Combining stem cell rejuvenation and senescence targeting to synergistically extend lifespan

Abstract

Why biological age is a major risk factor for many of the most important human diseases remains mysterious. We know that as organisms age, stem cell pools are exhausted while senescent cells progressively accumulate. Independently, induction of pluripotency via expression of Yamanaka factors (Oct4, Klf4, Sox2, c-Myc; OKSM) and clearance of senescent cells have each been shown to ameliorate cellular and physiological aspects of aging, suggesting that both processes are drivers of organismal aging. But stem cell exhaustion and cellular senescence likely interact in the etiology and progression of age-dependent diseases because both undermine tissue and organ homeostasis in different if not complementary ways. Here, we combine transient cellular reprogramming (stem cell rejuvenation) with targeted removal of senescent cells to test the hypothesis that simultaneously targeting both cell-fate based aging mechanisms will maximize life and health span benefits. We find that OKSM extends lifespan and show that both interventions protect the intestinal stem cell pool, lower inflammation, activate pro-stem cell signaling pathways, and synergistically improve health and lifespan. Our findings suggest that a combination therapy, simultaneously replacing lost stem cells and removing senescent cells, shows synergistic potential for anti-aging treatments. Our finding that transient expression of both is the most effective suggests that drug-based treatments in non-genetically tractable organisms will likely be the most translatable.

Keywords: aging; senescence; stem cells.

Figure 1

Constant expression of OKSM, Sen and OKSM-Sen led to increased stem cell proliferation over time. (A) esgGal4, UAS-GFP; tubGal80^ts> UAS-TdTomato control flies show a small number of stem cells (n < 10) and few enteroblasts during after seven days. Expression of OKSM (B), Sen (C) or both Sen and OKSM (D) led to an increase in both ISCs and EBs with the highest number of ISCs observed in the Sen condition as quantified (Right). On day 14, little change was observed in control files (E), but consistently higher numbers of ISCs were maintained in all three experimental conditions (F–H, quantified Right). Day 21 showed little change from day 14 with similar numbers of ISCs observed in the control flies (I) and consistently higher numbers in the three experimental conditions (J–L, quantified Right). By day 28, the number of ISCs was markedly decreased in control flies (M), while all three experimental conditions maintained ISC numbers (N–P, quantified Right).

Figure 2

Expression of OKSM, Sen and OKSM-Sen led to decreased SA-β-gal expression. Levels of b-galactosidase were assessed at day 40 in the following ubiquitous expression experiments using armGal4; tubGal80^ts to drive the expression of UAS-OKSM (A), UAS-Sen (B), combination of UAS-Sen and UAS-OKSM (C), and the control UAS-TdTomato at both day 14 (D) and day 40 (E). (F) Results were quantified from five experiments.

Figure 3

Gene expression changes in the Drosophila gut in Sen, OKSM and combination treatment. (A) Heatmap of gene expression in ubiquitous expression experiments with armGal4; tubGal80^ts > UAS-TdTomato (WT), armGal4; tubGal80^ts > UAS-OKSM (OKSM), armGal4; tubGal80^ts > UAS-Sen (Sen) and armGal4; tubGal80^ts > UAS-Sen; UAS-OKSM (OKSM-Sen) dissected midguts showing seven clusters with differing expression patterns across the four conditions. (B) Gene Ontology and KEGG Pathway Enrichment analysis of ubiquitous expression highlighting key signaling and metabolic pathways associated with the individual clusters (FDR < 10%). (C) Heatmap of gene expression changes in stem cell only expression experiments with esgGal4; tubGal80^ts > UAS-TdTomato (WT), esgGal4; tubGal80^ts > UAS-OKSM (OKSM), esgGal4; tubGal80^ts > UAS-Sen (Sen) and esgGal4; tubGal80^ts > UAS-Sen; UAS-OKSM (OKSM-Sen) dissected midguts again showing seven clusters with differing expression patterns across the four conditions. (D) Gene Ontology and KEGG Pathway Enrichment analysis of ISC only highlighting key signaling and metabolic pathways associated with the individual clusters (FDR < 10%). (E) Venn diagram showing overlap in genes from ubiquitous and ISC-restricted expression. (F) Gene Ontology and KEGG Pathway Enrichment analysis of the major pathways affected in both models.

Figure 4

Cycling OKSM and Sen expression leads to lifespan and health span extension while combined OKSM-Sen expression increases both. (A) Survival curve for armGal4; tubGal80^ts > UAS-TdTomato (TdTom), armGal4; tubGal80^ts > UAS-OKSM (OKSM), armGal4; tubGal80^ts > UAS-Sen (Sen) and armGal4; tubGal80^ts > UAS-Sen; UAS-OKSM (Sen; OKSM) where expression was limited to one twelve-hour period twice per week through a temperature shift. At 25°C the temperature sensitive Gal80 protein ceases to inhibit Gal4 from driving expression from UAS enhancers, allowing for a targeted expression window when flies were shifted from 18°C to 25°C. Expression of OKSM, Sen and Sen; OKSM in adult female flies resulted in increased lifespan as compared to control flies (TdTom). Mean and maximum lifespans are shown along with corrected p-values. (B) Survival curve for the same experiment but with 24 hours of expression twice per week induced by a temperature shift of 18°C to 25°C. There were similar benefits of expression but reduced compared to the 12-hour expression experiment. Mean and maximum lifespans are shown along with corrected p-values. (C) Survival curve for flies expressing OKSM, Sen and Sen; OKSM but maintained at 25°C throughout their lifespans. The overall lifespans are shorter due to the higher temperature, but in addition all three experimental conditions show detriments to both mean and maximum lifespans. Mean and maximum lifespans are shown along with corrected p-values. A P-value of 0 reflects P < 1.0 * 10⁻¹⁰.

Figure 5

Cycling OKSM expression maintains a larger pool of intestinal stem cells over time. We visualized and quantified stem cells using β-catenin-GFP (arm^GFSTF) which appeared enriched in ISCs. Flies cycled expression through temperature shifts from 18° to 25° for twelve hours twice per week. (A) arm^GFSTF; armGal4; tubGal80^ts > UAS-TdTomato control flies showed a small number of stem cells and few enteroblasts after four weeks. Expression of OKSM (B), Sen (C) or both Sen and OKSM (D) led to an increase in both ISCs with the highest number of ISCs observed in the OKSM condition as quantified (E, F). After eight weeks, the number of ISCs in the OKSM condition was much higher in the OKSM flies as compared to wildtype (G, H) and Sen (I) or Sen; OKSM (J) quantified (K, L). At twelve weeks, OKSM flies maintained high numbers of stem cells while the other conditions (M–P) showed fewer as quantified (Q, R).

Figure 6

Gompertz–Makeham mortality and survival analysis demonstrates decreased early mortality with compensating increased aging rate in Sen and OKSM interventions. Survival data for each intervention are shown together with the best-fit survival curve on the right and the corresponding mortality trajectories are shown on the left. Initial log(Mortality) parameter (A) can be read as the intersection between mortality curve and y-axis at age zero. The slope of the mortality curve is proportional to 1/MRDT. The dashed purple line illustrates hypothetical mortality trajectory assuming additivity of effects elicited by OKSM and Sen. (A, B) Mortality trajectories and survival curve for cohorts with continuous induction of OKSM, Sen or OKSM+Sen and controls. (C, D) Mortality trajectories and survival curve for cohorts with induction of OKSM, Sen or OKSM+Sen for 24 h every three days and matched controls. (E, F) Mortality trajectories and survival curve for cohorts with induction of OKSM, Sen or OKSM+Sen for 12 h every three days and matched controls. Note that flies subject to continuous induction (Panels A and B) are permanently kept at 25°C and are therefore aging more rapidly than flies cultured at 18°C and induced for only for short periods. The slope of the mortality trajectory of controls in A is therefore approximately two times as larger, compared to that of controls in panels C and E. For exact MRDT and A parameter values and associated confidence intervals see: (Supplementary Table 2).