MSCs-derived HGF alleviates senescence after AKI by modulating mitoSTAT3-controlled copper flux and respiration

Abstract

Background

Human umbilical cord-derived mesenchymal stem cells (hUC-MSCs) can reverse senescence after acute kidney injury (AKI) via maintaining mitochondrial homeostasis. Copper accumulation and STAT3 nuclear translocation promote senescence, but their mitochondrial localization in response to MSCs remains unclear.

Methods

C57 mice with renal unilateral ischemia reperfusion injury (uIRI) were renal capsular transplanted with hUC-MSCs for two weeks to assessed treatment efficacy. Then, RNA sequencing, protein co-immunoprecipitation, molecular docking, and molecular dynamic simulation were used to found the relationship between senescence, mitochondrial translocation of STAT3 (mitoSTAT3), and copper homeostasis. Furthermore, inhibition of cMet/HGFR, mitoSTAT3, or COX17 were used to validated their contact.

Results

HUC-MSCs improved renal function, reduced senescence markers (SA-β-gal, p53, p21, p16), and increased STAT3pSer727 and COX17 levels. RNA sequencing revealed that senescence regulation is associated with copper homeostasis and respiratory chain complex IV. Blocking MSCs-derived HGF via lentivirus decreased STAT3pSer727, COX17, and mt-Co1 (a key subunit of complex IV). Co-immunoprecipitation and molecular docking confirmed tight binding between STAT3pSer727 and COX17. Inhibiting cMet produced similar effects as HGF deficiency, with increasing mitochondrial copper and decreasing mt-Co1. In hypoxic renal tubular epithelial cells (RTECs), blocking HGF or cMet diminished STAT3 mitochondrial translocation, and inhibiting mitoSTAT3 decreased COX17 and mt-Co1. Furthermore, knockdown COX17 aggravated loss of complex IV activity, copper accumulation and RTECs senescence.

Conclusions

HUC-MSCs-derived HGF promotes STAT3 mitochondrial translocation via cMet, enhancing mitochondrial respiration and copper excretion through COX17, thereby reducing renal senescence after AKI.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04653-3.

Introduction

Acute kidney injury (AKI) is characterized by a rapid decline in glomerular filtration rate [1], with 94% of critically ill patients experiencing adverse renal outcome within 3 years [2]. Renal senescence following AKI promotes the deterioration of renal function, and mitochondrial dysfunction plays a significant role in this process.

CTR1 is responsible for cytosolic copper overload and COX17 mediates mitochondrial copper overload. CTR1 induce senescence of renal tubular epithelial cells (RTECs), possibly due to increasing pyruvate kinase activity [3]. While COX17 prevents renal deterioration by maintaining the activity of respiratory chain complex IV [4]. As the rate-limiting enzyme at the end of the respiratory chain, complex IV promotes ATP synthesis [5]. As the mitochondrial dysfunction closely associated with cellular senescence, although no one has investigated the correlation of COX17 and RTECs senescence, COX17 dysregulation might be more essential to incur the deterioration of renal function.

Since mesenchymal stem cells (MSCs) can regulate mitochondrial quality thereby reduce renal senescence, MSCs may hold the potential to alleviate mitochondrial copper overload [6]. MSCs or HGF can upregulated the STAT3pS727-induced mitochondrial translocation of STAT3 [7, 8]—play a role in inhibiting testicular premature senescence [9]. Also, mitochondrial STAT3 (mitoSTAT3) can enhance respiration by inhibiting the activity of succinate dehydrogenase (SDH) [10], which is related to copper metabolism. Copper transporters or chaperones such as ATOX1, ATP7B and CTR1 can activate STAT3 or its promoter [11, 12], but whether or not COX17 activate can STAT3 or its promoter remains unknown. the role of COX17 is unknown. We then made a hypothesis that mitoSTAT3 may interact with COX17, thereby affecting the structure and function of respiratory chain complex IV, ultimately affecting the RTECs senescence and renal outcomes. Since the close interact between copper transporters and mitochondria, mt-Co1, as a component of complex IV [13] and might be regulated by COX17. This study examined the role of mesenchymal stem cells (hUC-MSCs) - derived HGF in copper flux control and subsequent effects on renal senescence by viral transfection or antibody neutralization to inhibit hUC-MSCs-derived HGF to regulate gene expression.

Materials and methods

Animal model and treatment

Male C57BL/6 mice at 8 weeks of age were purchased from SPF (Beijing) Biotechnology Co., Ltd and housed at the Animal Center of Chinese PLA General Hospital. The mice with uniform body weights were subjected to temperature control (22–23 °C), humidity control, and a 12-h light/dark cycle. The mice were ear-tagged, and then randomly grouped by a random number table and computer for different purposes, with 6 per group in a double blinded manner. Unilateral renal ischemia-reperfusion injury (uIRI) was used to simulate acute kidney injury.

Firstly, mice were grouped as sham, uIRI + col, uIRI + col-MSC according to the need to observe the therapeutic effect of hUC-MSCs. Specifically, after anesthesia with 1% pentobarbital, the body temperature of the mice was maintained with a thermostatic pad during the 35 min of clamping of the left renal pedicle, and the renal blood flow was restored after the end of clamping. With the above operation, a mouse model of uIRI was established. hUC-MSCs obtained from Vcanbio Cell & Gene Engineering Corp., Ltd. (Tianjin, China) were mixed with 3 mg/ml rat tail collagen I (Life Technologies, USA). A 1/9 volume of sterile 1 M NaOH and 10×PBS was used to adjust the pH of the mixture. Renal subcapsular injection of 80 µl 2 × 10⁶ collagen-coated hUC-MSCs (col-MSC) was performed immediately after uIRI. After injection, the body temperature of the mice was allowed to coagulate collagen to colonize the hUC-MSCs under the renal capsule. The mixture was given again 7 days later. On day 14, the animals were euthanized and freshly harvested for analysis. After sampling, cervical dislocation was performed.

Secondly, to investigate the role of hUC-MSCs-derived HGF, mice were grouped as sham, uIRI, uIRI + MSC^ShNC, and uIRI + MSC^ShHGF. Specifically, hUC-MSCs were transfected with HGF-knockdown lentivirus (MSC^shHGF) or negative control (MSC^shNC) (HanBio Technology, China) at an MOI of 60. After 8 h, the fresh culture medium was replaced, cultured to 72 h, and then treated with 8 µg/ml puromycin. The infected and expanded hUC-MSCs were transplanted into mice as described above and samples were also collected on day 14.

Finally, every two-days mtcur-1 (EN300-188280, Enamine, Ukraine) or daily SGX-523 (S1112, Selleckchem, USA) were used based on hUC-MSCs treatment respectively. Mice were treated with SGX-523 (25 mg/kg) by oral gavage or mtcur-1 (10 mg/kg) by intraperitoneal injection. PBS or Methyl Cellulose (MC) (HY-125861, Med Chem Express, USA) was used as a drug control. Accordingly, mice were grouped as uIRI + MSC + PBS and uIRI + MSC + mtcur-1 to explore mitoSTAT3 inhibition on day 14. In addition, groups uIRI + MSC + MC and uIRI + MSC + SGX-523 were used to observe the effect of HGF receptor (cMet) inhibition at 14 days.

This study has been approved by the Institutional Animal Care and Use ethical Committee of the Chinese PLA General Hospital in 2022 (No. 2022-X18-30).

Renal function assay

Using Creatinine Assay kit (C011-2-1, Nanjing Jiancheng Bio, China) and Urea Assay Kit (C013-2-1, Nanjing Jiancheng Bio), the standard was diluted gradient, and microplate reader (Tecan, German) read the mice serum absorbance to fit the standard curve and calculate the Serum creatine (Scr) and Blood urea nitrogen (BUN) of the sample.

Sirus red staining

Kidney tissues were fixed with 4% paraformaldehyde, 2–3 mm Paraffin-embedded sections were sealed with neutral resins after deparaffinization, iron hematoxylin staining, Sirius red staining and dehydration for transparency. Red staining in the renal interstitial indicates fibrosis.

SA-β-gal staining

After fixed and embedded in OCT, 4 μm Fresh cryosections were incubated overnight at 37 °C with a senescence detection kit (GMS10012.3, GENMED, USA). The accumulation of senescence enzymes appeared as an inky green stain.

Tissue immunofluorescence

Briefly, underwent fixed and OCT removal, permeabilization with trixton-100, and blocking with serum, 4 μm cryosections frozen in OCT were exposed to primary antibodies overnight at 4 °C, followed by incubation with Cy3- or FITC-conjugated secondary antibodies at 37℃ for 1 h. After washing, the slides were sealed with an anti-fluorescence quencher containing DAPI. Required primary antibodies including: anti-STAT3pSer727 (ab32143, Abcam, USA), anti-Tomm20 (ab289670, Abcam), anti-COX17 (11464-1-AP, Proteintech, USA). A Leica TCS-SL confocal microscope and Olympus Fluoview 3000 were used for observation and image acquisition, respectively. Areas with strong fluorescence, excluding blood vessels and connective tissue, were identified as positive.

Electron microscopy

To observe mitochondrial morphology, 1 m³ kidney tissue were fixed with 2.5% glutaraldehyde, postfixed with 1% osmium tetroxide for 1 h, and embedded in Epon812. Then, ultrathin sections were prepared. Then, sections stained with uranyl acetate and lead citrate were observed using an H7650 transmission electron microscope (TEM) (Hitachi, Japan).

Transcriptomic and bioinformatics analysis

Samples were collected from sham, uIRI + col, and uIRI + COL-MSC groups. RNA of renal tissue was extracted and purified by Bohao Shanghai Biotechnology Corp., LTD. for mRNA sequencing, followed by sequencing library construction. Illumina NovaSeq6000 and PE150 (Pair-end 150 bp) sequencing modes were used for sequencing. Heatmap of differential gene, pathway enrichment, GO enrichment, GSEA enrichment and gene correlation analysis were performed using bioinformatics analysis tools in Hiplot Pro.

Co-immunoprecipitation

In brief, uIRI kidneys of 1/2 tissue block size was lysed using an immunoprecipitation kit (PK10007, Proteintech) according to the manufacturer’s instructions. Anti-STAT3pSer727 (ab32143, Abcam) was used as the bait protein. Combined with anti-COX17 (11464-1-AP, Proteintech), anti-ATP7B (NB100-360, Novus, USA), anti-CTR1 (13086, Cell Signaling Technology, USA), and other antibodies were used to detect protein immunoblotting. Then, anti-COX17 (11464-1-AP, Proteintech) was used as the bait protein. The expression of anti-STAT1 (9172, Cell Signaling Technology) and anti-STAT5 (12071-1-AP, Proteintech) were detected by Western blot.

Molecular docking

In this study, the binding pattern of STAT3 in its phosphorylated and non-phosphorylated form with COX17 proteins was predicted. The structure files of these proteins from the PDB database were extracted. After removing structural regions with predicted errors, the default configuration of ZDOCK 3.0.2 was used for docking research, and AMBER18 for energy minimization under the ff14SB force field. Finally, the binding energy assessment was implemented based on Prodigy and visualized by PyMOL 2.5.3.

Molecular dynamics simulation

The STAT3-COX17 and STAT3pSer727-COX17 protein complexes obtained from docking were used as the initial structures for all-atom molecular dynamics simulations, performed using AMBER 20 software. Under the joint adding of hydrogen atoms to each system by LEaP module, a truncated octahedral TIP3P solvent box at a distance of 10 Å from the system, and Na+/Cl- ions to balance the system charge, the topology and parameter files for the simulation were finally output. After calculating the long-range electrostatic interactions using the Particle Mesh Ewald (PME) method, constraining the bond lengths of hydrogen atoms by the SHAKE method, and temperature control by the Langevin algorithm, the trajectory was saved every 10 ps for subsequent analysis. The binding stability was compared by evaluating the binding energy, binding conformation, number of stable hydrogen bonds, Root Mean Square Fluctuation (RMSF), and Root Mean Square Deviation (RMSD) of COX17 before and after STAT3 phosphorylation.

Cell culture

Mouse primary renal tubular epithelial cells (mPRTECs) were cultured in DMEM/F12 (Gibico, USA) containing 10% serum with 5% CO2 at 37℃. A concentration of 1% was specified to mimic oxygen deprivation compared to normoxic conditions containing 21%O2. mPRTECs were subjected to 24 h oxygen deprivation in a glucose-free medium, then replaced with a fresh medium followed by 24 h normoxia to establish a cell hypoxia-reoxygenation model.

Antibody neutralization

hUC-MSCs were seeded in a suitable medium (YOCON Biology, Beijing, China). Then, the collected supernatants (MSC-CM) were concentrated and added to mPRTECs. Neutralizing antibodies against HGF (AF294, RD, USA) or cMet (AF276, RD) were added at a dilution ratio 1:500 to treat mPRTECs.

Inhibition of mitoSTAT3

Mtcur-1, a mitochondria-targeted STAT3 inhibitor, was used at a concentration of 10µM to limit mitochondrial STAT3.

Knockdown of COX17

To knockdown COX17, mPRTECs were seeded to reach 50% confluence for 24 h, then transfected with COX17 shRNA (Gene Pharm Bio Inc, China) for 4 h. The fresh medium was replaced for another 48 h, and then the medium containing puromycin was used to screen the stable cell line.

ATP detection

After sufficiently lysed and centrifuged, ATP production of the mPRTECs was determined through Assay Kit (S0027, Beyotime, China). The bioluminescence values were read by fitting a standard curve and using a microplate reader.

Measurement of respiratory chain complex IV activity

After mPRTECs were lysed and centrifuged, Cytochrome C Oxidase Activity Assay Kit (E-BC-K837-M, Elabscience) was used according to the manufacturer’s instructions.

Detection of copper ion content

Protein concentrations were determined by BCA (23225, Thermo, USA) in renal tissue directly or after cytoplasm and mitochondria isolation by extraction kit (C3601, Beyotime). Subsequently, Copper (Cu2+) Colorimetric Assay Kit (E-BC-K300-M, Elabscience) calculated copper ions level per gram of protein.

Live cell staining of mitochondrial permeability transition pore (mPTP)

This assay was achieved with the Mitochondrial Permeability Transition Pore Assay kit (KTA4002, Abbkine, China). After mPRTECs treatment, CalceinAM (positive control), fluorescence quenching solution, or Ionomycin (negative control) were added. The cells were incubated at 37℃ for 30 min in the dark, then changed to the medium for another 30 min, and finally replaced with Assay Buffer. Images were acquired using Alexa Fluo 488 fluorescence excitation under confocal microscopy.

Live cell staining of mitochondrial membrane potential

This assay was achieved with the Mitochondrial membrane potential assay kit (C2006, Beyotime). mPRTECs were stained with 1 µg/ml JC-1 dye at 37℃ for 35 min and then microscopically analyzed. Red fluorescence represents high potential, and the green is low. Their relative ratios were used for the overall assessment.

Live cell staining of intracellular and mitochondrial ROS

A cellular ROS assay kit (S0033, Beyotime) was used to assess intracellular ROS levels, whereas a mitochondrial superoxide indicator (M36007, Thermo) was used to determine mitochondrial ROS levels. mPRTECs were stained with 50 µM DCF-DA dye for 50 min or stained with 5 µM MitoSOX Red for 45 min and then microscopically analyzed.

Live cell staining of cuprous ion probe

mPRTECs were seeded in confocal plates and cultured in advance until treatment was completed. Copper sensor-1 (HY141511, Med Chem Express) was dissolved in DMSO, and cells were incubated at 5 µM working solution concentration for 30 min at 37℃. Using Alexa Fluo 546 fluorescence excitation, mPRTECs were microscopically analyzed after washing with PBS.

Immunofluorescence cell staining

The mPRTECs fixed with 4% paraformaldehyde were permeabilized with trixton-100, incubated with primary and secondary antibodies, stained with DAPI and photographed under a confocal microscope. Primary antibodies used in the experiments included: anti-STAT3pSer727 (ab32143, Abcam), anti-COX17 (11464-1-AP, Proteintech), anti-Tomm20 (ab289670, Abcam), anti-STAT3 (10253-2-AP, Proteintech), anti-P16 (ab54210, Abcam). After removing the blue channel marked by DAPI, the JACoP plug-in of image J software obtained Manders overlap coefficients for double staining. The Manders coefficient instead of Pearson’s coefficient is used here because when studying the co-localization of signal molecules in cells after a particular stimulus, the Manders overlap coefficient can more directly reflect the degree of spatial overlap between them in a particular situation, especially the localization of subcellular compartments.

Mitochondrial fractionation/subcellular fractionation

Mitochondrial protein was extracted by a kit (C3601, Beyotime). After cell homogenization, cell debris was removed with a centrifugal speed of 600 g and cytoplasmic and mitochondrial proteins were separated with a centrifugal acceleration of 11,000 g. The procedure was performed on ice throughout. Actin was used as a cytosolic internal loading control, whereas COX IV was used as a mitochondrial internal loading control.

Western blot analysis

The samples were lysed with RIPA (P0013B, Beyotime), and the concentrations were determined with BCA assay (23225, Thermo). Each sample was loaded to SDS-PAGE gels equally and subjected to vertical electrophoresis. Primary antibodies required for the assay included: anti-STAT3pSer727 (ab32143, Abcam), anti-mt-Co1 (AG2674, Beyotime), anti-STAT3 (10253-2-AP, Proteintech), anti-COX17 (11464-1-AP, Proteintech), anti-β-actin (66009-1-Ig, Proteintech), anti-α-Tubulin (80762-1-RR, Proteintech), anti-Collagen-I (ab270993, Abcam), anti-α-SMA (ab7817, Abcam), anti-Vimentin (ab92547, Abcam), anti-GAPDH (60004-1-Ig, Proteintech), anti-COXIV (AG8011, Beyotime), anti-P53 (2524, Cell Signaling Technology), anti-P16 (ab51243, Abcam), anti-P21 (ab109520, Abcam). Full length uncropped original western blots were presented in S1 File.

Statistical analysis

Data from 4 to 6 animal experiments or 3–4 cell experiments are presented as mean ± SD, and a 2-tailed t-test or ANOVA with multiple comparisons was used to make statistics in GraphPad Prism 8.0. The threshold for a significant difference was a P value of less than 0.05.

Results

hUC-MSCs reversed uIRI-induced renal senescence via STAT3pSer727, copper flux and mitochondrial respiration

To evaluate the efficacy of hUC-MSCs (Fig. 1A), renal function and fibrosis were assessed. Specifically, the uIRI-induced increasing Scr, BUN, and Sirus Red at day 14 were reduced following hUC-MSCs treatment (Fig. 1B-D). For senescence level, SA-β-gal activity was elevated post-uIRI but downregulated with hUC-MSC therapy (Fig. 1E). Surprisingly, STAT3pSer727 and COX17, which were slightly upregulated after uIRI, were markedly enhanced by hUC-MSCs (Fig. 1F-H). This observation prompted us to consider the association between STAT3pSer727 and COX17.

Fig. 1.

uIRI 14 days-induced renal senescence were inhibited by hUC-MSCs, while STAT3pSer727 and COX17 was upregulated. A Schematic illustrating the collection of animal specimens on day 14. Unilateral IRI was performed in C57BL/6 mice and given hUC-MSCs immediately in a renal capsule. MSCs were given again on day 7 (n = 6). B Serum creatinine levels after uIRI 14 days. C Blood urea nitrogen levels after uIRI 14 days. D Images of Sirus red staining. Scale bar, 20 μm. E Images of SA-β gal staining. Scale bar, 50 μm. F Western blotting of STAT3, STAT3pSer727 and COX17. G Quantification of protein level. The quantification of phosphor-STAT3 was normalized to total-STAT3. H Quantification of protein level of COX17. Data are presented as mean ± SD. **P < 0.01; *P < 0.05

RNA sequencing revealed that renal uIRI affected differential genes in senescence regulation, copper homeostasis (including copper transport, copper binding, and copper chaperone activity), and respiratory chain complex IV assembly (S1 Fig A-B). hUC-MSCs treatment led to partial enrichment of these genes (Fig. 2A), with strong correlations noted (S1 Fig C, Fig. 2B). Top enriched pathways related to senescence included tryptophan, fatty acid, β-alanine, and pyruvate metabolism (S1 Fig D, Fig. 2C), supporting the anti-senescence effect of hUC-MSCs. Differential GO enrichment highlighted cell aging, transporter activity, response to copper ion, respiratory electron transport chain, and regulation of protein localization (S1 Fig E, Fig. 2D). Notably, hUC-MSC treatment elevated mt-Co1, which was reduced by uIRI, amidst genes linked to respiratory chain electron transport (S1 Fig C, Fig. 2E). Despite uIRI down-regulating three (mt-Co1, mt-Co2, mt-Co3) of the four complex IV subunits (S1 Fig C), only mt-Co1 was upregulated by hUC-MSCs, indicating its importance in modulating respiratory chain complex IV activity.

Fig. 2.

The reduction in renal senescence by hUC-MSCs was correlated with copper homeostasis and mitochondrial respiration. RNA-seq of renal tissue after uIRI 14 days. uIRI + col (M group) and uIRI + col-MSCs (T group) were compared (n = 3). A heatmap of M group and T group. B Correlation heat map of differential gene correlation in senescence regulation and copper homeostasis. C Top 30 signaling pathway enrichment of M group and T group. D GO enrichment of M group and T group. E Enrichment of GSEA including mt-Co1. The lower part is the heat map of upregulated genes enriched around mt-Co1

HGF derived from hUC-MSCs modulated STAT3pSer727 and COX17 after renal uIRI

HGF mediated anti-senescence effect of hUC-MSCs after renal uIRI in our previous [6] and similar at this study (Fig. 3A). HGF knockdown was validated (Fig. 3B, C), and weakened the therapeutic benefits of hUC-MSCs including reversing renal function (Fig. 3D, E), fibrosis (Fig. 3F), and SA-β-gal activity (Fig. 3G). The present study revealed that the significant upregulation of STAT3pSer727 and COX17 induced by hUC-MSCs was impaired by inhibition of HGF signaling (Fig. 3H–N). Furthermore, immunofluorescence exhibited changes in Tomm20, a mitochondrial damage marker, showing increased levels post-uIRI but reduced after hUC-MSCs transplantation, whereas STAT3pSer727 increased. However, HGF knockdown in hUC-MSCs reversed these effects (Fig. 3O). Additionally, hUC-MSCs transplantation lowered the elevated renal copper levels after uIRI, and HGF inhibition did not significantly affect this outcome (Fig. 3P).

Fig. 3.

HGF mediates renal senescence and mitochondrial copper flux limitation of hUC-MSCs. A Schematic illustration of the experiment (n = 6). MSCs infected with shHGF lentivirus or shNC empty vector, were mixed with collagen and injected into uIRI mice and again 7 days later. B Western blotting of HGF expression in hUC-MSCs (passage 6). C ELISA detection of HGF released in hUC-MSCs (passage 6). D Serum creatinine levels after uIRI 14 days. E Blood urea nitrogen levels after uIRI 14 days. F Images of Sirus red staining. Scale bar, 20 μm. G Images of SA-β gal staining. Scale bar, 50 μm. H Western blotting of fibrosis and signaling pathway indicators. I–N Quantification of protein level. O Immunofluorescence staining images of STAT3pSer727 (Red) and Tom20 (Green). Scale bar, 20 μm. P Copper level of renal tissue. Data are presented as mean ± SD. **P < 0.01; *P < 0.05. ns, not significant

COX17 was identified as a tightly binding partner of STAT3pSer727

The synchronous changes of STAT3pSer727 and COX17 suggest their interaction. To validate it both at chemical and physical level, co-immunoprecipitation (Fig. 4A), molecular docking (Fig. 4B; Tables 1 and 2), and molecular dynamics simulations (Fig. 4C-F) were performed. Initially, immunoprecipitation confirmed the binding between STAT3pSer727 and COX17 in renal uIRI (Fig. 4A). Notably, STAT3pSer727 exhibiting higher binding energy (ΔG_bind showing a more considerable absolute value) (Tables 1 and 2), smaller Root Mean Square Error (Fig. 4C), and lower Root Mean Square Fluctuation (Fig. 4D) to COX17 than its non-phosphorylation. Additionally, the number of stable binding hydrogen bonds increased from 3 to 5 (Fig. 4E). During the dynamic simulation, the red phosphorylated ser727 group was consistently positioned within the hydrophobic pocket, indicating minor steric hindrance (Fig. 4F). Overall, our findings illustrate that COX17 binds more stably following STAT3 phosphorylation.

Fig. 4.

Phosphorylation of STAT3 at Ser727 promotes its binding to COX17. A Immuno-coprecipitation of STAT3pSer727 and COX17. IP: Immunoprecipitation. IB: Immunoblot. B Molecular docking patterns of STAT3 with COX17 before and after phosphorylation. Ser727 is centrally located in the binding pocket of the two molecules with little steric hindrance. C RMSD of STAT3 with COX17 before and after phosphorylation in molecular dynamics simulations. The black curve representing the combined deviation of the two fluctuates from within 10 to within 0.7 angstroms. D RMSF of STAT3 with COX17 before and after phosphorylation in molecular dynamics simulations. The curve fluctuation range was changed from 6 to 1 angstrom, and the main fluctuation range was reduced from 2 to 0.2 angstroms. E Number of hydrogen bonds of STAT3 with COX17 before and after phosphorylation in molecular dynamics simulations. The number of stable binding hydrogen bonds changed from 3 to 5. F Binding images of STAT3 in its non-phosphorylated and phosphorylated forms with COX17 extracted at different time points in molecular dynamics simulations. The red phosphorylated group representing ser727 was consistently located in the center of the binding pocket during kinetic simulations

Table 1.

Predicted binding energies and energy components (kcal/mol) for COX17/STAT3

System name	COX17/STAT3
ΔEvdw	− 71.48 ± 3.34
ΔEelec	− 168.50 ± 11.41
ΔGGB	207.95 ± 12.77
ΔGSA	− 9.99 ± 0.57
ΔGbind	− 42.03 ± 2.95

Table 2.

Predicted binding energies and energy components (kcal/mol) for COX17/STAT3pSer727

System name	COX17/STAT3pSer727
ΔEvdw	− 68.32 ± 3.44
ΔEelec	− 387.52 ± 24.07
ΔGGB	371.62 ± 20.53
ΔGSA	− 9.83 ± 0.58
ΔGbind	− 94.05 ± 11.16

ΔE_vdW: van der Waals energy. ΔE_elec: electrostatic energy. ΔG_GB: electrostatic contribution to solvation. ΔG_SA: non-polar contribution to solvation. ΔG_bind: binding free energy

Moreover, STAT3 can enter the mitochondrial matrix, allowing interaction with the copper transporter SLC25A3, another mitochondrial copper transporter found in the inner membrane [14]. Although molecular docking revealed strong binding between these proteins (S2 Fig), using STAT3 as a decoy in renal uIRI did not capture distinct copper transporters like SLC25A3, ATP7A/B, SCO1/2, COX11 and CTR1. Similarly, COX17 did not capture STAT1 and STAT5. The possibility of STAT3pSer727 binding to other intracellular copper transporters and COX17 binding to other STAT families were excluded, suggesting that the interaction between COX17 and STAT3pSer727 may be specific in copper homeostasis.

HGF/cMet-mediated hUC-MSCs modulate mitochondrial translocation of STAT3 (mitoSTAT3), copper flux, and respiration

To clarify the regulatory relationship between STAT3pSer727 and COX17, targeted inhibition of STAT3pSer727-induced mitoSTAT3 was performed using mtcur-1. In addition, SGX-523 was used to specifically inhibit cMet, a well-known receptor for HGF, thus providing more substantial support for the role of hUC-MSCs-derived HGF (Fig. 5A). Initially, inhibiting mitoSTAT3 or cMet elevated Scr, BUN (Fig. 5B-C), and fibronectin level (Fig. 5H-I), indicating worsened renal injury, along with decreased ATP content (Fig. 5D) and disrupted mitochondrial morphology (Fig. 5E). Subsequently, the anti-senescence effect of cMet/mitoSTAT3 in hUC-MSCs treatment was confirmed by the observation that mtcur-1 or SGX-523 heightened SA-β-gal activity (Fig. 5F) and the senescence protein level, including P53, P21, and P16 (Fig. 5H, J-L). Furthermore, the down-regulation of STAT3pSer727 stated the efficacy of mtcur-1, which mirrored the effects seen with the administration of SGX-523. Accordingly, both inhibitors decreased COX17 (Fig. 5G, N) and mt-Co1 (Fig. 5G, O) levels, suggesting that mitoSTAT3 and cMet sustain these proteins. Given the previously established role of COX17 in regulating mt-Co1 [4], and the observed enrichment of mt-Co1 following hUC-MSCs transplantation in RNA sequencing above, it is suggested that hUC-MSCs-derived HGF may influence mitochondrial respiration via cMet/mitoSTAT3, meriting further investigation. The altered COX17 expression points to its role in copper transport, particularly in mitochondria, as subcellular detection showed hUC-MSCs regulate mitochondrial copper through this pathway, not in the cytosol (Fig. 5P).

Fig. 5.

Inhibition of mitoSTAT3 attenuates the renal therapeutic effect of hUC-MSCs. A Schematic illustration of the experiment (n = 6). MSCs were divided into four groups based on two doses after uIRI, comparing once-daily administration of PBS or mtcur-1. In addition, MC or SGX-523 was administered every other day for comparison. B Serum creatinine levels after uIRI 14 days. C Blood urea nitrogen levels after uIRI 14 days. D ATP content of renal tissue. E Electron microscopy images. (×15 000 magnification). Scale bar, 2 μm. F Images of SA-β gal staining. Scale bar, 50 μm. G Immunofluorescence staining images of COX17 (Red). Scale bar, 20 μm H Western blotting of senescence and signaling pathway indicators. I–O Quantification of protein level. P Copper level of renal cytosol and mitochondria. Data are presented as mean ± SD. **P < 0.01; *P < 0.05. ns, not significant

Visualizing the subcellular localization of STAT3 in renal tissue is challenging, which is why cellular hypoxia/reoxygenation was employed to visualize its translocation (Figs. 6A and 7A). The opening of the mPTP, indicated by dim fluorescence, confirmed successful hypoxia-reoxygenation, while MSC-CM treatment reduced mPTP opening, with impaired fluorescence recovery upon HGF antibody neutralization (Fig. 6B). Notably, STAT3 levels, especially STAT3pSer727 in the mitochondria, decreased after oxygen deprivation but were rescued by MSC-CM, with restoration reversed by the HGF antibody (Fig. 6C–E)., Immunofluorescence subsequently revealed increased interaction between STAT3 and COX17 after MSC-CM treatment, but blocking HGF reduced this interaction, as evidenced by a decrease in the yellow overlap area and overlap coefficient (Fig. 6F, G). Additionally, oxygen deprivation-decreased ATP content and respiratory chain complex IV activity were upregulated by MSC-CM treatment, but attenuated by neutralizing HGF (Fig. 6H, I). These findings suggest that hUC-MSC-derived HGF regulates mitochondrial respiration in hypoxic RTECs.

Fig. 6.

hUC-MSCs derived HGF promotes mitochondrial translocation of STAT3 in oxygen-reoxygenated RTECs. A Schematic illustrating the cell treatment (n = 3–4). Renal tubular epithelial cells (mPRTECs) were isolated from the kidneys of wild-type C57 mice and cultured under normoxic or hypoxic conditions (OD). Based on hypoxia, Conditioned medium (CM) or HGF-neutralizing antibody (HGFAb) were added, respectively. B Expression of mPTP level (Green). C Western blotting images. D, E Quantification of protein level. F Immunofluorescence staining images of STAT3pSer727 (Green) and COX17 (Red). Scale bar, 20 μm. G Overlap coefficient of STAT3pSer727 and COX17. H Relative cellular ATP level. I Relative mitochondrial respiratory chain complex IV activity. Data are presented as mean ± SD. **P < 0.01; *P < 0.05. ns, not significant

Fig. 7.

cMet mediates hUC-MSCs-derived HGF to inhibit copper flux and maintain mitochondrial respiration. A Schematic illustrating the cell treatment (n = 3–4). mPRTECs were cultured under normoxic or hypoxic conditions (OD). Based on hypoxia, Conditioned medium (CM) or cMet-neutralizing antibody (cMetAb) were added, respectively. B Western blotting images. C–F Quantification of protein level. G) Immunofluorescence staining images of STAT3pSer727 (Red) and Tomm20 (Green). Scale bar, 20 μm. H Overlap coefficient of STAT3pSer727 and Tomm20. I Relative mitochondrial respiratory chain complex IV activity. J Copper sensor probe (Red). Data are presented as mean ± SD. **P < 0.01; *P < 0.05

Similar to HGF neutralization, cMet neutralization hindered the upregulation of STAT3pSer727, and increased COX17 and mt-Co1 following MSC-CM treatment (Fig. 7B–F). Given the localization of COX17 in the mitochondrial outer membrane, the outer membrane marker Tomm20 was used to demonstrate the displacement of STAT3pSer727. Immunofluorescence confirmed the mitochondrial translocation of STAT3pSer727 upon MSC-CM treatment, evidenced by a brighter yellow overlap area and a higher overlap coefficient. However, mitochondrial translocation of STAT3pSer727 was impaired after cMet neutralization (Fig. 7G, H), strongly indicating that both STAT3pSer727 and COX17 are regulated by hUC-MSCs-derived HGF. The subsequent role of COX17 was further confirmed. Additionally, respiratory chain complex IV activity was compromised after cMet neutralization (Fig. 7I), and oxygen deprivation-induced cellular toxic copper levels decreased with MSC-CM treatment but reversed with cMet neutralization (Fig. 7J). In summary, HGF/cMet from hUC-MSCs mediates mitochondrial translocation of STAT3, mitochondrial copper flux, and respiration.

mitoSTAT3-mediated hUC-MSCs adjust COX17 level and respiratory chain complex IV activity in hypoxia-reoxygenated RTECs

To investigate the role of mitoSTAT3 in effect of hUC-MSCs-derived HGF on RTECs, mtcur-1 was used for targeted inhibition in hypoxic RTECs treated with MSC-CM (Fig. 8A). Initially, Inhibition of mitoSTAT3 resulted in mPTP opening in hypoxic RTECs treated with MSC-CM, as evidenced by Calcein AM as a positive control and lonomycin as the negative (Fig. 8B). Additionally, mitochondrial membrane potential (MMP) decreased, indicated by increased green fluorescence and decreased red (Fig. 8C), along with elevated intracellular or mitochondrial ROS levels (Fig. 8D-E). Immunofluorescence showed that STAT3, which was normally in the cytoplasm during MSC-CM treatment, moved away from mitochondria upon mitoSTAT3 inhibition, with Tomm20 levels rising (Fig. 8F). These findings, along with the increased senescent protein level (P53, P21, and P16) and SASP factors (TGFβ and PDGF) following mitoSTAT3 inhibition (Fig. 8G-L), suggest an anti-senescence effect of mitoSTAT3. Furthermore, mt-Co1 protein levels and respiratory chain complex IV activity also decreased upon mitoSTAT3 inhibition (Fig. 8G, K, M). To dissect the link between COX17 and mitochondrial translocation of STAT3, subcellular protein isolation proved that mitochondrially localized COX17 translocated to the cytosol following mtcur-1 treatment (Fig. 8N), correlating with increased toxic cuprous levels (Fig. 8O) and impaired copper excretion.

Fig. 8.

Inhibiting mitoSTAT3 induced copper flux and RTECs senescence, also disrupted mitochondrial respiration in hUC-MSCs treatment. A Schematic illustrating the cell treatment (n = 3–4). In addition to the hypoxia (OD) and medium treatments (CM), mtcur-1 (mitochondrial STAT3-targeting inhibitor) was added or not. B Fluorescent probe of mitochondrial membrane potential. Monomers are shown in green and multimers in red. C Fluorescent probe for cellular ROS (Green). D Expression of mPTP level (Green). E Fluorescent probe for mitochondrial ROS (Red). F Immunofluorescence staining images of STAT3 (Red) and Tomm20 (Green). G Western blotting images. H–K Quantification of protein level. L Quantification of SASP mRNA level. M Relative respiratory chain complex IV activity. N Immunoblot images of isolated cytoplasmic and mitochondrial proteins. O Copper sensor probe (Red). Data are presented as mean ± SD. **P < 0.01; *P < 0.05

Similar to mitoSTAT3 inhibition, COX17 knockdown (Fig. 9A) resulted in increased mPTP opening (Fig. 9B), reduced mitochondrial membrane potential (Fig. 9C), elevated intracellular or mitochondrial ROS (Fig. 9D, E), higher P16 fluorescence value (Fig. 9F), and increased levels of senescent protein (P53, P21, and P16) (Fig. 9G–J) and SASP factors (IL-1β and PDGF) (Fig. 9L). In addition to suppressing senescence levels described above, COX17 appears to play an essential role in respiration, with down-regulation of mt-Co1 and complex IV activity observed after its knockdown (Fig. 9G, K, M). Toxic cuprous levels increased, indicating impaired copper excretion (Fig. 9M). The involvement of COX17 in regulating mt-Co1 was further confirmed with MSC-CM treatment. In summary, COX17 controls copper flux and mt-Co1-dependent respiration in hypoxic RTEC, which is corrected by hUC-MSCs through mitoSTAT3.

Fig. 9.

Knockdown of COX17 prevents hUC-MSCs from effectively regulating mitochondrial respiration and RTECs senescence. A Schematic illustrating the cell treatment (n = 3–4). In addition to the hypoxia (OD) and conditioned medium (CM) treatments, COX17^sh (COX17shRNA) was added or not. B Fluorescent probe of mitochondrial membrane potential. C Fluorescent probe for cellular ROS (Green). D Expression of mPTP level (Green). E Fluorescent probe for mitochondrial ROS (Red). F Immunofluorescence staining images of COX17 (Red) and P16 (Green). G Western blotting images. H–K Quantification of protein level. L Quantification of SASP mRNA level. M Relative respiratory chain complex IV activity. N Copper sensor probe (Red). Data are presented as mean ± SD. **P < 0.01; *P < 0.05

Discussion

hUC-MSCs secrete HGF to stimulate STAT3 translocation to the mitochondria, thereby interacting with COX17 on the mitochondrial outer membrane, ultimately activating the excretion of mitochondrial copper ions and controlling the activity of respiratory chain complex IV. Ultimately alleviating renal senescence after AKI.

Our previous work has demonstrated that HGF is an anti-aging component in MSCs, and the present study further revealed that HGF acts through inducing STAT3-mediated mitochondrial translocation and COX17 activation. And ultimately maintain mitochondrial homeostasis through copper excretion and respiration induction (Fig. 10). Targeting mitochondrial copper flux to improve renal outcome was early found in hUC-MSCs treatment. Furthermore, copper flux triggered by molecular mitochondrial translocation was organically integrated with cellular senescence.

Fig. 10.

A model depicting the mechanism of copper flux control from hUC-MSCs

Based on our previous research [6], the role of HGF derived from hUC-MSCs in inhibiting renal senescence after AKI was further demonstrated. The upregulation of STAT3pSer727 and COX17 was found simultaneously during this exploration, prompting us to consider their possible association. RNA sequencing reveals a correlation between senescence, copper homeostasis, and mitochondrial respiration in MSCs treatment, providing evidence for in-depth exploration. Then, STAT3pSer727, COX17, and a key subunit of Complex IV (mt-Co1) were also found to be modulated by HGF derived from hUC-MSCs. These findings suggest that hUC-MSCs-derived HGF could regulate copper transport and subsequent mitochondrial respiration. In support of this, results that inhibiting the known receptor for HGF (cMet) worsens renal senescence, aligns with reducing STAT3pSer727, COX17, and mt-Co1. Moreover, although the effect of HGF from MSCs on the total copper in kidney tissue had no statistical significance, the cytoplasmic and mitochondrial components were separated to confirm that inhibiting mitoSTAT3 or HGFR/cMet did not affect the cytoplasmic copper but the mitochondrial copper. cMet appears to protect against AKI by regulating copper levels predominantly in mitochondria rather than in the cytoplasm, corroborating that cMet focuses on regulating COX17, which is responsible for mitochondrial copper transport.

The established interaction between cMet and STAT3 [15] suggests that the regulation of STAT3 by HGF warrants further exploration. Our study shows that mitoSTAT3 induced by HGF derived from hUC-MSCs, lessens the adverse effects of its nuclear translocation. The observed discrepancy, where STAT3pS727 was slightly upregulated after uIRI but decreased in hypoxic RTECs might result from different cell types in renal tissue, as well as the unique microenvironment and stress response of RTECs to injury.

STAT3 has excellent potential to regulate copper homeostasis. Metabolomic studies highlight it in GSH formation, γ-glutamine cycle [16], and methionine homeostasis [17], both tied to copper balance. Following zinc exposure, STAT3 moves to lysosomes and mitochondria, helping mitigate lipotoxicity [18]. Given the link between copper homeostasis disorder and cuproptosis caused by protein-lipid acylation disorder [19], mitoSTAT3 may affect copper homeostasis. In addition, high copper levels sensitize cells to mPTP [20], contributing to a feedback loop that enhances cellular senescence due to increased ROS and decreased NAD+. Our research accordingly indicates that inhibiting mitoSTAT3 raises toxic cuprous levels, further opened mPTP, then induced cellular senescence. Additionally, prior research has indicated that STAT3 knockout disrupts mitochondrial respiratory chain complex I and indirectly affects the II-IV [21]. Our findings confirms that the copper-dependent activity of complex IV diminishes with mitoSTAT3 inhibition. Given its various functions such as preventing testicular senescence, reversing G0/G1 cell cycle arrest, or mitigating NLRP3-mediated oxidative stress and subsequent autophagy, mitoSTAT3 play a role in anti-senescence [9, 22]. But its inhibition in senescent CD4 + T cells reduces Th17-mediated inflammatory responses [23].In our study, mitoSTAT3 seemed to prevent excessive mPTP opening and maintain complex IV activity in our study, thereby reducing cellular senescence induced by toxic copper accumulation.

Both chemical and physical assessments indicate that STAT3pSer727, which induces mitoTSAT3, binds more stably to COX17 than its non-phosphorylated form does. Similar to the effects of STAT3, SLC25A3 regulates mPTP opening [24], mitochondrial respiration, and NLRP3-mediated oxidative stress [14], but is less prominent in copper transport because of its primary function as a phosphate transporter. Despite strong interaction between SLC25A3 and STAT3, the predominant function of COX17 in MSCs-regulated copper flux remains indisputable. Positioned on the outer mitochondrial membrane, COX17 interacts directly with the cytoplasm, thereby sensing copper flux both within mitochondria and across cellular compartments to maintain homeostasis. Interestingly, our RNA sequencing analysis showed no significant upregulation of COX17 at the mRNA level, possibly due to its reliance on STAT3 subcellular activation rather than transcriptional regulation. Lead exposure increases mitochondrial translocation of COX17, resulting in enhanced mitochondrial copper uptake and associated damage in Alzheimer’s disease [25]. In contrast, the present study revealed that inhibition of mitoSTAT3 prompted COX17 to move away from mitochondria, where it helps remove excess copper in hUC-MSCs treatment.

Although copper-mediated stress-induced premature senescence has recently been shown in preadipocytes, in-depth studies are limited to plants and less to mammals [26]. Copper transporters such as ATP7A play a role in senescence-related autophagy, and contribute to degenerative diseases [27]. In addition, copper chelation alleviates cellular senescence by affecting pyruvate dehydrogenase activity [28]. Our RNA sequencing analysis also enriched pyruvate metabolism, supporting the notion that hUC-MSCs modulate senescence through copper homeostasis. Moreover, p53/p21 signaling activation represents the DNA damage response during premature senescence, whereas p16 indicates the onset of chronic senescence. Knockdown of mitoSTAT3 or COX17 increased the expression of p16, p53 and p21 in MSCs. These results indicate that the impairment of mitochondrial copper excretion may aggravate DNA damage. UUO-induced conformational changes in COX17 also affected the steric hindrance of mt-Co1 [4]. Therefore, the excretion of mitochondrial copper by COX17 can maintain the activity of complex IV. Our findings suggest that COX17 knockdown leads to cellular senescence, decreasing mt-Co1, which inactivating the respiratory chain complex IV, leading to mitochondrial dysfunction and subsequent senescence in RTECs. Additionally, the maintenance of complex IV activity also requires the incorporation of an appropriate amount of copper. Thus, it is also essential to keep the mitochondrial copper homeostasis. After excluding the reactivity of other copper chaperones partners to STAT3pS727 through co-immunoprecipitation, the role of COX17-mediated mitochondrial copper transport was specifically emphasized.

STAT3 translocation to the endoplasmic reticulum regulates mitochondrial Ca2 + release [29], while its lysosomal localization aids autophagy [30]. Additionally, STAT3 is critical in modulating activity of ATP7B [11], which is the copper transporter in Golgi apparatus. Future research needs to explore the copper flux control of STAT3 within these organelles. Despite limited differential copper homeostasis proteins identified, it is important to note that cellular copper homeostasis is regulated primarily by mitochondria rather than cytoplasm. Consequently, omics sequencing of subcellular components could provide deeper insights into this regulation.

Conclusion

Our study revealed that hUC-MSCs-derived HGF promote the mitochondrial translocation of STAT3 and subsequent COX17-dependent mitochondrial copper excretion or respiration to reduce senescence following AKI. These findings may provide theoretical basis for the development of reno-protective drugs targeting mitochondrial copper and the improvement of organelle-oriented precision medicine.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

hUC-MSCs

Human umbilical cord-derived Mesenchymal Stem Cells

Col-MSCs

Collagen matrix-encapsulated MSCs

uIRI

Unilateral ischemia/reperfusion injury

HGF

Hepatocyte growth factor

mitoSTAT3

Mitochondrial translocation of STAT3

STAT3pSer727

Phosphorylation of STAT3 at Ser727

Author contributions

Kaiting Zhuang conceived studies, designed experiments, interpreted results, and wrote the article. Wenjuan Wang and Cheng Xu revised the article. Siyang Wang, Yanjun Liang, and Xumin Zheng constructed the mouse model and did the immunohistochemistry. Yuhao Chen and Yingjie Zhang prepared pathological specimen and captured images. Xiangmei Chen provided an experimental environment. Zhe Feng provide technical asistance. Guangyan Cai provided experimental funds.

Funding

This work was supported by the Natural Science Foundation of China (82170686), National Natural Science Foundation of China (No. 82200762), China Postdoctoral Science Foundation (No. 2024M754288), and Beijing Natural Science Foundation (7254312).

Data availability

The RNA-Seq data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE299500 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE299500).

Declarations

Ethics approval and consent to participate

hUC-MSCs used in this study were donated by Vcanbio Cell & Gene Engineering Corp. Ltd, which had confirmed that there was initial ethical approval for collection of human cells. The work described has been carried out in accordance with Declaration of Helsinki for experiments involving humans, and that informed consent was obtained for experimentation with human subjects. The animal experiments were conducted in line with the ARRIVE guidelines 2.0. All animal experiments were performed under ethical supervision and approved by the Institutional Animal Care and Use Committee of the Chinese PLA General Hospital in 2022 (Title research: Study on stem cell therapy for acute kidney injury. No. 2022-X18-39. Date of approval: April 18, 2022). This ethical approval also permitted performing experiments using the aforementioned hUC-MSCs.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.