2.1 ModRNA synthesis and formula

ModRNA was synthesised according to previously described methods.^{25, 26} In short, a linear DNA template was utilised for T7 RNA polymerase-mediated transcription. Uracil was replaced with N¹-methylpseudoridine in the process of in vitro transcription synthesis. RNA purification was performed using Ambion MEGAclear spin columns. Then, residual 5′-phosphates were removed using Antarctic Phosphatase (New England Biolabs). The purity and concentration of the modRNA was evaluated by NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and suspended at a concentration of 1 µg/µL for further use.

2.2 Cell culture

The acquisition and use of hADSCs were approved by the Clinical Ethics Committee of Shanghai Ninth People's Hospital (No. SH9H 2018-T22-1). In brief, hADSCs were cultured in mesenchymal stem cell medium (ScienCell, USA) containing 5% fetal bovine serum and 1% penicillin-streptomycin antibiotic. All cells were incubated in a humidified atmosphere of 5% CO2 at 37°C. Cells between passages 3 and 5 were used for all experiments.

2.3 ModRNA transfection and evaluation

ModRNA was transfected into hADSCs using Lipofectamine RNAiMAX Reagent (Invitrogen, California, USA). Liposome-modRNA complexes were produced by mixing the properly diluted modRNA and RNAiMAX reagents. After transfecting cells for 4 h, the complexes-containing culture medium were changed. In each well of a 6-well plate, 4 µg of modRNA and 10 µL of RNAiMAX reagent was used to transfect approximately 1 × 10⁶ hADSCs.

Transfection efficiency was evaluated by transfecting hADSCs with green fluorescent protein (GFP) modRNA. Twenty-four hours after transfection, the digested cells were detected using a CytoFLEX LX flow cytometer (Beckman Coulter, CA, USA).

Calcein acetoxymethyl ester (AM)/propidium iodide (PI) cytotoxicity and cell viability assay kits (Beyotime, China) were used to assess the toxicity of modRNA transfection to hADSCs. After staining with Calcein AM (1:1000) and PI (1:1000) for 15 min, hADSCs were observed using a fluorescence microscope (DMI3000 B, Leica, Heidelberg, Germany). The ImageJ 1.54f was used to perform statistical analyses.

2.4 Enzyme-linked immunosorbent assays (ELISAs)

VEGFA and bFGF ELISA kits (MultiSciences, China) were employed to quantify the expression kinetics of VEGFA and bFGF of hADSCs after transfection. Conditioned media were harvested 12, 24, 48, 72, 96 120 and 168 h after transfection for analysing the protein levels of VEGFA and bFGF.

For the in vivo expression of VEGFA and bFGF, samples were harvested at 24, 72 and 168 h post cell treatments, respectively. The tissues were homogenised and centrifuged prior to content detection following the kit instructions.

Blood was drawn into clean tubes and centrifuged at 3500 rpm for 15 min. The serum was obtained and used to detect the levels of cardiac Troponin I (cTnI) and creatine kinase-MB (CK-MB) using ELISA kits (Elabscience, China). All tests were repeated thrice for each group.

2.5 Tube formation assay

To assess the effects of the conditioned medium from hADSCs transfected with luciferase modRNA, VEGFA modRNA, and/or bFGF modRNA on the tube formation ability of human umbilical vein endothelial cells (HUVECs), HUVECs were collected and inoculated into 48-well plates coated with Matrigel (Corning, USA), then cultured in 300 µL of the abovementioned conditioned medium, respectively. Tubes were observed every 2 h. Subsequently, ImageJ was used to count the tubes and branches. All tests were repeated five times.

2.6 Transwell assay

To further evaluate the effects of the above-mentioned conditioned medium from hADSCs on the migratory ability of HUVECs, HUVECs were incubated in the upper chambers of the transwell plates with an 8-µm pore size (Corning, USA). The upper chamber of the transwell plates was inoculated with the medium of hADSCs transfected with different modRNAs, and the lower chamber was filled with medium containing 10% fetal bovine serum. After 24 h, migrated cells were fixed in 4% paraformaldehyde (PFA) and stained with 0.2% crystal violet (Aladdin, China). The migrated cells were observed using a fluorescence microscope (DMI3000 B, Leica, Heidelberg, Germany), and three randomly selected areas were used for statistical analysis.

2.7 Flow cytometry

H9C2 cells were seeded at a density of 1×10⁵ per well in a 12-well plate and cultured under normoxic conditions for 24 h to allow for cell adhesion. Subsequently, the culture medium of H9C2 cells was replaced with the conditional medium. Thereafter, the cells were cultured under hypoxic conditions (1% O₂) for 30 h. FITC-Annexin V apoptosis detection kits (Tonbo, China) were used to investigate apoptosis. The cells were analysed by flow cytometry (CytoFLEX, Beckman, Coulter). The data obtained were analysed using FlowJo 10.5 software (Tree Star Inc., Ashland, OR, USA).

2.8 Cell Counting Kit-8(CCK8) assay

To evaluate the proliferation of smooth muscle cells (SMCs), SMCs were seeded into 96-well plates at a density of 3×10³ cells/well and incubated in the conditioned medium of hADSCs transfected with different modRNAs. CCK8 (Dojindo, Kumamoto, Japan) assay was conducted at three time points (D0, D1, and D3). A microplate reader (Tecan i-control, Tecan, Switzerland) was used to record the optical densities at 450 nm.

2.9 Animal studies

Sprague-Dawley rats (male, 200–250 g) were provided by Shanghai Jihui Experimental Animal Co. (Shanghai, China). All animal experiments were approved by the Laboratory Animal Welfare Ethics Committee of the Shanghai Children's Medical Center (Approval No, SCMC-LAWEC-2019-009). The rats were randomly divided into six groups: (1) Sham group; (2) MI group; (3) hADSCs^Luc group (150-µL mixture of Matrigel (Corning) and hADSCs transfected with 20 µg Luciferase modRNA (5×10⁶ cells/rat) was injected into the left ventricular border area after MI); (4) hADSCs^VEGFA group (150-µL mixture of Matrigel (Corning) and hADSCs transfected with 20 µg VEGFA modRNA (5×10⁶ cells/rat) injected into the left ventricular border area after MI); (5) hADSCs^bFGF group (150-µL mixture of Matrigel (Corning) and hADSCs transfected with 20 µg bFGF modRNA (5×10⁶ cells/rat) injected into the left ventricular border area after MI); and (6) hADSCs^dual group (150-µL mixture of Matrigel (Corning) and hADSCs transfected with 20 µg VEGFA and 20 µg bFGF modRNA (5×10⁶ cells/rat) injected into the left ventricular border area after MI). The rat model was established according to previously described methods.¹⁶ Under continuous isoflurane anaesthesia, we performed left thoracotomy and permanent ligation of the left anterior descending coronary artery with a 6-0 thread to induce MI in the rats. During the operation, electrocardiogram (ECG) was employed to verity the successful establishment of the MI model.

For the treatment group, myocardial injections of the cells were administrated at three fixed positions at the infarct border zone 10 min following the ligation of the left anterior descending coronary artery in the rats. The modified cells were provided to the surgeon with assigned numbers only, without specific group information. The cells were injected into the recipient heart immediately after collection to maintain as much cell activity as possible. Subsequently, the chest was closed. To avoid immune rejection, all rats received 5 mg/kg/day methylprednisolone and 0.25 mg/kg/day tacrolimus every 12 h from 1 day before MI to the day they were sacrificed.

A total of 84 rats were used in the experiment. However, 6 rats died within 3 days of the model establishment. These rats showed no obvious group-specific distribution pattern. Since no subsequent data were collected from these rats, they were excluded for further analysis.

2.10 Triphenyl tetrazolium chloride (TTC) staining

TTC (MP BIO, California, USA) staining was used to assess the area of the MI. Dehydrogenase present in normal myocardial tissue can reduce TTC to red water-insoluble formazan, whereas the infarct area is white owing to a lack of dehydrogenase. Prior to staining, 2% TTC was fully dissolved in PBS. Twenty-four hours after MI, the rats were euthanised and their hearts were collected, fully washed with precooled PBS, and then sectioned into five pieces. After stained with 2% TTC solution, the heart slices were fixed with preheated 4% PFA for 30 min. Images of the stained heart slices were taken and the area of the left ventricular infarction was analysed using ImageJ.

2.11 Bioluminescence imaging

Bioluminescence imaging of the transplanted hADSCs was performed using an in vivo imaging system (VISQUE® Smart-LF, Vieworks, Korea). After the MI rat model was established, 5 ×10⁶ hADSCs transfected with VEGFA and bFGF modRNA were intramyocardially injected. After 24 h, rat hearts, livers, spleens, lungs and kidneys were harvested for imaging to evaluate the in vivo distribution of the injected cells.

2.12 Echocardiography

2.13 Epicardial activation mapping

The rats were anaesthetised 15 min and 4 weeks following MI surgery. Tracheal intubation and isoflurane were used to maintain anaesthesia. After left thoracotomy, 8 × 8 microelectrodes (MappingLab Inc., UK) were placed on the epicardial surface to record the epicardial electrical activity. The obtained active waveform was amplified using a filter amplifier (MappingLab Inc., UK) and then transmitted to a computer. All activation times were digitised and used to draw activation diagrams. Activation time was calculated as the maximum negative slope point of the active waveform. The conduction time (CT), conduction velocity (CV) and non-uniformity index were analysed using EMapScope 4.0 software (MappingLab Inc., UK).

2.14 Histology and immunohistochemical staining

Rat hearts were collected 1 and 4 weeks after MI. Haematoxylin-eosin staining (Solarbio, China) and Masson's trichrome staining (Solarbio, China) were performed. Scar tissue (%) and LV wall thickness of the heart were calculated using ImageJ.

A one-step terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) apoptosis detection kit (Servicebio) was used for TUNEL staining following the manufacturer's protocol. In short, the slices were hydrated with xylene and a series of graded ethanol solutions and stained with TUNEL reagent. Finally, the slices were counterstained with DAPI. Images were obtained using a laser scanning confocal microscope (TSC SP8, Leica).

For immunohistochemistry, antigen retrieval was performed using citric acid buffer (Yeasen, China). After being permeabilised with 0.2% Triton X-100 (Beyotime, China) and blocked with 5% bovine serum albumin, the heart slices were incubated with the primary antibodies at 4°C overnight and then stained with the secondary antibody at RT for 2 h. Finally, the samples were stained with DAPI (Yeasen, China). The primary antibodies used in this study included CD31 (ab182981, Abcam), Ki-67 (ab16667, Abcam), cTnT (ab8295, Abcam) and α-SMA (ab7817, Abcam). A confocal laser-scanning microscope (Leica; TSC SP8) was used to record the results.

2.15 Statistical analysis

All data are shown as mean ± standard deviation (SD). Prior to further statistical analyses, the normality of the data distribution was examined using Shapiro–Wilk test. Depending on the data type, Student's t-tests, one-way analyses of variance (ANOVAs) followed by Tukey's test, or two-way ANOVAs followed by the Bonferroni post hoc test (GraphPad Software, San Diego, CA, USA) were performed. The results were obtained from at least three independent experiments, and a value of p < .05 was deemed as statistically significant.

3.1 Efficient transfection of modRNA in hADSCs

The GFP reporter gene was used to verify the expression and transfection efficiency of modRNA in hADSCs. After 24 h, GFP protein expression level was detected using flow cytometry. The expression of GFP protein lasted for at least 5 days following transfection (Figure 1A). Furthermore, the transfection efficiency of modRNA reached 82.93% ± 1.5% within the first 24 h after transfection (Figure 1B and C). In addition, AM/PI staining showed that transfection with the modRNA had no significant effect on cell viability (Figure S1A and B).

To detect whether hADSCs were effectively transfected with VEGFA and bFGF modRNAs and determine their expression dynamics, we transfected luciferase, VEGFA and bFGF modRNAs and monitored protein expression levels in hADSCs. Since VEGFA and bFGF are secreted proteins, we used ELISAs to measure the levels of newly generated and accumulated proteins in the cell culture medium of hADSCs transfected with VEGFA and bFGF modRNAs. We found that hADSCs transfected with VEGFA or bFGF modRNA began to rapidly secrete VEGFA or bFGF protein, respectively, for the first 2 days and continued to express them for more than 5 days. However, by the seventh day, the expression of VEGFA and bFGF proteins in the transfected hADSCs diminished substantially (Figure 1D–G). Secretion in the transfection group was significantly higher than that of the control group at all time points (Figure 1D and E). These results indicate that modRNAs can efficiently transfect hADSCs and express the target protein rapidly and efficiently following transfection.

3.2 In vitro function of proteins secreted by modRNA-modified hADSCs

To evaluate the effects of VEGFA and bFGF modRNA transfection on HUVECs, we quantified the tube formation and migration abilities of HUVECs cultured under different conditions. Compared with the hADSCs^Luc group, the hADSCs^VEGFA, hADSCs^bFGF and hADSCs^dual groups promoted the formation of the tubular structure of HUVECs (Figure 2A–C). The ability of tube formation of the conditioned medium in the hADSCs^dual group was significantly higher than that in the other groups (Figure 2B and C). Consistent with the results of the tube-formation experiment, the transwell assay also showed that the conditioned medium of the hADSCs^dual group increased the migration ability of HUVECs more strongly than that of the other groups (Figure 2D and E). The above results indicated that hADSCs transfected with VEGFA and bFGF modRNAs, both showed strong angiogenic characteristics in vitro, and the combination of the two modRNAs was more efficacious, which further confirmed their therapeutic potential in heart therapy.

Next, conditioned media from different modRNA transfection groups were used to detect their effect on H9C2 apoptosis under hypoxia. We found that, compared to the hADSCs^Luc group, transfection with VEGFA or bFGF modRNA significantly reduced the percentage of apoptotic H9C2 cells under hypoxia. Compared to the hADSCs^VEGFA group, the hADSCs^bFGF and hADSCs^dual groups significantly inhibited H9C2 apoptosis (Figure 2F and G). These results suggest that the bFGF modRNA exerts a stronger inhibitory effect on H9C2 apoptosis than VEGFA modRNA.

To evaluate the effect of the conditioned medium of hADSCs transfected with VEGFA and bFGF modRNA on SMCs proliferation, cell counting was performed at 0, 1 and 3 days after SMC inoculation to quantify cell proliferation. The results showed that on the third day, the hADSCs^VEGFA, hADSCs^bFGF and hADSCs^dual groups significantly promoted the proliferation of SMCs and the hADSCs^dual group showed the most favourable results (Figure 2H).

3.3 hADSCs^dual improved cardiac function following MI in vivo

Next, we investigated the potential of modRNA-transfected hADSCs for MI treatment. Animal experimental procedures are shown in Figure 3A. All rats that underwent surgery experienced transmural MI. Typical changes were expressed in the ECG of rats at 5 min after ligation of the left anterior descending coronary artery (LAD). In comparison to the ECG prior to ligation, ECG of rats after LAD ligation exhibited an elevation of the ST segment (Figure S2A). Cardiac ultrasound detection showed that the ejection fraction in the MI group was significantly decreased to a range between 25% to 35% (Figure S2B). Compared with the sham group, 24 h after the model was established, CK-MB and cTnI serum levels in the MI group were remarkably elevated, with an increase of more than tenfold (Figure S2C and D). TTC staining showed that the colour boundary between the infarcted and normal myocardium was clear, and the percentages of the infarct areas of the left ventricle were stable at 37% (Figure S2E and F).

To evaluate the safety of hADSCs, a series of in vivo experiments were conducted. The in vivo distribution of hADSCs was examined 24 h after the transplantation procedure. The results demonstrated that at 24 h after the surgical procedure, the hADSCs were principally distributed within the heart, while no prominent presence of hADSCs was discernible in the liver, spleen, lungs or kidneys (Figure S3A). HE staining was also performed. The histological examination under microscopy did not show any overt pathological alterations (Figure S3B), which implied the satisfactory biocompatibility and safety of hADSCs transplantation, thus providing a basis for further investigation into their long-term behaviours and therapeutic efficacies.

The expression efficacies of VEGFA and bFGF modRNA in the infarcted zone are a prerequisite for their therapeutic efficacy. ELISA kits were employed to detect the expression levels of VEGFA and bFGF proteins in rat hearts. We measured the protein contents of VEGFA and bFGF at 24, 72 and 168 h postoperatively. The results indicated that during the first 72 h after transplantation, the expression of the target proteins in the hADSCs^dual group was significantly higher than that in the MI group (Figure 3B and C). However, at 168 h, there was no significant difference in the protein levels between the hADSCs^dual group and the MI group (Figure 3B and C).

To confirm whether modRNA-modified hADSCs affected LV function, we performed a series of echocardiographic measurements after MI surgery and cell treatment. Specifically, the following six groups were established: sham, MI, hADSCs^Luc, hADSCs^VEGFA, hADSCs^bFGF and hADSCs^dual groups. Echocardiography was performed on rats from different groups 1, 2 and 4 weeks after MI and treatment (Figure 3D). Echocardiography revealed that the LV function of rats in each group decreased significantly 1 week after the MI operation, and the extent of decline was comparable among all groups at this time point (Figure 3E). Cardiac ultrasound detection showed that there was no statistically significant difference in the LVEF and LVFS in each group in the first week after MI with/without cell transplantation (Figure 3E and H). However, 2 weeks after MI induction, LVEF and LVFS in rats treated with hADSCs^VEGFA and hADSCs^dual increased significantly in the hADSCs^VEGFA and hADSCs^dual groups compared with other MI groups (Figure 3F and I). At the fourth week, the LV function of rats in the hADSCs^dual group was significantly improved, LVEF increased from 32.71% ± .93% to 49.43% ± 2.90%, and LVFS increased from 15.94% ± .45% to 26.05% ± 1.79% (Figure 3G and J). The LV function in rats that received hADSCs^dual was found to be significantly higher than that of rats that received either hADSCs^VEGFA or hADSCs^bFGF treatments. The LV function of hADSCs^VEGFA improved slightly to 43.72% ± 3.46%, while LVFS increased to 22.51% ± 1.96% (Figure 3G and J). The LV function of the hADSCs^bFGF group slightly improved to 42.10% ± 2.80%, while LVFS increased to 21.38% ± 1.71% (Figure 3G and J). The treatment effects of modRNA-modified hADSCs all exceeded that of the non-cell-based treatments. Echocardiographic parameters increased significantly in rats receiving hADSCs^dual, illustrating the therapeutic potential of hADSCs^dual as a novel cardiovascular cell therapy for improving cardiac dysfunction after MI.

3.4 hADSCs^dual promoted left ventricular electrical conduction after MI in vivo

To evaluate changes in left ventricular electrical conductivity after MI surgery and treatment interventions, we measured the left ventricular electrical conduction at 15 min and 4 weeks post-MI. Compared to the sham group, all groups showed disordered conduction (Figure 4A), and the dispersion of electrical activity significantly increased at 15 min post-MI surgery (Figure 4B). At the fourth week, compared to the MI group and hADSCs^Luc group, rats receiving hADSCs^VEGFA, hADSCs^bFGF, or hADSCs^dual showed more orderly conductivity (Figure 4A), and the dispersion of electrical activity decreased (Figure 4B). Quantitative analysis of CT, CV, and dispersion of electrical activity in different groups was performed 15 min and 4 weeks after MI surgery and intervention treatment. We found that after MI surgery, the CT was prolonged, CV was significantly reduced, and conduction dispersion increased (Figure 4C–E). After treatment with hADSCs^VEGFA, hADSCs^bFGF or hADSCs^dual for 4 weeks, the CT of the left ventricle became shorter, the CV increased, and the dispersion of conduction decreased (Figure 4F–H). Compared to the other groups, the left ventricular conduction parameters of the hADSCs^dual group were significantly improved, supporting the therapeutic potential of hADSCs^dual as a new treatment method to improve left ventricular conductivity.

3.5 hADSCs^dual reduced cardiac fibrosis and preserved ventricular wall thickness

The hearts were harvested 4 weeks after modelling. Haematoxylin-eosin staining confirmed that, compared with other groups, the proportion of regenerated tissue in the heart of rats treated with hADSCs^dual was significantly increased (Figure 5A). The left ventricular scar area of rats, as detected by Masson staining, was significantly smaller in the hADSCs^dual group, and the ventricular wall thickness was well-maintained (Figure 5B and C). Quantitative analysis of the left ventricular infarction area and wall thickness of rats in each group also confirmed this finding (Figure 5D and E). Vimentin staining also revealed numerous fibroblasts in the infarcted zone in all groups 4 weeks post-MI (Figure 5F). However, compared to the MI group, the hADSCs^dual group showed fewer infiltrating fibroblasts (Figure 5F and G). Overall, hADSCs^dual reduced cardiac fibrosis, alleviated ventricular remodelling and inhibited the progression of myocardial injury induced by MI to a certain extent.

3.6 hADSCs^dual reduced apoptosis and promoted cell proliferation in the infarcted area

Previous studies have shown that MSCs exert their therapeutic effects primarily through paracrine signalling; therefore, we investigated the survival of hADSCs after implantation in vivo. To compare the engraftment between hADSCs^dual and hADSCs^Luc, we used CMDIL⁺ hADSCs to exert their therapeutic effects. hADSCs were found in all rats 1 and 4 weeks following MI surgery and hADSCs treatment (Figure S4A and B). Additionally, most hADSCs gathered in the infarct and border zones, suggesting minimal migration following injection. Interestingly, a significantly larger percentage of hADSCs pretreated with VEGFA and/or bFGF modRNA survived in the infarcted zone (Figure S4C and D).

To gain further insights into the mechanism of cardiac function improvement and fibrosis reduction, we first assessed the apoptosis of cells in the infarct region. Treatment with hADSCs^dual inhibited apoptosis in the infarcted zone. As shown by the TUNEL staining results, compared with the MI group, the proportion of apoptotic cells in the infarct zone in the hADSCs implantation group decreased to a certain extent 1 week following treatment (Figure 6A). Compared with the hADSCs^Luc group (9.54 ± 1.70%), the hADSCs^VEGFA group (6.88 ± 1.03%), hADSCs^bFGF group (4.39 ± 1.03%), and hADSCs^dual group (4.17 ± .95%) showed significantly fewer apoptotic cells in the infarcted area (Figure 6B). In addition, Ki-67 staining was performed to assess cell proliferation in the infarcted zone 1 and 4 weeks following MI surgery and treatment (Figure 6C and D). Compared to the hADSCs^Luc group, hADSCs^dual significantly promoted cell proliferation in the infarcted zone (Figure 6E and F). To further clarify the types of cells undergoing apoptosis and proliferation, immunostaining for cardiac Troponin T (cTnT) was carried out. Although we detected TUNEL-positive cells in the vicinity of residual cardiomyocytes expressing cTnT, along with a small number of cells showing co-expression of cTnT and Ki-67 (Figure S5A and B), the specific cell types were difficult to identify.

3.7 hADSCs^dual promoted stable vascular regeneration in vivo

To verify the capacity of hADSCs^dual to promote neovascularisation within the grafted region and whether hADSCs^dual can better promote angiogenesis in the infarct region, immunofluorescence double-staining with CD31 and α-SMA staining were conducted. The staining results showed that, compared with the MI and luciferase modRNA transfection groups, the number of blood vessels in the infarcted area was significantly greater after treatment (Figure 7A). hADSCs^VEGFA was superior to hADSCs^bFGF in promoting vascular regeneration, indicating that VEGFA played a more critical role than bFGF in promoting vascular regeneration. Interestingly, we observed that most of the newly formed vessels in the MI area co-expressed CD31 and α-SMA in the rats treated with hADSCs^dual, indicating mature blood vessels formation (Figure 7A, C and D). In addition, we found that the density of capillaries and mature vessels in rats receiving hADSCs^dual were significantly increased at the border zone of the MI area, when compared with those of all other treatments (Figure 7B, E and F).