Glycoprotein nonmetastatic melanoma protein B regulates lysosomal integrity and lifespan of senescent cells

Abstract

Accumulation of senescent cells in various tissues has been reported to have a pathological role in age-associated diseases. Elimination of senescent cells (senolysis) was recently reported to reversibly improve pathological aging phenotypes without increasing rates of cancer. We previously identified glycoprotein nonmetastatic melanoma protein B (GPNMB) as a seno-antigen specifically expressed by senescent human vascular endothelial cells and demonstrated that vaccination against Gpnmb eliminated Gpnmb-positive senescent cells, leading to an improvement of age-associated pathologies in mice. The aim of this study was to elucidate whether GPNMB plays a role in senescent cells. We examined the potential role of GPNMB in senescent cells by testing the effects of GPNMB depletion and overexpression in vitro and in vivo. Depletion of GPNMB from human vascular endothelial cells shortened their replicative lifespan and increased the expression of negative cell cycle regulators. Conversely, GPNMB overexpression protected these cells against stress-induced premature senescence. Depletion of Gpnmb led to impairment of vascular function and enhanced atherogenesis in mice, whereas overexpression attenuated dietary vascular dysfunction and atherogenesis. GPNMB was upregulated by lysosomal stress associated with cellular senescence and was a crucial protective factor in maintaining lysosomal integrity. GPNMB is a seno-antigen that acts as a survival factor in senescent cells, suggesting that targeting seno-antigens such as GPNMB may be a novel strategy for senolytic treatments.

Figure 1

GPNMB positively regulates cellular lifespan. (a) Replicative lifespan of HUVECs infected with a retroviral vector expressing shRNAs targeting GPNMB (sh-GPNMB1 or sh-GPNMB2) or a control vector (sh-Cont). Infected cells were passaged until the termination of replication and the number of population doublings was determined (n = 3 for sh-Cont, n = 4 for sh-GPNMB1, and n = 4 for sh-GPNMB2). **P < 0.01, sh-Cont vs. sh-GPNMB1; ##P < 0.01, sh-Cont vs. sh-GPNMB2. (b) Western blot analysis of p53 and p16 expression by HUVECs infected with a retroviral vector expressing shRNA for GPNMB (sh-GPNMB) or a control vector (sh-Cont). Original blots are presented in Supplementary Fig. 6. (c) SA-β-gal assay of HUVECs infected with a retroviral vector expressing shRNA for GPNMB (sh-GPNMB) or a control vector (sh-Cont) at passage 7. Representative photomicrographs are shown at × 200 magnification. Scale bar = 200 μm. The right graph displays quantification of SA-β-gal activity (n = 3 each). (d,e) SA-β-gal assay and PCR analysis of doxorubicin-induced senescent HUVECs infected with a GPNMB overexpression vector (GPNMB oe) or a control vector (Control) ((d) n = 4 for each; e, n = 3 for each). Representative photomicrographs are shown. Scale bar = 200 μm. The graphs display quantification of SA-β-gal activity (d) and expression of senescence makers (e). (f) Immunostaining for GPNMB in SK-MEL-2 at 4 °C and 37 °C. Plasma membrane was stained with WGA lectin (green), nuclei were labeled with Hoechst (blue), and lysosomes were labeled with Lysotracker (green). Scale bar (upper panel) = 100 μm. Scale bar (middle and lower panels) = 50 μm. The data were analyzed by repeated measures analysis followed by Tukey’s multiple comparison test (a) or by the two-tailed Student’s t test (c–e). *P < 0.05, **P < 0.01. The data are shown as the mean ± SD with plots of all individual data (a) or box and whisker plots (c–e).

Figure 2

GPNMB plays a critical role in lysosomal integrity. (a) Fluorescent imaging of GPNMB-mCherry (red) and lysosomes (LysoTracker Green; green). Scale bar = 25 μm. (b) Fluorescent imaging of lysosomes (LysoTracker Green; green), lysosomal pH (Lysosensor; blue), and lysosomal enzymatic activity (Lysosomal Intracellular Activity Assay; green) in HUVECs infected with a retroviral vector expressing shRNA for GPNMB (sh-GPNMB) or a control vector (sh-Cont). The right graphs display quantification of staining intensity (n = 3 each). Scale bar = 25 μm. (c) Transmission electron microscopy of HUVECs infected with a retroviral vector expressing shRNA for GPNMB (sh-GPNMB) or a control vector (sh-Cont). Scale bar = 2 μm for low-magnification images and 500 nm for high-magnification images. (d) Fluorescent imaging of GPNMB-mCherry (red) and mitochondria (MitoTracker; green) in HUVECs incubated for 8 h with (Starvation) or without (Cont) starvation for amino acids and serum. White arrow heads indicate co-localization of GPNMB expression with mitochondria. Scale bar = 20 μm. (e) Fluorescent imaging of mitophagy (Mitophagy dye; red) and mitochondria (MitoTracker; green), showing that introduction of siRNA for GPNMB (si-GPNMB) impairs mitophagy in senescent HUVECs after starvation for 8 h compared to cells infected with control siRNA (si-Cont). The right graph displays quantification of staining intensity (n = 4 each). Scale bar = 50 μm. (f) Fluorescent imaging of reactive oxygen species in mitochondria (MitoSox) in HUVECs after introduction of si-GPNMB or si-Cont. The right panel displays quantification of staining intensity (n = 4 each). Scale bar = 25 μm. The data were analyzed by the two-tailed Student’s t test and are presented as box and whisker plots (b,e,f). *P < 0.05, **P < 0.01.

Figure 3

GPNMB plays a critical role in lysosomal integrity. (a) GO terms for GPNMB-binding proteins identified by LC–MS/MS. (b) Binding of GPNMB to ATP6V1A in young and senescent HUVECs by immunoprecipitation (IP) with anti-GPNMB antibody and Western blot with anti-ATP6V1A antibody. Original blots are presented in Supplementary Fig. 7. (c) Expression of ATP6V0A3, ATP6V1A, and ATP6V1B2 in young and senescent HUVECs following treatment with si-GPNMB or si-Cont. Original blots are presented in Supplementary Fig. 8. (d) Association between the V_o domain and the V₁ domain in senescent HUVECs treated with si-GPNMB or si-Cont by IP with anti-ATP6V1B2 antibody and Western blot with anti-ATP6V0A3 antibody, anti-ATP6V1A antibody, or anti-ATP6V1B2 antibody. Original blots are presented in Supplementary Fig. 9. The graph on the right shows the quantification of ATP6V0A3 associated with ATP6V1B2 (n = 3) calculated by the following formula: IP si-GPNMB ATP6V0A3/INPUT si-GPNMB ATP6V0A3 vs. IP si-Cont ATP6V0A3/INPUT si-Cont ATP6V0A3. The data were analyzed by the two-tailed Student’s t test and are presented as box and whisker plots (d). *P < 0.05, **P < 0.01.

Figure 4

GPNMB expression is upregulated by lysosomal stress. (a) ATAC-seq of HUVECs. (b) ChIP-qPCR analysis of TFEB and TFEC transcription factors in young HUVECs and irradiation-induced senescent HUVECs (n = 4 each). The 10% of sonicated sample was used as an input control. The ratio of precipitated chromatin (%INPUT) was calculated as follows: %INPUT = 2^{(Ct (INPUT) – log}₂^{10 – Ct (IP))}*100. (c) Relative expression of the MITF/TFE transcription factors in young and replicative senescent HUVECs examined by qPCR (n = 3 each). (d) The left graph displays relative expression of GPNMB in replicative senescent HUVECs after introduction of siRNAs for the MITF/TFE transcription factors (si-MITF, si-TFEB, si-TFEC, and si-TFE3) or control siRNA (si-Cont) (n = 6 for si-Cont, n = 8 for si-MITF, n = 8 for si-TFEB, n = 8 for si-TFEC, and n = 8 for si-TFE3). Two outliers (○) were detected in the si-MITF and si-TFEB group by boxplot analysis and were excluded from statistical analysis. The right graph displays relative expression of GPNMB after introduction of a mixture of 4 siRNAs (si-MITF, si-TFEB, si-TFEC, and si-TFE3) or control siRNA (si-Cont). (n = 3 each). e, qPCR showing relative expression of GPNMB and the transcription factors MITF/TFEs in HUVECs treated for 24 h with 100 nmol l^–1 bafilomycin A1 (BafA1, n = 4) or the vehicle (Control, n = 4). The data were analyzed by the two-tailed Student’s t test (b–e), or by one-way ANOVA followed by Tukey’s multiple comparison test (d). *P < 0.05, **P < 0.01. The data are shown as box and whisker plots (b–e).

Figure 5

Gpnmb has a protective effect against vascular pathology. (a,b) Endothelium-dependent (left) and endothelium-independent (right) vasorelaxation were examined in the iliac arteries of Gpnmb knockout (Gpnmb KO) mice and littermate controls (WT) after feeding an HFD for 8 weeks (n = 9 each) (a), or in arteries from Gpnmb-overexpressing transgenic (Gpnmb Tg) mice and littermate controls (WT) after feeding the HFD for 8 weeks (n = 6 each) (b). (c,d) Laser Doppler perfusion imaging of hind limbs from Gpnmb KO mice and littermate controls (WT) (c), or Gpnmb Tg mice and littermate controls (WT), after 14 days of ischemia (d). The right graphs show quantification of blood flow recovery on Laser Doppler perfusion imaging at the indicated times (c, n = 9 each; d, n = 9 for WT and n = 6 for Tg). (e,f) Oil red O staining of thoraco-abdominal aortas from ApoE KO mice and ApoE KO/Gpnmb KO (DKO) mice (e), or ApoE KO mice and ApoE KO/Gpnmb Tg mice (f). The right graphs display quantification of the plaque area in the thoraco-abdominal aorta (e, n = 6 for ApoE KO and n = 4 for ApoE KO/Gpnmb KO; f, n = 6 each). Data were analyzed by repeated measures analysis followed by Tukey’s multiple comparison test (a–d), or by the two-tailed Student’s t test (e and f). *P < 0.05, **P < 0.01, NS (not significant). The data are shown as the mean ± SEM with plots of all individual data (a–d) or box and whisker plots (e,f).