Identification of HSP90 inhibitors as a novel class of senolytics

Abstract

Aging is the main risk factor for many chronic degenerative diseases and cancer. Increased senescent cell burden in various tissues is a major contributor to aging and age-related diseases. Recently, a new class of drugs termed senolytics were demonstrated to extending healthspan, reducing frailty and improving stem cell function in multiple murine models of aging. To identify novel and more optimal senotherapeutic drugs and combinations, we established a senescence associated β-galactosidase assay as a screening platform to rapidly identify drugs that specifically affect senescent cells. We used primary Ercc1 ^-/- murine embryonic fibroblasts with reduced DNA repair capacity, which senesce rapidly if grown at atmospheric oxygen. This platform was used to screen a small library of compounds that regulate autophagy, identifying two inhibitors of the HSP90 chaperone family as having significant senolytic activity in mouse and human cells. Treatment of Ercc1 ^-/∆ mice, a mouse model of a human progeroid syndrome, with the HSP90 inhibitor 17-DMAG extended healthspan, delayed the onset of several age-related symptoms and reduced p16^INK4a expression. These results demonstrate the utility of our screening platform to identify senotherapeutic agents as well as identified HSP90 inhibitors as a promising new class of senolytic drugs.The accumulation of senescent cells is thought to contribute to the age-associated decline in tissue function. Here, the authors identify HSP90 inhibitors as a new class of senolytic compounds in an in vitro screening and show that administration of a HSP90 inhibitor reduces age-related symptoms in progeroid mice.

Fig. 1

Development of a novel assay to screen for senotherapeutics. a Schematic diagram of the assay. Passage 2 primary mouse embryonic fibroblasts MEFs from DNA repair-deficient Ercc1 ^−/− mice are passaged at 20% O₂ to induce oxidative DNA damage. After 3 passages, 50% of the cells are senescent (red) and 50% remain non-senescent (yellow). Drugs are tested on these mixed cultures to determine if they affect senescent (C₁₂FDG positive) or non-senescent (C₁₂FDG negative) cells. b Representative images derived from three replicate experiments of p5 Ercc1 ^−/− MEF cultures measuring senescence-associated b-gal (SA-β-Gal) activity using colorimetric X-gal staining. Scale bar, 50 μm. c Representative flow cytometric histogram detecting SA-β-Gal activity using C₁₂FDG in senescent and non-senescent Ercc1 ^−/− MEF populations. The dotted line indicates the cut-off in intensity levels used to define senescent cells. d Representative image from IN Cell Analyzer 6000 to detect SA-β-Gal in p5 Ercc1 ^−/− MEFs using C₁₂FDG. Senescent cells are outlined in red (S), non-senescent cells are outlined in yellow (NS) and blue fluorescence indicates Hoechst-stained DNA in nuclei (N) used to obtain a total cell count. Scale bar, 50 μm. e Quantification of the senescent cell population in non-senescent (NS) and senescent (S) p5 Ercc1 ^−/− MEF cell cultures detected with X-gal and with C₁₂FDG by flow cytometry and INCell 6000 analyzer. Error bars indicate SD for n = 3. *p < 0.05, two-tailed Student’s t-test

Fig. 2

Detection of senescence markers in Ercc1 ^−/− MEFs. a Cell proliferation was measured in congenic WT and Ercc1 ^−/− MEFs from p2 through p5. n = 2. b Relative expression of p21^Cip1 and c p16^INK4a in p2 non-senescent (NS) and p5 senescent (S) Ercc1 ^−/− MEFs determined by qRT-PCR. Error bars indicate SD for n = 3. *p < 0.05, two-tailed Student’s t-test. d Protein expression levels of γH2AX in p2 NS and p5 S Ercc1 ^−/− MEFs. Changes in diameter e and volume f in p2 NS and p5 S Ercc1 ^−/− MEFs. Error bars indicate SD for n=3. * p < 0.05, two-tailed Student’s t-test. Ratio of cells expressing p16^INK4a g and IL-6 h in p2 NS and p5 S cultures determined by ViewRNA fluorescence in situ hybridization. Error bars indicate SD for n ⩾ =2. * p < 0.05, two-tailed Student’s t-test. i Representative FISH image of p16^INK4a and IL-6 of passage 5 S Ercc1 ^−/− MEFs

Fig. 3

Characterization of a novel, C₁₂FDG single-cell SA-ß-gal drug screening assay. a Scheme of possible outcomes after treating a senescent p5 Ercc1 ^−/− MEF culture with a potential senotherapeutic. Red=senescent cells, yellow=non-senescent cells. Senotherapeutics are characterized by either killing of senescent cells (senolytics) or by altering the senescent state of cells otherwise (senomorphics). After 48 h treating p5 senescent Ercc1 ^−/− MEFs with a drug, the relative number of senescent cells b and total cells c remaining is quantitated and plotted relative to untreated control cultures. Red bars indicate relative number of senescent cells, grey bars indicate relative number of total cells. Error bars indicate SD for n = 3. *p < 0.05 for senescent cells, two-tailed Student’s t-test

Fig. 4

Screening of a library of autophagy regulators yields senolytics. a Pie chart indicating the different functional groups of drugs in the autophagy library used in the screen. b The primary screen of all 97 autophagy regulators at 1 mM concentration. Plotted on the x-axis is the number of senescent cells in the drug-treated cultures relative to cells treated with vehicle only. On the y-axis is the fraction of total cells remaining after drug treatment relative to vehicle treated controls. Drugs that reduce the number of senescent cells > 50% can be found in the blue shaded area. Drugs that cause no change in cell number (senomorphics) are indicated by blue dots. Drugs that caused a decrease in total cell number to < 75% (potential senolytics) are indicated in red. Grey dots indicate drugs that lead to no significant change in cell senescence at the concentration used. c Pie chart indicating the functional groups of potential senescence-modulating drugs identified in the autophagy library. d Independent validation of the primary screen expressed as cell senescence and cell number relative to untreated control cultures (UT) of senescent cells. Known lysosomal inhibitors (lysosomal pH changing compounds, Fig. 4C) were excluded. All drugs were used at 1 μM, n = 3, graphed + SD. *p < 0.05, two-tailed Student’s t-test

Fig. 5

HSP90 inhibitors selectively kill senescent cells. a, b Celltox Green cytotoxicity assay. All potential senolytic drugs were added to cultures of a non-senescent (confluent) wild-type and b senescent Ercc1 ^−/− primary MEFs at 4 concentrations (0.03–1.00 μM). Error bars indicate SD for n = 3. c Graph depicting drugs that specifically kill senescent cells. Plotted is the viability of non-senescent WT vs. senescent Ercc1 ^−/− cells after treatment with each drug for 48 h at a 1 μM concentration. Cell toxicity was defined as cell viability < 75%. Only cells that kill senescent cells, without significant toxicity to quiescent, non-senescent cells, are considered as senolytics (indicated in yellow shaded area). d Dose response analysis of senolytic activity of 17-DMAG. Increasing concentrations of 17-DMAG (0.1–1000 nM) were tested and plotted against the fraction of remaining senescent (red) and non-senescent cells (yellow) after 48 h treatment. The EC₅₀ values of their senolytic potential were determined from their dose response curves using a 4-parameter curve fit analysis (graphpad). Error bars indicate SD for n = 3. e Flow cytometric analysis of cell death of senescent Ercc1 ^−/− MEF cell cultures treated with 17-DMAG via AnnexinV/7-AAD staining. MEF cells are either senescent (C₁₂FDG⁺; Top) or non-senescent (C₁₂FDG⁻, Bottom). Live cells were double negative for 7-AAD and AnnexinV (bottom left quadrant), early apoptotic cells were positive for Annexin V (bottom right quadrant), late apoptotic cells were positive for AnnexinV and 7-AAD (top left quadrant) and dead cells were positive for 7-AAD only (bottom right quadrant). f Quantification of the flow cytometry data. Apoptosis of senescent (red) and non-senescent (yellow) cells was calculated by summing up all AnnexinV-positive cells. Error bars indicate SD for n = 3, *p < 0.05, two-tailed Student’s t-test

Fig. 6

HSP90 inhibitors (HSP90inhs) are senolytic in two species and multiple cell types. a, b Representative dose response curves of seven HSP90 inhibitors. Eight concentrations (0.1–1000 nM) of each drug were tested and plotted against the relative fraction of remaining senescent a and non-senescent passage 5 ERCC-deficient MEFs b to determine their senolytic and cytotoxic potential. c Structure, origin, and IC₅₀ (HSP90a/b inhibition) of the HSP90 inhibitors tested in (A and B). EC₅₀ values of their cytotoxic potential for senescent (sen) and non-senescent (non-sen) cells were determined from their dose response curves in A and B using a 3-parameter curve fit analysis (graphpad), n = 2. d Effect of 17-DMAG on multiple cell types. Cell senescence was measured after treatment of senescent cultures of two different types of mouse cell (MEFs and mesenchymal stem cells) and two different human cell types (IMR90 primary myofibroblasts and WI38 primary lung fibroblasts) with 100 nM 17-DMAG. Senescence was induced by three methods: oxidative stress (murine cells), genotoxic stress (etoposide IMR90), and replicative stress (WI38, passage 30). All experiments were performed in triplicate. Error bars indicate  SD, *p < 0.05, two-tailed Student’s t-test. e Viability of HUVECs (human umbilical vein endothelial cells) treated with the HSP90 inhibitor ganetespib. Proliferating and senescent HUVECs were exposed to different concentrations of ganetespib (5–800 nM). After 72 h, the number of viable cells was measured. The red line denotes plating densities on day 0 of non-dividing senescent (set to 100%) as well as proliferating, non-senescent cells (also set to 100%). Plotted are the means ± SEM of five replicates at each concentration. Senescence was induced by 10 Gy ionizing radiation

Fig. 7

Multiple senescence markers are reduced in Ercc1 ^−/− MEFs after treatment with HSP90 inhibitors. a Diameter and volume of senescent Ercc1 ^−/− MEF cells were measured before (red) and after (orange) 1 µM 17-DMAG (17D) treatment for 24 h. Error bars indicate SD for n = 3, *p < 0.05, two-tailed Student’s t-test. b Relative mRNA expression of IL-6 and p16^INK4a determined by qRT-PCR was measured in untreated (UT) and 17-DMAG (17D) treated cells for 24 h. Error bars indicate SD for n = 3. * p < 0.05, two-tailed Student’s t-test. c Representative FISH images of p16^INK4a (green) and IL-6 (red) in senescent Ercc1 ^−/− MEFs with (17D) and without (UT) 17-DMAG treatment. d Quantification of the fraction of cells expressing p16^INK4a and IL-6 in p5 Ercc1 ^−/− MEF cells with (orange) and without (red) 17-DMAG treatment for 24 h determined by ViewRNA FISH. Error bars indicate SD for n = 2. *p < 0.05, two-tailed Student’s t-test. e Immunoblot detection of the DNA damage/senescence marker γH2AX in p5 Ercc1 ^−/− MEF cultures at several time points following exposure to 100 nM 17-DMAG. Semi-quantitative analysis of γH2AX expression relative to β-actin. Western blots were quantified with ImageJ

Fig. 8

Expression of heat shock proteins and HSP90 client proteins in senescent and non-senescent Ercc1 ^-/- MEFs. a Proposed model for how HSP90 promotes resistance to apoptosis in senescent, p5 Ercc1 ^−/− MEFs. b Immunoblot detection of HSP90, pAKT (Ser473), AKT, and actin in early passage, non-senescent cells (NS), intermediate passage (I), and late passage, senescent Ercc1-deficient MEFs (S) grown at 20% O₂. c Immunoblot detection of the same proteins in senescent, p5 Ercc1 ^−/− MEFs treated with 100 nM 17-DMAG at 6 and 24 h post-treatment

Fig. 9

Intermittent treatment of progeroid Ercc1 ^−/Δ mice with the senolytic HSP90 inhibitor 17-DMAG extends healthspan. a Schematic diagram of the in vivo treatment regiment. Animals were treated with 10 mg/kg 17-DMAG by oral gavage 3 times per week, every 3 weeks. Eight symptoms associated with frailty and aging were measured in the mice, 3 times each week. b Graphed is the age at onset of each symptom (appearance of a colored bar) and severity (height of the bar) for a sex-matched sibling pair of Ercc1 ^−/Δ mice treated with or without 17-DMAG. The composite height of the bar is an indication of the overall health (i.e., body condition score with a higher value being worse). c Comparison of age-related symptoms between cohorts of mice treated with the HSP90 inhibitor or vehicle only over time. The average fraction of total symptoms appearing in each age group is plotted. An increase in the percent of symptoms indicates a decrease in health. n = 6 mice per treatment group, error bars indicate SEM, *p < 0.05, **p < 0.01, ***p < 0.001. Relative expression levels of p16^INK4a in kidney d and liver e of Ercc1 ^−/Δ mice treated for 1 week by oral gavage with HSP90 inhibitor (10 mg/kg 17-DMAG oral gavage per treatment, 3×) compared to vehicle only. p16^INK4a was measured by qRT-PCR. n = 4 mice per group; error bars indicate SD, *p < 0.05, two tailed Student t-test