Engineered model of heart tissue repair for exploring fibrotic processes and therapeutic interventions

Yang, Pengcheng; Zhu, Lihang; Wang, Shiya; Gong, Jixing; Selvaraj, Jonathan Nimal; Ye, Lincai; Chen, Hanxiao; Zhang, Yaoyao; Wang, Gongxin; Song, Wanjun; Li, Zilong; Cai, Lin; Zhang, Hao; Zhang, Donghui

doi:10.1038/s41467-024-52221-9

Engineered model of heart tissue repair for exploring fibrotic processes and therapeutic interventions

Article
Open access
Published: 12 September 2024

Volume 15, article number 7996, (2024)
Cite this article

Download PDF

You have full access to this open access article

From

View current issue

Engineered model of heart tissue repair for exploring fibrotic processes and therapeutic interventions

Download PDF

Abstract

Advancements in human-engineered heart tissue have enhanced the understanding of cardiac cellular alteration. Nevertheless, a human model simulating pathological remodeling following myocardial infarction for therapeutic development remains essential. Here we develop an engineered model of myocardial repair that replicates the phased remodeling process, including hypoxic stress, fibrosis, and electrophysiological dysfunction. Transcriptomic analysis identifies nine critical signaling pathways related to cellular fate transitions, leading to the evaluation of seventeen modulators for their therapeutic potential in a mini-repair model. A scoring system quantitatively evaluates the restoration of abnormal electrophysiology, demonstrating that the phased combination of TGFβ inhibitor SB431542, Rho kinase inhibitor Y27632, and WNT activator CHIR99021 yields enhanced functional restoration compared to single factor treatments in both engineered and mouse myocardial infarction model. This engineered heart tissue repair model effectively captures the phased remodeling following myocardial infarction, providing a crucial platform for discovering therapeutic targets for ischemic heart disease.

Heart regeneration and repair after myocardial infarction: translational opportunities for novel therapeutics

Article 21 July 2017

Mechanisms Underlying Cardiomyocyte Development: Can We Exploit Them to Regenerate the Heart?

Article 03 June 2021

Pathophysiology and therapeutic potential of cardiac fibrosis

Article Open access 17 July 2017

Introduction

Ischemic heart disease (IHD) has affected nearly 197.2 million people globally over the past three decades, resulting in 9.1 million deaths¹. IHD mortality arises from arrhythmia or heart failure due to irreversible cardiac muscle damage leading to pathological remodeling². Following acute myocardial infarction (MI), the overall progression from myocardial injury to complete remodeling can be divided into three phases: death of stressed cardiomyocytes (CM), inflammatory response, and fibrotic scar formation, ultimately impairing myocardial contractility and electrophysiology³. Specifically, proteins released by apoptotic CM⁴ trigger inflammation, prompting resident cardiac fibroblasts (CF) to differentiate into myofibroblasts (myoCF), which enhances extracellular matrix (ECM) synthesis and collagen deposition in the myocardium scar. Consequently, the injured myocardium is replaced by the dense scar composed of ECM and CF, while surviving CM fails to achieve consistent electromechanical coupling, increasing reentry risk⁵. The fibrotic scar’s reduced amplitude and slow upstrokes of the action potential cause conduction blocks, with alternative pathways for electrical impulses involved in the re-entry contributing to arrhythmia⁶.

Human pluripotent stem cell-derived cardiac lineage cells, including CM⁷, CF⁸, and endothelial cells⁹, offer unlimited functional cell sources for fabricating human-engineered heart tissue (hEHT) that adequately mimic the structural and electrophysiological characteristics of adult myocardium^10,11. Spherical heart organoids can serve as 3D myocardial tissue to mimic cardiac development and disease^12,13. In this respect, Voges et al. found that cryoinjured heart organoids can regain contractility during healing¹⁴. Furthermore, Richards et al. developed a heart organoid model of MI by leveraging an oxygen-diffusion gradient, which emulated post-MI pathological features like fibrosis and metabolic shifts¹⁵. Despite advancements in the functionality of engineered myocardium and high-throughput drug testing platforms using cardiac organoids, there remains a need for the in vitro model that accurately represents myocardial injury, fibrosis, and electrophysiological dysfunction and allows for quantitative prediction of arrhythmia risk.

Here, we develop a hEHT injury repair model to simulate MI onset, progression, and recovery, focusing on myocardial remodeling. The remodeling process includes an early injury period (days 0–2) characterized by CM apoptosis and structural injury, followed by the fibrotic scar-forming period featuring CF activation (days 2–4). Electrophysiological dysfunction in the remodeled area is observed by day 6, indicated by reduced calcium wave conduction and activity. Transcriptomic analysis through bulk RNA sequencing (bulkRNA-seq) and single-cell RNA sequencing (scRNA-seq) revealed that cAMP, TGFβ1, and ROCK regulated the early injury-induced CM stress and CF mobilization, while the late period involves CM action potential modulation, CF proliferation, and ECM accumulation driven by WNT, PI3K, and NFκB. Also, scRNA-seq underscores the potential transitions of CF during myocardial remodeling. A mini hEHT injury repair model in a 96-well array is further developed to screen seventeen modulators from nine signaling pathways identified from transcriptomic analysis to assess their effects on myocardial remodeling. SB431542 (TGFβ inhibitor), Y27632 (ROCK inhibitor), and CHIR99021 (WNT activator) can prevent excessive fibrosis in the remodeled area. An integrated scoring system encompassing eight electrophysiological parameters evaluates the impact of SB431542, Y27632, and CHIR99021 on electrophysiological dysfunction recovery in the hEHT injury repair model. While Y27632, CHIR99021, and SB431542 partially restore the abnormal calcium activity, only their phased combination effectively rectifies the global electrophysiological dysfunction. The mouse MI model further confirms this phased combination’s superior efficacy in reducing fibrotic area and enhancing cardiac function post-MI. In conclusion, our hEHT repair model can mimic the phased myocardial remodeling and elucidate post-MI electrophysiological dysfunction, enabling rapid screening for potential anti-fibrotic and cardiac regeneration drug targets for the IHD treatment.

Results

Establishment of the hEHT injury repair model in vitro

Dissociated hPSC-derived cardiomyocytes (hPSC-CM) were encapsulated in fibrin-Matrigel hydrogel to form 7 × 7 mm hEHTs¹¹. The hEHTs spontaneously and rhythmically contracted after 3 days of culture (Supplementary Movie 1). Owing to its reproducibility, cryoinjury has been used to model MI in mice¹⁶, pigs¹⁷, and human cardiac organoids^13,14. To establish a standardized humanized myocardial injury repair model using hEHT (Fig. 1a), we designed and tested cryoinjury devices A, B, and C, differentiated by conical tip sizes (Supplementary Fig. 1a). These devices were used with a polydimethylsiloxane (PDMS) pedestal (Supplementary Fig. 1b) to assist cryoinjury device in localization and penetration of hEHT (Supplementary Fig. 1c). The tips were pre-cooled in liquid nitrogen to induce four cryoinjured areas per hEHT (Supplementary Fig. 1d). Blunter tips from devices A and B resulted in larger, more heterogeneity injury area (Supplementary Fig. 1e, f), occasionally leading to irreversible hEHT rupture (Supplementary Fig. 1g). Conversely, Device C, with its sharpest tip, produced a more consistent initial cryoinjured area (Supplementary Fig. 1e, f) without affecting hEHT spontaneous beating (Supplementary Movie 2). The initial cryoinjured area enlarged on day 1, increasing to the maximum on day 2. Between days 2 and 6, the enlarged cryoinjured area progressively remodeled (Fig. 1b, c), reflecting a phased remodeling of cryoinjured hEHT.

**Fig. 1: Establishment of hEHT injury repair model.**

Regional cell apoptosis and reactive oxygen species production are the critical initial events in myocardial remodeling post-MI¹⁸. We observed a significant increase in TUNEL⁺ cells (Supplementary Fig. 2a, b) and ROS production (Supplementary Fig. 2c, d) around the cryoinjured area on day 2. Destruction of the CM structure was observed by cTNT staining, reflecting the stress status of cryoinjured hEHT in the first two days post-injury (Supplementary Fig. 2e). Notably, only the cryoinjury but not puncture can induce dramatic cell apoptosis (Supplementary Fig. 2f, g) and ROS accumulation (Supplementary Fig. 2h, i) around the injured area. During remodeling, the TUNEL⁺ cells in the remodeled area significantly declined to normal levels by day 4 (Fig. 1d, e). The presence of CM (SAA⁺, MYH⁺), CF (VIM⁺ and SM22α⁺), and myoCF (αSMA⁺) was observed in normal hEHT¹¹ (Supplementary Fig. 2j). As known that αSMA⁺ myoCF play a significant role in myocardial remodeling, a substantial increase in αSMA expression localized in the stress fiber was found on days 2 and 4 of the remodeling (Fig. 1d, f). Hence, the remodeling of the cryoinjured hEHT can be divided into an early injury expansion phase (days 0–2) with cellular stress followed by CF activation phase (days 2–4). By day 6, the fibrotic scar in the remolded area was observed (Fig. 1g), resembling the early stress and late fibrosis post-MI¹⁹.

To visualize the remodeling in the hEHT injury repair model, we performed time-lapse imaging of hEHTs generated from hPSC-CM, while marking the whole population with mCherry lentivirus and marking CM with TNNT2-GCaMP5, such that non-CM fluoresce red and CM fluoresce green and red (Supplementary Movie 3). A spiral migration of non-CM along the edge of the cryoinjured area to refill the injured area was observed. Post-MI, CF differentiates into myoCF, followed by enhanced ECM synthesis with increased collagen I and III deposition in the fibrotic scar²⁰. As observed in our model, the remodeled areas were similarly characterized by the differential number of CF (VIM⁺ and SM22a⁺), with an increase in type I collagen and more fibronectin at day 6, along with fewer CM (cTNT⁺) (Fig. 1g, h). However, this fibrotic scar had no disruptions in the overall spontaneous beating of hEHT (Supplementary Movie 2). The above results demonstrated that our model could mimic the post-MI events, highlighting the major role of CF, which migrates from the edge of the injured area to the central injured region and secreted ECM to refill the injured area in the late scar-forming period²¹.

hEHT injury repair model exhibited arrhythmic-related electrophysiological malfunction post-fibrotic scar formation

MI frequently precipitates arrhythmias by impairing CM function and inducing fibrosis, which impedes cardiac electrical signal propagation²². Utilizing cardiac optical mapping with calcium-sensitive dye, we performed optical mapping on remodeled hEHTs (Fig. 2a and Supplementary Movie 4) to investigate tissue-level calcium wave propagation abnormalities around the remodeled areas on day 6 post-injury (Fig. 2b). Compared to uninjured areas, the remodeled area exhibited decreased calcium wave conduction velocity and increased directional propagation inhomogeneity owing to the fibrotic scar (Fig. 2c, d). In addition, calcium transients within the remodeled area had significantly lower amplitude, maximum upstroke velocity, and maximum recovery velocity, reflecting the diminished calcium signal in the remodeled area (Fig. 2e–g). Prolonged 50% and 80% of calcium signal duration (CSD50 and CAD80), and extended depolarization and repolarization activation times in the remodeled area were also noted (Fig. 2h). Our findings revealed that increased refilling of CF than CM in the remodeled area created an abnormal calcium activity, leading to a conduction block. These areas of slow conduction or conduction block will increase the risk of triggering reentrant arrhythmias in the whole hEHT (Supplementary Movie 5)²³. Corresponding studies in rat model of MI exhibited similar electrophysiological dysfunctions (Supplementary Fig. 3a). Specifically, the decline in calcium signal propagation velocity and directional homogeneity within the infarct area were observed (Supplementary Fig. 3b, c). The malfunction reflecting the changes in calcium signal waveform characteristics in the MI rats was similar to the remodeled hEHT (Supplementary Fig. 3d–f), emphasizing that the hEHT repair model can exhibit the process of injury onset, fibrosis development, and regionalized malfunction together to mimic the spatiotemporal landscapes of MI.

**Fig. 2: Regional calcium activity abnormalities in remodeled hEHT.**

hEHT injury repair model exhibited differential gene expression between the early and late periods of remodeling

We conducted bulkRNA-seq to further characterize the remodeling of injured hEHT by comparing cryoinjured (Cryo) and control (Con) hEHTs during the injury expansion (day 2, “Early”) and scar-forming (day 4, “Late”) periods (Supplementary Fig. 4a). Principal component analysis (PCA) revealed significant gene expression difference between these groups (Supplementary Fig. 4b). The comparison between Cryo and Con identified 1147 differentially expressed genes (DEGs) at day 2 and 556 DEGs at day 4 (Supplementary Fig. 4c). Notably, 151 DEGs were common across early and later periods, primarily enriched in terms related to response to hypoxic response and apoptosis regulation (Supplementary Fig. 4d). This included significant upregulation of hypoxic stress-related genes NDRG1²⁴ and anti-apoptotic genes FAIM2²⁵, demonstrating a collaborative anti-apoptotic response over time (Supplementary Fig. 4d) and correlating with decreased apoptosis from early to late period (Fig. 1d–f).

In the early period, the top 15 enriched Biological Process Gene Ontology (GO) terms for DEGs were mainly associated with response to hypoxia response, angiogenesis, and cell differentiation (Supplementary Fig. 4e). Conversely, the DEGs in the late period focused on CM’s action potential, cardiac contraction, and cell proliferation regulation (Supplementary Fig. 4e). KEGG pathway analysis suggested that DEGs in early and late periods were enriched for distinct signaling pathways, with MAPK, cAMP, and HIF-1 signaling prevalent in the early period, and TNF, NFκB enriched in the late period (Supplementary Fig. 4f), demonstrated that even shared GO terms like ECM organization, apoptosis regulation, cell adhesion, and inflammatory response (Supplementary Fig. 4e) were modulated by different signals during the remodeling of injured hEHT. These findings underscore the notable phenotypic differences between early and late periods of the remodeling process within the hEHT injury repair model, evidenced by increased ROS accumulation and cell apoptosis in the early period (Supplementary Fig. 2a–d), and CF activation (Fig. 1d–f), ECM deposition (Fig. 1g, h) in the late period.

hEHT injury repair model exhibited diversified cellular composition during remodeling

To elucidate cellular and signaling dynamics during repair, we performed scRNA-seq on Cryo and Con hEHT at Early and Late time points (Supplementary Fig. 4e). For both Cryo and Con groups, we identified four main cell types-CM, CF, epicardial cells (EP), and endothelial cells (ED) (Fig. 3a)-based on cell clustering followed by automatic annotation and expression of known cellular marker genes in the human heart (Supplementary Fig. 5a, b, Supplementary Table 2). Further debatching was performed using Harmony, and the data reached convergence after 4 iterations of flexibility, yet no injury-specific clusters emerged (Supplementary Fig. 5c, d). Meanwhile, the UMAP plots separated across varying groups illustrating cryoinjury didnot yield new cell subpopulation (Fig. 3a, b).

**Fig. 3: Transcriptomic analysis underlying the subcellular composition and signaling dynamics during the remodeling of injured hEHT.**

Changes in cell states are pivotal for the in vivo cardiac response following MI, encompassing cell stress, metabolic shift, CM dysfunction, and CF activation²⁶. To analyze CM and CF state shifts following cryoinjury, we subclustered these cell types across early and late remodeling periods, revealing four distinct CM subpopulations (CM1-CM4) (Fig. 3c–f) and five distinct CF subpopulations (CF1-CF5) (Fig. 3g–i) existed in both Cryo and Con hEHT. All CM subpopulations expressed sarcomere-related genes (MYH6, MYH7, TTN, MYH7B), but albeit at varying expression levels, with CM1 showing the lowest, particularly for MYH7 expression, which is the predominant isoform in the adult heart²⁷. The MYH6/MYH7 ratio signifies the transition of sarcomere components from fetal to an adult isoform²⁷, with CM1 displaying the lowest ratio and expressing cardiac development markers such as BMP4²⁸, highlighting its fetal traits (Fig. 3d–f). CM2 and CM3 shared similar sarcomere-related gene expression levels (Fig. 3e, f), whereas CM2 stood out with metabolic-related genes, including nuclear-encoded mitochondrial genes (NDUFA4, NDUFB11, COX7C, and COX5B), and the antioxidant gene CRYAB (Supplementary Fig. 6a, b), indicating CM2 as a high-energy subpopulation found in the adult heart²⁹. CM4 highly expressed division-associated genes (MKI67, CDK1) (Fig. 3d, e) suggested its possible role in regenerative repair³⁰.

Recent studies have identified multiple CF subpopulations in both healthy and diseased hearts³¹, crucial for myocardial remodeling. Similarly, five CF subpopulations (CF1, CF2, CF3, CF4, and CF5) were identified in the hEHT injury repair model (Fig. 3g). CF1 specifically overexpresses cell division-related genes, like MKI67, suggesting its proliferative subpopulation (Fig. 3h, i). CF2 exhibited significant expression of stress fiber markers ACTA2, characteristic of myoCF (Fig. 3h, i). CF3 showed enrichment in ECM organization and cell adhesion GO terms, with a high expression level of HTRA3³², known for maintaining CF quiescence (Fig. 3h, i). Intriguingly, CF5 expressed classical CF markers³³, such as DCN and LUM (Fig. 3i and Supplementary Fig. 5b), while also expressing markers associated with epithelial cell development, like UPK3B³⁴, and CXADR (Fig. 3h, i). Both CF4 and CF5 highly expressed CD44 (Fig. 3i), a cell-surface glycoprotein encoding gene that can regulate CF function following MI³⁵. In particular, DEGs in CF4 were mainly associated with epithelial-mesenchymal transition, indicating the transition status of CF4 between epicardial-like CF5 and other CF subpopulations. we further found that the VIM⁺/CD44⁺/CXADR^- CF, that is, CF4, accumulated in the remodeled area (Fig. 3j, k), whereas VIM⁺/CD44⁺/CXADR⁺ CF5 appeared only in the uninjured area (Fig. 3j, k), indicating a potential transition from epicardial-like CF to myoCF in myocardial remodeling.

The postnatal epicardium typically remains quiescent³⁶ but can reactivate following cardiac injury, leading to the upregulation of WT1 and EMT marker genes such as TBX18 or SNAI1³⁷. Upon injury, EP undergoes epithelial-to-mesenchymal transition and migrates to the injured area, contributing to post-injury repair³⁸. Notably, the EP in the hEHT injury repair model, immunostained by WT1, predominantly localized in the surface layer of hEHT, similar to the structural composition of myocardium in vivo, and also can contribute to the remodeled area (Supplementary Fig. 6c), indicating that EP migration might enhance repair post-injury. Meanwhile, some Tie1⁺ EC were irregularly distributed between uninjured and remodeled areas (Supplementary Fig. 6d), implying both EP and EC play roles in scar formation within our model, supported by animal models evidence³⁹. In addition to altering the spatial distribution of cell subpopulations, cryoinjury also led to a decrease in the total number of cells captured by scRNA-seq in hEHT (Supplementary Fig. 6e). Specifically, cryoinjury elevated the percentage of immature CM (CM1, CM2) and inactive CF (CF4, CF5), while all other cell subpopulations decreased in the early period (Supplementary Fig. 6e). This trend was completely reversed in the late period, with an increase in activated CF1, CF2 and also in proliferating CM3 (Supplementary Fig. 6e), which was consistent with the presence of different proportions of CF and CM in the remodeled area (Fig. 1g, h).

Signaling dynamic shift of different cell subpopulations in hEHT injury repair model

To explore the injury response of CM subpopulations, GO and KEGGs analysis of DEGs between Cryo and Con hEHT indicated that in the early period, primarily CM1 and CM4 underwent oxidative stress processes (Supplementary Fig. 7a), akin to the enriched gene ontology observed in bulk RNA-seq analysis, featuring genes SOD1⁴⁰ (Fig. 3l), PON2⁴¹, PXDN⁴² (Supplementary Fig. 7b). In the late period, enhanced metabolism and sarcomere remodeling were predominantly observed in CM2 and CM3, indicated by genes DIO3⁴³ and ARF1 (Supplementary Fig. 7c, d). Additionally, WNT signaling pathway genes like FRZB⁴⁴ (Fig. 3l), and TMEM88⁴⁵, exhibited differential regulation, being upregulated in the early and downregulated in late periods (Supplementary Fig. 7b, d). This suggested that the enhanced oxidative metabolism and functional recovery in CM2 may be modulated differently by the WNT signaling pathway during the remodeling. Moreover, transcription factor (TF) analysis indicated that upregulated TFs in four CM subpopulations correlated with hypoxic stress (STAT5A⁴⁶, BHLHE40) and anti-apoptosis (E2F6, E2F8, NFE2L1) in the early period (Supplementary Fig. 8a, b). TFs associated with cardiac development regulation (ETS1, SP2⁴⁷, MSX2) and CM proliferation (E2F7, E2F8) were more pronounced in the late period (Supplementary Fig. 8c, d), highlighting the role of stress in activating the intrinsic regenerative potential of CMs in the hEHT injury repair model.

The DEGs of CF were enriched in the TGFβ, MAPK, PI3K, and RHO signaling pathways (Supplementary Fig. 9a) in the early period, which were also identified in bulkRNA-seq. Notably, CF3 stood out with more DEGs, including TGFB1 (Fig. 3m) and RHOB⁴⁸ (Supplementary Fig. 9b), both related to CF activation and proliferation. As myoCF, CF2 showed elevated ACTA2 expression in the cryoinjury group, indicating heightened myoCF activation initiation in the early period for further scar formation in the late period (Fig. 3m). In the late period, CF in the cryoinjured group exhibited increased expression related to proliferation (CCNB2⁴⁹, LOX⁵⁰) and PI3K signaling pathway (CREB3L1⁵¹) (Fig. 3m and Supplementary Fig. 9d). Most CF subpopulations, apart from CF5, also showed elevated collagen secretion-associated genes, such as COL1A1, COL1A2, and TIPM2 (Fig. 3m and Supplementary Fig. 9d).

TF analysis for the CF subpopulation showed that the levels of proliferation-associated TFs E2F7⁵², E2F8, and E2F1⁵³ were decreased in proliferative CF1 in the early period (Supplementary Fig. 10a, b) but then increased in the late period (Supplementary Fig. 10c, d). Similarly, CF4 and CF5 highly expressed TF associated with cell fate transition in the early period, such as PBX4⁵⁴ and FOSL2⁵⁵ (Supplementary Fig. 10a, b), but had lower expression of MSX1 in the late period (Supplementary Fig. 10c, d), indicating that the injury-induced fate transition in inactive CF was progressively attenuated and eventually contribute to fibrosis. These findings indicated that the cryoinjury prompted CF to transition to a more active state characterized by enhanced cell proliferation and increased secretion of ECM components during the late period of the remodeling in the hEHT injury repair model, which mirrored the scar maturation process observed in vivo⁵⁶.

CF fate transition and interaction with CM in hEHT injury repair model

CF plays a critical role in pathological myocardial remodeling, leading us to investigate its fate transitions during injury repair. Pseudotime analysis identified two lineage trajectories: cell fate 1 from CF5 to CF1 and cell fate 2 from CF5 to CF3 (Supplementary Fig. 11a, b). Three gene sets were identified to modulate the cell fate transitions (Fig. 4a), including gene set 1, which was enriched in terms related to cell fate determination, oxidative stress, and WNT, Rho signaling pathways to promote cell fate 2 from CF5 to CF3. Gene set 3, which enriched for BMP, TGFβ, and MAPK signaling pathways, dominated the transition from CF5 to CF1 (Fig. 4b). Specifically, in gene set 3, the cytoskeleton-related genes SIPA1L1 and ACTA2 increased in cell fate 1 but decreased in cell fate 2, while the ECM-related genes POSTN and COL1A1 exhibited opposite patterns (Supplementary Fig. 11c), indicating the differential modulation in CF fate. To validate the in vivo relevance of CF fate transition, we further referenced the scRNA-seq data from the clinical heart with dilated cardiomyopathy (DCM) and ischemic cardiomyopathy (ICM) hearts³¹ (Supplementary Fig. 11d). Clustering and pseudotime analysis revealed that the quiescent CF in human MI hearts similarly transitioned to myoCF or contractile CF (Supplementary Fig. 11e, f), regulated by TGFβ, WNT, PI3K, and ERK signaling (Supplementary Fig. 11g). The CF in hEHT repair model exhibited similar fate transition patterns, with dynamic gene expression related to TGFβ signaling (JUN, FOS, COL3A1), ECM (ADAMTS1, COL16A1, MMP28), and stress response (COL1A1, BTG, FOSB) as in vivo (Fig. 4c, d).

**Fig. 4: Pseudotime analysis revealed that CF in the hEHT injury repair model exhibited similar cell fate transitions in vivo.**

Given CF’s highly plastic and dynamic nature, cell interaction analysis at both the major (Supplementary Fig. 12a) and subtype level (Supplementary Fig. 12b) revealed a notable alteration in the number of CF-CM interactions post-cryoinjury, especially between CF2/CF5 and CM, except mature CM3 (Supplementary Fig. 12c). To focus on the most representative interactions, the injury-specific cell interactions between CF and CM were further investigated at the subpopulation level (Supplementary Fig. 12d). In the early period of remodeling, the APOA2_ABCA1 interaction was enhanced between CF2/5 and CM2/4, which is known to regulate CM metabolism⁵⁷. The cryoinjury also strengthened interactions between CF5 and immature CM1/2 or proliferating CM4 via NRXN2_DAG1 (Supplementary Fig. 12e), which can promote CM proliferation⁵⁸. CF2 continuously expressed TGFB1 to induce CM stress⁵⁹ in both the early and late period (Supplementary Fig. 12e). Meanwhile, the detrimental effects of CF2 on CM were evidenced by the increased expression of THBS1 in the late period (Supplementary Fig. 12e), which can trigger CM autophagy⁶⁰. Conversely, FB5 exhibited a protective effect through increased DKK1_LRP6 interaction with CM (Supplementary Fig. 12e), leading to the inhibition of CM apoptosis⁶¹. The above scRNA-seq analysis demonstrated the dynamic cellular orchestration during remodeling of cryoinjured hEHT, especially in CF. These cellular dynamics were mainly dominated by serious signaling pathways such as TGFβ, WNT, RHO, PI3K, NFκB, and MAPK, which allowed us to gain a deeper understanding of the underlying mechanisms of myocardial pathological remodeling.

A mini hEHT-injury repair model for rapid screening of potential therapeutic targets for IHD

We developed a mini hEHT-injury repair model with a diameter of approximately 4 mm to expedite the evaluation of potential therapies against myocardial pathological remodeling. Utilizing 3D printing, we fabricated a mini-hEHT similar in shape to skeletonized 8-strip PCR tubes (Fig. 5a). A pedestal with a 0.3 mm diameter stainless steel needle (Fig. 5b) enabled batch, uniform cryoinjury in each mini-hEHT (Fig. 5c). Notably, the mini hEHT response to cryoinjury was consistent to that in hEHT. The injured area initially expanded in the early period (days 0–1) and began shrinking during the late period (days 1–5) (Fig. 5e, f), revealing a regionalized fibrotic area consisting of less CM and primarily populated by CF (Fig. 5g).

**Fig. 5: Establishment of the mini hEHT-injury repair model.**

Using the mini hEHT- injury repair model, we can evaluate the pro-repair of seventeen modulators from nine signaling pathways (Fig. 5d) identified from the bulkRNA-seq and scRNA-seq analyses (Supplementary Fig. 13a). Variation in the remodeling rate of cryoinjured mini-hEHT was first observed (Supplementary Fig. 13b), bFGF, Y27632, SB431542, 740YP, LY294002, and DMU212 significantly accelerated the remodeling (Supplementary Fig. 13c). Specifically, bFGF and LY294002 expanded the cryoinjured area in the early period, completing remodeling by day 4, while DMU212, 740YP, SB431542, and Y27632 directly shortened healing time to day 3 (Supplementary Fig. 13c, d). Thus, these interventions accelerated remodeling in different approaches, suggesting that they influenced remodeling outcomes by regulating CM or CF individually in the hEHT injury repair model.

Multidimensional evaluation of the effect of different signaling modulators on mini hEHT remodeling outcomes

We further developed a quantitative assessment strategy to analyze the effects of these modulators on CM and CF in the remodeled area by analyzing 3D images of the mini-hEHTs stained with CM and CF markers on day 6 (Supplementary Fig. 14a, Supplementary Movies 6 and 7). By subtracting the CM’s absence area from the maximum cryoinjured area at each Z-axis layer (Supplementary Fig. 14a, b), we can estimate the percentage of CM volume inside the remodeled for further comparison. We also assessed the CM eccentricity degree to gauge CM contribution to remodeling and measured CF stress fiber length to evaluate CF activation. (Supplementary Fig. 14c). Control experiments ruled out DMSO interference with these metrics (Supplementary Fig. 14d–f). With these three outcome measures, we then scored the impact of seventeen signaling modulators on mini-hEHT remodeling (Fig. 6a).

**Fig. 6: The evaluation of treatment effects on the mini hEHT-injury repair model.**

Among the seventeen candidates, CHIR99021, SCH772984, and Y27632 significantly increased CM fractional volume in the remodeled area from 40% in the control to over 60% (Supplementary Fig. 14g) and simultaneously reduced CM eccentricity (Supplementary Fig. 14h), indicating the beneficial effect on structural remodeling outcome. As for the impaction on CF, TGFβ1 significantly increased stress fibers length (Supplementary Fig. 14i). At the same time, PDTC, NS398, CHIR99021, SQ22536, 740YP, LY294002, and SB431542 inhibited CF activation (Supplementary Fig. 14i). Correlations among CM eccentricity, CF stress fiber length, and CM proportion in the remodeled area (Fig. 6b, c) suggested that CM and CF interact to influence structural outcomes. The overactive CF may prevent CM migration to the cryoinjured area, decreasing their proportion in the remodeled area. Thus, the multidimensional evaluating system based on these three interrelated indices, i.e., CF activation, CM eccentricity, and CM fractional volume in the remodeled area, identified Y27632, CHIR99021, and SB431542 possessed leading benefits on both myocardial regeneration and anti-fibrosis over other treatments (Fig. 6d).

Integrative scoring system to evaluate the functional recovery in the remodeled area

To understand the structural and electrophysiological shift clearly, we utilized the hEHT injury repair model to assess the effect of Y27632, CHIR99021, SB431542, and their combination (Fig. 7a). Consistent with the screen results based on mini-hEHT, Y27632, and CHIR99021 narrowed down the CM eccentricity and increased the CM volume fraction in the remodeled area (Supplementary Fig. 15a–d). SB431542 exhibited the most significant inhibition of CF activation but had only a tendency to increase CM volume fraction. In contrast, the phased combination of Y27632, CHIR99021, and SB431542 increased the CM volume fraction and inhibited fibrosis in the remodeled area (Supplementary Fig. 15a–d).

**Fig. 7: The integrative evaluation of treatment effects on the recovery of abnormal calcium activities in the hEHT repair model.**

Calcium activity abnormalities manifested as disrupted calcium signal conduction, driven by non-functional CF accumulation. We further evaluated the effect of these modulators on calcium activity (Supplementary Movie 8) in the hEHT injury repair model. We quantified the regional blockage of calcium wave conduction at baseline and under stimulation by isoproterenol (ISO) or increased electrical pacing (Fig. 7b). Only the phased combination of Y27632, CHIR99021, and SB431542 reduced the regional conduction block under three loading conditions (Fig. 7c). Moreover, the combination group can partially improve the calcium signal conduction homogeneity in the remodeled area, while largely restoring conduction velocity to normal level (Fig. 7d), where CHIR99021 uniquely enhancing conduction homogeneity under three loading conditions (Fig. 7d). Compared to uninjured areas, remodeled area exhibited reduced calcium signals across all conditions, with no treatment restoring this attenuation (Fig. 7e). The combination and CHIR99021 treatment could relatively rescue the most calcium activity parameters, represented by calcium signal amplitude and CSD80, to a normal level similar to that of the uninjured area in the control group (Fig. 7f).

To develop an integrative scoring system, an abnormal function area and ten parameters related to the calcium signal propagation or calcium signal wave were applied to evaluate the global characteristics of calcium activity in the remodeled area. Treatment rankings emerged from comparing the mean values of parameters with those in the uninjured area of the control group. The smaller the difference, the higher the ranking and the higher the score (Fig. 7d-f). To prevent similar highly correlated parameters from skewing scores⁶², we analyzed the correlation among ten parameters derived from the calcium conduction and signal waveform (Supplementary Fig. 15e). The resulting heatmap revealed that maximum upstroke and recovery velocity correlated most significantly with calcium signal amplitude. In addition, CSD80 and CSD50 had higher correlations due to their similar definitions, so the redundant parameters were removed from the composite scoring system (Supplementary Fig. 15e). Ultimately, our final integrative scoring system comprised eight informative parameters to evaluate treatment effects on functional recovery. This integrative scoring system demonstrated that the combination of Y27632, CHIR99021, and SB431542 effectively reduced abnormal functional areas (Fig. 7c), while also having the optimal effect in restoring conduction velocity, conduction direction homogeneity, and calcium signal amplitude and CSD, to achieve the highest scores across three loading conditions (Fig. 7g).

BulkRNA-seq revealed the antifibrotic/cardiac repair effects of Y27632/CHIR99021/SB431542

Given the significant effect of the combination treatment in promoting functional repair of injured hEHT, we performed bulkRNA-seq on CM and CF in cryoinjured hEHT sequentially treated with Y27632, CHIR99021, and SB431542 and in control hEHT correspondingly treated with DMSO (Supplementary Fig. 16a). PCA clustering analysis revealed distinct effects of these interventions on CM and CF (Supplementary Fig. 16b, c), with downregulation of LIMK2 and upregulation of LEF1, the downstream effector of ROCK and WNT signaling, and direct downregulation of TGFβ1 indicating the efficacy of the intervention at the corresponding target signaling (Supplementary Fig. 16d). Compared to Y27632 and SB431542, CHIR99021, applied during days 2–4 post-injury, yielded greater effects on both CM and CF (Supplementary Fig. 16e). These interventions also induced common effects, such as leading to the upregulation of STC1 in CM (Supplementary Fig. 16f), making it more antioxidant. Meanwhile, the sustained upregulation of TDO2⁶³ (Supplementary Fig. 16f) after the CHIR99021 treatment was shown to be cardioprotective.

Furthermore, Y27632 treatment specifically downregulated CHAC2⁶⁴ expression in CM to reduce ROS, while the specific upregulation of GPER1⁶⁵ can protect CM from various stressors (Supplementary Fig. 17a). Y27632 activated the pyruvate metabolism (LDHA, SLC2A1, HK2), enhanced the protection against hypoxic stress (NDNF, STC1), and increased prostaglandin synthesis (PTGDS, PTGES) in CM and CF (Supplementary Fig. 17a), demonstrating its potential role in metabolic reprogramming under stress. Additionally, the specific upregulation of ACKR3⁶⁶ and UCP2⁶⁷ in CF is cardioprotective (Supplementary Fig. 17a).

CHIR99021 treatment reduced the expression of the ferroptosis-related gene ALOX15⁶⁸ in CM while increasing the expression of the cardioprotection-related gene SSTR2⁶⁹, and also activation of C5AR1⁷⁰, which can promote cardiomyocyte proliferation after cardiac injury (Supplementary Fig. 17b). The consistent changes in ERK1/2 signaling-related genes (NPY, KDR, CD44, FGF19) in CM and CF, especially CD44 as a marker distinguishing between epicardial-like CF5 and transitional CF4, which was regionally highly expressed within the cryoinjured hEHT contributing to fibrotic scar formation (Fig. 3g, h) and was downregulated in the CHIR99021 group (Supplementary Fig. 17b). Meanwhile, a series of pro-inflammatory and pro-fibrotic mediators such as IL1B, IL34, IL17RE, SPP1, and MMP9 were downregulated in the CHIR99021 group (Supplementary Fig. 17b). The most studied MMP9⁷¹ will promote adverse cardiac remodeling post-MI, while it was upregulated in both early and late remodeling periods of cryoinjured hEHT (Supplementary Fig. 4d). Interestingly, CHIR99021 also specifically upregulated NPNT⁷² in CF, a gene associated with pro-cardiomyocyte division, and TBX5⁷³, a gene associated with promoting CF reprogramming to CM (Supplementary Fig. 17b).

SB431542 treatment significantly downregulated ECM synthesis genes in CM, such as TGFB1, COL1A1, COL5A1, ADAMTS8, and SMOC2 (Supplementary Fig. 17c), while the upregulation of EPO⁷⁴ was shown to play a critical protective role post-human MI. In addition to directly downregulating the myoCF marker ACTA2, SB431542 specifically downregulated the profibrotic-related genes SMYD3, and RASGRF1 (Supplementary Fig. 17c). Taken together, these findings demonstrated that the prominent functional repair of cryoinjured hEHT after combination treatment was jointly achieved by the enhancement of cellular stress resistance by Y27632, the reduction of CM apoptosis and promotion of CM proliferation by CHIR99021, and the direct antifibrotic effect of SB431542.

MI mice exhibited cardiac functional recovery to Y27632, CHIR99021, and SB431542 to treatment in combination

Undeniably, immature and proliferating CM subpopulations in our hEHT injury repair model can be essential for initiating regenerative repair after injury, as demonstrated in the cryoinjury model of neonatal mice¹⁶. The heart of neonatal mice with 15% direct apical excision can fully recover within 21 days, demonstrating its powerful regenerative capacity, but this capacity declines sharply with age⁷⁵. Considering that our model of myocardial injury repair focused more on fibrotic remodeling and screening for related therapeutic targets, we, therefore, performed transmural cryoinjury (penetrating the entire ventricular wall) in neonatal mice at 3 days postnatal age (P3) to promote cardiac healing with fibrotic scarring. Post-MI, the myocardial pathological remodeling process is multi-factor led and highly dynamic in vivo, which includes the early inflammatory phase, cardiac replacement fibrosis phase, and scar maturation phase³. Similarly, we validated the effects of Y27632, CHIR99021, and SB431542, these single-factor treatments, and their phased combination treatments on myocardial injury repair in the above cryoinjury model of neonatal mice (Supplementary Fig. 18a). After 12 days of treatment, the heart in the vehicle groups showed thinning of the ventricular wall and fibrotic scar filling (Supplementary Fig. 18b), whereas only the combination treatment can preserve ventricular structure and significantly reduce the fibrotic area (Supplementary Fig. 18c).

Subsequently, we developed the MI model in adult mice and evaluated the therapeutic effects of Y27632, CHIR99021, SB431542, and their combination (Fig. 8a). Echocardiography assessment (Fig. 8b) revealed that both the ejection fraction (EF) and fractional shortening (FS) significantly declined in all treated groups compared with the sham-operated group at day 1 post-MI. However, only the combined treatment enhanced cardiac function considerably after 21 days post-treatment compared to the vehicle group (Fig. 8c, d). The Masson staining analysis (Fig. 8e) further indicated that the combined treatment can significantly reduce the fibrotic area compared to the vehicle group (Fig. 8f). Notably, the phased single-factor treatment with CHIR99021 in days 7-14 (cardiac replacement fibrosis phase) post-MI also significantly reduced the formation of fibrotic scar (Fig. 8f) but failed to rescue the cardiac function (Fig. 8c, d). In contrast, sustained CHIR99021 treatment throughout the remodeling process (21 days post-MI) in adult MI mice (Supplementary Fig. 19a) failed to reduce the fibrotic area. Phased interventions with Y27632, CHIR99021, and SB431542 still facilitated functional recovery post-MI (Supplementary Fig. 19b–f), indicating the importance of symptomatic treatment of remodeling stages with different pathological features. The above results further validate that the potential value of the Y27632, CHIR99021, and SB431542 identified from the hEHT injury repair model, their phased combination can distinctively have a positive impact on myocardial pathological remodeling.

**Fig. 8: The combination treatment reduced the fibrotic area and enhanced cardiac function in the adult mouse model of MI.**

To further assess the impact of different treatments on the electrophysiological recovery of MI mice hearts, we analyzed the calcium activity in treated ventricles under four loading conditions using optical mapping (Supplementary Fig. 20a). In the control ventricle, as we found in the hEHT injury repair model, the area of disturbed calcium signaling progressively increased with increasing electrical pacing frequency and ISO loading (Supplementary Fig. 20b). In addition, we observed reduced calcium signaling amplitude and prolonged CSD80 in disturbed area, along with the occurrence of arrhythmias at high-frequency pacing (Supplementary Fig. 20c, d). The combination treatments significantly decreased abnormal functional areas (Supplementary Fig. 20b) and restored ventricular conduction velocity and conduction direction homogeneity to normal levels (Supplementary Fig. 20e). Based on the integrated scoring system developed to evaluate all characteristic parameters related to the calcium signaling activity, the combination treatment demonstrated the best effect in rescuing the electrophysiological dysfunction in the adult infarcted mice (Supplementary Fig. 20f).

Discussion

Current studies based on animal MI models⁷⁶ and human pluripotent stem cell-derived CF⁸ have provided valuable insights into the onset and progression of myocardial pathological remodeling. However, challenges like species differences and low drug screening throughput persist⁷⁷. Developing the hEHT model that mimics myocardial pathological remodeling for screening therapeutic targets that could normalize myocardial physiology following injury is crucial. Our hEHT repair model mimicked the phased remodeling process, including cellular stress, CF activation, and ECM deposition post-injury. The remodeled area was characterized by fewer CM, increased CF, and more type I collagen secretion and fibronectin. Our model also faithfully reproduced the electrophysiological dysfunction caused by fibrotic scar, represented by the abnormalities in calcium signaling activity in the remodeled area.

The evolution and transformation of CF act as the major events in driving the development of myocardial remodeling. Prior studies relied on animal MI models⁷⁸ or clinical samples⁷⁹. Clues on how human CF transits from a quiescent state to an active state and their role in fibrotic scarring are limited in animal models, and at the same time, clinical samples cannot encompass the pathological progression from healthy to scarred myocardium. In our hEHT repair model, transcriptomic analysis identified CF subpopulations in hEHT resembling both healthy and infarct myocardium: epithelial-like CF (CD44⁺, UPK3B⁺, LUM⁺), transition-state CF (CD44⁺), quiescent CF (HTRA3⁺, TCF21⁺), myoCF (ACTA2⁺) and proliferating CF (MKI67⁺). In cryoinjured hEHT, CF4 displayed notable cell migration and specification in the early remodeling period compared to control hEHT. myoCF showed enhanced proliferation and collagen assembly activity in the late period. These findings indicated that the CF might undergo the early stress response and late activation during myocardial remodeling, partly mimicking the post-MI phenotype in vivo. TGFβ signaling was upregulated in all CF subpopulations, while ROCK/Rho, MAPK/ERK, PI3K/AKT, WNT, NFκB, cAMP, and bFGF regulate myocardial fibrosis in various CF and CM subpopulations. Interestingly, previous studies noted myoCF conversion and epithelial-mesenchymal transition in fibrotic myocardium⁸⁰. Our annotation of CF subpopulations further supported these observations, emphasizing the need for further investigation into these identities’ transitions, and our model can facilitate the exploration of the markers of CF activation.

Our mini hEHT injury repair model allows rapid testing of modulators identified from transcriptome analysis. Y27632, CHIR99021, and SB431542 were identified to show anti-fibroblast activation and pro-cardiomyocyte in situ refilling. The ROCK inhibitor Y27632 was reported to improve cardiac function⁸¹ and also reduce apoptosis of transplanted CM⁸² in MI mice. In the hEHT injury repair model, the ROCK/RhoA signaling-related TF BHLHE40 was upregulated in CM (Supplementary Fig. 8b) and correlated with MI-induced inflammatory response⁸³. CHIR99021 enhanced hPSC-CM proliferation⁸⁴, evidenced by ETS1 upregulation in immature CM1 at the late remodeling period (Supplementary Fig. 8b), while ETS1 can directly promote cardiac development through WNT signaling⁸⁵. Meanwhile, SB431542 inhibits the TGF-β receptor and reduces CF activation-related gene expression⁸⁶. This contrasts with TGFβ signaling upregulation by myoCF in cryoinjured hEHT (Fig. 3k). In addition, myoCF also directly interacted with CM through TGFβ1(Supplementary Fig. 12c), potentially inducing CM stress⁸⁷. These evidences further support the potential anti-apoptotic, pro-cardiomyocyte proliferation, and anti-fibroblast activation effects of Y27632/CHIR99021/SB431542.

The myocardial pathological remodeling encompasses stress response, CF activation, and scar maturation post-MI. Ongoing fibrosis risks functional impairment, but phased intervention with different modulators can be more conducive in guiding the early fibrotic response to participate in myocardial defect repair while arresting persistent and excessive fibrosis. Notably, the phased combination of Y27632, CHIR99021, and SB431542 effectively reduced dysfunction area and improved calcium activity. Our studies in MI mice demonstrated better functional recovery with phased combination treatment targeting over single-factor treatments. Although the effect of other combinations of phased interventions and the regulatory mechanisms behind them are yet to be verified, this phased intervention strategy undoubtedly provided valuable insights for developing therapeutic regimens against myocardial injury.

Undeniably, the absence of immune cells and the immaturity of hPSC-CM in hEHT remains a noticeable challenge for our hEHT repair model to comprehensively mimic the myocardial pathological remodeling post-MI in adult hearts. Nonetheless, the occurrence of immune system-related terms from transcriptomic analysis still gave confidence for this engineered model to explore how immune cells respond to regional stress and regulate subsequent fibrosis mediated by CF with various identities. In summary, we have developed an engineered model of myocardial injury-repair simulating the phased pathological remodeling and functional impairment post-MI, based on which we can deeply analyze the cell fate regulation and subcellular transformation during myocardial remodeling. More importantly, this model provides a humanized screening platform for screening therapies targeting myocardial injury, promoting better recovery of fibrotic myocardium.

Methods

hPSC culture and cardiomyocyte differentiation

Human iPSC line WTC⁸⁸ (Coriell Institure, GM25256) and Human ESC line H9^89,90 (WiCell, WB68075) were used and seeded on Matrigel (Corning, 354277) coated plate at the cell density of 0.02 million/cm². The hPSC was maintained in mTeSR1 (STEMCELL Technologies, 85850) supplemented with 10 μM Y27632 (STEMCELL Technologies, 72302) on the first day and passaged every 5 days using Versene Solution (Thermo Fisher, 15040066). For inducing hPSC differentiated into CM, the mTeSR1 was changed into RPMI-1640 (Thermo Fisher, C11875500BT) supplemented with 0.2 mg/mL L-Ascorbic acid-2-phosphate sesquimagnesium salt hydrate (Sigma-Aldrich, A8960), B-27 supplement minus insulin (Shanghai BasalMedia, X064C5), and 5-8 μM CHIR99021 (STEMCELL Technologies, 72054) when the hPSC reached 80-90% confluence. After 48 hours, the medium was changed into the 1640/B27(-INS) medium. On day 3 of differentiation, 5 μM IWR1 (STEMCELL Technologies, 72564) was introduced into the 1640/B27(-INS) medium for another 48 hours. After that, the 1640/B27(-INS) medium was changed every 2 days. On day 9, beating CM can be observed. The differentiated cells on days 14-16 were dissociated into single cells using 2 mg/mL collagenase I (Sigma-Aldrich, C0130) for 1 hour and 0.25% trypsin/EDTA (Thermo Fisher, 25200056) for 5 minutes at 37 °C. After digestion termination using stop buffer (high glucose DMEM (Thermo Fisher, C11995500BT), supplemented with 50% fetal bovine serum (FBS) and 20 μg/mL DNase I (Millipore, 260913)), the harvested cells were used for fabricating hEHT.

hEHT fabrication and culture

Similar to the method as previously described in Refs. ^10,11, the 9 × 9 mm polydimethylsiloxane (PDMS, Dow Corning) square molds were treated with 0.2% Pluronic F127 (Sigma-Aldrich, P2443) for at least 1 hour and washed by DPBS for the next step. After dissociation, 1.0 × 10⁶ differentiated cells were added into 58 μL 1× cardiac medium consisting of low glucose DMEM (Thermo Fisher, 10567014), 2 μg/mL Vitamins B12 (Sigma-Aldrich, V2876), 2 mg/mL 6-Aminocaproic acid (Sigma-Aldrich, A2504), 1% Penicillin-Streptomycin (Thermo Fisher, 15140122), 10% FBS and 2.4 μL 50 U/mL thrombin (Sigma-Aldrich, T7201). Then the cardiac medium was mixed with the hydrogel solution consisting of 24 μL 2× cardiac medium, 24 μL 10 mg/mL fibrinogen (Sigma-Aldrich, F4883), and 12 μL Matrigel (Corning, 354277). The 120 μL cell-laden hydrogel was transferred to the 9 × 9 mm Nylon frames in the PDMS model and placed into the 37 °C incubator. After 30 min for polymerizing, the hEHT was transferred into the 24-well plate on a rocking platform and cultured in the hEHT medium consisting of RPMI 1640, B-27 supplement minus insulin (Shanghai BasalMedia, X064C5), 0.2 mg/mL ascorbic acid (Sigma-Aldrich, A8960), 2 mg/mL 6-Aminocaproic acid (Sigma-Aldrich, A2504), 1 mM sodium pyruvate (Thermo Fisher, 11360070), 0.1 mM non-essential amino acid (Thermo Fisher, 11140050), 0.45 μM 1-thioglycerol (Sigma-Aldrich, M6145), 5% FBS and 1% Penicillin-Streptomycin (Thermo Fisher, 15140122). 10 μM Y27632 (STEMCELL Technologies, 72302) were added into the hEHT medium for the first 24 hours. The hEHT medium was changed every 2 days.

Development and assessment of the hEHT injury repair model

After 4 days of cultivation, we designed and 3D-printed the cryoinjury device with a conical tip of 0.35 mm diameter and 8 mm height using LCD based SLA 3D Printer (Anycubic, Photon) to introduce standardized cryoinjury into hEHTs. After pre-cooling with liquid nitrogen for 15 seconds, the cryoinjury device easily penetrates and cryoinjury the hEHT with four physical holes of approximately 300–500 μm in diameter. For developing the mini hEHT injury repair model, the mini hEHT fabrication device (Fig. 5a) and cryoinjury pedestal (Fig. 5b) were designed by SolidWorks. The design files were exported to STL. Format and 3D printed by WeNext Technology Co., Ltd using nylon. The mini hEHTs were fabricated using the same method for fabricating hEHT and then batch cryoinjured by the 0.3 mm diameter stainless steel needle fixed in the 3D printed pedestal after 3 days of culture. The cryoinjured hEHTs or mini hEHTs were cultured in the hEHT medium for another 6 days for remodeling. The following compounds were added into the hEHT medium to intervene in the entire remodeling process of the cryoinjured hEHT: 25 ng/mL bFGF (MedChemExpress, HY-P7004), 10 μM Y27632 (STEMCELL Technologies, 72302), 10 ng/mL Rho activator II (Cytoskeleton, CN03-A), 5 μM CHIR99021 (STEMCELL Technologies, 72054), 5 μM IWR1 (STEMCELL Technologies, 72564), 10 ng/mL TGFβ1 (PeproTech 100-21-10UG), 5 μM SB431542 (STEMCELL Technologies, 72232), 5 μM Forskolin (MedChemExpress, HY-15371), 5 μM SQ22536 (MedChemExpress, HY-100396), 3 μM 740YP (MedChemExpress, HY-P0175), 5 μM LY294002 (MedChemExpress, HY-10108), 5 μM DMU212 (MedChemExpress, HY-137977), 3 μM SCH772984 (MedChemExpress, HY-50846), 5 ng/mL PMA (MedChemExpress, HY-18739), 5 μM PDTC (MedChemExpress, HY-18738), 5 μM Prostaglandin E2 (MedChemExpress, HY-101952), 5 μM NS398 (MedChemExpress HY-13913). The 10 μM Y27632 was added to the hEHT medium for combination treatment in the first 2 days. After 2 days, the Y27632 was removed, and the CHIR99021 was added for another 2 days. In the final 2 days, only the SB431542 was added to the hEHT medium. The 0.1% DMSO was added to the hEHT medium as the control, and all the medium was changed every 2 days.

pLV-mCherry and TroponinT-GCaMP5-Zeo lentivirus production

The pLV-mCherry (Addgene, 36084) and TroponinT (TNNT2)-GCaMP5-Zeo (Addgene, 46027) were packaged by co-transfection with the lentiviral packaging plasmid psPAX2 (Addgene, 12260), and the VSV-G envelope expressing plasmid pMD2.G (Addgene, 12259) in 293 T cells using Hieff Trans Liposomal Transfection Reagent (Yeasen, 40802ES03). After 48 hours, the culture supernatants were filtered using 0.22 μm filters (Millipore, SLGP033RB) and centrifuged using Amicon® Ultra-15 Centrifugal Filter Unit (Millipore, UFC910024) for harvesting lentivirus in Hanks’ solution (Sangon, B548148).

Time-lapse imaging of the remodeling of cryoinjured hEHT

For labeling the non-CM populations in the differentiated cells, the formulation of the 120 μL cell-laden hydrogel was changed into 1.0 × 10⁶ differentiated cells, 28 μL pLV-mCherry lentivirus, 28 μL TNNT2-GCaMP5 lentivirus, 2.4 μL 50 U/mL thrombin and the 60 μL hydrogel solution. After culture for 4 days, the labeled hEHT was cryoinjured and cultured for another 2 days to go through the stress period. Then, the cryoinjured hEHT was transferred to a 3.5 cm confocal dish with the hEHT medium. For the prolonged recording of the movement trajectory of the GFP-labeled CM and the mCherry-labeled but not GFP-labeled CF during the scar-forming process of cryoinjured hEHT, the 3.5 cm confocal dish was placed in the Heating Insert P Lab-Tek™ S compact (PECON) with the CO₂-Cover HP (PECON) for providing stable culture temperature and CO₂ concentration using CO₂ control device (Zeiss, CO₂ Module S) and temperature control unit (Zeiss, TempModule S). The whole device was assembled into the confocal system (Zeiss LSM980). The GFP/mCherry/phase-contrast images of the scar-forming region were taken every 30 minutes until the scar-forming was complete.

Immunofluorescence staining

The cells cultured in 3.5 cm confocal dishes or the hEHTs were fixed with 4% PFA for 30 minutes. After washing twice with DPBS, the samples were blocked and permeabilized in DPBS supplemented with 0.2% Triton X-100 and 5% donkey serum at 4 °C overnight. The following primary antibodies were used in this study: Vimentin (Abcam, 92547, 1:200), Cardiac Troponin T (Abcam, ab45932, 1:200), Cardiac Troponin I (Abcam, ab56357, 1:100), Transgelin (Abcam, ab10135, 1:200), Collagen I (Invitrogen, MA1-26771, 1:100), Fibronectin (Abcam, ab2413, 1:100), Sarcomeric α-actinin (Sigma-Aldrich, A7811, 1:200), Alpha smooth muscle actin (Abcam, ab7817, 1:200), Tie-1 (Proteintech, 19329-1-AP, 1:100), WT1 (ABclonal, A17006, 1:200), Fibroblasts antibody TE7 (Sigma-Aldrich, CBL271, 1:200), CD44 (Cell Signaling Technology, #3570, 1:200), CXADR (Invitrogen, MA5-29208, 1:100), MYH (Abcam, ab207926, 1:200). The samples were incubated with these primary antibodies at 4 °C overnight, followed by the species-appropriate AlexaFluor (Invitrogen) secondary antibodies (1:400 dilution) for 2 hours at room temperature. Dyes such as Cell-ROX (Thermo Fisher, C10422), Hoechst (Thermo Fisher, H21492), and TdT-mediated dUTP Nick End Labeling (TUNEL) apoptosis detection kit (Yeasen, 40307ES50) were also used according to manufacturer’s instructions in this study. After washing twice with DPBS, the confocal images were taken by Zeiss LSM980 and processed using ImageJ. For calculating the percentage of CM volume (P_cm) in the remodeled area of cryoinjured hEHT (Supplementary Fig. 14b), the calculation formula is

$${{{\rm{P}}}}_{{{\rm{cm}}}}=\frac{\left\{\left({{{\rm{Area}}}}_{{{\rm{c}}}}-{{{\rm{Area}}}}_{1}\right)+\cdots+\left({{{\rm{Area}}}}_{{{\rm{c}}}}-{{{\rm{Area}}}}_{{{\rm{n}}}-1}\right)+\left({{{\rm{Area}}}}_{{{\rm{c}}}}-{{{\rm{Area}}}}_{{{\rm{n}}}}\right)\right\}\times {{{\rm{Z}}}}_{{{\rm{n}}}}/({{\rm{n}}}-1)}{{{{\rm{Area}}}}_{{{\rm{c}}}}\times {{{\rm{Z}}}}_{{{\rm{n}}}}}$$

(1)

Optical mapping of hEHT

The calcium activity of remodeled hEHT was recorded as the established method⁹¹. Briefly, the hEHT were incubated with 2.5 μM Rhod2 at 37 °C for 20 minutes in the hEHT medium. After washing twice with DPBS, 5 μM blebbistatin (Sigma-Aldrich. B0560) was added into the hEHT medium to eliminate motion artifacts deriving from CM’s contractions during recordings. The remodeled hEHTs under different treatments were stimulated (8–10 V) by a bipolar platinum point-electrode at 1 Hz, and 1 or 1.5 Hz with 1 μM isoproterenol (Sigma-Aldrich, I0599990) supplemented in the hEHT medium. The high-speed imaging system (SciMedia, MiCAM02) with the 530 nm LED illumination system (SciMedia, LEX2-LE4-G) was applied to acquire 2048 frames of calcium activity images over 10 seconds for each recording.

Apical cryoinjury model in neonatal mice

This protocol was performed as previously described¹⁶ and approved by the Hubei University Animal Ethics and Welfare Committee (20240066). Mice were housed in a pathogen-free room with a 12-hour light/12-hour dark cycle, a temperature of 24 °C ± 2 °C, and a relative humidity of 40-70%. Briefly, the neonatal mice at 3 days of postnatal (P3) were placed on ice for 3–4 minutes. Once anesthetized, they were immediately transferred to the precooled operating table and taped down to secure the body. The left thoracotomy was performed for exposure of the heart between the 3rd and 4th intercostal space. A 1-mm diameter copper probe precooled in liquid nitrogen was applied to the left ventricle near the apex for 8-12 seconds to induce transmural cryoinjury. The chest wall and skin were progressively closed with 7-0 polypropylene sutures. After surgery, mice were placed on the 37°C-heating pad for recovery and randomly divided into five groups. Y27632, CHIR99021, and SB431542 were dissolved in DMSO at 5 mg of the small molecule per 1 kg body weight (5 mg/kg) and injected subcutaneously into mice after diluting in corn oil. For combination treatment, Y27632, CHIR99021, and SB431542 were injected subcutaneously on days 1, 5, and 9 after cryoinjury. For single-factor treatment, only a small molecule was injected subcutaneously once on day 1, day 5, or day 9 after cryoinjury, and DMSO was injected at the remaining time points. The vehicle group used Equal amounts of DMSO at all three injection times. Mice were sacrificed after 12 days of treatment, and their hearts were excised for further histological analysis (n = 9 mice for the vehicle; n = 7 mice for Y27632; n = 6 for CHIR99021 and SB431542; n = 9 mice for combination).

Mouse myocardial infarction model

Animal protocols were approved by the Hubei University Animal Ethics and Welfare Committee (20220053), and animal experiments were performed based on them. All procedures involving animals were performed according to the NIH guidelines. A mouse myocardial infarction model was performed as previously described⁹². Adult 8–9-week-old male C57/BL6 mice were anesthetized by intraperitoneal administration of tribromoethanol (400 mg/kg), intubated, and mechanically ventilated with a rodent respirator. The left anterior descending (LAD) coronary artery was permanently occluded using a 7-0 polypropylene suture. The chest was closed sequentially suturing rib and skin with a 5-0 suture. For non-infarcted controls, mice underwent the sham operation where the ligature around the left anterior descending coronary artery was not tied. Mice recovered from anesthesia under warm conditions with normal ventilation. One day after the MI model, the cardiac function was evaluated by echocardiography with the VINNO V6 Vet system. Infarcted mice were averagely divided into five groups: Vehicle (DMSO), Y27632, CHIR99021, SB431542, Y27632 + CHIR99021 + SB431542 (Combination) group, according to echo data with the comparable LVEF and LVFS. There were two types of drug administration methods for MI mice: 1) Sustained single-factor treatment: 3 mg/kg Y27632, 3 mg/kg CHIR99021, or 3 mg/kg SB431542 dissolved in DMSO were injected intraperitoneally every other day for three weeks in each single drug group. 2) Phased single-factor treatment, 5 mg/kg Y27632, or 5 mg/kg CHIR99021, or 5 mg/kg SB431542 were intraperitoneally injected once on day 0, day 7, or day 14 post-MI, and DMSO was injected at the remaining time points. The vehicle group used equal amounts of DMSO at all three injection times. For the combination treatment, these drugs were intraperitoneally injected one drug per week (Y27632 for the first week, CHIR99021 for the second week, and SB431542 for the third week). The vehicle group was treated with equivalent DMSO. After three weeks, the cardiac function of survival animals was re-evaluated by echocardiography, then animals were sacrificed, and hearts were excised for other analyses.

Histological analysis

The harvested hearts were fixed in 4% paraformaldehyde for 24-36 hours at 4 °C, halved 1-2 mm above the ligature knot, dehydrated in an ethanol and xylene series, and embedded in paraffin. For the MI mouse model, hearts were continuously sectioned at 5 μm thickness from the ligation site to the apex at an interval of 300 μm between each section group. For the cryoinjured mice, hearts were serially sectioned at 6 μm thickness from the cryoinjury site to the apex with a 150 μm interval between each section group. Three section groups were collected from each heart. Masson’s trichrome was performed using standard procedures⁹². Fibrotic area was calculated as the average percentage of collagen fiber area in each of the three left ventricular sections to the total area of the left ventricular wall. The percentage of the collagen fiber area was identified and calculated using the machine learning-based tool FibroSoft⁹³ and Image J software.

Optical mapping in rats and mice with myocardial infarction

All the rat animal experiments were performed in Henan Scope Research Institute of electrophysiology under the experimental animal usage guidelines of Henan Provincial Experimental Animal Center and were approved by the Animal Ethics Committee of Scope Research Institute of electrophysiology (SGll220821044). Similar to the methods described before⁹⁴, the 6–8-week-old male SD rats were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The adult male rats weighing 200-250 g were anesthetized with isoflurane (RWD, R510-22-10) and fixed in a supine position on the operating table. The small animal ventilator (ZSLAB, ZS-MV-HX) was used for maintaining their breathing, and the electrocardiogram leads (TECHMAN, BL-420N) were connected to monitor the rat’s electrocardiogram in real-time. After depilation and disinfection of the precordial area, a 2 cm incision was made at the second and third ribs of the sternum. The skin and muscle tissue were bluntly separated to expose the heart. The left anterior descending branch under the left auricle and pulmonary artery cone was ligated with a 5-0 nylon thread to cause acute ischemia. ST-segment elevation in the electrocardiogram represents successful ligation. The intercostal space, muscle, and skin were sutured sequentially, and the skin was disinfected with iodophor and sutured. After 4 weeks of feeding, the rats were anesthetized and euthanized to remove the heart. The heart was placed on the Langendorff perfusion apparatus (Scope, SG-LP011) and the aorta was retrogradely perfused with K-H solution containing 119 mM NaCl, 4 mM KCl, 1.8 mM CaCl₂-2H₂O, 1 mM MgCl₂-6H₂O, 1.2 mM KH₂PO₄, 25 mM NaHCO₃ and 10 mM glucose at 37 °C and 10 mL/min flow rate. After 15 minutes of stabilization, 1 mg/mL blebbistatin (Abcam, ab120425) was perfused into the heart to eliminate motion artifacts. Intracellular calcium changes were measured using 1 mg/mL RHOD-2AM (Abcam, ab142780). The entire heart was illuminated with 549 nm excitation light (MappingLab, LEDC-2001), and CMOS cameras (MappingLab, OMS-PCIE-2002) were used to capture the fluorescent signal emitted by the heart. The intracellular calcium signal under 5 Hz stimulation (stimulation at the apex of the heart with twice the threshold current) was recorded. For optical mapping in the heart of drug-treated MI mice, the main difference compared to infarcted rats was that the perfusion flow rate was reduced to 2.5 mL/min, and calcium signaling activity was recorded under four loading conditions (8 or 20 Hz electrical pacing with or without 1 μM isoproterenol).

Processing of data from optical mapping

The commercially available analysis software OMapScope 4.0 (MappingLab, Version 5.7.8) and BV-Ana Analyzer (BrainVision, Version 16.04.20) were applied for processing the optical mapping data from rat MI hearts and cryoinjured hEHTs. Gaussian spatial filtering (3 × 3 pixels) and median filtering (3 × 3 pixels) were used for the spatial alignment and processing of the calcium signal.

As shown in Fig. 2f, after normalizing using the background fluorescence(F/F_b), the maximum upstroke time or recovery time was determined as the time corresponding to the maximum rate of rise of the calcium signal depolarization or the maximum rate of decline of the calcium signal repolarization. The maximum depolarization slope or maximum repolarization slope was then defined as the maximum upstroke rate or maximum recovery rate. Calcium signal amplitude, and duration at 50% (CSD50) or 80% (CSD80) repolarization were also calculated. The maximum upstroke time was regarded as the excitation time of each pixel, and the spatial time arrays of a total of 20 × 20 pixels from the uninjured area or remodeled area in rat MI hearts and cryoinjured hEHTs were extracted as single ROI and further analyzed in the software Arrow Map Tool (Mapping Lab, Version 1.3) to evaluated conduction velocity (CV). The coefficient of variation of all CV in the 20 × 20-pixel matrix was calculated to evaluate the velocity value homogeneity. To evaluate the conduction direction homogeneity, first, calculate the average vector of the propagation direction of 400 pixels. Then, calculate the cosine similarity between the propagation direction of each of the 400 points and the average vector of the propagation direction in turn. Finally, take the average of the cosine similarity (cosθ) to evaluate whether the propagation direction is consistent or not (homogeneity level). The closer the average cosθ is to 1, the more consistent the propagation direction of all pixels in ROI. The closer the average cosθ is to 0 or even −1, the more inconsistent the direction. For developing the integrative scoring system, the average value of the calcium activity characteristics parameters in the uninjured areas of the control group was regarded as the criterion to rank the effeteness (average value of the corresponding parameter) of different treatments on functional recovery.

Bulk RNA sequencing

The hEHT and corresponding controls were collected on days 2 and 4 after cryoinjury and their total RNA was extracted using the RNeasy Fibrous Tissue Mini Kit (Qiagen 74704) according to the manufacturer’s instructions. Each group had 3 biological replicates. The cDNA libraries were constructed using KAPA Hyper Prep Kits (Roche KK8504) and sequenced on the Illumina Novaseq platform using the double-end sequencing (Paired-End) PE150 method. For bulkRNA-seq targeting CM and CF in treated hEHT (10 μM Y27632 on days 0–2 after cryoinjury, 5 μM CHIR99021 on days 2-4, 5 μM SB431542 on days 4-6) and corresponding control hEHT, these hEHTs were dissociated with 2 mg/ml collagenase type II and IV at 37 °C for 1 hour, followed by two washes with high-glucose DMEM. Cell clusters were further digested with 0.25% trypsin-EDTA at 37 °C for 4 minutes. After collecting cell precipitates by centrifugation at 300 × g for 5 minutes, these precipitates were resuspended in hEHT medium and cultured in uncoated cell culture plates. After 1 hour, unattached cells (CM) and attached cells (CF) in the plate were collected separately using TRIZOL for further cDNA library construction and sequencing. Raw reads were trimmed adapter sequences and removed low-quality reads by cutadapt (v1.10), and mapped to the GRCh37/hg19 reference genome by tophat2 (v2.0.13). After removing obvious outliers in each group via PCA analysis, the differentially expressed genes with log2|fold-change | ≥ 1 and P < 0.01 between cryoinjured hEHT and control hEHT were selected for Gene Ontology Biological Processes and KEGG pathway analysis via The Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/).

Single-cell RNA sequencing

The cryoinjured hEHTs under the early and late period of the remodeling process were dissociated by 2 mg/mL collagenase II (Sigma-Aldrich, C6885), and 2 mg/mL collagenase IV (Sigma-Aldrich, C5138) for 1 hour. After collagenases were removed, the dispersed hEHTs were further digested into single cells with 0.25% trypsin-EDTA at 37°C. Fluorescence-activated cell sorting was applied to obtain a single cell suspension, with a high percentage of live cells above 80%. Single-cell RNA-seq libraries were prepared according to the manufacturer’s instructions (Chromium Single Cell 3’ Reagent Kits User Guide (v3.1 Chemistry Dual Index)). Briefly, cells at a limiting dilution, barcoded single-cell gel beads, and partitioning oil were combined onto chromium next GEM chip to generate Gel Beads-in-emulsion (GEMs). Then, the gel beads were dissolved and mixed with a master mix containing reverse transcription reagents. Followed by demulsification, cDNA library construction was performed accordingly. All libraries prepared for this study were sequenced on NovaSeq 6000. Three biological replicates in each treatment condition were dissociated and split into 4 fractions to be sequenced in equal parts. The qualitative analysis of the sequenced data was performed first. The Phred quality score (Q) was defined as a property logarithmically related to the base-calling error probabilities (P) to measure the quality of each read base during the sequencing process. The formular was: Q = −10 log₁₀^P. The sequencing quality scores are shown in Supplementary Table 1.

scRNA-seq data processing

Data were generated and aggregated to the human genome (GRCh37, ftp://ftp.ensembl.org/pub/grch37/release-84/gtf/homo_sapiens/Homo_sapiens.GRCh37.82.gtf) by CellRanger software (10x Genomics) and raw data were further processed using the Seurat package in R⁹⁵. 35230 high-quality cells were extracted to identify distinct cell populations and subsequent downstream analyses. The following standards were adopted to perform the quality control process: (1) genes expressed by less than 3 cells were excluded; (2) cells that expressed fewer than 400 genes or the total number of molecules detected within a cell fewer than 1,000 were removed because it may be a dead or an empty droplet; (3) cells that expressed over 9,000 genes or the total number of molecules detected within a cell more than 10,000 were also excluded because it indicated that the “cell” could be a doublet or multiplet; (4) cells with more than a quarter of unique molecular identifiers (UMIs) were derived from the mitochondrial genome were not considered as well. The data were normalized using the “NormalizeData” function and further scaled using the “ScaleData” function of the Seurat package. To perform the linear dimensional reduction of this dataset, we first detected 2,000 highly variable genes using the “vst” method of the Seurat FindVariableGenes function, and then the selected genes were summarized by principal component analysis (PCA).

Identification of cell types and differentially expressed genes

The two-dimensional uniform manifold approximation and projection (UMAP) was calculated using the “RunUMAP” function and Graph-based clustering was conducted on cluster cells according to their gene expression profile using the “FindClusters” function of the Seurat package. Cell clusters in the resulting two-dimensional UMAP were annotated to known cell types based on previously reported marker genes³³, including DLK1, COL1A2, PDGFRA, GSN, LUM, COL1A1, DCN for CF, TBX18, WT1, ALDH1A2, UPK3B for epicardial cell, ERG, SOX7, COL4A1, GNG11, TIE1, CDH5, EMCN, RAMP2 for endothelial cell, and MYH6, PLN, MYL7, TTN, ENO3, NIK2-5, ACTN2, TNNC1, TNNT2 for CM. Additionally, we used the SingleR package⁹⁶ to automatically annotate the cell types. Cells expressing two or more markers simultaneously were removed. Second-level clustering was applied to both CM and CF, using 20 dimensions for clustering. A clustering resolution of 0.22 was used for CF, while a resolution of 0.5 was applied for CM. UMAP was utilized to visualize the result.

Differentially expressed gene analysis

Differentially expressed genes (DEG) of different cell clusters, as well as those of different samples, were identified using the “FindMarkers” function of the Seurat package. DEGs were selected only if the gene had an average log fold change 0.4 higher than the average expression in the other clusters and the adjusted p-value less than 0.05. Heatmaps were performed to show the DEGs of different cell clusters.

Transcription factors (TFs) analysis

For constructing gene regulatory networks, we used the pySCENIC package (version 0.9.1) with default parameters⁹⁷. Pre-processed, normalized data from Seurat were used to identify enriched TFs in different cell types from different groups. The GRNboost2 algorithm was used to infer gene-gene co-expression relationships between TFs and their potential targets. The regulon prediction step was performed using the databases from https://resources.aertslab.org/cistarget. Each regulon contained a TF and its target genes enriched for the TF’s motifs. The AUCell algorithm assessed the activity scores of each regulon within single cells.

Pseudotime trajectory inference

The Monocle R package (version 2.30.0) was used for the developmental trajectory inference⁹⁸. The analysis began by separating the cell clusters of interest using the “subset” command in Seurat. A CellDataSet object was then created using the “newCellDataSet” function in Monocle. After calculating size factors and estimating dispersions, genes were selected for pseudotime definition in a semi-supervised manner. Specifically, genes were filtered based on mean expression level and variability across cells, using the following criteria: mean expression >= 0.5 and dispersion empirical >= 1.25 * dispersion fit. Dimension reduction was performed using the “reduceDimension” function with the “DDRTree” method. Cells were ordered in pseudotime, and visualization was performed using the “plot cell trajectory” plot genes in “pseudotime” and “plot genes branched heatmap” functions.

Cell–cell interaction analysis

CellPhoneDB (version 5.0) was applied to study cell-cell interactions in different cell types from different groups⁹⁹. Pre-processed and normalized single-cell RNA sequencing data from Seurat were used as input. We used the latest ligand-receptor interaction database from CellPhoneDB (GitHub—ventolab/cellphonedb-data) to evaluate cellular interaction networks. Interactions were visualized using the kplots package.

Statistics and reproducibility

Each experiment contained at least 3 biological replicates or as indicated in the figure legends. GraphPad Prism 9 was used for statistical analysis. All data were reported as means± standard error of the mean (SEM). Experiments involving two groups were analyzed using unpaired two-tailed Student’s t-test. Sidak’s, Tukey’s, or Dunnett’s multiple comparisons were otherwise assessed by ANOVA and indicated in the corresponding Figure legend. p value less than 0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The raw data including bulk RNA sequencing and single-cell RNA sequencing related to the remodeling process generated in this study have been deposited in NCBI’s Gene Expression Omnibus (GEO) under the accession number GSE234717. Raw data for bulk RNA sequencing analysis that revealed the therapeutic mechanism of Y27632/CHIR99021/SB431542 have been deposited in the Genome Sequence Archive under the accession number HRA007850. The single-cell RNA sequencing dataset of human failing heart referenced in this study is available at GSE145154. All data needed to evaluate the conclusions in the paper are present within the paper and/or the Supplementary Information. Source data is available for Fig. 1c, e, f, h; 2b, c, d, g, h; 5f; 6b, c, d; 7b, c, d, f, g; 8c, d, f, and Supplementary Figs. 1f, g; 2b, d, g, i; 3b, c, e, f; 6e; 13b, c, d; 14d, e, f, g, h, i; 15b, c, d, e; 18c; 19c, d, f; 20a, b, d, e, f in the associated source data file. Source data are provided with this paper.

References

Safiri, S. et al. Burden of ischemic heart disease and its attributable risk factors in 204 countries and territories, 1990–2019. Eur. J. Prev. Cardiol. 29, 420–431 (2022).
Article PubMed Google Scholar
Nowbar, A. N., Gitto, M., Howard, J. P., Francis, D. P. & Al-Lamee, R. Mortality from ischemic heart disease. Circ. Cardiovasc Qual. Outcomes 12, e005375 (2019).
Article PubMed PubMed Central Google Scholar
Pesce, M. et al. Cardiac fibroblasts and mechanosensation in heart development, health and disease. Nat. Rev. Cardiol. 20, 309–324 (2023).
Article CAS PubMed Google Scholar
Piamsiri, C., Maneechote, C., Siri-Angkul, N., Chattipakorn, S. C. & Chattipakorn, N. Targeting necroptosis as therapeutic potential in chronic myocardial infarction. J. Biomed. Sci. 28, 25 (2021).
Article CAS PubMed PubMed Central Google Scholar
de Bakker, J. M. et al. Reentry as a cause of ventricular tachycardia in patients with chronic ischemic heart disease: electrophysiologic and anatomic correlation. Circulation 77, 589–606 (1988).
Article PubMed Google Scholar
Holmes, J. W., Laksman, Z. & Gepstein, L. Making better scar: Emerging approaches for modifying mechanical and electrical properties following infarction and ablation. Prog. Biophys. Mol. Biol. 120, 134–148 (2016).
Article PubMed Google Scholar
Lian, X. et al. Directed cardiomyocyte differentiation from human pluripotent stem cells by modulating Wnt/beta-catenin signaling under fully defined conditions. Nat. Protoc. 8, 162–175 (2013).
Article CAS PubMed Google Scholar
Zhang, H. et al. Generation of quiescent cardiac fibroblasts from human induced pluripotent stem cells for in vitro modeling of cardiac fibrosis. Circ. Res. 125, 552–566 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Paik, D. T. et al. Large-scale single-cell RNA-Seq reveals molecular signatures of heterogeneous populations of human induced pluripotent stem cell-derived endothelial cells. Circ. Res. 123, 443–450 (2018).
Article CAS PubMed PubMed Central Google Scholar
Shadrin, I. Y. et al. Cardiopatch platform enables maturation and scale-up of human pluripotent stem cell-derived engineered heart tissues. Nat. Commun. 8, 1825 (2017).
Article ADS PubMed PubMed Central Google Scholar
Zhang, D. et al. Tissue-engineered cardiac patch for advanced functional maturation of human ESC-derived cardiomyocytes. Biomaterials 34, 5813–5820 (2013).
Article CAS PubMed PubMed Central Google Scholar
Lewis-Israeli, Y. R. et al. Self-assembling human heart organoids for the modeling of cardiac development and congenital heart disease. Nat. Commun. 12, 5142 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Hofbauer, P. et al. Cardioids reveal self-organizing principles of human cardiogenesis. Cell 184, 3299–3317.e3222 (2021).
Article CAS PubMed Google Scholar
Voges, H. K. et al. Development of a human cardiac organoid injury model reveals innate regenerative potential. Development 144, 1118–1127 (2017).
CAS PubMed Google Scholar
Richards, D. J. et al. Human cardiac organoids for the modelling of myocardial infarction and drug cardiotoxicity. Nat. Biomed. Eng. 4, 446–462 (2020).
Article CAS PubMed PubMed Central Google Scholar
Polizzotti, B. D., Ganapathy, B., Haubner, B. J., Penninger, J. M. & Kuhn, B. A cryoinjury model in neonatal mice for cardiac translational and regeneration research. Nat. Protoc. 11, 542–552 (2016).
Article CAS PubMed PubMed Central Google Scholar
Hirano, A. et al. Cryoinjury-induced acute myocardial infarction model and ameroid constrictor-induced ischemic heart disease model in adult micro-mini pigs for preclinical studies. Transl. Med. Commun. 2, 1 (2017).
Article ADS Google Scholar
Peoples, J. N., Saraf, A., Ghazal, N., Pham, T. T. & Kwong, J. Q. Mitochondrial dysfunction and oxidative stress in heart disease. Exp. Mol. Med. 51, 1–13 (2019).
Article CAS PubMed Google Scholar
Talman, V. & Ruskoaho, H. Cardiac fibrosis in myocardial infarction-from repair and remodeling to regeneration. Cell Tissue Res. 365, 563–581 (2016).
Article CAS PubMed PubMed Central Google Scholar
Liu, M., Lopez de Juan Abad, B. & Cheng, K. Cardiac fibrosis: Myofibroblast-mediated pathological regulation and drug delivery strategies. Adv. Drug Deliv. Rev. 173, 504–519 (2021).
Article CAS PubMed PubMed Central Google Scholar
Bowers, S. L. K., Meng, Q. & Molkentin, J. D. Fibroblasts orchestrate cellular crosstalk in the heart through the ECM. Nat. Cardiovas. Res. 1, 312–321 (2022).
Article Google Scholar
Noordman, A. B. P. et al. Myocardial scar characterization and future ventricular arrhythmia in patients with ischemic cardiomyopathy and an implantable cardioverter-defibrillator. Front Cardiovasc Med. 8, 708406 (2021).
Article PubMed PubMed Central Google Scholar
Bui, A. H. & Waks, J. W. Risk stratification of sudden cardiac death after acute myocardial infarction. J. Innov. Card. Rhythm Manag. 9, 3035–3049 (2018).
Article PubMed PubMed Central Google Scholar
Sun, J. et al. Targeting the metastasis suppressor, NDRG1, using novel iron chelators: regulation of stress fiber-mediated tumor cell migration via modulation of the ROCK1/pMLC2 signaling pathway. Mol. Pharm. 83, 454–469 (2013).
Article CAS Google Scholar
Soliman, H. A. N., Toso, E. A., Darwish, I. E., Ali, S. M. & Kyba, M. Antiapoptotic Protein FAIM2 is targeted by miR-3202, and DUX4 via TRIM21, leading to cell death and defective myogenesis. Cell Death Dis. 13, 405 (2022).
Article CAS PubMed PubMed Central Google Scholar
Frangogiannis, N. G. Cardiac fibrosis. Cardiovasc Res. 117, 1450–1488 (2021).
Article CAS PubMed Google Scholar
Guo, Y. & Pu, W. T. Cardiomyocyte maturation: new phase in development. Circ. Res. 126, 1086–1106 (2020).
Article CAS PubMed PubMed Central Google Scholar
Cagavi, E. et al. Functional cardiomyocytes derived from Isl1 cardiac progenitors via Bmp4 stimulation. PLoS One 9, e110752 (2014).
Article ADS PubMed PubMed Central Google Scholar
Litvinukova, M. et al. Cells of the adult human heart. Nature 588, 466–472 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Mohamed, T. M. A. et al. Regulation of cell cycle to stimulate adult cardiomyocyte proliferation and cardiac regeneration. Cell 173, 104–116.e112 (2018).
Article CAS PubMed PubMed Central Google Scholar
Rao, M. et al. Resolving the intertwining of inflammation and fibrosis in human heart failure at single-cell level. Basic Res. Cardiol. 116, 55 (2021).
Article ADS PubMed Google Scholar
Ko, T. et al. Cardiac fibroblasts regulate the development of heart failure via Htra3-TGF-beta-IGFBP7 axis. Nat. Commun. 13, 3275 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Cui, Y. et al. Single-cell transcriptome analysis maps the developmental track of the human heart. Cell Rep. 26, 1934–1950.e1935 (2019).
Article CAS PubMed Google Scholar
Quijada, P. et al. Coordination of endothelial cell positioning and fate specification by the epicardium. Nat. Commun. 12, 4155 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Humeres, C. et al. Smad7 effects on TGF-beta and ErbB2 restrain myofibroblast activation and protect from postinfarction heart failure. J. Clin. Invest. 132, e146926 (2022).
Article CAS PubMed PubMed Central Google Scholar
Knight-Schrijver, V. R. et al. A single-cell comparison of adult and fetal human epicardium defines the age-associated changes in epicardial activity. Nat. Cardiovasc. Res. 1, 1215–1229 (2022).
Article PubMed PubMed Central Google Scholar
Quijada, P., Trembley, M. A. & Small, E. M. The role of the epicardium during heart development and repair. Circ. Res. 126, 377–394 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhou, B. & Pu, W. T. Epicardial epithelial-to-mesenchymal transition in injured heart. J. Cell Mol. Med. 15, 2781–2783 (2011).
Article CAS PubMed PubMed Central Google Scholar
Cahill, T. J., Choudhury, R. P. & Riley, P. R. Heart regeneration and repair after myocardial infarction: translational opportunities for novel therapeutics. Nat. Rev. Drug Discov. 16, 699–717 (2017).
Article CAS PubMed Google Scholar
van Deel, E. D. et al. Extracellular superoxide dismutase protects the heart against oxidative stress and hypertrophy after myocardial infarction. Free Radic. Biol. Med. 44, 1305–1313 (2008).
Article PubMed Google Scholar
Ye, Y. et al. PON2 ameliorates Ang II-induced cardiomyocyte injury by targeting the CANX/NOX4 signaling pathway. Immun. Inflamm. Dis. 11, e765 (2023).
Article CAS PubMed PubMed Central Google Scholar
Li, C. et al. PXDN reduces autophagic flux in insulin-resistant cardiomyocytes via modulating FoxO1. Cell Death Dis. 12, 418 (2021).
Article CAS PubMed PubMed Central Google Scholar
Canale, P. et al. Role of miR-133/Dio3 Axis in the T3-dependent modulation of cardiac mitoK-ATP Expression. Int J. Mol. Sci. 23, 6549 (2022).
Article CAS PubMed PubMed Central Google Scholar
Leyns, L., Bouwmeester, T., Kim, S. H., Piccolo, S. & De Robertis, E. M. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer. Cell 88, 747–756 (1997).
Article CAS PubMed PubMed Central Google Scholar
Lee, H. & Evans, T. TMEM88 inhibits Wnt signaling by promoting Wnt signalosome localization to multivesicular bodies. iScience 19, 267–280 (2019).
Article ADS PubMed PubMed Central Google Scholar
Heusch, G. et al. STAT5 activation and cardioprotection by remote ischemic preconditioning in humans: short communication. Circ. Res. 110, 111–115 (2012).
Article CAS PubMed Google Scholar
Cui, M. et al. Transcription factor NFYa controls cardiomyocyte metabolism and proliferation during mouse fetal heart development. Dev. Cell 58, 2867–2880.e2867 (2023).
Article CAS PubMed PubMed Central Google Scholar
Milia, J. et al. Farnesylated RhoB inhibits radiation-induced mitotic cell death and controls radiation-induced centrosome overduplication. Cell Death Differ. 12, 492–501 (2005).
Article CAS PubMed Google Scholar
Li, M. J. et al. Upregulation of CCNB2 and its perspective mechanisms in cerebral ischemic stroke and all subtypes of lung cancer: a comprehensive study. Front Integr. Neurosci. 16, 854540 (2022).
Article CAS PubMed PubMed Central Google Scholar
Wang, X. et al. LOX-1 in the maintenance of cytoskeleton and proliferation in senescent cardiac fibroblasts. J. Mol. Cell Cardiol. 60, 184–190 (2013).
Article CAS PubMed Google Scholar
Zou, Y. et al. DNA Hypermethylation of CREB3L1 and Bcl-2 associated with the mitochondrial-mediated apoptosis via PI3K/Akt pathway in human BEAS-2B cells exposure to silica nanoparticles. PLoS One 11, e0158475 (2016).
Article PubMed PubMed Central Google Scholar
Westendorp, B. et al. E2F7 represses a network of oscillating cell cycle genes to control S-phase progression. Nucleic Acids Res. 40, 3511–3523 (2012).
Article CAS PubMed Google Scholar
Liao, R. et al. E2F transcription factor 1 (E2F1) promotes the transforming growth factor TGF-beta1 induced human cardiac fibroblasts differentiation through promoting the transcription of CCNE2 gene. Bioengineered 12, 6869–6877 (2021).
Article CAS PubMed PubMed Central Google Scholar
Holowiecki, A. et al. Pbx4 limits heart size and fosters arch artery formation by partitioning second heart field progenitors and restricting proliferation. Development 147, dev185652 (2020).
Article CAS PubMed PubMed Central Google Scholar
Stellato, M. et al. The AP-1 transcription factor Fosl-2 drives cardiac fibrosis and arrhythmias under immunofibrotic conditions. Commun. Biol. 6, 161 (2023).
Article CAS PubMed PubMed Central Google Scholar
Teekakirikul, P. et al. Cardiac fibrosis in mice with hypertrophic cardiomyopathy is mediated by non-myocyte proliferation and requires Tgf-beta. J. Clin. Invest 120, 3520–3529 (2010).
Article CAS PubMed PubMed Central Google Scholar
Zhao, G. J., Yin, K., Fu, Y. C. & Tang, C. K. The interaction of ApoA-I and ABCA1 triggers signal transduction pathways to mediate efflux of cellular lipids. Mol. Med. 18, 149–158 (2012).
Article CAS PubMed Google Scholar
Morikawa, Y., Heallen, T., Leach, J., Xiao, Y. & Martin, J. F. Dystrophin-glycoprotein complex sequesters Yap to inhibit cardiomyocyte proliferation. Nature 547, 227–231 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Sorensen, D. W. & van Berlo, J. H. The role of TGF-beta signaling in cardiomyocyte proliferation. Curr. Heart Fail Rep. 17, 225–233 (2020).
Article PubMed PubMed Central Google Scholar
Zhou, J. et al. The long noncoding RNA THBS1-AS1 promotes cardiac fibroblast activation in cardiac fibrosis by regulating TGFBR1. JCI Insight 8, e160745 (2023).
Article PubMed PubMed Central Google Scholar
Wang, Y. et al. Low density lipoprotein receptor related protein 6 (LRP6) protects heart against oxidative stress by the crosstalk of HSF1 and GSK3beta. Redox Biol. 37, 101699 (2020).
Article CAS PubMed PubMed Central Google Scholar
Yang, H. et al. Deriving waveform parameters from calcium transients in human iPSC-derived cardiomyocytes to predict cardiac activity with machine learning. Stem Cell Rep. 17, 556–568 (2022).
Article CAS Google Scholar
Peck, K. A. et al. Extracellular vesicles secreted by TDO2-augmented fibroblasts regulate pro-inflammatory response in macrophages. Front Cell Dev. Biol. 9, 733354 (2021).
Article PubMed PubMed Central Google Scholar
Kaur, A. et al. ChaC2, an enzyme for slow turnover of cytosolic glutathione. J. Biol. Chem. 292, 638–651 (2017).
Article CAS PubMed Google Scholar
Pei, H. et al. G protein-coupled estrogen receptor 1 inhibits angiotensin II-induced cardiomyocyte hypertrophy via the regulation of PI3K-Akt-mTOR signalling and autophagy. Int J. Biol. Sci. 15, 81–92 (2019).
Article CAS PubMed PubMed Central Google Scholar
Duval, V., Alayrac, P., Silvestre, J. S. & Levoye, A. Emerging roles of the atypical chemokine receptor 3 (ACKR3) in cardiovascular diseases. Front Endocrinol. (Lausanne) 13, 906586 (2022).
Article PubMed Google Scholar
Cadenas, S. Mitochondrial uncoupling, ROS generation and cardioprotection. Biochim Biophys. Acta Bioenerg. 1859, 940–950 (2018).
Article CAS PubMed Google Scholar
Cai, W. et al. Alox15/15-HpETE aggravates myocardial ischemia-reperfusion injury by promoting cardiomyocyte ferroptosis. Circulation 147, 1444–1460 (2023).
Article CAS PubMed Google Scholar
Voros, I. et al. Somatostatin and its receptors in myocardial ischemia/reperfusion injury and cardioprotection. Front Pharm. 12, 663655 (2021).
Article CAS Google Scholar
Natarajan, N. et al. Complement receptor C5aR1 plays an evolutionarily conserved role in successful cardiac regeneration. Circulation 137, 2152–2165 (2018).
Article CAS PubMed PubMed Central Google Scholar
Lindsey, M. L. Assigning matrix metalloproteinase roles in ischaemic cardiac remodelling. Nat. Rev. Cardiol. 15, 471–479 (2018).
Article PubMed PubMed Central Google Scholar
Meng, F. & Martin, J. F. Embryonic ECM Protein SLIT2 and NPNT promote postnatal cardiomyocyte cytokinesis. Circ. Res. 127, 908–910 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhou, Y. et al. Single-cell transcriptomic analyses of cell fate transitions during human cardiac reprogramming. Cell Stem Cell 25, 149–164.e149 (2019).
Article CAS PubMed PubMed Central Google Scholar
Asaumi, Y. et al. Protective role of endogenous erythropoietin system in nonhematopoietic cells against pressure overload-induced left ventricular dysfunction in mice. Circulation 115, 2022–2032 (2007).
Article CAS PubMed Google Scholar
Lam, N. T. & Sadek, H. A. Neonatal heart regeneration: comprehensive literature review. Circulation 138, 412–423 (2018).
Article PubMed PubMed Central Google Scholar
Pillai, I. C. L. et al. Cardiac fibroblasts adopt osteogenic fates and can be targeted to attenuate pathological heart calcification. Cell Stem Cell 20, 218–232.e215 (2017).
Article CAS PubMed Google Scholar
Giacomotto, J. & Segalat, L. High-throughput screening and small animal models, where are we? Br. J. Pharm. 160, 204–216 (2010).
Article CAS Google Scholar
Ubil, E. et al. Mesenchymal-endothelial transition contributes to cardiac neovascularization. Nature 514, 585–590 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Koenig, A. L. et al. Single-cell transcriptomics reveals cell-type-specific diversification in human heart failure. Nat. Cardiovasc Res. 1, 263–280 (2022).
Article PubMed PubMed Central Google Scholar
Kanisicak, O. et al. Genetic lineage tracing defines myofibroblast origin and function in the injured heart. Nat. Commun. 7, 12260 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Dong, L. Y., Qiu, X. X., Zhuang, Y. & Xue, S. Y-27632, a Rho-kinase inhibitor, attenuates myocardial ischemia-reperfusion injury in rats. Int J. Mol. Med. 43, 1911–1919 (2019).
CAS PubMed Google Scholar
Zhao, M. et al. Y-27632 preconditioning enhances transplantation of human-induced pluripotent stem cell-derived cardiomyocytes in myocardial infarction mice. Cardiovasc Res. 115, 343–356 (2019).
Article CAS PubMed Google Scholar
Xu, Y. et al. A transient wave of Bhlhe41(+) resident macrophages enables remodeling of the developing infarcted myocardium. Cell Rep. 42, 113174 (2023).
Article CAS PubMed Google Scholar
Mills, R. J. et al. Functional screening in human cardiac organoids reveals a metabolic mechanism for cardiomyocyte cell cycle arrest. Proc. Natl Acad. Sci. USA 114, E8372–E8381 (2017).
Article CAS PubMed PubMed Central Google Scholar
Ruan, H. et al. Single-cell reconstruction of differentiation trajectory reveals a critical role of ETS1 in human cardiac lineage commitment. BMC Biol. 17, 89 (2019).
Article PubMed PubMed Central Google Scholar
Hall, C. et al. Chronic activation of human cardiac fibroblasts in vitro attenuates the reversibility of the myofibroblast phenotype. Sci. Rep. 13, 12137 (2023).
Article ADS PubMed PubMed Central Google Scholar
Frangogiannis, N. G. Transforming growth factor-beta in myocardial disease. Nat. Rev. Cardiol. 19, 435–455 (2022).
Article CAS PubMed Google Scholar
Miyaoka, Y. et al. Isolation of single-base genome-edited human iPS cells without antibiotic selection. Nat. Methods 11, 291–293 (2014).
Article CAS PubMed PubMed Central Google Scholar
Li, L. et al. Generation of high-performance human cardiomyocytes and engineered heart tissues from extended pluripotent stem cells. Cell Discov. 8, 105 (2022).
Article CAS PubMed PubMed Central Google Scholar
Zhang, D. et al. Highly efficient differentiation of human ES cells and iPS cells into mature pancreatic insulin-producing cells. Cell Res. 19, 429–438 (2009).
Article ADS CAS PubMed Google Scholar
Park, S. J. et al. Insights into the pathogenesis of catecholaminergic polymorphic ventricular tachycardia from engineered human heart tissue. Circulation 140, 390–404 (2019).
Article CAS PubMed PubMed Central Google Scholar
Pei, J. et al. LPA(2) contributes to vascular endothelium homeostasis and cardiac remodeling after myocardial infarction. Circ. Res. 131, 388–403 (2022).
Article CAS PubMed Google Scholar
Remes, A. et al. Adapted clustering method for generic analysis of histological fibrosis staining as an open source tool. Sci. Rep. 13, 4389 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Yu, C. et al. An intrapericardial injectable hydrogel patch for mechanical-electrical coupling with infarcted myocardium. ACS Nano 16, 16234–16248 (2022).
Article CAS PubMed Google Scholar
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018).
Article CAS PubMed PubMed Central Google Scholar
Aran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 20, 163–172 (2019).
Article CAS PubMed PubMed Central Google Scholar
Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).
Article CAS PubMed PubMed Central Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Article CAS PubMed PubMed Central Google Scholar
Efremova, M., Vento-Tormo, M., Teichmann, S. A. & Vento-Tormo, R. CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes. Nat. Protoc. 15, 1484–1506 (2020).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

H.Z. was funded by the National Key R&D Program of China (2019YFA0110401). D.Z. was funded by the National Key R&D Program of China (2021YFA1101902, 2018YFA0109100) and the National Natural Science Foundation of China (32171107). We also thank Yan Qi and Shutao Xia from Hubei University, and William T. Pu from Boston Children’s Hospital for reviewing the manuscript.

Author information

These authors contributed equally: Pengcheng Yang, Lihang Zhu.

Authors and Affiliations

State Key Laboratory of Biocatalysis and Enzyme Engineering, School of Life Sciences, Hubei University, Wuhan, China
Pengcheng Yang, Shiya Wang, Jonathan Nimal Selvaraj, Zilong Li, Lin Cai & Donghui Zhang
Department of Biological Repositories, Frontier Science Center for Immunology and Metabolism, Medical Research Institute, Zhongnan Hospital of Wuhan University, Wuhan University, Wuhan, China
Lihang Zhu
Center of Translational Medicine, The First Affiliated Hospital, Zhongshan School of Medicine, Sun Yat‐Sen University, Guangdong, China
Jixing Gong
Shanghai Institute for Congenital Heart Diseases, Shanghai Children’s Medical Center, National Children’s Medical Center, Shanghai, China
Lincai Ye & Hao Zhang
Key Laboratory of Birth Defects and Related Diseases of Women and Children of MOE, Department of Pediatrics, West China Second University Hospital, Sichuan University, Chengdu, China
Hanxiao Chen & Yaoyao Zhang
Henan SCOPE Research Institute of Electrophysiology Co. Ltd., Kaifeng, China
Gongxin Wang
Beijing Geek Gene Technology Co. Ltd., Beijing, China
Wanjun Song
Cardiovascular Research Institute, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
Donghui Zhang

Authors

Pengcheng Yang
View author publications
You can also search for this author in PubMed Google Scholar
Lihang Zhu
View author publications
You can also search for this author in PubMed Google Scholar
Shiya Wang
View author publications
You can also search for this author in PubMed Google Scholar
Jixing Gong
View author publications
You can also search for this author in PubMed Google Scholar
Jonathan Nimal Selvaraj
View author publications
You can also search for this author in PubMed Google Scholar
Lincai Ye
View author publications
You can also search for this author in PubMed Google Scholar
Hanxiao Chen
View author publications
You can also search for this author in PubMed Google Scholar
Yaoyao Zhang
View author publications
You can also search for this author in PubMed Google Scholar
Gongxin Wang
View author publications
You can also search for this author in PubMed Google Scholar
Wanjun Song
View author publications
You can also search for this author in PubMed Google Scholar
Zilong Li
View author publications
You can also search for this author in PubMed Google Scholar
Lin Cai
View author publications
You can also search for this author in PubMed Google Scholar
Hao Zhang
View author publications
You can also search for this author in PubMed Google Scholar
Donghui Zhang
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

P.Y. and L.Z. contributed equally to this study. L.C., D.Z., and H.Z. conceptualized the study. P.Y., L.Z., J.G., L.C., and D.Z. designed the experiments; P.Y., L.Z., S.W., and J.G. performed experimental studies in hPSC culture, model establishment, immunostaining, and optical mapping. L.C., P.Y., S.W., and Z.L. performed the analysis of the therapeutic efficacy of different compounds in MI mice. L.Z., H.C., W.S., and Y.Z. designed and assisted in single-cell RNA sequencing analysis. G.W. guided the calcium activities study in myocardial infarction rats and mice. P.Y., L.Z., S.W, J.N.S., L.Y., Y.Z., and D.Z. analyzed the data of experiments and organized figures. P.Y., L.Z., J.N.S., L.Y., L.C., and D.Z. wrote or edited the manuscript.

Corresponding authors

Correspondence to Lin Cai, Hao Zhang or Donghui Zhang.

Ethics declarations

Competing interests

For the development of the hEHT injury model and its application in drug screening, we have been granted a Chinese patent, with inventors Donghui Zhang, Pengcheng Yang, and Jixing Gong, and patent number ZL 2020 1 0620164. X. Besides that, the authors declare no competing interests.

Peer review

Peer review information

: Nature Communications thanks Anindita Das and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Supplementary Movie 4

Supplementary Movie 5

Supplementary Movie 6

Supplementary Movie 7

Supplementary Movie 8

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Yang, P., Zhu, L., Wang, S. et al. Engineered model of heart tissue repair for exploring fibrotic processes and therapeutic interventions. Nat Commun 15, 7996 (2024). https://doi.org/10.1038/s41467-024-52221-9

Download citation

Received: 03 February 2024
Accepted: 30 August 2024
Published: 12 September 2024
DOI: https://doi.org/10.1038/s41467-024-52221-9
Springer Nature Limited

Engineered model of heart tissue repair for exploring fibrotic processes and therapeutic interventions

Abstract

Similar content being viewed by others

Introduction

Results

Establishment of the hEHT injury repair model in vitro

hEHT injury repair model exhibited arrhythmic-related electrophysiological malfunction post-fibrotic scar formation

hEHT injury repair model exhibited differential gene expression between the early and late periods of remodeling

hEHT injury repair model exhibited diversified cellular composition during remodeling

Signaling dynamic shift of different cell subpopulations in hEHT injury repair model

CF fate transition and interaction with CM in hEHT injury repair model

A mini hEHT-injury repair model for rapid screening of potential therapeutic targets for IHD

Multidimensional evaluation of the effect of different signaling modulators on mini hEHT remodeling outcomes

Integrative scoring system to evaluate the functional recovery in the remodeled area

BulkRNA-seq revealed the antifibrotic/cardiac repair effects of Y27632/CHIR99021/SB431542

MI mice exhibited cardiac functional recovery to Y27632, CHIR99021, and SB431542 to treatment in combination

Discussion

Methods

hPSC culture and cardiomyocyte differentiation

hEHT fabrication and culture

Development and assessment of the hEHT injury repair model

pLV-mCherry and TroponinT-GCaMP5-Zeo lentivirus production

Time-lapse imaging of the remodeling of cryoinjured hEHT

Immunofluorescence staining

Optical mapping of hEHT

Apical cryoinjury model in neonatal mice

Mouse myocardial infarction model

Histological analysis

Optical mapping in rats and mice with myocardial infarction

Processing of data from optical mapping

Bulk RNA sequencing

Single-cell RNA sequencing

scRNA-seq data processing

Identification of cell types and differentially expressed genes

Differentially expressed gene analysis

Transcription factors (TFs) analysis

Pseudotime trajectory inference

Cell–cell interaction analysis

Statistics and reproducibility

Reporting summary

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Navigation