Targeted Apoptosis of Senescent Cells Restores Tissue Homeostasis in Response to Chemotoxicity and Aging

Abstract

The accumulation of irreparable cellular damage restricts healthspan after acute stress or natural aging. Senescent cells are thought to impair tissue function, and their genetic clearance can delay features of aging. Identifying how senescent cells avoid apoptosis allows for the prospective design of anti-senescence compounds to address whether homeostasis can also be restored. Here, we identify FOXO4 as a pivot in senescent cell viability. We designed a FOXO4 peptide that perturbs the FOXO4 interaction with p53. In senescent cells, this selectively causes p53 nuclear exclusion and cell-intrinsic apoptosis. Under conditions where it was well tolerated in vivo, this FOXO4 peptide neutralized doxorubicin-induced chemotoxicity. Moreover, it restored fitness, fur density, and renal function in both fast aging Xpd^TTD/TTD and naturally aged mice. Thus, therapeutic targeting of senescent cells is feasible under conditions where loss of health has already occurred, and in doing so tissue homeostasis can effectively be restored.

Keywords: FOXO4; IL6; LMNB1; Senescence; TP53; aging; apoptosis; cell-penetrating peptide; chemotherapy; tissue homeostasis.

Fig. 1. FOXO4 is elevated in senescent normal fibroblasts and ensures their viability

A) Schematic representation of the mRNA expression changes (Fig. S1A) of the cell-intrinsic apoptosis pathway(Tait and Green, 2010) between senescent and control (proliferating) IMR90 fibroblasts. Inset: Immunofluorescence for PUMA, BIM and BCL2. B) Volcano plot comparing transcriptional regulators in senescent vs. control IMR90. (See Fig. S1B for expression and p-values). Dark blue: associated with apoptosis. Inset, left: RNA expression of the FOXO cluster. N.D. Not detectable. Right: Protein levels of the FOXO cluster. FOXO1 was ectopically expressed as positive control. C) QPCR for changes in FOXO1, 3 and 4 mRNA after senescence-induction by 10Gy IR. p21^Cip1 (biphasic increase), p53 and ETS2 (biphasic decrease) are included as controls. D) Immunoblot for changes in FOXO3 and 4 protein levels after senescence-induction by 10Gy IR. E) The senescence-induced FOXO4 mRNA expression is successfully countered by two shRNAs. F) Cytochrome-C release assay (left) as measure for apoptosis in the conditions of E), quantified in a histogram (right). G) Induction of cleaved Caspase-3 after senescence induction in (mouse-specific) FOXO4-deprived wildtype or bax/bak^−/− BMK cells. H) AqueousOne viability (left) and colony density (right; see also Fig. S1C) of control and senescent IMR90 cells transduced with the short hairpins used in E).

Fig. 2. FOXO4 localizes to senescence-associated PML/DNA-SCARS, which contain active p53 and can be disrupted by FOXO4-DRI

A) FOXO4 foci and Senescence-Associated Heterochromatin Structures in senescent IMR90 (See also Fig. S2A–I). Bottom: Intensity plot (arbitrary units) of individual pixels measured by the indicated line. B) Quantification of cells containing ≥3 FOXO4 foci in time after senescence-inducing IR. C) FOXO4 foci in senescent cells transduced with the shRNAs against FOXO4 described in 1E). D) Structured Illumination Microscopic (SIM) image of the nucleus of a senescent IMR90 cell stained for FOXO4, 53BP1 and PML. Yellow arrow: Area processed for 3D surface-rendering (Insets). E) FOXO4 and Ser15-phosphorylated p53, assessed as in 2A. Note that for FOXO4 a different antibody (Sigma) was used. F+G) Sequence (H indicates predicted helix) and 3D structure of FOXO4 used for the design of FOXO4-DRI. The amino acids indicated in yellow in F) are shown as yellow spheres in the displayed structure of FOXO4 (3L2C, protein databank). Green aa in F) are not visualized in this 3D structure, but are part of the FOXO4-DRI sequence. Red aa in G) change most upon p53-interaction (Wang et al., 2008). See also Fig. S2J–L. H) ¹H,¹⁵N HSQC NMR spectrum of 15N-labelled recombinant FOXO4^86–206 incubated with increasing stoichiometric equivalents of recombinant p53 (60, 120, 240 or 300 μM, respectively). I) Experiment as in H), but with 1× or 2× stoichiometric equivalents of FOXO4-DRI (300 or 600 μM, respectively). J) Cellular uptake of FOXO4-DRI in senescent IMR90 visualized by an antibody against the HIV-TAT sequence. K) Quantification of the number of FOXO4/PML/53BP1-DNA-SCARS in control and senescent IMR90 incubated 3d with 25 M FOXO4-DRI and the pan Caspase-inhibitor QVD-OPH. # of small 53BP1 foci shown as control. Only infrequently FOXO4 foci were visible in control cells. L) Schematic representation of the p21^CIp1 (CDKN1a) promoter in which the canonical FOXO target sequence is flanked by two p53 binding sites. M+N) Quantification of nuclear p21^Cip1 intensity of senescent IMR90 treated as in K). N) Left: Immunoblot of senescent IMR90 cells incubated for the indicated time points with FOXO4-DRI and processed for Ser15-phosphorylated and total p53. Middle: Nuclear exclusion of pSer15-p53 in cell treated as in K+M). Right: Quantification of pSer15-p53 foci per nucleus of senescent IMR90.

Fig. 3. FOXO4-DRI selectively eliminates senescent cells through p53-mediated cell-intrinsic apoptosis

A) Viability assay of senescent and control IMR90 incubated with increasing doses of FOXO4-DRI (M). The selectivity index (SI50) reflects the differences in EC50 of a non-regression analysis for both groups. See also Fig. S3A. B) Real-time cell density measurement by xCELLigence of control and senescent IMR90 incubated with or without FOXO4-DRI (25 M). C) Viability assay comparing the effects of increasing doses of FOXO4-DRI and the same peptide in L-isoform, FOXO4-L. D) Viability assay comparing FOXO4-DRI, FOXO4-L, and an unrelated FOXM1-DRI peptide(Kruiswijk et al., 2016), at 6.25, 12.5 and 25 M, respectively. E) Viability assay comparing the pan-BCL inhibitor ABT-737 to FOXO4-DRI, when applied in three consecutive rounds at 1/3 the final concentration each (See also Fig. S3B+C). SI75 reflects differences in EC75 of a non-regression analysis for both groups. F) Viability assay comparing the effect of FOXO4-DRI on cells depleted for p53 by shRNA. See Fig. S3D for effects on p53 expression. G) Viability assay comparing the effect of FOXO4-DRI senescent cells incubated with pan-caspase inhibitors (20 M). H) Representative still images of real-time confocal-based imaging of senescent and control cells in the presence of a Caspase-3/7 activatable dye (green) and incubated with FOXO4-DRI. See also Mov. 3+4. Imaging started 8h after FOXO4-DRI addition.

Fig. 4. FOXO4-DRI counteracts Doxorubicin-induced senescence and chemotoxicity in vivo

A) SA-β-GAL assay to detect senescence in IMR90 7d after 2× treatment (1d in between) with 0.1 M Doxorubicin. B) Immunofluorescence for p16^ink4a and FOXO4 in control or Doxo-senescent IMR90. See also Fig. S4B. C) Quantification of the % of cells positive for p16^ink4a, IL1α, IL6 and FOXO4 foci after Doxorubicin-induced senescence. D) Viability assay comparing the effect of FOXO4-DRI on control and Doxo-senescent cells in vitro. SI determined as in Fig. 3A. E) Viability assay comparing ABT-737 vs. FOXO4-DRI on Doxo-senescent cells. F) Same as in E, but both added 3× at 1/3 the final concentration. See also Fig. S4C. G) Viability assay comparing effects of incubation of FOXO4-DRI for various time points prior, during or after Doxorubicin exposure (blue line) vs. cells already induced to senesce by Doxorubicin (green boxes). M=Mock. H) Representative bioluminescence image and quantification of p16-driven senescence (RLUC) in p16∷3MR mice (See Fig. S4A+D), treated as indicated with Doxorubicin, followed by FOXO4-DRI or Mock. I) Timeline of experiments in J-N. Doxorubicin: 2× i.p. at 10mg/kg. FOXO4-DRI: 3× i.v. at 5mg/kg, every other day (day 1, 3 and 5). J) Quantification of the % change in body weight of Doxorubicin-exposed mice treated with FOXO4-DRI or PBS, respectively. K) Quantification of the number of liver cells with ≥ 10 FOXO4 foci after Doxorubicin-exposure and treatment with PBS or FOXO4-DRI. L) Visualization and quantification of the % of liver cells from the mice in (K) expressing IL-6. M) Quantification of the Doxorubicin-induced increase in plasma AST levels as marker for liver damage. N) Quantification of the effects of PBS or FOXO4-DRI on Doxorubicin-induced plasma levels of AST and Urea as markers for liver and kidney damage, respectively. See also Fig. S4E+F.

Fig. 5. FOXO4-DRI decreases senescence and counters features of frailty in fast aging Xpd^TTD/TTD mice

A) Representative mice and quantification of p16^ink4a-driven RLUC Radiance in 26wk young wildtype and Xpd^TTD/TTD mice crossed into p16∷3MR. B) Left: Timeline for visualization of effects of FOXO4-DRI or PBS on p16-driven senescence by bioluminescence in Xpd^TTD/TTD. Middle: Representative visualization of p16^ink4a-driven senescence in the same Xpd^TTD/TTD-p16∷3MR mouse before and after FOXO4-DRI. Right: Quantification of the effects of FOXO4-DRI or PBS on senescence in a larger cohort of Xpd^TTD/TTD-p16∷3MR mice. C) Timeline for measuring the effects of FOXO4-DRI or PBS on hair density, behavior and running wheel activity in D–I. D) FOXO4-DRI improves fur appearance of Xpd^TTD/TTD mice. Left panels: Representative images of the same Xpd^TTD/TTD animal before and after treatment with FOXO4-DRI. Right panel: quantification of the average change in fur score (See also Fig. S5A). E) Quantification of abdominal temperature measured by infrared thermometer as measure for fur density of wt vs. Xpd^TTD/TTD mice (left) and the effects of FOXO4-DRI and PBS in the mice from D and Fig. S5A (right). F) Quantification of the response of the Xpd^TTD/TTD mice from D to gentle physical stimuli before and after treatment with FOXO4-DRI or PBS. Note that Xpd^TTD/TTD mice are generally relatively non-responsive. See also Fig. S5B. G) Quantification of the average distance run per day of wt. vs Xpd^TTD/TTD. H) Example of changes in running wheel behavior of a wt vs. Xpd^TTD/TTD mouse treated with FOXO4-DRI. Data normalized to 100% for respective running wheel activity at baseline. On day 10 a blood sample was taken, resulting in a transient decrease in activity. I) Quantification of the average change in running wheel activity in wildtype and Xpd^TTD/TTD mice after PBS or FOXO4-RI treatment.

Fig. 6. By targeting senescence, FOXO4-DRI counters loss of renal function of Xpd^TTD/TTD mice

A) Quantification of renal filtering capacity measured by plasma [Urea] in 13, 26 and 130w old wildtype mice and 13 and 26w Xpd^TTD/TTD mice. B–D) Visualization of senescence (SA-β-GAL), the major SASP factor IL-6 and FOXO4 foci in 26w + 130w wildtype and 26w Xpd^TTD/TTD old kidneys. Tubuli (T), Glomeruli (G). Inset C): Magnification of SA-β-GAL to reveal affected areas. Inset D) quantification of the % of renal cells expressing ≥10 FOXO4 foci. E) TUNEL assay to detect apoptosis in kidney sections of 130w old wt mice treated 3d with PBS or FOXO4-DRI. See Fig. S6A–D for pipeline and results with shFOXO4. F) Quantification of the % of platelets at time of sacrifice vs. baseline for wt and Xpd^TTD/TTD mice treated with PBS or FOXO4-DRI. See also Fig. S6E. G) Representative Images of kidneys from 26w wt or Xpd^TTD/TTD mice stained for LMNB1 loss. Quantified are the average number of nuclei per kidney positive for LMNB1 (at least 400nuclei per mouse). H) Viability assay on control or senescent IMR90 incubated with recombinant IL1α, IL1β or IL1 receptor antagonist (IL1-RA) 24h prior to exposure of FOXO4-DRI. I) Viability plot showing the effect of FOXO4-DRI on control and senescent IMR90 pretreated with Cortisol and LPS, prior to FOXO4-DRI treatment. J) Staining as in G), but for the SASP marker IL-6. Quantified is the average IL-6 intensity per kidney over at least 3 frames per mouse for at least 4 mice per group. K) Quantification of the % plasma [Urea] of three pooled cohorts of wt and Xpd^TTD/TTD mice (n=7–8 mice/treatment) after 30d treatment with PBS or FOXO4-DRI. Data are represented as mean +/− SEM. See also Fig. S4G. L) Experiment as in K), but using Ganciclovir (GCV) to mediate semigenetic clearance of senescent cells through the Thymidine Kinase expressed by the p16∷3MR construct. As GCV is i.p. administered, also FOXO4-DRI was i.p. administered in this experiment.

Fig. 7. By targeting senescence, FOXO4-DRI counters frailty and loss of renal function in naturally aged p16∷3MR mice

A) Quantification of p16^ink4a-driven RLUC radiance in 104w old p13∷3MR mice compared to 26w counterparts. Note there is a larger degree of spread in the signal, suggesting biological variation. B) Quantification of the % platelets at time of sacrifice/baseline of naturally aged p16∷3MR mice treated with PBS or FOXO4-DRI for 30d. Procedure as in Fig. 5C. C) Representative images and quantification of p16^ink4a-driven RLUC radiance of mice the from B). D) Example of fur density in FOXO4-DRI vs. Mock-treated male p16∷3MR mice. See also Fig. S7B. E) Quantification of the responsiveness of the mice in B–D treated with FOXO4-DRI or PBS. Analysis as in Fig. 5F. F–I) Quantification of the effects of FOXO4-DRI on LMNB1 loss and IL6 intensity in the kidneys and plasma [Urea] and [Creatinine] of the naturally aged p16∷3MR mice from B). G) Quantification of % plasma [Urea] and [Creatinine] of naturally aged (110+wk) p16∷3MR mice at 30d after i.p. injection with 3× 5mg/kg (every other day) FOXO4-DRI or 5×25mg/kg/day with GCV to clear senescent cells semigenetically.