Cardiomyocyte-targeting exosomes from sulforaphane-treated fibroblasts affords cardioprotection in infarcted rats

AngII (Cat. A9525), dimethyl sulfoxide (DMSO) (Cat. D8418) and L-SFN (PubChem CID: 5350) (Cat. S6317) were purchased from Sigma-Aldrich Chemical Co, Merk Life Science S.r.l., Milan, Italy. In accordance with the manufacturer’s instructions, SFN stock solution (48 mM) was prepared in DMSO, aliquoted and stored at − 20 °C. AngII stock solution (1 mM) was prepared in sterile distilled water, aliquoted and kept at − 20 °C. Repeated freeze–thaw cycles were avoided, and stock solutions were only stored for 6 weeks. The stock solutions were further diluted with the appropriate complete culture medium.

Mouse NIH/3T3 embryo fibroblasts (ATCC® CRL-1658™ kindly provided by Dr. Giuseppe Rainaldi, IFC-CNR, Pisa, Italy) are frequently used as in vitro models of adult cardiac fibroblasts [33, 34] and fibroblast-cardiomyocytes exosome-based crosstalk [35]. HL-1 mouse atrial cardiomyocytes (a kind gift of Professor W. Claycomb, LSUHSC, New Orleans, LA, USA), a cell line used to study cardiac physiology at the cellular level [36] and to assess the cardioprotective effects of exosomes [18, 19, 21, 37, 38], were cultured in Claycomb Medium (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy) [39] supplemented with 10% exosomes-depleted FBS, 100units/mL of penicillin, 100 µg/mL streptomycin, 250 ng/mL amphotericin B and 2 mM L-glutamine (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy). Murine immortalized cardiac endothelial cells (MCEC; Tebu-bio Srl, Magenta, Italy) were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy) containing 1 g/L glucose and supplemented with 10% (V/V) fetal bovine serum (FBS) (System Biosciences, Palo Alto, CA, USA), 100units/mL of penicillin, 100 µg/mL streptomycin, 250 ng/mL amphotericin B and 2 mM L-glutamine.

For 7 days (d), fibroblasts were treated with increasing doses of SFN (1.5—48uM) or 0.01% DMSO in complete medium in order to produce SFN-F-EXOs. To mimic an ischemic microenvironment, HL-1 [40] and MCEC [41] cells were treated with 100 nM AngII (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy) for at least 24 h. The treatment was carried out in exosome-free media, which consists of the appropriate growth medium supplemented with 10%(v/v) exosome-depleted FBS (System Biosciences, Palo Alto, CA, USA) whenever exosomes were isolated from cell-conditioned medium.

According to the manufacturer, the MTT test was used to assess the relative number of viable cells after each treatment. In brief, 1,000 fibroblasts/well were plated on a 96-well plate and allowed to adhere overnight before being treated with increasing dosages of SFN or DMSO (3 or 7d) the next day. After treatment, the culture media was removed, cells were washed with phosphate buffered saline (PBS) before being treated for 4 h at 37 °C with 500 μg/mL 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT, Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy). MTT was removed, and purple formazan crystals were dissolved in acidified isopropanol (100% isopropanol containing 0.31% (v/v) HCl) under steady agitation for 10 min (min) at room temperature (RT). A conventional plate-reader spectrophotometer was used to measure absorbance at 540 nm (signal) and 660 nm (background).

Exosomes were isolated from normal and SFN-treated NIH/3T3 conditioned medium using a differential centrifugation protocol previously described by us [18, 19, 21] and in compliance with the most recent Minimal Information for Studies of Extracellular Vesicles (“MISEV”) guideline [42].

As previously described by us [18, 19, 21, 37, 38], exosome pools were investigated using a NanoSight LM10 nanoparticle tracking analysis (NTA) device (NanoSight Ltd, UK) to evaluate particle size and concentration. In a nutshell, the exosome pellet was re-suspended in 100μL of filtered (0.2 μm filter) PBS. Prior to analysis, the exosome suspension was diluted ten times in PBS to match the instrument’s optimal detection range. The daily rate of exosome release was calculated by dividing the number of exosomes isolated from a P-100 plate (quantified by NTA) by the 70,000 initial cells seeded and the 7d-treatment duration.

HL-1 and MCEC were planted in either 8-well chamber slides (for fluorescence-based experiments) or 6-well plates (for protein or RNA extraction) depending on the assay. After 12 h, the culture medium was replaced with exosome-free complete growth medium with or without 100 nM AngII, and the cells were incubated for another 12 h. In accordance with the experimental protocol of each assay, a single dose of F-EXOs or SFN-F-EXOs suspended in similar volume of sterile PBS (100µL) was added directly to each well, and cells were incubated for additional time depending on specific assay, maintaining AngII in the culture medium of stressed cells. As a control, just PBS was used.

HL-1 cardiomyocytes were seeded in an 8-well chamber slide (8000 cells/well), allowed to adhere for 12 h, and then treated with/without 100 nM AngII for 12 h as previously described. Each well received 30 µg of F-EXOs or SFN-F-EXOs, and cells were cultured for another 12 h. As a non-treated control, PBS was added alone. The cell monolayer was rinsed with PBS the next day, fixed (4% paraformaldehyde, PFA, 15 min at room temperature) and permeabilized (0.1% TritonX-100, 2 × 2 min). To stain the filamentous actin of the cytoskeleton, cells were incubated for 1 h at RT with 1:300 dilution of Phalloidin-Atto 550 (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy). PBS was used to remove excess dye before mounting the gass slide with DAPI-containing Vectashield (DBA Italia srl, Segrate, Italy). The images were acquired using a fluorescence microscope (DM2500; Leica Microsystems srl, Milan, Italy) (20 × magnification). ImageJ software was used to manually quantify cell area, which was then expressed as pixels.

Superoxide (O2-) generation was measured by staining with 5 mol/L dihydroethidium (DHE, Invitrogen™, Thermo Fisher Scientific, Waltham, MA, USA) in PBS for 30 min at 37 °C [43]. HL-1 cardiomyocytes were seeded in an 8-well chamber slide (8,000 cells/well), allowed to adhere for 12 h, and then treated with/without 100 nM AngII for 12 h as previously described. Each well received 30 μg of F-EXOs or SFN-F-EXOs, and cells were cultured for another 12 h with/without 100 nM AngII. PBS alone was added as a control. The cell monolayer was washed three times with PBS before being mounted on glass slides were with DAPI-containing Vectashield (DBA Italia srl, Segrate, Italy). A fluorescence microscope (DM2500; Leica Microsystems srl, Milan, Italy) was used to capture the images (20 ×magnification). ImageJ software was used to quantify the fluorescence intensity of DHE.

HL-1 cardiomyocytes were seeded in an 8-well chamber slide (8000 cells/well), allowed to adhere for 12 h, and then treated with/without 100 nM AngII for 12 h as previously described. Each well received 30 µg of F-EXOs or SFN-F-EXOs, and cells were cultured for another 12 h. As a control, PBS was added alone. Acetylated histone H4 (AcH4) was stained with an antibody against acetylated H4 histone (Millipore) (overnight, at 4 °C), followed by Alexa-Fluor488 secondary antibody (Abcam, Cambridge, UK; 1:300). Images were acquired using a fluorescence microscope (DM2500; Leica Microsystems srl) (20 × magnification). Cells were fixed with 4% (w/v) PFA in PBS (10 min at RT) and permeabilized with 0.1% Triton-X100 (5 min at RT), with PBS washed in between steps. The cells were blocked in 1% (w/v) Bovine Serum Albumin (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy) (1 h at RT). Fluorescence intensity was quantified with ImageJ software.

To investigate exosome uptake from cardiac cells, HL-1 and MCEC cells were seeded in an 8-well chamber slide (15,000 cells/well), adhered for 12 h, and then pre-treated with/without 100 nM AngII for 12 h as previously described. The exosomal pellet obtained by ultracentrifugation was first fluorescently labeled by overnight incubation at 4 °C in the dark with constant shaking with a 1:1000 dilution of the lipophilic probe 4-(4-(Dihexadecylamino)styryl)-N-Methylpyridinium Iodide (DiA, Invitrogen™, Thermo Fisher Scientific, Waltham, MA, USA), in accordance with the manufacturer’s guidelines. Ultracentrifugation was used to remove the excess dye. After that, 100 µg of labeled exosomes were added to each chamber well and cells were cultured for another 3 h. After incubation, the media was removed, and the cells were washed with PBS and fixed for 10 min at RT with 4% (w/v) PFA. The cell monolayer was washed three times with PBS before being mounted on glass slides with DAPI containing Vectashield (DBA Italia srl, Segrate, Italy). Cells were observed with a fluorescence microscope (DM2500; Leica Microsystems srl, Milan, Italy) at 20 × magnification. Exosomal uptake was quantified as the ratio between DiA fluorescence to cell surface area using ImageJ software.

To extract cellular proteins, cells were detached with trypsin, pelleted by centrifugation (900 rpm for 5 min), washed once in PBS and re-suspended in 100μL radiommunoprecipitation assay (RIPA) buffer [50 mM Tris, 300 mM NaCl, 5mMEDTA, 1% (v / v) NP40, 0,1% (w/v) SDS, 0,5% (p/v) of sodium deoxycholate] containing protease inhibitors (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy). Cells lysates were incubated on ice for 30 min before being centrifuged at 10,000 g at 4 °C for 10 min to precipitate undigested cell membranes and collecting the supernatant (clear total cell lysate). To extract exosomal proteins, the ultracentrifuged F-EXO pellet was resuspended in 50μL RIPA buffer with protease inhibitors. After 30 min, the samples were sonicated three times on ice for 5 min each to increase digestion efficiency. The BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine protein concentration. SDS-PAGE was used to separate equal amounts of proteins on 10–15% (w/v) acrylamide-bis acrylamide gels under reducing and denaturing conditions. Proteins were transferred to a nitrocellulose membrane (Bio-Rad Laboratories, Inc., Italy) and probed overnight at 4 °C with the appropriate primary antibodies. Primary antibodies towards caspase 3 (Santa Cruz Biotechnology Inc., USA, sc-7148, lot number: J2815, 1:500), acetylated H4 histone (Millipore, Merk Life Science S.r.l., Milan, Italy, 06–866, lot number: 2664262, 1:1000), total H4 histone (Abcam, Cambridge, UK, ab10158, lot number: GR194493-1 and GR114730-1, 1:500), acetylated H3 histone (Acetyl K14) (Abcam, Cambridge, UK, Ab52946, lot number: GR149741-20, 1:1000), total H3 histone (Cell Signaling Technology Inc., Euroclone S.p.A., Pero, Italy, 4499, lot number: Lot-1 and Lot-9, 1:500), maspin (Santa Cruz Biotechnology Inc., USA sc-8543, lot number: D1009, 1:200), nuclear factor erythroid 2–related factor 2 (Nrf2) (Abcam, Cambridge, UK, Ab31163, lot number: GR298097-3, 1:500), heat shock protein 70 (Hsp70) (Enzo Life Sciences, Euroclone S.p.A., Pero, Italy, ADI-SPA-812F, lot number: 8101522, 1:500; Abcam, Cambridge, UK, Ab2787, GR3253215-4, 1:1000), CD81 (Santa Cruz Biotechnology, Inc., USA, sc-166029, lot number: F2416, 1:100), tumor susceptibility gene 101 (TSG101) (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy, T5701, lot number: 037M4775V, 1:1000) and β-tubulin (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy, T5201, lot number: 015M4819V, 1:500) were employed in this study. After incubation with the appropriate horseradish peroxidase-conjugated secondary antibodies (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy) for 1 h at RT, immunoreactive bands were detected by chemiluminescence (ECL substrate, Thermo Fisher Scientific, Waltham, MA, USA). In parallel, we used reversible Ponceau staining to ensure that the gels were loaded evenly.

RNA was extracted from cultured cells in order to evaluate the expression of certain genes using quantitative real-time PCR (RT-qPCR). Following the manufacturer's instructions, RNA was extracted using the AurumTM Total RNA Mini Kit (Bio-Rad Laboratories, Inc., Italy); 1 µg of RNA was retrotranscribed using the PrimeScript™ RT reagent kit with genomic DNA removal (Takara Bio, Saint-Germain-en-Laye, France). The cDNA was diluted 10 times in nuclease-free water before being used in qPCR. The qPCR was performed on 5µL-diluted cDNA using the QuantiTect SYBR Green PCR Master Mix (Qiagen, Milan, Italy) and according to the manufacturers’ instructions. The primers listed below were used: brain natriuretic peptide (BNP) FW = 5′-CGTTTGGGCTGTAACGCACT-3′, RV = 5′-TCACTTCAAAGGTGGTCCCAG-3′; sarcoplasmic/endoplasmic reticulum calcium (Ca2 +)-ATPase cardiac isoform 2a (SERCA2a) FW = 5′-CCTTTGCCGCTCATTTTCCAG-3′, RV = 5′-GGCTGCACACACTCTTTACC-3′; β-Actin FW = 5′-GGCACCACACCTTCTACAATG-3′, RV = 5′-GGGGTGTTGAAGGTCTCAAAC-3′. All primers were purchased from Sigma-Aldrich and used at a concentration of 500 nm. In accord with MIQE guidelines, relative quantification of transcript expression was performed using the 2-ΔΔCt comparative method [44]. β-Actin was used as housekeeping gene.

Twenty five healthy male RccHan^® Wistar outbred rats (10–12 weeks old, 200–300 g body weight) were subjected to non-reperfused myocardial infarction (MI) by closure of the left anterior descending coronary artery (LAD), as previously described by us [18, 19, 31, 32]. In brief, the overnight fasted animals were anesthetized with an intraperitoneal injection of ketamine (300 μl/animal, Ketanarkon^® 100; Streuli AG) + xylazine (15 μl/animal, Streuli AG), intubated, and mechanically ventilated while ECG and body temperature were monitored. 1–2% isoflurane (in 60% air, 40% oxygen) was used to maintain general anesthesia. The LAD was permanently tied off near its emergence (~ 2 mm) below the left atrium with 6–0 silk. Five rats (20%) died immediately after coronary ligation as a result of persistent ventricular fibrillation. Following coronary ligation, the surviving animals' left ventricular region bordering the infarcted area was injected three times with 100μL PBS (n = 7, vehicle), F-EXOs (n = 7, 10¹² nanoparticles), or SFN-F-EXOs (n = 6, 10¹² nanoparticles) before chest closure and suture. The quantity of exosomes to inject was determined by the number of nanovesicles measured per fibroblast in our tests, as well as the relative number of fibroblasts compared to other cardiac cell types [45], and it is consistent with earlier research that used exosome injection in rat hearts [19].

The surgical pain was controlled by daily administration of Meloxicam (2 mg/kg subcutaneus initial dose followed by 1 mg/kg subcutaneous for 2d, Mobic, Boehringer Ingelheim Pharma GmbH & Co KG, Biberach Germany). Rats were anesthetized with 2% isoflurane and sacrificed by lethal potassium chloride injection, which caused diastolic arrest of cardiac activity, 28 days after MI.

Transthoracic echocardiography was performed on fasting infarcted rats that had been sedated (14 mg/kg Xylazine + 40 mg/kg Zoletil 100 /animal; intraperitoneal) (PBS, n = 5; F-EXOs, n = 5; SFN-F-EXOs, n = 5) at 48 h (baseline) and 28d after MI to analyze global cardiac function [18, 19] using the Vevo® 2100 high-resolution Imaging System (FUJIFILM VisualSonics Inc., Toronto, Canada) equipped with a 13-24 MHz ultra-high frequency linear array transducer. Images were taken through the left parasternal window while heart rate and rhythm were monitored electrocardiographically. M-mode recordings were obtained by taking short-axis two-dimensional images of the left ventricle at the level of the papillary muscle. According to American Society of Echocardiography guidelines, the global left ventricular (LV) contractile function was measured from the 2D long-axis view and expressed as the ejection fraction (EF, %) during monitoring of heart rate and rhythm. LV fractional shortening (FS, %) refers to the percentage change in end-diastolic (LVEDD) and end-systolic (LVESD) left ventricular diameter as a parameter of how well the left ventricle is contracting itself and therefore reduces the size during systole; it was expressed as (LVEDD-LVESD /LVEDD) × 100. The changes in left ventricular ejection fraction (LVEF, %) were calculated as (LVEF(Day 28) − LVEF(baseline) / LVEF(baseline)) × 100.

Myocardial tissue was collected and fixed for 24–48 h in 10% formalin. The samples were dehydrated in an ethanol gradient series, cleared in xylene and embedded in paraffin wax (56 °C). Serial sections with thicknesses ranging from 5 to 7 μm were produced. Sections were deparaffinized in xylene and rehydrated using an ethanol series, as needed. Following PBS washing, slides were mounted with Vectasheild (DBA Italia srl, Segrate, Italy) and images were captured using a fluorescence microscope (DM2500; Leica Microsystems srl, Italy). Capillary density was measured in terms of capillary number/mm². Antigen retrieval for cardiomyocyte size assessment was accomplished using Lab Vision Trypsin enzymatic pretreatment (Thermo Fisher Scientific, Waltham, MA, USA). Before incubating with primary or secondary antibody, slides were rinsed in deionized water and PBS-tween 20 (PBST). Masson’s Trichrome staining with aniline blue (Bio-Optica Milan, Italy) was used for infarct size analysis according to manufacturer's instructions; the infarct scar zone (blue stained) was quantified by ImageJ software and expressed as percentage of left ventricle size, as previously described [32].

Capillary density in the infarct border zone was determined by staining with Wheat Germ Agglutinin (WGA, Invitrogen™, Thermo Fisher Scientific, Waltham, MA, USA) [46] for 30 min at a concentration of 100 μg/mL, as directed by the manufacturer.

Cardiomyocyte cell area was measured by immunofluorescence, labeling cardiomyocytes with a mouse anti-α-Sarcomeric Actinin (α-SA) antibody (Sigma-Aldrich, Merk Life Science S.r.l., Milan, Italy, A7811, lot number: 111M4845 and 055M4827V 1:100) overnight at 4 °C, followed by anti-mouse Alexa Fluor-568 (Abcam, Cambridge, UK) for 90 min at RT. After removing the excess secondary antibody the slides were mounted with DAPI containing Vectasheild (DBA Italia srl, Segrate, Italy). The LV infarct border zone was analyzed in each animal. Cross-sectional cell area was manually measured by ImageJ software and expressed as pixels (Additional file 1).

Graphical depiction and statistical analysis were performed using GraphPad Prism, commercially available software. The data are presented as the mean ± SEM of at least three independent replicates. To compare more than two treatment groups, one -way ANOVA with Tukey post-hoc test was used. To compare two independent groups, Student’s t-test for unpaired values was used. The threshold for statistical significance was chosen at P-value < 0.05.

Depending on the concentration, treatment time, and cell type, SFN has opposing effects on cell death and survival [47]. A preliminary analysis of the dose-time-response relationship of SFN in cultured murine fibroblasts is an important first step in determining the appropriate dose. For 3 and 7 days, cells were treated with SFN concentrations ranging from 1.5 to 48 µM. SFN exhibited no harmful effect on murine fibroblasts at low doses, while increased concentrations lowered cell viability (Fig. 1A-B). SFN treatment of fibroblasts for 3 days reduced cell viability from 12 µM to the highest lethal effect seen at 48 µM. (Fig. 1A). Fibroblasts remained alive after 7 days of treatment with 12 µM SFN (Fig. 1B). Interestingly, 7d treatment with a lower dose of SFN (1.5 µM) boosted cellular metabolic activity, which is consistent with previous findings [47, 48] (Fig. 1B). Conversely, SFN vehicle (DMSO 0.01% (v/v)) had no effect on cell viability at any time point, supporting SFN toxicity at higher doses (Fig. 1A-B). In the light of our findings and earlier research [48], we chose a 3 µM SFN dose and a 7d treatment time with medium changes every 3d to concentrate exosomes (Fig. 1C).

SFN is a dietary dose-dependent inhibitor of class I and II histone deacetylases (HDAC) [24]. As shown in Fig. 1D-E, 7d treatment of NIH/3t3 cells with 3 or 6 µM SFN did not significantly modify histone AcH4 and AcH3 levels. Our findings are consistent with those obtained in human fibroblasts treated with similar dose of SFN [48]. Similarly, 7d treatment of NIH/3t3 cells with 3 or 6 µM SFN did not change Nrf2 expression, as shown in Fig. 1F,

Exosomes were isolated from the conditioned medium of NIH/3t3 fibroblasts after 7 days of treatment with 3 µM SFN using repeated centrifugation and ultracentrifugation. NTA determined the size and concentration of exosomes in accordance with earlier researches [18, 19, 21, 37, 38]. The median exosome size for F-EXOs was 140.5 ± 6.2 nm and 158.0 ± 7 nm for SFN-F-EXOs, as shown in Fig. 2A, which is within the acceptable exosomal size range determined using this technique [49]. Further investigation revealed that the majority of the exosomes released by fibroblasts were between 100 and 160 nm in size, with the number of larger SFN-F-EXOs much higher than F-EXOs of comparable size (Additional file 1: Fig. S1). Indeed, SFN increased fibroblast exosome release by ~ 45% (7.42 ± 0.66 · 108 SFN-F-EXOs versus 5.13 ± 0.84 · 108 F-EXOs per 100 mm plate, p = 0.047). We also determined that SFN treated cells released 1,514 ± 378 exosomes each day, compared to untreated cells, which secreted 1,047 ± 484 exosomes per day.

Isolated and purified exosomes were also identified by the expression of recognized exosomal markers CD81, TSG101 and Hsp70 [12]. Of note, SFN-F-EXOs had larger amounts of TSG101 than F-EXOs for the same concentration of exosomes (~ twofold; p = 0.014). Although CD81 was expressed similarly in SFN-F-EXOs and F-EXOs, levels of Hsp70, an anti-apoptotic [21, 37, 50] and anti-fibrotic [51] exosomal protein, were higher in SFN-F-EXOs (~ 2.4-fold; p = 0.024) (Fig. 2B). Intracellular Hsp70 levels were also significantly elevated in SFN-treated fibroblasts (Fig. 2C).

AngII, as seen in Fig. 3A-B, increased the cardiomyocytes size compared to resting cells (PBS)(+ 34.3%, p = 0.0002). SFN-F-EXOs counteracted the hypertrophic response induced by AngII (-27.8% vs PBS, p < 0.0001; -17.2% vs F-EXOs, p = 0.02), whereas F-EXOs at identical concentration had no affect on the cell size of stressed cardiomyocytes (Fig. 3A-B).

As shown in Fig. 3C, BNP expression, an established marker of LV hypertrophy induced by AngII [52], was considerably reduced in AngII-stressed HL-1 treated with SFN-F-EXOs (-16%, p = 0.025), whereas a similar amount of F-EXO did not. Then, we assessed the expression of SERCA2a, a sarcoplasmic reticulum Ca2 + ATPase whose up-regulation restricts cardiomyocyte hypertrophy [53], to further analyze the anti-hypertrophic impact. Long-term treatment of AngII-stressed cardiomyocytes with SFN-F-EXOs raised SERCA2a expression compared to PBS (1.46-fold vs PBS, p = 7.18·10^–6) and to similar amount of F-EXOs (1.14-fold versus F-EXO, p = 0.044), which slightly enhanced SERCA2a levels compared to PBS (p = 0.001) (Fig. 3D).

AngII causes the production of reactive oxygen species (ROS), which are responsible for the development of heart hypertrophy [54]. Anion superoxide generation was significantly increased in HL-1 cardiomyocytes exposed to AngII (~ fivefold, p < 0.0001), as seen in Fig. 4A-B. The treatment of stressed cardiomyocytes with SFN-F-EXOs reduced the oxidative burst relative to untreated stressed cells (~ twofold decrease vs AngII, p < 0.0001; vs F-EXOs p = 0.001), while ROS levels remained somewhat higher than unstressed cells (vs. PBS p < 0.01). F-EXOs, on the other hand, have no anti-oxidant properties.

Prolonged exposure to AngII, a mediator of anoxic damage, causes HL-1 cardiomyocytes apoptosis [55]. AngII activated caspase-3 in stressed HL-1, as seen in Fig. 4C-D, which was significantly prevented by both SFN-F-EXOs and F-EXOs.

Long-term treatment with SFN-F-EXOs, as shown in Fig. 5A-B, raised AcH4 levels in murine cardiomyocytes exposed to AngII (1.46 fold vs PBS, p = 0.048; 1.52-fold vs F-EXOs, p = 0.025), whereas no significant changes were detected in cardiomyocytes treated with F-EXOs. Because histone deacetylase 1 (HDAC1) inhibition lowers pathological cardiac hypertrophy [56] and oxidative stress [57], we assessed the exosomal level of maspin, an HDAC1 inhibitor delivered by exosomes [58, 59]. Maspin was released from fibroblasts into exosomes, as seen in Fig. 5C-D, and its levels were unaffected by SFN treatment. As a result, we postulated that histone acetylation levels may be affected by the extent to which maspin-rich exosomes are incoroporated into cardiomyocytes. Indeed, we found that SFN-F-EXOs were taken up more efficiently by cardiomyocytes grown with (Fig. 5E-F) and without AngII (Additional file 1: Fig. S2) than F-EXOs (5.6-fold vs F-EXOs, p < 0.0002).

We studied exosome uptake by MCEC to better understand the cardiomyocyte tropism of exosomes, We found that endothelial incorporation of exosomes was substantially lower than that of cardiomyocytes (34-fold lower in average), but MCEC uptake of SFN-F-EXOs and F-EXOs was equivalent in both experimental settings (Additional file 1: Fig. S3).

The pre-surgery heart rate of all anesthetized rats was 398 ± 18 bpm during ECG monitoring, with no arrhythmia. To test the possible anti-remodeling effect of SFN-F-EXOs or F-EXOs, similar dose of exosomes (1·10¹² nanoparticles) was injected in the region bordering the infarct zone of rats with permanent LAD ligation. Five rats died suddenly during the first 24 h after surgery: 1/6 (15%) in the SFN-F-EXOs group, 2/7 (30%) in the F-EXOs and PBS groups. At 4 weeks, the surviving infarcted rats showed significant reduction in LVEF after treatment with PBS (-31.33%) or F-EXOs (− 28.7%) (Fig. 6B and E), whereas body weight loss was increased (− 5.5 ± 1%, p < 0.05) as compared to the pre-surgery state. Although SFN-F-EXOs group had significantly lower LVEF than PBS and F-EXOs groups at 48 h after MI (Fig. 6A), we observed better recovery of global cardiac function after intramyocardial injection in infarcted heart of SFN-F-EXOs rats (+ 36.54%) at 28d versus 48 h (baseline) compared to PBS and F-EXOs groups (Fig. 6E), without changing of heart rate (Fig. 6C-D).

The infarct scar size in rats treated with SFN-F-EXOs was considerably smaller compared to PBS (-46%, p = 0.039) and F-EXOs groups (− 47%, p = 0.044) (Fig. 6F-G), while the infarct scar size in F-EXOs animals was identical to PBS group (Fig. 6F-G). Indeed, a single intramyocardial injection of SFN-F-EXOs resulted in a significant decrease in cardiomyocyte cell area in the infarct border zone (-39.1% vs PBS, p = 0.0078; -36.4% versus F-EXOs, p = 0.027) (Fig. 7A). In contrast, F-EXO intramyocardial injection did not limit post-MI cardiomyocyte hypertrophy. Finally, the number of myocardial capillaries in the infarct border zone was comparable in all groups (Fig. 7B).

Our initial proof of concept study will support more in-depth research into the molecular mechanisms behind the effect of low-dose SFN on exosome biogenesis. The first characterization of the SFN-F-EXOs phenotype provided by our study strongly encourages future thorough investigations of exosomal cargo to establish quantitative production standards for clinical use. Additional experiments should be done to assess the clinical value of prolonged systemic administration of SFN-F-EXOs in large animal models of MI during hemodynamic monitoring, in order to develop a new therapeutic strategy for HF prevention that is easily transferable to the hospital setting, and even to validate a risk score model for predicting the cardiovascular outcome after treatment.

We demonstrated that continuous low-dose SFN treatment is a novel chemical-based method for producing anti-remodeling exosomes from a noncardiac fibroblast cell line. SFN increases the tropism of noncardiac exosomes towards cardiomyocytes, which higher internalization increases AcH4 levels, induces SERCA2a expression, lowers apoptotic oxidative stress and hinders the hypertrophic response. Our approach could pave the way for safe and cost-effective large-scale production of natural cardioprotective exosomes for therapeutic use without the need for vector-mediated gene editing of allogeneic fibroblasts to limit immunogenicity or for biomimetic exosomes to improve the distribution to the target cell. Furthermore, the cardioprotective procedure requiring a single myocardial injection of ready to use stable allogeneic SFN-F-EXOs, previously stored at − 80 °C avoiding freezing–thawing [80], can serve as good emergency cell-free treatment to rescue a damaged heart without waiting for the time-consuming autologous exosome generation, especially in older and frail patients.