iMSC-secreted factors preserve renal capillary networks and ameliorate fibrosis by interrupting the macrophage STING/CD8⁺ T cell axis in ischemic kidney disease

Abstract

Background

Renal fibrosis, the common endpoint of chronic kidney disease, is exacerbated by renal ischemia. While human iPSC-derived MSCs (iMSCs) hold therapeutic promise, their role in modulating the ischemic renal microenvironment, particularly through macrophage-immune crosstalk, remains unclear. This study assesses whether iMSCs alleviate renal fibrosis by targeting macrophage-STING to restore capillaries.

Methods

The protective effects of iMSC-conditioned medium (iMSC-CM) were assessed in H₂O₂‑induced HK‑2 and HUVEC models using CCK‑8, ROS staining, scratch assay, and tube formation assays. A mouse model of unilateral ischemia‑reperfusion (UIR) was established; mice received intravenous iMSCs or PBS. Kidneys were analyzed by histology, immunohistochemistry, qPCR, and Western blot. Transcriptomic sequencing revealed enrichment of immune-related pathways, including the STING pathway. To validate the role of STING, UIR mice were treated with the STING agonist DMXAA with or without iMSCs, and STING activation was assessed in LPS-stimulated macrophages in vitro. Macrophage polarization was evaluated by F4/80 co‑staining with CD86 (M1) or CD206 (M2).

Results

iMSC-CM mitigated H₂O₂-induced damage in HK-2 cells by promoting proliferation, reducing ROS levels, and suppressing fibrosis markers, and it promoted repair and angiogenesis in HUVECs. In vivo, iMSCs homed to injured kidneys, attenuated inflammation and fibrosis, notably reduced capillary rarefaction. Mechanistically, iMSC treatment profoundly inhibited STING activation in renal macrophages. This suppression disrupted a critical inflammatory axis: it reduced macrophage-derived IFN-β, leading to decreased infiltration of cytotoxic CD8⁺ T cells, which are detrimental to vascular endothelial cells. Consequently, the renal capillary network was preserved. Finally, STING agonist treatment abolished the anti-fibrotic benefits of iMSCs.

Conclusion

Collectively, this work reveals that iMSCs ameliorate renal fibrosis via the macrophage STING/CD8⁺ T cell axis, thereby preserving renal capillaries and highlighting a novel mechanism for iMSC-based therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-026-05099-x.

Background

Kidney disease is becoming a major global health burden and is projected to be the fifth leading cause of death by 2040. Globally, 850 million people have kidney disease, with approximately 2 million deaths annually from acute kidney injury (AKI), and 1.2 million deaths from chronic kidney disease (CKD) in 2017 [1, 2]. Given its high disease burden, leading nephrology societies have jointly called on the WHO to list kidney disease as a priority non‑communicable disease for prevention and control [3]. For most chronic kidney injury, renal fibrosis is regarded as the core pathological process and the ultimate common pathway of disease progression, characterized by persistent inflammatory cell infiltration, tubular atrophy, and progressive capillary rarefaction. At present, new antifibrotic agents such as pirfenidone have shown some efficacy in pulmonary fibrosis, but their therapeutic effects in renal fibrosis are very limited [4, 5]. Effective therapies for fibrosis in the kidney, liver, and heart remain lacking [6]. Accordingly, two urgent goals in kidney disease research are to decipher the molecular basis of renal fibrosis and to discover effective therapeutic interventions.

Mesenchymal stem cells (MSCs) are self-renewing, multipotent stromal cells with therapeutic potential for various diseases, including kidney injury. However, more research is needed to determine which source of MSC is best suited for therapeutic applications. The efficacy of MSCs varies depending on the source. Liu et al. found that BMSCs displayed higher EV yields than ASCs at earlier passages, while ASC-EVs were larger than BMSCs‐EVs at later passages [7]. Ali et al. reported that L-MSCs were superior to BM-MSCs in improving liver function, antioxidant status, and anti-fibrotic activity [8]. However, MSCs from adult tissues have limitations. Despite their broad tissue distribution, the limited availability or invasive harvesting of many tissue-specific MSCs restricts their clinical application [9]. The supply of MSCs from perinatal tissues (e.g., cord, amnion) is restricted, limiting large-scale production. Primary MSCs from aged donors also suffer from poor quality, diminished therapeutic efficacy, and rapid loss of expansion capacity and plasticity over passages. Additionally, MSCs derived from different donors and tissues show inherent heterogeneity, complicating standardized manufacturing. A solution to these limitations lies in the use of human iPSC-MSCs as a scalable, unlimited cell source for biomanufacturing. Li et al. reported that iMSCs offer an alternative MSC source with superior proliferation, enhanced paracrine support, and stronger immune regulation compared to parental MSCs [10]. iMSCs are readily available and address the heterogeneity of MSCs [11], and have therapeutic potential in the regeneration of blood vessels and periodontium, bone and cartilage, nerves, heart and liver [12–19].

The microenvironment of renal fibrosis is a complex network of multiple immune cells [20]. Macrophages are central regulators of innate immunity whose polarization state shapes the inflammatory response. Classically activated macrophages (M1 type) amplify the inflammatory response by releasing high concentrations of pro-inflammatory mediators (e.g., TNF-α, IL-1β, CCL2) while recruiting additional immune cells. In contrast, M2 (alternatively activated) macrophages balance the inflammatory response by exerting anti-inflammatory effects and participating in tissue repair. In the chronic ischemic microenvironment of the kidney, macrophages are often polarized into a pro-inflammatory M1 phenotype, promoting the occurrence and development of inflammation [21]. This dysregulation of innate immunity further activates adaptive immunity. Among the key events in this cascade are the recruitment and subsequent activation of CD8⁺ T cells at the lesion. Once activated, CD8⁺ T cells directly damage vulnerable renal tubular epithelial cells and vascular endothelial cells by releasing granzymes and effector cytokines (e.g., IFN-γ) [22–25]. Recent studies have revealed a detrimental positive feedback loop involving macrophages and CD8⁺ T cells: inflammatory mediators derived from macrophages promote functional initiation and chemotactic aggregation of CD8⁺ T cells, which in turn secrete IFN-γ, further reinforcing the M1‑dominant pro‑inflammatory state of macrophages [26–28]. This vicious cycle is the core obstacle that leads to persistent exacerbation of tissue damage and failure of vascular repair.

It is speculated that there is an upstream “master switch” that can be precisely regulated by iMSCs to break this vicious circle. Recent research has focused on STING, a key molecule in the innate immune signaling pathway. STING is a cytosolic DNA receptor that can be activated by abnormally present autologous or foreign DNA in the cytoplasm during cellular stress or injury. Upon activation, STING induces type I interferons (e.g., IFN‑β) and numerous pro‑inflammatory factors, serving as pivotal links between tissue damage and immune inflammation [29]. Studies have shown that activating the STING-IFNβ axis leads to more infiltration of CD8⁺ T cells, which synergizes with androgen deprivation therapy to produce the mechanism of anti-tumor immunity [30]. This provides us with a perspective on the possibility that iMSCs may target and inhibit the overactivation of STING in macrophages through their paracrine effects, thereby cutting off this harmful “immune-vascular” axis at the source. In this work, we found that STING signaling is involved in renal fibrosis.

Our results show that the STING‑TBK1‑IRF3 axis is upregulated after renal ischemia-reperfusion injury in C57BL/6 mice, and that inhibiting STING alleviates fibrosis. This effect is achieved by iMSCs via targeting the macrophage STING/CD8⁺ T cell axis, interrupting the harmful crosstalk that leads to capillary rarefaction. In summary, the macrophage-STING-CD8⁺ T cell axis can be used as a potential therapeutic intervention point for renal fibrosis.

Materials and methods

iMSC culture and conditioned medium preparation

Human iMSCs were obtained from Nuwacell Ltd. (Hefei, China) and maintained in ncMission hMSCs Medium. For conditioned medium (CM) collection, cells at passage 5 were grown to 90% confluence. The old medium was aspirated, and add fresh medium for another 24 h. The supernatant was then collected, centrifuged at 1000×g for 10 min to remove cell debris, and frozen at -80 °C for later use.

Flow cytometry

iMSCs were harvested and stained with fluorochrome‑conjugated antibodies: anti‑CD105‑APC, anti‑CD90‑PE, anti‑CD73‑APC, anti‑HLA‑DR‑FITC, anti‑CD34‑FITC, and anti‑CD45‑PE (all from BioLegend, USA; 0.25 µg per 10⁶ cells). The cells were incubated for 30 min in the dark, then washed with PBS (or corresponding buffer) and resuspended. Fluorescence was measured on a flow cytometer. Data were analyzed by FlowJo 10.8.

iMSCs differentiation

iMSC differentiation toward adipogenic, chondrogenic, and osteogenic lineages was assessed using manufacturer‑supplied protocols (Gibco, Thermo Fisher Scientific). After 21 or 28 days of culture, differentiation was detected by lineage‑specific staining: Oil Red O for adipocytes, Alcian Blue for chondrocytes, and Alizarin Red for osteoblasts (kits from Solarbio, China, and Cyagen Biosciences Inc., USA).

Cells culture and construction of cells model

Human renal proximal tubular epithelial cells (HK-2) and RAW264.7 macrophages were purchased from Procell (China) and maintained in their specific media. Human umbilical vein endothelial cells (HUVECs) from the same supplier were cultured in DMEM supplemented with 10% fetal bovine serum (FBS, Excell Bio, China) and 1% penicillin/streptomycin (Solarbio, China). All cells were incubated at 37 °C with 5% CO₂, with medium changed every other day and subcultured every three days.

Cellular injury was induced by exposing HK-2 and HUVECs to 100 µM H₂O₂ for 24 h, while RAW264.7 cells were stimulated with 20 ng/mL LPS (Sigma, USA) or 1 µg/mL DMXAA (Invivogen, France). Subsequently, injured HK-2/HUVECs and stimulated RAW264.7 cells were treated with iMSC-CM mixed 1:1 with regular culture medium.

Assessment of intracellular ROS levels

Intracellular ROS were detected in HK-2 and RAW264.7 cells using a commercially available ROS detection kit (Beyotime, China). Following 24 h of the indicated treatments, the cells were loaded with 2 mL of 0.1% DCFH-DA for 30 min at 37 °C. Rosup (50 µg/mL) served as a positive control. After loading, the cells were rinsed with serum-free RPMI 1640 (Gibco, USA) and visualized under a fluorescence microscope.

Cell scratch assay

HK-2 and HUVECs were seeded in 6-well plates and cultured until cell fusion. Subsequently, cells were either exposed to H₂O₂ or left untreated. Then, the damaged area was drawn on the cell monolayer with a sterile pipette tip. After washing the scratched monolayers three times with 1× PBS, the cells were cultured in medium containing 50% iMSC-CM, 10% FBS, or PBS, respectively, and incubated for 24 h. Wound closure was monitored by capturing images under a light microscope (Leica, Germany) at 0, 6, 12, and 24 h post-scratching. The distance between the wound edges (healing distance, in mm) was measured at each time point using ImageJ software. Three random fields per well were analyzed, and each experiment was repeated independently three times (n = 3).

Morphological assessment

HUVECs subjected to the respective treatments were fixed with 4% PFA after 24 h of H₂O₂ exposure, then stained with Giemsa (Solarbio, China) per the manufacturer’s protocol and examined by light microscopy.

Tube formation experiment

Matrigel (150 µL/well) was added to a pre‑cooled 48‑well plate and allowed to solidify. Treated HUVECs were seeded onto the gel and incubated for 6 h at 37 °C. Capillary‑like structures were imaged, and five random fields per well were analyzed using ImageJ. The following parameters were quantified: number of nodes, total tube length (µm), junctions, and branches. Experiments were performed in triplicate with three replicate wells per group, and repeated three times independently (n = 3).

Cell viability assay

HK-2 cells were seeded at 8,000 cells per well in 96-well plates and cultured in 100 µL of HK-2 medium containing 10% FBS. After exposure to H₂O₂ or iMSC-CM for 24 h, 10 µL of CCK‑8 reagent (Bimake, Netherlands) was pipetted into each well. Following 3 h of incubation, viable cells reduced the CCK-8 substrate to a colored formazan product, and the absorbance was measured by a microplate reader.

RT-qPCR analysis

For kidney tissue samples, approximately 50 mg of tissue was taken, and magnetic beads were added to crush using a tissue grinder, and homogenized with TRIzol reagent (Sigma, USA). In the cell sample, 1 × 10⁶ cells were mixed directly with TRIzol reagent to lyse the cells. Total RNA was isolated according to the manufacturer’s operating instructions and its concentration and purity were determined. The cDNA was then prepared using a reverse transcription kit using 1 µg RNA as a template (Transgen Bio, China). Quantitative PCR was carried out with SYBR Green master mix (Transgen Bio, China) under the following thermal profile: 94 °C for 30 s, then 40 cycles (94 °C/5 s, 60 °C/15 s, 72 °C/10 s). Ct values were read at the plateau phase. The relative expression value was calculated using the 2⁻^ΔΔCt formula, and β-actin was used as an internal reference for correction. See Supplementary Table S1 and S2 for primer sequences.

Animals experiments

Male C57BL/6 mice with a body weight of 20 ± 2 g were raised under specific pathogen-free conditions (24 °C, relative humidity 40–60%), and the mice were freely fed and watered. After a one‑week acclimation, unilateral ischemia‑reperfusion (UIR) injury was induced under isoflurane anesthesia. The core body temperature of mice was controlled at 38 ± 0.2 °C using a heating pad. A right flank incision was made, and blood flow through the renal pedicle was blocked by a microvascular clip for 30 min, as evidenced by the kidney turning dark purple. The clip was then removed to restore renal perfusion, confirmed by the return of normal kidney color. The incision was closed in layers. The mice in the sham operation group underwent the same surgical steps as the experimental group except that the renal pedicle was not clamped.

Immediately after reperfusion, iMSCs (1 × 10⁶ cells/100 µL PBS) were injected into a tail vein to administer treatment. Control mice received an equal volume of PBS. Based on transcriptomic findings, additional groups were included to validate STING involvement: UIR + DMXAA (STING agonist, 25 mg/kg, intraperitoneal injection on the 1st, 3rd, and 5th days after surgery) and UIR + iMSCs + DMXAA (same iMSC and DMXAA regimen). Thus, a total of five groups (n = 5 each) were used: sham, UIR + PBS, UIR + iMSCs, UIR + DMXAA, and UIR + iMSCs + DMXAA. At 14 days postoperatively, mice were deeply anesthetized with isoflurane and euthanized via carbon dioxide asphyxiation, after which the kidneys were isolated for subsequent experimental analyses. A 14‑day endpoint was selected based on our preliminary time‑course study showing established fibrosis and capillary rarefaction by this time, consistent with previous UIR fibrosis models.

The animal experiments in this study have passed the ethical review by the Animal Care and Use Committee of the Beijing Institute of Radiation Medicine (IACUC-DWZX-2024-P647).

In vivo biodistribution imaging

C57BL/6 mice aged 6–8 weeks were subjected to ischemia-reperfusion surgery, then assigned to receive tail vein injection of either DiO-labeled iMSCs or PBS vehicle (n = 2/group). iMSCs were pre-labeled with 10 µM DiO at 37 °C for 30 min, and 1 × 10⁶ labeled cells were injected per mouse. Prior to injection, mice were anesthetized with 3–5% isoflurane in oxygen (0.4–0.6 L/min) for 2 min until loss of righting reflex. Fourteen days after surgery, the mice were euthanized, and the distribution of DiO-labeled iMSCs in liver, spleen, heart, lung, kidney and other organs was observed using a liver imaging system (InVivo FX PRO, Bruker) and a probe confocal laser endoscope (pCLE; Cellvizio, Mauna Kea Technologies). For quantitative analysis, the average fluorescence intensity (radiance efficiency, p/s/cm²/sr) was measured in each organ using the imaging system’s software. Individual values from each mouse are shown as dots, with the mean indicated by a horizontal line (n = 2 per group). A control mouse subjected to the same UIR procedure but injected with unlabeled iMSCs was used to determine background fluorescence.

Histology and staining

Fourteen days after surgery, five mice per group were euthanized, and kidneys were removed aseptically. Fixation was performed by incubating tissues in 4% paraformaldehyde for ≥ 48 h, followed by dehydration in a graded ethanol series, paraffin embedding, and sectioning at 3 μm. Serial sections were then stained with H&E, PAS, and Masson. Additional sections were processed for F4/80 immunohistochemistry (Abcam, USA). Light microscopy was used to examine and capture images of all sections.

For quantitative analysis, tubular injury was assessed on PAS‑stained sections by evaluating the percentage of tubular necrosis, brush border loss, cast formation, and luminal dilation in five randomly selected fields per section (three sections per mouse). Collagen deposition on Masson‑stained sections was quantified as the percentage of positive area per field using ImageJ software. F4/80‑positive cells were counted in at least five fields per section at 100× magnification. All quantifications were performed blinded to the experimental groups.

Immunofluorescence

After dewaxing and permeabilization, the paraffin-embedded kidney sections were sealed with a blocking solution at room temperature for 1 h, and then incubated overnight with the primary antibody at 4 °C. The following antibodies were used: rabbit anti‑CD86, anti‑CD206, anti‑CD31, and anti‑CD8 (all from Cell Signaling Technology, USA). For macrophage co‑localization, sections were double‑stained with rat anti‑F4/80 (Abcam, USA) together with rabbit anti‑CD86 or anti‑CD206. The corresponding fluorescent secondary antibody (goat against rabbit or goat against rat, Proteintech, China) was applied to the washed sections and incubated for 1 h at room temperature and protected from light, followed by capping, and images were acquired using fluorescence microscopy.

Quantification was performed using ImageJ software. For each marker, a fixed intensity threshold was applied uniformly across all images. Five random fields per section were analyzed, and three non-adjacent sections per mouse (spaced ≥ 50 μm) were examined. All quantifications were carried out in a blinded manner (n = 5 mice per group). CD8⁺ cells were counted manually and expressed as cells/mm². CD31⁺ area fraction (%) was measured using the ImageJ area measurement tool. For double-stained sections, the percentage of CD86⁺ or CD206⁺ cells among F4/80⁺ macrophages was calculated based on manual cell counting.

Western blots

Cells (1 × 10⁶) or kidney tissue (50 mg) were lysed in RIPA/1% PMSF (Solarbio). Equal protein loads were resolved, blotted onto PVDF membrane, and probed with antibodies. The following antibodies were used: FN, vimentin, STING, p‑STING, IRF3, p‑IRF3, TBK1, p‑TBK1 (Cell Signaling Technology, USA); p65, p‑p65 (Abcam, USA); COL1 (Proteintech, China); β‑actin (ABclonal, China); CD86, CD206 (Cell Signaling Technology, USA); IFN‑β (Proteintech, China); and GAPDH (ABclonal, China). After HRP‑conjugated secondary antibody incubation, blots were developed using ECL (Thermo Fisher Scientific, China) and quantified with Scion Image (GenScript).

Transcriptome sequencing

Total RNA was isolated from mouse kidney tissues using MagZol Reagent (Magen, China). RNA quality (A260/A280 ratio and RIN) was evaluated on a NanoDrop spectrophotometer. High‑quality RNA samples were used to construct a paired‑end library with the ABclonal mRNA Seq Lib Prep Kit (ABclonal, China) according to the manufacturer’s instructions. Sequencing was performed on a NovaSeq 6000 system (PE150 read length), and bioinformatic analysis was conducted on the Illumina platform. All transcriptome data processing was carried out by Shanghai Zhongke New Life Biotechnology Co., Ltd.

Multiplex immunofluorescence (mIHC)

After treatment, RAW264.7 cells on coverslips were fixed (4% PFA, 15 min), permeabilized (0.2% Triton X-100, 10 min), blocked (3% BSA, 30 min), and quenched (3% H₂O₂, 15 min). Multiplex staining was performed using the Goat Anti-Mouse/Rabbit Multiplex IHC Detection Kit (Zen-Bioscience, China) according to the manufacturer’s instructions. Cells were first incubated with rabbit anti‑iNOS antibody (1:250) overnight at 4 °C, followed by HRP-conjugated secondary antibody (60 min, room temperature) and TSA-570 dye (1:500 with Enhancer, 10–15 min). After signal development, antibody complexes were stripped by microwave antigen retrieval in sodium citrate buffer (pH 6.0). The same procedure was repeated for mouse anti-CD206 antibody (1:200) using TSA-520 dye. Finally, TRITC‑phalloidin (Solarbio, China) was applied for 30 min and DAPI for 10 min to stain the cells. Coverslips were mounted and imaged using a confocal microscope (Leica, Germany).

Statistical analyses

The data are shown as the mean ± SD. Statistical significance between two groups was evaluated using the unpaired Student’s t-test, whereas comparisons among multiple groups were assessed by two-way ANOVA with Tukey’s post-hoc test. GraphPad Prism 10.0 software was used for all statistical calculations. A P value of less than 0.05 was taken to indicate a statistically significant difference.

Results

Characterization of iMSCs

iMSCs exhibited a typical spindle‑shaped, fibroblast‑like morphology and adhered to culture dishes. Their trilineage differentiation potential was confirmed by lineage‑specific staining: Alizarin Red for osteogenesis (mineralized nodules), Oil Red O for adipogenesis (lipid vacuoles), and Alcian Blue for chondrogenesis (proteoglycan synthesis) (Fig. 1A). Flow cytometry showed that > 98% of iMSCs expressed CD73, CD90, and CD105, whereas the hematopoietic markers CD45, CD11b, and HLA‑DR were detected in < 2% of cells, satisfying the ISCT criteria for MSC identification (Fig. 1B).

Fig. 1.

Characterization of iMSCs. A Differentiation of iMSCs toward osteogenic (Alizarin Red, calcium deposits), chondrogenic (Alcian Blue, proteoglycans), and adipogenic (Oil Red O, lipid vacuoles) lineages (10×; scale bar: 100 μm). B Immunophenotyping of iMSCs via flow cytometry (n = 3)

iMSC-CM attenuate H₂O₂-induced HK-2 cell damage

During the development of renal interstitial fibrosis, renal proximal tubular epithelial cells are the first cell type to sustain hypoxic and oxidative damage. In this study, 100 µM H₂O₂ was applied to create an in vitro oxidative injury model, as this concentration recapitulates pathological oxidative stress conditions and is widely applied in tubular injury research. A CCK‑8 assay showed that 100 µM H₂O₂ reduced HK‑2 cell viability by over 50% relative to untreated controls (Fig. 2A). We further investigated the paracrine protective effects of iMSC-derived conditioned medium (iMSC-CM) on H₂O₂-injured HK-2 cells by evaluating cell viability, migratory capacity, intracellular ROS generation, and the expression of fibrosis-associated genes. As shown in Fig. 2B, cell viability was markedly decreased by H₂O₂ relative to control conditions (P < 0.01), whereas iMSC-CM restored viability to near-control levels (P < 0.01 vs. H₂O₂ group). As shown in Fig. 2C, D and H₂O₂ significantly increased ROS production in HK-2 cells, whereas iMSC-CM markedly suppressed it. Wound healing was significantly impaired by H₂O₂, and iMSC-CM treatment effectively restored it (Fig. 2E and F). As shown in Fig. 2G and I, H₂O₂ significantly increased the mRNA levels of the fibrosis markers FN, α‑SMA, and COL1 in HK-2 cells, whereas iMSC‑CM treatment markedly reduced them.

Fig. 2.

iMSC-CM mitigated HK-2 cell damage. A Viability of HK-2 cells after exposure to increasing H₂O₂ concentrations (25–200 µM) and iMSC‑CM treatment, measured by CCK‑8 assay at 24 h. B HK-2 cells were treated with H₂O₂ (100 µM) followed by iMSC‑CM, and cell viability was assessed after 24 h using the CCK‑8 assay. C Staining for ROS was performed to assess oxidative stress levels from different groups (10×, scale bar: 200 μm). D Quantification of ROS⁺ cells per field. E Representative images of scratch wound healing in HK-2 cells treated with iMSC-CM or H₂O₂, taken at 0, 6, 12, and 24 h (10× magnification; scale bar: 200 μm). F Scratch wound healing ability of HK-2 cells in each group. G-I mRNA levels of COL1, FN, and α-SMA in HK-2 cells among three groups (n = 4 in each group). All data are mean ± SD. **p < 0.01,***p < 0.001,****p < 0.0001

iMSCs restore vascular damage

In CKD, reduced peritubular capillaries limit blood delivery to tubular epithelial cells, making PTCs a potential therapeutic target [31]. Giemsa staining showed that H₂O₂ treatment markedly altered HUVEC morphology, whereas iMSC-CM effectively restored it (Fig. 3A). Scratch assays demonstrated that H₂O₂ impaired wound healing, whereas iMSC‑CM accelerated wound closure, as shown by representative images (Fig. 3C) and time‑course analysis of wound closure percentage (Fig. 3D). Tube formation assay on Matrigel revealed that iMSC‑CM enhanced the angiogenic capacity of HUVECs. Representative images are shown in Fig. 3B. Quantitative analyses indicated that iMSC‑CM significantly increased the number of nodes (Fig. 3E), total tube length (Fig. 3F), number of junctions (Fig. 3G), and number of branches (Fig. 3H) compared to the H₂O₂‑treated group.

Fig. 3.

Protective effect of iMSC-CM on HUVECs. A HUVEC morphology. Cells were exposed to H₂O₂ in the presence or absence of iMSC‑CM and stained with Giemsa (20×, scale bar: 100 μm). B Tube formation assay. HUVECs were seeded on Matrigel with conditioned medium. Representative images at 6 h (10×; scale bar: 200 μm). C Scratch assay. Representative micrographs taken at 0, 6, 12, and 24 h after wounding (10×, scale bar: 200 μm). D Quantification of cell migration. Migration distance was calculated as the reduction in wound width from 0 to 24 h. Data are mean ± SD (n = 3). E-H Quantification of tube formation parameters: (E) number of nodes per field, (F) total tube length per field (µm), (G) number of junctions per field, (H) number of branches per field. All data are mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

iMSCs ameliorated UIR induced renal fibrosis in mice

We established a UIR mouse model of renal fibrosis to test the therapeutic efficacy of iMSCs. The biodistribution of iMSCs was examined by labeling cells with DiO. Representative fluorescence images showed that DiO‑labeled iMSCs accumulated in the lung, liver, and spleen, and preferentially homed to the ischemic kidney compared to the contralateral healthy kidney (Fig. 4A). Quantitative analysis confirmed significantly higher fluorescence intensity in the ischemic kidney (Fig. 4B). PAS staining revealed marked tubular injury (brush border loss, epithelial degeneration, and vacuolization) in UIR kidneys, and Masson staining showed extensive collagen deposition; iMSC treatment significantly attenuated both tubular damage and collagen accumulation (Fig. 4C-E). iMSCs also reduced KIM‑1 mRNA expression, a marker of proximal tubular injury (Fig. 4F). Furthermore, iMSCs downregulated the mRNA expression of α‑SMA, vimentin, TGF‑β, and fibronectin (FN) (Fig. 4G-J). Western blot showed that iMSCs decreased FN, collagen I, and vimentin expression at the protein level (Fig. 4K-N).

Fig. 4.

iMSCs reduce renal fibrosis in vivo. A Biodistribution of DiO-labeled iMSCs after tail vein injection. Representative fluorescence images of lung, heart, liver, spleen, and kidneys from two UIR mice injected with DiO-iMSCs and one control mouse without DiO (n = 2 per group). The ischemic kidney (right) shows stronger fluorescence than the contralateral kidney. B Quantification of organ fluorescence intensity. Data are individual values (dots) with mean (horizontal line); n = 2 per group. C Representative PAS - and Masson-stained kidney sections from sham, UIR + PBS, and UIR+iMSC groups (50×; scale bar: 100 μm). D Quantification of PAS staining: percentages of tubular necrosis, brush border loss, cast formation, and luminal dilation per field. E Quantification of Masson staining: collagen area percentage per field. F Relative mRNA expression of kidney injury marker KIM-1, as measured by RT-qPCR. G–J mRNA levels of α-SMA, vimentin, TGF‑β, and FN in kidney tissues. K FN, COL1, and vimentin were analyzed by Western blot, with β-actin as a loading control. L-N Quantification of FN, COL1, and vimentin band densities. All data are mean ± SD (n = 4 per group). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

iMSCs repressed renal inflammation in renal fibrosis mice

The inflammatory response serves a critical role in both the progression and resolution of UIR-induced renal injury. To evaluate the effect of iMSCs on renal inflammation, we first performed H&E staining to assess overall inflammatory cell infiltration. As shown in Fig. 5A, UIR kidneys exhibited extensive interstitial inflammatory cell accumulation, which was markedly reduced by iMSC treatment. Immunohistochemistry for F4/80 revealed that iMSCs significantly decreased the number of infiltrating macrophages (Fig. 5B, quantified in Fig. 5C). To further characterize macrophage polarization, triple immunofluorescence staining for F4/80, CD86, and CD206 was performed (Fig. 5D). Quantification showed that iMSC treatment significantly reduced the number of CD86⁺ F4/80⁺ macrophages (Fig. 5E) and increased the number of CD206⁺ F4/80⁺ macrophages (Fig. 5F). Additionally, qPCR detection showed that iMSCs markedly decreased the transcriptional levels of pro-inflammatory cytokines (IL-1β, IFN-γ, TNF-α, CCL2) in kidney tissues (Fig. 5G–J). Collectively, these data indicate that iMSCs suppress macrophage infiltration, shift the balance from M1 toward M2 polarization, and reduce the expression of pro‑inflammatory mediators in the ischemic kidney.

Fig. 5.

Intravenous injection of iMSCs reduced the inflammatory response of ischemia-reperfusion of the kidney. A H&E staining of kidney Sect.  (100×, scale bar: 50 μm). B F4/80 immunohistochemistry (100×, scale bar: 50 μm). C Quantitative analysis of F4/80-positive area (%). D Triple immunofluorescence staining of F4/80 (green), CD86 (red), and CD206 (purple) in renal tissues (100×, scale bar: 50 μm). E Percentage of CD86⁺F4/80⁺ cells among total F4/80⁺ macrophages (%). F Percentage of CD206⁺F4/80⁺ cells among total F4/80⁺ macrophages (%). G-J mRNA expression of IL-1β, IFN-γ, TNF-α, and CCL2 in different groups via qPCR. All data are mean ± SD. **p < 0.01, ***p < 0.001, ****p < 0.0001

iMSCs drive immune remodeling, improve angiogenesis, and inhibit CD8⁺ T cell-mediated cytotoxicity

Immunofluorescence analysis was performed to evaluate T cell infiltration and vascular density. As shown in Fig. 6A, UIR kidneys exhibited abundant CD8⁺ T cell infiltration, which was markedly reduced by iMSC treatment. Quantification confirmed that iMSCs significantly decreased the number of CD8⁺ cells per field (Fig. 6C). In parallel, CD31 immunostaining revealed sparse capillary networks in UIR kidneys, whereas iMSC treatment restored vessel density (Fig. 6B), with a significant increase in CD31 fluorescence intensity (Fig. 6D). As shown in Fig. 6E, CD8⁺ T cell density correlated negatively with CD31⁺ vessel area (r = -0.9013, P < 0.01) and positively with CD86⁺ M1 macrophage infiltration (r = 0.9489, P < 0.01) (Fig. 6F). These data suggest that CD8⁺ T cells may contribute to microvascular rarefaction and are closely associated with M1 macrophage accumulation.

Fig. 6.

iMSCs reduce CD8⁺ T cell infiltration and protect microvascular density in UIR kidneys. A CD8⁺ cells / mm² (150×, scale bar: 40 μm). B Immunofluorescence staining of CD31 (vascular endothelial marker) in renal tissues (100×, scale bar: 50 μm). C Quantitative percentage of CD8⁺ T cells (%). D Quantification of CD31 fluorescence intensity. E Correlation analysis between CD8^⁺ cell density and CD31⁺ vessel area. Each dot represents one animal (n = 5 per group, total n = 10). Pearson correlation coefficient. r = -0.9013, P < 0.0001. F Correlation analysis between CD8⁺ cell density and CD86⁺ M1 macrophage infiltration. r = 0.9489, P < 0.0001. All data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Transcriptomics reveal immune homeostasis disorders in renal fibrosis

To investigate the underlying mechanism through which iMSCs exert renoprotective effects against ischemia-reperfusion injury-induced renal fibrosis, RNA-seq analysis was conducted on renal tissues obtained from sham-operated and UIR-treated mice. A total of 1,760 differentially expressed genes (DEGs) were screened out (Fig. 7A). Distinct gene expression patterns between the two groups were visualized using volcano plots (Fig. 7B).

Fig. 7.

Transcriptomic analysis and qPCR validation of key pathways. A Distribution of differentially expressed genes in Sham and UIR renal samples (n = 3). B Volcano analysis of differentially expressed genes in UIR versus sham groups. C Top 20 enriched KEGG pathways with statistical significance (P < 0.05). D Functional annotation of top 30 GO entries in three major modules. E-N qPCR validation of selected genes in kidney tissues from sham, UIR + PBS, and UIR+iMSC groups (n = 4 per group). Expression levels of STING (E), TBK1 (F), IRF3 (G), IFN-β (H), CXCL10 (I), CCL5 (J), iNOS (K), CD86 (L), Arg-1 (M), and CD206 (N). All data are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

The DEGs were significantly enriched in immune‑associated pathways (P < 0.05) as determined by KEGG analysis. These included chemokine signaling, cell adhesion molecules, NOD‑like receptor signaling, cytokine‑cytokine receptor interaction, NK cell‑mediated cytotoxicity, and antigen processing/presentation (Fig. 7C).

The DEGs were enriched in biological processes linked to immunity, including immune system process, immune response, and its regulation, as indicated by GO analysis; cellular components including membrane, cell periphery, and plasma membrane; and molecular functions such as signaling receptor binding, molecular transducer activity, and immune receptor activity (Fig. 7D). Collectively, our observations underscore the importance of immune dysregulation in driving UIR‑mediated renal fibrosis.

Based on the transcriptomic findings, we further validated key genes involved in STING signaling, T‑cell recruitment, and macrophage polarization by qPCR. As shown in Fig. 7E–N, iMSC treatment significantly diminished the gene expression of STING, TBK1, IRF3, IFN‑β (STING pathway), CXCL10 and CCL5 (T‑cell chemoattractants), together with the M1-specific markers iNOS and CD86. In contrast, iMSC treatment led to increased expression of the M2 markers Arg1 and CD206. These results indicate that iMSCs suppress STING‑mediated inflammation, reduce T‑cell recruitment, and shift macrophage polarization from M1 toward M2 in the ischemic kidney.

iMSCs suppress oxidative stress and inflammation in macrophages by inhibiting the STING pathway, thereby reducing macrophage‑mediated chemotaxis of CD8⁺ T cells

RNA-seq results revealed that disrupted immune homeostasis dominates the progression of UIR-induced renal fibrosis. Notably, the STING–IFN-β signaling axis was markedly activated in ischemic renal tissues, accompanied by abnormal infiltration and functional disorder of CD8⁺ T cells. Macrophages act as pivotal upstream initiators of this immune cascade. Therefore, we further investigated whether iMSCs regulate macrophage inflammatory properties and subsequent STING/IFN-β-mediated CD8⁺ T cell dysfunction under inflammatory conditions in vitro.

As shown in Fig. 8A, LPS stimulation triggered excessive ROS accumulation in RAW264.7 cells, while iMSC-CM treatment significantly alleviated oxidative stress. Multiplex immunofluorescence showed that LPS promoted M1 pro-inflammatory macrophage polarization (elevated iNOS) and inhibited M2 anti-inflammatory phenotype (reduced CD206). In contrast, iMSC-CM effectively reversed this imbalance and redirected macrophages toward the M2 phenotype. Importantly, the STING agonist DMXAA largely abrogated the above protective effects of iMSC-CM, verifying the critical function of STING signaling within this process (Fig. 8B).

Fig. 8.

iMSCs suppress STING signaling in RAW macrophages. A ROS staining to assess intracellular oxidative stress in RAW cells (20×, scale bar: 200 μm). B Representative multiplexed immunohistochemistry (mIHC) staining images of five groups of RAW264.7 DAPI, TRITC Phalloidin, iNOS, and CD206 in individual and merged channels are shown (240×, scale bar: 20 μm). C Relative expression of total and phosphorylated STING, TBK1, IRF3 and p65 in RAW cells. D-J Western blot analysis of STING pathway components and macrophage polarization markers in RAW264.7 cells (n = 3 in each group). All data are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Western blot analysis and quantitative quantification (Fig. 8C-J) demonstrated that LPS stimulation strongly activated STING signaling, characterized by increased phosphorylation of STING, TBK1 and IRF3, followed by robust IFN-β production. Treatment with iMSC-CM markedly suppressed STING cascade activation and decreased downstream IFN-β secretion. Of note, DMXAA-mediated STING activation successfully restored STING cascade activity and IFN-β expression, thereby counteracting the immunomodulatory effects of iMSC-CM.

In summary, these in vitro findings demonstrate that iMSCs inhibit macrophage overactivation via suppressing the STING-IFN-β signaling axis. The restraint of macrophage-derived IFN-β further contributes to the regulation of renal CD8 + T cell immune homeostasis, which provides an upstream immune mechanism for iMSC to alleviate UIR-induced renal fibrosis.

iMSC alleviates renal fibrosis and inhibits the STING–IFN‑β signaling axis in vivo dependent on STING activation

To further verify the underlying mechanism by which iMSC regulates renal fibrosis via the STING–IFN-β axis in vivo, Masson staining and Western blot analysis were performed in sham, UIR, UIR+iMSC, UIR+DMXAA, and UIR+DMXAA+iMSC groups at 14 days after surgery.

Masson staining results (Fig. 9A) showed that extensive collagen deposition and obvious interstitial fibrosis were observed in the UIR group compared with the sham group. Consistent with the in vitro therapeutic effect, iMSC treatment significantly reduced fibrotic area and collagen accumulation in UIR kidneys. However, pharmacological activation of STING by DMXAA markedly aggravated renal interstitial fibrosis and collagen deposition. Moreover, the protective anti-fibrotic effects of iMSC were largely abolished by DMXAA intervention, indicating that STING activation mediates the pro-fibrotic progression of UIR injury.

Fig. 9.

iMSC ameliorates UIR-induced renal fibrosis via inhibiting the STING-IFN‑β signaling axis in vivo. A Masson staining for renal sections in Sham and each intervention group (100×, scale bar: 50 μm). B Western blot detection of pivotal molecules related to STING pathway and renal fibrosis. C-K Densitometric quantification of relative protein levels (n = 3 in each group). All data are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Western blot analysis (Fig. 9B) and quantitative statistical results (Fig. 9C-K) further confirmed the changes of STING-IFN-β cascade and fibrotic protein expression in each group. In contrast to sham-operated controls, the UIR model exhibited marked upregulation in the phosphorylation of STING, TBK1 and IRF3. Meanwhile, the expression of IFN-β, α-SMA and COL1 was also significantly upregulated. iMSC intervention markedly blunts overactivation of the STING‑IFN‑β cascade and reduces the abundance of fibrotic markers.

As expected, DMXAA treatment further enhanced STING pathway activation and renal fibrosis. Of note, combined treatment with DMXAA and iMSC reversed the inhibitory effect of iMSC on STING signaling, thereby restoring high levels of IFN-β and fibrotic marker expression.

Collectively, the in vivo findings illustrated that iMSCs alleviate UIR-triggered renal fibrotic lesions by suppressing the STING‑IFN‑β cascade. Moreover, the STING agonist DMXAA was capable of abolishing the renoprotective properties of iMSCs.

iMSC regulates CD8⁺ T cell infiltration, vascular endothelial function and macrophage polarization in UIR-induced renal injury via inhibiting the STING-IFN-β axis

To further confirm the immunomodulatory effects of iMSC mediated by the STING–IFN-β axis in vivo (consistent with the above in vivo pathway results), immunofluorescence staining was performed to detect CD8⁺ T cell infiltration, CD31⁺ vascular endothelial cells, and macrophage markers (CD86, F4/80, CD206) in renal tissues of each group at 14 days after surgery (Fig. 10).

Fig. 10.

Effects of iMSC on CD8⁺ T cell infiltration, vascular endothelial function and macrophage polarization in UIR-induced renal injury. A Immunofluorescence staining demonstrates CD8⁺ T cell infiltration within renal tissues (150×, scale bar: 40 μm). B Quantitative percentage of CD8⁺ T cells (%). C Immunofluorescence staining of CD31 (vascular endothelial marker) in renal tissues (100×, scale bar: 50 μm). D CD31 fluorescence intensity. E Multiplex immunofluorescence staining showing the co-expression of F4/80, CD86 and CD206 in renal tissues (100×, scale bar: 50 μm). F Percentage of CD86⁺F4/80⁺ cells among total F4/80⁺ macrophages (%). G Percentage of CD206⁺F4/80⁺ cells among total F4/80⁺ macrophages (%). All data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

As shown in Fig. 10A and B, UIR injury obviously elevated interstitial CD8⁺ T cell accumulation in renal tissues versus the Sham group, whereas iMSC intervention remarkably attenuated the infiltration of these immune cells. Further activation of STING via DMXAA notably aggravated CD8⁺ T cell recruitment and counteracted the protective action of iMSCs, suggesting that iMSCs modulate renal CD8⁺ T cell infiltration in a STING‑dependent manner.

CD31 immunofluorescence (Fig. 10C and D) showed that UIR significantly reduced CD31 fluorescence intensity (impaired vascular endothelial function), while iMSC enhanced CD31 expression to protect vascular integrity. DMXAA weakened CD31 expression and reversed the protective effect of iMSC, consistent with the STING-IFN-β axis regulation.

As illustrated by multiplex immunofluorescence co-staining (Fig. 10E) and subsequent quantitative assessment (Fig. 10F and G), UIR modeling promoted a pronounced M1-skewed macrophage polarization: CD86⁺ pro-inflammatory M1 macrophage populations were significantly upregulated, whereas CD206⁺ anti-inflammatory M2 macrophages were markedly downregulated. iMSC reversed this polarization bias, while DMXAA promoted M1 polarization and blocked the regulatory effect of iMSC, confirming iMSC modulates macrophage phenotype via the STING-IFN-β axis.

Collectively, consistent with the STING–IFN-β axis inhibition results, iMSC alleviates UIR-induced renal injury by reducing CD8⁺ T cell infiltration, protecting vascular endothelial function, and regulating macrophage M1/M2 polarization, all of which are reversed by STING activation.

Discussion

Renal fibrosis represents a common pathological endpoint of CKD, driven by complex interactions between innate immunity, adaptive immunity, and microvascular homeostasis [32–34]. Despite decades of study, no approved treatment for this process exists, highlighting the pressing requirement to discover new regulatory pathways and define targetable intervention points. The present study delineates a previously unrecognized role of iMSCs in mitigating renal fibrosis by disrupting a pathogenic cascade linking macrophage STING activation, CD8⁺ T cell recruitment, and microvascular rarefaction. Our findings suggest that iMSCs exert protective effects through dual mechanisms: a direct paracrine protection on tubular epithelial and endothelial cells (evidenced by reduced oxidative stress, improved viability, and enhanced tube formation in HK-2 and HUVEC experiments), and an indirect immunomodulatory effect by disrupting a pathogenic cascade involving macrophage STING activation, CD8⁺ T cell recruitment, and microvascular rarefaction. Together, these observations advance our understanding of iMSC-mediated renal protection and provide mechanistic insights into the immunometabolic basis of progressive kidney injury.

The kidney harbors a diverse immune microenvironment, where crosstalk between resident and infiltrating immune cells dictates the balance between injury and repair [35]. Macrophages are central regulators of innate immunity and display phenotypic plasticity, with M1 (pro‑inflammatory) and M2 (anti‑inflammatory) subsets driving distinct pathological outcomes. Numerous studies have established that M2 macrophages suppress CD8⁺ T cell-mediated cytotoxicity against renal tubular cells via the production of anti-inflammatory cytokines. Concurrently, other reports highlight interactions between M1 macrophages and CD4⁺ T cells. Macrophages serve as a primary source of T cell chemoattractants [36]. CD8⁺ T cells have a complex and still poorly understood function in renal fibrosis. Some evidence indicates a protective function against fibrosis, whereas their depletion has been shown to attenuate fibrotic progression [37]. However, in vivo experiment, CD8⁺ T cells knockout also impact other immune cells, such as dendritic cells, complicating interpretation [38]. Notably, CD8⁺ T cells have been shown to drive tubular rarefaction and worsen renal fibrosis [39]. We found that M1 macrophage infiltration strongly correlated with CD8⁺ T cell accumulation in the obstructed kidney, with M1‑derived IFN‑β serving as a key chemoattractant for CD8⁺ T cells. This aligns with recent reports identifying macrophages as a major source of T cell chemoattractants but further identifies the STING-IRF3-IFN-β axis as the critical molecular switch governing this interaction. Correspondingly, iMSC treatment not only reduced M1 (CD86⁺) macrophages but also increased M2 (CD206⁺) macrophages, shifting the M1/M2 balance toward a pro-repair phenotype.

It is noteworthy that CD8⁺ T cell-mediated microvascular injury emerges as a pivotal downstream event in our study. The negative correlation between CD8⁺ T cell infiltration and CD31⁺ vascular density supports the hypothesis that cytotoxic effector molecules and/or pro-inflammatory cytokines directly compromise endothelial cell survival and function maintenance. Subsequent vascular rarefaction exacerbates tissue hypoxia – a potent stimulus for fibroblast activation and extracellular matrix (ECM) deposition – thereby perpetuating a vicious cycle of “hypoxia-fibrosis-angiogenesis inhibition.” This finding underscores the loss of microvasculature as a core, likely irreversible, checkpoint in CKD progression. Indeed, while previous studies have linked CD8⁺ T cells to endothelial damage in autoimmune contexts, our work suggests that this axis may be a druggable target in renal fibrosis, although direct in vivo causality remains to be confirmed with genetic models.

MSCs are gaining increasing attention for their paracrine and immunomodulatory properties. Accumulating evidence suggests that MSC protection mainly arises from immunomodulation and paracrine effects, not direct differentiation [40].Giannasi et al., for example, found that ASC‑CM, in contrast to ASC‑EVs, strongly suppressed TNFα‑stimulated MMP activity in chondrocytes [41]. Consequently, MSC-CM has emerged as a promising cell-free therapeutic alternative. Our results demonstrate that iMSC-CM potently suppresses STING activation in macrophages, blocking downstream IRF3 phosphorylation and IFN-β secretion. This inhibition of the STING-IFN-β axis attenuates CD8⁺ T cell recruitment, preserves endothelial integrity, and ultimately mitigates fibrosis – effects that were reversed by the STING agonist DMXAA. This provides causal evidence identifying STING inhibition as a mechanistic keystone in iMSCs-mediated renal protection. Nevertheless, definitive causal validation will require genetic loss‑of‑function studies.

As a cytosolic DNA sensor, STING represents a crucial part of the innate immune system, has recently garnered significant attention in sterile inflammation and tissue repair [42]. Oxidative stress activates the STING‑IRF3 signaling pathway. STING, serving as a key signaling node, moves from the ER to the Golgi apparatus and there engages and activates TBK1. Activated TBK1 kinase further triggers phosphorylation modification of the transcription factor IRF3 [43]. This phosphorylation event serves as the essential molecular switch for IRF3 activation. The role of IRF3 is not isolated. It acts in synergy with other simultaneously triggered transcriptional regulators including NF-κB, to jointly boost IFN-β synthesis [44, 45]. Following pathway activation, both IFN-β mRNA and protein levels are significantly upregulated. The synthesized IFN-β protein is then secreted and acts on surrounding cells in autocrine and paracrine manners. IFN-β functions as a vital linkage between the innate and adaptive immune systems. It enhances the maturation and functional capacity of antigen-presenting cells, such as dendritic cells, and directly modulates the responses of T cells and B cells. This coordinated action ultimately establishes a long-lasting and antigen-specific adaptive immune protection [46].

Based on our correlative and pharmacological intervention data, our findings suggest a model in which STING activation contributes to CD8⁺ T cell recruitment and microvascular injury, and iMSCs may interrupt this axis by inhibiting STING. Direct in vivo causality will require further genetic studies. This pathological cycle is proposed as follows: (a) Renal injury leads to the release of DAMPs, thereby triggering STING pathway activation in macrophages; (b) STING‑IRF3 signaling drives IFN‑β secretion, recruiting and activating CD8⁺ T cells; (c) CD8⁺ T cells, via cytotoxic factors and cytokines, promote both endothelial apoptosis and vascular rarefaction; (d) diminished perfusion exacerbates hypoxia, further activating fibroblasts and promoting ECM deposition. By inhibiting STING, iMSCs disrupt this cycle at its origin, preventing downstream immune‑mediated microvascular dysfunction. This work integrates three distinct fields-STING biology, T cell immunology, and renal microvascular physiology-to provide a novel framework for understanding fibrotic progression and nominates STING as a potential therapeutic target.

How might the direct and indirect mechanisms work together in vivo? We propose that the direct paracrine protection of iMSCs on tubular and endothelial cells limits the initial oxidative damage, thereby reducing the release of DAMPs and dampening the upstream trigger of STING activation. Meanwhile, the immunomodulatory effect - inhibition of the macrophage STING‑CD8⁺ T cell axis - prevents the escalation of inflammation and subsequent microvascular injury. These two mechanisms likely act synergistically: direct cytoprotection lowers the inflammatory drive, while immune modulation interrupts the established vicious cycle, together preserving capillary networks and mitigating fibrosis.

Despite these insights, several limitations warrant consideration. First, the complexity of iMSC-CM necessitates identification of its active component(s). Fractionation studies combined with functional screening could pinpoint key effectors, potentially enabling the development of synthetic mimetics. Second, the mechanisms by which CD8⁺ T cells induce endothelial damage require clarification. In vitro co-culture models combined with knockout approaches would dissect the relative contributions of perforin/granzyme versus death receptor pathways. Third, while we focused on the STING-IRF3-IFN-β axis, STING also activates NF-κB, which drives TNF‑α and IL‑6 production, which are implicated in renal inflammation. Future studies should address whether iMSC-CM concurrently modulates NF-κB signaling and whether this contributes to its antifibrotic effects. Finally, translational studies in larger animal models are needed to validate STING as a therapeutic target and optimize iMSC-CM delivery.

Conclusion

This study shows that iMSCs attenuate ischemia‑reperfusion‑induced renal fibrosis by modulating renal inflammation and immune homeostasis. Mechanistically, iMSCs suppress the STING‑IRF3‑IFN‑β signaling pathway, reducing IFN‑β secretion and subsequent CD8⁺ T cell recruitment, thereby preserving renal capillary density. These findings highlight iMSCs as a promising cell‑based therapy for renal fibrosis and encourage further research toward clinical translation. A graphical summary is presented in Fig. 11. Figure 11 is a graphical summary of this study.

Fig. 11.

iMSCs treatment attenuates renal fibrosis and capillary rarefaction by targeting the STING pathway in macrophages.In the ischemic kidney (CKD model), pathogenic macrophages exhibit activated STING signaling, leading to the production of IFN-β. This promotes the infiltration and activation of cytotoxic CD8⁺ T cells, which contribute to capillary rarefaction (loss of CD31⁺ vessels) and fibrosis. Treatment with the STING agonist DMXAA exacerbates this pathway. Conversely, iMSCs-derived paracrine factors suppress STING activation in macrophages. This inhibition reduces IFN-β-mediated recruitment and activation of CD8⁺ T cells, thereby preserving the renal capillary network and mitigating fibrosis. BioRender was used to create this figure. (https://BioRender.com+)

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AKI

Acute kidney injury

CC

Cellular components

cDNA

Copy deoxyribonucleic acid

CKD

Chronic kidney disease

CM

Conditioned medium

CO₂

Carbon dioxide

COLⅠ

Type I collagen

DMXAA

5,6-Dimethylxanthenone-4-acetic acid

ECM

Extracellular matrix

FN

Fibronectin

GO

Gene ontology

H₂O₂

Hydrogen peroxide

HRP

Horseradish peroxidase

IFN

Interferon

IL

Interleukin

iMSCs

Induced pluripotent stem cells-derived mesenchymal stem cells

iPSCs

Induced pluripotent stem cells

IRF3

Interferon regulatory factor 3

ISCT

International society for cellular therapy

KEGG

Kyoto encyclopedia of genes and genomes

LPS

Lipopolysaccharide

MF

Molecular functions

MSCs

Mesenchymal stem cells

PBS

Phosphate buffer saline

PFA

Paraformaldehyde

PTC

Peritubular capillary

ROS

Reactive oxygen species

TBK1

TANK-Binding kinase 1

TNF-α

Tumor necrosis factor-alpha

UIR

Unilateral ischemia-reperfusion

WB

Western blot

Author contributions

FJX, QGA and QLC designed the research; MKL and YBZ performed the experiment and collected the data; XWH, YL and ZWL performed the statistical analysis; MKL and YBZ wrote the manuscript. XWH, YRJ and YMW assisted with the literature searches and revised the manuscript. YXL, LD and XCC made manuscript revisions and provided reagent support. All authors read and approved the final manuscript.

Funding

This study was supported by National Key Research and Development Program of China and Youth Independent Innovation Science Fund of PLA General Hospital (Nos. 2023YFC3605500, 2023YFC3605501; 22QNFC087).

Data availability

All datasets will be publicly released upon article publication. All other experimental materials are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Animal Ethics declaration: (1) Title of the approved project: Comparison of three animal models of transformation from AKI to CKD. (2) Ethics committee (institutional): Ethics Committee of the Animal Care and Use Committee of Beijing Institute of Radiation Medicine. (3) Ethics approval code: IACUC-DWZX-2024-P647. (4) Animal experiments were approved on August 12, 2024, and performed in compliance with the ARRIVE guidelines. Cell Ethics declaration: (1)We purchased iMSCs from Nuwacell Ltd. (Hefei, China). The company has provided formal documentation confirming that: The original human somatic cells (from which the iMSCs were reprogrammed) were collected under protocols reviewed and approved by an accredited Ethics Committee. All donors gave written informed consent for research use of their biological samples. Ethics approval code: ASSCR-YXLL-2019-03. (2)The HK-2 cell line (RRID: CVCL_0302) used in this study is a commercially available, immortalized cell line. This cell line was originally established and characterized by Ryan et al. The cells we used were purchased from [Procell Life Science & Technology Co.,Ltd.], which guarantees that its products are sourced in compliance with ethical standards. In accordance with our institution’s policy, no additional ethical approval is required for the use of such established, de-identified commercial cell lines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.