (n = 7 biologically independent animals for WT; n = 5 biologically independent animals for Tg). d Insulin tolerance test (ITT) in WT or Tie2-TERF2DN-Tg mice fed NC at the age of 20 weeks old (n = 7 biologically independent animals for WT; n = 6 biologically independent animals for Tg). e CDK inhibitor expression in mature adipocytes (MA) isolated from the WAT of WT or Tie2-TERF2DN-Tg mice (n = 6 biologically independent samples for WT; n = 5 biologically independent samples for Tg). f SA-β-Gal staining of subcutaneous (sWAT) or visceral epididymal WAT (eWAT) isolated from WT or Tie2-TERF2DN-Tg mice. g Immunoblotting for IRS-1 in the WAT isolated from WT or Tie2-TERF2DN-Tg mice (n = 7 biologically independent samples each). h Immunoblotting for the insulin signal pathway in the WAT, liver, and skeletal muscle of NC-fed WT or Tie2-TERF2DN-Tg mice with or without insulin treatment. A two-tailed Student’s t test was used for difference evaluation between the two groups (b–e, g). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Source data are provided as a Source Data file.

Fig. 6. EC-specific progeroid mice are metabolically healthy at the age of 10-week old.

a ITT in WT or Tie2-TRF2DN-Tg mice fed normal chow at the age of 10-week old (n = 7 biologically independent animals for WT; n = 5 biologically independent animals for Tg). b CDK inhibitor expression in stromal vascular fraction (SVF) or mature adipocytes (MA) isolated from WAT of WT or Tie2-TRF2DN-Tg mice at the age of 10-week old (n = 10 biologically independent samples each). c CDK inhibitor expression in ECs isolated from lung (n = 15 biologically independent samples for WT; n = 10 biologically independent samples for Tg), and WAT (n = 13 biologically independent samples for WT; n = 10 biologically independent samples for Tg) of WT or Tie2-TRF2DN-Tg mice at the age of 10-week old. d SA-β-Gal staining in EC and non-EC isolated from the lung of WT or Tie2-TRF2DN-Tg mice at the age of 10-week old. Bars: 100 μm. SA-β-Gal-positive cells were counted (n = 6 independent fields each). A two-tailed Student’s t test was used for difference evaluation between the two groups (b–d). Data are presented as mean ± s.e. *P < 0.05, ***P < 0.001, and ****P < 0.0001. Source data are provided as a Source Data file.

Fig. 7. EC senescence impairs metabolic health by enhancing adipocyte oxidative stress.

a Immunostaining for the 8-OHdG in the WAT of WT or Tie2-TERF2DN-Tg mice at the age of 20 weeks old. b ITT in WT or Tie2-TERF2DN-Tg mice with or without NAC treatment for 10 weeks beginning at 10-week old (n = 6 biologically independent animals for WT; n = 5 biologically independent animals for Tg). c CDK inhibitor expression in the WAT of control (vehicle-treated, n = 9 biologically independent samples for WT; n = 8 biologically independent samples for Tg) or NAC-treated (n = 5 biologically independent samples each) WT or Tie2-TERF2DN-Tg mice. Non-repeated ANOVA with post hoc analysis of Fisher’s PLSD was used for difference evaluation between the groups. d Immunostaining for the 8-OHdG in the WAT of WT or Tie2-TERF2DN-Tg mice treated with either vehicle or NAC. e Inflammatory gene expression in the WAT of control (vehicle-treated, n = 9 biologically independent samples for WT; n = 8 biologically independent samples for Tg) or NAC-treated (n = 5 biologically independent samples each) WT or Tie2-TERF2DN-Tg mice. A two-tailed Student’s t test was used for difference evaluation between the two groups (c, e). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ^#not significant. Bars: 100 μm. Source data are provided as a Source Data file.

Fig. 8. EC senescence impairs metabolic health through the SASP.

a ITT in recipient WT or Tie2-TERF2DN-Tg mice harboring BM of WT or Tg mice (n = 20 biologically independent animals for WT>WT group; n = 9 biologically independent animals for Tg>WT group; n = 9 biologically independent animals for WT>Tg group). b CDK inhibitor expression in the WAT isolated from recipient WT or Tie2-TERF2DN-Tg mice harboring BM of WT or Tg mice (n = 13 biologically independent samples for WT>WT group; n = 5 biologically independent samples for Tg>WT group; n = 6 biologically independent samples for WT>Tg group). c ITT in the recipient WT or Tie2-TERF2DN-Tg mice whose circulation was shared with WT (recipient–donor; WT–WT, n = 9 biologically independent animals) or Tie2-TERF2DN-Tg mice (n = 9 biologically independent animals for WT-Tg group; n = 6 biologically independent animals for Tg–Tg group). d ITT in the donor WT or Tie2-TERF2DN-Tg mice whose circulation was shared with WT (WT–WT, n = 12 biologically independent animals) or Tie2-TERF2DN-Tg mice (n = 7 biologically independent animals for WT-Tg group; n = 6 biologically independent animals for Tg–Tg group). Non-repeated ANOVA with post hoc analysis of Fisher’s PLSD was used for difference evaluation between the groups (a–d). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Source data are provided as a Source Data file.

Fig. 9. IL-1a orchestrates senescence-associated cytokine networks in EC.

a IL-1a concentration in culture medium derived from young or replicative senescent EC in the presence or absence of inflammasome activators, such as monosodium urate crystal (MSU, 200 or 400 μg/mL) and adenosine triphosphate (ATP, 2.5 or 5 mM) (n = 4 biologically independent samples for young; n = 3 biologically independent samples for senescent EC). b FACS analysis of cell-surface IL-1a in ECs isolated from the lung of young (7W) or aged (70W) mice (n = 6 biologically independent samples each). A two-tailed Student’s t test was used for difference evaluation between the two groups. c SASP factor expression in young or replicative senescent EC in the presence or absence of IL-1 receptor antagonist (n = 6 biologically independent samples each). d CDK inhibitor expression in 3T3-L1 adipocytes treated with the conditioned medium derived from young or replicative senescent EC with or without IL-1 receptor antagonist treatment (n = 6 biologically independent samples each). e CDK inhibitor expression in young control EC (GFP), premature senescent EC (TERF2DN), or premature senescent EC transfected with IL-1a siRNA (n = 4 biologically independent samples each). f SASP factor expression in young control EC (GFP), premature senescent EC (TERF2DN), or premature senescent EC transfected with IL-1a siRNA (n = 4 biologically independent samples each). Non-repeated ANOVA with post hoc analysis of Fisher’s PLSD was used for difference evaluation between the groups (a, c–f). Data are presented as mean ± s.e. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ^#not significant. Source data are provided as a Source Data file.