Human amniotic epithelial cells improve kidney repair in ischemia-reperfusion mouse model of acute kidney injury

Background: Acute kidney injury (AKI) is a common clinical disease with complex pathophysiology and very limited therapeutic choices. This prompts the need for novel therapy targeting multiple aspects of this disease. Human amnion epithelial cells (hAECs) are ideal alternative stem cell source for regenerative medicine. Increasing evidence suggests that hAEC-derived exosomes (hAECs-EXO) may act as novel cell– cell communicators. Accordingly, we assessed the therapeutic potential of hAECs in ischemia reperfusion mouse model of AKI and explored the underlying mechanisms. Methods: The hAECs were primary cultured and hAECs-EXO were isolated and characterized. An ischemic renal injury mouse model was established to mimic different severity of the kidney injury. Mouse blood creatinine level was used to assess renal function and kidney specimens were processed to detect cell proliferation, apoptosis and angiogenesis. Immune cells inltration was analyzed by ow cytometry. hAECs-EXO was used to treat hypoxia-reoxygenation (H/R) injured HK2 cells and mouse bone marrow-derived macrophages to evaluate their protective effect in vitro. Furthermore, hAEC exosomes were subjected to liquid chromatography-tandem mass spectrometry for proteomic proling. Results: We found that systematically administered hAECs could improve mortality and renal function in IRI mice; decrease the number of apoptotic cells; promote peritubular capillary regeneration and modulate kidney local immune response. However, hAECs showed very low kidney tissue integration. Exosomes isolated from hAECs recapitulated the renal protective effects of their parent cells. In vitro, hAECs-EXO protected HK-2 cells from H/R injury-induced apoptosis and promoted bone marrow-derived macrophage polarization toward M2 phenotype. Proteomic analysis on hAECs-EXO revealed proteins involved in extracellular matrix organization, growth factor signaling pathways, cytokine production and immunomodulation. These ndings demonstrated that paracrine of exosomes might be a key mechanism by hAECs mediating kidney functional recovery in AKI. Conclusions: We rst reported hAECs could improve mortality and renal repair in mice with ischemia-reperfusion injury. The anti-apoptotic, pro-angiogenetic, and immunomodulatory capabilities of hAECs at least partially, through paracrine pathways. The renoprotective effects of hAECs-EXO might be a promising clinical therapeutic tool, overcoming the weaknesses and risks associated with the use of native stem cells for patients with AKI.


Introduction
Acute kidney injury (AKI) is a common clinical disorder de ned by an abrupt decrease in glomerular ltration [1]. An episode of AKI is often associated with an increased risk of developing chronic kidney disease as well as an increased risk of both short-and long-term mortality [2,3]. Renal ischemia reperfusion (I/R) is one of the major causes of acute kidney injury (IRI-AKI), leading to its high morbidity and mortality [4]. Ischemia and hypoxia lead to tubular and endothelial necrosis and apoptosis. The tissue damage is exacerbated by the initiated in ammatory response after reperfusion and re-oxygenation of the kidney [5]. Because of the multifactorial pathophysiological mechanisms, there is no effective pharmacological therapy that prevents the evolution or reverses the injury once established [6]. With an increased incidence of AKI worldwide, there is an urgent need to develop more effective strategies for treating AKI.
In the past few years, stem cell therapy has been widely recognized as a potential treatment strategy for AKI. Human amniotic epithelial cells (hAECs) form the epithelial layer of amniotic membrane. As a stem cell, hAEC has attracted widespread attention in regenerative medicine due to its pluripotency, low mutation frequency, low immunogenicity, low tumorigenicity, rich sources, and no ethical risks [7]. Because of these advantageous characteristics, hAECs have obtained a good outcome after transplantation in many diseases, including lung injury [8], brain injury [9], and hepatic brosis [10]. However, it is unclear whether hAECs can promote AKI recovery.
Stem cell therapy was observed to have bene cial renal protection under a variety of kidney pathologies, yet differentiation of stem cells into enough number of cells to reconstitute renal parenchymal organs was undetected [11,12]. Therefore, studies have increasingly focused on the paracrine action of stem cells. Exosomes are major extracellular vesicles with a diameter in the range of 40-150 nm produced by multivesicular bodies and are important information carriers regulating both cellular function and gene expression via transferring cargo into recipient cells [13]. Recent studies have reported that hAEC-derived exosomes reduce liver brosis [14], restrict lung injury and enhance endogenous lung repair in bleomycinchallenged aged mice [15].
In this study, we aimed to explore whether hAECs or exosomes secreted from hAECs can reduce AKI mortality and improve renal function in an IRI mouse model, and to make a preliminary discussion on the therapeutic mechanisms of hAECs on AKI.

Material And Methods
Mice Male C57BL/6j mice (9-10 weeks) were purchased from Vital River Laboratory Animal Technology

Culture of hAECs
The human amniotic epithelial cells were provided by Shanghai iCELL Biotechnology Co., Ltd (Shanghai, China). Brie y, the human amniotic epithelial cells were isolated from fresh amnion membranes collected from healthy mothers after cesarean deliveries with written and informed consent. The procedure was approved by the Institutional Ethics Committee of the International Peace Maternity and Child Health Hospital, School of Medicine, Shanghai Jiao Tong University (Approval Number: [2014]11). The amnion was peeled from the collected placentas, then the blood and mucus were washed away with PBS carefully. The collected amnion tissue was then transferred into a 150 ml ask and dissociated with 50 ml trypsin for 30 min at 37℃. The digestive process was ended by adding 50 ml fresh culture medium.
The cell suspension was transferred through a 200 mesh stainless steel screen, centrifuged at 650 g for 5min. The cell pellet was then suspended in fresh culture medium. The cells of passage 1 were characterized by checking phenotype under light microscopy and detecting stem cell markers by ow cytometry (Supplementary Fig. 1).

Exosome isolation and characterization
Exosomes were obtained from the supernatants of hAECs through ultracentrifugation according to classical methods reported previously [16]. In brief, the complete culture medium (CM) from hAECs was collected and centrifuged at 2,000 g for 30 minutes and ltered through a 0.22um cell lter (Millipore, Billerica, MA, USA) to remove debris. The cell-free medium was then centrifuged at 20,000 g (Beckman Coulter, USA) for 30min at 4°C to remove micro-vesicles. The supernatants were discarded and the pellets were washed in PBS, after which the suspension was centrifuged at 100,000 g for 70min at 4°C to obtain exosomes. The ultrastructure of the exosomes was analyzed under a transmission electron microscope (Zeiss, Oberkochen, Germany). The protein levels of TSG101, Alix and Flotllin were detected using western blot. To determine the sizes of the puri ed vesicles, a nanoparticle tracking analysis (NTA) was performed using Zetaview software (Particle Metrix, Meerbusch, Germany).

Induction of AKI
Ischemia for 30-33 minutes at 37°C was induced in both kidneys using the ank approach as previously reported [17]. Sham operations were performed with exposure of both kidneys, but without induction of ischemia. 100μl of cell-free vehicle, hAECs (1×10 6 cells) or hAECs-exosomes (3×10 8 exosomes) was injected into injured mice via tail vein immediately after surgery. The animals were sacri ced for the subsequent experiments on day 1, day 2, day 3 and day 7 post surgery, respectively. All animal procedures were approved by the Institutional Animal Care and Use Committee of Beijing and were performed in accordance with the National Research Council Guide for the Care and Use of Laboratory Animals. Efforts were made to minimize animal suffering and limit the number of animals used in the study.
Plasma creatinine was determined by the picric acid method as previously reported [18]. Kidney tissue sections were xed with 10% buffered formalin followed by para n embedding and stained with periodic acid-Schiff. The degree of tubulointerstitial damage was scored semi-quantitatively by a renal pathologist who was blinded to the experimental groups. The scores were based on a 0 to 4+ scale, according to the percentage of the cortex and medullar junction region affected by loss of brush border and tubular necrosis and/or apoptosis (0 = no lesion, 1+ = <25%, 2+ = >25% to 50%, 3+ = >50% to 75%, 4+ = >75% to <100%,).
RNA extraction, reverse transcription and real-time RT-PCR A fraction of kidney was harvested and RNA was isolated using TRIzol® reagent following the manufacturer's instructions (Ambion, Thermo Fisher Scienti c Inc., Waltham, MA, USA). RNA concentrations were determined by photometric measurements. cDNA was synthesized from 2 µg total RNA using FastKing RT Enzyme (KR118; TIANGEN Biotech, Beijing, China) for real-time RT-PCR. The primers used for PCR analyses were listed in Supplementary Table 1. Real-time PCR reagents were prepared from the SYBR Green PCR Master Mix (FP209; TIANGEN Biotech, Beijing, China), and all PCR analyses were performed on an ABI Vii7 system.

Immunohistochemistry staining
Immunohistochemistry staining of the kidney was performed on para n sections. The primary antibodies included rabbit anti-Ki67 (1:400, Cat. No. 9129, Cell Signaling Technology); rabbit anti-CD31 (1:100, Cat. No.28364, Abcam). The slides were then exposed to DAB-labeled secondary antibodies. The staining was examined using microscope (Leica, Germany). The number of positive cells was counted in 7-8 high-powered elds from the outer medulla in each kidney examined. Apoptosis in kidney tissues was detected on para n sections by the in situ terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) method following the standard protocol (Beyotime, China).

Western blot analysis
Samples were lysed on ice in lysis buffer supplemented with protease inhibitors. Aliquots of cell lysates were boiled in SDS-PAGE sample buffer, fractionated on 12% SDS-PAGE gel, and transferred to PVDF membrane. The membranes were blocked with 5% milk in TBST (Tris-buffered saline, 10 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 0.1% Tween-20) for 1 hour at room temperature and incubated with primary antibodies at 4°C overnight. The following primary antibodies and dilutions were used: rabbit anti-Alix The relative intensity of the protein bands was quanti ed by digital densitometry using ImageJ software (NIH, Bethesda, MD, USA). The level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal standard.

Flow cytometry
Mice were anesthetized, sacri ced, and perfused with ice-cold PBS via the left ventricle for 2 minutes.
Kidneys were minced and incubated with collagenase type IA (1mg/ml, Sigma-Aldrich) in DMEM/F12 for 25 minutes at 37°C with constant shaking. Digestion was stopped by the addition of ice-cold FBS. The digested kidney tissue suspension was passed through a 100-μm cell strainer (Thermo Fisher) and centrifuged at 1,000 g for 5 minutes at 4°C. The pellet was then incubated with ACK Lysing Buffer (0.15 M NH4Cl, 10 mM KHCO3, and 0.1 mM Na2 EDTA) for 5 minutes at room temperature to remove red blood cells, and centrifuged twice at 1,000 g for 5 minutes at 4°C. The pellet was then washed with PBS. Cells were incubated with F4/80 (PE/cy7, Cat. No.123113, BioLegend) and CD206 (AF647, Cat. No. 141712, BioLegend) antibodies. Data were acquired using BD FACSVerse and analyzed using FlowJo software.
Bone marrow-derived macrophages were isolated and cultured as described previously [19]. In brief, bone marrow cells were collected from the tibia and femur of 8 weeks old C57BL/6J mice. Cells were cultured in L929 conditioned medium (RPMI-1640 with 10% heat inactivated fetal bovine serum, 15% L929 sup, and 100 U/mL penicillin-streptomycin) and allowed to attach for 2 days. A fraction of cells was collected before polarization. The rest of cells were continued to culture in hAECs-EXO conditioned medium (RPMI-1640 with 10% heat inactivated fetal bovine serum,1x10 8 exosomes/ml, and 100 U/mL penicillinstreptomycin) for 7 days. On day 8, the media was removed and the cells were lysed in TRIzol® reagent and quanti ed M1 and M2 macrophage marker gene expression by reverse transcription-PCR. The primers used for PCR analyses were listed in Supplementary Table 1. Shotgun proteomics followed by label-free quanti cation Peptide preparation for proteomic analysis was performed as described previously [20]. Brie y, 300ug exosome proteins isolated from hAECs were used to perform the analysis with triplicates. Proteins were separated using 4-15% Mini-PROTEAN ® TGX TM Precast Protein Gel (BioRad, Hercules, CA, USA), stained with Coomassie Brilliant Blue R-250 staining solution (BioRad, Hercules, CA, USA). Total proteins resolved using mono-dimensional gel electrophoresis was subjected to protein-in-gel digestion. The peptides of each sample were desalted by a C 18 Cartridge, concentrated by vacuum centrifugation and reconstituted in 40 uL of 0.1% formic acid.
The MS experiments were performed on a Q-Exactive HF Hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scienti c, Hemel Hempstead, UK) coupled to an Easy nLC nano-liquid chromatography (Thermo Fisher Scienti c, Hemel Hempstead, UK). The desalted peptides were separated on a 15cm long separation column with an inner diameter of 50 um (Acckaim PepMap RSLC, nano viper, P/N164943, Thermo Fisher Scienti c, USA) using gradient of buffer A (Ultrapure water, 0.1% formic acid) and buffer B (Acetonitrile, 0.1% formic acid) at a owrate of 300 nL min-1. The chromatographic gradient was set to provide a linear increase from 3% to 80% buffer B in 110 min, for a total run time of 120 min. MS data was acquired on data dependent mode dynamically choosing the top ten most abundant precursor ions from the survey scan (350-1800 m/z) for fragmentation and MS/MS analysis. Precursors with a charged state of +1 were rejected and the dynamic exclusion duration was set as 25 s. MS raw data were processed using MaxQuant software version 1.5.5.1 according to the standard work ow with the built-in search engine Andromeda [21]. Proteins were identi ed by searching against the Uniprot human reference proteome (Uniprot_HomoSapiens_20386_20180905) database. Carbamidomethylation of cysteines was set as xed modi cation, while protein N-terminal acetylation and methionine oxidation were de ned as variable modi cations for peptide search. The false discovery rates (FDR) for peptide and protein identi cations were set to 0.01%. A maximum of two missed cleavages were allowed for tryptic digestion. The MaxLFQ label-free quantitation method [22] with retention time alignment and match-between-runs feature in MaxQuant was applied to extract the maximum possible quanti cation information. Protein abundance was calculated based on normalized spectral intensity (LFQ intensity).
The gene ontology (GO) annotation in categories such as cellular component, biological process, molecular functions of the signi cantly expressed proteins was carried out using DAVID functional enrichment analysis tool (https://david.ncifcrf.gov/).

Statistical analysis
Statistical analysis was performed using SPSS 25.0 statistical software (SPSS, Chicago, IL). Normally distributed variables are expressed as mean ± standard deviation and compared using a t-test. Nonnormally distributed non-parametric variables are expressed as median and interquartile range and are compared between groups using the Mann-Whitney U test. All P values were two-tailed, and P <0.05 was considered statistically signi cant.

Results hAECs ameliorated acute renal IRI
We established an ischemic renal injury mice model by changing the clamping time of the bilateral renal pedicle to control the severity of the injury. The seven-day mortality rate in the 33min-IRI group reached to 90% (Fig. 1a) (n = 21), and the blood creatinine level on the rst day after surgery was 1.95 ± 0.15mg / dl (Fig. 1b), which was in line with the characteristics of clinically severe AKI. There was no death in the 30min-IRI group (Fig. 1a), and the postoperative day 1 serum creatinine was 0.97 ± 0.3mg / dl, which was corresponding to clinically moderate AKI (Fig. 1b). The mice subjected to renal IRI showed pronounced renal pathological damage compared to mice in the sham group. The main features were tubular cell swelling, widespread tubular dilation and degeneration, nuclear condensation, and in ammatory cell in ltration (Fig. 1c). Therefore, we successfully established a mouse model of IRI-AKI. Due to the signi cant natural death of the 33min-IRI group, the severe AKI model was used to observe the seven-day mortality of mice, while the 30min-IRI group of moderate AKI was mainly used to assess mouse kidney damage and subsequent mechanism study. We next examined the effects of hAECs on IRI renal damage. Injection of hAECs reduced the seven-day mortality rate from 90% to 47.4% (p <0.05) in the 33min-IRI severe AKI group (Fig. 1d). Serum creatinine levels were decreased in hAECs group on day 1 and day 2 post-surgery (Fig. 1e). The proportion of necrotic and damaged tubules tended to be lower in the cortex and outer medullar regions of the hAECs treated group than in the Vehicle control ( Fig. 1 f and g). These ndings indicate that hAECs can promote renal function and enhance mice survival in IRI-AKI.

hAECs-EXO recapitulated hAECs' protective effect on renal IRI
We examined hAECs organ engraftment after transplantation and found very low kidney tissue integration of hAECs from the beginning of injection till 7 days post injection despite a positive therapeutic signal in the IRI-AKI mice (Supplementary Figure2). Except the direct repopulation to the injured tissues, stem cells also exerted their therapeutic effect by paracrine/endocrine of extracellular vesicles such as exosomes [23]. We isolated exosomes from hAEC conditioned medium and tested their effects on IRI-AKI. The morphological assessment revealed the typical cup-shaped morphology of exosomes as determined by transmission electron microscopy (Fig. 2a). Size analysis using a nanoparticle tracking system showed a peak size of 50-150 nm (Fig. 2b). Western blot results showed the presence of exosomal markers, including Flotillin, TSG101 and Alix in the hAEC exosomes (Fig. 2c).
hAECs and hAECs-EXO protected kidney from AKI by reducing apoptosis and stimulating cell proliferation On ischemic insults, tubular cells often undergo apoptosis or necrosis. When TUNEL-stained kidney tissues were observed by uorescence microscopy, the number of apoptotic cells in the IRI group was signi cantly increased compared with that in the sham group (Fig. 3a). However, as shown in Fig. 3b, hAECs or hAECs-EXO treatment signi cantly decreased the number of apoptotic cells. Moreover, in vitro, HK-2 cells were cultured in a hypoxic environment to simulate ischemia within the tissue, and proteins from the HK-2 cells treated with or without hypoxia-reoxygenation (H/R) injury were collected. Western blot analysis showed that the expression level of apoptosis-related marker cleaved caspase-3 in the H/R group was signi cantly higher than that in the control group, while hAECs-EXO treatment signi cantly suppressed cleaved casapase-3 expression (Fig. c and d).
Meanwhile, hAECs and hAECs-EXO stimulated dramatic tubular cell proliferation two days after surgery as shown by anti-Ki67 staining (Fig 3e and f). Similarly, in H/R stimulated HK-2 cells, the expression of PCNA was signi cantly higher in hAECs-EXO treated group than H/R group (Fig. 3 g and h). Taken together, these observations demonstrated that hAECs exosomes enhanced cell proliferation and attenuated apoptosis in renal epithelial cells under H/R injury.

hAECs and hAECs-EXO promoted angiogenesis
The tubular injuries in AKI are accompanied by renal vascular impairment. We next investigated the change of peritubular capillaries (PTCs) in IRI kidneys by immunohistochemistry and Western blot analysis using PTC endothelial cell marker, CD31. CD31+ capillaries were found to be intact in the sham kidney. Ischemia induced signi cant PTCs loss in the kidneys during the 7-day course of reperfusion (p < 0.05). However, the number of microvascular capillaries was highly increased by hAECs transplantation or hAECs-EXO administration (p<0.05) post ischemic injury (Fig. 4 a-d). We also determined the mRNA levels of the angiogenesis-related genes (Egf, Fgf, Hgf, Igf-1, Pdgf, and Vegf) in the IRI kidneys. The expression of these regeneration markers was very weak in sham kidneys and IRI induced a slight increase in the gene expression. hAECs and hAECs-EXO treatment signi cantly upregulated the expression of these genes, with only no obvious difference in Egf expression between vehicle control group and hAECs or hAECs-EXO treatment groups (Fig. 4e).

hAECs and hAECs-EXO increased M2 macrophage in ltration with a concomitant modulation of in ammatory cytokines
Stem cell therapy has been reported to prevent renal injuries via immune regulation [24,25]. Macrophages are important immune cells with high heterogeneity and plasticity, and the imbalance of M1/M2 macrophages phenotypes contributes to the in ammation or tissue repair in acute kidney injury. Ischemia reperfusion promoted immune cell in ltration into the kidney as shown by the increased percentage of macrophage compared to the sham group (Fig.5a). A continuous increase of macrophage population was observed during the 7-day course of reperfusion in hAECs and hAECs-EXO treated kidney (Fig. 5b), in consistent with the high levels of macrophage attractive chemokines in local kidney (Fig.5c).
The phenotype of local macrophages was determined by ow cytometry (Fig. 6a). I/R animals exhibited prominent in ltration of CD206+/F4/80+ M2 type macrophages in the kidneys on day 2 and afterwards after hAECs transplantation or hAECs-EXO administration (Fig. 6b). Concomitantly, postischemic kidneys of the hAEC or hAECs-EXO-treated groups expressed higher levels of anti-in ammatory cytokines such as IL4 and IL13, and lower levels of proin ammatory cytokines such as IFNγ and TNFα, than those of the vehicle-treated group (Fig. 6c). In vitro, using hAECs-EXO as supplement to culture bone marrow-derived macrophages for 7 days, the gene expression of M2 markers including Cd206, Cd163, Il4rα, and Arg1 was signi cantly increased, with a downregulation of gene expression of M1 markers such as Cd86, Ifnγ, Tnfα and iNos (Fig. 6d). Thus, we postulated that hAECs could alleviate IRI probably through promoting M2 macrophage polarization and inhibiting the systemic in ammation.

Proteomic pro ling of hAECs-EXO
Above data revealed that hAEC-derived exosomes displayed renal protective functions similar to parent cells, indicating the paracrine pathway played an important role in the process of hAEC-mediated kidney tissue functional recovery. A comprehensive analysis of the proteome on hAECs-EXO allowed an overall identi cation of 171 proteins (Supplementary table 1). The functional annotation of these proteins was performed using DAVID functional enrichment analysis tool. Predictably, a large proportion of exosome proteins was annotated as either of extracellular or membrane origin, with a sizable contribution to the cytosol and ER lumen (Fig. 7a). The molecular functions of these exosome proteins were mainly involved in protein binding, integrin binding, heparin binding, cadherin binding as well as extracellular matrix structural constituent (Fig. 7b). The biological process category showed "extracellular matrix organization", "cell adhesion", "regulation of complement activation", "leukocyte migration", "innate immune response", as well as "angiogenesis" and "positive regulation of cell proliferation" as the most enriched terms (Fig. 7c).

Discussion
Acute kidney injury is the rapid onset of decreased kidney function that ultimately increases mortality and morbidity. Despite the ever-increasing prevalence of acute kidney disease, there is a lack of potential therapeutic agents, leaving us with the option of costlier replacement therapies. In recent years, stem cell therapy has drawn much attention for the treatment of AKI [26]. As a main player, MSCs derived from different sources like adipose, bone-marrow and human umbilical cord have shown renoprotective effects in several models of toxin or IRI-induced AKI [27][28][29]. Meanwhile, numerous controversies have yet arisen regarding safety, cost, availability, and ethical considerations of MSCs [30]. In seeking of alternative cell sources, human amniotic epithelial cells are one of the most promising. In the current study, we observed a decrease of mortality rate after hAEC treatment in our severe IRI-AKI mice model.
Ischemia-reperfusion injuries leading to AKI have been associated with accumulation of reactive oxygen species (ROS) and apoptosis, loss of peritubular capillaries and migration of pro-in ammatory immune cells to the damage tissue [4,31]. In our study, we demonstrated that hAEC therapy could effectively reduce kidney injury in animal model of IRI induced AKI. Intravenous injection of hAECs attenuated the loss of peritubular capillaries and tubular cell apoptosis and necrosis, and promoted cell proliferation in the IRI injured kidney. hAECs reprogramed macrophages to shift from a pro-in ammatory M1 to antiin ammatory M2 state. This shift was associated with the increased levels of IL4, IL13 and decreased levels of TNFα and IFNγ, which in turn helped to reduce the in ammatory response. Taken together, these anti-in ammatory, anti-apoptotic, and pro-angiogenetic effects of hAECs were associated with both the rapid recovery of kidney function and the enhanced survival in the mice with IRI kidney injury.
It was originally believed that stem cells provided a therapeutic basis for regeneration by engrafting at the site of injury via transdifferentiation. However, further investigation revealed that engraftment of stem cells to the site of injury was very rare and therefore was not likely to be the sole cause of regeneration after AKI [32]. In our case, we found very few kidney tissue integration of hAECs after transplantation. The majority of hAECs were concentrated in the lung and heart. Growing evidence in regenerative medicine supports the hypothesis that stem cells exert their therapeutic effect by a paracrine/endocrine manner rather than a direct repopulation of the injured tissues [23]. Exosomes secreted from stem cells could act as major transporters in cell-cell communication to deliver bioactive molecules from original cells to the recipient cells [33]. It has been reported that hAEC-derived exosomes can restore ovarian function in chemotherapy-induced premature ovarian failure by transferring microRNAs against apoptosis [34]. In our study, we found that injection of exosomes derived from hAECs recapitulated hAECs' protective effects on renal ischemia reperfusion injury, limiting apoptosis, enhancing proliferation, regulating angiogenesis and immune response, and thus promoting kidney repair.
We characterized the molecular compositions of hAECs-EXO by proteomic approach to nd out the speci c mechanisms of hAEC-based therapy in AKI. Cellular processes including "extracellular matrix organization", "cell adhesion", "leukocyte migration" as well as "angiogenesis" and "positive regulation of cell proliferation" were enriched in hAECs-EXO. Proteins involved in the IGF signaling, HIF signaling, Integrin signaling, Wnt signaling, and TGFβ signaling were detected in hAECs-EXO. ECM proteins are also abundant in hAECs-EXO. It has been shown that ECM processing and cell-ECM interactions are major determinants in the regulation of signaling pathways that drive kidney repair [35,36]. In addition to the classical ECM structural proteins, we found that several Matricellular Proteins (MCPs) such as TNXB, THBS1, THBS3 were detected in hAEC-EXO with high abundance. MCPs are ECM bound nonstructural proteins that interact with integrins, growth factor receptors, and growth factors to modulate their function and activity [37,38]. In our proteomic analysis, Tenascin X (TNXB) as a typical matricellular protein stands out. An in vitro study has shown that TNX physically interacts with VEGF-B and enhances the ability of VEGF-B to stimulate endothelial cell proliferation [39], suggesting a role for TNXB in the regulation of angiogenesis. It was found that in normal kidney tenascin expression was limited to the medullary interstitium [40]. Chen et.al. reported that Tenascin-C (TNC), a member of the tenascin family, was speci cally induced at sites of injury and recruited Wnt ligands, thereby creating a favorable microenvironment for tubular repair and regeneration after ischemia reperfusion-induced AKI [41]. Thus, it would be interesting to explore the in vivo role of TNXB on kidney regeneration in IRI models, leading to an elucidation of the molecular mechanism of hAECs therapy for AKI.

Conclusion
In summary, the above ndings advance our knowledge of the therapeutic potential of hAECs in ischemia reperfusion-induced AKI mouse model, pointing to hAECs as an attractive candidate for promoting renal tissue repair, possibly through the limitation of apoptosis, promotion of angiogenesis and immunomodulation. Our data also demonstrate the renoprotective effects of hAEC-derived exosomes, which might be a promising clinical therapeutic tool, overcoming the weaknesses and risks associated with the use of native stem cells for patients with AKI.

Declarations
Ethics approval and consent to participate Human amniotic epithelial stem cells were isolated from fresh amnion membranes collected from healthy mothers after cesarean deliveries with written and informed consent. The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
Li Yang has a patent pending for the use of hAECs in repair of the injured kidney after AKI. Shanghai iCELL Biotechnology company declared no potential con icts of interest with respect to the research, authorship, and/or publication of this article.