Human Umbilical Cord Stem Cells May Improve Memory by Fixing Aged Brain Immune Cells Through Blocking a Specific Cellular Pathway

All experimental procedures were approved by the Animal Ethics Committee of Central South University (permit number: CSU‐2023‐0480). Twelve male C57BL/6 mice (8 weeks, 20.25 ± 0.6 g) were procured from Hunan Slaike Jingda Laboratory Animal Co. LTD (permit number: SCXK (Xiang) 2020‐0019). Mice were raised with free access to food and water in a temperature‐controlled room (23°C ± 2°C) with a normal dark–light (12 h: 12 h) cycle. Until reaching 18 months of age, the mice were fed a standard diet and then randomly allocated into two groups: the old‐vehicle group (n = 6) and the old‐MSC group (n = 6). The old‐MSC group was intravenously injected with hUC‐MSCs (1 × 10⁶ cells/200 μL/mouse) via the caudal vein once a week for four consecutive weeks. Meanwhile, the old‐vehicle group was injected with hUC‐MSCs vehicle (200 μL/mouse) in the same way. Additionally, six male C57BL/6 mice (8 weeks, 20.19 ± 0.8 g) were obtained from the same supplier and served as the young group. The young group received a standard diet without any intervention.

BV2 cells (a mouse microglia cell line) and HT22 cells (a mouse neuron cell line) were purchased from Procell Life Science & Technology Co. Ltd. and cultured in Dulbecco's Modified Eagle Medium (DMEM, Gibco) supplemented with 10% Fetal Bovine Serum (FBS, Invitrogen) in an atmosphere of 5% CO₂ at 37°C.

hUC‐MSCs were supplied by the National Engineering Research Center of Human Stem Cells and the Guangxiu Tech Biotechnology Co. The preparation was carried out in a Good Manufacturing Practice (GMP) laboratory for medical products, as previously reported. Briefly, the master cell bank at passage 1 and the working cell bank at passage 4 were identified by the criteria suggested by the International Society for Cellular Therapy (ISCT): (1) plastic adherent in standard culture conditions using tissue culture flasks; (2) ≥ 95% of the MSC population express CD105, CD73, and CD90 and lack expression of (≤ 2% positive) CD45, CD34, CD11b, CD19, and HLA‐DR as measured by flow cytometry (BD Accuri C6); (3) differentiation potential to osteoblasts, adipocytes, and chondroblasts under standard in vitro differentiating conditions. Moreover, each batch of the master cell bank and the working cell bank was tested for sterility, human virus, G‐banding karyotype analysis, tumorigenicity, immunomodulatory capacity, telomere length, and telomerase activity. In preparation for infusion, frozen UMSCs at passage 4 were quickly thawed, washed, and suspended in normal saline supplemented with human serum albumin. For each MSC product, cell counts and viability were examined using trypan blue staining by a cell counter (Countstar) as product release criteria (Cheng et al. 2022).

hUC‐MSCs possess inherent immunosuppressive properties and low immunogenicity, which enable them to evade a robust host immune response following transplantation into mice (Naji et al. 2019). Our previous work has also demonstrated that hUC‐MSC transplantation in mice does not induce an abnormal immune response; instead, it suppresses excessive immune response, supporting their safety and efficacy in cross‐species applications (Peng et al. 2025).

To establish a cellular model of senescent microglia and neurons, BV2 and HT22 cells were plated in 24‐well plates for 24 h and subsequently treated with hydrogen peroxide (H₂O₂, Merck) at concentrations of 50, 80, 100, 120, and 150 μM for an additional 24 h. Then, the optimal concentration for inducing cellular senescence was determined using BODIPY staining and Senescence‐Associated β‐galactosidase staining. After the senescent cell models were successfully constructed, senescent microglia were co‐cultured with hUC‐MSCs by using a Transwell system (0.4 μm pore) for 24 h. Meanwhile, senescent microglia were treated with 10 μM JSH‐23 (NF‐κB inhibitor) for 24 h to evaluate the role of NF‐κB in lipid droplet‐accumulating microglia.

To investigate whether hUC‐MSCs alleviate senescent microglia‐induced neuronal damage by improving senescent microglial function, we divided the HT22 cells into four groups: (1) ctrl group: HT22 cells without any treatment; (2) H₂O₂ group: senescent neurons group: cells treated with 80 μM H₂O₂ for 24 h; (3) B‐H group: senescent microglia–neuron co‐culture group (BV2‐H₂O₂/HT22‐H₂O₂): BV2 cells were treated with H₂O₂ for 24 h, then co‐cultured with H₂O₂‐induced senescent HT22 neurons for another 24 h; (4) B‐M group: hUC‐MSC‐preconditioned senescent microglia–neuron co‐culture group (BV2‐H₂O₂‐MSC/HT22‐H₂O₂): senescent BV2 cells (H₂O₂‐treated) were co‐cultured with hUC‐MSCs for 24 h, then co‐cultured with senescent HT22 neurons for an additional 24 h.

The T‐maze test was respectively performed on mice before and one week after the final injection of hUC‐MSCs. The T‐maze apparatus consists of two goal arms (22 × 5 × 5 cm) and a stem that is perpendicular to them, measuring 45 cm in length, with the same width and height as the goal arms. The approach alley contains a 5 × 5 cm starting box. At the start of the experiment, the mouse was placed in the stem. The gate was then opened, allowing the mouse to leave the main arm and enter one of the goal arms (defined as all four limbs crossing into the arm). After each entry, the mouse was immediately returned to the main arm and kept there for 5 s. This process was repeated 9 times, with the number of entries into each arm being recorded. The success rate was calculated by dividing the number of alternating choices within the same experimental session by the total number of choices made.

The novel object recognition test (NOR) test was respectively performed on mice before and one week after the final injection of hUC‐MSCs. The NOR apparatus consists of an open‐field box (25 × 25 × 30 cm), two identical objects, and a new object, situated in a quiet room. The NOR test consists of two steps: the familiarization phase and the test phase. During the 5‐min familiarization phase, two identical objects are placed in two non‐opposite corners of the box, equidistant from the periphery. The animal is then placed at the center of the box to explore freely for 5 min. After a 60‐min intertrial interval, the test phase begins, where one of the two identical objects is replaced with a novel object. Each mouse performs the test individually, and the box and objects are cleaned with 75% alcohol after each test. Exploration behavior is defined as the animal turning its nose toward the objects at a distance of ≤ 3 cm or touching them with the nose or forepaw, while passing by, climbing, or sitting on the object is not included. The recognition index (RI) is calculated as the percentage of time spent exploring the novel object over the total time spent exploring both objects.

The Senescence‐Associated β‐galactosidase (SA‐β‐gal) staining kit (Beyotime) was used to detect SA‐β‐gal activity in brain tissues and BV2 or HT22 cells. Brain tissues frozen sections and cells were fixed for 15 min with fixative solution at room temperature, followed by incubation in staining solution overnight at 37°C according to the manufacturer's protocol. Images were captured by Nikon ECLIPSE TS100 microscopy, and the number of SA‐β‐gal‐positive cells was quantified using ImageJ software.

Tissue sections were fixed in 4% paraformaldehyde and embedded in paraffin. 5 μm thick sections of brain tissue were baked in an oven at 65°C for 2 h, deparaffinized with xylene, dehydrated with gradient ethanol, and repaired with citrate repair solution for antigen. Next, peroxidase activity was blocked with 3% H₂O₂ and non‐specific antigens were blocked with 5% donkey serum for 1 h. Then sections were incubated overnight at 4°C with guinea pig anti‐PLIN2 antibody (Fitzgerald LLC, 1:200) and rabbit anti‐IBA1 antibody (Wako, 1:100). After washing with PBS three times, tissue sections were incubated with Alexa Fluor 488‐Donkey Anti‐Rat IgG (H + L) (Abcam, 1:500) and Alexa Fluor 647‐Goat anti‐Guinea Pig IgG (H + L) (ThermoFisher, 1:500) for 1 h. Nuclei were counterstained with DAPI (Sigma, 1:1000) at room temperature for 3 min. Images were captured by ZEISS Axioscope 5 fluorescence microscope.

After euthanasia, the mice were sequentially perfused intracardially with normal saline and 4% paraformaldehyde (PFA). The hippocampal tissues were then immersed in an electron microscopy fixative solution. The tissues were embedded in resin, and ultrathin sections were prepared using a Leica EM UC7 ultramicrotome. The sections were stained with heavy metals (lead and uranium) and imaged with a Hitachi HT7700 transmission electron microscope at 80 kV.

BODIPY is a neutral lipid dye. BV2 cells were seeded in a 24‐well plate at 2 × 10⁴ peer well. Following H₂O₂ treatment and co‐culture with hUC‐MSCs, the cells were fixed in 4% PFA for 30 min at room temperature. After washing with DPBS, the cells were incubated in DMEM containing 2 μM BODIPY 493/503 (ThermoFisher Scientific) at 37°C for 10 min. Subsequently, the nuclei were counterstained with DAPI (Sigma‐Aldrich) at room temperature for 3 min. Fluorescent images were captured using a ZEISS Axioscope5 fluorescence microscope, and the mean fluorescence intensity of BODIPY was analyzed using ImageJ software.

BV2 cells were seeded in a 24‐well plate at 2 × 10⁴ per well. After H₂O₂ treatment and hUC‐MSCs co‐culture, cells were incubated with 100 μg/mL pHRodo Green Zymosan Bioparticles (ThermoFisher Scientific) in DMEM at 37°C for 15 min. The nuclei were counterstained with Hoechst (Beyotime) at 37°C for 5 min. Fluorescent images were acquired using a ZEISS Axioscope5 fluorescence microscope, and the mean fluorescence intensity of zymosan was analyzed using ImageJ software.

RNA sequencing of hippocampal tissues was performed by Beijing Genomics Institute (BGI). Briefly, RNA samples were denatured, and mRNA were enriched using oligo (dT)‐attached magnetic beads. After fragmentation, cDNA synthesis was performed, followed by end‐repair, A‐tailing, and PCR amplification. The final DNA nanoballs (DNBs) were loaded into patterned nanoarrays and sequenced using combinatorial Probe‐Anchor Synthesis (cPAS). The sequencing data was filtered using SOAPnuke, and gene expression levels were calculated with RSEM (v1.3.1). Differential expression analysis was performed using DESeq2 (v1.4.5), and KEGG enrichment analysis was conducted using Phyper based on the Hypergeometric test, with a significance threshold of p value ≤ 0.05.

Hippocampal tissues and BV2 cells were homogenized in lysis buffer (RIPA [Beyotime]: PMSF [BOSTER] 100:1) on ice for 30 min, followed by centrifugation at 12000 rpm for 20 min at 4°C. Protein concentration was measured by BCA Protein Assay Kit (Beyotime). Proteins were denatured by boiling at 100°C for 10 min and then separated by 10% or 12% SDS‐PAGE and transferred onto polyvinylidene fluoride (PVDF, Millipore) membranes. Membranes were blocked with 5% skim milk powder in TBST for 2 h, washed with TBST, and incubated overnight at 4°C with primary antibodies. After washing with TBST, membranes were further incubated with HRP‐conjugated secondary antibodies for 2 h at room temperature. Super Signal Enhancer Chemiluminescent (ECL plus, Nature biosciences) was used to detect chemiluminescence with the Fusion FX5 Chemiluminescence Imaging System.

The following primary and secondary antibodies were used: rabbit anti‐p16 (CST, 1:1000), rabbit anti‐p21 (Abcam, 1:1000), rabbit anti‐γ‐H2A.X (CST, 1:1000), guinea pig anti‐PLIN2 (Fitzgerald LLC, 1:1000), rabbit anti‐SREBP1 (Abcam, 1:1000), rabbit anti‐p65 (CST, 1:1000), rabbit anti‐p‐p65 (CST, 1:1000), and rabbit anti‐β‐actin (Abcam, 1:3000), goat anti‐mouse IgG H&L (Abcam, 1:1000), goat anti‐rabbit IgG H&L (Abcam, 1:1000).

Total RNA from hippocampal tissues and BV2 or HT22 cells was extracted using the TRIzol reagent (Apex BIO). cDNA synthesis was performed with the Go Script Reverse Transcription system (Promega). Real‐time PCR was performed with SYBR GREEN 480 (Roche). β‐actin was used as the reference gene, and gene expression levels were calculated using the 2^−ΔΔCt method. The following primers were used: β‐actin forward primer (5′‐GTGACGTTGACATCCGTAAAGA‐3′) and reverse primer (5′‐GCCGGACTCATCGTACTCC‐3′); Cdkn2a (p16Ink4a) forward primer (5′‐CCCAACGCCCCGAACT‐3′) and reverse primer (5′‐GCAGAAGAGCTGCTACGTGAA‐3′); Cdkn1a (p21Cip1) forward primer (5′‐GTCAGGCTGGTCTGCCTCCG‐3′) and reverse primer (5′‐CGGTCCCGTGGACAGTGAGCAG‐3′); TNF‐α forward primer (5′‐GTCTGTATCCTTCTAACTTA‐3′) and reverse primer (5′‐TCTTGTGTTTCTGAGTAG‐3′); TGF‐β forward primer (5′‐CTGCAAGAGACTTCCATCCAG‐3′) and reverse primer (5′‐TGGTTTTCTCATAGATGGCG‐3′); IL‐6 forward primer (5′‐CTGCAAGAGACTTCCATCCAG‐3′) and reverse primer (5′‐AGTGGTATAGACAGGTCTGTTGG‐3′); Cxcl2 forward primer (5′‐CCAACCACCAGGCTACAGG‐3′) and reverse primer (5′‐GCGTCACACTCAAGCTCTG‐3′).

Data were analyzed using GraphPad Prism 8.0 Software and presented as the mean ± standard deviation. Significant differences between groups were evaluated by one‐way ANOVA following normality and lognormality tests.

To evaluate the therapeutic potential of hUC‐MSCs on the age‐related cognitive decline, hUC‐MSCs were injected intravenously into aged mice (Figure S1a). Cognitive assessments using the T‐maze and NOR tests revealed significant improvements following hUC‐MSCs treatment. Aged mice demonstrated impaired performance in the T‐maze test (46.67% ± 2.89%) compared to young controls (76.67% ± 3.82%; p < 0.05), with hUC‐MSCs treatment restoring performance to youthful levels (73.33% ± 4.16%; p < 0.01 vs. aged group) (Figure 1a). Consistent with these findings, the NOR test revealed a reduced recognition index in aged mice (0.45 ± 0.03 vs. young: 0.67 ± 0.05; p < 0.01), which was significantly improved by hUC‐MSCs administration (0.64 ± 0.04; p < 0.01 vs. aged) (Figure 1b). Notably, total exploration times remained comparable between old and hUC‐MSCs treatment groups, though both were reduced relative to young controls (Figure S1b). Furthermore, the elevated SA‐β‐gal activity was observed in the aged hippocampus. hUC‐MSCs administration effectively attenuated this senescence‐associated phenotype, significantly reducing SA‐β‐gal activity compared to the old group (Figure 1c, Figure S1c).

Consistent with the observed phenotypic changes, the mRNA expression levels of senescence‐associated markers Cdkn2a (p16^INK4a) and Cdkn1a (p21^WAF1) were significantly upregulated in the aged hippocampus compared to young controls, which were decreased after hUC‐MSCs administration (Figure 1d). Western blot analysis confirmed corresponding increases in p16^INK4a (1.24‐fold) and p21^WAF1 (1.2‐fold) protein expression in aged mice, with significant reductions post‐treatment (Figure 1e, Figure S1d). Moreover, the mRNA levels of senescence‐associated secretory phenotype (SASP) markers, such as interleukin‐6 (il‐6), tumor necrosis factor‐α (tnf‐α), transforming growth factor‐β (tgf‐β), and Cxcl2, exhibited parallel changes in the aged hippocampus, all of which were attenuated by hUC‐MSCs treatment (Figure 1f). These findings demonstrate that hUC‐MSCs treatment not only rescues age‐related cognitive deficits but also systemically decreases the expression of senescence‐related markers, including cell cycle arrest and SASP in the aged hippocampus.

To investigate the effects of hUC‐MSCs on hippocampal microglial activation, we assessed microglial morphology using immunofluorescence. Results showed that the number of IBA1‐positive microglia was increased in the aged hippocampus compared to young controls, and these cells exhibited enlarged cell bodies and reduced branching, consistent with activated microglia (Figure 2a, Figure S2a,b). hUC‐MSCs transplantation decreased the number of microglia and restored ramified morphology (Figure 2a, Figure S2a,b). Transmission Electron Microscopy (TEM) ultrastructural analysis revealed that LDs were accumulated in the microglia of the aged hippocampus, which exhibited direct membrane contact with lysosomes (red arrow) (Figure 2b). hUC‐MSCs treatment significantly reduced LD size, with LDs fully encapsulated within lysosomal compartments (red arrow) (Figure 2b). The phospholipid monolayer of LDs is enriched with lipoproteins, among which perilipins (PLIN 1–5) are the most abundant. PLIN2 is widely present in non‐adipose cells and localizes to the LD surface, making it a universal marker for LDs (Conte et al. 2021). To specifically detect the content of LDs in microglia of the aged hippocampus, we performed immunofluorescence staining using PLIN2 (LD marker) and IBA1 (microglial marker). Consistent with TEM observations, PLIN2⁺ LDs were accumulated in IBA1⁺ microglia in the aged hippocampus, which was attenuated by hUC‐MSCs transplantation (Figure 2c). Western blot analysis confirmed the increase of the PLIN2 protein levels in the aged hippocampus, with hUC‐MSCs transplantation reducing the expression (Figure 2d). These results collectively demonstrate that hUC‐MSCs restore microglial quiescence in the aged hippocampus and enhance the lysosomal‐mediated lipid clearance.

To investigate the anti‐senescence effect of hUC‐MSCs on microglia, we established an H₂O₂‐induced senescent microglia model in vitro. Dose–response analysis (50, 80, 100, 120, and 150 μM H₂O₂) revealed cell viability changes: 80–100 μM H₂O₂ induced cell proliferation, while ≥ 120 μM H₂O₂ caused cytotoxicity (Figure S3a). SA‐β‐gal activity revealed dose‐dependent increases from 80 to 100 μM (Figure S3b). Based on these results, 80 μM H₂O₂ was selected for subsequent senescence modeling (Figure 3a). Senescent BV2 cells were co‐cultured with hUC‐MSCs; a significant decrease in SA‐β‐gal activity was observed (Figure 3b). Western blot analysis demonstrated that hUC‐MSCs mediated the suppression of p21 and γ‐H2AX expression compared to senescence controls (Figure 3c, Figure S3c). hUC‐MSCs also downregulated p16 expression and suppressed SASP expression such as il‐1β, il‐6, tgf‐β, and Cxcl2 (Figure 3d,e). These in vitro results confirm that hUC‐MSCs exert potent anti‐senescence effects, effectively counteracting H₂O₂‐induced microglial senescence.

To determine whether hUC‐MSCs restore senescent microglial function, we assessed LD content and phagocytic capacity in microglial senescence in vitro. BODIPY staining demonstrated H₂O₂ induced dose‐dependent LD accumulation (Figure S3d). Senescent microglia were co‐cultured with hUC‐MSCs; a significant decrease in BODIPY fluorescence intensity was observed (Figure 3f,i), indicating reduced LD accumulation. This finding was further confirmed by immunofluorescence and western blot analysis of PLIN2 expression (Figure 3g,j, Figure S3). Phagocytic assay revealed hUC‐MSCs significantly restored senescent microglial impaired phagocytic function (Figure 3h,k). These findings demonstrate that hUC‐MSCs reduce LD accumulation and improve phagocytic capacity in senescent microglia.

Microglia, as the resident immune cells of the central nervous system, are vital for maintaining neural homeostasis. To investigate the detrimental effects of senescent microglia on aging neurons, we established an H₂O₂‐induced neuronal senescence model in vitro and co‐cultured it with senescent microglia (Figure 4a). Senescent neurons exhibited a significant increase in p21 mRNA expression, and this increase was further elevated after co‐culture with senescent microglia (Figure 4b). SA‐β‐gal activity revealed a significant increase in senescent neurons, which remained unaltered by co‐culture with senescent microglia (Figure 4c,d). Notably, co‐culture with senescent microglia markedly upregulated mRNA expression of the proinflammatory factors tnf‐α and il‐6 in senescent neurons compared to the non‐co‐cultured group (Figure 4e). Furthermore, senescent neurons co‐cultured with senescent microglia (B‐H group) showed a trend toward reduced apoptosis compared to senescent neurons alone (H₂O₂ group): B‐H = 11.97% ± 1.198%, H₂O₂ = 14.21% ± 2.544%, p = 0.66 (Figure 4f). These findings suggest that senescent microglia can exacerbate neuronal senescence and neuroinflammation, thereby contributing to age‐related cognitive decline.

To assess whether hUC‐MSCs alleviate senescent microglia‐induced neuronal damage by improving senescent microglial function, we co‐cultured senescent neurons with senescent microglia that had been pre‐cocultured with hUC‐MSCs. Results showed that senescent neurons exhibited a significant decrease in p21 mRNA expression (Figure 4b) and SA‐β‐gal activity after co‐culture with senescent microglia that had been pre‐cocultured with hUC‐MSCs (Figure 4c,d). Meanwhile, senescent microglia pre‐cocultured with hUC‐MSCs suppressed neuroinflammation by reducing the mRNA expression of tnf‐α and il‐6 in senescent neurons (Figure 4e). Senescent neurons also exhibited an increase in apoptosis after co‐culture with senescent microglia that had been pre‐cocultured with hUC‐MSCs (B‐M: 18.70 ± 1.108 vs. B‐H: 11.97 ± 1.198; p = 0.02) (Figure 4f), indicating hUC‐MSC‐preconditioned senescent microglia promoted clearance of senescent neurons. These results indicate that hUC‐MSCs can alleviate senescent microglia‐induced neuronal damage by enhancing senescent microglial function to attenuate neuronal senescence, mitigate neuroinflammation, and promote senescent neuron clearance.

To explore the anti‐aging mechanisms of hUC‐MSCs, we performed RNA sequencing on mouse hippocampus. As shown in Figure 5a, hUC‐MSCs‐treated aged hippocampal samples cluster closer to young samples in principal component analysis (PCA) space, while untreated aged samples are distinct from both groups, indicating partial restoration of the "young‐like" transcriptional profile. Meanwhile, a total of 760 differentially expressed genes (DEGs) were identified in the aged hippocampus compared to young controls, with 429 upregulated and 331 downregulated (Figure 5b). Comparing hUC‐MSC‐treated aged hippocampus (Old‐MSC group) with untreated aged hippocampus (Old‐vehicle group), 261 DEGs were identified, including 162 upregulated and 99 downregulated genes (Figure 5c). GO Biological Processes (BP) enrichment showed that DEGs between untreated aged and young groups (760 DEGs) are enriched in age‐related pathways such as inflammatory response (Figure 5d). Although only the 261 DEGs were identified between hUC‐MSCs‐treated and untreated aged groups, they were significantly enriched in senescence‐associated processes, including immune response, inflammatory response, and lipid metabolism (Figure 5e). Additionally, we also analyzed the DEGs between hUC‐MSCs‐treated aged and young groups, identifying 750 DEGs (518 upregulated, 232 downregulated, Figure 5f). These DEGs were enriched in structure‐related gene networks (e.g., embryonic axis patterning, cytoplasmic organization) and metabolic regulation (e.g., glucose metabolism) (Figure 5g). These findings indicate that hUC‐MSCs shift the aging hippocampus toward a healthier state by regulating senescence‐associated pathways (inflammation, lipid metabolism) but do not fully restore structural or metabolic gene networks. KEGG pathway enrichment analysis revealed that DEGs (Old‐MSC vs. Old‐vehicle) were enriched in cell signaling and inflammation‐related pathways, particularly the NF‐κB signaling pathway (Figure 5h). Evidence indicates that NF‐κB could promote the secretion of inflammatory factors and upregulate the expression of sterol regulatory element‐binding protein 1 (SREBP1) to facilitate the formation of LDs (Yan and Horng 2020). Consistent with this, we found the expression of SREBP1 was significantly increased in aged hippocampus and markedly reduced after hUC‐MSCs transplantation (Figure 5i). Meanwhile, hUC‐MSCs transplantation significantly reduced the p‐p65/p65 ratio in the aged hippocampus, suggesting suppression of NF‐κB pathway activation (Figure 5i). These data suggest that hUC‐MSCs ameliorate age‐related cognitive decline via inhibition of the NF‐κB‐SREBP1 pathway in the aged hippocampus.

To investigate whether hUC‐MSCs attenuate microglial senescence via inhibition of the NF‐κB‐SREBP1 pathway, we employed JSH‐23 to inhibit NF‐κB nuclear translocation in senescent microglia (Figure 6a). As shown in Figure 6b,c, JSH‐23 significantly inhibited NF‐κB nuclear translocation in senescent microglia, with a similar effect observed with hUC‐MSCs treatment. Western blot analysis further revealed that both JSH‐23 and hUC‐MSCs treatment significantly reduced the p‐p65/p65 ratio and SREBP1 protein level in senescent microglia (Figure 6d), indicating hUC‐MSCs inhibited the NF‐κB‐SREBP1 pathway. Consistent with the effects of hUC‐MSCs, JSH‐23 markedly reduced the proportion of SA‐β‐gal positive cell rate (Figure 6e,f) and suppressed the expression of SASP, including il‐6, tnf‐α, tgf‐β, and Cxcl2 in senescent microglia (Figure 6g). These findings suggest that hUC‐MSCs attenuate microglial senescence through inhibition of the NF‐κB‐SREBP1 axis.

BODIPY staining, PLIN2 immunofluorescence, and oil red O staining demonstrated that JSH‐23 significantly reduced LD accumulation in senescent microglia (Figure 6h,i, Figure S4a–c), recapitulating the LD‐reducing effects observed with hUC‐MSC co‐culture. Additionally, JSH‐23 also significantly enhanced the phagocytic capacity in senescent microglia (Figure 6j). These data demonstrate that hUC‐MSCs reduce LD accumulation and enhance phagocytic capacity in senescent microglia via inhibition of the NF‐κB‐SREBP1 pathway.