Effects of Fisetin Treatment on Cellular Senescence of the Brain and Other Multiple Organs of Old Sheep

1. Introduction

Fisetin (3, 7, 3', 4' tetrahydroxyflavone) is a natural flavonoid that exists in strawberries, cucumber, apples, onions and so on [1](36558979). Early study by Pamela Maher et al showed that fisetin activated ERK and induced cAMP response element-binding protein (CREB) phosphorylation in rat hippocampal slices, facilitated long-term potentiation in rat hippocampal slices, and enhanced object recognition in mice[2]. Fisetin also possesses neurotrophic effects by promoting rat cortical neurons cell survival and generating long neurites via promoting proteasome activity not through activation of pERK or increase glutathione[3,4]. Cumulative evidence of this research team indicate fisetin is a novel neuroprotective and cognition-enhancing molecule. Fisetin not only has direct antioxidant activity, but also can increase the intracellular levels of glutathione (GSH), the major intracellular antioxidant. Fisetin maintains mitochondrial function in the presence of oxidative stress. In addition, it has anti-inflammatory activity against microglial cells and inhibits the activity of 5-lipoxygenase, and consequently reducing the production of lipid peroxides and their pro-inflammatory by-products. These wide ranges of actions suggest that fisetin has potential to reduce the age-related decline in brain function[5]. Fisetin can increase GSH level mainly by activating transcription factor 4 (ATF4) under basal conditions while activating both ATF 4 and NF-E2-related factor 2 (NRF2) under oxidative stress[6]. In animal model, oral administration of fisetin to APPswe/PS1dE9 double transgenic Alzheimer’s disease (AD) mice from 3 to 12 months of age prevented the development of learning and memory deficits. This correlated with an increase in ERK phosphorylation and a decrease in protein carbonylation, a marker of oxidative stress. In addition, fisetin decreased the levels of p25, the cyclin-dependent kinase 5 (Cdk5) activator p35 cleavage product, in both control and AD brains. It is known elevated levels of p25 relative to p35 cause dysregulation of Cdk5 activity leading to neuroinflammation and neurodegeneration. These effects were also mediated by its anti-inflammatory effects, including changes in eicosanoid synthesis, and the maintenance of markers of synaptic function in the AD mice [7]. Fisetin was also found to reduce cognitive deficits in old senescence-accelerated prone 8 (SAMP8) mice, a model for sporadic AD and dementia, while restoring multiple markers associated with impaired synaptic function, stress, and inflammation [8]. Fisetin and other flavonoids also function as mitochondrial uncouplers to mitigate neurodegeneration in aged C. elegans, possibly via a PINK1/Parkin mitophagy process [9]. Further, fisetin boosted mental health in the aged animals. Supplementing fisetin (oral 20mg/kg/BW/day) for four weeks improved relative electroencephalograph α-power, β-power, and multi-unit activity (MUA) count in aged rats. Fisetin treated aged rats also showed significantly improved cognitive and behavioral performances than non-treated aged rats [10]. Further, fisetin can inhibit aggregation of the tau fragment, K18, and can disaggregate tau K18 filaments in vitro and prevent the formation of tau aggregates in cells [11].

Fisetin served as caloric restriction mimetics. Fisetin, administrated (15mg/kg b.w., orally) to young D-gal induced aged (D-gal 500mg/kg b.w subcutaneously) or naturally aged rats for 6 weeks significantly decreased the level of pro-oxidants and increased the level of antioxidants. Furthermore, fisetin also ameliorated mitochondrial membrane depolarization, apoptotic cell death and impairments in the activities of synaptosomal membrane-bound ion transporters in aged rat brain. Fisetin up-regulated the expression of autophagy genes (Atg-3 and Beclin-1), sirtuin-1 and neuronal markers (NSE and Ngb), and down-regulated the inflammatory (IL-1β and TNF-α) and Sirt-2 genes mRNA respectively in aged rat brain [11]. Same treatment regimen also suppresses aging-induced increases in the levels of reactive oxygen species, eryptosis, lipid peroxidation, and protein oxidation in rat erythrocytes [12]. Fisetin prevented the D-gal-mediated reactive oxygen species (ROS) accumulation, by regulating the endogenous antioxidant mechanisms, such as Sirt1/Nrf2 signaling, suppressed the activated p-JNK/NF-kB pathway, and its downstream targets, such as inflammatory cytokines[13]. Fisetin enhanced phosphorylation and nuclear translocation of Nrf2, which subsequently activated antioxidant enzyme heme oxygenase-1 (HO-1) in RPE cells, and therefore contributing to the amelioration of oxidative stress-induced ocular disorders [14]. Similar effect was observed in C2C12 myoblasts cells [15].

More recently, fisetin was found to be a potent senolytic agent. Acute or intermittent treatment of progeroid and old mice with fisetin reduced senescence markers in multiple tissues. Fisetin reduced senescence in a subset of cells in murine and human adipose tissue. Administration of fisetin to wild-type aged mice restored tissue homeostasis, reduced age-related pathology, and extended median and maximum lifespan [16]. Senescent cells accumulate with ageing and are resistant to apoptosis and have up-regulation of anti-apoptotic pathways which defend them against their own inflammatory senescence-associated secretory phenotype (SASP), allowing them to survive, but killing neighboring cells. Because senescent cells take weeks to reaccumulate, senolytics can be administered intermittently - a 'hit-and-run' approach [17]. Fisetin pretreatment or treatment of old mice after COVID19 viral infection significantly reduced mortality, cellular senescence, and inflammatory markers and increased antiviral antibodies [18]. There may be crosstalk between fisetin’s anti-cell senescent and neuroprotection[19]. Fisetin treatment also has been shown to restore muscle stem cells function in muscular dystrophy mice or progeria mice by targeting senescent cells in the diseased or aged muscle tissues [20,21]. Our recent work showed fisetin selectively attenuated markers of senescence in a dose-dependent manner while maintaining the differentiation potential of the expanded human adipose stem cells in vitro [22].

Fisetin also suppressed renal tubular senescence and attenuated renal fibrosis and improved tubular repair, as indicated by restoration of tubular regeneration and renal function[23].. Fisetin treatment reduced renal fibrosis by inhibiting the phosphorylation of SMAD3, oxidative damage, inflammation, apoptotic cell death, and accumulation of profibrotic M2 macrophages in the obstructed kidneys. In cultured human proximal tubular cells, fisetin treatment inhibited TGF-β1-induced phosphorylation of SMAD3 and SMAD2[24].

Fisetin also demonstrated the potent cardioprotective efficacy against Ang-II induced apoptosis in H9c2 cells and in spontaneous hypertension rat model. Fisetin treatment significantly reduced the apoptotic nuclei number and apoptotic proteins such as TNF- α, Fas L, FADD, Cleaved caspase-3 and Cleaved PARP and increased the cell survival and anti-apoptotic proteins like Bcl-2, Bcl-x_L, p-IGF1R, p-PI3K and p-AKT in both in vitro and in vivo models[25]. Further, fisetin treatment reduced both the atherosclerosis plaque and the lipid accumulation in the aortic sinus in apoE-/- mice fed with a high-fat diet, and the expressions of PCSK9, LOX-1 and aging markers, including p53, p21 and p16 were downregulated [26]. In vascular smooth muscle cell (VSMC) cells, fisetin inhibited cellular senescence induced by the PTEN-PKCδ-NOX1-ROS signaling pathway after hydrogen peroxide (H2O2) treatment, and this anti-aging effect was attributed to reduced ROS production caused by suppressing NOX1 activation[27].

Taken together, fisetin has been shown broad beneficial effects on age related disease. However, majority of the previous studies used mice. Therefore, the goal of this study is to take advantage of large animal model’s translational value and investigate whether systemic fisetin treatment of aged sheep would have beneficial to multiple organs via targeting cell senescent cells.

2. Materials and Methods

2.1. Animal use ethics

This study was approved by Institutional animal care and use committee (IACUC) of Colorado State University (#3159). 8 years old Nordic sheep of both male and females were divided into two groups (N=6/group), fisetin group received fisetin at 100mg/kg in DMSO and 10 mM in ethanol by intravenous infusion two days every week for 8 weeks, the control group received vehicle intravenous infusion using same regimen. Sheep were sacrificed at 8 weeks following the protocol and several organ tissues were harvested and dissected for histology or Q-PCR analysis.

2.2. Histology

Brain tissues were harvested after 8 weeks, left hemisphere were dissected and fixed in neutral buffered formalin for 9 days and then midbrain was sliced into three different level and processed using 15% sucrose 30% sucrose in PBS until tissues sink, and embedded in NEG freezing medium. Cerebellum was also harvested and sliced in the middle and embedded in NEG freezing medium after 9 days of fixation. Furthermore, hippocampus was dissected (after 6 months fixation) and processed using 15% sucrose and 30% sucrose and embedding in freezing medium and snap frozen in liquid nitrogen. Cerebral brain cortex (mid brain level), cerebellum and hippocampus sections were cut into 8µm thickness using Leica Cryostat for histology and immunofluorescent staining. The right hemispheres were used to collect fresh cerebral brain cortex and frozen in -80^oC for isolation RNA and real time quantitative polymerase chain reaction (Q-PCR). H&E staining were performed using AnaTech extra-strength Hematoxylin and eosin Y in 75% alcohol following the manufactures protocol.

2.3. Senescent associated β-galactosidase (SA-β-Gal) staining

Brain cortex and cerebellum cryosections were used to perform SA-β-Gal staining using Senescence beta-galactosidase Staining kits (#9860, Cell Signaling Technology). The staining procedure followed the manufacture’s protocol. Blue staining indicated positive cells. Both intense blue staining and light blue that can be distinguished from surrounding cells were counted positive. Positive cells were counted using Image J. Positive cells number in 200X microscope field was used to compare between groups. Hippocampus tissues SA-β-Gal staining did not work due to long time fixation before dissecting for making cryosections.

2.4. Immunofluorescent staining.

To detect which cells become senescent in the brain tissues, immunofluorescent (IF) staining of P16ink4A (P16) and colocalized with astrocytes markers glial fibrillary acidic protein (GFAP), neuronal marker NeuN and microglia marker ionized calcium-binding adaptor molecule 1(IBA1) was performed for the brain cortex, and dissected hippocampus. IF was performed using the same principle and procedure as previously described[28]. Briefly, sections slides were dried at room temperature and refixed in neutral buffered formalin for 8 minutes to avoid section come off during antigen retrieval. Heat retrieval in pH=6.0 citrate buffer at 92^oC water bath were performed and cool 10 minutes. Slides were washed with PBS for three times and blocking with 5% donkey serum at ambient temperature for 1 hour. Then primary antibodies diluted in 5% donkey serum were added on the sections and incubated at 4⁰C overnight. The primary antibodies dilutions are rabbit anti-GFAP (ab68428, Abcam, 1:100), rabbit anti-IBA1(10904-1-AP, Proteintech 1:200 dilution), rabbit anti-NeuN (ab177487, 1:2000), mouse anti-P16 (SC1661, Santa Cruz Biotechnology,1:50). For dissected hippocampus staining, we used mouse anti-P16 from Thermofisher Scientific (MA5-17142) because the anti-P16 antibody from Santa-Cruz Biotechnology did not work for hippocampus sections. After primary antibody incubation, slides were washed with PBS for three times. Then sections were incubated with donkey anti-rabbit-594 (711-585-152, Jackson ImmunoReseach Laboratory Inc.) and donkey anti-mouse 488 (715-545-150, Jackson ImmunoResearch Laboratory Inc.) at 1:200 dilution and incubated at ambient temperature for 2 hrs. After secondary antibody incubation, slides were washed three times and nuclear were stained with 4′,6-diamidino-2-phenylindole (DAPI, Thermofisher Invitrogen) at 1ug/ml in PBS for 10 minutes. Slides were finally washed three times with PBS and three times of deionized water and mounted with VectaMount® AQ Aqueous Mounting Medium (H-5501-60, Vector Laboratories). Immunofluorescent images were captured using Nikon Ti microscope at 200X magnification and cell numbers were quantified with Image J software.

2.5. Brain cortex, heart, spleen tissue and bone marrow harvest

Brain cortex tissues from the same location were cut and snap frozen in liquid nitrogen and subsequently store in -80^oC for later RNA isolation. Left ventricle muscle of the heart was cut at the same location and snap frozen in liquid nitrogen and stored in -80^oC for isolation of RNA and Q-PCR analysis. Spleen tissues from the same location of each sheep were cut and frozen in -80^oC for RNA extraction. Briefly, heart muscle or spleen tissues were rinsed with PBS and similar size around 0.5g of tissues were transferred to 2 ml tube and cut with scissors to around 1mm size, and then were homogenized with IKA-T10 tissue dispenser in 2ml TRIzol™ Reagent (15596026, Themofisher Scientific) until no visible tissue chunk left. Tissues were homogenized using 30 second homogenization and then 30 second on ice. After homogenization, tissues lysates were centrifuged at 12000g for 5 minutes to remove fat or tissue debris. Supernatant were transferred to a new tube for RNA isolation following the protocol by manufacturer. Furthermore, bone marrow was aspirated from sheep iliac crest and nucleated cells were isolated using gradient centrifugation using Percoll gradient reagent and nucleated cells were used for RNA extraction using Trizol reagent as indicated below.

2.6. RNA isolation, cDNA synthesis and semi-quantitative and Quantitative -PCR (Q-PCR).

All RNAs were isolated using the protocol provided by the manufacturer (#15596026,Thermofisher Scientific). Total RNA was air dried for 7 minutes and resuspended in DNase/RNase free water using appropriate volume based on the RNA pellet size and allowed RNA to dissolve for 10 minutes at room temperature. RNA concentrations were measured with Nanodrop plate equipped in Tecan Plate Reader (Switzerland). First strand cDNA synthesis was then conducted using iScript™ Reverse Transcription Supermix, 100 x 20 µl rxns, 400 µl (1708841, BioRad) in 20ul reaction using 25^oC 5 minutes, 46^oC 30 minutes and 95^oC for 1 minute. After cDNA synthesis, cDNA was diluted using DNase/RNase free H2O to 10 ng/µl. Subsequently semi-quantitative or Q-PCR was performed. For the brain tissues, we performed both semi-quantitative-PCR and Q-PCR. Semi-quantitative PCR was performed using GoTaq® G2 Master Mixes(M7823). PCR products were electrophoresed using 1.5% agarose gel (BioRad) and gel images were captured using ChemiDoc MP System (BioRad). Housekeeping gene and target band density were quantified with Image Lab software (BioRad). Q-PCR was performed using SsoAdvanced Universal SYBR® Green Supermix, 1,000 x 20 µl rxns, 10 ml (10 x 1 ml) (1725272, BioRad) using 10µl reaction using 1µl diluted cDNA. Q-PCR was conducted on ABI system. The analyzed genes included Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), GFAP, lysomal β-galactosidase (GLB1), super oxide dismutase 1 (SOD1), catalase (CAT), neurofilament heavy chain (NEFH), neurofilament light chain (NEFL), interleukin 10 (IL10), interleukin 8 (IL8), P21, P53, NOD-, LRR- and pyrin domain-containing protein 3(NLRP3), Triggering Receptor Expressed On Myeloid Cells 2 (TREM2). Gene expression levels were expressed as fold change of control (2^{-Delta-Delra CT}) methods. All primers were designed using primer 3 input free online software [29,30](22730293, 17379693). Primer information are shown on Table 1:

2.7. Statistical analysis: All data was analyzed with two group t test using Graphpad Prism 9. P<0.05 was considered statistically significant.

3. Results

3.1. Fisetin treatment did not significantly affect sheep brain general morphology.

Gross images showed no obvious differences on the cerebral brain cortex (midbrain) between fisetin treated sheep and vehicle treated sheep (Figure 1A). H&E staining did not reveal significant difference between fisetin treated and vehicle treated sheep brain for the cerebral cortex (Figure 1B), cerebellum (Figure 1C) and different regions of the hippocampus (Figure 1D).

3.2. Fisetin treatment decreased SA- β-Gal positive cells.

We found many SA-β-Gal positive cells in the gray matter of the control sheep’s cerebral brain cortex tissues which indicated the presence of senescent cells at this age of sheep (Blue staining, insets highlighted positive cells (Figure 2A). After Fisetin treatment, we still observed many positive cells. These positive cells are mainly large neurons which have loose large nuclear that not stained. Fisetin treatment showed trend of decreased SA-β-Gal positive cells compared to control group (P=0.1449) (Figure 2B). We also found many SA-β-Gal positive cells in the white matter of cerebral brain cortex, but relative fewer than gray matter. These SA-β-Gal positive cells were also mostly neurons. Fisetin treatment significantly decreased the number of SA-β-Gal positive cells in white matter compared to control (P=0.0306, Figure 2 C-D). Furthermore, we also found SA-β-Gal positive cells in the gray matter of the cerebellum, but not as many as in brain cortex region. These positive cells are also mainly large neurons (Figure 2E). Fisetin treatment showed trend of reduction of SA-β-Gal positive cells compared to control group (P=0.1804, Figure 2F). Further, SA-β-Gal positive cells were also found in the white matter of cerebellum and they are also mainly large neurons (Figure 2G). Fisetin treatment did not significantly decreased the number of SA-β-Gal positive cells in the white matter of cerebellum (Figure 2H).

3.3. Effects of fisetin treatment on the cellular senescence of neurons, astrocytes, and microglial cells of cerebral cortex.

Since we found many senescent cells in the aged sheep both in cerebral brain cortex and cerebellum, we want to identify which cells become cell senescent and if fisetin treatment decreased the senescent cells burden. We performed P16 staining, a hall marker of senescent cells and colocalized with different neuronal cells markers. We first tested P16 and colocalized with NEUN, a neuronal marker. We found many intense NEUN positive red stained cells in the gray matter while P16 staining localized in the nuclei of both neurons and non-neurons. P16 staining in the neurons are relative defused due to the loose nuclei structure of neurons (green channel). Colocalized cells appeared orange color (Figure 3A gray matter). Quantification indicated fisetin treatment significantly decreased P16⁺/NEUN⁺ cells/Total NEUN⁺ cells percentage in the NEUN positive neurons in the brain cortex gray matter (P=0.0079) (Figure 3B). Furthermore, in the white matter, very few NEUN⁺ cells are detected while many small P16⁺ green cells that are not colocalized with NEUN were detected (Figure 3A white matter). Therefore, we counted the NEUN^-P16⁺cells/Total DAPI cells percentage. Fisetin treatment significantly decreased the NEUN^-P16⁺cells/Total DAPI cells percentage in the white matter (P=0.0002) (Figure 3C). We further performed P16 staining and colocalized with astrocytes marker GFAP. GFAP⁺ astrocytes were stained red with many processes with small cytoplasm in the gray matter (Figure 3D gray matter). Fisetin treatment significantly reduced the P16⁺/GFAP⁺/Total GFAP⁺ cells percentage of astrocytes in the gray matter (P=0.0255) (Figure 3E). In the white matter, many more GFAP⁺ cells are detected (Figure 3D white matter). Fisetin treatment significantly decreased P16⁺/GFAP⁺/Total GFAP⁺ cells percentage of astrocytes in the white matter(P=0.0521) (Figure 3F). Additionally, we performed double staining of P16 and colocalized with microglia markers IBA1. IBA1⁺ cells were stained in red and are small cells with some process and some were colocalized with P16 (Figure 3G gray matter). Fisetin treatment significantly decreased P16⁺/IBA1⁺/Total IBA1⁺ cells percentage in the microglia cells (P=0.0407) (Figure 3H). Furthermore, in the white matter, many more IBA1⁺ (red) and P16⁺ (green) small cells with some colocalized were found (Figure.3G white matter). Fisetin treatment significantly reduced the P16⁺/IBA1⁺/Total IBA1⁺ percentage in microglia cells in the white matter (P=0.0004) (Figure 3I).

3.4. Effects of Fisetin treatments on cellular senescent of neurons, astrocytes, and microglial cells of hippocampus.

To further determine if fisetin treatment affects important brain area hippocampus, we dissected hippocampus from the cerebral brain cortex after fixation and processed and cryosection were cut for double immunofluorescent staining. First, we performed NEUN and P16 (MA5 17142, Thermofisher Scientific) double immunofluorescent staining. We choose Thermofisher P16 antibody for all hippocampus staining because SC1661 P16 antibody from Santa-Cruz (SC1661) did not work for hippocampus likely because we dissected the hippocampus after 6 months fixation in NBF after failure to locate hippocampus using whole cerebral brain section. NEUN/P16 double staining showed strong NEUN positive staining in very organized manner in Cornu Ammonis (CA) while large NEUN⁺ cells were scattered in non-CA area of hippocampus (Figure 4A merged images). P16 was stained in green clearly in the nuclei (Figure 4A green channel). Large number of NEUN⁺/P16⁺cells (orange or yellow colocalized cells) were detected in CA area in both control and fisetin group (Figure 4A merged channel). Fisetin treatment did not significantly change P16⁺NEUN⁺/Total NEUN⁺ cells percentage in CA area (Figure 4B). We also performed double staining of P16 and GFAP to detect senescent cells in the astrocytes in hippocampus. GFAP⁺ cells are stained in red color with many red processes and located in non-CA area with some processes penetrat CA area (Figure 4C). Therefore, we count P16⁺/GFAP⁺ cells in the non-CA area. Fisetin treatment showed trend of reduction of P16⁺/GFAP⁺/Total GFAP⁺ cells percentage in the non-CA area of hippocampus (P=0.1076) (Figure 4D). Furthermore, we performed double immunoflurescent staining of IBA1 and P16 to detect microglia cells senescent in the hippocampus. IBA1 was stained in red in small cells with few processes. P16 was stained green in the nuclei. We counted P16⁺/IBA1⁺ cells in the non-CA area because IBA1⁺ cells are not expressed by neurons in CA area (Figure 4E). Fisetin treatment significantly decreased P16⁺IBA1⁺/Total IBA1⁺ cells percentage in the microglia cells in the non-CA area of hippocampus (P=0.0102) (Figure 4F).

3.5. Effects of fisetin treatment on gene expression of sheep brain cortex.

Semi-quantitative-PCR indicated fisetin treatment did not significantly change gene expression of antioxidant genes SOD1 and CAT, astrocytes marker (GFAP), neuronal markers (NEFH, NEFL), SASP markers interleukin 10 and 8 (IL10, IL8) and senescent marker gene galactosidase beta 1 (GLB1) (Figure 5 A-B). We further performed Q-PCR analysis. We found fisetin treatment did not significantly affect SOD1 and CAT expression (Figure 5C). Fisetin treatment did not affect neuronal marker NEFL and NEFH and astrocytes marker GFAP (Figure 5D). Fisetin treatment did not significantly changed GLB1 gene expression as well as P53 expression, but showed a trend of decrease P21 expression (P=0.127). Fisetin treatment did not significantly change SASP gene IL10 and IL8 expression (Figure 5F). Furthermore, fisetin treatment did not significantly affect inflammasome gene NLRP3, but showed a trend of decreasing TREM2, another inflammasome gene (Figure 5G).

3.6. Effect of fisetin treatment on gene expression of heart tissues.

To further investigate if fisetin treatment affects gene expression of other tissues, we harvest ventricle muscle of heart tissues and performed Q-PCR analysis. We found fisetin treatment did not significantly change gene expression of SOD1, CAT in the sheep’s heart ventricle muscle tissues (Figure 6A). Among the senescent related genes, we found fisetin showed a trend of down-regulating GLB1 (P=0.192), but not affecting P21 and P53 expression (Figure 6B). Furthermore, fisetin treatment demonstrated a trend of increase of IL10 (P=0.0695) and IL8 (P=0.0691(Figure 6C). Finally, fisetin treatment showed a trend of up-regulating inflammasome gene NLRP3, with no effect on TREM2 expression of heart tissues (Figure 6D).

3.7. Effects of Fisetin treatment on sheep spleen tissues gene expressions

We also harvested spleen tissues, an important immune organ and extract total RNA for gene expression analysis. We found SOD1 was significantly decreased (P=0.017) and CAT was significantly increased (P=0.053) by fisetin treatment (Figure 7A). We also found fisetin treatment showed a trend of decrease for senescent hall marker gene GLB1 (P=0.083) (Figure 7B). No significant changes were found for P21 and P53 mRNA expression in the spleen tissues (Figure 7B). We did not find significant changes for SASP genes IL10 and IL8 (Figure 7C). Fisetin treatment did not significant change the expression of NLRP3, but demonstrated a trend of increase for TREM2 (P=0.098) (Figure 7D).

3.8. Effects of fiestin treatment on gene expression of bone marrow nucleated cells.

Furthermore, we isolated bone marrow nucleated cells from iliac crest and performed Q-PCR to investigate if fisetin treatment affects bone marrow cell senescent. Fisetin treatment showed a trend of decreasing SOD1, but not affecting CAT expression (Figure 8A). We did not observe significant differences between fisetin treatment and control for GLB1 and P21 expression (Figure 8B). However, we found fisetin treatment significantly increased P53 expression in the bone marrow (P=0.046) (Figure 8C). No statistical significance was found for IL10 and IL8 expression between fisetin treatment and control in bone marrow (Figure 8D). Additionally, fisetin treatment did not change inflammasome gene expression level of NLRP3, but showed a trend of increase in TREM2(P=0.073) (Figure 8E).