Introduction

Cardiovascular disease and myocardial infarction (MI) remain major global health burdens, and their relationship to the escalating obesity epidemic and an increasingly unhealthy lifestyle means their frequency is unlikely to decline in the coming years (Braunwald 2015). MI, often occurring as a consequence of coronary heart disease, can result in massive cardiomyocyte (CM) loss, perhaps up to 1 billion cells (Murry et al. 2006) and is often followed by fibrotic tissue and scar formation, severely limiting the functional capacity of the heart and leading to heart failure. The holy grail for cardiovascular therapeutics would be the ability to stimulate cardiac regeneration and replace the myocardium lost to cell death; however, this remains elusive and poorly understood, with regeneration in mammalian models restricted to early developmental stages (reviewed by Murry et al. 2006; Cahill et al. 2017). Current treatment options for cardiovascular disease range from pharmaceutical to transplant; however, there is no esca** that each strategy has associated drawbacks (reviewed by Hashimoto et al. 2018).

The ability to somehow stimulate the production of new CMs, for example via the use of stem cell or cell replacement therapies, has long been the ultimate goal (reviewed by Murry et al. 2006; Hashimoto et al. 2018). However, the limited success of CM-focused therapies has highlighted that these approaches are too narrow and overly focused on a single cell type. To develop an approach that could potentially stimulate endogenous regeneration or provide additional support to a CM replacement strategy, it will also be essential to harness a pro-regenerative microenvironment, widening the scope to encompass diverse non-myocytes and the extracellular milieu. To study this complexity, it seems logical to turn to a model system that offers both a natural regenerative ability and a complete in vivo system, criteria that are met (almost uniquely) by the zebrafish.

Surgical resection of the apex of the ventricle has been in use for almost twenty years to study the mechanisms involved in regenerating cardiac tissue in the adult zebrafish (Poss 2002). Additionally, genetic cell ablation models have been utilised to reduce CM number and trigger regeneration (Wang et al. 2011). More recently, the development of a cryoinjury model, whereby a liquid nitrogen-cooled probe is placed on the ventricle to induce localised cell death (Chablais et al. 2011; Schnabel et al. 2011; Gonzalez-Rosa et al. 2011), has provided a new aperture to view the resulting regeneration and is more representative of the cellular damage resulting from MI. Whichever injury model is chosen, studies into the complex injury response in this model are providing a wealth of information on the cellular and molecular mechanisms required to rebuild the heart after damage and are increasingly highlighting the need to widen the scope beyond CMs to the microenvironment surrounding these cells.

Current therapeutic outlook

Despite the relatively simple architecture of the adult mammalian heart, repair is complex, and regeneration has proven unattainable thus far. As regeneration is currently impossible, and heart transplant is limited by practical considerations (not least the lack of donors but also the associated surgical complexities) (Yacoub 2015), the therapeutic focus is treatment rather than cure, specifically in preventing the progression of ischaemic heart disease to heart failure (Sacks et al. 2014). Cardioprotective treatment options can improve blood supply (e.g. revascularisation by thrombolysis or bypass surgery) and pharmacological interventions can decelerate cardiac remodelling (e.g. ACE inhibitors and β-blockers), whereas more advanced stages of heart failure benefit from mechanical support therapies (such as left-ventricular assist devices or cardiac resynchronisation therapy) (Reviewed by Hashimoto et al. 2018). Despite this broad range of therapeutic interventions, ischaemic heart failure and its associated adverse cardiac remodelling remain major challenges for health services worldwide.

Cell-replacement therapies (using, for example, direct application of (often exogenous) stem cells or re-differentiated CMs derived from somatic cells via induced pluripotent stem cells (iPSCs)) aim to directly address the problem of insufficient CM proliferation, which is widely considered to be the limiting factor in heart regeneration. These therapies, which involve growing vast numbers of cells in vitro (potentially more than the 1 billion CMs that may be lost following MI) and injecting them directly into the injured heart, have shown promise in numerous pre-clinical settings. However, results for this relatively new strategy are often inconsistent and fail to achieve marked improvements in cardiac function (reviewed by Müller et al. 2018; Tehzeeb et al. 2019), and studies in animal models have revealed concerning and prevalent side effects including arrhythmias and tachycardia (Shiba et al. 2016; Liu et al. 2018; and reviewed by Chen et al. 2020). These inconsistencies may be partly explained by a recent and important publication from Vagnozzi et al. (2020) which has found that direct chemical stimulation of the innate immune system produces an outcome comparable to treatment with cell therapies previously reported to be reparative in models of cardiac ischaemic injury (Vagnozzi et al. 2020). This comprehensive report has demonstrated that beneficial effects thought to result from stem cell therapies are chiefly mediated by an acute regional accumulation of specific macrophage populations and an inflammatory wound healing response triggered by the cell transplant, rather than direct proliferation of the transplanted cells. These findings are also cohesive with an emerging research trend to further delineate the role of the immune response in cardiac repair and regeneration, which will be the focus of this review.

Cardiomyocyte proliferation

In recent years, the scant proliferative capacity of CMs has been highlighted as the major limiting factor in the process of heart regeneration following cardiac injury/MI. In the mature mammalian heart, CM proliferation is poor, and progenitors are present in low numbers (reviewed by Cahill et al. 2017; Hashimoto et al. 2018), with the post-natal heart mainly increasing in size via cardiac hypertrophy (Alkass et al. 2015). The composition of the adult myocardium is noteworthy when considering its regenerative potential. Despite the fact that approximately 90% of the myocardial mass is composed of CMs, these cells only represent ~ 30% of the total number of cells, the rest being predominantly composed of endothelial cells (~ 43%), fibroblasts (~ 20%) and leukocytes (~ 7%) (Reiss et al. 1996; Pinto et al. 2016). About 1% are CM progenitor cells (or stem cells), and this, coupled with the fact that the majority of adult CMs are quiescent, is thought to account for the fact that the adult human heart has a very minimal regenerative capacity post-injury, with CMs estimated to have an annual renewal rate ranging from 1% in early adulthood (25 years) to as little as 0.45% in later life (75 years) (Murry et al. 2006). Considering the cellular damage post-MI can be in the region of a loss of 25% of cells (Murry et al. 2006), it is clear that normal turnover and replacement operates at a substantial deficit and is insufficient to restore a functional myocardium post-infarction (Bergmann et al. 2009; Senyo et al. 2013). However, if only 30% of the total number of cardiac cells are CMs, it seems restrictive to narrow our focus for regenerative capacity to this single cell type.

Despite the absence of robust regeneration in the adult mammalian heart, cardiac regeneration is observed, to varying degrees, in neonatal mammals. Murine and porcine hearts are capable of efficient regeneration, providing the injury occurs in the first 2 days post-natally, and the magnitude of the injury is not too great (< 15%) (Porrello et al. 2011; Bryant et al. 2015; Notari et al. 2018; Ye et al. 2018; Zhu et al. 2018). Use of an adult zebrafish cryoinjury model also supports this finding that even in highly regenerative systems the potential is not inexhaustible. Injury resulting in up to 20% ventricular cell death (by area) seems to be well tolerated and completely resolved by 60-day post-injury (dpi) (Chablais et al. 2011; Bevan et al. 2020) yet increasing the injury area by 5% seems to push the regenerative capacity of the zebrafish heart beyond its limits, with scar resolution incomplete even at 130 dpi (Gonzalez-Rosa et al. 2011). Additionally, a recent report demonstrates that repeated cryoinjuries limit the regenerative ability of adult zebrafish (Bise et al. 2020). Interestingly, CM proliferation is activated after each injury (although this becomes less efficient over time), but collagen deposition is exacerbated by each injury and scar removal gradually fails, further suggesting that regenerative capacity has limits (Bise et al. 2020).

Zebrafish injury models have also been instrumental in clarifying the source of new heart muscle cells in a regenerative context. Several studies have exploited lineage-tracing methods to show that existing mature CMs are the source of new cardiac muscle and stem/progenitor cells have no significant involvement in this process (Jopling et al. 2010 and reviewed by Kikuchi 2015). Morphologically, the dividing zebrafish CMs change their contractile state in a manner similar to structural alterations observed to facilitate proliferation in murine cells, such as disassembly of sarcomeres, which may indicate conservation of the underlying processes required for CM proliferation (Ahuja et al. 2004; Jopling et al. 2010; and reviewed by Kikuchi 2015).

There is some CM proliferation in adult mammals and this likely arises from existing CMs (as in zebrafish); however, this occurs at a much lower frequency and the signature of the CMs retaining this proliferative ability has not yet been fully determined (Bergmann et al. 2009; Jopling et al. 2010; Senyo et al. 2013; Kikuchi 2015). Evidence has emerged in recent years to suggest that ploidy is an important factor. The majority of zebrafish CMs are mononucleated and diploid; however, these properties are lost in mammalian cells which become binucleated or polyploid soon after birth (Soonpaa et al. 1996; Kikuchi et al. 2010; Mollova et al. 2013; Ye et al. 2016). This hypothetical relationship has been further substantiated as the prevalence of mononuclear diploid CMs has been shown to correlate with functional recovery and CM proliferation after coronary artery ligation in mice and CM ploidy can generally predict regenerative potential across vertebrate species (Patterson et al. 2017; Hirose et al. 2019). Gonzalez-Rosa et al. have recently shown direct evidence of the importance of ploidy, as experimental polyploidisation of zebrafish CMs is sufficient to inhibit proliferative potential in this highly regenerative model, and a substantial proportion of diploid CMs (approximately 75%) is required to support regeneration, leading the authors to propose stimulation of the rare diploid cells in the human heart as a method to boost heart regeneration (González-Rosa et al. 2018).

The importance of CM replacement for complete regeneration to occur is clear, but the value of the interplay between these cells and their microenvironment is just beginning to be elucidated. The study of many different in vivo models of cardiovascular disease and injury is hel** researchers look beyond CMs to the myriad of cells and processes that co-activate, co-ordinate and co-regulate regeneration of the heart. These processes include signalling mechanisms for cell recruitment (reviewed by Sanz-Morejón and Mercader 2020), phagocytic immune cell removal of debris (de Preux Charles et al. 2016; Lai et al. 2017; Bevan et al. 2020), fibroblasts laying down ECM (Simões et al. 2020) and MMP breakdown of ECM to permit the infiltration of new vasculature (Bellayr et al. 2009; Marín-Juez et al. 2016; Xu et al. 2000). Further to this, their genetic tractability has proven an adaptable asset during an extraordinary transitional period in genome editing, from early labour-intensive reverse genetics approaches like ENU-screening through to the now ubiquitous CRISPR-cas9 system (Koster and Sassen 2015; Sertori et al. 2016). As a non-mammalian model, there are genetic discrepancies, with the poor conservation of some alleles and duplication of approximately 20% of genes (Postlethwait 2000) sometimes making direct comparisons difficult. However, 70% of protein-coding human genes are related to genes found in zebrafish, and 84% of disease-related genes have a zebrafish equivalent (Howe et al. 2013) meaning they have become an important model of human disease (reviewed by Lieschke and Currie 2007). Antibody availability has historically been poor (though in recent years the prevalence of custom synthesis services and the popularity of the zebrafish as a model system has led to some improvements on this score), yet amenability to genome-editing has also facilitated the generation of numerous transgenic lines which are both a powerful research tool in their own right and partially compensate for the lack of commercially available antibodies.

In addition to these well-known characteristics, the key attribute for zebrafish in this research context is their near-unique ability to completely regenerate the adult heart post-injury, thus providing a cellular and molecular map for the processes from repair to regeneration (Poss 2002; Chablais et al. 2011; Schnabel et al. 2011; Gonzalez-Rosa et al. 2011; Bevan et al. 2020). As described above, three main models of cardiac injury have been described in adult zebrafish, (1) cardiac resection (Poss 2002), (2) cryoinjury (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011) and (3) genetic ablation of CMs (Wang et al. 2011). The resection model, which involves the surgical removal of the ventricular apex, was used in the first landmark study describing the regenerative ability of the adult zebrafish heart (Poss 2002). First published in 2011, the cryoinjury model arguably provides the most representative model for the ischaemia-induced cell death associated with infarction, resulting in extensive scar formation (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011; Bevan et al. 2020). Although these initial cryoinjury studies report some discrepancies in the “completeness” of the regeneration process, this is likely due to differences in the extent of the injury inflicted (20% ventricular area vs 25%, Chablais et al. 2011; González-Rosa et al. 2011, respectively), and overall, these studies largely correlate on the major events post-injury. Around the same time as the description of the cryoinjury model, it was shown that adult zebrafish could also recover from the loss of up to 60% of their CMs, ablated using a genetic Cre/LoxP-driven cytotoxic diptheria toxin A expression system (Wang et al. 2011). Although this system further highlights the remarkable regenerative capacity of the zebrafish and allows a less invasive method to study CM replacement, it is arguably the least representative of human cardiac damage as it is not localised, resulting in the random loss of CMs throughout the ventricle and, therefore, does not elicit the same targeted fibrotic, scarring and angiogenic responses, which are all important aspects of the repair to regeneration transition.

It is important to note that following cardiac damage, such as MI (humans) or cryoinjury (zebrafish), the initial phases from injury to repair and scarring are conserved, further strengthening the value of the zebrafish as a powerful tool to understand the limiting factors preventing mammalian regeneration (Fig. 1 and reviewed by Giardoglou and Beis 2019). Both human and zebrafish repair involves an initial inflammatory phase (defined by the recruitment of immune cells and the clearance of cellular debris by phagocytosis) followed by a reparative phase characterised by deposition of collagen and other extracellular matrix (ECM) components and scar formation (Dobaczewski et al. 2011; Chablais and Jazwinska 2012). In humans, this collagenous matrix develops to form a mature scar that is never resolved; however, in zebrafish the deposed collagen is rapidly remodelled and replaced with new myocardium (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011; Hortells et al. 2019).

Fig. 1
figure 1

Human and zebrafish heart repair. Phases of repair/regeneration in human and zebrafish hearts post-injury, showing that initial inflammatory and scarring responses are similar, but the final stages diverge, with humans exhibiting persistent scar tissue and poor renewal of CMs. Zebrafish models of MI (e.g. cryoinjury shown here) exhibit a regenerative phase of scar resolution and CM proliferation, terminating in a return to healthy myocardium

Full size image

Cryoinjury results in cell death within the ventricle wall (with apoptotic cells also detectable in the lumen of coronary vessels) which peaks at approximately 4 dpi and decreases progressively to below 0.5% at 60 dpi (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011). This apoptotic peak is concomitant with the initial inflammatory response and the commencement of neovascularisation with existing coronary vessels sprouting into the injury area (González-Rosa et al. 2011; Marín-Juez et al. 2019). Extensive fibrin accumulation in the injury areas is also seen at 4 dpi (Chablais et al. 2011; Schnabel et al. 2011) but is mostly eliminated by 14–21 dpi (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011). Extensive CM (and other cell) proliferation is observed during these initial phases of the injury response, peaking within the first week (Chablais et al. 2011; González-Rosa et al. 2011; Schnabel et al. 2011). By 21 dpi, vessel coverage of the injured area is complete, and this re-vascularisation of the injured area is so rapid that a mere 40 days after injury the vessels are indistinguishable between controls and injured hearts (González-Rosa et al. 2011; Marín-Juez et al. 2019).

The deposition, remodelling and maturation of the collagen-rich scar are the final phase in the mammalian repair process; however, the resolution of this tissue occurs rapidly in the zebrafish, with scar clearance and regeneration of the cardiac tissue being completed within 60–130 dpi depending on the size of the initial injury (Bevan et al. 2020; Schnabel et al. 2011; Chablais et al. 2011; González-Rosa et al. 2011; and reviewed by Dittrich and Lauridsen 2019). Interestingly, comparative analyses between cryoinjury and resection injury responses indicate differences in the degree of apoptosis, inflammation and scarring, further highlighting variation between injuries (Chablais et al. 2011; Simões et al. 2020). This regenerative capacity in an adult in vivo system provides an unparalleled opportunity to study the molecular and cellular processes of regeneration in an intact environment.

Emerging roles for inflammatory populations

There has been a recent shift to focus on inflammation as a crucial aspect of the regeneration process. Zebrafish have many innate and adaptive immune cell populations which are analogous to mammals and can be found in similar ratios within the healthy and injured zebrafish heart (Herbomel et al. 1999; Wittamer et al. 2011; Dee et al. 2016; Bevan et al. 2020). Our group and others have described the systematic recruitment and expansion of innate and adaptive immune cell populations and have begun to uncover their roles in the adult zebrafish heart during the characteristic phases of the repair and regeneration timeline (de Preux Charles et al. 2016; Lai et al. 2017; Hui et al. 2017; Sanz-Morejón et al. 2019; Bevan et al. 2020; Simões et al. 2020) (Fig. 2). This has shown the direct parallels that exist between zebrafish and mammalian cardiac repair, with timely induction and resolution of the inflammatory response being essential to achieve normal regeneration in the zebrafish heart and to regulate many other aspects of repair, including revascularisation, CM proliferation and scar deposition and resolution (Bevan et al. 2020; Simões et al. 2020). Recent observations have also illustrated that without tight regulation of the inflammatory response, excessive immune cell accumulation is inhibitory to regeneration, regardless of the presence of proliferating cells and the ECM (Xu et al. 2019). This demonstrates that inflammatory cells have critical roles in coupling CM proliferation and scar resolution, both of which are central to an efficient regenerative outcome. A thorough characterisation of immune cell types during all the repair and regeneration stages and a full understanding of the differences between zebrafish and mammals will be essential to understanding their diverse roles after injury. Here we introduce the main immune cell types that have been studied in the context of cardiac regeneration.

Fig. 2
figure 2

Immune cell population and ECM dynamics during zebrafish cardiac repair/regeneration. Diagram adapted from data published by Bevan et al. (2020), showing relative waves of immune cell populations and collagen (specifically Collagen I) deposition as detected over a 60-day time period following cardiac cryoinjury. A colour-coded legend, including markers used to define the relative populations, is shown below the graph

Full size image

Granulocytes

Granulocytes are key components of the innate immune system and include neutrophils and eosinophils (reviewed by Lin and Loré 2017) whose roles and dynamics following cardiac injury in the adult zebrafish are beginning to be defined (Lai et al. 2017; Xu et al. 2018; Bevan et al. 2020).

Neutrophils are amongst the “first responders” to tissue damage and are rapidly mobilised and recruited to the cryoinjured heart from 6 hours post injury (hpi) (Bevan et al. 2020). Neutrophil numbers peak during the first 24 hpi (Fig. 2), during which time they are major contributors to the pro-inflammatory phase of repair (Lai et al. 2017; Xu et al. 2018; Bevan et al. 2020). In addition to the phagocytosis of necrotic tissue, recruited neutrophils secrete pro-inflammatory mediators, reactive oxygen species (ROS) and cytokines which recruit additional immune cells and promote myofibroblast differentiation and angiogenesis, which are essential for later stages of regeneration (Lai et al. 2017; Xu et al. 2018; Bevan et al. 2020). However, the release of ROS by neutrophils, as well as infiltrating monocytes, can cause further tissue necrosis and scarring (Bonaventura et al. 2019). Neutrophil retention has been shown to prolong the inflammatory period leading to delayed scar regression and reduced CM proliferation in heart injury models (Robertson et al. 2014; Lai et al. 2017; Xu et al. 2019; Vivien et al. 2019) (summarised in Fig. 3 and Table 1). The importance of all the steps in the cascade should not be underestimated, as each additional piece of the puzzle contributes to rebuilding the heart. The interplay between the physical and molecular processes that orchestrate to regenerate the heart is still incompletely understood; however, the zebrafish presents an invaluable system to study the interplay of these different cells and processes. This regenerative adult in vivo model is amenable to high-throughput genetic editing approaches and live-imaging is unparalleled in its potential to unravel the complexities of the therapeutic holy grail that is human heart regeneration.

Fig. 3
figure 3

Heart repair and regeneration requires a concerted effort by many different cell types. Representation of the numerous cell populations required throughout heart repair and regeneration in the adult zebrafish. Sample publications detailing the roles of these different cell types are listed in Table 1

Full size image
Table 1 Components of the cardiac microenvironment important for regeneration
Full size table

Summary

In this review, we have covered the advantages of the zebrafish model for cardiac regeneration research with a particular focus on the roles of different cell types within the cardiac microenvironment. Many of these roles are being uncovered by studying zebrafish and their remarkable natural regenerative ability. There is no doubt that the recovery of CM number is crucial to the regeneration of the heart and the return to full functional capacity, but studies in zebrafish and other regenerative models are revealing the plethora of signalling and other functions delivered by immune cells, epicardial cells, nerves, fibroblasts and endothelial cells that are crucial to supporting this regenerative outcome. In our minds, future therapeutic approaches will need to incorporate the protection or replacement of these other cell types, as well as CMs, to be effective. Groups studying the zebrafish are leading the way in this investigation into the complex cardiac microenvironment.