Comparisons of miRNA profiles of exosomes derived from human iPSCs, ADSCs, and BMSCs and effects on chondrocyte function

Abstract

Aims

This study aimed to identify and compare the microRNA (miRNA) profiles of exosomes derived from human induced pluripotent stem cells (iPSCs), bone marrow mesenchymal stem cells (BMSCs), and adipose tissue-derived stem cells (ADSCs) (hiPSC-Exos, hBMSC-Exos, and hADSC-Exos), and their functional effects on human articular chondrocytes (hACs).

Methods

hiPSC-Exos, hBMSC-Exos, and hADSC-Exos were collected from the appropriate cells cultured in 10% bovine exosome-depleted fetal bovine serum (de-Exo-FBS) for 48 hours. Next-generation sequencing (NGS) and bioinformatics were used to analyze the small RNA profiles of these exosomes. The biological functions of hACs were examined after a 12-day treatment with exosomes.

Results

hBMSC-Exos and hADSC-Exos had similar miRNA profiles but were largely different from hiPS-Exos. There were 17 highly expressed miRNAs in hiPSC-Exos, 13 miRNAs in hADSC-Exos, and 11 miRNAs in hBMSC-Exos. Among them, seven miRNAs overlapped between the hBMSC-Exos and hADSC-Exos, and only three of them (hsa-miR-16-5p, hsa-miR-25-3p, and hsa-miR-93-5p) overlapped among all three exosomes. The putative target genes of the three overlapping exosomal miRNAs, and high-scoring target genes, including MAN2A1, ZNFX1, PHF19, GPR137C, ENPP5, B3GALT2, FNIP1, PKD2, and FBXW7, were identified. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses revealed that these genes are involved in cell growth, bone ossification, and cartilage development/differentiation, possibly via the mitogen-activated protein kinase (MAPK) signalling pathway. Accordingly, we confirmed the biological effect on cartilage differentiation and found that hiPSC-Exos, hBMSC-Exos, and hADSC-Exos maintained hAC viability, prevented senescence, promoted the formation of a normal cartilage matrix (glycosaminoglycan and type II collagen), and downregulated fibrocartilage matrix (type I collagen) in normal hACs. Comparatively, hBMSC-Exos had the greatest effect on hAC function.

Conclusion

Bioinformatics revealed differences and possible mechanisms of action of exosomes derived from pluripotent hiPSCs, multipotent hADSCs, and multipotent hBMSCs, and these exosomes effectively suppressed cell senescence and promoted normal functional extracellular matrix formation in hACs. Further investigations of the different functions of exosomes from pluripotent-hiPSCs other than those from multipotent-hMSCs are needed.

Cite this article: Bone Joint Res 2025;14(8):696–712.

Article focus

• This study identified and compared the microRNA (miRNA) profiles of exosomes derived from human induced pluripotent stem cells (iPSCs), bone marrow mesenchymal stem cells (BMSCs), and adipose tissue-derived stem cells (ADSCs) (hiPSC-Exos, hBMSC-Exos, and hADSC-Exos), and their functional effects on human articular chondrocytes (hACs).

Key messages

• hiPSC-Exos, hBMSC-Exos, and hADSC-Exos maintained hAC viability, prevented senescence, promoted the formation of a normal cartilage matrix (glycosaminoglycan and type II collagen), and downregulated fibrocartilage matrix (type I collagen) in normal hACs.

• The most highly expressed miRNAs found in the hiPSC-Exos, hADSC-Exos, and hBMSC-Exos were miR-16-5p, miR-25-3p, and miR-93-5p.

• hBMSC-Exos exhibited the greatest efficacy, while hiPSC-Exos showed the lowest efficacy in maintaining normal chondrocyte function.

Strengths and limitations

• This study applies an integrated miRNA profile from hiPSC-Exos, hBMSC-Exos, and hADSC-Exos. These exosomes from stem cells with different hierarchical potencies may exert different biological effects.

• This study only focuses on examining chondrocyte functions by these exosomes, but other putative effects such as bone, neuron development, and vesicle biogenesis also need to be explored in the future.

Introduction

Stem cells have the potential to differentiate into multiple cell lineages, possess a strong capacity for self-renewal,¹ and have been clinically used to repair and regenerate tissue defects. To date, stem cells used for clinical treatment are mainly derived from embryonic, fetal, and adult tissues and from differentiated somatic cells reprogrammed into induced pluripotent stem cells (iPSCs).² However, these sources for cell therapy have limitations, such as ethical and legal concerns around using embryonic and fetal tissues, the limited ability of adult stem cells to differentiate into specific cell types, and possible effects of mutations on the use of iPSCs. In comparison, adult stem cells have fewer adverse effects and thus serve as the most advantageous stem cell type for cell therapy and tissue engineering.²

Stem cell potency refers to the varying abilities of stem cells to differentiate into specialized cell types, and stem cells are usually categorized as multipotent, pluripotent, or totipotent.³ Generally, these cells have the potential to differentiate into more cell types that can support various therapeutic applications. The generation of iPSCs that can differentiate into all cell types in the body is an ideal strategy to widen the application range of cell therapies.⁴ However, iPSCs are generated via the ectopic expression of transcription factors, so the risk of mutation and potential tumorigenicity still limit clinical use.

In contrast, adult stem cells are usually multipotent and are responsible for generating new tissue in response to injury, disease, or regular maintenance in the body.⁵ Among adult stem cells, bone marrow mesenchymal stem cells (BMSCs) and adipose tissue-derived stem cells (ADSCs) can be stimulated to differentiate into many mesodermal cell types, e.g. articular chondrocytes, osteoblasts, cardiomyocytes, endothelial cells, and fibroblasts.² Among them, articular cartilage is a highly specialized tissue without vessels, nerves, or lymph nodes, which limits its repair potential. Therefore, the regeneration of cartilage defects remains a major unmet medical need.⁶ Recently, the potential of BMSCs and ADSCs to treat cartilage lesions and osteoarthritis (OA) has been widely demonstrated in experimental animals^7,8 and in clinical trials.^9-11 In addition, ADSCs are relatively easily harvested in greater quantities with less damage to the donor site, and have a greater proliferative capacity than BMSCs.^12,13 Moreover, ADSCs have been reported to undergo senescence later than BMSCs.¹⁴ Nevertheless, previous reports have shown that less than 1% of transplanted mesenchymal stem cells (MSCs) reach the target tissue, and some of them differentiate into unwanted fibro-like tissue or undergo hypertrophic cell changes at the cartilage repair site.^15,16 These previous reports, along with the unique benefits and limitations of these stem cells, indicate that the type of stem cell for regenerating articular cartilage still needs to be investigated.

Exosomes are extracellular vesicles (EVs) that are 40 to 160 nm (average ~100 nm) in diameter with an endosomal origin, and carry a variety of cargo including messenger RNAs (mRNAs), microRNAs (miRNAs), long noncoding RNAs, DNA, and proteins.¹⁷ These vesicles are released from most cells and can be isolated from most body fluids.^18-20 Recently, exosomes derived from MSCs have received increased attention due to studies showing the therapeutic effects of these materials in treating OA.^21-23 Moreover, exosomes did not induce any severe immune reactions observed in mice by repeated administration of mouse or human cell-derived exosomes for extended periods.^24,25 Replacing living cells with acellular exosomes may address safety concerns. Moreover, acellular MSC exosomes may serve as a biopharmacological product with clinical efficacy superior to that of allogenic MSC treatment for orthopaedics.^26,27 However, the underlying mechanisms of action of exosomes remain unclear. Exosomes play roles in intercellular communication, organ homeostasis, and disease.²⁵ miRNAs are endogenio²⁸hether exosomes derived from stem cells with varying hierarchical potencies have different biological effects on recipient cells through the miRNAs they carry.

Accordingly, in this study, we aimed to identify and compare highly expressed exosomal miRNAs from three kinds of stem cells (hiPSCs, hADSCs, and hBMSCs) using small RNA sequencing technology, and to further predict the target genes and possible pathways associated with cell differentiation and senescence in the musculoskeletal system. Moreover, the effects of these exosomes on the expression of functional and degenerative genes in human articular chondrocytes (hACs) were compared to confirm the findings from the miRNA analysis.

Methods

hiPSCs, hADSCs, hBMSCs, and hACs

The human iPSC line (IBMS-iPSC-01-02_feeder-free), derived fom the peripheral blood mononuclear cells (PBMCs) of a 33-year-old female, was generated and authenticated by the Taiwan Human Disease iPSCs Service Consortium. The characterization and validation (pluripotency markers, genome stability, cell identity, contaminant testing, and differentiation potential) of the hiPSCs were performed (Supplementary Figure c), and the cells were cultured in a feeder-free culture system with Essential 8 (E8) medium (Thermo Fisher Scientific, USA) on Matrigel-coated plates (final 200-fold dilution, #354230; Corning, USA). The medium was exchanged every day. The cultured cells were incubated in a humidified atmosphere of 5% CO₂ at 37°C. The hiPSCs were subcultured when the cells were 70% to 80% confluent by using Accutase (Thermo Fisher Scientific) at a 1:8 to 1:12 split ratio every three to five days. A ROCK inhibitor (10 µM, Y-27632) (TargetMol Chemicals Inc., USA) was used to promote cell adhesion. The hADSCs and hBMSCs were isolated from subcutaneous adipose and bone marrow tissue and consisted of four young patients (mean age 36.4 years (SD 1.2)) as described previously.²⁹ These stromal cells were characterized according to the International Society for Cell & Gene Therapy (ISCT) guidelines (Supplementary Figure c). Human adipose tissue was obtained from patients with approval from the Ethics Committee of Kaohsiung Medical University Hospital (KMUHIRB-E(II)-20220292). After informed consent was obtained, subcutaneous adipose tissue from the gluteal area was collected from the patients during plastic surgery. After 3 g of subcutaneous adipose tissue was isolated from humans, the adipose tissue was minced with scissors. The minced adipose tissue was then digested with 1 mg/ml type IA collagenase (125 U/mg) at 37°C under 5% CO₂ for 24 hours. The digested tissue was centrifuged at 1,000 rpm for five minutes, and the pellet was washed twice with phosphate-buffered saline (PBS). The pellet was then resuspended in K-NAC medium composed of Keratinocyte-SFM Basal Medium (Gibco-BRL, Thermo Fisher Scientific, USA) supplemented with 25 mg of bovine pituitary extract (BPE; Gibco-BRL), 2.5 µg of human recombinant epidermal growth factor (rEGF; Gibco-BRL), 2 mM N-acetyl-L-cysteine, 0.2 mM L-ascorbic acid 2-phosphate sesquimagnesium salt, and 5% fetal bovine serum (FBS),³⁰ and then plated in a 100 mm culture dish. K-NAC medium was then used to isolate and expand hADSCs. hBMSCs were isolated, prepared, and cultured as described previously.³¹ hBMSC aspirations were obtained with the approval of the Institutional Review Board of Taipei Veterans General Hospital (2016-11-006C). The cells were seeded at 4 × 10^3 cells/cm² and grown in complete culture medium (CCM) composed of α-minimal essential medium (α-MEM) (Gibco-BRL) supplemented with 16.6% FBS, 100 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine, with the medium changed twice per week. hACs, purchased from Clonetics, and Normal Human Chondrocytes-Knee (NHAC-kn; Lonza, USA) were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with: 100 mg/ml ascorbic acid; 1% nonessential amino acids; 1% penicillin‒streptomycin; 0.1% insulin, transferrin, and selenium (ITS; Corning); and 10% FBS.

Exosome induction and isolation

The following cell culture conditions were used to induce hADSC-derived exosomes (hADSC-Exos) and hBMSC-derived exosomes (hBMSC-Exos): hADSCs and hBMSCs were suspended in DMEM supplemented with 1% penicillin‒streptomycin and 10% bovine exosome-depleted FBS (de-Exo-FBS; Gibco), and then a total of 1.1 × 10^6 cells were seeded in 100 mm culture dishes for 48 hours. The hiPSC-derived exosomes (hiPSC-Exos) were generated as follows: the hiPSCs were cultured in an E8 medium, and once the iPSCs reached 50% confluence, the E8 medium was collected. This medium, now referred to as conditioned medium (CM), can be used for exosome isolation and purification. For CM collection, we used hADSCs at passages 6-8 (P6 to P8) and hBMSCs at passage 3 (P3). Thereafter, the CM from which the exosomes had been secreted was collected. Exosomes were isolated from the CM through ultracentrifugation as described previously with little modification.³² In brief, the CM was centrifuged at 20,000× g and 4°C for ten minutes to remove large cell fragments. The supernatant was then centrifuged at 50,000× g for one hour to remove residual organelles and nuclei. This step was followed by centrifugation at 180,000× g and 4°C for 1.5 hours to obtain a pellet containing the exosome, and the supernatant was discarded. The exosomes were then resuspended in PBS to eliminate contaminating proteins, followed by centrifugation at 180,000× g and 4°C for 1.5 hours. The pellet containing the exosomes was resuspended in 1 ml of PBS. Then, the supernatant was filtered through a 0.22 μm filter sterilizer (Steritop; Millipore, USA) to sterilize the purified exosomes. The isolated exosomes stored at 4°C were used within two weeks. The exosomes were characterized according to the guidelines of MISEV2014 and MISEV2018, including analysis of exosome size and number by nanoparticle tracking analysis (NTA), morphology by transmission electron microscopy (TEM), and protein composition by western blot as below.³³

Nanoparticle tracking analysis of exosome size and number

To determine the exosome size and number, we performed NTA on a ZetaView BASIC NTA Nanoparticle Tracking Video Microscope PMX120 (Particle Metrix, Germany) equipped with a 532 nm green laser. The isolated hiPSC-Exo, hADSC-Exo, and hBMSC-Exo samples (1 µl) were diluted to 1,000 µl with freshly filtered PBS (pore size: 0.22 µm) and loaded into the detection chamber using a syringe. Our manually adjusted camera setup was maintained for all samples with a slider shutter at 650 and a slider gain at 50. Next, five 30-second videos were recorded. The detection threshold was set at 6, and the blur and maximum jump distances were automatically set. The temperature was maintained at 25°C. The data were processed using ZetaView (version 8.05.14 SP7; Particle Metrix).

Transmission electron microscopy

To investigate the exosomal morphology of the hiPSCs, hADSCs, and hBMSCs, we used TEM. In brief, a 10 µl drop of the exosome suspension (10^11 particles/ml) was added to a copper grid and air-dried. Then, the samples were fixed with 3% glutaraldehyde for two hours, washed with PBS, and negatively stained with 2% uranyl acetate for 30 seconds. The exosome samples were then observed under a transmission electron microscope (Hitachi, Japan), and images were acquired using a digital camera (Olympus, Japan).

Western blot analysis

The cells and exosomes were lysed in radioimmunoprecipitation assay buffer (Invitrogen, Thermo Fisher Scientific) supplemented with protease inhibitor cocktails (P8340; Sigma-Aldrich, USA). Next, 20 μg of protein were separated using NuPAGE Novex 4% to 12% Bis-Tris gels (Invitrogen) and transferred to polyvinylidene difluoride nylon membranes (Sigma-Aldrich). The membranes were blocked with 5% non-fat dry milk in Tris-buffered saline for one hour and then incubated with primary antibodies against CD9 (ab223052, dilution 1:1,000), CD63 (ab134045, dilution 1:1,000), CD81 (66866-1-lg, dilution 1:1,000), ALIX (12422-1-AP, dilution 1:1,000), TSG101 (28283-1-AP, dilution 1:1,000), and β-actin (20536-1-AP; 1:1,000; Proteintech, USA) at 4°C overnight. Specific bands were detected using a horseradish peroxidase-conjugated secondary antibody (1:5,000) for one hour and then visualized with an enhanced chemiluminescence western blot system (Pierce, USA).

Treatment with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos

For hiPSC-Exo, hADSC-Exo, and hBMSC-Exo treatment, the solution was diluted with basal medium (DMEM supplemented with 1% penicillin‒streptomycin) immediately before treatment. The hACs were seeded at 10^5 cells/six-well plate, and the culture medium was removed after 24 hours and treated with or without hiPSC-Exos, hADSC-Exos, or hBMSC-Exos (10^9 particles/ml) for 12 days. The hACs were divided into groups and incubated in three media: 1) Ctrl group (basal medium alone); 2) hiPSC-Exo group (basal medium + 10^9 particles/ml hiPSC-Exos); 3) hADSC-Exo group (basal medium + 10^9 particles/ml hADSC-Exos); and 4) hBMSC-Exo group (basal medium + 10^9 particles/ml hBMSC-Exos). The medium was changed every two days, and the supernatant and hACs were collected for further experimental analyses of senescence-associated β-galactosidase (SA-β-gal), enzyme-linked immunosorbent assay (ELISA), and 1,9-dimethylmethylene blue (DMMB) at Day 12.

Senescence-associated β-galactosidase

SA-β-gal staining was used to verify that treatment with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos reduced hAC senescence. The hACs (passage 8) were grown in six-well plates, washed, fixed, and stained with a SA-β-gal cellular senescence assay kit (Cell Biolabs Inc., USA) and a negative control figure shown in Supplementary Figure b, according to the manufacturer’s protocol and as previously described. The sections were examined under a microscope (Olympus).

Enzyme-linked immunosorbent assay

Exosome-treated hACs and culture supernatants were collected to determine the concentrations of various proteins in the extracellular matrix (ECM) (type I and type II collagen) by using an ELISA. For each assay, 100 μl of the supernatant was added to each well of a 96-well ELISA plate from express ELISA kits for human type I collagen and type II collagen (RK0114 and RK01152 (ABclonal, USA), respectively), and sequentially treated with detection antibodies and chromogenic substrate in accordance with the manufacturer’s instructions. The optical density (OD) was then read at 420 nm. These protein concentrations were determined on the basis of a standard curve produced using the serially diluted reference samples provided in the ELISA kits and presented in picograms per millilitre.

Detection of sulphated GAG synthesis by the DMMB assay

The hACs were collected and digested with 300 μg/ml papain at 60°C for 18 hours. The DNA content and sulphated glycosaminoglycan (GAG) accumulation of hACs were quantified spectrofluorometrically by using 33,258 Hoechst and a Blyscan Glycosaminoglycan Assay (Biocolor, UK), respectively. The standard curve for the DMMB assay was generated using aqueous chondroitin sulphate C (Sigma-Aldrich), with a concentration range of 0 to 25 µg/µl.

MTS assays after hiPSC-Exo, hADSC-Exo, and hBMSC-Exo treatment

To determine cell viability and proliferation after exosome treatment, we used MTS assays. hACs were seeded at a low cell density (1 × 10^4 cells/well) in a six-well plate and treated with or without exosomes (10^9 particles/ml) for one, three, or five days; the medium was changed once every two days. On the day of cell harvest, the culture medium was removed, followed by the addition of 2 ml of fresh medium containing the MTS reagent (10% v/v, CellTiter 96 AQueous One Solution Cell Proliferation Assay; Promega, USA), and the mixture was allowed to react for 3.5 hours. Finally, 200 µl of the supernatant was added to a 96-well plate, and the OD was read at 450 nm.

miRNA extraction and library construction for NGS

Total RNA was extracted from hiPSC-Exo, hADSC-Exo, and hBMSC-Exo (10^11 particles) with TRIzol Reagent (Invitrogen, USA), according to the manufacturer’s instructions. The purified RNA was quantified at OD260 nm by using an ND-1000 spectrophotometer (Nanodrop Technology, USA) and qualified by using a Bioanalyzer 2100 with an RNA 6000 Nano kit (Agilent Technology, USA). Sample libraries were prepared using the QIAseq miRNA Library Kit (Qiagen, USA) according to the manufacturer’s protocols. Specially designed 3’ and 5’ adapters were ligated to mature miRNAs using the QIAseq miRNA next-generation sequencing (NGS) 3’ and 5’ ligase. The ligation reaction was incubated for one hour at 28°C, followed by 20 minutes at 65°C, and then held at 4°C. The ligated miRNAs were reverse transcribed into complementary DNA (cDNA) using a reverse transcription (RT) primer with integrated unique molecular indexes (UMI) assignment and the QIAseq miRNA NGS RT Enzyme. This reaction was incubated for one hour at 50°C followed by 15 minutes at 70°C, and then held at 4°C. No libraries were prepared from adapter–dimers. The cDNA was cleaned up using a streamlined magnetic bead-based method. A total of 143 µl of QMN beads were added to the reaction and incubated for five minutes at room temperature. The beads were then washed twice with 200 µl of 80% ethanol. After washing, the beads were air-dried on a magnetic stand for ten minutes at room temperature. The cDNA was eluted by adding 17 µl of nuclease-free water to the tubes. Library amplification was performed using the QIAseq miRNA Library Kit (Qiagen), following the manufacturer’s instructions. Library reactions were prepared using 2 μl of universal forward primer (5’-AACTGTAGGCACCATCAAT-3’) and 2 μl of indexing reverse primers, each containing a unique six-nucleotide multiplex sequence tag (hADSC-Exo: ACAGTG; hBMSC-Exo: GCCAAT; hiPSC-Exo: CCAACA). The reaction mixture also included 15 μl of cDNA Cleanup input, 3 μl of HotStarTaq DNA Polymerase, 15 μl of library buffer, and 42 μl of nuclease-free water, with a final reaction volume of 80 μl. Polymerase chain reaction (PCR) was performed on a thermal cycler with the following conditions: an initial denaturation at 95°C for 15 minutes, followed by 20 cycles of 95°C for 15 seconds, 60°C for 30 seconds, and 72°C for 15 seconds. A final extension step was carried out at 72°C for two minutes. The library cleanup process followed the same magnetic bead-based method described for cDNA cleanup. After eluting the DNA into new tubes, the miRNA sequencing library was ready for quality control (QC) and subsequent NGS. Libraries were sequenced on an Illumina instrument (75-cycle single-end read, 75SE; Illumina, USA). Sequencing data were processed using the Illumina software BCL2FASTQ v2.20. The small RNA sequences generated were subjected to a filtering process to obtain qualified reads. The total read length of the original sequence was 82,471,001, and the mean read length of each sample was 27,490,333 (SD= 65,005,898). Trimmomatic was used to clip the 3′ adaptor sequence, trim or remove the reads according to the quality score, and discard trimmed reads shorter than 18 nucleotides.³⁴ After low-quality data were filtered, the qualified reads were analyzed using miRDeep2³⁵ and RFam (including signal recognition particle RNA (SRP RNA), small nuclear RNA (snRNA), pre-miRNA, small nucleolar RNAs (snoRNA), Y RNA, and transfer RNA (tRNA)), to align the reads to the reference genome downloaded from UCSC (University of California, USA). Only reads that mapped perfectly to the genome five or fewer times were used for miRNA detection, since miRNAs usually map to few genomic locations.

Enrichment analysis

Enrichment analysis was conducted using the R package clusterProfiler (R Foundation for Statistical Computing, Austria).³⁶ To identify significantly enriched terms with a higher-than-expected number of genes based on random distribution, a hypergeometric test was applied.³⁷ For multiple comparison correction and to control the false discovery rate (FDR), Benjamini-Hochberg (BH) adjustment was used, enhancing the reliability of identifying specific functional categories, such as Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, by reducing the likelihood of false positives associated with multiple testing.³⁸

Statistical analysis

The data from our in vitro studies are presented as the means ± SEMs. All experiments were repeated at least three times. Before conducting the analysis of variance (ANOVA) or independent-samples t-test, the normality of the data was assessed using the Shapiro-Wilk test in SPSS. Results confirmed that all samples across each treatment followed a normal distribution. Statistical significance was evaluated by carrying out a one-way ANOVA using SPSS Statistics, and multiple comparisons were performed using the Scheffe test. p < 0.05 was considered to indicate statistical significance.

Results

Identification of hiPSC-, hADSC-, and hBMSC-derived exosomes

To identify exosomes from stem cells with different origins and different potencies, we assessed the characteristics of the exosomes secreted from the hiPSCs, hADSCs, and hBMSCs through TEM, western blotting, and NTA. TEM revealed that the hiPSC-Exos, hADSC-Exos, and hBMSC-Exos successfully isolated from CM displayed a typical rotund and cup shape (Figure 1a). CD9, CD63, CD81 (cytosolic marker), ALIX, TSG101 (lipid-bound protein markers), and β-actin (negative marker) are signature exosomal proteins, and their expression further demonstrated that these vesicles are exosomes (Figure 1b). NTA was used to quantify the total amount of exosomal particles and the density at approximately 2-3E + 10 particles/ml and showed no significant differences among the groups (p > 0.05 for all three groups, Scheffe test) (Figure 1c). NTA also revealed that the mean vesicle diameter was approximately 130 nm (iPSC 134.9 nm (SD 47.5); ADSC 128.48 nm (SD 47.27); BMSC 127.6 nm (SD 49.9)); the particle size distribution did not significantly differ among these three types of stem cells (p > 0.05 for all three groups, Scheffe test) (Figure 1d).

Fig. 1.

Exosome identification. a) Transmission electron microscopy (TEM) image showing the morphology of exosomes from human induced pluripotent stem cells (hiPSCs), adipose tissue-derived stem cells (hADSCs), and bone marrow mesenchymal stem cells (hBMSCs) (scale bar = 20 nm). b) Western blotting was used to analyze the exosomal markers CD9, CD63, CD81, ALIX, and TSG101 and the negative control marker β-actin. Full-length blots/gels are presented in Supplementary Figure a. c) The total amount of exosomes secreted by hiPSCs, hADSCs, and hBMSCs was determined by nanoparticle tracking analysis (NTA). p > 0.05 for all three groups (Scheffe test). d) The particle size of the exosomes.

Length distribution, proportion of noncoding RNAs, and correlation of miRNAs

Exosomal samples derived from three kinds of stem cells, hiPSC-Exos, hADSC-Exos, and hBMSC-Exos, were evaluated. Exosomal miRNA comparison, identification, and prediction analyses were performed on the valid sequencing data, and a total of 46,452,816 valid reads (Quality Value (QV) > 30, 15-30 bp) were identified (Figure 2a). Further analysis revealed that 3.87%, 2.16%, and 12.71% of all the screened sequences were miRNAs, while 96.13%, 97.84%, and 87.29% of the total sequencing data were Rfam RNAs in hiPSC-Exos, hADSC-Exos, and hBMSC-Exos, respectively. Among them, ribosomal RNA (rRNA) was abundantly expressed in the three kinds of stem cell exosomes (38.65%, 94.07%, and 59.76% in hiPSC-Exos, hADSC-Exos, and hBMSC-Exos, respectively). In addition, 52% of the total snRNA was enriched in hiPSC-Exos, and 10.86% of the total Y RNA was highly expressed in hBMSC-Exos (Figure 2b, Supplementary Tables i and ii). Next, we compared the exosomal miRNA expression in each of the two groups of the three kinds of stem cells. The miRNA distribution plot showed linear regression in hADSC-Exos and hBMSC-Exos, indicating a strong correlation in exosomal miRNA expression between hADSC-Exos and hBMSC-Exos (R² = 0.934). However, there was a marked difference between hiPSC-Exos with hADSC-Exos and hBMSC-Exos; the miRNA distribution plot showed a spread shape, and the figure also indicated a weak correlation between hiPSC-Exos with hADSC-Exos (R² = 0.0008) and hBMSC-Exos (R² = 9E-06) (Figure 2c). The statistical results showed that the exosomal miRNAs were very similar between hADSC-Exos and hBMSC-Exos, but the exosomal miRNAs of hiPSC-Exos were unique compared to those of hADSC-Exos and hBMSC-Exos. The results also revealed the top 15 miRNAs in exosomes derived from three types of stem cells (Figure 2d). Therefore, we subsequently focused on the differential expression and comparisons of exosomal miRNAs among these three types of stem cells.

Fig. 2.

Length distribution, proportion of noncoding RNAs, and correlation of microRNAs (miRNAs) in induced pluripotent stem cell (iPSC)-Exos, adipose tissue-derived stem cell (ADSC)-Exos, and bone marrow mesenchymal stem cell (BMSC)-Exos. a) RNA sequence length distribution. b) The proportion of noncoding RNAs, including mature microRNAs (miRNAs), ribosomal RNA (rRNAs), transfer RNAs (tRNAs), Y RNAs, small nucleolar RNAs (snoRNAs), pre-miRNAs, small nuclear RNA (snRNAs), signal recognition particle RNAs (SRP RNAs). c) Linear regression analysis of the correlation of miRNAs between hADSC-Exos and hBMSC-Exos, between hADSC-Exos and hiPSC-Exos, and between hBMSC-Exos and hiPSC-Exos. d) The top 15 miRNAs in hiPSC-Exos, hADSC-Exos, and hBMSC-Exos. rRNA, ribosomal RNA; siRNA, small nuclear RNA.

Heatmap analysis of miRNAs and miRNA-target gene interaction networks

The miRNA expression profile used reads per million miRNA reads (RPM) as a normalization tool to compare the quantitative miRNA expression in three samples, ranging from the highest to ~5,000 RPM defined as high expression. Heatmap analysis revealed highly expressed exosomal miRNAs in hiPSCs and compared the exosomal miRNA expression between hADSC-Exos and hiPSC-Exos, and between hBMSC-Exos and hiPSC-Exos. The most highly expressed miRNAs are represented in red, and conversely, the miRNAs with the weakest expression are represented in blue. We identified several miRNAs with the highest expression in hiPSC-Exos but the lowest expression in hADSC-Exos or hBMSC-Exos, such as hsa-miR-302b-3p, hsa-miR-20b-5p, hsa-miR-302d-3p, and hsa-miR-302a-3p. Moreover, several highly expressed miRNAs, such as hsa-miR-16-5p, hsa-miR-93-5p, and hsa-miR-25-3p, coexisted in these exosomes (Figure 3a). Subsequently, we increased the screening strategy to ~10,000 RPM to decrease the number of highly expressed exosomal miRNAs from 33 to ~15 miRNAs. The target genes of these highly expressed miRNAs were predicted using TargetScan (version 7.2; targetscan.org), miRanda (microrna.org), and miRDB (mirdb.org), and 5,952, 5,332, and 5,112 target mRNAs were predicted in hiPSCs, hADSCs, and hBMSCs, respectively (Figure 3b). To reduce the number of predicted miRNA targets, we increased the target score threshold to 99 and analyzed the exosomal miRNA-target gene interaction networks using Cytoscape.³⁹ The independent putative hiPSC, hADSC, and hBMSC exosomal miRNAs and their targets were integrated and are presented in Figure 3c. We found that the miRNAs of hiPSCs greatly differed from those of hADSCs and hBMSCs, but the miRNAs of hADSCs and hBMSCs were very similar (Figure 3c). Moreover, we examined the common miRNAs in the three kinds of exosomes and found that the most common exosomal miRNAs in hiPSCs, hADSCs, and hBMSCs were hsa-miR-16-5p, hsa-miR-25-3p, and hsa-miR-93-5p, and their target genes were PHF19, CPR137C, PKD2, ZNFX1, ENPP5, B3GALT2, FBXW7, CD69, FNIP1, and MAN2A1 . According to these results, we further analyzed the biological functions and signalling regulatory pathways of individual or common miRNAs and target genes in hiPSCs, hADSCs, and hBMSCs.

Fig. 3.

Heatmap analysis of microRNAs (miRNAs) and miRNA-target gene interaction networks in induced pluripotent stem cell (iPSC)-Exos, adipose tissue-derived stem cell (ADSC)-Exos, and bone marrow mesenchymal stem cell (BMSC)-Exos. a) Heatmap. b) The highly expressed miRNAs in human iPSCs (hiPSCs), hADSCs, and hBMSCs and their target genes. c) The exosomal miRNA-target gene interaction network in hiPSCs, hADSCs, and hBMSCs.The exosomal miRNAs and targets that did not overlap in the three stem cell-derived exosomes are marked in green, and those that overlapped in the hADSCs and hBMSCs are shown in blue.

GO and KEGG enrichment of highly expressed miRNAs

We performed GO analysis to identify the biological processes, cellular components, and molecular functions associated with the identified exosomal miRNA target genes in hiPSCs, hADSCs, and hBMSCs (Figure 4a). The highly expressed exosomal miRNA target genes in hiPSC-Exos were significantly enriched in 1,832 GO pathways (the p-value threshold of < 0.05 was used as a selection criterion). Among these pathways, 1,491 GO pathways were associated with biological processes, 213 were associated with cellular components, and 128 were associated with molecular functions. The major target genes were associated with neuroprojection organization, extracellular exosome biogenesis, EV biogenesis, post-synaptic endosomes, axonal growth cones, nuclear import signalling, and receptor activity. The highly expressed exosomal miRNA target genes in hADSC-Exos were significantly enriched in 1,493 GO pathways (the p-value threshold of < 0.05 was used as a selection criterion), among which 1,240 GO pathways were associated with biological processes, 154 with cellular components, and 99 with molecular functions. The major target genes of the hADSC-Exos were associated with collagen fibril organization, chondrocyte development, bone resorption, cartilage development, complex of collagen trimers, platelet-derived growth factor receptor binding, and glucocorticoid receptor binding. In addition, the highly expressed exosomal miRNA target genes in the hBMSC-Exos were significantly enriched in 1,336 GO pathways (the p-value threshold of < 0.05 was used as a selection criterion). Among these pathways, 1,153 GO pathways were associated with biological processes, 104 were associated with cellular components, and 79 were associated with molecular functions. The major target genes of the hBMSC-Exo miRNAs were associated with axonogenesis, chondrocyte differentiation, ossification, collagen trimer complex formation, neurotrophin receptor binding, and vascular endothelial growth factor binding. These results indicated that the GO pathways of exosomal miRNA target genes were more diverse in hiPSCs than in hADSCs and hBMSCs. The common biological functions of these GO pathways were chondrocyte differentiation, cartilage development, and ageing. Finally, the target genes were subjected to KEGG pathway enrichment analyses on ClusterProfiler (version 3.18.1; the p-value threshold of < 0.05, q < 0.05 was used as a selection criterion).³⁶ The target genes were significantly enriched in 113, 111, and 102 KEGG pathways in hiPSCs, hADSCs, and hBMSCs, respectively. The results indicated that the hiPSC-Exos miRNAs significantly affected the following processes: mitogen-activated protein kinase (MAPK) signalling, Hippo signalling, signalling pathway regulating pluripotency of stem cells, longevity regulating pathway, PI3K-Akt, Wnt signalling, focal adhesion, cell senescence, and cell cycle. The hADSC-Exos miRNAs significantly affected the following processes: MAPK signalling, PI3K-Akt, focal adhesion, cellular senescence, signalling pathway regulating pluripotency of stem cells, etc. The hBMSC-Exos miRNAs significantly affected MAPK signalling, signalling pathway regulating pluripotency of stem cells, FoxO signalling, and Wnt signalling. MAPK and regulating pluripotency of stem cells are the major and common signalling pathways in the three groups. The database included GO terms and KEGG enrichment pathways for Annotation, Visualization, and Integrated Discovery (DAVID)^37,40-42 (Figures 4a and 4b).

Fig. 4.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses of the significantly highly expressed microRNAs (miRNAs) in induced pluripotent stem cell (iPSC)-Exos, adipose tissue-derived stem cell (ADSC)-Exos, and bone marrow mesenchymal stem cell (BMSC)-Exos. a) GO enrichment summary of exosomal miRNA target genes according to GO term classification. b) KEGG enrichment histogram for analyzing exosomal miRNA target genes. MAPK, mitogen-activated protein kinase.

The overlapping highly expressed exosomal miRNAs and their target genes

The overlap region of the three circles represents 228 common exosomal miRNAs in hiPSCs, hADSCs, and hBMSCs (Figure 5a). There were 272, 32, and 45 exosomal miRNAs that did not overlap in hiPSCs, hADSCs, and hBMSCs, respectively, indicating that hiPSC-derived exosomes have more unique miRNAs than those from hADSCs and hBMSCs. We selected these overlapping and highly expressed miRNAs (RPM > 10,000) from three samples, and three of the top 228 miRNAs met these criteria (Figure 5b). The three overlapping miRNAs were hsa-miR-16-5p, hsa-miR-93-5p, and hsa-miR-25-3p. We also used TargetScan, miRanda, and miRDB software to predict the target genes of these three miRNAs, and 2,792 exosomal miRNA target genes were predicted for GO and KEGG enrichment analysis. The commonly expressed exosomal miRNA target genes in three samples were significantly enriched in 580 GO pathways (the p-value threshold of < 0.05 was used as a selection criterion), among which 463 GO pathways were associated with biological processes, 66 were associated with cellular components, and 51 were associated with molecular functions (Figure 5c). Moreover, the target genes were significantly enriched in 69 KEGG pathways (Figure 5d). We found that these six overlapping miRNA target genes can be used to focus on common effects in individual stem cell exosomes. The primary biological results of the GO and KEGG pathway analyses were related to bone/cartilage development and MAPK signalling. The predicted miRNA-gene interaction network was analyzed and is presented in Figure 5e.

Fig. 5.

The overlapping highly expressed exosomal microRNAs (miRNAs) and their target genes were analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). a) The intersection of exosomal miRNAs derived from human induced pluripotent stem cells (hiPSCs), adipose tissue-derived stem cells (hADSCs), and bone marrow mesenchymal stem cells (hBMSCs). b) The highly expressed miRNAs in three samples and their target genes. c) Network analysis of the top six commonly expressed exosomal miRNAs and their target genes in hiPSCs, hADSCs, and hBMSCs. The orange strips in the network indicate exosomal miRNAs, whereas the green ovals indicate their target genes. d) GO enrichment summary of exosomal miRNA target genes by GO term classification. e) KEGG enrichment histogram for analyzing exosomal miRNA target genes. MAPK, mitogen-activated protein kinase; RPM, reads per million miRNA reads.

Exosomes improved chondrocyte function and reduced cell senescence

Bioinformatics analysis revealed that these exosomal miRNAs promoted chondrocyte differentiation, increased ECM synthesis, and reduced ageing; thereafter, we investigated exosome uptake by hACs to confirm the biological effects of exosomal miRNAs, especially their chondrogenic functions, from hiPSCs, hADSCs, and hBMSCs. First, after treatment of hACs with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos for one, three, or five days, there was no statistically significant difference in cell viability among the three exosome treatment groups (p > 0.05 compared to the control group, independent-samples t-test; Figure 6a). Healthy articular cartilage is mainly composed of hyaline cartilage, which contains type II collagen and negatively charged proteoglycans (GAGs) as the primary matrix components, and these components play a crucial role in maintaining cartilage function. However, in early OA, the surface of articular cartilage can become impaired, and type I collagen, which is typically found in fibrocartilage, may begin to replace type II collagen. Therefore, we measured the GAG content and type I and II collagen content using DMMB assays and ELISAs. Treatment with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos for 12 days significantly increased the GAG content (p = 0.026, p = 0.016, and p = 0.018 compared to the control group, respectively; independent-samples t-test) and type II collagen (p < 0.001 compared to the control group, two-way ANOVA and Scheffe’s method) but suppressed type I collagen (p = 0.022, p = 0.003, p < 0.001 compared to the control group, respectively; two-way ANOVA and Scheffe’s method) in in vitro cultured hACs (Figures 6b to 6d). In addition, the cell senescence pathway was identified by the GO and KEGG pathway analyses. To verify the biological function of these exosomes, we detected the most common marker of senescence, SA-β-gal. Treatment with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos for 12 days significantly inhibited the number of SA-β-Gal-positive cells compared with that in the control group (p < 0.001 compared to the control group, two-way ANOVA and Scheffe’s method) (Figure 6e). Among them, hBMSC-Exos had a better effect than the other treatments (p < 0.001 compared to the ADSC-Exos and iPSC-Exos group, two-way ANOVA and Scheffe’s method). In summary, we demonstrated that exosomes derived from hiPSCs, hADSCs, and hBMSCs improved several biological functions, including ECM secretion, cartilage development, and cell senescence in cultured hACs.

Fig. 6.

Comparison of human articular chondrocytes (hAC) functions after treatment with human induced pluripotent stem cell (hiPSC)-Exos, adipose tissue-derived stem cell (hADSC)-Exos, or bone marrow mesenchymal stem cells (hBMSC)-Exos. a) MTS assays, b) 1,9-dimethylmethylene blue (DMMB) assays, and c) type II and d) type I collagen enzyme-linked immunosorbent assays (ELISAs) were used to measure cell viability, glycosaminoglycan (GAG) content, and collagen protein levels in hACs treated with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos (10^9 particles/ml) for 12 days. The data are presented as the optical density (OD) ratio and protein concentration, and each column represents the mean ± standard error of the mean (SEM) of six replicate cultures. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group; #p < 0.05, ##p < 0.01, ###p < 0.001 compared with the treatment groups, two-way analysis of variance (ANOVA) and Scheffe’s method. e) Senescence-associated β-galactosidase (SA-β-Gal) staining was used to detect hAC senescence in hACs treated with hiPSC-Exos, hADSC-Exos, or hBMSC-Exos (10^9 particles/ml) for 12 days. The upper panel shows that the SA-β-Gal-positive cells in the control, hiPSC-Exo, hADSC-Exo, and hBMSC-Exo treatment groups are blue (scale bar = 150 µm). The quantification data are presented as the ratio of SA-β-Gal-positive cells/total cells in the lower panel, and each column represents the mean ± SEM of ten replicate cultures. ***p < 0.001, compared with the control. ###p < 0.001, compared with the treatment groups. ns, non-significant.

Discussion

In this study, we hypothesized that exosomes from stem cells with different hierarchical potencies may exert their biological effects through the miRNAs they carry. We found that the miRNA profiles of hBMSC-Exos and hADSC-Exos were similar but largely different from those of hiPSC-Exos. Putative targets of highly expressed miRNAs in hBMSC-Exos and hADSC-Exos were functionally enriched in collagen fibril organization, comprising structural changes in the ECM, cartilage development, and senescence. Interestingly, the same but more extensive functions, including neuron projection organization, axonal growth corn, wound healing, and transcriptional activity, were also identified in hiPSC-Exos. In this study, we also focused on the intersection of highly expressed exosomal miRNAs and their common biological functions in these three kinds of stem cells. The common miRNAs are hsa-miR-16-5p, hsa-miR-93-5p, and hsa-miR-25-3p, and their putative targets included MAN2A1, ZNFX1, PHF19, GPR137C, ENPP5, B3GALT2, FNIP1, PKD2, and FBXW7. The common functional roles of these predicted genes are mainly in cell growth, bone ossification, and cartilage development/differentiation, which may be mediated through the MAPK signalling pathway. Based on this bioinformatic analysis of miRNA profiles of exosomes from different origins, we further found that three kinds of stem cell-derived exosomes maintained the viability of hACs, promoted chondrogenic functions, and inhibited senescence. Notably, among the tested exosomes, hBMSC-Exos exhibited the greatest efficacy, while hiPSC-Exos showed the lowest efficacy in maintaining normal chondrocyte function.

Previous studies have shown that iPSCs can be induced to differentiate into a range of functional cell types, leading to several clinical trials for treatment of various diseases.⁴ However, there are still many concerns regarding the safety of iPSCs and the disadvantages of iPSCs compared to MSCs. Therefore, safety concerns and immunosuppression issues make MSCs superior for regenerative applications. Currently, MSC clinical trials are mainly Phase I or II trials, and more Phase III studies are needed to prove their therapeutic efficacy.⁴³ Exosomes carry cargoes of specific miRNAs, small molecules, and proteins; these biomolecules delivered into recipient cells may replace the partial functions of transplanted stem cells; and exosomes have recently emerged as a promising cell-free tissue engineering therapy. Through small RNA sequence analysis, we found that hiPSC-Exos contain a greater quantity of noncoding RNA fragments than both hBMSC-Exos and hADSC-Exos. These noncoding RNAs included tRNA, snRNA, Y RNA, SRP RNA, snoRNA, and miRNA. Notably, the snRNAs constituted 55.9% of the noncoding RNAs in the hiPSC-Exos but not in the hBMSC-Exos (3.04%) or hADSC-Exos (0.55%). Because snRNAs are crucial spliceosome components⁴⁴ and are known for their high transcriptional activity in hiPSCs,⁴⁵ the abundance of snRNAs in hiPSC-Exos found in this study indicated that these molecules may contribute to the increased transcriptional ability of recipient cells. In contrast to hiPSC-Exos, miRNAs exhibited a highly similar profile in hADSC-Exos and hBMSC-Exos.

The most highly expressed miRNAs found in the hiPSC-Exos, hADSC-Exos, and hBMSC-Exos were miR-16-5p, miR-25-3p, and miR-93-5p, respectively. The functions of these three miRNAs’ putative target genes, such as PDK2, FBXW7, ZNFX1, PHF19, and ENPP5, were predicted to be associated with the promotion of osteogenic and chondrogenic differentiation. Previous studies have also reported the regulatory effects of these target genes on articular chondrocytes.^46-50 For example, pyruvate dehydrogenase kinase (PDK)-2 significantly increased reactive oxygen species (ROS), oxidative stress, and senescence in chondrocytes and in the articular cartilage of a DMM-induced OA model in vivo. PDK2 deficiency alleviated DMM-induced cartilage destruction.⁴⁶ Accordingly, miR-93-5p may deplete PDK2 to alleviate cartilage degradation in OA. Because Ecto-nucleotide pyrophosphatase/phosphodiesterase (ENPP)-5 plays a role in the regulation of bone mineralization,⁴⁷ miR-93-5p may downregulate ENPP5 to suppress bone mineralization. Zinc finger NFX1-type containing 1 (ZNFX1) antisense RNA 1 (ZFAS1) was shown to promote chondrocyte proliferation and reduce oxidative stress, and is a promising therapeutic target for OA.⁴⁸ miR-93-5p may facilitate proliferation and repress oxidative stress, inflammation, and apoptosis in osteoarthritic chondrocytes by suppressing ZNFX1. F-box and WD repeat domain-containing-7 (Fbxw7) reportedly mediates ubiquitylation to control osteogenesis and inhibit chondrogenesis.⁴⁹ miR-25-3p-mediated Fbxw7 ablation may promote chondrogenesis. Moreover, PHD finger protein 19 (PHF19) is associated with a risk of rheumatoid arthritis,⁵⁰ and the associations between miRNAs and their putative regulatory genes need further investigation. In this study, KEGG signalling pathway enrichment analysis revealed that the common exosomal miRNAs mainly play a role in the MAPK, PI3K-Akt, and Hippo signalling pathways, cellular senescence, and longevity-regulating pathways. Accordingly, we found that the common effect of these three tested exosomes may be related to cell differentiation and ageing.

In contrast, we found that hADSC-Exos and hBMSC-Exos shared seven common miRNAs, suggesting that hADSC-Exos and hBMSC-Exos might have more similar functions than hiPSC-Exos. The miRNA profiles of hADSC-Exos and hBMSC-Exos were more capable of regulating collagen fibril organization and cartilage and bone development through the MAPK, PI3K-Akt, or FoxO pathways. Since hADSCs and hBMSCs originate from the mesoderm, they might have similar potential to promote the differentiation of cartilage and bone. In contrast to the miRNA profile of hiPSC-Exos, the miRNA profile of hiPSC-Exos may have a stronger effect on neuron development and vesicle biogenesis than that of hADSC-Exos or hBMSC-Exos. The miRNA targets in hiPSC-Exos were found to be involved in Hippo signalling, Wnt signalling, cell junctions, and the p53 signalling pathway, indicating that they may influence cell senescence and metabolism. Moreover, the miR-302 family was found to be significantly enriched in hiPSC-Exos, indicating that it may play a pivotal role in influencing stem cell proliferation, differentiation, and reprogramming.^51-54

In comparison with a previous study by Bi et al⁵⁵ on hiPSC-Exos, human embryonic stem cell (hESC)-Exos, and human umbilical cord mesenchymal stem cell (hUCMSC)-Exos, our findings revealed that the miRNA profiles of hiPSC-Exos shared the same highly expressed miRNAs (miR-19b-3p, miR-20a-5p, miR-302d-3p, miR-302a-3p, miR-93-5p, miR-92a-3p, miR-148a-3p, and miR-21-5p) with those reported in their study. The highly expressed miRNAs in the hADSC-Exos and hBMSC-Exos in our study overlapped with seven miRNAs found in the hUC-MSC-Exos (miR-21-5p, let-7b-5p, miR-191-5p, miR-199a-3p, let-7i-5p, miR-100-5p, and miR-25-3p) in Bi et al’s study. These findings indicated that the miRNA profiles of exosomes from a similar hierarchy are more similar. Bi et al’s study indicated that the putative targets of miRNAs, hESC-Exos, and hiPSC-Exos play major roles in regulating development, metabolism, and anti-aging processes.⁵⁵ This finding is similar to our findings in hiPSC-Exos. Bi et al’s research did not specify the ages of the hESCs, hiPSCs, and hUC-MSCs in their study, suggesting that age is unlikely to be a confounding factor affecting the bioinformatics results.

Functional analysis of miRNA typically relies on computational target gene prediction followed by GO and KEGG pathway analysis. The limitations of these methods are that they do not consider the tripartite network topology between miRNAs, target genes, and functional annotations. Moreover, the highly false-positive nature of sequence-based target prediction algorithms causes the propagation of annotation errors. Integrating miRNA and mRNA expression data can strengthen the target predicted and reduce false-positive results in exosome-treated chondrocytes. It is necessary to validate miRNA-target interactions in future studies. Notably, several parameters for cell culture conditions could influence exosome yields or biological functions, such as passage number, cell seeding density, or CM collection time.⁵⁶ A previous study indicated that increasing the collection frequency and decreasing cell seeding density could increase exosome production, while a passage number beyond 4 results in lower pro-vascularization bioactivity.⁵⁷ Unfortunately, when comparing the internal exosome content, such as miRNA, proteins, etc., from different kinds of MSCs, controlling such factors such as passage number and cell seeding density is challenging due to different cell isolation procedures and seeding density requirements. In this study, prior to exosome collection in cultures, we have checked the characterization markers of hADSCs and hBMSCs to ensure that the marker gene expressions and phenotypes are in line with ISCT guidelines. As hADSCs were able to maintain their characteristics longer than hBMSCs in our system, we used hADSCs at passages 6-8 (P6-P8) and hBMSCs at passage 3 (P3). This is still a limitation, however, for a study comparing internal exosome content among different types of cells.

Several studies have reported that MSC-derived exosomes were administered in patients without side effects and safety concerns.^21-25 The potential off-target effects of exosomal miRNAs may be due to a specific miRNA targeting several mRNAs, thus affecting related cellular pathways. Therefore, efficiently delivering miRNA to the precise location without undesired effects is crucial for exosome treatments clinically. This is particularly the case for the joint capsule as it resembles a sac-like envelope enclosing the joint. For future clinical administration, intra-articular (IA) injections of exosomes may allow easier control of the dose and location; however, the effect on cells other than chondrocytes, such as synovial cells, should also be considered.

The potential clinical application of stem cell-derived exosomes in the future could be used for cartilage defects or osteoarthritic treatments. Firstly, optimization of exosome production, a high-throughput cellular source, a reproducible and large-scale production, and a standardized isolation protocol are required. Secondly, bioengineering approaches to improve delivery methods include loading exosomes with specific therapeutic molecules or modifying their surface properties to improve the targeting and uptake into chondrocytes. Additionally, more pre-clinical studies should be undertaken before clinical translation, including the efficacy of exosomes and detailed effective mechanisms, as well as long-term safety and efficacy comparisons with other emerging treatments for OA.

In conclusion, in this study we demonstrated that hBMSC-Exos and hADSC-Exos have similar miRNA profiles but are largely different from hiPSC-Exos. However, the common biological functions of these three kinds of stem cell-derived exosomes can effectively suppress cell senescence and improve chondrocyte functions, especially hBMSC-Exos, which have a better chondrogenic ability than hiPSC-Exos and hADSC-Exos in vitro. This finding suggests that the common effects of the three kinds of exosomes on chondrocytes have a high possibility of acting through the common miRNAs, but the other mechanisms still need further investigation. Using exosomes derived from stem cells and delivering miRNAs or other noncoding RNAs may be potential mechanisms for improving normal functions and suppressing the senescence of hACs. Overall, the application of stem cell exosomes may be a novel therapeutic approach for cartilage regeneration.

Data Availability

All data generated or analyzed during this study are included in the published article and/or in the supplementary material.