Cardiac fibroblasts in heart failure and regeneration

In heart disease patients, myocyte loss or malfunction invariably leads to fibrosis, involving the activation and accumulation of cardiac fibroblasts that deposit large amounts of extracellular matrix. Apart from the vital replacement fibrosis that follows myocardial infarction, ensuring structural integrity of the heart, cardiac fibrosis is largely considered to be maladaptive. Much work has focused on signaling pathways driving the fibrotic response, including TGF-β signaling and biomechanical strain. However, currently there are very limited options for reducing cardiac fibrosis, with most patients suffering from chronic fibrosis. The adult heart has very limited regenerative capacity. However, cardiac regeneration has been reported in humans perinatally, and reproduced experimentally in neonatal mice. Furthermore, model organisms such as the zebrafish are able to fully regenerate their hearts following massive cardiac damage into adulthood. Increasing evidence points to a transient immuno-fibrotic response as being key for cardiac regeneration to occur. The mechanisms at play in this context are changing our views on fibrosis, and could be leveraged to promote beneficial remodeling in heart failure patients. This review summarizes our current knowledge of fibroblast properties associated with the healthy, failing or regenerating heart. Furthermore, we explore how cardiac fibroblast activity could be targeted to assist future therapeutic approaches.

Historically, fibroblast were considered to be a highly numerous, if not the most numerous, cell type (Souders et al., 2009).A recent study relying on robust fibroblast markers placed their number at only 11% of total cells in mouse heart, substantially less than previous estimations (Pinto et al., 2016).Zebrafish heart is composed of a far lower proportion of epicardial cells/fibroblasts, at just 1%-2% of all cells (Rolland et al., 2021).Concerning human heart, estimates vary between 10% and 30% of cells, but approaches such as quantification of immune-labelled fibroblasts is less efficient than in mice (Pinto et al., 2016).Diversity in CF proportions among species likely reflect different requirements for ECM content linked to physiological characteristics such as size and mechanical stress, and likely influences response to injury.
Much of the current knowledge on CF function derives from studies focusing on fibrosis and HF.However, recent studies in the field of cardiac repair and regeneration have changed our view on this process.Indeed, both regenerative and non-regenerative responses include inflammatory and proliferative phases, the latter involving fibroblast proliferation and ECM deposition.However, contrary to HF, regeneration is associated with a transient fibrotic response (Sánchez-Iranzo et al., 2018).Exploiting the signaling driving fibrosis resolution could offer new therapeutic possibilities for regulating fibrosis in HF patients.In this review, we discuss properties of fibroblasts and the fibrotic process in HF and regeneration, highlighting common and context-specific features.
2 Fibroblasts in the healthy heart

Markers
Fibroblasts have been defined as widely distributed mesenchymal cells that produce extracellular matrix constituents, notably Collagen Type I (Souders et al., 2009).However, the relative abundance, origins and biological properties CFs have remained controversial due to a lack of consensus over definitive markers.Several studies using putative markers such as FSP1 or CD90/thy-1, now known to be expressed by other lineages (Moore-Morris et al., 2016), pointed to endothelial-tomesenchymal transition and recruitment of circulating progenitors as major sources of fibroblasts in HF (Zeisberg et al., 2007).More recently, the tyrosine kinase receptor, platelet-derived growth factor receptor alpha polypeptide (PDGFRα) and the bHLH transcription factor, Tcf21, have been widely accepted as a comprehensive markers for fibroblasts in both mice and zebrafish (Smith et al., 2011;Acharya et al., 2012;Sánchez-Iranzo et al., 2018) (see Table 1).Furthermore, the development of single cell/nuclei sequencing has revealed that fibroblasts form a clearly identifiable cell-type and validated the specificity of markers such as PDGFRα for the fibroblast lineage in humans (Litviňuková et al., 2020).
Fibroblast activation has been explored in common murine models of HF, notably coronary artery ligation and transverse aortic constriction, discussed in more detail below.In response to cardiac injury or stress, fibroblasts undergo activation, including fibroblast-to-myofibroblast transition.As well as for morphological reasoning, the term myofibroblasts was coined due to their de novo expression of contractile genes, including ACTA2, which codes for α smooth muscle actin (α-SMA) (Gabbiani et al., 1971).α-SMA expression characterizes myofibroblasts, and allows these cells to physically contribute to scar tissue remodeling, but it is worthwhile noting that it is also used to mark for vascular smooth muscle cells (vSMCs) and is not expressed by all activated fibroblasts (Moore- Morris et al., 2015).Fibroblast activation protein α (FAP), a membrane-bound serine protease, has also been shown to be associated with cardiac myofibroblasts (Tillmanns et al., 2015).Another key marker for myofibroblasts is the matricellular protein periostin, which is associated with a strong fibrotic response after injury (Snider et al., 2009).It is important to note that the use of (secreted) ECM constituents as markers for fibroblasts can lead to mislabeling, which can be overcome at least in animal models by the use of reporter lines (see Table 1).

Origins
During development, the majority of CFs derive from epithelialto mesenchymal transition of the epicardium, as first shown in the avian system by retroviral tagging of proepicardial progenitors in chick embryos (Mikawa and Gourdie, 1996).PDGFR signaling has been shown to play a key role in this process (Smith et al., 2011).Genetic lineage-tracing methods have confirmed that a majority of fibroblasts derive from epicardium but also revealed that a subset of fibroblasts, enriched in the interventricular septum, are derived from embryonic endocardium (Moore-Morris et al., 2014;Han et al., 2023).Importantly, these developmentally-derived populations give rise to the vast majority of fibroblasts in fibrotic conditions (Ali et al., 2014;Moore-Morris et al., 2014;Han et al., 2023)

Fibrosis in HF models
HF is defined as a clinical syndrome resulting from structural and/or functional changes resulting in cardiac dysfunction, that can be associated with preserved (HFpEF) or reduced (HFrEF) ejection fraction (EF) (Ponikowski et al., 2016).Multiple disorders and diseases affecting heart development and physiology lead to HF, which is invariably associated with fibrosis of the perivascular and interstitial areas of the myocardium.Murine HF models have been extensively used to explore the cellular and molecular mechanisms driving fibrosis.

Myocardial infarction
Coronary artery ligation is used to induce acute ischemic injury or chronic myocardial infarction (MI) by temporarily or permanently occluding the left anterior descending (LAD) coronary artery (Xu et al., 2014).This results in myocardial ischemia and the gradual replacement of dead myocardial tissue by an infarct scar.The fibroblasts that form the infarct scar have been shown to derive from resident fibroblasts (Ruiz-Villalba et al., 2015;Moore-Morris et al., 2018), as well as de novo epicardial-EMT (Zhou et al., 2011).Collagen-GFP + ;CD45 + fibrocytes, circulating cells of hematopoietic origin, have also been observed at the epicardial surface of the heart as a result of the surgical procedure (Moore-Morris et al., 2018).During the infarction process, fibroblasts transition between different states.Notably, scar formation results from fibroblast proliferation and α-SMA + myofibroblast transition during the first week after infarction.
Following this, α-SMA expression and collagen production is reduced, as the scar matures (Fu et al., 2018).Importantly, increased left ventricular wall rupture has been linked to reduced cardiac fibroblast activity (Ruiz-Villalba et al., 2020;Murtha et al., 2023).

Transverse aortic constriction
Transverse aortic constriction (TAC) is the most common experimental model used to achieve pressure overload-induced cardiac hypertrophy and HF (Rockman et al., 1991).During the surgical procedure, surgical suture is tied around the transverse aortic arch, resulting in left-sided HF through pressure overload.Initially, depending on the tightness of the aortic ligation, the heart undergoes a hypertrophic compensatory response, that can be associated with a conserved or slightly improved ejection fraction, but ultimately develops signs of HF (Richards et al., 2019).As in the MI model, resident fibroblast populations undergo a proliferative response, expanding in perivascular and interstitial areas (Moore-Morris et al., 2014).Interestingly, resident fibroblasts of endocardial origin have recently been shown to be highly responsive to pressure overload compared with epicardium-derived fibroblasts, and specifically targeting endocardium-derived fibroblasts alleviates pathological remodeling (Han et al., 2023).

Fibrosis in regeneration models
Regeneration refers to the restoration of normal organ architecture and function following injury.In many organs, such as the liver or skin, fibrosis is a transient process that naturally recedes during healing.The adult mammalian heart has a very limited regenerative capacity, and most of our understanding of cardiac regenerative biology has come from animal models such as the zebrafish (Hortells et al., 2019).Characterizing the fibrotic process in these models is currently an intense area of research.

Adult teleost fish heart regeneration
The adult zebrafish heart has remarkable regenerative competence, being able to fully regenerate after enduring up to 20% ventricular amputation, a feat in which the adult mammalian heart cannot compare (Poss et al., 2002).Upon cardiac injury, rather than regenerating, mammalian hearts develops chronic fibrosis, a response which is unfortunately maladaptive as it results in permanently altered cardiac form and function, and ultimately, HF (Talman and Ruskoaho, 2016).Conversely, in the zebrafish, fibrosis represents an integrative part of the regenerative response.In zebrafish cardiac regeneration, it is well accepted that new cardiomyocytes derive from pre-existing cardiomyocytes that undergo dedifferentiation, proliferation and redifferentiation (Jopling et al., 2010).Recently, genetic deletion experiments of collagen-producing cells revealed that fibroblast-mediated fibrosis was essential for cardiac regeneration in this model (Sánchez-Iranzo et al., 2018).It is currently unclear to what extent fibroblasts themselves determine the hearts capacity to regenerate, i.e., whether CFs in myocardium with regenerative capacity possess specific properties as compared to CFs in nonregenerative contexts.Interestingly, analysis of the regenerative response in two populations of Astyanax mexicanus fish has provided novel insight into this question.Indeed, following apical resection, myocyte proliferation was equivalent in both the nonregenerative cave-dwelling and regenerative surface populations.However, the former presented heightened immune and fibrotic responses, suggesting that modulating non-myocyte behavior will be equally as important as promoting myocyte proliferation to achieve regeneration in HF patients (Stockdale et al., 2018).In zebrafish, single cell RNA-sequencing (scRNA-seq) analysis revealed that ɑSMA was not significantly upregulated in CFs during the regenerative response, in contrast to CFs following myocardial infarction (Rolland et al., 2023).Hence, comparative analyses between models may reveal specific features of fibroblasts that are essential for the regenerative response.

Neonatal mouse heart regeneration
Cardiac regeneration following apical resection has been reported to occur un neonatal mice (1 day postnatal), with resected animals displaying normal structure and function by 3week of age (Porrello et al., 2011).This recovery has been associated with cardiomyocyte proliferation, as well as a transient fibrotic response (Porrello et al., 2011), similar to what is observed in adult zebrafish (Poss et al., 2002).In a neonatal MI model, a more robust fibrotic response was observed when MI was induced at a non-regenerative stage (P7) compared to the 1 day time point, notably including increased fibroblast proliferation (Wang Z. et al., 2020).
Fibrotic responses occur in all of these models, but are different in nature.In summary, depending on the context, fibrosis is an essential part of response to insult, as for zebrafish heart regeneration, or a maladaptive process, as in non-ischemic HF patients.Regeneration appears to be dependent on cardiomyocyte proliferation, but also on specific properties of cardiac fibroblasts that support the formation of new myocardium (Figure 1).A better understanding of the signaling regulating fibrosis would likely be beneficial for novel therapeutic approaches aimed at alleviating HF.

Signaling regulating cardiac fibrosis
A myriad of cell-signaling pathways have been implicated in the activation of fibroblasts.Activated fibroblasts, or myofibroblasts, are known to secrete large amounts of ECM components, including collagens and proteoglycans, that accumulate in interstitial and perivascular areas (Souders et al., 2009).Novel therapeutic targeting of these processes is of intense clinical interest due to the paucity of current treatments available to combat cardiac fibrosis.
Transforming growth factor-beta (TGF-β) proteins bind to cellsurface receptors and regulate key cellular processes, notably by activating Smad proteins that translocate to the nucleus and regulate transcription (Derynck and Zhang, 2003).TGF-β signaling is known to promote ECM deposition and fibrosis (Border and Noble, 1994), and TGF-β neutralization has been shown to prevent CF activation and reduce fibrosis (Kuwahara et al., 2002).More recently, conditional deletion of TGF-β receptors and Smads in CFs markedly reduced fibrosis in a pressure overload model (Khalil et al., 2017).However, TGF-β signaling has also been shown to be essential for cardiac regeneration in zebrafish, as chemical inhibition inhibited recovery from cryoinjury (Chablais and Jazwinska, 2012).Furthermore, TGF-β signaling has been shown to drive zebrafish valve regeneration, notably by promoting mesenchymal progenitor proliferation and differentiation (Bensimon-Brito et al., 2020).Hence, studies focusing on this pathway have revealed that TGF-β signaling is responsible for promoting excessive fibroblast activation in the context of heart disease, but is essential for the fibrotic response required in the context of cardiac regeneration.
Wingless/Int (WNT) signaling plays a key role during development, and involves β-catenin dependent and independent, non-canonical, pathways (Logan and Nusse, 2004).Wnt signaling is notably known to regulate cellular polarity and motility.Conditional deletion of β-catenin in CFs lead to reduced matrix deposition by fibroblasts following pressure overload in mice (Xiang et al., 2017).Canonical WNT signaling has been shown to be required for zebrafish heart regeneration by promoting cardiomyocyte proliferation (Bertozzi et al., 2022).Whether Wnt signaling in fibroblasts is required for scar absorption in this context is currently unclear.
Of late, the little studied interleukin (IL)-11 has captured great interest, being proposed as a newer mediator in the pathogenesis of fibrotic diseases (Fung et al., 2022).IL-11 is a multifunctional member of the IL-6 cytokine family and is known to play a pivotal pro-fibrotic role in various inflammatory diseases (Fung et al., 2022).It was demonstrated in primary human fibroblasts that TGF-β1 is a potent stimulator of IL-11 gene transcription, and inhibiting IL-11 signaling in fibroblasts in vivo lead to reduced fibrosis (Schafer et al., 2017).In zebrafish, IL-11/STAT3 signaling has been shown to limit fibrosis and promote cardiac regeneration (Allanki et al., 2021).
Fibroblasts are highly sensitive to mechanical and bio molecular cues, and have been referred to as "sentinel cells" (Smith et al., 1997).Notably, increased stiffness leads to fibroblast-to-myofibroblast transition (van Putten et al., 2016).Knockdown of mechanosensitive channel TRPV4 in CFs has been shown to impair TGF-β1 mediated myofibroblast transition (Adapala et al., 2013).Furthermore, Piezo1 signaling in CFs has been shown to be upstream of pro-inflammatory IL6 secretion (Blythe et al., 2019).Hence, transitions between different fibroblast states are highly dependent on mechanosensing, and further characterizing mechanoreceptor functions in regenerative and non-regenerative contexts may reveal novel therapeutic targets.

Conclusion/perspectives
Fibrosis has long been considered a pathologic process, and the roles of fibroblasts in disease and regeneration contexts were largely unknown.The identification of specific fibroblast markers, the generation of genetic tools and the development of single cell/ nuclei-sequencing have enabled tremendous progress in the characterization of CFs.Moreover, through studies using multiple models of HF and regeneration, it transpires that cardiac fibrosis is a multi-phased process: required in the initial response to insult, but often detrimental in its chronic form.Hence, in the future, regulating rather than inhibiting cardiac fibroblast activity may well be decisive for improving HF patient treatment.

TABLE 1
Major fibroblast markers, genetic tools and antibodies for their identification and lineage tracing in mice and zebrafish.
a ColI and periostin are secreted ECM, constituents for which reporter lines, but not immunostaining, accurately labels fibroblasts.Ab, antibody.VSMCs, vascular smooth muscle cells.
Morris et al., 2014)04) the inner most layer of the heart chambers.During midgestation, endocardial EMT contributes mesenchymal cells to the forming atrioventricular and semilunar valves(de Lange et al., 2004).A subset of fibroblasts produced during formation of the atrioventircular valves migrate invests the ventricular septum and sub-endocardial myocardium, complementing the distribution of the epicardium-derived fibroblasts (Moore-Morris et al., 2014).Lastly, the cardiac neural crest, a heterogeneous cell population originating from the dorsal neural tube, has been reported to provide a small proportion of fibroblasts to the atria and ventricles(Ali et al. (Moore-Morris et al., 2014;Ruiz-Villalba et al., 2015)re of epicardial origin, but smaller subsets are derived from endocardium and neural crest.Similarly, genetic lineage tracing in zebrafish also showed that fibroblasts are essentially of epicardial origin(Sánchez-Iranzo et al., 2018).Fibrosis, the excessive accumulation of fibroblasts, results from the activation and proliferation of these resident fibroblast lineages(Moore-Morris et al., 2014;Ruiz-Villalba et al., 2015).