Exosomal hsa_circ_0007047 Attenuates Pos-myocardial Infarction Remodeling by Promoting Angiogenesis via miR-1178-3p/PDPK1 Axis

Emerging studies indicate that exosomes and their inner noncoding RNAs, especially circular RNAs (circRNAs), play key roles in gene regulatory network and cardiovascular repair. However, our understanding of exosomal circRNAs on cardiac remodeling after myocardial infarction (MI) remains limited. In the present study, exosomes were harvested from the serum of patients with and without postinfarction cardiac remodeling. The results showed that the level of hsa_circ_0007047 was signicantly downregulated in serum exosome of patients with the adverse cardiac remodeling when compared with those without post-MI remodeling or normal subjects. Loss-of-function approaches in vitro established that exosomal hsa_circ_0007047 robustly promoted angiogenesis and stimulated of cultured human vascular smooth muscle cells proliferation and migration. Accordingly, overexpression of exosomal hsa_circ_0007047 in mice signicantly attenuated MI-induced myocardial brosis and left ventricular dysfunction, accompanied by a larger functional capillary network at the border zone. Further exploration of the downstream target gene indicated that hsa_circ_0007047 acts as a competing endogenous RNA by directly binding to miR-1178-3p and thereby inducing transcription of its target gene phosphoinositide-dependent kinase-1 (PDPK1), a critical positive regulatory factor of angiogenesis. Together, our results revealed that exosomal hsa_circ_0007047 attenuated detrimental post-MI remodeling via miR-1178-3p/PDPK1 axis, which facilitated revascularization, ultimately improved the cardiac function.


Introduction
Despite current pharmacological and technological treatment strategies, progression to heart failure occurs in up to one-third of myocardial infarction (MI) patients because of adverse cardiac remodeling 1 .
Pathological left ventricular (LV) remodeling after MI is one of the most important risk factors for the development of complicated ventricular arrhythmia, congestive heart failure and even ultimate cardiogenic death 2 . On the other hand, it has been showed that ampli ed myocardial angiogenesis, facilitating neovascularization is mainly responsible for structural preservation of infarcted myocardium and maintaining its basic function 3 . Therefore, promotion of myocardial angiogenesis has been proposed as one of e cient therapeutic approaches for cardiac dysfunction and deleterious remodeling after MI.
Meaningfully, recent studies aimed at this phenomenon explored whether exosomes, the membranous vesicular bodies typically 30-100 nm in diameter, might be also involved in its regulation 4 . Exosomes contain multiple proteins, mRNAs and noncoding RNAs, contributing to the cellular communication and regulation of the multiple processes. The number and content of heterogenous exosomes not only affect the physiological state of different normal cells, but they could be also linked to many pathological pathways 5 . Importantly, it has been already demonstrated that exosomes play crucial roles in the development of cardiovascular diseases, speci cally, in the regulation of the post-MI cardiac remodeling 6 . Thus, it has been shown that the CD4-activated exosomes promote the post-ischemic cardiac brosis, through miR-142-3p/Wnt signaling cascade-mediated activation of interstitial myo broblasts 7 . This nding suggested that pharmacologic targeting of miR-142-3p in CD4-activated exosomes may hold promise for alleviation of the post-MI cardiac remodeling. Moreover, application of exosomes isolated from the plasma of ischemia-conditioned rats improved cardiac function and angiogenesis after MI through targeting the 70-kDa heat shock protein 8 . Song et al., also showed that application of exosomes derived from the umbilical blood mesenchymal stem cells attenuate myocardial injury by inhibiting ferroptosis in mice with experimental MI 9 . In addition, the use of certain exosome-derived noncoding RNAs, such as the miRNAs and lncRNAs has been proven bene cial for healing of human post-MI myocardium. For example, exosomal lncRNA AK139128 derived from hypoxic cardiomyocytes promoted apoptosis and inhibited proliferation of cardiac broblasts 10 . LncRNA KLF3-AS1 isolated from human mesenchymal stem cells-derived exosomes ameliorates pyroptosis of cardiomyocytes and alleviated the outcome of MI through the induction of the miR-138-5p/sirtuin1 axis 11 . Moreover, exosomes isolated from coronary serum of patients with MI promote myocardial angiogenesis through the miRNA-143/insulin-like growth factor-I receptor pathway 12 . Together, the above-mentioned results indicate that non-coding RNAs may indeed play crucial roles in the regulation of different processes contributing to the nal clinical outcome of patients a icted with MI.
More precisely, the circular non-coding RNAs (circRNAs), which are widely present in eukaryotic cells and participate in the regulation of the transcription and post-transcriptional expression of multiple genes contributing to normal cardiac functions, might also regulate the development of certain cardiac pathologies. Meaningfully, it has been proposed that certain changes of circRNAs levels might be considered as a potential biomarker of the therapeutic e ciency of cardiovascular diseases 13 . Indeed, circRNA HIPK3 was demonstrated to contribute to cardiac regeneration after experimental MI in mice by binding to Notch1 and miR-133a 14 . Elevation of circRNA 010567 also associated with the improvement of cardiac function and alleviation the myocardial brosis of MI rats through inhibiting transforming growth factor β1 15 . Moreover, circRNA Ttc3 in cardiomyocytes counteracted the hypoxia-induced ATP depletion and the deterioration of cardiac dysfunction, by sponging miR-15b in a rat model of MI 16 .
Exsomal circRNA 0001273 derived from human umbilical cord mesenchymal stem cells remarkably inhibited the occurrence of myocardial cell apoptosis and subsequently promoted MI repair in ischemic environment 17 . However, other putative regulatory mechanism in which exosome-derived circRNAs would contribute the postinfarct cardiac remodeling, remain to be better explored.
To provide a basis for the further study of the molecular pathogenesis of LV remodeling after MI and to identify potential therapeutic targets for this disease, we compared the expression pro les of exosomal circRNAs in patients with and without cardiac remodeling using a high-throughput RNA sequencing. First, we tested function of exosomal hsa_circ_0007047, which has already been related to LV remodeling, in vitro and in vivo using silencing and overexpression strategies. Then, we explored the mechanisms underlying the actions of exosomal hsa_circ_0007047 during the LV remodeling after MI. We anticipated that the obtained results would improve our understanding of exosomal circRNAs regulation during development of cardiovascular diseases.

Results
Exosome isolated from serum of patients with post-MI cardiac remodeling contains the signi cantly downregulated levels of hsa_circ_0007047 Our initial whole transcriptome sequencing analysis was aimed at the exploration of circRNAs expression in exosomes derived from sera of patients with and without postinfarct cardiac remodeling. The obtained results indicated that three differentially expressed exosomal circRNAs (hsa_circ_0000212, hsa_circ_0089282 and hsa_circ_0007047) could be detected in the comparison between postinfarct cardiac remodeling (CR) vs. control groups and also between CR vs. non-postinfarct cardiac remodeling (N-CR) group (Fig. 1A, Supplementary Table 4).
Importantly, we fund that while the expression of hsa_circ_0000212 and hsa_circ_0089282 was signi cantly higher, the expression of hsa_circ_0007047 was downregulated in both CR vs. N-CR group, and CR vs. N-CR groups ( Fig. 1B and C). In order to con rm these results, qRT-PCR was performed to detect the expression of these circRNAs by using divergent primers in the serum exosomes of patients with and without post-MI cardiac remodeling. The isolated exosomes were identi ed by TEM and NTA.
The results showed that the diameter of exosomes is between 100-150 nm, which presented the typical exosomal morphology (Fig. 1D). Then, we additionally found that the exosomal biomarkers, CD6, CD9 and TSG101 were also highly detected in all groups (Fig. 1E). Consistent with the transcriptome analysis, qRT-PCR results showed that the expression of hsa_circ_0000212 and hsa_circ_0089282 was signi cantly elevated while the expression of hsa_circ_0007047 was downregulated in the comparison between both the CR vs. N-CR groups and in CR vs. control groups (Fig. 1F).

Characteristics of hsa_circ_0007047
To learn whether hsa_circ_0007047 functions act as circRNA during the postinfarct cardiac remodeling, rst the ring structure of hsa_circ_0007047 was detected ( Fig. 2A). The speci c products of hsa_circ_0007047 were probed with the divergent and convergent primers. The agarose electrophoresis assay demonstrated that hsa_circ_0007047 could be ampli ed from both cDNA and gDNA template by using convergent primers while it only could be ampli ed from cDNA template by using divergent primers ( Fig. 2B). However, the GAPDH could only be ampli ed by the convergent primers, using cDNA and gDNA as the template.
Moreover, the reverse shear site of hsa_circ_0007047 was con rmed by Sanger sequencing of the ampli ed product of hsa_circ_0007047 (Fig. 2C). Finally, qRT-PCR was performed with cDNA as templates after the RNA was digested by RNA exonuclease R before reverse transcription, GAPDH as the negative control. The results showed that the expression of hsa_circ_0007047 was not changed in RNase R(+) group compared to the RNase R(-) group. Meaningfully, the expression of GAPDH was signi cantly downregulated in RNase R(+) group compared to the RNase R(-) group, indicating that hsa_circ_0007047 had a relatively stable structure (Fig. 2D).
The hsa_circ_0007047 promote the new angiogenesis, migration and proliferation of cultured VSMCs It has been previously demonstrated that angiogenesis plays crucial roles in the postinfarct cardiac remodeling 18 . Therefore, we now explored whether hsa_circ_0007047 would affect angiogenesis of CMECs and the proliferation and migration of VSMCs. The gain or loss function mutations of hsa_circ_0007047 were achieved by co-incubation with the exosomes harboring with overexpression (Over-) or suppression (Sh-) vector of hsa_circ_0007047. The qRT-PCR analysis showed that the expression of hsa_circ_0007047 was successfully overexpressed or suppressed respectively (Fig. 3A).
Meaningfully, we have noticed that the numbers of capillary-like structures were signi cantly decreased in Sh-hsa_circ_0007047-treated cultures and increased in the Over-hsa_circ_0007047-treated group, when compared to the NC group (Fig. 3B). Further analysis showed that the CMECs-formed tubes were signi cantly shorter in Sh-hsa_circ_0007047 group than that in the NC group, but elongated in the Over-hsa_circ_0007047 group. These results indicated that the hsa_circ_0007047 could promote angiogenesis of CMECs. Results of additional CCK-8 assay also demonstrated that the cell viability of VSMCs was signi cantly increased in the Over-hsa_circ_0007047 group, as compared to the NC group, while it was signi cantly suppressed in cultures treated with the Sh-hsa_circ_0007047 at 48 h and 72 h, respectively ( Fig. 3C). Results of cell cycle analysis showed that the number of VSMCs staged in the G1 phase was signi cantly decreased in the Over-hsa_circ_0007047-treated cultures, while the number of VSMCs progressed to the S and G2 phases was signi cantly increased in comparison with the NC group. Also, the treatment with Sh-hsa_circ_0007047 induced just opposite results when compared to the NC group (Fig. 3D). In addition, we found that while Over-hsa_circ_0007047 promoted VSMCs migration, while inhibiting the migration by Sh-hsa_circ_0007047 (Fig. 3E). Jointly, our novel results suggested that overexpression of hsa_circ_0007047 promoted the angiogenesis of CMECs, as well as increased the proliferation and migration of VSMCs.
hsa_circ_0007047 acted as a molecular sponge for miR-1178-3p We used the CircInteractome program to predict with which miRNA the hsa_circ_0007047 would interact. The results showed that there is a binding site of miR-1178-3p in the 3'UTR, suggesting that the expression of hsa_circ_0007047 might be regulated by miR-1178-3p (Fig. 4A). To con rm this assumption, dual luciferase reporter analysis was performed. The results showed that the relative luciferase activity was signi cantly reduced in the hsa_circ_0007047-WT-pmiGLO group compared to the pmiGLO group treated with miR-1178-3p mimics. However, no signi cant changes were found in the hsa_circ_0007047-Mut-pmiGLO group, when compared with the pmiGLO group (Fig. 4B). Further analysis showed that the expression of miR-1178-3p was signi cantly elevated in the exosomes isolated from serum of patients with postinfarct cardiac remodeling, when compared with non-postinfarct cardiac remodeling and control group (Fig. 4C). In addition, we have noticed that the expression of miR-1178-3p was signi cantly decreased in Over-hsa_circ_0007047 VSMCs while increased in the Sh-hsa_circ_0007047 (Fig. 4D). These results demonstrated that hsa_circ_0007047 is a direct target of miR-1178-3p.

miR-1178-3p inhibited the angiogenesis of CMECs and the proliferation and migration of VSMCs
Considering the demonstrated interaction between the hsa_circ_0007047 and the miR-1178-3p, we next con rmed the potential role of miR-1178-3p in the CMECs angiogenesis and the proliferation and migration of VSMCs by assessing the loss and gain of miR-1178-3p function mutations. Results of qRT-PCR analysis showed that the expression of miR-1178-3p was successfully overexpressed or suppressed ( Fig. 5A). Tube formation analysis showed that the number of capillary-like structures was signi cantly decreased in miR-1178-3p mimics group while increased in miR-1178-3p inhibitor group compared to the NC group. Further quanti cation analysis showed that the tube length of CMECs was signi cantly shorter in miR-1178-3p mimics group than that in the NC group, while it was longer in CMECs treated with miR-1178-3p inhibitor (Fig. 5B). The above results jointly indicated that miR-1178-3p could inhibit angiogenesis of CMECs.
Results of the CCK-8 assay demonstrated that the viability of VSMCs was signi cantly decreased in the miR-1178-3p mimics group, when compared to the NC group, while it was dramatically elevated in miR-1178-3p inhibitor group (Fig. 5C). Cell cycle analysis showed that the number of VSMCs entering the G1 phase was signi cantly increased in miR-1178-3p mimics-treated group, while the number of VSMCs entering the S and G2 phase was signi cantly decreased, as compared to the NC group. Meaningfully, just opposite results were detected in cultures of VSMCs treated with miR-1178-3p inhibitor (Fig. 5D).
Furthermore, we showed that overexpressing the miR-1178-3p in cultured VSMCs inhibited their migration. In contrast, suppressing of the miR-1178-3p expression promoted migration of cultured VSMCs (Fig. 5E). In addition, we found that the inhibitory the effects of miR-1178-3p on the angiogenesis of CMECs, and the proliferation and migration of VSMCs could be partially reversed by the coincubation of these cultured cells with the exosomes harboring the Over-expressed hsa_circ_0007047 or Sh-hsa_circ_0007047 plasmid in miR-1178-3p mimics or inhibitor groups, respectively ( Fig. 5B-E). These ndings suggested that overexpression of miR-1178-3p inhibited CMECs angiogenesis, and VSMCs proliferation and migration.

PDPK1 is identi es as the target of miR-1178-3p
To identify the downstream regulator of miR-1178-3p, three online databases, TargetScan7.2, miRwalk and miRDB, were used to predict its targets (Supplementary Table 5). As shown in Fig. 6A, we found three common targets from the three databases, including cholinergic receptor nicotinic beta 4 (CHRNB4), Stonin 2 (STON2) and PDPK1. The subsequent qPCR and western blotting demonstrated that the expression of PDPK1 was signi cantly downregulated while the expression of CHRNB4 and STON2 was upregulated in the serum of CR group compared with N-CR and control group ( Fig. 6B and C). The expression of PDPK1 was most downregulated among the three genes and was selected for further study. Furthermore, the consecutive results of the luciferase reporter assays, indicated that cotransfection of the PDPK1-WT pmiGLO reporter plasmids and miR-1178-3p mimic predominantly reduced the luciferase activity. Conversely, co-transfection of the PDPK1-Mut pmiGLO reporter plasmids with miR-1178-3p mimic showed no obvious effect on the luciferase activity (Fig. 6D). In addition, we found that miR-1178-3p overexpression reduced PDPK1 expression while miR-1178-3p suppression increased PDPK1 expression both at mRNA and protein levels ( Fig. 6E and 6F). These results indicated that PDPK1 is targeted by miR-1178-3p.
miR-1178-3p-mediated PDPK1 regulate CMECs angiogenesis and VSMCs proliferation and migration The next set of experiments was aimed at exploring how the loss or gain of PDPK1 expression would affect the functions of the cultured cells. Results of qRT-PCR and western blotting analysis showed that the PDPK1 could be successfully overexpressed or suppressed in cultured CMECs ( Fig. 7A and B). The consecutive tube formation analysis indicated that overexpression PDPK1 signi cantly increased the number of capillary-like structures and the tube length, while suppression PDPK1 greatly decreased the number of capillary-like structures and the CMECs tube length as compared to the control group (Fig. 7C).
The CCK-8 assay demonstrated that the cell viability of VSMCs was signi cantly elevated in the Over-PDPK1 group compared to the NC group, while it was dramatically decreased in Si-PDPK1 group at 48 h and 72 h incubations, respectively (Fig. 7D). Cell cycle analysis further indicated that the number of VSMCs being in the G1 phase was signi cantly decreased in the Over-PDPK1 group while the number of VSMCs that entered the S and G2 period was signi cantly increased as compared to the NC group. Meaningfully, just the opposite results were detected in the Si-PDPK1 group compared to the NC group ( Fig. 7E). Furthermore, we showed that the overexpression of PDPK1 could promote the migration of VSMCs, while the suppression of PDPK1 inhibited migration of VSMCs (Fig. 7F). In addition, we found that the effects of PDPK1 on the angiogenesis of CMECs, proliferation and migration of VSMCs could be reversed by adding miR-1178-3p mimics to the over-PDPK1 group, or by treatment of the si-PDPK1 group with the miR-1178-3p inhibitors ( Fig. 7C-F). Altogether, these results indicated that miR-1178-3p-induced PDPK1 regulated the angiogenesis of CMECs and the proliferation and migration of VSMCs.
hsa_circ_0007047 alleviated postinfarct cardiac remodeling in vivo To further con rm the bene cial effects of the exosomal hsa_circ_0007047 in the postinfarct ventricular remodeling, exosomes containing overexpressing or suppressing hsa_circ_0007047 plasmid were injected into the tail vein of mice after induction of their experimental MI. Importantly, we recorded that the injection of exosomes containing Over-hsa_circ_0007047 signi cantly decreased theirs LVEDD and LVESD, and improved LVEF and LVFS (P < 0.05). In contrast, the injection of exosomes containing Sh-hsa_circ_0007047 to the parallel MI mice induced the opposite effects (Fig. 8A). Even the initial comparison of the myocadiac sections from all experimental groups, stained with the H&E, clearly illustrated the exclusive bene cial effects of treatment with the Over-hsa_circ_0007047. Analysis of the parallel histologic sections, stained with the Masson's method or with the Sirius Red further con rmed a highly reduced degree of cardiac brosis in the myocardium of the Over-hsa_circ_0007047 Exo-treated group, as compared to the untreated MI group. In contrast, we have noticed that administration of the exosome of Sh-hsa_circ_0007047 aggravated the postinfarct ventricular remodeling (Fig. 8B).
Further histologic analysis also demonstrated that treatment with the Over-hsa_circ_0007047 exosomes led to a signi cant increase in capillary density as compared to the untreated MI group. In contrast the density of capillaries was signi cantly decreased in the Sh-hsa_circ_0007047 exosomes group (Fig. 8B).
Results of western blotting analysis and immuno uorescence staining assay of the parallel tissue samples additionally demonstrated that the expression of HIF-1α, VEGF and VEGFR were signi cantly elevated in the Over-hsa_circ_0007047 exosomes-treated group, while decreased in the Sh-hsa_circ_0007047 exosomes-treated group as compared with the NC group ( Fig. 8C & D). These results were also con rmed by immuno uorescence staining assays. These results demonstrated that hsa_circ_0007047 functions as an effective suppressor of the post-MI deleterious ventricular remodeling.
The next series of experiments was aimed to investigate whether hsa_circ_0007047 could affect the expression of miR-1178-3p and PDPK1, in heart tissues. Meaningfully, we found that administration of exosomes containing overexpressing hsa_circ_0007047 plasmid greatly decreased levels of miR-1178-3p and increased PDPK1 levels. However, treatment with the exosomes containing siRNA of hsa_circ_0007047 signi cantly increased the expression of miR-1178-3pm and decreased PDPK1 levels in vivo (Fig. 9A & B). The additional western blotting and immuno uorescence staining also con rmed the already mentioned expression changes of PDPK1 in different groups ( Fig. 9C & D). Therefore, these results demonstrated that hsa_circ_0007047 attenuated the post-MI ventricular remodeling via regulation of miR-1178-3p-PDPK1 axis in vivo.

Discussion
In the present study, exosomal hsa_circ_0007047 was identi ed as a novel regulator of the postinfarct ventricular remodeling. Function and mechanism analysis revealed that exosomal hsa_circ_0007047 alleviated adverse postinfarct cardiac remodeling by promoting angiogenesis via regulating miR-1178-3p/PDPK1 axis.
The cardiac remodeling after MI refers to a cascade of structural and functional changes in cardiomyocytes and intercellular substance. An extremely complex pathogenesis of this process includes myocardial hypertrophy, brosis, in ammation, autophagy and metabolic malfunction 19 . It has been recently demonstrated that exosomes are considered as the main mediators of intercellular communication in the myocardium, which may attenuate the development of detrimental structural changes and the consequent heart failure after MI 20 . Thus, the presence of exosomes likely contributes to inhibition of brosis, promotion of angiogenesis and alleviation of in ammation and pyroptosis, by translocation of intercellular molecules, proteins, and organelles. Particular attention has been paid recently to the exosomal circRNAs, after the discovery that the circular form of the non-coding RNA, such as circFndc3b, circRNA HIPK3 and circRNA CDR1, could contribute to bene cial regulation of post-MI ventricular remodeling 21,22 . Further studies exploring the mechanistic involvement of the diverse exosomes in the myocardial healing demonstrated that hypoxia-elicited mesenchymal stem cell derived exosomes facilitated cardiac repair through miR-125b-mediated amelioration of apoptosis and pyroptosis of cardiomyocytes within the infarct region of murine MI model 23 .
Results of other studies also indicated that angiogenesis plays a crucial role in ameliorating the adverse myocardial remodeling, thereby, contributing to the improvement of cardiac function and preventing the heart failure 24 . Silencing of angiogenesis inhibiting factor epiregulin disrupted the ERK1/2 signaling and promoted LV remodeling 25  On the other hand, numerous studies demonstrated that exosomes play a regulatory role in angiogenesis during cardiac remodeling after MI, by harboring certain non-coding RNAs or metabolic proteins 29 . Thus, exosomes derived from Akt-modi ed human umbilical cord mesenchymal stem cells improve cardiac regeneration and promote angiogenesis via activating platelet-derived growth factor D 30 . Exosomal miR-132 and miR-146a delivered by mesenchymal stem or adipose-derived stem cells attenuated the ischemic myocardial damage by inducing angiogenesis in MI, respectively 31,32 . Importantly, results of our present study revealed that the newly identi ed exosomal hsa_circ_0007047, which promoted angiogenesis, proliferation, and migration of vascular smooth muscle cells in vitro and could alleviate ventricular remodeling in murine MI model.
HIF-1α, VEGF and VEGFR have been previously indicated as the angiogenetic factors after MI 33,34 . We found that the expression of these factors in the murine myocardium subjected to experimental MI was signi cantly elevated after treatment with Over-hsa_circ_0007047 exosomes but decreased after treatment with Sh-hsa_circ_0007047 exosomes. Therefore, these results demonstrated that hsa_circ_0007047 functions as a potential suppressor of the adverse postinfarct cardiac remodeling by promoting angiogenesis.
In the present study, bioinformatics analysis and dual luciferase reporter system demonstrated that miR-1178-3p not only binds to hsa_circ_0007047 but also to PDPK1, the leucyl-speci c aminopeptidase activated factor, contributing to the VEGF-dependent activation of S6K, and the ultimate stimulation of endothelial cell proliferation and angiogenesis 35 . Interestingly, we have also established that the overexpression of miR-1178-3p inhibited angiogenesis, proliferation, and migration of VSMCs, suggesting that miR-1178-3p could inhibit the overzealous angiogenesis in cardiac remodeling after MI. On the other hand, it has been shown that the ablation of PDPK1 in cultured vascular endothelial cells enhanced insulin sensitivity and suppressed angiogenesis 36 . These results demonstrated that PDPK1 act as a positive regulator in angiogenesis. Consistent with these ndings, we revealed that overexpression of PDPK1 promoted angiogenesis, proliferation and migration of VSMCs. In addition, we found that the function of miR-1178-3p and PDPK1 on the angiogenesis, proliferation and migration of VSMCs could be reversed by hsa_circ_0007047 and miR-1178-3p, respectively. Altogether, these results indicated that hsa_circ_0007047 alleviated postinfarct ventricular remodeling by promoting angiogenesis via miR-1178-3p-mediated PDPK1 expression.
In this present study, we identi ed a novel exosomal hsa_circ_0007047, which was profoundly downregulated in patients with detrimental ventricular remodeling after MI. Meaningfully, we have further established that exosomal hsa_circ_0007047 alleviated postinfarct remodeling by promoting angiogenesis and expanding collateral network via regulating miR-1178-3p/PDPK1 axis, that consequently preserved function of the border zone of infarcted myocardium. These ndings suggest that overexpressing or application of exosomal hsa_circ_0007047 should be further investigated as a putative therapeutic target in ischemic heart failure patients.

Clinical specimens and ethical statement
The clinical serum specimens were obtained from MI patients with (n = 10) or without (n = 10) cardiac remodeling at Guangdong Provincial Hospital of Chinese Medicine following the inclusion and exclusion criteria (Supplementary Table 1). The group of healthy individuals (n = 10) respectively matching with our patients, served as an additional control. All participants of our study have signed informed consent, and the entire project has been approved by the ethics committee of Guangdong Provincial Hospital of Chinese Medicine (B2015-129-01) and conducted in accordance with the Declaration of Helsinki and its text revisions. The clinical parameters of each group were summarized in Supplementary Table 2.

Isolation and identi cation of human exosomes
Exosomes were isolated from human serum by using ExoQuick exosome precipitation solution (SBI, CA, USA) following the user manual. Brie y, serum was centrifuged at 3000 g for 15 min and then add the 63 µL ExoQuick exosome precipitation solution and refrigerate the mixture 30 min at 4°C. And then the exosome was resuspended by 100 µL of sterile PBS after centrifugation at 1500 g for 30 min. Particle size, morphology and the total amount of the exosomes were then identi ed by transmission electron microscope (TEM) (Philips TECNAI 20, Netherland) and by the nanoparticle tracking analysis (NTA), respectively. Exosome protein markers CD6, CD9 and tumor susceptibility gene 101 (TSG101) were identi ed by western blot assay.

Cell cultures and transfection of human VSMCs and CMECs
Human vascular smooth muscle cells (VSMCs) and cardiac microvascular endothelial cells (CMECs) were purchased from the American Type Culture Collection (ATCC, Bethesda, MD, USA). VSMCs were cultured in the 90% Dulbecco's modi ed Eagle's medium (DMEM, 12430054, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA), 100ug/mL penicillin and 100ug/mL streptomycin in a 5% CO 2 -contained incubator under 95% saturation humidity at 37℃. For the function analysis of hsa_circ_0007047, the overexpressing or suppressing plasmids of hsa_circ_0007047 were co-incubated with the serum exosomes and then co-incubated with VSMCs. For further analysis of the phosphoinositide-dependent kinase-1 (PDPK1) or miR-1178-3p effects, the (60-80%) con uent cell cultures were transfected with the overexpressing or suppressing of pCDNA3.1-PDPK1 plasmids, and with the miR-1178-3p mimics or its inhibitor, using the Lipofectamine™ 2000 transfection reagents (52887; Invitrogen, USA). The quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and western blots were additionally used to detect the overexpression or silencing e ciency.
Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) Relative expression levels of targeted genes were calculated using the 2 −△△Ct methods, which was normalized to GAPDH (for PDPK1 and hsa_circ_0007047) and to U6 (for miR-1178-3p). The nuclear and cytoplasmic RNA were extracted using nuclear and cytoplasmic RNA puri cation kit (Fisher scienti c). The samples of DNA or cDNA were ampli ed by divergent or convergent primers of hsa_circ_0007047, respectively. The ampli cation process were run as follows: 95°C for 5 min, followed by 32 cycles at 94°C for 15 s, 60°C for 30 min, and 72°C for 30s. The nal products were then detected by 1% Agarose gel electrophoresis. All primers used in the present study were synthesized by General Biol company (General Biol Co., Ltd, Anhui, China), and the detailed information is listed in Supplementary Table 3.

Western blotting
Total protein was extracted from the serum or heart tissues using RIPA lysis buffer (P0013, Beyotime, Shanghai, China) according to the manufacturer's instructions as described previously 37

Cell migration assays
VSMCs from different groups were harvested after culturing in serum-free medium for 16 h, resuspended in serum-free medium and transferred to two-layered Transwell chambers pre-coated with Matrigel (BD Biosciences) and incubated at 37°C for the following 24 h. Both membranes were then xed with 100% methanol and stained with 1% toluidine blue. The numbers of cells attached to each membrane were counted under a light microscope (Zeiss710, Germany).

Tube formation assay
The growth factor reduced Matrigel (BD Biosciences) was thawed on ice and the 300 µL samples of this preparation were plated into 24-well plates and incubated for 30 min at 37°C to allow polymerization.
CMECs were suspended in 0.2% endothelial growth basal medium (EBM), and 5×10 4 cells of CMECs were added to Matrigel-coated wells. To assess the potential in uence of hsa_circ_0007047, miR-1178-3p and PDPK1 in VSMCs, cultures of CMECs were co-incubated with exosomes contained overexpressing or silencing hsa_circ_0007047, with miR-1178-3p mimics/inhibitors or overexpressing/silencing the PDPK1 plasmids. Then, the in uence of these diverse preparations on the initiation of cellular tubes formation was monitored for 12 h at 37°C under a phase contrast microscope (×4) (Nikon TS100). Tube lengths were quanti ed using the Image J software (National Institutes of Health).

Luciferase reporter assay
For hsa_circ_0007047, the miRNA targets of hsa_circ_0007047 was predicted by CircInteractome. For the downstream regulator of miR-1178-3p, online databases TargetScan7.2, miRwalk and miRDB were used to predict the targets of miR-1178-3p. And then the fragment of hsa_circ_0007047 or PDPK1 were cloned into the pmiGLO vector containing the wild type sequence and mutant binding sequence, respectively. The hsa_circ_0007047 or PDPK1 pmiGLO vector and miR-1178-3p mimcis were co-transfected into the VSMCs when the con uence reached at 60-70% by using the Lipofectamine™ 2000 transfection reagents (52887, Invitrogen, USA). Cells were washed twice with phosphate buffered saline and lysed using the passive lysis buffer after cultured for 48 h after transfection. The luciferase activity was evaluated using the Dual-Luciferase Reporter Assay System (Progema). The primers used for vector construction were listed in Supplementary Table 3.

Model of experimental MI in mice
Wild-type male C57BL/6J mice (

Echocardiography
Echocardiographic assessment was performed as previously described assess the relevant heart's actions 39 . Brie y, the LV function in C57BL/6J mice was conducted just before sacri ce using the Acuson Sequoia C512 system equipped with a 15L8 linear array transducer with 30 MHz. Mice were anesthetized with 1.5% iso urane mixed with oxygen and placed in a supine position on a heating pad.
Short-axis measurements were used to capture M-mode tracing at the level of the papillary muscles with a 25-mm signal depth. Three to six consecutive cardiac cycles were measured using M-mode tracings with the accompanying software.

Histopathological changes analysis
Hematoxylin and eosin (H&E), Masson and Sirius red staining were performed to evaluate the histopathological changes in vivo as described previously 40 . Brie y, heart tissue was sliced into 8-µmthickness sections and xed with 4% paraformaldehyde at room temperature. The tissue sections were

Statistical analysis
Data are presented as the mean ± standard deviation (SD) from at least three independent experiments performed in triplicate. Analyses were performed using Prism 8.1.2 (GraphPad Software Inc.) by unpaired t-test and one-way analysis of variance (ANOVA) between two groups and more than two groups, respectively. P values < 0.05 was de ned as the level of statistical signi cance.  Characteristics of hsa_circ_0007047. A, Schematic diagram of the ring structure of hsa_circ_0007047; B, Agarose electrophoresis assay to ampli ed the speci c products of hsa_circ_0007047 from both cDNA and gDNA template by using convergent and divergent primers, GAPDH served as control; C, The reverse shear site of hsa_circ_0007047 was con rmed by Sanger sequencing of the ampli ed product of hsa_circ_0007047; D, qRT-PCR was used to detect the expression of hsa_circ_0007047 by using cDNA as templates with the RNA was digested by RNA exonuclease R before reverse transcription, GAPDH served as the negative control. Data are mean±SD. **P <0.01, RNase R+ group vs RNase R-group.      The expression changes of miR-1178-3p and PDPK1 in vivo treatment with exosomes harboring hsa_circ_0007047. A, The expression of miR-1178-3p was detected by qRT-PCR, U6 act as control; B and C, The expression of PDPK1 was detected by qRT-PCR and western blotting, respectively, GAPDH served as control; D, Immuno uorescence staining was used to detect the expression of PDPK1 in heart tissues.