P2Y₂ knockout (KO) CD1/C57BL/6J mice were a kind gift from Dr. B. Koller (University of North Carolina, Chapel Hill, United States) (Homolya et al., 1999). 12- to 16-week-old male and female wild type (WT) and P2Y₂ KO mice were used randomly for cADSC isolation and ischemia experiments.

cADSC were isolated from the stromal vascular fraction of wild type (WT) and P2Y₂ KO mouse cardiac adipose tissue and their purity was checked by flow cytometry as previously described (Nagata et al., 2016). The cardiac adipose tissue, located between the visceral pericardium and the epicardium (including epi- and pericardic fat depots) was freshly harvested. Adipose tissue was washed in phosphate-buffered saline (PBS) and minced, followed by a digestion in collagenase A (2.5 g/L) at 37°C for 45 min. The digested tissue was filtered through a 100 μm cell strainer (Corning) and centrifuged at 500 g for 5 min. The supernatant containing adipocytes and debris was discarded. A red blood cell lysis was performed on the pelleted cells by an osmotic shock using ammonium-chloride-potassium (ACK) lysing buffer containing 0.15 M ammonium chloride, 0.01 M potassium bicarbonate and 0.0001 M disodium EDTA. After a last centrifugation at 500 g for 5 min, the resultant pellet, named the stromal vascular fraction, was suspended in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin, plated at a density of 5 × 10⁴ per cm² and incubated at 37°C in a humidified 95% O₂/5% CO₂ atmosphere. cADSC were selected by their high plastic adhesion, non-adherent cells were removed after 24 h, and the culture medium was subsequently replaced every 2 days. The cADSC cultures were not stimulated or stimulated daily with UTP (100 µM) for 7 days in the DMEM medium described above. A fraction of cADSC was used to check their purity before transplantation by flow cytometry using CD90, C105, and CD44 markers, as previously described (Nagata et al., 2016).

WT mice were anesthetized with a mixture of Dormazolam (midazolam, 5 mg/kg), Domitor (medetomidine hydrochloride, 0.5 mg/kg) and Fentadon (fentanyl, 0.05 mg/kg), intubated and mechanically ventilated (rate 130 stroke/min, tidal volume 0.13 ml). Optical magnification loop was used for better visualization of the operation field. A left side thoracotomy was performed between the third and the forth rib, and the pericardium was incised. Once the heart was exposed, MI was induced by the permanent ligation of the left anterior descending artery (LAD) proximal to its bifurcation from the main stem. Successful performance of coronary occlusion was confirmed by blanching of the myocardium distal to the coronary ligation. The thoracic incision was then closed with a 5-0 silk suture at the muscle tissue and a 7-0 silk suture at the skin. An antagonist cocktail of Naloxon (naloxone hydrochloride, 1.2 mg/kg), Anexate (flumazenil, 0.5 mg/kg) and Atipam (atipamezole hydrochloride, 2.5 mg/kg) was finally administered to the mice to reverse the general anesthesia. An intraperitoneal injection of Temgesic (buprenorphine, 0.3 mg/kg) was administered after the surgery and the next morning. Three days after LAD, mice underwent a second thoracotomy, as described above. This was followed by intramyocardial cADSC injections. Briefly, cADSC cultured in DMEM for 7 days, untreated or stimulated daily with UTP (100 µM), were detached with trypsin/EDTA and resuspended in PBS. These cells were then injected at a concentration of 10⁴ cells/µl at 3 different sites along the infarct border zone with a final volume of 10 μl at each site (3 × 10⁵ cells/heart). Sham/control group mice underwent the same experimental procedure excepted that they were subjected only to PBS injection. 14 days after injection, hearts were perfused with PAF 4%, collected, fixed in PAF 4%, and either immerged in isopropanol 70% and embedded in paraffin or immerged in a 10%, 20%, and finally a 30% sucrose solution and embedded in Tissue-Tek OCT compound (Sakura Finetek) and frozen at −80°C.

Paraffin cross-sections (8 µm) of infarcted hearts were cut, fixed in Bouin’s solution (Sigma-Aldrich) and stained with Masson’s trichrome (Sigma-Aldrich), according to the manufacturer’s protocol. Sections were counterstained with hematoxiline and mounted. Images of whole hearts were acquired with NanoZoomer-SQ (Hamamatsu) at 0.23 μm/pixel. Fibrosis was quantified as the relative area of the blue staining (collagen) on five sections per ischemic heart compared to the left ventricle surface, using ImageJ software.

The vascularity of the ischemic myocardium was assessed by staining with specific markers recognizing the presence of capillaries and arterioles. Briefly, 8 µm-thick heart cryosections were stained with an anti-isolectin B4 antibody—Alexa Fluor ™ 647 conjugate (1/400 ON at 4°C) (Invitrogen, Merelbeke, Belgium), an anti-α-smooth muscle actin antibody—Cy3™ conjugate (1/400 1 h at RT) (Sigma-Aldrich, St. Louis, MO, United States) and Hoechst (1/2000 1 min at RT) (Thermo Scientific, Merelbeke, Belgium). The area of the infarct border zone was determined as the 0.5 mm region of histologically intact myocardium surrounding the infarct-related fibrocellular region. Capillary and arteriole density was quantified in this peri-infarct myocardium. The final data were expressed as the number of capillaries or arterioles (<20 µm in diameter) per square millimeter, the percentage is referring to the ratio between control condition (PBS injection) and cADSC injection conditions. These analyses were performed using ImageJ software by examining 30 counting surfaces/field for capillary and arteriole density on a total of five sections per heart, at ×10 magnification, in a blinded fashion. Sections were counterstained with Hoechst to visualize the entire population of cell nuclei within each myocardial section.

Images were acquired at LiMiF (Université Libre de Bruxelles, Faculté de Médecine, Bruxelles, Belgique), on an Axio Observer Z1 inverted microscope (Zeiss) equipped with a Zeiss Axiocam 702 mono camera using a EC Plan NeoFluar ×10/0.3 dry objective (Zeiss). The microscope is equipped with an HBO 100 light source. Three channels were recorded with narrow band-pass filter sets (Zeiss) #49 (ex. 335–383 nm, em. 420–470 nm), #43 (ex. 538–562 nm, em. 570–640 nm), and #50 (ex. 625–655 nm, em. 665–715 nm) used to visualize blue, red and far-red fluorochromes, respectively. Images of 2.3 pixels (Axiocam 702 m) were acquired and recorded as 16-bit czi files. The field of view is defined by the ×10 objective resulting in 1.13 mm by 712.58 micron images. The pixel scaling results in 0.586 micron by 0.586 micron. Settings were kept identical for all conditions. Single plane images were displayed using Zen (Blue Edition) software (Zeiss) and exported as uncompressed TIF images.

RNA-sequencing experiments were performed on WT or P2Y₂ KO cADSCs cultured in DMEM for 7 days, untreated or treated every 24 h with UTP (100 µM). RNA was extracted using the RNeasy Mini Kit (Qiagen) after cell lysis using TRI Reagent Solution (Invitrogen). 1 µg/50 µl of RNA was engaged and the quality was checked using a Bioanalyzer 2100 (Agilent technologies). RNA-sequencing experiments were performed on two different pools of RNA from two cADSC cultures per condition. Complementary DNA (cDNA) libraries were obtained using the TruSeq Stranded mRNA Library Prep kit (NuGen) following manufacturer recommendations. The multiplex libraries (18 pM) were loaded on flow cells and sequences were produced using a HiSeq PE Cluster Kit v4 and TruSeq SBS Kit v3-HS from a HiSeq 1500 (Ilumina). Approximately 25 million paired-end reads per sample were mapped against the mouse reference genome (GRCm38.p4/mm10) using STAR software to generate read alignments for each sample. Annotations Mus_musculus GRC38.87.gtf were obtained from ftp.Ensembl.org. After transcripts assembling, gene level counts were obtained using HTSeq.

A gene enrichment analysis was performed with DAVID software on the RNA-sequencing data to determine which biological processes were enriched for differentially expressed genes. Only genes with CPM > 0.5 and a fold change >2 or <0.5 were considered. The modified Fisher Exact p-value or EASE score was reported by the software and indicated as logarithmic p-value.

Total mRNA was extracted from cADSC, cultured for 7 days in DMEM and untreated or treated daily with UTP (100 µM), as described above. RNA samples were eluted with RNase-Free water and RNA concentration was quantified by NanoDrop. Reverse transcription was performed with 500 ng mRNA, 0.5 µl random hexamers, 1 µl dNTPs (10 mM), 4 µl ×5 First Strand Buffer (250 mM Tris-HCl, pH 8.3, 375 mM KCl, 15 mM MgCl₂), 2 µl DTT (0.1 M), 1 µl RNaseOUT ribonuclease inhibitor (40 U/µl), 1 µl SuperScript II reverse transcriptase (200 U/µl) and RNase-Free water up to a volume of 20 µl. Samples were incubated at 42°C for 50 min to synthesize cDNA and at 70°C for 10 min to stop the reaction. Quantitative polymerase chain reaction (qPCR) experiments were performed using 5 ng cDNA, specific primers (5 µM) for UTP target genes and P2Y receptors, 2x KAPA SYBR FAST qPCR Master Mix [KAPA SYBR FAST DNA Polymerase, reaction buffer, dNTPs, SYBR Green I dye, and MgCl2 (2.5 mM)], 50x KAPA SYBR FAST ROX Low and RNase-Free water up to a 20 µl volume on a 7500 Fast Real-Time PCR System (Applied Biosystems). Expression values for each gene were normalized to the expression of the housekeeping gene Rpl32. Results were expressed as 2^−ΔCt. The specific primers for Ccl5, Cxcl5, Ccl12, Saa3, and Rpl32 genes and for P2ry2 and P2ry4 genes are shown in Table 1.

Supernatants from cADSC cultured in DMEM for 7 days, untreated or treated daily with UTP (100 µM), were collected. CCL5, CXCL5, CCL12, and SAA3 protein levels were quantified by ELISA following the manufacturer’s procedure (R&D Systems). Total proteins were quantified using Pierce™ 660 nm Protein Assay Reagent (Thermo Scientific) supplemented with Ionic Detergent Compatibility Reagent (Thermo Scientific) to normalize ELISA results.

All data are expressed as mean ± SEM, and statistical analysis was performed with Prism Software (version 6; GraphPad, San Diego, CA, United States). Endpoint comparisons between 2 groups were performed using unpaired 2-tailed Student’s t-test. For multiple comparisons, false discovery rate according Benjamini-Hochberg was applied to control for type I false positive errors, FDR was set to 0.05. For parallel repeated-measures studies, 2-way ANOVA was used with Bonferroni post-hoc evaluation to determine the significance for individual time points. A 2-tailed p < 0.05 was considered as significant.

Stromal cells were isolated from cardiac adipose tissue of WT and P2Y₂ KO mice. Non-adherent cells were removed 24 h after the plastic adherence step. Flow cytometry experiments were performed with adherent stromal cells using CD90, CD105, and CD44 as markers of mesenchymal stem cells, and CD26 as fibroblast marker. More than 90% of isolated cells were CD90⁺ CD105⁺ CD44⁺ CD26⁻, and they were cultured for 7 days in proliferation medium (data not shown). These cADSC were then counted for intramyocardial injections (3 × 10 µl containing 10⁵ cells each). LAD ligation was performed on WT mice and intramyocardial injections of PBS (control), WT cADSC and P2Y₂ KO cADSC were performed in the peri-infarct border zone, 3 days after LAD ligation. 14 days after injection, hearts were harvested and embedded in paraffin for further analysis.

Five heart sections per transplanted ischemic heart were stained with Masson’s trichrome to evaluate cardiac fibrosis. Fibrosis was quantified by calculating the area stained blue, expressed as a percentage of the left ventricle’s total area (Figures 1A,B). We observed a strong reduction of cardiac fibrosis in ischemic heart transplanted with WT cADSC compared to mice injected with PBS. Collagen stained area was reduced from 24.0 ± 1.3% of left ventricle area in PBS-injected hearts to 9.0± 1.3% of left ventricle area in WT cADSC-injected hearts (decrease of 62.5 ± 8.5%; mean ± SEM; ***: p < 0.001) (Figure 1A). Very interestingly this cardiac fibrosis reduction was not observed after transplantation of cADSC isolated from P2Y₂ KO mice (Figures 1A,B). Effectively collagen stained area observed in PBS-injected ischemic mice (24.0 ± 1.3%; mean ± SEM), was comparable after P2Y₂ KO ADSC transplantation (22.3 ± 1.0%; mean ± SEM). We observed comparable collagen blue staining for the fibrotic area of ischemic hearts injected with P2Y₂ KO cADSC and PBS-injected ischemic hearts (Figure 1A, ×20 magnification). Collagen stainings are shown inside the fibrotic and remote areas of ischemic hearts injected with PBS, WT cADSCs or P2Y₂ KO cADSC (Figure 1A, ×80 magnification).

The inhibitory effect of undifferentiated cADSC transplantation on post-ischemic cardiac fibrosis is thus dependent on their P2Y₂ receptor expression. Reduced cardiac fibrosis was observed without any required pretreatment of cADSC. The release of endogenous extracellular nucleotides currently observed during ischemia could thus regulate the cardioprotective action of transplanted cADSC through activation of P2Y₂ receptors expressed at their surface.

RNA-sequencing experiments were assessed on undifferentiated cADSC cells, unstimulated or treated daily with UTP (100 µM) during 7 days of culture in proliferation medium. RNA samples were extracted from two different cADSC cultures per condition and pooled for RNA-sequencing experiments. Among the 1,359 differentially regulated genes, 598 genes were upregulated by UTP (fold change > 2) and 761 were downregulated by UTP (fold change <0.5). MA-plot representation provides an overview of regulated-gene distribution within two-sample comparison (Figure 2A). We observed inter alia upregulation of Ccl5, Ccl22, Cxcl5, and downregulation of Ccl12 chemokine genes by UTP in ADSC (Figure 2A). A gene ontology (GO) enrichment analysis with DAVID software on the RNA-sequencing data obtained from UTP-treated ADSC versus untreated ADSC revealed that several biological processes linked to innate immune response, inflammation, as well as signal/receptor transduction and transmembrane/ion transport were enriched (Figure 2B). The modified Fisher Exact p-value or EASE score was reported by the software and indicated as logarithmic p-value.

We have previously demonstrated differential involvement of both mouse P2Y₂ and P2Y₄ receptors for UTP and ATP, in cADSC adipogenic differentiation thanks to P2Y₂ and P2Y₄ knockout mice (Lemaire et al., 2017; Negri et al., 2019). The absence of a clear detection of membrane-bound P2Y₂ receptors remains a limitation. Our use of mouse monoclonal anti-P2Y₂ antibody (H-5) from Santa Cruz Biotechnology (sc-518121) at 1:500 dilution was previously unsuccessful in mouse cADSC (data not shown). A comparable weak staining of cADSC was obtained with phycoerythrin-conjugated m-IgGκ secondary antibody alone (sc-516141; Santa Cruz Biotechnology) (data not shown).

We decided to compare UTP-regulated genes in WT and P2Y₂ KO cADSC to identify more precisely genes and associated-biological processes specific of P2Y₂ activation in these cells, and potentially linked to our cardiac fibrosis data using WT and P2Y₂ KO ADSC (Figure 1). UTP is used as a stimulating agent instead of ATP to avoid regulation of genes related to ADP or adenosine receptor activation after ATP degradation. We have not detected expression of P2Y₆ UDP receptor in cADSC (data not shown). A UTP-treatment during 7 days with UTP would thus activate only P2Y₂ and P2Y₄ receptors in cADSC.

RNA-sequencing experiments were performed using P2Y₂ KO cADSC, unstimulated or treated daily with UTP (100 µM). UTP gene expression profile was compared in both WT and P2Y₂ KO cADSC, only genes with CPM >0.5 and a fold change >2 or <0.5 being considered. Among the 1,359 regulated genes identified in WT cADSC (Figure 2A), we observed that 977 genes were no more regulated in P2Y₂ KO cADSC whereas 382 genes were still regulated by UTP in P2Y₂ KO cADSC. A detailed GO enrichment analysis was made for these 977 UTP target genes only identified in WT ADSC and not in P2Y₂ KO ADSC (Figure 3A). Besides the important number of genes involved in immune and inflammatory responses, GO analysis revealed enriched biological processes linked to multicellular organism development, cell differentiation and adhesion, angiogenesis and response to hypoxia (Figure 3A). Table 2 shows UTP target genes linked to processes moderately enriched but potentially related to our model of cardiac ischemia. Among the 977 UTP target genes, we identified genes linked to extracellular matrix organization and collagen catabolic process such as matrix metalloproteinases Mmp1b, Mmp9, and Mmp25 and collagen types Col10a1, Col17a1, and Col25a1, and genes related to macrophage chemotaxis such as Ccr7 chemokine receptor and colony-stimulating factor 1 receptor (Csf1r) (Table 2). We also identified UTP target genes involved in macrophage differentiation into pro-fibrotic foam cell such as Csf2, macrophage scavenger receptor Msr1 and apolipoprotein B (ApoB), as well as genes linked to response to hypoxia such as Agtr1b, purinergic receptor P2X₂, adenosine A1 receptors and uncoupling protein Ucp3 (Table 2). Interestingly, we identified a significant number of UTP target genes linked to angiogenesis regulation such as Wnt7a, Ramp3, Mmp9, Nos3, and Adam8 (Table 2).

Among all the UTP-regulated genes, we studied further the up-regulation of three pro-angiogenic genes, Ccl5, Cxcl5, and Saa3, and the down-regulation of the pro-fibrotic chemokine Ccl12. We confirmed up-regulation of Ccl5, Cxcl5, and Saa3, and downregulation of Ccl12 mRNAs in response to UTP in cADSC by qPCR experiments (Figure 3B). CCL5, CXCL5, CCL12, and SAA3 protein levels were quantified by ELISA in the supernatants of UTP-treated or untreated undifferentiated cADSC. We observed that UTP treatment increased the release of CCL5, CXCL5, and SAA3, and decreased the release of CCL12 by cADSC (Figure 3C).

SAA3 is a key inflammatory adipocyte-derived factor which is the murine ortholog of human serum amyloid A promoting angiogenesis in many diseases. (Lv et al., 2016; Tannock et al., 2018) Interestingly, CCL5 chemokine was previously reported as a major regulator of ADSC migration and post-ischemic reparative capacities. (Kauts et al., 2013; Kimura et al., 2014) UTP-mediated release of CCL5, CXCL5, and SAA3 by cADSC, as well as the consistent number of UTP-regulated genes involved in angiogenesis regulation (Table 2), led us to investigate the potential action of UTP-treated cADSC transplantation on cardiac revascularization.

WT and P2Y₂ KO cADSC were cultured without or with daily stimulation of UTP (100 µM) for 7 days in proliferation medium. Flow cytometry experiments revealed a comparable amount, between 92% and 95%, of CD90⁺ CD105⁺ CD44⁺ CD26⁻ cells in all the cell cultures (data not shown). We have performed qPCR experiments to quantify P2Y₂ and P2Y₄ mRNA in wild type cADSC, after stimulation or not with 100 µM during 7 days (Figure 4A). We observed a high level of expression of P2Y₂ mRNAs compared to weakly detected P2Y₄ receptor mRNAs, in cADSC after 7 days of culture in the proliferation medium (Figure 4A). We also observed that sustained stimulation with 100 µM UTP during 7 days did not induce down-regulation of P2Y₂ receptor mRNAs (Figure 4A).

Intramyocardial injections of PBS (control), untreated WT or P2Y₂ KO cADSC, and UTP-treated WT or P2Y₂ KO cADSC were performed in ischemic mice, 3 days after LAD ligation. Immunofluorescent stainings using anti-isolectin B4 and anti-α-SMA antibodies were performed on heart sections, 14 days after cASC injection, to quantify capillary and arteriole density respectively, in infarct border zone at different levels of the ischemic heart (Figure 4B). Analysis of peri-infarct microvasculature 14 days after cADSC injection revealed a higher capillary density in heart transplanted with untreated cADSC compared to PBS-injected mice (increase of 15.2 ± 2.8%; mean ± SEM; ***: p < 0.001) (Figures 4C,D). Capillary density was even more enhanced in heart transplanted with UTP-treated cADSC compared to PBS-injected mice (increase of 27.2 ± 4.2%; mean ± SEM; ***: p < 0.001) (Figure 4D). Interestingly, capillary density was similar in P2Y₂ KO cADSC-transplanted mice and in PBS-injected mice (Figure 4D). Moreover, the effect of UTP treatment on the capacity of cADSC to increase capillary density, was not observed using UTP-treated P2Y₂ KO cADSC (Figure 4D). No significant effect of UTP-treated or untreated ADSC transplantation was observed on arteriole density, defined by the number of α-SMA positive vessels with a diameter <20 µm (Figures 4E,F).