Molecular Mechanisms of Exosomes From Human iPSC-Cardiomyocytes and Mesenchymal Stem Cells in Restoring the Injured Myocardium.
Oct 23, 2025DOI: 10.1161/jaha.124.037005Journal ArticleAHA
Written by: Eileen TzngNathan BayardoGentaro IkedaNirmal VadgamaHiroyuki TakashimaJennifer LyonsHirotaka IekiMihoko BennettMichael SnyderJoseph C WuConnor G O'BrienPhillip C Yang
Animal care and interventions were overseen by the Stanford University Administrative Panel on Laboratory Animal Care. Studies were carried out in accordance with the Laboratory Animal Welfare Act, while adhering to the “Guide for the Care and Use of Laboratory Animals” (www.nap.edu/catalog/5140.html) and Animals in Research: Reporting In Vivo Experiments guidelines.20 Administrative Panel on Laboratory Animal Care protocol number 16066 described the research question, experimental design, and procedural implementation. No criteria were set for including or excluding animals. One sample was omitted from transcriptional analysis due to suspected batch effect introduced during upstream processing of tissue samples. All research was performed at Stanford University in Stanford, California. The data that support the findings of this study are available from the corresponding author upon reasonable request.Exosomes were harvested from StemPro BM Mesenchymal Stem Cells (ThermoFisher Scientific, Waltham, MA) derived from human bone marrow. Cells were cultured to 60% confluence using MesenPro Media (12746012; ThermoFisher Scientific) in flat‐surfaced 75 cm2 Tissue Culture Flasks. MSCs were washed with PBS to remove xenogenic exosomes from MesenPro serum, and incubated in high glucose DMEM containing 10% exosome free FBS for 48 hours.Human iPSCs were expanded with Essential 8 pluripotent stem cell medium (E‐8 media; Life Technologies, Carlsbad, CA) on Matrigel (Corning, NY) coated 6‐well plates (9.6 cm2 per well). At 80% confluence, media was substituted for differentiation medium (RPMI 1640, GlutaMAX Supplement, and B‐27 Supplement Minus Insulin, ThermoFisher Scientific A1895601) and treated with 6 μM GSK inhibitor CHIR for 48 hours for mesodermal specification. Media was replaced with differentiation medium without CHIR on day 3. Cardiac differentiation was initiated by adding 2 μM C59 (Wnt inhibitor) in differentiation medium for days 4 to 6. Upon observation of contractions, cells were cultured in metabolic selection medium (RPMI 1640 minus glucose, GlutaMAX Supplement, with B‐27 Supplement 17504044; ThermoFisher Scientific) for 5 days. Following selection, cells were allowed to rest for 48 hours in maintenance media (RPMI 1640 medium with B‐27 Supplement), trypsinized, and replated onto Matrigel coated plates. iCMs were matured for 30 days and assessed for ≥80% purity with flow cytometry before harvesting supernatants.Exosome isolation was facilitated by polyethylene glycol precipitation according to previously validated protocols.11, 13 Briefly, cell debris was removed from supernatant by centrifugation at >1000 for 10 minutes, followed by vacuum filtration through a 0.22‐μm syringe filter. The filtrate was then incubated in 25% PEG‐8000 at 4 °C for 16 hours. The resulting mixture was then centrifuged at 1500 for 30 minutes at 4 °C. Supernatant was aspirated, leaving 5 mL remaining and finally centrifuged at 300 for 5 minutes. The residual supernatant was removed, and pellets were resuspended in filtered PBS and stored in −80 °C until administration. Exosome isolates were administered within 1 week of isolation.
g
g
g
NTA was used to quantify the concentration of particles in resuspended pellets, enabling preparation of aliquots at the desired injection concentration. Three measurements of the same sample were imaged using a 640 nm laser with a NanoSight LM20 (NanoSight Ltd, Malvern, UK). Particle size and velocity in the fixed chamber was used to determine the mean size and SD values of the particles using NTA 3.1 Build 3.4.4 (NanoSight Ltd).
All procedures were conducted in the Stanford Translational Research Laboratory located in the Falk Cardiovascular Research Building. Animals underwent a 48‐ to 72‐hour acclimation period following arrival at the facility. All procedures were performed under sufficient anesthesia, using inhaled isoflurane (1%–2%) and oxygen (1–2 L) to maintain animals free of pain and distress. This method is consistent with the recommendation of the Panel on Euthanasia of the American Veterinary Medical Association. Metamizol was added to the drinking water for the first 3 postoperative days and any other manipulations were minimized in this period. In addition, discomfort, distress, pain, and injury were minimized by preanesthesia with intramuscular injection of telazol (6 mg/kg) followed by endotracheal intubation and maintenance of general anesthesia with 1% to 2% isoflurane. If necessary, the animals were managed by the animal facilities of the Department of Comparative Medicine and left to recover, provided with a heating source, under continuous veterinarian supervision, and daily rounds. There were no expected or unexpected adverse events during the course of animal care and monitoring. Humane end points were set for all animals for this study: During the procedure: low blood pressure caused by excessive blood loss and sustained resuscitation is unsuccessfulDuring or after recovery: Anorexia: Rapid deterioration of body condition with >15% to 20% weight loss over 3 to 4 daysAmbulation: Inability to rise/ambulate and access food or water secondary to irreversible causeMentation: Devastating neurologic injury resulting in permanent or overly prolonged (>1–7 days) quadriplegia, paraplegia, or hemiplegiaAnimals that remain comatose or moribund following surgery or do not wake from anesthesiaVocalization: Monitoring signs of distress or painRespiration: Prolonged labored respiration not amenable to diuresis with animal in obvious distressWound healing: Necrotizing wound infection or wound infection that fails to respond to drainage or antibioticsHeart rhythm: Following 15 minutes of unsuccessful resuscitation (no return of spontaneous circulation) with drugs, electrolyte correction, cardiopulmonary resuscitation, or defibrillation on animals with a witnessed cardiorespiratory arrest/malignant arrhythmia.
During the procedure: low blood pressure caused by excessive blood loss and sustained resuscitation is unsuccessful
During or after recovery: Anorexia: Rapid deterioration of body condition with >15% to 20% weight loss over 3 to 4 daysAmbulation: Inability to rise/ambulate and access food or water secondary to irreversible causeMentation: Devastating neurologic injury resulting in permanent or overly prolonged (>1–7 days) quadriplegia, paraplegia, or hemiplegiaAnimals that remain comatose or moribund following surgery or do not wake from anesthesiaVocalization: Monitoring signs of distress or painRespiration: Prolonged labored respiration not amenable to diuresis with animal in obvious distressWound healing: Necrotizing wound infection or wound infection that fails to respond to drainage or antibioticsHeart rhythm: Following 15 minutes of unsuccessful resuscitation (no return of spontaneous circulation) with drugs, electrolyte correction, cardiopulmonary resuscitation, or defibrillation on animals with a witnessed cardiorespiratory arrest/malignant arrhythmia.
Anorexia: Rapid deterioration of body condition with >15% to 20% weight loss over 3 to 4 days
Ambulation: Inability to rise/ambulate and access food or water secondary to irreversible cause
Mentation: Devastating neurologic injury resulting in permanent or overly prolonged (>1–7 days) quadriplegia, paraplegia, or hemiplegia
Animals that remain comatose or moribund following surgery or do not wake from anesthesia
Vocalization: Monitoring signs of distress or pain
Respiration: Prolonged labored respiration not amenable to diuresis with animal in obvious distress
Wound healing: Necrotizing wound infection or wound infection that fails to respond to drainage or antibiotics
Heart rhythm: Following 15 minutes of unsuccessful resuscitation (no return of spontaneous circulation) with drugs, electrolyte correction, cardiopulmonary resuscitation, or defibrillation on animals with a witnessed cardiorespiratory arrest/malignant arrhythmia.
Any pain or distress that animals may experience was treated with appropriate anesthetics and analgesics. Daily monitoring signs may have also included body temperature, hydration status, along with urination and defecation.
The sample size for our study was informed by a preliminary investigation that included a human amniotic MSC treatment arm and a normal saline control arm, using 5 to 7 Yorkshire swine per group. In this prior study, we observed statistically significant improvements in key functional and structural outcomes, including LVEF, left ventricular (LV) end‐diastolic diameter, and peri‐infarct region viability, in the human amniotic MSC–treated group compared with controls.
Specifically, LVEF showed a significant increase in the treatment group, with a mean improvement of 14% and an SD of 6%. The calculated effect size was Hedges's g=2.627 (95% CI, 1.135–4.120), indicating a large effect and suggesting a meaningful clinical impact of human amniotic MSC therapy. While this specific analysis was not independently replicated, the preliminary findings provide a strong rationale for the anticipated efficacy of iCM‐Exo treatment in the present study.
With approval and under the guidelines of Stanford Veterinary Service Center and Stanford Administrative Panel on Laboratory Animal Care (No. 16066), 15 female Yorkshire swine (Sus scrofa domestica, aged 3–5 months for duration of model) were anesthetized with isoflurane and subject to myocardial infarction using a 10‐mm over‐the‐wire angioplasty balloon inserted through the femoral artery and placed in the proximal left anterior descending coronary artery at the first diagonal branch and inflated for 60 minutes and subsequently reperfused. Left anterior descending coronary artery occlusion was confirmed by contrast agent using the Allura Biplane Fluoroscopy unit (Philips, Eindhoven, Netherlands), and confirmation was repeated at 10‐minute intervals during the occlusion period. Immediately following deflation, treatment was delivered to the peri‐infarct region. An interventionalist and a magnetic resonance imaging (MRI) technician were blinded to the treatment group during delivery of treatment, MRI analysis of cardiac images, and group analysis. The outcome measures assessed are LVEF, LV volumes, infarct size, viability of the peri‐infarct region, and changes in gene expression in treated porcine myocardium.
The immunosuppressant cyclosporine A (10 mg/kg oral administration, Sandimmune formulation; Novartis Pharma AG, Basel, Switzerland) was administered to animals 3 days before and 2 days after the intramyocardial injection procedure. At the beginning of the procedure, a 1‐time dose of 15 mg/kg methylprednisolone sodium succinate was administered intravenously. Treatments were delivered to the peri‐infarct region. Ten 0.5‐mL injections containing PBS as a control treatment (n=4) or 5×1011 exosomes isolated from either iCMs (n=6) or MSCs (n=5) were delivered transendocardially to the peri‐infarct region using Helix Helical Infusion System, or a Morph Universal Deflectable Guide Catheter (BioCardia Inc, Sunnyvale, CA) as previously described.5 Fluoroscopy was used to visually administer and map injection sites surrounding the infarct region. Randomization was not used to allocate experimental units to control and treatment groups. The interventionalist was blinded to the treatment group during delivery of treatment.
Animals underwent cardiac MRI at weeks 2 and 4 after treatment to measure myocardial function and fibrosis as described previously.21, 22 A Signa 3T EXCITE scanner (GE Healthcare, Chicago, IL) and a phased array 2 channel surface coil (Rapid MR International, Columbus, OH) was used for cardiac imaging. Briefly, steady‐state free procession (repetition time, 3.4 ms; echo time, min‐full; flip angle 45°; thickness 8 mm; matrix 224×224; and field of view, 28 cm) was used to acquire cine images. Right and left anterior‐oblique ventriculograms were generated to identify the peri‐infarct region with manganese‐enhanced MRI and delayed gadolinium enhancement MRI. Manganese‐enhanced MRI was obtained using fast gradient echo‐inversion recovery sequence (repetition time, 6.2 ms; echo time, 2.9 ms; flip angle, 15°; thickness, 8 mm; matrix, 224×192; field of view, 28 cm; inversion time, 300–600 ms) 25 to 40 minutes after intravenous infusion of EVP1001‐1 at 20.3 μmol/kg (2.0 μmol/kg per min) (Eagle Vision Pharmaceutical Corp; Downington, PA). After a 30‐minute washout period, 0.2 mmol/kg of gadolinium‐DTPA (Magnevist, Bayer Health Care Pharma AG, Berlin, Germany) was administered. Delayed gadolinium enhancement MRI was obtained using fast gradient echo‐inversion recovery sequence (inversion time, 250–300 ms) 10 to 25 minutes later. Manganese‐enhanced MRI is expressed as a percentage of the entire left ventricle. The MRI technician was blinded to the treatment group during acquisition and analysis of cardiac images. The outcome measures assessed are LVEF, LV volumes, infarct size, and myocardial viability.
At week 4, following the final cardiac MRI, swine were euthanized with an intravenous potassium chloride injection. Explanted hearts were stained with Evans Blue and tetrazolim chloride to delineate the peri‐infarct region.21 One percent Evans blue dye was injected into the left (60 mL) and right (30 mL) coronary arteries to delineate Evans blue–negative (ischemic region) versus Evans blue–positive perfused (normal) myocardium after occlusion at the same left anterior descending coronary artery site used in the initial infarction procedure. The excised heart was then cut into 1 cm‐thick cross‐sectional slice, incubated in 1% tetrazolim chloride solution, and fixed with 10% formalin. Tetrazolim chloride stains viable myocardium red and leaves nonviable myocardium white. The peri‐infarct region was designated as viable (tetrazolim chloride–positive) but ischemic (Evans blue–negative) myocardium, of which tissue from this specified region was used for subsequent transcriptomic studies.
Total RNA from the peri‐infarct region of explanted hearts was isolated using the Qiagen Rneasy Plus Minikit (74134; Qiagen, Hilden, Germany), according to the manufacturer's instructions. RNA concentration and purity were assessed using the NanoDrop spectrophotometer (ThermoFisher Scientific), and RNA integrity was evaluated using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA).
RNA samples were submitted to Novogene (Sacramento, CA) for RNA sequencing using the Novaseq 6000 platform (Illumina, San Diego, CA) with an S4 flow cell. For library preparation, 1 μg of high‐quality RNA per sample was used as input material. Sequencing libraries were constructed using the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, Los Angeles, CA), following the manufacturer's protocol.
Conditioned media from the iCM‐Exo and MSC‐Exo treatment groups (n=2 biological replicates per group) were submitted to System Biosciences (Palo Alto, CA) for Exo‐NGS services. Extracellular vesicles were isolated using the ExoQuick Exosome Isolation Kit (System Biosciences), followed by RNA purification with the System Biosciences spin column‐based RNA Purification Kit. Exosomal RNA quality and quantity were assessed with Agilent RNA 6000 Pico chips on the Agilent 2100 Bioanalyzer. Next‐generation sequencing was performed on the Illumina NextSeq platform, with each sample sequenced to a depth of approximately 10 to 15 million single‐end reads (75 bp read length).
Raw RNA‐seq reads were subjected to quality control using FastQC version 0.11.9 and trimmed for adapter sequences and low‐quality bases using Trimmomatic version 0.39. Clean reads were aligned to the Sus scrofa reference genome (Sscrofa11.1) for porcine tissue RNA sequencing and to the human reference genome (GRCh38) for exosomal RNA sequencing using STAR version 2.7.9a. Gene‐level counts were generated with featureCounts version 2.0.1 and used for downstream differential expression analysis.
Differential expression analysis between treatment groups was performed in R version 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria) using the DESeq2 package.23 Genes or miRNAs with low expression were filtered out based on a counts‐per‐million threshold of ≤1 in <2 samples. Significant features were identified using an adjusted P value (false discovery rate [FDR] ≤0.05) as the threshold for significance. Fold changes were calculated as log2‐transformed values to assess upregulated and downregulated features between treatment groups. For porcine peri‐infarct tissue, pathway enrichment analysis of differentially expressed genes (DEGs) was conducted using ShinyGO version 0.81.24 Gene Ontology Biological Processes (GO:BP) terms were identified to explore functional pathways associated with DEGs. Default settings were applied, including an adjusted P value threshold (FDR ≤0.05) for significance. For the miRNA analysis comparing iCM‐Exo and MSC‐Exo, pathway enrichment was performed using DIANA‐miRPath version 4.025 with the microT‐CDS algorithm for target prediction. Both “Genes Union” and “Pathways Union” methods were used to identify biological processes and pathways enriched by the miRNA targets. Enrichment significance was determined using Fisher's exact test with FDR correction (q≤0.05). To improve reliability, the coefficient of variation was calculated for each miRNA across biological replicates within each group. miRNAs with a coefficient of variation >20% were excluded from the analysis, retaining only those with consistent expression across samples for downstream analyses. Heatmaps of differentially expressed genes and miRNAs were generated using the online tool Morpheus, while hierarchical clustering dendrograms and bar plots were created using the ggplot2 package in R.
Data are expressed as the median with interquartile range (IQR). The Kruskal–Wallis test was used to assess differences among ≥3 groups, and Dunn's multiple comparisons test was then used to account for small sample sizes. Analyses were performed using Prism version 10.1.2 for Windows (GraphPad Software, Boston, Massachusetts; www.graphpad.com). The assumptions for the Kruskal–Wallis test, including homoscedasticity, were evaluated using residual plots, homoscedasticity plots, and QQ plots. A P value <0.05 was considered statistically significant.
Results
Porcine hearts were treated with iCM‐Exo (n=6), MSC‐Exo (n=5), or PBS (n=4). Cardiac MRI at weeks 2 and 4 after injury measured the following: LVEF (percentage), LV end‐systolic volume (mL), LV end‐diastolic volume (mL), scar size (percentage) via delayed gadolinium enhancement MRI (DEMRI), and myocardial viability (percentage) via manganese‐enhanced MRI (MEMRI).
LVEF decreased significantly less at week 2 by iCM‐Exo group compared with the PBS group (median: PBS, 21.74% [IQR, 18.67%–23.82%] versus iCM‐Exo, 32.18% [IQR, 28.11%–34.97%]; P<0.01; Figure 1A). A significant difference persisted at week 4 (median: PBS, 19.89% [IQR, 14.19%–21.91%] versus iCM‐Exo, 33.30% [IQR, 30.33%–36.00%]; P<0.01; Figure 1A). Furthermore, LVEF delta (the difference in LVEF between weeks 2 and 4) was significantly higher in the iCM‐Exo group than in the PBS group (median: PBS, −3.38% [IQR, −5.08% to 0.21%] versus iCM‐Exo, 1.27% [IQR, 0.53%–1.93%]; P<0.05; Figure 1A). Significant differences in LV end‐diastolic volume (median: PBS, 164.1 mL [IQR, 141.3–206.9 mL] versus iCM‐Exo, 110.4 mL [IQR, 108.2–121.3 mL]; P<0.05; Figure 1B) and in LV end‐systolic volume (median: PBS, 127.4 mL [IQR, 109.8–167.0 mL] versus iCM‐Exo, 75.21 mL [IQR, 72.71–83.34 mL]; P<0.05; Figure 1C) were observed only at week 2 in the iCM‐Exo group. iCM‐Exo treatment showed significantly reduced scar size compared with PBS only at week 4 (median: PBS, 34.94% [IQR, 33.83%–37.38%] versus iCM‐Exo, 26.55% [IQR, 20.20%–32.31%]; P<0.05; Figure 1D). These results demonstrate that iCM‐Exo treatment mitigated cardiomyocyte injury and stabilized cardiac function. MSC‐Exo treatment did not show the same mitigation of myocardial injury but demonstrated a trend toward benefit.
A, Left ventricular ejection fraction, (B) left ventricular end diastolic volume, (C) left ventricular end systolic volume, and (D) infarct size as an area percentage of delayed gadolinium enhancement MRI‐positive signal with a representative of infarct size. Statistical significance was determined by Kruskal–Wallis (iCM‐Exo, n=6; MSC‐Exo, n=5; PBS n=4) with multiple comparisons. Median with interquartile range is shown. Only significance between control and treatment groups at weeks 2 and 4 are shown (P<0.05, P*<0.01). iCM‐Exo indicates exosomes derived from induced pluripotent stem cell–derived cardiomyocytes; LVEDV, left ventricular end‐diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end‐systolic volume; MRI, magnetic resonance imaging; and MSC‐Exo, exosomes derived from mesenchymal stem cells.
At weeks 2 and 4, Yorkshire pigs were given gadolinium and manganese contrast agents to assess scar size and viable tissue, respectively. Cardiac viability is determined by the difference in T1 values observed before and after maganese uptake.21 Peri‐infarct region is dilineated as the area that appears as scar tissue on DEMRI but shows maganese uptake on MEMRI, indicating presence of surviving cardiomyocytes among damaged tissue. The resultant area of this difference (measured as percentage PIR Viability) is correlated with LV function. Manganese‐enhanced MRI showed that iCM‐Exo treatment significantly increased myocardial viability at week 4 (median: PBS, 79.21% [IQR, 77.82%–81.14%] versus iCM‐Exo, 92.48% [IQR, 87.72%–95.46%]; P<0.05; Figure 2A). MSC‐Exo treatment did not yield significant improvement in viability at either time point. Injections of iCM‐Exo into the peri‐infarct region indicated significant rescue of the injured but viable cells, improving myocardial function and morphology. Both treatment groups showed trends toward improved peri‐infarct viability compared with PBS (week 2: PBS versus iCM‐Exo, P>0.99; PBS versus MSC‐Exo, P>0.99; week 4: PBS versus iCM‐Exo, P=0.32; PBS versus MSC‐Exo, P>0.99; delta: PBS versus iCM‐Exo, P=0.92; PBS versus MSC‐Exo, P=0.39; Figure 2B).
A, Manganese‐enhanced MRI percentage viability and (B) peri‐infarct region percentage viability. Statistical significance was determined by Kruskal–Wallis (iCM‐Exo, n=5; MSC‐Exo, n=4; PBS n=3) with multiple comparisons. Median with interquartile range is shown. Only significance between control and treatment groups within each week are shown (*P<0.05). iCM‐Exo indicates exosomes derived from induced pluripotent stem cell‐derived cardiomyocytes; MEMRI, manganese‐enhanced MRI; MRI, magnetic resonance imaging; MSC‐Exo, exosomes derived from mesenchymal stem cells; and PIR, peri‐infarct region.
Furthermore, DEMRI showed that iCM‐Exo–treated pigs had significantly less scar burden at week 4. This finding was consistent with the detection of significantly enhanced myocardial viability in the iCM‐Exo treatment groups, demonstrating sustained restoration of the peri‐infarct region, cardiac function, volumetric remodeling, and morphology in the porcine chronic HF model at week 4 (Figure 1B and 1C).
Differential gene expression analysis of porcine peri‐infarct tissue provided insights into the mechanisms by which iCM‐Exo and MSC‐Exo treatments contribute to cardiac repair. The principal component analysis reveals clustering of samples within each group, indicating consistent gene expression profiles among replicates. While there is some divergence, the overall clustering suggests good reproducibility and biological distinction between the groups (Figure S1A). The correlation matrix revealed the degree of relatedness between RNA sequencing samples from iCM‐Exo and MSC‐Exo groups on the basis of their normalized gene expression profiles. The high correlation values observed within the iCM subsets suggest these samples are biologically distinct. In contrast, MSC and PBS appeared to be less distinct from one another (Figure S1B).
Significant DEGs were identified in porcine tissue of both treatment groups, with 457 DEGs detected in the MSC versus PBS comparison and 8683 DEGs in the iCM versus PBS comparison, without applying a fold‐change threshold. In the MSC‐Exo treatment group, 185 genes were upregulated while 272 genes were downregulated. In contrast, the iCM‐Exo treatment induced a broader transcriptional response, with 4453 upregulated and 4230 downregulated genes (Figure 3A). These results suggest a more extensive transcriptional impact of iCM‐Exo compared with MSC‐Exo, potentially reflecting differences in the miRNA and protein cargo of their respective exosomes. This is further supported by the dendrogram of sample relatedness and the correlation matrix, which demonstrated greater biological distinction of iCM‐Exo‐treated samples compared with MSC‐Exo and PBS groups (Figure S1C). Of note, PBS and MSC‐Exo groups exhibited closer similarity to one another than to iCM‐Exo–treated samples. The dendrogram provided a hierarchical clustering of the samples, offering a visual representation of their relatedness. Samples that clustered tightly together exhibited closer relationships, and the clustering pattern aligned with the correlation matrix, confirming the grouping of related samples. iCM samples formed distinct branches whereas PBS and MSC samples were more closely related. This may explain why the MSC versus PBS comparison yielded fewer DEGs.
(A) Venn diagrams of differentially expressed genes between iCM‐Exo, MSC‐Exo, and PBS treatments. These results highlight distinct transcriptional responses of iCM‐Exo and MSC‐Exo while identifying shared molecular signatures relevant to cardiac repair mechanisms. Dot plot of top enriched GO:BP enrichment analysis of DEGs associated with (B) upregulated and (C) downregulated genes in the iCM‐Exo treatment group and top enriched GO:BP terms associated with (D) upregulated and (E) downregulated genes in the MSC‐Exo treatment groups. The x axis represents fold enrichment, while the y axis lists the enriched biological processes. Dot size indicates the number of genes involved in each process, and dot color represents the statistical significance of enrichment (−log10[FDR]). FDR indicates false discovery rate; GO:BP, Gene Ontology Biological Process; iCM, induced pluripotent stem cell–derived cardiomyocyte; and MSC, mesenchymal stem cell.
To further dissect the effects of iCM‐Exo and MSC‐Exo treatments, DEGs were compared between the two exosome groups. Among the DEGs identified, 397 were shared between iCM‐Exo and MSC‐Exo treatments, representing 86.9% of the DEGs in the MSC‐Exo group.
Interestingly, five genes exhibited opposite regulation between the treatments. For example, FBN2 and ENOX1 were upregulated in the MSC‐Exo group but downregulated in the iCM‐Exo group, whereas ANGPTL7 and RAPGEF5 displayed the reverse pattern. These findings may point to distinct biological priorities for each exosome type and mechanistic differences in their therapeutic effects.
GO:BP analysis revealed both shared and unique pathways modulated by iCM‐Exo and MSC‐Exo treatments, providing further insight into their mechanisms of action (Figure 3B through 3E). Shared pathways included upregulated processes such as collagen fibril organization, generation of neurons, and regulation of cell adhesion, suggesting the importance of extracellular matrix remodeling, neurovascular integration, and enhanced cellular connectivity in cardiac repair. Similarly, shared downregulated pathways included myofibril assembly, muscle system process, and protein metabolic process, reflecting a mutual suppression of maladaptive hypertrophy and preservation of myocardial architecture. Distinct pathways were also identified for each treatment. iCM‐Exo uniquely upregulated pathways such as positive regulation of organelle organization, apoptotic processes, and cellular response to lipids, suggesting a focus on metabolic reorganization and structural optimization. In contrast, unique downregulated pathways for iCM‐Exo included mitochondrial RNA processing and sarcoplasmic reticulum calcium ion transport, which may mitigate oxidative stress and calcium overload in failing cardiomyocytes. MSC‐Exo demonstrated unique upregulation of pathways related to fibroblast proliferation, extracellular structure organization, and response to reactive oxygen species, consistent with their role in anti‐inflammatory signaling and matrix stabilization. Unique downregulated pathways included actin–myosin filament sliding, cardiac muscle hypertrophy, and lateral ventricle development, suggesting suppression of maladaptive remodeling under chronic stress conditions. The convergence of shared pathways demonstrates a therapeutic strategy centered on extracellular matrix remodeling, neurovascular support, and attenuation of hypertrophy. However, the divergence in unique pathways highlights distinct molecular priorities: iCM‐Exo focuses on cardiomyocyte‐specific processes such as metabolic optimization and structural reorganization while MSC‐Exo emphasizes anti‐inflammatory signaling and matrix stabilization. These complementary mechanisms suggest the potential for synergistic effects if combined. Further investigation is warranted to explore the opposing regulation of specific genes and their implications for therapeutic efficacy.
A direct comparison of the transcriptional profiles of iCM‐Exo and MSC‐Exo treatments revealed distinct regulatory priorities (Figure 4A). Upregulated pathways in iCM‐Exo treatment (and thus downregulated in MSC‐Exo treatment) were predominantly related to immune system processes, including adaptive immune response, leukocyte activation, and regulation of immune response, suggesting immune activation and tissue repair. Conversely, downregulated pathways in iCM‐Exo treatment (and thus upregulated in MSC‐Exo treatment) were linked to mitochondrial bioenergetics, such as oxidative phosphorylation, aerobic respiration, and the electron transport chain, reflecting a metabolic shift that reduces oxidative stress and supports cell survival in ischemic tissue. The Kyoto Encyclopedia of Genes and Genomes hierarchical clustering tree and network diagram (Figure 4B) highlight the distinct biological priorities of each treatment with iCM‐Exo focusing on immune and metabolic regulation and MSC‐Exo emphasizing structural stabilization and anti‐inflammatory effects. Together, these results reveal distinct mechanisms of action, offering a multifaceted approach to myocardial repair.
A, Unique and overlapping pathways between iCM‐Exo and MSC‐Exo treatments by regulation type. Bar graph depicting the number of pathways uniquely enriched or shared between iCM‐Exo and MSC‐Exo treatments, categorized as upregulated (blue) or downregulated (red). Overlapping pathways represent shared biological processes modulated by both exosome types while iCM Unique and MSC Unique pathways reflect distinct mechanisms specific to each treatment. The analysis reveals a larger number of pathways uniquely associated with iCM‐Exo treatment, particularly in upregulated pathways, highlighting its broader transcriptional impact compared with MSC‐Exo. B, Hierarchical clustering of top 10 upregulated and downregulated KEGG Pathways in iCM vs MSC treatments. This hierarchical clustering tree visualizes the top 10 KEGG pathways enriched in porcine tissue treated with iCM‐Exo vs MSC‐Exo. Upregulated pathways are indicated by red dots, and downregulated pathways by blue dots. The tree shows the relative significance and gene overlap among the pathways with smaller P values indicating more significant enrichments. KEGG pathway network analysis of iCM‐Exo vs MSC‐Exo treatments in porcine tissue. This network diagram illustrates the KEGG pathways enriched in porcine tissue treated with iCM‐Exo vs MSC‐Exo. Pathway nodes are connected if they share ≥30% genes. Green nodes represent upregulated pathways, while red nodes represent downregulated pathways. The color intensity of the nodes indicates the significance of enrichment, with darker nodes corresponding to more significantly enriched gene sets. The size of each node reflects the number of genes within the pathway, with larger nodes representing larger gene sets. The thickness of the edges between nodes represents the extent of gene overlap between pathways, with thicker edges indicating greater overlap. iCM‐Exo indicates exosomes derived from induced pluripotent stem cell–derived cardiomyocytes; KEGG, Kyoto Encyclopedia of Genes and Genomes; and MSC‐Exo, exosomes derived from mesenchymal stem cells.
To provide a global perspective, hierarchical clustering and heatmap analysis were performed on the top 2000 genes ranked by SD across all samples (Figure 5A). This unsupervised approach, which does not directly group samples, revealed that iCM‐Exo–treated samples are transcriptionally distinct from MSC‐Exo– and PBS‐treated groups, which were more similar to each other. This observation is consistent with the correlation matrix and principal component analysis, further showing the broader transcriptional impact of iCM‐Exo compared with MSC‐Exo and PBS. This hierarchical clustering analysis supports the findings of prior analyses in that the iCM‐Exo treatment uniquely upregulated GO:BP terms related to immune response. Conversely, iCM‐Exo downregulated pathways related to heart contraction and mitochondrial bioenergetics, indicating metabolic adaptations that reduce oxidative stress and promote cardiomyocyte survival.
(A) GO and biological processes are shown by the heatmap displaying hierarchical clustering of the top 2000 genes ranked by standard deviation across all samples and highlighting the transcriptional differences between iCM‐Exo, MSC‐Exo, and PBS treatments. Gene expression is centered by subtracting the average expression for each gene with colors representing upregulated (red) and downregulated (blue) genes. Hierarchical clustering was performed using Pearson's correlation coefficient (1−r) as the distance metric and average linkage as the clustering method. K‐means clustering (k=2) was applied, as determined by the elbow plot. Adjacent to the heatmap, hierarchical trees illustrate enriched GO:BPs for upregulated and downregulated genes in each cluster. (B) Heatmap of differentially expressed miRNAs between iCM‐Exo and MSC‐Exo treatments. Heatmap illustrating the expression patterns of miRNAs that were significantly differentially expressed between iCM‐Exo and MSC‐Exo groups. Rows represent individual miRNAs, and columns represent biological replicates of each exosome type. Expression levels are color‐coded, with red indicating upregulation, blue indicating downregulation, and white representing intermediate expression levels. To ensure data robustness, miRNAs with a coefficient of variation >20% between replicates were excluded from the analysis. Hierarchical clustering reveals distinct miRNA signatures for each exosome type, reflecting their unique molecular profiles and potential functional roles in cardiac repair. GO indicates Gene Ontology; GO:BP, Gene Ontology Biological Processes; iCM, induced pluripotent stem cell–derived cardiomyocyte; iCM‐Exo, exosomes derived from induced pluripotent stem cell–derived cardiomyocytes; miRNA, micro RNA; MSC, mesenchymal stem cell; and MSC‐Exo, exosomes derived from mesenchymal stem cells.
To complement the transcriptomic analysis of the porcine peri‐infarct tissue, exosomal next‐generation sequencing was performed to investigate the molecular contents of MSC‐Exo and iCM‐Exo. Distinct groups of significantly expressed miRNAs were identified between the two exosome types, reflecting their unique molecular signatures (Figure 5B). Functional pathway analysis of these differentially expressed miRNAs was conducted using DIANA‐miRPath version 4.0,25 highlighting the distinct functional roles of MSC‐Exo and iCM‐Exo in modulating cellular processes relevant to cardiac repair. For miRNAs upregulated in iCM‐Exo (and thus downregulated in MSC‐Exo), the top 20 enriched pathways involved transcriptional regulation, intracellular signal transduction, protein phosphorylation, and cell differentiation (Figure 6A). Specific pathways directly related to cardiac biology, such as heart development and angiogenesis, were also enriched, suggesting that iCM‐Exo promotes cardiomyocyte regeneration and vascular integration. In contrast, miRNAs upregulated in MSC‐Exo (and thus downregulated in iCM‐Exo) were enriched in pathways related to transcriptional control, protein phosphorylation, and atrial septum development (Figure 6B). Processes such as Wnt signaling, embryonic development, and neuroinflammatory responses were also prominent, suggesting an additional reparative role for MSC‐Exo that includes extracellular matrix remodeling and anti‐inflammatory effects. Moreover, enriched pathways such as pre‐miRNA processing and protein glycosylation indicated potential regulatory roles in intercellular communication and stabilization of tissue architecture. Both data sets showed enriched pathways related to nervous system development and axon guidance. These findings may point to the shared developmental origins and signaling pathways between the nervous system and cardiac tissue. In the context of cardiac repair, these pathways may enhance vascular integration, extracellular matrix remodeling, and neurovascular coupling, ultimately promoting functional recovery in the injured myocardium. These miRNAs may direct future studies into whether individual or combination of miRNAs can be effective therapies by targeting specific cardiac restorative pathways. Overall, the analyses revealed that iCM‐Exo and MSC‐Exo miRNAs regulate distinct but complementary processes involved in cardiac repair. iCM‐Exo primarily upregulated pathways associated with regeneration and vascularization, while MSC‐Exo focused on structural stabilization and anti‐inflammatory mechanisms. Both exosome types shared enriched pathways critical for heart development and repair, suggesting complementary roles in balancing structural remodeling and functional recovery in the injured myocardium. It could be argued that iCM‐Exo is better for promoting regeneration and vascular remodeling while MSC‐Exo provides anti‐inflammatory and structural support.
(A) iCM‐Exo upregulation of miRNAs is shown in the bar plot depicting the top 20 enriched GO:BP terms derived from the Genes Union method. The x axis represents the −log10(FDR) of each pathway, indicating the significance of enrichment, while the y axis lists the enriched terms. The number of target genes contributing to each pathway is annotated beside each bar. Heatmap displaying enriched pathways identified via the Pathways Union method. Each cell represents the significance (−log10[FDR]) of a specific miRNA‐term pair, allowing visualization of how individual miRNAs contribute to enriched pathways. The hierarchical clustering highlights shared and distinct pathway involvement among miRNAs upregulated in iCM‐Exo. (B) iCM‐Exo downregulation of miRNAs is shown by the bar plot depicting the top 20 enriched Gene Ontology Biological Process (GO:BP) terms derived from the Genes Union method. The x axis represents the −log10(FDR) of each pathway, indicating the significance of enrichment, while the y axis lists the enriched terms. The number of target genes contributing to each pathway is annotated beside each bar. Heatmap displaying enriched pathways identified via the Pathways Union method. Each cell represents the significance (−log10[FDR]) of a specific miRNA‐term pair, allowing visualization of how individual miRNAs contribute to enriched pathways. The hierarchical clustering highlights shared and distinct pathway involvement among miRNAs upregulated in iCM‐Exo. FDR, false discovery rate; GO:BP, Gene Ontology Biological Processes; iCM‐Exo, exosomes derived from induced pluripotent stem cell‐derived cardiomyocytes; miRNA, micro RNA.
Discussion
In this study, we demonstrated significant preservation of cardiac function, viability, and mitigation of remodeling in injured porcine myocardium treated with iCM‐Exo and a trend toward improvement with MSC‐Exo. iCM‐Exo treated hearts had significant increase in viable myocardium at week 4 while iCM‐Exo significantly decreased the scar size at week 4. Exosome treatments improved the heart function and morphology by increasing the viable myocardium in peri‐infarct region, resulting in sustained recovery over the 4‐week period. Both treatments share pathways contributing to shared degrees of improvement while distinct pathways drive differences in functional outcomes. The PBS treated control hearts deteriorated significantly over time when compared with the exosome‐treated hearts. The marked improvements seen in cardiac function are conferred by miRNA carried by the exosomes, with shared and distinct mechanisms of actions.
iCM‐Exo and MSC‐Exo treatments have shared and distinct mechanisms of actions where iCM‐Exo appear to have more extensive transcriptional impact. PBS and MSC‐Exo groups show closer similarity to one another than iCM‐Exo‐treated samples. Shared pathways in extracellular matrix remodeling, suppression of hypertrophy, and more suggest contributions in baseline improvement seen in both treatments while unique pathways explain differences in cardiac improvement with iCM‐Exo treatments and MSC‐Exo treatments. Our transcriptional analysis can begin to provide insight into how genes are impacting cardiac improvement and what exosomal cargo is contributing to which pathways. Genes such as FBN2,26, 27 ENOX1,28, 29, 30 ANGPTL7,31, 32 and RAPGEF5 33, 34 may contribute to observed transcriptional changes while identifying possible future targets of interest. The role of each gene in cardiac improvement requires further study.
Given the distinct miRNA profiles of iCM‐Exo and MSC‐Exo, understanding how their cargo drives observed cardiac outcomes requires further study and remains of interest for translational applications, especially when miRNA have multiple targets. Several miRNA of interest include both down‐ and upregulation in the let‐7 family in iCM‐Exo and MSC‐Exo, respectively, which can target cellular respiration with putative targets mapping to the mitochondrial genome.35, 36 hsa‐miR‐1‐3p is also upregulated in iCM‐Exo and downregulated in MSC‐Exo with roles in targeting cellular proliferation, regulating cardiac and skeletal muscle, and targeting the Notch pathway,37, 38 and is activated during mitochondrial biogenesis and enhances translation of bioenergetic proteins.35, 39
The coronary anatomy of a pig most closely mimics the human anatomy due to the lack of collateral circulation, generating a myocardial injury model most representative of clinically relevant ischemia reperfusion pathology. This injury model allows the most meaningful evaluation of delivery into the peri‐infarct region to assess LV arrhythmia inducibility and LV remodeling. Thus, advanced cardiac MRI to monitor therapeutic efficacy of novel biologics in a porcine myocardial injury model is clinically relevant and necessary for clinical translation. This study will provide the necessary foundation to move forward with a clinical trial of biologics for peri‐infarct injury. Furthermore, robust electrophysiological studies and collection of myocardial tissue and blood samples are feasible in a Yorkshire swine model. Altogether, exosomes from both MSCs and iCMs were shown to improve cardiac function in an acute porcine myocardial injury model, and these biologics could be translated to treat patients with HF. The findings highlight the potential for exosome‐based therapies in HF and the need for further assessment of combinatorial approaches to leverage complementary mechanisms.
Limitations in this study include but are not limited to a small sample size in a xenogeneic model in which there may be unknown immunogenic effects from the human products applied in a porcine model. Furthermore, pig myocardium was collected 4 weeks after injection. To elucidate the immediate response in the peri‐infarct myocardium, tissue samples should be harvested hours to days after injection for analysis. We also do not test direct comparisons with combined treatments.
In an acute porcine myocardial infarction model, we investigated the effect of MSC‐Exo and iCM‐Exo treatments on injured myocardium. Both iCM‐Exo and MSC‐Exo treatment can stabilize function and morphology of injured myocardium. Comparatively, PBS‐treated control hearts deteriorated significantly over time. Shared pathways, including but not limited to extracellular matrix remodeling and suppression of hypertrophy, contribute to baseline improvement seen between iCM‐Exo and MSC‐Exo treatments, while distinct pathways drive differential efficacy. The contribution of specific exosomal cargo to observed cardiac improvement requires further study, and understanding the complementary therapeutic effect of combination treatments would be of interest.
Sources of Funding
This study was funded by the Intervalien Foundation, Mountain View, California. N.V. was supported by a American Heart Association postdoctoral fellowship (23POSTCHF1031213).
Disclosures
The authors have reported that they have no relationships relevant to the contents of this paper to disclose. There are no funding relationships with industry.
Supporting information
Figure S1
CO
Dr. Connor O'Brien
Cardiology · Cardiac Critical CareUniversity of California San Francisco Medical Centerconnor
