Inhalable Stem Cell Exosomes Promote Heart Repair After Myocardial Infarction

Abstract

Background: Exosome therapy shows potential for cardiac repair after injury. However, intrinsic challenges such as short half-life and lack of clear targets hinder the clinical feasibility. Here, we report a noninvasive and repeatable method for exosome delivery through inhalation after myocardial infarction (MI), which we called stem cell-derived exosome nebulization therapy (SCENT).

Methods: Stem cell-derived exosomes were characterized for size distribution and surface markers. C57BL/6 mice with MI model received exosome inhalation treatment through a nebulizer for 7 consecutive days. Echocardiographies were performed to monitor cardiac function after SCENT, and histological analysis helped with the investigation of myocardial repair. Single-cell RNA sequencing of the whole heart was performed to explore the mechanism of action by SCENT. Last, the feasibility, efficacy, and general safety of SCENT were demonstrated in a swine model of MI, facilitated by 3-dimensional cardiac magnetic resonance imaging.

Results: Recruitment of exosomes to the ischemic heart after SCENT was detected by ex vivo IVIS imaging and fluorescence microscopy. In a mouse model of MI, SCENT ameliorated cardiac repair by improving left ventricular function, reducing fibrotic tissue, and promoting cardiomyocyte proliferation. Mechanistic studies using single-cell RNA sequencing of mouse heart after SCENT revealed a downregulation of Cd36 in endothelial cells (ECs). In an EC-Cd36^fl/- conditional knockout mouse model, the inhibition of CD36, a fatty acid transporter in ECs, led to a compensatory increase in glucose utilization in the heart and higher ATP generation, which enhanced cardiac contractility. In pigs, cardiac magnetic resonance imaging showed an enhanced ejection fraction (Δ=11.66±5.12%) and fractional shortening (Δ=5.72±2.29%) at day 28 after MI by SCENT treatment compared with controls, along with reduced infarct size and thickened ventricular wall.

Conclusions: In both rodent and swine models, our data proved the feasibility, efficacy, and general safety of SCENT treatment against acute MI injury, laying the groundwork for clinical investigation. Moreover, the EC-Cd36^fl/- mouse model provides the first in vivo evidence showing that conditional EC-CD36 knockout can ameliorate cardiac injury. Our study introduces a noninvasive treatment option for heart disease and identifies new potential therapeutic targets.

Keywords: CD36; exosomes; extracellular vesicles; inhalation; myocardial infarction.

Figure 1.. Feasibility and therapeutic efficacy of SCENT in mouse with MI injury.

A. Schematic of the concept of SCENT in treating MI. A representative TEM image showing the morphology of LSC-Exo before inhalation. Created with BioRender.com. B. Size analysis of fresh exosome particles by ZetaView. C. Representative dSTORM image of a single exosome, fluorescently labelled with CD81, CD63, CD9 antibodies. Scale bar, 100 nm. D. Representative ex vivo IVIS imaging of the hearts and lungs from mice at different time points. n = 3 animals per group. E. Quantification of fluorescence signals in (D) after exosome inhalation. Data are mean ± SD. F. Schematic showing the mouse study design. Created with BioRender.com. PARI Trek^® S is a trademark of PARI GmbH and its affiliates. G. Summary of survival rates of the animals in MI+SCENT group, MI+PBS group and MI group. MI+SCENT group: MI mice with SCENT treatment (LSC-Exo inhalation); MI+PBS group: MI mice with PBS inhalation; MI group: MI mice without any treatment. H. Representative M-mode images of echocardiography measurement of cardiac function. I. Quantitative analysis of LV-EF and LV-FS. Data were expressed as mean ± SD. n = 8 animals for each group. Statistical analysis was performed by two-way mixed-effects ANOVA, followed by Sidak’s test. J. Masson trichrome staining was performed to show the histological improvement of the infarct area after SCENT. Scale bar, 500 μm. K-L. Quantitative analysis of infarct size (K) and wall thickness (L). Data are mean ± SD. n = 8 animals for each group. Statistical analysis was performed by one-way ANOVA, followed by Tukey’s test. M. Representative immunofluorescent images of Ki67 to show the proliferation of cardiomyocytes after SCENT and quantitative analysis. Data are expressed as mean ± SD. n = 8 animals for each group. Statistical analysis was performed by two tailed unpaired t test. Scale bar, 50 μm. N. Quantitative analysis of apoptotic cells by TUNEL assay. Data are expressed as mean ± SD. n = 8 animals for each group. Statistical analysis was performed by two tailed unpaired t test.

Figure 2.. Single-cell transcriptome profiling of whole mouse hearts in SCENT and control.

A. Schematic showing the process of sample collection and droplet-based single-cell RNA-sequencing by 10x Genomics. Created with BioRender.com. B-C. Uniform manifold approximation and projection (UMAP) embedding of 19, 568 single cells delineate 2 groups (B) and 8 cell types (C). D. Dot plot of markers representing the gene expression signature for individual cell type. E. Feature plot of marker genes. F. UMAP embedding of 4 endothelial populations. Left, colored by groups; Right, colored by seurat cluster. G. Volcano plot showing differentially expressed genes of endothelial cells between the Control and SCENT groups. H. Dot plot of representative differentially expressed genes of endothelial cells between the Control and SCENT groups.

Figure 3.. CD36 low expression in ECs ameliorates the MI injury.

A. Violin plots of expression of CD36 in endothelial cells in Control and SCENT groups. B. Feature plot of CD36 expression in endothelial cells. C. Violin plots of expression of CD36 expression in endothelial cells in different seurat clusters. D. Distribution of endothelial cell populations after subclustering analysis. E. Violin plots showing expression of related genes in the CD36 pathway (Fabp4, Lpl, Hspg2, Gpihbp1) in endothelial cells in Control and SCENT groups. F. Representative immunofluorescent images of CD36 to show the protein expression level and quantitative analysis. Scale bar, 50 μm. Data are expressed as mean ± SD. n = 6 animals for each group. Statistical analysis was performed by one-way ANOVA, followed by Tukey’s test. G. Schematic showing the breeding strategy of Cd36^EC-hKO mouse model by Cre-loxP system. Created with BioRender.com. H. Representative immunofluorescent images of CD36 to show the protein expression level and quantitative analysis. Scale bar, 50 μm. Data are expressed as mean ± SD. n = 3 animals for each group. Statistical analysis was performed by two tailed unpaired t test. I. Quantitative analysis of LV-EF and LV-FS. Data are expressed as mean ± SD. n = 6 mice for each group. Statistical analysis was performed by two-way mixed-effects ANOVA, followed by Sidak’s test. J. Representative immunofluorescent images of Ki67 to show the proliferation of cardiomyocytes after SCENT and quantitative analysis. Scale bar, 50 μm. Data are expressed as mean ± SD. n = 6 animals for each group. Statistical analysis was performed by two tailed unpaired t test. K. Quantitative analysis of glucose concentration in mouse whole hearts from Cd36^EC-hKO or WT group. Data are expressed as mean ± SD. n = 3 animals for each group. Statistical analysis was performed by two tailed unpaired t test. L. Consumption of glucose into live mouse heart in 60 mins. Data are expressed as mean ± SD. n = 4 animals for each group. Statistical analysis was performed by two tailed unpaired t test. M. Uptake rate of glucose into live mouse heart was measured. Data are expressed as mean ± SD. n = 3 animals for each group. Statistical analysis was performed by two tailed unpaired t test. N. Quantitative analysis of ATP concentration in mouse heart. Data are expressed as mean ± SD. n = 3 animals for each group. Statistical analysis was performed by two tailed unpaired t test. O. Schematic illustration of proposed mechanism of CD36 deficiency in cardioprotection. Created with BioRender.com.

Figure 4.. SCENT provides cardioprotection in swine with ischemic heart injury.

A. Schematic showing the swine study design. Created with BioRender.com. B. Angiography of pigs during MI surgery through balloon occlusion procedure. Arrow points the balloon placement. C. Representative electrocardiography during MI surgery, at Day 3 post-MI and Day 28 post-MI. Arrows point the ST segment elevation. D. Representative MRI images of pig hearts at base, middle and apex location, at Pre-MI, Day 3 post-MI, Day 28 post-MI. Arrows point the MI infarct areas. E. Quantitative analysis of total MI area. Data are expressed as mean ± SD. Statistical analysis was performed by two-way mixed-effects ANOVA, followed by Sidak’s test. F. Quantitative analysis of wall thickness based on MRI images. Statistical analysis was performed by two-way mixed-effects ANOVA, followed by Sidak’s test. G. Quantitative analysis of LV-EF, LV-FS based on MRI. Data are mean ± SD. Statistical analysis was performed by two-way mixed-effects ANOVA, followed by Sidak’s test. H. Representative immunofluorescent images of pHH3 to show the proliferation of cardiomyocytes after SCENT and quantitative analysis. Scale bar, 50 μm. Data are expressed as mean ± SD. Statistical analysis was performed by two tailed unpaired t test. I. Representative immunofluorescent images of vWF to show neovascularization after SCENT and quantitative analysis. Scale bar, 100 μm. Data are expressed as mean ± SD. Statistical analysis was performed by two tailed unpaired t test.