COPI vesicle formation and N-myristoylation are targetable vulnerabilities of senescent cells

Drugs that selectively kill senescent cells (senolytics) improve the outcomes of cancer, fibrosis and age-related diseases. Despite their potential, our knowledge of the molecular pathways that affect the survival of senescent cells is limited. To discover senolytic targets, we performed RNAi screens and identified coatomer complex I (COPI) vesicle formation as a liability of senescent cells. Genetic or pharmacological inhibition of COPI results in Golgi dispersal, dysfunctional autophagy, and unfolded protein response-dependent apoptosis of senescent cells, and knockdown of COPI subunits improves the outcomes of cancer and fibrosis in mouse models. Drugs targeting COPI have poor pharmacological properties, but we find that N-myristoyltransferase inhibitors (NMTi) phenocopy COPI inhibition and are potent senolytics. NMTi selectively eliminated senescent cells and improved outcomes in models of cancer and non-alcoholic steatohepatitis. Our results suggest that senescent cells rely on a hyperactive secretory apparatus and that inhibiting trafficking kills senescent cells with the potential to treat various senescence-associated diseases.

Source gel images for Supplemental Figure 14 Uncropped images of gels and western blots shown in Supplemental Figure 14

Supplemental Figure 1 .
Setting up RNAi screens for senolytic targets.a,Quantification of the percentage of cells with 3 or more cytoplasmic foci of fluorescent siGLO Red (Right) in control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells 72h after reverse transfection at day 6 after senescence induction.Cells were transfected with varying volumes of Dharmafect 1 per well.Representative IF images are shown (left).(n=3).Scale bar, 100µm.b, Quantification of GFP nuclear intensity (right) relative to un-transfected control for control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells 72h after reverse transfection with GFP siRNAs on day 6 post-senescence induction cells.(n=2) Representative IF images shown (left).Scale bar, 100µm.c, Survival of control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells 72h after reverse transfection with the indicated siRNAs on day 6 post-senescence induction.Pools of 4 individual siRNAs targeting BCL2L1 (pool L1), BCL2L2 (pool L2), or 4 siRNAs each against BCL2L1 and BC2L2 (pool L1+L2) were used.Treatment of cells with 1µM ABT-263 for 72 h starting 6 days after senescence induction was used as a senolytic control (ABT-263).(n=3).Data throughout the figure is represented as mean±SD where applicable.Statistical tests throughout figure were performed using unpaired, two-tailed, student's t-test.N represents independent experiments.Supplemental Figure 2. Knockdown of COPB2 and COPG1 using shRNAs.a, Levels of COPB2 after knockdown with three independent shRNAs.(Left) qRT-PCR (n=3) One-way ANOVA.(Right) Panel showing immunoblot representative of two independent experiments.Immunoblot of GAPDH is included as a loading control.b, Levels of COPG1 after knockdown with three independent shRNAs.(n=3).One-way ANOVA.c, Quantification of cell survival of control (DMSO) and doxorubicin-induced senescent (Doxo) cells.(n=4) Unpaired two-tailed Student's t-test.Data represented as mean±SD throughout figure.N represents independent experiments.Supplemental Figure 3. COPI inhibition causes apoptosis in senescent cells.a, Quantification of percentage cell survival of control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells treated in parallel with 20 µM of pan-caspase inhibitor (Q-VD-OPh) and either 1µM ABT-263, 2.5µM golgicide A (GCA) or 150nM brefeldin A (BFA) for 72h.(n=5).Ordinary Two-way ANOVA.b, Quantification of percentage cell survival of senescent (4OHT) treated in parallel with inhibitors to inhibit pyroptosis (10µM Z-YVAD-FMK, 20µM VX-765), necroptosis (10µM Nec-1) or ferroptosis (1µM Liproxstatin) and either 2.5µM golgicide A (GCA) or 150nM brefeldin A (BFA) for 72h.(n=3).Comparisons are to the corresponding senescent cells treated with DMSO (blue bars).Ordinary Two-way ANOVA.c, Caspase-3/7 activity in control (DMSO) or oncogene-induced senescent (4OHT) cells after reverse transfection with COPB2 siRNAs 6 days after senescence induction (n=2).d, Quantification of cells positive for cleaved caspase 3 as assessed by IF analysis (n=4).Unpaired, two-tailed, Student's t-test.Data is represented throughout the figure as mean ± SD.N represents number of independent experiments.Supplemental Figure 4. GBF1 inhibitors are senolytic.a, Induction of senescence by bleomycin and irradiation in IMR90 cells.Quantification of the percentage of IMR90 cells positive for SA-β-gal staining (middle) or BrdU incorporation (right) 6 days after treatment of cells with 33µM bleomycin or 20Gy Irradiation.(n=3).Representative image of SA-β-gal staining shown (left).Scale bar, 100µm.One-way ANOVA.b, Bleomycin-induced senescence in NHLF cells.Quantification of the percentage of NHLF cells positive for SA-β-gal staining (middle) or BrdU incorporation (right) 6 days after treatment of cells with 50 µg/ml bleomycin (n=3).Representative image of SA-βgal staining shown (left).Scale bar, 50µm.Unpaired, two-tailed Student's t-test.c, Percentage cell survival in either control (DMSO) or bleomycin-treated (Bleo) NHLF cells 72h after treatment on day 7 with brefeldin A (left) or golgicide A (GCA, right).(n=3).Unpaired, two-tailed.Student's t-test.d, Bleomycin-induced senescence in PBECs.Quantification of the percentage of PBECs staining positive for SA-β-gal activity (middle) or BrdU incorporation (right) 6 days after treatment with 100ng/ml bleomycin (n=3).Representative image of SA-β-gal staining shown (left).Scale bar, 50µm.Unpaired, two-tailed.Student's t-test.e, Percentage cell survival in either control (DMSO) or bleomycin-treated (Bleo) PBECs 72h after treatment on day 7 with brefeldin A (left) or golgicide A (GCA, right).(n=3).Unpaired, two-tailed.Student's ttest.. f, Quantification of percentage cell survival in either PBECs staining negative (p16 INK4a negative) or positive (p16 INK4a positive) for p16 Ink4a 72h after treatment with brefeldin A (n=3).Representative images of p16 INK4a IFs are shown (left).Unpaired, two-tailed.Student's t-test.Data is represented throughout the figure as mean±SD.N represents number of independent experiments.Supplemental Figure 5. Glucocorticoid treatment or PTBP1 depletion downregulate SASP without preventing the growth arrest associated with senescence.a Percentage of IMR90 ER:RAS cells positive for BrdU incorporation on day 6 post senescence induction.Cells were treated on day 4 with 10 µM beclomethasone (Bec) or on day 0 with 10 µM triamcinolone (Tri) (n=3).One-way ANOVA.b, Representative immunofluorescence images of IMR90 ER:RAS cells treated with beclomethasone or triamcinolone, fixed and stained 6 days postinduction.Scale Bar, 100µm.c, Quantification of percentage cells positive for p16 INK4a

(
left) and p21 CIP1 (right) on day 6 after senescence induction (n=3).Data represented as mean±SD.One-way ANOVA.d,Percentage of cells positive for IL6 (Left) and IL8 (Right) on day 10 after senescence induction (n=3) One-way ANOVA, Dunnett's correction.e, Relative mRNA levels of PTPB1 on day 4 following transduction with shRNAs.(n=2).p21 CIP1 .f, Percentage of cells positive for BrdU incorporation (left), p16 INK4a (middle), or p21 CIP1 (right) on day 6 after senescence induction (n=4).Doxycycline was added on day 0 to induce shRNAs.One-way ANOVA, Dunnett's correction.g, Representative IF images of cells from the experiment described in e-f.Scale Bar, 100µm.h, SASP inhibition caused by the knockdown of PTBP1 prevents the senolysis induced by COPB2 depletion.Quantification of cell survival of senescent (4OHT) and control (DMSO) IMR90 ER:RAS cells infected with the indicated shRNAs (n=4).Data represented as mean±SD.Statistical comparisons of 4OHT Vector vs. 4OHT shPTBP1 shRNAs are shown.Two-way ANOVA.Data throughout the figure is represented as mean ± SD.N throughout figure represents independent experiments.Supplemental Figure 6.COPB2 depletion activates the unfolded protein response.a-b, GSEA plot showing enrichment of the indicated signatures after COPB2 knockdown with either shCOPB2.1 (a, n=3) or shCOPB2.2(b, n=3) in senescent IMR90 ER:RAS cells.c-d, GSEA plot showing enrichment of the indicated signatures after COPB2 knockdown with either shCOPB2.1 (a, n=3) or shCOPB2.2(b, n=3) in IMR90 cells undergoing bleomycin-induced senescence.NES, normalized enrichment score; FDR, false discovery rate.N throughout figure represents independent experiments.Supplemental Figure 7. Effect of GBF1 inhibitors on SASP production, secretion, and intracellular accumulation.a, Representative IF images showing IL8 treatment with 1.25µM golgicide A (GCA).(n=3, quantification in Figure 4c) Scale bar, 100µm.b, Quantification of intracellular levels of SASP factors as assessed by IF at either 24h (left) or 48h (right) following treatment of IMR90 ER:RAS cells with either 1.25µM golgicide A (GCA) or 150nM brefeldin A (BFA), 7 days after senescence induction.(n=3).Two-way ANOVA.c, Quantification of mRNA levels for the indicated SASP factors 24h after treatment of IMR90 ER:RAS with either 1.25µM golgicide A (GCA) or 150nM brefeldin A (BFA), 7 days after senescence induction.(n=3).Twoway ANOVA.d, Fold change (relative to senescent cells) of secreted SASP levels 24h after treatment of day 7 IMR90 ER:RAS with either 1.25µM golgicide A (GCA) or 150nM brefeldin A (BFA).(n=3) as quantified by ELISA.Two-way ANOVA.Data throughout the figure are represented as mean±SD.N throughout figure represents independent experiments.Supplemental Figure 8. COPI inhibition results in an unfolded protein response in cells undergoing bleomycin-induced senescence.a -b, Representative IF images of XBP1 staining (a) and quantification (b) in either control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells treated with either 1.25µM golgicide A (GCA) or 150nM brefeldin A (BFA) for 48h.(n=3 (-), GCA & BFA; n=2, ABT-263).Unpaired, two-tailed, t-test.Scale bar, 100µm.c-d, Quantification of percentage of positive cells staining for nuclear XBP1 (c), or ATF6 (d) in control (DMSO) or therapy-induced senescent (Bleomycin) IMR90 cells treated on day 7 for 48h with either 1.25µM golgicide A (GCA) or 150nM brefeldin A (BFA) for 48h.(n=3) Unpaired, two-tailed Student's t-test.Data represented as mean ± SD.N represents number of independent experiments throughout figure.Supplemental Figure 9. Washout experiments with GBF1 inhibitors.a Heatmap showing expression of SASP factors in control (DMSO), oncogene-induced senescent (4OHT), and therapy induced senescence (BLEO) IMR90 cells.Data are shown as row z-score normalized.b, Percentage cell survival of control (DMSO) or bleomycintreated (BLEO) IMR90 cells treated with 1µM ABT-263 (ABT), 2.5µM golgicide A (GCA) and 150nM brefeldin A (BFA) on day 7 post-induction for 24h (top, n=3) or 48h (bottom, n=5) followed by drug washout and fixation 72h after initial treatment.Unpaired, two-tailed Student's t-test.c, Percentage cell survival of control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells treated with 1µM ABT-263 (ABT), 2.5µM golgicide A (GCA) and 150nM brefeldin A (BFA) on day 7 post-induction for 24h, followed by drug washout and fixation 72h after initial treatment.(n=5).Unpaired, twotailed Student's t-test.Data represented as mean ± SD throughout figure.N represents independent experiments.Supplemental Figure 10.Effect of PERK inhibitors on senescence.a-c, Quantification of BrdU (a), p16 INK4a (b), and p21 CIP1 (c) staining by IF of either control (DMSO) or senescent (4OHT) IMR90 ER: RAS cells treated with either 1µM GSK2656157 or 1µM GSK2606414 on day 4 post senescence induction.Cells were fixed 6 days post senescence induction.(n=3) One-way ANOVA, Dunnett's Correction.Data represented as mean ± SD. d, Representative IF images of cells of the experiment described in a-c.Scale Bar, 100µm.N represents independent experiments.Supplemental Figure 11.Induction of senescence in A549 and SK-HEP-1 cells.a, Representative images of SA-β-gal staining in control (DMSO) or therapy-induced senescent (Etoposide) A549 cells 6 days after treatment with etoposide.Scale bar, 100µm.b, Quantification of percentage positive control (DMSO) or therapy induced senescent (etoposide) A549 cells for SA-β-gal staining (left), BrdU incorporation (middle) or p21 CIP1 staining s(right) on day 6.(n=3).Unpaired, two-tailed Student's ttest.c, Representative images of SA-β-gal staining in control (DMSO) or therapyinduced senescent (Etoposide) SK-HEP-1 cells 6 days after treatment with etoposide.Scale bar, 100 µm.d, Quantification of percentage positive control (DMSO) or therapyinduced senescent (Etoposide) SK-HEP1 cells for SA-β-gal staining (left), BrdU incorporation (middle) or p21 CIP1 staining s(right) on day 6.(n=3).Unpaired, Student's t-test.Data represented as mean ± SD.N represents independent experiments.Supplemental Figure 12.Irradiation causes senescence in HFFF2 cells.a, Representative images of SA-β-gal staining (left) in either control (Mock) or irradiationinduced senescent (Irr.20Gy) HFFF2 cells.Quantification of percentage cells positive for BrdU incorporation (middle) or SA-β-gal staining (right) is shown.(n=3).Unpaired, two-tailed Student's t-test.b-c, Relative mRNA levels of COPB2 (b) and COPA (c) following transduction of HFFF2 cells with the corresponding shRNAs again each.(n=3) One-way ANOVA.d, Quantification of percentage cell survival in either control (DMSO) or therapy-induced senescent (Doxo).IMR90 cells following transduction with inducible shRNAs against COPA.shRNAs were induced 7 days after induction of senescence and cells fixed 10 days after doxycycline addition.(n=3) Unpaired, twotailed Student's t-test.e, Senolytic activity of COPA depletion during OIS in IMR90 ER: RAS cells (n=3).shRNAs were induced 7 days after induction of senescence and cells fixed 10 days after doxycycline addition.(n=3) Unpaired, two-tailed Student's t-test.f, Tumor growth curves of a second experiment (Experiment B) showing the tumour volume monitored over time (IR=irradiation).Data represented as mean ± SEM for all mice in each group.(n=7 per group).This experiment and the one shown in Fig 5e (Experiment A) were pooled in the analysis shown in Extended Data Figure 5d.Data represented as mean ± SD throughout the figure unless otherwise stated.N represents independent experiments or mice throughout figure.Supplemental Figure 13.NMT inhibitors are senolytic.a, Heatmap showing expression of SASP factors in either control (DMSO) or senescent (4OHT) IMR90 ER:RAS cells treated with 300nM IMP1088 or 1.5µM DDD86481.Data displayed as row-z-score normalized.(n=3) b-c, GSEA plot showing enrichment of the indicated signatures in senescent cells treated with 300nM IMP1088 (b) or 1.5µM DDD86481 (c).NES, normalized enrichment score; FDR, false discovery rate.d, Caspase-3/7 activity in control (DMSO) or oncogene-induced senescent (4OHT) cells after treatment with DMSO or 300nM IMP1088.Measures start 7 days after senescence induction (n=2).Data represented as mean±SD.N represents independent experiments throughout figure.Supplemental Figure 14.Expression of IpaJ wt targets ARF1.a, In-gel visualization of the effect of NMTi (IMP-1088), IpaJ wild-type (WT), and inactive mutant (C64A) on N-myristoylation of proteins with YnMyr.Each lane shows an independent sample, gel was run once.b-d, Densitometric intensity quantified of ARF1 (b), ARL1 (c), and PPM1b (d) normalized to TUBA levels.Ordinary One-way ANOVA, Dunnett's correction.(n=2 independent samples for controls, n=3 independent samples for all other groups).Each western blot includes the stated biological replicates per condition and was run once.Data represented as mean±SD.e, Immunoblots of ARF1, ARL1, PPM1B, and TUBA were used for the quantification shown in b-d.Immunoblot of TUBA is included as a sample processing control.Supplemental Figure 15.Induction of senescence in HCT116 and MCF7 cancer cells.a, c, Representative images of three independent experiments of SA-β-gal staining in HCT116 cells 6 days after treatment with DMSO (control, a and c), etoposide (a), or doxorubicin (c).Scale bar, 100µm.b, d, Quantification of percentage positive control (DMSO) or therapy-induced senescent (etoposide, b or doxorubicin, d) HCT116 cells for SA-β-gal staining (left), BrdU incorporation (middle) or p21 CIP1 staining (right) on day 6.(n=3).Unpaired, two-tailed Student's t-test.Data represented as mean±SD.e, g, Representative images of three independent experiments of SA-β-gal staining in MCF7 cells 6 days after treatment with DMSO (control, e, and g), etoposide (e), or doxorubicin (g).Scale bar, 100µm.f, h, Quantification of percentage positive control (DMSO) or therapy-induced senescent (etoposide, f or doxorubicin, h) MCF7 cells for SA-β-gal staining (left), BrdU incorporation (middle) or p21 CIP1 staining (right) on day 6.(n=3).Unpaired, two-tailed Student's t-test.Data represented as mean ± SD.Supplemental Figure 16.NMTi are well tolerated in vivo.a Analysis of markers of liver function for C57/BL6 mice given 1 round of treatment with 10mg/kg DDD86481 or 25 mg/kg IMP1320.Blood was collected 7 days after.(n=5 mice for Vehicle and DDD86481, n=4 mice for IMP1320).One-way ANOVA, Dunnett's correction.b, Blood glucose levels for C57/BL6 mice given 1 round of treatment with 10mg/kg DDD86481 or 25 mg/kg IMP1320 and blood collected 7 days after.(n=5 mice for Vehicle and DDD86481, n=4 mice for IMP1320).One-way ANOVA.c, Blood insulin levels for C57/BL6 mice given 1 round of treatment with 10mg/kg DDD86481 and blood collected 7 days after.(n=5 mice for Vehicle, n=6 mice for DDD86481) Unpaired, twotailed, Student's t-test.d, Blood cell counts for C57/BL6 mice given 1 round of treatment with 10mg/kg DDD86481 or 25 mg/kg IMP1320 and blood collected 1 and 7 days after.(day 1, n=5 mice per group; day 7, n=3 per group).Two-way ANOVA.Data represented as mean ± SD.N represents number of mice throughout the figure.
Source gel images for Supplemental Figure 2aUncropped images of western blots shown in Supplemental Figure2a.The molecular weights (kDa) of size markers are indicated.Uncropped images of western blots shown inSupplemental Figure2a.The molecular weights (kDa) of size markers are indicated.
. The molecular weights (kDa) of size markers are indicated.Red boxes mark the cropped images shown in the Figure.

gel images for Supplemental Figure 14
McHugh et al.Source