Neutrophils facilitate the epicardial regenerative response after zebrafish heart injury

Abstract

Despite extensive studies on endogenous heart regeneration within the past 20 years, the players involved in initiating early regeneration events are far from clear. Here, we assessed the function of neutrophils, the first-responder cells to tissue damage, during zebrafish heart regeneration. We detected rapid neutrophil mobilization to the injury site after ventricular amputation, peaking at 1-day post-amputation (dpa) and resolving by 3 dpa. Further analyses indicated neutrophil mobilization coincides with peak epicardial cell proliferation, and recruited neutrophils associated with activated, expanding epicardial cells at 1 dpa. Neutrophil depletion inhibited myocardial regeneration and significantly reduced epicardial cell expansion, proliferation, and activation. To explore the molecular mechanism of neutrophils on the epicardial regenerative response, we performed scRNA-seq analysis of 1 dpa neutrophils and identified enrichment of the FGF and MAPK/ERK signaling pathways. Pharmacological inhibition of FGF signaling indicated its’ requirement for epicardial expansion, while neutrophil depletion blocked MAPK/ERK signaling activation in epicardial cells. Ligand-receptor analysis indicated the EGF ligand, hbegfa, is released from neutrophils and synergizes with other FGF and MAPK/ERK factors for induction of epicardial regeneration. Altogether, our studies revealed that neutrophils quickly motivate epicardial cells, which later accumulate at the injury site and contribute to heart regeneration.

Summary statement

Injury-site neutrophils associate with and stimulate epicardial cells to support mechanisms of zebrafish cardiac regeneration.

1. Introduction

Myocardial infarction creates irreparable damage by forming non-contractile, collagenous scar tissue at the injury site instead of new cardiac tissue, which contributes to heart failure development (Pfeffer and Braunwald, 1990; Jessup and Brozena, 2003; Kehat and Molkentin, 2010). Although mammalian hearts have previously been considered as a post-mitotic organ, recent reports indicate low levels of cardiomyocyte (CM) turnover occur throughout human life, but adult mammals cannot efficiently restore lost contractile tissues (Bergmann et al., 2009, 2015; Senyo et al., 2013; Lam and Sadek, 2018; Haubner et al., 2016). Currently, no regenerative therapy is available to treat or prevent human heart failure after ischemia. As adult zebrafish demonstrate a robust heart regenerative response to injury, we can examine their endogenous regenerative mechanisms to identify cell types and factors capable of benefiting heart regeneration, which may help develop regenerative therapy strategies to treat or prevent human heart failure.

After heart injury, alarmins, cytokines, chemokines, and damage-associated molecular patterns quickly mobilize neutrophils to the ischemic injury site (Ma et al., 2013; Ma, 2021; Bonaventura et al., 2016, 2020). Adverse cardiac repair has long been attributed to neutrophils due to their release of damaging reactive oxygen species, granules, and formation of neutrophil extracellular traps (NETs), which exacerbate ischemic injury by enhancing injury site inflammation, degrading the extracellular matrix, and damaging CMs (Ma et al., 2013; Ma, 2021; Bonaventura et al., 2016, 2020). However, clinical anti-neutrophil therapies fail to reduce infarct size (Vinten-Johansen, 2004), neutrophil depletion in murine myocardial infarction models contributes to heart failure development (Horckmans et al., 2017), and delayed neutrophil clearance in zebrafish cryoinjury models hinders myocardial recovery (Lai et al., 2017; Xu et al., 2019). These studies revealed a requirement for neutrophils in heart repair and regeneration by clearing apoptotic tissue and releasing pro-angiogenesis factors like VEGF (Horckmans et al., 2017; Ma, 2021; Taichman et al., 1997), suggesting a more complex role for these cells. A detailed mechanistic understanding of their role still needs to be determined, as there are few analyses of the neutrophil response within regeneration models.

Non-CM cardiac tissues have recently been demonstrated to be critical for heart regeneration after injury. One of these players is the epicardium, which envelops the outer surface of the heart, and rapidly proliferates and accumulates in the injury site after heart damage (Lepilina et al., 2006; Wang et al., 2015; Kikuchi et al., 2011a; Zhou et al., 2011a; Cai et al., 2019; Cao et al., 2017). Epicardial cells do not contribute to new CMs during heart development and regeneration (Kikuchi et al., 2011a), but are required for heart regeneration as genetic ablation of the epicardium blocks myocardial restoration and causes scar formation (Kikuchi et al., 2011a; Wang et al., 2013, 2015; Sun et al., 2022a; Cao and Poss, 2018). Critical epicardial subpopulations with roles in the facilitation of CM proliferation, coronary growth guidance, and extracellular matrix remodeling have also been identified through scRNA-seq analyses and functional analyses (Sun et al., 2022a, 2023). Although these studies reveal the essential role of the epicardium on successful heart repair/regeneration, the mechanism for regulating the early epicardial regenerative response remains unclear. As neutrophil mobilization and the epicardial response quickly occur in the wound, and leukocytes and epicardial cells associate with one another in the damaged mammalian heart (Huang et al., 2012; Ramjee et al., 2017; Pinto et al., 2014; Stevens et al., 2016; FitzSimons et al., 2020), we speculated that neutrophils may have a role in regulating the early epicardial regeneration response after heart injury.

In this study, we first examined neutrophils in zebrafish surgery resected hearts and detected quick neutrophil mobilization to the injury site with peak neutrophil recruitment at 1 dpa and resolution at 3 dpa. Pharmacological inhibition of neutrophil recruitment by the CDK9 inhibitor AT7519 impaired myocardial regeneration. We then found that epicardial expansion coincides with the dynamic neutrophil response and activated epicardial cells associate with neutrophils in the wound. After depleting neutrophils in the injury site, we observed blocked epicardial cell expansion, proliferation, and activation. Further, we performed scRNA-seq analysis of neutrophils at 1 dpa and identified clusters enriched with the FGF and MAPK/ERK signaling pathways. We then found that inhibiting the FGF signaling activity with antagonist blocked epicardial cell proliferation, while neutrophil depletion limited the activation of MAPK/ERK signaling in epicardial cells. Ligand-receptor analysis suggested that neutrophils secrete the EGF ligand heparin-binding EGF-like growth factor a (hbegfa) towards target epicardial cells, synergizing with other FGF and MAPK/ERK factors to induce epicardial regeneration. Altogether, our studies revealed that neutrophils are critical for the early epicardial regenerative response after heart damage.

2. Results

2.1. Neutrophil depletion inhibits cardiomyocyte proliferation during regeneration

Previous studies revealed rapid neutrophil infiltration to infarcted mammalian hearts (Kyne et al., 2000; Mocatta et al., 2007; Arruda-Olson et al., 2009; Dogan et al., 2009; Chia et al., 2009; Guasti et al., 2011; Akpek et al., 2012; Zhang et al., 2018; Yan et al., 2013; Rusinkevich et al., 2019) and quick mobilization toward cryoinjured zebrafish hearts (Lavine et al., 2014; Aurora et al., 2014; Lai et al., 2017; Xu et al., 2018, 2019; Bevan et al., 2020). To fully characterize neutrophils in regenerating heart models, we examined zebrafish neutrophils over time with the lyz:EGFP neutrophil reporter after ventricular resection surgery (Hall et al., 2007; Ellett et al., 2011). We observed rapid neutrophil mobilization starting from 6 h post-amputation (hpa). Neutrophils continued to amass at the injury site and peak at 1-day post-amputation (dpa), and then they quickly resolved, significantly dropping in numbers from 2 to 3 dpa (Fig. 1A and B).

Fig. 1. Neutrophil depletion inhibits cardiomyocyte proliferation.

(A) Visualization of lyz:EGFP⁺ cells in sections from lyz:EGFP transgenic hearts with no injury (top) and at 1 dpa (bottom). Brackets represent the regenerating area, dotted white line the injury site. Scale bar, 100 μm. (B) Quantification of lyz:EGFP ⁺ cells in the regenerating area from (A). No injury (n = 6), 6 hpa (n = 9), 1 dpa (n = 7), 2 dpa (n = 10), and 3 dpa (n = 8). The experiments were repeated at least once. Kruskal-Wallis ANOVA test. (C) Visualization of lyz:EGFP⁺ cells in sections from lyz:EGFP transgenic hearts with transient (4–6 hpa) DMSO (top) or AT7519-treatment (bottom) at 1 dpa. Brackets represent the injury area, dotted white line the injury site. Scale bar, 100 μm. (D) Quantification of lyz⁺ neutrophils in the regenerating area with transient DMSO or AT7519 treatment from (C). n = 10 animals for DMSO and AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (E) Visualization of mpeg1:EGFP ⁺ cells (macrophages) in sections from mpeg1:EGFP transgenic hearts with transient (4–6 hpa) DMSO (top) or AT7519-treatment (bottom) at 1 dpa. Brackets represent the injury area, dotted white line the injury site. Scale bar, 100 μm. (F) Quantification of mpeg1⁺ macrophages in the regenerating area with transient DMSO or AT7519 treatment from (E). n = 7 animals for DMSO and 6 animals for AT7519 groups. The experiments were conducted once. Mann-Whitney Rank Sum test. (G) Visualization of cardiomyocyte proliferation in sections from injured ventricles of EK WT animals at 7 dpa assessed with Mef2 and PCNA antibody staining in DMSO (left) or AT7519 (right) treated animals. Brackets represent the regenerating area, dotted white line the injury site. Dotted white squares enlarged in upper right-hand panels. Scale bar, 100 μm; upper right-hand panel scale bar, 50 μm. (H) Quantification of cardiomyocyte proliferation index at 7 dpa from (G) after vehicle DMSO or AT7519 treatment from 4 to 6 hpa. n = 13 animals for DMSO and 12 animals for AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test.

To examine the role of neutrophils during zebrafish heart regeneration, we employed AT7519, a CDK9 inhibitor that promotes neutrophil apoptosis, reduces neutrophil infiltration, and resolves inflammation (Rossi et al., 2006; Loynes et al., 2010; Leitch et al., 2012; Wang et al., 2012; Lucas et al., 2014; Hoodless et al., 2016; Kaveh et al., 2022). This antagonist has successfully been used to specifically deplete neutrophils in a zebrafish larvae model of cardiac injury (Kaveh et al., 2022). We treated injured lyz:EGFP fish from 4 to 6 hpa with control DMSO or AT7519 and collected their hearts at 1 dpa. The transient AT7519 treatment significantly reduced neutrophil numbers by 60 % at the injury site compared to vehicle controls (Fig. 1C and D). We also analyzed other immune cells like macrophages after AT7519 treatment and didn’t detect a significant difference in macrophage numbers in the injury site in comparison to control hearts (Fig. 1E and F; Supp Fig. 1), a similar result as previously observed in the zebrafish larvae model (Kaveh et al., 2022). To determine the mechanism of action for transient AT7519 treatment reducing neutrophil numbers in injured adult zebrafish hearts, we also examined whether AT7519 treatment induced neutrophil or macrophage apoptosis by conducting TUNEL assays. We found that transient AT7519 treatment had no significant effect on the percentage of apoptotic neutrophils or macrophages (Supp. 2A-D), suggesting transient AT7519 impacts neutrophil infiltration by reverse migration and not apoptosis in injured adult zebrafish hearts.

As transient 4–6 hpa AT7519-treatment hindered neutrophil recruitment to the injury site, we then analyzed the effect of neutrophil loss on cardiac regeneration. To do that, we performed Mef2/PCNA staining on 7 dpa hearts with vehicle- or AT7519-treatment. The results demonstrated a 63 % lowered CM proliferation index at 7 dpa with neutrophil loss compared to control hearts (Fig. 1E and F), suggesting neutrophils are required for heart regeneration. We also examined the effect of transient drug treatment at 3 dpa, when neutrophils are resolved, and CM proliferation induction occurs. We observed no significant effect on the CM proliferation index at 7 dpa in control versus AT7519-treated hearts (Supp Fig. 3E and F), indicating that AT7519 treatment impacts CM proliferation by depleting neutrophils and does not interfere with CM cell cycling by non-specific drug effects. Moreover, we also analyzed the effect of AT7519 treatment on the vasculature with our deltaC:EGFP reporter (Sun et al., 2022b) and TUNEL staining (Supp Fig. 2G and H). We observed no significant difference in the number of apoptotic coronary endothelium in control DMSO vs. AT7519-treated hearts (Supp Fig. 2G and H), further suggesting that transient AT7519 treatment does not generate an apoptotic environment in the recovering heart.

2.2. Neutrophils associate with activated, expanding epicardial cells

As neutrophils localize to the epicardial surface in the infarct region after MI/reperfusion injury in mice (Huang et al., 2012) and the epicardium is required for myocardial regeneration (Wang et al., 2015; Sun et al., 2022a), we speculated that neutrophils contribute to the epicardial regenerative response. To explore the mechanism of neutrophil loss on blocking heart regeneration, we then examined the spatial-temporal correlation of neutrophils with epicardial cells after zebrafish heart injury. By performing PCNA staining from 1 to 3 dpa in tcf21:nucEGFP fish (Wang et al., 2011), we found epicardial cell proliferation peaks at 1 dpa (Fig. 2A and B), which coincides with the dynamic neutrophil mobilization after heart injury. We further visualized significant association of epicardial cells with neutrophils in the cardiac wound, with AT7519 treatment reducing their association (Fig. 2C and D). Moreover, we found that proliferating epicardial cells were closely associated with neutrophils at the injury site, and AT7519 treatment hindered their association (Fig. 2E and F). We also performed immunostaining with raldh2, which is a critical enzyme for retinoic acid signaling and is induced in the epicardium (Lepilina et al., 2006; Kikuchi et al., 2011b). Our immunostaining analyses revealed a clear association of neutrophils with raldh2⁺ epicardial cells in the injury site (Fig. 2G), suggesting neutrophils reside near activated epicardial cells after zebrafish cardiac injury.

Fig. 2. Neutrophils associate with activated epicardial cells after heart injury.

(A) Visualization of epicardial cell proliferation in sections from 1, 2, and 3 dpa tcf21:nucEGFP animals assessed with PCNA antibody staining. Dotted white line represents the injury site. Scale bar, 50 μm. Yellow dotted box, enlarged in upper left panels. Arrowheads represent double positive tcf21⁺/PCNA⁺ epicardial cells. Scale bar, 25 μm. (B) Quantification of the epicardial cell proliferation index in the lateral wound edges from (A). n = 8 animals for 1 and 2 dpa, 7 animals for 3 dpa. Kruskal-Wallis ANOVA test. (C) The association of tcf21⁺ cells with mpx ⁺ neutrophils assessed by mpx antibody staining in injured ventricles of tcf21:nucEGFP animals at 1 dpa. Scale bar, 100 μm. Dashed white rectangle is enlarged to the right. Arrowheads represent tcf21⁺ cells associating with mpx ⁺ cells. Enlarged panel scale bar, 10 μm. n = 12 animals. (D) Quantification of the association of tcf21⁺ cells with mpx ⁺ neutrophils, assessed by mpx antibody staining in injured ventricles of tcf21:nucEGFP animals at 1 dpa with DMSO or AT7519 treatment from 4 to 6 hpa. n = 12 animals for DMSO and 14 animals for AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (E) The association of proliferating (PCNA⁺) tcf21⁺ cells with mpx ⁺ neutrophils assessed by mpx and PCNA antibody staining in injured ventricles of tcf21:nucEGFP animals at 1 dpa. Scale bar, 100 μm. Dashed white rectangle is enlarged to the right. Arrowhead represents tcf21⁺/PCNA⁺ cells associating with mpx⁺ cells. Enlarged panel scale bar, 10 μm. n = 12 animals. (F) Quantification of the association of proliferating (PCNA⁺) tcf21⁺ cells with mpx ⁺ neutrophils assessed by mpx antibody staining in injured ventricles of tcf21:nucEGFP animals at 1 dpa with DMSO or AT7519 treatment from 4 to 6 hpa. n = 11 animals for DMSO and 13 animals for AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (G) The correlation of lyz⁺ cells with raldh2⁺ epicardial cells (tcf21⁺) assessed by raldh2 antibody staining in injured ventricles of lyz:EGFP; tcf21:mCherry animals at 1 dpa. Scale bar, 50 μm. Dashed white rectangle area is enlarged to the right. Arrowheads represent lyz ⁺ cells associating with raldh2⁺ epicardial (tcf21⁺) cells in the regenerating area. Enlarged panel scale bar, 25 μm. n = 6 animals.

The neutrophil-epicardial cell interaction is critical for epicardial cell expansion, proliferation, and activation.

As we observed that neutrophil depletion limited CM proliferation and a close association of neutrophils with activated proliferating epicardial cells, we speculated that interference with the epicardial cell-neutrophil interaction during conditions of neutrophil depletion hinders the epicardial regeneration response. To test this possibility, we examined the effect of neutrophil depletion on the epicardium. We first analyzed the effect of neutrophil loss on epicardial cell accumulation in the injury site at 7 dpa in tcf21:nucEGFP hearts. We found that neutrophil loss limited the epicardial wound coverage at 7 dpa with a 41 % reduced EGFP⁺ cell number with AT7519 treatment compared to vehicle control (Fig. 3A and B). We then examined the effect of neutrophil depletion on epicardial cell proliferation. When analyzing GFP/PCNA stained 1 dpa tcf21:nucEGFP hearts, we observed a significantly hindered epicardial cell proliferation index by 64 % in the injury site of AT7519-treated animals in comparison to vehicle control (Fig. 3C and D). As neutrophils closely associate with raldh2⁺ activated epicardial cells at 1 dpa, we also analyzed the effect of neutrophil loss on raldh2⁺ activation in the injury site by raldh2 antibody staining in vehicle- or AT7519-treated 1 dpa tcf21:nucEGFP fish. We observed a significant reduction of 49 % for the raldh2 signals in the injury site of neutrophil-depleted hearts compared to vehicle control (Fig. 3E and F). Our observations of a significant reduction in epicardial cell expansion, proliferation, and activation under conditions of neutrophil loss indicate that neutrophils are critical for the epicardial regenerative response after heart damage.

Fig. 3. Neutrophil depletion hinders epicardial cell expansion, proliferation, and activation.

(A) Images of epicardial cell expansion at 7 dpa in tcf21:nucEGFP animals treated with vehicle DMSO (top) or AT7519 (bottom) from 4 to 6 hpa. Brackets represent the regenerating area, dotted white line the injury site. Scale bar, 100 μm. (B) Quantification of tcf21:nucEGFP⁺ cells from (A) in the regenerating area at 7 dpa. n = 18 animals for DMSO and 16 animals for AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (C) Visualization of epicardial cell proliferation in sections from 1 dpa tcf21:nucEGFP animals assessed with PCNA antibody staining in control DMSO (top) or AT7519 (bottom) treated hearts. Brackets represent the regenerating area, dotted white line the injury site. Scale bar, 100 μm. White dotted box, enlarged in upper right panels. Arrowheads represent double positive tcf21⁺/PCNA⁺ epicardial cells. Enlarged upper right panel scale bar, 25 μm. (D) Quantification of the epicardial cell proliferation index in the lateral wound edges from (C) with DMSO or AT7519 treatment. n = 13 animals for DMSO and 16 animals for AT7519 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (E) Visualization of raldh2⁺ signaling in/near epicardial cells in sections from 1 dpa tcf21:nucEGFP animals assessed with raldh2 antibody staining in control DMSO (top) or AT7519 (bottom) treated hearts. Scale bar, 50 μm. (F) Quantification of raldh2⁺ signals in the lateral wound edges after vehicle DMSO or AT7519 treatment from 4 to 6 hpa from (E). n = 7 animals in DMSO and AT7519 groups. The experiments were conducted once. Mann-Whitney Rank Sum test.

To confirm that our observations were the result of neutrophil loss altering the epicardial response and not an artefact of AT7519 treatment inducing epicardial cell death, we also performed TUNEL assays with tcf21:nucEGFP 1 dpa DMSO-control and AT7519-treated hearts. We detected no significant difference in the levels of apoptotic epicardial cells in control or AT7519-treated hearts (Supp Fig. 2E and F), suggesting that drug treatment did not induce epicardial cell death. Further, we also analyzed the effect of transient (2-h) AT7519 treatment on epicardial cell proliferation in non-injured adult tcf21:nucEGFP ex vivo hearts. No significant difference occurred in the epicardial cell proliferation index of uninjured adult hearts from control or AT7519 treated hearts (Supp Fig. 3A and B), suggesting transient AT7519 treatment doesn’t create an anti-proliferative environment. Moreover, we analyzed the effect of transient drug treatment at 3 dpa, a time point when neutrophils are largely resolved from the injury site, and we observed no significant effect on the epicardial cell expansion at 7 dpa in control versus AT7519-treated hearts (Supp Fig. 3C and D). These observations further suggest that the impact of AT7519-treatment on epicardial cell expansion, proliferation, and activation are the result of neutrophil depletion and not the generation of an apoptotic, anti-proliferative tissue environment.

2.3. Injury site neutrophils have different functional clusters

To further analyze the molecular nature of recruited neutrophils on heart regeneration, we conducted scRNA-seq of zebrafish neutrophils after heart injury. To do this, we amputated lyz:EGFP fish, extracted the hearts at 1 dpa, collected the wounded region, digested, isolated EGFP⁺ cells by FACS, and performed 10x Genomics’ scRNA-sequencing (Fig. 4A). After analyzing the scRNA-seq dataset, we identified 4 neutrophil clusters (Fig. 4B). We then confirmed that our clusters represented neutrophils, by generating violin plots for known zebrafish neutrophil markers (lyz, mpx, npsn, cpa5, mmp9, rac2, srgn, cybb, cepb1, lta4h, spi1b, and coro1a) from our 1 dpa neutrophil scRNA-seq dataset, as well as violin plots of other leukocyte markers (mpeg1.1 for macrophages, lck for T-cells, and itgae.1 for dendritic cells) (Supp Fig. 4) (Hall et al., 2007; Renshaw et al., 2006; Di et al., 2017; Rougeot et al., 2019; Tell et al., 2012; Huang et al., 2021; Jin et al., 2016; Poplimont et al., 2020; Keightley et al., 2017; Li et al., 2012; Ellett et al., 2011; Langenau et al., 2004; Lewis et al., 2014). These macrophage, T-cell, and dendritic cell markers are expressed in few to no cells in each neutrophil cluster. Markers for other leukocytes like pax5 (B-cells), gata2a (eosinophils), and nkl.2 (natural killer cells) had no expression in our dataset, as did markers for other cell types like erythrocytes (cahz, slc4a1a) and thrombocytes (itga2b) (Ma et al., 2021; Lai et al., 2017; Schebesta et al., 2007).

Fig. 4. Injury site neutrophils are composed of different functional clusters.

(A) Schematic of scRNA-seq from zebrafish 1 dpa lyz:EGFP hearts after ventricular resection surgery. (B) Uniform manifold approximation and projection (UMAP) clustering of lyz⁺ single-cells from adult 1 dpa hearts. (C) Gene ontology analysis of lyz:EGFP ⁺ cell clusters. (D) Heatmap of the top 10 markers for neutrophils from adult 1 dpa lyz:EGFP hearts. (E) Feature plots of neutrophil cluster 1 enriched genes: cxcr4b, nrros, raraa, and sgk1 expression in neutrophil clusters from 1 dpa hearts. (F) Feature plots of neutrophil cluster 2 enriched genes: gsdmea, f3b, sdc4, and lacc1 expression in neutrophil clusters from 1 dpa hearts. (G) Feature plots of neutrophil cluster 3 enriched genes: nr4a1, nocta, egr1, and trib3 expression in neutrophil clusters from 1 dpa hearts. (H) Feature plots of neutrophil cluster 4 enriched genes: ikzf1, afap1l1b, lpar5a, and sall1a expression in neutrophil clusters from 1 dpa hearts.

Based on gene ontology analysis, cluster 1 represents an apoptosis subpopulation with roles in leukocyte differentiation and migration (Fig. 4C and D). Of interest, the chemokine receptor cxcr4b, which has roles in immune cell migration, inflammation, and tissue regeneration is particularly enriched in clusters 1 and 4 (Sommer et al., 2020) (Fig. 4E). Other genes enriched in cluster 1 with overlap in clusters 2–4 include: the negative reactive oxygen species regulator nrros which activates TGF-β1 in macrophages (Noubade et al., 2014; Qin et al., 2018), the retinoic acid receptor raraa (Hale et al., 2006), and the anti-apoptotic gene serum/glucocorticoid regulated kinase 1 sgk1 which regulates neutrophil survival (Burgon et al., 2014) (Fig. 4E). Cluster 2 depicts a dynamic subpopulation enriched with response to external biotic stimulus, wounding, and cytokine genes in addition to genes for necroptosis and MAPK, NF-kappaB, and C-type lectin receptor signaling genes (Fig. 4C). Enriched cluster 2 genes have overlap with cluster 4 and include: the pryoptosis activator gasdermin E (gsdmea) (Chen et al., 2021), coagulation factor IIIb (f3b) with roles in vascularization during development (Zhou et al., 2011b), the cell signaling activator syndecan 4 (sdc4) which acts as an FGF co-receptor and independent ligand receptor for FGFs, VEGFs, and PDGFs (Elfenbein and Simons, 2013), and laccase (lacc1), which regulates macrophage metabolism (Fig. 4F) (Cader et al., 2020). Cluster 3 corresponds to a population with elevated tissue regeneration, glycolysis, insulin signaling, and ERK1-ERK2 MAPK cascade genes (Fig. 4C). Further analysis of cluster 3 demonstrated elevated expression of the nuclear receptor nr4a1 (Fig. 4G), which functions in cell metabolism, proliferation, apoptosis, and inflammation (Wu and Chen, 2018; Xu et al., 2022). Other cluster 3 enriched genes include the circadian regulator nocturin a (nocta) (Yang et al., 2017); the extracellular matrix regulator early growth response 1 (egr1) which supports tissue development, homeostasis, and repair (Havis and Duprez, 2020); and tribbles pseudokinase 3 (trib3) which behaves as either a pro-proliferation or pro-apoptosis stress gene through MAPK/ERK signaling depending on its environment (Ord and Ord, 2017; Cao et al., 2021). Analysis of cluster 4 revealed an endocytosis/phagosome enriched subpopulation with genes for TNFR1, IL-4, and NF-kappaB signaling (Fig. 4C). In cluster 4 we observed enhanced expression for the following genes: the zinc finger transcription factor ikzf1, which regulates leukocyte differentiation (Hess et al., 2022); actin filament associated protein (afap1l1b) for podosome formation (Snyder et al., 2011); lysophosphatidic acid receptor 5a (lpar5a) with roles in wound healing by regulating the immune response, extracellular matrix remodeling, and in the proliferation/recruitment of epithelial cells (Dacheux et al., 2023); and spalt-like transcription factor 1a (sall1a) which induces angiogenesis and positively regulates the FGF signaling pathway (Harvey and Logan, 2006; Yamamoto et al., 2010) (Fig. 4H).

2.4. Neutrophils control the epicardial response through FGF signaling

Based on analysis of our neutrophil scRNA-seq dataset, we identified neutrophil clusters enriched with FGF signaling pathway components (Fig. 4C–F) (Farooq et al., 2021; Khosravi et al., 2021). Previous reports demonstrated the role of FGF signaling in zebrafish fin, skeletal muscle, and heart regeneration (Kirchgeorg et al., 2018; Saera-Vila et al., 2016; Poss et al., 2000; Tahara et al., 2021). Moreover, FGF signaling within the zebrafish myocardium is critical for epicardial morphological changes and neovascularization after cardiac injury, while FGF10 overexpression in the myocardium facilitated epicardial accumulation in neonatal hearts (Lepilina et al., 2006; Rubin et al., 2013). However, whether FGF signaling affects early cardiac mechanisms like epicardial expansion in zebrafish has not been examined. To test this possibility, we examined epicardial cells after FGF signaling inhibition with SU5402 treatment (Kirchgeorg et al., 2018; Saera-Vila et al., 2016; Poss et al., 2000). We treated tcf21:nucEGFP fish with DMSO or SU5402 from 27 to 32 hpa and analyzed epicardial proliferation at 32 hpa (1 dpa). We detected a significant reduction in the epicardial proliferation index at 1 dpa by 40 % with SU5402 treatment compared to vehicle control (Fig. 5A and B), indicating the requirement of the FGF signaling pathway for the early epicardial response.

Fig. 5. Neutrophils promote epicardial expansion through FGF and MAPK/ERK signaling.

(A) Visualization of epicardial cell proliferation in sections from 1 dpa tcf21:nucEGFP animals assessed with PCNA antibody staining in control DMSO (top) or FGF signaling inhibitor SU5402 (bottom) treated hearts. Brackets represent the regenerating area, dotted white line the injury site. Scale bar, 100 μm. White dotted box, enlarged in upper right panels. Arrowheads represent double positive tcf21⁺/PCNA⁺ epicardial cells. Enlarged upper right panel scale bar, 25 μm. (B) Quantification of the epicardial cell proliferation index in the lateral wound edges from (A) with DMSO or SU5402 treatment. n = 11 animals for DMSO and 13 animals for SU5402 groups. The experiments were repeated at least once. Mann-Whitney Rank Sum test. (C) Visualization of MAPK/ERK signaling activation with ERK-P antibody staining in injured 1 dpa tcf21:nucEGFP hearts. Scale bar, 100 μm. White dotted box, enlarged in panels to the right. Arrowheads represent double positive tcf21⁺/ERK-P⁺ epicardial cells. Enlarged panel scale bar, 10 μm. n = 7 animals. (D) The activation of MAPK/ERK signaling with ERK-P signals assessed by ERK-P antibody staining in injured ventricles of tcf21:nucEGFP animals at 1 dpa with DMSO (top) or AT7519 (bottom) treatment at 4–6 hpa. Dotted white line represents the injury site. Scale bar, 100 μm. (E) Quantification of ERK-P⁺ signals in the lateral wound edges after DMSO or AT7519 treatment from (D). n = 13 DMSO and AT7519 animals in each group. The experiments were repeated at least once. Mann-Whitney Rank Sum test.

MAPK/ERK signaling is downstream of the FGF pathway, regulates organ regeneration (Wen et al., 2022), and is induced in the regenerating zebrafish epicardium (Han et al., 2014; Liu and Zhong, 2017). As epicardial progenitor cells proliferate from activation of the MAPK/ERK pathway (Deng et al., 2017), we speculated that the neutrophil-epicardium interaction may involve MAPK/ERK signaling. We examined 1 dpa tcf21:nucEGFP hearts by performing ERK-P antibody immunostaining, which indicates activation of the MAPK/ERK signaling pathway. We observed colocalization of ERK-P signals with tcf21⁺ epicardial cells (Fig. 5C), suggesting activation of the MAPK/ERK signaling pathway in early epicardial cells. Next, we examined MAPK/ERK signaling activation through ERK-P expression in the injury site after depleting neutrophils from 4 to 6 hpa. We performed ERK-P immunostaining on 1 dpa tcf21:nucEGFP fish hearts with DMSO or AT7519 treatment. The results revealed an 85 % lower number of ERK-P⁺ signals in the injury site of AT7519 treated hearts compared to vehicle control (Fig. 5D and E), indicating a requirement of neutrophils for efficient MAPK/ERK activation in early epicardial cells.

2.5. Communication between neutrophils and epicardial cells during zebrafish heart regeneration

To identify potential ligand-receptor interactions involved in the communication between neutrophils and epicardial cells, we analyzed the outgoing patterns of secreting cells and incoming patterns of target cells at 2 days post-injury (dpi) from a published scRNA-seq dataset of non-myocytes from injured zebrafish hearts (Ma et al., 2021) with CellChat (Fig. 6A and B). We selected this dataset, as it includes both neutrophils and fibroblasts (epicardial cells) from 2 dpi hearts, which will more closely represent 1 dpa epicardial cells. In these analyses, epicardial cells and fibroblasts are combined into one group as the epicardium contributes to cardiac fibroblasts, these cells have similar gene expression patterns, and separating these cell types is still considered controversial (Quijada et al., 2020; TALLQUIST, 2019; Hesse et al., 2021). Currently, there is no available dataset of epicardial cells from 1 dpa regenerating hearts for direct comparison with our 1 dpa neutrophil dataset. In the analyzed 2 dpi dataset, we detected 1 primary signaling pathway from zebrafish neutrophils to fibroblasts/epicardial cells (Fib/Epi) (Fig. 6A and B), the epidermal growth factor (EGF) signaling pathway. Further analysis of this pathway revealed that both neutrophils and lymphatic endothelial cells secrete the EGF factor heparin-binding EGF-like growth factor a (hbegfa), while thrombocytes secrete hbegfb, which binds to the Fib/Epi cell receptor epidermal growth factor receptor a (egfra) for EGF signaling at 2 dpi (Fig. 6C–E). Analysis of our 1 dpa neutrophil scRNA-seq dataset demonstrated expression of the EGF ligand hbegfa in all neutrophil clusters (Fig. 6F), further implicating EGF signaling as a possible supportive factor for neutrophils to aid epicardial regeneration. Release of the EGF factor hbegfa from neutrophils is particularly intriguing, as hbegfa has previously demonstrated roles in zebrafish tissue regeneration, like retinal and spinal cord regeneration, by synergizing with growth factors from the FGF, IGF, and insulin signaling pathways as well as cytokines to induce cell proliferation pathways like the MAPK/ERK cascade (Wan et al., 2014; Dao et al., 2018; Cigliola et al., 2023). Of note, we observed enrichment of insulin, MAPK/ERK, and FGF signaling factors in our neutrophil scRNA-seq dataset in addition to hbegfb expression in all neutrophil clusters (Fig. 6F), indicating facilitation of the epicardial regenerative response by neutrophils likely occurs through the cooperation of several growth factors and cytokines to induce epicardial cell proliferation. Moreover, HB-EGF released from myeloid cells induces the proliferation of epithelial cells to promote pancreatic recovery in mice, further implicating immune cells in the release of growth secretory factors for epithelial cell proliferation in vertebrates (Wen et al., 2019).

Fig. 6. Communication between neutrophils and epicardial cells during zebrafish heart regeneration.

(A) Alluvial plot showing the inferred outgoing communication patterns of non-myocyte cell populations from injured zebrafish hearts at 2 days post-injury, which shows the correspondence between the inferred latent patterns and cell groups, as well as signaling pathways. The thickness of the flow indicates the contribution of the cell group or signaling pathway to each latent pattern. EC, endothelial cell; Fib, fibroblast; LEC, lymphatic endothelial cell; Mes, mesenchymal cells; Neutro, neutrophils; T/NK, T-cells and natural killer cells; Thromb, thrombocytes. MK, midkine; MIF, macrophage migration inhibitory factor; CXCL, chemokine (C-X-C motif) ligand, ANGPTL, angiopoietin-like proteins; EGF, epidermal growth factor; VEGF, vascular endothelial growth factor; PDGF, platelet-derived growth factor. (B) Alluvial plot showing the inferred incoming communication patterns of non-myocyte cell populations from injured zebrafish hearts at 2 days post-injury, which shows the correspondence between the inferred latent patterns and cell groups, as well as signaling pathways. The thickness of the flow indicates the contribution of the cell group or signaling pathway to each latent pattern. EC, endothelial cell; Fib, fibroblast; LEC, lymphatic endothelial cell; Mes, mesenchymal cells; Neutro, neutrophils; T/NK, T-cells and natural killer cells; Thromb, thrombocytes. MK, midkine; MIF, macrophage migration inhibitory factor; CXCL, chemokine (C-X-C motif) ligand, ANGPTL, angiopoietin-like proteins; EGF, epidermal growth factor; VEGF, vascular endothelial growth factor; PDGF, platelet-derived growth factor. (C) Circle plot showing the inferred EGF signaling pathway network in the regenerating zebrafish heart. The width of the edges represents the interaction strength. Arrow indicates the target cell. EC, endothelial cell; Fib, fibroblast; LEC, lymphatic endothelial cell; Mes, mesenchymal cells; Neutro, neutrophils; T/NK, T-cells and natural killer cells; Thromb, thrombocytes. (D) Heatmap for the inferred EGF signaling pathway network, displaying its relative importance of each cell group ranked according to the computed four network centrality measures in the regenerating zebrafish heart. Edge width represents the communication probability, and the edge colors are consistent with the color of the sender cell type. EC, endothelial cell; Fib, fibroblast; LEC, lymphatic endothelial cell; Mes, mesenchymal cells; Neutro, neutrophils; T/NK, T-cells and natural killer cells; Thromb, thrombocytes. (E) Violin plot demonstrating the expression levels of the ligands (hbegfa, hbegfb) and receptor (egfra) involved in the EGF signaling pathway network in the regenerating zebrafish heart for each non-myocyte cell type. (F) Violin plot demonstrating the expression levels of the EGF signaling pathway ligand hbegfb in 1 dpa zebrafish neutrophils.

3. Discussion

In this study, we examined zebrafish neutrophils after heart injury and identified their association with epicardial cells. Our work reveals that neutrophils are critical for the epicardial regenerative response during heart regeneration.

As neutrophil depletion in our study relied on pharmaceutical treatment with the CDK9 inhibitor AT7519, which previously demonstrated anti-proliferation and apoptosis induction characteristics (Squires et al., 2009; Santo et al., 2010), we performed additional assays to exclude the possibility of non-specific drug effects leading to misinterpretation of our results (Supp Fig. 2A-H; Supp Fig.3 A-F). Although our data indicate that transient AT7519 treatment does not create an apoptotic, anti-proliferative tissue environment, we aim to continue investigating the neutrophil-epicardium interaction during cardiac repair with genetic approaches such as the MTZ-NTR system for specific neutrophil depletion, which was previously resistant to neutrophil ablation (Hall et al., 2022), to identify critical neutrophil functions and further explore other neutrophil-cardiac cell interactions.

AT7519 incubation depleted cardiac injury-site neutrophils by reverse migration without altering macrophage retention, as observed in a recent model of larval heart regeneration (Kaveh et al., 2022). However, we observed differences in the regenerative response with neutrophil loss in the adult ventricular resection model compared to the laser-dissected larval zebrafish model. Adult zebrafish experienced a hindered epicardial regenerative response, contributing to a limited myocardial recovery, while larval zebrafish demonstrated elevated CM expansion and wound closure in response to neutrophil loss. Enhanced cardiac regeneration occurred in larval zebrafish from elevated tnf expression in cardiac wound macrophages, which have a pro-regenerative phenotype conducive to heart repair (Kaveh et al., 2022; Nguyen-Chi et al., 2017; Tsarouchas et al., 2018; Gurevich et al., 2018; Cavone et al., 2021), while tnf ⁺ macrophages have a pro-inflammatory phenotype and facilitate scar deposition in adult zebrafish (Bevan et al., 2020), implicating a functional transition in tnf⁺ macrophages during development (Kaveh et al., 2022). Therefore, the observed cardiac regeneration differences in AT7519 treatment between these studies likely result from alterations in the immune environment and genetic landscape of adult versus developing zebrafish hearts and the different injury methods utilized.

The activated epicardium swiftly proliferates, migrates, encloses the wounded region, and releases paracrine signals to support heart regeneration (Lepilina et al., 2006; Wang et al., 2015; Kikuchi et al., 2011a; Zhou et al., 2011a; Cai et al., 2019; Cao et al., 2017). These epicardial events are regulated by several factors like hypoxia, oxytocin, C/EBP, Tβ4, macrophages, and hedgehog signaling during development and regeneration in zebrafish and murine neonates (Wang et al., 2015; Wasserman et al., 2022; Huang et al., 2012; Pinto et al., 2014; Vieira et al., 2017; Bruton et al., 2022). Recently, we identified critical epicardial subtypes like hapln1a⁺ cells, which deposit hyaluronic acid to facilitate CM proliferation and pre-lead coronary extensions (Sun et al., 2022a, 2023). These studies highlight the importance of the epicardium during heart regeneration, however, the regulation of early epicardial regenerative responses remains far from clear. In this study, we observed the association of neutrophils with activated, expanding epicardial cells, and demonstrated that interfering with the neutrophil response blocked the epicardial regenerative response. The capacity for leukocytes to facilitate epicardial cell activation during heart regeneration has long been speculated and was only recently indicated by a study revealing that macrophages activate the epicardium to induce CM proliferation in larval zebrafish (Bruton et al., 2022). Our work provides further evidence for the regulatory role of leukocytes on early epicardial responses. Further, the close association of neutrophils with epicardial cells and the suggestion of the epicardium regulating the immune response by controlling infiltration of neutrophils, macrophages, and T-reg cells (Huang et al., 2012; Balmer et al., 2014; Pinto et al., 2014; Ramjee et al., 2017; Sanz-Morejón et al., 2019), indicates that cross talk between epicardial cells and leukocytes are critical for heart regeneration initiation and progression.

Our study indicated neutrophils facilitate the epicardial regenerative response potentially through FGF signaling acting synergistically with the EGF, IGF, and insulin pathways to activate downstream effectors like the MAPK/ERK signaling pathway, as observed with retinal regeneration (Wan et al., 2014). Previously, myocardial FGF signaling supported the regenerating epicardium in injured murine neonatal hearts and facilitated neovascularization after zebrafish cardiac injury (Lepilina et al., 2006; Rubin et al., 2013). Our observations of neutrophil clusters enriched with FGF signaling pathway factors and that depleting neutrophils blocked epicardial ERK-P activation and epicardial expansion indicate neutrophils are also an important source of FGF activity after heart injury, especially in the early epicardial regenerative response. However, as AT7519 can activate Gsk3-beta, which negatively regulates ERK (Santo et al., 2010), it is possible that the observed dampening of ERK-P expression is an artefact of AT7519-treatment and does not directly result from neutrophil depletion. As FGF signaling is well-known to play critical roles in mechanisms of tissue repair and regeneration (Kirchgeorg et al., 2018; Saera-Vila et al., 2016; Poss et al., 2000; Tahara et al., 2021; Farooq et al., 2021; Khosravi et al., 2021), our work further establishes FGF signaling in epicardial stimulation quickly after heart injury and suggests additional cell types, like neutrophils, contribute to epicardial activation through this pathway. The fact that rat inflammatory cells upregulate FGF-1 in response to myocardial infarction (Zhao et al., 2011) and neutrophils associate with the epicardium after heart injury in mice models of MI (Huang et al., 2012), combined with our observations on the close interaction of neutrophils with epicardial cells in the adult zebrafish injured heart, suggest the neutrophil-epicardium interaction and FGF signaling for epicardial regeneration are a conserved interaction of leukocytes activating early epicardial cells among vertebrates.

4. Materials and methods

4.1. Zebrafish and heart injuries

Outbred EK or EK/AB mixed background zebrafish (4–10 months of age) were used for ventricular resection surgeries as previously described (Poss et al., 2002). Animal density was maintained at ~4 fish per liter in all experiments. AT7519 (MedChemExpress, catalog no. HY-50940) was dissolved in DMSO as a 65.4 mM stock solution. To inhibit neutrophil recruitment, animals were treated for 2 h from 4 to 6 h post-amputation (hpa) with 50 μM AT7519 or 0.08 % DMSO in 600 mL fish water in the dark. SU5402 (Selleck Chemicals, catalog no. S7667) was dissolved in DMSO as a 17 mM stock. To inhibit FGF signaling, animals were treated for 5 h from 27 hpa to 32 hpa with 17 μM SU5402 or 0.1 % DMSO in 600 mL fish water in the dark, as described (Kirchgeorg et al., 2018; Saera-Vila et al., 2016; Poss et al., 2000). Transgenic strains described elsewhere include Tg(lyz:EGFP)^nz117Tg (Hall et al., 2007; Ellett et al., 2011), Tg(mpeg1:EGFP)^gl22Tg (Ellett et al., 2011), Tg (tcf21:nucEGFP)^pd41Tg (Wang et al., 2011), Tg(tcf21:mCherry-NTR)^pd108Tg (Wang et al., 2015), and Tg(deltaC:EGFP)^em11Tg (Sun et al., 2022b). All transgenic strains were analyzed as hemizygotes. All animal procedures were performed in accordance with Emory University guidelines.

4.2. Histological methods

Histological analyses were performed on 10 μm cryosections of paraformaldehyde-fixed hearts. For immunostaining of heart sections: slides were placed in caplin jars with citrate buffer (10 mM Citric Acid, 0.05 % Tween 20, pH 6.0) and boiled at 98 °C for 30 min in a 1 L beaker with water, allowed to cool for 30 min, and then were washed 4x for 5 min with PBS plus 0.1 % Tween-20. Samples were then blocked with 2 % horse serum, 1 % DMSO, 10 % heat inactivated new calf serum and 0.1 % Tween-20 in PBS for 1 h at 37 °C. Primary antibodies were diluted in PBS plus 1 % DMSO, 10 % heat inactivated new calf serum, and 0.1 % Tween-20 and incubated with hearts overnight at 4 °C. Samples were then washed with PBS plus 0.1 % Tween-20 and incubated with the secondary antibody diluted in the PBS plus 1 % DMSO, 10 % heat inactivated new calf serum, and 0.1 % Tween-20 for 1 h at 37 °C. Samples were mounted in Vectashield vibrance antifade mounting medium with DAPI (Vector Laboratories, H-1800-10). For IB4 immunostaining, slides were washed 4x for 5 min with PBS plus 0.1 % Tween-20. Samples were then blocked with 2 % horse serum, 1 % DMSO, 10 % heat inactivated new calf serum and 0.1 % Tween-20 in PBS for 1 h at 37 °C. The IB4 antibody was diluted in PBS plus 1 % DMSO, 10 % heat inactivated new calf serum, and 0.1 % Tween-20 and incubated with hearts at RT for 3 h or 4 °C overnight. Samples were then washed with PBS plus 0.1 % Tween-20 and mounted in Vectashield vibrance antifade mounting medium with DAPI. Primary antibodies used in this study include: IB4 (1:100; Vector labs, DL-1208-5), anti-aldh1a2 (raldh2, 1:100; rabbit; GeneTex: GTX124302), anti-mpx (1:100; rabbit; GeneTex: GTX128379), anti-Mef2 (1:100; rabbit; Santa Cruz Biotechnology: sc-313), anti-PCNA (1:200; mouse; Sigma-Aldrich: P8825), anti-GFP (1:400; chicken; Aves Labs: GFP-1020), and anti-ERK-P (1:100; rabbit; Cell Signaling Technology: 4370). Alexa Fluor secondary antibodies used include 488 (1:200; goat anti-Chicken; Invitrogen A-11039), 594 (1:200; goat anti-mouse; Invitrogen: A-11005 or 1:200; goat anti-rabbit; Invitrogen: A-11012), and 633 (1:200; goat anti-mouse; Invitrogen: A-21050 or 1:200; goat anti-rabbit; Invitrogen: A-21070). For detection of apoptotic cells, the Click-iT^™ Plus TUNEL Assay (ThermoFisher: C10619) kit was used according to the manufacturer’s instructions.

4.3. Quantification of epicardial cell proliferation index

To calculate epicardial cell proliferation indices, 10 μM ventricular sections from 1 to 3 dpa tcf21:nucEGFP fish were stained with antibodies against PCNA and GFP. Ventricle images after staining were captured using a 20x objective lens (1024 × 1024 pixels). To calculate proliferating epicardial cell numbers in the injured area, we imaged the 3 ventricular sections with the largest injuries, and quantified tcf21⁺ and tcf21⁺/PCNA⁺ cells within a defined region (250 pixels × 100 pixels) in the lateral wound edges of both sides of the injury site. The percentages of tcf21⁺/PCNA⁺ cells from the 3 selected sections were averaged to determine the proliferation index for each animal.

4.4. Quantification of cardiomyocyte proliferation index

To calculate the cardiomyocyte proliferation indices, 10 μM ventricular sections from 7 dpa hearts were stained with antibodies against Mef2 and PCNA. Ventricle images after staining were captured using a 20x objective lens (1024 × 1024 pixels). For resection injuries, we imaged the 3 ventricular sections with the largest injuries, and quantified Mef2⁺ and Mef2⁺/PCNA⁺ cells within a defined region, typically including almost all Mef2⁺/PCNA⁺cells near the injury site (1024 ×216 pixels). The percentages of Mef2⁺/PCNA⁺ cells from the 3 selected sections were averaged to determine a proliferation index for each animal.

4.5. Quantification of neutrophils, macrophages, epicardial cells, raldh2 and ERK-P signals

To quantify neutrophils in the injured heart area, ventricle images from lyz:EGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Neutrophil totals were determined manually within and near the injury site (1024 × 350 pixels) and averaged from 3 sections of each ventricle. To quantify total macrophages in the injured heart area, ventricle images from mpeg1:EGFP hearts or after IB4 staining were captured using a 20x objective lens (1024 × 1024 pixels). Macrophage totals were determined manually within and near the injury site (1024 × 350 pixels) and averaged from 3 sections of each ventricle. To quantify epicardial cells in the injured heart area at 7 dpa, ventricle images from tcf21:nucEGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Epicardial cells were determined manually within the wound area (216 pixels in vertical) and the ratio of EGFP⁺ epicardial cells versus the length of the apical wound was calculated and averaged from 3 sections of each ventricle (Wang et al., 2015). To quantify raldh2 signals in the lateral wound edges, images of ventricles after raldh2 antibody staining were captured using a 20x objective lens (1024 × 1024 pixels). The raldh2 signal area was measured by ImageJ software for signals in the lateral wound edges (250 × 100 pixels) and averaged from 3 sections of each ventricle. To quantify ERK-P signals in the lateral wound edges, images of ventricles after ERK-P antibody staining were captured using a 20x objective lens (1024 × 1024 pixels). The ERK-P signal area was measured by ImageJ software for signals in the lateral wound edges (250 × 100 pixels) and averaged from 3 sections of each ventricle.

4.6. Quantification of apoptotic neutrophils, macrophages, epicardial cells, and vascular endothelium

To quantify apoptotic neutrophils in the regenerating area, ventricle images from lyz:EGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Apoptotic neutrophil totals (TUNEL⁺/lyz⁺) and neutrophils (lyz⁺) were determined manually within and near the injury site (1024 × 350 pixels). The percentages of TUNEL⁺/lyz⁺ cells per total lyz⁺ cells from the 3 selected sections were averaged to determine an apoptotic cell index for each animal. To quantify apoptotic macrophages in the regenerating area, ventricle images from mpeg1:EGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Apoptotic macrophage totals (TUNEL⁺/mpeg1⁺) and macrophages (mpeg1⁺) were determined manually within and near the injury site (1024 × 350 pixels). The percentages of TUNEL⁺/mpeg1⁺ cells per total mpeg1⁺ cells from the 3 selected sections were averaged to determine an apoptotic cell index for each animal. To quantify apoptotic epicardial cells in the regenerating area, ventricle images from tcf21:nucEGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Apoptotic epicardial cell totals (TUNEL⁺/tcf21⁺) and epicardial cells (tcf21⁺) were determined manually within a defined region (250 pixels × 100 pixels) in the lateral wound edges of both sides of the injury site. The percentages of TUNEL⁺/tcf21⁺ cells per total tcf21⁺ cells from the 3 selected sections were averaged to determine an apoptotic cell index for each animal. To quantify apoptotic vascular endothelial cells in the regenerating area, ventricle images from deltaC:EGFP hearts were captured using a 20x objective lens (1024 × 1024 pixels). Apoptotic vascular endothelial cell totals (TUNEL⁺/deltaC⁺) and vascular endothelial cells (deltaC⁺) were determined manually within a defined region (250 pixels × 100 pixels) in the lateral wound edges of both sides of the injury site. The percentages of TUNEL⁺/deltaC⁺ cells per total deltaC⁺ cells from the 3 selected sections were averaged to determine an apoptotic cell index for each animal.

4.7. Quantification of neutrophils associating with epicardial cells and proliferating epicardial cells

To calculate the association of neutrophils with epicardial cells, 10 μM ventricular sections from tcf21:nucEGFP fish were stained with mpx antibody. Ventricle images after staining were captured using a 20x objective lens (1024 × 1024 pixels). To calculate neutrophils associating with epicardial cells, we imaged the 3 ventricular sections with the largest injuries and quantified total tcf21⁺ cells and tcf21⁺ cells adjacent to mpx signals within a defined region (250 pixels × 100 pixels) in the lateral wound edges of both sides of the injury site. The percentages of tcf21⁺cells adjacent to mpx⁺neutrophils per total tcf21⁺cells from the 3 selected sections were averaged to determine the association of neutrophils with epicardial cells in each animal. To calculate the association of neutrophils with proliferating epicardial cells, 10 μM ventricular sections from tcf21:nucEGFP fish were stained with mpx and PCNA antibodies. Ventricle images after staining were captured using a 20x objective lens (1024 × 1024 pixels). To calculate neutrophils associating with proliferating epicardial cells, we imaged the 3 ventricular sections with the largest injuries and quantified total tcf21⁺/PCNA⁺ cells and tcf21⁺/PCNA⁺ cells adjacent to mpx signals within a defined region (250 pixels × 100 pixels) in the lateral wound edges of both sides of the injury site. The percentages of tcf21⁺/PCNA⁺ cells adjacent to mpx⁺ neutrophils per total tcf21⁺/PCNA⁺ cells from the 3 selected sections were averaged to determine the association of neutrophils with proliferating epicardial cells for each animal.

4.8. Heart culture

The ex vivo heart culture was performed as previously described (Sun et al., 2022a, 2023; Wang et al., 2015). In summary, hearts were rinsed several times in PBS after collection and then placed in DMEM medium plus 10 % fetal bovine serum, 1 % non-essential amino acids, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL Primocin^™ (InvivoGen) and 50 μM 2-Mercaptoethanol (DMEM⁺⁺). Hearts were cultured at 28 °C in an incubator with 5 % CO₂. AT7519 (MedChemExpress, catalog no. HY-50940) was dissolved in DMSO to a stock concentration of 65.4 mM. AT7519 was used at a final concentration of 50 μM for ex vivo culture and 0.08 % DMSO for control-treated hearts. Control or AT7519-treatment occurred for a transient 2-h period, then the hearts were washed and placed in fresh DMEM media and cultured for 4 h. After 4 h, hearts were fixed in 4 % paraformaldehyde overnight at 4 °C.

4.9. Single-cell RNA-sequencing

To prepare lyz⁺ cells for single-cell RNA-sequencing analyses we raised 120 lyz:EGFP fish to the adult stage. The hearts of adult fish at 6 months old (3.5–4 cm length) were extracted at 1-day post-amputation (dpa) and the wounded apex region was collected. The heart samples were digested with 0.26 U/mL Liberase^™ Thermolysin Medium (TM) based on a previously published protocol (Spanjaard et al., 2018). Dissociated cells were spun down and live EGFP⁺ cells were sorted by flow cytometry. To ensure at least a 95 % cell viability following the entire procedure, we gated viable cells by negative SYTOX^™ Red fluorescence and examined cell viability after cell sorting. Isolated cells were sent to the Emory Integrated Genomics Core (EIGC) center for 10x single-cell RNA-sequencing. Isolated EGFP⁺ cells were prepared in a single-cell suspension and counted using a Countess (ThermoFisher) system. The loaded Single Cell 3’ Chip was placed on a 10x Genomics Chromium Controller Instrument (10x Genomics, Pleasanton, CA, USA) to generate single-cell gel beads in emulsion (GEMs). Single-cell RNA-seq libraries were prepared using the Chromium Single Cell 3’ Library & Cell Bead Kit v3.1 (Cat. No. 1000128, 1000127, 120262; 10x Genomics) according to the manufacturer’s protocol. Libraries were sequenced with an Illumina NextSeq550 using mid-output 150-cycle kits according to manufacturer specifications. The newly generated scRNA-seq data were demultiplexed, aligned, and quantified using Cell Ranger Single-Cell Software. Preliminary filtered data generated from Cell Ranger were used for downstream analysis by the Seurat R package according to standard workflow.

4.10. Cell-cell interaction analysis

A previously published scRNAseq dataset of the regenerating zebrafish heart was obtained from the authors (GSE145979_2_dpi) (Ma et al., 2021). Downstream analysis was performed by the Seurat R package (4.3.0.1) according to standard workflow (Hao et al., 2021). The different cell types were identified based on the published paper (Ma et al., 2021). Cell-cell interaction analysis was performed with R package CellChat (v1.6.1) (Jin et al., 2021). Briefly, we created CellChat object using “createCellChat” function, followed by the recommended preprocessing functions with default parameters for the analysis of individual datasets. CellChatDB.zebrafish were used as the database for inferring cell-cell communication. We computed the communication probability by the “computeCommunProb” function. Function “netAnalysis_computeCentrality” was used to calculate network centrality scores. Functions such as “netVisual_circle”, “netVisual_aggregate”, “netVisual_bubble”, “netAnalysis_signalingRole_network”, “netAnalysis_signalingRole_scatter”, and “netAnalysis_signalingRole_heatmap” were used to generate different plots.

4.11. Statistical analysis

All data are presented as mean ± standard error of the mean (SEM). All statistical analyses were performed using GraphPad Prism 7 software. The Kruskal-Wallis test was used for assessing statistical differences between 3 or more groups. The Mann-Whitney Rank Sum test was used for assessing statistical differences between the 2 groups. Results with P values < 0.05 were considered statistically significant.

Data availability

Data will be made available on request.