Fisetin Attenuates Cellular Senescence Accumulation During Culture Expansion of Human Adipose-Derived Stem Cells

Abstract

Mesenchymal stem cells (MSCs) have long been viewed as a promising therapeutic for musculoskeletal repair. However, regulatory concerns including tumorgenicity, inconsistencies in preparation techniques, donor-to-donor variability, and the accumulation of senescence during culture expansion have hindered the clinical application of MSCs. Senescence is a driving mechanism for MSC dysfunction with advancing age. Often characterized by increased reactive oxygen species, senescence-associated heterochromatin foci, inflammatory cytokine secretion, and reduced proliferative capacity, senescence directly inhibits MSCs efficacy as a therapeutic for musculoskeletal regeneration. Furthermore, autologous delivery of senescent MSCs can further induce disease and aging progression through the secretion of the senescence-associated secretory phenotype (SASP) and mitigate the regenerative potential of MSCs. To alleviate these issues, the use of senolytic agents to selectively clear senescent cell populations has become popular. However, their benefits to attenuating senescence accumulation in human MSCs during the culture expansion process have not yet been elucidated. To address this, we analyzed markers of senescence during the expansion of human primary adipose-derived stem cells (ADSCs), a population of fat-resident MSCs commonly used in regenerative medicine applications. Next, we used the senolytic agent fisetin to determine if we can reduce these markers of senescence within our culture-expanded ADSC populations. Our results indicate that ADSCs acquire common markers of cellular senescence including increased reactive oxygen species, senescence-associated β-galactosidase, and senescence-associated heterochromatin foci. Furthermore, we found that the senolytic agent fisetin works in a dose-dependent manner and selectively attenuates these markers of senescence while maintaining the differentiation potential of the expanded ADSCs.

Graphical abstract

Graphical Abstract.

Significance Statement.

The accumulation of dysfunctional, senescent cells throughout aging is not confined to specific tissues and cell types, but instead affects the whole body, including stem cells. Similarly, during culture expansion, stem cells accumulate senescence while concurrently losing their regenerative potential. In this study, we found that fisetin (a well-known senotherapeutic agent) can reduce the number of senescent cells during stem cell expansion. The current results indicate that fisetin may be used not only as a promising therapeutic to remove senescent cells in stem cell isolates from older individuals but also to reduce the accumulation of senescence during culture expansion.

Introduction

In the 1960s and 1970s, Dr. A.J. Friedenstein first identified spindle-shaped, plastic adherent cells from bone marrow.¹ Later work would demonstrate both the osteogenic and chondrogenic potential of these cells,^2,3 and in 1991 this cell population would first be termed mesenchymal stem cells (MSCs) by Dr. Arnold Caplan.⁴ Building on these seminal pieces of work, further evidence has demonstrated MSCs promise as a regenerative product across musculoskeletal tissues, nervous tissue, skin, and myocardium, among others.⁵ Recent evidence suggests that MSCs are a particularly promising therapeutic candidate due to their broad distribution in the body, diverse multilineage differentiation potential, and unique immunomodulatory characteristics.^5,6 Although both significant progress have been made in MSC-based therapeutic approaches, there are still several risks such as tumorigenicity and inconsistencies in preparation techniques and clinical outcomes that must be resolved to designate MSC-based therapies as reliable and effective treatment strategies to repair tissue after injury and disease. Therefore, improved and standardized techniques for MSC isolation, culture, and delivery are crucial to advancing our understanding of the mechanisms underlying the regenerative benefits of MSCs in vivo.

Originally discovered in bone marrow, MSCs have since been found in skeletal muscle,⁷ dental,⁸ umbilical cord,⁹ peripheral blood,¹⁰ and adipose tissues.¹¹ However, bone marrow-derived MSCs (BMSCs), among others, are difficult to harvest and isolate, highlighting the need to identify novel, easily harvested repositories of MSCs.¹² In this way, adipose tissue has evolved as a promising source of MSCs. Otherwise known as adipose-derived stem cells (ADSCs), MSCs from adipose tissue, exhibit similar multipotent potential as BMSCs but are significantly easier to harvest^12,13 from subcutaneous fat. In healthy individuals, adipose tissue is systemically distributed, constitutes 10%-30% of body weight, and lends a high cellular isolation yield per gram of harvested tissue.^12,14 Thus, ADSC isolation is both less invasive and more efficient than that of BMSCs, establishing ADSCs as an exciting biologic treatment option and regenerative therapy for age-related musculoskeletal conditions. The therapeutic potential of ADSCs for musculoskeletal conditions such as osteoarthritis,¹⁵ bone defects,¹⁶ and tendinopathy¹⁷ have been investigated through a variety of pre-clinical work and clinical trials. Initial results from these ADSC-based clinical trials are extremely promising; however, there are still limitations to their therapeutic potential in vivo.

To be considered a potential therapy for musculoskeletal repair, MSCs and ADSCs alike require processing beyond the current FDA guidelines for minimal manipulation and clinical administration.¹⁸ This is primarily due to the isolation and culture expansion required to create an administrable population of cells. Unfortunately, once removed from the body and cultured in vitro, previous research has shown that due to the in vitro culture conditions MSCs acquire new phenotypes including a modulated secretome, transcriptome, and surface epitope profile.^19,20 Similarly, due to differences in donor lifestyle, medical history, and age MSCs may not always elicit the same regenerative properties from donor to donor.²¹ These cell culture-acquired phenotypes and donor-dependent differences have been shown to result in a loss of proliferative and differentiation potential, as well as increased markers of senescence throughout the expansion process.²²

The accumulation of senescence in vitro was first observed in 1961 when Dr. Fredrick Hayflick observed human fibroblasts had a limited proliferative period during culture expansion.²³ Biologically, senescence has several roles, including a protective mechanism to prevent the replication of damaged and/or mutated DNA.²⁴ However, senescence is increasingly accepted to be a driving mechanism of aging where stressed and damaged cells enter a non-proliferative state.²⁵ Senescent cell accumulation has also been shown to result in increased reactive oxygen species, senescence-associated β-galactosidase, senescence-associated heterochromatin foci, and a senescence-associated secretory phenotype (SASPs); which promote inflammation, degrade proteins, and leads to tissue dysfunction and further senescence induction to surrounding cells and tissues.²² Furthermore, the delivery of aged and/or senescent cells has been shown to diminish the maintenance, regeneration, and repair of musculoskeletal tissues in vivo.^26,27 Thus, the regenerative potential of MSCs is limited by issues with the conflicting goals of establishing a clinically effective MSC population via in vitro isolation and expansion while concurrently preventing harmful senescence accumulation.

The loss in the therapeutic potential of MSCs due to aging and culture expansion has given rise to the idea that senolytic agents may be used to rescue previously unusable donations by attenuating senescence and producing an enriched stem cell population. Moreover, numerous naturally occurring senolytic agents such as piperlongumine, dasatinib/quercetin, and fisetin have been found to exhibit senolytic activity by targeting anti-apoptotic pathways known to be upregulated in senescent cells.^28-31 Concurrently, several reports have found significant improvement in healthspan indices in vivo following treatment with senolytic agents.³² Our group has demonstrated that fisetin, in particular, may be used to improve age-related cartilage and bone degeneration in a progeria mouse model.³³ These results further validate the use of fisetin as a senotherapeutic that extends health and lifespan in mice.³⁴ In the in vitro setting, fisetin has been shown to attenuate senescence in several cell and tissue types; however, no study has yet demonstrated its effects on senescence in ADSCs,^34,35 a common and easily obtainable source of autologous stem cells.^12,14

With this knowledge, we sought to investigate the effects of fisetin on senescence accumulation in MSCs using culture-expanded human ADSCs. It was hypothesized that fisetin, a senolytic agent and antioxidant, can reduce senescent cell accumulation in ADSCs undergoing culture expansion while maintaining their regenerative potential. Ultimately, positive findings may highlight novel senotherapeutic strategies to improve the efficacy of stem cell-based therapies.

Materials and Methods

Cell Culture and Expansion

Banked adipose-derived stem cells (ADSCs) were purchased from CellTex Therapeutics (CellTex, catalog # N/A) and cultured in growth media containing DMEM-F:12 (Gibco, catalog # 11320033) with 10% Fetal Bovine Serum (FBS; Gibco, catalog # 10437028) and 1% penicillin/streptomycin (P/S; Gibco, catalog # 15140148). Four donors were used with ages ranging from 10 to 80 years of age. All cells were maintained in tissue culture-treated T75 flasks (Corning, catalog # 10-126-11) at 37°C in humidified air containing 5% CO₂. All cells were received at passage 1 and subsequently passaged up to 18 times. At 70%-90% confluency, the cells were passaged and reseeded at a 1:4 dilution.

Determination of Cell Number and Proliferation

Cell number was determined by removing the cells using TrypLE (Gibco, catalog # 12605028) and conducting direct cell counts using the Countess II automated cell counter. The metabolic activity of the cells was quantified by treating cells with the PrestoBlue Cell Viability Reagent (Invitrogen, catalog # A13262) for 2 h and measuring the optical density at 570 nm per manufacturer’s protocol. The growth rate was determined using the following equation:

$${\displaystyle \begin{array}{l}{\displaystyle \frac{\mathrm{l}\mathrm{n}\left(\frac{FinalCellCount}{InitialCellCount}\right)}{NumberofDays}}\end{array}}$$

Reactive Oxygen Species and C₁₂FDG Quantification

Cells positive for reactive oxygen species and senescence-associated β-galactosidase were quantified using flow cytometry. Senescence-associated β-galactosidase was evaluated using the C₁₂FDG substrate which becomes fluorescent when hydrolyzed by SA-β-galactosidase. ADSCs were tested at passages 4, 6, and 8 after being plated in 12-well plates (Genesee, catalog # 25106) in triplicate at a density of 40 000 cells/well. The cells were treated with 100 nM of bafilomycin A1 (Cell Signaling, catalog # 54645) for 1 h at 37°C and 5% CO₂. Next, 33µM C₁₂FDG (Thermo Fisher, catalog # D2893) was added and incubated for 2 h. After 1.5 h the CellROX reagent (Thermo Fisher, catalog # C10422) was also added for the remaining 30 min. The cells were then washed with PBS and collected from the plate using the TrypLE reagent. Using the Guava EasyCyte flow cytometer, cells positive for C₁₂FDG and reactive oxygen species were quantified using unstained cells to set gate thresholds for fluorescence.

Senescence-Associated Heterochromatin Foci Staining

Cells were plated into chamber slides for senescence-associated heterochromatin foci (SAHF) staining at a density of 5000 cells/cm². Cells were seeded in triplicate for each group and incubated to adhere overnight. The following day, cells were fixed with cold 4% paraformaldehyde (Alfa Aeser, Catalog # J19943K2) for 15 min, and washed 3 more times using PBS. The cells were then blocked using 10% donkey serum (DS; Jackson Immuno, catalog # 017000121) and 0.3% Triton X-100 (Fisher, catalog # BP151-500) in PBS for 1 h followed by overnight incubation at 4°C with the γ-H2AX antibody (Millipore Sigma, catalog # 05636I) using a 1:150 dilution and the H3K9me3 antibody (Millipore Sigma, catalog # 07442) using a 1:250 dilution. Both antibodies were added to a dilution buffer containing 1X PBS with 1% DS, 1% bovine serum albumin (BSA; Sigma, catalog # A9647-100G), and 0.3% Triton X-100. Following antibody conjugation, the cells were washed 3 times with wash buffer containing 1X PBS with 0.1% BSA. The cells were then incubated with 1:400 diluted Alexa Flour 488 and 594 (Invitrogen, catalog # A21206 and A21203, respectively) secondary antibodies for 1 h at room temperature. The cells were washed three more times and incubated with 1 µg/mL of DAPI (Sigma, catalog # D9542-10MG) for 10 min before being washed with PBS. Five images were taken per well and 3 wells per group were imaged using the Nikon Eclipse Ni-U microscope. Cells positive for the γ-H2AX and H3K9me3 antibodies were quantified using the ImageJ software.

Fisetin Treatment

Cells were seeded in triplicate for each group using previously described culture methods and seeding densities. After the cells adhered to the flask, the media was changed. For the untreated group, normal ADSC growth media was used (DMEM:F-12 with 10% FBS and 1% penicillin/streptomycin), and for the treated groups 25µM, 50µM, or 100µM of fisetin (Selleckchem, catalog # S2298) was added. Cells were treated for 24-h, after which the wells were washed, and normal growth media was added.

Adipogenic and Osteogenic Differentiation

ADSC differentiation was carried out using the Lonza Adipogenesis and Osteogenesis kits (Lonza, catalog # PT3004 and PT3002, respectively) according to the manufacturer’s instructions. 20 000 cells were seeded into each well of a 24-well plate (Genesee, catalog # 25-107). Cells were grown to confluency using normal growth media. Once all wells reached 90-100% confluency the media was changed to differentiation media. For adipogenic differentiation, the Lonza adipogenic induction media was added for 3 days followed by 2 days with the Lonza adipogenic maintenance media. This process was repeated for a total of 15 days. For osteogenic differentiation, fresh Lonza osteogenic induction media was added every 3 days for a total of 15 days.

Oil Red O Staining

Following adipogenic differentiation, the cells were fixed with cold 4% paraformaldehyde (Alfa Aeser, Catalog #J19943K2) for 15 min and washed twice with water. The cells were then incubated with 60% isopropanol for 5 min followed by incubation with 500 μL of oil red O working solution for 15 min at room temperature. Working solution was made from 3 parts oil red O stock in 2 parts water. The working solution was thoroughly mixed and filtered through Whatman No. 1 filter paper prior to use. Oil red O stock was made from 60 mg oil red O (Sigma Aldrich, catalog # MAK194C) in 20 mL of 100% isopropanol. Following incubation, the cells were washed 3 times with distilled water. The cells were then imaged using the Nikon Eclipse Ni-U microscope. Following imaging the dye was collected by adding 500 μL of isopropanol per well and incubated on a rocker for 15 min. Two hundred microliter of dye from each well was then collected into a 96-well plate and the optical density was measured at 492 nm.

Alizarin Red Staining

Following osteogenic differentiation, the cells were fixed with cold 4% paraformaldehyde (Alfa Aeser, Catalog # J19943K2) for 15 min and washed twice with water. The cells were then incubated at room temperature for 30 min with a 2% alizarin red (Sigma Aldrich, catalog #A5533) solution in water with a pH corrected to 4.1-4.3 using ammonium hydroxide. The cells were then washed 5 times with water and imaged using the Nikon Eclipse Ni-U microscope. To quantify the alizarin red staining the dye was solubilized by incubating the cells for 15 min using a 10% (w/v) cetylpyridinium chloride (CPC) (Sigma-Aldrich, Catalog # C0732) in PBS (pH 7.4) with moderate shaking. The final solutions were added to a 96 well plate and the optical density was measured at 570 nm. The optical density of CPC alone was subtracted from each of the sample measurements.

RNA Isolation, Reverse Transcription, and Quantitative PCR

Total RNA was extracted using the Trizol reagent (Invitrogen, catalog # 15596026) according to the manufacturer’s instructions. Following RNA isolation, reverse transcription was performed using the qScrpit cDNA synthesis kit (VWR, catalog # 95048-100) according to the manufacturer’s instructions. Quantitative real-time PCR was performed using the PerfecCTa SYBR Green FastMix (VWR, catalog # 101414-280) on the Step-One Plus Real-Time PCR system (Applied Biosystems, catalog # 4376598). All results were normalized to the Gapdh reference gene. Primer sequences were as follows: Gapdh F: GGAGCGAGATCCCTCCAAAAT R: GGCTGTTGTCATACTTCTCATGG, PPARγ F: TACTGTCGGTTTCAGAAATGCC R: GTCAGCGGACTCTGGATTCAG, C/EBPα F: ACTCCAGGGGTGAACGGAAT R: CATGGGCGAACTCTTTTTGCT, FABP4 F: ACTGGGCCAGGAATTTGACG R: CTCGTGGAAGTGACGCCTT, Col1a F: AGGGCTCCAACGAGATCGAGATCCG R: TACAGGAAGCAGACAGGGCCAACGTCG, ALP F: GACCTCCTCGGAAGACACTC R: TGAAGGGCTTCTTGTCTGTG, OC F: GGCGCTACCTGTATCAATGG R: GTGGTCAGCCAACTCGTCA.

Statistics

All statistical comparisons and figures were generated using the PRISM 9 analysis software. Triplicates for each donor were used with multiple donors, of different ages, for each experiment. t-Tests were completed for all comparisons between 2 groups. One-way ANOVAs were completed to test for significant differences between 3 or more groups, and 2-way ANOVAs were completed to compare multiple groups over time. Tukey’s honestly significant difference (HSD) post-hoc testing was completed in both scenarios. Values with a P < .05 were considered to be statistically significant.

Results

Adipose-Derived Stem Cells Accumulate Markers of Cellular Senescence During Culture Expansion

Following extended culture expansion of banked human adipose-derived stem cells (ADSCs) one-way ANOVA results found that the growth rate significantly decreased over time (Fig. 1A). This was particularly evident after passage 6 when the growth rate significantly declined when compared to passage 8 (P < .05). This decrease was also found to be significantly different between passages 4 and 8 (P < .01). These changes were mirrored by a significant increase in the doubling time between passages 4-8 and 6-8 (P = .005 and .02, respectively, Fig. 1B). Concurrent with the changes in growth rate, one-way ANOVA results found that the number of cells positive for reactive oxygen species (ROS) was significantly increased with passaging (Fig. 1C). Between passage 4 and 6 a 4.4% ROS increase was observed (P < .0001) followed by a 5.7% increase when compared to passage 8 cells (P = .0008). Furthermore, senescence-associated marker β-galactosidase was upregulated in ADSCs after culture expansion (Fig. 1D). Here, we observed a significant increase between passage 4 and 6 (Δ = 15.48%, P = .018), 4 and 8 (Δ = 43.46%, P = .003), and 6 and 8 (Δ = 27.98%, P < .0001).

Figure 1.

Adipose-derived stem cells lose function and acquire markers of senescence during culture expansion. (A) Growth rate significantly decreases with culture expansion. (B) Doubling time significantly increases with culture expansion. (C) Cells positive for reactive oxygen species increase with culture expansion. (D) Cells positive for the C₁₂FDG marker increase with culture expansion. (*P < .05, **P < .01, ***P < .001, ****P < .0001).

Senescence-associated heterochromatin foci (SAHF) are a common feature during senescence progression; thus we next investigated the changes observed between low and high passage ADSCs (Fig. 2). When comparing passages 4 and 18 ADSCs, we observed that passage 18 presented significantly more cells positive for H3K9me3 than ADSCs at passage 4 (Δ = 31.17%, P = .014, Fig. 2A). Similarly, within the cells positive for H3K9me3, significantly more foci were found within passage 18’s nuclei when compared to passage 4 (Δ = 6.52 foci/nuclei, P = .014, Fig. 2B). When investigating the changes associated with SAHF γ-H2AX an increase in γ-H2AX positive cells at passage 18 cells was observed, albeit not significantly (Δ = 16.00%, P = .078, Fig. 2C). However, within the nuclei positive for γ-H2AX, significantly more foci were observed in ADSCs at passage 18 (Δ = 5.45 foci/nuclei, P = .0043, Fig. 2D) when compared to passage 4 ADSCs.

Figure 2.

Senescence-associated heterochromatin foci increase in high passage ADSC. (A-B) Cells positive for H3K9 and the number of H3K9 foci per nuclei are upregulated with passaging. (C-D) Cells positive for γ-H2AX and the number of γ-H2AX foci per nuclei are upregulated with passaging. (E-F) Representative immunofluorescence images of low and high passage senescence associated heterochromatin foci. (*P < .05, **P < .01, ***P < .001).

Fisetin Attenuates Cellular Senescence in a Dose-Dependent Manner in ADSCs

After observing the accumulation of senescent cells during culture expansion on ADSCs, we next looked to determine the potential use of fisetin in eliminating senescent cells during culture expansion. To determine this, we used the PrestoBlue cell viability reagent to determine the relative number of cells removed following treatment with increasing concentrations of fisetin (Fig. 3A). Following a 24-h treatment and 24-h recovery, we found a significant decrease in cell viability with increasing concentrations of fisetin. To determine if the reduction in cell number corresponded with a reduction in potential senescent cells, we used the C₁₂FDG substrate to quantify the number of cells presenting senescence-associated β-galactosidase (Fig. 3B). These results correlated with the reduction in cell viability indicating that fisetin selectively targeted senescent cells. From these results, we found that the 50 µM fisetin concentration would be optimal as it removed 43.7% of senescent cells while cell viability was only reduced by 17.5% compared to untreated cells (Fig. 3A-3C). Comparing this to the 100 µM concentration, no significant difference was found in the reduction of cells positive for C₁₂FDG when compared to the 50 µM concentration (Δ = 4.7%, P = .126). However, there was a significant difference in the total cell viability when comparing the 50 µM and 100 µM fisetin concentrations (Δ = 29.9%, P < .0001), leading us to believe that 100 µM of fisetin may have also eliminated non-senescent proliferating cells.

Figure 3.

Effect of flavonoids on ADSC cell death and senescence. (A) Fisetin induces cell death with increasing concentration when compared to untreated ADSCs. (B) Fisetin reduces C₁₂FDG positive ADSC populations with increasing concentration. (C) Cell viability is plotted against the corresponding reduction in C₁₂FDG following fisetin treatment (****P < .0001).

Based on the results of Fig. 3 we used 50 µM of fisetin to determine if fisetin would attenuate other markers of senescence at both low and high passage from multiple ADSC samples. Here we found that again 50 µM of fisetin reduced the number of cells with elevated reactive oxygen species by 73.8% in passage 4 cells and 78.4% in passage 10 cells (P < .0001, Fig. 4B). Similarly, the number of cells positive for β-galactosidase, as determined by C₁₂FDG quantification, was reduced by 40.9% in passage 4 cells and 49.0% in passage 10 cells (P < .0001, Fig. 4F).

Figure 4.

Fisetin treatment reduces reactive oxygen species and C₁₂FDG. Cells were treated with 50 µM of Fisetin and tested for ROS and C₁₂FDG using flow cytometry. (A) ROS positive untreated ADSCs compared to treated. (B) ROS reduction relative to untreated at passage 4 (p4) and passage 10 (p10). (E) C₁₂FDG positive untreated ADSCCs compared to treated. (F) C₁₂FDG reduction relative to untreated at p4 and p10. Black peaks in histograms represent unlabeled cells while white and green peaks represent untreated and treated, respectively (****P < .0001).

Following ROS and C₁₂FDG quantification, we determined the number of ADSCs positive for H3K9me3 and γ-H2AX SAHF following treatment with 50µM of fisetin. Here, we again observed a significant decrease with 58.2% and 57.3% fewer cells presenting H3K9me3 SAHF at passages 4 and 18, respectively (P < .001, Fig. 5A). Furthermore, within the cells positive for H3K9me3 foci following treatment with fisetin, the number of foci present was significantly reduced in passages 4 and 18 (Δ = 68.0% P = .0086 and Δ = 73.7% P = .0037, respectively, Fig. 5B). These results were conserved when quantifying the number of γ-H2AX foci. Here we observed 36.3% (P = .0004) and 37.6% (P = .0002) fewer nuclei positive for γ-H2AX foci when compared to untreated controls at passages 4 and 18 respectively (Fig. 5C). Again, among the nuclei presenting γ-H2AX foci a significant reduction in the number of foci/nuclei was observed in both passages 4 and 18 ADSCs after treatment (Δ = 55.5% P = .007 and Δ = 59.4% P = .003, respectively, Fig. 5D).

Figure 5.

Fisetin treatment reduces senescence associated heterochromatin foci. Cells were treated with 50 µM of fisetin and tested for H3K9 and γ-H2AX. (A–B) H3K9 positive cells and H3K9 foci/nuclei are reduced following fisetin treatment. (C-D) γ-H2AX positive cells and γ-H2AX foci/nuclei are reduced following fisetin treatment. (E-H) Representative immunofluorescence images of passage 4 and 18 treated and untreated cells. H3K9 = green, γ-H2AX = red, and nuclei = blue. (**P < .01, ***P < .001).

Fisetin Treatment Maintains Adipose-Derived Stem Cell Adipogenic and Osteogenic Differentiation Capacity

To determine the effects of fisetin on the multipotent differentiation of ADSCs, we performed adipogenic and osteogenic differentiation following treatment with 50 µM of fisetin on low passage (p4) and high passage (p10) cells. We found that both adipogenesis and osteogenesis were largely unaffected by fisetin treatment. Adipogenic transcripts PPARγ, C/EBPα, and FABP4 were all not significantly different following 15 days of adipogenic differentiation when compared to untreated low passage cells (P = .380, .255, and .697, respectively, Fig. 6A–6C). However, a significant decrease in PPARγ expression was noted in high passage cells (P = .045, Fig. 7A). Adipogenic markers C/EBPα and FABP4 were not significantly different after fisetin treatment in high passage cells (P = .289, .739, respectively, Fig. 7B–7D). Similarly, oil red o staining for lipogenesis showed no significant change in lipid vesicle production between fisetin-treated and untreated low or high passage cells (P = .833, 0.637, respectively, Fig. 7E, 7F). The osteogenic potential was also not altered by fisetin treatment as determined by qPCR and alizarin red staining in both low- or high-passage cells. Following 15 days of differentiation, collagen type 1 (Col1a), osteocalcin (OC), and alkaline phosphatase (ALP) were all similarly expressed between treated and untreated low passage cells (P = .357, .358, and .591, respectively, Fig. 6G–6I) and high passage cells (P = .138, .867, and .54, respectively, Fig. 7G–7I). Alizarin red staining was also unaltered between treated and untreated low or high passage cells (P = .076, .711, respectively, Figs. 6J, 7J).

Figure 6.

Fisetin treatment maintains low passage (p4) adipogenic and osteogenic potential. (A-C) Adipogenic markers PPARγ, C/EBPa, and FABP4 are not significantly affected by fisetin treatment. (D) Lipid production is not significantly altered after 50 μM Fisetin treatmen. (E, F) Representative oil red O images following 15 days of differentiation. (G-I) Osteogenic markers Col1a, ALP, and OC are not significantly affected by fisetin treatment. (J) Calcium production is not significantly altered after 50 uM fisetin treatment. (K, L) Representative alizarin red images following 15 days of differentiation.

Figure 7.

Fisetin treatment maintains high passage (p10) Adipogenic and osteogenic potential. (A-C) Adipogenic markers PPARγ, C/EBPa and FABP4 are not significantly affected by fisetin treatment. (D) Lipid production is not significantly altered after 50 μM Fisetin treatment. (E, F) Representative oil red O images following 15 days of differentiation. (G-I) Osteogenic markers Col1a, ALP, and OC are not significantly affected by fisetin treatment. (J) Calcium production is not significantly altered after 50 μM Fisetin treatment. (K, L) Representative alizarin red images following 15 days of differentiation.

Discussion

The regenerative potential of human mesenchymal stem cells shows promise in the field of musculoskeletal tissue regeneration with several reports demonstrating the potential of MSCs to promote musculoskeletal tissue healing and repair.^36,37 However, for many regenerative medicine therapies, MSCs are often cultivated for several passages in vitro to achieve the number of cells required for transplantation.³⁸ It has been demonstrated that culture expansion affects the regenerative potential of adult stem cells.^39,40 Cellular senescence is thought to be a fundamental driver of aging and a major contributor to age-associated decline and loss of physiological reserve.²⁵ Senescent cell accumulation is not only a fundamental property of aging but also promotes several age-related pathologies and can affect a person’s ability to withstand stress, recuperate from injury, and operate at peak mental capacity.^24,25,41 There are, however, drugs and supplements (“senolytics”) that can selectively target senescent cells, reducing their harmful effects, even in older individuals.^29,31 Here, we aimed at characterizing cellular senescence during the expansion of human adipose-derived stem cells (ADSCs) and tested whether treatment with the senolytic fisetin could reduce the accumulation of senescence during culture expansion.

To assess the impacts of culture expansion on senescence accumulation in MSCs and ADSCs alike, we investigated common markers associated with senescence at different passages of ADSCs. Due to the elusive pathophysiology of senescence both in vitro and in vivo, it was known that not all markers referenced in the literature may be present. In previous studies it is been shown that the p53 pathway triggers the cellular senescence process in MSCs,⁴² thereby resulting in the accumulation of non-proliferative senescent MSCs.^22,43,44 Furthermore, the expression of reactive oxygen species, senescence-associated β-galactosidase, DNA damage, nuclear foci such as γ-H2AX, and a complex of growth factors (ie, proteases and cytokines), collectively called the senescence-associated secretory phenotypes (SASPs), have shown to be hallmark markers of senescence.⁴⁵

In the present study, we identified several of these canonical markers of senescence, including reactive oxygen species, senescence-associated β-galactosidase, and senescence-associated heterochromatin foci, to be associated with human ADSCs undergoing extended culture expansion. These findings correlate with previous studies which identified senescence associated changes in ADSCs as a result of age and culture expansion.^19,46,47 Importantly, to our knowledge only one other study exists to demonstrate antioxidant’s effects on attenuating senescence in ADSCs.⁴⁸ Furthermore, the present study is the first of its kind to demonstrate the use of the senolytic agent fisetin to attenuate these changes in ADSCs by demonstrating a reduction in reactive oxygen species, senescence-associated β-galactosidase, and senescence-associated heterochromatin foci. These results also correlate with similar studies using fisetin in other cell and tissue types. These studies have shown the mechanism of action for fisetin to be multi-nodal, targeting senescence associated pathways such as SIRT1,⁴⁹ BCL-2/BCL-X_L,^30,50 HIF-1α,⁵¹ p53/MDM2,^30,52 and AKT^30,53 leading to elimination of senescent cells and reduction in SASP driven inflammation in vitro and in vivo. While our study did not investigate the effects of fisetin on the ADSC’s SASP we believe fisetin treatments would also attenuate the SASP as this has been shown to be true in ex vivo adipose tissue treatments as well as several other tissue types.^30,34,54

These findings correlate with several other groups investigations into the effects of in vitro culture expansion on the accumulation of senescence in ADSCs.^47,55 Since MSCs attain replicative senescence during in vitro culture expansion it is of critical importance to address these challenges and standardize the regulatory criteria as cellular senescence can impair the ability of MSCs to suppress inflammation and/or reduce their therapeutic efficacy in tissue repair. It has been shown that low passage MSCs are the most clinically efficacious.^56,57 The study by Wagner et al. showed that culture expansion of MSCs accumulated gradual changes in the global gene and miRNA expression.⁴⁴ Furthermore, the study findings concluded that the senescent-associated changes that occurred in gene and protein expressions were not only associated with culture expansion but were also observed at the beginning of in vitro culture. Hence, identifying, characterizing, and/or eliminating senescent cells in early MSC cultures can be a critical step to ensuring the best MSC product for cell-based therapy. In the present study, we found this theory to hold true as many markers of senescence were in at passage 4. This may be due to variability in donor age, as both young and aged donor ADSCs were used. However, when stratified for age, there was no significant difference in baseline markers of senescence among the ADSC isolates.

Importantly, the present results also work to further support fisetin’s use as a hit-and-run treatment. Previous research has shown that prolonged treatment of MSCs with fisetin can inhibit osteogenesis.^58,59 Similar findings have also been made for fisetin’s effects on adipogenesis.⁶⁰ However, our study highlights the use of a single treatment of fisetin allowing for the recovery and maintenance of ADSC differentiation capacity at both low and high passage, while also reducing the senescence accumulation during the culture expansion process. These results correlate with previous literature to suggest fisetin is best used as a hit-and-run treatment, reducing the risk of side effects.^34,61,62 Interestingly, it was hypothesized that fisetin may be more efficacious in improving the differentiation capacity of more highly senescent high passage ADSC populations; however, this was not evident in the present study. In fact a slight decrease in adipogenic differentiation was noted with a significant decrease in PPARγ expression after fisetin treatment in high passage cells. The reason for this is likely multifactorial but may be in part due to activation of SIRT-1 and its deacetylation of PPARγ in the early stages of adipogenesis.⁶³ Further development of this research is required and remains to be validated in vivo, but these results are promising in that through the suppression of senescence accumulation and increased adipogenesis a reduction in the inflammatory mediators of senescence may be elicited.

Targeting and safely eliminating senescent cells with senolytic agents is now widely studied as cellular senescence is emerging as the conduit to a multitude of age-related disease conditions such as osteoarthritis,²⁷ Alzheimer’s disease,⁶⁴ obesity,⁶⁵ pulmonary fibrosis,⁶⁶ and chronic kidney disease.⁶⁷ Senolytics are a class of drugs that selectively clear or eliminate senescence by inducing apoptosis in senescent cells. In particular, dasatinib, quercetin, fisetin, and navitoclax have all been shown to eliminate senescence in many preclinical models by delaying or alleviating frailty, cancer, and cardiovascular disease.^34,68,69

Of the numerous senolytic agents discovered to date, fisetin is of particular interest due to its established antioxidant, anticarcinogenic, anti-inflammatory, and apoptotic qualities.⁷⁰ Fisetin is a bioactive flavonoid molecule found in fruits and vegetables such as cucumber, apple, grape, and onion, with the highest concentration being found in strawberries.⁷⁰ Aged C57BL/6 mice treated orally at 22-24 months with 100 mg/kg fisetin for 5 days showed a reduction in senescent cells in white adipose tissue.³⁴ Additionally, fisetin treatment of mice, at 85 weeks of age, significantly prolonged the lifespan of these mice by an additional 3 months. This research has given rise to the Alleviation by Fisetin of Frailty, Inflammation and Related Measures in Older Adults (AFFIRM-LITE) clinical trial. Currently, in the recruiting phase, the trial hopes to recruit 40 participants between the ages of 70 and 90 to take an oral 2-day dose of placebo or fisetin at 20 mg/kg/day with analysis focusing on markers of frailty, inflammation, insulin resistance, and bone metabolism.⁷¹

In conclusion, typically, freshly isolated stem cell numbers are limited, and it is necessary to expand their populations in vitro before clinical use. To examine the characteristics and safety of long-term cultured MSCs, the current study evaluated the effects of long-term culture on MSC senescence accumulation. These results indicate that MSCs undergo replicative senescence during long-term culture in vitro, as demonstrated by decreases in growth rate with concurrent increases in reactive oxygen species, senescence-associated β-galactosidase, and senescence-associated heterochromatin foci. However, in all cells, baseline levels of all markers were noted, indicating that the quality of MSC preparations should be carefully assessed before clinical application. To mitigate these markers, this study also demonstrated the use of single treatments of the senolytic agent fisetin in removing senescent cell populations in both low and high passage preparations without affecting their multipotent differentiation potential; providing an important insight into enriching freshly isolated orthobiologics, such as stem cells or bone marrow concentrate, by treating the biologic product with fisetin before banking or administration to the patient. While the current results are promising for future improvements to stem cell therapies, further research will be necessary to understand the mechanisms that regulate the replicative senescence of stem cells and how to ensure MSCs remain in a senescent-free state during culture expansion.

Funding

This study was generously supported through a philanthropic donation from the Borgen family and the Linda and Mitch Hart Foundation.

Conflict of Interest

J.H. declared annual royalty payments from Cook Myosite, Inc. All of the other authors declared no potential conflicts of interest for the current study. All authors are or were paid employees of the non-profit Steadman Philippon Research Institute (SPRI). SPRI exercises special care to identify any financial interests or relationships related to research conducted here. During the past calendar year, SPRI has received grant funding or in-kind donations from Arthrex, DJO, MLB, Ossur, Siemens, Smith & Nephew, XTRE, and philanthropy. These funding sources provided no support for the work presented in this manuscript unless otherwise noted.

Author Contributions

M.M.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript. A.L.N.: collection and assembly of data, data analysis and interpretation, and manuscript writing. A.G.: manuscript writing. J.B., J.M., C.H.: collection of data.H.K.: data analysis and interpretation, manuscript writing. W.S.H.: conception and design, manuscript writing. S.R.: conception and design, administrative support, manuscript writing. J.H.: conception and design, financial support, manuscript writing, and final approval of manuscript.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.